#asahi-dev · 2024-10-28 19:52
kettenisso we get an m4 imac but nothing else yet?#asahi-dev · 2024-08-30 11:12
kettenisOpenBSD driver does that too, so that shouldn't be an issue either#asahi-dev · 2024-08-30 11:11
kettenisah, yes, 32-bit write for output characters as well#asahi-dev · 2024-08-30 10:27
kettenisThe OpenBSD driver already does that so I expect no issues#asahi-dev · 2024-08-30 10:26
kettenisnickchan: the main (only?) difference that seems to make is that it'll use a 32-bit read to get the input characters#asahi-dev · 2024-08-06 15:35
kettenisright, make upstreaming even harder since now you have rust bindings to bikeshed about ;)#asahi-dev · 2024-07-14 15:29
kettenisright#asahi-dev · 2024-07-14 15:25
kettenisguess it should only matter for the first-stage m1n1#asahi-dev · 2024-07-14 15:24
kettenisah, but we support both and some of us may still be in the habit of using macho?#asahi-dev · 2024-06-25 12:15
kettenisor rather 833#asahi-dev · 2024-06-25 12:14
ketteniserh, 56-bit wraps after 776 days according to my calculation#asahi-dev · 2024-06-25 12:10
kettenisand the counter still shouldn't overflow in the lifetime of the hardware, so why not?#asahi-dev · 2024-06-25 12:09
kettenis1ns resolution is useful to have when micro-benchmarking#asahi-dev · 2024-06-25 12:08
kettenisbox ticking?#asahi-dev · 2024-06-25 12:07
kettenischadmed: ARM requires the system timer to run at 1GHz in ARMv8.6 and later#asahi-dev · 2024-06-19 16:00
kettenisa few extra bytes are not going to hurt on this hardware#asahi-dev · 2024-06-19 16:00
kettenissorry, but such "optimizations" are pretty pointless#asahi-dev · 2024-06-19 09:40
kettenisSo implementing a UCLASS_RNG driver for SEP might be worth the effort#asahi-dev · 2024-06-19 09:36
kettenisImplementing EFI_RNG_PROTOCOL would be good since it would allow the Linux kernel efistub to randomize the physical address of the kernel#asahi-dev · 2024-06-19 09:33
ketteniszzywysm: kaslr-seed is set by m1n1 and makes it through to the kernel#asahi-dev · 2024-05-31 20:54
kettenisyou probably could do this with a usb ethernet device#asahi-dev · 2024-05-31 20:47
kettenisnot sure how useful it would be, given that we thethered boot support over USB#asahi-dev · 2024-05-31 20:46
kettenisjannau: no I didn't (if I was aware that such a driver existed I defenitely forgot about it)#asahi-dev · 2024-05-14 10:20
kettenisSME without SVE seems really weird#asahi-dev · 2024-04-29 10:23
kettenis?#asahi-dev · 2024-04-29 10:23
kettenis** Booting bootflow '<NULL>' with efi_mgr #asahi-dev · 2024-04-29 10:22
kettenisthat is indeed strange; what is the actual bootflow that gets selected?#asahi-dev · 2024-04-29 10:11
kettenishmm, it isn't clear to me what problems you folks are seeing with the new u-boot boot flow stuff#asahi-dev · 2024-04-12 05:10
kettenisjannau: been a bit busy, but I'll get to it in the coming days#asahi-dev · 2024-04-07 08:12
kettenisjanneg: I pushed a change to the asahi branch that should make us boot from the designated EFI system partition for the autogenerated EFI boot options#asahi-dev · 2024-04-04 11:16
ketteniswhich isn't explicitly forbidden by the UEFI standard#asahi-dev · 2024-04-04 11:16
kettenisbut yes, the most important thing for asahi is the multiple-ESP thing#asahi-dev · 2024-04-04 11:13
kettenis(there are other differences between those tow though)#asahi-dev · 2024-04-04 11:13
ketteniswe neuter the DTB loading for apple m1 so to some extent bootmeth_efi_mgr and bootmeth_efi are equivalent for us#asahi-dev · 2024-04-04 09:36
kettenisyou'd think that would be the sensible thing to do, but the current upstream implementation just iterates over the partitions by number#asahi-dev · 2024-04-04 09:03
kettenisbut I still need to do a bit more testing before I submit that upstream#asahi-dev · 2024-04-04 09:02
kettenisI have a diff to make the efi boot manager code prefer booting from the designated EFI partition instead of just any partition#asahi-dev · 2024-04-04 08:58
kettenisI dropped the code that would seed the Boot#### and BootOrder variables#asahi-dev · 2024-04-04 08:53
kettenisjanneg: now where did I store that fixup...#asahi-dev · 2024-03-29 14:44
kettenisyou mean if you upgrade to macOS 14 firmware the ascwrap-v4s become ascwrap-v6s?#asahi-dev · 2024-03-29 14:18
kettenisbut that might cause isses further down the line when we actually try to shutdown clusters#asahi-dev · 2024-03-29 14:17
kettenisI soppose m1n1 could repair this by starting the first e-core and parking the p-core#asahi-dev · 2024-03-29 14:07
kettenisthe DT doesn't really have CPU numbers; just the MPIDR register values#asahi-dev · 2024-03-29 14:02
kettenisother arm64 hardware that is#asahi-dev · 2024-03-29 14:02
kettenisjanneg: I'm pretty sure there is other hardware that boots on the CPU that isn't the first one listed in the device tree#asahi-dev · 2024-03-17 23:16
kettenisthe JIT pages don't have the magic bit set that enforces BTI#asahi-dev · 2024-03-17 23:16
kettenisah, I think I know why the BTI llvmpipe issue doesn't happen with Linux#asahi-dev · 2024-03-17 21:02
kettenisand since the efi boot variables pre-seeding isn't upstream yet, that's fine#asahi-dev · 2024-03-17 21:00
kettenisyeah, I saw it#asahi-dev · 2024-03-17 20:53
kettenisme neither until some other OpenBSD user ran glxgears on an M2 laptop ;)