cc: Avoid scheduling empty raster task queue whenever possible.

cc: Avoid scheduling empty raster task queue whenever possible.

Track if raster worker pool instances have pending tasks and
avoid scheduling of empty task queue unless raster worker pool
has pending tasks.

BUG=269841
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=252280

[vmpstr, 2014-02-19 23:04:40]

https://codereview.chromium.org/172783002/diff/30001/cc/resources/raster_work...
File cc/resources/raster_worker_pool_delegate.cc (right):

https://codereview.chromium.org/172783002/diff/30001/cc/resources/raster_work...
cc/resources/raster_worker_pool_delegate.cc:18:
has_pending_tasks_(raster_worker_pools_.size(), false),
I'd prefer "num_raster_worker_pools" here to make sure that accidental
reordering of member variables doesn't break this.

https://codereview.chromium.org/172783002/diff/30001/cc/resources/raster_work...
cc/resources/raster_worker_pool_delegate.cc:56: has_pending_tasks_[i] = true;
Should this be has_pending_tasks_[i] = !raster_queue[i].items.empty()? That
would avoid scheduling empty raster tasks multiple times before we have a chance
to process the raster completion. (and assign call below would not be needed)

[vmpstr, 2014-02-19 23:05:24]

https://codereview.chromium.org/172783002/diff/30001/cc/resources/raster_work...
File cc/resources/raster_worker_pool_delegate.cc (right):

https://codereview.chromium.org/172783002/diff/30001/cc/resources/raster_work...
cc/resources/raster_worker_pool_delegate.cc:56: has_pending_tasks_[i] = true;
On 2014/02/19 23:04:40, vmpstr wrote:
> Should this be has_pending_tasks_[i] = !raster_queue[i].items.empty()? That
> would avoid scheduling empty raster tasks multiple times before we have a
chance
> to process the raster completion. (and assign call below would not be needed)

Actually you still would need the assign, nevermind.

[reveman, 2014-02-20 00:11:53]

Renamed has_pending_tasks_ to did_finish_running_tasks_pending_ as discussed.

https://codereview.chromium.org/172783002/diff/30001/cc/resources/raster_work...
File cc/resources/raster_worker_pool_delegate.cc (right):

https://codereview.chromium.org/172783002/diff/30001/cc/resources/raster_work...
cc/resources/raster_worker_pool_delegate.cc:18:
has_pending_tasks_(raster_worker_pools_.size(), false),
On 2014/02/19 23:04:40, vmpstr wrote:
> I'd prefer "num_raster_worker_pools" here to make sure that accidental
> reordering of member variables doesn't break this.

Done.

https://codereview.chromium.org/172783002/diff/90001/cc/resources/raster_work...
File cc/resources/raster_worker_pool_delegate.cc (right):

https://codereview.chromium.org/172783002/diff/90001/cc/resources/raster_work...
cc/resources/raster_worker_pool_delegate.cc:125:
!did_finish_running_tasks_pending_count_) {
Moved this here as there's no guarantee that required-for-activation arrives
first.

[vmpstr, 2014-02-20 01:03:44]

lgtm

[reveman, 2014-02-20 01:05:07]

The CQ bit was checked by reveman@chromium.org

[commit-bot: I haz the power, 2014-02-20 01:46:38]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/reveman@chromium.org/172783002/90001

[commit-bot: I haz the power, 2014-02-20 05:17:10]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/reveman@chromium.org/172783002/90001

[commit-bot: I haz the power, 2014-02-20 09:17:05]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/reveman@chromium.org/172783002/90001

[commit-bot: I haz the power, 2014-02-20 12:33:27]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/reveman@chromium.org/172783002/90001

[commit-bot: I haz the power, 2014-02-20 17:06:37]

Change committed as 252280