#
✕ nick:marcan
#asahi-dev · 2021-07-19 18:47
marcanthat will definitely make the kernel panic if not disabled
#asahi-dev · 2021-07-19 18:47
marcanI saw someone run into that, they still had KTRR/CTRR enabled
#asahi-dev · 2021-07-19 18:47
marcanoh also, make sure you have all the bputil security stuff disabled!
#asahi-dev · 2021-07-19 18:46
marcanterrible idea once things work, since it makes everything super slow
#asahi-dev · 2021-07-19 18:46
marcanthat will add a nasty exception handler patch which might catch early boot problems for you
#asahi-dev · 2021-07-19 18:46
marcanyou can look for a call to #self.patch_exception_handling() in hv.py and uncomment that (it's hacky)
#asahi-dev · 2021-07-19 18:45
marcanI actually haven't tried a retail kernel, though I'd hope it would work. I might try that later this week, since I just upgraded to 12.0 and I might not bother with the dev kernel thing if I don't have to
#asahi-dev · 2021-07-19 18:44
marcanif you see a progress bar under the rainbow logo then it's working; if you don't see that then something likely went wrong
#asahi-dev · 2021-07-19 18:44
marcanif you don't have any active tracers in the hypervisor, it's actually normal for there to be next to no logging; without any tracers, there is very little the hypervisor is doing after all.
#asahi-dev · 2021-07-19 18:43
marcan/dev/m1n1-sec would then be where the xnu serial output goes
#asahi-dev · 2021-07-19 18:43
marcanif you're on Linux, you can add udev rules like these to have it properly identified: https://mrcn.st/p/UKqxUOw2
#asahi-dev · 2021-07-19 18:41
marcanDanielHuisman[m]: you need to open the secondary USB interface with e.g. picocom to see serial output
#asahi-dev · 2021-07-19 06:50
marcanyeah, the SEP does some HDCP things
#asahi-dev · 2021-07-19 00:44
marcanI mean I briefly worked for a black company :p
#asahi-dev · 2021-07-18 15:32
marcanthe set_block_dcp stuff looks like float arguments, which is interesting
#asahi-dev · 2021-07-18 15:30
marcanand that's it, then it starts submitting surface updates
#asahi-dev · 2021-07-18 15:30
marcanthen it stats doing swaps; the first time it does some interesting things with configuring the color matrices etc, there's also a generic "set_block_dcp" which might well be raw reg writes
#asahi-dev · 2021-07-18 15:29
marcanthen finally more easy no-arg calls, a writeDebugInfo thing we can surely ignore, some power management stuff, and finally IOMobileFramebufferAP::set_digital_out_mode which again causes a big prop thrash
#asahi-dev · 2021-07-18 15:28
marcanthen there's thrashing because for some reason it detects a display hotplug (this might be related to HDCP crap or my capture card)
#asahi-dev · 2021-07-18 15:28
marcanand that's the end of init
#asahi-dev · 2021-07-18 15:28
marcanthen there's the big dcpav prop dump
#asahi-dev · 2021-07-18 15:27
marcanand that stuff is more init calls with no args, some memory management thing, and that's about it
#asahi-dev · 2021-07-18 15:27
marcanhttps://mrcn.st/p/W4ijTdUF like the beginning of that boils down to "return true" crap, a bunch of property updates, and then another big nested callback where the AP starts doing stuff
#asahi-dev · 2021-07-18 15:26
marcanand a lot of the callbacks are bullshit, like registering/looking up iokit stuff that just has to return "true"
#asahi-dev · 2021-07-18 15:25
marcansven: to be fair it's not *that* bad. I think we can ignore a lot of the crap macos calls
#asahi-dev · 2021-07-18 15:10
marcan(without docs)
#asahi-dev · 2021-07-18 15:10
marcanboth of those are on my list of things I sure hope I don't have to touch because I don't even know how you would cleanroom that
#asahi-dev · 2021-07-18 15:09
marcan:p
#asahi-dev · 2021-07-18 15:09
marcansven: you wanna sign up to write the displayport training and bandwidth calculation code?
#asahi-dev · 2021-07-18 15:04
marcanI did paste the RPC dump there
#asahi-dev · 2021-07-18 15:03
marcanso far nobody hates me in #dri-devel at least
#asahi-dev · 2021-07-18 15:03
marcanI mean... it's either this or no M1 support
#asahi-dev · 2021-07-18 15:00
marcanit is completely evident this was made for macos and only for macos
#asahi-dev · 2021-07-18 15:00
marcanbut we don't have that luxury :p
#asahi-dev · 2021-07-18 15:00
marcanof course it would be
#asahi-dev · 2021-07-18 14:56
marcanalyssa: *shrug* nothing surprises me any more
#asahi-dev · 2021-07-18 14:37
marcanseems the RPC system for those is called "EPIC" and I hope it's less insane than the main DCP one
#asahi-dev · 2021-07-18 14:36
marcanlook for all the *Proxy.kext ones too; those are proxies to specific DCP endpoints
#asahi-dev · 2021-07-18 14:33
marcanAppleMCDP29XX.kext is in there too etc
#asahi-dev · 2021-07-18 14:31
marcanyou can piece together the DCP firmware from existing kexts, e.g. parts of IOAVFamily.kext are in there too
#asahi-dev · 2021-07-18 14:30
marcanwhich have the other half of the driver
#asahi-dev · 2021-07-18 14:30
marcanalso see IOMobileGraphicsFamily.kext vs IOMobileGraphicsFamily-DCP.kext
#asahi-dev · 2021-07-18 14:29
marcanha
#asahi-dev · 2021-07-18 14:29
marcanI don't think DCP has anything to do with the touch bar
#asahi-dev · 2021-07-18 14:28
marcanso that makes sense
#asahi-dev · 2021-07-18 14:28
marcanIIRC the display hardware for that thing is different
#asahi-dev · 2021-07-18 14:28
marcanah, I was wondering about DFR
#asahi-dev · 2021-07-18 14:27
marcansee AppleMobileDispH13G-DCP.kext vs AppleMobileDispH13G.kext
#asahi-dev · 2021-07-18 14:26
marcan(even if no swap requests were made)
#asahi-dev · 2021-07-18 14:26
marcanone good reason, though, would be to allow the main CPU to go to sleep for extended periods of time. I wouldn't at all be surprised if DP/HDMI protocol bullshit required the DCP to wake up for every frame under some circumstances.

plus ›