0

I apologize for the crosspost, but I posted this question and some related ones in askubuntu and they just aren't getting any traction there, even after I took my meager rep there and applied a bounty. Can someone here help?


I've been struggling with getting my nvidia GPU to work, as detailed in this question and this question I posted recently. Everything was working fine until a few weeks ago, when suddenly, with no obvious associated event, the nVidia card stopped producing output under linux (works fine in Windows). Since then, I've done a lot of digging and found this incredibly in depth answer which listed all sorts of troubleshooting tools and tips.

One of the tips was to list the driver parameters using the following command:

sudo sh -c "ls /sys/module/<driver-module-from-lshw>/parameters/ | xargs -L1 -i{} sh -c 'echo -n {} \" \" ; cat /sys/module/<driver-module-from-lshw>/parameters/{}'"

I have two video cards - one is radeon, the other nvidia. The output of the above command with the radeon driver generates many parameters. The nvidia one gives a single parameter, which seems strange to me.

Shouldn't a working driver have more parameters? I'm asking this to see if this is a clue for my larger problem.

At the very least, can someone post samples of their working nvidia driver parameters?

$ driver=nvidia
$ sudo sh -c "ls /sys/module/$driver/parameters/ | xargs -L1 -i{} sh -c 'echo -n {} \" \" ; cat /sys/module/$driver/parameters/{}'"
nv_cap_enable_devfs  1
--------------------------------------------
$ driver=radeon
$ sudo sh -c "ls /sys/module/$driver/parameters/ | xargs -L1 -i{} sh -c 'echo -n {} \" \" ; cat /sys/module/$driver/parameters/{}'"
agpmode  -1
aspm  -1
audio  -1
auxch  -1
backlight  -1
bapm  -1
benchmark  0
cik_support  1
connector_table  0
deep_color  0
disp_priority  0
dpm  -1
dynclks  -1
fastfb  0
gartsize  1024
hard_reset  0
hw_i2c  0
lockup_timeout  10000
modeset  1
msi  -1
mst  0
no_wb  0
pcie_gen2  -1
r4xx_atom  0
runpm  -1
si_support  1
test  0
tv  1
use_pflipirq  2
uvd  1
vce  1
vm_block_size  12
vm_size  8
vramlimit  0

System details:

  • Ryzen 5, no integrated graphics
  • B450 Tomahawk Max motherboard
  • ADATA SX8100 512 GB SSD
  • Nvidia GeForce 1660 main GPU
  • Radeon HD 5450 secondary GPU
  • Dual boot Ubuntu 20.04 and windows 10
  • UEFI firmware
  • No overclocking or other tweaks
  • no xorg.conf, using auto-config

in case it's helpful, I'll post the output of various commands in the referenced post to give details of my system:

sudo sh -c "lsb_release -sd; dmidecode -s system-product-name; echo ==; lshw -c display; echo ==; xrandr --verbose; echo ==; cat /etx/X11/xorg.conf"

cat /var/log/Xorg.0.log

ls /sys/class/drm/*/edid | xargs -i{} sh -c "echo {}; parse-edid < {}"

xtrace xrandr --output HDMI-1-0 --mode 1920x1080 --right-of HDMI-1

modinfo nvidia

cat /proc/driver/nvidia/version gives:

NVRM version: NVIDIA UNIX x86_64 Kernel Module  455.38  Thu Oct 22 06:06:59 UTC 2020
GCC version:  gcc version 10.2.0 (Ubuntu 10.2.0-13ubuntu1)
11
  • Check with "lsmod | grep nvidia", you may be looking for another nvidia module. – Gerard H. Pille Dec 24 '20 at 0:28
  • @GerardH.Pille: lsmod | grep nvidia lists these modules: nvidia, nvidia_modeset, nvidia_drm, nvidia_uvm, i2c_nividia_gpu. lspci -vnn shows that the kernel modules in use for my nvidia card is "nvidia". listed kernel modules are "nvidiafb, nouveau, nvidia_drm, nvidia". I'm not sure which other module I should be looking for? – rothloup Dec 24 '20 at 2:52
  • Where is nouveau "listed"? Run modinfo for the nvidia_drm and _modeset. Why do you expect more parameters? Never once did I set a kernel parameter for a proprietary nvidia driver before I got myself a laptop with hybrid amd/nvidia graphics. Even there it was no use, can't get the very expensive tu116m to work on linux. – Gerard H. Pille Dec 24 '20 at 3:27
  • @GerardH.Pille: in the output of lspci -vnn, it lists "Kernel driver in use: nvidia" followed by "Kernel modules: nvidiafb, nouveau, nvidia_drm, nvidia" on the next line. and I'm not trying to "set" a kernel parameter. I'm trying to diagnose if there is something wrong with my driver, since my card isn't producing any output. It was working about a month ago, and works under windows. See my linked questions in the OP. – rothloup Dec 24 '20 at 5:22
  • @GerardH.Pille: 'modinfo nvidia_drm': paste.ubuntu.com/p/tWtXFjJnrc 'modinfo nvidia_modeset': paste.ubuntu.com/p/rVG2GF2dNJ what am I looking for in the output of this command? – rothloup Dec 24 '20 at 5:25
1

According to a gentoo discussion forum this may be a fix:

option nvidia-drm modeset=1

put this in a /etc/modprobe.d/XXX.conf file.

1
  • I tried this, both in the modprobe file and by passing "nvidia-drm.modeset=1" directly in the kernel parameters. It appeared to make no impact. – rothloup Dec 26 '20 at 21:49

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.