cc: Make TileManager task limit configurable

cc: Make TileManager task limit configurable

Currently TileManager will only schedule up to a hardcoded limit of 32
tasks. Instead, make this a layer tree setting.

This will help impl-side painting for layout tests, so that one
ManageTiles + WaitForAllScheduledTasks can guarantee all visible tiles
are ready.

R=vmpstr@chromium.org
BUG=381919
Committed: https://crrev.com/06a2caf2033efcc3c93802a69fd6e32c6bf194e1
Cr-Commit-Position: refs/heads/master@{#300941}

[vmpstr, 2014-10-22 22:39:28]

lgtm

[enne (OOO), 2014-10-22 22:42:39]

reveman: I talked about this a bunch with vmpstr over chat (which is why there's
so little discussion here), but would appreciate your thoughts on this change
too before I land it.

[reveman, 2014-10-23 15:18:18]

On 2014/10/22 22:42:39, enne wrote:
> reveman: I talked about this a bunch with vmpstr over chat (which is why
there's
> so little discussion here), but would appreciate your thoughts on this change
> too before I land it.

A lot of code has been tuned to perform well with 32 as a limit and I'd rather
not change this in the general case. Also not a fan of treating NOW tiles in a
special way here. Could we move kScheduledRasterTasksLimit to LayerTreeSettings
and just set it to INT_MAX for layout tests instead?

[enne (OOO), 2014-10-23 17:33:22]

On 2014/10/23 at 15:18:18, reveman wrote:
> A lot of code has been tuned to perform well with 32 as a limit and I'd rather
not change this in the general case. Also not a fan of treating NOW tiles in a
special way here. Could we move kScheduledRasterTasksLimit to LayerTreeSettings
and just set it to INT_MAX for layout tests instead?

Sure, PTAL.

[enne (OOO), 2014-10-23 17:51:52]

enne@chromium.org changed reviewers:
+ reveman@chromium.org

[reveman, 2014-10-23 19:17:39]

lgtm with nits

https://codereview.chromium.org/672673003/diff/40001/cc/resources/tile_manager.h
File cc/resources/tile_manager.h (right):

https://codereview.chromium.org/672673003/diff/40001/cc/resources/tile_manage...
cc/resources/tile_manager.h:241: size_t scheduled_raster_task_limit_;
nit: const size_t

https://codereview.chromium.org/672673003/diff/40001/cc/test/fake_tile_manage...
File cc/test/fake_tile_manager.cc (right):

https://codereview.chromium.org/672673003/diff/40001/cc/test/fake_tile_manage...
cc/test/fake_tile_manager.cc:72: static_cast<size_t>(-1)) {
nit: std::numeric_limits<size_t>::max() instead?

https://codereview.chromium.org/672673003/diff/40001/cc/test/fake_tile_manage...
cc/test/fake_tile_manager.cc:82: static_cast<size_t>(-1)) {
std::numeric_limits<size_t>::max()?

https://codereview.chromium.org/672673003/diff/40001/cc/trees/layer_tree_sett...
File cc/trees/layer_tree_settings.cc (right):

https://codereview.chromium.org/672673003/diff/40001/cc/trees/layer_tree_sett...
cc/trees/layer_tree_settings.cc:70: scheduled_raster_task_limit(32u),
nit: s/32u/32/ as we're not using 'u' suffix for other fields with size_t type

[enne (OOO), 2014-10-23 19:29:35]

https://codereview.chromium.org/672673003/diff/40001/cc/resources/tile_manager.h
File cc/resources/tile_manager.h (right):

https://codereview.chromium.org/672673003/diff/40001/cc/resources/tile_manage...
cc/resources/tile_manager.h:241: size_t scheduled_raster_task_limit_;
On 2014/10/23 19:17:38, reveman wrote:
> nit: const size_t

Done.

https://codereview.chromium.org/672673003/diff/40001/cc/test/fake_tile_manage...
File cc/test/fake_tile_manager.cc (right):

https://codereview.chromium.org/672673003/diff/40001/cc/test/fake_tile_manage...
cc/test/fake_tile_manager.cc:72: static_cast<size_t>(-1)) {
On 2014/10/23 19:17:39, reveman wrote:
> nit: std::numeric_limits<size_t>::max() instead?

Done.

https://codereview.chromium.org/672673003/diff/40001/cc/test/fake_tile_manage...
cc/test/fake_tile_manager.cc:82: static_cast<size_t>(-1)) {
On 2014/10/23 19:17:38, reveman wrote:
> std::numeric_limits<size_t>::max()?

Done.

https://codereview.chromium.org/672673003/diff/40001/cc/trees/layer_tree_sett...
File cc/trees/layer_tree_settings.cc (right):

https://codereview.chromium.org/672673003/diff/40001/cc/trees/layer_tree_sett...
cc/trees/layer_tree_settings.cc:70: scheduled_raster_task_limit(32u),
On 2014/10/23 19:17:39, reveman wrote:
> nit: s/32u/32/ as we're not using 'u' suffix for other fields with size_t type

Done.

[enne (OOO), 2014-10-23 19:29:47]

The CQ bit was checked by enne@chromium.org

[commit-bot: I haz the power, 2014-10-23 19:30:43]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/672673003/60001

[commit-bot: I haz the power, 2014-10-23 20:34:51]

Committed patchset #4 (id:60001)

[commit-bot: I haz the power, 2014-10-23 20:35:21]

Patchset 4 (id:??) landed as
https://crrev.com/06a2caf2033efcc3c93802a69fd6e32c6bf194e1
Cr-Commit-Position: refs/heads/master@{#300941}