On Sun, 30 Aug 2009, Tilman Schmidt wrote: > While working on the driver for the device: > > ID 0681:0002 Siemens Information and Communication Products Gigaset 3075 Passive ISDN > > I have repeatedly encountered an error state I can only clear > by disconnecting and reconnecting the USB cable. The device > just stops responding to commands. (Specifically, the USB OUT > transfer sending the command to the device is successful but > the expected reply from the device is never signalled.) This > state persists even if I unload the driver and load it again: > the driver probes the device just fine, switches the altsetting > successfully, but none of the initialization commands get any > response. When I unplug and replug the USB cable, however, the > device immediately starts responding normally again. > > I should mention that this device is self-powered, so unplugging > the USB cable does not power down the device. In fact, its main > function (ISDN PABX and DECT base) is unaffected by the error or > by the unplugging. > > I'd like to spare the users of the driver the hassle of pulling > the plug. So what does disconnecting the USB cable do that even > unbinding the driver doesn't? Unplugging the cable removes the +5V power line. Even though your device is self-powered, it may monitor this voltage. More significantly, when the cable is plugged back in again the system will reset the device. Unloading and reloading a driver does not do this. > What can I do programmatically > from within the driver to get as near to an actual physical > disconnect as possible? You can call usb_reset_device() or usb_queue_reset_device(). See the kerneldoc for those functions in drivers/usb/core/hub.c; using them correctly can be a little tricky. 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
