From mboxrd@z Thu Jan 1 00:00:00 1970 From: Daisuke Nishimura Subject: Re: [RFC][PATCH] another swap controller for cgroup Date: Mon, 17 Mar 2008 17:15:16 +0900 Message-ID: <47DE2894.6010306@mxp.nes.nec.co.jp> References: <20080317020407.8512E1E7995@siro.lan> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <20080317020407.8512E1E7995-Pcsii4f/SVk@public.gmane.org> List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: containers-bounces-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org Errors-To: containers-bounces-cunTk1MwBs9QetFLy7KEm3xJsTq8ys+cHZ5vskTnxNA@public.gmane.org To: yamamoto-jCdQPDEk3idL9jVzuh4AOg@public.gmane.org Cc: minoura-jCdQPDEk3idL9jVzuh4AOg@public.gmane.org, Linux MM , Linux Containers , Hugh Dickins , Balbir Singh List-Id: containers.vger.kernel.org Hi, Yamamoto-san. I'm reviewing and testing your patch now. I think your implementation is better because: - the group to be charged is determined correctly at the point of swapout, without fixing the behavior of move_task of memcg. (I think the behavior of move_task of memcg should be fixed anyway.) - the group to be uncharged is remembered in page struct of pmd, so there is no need to add array of pointers to swap_info_struct. > - anonymous objects (shmem) are not accounted. IMHO, shmem should be accounted. I agree it's difficult in your implementation, but are you going to support it? Thanks, Daisuke Nishimura.