On Tue, Dec 04, 2001 at 08:08:55PM -0500, Martin Hamilton wrote:
> 
> Am I right in thinking that usb_serial_probe() in usbserial.c is the
> function that needs to be frobbed in order to influence the default
> endpoint detection ?  I was thinking about a quick kludge based on
> subverting this assignment...
> 
>     interrupt_in_endpoint[num_interrupt_in] = endpoint;
> 
>  ... so that whereas currently (for the Belkin)
> 
>     interrupt_in_endpoint[0] is the control/exceptions endpoint, and
>     interrupt_in_endpoint[1] is the regular data transfer one
> 
> We swap the two around, i.e.
> 
>     interrupt_in_endpoint[0] is the regular data transfer endpoint, and
>     interrupt_in_endpoint[1] is the control/exceptions one
> 
> Would this be enough to fool the kernel into using the right ones?

Sounds like a place to try.

> If a hard coded kludge like this works, we can have a think about how
> best to incorporate this into the proper driver!

The logic should go into the mct_u232 driver itself.  It knows about the
device specific stuff.

> Presumably you'd prefer to keep the driver specific stuff out of
> usbserial.c and usb-serial.h as much as possible ?

Exactly :)

thanks,

greg k-h

_______________________________________________
[EMAIL PROTECTED]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-users

Reply via email to