Files
punktfunk/packaging
enricobuehler 6a506a8fa9
windows-drivers / probe-and-proto (pull_request) Successful in 30s
ci / bun-nix (pull_request) Successful in 56s
ci / docs-site (pull_request) Successful in 1m16s
ci / web (pull_request) Successful in 1m17s
ci / rust-arm64 (pull_request) Successful in 1m31s
windows-drivers / driver-build (pull_request) Successful in 1m47s
android / android (pull_request) Successful in 4m40s
ci / rust (pull_request) Successful in 9m54s
apple / swift (pull_request) Failing after 13m27s
apple / screenshots (pull_request) Skipped
fix(vdisplay/driver,pf-frame): no punktfunk process holds REALTIME GPU priority by default
The RX 9070 XT field A/B (2026-08-11/12 logs) convicted BOTH of our REALTIME
GPU-scheduling levers of generating the metronomic capture-stall class the
stall program has chased for weeks — compose-silence holes of 150-800 ms in
which ETW shows NO process presenting while the GPU stays responsive:

- the vdisplay driver's IddCxSetRealtimeGPUPriority raise beat at ~1.75-1.78 s
  (PFVD_NO_RT_GPU=1 alone removed that metronome: ~0.35 stalls/s metronomic ->
  10 sparse aperiodic over 3.9 min);
- the host auto-gate's HIGH->REALTIME upgrade (pf-frame dxgi.rs, T2.3) beat at
  ~3.58 s in the AV1 sessions where it promoted (vram_pct=1, 12:59:26); pinning
  PUNKTFUNK_GPU_PRIORITY_CLASS=high removed that residual too (13:45 session:
  zero metronomic, stall rate at the clean-run baseline).

Neither period matches any punktfunk clock: the full periodic-actor census
(driver: event-paced drain + 16 ms E_PENDING wait, 33 ms cursor poll, 3 s
watchdog reap; host: 250 ms descriptor poll, 5/50/100 ms probes + ~2 s scanline
retarget, 2 s VRAM gate, 2 s exclusive re-assert, 3.33 s pinger, 1 s stats,
~1 Hz phase-lock, fps/2 LTR marks) has nothing in the 1.69-2.29 s band, and
every host-side actor ran unchanged in the A/B that killed the fast metronome.
The periodicity is emergent from holding an unreachable-priority queue against
the WDDM scheduler on this AMD family (the period even differs by which of our
processes holds REALTIME); it is not a punktfunk cadence being amplified, so
there is nothing punktfunk-periodic to fix - the fix is to stop holding
REALTIME by default, which is also canonical parity (no shipping IDD raises
it, and HIGH was the class that delivered the original Sunshine-parity encode
win).

- Driver: PFVD_NO_RT_GPU (default-ON, opt-OUT) becomes the PFVD_RT_GPU ladder,
  default OFF on every vendor: unset = no raise (canonical IDD behavior);
  =thread = SetGPUThreadPriority(+7), a graduated in-band middle rung for field
  A/B (not default: unmeasured here, and the host measured the same call as "no
  help" for its own starvation case); anything else = the old REALTIME DDI.
  PFVD_NO_RT_GPU stays recognized and WINS over the opt-in, so the field boxes
  that carry it through the default-ON era keep meaning OFF. Both directions
  remain A/B-able without a rebuild (machine env + device restart). The CPU
  half of the original branch-2 hardening (MMCSS / TIME_CRITICAL) is untouched
  - it addressed the delivery holes that were actually observed.
- Host: PUNKTFUNK_GPU_PRIORITY_CLASS default auto -> high. `auto` (the gated
  REALTIME upgrade) stays available as an explicit opt-in, `realtime` still
  pins; unrecognized values now land on the HIGH default instead of silently
  opting into the gate - a typo must not buy the hazard. The VRAM/HAGS gate
  machinery is unchanged for `auto`; it guards the NVENC-hang hazard but cannot
  see this one.
- stall.rs: the no-OS-event METRONOMIC warning now carries rt_gpu_driver /
  rt_gpu_host fields (the machine-env state of both levers) and names clearing
  them as the FIRST cure, ahead of the display-hardware suspects - a field log
  self-answers the triage question this program just spent a week on.

No console policy axis for the driver knob: the lever is default-safe now, the
driver reads config at WUDFHost scope where machine env already matches the
device-restart lifecycle, and a policy axis would need pf-driver-proto churn
(or a device-key registry write) for an experimental lever that only exists to
be A/B-ed. If the `thread` rung ever proves out as a default-worthy raise,
that is the moment to revisit.
2026-08-12 13:57:20 +02:00
..

Packaging punktfunk for Fedora / Bazzite

The punktfunk host links system FFmpeg (NVENC on NVIDIA, VAAPI on AMD/Intel, with a GPU-less software-H.264 fallback), PipeWire and Opus. This page covers packaging it for the Fedora Atomic / Bazzite world (rpm-ostree + bootc), where most of those deps are already present; the NVIDIA-specific notes below apply to the NVENC path.

👉 Ubuntu/Debian hosts install via apt from Gitea's package registry — see debian/README.md (apt update && apt upgrade for new builds).

👉 End-to-end Bazzite setup walkthrough (install → udev/group → host.env → service → firewall → verify → troubleshooting): bazzite/README.md. This file is the higher-level packaging rationale.

packaging/
  rpm/punktfunk.spec      # the RPM (builds punktfunk-host from source with cargo)
  bazzite/host.env        # gamescope-default config for a Bazzite appliance
  bazzite/README.md       # step-by-step Bazzite setup guide
  bazzite/*sysext*.sh     # the no-layering path: build/install/publish the systemd-sysext
  bootc/Containerfile     # bake punktfunk into a Bazzite-based atomic image
  copr/                   # COPR build-from-SCM settings

The other packaging targets have their own READMEs: debian/ (apt), arch/ (pacman binary repo + PKGBUILD + SteamOS sysext), flatpak/ (the client), windows/ (host installer + drivers), plus kde/ and linux/ helpers. NixOS / Nix users get a flake (flake.nix at the repo root) with reproducible punktfunk-host, -client, -web, -scripting and -gamescope packages plus a services.punktfunk NixOS module — see nix/README.md.

What's needed beyond base Fedora

Dependency Where it comes from
ffmpeg-libs with NVENC RPM Fusion nonfree (ffmpeg, not ffmpeg-free)
NVIDIA driver (libnvidia-encode, libEGL_nvidia) Bazzite -nvidia images ship it; plain Fedora: akmod-nvidia + xorg-x11-drv-nvidia-cuda
gamescope, PipeWire, wireplumber Bazzite ships these; plain Fedora: dnf install gamescope pipewire wireplumber
opus, libei Fedora base / updates

On Bazzite the only genuinely new runtime bits are ffmpeg-libs (RPM Fusion) + opus + libei — the rest of the stack is already there. The default backend is gamescope (packaging/bazzite/host.env), which the host spawns headless per session — no desktop login.

On Bazzite / Fedora Atomic the recommended install is the systemd-sysext image — rpm-ostree layering is a last resort per the Bazzite docs (it slows every OS update and can block upgrades), while a sysext overlays /usr at runtime, survives OS updates, and updates in one command with no reboot. CI wraps the same RPMs below into the image, so content and channels are identical.

curl -fsSLO https://git.unom.io/unom/punktfunk/raw/branch/main/packaging/bazzite/punktfunk-sysext.sh
sudo bash punktfunk-sysext.sh install     # then: sudo punktfunk-sysext update | status | remove

Full walkthrough (incl. the F43→F44 rebase behavior and migration off layering): bazzite/README.md.

Option B — Gitea RPM registry (per-host, rpm-ostree layering)

The host's RPM is published to unom's self-hosted Gitea RPM registry (CI builds it on every push), mirroring the Debian/apt setup. Add one repo file, install, and track updates with rpm-ostree upgrade — no COPR account needed. Full guide: rpm/README.md.

# GPG-signed pkgs + Gitea-signed metadata → gpgcheck=1, repo_gpgcheck=1 (see rpm/README.md)
sudo tee /etc/yum.repos.d/punktfunk.repo >/dev/null <<'REPO'
[gitea-unom-bazzite]
name=punktfunk (unom, Bazzite)
baseurl=https://git.unom.io/api/packages/unom/rpm/bazzite
enabled=1
gpgcheck=1
repo_gpgcheck=1
gpgkey=https://git.unom.io/api/packages/unom/rpm/repository.key
       https://git.unom.io/api/packages/unom/generic/punktfunk-keys/1/RPM-GPG-KEY-punktfunk
REPO
rpm-ostree install punktfunk && systemctl reboot
# updates:  rpm-ostree upgrade && systemctl reboot

Option C — COPR (per-host, rpm-ostree install)

  1. Create a COPR project, enable build-from-SCM pointing at this repo, spec path packaging/rpm/punktfunk.spec (see copr/README.md). Under External Repositories add RPM Fusion nonfree so ffmpeg-devel resolves at build time.
  2. On the Bazzite host:
    # RPM Fusion (for the NVENC ffmpeg) — usually already enabled on Bazzite
    rpm-ostree install \
      https://mirrors.rpmfusion.org/free/fedora/rpmfusion-free-release-$(rpm -E %fedora).noarch.rpm \
      https://mirrors.rpmfusion.org/nonfree/fedora/rpmfusion-nonfree-release-$(rpm -E %fedora).noarch.rpm
    # enable the COPR + install punktfunk
    sudo wget -O /etc/yum.repos.d/_copr_punktfunk.repo \
      https://copr.fedorainfracloud.org/coprs/enricobuehler/punktfunk/repo/fedora-$(rpm -E %fedora)/
    rpm-ostree install punktfunk
    systemctl reboot
    

Option D — bootc (image-based, atomic)

Layer punktfunk into a Bazzite image once, then rebase any number of hosts onto it — no per-host drift. See bootc/Containerfile:

podman build -t ghcr.io/<you>/bazzite-punktfunk -f packaging/bootc/Containerfile .
podman push  ghcr.io/<you>/bazzite-punktfunk
# on the target:
sudo bootc switch ghcr.io/<you>/bazzite-punktfunk && systemctl reboot

First-run setup (all options)

ujust add-user-to-input-group           # virtual gamepads need /dev/uinput (then re-login).
                                        # On Bazzite use ujust, NOT `usermod -aG input` (atomic OS — it won't stick).
mkdir -p ~/.config/punktfunk
cp /usr/share/punktfunk/host.env.bazzite ~/.config/punktfunk/host.env   # edit (gamescope app, etc.)
systemctl --user enable --now punktfunk-host

# Management web console (pairing + status) — pulled in by default (the host RPM Recommends it;
# `--no-install-recommends` / headless-only boxes can skip it). Enable it and read the login password:
systemctl --user enable --now punktfunk-web
journalctl --user -u punktfunk-web-init | sed -n 's/.*password generated: //p'   # then open https://<host-ip>:47992

Pair a stock Moonlight client (mDNS-discovered), or connect the native punktfunk/1 client — via the web console at https://<host-ip>:47992 or directly.

⚠️ COPR caveat: COPR's mock chroot has no bun, so a COPR build produces only punktfunk + punktfunk-clientnot punktfunk-web. For the console on a COPR/bootc host, install from the Gitea RPM registry (Option B — it carries punktfunk-web; the sysext image includes it too), which is also why bootc/Containerfile installs from there rather than COPR.

Why not Flatpak (for the HOST)?

The host needs unsandboxed access the zero-copy NVENC path, /dev/uinput, the PipeWire graph and the compositor's privileged protocols — a Flatpak sandbox fights all of these. An RPM (or the bootc layer) installs into the host system where those just work.

👉 The client is a different story — it IS shipped as a Flatpak (the only viable Steam Deck install path: SteamOS /usr is read-only and lacks libadwaita/libSDL3). See flatpak/README.md. The client sandbox only needs the GPU render node, Wayland, PipeWire audio, the network and hidraw — all expressible as finish-args.

Building the SRPM/RPM locally (Fedora only)

git archive --format=tar.gz --prefix=punktfunk-0.3.0/ -o ~/rpmbuild/SOURCES/punktfunk-0.3.0.tar.gz HEAD
rpmbuild -ba packaging/rpm/punktfunk.spec     # needs the BuildRequires from the spec
# (0.3.0 = the spec's default %{pf_version}; the prefix and tarball name must match it)

(Not buildable on Debian/Ubuntu — use a Fedora toolbox/container or COPR.)