From mboxrd@z Thu Jan 1 00:00:00 1970 From: Andrew Morton Subject: Re: - memory-cgroup-enhancements-add-pre_destroy-handler.patch removed from -mm tree Date: Mon, 5 Nov 2007 23:50:15 -0800 Message-ID: <20071105235015.eb968ad7.akpm@linux-foundation.org> References: <200711060721.lA67LTeo012008@imap1.linux-foundation.org> <20071106164447.98fa9032.kamezawa.hiroyu@jp.fujitsu.com> Reply-To: linux-kernel@vger.kernel.org Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Return-path: Received: from smtp2.linux-foundation.org ([207.189.120.14]:40487 "EHLO smtp2.linux-foundation.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751441AbXKFHvY (ORCPT ); Tue, 6 Nov 2007 02:51:24 -0500 In-Reply-To: <20071106164447.98fa9032.kamezawa.hiroyu@jp.fujitsu.com> Sender: mm-commits-owner@vger.kernel.org List-Id: mm-commits@vger.kernel.org To: KAMEZAWA Hiroyuki Cc: a.p.zijlstra@chello.nl, balbir@linux.vnet.ibm.com, dev@sw.ru, ebiederm@xmission.com, herbert@13thfloor.at, menage@google.com, nickpiggin@yahoo.com.au, rientjes@google.com, svaidy@linux.vnet.ibm.com, xemul@openvz.org, mm-commits@vger.kernel.org On Tue, 6 Nov 2007 16:44:47 +0900 KAMEZAWA Hiroyuki wrote: > On Mon, 05 Nov 2007 23:14:40 -0800 > akpm@linux-foundation.org wrote: > > > > > The patch titled > > memory cgroup enhancements: add pre_destroy handler > > has been removed from the -mm tree. Its filename was > > memory-cgroup-enhancements-add-pre_destroy-handler.patch > > > > This patch was dropped because I can't get it to compile - new version after next -mm, please > > > > Hmm, ok. I'll reschedule this and next one. Thanks. > But is there objections to pre_destroy handler ? Dunno, I haven't been paying much attention. Has anyone else? It's a bit suspicious that nobody noticed that this would cause obvious build errors. More reviewing effort please, guys. > If we should deal with reference count of following in another way, > == > page->page_cgroup->memory_cgroup, memory_cgroup contains css. > page_cgroup has ref to css. > == > I'll consider something. (for example, add refcnt to memory_cgroup itself.) > > Changing above may allow us to guarantee that if no task, css's refcnt is 0.