On Sat, 2016-11-05 at 00:32 +0200, Imre Deak wrote:
> On Fri, 2016-11-04 at 21:01 +, Chris Wilson wrote:
> > On Fri, Nov 04, 2016 at 10:33:24PM +0200, Imre Deak wrote:
> > > On Thu, 2016-11-03 at 21:14 +, Chris Wilson wrote:
> > > > Where is that guaranteed? I thought we only serialised wit
On Fri, 2016-11-04 at 21:01 +, Chris Wilson wrote:
> On Fri, Nov 04, 2016 at 10:33:24PM +0200, Imre Deak wrote:
> > On Thu, 2016-11-03 at 21:14 +, Chris Wilson wrote:
> > > Where is that guaranteed? I thought we only serialised with the
> > > pm
> > > interrupts. Remember this happens befor
On Fri, Nov 04, 2016 at 10:33:24PM +0200, Imre Deak wrote:
> On Thu, 2016-11-03 at 21:14 +, Chris Wilson wrote:
> > Where is that guaranteed? I thought we only serialised with the pm
> > interrupts. Remember this happens before rpm suspend, since
> > gem_idle_work_handler is responsible for dro
On Thu, 2016-11-03 at 21:14 +, Chris Wilson wrote:
> On Thu, Nov 03, 2016 at 10:57:23PM +0200, Imre Deak wrote:
> > On Thu, 2016-11-03 at 18:59 +, Chris Wilson wrote:
> > > On Thu, Nov 03, 2016 at 06:19:37PM +0200, Imre Deak wrote:
> > > > We assume that the GPU is idle once receiving the s
On Thu, Nov 03, 2016 at 10:57:23PM +0200, Imre Deak wrote:
> On Thu, 2016-11-03 at 18:59 +, Chris Wilson wrote:
> > On Thu, Nov 03, 2016 at 06:19:37PM +0200, Imre Deak wrote:
> > > We assume that the GPU is idle once receiving the seqno via the last
> > > request's user interrupt. In execlist m
On Thu, 2016-11-03 at 18:59 +, Chris Wilson wrote:
> On Thu, Nov 03, 2016 at 06:19:37PM +0200, Imre Deak wrote:
> > We assume that the GPU is idle once receiving the seqno via the last
> > request's user interrupt. In execlist mode the corresponding context
> > completed interrupt can be delaye
On Thu, Nov 03, 2016 at 06:19:37PM +0200, Imre Deak wrote:
> We assume that the GPU is idle once receiving the seqno via the last
> request's user interrupt. In execlist mode the corresponding context
> completed interrupt can be delayed though and until this latter
> interrupt arrives we consider
On 03/11/2016 16:19, Imre Deak wrote:
We assume that the GPU is idle once receiving the seqno via the last
request's user interrupt. In execlist mode the corresponding context
completed interrupt can be delayed though and until this latter
interrupt arrives we consider the request to be pending
We assume that the GPU is idle once receiving the seqno via the last
request's user interrupt. In execlist mode the corresponding context
completed interrupt can be delayed though and until this latter
interrupt arrives we consider the request to be pending on the ELSP
submit port. This can cause a