#asahi-dev · 2022-03-22 12:31
marcansame with PMP and everything else#asahi-dev · 2022-03-22 12:31
marcanGFX also will, though I'm dreading having to do more refactoring of mailbox if it ends up being a performance bottleneck there, since there it actually matters#asahi-dev · 2022-03-22 12:30
marcane.g. SMC will reuse all of the RTKit stuff as-is from the NVMe submission, but instead builds up a key/value interface over a user endpoint#asahi-dev · 2022-03-22 12:30
marcanyes#asahi-dev · 2022-03-22 12:30
marcanSEP, in particular#asahi-dev · 2022-03-22 12:30
marcanby the way, there are mailboxes that *don't* do RTKit#asahi-dev · 2022-03-22 12:30
marcanrtkit is just apple's RTOS; we use the name to refer to the shared framework it implements over the mailbox#asahi-dev · 2022-03-22 12:28
marcanmight as well just accept that there's no good way to do it and go monolith; at least it won't be a gpu+display monolith like x86 GPUs#asahi-dev · 2022-03-22 12:28
marcanit'd just end up being a pile of drivers with a ton of incomprehensible cross-dependencies#asahi-dev · 2022-03-22 12:28
marcan*but* display controllers are a tightly tied mess already, so I don't see using different drivers for that in linux buying us anything really#asahi-dev · 2022-03-22 12:27
marcanthe only copro that does have a ton of interfaces some of which are logically separate is DCP; that one has sub interfaces for dealing with things like DisplayPort hardware and the DP2HDMI etc, besides the main giant dcpep#asahi-dev · 2022-03-22 12:27
marcanand the existing abstraction does *not* work properly, IIRC due to how it handles queuing, if we try to expose the different endpoints as different channels#asahi-dev · 2022-03-22 12:26
marcanarnd: our conclusion re: mailbox was that because Apple's coprocessors generally only do one thing, trying to use the mailbox abstraction to provide the multiplexing is just a waste of time#asahi-dev · 2022-03-22 12:25
marcanyeah#asahi-dev · 2022-03-22 12:24
marcanRTKit is really a higher level interface and needs to deal with things like buffer/memory management too#asahi-dev · 2022-03-22 12:24
marcanbut trying to use it above RTKit is something we thought of, and decided would be an even worse idea because it clashes even more with the mailbox design#asahi-dev · 2022-03-22 12:23
marcanbut now we're stuck with it, fine, as a dumb low-level abstraction on top of the raw mailbox hardware#asahi-dev · 2022-03-22 12:23
marcanit is poorly designed, and different drivers interpret its design in different, incompatible ways#asahi-dev · 2022-03-22 12:23
marcanmaz: if you want my personal opinion on mailbox, using that subsystem at all was a mistake because it just *sucks*#asahi-dev · 2022-03-22 12:11
marcanjoske: IIRC pacstrap really wants the target dir to be a mountpoint#asahi-dev · 2022-03-22 10:23
marcan(that's the only package needed since otherwise we start with a fully baked ALARM tarball)#asahi-dev · 2022-03-22 10:23
marcanso the host needs to have it installed first#asahi-dev · 2022-03-22 10:23
marcanincluding repos etc#asahi-dev · 2022-03-22 10:23
marcanjoske: pacstrap pulls it from the host before all those scripts start#asahi-dev · 2022-03-22 01:18
marcanhaving a UI can be nice though, depending on the case at hand#asahi-dev · 2022-03-22 01:16
marcanand the protocols are very heterogeneous, DCP already uses two different IPC mechanisms (this is only one of them) plus other stuff#asahi-dev · 2022-03-22 01:16
marcanso writing wireshark modules would be somewhat duplicative if we're going to write that anyway#asahi-dev · 2022-03-22 01:15
marcanbut the thing about the python is that it's not just a dissector, we often use the same frameworks to then build actual clients, e.g. DCP#asahi-dev · 2022-03-22 01:15
marcancute :)#asahi-dev · 2022-03-22 00:12
marcanlook at the nullable logic in Method to see how that works#asahi-dev · 2022-03-22 00:12
marcanthat should be part of the InPtr#asahi-dev · 2022-03-22 00:10
marcancould be a pointer null flag#asahi-dev · 2022-03-22 00:10
marcanthe 00 might be relevant#asahi-dev · 2022-03-22 00:10
marcan3 bytes of padding to align to 4 bytes are normal too#asahi-dev · 2022-03-22 00:09
marcannicolas17: yes, it's leftover garbage#asahi-dev · 2022-03-21 20:16
marcancool#asahi-dev · 2022-03-21 20:12
marcanthanks!#asahi-dev · 2022-03-21 20:12
marcancool :)#asahi-dev · 2022-03-21 20:08
marcankettenis: you can send it for me if you like :)#asahi-dev · 2022-03-21 20:07
marcanI honestly don't remember off the top of my head how this works, so... :-)#asahi-dev · 2022-03-21 20:04
marcannicolas17: feel free to send a PR :)#asahi-dev · 2022-03-21 19:08
marcanthis is from january, probably 12.0.1?#asahi-dev · 2022-03-21 19:08
marcanhttps://hub.marcan.st/t/dcp-m1pro.log#asahi-dev · 2022-03-21 19:06
marcanso there is some duplication there#asahi-dev · 2022-03-21 19:06
marcaninstead trace_dcp.py has its own, dumber, parser that only looks at function names#asahi-dev · 2022-03-21 19:06
marcanthey could be parsed live, but I never bothered integrating that#asahi-dev · 2022-03-21 19:05
marcanyes#asahi-dev · 2022-03-21 19:05
marcanso they would not help you for any other ASC#asahi-dev · 2022-03-21 19:05
marcannicolas17: FWIW the DCP traces are a protocol only used by DCP#asahi-dev · 2022-03-21 19:03
marcanI added that feature because there was a lot of iteration in the DCP protocol/marshaling parsing and I didn't want to have to reboot macOS under the HV every time I changed something