Bug 54509 - sna: screen corruption with glx-dock (cairo-dock)
Summary: sna: screen corruption with glx-dock (cairo-dock)
Alias: None
Product: Mesa
Classification: Unclassified
Component: Drivers/DRI/i915 (show other bugs)
Version: 8.0
Hardware: All All
: medium normal
Assignee: Ian Romanick
QA Contact:
Keywords: licence
Depends on:
Reported: 2012-09-04 20:35 UTC by Matthew Monaco
Modified: 2019-09-18 19:35 UTC (History)
1 user (show)

See Also:
i915 platform:
i915 features:

Corrupt Screen (2.56 MB, image/jpeg)
2012-09-04 20:36 UTC, Matthew Monaco
Normal Screen (2.55 MB, image/jpeg)
2012-09-04 20:36 UTC, Matthew Monaco
dmesg.log (52.79 KB, text/plain)
2012-09-04 20:36 UTC, Matthew Monaco
Xorg.log (45.12 KB, text/plain)
2012-09-04 20:37 UTC, Matthew Monaco

Description Matthew Monaco 2012-09-04 20:35:20 UTC
I've been getting screen corruption since SNA was first introduced. The latest version of xf86-video-intel that I've tried is 2.20.6.

Some details about my system:
 Arch Linux x86_64
 cairo-dock v3.0.2
 xcompmgr v1.1.6 (run as xcompmgr -c)
 mesa 8.0.4
 gpu: i915 (core i3)

The corruption is more or less as shown in the screenshot. Sometimes it's more colorful, sometimes it's completely black, but always covering about the same area in the top left corner. It happens almost every time I click on a particular menu button.

It only happens in SNA mode. It doesn't not happen with the dock is using the cairo backend (with or without SNA), only the opengl backend.

The screen goes back to normal if the corrupted area is refreshed; I usually just hit F11 in Firefox or gnome-terminal...

Please let me know what else I can provide =)
Comment 1 Matthew Monaco 2012-09-04 20:36:00 UTC
Created attachment 66632 [details]
Corrupt Screen
Comment 2 Matthew Monaco 2012-09-04 20:36:26 UTC
Created attachment 66634 [details]
Normal Screen
Comment 3 Matthew Monaco 2012-09-04 20:36:54 UTC
Created attachment 66635 [details]
Comment 4 Matthew Monaco 2012-09-04 20:37:10 UTC
Created attachment 66636 [details]
Comment 5 Chris Wilson 2012-09-04 20:47:39 UTC
To be the bearer of bad news, that is looks just like a mesa bug. Cat ~/.xsession-errors would be a good one to check.
Comment 6 Matthew Monaco 2012-09-04 20:52:39 UTC
Sorry, I thought that since it only happened with SNA...

I can reproduce the bug without any new output in .xsession-errors.

However, if click the dock applet, and refresh the screen a few times, I eventually see:

error 8: BadMatch (invalid parameter attributes) request 152 minor 6 serial 56132
error 9: BadDrawable (invalid Pixmap or Window parameter) request 149 minor 4 serial 56133
error 4: BadPixmap (invalid Pixmap parameter) request 54 minor 0 serial 56183
Comment 7 Chris Wilson 2012-09-04 20:58:37 UTC
Using cairo-dock-3.0.2 on debian and cairo-dock -o, it renders incorrectly using UXA or SNA, so I'm not convinced that it is bug free to begin with... And since it can't keep track of its own Pixmaps....
Comment 8 Chris Wilson 2012-09-04 21:02:42 UTC
If you think it is worth persuing, your best bet is to try recompiling xf86-video-intel with --enable-debug (and perhaps --enable-debug=full) to see if that catches anything obvious.
Comment 9 Matthew Monaco 2012-09-04 21:30:46 UTC
I definitely /cannot/ reproduce the corruption with UXA. However, I still see those three lines after a bit of mucking with the dock in .xsession-errors. I don't exactly follow your meaning, do you now think it might be a cairo-dock bug rather than mesa?

I'll recompile and post Xorg.log soon. It's definitely worth pursuing, I'm just not sure where =)
Comment 10 Chris Wilson 2012-09-04 21:39:54 UTC
The garbage you have on your screen looks like a mesa bug, as if it is keeping stale data around in its state tracker. The misrendering I see (where the wrong icons are drawn at the wrong time) is a cairo-dock bug. The errors reported for invalid pixmap is a cairo-dock bug.
Comment 11 Matthew Monaco 2012-09-04 21:53:20 UTC
Don't know if this helps anymore, but I recompiled so I'm posting it:


(nb: I've never used or heard of wikisend before, but it is ok with a 16M file)
Comment 12 Matthew Monaco 2012-09-06 18:29:40 UTC
I'm switching the product to mesa, might someone from mesa take a look at the corruption to see if it might be a mesa problem?
Comment 13 Elizabeth 2017-10-27 20:25:34 UTC
Hello Matthew, is this still valid with latest SW?? Thank you.
Comment 14 GitLab Migration User 2019-09-18 19:35:13 UTC
-- GitLab Migration Automatic Message --

This bug has been migrated to freedesktop.org's GitLab instance and has been closed from further activity.

You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.freedesktop.org/mesa/mesa/issues/716.

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.