I am running Fedora 40 KDE using Wayland and an NVIDIA GeForce RTX 2060. Everything was running great until all the kernel updates that followed 6.8.7. Since then, every single kernel causes a black screen after the GRUB menu. The system is not frozen, as far as I know, as the keyboard’s num/caps/scroll lock keys still work. I am basically stuck on 6.8.7, which is infuriating, as I specifically chose Fedora for new kernels and regular updates. My system is running BTRFS and is encrypted with LUKS.
What I have tried:
Disabled Fast Boot (off by default).
Disabled Secure Boot (off by default).
Disabled iGPU.
Slowly typing in my LUKS password at the black screen, hoping it will progress SDDM.
Waiting 20+ minutes at the black screen.
Try to access a tty with the CTRL+F keys. Can’t get to any of them.
journalctl looks normal to me except for a wall of text related to PipeWire.
I changed the GRUB config, removing rhgb and changing quiet to noquiet, and was able to boot. I hit enter after 20 seconds to bring up the LUKS prompt, since nothing was happening. There were no errors in GRUB prior to this, it just listed the connected USB devices. No errors in systemd after entering LUKS password.
The only thing I can think of is that the NVIDIA drivers are being loaded before LUKS.
I don’t use LUKS at all (never have any need) so I don’t know what the normal modus operandi is for it. Hopefully someone with experience of what is normally does and looks like can chime in.
I’m pleased you get to be able to boot a newer kernel; shows there’s nothing inherently wrong with the installation; all we need to do now is work out what the root cause is!
One thing that has been noticed with the newer 6.11 kernels is that sometimes it is necessary to have a special environment variable to manage the black screen issue.
There are several threads about that.
The fix for many is to edit the /etc/environment file and insert a line GSK_RENDERER=nglthen reboot.
Sorry for the super late reply, but basically everything was working until a restart this morning, so I thought I would post an update.
I was using @anothermindbomb suggestion of removing rhgb and quiet up to the middle of February. It worked perfectly fine, but the boot process is much slower with all the init information.
I then switched to what @computersavvy suggested, adding GSK_RENDERER=ngl and adding back rhgb and quiet my grub config. That worked just fine until last night, when I upgraded from kernel 6.13.5 to 6.13.9. Upon restart this morning, I was greeted with the infamous black screen and could do nothing. On top of that, the other 2 installed kernels, now too, present a black screen. I was forced to make a live USB, chroot in, remove rhgb and quiet, then update GRUB. Now I am running the latest kernel without issue.
For anyone who uses NVIDIA and LUKS, I suggest you remove rhgb and quiet and save yourself the pain I had to go through.
I’m not sure if Fedora can resolve this issue themselves, or if LUKS needs an update, or NVIDIA is to blame, but I won’t be playing with my grub settings again.
That is one of the first things I do for every new install. I can why “rhgb quiet” would be useful for something like a cubicle farm, but for machine managed by the user, it save time in the long run.
I never had to touch those or any other GRUB settings on my Ubuntu or Arch systems, BUT those systems don’t use NVIDIA, which seems to be the main factor here.
