Mir absolute pointer support?

[Daniel van Vugt]

Thomas,

Yes, log a bug please.

All evdev motion should work (eventually) to some usable extent. 
Although I have observed some laptop touchpads simply don't work (e.g. 
in the presence of a TrackPoint). Sounds like the same kind of issue. We 
do need to revisit Mir's range of motion device support and it would be 
useful to start documenting which devices need work.

- Daniel


On 02/10/14 17:26, Thomas Hellstrom wrote:
> Hi!
>
> I'm putting together a VMMouse driver for the linux kernel. In the
> preferred mode, the mouse driver sends absolute XY position events: The
> idea behind the VMMouse is to be able to use the host cursor to avoid
> cursor lag, and then we need to explicitly specify the cursor position
> to the guest Virtual Machine.
>
> The idea with a kernel driver is twofold:
> 1) To be able to run Xorg rootless
> 2) To enable vmmouse functionality on Mir and Wayland.
>
> The mouse actually shows up as two input devices. One sending only
> relative movement events and wheel events, one sending only absolute
> position events. The idea is that the display server binds to both input
> devices and multiplexes the events, although relative movement events
> are very rarely used.
>
> This works fine with Xorg and XMir. The Xorg evdev driver recognizes the
> absolute device as a touchscreen and the relative device as a mouse.
>
> However, when I run the Mir demos, it's obvious that no events reach Mir
> from the absolute input device. Only mouse wheel motion that is sent
> through the relative device.
>
> Before I file a bug, I'd like to know whether absolute pointers are not
> supported by Mir yet or if there's some other kind of filtering going on
> (for example both the relative and absolute input devices share the same
> name and vendor etc.)
>
> Thanks,
> Thomas
>
>