Since upgrading to rocky 9.6 from 9.5 (fortunately only on a trial system) the display has become unusably slow, e.g. on moving mouse from one window to another it takes 5 seconds to register it’s in a different window, on typing a second or two goes by after each keypress to register on the display…

This issue has now been fixed by Jocelyn Falempe of RedHat, by reverting a patch which had ignored which parts of a window needed refreshing and refreshed the whole lot every time. The fixed mesa packages are version 24.2.8-3.el9_6 - I have copies if needed by anybody. However, the patch was needed…

[image] rfs: Xorg typically used around 1% of a cpu at most when the screen was static with no mouse or keyboard activity, on 9.6 it never falls below 30%, and if there is any mouse movement or keyboard activity it goes up to 99% and sticks there until everything is still again. is there anyt…

Any hardware information? Is this an SBC with strange graphics hardware or something?

Thank you Tony for the sensible question. I could find nothing helpful in the log files with the exception of a single occurrence of a SELinux AVC complaining that Xorg-wrap couldn’t get sysadmin capability (but I mean single once on one of many boots). There is no Xwrapper.config file saved in /etc…

Here is my attempt to attach ~/.local/share/xorg/Xorg.0.log (but only part is included due to reply length limitation): [ 789.958] (--) Log file renamed from "/home/rfs/.local/share/xorg/Xorg.pid-4522.log" to "/home/rfs/.local/share/xorg/Xorg.0.log" [ 789.958] X.Org X Server 1.20.11 X Protocol…

Thank you Brian likewise. No, it’s an Intel server - entirely normal apart from being a tad old. The video chip is mgag200se; this is driven by the mgag200 kernel module. I can add a bit more info gleaned from further experiments: Unplugging mouse and keyboard makes no difference. I think that th…

[image] rfs: I think that the occurrence of long strings of false-input fffffffff is due to the system registering keypress of the f key but not registering the keyrelease. The only place I see long strings of f’s is in this “modeset” line, that would not indicate a keyboard input “790.156] …

I can’t see why it would matter since your setup apparently works fine with 9.5, but have you tried a different keyboard? I don’t think it’ll fix it but it’ll take you ten seconds to try it.

Tony, I think that the occurrence of long strings of false-input fffffffff is due to the system registering keypress of the f key but not registering the keyrelease. Misunderstanding here. I’m not referring to (non-existent) strings of false-input fffff in the log file. What’s happening is th…

Frank, I can’t see why it would matter since your setup apparently works fine with 9.5, but have you tried a different keyboard? I don’t think it’ll fix it but it’ll take you ten seconds to try it. Might be true if I had one, but I haven’t. Thanks, Roger.

[image] rfs: sometimes when I type an f, or an F, or a f, or a f, I get long strings of ordinary fffffff appearing in whatever window has the input focus Let me understand this clearly (because I misunderstood what you wrote earlier): of ALL the keys on the keyboard, ONLY the key “F” (shift…

Crushingly slow display since upgrade to 9.6

Rocky Linux Help & Support

gerry666uk August 4, 2025, 6:17pm 18

Is there a bugzilla report about this?

RL9.6_update and NVIDIA nvidia-570.172.08 no wayland

Topic		Replies	Views
Bug in i915 video driver? Rocky Linux Help & Support	6	3056	November 13, 2022
X doesn't start, black screen after logging on through GDM Rocky Linux Help & Support	0	1753	March 7, 2023
RL9.6_update and NVIDIA nvidia-570.172.08 no wayland Rocky Linux Help & Support	55	1324	October 25, 2025
GNOME/Wayland crashes after clean v9.4 Workstation installation Rocky Linux Help & Support rocky-linux-9	9	1048	August 16, 2024
Change Display Resolution i Rocky Linux 9 Rocky Linux Help & Support	16	8956	November 30, 2022

Crushingly slow display since upgrade to 9.6

Related topics