On Fri, 16 Jun 2000, Greywolf wrote:

> On Fri, 16 Jun 2000, matthew green wrote:
> 
> #    ...we have a _working_ hme driver now?
> #    
> #    That's WAY cool.
> # 
> # 
> # it works for sbus only.  PCI is still busted.
> 
> Are we having problems talking to the PCI bus in general?  That's what
> it seems like to me.  The PCI layout must be hella weird if we can't talk
> to the PCI stuff which works in other places...

Sun does use custom PCI controllers and relies heavily on the OBP for
device probing, so it's not at straight-forward as just another intel
box.  Then there's the endianness issue, and the IOMMU and DVMA....

Eduardo Horvath