#asahi-dev · 2024-10-26 12:51
chaos_princessthe hypervisor part is only for exploring how macos does things and for some debugging, correct.#asahi-dev · 2024-10-26 12:50
chaos_princessyea, but who cares for exploratory code, you are gonna write real drivers in c/rust anyway#asahi-dev · 2024-10-26 12:49
chaos_princessno, why?#asahi-dev · 2024-10-26 12:44
chaos_princessfor example here is the entry point of the webcam driver that runs under m1n1 itself https://github.com/AsahiLinux/m1n1/blob/main/proxyclient/experiments/isp.py#asahi-dev · 2024-10-26 12:43
chaos_princessor you can pretend to be macos too, several drivers were first prototyped as python scripts in m1n1, and then rewritten to a real kernel driver#asahi-dev · 2024-10-26 12:42
chaos_princessit runs macos in a very thin vm that passes through almost all hardware#asahi-dev · 2024-10-26 12:41
chaos_princesshttps://github.com/AsahiLinux/m1n1/blob/main/proxyclient/m1n1/hw/sep.py - there is a beginning of a sep tracer here#asahi-dev · 2024-10-26 12:39
chaos_princessThe recommended approach is to use m1n1 (a hypervizor/debugger that can load python scripts) to intercept the macos->sep communication and figure it out from there#asahi-dev · 2024-10-26 12:38
chaos_princessthat is one approach, but it is somewhat recommended against, because it is not open source, and there might be legal issues if you dont know how to do it properly.#asahi-dev · 2024-10-26 12:34
chaos_princesssimilar deal for external touch id on apple keyboards, except over bt or something instead of over SPI#asahi-dev · 2024-10-26 12:33
chaos_princessfor that, you need SEP, the touch id sensor will only talk to it, and the only thing that the driver can do is initialization and then passing through encrypted buffers between sep and the sensor#asahi-dev · 2024-10-26 12:15
chaos_princessis there any hardware feature that you personally find missing?#asahi-dev · 2024-10-25 21:41
chaos_princesshttps://social.treehouse.systems/@chaos_princess/113370229410416973 - hmm, what do we have here :P#asahi-dev · 2024-10-25 16:58
chaos_princessIt is p.pmgr_adt_clocks_enable("adt path here") in m1n1 shell, not sure about how to do it with raw register writes#asahi-dev · 2024-10-25 13:13
chaos_princessno, you need to turn it on via pmgr#asahi-dev · 2024-10-25 12:56
chaos_princessdid you turn on the power to it?#asahi-dev · 2024-10-24 14:55
chaos_princessno i mean, what extra stuff does the exclave thing do, compared to what gl2 did previously?#asahi-dev · 2024-10-24 14:53
chaos_princesshow is it different from the old ppl/whatever it is called now?#asahi-dev · 2024-10-24 14:48
chaos_princessIs there a writeup on exclaves anywhere, cause eclecticlight's is kinda... light on the details. #asahi-dev · 2024-10-15 15:01
chaos_princessall of those are electron wrappers, just use the web version#asahi-dev · 2024-10-01 17:09
chaos_princesskinda? i think the name for the device class is "set-top box"#asahi-dev · 2024-10-01 17:07
chaos_princesst7000-handheld/t7000-desktop :P#asahi-dev · 2024-09-30 07:15
chaos_princessIs adp3.2 the arm dtk?#asahi-dev · 2024-09-27 15:32
chaos_princesstry searching in already existing device trees for other devices for "edp"#asahi-dev · 2024-09-27 15:32
chaos_princessthat - yes, but you probably do not need to describe the display itself#asahi-dev · 2024-09-27 15:20
chaos_princessbut i do not think you even need special dp support, iboot should initialize everything and then you treat it as just a display#asahi-dev · 2024-09-27 15:19
chaos_princessthere aren't, asahi specifically has effectively no dp support#asahi-dev · 2024-09-27 15:12
chaos_princessyea, thats probably edp#asahi-dev · 2024-09-27 15:10
chaos_princessare you sure it is real displayport and not embedded displayport?#asahi-dev · 2024-09-26 16:08
chaos_princesshttps://github.com/AsahiLinux/linux/blob/bd0a1a7d465fcb60685a2360565ed424bafff354/drivers/gpu/drm/bridge/samsung-dsim.c - you might try writing to registers as if it is this one and seeing if that works#asahi-dev · 2024-09-26 16:08
chaos_princessif it is samsung, the register layout should be completely different#asahi-dev · 2024-09-26 15:54
chaos_princesstry writing the packet data to offset 0x38, and header to 0x34#asahi-dev · 2024-09-26 15:48
chaos_princesscould be the same one then, but a different version#asahi-dev · 2024-09-26 15:48
chaos_princessoh, right i misread#asahi-dev · 2024-09-26 15:45
chaos_princessmight be a different ip block#asahi-dev · 2024-09-26 15:43
chaos_princessread 4 bytes at offset 0, what is there?#asahi-dev · 2024-09-26 06:51
chaos_princesssure go ahead then#asahi-dev · 2024-09-26 06:50
chaos_princessno, i mean, do i need to just say "sure go ahead", or do you want me to formally send that to mailing list#asahi-dev · 2024-09-26 06:36
chaos_princessIs there something specific i need to do?#asahi-dev · 2024-09-25 21:51
chaos_princessnickchan: m1 is ok with 2 byte writes#asahi-dev · 2024-09-25 14:08
chaos_princessWasnt able to check yesterday#asahi-dev · 2024-09-24 16:41
chaos_princessnow im kinda sad that my iphone 10 went spicy pillow and shattered the screen some time ago :P#asahi-dev · 2024-09-24 16:19
chaos_princesskk#asahi-dev · 2024-09-24 16:10
chaos_princessbut it does write 16 bits, no?#asahi-dev · 2024-09-24 16:10
chaos_princessoh?#asahi-dev · 2024-09-24 16:09
chaos_princesssince you are touching it all anyway, it might make sense to use this helper https://elixir.bootlin.com/linux/v6.11/source/drivers/gpu/drm/drm_mipi_dsi.c#L1339#asahi-dev · 2024-09-24 16:01
chaos_princesssure, will test in a bit#asahi-dev · 2024-09-24 15:39
chaos_princess99% sure it will, from poking at it in general, it implements a lot of mipi spec#asahi-dev · 2024-09-24 15:38
chaos_princess*11#asahi-dev · 2024-09-24 15:38
chaos_princessmakes sense, log2(2048) is 1