#asahi-dev · 2026-06-28 17:03
yukawe never power on dispext_sys without also powering dispext_fe#asahi-dev · 2026-06-28 17:02
yukabut I guess we treat them all the same#asahi-dev · 2026-06-28 17:02
yukait still sounds wrong to have pmp report of DISPEXT_SYS technically report the state of dispext_fe#asahi-dev · 2026-06-28 17:01
chaos_princesspowerup = pmgr, then pmp, then driver. powerdown in reverse#asahi-dev · 2026-06-28 17:00
yuka"after" as in run-time order?#asahi-dev · 2026-06-28 17:00
chaos_princessso, 1#asahi-dev · 2026-06-28 17:00
chaos_princesspmp stuff should be after all pmgr stuff#asahi-dev · 2026-06-28 17:00
yukawhich would mean having ifdef APPLE_USE_PMP in the -pmgr.dtsi#asahi-dev · 2026-06-28 17:00
yuka2. replace the dependency of ps_disp*_fe on ps_disp*_cpu to pmp_report_disp*#asahi-dev · 2026-06-28 16:59
yukaeven though this is not actually what we want to signal (the id1 != 0 is still on disp_cpu not disp_fe)#asahi-dev · 2026-06-28 16:58
yuka1. Make pmp_report_disp* depend on ps_disp*_fe#asahi-dev · 2026-06-28 16:58
yukaI see two paths for this:#asahi-dev · 2026-06-28 16:57
yukait's fe -> cpu and not cpu -> fe#asahi-dev · 2026-06-28 16:56
yukathe disp/dispext power domain dependencies were inverted on t6031#asahi-dev · 2026-06-28 16:56
yukablah#asahi-dev · 2026-06-28 16:40
chaos_princessand in fact always-on breaks something#asahi-dev · 2026-06-28 16:40
yukapresumably#asahi-dev · 2026-06-28 16:40
yukabecause it starts out as enabled and we don't touch it#asahi-dev · 2026-06-28 16:40
chaos_princessi think it just never connects to it or something#asahi-dev · 2026-06-28 16:40
chaos_princessno node also works#asahi-dev · 2026-06-28 16:40
yuka(wait, shouldn't it be an always-on reporting node then, instead of no node?)#asahi-dev · 2026-06-28 16:39
yukathen I'll not include pmp_report_gfx on M3*#asahi-dev · 2026-06-28 16:39
yukaAh oki#asahi-dev · 2026-06-28 16:39
chaos_princesspmp does not work there for unrelated reasons#asahi-dev · 2026-06-28 16:39
yukajust nobody tested it?#asahi-dev · 2026-06-28 16:39
chaos_princessit doesn't#asahi-dev · 2026-06-28 16:39
yukahow does that work on t8112 then#asahi-dev · 2026-06-28 16:38
yukaOh, huh#asahi-dev · 2026-06-28 16:38
chaos_princesswe can't reinit the gpu, so we just don't let pmp power it down#asahi-dev · 2026-06-28 16:38
yukadid you trace it and determine the adt lies?#asahi-dev · 2026-06-28 16:38
yukawhy?#asahi-dev · 2026-06-28 16:38
yukachaos_princess: t600x and t602x have GFX notify_pmp, but only t8112 has it wired up in our dts#asahi-dev · 2026-06-28 16:15
yukado they actually check their moderation backlog or do I have to re-send it?#asahi-dev · 2026-06-28 16:15
yuka> Post by non-member to a members-only list#asahi-dev · 2026-06-28 16:15
yukau-boot mailing list...#asahi-dev · 2026-06-28 15:50
nickchani probably should write a binding for it and send it upstream#asahi-dev · 2026-06-28 15:50
nickchani could make use the apple,dma-range property in the downstream dart driver#asahi-dev · 2026-06-28 15:04
yukathe uboot dummy driver was too easy, it turns out when I tested it I still had the hack enabled#asahi-dev · 2026-06-28 14:13
chaos_princessit is not ever getting a "real" pmp driver anyway#asahi-dev · 2026-06-28 14:13
chaos_princessto u-boot? sure.#asahi-dev · 2026-06-28 14:10
yukashould I send that directly upstream?#asahi-dev · 2026-06-28 14:10
yukaworks!#asahi-dev · 2026-06-28 14:07
yukaI'm looking into a dummy driver for uboot now#asahi-dev · 2026-06-28 13:55
chaos_princess1-ish W sounds about right#asahi-dev · 2026-06-28 13:54
yukaI would say success#asahi-dev · 2026-06-28 13:54
yukaokay, now same setup, in TTY, some background services stopped, but *without pmp*: 5.48W#asahi-dev · 2026-06-28 13:48
chaos_princesson j316 the gpio is connected to pmic, so that is probably via smc#asahi-dev · 2026-06-28 13:45
yukaor which kind of gpio#asahi-dev · 2026-06-28 13:45
yukais it usually an smc gpio?