Re: Mousedev or hiddev problem, was: Re: USB mice do not work on 2.6.12-rc5-git9, -rc5-mm1, -rc5-mm2

[Vojtech Pavlik <vojtech@suse.cz>]

On Wed, Jun 08, 2005 at 06:55:07AM +0200, Zoltan Boszormenyi wrote:
> Vojtech Pavlik írta:
> >On Tue, Jun 07, 2005 at 12:04:15AM +0200, Zoltan Boszormenyi wrote:
> >
> >>Hi,
> >>
> >>Zoltan Boszormenyi írta:
> >>
> >>>All the -bk7+ kernels I tried produced the same strange bug
> >>>on my system: after gpm started I was able to move the
> >>>pointer on the screen but when X started up, it's pointer froze.
> >>
> >>it turned out that there is nothing wrong with USB on my system.
> >>
> >>But someone broke the /dev/input/mouseX <-> USB mouse interaction
> >>in 2.6.11-bk7 and my two-headed system with two X servers were
> >>manually set up to use the distinct mouse devices so the two heads
> >>do not interfere.
> >>
> >>No wonder gpm works, it reads /dev/input/mice. Starting only
> >>one X and using /dev/input/mice I found no problems. Setting it
> >>back to /dev/input/mouse0, the mouse pointer is dead again.
> >>
> >>Someone deserves a mousebite...
> >
> > 
> >Most likely it's because the keyboards are now identified as having
> >mouse capabilities, too, and changing the numbers. Check
> >/proc/bus/input/devices.
> 
> Thanks for the enlightenment, now I have to use /dev/input/mouse2 and
> mouse3 for the two X servers.
> 
> BTW, where is it documented? Was it this changeset?
> 
> <vojtech@suse.cz>
>         input: Fix keyboard scrollwheel support, add horizontal
>                wheel support, and enable both by default.
> 
>         Signed-off-by: Vojtech Pavlik <vojtech@suse.cz>
 
Yes, it was this one, and it'll be disabled again for 2.6.12, since it
causes headaches like yours.

It wasn't documented, because the mouse# numbers were never considered
to be stable (they change if you plug/unplug mice), and thus I didn't
expect anyone using them without having udev make them stable. My wrong.

-- 
Vojtech Pavlik
SuSE Labs, SuSE CR