From mboxrd@z Thu Jan 1 00:00:00 1970 From: Glauber Costa Subject: Re: [RFC] slub: show dead memcg caches in a separate file Date: Tue, 8 May 2012 08:55:14 -0300 Message-ID: <4FA909A2.2010203@parallels.com> References: <1336070841-1071-1-git-send-email-glommer@parallels.com> <4FA89348.6070000@parallels.com> Mime-Version: 1.0 Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: Sender: cgroups-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org List-ID: Content-Type: text/plain; charset="us-ascii"; format="flowed" To: Pekka Enberg Cc: Suleiman Souhlal , cgroups-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, linux-mm-Bw31MaZKKs3YtjvyW6yDsg@public.gmane.org, devel-GEFAQzZX7r8dnm+yROfE0A@public.gmane.org, Christoph Lameter , Michal Hocko , Kamezawa Hiroyuki , Johannes Weiner , Frederic Weisbecker On 05/08/2012 02:42 AM, Pekka Enberg wrote: > On Tue, May 8, 2012 at 6:30 AM, Glauber Costa wrote: >> But there is another aspect: those dead caches have one thing in common, >> which is the fact that no new objects will ever be allocated on them. You >> can't tune them, or do anything with them. I believe it is misleading to >> include them in slabinfo. >> >> The fact that the caches change names - to append "dead" may also break >> tools, if that is what you are concerned about. >> >> For all the above, I think a better semantics for slabinfo is to include the >> active caches, and leave the dead ones somewhere else. > > Can these "dead caches" still hold on to physical memory? If so, they > must appear in /proc/slabinfo. Yes, if they didn't, I would show them nowhere, instead of in a separate file. But okay, that's why I sent a separate RFC for that part. I will revert this behavior.