Add configurations for nvidia microarchitectures

[BCNelson]

And add logic to enable the opensource drivers if the NVIDIA driver package and micro architecture supports it

Things done

• Tested the changes in your own NixOS Configuration
• Tested the changes end-to-end by using your fork of nixos-hardware and
  importing it via <nixos-hardware> or Flake input

[Mic92]

Making this a public api, would work against #992 a bit. Progress on that has been slow, but we made good progress on https://github.com/numtide/nixos-facter so I would prefer to not expose stuff in the flake.nix, which we later on want to deprecate again. The alternative would be to make this part of NixOS as well...

With that we could than automatically detect what nvidia generation is installed in a machine.

[eg-ayoub]

I have a small question about this,
considering the configs for other devices would still show an error until they set the architecture,
would it be possible to have a meaningful error message in this case ?
Even if one is integrated into nixpkgs, will it guarantee users will know how to fix the nixos-hardware part properly when making their PRs ?

[Mic92]

It's difficult the way this was implemented in nixpkgs to have such an error message.

[Lyndeno]

Can we add the Pascal dependency to the XPS 9560 while we are at it?

[Lyndeno]

I have a small question about this, considering the configs for other devices would still show an error until they set the architecture, would it be possible to have a meaningful error message in this case ? Even if one is integrated into nixpkgs, will it guarantee users will know how to fix the nixos-hardware part properly when making their PRs ?

I suggest we add the architecture to each model that uses nvidia before this is merged

[Mic92]

I have a small question about this, considering the configs for other devices would still show an error until they set the architecture, would it be possible to have a meaningful error message in this case ? Even if one is integrated into nixpkgs, will it guarantee users will know how to fix the nixos-hardware part properly when making their PRs ?

I suggest we add the architecture to each model that uses nvidia before this is merged

I think this might be a lot of work for a single person. This is better parallelized after merging.

[Lyndeno]

I think this might be a lot of work for a single person. This is better parallelized after merging.

True, I can take a look at affected modules once this is merged.

[Mic92]

here is a checklist: #1100