Re: new procfs memory analysis feature

[Paul Cameron Davies]

On Fri, 8 Dec 2006, Jeremy Fitzhardinge wrote:

> I looked at implementing linear pagetable mappings for x86 as a way of
> getting rid of CONFIG_HIGHPTE, and to make pagetable manipulations
> generally more efficient.  I gave up on it after a while because all the
> existing pagetable accessors are not suitable for a linear pagetable,
> and I didn't want to have to introduce a pile of new pagetable
> interfaces.  Would the PTI interface be helpful for this?

Yes.  The PTI is a useful vehicle for experimentation with page tables.
The PTI has two components.  The first component provides for architectural
and implementation independent page table access.  The second component
provides for architecture dependendent access, but I have only done this 
for IA64.  However, abstracting out the page table implementation for the 
arch dependent stuff on x86 would enable experimentation with
implementing linear page table mappings for x86, while leaving
the current implementation in place as an alternative page table.

Cheers

Paul Davies
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/