new gadget - Doro 212IPC usb telephone

[Hans Petter Selasky]

On Saturday 21 May 2005 08:13, Torfinn Ingolfsen wrote:
> On Fri, 20 May 2005 15:29:46 -0400
>
> Mathew Kanner <mat at cnd.mcgill.ca> wrote:
> >  Don't load the UHID driver, let ugen get the device.  In
> > theory it would be nice to have dual devices, that is, a ugen and a
> > uhid/umass/whatever at the same time for the device.
>
> Hmm, I'm just brainstorming here. For the 'udesc_dump' case, wouldn't
> that just mean that uhid (and other u* (usb) drivers) would have sort of
> a "fall through" to ugen (or at least ugen functionality)?
> I mean, the device descriptor is supported on all devices, right?
> So uhid and other usb drivers would pass any requests they don't support
> natively to ugen.

The ugen driver is probed last, when no other usb drivers have been found.

>
> Oh well, it probably would complicate things in the usb drivers.

The problem is that pipes are setup/stopped/aborted and not transfers. Then 
ugen might infer with transfers that were not started by ugen. Also the 
setting of configuration index and alternate index has got to be under some 
restrictions.

If you got device "snd_uaudio" loaded or compiled into the kernel, 
a /dev/dspX.X" device should be created for your telephone. The problem is 
that "bInterfaceNumber 1" requires setting of "bAlternateSetting 1" to work:

 Standard Interface Descriptor:
   bLength            9
   bDescriptorType    04
   bInterfaceNumber   1
   bAlternateSetting  0
   bNumEndpoints      0
   bInterfaceClass    01
   bInterfaceSubClass 02
   bInterfaceProtocol 00
   iInterface         0

 Standard Interface Descriptor:
   bLength            9
   bDescriptorType    04
   bInterfaceNumber   1
   bAlternateSetting  1
   bNumEndpoints      1
   bInterfaceClass    01
   bInterfaceSubClass 02
   bInterfaceProtocol 00
   iInterface         0

So the uaudio driver (see /sys/dev/sound/usb) might have to be changed.

--HPS