SUMMARY kwin according to amdgpu top is eating up almost all of the GFX resources on my gpu causing the whole desktop to act extremely slowly this dose not normally happen but recently started happening. this issue is also not constant it goes in and out and stops affter a reboot but it will show up again. this also makes spectical unuseable so i can't take screenshots. STEPS TO REPRODUCE 1. use kde plasma OBSERVED RESULT 100% gpu useage casuing slow down system wide and making games unplayable EXPECTED RESULT kwin to use a normal amount of gpu. SOFTWARE/OS VERSIONS ``` c2@gentoobtw ~ $ fastfetch -/oyddmdhs+:. c2@gentoobtw -odNMMMMMMMMNNmhy+-` ------------ -yNMMMMMMMMMMMNNNmmdhy+- OS: Gentoo Linux x86_64 `omMMMMMMMMMMMMNmdmmmmddhhy/` Kernel: Linux 6.14.9-gentoo-dist omMMMMMMMMMMMNhhyyyohmdddhhhdo` Uptime: 47 mins .ydMMMMMMMMMMdhs++so/smdddhhhhdm+` Packages: 1969 (emerge), 13 (flatpak) oyhdmNMMMMMMMNdyooydmddddhhhhyhNd. Shell: bash 5.2.37 :oyhhdNNMMMMMMMNNNmmdddhhhhhyymMh Display (LG FULL HD): 1080x1920 @ 60 Hz in 22" [External] .:+sydNMMMMMNNNmmmdddhhhhhhmMmy Display (XV275K): 3840x2160 @ 60 Hz (as 3072x1728) in 27" [External] /mMMMMMMNNNmmmdddhhhhhmMNhs: Display (LG FULL HD): 1080x1920 @ 60 Hz in 22" [External] `oNMMMMMMMNNNmmmddddhhdmMNhs+` Display (Odyssey G7): 3840x2160 @ 120 Hz in 31" [External, HDR] * `sNMMMMMMMMNNNmmmdddddmNMmhs/. DE: KDE Plasma 6.3.5 /NMMMMMMMMNNNNmmmdddmNMNdso:` WM: KWin (Wayland) +MMMMMMMNNNNNmmmmdmNMNdso/- WM Theme: Infinity-Solid-Aurorae-6 yMMNNNNNNNmmmmmNNMmhs+/-` Theme: Breeze (Custum_blues_and_purples) [Qt], Breeze-Dark [GTK2], Breeze [GTK3] /hMMNNNNNNNNMNdhs++/-` Icons: Vivid-Dark-Icons [Qt], Vivid-Dark-Icons [GTK2/3/4] `/ohdmmddhys+++/:.` Font: Noto Sans (10pt) [Qt], Noto Sans (10pt) [GTK2/3/4] `-//////:--. Cursor: Vivid-Dark-Icons (24px) Terminal: konsole 24.12.3 CPU: AMD Ryzen 9 9950X3D (32) @ 5.76 GHz GPU 1: AMD Radeon RX 7900 XTX [Discrete] GPU 2: AMD Radeon Graphics [Integrated] Memory: 10.46 GiB / 91.96 GiB (11%) Swap: 0 B / 16.00 GiB (0%) Disk (/): 667.82 GiB / 914.07 GiB (73%) - xfs Disk (/run/media/hdd): 8.73 TiB / 10.83 TiB (81%) - ext4 Disk (/run/media/nvme2): 196.79 GiB / 232.88 GiB (85%) - btrfs Disk (/run/media/ssds): 213.56 GiB / 476.46 GiB (45%) - xfs Local IP (enp14s0): 10.0.0.236/24 Locale: en_US.UTF-8 c2@gentoobtw ~ $ ``` Qt Version: 6.8.3 ADDITIONAL INFORMATION
``` c2@gentoobtw ~ $ err Jun 05 09:57:12 gentoobtw kernel: hub 10-0:1.0: config failed, hub doesn't have any ports! (err -19) Jun 05 09:57:12 gentoobtw kernel: amdgpu: Overdrive is enabled, please disable it before reporting any bugs unrelated to overdrive. Jun 05 09:57:15 gentoobtw kernel: Bluetooth: hci0: No support for _PRR ACPI method Jun 05 09:57:28 gentoobtw sddm-helper[1361]: gkr-pam: unable to locate daemon control file Jun 05 09:57:30 gentoobtw bluetoothd[1120]: Failed to remove UUID: Authentication Failed (0x05) Jun 05 09:57:30 gentoobtw kernel: Bluetooth: hci0: Opcode 0x0c24 failed: -110 Jun 05 09:57:30 gentoobtw kernel: Bluetooth: hci0: command 0x0c24 tx timeout Jun 05 09:57:30 gentoobtw kernel: Bluetooth: hci0: Resetting usb device. Jun 05 09:57:31 gentoobtw kernel: Bluetooth: hci0: No support for _PRR ACPI method Jun 05 09:57:32 gentoobtw bluetoothd[1120]: src/adv_monitor.c:btd_adv_monitor_power_down() Unexpected NULL btd_adv_monitor_manager> Jun 05 10:08:18 gentoobtw pipewire-pulse[1755]: mod.protocol-pulse: PID 2099 .flatpak-info parsing failed: Operation not supported Jun 05 10:18:13 gentoobtw pipewire-pulse[1755]: mod.protocol-pulse: PID 2099 .flatpak-info parsing failed: Operation not supported c2@gentoobtw ~ $ journalctl -xb | grep kwin Jun 05 09:57:29 gentoobtw kwin_wayland_wrapper[1473]: No backend specified, automatically choosing drm Jun 05 09:57:30 gentoobtw kwin_wayland_wrapper[1639]: The XKEYBOARD keymap compiler (xkbcomp) reports: Jun 05 09:57:30 gentoobtw kwin_wayland_wrapper[1639]: > Warning: Could not resolve keysym XF86RefreshRateToggle Jun 05 09:57:30 gentoobtw kwin_wayland_wrapper[1639]: > Warning: Could not resolve keysym XF86Accessibility Jun 05 09:57:30 gentoobtw kwin_wayland_wrapper[1639]: > Warning: Could not resolve keysym XF86DoNotDisturb Jun 05 09:57:30 gentoobtw kwin_wayland_wrapper[1639]: Errors from xkbcomp are not fatal to the X server Jun 05 09:57:30 gentoobtw kwin_wayland_wrapper[1669]: The XKEYBOARD keymap compiler (xkbcomp) reports: Jun 05 09:57:30 gentoobtw kwin_wayland_wrapper[1669]: > Warning: Unsupported maximum keycode 708, clipping. Jun 05 09:57:30 gentoobtw kwin_wayland_wrapper[1669]: > X11 cannot support keycodes above 255. Jun 05 09:57:30 gentoobtw kwin_wayland_wrapper[1669]: > Warning: Could not resolve keysym XF86RefreshRateToggle Jun 05 09:57:30 gentoobtw kwin_wayland_wrapper[1669]: > Warning: Could not resolve keysym XF86Accessibility Jun 05 09:57:30 gentoobtw kwin_wayland_wrapper[1669]: > Warning: Could not resolve keysym XF86DoNotDisturb Jun 05 09:57:30 gentoobtw kwin_wayland_wrapper[1669]: Errors from xkbcomp are not fatal to the X server Jun 05 09:57:30 gentoobtw kwin_wayland_wrapper[1673]: Initializing "/usr/lib64/qt6/plugins/plasma/kcms/systemsettings/kcm_fonts.so" Jun 05 09:57:30 gentoobtw kwin_wayland_wrapper[1673]: Initializing "/usr/lib64/qt6/plugins/plasma/kcms/systemsettings/kcm_style.so" Jun 05 09:57:31 gentoobtw kwin_wayland_wrapper[1473]: Module 'org.kde.kwin.decoration' does not contain a module identifier directive - it cannot be protected from external registrations. Jun 05 09:57:31 gentoobtw kwin_wayland_wrapper[1473]: kf.windowsystem: static bool KX11Extras::mapViewport() may only be used on X11 Jun 05 09:57:31 gentoobtw kwin_wayland_wrapper[1473]: kf.windowsystem: static bool KX11Extras::mapViewport() may only be used on X11 Jun 05 09:57:32 gentoobtw kwin_wayland_wrapper[1473]: <Unknown File>:97: TypeError: Cannot read property 'translucencyWindowTypeAnimation' of undefined Jun 05 09:57:33 gentoobtw kwin_wayland_wrapper[1473]: kf.windowsystem: static bool KX11Extras::mapViewport() may only be used on X11 Jun 05 09:59:50 gentoobtw kwin_wayland_wrapper[1473]: <Unknown File>:97: TypeError: Cannot read property 'translucencyWindowTypeAnimation' of undefined Jun 05 10:00:44 gentoobtw dbus-daemon[1395]: [session uid=1000 pid=1395 pidfd=5] Activating via systemd: service name='org.kde.spectacle' unit='app-org.kde.spectacle.service' requested by ':1.13' (uid=1000 pid=1473 comm="/usr/bin/kwin_wayland --wayland-fd 7 --socket wayl") Jun 05 10:16:21 gentoobtw kwin_wayland_wrapper[1473]: kf.windowsystem: static bool KX11Extras::mapViewport() may only be used on X11 Jun 05 10:16:22 gentoobtw kwin_wayland_wrapper[1473]: <Unknown File>:97: TypeError: Cannot read property 'translucencyWindowTypeAnimation' of undefined Jun 05 10:16:23 gentoobtw kwin_wayland_wrapper[1473]: kf.windowsystem: static bool KX11Extras::mapViewport() may only be used on X11 Jun 05 10:16:32 gentoobtw kwin_wayland_wrapper[1473]: kf.windowsystem: static bool KX11Extras::mapViewport() may only be used on X11 Jun 05 10:17:25 gentoobtw kwin_wayland_wrapper[1473]: kf.windowsystem: static bool KX11Extras::mapViewport() may only be used on X11 Jun 05 10:17:26 gentoobtw kwin_wayland_wrapper[1473]: kf.windowsystem: static bool KX11Extras::mapViewport() may only be used on X11 Jun 05 10:17:36 gentoobtw kwin_wayland_wrapper[1473]: <Unknown File>:97: TypeError: Cannot read property 'translucencyWindowTypeAnimation' of undefined Jun 05 10:17:38 gentoobtw kwin_wayland_wrapper[1473]: <Unknown File>:97: TypeError: Cannot read property 'translucencyWindowTypeAnimation' of undefined Jun 05 10:27:23 gentoobtw dbus-daemon[1395]: [session uid=1000 pid=1395 pidfd=5] Activating via systemd: service name='org.kde.spectacle' unit='app-org.kde.spectacle.service' requested by ':1.13' (uid=1000 pid=1473 comm="/usr/bin/kwin_wayland --wayland-fd 7 --socket wayl") Jun 05 10:35:05 gentoobtw sudo-rs[19507]: c2 : TTY=/dev/pts/6 ; PWD=/home/c2 ; USER=root ; COMMAND=/usr/bin/taskset -c 2-31 emerge -s kwin Jun 05 10:35:14 gentoobtw sudo-rs[19606]: c2 : TTY=/dev/pts/6 ; PWD=/home/c2 ; USER=root ; COMMAND=/usr/bin/taskset -c 2-31 emerge -av kde-plasma/kwin Jun 05 10:36:44 gentoobtw dbus-daemon[1395]: [session uid=1000 pid=1395 pidfd=5] Activating via systemd: service name='org.kde.spectacle' unit='app-org.kde.spectacle.service' requested by ':1.13' (uid=1000 pid=1473 comm="/usr/bin/kwin_wayland --wayland-fd 7 --socket wayl") c2@gentoobtw ~ $ ```
also kwin is doing this sometimes on boot
what i emnt in the last comment is that for some reason sometimes on start up the pannel in plasma dosne't show up on some of my monitors and the there was basicaly a black screen minus the cursor on 2 of the screens
(In reply to testc2n14 from comment #3) > what i emnt in the last comment is that for some reason sometimes on start > up the pannel in plasma dosne't show up on some of my monitors and the there > was basicaly a black screen minus the cursor on 2 of the screens This sounds like a different bug. Please open a new bug report for it. Thanks.
I confirm I'm seeing the same issue on 6.3.91-3 on Arch with an intel gpu, it happens sometimes, usually triggered by notifications it seems, and all I have to do is show the panel / press the window key, and the usage goes away.
for me the panel becomes totally unresponsive a good amount the time the only solution i have found is to either kill the kwin presses or to reboot. i think it may have had something to do with a dual GPU setup ie a igpu and a graphics card across multiple monitors
i am still encouting a bug where kwin eats 30% of my gpu until reboot, killing the kwin prosses stops it for 5 minutes. it will constatly use 28-32 % GFX on my gpu constatly ther is no downtime i can walk away for a full day come back and it's still using 30 gpu on idle i have a 7900 xtx so thats alot of gpu it's eating
https://drive.proton.me/urls/6MMTE53J5G#ltM7uQYBTxJx
Created attachment 182307 [details] amdgpu_top output Please attach screenshots directly to bug reports. External links can stop working. Thanks.
so sense someone else has this issue and no soultion has been purposed here i am going to file a bug report for gentoo
sorry i didn't mean to send last message
but i am still having this issue on kde plasma 6.4.2 and replicated the issue on the stable version of the package on gentoo
hey could someone move this up in piority it;s been over a month and this issue is getting unbearable with the heat outside my ac can't deal with my gpu randomly pulling 100w extra for no reason. i have a log where i made a systemD service to check for spikes in gpu useage in this log i have everything 5 minutes befor the spike and 2 minutes affter https://960205.bugs.gentoo.org/attachment.cgi?id=938628 also i made a post ont he gentoo bug tracker https://bugs.gentoo.org/960205
Thanks for the logs, that helps. I noticed there are a lot of coredumps. I'm not sure if things are crashing because of the high system resource usage, or the resource usage is being impacted by all the crashes. There are 847 unique crash files referenced in those logs, which indicate the coredump processor is processing them. This would definitely contribute to system load. The first example: Jul 19 18:28:29 gentoobtw drkonqi-coredump-processor[12327]: "/home/c2/.local/share/Steam/ubuntu12_32/steam" 4226 "/var/lib/systemd/coredump/core.steam.1000.230236a542f84bc9ab20a50c18c8ef5e.4226.1749083577000000.zst" Some of the coredumps are for kwin_wayland itself. 2494:Jul 19 18:28:29 gentoobtw drkonqi-coredump-processor[12327]: "/usr/bin/kwin_wayland" 1500 "/var/lib/systemd/coredump/core.kwin_wayland.1000.06f6fa6f06d5410ba62bf29e50831b4f.1500.1749803367000000.zst" 2521:Jul 19 18:28:29 gentoobtw drkonqi-coredump-processor[12327]: "/usr/bin/kwin_wayland" 1496 "/var/lib/systemd/coredump/core.kwin_wayland.1000.0f3b98fbcc9147238f73f515738c984f.1496.1749937390000000.zst" 2524:Jul 19 18:28:29 gentoobtw drkonqi-coredump-processor[12327]: "/usr/bin/kwin_wayland" 1047070 "/var/lib/systemd/coredump/core.kwin_wayland.1000.0f3b98fbcc9147238f73f515738c984f.1047070.1749937546000000.zst" Have you noticed things crashing before or after the high system load? Also, I would recommend reporting at least the kwin crashes in a separate bug report.
Also, we will need the performance information they asked for in the Gentoo report 4) Run `perf top -g` for a few minutes when it is misbehaving and take a screenshot. 5) If 4) doesn't prove useful, you can do `perf record -p PID_OF_KWIN -g` for a few minutes or more, and then ^C, then perf report -g to see what it was doing.
hey so it capture the data and it seems to have gotten info on the things it's doing that are using up so much prepossessing power but idk what any of it means. posting as a proton drive link because file size was too big for the gentoo bug report https://drive.proton.me/urls/JNBC0PVJC4#6AFTaC2GVDcm https://drive.proton.me/urls/TDCZ5551SG#hf9TvsmJi1fr
(In reply to testc2n14 from comment #16) > posting as a proton drive link because file size was too big for the gentoo > bug report > https://drive.proton.me/urls/JNBC0PVJC4#6AFTaC2GVDcm > https://drive.proton.me/urls/TDCZ5551SG#hf9TvsmJi1fr Please attach the files directly to this bug report. Files on other systems can go missing. Thanks.
i have recently added a second gpu and kwin has been going crazy and eating up 100% gpu for seemingly no reason. i will add attachments hsowing the issue. also for mutiple of the reports the onyl grpahical apps running are lact and kde konsol
Created attachment 183928 [details] per report of kwin
Created attachment 183929 [details] perf -g
Created attachment 183930 [details] amdgpu_top useage show case
Created attachment 183931 [details] lact prosses showcase
Thanks for attaching those reports. The kwin developers can take a closer look.
hey experiencing more issues this time mostly not it constantly using a large amount of gpu resources but only when i am doing something like 80% GPU usage for moving my mouse when some windows are maximized but for others it's only 30% or sometimes it only hangs every once and a while with a quick spike of 25% but on my desktop it's 1-5% usage moving around the mouse but if if i use the select tool 100% usage. video showing the issue https://youtu.be/yqpcmgxuNgk FYI i only have accesses to my computer today earliest i might be able to have it on wednessday if my school lets me have it on campus again but worse case is when ever my parents come to visit me again. for some reason the file uploader for gentoo bug report wouldn't let me upload any file with this text in it so i am just coping and pasting it Samples: 104K of event 'cycles:P', Event count (approx.): 115644933480 Children Self Command Shared Object Symbol + 22.53% 0.01% kwin_wayland [kernel.kallsyms] [k] 0xffffffff9aa0012f + 22.43% 0.03% kwin_wayland libQt6Widgets.so.6.9.1 [.] QApplicationPrivate::notify_helper + 21.96% 0.03% kwin_wayland [kernel.kallsyms] [k] 0xffffffff9bfe1eea + 19.97% 0.00% kwin_wayland libQt6Core.so.6.9.1 [.] 0x00007fe5185732ca + 18.16% 0.00% kwin_wayland [unknown] [.] 0x7f8007894800140d + 16.83% 0.00% kwin_wayland [unknown] [.] 0x0000000100000001 + 16.81% 0.00% kwin_wayland libQt6Core.so.6.9.1 [.] QTimerEvent::~QTimerEvent + 16.77% 0.02% kwin_wayland libQt6Core.so.6.9.1 [.] QTimer::timerEvent + 16.64% 0.01% kwin_wayland libkwin.so.6.4.4 [.] KWin::RenderLoop::frameRequested + 16.44% 0.05% kwin_wayland libkwin.so.6.4.4 [.] KWin::Compositor::composite + 12.89% 0.10% kwin_wayland libc.so.6 [.] ioctl + 12.05% 0.01% kwin_wayland libkwin.so.6.4.4 [.] KWin::Compositor::paintPass + 12.03% 0.01% kwin_wayland libkwin.so.6.4.4 [.] KWin::SceneDelegate::paint + 11.99% 0.01% kwin_wayland libkwin.so.6.4.4 [.] KWin::WorkspaceScene::paint + 11.73% 0.01% kwin_wayland libkwin.so.6.4.4 [.] KWin::EffectsHandler::paintScreen + 11.56% 0.01% kwin_wayland libkwin.so.6.4.4 [.] KWin::WorkspaceScene::paintWindow + 11.54% 0.01% kwin_wayland libkwin.so.6.4.4 [.] KWin::EffectsHandler::paintWindow + 11.49% 0.01% kwin_wayland libkwin.so.6.4.4 [.] KWin::EffectsHandler::drawWindow + 9.73% 0.00% DP-4 [unknown] [k] 0xe9097400387f8348 + 9.73% 0.00% DP-4 libQt6Core.so.6.9.1 [.] QThread::qt_metacast + 9.73% 0.00% DP-4 [unknown] [.] 0x0000560d4346fab0 + 9.73% 0.00% DP-4 libQt6Core.so.6.9.1 [.] 0x00007fe5184b7343 + 9.73% 0.00% DP-4 libc.so.6 [.] pthread_once + 9.73% 0.00% DP-4 libc.so.6 [.] 0x00007fe5178b2469 + 9.73% 0.00% DP-4 libkwin.so.6.4.4 [.] std::__future_base::_State_baseV2::_M_do_set + 9.73% 0.00% DP-4 libkwin.so.6.4.4 [.] std::_Function_handler<std::unique_ptr<std::__future_base::_Result_base, std::__future_base::_Result_base::_Deleter> (), std::__future_base::_Task_setter<std::unique_ptr<std::__future_base::_Result<void>, std::__future_base::_Result_base::_Deleter>, std::thread::_Invoker<std::tuple<QThread::create<KWin::DrmCommitThread::DrmCommitThread(KWin::DrmGpu*, QString const&)::{lambda()#1}>(KWin::DrmC + 8.82% 0.00% kwin_wayland [kernel.kallsyms] [k] 0xffffffff9b128f13 + 8.55% 0.00% kwin_wayland libkwin.so.6.4.4 [.] KWin::OffscreenEffect::drawWindow + 8.23% 0.00% DP-4 libc.so.6 [.] 0x00007fe5178b5672 + 8.19% 0.00% kwin_wayland libc.so.6 [.] 0x00007fe5178b5672 + 7.91% 0.02% DP-4 [kernel.kallsyms] [k] 0xffffffff9aa0012f + 7.78% 0.00% kwin_wayland [amdgpu] [k] 0xffffffffc0a00052 + 7.77% 0.00% kwin_wayland libkwin.so.6.4.4 [.] KWin::WorkspaceScene::paintGenericScreen + 7.63% 0.00% kwin_wayland [kernel.kallsyms] [k] 0xffffffff9bb35f33 + 7.38% 0.00% kwin_wayland [kernel.kallsyms] [k] 0xffffffff9bb35b6b + 7.34% 0.11% DP-4 [kernel.kallsyms] [k] 0xffffffff9bfe1eea + 6.76% 0.00% kwin_wayland [kernel.kallsyms] [k] 0xffffffff9aa012a6 + 6.74% 0.00% kwin_wayland [kernel.kallsyms] [k] 0xffffffff9bfe73b9 + 6.34% 6.16% kwin_wayland libc.so.6 [.] srandom_r + 6.12% 0.00% kwin_wayland [kernel.kallsyms] [k] 0xffffffff9b028efc + 6.07% 0.00% kwin_wayla:cs0 [kernel.kallsyms] [k] 0xffffffff9aa0012f + 5.98% 0.00% kwin_wayla:cs0 [kernel.kallsyms] [k] 0xffffffff9bfe1eea + 5.68% 0.00% kwin_wayland [kernel.kallsyms] [k] 0xffffffff9acd87b4 + 5.66% 0.00% kwin_wayland [kernel.kallsyms] [k] 0xffffffff9b02d7e3 + 5.63% 0.00% kwin_wayla:cs0 [unknown] [.] 0000000000000000 + 5.62% 0.00% kwin_wayla:cs0 libgallium-25.1.8.so [.] 0x00007fe50b4638db + 5.59% 0.00% kwin_wayland [amdgpu] [k] 0xffffffffc0a2f701 + 5.54% 0.00% kwin_wayland [ttm] [k] 0xfffffffc0641b52 + 5.15% 0.00% kwin_wayla:cs0 libgallium-25.1.8.so [.] 0x00007fe50b3e60c5 + 5.14% 0.04% kwin_wayla:cs0 libc.so.6 [.] ioctl + 5.00% 0.00% kwin_wayland [unknown] [k] 0x0000000800000008 + 5.00% 0.00% kwin_wayland libgallium-25.1.8.so [.] 0x00007fe50ae52b79 + 5.00% 0.00% kwin_wayland [kernel.kallsyms] [k] 0xffffffff9b021ece + 4.95% 0.00% kwin_wayla:cs0 [kernel.kallsyms] [k] 0xffffffff9b128f13 + 4.84% 0.00% kwin_wayla:cs0 [amdgpu] [k] 0xffffffffc0a00052 + 4.84% 0.00% kwin_wayland [kernel.kallsyms] [k] 0xffffffff9ace51e1 + 4.83% 0.00% kwin_wayla:cs0 libgallium-25.1.8.so [.] 0x00007fe50bd19b0c + 4.83% 0.00% kwin_wayla:cs0 libgallium-25.1.8.so [.] 0x00007fe50bd1eecb + 4.80% 0.00% kwin_wayland libgallium-25.1.8.so [.] 0x00007fe50af2f5d6 + 4.80% 0.00% kwin_wayland libgallium-25.1.8.so [.] 0x00007fe50af2f099 + 4.77% 0.00% kwin_wayla:cs0 [unknown] [k] 0x4810c0834800444e + 4.77% 0.00% kwin_wayla:cs0 libLLVM.so.20.1 [.] 0x00007fe509e70a30 + 4.77% 0.00% kwin_wayla:cs0 [kernel.kallsyms] [k] 0xffffffff9bb35f33 + 4.70% 0.00% kwin_wayland [unknown] [.] 0000000000000000 + 4.70% 0.00% kwin_wayla:cs0 [kernel.kallsyms] [k] 0xffffffff9bb35b6b + 4.69% 0.00% kwin_wayland [kernel.kallsyms] [k] 0xffffffff9ace4cef + 4.44% 0.00% kwin_wayland [kernel.kallsyms] [k] 0xffffffff9b12ce6e + 4.43% 0.01% DP-4 [kernel.kallsyms] [k] 0xffffffff9ae09a6e + 4.36% 0.00% kwin_wayland libQt6Core.so.6.9.1 [.] QMetaCallEvent::~QMetaCallEvent + 4.35% 0.00% kwin_wayland [unknown] [.] 0xfc894954410020ab + 4.35% 0.00% kwin_wayland [unknown] [.] 0x000000010000002b + 4.26% 0.02% kwin_wayland libQt6Core.so.6.9.1 [.] QObject::event + 4.24% 0.02% DP-4 [kernel.kallsyms] [k] 0xffffffff9adfc830 + 4.22% 0.00% kwin_wayland libc.so.6 [.] 0x00007fe5179a9039 + 4.10% 0.00% kwin_wayland [kernel.kallsyms] [k] 0xffffffff9b10dd51 + 4.04% 0.00% kwin_wayland [kernel.kallsyms] [k] 0xffffffff9b02ca30
is there any thing else i could posibly do to either fix the issue or gain more info about it, would changing use flags or mabey using the binary package work?, this is getting bad enough to where i just want to switch to either sway or hyprland
Unfortunately, at this point the KWin devs will need to take a look.
(In reply to TraceyC from comment #26) > Unfortunately, at this point the KWin devs will need to take a look. and kind of a time line on when they will see it or no? is there anything i could do to speed up the prosses
Created attachment 184939 [details] kwin memoery leak so i am getting memeroy leaks now, i was doing nothing weird just a single monitor pluged into my 7900 xtx
Sep 13 16:47:09 gentoobtw kernel: Out of memory: Killed process 1836 (kwin_wayland) total-vm:137176584kB, anon-rss:87179120kB, file-rss:212692kB, shmem-rss:156288kB, UID:1000 pgtables:202020kB oom_score_adj:200
(In reply to testc2n14 from comment #25) > is there any thing else i could posibly do to either fix the issue or gain > more info about it, would changing use flags or mabey using the binary > package work? Try it. As is, this isn't really actionable, we need to find out what's triggering the issue. If you have any plugins / effects / scripts enabled, try disabling them as well and see if one is the cause. Or as a bigger hammer, try a completely new user account, so that all the configs and everything have default values.
(In reply to Zamundaaa from comment #30) > (In reply to testc2n14 from comment #25) > > is there any thing else i could posibly do to either fix the issue or gain > > more info about it, would changing use flags or mabey using the binary > > package work? > Try it. As is, this isn't really actionable, we need to find out what's > triggering the issue. > If you have any plugins / effects / scripts enabled, try disabling them as > well and see if one is the cause. Or as a bigger hammer, try a completely > new user account, so that all the configs and everything have default values. yeah i have no plugins or scirpts on anything, my setup is pretty much default i messed around with with setitngs mainly turning stuff like wobly windows on adjusting the colors to be just right but i am mainly using just breeze. i assume we could find the source issue in the perflogs or journalctl logs. also it's going to be complicated to do anything for the next couple weeks as my parent's aren't going to be in but i have ssh mostly setup (i need to configure the netowrk and router and my server it's a whole thing) so i can ssh into the server and hopefully setup some graphical remote acsses software from ssh.
🐛🧹 ⚠️ 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!
(In reply to Bug Janitor Service from comment #32) > 🐛🧹 ⚠️ 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! the info may not been in the title post but is in the comments
@testc2n14@proton.me Not sure what's causing it but I remember experiencing a similar issue. My system broke after installing an app that did its own instillation of xdg-desktop-portal-gtk. It was running along side xdg-desktop-portal-kde and causing issues. Could be completely unrelated to your issue. What's the output of: ps aux | grep xdg-desktop-portal
(In reply to yelsinsepulveda from comment #34) > @testc2n14@proton.me > > Not sure what's causing it but I remember experiencing a similar issue. My > system broke after installing an app that did its own instillation of > xdg-desktop-portal-gtk. It was running along side xdg-desktop-portal-kde and > causing issues. Could be completely unrelated to your issue. > > What's the output of: > ps aux | grep xdg-desktop-portal yeah i don't think that would be the cause of my issue, heres the output of the command ``` c2@gentoobtw ~ $ ps aux | grep xdg-desktop-portal c2 1372 0.0 0.0 612796 17676 ? Ssl 09:21 0:02 /usr/libexec/xdg-desktop-portal c2 1664 0.0 0.2 1926340 283680 ? Ssl 09:21 0:13 /usr/libexec/xdg-desktop-portal-kde c2 1848 0.0 0.0 497916 34276 ? Ssl 09:21 0:00 /usr/libexec/xdg-desktop-portal-gtk c2 232035 0.0 0.0 268538392 36940 ? S 19:11 0:00 /usr/libexec/kf6/kioworker /usr/lib64/qt6/plugins/kf6/kio/tags.so tags local:/run/user/1000/xdg-desktop-portal-kdeerCBSw.22.kioworker.socket c2 232043 2.1 0.2 1455928 233764 ? SNl 19:11 0:04 /usr/libexec/kf6/kioworker /usr/lib64/qt6/plugins/kf6/kio/thumbnail.so thumbnail local:/run/user/1000/xdg-desktop-portal-kdeymtNwb.28.kioworker.socket c2 232245 0.5 0.0 177552 39560 ? Sl 19:11 0:01 /usr/libexec/kf6/kioworker /usr/lib64/qt6/plugins/kf6/kio/kio_trash.so trash local:/run/user/1000/xdg-desktop-portal-kdeRTophk.29.kioworker.socket c2 232248 0.0 0.0 177108 38816 ? Sl 19:11 0:00 /usr/libexec/kf6/kioworker /usr/lib64/qt6/plugins/kf6/kio/kio_trash.so trash local:/run/user/1000/xdg-desktop-portal-kdeZlJdaM.30.kioworker.socket c2 233051 0.0 0.0 8992 6280 pts/3 S+ 19:15 0:00 grep --color=auto xdg-desktop-portal c2@gentoobtw ~ $ ```
This is still in NEEDSINFO because we're still waiting on information that Zamundaaa asked for. - Try a completely new user account, so that all the configs and everything have default values. I did read that your normal user is pretty close to defaults, but this is still something that needs to be tried. As he said, we need to narrow down what's causing the issue before we can try to fix it.
(In reply to TraceyC from comment #36) > This is still in NEEDSINFO because we're still waiting on information that > Zamundaaa asked for. > > - Try a completely new user account, so that all the configs and everything > have default values. I did read that your normal user is pretty close to > defaults, but this is still something that needs to be tried. > > As he said, we need to narrow down what's causing the issue before we can > try to fix it. ok i will do that but first i will see if i can get the issue to trigger if i am just sitting on a blank desktop and never open up any apps, i don't want to deal with the pain of getting all my stuff wokring on a new user
hey uh leavving this bug report kde bysdie just this issue has been way way too buggy for me on mutiple diffrent devices and i just can;t deal with it, setting up hyrland seems like it;s jsut goign to be way less effort and easier, i just want a stable de with hdr that dosen;t constantly break for no reason
(In reply to François Guerraz from comment #5) > I confirm I'm seeing the same issue on 6.3.91-3 on Arch with an intel gpu, > it happens sometimes, usually triggered by notifications it seems, and all I > have to do is show the panel / press the window key, and the usage goes away. Are you still seeing this problem with plasma 6.4.5 or later? If so, could you run `sudo perf top -g` until the system misbehaves and take a screenshot, then attach that to this report? We're trying to find something to narrow down the root cause to make this actionable. Thanks.
(In reply to TraceyC from comment #39) > (In reply to François Guerraz from comment #5) > > I confirm I'm seeing the same issue on 6.3.91-3 on Arch with an intel gpu, > > it happens sometimes, usually triggered by notifications it seems, and all I > > have to do is show the panel / press the window key, and the usage goes away. > > Are you still seeing this problem with plasma 6.4.5 or later? If so, could > you run `sudo perf top -g` until the system misbehaves and take a > screenshot, then attach that to this report? We're trying to find something > to narrow down the root cause to make this actionable. Thanks. yeah still seeing the issue also i am pretty sure i already upoaded that data as i did a good amount of stuff with perf
(In reply to testc2n14 from comment #40) > yeah still seeing the issue > > also i am pretty sure i already upoaded that data as i did a good amount of > stuff with perf I do see all the attachments and comments you provided with perf data and other debugging. I was asking François for data from their system.
I see this many times. Operating System: Fedora Linux 42 KDE Plasma Version: 6.5.2 KDE Frameworks Version: 6.19.0 Qt Version: 6.9.3 Kernel Version: 6.17.7-200.fc42.x86_64 (64-bit) Graphics Platform: Wayland Processors: 8 × AMD FX(tm)-8320 Eight-Core Processor Memory: 32 GiB of RAM (31.3 GiB usable) Graphics Processor: NVIDIA GeForce GTX 960 Usually when I try switch user and kwin tries talking to the video card. kernel: [drm:nv_drm_gem_alloc_nvkms_memory_ioctl [nvidia_drm] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Failed to allocate NVKMS memory for GEM object .... The following three lines just repeat over and over in the journal kwin_core: Rendering a layer failed! kwin_core: src QRectF(0,0 1920x1800) -> dst QRect(0,0 1920x1800) kwin_wayland_drm: Failed to create framebuffer: Invalid argument I am not sure if this is related to video memory or another issue. Sometimes my system will work for days, other times just hours before getting a black screen when I lock the screen to switch to a different session. I can ssh into the system and look at log files. System worked well under X11 but I have had issues with Wayland.
(In reply to Robin Laing from comment #43) > I see this many times. Just to confirm, are you also seeing 100% GPU usage?
Created attachment 187057 [details] ~93% gpy yseage on the 6600 on idle just wanted to add that this is still happening, as of right now no displays are plugged into he 7900 xtx and i have 3 displays plugged into the 6600
really didn't want ot spend my entire thanks giving breaks witchign to hyprland but that seems alot easier then dealing with this bug
Setting this to NEEDSINFO pending a response from Robin
I changed my system configuration 2 days ago. Switched from SDDM to GDM which has helped save session issues, so I cannot honestly comment on this bug anymore. Since I made the change, I am feeling that much of the lockup issues are related to sddm and how it calls kwin. I won't be able to do any testing of this installation for a few days. What I found out before is kwin starts acting up when the GPU ram is getting used up. System becomes very unresponsive. It doesn't need to be 100% but high usage. When I tried to switch user, the system would move to a crawl and that is when I came in to report the issue. Since that point, I have found that kscreenlocker_greet is using up 196MiB for the locked session when I open a second session. It is presently the largest GPU usage at 10%. Followed by an open Thunderbird session and then Plasmashell. This is using nvtop to monitor. If I lock my second session, kscreenlocker_greet for the second session uses up about the same amount of GPU ram. I used to be able to open 3 responsive sessions with F41. Now I am limited to 2 sessions, even if I close off most apps.
removed my self from the cc list sense i have pretty much fully switched to hyprland and i can not replicate the issue there.
I just wanted to add. Installed a new video card today and noticed that vram usage is about 1Gig per session. My old card only had 2Gig of vram and may have helped cause issues. If I have anymore related issues, I will add comments.
Thanks for keeping us updated, Robin. I'll mark this NEEDSINFO so it reminds us in 15 days.
A few days later. I am still under the impression that kwin_wayland is an issue but kscreenlocker it worse. Presently, on my system A session that is open and locked, kscreenlocker is using 228MiB of video ram using nvtop. kwin_wayland is using 222MiB of vram. Firefox with over 1000 tabs is at 187MiB. Second account open is running java apps. kwin_wayland for that account is only using 31MiB. I cannot understand why kscreenlocker needs so much vram when the display isn't even being on. Background is a solid colour, don't even have the clock display. This is an strace of the kscreenlocker process. No clock with the display but lots of localtime calls. ppoll([{fd=4, events=POLLIN}, {fd=6, events=POLLIN}, {fd=40, events=POLLIN}, {fd=45, events=POLLIN}, {fd=47, events=POLLIN}, {fd=48, events=POLLIN}, {fd=288, events=POLLIN}], 7, {tv_sec=0, tv_nsec=999000000}, NULL, 8) = 0 (Timeout) newfstatat(AT_FDCWD, "/etc/localtime", {st_mode=S_IFREG|0644, st_size=2332, ...}, 0) = 0 newfstatat(AT_FDCWD, "/etc/localtime", {st_mode=S_IFREG|0644, st_size=2332, ...}, 0) = 0 newfstatat(AT_FDCWD, "/etc/localtime", {st_mode=S_IFREG|0644, st_size=2332, ...}, 0) = 0 newfstatat(AT_FDCWD, "/etc/localtime", {st_mode=S_IFREG|0644, st_size=2332, ...}, 0) = 0 newfstatat(AT_FDCWD, "/etc/localtime", {st_mode=S_IFREG|0644, st_size=2332, ...}, 0) = 0 newfstatat(AT_FDCWD, "/etc/localtime", {st_mode=S_IFREG|0644, st_size=2332, ...}, 0) = 0 newfstatat(AT_FDCWD, "/etc/localtime", {st_mode=S_IFREG|0644, st_size=2332, ...}, 0) = 0 newfstatat(AT_FDCWD, "/etc/localtime", {st_mode=S_IFREG|0644, st_size=2332, ...}, 0) = 0 ppoll([{fd=4, events=POLLIN}, {fd=6, events=POLLIN}, {fd=40, events=POLLIN}, {fd=45, events=POLLIN}, {fd=47, events=POLLIN}, {fd=48, events=POLLIN}, {fd=288, events=POLLIN}], 7, {tv_sec=0, tv_nsec=0}, NULL, 8) = 0 (Timeout) An strace of kwin_wayland is very busy. Video ram was an issue for my problems. Before Wayland, I could have multiple accounts open on this computer and the old video card. Will have to look at kscreenlocker for any bugs. With an nvdia video card, looking at the ram usage is important. I am starting to feel that Wayland requires much more vram than X did. FWIW, checking, via ssh, a machine that has gone into monitor off mode and only one account open. kscreenlocker is using 286MiB . kwin_wayland is only using 123MiB. Plasmashell is at 148MiB. One account open and locked with monitors off and the Vram usage is still 1.58G Of course, there are multiple firefox windows open. One of the large memory usage items is /usr/bin/maliit-keyboard which is almost as much as firefox. If there is anything that I can test, please request and may require details on how to test.