[PATCH 0/2] Synaptics absolute coordinate motion event support

[Peter Hutterer]

On Mon, Oct 11, 2010 at 12:39:03PM -0400, Joe Shaw wrote:
> On Sun, Oct 10, 2010 at 11:51 PM, Peter Hutterer
> <peter.hutterer at who-t.net> wrote:
> > On Fri, Oct 08, 2010 at 03:20:40PM -0400, Joe Shaw wrote:
> >> Attached are a pair of patches which add support for absolute
> >> coordinate motion events to the synaptics driver.  The first patch
> >> allows Absolute to be passed into XSetDeviceMode() and the latter posts
> >> absolute motion events using the raw X and Y coordinates rather than
> >> computing deltas.
> >
> > I do wonder how many users actually use their touchpad in absolute mode.
> 
> Indeed, I doubt many, but these patches aren't academic.  I'm working
> on a product which does use a touchpad in absolute mode.

do you use the synaptics features (two-finger scrolling, tapping, etc.)? 
if not, evdev may be the better choice.

> > either way, with these patches the cursor jumps back to 0/screen height
> > because the last four events sent by the device after releasing are
> > xmin/ymax, resetting the cursor. Something's not quite right there, can you
> > please check this again? thanks.
> 
> I saw something like this too on my device, but since we're also
> working on the hardware I just fixed the firmware to not send an X,Y
> of 0,0 when BTN_TOUCH is 0.
> 
> Do you see the xmin/ymax issue with evtest as well, or only through
> "xinput test"?

yes, I get x 0 and y <max> values in the same event that sets BTN_TOUCH 
to 0. which indicates a bug in the synaptics driver, we should just ignore
that packet then.

Cheers,
  Peter

> >> I also have patches against the 1.2 branch -- this is the version that
> >> I am currently using -- so it'd be good those could get applied too.
> >
> > 1.2 won't get new features. it's kinda the point of a stable branch.
> 
> Fair enough, I'll just patch it in.
> 
> Thanks,
> Joe
>