cc: Make single-thread renderer compositor always synchronous.

cc: Make single-thread renderer compositor always synchronous.

Single-thread compositing is only used in layout tests which also wants
synchronous compositing. So remove the optionality of synchronous
single-thread in the renderer and just always use the two together when
the thread is disabled there.

Also move the decision to use zero copy in synchronous single threaded
mode out to the browser process and just DCHECK it in cc, so that the
browser can make decisions based on using zero or one copy
authoritively.

R=enne
BUG=489447
Committed: https://crrev.com/16e072dbe4cb807dec7f17340f62a3516b852c17
Cr-Commit-Position: refs/heads/master@{#330860}

[danakj, 2015-05-20 18:56:07]

danakj@chromium.org changed reviewers:
+ piman@chromium.org

[danakj, 2015-05-20 18:56:07]

+piman for content/

[enne (OOO), 2015-05-20 19:05:23]

enne@chromium.org changed reviewers:
- piman@chromium.org

[enne (OOO), 2015-05-20 19:05:24]

lgtm

https://codereview.chromium.org/1149803002/diff/20001/content/renderer/gpu/re...
File content/renderer/gpu/render_widget_compositor.cc (right):

https://codereview.chromium.org/1149803002/diff/20001/content/renderer/gpu/re...
content/renderer/gpu/render_widget_compositor.cc:320:
settings.single_thread_proxy_scheduler = false;
Can you leave a comment here that this means that the two states for the
renderer are multi-threaded+scheduler and single-threaded+synchronous, and that
this flag has no effect for threaded mode?

My worry here is somebody will see this and thinks this implies always using
single thread synchronous?

[danakj, 2015-05-20 19:07:38]

https://codereview.chromium.org/1149803002/diff/20001/content/renderer/gpu/re...
File content/renderer/gpu/render_widget_compositor.cc (right):

https://codereview.chromium.org/1149803002/diff/20001/content/renderer/gpu/re...
content/renderer/gpu/render_widget_compositor.cc:320:
settings.single_thread_proxy_scheduler = false;
On 2015/05/20 19:05:23, enne wrote:
> Can you leave a comment here that this means that the two states for the
> renderer are multi-threaded+scheduler and single-threaded+synchronous, and
that
> this flag has no effect for threaded mode?
> 
> My worry here is somebody will see this and thinks this implies always using
> single thread synchronous?

Done.

[piman, 2015-05-20 19:10:29]

piman@chromium.org changed reviewers:
+ piman@chromium.org

[piman, 2015-05-20 19:10:30]

LGTM, but issue title needs a typo fix (syncronous->synchronous)

[danakj, 2015-05-20 19:12:11]

Fixed, thanks.

[danakj, 2015-05-20 19:12:15]

The CQ bit was checked by danakj@chromium.org

[danakj, 2015-05-20 19:12:15]

The patchset sent to the CQ was uploaded after l-g-t-m from enne@chromium.org
Link to the patchset: https://codereview.chromium.org/1149803002/#ps40001
(title: "syncsinglethread: .")

[commit-bot: I haz the power, 2015-05-20 19:12:52]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1149803002/40001

[reveman, 2015-05-20 19:13:32]

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

[reveman, 2015-05-20 19:13:32]

https://codereview.chromium.org/1149803002/diff/20001/cc/trees/layer_tree_hos...
File cc/trees/layer_tree_host_impl.cc (right):

https://codereview.chromium.org/1149803002/diff/20001/cc/trees/layer_tree_hos...
cc/trees/layer_tree_host_impl.cc:2132: // draw when raster is complete. 
Therefore, it must use one of zero
fyi, tiles are ready to draw when raster is complete in the one-copy case too.
the problem is that one-copy currently rely on being able to communicate with
the compositor thread from worker threads. this problem will go away when
one-copy uses a separate context just like gpu raster. just filed
crbug.com/490295 for that. please add a TODO here and/or in
IsZeroCopyUploadEnabled().

[danakj, 2015-05-20 19:16:30]

The CQ bit was checked by danakj@chromium.org

[danakj, 2015-05-20 19:16:31]

The patchset sent to the CQ was uploaded after l-g-t-m from enne@chromium.org,
piman@chromium.org
Link to the patchset: https://codereview.chromium.org/1149803002/#ps60001
(title: "syncsinglethread: todo")

[danakj, 2015-05-20 19:16:34]

https://codereview.chromium.org/1149803002/diff/20001/cc/trees/layer_tree_hos...
File cc/trees/layer_tree_host_impl.cc (right):

https://codereview.chromium.org/1149803002/diff/20001/cc/trees/layer_tree_hos...
cc/trees/layer_tree_host_impl.cc:2132: // draw when raster is complete. 
Therefore, it must use one of zero
On 2015/05/20 19:13:32, reveman wrote:
> fyi, tiles are ready to draw when raster is complete in the one-copy case too.
> the problem is that one-copy currently rely on being able to communicate with
> the compositor thread from worker threads. this problem will go away when
> one-copy uses a separate context just like gpu raster. just filed
> crbug.com/490295 for that. please add a TODO here and/or in
> IsZeroCopyUploadEnabled().

Done.

[commit-bot: I haz the power, 2015-05-20 19:18:06]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1149803002/60001

[reveman, 2015-05-20 19:26:55]

On 2015/05/20 at 19:16:34, danakj wrote:
>
https://codereview.chromium.org/1149803002/diff/20001/cc/trees/layer_tree_hos...
> File cc/trees/layer_tree_host_impl.cc (right):
> 
>
https://codereview.chromium.org/1149803002/diff/20001/cc/trees/layer_tree_hos...
> cc/trees/layer_tree_host_impl.cc:2132: // draw when raster is complete. 
Therefore, it must use one of zero
> On 2015/05/20 19:13:32, reveman wrote:
> > fyi, tiles are ready to draw when raster is complete in the one-copy case
too.
> > the problem is that one-copy currently rely on being able to communicate
with
> > the compositor thread from worker threads. this problem will go away when
> > one-copy uses a separate context just like gpu raster. just filed
> > crbug.com/490295 for that. please add a TODO here and/or in
> > IsZeroCopyUploadEnabled().
> 
> Done.

thanks!

[commit-bot: I haz the power, 2015-05-20 19:44:06]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2015-05-20 19:44:07]

Try jobs failed on following builders:
  linux_chromium_gn_dbg on tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[danakj, 2015-05-20 20:20:19]

The CQ bit was checked by danakj@chromium.org

[danakj, 2015-05-20 20:20:20]

The patchset sent to the CQ was uploaded after l-g-t-m from enne@chromium.org,
piman@chromium.org
Link to the patchset: https://codereview.chromium.org/1149803002/#ps80001
(title: "syncsinglethread: .")

[commit-bot: I haz the power, 2015-05-20 20:21:45]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1149803002/80001

[commit-bot: I haz the power, 2015-05-20 20:50:24]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2015-05-20 20:50:25]

Try jobs failed on following builders:
  cast_shell_linux on tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/cast_shell_linu...)

[danakj, 2015-05-20 21:14:15]

The CQ bit was checked by danakj@chromium.org

[danakj, 2015-05-20 21:14:16]

The patchset sent to the CQ was uploaded after l-g-t-m from enne@chromium.org,
piman@chromium.org
Link to the patchset: https://codereview.chromium.org/1149803002/#ps100001
(title: "syncsinglethread: zerocopyforunittests")

[commit-bot: I haz the power, 2015-05-20 21:15:09]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1149803002/100001

[commit-bot: I haz the power, 2015-05-20 22:23:07]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2015-05-20 22:23:09]

Try jobs failed on following builders:
  linux_chromium_gn_rel on tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[danakj, 2015-05-20 22:30:56]

The CQ bit was checked by danakj@chromium.org

[danakj, 2015-05-20 22:30:56]

The patchset sent to the CQ was uploaded after l-g-t-m from enne@chromium.org,
piman@chromium.org
Link to the patchset: https://codereview.chromium.org/1149803002/#ps120001
(title: "syncsinglethread: .")

[enne (OOO), 2015-05-20 22:31:10]

o/~ "If you're gonna turn on the scheduler, you'd better use zero copy!"

[enne (OOO), 2015-05-20 22:32:15]

o/~ "If you're gonna turn on the scheduler, you'd better use zero copy!"

http://i.imgur.com/EM3k1Px.gif

[commit-bot: I haz the power, 2015-05-20 22:32:42]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1149803002/120001

[commit-bot: I haz the power, 2015-05-21 00:31:29]

Committed patchset #7 (id:120001)

[commit-bot: I haz the power, 2015-05-21 00:32:22]

Patchset 7 (id:??) landed as
https://crrev.com/16e072dbe4cb807dec7f17340f62a3516b852c17
Cr-Commit-Position: refs/heads/master@{#330860}