On Thu, Nov 15, 2018 at 01:57:31PM +0100, Boris Brezillon wrote: > What we could do though, is expose I3C devices that do not have a > driver in kernel space, like spidev does. Yes, this is much safer and more robust especially if the bus has enumeration requirements like you're saying it does. It's much easier and more sensible to do this if the bus can be enumerated than it is with SPI, USB would be a good example here I guess (though I've never looked at the details of how USB does this on the implementation side so this may be me speaking in blissful ignorance). I do agree that there will be use cases that turn up for this.
Attachment:
signature.asc
Description: PGP signature
