#asahi-re · 2026-01-26 21:22
sveni've tried to look at iboot before and it's just painful#asahi-re · 2026-01-26 21:22
svenyeah, i think this is worth a try#asahi-re · 2026-01-26 21:22
amarioguybeing straight up here, my binary RE of iBoot for the register was hitting a dead-end personally so i think that's gonna be a more productive approach than bashing my head for hours trying to figure out what some register maps to#asahi-re · 2026-01-26 21:20
sventhat was my point :D#asahi-re · 2026-01-26 21:20
svenyeah#asahi-re · 2026-01-26 21:20
amarioguywe only care about read here, because we can always test such an MSR on an m3 machine or smth#asahi-re · 2026-01-26 21:19
svenbut worth a try anyway#asahi-re · 2026-01-26 21:19
svenmight just be that the msr that disables everything also disappears itself :D#asahi-re · 2026-01-26 21:19
svenbut i think they could all be read on M1/M2 at least#asahi-re · 2026-01-26 21:19
svenhm.. i know that certain msrs can't be _written_ to anymore once GL2 is enabled from EL2#asahi-re · 2026-01-26 21:15
amarioguyif we're pretty sure it's an MSR (which i don't want to rule out) then that might be a way to go about it#asahi-re · 2026-01-26 21:14
amarioguywe can at least infer the state difference if it is MSR bound or not#asahi-re · 2026-01-26 21:13
amarioguymostly bringing this up because even if tracing isn't gonna be possible...we can just dump all MSRs can we?#asahi-re · 2026-01-26 21:13
amarioguymaking m1n1 boot as Mach-O again#asahi-re · 2026-01-26 21:12
amarioguyi think one thing we should try doing#asahi-re · 2026-01-26 21:12
amarioguythinking about it a bit more#asahi-re · 2026-01-26 21:12
amarioguyhmmm#asahi-re · 2026-01-22 14:42
chaos_princessit is embedded in the kext#asahi-re · 2026-01-22 14:39
alex0(ignore that lmao, wanted to disable trydactyl but failed lmao), hello there, just wanted to ask where i can find the avd firmware. i found out this project https://github.com/eiln/avd so i curious about it #asahi-re · 2026-01-22 14:38
Shiz:(#asahi-re · 2026-01-22 14:37
alex0:mode ignore#asahi-re · 2026-01-18 01:07
amarioguymakes so many function flows hard to follow...even if ik it's probably better speed wise#asahi-re · 2026-01-18 01:06
amarioguyfunction outlining is the *worst*#asahi-re · 2026-01-15 16:34
sptmworld over: arrogant pricks with personality disorders.#asahi-re · 2026-01-15 16:34
sptmbut way to go -- helping reinforce the stereotype of german open source programmers the#asahi-re · 2026-01-15 16:34
sptmmpt to demonstrate something useful, for the sole purpose of helping out#asahi-re · 2026-01-15 16:34
sptmand yes, anonymous, for my own reasons, and also because the only motivation was to atte#asahi-re · 2026-01-15 16:34
sptmnspecting it just like the one that exists today#asahi-re · 2026-01-15 16:34
sptmmpilation. there is nothing wrong with using this technique on your own m1n1 build and i#asahi-re · 2026-01-15 16:34
sptmhell, the eu even explicitly have a section about this in their legislation, titled deco#asahi-re · 2026-01-15 16:34
sptmalso, I'm pretty sure every civilised country on earth lets you build tools incorporating bits that were reverse engineered, to inspect other tools to produce an end product that is not contaminated#asahi-re · 2026-01-15 16:34
sptme path" but instead something somoene could iterate on easily#asahi-re · 2026-01-15 16:34
sptmwriting example code, and then having it in a form that is BUILDABLE in m1n1 as an example for people working on it so that it was not just a "hey maybe this might be a workabl#asahi-re · 2026-01-15 16:34
sptmand, "slop", I put at least 24 hours of my free time into trying to understand the system, documenting what I found#asahi-re · 2026-01-15 16:33
sptmoducts, I just sometimes have to use them)#asahi-re · 2026-01-15 16:33
sptmsven: it was untested because I don't have a spare M4 lying around (I don't buy apple pr#asahi-re · 2026-01-15 08:17
svenalso reads like asm2c which is precisely what we need to avoid#asahi-re · 2026-01-15 08:16
svenfl0_id: nah, you're right. it's essentially **untested** AI slop by an anonymous contributor without even a nickname#asahi-re · 2026-01-15 08:05
svenhrm, i'd really expect this to be a msr since it disables something inside the cpu and that's the easiest place to hook up that up#asahi-re · 2026-01-15 00:34
amarioguytho that callback soup still worries me a bit...#asahi-re · 2026-01-15 00:31
amarioguychaos_princess: in case you were following along, function that i found this in is 0x10081AF2F64, offset 0x4B4#asahi-re · 2026-01-15 00:30
amarioguy0x2100F8100#asahi-re · 2026-01-15 00:30
amarioguylooks like there's this register that seems to be in the same space as the RVBAR registers#asahi-re · 2026-01-15 00:30
amarioguyok ok, so after doing some more digging the past day or so#asahi-re · 2026-01-15 00:27
amarioguyi'm thinking the presence of some of the segments (that are not present in the raw image case) are what signal a mach-o/sptm compatible boot object vs a raw one#asahi-re · 2026-01-15 00:27
amarioguyat least not immediate#asahi-re · 2026-01-15 00:27
amarioguychaos_princess, nnym: just checked both LPs, yea doesn't seem to be anything differentiating it#asahi-re · 2026-01-15 00:11
fl0_idsorry, glad I am wrong then#asahi-re · 2026-01-14 23:54
sptmand the XN page, fault, replace GENTER with another faulting instruction, then trap there and do the work SPTM would've done (all inputs are in the registers) -- is this invalid?#asahi-re · 2026-01-14 23:52
sptmi spent 12 hours in ghidra, but yes I got it to write up my notes