Framework NixOS (Linux) users self-help

[ahoneybun]

You may want to help this project so that others can add it to their configs so they have the fixes/changes for the best experience:

GitHub - NixOS/nixos-hardware: A collection of NixOS modules covering...

[TibialCuriosity]

I’d be happy to help with this project! I’ll have a look to see what issues are open. Is there any other way I can assist?

[kirelagin]

Yep, I’ve applied the proposed patch and, so far, looks like it fixed it for me.

[Master_Chase]

@Master_Chase

Does anyone else have WiFi hand off issues? My Framework will hang on to an AP that has so little signal I can’t even send/receive pings rather than switch to the one next to it as I move around my work. I have to toggle airplane mode or click the WiFI name in the drop down for it to switch (I’ve tried letting it wait over 5 minutes).

AMD 7840 with the RZ616 wifi card would switching to the ax210 card fix the issue?

@kirelagin

Yes! I was thinking just earlier today that I should post somewhere on the forum. (Although it is offtopic specifically here, as this has nothing to do with NixOS.)

I have two APs with 802.11k and 802.11v configured, and roaming works flawlessly with all other devices (phones, macbooks, my other Dell laptop), but the AMD framework stubbornly stays connected to the AP with an absolutely unusable signal level when I am already right next to the better AP (while other devices roam mid-way).

Got tired of the WiFi handoff issue and bought an Intel AX210 off ebay. After swapping the WiFi card my handoff issue is gone.

[kirelagin]

Oh that’s great to know, I should do the same then. I was worried that there might be a legit reason why they didn’t go with it in the first place, but probably there is none…

[blinry]

As of today, PPD 0.20 has landed in nixpkgs, making it easier to get great power savings on AMD!

I tried to throw together an overlay for Mario’s latest improvements, but the build system changed a bit, and it still seems to complain about the bash-completion dependency. Maybe someone can fix it up.

nixpkgs.overlays = [
  (final: prev: {
    power-profiles-daemon = prev.power-profiles-daemon.overrideAttrs (
      old: {
        nativeBuildInputs = old.nativeBuildInputs ++ (with pkgs; [
          cmake
          pylint
          argparse-manpage
          bash-completion
        ]);
        src = prev.fetchFromGitLab {
          domain = "gitlab.freedesktop.org";
          owner = "upower";
          repo = "power-profiles-daemon";
          rev = "f8bea7c205afc4b3101edc31b8e9b35780b0ceb6";
          sha256 = "sha256-+TAZNZOChG4FASdGpBL3mQsFZgUhLrmWFpbxHjTyWgE=";
        };
      }
    );
  }
  )
];

[kirelagin]

I did it like this, and I am not proud of it (of the Python part):

      nativeBuildInputs = (lib.filter (d: !(lib.hasPrefix "python" d.name)) old.nativeBuildInputs) ++ [
        (pkgs.python3.pythonOnBuildForHost.withPackages (ps: with ps; [
          pygobject3
          dbus-python
          python-dbusmock
          pylint
          argparse-manpage
          shtab
        ]))
        pkgs.cmake
      ];
      buildInputs = old.buildInputs ++ [
        pkgs.bash-completion
      ];

[tlvince]

Here’s another way to do it (by disabling those options and avoiding rebuilds):

services.power-profiles-daemon.package = pkgs.power-profiles-daemon.overrideAttrs {
  version = "f8bea7c205afc4b3101edc31b8e9b35780b0ceb6";
  src = pkgs.fetchFromGitLab {
    domain = "gitlab.freedesktop.org";
    owner = "upower";
    repo = "power-profiles-daemon";
    rev = "f8bea7c205afc4b3101edc31b8e9b35780b0ceb6";
    sha256 = "sha256-+TAZNZOChG4FASdGpBL3mQsFZgUhLrmWFpbxHjTyWgE=";
  };
  mesonFlags =
    pkgs.power-profiles-daemon.mesonFlags
    ++ [
      "-Dbashcomp=disabled"
      "-Dmanpage=disabled"
      "-Dpylint=disabled"
    ];
};

[paperd]

Worth noting that these PPD updates do not seem to work on the latest stable, 23.11.

[kjhoerr]

@paperd

Worth noting that these PPD updates do not seem to work on the latest stable, 23.11.

I’m running PPD 0.20 on 23.11 stable, so it may just have to do with tweaking the config.

This was the override that I was using for a bit to get things working:

        (final: prev: {
            power-profiles-daemon = prev.power-profiles-daemon.overrideAttrs(
              old: {
                version = "0.20";
                nativeBuildInputs = old.nativeBuildInputs ++ [ pkgs.argparse-manpage ];
                src = prev.fetchFromGitLab {
                  domain = "gitlab.freedesktop.org";
                  owner = "upower";
                  repo = "power-profiles-daemon";
                  rev = "0.20";
                  sha256 = "sha256-8wSRPR/1ELcsZ9K3LvSNlPcJvxRhb/LRjTIxKtdQlCA=";
                };
                postPatch = ''
                  patchShebangs --build \
                    src/powerprofilesctl \
                    tests/integration-test.py \
                    tests/unittest_inspector.py
                '';
              });
            }
        )

Although pkgs refers to nixpkgs-unstable, and actually if you have a nixpkgs-unstable input overriding is way simpler now that the PPD 0.20 update is merged. This is what I have now:

        # Use nixpkgs-unstable PPD with latest source/inputs
        (_: _: { power-profiles-daemon = pkgs.power-profiles-daemon; })

[Alex_Shpilkin]

Worth asking for a backport to 23.11 maybe rather than jumping through those hoops? I didn’t do it as I both run unstable and have very little understanding of how PPD works and what it might have compatibilities with; but NixOS releases are not Debian stable, and requests to backport version bumps are usually accepted if there are no issues with them.

[bayesian]

Does this mean that once we either backport to 23.11 or hit 24.05 that this would be all we need for @Mario_Limonciello’s fixes?

  # AMD has better battery life with PPD over TLP:
  # https://community.frame.work/t/responded-amd-7040-sleep-states/38101/13
  services.power-profiles-daemon.enable = lib.mkDefault true;

Already in NixOS/nixos-hardware.

[kirelagin]

#TIL that the preventWakeOnAC workaround from nixos-hardware is no longer needed with kernel >= 6.7 because of

[TRACKING] Framework AMD Ryzen 7040 Series lid wakeup behavior feedback

I’ve sent this patch series to the mailing list to add a workaround in the kernel for the EC behavior (at least with BIOS 03.03). I would expect there is no need for a udev rule with this in place. [PATCH 0/4] Add a workaround for Framework 13 spurious IRQ1 (kernel.org) If you prefer the old behavior (which allows “real” wake from keyboard to work) and have this patch, there is a module parameter you can put on your kernel command line amd_pmc.disable_workarounds=1.

[Strixpyrr]

I’m getting the “No devices available” error when enrolling fingerprints on my Intel 12th gen Framework, despite it working fine in Manjaro. I’ve tried clearing the old fingerprints, updating the firmware (even though the guide said it’s only needed for 13th gen and AMD), nothing is working.

My config contains:

services.fprintd.enable = true;
services.fprintd.tod.enable = true;
services.fprintd.tod.driver = pkgs.libfprint-2-tod1-goodix;
security.pam.services.login.fprintAuth = true;

Any ideas?

[Quinn_Dougherty]

I’d love two separate nixos threads for AMD and intel, things are getting pretty messy

[Francesco_Basile]

Does anyone else have a problem with the cpu frequency? It appears to be stuck at 400Hz no matter what.
intel_pstate is active and governor is ‘performance’.
I don’t think I had this problem yesterday, I only noticed after adding a TLP configuration but even reverting that doesn’t make a difference

[Alexander_Taylor-Smi]

Good evening all,

Hoping to get a few pointers on where to start with a few issues I’m having.

I’m a bit of a newer Linux user as I’ve used Fedora for about a year, but never had to tinker as it’s always worked / never gotten in the way of my needs, but this weekend I decided to learn something new with my flash new toy.

So here are my issues:

• When connecting the dock, either 1)The external screen wakes, and no peripherals work, or 2) the screen is not recognised but all my peripherals work
• The HDMI expansion card does nothing. Not so much as an audible device tone.

Some context, I have run my laptop in dual boot with Fedora 39 for about a fortnight and had neither of the above issues.

Things I have tried:

• nixos-rebuild switch --upgrade
• changing to unstable branch
• applying the nixos-hardware config
• trying a different Desktop Environment
• quick google searches before realising I’m in too deep and don’t have enough established knowledge of linux or Nix
• going back to Fedora as a sanity check

I appreciate this is probably not the best initial place to ask this, but I figured it was contextually, worth doing as we’re all on (somewhat) similar platforms. I’m happy to go and read and/or play, but I don’t know where to start.

Any hints, tips or ideas would be greatly appreciated.

Cheers,

[Matt_Hartley]

Fingerprint related stuff now has its own thread.

[John_Evans]

I thought I’d share my config here for my 12th-gen Intel Framework laptop

This is a flake-based config with home-manager as a module. It includes the recommended TLP config from Framework. I’m currently running a pre-release version of System76s COSMIC desktop environment (which is awesome!), but also have configuration for KDE Plasma 6. I also have Firefox and Neovim configured through home-manager.

GitHub - thejevans/nix-config: My configuration for the Nix package manager and...

[Matt_Hartley]

I am in the process of working with the NixOS community right now (literally in chat in another tab) for a community support program.

Soon, once we get everything dialed in, you will see NixOS recommendations in the nearish future.