#
✕ nick:sven
#asahi-dev · 2026-01-13 16:27
svendo you still have the panic serial output somewhere?
#asahi-dev · 2026-01-13 16:27
svenhuh
#asahi-dev · 2026-01-13 16:20
svenPR for the serial fix: https://github.com/AsahiLinux/m1n1/pull/531
#asahi-dev · 2026-01-13 16:20
svenwhat was the issue?
#asahi-dev · 2026-01-13 16:20
svennice
#asahi-dev · 2026-01-13 16:20
sven:D
#asahi-dev · 2026-01-13 14:48
svenit may be that m4+ or whatever used the dockchannel UART as default with an earlier macOS release already
#asahi-dev · 2026-01-13 14:47
sven:D
#asahi-dev · 2026-01-13 14:47
svenI’m not sure when that was added, all i know is that even m1 already has a dockchannel UART node
#asahi-dev · 2026-01-13 14:24
svenweird :D
#asahi-dev · 2026-01-13 13:59
svenhave you tried chainloading xnu? I think that used to work as well and would rule out issues with the hv
#asahi-dev · 2026-01-13 13:58
svenand I’m not sure if it removes anything the hv needs
#asahi-dev · 2026-01-13 13:58
svenbut that also removes usb
#asahi-dev · 2026-01-13 13:58
svenBRINGUP
#asahi-dev · 2026-01-13 13:51
svenat this point nothing in m1n1 should touch dcp i think :/
#asahi-dev · 2026-01-13 13:51
svenhmmm
#asahi-dev · 2026-01-13 13:26
svenweird
#asahi-dev · 2026-01-13 13:15
svenyapz337: yes
#asahi-dev · 2026-01-13 13:12
svenmaybe they don’t have an error path for “coproc never comes up” and the thread just waits forever
#asahi-dev · 2026-01-13 13:11
svenNice, go on then and ignore my suggestions :D
#asahi-dev · 2026-01-13 13:10
svenhave you tried disabling all dcp code in m1n1? Iirc it talks to the iboot endpoint
#asahi-dev · 2026-01-13 13:09
svenweird
#asahi-dev · 2026-01-13 13:08
svenhuh, does it work when you boot macOS as intended?
#asahi-dev · 2026-01-13 13:05
svenright, that's something else then and likely not related to when the monitor was plugged in
#asahi-dev · 2026-01-13 13:04
svenis hdmi not working on macos or on linux?
#asahi-dev · 2026-01-13 13:02
svenstill sounds like progress :)
#asahi-dev · 2026-01-13 13:01
svenon m1/m2? would be good to submit a PR to fix it either way :)
#asahi-dev · 2026-01-13 11:38
svenlinux still uses the uart that works with both hv and cs
#asahi-dev · 2026-01-13 11:37
svenno, that's macos only
#asahi-dev · 2026-01-13 11:02
svenyapz337: good to hear, this was completely untested :D will open the m1n1 PR with that change later
#asahi-dev · 2026-01-13 11:01
svenfrom boot to first admin command, includes sart and asc as well
#asahi-dev · 2026-01-13 11:01
svenintegralpilot: https://f.svenpeter.dev/eed5d7f6e27e2f2f0c9dbf6d4758a2e630f71a1d8cdc4efa3a4fffc970461c0d_nvme_trace.log
#asahi-dev · 2026-01-13 10:06
svenself.adt["defaults"].serial_device = getattr(self.adt["/arm-io/uart0"], "AAPL,phandle") in setup_adt() in proxyclient/m1n1/hv/__init__.py might be enough to get xnu serial output back
#asahi-dev · 2026-01-13 09:02
svenyeah
#asahi-dev · 2026-01-13 08:54
sventhough I think either doing the ADT patching in run_guest or just adding virt_dockchannel.py to the HV is probably more important and easier
#asahi-dev · 2026-01-13 08:54
svengood point
#asahi-dev · 2026-01-13 08:50
svenThat’s where XNU’s output now ends up unless you mess with the ADT on newer macOS versions
#asahi-dev · 2026-01-13 08:50
svenyes
#asahi-dev · 2026-01-13 08:49
svenguess that’s another yak to shave: figure out if that new UART ends up in that usb debug thingy and then maybe also write a simple Linux driver
#asahi-dev · 2026-01-13 08:48
svenyeah
#asahi-dev · 2026-01-13 08:47
svenyeah, that’s why I suspect the dockchannel UART is somewhere in there as well
#asahi-dev · 2026-01-13 08:47
svenjust needs someone to implement it
#asahi-dev · 2026-01-13 08:47
svensure, it’s looks like just another serial driver
#asahi-dev · 2026-01-13 08:46
svenmaybe in that strange USB debug thing?
#asahi-dev · 2026-01-13 08:46
svenI also wonder where the other end of that dock channel UART is
#asahi-dev · 2026-01-13 08:42
svenshould be easy to virtualize that
#asahi-dev · 2026-01-13 08:42
svenand it looks fairly simple (which now that i think about it make perfect sense for what's essentially a UART)
#asahi-dev · 2026-01-13 08:41
svenlooks like the full dockchannel serial implementation is in pexpert/arm/pe_serial.c in open source xnu
#asahi-dev · 2026-01-11 23:39
svenbut that's probably not it then
#asahi-dev · 2026-01-11 23:39
svenokay, i was wondering if they maybe changed something subtle about gxf/sprr that SPTM needs with m3 which now breaks the hv in strange ways

plus ›