Fix Dummy Output on IPU7-camera laptops caused by a WirePlumber v4l2 hang
Most of the record holds. Issue #12720 and its two comments (X1 Carbon Gen 12, Dell XPS 14 DA14260) support the symptom, the trigger and the user-copy workaround. The intel-ipu7-camera 1.0.6-2 package downloaded from pkgs.omarchy.org ships hide-ipu7-v4l2.conf with device.disabled = true. install/hardware/intel/ipu7-camera.sh on quattro installs it when OVTI08F4 is present. On this workstation (wireplumber 0.5.17-1) line 42 of create-device.lua is the bare return. I ran the record's sed on a scratch copy: it adds exactly one line, and the result is byte-identical to the 0.5.18 tag's create-device.lua. The WirePlumber locations doc confirms $XDG_DATA_HOME/wireplumber is searched first, omarchy restart audio exists (bin/omarchy-restart-audio), and the 0.5.18 NEWS entry (#1004, !894) matches the record. Arch extra has had 0.5.18-1 since 2026-09-30. stable-mirror.omarchy.org still serves wireplumber 0.5.17-1 today, 2026-10-05. The version range is wrong. The tagged sources show 0.5.0 through 0.5.15 run this step as a SimpleEventHook with no transition, so a bare return there is harmless. The AsyncEventHook, and with it the hang, first appears in 0.5.16. So only 0.5.16 and 0.5.17 are affected. The record's check ("0.5.17 or older" plus a return after the log line) would match 0.5.15 too, and on that version the sed would insert a call on an undefined transition. Cause and fix are corrected to 0.5.16 and 0.5.17, and the check now also looks for AsyncEventHook. Not exercised: no WirePlumber restart and no IPU7 hardware here.
The user copy shadows the packaged script for as long as it exists, including after WirePlumber upgrades. Left in place after 0.5.18 arrives, it runs a 0.5.17 script against a newer daemon, which can break camera or audio policy in ways that are hard to trace back. Delete it as soon as pacman -Q wireplumber reports 0.5.18 or newer.
Symptom
Fresh Omarchy 4.0.4 install on a laptop with an Intel IPU7 camera (reported on a ThinkPad X1 Carbon Gen 12 and a Dell XPS 14 DA14260) has no sound at all. wpctl status lists only Dummy Output, pactl list cards shows the SOF card with Active Profile: off, paplay ends with Stream error: Timeout, the bar shows audio muted and the volume keys do nothing. Restarting WirePlumber or running omarchy restart audio puts the card straight back to Off. The kernel side looks healthy: SOF driver bound, sof-firmware loaded, UCM present.
Cause
Two layers. The trigger is /usr/share/wireplumber/wireplumber.conf.d/hide-ipu7-v4l2.conf, shipped by the intel-ipu7-camera package from the omarchy repo, which Omarchy installs automatically when the OVTI08F4 ACPI device is present. It sets device.disabled = true on the IPU7's raw V4L2 nodes, which is a reasonable thing to do. The defect is in WirePlumber 0.5.16 and 0.5.17. Those releases run the V4L2 device-creation step in scripts/monitors/v4l2/create-device.lua as an AsyncEventHook, and its disabled branch logs a notice and returns without calling transition:advance (). The transition never completes, the event hook never reports done, and WirePlumber's event dispatcher stops dispatching anything for the rest of the session. All policy starves behind it: profile selection (so the card stays on ACP's off default and creates no nodes), default-node selection and stream linking. 0.5.15 and earlier run the same step as a SimpleEventHook with no transition, so the bare return there is harmless. Confirmed on this workstation (wireplumber 0.5.17-1): line 42 of /usr/share/wireplumber/scripts/monitors/v4l2/create-device.lua is a bare return. WirePlumber 0.5.18 fixes it (NEWS: "made the v4l2 monitor advance its transition when a device is disabled", #1004, !894), and the 0.5.18 tag's copy of the file carries the missing line. Any device.disabled = true rule on a V4L2 device triggers the same hang on 0.5.16 or 0.5.17, not only on IPU7 machines.
Fix
**1. Confirm this is the cause.**
pacman -Q wireplumber intel-ipu7-camera
ls /usr/share/wireplumber/wireplumber.conf.d/hide-ipu7-v4l2.conf
journalctl --user -b -u wireplumber | grep -i 'V4L2 device .* disabled'
grep -n 'EventHook' /usr/share/wireplumber/scripts/monitors/v4l2/create-device.lua
sed -n 38,44p /usr/share/wireplumber/scripts/monitors/v4l2/create-device.lua
It applies only when all three hold: wireplumber is 0.5.16 or 0.5.17, the file declares AsyncEventHook, and the log:notice (... " disabled") line is followed directly by return. On 0.5.15 or older the file declares SimpleEventHook, this defect does not exist, and the edit below would break the script. Stop there and look elsewhere.
**2. Shadow the broken script with a corrected user copy.** WirePlumber looks up scripts in $XDG_DATA_HOME/wireplumber/scripts before /usr/share/wireplumber/scripts, so nothing under /usr is touched and no package is rebuilt:
mkdir -p ~/.local/share/wireplumber/scripts/monitors/v4l2
cp /usr/share/wireplumber/scripts/monitors/v4l2/create-device.lua \
~/.local/share/wireplumber/scripts/monitors/v4l2/
sed -i '/ disabled")$/a\ transition:advance ()' \
~/.local/share/wireplumber/scripts/monitors/v4l2/create-device.lua
sed -n 38,45p ~/.local/share/wireplumber/scripts/monitors/v4l2/create-device.lua
The block must now read:
if cutils.parseBool (properties ["device.disabled"]) then
log:notice ("V4L2 device " .. properties["device.name"] .. " disabled")
transition:advance ()
return
end
**3. Restart audio.**
Omarchy 4:
omarchy restart audio
Plain Arch:
systemctl --user restart wireplumber.service pipewire.service pipewire-pulse.service
**4. Remove the override once the fixed WirePlumber is installed.** WirePlumber 0.5.18 has been in Arch extra since 2026-09-30, so on plain Arch a normal full upgrade fixes this and the override is not needed. On Omarchy 4 omarchy update pulls Arch packages from stable-mirror.omarchy.org, and as of 2026-10-05 that snapshot still served wireplumber 0.5.17-1, so omarchy update does not fix it yet. Check after each update:
pacman -Q wireplumber
# once it reports 0.5.18 or newer:
rm ~/.local/share/wireplumber/scripts/monitors/v4l2/create-device.lua
omarchy restart audio
Do not remove intel-ipu7-camera to get sound back. The drop-in is not the defect, and removing the package takes the camera with it.
Verify
wpctl status lists the real speaker and microphone nodes instead of only Dummy Output, pactl list cards | grep 'Active Profile' shows a HiFi profile rather than off, paplay /usr/share/sounds/alsa/Front_Center.wav plays (if alsa-utils ships that file on your system, otherwise any audio file), and the volume keys move the level and show the OSD.
Sources
- https://github.com/omacom/omarchy/issues/12720
- https://gitlab.freedesktop.org/pipewire/wireplumber/-/raw/0.5.18/NEWS.rst
- https://gitlab.freedesktop.org/pipewire/wireplumber/-/raw/0.5.18/src/scripts/monitors/v4l2/create-device.lua
- https://pipewire.pages.freedesktop.org/wireplumber/daemon/locations.html
- https://archlinux.org/packages/extra/x86_64/wireplumber/json/
- https://gitlab.freedesktop.org/pipewire/wireplumber/-/raw/0.5.15/src/scripts/monitors/v4l2/create-device.lua
- https://gitlab.freedesktop.org/pipewire/wireplumber/-/raw/0.5.16/src/scripts/monitors/v4l2/create-device.lua