Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <20221119075615.723290-1-syoshida@redhat.com>
In-Reply-To: <20221119075615.723290-1-syoshida@redhat.com>
From:   Eric Dumazet <edumazet@google.com>
Date:   Sat, 19 Nov 2022 10:31:38 -0800
Message-ID: <CANn89iJaCRMb-vSrBOV_zbjxq8Gpg7K3d031AECmEqSN-XWpkA@mail.gmail.com>
Subject: Re: [PATCH] net: tun: Fix use-after-free in tun_detach()
To:     Shigeru Yoshida <syoshida@redhat.com>
Cc:     davem@davemloft.net, kuba@kernel.org, pabeni@redhat.com,
        netdev@vger.kernel.org, linux-kernel@vger.kernel.org,
        syzkaller-bugs@googlegroups.com
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Fri, Nov 18, 2022 at 11:56 PM Shigeru Yoshida <syoshida@redhat.com> wrote:
>
> syzbot reported use-after-free in tun_detach() [1].  This causes call
> trace like below:
>
> ==================================================================
> BUG: KASAN: use-after-free in notifier_call_chain+0x1da/0x1e0
> ...
> Call Trace:

Please include a symbolic stack trace, I think syzbot has them.

>  <TASK>
>  dump_stack_lvl+0x100/0x178
>  print_report+0x167/0x470
>  ? __virt_addr_valid+0x5e/0x2d0
>  ? __phys_addr+0xc6/0x140
>  ? notifier_call_chain+0x1da/0x1e0
>  ? notifier_call_chain+0x1da/0x1e0
>  kasan_report+0xbf/0x1e0
>  ? notifier_call_chain+0x1da/0x1e0
>  notifier_call_chain+0x1da/0x1e0
>  call_netdevice_notifiers_info+0x83/0x130
>  netdev_run_todo+0xc33/0x11b0
>  ? generic_xdp_install+0x490/0x490
>  ? __tun_detach+0x1500/0x1500
>  tun_chr_close+0xe2/0x190
>  __fput+0x26a/0xa40
>  task_work_run+0x14d/0x240
>  ? task_work_cancel+0x30/0x30
>  do_exit+0xb31/0x2a40
>  ? reacquire_held_locks+0x4a0/0x4a0
>  ? do_raw_spin_lock+0x12e/0x2b0
>  ? mm_update_next_owner+0x7c0/0x7c0
>  ? rwlock_bug.part.0+0x90/0x90
>  ? lockdep_hardirqs_on_prepare+0x17f/0x410
>  do_group_exit+0xd4/0x2a0
>  __x64_sys_exit_group+0x3e/0x50
>  do_syscall_64+0x38/0xb0
>  entry_SYSCALL_64_after_hwframe+0x63/0xcd
>
> The cause of the issue is that sock_put() from __tun_detach() drops
> last reference count for struct net, and then notifier_call_chain()
> from netdev_state_change() accesses that struct net.
>
> This patch fixes the issue by calling sock_put() from tun_detach()
> after all necessary accesses for the struct net has done.
>
> Link: https://syzkaller.appspot.com/bug?id=96eb7f1ce75ef933697f24eeab928c4a716edefe [1]
> Signed-off-by: Shigeru Yoshida <syoshida@redhat.com>

Please add a Fixes: tag, once you identified which commit added this bug.

> ---
>  drivers/net/tun.c | 6 +++++-
>  1 file changed, 5 insertions(+), 1 deletion(-)
>
> diff --git a/drivers/net/tun.c b/drivers/net/tun.c
> index 7a3ab3427369..ce9fcf4c8ef4 100644
> --- a/drivers/net/tun.c
> +++ b/drivers/net/tun.c
> @@ -686,7 +686,6 @@ static void __tun_detach(struct tun_file *tfile, bool clean)
>                 if (tun)
>                         xdp_rxq_info_unreg(&tfile->xdp_rxq);
>                 ptr_ring_cleanup(&tfile->tx_ring, tun_ptr_free);
> -               sock_put(&tfile->sk);
>         }
>  }
>
> @@ -702,6 +701,11 @@ static void tun_detach(struct tun_file *tfile, bool clean)
>         if (dev)
>                 netdev_state_change(dev);
>         rtnl_unlock();
> +
> +       if (clean) {
> +               synchronize_rcu();
> +               sock_put(&tfile->sk);
> +       }
>  }
>
>  static void tun_detach_all(struct net_device *dev)
> --
> 2.38.1
>