10

I've read this [SOLVED] cannot set nvidia-drm.modset=1, and this nvidia-drm Failed to grab modeset ownership, and this [Dual system, Ubuntu 20.04]nvidia-gpu: i2c timeout error; ucsi_ccg: i2c_transfer failed -110, ucsi_ccg_init failed and many others. The first article says to "edit /etc/modprobe.d/nvidia.conf or /etc/modprobe.d/nvidia-kernel-common.conf" and add the line options nvidia-drm modeset=1. I don't have either of those files but there is an /etc/modprobe.d/nvidia-graphics-drivers-kms.conf which already has that line present:

# This file was generated by nvidia-driver-535
# Set value to 0 to disable modesetting
options nvidia-drm modeset=1

Yet, I continue to get the drm error in dmesg:

[   17.478652] [drm:nv_drm_master_set [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000900] Failed to grab modeset ownership
[   20.149255] systemd-journald[572]: File /var/log/journal/10b971ac1a304176906b1f6a23827476/user-1000.journal corrupted or uncleanly shut down, renaming and replacing.
[   20.706536] [drm:nv_drm_master_set [nvidia_drm]] *ERROR* [nvidia-drm] [GPU ID 0x00000900] Failed to grab modeset ownership

What I don't know is how much of an issue this is. I'm am investigating an infrequent failure to resume from suspend issue so, while I'd like to get rid of this message, it would be helpful to know if this error might be related to the issue resume issue.

AdvApp
  • 347
  • 1
  • 5
  • 15

3 Answers3

7
sudo nano /etc/default/grub
Add kernel option to the line GRUB_CMDLINE_LINUX_DEFAULT
GRUB_CMDLINE_LINUX_DEFAULT="nvidia-drm.modeset=1 nvidia-drm.fbdev=1"
sudo update-initramfs -u # probably isn't needed but did it anyway. shouldn't harm anything
sudo update-grub 
reboot
sudo cat /sys/module/nvidia_drm/parameters/modeset 
and you should see "Y"
if you 
sudo dmesg | grep drm
errors should be gone.
0

To deal with the i2c error create this file:

sudo nano /etc/modprobe.d/blacklist_i2c-nvidia-gpu.conf 

and add this line to it:

blacklist i2c_nvidia_gpu

Save and exit, then reboot.

sudo dmesg | grep i2c

The error should be gone.

zx485
  • 2,865
0

According to nvdia, this is a harmless message. Also, another message states that you can get rid of it by adding fbdev=1 parameter to nvidia-drm.

JRo
  • 73