Hi Helen,
On 13/08/24 01:47, Helen Mae Koike Fornazier wrote:
Hi Vignesh,
Thanks for your patch.
---- On Mon, 12 Aug 2024 08:20:28 -0300 Vignesh Raman wrote ---
> We have enabled PROVE_LOCKING (which enables LOCKDEP) in drm-ci.
> This will output warnings when kernel locking errors are encountered
> and will continue executing tests. To detect if lockdep has been
> triggered, check the debug_locks value in /proc/lockdep_stats after
> the tests have run. When debug_locks is 0, it indicates that lockdep
> has detected issues and turned itself off. So check this value and
> exit with an error if lockdep is detected.
Should we exit with an error? Or with a warning? (GitLab-CI supports that).
Well, I guess it is serious enough.
I think we can exit with an error since we check the status at the end
of the tests.
Should we also track on the xfail folder? So we can annotate those errors as well?
Do you mean reporting this error in expectation files?
Did you have an entire pipeline with this? To see if everything is still green?
Yes. https://gitlab.freedesktop.org/vigneshraman/linux/-/jobs/62177711
This is a test branch in which I reverted a fix for the lockdep issue.
We see 'WARNING: bad unlock balance detected!' in logs and pipeline is
still green.
Regards,
Vignesh
Helen
>
> Signed-off-by: Vignesh Raman vignesh.raman@xxxxxxxxxxxxx>
> ---
>
> v1:
> - Pipeline link to show lockdep_stats before and after tests,
> https://gitlab.freedesktop.org/vigneshraman/linux/-/pipelines/1246721
>
> ---
> drivers/gpu/drm/ci/igt_runner.sh | 11 +++++++++++
> 1 file changed, 11 insertions(+)
>
> diff --git a/drivers/gpu/drm/ci/igt_runner.sh b/drivers/gpu/drm/ci/igt_runner.sh
> index f38836ec837c..d2c043cd8c6a 100755
> --- a/drivers/gpu/drm/ci/igt_runner.sh
> +++ b/drivers/gpu/drm/ci/igt_runner.sh
> @@ -85,6 +85,17 @@ deqp-runner junit \
> --limit 50 \
> --template "See https://$CI_PROJECT_ROOT_NAMESPACE.pages.freedesktop.org/-/$CI_PROJECT_NAME/-/jobs/$CI_JOB_ID/artifacts/results/{{testcase}}.xml";
>
> +# Check if /proc/lockdep_stats exists
> +if [ -f /proc/lockdep_stats ]; then
> + # If debug_locks is 0, it indicates lockdep is detected and it turns itself off.
> + debug_locks=$(grep 'debug_locks:' /proc/lockdep_stats | awk '{print $2}')
> + if [ "$debug_locks" -eq 0 ]; then
> + echo "LOCKDEP issue detected. Please check dmesg logs for more information."
> + cat /proc/lockdep_stats
> + ret=1
> + fi
> +fi
> +
> # Store the results also in the simpler format used by the runner in ChromeOS CI
> #sed -r 's/(dmesg-warn|pass)/success/g' /results/results.txt > /results/results_simple.txt
>
> --
> 2.43.0
>
>
