Good afternoon, a few minutes ago my Fedora 38 Workstation system updated the kernel to version 6.3.11. When I reboot the system I see that the Nvidia drivers stopped working. At first I didn’t notice it, in fact I thought the new kernel had fixed the screen flickering I’ve been seeing since I upgraded to Fedora 38, but I immediately saw that the GPU went to full load without having any program running. So I opened Nvidia Settings and noticed what is shown in the following image:
I opened the terminal and ran the following commands which gave me as a result:
lspci -k | grep -A 2 -i "VGA"
09:00.0 VGA compatible controller: NVIDIA Corporation TU104 [GeForce RTX 2080 Rev. A] (rev a1)
Subsystem: Gigabyte Technology Co., Ltd Device 37b1
Kernel driver in use: nouveau
ERROR: NVIDIA driver is not loaded
(nvidia-settings:9214): GLib-GObject-CRITICAL **: 14:31:06.242: g_object_unref: assertion 'G_IS_OBJECT (object)' failed
** (nvidia-settings:9214): CRITICAL **: 14:31:06.243: ctk_powermode_new: assertion '(ctrl_target != NULL) && (ctrl_target->h != NULL)' failed
modprobe: FATAL: Module nvidia not found in directory /lib/modules/6.3.11-200.fc38.x86_64
modinfo -F version nvidia
modinfo: ERROR: Module nvidia not found.
What do those messages mean, is there any way to resolve them?
Thanks for help.
Could it be a similar issue to this one?
You should mention the model of your NVIDIA GPU and how you installed the drivers (e.g, rpmfusion-nonfree, NVIDIA, or ?). Sometimes the kernel is released before NVIDIA has updated their drivers, and rpmfusion takes some time to create Fedora packages. Support for older GPU’s may take longer or could be discontinued for newer kernels.
The RTX 2080 should not have done that.
How long did you wait after the update before you did the reboot?
Impatience often causes this kind of problem because the system has to compile the new kernel modules for the nvidia drivers after the kernel is updated, and rebooting too soon may cause corruption in that process.
One may recover by doing
sudo dnf remove kmod-nvidia-6.3.11* then after that completes run
sudo akmods --force to rebuild the modules. Waiting at least 2 minutes after that completes then rebooting should properly load the drivers on the new kernel. (note that I am assuming you installed the drivers from rpmfusion. If installed from another source the process is different.)
The nvidia 535.54.03 driver from rpmfusion works well with the 6.3.11 kernel for me.
Hi @bryanmoore, Yes, that’s the same problem I had! I just corrected it. Thank you very much!
Yes, I rebooted the system almost immediately after updating the kernel. My mistake, I admit it. Fortunately I solved it by reinstalling the drivers.
What I do see that remains, is the screen flickering. This is how I have the GRUB configured. Is there anything else I should do? Thanks for help.
GRUB_CMDLINE_LINUX="rd.driver.blacklist=nouveau modprobe.blacklist=nouveau resume=UUID=fce70eb5-ff76-4cb4-b62f-3fc81d8da9ea rhgb quiet nvidia-drm.modeset=1"
This seems to be a common issue—you can check out threads like this one—and there doesn’t seem to be a solution…