feat(pad-audio): Linux hosts stream pad audio — the per-pad PipeWire sink (WP3) #182
No Assignees
Notifications
Due Date
No due date set.
Dependencies
No dependencies set.
Reference: unom/punktfunk#182
Reference in New Issue
Block a user
Closes the last hole in DualSense support on Linux hosts: the 0xD1 pad-audio plane was Windows-only —
host_cap()answeredfalseandspawn()was a stub off-Windows, so a tier-A client (Android USB DS5) against a Linux host negotiated the cap off and stayed on wire rumble.What ships
audio/linux/pad_sink.rs— oneAudio/Sinkstream node per DualSense-family pad, minted with the identity the matchers read: ALSA-stylenode.namecarrying the pad's pairing MAC,node.description "Wireless Controller",device.bus/vendor.id/product.id/form_factorproplist, per-pad serial. No default-sink claim,priority.session 50. Theprocess()callback is the capture (we own the sink).PUNKTFUNK_PAD_SINK_NAME/_DESCoverride the strings for field debugging ({pad}/{mac}expand).native/pad_audio.rs— the framer/gate/lane machinery compiles on Linux; the streamer thread is generic over the capturer (same open-with-backoff); Linuxhost_cap= client asked +PUNKTFUNK_PAD_AUDIO≠0 + a reachable PipeWire socket; Linuxspawnmints the sink lazily in-thread.spawngains anedgeflag for the Edge identity.pad-sink-testdevtest — mints one sink and captures from it with per-pair peak reporting (speaker pair vs coil pair), no client needed.PUNKTFUNK_PAD_AUDIO{,_SLOTS}lose the "(Windows)" marker; the roadmap non-goal narrows to Bluetooth client pads.The AUX0–3 channel shape (second commit)
A field report ("GE-Proton 11-5 fixed DualSense audio — you only need the controller audio set to Pro Audio") led into GE's
proton-ds5-hapticpatch set: GE ≥11-4 opens the pad's node through its bundled pipewire-alsa plugin aspipewire:NODE=<name>withaux_channels=1, and its pulse fallback forces aPA AUX0..3map withstream.dont-remix— "the hidden PipeWire parent for a DualSense output exposes AUX0 through AUX3". Pro Audio is what makes a real pad present that shape; our sink mints it directly, so no profile step exists for users. A positioned FL FR RL RR sink would put those writers through position channelmix instead of index passthrough.Also confirmed from the patch set: device matching is bus/vid/pid proplist or the Sony name substrings (we carry both), and on GE ≥11-4 the MMDevice ContainerId is synthesized from the wine-side HID USB parent — the pure-PipeWire-node
GUID_NULLrisk from the design doc no longer applies.Validation (on-glass, Bazzite 44 / .41)
cargo build --release --locked(nvenc+vulkan-encode),clippy --all-targets -D warnings, pad_audio+pad_sink tests 11/11,fmt— in a fedora:44 container on the box.The 0xD1 plane was Windows-host-only: host_cap() answered false and spawn() was a stub everywhere else, so an Android tier-A client against a Linux host negotiated the cap off and stayed on wire rumble. The whole downstream machinery (framer, silence gate, lanes, 0xD1 send) was already capture- agnostic — only the capturer was WASAPI. - audio/linux/pad_sink.rs: one Audio/Sink stream node per DualSense-family pad, minted with the identity the matchers read (ALSA-style node.name with the pad's pairing MAC, description "Wireless Controller", bus/vendor/ product/form-factor proplist, per-pad serial), 4-ch F32 48 kHz FL FR RL RR, no default-sink claim, priority.session 50. The process() callback IS the capture. PUNKTFUNK_PAD_SINK_NAME/_DESC override the strings for field debugging ({pad}/{mac} expand). - native/pad_audio.rs: the shared logic and lanes compile on Linux; pad_audio_thread is generic over the capturer (open-with-backoff kept); host_cap() Linux arm = client asked + PUNKTFUNK_PAD_AUDIO + a reachable PipeWire socket; spawn() Linux arm mints the sink lazily in the streamer thread. spawn() gains an edge flag (Edge identity; ignored on Windows). - devtest pad-sink-test: mint one sink and capture from it, no client — the WP3 on-glass gate. Verified on a Bazzite 44 host: identity served through pipewire-pulse, rear-pair (voice-coil) tone captured bit-exact over both the native and pulse legs. - docs: PUNKTFUNK_PAD_AUDIO{,_SLOTS} are no longer (Windows); the roadmap non-goal narrows to Bluetooth client pads. Gates (fedora:44 container, natively on the .41 box): cargo build --release --locked (nvenc+vulkan-encode), clippy --all-targets -D warnings, cargo test pad_audio+pad_sink 11/11, cargo fmt.