From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail203.messagelabs.com (mail203.messagelabs.com [216.82.254.243]) by kanga.kvack.org (Postfix) with ESMTP id B84E36B0023 for ; Fri, 28 Oct 2011 04:42:49 -0400 (EDT) Received: by ggnh4 with SMTP id h4so4499222ggn.14 for ; Fri, 28 Oct 2011 01:42:47 -0700 (PDT) MIME-Version: 1.0 In-Reply-To: References: <1319384922-29632-1-git-send-email-gilad@benyossef.com> <1319384922-29632-5-git-send-email-gilad@benyossef.com> Date: Fri, 28 Oct 2011 10:42:47 +0200 Message-ID: Subject: Re: [PATCH v2 4/6] mm: Only IPI CPUs to drain local pages if they exist From: Gilad Ben-Yossef Content-Type: multipart/alternative; boundary=bcaec52c5ea9a02f7d04b057df44 Sender: owner-linux-mm@kvack.org List-ID: To: Christoph Lameter Cc: linux-kernel@vger.kernel.org, Peter Zijlstra , Frederic Weisbecker , Russell King , linux-mm@kvack.org, Pekka Enberg , Matt Mackall , Sasha Levin --bcaec52c5ea9a02f7d04b057df44 Content-Type: text/plain; charset=ISO-8859-1 On Fri, Oct 28, 2011 at 6:10 AM, Christoph Lameter wrote: > On Sun, 23 Oct 2011, Gilad Ben-Yossef wrote: > > > +/* Which CPUs have per cpu pages */ > > +cpumask_var_t cpus_with_pcp; > > +static DEFINE_PER_CPU(unsigned long, total_cpu_pcp_count); > > This increases the cache footprint of a hot vm path. Is it possible to do > the same than what you did for slub? Run a loop over all zones when > draining to check for remaining pcp pages and build the set of cpus > needing IPIs temporarily while draining? > > Sounds like a good idea. I will give it a shot. Thanks, Gilad -- Gilad Ben-Yossef Chief Coffee Drinker gilad@benyossef.com Israel Cell: +972-52-8260388 US Cell: +1-973-8260388 http://benyossef.com "I've seen things you people wouldn't believe. Goto statements used to implement co-routines. I watched C structures being stored in registers. All those moments will be lost in time... like tears in rain... Time to die. " --bcaec52c5ea9a02f7d04b057df44 Content-Type: text/html; charset=ISO-8859-1 Content-Transfer-Encoding: quoted-printable

On Fri, Oct 28, 2011 at 6:10 AM, Christoph Lameter <cl@gentwo.org> wrote:
On Sun, 23 Oct 20= 11, Gilad Ben-Yossef wrote:

> +/* Which CPUs have per cpu pages =A0*/
> +cpumask_var_t cpus_with_pcp;
> +static DEFINE_PER_CPU(unsigned long, total_cpu_pcp_count);

This increases the cache footprint of a hot vm path. Is it possible t= o do
the same than what you did for slub? Run a loop over all zones when
draining to check for remaining pcp pages and build the set of cpus
needing IPIs temporarily while draining?


Sounds like a good idea. I will give i= t a shot.

Thanks,
Gilad=A0


--
Gilad Ben-Yossef
Chief Coffee = Drinker
gilad@benyossef.co= m
Israel Cell: +972-52-8260388
US Cell: +1-973-8260388
http://benyossef.com
"I've seen things you people wouldn't believe. Goto statement= s used to implement co-routines. I watched C structures being stored in reg= isters. All those moments will be lost in time... like tears in rain... Tim= e to die. "

--bcaec52c5ea9a02f7d04b057df44-- -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@kvack.org. For more info on Linux MM, see: http://www.linux-mm.org/ . Fight unfair telecom internet charges in Canada: sign http://stopthemeter.ca/ Don't email: email@kvack.org