netdev - Re: [PATCH net v2] tun: Check tun device queue status in tun_chr_write

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CACGkMEsXYAHTb40jbtr35=O2NgJHHNkC_E2b8bqxygrmLOtRbQ@mail.gmail.com>
Date:   Fri, 16 Sep 2022 09:56:45 +0800
From:   Jason Wang <jasowang@...hat.com>
To:     Liu Jian <liujian56@...wei.com>
Cc:     davem <davem@...emloft.net>, Eric Dumazet <edumazet@...gle.com>,
        Jakub Kicinski <kuba@...nel.org>,
        Paolo Abeni <pabeni@...hat.com>,
        Alexei Starovoitov <ast@...nel.org>,
        Daniel Borkmann <daniel@...earbox.net>,
        Jesper Dangaard Brouer <hawk@...nel.org>,
        John Fastabend <john.fastabend@...il.com>,
        netdev <netdev@...r.kernel.org>
Subject: Re: [PATCH net v2] tun: Check tun device queue status in tun_chr_write_iter

On Thu, Sep 15, 2022 at 8:34 PM Liu Jian <liujian56@...wei.com> wrote:
>
> syzbot found below warning:
>
> ------------[ cut here ]------------
> geneve0 received packet on queue 3, but number of RX queues is 3
> WARNING: CPU: 1 PID: 29734 at net/core/dev.c:4611 netif_get_rxqueue net/core/dev.c:4611 [inline]
> WARNING: CPU: 1 PID: 29734 at net/core/dev.c:4611 netif_receive_generic_xdp+0xb10/0xb50 net/core/dev.c:4683
> Modules linked in:
> CPU: 1 PID: 29734 Comm: syz-executor.0 Not tainted 5.10.0 #5
> Hardware name: linux,dummy-virt (DT)
> pstate: 60400005 (nZCv daif +PAN -UAO -TCO BTYPE=--)
> pc : netif_get_rxqueue net/core/dev.c:4611 [inline]
> pc : netif_receive_generic_xdp+0xb10/0xb50 net/core/dev.c:4683
> lr : netif_get_rxqueue net/core/dev.c:4611 [inline]
> lr : netif_receive_generic_xdp+0xb10/0xb50 net/core/dev.c:4683
> sp : ffffa00016127770
> x29: ffffa00016127770 x28: ffff3f4607d6acb4
> x27: ffff3f4607d6acb0 x26: ffff3f4607d6ad20
> x25: ffff3f461de3c000 x24: ffff3f4607d6ad28
> x23: ffffa00010059000 x22: ffff3f4608719100
> x21: 0000000000000003 x20: ffffa000161278a0
> x19: ffff3f4607d6ac40 x18: 0000000000000000
> x17: 0000000000000000 x16: 00000000f2f2f204
> x15: 00000000f2f20000 x14: 6465766965636572
> x13: 20306576656e6567 x12: ffff98b8ed3b924d
> x11: 1ffff8b8ed3b924c x10: ffff98b8ed3b924c
> x9 : ffffc5c76525c9c4 x8 : 0000000000000000
> x7 : 0000000000000001 x6 : ffff98b8ed3b924c
> x5 : ffff3f460f3b29c0 x4 : dfffa00000000000
> x3 : ffffc5c765000000 x2 : 0000000000000000
> x1 : 0000000000000000 x0 : ffff3f460f3b29c0
> Call trace:
>  netif_get_rxqueue net/core/dev.c:4611 [inline]
>  netif_receive_generic_xdp+0xb10/0xb50 net/core/dev.c:4683
>  do_xdp_generic net/core/dev.c:4777 [inline]
>  do_xdp_generic+0x9c/0x190 net/core/dev.c:4770
>  tun_get_user+0xd94/0x2010 drivers/net/tun.c:1938
>  tun_chr_write_iter+0x98/0x100 drivers/net/tun.c:2036
>  call_write_iter include/linux/fs.h:1960 [inline]
>  new_sync_write+0x260/0x370 fs/read_write.c:515
>  vfs_write+0x51c/0x61c fs/read_write.c:602
>  ksys_write+0xfc/0x200 fs/read_write.c:655
>  __do_sys_write fs/read_write.c:667 [inline]
>  __se_sys_write fs/read_write.c:664 [inline]
>  __arm64_sys_write+0x50/0x60 fs/read_write.c:664
>  __invoke_syscall arch/arm64/kernel/syscall.c:36 [inline]
>  invoke_syscall arch/arm64/kernel/syscall.c:48 [inline]
>  el0_svc_common.constprop.0+0xf4/0x414 arch/arm64/kernel/syscall.c:155
>  do_el0_svc+0x50/0x11c arch/arm64/kernel/syscall.c:217
>  el0_svc+0x20/0x30 arch/arm64/kernel/entry-common.c:353
>  el0_sync_handler+0xe4/0x1e0 arch/arm64/kernel/entry-common.c:369
>  el0_sync+0x148/0x180 arch/arm64/kernel/entry.S:683
>
> This is because the detached queue is used to send data. Therefore, we need
> to check the queue status in the tun_chr_write_iter function.
>
> Fixes: cde8b15f1aab ("tuntap: add ioctl to attach or detach a file form tuntap device")

Not sure this deserves a stable.

> Signed-off-by: Liu Jian <liujian56@...wei.com>
> ---
> v1->v2: add fixes tag
>  drivers/net/tun.c | 5 +++++
>  1 file changed, 5 insertions(+)
>
> diff --git a/drivers/net/tun.c b/drivers/net/tun.c
> index 259b2b84b2b3..261411c1a6bb 100644
> --- a/drivers/net/tun.c
> +++ b/drivers/net/tun.c
> @@ -2019,6 +2019,11 @@ static ssize_t tun_chr_write_iter(struct kiocb *iocb, struct iov_iter *from)
>         if (!tun)
>                 return -EBADFD;
>
> +       if (tfile->detached) {

tfile->detached is synchronized through rtnl_lock which is probably
not suitable for the datapath. We probably need to rcuify this.

> +               tun_put(tun);
> +               return -ENETDOWN;

Another question is that can some user space depend on this behaviour?
I wonder if it's more safe to pretend the packet was received here?

Thanks

> +       }
> +
>         if ((file->f_flags & O_NONBLOCK) || (iocb->ki_flags & IOCB_NOWAIT))
>                 noblock = 1;
>
> --
> 2.17.1
>