Re: Adding an interrupt locator to the GPIO, I2C and SPI buses

[Manuel Bouyer <bouyer%antioche.eu.org@localhost>]

On Fri, Apr 19, 2013 at 06:22:58PM +0100, David Laight wrote:
> On Fri, Apr 19, 2013 at 10:01:46AM +0200, Manuel Bouyer wrote:
> > 
> > That would be a problem if the GPIO module has less interrupt lines than
> > interrupt sources (as in the AM335x: 32 GPIO per module but only
> > 2 interrupt lines to the PIC): 
> > the interrupt would then be shared and drivers would have to deal
> > with that, eventually with callback functions to the gpio driver
> > to know which pin really generated the interrupt.
> 
> I'd have thought the generic code ought to allow for interrupt
> multiplexors - so the driver for a gpio block can attach its own
> interrupt hander to the cpu (or whatever its interrupt line is
> connected to) and then create a whole load of extra interrupt
> sources that other drivers can attach to.

that would be nice (and I would have use for other cases than this one),
but we're not close to this yet. This would need an MI interface and
change to all implementations, this needs carefull design.

I think an interface specific to GPIO would be good enough for the time
being, as this seems the common case (on the beaglebone there's also
various devices interrupting via GPIOs).

-- 
Manuel Bouyer <bouyer%antioche.eu.org@localhost>
     NetBSD: 26 ans d'experience feront toujours la difference
--