You are not logged in.
@seth I just tested that.
Added
options nvidia-drm modeset=1
to /etc/modprobe.d/nvidia.conf
Tryed booting with "initcall_blacklist=simpledrm_platform_driver_init" kernel parameter instead of "nvidia_drm.modeset=1"
Result: all I got is a black screen. Also when i reboot i get the following message:
I cannot boot now, so cannot do further testing for the moment, seem something is broken, do you know what happened ?
Offline
There's a kernel error lost above the left-top corner of your monitor and you cannot mount the root device.
If this isn't because of the blocked simpledrm_platform_driver_init you might have screwed your initramfs and lack some kernel module to access the root partition.
Or, w/ the recent shenanigans, maybe some critical firmware?
Also "lsblk -f" to make sure the UUID is correct.
Any chance to get a better display of the kernel error/warning (the stuff above the end-trace)
Offline
Sorry guys, knocked out for the moment. I'm prepping a bootable recovery usb with the arch iso, so i can chroot and mkinitcpio to regenerate the initramfs. Will be back as soon as possible, for the moment being, it seems that removing the "nvidia_drm.modeset=1" kernel hack will break compatibility with our 390xx cards
UPDATE: my system is back online
By the way, what broke my system was:
Added
options nvidia-drm modeset=1
to /etc/modprobe.d/nvidia.conf
Tryed booting with "initcall_blacklist=simpledrm_platform_driver_init" kernel parameter instead of "nvidia_drm.modeset=1"
Result: all I got is a black screen. Also when i reboot i got the message:
failed to mount UUID=xxxx on real root. You are now being dropped into an emergency shell
This was solved by using usb stick to boot into arch iso and then issued the following commands
mount /dev/sda2 /mnt
Threw the following message:
BTRFS error devide sda2 in btrfs_replay_log errno:-5 IO failure (Failed to recover log tree)
So I had to run:
btrfs rescue zero-log /dev/sda2
mount -o rw,usebackuproot,rescue=all /dev/sda2 /mnt
mount --bind /dev /mnt/dev
mount --bind /proc /mnt/proc
mount --bind /sys /mnt/sys
mount --bind /run /mnt/run
mkdir -p /mnt/boot/efi
mount /dev/sda1 /mnt/boot/efi
chroot /mnt /bin/bash
mkinitcpio -P
exit
umount -R /mnt
reboot
also removed the file
/etc/modprobe.d/nvidia.conf
and reverted to the original kernel parameters, that now look like this:
linux /boot/vmlinuz-linux root=UUID=40f88f1c-b02d-4aa0-a584-6774c1d2d3e0 rw loglevel=3 quiet nvidia-drm.modeset=1 rcutree.rcu_idle_gp_delay=1 zswap.enabled=0
instead of
linux /boot/vmlinuz-linux root=UUID=40f88f1c-b02d-4aa0-a584-6774c1d2d3e0 rw loglevel=3 quiet initcall_blacklist=simpledrm_platform_driver_init rcutree.rcu_idle_gp_delay=1 zswap.enabled=0
I assume my nvidia 390xx card required the nvidia-drm.modeset=1 kernel parameter, but what may have been the real issue that broke the partition ?
probably adding
initcall_blacklist=simpledrm_platform_driver_init
disabled part of the framebuffer used during early init.
It looks like system froze or panicked during early boot and that's what broke my sda2 partition
Last edited by canolucas (2025-06-25 10:05:27)
Offline