[PATCH] ppc64 memory_present calls

[Joel Schopp]

> What about the actual memory_present() discontig implementation?  We'll
> need to push that ahead of sparsemem anyway, so we might as well do them
> at the same time.

DISCONTIG seems to work perfectly fine without it.  Do you mean moving 
some existing discontig code into something called memory_present() or 
defining memory_present to {} for the discontig case?  If it is the 
latter we seem pretty well covered by include/linux/mmzone.h