Skip to content

arm64: dts: apple: t6050: describe the headphone jack codec on every board - #2277

Draft
RyanTheTide wants to merge 4 commits into
integration/aurorafrom
dtfill-t6050/jack
Draft

RyanTheTide wants to merge 4 commits into
integration/aurorafrom
dtfill-t6050/jack

Conversation

@RyanTheTide

@RyanTheTide RyanTheTide commented Nov 19, 2026 •

Copy link
Copy Markdown
Contributor

Problem

The 3.5 mm headphone jack's CS42L84 codec was described only on J714s, J775d, K114c and K116s, each with its own copy of the same node. J714c, J716c, J716s, J873s, K114s and K116c had no codec, and J714c, J716c, J716s and the K boards disabled i2c2. Every one of these models has a 3.5 mm jack, and each board's ADT has the codec at the same place.

Change

  • New t6050-jack.dtsi: enables i2c2 and adds jack_codec: codec@4b (cirrus,cs42l84), reset on AP pin 139 (active high), interrupt on AP pin 134 (level low). Values from each board's ADT; they are the same on all of them.
  • J714s, J775d: include it in place of the local copy.
  • t6050-k114-k116.dtsi: include it in place of the i2c2 disable, and expose pins 134 and 139 in pinctrl_ap. K114c and K116s drop their copies (they keep their own pin lists, which already expose both pins). K114s and K116c get the codec through the shared file.
  • J714c, J716c, J716s: include it in place of the i2c2 disable, and expose pins 134 and 139.
  • J873s: include it (its pinctrl_ap has no reserved ranges).
  • J775c is not touched: its full board description in arm64: dts: apple: support M5 Max Mac Studio (J775c) #2102 already carries the codec.

No sound card node is added. The jack codec probes on its own on the new boards; it reports a plug and plays only through a macaudio card whose jack link names &jack_codec, and that card also needs the AOP audio service (aop_audio), which is not enabled on these boards yet. No speaker amps are added.

Testing

  • Full kernel tree make dtbs CHECK_DTBS=y (dtschema) for every T6050 DTB on this branch: builds clean, no dtc warnings, and no schema message that integration/aurora does not already have.
  • DTB build for all 11 T6050 boards: all OK, 0 warnings before and after.
  • Byte-identical DTBs: J714s, K114c, K116s, J775c.
  • J775d: only gains an explicit status = "okay" on i2c2, which was already enabled by default.
  • Changed: J714c, J716c, J716s, J873s, K114s, K116c (codec node, i2c2 enabled, pinctrl_ap reserved ranges).
  • dt-validate against the processed binding schema, before and after, on the 7 changed DTBs: 0 new messages. cirrus,cs42l84 has a binding.

Still to run

Owners only; no headphone playback is run as part of this PR.

  • Every changed board: dmesg | grep -i cs42l84 shows Cirrus Logic CS42L84, Revision: .. and no probe error; ls /sys/bus/i2c/drivers/cs42l84/ lists the codec on i2c2 at 0x4b.
  • J714s, J775d, K114c, K116s (have a card; regression check, their DTBs are unchanged): evtest on the jack input device shows SW_HEADPHONE_INSERT 1 on plug and 0 on unplug. Same row for the other boards once they have a macaudio card and aop_audio enabled.
  • J714c (owner XY55bqg6iAw8ttP), J716c (erikwb), J716s (rolotrealanis98), K114s, K116c and J873s (no owner yet): probe row above.
Per-board values and DTB diffs

All boards, from each board's ADT: bus i2c2, address 0x4b, reset AP pin 139 active high, interrupt AP pin 134 level low.

pinctrl_ap gpio-reserved-ranges, before -> after:

Board Before After
J714c <0 80>, <81 25>, <108 113> <0 80>, <81 25>, <108 26>, <135 4>, <140 81>
J716c, J716s <0 80>, <81 140> <0 80>, <81 53>, <135 4>, <140 81>
K114s, K116c (shared dtsi) <0 221> <0 134>, <135 4>, <140 81>
J873s none none

DTB diff (J714c; J716c, J716s, K114s, K116c are the same apart from the reserved ranges, J873s has only the i2c2 part):

 			reg = <0x05 0x5018000 0x00 0x4000>;
-			status = "disabled";
+			status = "okay";
+
+			codec@4b {
+				#sound-dai-cells = <0x00>;
+				compatible = "cirrus,cs42l84";
+				interrupts-extended = <0x71 0x86 0x08>;
+				reg = <0x4b>;
+				reset-gpios = <0x71 0x8b 0x00>;
+				sound-name-prefix = "Jack";
+			};
...
-			gpio-reserved-ranges = <0x00 0x50 0x51 0x19 0x6c 0x71>;
+			gpio-reserved-ranges = <0x00 0x50 0x51 0x19 0x6c 0x1a 0x87 0x04 0x8c 0x51>;

J775d:

 			reg = <0x05 0x5018000 0x00 0x4000>;
+			status = "okay";

The original node is DjDeveloperr's J714s node, also carried by Jeff Queisser (J775d), Joshua Warren (K114c) and Jens Axboe (K116s).

Assisted-by: LLM (Claude:Opus 5.5)

…board

The CS42L84 headphone codec was described only on J714s, J775d, K114c and
K116s, each with its own copy of the same node. Every T6050 board's ADT
has the codec at the same place: i2c2, address 0x4b, reset on AP pin 139
(active high) and interrupt on AP pin 134 (level low), and every model has
a 3.5 mm jack.

Move the node into a new shared t6050-jack.dtsi, which also enables i2c2,
and include it from J714s, J775d and t6050-k114-k116.dtsi in place of the
copies. Add it to J714c, J716c, J716s and J873s, and to K114s and K116c
through t6050-k114-k116.dtsi. On the boards that limit pinctrl_ap's
exposed pins (J714c, J716c, J716s, K114s, K116c) open pins 134 and 139.
J714c, J716c and J716s no longer disable i2c2.

The J714s, K114c and K116s DTBs are byte-identical. The J775d DTB only
gains an explicit status = "okay" on i2c2, which was already enabled by
default. J775c is left alone: its full board description is in review
separately and already carries the codec.

These boards have no sound card node yet, so the codec probes on its own;
the jack reports a plug only through a macaudio card that links it.

The original node was written by DjDeveloperr for the J714s, and copied
by Jeff Queisser (J775d), Joshua Warren (K114c) and Jens Axboe (K116s).

Signed-off-by: Ryan Murray <ryan@aurorasilicon.org>
@RyanTheTide

Copy link
Copy Markdown
Contributor Author

Blockers: 1.

At the head the values are right: the reset (AP 139, active high) and interrupt (AP 134, level low) match the J714s/K114c/K116s node, and every new pin list exposes 134 and 139 while 112 stays reserved (J714c <108 26>, <135 4>, <140 81>; J716c/J716s <81 53>, <135 4>, <140 81>). No card names the codec on the new boards, so they gain a plug-detect device and no output path.

  1. The PR conflicts with integration/aurora in t6050-j775d.dts. ebaf49c moved the Mac Studio audio into t6050-studio-audio.dtsi, which carries its own copy of the jack codec (t6050-studio-audio.dtsi:71-80) for both J775c and J775d. On rebase, drop the J775d hunk and have t6050-studio-audio.dtsi include t6050-jack.dtsi in place of its copy. That also covers J775c, which this PR leaves out, and then the J775c/J775d DTBs should be identical apart from the explicit i2c2 status.
  2. Heads-up: arm64: dts: apple: t6050: share the K boards' N1 and USB-C ports, add K114s and K116c #2280 rewrites the same K-board &pinctrl_ap list (t6050-k114-k116.dtsi:144 here). The second to land needs <0 80>, <81 53>, <135 4>, <140 81> so that pins 80, 134 and 139 are all exposed.
  3. Both CI jobs failed on the runner's apt mirror; rerun after the rebase.

HW row: on the rebased head, each changed board (J714c, J716c, J716s, K114s, K116c, J873s) by its owner: Cirrus Logic CS42L84, Revision in dmesg, codec bound at i2c2-0x4b, no IRQ storm (/proc/interrupts count flat with nothing plugged). J714s or K116s regression: SW_HEADPHONE_INSERT 1/0 on plug/unplug. 0 AVC, 0 failed units, no new kernel errors.

Assisted-by: LLM (Claude:Opus 5.5)

t6050-j775d.dts: took integration's side (its audio moved to t6050-studio-audio.dtsi); the jack include follows in the next commit.

Signed-off-by: Ryan Murray <ryan@aurorasilicon.org>
The Mac Studio's audio moved into t6050-studio-audio.dtsi, which carries its
own copy of the CS42L84 node. Include t6050-jack.dtsi there instead, as the
other T6050 boards do. The J775d DTB only gains an explicit status = "okay"
on i2c2.

Signed-off-by: Ryan Murray <ryan@aurorasilicon.org>
…he jack pins

The K board N1/USB-C change opens pin 80 (PCIe port 0's PERST#) in the shared
K114/K116 pinctrl_ap list, and this branch opens 134 and 139 for the jack.
Use the union, <0 80>, <81 53>, <135 4>, <140 81>, in both so they land in
either order. Pin 80 reads cleanly; only pin 112 takes an SError.

Signed-off-by: Ryan Murray <ryan@aurorasilicon.org>
@RyanTheTide

Copy link
Copy Markdown
Contributor Author

Merged integration/aurora (f5f023b); conflicts in t6050-j775d.dts resolved (Studio audio now in t6050-studio-audio.dtsi, which includes t6050-jack.dtsi in a follow-up commit; K114/K116 pinctrl_ap list set to the union with #2280: <0 80>, <81 53>, <135 4>, <140 81>), all eleven T6050 board DTBs build with unchanged dtc warnings.

@RyanTheTide

Copy link
Copy Markdown
Contributor Author

The conflict is resolved and CI is green on 8da4cc7; still needed are probe rows on that head from each changed board (J714c, J716c, J716s, K114s, K116c, J873s, and J775c/J775d through the shared Studio audio file: CS42L84 bound at i2c2-0x4b, no interrupt storm), plus a J714s or K116s plug/unplug regression row, then it can leave draft.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant