Hi, I am using 5.10.41 with RT patchsets and consistently seeing GPU hangs as shown below. ----- i915 0000:00:02.0: [drm] Resetting rcs0 for preemption time out i915 0000:00:02.0: [drm] abcd_plannings[13636] context reset due to GPU hang i915 0000:00:02.0: [drm] GPU HANG: ecode 9:1:85dffffb, in abcd_plannings [13636] ----- The GPU hangs only when using the PREEMPT_RT. Without PREEMPT_RT, the driver is stable. Raised an issue with intel. https://gitlab.freedesktop.org/drm/intel/-/issues/3625 The response to the issues raised for drm hangs seems to take forever. So, I want to see if i can do something to figure out the issue. Checking from the past history, seeing commits like 1dfffa0051eae890ce36924651ecff60df5d779e, i see that some adaptations are necessary for DRM driver to work as desired on PREEMPT_RT. I am hoping something like that is needed to fix this issue too. Any directions on how I can debug this further? I have a kdump solution in place. Currently I am planning to trigger a panic right after a hang is detected and see from the dump if there is any clue as to what the real issue could be. Any pointers/probable fixes/suggestions will be very helpful! Thank you -Manty
