[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <Pine.LNX.4.64.0709181212380.3513@schroedinger.engr.sgi.com>
Date: Tue, 18 Sep 2007 12:30:20 -0700 (PDT)
From: Christoph Lameter <clameter@....com>
To: Andi Kleen <ak@...e.de>
cc: Lee Schermerhorn <Lee.Schermerhorn@...com>,
linux-kernel <linux-kernel@...r.kernel.org>,
akpm@...ux-foundation.org, eric.whitney@...com,
Mel Gorman <mel@....ul.ie>
Subject: Re: [PATCH] 2.6.23-rc6: Fix NUMA Memory Policy Reference Counting
On Tue, 18 Sep 2007, Andi Kleen wrote:
> > The patch does require concurrent increments and decrements in the main
> > fault patch. The potential is to create another bouncing cacheline for
> > concurrent faults. This looks like it would cause a performance issue.
>
> While may be true correctness is always more important than performance.
> So I think this is the right thing for .23. Any performance improvements
> if needed can come later.
Is there any real issue as a result of the refcounting issues? It seems
that performance improvements are not possible unless one would modify how
memory policies work using RCU or so.
Ok the effect is limited. get_vma_policy() only increments refcounts
for foreign or shared policies. So it seems that it is fine if one stays
away from shmem.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists