On Thu, Jul 22, 2010 at 9:09 PM, Adam Kropelin <akropel1@xxxxxxxxxxxxxxxx> wrote: >> From what I read, I think you are right. And it seems to me that it is not >> a very good idea to have a USB Composite device with multiple >> configurations. What is the use case for that? > > I don't know. My device has only a single config into which it places > several interfaces grouped by IADs. I had expected Linux to simply > ignore the RNDIS interfaces if couldn't utilize them, That is my expectation as well. I think the original code is flawed. Why reject the configuration at all? It is just there will be no rndis support from the host Linux OS. But it is still a valid USB device. There are many USB device out there without kernel driver support yet they are all valid USB device. > but instead it > refuses to select the config at all, rendering the other interfaces > inaccessible. When I attempted to work around the issue by providing a > second fallback config to satisfy Linux, Windows broke due to the > issue outlined in that blog post. Catch 22. I do not agree that Windows is broken here... -- Xiaofan http://sourceforge.net/projects/libusb-win32/ -- 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