Hey, On 7/9/26 13:02, Sebastian Andrzej Siewior wrote: > On 2026-07-09 11:53:52 [+0200], Maarten Lankhorst wrote: >> Hello Sebastian, > Hi Maarten, > >> The test results are available, mostly looks good except 1 spurious >> incomplete. >> >> I tried reproducing it, but was unable to. It seems that pre-existing >> incompletes >> on the specific IGT testcase already happened before the changes. >> >> Do you want to review the patches you didn't author, so I can move forward? >> Even if you are uncomfortable to review the signalers_lock changes, it should >> be possible to have the rest reviewed and merged upstream. > > This looks good to me. The condition in 5/7 is the as in 3/7. Not sure > if this should be merged into a helper or if it should be passed from > the caller. But it seems the information of the calling context is lost. > > Please move forward ;)
Yeah the problem was similar, I believe neither check should exist, instead the caller should give this information like the comment mentions. It's best not to try to change i915 too much though, I can try without both 3/7 and 5/7 now that the plan is to disable some problematic perf events and not run selftests in various conditions that should no longer happen. Kind regards, ~Maarten Lankhorst
