ipv6: introduce RT6_LOOKUP_F_DST_NOREF flag in ip6_pol_route()

This new flag is to instruct the route lookup function to not take
refcnt on the dst entry. The user which does route lookup with this flag
must properly use rcu protection.
ip6_pol_route() is the major route lookup function for both tx and rx
path.
In this function:
Do not take refcnt on dst if RT6_LOOKUP_F_DST_NOREF flag is set, and
directly return the route entry. The caller should be holding rcu lock
when using this flag, and decide whether to take refcnt or not.

One note on the dst cache in the uncached_list:
As uncached_list does not consume refcnt, one refcnt is always returned
back to the caller even if RT6_LOOKUP_F_DST_NOREF flag is set.
Uncached dst is only possible in the output path. So in such call path,
caller MUST check if the dst is in the uncached_list before assuming
that there is no refcnt taken on the returned dst.
Signed-off-by: Wei Wang <weiwan@google.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Acked-by: Mahesh Bandewar <maheshb@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

include/net/ip6_route.h

@@ -36,6 +36,7 @@ struct route_info {
 #define RT6_LOOKUP_F_SRCPREF_PUBLIC	0x00000010
 #define RT6_LOOKUP_F_SRCPREF_COA	0x00000020
 #define RT6_LOOKUP_F_IGNORE_LINKSTATE	0x00000040
+#define RT6_LOOKUP_F_DST_NOREF		0x00000080
 
 /* We do not (yet ?) support IPv6 jumbograms (RFC 2675)
  * Unlike IPv4, hdr->seg_len doesn't include the IPv6 header

net/ipv6/route.c

@@ -1391,9 +1391,6 @@ static struct rt6_info *rt6_get_pcpu_route(const struct fib6_result *res)
 
 	pcpu_rt = this_cpu_read(*res->nh->rt6i_pcpu);
 
-	if (pcpu_rt)
-		ip6_hold_safe(NULL, &pcpu_rt);
-
 	return pcpu_rt;
 }
 
@@ -1403,12 +1400,9 @@ static struct rt6_info *rt6_make_pcpu_route(struct net *net,
 	struct rt6_info *pcpu_rt, *prev, **p;
 
 	pcpu_rt = ip6_rt_pcpu_alloc(res);
-	if (!pcpu_rt) {
-		dst_hold(&net->ipv6.ip6_null_entry->dst);
-		return net->ipv6.ip6_null_entry;
-	}
+	if (!pcpu_rt)
+		return NULL;
 
-	dst_hold(&pcpu_rt->dst);
 	p = this_cpu_ptr(res->nh->rt6i_pcpu);
 	prev = cmpxchg(p, NULL, pcpu_rt);
 	BUG_ON(prev);
@@ -2189,9 +2183,12 @@ struct rt6_info *ip6_pol_route(struct net *net, struct fib6_table *table,
 			       const struct sk_buff *skb, int flags)
 {
 	struct fib6_result res = {};
-	struct rt6_info *rt;
+	struct rt6_info *rt = NULL;
 	int strict = 0;
 
+	WARN_ON_ONCE((flags & RT6_LOOKUP_F_DST_NOREF) &&
+		     !rcu_read_lock_held());
+
 	strict |= flags & RT6_LOOKUP_F_IFACE;
 	strict |= flags & RT6_LOOKUP_F_IGNORE_LINKSTATE;
 	if (net->ipv6.devconf_all->forwarding == 0)
@@ -2200,23 +2197,15 @@ struct rt6_info *ip6_pol_route(struct net *net, struct fib6_table *table,
 	rcu_read_lock();
 
 	fib6_table_lookup(net, table, oif, fl6, &res, strict);
-	if (res.f6i == net->ipv6.fib6_null_entry) {
-		rt = net->ipv6.ip6_null_entry;
-		rcu_read_unlock();
-		dst_hold(&rt->dst);
-		return rt;
-	}
+	if (res.f6i == net->ipv6.fib6_null_entry)
+		goto out;
 
 	fib6_select_path(net, &res, fl6, oif, false, skb, strict);
 
 	/*Search through exception table */
 	rt = rt6_find_cached_rt(&res, &fl6->daddr, &fl6->saddr);
 	if (rt) {
-		if (ip6_hold_safe(net, &rt))
-			dst_use_noref(&rt->dst, jiffies);
-
-		rcu_read_unlock();
-		return rt;
+		goto out;
 	} else if (unlikely((fl6->flowi6_flags & FLOWI_FLAG_KNOWN_NH) &&
 			    !res.nh->fib_nh_gw_family)) {
 		/* Create a RTF_CACHE clone which will not be
@@ -2224,40 +2213,38 @@ struct rt6_info *ip6_pol_route(struct net *net, struct fib6_table *table,
 		 * the daddr in the skb during the neighbor look-up is different
 		 * from the fl6->daddr used to look-up route here.
 		 */
-		struct rt6_info *uncached_rt;
+		rt = ip6_rt_cache_alloc(&res, &fl6->daddr, NULL);
 
-		uncached_rt = ip6_rt_cache_alloc(&res, &fl6->daddr, NULL);
-
-		rcu_read_unlock();
-
-		if (uncached_rt) {
-			/* Uncached_rt's refcnt is taken during ip6_rt_cache_alloc()
-			 * No need for another dst_hold()
+		if (rt) {
+			/* 1 refcnt is taken during ip6_rt_cache_alloc().
+			 * As rt6_uncached_list_add() does not consume refcnt,
+			 * this refcnt is always returned to the caller even
+			 * if caller sets RT6_LOOKUP_F_DST_NOREF flag.
 			 */
-			rt6_uncached_list_add(uncached_rt);
+			rt6_uncached_list_add(rt);
 			atomic_inc(&net->ipv6.rt6_stats->fib_rt_uncache);
-		} else {
-			uncached_rt = net->ipv6.ip6_null_entry;
-			dst_hold(&uncached_rt->dst);
-		}
+			rcu_read_unlock();
 
-		return uncached_rt;
+			return rt;
+		}
 	} else {
 		/* Get a percpu copy */
-
-		struct rt6_info *pcpu_rt;
-
 		local_bh_disable();
-		pcpu_rt = rt6_get_pcpu_route(&res);
+		rt = rt6_get_pcpu_route(&res);
 
-		if (!pcpu_rt)
-			pcpu_rt = rt6_make_pcpu_route(net, &res);
+		if (!rt)
+			rt = rt6_make_pcpu_route(net, &res);
 
 		local_bh_enable();
-		rcu_read_unlock();
-
-		return pcpu_rt;
 	}
+out:
+	if (!rt)
+		rt = net->ipv6.ip6_null_entry;
+	if (!(flags & RT6_LOOKUP_F_DST_NOREF))
+		ip6_hold_safe(net, &rt);
+	rcu_read_unlock();
+
+	return rt;
 }
 EXPORT_SYMBOL_GPL(ip6_pol_route);