#
✕ nick:kettenis
#asahi-dev · 2022-02-08 20:15
kettenisheh, I overlooked that as well
#asahi-dev · 2022-02-08 13:12
kettenisheh, yes, I need to write something up
#asahi-dev · 2022-02-08 13:12
kettenislet me see if I can join there
#asahi-dev · 2022-02-08 13:10
kettenisarnd: thanks, I'll have a look at that
#asahi-dev · 2022-02-08 13:07
kettenisso EFI already provides a way to mark resident code in its memory map
#asahi-dev · 2022-02-08 13:03
ketteniswe just have to make sure the assumptions on things like the MMU state and interrupts match what ardb was thinking about for the UEFI interface
#asahi-dev · 2022-02-08 13:02
kettenisright
#asahi-dev · 2022-02-08 13:02
kettenishaving m1n1 scribble something into the device tree and making u-boot turn that into the appropriate EFI table shouldn't be too difficult
#asahi-dev · 2022-02-08 13:00
kettenisand you'd lose the functionality if you booted straight from m1n1
#asahi-dev · 2022-02-08 12:56
kettenisbut I think you don't want to tie this to efi
#asahi-dev · 2022-02-08 12:56
kettenistying this to EFI would restrict the paint colors somewhat
#asahi-dev · 2022-02-08 12:53
kettenis(currently it is all ACPI-specific)
#asahi-dev · 2022-02-08 12:53
kettenis(he's starting to look at suspend/resume, starting with creating a better way to hook that functionality up in a machine-independent way)
#asahi-dev · 2022-02-08 12:52
kettenisbut I think I can transition to new bindings in this area without too much disruption
#asahi-dev · 2022-02-08 12:52
kettenisso I have to be a little bit more careful with breaking stuff ;)
#asahi-dev · 2022-02-08 12:51
kettenistheo has an m1 macbook air now
#asahi-dev · 2022-02-08 12:51
kettenisyup
#asahi-dev · 2022-02-08 12:50
kettenisok, that would work
#asahi-dev · 2022-02-08 12:50
kettenisand
#asahi-dev · 2022-02-08 12:50
kettenisinfo-rtc_scrpad = 5137 / 0x1411
#asahi-dev · 2022-02-08 12:50
kettenisinfo-rtc = 5126 / 0x1406
#asahi-dev · 2022-02-08 12:49
kettenisso a simple base wouldn't work
#asahi-dev · 2022-02-08 12:49
kettenishmm, the layout of the RTC registersers is different
#asahi-dev · 2022-02-08 12:32
kettenisbut the j293 ADT doesn't
#asahi-dev · 2022-02-08 12:31
kettenisthe j314 ADT mentions CLKM
#asahi-dev · 2022-02-08 12:24
kettenisso doing the rtc and "bootmode scratchpad" would be enough for now
#asahi-dev · 2022-02-08 12:19
kettenisthe way the RTC works is a bit strange; bet Apple had a huge finger in the design of that part
#asahi-dev · 2022-02-08 12:17
kettenis:D
#asahi-dev · 2022-02-08 12:15
kettenismy understanding of SPMI is very limited ;)
#asahi-dev · 2022-02-08 12:15
kettenisAnd we probably should not add the SPMI controllers that are used by the coprocessors to the device tree
#asahi-dev · 2022-02-08 12:14
kettenisI suspect some of that is managed by the SMC and DCP
#asahi-dev · 2022-02-08 12:13
kettenisoh, they're different chips sitting on the SPMI bus
#asahi-dev · 2022-02-08 12:11
kettenisand we may need them to do better power management I suppose?
#asahi-dev · 2022-02-08 12:10
kettenisso those would become subnodes I suppose?
#asahi-dev · 2022-02-08 12:07
kettenisbut the slave can't be accessed directly from the AP I guess?
#asahi-dev · 2022-02-08 12:05
kettenisthe PMU is a separate chip isn't it?
#asahi-dev · 2022-02-08 12:05
kettenisand "apple,maverick-pmu", "apple,pmu" for the m1 pro/max machines?
#asahi-dev · 2022-02-08 12:05
kettenisso "apple,sera-pmu", "apple,pmu" for the m1 machines
#asahi-dev · 2022-02-08 12:04
kettenisat least for the top node
#asahi-dev · 2022-02-08 12:04
ketteniswe'd probably still want to keep a scheme with two compatibles
#asahi-dev · 2022-02-08 12:02
kettenisso the RTC could simply be an "apple,pmu-rtc" node
#asahi-dev · 2022-02-08 12:01
ketteniswould a simple "reg" property be sufficient for the subnodes?
#asahi-dev · 2022-02-08 11:59
ketteniswhich works best if there are limited in scope
#asahi-dev · 2022-02-08 11:58
kettenisso the question is if we can create sub-drivers that are somewhat future-proof
#asahi-dev · 2022-02-08 11:57
kettenisthat appears to be future-proof
#asahi-dev · 2022-02-08 11:57
kettenisso the way this implemented now is that you have a base driver which really only provides the regmap
#asahi-dev · 2022-02-08 11:56
kettenisyes indeed
#asahi-dev · 2022-02-08 11:15
kettenis(and yes have some food first)
#asahi-dev · 2022-02-08 11:15
kettenis(streaming sessions don't work too well for that)
#asahi-dev · 2022-02-08 11:14
kettenismarcan: we should discuss pmu device tree bindings

plus ›