multi-touch-dev team mailing list archive

[Chase Douglas <chase.douglas@xxxxxxxxxxxxx>]

On Fri, 2010-10-08 at 08:23 -0600, Duncan McGreggor wrote:
> Hey folks,
> 
> Chase and I have been discussing issues in our struggle to support MT on
> Synaptics hardware. We tried to make some appropriate decisions, but
> ultimately, to go any further, we need input from the MT community, from
> Xorg hackers, and kernel hackers. We need consensus.
> 
> Chase can follow up with a more detailed email, but the problems boil
> down to the following:
>  * should MT touches be disabled when they occur over a button area?
>  * should they just be disabled during a button press?
>  * should this be done in X?
>  * or in the kernel?
>  * we don't know the full Synaptics protocol; does anyone know ways of
> discovering if a touchpad has integrated or standalone buttons?
> 
> Chase, can you follow up with more technical details for folks and links
> to patches you've experimented with? And forward this message to the
> appropriate lists? There are some potentially subtle engineering issues
> at play here (as you've noted) and we need more input before we go any
> further.

Sure. First, the patches:

http://kernel.ubuntu.com/git?p=cndougla/ubuntu-natty.git;a=shortlog;h=refs/heads/multitouch-synaptics

The juicy bits are in the last three commits. The controversial bits are
in the last commit, which essentially ignores (or causes active touches
to vanish) if they occur inside a button area of an integrated trackpad
while a button is held down. This seems to work fairly well on my Dell
Mini 1012 when I modprobe psmouse like:

modprobe psmouse synaptics_multitouch=1 synaptics_button_thresh=4100

I'm going to ask around on the xorg-devel and linux-input mailing lists
on what others think the proper solution to this issue is. Feel free to
chime in here or elsewhere with your thoughts.

Thanks,

-- Chase