[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <alpine.DEB.2.00.0912150836190.16754@router.home>
Date: Tue, 15 Dec 2009 08:46:37 -0600 (CST)
From: Christoph Lameter <cl@...ux-foundation.org>
To: Eric Dumazet <eric.dumazet@...il.com>
cc: Tejun Heo <tj@...nel.org>, linux-kernel@...r.kernel.org,
Pekka Enberg <penberg@...helsinki.fi>,
Mel Gorman <mel@....ul.ie>,
Mathieu Desnoyers <mathieu.desnoyers@...ymtl.ca>
Subject: Re: [this_cpu_xx V7 6/8] Make slub statistics use this_cpu_inc
On Tue, 15 Dec 2009, Eric Dumazet wrote:
> Le 14/12/2009 23:03, Christoph Lameter a écrit :
> > this_cpu_inc() translates into a single instruction on x86 and does not
> > need any register. So use it in stat(). We also want to avoid the
> > calculation of the per cpu kmem_cache_cpu structure pointer. So pass
> > a kmem_cache pointer instead of a kmem_cache_cpu pointer.
> >
>
> Hmm, last time I checked, [__]this_cpu_inc were not using "inc" on x86,
> but the default [__]this_cpu_add((pcp), 1)
Well the add 1 is still a single instruction so its okay. We can easily
add support for inc on x86 if makes a difference?
Powered by blists - more mailing lists