From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ingo Molnar Subject: Re: [PATCH] cgroup: Fix root_count when mount fails due to busy subsystem Date: Fri, 23 Jan 2009 11:22:53 +0100 Message-ID: <20090123102253.GF15188@elte.hu> References: <20090123004703.25103.29754.stgit@menage.corp.google.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Content-Disposition: inline In-Reply-To: <20090123004703.25103.29754.stgit@menage.corp.google.com> Sender: linux-kernel-owner@vger.kernel.org To: Paul Menage Cc: akpm@linux-foundation.org, serue@us.ibm.com, linux-kernel@vger.kernel.org, containers@lists.osdl.org, Peter Zijlstra List-Id: containers.vger.kernel.org * Paul Menage wrote: > cgroup: Fix root_count when mount fails due to busy subsystem > > root_count was being incremented in cgroup_get_sb() after all error > checking was complete, but decremented in cgroup_kill_sb(), which can be > called on a superblock that we gave up on due to an error. This patch > changes cgroup_kill_sb() to only decrement root_count if the root was > previously linked into the list of roots. > > Signed-off-by: Paul Menage > > --- > > I was actually surprised to find that list_del() doesn't crash when > run on an unattached list_head structure. > > kernel/cgroup.c | 6 ++++-- > 1 files changed, 4 insertions(+), 2 deletions(-) > > diff --git a/kernel/cgroup.c b/kernel/cgroup.c > index adcd0bb..9ce27e8 100644 > --- a/kernel/cgroup.c > +++ b/kernel/cgroup.c > @@ -1115,8 +1115,10 @@ static void cgroup_kill_sb(struct super_block *sb) { > } > write_unlock(&css_set_lock); > > - list_del(&root->root_list); > - root_count--; > + if (!list_empty(&root->root_list)) { > + list_del(&root->root_list); > + root_count--; > + } That's ugly. It is _much_ cleaner to always keep the link head consistent - i.e. initialize it with INIT_LIST_HEAD() and then remove from it via list_del_init(). That way the error path will do the right thing automatically, and there's no need for that ugly "if !list_empty" construct either. Ingo