Re: dmesg and ofctl -p output as instructed

[Julian Coleman <jdc%coris.org.uk@localhost>]

Hi,

> > perhaps we can re-work the whole pci_intr_map() part of the pc
> > such that we can call it before OF_mapintr(), or perhaps add a
> > new vector to the pc to check for thiss, before calling
> > OF_mapintr().
> 
> I considered that, but a commented check based on the node passed to 
> OF_mapintr and doing the strcmp Julian suggested is probably good enough
> and a lot simpler.
> 
> Either is fine for me (as is cleaning up the "cookie" mess into an abstract
> interface with proper functions that could be used to retrieve the pbm/softc
> and where we could add a predicate for the test at hand; but that would be
> a lot of churn for purely aesthetical reasons - as long as we don't see 
> machines with different pci roots).

I think that we could always call pa->pa_pc->spc_find_ino() in pci_intr_map(),
do the fully-specified check for psycho, and call OF_mapintr() in all cases.
However, I'm not sure how to handle the EBUS case.  Perhaps we could add the
grandparent (PCI controller) node to ebus_softc, and then reference that in
ebus_find_ino()?  Alternatively, add a spc_find_ino() to ebus_softc and call
that like we do in the PCI case?

Regards,

J

-- 
   My other computer runs NetBSD too   -         http://www.netbsd.org/