I’m seeing a shutdown-time plasmashell crash on Arch Linux with KDE Plasma 6.7.3 and Wayland.
Important clarification: the system power-off is requested first by KDE’s `plasma-shutdown` helper. The plasmashell crash happens afterward, while the desktop is shutting down. So I do not think plasmashell is causing the computer to power
off; rather, it appears to crash during teardown (the orderly stopping of desktop processes).
Environment:
- Arch Linux, kernel 7.1.4-arch1-1
- KDE Plasma 6.7.3
- Wayland
- Dual NVIDIA RTX 5060 Ti GPUs
Confirmed July 28 sequence:
08:45:43:
systemd-logind: poweroff requested from client PID 428054 (‘plasma-shutdown’)
08:45:43:
plasmashell: Akonadi server running without control process
plasmashell: QLocalSocket: Remote closed …
08:45:43:
panel.colorizer: TypeError: Cannot read property ‘cfg’ of null
08:45:45:
kernel: plasmashell[4924]: segfault at 108 … in libPlasma.so.6.7.3
systemd: plasma-plasmashell.service: Main process exited, code=dumped, status=11/SEGV
A second shutdown on July 31 had the same ordering: `plasma-shutdown` requested power-off first, followed by Akonadi socket closures, panel-colorizer cleanup errors, and process-destruction messages. That second event did not produce a
plasmashell coredump.
The crash address is `0x108`, which looks consistent with a null-pointer dereference (attempting to access data through a missing object).
I also have a separate issue where newly opened windows sometimes appear smaller rather than maximized. I do not yet know whether that is related, so I am treating it as separate unless there is a known connection.
Questions:
- Is this a known Plasma 6.7.3 shutdown-time crash in libPlasma?
- Has anyone seen a `plasmashell` SIGSEGV after `plasma-shutdown` initiates power-off?
- Could a third-party panel widget’s teardown behavior contribute to this, even though the `cfg` error is logged after shutdown starts?
- Would upgrading Plasma be the recommended first test?
I can provide the relevant journal excerpts and coredump metadata if useful.