On Tue, Nov 08, 2011 at 02:33:02PM -0500, David Miller wrote: > From: David Miller <davem@xxxxxxxxxxxxx> > Date: Tue, 08 Nov 2011 14:19:50 -0500 (EST) > > > I suspect that your real problem has nothing to do with UDP or RAW, > > but rather the issue is that entries already in the routing cache > > with a NULL peer need to be refreshed with peer information created > > in another context. Yes, that's the problem. > > So you want something like this patch: > Originally, I wanted to fix it with the patch below. Given the fact that dst->obsolete is not null, this should do the same like your patch for output routes. During the tests with this patch I noticed a problem with that. Unfortunately I can't remember what it was... I'll do some investigating, perhaps I can get it back to my mind. I did some quick tests with this and with your patch and both seem to fix the problem at the first glance. diff --git a/net/ipv4/route.c b/net/ipv4/route.c index 511f4a7..ac189c9 100644 --- a/net/ipv4/route.c +++ b/net/ipv4/route.c @@ -2723,7 +2723,7 @@ struct rtable *__ip_route_output_key(struct net *net, struct flowi4 *flp4) !((rth->rt_key_tos ^ flp4->flowi4_tos) & (IPTOS_RT_MASK | RTO_ONLINK)) && net_eq(dev_net(rth->dst.dev), net) && - !rt_is_expired(rth)) { + (rth = (struct rtable *) dst_check(&rth->dst, 0)) && rth) { dst_use(&rth->dst, jiffies); RT_CACHE_STAT_INC(out_hit); rcu_read_unlock_bh(); -- 1.7.0.4 -- To unsubscribe from this list: send the line "unsubscribe netdev" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html