No external displays detected and X Server Settings blank on Ubuntu, Driver 470

nvidia-bug-report.log.gz (146.9 KB)

This morning I was working on my laptop, restarted and suddenly the external monitor (DisplayPort) stopped working. It isn’t detected at all. I tested the HDMI port and that is also not working. To my knowledge both ports are hardwired to the Nvidia GPU. When I open the Nvidia X Server Settings program, it’s a blank screen. I have never used that before so I don’t know if that problem showed up at the same time.

I tried purging and reinstalling the 470 drivers twice. I also tried upgrading to the (untested) 490 driver. When I disable the Nvidia driver and use the XOrg driver, the external monitors work. I also tried running “prime-select nvidia”.

Information about my machine/OS:

Intel Core i7-10875H
Intel UHD Graphics (CML GT2)
TU104M (RTX 2080 SUPER Mobile /MaxQ)
Ubuntu 21.10 (64 bit)
Linux Kernel 5.11.0-18-generic
GNOME Version 40.4.0
Wayland Windowing System

Looking at the logs, it can see the external monitor connected:

[ 17.191] (–) NVIDIA(GPU-0):
[ 17.191] (–) NVIDIA(GPU-0): LG Electronics LG HDR DQHD (DFP-3): connected
[ 17.192] (–) NVIDIA(GPU-0): LG Electronics LG HDR DQHD (DFP-3): Internal DisplayPort
[ 17.192] (–) NVIDIA(GPU-0): LG Electronics LG HDR DQHD (DFP-3): 2660.0 MHz maximum pixel clock

Don’t know why it says “Internal Displayport”, but the LG DQHD is the external monitor.

I also saw this:

Nov 12 06:46:08 DanUbuntuLaptop kernel: [170229.432062] NVRM: API mismatch: the client has the version 470.82.00, but
Nov 12 06:46:08 DanUbuntuLaptop kernel: [170229.432062] NVRM: this kernel module has the version 470.74. Please
Nov 12 06:46:08 DanUbuntuLaptop kernel: [170229.432062] NVRM: make sure that this kernel module and all NVIDIA driver
Nov 12 06:46:08 DanUbuntuLaptop kernel: [170229.432062] NVRM: components have the same version.

After that there are several places complaining that it couldn’t communicate with the drivers.

I was able to resolve this. The “mismatch” lines in the bug report spurred me to check “dkms status”, which confirmed the kernel version was 5.13.0-21-generic, which is different than what was running. I then determined that 5.13.0-21-generic had been installed (probably when I upgrade Ubuntu from 21.04 to 21.10), but the grub entry never got updated, and running grub-update didn’t fix it. I have a multiboot setup on my laptop that includes Ubuntu, OpenSUSE and Windows. The OpenSUSE grub is what is used for selecting the OS. I needed to boot into OpenSUSE and update the bootloader from there. Once I did that, the boot entry was set to the correct kernel version, then I reinstalled the NVIDIA drivers and resolved the issues.

Moral of the story: if you have these issues, compare “uname -r” to “dkms status” and make sure they match.

This topic was automatically closed 14 days after the last reply. New replies are no longer allowed.