Skip to content
Merged
Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
22 commits
Select commit Hold shift + click to select a range
4214e25
docs: J81 audio bring-up notes
ricktav Nov 19, 2026
aca8ad5
j81: stage the A8X SmartIO firmware
ricktav Nov 19, 2026
562b6f6
kernel: t7001-air2: enable I2C1 for the speaker amplifiers
ricktav Nov 19, 2026
8281202
kernel: SmartIO on the A8 and A8X (AppleA7IOP version 1)
ricktav Nov 19, 2026
9823945
kernel: MCA version 1 (A8X) in the S8000 MCA driver
ricktav Nov 19, 2026
f8415e9
kernel: Maxim MAX98721 speaker amplifier driver
ricktav Nov 19, 2026
6dda863
docs: J81 audio: status after the first driver pieces
ricktav Nov 19, 2026
8ef38f4
kernel: the T7001 (A8X) NCO
ricktav Nov 20, 2026
78afb2f
kernel: t7001: the speaker audio path in the device tree
ricktav Nov 20, 2026
5f684fe
docs: J81 audio: the NCO and the device tree nodes
ricktav Nov 20, 2026
8b00b61
kernel: t7001-air2: enable the speaker link
ricktav Nov 20, 2026
7e46a98
j81 audio: fixes from the first hardware probe
ricktav Nov 20, 2026
df3bb84
kernel: t7001-air2: the speakers (amplifiers and sound card)
ricktav Nov 20, 2026
9129c9f
docs: J81 audio: the MCA3 boot freeze, bisected
ricktav Nov 20, 2026
9f91789
docs: J81 audio: first sound from both speaker amplifiers
ricktav Nov 20, 2026
7078111
docs: J81 audio: the speaker channel map
ricktav Nov 20, 2026
b40de77
kernel: J81 speakers: name the channels; fix the audio-graph-card2 mu…
ricktav Nov 20, 2026
f013d89
docs: J81 audio: status
ricktav Nov 20, 2026
c917bbb
system: UCM profile for the iPad Air 2 speakers
ricktav Nov 20, 2026
17aa5eb
j81: the speakers work
ricktav Nov 20, 2026
66e06b8
j81: the speakers work (from #192)
ricktav Nov 22, 2026
138ab2c
Merge main and Orchard's integration of #192
henrikklee Nov 22, 2026
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 1 addition & 1 deletion devices/j81/device.toml
Original file line number Diff line number Diff line change
Expand Up @@ -53,7 +53,7 @@ resume = "untested"
usb_serial = "working"
usb_network = "working"
usb_host = "missing"
audio_output = "missing"
audio_output = "working" # Both speakers (MAX98721 on MCA3); tested 2026-11-20. The headphone jack (CS42L81) is not supported yet.
audio_input = "missing"
camera = "missing"
video_decode_hevc = "missing"
19 changes: 19 additions & 0 deletions devices/j81/firmware.json
Original file line number Diff line number Diff line change
Expand Up @@ -133,6 +133,14 @@
"rootfs": "usr/sbin/BlueTool"
},
"device": "/usr/sbin/BlueTool"
},
"kernelcache": {
"ipsw": {
"kernelcache": true
},
"device": {
"kernelcache": "active"
}
}
},
"outputs": [
Expand Down Expand Up @@ -318,6 +326,17 @@
"input": "BlueTool"
}
]
},
{
"destination": "lib/firmware/apple/t7000-smartio.bin",
"sha256": "5fc96ee702d88b9a8f6a2526f566dd266c6c0d463b10490ba288db8e77434bd3",
"steps": [
{
"step": "smartio-from-kernelcache@2",
"input": "kernelcache",
"kext": "com.apple.driver.AppleT7000SmartIO"
}
]
}
]
}
137 changes: 137 additions & 0 deletions docs/j81-audio.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,137 @@
# J81 (iPad Air 2) audio bring-up

Working notes for speaker and headphone audio on the iPad Air 2 (J81, A8X /
T7001). Status: **both speakers play**, from a RAM boot and from an installed
Omarchy through PipeWire (`audio_output` is `working`). The headphone codec
and the microphones are not supported yet (`audio_input` stays `missing`).

Sources:
- Apple's device tree `Firmware/all_flash/DeviceTree.j81ap.im4p`
(sha256 `3021ec21…a48e`, as pinned in `devices/j81/boot.json`).
- The kernelcache `kernelcache.release.ipad5b` (19H422, sha256 `69120a77…9f59`;
matches the BuildManifest digest). Both are read from the pinned public
IPSW by byte range.

## Hardware (from the ADT)

| Block | ADT | Notes |
| --- | --- | --- |
| SmartIO | `/arm-io/sio`, `iop,s5l8960x`, reg `0x20ae00000` | DMA parent of every MCA and of spi1 |
| MCA0 | `0x20a0a0000`, SIO endpoints 0x28/0x29 | headphone/mic codec |
| MCA1 | `0x20a0a4000` | voice |
| MCA3 | `0x20a0ac000`, SIO endpoints 0x2e/0x2f | **speakers**; TDM, 8 slots (like J71b) |
| MCA4 | `0x20a0b0000` | Bluetooth |
| I2S switch | `0x20a003000`, `i2s-switch,t7001` | same class as S8000 |
| Speaker amps | 2× MAX98721 on i2c1, 0x31 and 0x34 | IRQ GPIO 8 / 9; `function-alive` GPIO 101; calibration in syscfg `SpCl` |
| Headset codec | CS42L81 on spi1 (`0x20a084000`) | CS GPIO 157, IRQ GPIO 116 |

The MCAs report `mca-version` 1. The J81 Linux DT already has the PMGR
power states (`ps_sio`, `ps_sio_p`, `ps_mca0`–`4`, `ps_spi1`, `ps_i2c1`) and a
disabled i2c1 node; SIO, the I2S switch, spi1, the MCAs, the codec, the amps
and the sound card are missing.

## What the A9/A10 audio path gives us, and what is new

The J81 is closest to the S8000/T8010 path (SmartIO DMA + MCA), not to the A5
(Cascadia) path.

**SmartIO firmware.** `com.apple.driver.AppleT7000SmartIO` carries it in
`__DATA,__firmware` (0x788e0 bytes, sha256 `5fc96ee7…4bd3`, a 32-bit ARM
image, `KF-SIO`). `orchard.apple.smartio.extract()` already extracts it; the
recipe needs the bundle in `smartio.BUNDLES` and a
`smartio-from-kernelcache@2` step in `devices/j81/firmware.json`. The message
protocol matches A9/A10 (ready message op 0x6d, protocol version 3).

**SmartIO registers (new work).** The J81's SIO is an A7 IOP version 1 (iOS:
AppleA7IOPV1), while S8000/T8010 use version 2. `apple-sio-core` needs a V1
register backend. The J81 ANS (also V1, already working through `asp`) shows
part of the layout (remap registers 0x08–0x1c, CPU control 0x28, mailbox at
+0x1000); the queue, interrupt-status and timebase registers are not known
yet.

**MCA v1 (new work, small).** iOS's `AppleAE2Audio` handles MCA versions 0, 1,
4 and 5 in one driver. Its v4 register offsets match `s8000-mca.c`; its v0/v1
offsets match the A5 Cascadia `s5l-audio.c` map (for example TX_CFG 0x18,
RX_CFG 0x2c, FRAME 0x0c; FIFO size register 0x04 reads 0x80). Plan: a v1
register map inside the `s8000-mca.c` framework.

**MAX98721 amps (new work).** iOS's `AppleMAX98721Amp` uses 8-bit registers
that match the upstream MAX98925 map. Init, as iOS does it:
- FORMAT `0x20` = 0x57 / 0x97 / 0xd7 (16/24/32-bit, TDM, one bit of delay).
- MAP `0x11`–`0x19` = `c0 00 00 f0 00 ab 89 00 00`; `0x22`–`0x26` = 0; `0x2b` = 2.
- Per amp from the ADT `speaker-config`: `0x2c`, `0x2e`, `0x2f`, `0x30`–`0x35`, `0x37`.
- Enable: `0x36` = `0xb0` | per-amp bits, then `0x38` = 0x80, wait 1.5 ms.

Base: upstream `max98925.c` with a new compatible, the MAX98721 revision ID
(unknown yet; upstream accepts 0x51/0x80), the ADT values and TDM slots.

**NCO (new work, small).** The ADT's `/arm-io/nco` (`nco,t7001`,
`nco,s5l8960x`) is the block the T8010 driver variant already handles (iOS:
`AppleS5L8960XNCO`, same control bits). The only difference is the channel
stride: iOS reads `pmgr-nco-page-size` and falls back to 0x1000, and the J81
ADT has no such property. Its window is 0x5000 bytes, so five channels 0x1000
apart; MCA0, 1, 3 and 4 use channels 0, 1, 3 and 4 (`function-mclk_frequency`).
The reference rate is in `/arm-io` `clock-frequencies`, which the IPSW ADT
leaves zero for iBoot to fill in; it has to be read from the live ADT (m1n1
exposes it as MTD `adt`). A running J81 reports 342857142 Hz (2.4 GHz / 7)
at index 10 (clock-id 0x10a minus 0x100), the rate iOS special-cases for
precision.

**I2S switch.** iOS binds `AppleAE2I2SSwitch3` to `i2s-switch,s5l8960x`, the
layout `s8000-mca.c` programs. MCA3's route is `aud3` to `pin3`.

**Headset codec.** CS42L81 has a driver on the A9/A10 path; the J81 MCA0
record starts like J71t's.

## Plan

1. **Read-only probe** (no audio driver): enable i2c1 with no clients, read the
amps' revision register `0xff` and dump `0x00`–`0x3a` at 0x31 and 0x34.
Read PMGR power states for MCA0/MCA3/SIO, and only for powered blocks the
MCA version/FIFO registers and SIO CPU control. No writes.
2. SmartIO firmware recipe for the J81.
3. SmartIO V1 register backend; SIO reaches "ready".
4. MCA v1 and the DT nodes; validate the TDM stream with the amp outputs off.
5. MAX98721 driver; low-gain `speaker-test`, one amp at a time.
6. Headphones and microphones through CS42L81.

## Status

| Piece | State |
| --- | --- |
| SmartIO firmware (`smartio-from-kernelcache@2`, `AppleT7000SmartIO`) | done |
| SmartIO V1 register backend (`apple,t7000-sio`) | works on a J81: firmware boots, setup completes, 4 DMA channels |
| MCA v1 (`apple,t7001-mca`, MCA3 speaker link) | plays on a J81: TDM 8×16 in a 256-bit frame, bit clock from NCO channel 3 (24.576 MHz at 48 kHz) |
| MAX98721 driver | both amplifiers play on a J81 (0x31 slot 0, 0x34 slot 1) |
| NCO (`apple,t7001-nco`, 0x1000 channel stride) | written, builds |
| DT: SIO, I2S switch, MCA3, NCO in `t7001.dtsi` (disabled); SmartIO enabled on the Air 2 | written |
| NCO reference rate | 342857142 Hz (live ADT `clock-frequencies[10]`); NCO, I2S switch and MCA3 enabled on the Air 2 |
| DT: amplifiers and sound card (`audio-graph-card2`, two codecs on one link) | card "J81 Speakers" registers and plays; parsing the `multi` node logs OF refcount warnings in audio-graph-card2, to be fixed |

## Speaker safety

iOS protects the speakers in software (amp current/voltage feedback plus the
`SpCl` calibration); Linux will have neither at first. Use the ADT's hardware
limiter and boost settings, start with low gain, test one amp at a time, and
check the stream with the speaker outputs disabled first: a wrong TDM slot or
format can produce full-scale noise (J71b's plain-I2S attempt gave loud
static).

## Test log

| Date | Step | Result |
| --- | --- | --- |
| 2026-11-19 | ADT and kernelcache analysis (offline) | hardware above identified; no device access |
| 2026-11-19 | SmartIO firmware recipe | `t7000-smartio.bin` extracted, sha256 matches |
| 2026-11-19 | SmartIO V1 backend, MCA v1, MAX98721 driver (offline) | build as modules for apple-4k; patch audit passes for apple-4k and apple-16k; not run on hardware |
| 2026-11-20 | Live ADT read from Omarchy (read-only) | NCO reference 342857142 Hz |
| 2026-11-20 | RAM probe: amplifiers (read-only) | both answer at 0x31/0x34, revision 0x80; registers hold iOS's MAP values, global enable (0x38) and channel enables (0x36) off |
| 2026-11-20 | RAM probe: SmartIO | firmware boots; first build failed setup with -EBUSY: the send FIFO stays disabled on V1 (TX status 0x00020000) until the AP enables it; fixed in the V1 backend |
| 2026-11-20 | RAM probe: NCO and MCA3 | both bind; MCA version register reads 0x00010002 (version in the upper half on v1), check fixed |
| 2026-11-20 | RAM boots with the speaker DT | 4 of 5 freeze within ~1-12 s of /init (no console output, no USB), before any audio driver loads (they are blacklisted). Bisected by enabling one node at a time, one boot each: main kernel, +I2C1+SIO, +NCO, +I2S switch all boot; +MCA3 freezes. On a running system MCA3 can be powered (PMGR 0x20e020148) and its registers read (0x00010002) without a hang, so the suspect is how the boot treats the enabled MCA3 node: with a consumer that never probes, genpd keeps the domain in its boot state (sync_state pending) |
| 2026-11-20 | NCO variant and DT nodes (offline) | NCO layout from iOS's `AppleS5L8960XNCO`; patch audit passes for apple-4k and apple-16k; not run on hardware |
| 2026-11-20 | RAM boots with the speaker DT, DFU entered from Recovery | 2 of 2 boot normally; PMGR at handoff: MCA0-4 `0x200` (off), SIO on. The earlier freezes all followed DFU entered straight from iPadOS, so they may be the same bootchain problem rather than MCA3; open until more boots confirm it |
| 2026-11-20 | First playback (SmartIO firmware loaded by hand, modules loaded in order) | 440 Hz at -20 dBFS (-32 dBFS after the driver's attenuation) heard from amp0 at gain 8; 660 Hz in channel 1 heard from amp1. During playback: hw_ptr advances at 48000 frames/s, MCA3 TX_CFG 0x109 and FRAME 0x0f107007 (running), no underflow; i2s-switch aud3/pin3 routed with output enables; both amps global enable 0x80, blocks 0xb0, no status flags. A -30 dBFS tone at gain 4 was inaudible |
| 2026-11-20 | Stereo test (portrait, Home button down) | channel 0 / amp0 (0x31, slot 0) is the bottom-left speaker, channel 1 / amp1 (0x34, slot 1) the bottom-right; both together play cleanly |
| 2026-11-20 | Installed Omarchy (platform layer from this branch, SmartIO firmware copied in, `boot-entries sync`) | card and drivers come up on boot, no refcount warnings; without a UCM profile PipeWire plays silence (both gain controls start at 0); with `ucm2/conf.d/J81_Speakers` WirePlumber offers "Built-in Audio Speakers", sets both gains to 22, and `pw-play` is heard |
| 2026-11-20 | Lesson | reading MCA3 offset 0x64 (not a v1 register) resets the SoC; read only the registers the driver names |
4 changes: 3 additions & 1 deletion kernel/configs/apple-16k/config
Original file line number Diff line number Diff line change
Expand Up @@ -4625,6 +4625,7 @@ CONFIG_SND_SOC_DMIC=m
CONFIG_SND_SOC_MAX98357A=m
# CONFIG_SND_SOC_MAX98504 is not set
# CONFIG_SND_SOC_MAX9867 is not set
CONFIG_SND_SOC_MAX98721=m
# CONFIG_SND_SOC_MAX98927 is not set
# CONFIG_SND_SOC_MAX98520 is not set
# CONFIG_SND_SOC_MAX98373_I2C is not set
Expand Down Expand Up @@ -4763,7 +4764,8 @@ CONFIG_SND_SOC_SIMPLE_AMPLIFIER=m
CONFIG_SND_SIMPLE_CARD_UTILS=m
CONFIG_SND_SIMPLE_CARD=m
# CONFIG_SND_AUDIO_GRAPH_CARD is not set
# CONFIG_SND_AUDIO_GRAPH_CARD2 is not set
CONFIG_SND_AUDIO_GRAPH_CARD2=m
# CONFIG_SND_AUDIO_GRAPH_CARD2_CUSTOM_SAMPLE is not set
# CONFIG_SND_TEST_COMPONENT is not set
# end of Generic drivers

Expand Down
4 changes: 3 additions & 1 deletion kernel/configs/apple-4k/config
Original file line number Diff line number Diff line change
Expand Up @@ -4625,6 +4625,7 @@ CONFIG_SND_SOC_CS42L81=m
CONFIG_SND_SOC_MAX98357A=m
# CONFIG_SND_SOC_MAX98504 is not set
# CONFIG_SND_SOC_MAX9867 is not set
CONFIG_SND_SOC_MAX98721=m
# CONFIG_SND_SOC_MAX98927 is not set
# CONFIG_SND_SOC_MAX98520 is not set
# CONFIG_SND_SOC_MAX98373_I2C is not set
Expand Down Expand Up @@ -4763,7 +4764,8 @@ CONFIG_SND_SOC_SIMPLE_AMPLIFIER=m
CONFIG_SND_SIMPLE_CARD_UTILS=m
CONFIG_SND_SIMPLE_CARD=m
# CONFIG_SND_AUDIO_GRAPH_CARD is not set
# CONFIG_SND_AUDIO_GRAPH_CARD2 is not set
CONFIG_SND_AUDIO_GRAPH_CARD2=m
# CONFIG_SND_AUDIO_GRAPH_CARD2_CUSTOM_SAMPLE is not set
# CONFIG_SND_TEST_COMPONENT is not set
# end of Generic drivers

Expand Down
Original file line number Diff line number Diff line change
@@ -0,0 +1,36 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: ricktav <82380971+ricktav@users.noreply.github.com>
Date: Thu, 19 Nov 2026 20:42:11 +0100
Subject: [PATCH] arm64: dts: apple: t7001-air2: Enable I2C1

I2C1 carries the iPad Air 2's two speaker amplifiers (MAX98721, at 0x31
and 0x34 in Apple's device tree). Enable the bus without clients so the
amplifiers can be identified from user space before a codec driver is
written.

Signed-off-by: ricktav <82380971+ricktav@users.noreply.github.com>
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
---
arch/arm64/boot/dts/apple/t7001-air2.dtsi | 9 +++++++++
1 file changed, 9 insertions(+)

diff --git a/arch/arm64/boot/dts/apple/t7001-air2.dtsi b/arch/arm64/boot/dts/apple/t7001-air2.dtsi
index d4b114f73..6f7db2f1c 100644
--- a/arch/arm64/boot/dts/apple/t7001-air2.dtsi
+++ b/arch/arm64/boot/dts/apple/t7001-air2.dtsi
@@ -174,6 +174,15 @@ rtc_offset: rtc-offset@4 {
};
};

+/*
+ * Speaker amplifiers (two MAX98721, at 0x31 and 0x34 in Apple's device
+ * tree). No driver binds them yet; the bus is enabled so they can be
+ * identified from user space.
+ */
+&i2c1 {
+ status = "okay";
+};
+
&serial5 {
pinctrl-0 = <&uart5_pins>;
pinctrl-names = "default";
Original file line number Diff line number Diff line change
@@ -0,0 +1,107 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: ricktav <82380971+ricktav@users.noreply.github.com>
Date: Fri, 20 Nov 2026 09:16:24 +0100
Subject: [PATCH] arm64: dts: apple: t7001: Add the speaker audio path

Describe the blocks the iPad Air 2's speakers are played through, all
disabled: the SmartIO DMA coprocessor (AppleA7IOP version 1, endpoints
from Apple's device tree), the i2s-switch, MCA3 (the speaker link, NCO
channel 3, MCLK word at 0x20a00200c) and the NCO with its reference
clock. The NCO reference rate is filled in by iBoot in Apple's device
tree, so the board sets it.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Signed-off-by: ricktav <82380971+ricktav@users.noreply.github.com>
---
arch/arm64/boot/dts/apple/t7001.dtsi | 61 ++++++++++++++++++++++++++++
1 file changed, 61 insertions(+)

diff --git a/arch/arm64/boot/dts/apple/t7001.dtsi b/arch/arm64/boot/dts/apple/t7001.dtsi
index 892ff78bd..8716534b6 100644
--- a/arch/arm64/boot/dts/apple/t7001.dtsi
+++ b/arch/arm64/boot/dts/apple/t7001.dtsi
@@ -30,6 +30,17 @@ clkref: clock-ref {
clock-output-names = "clkref";
};

+ /*
+ * The NCO's reference. Its rate is in the clock-frequencies table that
+ * iBoot fills in (the NCO's first clock-id); the IPSW ADT leaves it 0,
+ * so each board sets it.
+ */
+ nco_clkref: clock-ref-nco {
+ compatible = "fixed-clock";
+ #clock-cells = <0>;
+ clock-output-names = "nco_ref";
+ };
+
spi3_clk: clock-spi3 {
compatible = "fixed-clock";
#clock-cells = <0>;
@@ -165,6 +176,47 @@ ans_mbox: mbox@208041000 {
power-domains = <&ps_ans>;
};

+ /* ADT i2s-switch,t7001: AppleAE2I2SSwitch3, as on S8000 */
+ i2s_switch: syscon@20a003000 {
+ compatible = "apple,t7001-i2s-switch", "syscon";
+ reg = <0x2 0x0a003000 0x0 0x1000>;
+ status = "disabled";
+ };
+
+ mca3: mca@20a0ac000 {
+ compatible = "apple,t7001-mca";
+ reg = <0x2 0x0a0ac000 0x0 0x4000>,
+ <0x2 0x0a00200c 0x0 0x4>;
+ reg-names = "mca", "mclk";
+ interrupt-parent = <&aic>;
+ interrupts = <AIC_IRQ 172 IRQ_TYPE_LEVEL_HIGH>;
+ power-domains = <&ps_mca3>;
+ clocks = <&nco 3>;
+ clock-names = "bitclk";
+ dmas = <&sio 0x2e>;
+ dma-names = "tx";
+ #clock-cells = <0>;
+ clock-output-names = "mca3-mclk";
+ #sound-dai-cells = <0>;
+ status = "disabled";
+ };
+
+ sio: dma-controller@20ae00000 {
+ compatible = "apple,t7000-sio";
+ reg = <0x2 0x0ae00000 0x0 0x2000>;
+ reg-names = "a7iop";
+ interrupt-parent = <&aic>;
+ interrupts = <AIC_IRQ 149 IRQ_TYPE_LEVEL_HIGH>,
+ <AIC_IRQ 148 IRQ_TYPE_LEVEL_HIGH>,
+ <AIC_IRQ 151 IRQ_TYPE_LEVEL_HIGH>,
+ <AIC_IRQ 150 IRQ_TYPE_LEVEL_HIGH>;
+ power-domains = <&ps_sio>;
+ firmware-name = "apple/t7000-smartio.bin";
+ dma-channels = <64>;
+ #dma-cells = <1>;
+ status = "disabled";
+ };
+
serial0: serial@20a0c0000 {
compatible = "apple,s5l-uart";
reg = <0x2 0x0a0c0000 0x0 0x4000>;
@@ -341,6 +393,15 @@ pmgr: power-management@20e000000 {
reg = <0x2 0xe000000 0 0x24000>;
};

+ /* Five channels 0x1000 apart; MCA0-4 use 0, 1, 3 and 4 */
+ nco: clock-controller@20e011000 {
+ compatible = "apple,t7001-nco";
+ reg = <0x2 0x0e011000 0x0 0x5000>;
+ clocks = <&nco_clkref>;
+ #clock-cells = <1>;
+ status = "disabled";
+ };
+
wdt: watchdog@20e027000 {
compatible = "apple,t7000-wdt", "apple,wdt";
reg = <0x2 0x0e027000 0x0 0x1000>;
--
2.56.0

Original file line number Diff line number Diff line change
@@ -0,0 +1,33 @@
From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
From: ricktav <82380971+ricktav@users.noreply.github.com>
Date: Fri, 20 Nov 2026 09:16:25 +0100
Subject: [PATCH] arm64: dts: apple: t7001-air2: Enable SmartIO

Start the SmartIO coprocessor, the DMA engine of the speaker link. Its
firmware is staged from the kernelcache (apple/t7000-smartio.bin).

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Signed-off-by: ricktav <82380971+ricktav@users.noreply.github.com>
---
arch/arm64/boot/dts/apple/t7001-air2.dtsi | 5 +++++
1 file changed, 5 insertions(+)

diff --git a/arch/arm64/boot/dts/apple/t7001-air2.dtsi b/arch/arm64/boot/dts/apple/t7001-air2.dtsi
index 613dd1bea..1d33fef01 100644
--- a/arch/arm64/boot/dts/apple/t7001-air2.dtsi
+++ b/arch/arm64/boot/dts/apple/t7001-air2.dtsi
@@ -193,6 +193,11 @@ &i2c1 {
status = "okay";
};

+/* SmartIO: the DMA engine of every MCA and of spi1 */
+&sio {
+ status = "okay";
+};
+
&serial5 {
pinctrl-0 = <&uart5_pins>;
pinctrl-names = "default";
--
2.56.0

Loading
Loading