Bug 101565 - Intermittent graphical corruption since "radeonsi: don't emit partial flushes at the end of IBs (v2)" (c9040dc9)
Summary: Intermittent graphical corruption since "radeonsi: don't emit partial flushes...
Status: RESOLVED FIXED
Alias: None
Product: Mesa
Classification: Unclassified
Component: Drivers/Gallium/radeonsi (show other bugs)
Version: git
Hardware: Other All
: medium normal
Assignee: Default DRI bug account
QA Contact: Default DRI bug account
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2017-06-23 14:03 UTC by Grigori Goronzy
Modified: 2017-06-23 17:47 UTC (History)
0 users

See Also:
i915 platform:
i915 features:

Attachments
Example of graphical corruption (387.82 KB, image/jpeg)
2017-06-23 14:03 UTC, Grigori Goronzy 		Details
View All

Description Grigori Goronzy 2017-06-23 14:03:45 UTC 
Created attachment 132156 [details]
Example of graphical corruption

Commit c9040dc9 seems to cause graphical corruption on VI for me. It looks like some caching issue, since the corruption is intermittent and mostly appears only for a single frame. Maybe some assumptions made about the kernel's handling of IBs aren't true?

I can trigger issues reliably by running some OpenCL load in the background (for instance, an Ethereum miner). That will make X (glamor) and gnome-shell rendering glitch randomly with high probability.

Reverting the commit fixes the problem completely.

Attached is a screenshot, showing what kind of artefacts I encounter. (Using the screenshot tool doesn't work as it'll rerender everything before capture)
Comment 1 Marek Olšák 2017-06-23 17:47:30 UTC 
I reverted the commit.

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.