On Tuesday 18 August 2009 17:43:23 Alan Stern wrote: > On Tue, 18 Aug 2009, Laurent Pinchart wrote: > > Hi everybody, > > > > this patch series introduces an endpoint claim/release interface for > > function drivers to dispatch requests targeted at endpoints to the > > correct function driver. > > It's not clear why this should be needed. Don't function drivers > already allocate endpoints amongst themselves as part of configuration > initialization? Do they, but they don't establish an endpoint -> function ownership link. > After all, we can't have two different interfaces trying to use the same > endpoint. Except in different configurations. Actually it seems multiple configurations are not well supported today. The infrastructure is there, but function drivers allocate their endpoints globally at the gadget level and not at the configuration level. > If this is a matter of figuring out which function driver owns an > endpoint, can't that be done by searching through the endpoint > descriptors for each function's active interfaces? In that case I would have to loop over all functions for the active configuration (easy) and parse all descriptors for the correct speed to locate endpoint descriptors. That should be possible, would the overhead be ok ? -- Laurent Pinchart -- 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
