Date: Wed, 21 Jun 2000 17:21:44 -0700
From: Richard Henderson <rth@twiddle.net>
> To the point, I'd like it if we all could agree to adopt
> something which addresses everyone's known concerns. Just like
> we did for the pci DMA mapping infrastructure. :-)
Indeed. Aside from the mmap64 issue, what else are you missing?
That's about it. In fact I think it's nicer if the user doesn't
need to know how big physical addresses are, nor anything about
them anyways, just to mmap device regions.
The PPC folks indicate they need to be able to tell the app some
kind of shift value for forming the address in load/store instruction
on the PCI I/O and MEM regions it maps. Aparently, this shift value
is different on different PPC machines which is why the app needs to
obtain it as opposed to just knowing when it is __powerpc__ what to
use.
Later,
David S. Miller
davem@redhat.com
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Fri Jun 23 2000 - 21:00:23 EST