Minor typo observed when cheking the "Prepared shared repository" step logs in
autobuilder web interface:
====================================================================================================
Intially fetching repo poky (1675810261.1)
====================================================================================================
Signed-off-by: Alexis Lothoré <alexis.lothore@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Current regression reports do not contain information about versions compared
when generating reports. While it is still possible to get the information by
searching the autobuilder log, it is not convenient. Moreover, future
developments will allow to generate multiple reports (with different bases for
comparison) in a single build.
As a consequence, embed target and base revisions in the report header
Signed-off-by: Alexis Lothoré <alexis.lothore@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Since we are now force-fetching base revisions and target revisions for
regression report generation, we can make testresults clone even more "shallow"
to increase clone speed in CI pipelines
Signed-off-by: Alexis Lothoré <alexis.lothore@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
If we try to run send-qa-email to simulate past releases (for example, for
development or debugging), the execution will very likely fail because the
target revision to examine (ie: the poky revision) is too old, and as a
consequence is not contained in the testresults shallow clone anymore (because
testsresults history keeps moving forward as builds are triggered on
autobuilder). As a consequence, force-fetch the "target" revision, as it is
already done for the "base" revision
Signed-off-by: Alexis Lothoré <alexis.lothore@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
For the month of January 2023, the distribution of dd times has a long
tail that extends to 13 seconds with 2 events exceeding the current
limit of 30 seconds.
Reduce the timeout to 3 seconds based on the observed distribution of
dd times, which would result in the timout triggering about 20 times a month.
That's enough data to be useful but not so much that it's overwhelming the
logging or the people who will analyze it. It also avoids the rapid increase
in the tail of the distribution which starts to rise exponentially under 2 seconds.
It's sensible response time for people to expect the system to have.
Signed-off-by: Randy MacLeod <Randy.MacLeod@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The "previous version" computation bring many edge cases depending on the
version under release. Add a basic test suite to validate currently implemented
computation strategy and to prevent mistakes when ediiting it in the future
Signed-off-by: Alexis Lothoré <alexis.lothore@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Instead of only generating regressions reports against HEAD of relevant branch, compute
most relevant tag (ie : release) against which we can check for regressions. General rules
introduced are the following :
- milestone release is checked against previous milestone if possible, otherwise
against major release
- point release is checked against previous point release if possible,
otherwise against major release
- major release is checked against previous major release
- a non release build is checked against base branch
Examples :
- 4.1.2.rc1 is checked against yocto-4.1.1
- 4.1.2 is checked against yocto-4.1.1
- 4.1.1.rc1 is checked against yocto-4.1
- 4.1.1 is checked against yocto-4.1
- 4.1 is checked against yocto-4.0
- 4.1.rc4 is checked against yocto-4.0
- 4.1_M2.rc1 is checked against 4.1_M1
- 4.1_M2 is checked against 4.1_M1
- 4.1_M1.rc1 is checked against yocto-4.0
- 4.1_M1 is checked against yocto-4.0
Signed-off-by: Alexis Lothoré <alexis.lothore@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This wrapping allows importing send_qa_email.py in another script without
executing the QA emailing routine. This is useful for unit testing the script
internal functions
Signed-off-by: Alexis Lothoré <alexis.lothore@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Merely renaming send-qa-email to send_qa_email.py would imply to update
yocto-autobuilder2 scripts, which can be daunting. In order to preserve
compatibility with current configuration, add a symlink pointing to the
new script
Signed-off-by: Alexis Lothoré <alexis.lothore@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This rename fixes send-qa-email regarding python guidelines, which allows to
import the script content in another script (for unit testing purpose for
example)
Signed-off-by: Alexis Lothoré <alexis.lothore@bootlin.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The pub/non-release directories have aged out and no longer work.
Switch to a stable location and bump the version to 4.1.2 as well.
Signed-off-by: Michael Halstead <mhalstead@linuxfoundation.org>
Currently we use 56 or 24 cores on the various workers but this is
pulling machines over with OOM issues, particularly in oe-selftest.
Lower the parallelism for parsing to try and help avoid this as we
don't really need maximum speed for parsing on the autobuilder.
Keep the same numbers for the perf workers as we want to use all
cores there.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This will be particularly useful for testing Y2038 issues,
as that is the only 32 bit target that can be run at native speed
on the AB cluster.
Whether a-full should get fast ptests or full ptests can be decided
after collecting AB execution times for both.
Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Bitbake is going to upgrade the minimum python to 3.8, so Ubuntu 18.04
needs to use buildtools tarball to remain compatible
Signed-off-by: Joshua Watt <JPEWhacker@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Many of the oe-selftest test cases boot up an image in qemu, but if the
host can't use KVM to accelerate the guest then the tests take a lot
longer and may randomly timeout.
Now that oe-selftest can run for qemuarm64[1], switch the
oe-selftest-armhost to use qemuarm64 so that KVM can be used. This
should mean faster and more reliable tests, which also cover another
architecture.
[1] oe-core 87fcee9fed915ff3af528b976d7d44737f9f1a95
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
It happened that the git repositories were dirty and resulted in
incorrect files being used. Let's use git clean -ffdx to force a
completely clean git repositories before and after checking out a branch
so that nothing is left from or to another branch build. Also git reset
--hard to undo changes to tracked files.
Cc: Quentin Schulz <foss+yocto@0leil.net>
Reported-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com> #https://lore.kernel.org/yocto-docs/e50abe3c777e4a23a752a3ec25ad0b2a@axis.com/
Signed-off-by: Quentin Schulz <quentin.schulz@theobroma-systems.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
We disabled meta-gplv2 for master, ensure the repo isn't used in the
collection or QA scripts.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
With the recent template config changes to oe-core its content is no longer
backwards compatible.
'-f' is to guarantee future compatibility with build setups where the file
is not present.
Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Archive the buildstats of every failed build and 1% of random builds. Convert
the time-stamped buildstats directory to a compressed tarball using the
hostname as a prefix (to the file name only) so that one can identify the
source machine. Move these tarballs to the directory:
testresults/<build_name>/buildstats/
The archiving is performed during the "collect results" step.
Signed-off-by: Aryaman Gupta <aryaman.gupta@windriver.com>
Signed-off-by: Randy MacLeod <Randy.MacLeod@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Prevent severe system overload by setting the pressure limits for CPU
and IO at 10000.
Signed-off-by: Aryaman Gupta <aryaman.gupta@windriver.com>
Signed-off-by: Randy Macleod <Randy.Macleod@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Build and test core-image-full-cmdline with the linux-yocto-rt kernel,
adding the new rt test to verify that the kernel is actually the
PREEMPT_RT version expected.
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Master branch of the yocto-docs now autogenerates the releases.rst file
and the file therefore does not exist anymore.
Signed-off-by: Quentin Schulz <foss+yocto@0leil.net>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
`git checkout <pathspec>` without <tree-ish> only undo the changes made
to pathspec compared to HEAD. It is not guaranteed that HEAD is on
latest commit of master branch from upstream when running this script,
therefore let's add the <tree-ish> argument to specify which branch to
use to make sure set_versions.py is the latest, otherwise getlatest
might return the wrong information.
Cc: Quentin Schulz <foss+yocto@0leil.net>
Signed-off-by: Quentin Schulz <quentin.schulz@theobroma-systems.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
A dedicated domain for hashserv allows us to move it to dedicated
infrastructre.
Signed-off-by: Michael Halstead <mhalstead@linuxfoundation.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
meta-gplv2 is deprecated, so remove it from the autobuilder in master.
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The meta-gplv2 layer is deprecated in master, so remove it from the
check-layer job.
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Although bitbake.conf sets the default SDKMACHINE to the build
architecture, config.json resets that to i686.
As oe-selftest assumes that the SDKs it builds are usable on the host
machine, we should set SDKMACHINE=aarch64 in the oe-selftest-armhost
build.
A follow-up more invasive patch to clean up the SDKMACHINE assignments
is in progress, once it has been verified to not cause regressions.
Signed-off-by: Ross Burton <ross.burton@arm.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
We could do with wider exposure of testing of this so enable on the
autobuilder by default.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
We may not have fetched the master revision if a different one was being
built so using origin/master works consistently in all cases.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
