[RyanTheTide]

Preliminary device trees for the MacBook Pros with the PROLED XDR display: K114c and K116c (M5 Max), K114s and K116s (M5 Pro). Nobody on the team has one of these yet, so none of this has run on hardware. Every board value comes from the boards' own ADT (macOS 27.2, 26B8125), checked against J716c's.

This is stacked on #935 and #945 (T6050 SoC and J716c). The first 12 commits are theirs and drop out once they merge. Only the last 7 belong here.

What's in it

• Bindings for the four boards, plus apple,t6050-pmgr-pwrstate (the T6050 trees already use it) and apple,t6050-fpwm.
• The three T6050 FPWM channels and their power states, disabled in the SoC file. FPWM1 and FPWM2 power states stay off unless a board uses them, because J716c's ADT ties FPWM1's state to its indicator controller.
• The boards, in a shared t6050-k114-k116.dtsi that follows J716c. CPUs, NVMe, SMC, boot framebuffer, console, keyboard and trackpad on the MTP, keyboard backlight on FPWM0.
• The rear Apple logo as a pwm-leds LED (apple::logo) on FPWM1 at the ADT's 25 kHz, with default-state = "keep" so it stays however the boot loader left it.
• The touchscreen interface ("kiwi" on the MTP transport), described but disabled. dockchannel-hid now ignores interfaces whose node is disabled.

The M5 Pro boards reuse the Max's CPU list. The static ADT lists the same CPUs for both, and the loader drops cores a part doesn't have using the ADT the Mac boots with. That needs confirming on a real M5 Pro boot.

Left out on purpose

• PCIe and the N1 (Wi-Fi, Bluetooth). The N1's platform ID needs the board ID, which isn't in the static ADT.
• DCP for the OLED panel. Boot framebuffer only, same as J716c today.

Waiting on hardware or follow-ups

• Touchscreen. It doesn't send finished HID touches, it sends raw sensor frames and the Mac computes contacts in user space. Plan is firmware load through the existing transport once its sequence is specified, raw frames through hidraw, and a separate user-space service that turns frames into MT contacts. The firmware comes from the user's own macOS at install (Admit the PROLED MacBook Pros (K114/K116, M5 Pro/Max) as experimental aurora-installer#107 collects it unconverted).
• Logo brightness policy (the ADT's nits mapping and per-device syscfg calibration) belongs in user space with values the installer extracts. The ADT's use-enable-pulse isn't handled yet.
• FPWM2 (mesa-backlight, 100 Hz) is described but left off until we know what it lights.

Tests (buildbox, arm64 cross)

• make W=1 for the four K boards plus J716c and J775D: clean.
• dt_binding_check on arm/apple.yaml, power/apple,pmgr-pwrstate.yaml, pwm/apple,s5l-fpwm.yaml, input/apple,dockchannel-hid.yaml: clean.
• CHECK_DTBS=y on the K boards: 66 warnings each, all inherited from the T6050 base series (J716c has 71). Nothing new from these commits.
• W=1 build of drivers/hid/dockchannel-hid, pwm-apple and leds-pwm: clean.
• No KUnit added. The driver change is one availability check and the existing tests don't cover OF lookups.
• No boot test. There's no K board to boot, and the shared-code change (the disabled-node check) can't change behaviour on any existing board since none of them has a disabled interface node.

The bindings and the SoC FPWM commits don't depend on the hardware and can merge once #935 and #945 do. The board commit needs a hardware test.

🤖 Generated with Claude Code

[RyanTheTide]

Review for the T6050 lane. With #935, #945 and #1167 merged, this one now carries only its own commits on top of integration.

What I checked

• make W=1 and CHECK_DTBS=y on the four K boards and the J716s after the merge: no warnings.
• The compiled J716s tree before and after this PR: every node and property is the same, this PR only adds the disabled FPWM1/FPWM2 channels and their power states.

What I pushed

• A merge of current integration. arm/apple.yaml conflicted: the K boards now sit in the M5 Pro and M5 Max entries next to the J714/J716 boards. The pwm and pmgr-pwrstate bindings already had T6050 from the earlier merges.
• J716s: arm64: dts: apple: add the J716s (MacBook Pro 16-inch, M5 Pro) #1167 described FPWM0 and its power state in the board file. With this PR's SoC-level FPWM nodes that clashes, so the J716s now just enables &fpwm0. The compiled node is the same as the one Rolando tested.
• Dropped the default bootargs = "maxcpus=1" from t6050-k114-k116.dtsi, as on the other T6050 boards.

Build-tested, unbooted. Nobody on the team has a K board, so this goes in as preliminary and untested, which is what aurora-silicon/aurora-installer#107 and aurora-silicon/aurora-silicon.github.io#16 say too.