On Tue, 9 Jun 2015, Stefan Koch wrote: > Am Dienstag, den 09.06.2015, 10:31 -0400 schrieb Alan Stern: > > On Tue, 9 Jun 2015, Stefan Koch wrote: > > > > > Hi > > > > > > I think whitespaces should be ok now. I have tested now three mail > > > clients... > > > > > > This patch series enables the interface authorization. The description > > > is improved now. Is the series correct now? > > > > This patch series is much, much more complicated than necessary. > > > > > The first patch allows to handle the probing of drivers > > > manually. > > > > Why do you do that? It's completely unrelated to the problem of > > authorizing USB interfaces. > To prevent drivers from binding. But you don't care about all drivers in general; you only care about drivers binding to USB interfaces. So you should change only the USB binding routines, not the routines in drivers/base/bus.c. > > > And the third patch introduces the usb interface authorization. This is > > > the main concern of this kernel modification. > > > > This patch also is more complicated than it should be. Why do you want > > to unregister and then re-register the interfaces? All you need to do > > is unbind drivers and prevent other drivers from binding. > Do you see a much simplier method to prevent drivers from binding as > with patch 1? Yes, and you already included it in your 3/4 patch -- the changes you made to usb_probe_interface() and usb_claim_interface(). > Another approach could be to check in bus_probe_device() (bus.c) if the > bus is "usb" and then use to_usb_interface() or so on and check the intf > auth status bit. Again, why bother? It's silly to check for the usb bus in bus_probe_device(), when usb_probe_interface() already _knows_ that the bus is usb. 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
