Bug 458936 - Stability issues related to notifications (or images in them?) on debian testing/unstable
Summary: Stability issues related to notifications (or images in them?) on debian test...
Status: RESOLVED WORKSFORME
Alias: None
Product: plasmashell
Classification: Plasma
Component: Notifications (show other bugs)
Version: 5.25.4
Platform: Debian unstable Linux
: NOR major
Target Milestone: 1.0
Assignee: Plasma Bugs List
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2022-09-09 20:25 UTC by spam
Modified: 2023-05-27 03:45 UTC (History)
4 users (show)

See Also:
Latest Commit:
Version Fixed In:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description spam 2022-09-09 20:25:18 UTC
SUMMARY
***
NOTE: If you are reporting a crash, please try to attach a backtrace with debug symbols.
See https://community.kde.org/Guidelines_and_HOWTOs/Debugging/How_to_create_useful_crash_reports
***
possibly related to https://bugs.kde.org/show_bug.cgi?id=457847

If you leave system working while notifications are happening there's a chance that plasmashell will lock up, prolonging this will cause whole system to mostly lock up. It seems that some (rogue? broken icon?) notification may cause plasmashell to lockup entirely and take some other processes (i.e. ones generating notifications) talking to it a hostage.

STEPS TO REPRODUCE
1. Have something generating notifications preferably with custom icons (i.e. microblogging pleroma-based websites, IM like matrix or slack).
2. Make sure you actually have plasmashell notifications appearing on screen (i.e. there's notification widget and DND isn't turned on)
3. Go on about your day

OBSERVED RESULT

Eventually, plasmashell will lockup and become unresponsive (especially noticeable if you have seconds in time widget), there will be an artifact like a black rectange or hall-of-mirrors effect or similar depending whether composition is on or off. Killing plasmashell helps but has high chance of crashing whatever was generating notifications (i.e. chromium). If this state persists, system will most likely lock up (cpu usage spike? kwin locking up?) stop responding to (newly plugged?) USB keyboards/mice, trying to switch to TTY yields black screen with non-blinking cursor. If system goes to sleep in this state it will be still broken when woken up but lock screen seems to work fine.

EXPECTED RESULT

Your day is just about the same as before

SOFTWARE/OS VERSIONS
Windows: 
macOS: 
Linux/KDE Plasma: kernel 5.18.0-4-amd64
(available in About System)
KDE Plasma Version: 5.25.4
KDE Frameworks Version:  5.97.0
Qt Version: 5.15.4

ADDITIONAL INFORMATION

Using X11 (kwin-x11).

I have two devices where this happens - a dell laptop with intel graphics and amd desktop, and this happens on both of them, so I don't think it's a MESA issues, not to mention this happens with and without compositing. However, turning DND mode on helped tremendously - the notifications that go through are "critical" ones and on my system those are from psi-notify daemon, which doesn't use icons and system is much more stable, even with sleeps. 

My guess is that notification happens, most likely coming from microblogging or IM, which tends to use user avatar as an icon, icon fails to render properly and plasma doesn't know how to handle it and locks up, while it's locked up it cannot respond to other requests to notifications and/or possibly leaks memory. Debugging it in this state is near impossible, but I managed to ssh into locked system once and had to kill/restart sddm to make it work again, but usually the only solution is a hard-reset with a hardware button. No crash dumps for plasmashell in /var/crash only chromium crashes there, but next time it happens i might try to get some + get some logs. Nothing out of ordinary in .xsession-errors either.
Comment 1 dmitry.myachin 2022-11-09 13:07:20 UTC
have same problem in Arch Linux

Linux ummini 6.0.7-arch1-1 #1 SMP PREEMPT_DYNAMIC Thu, 03 Nov 2022 18:01:58 +0000 x86_64 GNU/Linux
plasmashell 5.26.3
Comment 2 spam 2022-11-09 17:56:37 UTC
this seem to be only happening on X11, not Wayland
Comment 3 Aaron Rainbolt 2023-04-13 19:37:19 UTC
This is happening to me on Plasma 5.27.4 on Kubuntu 22.04 (using an experimental backports repo to get the newer version of Plasma). Usually it takes out kwin_x11 first, then Plasmashell shortly thereafter. I have to kill and relaunch both to get things stable again - just restarting kwin_x11 usually results in Plasmashell locking up later on.

Enabling permanent Do Not Disturb works around the issue for me, but is less than ideal.
Comment 4 Aaron Rainbolt 2023-04-13 19:45:48 UTC
(Also I believe Element (a Matrix chat client) is good at causing this problem on my system.)
Comment 5 Nate Graham 2023-04-27 23:28:03 UTC
Can you get a backtrace of what Plasma is doing when it's hung in this state?

1. Run gdb attach -p $(pidof plasmashell) in a terminal window while it's hung
2. If it says "--Type  for more, q to quit, c to continue without paging--" at the bottom of the terminal window, press the "c" key.
3. When you see a prompt that says "(gdb)", type "bt" and press the enter key.
4. If it says "--Type  for more, q to quit, c to continue without paging--" at the bottom of the terminal window, press the "c" key.
5. Copy-and-paste the contents of the terminal window into a comment here.

Thanks!
Comment 6 Bug Janitor Service 2023-05-12 03:46:06 UTC
Dear Bug Submitter,

This bug has been in NEEDSINFO status with no change for at least
15 days. Please provide the requested information as soon as
possible and set the bug status as REPORTED. Due to regular bug
tracker maintenance, if the bug is still in NEEDSINFO status with
no change in 30 days the bug will be closed as RESOLVED > WORKSFORME
due to lack of needed information.

For more information about our bug triaging procedures please read the
wiki located here:
https://community.kde.org/Guidelines_and_HOWTOs/Bug_triaging

If you have already provided the requested information, please
mark the bug as REPORTED so that the KDE team knows that the bug is
ready to be confirmed.

Thank you for helping us make KDE software even better for everyone!
Comment 7 Bug Janitor Service 2023-05-27 03:45:41 UTC
This bug has been in NEEDSINFO status with no change for at least
30 days. The bug is now closed as RESOLVED > WORKSFORME
due to lack of needed information.

For more information about our bug triaging procedures please read the
wiki located here:
https://community.kde.org/Guidelines_and_HOWTOs/Bug_triaging

Thank you for helping us make KDE software even better for everyone!