Bug 92573 - Slow mouse movements on rotated screen with dual GPU setup (setprovideroutputsource)
Summary: Slow mouse movements on rotated screen with dual GPU setup (setprovideroutput...
Status: RESOLVED MOVED
Alias: None
Product: xorg
Classification: Unclassified
Component: Server/General (show other bugs)
Version: unspecified
Hardware: Other Linux (All)
: medium enhancement
Assignee: Xorg Project Team
QA Contact: Xorg Project Team
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2015-10-21 15:10 UTC by Jan Burgmeier
Modified: 2018-12-13 22:34 UTC (History)
1 user (show)

See Also:
i915 platform:
i915 features:


Attachments
Xorg server logfile (281.13 KB, text/plain)
2015-10-21 15:10 UTC, Jan Burgmeier
no flags Details
dmesg output (50.31 KB, text/plain)
2015-10-21 15:10 UTC, Jan Burgmeier
no flags Details
radeon: Pass TRUE to drmmode_set_desired_modes for GPU screens (506 bytes, patch)
2015-10-27 08:44 UTC, Michel Dänzer
no flags Details | Splinter Review

Description Jan Burgmeier 2015-10-21 15:10:17 UTC
Created attachment 119040 [details]
Xorg server logfile

Hi,

My setup is the following:

PC: HP T620 Plus with two AMD graphics card
Monitors: 3 Dell 4k Monitos running at 1920x1080. Two are attached to the primary card and one to the secondary card.

When moving the mouse inside a window (firefox, xterm, ...) on a rotated screen on it is very slow. Moving the mouse outside a windows on the background is fast as it should be. The mouse also disappears when it is not moving.

The following versions were used:
Xorg-1.17.2
xf86-video-ati latest git version (ce9914a)
mesa-11.0.0
libdrm-2.4.64
libLLVM-3.7

It does not matter if using DRI2 or DRI3.
It does not matter if ColorTiling2D is on or off.

Kind regards
Jan Burgmeier
Comment 1 Jan Burgmeier 2015-10-21 15:10:44 UTC
Created attachment 119041 [details]
dmesg output
Comment 2 Alex Deucher 2015-10-21 15:20:55 UTC
Prime uses software cursors which are slower than hw cursors.  I believe there were some patches to fix this in the X server, but I'm not sure they were ever committed.
Comment 3 Jan Burgmeier 2015-10-21 15:56:39 UTC
Thanks for the info. I found the following patch series on the mailing list: http://lists.x.org/archives/xorg-devel/2015-June/046628.html 
As far as I could all of them except the HW cursor patch are included in git and will be released with 1.18. I try to backport them to 1.17 and see if the problem is fixed.
Comment 4 Michel Dänzer 2015-10-22 02:51:03 UTC
Note that if you're talking about the actual speed of the cursor movement, that sounds like an input issue rather than anything graphics related.
Comment 5 Michel Dänzer 2015-10-22 02:53:56 UTC
Might be fixed by http://lists.x.org/archives/xorg-devel/2015-June/046633.html ?
Comment 6 Michel Dänzer 2015-10-22 02:54:19 UTC
Sorry, wrong bug.
Comment 7 Jan Burgmeier 2015-10-22 13:26:36 UTC
This bug is fixed with the following patch:

http://lists.x.org/archives/xorg-devel/2015-June/046634.html

But I think it is not merged in current git master. Is there a reason for it?
Comment 8 Jan Burgmeier 2015-10-22 13:59:17 UTC
Some more testing showed that the cursor position is not accurate when the mouse cursor is changed (e.x: input cursor). The changed cursor top positon is at the bottom of the old cursor.
Comment 9 Jan Burgmeier 2015-10-23 09:51:13 UTC
Without the patch noted in comment #7 and reverted commit 673e1c7637687c74fc9bdeeeffb7ace0d04b734f (see other bug report: https://bugs.freedesktop.org/show_bug.cgi?id=92632). All problems are gone. No perfermance issues, no hidden cursor and the position with other cursor bitmaps is also fine.
Comment 10 Michel Dänzer 2015-10-27 08:44:43 UTC
Created attachment 119216 [details] [review]
radeon: Pass TRUE to drmmode_set_desired_modes for GPU screens

Does this xf86-video-ati patch help by any chance?
Comment 11 Jan Burgmeier 2015-10-28 08:18:31 UTC
Maybe this is related or a dublicate of Bug https://bugs.freedesktop.org/show_bug.cgi?id=92632 because the problem is also solved when this commit is reverted.

If wished I can further test it. Should I apply this patch against git master? And with or without the xorg patch for PRIME HW cursor support
Comment 12 GitLab Migration User 2018-12-13 22:34:15 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/xserver/issues/487.


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.