Multiple issues with mobile RTX 4060 and NVIDIA drivers 550.78

This is a continuation of Laptop's RTX 4060 is basically unusable (Xid 62 and System is not in ready state)

Issues with NVIDIA drivers 550.78:

  • Only a single display resolution is available, the native one.
  • At 60Hz display output is extremely jerky, at 240Hz it’s fine but idle power consumption is 1W higher (4.5W vs 5.5W). Under Windows the GPU idles at 3.9W.
  • After (software suspend) resume the Xorg server is kinda dead. You see a black background and the mouse pointer which you can move. The only workaround is to press Ctrl + Alt + Backspace to restart Xorg. Switching to Linux console is not possible either. They are all dead.
  • Under Windows the monitor is GSync compatible, under Linux it’s GSync uncertified or something like that and it doesn’t use/enable VRR.

In short it’s all quite depressing. This laptop model was released over nine months ago. It’s basically unusable under Linux.

I’m puzzled whether you didn’t read here at all or you read here just to buy the second worst laptop just to bug. The worst of all being intel based legions.

It’s based on AMD :-(

And sorry I just bought prolly the cheapest one.

Yes. That’s why you’re lucky, the intel ones are complete wrecks. It also doesn’t matter how much you paid, the legion series are classic “Works With Windows” systems, even if you paid $3k.

This was always the issue with nvidia-only notebooks, the nvidia driver provides scaled resolutions only through randr 1.1
xrandr --q1 shows them. Quite unusable. Don’t know if Wayland handles this better.

Did you check if it even outputs at 60Hz, eg using glxgears?

It’s like something like that, I suspect:
https://forums.developer.nvidia.com/t/gsync-screen-of-the-alienware-m18-r2-is-not-detected/283922/4?u=generix
Don’t know why it rejects the 60Hz as minimum refresh rate. Should work but doesn’t.

Please run nvidia-bug-report.sh as root and attach the resulting nvidia-bug-report.log.gz file to your post.

Thanks for letting me know. I’ve given the laptop back to its owner who will be using it with Windows exclusively, so I’m unable to investigate or debug further.

Why doesn’t NVIDIA (want to) fix this? Or it’s an intended behavior?

aplattner once explained it long ago though I didn’t fully got the point of it. It was: " Note that unlike other modesetting interfaces, RandR 1.2 provides control over the actual physical mode timings being sent to the monitor, so only supported timings are listed."
Which doesn’t keep the other drm drivers like i915, amdgpu to provide scaled, fake resolutions.

Why doesn’t the NVIDIA driver provide fake resolutions for Xorg on laptops, @aplattner @aritger @ekurzinger ?

Like @generix said, Intel and AMD don’t have this issue (and I can confirm it as I own both Intel and AMD based laptops).

It has been a long time since I’ve looked in this area, but I think the issue is described in the “Why doesn’t the NVIDIA X driver make more display resolutions and refresh rates available via RandR?” FAQ here:

http://us.download.nvidia.com/XFree86/Linux-x86_64/550.78/README/faq.html

But, yes, it is a fair point that other drivers provide fake scaled resolutions. For what it is worth, we do have similar fake scaling in the nvidia-settings display configuration page, and you should be able to configure whatever scaling you like on the xrandr commandline using the --scale family of xrandr options. But, yes, it isn’t great that different drivers behave differently in this regard.

For the other issues mentioned at the start of this thread:

  • I don’t know off hand about the idle power consumption or suspend/resume issues.

  • We’re aware of the notebook gsync issues, and those should be addressed in a future release.

Now that’s a lot of words for a dud. Rather sounds like nvida isn’t really interested in graphics anymore.

Only now putting out
“Run Time D3 (RTD3) with the open kernel modules will only be supported on Ampere and later (with the closed kernel modules, we are able to support RTD3 on Turing, too). I doubt we will be able to enable RTD3 on Turing with the open kernel modules.”
After users having to fiddle out this by themselves is infuriating, resembling intel arrogance.