#
✕ nick:marcan
#asahi · 2021-08-09 17:50
marcan(except SEP which won't use that lib probably)
#asahi · 2021-08-09 17:50
marcanso it's going to be 1:1 an RTKit-lib driver and the mailbox, in every case
#asahi · 2021-08-09 17:50
marcandue to the architecture mismatch
#asahi · 2021-08-09 17:50
marcanthere will be a single endpoint
#asahi · 2021-08-09 17:49
marcanbut we can't use those endpoints
#asahi · 2021-08-09 17:48
marcanmessages would be 64+8 bits, probably 64+64 in practice
#asahi · 2021-08-09 17:48
marcanstandard queueing stuff
#asahi · 2021-08-09 17:48
marcanif the FIFO is full we block until IRQ
#asahi · 2021-08-09 17:48
marcanwell sure, but that's just a standard FIFO thing, nothing about specific messages
#asahi · 2021-08-09 17:47
marcanit's a fire and forget message
#asahi · 2021-08-09 17:47
marcanyes but why do we need to wait for anything?
#asahi · 2021-08-09 17:47
marcanpipcet[m]: is there any case where we need to know that message transmission is done? that sounds wrong.
#asahi · 2021-08-09 17:46
marcanalso, if it's instantiated separately, that still leaves the issue of the cpu boot and other things we might need to do, which might end up outside of the mailbox driver's API
#asahi · 2021-08-09 17:45
marcanmight make sense to use, but I need to look at the framework to convince myself
#asahi · 2021-08-09 17:44
marcanjust the FIFO?
#asahi · 2021-08-09 17:44
marcanwhat is the benefit of using the mailbox interface?
#asahi · 2021-08-09 16:55
marcanyeah, we justr need a separate hook for SART
#asahi · 2021-08-09 16:53
marcanthere's many ways to do this once it's our own code everywhere
#asahi · 2021-08-09 16:53
marcanjust tell the library not to use a DART but something else instead
#asahi · 2021-08-09 16:53
marcanexactly
#asahi · 2021-08-09 16:52
marcanthe drivers would be DCP, ANS, etc
#asahi · 2021-08-09 16:52
marcanit would just be a library
#asahi · 2021-08-09 16:52
marcanit doesn't even really need to be a driver
#asahi · 2021-08-09 16:52
marcanand as part of that they instantiate the default endpoints, or not
#asahi · 2021-08-09 16:52
marcanlike, we can have the core mailbox/ep handling stuff, that is instantiated by downstream drivers that actually handle devices like DCP
#asahi · 2021-08-09 16:51
marcanyeah, so all those can be variations in one driver
#asahi · 2021-08-09 16:48
marcanand apple stuff is never going to use another mailbox driver
#asahi · 2021-08-09 16:48
marcannothing other than apple stuff is ever going to use the ASC driver
#asahi · 2021-08-09 16:48
marcanneither seems to be the case here
#asahi · 2021-08-09 16:48
marcanthose abstractions are only useful when they either provide enough useful primitives to reduce code, or existing higher level drivers can hook into the interface
#asahi · 2021-08-09 16:47
marcannothing says we have to use kernel abstractions
#asahi · 2021-08-09 16:47
marcanit doesn't seem useful to use an abstraction for us
#asahi · 2021-08-09 16:47
marcansven: honestly I'm leaning towards ignoring mailbox/etc and just doing our own driver
#asahi · 2021-08-09 12:48
marcanit's basically a fork IIRC
#asahi · 2021-08-09 12:48
marcankind of
#asahi · 2021-08-09 12:45
marcanbut sure, I'd be happy to have some of the gentoo ARM folks around
#asahi · 2021-08-09 12:44
marcanis gentoo even "big" any more? :-)
#asahi · 2021-08-09 12:42
marcanbut gentoo is easy, just have an overlay
#asahi · 2021-08-09 12:42
marcanchadmed: I'm a gentoo user fwiw :)
#asahi · 2021-08-09 12:42
marcanbut we'll see
#asahi · 2021-08-09 12:42
marcanthat driver has a bunch of fake FIQ DMA stuff we are definitely not using, and doesn't mention FIFOs which apple's controller has, so this might be a good case for writing a new driver rather than trying to shoehorn apple mode into that one
#asahi · 2021-08-09 12:36
marcanjannau: cool!
#asahi · 2021-08-09 12:36
marcanI'll just hvtrace it and see how close it is
#asahi · 2021-08-09 12:35
marcanspi-s3c24xx.c matches at least *one* of the register names
#asahi · 2021-08-09 12:34
marcanstrings suggests it does dma, but there's no DART... I'm going to assume they don't actually use that here
#asahi · 2021-08-09 12:32
marcanheh, ok
#asahi · 2021-08-09 12:30
marcandid we ever figure out if spi is a thing that already has a similar driver? IIRC it was "samsung stuff" but I wonder if it's like UART or not
#asahi · 2021-08-09 12:29
marcanack
#asahi · 2021-08-09 12:28
marcanwe've had enough chitchat, time to throw implementations at people and figure out if they stick
#asahi · 2021-08-09 12:27
marcanalright, then it's gonna be whatever I can come up with :p

plus ›