On Wed, Jun 23, 2010 at 12:26:09PM -0300, Otavio Salvador wrote: > Hello, > > On Wed, Jun 23, 2010 at 11:36 AM, Greg KH <greg@xxxxxxxxx> wrote: > > On Wed, Jun 23, 2010 at 10:06:32AM +0200, Oliver Neukum wrote: > >> Well, no. CDC-ACM is clearly modeled on a modem. It has notions > >> like learning about loss of carrier and other stuff and encapsulated > >> commands to an external modem. It is not just a serial driver. > >> > >> If you have a device that uses none of the extended features of cdc-acm, > >> you are better of writing a minimal serial driver. > > > > True, but you could just ignore all the modem stuff, and send and > > receive data from the device, right? That is probably what happens on > > Windows, as the device was made to work with the Microsoft-provided > > driver. > > So it would be nice to have a common decision about it; if it works as > CDC ACM should it be let as it or we should add a generic driver for > it? I say let it be. Again, the device does work properly, right? That's the most important thing here. > I think we could let it work _but_ find a way to export the > NOT_A_MODEM quirk somewhere. That would allow the system to handle it > differently. What would "handle it differently" look like? What would you decide is a feature that you don't want to support on such a device? Baud settings? Modem control lines? Something else? Those are all things that some userspace programs do check and care about. The device manufacturer made the device to match the spec, so unless it does not follow the spec in some manner, I don't see the benifit of using a different driver. thanks, greg k-h -- 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
