Re: Long Format VHPT patches

From: Ian Wienand <ianw_at_gelato.unsw.edu.au>
Date: 2004-08-31 15:53:32
On Thu, Aug 26, 2004 at 01:19:49PM -0700, Christoph Lameter wrote:
> One issue that bothers me is that it is no longer possible to swap pte,
> pmd and pgd entries via cmpxchg. My page fault scalability patches are based
> on that ability. The Itanium can only do a cmpxchg with a 64 bit value.
> The long VHPT entries are 32 bytes long.

As others have said, for the initial patch this shouldn't be an issue
(we have started making sure that the lvhpt patches will apply ontop
of your patches too).

However, in the longer term the length a pte is.  To do anything
interesting with the long format vhpt we are going to need more bits
in a PTE for things like protection keys and page sizes.  We have some
preliminary patches at

http://www.gelato.unsw.edu.au/cgi-bin/viewcvs.cgi/cvs/kernel/pteextension/

which do this, and we are building ontop of this approach.  The
essence of the change is 

-typedef struct { unsigned long pte; } pte_t;
+typedef struct { unsigned long pte; unsigned long itir; } pte_t;

I would like to know from the experienced hackers what the chances of
getting a change like this in are, or if anyone has any other ideas.

-i

-
To unsubscribe from this list: send the line "unsubscribe linux-ia64" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Received on Tue Aug 31 01:54:00 2004

This archive was generated by hypermail 2.1.8 : 2005-08-02 09:20:30 EST