System Freeze When Moving Cursor to External Monitor on Fedora 43 with KDE Plasma 6.6.3 (Wayland)

Good morning everyone,

I need help with an issue that started after performing system updates. My setup is:

  • Operating System: Fedora Linux 43
  • KDE Plasma Version: 6.6.3
  • KDE Frameworks Version: 6.24.0
  • Qt Version: 6.10.2
  • Kernel Version: 6.19.10-200.fc43.x86_64 (64-bit)
  • Graphics Platform: Wayland

Before the update, everything worked fine. However, now, whenever I connect an external monitor to my laptop and move the cursor toward it, the system freezes completely.

Does anyone know what might be causing this or how I should diagnose and fix it?

Any guidance would be greatly appreciated.

I encountered the same problem after yesterday’s update. I have an nvidia 1650 video card with nvidia kernel module open.

It sounds similar to this issue:

@ulysses Which version of the driver is installed? I believe (from RPMFusion bug tracker) that 580.142 was released for a while and then rolled back a couple of days ago.

@sumerald Are you also using an Nvidia GPU?

1 Like

The freezes stop after a while. In journalctl I see this:

Apr 01 19:29:43 fedora kwin_wayland[2386]: Invalid framebuffer status:  "GL_FRAMEBUFFER_INCOMPLETE_MISSING_ATTACHMENT"

Hello, P G. You already helped me last time with Nvidia problems. I’m grateful to you. Well done.

ulysses@fedora ~ [127]
āÆ rpm -qa | grep nvidia
nvidia-gpu-firmware-20260309-1.fc43.noarch
xorg-x11-drv-nvidia-cuda-libs-580.142-1.fc43.x86_64
nvidia-modprobe-580.142-1.fc43.x86_64
nvidia-persistenced-580.142-1.fc43.x86_64
xorg-x11-drv-nvidia-kmodsrc-580.142-1.fc43.x86_64
xorg-x11-drv-nvidia-cuda-libs-580.142-1.fc43.i686
xorg-x11-drv-nvidia-libs-580.142-1.fc43.i686
xorg-x11-drv-nvidia-libs-580.142-1.fc43.x86_64
akmod-nvidia-580.142-1.fc43.x86_64
xorg-x11-drv-nvidia-cuda-580.142-1.fc43.x86_64
xorg-x11-drv-nvidia-580.142-1.fc43.x86_64
nvidia-settings-580.142-1.fc43.x86_64
xorg-x11-drv-nvidia-power-580.142-1.fc43.x86_64
kmod-nvidia-6.19.9-200.fc43.x86_64-580.142-1.fc43.x86_64
kmod-nvidia-6.19.10-200.fc43.x86_64-580.142-1.fc43.x86_64
ulysses@fedora ~  
ulysses@fedora ~ [1]
āÆ rpm -qi nvidia-gpu-firmware
Name        : nvidia-gpu-firmware
Version     : 20260309
Release     : 1.fc43
Architecture: noarch
Install Date: Sat 14 Mar 2026 11:11:54 AM MSK
Group       : Unspecified
Size        : 105915219
License     : LicenseRef-Callaway-Redistributable-no-modification-permitted
Signature   :
              RSA/SHA256, Tue 10 Mar 2026 05:03:52 PM MSK, Key ID 829b606631645531
Source RPM  : linux-firmware-20260309-1.fc43.src.rpm
Build Date  : Tue 10 Mar 2026 02:08:48 PM MSK
Build Host  : buildvm-ppc64le-32.rdu3.fedoraproject.org
Packager    : Fedora Project
Vendor      : Fedora Project
URL         : http://www.kernel.org/
Bug URL     : https://bugz.fedoraproject.org/linux-firmware
Summary     : Firmware for NVIDIA GPUs
Description :
Firmware for NVIDIA GPUs.
ulysses@fedora ~  
āÆ rpm -qi xorg-x11-drv-nvidia  
Name        : xorg-x11-drv-nvidia
Epoch       : 3
Version     : 580.142
Release     : 1.fc43
Architecture: x86_64
Install Date: Mon 30 Mar 2026 08:29:26 PM MSK
Group       : Unspecified
Size        : 177609054
License     : Redistributable, no modification permitted
Signature   :
              RSA/SHA256, Wed 18 Mar 2026 01:24:36 PM MSK, Key ID 6a2af96194843c65
              RSA/SHA256, Wed 18 Mar 2026 01:24:34 PM MSK, Key ID 6a2af96194843c65
Source RPM  : xorg-x11-drv-nvidia-580.142-1.fc43.src.rpm
Build Date  : Wed 18 Mar 2026 02:39:28 AM MSK
Build Host  : buildvm-01.online.rpmfusion.net
Packager    : RPM Fusion
Vendor      : RPM Fusion
URL         : https://www.nvidia.com/
Bug URL     : https://bugz.rpmfusion.org/xorg-x11-drv-nvidia
Summary     : NVIDIA's proprietary display driver for NVIDIA graphic cards
Description :
This package provides the most recent NVIDIA display driver which allows for
hardware accelerated rendering with current NVIDIA chipsets series.
Fermi and Kelper GPUs NOT supported by this release.


For the full product support list, please consult the release notes
http://download.nvidia.com/XFree86/Linux-x86_64/580.142/README/index.html


Please use the following documentation:
http://rpmfusion.org/Howto/NVIDIA 
ulysses@fedora ~
āÆ rpm -qi kmod-nvidia-6.19.10-200.fc43.x86_64  
Name        : kmod-nvidia-6.19.10-200.fc43.x86_64
Epoch       : 3
Version     : 580.142
Release     : 1.fc43
Architecture: x86_64
Install Date: Tue 31 Mar 2026 06:10:25 PM MSK
Group       : Unspecified
Size        : 36772697
License     : Redistributable, no modification permitted
Signature   : (none)
Source RPM  : nvidia-kmod-580.142-1.fc43.src.rpm
Build Date  : Tue 31 Mar 2026 06:09:25 PM MSK
Build Host  : fedora
URL         : https://www.nvidia.com/
Summary     : nvidia kernel module(s) for 6.19.10-200.fc43.x86_64
Description :
This package provides the nvidia kernel modules built for the Linux
kernel 6.19.10-200.fc43.x86_64 for the x86_64 family of processors.

I read in the news that KDE updated the interaction with the video buffer or something like that. I’m not very strong in this topic. Could this be a KDE bug?

Looks like you have the 580.142 version that causes problems in this scenario. (btw it seems to be specifically about having a mix of internal and external monitors. Other multiple monitor scenarios are fine - I have two monitors connected to my desktop, and I don’t get any problems.)

Try this:

sudo dnf distro-sync --allowerasing

Hopefully you’ll see the Nvidia packages revert to 580.126.18, picking up the rollback that RPMFusion did.

Then wait a few minutes before rebooting.

My impression from this thread is that Nvidia think it’s a bug on their side.

1 Like

Unfortunately, the command you suggested does nothing. Nothing to do.
My akmod-nvidia was built successfully, unlike other guys.
I tried this and the freezes disappeared at startup:

āÆ sudo akmods --rebuild --verbose
[sudo] password for ulysses:
Checking kmods exist for 6.19.10-200.fc43.x86_64 [OK]
Building and installing nvidia-kmod [OK]

I don’t know, maybe it’s a coincidence.
In any case, I don’t want to downgrade and make the system more fragile just yet. I’ll wait for a fix from the developers.

I assumed only the closed kernel modules had that bug? But I did not check the kernel-open source.
@ulysses does adding KWIN_FORCE_SW_CURSOR=1 to /etc/environment and restart prevent the issue?

According to my observations, this has nothing to do with it. Perhaps the problem is probabilistic and does not occur every reboot.

Good point. This person had the issue on a 1650, so they might have been running either the closed or the open kernel.

The other reports I’ve seen have been on 10series, so must be closed kernel.

the specific code is also in the 580.142 kernel-open drivers but apparently it does not lead to a hard crash like the closed kernel modules.

2 Likes

My configuration:
Operating System: Fedora Linux 43
KDE Plasma Version: 6.6.3
KDE Frameworks Version: 6.24.0
Qt Version: 6.10.2
Kernel Version: 6.19.10-200.fc43.x86_64 (64-bit)
Graphics Platform: Wayland
Processors: 12 Ɨ AMD Ryzen 5 5600H with Radeon Graphics
Memory: 16 GiB of RAM (15.0 GiB usable)
Graphics Processor 1: AMD Radeon Graphics
Graphics Processor 2: NVIDIA GeForce GTX 1650
Manufacturer: HP
Product Name: HP Pavilion Gaming Laptop 15-ec2xxx

āÆ modinfo -l nvidia
Dual MIT/GPL

I’m using a laptop with an external monitor connected to DisplayPort.
The problem is reproduced unstable. But it is there. I restarted the computer five times, and saw freezes three times within the first 5 - 6 seconds. Then everything works fine.
As I understand it, this is either a KDE or nvidia bug. akmods is rebuilt without errors, dmesg is also fine. But in journalctl I see this consistently:

Apr 01 20:26:33 fedora kwin_wayland[2351]: Invalid framebuffer status:  "GL_FRAMEBUFFER_INCOMPLETE_MISSING_ATTACHMENT"
Apr 01 20:26:33 fedora kwin_wayland[2351]: 0x500: GL_INVALID_ENUM error generated. Invalid <face>.
Apr 01 20:26:33 fedora kwin_wayland[2351]: Invalid framebuffer status:  "GL_FRAMEBUFFER_INCOMPLETE_MISSING_ATTACHMENT"
Apr 01 20:26:33 fedora kwin_wayland[2351]: 0x500: GL_INVALID_ENUM error generated. Invalid <face>.
Apr 01 20:26:33 fedora kwin_wayland[2351]: Invalid framebuffer status:  "GL_FRAMEBUFFER_INCOMPLETE_MISSING_ATTACHMENT"
Apr 01 20:26:33 fedora kwin_wayland[2351]: 0x500: GL_INVALID_ENUM error generated. Invalid <face>.
Apr 01 20:26:33 fedora kwin_wayland[2351]: Invalid framebuffer status:  "GL_FRAMEBUFFER_INCOMPLETE_MISSING_ATTACHMENT"
Apr 01 20:26:33 fedora kwin_wayland[2351]: 0x500: GL_INVALID_ENUM error generated. Invalid <face>.
Apr 01 20:26:33 fedora kwin_wayland[2351]: Invalid framebuffer status:  "GL_FRAMEBUFFER_INCOMPLETE_MISSING_ATTACHMENT"
Apr 01 20:26:33 fedora kwin_wayland[2351]: 0x500: GL_INVALID_ENUM error generated. Invalid <face>.
Apr 01 20:26:33 fedora kwin_wayland[2351]: Invalid framebuffer status:  "GL_FRAMEBUFFER_INCOMPLETE_MISSING_ATTACHMENT"
Apr 01 20:26:33 fedora kwin_wayland[2351]: 0x500: GL_INVALID_ENUM error generated. Invalid <face>.
Apr 01 20:26:33 fedora kwin_wayland[2351]: Invalid framebuffer status:  "GL_FRAMEBUFFER_INCOMPLETE_MISSING_ATTACHMENT"
Apr 01 20:26:33 fedora kwin_wayland[2351]: 0x500: GL_INVALID_ENUM error generated. Invalid <face>.

Good evening, I wanted to thank you for your responses. At the moment, I’m still learning about Linux and I’m a little afraid of breaking the system, but I at least found a temporary solution:

sudo dnf remove ā€˜nvidia’ --exclude nvidia-gpu-firmware
sudo dnf reinstall xorg-x11-drv-nouveau
sudo dracut --force

I know I will lose some performance, but at least I can use the screens and continue with my studies. Perhaps my solution will be useful to others in the same situation.

Again, thank you very much for your help.

How did you install the Nvidia drivers? I ask that because I see you have installed more than what I have:

The ones in red are the ones I also have in my setup, being that I use a newer driver: 595.58.03-1 coming from Fedora Kinoite 44 beta.
I have no i686 libs, but that should not be any problem I guess. When the system does not need those libs they don’t hurt the system.
What I also don’t have are the kmod-nvidia lines.
Can you describe in detail how you installed the Nvidia drivers?

I use this method, copied from my own installation manual.
This is however only meant for an atomic distro. So if you don't use one then this is NOT for you:

Install RPM Fusion repositories and Nvidia drivers
1.  Install the versioned Free and Nonfree RPM Fusion repos:

1a. rpm-ostree install \
    https://mirrors.rpmfusion.org/free/fedora/rpmfusion-free-release-$(rpm -E %fedora).noarch.rpm \
    https://mirrors.rpmfusion.org/nonfree/fedora/rpmfusion-nonfree-release-$(rpm -E %fedora).noarch.rpm
1b. reboot

2.  To replace the versioned RPM Fusion repos that were previously installed with the unversioned repos:

2a. rpm-ostree update \
    --uninstall rpmfusion-free-release \
    --uninstall rpmfusion-nonfree-release \
    --install rpmfusion-free-release \
    --install rpmfusion-nonfree-release                             # These 5 lines form 1 command
2b. reboot

3.  Install Nvidia driver
3a. rpm-ostree install akmod-nvidia xorg-x11-drv-nvidia -y			#install drivers
3b. reboot										#wait with reboot for at least 5 minutes, CPU @ 0%

4.  Blacklist Nova and Nouveau drivers
4a. rpm-ostree kargs --append=rd.driver.blacklist=nouveau,nova-core --append=modprobe.blacklist=nouveau,nova-core --append=nvidia-drm.modeset=1 --append=initcall_blacklist=simpledrm_platform_driver_init
4b. reboot

5. Install CUDA driver
5a. rpm-ostree install  xorg-x11-drv-nvidia-cuda -y
5b. reboot

Many reboots but it gets the thing going.

Those lines seem fine - it’s normal to see them on a non-atomic Fedora.

1 Like

This seems excessiv to me. Are reboots 3b, 4b really mandatory? That’s worse than the other OS from Redmond.

Hello. I installed according to the instructions with rpm-fusion. akmod-nvidia, as far as I understand, is a thing that rebuilds open Nvidia kernel modules for each new kernel version.

I repeat, I probably wrote the wrong topic. Because I only experience freezes when the system starts for 5-6 seconds. Perhaps this is due to the same bug that causes more serious problems for other guys, or maybe not. In any case, my problem is not so annoying that I take any non-trivial steps to solve it. I’ll wait for a fix from upstream.

I remember when I first installed Fedora 42 KDE a year ago, the freezes were simply terrifying. Thanks to @pg-tips for helping me solve the problem. Nevertheless, even after help from P G, the interface was unsmooth and jerky, but at least usable. During the life cycle of Fedora 42 and 43, amazing work was done. The interface has become much smoother, although not perfect, compared to my other machine, which has video graphics from Intel. So, the problem I’m facing now doesn’t look so terrible. So I am ready to just wait and help in finding it if there is a request for it.

I see now that I have used the instructions for an atomic distro since that is the one I use.
Forget what I wrote if you are not using an atomic distro like Kinoite.
I wil change the post so it is only for atomic distro’s.
Sorry.