DESCRIPTION Using the Overview search to perform a simple calculator expression (e.g., 2026-17) and then exiting the search - by pressing Esc, Super, clicking the result, or clearing the search - causes Plasma to crash and restart. Multiple, but not all, running applications are terminated, while Plasma itself recovers. This issue is reliably reproducible and appears isolated to the calculator integration used by KRunner inside the Overview effect. The system is running a 3‑monitor configuration, and all screens briefly go black during the crash. My original reaction was to toot about it, see my thread https://layer8.space/@platymew/116528146718771455. Later someone suggested, i'd file a proper bug report, this one here. I didn't feel the need to reboot the box since encountering those crashes, 2-ish days ago. Let me know if you really-really need additional logs or debug traces, because i'm reluctant to crash it intentionally, while i've got all this precious state accumulated in memory - the box is "warm" and currently being used for important tasks. STEPS TO REPRODUCE 1. Press Super to open the Overview effect (spanning 3 monitors). I think, that's the default setting for those hot corners. 2. Type a simple calculator expression, e.g. 2026-17. 3. Trigger any of the following: Press Esc, Super again, Click the calculator result, or Click the “clear search” (backspace) icon. I have not tried the hot corner trigger, yet. 4. Plasma freezes momentarily, all three monitors go black, and the session restarts. OBSERVED RESULT Calculator expressions evaluate normally. Plasma momentarily crashes and restarts. All three monitors go black during the crash. Several applications are terminated, including: Firefox, Chromium, Thunderbird, GNUcash, Steam (native), Terminator, virt‑manager window (VMs remain running) Other applications remain unaffected (e.g., Dolphin, Kate, Clock, Nextcloud tray). The desktop recovers and remains stable otherwise. EXPECTED RESULT Exiting the Overview search should not crash the Plasma session. Running applications should remain unaffected. Multi‑monitor setups should behave identically to single‑monitor setups. SOFTWARE/OS VERSIONS Operating System (available in the Info Center app, or by running `kinfo` in a terminal window): Manjaro KDE Plasma Version: 6.6.4 KDE Frameworks Version: 6.25.0 Qt Version: 6.11.0 Graphics Platform: Wayland, 3 monitors ADDITIONAL INFORMATION The issue is not an OOM event (system has 64 GB RAM, and htop reports a 3rd in apps + a 3rd in caches). Non‑calculator search terms behave normally. Pressing Enter to “run” the calculator result does nothing (possibly a separate bug) - i expected it to open the expression in the calculator app. Strongly suggests a bug in the KRunner Calculator Runner when used inside the Overview effect.
I cannot reproduce it, also on KDE Plasma 6.6.4, however, just with a single monitor. Did you try whether a multiple monitor configuration is required to reproduce the issue? Did you try reproducing the issue with krunner started with alt+space? Is anything being logged at the time the error occurs? (follow the current system log with `journalctl -f`)
(In reply to Tobias Zwick from comment #1) > Did you try whether a multiple monitor configuration is required to > reproduce the issue? > Did you try reproducing the issue with krunner started with alt+space? I'll answer the 1st 2 questions tomorrow, when i'm prepared to intentionally crash the thing. > Is anything being logged at the time the error occurs? (follow the current > system log with `journalctl -f`) For now, i'll attach the journal for 3 calculator crashes
Created attachment 192143 [details] journal log of 3 calculator crashes
The journal log indeed shows that the calculator produces a segfault and then somehow takes Wayland with it. This is the reason why all applications running on Wayland also exited (some applications tried to reconnect to Wayland). A snippet from the logs: Mai 06 16:35:16 um780xtx kernel: calculator[679649]: segfault at 10 ip 00007fa1bdae5e48 sp 00007fa19ac7e5d8 error 4 in libgmp.so.10.5.0[70e48,7fa1bda86000+7e000] likely on CPU 4 (core 4, socket 0) Mai 06 16:35:16 um780xtx kernel: Code: 48 89 de ff 15 01 4d 03 00 48 89 c1 8b 83 c0 09 00 00 89 81 c0 09 00 00 5b 41 5c 5d c3 0f 1f 44 00 00 f3 0f 1e fa 48 8b 47 18 <ff> 60 10 0f 1f 44 00 00 f3 0f 1e fa 48 8b 46 18 ff 60 18 0f 1f 44 Mai 06 16:35:16 um780xtx systemd-coredump[679678]: Process 676041 (kwin_wayland) of user 1000 terminated abnormally with signal 11/SEGV, processing... I'll set to confirmed because this might already be enough information for developers (I am just a fellow bug reporter, trying to help with triage).
I've still not rebooted the box, since encountering the 1st crash. I've unplugged the non-primary monitors from the GPU, single-monitor mode now. Clicking on calculator result in overview logged this, but didn't crash. It doesn't seem to be reproducible with only 1 monitor. Every time i return from Overview, the following gets logged. Mai 11 15:47:38 um780xtx kwin_wayland[2511871]: Could not delete texture because no context is current Mai 11 15:47:41 um780xtx kwin_wayland[2511871]: Could not delete texture because no context is current Mai 11 15:47:43 um780xtx kwin_wayland[2511871]: Could not delete texture because no context is current Mai 11 15:47:48 um780xtx kwin_wayland[2511871]: Could not delete texture because no context is current Krunner via Alt+Space doesn't seem to produce the fault condition, neither in multi-, nor in single-monitor mode. Plugging the non-primary monitors back in produced this log, see KDE_bug_report_519923-multiMon-plugIn.log It seems, running multiple monitors is a requirement to encountering this kind of crash, see log KDE_bug_report_519923-multiMon-crash.log I've rebooted the box now - technically shutdown/poweroff + cold boot, with 3 monitors. This is still logged, every time i exit the Overview. Mai 11 16:10:18 um780xtx kwin_wayland[1650]: Could not delete texture because no context is current The issue remains reproducible after rebooting, see KDE_bug_report_519923-crashAfterReboot-3mon.log. I haven't encountered this kind of crash before the big Manjaro update (i think, on May 1st). There were only updates for Chromium and Firefox, since.
Created attachment 192186 [details] Plugging the non-primary monitors back in produced this log
Created attachment 192187 [details] reproducible crash after plugging the monitors back in
Created attachment 192188 [details] The issue remains reproducible after rebooting
Created attachment 192206 [details] Demo screen recording Reproduction requires more than 1 active monitor. This demo shows a virtual machine in virt-manager, with a second VirtIO Video device. It's a freshly installed Manjaro, updated like my daily driver, on which i've encountered this crash 1st.
Pasting inline the part of the backtrace that seems more relevant: Stack trace of thread 2511871: #0 0x00007f142dde2241 n/a (libQt6Quick.so.6 + 0x3e2241) #1 0x00007f142de3cdc3 n/a (libQt6Quick.so.6 + 0x43cdc3) #2 0x00007f142bbcd5d2 _ZN14QObjectPrivate14deleteChildrenEv (libQt6Core.so.6 + 0x1cd5d2) #3 0x00007f142bbd3e4e _ZN7QObjectD2Ev (libQt6Core.so.6 + 0x1d3e4e) #4 0x00007f1423766d76 n/a (libQt6QuickTemplates2.so.6 + 0x166d76) #5 0x00007f142bbcd5d2 _ZN14QObjectPrivate14deleteChildrenEv (libQt6Core.so.6 + 0x1cd5d2) #6 0x00007f142bbd3e4e _ZN7QObjectD2Ev (libQt6Core.so.6 + 0x1d3e4e) #7 0x00007f142376d7da n/a (libQt6QuickTemplates2.so.6 + 0x16d7da) #8 0x00007f142bbcd5d2 _ZN14QObjectPrivate14deleteChildrenEv (libQt6Core.so.6 + 0x1cd5d2) #9 0x00007f142bbd3e4e _ZN7QObjectD2Ev (libQt6Core.so.6 + 0x1d3e4e) #10 0x00007f13e4c9e576 n/a (libQt6QuickLayouts.so.6 + 0x2f576) #11 0x00007f142bbcd5d2 _ZN14QObjectPrivate14deleteChildrenEv (libQt6Core.so.6 + 0x1cd5d2) #12 0x00007f142bbd3e4e _ZN7QObjectD2Ev (libQt6Core.so.6 + 0x1d3e4e) #13 0x00007f142de48ca2 n/a (libQt6Quick.so.6 + 0x448ca2) #14 0x00007f142bbcd5d2 _ZN14QObjectPrivate14deleteChildrenEv (libQt6Core.so.6 + 0x1cd5d2) #15 0x00007f142bbd3e4e _ZN7QObjectD2Ev (libQt6Core.so.6 + 0x1d3e4e) #16 0x00007f142376c666 n/a (libQt6QuickTemplates2.so.6 + 0x16c666) #17 0x00007f142ad43027 n/a (libQt6QmlModels.so.6 + 0x8d027) #18 0x00007f142ad3af8d _ZN17QQmlDelegateModelD1Ev (libQt6QmlModels.so.6 + 0x84f8d) #19 0x00007f142ad3b045 _ZN17QQmlDelegateModelD0Ev (libQt6QmlModels.so.6 + 0x85045) #20 0x00007f142dec48c6 _ZN14QQuickItemViewD2Ev (libQt6Quick.so.6 + 0x4c48c6) #21 0x00007f142de4af66 n/a (libQt6Quick.so.6 + 0x44af66) #22 0x00007f142bbcd5d2 _ZN14QObjectPrivate14deleteChildrenEv (libQt6Core.so.6 + 0x1cd5d2) #23 0x00007f142bbd3e4e _ZN7QObjectD2Ev (libQt6Core.so.6 + 0x1d3e4e) #24 0x00007f142de48ca2 n/a (libQt6Quick.so.6 + 0x448ca2) #25 0x00007f142bbd451c _ZN7QObject5eventEP6QEvent (libQt6Core.so.6 + 0x1d451c) #26 0x00007f142cd02450 _ZN19QApplicationPrivate13notify_helperEP7QObjectP6QEvent (libQt6Widgets.so.6 + 0x102450) #27 0x00007f142bb76c98 _ZN16QCoreApplication15notifyInternal2EP7QObjectP6QEvent (libQt6Core.so.6 + 0x176c98) #28 0x00007f142bb77070 _ZN23QCoreApplicationPrivate16sendPostedEventsEP7QObjectiP11QThreadData (libQt6Core.so.6 + 0x177070) #29 0x00007f142bd55cbd _ZN20QEventDispatcherUNIX13processEventsE6QFlagsIN10QEventLoop17ProcessEventsFlagEE (libQt6Core.so.6 + 0x355cbd) #30 0x00007f142c937913 _ZN23QUnixEventDispatcherQPA13processEventsE6QFlagsIN10QEventLoop17ProcessEventsFlagEE (libQt6Gui.so.6 + 0x737913) #31 0x00007f142bb845d6 _ZN10QEventLoop4execE6QFlagsINS_17ProcessEventsFlagEE (libQt6Core.so.6 + 0x1845d6) #32 0x00007f142bb7a311 _ZN16QCoreApplication4execEv (libQt6Core.so.6 + 0x17a311) #33 0x0000557f88296937 n/a (/usr/bin/kwin_wayland + 0x3d937) #34 0x00007f142b2276c1 n/a (libc.so.6 + 0x276c1) #35 0x00007f142b2277f9 __libc_start_main (libc.so.6 + 0x277f9) #36 0x0000557f8829ce25 n/a (/usr/bin/kwin_wayland + 0x43e25)
I'm not able to reproduce this on Plasma 6.6.4 or built from git-master, with multiple monitors, but with Qt 6.10.3 Since the crash looks like it's in QtQuick this looks like a potential Qt regression
During today's Manjaro update, the KDE versions got bumped, but the issue remains reproducible in my test VM. These versions are also affected. KDE Plasma Version: 6.6.5 KDE Frameworks Version: 6.26.0 Qt Version: 6.11.1 Graphics Platform: Wayland, 2 VirtIO monitors
> perform a simple calculator expression (e.g., 2026-17) and then exiting the search - by pressing Esc I can reproduce the issue. This crashes KWin every time for me. I have two monitors, one on a laptop and one standalone. - Laptop screen: 2880x1800, 175% scale - Standalone monitor: 2650x1400, 100% scale
Here's a better backtrace: #0 __pthread_kill_implementation (threadid=<optimized out>, signo=signo@entry=11, no_tid=no_tid@entry=0) at pthread_kill.c:44 #1 0x00007f3a76efd363 in __pthread_kill_internal (threadid=<optimized out>, signo=11) at pthread_kill.c:89 #2 0x00007f3a76ea17d0 in __GI_raise (sig=11) at ../sysdeps/posix/raise.c:26 #3 0x00007f3a7a4e6469 in KCrash::defaultCrashHandler (sig=11) at /usr/src/debug/kde-banana-kcrash-git/kcrash/src/kcrash.cpp:623 #4 0x00007f3a76ea18f0 in <signal handler called> () at /usr/lib/libc.so.6 #5 0x00007f39ca1c9e48 in __gmp_randclear (rstate=0x7f39915f5d00 <randstate>) at rand/randclr.c:36 #6 0x00007f39910a1638 in Calculator::~Calculator (this=0x7f39700017b0, this=<optimized out>) at /usr/src/debug/libqalculate/libqalculate/libqalculate/Calculator.cc:701 #7 0x00007f39910a1ed6 in Calculator::~Calculator (this=0x7f39700017b0, this=<optimized out>) at /usr/src/debug/libqalculate/libqalculate/libqalculate/Calculator.cc:706 #8 0x00007f39d4c0b6ed in QalculateEngine::~QalculateEngine (this=0x7f3970001780) at /home/nate/kde/src/plasma-workspace/runners/calculator/qalculate_engine.cpp:80 #9 0x00007f39d4c0b77e in QalculateEngine::~QalculateEngine (this=0x7f3970001780) at /home/nate/kde/src/plasma-workspace/runners/calculator/qalculate_engine.cpp:83 #10 0x00007f39d4c0df64 in std::default_delete<QalculateEngine>::operator() (this=<optimized out>, __ptr=<optimized out>) at /usr/include/c++/16.1.1/bits/unique_ptr.h:86 #11 std::unique_ptr<QalculateEngine, std::default_delete<QalculateEngine> >::~unique_ptr (this=0x560022ad9ad8) at /usr/include/c++/16.1.1/bits/unique_ptr.h:398 #12 CalculatorRunner::~CalculatorRunner (this=0x560022ad9ac0) at /home/nate/kde/src/plasma-workspace/runners/calculator/calculatorrunner.cpp:41 #13 0x00007f39d4c0dfde in CalculatorRunner::~CalculatorRunner (this=0x560022ad9ac0) at /home/nate/kde/src/plasma-workspace/runners/calculator/calculatorrunner.cpp:41 #14 0x00007f3a7769cc6a in QObject::event (this=<optimized out>, e=<optimized out>) at /usr/src/debug/qt6-base/qtbase/src/corelib/kernel/qobject.cpp:1464 #15 0x00007f3a788da11f in QApplicationPrivate::notify_helper (this=<optimized out>, receiver=0x560022ad9ac0, e=0x7f39702d7130) at /usr/src/debug/qt6-base/qtbase/src/widgets/kernel/qapplication.cpp:3276 #16 0x00007f3a77638658 in QCoreApplication::notifyInternal2 (receiver=0x560022ad9ac0, event=event@entry=0x7f39702d7130) at /usr/src/debug/qt6-base/qtbase/src/corelib/kernel/qcoreapplication.cpp:1114 #17 0x00007f3a77638a61 in QCoreApplication::sendEvent (receiver=<optimized out>, event=0x7f39702d7130) at /usr/src/debug/qt6-base/qtbase/src/corelib/kernel/qcoreapplication.cpp:1559 #18 QCoreApplicationPrivate::sendPostedEvents (receiver=0x0, event_type=52, data=0x56002498f170) at /usr/src/debug/qt6-base/qtbase/src/corelib/kernel/qcoreapplication.cpp:1914 #19 0x00007f3a7782d12a in QCoreApplication::sendPostedEvents (receiver=0x0, event_type=52) at /usr/src/debug/qt6-base/qtbase/src/corelib/kernel/qcoreapplication.cpp:1767 --Type <RET> for more, q to quit, c to continue without paging--c #20 QThreadPrivate::finish (this=0x560023a37ad0, calledFromExit=false) at /usr/src/debug/qt6-base/qtbase/src/corelib/thread/qthread_unix.cpp:490 #21 0x00007f3a7782d506 in operator() (__closure=0x0, arg=0x560022a7b7f0) at /usr/src/debug/qt6-base/qtbase/src/corelib/tools/qscopedpointer.h:111 #22 _FUN () at /usr/src/debug/qt6-base/qtbase/src/corelib/thread/qthread_unix.cpp:429 #23 __pthread_cleanup_class::~__pthread_cleanup_class (this=<synthetic pointer>, this=<optimized out>) at /usr/include/pthread.h:578 #24 QThreadPrivate::start (arg=0x560022a7b7f0) at /usr/src/debug/qt6-base/qtbase/src/corelib/thread/qthread_unix.cpp:471 #25 0x00007f3a76efb1b9 in start_thread (arg=<optimized out>) at pthread_create.c:454 #26 0x00007f3a76f8021c in __GI___clone3 () at ../sysdeps/unix/sysv/linux/x86_64/clone3.S:78
Git commit d08f6052e86953bfacb8acb8a192bf3c3a1049d5 by Alexander Lohnau. Committed on 25/05/2026 at 08:58. Pushed by alex into branch 'master'. plugins/overview: Reuse KRunner manager This will allow for more efficient querying, because we do not query twice and some plugins have a instance-local cache. Also, we do not run into threading issues, for example in libqalculate Related: bug 483147 M +1 -0 .kde-ci.yml M +5 -0 CMakeLists.txt M +3 -1 src/plugins/CMakeLists.txt M +1 -0 src/plugins/overview/CMakeLists.txt M +1 -2 src/plugins/overview/kcm/overvieweffectkcm.cpp M +11 -3 src/plugins/overview/overvieweffect.cpp M +5 -2 src/plugins/overview/overvieweffect.h M +12 -2 src/plugins/overview/qml/Main.qml https://invent.kde.org/plasma/kwin/-/commit/d08f6052e86953bfacb8acb8a192bf3c3a1049d5