| 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
| ||
|
Message-ID: <CAF2d9jgKfEmAhO8_FMARirQ8DRqdhTtZ0xxOXUorry3oqNK_hA@mail.gmail.com>
Date: Tue, 19 May 2015 16:33:21 -0700
From: Mahesh Bandewar <maheshb@...gle.com>
To: Konstantin Khlebnikov <khlebnikov@...dex-team.ru>
Cc: linux-netdev <netdev@...r.kernel.org>,
"David S. Miller" <davem@...emloft.net>,
Jiri Benc <jbenc@...hat.com>,
Hannes Frederic Sowa <hannes@...essinduktion.org>
Subject: Re: [PATCH 2/3] ipvlan: grab rcu_read_lock on xmit path
On Thu, May 14, 2015 at 6:56 AM, Konstantin Khlebnikov
<khlebnikov@...dex-team.ru> wrote:
>
> ipvlan_start_xmit() is called with rcu_read_lock_bh() while its internal
> structures requre normal rcu_read_lock().
>
> Signed-off-by: Konstantin Khlebnikov <khlebnikov@...dex-team.ru>
>
> ---
>
> [ 802.945151] ===============================
> [ 802.945160] [ INFO: suspicious RCU usage. ]
> [ 802.945164] 4.1.0-rc3+ #71 Not tainted
> [ 802.945165] -------------------------------
> [ 802.945167] drivers/net/ipvlan/ipvlan.h:103 suspicious rcu_dereference_check() usage!
> [ 802.945168]
> [ 802.945168] other info that might help us debug this:
> [ 802.945168]
> [ 802.945170]
> [ 802.945170] rcu_scheduler_active = 1, debug_locks = 1
> [ 802.945173] 3 locks held by ping6/3813:
> [ 802.945174] #0: (sk_lock-AF_INET6){+.+.+.}, at: [<ffffffff81880f72>] rawv6_sendmsg+0x512/0xbb0
> [ 802.945197] #1: (rcu_read_lock_bh){......}, at: [<ffffffff8185e577>] ip6_finish_output2+0x57/0x790
> [ 802.945205] #2: (rcu_read_lock_bh){......}, at: [<ffffffff8177a68b>] __dev_queue_xmit+0x4b/0x930
> [ 802.945218]
> [ 802.945218] stack backtrace:
> [ 802.945221] CPU: 2 PID: 3813 Comm: ping6 Not tainted 4.1.0-rc3+ #71
> [ 802.945222] Hardware name: OpenStack Foundation OpenStack Nova, BIOS Bochs 01/01/2011
> [ 802.945224] 0000000000000001 ffff8800db7b7888 ffffffff819de6f8 0000000000000007
> [ 802.945226] ffff8800db738000 ffff8800db7b78b8 ffffffff810a7f92 ffff880214fc8c00
> [ 802.945229] ffff88021595b000 ffff88021595a000 ffff880214adf800 ffff8800db7b7968
> [ 802.945232] Call Trace:
> [ 802.945248] [<ffffffff819de6f8>] dump_stack+0x4c/0x65
> [ 802.945253] [<ffffffff810a7f92>] lockdep_rcu_suspicious+0xe2/0x130
> [ 802.945265] [<ffffffff815e24ac>] ipvlan_queue_xmit+0x17c/0x5a0
> [ 802.945268] [<ffffffff810aa5a4>] ? __lock_is_held+0x54/0x70
> [ 802.945271] [<ffffffff815e300c>] ipvlan_start_xmit+0x1c/0x50
> [ 802.945272] [<ffffffff8177a117>] dev_hard_start_xmit+0x2f7/0x820
> [ 802.945274] [<ffffffff817797b6>] ? netif_skb_features+0xf6/0x1d0
> [ 802.945276] [<ffffffff81779b84>] ? validate_xmit_skb.isra.99.part.100+0x24/0x2c0
> [ 802.945278] [<ffffffff8177ad04>] __dev_queue_xmit+0x6c4/0x930
> [ 802.945280] [<ffffffff8177a68b>] ? __dev_queue_xmit+0x4b/0x930
> [ 802.945281] [<ffffffff810a963a>] ? mark_held_locks+0x6a/0x90
> [ 802.945283] [<ffffffff8177af8e>] dev_queue_xmit_sk+0xe/0x10
> [ 802.945285] [<ffffffff8185e824>] ip6_finish_output2+0x304/0x790
> [ 802.945287] [<ffffffff81860dde>] ? ip6_finish_output+0x9e/0x1e0
> [ 802.945288] [<ffffffff81860dde>] ip6_finish_output+0x9e/0x1e0
> [ 802.945290] [<ffffffff81860fdb>] ip6_output+0xbb/0x180
> [ 802.945302] [<ffffffff818a7d0a>] ? ip6_find_1stfragopt+0x9a/0xa0
> [ 802.945304] [<ffffffff81860d40>] ? ip6_fragment+0xe80/0xe80
> [ 802.945306] [<ffffffff818a805c>] ip6_local_out_sk+0x2c/0x70
> [ 802.945308] [<ffffffff818a80b0>] ip6_local_out+0x10/0x20
> [ 802.945309] [<ffffffff81861571>] ip6_send_skb+0x31/0xd0
> [ 802.945311] [<ffffffff81861644>] ip6_push_pending_frames+0x34/0x40
> [ 802.945313] [<ffffffff81881368>] rawv6_sendmsg+0x908/0xbb0
> [ 802.945328] [<ffffffff810aa5a4>] ? __lock_is_held+0x54/0x70
> [ 802.945340] [<ffffffff81817e9e>] inet_sendmsg+0x10e/0x1f0
> [ 802.945343] [<ffffffff81817d90>] ? inet_recvmsg+0x200/0x200
> [ 802.945351] [<ffffffff81758275>] sock_sendmsg+0x45/0x50
> [ 802.945354] [<ffffffff8175a719>] SYSC_sendto+0xd9/0x110
> [ 802.945357] [<ffffffff8175ac99>] SyS_sendto+0x9/0x10
> [ 802.945362] [<ffffffff819ebfae>] system_call_fastpath+0x12/0x76
> ---
> drivers/net/ipvlan/ipvlan_main.c | 4 ++++
> 1 file changed, 4 insertions(+)
>
> diff --git a/drivers/net/ipvlan/ipvlan_main.c b/drivers/net/ipvlan/ipvlan_main.c
> index 77b92a0fe557..0cafd3e6f02d 100644
> --- a/drivers/net/ipvlan/ipvlan_main.c
> +++ b/drivers/net/ipvlan/ipvlan_main.c
> @@ -173,6 +173,7 @@ static int ipvlan_stop(struct net_device *dev)
> return 0;
> }
>
> +/* called with rcu_read_lock_bh. */
> static netdev_tx_t ipvlan_start_xmit(struct sk_buff *skb,
> struct net_device *dev)
> {
> @@ -180,6 +181,7 @@ static netdev_tx_t ipvlan_start_xmit(struct sk_buff *skb,
> int skblen = skb->len;
> int ret;
>
> + rcu_read_lock();
I don't believe this is correct. The correct way would be the way it
is fixed in the following patch -
https://patchwork.ozlabs.org/patch/471481/
> ret = ipvlan_queue_xmit(skb, dev);
> if (likely(ret == NET_XMIT_SUCCESS || ret == NET_XMIT_CN)) {
> struct ipvl_pcpu_stats *pcptr;
> @@ -193,6 +195,8 @@ static netdev_tx_t ipvlan_start_xmit(struct sk_buff *skb,
> } else {
> this_cpu_inc(ipvlan->pcpu_stats->tx_drps);
> }
> + rcu_read_unlock();
> +
> return ret;
> }
>
>
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Powered by blists - more mailing lists