re: dmesg and ofctl -p output as instructed

Martin Husemann writes:
> On Sun, Dec 20, 2015 at 10:38:45AM +0000, Julian Coleman wrote:
> > Agreed.  However, we'd need to know that we were attached to a psycho
> > controller in pci_intr_map(), and (as you mention) that's not easy to do
> > nicely [*].
> 
> Not nicely, but the same as you did in OF_mapintr should work?
> Or somehow identify pa->pa_pc->cookie?

IMO, this is more gross than the current proposal or putting
the first PCI controller name in a global.

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().

actually, perhaps the last is a fine idea.


.mrg.