Sun has consistently fouled up ebus/pci implementations. Be very careful
because the prom changes have changed how this is interpreted at least twice.


On Thu, 18 Oct 2001, matthew green wrote:

>    
>    (Sun deserves a brick to the head for this one -- this seems clearly
>    out of line with the PCI spec.)
> 
> 
> 
> this will be an artifact of the way the "hme" ic is implemented,
> (in PCIO) as function 1 of a dual-function PCI ic.  obviously
> the ebus on this ic isn't active .... there is no node.
> 
> 
> .mrg.
>