#
✕ nick:marcan
#asahi-dev · 2021-12-23 09:52
marcanRWCV
#asahi-dev · 2021-12-23 09:52
marcanI think I found the antenna SKU; it's another ACPI method
#asahi-dev · 2021-12-23 09:28
marcanstill need to figure out where the antenna sku comes from, I just hacked it in for now
#asahi-dev · 2021-12-23 09:28
marcanjust got the 4364B3 in this bali to work
#asahi-dev · 2021-12-23 09:28
marcan... but it's necessary on at least some T2 macs, without that entropy block the cards won't boot
#asahi-dev · 2021-12-23 09:27
marcanthat's cargo cult on M1 machines, since those wifi chipsets have hardware RNGs
#asahi-dev · 2021-12-23 09:27
marcanha, so that random bytes thing corellium was doing?
#asahi-dev · 2021-12-23 08:20
marcanso, now just the ACPI shenanigans for module-instance and antenna-sku left to make T2 macs work properly
#asahi-dev · 2021-12-23 08:19
marcanso it turns out 1) the core was wrong, as I said, 2) you need to flip an OTP enable bit, which I'd found but did not understand went along with the shadow area, *and* the arm64 macOS driver does not do that but the x86 build does, 3) you have to read the OTP area using 16-bit accesses on these chips
#asahi-dev · 2021-12-23 08:18
marcansigh... finally managed to read the OTP on 4364
#asahi-dev · 2021-12-23 04:24
marcanI mean chrome uses some fancy kernel stuff like seccomp and namespaces for sandboxing
#asahi-dev · 2021-12-23 04:23
marcanas in .config stuff
#asahi-dev · 2021-12-23 04:23
marcannot just missing kernel features?
#asahi-dev · 2021-12-23 04:22
marcan(which is why RHEL will never run their default kernels on M1s... that and they also disable DT support because apparently they want the whole world to run on ACPI)
#asahi-dev · 2021-12-23 04:21
marcan*RHEL
#asahi-dev · 2021-12-23 04:21
marcanbut 64K is common and RHLE defaults to that
#asahi-dev · 2021-12-23 04:21
marcanand I think almost no other system supports 16K?
#asahi-dev · 2021-12-23 04:21
marcanno, this one does not support 64K
#asahi-dev · 2021-12-23 04:20
marcanprobably not
#asahi-dev · 2021-12-23 04:19
marcanchromium is similarly broken
#asahi-dev · 2021-12-23 04:19
marcan"chrome is broken on Asahi Linux by choice" is a pretty bad look for them :p
#asahi-dev · 2021-12-23 04:18
marcanfeel free to :)
#asahi-dev · 2021-12-23 04:18
marcanno idea
#asahi-dev · 2021-12-23 04:18
marcanyeah well, let's see once enough people complain :-)
#asahi-dev · 2021-12-23 04:18
marcanmy current plan for chrome is to just do our first dev release with a big "chrome doesn't work, please file upstream bugs for 16K page support" known issue :p
#asahi-dev · 2021-12-23 04:16
marcanany non-emulation software that doesn't work on 16K pages is broken and needs to be fixed
#asahi-dev · 2021-12-23 04:16
marcanas far as I'm concerned we're doing 4K only because distros default to it (so they don't have to ship a separate kernel once everything is upstream) and for FEX to work properly; I still expect 16K to work for everything else, and will be shipping 16K kernels for our ALARM remix until 4K is stable and FEX makes sense
#asahi-dev · 2021-12-23 04:15
marcanit not working on 16K is stupid
#asahi-dev · 2021-12-23 04:15
marcanno, people should be filing bugs for chrome
#asahi-dev · 2021-12-23 04:14
marcanwhich is still a ways off to be useful anyway
#asahi-dev · 2021-12-23 04:14
marcanbecause x86 emulation
#asahi-dev · 2021-12-23 04:14
marcan*mixing
#asahi-dev · 2021-12-23 04:14
marcanpoke sven :p
#asahi-dev · 2021-12-23 04:14
marcanthe kernel needs a bunch of patches that aren't upstreamed yet to support mising 4K CPU and 16K IOMMU pages
#asahi-dev · 2021-12-22 20:25
marcananyway, more tomorrow, sleep now
#asahi-dev · 2021-12-22 20:10
marcanOTOH I'm looking at the arm64 driver... maybe I should look at the x86 one...
#asahi-dev · 2021-12-22 20:10
marcanand yet I finally found the OTP readback code in the apple driver and there's nothing special there
#asahi-dev · 2021-12-22 20:09
marcanand there's an "otp instead" bit and flipping it on just causes read commands to never complete
#asahi-dev · 2021-12-22 20:09
marcanI found a way to dump SROM via index/value, which is apparently related, but that doesn't contain the apple info
#asahi-dev · 2021-12-22 20:09
marcanthis is corroborated by a bunch of references to this in that android driver I found (which also parses OTP for different reasons)
#asahi-dev · 2021-12-22 20:08
marcanor if it is it's blank
#asahi-dev · 2021-12-22 20:08
marcan... but it isn't
#asahi-dev · 2021-12-22 20:08
marcanit's *supposed* to be at 0x800 in chipcommon on that one
#asahi-dev · 2021-12-22 20:08
marcanbut indeed as Redecorating[m] hypothesized, the core was wrong
#asahi-dev · 2021-12-22 20:07
marcanghidra :p
#asahi-dev · 2021-12-22 20:07
marcantoo bad it doesn't apply to x86... I really wish I had a hypervisor like m1n1 here, that would've probably given me the answer in minutes :p
#asahi-dev · 2021-12-22 20:06
marcanthanks :)
#asahi-dev · 2021-12-22 20:05
marcanin principle the patches should work as is if you disable the OTP codepath and manually pick the right firmware, but the entire point of this exercise is not having to do that
#asahi-dev · 2021-12-22 20:04
marcanspent all day trying to read the OTP on a BCM4364 and failing
#asahi-dev · 2021-12-22 20:04
marcanI'm trying to figure out firmware selection

plus ›