ati_remote module responding incorrectly

[bjsnider]

Binary package hint: linux-image-2.6.15-25-k7

With this kernel booted, the ATI Remote Wonder fails to operate correctly. In all previous kernels, the hardware worked fine. In this case, the keys no longer produce a single event, but that event repeated 2-5 times. So, when I press the numerical 5 key, what i get is 555 or 5555 instead of 5.

[Gaëtan Petit (gaetanp)]

i can confirm this
but it appear for a long time
i notice this in breezy

Mapping is also incorrect 1 return : &&&&
2 return : éééé
3 return : """""

etc etc.

I set this to confirmed

[Gaëtan Petit (gaetanp)]

sorry i use both 386 and 686-smp

[Chris Jones (cmsj)]

This is caused by the change from HZ=1000 to HZ=250 - the driver blindly assumes that HZ is 1000.
There is a patch against 2.6.17 at: http://lkml.org/lkml/2006/7/2/89 but it does not cleanly apply to 2.6.15. i will have a look at porting it.

[bjsnider (brandonjsnider-deactivatedaccount)]

This may sound simplistic, but why not just go back to the prevous version of the module, the one that has worked up until now?

[Chris Jones (cmsj)]

Because, if I'm right, this is not due to a change in the module, but a global change in the kernel (ie dropping the default timer from 1000HZ to 250HZ), so one either puts that change back (which hurts laptop people) or fix the driver to behave correctly instead of having it just happen to work if you get certain settings right.

[Chris Jones (cmsj)]

Ok, the patch seems to port pretty easily (but I have to concede that I'm fumbling around things I don't fully understand) and from my initial testing seems to work absolutely fine.
I will attach it to this bug shortly.

[Chris Jones (cmsj)]

Backported to 2.6.15 from patch posted to lkml by Marko Macek against 2.6.17 (so their patch may change subsequently)

[Ben Collins (ben-collins)]

Patch was reversed, but applied. Thanks