#asahi-dev · 2026-10-04 17:24
svenmostly SPTM stuff #asahi-dev · 2026-10-04 17:20
svenit‘s still a it of mess and requires binary patches to xnu and SPTM and I hope to get rid of most of them #asahi-dev · 2026-10-04 17:19
svenyeah, probably 1-2 more weeks until the entire thing is upstream #asahi-dev · 2026-10-03 11:09
sventwo minis and a MacBook so far… ordered another one for m5 though #asahi-dev · 2026-10-03 09:30
sven5 seconds or so after the cores exited the hv it would reboot #asahi-dev · 2026-10-03 09:30
svenyes#asahi-dev · 2026-10-03 09:11
svenhuh, when I tried on my m1 it was much quicker #asahi-dev · 2026-10-03 08:57
svenwhich also means the suspend somehow failed #asahi-dev · 2026-10-03 08:56
svenwhen I looked into it and failed I noticed it always rebooted after 5 or so seconds #asahi-dev · 2026-10-03 08:56
svenmight even be the normal one #asahi-dev · 2026-10-03 08:56
svenyeah, there’s some watchdog somewhere #asahi-dev · 2026-10-03 07:11
svenofc you are in the history for those files, just use git log --follow#asahi-dev · 2026-10-02 19:25
svenwould be nice to at least get that in#asahi-dev · 2026-10-02 19:25
sveni kinda hope i'll be able to finish at least the typec subsystem part and get that out tomorrow#asahi-dev · 2026-10-02 19:25
sventhere's probably still enough time if you send it out sometime next week for this cycle, but there's no rush. there's always the next one ;)#asahi-dev · 2026-10-02 19:21
sveni (or you) can send another smaller PR based on the first one #asahi-dev · 2026-10-02 19:21
sveni'm just going on vacation next week and figured i'd already send out whatever we have until now :D#asahi-dev · 2026-10-02 19:00
svenyeah, my m4 pro only has three dcpext nodes#asahi-dev · 2026-10-02 18:57
sven*my#asahi-dev · 2026-10-02 18:57
svenif you give me a MMIO addreess i can try and trace them on me m4 pro tomorrow#asahi-dev · 2026-10-01 06:14
sveniirc 24.x was still broken #asahi-dev · 2026-10-01 06:14
svenUpdate the target volume to the latest 26#asahi-dev · 2026-09-30 09:17
svenSee my open PR for what probably are the EL1 ones we should use #asahi-dev · 2026-09-30 09:16
svenoh, yeah, that uses the EL2 variant of those regs #asahi-dev · 2026-09-30 09:15
svenhuh, that would be hvc #0 though which I don’t think I actually use #asahi-dev · 2026-09-30 09:13
svenit probably really shouldn’t be enabled unconditionally based on locked sysregs#asahi-dev · 2026-09-30 09:13
svenI’d blame the sprr emu doing something dumb then #asahi-dev · 2026-09-30 09:12
svenis that an unhandled hvc?#asahi-dev · 2026-09-30 08:58
svenbut that also shouldn’t break anything #asahi-dev · 2026-09-30 08:58
svenI think right now I enable sprr/gxf emulation by default whenever sysregs are locked. maybe that should just be a flag instead #asahi-dev · 2026-09-30 08:57
svenhuh#asahi-dev · 2026-09-30 08:28
svenalso tested sptm on 27.x beta on my old m1 so presumably that will work on m4 pro as well #asahi-dev · 2026-09-30 08:28
svenI did hv bringing with 26.7 on m4 pro so that should work #asahi-dev · 2026-09-30 07:07
svenm4 pro works for me, I have NO_DISPALY in config.h but it should work even without that #asahi-dev · 2026-09-29 18:43
svennice, gonna try tomorrow then#asahi-dev · 2026-09-29 18:02
svenwhile the CTRR are just 0x8000000000000001#asahi-dev · 2026-09-29 18:02
sventhat's how i e.g. have to init one of them to make sptm happy#asahi-dev · 2026-09-29 18:02
svensptm_cpus[i].shadow[HV_SPTM_ACC_CTXR_A_CTL_EL2] = 0xc000000000aa019a;#asahi-dev · 2026-09-29 18:02
sveni don't know, i just know their names but the control register is more complicated#asahi-dev · 2026-09-29 18:01
sventhey have the same values on m4 pro so it should work#asahi-dev · 2026-09-29 18:00
svencan you try using the _EL1 variant and see if that still works?#asahi-dev · 2026-09-29 17:59
svenyuka: can't test linux boot yet because i don't have device trees for M4 but I found the names for the weird CTRR registers in SPTM and I think also their _EL1 variant https://github.com/AsahiLinux/m1n1/pull/688/changes#asahi-dev · 2026-09-28 15:37
svenfeels like i'm hitting even more weird timing edge cases than before on this m4 pro compared to my old m1#asahi-dev · 2026-09-28 15:35
svenin other cursed news, i occasionally see hv rendezvous failures now. wasn't that smp issue bad enough? :(#asahi-dev · 2026-09-28 08:32
svenhuh, odd#asahi-dev · 2026-09-28 08:28
svenhttps://github.com/AsahiLinux/m1n1/blob/c42cf43d03887208e70ec5dcc4da0b2af3699c54/src/chickens.c#L228 something weird like that?#asahi-dev · 2026-09-27 20:49
svenpossibly, yeah#asahi-dev · 2026-09-27 20:48
svenhttps://github.com/AsahiLinux/m1n1/commit/fabe27e3f14891674a2fc42e9fbd2b51542230cd https://github.com/AsahiLinux/m1n1/commit/8fa422c0d3c11b3cc09e758c80d88ad3f5bf9a6a that sounds similar, so i'm probably just missing some new carveouts or whatever that have to be mapped differently#asahi-dev · 2026-09-27 20:35
sventhat just leaves that weird AMCC panic then#asahi-dev · 2026-09-27 20:32
svenxnu thinks the cluster is offline while it's actually stuck in that wfi loop and then tries to bring it back up which can't happen because it's been powered on the entire time :D