It would be awesome if the QA team could help us bisect this.

Tested this issue on MacPro 10.11.6 Intel Iris Pro Graphis using chrome latest canary M55-55.0.2851.0. Observed no rendering issues while loading the page.

This issue looks specific to NVIDIA GPU, Requesting someone from MTV team to look in to it.

Confirmed. This really looks like backdrop filter + DSF = 2. Jayadev, can you please take a look?

I don't think this is backdrop filter + DSF. Backdrop filters are enabled only with  "enable-experimental-web-platform-features" flag. I can repro this on MacBook Pro Retina even with that flag disabled. Also, I can't repro it on my Linux workstation even with "force-device-scale-factor=2". So, I don't think its DSF either.

I think we need a bisect to figure out what's going wrong.  

I seem to recall when triaging this that it looked related to the enabling of the Core Profile in Issue 635081, but I'm not sure. A bisect may or may not turn up the revision 1eb4627c819627991316dc07225cb76d7f3b7e35 which enabled it.

Actually this happens without core profile also (through --use-gl=desktop), if I remember correctly

Able to reproduce the issue on MacBook Pro (Retina, Mid 2012) with Graphics with NVIDIA GeForce GT 650M 1024 MB. Issue is a regression broken in M55. Below are bisect details:

Bisect info:
============
55.0.2844.0 - Good Build
55.0.2845.0 - Bad Build

Running Bisect tool invoked chromium builds with all good revisions. From the below change Log:

https://chromium.googlesource.com/chromium/src/+log/55.0.2844.0..55.0.2845.0?pretty=fuller&n=10000

Suspecting changes #415456 or #415384 could be the possible culprits:

Change Log:
===========
https://chromium.googlesource.com/chromium/src/+/d89cde2c1cfa24e74c97efa88170ed80637f2ae5

or

https://chromium.googlesource.com/chromium/src/+/324b5a97209802eac4b832f06e8ac6b83255cca2

issue is not observed on Windows or Ubuntu OS. Issue is also seen on canary version 55.0.2859.0 on MAC 10.11.6. Removing Bisect label.

Both suspect CLs are only changing tests, so they cannot be the cause for this bug. The bisect has >400 CLs in it. Is is possible to get a narrower bisect ? 

From the CL: https://chromium.googlesource.com/chromium/src/+log/55.0.2844.0..55.0.2845.0?pretty=fuller&n=10000

Suspecting Commit - 690c132cbec6ad770573f9157105fc6a041abd52	/   bdf27542f50eae092db4293a2ae5722ec924c30f	

https://codereview.chromium.org/2286843002   /  https://codereview.chromium.org/2294463002

@lukasza / wfh: Could you please look into the issue, and if it has nothing to do with your changes and if possible please do assign it to the concerned owner.

Thank you.

There's no way those CLs are related.

jaydaskia@: could you please run a bisect yourself? The most difficult part will probably be probing back into the continuous build archive at https://commondatastorage.googleapis.com/chromium-browser-snapshots/index.html?prefix=Mac/ to find the last known good revision, but https://www.chromium.org/developers/bisect-builds-py can be easily used to track down the cause.

I've tested that the issue repros at r415440, but doesn't repro one revision earlier.  Description of this CL also talks about clipping, so I think this is indeed the source of the regression.

Thank you lukasza@ for tracking this down.

I think cc doesn't handle clip parent correctly when a layer draws into a render surface.
I made a minimal repro here: http://jsbin.com/xoxepeh/
Rendering artifact can be reproduced on Linux with --force-device-scale-factor=2 (anything other than 1 will do).

Reopening this as I had to revert my CL (https://bugs.chromium.org/p/chromium/issues/detail?id=660047) 

**** Bulk edit -  please ignore if not applicable ****

A friendly reminder that M55 Stable is launch is coming soon! Your bug is labelled as Stable ReleaseBlock, pls make sure to land the fix and get it merged into the release branch ASAP so it gets enough baking time in Beta (before Stable promotion). Thank you!

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/b874eddcc31db95ab073c278a880cab0d59e655c

commit b874eddcc31db95ab073c278a880cab0d59e655c
Author: jaydasika <jaydasika@chromium.org>
Date: Wed Nov 02 22:34:28 2016

cc : Reland move screen space scale factor to root transform node

This is relanding the CL :
https://codereview.chromium.org/2408243002
If DT is the device transform, DSF is matrix scaled by device scale
factor and page scale factor for root, RT is matrix translated by root's
position, the screen space transform of root layer should be DT*DSF*RT,
https://codereview.chromium.org/2408243002 changed this to DSF*DT*RT and
that caused  https://crbug.com/660047 . This CL fixes it.

BUG= 644582 
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel

Review-Url: https://codereview.chromium.org/2473453002
Cr-Commit-Position: refs/heads/master@{#429425}

[modify] https://crrev.com/b874eddcc31db95ab073c278a880cab0d59e655c/cc/trees/draw_property_utils.cc
[modify] https://crrev.com/b874eddcc31db95ab073c278a880cab0d59e655c/cc/trees/effect_node.cc
[modify] https://crrev.com/b874eddcc31db95ab073c278a880cab0d59e655c/cc/trees/layer_tree_host_common.cc
[modify] https://crrev.com/b874eddcc31db95ab073c278a880cab0d59e655c/cc/trees/layer_tree_host_common_unittest.cc
[modify] https://crrev.com/b874eddcc31db95ab073c278a880cab0d59e655c/cc/trees/property_tree.cc
[modify] https://crrev.com/b874eddcc31db95ab073c278a880cab0d59e655c/cc/trees/property_tree.h
[modify] https://crrev.com/b874eddcc31db95ab073c278a880cab0d59e655c/cc/trees/property_tree_builder.cc
[modify] https://crrev.com/b874eddcc31db95ab073c278a880cab0d59e655c/cc/trees/property_tree_unittest.cc
[modify] https://crrev.com/b874eddcc31db95ab073c278a880cab0d59e655c/third_party/WebKit/Source/platform/graphics/compositing/PaintArtifactCompositor.cpp

Your change meets the bar and is auto-approved for M55 (branch: 2883)

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/3a760754ce49265f5fa7ce866dc846d50ee0890e

commit 3a760754ce49265f5fa7ce866dc846d50ee0890e
Author: jaydasika <jaydasika@chromium.org>
Date: Fri Nov 04 19:04:58 2016

cc : Reland Move screen space scale factor to root transform node

This CL computes the screen space scale factor which is
combined form scale factors of device transform, device
scale factor and page scale factor(if required), and stores
it on the root transform node as its surface contents scale.
This also implicitly fixes a clipping bug.
TransformTree::ComputeTransforms(a, b) should return the
transform between a and b without the surface contents
scale. But, since screen space scale was baked into the
local transform of the contents root node (before this CL),
ComputeTransform(a, root) was having the scale baked in.
(and this caused the bug)

CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel

Review-Url: https://codereview.chromium.org/2408243002
Cr-Commit-Position: refs/heads/master@{#427366}
(cherry picked from commit 61fa0fb418a3f77665e0bc1be88ec5a2984c6dc6)

Committed: https://chromium.googlesource.com/chromium/src/+/b96a29859b70eea11eacefef82b25318d43700fc

patch from issue 2452023003 at patchset 1 (http://crrev.com/2452023003#ps1)

NOPRESUBMIT=true
NOTREECHECKS=true
NOTRY=true
BUG= 644582 

Review-Url: https://codereview.chromium.org/2464103005
Cr-Commit-Position: refs/branch-heads/2883@{#460}
Cr-Branched-From: 614d31daee2f61b0180df403a8ad43f20b9f6dd7-refs/heads/master@{#423768}

[modify] https://crrev.com/3a760754ce49265f5fa7ce866dc846d50ee0890e/cc/trees/draw_property_utils.cc
[modify] https://crrev.com/3a760754ce49265f5fa7ce866dc846d50ee0890e/cc/trees/effect_node.cc
[modify] https://crrev.com/3a760754ce49265f5fa7ce866dc846d50ee0890e/cc/trees/layer_tree_host_common.cc
[modify] https://crrev.com/3a760754ce49265f5fa7ce866dc846d50ee0890e/cc/trees/layer_tree_host_common_unittest.cc
[modify] https://crrev.com/3a760754ce49265f5fa7ce866dc846d50ee0890e/cc/trees/layer_tree_impl.cc
[modify] https://crrev.com/3a760754ce49265f5fa7ce866dc846d50ee0890e/cc/trees/property_tree.cc
[modify] https://crrev.com/3a760754ce49265f5fa7ce866dc846d50ee0890e/cc/trees/property_tree.h
[modify] https://crrev.com/3a760754ce49265f5fa7ce866dc846d50ee0890e/cc/trees/property_tree_builder.cc
[modify] https://crrev.com/3a760754ce49265f5fa7ce866dc846d50ee0890e/cc/trees/property_tree_unittest.cc
[modify] https://crrev.com/3a760754ce49265f5fa7ce866dc846d50ee0890e/third_party/WebKit/Source/platform/graphics/compositing/PaintArtifactCompositor.cpp

**** Bulk edit -  please ignore if not applicable ****

A friendly reminder that M55 Stable is launch is coming soon! Your bug is labelled as Stable ReleaseBlock, pls make sure to land the fix and get it merged into the release branch ASAP so it gets enough baking time in Beta (before Stable promotion). Thank you!

Also due to Thanksgiving holidays in US, please make sure all fixes are ready and merged to M55 latest by 5:00 PM PT Friday, 11/18/16.

Tested the issue on MacBook Pro (Retina, Mid 2012) with Graphics with NVIDIA GeForce GT 650M 1024 MB using chrome version 55.0.2883.44 with the URL http://www.apple.com/macos/sierra-preview/.Page rendered without any issues.
Please find the attached screen shot for the same.

Adding TE-Verified labels.

Thanks,

Deleted: Screen Shot 2016-11-09 at 10.26.39 AM.png 10.7 MB

	Screen Shot 2016-11-09 at 10.26.39 AM.png 10.7 MB View Download

I can reproduce a similar issue on Chrome Beta 57.0.2987.37

A friendly reminder that M57 Stable is launch is coming VERY soon! Your bug is labelled as Stable ReleaseBlock, pls make sure to land the fix and get it merged into the release branch (2987) ASAP so it gets enough baking time in Beta (before Stable promotion). Thank you!

Victor, I filed a new bug  Issue 693213  about the regression since this particular issue was fixed months ago. Marking this fixed again.