Kernel 6.1.19 doesn't display boot messages, booting into older kernel does

have you try without theses options ?

By now I’ve completely stripped it down and removed almost all the options and it still isn’t working. I even uninstalled grub, deleted /etc/default/grub and did a fresh install of grub and still no luck. At this point I’m almost ready to give up if it weren’t for the fact that it works with the Live-System from USB (but this seems to be a very different grub setup)

SUCCESS! (Well, kind of…)
I think I finally figured out the problem, but don’t have a solution yet. The culprit are the non-free nvidia-drivers. Since the new kernel they not only break my startup messages but seemingly all low level displays. For example I noticed that I also can’t access a text console with Ctrl+Alt+F2 anymore. Once I switched to the open source drivers the text displays worked as the supposed to. But now I’ve to do without graphics acceleration and desktop effects.

So: Where do I go from here? Could this be some kind of wrong setting that I can fix? The system asking for a resolution that the nvidia-driver doesn’t support or anything like this? Or is this a bug I need to report (if so: where?)

is it nvidia-setting missing ?
can you check pacman -Qs nvidia-setting ?

see [Testing Update] 2023-03-25 - Kernels, Firefox, Deepin, GNOME, NVIDIA, Wine - #3 by stephane

Hu, no.

nvidia-settings was indeed not installed. I’ve added it but that didn’t fix my problem. Do you have a particular setting in mind that could help me?

Now that you figured out it is video driver related, maybe provide journal of an unsuccessful boot after reinstalling Nvidia drivers?

I’ve pasted a Journal for a boot with enabled nvidia drivers here: https://pastebin.com/wDUEcfn4
Since I’ve no idea what to look for I’ve pasted the whole thing. But of cause I’ll be happy to provide a more “focused” journal, if you could help me what to look for/how to create it.

I don’t know if this is related to your problem, but I noticed this:

Mai 14 21:54:32 roedelkiste kernel: NVRM: The NVIDIA probe routine was not called for 1 device(s).
Mai 14 21:54:32 roedelkiste kernel: NVRM: This can occur when a driver such as: 
                                    NVRM: nouveau, rivafb, nvidiafb or rivatv 
                                    NVRM: was loaded and obtained ownership of the NVIDIA device(s).
Mai 14 21:54:32 roedelkiste kernel: NVRM: Try unloading the conflicting kernel module (and/or
                                    NVRM: reconfigure your kernel without the conflicting
                                    NVRM: driver(s)), then try loading the NVIDIA kernel module
                                    NVRM: again.

Do you have a file in /etc/modprobe.d to blacklist nouveau module? I have one created automatically by mhwd. File is mhwd-gpu.conf:

##
## Generated by mhwd - Manjaro Hardware Detection
##
 
blacklist nouveau
blacklist ttm
blacklist drm_kms_helper
blacklist drm

Another thing is that you seem to have a little complex configuration and you have two graphics cards, where the old one (with a legacy driver) is connected to your only monitor and the new one is used only for a virtual machine. Is this right? I didn’t see any place where you mention quite a relevant thing like this.

You have a graphics problem and I would try to debug it removing everything that could mess with the graphics and probably a second graphics card is clearly something that could affect in this case

Do you have a file in /etc/modprobe.d to blacklist nouveau module? I have one created automatically by mhwd. File is mhwd-gpu.conf:

Yes, I’ve the same file with the same content.

Another thing is that you seem to have a little complex configuration and you have two graphics cards, where the old one (with a legacy driver) is connected to your only monitor and the new one is used only for a virtual machine. Is this right?

Almost, the second card is also connected to a monitor and I checked its output during the boot process. There is nothing there. But yes, after that it is only used for the virtual machine. Sorry, that I forgot to mention this. But since it was working with the old Kernel, I didn’t think that this might be a problem.

I tried removing the second card but it didn’t solve the problem :frowning:

So by accident I stumbled upon a solution for my problem. It looks like Kernel 6 has a problem with loading the vfio_pci kernel module early, that are needed to pass a GPU to a virtual machine.

This seems to block graphics output until the regular graphics drivers are loaded. See this forum thread

The solution seems to be either to also load the graphics driver early (add them to the module-list in /etc/mkinitcpio.conf) or load vfio_pci via modprobe.conf instead. For a little more details see the Arch Wiki. For me the solution was to:

  1. remove all vfio-entries from /etc/mkinitcpio.conf
  2. add the following to /etc/modprobe.d/vfio.conf

softdep nouveau pre: vfio-pci
softdep nvidia pre: vfio-pci
softdep drm pre: vfio-pci

  1. rebuild with sudo mkinitcpio -P
  2. reboot

This topic was automatically closed 2 days after the last reply. New replies are no longer allowed.