diff mbox

[net] bridge: netlink: fix slave_changelink/br_setport race conditions

Message ID 1437563020-23506-1-git-send-email-razor@blackwall.org
State Accepted, archived
Delegated to: David Miller
Headers show

Commit Message

Nikolay Aleksandrov July 22, 2015, 11:03 a.m. UTC 
From: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>

Since slave_changelink support was added there have been a few race
conditions when using br_setport() since some of the port functions it
uses require the bridge lock. It is very easy to trigger a lockup due to
some internal spin_lock() usage without bh disabled, also it's possible to
get the bridge into an inconsistent state.

Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
Fixes: 3ac636b8591c ("bridge: implement rtnl_link_ops->slave_changelink")
---
 net/bridge/br_netlink.c | 10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)

Comments

Jiri Pirko July 22, 2015, 12:52 p.m. UTC | #1 
Wed, Jul 22, 2015 at 01:03:40PM CEST, razor@blackwall.org wrote:
>From: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
>
>Since slave_changelink support was added there have been a few race
>conditions when using br_setport() since some of the port functions it
>uses require the bridge lock. It is very easy to trigger a lockup due to
>some internal spin_lock() usage without bh disabled, also it's possible to
>get the bridge into an inconsistent state.
>
>Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
>Fixes: 3ac636b8591c ("bridge: implement rtnl_link_ops->slave_changelink")

Looks fine to me, thanks for this fix.

Reviewed-by: Jiri Pirko <jiri@resnulli.us>
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
David Miller July 26, 2015, 11:28 p.m. UTC | #2 
From: Nikolay Aleksandrov <razor@blackwall.org>
Date: Wed, 22 Jul 2015 13:03:40 +0200

> From: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
> 
> Since slave_changelink support was added there have been a few race
> conditions when using br_setport() since some of the port functions it
> uses require the bridge lock. It is very easy to trigger a lockup due to
> some internal spin_lock() usage without bh disabled, also it's possible to
> get the bridge into an inconsistent state.
> 
> Signed-off-by: Nikolay Aleksandrov <nikolay@cumulusnetworks.com>
> Fixes: 3ac636b8591c ("bridge: implement rtnl_link_ops->slave_changelink")

Applied and queued up for -stable, thanks.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
diff mbox

Patch

diff --git a/net/bridge/br_netlink.c b/net/bridge/br_netlink.c
index 364bdc98bd9b..3da5525eb8a2 100644
--- a/net/bridge/br_netlink.c
+++ b/net/bridge/br_netlink.c
@@ -693,9 +693,17 @@  static int br_port_slave_changelink(struct net_device *brdev,
 				    struct nlattr *tb[],
 				    struct nlattr *data[])
 {
+	struct net_bridge *br = netdev_priv(brdev);
+	int ret;
+
 	if (!data)
 		return 0;
-	return br_setport(br_port_get_rtnl(dev), data);
+
+	spin_lock_bh(&br->lock);
+	ret = br_setport(br_port_get_rtnl(dev), data);
+	spin_unlock_bh(&br->lock);
+
+	return ret;
 }
 
 static int br_port_fill_slave_info(struct sk_buff *skb,