Possible Wayland bug introduced with latest nvidia-dkms-575.57.08

seth · 2025-06-12 13:08:34

tekstryder wrote:

have the same GP107 (1050 Ti) gen card as you and zero issue.

nb.

The OP wrote:

And my secondary display doesn't work in Wayland anymore, however my third display, which is connected via Displayport chaining to the second one works. … SDDM correctly outputs to all 3 displays, X11 as well. It is just when I login to Wayland that it doesnt show any image.

https://bbs.archlinux.org/viewtopic.php … 5#p2245875

my compositor (picom) still has nice transparency blur effects and all, display still 3840x2160 @ 120Hz … But I tried to open minecraft and it is broken. … other games with varying luck … Very hit or miss.

There're multiple regression reports, but they vary in symptoms and probably GPU generation.
I will however say that nvidia officially still supports those devices, so this is simply a bug in the driver that will likely get addressed and not be ignored because the VBIOS is too old.

Edit: the caveat is btw specific to Hopper GPUs (datacenter version of Ada) and doesn't cover Ampere chips at all. Ie. doesn't apply here.

Last edited by seth (2025-06-12 13:13:04)

phpTANKER · 2025-06-12 13:19:10

tekstryder wrote:

seth wrote:
yeah - the 575xx driver has some serious issues.
phpTANKER wrote:
My old 1080 Ti is on VBIOS 86.02.39.00.23 and I don't think I can update it..
You can't. And you don't need nor want to.
Lots of FUD being thrown around in various threads.
I have the same GP107 (1050 Ti) gen card as you and zero issue.
Did you get yours straightened out? I can't even tell from this mangled thread with so many tangents.

Never got it sorted out. I tried:
- Blacklisting nova_core, nouveau (mkinitcpio -P after)
- Adding nouveau modeset=0 to modprobe.d and kernel parameters
- Adding kernel parameters and verified that they were loaded: vidia-drm.modeset=1 nvidia-drm.fbdev=0 + nvidia.NVreg_EnableGpuFirmware=0 as suggested later.
The only thing I didn't try was adding nvidia.NVreg_EnableGpuFirmware=0 to modprode and running mkinitcpio again. I will take a look later tonight.

seth · 2025-06-12 13:34:23

The only thing I didn't try was adding nvidia.NVreg_EnableGpuFirmware=0 to modprode and running mkinitcpio again.

That's a kernel parameter. They all are. They have nothing to do w/ the iniramfs, you pass them at the commandline

Can you switch the monitors around, ie. check whether it's related to the monitor or the used output?

tekstryder · 2025-06-12 13:38:40

Right. I'm following the other threads as well.

Neither GSP (there is none) nor datacenter GPU/drivers are relevant to this thread.

seth wrote:

The OP wrote:
And my secondary display doesn't work in Wayland anymore, however my third display, which is connected via Displayport chaining to the second one works. … SDDM correctly outputs to all 3 displays, X11 as well. It is just when I login to Wayland that it doesnt show any image.

Okay, so nothing has changed from original symptom.

seth wrote:

https://bbs.archlinux.org/viewtopic.php … 5#p2245875

I don't see the correlation. Maybe I'm missing something.

seth wrote:

There're multiple regression reports, but they vary in symptoms and probably GPU generation.

That's fine. They should be reported here as well:

• 575 release feedback & discussion

Tangent: I've been having much success recently in having my issues reported upstream acked by devs/mods, assigned internal tracking numbers, and some fixed and/or worked around.
Most recently this one, which is kind of a biggie, and I'll be updating the relevant Arch GitLab bug with my new information.

I see OP has replied...

Last edited by tekstryder (2025-06-12 13:52:17)

tekstryder · 2025-06-12 13:43:51

phpTANKER wrote:

Never got it sorted out. I tried:
- Blacklisting nova_core, nouveau (mkinitcpio -P after)
- Adding nouveau modeset=0 to modprobe.d and kernel parameters
- Adding kernel parameters and verified that they were loaded: vidia-drm.modeset=1 nvidia-drm.fbdev=0 + nvidia.NVreg_EnableGpuFirmware=0 as suggested later.
The only thing I didn't try was adding nvidia.NVreg_EnableGpuFirmware=0 to modprode and running mkinitcpio again. I will take a look later tonight.

- Get rid of any blacklisting. nvidia-utils handles all of this that's required.
- Purge anything you added nouveau related. Again handled by nvidia-utils.
- GPU firmware is irrelevant to our generation card. Remove that.

Try to isolate each display and check the result. E.g..move one of the working displays into the daisy chain to replace the non-functioning one.

Are the displays all of disparate native resolutions?

EDIT: ninja'd by:

seth wrote:

Can you switch the monitors around, ie. check whether it's related to the monitor or the used output?

Last edited by tekstryder (2025-06-12 13:48:00)

phpTANKER · 2025-06-25 18:36:57

Sorry for the very late responses:

- Get rid of any blacklisting. nvidia-utils handles all of this that's required.
Done

- Purge anything you added nouveau related. Again handled by nvidia-utils.
Done

- GPU firmware is irrelevant to our generation card. Remove that.
Done

Try to isolate each display and check the result. E.g..move one of the working displays into the daisy chain to replace the non-functioning one.
Done, the issue seems to be the 2nd display attached to the daisy chain, no matter which one.

Are the displays all of disparate native resolutions?
Yes, but they are not the same resolutions: (DP-1: 1920x1080 144Hz @ 100% Scale, DP-2-3-CHAIN: 2560x1440 60Hz @ 130% Scale)

Last edited by phpTANKER (2025-06-25 22:45:51)

seth · 2025-06-25 21:39:40

And my secondary display doesn't work in Wayland anymore, however my third display, which is connected via Displayport chaining to the second one works.
…
the issue seems to be the 2nd display attached to the daisy chain, no matter which one.

The 2 chained ones are, but my main one differs (DP-1: 1920x1080 144Hz @ 100% Scale, DP-2-3-CHAIN: 2560x1440 60Hz @ 130% Scale)

Please post your Xorg log, https://wiki.archlinux.org/title/Xorg#General and

for OUT in /sys/class/drm/card*; do echo $OUT; edid-decode $OUT/edid; echo "================="; done

You'll need https://archlinux.org/packages/extra/x86_64/v4l-utils/
Also test the behavior of a different WL compositor like weston, sway or hyprland and nb. https://wiki.archlinux.org/title/KDE#Un … lution_set

phpTANKER · Yesterday 01:07:31

seth wrote:

And my secondary display doesn't work in Wayland anymore, however my third display, which is connected via Displayport chaining to the second one works.
…
the issue seems to be the 2nd display attached to the daisy chain, no matter which one.
The 2 chained ones are, but my main one differs (DP-1: 1920x1080 144Hz @ 100% Scale, DP-2-3-CHAIN: 2560x1440 60Hz @ 130% Scale)
Please post your Xorg log, https://wiki.archlinux.org/title/Xorg#General and
for OUT in /sys/class/drm/card*; do echo $OUT; edid-decode $OUT/edid; echo "================="; done
You'll need https://archlinux.org/packages/extra/x86_64/v4l-utils/
Also test the behavior of a different WL compositor like weston, sway or hyprland and nb. https://wiki.archlinux.org/title/KDE#Un … lution_set

Maybe I'm missing something but I'm not sure I have any Xorg logs, both ~/.local/share/xorg and grep -e Log -e tty Xorg.0.log/Xorg.*.log return nothing.

I'll try the for loop once I have the chance to try the drivers again. It's a bit difficult since it's my daily driver desktop.

seth · Yesterday 05:15:44

grep -e Log -e tty Xorg.0.log/Xorg.*

is a (useless) example on how to filter random stuff out of that log - and assumes teh log to be in the $PWD

cat /var/log/Xorg.*.log ~/.local/share/xorg/Xorg.*.log

*read* the articles, don't blindly copypaste the blue boxes.

Arch Linux

#26 2025-06-12 13:08:34

Re: Possible Wayland bug introduced with latest nvidia-dkms-575.57.08

#27 2025-06-12 13:19:10

Re: Possible Wayland bug introduced with latest nvidia-dkms-575.57.08

#28 2025-06-12 13:34:23

Re: Possible Wayland bug introduced with latest nvidia-dkms-575.57.08

#29 2025-06-12 13:38:40

Re: Possible Wayland bug introduced with latest nvidia-dkms-575.57.08

#30 2025-06-12 13:43:51

Re: Possible Wayland bug introduced with latest nvidia-dkms-575.57.08

#31 2025-06-25 18:36:57

Re: Possible Wayland bug introduced with latest nvidia-dkms-575.57.08

#32 2025-06-25 21:39:40

Re: Possible Wayland bug introduced with latest nvidia-dkms-575.57.08

#33 Yesterday 01:07:31

Re: Possible Wayland bug introduced with latest nvidia-dkms-575.57.08

#34 Yesterday 05:15:44

Re: Possible Wayland bug introduced with latest nvidia-dkms-575.57.08

Board footer