duncan.sands at math.u-psud.fr
Thu Oct 7 13:33:45 EDT 2004
> Heh, OK. We do seem to get 'down' interrupts though, just not 'up'
> interrupts -- so what I have there ought to leave polling on for the Rev
> 4 modems, while dropping back to interrupts for the older ones where it
> actually works.
> Or do you think we should base it on the revision rather than the
> behaviour -- or just leave the timer running always, even on the older
Why not just always use polling? That's what the windows drivers do by
the way. IIRC it was one of the linux guys who noticed experimentally
one day that this interrupt urb completes when the line state changes,
and decided that it would be a cool way of determining line state (I guess
he didn't like polling).
More information about the Usbatm