#asahi-dev · 2026-04-13 18:34
tryhard[m]pmp_v1_report.rs#asahi-dev · 2026-04-13 18:33
chaos_princesswhat is your file name#asahi-dev · 2026-04-13 18:33
tryhard[m]module_platform_driver! {... (full message at <https://matrix.org/oftc/media/v1/media/download/Ad_sTli1KDff3Chfii4SLLk1FxbjJeQwCuGNWoFKYIZ5ysfFz3HffkKvSQkX_cp2hNvOPemyo8VCukhuIDSkDmdCed0DeGNAAG1hdHJpeC5vcmcvbEdHU1ZvRlNLY0tBZVN3a210c2Z0YVRC>)#asahi-dev · 2026-04-13 18:32
tryhard[m]umm anybody has any idea for this error: ERROR: modpost: missing MODULE_LICENSE() in drivers/pmdomain/apple/pmp_v1_report.o?#asahi-dev · 2026-04-13 02:56
Advert[m]https://t.me/+c-LC9ed_hBgwYmZh#asahi-dev · 2026-04-12 19:56
tryhard[m]thx#asahi-dev · 2026-04-12 19:55
tryhard[m]uh yes i thought like the parent needs it for some reason#asahi-dev · 2026-04-12 19:54
chaos_princessso the child driver can get it#asahi-dev · 2026-04-12 19:54
tryhard[m]what for?#asahi-dev · 2026-04-12 19:53
chaos_princessand the parent driver puts a reference to itself into it's driver data#asahi-dev · 2026-04-12 19:52
tryhard[m]so its basically communicating through the parent driver data#asahi-dev · 2026-04-12 19:49
chaos_princesshttps://github.com/AsahiLinux/linux/blob/asahi/sound/soc/apple/aop_audio.rs#L676-L679 - you do something like this in probe, store the trait object, and then just call methods on it#asahi-dev · 2026-04-12 19:48
tryhard[m]say some power entry wants to send something#asahi-dev · 2026-04-12 19:48
tryhard[m]how does it actually get to the parent?#asahi-dev · 2026-04-12 19:46
chaos_princessit is based on a real thing that exists in aop firmware, but my point is that you can just make stuff up#asahi-dev · 2026-04-12 19:45
tryhard[m]so epicservice exists only for the communication between the parent driver and the childs?#asahi-dev · 2026-04-12 19:44
chaos_princessEPIC exists for like 2 devices, aop and dcp, nothing else has it, so for all intents and purposes, it is just a thing 2 drivers use to communicate#asahi-dev · 2026-04-12 19:43
chaos_princessyou can put whatever you want there#asahi-dev · 2026-04-12 19:43
chaos_princessso what#asahi-dev · 2026-04-12 19:42
tryhard[m]so yeah it uses this epicservice#asahi-dev · 2026-04-12 19:41
chaos_princessthat is the driver, and the shared trait is here https://github.com/AsahiLinux/linux/blob/asahi/rust/kernel/soc/apple/aop.rs#L24#asahi-dev · 2026-04-12 19:40
tryhard[m]in SndSocAopDriver?#asahi-dev · 2026-04-12 19:38
chaos_princessand there is a trait under rust/kernel that defines what calls are there#asahi-dev · 2026-04-12 19:38
chaos_princessif you look at aop_audio's initialization, there is a step that pulls out a reference to aop, and then just calls methods on it.#asahi-dev · 2026-04-12 19:36
tryhard[m]but that service exists anyway, not because they need to communicate#asahi-dev · 2026-04-12 19:36
tryhard[m]so what is the best approach for them to communicate? aop and aop_audio seems to be using a shared service#asahi-dev · 2026-04-12 19:34
tryhard[m]yeah that makes sense#asahi-dev · 2026-04-12 19:34
tryhard[m]uh i understand you now#asahi-dev · 2026-04-12 19:34
chaos_princessotherwise - you have to do it yourself#asahi-dev · 2026-04-12 19:34
chaos_princessif you do separate device for each entry, you get the power sequencing for free. #asahi-dev · 2026-04-12 19:34
tryhard[m]ok#asahi-dev · 2026-04-12 19:33
chaos_princesshttps://github.com/AsahiLinux/linux/blob/asahi-wip/arch/arm64/boot/dts/apple/t600x-die0.dtsi#L76 - this#asahi-dev · 2026-04-12 19:33
tryhard[m]if i have the bindings then linux would still do it#asahi-dev · 2026-04-12 19:33
tryhard[m]what do you mean#asahi-dev · 2026-04-12 19:32
chaos_princessyou have to do it manually from the driver instead of getting linux core to do it for you#asahi-dev · 2026-04-12 19:31
tryhard[m]if i*#asahi-dev · 2026-04-12 19:31
tryhard[m]ok but i have the genpd bindings then i can just use it in the same driver#asahi-dev · 2026-04-12 19:30
chaos_princessmaybe? the reason i split the reporting entries into separate devices is so i get the power ordering "for free" via genpd. if you look at the device trees for pmp v2, the entries themselves have power-domains entries that ensure the order of pmgr -> pmp -> device itself#asahi-dev · 2026-04-12 19:29
tryhard[m]which probably have many compiling errors but its a start#asahi-dev · 2026-04-12 19:29
tryhard[m]i have the bindings so...#asahi-dev · 2026-04-12 19:29
tryhard[m]yeah#asahi-dev · 2026-04-12 19:28
chaos_princessso everything is one big driver?#asahi-dev · 2026-04-12 19:28
tryhard[m]chaos_princess: could there be some kind of problem if i just merge the init driver and the report driver into one?#asahi-dev · 2026-04-12 18:46
yukaHowever, the watchdog is still rebooting after ~120s even though I added the wdt node :(#asahi-dev · 2026-04-12 18:45
yukafor serial0#asahi-dev · 2026-04-12 18:45
yukaI was missing the alias in the device tree#asahi-dev · 2026-04-12 18:45
yukaSo on my t8132 I made some progress in the sense that I figured out why the serial console wasn't properly working except with earlycon=s5l,...#asahi-dev · 2026-04-12 17:30
inmary[m]https://t.me/+pa-CiKYv9-5lNWM8#asahi-dev · 2026-04-12 17:30
inmary[m]Get in touch with this platform for greatness you’ll definitely thank me later
ℹ️❤️
👇🏻👇🏻👇🏻#asahi-dev · 2026-04-12 17:17
bennofsok my "fix" for the dp-altmode problem is now to patch the kernel to retry calling the oob hotplug hook until the dcp is available. this fixes the bootup issue