On Fri, Nov 23, 2007 at 08:19:06PM +0100, Matthias Drochner wrote:
> -aprint_error() conversions -- this looks harmless
> -struct device -> device_t -- looks harmless too, and even
>  if there was a mistake it should rather cause a panic
> -callout_reset() -> callout_schedule() -- at a first glance
>  this looks correct too

The change for callout are strictly equivalent. I can't find anything
wrong with the rest either. Some cosmetic issues, but nothing that would
explain your issues. The driver works fine for me.

Can you check if wpi_intr is spinning?

Joerg