From mboxrd@z Thu Jan  1 00:00:00 1970
From: Patrick McHardy <kaber@trash.net>
Subject: [IPV4]: route: fix crash ip_route_input
Date: Wed, 06 Feb 2008 14:36:57 +0100
Message-ID: <47A9B7F9.3090808@trash.net>
Mime-Version: 1.0
Content-Type: multipart/mixed;
 boundary="------------060405040509010209010802"
Cc: Netfilter Development Mailinglist
	<netfilter-devel@vger.kernel.org>,
	"Denis V. Lunev" <den@openvz.org>,
	Linux Netdev List <netdev@vger.kernel.org>
To: "David S. Miller" <davem@davemloft.net>
Return-path: <netfilter-devel-owner@vger.kernel.org>
Sender: netfilter-devel-owner@vger.kernel.org
List-Id: netdev.vger.kernel.org

This is a multi-part message in MIME format.
--------------060405040509010209010802
Content-Type: text/plain; charset=ISO-8859-15; format=flowed
Content-Transfer-Encoding: 7bit


--------------060405040509010209010802
Content-Type: text/x-patch;
 name="05.diff"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline;
 filename="05.diff"

commit dad61a4af7d23146ce67ec2f069f6ba9b75a578d
Author: Patrick McHardy <kaber@trash.net>
Date:   Wed Feb 6 14:35:11 2008 +0100

    [IPV4]: route: fix crash ip_route_input
    
    ip_route_me_harder() may call ip_route_input() with skbs that don't
    have skb->dev set for skbs rerouted in LOCAL_OUT and TCP resets
    generated by the REJECT target, resulting in a crash when dereferencing
    skb->dev->nd_net. Since ip_route_input() has an input device argument,
    it seems correct to use that one anyway.
    
    Bug introduced in b5921910a1 (Routing cache virtualization).
    
    Signed-off-by: Patrick McHardy <kaber@trash.net>

diff --git a/net/ipv4/route.c b/net/ipv4/route.c
index 8842ecb..525787b 100644
--- a/net/ipv4/route.c
+++ b/net/ipv4/route.c
@@ -2041,7 +2041,7 @@ int ip_route_input(struct sk_buff *skb, __be32 daddr, __be32 saddr,
 	int iif = dev->ifindex;
 	struct net *net;
 
-	net = skb->dev->nd_net;
+	net = dev->nd_net;
 	tos &= IPTOS_RT_MASK;
 	hash = rt_hash(daddr, saddr, iif);
 

--------------060405040509010209010802--