On Sun, May 3, 2009 at 20:41, Alan Stern <stern@xxxxxxxxxxxxxxxxxxx> wrote: > On Sun, 3 May 2009, Kay Sievers wrote: > >> I thought about something like this, which gets rid of the additional >> class, adds the endpoints to the usb subsystem directly, and also does >> not request a device node until usbfs2 is eventually implemented. >> >> At the time these nodes get enabled, we must also find a naming scheme >> for them, instead of letting them show up in the root of /dev, like the >> current dead ones do. >> >> Quick patch to illustrate below. > > If you haven't already done so, you should check all the other places > where usbcore uses is_usb_device(). They are likely to get confused if > endpoint thingies start showing up. Yeah, right. A quick look did not reveal any other places. > For example, we don't want to > allow drivers to bind to endpoints (not yet). In fact, you might want > to add a usb_endpoint_type, in addition to usb_device_type and > usb_interface_type. That's the next thing to do, if we do this change. It's also useful that it get added to the event environment that way, and we don't need to match on the device name then to find the right device-type. What do you think about the name to choose? It's: 4-0:1.0-ep81 now. But it could be: 4-0:1.0-ep_81 to match closer to the old name, or anything else, or David liked to add direction to the name, if I remember that correctly? Thanks, Kay -- 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
