2021-07-16 01:39:48

by Wangshaobo (bobo)

[permalink] [raw]
Subject: Re: [PATCH] Bluetooth: fix use-after-free error in lock_sock_nested()


在 2021/7/15 12:57, Wangshaobo (bobo) 写道:
>
> 在 2021/7/15 5:50, Luiz Augusto von Dentz 写道:
>> Hi,
>>
>> On Tue, Jul 13, 2021 at 8:20 PM Wang ShaoBo
>> <[email protected]> wrote:
>>> use-after-free error in lock_sock_nested() is reported:
>>>
>>> [  179.140137][ T3731]
>>> =====================================================
>>> [  179.142675][ T3731] BUG: KMSAN: use-after-free in
>>> lock_sock_nested+0x280/0x2c0
>>> [  179.145494][ T3731] CPU: 4 PID: 3731 Comm: kworker/4:2 Not
>>> tainted 5.12.0-rc6+ #54
>>> [  179.148432][ T3731] Hardware name: QEMU Standard PC (i440FX +
>>> PIIX, 1996), BIOS 1.13.0-1ubuntu1.1 04/01/2014
>>> [  179.151806][ T3731] Workqueue: events l2cap_chan_timeout
>>> [  179.152730][ T3731] Call Trace:
>>> [  179.153301][ T3731]  dump_stack+0x24c/0x2e0
>>> [  179.154063][ T3731]  kmsan_report+0xfb/0x1e0
>>> [  179.154855][ T3731]  __msan_warning+0x5c/0xa0
>>> [  179.155579][ T3731]  lock_sock_nested+0x280/0x2c0
>>> [  179.156436][ T3731]  ? kmsan_get_metadata+0x116/0x180
>>> [  179.157257][ T3731]  l2cap_sock_teardown_cb+0xb8/0x890
>>> [  179.158154][ T3731]  ? __msan_metadata_ptr_for_load_8+0x10/0x20
>>> [  179.159141][ T3731]  ? kmsan_get_metadata+0x116/0x180
>>> [  179.159994][ T3731]  ? kmsan_get_shadow_origin_ptr+0x84/0xb0
>>> [  179.160959][ T3731]  ? l2cap_sock_recv_cb+0x420/0x420
>>> [  179.161834][ T3731]  l2cap_chan_del+0x3e1/0x1d50
>>> [  179.162608][ T3731]  ? kmsan_get_metadata+0x116/0x180
>>> [  179.163435][ T3731]  ? kmsan_get_shadow_origin_ptr+0x84/0xb0
>>> [  179.164406][ T3731]  l2cap_chan_close+0xeea/0x1050
>>> [  179.165189][ T3731]  ? kmsan_internal_unpoison_shadow+0x42/0x70
>>> [  179.166180][ T3731]  l2cap_chan_timeout+0x1da/0x590
>>> [  179.167066][ T3731]  ? __msan_metadata_ptr_for_load_8+0x10/0x20
>>> [  179.168023][ T3731]  ? l2cap_chan_create+0x560/0x560
>>> [  179.168818][ T3731]  process_one_work+0x121d/0x1ff0
>>> [  179.169598][ T3731]  worker_thread+0x121b/0x2370
>>> [  179.170346][ T3731]  kthread+0x4ef/0x610
>>> [  179.171010][ T3731]  ? process_one_work+0x1ff0/0x1ff0
>>> [  179.171828][ T3731]  ? kthread_blkcg+0x110/0x110
>>> [  179.172587][ T3731]  ret_from_fork+0x1f/0x30
>>> [  179.173348][ T3731]
>>> [  179.173752][ T3731] Uninit was created at:
>>> [  179.174409][ T3731]  kmsan_internal_poison_shadow+0x5c/0xf0
>>> [  179.175373][ T3731]  kmsan_slab_free+0x76/0xc0
>>> [  179.176060][ T3731]  kfree+0x3a5/0x1180
>>> [  179.176664][ T3731]  __sk_destruct+0x8af/0xb80
>>> [  179.177375][ T3731]  __sk_free+0x812/0x8c0
>>> [  179.178032][ T3731]  sk_free+0x97/0x130
>>> [  179.178686][ T3731]  l2cap_sock_release+0x3d5/0x4d0
>>> [  179.179457][ T3731]  sock_close+0x150/0x450
>>> [  179.180117][ T3731]  __fput+0x6bd/0xf00
>>> [  179.180787][ T3731]  ____fput+0x37/0x40
>>> [  179.181481][ T3731]  task_work_run+0x140/0x280
>>> [  179.182219][ T3731]  do_exit+0xe51/0x3e60
>>> [  179.182930][ T3731]  do_group_exit+0x20e/0x450
>>> [  179.183656][ T3731]  get_signal+0x2dfb/0x38f0
>>> [  179.184344][ T3731]  arch_do_signal_or_restart+0xaa/0xe10
>>> [  179.185266][ T3731]  exit_to_user_mode_prepare+0x2d2/0x560
>>> [  179.186136][ T3731]  syscall_exit_to_user_mode+0x35/0x60
>>> [  179.186984][ T3731]  do_syscall_64+0xc5/0x140
>>> [  179.187681][ T3731] entry_SYSCALL_64_after_hwframe+0x44/0xae
>>> [  179.188604][ T3731]
>>> =====================================================
>>>
>>> In our case, there are two Thread A and B:
>>>
>>> Context: Thread A:              Context: Thread B:
>>>
>>> l2cap_chan_timeout()            __se_sys_shutdown()
>>>    l2cap_chan_close()              l2cap_sock_shutdown()
>>>      l2cap_chan_del()                l2cap_chan_close()
>>>        l2cap_sock_teardown_cb() l2cap_sock_teardown_cb()
>>>
>>> Once l2cap_sock_teardown_cb() excuted, this sock will be marked as
>>> SOCK_ZAPPED,
>>> and can be treated as killable in l2cap_sock_kill() if sock_orphan()
>>> has
>>> excuted, at this time we close sock through sock_close() which end
>>> to call
>>> l2cap_sock_kill() like Thread C:
>>>
>>> Context: Thread C:
>>>
>>> sock_close()
>>>    l2cap_sock_release()
>>>      sock_orphan()
>>>      l2cap_sock_kill()  #free sock if refcnt is 1
>>>
>>> If C completed, Once A or B reaches l2cap_sock_teardown_cb() again,
>>> use-after-free happened.
>>>
>>> We should set chan->data to NULL if sock is freed, for telling teardown
>>> operation is not allowed in l2cap_sock_teardown_cb(), and also we
>>> should
>>> avoid killing an already killed socket in l2cap_sock_close_cb().
>>>
>>> Signed-off-by: Wang ShaoBo <[email protected]>
>>> ---
>>>   net/bluetooth/l2cap_sock.c | 14 ++++++++++++--
>>>   1 file changed, 12 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/net/bluetooth/l2cap_sock.c b/net/bluetooth/l2cap_sock.c
>>> index c99d65ef13b1..ddc6a692b237 100644
>>> --- a/net/bluetooth/l2cap_sock.c
>>> +++ b/net/bluetooth/l2cap_sock.c
>>> @@ -1215,14 +1215,18 @@ static int l2cap_sock_recvmsg(struct socket
>>> *sock, struct msghdr *msg,
>>>    */
>>>   static void l2cap_sock_kill(struct sock *sk)
>>>   {
>>> +       struct l2cap_chan *chan;
>>> +
>>>          if (!sock_flag(sk, SOCK_ZAPPED) || sk->sk_socket)
>>>                  return;
>>>
>>>          BT_DBG("sk %p state %s", sk, state_to_string(sk->sk_state));
>>>
>>>          /* Kill poor orphan */
>>> -
>>> -       l2cap_chan_put(l2cap_pi(sk)->chan);
>>> +       chan = l2cap_pi(sk)->chan;
>>> +       l2cap_chan_put(chan);
>
> There is a problem here, the above sentence `l2cap_chan_put(chan)`
> should put after
>
> following sentence.
>
>>> +       if (refcount_read(&sk->sk_refcnt) == 1)
>>> +               chan->data = NULL;
>> Instead of checking if it is the last reference here, wouldn't it be
>> better to reset the chan->data to NULL on l2cap_sock_destruct?
>
> Hi,
>
> In my case it looks OK, this is the diff:
>
> diff --git a/net/bluetooth/l2cap_sock.c b/net/bluetooth/l2cap_sock.c
> index f1b1edd0b697..32ef3328ab49 100644
> --- a/net/bluetooth/l2cap_sock.c
> +++ b/net/bluetooth/l2cap_sock.c
> @@ -1500,6 +1500,9 @@ static void l2cap_sock_close_cb(struct
> l2cap_chan *chan)
>  {
>         struct sock *sk = chan->data;
>
> +       if (!sk)
> +               return;
> +
>         l2cap_sock_kill(sk);
>  }
>
> @@ -1508,6 +1511,9 @@ static void l2cap_sock_teardown_cb(struct
> l2cap_chan *chan, int err)
>         struct sock *sk = chan->data;
>         struct sock *parent;
>
> +       if (!sk)
> +               return;
> +
>         BT_DBG("chan %p state %s", chan, state_to_string(chan->state));
>
>         /* This callback can be called both for server (BT_LISTEN)
> @@ -1700,6 +1706,7 @@ static void l2cap_sock_destruct(struct sock *sk)
>         BT_DBG("sk %p", sk);
>
>         if (l2cap_pi(sk)->chan)
> +              l2cap_pi(sk)->chan->data = NULL;
>                  l2cap_chan_put(l2cap_pi(sk)->chan);
>
> But if it has potential risk if l2cap_sock_destruct() can not be
> excuted in time ?
>
> sk_free():
>
>         if (refcount_dec_and_test(&sk->sk_wmem_alloc)) //is possible
> this condition false ?
>
>               __sk_free(sk)   -> ... l2cap_sock_destruct()
>
Dear Luiz,

Not only that, if l2cap_sock_kill() has put 'l2cap_pi(sk)->chan', how
does we avoid re-puting 'l2cap_pi(sk)->chan' if l2cap_sock_destruct()
work postponed? this will cause underflow of chan->refcount; this PATCH
4e1a720d0312 ("Bluetooth: avoid killing an already killed socket") also
may not work in any case because only sock_orphan() has excuted can this
sock be killed, but if sco_sock_release() excute first, for this sock
has been marked as SOCK_DEAD, this sock can never be killed. So should
we think put chan->data = NULL in xx_sock_kill() is a better choice ?

- WangShaoBo