#
✕ nick:kettenis
#asahi · 2021-12-06 18:01
ketteniswhat tree should I generate that against?
#asahi · 2021-12-06 18:01
kettenismarcan: I have a small diff for the wifi mac address
#asahi · 2021-12-06 17:34
kettenis#1 would be where I would start
#asahi · 2021-12-06 17:33
kettenis2: use the m1n1 hv to run macOS and see what it does
#asahi · 2021-12-06 17:32
kettenis1: use m1n1 to poke at the hardware and see if you can figure out how it works
#asahi · 2021-12-06 17:32
kettenisjmr2: there are basically two approaches
#asahi · 2021-12-06 13:30
kettenisso that could be used to make multiple ESPs work
#asahi · 2021-12-06 13:30
kettenisnot sure how GRUB uses that information, but in principle it could use this to look up the right configuration file
#asahi · 2021-12-06 13:29
kettenismarcan: EFI applications (such as GRUB) actually get told the partition they were started from
#asahi · 2021-12-03 11:04
kettenisif you search for threads with apple in the subject you pretty much catch everything that is relevant
#asahi · 2021-12-03 11:04
kettenisand of course the earlier months
#asahi · 2021-12-03 11:02
kettenishttp://lists.infradead.org/pipermail/linux-arm-kernel/2021-December/thread.html
#asahi · 2021-12-03 10:59
kettenisno, we just focus on getting stuff done instead on reporting ;)
#asahi · 2021-12-02 15:25
kettenisexactly; it is more like a BIOS update ;)
#asahi · 2021-12-02 12:17
kettenisthe hard reality is that it probably isn't very useful to test something that deviates significantly from that branch
#asahi · 2021-12-02 11:09
kettenismarcan: heh
#asahi · 2021-12-02 11:09
kettenisor change the device tree to match what the corellium drivers use
#asahi · 2021-12-02 11:08
kettenisyou can either change your kernel to use the ones in "my" device tree
#asahi · 2021-12-02 11:08
kettenisbecause the correllium compatibles make no sense
#asahi · 2021-12-02 11:08
kettenisthe spi and keyboard bindings in that tree are what I currently use in openbsd
#asahi · 2021-12-02 11:06
kettenisI want to change it eventually, but I'm waiting on sven to settle on the rtkit/sart/nvme bindings ;)
#asahi · 2021-12-02 11:04
kettenismarcan: well, OpenBSD currently relies on u-boot starting the corporcessor
#asahi · 2021-12-02 11:02
kettenisI think sven is working on making that a little bit more robust
#asahi · 2021-12-02 11:02
ketteniss/driver/drive/
#asahi · 2021-12-02 11:01
kettenisthe kernel might not see the nvme driver because u-boot already initialized the coprocessor
#asahi · 2021-12-02 11:01
kettenisnot sure the "legacy" boot path works; I've only tested the EFI boot path
#asahi · 2021-12-02 10:45
kettenismps: does the nvme disk now show up in u-boot?
#asahi · 2021-12-02 10:44
kettenisinstall a cross-compiler and make sure you set the CROSS_COMPILE environment variable properly
#asahi · 2021-12-01 21:56
kettenisthe changes I just pushed to my apple-m1-m1n1-nvme u-boot branch may fix nvme with the macos 12 firmware
#asahi · 2021-12-01 19:22
kettenisbut yes, that's what I was thinking of
#asahi · 2021-12-01 19:22
kettenisalso need the syscon node that sits on top of the power domains
#asahi · 2021-12-01 19:18
kettenisyup, the problem is that u-boot sees the power-domain property, tries to find a driver, fails and then ends up without a serial console
#asahi · 2021-12-01 19:15
kettenisbut that is probably a bigger mainenance burden
#asahi · 2021-12-01 19:14
kettenisalternatively, for these u-boot,dm-pre-reloc properties we could write code that add them
#asahi · 2021-12-01 19:13
kettenisu-boot solves this by having *-u-boot.dtsi files that get included automatically
#asahi · 2021-12-01 19:09
kettenisif we store the firmware on the ESP our firmware update tool should be able to grab it from there
#asahi · 2021-12-01 19:09
ketteniswill be fine
#asahi · 2021-12-01 19:08
kettenisheh
#asahi · 2021-12-01 19:07
kettenisthis is needed to make sure the power domains needed for the serial console can be found in the early "pre-relocation" phase of u-boot
#asahi · 2021-12-01 19:06
kettenismarcan: for u-boot we need to annotate a few nodes in the device tree with a "u-boot,dm-pre-reloc" proprty
#asahi · 2021-12-01 17:07
kettenisthere really shouldn't be a need to "port" debian-installer
#asahi · 2021-12-01 15:20
kettenisno, it means I should fix the code
#asahi · 2021-12-01 15:07
kettenisthat may explain why nvme isn't working for you
#asahi · 2021-12-01 14:57
kettenisare you using firmware from macos 12.0?
#asahi · 2021-11-30 21:50
kettenisdoesn't really matter; everything that needs to be built has been built at that point
#asahi · 2021-11-30 14:00
kettenisno reason to believe it shouldn't work on the air
#asahi · 2021-11-30 13:59
kettenisu-boot works on 13" macbook pro
#asahi · 2021-11-29 17:36
kettenisthat's pretty good I'd say
#asahi · 2021-11-28 19:45
ketteniss/synches/synched
#asahi · 2021-11-28 19:44
kettenissynches the device trees on the apple-m1-m1n1-nvme branch of my u-boot repo with marcan's t6000/bringup-work branch

plus ›