#
#asahi-dev · 2026-06-13 09:19
yukaok, this matches what I read from tuxvdmtool
#asahi-dev · 2026-06-13 09:18
jannaumy understanding is that DVEn is "Perform Action" on the local controller. so its data should be just 0x1824606 for configuring debugusb
#asahi-dev · 2026-06-13 09:13
yukaat least inside m1n1, I get responses of cmd_status = 0x444d4321 which is just TPS_CMD_INVALID with different endianness
#asahi-dev · 2026-06-13 09:12
yukaAlso I think the spmi interface might have different endianness
#asahi-dev · 2026-06-13 09:12
yukaYes, I know, but is there any documentation on what values I should pass to dven?
#asahi-dev · 2026-06-13 09:11
jannauyuka: devn is for acting on the local usb-pd controller, cmp serial() in tuxvdmtool
#asahi-dev · 2026-06-13 09:09
yukaalso, 0x03 0x06 would be for entering serial, not kis
#asahi-dev · 2026-06-13 09:09
yukaalso, these 32-bit vdms are passed as little-endian, while your patch seems to have them written out as bytes (but without applying the endianness conversion)
#asahi-dev · 2026-06-13 09:08
yuka(0x1840306)
#asahi-dev · 2026-06-13 09:08
yukathis effectively omits the "perform action" part and only passes the action
#asahi-dev · 2026-06-13 09:08
yukahttps://github.com/AsahiLinux/tuxvdmtool/blob/4f3fed6a8c7b6fc6f656efd8df7620b1dbcb30ee/src/cd321x.rs#L233
#asahi-dev · 2026-06-13 09:07
yukatuxvdmtool only passes one u32 vdm command
#asahi-dev · 2026-06-13 09:07
yukaHowever I'm also confused about the use of DVEn
#asahi-dev · 2026-06-13 09:07
yukaso it couldn't have worked
#asahi-dev · 2026-06-13 09:07
yukafirst I noticed last time I tried this, I didn't change the key
#asahi-dev · 2026-06-13 09:06
yukahttps://f.svenpeter.dev/e4731ec11516d79b9c675329ebf55b38134a22d30fa2251318b01867e9e61aee_enter-kis.patch
#asahi-dev · 2026-06-13 09:06
yukasven: you posted this "enter KIS" diff for m1n1 a while ago
#asahi-dev · 2026-06-13 01:26
NiklasSheth[m]I sent a DM to your matrix, lmk if that didn't go through
#asahi-dev · 2026-06-13 01:15
StanfordAppliedCyber[m]n1776[m]: @appliedcyber:matrix.org i also have no clue how this works
#asahi-dev · 2026-06-13 01:12
NiklasSheth[m]Sorry I'm new to Matrix/IRC
#asahi-dev · 2026-06-13 01:11
NiklasSheth[m]I tried but I think the bridge is causing problems, do you have a direct Matrix handle? I'm @n1776:matrix.org
#asahi-dev · 2026-06-13 00:50
StanfordAppliedCyber[m]n1776[m]: send me a dm, matrix is giving me issues
#asahi-dev · 2026-06-13 00:22
NiklasSheth[m]StanfordAppliedCyber[m]: Nice, send me the docs in a DM when you get a chance!
#asahi-dev · 2026-06-12 23:25
StanfordAppliedCyber[m]chaos_princess: yes
#asahi-dev · 2026-06-12 23:15
StanfordAppliedCyber[m]<NickChan[m]> "you probably should patch..." <- too difficult-- the kernel works fine, userspace has too many places and graceful fallback paths
#asahi-dev · 2026-06-12 13:26
floklioh, very cool
#asahi-dev · 2026-06-12 13:09
Sobek[m]* https://github.com/AsahiLinux/asahi-installer/pull/428 🥳
#asahi-dev · 2026-06-12 13:09
Sobek[m]https://github.com/AsahiLinux/asahi-installer/pull/428
#asahi-dev · 2026-06-12 11:30
NickChan[m]the apple counters still work since it's read-only, but the problem now is the arm counters can't be redirected so it may have an unexpected frequency
#asahi-dev · 2026-06-12 11:29
NickChan[m]<StanfordAppliedCyber[m]> "some of the registers were at..." <- you probably should patch every counter sysreg access to the same counter register
#asahi-dev · 2026-06-12 11:25
NickChan[m]since that's merely a read
#asahi-dev · 2026-06-12 11:25
NickChan[m]Though I *think* it should be possible to access the 24 MHz counter
#asahi-dev · 2026-06-12 11:24
NickChan[m]It's locked on m4 according to yuka's logs
#asahi-dev · 2026-06-12 11:17
chaos_princessis that one also locked in raw mode?
#asahi-dev · 2026-06-12 11:17
StanfordAppliedCyber[m]basically every apple sysreg is locked on the neo
#asahi-dev · 2026-06-12 11:16
chaos_princessdon't? just set SYS_IMP_APL_AGTCNTRDIR_EL1
#asahi-dev · 2026-06-12 11:16
StanfordAppliedCyber[m]the driver is still unhappy about something not being freed fast enough but it's alive, idgaf, cry me a river applesunrisewlan
#asahi-dev · 2026-06-12 11:16
StanfordAppliedCyber[m]some commpage field just fixed everything
#asahi-dev · 2026-06-12 11:16
StanfordAppliedCyber[m]solution: more lying to userspace
#asahi-dev · 2026-06-12 11:16
StanfordAppliedCyber[m]some of the registers were at 1ghz and some where at 24mhz
#asahi-dev · 2026-06-12 11:15
StanfordAppliedCyber[m]yeah that's the exact calculation
#asahi-dev · 2026-06-12 11:14
chaos_princess1ghz/24mhz ≈ 41.6
#asahi-dev · 2026-06-12 10:59
StanfordAppliedCyber[m]i can literally see the loginwindow process
#asahi-dev · 2026-06-12 10:59
StanfordAppliedCyber[m]final issue to fix: why isnt windowserver being spawned
#asahi-dev · 2026-06-12 10:58
StanfordAppliedCyber[m]wlan issue fixed, tl;dr clocks were off so everything was working properly but the timeout was artifically low (by a factor of ~42x) --> dext was timing out despite everything working perfectly fine
#asahi-dev · 2026-06-12 09:35
chaos_princessnot really beyond "just read the adt i guess"
#asahi-dev · 2026-06-12 09:34
StanfordAppliedCyber[m]i know generally what it looks like but it written out explicitly seems nice
#asahi-dev · 2026-06-12 09:34
StanfordAppliedCyber[m]or really any device
#asahi-dev · 2026-06-12 09:34
StanfordAppliedCyber[m]does anyone have a dependency tree of all the apple devices? ie wifi depends on pcie depends on pmgr (or whatever)
#asahi-dev · 2026-06-12 09:32
ScottJones[m]ScottJones[m]: Correction on my PMP data point: turns out PMP never booted — my m1n1 predates PR #545's dt_set_pmp, so the define silently no-op'd. Gist updated; will re-test with new m1n1.

plus ›