I have a Dell Precision 7520 with a Quadro M1200. I’m running a clean installation of KUbuntu 18.04. I’ve added the latest NVidia driver (396.45) from https://launchpad.net/~graphics-drivers/+archive/ubuntu/ppa.
My external monitors are both Dell P2715Q (4k) monitors.
I’ve searched around, and not found anything that has really even pointed me in the direction of a solution…
“Out of the box” only one monitor is displaying anything. The second is just black. It is recognized by NVida – I can see it in the NVidia Settings app, where it shows up as DP-5, and is part of X screen 0. The laptop monitor and working display show up in this tool as “PRIME Display”; they are also part of X screen 0.
In the KDE display settings, my laptop monitor shows up as eDP-1-1 and the working monitor as DP-1-1.
The 2nd monitor is enabled – I can see this both in the NVidia settings & KDE settings.
I can tell that something about the non-working monitor is working, because windows can be placed there, even though I don’t see anything…
I had this all working with 16.04, but due to a confluence of events, I had to re-install everything.
Any pointers would be greatly appreciated.
nvidia-bug-report.log-2018-07-30.gz (130 KB)
Logs look normal, monitor is detected and added. Check if it works when you disable KDE’s kscreen2 service.
No change, unfortunately.
Did you check if enabling prime sync helps, i.e. create
/etc/modprobe.d/99-nvidia-modeset.conf
options nvidia-drm modeset=1
run
sudo update-initramfs
afterwards, reboot.
Still no luck :-(
I’m not sure how to verify that configuration got applied, though…
I also just remembered one other thing… Based on another post (I believe from you) indicating that the generated /usr/share/X11/xorg.conf.d/11-nvidia-prime.conf file has a bad path (it does – it refers to “ModulePath “/x86_64-linux-gnu/nvidia/xorg””), I went ahead and created a symlink from /x86_64-linux-gnu to /usr/lib/x86_64-linux-gnu. (That is the path used in /usr/share/X11/xorg.conf.d/10-nvidia.conf.)
The other thing I’ve noticed is that when I use “nvidia-select” to change to “intel” and reboot the scale of my windows and text is much nicer (smaller, cleaner) than when I change to “nvidia”. I’m assuming this is a different issue, but I thought I’d add that here just in case it sheds any light on what may be happening…
Thanks so much for any other ideas you may have…
jay
To check if the modeset=1 option is applied,
sudo cat /sys/module/nvidia_drm/parameters/modeset
should return ‘Y’.
Just to make sure, did you swap cables/monitors to rule out a defective monitor?
To check if this a plain driver/xserver bug, create a ~/.xinitrc
xrandr --setprovideroutputsource modesetting NVIDIA-0
xrandr --auto
xterm &
exec xterm
switch to vt, stop sddm and start a plain xserver using
startx
then see if you can move the mouse on all three monitors.
One more bug I came across is, when wayland is running it sometimes prevents the use of monitors connected to the nvidia gpu on prime setups. Maybe check if sddm (or whatever DM you’re using) is running in wayland mode.
To check if the connector isn’t broken, a test case would be to run an xserver on just the nvidia, avoiding prime.
Have this as /etc/X11/xorg.conf
Section "ServerLayout"
Identifier "layout"
Screen 0 "nvidia" 0 0
EndSection
Section "Screen"
Identifier "nvidia"
Device "nvidia"
EndSection
Section "Device"
Identifier "nvidia"
Driver "nvidia"
VendorName "NVIDIA Corporation"
BusID "PCI:1:0:0"
Option "AllowEmptyInitialConfiguration"
EndSection
and a ~/.xinitrc
exec xterm
and use startx
This should give you an xterm on just the monitor connected to the nvidia gpu.
OK… I am a bit embarrassed… It seems to be something with the docking station I’m using. I verify that I can switch the plugs and then the monitor which was working goes black, and the other one starts working.
It is weird, because clearly both monitors are being detected.
My next step will be to get the correct adapters and verify that if I plug the cables directly into the laptop they both work.
THANK you for your time!
jay