[AxxzyWasTaken]

display-gate.c enables the T6041 display nodes only on root compatible apple,t6040, the M4 Pro. The M4 Max (J616C, root apple,t6041) has the same display block at the same addresses, and aura#67 publishes the same apple,t6041-handoff for it. With no gate entry for that root, the handoff is ignored: the nodes stay disabled, appledrm has nothing to bind, the panel stays on simpledrm, and apple-mcc waits forever for /display-subsystem.

This adds an entry for apple,t6041 that shares gate_t6041_run(). Nothing changes on other machines, and without the marker the gate does nothing, as on the M4 Pro.

Testing (J616C, 14-core M4 Max, one-shot boot, no desktop session): this change on a 7.3 kernel with #914, #917, aura#67 and aura#68. The gate logs display handoff accepted and enables stages 0 and 1, apple-dcp binds, DRM initialises (Initialized apple 1.0.0, 3456x2234 at 120 Hz, fb0: appledrmfb), and apple-mcc stops deferring and enables the system level cache on 24 planes. Nothing is left in devices_deferred. Without this change, the same boot leaves all the display nodes disabled. Nobody logged in or ran a compositor on it, and external displays weren't tested. The gate also powers up the external display chain on this machine; on this boot the dcpext nodes came up as devices, but no external display was attached.

The build here is on our 7.3 trial branch, not on integration/aurora as such: the same hunks apply there, and this branch hasn't been built separately.

Needs #914 for the J616C device tree. This Mac is the only one tested.

[RyanTheTide]

Merge train (MTL, 06/11): this is superseded on integration/aurora. astaff's series already added a gate_t6041_max entry for root apple,t6041 in display-gate.c with its own gate_t6041_max_run() (no PMP on the M4 Max, external chains follow the boot loader's firmware memory), so this now conflicts. If the merged gate is missing something you need on your J616C, a small fix on top of it would be the way to go and I'll gate it.

[AxxzyWasTaken]

Closing: superseded by the gate_t6041_max entry from astaff's series on integration/aurora. Thanks.