Add display list fallback reason histograms for 2D canvas

Add display list fallback reason histograms for 2D canvas

Two new histogram to track the reasons for which a) display-list-2d-
canvas disables itself internally, and b) GPU accelerated canvas
stops deferring and batching draw operations.

BUG=577896
Committed: https://crrev.com/bce6e17e44e6ce516ed63b6360823dea10461b3e
Cr-Commit-Position: refs/heads/master@{#373861}

[Justin Novosad, 2016-01-20 22:20:46]

junov@chromium.org changed reviewers:
+ asvitkine@chromium.org, senorblanco@chromium.org

[Justin Novosad, 2016-01-20 22:20:47]

PTAL.
asvitkine: histogram stuff (calls are in Canvas2DLayerBridge.cpp and
RecordingImageBufferSurface.cpp)
senorblanco: everything else

[Stephen White, 2016-01-21 21:02:31]

LGTM w/nits

https://codereview.chromium.org/1609343004/diff/20001/tools/metrics/histogram...
File tools/metrics/histograms/histograms.xml (right):

https://codereview.chromium.org/1609343004/diff/20001/tools/metrics/histogram...
tools/metrics/histograms/histograms.xml:3594: +    The reasons why a GPU
accelerated canvas stope deferring it rendering
Nit: deferring it rendering -> deferring its rendering

https://codereview.chromium.org/1609343004/diff/20001/tools/metrics/histogram...
tools/metrics/histograms/histograms.xml:57953: +      label="Canvas state stack
too large (unballanced save/restore?)."/>
Nit: unballanced -> unbalanced

https://codereview.chromium.org/1609343004/diff/20001/tools/metrics/histogram...
tools/metrics/histograms/histograms.xml:57967: +  <int value="15" label="Darwing
a texture backed pattern."/>
Nit: suggest hyphenating -> texture-backed

https://codereview.chromium.org/1609343004/diff/20001/tools/metrics/histogram...
tools/metrics/histograms/histograms.xml:57976: +      label="Canvas state stack
too large (unballanced save/restore?)."/>
Nit: unballanced -> unbalanced

[Alexei Svitkine (slow), 2016-01-21 22:12:31]

asvitkine@chromium.org changed reviewers:
+ jwd@chromium.org
- asvitkine@chromium.org

[Alexei Svitkine (slow), 2016-01-21 22:12:31]

Re-directing histograms review to Jesse.

[jwd, 2016-01-22 21:22:20]

LGTM with Stephen's nits and mine.

https://codereview.chromium.org/1609343004/diff/20001/tools/metrics/histogram...
File tools/metrics/histograms/histograms.xml (right):

https://codereview.chromium.org/1609343004/diff/20001/tools/metrics/histogram...
tools/metrics/histograms/histograms.xml:3594: +    The reasons why a GPU
accelerated canvas stope deferring it rendering
stope -> stops? stopped?

[Justin Novosad, 2016-02-05 17:07:15]

The CQ bit was checked by junov@chromium.org

[Justin Novosad, 2016-02-05 17:07:15]

The patchset sent to the CQ was uploaded after l-g-t-m from
senorblanco@chromium.org, jwd@chromium.org
Link to the patchset: https://codereview.chromium.org/1609343004/#ps40001
(title: "git cl web")

[commit-bot: I haz the power, 2016-02-05 17:07:27]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1609343004/40001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1609343004/40001

[commit-bot: I haz the power, 2016-02-05 18:52:09]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2016-02-05 18:53:15]

Description was changed from

==========
Add display list fallback reason histograms for 2D canvas

Two new histogram to track the reasons for which a) display-list-2d-
canvas disables itself internally, and b) GPU accelerated canvas
stops deferring and batching draw operations.

BUG=577896
==========

to

==========
Add display list fallback reason histograms for 2D canvas

Two new histogram to track the reasons for which a) display-list-2d-
canvas disables itself internally, and b) GPU accelerated canvas
stops deferring and batching draw operations.

BUG=577896

Committed: https://crrev.com/bce6e17e44e6ce516ed63b6360823dea10461b3e
Cr-Commit-Position: refs/heads/master@{#373861}
==========

[commit-bot: I haz the power, 2016-02-05 18:53:16]

Patchset 3 (id:??) landed as
https://crrev.com/bce6e17e44e6ce516ed63b6360823dea10461b3e
Cr-Commit-Position: refs/heads/master@{#373861}

[please use gerrit instead, 2016-02-05 23:12:15]

This patch may introduced a build bot failure in webkit_tests:

https://build.chromium.org/p/chromium.webkit/builders/WebKit%20Win%20Oilpan%2...

Regressions: Unexpected timeouts (1)
virtual/gpu/fast/canvas/canvas-createImageBitmap-drawImage-flipY.html[Timeout]

If the failure persists, I will revert this patch speculatively.

[Justin Novosad, 2016-02-08 21:34:54]

That is a new test that was added just a few days ago:
https://codereview.chromium.org/1631733003/

[Justin Novosad, 2016-02-08 21:39:19]

On 2016/02/08 21:34:54, Justin Novosad wrote:
> That is a new test that was added just a few days ago:
> https://codereview.chromium.org/1631733003/

xidachen@ is working on marking it as slow. The test it was derived from is
already marked as slow.