ipv4: call dst_hold_safe() properly

This patch checks all the calls to
dst_hold()/skb_dst_force()/dst_clone()/dst_use() to see if
dst_hold_safe() is needed to avoid double free issue if dst
gc is removed and dst_release() directly destroys dst when
dst->__refcnt drops to 0.

In tx path, TCP hold sk->sk_rx_dst ref count and also hold sock_lock().
UDP and other similar protocols always hold refcount for
skb->_skb_refdst. So both paths seem to be safe.

In rx path, as it is lockless and skb_dst_set_noref() is likely to be
used, dst_hold_safe() should always be used when trying to hold dst.

In the routing code, if dst is held during an rcu protected session, it
is necessary to call dst_hold_safe() as the current dst might be in its
rcu grace period.
Signed-off-by: Wei Wang <weiwan@google.com>
Acked-by: Martin KaFai Lau <kafai@fb.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

include/net/route.h

@@ -190,7 +190,9 @@ static inline int ip_route_input(struct sk_buff *skb, __be32 dst, __be32 src,
 	rcu_read_lock();
 	err = ip_route_input_noref(skb, dst, src, tos, devin);
 	if (!err)
-		skb_dst_force(skb);
+		skb_dst_force_safe(skb);
+	if (!skb_dst(skb))
+		err = -EINVAL;
 	rcu_read_unlock();
 
 	return err;

net/ipv4/route.c

@@ -2234,11 +2234,9 @@ static struct rtable *__mkroute_output(const struct fib_result *res,
 		rth = rcu_dereference(*prth);
 
 rt_cache:
-		if (rt_cache_valid(rth)) {
-			dst_hold(&rth->dst);
+		if (rt_cache_valid(rth) && dst_hold_safe(&rth->dst))
 			return rth;
 	}
-	}
 
 add:
 	rth = rt_dst_alloc(dev_out, flags, type,