#asahi-dev · 2021-11-08 10:54
marcanyeah, works fine if I switch to wfe mode, welp#asahi-dev · 2021-11-08 10:52
marcanmaybe it keeps retrying the same CPU?#asahi-dev · 2021-11-08 10:52
marcanand I guess that counts as something like "hey, I'm good to take an IRQ" and somehow it never times out#asahi-dev · 2021-11-08 10:51
marcanit happens because the CPUs are in a wfi loop#asahi-dev · 2021-11-08 10:51
marcanah lol, I figured it out#asahi-dev · 2021-11-08 10:33
marcanah, I think the 8 "delays" registers are actually config of some sort, let's see#asahi-dev · 2021-11-08 10:25
marcankinda, I was going to do pcie, then realized the change broke the HV#asahi-dev · 2021-11-08 10:20
marcanoh huh, it's not broadcasting now, hm#asahi-dev · 2021-11-08 10:08
marcanwonder if it breaks when IRQs get delivered to CPUs that the guest hasn't started yet? but the IRQ controller is supposed to time out and send them to all CPUs anyway...#asahi-dev · 2021-11-08 10:06
marcanoh, this is weird. doing the msr(s3_4_c15_c10_4, 0) thing at init actually breaks the hypervisor.#asahi-dev · 2021-11-06 04:30
marcan(also whatever SEP shenanigans are involved with OS identity)#asahi-dev · 2021-11-06 04:29
marcan(among other things because multiple OSes sharing a container would have to share firmware version, and who says all distros you have installed have the same kernel support for firmware? if you want to update you'd have to first make sure all your distros support the new version, which kinda sucks)#asahi-dev · 2021-11-06 04:29
marcanI still think there are reasons to use the native boot picker and do it this way ;)#asahi-dev · 2021-11-06 04:28
marcankettenis: if we do what I said of using multiple ESPs, then we don't really need nvram; each OS would be the default boot path in its ESP#asahi-dev · 2021-11-05 08:31
marcanpython script#asahi-dev · 2021-11-05 08:29
marcanthat's *why* we're using the whole u-boot/efi thing#asahi-dev · 2021-11-05 08:29
marcanyes, we won't provide grub either#asahi-dev · 2021-11-05 08:28
marcanfor Arch I will take care of that myself, as the reference distro#asahi-dev · 2021-11-05 08:28
marcanagraf: also we don't intend to own the kernel at all; if you want to be on our installer list you'd better be building kernels from usable trees for those images#asahi-dev · 2021-11-05 08:27
marcanagraf: one issue is that I want to be able to stream images directly to the target disk, but I already built myself a seekable/cached fd view over HTTP for the ipsw in the installer... so that will work nicely for that use case too#asahi-dev · 2021-11-05 08:13
marcanresizing things down later to make space for more stuff is more of a pain#asahi-dev · 2021-11-05 08:13
marcanit's important for me :p#asahi-dev · 2021-11-05 08:12
marcanthink e.g. planning to install a second distro#asahi-dev · 2021-11-05 08:12
marcanthis is actually important to do in the installer, because users need to be able to choose how much space to allocate to linux and how much space to leave unallocated for whatever reason#asahi-dev · 2021-11-05 08:11
marcan(replace resize2fs with the appropriate dance for the fs/volume management in use)#asahi-dev · 2021-11-05 08:11
marcanthe actual partition would be created as full size already by the installer#asahi-dev · 2021-11-05 08:11
marcannah, all they have to do is resize2fs#asahi-dev · 2021-11-05 08:10
marcanand I don't want static PARTUUIDs either#asahi-dev · 2021-11-05 08:10
marcanbut I don't want distros editing the GPT on first boot#asahi-dev · 2021-11-05 08:09
marcanI've used them for that#asahi-dev · 2021-11-05 08:09
marcanyes#asahi-dev · 2021-11-05 08:08
marcananyway, I would at least randomize the PARTUUIDs (or rather: diskutil will do that for me anyway)#asahi-dev · 2021-11-05 08:08
marcansure#asahi-dev · 2021-11-05 08:06
marcanyeah, *that* could work :)#asahi-dev · 2021-11-05 08:06
marcancloning UUIDs is a terrible idea#asahi-dev · 2021-11-05 08:06
marcanI would much, much prefer to have proper unique UUIDs#asahi-dev · 2021-11-05 08:06
marcansure, but once you use UUIDs now the device is unknown and conflicts are global#asahi-dev · 2021-11-05 08:05
marcanagraf: for dd-able images you can just forgo UUIDs and use device paths or whatever; that's typical on SBCs but won't work for us#asahi-dev · 2021-11-05 08:04
marcansven: yeah, go ahead and add t6k#asahi-dev · 2021-11-05 08:04
marcanpeople are going to want to use these things a lot more than they use SBCs, and we also need to assume less experienced users will be more common ;)#asahi-dev · 2021-11-05 08:03
marcanyes, and those are SBCs, not real computers#asahi-dev · 2021-11-05 08:03
marcanthat just sets users up for trouble in the future#asahi-dev · 2021-11-05 08:03
marcanbut yeah, I don't like hardcoded UUIDs#asahi-dev · 2021-11-05 08:03
marcanlocal dual-boot is important#asahi-dev · 2021-11-05 08:03
marcanbut that's an implementation detail#asahi-dev · 2021-11-05 08:03
marcananyway, I'm generally on board with the general flow, which isn't much different from my idea except I was thinking that instead of a disk image I'd use some json/whatever description of the partitions and individual files#asahi-dev · 2021-11-05 08:02
marcanthough I think for images like this usually there would be a startup script to do it on first boot#asahi-dev · 2021-11-05 08:01
marcanI mean whatever the distro rootfs partition is#asahi-dev · 2021-11-05 08:01
marcanbut the issue with copying the PARTUUID/GUID as-is is that it breaks with dual-boot, and gets really confusing if you have external disks with installs too#asahi-dev · 2021-11-05 08:00
marcanwe'll deal with u-boot, that's not a problem