By changing the default code generation of GCC we're inadvertently
breaking the GCC test suite, which has ~120K+ more failures when run for
aarch64 compared to x86-64.
This was because the generated code fragments included the BTI
instructions, which the test case wasn't expecting. We can't tell the
tests globally to run without branch protection, as that will break the
tests which also turn it on.
Remove the enabling of branch protection by standard in GCC, we'll
enable it in the tune files instead.
(From OE-Core rev: 81351d984bc6238e9b901ae83c0de6e3d3463c77)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit bb4b9017db6a893ed054a2d2ad4cc671dec09c42)
Signed-off-by: Steve Sakoman <steve@sakoman.com>
Also remove old CVE_CHECK_IGNOREs which are no longer needed due to CPE
updates.
This is a backport from master. Mickledore had one extra CVE patch that
was not on master at the time of upgrade, so it had to be manually
removed here.
(From OE-Core rev: 309b58071d14406ccdf90342f0a33285dc83c87c)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Natasha Bailey <nat.bailey@windriver.com>
Signed-off-by: Steve Sakoman <steve@sakoman.com>
Updating to the latest korg -stable release that comprises
the following commits:
d54cfc420586 Linux 5.15.120
c06edf13f4cf nubus: Partially revert proc_create_single_data() conversion
6e65fa33edf5 parisc: Delete redundant register definitions in <asm/assembly.h>
b4d8f8900021 drm/amdgpu: Validate VM ioctl flags.
26eb191bf5a0 scripts/tags.sh: Resolve gtags empty index generation
989b4a753c7e perf symbols: Symbol lookup with kcore can fail if multiple segments match stext
87f51cf60e3e Revert "thermal/drivers/mediatek: Use devm_of_iomap to avoid resource leak in mtk_thermal_probe"
6a28f3490d3d HID: logitech-hidpp: add HIDPP_QUIRK_DELAYED_INIT for the T651.
67ce7724637c HID: wacom: Use ktime_t rather than int when dealing with timestamps
347732317749 bpf: ensure main program has an extable
d874cf9799a9 can: isotp: isotp_sendmsg(): fix return error fix on TX path
27d03d15bb8b x86/smp: Use dedicated cache-line for mwait_play_dead()
d6c745ca4fc5 x86/microcode/AMD: Load late on both threads too
9052349685e9 drm/amdgpu: Set vmbo destroy after pt bo is created
796481bedc3e mm, hwpoison: when copy-on-write hits poison, take page offline
6713b8f11aa0 mm, hwpoison: try to recover from copy-on write faults
b46021ab8304 mptcp: consolidate fallback and non fallback state machine
42ff95b4bd11 mptcp: fix possible divide by zero in recvmsg()
(From OE-Core rev: 5209c9db1a918aa06c5d37163742efc59bb4503b)
Signed-off-by: Bruce Ashfield <bruce.ashfield@gmail.com>
Signed-off-by: Steve Sakoman <steve@sakoman.com>
Update webkitgtk from 2.38.5 to 2.38.6. It fixes CVE-2023-27932 and
CVE-2023-27954.
What’s new in the WebKitGTK 2.38.6 release?
* Enable the Asynchronous Clipboard API to make certain pages work (e.g.
GithHub started recently requiring it).
* Support :has() CSS selectors in content filters.
* Apply basic font properties as font variation settings.
* The Bubblewrap sandbox no longer requires setting an application
identifier via GApplication to operate correctly. Using GApplication
is still recommended, but optional.
* Improvements to the GStreamer multimedia playback, in particular
around MSE, WebRTC, and seeking.
* Fix the build with journald support enabled when using elogind instead
of the systemd libraries.
* Fix the build with Link-Time Optimization enabled (-flto=auto).
* Fix context menus not working in the remote Web Inspector.
* Fix usage of the remote Web Inspector over HTTP.
* Fix debug logs not being emitted in release builds.
* Fix several crashes and rendering issues.
(From OE-Core rev: 34bd08b121f1653907b45eb8d6192a77ef638005)
Signed-off-by: Kai Kang <kai.kang@windriver.com>
Signed-off-by: Steve Sakoman <steve@sakoman.com>
A vulnerability in the lsi53c895a device affects the latest version
of qemu. A DMA-MMIO reentrancy problem may lead to memory corruption
bugs like stack overflow or use-after-free.
References:
https://nvd.nist.gov/vuln/detail/CVE-2023-0330
Upstream patches:
b987718bbb
(From OE-Core rev: 3be283237a9e473112430bcb614b25b7ff7e64b2)
Signed-off-by: Archana Polampalli <archana.polampalli@windriver.com>
Signed-off-by: Steve Sakoman <steve@sakoman.com>
- Kirkstone is now a 4 year LTS
- Add future Scarthgap LTS
- Update document metadata: license, title
(From yocto-docs rev: 408b722ca330155e770262c7db8494ff897cfdb1)
Signed-off-by: Michael Opdenacker <michael.opdenacker@bootlin.com>
Signed-off-by: Steve Sakoman <steve@sakoman.com>
The intent behind these functions was to dump the system state when issues occured
but it has never really worked as we'd planned. Regular monitoring as the build
runs has largely replaced this as that allows a trend to be seen rather than a spot
value which was never really useful. The code is bitrotting and not functioning
correctly so drop it.
[YOCTO #13872]
RP: Reword commit message
(From OE-Core rev: 8d1bc34cffdd9f054e51db4e880747c79bf834fe)
Signed-off-by: Thomas Roos <throos@amazon.de>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit dea37ba49a236029da73d5cfbfc069bffc38b508)
Signed-off-by: Steve Sakoman <steve@sakoman.com>
Instaed of allways create the directories and removing it at the if they are
not used, we can just do it when there are modules configuration to be created.
So the best thing to do is install the directories only when necessary.
(From OE-Core rev: 455baf41550431c22047fe718c8eaae71924b23f)
Signed-off-by: Jose Quaresma <jose.quaresma@foundries.io>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 71460993f350bca3d5a22115fd5551696f955c9f)
Signed-off-by: Steve Sakoman <steve@sakoman.com>
native and nativesdk classes are special and must be inherited last :
put them at the end of the gathered classes to inherit.
(From OE-Core rev: 6de7a92c4366f8b15ff97e53be91db31348b3922)
Signed-off-by: Yoann Congal <yoann.congal@smile.fr>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit a6614fd800cbe791264aeb102d379ba79bd145c2)
Signed-off-by: Steve Sakoman <steve@sakoman.com>
There are four directories in which scripts can be placed which will
always be run for any interface during certain phases of ifup and ifdown
commands:
/etc/network/if-pre-up.d/
/etc/network/if-up.d/
/etc/network/if-down.d/
/etc/network/if-post-down.d/
Even if there are no scripts in these directories, ifup and ifdown
commands will also search these directories by using run-parts command.
Install these directories to fix the following runtime errors:
$ cat /etc/network/interfaces
auto lo
iface lo inet loopback
$ ifdown lo
ifdown: interface lo not configured
$ ifup lo
run-parts: failed to open directory /etc/network/if-up.d: No such file or directory
ifup: failed to bring up lo
(From OE-Core rev: c248473e2d60cfa67c64aa586b404119dec728ff)
Signed-off-by: Yi Zhao <yi.zhao@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 277bc7ab1fedd81f4df578e544ec381c819a10f9)
Signed-off-by: Steve Sakoman <steve@sakoman.com>
2 issues:
- the .so extension is hard-coded, and therefore the libxcryt package compiled with
meta-darwin is empty, because the dylib files are not contained in FILES_${PN}
- nothing actually produces a file libcrypt-*.so (the symlink file is libcrypt.so, without dash), thus
defining FILES:${PN} manually to contain libcrypt-*.so has no effect.
(From OE-Core rev: 06f8aaf8ee03b07f79e1a17708b5946782e389bd)
Signed-off-by: Etienne Cordonnier <ecordonnier@snap.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 7ed6bfa2428b4f1ba7f09d6e9e67c462ff355153)
Signed-off-by: Steve Sakoman <steve@sakoman.com>
This needs to be done for any item that is linked under rustc,
and not just rust itself. Latest python-cryptography exposes the issue.
(From OE-Core rev: 967d847a9815df43d0c92ca61cc544e1fe5dcc03)
Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit d3811228747590ea06e8d68be4785d45ec9c478f)
Signed-off-by: Steve Sakoman <steve@sakoman.com>
cargo_do_compile runs only if the recipe is built using cargo
as the top level tool. Some recipes hide usage of cargo inside setuptools
(or autoconf) and use do_compile definitions specific to those,
and so the environment isn't properly set up.
This was exposed by latest versions of python3-cryptography.
(From OE-Core rev: a1946efdbec608d47f9e992c1b5cf3c671a204fc)
Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 9f4ff643a028d7f5670d80861f2ce19ca2d90faa)
Signed-off-by: Steve Sakoman <steve@sakoman.com>
There was a race condition during the Makefile execution in
between the assemble and compile targets, only the assemble
target had a dependency on creating the build directory.
If the compile target was executed first, an error was thrown
by bitbake since the build directory did not exist yet:
| Assembler messages:
| Fatal error: can't create build/hello_baremetal_aarch64.o:
No such file or directory
Update the SRCREV to reflect the latest changes serializing
the makefile targets to avoid such race condition from happening.
[YOCTO #15146]
(From OE-Core rev: fe637f795238685f5d8e4e3dbf48235dabb1cc17)
Signed-off-by: Alejandro Enedino Hernandez Samaniego <alejandro@enedino.org>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 9a999af292c7b3dd3bb7c0722cc31624c425f432)
Signed-off-by: Steve Sakoman <steve@sakoman.com>
For the x11 backend package config, drop the redundant dependencies
libxcb and cairo. The former is listed twice in the package config,
while the latter is also listed globally.
For the xwayland package config, add the missing dependencies libxcb
and libxcursor. These dependencies are hidden when the x11 backend is
enabled by default. When the default setting is overridden to disable
the x11 backend, the following error occurs:
```
| Run-time dependency xcursor found: NO (tried pkgconfig and cmake)
|
| ../git/xwayland/meson.build:34:2: ERROR: Problem encountered: xwayland requires xcursor which was not found. Or, you can use '-Dxwayland=false'.
```
(From OE-Core rev: d8d60e0a0e838c332f044ecee89b8ad2acae9567)
Signed-off-by: Tom Hochstein <tom.hochstein@nxp.com>
Signed-off-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 49f4dd391b408a850d13010c90492c621b7912c3)
Signed-off-by: Steve Sakoman <steve@sakoman.com>
Changelog:
===========
* Fix: segmentation fault on filter interpretation in "switch" mode
* Fix: `ip` context is expressed as a base-10 field
* Fix: c99: use __asm__ __volatile__
* Fix: c99: static assert: clang build fails due to multiple typedef
* Fix: Reevaluate LTTNG_UST_TRACEPOINT_DEFINE each time tracepoint.h is included
* Fix: trace events in C++ constructors/destructors
* Fix: trace events in C constructors/destructors
* Fix: use unaligned pointer accesses for lttng_inline_memcpy
(From OE-Core rev: c4bfec4b2f092bd76292ad8a03917514626a3a8f)
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 656470b4b0db579308d218d1ece77bdacd168d14)
Signed-off-by: Steve Sakoman <steve@sakoman.com>
Bugfixes:
==========
- Security fixes for flacparse, dvdspu, and subparse
- d3d11videosink: Fix error on pause and play
- decklink: Correctly handle SDK strings on macOS and free strings
after usage on Linux
- filesink: Fix buffered mode writing of buffer lists and buffers with
multiple memories
- gldownload: handle passthrough without a critical
- h265parse: Fix framerate handling regression
- oggdemux: vp8 fixes
- mp4mux, qtmux, qtdemux: Opus audio mapping fixes
- pngdec: Fix wrong colours output from 16bit RGB images
- ptp clock: Work around ptpd bug in default configuration
- srtpdec: fix critical warnings on shutdown
- v4l2src: fix support for bayer format
- v4l2videoenc: support force-keyframe event in v4l2 encoder
- vtenc: apply DTS offset to ensure DTS <= PTS
- gst-python: allow more functions to be called before gst_init()
- cerbero: fix vaapi variant; add qt6 build on windows; ensure errors
on unguarded use of new APIs, require macOS 10.13
- packages: ship codecalpha, rtponvif, dvbsubenc, switchbin,
videosignal plugins; fix pango crash on 32-bit windows
- various bug fixes, memory leak fixes, and other stability and
reliability improvements
(From OE-Core rev: a071696747ece93409d533a2e966cd59195a056d)
Signed-off-by: Wang Mingyu <wangmy@fujitsu.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit d7bf88db7f2f55540bf7891474a3a28f5681ddbb)
Signed-off-by: Steve Sakoman <steve@sakoman.com>
* Noteworthy changes in release 3.10 (2023-05-21) [stable]
** Bug fixes
cmp/diff can again work with file dates past Y2K38
[bug introduced in 3.9]
diff -D no longer fails to output #ifndef lines.
[bug#61193 introduced in 3.9]
Remove the comment addition from the patch body, as it
increases likelyhood of rebase conflicts, and repeats what
the commit says.
(From OE-Core rev: 21e40166870fadee986fb36be80019d3bcdb69e5)
Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 925155acc6922f7e9df2afa45e79ad1b2c57ba24)
Signed-off-by: Steve Sakoman <steve@sakoman.com>
* CVE-2023-33551.patch is for fix CVE-2023-33551.
* CVE-2023-33552-3.patch is upstream patch for fix CVE-2023-33552, since
the patch context is changed, CVE-2023-33552-1.patch and
CVE-2023-33552-2.patch is backported.
(From OE-Core rev: 77b5c2f68ea0eef45e77936ce3e7caafa7bd3f54)
Signed-off-by: Changqing Li <changqing.li@windriver.com>
Signed-off-by: Steve Sakoman <steve@sakoman.com>
Combine two commits from git.ghostscript.com/ghostpdl.
The first incompletely addresses the CVE. The second
addresses defects in the first.
(From OE-Core rev: 8173939cfa4d0be8c0ffaa45e371945b1f7a16ae)
Signed-off-by: Joe Slater <joe.slater@windriver.com>
Signed-off-by: Steve Sakoman <steve@sakoman.com>
(From OE-Core rev: 11b6e64c07df043441824511c931fb0bc9673adc)
Signed-off-by: Chee Yang Lee <chee.yang.lee@intel.com>
Signed-off-by: Steve Sakoman <steve@sakoman.com>
There were vestigal remains of API key support which could be removed,
but as using an API key - in theory - gives the user larger rate limits
it's probably wise to expose it.
If the user has an API key, then set NVDCVE_API_KEY.
(From OE-Core rev: 64784e90c5ff559f4da6faadb970cc7aff549592)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit a542de684282bfec79f24ae2f1a2027ffde319d8)
Signed-off-by: Steve Sakoman <steve@sakoman.com>
Add a note of what range we're fetching, and use bb.note() instead of
debug() as messages about retrying shouldn't really be considered debug
logging.
(From OE-Core rev: cc82df9c0ff8ba17196b578e0182a5c726cfbbff)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit b64a869b9c5e1d504f1011da16b5c5ff721afbf0)
Signed-off-by: Steve Sakoman <steve@sakoman.com>
Current 503 errors seem to last several seconds.
In most cases there are two errors and third request succeeds.
However sometimes the outage takes more than time needed
for two retries and third one also fails.
Extend retry count from 3 to 5 to improve the probablity
that the fetcher succeeds.
(From OE-Core rev: 5ca193e287c54bcb17f26e82e61e9446c21cd7e4)
Signed-off-by: Peter Marko <peter.marko@siemens.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit f4d118af2360cff7f234102fd5e4b65a6f4146a6)
Signed-off-by: Steve Sakoman <steve@sakoman.com>
Last couple days it is not possible to update NVD DB as servers
are returning lot of errors.
Mostly "HTTP Error 503: Service Unavailable" is observed but
sporadially also some others.
Retrying helps in most cases, so extend retries to all errors.
Additionally add sleep which is recommended by NVD between requests.
These retries are already implemented between successful requests,
but giving servers time between failed ones is important, too.
(From OE-Core rev: 75d2c50a10b8b506096b43a15bca3f06c6734e2f)
Signed-off-by: Peter Marko <peter.marko@siemens.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 88dad8f198baa80af5ab576498f4df6ed639d551)
Signed-off-by: Steve Sakoman <steve@sakoman.com>
After upgrade to soon-to-be-released kirkstone 4.0.11 CVE annotations got broken.
Anything which has only cvssV3 does not resolve properly.
Fix the API fields used to extract it.
i0.0 score is now at level of NVD DB 1.1.
All CVEs with UNKNOWN vector are not present in NVD DB 1.1.
NVD API 1.1:
sqlite> select vector, count(vector) from nvd group by vector;
ADJACENT_NETWORK|4776
LOCAL|32146
NETWORK|167746
PHYSICAL|185
sqlite> select scorev3, count(scorev3) from nvd group by scorev3;
0.0|73331
1.8|7
1.9|3
...
NVD API 2.0 (broken):
sqlite> select vector, count(vector) from nvd group by vector;
ADJACENT_NETWORK|4587
LOCAL|26273
NETWORK|150421
UNKNOWN|24644
sqlite> select scorev3, count(scorev3) from nvd group by scorev3;
0.0|205925
NVD API 2.0 (fixed):
sqlite> select vector, count(vector) from nvd group by vector;
ADJACENT_NETWORK|5090
LOCAL|32322
NETWORK|168004
PHYSICAL|213
UNKNOWN|511
sqlite> select scorev3, count(scorev3) from nvd group by scorev3;
0.0|73841
1.8|7
1.9|3
...
(From OE-Core rev: babf20391305e514ddcd24be261b9b8487c8c767)
Signed-off-by: Peter Marko <peter.marko@siemens.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 61a5857efdcc0f49c69c0deb24fce99007aeef19)
Signed-off-by: Steve Sakoman <steve@sakoman.com>
When requesting updates in a specific range, use the actual current time
and database mtime instead of truncating to midnight, and explicitly set
the timezone to UTC so that NIST don't treat the timestamps as _their_ local
time when they're _our_ local time.
(From OE-Core rev: aa03556732b295fcf4bc2de11e3bc2e2b364580e)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 9aa0ec37f5f74252588d2494a71c71a7d8e68df9)
Signed-off-by: Steve Sakoman <steve@sakoman.com>
Some CVEs, such as CVE-2013-6629, list multiple configurations which are
vulnerable. The current JSON parser only considers the first
configuration.
Instead, consider every configuration. We don't yet handle the AND/OR
logical operators, but this is a step in the right direction.
(From OE-Core rev: a2d50c0fd1e1be869d8786b920f8b428a3292ed1)
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Alexandre Belloni <alexandre.belloni@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit e1bf4f6dd686055fe9a8bdcc3f739eac2807bae0)
Signed-off-by: Steve Sakoman <steve@sakoman.com>
Needs to go to master branch before stable branches.
This reverts commit 7702dc8fc6c9b34647067ffabbc0e24d6109abe7.
(From OE-Core rev: ac4c9f5aa967507d028caa3ee70f3fce580f9a09)
Signed-off-by: Steve Sakoman <steve@sakoman.com>
Remove log_lock.pid which maybe created during do_rootfs. In commit
[dnf: only write the log lock to root for native dnf],
native dnf changed to write log lock to root, and target dnf still
use /var/log, so log_lock.pid need to be removed post do_rootfs.
(From OE-Core rev: 595fbe4c9ad25e52e88d7bcf1d1864fe5ec324a5)
Signed-off-by: Changqing Li <changqing.li@windriver.com>
Signed-off-by: Steve Sakoman <steve@sakoman.com>
From commit 742a1b7124, log_lock.pid is writen to root, but following
file is not changed, which will make it never deleted, and an unexpected
file exist in root dir after boot target.
$ tail -n 1 etc/tmpfiles.d/dnf.conf
r /var/log/log_lock.pid
Besides, root dir may be read-only, so it is better still
keep the log_lock.pid under /var/log, only write the log lock to root
for native dnf for fixing issue mentioned in 742a1b7124
(From OE-Core rev: 207cc7acef40f92b5de306aa6f035258c0eec828)
Signed-off-by: Changqing Li <changqing.li@windriver.com>
Signed-off-by: Steve Sakoman <steve@sakoman.com>
This fix simplifies the code to drop the pthread and mutexs which should
address some of the buffering issues we were seeing in some test case
failures.
(From OE-Core rev: c0ebc132c279d3195aa8ed086d596ed1b000cc2a)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit bffa6d381d7ea0a6aad19d61cd973c82aadef070)
Signed-off-by: Steve Sakoman <steve@sakoman.com>
Currently the code can race as there is a read/write thread handling the stdio but
there is no guarantee that when the process exits, the thread has handled all the
data. This results in output where "END:" isn't actually at the end of the logs
but somewhere in the middle of the output.
Update to a revision with this fix.
(From OE-Core rev: d73b83da26722890dac8ac9bea19890927b9ea3b)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 6608ee551ed7e8864cdeabfb09832621313d3191)
Signed-off-by: Steve Sakoman <steve@sakoman.com>
Pulls in:
utils: Ensure buffers are flushed after child exits
We currently wait for the child to exit but we don't flush the buffers.
This can mean the output ends up out of sync and the END: line isn't at
the end of the logs.
We've recently seen a lot of issues related to this on the autobuilder.
Add in a flush call for all fds to ensure buffers are in sync. This
does appear to improve warnings on the autobuilder now we started detecting
the issue.
(From OE-Core rev: 0c484c5aa054f4a8f7794bba18b1d0d168dbb941)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit f051de5d5da7e9a2f4137013e24589b594ff6d35)
Signed-off-by: Steve Sakoman <steve@sakoman.com>
The first time logrotate runs it reports an error:
error: state file /var/lib/logrotate.status is
world-readable and thus can be locked from other
unprivileged users. Skipping lock acquisition...
This check was added with
1f76a381e2
This error is only reported once as logrotate removes
the world-readable permissions if this happens.
Since logrotate creates this file if it does not exist,
there should be no need to install it in the first place.
(From OE-Core rev: d081071c99a814e43cdd2c3ca92892c10b631692)
Signed-off-by: Jermain Horsman <jermain.horsman@nedap.com>
Signed-off-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
(cherry picked from commit 8169cd2d18f1569e4357f082adbef492710e8c36)
Signed-off-by: Steve Sakoman <steve@sakoman.com>
