New issue
Advanced search Search tips

Issue 886970 link

Starred by 5 users

Issue metadata

Status: ExternalDependency
Owner:
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Mac
Pri: 2
Type: Bug

Blocking:
issue 662644


Show other hotlists

Hotlists containing this issue:
webgl-conformance-all


Sign in to add a comment

Invalid framebuffer alpha

Reported by a.kunit...@redbrixwall.com, Sep 19

Issue description

UserAgent: Mozilla/5.0 (Macintosh; Intel Mac OS X 10_13_6) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/69.0.3497.100 Safari/537.36

Steps to reproduce the problem:
1. open attachment

What is the expected behavior?
square must be green

What went wrong?
square is red

Did this work before? N/A 

Does this work in other browsers? Yes

Chrome version: 69.0.3497.100  Channel: stable
OS Version: OS X 10.13.6
Flash Version: 

https://jsfiddle.net/61qkufer/248/
1) if render to texture on next frame, is ok
2) framebuffer not empty, just alpha is 0 

MacBook Pro (Retina, 15-inch, Mid 2015) Graphics Intel Iris Pro 1536 MB
 
chrome-bug.html
4.5 KB View Download
Screen Shot 2018-09-19 at 20.46.57.png
146 KB View Download
Labels: Needs-Triage-M69
Components: Internals>GPU>VendorSpecific
Labels: GPU-Intel
Green on:
Chrome version	Chrome/69.0.3497.100
Operating system	Mac OS X 10.13.6
Machine model name	MacBookAir
Machine model version	7.2
GL_VENDOR	Intel Inc.
GL_RENDERER	Intel(R) HD Graphics 6000
GL_VERSION	4.1 INTEL-10.36.19

So I guess it's very hardware specific.
Cc: yunchao...@intel.com yang...@intel.com
Yang, Yunchao: could you please triage this bug on whatever Mac hardware you have available and see how many GPU types it affects?

Owner: yang...@intel.com
Sure, we will triage. 
Labels: Triaged-ET Target-71 M-71 FoundIn-71
Status: Untriaged (was: Unconfirmed)
Able to reproduce the issue on MacBook Pro (Retina, 15-inch, Mid 2015) 10.13.3 using chrome reported version #69.0.3497.100 and latest canary #71.0.3556.0. Issue is specific to OS-mac.
This is a non-regression issue as it is observed from M60 old builds. 

Hence, marking it as untriaged to get more inputs from dev team.

Thanks...!!
I can only reproduce the issue on MacBook Pro (Retina, 15-inch, Mid 2015, Intel Iris Pro). 

Here are the devices I have tested (macOS version is 10.13.6)


Macbook Pro Mid 2015 (15-inch) 10.13.6, Intel Iris Pro)	69.0.3497.100 	Red
Macbook Pro 2017 (13-inch, Iris Graphics 650)			69.0.3497.100	Green
Macbook Pro 2018 (13-inch, Iris Plus Graphics 655)		70.0.3514.2 	Green
Macbook Early 2016 (12-inch, HD Graphics 515) 			69.0.3497.100	Green
Macbook Air Early 2015 (13-inch, HD Graphics 6000) 		69.0.3497.100	Green
Macbook Pro 2016 (13-inch, Iris Graphics 550) 			69.0.3497.100	Green
Macbook Pro 2018 (15-inch, UHD Graphics 630) 			69.0.3497.100	Green
Hm... Do we have different workarounds on that device in particular? Or is it a hardware/driver bug on Iris Pro only, that Chrome happens to hit because it does something else differently from the other browsers?
Owner: jiawei.s...@intel.com
Status: Assigned (was: Untriaged)
Could someone please also contribute a WebGL conformance test to https://github.com/KhronosGroup/WebGL – perhaps under conformance/rendering – that catches this alpha blending error? It looks like RGB textures are broken on this driver. Thanks.

I have written a conformance test based on the case provided here:
https://github.com/KhronosGroup/WebGL/pull/2716
Shall we write a Radar about this issue to Apple?
Yes, please. Would you mind writing a draft? Thanks.

Status: ExternalDependency (was: Assigned)
Thanks Jiawei for writing the conformance test and text for the Radar. Filed Radar 45235525 about it.

Submitter, I recommend you allocate RGBA instead of RGB textures as a workaround. We would prefer to not work around this bug in the browser; it will be a lot of code for one particular GPU type on macOS, and hopefully Apple and Intel will collaborate to fix it.

I believe is a chromium/angle bug on working with alpha channel on frame buffers, in our product we have more strangle result with invalid alpha-channel values at any GPU, (Chrome only): Transparent canvas on each 3-th frame (on canvas configured without alpha), but reproduce on POC code too hard. For now we use workaround: clean alpha channel on end of frame.
Project Member

Comment 16 by bugdroid1@chromium.org, Oct 24

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/319faa6f9fc6bb050759c111202dbb61ba99086d

commit 319faa6f9fc6bb050759c111202dbb61ba99086d
Author: Kenneth Russell <kbr@chromium.org>
Date: Wed Oct 24 17:21:44 2018

Roll WebGL 6d2f3f4..0d55c88

https://chromium.googlesource.com/external/khronosgroup/webgl.git/+log/6d2f3f4..0d55c88

Includes suppressions for the following failures:

  conformance/rendering/rendering-sampling-feedback-loop.html
  conformance2/textures/misc/tex-mipmap-levels.html
  conformance2/textures/image_bitmap_from_video/
    tex-2d-rgb565-rgb-unsigned_short_5_6_5.html
  deqp/functional/gles3/lifetime.html
  deqp/functional/gles3/shaderoperator/common_functions_*.html

Bug: 660844,  739604 , 766918, 849576, 886970,  890002 , 890539,  898350 ,  898351 ,  angleproject:2832 
Cq-Include-Trybots: luci.chromium.try:win_optional_gpu_tests_rel;luci.chromium.try:mac_optional_gpu_tests_rel;luci.chromium.try:linux_optional_gpu_tests_rel;luci.chromium.try:android_optional_gpu_tests_rel;luci.chromium.try:linux_angle_rel_ng;luci.chromium.try:win_angle_rel_ng
Change-Id: I85be441fa86090e5ed056a9366452248985693e3
Reviewed-on: https://chromium-review.googlesource.com/c/1297570
Reviewed-by: James Darpinian <jdarpinian@chromium.org>
Commit-Queue: Kenneth Russell <kbr@chromium.org>
Cr-Commit-Position: refs/heads/master@{#602377}
[modify] https://crrev.com/319faa6f9fc6bb050759c111202dbb61ba99086d/DEPS
[modify] https://crrev.com/319faa6f9fc6bb050759c111202dbb61ba99086d/content/test/gpu/gpu_tests/webgl2_conformance_expectations.py
[modify] https://crrev.com/319faa6f9fc6bb050759c111202dbb61ba99086d/content/test/gpu/gpu_tests/webgl_conformance_expectations.py
[modify] https://crrev.com/319faa6f9fc6bb050759c111202dbb61ba99086d/content/test/gpu/gpu_tests/webgl_conformance_revision.txt

More information: it's a regression in the Mac OpenGL driver for Haswell GPUs in 10.13. The test passes on 10.12. A suppression for the failure on 10.13 is incoming.

Project Member

Comment 18 by bugdroid1@chromium.org, Oct 24

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/629399230ca1367b962a2dfa89b89c4a95ede1ee

commit 629399230ca1367b962a2dfa89b89c4a95ede1ee
Author: Kenneth Russell <kbr@chromium.org>
Date: Wed Oct 24 22:09:21 2018

Suppress canvas-alpha-bug failure on Mac Intel Haswell GPUs.

conformance/rendering/canvas-alpha-bug.html is failing on this
configuration on 10.13 and later. Rather than specialize the
suppression for multiple OS versions, suppress it on this GPU type.

Tbr: jdarpinian@chromium.org
No-Try: True
Bug: 886970
Change-Id: Ifeeaa59aca1c232b3d3eda6b2a82b35451c8a2bc
Reviewed-on: https://chromium-review.googlesource.com/c/1298199
Commit-Queue: Kenneth Russell <kbr@chromium.org>
Reviewed-by: Kenneth Russell <kbr@chromium.org>
Cr-Commit-Position: refs/heads/master@{#602479}
[modify] https://crrev.com/629399230ca1367b962a2dfa89b89c4a95ede1ee/content/test/gpu/gpu_tests/webgl2_conformance_expectations.py
[modify] https://crrev.com/629399230ca1367b962a2dfa89b89c4a95ede1ee/content/test/gpu/gpu_tests/webgl_conformance_expectations.py

Blocking: 662644
Labels: webgl-conformance
Square is red

MacBook PRO, Late 2013 A1502 (13-inch, Intel Iris Graphics 6100) 10.14.1
Square is BLUE

iMac (27-inch, Mid 2010) ATI Radeon HD 5750 1024MB 
MacOS 10.11.6
Chrome 71.0.3578.80

Sign in to add a comment