[Sugar-devel] XO-1 touchpad [EduJAM day 2 Tour of Uruguay; Sugar-devel Digest, Vol 31, Issue 6]

[Yioryos Asprobounitis]

> Message: 6
> Date: Tue, 3 May 2011 17:47:23 +1000
> From: James Cameron <quozl at laptop.org>
> To: David Farning <dfarning at activitycentral.com>
> Cc: 'IAEP' <iaep at lists.sugarlabs.org>,   
> 'Sugar Devel'
>     <sugar-devel at lists.sugarlabs.org>
> Subject: Re: [Sugar-devel] EduJAM day 2 Tour of Uruguay
> Message-ID: <20110503074723.GF3717 at us.netrek.org>
> Content-Type: text/plain; charset=us-ascii
> 
> On Mon, May 02, 2011 at 04:24:53PM -0300, David Farning
> wrote:
> > >From a tech point of view, the big question have
> been:
> > 4. questions about why touch pad behavior has changed
> in the latest
> > release with out giving teachers a heads up.
> 
> I don't know what release they are using, or were using.
> 
> I don't know if they mean XO-1.5 or XO-1.
> 
> XO-1.5 touchpad behaviour was changed; tap to click was
> disabled or
> enabled.  Default varied by release, and the default
> should be
> configurable.
> 
> XO-1 touchpad behaviour was changed.  Latest OLPC
> release on XO-1 is
> 10.1.3, which has no touchpad change compared to 10.1.2, so
> I presume
> they were on 8.2.1 or a derivative previously?
> 

I was loging serio-related messages for some time.

Post os802/2.6.25-kernel builds on the XO-1 have a considerable higher frequency of miscalibration symptoms (mostly big jumps and soft packet spews). 

This is true for both fedora and non-fedora builds and although the automatic recalibration is a big step forward they are still user-noticeable and very disruptive.

On average they occur every 5 minutes (though the frequency may vary) and quite often are staggered (2-3 events in a minute).  

The frequency appears to be unrelated to the system load or specific applications/activities. 
Though I never really checked, I have the _sense_ that the frequency may also increase in very low memory conditions.
 
Interestingly they almost never happen the first 10 minutes after boot.
So it may be a function of time.
Power-cycling the touchpad and recalibrating every 7 minutes appears to decrease the spontaneous miscalibration symptoms to ~1/hour.

Of course scheduled recalibrations are equally disruptive for the user. 
However, I think it may points to an underlaying problem that accumulates over time (leak of some sort ?). 

This may not be directly related to the new driver, but could be related to the EC code or the kernel configuration or something else, though I'm afraid I have no idea how to test further.