Difference in kernel modules from `nixos-generate-config` and `facter`

[pimvh]

Describe the bug

Hi! Thanks for this cool new way of defining hardware config. I am not sure, but I think I ran into a bug.

I am not using nixos-hardware on my machine (as my laptop not defined in there), and using facter I observed changes in my available kernel modules.

In short, the following kernel modules did not appear when using facter instead of the generated hardware-configuration.nix:

> "usb_storage"
> "sd_mod"

Hardware config

nix hardware scan config:

# Do not modify this file!  It was generated by ‘nixos-generate-config’
# and may be overwritten by future invocations.  Please make changes
# to /etc/nixos/configuration.nix instead.
{
  config,
  lib,
  pkgs,
  modulesPath,
  ...
}:

{
  imports = [
    (modulesPath + "/installer/scan/not-detected.nix")
  ];

  boot.initrd.availableKernelModules = [
    "xhci_pci"
    "thunderbolt"
    "nvme"
    "usb_storage"       # < -- missing in facter
    "sd_mod"            # < -- missing in facter
  ];
  boot.initrd.kernelModules = [ ];
  boot.kernelModules = [ "kvm-intel" ];
  boot.extraModulePackages = [ ];

  fileSystems."/" = {
    device = "/dev/disk/by-uuid/...";
    fsType = "ext4";
  };

  boot.initrd.luks.devices."luks-..".device = "/dev/disk/by-uuid/...";

  fileSystems."/boot" = {
    device = "/dev/disk/by-uuid/...";
    fsType = "vfat";
  };

  swapDevices = [ ];

  # Enables DHCP on each ethernet and wireless interface. In case of scripted networking
  # (the default) this is the recommended approach. When using systemd-networkd it's
  # still possible to use this option, but it's recommended to use it in conjunction
  # with explicit per-interface declarations with `networking.interfaces.<interface>.useDHCP`.
  networking.useDHCP = lib.mkDefault true;
  # networking.interfaces.wlp0s20f3.useDHCP = lib.mkDefault true;

  nixpkgs.hostPlatform = lib.mkDefault "x86_64-linux";
  hardware.cpu.intel.updateMicrocode = lib.mkDefault config.hardware.enableRedistributableFirmware;
}

generates the following kernel modules:

nixosConfigurations.<< host >>.config.boot.kernelModules
[
  "kvm-intel"
  "bridge"
  "macvlan"
  "tap"
  "tun"
  "tun"
  "loop"
  "atkbd"
  "ctr"
  "msr"
]

nixosConfigurations.nolte.config.boot.initrd.availableKernelModules
[
  "xhci_pci"
  "thunderbolt"
  "nvme"
  "usb_storage"
  "sd_mod"
  "ext2"
  "ext4"
  "dm_mod"
  "dm_crypt"
  "cryptd"
  "input_leds"
  "aes"
  "aes_generic"
  "blowfish"
  "twofish"
  "serpent"
  "cbc"
  "xts"
  "lrw"
  "sha1"
  "sha256"
  "sha512"
  "af_alg"
  "algif_skcipher"
  "ecb"
  "ahci"
  "sata_nv"
  "sata_via"
  "sata_sis"
  "sata_uli"
  "ata_piix"
  "pata_marvell"
  "nvme"
  "sd_mod"
  "sr_mod"
  "mmc_block"
  "uhci_hcd"
  "ehci_hcd"
  "ehci_pci"
  "ohci_hcd"
  "ohci_pci"
  "xhci_hcd"
  "xhci_pci"
  "usbhid"
  "hid_generic"
  "hid_lenovo"
  "hid_apple"
  "hid_roccat"
  "hid_logitech_hidpp"
  "hid_logitech_dj"
  "hid_microsoft"
  "hid_cherry"
  "hid_corsair"
  "pcips2"
  "atkbd"
  "i8042"
  "rtc_cmos"
]

Using facter

With the following relevant config:

...
modules = [
    inputs.nixos-facter-modules.nixosModules.facter
    # configure the facter report
    { config.facter.reportPath = "${self}/inventory/${systemAttrs.hostname}/facter.json"; }
]
...

I observed the following diff:

Version changes:
[C.]  nix-community/nixos-facter#1  udev-rules  <none> x2 -> <none>
Removed packages:
[R.]  nix-community/nixos-facter#1  extra-utils    <none>
[R.]  nix-community/nixos-facter#2  initrd-fsinfo  <none>
[R.]  nix-community/nixos-facter#3  keymap         <none>
[R.]  nix-community/nixos-facter#4  link-units     <none>
[R.]  nix-community/nixos-facter#5  stage          1-init.sh
Closure size: 2605 -> 2599 (14 paths added, 20 paths removed, delta -6, disk usage -15.5MiB).

nix-repl> nixosConfigurations.nolte.config.boot.kernelModules
[
  "kvm-intel"
  "bridge"
  "macvlan"
  "tap"
  "tun"
  "tun"
  "loop"
  "atkbd"
  "ctr"
  "msr"
]



nixosConfigurations.nolte.config.boot.initrd.availableKernelModules
[
  "thunderbolt"
  "xhci_pci"
  "nvme"
  "ext2"
  "ext4"
  "dm_mod"
  "dm_crypt"
  "cryptd"
  "input_leds"
  "aes"
  "aes_generic"
  "blowfish"
  "twofish"
  "serpent"
  "cbc"
  "xts"
  "lrw"
  "sha1"
  "sha256"
  "sha512"
  "af_alg"
  "algif_skcipher"
  "ecb"
  "ahci"
  "sata_nv"
  "sata_via"
  "sata_sis"
  "sata_uli"
  "ata_piix"
  "pata_marvell"
  "nvme"
  "sd_mod"
  "sr_mod"
  "mmc_block"
  "uhci_hcd"
  "ehci_hcd"
  "ehci_pci"
  "ohci_hcd"
  "ohci_pci"
  "xhci_hcd"
  "xhci_pci"
  "usbhid"
  "hid_generic"
  "hid_lenovo"
  "hid_apple"
  "hid_roccat"
  "hid_logitech_hidpp"
  "hid_logitech_dj"
  "hid_microsoft"
  "hid_cherry"
  "hid_corsair"
  "pcips2"
  "atkbd"
  "i8042"
  "rtc_cmos"
]

If I am correct, the facter hardware scan does not include the following 'extra' nixosConfigurations.<< host >>.config.boot.initrd.availableKernelModules, which appear when using hardware-configuration.nix:

> "usb_storage"
> "sd_mod"

Expected behavior

I expect the kernel modules to remain the same, as with hardware-configuration.nix. It might have to do with kernel modules which are generated on initial install via a USB stick, as when I re-generate the config these kernel modules seem to match the ones generated by facter, the missing kernel modules might indeed then be superfluous. I wonder whether this is expected behaviour. Regenerating it generates the following relevant config:

 boot.initrd.availableKernelModules = [ "xhci_pci" "thunderbolt" "nvme" ];

System information

HP HP ZBook Firefly 14 inch G10 Mobile Workstation PC Linux ... 6.6.56 nix-community/nixos-facter#1-NixOS SMP PREEMPT_DYNAMIC Thu Oct 10 10:50:06 UTC 2024 x86_64 GNU/Linux

And I am running NixOS unstable. I can share the full report if you want, but I am not sure whether that is relevant.

[brianmcgee]

Hi, thanks for the detailed feedback. I will have a look in the next few days.

[brianmcgee]

@pimvh, it's been longer than I expected, but I can finally look at this.

Please re-run with the latest and greatest, and let me know if the kernel modules are still missing.