02:15 <monthebiff> morning all asahi o/

13:51 <Andjrfi> Hi. i'm experiencing some problem with DCP on M1 macbook air (A2337). trying to boot MacOs 15.5 (24F74) under the hypervisor with run_guest.py (development, but tried release as well and got the same issue). i get the following panic string: "No response received in 20s, DCP not started? "

13:52 <Andjrfi> main branch on m1n1

13:52 <chaos_princess> you need to use 13.5, dcp protocols change between versions

13:54 <Andjrfi> newer MacOs versions are not supported? only 13.5?

13:55 <chaos_princess> pretty much. asahi pretends to be 13.5, there really isn't much of a reason to use newer os-paired firmware on m1

14:06 <Andjrfi> So assuming I wanna boot a recent-ish XNU (Sequoia/Tahoe) via m1n1, I'd have to embed a matching DCP firmware within m1n1? How much effort would that entail?

14:07 <chaos_princess> it is impossible as the dcp firmware is loaded by iboot2 which checks signature

14:08 <chaos_princess> actually, wait a sec, that shouldn't even matter

14:08 <chaos_princess> are you tracing dcp?

14:08 <Andjrfi> Nope

14:09 <chaos_princess> and thats a xnu panic, not m1n1

14:09 <Andjrfi> Just trying to get XNU to properly boot (and show a working maCOS UI)

14:09 <Andjrfi> Yeah, XNU panic (in RTBuddy kext)

14:09 <chaos_princess> and to make sure, the kernel you are loading and macos rootfs are matching?

14:10 <Andjrfi> I'll double-check, but 99% yes

14:10 <chaos_princess> should work then i guess

14:18 <Andjrfi> Yeah, it's the same version exactly. same problem happens with dev+release kernels (macOS 15.5 / xnu-11417.121.6~2)

14:19 <Andjrfi> And I also have a single macOS partition on this M1, so it has to be the same filesystem :|

14:20 <chaos_princess> there is a chance that something broke in 15.x, since nobody cares about this usecase (it is not useful right now, or maybe at all 15.x is kind of a lost version for asahi purposes)

14:20 <chaos_princess> mind trying 14.x?

14:24 <jannau> that's probably caused by m1n1 trying to work around on sequoia bugs. try changing FW_MAX to V15_3 in https://github.com/AsahiLinux/m1n1/commit/277e8e71de92dbb2ff0dba9d49b2b27a97e23a08

14:26 <jannau> issue is either touching dcp at all when booting or not putting it back into the correct sleep state

14:26 <Andjrfi> jannau: I'll give it a try, thanks! If nothing works I'll check 14.x for sanity

14:27 <jannau> Andjrfi: a possible workaround would be to apply that workaround only when OS FIRMWARE version is 13.5 or older

16:44 <Andjrfi> I changed the if-statement so that i wont enter that case with the workaround. I then got a couple of panics in AppleS5L8940XI2Controller and AppleARMIICController, and now it booted fine (i didnt change anything else). so it would seem that the workaround was indeed the issue on 15.5, but it seems like there is another issue which is not deterministic (or is this the issue the workaround was meant to solve??).

16:45 <Andjrfi> Thanks for the help ^_^