Re: [PATCH 2.6.11] IBM TrackPoint support

[Dmitry Torokhov <dmitry.torokhov@gmail.com>]

On Mon, 14 Mar 2005 08:40:22 -0500, Stephen Evanchik <evanchsa@gmail.com> wrote:
> On Mon, 14 Mar 2005 13:19:56 +0100, Vojtech Pavlik <vojtech@suse.cz> wrote:
> > How much does it interpret the stream in non-transparent mode? Are
> > commands also passed through in soft transparent mode?
> >
> > I'm asking because we might want to implement a passthrough port
> > similarly to what the Synaptics driver does and allow extended protocol
> > mice to be connected to the external mouse port.
> 
> I originally thought that I could implement something similar to the
> Synaptics driver. Unfortunately, while in transparent mode bytes are
> relayed unmodified with the TrackPoint controller disabled. In other
> words, no simultaneous usage.
> 
> That doesn't mean extended protocol mice couldn't be supported in
> transparent mode however. I didn't find it particularly useful given
> the TrackPoint itself would be disabled.
> 

Here is my take on it (now that I have skimmed the TrackPint spec) -
transparent mode is to be used only when querying the external device.
This way trackpoint does not interfere with data stream at all and the
kernel gets a chance to know exactly whta is behing the trackpoint -
Logitech, explorer, something more exotic... Once identification is
done transparent mode should be cancelled.

Bit 3 can be used to de-multiplex 2 streams; hopefully trackpoint is
able to rely packets longer than 3 bytes from the external device.

As far as I can see there is no point of exporting transparent mode to
the userspace via sysfs. I also do not think that we need to export
middle_button_disable as it is "..for compatibility with older
software expecting this bit be always 0" and we do not have such an
issue. Also, if we implement pass-through port, then ext_dev is also
not needed since user can either unbind the driver from pass-through
port or just ignore the secondary input device in his/her config.

-- 
Dmitry