[m0d16l14n1]

Draft at 0434a61e4a8fa71b643858ea053c05e79ce3e4b5. Normal merge of preceding09e6fbbc and main673f6184, retaining both histories. This shares J813’s T8142 SoC description and adds the J815 board resources in15 DT/binding files: power/PMP, USB, panel/GPU, AVD/ISP, N1, AOP/I2C, ALS, SMC sensors and SEP/Touch ID. The shared PMP keeps its SRAM supplier and always-on property. J815 sound and AOP audio remain disabled in this parent.

The J813 N1 factory-address targets and ALS node carry @gig3m’s supplied patches. The loader aliases target the proper six-byte address placeholders; experimental Bluetooth options are not enabled. T8142’s GPU binding retains its actual ordered seven memory regions, including the final fw-data; main’s ISP ordering and other family binding updates are retained.

The main integration resolves the J815 EOF conflict by keeping its complete prior board description as a byte-exact prefix and appending main’s complete MCC block. J815 retains main’s MCC enablement. J813’s extraction into the shared file now inherits the exact MCC geometry disabled; main’s RTC/rtc-offset additions are also retained. A later J700-only main update changed none of the reviewed or tested inputs.

Schema validation found a duplicate apple,agx-g17s entry already present in main. Removed only its redundant trailing occurrence from one enum; the allowed set and other conditions are unchanged. Main’s pinctrl GPIO-hog acceptance remains.

Local validation:

• Six board/layer symbolic three-way property checks pass, retaining PMP, factory aliases and ALS.
• Eight exact-tree binding documents pass metaschema/reference checks. Five actual T8142 memory-ABI cases pass;560 finite non-target comparisons produce the same results as current main.
• Three actual SEP KUnit bodies pass in a host Rust extraction on the composed source. This is not a local kernel/DTB or arm64 QEMU build.
• Independent source review found no concrete regression. Earlier incomplete EOF/source-model candidates and the inherited schema failure are retained in local evidence; no invalid candidate was published.

Current0434a61 Host43378989747/job131732089962 passed the actual tests. DT43378989936/job131732091047 built137 DTBs on both base and head (make rc=0/0), with402/394 warning lines, one new/nine removed. All seven changed bindings, including the repaired AGX schema, and ADT’s19 tests/six baseline comparisons passed without regression. The sole new warning is the retained J815 disp0_piodma alias; no new MCC, RTC, N1, ALS or GPU warning appeared. The overall DT check is failed, not green. No unchanged-source retry or warning filter was applied.

Keep draft: Aura#165 needs affected T8142 loader boot qualification and compatible-loader adoption before coordinated canonical-alias migration and a fresh green DT result. Do not filter the warning or retry unchanged source.

J813 evidence remains attributed to the owner: 6828183980 reports five boots/daily use of the97f feature on2b05915f, firmware26.6.2, with N1/local display and companion changes. It closes the original shared-DT owner check at that composition scope; GPU/SEP were untested there, and Bluetooth LE pairing required opt-in flags. 6837248959 separately reports J81327.0-stub ALS readings with installer-supplied calibration. VD6286 calibration stays in the ALS service-property path, without a CT817 firmware-name.

The later Touch ID/profile work and explicit guard1 diagnostic correction are tracked separately in draft #1732 at98e509f0; its net follow-up remains two files. Ryan’s source review5883010555 is COMMENTED, not formal approval; utility model coverage is a separate aurora-xart PR.

Original J815 report predates these integrations:27.0 stub/Aura Stage2, Plasma panel, GLES3.2/Vulkan with #1413, AVD H.264/HEVC, camera, Wi-Fi/BT, ALS, SMC and Touch ID. #1413 and native Centauri#1530 are landed source prerequisites. No boot of this complete new tree, loader/default change, audio playback or physical rollout qualification was performed here.

Original bring-up: @m0d16l14n1, assisted by Claude Opus5.5. Owner follow-ups: @gig3m, assisted by Claude Fable5.2. Integration/schema review assisted by Codex.

[DjDeveloperr]

Codex (GPT-6.1 Sol), reviewing on @DjDeveloperr’s behalf.

Keep draft: the J815 results do not close the explicit J813 boot gate for extracting/reworking the shared DT. Brightness/camera/USB3/backlight/decoder are separate outcomes and a source merge cannot blanket-qualify them. CI approval is now unblocked so schema/build results can be evaluated; an owner-approved J813 regression row is still needed for the shared change. No new hardware or media upload is requested here.

[DjDeveloperr]

Codex (GPT-6.1 Sol), reviewing on @DjDeveloperr’s behalf.

The DT retry actually ran: baseline412 warning lines versus head442, with30 new J815 warnings. Both DT builds returned0, and ADT coverage passed its19 tests and six board comparisons. This was a real schema/profile failure, not the earlier apt setup problem.

I pushed three ancestry-preserving follow-ups at ec064e9:

• Narrow T8142 I2C/SPI/pinctrl/setup-mailbox/ALS descriptions, the eight-domain ISP profile versus T8140’s fourteen, and a dedicated two-mailbox/six-region GPU profile. Other profile constraints are retained.
• Thomas Glanzmann’s byte-exact macaudio/TAS schema prerequisites from #1423 (original commits9afc945d0cdb and bb9ad0e), with attribution. No driver adoption or partial merge-parent edge is implied, and #1423’s branch is untouched.
• An explicit native T8142 Centauri binding profile with the PCI1900 fallback on J815 only. The generic driver still selects by host/PCI ID and its board table (#1530); no firmware/platform/protocol values were guessed. The older T6050 PCI-only profile retains all three wire27 requirements.

Validation: all eight changed primary schemas pass strict dtschema2026.9 document/reference checks;45 targeted processed-profile controls pass, including reproduced baseline rejections and strict negative cases. The native N1 binding has15 passing positive/negative controls; the exact audio prerequisites have document/profile controls. The final diff changes11 schema files and J815’s endpoint-compatible tuple only: J813/common DTS, drivers, resources, aliases and tests are unchanged. This is not a full new CHECK_DTBS or physical boot result; fresh CI is required.

Keep draft/unmerged. The remaining alias warning is a real producer/consumer contract: current T8142 aura hardcodes disp0_piodma. Draft aura#165 supplies canonical disp0-piodma plus legacy fallback, with production/libfdt regression controls; a J815 alias rename must wait for compatible-loader adoption and qualification. The explicit J813 shared-DT boot gate remains open, and the original J815 report is not an exact-new-head boot. No hardware, playback, loader/default or image action was performed.

[DjDeveloperr]

Codex (GPT-6.1 Sol), reviewing on @DjDeveloperr’s behalf.

Resolved the current-main conflicts in b180163 with a normal integration merge. The resolution preserves the firmware-27 DCP provider clocks, shared CPU/OPP/actuator data, T6050 mailbox/ALS/LAS profiles, and main’s ISP heap/text/data memory ABI and 14–16 power-domain profile, alongside T8142’s distinct eight-domain profile. Review caught and corrected an initial candidate that dropped the optional ISP text/data regions before publication.

All 11 schema documents, 65 targeted ABI controls, 15 full Centauri controls and 10 audio controls pass. The symbolic J813 source comparison preserves 307 nodes and 1,377 explicit properties, with redundant interrupt-parent entries inherited identically. This does not establish compiled-DTB or physical boot equivalence.

Keep draft: fresh integrated-head host/DT CI and the author-required J813 shared-extraction boot remain gates. The existing display alias remains intact; aura#165 must be qualified/adopted before migration. The earlier EC DT log is unavailable (server errors), so I am not claiming a verified remaining-warning count.

[DjDeveloperr]

Codex (GPT-6.1 Sol), reviewing on @DjDeveloperr’s behalf.

Corrected the T8142 GPU binding in 9122abe. My earlier strict profile omitted the existing seventh fw-data crash-log region. The actual current J813/J815 tuple reproduces the old failure; the corrected ordered tuple passes 32 controls, with 768 comparisons confirming unchanged acceptance across 16 non-T8142 compatible profiles and strict meta/reference validation passing. Other profiles retain max six regions. No DT/GPU driver/firmware resource value changed.

Concrete compiled evidence comes from #1534’s ea5ab50 run: tools installed, base/head make succeeded with 137 DTBs, base400/head395 warning lines, five new warnings (four GPU cardinality, one J815 underscore alias), ten removed; binding docs and ADT coverage passed. Fresh CI must confirm the new schema result.

The legacy alias remains intentionally unchanged pending aura165 hardware qualification and compatible-loader adoption. Keep draft and the J813 boot gate; no warning filter, premature alias rename or new hardware action.

[DjDeveloperr]

Codex (GPT-6.1 Sol), reviewing on @DjDeveloperr’s behalf.

Current-head CI has now executed: Host run 43107102188 passed its actual tests. DT run 43107102060 built all 137 DTBs successfully on both base and head; warnings fell from 400 to 391 (10 removed, one new). The corrected seven-region T8142 GPU schema passed binding validation, and all four earlier GPU warnings are gone.

The sole new warning is t8142-j815.dtb aliases/disp0_piodma, which violates the standard alias-name regex. Keep the loader compatibility gate intact: a canonical alias migration must wait for the affected-hardware boot/adoption qualification of aura#165; do not rename the alias or filter this warning merely to turn CI green. ADT coverage passed. This remains draft; the existing physical/loader qualification requirements are unchanged.

[DjDeveloperr]

Resolved the current-main conflict as normal merge 97f1472, parents 9122abe/main68830200; no history rewrite. The ISP binding uses main's full version, preserving the Neo sensor-domain order and T8142 eight-domain case, so that already-landed binding disappears from the residual. All added/removed feature lines in the remaining 15 DT/binding files match the preceding head.

The complete bindings directory is identical to the validated #1534 c83c100 tree: eight document metaschema/reference checks and 32 candidate/32 main GPU controls pass on those identical inputs. No repeated schema test, new DTB compile or hardware boot is claimed. The parent still disables J815 sound/AOP audio, and the underscore loader alias is retained without warning filtering. #1534's separate audio DT feature remains unchanged; no duplicate source/audio build or child branch update is needed for this source-preserving parent integration.

Keep draft for fresh Host/DT CI, Aura #165 affected-loader boot and compatible-loader adoption before alias migration, and the author-required J813 shared-DT boot. The original J815 hardware report retains its pre-integration scope; no new machine or hardware action is requested.

[gig3m]

A J813 owner's check for the first commit, which the description asks for.

A J813 (MacBook Air 13-inch M5, Mac17,3, 26.6.2 firmware) has booted this PR at 97f1472, carried on 2b05915, since 2026-11-11. The Stage 2 device tree is built from the same tree (aura 1.0.0b10). Five boots so far and daily use: internal display, NVMe, keyboard and trackpad, USB, speakers and microphones, Wi-Fi (with #1530), RTC (with #1538) and s2idle (with #1527) all work. I noticed no regression against the earlier build of the same base without this PR.

Limits: this is the 26.6.2 firmware, so the GPU is not in use here, and the tree also carries the J813 block below and a local display format patch.

One gap for J813: it needs the same N1 nodes this PR gives t8142-j815.dts. Without them the loader has nowhere to put the factory addresses: Bluetooth: hci0: No factory local-bd-address from the loader, bluetoothd ignores the controller, and Wi-Fi comes up on a non-factory MAC. With the block below added to t8142-j813.dts (same PCI IDs as J815: 106b:1900, 106b:1902, 106b:1903), hci0 logs Native startup accepted; factory address in use and Wi-Fi uses its factory MAC. With btcentauri loaded with enable_acl=1 arm_auxiliary_rings=1, a Bluetooth LE mouse then pairs and works; with the defaults it does not pair. The block could equally go in the shared t8142-j813-j815.dtsi. I left out the sep alias; SEP is untested here.

diff --git c/arch/arm64/boot/dts/apple/t8142-j813.dts w/arch/arm64/boot/dts/apple/t8142-j813.dts
index 020cd3d6fe28..5b0c6299ae5b 100644
--- c/arch/arm64/boot/dts/apple/t8142-j813.dts
+++ w/arch/arm64/boot/dts/apple/t8142-j813.dts
@@ -121,6 +121,49 @@ &dockchannel {
 #include "t8142-j813-n1.dtsi"
 #include "t8142-j813-audio.dtsi"
 
+/*
+ * N1 functions behind apcie port 0 (0000:00:00.0 -> 01:00.{0,1,2}), so the
+ * loader has somewhere to put the factory addresses: it copies ADT /chosen
+ * mac-address-wifi0 and mac-address-bluetooth0 into the wifi0/bluetooth0
+ * alias targets. Same nodes as t8142-j815.dts (same N1, same PCI IDs).
+ */
+/ {
+	aliases {
+		wifi0 = &n1_alpha;
+		bluetooth0 = &n1_beta;
+	};
+};
+
+&pcie0 {
+	n1_port: pci@0,0 {
+		device_type = "pci";
+		reg = <0x0 0x0 0x0 0x0 0x0>;
+		bus-range = <1 1>;
+		#address-cells = <3>;
+		#size-cells = <2>;
+		ranges;
+
+		n1_control: control@0,0 {
+			compatible = "apple,t8142-centauri-control", "pci106b,1900";
+			reg = <0x10000 0x0 0x0 0x0 0x0>;
+		};
+
+		n1_alpha: wifi@0,1 {
+			compatible = "pci106b,1902";
+			reg = <0x10100 0x0 0x0 0x0 0x0>;
+			/* Filled by the loader */
+			local-mac-address = [00 00 00 00 00 00];
+		};
+
+		n1_beta: bluetooth@0,2 {
+			compatible = "pci106b,1903";
+			reg = <0x10200 0x0 0x0 0x0 0x0>;
+			/* Filled by the loader */
+			local-bd-address = [00 00 00 00 00 00];
+		};
+	};
+};
+
 &mtp_hid {
 	multi-touch {
 		firmware-name = "apple/tpmtfw-j813.bin";

Disclosure: the testing, the patch and the first draft of this text were done with an LLM coding assistant (Claude Code). I ran the hardware side and have read what is posted here.

[DjDeveloperr]

Accepted @gig3m's6828183980 J813 owner check at its stated scope: five boots/daily use of the97f14725 feature on2b05915f, firmware26.6.2, same-tree Aura1.0.0b10 DT, supplied N1 block/local display-format patch and the named Wi-Fi/RTC/s2idle companions. This closes the first-commit shared-extraction owner-check item. It does not establish isolated newly published-tree, GPU or SEP qualification.

Carried the supplied N1 block exactly in223d7d7, with attribution in the commit, and atomically carried the updated parent into #1534 atdf3087d0. The native control tuple/PCI functions are exact J815 topology. Bounded symbolic parsing preserves all307 original J813 nodes/properties and adds only four PCI nodes and two alias properties resolving to Wi-Fi/Bluetooth address placeholders. J815 source, SEP alias and Bluetooth experimental options are untouched. The owner's BLE pairing result required enable_acl=1/arm_auxiliary_rings=1; default pairing remains unqualified. No new root build, DTB compile, hardware, playback or loader action occurred.

Fresh CI is required on the two published layers. Keep draft: Aura165 affected-loader qualification/adoption is still missing, so disp0_piodma stays unchanged and its warning is not filtered. The J813 result on legacy Aura1.0.0b10 does not qualify that companion alias migration.

[DjDeveloperr]

Integrated current main after #1527 merged, as normal mergee8fb89d9. Its PMP always-on addition conflicted with this stack's PMGR restructuring. The resolution retains the parent layout and copies main's complete ps_pmp properties/comment into it at the proper indentation; the SRAM supplier and always-on flag are preserved. No other text on that conflict file changes from the preceding parent. All14 other feature added/removed line sets remain exact, current cpuidle source is exact main, and the J813 N1 DTS is unchanged.

Symbolic checks for J813/J815 in both stack layers confirm the one PMP node at0x598, its SRAM supplier, always-on and factory aliases. #1534 inherits the same parent as260d5e15 and retains its exact J815 audio net patch. Both refs were published atomically without rewriting history. These are source checks only, with fresh hosted CI required.

The accepted J813 owner-check report already included #1527 in its composition; that gate stays closed at its stated scope. Keep draft for Aura165 affected-loader qualification/adoption and coordinated alias migration; no alias rename/filter or new hardware/loader/playback action occurred.

[gig3m]

Touch ID works on J813 with the J815 SEP block from this PR, plus one line in the SEP profile.

J813 (Mac17,3, 27.0 stub), this PR at 223d7d7 on 9219565, aura 1.0.0b12, aurora-sep 1-25, xart_guard=1:

• read-only attach first (xart_writes=0 provision_keybag=0): 12 endpoints, key store open
• writes on: keybag provisioned, one finger enrolled, matches at the lock screen
• macOS booted next: its Touch ID still works
• back in Linux: Touch ID: restored 1 enrolled identity/identities; enrolment survived reboot, matches again
• 0 guard refusals; the guard list stayed empty

The J813 ADT (27.0) carries the values the T8142 profile comment gives for J815: iop-sep,ascwrap-v6 with sepfw-booted, dart-sep vm-base-0 0x10000004000 / vm-size-0 0xffff0000, Mesa on spi2 at 0x3a5108000, 8 MHz, CS 0, power on NUB GPIO 36, data-ready on AP GPIO 200.

Two changes:

1. t8142-j813.dts: the J815 SEP and sensor block as it is in this PR, with apple/mesacal-j813.bin and apple/msrk-j813.bin. The compiled sep, mbox, spi2, fingerprint@0 and sep_dart nodes equal J815's, so it could go in t8142-j813-j815.dtsi instead.
2. profile.rs: accept apple,j813 for the T8142 profile. Without it the module logs no platform profile for this SoC; the SEP stays disabled.

diff --git a/arch/arm64/boot/dts/apple/t8142-j813.dts b/arch/arm64/boot/dts/apple/t8142-j813.dts
index 5b0c6299ae5b..5889b852793a 100644
--- a/arch/arm64/boot/dts/apple/t8142-j813.dts
+++ b/arch/arm64/boot/dts/apple/t8142-j813.dts
@@ -131,6 +131,12 @@ / {
 	aliases {
 		wifi0 = &n1_alpha;
 		bluetooth0 = &n1_beta;
+		/*
+		 * The loader finds the SEP through this alias: it reserves SEPFW
+		 * and copies the boot's local-policy and iBoot manifests onto
+		 * the node.
+		 */
+		sep = &sep;
 	};
 };
 
@@ -171,3 +177,94 @@ multi-touch {
 };
 
 /* Unexposed chassis port instances remain described and disabled. */
+
+/*
+ * Secure Enclave and the Touch ID sensor. All values are from the J813 ADT
+ * (macOS 27.0) and are the same as J815's:
+ *  - sep: iop-sep,ascwrap-v6 at 0x30a600000 (0x88000), sepfw-booted, so
+ *    apple_sep warm-attaches; ASC mailbox at +0x8000, AIC 445-448.
+ *  - dart-sep: t8110 at 0x308b80000, irq 455; vm-base-0 0x10000004000,
+ *    vm-size-0 0xffff0000.
+ *  - spi2 at 0x3a5108000, irq 1241; PMGR device SPI2 at +0x440, parent SPI_P.
+ *  - mesa: "biosensor,mesa", sensor-id 0x3356, 8 MHz, CS0; power
+ *    function-mesa_pwr = NUB GPIO pin 36; data-ready interrupt on AP GPIO
+ *    pin 200.
+ * The SEP and its DART stay disabled here: apple_sep enables them at module
+ * init (drivers/soc/apple/dt.rs).
+ */
+&pmgr {
+	ps_spi2: power-controller@440 {
+		compatible = "apple,t8142-pmgr-pwrstate", "apple,t8103-pmgr-pwrstate";
+		reg = <0x440 4>;
+		#power-domain-cells = <0>;
+		#reset-cells = <0>;
+		power-domains = <&ps_spi_p>;
+		label = "spi2";
+	};
+};
+
+&{/soc} {
+	sep: sep@30a600000 {
+		compatible = "apple,sep";
+		reg = <0x3 0x0a600000 0x0 0x88000>;
+		mboxes = <&sep_mbox>;
+		mbox-names = "mbox";
+		iommus = <&sep_dart 0>;
+		status = "disabled";
+	};
+
+	sep_mbox: mbox@30a608000 {
+		compatible = "apple,t8142-asc-mailbox", "apple,asc-mailbox-v4";
+		reg = <0x3 0x0a608000 0x0 0x4000>;
+		interrupt-parent = <&aic>;
+		interrupts = <AIC_IRQ 445 IRQ_TYPE_LEVEL_HIGH>,
+			     <AIC_IRQ 446 IRQ_TYPE_LEVEL_HIGH>,
+			     <AIC_IRQ 447 IRQ_TYPE_LEVEL_HIGH>,
+			     <AIC_IRQ 448 IRQ_TYPE_LEVEL_HIGH>;
+		interrupt-names = "send-empty", "send-not-empty",
+				  "recv-empty", "recv-not-empty";
+		#mbox-cells = <0>;
+		status = "okay";
+	};
+
+	spi2: spi@3a5108000 {
+		compatible = "apple,t8142-spi", "apple,t8103-spi";
+		reg = <0x3 0xa5108000 0x0 0x4000>;
+		interrupt-parent = <&aic>;
+		interrupts = <AIC_IRQ 1241 IRQ_TYPE_LEVEL_HIGH>;
+		clocks = <&clkref>;
+		power-domains = <&ps_spi2>;
+		#address-cells = <1>;
+		#size-cells = <0>;
+		status = "okay";
+
+		mesa: fingerprint@0 {
+			compatible = "apple,mesa-fingerprint";
+			reg = <0>;
+			spi-max-frequency = <8000000>;
+			spi-cpol;
+			spi-cs-setup-delay-ns = <20>;
+			spi-cs-hold-delay-ns = <20>;
+			enable-gpios = <&pinctrl_nub 36 GPIO_ACTIVE_HIGH>;
+			interrupt-parent = <&pinctrl_ap>;
+			interrupts = <200 IRQ_TYPE_EDGE_RISING>;
+			/* Per-device files, built from the Mac's own factory records. */
+			firmware-name = "apple/mesacal-j813.bin", "apple/msrk-j813.bin";
+			status = "okay";
+		};
+	};
+};
+
+&sep_dart {
+	apple,dma-range = /bits/ 64 <0x10000004000 0xffff0000>;
+	apple,dma-offset = /bits/ 64 <0x0>;
+};
+
+/* Expose the sensor's two pins next to the ones sound uses. */
+&pinctrl_ap {
+	gpio-reserved-ranges = <0 100>, <101 49>, <152 48>, <201 23>;
+};
+
+&pinctrl_nub {
+	gpio-reserved-ranges = <0 18>, <19 17>, <37 3>;
+};
diff --git a/drivers/soc/apple/profile.rs b/drivers/soc/apple/profile.rs
index ac5f573455b9..504fccc20c14 100644
--- a/drivers/soc/apple/profile.rs
+++ b/drivers/soc/apple/profile.rs
@@ -710,8 +710,9 @@ fn select(
     if has(b"apple,t8132") && (has(b"apple,j713") || has(b"apple,j715")) {
         return Ok(&T8132);
     }
-    // Only the J815, whose ADT was read; the J813's sensor bus is unchecked.
-    if has(b"apple,t8142") && has(b"apple,j815") {
+    // The J815 and the J813, whose ADTs were read (macOS 27.0 stub firmware):
+    // the J813's SEP, its DART, spi2 and the Mesa node carry the J815's values.
+    if has(b"apple,t8142") && (has(b"apple,j815") || has(b"apple,j813")) {
         return Ok(&T8142);
     }
     Err(ENODEV)

Not tested: deep sleep (s2idle only here), more than one finger, a lockout. aurora-xart status still prints "no profile for this Mac" on J813; its model list is separate from the driver's.

Assisted-by: Claude Fable 5.2

[gig3m]

The J815 als node from this PR also works on J813.

J813 (Mac17,3, 27.0 stub), same tree as my Touch ID comment. The J813 ADT has /arm-io/aop/iop-aop-nub/als. With the node below, aop_als binds and iio:device0 (aop-sensors-als) reports: in_illuminance_input from about 2 (covered) to about 1200 in a bright room, in_colortemp_input about 6800. The calibration (apple/aop-als-cal.bin, 193 bytes) was already in the vendor firmware from the installer.

Being identical to J815's, it could go in t8142-j813-j815.dtsi.

diff --git a/arch/arm64/boot/dts/apple/t8142-j813.dts b/arch/arm64/boot/dts/apple/t8142-j813.dts
index 5889b852793a..ef47c181c23a 100644
--- a/arch/arm64/boot/dts/apple/t8142-j813.dts
+++ b/arch/arm64/boot/dts/apple/t8142-j813.dts
@@ -121,6 +121,18 @@ &dockchannel {
 #include "t8142-j813-n1.dtsi"
 #include "t8142-j813-audio.dtsi"
 
+/*
+ * Ambient light sensor, the AOP's "als" service: an ST VD6286, as on J815
+ * (same ADT node, /arm-io/aop/iop-aop-nub/als). aop_als sends it the
+ * per-device calibration (apple/aop-als-cal.bin). No firmware-name: that
+ * names the CT817 setup-port message, which this sensor does not take.
+ */
+&aop {
+	aop_als: als {
+		compatible = "apple,t8142-aop-als", "apple,aop-als";
+	};
+};
+
 /*
  * N1 functions behind apcie port 0 (0000:00:00.0 -> 01:00.{0,1,2}), so the
  * loader has somewhere to put the factory addresses: it copies ADT /chosen

One cosmetic thing: apple_aop still logs no ALS calibration named in the device tree; the sensor stays off at boot, although this sensor needs no firmware-name and works.

Assisted-by: Claude Fable 5.2

[DjDeveloperr]

Accepted @gig3m's6837017905 J813 Touch ID report at its stated27.0-stub/223d-on92195659/Aura1.0.0b12/aurora-sep1-25 composition scope, including the additional DT/profile patch. It adds positive attach, single-finger enrol/match, macOS preservation and Linux reboot/persistence evidence. It does not retroactively qualify the unchanged public J813 DT or prove IRQ-only capture, rejection/lockout/multifinger/deep-sleep behavior.

Carried the supplied two-file follow-up separately in DRAFT#1732 at70035185, stacked on this e8fb89d parent. This keeps1532 DT-only. The SEP/mailbox/SPI/DART/GPIO topology was independently reviewed against J815 and the patch applies with exact contexts; board-specific calibration paths/audio pins remain correct. Seven KUnit selector cases were added; local command creation failed EAGAIN, so no local Rust/kernel test pass is claimed and hosted checks are required.

Critical policy scope: existing xart_writes=1/provision_keybag=1/guard-1 defaults are unchanged, but admitting J813 makes it eligible for that writable/provisioning behavior, with inherited guard1. The draft is not read-only by default and needs SEP/XART policy review. aurora-xart model reporting and the T8142/J815 diagnostic label remain separate limitations. No physical/XART/enrolment/provision/loader/default action occurred here; this report does not clear Aura165 alias qualification/adoption.

[DjDeveloperr]

Codex, reviewing on @DjDeveloperr’s behalf.

Carried @gig3m’s supplied J813 ALS node exactly in09e6fbbc, from report6837248959. The owner’s27.0-stub/previously stated composition produced illuminance about2–1200 and color temperature about6800; that is accepted at the reported scope, without claiming automatic brightness, sleep coverage or a boot of this newly integrated head.

Source review confirms the compatible pair matches the existing J815 node and binding. The VD6286 path in aop_als loads apple/aop-als-cal.bin as a service property. No firmware-name is needed: the AOP setup-port calibration path is for CT817, and its existing “sensor stays off” message is misleading for this case. No driver/calibration content is changed here.

Bounded symbolic parsing preserves all311 existing J813 nodes and their properties and adds only /soc/aop@390c00000/als, with properties equal to J815. Normal merges78656b43 (#1534) and9053c517 (#1732) carry this parent update while preserving the whole J815 audio DTS and the two-file SEP/profile net delta. Both actual three-way merge trees match the prepared trees; all histories remain intact.

Fresh Host/DT CI is queued; these are source checks, not a local DTB/kernel/hardware pass. Keep the drafts and Aura165 affected-loader qualification/adoption/alias gate. The now-merged Centauri#1530 source prerequisite is recorded; SEP policy/utility gates remain separate.