#asahi-dev · 2021-07-18 14:25
marcanand the fact that there is a whole driver kext for *M1* silicon that does not use DCP suggests that was certainly in the cards at some point#asahi-dev · 2021-07-18 14:25
marcanalyssa: we do not#asahi-dev · 2021-07-12 16:59
marcanframebuffer is one that would be obvious if it weren't coherent#asahi-dev · 2021-07-12 16:58
marcanyeah, I didn't find any noncoherence#asahi-dev · 2021-07-12 16:29
marcan*even more#asahi-dev · 2021-07-12 16:29
marcan(and they also support it in kernelspace because not doing so would be even weird and architecture-breaking and bizarre)#asahi-dev · 2021-07-12 16:29
marcanthey only care about 16K systemwide, except they needed 4K in userspace in the CPU for Rosetta. that's it.#asahi-dev · 2021-07-12 16:29
marcanthe 4K/16K page thing is how it is because it's exactly what they need#asahi-dev · 2021-07-12 16:28
marcanif asahi gets big enough that people rely on it, I expect apple to tacitly try not to break us too badly for no good reason#asahi-dev · 2021-07-12 16:27
marcanthey clearly aren't trying to trip us up#asahi-dev · 2021-07-12 16:27
marcanmacOS updates so far have made our lives *easier*, not *harder:#asahi-dev · 2021-07-12 16:26
marcanand I'm fine with that#asahi-dev · 2021-07-12 16:26
marcan"here's code execution, have fun" is their take#asahi-dev · 2021-07-12 16:26
marcanasahi isn't something they want to hinder at all#asahi-dev · 2021-07-12 16:26
marcanbut they also aren't going to spend engineering effort into helping things that only we care about#asahi-dev · 2021-07-12 16:26
marcanthey built the entire bootpolicy system so things like asahi can exist#asahi-dev · 2021-07-12 16:25
marcanchadmed: apple isn't deliberately breaking us in any way#asahi-dev · 2021-06-22 08:22
marcanlater I think all it does is invals#asahi-dev · 2021-06-22 08:22
marcansven: during init only#asahi-dev · 2021-06-22 01:06
marcanI do see osx poking registers we don't know about#asahi-dev · 2021-06-22 01:06
marcansven: I don't think I saw anything weird yet#asahi-dev · 2021-06-21 17:42
marcanbloom: nn :p#asahi-dev · 2021-06-21 17:42
marcansee also https://github.com/AsahiLinux/m1n1/blob/main/tools/apple_regs.json everything that mentions "pointer authentication"#asahi-dev · 2021-06-21 17:40
marcanapple documented this, then deleted the file in 11.2+, because of course they didn't *mean* to document it...#asahi-dev · 2021-06-21 17:40
marcan"Key generation on Apple CPUs"#asahi-dev · 2021-06-21 17:40
marcanc1truz[m]: https://opensource.apple.com/source/xnu/xnu-7195.50.7.100.1/doc/pac.md#asahi-dev · 2021-06-21 17:38
marcanprobably more things#asahi-dev · 2021-06-21 17:38
marcanthey support extra kernel keys, per-boot diversification, and they also did the whole EL1/EL2/EL12 VHE thing with the entire register set (standard ones and their add-ons)#asahi-dev · 2021-06-21 15:17
marcanthe ones that don't need massive HV changes but do require Apple hardware (and the right bits flipped by the HV)#asahi-dev · 2021-06-21 15:17
marcanthat's surprising, I imagined they would at least still want AMX/Apple PAC and things like that#asahi-dev · 2021-06-21 13:53
marcanc1truz[m]: nice! I assume that one doesn't need GXF/SPRR? I can't imagine Linux playing nicely with that :)#asahi-dev · 2021-06-20 05:38
marcanc1truz[m]: I don't think many of us are testing desktop environments at this stage (not much point until GPU drivers start to exist), but that sounds like a bug in xorg#asahi-dev · 2021-06-19 00:38
marcanc1truz[m]: \o/#asahi-dev · 2021-06-12 14:26
marcanof course it is, that's where all the vdmtool stuff is implemented :p#asahi-dev · 2021-06-12 12:33
marcanjannau: yes, there are definitely differences#asahi-dev · 2021-06-10 12:51
marcanthough there are probably good reasons to move more dev talk here :)#asahi-dev · 2021-05-31 02:36
marcanbloom: done :)#asahi-dev · 2021-05-31 02:11
marcansaw it, let me wake up :>#asahi-dev · 2021-05-31 01:31
marcanif so then I guess it is the kernel which sets this up#asahi-dev · 2021-05-31 01:31
marcanwould be interesting to see if removing that makes it work under the HV#asahi-dev · 2021-05-31 01:31
marcanthis is probably wrong, I bet there are EL1 regs I'm not forwarding to#asahi-dev · 2021-05-31 01:31
marcanhacr.TRAP_CTRR = 1#asahi-dev · 2021-05-31 01:31
marcanthough, actually#asahi-dev · 2021-05-31 01:30
marcan*does work#asahi-dev · 2021-05-31 01:30
marcanwhat I find weird is that m1n1 chainloading doesn't work; that suggests CTRR has the addresses locked but isn't enforced yet#asahi-dev · 2021-05-31 01:30
marcanI mean that's *why* it fails under the HV, the registers are already set up#asahi-dev · 2021-05-31 01:29
marcanand for CTRR it should be iBoot#asahi-dev · 2021-05-31 01:29
marcanCTRR is a newer version of KTRR, right?#asahi-dev · 2021-05-30 18:38
marcanEmantor: nah, the point of ctrr is iBoot should lock it down already#asahi-dev · 2021-05-30 17:34
marcanalso heck, how does chainloading even work with ctrr?