Hi, On Mon, Aug 08, 2011 at 01:44:14PM +0300, Heikki Krogerus wrote: > > Currently we have the situation that host and device driver are > > completely unaware of each other, at least on most architectures. > > >From what we have in the tree I like the msm usb otg best, apart > > from the fact that it's a platform specific solution. It's even > > not complicsted to make it work on i.MX and therefore probably > > on other architectures aswell. > > I think these kind of things should be considered after we have the > basic phy support separated from otg, so not at this stage. I'm not > sure how it would work with Felipe's idea of keeping everything as > flexible as possible. > > I can only point out that there should not be anything preventing > something like pci card or sdio card providing a controller and a > transceiver to the system, so relying on platform data may not be > enough. How would you guys feel if the transceiver would be matched in several ways? ULPI transceivers can be matched at runtime, as you can always read the product and the vendor id. Other interfaces will need to be handled differently. I guess with UTMI, there is no need for transceiver driver, or it's enough to have one for all of them. Well, in most cases... :). We would in any case always have what Sascha suggested as the last way of matching the transceiver. What do you guys think? -- heikki -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
