Black screen on Nvidia after updating to 20260428

@Micha @MartianStorm about broken wifi. I see it got mentioned on a mailing list here. Should be fixed soon if you have installed tumbleweed long ago and still use AppArmor instead of SELinux.

Also need to migrate to selinux somehow.

I referred to the 3 dot bug, just to make sure. Was meant to be a reply.

1 Like

There are instructions on how to do that. You’ll need to mess around in terminal to do it. IDK if it can be done through yast.

Before switching to SELinux try changing kernel parameter to apparmor=0 in bootloader settings. If it works then your wifi is blocked by AppArmor.

I edited my GRUB configuration by adding the line ā€œinitcall_blacklist=sysfb_init,ā€ then reinstalled SDDM and rebooted.

For me, the command in GRUB didn’t change anything:

  • KWin still keeps restarting the visual effects in a loop.

  • The SDDM login manager doesn’t let me enter the password right away: it always restarts once or twice before allowing me to enter the password.

I’m still using the G06 driver and kernel 7.0.6

I know you said it was an issue with the Nvidia driver related to the 7.0.X kernel,

but I was hoping your solution would work.

I’ll look into it. Do you happen to know if your system is booting in BIOS or UEFI mode? You can check if with file /sys/firmware/efi command. If it says that file is not found then you’re using BIOS mode, if it exists then your system is UEFI.

Huge apology: rookie mistake!
I forgot to run the command to update GRUB!

So, what I wrote earlier is incorrect.

  • SDDM is working properly again
  • KWin no longer restarts.

At this point, everything is back to normal.

Just one thing (not too bothersome): I no longer have access to the ESC key during the boot phase.

Shutdown and reboot don’t work

1 Like

That is to be expected with this workaround. Unfortunately you won’t see TTY until Nvidia driver kicks in.

Thanks, I will try, but I cannot test this before end of next week.

1 Like

@Lioli7k Hi, it would be interested to see the output from;

lsgpu
ls -1 /sys/class/drm/*/edid

On a working system and a non-working one…

I’ve been playing around on the MacBook Pro I have with hyprland, it has an older kepler-2 gpu and did duplicate the issue with the G05 driver, it looses the plot as it shows up as a ghost device, eg Uknown-1 instead of eDP-2…

Using the nouveau driver works fine driving eDP-2…

/sys/class/drm/card0-DP-1/edid
/sys/class/drm/card0-DP-2/edid
/sys/class/drm/card0-eDP-1/edid
/sys/class/drm/card0-HDMI-A-1/edid

The output on my machine which probably would not be working without my workaround. 595.71.05_k7.0.5_1-2.3. Not really very easy to get the output when the system is affected because not even a tty session is available unless maybe a start job is created to generate it.

1 Like

I found that my card (5070 Ti) won’t work when allowing the firmware to be installed. It tries to force install itself every time there is an update for those packages as part of the all firmware package. I think it was designed only for older cards and will interfere with the new drivers.

I remove an lock out these packages: kernel-firmware-nvidia kernel-firmware-nvidia-gspx-G06.

zypper rm kernel-firmware-nvidia kernel-firmware-nvidia-gspx-G06
zypper al kernel-firmware-nvidia kernel-firmware-nvidia-gspx-G06

Doing this allows the nvidia card to work. With those installed the driver can’t communicate with the card for some reason.

You shouldn’t use G06 at all with a 5070 Ti…G07 is the correct driver…

And it is clear why the other G06 drivers stop working when installing this package. The G06 driver is at v580 and kernel-firmware-nvidia-gspx-G06 is at v550.

It already helps to uninstall kernel-firware-all. This helps to better maintain and administrate your system, as you can choose which firmware packages are needes on your system and which can be uninstalled. This leads also to a decrease of update sizes.

I am using the G07 driver. I was using the G06 driver prior to when G07 was available though I notice the G07 driver works much better than G06 did with newer cards.

There is no kernel-firmware-nvidia-gspx-G07. It is likely no longer needed since G07. But every time there is an update to kernel-firmware-all it would force reinstall the G06 firmware even after I uninstalled it (even without the G06 driver). When it did install that firmware and I restarted, the nvidia card no longer worked (even prior to G07 driver availability) so I uninstalled it and locked it out to prevent it from interfering. I believe that firmware is meant for cards prior to Turing since it seemed to not cause an issue until I got a Turing or later card. I’ve not uninstalled kernel-firmware-all because I’ve not run into any other driver/firmware issues outside of nvidia, so I just let it install the defaults.

New kernel 7.0.7, rebuild G07 kmp… Still no issues here with GNOME and the Turing based RTX 4000.

Hopefully something is fixed for others after this dup?

1 Like

I have installed the new kernel.
I had modified my grub with ```
initcall_blacklist=sysfb_init

But to see if all is ok with kde, i would know how to remove the command

Still have issues with the latest update; running 2070super
fallback to 6.18 for now…

I have rebooted and used ā€œEā€ in grub to cancel initcall_blacklist=sysfb_init.
Kernel 7.0.7 default in use.
The problem is not solved.
Kwin reboot the display effects and sddm is not stable.

Kernel 7.0.7 with nvidia-open-driver-G07-signed-kmp-default 595.71.05_k7.0.7_1-2.4 still has the same issue on my system as well.