From mboxrd@z Thu Jan 1 00:00:00 1970 From: Stefan Hajnoczi Subject: Re: [PATCH 3.8-rc] tuntap: refuse to re-attach to different tun_struct Date: Thu, 10 Jan 2013 12:53:12 +0100 Message-ID: <20130110115312.GC30885@stefanha-thinkpad.redhat.com> References: <1357804788-19976-1-git-send-email-stefanha@redhat.com> <20130110102319.GB13451@redhat.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Cc: "David S. Miller" , Jason Wang , netdev@vger.kernel.org, linux-kernel@vger.kernel.org To: "Michael S. Tsirkin" Return-path: Content-Disposition: inline In-Reply-To: <20130110102319.GB13451@redhat.com> Sender: linux-kernel-owner@vger.kernel.org List-Id: netdev.vger.kernel.org On Thu, Jan 10, 2013 at 12:23:19PM +0200, Michael S. Tsirkin wrote: > On Thu, Jan 10, 2013 at 08:59:48AM +0100, Stefan Hajnoczi wrote: > > Multiqueue tun devices support detaching a tun_file from its tun_struct > > and re-attaching at a later point in time. This allows users to disable > > a specific queue temporarily. > > > > ioctl(TUNSETIFF) allows the user to specify the network interface to > > attach by name. This means the user can attempt to attach to interface > > "B" after detaching from interface "A". > > > > The driver is not designed to support this so check we are re-attaching > > to the right tun_struct. Failure to do so may lead to oops. > > > > Signed-off-by: Stefan Hajnoczi > > --- > > This fix is for 3.8-rc. > > > > drivers/net/tun.c | 2 ++ > > 1 file changed, 2 insertions(+) > > > > diff --git a/drivers/net/tun.c b/drivers/net/tun.c > > index fbd106e..cf6da6e 100644 > > --- a/drivers/net/tun.c > > +++ b/drivers/net/tun.c > > @@ -491,6 +491,8 @@ static int tun_attach(struct tun_struct *tun, struct file *file) > > err = -EINVAL; > > if (rcu_dereference_protected(tfile->tun, lockdep_rtnl_is_held())) > > goto out; > > + if (tfile->detached && tun != tfile->detached) > > + goto out; > > > > err = -EBUSY; > > if (!(tun->flags & TUN_TAP_MQ) && tun->numqueues == 1) > > -- > > 1.8.0.2 > > > I agree this is a bug but even with this patch, we still allow: > > SETIFF > SETQUEUE (DISABLED) > SETIFF > > Originally the rule always was that repeated setiff calls fail with > EINVAL. We broke that when we set tun to NULL. It's probably worth > preserving that, even if queue is disabled. Applying something like the below > instead will address this concern, won't it? Sounds good. Stefan