Playing with touch screen events

Andy Green andy at
Tue Oct 21 04:00:37 CEST 2008

Hash: SHA1

Somebody in the thread at some point said:
| Andy Green wrote:
|> Somebody in the thread at some point said:
|>> what we need is a better touchscreen.. or a very very good filter
|>> between the ts
|>> device and x events so every app gets reliable events. tslib i think
|>> has such
|>> an infrastructure - BUT i never got time to play with it. :)
|> I think these behaviours can also be trouble in the driver... "pen up
|> and down" is a separate interrupt than sample from ADC and it wouldn't
|> surprise me if we can improve it.  There was a patch about it from Dima
|> Kogan recently
| I can confirm that this fixes a lot of scrolling (and mouse moves at
| all) issues...

This reminds us that packaged kernel can be some distance away from our
current kernel in git, I guess it only makes sense to explore current
git kernel behaviours for touchscreen.

|> Actual position averaging is pretty solved by the "median" patch stuff,
|> there's some question about whether it's acceptable upstream I guess we
|> are going to have to find out about.
| I don't understand, is this usable in the moko or not? When you say
| upstream you mean a merging with the standard linux upstream code or
| with the openmoko stable?

I didn't put it on stable branch yet because of a thread where folks
were saying that upstream did not believe in any processing of
touchscreen data in the driver.  But, the driver is in a unique position
to reduce the amount of and increase the quality of information passed
to the input subsystem.

Since it's really effective and I naturally like it since I did it, I
would like to use it on our normal kernel, but it can conflict with this
~ get the kernel upstream meme.

- -Andy
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Using GnuPG with Fedora -


More information about the devel mailing list