net: sched: fix crash when deleting secondary chains

If you flush (delete) a filter chain other than chain 0 (such as when
deleting the device), the kernel may run into a use-after-free. The
chain refcount must not be decremented unless we are sure we are done
with the chain.

To reproduce the bug, run:
    ip link add dtest type dummy
    tc qdisc add dev dtest ingress
    tc filter add dev dtest chain 1  parent ffff: flower
    ip link del dtest

Introduced in: commit f93e1cdc ("net/sched: fix filter flushing"),
but unless you have KAsan or luck, you won't notice it until
commit 0dadc117 ("cls_flower: use tcf_exts_get_net() before call_rcu()")

Fixes: f93e1cdc ("net/sched: fix filter flushing")
Acked-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: Roman Kapl <code@rkapl.cz>
Signed-off-by: David S. Miller <davem@davemloft.net>

net/sched/cls_api.c

@@ -205,13 +205,14 @@ static void tcf_chain_head_change(struct tcf_chain *chain,
 
 static void tcf_chain_flush(struct tcf_chain *chain)
 {
-	struct tcf_proto *tp;
+	struct tcf_proto *tp = rtnl_dereference(chain->filter_chain);
 
 	tcf_chain_head_change(chain, NULL);
-	while ((tp = rtnl_dereference(chain->filter_chain)) != NULL) {
+	while (tp) {
 		RCU_INIT_POINTER(chain->filter_chain, tp->next);
-		tcf_chain_put(chain);
 		tcf_proto_destroy(tp);
+		tp = rtnl_dereference(chain->filter_chain);
+		tcf_chain_put(chain);
 	}
 }