[Solved] Nvidia Quadro K600 nvidia-470xx-dkms driver blank screen

stiw47 · 2024-02-23 10:35:30

For few years, my Quadro K600 worked with just installing nvidia-470xx-dkms and nvidia-470xx-utils from AUR:

$ yay -S nvidia-470xx-dkms

I don't even remember that I was dealing with set grub cmdline parameter and mkiiniticpio modules/hooks before it broke - it simply worked. Few months ago, with never kernels, display stopped working. When I boot my PC, I can see kernel messages until certain point when SDDM should appear. At this point, instead of SDDM which appears before (and should star KDE), now I'm getting blank screen. This is actually home server and I have another desktop PC, laptop, etc., so I did not care too much, but since it is already connected to TV via DP to HDMI cable, I would like to bring desktop back if possible.

My Archlinux is up to date.
Regarding DE, didn't configured anything special. Just followed basic steps - installed xorg, installed plasma and sddm, and enabled sddm.service - and this worked for few years.

All few years while it worked, I used linux-lqx kernel - not working any more, there is no desktop after kernel boot messages.
Already tried with default kernel.
Currently, in this latest try, I am on linux-lts:

[stiw47@archmedia ~]$ uname -a
Linux archmedia 6.6.17-1-lts #1 SMP PREEMPT_DYNAMIC Sat, 17 Feb 2024 08:28:50 +0000 x86_64 GNU/Linux

I have installed nvidia-470xx stuffs from AUR. On top, this latest time, I even didn't used yay, installed everything by manually downloading PKGBUILD, then makepkg -s, then sudo pacman -U *.pkg.tar.zst

[stiw47@archmedia ~]$ pacman -Q | grep nvidia
lib32-nvidia-470xx-utils 470.239.06-1
lib32-opencl-nvidia-470xx 470.239.06-1
nvidia-470xx-dkms 470.239.06-1
nvidia-470xx-settings 470.223.02-1
nvidia-470xx-settings-debug 470.223.02-1
nvidia-470xx-utils 470.239.06-1
opencl-nvidia-470xx 470.239.06-1

And this is my card:

[stiw47@archmedia ~]$ lspci | grep -E 'VGA|3D'
01:00.0 VGA compatible controller: NVIDIA Corporation GK107GL [Quadro K600] (rev a1)

I should have not remains from nouveau driver any more (since I also tried to make it work with it). As far as I remember, it worked with nouveau few years ago, but there were some glitches, so I switched to nvidia-470xx-dkms then, and was happy for few years:

[stiw47@archmedia ~]$ pacman -Q | grep nouveau
[stiw47@archmedia ~]$ pacman -Q | grep xf86
lib32-libxxf86vm 1.1.5-1
libxxf86vm 1.1.5-1
xf86-input-libinput 1.4.0-1
xf86-video-intel 1:2.99.917+923+gb74b67f0-1
xf86-video-vesa 2.6.0-1

This is my grub cmdline:

[stiw47@archmedia ~]$ grep GRUB_CMDLINE_LINUX_DEFAULT /etc/default/grub
GRUB_CMDLINE_LINUX_DEFAULT="loglevel=3 noquiet nvidia-drm.modeset=1"

I've done grub update after adding modeset=1 parameter:

[stiw47@archmedia ~]$ sudo grub-mkconfig -o /boot/grub/grub.cfg 
Generating grub configuration file ...
Found linux image: /boot/vmlinuz-linux-lts
Found initrd image: /boot/initramfs-linux-lts.img
Found fallback initrd image(s) in /boot:  initramfs-linux-lts-fallback.img
Found linux image: /boot/vmlinuz-linux
Found initrd image: /boot/initramfs-linux.img
Found fallback initrd image(s) in /boot:  initramfs-linux-fallback.img
Warning: os-prober will not be executed to detect other bootable partitions.
Systems on them will not be added to the GRUB boot configuration.
Check GRUB_DISABLE_OS_PROBER documentation entry.
Adding boot menu entry for UEFI Firmware Settings ...
done

This is from mkinitcpio.conf:

[stiw47@archmedia ~]$ grep -E 'MODULES=\(|HOOKS=\(' /etc/mkinitcpio.conf | grep -v \#
MODULES=(nvidia nvidia_modeset nvidia_uvm nvidia_drm)
HOOKS=(base udev autodetect modconf keyboard keymap consolefont block filesystems fsck)

And I done mkinitcpio -P after changes, it shown no errors:

[stiw47@archmedia ~]$ sudo mkinitcpio -P
==> Building image from preset: /etc/mkinitcpio.d/linux-lts.preset: 'default'
==> Using default configuration file: '/etc/mkinitcpio.conf'
  -> -k /boot/vmlinuz-linux-lts -g /boot/initramfs-linux-lts.img --microcode /boot/*-ucode.img
==> Starting build: '6.6.17-1-lts'
  -> Running build hook: [base]
  -> Running build hook: [udev]
  -> Running build hook: [autodetect]
  -> Running build hook: [modconf]
  -> Running build hook: [keyboard]
==> WARNING: Possibly missing firmware for module: 'xhci_pci'
  -> Running build hook: [keymap]
  -> Running build hook: [consolefont]
==> WARNING: consolefont: no font found in configuration
  -> Running build hook: [block]
  -> Running build hook: [filesystems]
  -> Running build hook: [fsck]
==> Generating module dependencies
==> Creating zstd-compressed initcpio image: '/boot/initramfs-linux-lts.img'
==> Image generation successful
==> Building image from preset: /etc/mkinitcpio.d/linux-lts.preset: 'fallback'
==> Using default configuration file: '/etc/mkinitcpio.conf'
  -> -k /boot/vmlinuz-linux-lts -g /boot/initramfs-linux-lts-fallback.img -S autodetect --microcode /boot/*-ucode.img
==> Starting build: '6.6.17-1-lts'
  -> Running build hook: [base]
  -> Running build hook: [udev]
  -> Running build hook: [modconf]
  -> Running build hook: [keyboard]
==> WARNING: Possibly missing firmware for module: 'xhci_pci'
  -> Running build hook: [keymap]
  -> Running build hook: [consolefont]
==> WARNING: consolefont: no font found in configuration
  -> Running build hook: [block]
==> WARNING: Possibly missing firmware for module: 'qla2xxx'
==> WARNING: Possibly missing firmware for module: 'qed'
==> WARNING: Possibly missing firmware for module: 'qla1280'
==> WARNING: Possibly missing firmware for module: 'aic94xx'
==> WARNING: Possibly missing firmware for module: 'wd719x'
==> WARNING: Possibly missing firmware for module: 'bfa'
  -> Running build hook: [filesystems]
  -> Running build hook: [fsck]
==> Generating module dependencies
==> Creating zstd-compressed initcpio image: '/boot/initramfs-linux-lts-fallback.img'
==> Image generation successful
==> Building image from preset: /etc/mkinitcpio.d/linux.preset: 'default'
==> Using configuration file: '/etc/mkinitcpio.conf'
  -> -k /boot/vmlinuz-linux -c /etc/mkinitcpio.conf -g /boot/initramfs-linux.img
==> Starting build: '6.7.5-arch1-1'
  -> Running build hook: [base]
  -> Running build hook: [udev]
  -> Running build hook: [autodetect]
  -> Running build hook: [modconf]
  -> Running build hook: [keyboard]
==> WARNING: Possibly missing firmware for module: 'xhci_pci'
  -> Running build hook: [keymap]
  -> Running build hook: [consolefont]
==> WARNING: consolefont: no font found in configuration
  -> Running build hook: [block]
  -> Running build hook: [filesystems]
  -> Running build hook: [fsck]
==> Generating module dependencies
==> Creating zstd-compressed initcpio image: '/boot/initramfs-linux.img'
==> Image generation successful
==> Building image from preset: /etc/mkinitcpio.d/linux.preset: 'fallback'
==> Using configuration file: '/etc/mkinitcpio.conf'
  -> -k /boot/vmlinuz-linux -c /etc/mkinitcpio.conf -g /boot/initramfs-linux-fallback.img -S autodetect
==> Starting build: '6.7.5-arch1-1'
  -> Running build hook: [base]
  -> Running build hook: [udev]
  -> Running build hook: [modconf]
  -> Running build hook: [keyboard]
==> WARNING: Possibly missing firmware for module: 'xhci_pci'
  -> Running build hook: [keymap]
  -> Running build hook: [consolefont]
==> WARNING: consolefont: no font found in configuration
  -> Running build hook: [block]
==> WARNING: Possibly missing firmware for module: 'qla2xxx'
==> WARNING: Possibly missing firmware for module: 'qed'
==> WARNING: Possibly missing firmware for module: 'qla1280'
==> WARNING: Possibly missing firmware for module: 'aic94xx'
==> WARNING: Possibly missing firmware for module: 'wd719x'
==> WARNING: Possibly missing firmware for module: 'bfa'
  -> Running build hook: [filesystems]
  -> Running build hook: [fsck]
==> Generating module dependencies
==> Creating zstd-compressed initcpio image: '/boot/initramfs-linux-fallback.img'
==> Image generation successful

Also added nvidia.hook to /etc/pacman.d/hooks/, but at this stage of driver installation, I think this is not mandatory for driver to works:

[stiw47@archmedia ~]$ cat /etc/pacman.d/hooks/nvidia.hook 
[Trigger]
Operation=Install
Operation=Upgrade
Operation=Remove
Type=Package
Target=nvidia-470xx-dkms
Target=linux-lts
# Adjust line(6) above to match your driver, e.g. Target=nvidia-470xx-dkms
# Change line(7) above, if you are not using the regular kernel For example, Target=linux-lts

[Action]
Description=Update Nvidia module in initcpio
Depends=mkinitcpio
When=PostTransaction
NeedsTargets
Exec=/bin/sh -c 'while read -r trg; do case $trg in linux) exit 0; esac; done; /usr/bin/mkinitcpio -P'

Rebooted after that, and as said on beginning, there is blank screen after kernel boot messages - SDDM and KDE not appears. Before this broke, at this point, SDDM and KDE would appear, and I did not changed anything in mean time, just done regular updates with sudo pacman -Syu.

Card is not broken, because if I remove all above mentioned stuff (nvidia-470xx-****, grub cmdline, revert mkinitcpio.conf to default, etc.), and boot without any graphics driver installed, I'm getting my desktop, but in fixed low resolution. But I have to add grub cmdline parameter nomodeset in this case:

[stiw47@archmedia ~]$ grep GRUB_CMDLINE_LINUX_DEFAULT /etc/default/grub
GRUB_CMDLINE_LINUX_DEFAULT="loglevel=3 noquiet nomodeset"

This is from dmesg, grep'ed for nvidia:

[root@archmedia ~]# dmesg -T | grep nvidia
[Fri Feb 23 10:23:06 2024] Command line: BOOT_IMAGE=/boot/vmlinuz-linux-lts root=UUID=02f3913c-43a6-478e-8dbf-191cd9e0aa01 rw loglevel=3 noquiet nvidia-drm.modeset=1
[Fri Feb 23 10:23:06 2024] Kernel command line: BOOT_IMAGE=/boot/vmlinuz-linux-lts root=UUID=02f3913c-43a6-478e-8dbf-191cd9e0aa01 rw loglevel=3 noquiet nvidia-drm.modeset=1
[Fri Feb 23 10:23:07 2024] nvidia: loading out-of-tree module taints kernel.
[Fri Feb 23 10:23:07 2024] nvidia: module license 'NVIDIA' taints kernel.
[Fri Feb 23 10:23:07 2024] nvidia: module verification failed: signature and/or required key missing - tainting kernel
[Fri Feb 23 10:23:07 2024] nvidia: module license taints kernel.
[Fri Feb 23 10:23:08 2024] nvidia-nvlink: Nvlink Core is being initialized, major device number 238
[Fri Feb 23 10:23:08 2024] nvidia 0000:01:00.0: vgaarb: VGA decodes changed: olddecodes=io+mem,decodes=none:owns=io+mem
[Fri Feb 23 10:23:09 2024] nvidia-modeset: Loading NVIDIA Kernel Mode Setting Driver for UNIX platforms  470.239.06  Sat Feb  3 06:03:51 UTC 2024
[Fri Feb 23 10:23:09 2024] nvidia_uvm: module uses symbols nvUvmInterfaceDisableAccessCntr from proprietary module nvidia, inheriting taint.
[Fri Feb 23 10:23:09 2024] nvidia-uvm: Loaded the UVM driver, major device number 236.
[Fri Feb 23 10:23:09 2024] [drm] [nvidia-drm] [GPU ID 0x00000100] Loading driver
[Fri Feb 23 10:23:10 2024] caller _nv000720rm+0x1ad/0x200 [nvidia] mapping multiple BARs
[Fri Feb 23 10:23:12 2024] [drm] Initialized nvidia-drm 0.0.0 20160202 for 0000:01:00.0 on minor 0
[Fri Feb 23 10:23:27 2024] [drm:drm_new_set_master] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Failed to grab modeset ownership
[Fri Feb 23 10:23:27 2024] [drm:drm_new_set_master] *ERROR* [nvidia-drm] [GPU ID 0x00000100] Failed to grab modeset ownership

Full dmesg, from reboot until point when desktop not appeared (blank screen), could be seen here.

Yeah, I am aware, it could be that this Quadro K600 could go to trash (or to be sold for some cheap money), but if someone have some advice, I would be thankful.
Please note that I am not such an expert for troubleshooting of this exotic type, so please be gently ?

This is some additional info, please let me know if you need more info.

Thank you very much!

plasma packages:

[stiw47@archmedia ~]$ pacman -Q | grep plasma
kdeplasma-addons 5.27.10-2
plasma-bigscreen 5.27.10-1
plasma-browser-integration 5.27.10-1
plasma-desktop 5.27.10-1
plasma-disks 5.27.10-1
plasma-firewall 5.27.10-1
plasma-framework5 5.115.0-1
plasma-integration 5.27.10-1
plasma-nano 5.27.10-1
plasma-nm 5.27.10-1
plasma-pa 5.27.10-1
plasma-remotecontrollers 5.27.10-1
plasma-sdk 5.27.10-1
plasma-settings 23.01.0-5
plasma-systemmonitor 5.27.10-1
plasma-thunderbolt 5.27.10-1
plasma-vault 5.27.10-1
plasma-wayland-session 5.27.10-2
plasma-welcome 5.27.10-1
plasma-workspace 5.27.10-2
plasma-workspace-wallpapers 5.27.10-1

SDDM:

[stiw47@archmedia ~]$ pacman -Q | grep sddm
sddm 0.20.0-4
sddm-kcm 5.27.10-1

Xorg:

[stiw47@archmedia ~]$ pacman -Q | grep xorg
xorg-bdftopcf 1.1.1-1
xorg-docs 1.7.2-2
xorg-font-util 1.4.1-1
xorg-fonts-100dpi 1.0.4-2
xorg-fonts-75dpi 1.0.4-1
xorg-fonts-alias-100dpi 1.0.5-1
xorg-fonts-alias-75dpi 1.0.5-1
xorg-fonts-encodings 1.0.7-1
xorg-iceauth 1.0.9-1
xorg-mkfontscale 1.2.2-1
xorg-server 21.1.11-1
xorg-server-common 21.1.11-1
xorg-server-devel 21.1.11-1
xorg-server-xephyr 21.1.11-1
xorg-server-xnest 21.1.11-1
xorg-server-xvfb 21.1.11-1
xorg-sessreg 1.1.3-1
xorg-setxkbmap 1.3.4-1
xorg-smproxy 1.0.7-1
xorg-twm 1.0.12-1
xorg-util-macros 1.20.0-2
xorg-x11perf 1.6.2-1
xorg-xauth 1.1.2-1
xorg-xbacklight 1.2.3-3
xorg-xcmsdb 1.0.6-1
xorg-xcursorgen 1.0.8-1
xorg-xdpyinfo 1.3.4-1
xorg-xdriinfo 1.0.7-1
xorg-xev 1.2.5-1
xorg-xgamma 1.0.7-1
xorg-xhost 1.0.9-1
xorg-xinit 1.4.2-1
xorg-xinput 1.6.4-1
xorg-xkbcomp 1.4.7-1
xorg-xkbevd 1.1.5-1
xorg-xkbutils 1.0.6-1
xorg-xkill 1.0.6-1
xorg-xlsatoms 1.1.4-1
xorg-xlsclients 1.1.5-1
xorg-xmessage 1.0.6-1
xorg-xmodmap 1.0.11-1
xorg-xpr 1.1.0-1
xorg-xprop 1.2.7-1
xorg-xrandr 1.5.2-1
xorg-xrdb 1.2.2-1
xorg-xrefresh 1.0.7-1
xorg-xset 1.2.5-1
xorg-xsetroot 1.1.3-1
xorg-xvinfo 1.1.5-1
xorg-xwayland 23.2.4-2
xorg-xwd 1.0.9-1
xorg-xwininfo 1.1.6-1
xorg-xwud 1.0.6-1
xorgproto 2023.2-1

mesa:

[stiw47@archmedia ~]$ pacman -Q | grep mesa
lib32-mesa 1:24.0.1-1
libva-mesa-driver 1:24.0.1-1
mesa 1:24.0.1-1
mesa-utils 9.0.0-3
opencl-clover-mesa 1:24.0.1-1
opencl-rusticl-mesa 1:24.0.1-1

Last edited by stiw47 (2024-02-23 22:09:00)

seth · 2024-02-23 15:42:42

Please remove the oversized images or replace them w/ links (but they're not relevant, you'll get at best XGA on the VESA driver w/ nomodeset)

Can you still boot the multi-user.target w/ the nvidia GPU enabled? (2nd link below)
Ideally post your Xorg log, https://wiki.archlinux.org/title/Xorg#General from there after failing to start SDDM (so we're getting the xorg log from a bad run and not the nomodeset/VESA one)

There's an open thread that lead to interference between gnome, nvidia (545xx) and mesa 24 - you could try to downgrade the latter to the previous version in isolation
If this is the problem, you'll likely find mesa_EGL and zink references in your "sudo journalctl -b -1" (last boot, assuming that was a bad one)

stiw47 · 2024-02-23 18:25:09

Thank you for the answer, I removed the images.

My default target is graphical.target

[stiw47@archmedia ~]$ systemctl get-default
graphical.target

But as said, I'm using this PC as home server, so in real life 90% of the time I'm accessing it either via SSH, or from other devices through Home Assistant/Nextcloud/Jellyfin and other services clients. Please let me know if you still need that I enable multiuser.target for some reason or this question was more like "can the PC boot, in order to get logs, etc."? Anyway, PC is up and running, without GUI, and with Nvidia GPU enabled (nvidia-470xx-dkms driver is installed at the moment).

Xorg.0.log can be found here: https://pastebin.com/Ac0nu5ux I don't understand it. I mean, I can see a lot of that about drivers, but have no idea what to do further. Thank you in advance.
It is from bad run, without nomodeset, and with nvidia-drm.modeset=1, and blank screen when SDDM should appear.

Will try to find the thread you mentioned, but I am on KDE not gnome, and my GPU cannot use newer driver, it have to use 470. And there is no any of mesa_EGL in journalctl since bad boot:

[root@archmedia ~]# journalctl -b -1 | grep mesa_EGL
[root@archmedia ~]#

Last edited by stiw47 (2024-02-23 18:29:09)

seth · 2024-02-23 20:51:59

[    11.780] (II) NVIDIA(0): Detected PCI Express Link width: 16X
[    11.784] (--) NVIDIA(GPU-0): CRT-0: disconnected
[    11.784] (--) NVIDIA(GPU-0): CRT-0: 400.0 MHz maximum pixel clock
[    11.784] (--) NVIDIA(GPU-0): 
[    11.788] (--) NVIDIA(GPU-0): DFP-0: disconnected
[    11.788] (--) NVIDIA(GPU-0): DFP-0: Internal TMDS
[    11.788] (--) NVIDIA(GPU-0): DFP-0: 330.0 MHz maximum pixel clock
[    11.788] (--) NVIDIA(GPU-0): 
[    11.866] (--) NVIDIA(GPU-0): DFP-1: disconnected
[    11.866] (--) NVIDIA(GPU-0): DFP-1: Internal TMDS
[    11.866] (--) NVIDIA(GPU-0): DFP-1: 165.0 MHz maximum pixel clock
[    11.866] (--) NVIDIA(GPU-0): 
[    11.866] (--) NVIDIA(GPU-0): DFP-2: disconnected
[    11.866] (--) NVIDIA(GPU-0): DFP-2: Internal DisplayPort
[    11.866] (--) NVIDIA(GPU-0): DFP-2: 960.0 MHz maximum pixel clock
[    11.866] (--) NVIDIA(GPU-0): 
[    11.866] (==) NVIDIA(0): 
[    11.866] (==) NVIDIA(0): No modes were requested; the default mode "nvidia-auto-select"
[    11.866] (==) NVIDIA(0):     will be used as the requested mode.
[    11.866] (==) NVIDIA(0): 
[    11.866] (--) NVIDIA(0): No enabled display devices found; starting anyway because
[    11.866] (--) NVIDIA(0):     AllowEmptyInitialConfiguration is enabled
[    11.866] (II) NVIDIA(0): Validated MetaModes:
[    11.866] (II) NVIDIA(0):     "NULL"

There's no output detected on the GPU, reading your OP again

but since it is already connected to TV via DP to HDMI cable, I would like to bring desktop back if possible

this is likely the cause and either the adapter at fault or the TV shows up too late for the handshake.

Does it help to
1. re-attach the TV after the boot?
2. boot the multi-user.target first, then start X11 w/ a delay

stiw47 · 2024-02-23 22:08:36

I cannot believe such a king you are, and such an idiot I am ?. I connected my LG ultrawide monitor with rela and qulity DP to DP cable, and it is working. So it seems this junk of DP to HDMI cable is broken (again). It is not adapter btw, but just simple cable with DP on one end, and HDMI on another. Main problem is that this GPU has no HDMI output, and as we already knows, most of the TVs have HDMI only, not DP, and in my hud it is hard to find good and quality this type of cable. It is second one broken. With the first one I had a luck it was visible damaged. Thank you for your time, will see tomorrow to buy some cable.

seth · 2024-02-23 22:15:40

Yeah, that still counts as an adapter

The HDCP handshake is fragile by design - make the cable as short as you can get away with and sure that its sufficiently shielded from any interference.
If there's any error during the handshake, it cannot be corrected but will cause a timeout and complete failure - and the adapter transition doesn't make things any more reliable…

Arch Linux

#1 2024-02-23 10:35:30

[Solved] Nvidia Quadro K600 nvidia-470xx-dkms driver blank screen

#2 2024-02-23 15:42:42

Re: [Solved] Nvidia Quadro K600 nvidia-470xx-dkms driver blank screen

#3 2024-02-23 18:25:09

Re: [Solved] Nvidia Quadro K600 nvidia-470xx-dkms driver blank screen

#4 2024-02-23 20:51:59

Re: [Solved] Nvidia Quadro K600 nvidia-470xx-dkms driver blank screen

#5 2024-02-23 22:08:36

Re: [Solved] Nvidia Quadro K600 nvidia-470xx-dkms driver blank screen

#6 2024-02-23 22:15:40

Re: [Solved] Nvidia Quadro K600 nvidia-470xx-dkms driver blank screen

Board footer