[rolotrealanis98]

The M5 Pro MacBook Pro 16-inch: one T6050 die, like the J716c, on t6050-max.dtsi.

The first 13 commits are #945 (with #935) as they stand today, cherry-picked onto integration/aurora (only the qspi binding needed a conflict fix: enum [t6050-qspi, t8152-qspi] + const t8140-qspi); they drop out once #945 lands.

New here:

• t6050-max: list the 18 CPUs in the ADT's order. m1n1's dt_set_cpus() matches CPU nodes by position and stopped on the J716s with "DT CPU 5 MPIDR mismatch: 0x40100 != 0x40005". J716c and J714c use the same file.
• t6050: the SMC's second GPIO bank as gpio-low (apple,smc-low-gpio), as on T6030/T6031.
• dt-bindings: apple,j716s; t6050 compatibles for s5l-fpwm, spmi, pmgr-pwrstate and pmgr (the T6050 trees already use them); apple,t6050-cluster-cpufreq with the T8112 fallback.
• t6050-j716s.dts: N1 on PCIe port 0 (platform ID 0x000c6050), keyboard backlight (FPWM0), left-front Type-C (ATC1) and right Type-C (ATC2, SN201202x on nub-spmi-a1) for USB, SMC sensors (fan mode keys F0md/F1md, temperature discovery), trackpad AFE reset on SMC GPIO gp1c, the T8112 cpufreq fallback. ATC3 and nub-spmi-a2 off (not in this ADT); the debug port (ATC0) is not described.
• j716s.config: the boot-path drivers built in (the root file system has no modules).

Tested on J716s: this exact DTB plus a local node for the debug-port gadget (our tether), on integration/aurora 7.3 with the T6050 series: NVMe root, keyboard and trackpad (DockChannel MTP), USB3 on both Type-C ports (SuperSpeed Plus on the right port, hot-plugged), network, SMC battery, fans and temperatures, the N1 adopted with Wi-Fi and Bluetooth, 18 CPUs scaling through the T8112 fallback, s2idle. Not booted on J716c or J714c.

Checks: make dtbs clean, W=1 clean for the four T6050 boards, dt_binding_check clean for the touched bindings. dtbs_check on t6050-j716s.dtb: 53 lines, none from a node this file adds; the rest are the T6050 bindings that live in #936, #938, #939, #941, #946 and #481 (J716c on the same tree: 58).

Dependencies: #945 (contained). Runtime: #936 (NVMe), #939 (combo PHY, dwc3), #941 (SMC), #937/#938 (N1). cpufreq: with #481, its T6050 entry takes over from the fallback; if #481 lands first, the fallback override and its binding line can go instead.

Open questions: whether you prefer this stacked on the j716c/dt branch instead of carrying #945; a panel node for the boot framebuffer (the file has none, so the connector reports as Unknown rather than eDP); trackpad AFE reset via gp1c (tested) vs the pcio template route once #941 lands.

[RyanTheTide]

Thanks for this, and for booting it on a real J716s. Ryan wants the T6050 trees in, so this PR is in scope and I've pushed our fixups on top. Your commits stay as they are.

What I checked

• The J716s file against the J716s ADT: the register windows and interrupts of everything it adds (PCIe port 0, the HPM buses, FPWM0) match.
• The CPU order. The J775D ADT lists its cores the same way, cluster by cluster and die by die, so t6050-ultra.dtsi already follows it and needs no change.
• make W=1 and CHECK_DTBS=y on the J716s, J716c and J775D after the merge: no warnings.

What I pushed

• A merge of current integration, which now has arm64: dts: apple: add the T6050 (M5 Ultra) SoC and J775D Mac Studio device trees #935 and arm64: dts: apple: add the J716c (MacBook Pro 16-inch, M5 Max) #945 with their review fixups. The copies of those commits in this branch have different hashes from the merged ones, so git sees them as separate commits with the same content. I kept your CPU order and the gpio-low bank and took the reviewed versions of the other T6050 files.
• Dropped arch/arm64/configs/j716s.config. The tree builds every Mac from defconfig plus asahi.config and keeps no per-board fragments. If the J716s needs an option asahi.config lacks, a PR adding it there is welcome. Building the boot path in rather than as modules is a choice for your distro config.
• Dropped the default bootargs = "maxcpus=1", as on the other T6050 boards.

On your open questions

• Carrying arm64: dts: apple: add the J716c (MacBook Pro 16-inch, M5 Max) #945 instead of stacking was fine, the merge sorted it out.
• The cpufreq override with the T8112 fallback is fine for now. The binding has both forms. Once cpufreq: apple: Add the T6050 thermal-owned profile #481 lands the board override can go.
• A panel node for the boot framebuffer and the trackpad reset route can be follow-ups.

Build-tested here, and you've booted it on the J716s, so this goes in once CI is green.

[rolotrealanis98]

Thanks for the review and the fixups. Dropping j716s.config is fine; the built-in boot path stays in our distro config, and if we find an option asahi.config lacks we'll send it there. I'll follow up with the panel node for the boot framebuffer and the trackpad reset route, and rebase #1183 once this is in.