On Wed, Mar 20, 2013 at 10:42:50PM +0400, Sergei Organov wrote: > Johan Hovold <jhovold@xxxxxxxxx> writes: > [...] > > Just for the record, there's really nothing wrong with the usb-serial > > throughput. The days of a single read and write urb are long gone > > (2.6.32 if I remember correctly). > > That's what I remember as well. Since some version I was not able to > significantly out-perform the usb-serial with my own driver, so I saw no > reason to submit/support one. > > > The reason why one shouldn't use the generic driver for a "real" > > usb-serial device is that you cannot control baudrates, etc, and of > > course that the device-driver matching isn't automatic. > > But these are 2 entirely different reasons. While the former can't be > solved with generic driver, the latter sure can. So the question is: why > not? Nobody cared, or there is some reason not to? I guess largely for historical reasons. But you're right (and it has been up for discussion before), if we end up with a bunch of such "dumb" devices, I think we should use a single driver for those at least until the need arises to differentiate them. I'll have a look at this shortly. Thanks, Johan -- 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
