On Thu, 18 Feb 2010, Jari Vanhala wrote:

On Wed, 2010-02-17 at 13:41 +0100, ext Jiri Kosina wrote:

quoted

On Wed, 17 Feb 2010, Jari Vanhala wrote:
As the device doesn't seem to be registering itself on the HID bus at all, 
I am not sure we want to put it together with all the HID drivers. Maybe 
drivers/input/misc would be more appropriate?

I felt like I forgot something.. Well, moving it to input/misc is easy
(commit attached), but I (still) wanted to try getting it to use
hid-functions. Results weren't that nice. Looking hid-core just shows,
it's thought to connect (hotplug) usb or bt devices and i2c doesn't fit
nicely to that. Making it to use hid_allocate_device(), hid_connect()
etc takes actually more code..

Well, the major differentiator should be whether the "on-the-wire" the 
protocol is using is (or at least tries to be in some aspects) 
HID standard compliant.

Which, as far as I understand, is not the case here.

Leaving it as-is in hid would look little odd, so probably moving it to
drivers/input/misc is best thing. What you think? Dmitry?

Thanks,

-- 
Jiri Kosina
SUSE Labs, Novell Inc.