Merge branch 'fib_trie-next'

Alexander Duyck says:

====================
ipv4/fib_trie: Cleanups to prepare for introduction of key vector

This patch series is meant to mostly just clean up the fib_trie to prepare
it for the introduction of the key_vector.  As such there are a number of
minor clean-ups such as reformatting the tnode to match the format once the
key vector is introduced, some optimizations to drop the need for a leaf
parent pointer, and some changes to remove duplication of effort such as
the 2 look-ups that were essentially being done per node insertion.

v2: Added code to cleanup idx >> n->bits and explain unsigned long logic
    Added code to prevent allocation when tnode size is larger than size_t
====================
Signed-off-by: David S. Miller <davem@davemloft.net>

include/net/ip_fib.h

@@ -185,6 +185,7 @@ struct fib_table {
 	u32			tb_id;
 	int			tb_default;
 	int			tb_num_default;
+	struct rcu_head		rcu;
 	unsigned long		tb_data[0];
 };
 
@@ -206,12 +207,16 @@ void fib_free_table(struct fib_table *tb);
 
 static inline struct fib_table *fib_get_table(struct net *net, u32 id)
 {
+	struct hlist_node *tb_hlist;
 	struct hlist_head *ptr;
 
 	ptr = id == RT_TABLE_LOCAL ?
 		&net->ipv4.fib_table_hash[TABLE_LOCAL_INDEX] :
 		&net->ipv4.fib_table_hash[TABLE_MAIN_INDEX];
-	return hlist_entry(ptr->first, struct fib_table, tb_hlist);
+
+	tb_hlist = rcu_dereference_rtnl(hlist_first_rcu(ptr));
+
+	return hlist_entry(tb_hlist, struct fib_table, tb_hlist);
 }
 
 static inline struct fib_table *fib_new_table(struct net *net, u32 id)
@@ -222,15 +227,19 @@ static inline struct fib_table *fib_new_table(struct net *net, u32 id)
 static inline int fib_lookup(struct net *net, const struct flowi4 *flp,
 			     struct fib_result *res)
 {
-	int err = -ENETUNREACH;
+	struct fib_table *tb;
+	int err;
 
 	rcu_read_lock();
 
-	if (!fib_table_lookup(fib_get_table(net, RT_TABLE_LOCAL), flp, res,
-			      FIB_LOOKUP_NOREF) ||
-	    !fib_table_lookup(fib_get_table(net, RT_TABLE_MAIN), flp, res,
-			      FIB_LOOKUP_NOREF))
-		err = 0;
+	for (err = 0; !err; err = -ENETUNREACH) {
+		tb = fib_get_table(net, RT_TABLE_LOCAL);
+		if (tb && !fib_table_lookup(tb, flp, res, FIB_LOOKUP_NOREF))
+			break;
+		tb = fib_get_table(net, RT_TABLE_MAIN);
+		if (tb && !fib_table_lookup(tb, flp, res, FIB_LOOKUP_NOREF))
+			break;
+	}
 
 	rcu_read_unlock();
 
@@ -249,28 +258,33 @@ int __fib_lookup(struct net *net, struct flowi4 *flp, struct fib_result *res);
 static inline int fib_lookup(struct net *net, struct flowi4 *flp,
 			     struct fib_result *res)
 {
-	if (!net->ipv4.fib_has_custom_rules) {
-		int err = -ENETUNREACH;
-
-		rcu_read_lock();
-
-		res->tclassid = 0;
-		if ((net->ipv4.fib_local &&
-		     !fib_table_lookup(net->ipv4.fib_local, flp, res,
-				       FIB_LOOKUP_NOREF)) ||
-		    (net->ipv4.fib_main &&
-		     !fib_table_lookup(net->ipv4.fib_main, flp, res,
-				       FIB_LOOKUP_NOREF)) ||
-		    (net->ipv4.fib_default &&
-		     !fib_table_lookup(net->ipv4.fib_default, flp, res,
-				       FIB_LOOKUP_NOREF)))
-			err = 0;
-
-		rcu_read_unlock();
-
-		return err;
+	struct fib_table *tb;
+	int err;
+
+	if (net->ipv4.fib_has_custom_rules)
+		return __fib_lookup(net, flp, res);
+
+	rcu_read_lock();
+
+	res->tclassid = 0;
+
+	for (err = 0; !err; err = -ENETUNREACH) {
+		tb = rcu_dereference_rtnl(net->ipv4.fib_local);
+		if (tb && !fib_table_lookup(tb, flp, res, FIB_LOOKUP_NOREF))
+			break;
+
+		tb = rcu_dereference_rtnl(net->ipv4.fib_main);
+		if (tb && !fib_table_lookup(tb, flp, res, FIB_LOOKUP_NOREF))
+			break;
+
+		tb = rcu_dereference_rtnl(net->ipv4.fib_default);
+		if (tb && !fib_table_lookup(tb, flp, res, FIB_LOOKUP_NOREF))
+			break;
 	}
-	return __fib_lookup(net, flp, res);
+
+	rcu_read_unlock();
+
+	return err;
 }
 
 #endif /* CONFIG_IP_MULTIPLE_TABLES */

include/net/netns/ipv4.h

@@ -7,6 +7,7 @@
 
 #include <linux/uidgid.h>
 #include <net/inet_frag.h>
+#include <linux/rcupdate.h>
 
 struct tcpm_hash_bucket;
 struct ctl_table_header;
@@ -38,9 +39,9 @@ struct netns_ipv4 {
 #ifdef CONFIG_IP_MULTIPLE_TABLES
 	struct fib_rules_ops	*rules_ops;
 	bool			fib_has_custom_rules;
-	struct fib_table	*fib_local;
-	struct fib_table	*fib_main;
-	struct fib_table	*fib_default;
+	struct fib_table __rcu	*fib_local;
+	struct fib_table __rcu	*fib_main;
+	struct fib_table __rcu	*fib_default;
 #endif
 #ifdef CONFIG_IP_ROUTE_CLASSID
 	int			fib_num_tclassid_users;

net/ipv4/fib_frontend.c

@@ -89,17 +89,14 @@ struct fib_table *fib_new_table(struct net *net, u32 id)
 
 	switch (id) {
 	case RT_TABLE_LOCAL:
-		net->ipv4.fib_local = tb;
+		rcu_assign_pointer(net->ipv4.fib_local, tb);
 		break;
-
 	case RT_TABLE_MAIN:
-		net->ipv4.fib_main = tb;
+		rcu_assign_pointer(net->ipv4.fib_main, tb);
 		break;
-
 	case RT_TABLE_DEFAULT:
-		net->ipv4.fib_default = tb;
+		rcu_assign_pointer(net->ipv4.fib_default, tb);
 		break;
-
 	default:
 		break;
 	}
@@ -132,13 +129,14 @@ struct fib_table *fib_get_table(struct net *net, u32 id)
 static void fib_flush(struct net *net)
 {
 	int flushed = 0;
-	struct fib_table *tb;
-	struct hlist_head *head;
 	unsigned int h;
 
 	for (h = 0; h < FIB_TABLE_HASHSZ; h++) {
-		head = &net->ipv4.fib_table_hash[h];
-		hlist_for_each_entry(tb, head, tb_hlist)
+		struct hlist_head *head = &net->ipv4.fib_table_hash[h];
+		struct hlist_node *tmp;
+		struct fib_table *tb;
+
+		hlist_for_each_entry_safe(tb, tmp, head, tb_hlist)
 			flushed += fib_table_flush(tb);
 	}
 
@@ -665,10 +663,12 @@ static int inet_dump_fib(struct sk_buff *skb, struct netlink_callback *cb)
 	s_h = cb->args[0];
 	s_e = cb->args[1];
 
+	rcu_read_lock();
+
 	for (h = s_h; h < FIB_TABLE_HASHSZ; h++, s_e = 0) {
 		e = 0;
 		head = &net->ipv4.fib_table_hash[h];
-		hlist_for_each_entry(tb, head, tb_hlist) {
+		hlist_for_each_entry_rcu(tb, head, tb_hlist) {
 			if (e < s_e)
 				goto next;
 			if (dumped)
@@ -682,6 +682,8 @@ static int inet_dump_fib(struct sk_buff *skb, struct netlink_callback *cb)
 		}
 	}
 out:
+	rcu_read_unlock();
+
 	cb->args[1] = e;
 	cb->args[0] = h;
 
@@ -1117,14 +1119,34 @@ static void ip_fib_net_exit(struct net *net)
 
 	rtnl_lock();
 	for (i = 0; i < FIB_TABLE_HASHSZ; i++) {
-		struct fib_table *tb;
-		struct hlist_head *head;
+		struct hlist_head *head = &net->ipv4.fib_table_hash[i];
 		struct hlist_node *tmp;
+		struct fib_table *tb;
+
+		/* this is done in two passes as flushing the table could
+		 * cause it to be reallocated in order to accommodate new
+		 * tnodes at the root as the table shrinks.
+		 */
+		hlist_for_each_entry_safe(tb, tmp, head, tb_hlist)
+			fib_table_flush(tb);
 
-		head = &net->ipv4.fib_table_hash[i];
 		hlist_for_each_entry_safe(tb, tmp, head, tb_hlist) {
+#ifdef CONFIG_IP_MULTIPLE_TABLES
+			switch (tb->tb_id) {
+			case RT_TABLE_LOCAL:
+				RCU_INIT_POINTER(net->ipv4.fib_local, NULL);
+				break;
+			case RT_TABLE_MAIN:
+				RCU_INIT_POINTER(net->ipv4.fib_main, NULL);
+				break;
+			case RT_TABLE_DEFAULT:
+				RCU_INIT_POINTER(net->ipv4.fib_default, NULL);
+				break;
+			default:
+				break;
+			}
+#endif
 			hlist_del(&tb->tb_hlist);
-			fib_table_flush(tb);
 			fib_free_table(tb);
 		}
 	}