On Mar 20, 2007, at 11:12 PM, Michael Lorenz wrote:

> Yes, sort of. In doubt it's what bus_space_mmap() returns ( and the  
> man page says 'physical address or opaque token' ). If at all then  
> vtophys should do The Right Thing and we shouldn't have to atop()  
> anything.

No, vtophys() will not necessarily do the right thing.  The only thing  
you can do with a bus_space_mmap() return value is call  
pmap_phys_address() with the token in order to map it to a paddr_t.

-- thorpej