Bug#682413: USB TrackPoint keyboards unusable since synaptics-usb was included

[Andrew Deason <adeason@dson.org>]

found 682413 linux/3.10.7-1
tags 682413 patch
thanks

I have encountered this, as well. After looking into it for a little
while, it seems that the synaptics_usb module puts the relevant device
in non-HID mode, and reads the raw data from the force sensor on the
trackpoint, and presents that force data as relative coordinates to the
linux input subsystem. That is incorrect, and makes for some pretty
weird results, as people in this bug report have probably seen.

I've submitted a few patches to linux upstream to try to improve this:

 - This patch: <http://article.gmane.org/gmane.linux.kernel.input/31877>
   doesn't solve the underlying issue, but fixes one of the issues
   causing seeming hypersensitivity in one direction. This is what
   causes the cursor to seem to move when you click one of the
   trackpoint mouse buttons.

 - This patch: <http://article.gmane.org/gmane.linux.kernel.input/31935>
   should fix the underlying issue and translates the readings from the
   force sensor into relative cursor movements.

 - This patch: <http://article.gmane.org/gmane.linux.kernel.input/31936>
   lets you put the device in HID mode via a module parameter, so you
   don't need to recompile the module to use it as a regular HID mouse.

If anyone here would like to test these out and see how the trackpoint
behaves, it would be appreciated. If you don't know how to run such
patches, I can provide info on how to do that. (Providing a debian
package I'm not sure would work, since we already have a synaptics_usb
driver in the regular kernel package... and I don't want to have to
generate an entirely new kernel package just for patching this one
driver.)

-- 
Andrew Deason
adeason@dson.org