Bug 470661 - [aarch64] plasmashell crashes on 1st start on aarch64
Summary: [aarch64] plasmashell crashes on 1st start on aarch64
Status: RESOLVED UPSTREAM
Alias: None
Product: plasmashell
Classification: Plasma
Component: generic-crash (show other bugs)
Version: 5.27.5
Platform: openSUSE Linux
: NOR crash
Target Milestone: 1.0
Assignee: Plasma Bugs List
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2023-06-05 11:42 UTC by Guillaume GARDET
Modified: 2023-06-09 16:52 UTC (History)
2 users (show)

See Also:
Latest Commit:
Version Fixed In:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Guillaume GARDET 2023-06-05 11:42:04 UTC
Reported first on openSUSE bugzilla: https://bugzilla.opensuse.org/show_bug.cgi?id=1211628
I bring this bug upstream to reach people knowing more on plasma.

SUMMARY
plasmashell crashes on 1st start on (unreleased) openSUSE Tumbleweed aarch64 with the following trace from coredumpctl gdb.


#0  0x0000ffff98f66618 in __pthread_kill_implementation () from /lib64/libc.so.6
#1  0x0000ffff98f1e8c0 in raise () from /lib64/libc.so.6
#2  0x0000ffff9b8e79d8 in KCrash::defaultCrashHandler(int) () from /lib64/libKF5Crash.so.5
#3  <signal handler called>
#4  0x0000ffff98f66618 in __pthread_kill_implementation () from /lib64/libc.so.6
#5  0x0000ffff98f1e8c0 in raise () from /lib64/libc.so.6
#6  <signal handler called>
#7  0x0000ffff98fc6090 in poll () from /lib64/libc.so.6
#8  0x0000ffff97c125b0 in ?? () from /lib64/libglib-2.0.so.0
#9  0x0000ffff97c126e8 in g_main_context_iteration () from /lib64/libglib-2.0.so.0
#10 0x0000ffff9963f968 in QEventDispatcherGlib::processEvents (this=0xffff50000b70, flags=...)
    at kernel/qeventdispatcher_glib.cpp:423
#11 0x0000ffff995d4a30 in QEventLoop::exec (this=this@entry=0xffff7c94e658, flags=..., flags@entry=...)
    at ../../include/QtCore/../../src/corelib/global/qflags.h:69
#12 0x0000ffff993cb3d0 in QThread::exec (this=this@entry=0xaaaaf56b5f10)
    at ../../include/QtCore/../../src/corelib/global/qflags.h:121
#13 0x0000ffff9af60554 in QQmlThreadPrivate::run (this=0xaaaaf56b5f10)
    at /usr/src/debug/qtdeclarative-everywhere-src-5.15.9+kde25/src/qml/qml/ftw/qqmlthread.cpp:155
#14 0x0000ffff993ccba4 in operator() (__closure=<optimized out>) at thread/qthread_unix.cpp:350
#15 (anonymous namespace)::terminate_on_exception<QThreadPrivate::start(void*)::<lambda()> > (t=...)
    at thread/qthread_unix.cpp:287
#16 QThreadPrivate::start (arg=0xaaaaf56b5f10) at thread/qthread_unix.cpp:310
#17 0x0000ffff98f649e4 in start_thread () from /lib64/libc.so.6
#18 0x0000ffff98fd085c in thread_start () from /lib64/libc.so.6

After this crash, plasmashell is reloaded automatically, but part of the default openSUSE theme is not applied, namely the bottom bar which is missing icons/applets and is smaller than expected.

This started zhen ze updated from 5.27.4 to 5.27.5 and it happens in X11 only.

STEPS TO REPRODUCE
Install unreleased openSUSE Tumbleweed for aarch64. 
X11 only.

OBSERVED RESULT
Theme is not applied properly and we get traces in journal.

EXPECTED RESULT
Theme should be applied properly and no crash should happen.

SOFTWARE/OS VERSIONS
Linux/KDE Plasma: openSUSE Tumbleweed
(available in About System)
KDE Plasma Version: 5.27.5 
KDE Frameworks Version: 5.106.0 
Qt Version: 5.15.9

ADDITIONAL INFORMATION
Comment 1 Nate Graham 2023-06-05 18:30:03 UTC
Backtrace does not contain any evidence of KDE code. Possibly a Qt issue or a setup/packaging issue. You said you're using "(unreleased) openSUSE Tumbleweed aarch64"; can you describe the hardware and software setup a bit more?
Comment 2 Guillaume GARDET 2023-06-06 07:50:35 UTC
(In reply to Nate Graham from comment #1)
> Backtrace does not contain any evidence of KDE code. Possibly a Qt issue or
> a setup/packaging issue. You said you're using "(unreleased) openSUSE
> Tumbleweed aarch64"; can you describe the hardware and software setup a bit
> more?

The problem is under qemu. The screenshot showing the issue can be seen here: https://openqa.opensuse.org/tests/3337754#step/opensuse_welcome/4
and can be compared to https://openqa.opensuse.org/tests/3279201#step/opensuse_welcome/4

I wrote unreleased, because QA is blocking the updates delivered to users due to this bug.
Comment 3 Guillaume GARDET 2023-06-06 11:20:26 UTC
And the journal shows:

May 23 02:13:20.779784 ip-10-0-2-15.eu-west-1.compute.internal plasmashell[1539]: 25 -- exe=/usr/bin/plasmashell
May 23 02:13:20.790901 ip-10-0-2-15.eu-west-1.compute.internal plasmashell[1539]: 13 -- platform=xcb
May 23 02:13:20.791043 ip-10-0-2-15.eu-west-1.compute.internal plasmashell[1539]: 11 -- display=:0
May 23 02:13:20.791106 ip-10-0-2-15.eu-west-1.compute.internal plasmashell[1539]: 20 -- appname=plasmashell
May 23 02:13:20.791166 ip-10-0-2-15.eu-west-1.compute.internal plasmashell[1539]: 17 -- apppath=/usr/bin
May 23 02:13:20.791216 ip-10-0-2-15.eu-west-1.compute.internal plasmashell[1539]: 10 -- signal=11
May 23 02:13:20.791279 ip-10-0-2-15.eu-west-1.compute.internal plasmashell[1539]: 9 -- pid=1539
May 23 02:13:20.791336 ip-10-0-2-15.eu-west-1.compute.internal plasmashell[1539]: 18 -- appversion=5.27.5
May 23 02:13:20.791394 ip-10-0-2-15.eu-west-1.compute.internal plasmashell[1539]: 19 -- programname=Plasma
May 23 02:13:20.791450 ip-10-0-2-15.eu-west-1.compute.internal plasmashell[1539]: 31 -- bugaddress=submit@bugs.kde.org
May 23 02:13:20.791511 ip-10-0-2-15.eu-west-1.compute.internal plasmashell[1539]: KCrash: Application 'plasmashell' crashing...
May 23 02:13:20.826815 ip-10-0-2-15.eu-west-1.compute.internal plasmashell[1539]: KCrash: Attempting to start /usr/libexec/drkonqi
Comment 4 Nate Graham 2023-06-06 15:17:32 UTC
Any ideas, Fabian?
Comment 5 Guillaume GARDET 2023-06-09 15:48:01 UTC
So, this is this Qt issue: https://bugreports.qt.io/browse/QTBUG-111935

And the (tested) fix is: https://code.qt.io/cgit/qt/qtdeclarative.git/commit/?id=15ec024152a1d1d99a4934f7b2408e7af7b2552a

Which should have been pushed to 5.15 branch, but is not in yet apparently.
Comment 6 Nate Graham 2023-06-09 16:18:00 UTC
Thanks for following up!
Comment 7 Nate Graham 2023-06-09 16:52:59 UTC
Thanks for following up!