On Tue, Jun 23, 2020 at 03:53:22PM +0200, Lars Povlsen wrote: > Mark Brown writes: > >If there's a mux that needs to be handled specially that mux should be > >described in DT on the relevant boards, there shouldn't just be > >something hard coded in the controller driver. > I looked at the spi-mux driver, but that is more for muxing the CS's, as > I understand - not the actual bus segment. I could use it, but it would It doesn't matter that much exactly what signals get switched I think, we can't really tell by the time we get back to the controller. > require encoding the bus segment into the CS (double the normal > range). Also, selecting the bus interface is tightly coupled to the > controller - its not an externally constructed board mux. It sounds like this controller should be describing a mux all the time - if there's completely separate output buses that you can switch between then we'll need to know about that if someone wires up a GPIO chip select.
Attachment:
signature.asc
Description: PGP signature
