Running F43 with Plasma/KDE, kernel is 7.0.13 (I think). I have an Nvidia card installed.
It was a clean install a couple of months ago and was working okay. I installed a bunch of Plasma updates this morning. Now all I get is a black screen with a blinking cursor in the upper left corner of the screen. (It sat and blinked for about 4 hours while I did something else)
I tried booting to another kernel - no luck. I tried booting to the rescue image - no luck.
Looking through the posts, others have managed to reach a console screen to make changes. I don’t seem able to get to the console. Any suggestions other than a re-install?
Boot with the kernel parameter systemd.unit=emergency.target, then view your journalctl logs for the previous boot (-b -1) and see what was complaining bitterly.
Given that no previous kernels worked I suspect it’s something which is causing KDE or the greeter to soil itself - might be enough to just remove the contents of ~/.cache. It won’t tell you what precisely was the culprit but it may well you get back into a normal session a little sooner.
If you have a Live Installer USB you can check drive “health” and make sure the system drive isn’t full. After mounting the root partition of your system drive you can run journalctl in a terminal to view the journal entries for boots that failed:
-D DIR, --directory=DIR
Takes a directory path as argument. If specified, journalctl will operate
on the specified journal directory DIR instead of the default runtime and
system journal paths.
Look for <root mount location>/var/log/journal/<ID string> and use that with the -D directory option.
"no backend specified, automatically choosing drm\n
“Accepting client connection on sockets: QList(“Wayland-0\”)\n”
"kwin_core: Failed to open /dev/dri/card0 device (in yellow)
watchdog: watchdog0: watchdog did not stop (in red)
a lot of entries in dark blue that I can’t read against the black screen
type ctrl+alt-f3 when you get the black screen with blinking cursor.
You should then see a login prompt. You may need to wait a short while for it to appear.
I got to the console and looked at the journal. I’m not sure quite what I’m looking for. There is a kernel section where it looks like nouveau is negotiating to use the Nvidia driver, but could be normal. I thought I’d removed nouveau drivers but this might be unrelated.
computer sitting at console prompt started giving output
nouveau 0000:04:00.0: dim: DDC responded, but no EDID for HDMI-A-!
You need to install the nvidia drivers from rpmfusion.
If you already did that then you need to rebuild the nvidia drivers.
I think the command is sudo akmods --rebuild.
If that is successful reboot.
Normally you should not attempt to remove nouveau or the nvidia-gpu-firmware.noarch package. Nouveau is used as a fallback when Nvidia has an (unfortunately all to common) issue.
It might help us understand your issue if you could post details of what you did to “remove nouveau” and also what you did to “restore” it. If you can boot using nouveau then you should be better “protected” against problems with the RPMFusion packages.
It would help if you showed the exact command you used and the exact output it produced, preferably cut directly from the terminal and pasted directly into here as pre-formatted text.
I can see that you say you installed org-x11-drv-nvidia-580xx and earlier you used the correct name of xorg-x11-drv-nvidia-580xx so I assume that’s a typo in your posting, but without seeing precisely what command you executed and precisely what it returned, assuming is all we can do.
Ah, my bad. In F43 those packages aren’t named akmod-nvidia-580xx and xorg-x11-drv-nvidia-580xx. You can just use akmod-nvidia and xorg-x11-drv-nvidia, which in F43 give you version 580 of the driver.
It’s not immediately obvious why this happened, but you could try reinstalling them, see if it works, and if not then at least it should be possible to get logs to see what’s going on.
Please read and follow the instructions in the common issues topic that is directly related to that issue.
Note that the 580xx packages are only in the rpmfusion-nonfree-updates repo for fedora 44, and that at this exact moment even those are broken with one of the dependent packages not properly updated to match the rest.
However, the latest 580xx driver packages are in the rpmfusion-nonfree-updates-testing repo and they appear to work properly
You can enable the rpmfusion repos then install the 580xx drivers before upgrading to f44 with this command. sudo dnf install akmod-nvidia-580xx xorg-x11-drv-nvidia-580xx-cuda --releasever=44 --enablerepo rpmfusion-nonfree-updates-testing.
If the driver is already installed then substitute upgrade for install in the command above