https://bugs.kde.org/show_bug.cgi?id=526801

            Bug ID: 526801
           Summary: Resident but idle plasma-shutdown keeps owning
                    org.kde.Shutdown, so the launcher's Shut Down entry
                    silently does nothing
    Classification: Plasma
           Product: plasmashell
      Version First 6.7.5
       Reported In:
          Platform: Fedora RPMs
                OS: Linux
            Status: REPORTED
          Severity: normal
          Priority: NOR
         Component: Session Management
          Assignee: [email protected]
          Reporter: [email protected]
                CC: [email protected]
  Target Milestone: 1.0

**Summary:** a resident but idle `plasma-shutdown` keeps owning
`org.kde.Shutdown`,
so `ksmserver-logout-greeter` refuses to start and *Shut Down* in the
Application
Launcher silently does nothing.

## Environment

- Fedora 44, KDE Plasma 6.7.5 (`plasma-workspace-6.7.5-1.fc44`)
- Wayland session, systemd user session (not remote, active, `CanPowerOff` =
yes)
- No inhibitors: `systemd-inhibit --list` only shows sleep-related `delay`
locks
  plus PowerDevil's key-handling `block`.

## Steps to reproduce

1. Log in to a Plasma 6.7.5 Wayland session.
2. Get `org.kde.Shutdown` D-Bus-activated **without** calling any method on it.
   A reliable trigger is launching the input method selector `im-chooser` 1.7.6
   (Fedora package `im-chooser`): `libimchooseui.so` creates a `GDBusProxy` for
   `org.kde.Shutdown` / `/Shutdown` with `flags = 0`
(`G_DBUS_PROXY_FLAGS_NONE`)
   and then only sends `org.freedesktop.DBus.Peer.Ping`. Because
   `G_DBUS_PROXY_FLAGS_DO_NOT_AUTO_START` is not set, creating the proxy starts
   the service.
3. Confirm the service is now resident and idle:
   `systemctl --user status 'dbus-:[email protected]'`
   → `Active: active (running)` for tens of minutes, `Main PID` =
`plasma-shutdown`,
   cumulative CPU time a few tens of ms.
4. Click *Shut Down* in the Application Launcher (Kickoff).

## Actual result

Nothing happens at all: no confirmation dialog, no error message, no request to
logind. Repeating the click does not help. The user journal shows, once per
click:

```
systemd[1325]: Started dbus-:[email protected].
ksmserver-logout-greeter[496120]: org.kde.Shutdown D-Bus service is already
registered; quitting
... (repeated at 11:10:23, 11:10:28, 11:10:55, 11:10:57, 11:11:00, 11:12:11)
plasma_waitforname[497561]: WaitForName: Service was not registered within
timeout
```

In this boot, `plasma-shutdown` was started exactly once (10:23:11) and never
exited. In earlier sessions the same unit was started and stopped within
seconds
while an actual logout/shutdown was in progress, i.e. residency is not
intended.

Workaround that restores the button without rebooting:

```bash
systemctl --user stop 'dbus-:[email protected]'
```

## Expected result

Clicking *Shut Down* shows the confirmation dialog and shuts down / reboots.
Specifically, one or both of:

1. `plasma-shutdown` should exit when it has no work to do (it is a
   D-Bus-activated helper, so nothing should keep it alive indefinitely), and
2. `ksmserver-logout-greeter` should not treat mere name ownership as "a logout
   is already in progress" — it should check whether an actual session-end
request
   is pending, and otherwise take over the name.

## Additional information

- `org.kde.Shutdown` is declared activatable in
  `/usr/share/dbus-1/services/org.kde.Shutdown.service`
(`Exec=/usr/bin/plasma-shutdown`),
  providing `logout`, `logoutAndReboot`, `logoutAndShutdown`, `saveSession`.
- Because `plasma-shutdown` provides *logout* as well, this is not limited to
  shutdown: any unfulfilled logout path through that name can wedge the
session's
  leave/session-management UI.
- Triggering a third-party probe like this also happens at login time on some
  systems, so the failure mode may look intermittent.
- Related reports: bug 477056, bug 488277 (`plasmashell / Session Management`);
  bug 525919 (`Discover`: "KWin failed to complete logout" after
  `plasma-shutdown` was invoked).

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

Reply via email to