I’m on Debian Sid, and today upon boot up Xorg failed to start. I checked logs and saw it was not picking up the NVIDIA driver. I temporarily uninstalled all nvidia* packages from my system and installed xserver-xorg-video-nouveau. This got my Xorg back up and running, but now came the task of troubleshooting why NVIDIA’s propietary module was failing to get loaded.
Upon reinstalling the nvidia-driver metapackage, I did some tests, and this is the relevant info I found
$ sudo dkms status
nvidia-current, 418.74, 4.19.0-5-amd64, x86_64: installed
$ sudo /sbin/modprobe -v nvidia
install modprobe -i nvidia-current $CMDLINE_OPTS
insmod /lib/modules/4.19.0-5-amd64/updates/dkms/nvidia-current.ko
modprobe: ERROR: could not insert 'nvidia_current': Invalid argument
modprobe: ERROR: ../libkmod/libkmod-module.c:979 command_do() Error running install command for nvidia
modprobe: ERROR: could not insert 'nvidia': Operation not permitted
Does anyone know what might be causing this?
Running the modprobe above without sudo gives me a “Key expired” message on the nvidia module, which would suggest a conflict with Secure Boot, but that is disabled on my BIOS (I’ve double checked). I already tried booting on different kernel versions (4.19.0-5, 4.19.0-4, 4.19.0-3) and all gave the same error.
I’m very confused since it was working just fine up until last night, yet I didn’t change any settings related to this and I’m not sure if aptitude has a way to check which packages specifically I updated yesterday. If it matters, I have a GTX 760.
Nvidia driver always* break after a system update. I believe the proper steps are to uninstall the drivers completely, update the system then reinstall the drivers after that.
I don’t run Linux or nVidia so can’t comment further but this appear a lot around here and that seems to be the main suggestion I think.
*I don’t know if of is always but it sure seems like it.
Good news. I sent a bug report and the maintainers got back to me.
They suggested sudo dpkg-reconfigure nvidia-kernel-dkms . This still made me unable to load the nvidia module However, they also told me how to check the apt logs to see what exactly I had upgraded the day before the bug (/var/log/apt/history.log). The only seemingly related package was the new kernel versions. So I upgraded once again today, and sure enough there was a new version for the kernel packages. Did a reboot, and now everything is fine. The nvidia module is loaded automatically again
TL;DR nvidia-driver was not at fault, my best guess is that some bug slipped through to the kernel
It is not a bug. When you upgrade the kernel, you need to ensure that the NV KMS driver supports the kernel version. The same goes if you upgrade XORG server. It is part of the price of using team green’s official driver. It is recommended to check with the NV site to see if they support the version of X and Linux Kernel that you want to upgrade to.
Sometimes a reconfigure of the NV dkms will work when there is a minor kernel version bump, but still, you would need to do a dpkg-reconfigure to resetup the package and generate the kernel modules.