Almalinux 9.5 new kernel update Wayland not in function Only X11

1

HI,
After update Almalinux 9.5 to new Kernel 5.14.0-503.34.1.el9_5.x86_64 wayland is not in function and no operative. Only X11 choice is running ( Kde or Gnome).
NVIDIA GPU GeForce RTX 3080/PCIe/SSE2 is is running correctly.
Any one as some idea ?
Regards
Henri

2

Wayland and NVIDIA in general have a lot of issues currently. Check the logs to see whatā€™s wrong. Share it here if you need guidance

1 Like
3

Hello,
I checked the logs, and there are no errors. Wayland canā€™t start because ā€œ/usr/lib/udev/rules.d/61-gdm.rules,ā€ due to the NVIDIA driver settings, is running the x11 server!
I tried forcing Wayland to boot; the result was a black screen and the graphics process was frozen.
I implemented the new NVIDIA direct driver ā€œNVIDIA-Linux-x86_64-570.133.07.run,ā€ with the same result.
Iā€™m keeping this driver because it seems more efficient than the NVIDIA procedures indicated in the Almalinux wiki (both in terms of installation and functionality)!
I installed the new 9.6 update for testing. The installation was perfect. Overall, it works perfectly. All my applications previously installed on 9.5 are fine. However, I had to uninstall and reinstall the nvidia driver for the same reasons indicated previously. Operation is normal but with more fluidity and optimization for KDE! (System Info)
< Operating System: AlmaLinux 9.6
KDE Plasma Version: 5.27.11
KDE Frameworks Version: 5.115.0
Qt Version: 5.15.9
Kernel Version: 5.14.0-570.5.1.el9_6.x86_64 (64-bit)
Graphics Platform: X11
Processors: 12 Ɨ IntelĀ® XeonĀ® E-2276G CPU @ 3.80GHz
Memory: 46.5 GB of RAM
GPU: NVIDIA GeForce RTX 3080/PCIe/SSE2
Manufacturer: Dell Inc.
Product Name: PowerEdge T340 >
Iā€™m still seeing a bug with incomplete repolist display!
Kind regards,
Henri

4

Has multiple conditions. E.g.

# disable Wayland on Hi1710 chipsets
# suspend/resume services are necessary for working wayland support
# If this is a hybrid graphics laptop with vendor nvidia driver, disable wayland
# Disable wayland in situation where we're in a guest with a virtual gpu and host passthrough gpu
# For versions above 470 but lower than 510 prefer Xorg, above 510, prefer Wayland.

I have installed packages:

nvidia-driver nvidia-driver-cuda nvidia-kmod-common nvidia-modprobe nvidia-settings nvidia-xconfig dnf-plugin-nvidia kernel-devel-matched dkms

from stream ā€˜nvidia-driver:570-dkmsā€™ of ā€˜cuda-rhel9-x86_64ā€™ repo (the ā€˜dkmsā€™ is from EPEL). Current package version is 570.124.06-1.el9.

With these versions/packages all the 61-gdm.rules conditions allow Wayland.

The CUDA repo definition:

[cuda-rhel9-x86_64]
baseurl = https://developer.download.nvidia.com/compute/cuda/repos/rhel9/x86_64
enabled = 1
gpgcheck = 1
gpgkey = https://developer.download.nvidia.com/compute/cuda/repos/rhel9/x86_64/D42D0685.pub
name = cuda-rhel9-x86_64
5

Hello Jukka;
Iā€™ve already tested the proposed oficial solution. In my case, it doesnā€™t work. The X server hangs. Installing the native NVIDIA driver (570.133.07) works perfectly for both graphics and parallel computing. The X server only starts under X11 and gdm!
I sent a question to NVIDIA about this as a ā€œmessage in a bottle.ā€
To be continuedā€¦
In this case, this allows me to continue my LLM project despite everything!
Best regards,
PS: in my congig, Tests of Release 9.6 are OK !
Henri