Bug 92632 - Huge performance drop and inaccurate cursor
Summary: Huge performance drop and inaccurate cursor
Status: RESOLVED MOVED
Alias: None
Product: xorg
Classification: Unclassified
Component: Driver/Radeon (show other bugs)
Version: unspecified
Hardware: Other Linux (All)
: medium normal
Assignee: xf86-video-ati maintainers
QA Contact: Xorg Project Team
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2015-10-23 09:34 UTC by Jan Burgmeier
Modified: 2019-11-19 07:53 UTC (History)
0 users

See Also:
i915 platform:
i915 features:


Attachments
dmesg output (44.33 KB, text/plain)
2015-10-23 09:34 UTC, Jan Burgmeier
no flags Details
Output of lspci -v (7.10 KB, text/plain)
2015-10-23 09:34 UTC, Jan Burgmeier
no flags Details
Xorg server logfile (43.91 KB, text/plain)
2015-10-23 09:34 UTC, Jan Burgmeier
no flags Details
xrandr --verbose output while running with git master (problem occures) (3.99 KB, text/plain)
2015-10-28 08:13 UTC, Jan Burgmeier
no flags Details
patch to revert commit 673e1c76. (6.28 KB, text/plain)
2015-10-28 08:14 UTC, Jan Burgmeier
no flags Details
Xorg log while running with patch to revert commit 673e1c76. (46.85 KB, text/plain)
2015-10-28 08:17 UTC, Jan Burgmeier
no flags Details

Description Jan Burgmeier 2015-10-23 09:34:18 UTC
Created attachment 119134 [details]
dmesg output

Hi all,

I'm facing a huge performance drop with latest git version of xf86-video-ati. Creating and moving windows is very slow. There is also another problem where the mouse cursor is inaccurate positioned by some pixels when it is changed (input cursor, windows resize cursor, ...)
This does not happen with the last released version 7.5.0.

I bisected the driver and found the bad commit before that commit or when it is reverted both problems do not appear:

673e1c7637687c74fc9bdeeeffb7ace0d04b734f is the first bad commit
commit 673e1c7637687c74fc9bdeeeffb7ace0d04b734f
Author: Michel Dänzer <michel.daenzer@amd.com>
Date:   Thu Apr 2 17:54:33 2015 +0900

    Defer initial modeset until the first BlockHandler invocation
    
    This ensures that the screen pixmap contents have been initialized when
    the initial modes are set.
    
    Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=27757
    
    Reviewed-by: Alex Deucher <alexander.deucher@amd.com>

Kind regards
Jan Burgmeier
Comment 1 Jan Burgmeier 2015-10-23 09:34:38 UTC
Created attachment 119135 [details]
Output of lspci -v
Comment 2 Jan Burgmeier 2015-10-23 09:34:51 UTC
Created attachment 119136 [details]
Xorg server logfile
Comment 3 Michel Dänzer 2015-10-23 09:46:10 UTC
Does this still happen with http://cgit.freedesktop.org/xorg/driver/xf86-video-ati/commit/?id=548e97b3b7d1e94075a54ca2bb4eb683025098a7 ?

If yes, what desktop environment are you running?
Comment 4 Jan Burgmeier 2015-10-23 10:01:03 UTC
No it is not fixed.

The desktop environment is gtk2 based Mate 1.6.0
Comment 5 Michel Dänzer 2015-10-27 08:48:48 UTC
MATE 1.10.2 seems to work fine for me, not sure what's wrong on your end. :( Do you have compositing enabled in the MATE settings?

Can you also attach the output of xrandr when the problem occurs, the patch you use for reverting the bisected commit, and the Xorg log file from running with that patch?
Comment 6 Jan Burgmeier 2015-10-28 08:13:22 UTC
Created attachment 119242 [details]
xrandr --verbose output while running with git master (problem occures)
Comment 7 Jan Burgmeier 2015-10-28 08:14:23 UTC
Created attachment 119243 [details]
patch to revert commit 673e1c76.
Comment 8 Jan Burgmeier 2015-10-28 08:17:37 UTC
Created attachment 119244 [details]
Xorg log while running with patch to revert commit 673e1c76.
Comment 9 Martin Peres 2019-11-19 07:53:23 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/xorg/driver/xf86-video-ati/issues/142.


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.