Make the audio stack reliable on headless boots - #79
Open
olaservo wants to merge 1 commit into
Open
Conversation
PipeWire and WirePlumber are user services and the image does not enable lingering for pollen, so user@1000 runs only while a login session exists. A headless boot has no audio server, and the stack is torn down on SSH disconnect. Daemon audio falls back to pulsesink whenever ~/.asoundrc does not resolve, and then blocks on a pulse socket that never appears. Bluetooth audio is PipeWire-only, so it cannot work on a headless boot at all. The image's own 40-force-monitor-bluez.conf never loads, because WirePlumber is not running to read it. Observed on a Reachy Mini Wireless: this starved the 50 Hz control loop and ended in an fd leak and a SIGABRT restart loop. Three changes in stage2/05-reachy-mini: * Enable lingering for pollen, so user@1000 starts at boot. * 41-force-monitor-alsa.conf is the ALSA-side equivalent of the pollen-robotics#55 bluez fix. monitor.alsa sits behind an optional reserve-device -> support.dbus chain that can drop the monitor at a session-less startup, leaving WirePlumber active with an empty device graph while aplay -l sees every card. Disable reserve-device and mark monitor.alsa required. In some cases this turns a silent failure into an exit: if the monitor cannot load, WirePlumber stops instead of running device-less, so systemd restarts it rather than leaving an empty graph. * wireplumber-alsa-watchdog.{timer,service}: 45 s after the user manager starts, restart the audio stack if the graph holds no ALSA device. The check greps for ALSA because v4l2 and Bluetooth objects can populate the graph while the cards are missing. Healthy boots are untouched. The delay is in the timer rather than an ExecStartPre sleep: a sleeping oneshot wanted by default.target holds user@1000 in activating for the whole delay on every boot. Tested on a Reachy Mini Wireless (v0.2.7, trixie, wireplumber 0.5.8-2). Headless power-cycles bring up a full graph and the watchdog no-ops. Disabling monitor.alsa reproduces the empty graph, and the watchdog detects and recovers it. Assisted-by: Claude:claude-fable-5
olaservo
force-pushed
the
fix/headless-boot-audio-stack
branch
from
August 15, 2026 01:18
f852cbb to
5ad4a1c
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
This addresses an issue I ran into while testing Bluetooth on a Reachy Mini Wireless:
pollen, souser@1000runs only while a login session exists.40-force-monitor-bluez.confnever loads, because WirePlumber is not running to read it.pulsesinkwhen~/.asoundrcdoes not resolve, and then blocks on a pulse socket that never appears. That starved the 50 Hz control loop and ended in an fd leak and a SIGABRT restart loop.Three changes in
stage2/05-reachy-mini:pollen.41-force-monitor-alsa.conf: the ALSA-side equivalent of the Add support for connecting bluetooth devices #55 bluez fix, for a related failure where WirePlumber stays active with an empty device graph whileaplay -lsees every card. In some cases this turns a silent failure into an exit.wireplumber-alsa-watchdog.{timer,service}: restarts the audio stack 45 s after startup if the graph holds no ALSA device. Healthy boots are untouched.Tested on a Reachy Mini Wireless (v0.2.7, trixie, wireplumber 0.5.8-2).