#
✕ nick:marcan
#asahi · 2021-05-08 17:13
marcanbut the actual faulting instruction there is a null deref
#asahi · 2021-05-08 17:10
marcan"Invalid kernel stack pointer" is supposedly a stack pointer issue though
#asahi · 2021-05-08 17:07
marcanso it might be a legitimate PAC'ed LR
#asahi · 2021-05-08 17:07
marcanPAC codes go in the upper bits
#asahi · 2021-05-08 17:07
marcanI know
#asahi · 2021-05-08 17:01
marcanah, maybe this is a PAC thing; it's entirely possible the PAC registers are VHE'd and I'm passing through to the wrong ones
#asahi · 2021-05-08 17:01
marcanInvalid kernel stack pointer (probable corruption). at pc 0xfffffe00113dd374, lr 0xc1ff7e00112837e8 (saved state: 0xfffffe0015083cb0)
#asahi · 2021-05-08 16:32
marcanah, it only uses stack args apparently
#asahi · 2021-05-08 16:27
marcanxnu's va_list doesn't seem to match the arm standard one... :/
#asahi · 2021-05-08 16:19
marcanI'm writing a panic decoder in python now
#asahi · 2021-05-08 16:19
marcanI already have the panic hooked because it triggers a debugger trap and I have all exceptions hooked
#asahi · 2021-05-08 16:19
marcanyrlf: there is no framebuffer yet either
#asahi · 2021-05-08 15:56
marcanso macos is panicing now... trying to figure out how to decode the panic from the hypervisor, since serial isn't initialized yet
#asahi · 2021-05-08 14:17
marcanbefore any stack accesses
#asahi · 2021-05-08 14:17
marcansetting pan to 0 as the first instruction of the EL2 handlers fixes it
#asahi · 2021-05-08 14:17
marcanso everything explodes
#asahi · 2021-05-08 14:17
marcanand since m1n1 maps everything as EL0-accessible, EL2 can't access any data, not even the stack
#asahi · 2021-05-08 14:16
marcanthe bit isn't cleared when entering EL2
#asahi · 2021-05-08 14:16
marcanj`ey: EL2 faults
#asahi · 2021-05-08 13:59
marcandidn't realize PAN leaks up to EL2
#asahi · 2021-05-08 13:59
marcanit was PAN.
#asahi · 2021-05-08 12:42
marcanah no, I think that's just the USB buffering
#asahi · 2021-05-08 11:18
marcanwell that's weird on a skip...
#asahi · 2021-05-08 11:18
marcanSkip: msr s3_5_c15_c4_0, x13 = d
#asahi · 2021-05-08 11:18
marcanTTY> hv_exc_sync
#asahi · 2021-05-08 11:18
marcanPass: mrs x13, s3_5_c15_c4_0 = d
#asahi · 2021-05-08 11:18
marcanTTY> ret
#asahi · 2021-05-08 11:18
marcanTTY> hv_exc_sync
#asahi · 2021-05-08 11:18
marcanit's not the guest borking me, it's a bug somewhere
#asahi · 2021-05-08 11:18
marcanthe hv isn't even returning from the last exception
#asahi · 2021-05-08 11:18
marcanduh
#asahi · 2021-05-08 09:01
marcansven: yeah, seems they don't :(
#asahi · 2021-05-08 08:57
marcanwell, let's find out
#asahi · 2021-05-08 08:57
marcan(like I did)
#asahi · 2021-05-08 08:57
marcansven: what if I set a bunch of HACR_EL2 bits?
#asahi · 2021-05-08 08:57
marcanwhen you're giving the guest access to 95% of the hardware directly, having a massive OS in the way without being able to use most of its features just increases the bug/problem surface
#asahi · 2021-05-08 08:57
marcanit's possible, I just don't really see the point :)
#asahi · 2021-05-08 08:55
marcanyou lose any benefits you'd get from doing this under linux
#asahi · 2021-05-08 08:55
marcan... so you're loading a big blob of linux and qemu and everything into a tmpfs... without any network, storage, or anything... what's the point then?
#asahi · 2021-05-08 08:55
marcanm1n1 does
#asahi · 2021-05-08 08:55
marcanlinux also doesn't have a polling driver for usb
#asahi · 2021-05-08 08:55
marcanpresuambly you want a filesystem
#asahi · 2021-05-08 08:55
marcanfor serial, yes
#asahi · 2021-05-08 08:54
marcanif not nvme, then USB
#asahi · 2021-05-08 08:54
marcanpresumably you want linux to have some kind of working storage
#asahi · 2021-05-08 08:54
marcanwhat are you going to do, patch the nvme driver etc to use polling instead of IRQs?
#asahi · 2021-05-08 08:54
marcanif you're going to cripple Linux to this point, I fail to see what that buys you over just using the HV in m1n1
#asahi · 2021-05-08 08:53
marcanit needs AIC if you want any non-polled IO
#asahi · 2021-05-08 08:53
marcancoexistence is unlikely to work
#asahi · 2021-05-08 08:53
marcanlinux needs AIC

plus ›