This is Mac specific. It does not repro on linux or Windows.

Bisect is needed. Adding NextAction to make sure that occurs.

Taking a look.

Did a bisect, the change which caused this behavior is:

crrev.com/2265273002 - Re-enable promotion of RenderPassDrawQuads to CALayers

Not necessarily sure that we're applying opacity twice - this may be a case that we're just supplying the opacity to CA in the wrong way or with the wrong value.

It's likely that the issue is in the CA path, as other OSes don't show this issue.

Also note that this is independent of GPU raster (Hardware acceleration in the bug title refers to hardware accelerated compositing).

erikchen@, can you take a look?

ericrk: Thanks for tracking this down. I bet CopyRenderPassDrawQuadToOverlayResource applies the opacity, and then 
"""
4045   gl_->ScheduleCALayerSharedStateCHROMIUM(                                      
4046       ca_layer_overlay->shared_state->opacity, is_clipped, clip_rect,           
4047       sorting_context_id, gl_transform);  
"""

applies the opacity again.

This bug is present on M54 (stable). This seems unlikely to be worth a refresh, but we should merge to M55.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/4a9cceac5f1883f48b564ad277d3d39cdd9f8615

commit 4a9cceac5f1883f48b564ad277d3d39cdd9f8615
Author: erikchen <erikchen@chromium.org>
Date: Tue Nov 01 18:33:25 2016

mac: Fix bug where opacity is applied twice on RenderPassDrawQuads.

When using the CoreAnimation compositor, RenderPassDrawQuads are copied to an
IOSurface. The opacity is already applied there, so there's no need for the
CoreAnimation Compositor to apply it again.

BUG= 660600 
CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel

Review-Url: https://codereview.chromium.org/2467643002
Cr-Commit-Position: refs/heads/master@{#429060}

[modify] https://crrev.com/4a9cceac5f1883f48b564ad277d3d39cdd9f8615/cc/output/gl_renderer.cc
[modify] https://crrev.com/4a9cceac5f1883f48b564ad277d3d39cdd9f8615/content/test/data/gpu/filter_effects.html
[modify] https://crrev.com/4a9cceac5f1883f48b564ad277d3d39cdd9f8615/content/test/gpu/gpu_tests/pixel_test_pages.py

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/+/8bebab8c7bad3e5c3e76ed8b83963ccd561dd522

commit 8bebab8c7bad3e5c3e76ed8b83963ccd561dd522
Author: erikchen <erikchen@chromium.org>
Date: Wed Nov 02 18:06:08 2016

[Merge to 2883] mac: Fix bug where opacity is applied twice on RenderPassDrawQuads.

> When using the CoreAnimation compositor, RenderPassDrawQuads are copied to an
> IOSurface. The opacity is already applied there, so there's no need for the
> CoreAnimation Compositor to apply it again.
>
> BUG= 660600 
> CQ_INCLUDE_TRYBOTS=master.tryserver.blink:linux_precise_blink_rel
>
> Review-Url: https://codereview.chromium.org/2467643002
> Cr-Commit-Position: refs/heads/master@{#429060}
> (cherry picked from commit 4a9cceac5f1883f48b564ad277d3d39cdd9f8615)

Review URL: https://codereview.chromium.org/2469893003 .

Cr-Commit-Position: refs/branch-heads/2883@{#423}
Cr-Branched-From: 614d31daee2f61b0180df403a8ad43f20b9f6dd7-refs/heads/master@{#423768}

[modify] https://crrev.com/8bebab8c7bad3e5c3e76ed8b83963ccd561dd522/cc/output/gl_renderer.cc
[modify] https://crrev.com/8bebab8c7bad3e5c3e76ed8b83963ccd561dd522/content/test/data/gpu/filter_effects.html
[modify] https://crrev.com/8bebab8c7bad3e5c3e76ed8b83963ccd561dd522/content/test/gpu/gpu_tests/pixel_test_pages.py

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/18cdab375861899edb98261fec71f14360a1934a

commit 18cdab375861899edb98261fec71f14360a1934a
Author: erikchen <erikchen@chromium.org>
Date: Wed Nov 02 19:04:35 2016

Update expectations for filter effect pixel tests.

Reference images look correct:
gs://chromium-gpu-archive/reference-images/Pixel_CSSFilterEffects_NoOverlays_v3_mac_8086_0a2e_msaa.png
gs://chromium-gpu-archive/reference-images/Pixel_CSSFilterEffects_v3_mac_8086_0a2e_msaa.png

BUG= 660600 ,  581526 
CQ_INCLUDE_TRYBOTS=master.tryserver.chromium.linux:linux_optional_gpu_tests_rel;master.tryserver.chromium.mac:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel;master.tryserver.chromium.android:android_optional_gpu_tests_rel

Review-Url: https://codereview.chromium.org/2472533002
Cr-Commit-Position: refs/heads/master@{#429355}

[modify] https://crrev.com/18cdab375861899edb98261fec71f14360a1934a/content/test/gpu/gpu_tests/pixel_expectations.py

 Issue 662072  has been merged into this issue.

 Issue 662764  has been merged into this issue.

Tested the issue on MacBook Pro (Retina, Mid 2012) using chrome version 55.0.2883.44 with the URL http://codepen.io/janpaul123/pen/QKReyb
Observed that both images are looking same.
Please find the attached screen shot for the same.

Adding TE-Verified labels.

Thanks,

Deleted: Screen Shot 2016-11-09 at 3.03.27 PM.png 82.2 KB

	Screen Shot 2016-11-09 at 3.03.27 PM.png 82.2 KB View Download

 Issue 662532  has been merged into this issue.

 Issue 669477  has been merged into this issue.