Merge branch 'rhash-raw-walkers-remove-part-1'

Herbert Xu says:

====================
rhashtable: Get rid of raw table walkers part 1

This series starts the process of getting rid of all raw rhashtable
walkers (e.g., using any of the rht_for_each helpers) from the
kernel.

We need to do this before I can fix the resize kmalloc failure issue
by using multi-layered tables.

We should do this anyway because almost all raw table walkers are
already buggy in that they don't handle multiple rhashtables during
a resize.
====================

Dave/Tomas, please keep an eye out for any new patches that try
to introduce raw table walkers and nack them.
Signed-off-by: David S. Miller <davem@davemloft.net>

MAINTAINERS

@@ -9936,6 +9936,7 @@ F:	net/rfkill/
 
 RHASHTABLE
 M:	Thomas Graf <tgraf@suug.ch>
+M:	Herbert Xu <herbert@gondor.apana.org.au>
 L:	netdev@vger.kernel.org
 S:	Maintained
 F:	lib/rhashtable.c

include/linux/rhashtable.h

@@ -173,7 +173,7 @@ struct rhashtable_walker {
 struct rhashtable_iter {
 	struct rhashtable *ht;
 	struct rhash_head *p;
-	struct rhashtable_walker *walker;
+	struct rhashtable_walker walker;
 	unsigned int slot;
 	unsigned int skip;
 };
@@ -346,8 +346,8 @@ struct bucket_table *rhashtable_insert_slow(struct rhashtable *ht,
 					    struct bucket_table *old_tbl);
 int rhashtable_insert_rehash(struct rhashtable *ht, struct bucket_table *tbl);
 
-int rhashtable_walk_init(struct rhashtable *ht, struct rhashtable_iter *iter,
-			 gfp_t gfp);
+void rhashtable_walk_enter(struct rhashtable *ht,
+			   struct rhashtable_iter *iter);
 void rhashtable_walk_exit(struct rhashtable_iter *iter);
 int rhashtable_walk_start(struct rhashtable_iter *iter) __acquires(RCU);
 void *rhashtable_walk_next(struct rhashtable_iter *iter);
@@ -906,4 +906,12 @@ static inline int rhashtable_replace_fast(
 	return err;
 }
 
+/* Obsolete function, do not use in new code. */
+static inline int rhashtable_walk_init(struct rhashtable *ht,
+				       struct rhashtable_iter *iter, gfp_t gfp)
+{
+	rhashtable_walk_enter(ht, iter);
+	return 0;
+}
+
 #endif /* _LINUX_RHASHTABLE_H */

lib/rhashtable.c

@@ -489,10 +489,9 @@ struct bucket_table *rhashtable_insert_slow(struct rhashtable *ht,
 EXPORT_SYMBOL_GPL(rhashtable_insert_slow);
 
 /**
- * rhashtable_walk_init - Initialise an iterator
+ * rhashtable_walk_enter - Initialise an iterator
  * @ht:		Table to walk over
  * @iter:	Hash table Iterator
- * @gfp:	GFP flags for allocations
  *
  * This function prepares a hash table walk.
  *
@@ -507,30 +506,22 @@ EXPORT_SYMBOL_GPL(rhashtable_insert_slow);
  * This function may sleep so you must not call it from interrupt
  * context or with spin locks held.
  *
- * You must call rhashtable_walk_exit if this function returns
- * successfully.
+ * You must call rhashtable_walk_exit after this function returns.
  */
-int rhashtable_walk_init(struct rhashtable *ht, struct rhashtable_iter *iter,
-			 gfp_t gfp)
+void rhashtable_walk_enter(struct rhashtable *ht, struct rhashtable_iter *iter)
 {
 	iter->ht = ht;
 	iter->p = NULL;
 	iter->slot = 0;
 	iter->skip = 0;
 
-	iter->walker = kmalloc(sizeof(*iter->walker), gfp);
-	if (!iter->walker)
-		return -ENOMEM;
-
 	spin_lock(&ht->lock);
-	iter->walker->tbl =
+	iter->walker.tbl =
 		rcu_dereference_protected(ht->tbl, lockdep_is_held(&ht->lock));
-	list_add(&iter->walker->list, &iter->walker->tbl->walkers);
+	list_add(&iter->walker.list, &iter->walker.tbl->walkers);
 	spin_unlock(&ht->lock);
-
-	return 0;
 }
-EXPORT_SYMBOL_GPL(rhashtable_walk_init);
+EXPORT_SYMBOL_GPL(rhashtable_walk_enter);
 
 /**
  * rhashtable_walk_exit - Free an iterator
@@ -541,10 +532,9 @@ EXPORT_SYMBOL_GPL(rhashtable_walk_init);
 void rhashtable_walk_exit(struct rhashtable_iter *iter)
 {
 	spin_lock(&iter->ht->lock);
-	if (iter->walker->tbl)
-		list_del(&iter->walker->list);
+	if (iter->walker.tbl)
+		list_del(&iter->walker.list);
 	spin_unlock(&iter->ht->lock);
-	kfree(iter->walker);
 }
 EXPORT_SYMBOL_GPL(rhashtable_walk_exit);
 
@@ -570,12 +560,12 @@ int rhashtable_walk_start(struct rhashtable_iter *iter)
 	rcu_read_lock();
 
 	spin_lock(&ht->lock);
-	if (iter->walker->tbl)
-		list_del(&iter->walker->list);
+	if (iter->walker.tbl)
+		list_del(&iter->walker.list);
 	spin_unlock(&ht->lock);
 
-	if (!iter->walker->tbl) {
-		iter->walker->tbl = rht_dereference_rcu(ht->tbl, ht);
+	if (!iter->walker.tbl) {
+		iter->walker.tbl = rht_dereference_rcu(ht->tbl, ht);
 		return -EAGAIN;
 	}
 
@@ -597,7 +587,7 @@ EXPORT_SYMBOL_GPL(rhashtable_walk_start);
  */
 void *rhashtable_walk_next(struct rhashtable_iter *iter)
 {
-	struct bucket_table *tbl = iter->walker->tbl;
+	struct bucket_table *tbl = iter->walker.tbl;
 	struct rhashtable *ht = iter->ht;
 	struct rhash_head *p = iter->p;
 
@@ -630,8 +620,8 @@ void *rhashtable_walk_next(struct rhashtable_iter *iter)
 	/* Ensure we see any new tables. */
 	smp_rmb();
 
-	iter->walker->tbl = rht_dereference_rcu(tbl->future_tbl, ht);
-	if (iter->walker->tbl) {
+	iter->walker.tbl = rht_dereference_rcu(tbl->future_tbl, ht);
+	if (iter->walker.tbl) {
 		iter->slot = 0;
 		iter->skip = 0;
 		return ERR_PTR(-EAGAIN);
@@ -651,7 +641,7 @@ void rhashtable_walk_stop(struct rhashtable_iter *iter)
 	__releases(RCU)
 {
 	struct rhashtable *ht;
-	struct bucket_table *tbl = iter->walker->tbl;
+	struct bucket_table *tbl = iter->walker.tbl;
 
 	if (!tbl)
 		goto out;
@@ -660,9 +650,9 @@ void rhashtable_walk_stop(struct rhashtable_iter *iter)
 
 	spin_lock(&ht->lock);
 	if (tbl->rehash < tbl->size)
-		list_add(&iter->walker->list, &tbl->walkers);
+		list_add(&iter->walker.list, &tbl->walkers);
 	else
-		iter->walker->tbl = NULL;
+		iter->walker.tbl = NULL;
 	spin_unlock(&ht->lock);
 
 	iter->p = NULL;

net/netlink/diag.c

@@ -63,43 +63,75 @@ static int sk_diag_fill(struct sock *sk, struct sk_buff *skb,
 static int __netlink_diag_dump(struct sk_buff *skb, struct netlink_callback *cb,
 				int protocol, int s_num)
 {
+	struct rhashtable_iter *hti = (void *)cb->args[2];
 	struct netlink_table *tbl = &nl_table[protocol];
-	struct rhashtable *ht = &tbl->hash;
-	const struct bucket_table *htbl = rht_dereference_rcu(ht->tbl, ht);
 	struct net *net = sock_net(skb->sk);
 	struct netlink_diag_req *req;
 	struct netlink_sock *nlsk;
 	struct sock *sk;
-	int ret = 0, num = 0, i;
+	int num = 2;
+	int ret = 0;
 
 	req = nlmsg_data(cb->nlh);
 
-	for (i = 0; i < htbl->size; i++) {
-		struct rhash_head *pos;
+	if (s_num > 1)
+		goto mc_list;
 
-		rht_for_each_entry_rcu(nlsk, pos, htbl, i, node) {
-			sk = (struct sock *)nlsk;
+	num--;
 
-			if (!net_eq(sock_net(sk), net))
-				continue;
-			if (num < s_num) {
-				num++;
+	if (!hti) {
+		hti = kmalloc(sizeof(*hti), GFP_KERNEL);
+		if (!hti)
+			return -ENOMEM;
+
+		cb->args[2] = (long)hti;
+	}
+
+	if (!s_num)
+		rhashtable_walk_enter(&tbl->hash, hti);
+
+	ret = rhashtable_walk_start(hti);
+	if (ret == -EAGAIN)
+		ret = 0;
+	if (ret)
+		goto stop;
+
+	while ((nlsk = rhashtable_walk_next(hti))) {
+		if (IS_ERR(nlsk)) {
+			ret = PTR_ERR(nlsk);
+			if (ret == -EAGAIN) {
+				ret = 0;
 				continue;
 			}
+			break;
+		}
 
-			if (sk_diag_fill(sk, skb, req,
-					 NETLINK_CB(cb->skb).portid,
-					 cb->nlh->nlmsg_seq,
-					 NLM_F_MULTI,
-					 sock_i_ino(sk)) < 0) {
-				ret = 1;
-				goto done;
-			}
+		sk = (struct sock *)nlsk;
 
-			num++;
+		if (!net_eq(sock_net(sk), net))
+			continue;
+
+		if (sk_diag_fill(sk, skb, req,
+				 NETLINK_CB(cb->skb).portid,
+				 cb->nlh->nlmsg_seq,
+				 NLM_F_MULTI,
+				 sock_i_ino(sk)) < 0) {
+			ret = 1;
+			break;
 		}
 	}
 
+stop:
+	rhashtable_walk_stop(hti);
+	if (ret)
+		goto done;
+
+	rhashtable_walk_exit(hti);
+	cb->args[2] = 0;
+	num++;
+
+mc_list:
+	read_lock(&nl_table_lock);
 	sk_for_each_bound(sk, &tbl->mc_list) {
 		if (sk_hashed(sk))
 			continue;
@@ -116,13 +148,14 @@ static int __netlink_diag_dump(struct sk_buff *skb, struct netlink_callback *cb,
 				 NLM_F_MULTI,
 				 sock_i_ino(sk)) < 0) {
 			ret = 1;
-			goto done;
+			break;
 		}
 		num++;
 	}
+	read_unlock(&nl_table_lock);
+
 done:
 	cb->args[0] = num;
-	cb->args[1] = protocol;
 
 	return ret;
 }
@@ -131,20 +164,20 @@ static int netlink_diag_dump(struct sk_buff *skb, struct netlink_callback *cb)
 {
 	struct netlink_diag_req *req;
 	int s_num = cb->args[0];
+	int err = 0;
 
 	req = nlmsg_data(cb->nlh);
 
-	rcu_read_lock();
-	read_lock(&nl_table_lock);
-
 	if (req->sdiag_protocol == NDIAG_PROTO_ALL) {
 		int i;
 
 		for (i = cb->args[1]; i < MAX_LINKS; i++) {
-			if (__netlink_diag_dump(skb, cb, i, s_num))
+			err = __netlink_diag_dump(skb, cb, i, s_num);
+			if (err)
 				break;
 			s_num = 0;
 		}
+		cb->args[1] = i;
 	} else {
 		if (req->sdiag_protocol >= MAX_LINKS) {
 			read_unlock(&nl_table_lock);
@@ -152,13 +185,22 @@ static int netlink_diag_dump(struct sk_buff *skb, struct netlink_callback *cb)
 			return -ENOENT;
 		}
 
-		__netlink_diag_dump(skb, cb, req->sdiag_protocol, s_num);
+		err = __netlink_diag_dump(skb, cb, req->sdiag_protocol, s_num);
 	}
 
-	read_unlock(&nl_table_lock);
-	rcu_read_unlock();
+	return err < 0 ? err : skb->len;
+}
+
+static int netlink_diag_dump_done(struct netlink_callback *cb)
+{
+	struct rhashtable_iter *hti = (void *)cb->args[2];
+
+	if (cb->args[0] == 1)
+		rhashtable_walk_exit(hti);
 
-	return skb->len;
+	kfree(hti);
+
+	return 0;
 }
 
 static int netlink_diag_handler_dump(struct sk_buff *skb, struct nlmsghdr *h)
@@ -172,6 +214,7 @@ static int netlink_diag_handler_dump(struct sk_buff *skb, struct nlmsghdr *h)
 	if (h->nlmsg_flags & NLM_F_DUMP) {
 		struct netlink_dump_control c = {
 			.dump = netlink_diag_dump,
+			.done = netlink_diag_dump_done,
 		};
 		return netlink_dump_start(net->diag_nlsk, skb, h, &c);
 	} else