You are not logged in.
Pages: 1
The problem affects just SOME of the workstations we have. It appeared approx. 6 months ago and persist with subsequent upgrades:
Hardware (in short): nVidia Pascal (GP107 / GTX 1050), AMD Matisse/Vermeer motherboard, BenQ XL2410T monitor.
Config (in short): up-to-date, XFCE and MATE, X11 (no Wayland) LightDM, nvidia drivers (I tried also 550xx from AUR).
If I press Ctrl-Alt-Fx for switching to a tty then a black screen appears and it takes ~ 2-3 minutes! before it switches to a tty. Back (Ctrl-Alt-F7) it takes the same time. Moreover - if I have switched from a LidghtDM or MATE I finally return to my session; when I switch from and XFCE session to a tty and back - after a loooong time a black screen with a cursor appears; nothing more.
Absolutely no errors in logs nor dmesg nor journald.
I've tried 1.7 from https://wiki.archlinux.org/title/NVIDIA/Troubleshooting - nothing.
I've tried 5.6 and 5.10 from https://wiki.archlinux.org/title/Xfce#Troubleshooting - nothing.
Any ideas welcome.
Offline
Try adding
nvidia-drm.modeset=1 nvidia-drm.fbdev=0to your kernel parameters (the cmdline, not modprobe!) and check whether you can reproduce.
Offline
Guess I'm expected to update the Wiki regarding this per:
• https://gitlab.archlinux.org/archlinux/ … ote_218166
We waited in hopes that nVidia would fix the fbdev issues for older GPUs, but seems that's not in the cards (pun).
Dragging my feet at the weight of responsibility, but it will likely be as simple as stating those kernel parameter are required for Pascal and older.
I try keep the thread current on the nVidia forums as well. No ack from nVidia.
• Nvidia driver isn’t compatible with simpledrm so boot output and TTYs are blank
Offline
Try adding
nvidia-drm.modeset=1 nvidia-drm.fbdev=0to your kernel parameters (the cmdline, not modprobe!) and check whether you can reproduce.
Doesn't change anything :-(
Offline
Please post your complete system journal for the boot:
sudo journalctl -b | curl -F 'file=@-' 0x0.stafter causing this and the boot should also be with the suggested kernel parameters.
Offline
Offline
First off: "vga=0x318" - why's that there?
If that's not it:
Feb 07 10:02:49 localhost kernel: vesafb: mode is 1024x768x32, linelength=4096, pages=1
Feb 07 10:02:49 localhost kernel: vesafb: scrolling: redraw
Feb 07 10:02:49 localhost kernel: vesafb: Truecolor: size=8:8:8:8, shift=24:16:8:0
Feb 07 10:02:49 localhost kernel: vesafb: framebuffer at 0xf1000000, mapped to 0x0000000006247519, using 6144k, total 16384k
Feb 07 10:02:49 localhost kernel: fb0: VESA VGA frame buffer deviceNo efifb nor simpledrm (the latter because it's nuked by the parameter), no indication of a nother VGA device.
Also no nvidia/drm errors.
Is the journal supposed to cover such incident?
Is this still a problem w/ the 570xx driver (and 6.13 kernel)?
Offline
"vga=0x318" - the same partition is cloned to many computers. Some use monitors for which forcing this text mode is necessary. But I tried to comment it out and it doesn't help.
I guess the reason may be some obsolete, hard to be found, blocking package or systemd component. But I have no idea how to identify it.
I'll try the newest driver tomorrow. But, honestly, I wouldn't expect any change...
Offline
The nvidia driver always had trouble switching the framebuffer between GUI and console - it's odd that the behavior doesn't change between the vga and framebuffer console, though.
If you don't care for a framebuffer console anyway and are using grub try https://wiki.archlinux.org/title/GRUB/T … ramebuffer.
Alternatively flip to "nvidia_drm.fbdev=1"
Offline
Pages: 1