On Thu, Apr 05, 2012 at 09:08:20PM -0400, Dave Mielke wrote: > [quoted lines by Greg KH on 2012/04/05 at 17:40 -0700] > > >It does matter, > > Then my question becomes how is it to be done for the CP2110? It's already done in the driver, right? > >what's wrong with using the in-kernel driver for this device? > > One reason is that the /dev/ttyUSBn path isn't predictable. Even if it's made > predictable on Linux, it still mightn't be on other platforms. Then that is up to those other platforms. And the /dev path is quite predictable, look at /dev/serial/ and use the symlinks there that are provided for you. Unless your device is "dumb" and doesn't have a serial number, then you might have problems, but you can work around them by looking at physical positions and the like. > >You shouldn't have to mess with usbfs or HID reports or > >anything else like this at all, it should all "just work" for you. > > Yes, in theory, that's correct. The problem is, however, that this code must > work on many platforms (e.g. we're now working on making it even work within > grub) so it's just far easier, from my perspective, to have a generic layer of > basic USB operations which works everywehre, and then to use that. I totally fail to see how this pertains to Linux, given that we already support this device just fine, and you shouldn't be writing userspace serial tty drivers anyway, as no one will be able to access the device that way (unless you are going to emulate the whole POSIX serial port api from userspace as well, and surely you wouldn't do that...) sorry, 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
