[7robertodantas]

Written with an AI assistant (Claude) on @7robertodantas's behalf; every hardware step was run and checked on the owner's J575c.

The J575c (Mac Studio, M4 Max) speaker was off in the DT and in macaudio: the J575c/J773 variant never bound the speaker link, because the speaker had not been brought up and the J773s ADT's speaker stream lists no receive slots, so no V/I sense was expected.

What the hardware does. The J575c ADT's iop-audio-controller/audio-speaker also lists one playback slot and no receive slots. The firmware runs the stream as on the J616s anyway: six 32-bit playback slots and twelve 16-bit sense slots in a 250-bit frame at 12 MHz, on base-ns ADMAC TX0/RX0.

• A one-channel stream drains six times faster than 48 kHz (about 290,000 frames/s); a six-channel one runs at 48 kHz, playback and sense in lockstep.
• The amp's current and voltage come back on sense channels 0 and 1. With a 440 Hz tone at -40 dB on Speaker Volume (about 0.2 mW into the speaker), both read the tone (100 and 91 LSB), the voltage 14 degrees ahead of the current, |Z| 3.4 ohm at speakersafetyd's SN012776 scales (3.75 A, 14 V).

So the J575c speaker can run under speakersafetyd with the Mac Studio speaker model (j575.conf, as j375.conf/j475.conf).

The change.

1. ASoC: apple: macaudio: a speaker front-end that carries more slots than amps (speaker_fe_channels, the J713's six-slot stream) no longer merges its back-end's channel limits. ASoC merges a lone codec's own limits (two channels on an SN012776) but not several codecs', so with one amp the six-slot front-end came out empty and every open failed with -EINVAL.
2. ASoC: apple: macaudio: the J575c gets its own variant: the J616s stream format and frame (I2S, inverted frame clock, bclk ratio 250), the front-end on all six slots as on the J713, the amp on the first two (it downmixes them). The J773 keeps its speaker off.
3. dt-bindings: sound: apple,t6041-aop-audio: allow spkr, sense, cout, cin (speakers and jack, no microphones).
4. t6041-j575c: enable i2c1, the amp and the card; the AOP audio node gets the speaker stream and its sense. The amp's address, interrupt and shutdown line, guessed until now, are the ADT's. The amp drops its "Speaker" name prefix: like the J375's single amp its controls are the card's plain Speaker Volume, Amp Gain Volume, VSENSE Switch, ISENSE Switch, which macaudio's mono fixups, the Mac Studio UCM profile and speakersafetyd's model use (with the prefix the card's control fixup fails and it does not bind).

No AOP driver change. The boot loader passes the amp's ADT configuration (apple,speaker-config 00 14 00 02, apple,bop-config, apple,amp-dcblocker-config; aura kboot, 1.0.0b5 and later). Without a protection daemon macaudio holds the amp at its lowest step (Speaker Volume 1, -100 dB).

Userspace, in separate PRs:

• macaudio: the J575c headphone jack and speaker alsa-ucm-conf-asahi#15: the J575c speaker verb (six channels, no jack codec on the card) and the J575c jack profile.
• speakersafetyd: the J575c (Mac Studio M4 Max) speaker profile alarm-aurora#267: speakersafetyd tries apple/j575c.conf before apple/j575.conf, and ships j575c.conf (j575.conf with the twelve-channel sense).

Tested on a J575c: linux-aurora 7.3.0 built from integration/aurora 1943af8 with #1546 and #1565, these commits' macaudio as a module, and these DT commits on that tree (with #1499) as a one-shot aura Stage 2 trial (aura 1.0.0b8 + aura#160). The PR is on today's integration/aurora (ea10c4d), where macaudio.c is unchanged.

Test	Result
Card binds	pass: AppleJ575 (PCM 1 "Secondary", PCM 2 "Speaker Sense") next to AppleJack, no amp faults
No daemon	pass: Speaker Volume held at 1 (-100.5 dB)
Stock speakersafetyd 2.0.1 with its j575.conf	fails safe: panics at snd_pcm_hw_params_set_channels (two channels asked of the twelve-channel sense PCM), speaker stays at the floor
speakersafetyd with the twelve-channel profile	pass: 1 speaker, amp gain 21 dBV, max power 0.94 W, min gain -17.65 dB; takes the volume lock, models from the sense
V/I sense at -40 dB, 440 Hz	pass: as above
PipeWire (aurora-audio's graph-j575.json, "Mac Studio Speaker")	pass: the owner confirmed sound from the built-in speaker
Headphone jack card alongside	pass (with the UCM jack profile)
dt-validate (dtschema 2026.9) on the J575c DTB	0 new warnings against integration/aurora; the binding passes dt-doc-validate
checkpatch --strict	0 errors (warnings only for the Co-Authored-By: trailer's case)

Not tested: a cold boot with the speaker enabled (the trials were warm reboots), suspend and resume with the speaker playing, long loud playback (speakersafetyd's governor reducing gain), the J773 (unchanged here).

🤖 Generated with Claude Code

[RyanTheTide]

Checked the macaudio change and the J575c config. The speaker comes up at the floor and only rises under speakersafetyd, and the V/I sense test backs the model. One question for a J713 or J715 owner: the speaker front-end no longer merges the back-end's channel limits on those boards either, so a quick speaker open there would confirm nothing changed. Approved.