support absolute input devices

[dsd]

In the x86-3.1 kernel (and upstream), the hgpk_mode selection of PenTablet mode now reports absolute coordinates to userspace.

kbdshim should be extended to pick up these devices and to convert the absolute motions to relative ones, hence supporting rotation and user activity detection.

[pgf]

what's the status of this? is it still necessary?

[dsd]

Yes - when the HGPK pad is put into PenTablet mode it reappears as an absolute input device, so we need to handle it still.

[dsd]

patch to implement this

[dsd]

Patch attached above.

[dsd]

Paul committed this, thanks.

[dsd]

Fixed in kbdshim-26. Peter, please package this. Source tarball is http://dev.laptop.org/~pgf/rpms/srpms/olpc-kbdshim-26-git74c4c20.tar.gz

[pbrobinson]

olpc-kbdshim-26-1.fc17 built for x86 and ARM, should be in the next fedora repo compose for builds

[dsd]

Test in 12.1.0 build 8.

[greenfeld]

Tested with 12.1.0 build 8 with an ALPS touchpad per pgf's suggestion. {Touchscreens take a different input path.}

When rotated to one of the two sideways orientations, the touchpad on an XO-1 with ALPS is not rotated to match.

[dsd]

Did you test the touchpad in touchpad mode or in pentablet mode?

The touchpad mode is known broken (#11798), the pentablet one (this ticket) should work.

[greenfeld]

I did not know that the different modes of the ALPS touchpad reported differently back to the OS.

In non-capacitive mode the ALPS touchpad rotates its output with the display in 12.1.0 os8 on a XO-1.