docs(audio): only Pro Audio is a saved pin, and the laptop card lost its HiFi profile

Two causes, one symptom. WirePlumber's own policy script never picks pro-audio
(it skips it by name), so both cards sitting on it can only come from the saved
default-profile file -- which is consulted before the automatic pick and is
therefore why it never self-corrects.

The laptop codec is a different story: ALSA UCM still lists its HiFi verb and
the kernel brought Jack Out / Speaker / HDMI up clean at boot, so the profile
list PipeWire built at enumeration is what is short. That needs a device
re-probe, not a setting.

Names the risk before the fix: pro-audio is the profile that opens the capture
PCM which hung the DSP in July, and the node guard written then is no longer on
disk.

docs/2026-09-22-audio-profile-stuck.md

+# Why pavucontrol shows only "Pro Audio", and why it never fixes itself
+
+**Status:** diagnosis, 2026-09-22. Read-only — nothing here has been changed.
+PLN: "why i see only pro audio? Can parvagues tooling ensure we have proper
+system and pavucontrol showing Speaker Headphones HDMI 1/2/3 etc as usual
+never stuck on pro audio without autofix?"
+
+## The state of the box
+
+Two ALSA cards, no USB interface attached:
+
+| card | device | profiles offered | active |
+|---|---|---|---|
+| `alsa_card.pci-0000_01_00.1` | NVIDIA HDMI (dGPU) | `off`, 6× HDMI stereo/surround, `pro-audio` | **pro-audio** |
+| `alsa_card.pci-0000_00_1f.3-platform-sof_sdw` | laptop codec (Speaker / Jack / HDMI) | `off`, `pro-audio` **only** | **off** |
+
+So: Speaker and Headphones are not "hidden", they are **not offered** — the SOF
+card's `HiFi` profile is absent from PipeWire's list entirely. And the NVIDIA
+card's HDMI profiles are all `available: no`, because no display with audio is
+plugged in.
+
+## Two separate causes
+
+### 1. A saved pin, which is exactly why it never autofixes
+
+`~/.local/state/wireplumber/default-profile`:
+
+```
+alsa_card.pci-0000_00_1f.3-platform-sof_sdw=off
+alsa_card.pci-0000_01_00.1=pro-audio
+```
+
+WirePlumber's `policy-device-profile.lua` consults that file **first**
+(`findDefaultProfile`) and only falls through to its own pick if the saved
+profile is unavailable. Both entries are available, so they win on every boot,
+forever. This is the "stuck" — it is a preference, not a fault.
+
+It is also proof the pin was set **by hand** (pavucontrol/pactl), because
+`findBestProfile` skips `pro-audio` by name:
+
+```lua
+elseif profile and profile.name ~= "pro-audio" then
+```
+
+WirePlumber will never choose Pro Audio on its own. Something chose it once.
+
+Remove the two lines and the automatic pick takes over: the NVIDIA card goes to
+`off` while nothing is plugged, and to `Digital Stereo (HDMI)` (priority 5900)
+the moment a display appears. That is the "as usual" behaviour.
+
+### 2. The laptop card lost its HiFi profile, and ALSA is not to blame
+
+The layers below PipeWire are healthy:
+
+```
+$ alsaucm -c sof-soundwire list _verbs
+  0: HiFi
+    Play HiFi quality Music
+
+$ aplay -l
+card 1: sofsoundwire, device 0: Jack Out      <- Headphones
+card 1: sofsoundwire, device 2: Speaker
+card 1: sofsoundwire, device 5/6/7: HDMI 1/2/3
+```
+
+and the kernel brought the card up clean at boot (20:12:54: Headset Jack,
+HDMI/DP pcm=5,6,7 registered, **no IPC error, no DSP timeout** — none of the
+2026-07-26 signature). So every sink he wants is present at the ALSA layer and
+PipeWire simply did not build the profile when it enumerated the card, ~11.5 s
+into the boot.
+
+**A profile list is built once, at device creation.** Getting `HiFi: Default`
+back therefore needs the device re-probed, which means restarting WirePlumber.
+Per #39's own learnings that restart resets sink volume to the device default
+(87%) and drops any `fold-orbits` links — a calm-evening move, before a
+rehearsal, never at a venue.
+
+## The risk that has to be named first
+
+`pro-audio` on the SOF card is **the dangerous profile**, not the safe one: it
+opens the raw PCMs including `pcm4` (Microphone), the exact PCM that hung the
+SOF DSP on 2026-07-26 and put WirePlumber in uninterruptible D-state, unkillable,
+taking scsynth down mid-session and needing a reboot (archive #39).
+
+The guard written then — `~/.config/wireplumber/wireplumber.conf.d/51-disable-sof-dmic.conf`,
+`node.disabled` on the capture node — **is gone**. That directory now holds only
+`50-alsa-config.conf`; its mtime is Sep 20 19:31. Nothing in this repo writes
+there (checked: `grep -rn wireplumber tools/` finds only documentation and the
+gig-up warning), so it was removed outside the tooling.
+
+Order of operations, therefore:
+
+1. rewrite the DMIC guard, matching the node **by name** (`HiFi__Mic__source`) —
+   the archive's "card3" is stale by its own "the ALSA card number is not
+   stable" learning;
+2. clear the two `default-profile` pins;
+3. restart WirePlumber once, on a calm evening, and check `HiFi` appears;
+4. only then set the SOF card to HiFi.
+
+## Gig relevance
+
+With the UMC202HD unplugged and the SOF card `off`, there is **no analog sink at
+all** on this box — `gig-up --headphones` has nothing to fold onto. That is fine
+for a venue (the UMC is the gig interface) and blocking for a train rehearsal.
+
+## What tooling can honestly promise
+
+Following `check-preload.sh`'s shape — check by default, mutate only on an
+explicit flag, and the gate never passes the flag:
+
+- **check**: a card pinned to a profile while a higher-priority one is
+  available; a card whose UCM verbs exist but whose PipeWire profile list does
+  not offer them. `ADVISE`, not `BLOCK` — a pinned card is a preference, and the
+  gate must not refuse to launch over one.
+- **--fix**: for the NVIDIA card, clear the pin and set the best available
+  profile. That is a complete fix.
+- **--fix cannot restore HiFi on the SOF card**, because that needs a device
+  re-probe. The tool must say so in words rather than pretend, and must never
+  restart WirePlumber on its own during a set.