Bug 433342

Summary: Plasmashell gets stuck after turning on the monitor
Product: [Plasma] plasmashell Reporter: Leyenda <neopytek>
Component: generalAssignee: David Edmundson <kde>
Status: RESOLVED DUPLICATE    
Severity: normal CC: nate, plasma-bugs
Priority: NOR Keywords: wayland
Version: 5.22.0   
Target Milestone: 1.0   
Platform: Neon   
OS: Linux   
Latest Commit: Version Fixed In:
Sentry Crash Report:
Attachments: Backtrace of dr konqi

Description Leyenda 2021-02-20 17:01:53 UTC
SUMMARY
Plasmashell gets stuck after turning on the display screen, all the panels are closed, only the windows are open but unused and you have to do alt+f2 to kill Plasmashell and run it again.

STEPS TO REPRODUCE
1. Open session on Wayland
2. Open some some Windows
3. Turn your display screen off
4. Turn your display screen on
5. All the panels are closed, only the windows are open and they don't respond, Plasmashell is running

OBSERVED RESULT
All the panels are closed, only the windows are open and they don't respond

EXPECTED RESULT
All the panels on their place and responsive windows

SOFTWARE/OS VERSIONS
Windows: 
macOS: 
Linux/KDE Plasma: 
(available in About System)
KDE Plasma Version:  5.21.0
KDE Frameworks Version: 5.80
Qt Version: 5.15.2

ADDITIONAL INFORMATION
Comment 1 Nate Graham 2021-03-20 01:18:31 UTC
Did Plasma crash? Or hang? Or something else? Are you using a desktop with only one screen?
Comment 2 Leyenda 2021-03-21 14:48:13 UTC
(In reply to Nate Graham from comment #1)
> Did Plasma crash? Or hang? Or something else? Are you using a desktop with
> only one screen?

It is something strange, I use only one screen, if I turn off the screen and turn on again, all the apps crash. I even can't send a bug report cause dr knoqui crashes too. Plasmashell is running but all panels are gone. I have to kill plasmashell process and rerun it with alt+f2.

I have to reinstall my Neon, maybe its a only a particular problem with my system.
Comment 3 Nate Graham 2021-03-21 16:07:29 UTC
Yes, right now all apps die when KWin crashes, so it sounds like that's what's happening. Can you get a backtrace for KWin? Please add ExternalSizeMax=6G" to the file /etc/systemd/coredump.conf, then reboot, then reproduce the crash, then attach the backtrace, receieved using coredumpctl as outlined in https://community.kde.org/Guidelines_and_HOWTOs/Debugging/How_to_create_useful_crash_reports#Retrieving_a_backtrace_using_coredumpctl
Comment 4 Leyenda 2021-03-21 17:17:40 UTC
Added line
rebooted
repeated the steps
and Kwin is not listed on coredumpctl output, only dr Konqui, which produces that bt:

#0  __GI_raise (sig=<optimized out>) at ../sysdeps/unix/sysv/linux/raise.c:50
#1  0x00007f3fb3d3afcc in KCrash::defaultCrashHandler(int) () from /lib/x86_64-linux-gnu/libKF5Crash.so.5
#2  <signal handler called>
#3  __GI_raise (sig=sig@entry=6) at ../sysdeps/unix/sysv/linux/raise.c:50
#4  0x00007f3fb24f5859 in __GI_abort () at abort.c:79
#5  0x00007f3fb2950bf7 in qt_message_fatal (context=..., message=<synthetic pointer>...)
    at global/qlogging.cpp:1914
#6  QMessageLogger::fatal (this=this@entry=0x7ffc17a1b220, 
    msg=msg@entry=0x7f3fb14220b8 "The Wayland connection broke. Did the Wayland compositor die?")
    at global/qlogging.cpp:893
#7  0x00007f3fb1365d65 in QtWaylandClient::QWaylandDisplay::checkError (this=<optimized out>)
    at /usr/include/x86_64-linux-gnu/qt5/QtCore/qlogging.h:90
#8  QtWaylandClient::QWaylandDisplay::checkError (this=<optimized out>) at qwaylanddisplay.cpp:204
#9  0x00007f3fb137716c in QtWaylandClient::QWaylandDisplay::flushRequests (this=0x56329ecf4980)
    at qwaylanddisplay.cpp:222
#10 0x00007f3fb2ba8730 in doActivate<false> (sender=0x56329ed715e0, signal_index=4, argv=0x7ffc17a1b330)
    at ../../include/QtCore/../../src/corelib/kernel/qobjectdefs_impl.h:395
#11 0x00007f3fb2ba1ac7 in QMetaObject::activate (sender=sender@entry=0x56329ed715e0, 
    m=m@entry=0x7f3fb2e0b1c0 <QAbstractEventDispatcher::staticMetaObject>, 
    local_signal_index=local_signal_index@entry=1, argv=argv@entry=0x0) at kernel/qobject.cpp:3946
#12 0x00007f3fb2b6e287 in QAbstractEventDispatcher::awake (this=this@entry=0x56329ed715e0)
    at .moc/moc_qabstracteventdispatcher.cpp:149
#13 0x00007f3fb2bcc01c in QEventDispatcherGlib::processEvents (this=0x56329ed715e0, flags=...)
    at kernel/qeventdispatcher_glib.cpp:430
#14 0x00007f3fb2b701eb in QEventLoop::exec (this=this@entry=0x7ffc17a1b450, flags=..., flags@entry=...)
    at ../../include/QtCore/../../src/corelib/global/qflags.h:141
#15 0x00007f3fb2b78394 in QCoreApplication::exec () at ../../include/QtCore/../../src/corelib/global/qflags.h:121
Comment 5 Nate Graham 2021-03-21 17:50:34 UTC
I'm surprised to hear that there's no kwin_wayland backtrace, because the symptoms you describe line up perfectly with the compositor crashing.

How about plasmashell? Is there a crash log for that?
Comment 6 Leyenda 2021-03-21 17:51:54 UTC
Created attachment 136925 [details]
Backtrace of dr konqi

Dr konqi back trace
Comment 7 Leyenda 2021-03-21 17:56:43 UTC
(In reply to Nate Graham from comment #5)
> I'm surprised to hear that there's no kwin_wayland backtrace, because the
> symptoms you describe line up perfectly with the compositor crashing.
> 
> How about plasmashell? Is there a crash log for that?

No, thereisn't a plasmashell line:

Sun 2021-03-21 18:10:08 CET    2420  1000  1001   6 present   /usr/lib/x86_64-linux-gnu/libexec/drkonqi
Sun 2021-03-21 18:47:03 CET    9073  1000  1001   6 present   /usr/lib/x86_64-linux-gnu/libexec/drkonqi
Sun 2021-03-21 18:48:36 CET    2171  1000  1001   6 present   /usr/bin/kwalletd5
Sun 2021-03-21 18:48:37 CET    1998  1000  1001   6 present   /usr/lib/x86_64-linux-gnu/libexec/kactivitymanagerd
Sun 2021-03-21 18:48:38 CET   10055  1000  1001   6 present   /usr/bin/systemmonitor
Sun 2021-03-21 18:48:38 CET   10160  1000  1001   6 present   /usr/bin/kded5
Sun 2021-03-21 18:48:38 CET   10161  1000  1001   6 present   /usr/lib/x86_64-linux-gnu/libexec/drkonqi
Sun 2021-03-21 18:48:38 CET    2058  1000  1001   6 present   /usr/bin/kded5
Sun 2021-03-21 18:48:39 CET    1945  1000  1001   6 present   /usr/bin/krunner
Sun 2021-03-21 18:48:40 CET   10280  1000  1001   6 present   /usr/lib/x86_64-linux-gnu/libexec/drkonqi
Sun 2021-03-21 18:48:41 CET    9396  1000  1001   6 present   /usr/lib/x86_64-linux-gnu/libexec/drkonqi
Sun 2021-03-21 18:48:47 CET    2020  1000  1001   6 present   /usr/lib/x86_64-linux-gnu/libexec/baloorunner
Sun 2021-03-21 18:48:48 CET    2219  1000  1001   6 present   /usr/lib/x86_64-linux-gnu/libexec/kdeconnectd
Comment 8 Leyenda 2021-03-21 18:01:06 UTC
2420 is the line for bt I posted on comment 4. Then I did try again, System Monitor was running when I tuned off/on the screen, that is why it appears on last coredumpctl lines
Comment 9 Leyenda 2021-03-21 18:09:11 UTC
A video showing the problem

https://sendvid.com/woq1sla0
Comment 10 Leyenda 2021-03-21 18:21:29 UTC
I don't want to waste your time, I think it could be some problem with my system, some wrong update or because I installed some weird packages, something like that.
When I have time I'll do a fresh install of the system and try if it happens again.
Comment 11 Nate Graham 2021-03-21 18:54:15 UTC
OK, thanks. Let me know when you're got that kwin_wayland backtrace :) Then we can re-open this.
Comment 12 Leyenda 2021-03-28 11:37:17 UTC
I did a fresh installation of Neon Testing. Installed plasma-workspace-wayland and did try again.Turned off the screen and turned on, again all windows and apps crashed, panels disappeared and background went to black. Some minutes after panels and background are restored themselves, but plasmoid on panels works weird and wrong and had to exit and login again.
Comment 13 Leyenda 2021-04-10 09:27:29 UTC
Maybe it is the same problem than bug 426293?
Comment 14 Nate Graham 2021-04-12 15:26:37 UTC
Not the same problem.
Comment 15 Leyenda 2021-06-10 13:12:05 UTC
I think its the same problem as https://bugs.kde.org/show_bug.cgi?id=420160
Comment 16 Nate Graham 2021-06-10 15:35:56 UTC
Ah yup.

*** This bug has been marked as a duplicate of bug 420160 ***
Comment 17 Nate Graham 2021-10-04 14:25:57 UTC

*** This bug has been marked as a duplicate of bug 438839 ***