Builder for nvidia-x11-550.78-6.10.drv failed with exit code 2

Appreciate the sentiment, hopefully amd/intel release a less humongous/power hungry model at some point… :stuck_out_tongue:

Modesetting is mandatory for wayland.

The powermanagement setting is required to support sleep.

Otherwise this looks correct. For the record, what is the exact set of problems (including whether nvidia-settings works) you get after a nixos-rebuild boot --use-remote-sudo && systemctl reboot with:

 # fuck nvidia
  services.xserver.videoDrivers = ["nvidia"];
  hardware.nvidia = {
    open = true;
    nvidiaSettings = true;
    modesetting.enable = true;
    powerManagement.enable = true;

    package = config.boot.kernelPackages.nvidiaPackages.mkDriver {
      version = "555.58.02";
      sha256_64bit = "sha256-xctt4TPRlOJ6r5S54h5W6PT6/3Zy2R4ASNFPu8TSHKM=";
      sha256_aarch64 = "sha256-wb20isMrRg8PeQBU96lWJzBMkjfySAUaqt4EgZnhyF8=";
      openSha256 = "sha256-8hyRiGB+m2hL3c9MDA/Pon+Xl6E788MZ50WrrAGUVuY=";
      settingsSha256 = "sha256-ZpuVZybW6CFN/gz9rx+UJvQ715FZnAOYfHn5jt5Z2C8=";
      persistencedSha256 = "sha256-a1D7ZZmcKFWfPjjH1REqPM5j/YLWKnbkP9qfRyIyxAw=";
    };
  };

I also force the GBM backend and the GLX library, but I think that’s working around old wlroots bugs:

    # Required to run the correct GBM backend for nvidia GPUs on wayland
    GBM_BACKEND = "nvidia-drm";
    # Apparently, without this nouveau may attempt to be used instead
    # (despite it being blacklisted)
    __GLX_VENDOR_LIBRARY_NAME = "nvidia";

And finally I set the "nvidia-drm.fbdev=1" kernel arg to work around this bug, but that should have been fixed in the driver/kernel by now.

I don’t seem to set services.xserver.videoDrivers at all in my config, as a side note, I don’t understand how that works… My config checks if it’s in that list to switch the sway flag, though, and that’s definitely working, so I guess it ends up in there by default somehow. Either way, that setting should not be harmful.