On Mon, 11 Aug 2014, Felipe Balbi wrote: > On Mon, Aug 11, 2014 at 09:44:05AM -0400, Alan Stern wrote: > > On Mon, 11 Aug 2014, Felipe Balbi wrote: > > > > > > > I don't know what exactly problem you met, but current pullup dp during loading > > > > > gadget driver behavior is not suitable for webcam and android use case even vbus > > > > > is there. > > > > > > > > Why not? What's wrong with it? > > > > > > g_uvc needs a userspace counterpart. If we connect to the host before > > > userspace is ready, we might connect as a non-usable function. > > > > Okay, but can't that be fixed in g_uvc? Have it disable the pullup > > until the userspace component is ready, then enable the pullup. > > that might be too late, although propbably not catastrophic. It looks like the UDC core needs to be improved a little. Right now, udc_bind_to_driver() unconditionally calls usb_gadget_connect(), thereby overriding any calls the gadget driver may have made during its bind callback. What we should do instead is call usb_gadget_connect() only if the gadget driver didn't try to change the pullup setting. > > I'm not sure how this could be made to work in a composite gadget, > > though. What if there were two uvc functions? The pullup would have > > you can refcount the disables, right ? Not according to the kerneldoc in gadget.h. Alan Stern -- 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