Enable async upload egl with mailbox synchronizer

Enable async upload egl with mailbox synchronizer

With the workaround in r286800, mailbox synchronizer skips tile
textures, so is compatible with async transfer egl.

This is a partial revert of r286501.

BUG=399311
Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=289172

[boliu, 2014-08-11 15:44:30]

PTAL.

Daniel is on vacation so sending you all the webview gpu related changes.
Thanks!

[boliu, 2014-08-12 15:30:12]

Shopping for owners who are not away at siggraph.

vmiura, would you feel comfortable reviewing this? This partially reverts
https://codereview.chromium.org/413103002/

[vmiura, 2014-08-12 21:39:16]

On 2014/08/12 15:30:12, boliu wrote:
> Shopping for owners who are not away at siggraph.
> 
> vmiura, would you feel comfortable reviewing this? This partially reverts
> https://codereview.chromium.org/413103002/

In,
https://codereview.chromium.org/429983002/diff/20001/gpu/command_buffer/servi...
there is a TODO to remove this.

Is it OK to depend on this change as far as you know?

[boliu, 2014-08-12 21:58:38]

On 2014/08/12 21:39:16, vmiura wrote:
> On 2014/08/12 15:30:12, boliu wrote:
> > Shopping for owners who are not away at siggraph.
> > 
> > vmiura, would you feel comfortable reviewing this? This partially reverts
> > https://codereview.chromium.org/413103002/
> 
> In,
>
https://codereview.chromium.org/429983002/diff/20001/gpu/command_buffer/servi...
> there is a TODO to remove this.
> 
> Is it OK to depend on this change as far as you know?

It is ok in the sense that m37 webview is already depending on r286800 to avoid
too many fences.

I think proper fix is going to be quite involved, need to re-architect how
webview runs GL, and switch to using per frame fences rather than per mailbox.

[vmiura, 2014-08-12 23:13:18]

On 2014/08/12 21:58:38, boliu wrote:
> On 2014/08/12 21:39:16, vmiura wrote:
> > On 2014/08/12 15:30:12, boliu wrote:
> > > Shopping for owners who are not away at siggraph.
> > > 
> > > vmiura, would you feel comfortable reviewing this? This partially reverts
> > > https://codereview.chromium.org/413103002/
> > 
> > In,
> >
>
https://codereview.chromium.org/429983002/diff/20001/gpu/command_buffer/servi...
> > there is a TODO to remove this.
> > 
> > Is it OK to depend on this change as far as you know?
> 
> It is ok in the sense that m37 webview is already depending on r286800 to
avoid
> too many fences.
> 
> I think proper fix is going to be quite involved, need to re-architect how
> webview runs GL, and switch to using per frame fences rather than per mailbox.

LGTM, thanks.

[boliu, 2014-08-12 23:14:38]

The CQ bit was checked by boliu@chromium.org

[commit-bot: I haz the power, 2014-08-12 23:52:25]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/boliu@chromium.org/463563002/1

[commit-bot: I haz the power, 2014-08-13 03:06:00]

Change committed as 289172