https://bugs.kde.org/show_bug.cgi?id=523285

--- Comment #3 from Oleksandr <[email protected]> ---
Clarification: there is no crash, so no backtrace or coredump exists.

- coredumpctl list plasmalogin-helper → "No coredumps found"
- systemctl status plasmalogin after reproduction shows the service active
(running), started cleanly (ExecStartPre → code=exited, status=0/SUCCESS)

The summary described the recovery state, not the defect (I have since renamed
it). The actual bug:

- Restarting plasmalogin.service leaves a black screen on the active VT.
- The greeter does start, but only renders on a secondary VT (reachable via
Ctrl+Alt+F2).
- Switching back to the original VT returns to the black screen; it is never
reclaimed.

So the greeter process starts correctly but never takes over the active VT
after a service restart. This is a VT/session-handoff issue, not a crash, which
is why the standard backtrace workflow does not apply.

See the attached file plasmalogin-crash.log for the full journal from a clean
reproduction (fresh boot, then systemctl restart plasmalogin). The relevant
lines:

лип 21 13:09:55 plasmalogin[1804]: Auth: plasmalogin-helper exited with 255
лип 21 13:10:01 plasmalogin[6475]: Auth: plasmalogin-helper exited with 64
лип 21 13:10:01 plasmalogin-helper[6817]:
pam_unix(plasmalogin-greeter:session): session opened for user plasmalogin
лип 21 13:10:02 plasmalogin-helper[6817]: Starting Wayland user session:
".../startplasma-login-wayland"
лип 21 13:10:40 plasmalogin-helper[7210]: pam_unix(plasmalogin:session):
session opened for user admin
лип 21 13:10:40 plasmalogin-helper[7210]: Starting Wayland user session:
".../startplasma-wayland"
лип 21 13:10:45 plasmalogin[6475]: Auth: plasmalogin-helper exited with 255

plasmalogin-helper exits 255 on the restart/teardown path and again at
13:10:45, with a 64 in between - all clean error exits, no signal, no coredump.
The greeter session and my subsequent login (13:10:40, on the F2 VT) open
normally, so authentication and session startup work; the failure is around the
helper's exit / VT handoff.

Environment: Wayland session, AMD (amdgpu, RX 6800 XT), plasma-workspace 6.7.2
(Nobara/Fedora 44). plasma-workspace debuginfo is installed. Since nothing
crashes there is no coredump to capture, but I am happy to enable verbose
plasmalogin logging if you tell me which categories would help.

Full untruncated log with complete PAM output attached to the bug as
plasmalogin-crash.log.

-- 
You are receiving this mail because:
You are watching all bug changes.

Reply via email to