sync compsitor: Send software reply in swap

sync compsitor: Send software reply in swap

Similar to r367111, send the reply to software draw in swap as well, and
allow PrepareTiles to run in parallel.

BUG=545633
Committed: https://crrev.com/b04b149490d47137bf9831a55ff8b7a15adb2cf7
Cr-Commit-Position: refs/heads/master@{#367201}

[boliu, 2015-12-30 18:22:40]

boliu@chromium.org changed reviewers:
+ hush@chromium.org

[boliu, 2015-12-30 18:22:41]

and for software draw

[hush (inactive), 2015-12-30 21:51:09]

lgtm % some comments

https://codereview.chromium.org/1553723003/diff/1/content/renderer/android/sy...
File content/renderer/android/synchronous_compositor_proxy.cc (right):

https://codereview.chromium.org/1553723003/diff/1/content/renderer/android/sy...
content/renderer/android/synchronous_compositor_proxy.cc:291:
base::AutoReset<IPC::Message*> scoped_hardware_draw_reply(
name scoped_software_draw_reply?

https://codereview.chromium.org/1553723003/diff/1/content/renderer/android/sy...
content/renderer/android/synchronous_compositor_proxy.cc:298:
SendDemandDrawSwReply(false, &empty_frame, reply_message);
there seems to be an inconsistency between the IPC APIs of software draw and
hardware draw.
In hardware draw, we don't send the "success" bit. In software draw, we do.

In fact, we can still infer success bit from the emptiness of the frame in
software draw mode right?

https://codereview.chromium.org/1553723003/diff/1/content/renderer/android/sy...
content/renderer/android/synchronous_compositor_proxy.cc:350:
DCHECK(!(hardware_draw_reply_ && software_draw_reply_));
the above 2 lines can be combined into an xor on the 2 booleans:
!hardware_draw_reply_  != !software_draw_reply_

[boliu, 2015-12-30 22:17:46]

https://codereview.chromium.org/1553723003/diff/1/content/renderer/android/sy...
File content/renderer/android/synchronous_compositor_proxy.cc (right):

https://codereview.chromium.org/1553723003/diff/1/content/renderer/android/sy...
content/renderer/android/synchronous_compositor_proxy.cc:291:
base::AutoReset<IPC::Message*> scoped_hardware_draw_reply(
On 2015/12/30 21:51:09, hush wrote:
> name scoped_software_draw_reply?

Oops. Done

https://codereview.chromium.org/1553723003/diff/1/content/renderer/android/sy...
content/renderer/android/synchronous_compositor_proxy.cc:298:
SendDemandDrawSwReply(false, &empty_frame, reply_message);
On 2015/12/30 21:51:09, hush wrote:
> there seems to be an inconsistency between the IPC APIs of software draw and
> hardware draw.
> In hardware draw, we don't send the "success" bit. In software draw, we do.
> 
> In fact, we can still infer success bit from the emptiness of the frame in
> software draw mode right?

True. Previously the success bit was so that I could write *result = false at
the beginning and not worry about early outs. I guess this patch solved that by
factoring out DoDemandDrawSw.

Don't want to touch IPC messages in this patch though.

https://codereview.chromium.org/1553723003/diff/1/content/renderer/android/sy...
content/renderer/android/synchronous_compositor_proxy.cc:350:
DCHECK(!(hardware_draw_reply_ && software_draw_reply_));
On 2015/12/30 21:51:09, hush wrote:
> the above 2 lines can be combined into an xor on the 2 booleans:
> !hardware_draw_reply_  != !software_draw_reply_

Those aren't booleans though

[boliu, 2015-12-30 22:18:08]

The CQ bit was checked by boliu@chromium.org

[boliu, 2015-12-30 22:18:08]

The patchset sent to the CQ was uploaded after l-g-t-m from hush@chromium.org
Link to the patchset: https://codereview.chromium.org/1553723003/#ps20001
(title: "review")

[commit-bot: I haz the power, 2015-12-30 22:18:12]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1553723003/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1553723003/20001

[commit-bot: I haz the power, 2015-12-30 23:25:47]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2015-12-30 23:27:00]

Description was changed from

==========
sync compsitor: Send software reply in swap

Similar to r367111, send the reply to software draw in swap as well, and
allow PrepareTiles to run in parallel.

BUG=545633
==========

to

==========
sync compsitor: Send software reply in swap

Similar to r367111, send the reply to software draw in swap as well, and
allow PrepareTiles to run in parallel.

BUG=545633

Committed: https://crrev.com/b04b149490d47137bf9831a55ff8b7a15adb2cf7
Cr-Commit-Position: refs/heads/master@{#367201}
==========

[commit-bot: I haz the power, 2015-12-30 23:27:01]

Patchset 2 (id:??) landed as
https://crrev.com/b04b149490d47137bf9831a55ff8b7a15adb2cf7
Cr-Commit-Position: refs/heads/master@{#367201}