Bug 79414 - GFX_OGL turing-fluid demo not working as expected with HDMI+eDP display cloning
Summary: GFX_OGL turing-fluid demo not working as expected with HDMI+eDP display cloning
Status: RESOLVED DUPLICATE of bug 87345
Alias: None
Product: Mesa
Classification: Unclassified
Component: Drivers/DRI/i965 (show other bugs)
Version: unspecified
Hardware: x86-64 (AMD64) Linux (All)
: medium major
Assignee: Ian Romanick
QA Contact: Intel 3D Bugs Mailing List
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2014-05-29 14:59 UTC by Joe Konno
Modified: 2015-02-09 04:17 UTC (History)
5 users (show)

See Also:
i915 platform:
i915 features:


Attachments
turing-fluid demo corruption (1.64 MB, image/png)
2014-05-29 14:59 UTC, Joe Konno
Details

Description Joe Konno 2014-05-29 14:59:31 UTC
Created attachment 100107 [details]
turing-fluid demo corruption

GFX_OGL turing-fluid demo shows substantial display corruption (attached) with a GL-composited window manager, the Chrome web browser, and a cloned HDMI+eDP display configuration. The corruption becomes worse over execution time.

This has been reproduced on multiple hardware configurations (where HDMI+eDP is possible) and multiple Linux distros.

Steps to Reproduce:

1. Boot to a GL-composited window manager

2. Connect HDMI +eDP, display clone (not extend)

3. Open Google Chrome web browser

4. Run the below WebGL Demo:
   http://cake23.de/turing-fluid.html

5. Check the behavior of colors over the Window and click on the same

Attachments:

turing fluid demo corruption) Screenshot taken from a Chromium OS build, also seen by others on differing hardware configurations and multiple linux distros.
Comment 1 Joe Konno 2014-05-29 15:30:52 UTC
Comment on attachment 100107 [details]
turing-fluid demo corruption

Added image/png mime type.
Comment 2 Kenneth Graunke 2014-05-29 15:39:47 UTC
http://people.freedesktop.org/~kwg/fluid.trace

I replayed an apitrace on Intel and nVidia a while back.  IIRC, 1x1 green pixels appear in the starting image.  On nVidia, those get blurred/filtered out near the start.  On Intel, they don't, and so the actual algorithm of the demo seems to perpetuate them in odd ways.

I haven't looked into it further, so someone is free to pick this bug up.
Comment 3 Jason Ekstrand 2015-02-09 04:17:02 UTC

*** This bug has been marked as a duplicate of bug 87345 ***


Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.