Zero-budget TileManager in resourceless software mode

Zero-budget TileManager in resourceless software mode

In ResourceManager, add Reinitialize which essentially
calls Initialize again. Make sure the default resource
type is updated again to be correct.

In LayerTreeHostImpl, set bytes_limit_when_visible_ to 0
in resourceless software mode. Set back default value in
DeferredInitialize.

In TextureLayerImpl, release the resource on
DidLoseOutputSurface if not using mailboxes. For
mailboxes, the resource cannot be recreated by simply
calling CreateResourceFromTextureMailbox again.

BUG=245935
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=207424

[boliu, 2013-06-14 04:48:33]

Dana, could you give this a first pass. There are still some clean up TODO
comments in LTHI.

https://codereview.chromium.org/16700003/diff/12001/cc/layers/texture_layer_i...
File cc/layers/texture_layer_impl.cc (right):

https://codereview.chromium.org/16700003/diff/12001/cc/layers/texture_layer_i...
cc/layers/texture_layer_impl.cc:134: if (output_surface_lost || !uses_mailbox_)
{
For mailbox case, is it legal to save whether mailbox is active in a instance
var, then call CreateResourceFromTextureMailbox at draw time? Then I think we
wouldn't need to make this distinction here and this patch would become way
smaller.

[boliu, 2013-06-14 20:46:33]

In texture layer, release texture_id resource, but don't release mail box
resource. I think other layers are ok due to restricting DeferredInitialize to
impl_side_painting and solid_color_scrollbars.

Reverted renaming DidLoseOutputSurface. Now majority of change is in LTHI.

[boliu, 2013-06-17 20:08:58]

Dana: friendly ping!

[danakj, 2013-06-19 18:38:12]

LGTM overall, few nits

https://codereview.chromium.org/16700003/diff/20001/cc/layers/texture_layer_i...
File cc/layers/texture_layer_impl.cc (right):

https://codereview.chromium.org/16700003/diff/20001/cc/layers/texture_layer_i...
cc/layers/texture_layer_impl.cc:133: if (external_texture_resource_ &&
!uses_mailbox_) {
I think you could DCHECK that external texture resource is 0 here. I don't think
you can get here in between WillDraw and DidDraw, and I think it'd be wrong if
you did.

https://codereview.chromium.org/16700003/diff/22001/cc/trees/layer_tree_host_...
File cc/trees/layer_tree_host_impl.cc (right):

https://codereview.chromium.org/16700003/diff/22001/cc/trees/layer_tree_host_...
cc/trees/layer_tree_host_impl.cc:1471: if (!visible_ && renderer_)
nit: can we just always call SetVisible() here if (renderer_)? The renderers
have an early out when it doesn't change, rather than relying on their default
value being true.

https://codereview.chromium.org/16700003/diff/22001/cc/trees/layer_tree_host_...
File cc/trees/layer_tree_host_impl.h (right):

https://codereview.chromium.org/16700003/diff/22001/cc/trees/layer_tree_host_...
cc/trees/layer_tree_host_impl.h:402: void EnforceZeroBudget(bool on);
nit: bool zero_budget

https://codereview.chromium.org/16700003/diff/22001/cc/trees/layer_tree_host_...
File cc/trees/layer_tree_host_impl_unittest.cc (right):

https://codereview.chromium.org/16700003/diff/22001/cc/trees/layer_tree_host_...
cc/trees/layer_tree_host_impl_unittest.cc:5875:
host_impl_->InitializeRenderer(output_surface.PassAs<OutputSurface>());
EXPECT_TRUE() ?

https://codereview.chromium.org/16700003/diff/22001/cc/trees/thread_proxy.cc
File cc/trees/thread_proxy.cc (left):

https://codereview.chromium.org/16700003/diff/22001/cc/trees/thread_proxy.cc#...
cc/trees/thread_proxy.cc:613:
DCHECK_GT(layer_tree_host_impl_->memory_allocation_limit_bytes(), 0u);
keep this if (!impl side painting) ?

[boliu, 2013-06-19 19:32:08]

All done.

[commit-bot: I haz the power, 2013-06-19 19:33:04]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/[email protected]/16700003/31001

[commit-bot: I haz the power, 2013-06-20 13:55:23]

Retried try job too often on win_rel for step(s) chrome_frame_tests
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=win_rel&nu...

[commit-bot: I haz the power, 2013-06-20 13:58:28]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/[email protected]/16700003/31001

[commit-bot: I haz the power, 2013-06-20 13:58:57]

Change committed as 207424