I’m new to Linux so I’ll describe my situation and context as best I can.
I’m dual booting windows 11 and fedora 43 kde on separate NVMe drives. I followed an online guide and removed my Windows drive while setting up the one with fedora then reinstalled the windows drive so I choose the os in the bios. Ran into a few issues but was able to resolve those on my own until I tried to install my Nvidia drivers. After attempting to install my drivers and restarting it booted up normal the first time and took me to the sign in screen but once I sign it it’s just a glitchy black screen my mouse is the only things visible. I tried using the keyboard shortcuts to open the terminal that didn’t work and I figured out none of the keyboard shortcuts are working for some reason idk how I messed up installing my drivers so bad that it caused that issue. But I can go to the top left corner to access desktop overview which causes the black screen to glitch out sometime flashing large red text or distorted icons and I thought this view would be my solution because there’s a search bar up there but it was false hope I tried opening the konsole from there but It won’t work in fact I could get anything to open other than the alert browser and problem reporting. I couldn’t even get it to shut down I had to hold the power button and upon restarting this time I got the message “Nvidia kernal module missing. Falling back to nouveau” but everything else was still the same. I tried reverting to a previous version of the kernal but it was having the exact same issue. I tried entering like rescue or emergency mode in the grub menu and was told I couldn’t access the console and to see sulogin(8). At this point I was completely stuck and decided to start fresh again with a new install of fedora so I once again removed the drive with windows freshly installed fedora went through the setup process again it went much smoother this time until once again I tried to install the nvidia drivers I watched more tutorials this time and tried a different method and once again got the same issue which is where I’m at now asking for any idea on what to do or how to fix this because I really don’t want to have to take my windows NVMe out again to do a fresh boot to have potentially the same issue happen again.
I have also been getting the message on startup [0.073729] RDseed32 is broken. Disabling the coresponding CPUID bit. But from what I read online I don’t think that’s related to my current issue but I’m trying to include all the details I can think of here.
My system info is a Ryzen 9800x3d, Nvidia 5080, Msi mag x870 tomahawk wifi, 32gb ddr5, and both windows and fedora are on separate 2tb nvme pcie 5.0
Sounds very much like your nvidia drivers are not built and it’s running the nouveau drivers as a fallback.
Boot up.
Switch to a normal terminal with Ctrl+Alt+F5
Sign in
Follow the instructions here to rebuild your nvidia drivers.
Ensure you wait a few minutes before rebooting to allow the drivers time to rebuild - this is what seems to get most people.
Don’t sweat the RDSEED32 error right now - it’s got nothing to do with this specific issue. There may even be a new BIOS from MSI by now to installed to sort it out.
Sorry I’m just now responding I haven’t had a chance to mess with it till now but I really appreciate your advice. At first ctrl alt f5 wasn’t working but I thought maybe it’s my keyboard so I plugged in my old keyboard and I was able to get it to work. However when I tried to uninstall the drivers using “dnf remove xorg-x11-drv-nvidia*” which I got from your link it told me the requested operation requires superuser privileges but this is the only user I made. Am I wrong in assuming there are drivers to uninstall did I just fail to install them at all in the first place or do I even need to uninstall them if I build new ones will they override the old ones. Sorry if these are stupid questions I’m very new to Linux.
There should be no need to remove anything. As you have a 5080, the only instructions you need to follow from that article are these
The message about needing superuser privs is handled by the sudo command… so reading the link above, there are 3 commands to be executed (I’ll list the text from the article that I’m referring to so you can follow along):
Current GeForce/Quadro/Tesla
Supported on current stable Xorg server release.
This driver is suitable for any GPU found in 2014 and later.
/!\ The 510+ driver is available by default on Fedora 34+ and later and has dropped support for some older Kepler GPU.
sudo dnf update -y # and reboot if you are not on the latest kernel
sudo dnf install akmod-nvidia # rhel/centos users can use kmod-nvidia instead
sudo dnf install xorg-x11-drv-nvidia-cuda #optional for cuda/nvdec/nvenc support
/!\ After the RPM transaction ends, please remember to wait until the kmod has been built. This can take up to 5 minutes on some systems.
Once the module is built, "modinfo -F version nvidia" should outputs the version of the driver such as 440.64 and not modinfo: ERROR: Module nvidia not found.
Follow those three sudo commands, above and let us know how you got on. You should get no errors spat back at you, if you do, post the error in here and we’ll get this sorted out. The latest version of the driver you’re going to build is 580.119.02.
After trying to execute the first sudo it gave me a bunch of “Curl error (6):” there’s more text after this but it’s to long for me to type so I’ll try to attach an image
I ran the command it had me put in my password so I did but didn’t output any response afterwards not sure if it was supposed to but then I tried the update command again and it gave me the same curl errors.
Would connecting to ethernet be beneficial I’m currently using the wifi built into my motherboard and it was working before this whole thing started.
Francis has already covered this overnight, but as he says, in order to download packages from the repositories you have to have a working connection to the internet in some fashion - ethernet or wired - doesn’t matter but it must be up and configured.
I don’t know if this is possible in this particular case, I have never used it, but can it be that secure boot is blocking the correct installation of the Nvidia drivers? It is a PC which also runs Windows 11 so secure boot should be enabled.
Just an idea.
It doe not block installation of the driver, but certainly does block loading the driver at boot time.
The fix is shown here as well as in the file /usr/share/doc/akmods/README.secureboot.
I got the error “Connection activation failed: No suitable device found for this connection (device enp8s0 not available because profile is not compatible with device (mismatching interface name))”
When it produced the list of connections it listed my wifi twice with the same name but different UUIDs. I checked and do not have a long enough Ethernet cable but will be picking up one tomorrow if I can’t get the wifi to connect.
Edit: I typed in the UUID for the second listing of my wifi and was able to connect
You should be able to clean out all the errors and start new if you were to go into the directory /etc/NetworkManager/system-connections/ and delete all those file there. When you restart and attempt to connect your wifi it should create a new connection profile in that same directory.
I don’t know which DE you are using – Maybe gnome? If so then when you open the settings → network (or wifi) under the ‘identity’ tab there is an option named ‘cloned address’ where you have 5 different options to select. If that is set to ‘random’ you will have a new MAC address presented every time you connect anywhere and that may lead to the type issue you show. I use one of three options everywhere for wifi.
‘random’, ‘stable per SSID’, or ‘stable’ for my laptop (which may connect away from home, and ‘permanent’ for my daily driver at home.
Your use case may determine which option to use, but you should not have 2 connection files with the same name & SSID on your system.
So thanks to the help of everyone in this thread I was able to connect to my wifi and run the commands listed on the nvidia rpm fusion link provided earlier and it all seemed to work I thought I waited 15 minutes after running the command just to be safe and checked the driver version installed and it was the correct one so I rebooted and sadly my screen was still black after signing in but then I saw this comment about disabling secure boot. So I tried to follow the link you provided I turned on my PC hit f11 to enter the bios boot menu chose the drive with fedora it took me to the grub menu then chose the UEFI firmware setting and that just took my to my bios settings and I couldn’t find anything relating to secure boot in there so am I missing something or misunderstanding the instructions.
For my systems the ‘secure boot’ setting is under the ‘security’ tab in the bios setup menu.
Bios menus vary by brand and version so you may need to search for the setting to enable or disable it.
Following the steps in the file I listed above should allow loading the driver even when secure boot is enabled, so that probably would be the preferred option.
I was able to find mine after more digging and it finally worked! This thread has been so helpful so thank you and everyone else who bothered to help out a stranger who was in way over his head lol.
