#asahi-dev · 2026-08-02 18:27
chaos_princessdon't poke at smc, it is most likely probably not the right thing to do#asahi-dev · 2026-08-02 18:26
svenso you have to talk to smc?!#asahi-dev · 2026-08-02 18:26
unifiedgravity[m]SMC key: pmVC#asahi-dev · 2026-08-02 18:26
unifiedgravity[m]value: 0x101#asahi-dev · 2026-08-02 18:26
unifiedgravity[m]Rail is VDD_CIO power#asahi-dev · 2026-08-02 18:25
unifiedgravity[m]Uh I guess managed to stop it giving the async serror and just give a stack trace, but then the installation was hanging on the update-m1n1 and one time I rebooted and then after skip proxy it would now stay on asahi logo rather than going stage 2#asahi-dev · 2026-08-02 18:24
svenjust different tunables but m1n1 already takes care of that#asahi-dev · 2026-08-02 18:24
svenand no, PHY is the same#asahi-dev · 2026-08-02 18:24
svenwhat's rail?#asahi-dev · 2026-08-02 18:24
unifiedgravity[m]PHY + rail + ACIO register/power is the part with the differences#asahi-dev · 2026-08-02 18:22
svenhow did you approach bringup on the m1 pro? and how did you corrupt the installation?#asahi-dev · 2026-08-02 18:22
unifiedgravity[m]Weird#asahi-dev · 2026-08-02 18:21
svenuh, there's something wrong then. M1 pro is more or less the same as t8103#asahi-dev · 2026-08-02 18:21
unifiedgravity[m]should come out good soon.#asahi-dev · 2026-08-02 18:21
unifiedgravity[m]Okay I took a big chunk out of the M1 Pro USB4 bringup today; it is quite different from the t8103 M2 work that Sven did. Managed to advance the crash along but then corrupted my whole Asahi installation in the lockup and power cycling so calling it a day. Tomorrow I might get the M2 back out; I never managed to sell it. But yeah it's the acio which is the pain in the butt but I added a lot of dtb stuff related to it; hopefully#asahi-dev · 2026-08-02 17:28
svenjust finding it the first time was "fun"#asahi-dev · 2026-08-02 17:27
svenbut as i said, if you know it's there and if you know it's somewhere in PMGR you can also easily see it in the hv#asahi-dev · 2026-08-02 17:26
sventhat's the weird address#asahi-dev · 2026-08-02 17:26
svenApplePMGR::enableCioReconfig#asahi-dev · 2026-08-02 16:48
jannauyou might need an old m1n1 with that (not sure if we already removed code required for it)#asahi-dev · 2026-08-02 16:43
chaos_princessok, will try it out in a bit#asahi-dev · 2026-08-02 16:39
svenhttps://github.com/AsahiLinux/linux/commits/atcphy-20230219/ maybe#asahi-dev · 2026-08-02 16:38
svenlet me check#asahi-dev · 2026-08-02 16:38
chaos_princessdo you have the working branch anywhere?#asahi-dev · 2026-08-02 16:38
sveni'm really out of ideas wtf we are doing wrong#asahi-dev · 2026-08-02 16:37
chaos_princessdidn't help#asahi-dev · 2026-08-02 16:37
sven(where won't work means "assert %d failed" followed by a panic)#asahi-dev · 2026-08-02 16:37
chaos_princessSo, since the old phy code ran the phy reconfig in workqueue, one of my experiments involved also putting it into a workqueue and adding msleeps in random places#asahi-dev · 2026-08-02 16:36
svenyou have to take all of acio down when you disconnect a cable or the next connection won't work iirc#asahi-dev · 2026-08-02 16:35
chaos_princessspecifically the acio, not the m3 coproc, that one will not work#asahi-dev · 2026-08-02 16:35
chaos_princessalso acio can be started up before the phy. Not sure how that helps, but it seems to work#asahi-dev · 2026-08-02 16:34
chaos_princessi love(*) pmgr#asahi-dev · 2026-08-02 16:33
svenbut i got it because i enabled all debug i could find and saw something about a reconfig clock in xnu's output#asahi-dev · 2026-08-02 16:33
svenyou can trace it if you know what you're looking for#asahi-dev · 2026-08-02 16:32
svenyup#asahi-dev · 2026-08-02 16:32
chaos_princess...#asahi-dev · 2026-08-02 16:32
svenit's hardcoded in the pmgr kext :/#asahi-dev · 2026-08-02 16:32
chaos_princessyes#asahi-dev · 2026-08-02 16:32
svenis that the one somewhere in the middle of pmgr?#asahi-dev · 2026-08-02 16:31
chaos_princessgithub.com/AsahiLinux/linux/commit/f0daf084ff8f5717ec92bff380c816f9fb991475#diff-845c07f30f0f94856b22d7b50c0ff3d27a961abeffefcdfab01f2e8cd64569ccR1707 - this one#asahi-dev · 2026-08-02 16:31
svenfor acio?#asahi-dev · 2026-08-02 16:31
chaos_princesslike, i can maybe see where the second and third came from, but the first gets no hits in adt#asahi-dev · 2026-08-02 16:30
svenwhich one specifically?#asahi-dev · 2026-08-02 16:30
chaos_princesshow did you derive the addresses in dts#asahi-dev · 2026-08-02 16:30
chaos_princesssven: oh, right, i should have just asked#asahi-dev · 2026-08-02 16:30
svenchaos_princess: how is tbt coming along btw? ;p#asahi-dev · 2026-08-02 15:41
aviv_beHello yall, I'm a beginner linux developer and would like to join the development on Asahi. For the past few months I have been messing around with linux kernel development, writing drivers and a ram disk, and also worked on a large open source project through university. I have an m3 pro macbook pro for development.#asahi-dev · 2026-08-02 15:15
svenbut i'm not sure why this matters, just introduce a new compatible, add apple,firmware-abi (or however it's called for the gpu) and then just change behavior for macos 15+#asahi-dev · 2026-08-02 15:12
svenNULL_PRP_PTR_ERR F2H %u SL[%u] ERR 0x%x HIX %u SEG %u TAG %u OP %u DRAM %u LBA 0x%x secOffset 0x%x dp[%x, %x, %x, %x] mmu[%x, %x, %x, %x, %x, %x, %x] <-- pretty sure that's the error message i saw without that bit