#asahi · 2022-03-07 14:25
kettenis(which isn't the case for U-Boot yet for these Apple systems)#asahi · 2022-03-07 14:25
kettenisthe most notable feature right now is better address space randomization if the EFI random number generator protocol is supported#asahi · 2022-03-07 14:24
kettenismps: some arm64 linux kernel features are only offered by the EFISTUB#asahi · 2022-03-04 09:54
kettenisupdated u-boot with the latest device trees as well#asahi · 2022-03-02 23:08
kettenisupdated the device trees in u-boot to sync them with the linux repo#asahi · 2022-03-02 16:04
kettenisI have huge respect for the guy who keeps it running on OpenBSD#asahi · 2022-03-02 16:03
kettenisaxboe: and 95% of those ifdefs are unecessary if you just think a little bit harder and pick a solution that's actually portable#asahi · 2022-03-02 16:01
kettenisaxboe: these days multi-arch means Windows/Linux/macOS on x86 and Android on arm64 isn't it? ;)#asahi · 2022-03-01 14:51
kettenisat this point they're straight copies of the ones from asahi linux#asahi · 2022-03-01 14:51
kettenisI'll add them#asahi · 2022-03-01 14:50
ketteniscool, thanks!#asahi · 2022-03-01 14:50
kettenisdoes it also put the firmware tar file on the ESP?#asahi · 2022-02-28 22:41
kettenisnope, but I did start some discussions#asahi · 2022-02-28 13:02
kettenisit's worth a try#asahi · 2022-02-28 12:57
ketteniswhere it says # enable and route debug exceptions to EL2#asahi · 2022-02-28 12:57
kettenisSee proxyclient/m1n1/hv.py#asahi · 2022-02-28 12:56
kettenisin the end I disabled the access to MDSCR_EL1 in my OpenBSD kernel when I was testing it under the hypervisor#asahi · 2022-02-28 12:55
kettenisbut there was a worry that this would interfere with single-stepping in the hypervisor itself#asahi · 2022-02-28 12:54
kettenisIt was suggested that the HV could set a bit somewhere that makes it not trap when MDSCR_EL1 is accessed#asahi · 2022-02-28 12:53
kettenisI ran into the same issue with OpenBSD#asahi · 2022-02-27 16:18
kettenisu-boot-nodtb.bin isn't compressed, so it has to be last#asahi · 2022-02-27 16:17
kettenisand any uncompressed binary payload has to be last#asahi · 2022-02-27 10:13
kettenisjannau: the solution to that is to use a different compatible for the DARTs that need their mappings retained#asahi · 2022-02-26 14:46
kettenisthat's why I think running the normal OS installer makes most sense for distributions that provide one#asahi · 2022-02-26 14:13
kettenisthe issue isn't unique to Apple M1 systems; basically all arm64 SBCs have the same issue#asahi · 2022-02-26 14:12
ketteniswhat I'm trying to say is that there is a way to find this out and do the right thing instead of controlling this by passing stuff along from the install script#asahi · 2022-02-26 14:11
kettenisnot sure if the Linux kernel advertises the situation (not supporting persistent EFI variables) to userland#asahi · 2022-02-26 14:10
kettenisright#asahi · 2022-02-26 14:10
kettenisgrub itself, or grub-install running in the kernel?#asahi · 2022-02-26 14:09
kettenisit supports them at boot time#asahi · 2022-02-26 14:08
kettenisGlanzmann: U-Boot's EFI implementation does indicate that runtime support for EFI variables isn't there#asahi · 2022-02-26 13:19
kettenisheh, to me a UUID is an 128-bit integer that is insanely formatted such that everybody gets its endianness wrong ;)#asahi · 2022-02-26 13:12
kettenisDid I see correctly that you set the volume ID in the VFAT filesystem instead of the UUID in the GPT?#asahi · 2022-02-25 16:49
kettenisit'd be great if somebody using linux could test this#asahi · 2022-02-25 16:49
kettenisthis implements marcan's suggestion to keep defaulting to serial when running under the hypervisor#asahi · 2022-02-25 16:48
kettenis(which will always be true on laptops)#asahi · 2022-02-25 16:48
kettenisthat is my attempt to make u-boot set the default output path for the OS to the framebuffer if a keyboard is connected to the machine#asahi · 2022-02-25 16:48
kettenishttps://github.com/kettenis/u-boot/tree/framebuffer#asahi · 2022-02-22 18:12
kettenisI still think I prefer the "install u-boot using the Asahi installer and pretend this is a normal machine afterwards" approach#asahi · 2022-02-22 18:08
kettenisthe real question is how we'd handle the maintenance burden of updating the installer for each new release#asahi · 2022-02-22 18:07
kettenisFreeBSD an NetBSD do things in a different way I think#asahi · 2022-02-22 18:06
ketteniswhat may make sense is to copy bsd.rd and BOOTAA64.EFI into the ESP#asahi · 2022-02-22 18:05
kettenisNot keen on that#asahi · 2022-02-22 18:05
kettenisoh, you mean an OpenBSD rootfs?#asahi · 2022-02-22 18:01
kettenishow would a filesystem image of some sort help you?#asahi · 2022-02-22 18:00
kettenisthat doesn't make sense#asahi · 2022-02-22 17:57
kettenisso as soon as you push your current work I'll have a closer look#asahi · 2022-02-22 17:56
kettenisTheo wants to look into copying the wifi firmware into the installer from the ESP such that we can have working wifi in the OpenBSD installer#asahi · 2022-02-22 17:53
kettenisah, right#asahi · 2022-02-22 17:53
kettenissorry, I lost you, what asahi-specific bits are you thinking of?