Disable paint invalidatin optimization about SVG transform change

Disable paint invalidatin optimization about SVG transform change

The optimization prunes SVG subtree paint invalidation walking if
the container's transform changed. However, the transform changed
flag is only updated during layout causing:
- If multiple layouts before an invalidation, later layouts may
  clear the flag set by previous layout;
- For paint-only invalidation (after we use unified invalidation)
  the flag may remain uncleared if there is no layout causing
  incorrect pruning of tree walking.

SvgCubics doesn't show visible performance change with this patch.

BUG=394619
Committed: https://src.chromium.org/viewvc/blink?view=rev&revision=179076

[fs, 2014-07-24 08:48:42]

If this doesn't (noticeably) degrade performance, then I'm all for it.

[dsinclair, 2014-07-24 17:20:58]

lgtm

[Julien - ping for review, 2014-07-28 18:04:07]

lgtm

fs, the idea is to see if this kills our performance. We think that with the
invalidation-after-compositing-update, we may still trigger a lot of layout
(e.g. in SVGCubics) but we hope to only trigger a limited number of invalidation
phases.

[Julien - ping for review, 2014-07-28 18:05:28]

https://codereview.chromium.org/412083003/diff/1/Source/core/rendering/svg/Re...
File Source/core/rendering/svg/RenderSVGModelObject.cpp (left):

https://codereview.chromium.org/412083003/diff/1/Source/core/rendering/svg/Re...
Source/core/rendering/svg/RenderSVGModelObject.cpp:158: return;
Actually, we should re-enable the ASSERT in RenderObject for SVG sub-trees in
RenderObject::assertSubtreeClearedPaintInvalidationState.

[Xianzhu, 2014-07-28 20:09:38]

https://codereview.chromium.org/412083003/diff/1/Source/core/rendering/svg/Re...
File Source/core/rendering/svg/RenderSVGModelObject.cpp (left):

https://codereview.chromium.org/412083003/diff/1/Source/core/rendering/svg/Re...
Source/core/rendering/svg/RenderSVGModelObject.cpp:158: return;
On 2014/07/28 18:05:27, Julien Chaffraix - PST wrote:
> Actually, we should re-enable the ASSERT in RenderObject for SVG sub-trees in
> RenderObject::assertSubtreeClearedPaintInvalidationState.

Done.

[Xianzhu, 2014-07-28 20:09:41]

The CQ bit was checked by wangxianzhu@chromium.org

[commit-bot: I haz the power, 2014-07-28 20:10:00]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/wangxianzhu@chromium.org/412083003/20001

[commit-bot: I haz the power, 2014-07-28 21:21:37]

FYI, CQ is re-trying this CL (attempt #1).
The failing builders are:
  android_blink_compile_rel on tryserver.blink
(http://build.chromium.org/p/tryserver.blink/builders/android_blink_compile_re...)
  linux_blink_dbg on tryserver.blink
(http://build.chromium.org/p/tryserver.blink/builders/linux_blink_dbg/builds/1...)
  linux_blink_rel on tryserver.blink
(http://build.chromium.org/p/tryserver.blink/builders/linux_blink_rel/builds/1...)
  mac_blink_rel on tryserver.blink
(http://build.chromium.org/p/tryserver.blink/builders/mac_blink_rel/builds/17384)
  win_blink_compile_dbg on tryserver.blink
(http://build.chromium.org/p/tryserver.blink/builders/win_blink_compile_dbg/bu...)
  win_blink_rel on tryserver.blink
(http://build.chromium.org/p/tryserver.blink/builders/win_blink_rel/builds/19285)
  linux_gpu_triggered_tests on tryserver.chromium.gpu
(http://build.chromium.org/p/tryserver.chromium.gpu/builders/linux_gpu_trigger...)

[commit-bot: I haz the power, 2014-07-28 21:46:47]

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

[commit-bot: I haz the power, 2014-07-28 21:46:48]

Try jobs failed on following builders:
  linux_blink_rel on tryserver.blink
(http://build.chromium.org/p/tryserver.blink/builders/linux_blink_rel/builds/1...)

[Xianzhu, 2014-07-28 22:20:10]

The CQ bit was checked by wangxianzhu@chromium.org

[commit-bot: I haz the power, 2014-07-28 22:20:35]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/wangxianzhu@chromium.org/412083003/40001

[commit-bot: I haz the power, 2014-07-29 00:38:15]

Change committed as 179076