#asahi-dev · 2025-03-29 05:10
chaos_princesssomewhat annoying that we can't just dynamicuser it due to blackbox#asahi-dev · 2025-03-29 04:23
chaos_princessYou need to install the profile for it to work: https://developer.apple.com/services-account/download?path=/OS_X/OS_X_Logs/Bluetooth_macOS.mobileconfig#asahi-dev · 2025-03-27 22:38
chaos_princessbut, like, why is there a dma attach fail first, and then a aop attach fail?#asahi-dev · 2025-03-27 22:38
chaos_princesslike, ok, i got that wrong at first and dma channels are probed after aop devices are attached#asahi-dev · 2025-03-27 22:38
chaos_princessi am so confused too#asahi-dev · 2025-03-27 22:32
chaos_princessok, how many times is probe called#asahi-dev · 2025-03-27 22:32
chaos_princessaaaa#asahi-dev · 2025-03-27 22:32
chaos_princessoh#asahi-dev · 2025-03-27 22:26
chaos_princesscant be, otherwise https://github.com/AsahiLinux/linux/blob/asahi/sound/soc/apple/aop_audio.rs#L377 wouldnt get hit #asahi-dev · 2025-03-27 22:16
chaos_princesslike, i do not see the way to get the output above w/o calling probe() twice?#asahi-dev · 2025-03-27 22:15
chaos_princessbut "Unable to attach device 1885629744..." is an error from line https://github.com/AsahiLinux/linux/blob/asahi/sound/soc/apple/aop_audio.rs#L673#asahi-dev · 2025-03-27 22:15
chaos_princessi just do not understand how you got the output above then. "Unable to get dma channel" is printed from https://github.com/AsahiLinux/linux/blob/asahi/sound/soc/apple/aop_audio.rs#L377 , which is called here https://github.com/AsahiLinux/linux/blob/asahi/sound/soc/apple/aop_audio.rs#L595 and since it got an error, it should exit both SndSocAopData::new and SndSocAopDriver::probe.#asahi-dev · 2025-03-27 22:02
chaos_princessonly once and fails?#asahi-dev · 2025-03-27 22:02
chaos_princessinteresting#asahi-dev · 2025-03-27 22:01
chaos_princessi wouldn't be#asahi-dev · 2025-03-27 21:58
chaos_princess(without your fixes, ofc)#asahi-dev · 2025-03-27 21:57
chaos_princessjannau: mind adding a debug print to https://github.com/AsahiLinux/linux/blob/asahi/sound/soc/apple/aop_audio.rs#L517, and telling me how many times that gets called?#asahi-dev · 2025-03-27 21:42
chaos_princessok, yep, thats a case of someone (me) not handling eprobe_defer correctly. would be annoying to fix though unfortunatelty#asahi-dev · 2025-03-27 21:36
chaos_princessif so, that is enough#asahi-dev · 2025-03-27 21:36
chaos_princess"Unable to get dma channel", something something?#asahi-dev · 2025-03-27 21:35
chaos_princesscan you still look in dmesg, and how exactly is it failing?#asahi-dev · 2025-03-27 21:32
chaos_princessanything else in dmesg?#asahi-dev · 2025-03-27 21:28
chaos_princessand, like, does it matter? if platform device is created before admac comes up, should it not just probe later?#asahi-dev · 2025-03-27 21:27
chaos_princessshould it not do that automatically?#asahi-dev · 2025-03-27 21:26
chaos_princessno, never seen that one before, does aop handle EPROBE_DEFER incorrectly?#asahi-dev · 2025-03-24 17:45
chaos_princessm2max, i was too lazy to type out in full#asahi-dev · 2025-03-24 17:44
chaos_princesspro/max#asahi-dev · 2025-03-24 11:40
chaos_princessnp, its just i thought someone tested it on a m2 pro/max device#asahi-dev · 2025-03-24 11:17
chaos_princessif he trusts me with a ssh session, that would be even better#asahi-dev · 2025-03-24 11:17
chaos_princessso we can test on that#asahi-dev · 2025-03-24 11:17
chaos_princessactually, neal is getting a m2 soon, right#asahi-dev · 2025-03-24 11:16
chaos_princessthough i never uploaded my patched tracer#asahi-dev · 2025-03-24 11:16
chaos_princessIf they are competent, maybe try capturing the m1n1 trace#asahi-dev · 2025-03-24 11:14
chaos_princessanyway, stab in the dark, but try changing this to 0xf or something https://github.com/AsahiLinux/linux/blob/asahi/drivers/soc/apple/aop.rs#L32#asahi-dev · 2025-03-24 11:13
chaos_princessI thought janne had one#asahi-dev · 2025-03-24 11:12
chaos_princessbut didnt we test m2pro already?#asahi-dev · 2025-03-24 11:12
chaos_princessThat sounds endpoint related#asahi-dev · 2025-03-24 11:10
chaos_princessor audio subdev#asahi-dev · 2025-03-24 11:10
chaos_princessIs it aop device itself?#asahi-dev · 2025-03-24 11:08
chaos_princessif you do not start all of them, it just sits there doing nothing#asahi-dev · 2025-03-24 11:08
chaos_princessCould there be more endpoints than on 314#asahi-dev · 2025-03-24 11:05
chaos_princesswhich machine/fw?#asahi-dev · 2025-03-24 09:22
chaos_princessand from a cursory reading, i do not think there is anything that prevents them requiring a nda to disclose that#asahi-dev · 2025-03-24 09:21
chaos_princessbut who knows, i am not expecting much tbh#asahi-dev · 2025-03-24 09:21
chaos_princessmy idea was to go via the "this info is needed for debugging performance, since apple can optimize their apps better by knowing what is happening after the compiler"#asahi-dev · 2025-03-24 09:17
chaos_princessdont care and that is arm politics, i am more interested in getting the gpu/ane isa that way#asahi-dev · 2025-03-24 09:13
chaos_princessthe idea is to maybe use the dma on iphone to indirectly benefit the mac#asahi-dev · 2025-03-23 17:53
chaos_princessfrom what i remember from reading arm64 spec, you can disable translation in el2#asahi-dev · 2025-03-23 15:51
chaos_princesscould it be that you are jumping to a no-exec page then?#asahi-dev · 2025-03-23 15:51
chaos_princessplease use a pastebin next time