Re: 2.6 input drivers FAQ

[Vojtech Pavlik <vojtech@suse.cz>]

On Mon, Feb 02, 2004 at 03:28:49PM -0500, Valdis.Kletnieks@vt.edu wrote:
> On Mon, 02 Feb 2004 21:18:13 +0100, Vojtech Pavlik said:
> 
> > Because normally the X server reads them in very quick succession and if
> > you don't make a very short click, the sequence looks like this:
> > 
> > push1 push2 release1 release2, which is fine, because X interprets that
> > as just a push and a release.
> > 
> > If there is disk activity or something else that causes the scheduling
> > to be delayed, it's push1 release1 push2 release2, which counts as a
> > doubleclick.
> > 
> > Hence sporadic doubleclicking.
> 
> Well.. that would explain things except for the single /dev/psaux I have.
> 
> Could a similar timing hole happen if the system submerged into SMM
> code for a battery check or similar? (I know, that *should* cause
> lost events not duplicated, but....)

I don't see a way how this could happen. One thing you could try - run
evtest on the mouse event device in a window/on another vc, and when the
doubleclick happens, examine the last sreenful of events.

> > For movement, of course, you get twice the mouse speed, but usually most
> > people just adjust the acceleration settings and are done with that.
> 
> Haven't seen this.

Most likely your problem comes from elsewhere. Did you try another
mouse?

-- 
Vojtech Pavlik
SuSE Labs, SuSE CR