#
✕ nick:marcan
#asahi-dev · 2023-05-01 10:49
marcan(the payload binary)
#asahi-dev · 2023-05-01 10:49
marcanif you send it to me I can try to make sense of it
#asahi-dev · 2023-05-01 10:49
marcanoh, like you made the payload yourself?
#asahi-dev · 2023-05-01 10:46
marcanare you using earlycon?
#asahi-dev · 2023-05-01 10:45
marcanbut probably in the kernel before it installs its own exception handlers, so it goes back to m1n1
#asahi-dev · 2023-05-01 10:45
marcanand not in m1n1
#asahi-dev · 2023-05-01 10:45
marcanthat's a NULL deref
#asahi-dev · 2023-05-01 10:44
marcanthere are very few commits to the good commit so it should be very obvious what is wrong from the guest exception
#asahi-dev · 2023-05-01 10:38
marcanthe guest exception would cause a panic which then triggers the hvc
#asahi-dev · 2023-05-01 10:37
marcancan you paste the guest exception?
#asahi-dev · 2023-05-01 10:37
marcanonly the HV would catch that
#asahi-dev · 2023-05-01 10:37
marcanyeah but it can't be that same exception because it's an hvc
#asahi-dev · 2023-05-01 10:36
marcanchadmed: same exception from the guest in the vUART?
#asahi-dev · 2023-05-01 10:36
marcanremember we don't guarantee ABI compat for the uart protocol, I only actively try to keep chainload.py working
#asahi-dev · 2023-05-01 10:35
marcanyeah, it would :p
#asahi-dev · 2023-05-01 10:35
marcanprobably means m1n1 is panicking which means you should be getting a message over the vUART
#asahi-dev · 2023-05-01 10:35
marcanchadmed: um, that hvc is a reboot() call
#asahi-dev · 2023-05-01 10:33
marcanI haven't had to manually invoke that reset script in a looooong time :)
#asahi-dev · 2023-05-01 10:32
marcanmaz: m1n1.setup already does the PMU counter reset, so almost all m1n1 scripts automatically do it, including chainload.py
#asahi-dev · 2023-05-01 04:58
marcanpaste the whole error?
#asahi-dev · 2023-05-01 04:58
marcanokay then I don't know :p
#asahi-dev · 2023-05-01 04:46
marcanFAR will often be the UART because of the vUART emulation going on all the time
#asahi-dev · 2023-05-01 04:45
marcanlook at the L2C regs instead
#asahi-dev · 2023-05-01 04:45
marcanif this is an SError, the FAR is irrelevant
#asahi-dev · 2023-05-01 04:45
marcanchadmed: wait are you looking at the right number?
#asahi-dev · 2023-05-01 04:45
marcanpovik: nice :D
#asahi-dev · 2023-05-01 04:44
marcanthat usually means some UART power domain is getting turned off, though the HV should disallow that
#asahi-dev · 2023-04-30 16:41
marcanjannau: yeah, that's the only blocker as far as I'm concerned
#asahi-dev · 2023-04-30 14:56
marcanI *want* to say the spihid stuff might finally be stable now :)
#asahi-dev · 2023-04-30 14:56
marcanpushed
#asahi-dev · 2023-04-30 14:29
marcanmaybe that's a better idea
#asahi-dev · 2023-04-30 14:29
marcanwell I guess I could mutate the input device
#asahi-dev · 2023-04-30 14:28
marcannot that there's any power disadvantage to doing that right now, it does not actually shut down anything in the LL driver, so meh
#asahi-dev · 2023-04-30 14:28
marcanso instead I guess we just keep the LL device open
#asahi-dev · 2023-04-30 14:28
marcanwell... there's no hook for that, lovely
#asahi-dev · 2023-04-30 14:02
marcanI guess I have to make magicmouse send the init command unconditionally on every device open
#asahi-dev · 2023-04-30 14:01
marcanthe desktop environment closes the trackpad device when you close the lid. that makes sense but... yeah, then I lose the "omg reset" report sometimes.
#asahi-dev · 2023-04-30 14:01
marcanlol okay.
#asahi-dev · 2023-04-30 14:01
marcan[ 65.477961] apple_ll_close
#asahi-dev · 2023-04-30 13:47
marcannow I'm trying to figure out why the reset report never makes it to hid-magicmouse, but only sometimes (usually the second time it happens) :/
#asahi-dev · 2023-04-30 13:46
marcanjannau: it's asahi-wip plus spi debug stuff
#asahi-dev · 2023-04-30 13:20
marcancould be the cursor
#asahi-dev · 2023-04-30 13:20
marcanit happened when I first moved the mouse after plasma startup
#asahi-dev · 2023-04-30 13:08
marcanjannau: https://mrcn.st/p/z4Z89K8m
#asahi-dev · 2023-04-30 13:07
marcanumm... something else blew up...
#asahi-dev · 2023-04-30 13:06
marcanvalidating and will push in a bit.
#asahi-dev · 2023-04-30 13:06
marcanfixed the stupid trackpad stops working issue. I think what happens is sometimes the Broadcom touch controller freaks out, and then the STM times out and resets it. when that happens hid-magicmouse gets a report 0x60 and it has to re-send the HID report init/change packet to go back to regular MT mode.
#asahi-dev · 2023-04-29 08:05
marcanit's just a copy paste and they decided it wasn't worth it
#asahi-dev · 2023-04-29 08:05
marcanyup
#asahi-dev · 2023-04-29 08:01
marcanthey just dropped it

plus ›