#asahi-dev · 2026-01-24 04:56
integralpilotif you remove all arguments about serial (just have the arg "debug") and just have it write to the iBoot framebuffer, do you see anything?#asahi-dev · 2026-01-24 04:55
integralpilotyay!#asahi-dev · 2026-01-24 02:21
fl0_id__https://www.pastery.net/mhtuha/#asahi-dev · 2026-01-24 02:19
fl0_id__I messed with the chickens again a bit and I got further...#asahi-dev · 2026-01-23 22:45
fl0_id__if I just && it seems to not reconnect or sth. or yeah, tbh it seems like unreliable evice names https://www.pastery.net/vynadt/#asahi-dev · 2026-01-23 22:43
fl0_id__now I think I do.. just if I plugged in CS later, then it was usually ACM2, but when rebooting, it would be ACM0#asahi-dev · 2026-01-23 22:42
noopwafeloh, you don't get reliable device names?#asahi-dev · 2026-01-23 22:41
noopwafelwhat is the problem if you just && it..?#asahi-dev · 2026-01-23 22:40
fl0_id__is there a trick to chainloading m1n1 before linux.py? it seems that is a problem if I just && it, but I need a newer m1n because I disabled cpufreq there. if I just use linux.py, it fails on cpufreq#asahi-dev · 2026-01-23 22:33
fl0_id__https://www.pastery.net/pjzprs/#asahi-dev · 2026-01-23 22:33
fl0_id__at least I have backtrace symbols back...forgot I messed around in proxyclient/m1n1/hv/__init__.py with symbol offset...#asahi-dev · 2026-01-23 22:28
fl0_id__mmh the cs output seems to be connected now properly. but no m1n1 or linux output there. https://www.pastery.net/bydmnv/#asahi-dev · 2026-01-23 22:09
jannauyou can keep cs connected. and use /dev/m1n1 / dev/m1n1-sec for everything m1n1 related#asahi-dev · 2026-01-23 22:08
fl0_id__yes#asahi-dev · 2026-01-23 22:07
jannaudid you "install" udev/80-m1n1.rules#asahi-dev · 2026-01-23 22:06
fl0_id__or should I then use the CS micro usb for m1n1?#asahi-dev · 2026-01-23 22:06
fl0_id__thx for the feedback. I meant also like when to connect the CS. Like if I just leave it connected when restarting, it seems to sometimes be ACM0, and then m1n1 doesn't work to chainload etc#asahi-dev · 2026-01-23 21:55
svenunfortunately not enough to bring out the other end of the dockchannel serial #asahi-dev · 2026-01-23 21:55
sventhere’s also code to talk to that debugusb thing in libirecovery #asahi-dev · 2026-01-23 21:54
svenmight be another alternative for an easy hardware serial #asahi-dev · 2026-01-23 21:54
svenif someone is bored: there’s apparently a AppleKIS.kext in 12.3 and 12.3.1 and we know how to put the Mac’s into KIS/DebugUSB mode and could do that from m1n1#asahi-dev · 2026-01-23 21:42
svenso if that doesn't work there's something wrong with the CS or with your USB cable#asahi-dev · 2026-01-23 21:42
svenyeah, m1n1 will send all its output over the hardware serial even before usb is up#asahi-dev · 2026-01-23 21:39
yukahmm, if you want we can test your CS/cable/... vs mine some time. but you should definitely get output from m1n1 when you restart (Ctrl+_ '!' on the CS)#asahi-dev · 2026-01-23 21:38
fl0_id__I think so... though looking at them again, the control messages are wrong. but it was giving the correct one with the same cable#asahi-dev · 2026-01-23 21:33
yukathe usb-c cable needs to have all the pins wired up#asahi-dev · 2026-01-23 21:33
yukaare you sure your cable is good?#asahi-dev · 2026-01-23 21:32
fl0_id__yuka, no, the control messages are there#asahi-dev · 2026-01-23 21:32
yukathere should be at least the control messages from the CS#asahi-dev · 2026-01-23 21:32
yukano output at all?#asahi-dev · 2026-01-23 21:29
fl0_id__yuka, I'm prob doing sth wrong with in regards to which serial port goes where...? the CS gets recognized, and should be on ACM2, but there's no output. I now only used /dev/m1n1 (ACM0) to chainload and load kernel, not the ACM1 port#asahi-dev · 2026-01-23 21:21
fl0_id__yes. it got an input/output error "read zero bytes" after, but not sure if that is from resetting myself or not#asahi-dev · 2026-01-23 21:19
yukais the kfence: the last line you see?#asahi-dev · 2026-01-23 21:19
yukathe second log looks like it is cut short#asahi-dev · 2026-01-23 21:19
fl0_id__ah ok... makes sense#asahi-dev · 2026-01-23 21:19
yukayou need the physical serial to observe baremetal boot#asahi-dev · 2026-01-23 21:18
yukaare you connecting to m1n1 by the virtual uart serial?#asahi-dev · 2026-01-23 21:18
fl0_id__is NUMA: Faking ... a potential error or does it not matter at this stage?#asahi-dev · 2026-01-23 21:17
fl0_id__my m1n1 backtrace is still missing symbols so not that helpful#asahi-dev · 2026-01-23 21:17
fl0_id__with hv and the right address/reg also added to the reg property, it continues a bit further https://www.pastery.net/urfxmv/ (and then seemingly hangs)#asahi-dev · 2026-01-23 21:16
fl0_id__yuka, without hv, it fails much earlier with a serial exception. https://www.pastery.net/ynxkxz/#asahi-dev · 2026-01-23 20:44
fl0_id__ah ok. I'll try that#asahi-dev · 2026-01-23 20:41
yukaif you declare all cpus in the device tree and then comment out the call to smp_start_secondaries() in linux.py, it also works nicely by pruning all cpus except the boot cpu#asahi-dev · 2026-01-23 20:41
yukacan you try baremetal (linux.py)?#asahi-dev · 2026-01-23 20:41
fl0_id__yes#asahi-dev · 2026-01-23 20:41
yukaAh you're booting in the hv right?#asahi-dev · 2026-01-23 20:40
fl0_id__oops didn't even notice that...#asahi-dev · 2026-01-23 20:40
svencache shouldn't matter#asahi-dev · 2026-01-23 20:40
yukaThe cpu pruning feature of m1n1 is very nice#asahi-dev · 2026-01-23 20:40
fl0_id__yeah that's waht I thought. In an earlier more complete device tree I had this for all, but cannot find it atm (it didn't work at the time for other reasons)