Bug 505760

Summary: Plasmashell Crashes and Restarts When Playing Music
Product: [Plasma] plasmashell Reporter: Claire <accounts>
Component: generic-crashAssignee: Plasma Bugs List <plasma-bugs-null>
Status: RESOLVED DOWNSTREAM    
Severity: crash CC: kde, nate
Priority: HI Keywords: drkonqi
Version First Reported In: 6.3.5   
Target Milestone: 1.0   
Platform: Fedora RPMs   
OS: Linux   
Latest Commit: Version Fixed In:
Sentry Crash Report: https://crash-reports.kde.org/organizations/kde/issues/197923/events/8c2d41bdef244ed1bbade00e4aea1c46/
Attachments: New crash information added by DrKonqi

Description Claire 2025-06-19 00:56:37 UTC 
Application: plasmashell (6.3.5)

ApplicationNotResponding [ANR]: false
Qt Version: 6.9.1
Frameworks Version: 6.14.0
Operating System: Linux 6.14.11-300.fc42.x86_64 x86_64
Windowing System: Wayland
Distribution: "Fedora Linux 42 (KDE Plasma Desktop Edition)"
DrKonqi: 6.3.5 [CoredumpBackend]

-- Information about the crash:
When I play music via the Strawberry music player, plasmashell sometimes completely crashes and restarts. Some notable details:

1. This happens once per boot maximum (sometimes it does not happen). After plasmashell recovers, I can listen to music as normal.
2. I have PlasMus Toolbar widget active on my app panel. I suspect there might be something there, given the stack trace shows plasma widgets/addons as well as pipewire.

I've disabled scrolling text and lyrics display for PlasMus Toolbar -- I'm wondering if somehow that was the cause, but I can't be sure. Hopefully this stack trace makes more sense to you than me.

The crash can be reproduced sometimes.

-- Backtrace (Reduced):
#5  std::__atomic_base<QMutexPrivate*>::load (this=0x838b48fff5097430, __m=std::memory_order_relaxed) at /usr/include/c++/15/bits/atomic_base.h:828
#6  std::atomic<QMutexPrivate*>::load (this=0x838b48fff5097430, __m=std::memory_order_relaxed) at /usr/include/c++/15/atomic:587
#7  QAtomicOps<QMutexPrivate*>::loadRelaxed<QMutexPrivate*> (_q_value=<error reading variable: Cannot access memory at address 0x838b48fff5097430>) at /usr/src/debug/qt6-qtbase-6.9.1-1.fc42.x86_64/src/corelib/thread/qatomic_cxx11.h:202
#8  QBasicAtomicPointer<QMutexPrivate>::loadRelaxed (this=0x838b48fff5097430) at /usr/src/debug/qt6-qtbase-6.9.1-1.fc42.x86_64/src/corelib/thread/qbasicatomic.h:170
#9  QBasicMutex::fastTryLock (this=0x838b48fff5097430) at /usr/src/debug/qt6-qtbase-6.9.1-1.fc42.x86_64/src/corelib/thread/qmutex.h:80


Reported using DrKonqi
Comment 1 Claire 2025-06-19 00:56:40 UTC 
Created attachment 182373 [details]
New crash information added by DrKonqi

DrKonqi auto-attaching complete backtrace.
Comment 2 David Edmundson 2025-06-19 08:44:07 UTC 
Trace is weird, it shows mpris stuff running in the DBus thread, not the main thread.

QObjectBindableProperty seems to be doing something very very stupid.

I think we may have to try and repro this in a small case and report to Qt.
Comment 3 Nate Graham 2025-07-29 15:23:30 UTC 
Claire, did disabling "PlasMus Toolbar" help?
Comment 4 Nate Graham 2025-07-31 23:05:37 UTC 
.
Comment 5 Claire 2025-08-01 02:17:13 UTC 
(In reply to Nate Graham from comment #3)
> Claire, did disabling "PlasMus Toolbar" help?

Apologies for the delay. I haven't had the crash even with PlasMus toolbar recently, weirdly enough. I did disable a feature of the PlasMus toolbar though; specifically, there's a feature that causes text to scroll across the panel, showing the name of the song (or sometimes lyrics). Disabling this seems to have caused it to no longer happen, but that could also be coincidence. The toolbar has also updated a few times in the intervening time, so maybe it was fixed?
Comment 6 Nate Graham 2025-08-07 15:48:40 UTC 
If you turn that feature back on, does the issue come back?
Comment 7 Claire 2025-08-08 04:38:01 UTC 
(In reply to Nate Graham from comment #6)
> If you turn that feature back on, does the issue come back?

I can try it out in the morning. Also, apologies, the widget is PlasMusic toolbar. Here is the location: https://store.kde.org/p/2128143
Comment 8 Claire 2025-08-08 17:57:40 UTC 
Seems fixed.
Comment 9 Nate Graham 2025-08-08 18:10:03 UTC 
Great, it sounds like the widget was to blame all along. Until we can sandbox widgets and run them in another process, this kind of issue will always be possible, unfortunately.