From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752921Ab2I3IYN (ORCPT ); Sun, 30 Sep 2012 04:24:13 -0400 Received: from mail-pb0-f46.google.com ([209.85.160.46]:56715 "EHLO mail-pb0-f46.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752761Ab2I3IYI (ORCPT ); Sun, 30 Sep 2012 04:24:08 -0400 Date: Sun, 30 Sep 2012 17:23:58 +0900 From: Tejun Heo To: Glauber Costa Cc: Mel Gorman , Michal Hocko , linux-kernel@vger.kernel.org, cgroups@vger.kernel.org, kamezawa.hiroyu@jp.fujitsu.com, devel@openvz.org, linux-mm@kvack.org, Suleiman Souhlal , Frederic Weisbecker , David Rientjes , Johannes Weiner Subject: Re: [PATCH v3 04/13] kmem accounting basic infrastructure Message-ID: <20120930082358.GG10383@mtj.dyndns.org> References: <50635B9D.8020205@parallels.com> <20120926195648.GA20342@google.com> <50635F46.7000700@parallels.com> <20120926201629.GB20342@google.com> <50637298.2090904@parallels.com> <20120927120806.GA29104@dhcp22.suse.cz> <20120927143300.GA4251@mtj.dyndns.org> <20120927144307.GH3429@suse.de> <20120927145802.GC4251@mtj.dyndns.org> <50649B4C.8000208@parallels.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <50649B4C.8000208@parallels.com> User-Agent: Mutt/1.5.21 (2010-09-15) Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Hello, Glauber. On Thu, Sep 27, 2012 at 10:30:36PM +0400, Glauber Costa wrote: > > But that happens only when pages enter and leave slab and if it still > > is significant, we can try to further optimize charging. Given that > > this is only for cases where memcg is already in use and we provide a > > switch to disable it globally, I really don't think this warrants > > implementing fully hierarchy configuration. > > Not totally true. We still have to match every allocation to the right > cache, and that is actually our heaviest hit, responsible for the 2, 3 % > we're seeing when this is enabled. It is the kind of path so hot that > people frown upon branches being added, so I don't think we'll ever get > this close to being free. Sure, depening on workload, any addition to alloc/free could be noticeable. I don't know. I'll write more about it when replying to Michal's message. BTW, __memcg_kmem_get_cache() does seem a bit heavy. I wonder whether indexing from cache side would make it cheaper? e.g. something like the following. kmem_cache *__memcg_kmem_get_cache(cachep, gfp) { struct kmem_cache *c; c = cachep->memcg_params->caches[percpu_read(kmemcg_slab_idx)]; if (likely(c)) return c; /* try to create and then fall back to cachep */ } where kmemcg_slab_idx is updated from sched notifier (or maybe add and use current->kmemcg_slab_idx?). You would still need __GFP_* and in_interrupt() tests but current->mm and PF_KTHREAD tests can be rolled into index selection. Thanks. -- tejun