[Meego-kernel] [PATCH 0/2] Usb gadget patches for Intel Medfield

[Greg KH]

On Fri, Oct 22, 2010 at 08:55:51PM -0700, Bastian, Waldo wrote:
> > But why does this have to be a platform specific gadget?
> 
> The ACM gadget driver is developed with a specific platform in mind
> because it was needed there, but the code itself isn't platform
> specific.

Good, it shouldn't be :)

> The patch for the Nokia gadget driver takes out eem/ecm function on
> the Medfield platform because the Medfield platform has a limitation
> on the number of USB endpoints that are supported.

So you just broke that driver for the nokia platform?

> > What's wrong with the acm gadget support we already have in the main
> > kernel tree? Why duplicate that work here?
> 
> It extends the kernel functionality. The kernel version provides only
> 1 ACM channel, this composite gadget driver provides 2 separate ACM
> channels. 
> 
> It's purpose in live is to provide access to modem debug functionality
> for which two channels are required: one channel is used
> bi-directionally for sending commands to the modem the other channel
> is used to stream modem debug information off the device. See also
> James patch for modem trace capture which hooks up with the gadget
> driver to forward the information from the modem.

You need to run this by the developers on the linux-usb mailing list
before assuming that this the correct thing to do.

I really don't see why you can't use the in-kernel gadget driver for
this type of functionality as it provides these types of "channels" for
you to use already.

Again, why can't you use the in-kernel code for this today?  It already
provides a tty port, which you can use for whatever you need.

Please follow this up on linux-usb at vger.kernel.org, and post the code
there for discussion.

thanks,

greg k-h