Bug 507934

Thanks for the bug report, and the plasmashell logs.

As the message at the top of the summary says, if something crashed, we need a backtrace of it so we can figure out what's going on. Can you please attach a backtrace of the crash using the coredumpctl command-line program, as detailed in https://community.kde.org/Guidelines_and_HOWTOs/Debugging/How_to_create_useful_crash_reports#Retrieving_a_backtrace_using_coredumpctl ?

Thanks.

A backtrace isn't going to help you diagnose a protocol error.

The issue is related to Qt 6.10

(In reply to TraceyC from comment #1)
> Thanks for the bug report, and the plasmashell logs.
> 
> As the message at the top of the summary says, if something crashed, we need
> a backtrace of it so we can figure out what's going on. Can you please
> attach a backtrace of the crash using the coredumpctl command-line program,
> as detailed in
> https://community.kde.org/Guidelines_and_HOWTOs/Debugging/
> How_to_create_useful_crash_reports#Retrieving_a_backtrace_using_coredumpctl ?
> 
> Thanks.

I don't have a backtrace because I had coredumps disabled. If it happens again I'll attach it though. For some reason the KDE Crash tool doesn't trigger for Plasma Shell. I'm not sure why (probably an issue on my end somewhere. I do have KCrash installed)

Created attachment 186273 [details]
output from: env WAYLAND_DEBUG=1 plasmashell --replace 2>&1 | lz4c -c0 - plasma-shell.log.lz4

I'm also getting the same issue. My plasmashell log is attached. It doesn't seem easy to reproduce. Sometimes I can go hours before it reoccurring again. The easiest way for me to reproduce so far is by switching between activities or interacting with the task manager panel (either via clicking on the application launcher widget on the panel, clicking on the system tray widget or interacting with the icon-only task manager).

Details about my system:
Operating System: Arch Linux 
KDE Plasma Version: 6.5.0
KDE Frameworks Version: 6.19.0
Qt Version: 6.10.0
Kernel Version: 6.12.55-1-lts (64-bit)
Graphics Platform: Wayland
Processors: 16 × AMD Ryzen 7 5700X 8-Core Processor
Memory: 64 GiB of RAM (62.7 GiB usable)
Graphics Processor: NVIDIA GeForce RTX 4070

Created attachment 186274 [details]
gdb attach

To add to my previous comment, attached is a gdb log from attaching to the plasmashell process. But there really isn't anything in there. When the "xdg_surface#353: error 3: attached a buffer before configure event" happens, the plasmashell simply dies with no exception or anything, and therefore there is no ability to see the backtrace.

I'm not sure where to go from here but I'm happy to get some pointers to further help investigate this issue.

Created attachment 186314 [details]
gdb break on _exit()

Attached is a gdb log from breaking on _exit().

I think I could reproduce it pretty easily now. The trick is to press and hold the shortcuts configured for switching activities. Doing so switches the activities pretty quickly and at some point (typically not more than 10 seconds in my experience from the onset of holding the shortcut keys for switching activities), it will cause plasmashell to die with an exit code of -1.

Operating System: Arch Linux 
KDE Plasma Version: 6.5.1
KDE Frameworks Version: 6.19.0
Qt Version: 6.10.0
Kernel Version: 6.12.56-1-lts (64-bit)
Graphics Platform: Wayland
Processors: 16 × AMD Ryzen 7 5700X 8-Core Processor
Memory: 64 GiB of RAM (62.7 GiB usable)
Graphics Processor: NVIDIA GeForce RTX 4070

Weirdly, I can't reproduce the issue by doing that.

(In reply to Nate Graham from comment #7)
> Weirdly, I can't reproduce the issue by doing that.

Thanks for looking into this, Nate!

It bears mentioning that, all this while, I had explicit sync disabled with __NV_DISABLE_EXPLICIT_SYNC=1 to work around a separate fd leak bug in the nvidia driver: https://forums.developer.nvidia.com/t/fd-leak-with-explicit-sync-and-kde-plasma/317293/26

With explicit sync enabled, I couldn't reproduce the issue as well.

On my side I'm using AMD hardware. I haven't seen these crashes in a while now though so maybe something changed in Mesa. I don't know.

Using NVidia without explicit sync is not a supported configuration, and completely expected to cause bad problems.