Re: [PATCH 2/5] drm/i915/selftests: Flush the active callbacks

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Quoting Chris Wilson (2019-11-22 13:13:54)
> Quoting Chris Wilson (2019-11-22 13:09:00)
> > We complete the sync before the interrupt handler's irq work has
> > finished executing the callback to mark the barrier as completed.
> > So when we look at whether the engine is parked or not, we see the state
> > before the request has finished processing and find it still active.
> 
> Note to self, it does not belong inside i915_request_retire() as we
> take a reference to the request in the breadcrumb handler. I have asked
> myself that more than once and I keep forgetting the answer.

Plus there's that whole serialisation dance we do with the irq handler
to make sure we're on the right side of the reference barrier.
-Chris
_______________________________________________
Intel-gfx mailing list
Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/intel-gfx




[Index of Archives]     [AMD Graphics]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite News]     [Linux Kernel]     [Linux SCSI]

  Powered by Linux