[patch 28/28] udp: Fix rcv socket locking

[Greg KH <gregkh@suse.de>]

[-- Attachment #1: 0006-udp-Fix-rcv-socket-locking.patch --]
[-- Type: text/plain, Size: 3998 bytes --]

2.6.25-stable review patch.  If anyone has any objections, please let us
know.

------------------
From: Herbert Xu <herbert@gondor.apana.org.au>

[ Upstream commits d97106ea52aa57e63ff40d04479016836bbb5a4e and
   93821778def10ec1e69aa3ac10adee975dad4ff3 ]

The previous patch in response to the recursive locking on IPsec
reception is broken as it tries to drop the BH socket lock while in
user context.

This patch fixes it by shrinking the section protected by the
socket lock to sock_queue_rcv_skb only.  The only reason we added
the lock is for the accounting which happens in that function.

Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>

---
 net/ipv4/udp.c |   57 +++++++++++++++++++++++++++++++++------------------------
 net/ipv6/udp.c |    6 +++---
 2 files changed, 36 insertions(+), 27 deletions(-)

--- a/net/ipv4/udp.c
+++ b/net/ipv4/udp.c
@@ -956,6 +956,27 @@ int udp_disconnect(struct sock *sk, int 
 	return 0;
 }
 
+static int __udp_queue_rcv_skb(struct sock *sk, struct sk_buff *skb)
+{
+	int is_udplite = IS_UDPLITE(sk);
+	int rc;
+
+	if ((rc = sock_queue_rcv_skb(sk, skb)) < 0) {
+		/* Note that an ENOMEM error is charged twice */
+		if (rc == -ENOMEM)
+			UDP_INC_STATS_BH(UDP_MIB_RCVBUFERRORS,
+					 is_udplite);
+		goto drop;
+	}
+
+	return 0;
+
+drop:
+	UDP_INC_STATS_BH(UDP_MIB_INERRORS, is_udplite);
+	kfree_skb(skb);
+	return -1;
+}
+
 /* returns:
  *  -1: error
  *   0: success
@@ -1046,14 +1067,16 @@ int udp_queue_rcv_skb(struct sock * sk, 
 			goto drop;
 	}
 
-	if ((rc = sock_queue_rcv_skb(sk,skb)) < 0) {
-		/* Note that an ENOMEM error is charged twice */
-		if (rc == -ENOMEM)
-			UDP_INC_STATS_BH(UDP_MIB_RCVBUFERRORS, is_udplite);
-		goto drop;
-	}
+	rc = 0;
 
-	return 0;
+	bh_lock_sock(sk);
+	if (!sock_owned_by_user(sk))
+		rc = __udp_queue_rcv_skb(sk, skb);
+	else
+		sk_add_backlog(sk, skb);
+	bh_unlock_sock(sk);
+
+	return rc;
 
 drop:
 	UDP_INC_STATS_BH(UDP_MIB_INERRORS, is_udplite);
@@ -1091,15 +1114,7 @@ static int __udp4_lib_mcast_deliver(stru
 				skb1 = skb_clone(skb, GFP_ATOMIC);
 
 			if (skb1) {
-				int ret = 0;
-
-				bh_lock_sock_nested(sk);
-				if (!sock_owned_by_user(sk))
-					ret = udp_queue_rcv_skb(sk, skb1);
-				else
-					sk_add_backlog(sk, skb1);
-				bh_unlock_sock(sk);
-
+				int ret = udp_queue_rcv_skb(sk, skb1);
 				if (ret > 0)
 					/* we should probably re-process instead
 					 * of dropping packets here. */
@@ -1192,13 +1207,7 @@ int __udp4_lib_rcv(struct sk_buff *skb, 
 			uh->dest, inet_iif(skb), udptable);
 
 	if (sk != NULL) {
-		int ret = 0;
-		bh_lock_sock_nested(sk);
-		if (!sock_owned_by_user(sk))
-			ret = udp_queue_rcv_skb(sk, skb);
-		else
-			sk_add_backlog(sk, skb);
-		bh_unlock_sock(sk);
+		int ret = udp_queue_rcv_skb(sk, skb);
 		sock_put(sk);
 
 		/* a return value > 0 means to resubmit the input, but
@@ -1493,7 +1502,7 @@ struct proto udp_prot = {
 	.sendmsg	   = udp_sendmsg,
 	.recvmsg	   = udp_recvmsg,
 	.sendpage	   = udp_sendpage,
-	.backlog_rcv	   = udp_queue_rcv_skb,
+	.backlog_rcv	   = __udp_queue_rcv_skb,
 	.hash		   = udp_lib_hash,
 	.unhash		   = udp_lib_unhash,
 	.get_port	   = udp_v4_get_port,
--- a/net/ipv6/udp.c
+++ b/net/ipv6/udp.c
@@ -373,7 +373,7 @@ static int __udp6_lib_mcast_deliver(stru
 					uh->source, saddr, dif))) {
 		struct sk_buff *buff = skb_clone(skb, GFP_ATOMIC);
 		if (buff) {
-			bh_lock_sock_nested(sk2);
+			bh_lock_sock(sk2);
 			if (!sock_owned_by_user(sk2))
 				udpv6_queue_rcv_skb(sk2, buff);
 			else
@@ -381,7 +381,7 @@ static int __udp6_lib_mcast_deliver(stru
 			bh_unlock_sock(sk2);
 		}
 	}
-	bh_lock_sock_nested(sk);
+	bh_lock_sock(sk);
 	if (!sock_owned_by_user(sk))
 		udpv6_queue_rcv_skb(sk, skb);
 	else
@@ -499,7 +499,7 @@ int __udp6_lib_rcv(struct sk_buff *skb, 
 
 	/* deliver */
 
-	bh_lock_sock_nested(sk);
+	bh_lock_sock(sk);
 	if (!sock_owned_by_user(sk))
 		udpv6_queue_rcv_skb(sk, skb);
 	else

--