Hey, On 7/6/26 14:20, Sebastian Andrzej Siewior wrote: > On 2026-07-06 13:48:16 [+0200], Maarten Lankhorst wrote: >> All fixes to make the core i915 module (without display) working correctly on >> PREEMPT_RT. >> >> Some fixes are still needed, as the selftests still fail. But otherwise the >> test results were looking good on PREEMPT_RT. >> >> It looks like the patch that added a spinlock caused a regression on >> !PREEMPT_RT, >> and may even have caused similar test failures on PREEMPT_RT, so to fix it >> I removed RCU entirely, and converted to a locked implementation instead. >> >> This time at least, i915 live selftests pass on !PREEMPRT_RT for at least 1 >> machine >> with same config as CI. > > I saw the version counter overflowed and is at 4 here. I appreciate all > the work. May I ask how are things?
I'm fixing the bugs in core i915 with this specific PREEMPT_RT series, and left out the display changes. This showed that the failures in i915's selftests were at least partially related to the PREEMPT_RT changes in the last patch. Hopefully v4 passes CI and if you want, you can review those changes so they can be merged. Looks like sashiko is great at finding concurrency bugs, changing from RCU back to a normal lock is non-trivial. It unfortunately found 1 more in add_signaling_context for a missing !list_empty(&ce->signal_link), but should be safe to review regardless. Kind regards, ~Maarten
