Sometimes various windows content (mostly Firefox windows but others too though I don’t recall which others there where) on my Manjaro KDE laptop kind of hangs and jumps around between various past states. Moving / resizing those windows makes them show their current content again, but as soon as they’re left still they again only display their last few states from when it was still moved / resized.
The only workaround I’ve found is closing and reopening the window. I haven’t identified a trigger that causes a window to adapt this behavior.
Has anybody else experienced something like this and maybe knows how to fix it? Or is there anybody with knowledge of KDE or it’s for this error description relevant components who can tell me how to debug this problem?
did you try different kernels? in your case, since you have a newer hardware, install the 6.0 kernel - not the rc one - from system settings/kernel, reboot… it should automatically boot with it…
if you cant boot for what ever reason with the 6 kernel, just select the 5.15 in grub menu …
okey. I switched to kernel 6.0.2-2 and rebooted. Works fine, and I haven’t seen the issue so far, though it wasn’t super frequent before either. So I’ll have to use the system for a few hours to see if that really fixed it.
I’ll update this post later.
UPDATE: I just had a very similar problem with a terminal (=“konsole”) window, it jittered around though the jittering states seemed to all be coming from the last few seconds, unlike before where the jittering states remained the same old states until the window was moved or resized.
we will need logs, so open this file: kate /etc/environment
and put there this line: QT_LOGGING_RULES='*=false'
save the file
this will disable those qt/kwin logs, so that the logs are more readable
reboot and test, and when it happens again provide logs: journalctl -b0 -p5 --no-pager
Today around ~16:00 I’ve observed a similar issue, I was moving the eclipse startup splash window while it got closed which resulted in it’s rendering getting “stuck” on all my virtual displays, never closing, even after closing eclipse.
I’ve restarted the plasmashell @ 17:05, but that didn’t get rid of that stuck splash screen.
It can’t be interacted with anymore, I can put other windows over it but if I minimize everything it hovers over the desktop but all mouse events get handled by the desktop behind it.
The logs you’ve asked for don’t seem to contain anything related:
Nov 13 15:31:47 host systemd-xdg-autostart-generator[1336822]: /home/user/.config/autostart/org.moson.matray.desktop: not generating unit, error parsing Exec= line: No such file or directory
Nov 13 15:31:47 host systemd[1345]: /run/user/1000/systemd/generator.late/app-xbindkeys@autostart.service:18: WorkingDirectory= path is not absolute, ignoring:
Nov 13 15:42:45 host firefox[8864]: g_object_ref: assertion 'G_IS_OBJECT (object)' failed
Nov 13 15:42:47 host kernel: audit: type=1334 audit(1668350567.211:328): prog-id=33 op=LOAD
Nov 13 15:42:47 host kernel: audit: type=1334 audit(1668350567.211:329): prog-id=34 op=LOAD
Nov 13 15:42:47 host kernel: audit: type=1334 audit(1668350567.211:330): prog-id=35 op=LOAD
Nov 13 15:42:47 host kernel: audit: type=1130 audit(1668350567.308:331): pid=1 uid=0 auid=4294967295 ses=4294967295 subj=unconfined msg='unit=systemd-hostnamed comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success'
Nov 13 15:42:58 host firefox[8864]: g_object_ref: assertion 'G_IS_OBJECT (object)' failed
Nov 13 15:43:11 host firefox[8864]: g_object_ref: assertion 'G_IS_OBJECT (object)' failed
Nov 13 15:43:24 host firefox[8864]: g_object_ref: assertion 'G_IS_OBJECT (object)' failed
Nov 13 15:43:55 host kernel: audit: type=1131 audit(1668350635.665:332): pid=1 uid=0 auid=4294967295 ses=4294967295 subj=unconfined msg='unit=systemd-hostnamed comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success'
Nov 13 15:43:55 host kernel: audit: type=1334 audit(1668350635.761:333): prog-id=0 op=UNLOAD
Nov 13 15:43:55 host kernel: audit: type=1334 audit(1668350635.761:334): prog-id=0 op=UNLOAD
Nov 13 15:43:55 host kernel: audit: type=1334 audit(1668350635.761:335): prog-id=0 op=UNLOAD
Nov 13 15:44:30 host pamac-tray-plas[1772]: updates_checker.vala:71: check updates
Nov 13 15:44:36 host pamac-tray-plas[1772]: updates_checker.vala:101: 2 updates found
Nov 13 16:44:30 host pamac-tray-plas[1772]: updates_checker.vala:71: check updates
Nov 13 16:44:36 host pamac-tray-plas[1772]: updates_checker.vala:101: 2 updates found
Nov 13 17:05:30 host systemd[1345]: plasma-plasmashell.service: Consumed 52min 19.111s CPU time.
I had 2 or 3 freezes, where the mouse was still moveable but nothing reacted to mouse or keyboard. I was able to switch to TTY2 and back once, but afterwards not even that worked anymore. The system was still running, ssh-server worked and accepted connections.
no, it’s not just eclipse. but it happens very rarely.
logging out + in will close all my open windows, which is something I don’t like to do too often, and especially not right now while I have an ongoing long running gui based task now.
Sadly, now that other issue I was talking about yesterday hit again, the screen froze, only the mouse pointer still moves. clicking doesn’t do anything. SSH still works, but the desktop is toast.
Do you have any ideas how to fix this? Years ago it always helped to kill and restart plasmashell, but that trick doesn’t help me now sadly.
looking at journalctrl -e via ssh doesn’t give me anything. The freeze happened somewhere in between me opening kate:
ov 14 14:11:06 xps15-7590 rtkit-daemon[1440]: Supervising 14 threads of 11 processes of 1 users.
Nov 14 14:11:40 xps15-7590 rtkit-daemon[1440]: Supervising 14 threads of 11 processes of 1 users.
Nov 14 14:11:40 xps15-7590 rtkit-daemon[1440]: Supervising 14 threads of 11 processes of 1 users.
Nov 14 14:11:54 xps15-7590 systemd[1345]: Started Kate - Advanced Text Editor.
Nov 14 14:11:55 xps15-7590 systemd[1345]: Started Kate - Advanced Text Editor.
Nov 14 14:12:40 xps15-7590 rtkit-daemon[1440]: Supervising 14 threads of 11 processes of 1 users.
Nov 14 14:12:40 xps15-7590 rtkit-daemon[1440]: Supervising 14 threads of 11 processes of 1 users.
and me trying if using the fn keys to increase / decrease sound volume & screen brighthness still work (which they do):
Nov 14 14:13:30 xps15-7590 kernel: ACPI BIOS Error (bug): Could not resolve symbol [\_SB.PCI0.PEG0.PEGP.BRT6.LCD], AE_NOT_FOUND (20220331/psargs-330)
Nov 14 14:13:30 xps15-7590 kernel: ACPI Error: Aborting method \_SB.PCI0.PEG0.PEGP.BRT6 due to previous error (AE_NOT_FOUND) (20220331/psparse-529)
Nov 14 14:13:30 xps15-7590 kernel: ACPI Error: Aborting method \EV5 due to previous error (AE_NOT_FOUND) (20220331/psparse-529)
Nov 14 14:13:30 xps15-7590 kernel: ACPI Error: Aborting method \SMEE due to previous error (AE_NOT_FOUND) (20220331/psparse-529)
At that moment where that freeze which was fixed by restarting kwin_x11 happened (~ around 14:13 today) the command you suggest journalctl -b0 -p4 --no-pager gives less than what I posted at 14:35 today.
Do you want me to post a pastebin of the whole log, which starts at Nov 12 19:44:20, which apparently was the last time I rebooted my laptop?
Restarting kwin_x11 by the way also got rid of that hung eclipse flash screen.
Today I had another complete desktop hang. I was able to get the windows responsive to mouse actions again by restarting kwin_x11 but the plasma-panels didn’t work anymore and nothing reacted to keyboard input. I’ve additionally restarted plasmashell, but that didn’t help either.
Didn’t have too much time to debug, therefore I just rebooted the machine.
i use prime-run to start gpu intensive tasks like games, though I don’t do that often, something like once a month? Though probably with several months in between usages but if I use it then repeatably for days / weeks on end. I did install unigine-superposition about a week ago to see if prime-run is still working as it did: it does.
What else is there to restart when KDE has problems?
I now have quite frequent hangs of the dekstop and my panels (like once every 2 days) that can’t be fixed by restarting kwin_x11 or plasmashell.