Always paint background and shadow separately

Always paint background and shadow separately

The logic to piggyback shadow on background painting is complex and
fragile. For the bug, we skip painting background-color which is
obscured by background-image but we also slip shadow piggybacked on
painting background. The logic was added in WebKit for some graphics
library that can draw combined background and shadow faster.

It's confirmed that Skia doesn't benefit from the logic. Local perf
test (containing hundreds of divs with simple background and shadow)
showed this CL doesn't affect rasterization performance, and degrade
record performance by 0% to 40% because of more drawing operations to
handle. Considering that the test is an extreme example of background+
shadow which is unlikely to exist in real pages, the performance
impact of this CL on real pages will be acceptable.

Telemetry-cluster showed no regression on 10k web pages.

Remove the logic to simply code and fix the bug.

BUG=660187
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2

Committed: https://crrev.com/eac477641ec5fdf9e3b9ec52d46fc3a908e8503e
Cr-Commit-Position: refs/heads/master@{#429738}

[Xianzhu, 2016-11-01 19:07:50]

Description was changed from

==========
Always paint background and shadow separately

The logic to piggyback shadow on background painting is complex and
fragile. For the bug, we skip painting background-color which is
obscured by background-image but we also slip shadow piggybacked on
painting background.

BUG=660187
==========

to

==========
Always paint background and shadow separately

The logic to piggyback shadow on background painting is complex and
fragile. For the bug, we skip painting background-color which is
obscured by background-image but we also slip shadow piggybacked on
painting background.

BUG=660187
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2
==========

[Xianzhu, 2016-11-01 19:08:04]

The CQ bit was checked by wangxianzhu@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2016-11-01 19:08:28]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-11-01 19:41:22]

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

[commit-bot: I haz the power, 2016-11-01 19:41:23]

Dry run: Try jobs failed on following builders:
  linux_layout_tests_slimming_paint_v2 on master.tryserver.chromium.linux
(JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_layout_te...)

[Xianzhu, 2016-11-03 00:51:23]

Description was changed from

==========
Always paint background and shadow separately

The logic to piggyback shadow on background painting is complex and
fragile. For the bug, we skip painting background-color which is
obscured by background-image but we also slip shadow piggybacked on
painting background.

BUG=660187
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2
==========

to

==========
Always paint background and shadow separately

The logic to piggyback shadow on background painting is complex and
fragile. For the bug, we skip painting background-color which is
obscured by background-image but we also slip shadow piggybacked on
painting background. The logic was added in WebKit for some graphics
library that can draw combined background and shadow faster.

It's confirmed that Skia doesn't benefit from the logic. Local perf
test (containing hundreds of divs with simple background and shadow)
showed this CL doesn't affect rasterization performance, and degrade
record performance by 0% to 40% because of more drawing operations to
handle. Considering that the test is an extreme example of background+
shadow which is unlikely to exist in real pages, the performance
impact of this CL on real pages will be acceptable.

Telemetry-cluster showed no regression on 10k web pages.

Remove the logic to simply code and fix the bug.

BUG=660187
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2
==========

[Xianzhu, 2016-11-03 01:08:53]

wangxianzhu@chromium.org changed reviewers:
+ pdr@chromium.org, trchen@chromium.org

[chrishtr, 2016-11-03 17:56:42]

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

[chrishtr, 2016-11-03 17:56:42]

lgtm

[Xianzhu, 2016-11-03 18:45:43]

Patchset #2 (id:20001) has been deleted

[Xianzhu, 2016-11-03 18:50:38]

The CQ bit was checked by wangxianzhu@chromium.org

[Xianzhu, 2016-11-03 18:50:39]

The patchset sent to the CQ was uploaded after l-g-t-m from
chrishtr@chromium.org
Link to the patchset: https://codereview.chromium.org/2464053003/#ps40001
(title: "git cl set-commit")

[commit-bot: I haz the power, 2016-11-03 18:51:29]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[Xianzhu, 2016-11-03 18:51:40]

The CQ bit was checked by wangxianzhu@chromium.org

[commit-bot: I haz the power, 2016-11-03 18:52:14]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-11-03 19:44:47]

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

[commit-bot: I haz the power, 2016-11-03 19:44:48]

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

[Xianzhu, 2016-11-03 19:50:05]

The CQ bit was checked by wangxianzhu@chromium.org

[commit-bot: I haz the power, 2016-11-03 19:51:00]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-11-03 22:40:12]

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

[commit-bot: I haz the power, 2016-11-03 22:40:13]

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

[Xianzhu, 2016-11-03 22:44:06]

The CQ bit was checked by wangxianzhu@chromium.org

[commit-bot: I haz the power, 2016-11-03 22:44:55]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-11-04 00:40:02]

Description was changed from

==========
Always paint background and shadow separately

The logic to piggyback shadow on background painting is complex and
fragile. For the bug, we skip painting background-color which is
obscured by background-image but we also slip shadow piggybacked on
painting background. The logic was added in WebKit for some graphics
library that can draw combined background and shadow faster.

It's confirmed that Skia doesn't benefit from the logic. Local perf
test (containing hundreds of divs with simple background and shadow)
showed this CL doesn't affect rasterization performance, and degrade
record performance by 0% to 40% because of more drawing operations to
handle. Considering that the test is an extreme example of background+
shadow which is unlikely to exist in real pages, the performance
impact of this CL on real pages will be acceptable.

Telemetry-cluster showed no regression on 10k web pages.

Remove the logic to simply code and fix the bug.

BUG=660187
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2
==========

to

==========
Always paint background and shadow separately

The logic to piggyback shadow on background painting is complex and
fragile. For the bug, we skip painting background-color which is
obscured by background-image but we also slip shadow piggybacked on
painting background. The logic was added in WebKit for some graphics
library that can draw combined background and shadow faster.

It's confirmed that Skia doesn't benefit from the logic. Local perf
test (containing hundreds of divs with simple background and shadow)
showed this CL doesn't affect rasterization performance, and degrade
record performance by 0% to 40% because of more drawing operations to
handle. Considering that the test is an extreme example of background+
shadow which is unlikely to exist in real pages, the performance
impact of this CL on real pages will be acceptable.

Telemetry-cluster showed no regression on 10k web pages.

Remove the logic to simply code and fix the bug.

BUG=660187
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2
==========

[commit-bot: I haz the power, 2016-11-04 00:40:03]

Committed patchset #2 (id:40001)

[commit-bot: I haz the power, 2016-11-04 00:45:00]

Description was changed from

==========
Always paint background and shadow separately

The logic to piggyback shadow on background painting is complex and
fragile. For the bug, we skip painting background-color which is
obscured by background-image but we also slip shadow piggybacked on
painting background. The logic was added in WebKit for some graphics
library that can draw combined background and shadow faster.

It's confirmed that Skia doesn't benefit from the logic. Local perf
test (containing hundreds of divs with simple background and shadow)
showed this CL doesn't affect rasterization performance, and degrade
record performance by 0% to 40% because of more drawing operations to
handle. Considering that the test is an extreme example of background+
shadow which is unlikely to exist in real pages, the performance
impact of this CL on real pages will be acceptable.

Telemetry-cluster showed no regression on 10k web pages.

Remove the logic to simply code and fix the bug.

BUG=660187
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2
==========

to

==========
Always paint background and shadow separately

The logic to piggyback shadow on background painting is complex and
fragile. For the bug, we skip painting background-color which is
obscured by background-image but we also slip shadow piggybacked on
painting background. The logic was added in WebKit for some graphics
library that can draw combined background and shadow faster.

It's confirmed that Skia doesn't benefit from the logic. Local perf
test (containing hundreds of divs with simple background and shadow)
showed this CL doesn't affect rasterization performance, and degrade
record performance by 0% to 40% because of more drawing operations to
handle. Considering that the test is an extreme example of background+
shadow which is unlikely to exist in real pages, the performance
impact of this CL on real pages will be acceptable.

Telemetry-cluster showed no regression on 10k web pages.

Remove the logic to simply code and fix the bug.

BUG=660187
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_layout_tests_slimming_paint_v2

Committed: https://crrev.com/eac477641ec5fdf9e3b9ec52d46fc3a908e8503e
Cr-Commit-Position: refs/heads/master@{#429738}
==========

[commit-bot: I haz the power, 2016-11-04 00:45:01]

Patchset 2 (id:??) landed as
https://crrev.com/eac477641ec5fdf9e3b9ec52d46fc3a908e8503e
Cr-Commit-Position: refs/heads/master@{#429738}