Hi,
My calculation was wrong.
It is certain that (pseudo-)physical memory "256GB-" and
"-4TB" exits.
"256GB-" and "-16TB".
And it is always sparse actually...
So my patch always consumes large amount of memory...
It is not good. (though the patch is usefull actually now)
As I said:
I intended to modify as less existent code as possible.
But
this may be wrong.
I hope my patch helps to understand IA64's phys-to-machine
mapping structure...
Thanks.
--
Itsuro ODA <oda(a)valinux.co.jp>