Bug 497647 - Mouse clicks randomly stop having any effect, move still works
Summary: Mouse clicks randomly stop having any effect, move still works
Status: REPORTED
Alias: None
Product: kwin
Classification: Plasma
Component: libinput (show other bugs)
Version: 6.2.4
Platform: NixOS Linux
: NOR normal
Target Milestone: ---
Assignee: KWin default assignee
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2024-12-18 12:12 UTC by Leo Gaspard
Modified: 2024-12-29 09:50 UTC (History)
0 users

See Also:
Latest Commit:
Version Fixed In:
Sentry Crash Report:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Leo Gaspard 2024-12-18 12:12:15 UTC
SUMMARY

At seemingly random times, mouse clicks stop having any impact, both on the UI and on the applications. The mouse still moves fine. I also tried mapping some keyboard keys to mouse click and mouse move, and my keyboard is able to move the mouse but not to click when the issue happens. Scrolling also fails.

`sudo libinput debug-events` does show the mouse click events when clicking, but it has no impact on KDE. This leaves only keyboard navigation to use the computer. Unfortunately I forgot to screenshot the exact messages when this last happened, but it showed basically the same thing as it now shows, which is POINTER_BUTTON BTN_LEFT pressed and then released on two lines.

Keyboard-navigating to the "Sleep" option, sleeping and resuming seems to solve the issue, until the next time it happens.

I switched to KDE a few days ago, and this has already happened to me a few times.

STEPS TO REPRODUCE

Unfortunately unknown, but please let me know if there are any other commands I should run while the bug happens in order to give you additional information!

SOFTWARE/OS VERSIONS

Operating System: NixOS 25.05
KDE Plasma Version: 6.2.4
KDE Frameworks Version: 6.8.0
Qt Version: 6.8.0
Kernel Version: 6.6.64 (64-bit)
Graphics Platform: Wayland
Processors: 16 × AMD Ryzen 9 PRO 6950H with Radeon Graphics
Memory: 30.6 Gio of RAM
Graphics Processor: AMD Radeon Graphics
Comment 1 Leo Gaspard 2024-12-19 07:48:27 UTC
The issue just happened again, this time actually while going out of sleep mode. Going to sleep and resuming still fixed the issue. During that time, I used the keyboard to run the `libinput debug-events` command, and here is exactly what is logged on a mouse click that has no impact on the UI:

 event18  POINTER_BUTTON          +1.068s	BTN_LEFT (272) pressed, seat count: 1
 event18  POINTER_BUTTON          +1.140s	BTN_LEFT (272) released, seat count: 0

If you know any other intermediary point I could check, or a way to trace all input events received by plasma to understand how they're being handled exactly, I'll be happy to test them out next time the issue happens!
Comment 2 Leo Gaspard 2024-12-24 12:33:54 UTC
I actually just found something interesting: the issue happened again, but this time Firefox still has the mouse buttons working. And if I switch to another workspace, click, drag down and release, nothing happens in the application. But if I switch back to Firefox, it has selected text on the current web page.

If I open a new firefox window (hoping it being the same process would help) then click on it, it doesn't do anything. But if the other firefox window was underneath it (the one who does receive clicks), then the other firefox window gets pulled to top of screen (above the new window) upon clicking on the new firefox window.

This issue might be due to one window eventually taking control of the mouse clicks and not releasing it? But it's still strange that Plasma would not receive clicks on eg. its bottom application bar, but would still make the only mouse-working window the topmost window upon clicking on another window that's supposed to be above it.

I'm running on wayland, in case it could help.
Comment 3 Leo Gaspard 2024-12-24 12:38:50 UTC
Oh well. Just after writing this message, I randomly selected text in the bug report text, and then clicked elsewhere to unselect… and the "click elsewhere" did not do anything. It seems like the wayland compositor stopped transmitting events to the firefox window too. Maybe focus went to some widget that I could not identify, but I was unable to do anything else than put to sleep using the keyboard and resume, which led me to here.
Comment 4 Leo Gaspard 2024-12-24 15:12:20 UTC
I finally noticed that the "Debugging Plasma" wiki page was not the one I should have followed, and found the following interesting things in my logs:

déc. 24 14:08:02 coegni kwin_wayland[2739]: kf.windowsystem: static bool KX11Extras::mapViewport() may only be used on X11
déc. 24 14:08:02 coegni kwin_wayland[2739]: kwin_scene_opengl: 0x2: GL_INVALID_VALUE in glTexStorage2D(width, height or depth < 1)
déc. 24 14:08:02 coegni kwin_wayland[2739]: kwin_scene_opengl: Invalid framebuffer status:  "GL_FRAMEBUFFER_INCOMPLETE_ATTACHMENT"
déc. 24 14:08:02 coegni kwin_wayland[2739]: kwin_scene_opengl: 0x2: GL_INVALID_OPERATION in glBindFramebuffer(non-gen name)
déc. 24 14:08:02 coegni kwin_wayland[2739]: kwin_scene_opengl: 0x2: GL_INVALID_FRAMEBUFFER_OPERATION in glClear(incomplete framebuffer)
déc. 24 14:08:06 coegni kwin_wayland[2739]: kf.windowsystem: static bool KX11Extras::mapViewport() may only be used on X11
déc. 24 14:08:06 coegni kwin_wayland[2739]: kwin_scene_opengl: 0x2: GL_INVALID_VALUE in glTexStorage2D(width, height or depth < 1)
déc. 24 14:08:06 coegni kwin_wayland[2739]: kwin_scene_opengl: Invalid framebuffer status:  "GL_FRAMEBUFFER_INCOMPLETE_ATTACHMENT"
déc. 24 14:08:06 coegni kwin_wayland[2739]: kwin_scene_opengl: 0x2: GL_INVALID_OPERATION in glBindFramebuffer(non-gen name)
déc. 24 14:08:06 coegni kwin_wayland[2739]: kwin_scene_opengl: 0x2: GL_INVALID_FRAMEBUFFER_OPERATION in glClear(incomplete framebuffer)
déc. 24 14:42:52 coegni kwin_wayland[2739]: kwin_libinput: Libinput: event2  - Logitech MX Vertical: client bug: event processing lagging behind by 21ms, your system is too slow
déc. 24 14:48:44 coegni kwin_wayland[2739]: kf.windowsystem: static bool KX11Extras::mapViewport() may only be used on X11
déc. 24 14:48:44 coegni kwin_wayland[2739]: kwin_scene_opengl: 0x2: GL_INVALID_VALUE in glTexStorage2D(width, height or depth < 1)
déc. 24 14:48:44 coegni kwin_wayland[2739]: kwin_scene_opengl: Invalid framebuffer status:  "GL_FRAMEBUFFER_INCOMPLETE_ATTACHMENT"
déc. 24 14:48:44 coegni kwin_wayland[2739]: kwin_scene_opengl: 0x2: GL_INVALID_OPERATION in glBindFramebuffer(non-gen name)
déc. 24 14:48:44 coegni kwin_wayland[2739]: kwin_scene_opengl: 0x2: GL_INVALID_FRAMEBUFFER_OPERATION in glClear(incomplete framebuffer)
déc. 24 14:54:30 coegni kwin_wayland[2739]: kwin_libinput: Libinput: event2  - Logitech MX Vertical: client bug: event processing lagging behind by 21ms, your system is too slow
déc. 24 14:57:55 coegni kwin_wayland[2739]: kwin_libinput: Libinput: event2  - Logitech MX Vertical: client bug: event processing lagging behind by 21ms, your system is too slow
déc. 24 15:06:34 coegni kwin_wayland[2739]: kwin_libinput: Libinput: event2  - Logitech MX Vertical: client bug: event processing lagging behind by 25ms, your system is too slow
déc. 24 15:11:29 coegni kwin_wayland[2739]: kwin_libinput: Libinput: event2  - Logitech MX Vertical: client bug: event processing lagging behind by 26ms, your system is too slow
déc. 24 15:11:29 coegni kwin_wayland[2739]: kwin_libinput: Libinput: event2  - Logitech MX Vertical: WARNING: log rate limit exceeded (5 msgs per 60min). Discarding future messages.
déc. 24 15:21:22 coegni kwin_wayland[2739]: kf.windowsystem: static bool KX11Extras::mapViewport() may only be used on X11
déc. 24 15:21:22 coegni kwin_wayland[2739]: kwin_scene_opengl: 0x2: GL_INVALID_VALUE in glTexStorage2D(width, height or depth < 1)
déc. 24 15:21:22 coegni kwin_wayland[2739]: kwin_scene_opengl: Invalid framebuffer status:  "GL_FRAMEBUFFER_INCOMPLETE_ATTACHMENT"
déc. 24 15:21:22 coegni kwin_wayland[2739]: kwin_scene_opengl: 0x2: GL_INVALID_OPERATION in glBindFramebuffer(non-gen name)
déc. 24 15:21:22 coegni kwin_wayland[2739]: kwin_scene_opengl: 0x2: GL_INVALID_FRAMEBUFFER_OPERATION in glClear(incomplete framebuffer)

[Around 15:50: the issue happened, no new logs since then]

[After the resume from suspend:]
déc. 24 15:59:14 coegni kwin_wayland[2739]: kwin_libinput: Libinput: client bug: timer event22 debounce: scheduled expiry is in the past (-3527ms), your system is too slow
déc. 24 15:59:14 coegni kwin_wayland[2739]: kwin_libinput: Libinput: client bug: timer event22 debounce short: scheduled expiry is in the past (-3540ms), your system is too slow

In particular, the Logitech MX Vertical is the mouse I'm using when the issue happens, though when it happens the touchpad's buttons are affected too. And actually, event22 is "TPPS/2 Elan TrackPoint", so AFAIU one of my laptop's mouse inputs.

Unfortunately I didn't find a way to disable log rate limiting so I can confirm the issue is indeed due to the "system too slow" warning message.

Also, it stays weird to me that the mouse is still moving fine, and only button clicks have no impact.
Comment 5 Leo Gaspard 2024-12-29 09:50:14 UTC
Interesting data point: I just spent a few days away from home, with the computer not plugged into its dock, and the issue did not reproduce.

I guess I'll try to bisect plugging which device is causing this, though it'll be painful as my home office setup is pretty hard-wired nowadays.