#asahi · 2021-02-06 07:54
marcanthe initramfs is optional of course#asahi · 2021-02-06 07:53
marcanlet me know how it works :)#asahi · 2021-02-06 07:53
marcanthere's a mach-o trick (and iBoot loading quirk) used to make this work without touching the mach-o headers, just with a simple cat#asahi · 2021-02-06 07:53
marcanthe kernel & initramfs should be compressed (because otherwise m1n1 can't tell where they end)#asahi · 2021-02-06 07:52
marcanbasically just `cat m1n1.macho kernel dtb initramfs > m1n1-with-stuff.macho`#asahi · 2021-02-06 07:52
marcanusage here: https://github.com/AsahiLinux/docs/wiki/Developer-Quickstart#direct-payload-booting#asahi · 2021-02-06 07:52
marcanjust pushed m1n1 support for built-in payloads, for those without serial#asahi · 2021-02-05 17:20
marcanI'll test your patch tomorrow#asahi · 2021-02-05 17:20
marcananyway, sleep :)#asahi · 2021-02-05 17:19
marcanbut yeah#asahi · 2021-02-05 17:19
marcanin this case we wouldn't be able to do that case without saving the timers for this reason#asahi · 2021-02-05 17:17
marcaneven on VM->HV-EL0 userspace context switch?#asahi · 2021-02-05 17:16
marcanI guess you need to do that to context switch VMs anyway...#asahi · 2021-02-05 17:16
marcanwhich would mean when context switching to TGE=1 we need to take over the timers and save state (or at least mask them)#asahi · 2021-02-05 17:16
marcanthe timers confuse me now though, because... okay, so we have HYP and non-HYP timers, non-HYP do this automasking thing... but not with TGE=1... so now with TGE=1 the hypervisor "owns" all timers and has to deal with them?#asahi · 2021-02-05 17:14
marcanso how do you know which EL0 is which#asahi · 2021-02-05 17:14
marcanthe "problem" with ARM virtualization is that EL0 is used by *both* the level 2 HV *and* the guest#asahi · 2021-02-05 17:13
marcanTGE is so that guests can go EL0-EL1#asahi · 2021-02-05 17:10
marcan(except for places where I've already seem the CPU speculate stuff into madness, and had to add them)#asahi · 2021-02-05 17:10
marcanthis is why m1n1 is fun, it's so dumb and simple you get to experience the wonders if "no barriers, anywhere" :)#asahi · 2021-02-05 17:09
marcanmaz: that would make sense#asahi · 2021-02-05 17:09
marcanmodwizcode: good question...#asahi · 2021-02-05 17:08
marcanit's just interesting that isb kicked the fiq into working again#asahi · 2021-02-05 17:08
marcananyway, I'll look for that mask register#asahi · 2021-02-05 17:08
marcanit's just that the way that mask "stuck" when I flipped TGE seemed to be a quirk of the instruction pipeline#asahi · 2021-02-05 17:07
marcanoh in guest mode it masks itself#asahi · 2021-02-05 17:07
marcanthis is really sounding like those timers are wired deep into the instruction pipeline...#asahi · 2021-02-05 17:07
marcanneeds the isb though, running a bunch of code does nothing. so isb "kicks" something into re-evaluating that FIQ state in light of the new TGE setting#asahi · 2021-02-05 17:06
marcanturns out it sticks by accident. an isb makes it fire a FIQ again (and FIQ storm if not handled, as you would expect from TGE=1 mode)#asahi · 2021-02-05 17:05
marcanmaz: you know how I was talking earlier about the timers? and how guest timer FIQs get masked somewhere and only delivered once, and that mask sticks when you turn on TGE?#asahi · 2021-02-05 17:05
marcanoh this is a fun quirk#asahi · 2021-02-05 16:46
marcanactually let me try something silly...#asahi · 2021-02-05 16:46
marcanI think AIC really is just a global block#asahi · 2021-02-05 16:44
marcanwell, there is one per CPU, and one special address that maps to your CPU#asahi · 2021-02-05 16:44
marcanopticron: yeah, there is a single "reason" register that returns the lowest firing irq number when read, and simultaneously masks it#asahi · 2021-02-05 16:43
marcanI suspect they may be abusing the top address lines for this, though I'm not sure#asahi · 2021-02-05 16:43
marcanI wonder how that's done in hardware#asahi · 2021-02-05 16:43
marcanmodwizcode: this is why AIC knows the current CPU and displays it its registers#asahi · 2021-02-05 14:20
marcanbbl, dinner :)#asahi · 2021-02-05 14:19
marcanit's so beautifully simple#asahi · 2021-02-05 14:19
marcanI really like it#asahi · 2021-02-05 14:19
marcanyeah#asahi · 2021-02-05 14:19
marcanand I guess this is the first arm64 platform that can't do that in hardware#asahi · 2021-02-05 14:19
marcanlinux sends 7 different messages via IPIs#asahi · 2021-02-05 14:18
marcan:)#asahi · 2021-02-05 14:18
marcanyou need at least 8 targets#asahi · 2021-02-05 14:18
marcanI mean how would two true IPIs work#asahi · 2021-02-05 14:18
marcanit's two *per-cpu* IPIs#asahi · 2021-02-05 14:17
marcanbut clearly "self" is intended as self#asahi · 2021-02-05 14:17
marcanwhich is why they are really just two IPIs