From: David S. Miller <davem@davemloft.net>
Date: Sat, 5 Mar 2011 05:35:25 +0000 (-0800)
Subject: ipv4: Set rt->rt_iif more sanely on output routes.
X-Git-Tag: v2.6.39-rc1~468^2~153
X-Git-Url: https://openfabrics.org/gitweb/?a=commitdiff_plain;h=1018b5c01636c7c6bda31a719bda34fc631db29a;p=~emulex%2Finfiniband.git

ipv4: Set rt->rt_iif more sanely on output routes.

rt->rt_iif is only ever inspected on input routes, for example DCCP
uses this to populate a route lookup flow key when generating replies
to another packet.

Therefore, setting it to anything other than zero on output routes
makes no sense.

Signed-off-by: David S. Miller <davem@davemloft.net>
---

diff --git a/net/ipv4/route.c b/net/ipv4/route.c
index 9794a2c6023..602473c9201 100644
--- a/net/ipv4/route.c
+++ b/net/ipv4/route.c
@@ -2381,7 +2381,7 @@ static struct rtable *__mkroute_output(const struct fib_result *res,
 	rth->fl.mark    = oldflp->mark;
 	rth->rt_dst	= fl->fl4_dst;
 	rth->rt_src	= fl->fl4_src;
-	rth->rt_iif	= oldflp->oif ? : dev_out->ifindex;
+	rth->rt_iif	= 0;
 	/* get references to the devices that are to be hold by the routing
 	   cache entry */
 	rth->dst.dev	= dev_out;