cc: Make property trees support resourceless draw mode

cc: Make property trees support resourceless draw mode

Property trees currently rely on render target information
computed at tree-building time (on the main thread) when
computing draw properties. However, in resourceless draw
mode (used by Android WebView), only the root render surface
is used. Since resourceless draw mode can be arbitrarily
enabled and disabled on the compositor thread, we cannot
know at tree building time whether it will be used. This
CL makes property trees support resourceless draw mode
without requiring trees to be rebuilt when the draw mode
changes (though they do still need to be updated).

Most of the changes are for computing clipping-related
properties, since render surfaces change the way clips are
propagated to descendants and also change the space in
which drawable_content_rect and clip_rect are expressed.

BUG=524106
CQ_INCLUDE_TRYBOTS=tryserver.blink:linux_blink_rel

Committed: https://crrev.com/51d73f712c256fa9e6e5bb320eac2ec7ba7b92a3
Cr-Commit-Position: refs/heads/master@{#354831}

[ajuma, 2015-10-16 18:37:06]

ajuma@chromium.org changed reviewers:
+ jaydasika@chromium.org

[jaydasika, 2015-10-16 21:27:10]

https://codereview.chromium.org/1387003002/diff/140001/cc/trees/property_tree...
File cc/trees/property_tree_builder.cc (right):

https://codereview.chromium.org/1387003002/diff/140001/cc/trees/property_tree...
cc/trees/property_tree_builder.cc:159: if (!requires_node) {
In the resourceless draw mode case, can we not just create clip nodes when
layer_clips_subtree = true as we don't have to bother about render surfaces?

[ajuma, 2015-10-16 22:40:41]

https://codereview.chromium.org/1387003002/diff/140001/cc/trees/property_tree...
File cc/trees/property_tree_builder.cc (right):

https://codereview.chromium.org/1387003002/diff/140001/cc/trees/property_tree...
cc/trees/property_tree_builder.cc:159: if (!requires_node) {
On 2015/10/16 21:27:09, jaydasika wrote:
> In the resourceless draw mode case, can we not just create clip nodes when
> layer_clips_subtree = true as we don't have to bother about render surfaces?

It's true that these are the only clips we have to apply in resourceless mode.
But when we're building trees, we need trees that can support both modes without
requiring a rebuild when the mode changes (since the mode can change arbitrarily
on the compositor thread). So, for example, a single set of trees built on the
main thread and then committed to the compositor thread might be used for a
resourceless draw, then a regular draw, and then another resourceless draw.

[jaydasika, 2015-10-16 23:13:57]

lgtm

Thanks for all the comments.

https://codereview.chromium.org/1387003002/diff/140001/cc/trees/draw_property...
File cc/trees/draw_property_utils.cc (right):

https://codereview.chromium.org/1387003002/diff/140001/cc/trees/draw_property...
cc/trees/draw_property_utils.cc:447: parent_clip_node->id == 0 ||
This will we a no-op after this lands:
htts://crrev.com/1409853002
(Because, we will be iterating from 1 and when id=1, we don't reach here). You
can remove it in a later CL, just wanted you make a note.

https://codereview.chromium.org/1387003002/diff/140001/cc/trees/draw_property...
cc/trees/draw_property_utils.cc:553: : false;
Same as above, this can be removed.

[ajuma, 2015-10-19 17:46:20]

https://codereview.chromium.org/1387003002/diff/140001/cc/trees/draw_property...
File cc/trees/draw_property_utils.cc (right):

https://codereview.chromium.org/1387003002/diff/140001/cc/trees/draw_property...
cc/trees/draw_property_utils.cc:447: parent_clip_node->id == 0 ||
On 2015/10/16 23:13:57, jaydasika wrote:
> This will we a no-op after this lands:
> htts://crrev.com/1409853002
> (Because, we will be iterating from 1 and when id=1, we don't reach here). You
> can remove it in a later CL, just wanted you make a note.

Done.

https://codereview.chromium.org/1387003002/diff/140001/cc/trees/draw_property...
cc/trees/draw_property_utils.cc:553: : false;
On 2015/10/16 23:13:57, jaydasika wrote:
> Same as above, this can be removed.

Done.

[ajuma, 2015-10-19 17:46:25]

The CQ bit was checked by ajuma@chromium.org

[ajuma, 2015-10-19 17:46:25]

The patchset sent to the CQ was uploaded after l-g-t-m from
jaydasika@chromium.org
Link to the patchset: https://codereview.chromium.org/1387003002/#ps160001
(title: "Rebased")

[commit-bot: I haz the power, 2015-10-19 17:47:58]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1387003002/160001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1387003002/160001

[commit-bot: I haz the power, 2015-10-19 19:44:05]

Committed patchset #9 (id:160001)

[commit-bot: I haz the power, 2015-10-19 19:45:19]

Patchset 9 (id:??) landed as
https://crrev.com/51d73f712c256fa9e6e5bb320eac2ec7ba7b92a3
Cr-Commit-Position: refs/heads/master@{#354831}