Prefer to use the OES_depth24 extension for WebGL's depth-only back buffers.

Prefer to use the OES_depth24 extension for WebGL's depth-only back buffers.

Instead of using the OES_packed_depth_stencil extension to obtain 24-bit
depth buffers, prefer the OES_depth24 extension. This avoids a situation
where the depth buffer is cleared but the stencil buffer must be
preserved, leading to performance problems on some GPUs.

BUG=584807
Committed: https://crrev.com/2374479265b26ba0a4a658582b068a4526220774
Cr-Commit-Position: refs/heads/master@{#374562}

[Ken Russell (switch to Gerrit), 2016-02-06 03:26:18]

kbr@chromium.org changed reviewers:
+ bajones@chromium.org, zmo@chromium.org

[Ken Russell (switch to Gerrit), 2016-02-06 03:26:19]

Mo, Brandon: please review.

Stephane: could you try patching this into your Chromium workspace and make sure
it fixes the performance problem? Thanks.

[marcheu, 2016-02-06 03:54:28]

On 2016/02/06 03:26:19, Ken Russell wrote:
> Mo, Brandon: please review.
> 
> Stephane: could you try patching this into your Chromium workspace and make
sure
> it fixes the performance problem? Thanks.

I will have to get a device first, it arrived in MTV today but not in my hands
yet :)

[marcheu, 2016-02-06 04:13:23]

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

[marcheu, 2016-02-06 04:13:24]

patch lgtm in any case

[Zhenyao Mo, 2016-02-07 18:16:21]

On 2016/02/06 04:13:24, marcheu wrote:
> patch lgtm in any case

The code looks good, but it seems you will need to update the unittest.

[Ken Russell (switch to Gerrit), 2016-02-09 01:37:43]

On 2016/02/07 18:16:21, Zhenyao Mo wrote:
> On 2016/02/06 04:13:24, marcheu wrote:
> > patch lgtm in any case
> 
> The code looks good, but it seems you will need to update the unittest.

Unittest has been updated.

[Zhenyao Mo, 2016-02-09 17:28:09]

lgtm

[Zhenyao Mo, 2016-02-09 17:29:10]

Just realized this might conflict with
https://codereview.chromium.org/1673113002/.

[Ken Russell (switch to Gerrit), 2016-02-09 23:56:30]

On 2016/02/09 17:29:10, Zhenyao Mo wrote:
> Just realized this might conflict with
> https://codereview.chromium.org/1673113002/.

Thanks. Rebased it and re-tested. CQ'ing.

[Ken Russell (switch to Gerrit), 2016-02-09 23:56:34]

The CQ bit was checked by kbr@chromium.org

[Ken Russell (switch to Gerrit), 2016-02-09 23:56:34]

The patchset sent to the CQ was uploaded after l-g-t-m from
marcheu@chromium.org, zmo@chromium.org
Link to the patchset: https://codereview.chromium.org/1675973002/#ps80001
(title: "Rebased.")

[commit-bot: I haz the power, 2016-02-09 23:56:50]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1675973002/80001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1675973002/80001

[commit-bot: I haz the power, 2016-02-10 01:11:46]

Committed patchset #5 (id:80001)

[commit-bot: I haz the power, 2016-02-10 01:12:57]

Description was changed from

==========
Prefer to use the OES_depth24 extension for WebGL's depth-only back buffers.

Instead of using the OES_packed_depth_stencil extension to obtain 24-bit
depth buffers, prefer the OES_depth24 extension. This avoids a situation
where the depth buffer is cleared but the stencil buffer must be
preserved, leading to performance problems on some GPUs.

BUG=584807
==========

to

==========
Prefer to use the OES_depth24 extension for WebGL's depth-only back buffers.

Instead of using the OES_packed_depth_stencil extension to obtain 24-bit
depth buffers, prefer the OES_depth24 extension. This avoids a situation
where the depth buffer is cleared but the stencil buffer must be
preserved, leading to performance problems on some GPUs.

BUG=584807

Committed: https://crrev.com/2374479265b26ba0a4a658582b068a4526220774
Cr-Commit-Position: refs/heads/master@{#374562}
==========

[commit-bot: I haz the power, 2016-02-10 01:12:58]

Patchset 5 (id:??) landed as
https://crrev.com/2374479265b26ba0a4a658582b068a4526220774
Cr-Commit-Position: refs/heads/master@{#374562}

[phoglund_chromium, 2016-02-16 10:15:41]

A revert of this CL (patchset #5 id:80001) has been created in
https://codereview.chromium.org/1698883004/ by phoglund@chromium.org.

The reason for reverting is: Speculative revert: appears to break
printing/webgl-oversized-printing.html on Linux MSAN:
https://build.chromium.org/p/chromium.webkit/builders/WebKit%20Linux%20MSAN/b....