On 04/10/2022 16:13, Ceraolo Spurio, Daniele wrote:
On 10/4/2022 4:14 AM, Tvrtko Ursulin wrote:
On 03/10/2022 13:16, Tvrtko Ursulin wrote:
From: Tvrtko Ursulin tvrtko.ursulin@intel.com
Patch which added graceful exit for non-persistent contexts missed the fact it is not enough to set the exiting flag on a context and let the backend handle it from there.
GuC backend cannot handle it because it runs independently in the firmware and driver might not see the requests ever again. Patch also missed the fact some usages of intel_context_is_banned in the GuC backend needed replacing with newly introduced intel_context_is_schedulable.
Fix the first issue by calling into backend revoke when we know this is the last chance to do it. Fix the second issue by replacing intel_context_is_banned with intel_context_is_schedulable, which should always be safe since latter is a superset of the former.
v2: * Just call ce->ops->revoke unconditionally. (Andrzej)
CI is happy - could I get some acks for the GuC backend changes please?
I think we still need to have a longer conversation on the revoking times, but in the meantime this fixes the immediate concerns, so:
Acked-by: Daniele Ceraolo Spurio daniele.ceraolospurio@intel.com
Thanks, I've pushed it so should unbreak 6.0 via stable.
For follow up work I am okay either with a fixes 20ms timeout (this was enough for users which originally reported it), or go with fully configurable? Latter feels a bit over the top since it would then me a kconfig and sysfs to align with the normal preempt timeout.
Regards,
Tvrtko