#
✕ nick:sven
#asahi-dev · 2026-05-17 19:25
svenor almost verbatim i think
#asahi-dev · 2026-05-17 19:25
svenweird. in the long response they were verbatim for m1 but i suspected an offset for m3
#asahi-dev · 2026-05-17 19:25
svenbut maybe that was a red herring then
#asahi-dev · 2026-05-17 19:25
svenbecause at least for m1 the address was also in the very long response
#asahi-dev · 2026-05-17 19:24
sveninteresting. must've overlooked that then
#asahi-dev · 2026-05-17 19:23
svenis that the very large one?
#asahi-dev · 2026-05-17 19:15
svenmaybe whatever co-proc runs that debugusb stuff is just busy with something sometimes? or maybe it's just slow? :/
#asahi-dev · 2026-05-17 18:42
sventhat code is fairly dumb so should hopefully be easy enough to spot the bug
#asahi-dev · 2026-05-17 18:41
svencould be
#asahi-dev · 2026-05-17 18:40
sveni'll try getting some with wireshark and the latest macos
#asahi-dev · 2026-05-17 18:40
svendon't have them because i couldn't easily get the m3 pro into debugusb mode while having my usb sniffer connected. the m3 is my $work machine with MDM crap :(
#asahi-dev · 2026-05-17 13:51
sven*replay
#asahi-dev · 2026-05-17 13:51
svenbut that would prove that the issue is inside atcphy somewhere
#asahi-dev · 2026-05-17 13:51
sventhat's gonna be annoying
#asahi-dev · 2026-05-17 13:51
svenone thing i wanted to try is reply the macos atcphy/dwc3/etc. trace while commenting out all of that inside linux and just triggering the acio manually somehow but ugh...
#asahi-dev · 2026-05-17 13:50
svenyeah, i tried hacking around quite a bit and never got anywhere as well :/
#asahi-dev · 2026-05-17 13:47
svenso that leaves something subtle in atcphy i think :(
#asahi-dev · 2026-05-17 13:46
svenwhich i think is confirmed with that new acio log message now
#asahi-dev · 2026-05-17 13:46
sveni think it's atcphy because i went through all commits that touched thunderbolt/ since the last branch that worked and there wasn't anything suspicious there
#asahi-dev · 2026-05-17 13:44
sveni think this is an issue inside atcphy fwiw
#asahi-dev · 2026-05-17 13:41
svenpity
#asahi-dev · 2026-05-17 13:41
svenaw... i was hoping they included all the asserts now
#asahi-dev · 2026-05-17 13:37
svencan you run strings on the acio fw?
#asahi-dev · 2026-05-17 13:37
svenmore verbose logging would've saved me a lot of headaches when i did the initial bringup
#asahi-dev · 2026-05-17 13:36
svenguess i'll have to upgrade my mini to that one
#asahi-dev · 2026-05-17 13:35
svenhah, so that unknown apple VSE thing is actually called "Lane init parmas" :D
#asahi-dev · 2026-05-17 13:32
svenwhich macos firmware is that?
#asahi-dev · 2026-05-17 13:32
svenacio used to only contain a single "assert %d violated" string
#asahi-dev · 2026-05-17 13:32
sventhat's new
#asahi-dev · 2026-05-17 13:32
svenhuh, yeah
#asahi-dev · 2026-05-17 13:26
sveni.e. after writing to that apple VSE register nothing happened
#asahi-dev · 2026-05-17 13:26
svenso before that
#asahi-dev · 2026-05-17 13:26
svenjannau: uh.. i don't quite remember, but I think it didn't even see the connection
#asahi-dev · 2026-05-17 10:11
svenbut I can take a look once im home
#asahi-dev · 2026-05-17 10:11
svenI don’t think I have a dump for the m3 pro, I just know that getting the address from that 0x10000 blob didn’t work
#asahi-dev · 2026-05-17 08:29
svenyeah, the config one is enable_portal and you enable RPM and PPM there
#asahi-dev · 2026-05-17 08:28
svenand for writing i definitely needed the correct address
#asahi-dev · 2026-05-17 08:27
svenyou can also read/poll with a PAR if you know the correct address
#asahi-dev · 2026-05-17 08:27
sventhe PPM one enables reading I think
#asahi-dev · 2026-05-17 08:26
sventhe config one is probably the PCW that does enable_portal
#asahi-dev · 2026-05-17 08:26
svenKISPortal = Enum(Int8ul, Config=0x01, RSM=0x10, PAM=0x11, PPM=0x13)
#asahi-dev · 2026-05-17 08:25
sven*enable_portal
#asahi-dev · 2026-05-17 08:25
sveni guess one of those is enable_port
#asahi-dev · 2026-05-17 08:23
svenwhat commands are those?
#asahi-dev · 2026-05-17 08:23
sveni can't read that hex by heart :D
#asahi-dev · 2026-05-17 08:23
svenhow do you write to the serial? i had to give it an address there
#asahi-dev · 2026-05-17 08:21
svenand the m1 address comes from somewhere inside par(seq=0, portal=0, offset=13, addr=0, len_=0x10000) but when i used the address that m3 returned it didn't work
#asahi-dev · 2026-05-17 08:21
svena = par(seq=0x1337, portal=KISPortal.PAM, offset=0, addr=0x23d134014, len_=4) <-- iirc that address was different on m3
#asahi-dev · 2026-05-17 08:19
svenhttps://f.svpe.de/e4731ec11516d79b9c675329ebf55b38134a22d30fa2251318b01867e9e61aee_enter-kis.patch i think that's what i used to enter debugusb from m1n1
#asahi-dev · 2026-05-17 08:18
svenhrm, m3 had a different address in the PCR or whatever thing

plus ›