On 16/07/2020 09:47, Chris Wilson wrote:
Quoting Tvrtko Ursulin (2020-07-16 09:41:17)
Right I missed dma_fence_is_signaled calls i915_request_completed.
In this case the remaining question is do we care about wait ioctl
potentially returning before the hypothetical sync fence for the same
Quoting Tvrtko Ursulin (2020-07-16 09:41:17)
> Right I missed dma_fence_is_signaled calls i915_request_completed.
>
> In this case the remaining question is do we care about wait ioctl
> potentially returning before the hypothetical sync fence for the same
> request is signaled? This seems like
On 15/07/2020 15:47, Chris Wilson wrote:
Quoting Chris Wilson (2020-07-15 15:47:15)
Quoting Tvrtko Ursulin (2020-07-15 13:26:23)
On 15/07/2020 13:06, Tvrtko Ursulin wrote:
On 15/07/2020 11:50, Chris Wilson wrote:
Currently, we use i915_request_completed() directly in
i915_request_wait()
Quoting Chris Wilson (2020-07-15 15:47:15)
> Quoting Tvrtko Ursulin (2020-07-15 13:26:23)
> >
> > On 15/07/2020 13:06, Tvrtko Ursulin wrote:
> > >
> > > On 15/07/2020 11:50, Chris Wilson wrote:
> > >> Currently, we use i915_request_completed() directly in
> > >> i915_request_wait() and follow up
Quoting Tvrtko Ursulin (2020-07-15 13:26:23)
>
> On 15/07/2020 13:06, Tvrtko Ursulin wrote:
> >
> > On 15/07/2020 11:50, Chris Wilson wrote:
> >> Currently, we use i915_request_completed() directly in
> >> i915_request_wait() and follow up with a manual invocation of
> >> dma_fence_signal().
On 15/07/2020 13:06, Tvrtko Ursulin wrote:
On 15/07/2020 11:50, Chris Wilson wrote:
Currently, we use i915_request_completed() directly in
i915_request_wait() and follow up with a manual invocation of
dma_fence_signal(). This appears to cause a large number of contentions
on i915_request.lock
On 15/07/2020 11:50, Chris Wilson wrote:
Currently, we use i915_request_completed() directly in
i915_request_wait() and follow up with a manual invocation of
dma_fence_signal(). This appears to cause a large number of contentions
on i915_request.lock as when the process is woken up after the
Currently, we use i915_request_completed() directly in
i915_request_wait() and follow up with a manual invocation of
dma_fence_signal(). This appears to cause a large number of contentions
on i915_request.lock as when the process is woken up after the fence is
signaled by an interrupt, we will