From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753626Ab2GIXNj (ORCPT ); Mon, 9 Jul 2012 19:13:39 -0400 Received: from mail.linuxfoundation.org ([140.211.169.12]:38640 "EHLO mail.linuxfoundation.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751510Ab2GIXNh (ORCPT ); Mon, 9 Jul 2012 19:13:37 -0400 Date: Mon, 9 Jul 2012 16:13:36 -0700 From: Andrew Morton To: Glauber Costa Cc: , , Greg Thelen , Serge Hallyn , Tejun Heo , "Eric W. Biederman" , Greg Kroah-Hartman Subject: Re: [PATCH v2] fail dentry revalidation after namespace change Message-Id: <20120709161336.0ec23592.akpm@linux-foundation.org> In-Reply-To: <1341565747-15374-1-git-send-email-glommer@parallels.com> References: <1341565747-15374-1-git-send-email-glommer@parallels.com> X-Mailer: Sylpheed 3.0.2 (GTK+ 2.20.1; x86_64-pc-linux-gnu) Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Fri, 6 Jul 2012 13:09:07 +0400 Glauber Costa wrote: > When we change the namespace tag of a sysfs entry, the associated dentry > is still kept around. readdir() will work correctly and not display the > old entries, but open() will still succeed, so will reads and writes. > > This will no longer happen if sysfs is remounted, hinting that this is a > cache-related problem. > > I am using the following sequence to demonstrate that: > > shell1: > ip link add type veth > unshare -nm > > shell2: > ip link set veth1 > cat /sys/devices/virtual/net/veth1/ifindex > > Before that patch, this will succeed (fail to fail). After it, it will > correctly return an error. Differently from a normal rename, which we > handle fine, changing the object namespace will keep it's path intact. > So this check seems necessary as well. > > ... > > --- a/fs/sysfs/dir.c > +++ b/fs/sysfs/dir.c > @@ -307,6 +307,7 @@ static int sysfs_dentry_revalidate(struct dentry *dentry, struct nameidata *nd) > { > struct sysfs_dirent *sd; > int is_dir; > + int type; > > if (nd->flags & LOOKUP_RCU) > return -ECHILD; > @@ -326,6 +327,13 @@ static int sysfs_dentry_revalidate(struct dentry *dentry, struct nameidata *nd) > if (strcmp(dentry->d_name.name, sd->s_name) != 0) > goto out_bad; > > + /* The sysfs dirent has been moved to a different namespace */ > + type = KOBJ_NS_TYPE_NONE; > + if (sd->s_parent) > + type = sysfs_ns_type(sd->s_parent); > + if (type && (sysfs_info(dentry->d_sb)->ns[type] != sd->s_ns)) eww, the code is assuming that KOBJ_NS_TYPE_NONE has a value of zero. Don't do that; it smells bad. I renamed my version of this patch to "sysfs: fail dentry revalidation after namespace change", as carefully explained in section 15 of the excellent Documentation/SubmittingPatches, then queued this: From: Andrew Morton Subject: sysfs-fail-dentry-revalidation-after-namespace-change-fix don't assume that KOBJ_NS_TYPE_NONE==0. Also save a test-n-branch. Cc: Eric W. Biederman Cc: Glauber Costa Cc: Greg Kroah-Hartman Cc: Tejun Heo Signed-off-by: Andrew Morton --- fs/sysfs/dir.c | 8 +++++--- 1 file changed, 5 insertions(+), 3 deletions(-) diff -puN fs/sysfs/dir.c~sysfs-fail-dentry-revalidation-after-namespace-change-fix fs/sysfs/dir.c --- a/fs/sysfs/dir.c~sysfs-fail-dentry-revalidation-after-namespace-change-fix +++ a/fs/sysfs/dir.c @@ -329,10 +329,12 @@ static int sysfs_dentry_revalidate(struc /* The sysfs dirent has been moved to a different namespace */ type = KOBJ_NS_TYPE_NONE; - if (sd->s_parent) + if (sd->s_parent) { type = sysfs_ns_type(sd->s_parent); - if (type && (sysfs_info(dentry->d_sb)->ns[type] != sd->s_ns)) - goto out_bad; + if (type != KOBJ_NS_TYPE_NONE && + sysfs_info(dentry->d_sb)->ns[type] != sd->s_ns) + goto out_bad; + } mutex_unlock(&sysfs_mutex); out_valid: _