[RFC PATCH] Initial libudev input-hotplug support

Ping pinglinux at gmail.com
Tue Oct 6 14:07:33 PDT 2009

Thank you, Matthew, for picking up this topic for Wacom.  Please see my
comments in line.

On Tue, Oct 6, 2009 at 11:25 AM, Matthew Garrett <mjg59 at srcf.ucam.org>wrote:

> I think that this will currently break wacom. Wacom is set up to provide
> one input device per input mechanism - so typically we may have pen,
> stylus and eraser. The driver insists on each of these being declared as
> separate instances, so in hal we have a callout for wacoms that
> generates extra devices in the hal namespace. Each of these devices then
> corresponds to one of the input mechanisms. For this to work in a udev
> world, we either need to have a mechanism for a single udev event to
> generate (or remove) multiple input devices

If we go with this approach, I'd expect that requires more work on server
side. But xf86-input-wacom could stay unchanged.

> or alternatively to fix up
> wacom so it can be told to create multiple X input devices in a single
> instance.

Does xserver have mechanism to create multiple X input devices in a single
instance now? If not, this approach would need both server and driver work.
With the driver's current status (I mean out of sync with xserver), adding
more work now would not make too big a difference. We could design this
support into the current rework.

So, I don't have a preference for which way to go. I hope we choose an
approach that would have a longer lifespan.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://lists.x.org/archives/xorg-devel/attachments/20091006/8581dd39/attachment.html 

More information about the xorg-devel mailing list