aura: Handle released widget when GPU channel gets established

aura: Handle released widget when GPU channel gets established

Allow this sequence:
GpuProcessTransporFactory::CreateOutputSurface()
  EstablishGpuChannel()
Compositor::ReleaseAcceleratedWidget()
GpuProcessTransportFactory::EstablishedGpuChannel()

BUG=513540
Committed: https://crrev.com/aa1dc3494905b81e33c1c7df2f09fdc492589d79
Cr-Commit-Position: refs/heads/master@{#355121}

[no sievers, 2015-10-16 20:17:18]

sievers@chromium.org changed reviewers:
+ piman@chromium.org

[no sievers, 2015-10-16 20:18:21]

I think I missed this corner case. Is that ok to remove the DCHECK() here and
return?

Compositor already tracks |output_surface_requested_|, so the next
SetAcceleratedWidget() would cause CreateOutputSurface() again.

[piman, 2015-10-20 01:51:49]

https://codereview.chromium.org/1406993003/diff/1/content/browser/compositor/...
File content/browser/compositor/gpu_process_transport_factory.cc (right):

https://codereview.chromium.org/1406993003/diff/1/content/browser/compositor/...
content/browser/compositor/gpu_process_transport_factory.cc:233:
PerCompositorData* data = per_compositor_data_[compositor.get()];
Can you use per_compositor_data_.find(compositor.get()), so that you avoid
adding a NULL entry that nothing will clear later (logical leak)?

[no sievers, 2015-10-20 17:01:49]

Patchset #2 (id:20001) has been deleted

[no sievers, 2015-10-20 17:16:14]

On 2015/10/20 01:51:49, piman (slow to review) wrote:
>
https://codereview.chromium.org/1406993003/diff/1/content/browser/compositor/...
> File content/browser/compositor/gpu_process_transport_factory.cc (right):
> 
>
https://codereview.chromium.org/1406993003/diff/1/content/browser/compositor/...
> content/browser/compositor/gpu_process_transport_factory.cc:233:
> PerCompositorData* data = per_compositor_data_[compositor.get()];
> Can you use per_compositor_data_.find(compositor.get()), so that you avoid
> adding a NULL entry that nothing will clear later (logical leak)?

oops, my bad. done.

[piman, 2015-10-20 17:43:20]

lgtm

[no sievers, 2015-10-20 18:01:44]

The CQ bit was checked by sievers@chromium.org

[commit-bot: I haz the power, 2015-10-20 18:04:41]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1406993003/40001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1406993003/40001

[commit-bot: I haz the power, 2015-10-20 18:55:52]

Committed patchset #2 (id:40001)

[commit-bot: I haz the power, 2015-10-20 18:56:56]

Patchset 2 (id:??) landed as
https://crrev.com/aa1dc3494905b81e33c1c7df2f09fdc492589d79
Cr-Commit-Position: refs/heads/master@{#355121}