Created attachment 172100 [details] Backtrace #SUMMARY Hello, Upgrade to 6.1.3. I experienced insisting crashes for KWIN when alt-tabbing today after resuming from suspend. I had 2 instances in less than 20 mins. #STEPS TO REPRODUCE 1. Suspend / Resume? 2. Open a few applications on the same desktop view 3. Alt-tab a few times #OBSERVED RESULT Kwin crashes #EXPECTED RESULT Nominal operation #SOFTWARE/OS VERSIONS Operating System: Arch Linux KDE Plasma Version: 6.1.3 KDE Frameworks Version: 6.4.0 Qt Version: 6.7.2 Kernel Version: 6.8.9-zen1-1-zen (64-bit) Graphics Platform: Wayland Processors: 16 × AMD Ryzen 7 5800X 8-Core Processor Memory: 62.7 Gio of RAM Graphics Processor: AMD Radeon RX 7900 XTX Manufacturer: Micro-Star International Co., Ltd Product Name: MS-7B86 System Version: 3.0 #ADDITIONAL INFORMATION Cleaning up "~/.cache/kwin/" , "~/.cache/qtshadercache*" and re-logging seems to have stabilized things.
Cannot Reproduce. Operating System: Manjaro Linux KDE Plasma Version: 6.1.3 KDE Frameworks Version: 6.4.0 Qt Version: 6.7.2 Kernel Version: 6.10.2-1-MANJARO (64-bit) Graphics Platform: Wayland
Similar issue on arch as well: `ooo/ OS: Arch Linux x86_64 `+oooo: Kernel: 6.10.2-arch1-1 `+oooooo: Uptime: 15 mins `/:-:++oooo+: Packages: 1535 (pacman) `/++++/+++++++: Shell: zsh 5.9 `/++++++++++++++: Resolution: 3840x1600 `/+++ooooooooooooo/` DE: Plasma 6.1.3 ./ooosssso++osssssso+` WM: kwin .oossssso-````/ossssss+` Theme: Breeze-Dark [GTK2], Breeze [GTK3] -osssssso. :ssssssso. Icons: breeze-dark [GTK2/3] :osssssss/ osssso+++. Terminal: pipenv /ossssssss/ +ssssooo/- CPU: AMD Ryzen 9 5900X (24) @ 3.700GHz `/ossssso+/:- -:/+osssso+- GPU: NVIDIA GeForce RTX 2080 Ti Rev. A `+sso+:-` `.-/+oso: Memory: 5478MiB / 31999MiB
(In reply to dms.jonathan from comment #2) > Similar issue on arch as well: > > `ooo/ OS: Arch Linux x86_64 > `+oooo: Kernel: 6.10.2-arch1-1 > `+oooooo: Uptime: 15 mins > `/:-:++oooo+: Packages: 1535 (pacman) > `/++++/+++++++: Shell: zsh 5.9 > `/++++++++++++++: Resolution: 3840x1600 > `/+++ooooooooooooo/` DE: Plasma 6.1.3 > ./ooosssso++osssssso+` WM: kwin > .oossssso-````/ossssss+` Theme: Breeze-Dark [GTK2], Breeze > [GTK3] > -osssssso. :ssssssso. Icons: breeze-dark [GTK2/3] > :osssssss/ osssso+++. Terminal: pipenv > /ossssssss/ +ssssooo/- CPU: AMD Ryzen 9 5900X (24) @ > 3.700GHz > `/ossssso+/:- -:/+osssso+- GPU: NVIDIA GeForce RTX 2080 Ti > Rev. A > `+sso+:-` `.-/+oso: Memory: 5478MiB / 31999MiB For some reason, software rendering was activated. I deactivated with the system tray account of QT and it was fix on next login. Sorry for noise
Hmm, the backtrace doesn't make much sense. It's saying that this is the crashing thread: Stack trace of thread 90033: #0 0x00007cbeb231abb0 ppoll (libc.so.6 + 0x10bbb0) #1 0x00007cbeb2ab93ab _Z12qt_safe_pollP6pollfdm14QDeadlineTimer (libQt6Core.so.6 + 0x2b93ab) #2 0x00007cbeb2ac60f3 _ZN20QEventDispatcherUNIX13processEventsE6QFlagsIN10QEventLoop17ProcessEventsFlagEE (libQt6Core.so.6 + 0x2c60f3) #3 0x00007cbeb33f9693 _ZN23QUnixEventDispatcherQPA13processEventsE6QFlagsIN10QEventLoop17ProcessEventsFlagEE (libQt6Gui.so.6 + 0x5f9693) #4 0x00007cbeb294f01e _ZN10QEventLoop4execE6QFlagsINS_17ProcessEventsFlagEE (libQt6Core.so.6 + 0x14f01e) #5 0x00007cbeb294934d _ZN16QCoreApplication4execEv (libQt6Core.so.6 + 0x14934d) #6 0x00005856153e2015 n/a (kwin_wayland + 0x3d015) #7 0x00007cbeb2234e08 n/a (libc.so.6 + 0x25e08) #8 0x00007cbeb2234ecc __libc_start_main (libc.so.6 + 0x25ecc) #9 0x00005856153e7fb5 n/a (kwin_wayland + 0x42fb5) But it doesn't look like a crashing thread to me. Can you following the instructions in https://community.kde.org/Guidelines_and_HOWTOs/Debugging/How_to_create_useful_crash_reports#Retrieving_a_backtrace_using_coredumpctl to get a new backtrace that shows only the crashing thread? Maybe it will be different this time and show something more actionable.
Unfortunately, for some reason kwin coredumps are all "unavailable" from coredumpctl perspective.
(In reply to temptempor from comment #5) > Unfortunately, for some reason kwin coredumps are all "unavailable" from > coredumpctl perspective. While I had access, I followed: https://community.kde.org/Guidelines_and_HOWTOs/Debugging/How_to_create_useful_crash_reports#Generate_backtrace_from_core Is this not relevant anymore?
There are multiple ways to get a backtrace. Hopefully that one still works, but I generally direct people to https://community.kde.org/Guidelines_and_HOWTOs/Debugging/How_to_create_useful_crash_reports#Retrieving_a_backtrace_using_coredumpctl because it's easier.
🐛🧹 ⚠️ This bug has been in NEEDSINFO status with no change for at least 15 days. Please provide the requested information, then set the bug status to REPORTED. If there is no change for at least 30 days, it will be automatically closed as RESOLVED WORKSFORME. For more information about our bug triaging procedures, please read https://community.kde.org/Guidelines_and_HOWTOs/Bug_triaging. Thank you for helping us make KDE software even better for everyone!
Hi all! It appears i'm having the same issue and it's quite bad, i have to avoid alt-tabbing altogether just to avoid it. Crash Handler however made a backtrace that i'll attach as kde6_alttab_crash_gavlig.txt. Also the repro is quite easy for me: - open an gpu intensive application (in my case it's a Bevy Engine app i'm woring on, it get's rendered via Vulkan, but it can also be firefox rendering some youtube videos for example) - start spamming alt-tab - sometimes it crashes right away, sometimes you need to do it 2-3 times. (if my backtrace is lacking something, i'm ready to get more info, just let me know) here is my system info: -` .o+` ----------------- `ooo/ OS: Arch Linux x86_64 `+oooo: Kernel: Linux 6.10.4-arch2-1 `+oooooo: Uptime: 12 mins -+oooooo+: Packages: 1811 (pacman) `/:-:++oooo+: Shell: bash 5.2.32 `/++++/+++++++: Display (SAMSUNG): 3840x2160 @ 60 Hz (] `/++++++++++++++: Display (LG HDR 4K): 3840x2160 @ 60 Hz* `/+++ooooooooooooo/` DE: KDE Plasma 6.1.4 ./ooosssso++osssssso+` WM: KWin (Wayland) .oossssso-````/ossssss+` WM Theme: Breeze -osssssso. :ssssssso. Theme: Breeze (Dark) [Qt], Breeze-Dark] :osssssss/ osssso+++. Icons: breeze-dark [Qt], breeze-dark [] /ossssssss/ +ssssooo/- Font: Noto Sans (10pt) [Qt], Noto Sans] `/ossssso+/:- -:/+osssso+- Cursor: breeze (24px) `+sso+:-` `.-/+oso: Terminal: WezTerm 20240203-110809-50462 `++:. `-/+/ Terminal Font: JetBrains Mono .` `/ CPU: AMD Ryzen 9 7900X (24) @ 5.73 GHz GPU 1: AMD Radeon RX 7900 XT [Discrete] GPU 2: AMD Raphael [Integrated] Memory: 5.42 GiB / 30.56 GiB (18%) Swap: 0 B / 16.07 GiB (0%) Disk (/): 672.11 GiB / 915.15 GiB (73%s Local IP (eno1): lll.lll.l0l.lll/24 Locale: en_US.UTF-8
Created attachment 172644 [details] stack trace from Crash Handler
gavlig@gmail.com, your backtrace indicates a different issue: a crash in the AMD graphics drivers. Please read https://docs.mesa3d.org/bugs.html and submit a bug report for the Mesa developers. Thanks!
(In reply to Nate Graham from comment #11) > gavlig@gmail.com, your backtrace indicates a different issue: a crash in the > AMD graphics drivers. Please read https://docs.mesa3d.org/bugs.html and > submit a bug report for the Mesa developers. Thanks! Thank you Nate, I'll do that!
As for me, this hasn't happened again so far but getting a backtrace in this kind of situation is not easy: Every single application running on the desktop crashes and generates a coredump, this usually overwrites kwin's own coredump. Last time before it disappeared, I set 8GB of available space for coredumps but Kwin's was still overwritten.
Created attachment 173093 [details] Backtrace_with_coredump Hello, This happened and I was able to capture some stuff.
Unfortunately, I have to use Mesa git to deal with some other problem and can't get the debug symbols for the build I'm using...
🐛🧹 This bug has been in NEEDSINFO status with no change for at least 30 days. Closing as RESOLVED WORKSFORME.
(In reply to Bug Janitor Service from comment #17) > 🐛🧹 This bug has been in NEEDSINFO status with no change for at least 30 > days. Closing as RESOLVED WORKSFORME. Hey! Im using Bazzite - Fedora based. I think I had the same issue, when pressing alt + tab long enough that the "task switcher" would be shown, it crashed the desktop. I figured it has something to do with that component. I remembered that I switched the color theme from Vapor to Breeze Dark, when this bacame a problem. I switched it back to Vapor and now it's working as before. Might be worth a try for anyone searching
Can you reproduce that, temptempor?
(In reply to Nate Graham from comment #19) > Can you reproduce that, temptempor? Well it is incredibly random. I stayed on Plasma 6.2.3 due to KMSGRAB issues and the lack of an equivalent solution for now (HDR10+ capture, ability to whitelist application and binaries that we want to allow grabbing screen without user prompt). I had this happened earlier this week. But still no conclusive coredump to be analyzed. When it happens in a row, I usually get rid of it by deleting everything in ~/.cache that is more or less related to KDE shell (kcmshell*,qtshadercache*,kwin,ksycoca,ksplash*,etc.). It's usually a good stop gap. But I agree that when it happens, it *seems* it's when the window selector shows up.