From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-qg0-f48.google.com (mail-qg0-f48.google.com [209.85.192.48]) by kanga.kvack.org (Postfix) with ESMTP id B30AA6B0035 for ; Wed, 28 May 2014 19:45:08 -0400 (EDT) Received: by mail-qg0-f48.google.com with SMTP id i50so19826178qgf.21 for ; Wed, 28 May 2014 16:45:08 -0700 (PDT) Received: from qmta05.emeryville.ca.mail.comcast.net (qmta05.emeryville.ca.mail.comcast.net. [2001:558:fe2d:43:76:96:30:48]) by mx.google.com with ESMTP id k10si22590269qaj.33.2014.05.28.16.45.07 for ; Wed, 28 May 2014 16:45:07 -0700 (PDT) Date: Wed, 28 May 2014 18:45:04 -0500 (CDT) From: Christoph Lameter Subject: Re: [PATCH] page_alloc: skip cpuset enforcement for lower zone allocations (v3) In-Reply-To: <20140528224324.GA1132@amt.cnet> Message-ID: References: <20140523193706.GA22854@amt.cnet> <20140526185344.GA19976@amt.cnet> <53858A06.8080507@huawei.com> <20140528224324.GA1132@amt.cnet> Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-linux-mm@kvack.org List-ID: To: Marcelo Tosatti Cc: Li Zefan , linux-mm@kvack.org, linux-kernel@vger.kernel.org, Lai Jiangshan , Mel Gorman , Tejun Heo , David Rientjes , Andrew Morton On Wed, 28 May 2014, Marcelo Tosatti wrote: > diff --git a/mm/page_alloc.c b/mm/page_alloc.c > index 5dba293..dfea3dc 100644 > --- a/mm/page_alloc.c > +++ b/mm/page_alloc.c > @@ -2698,6 +2698,7 @@ __alloc_pages_nodemask(gfp_t gfp_mask, unsigned int order, > unsigned int cpuset_mems_cookie; > int alloc_flags = ALLOC_WMARK_LOW|ALLOC_CPUSET|ALLOC_FAIR; > struct mem_cgroup *memcg = NULL; > + nodemask_t *cpuset_mems_allowed = &cpuset_current_mems_allowed; Why do you need this one? > gfp_mask &= gfp_allowed_mask; > > @@ -2726,9 +2727,14 @@ __alloc_pages_nodemask(gfp_t gfp_mask, unsigned int order, > retry_cpuset: > cpuset_mems_cookie = read_mems_allowed_begin(); > > +#ifdef CONFIG_NUMA > + if (gfp_zone(gfp_mask) < policy_zone) > + cpuset_mems_allowed = NULL; nodemask = &node_states[N_ONLINE]; > +#endif > + > /* The preferred zone is used for statistics later */ > first_zones_zonelist(zonelist, high_zoneidx, > - nodemask ? : &cpuset_current_mems_allowed, > + nodemask ? : cpuset_mems_allowed, Skip this? > &preferred_zone); > if (!preferred_zone) > goto out; > Why call __alloc_pages_nodemask at all if you want to skip the node handling? Punt to alloc_pages() -- 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/ . Don't email: email@kvack.org