[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <6AADFAC011213A4C87B956458587ADB4021FE16C@dggeml532-mbs.china.huawei.com>
Date: Tue, 30 Apr 2019 05:11:45 +0000
From: "weiyongjun (A)" <weiyongjun1@...wei.com>
To: Cong Wang <xiyou.wangcong@...il.com>,
"Michael S. Tsirkin" <mst@...hat.com>
CC: yuehaibing <yuehaibing@...wei.com>,
David Miller <davem@...emloft.net>,
Jason Wang <jasowang@...hat.com>,
Eric Dumazet <edumazet@...gle.com>,
"Jesper Dangaard Brouer" <brouer@...hat.com>,
"Li,Rongqing" <lirongqing@...du.com>,
Nicolas Dichtel <nicolas.dichtel@...nd.com>,
Chas Williams <3chas3@...il.com>,
"wangli39@...du.com" <wangli39@...du.com>,
LKML <linux-kernel@...r.kernel.org>,
Linux Kernel Network Developers <netdev@...r.kernel.org>
Subject: RE: [PATCH] tun: Fix use-after-free in tun_net_xmit
> Network Developers <netdev@...r.kernel.org>
> Subject: Re: [PATCH] tun: Fix use-after-free in tun_net_xmit
>
> On Mon, Apr 29, 2019 at 7:55 AM Michael S. Tsirkin <mst@...hat.com>
> wrote:
> > The problem seems real enough, but an extra synchronize_net on
> tun_attach
> > might be a problem, slowing guest startup significantly.
> > Better ideas?
>
> Yes, I proposed the following patch in the other thread.
>
> diff --git a/drivers/net/tun.c b/drivers/net/tun.c
> index e9ca1c088d0b..31c3210288cb 100644
> --- a/drivers/net/tun.c
> +++ b/drivers/net/tun.c
> @@ -3431,6 +3431,7 @@ static int tun_chr_open(struct inode *inode,
> struct file * file)
> file->private_data = tfile;
> INIT_LIST_HEAD(&tfile->next);
>
> + sock_set_flag(&tfile->sk, SOCK_RCU_FREE);
> sock_set_flag(&tfile->sk, SOCK_ZEROCOPY);
>
> return 0;
This patch should not work. The key point is that when detach the queue
with index is equal to tun->numqueues - 1, we do not clear the point
in tun->tfiles:
static void __tun_detach(...)
{
...
**** if index == tun->numqueues - 1, nothing changed ****
rcu_assign_pointer(tun->tfiles[index],
tun->tfiles[tun->numqueues - 1]);
....
}
And after tfile free, xmit have change to get and use the freed file point.
Regards
Powered by blists - more mailing lists