[PATCH] lockdep: annotate nfs/nfsd in-kernel sockets

Stick NFS sockets in their own class to avoid some lockdep warnings.  NFS
sockets are never exposed to user-space, and will hence not trigger certain
code paths that would otherwise pose deadlock scenarios.

[akpm@osdl.org: cleanups]
Signed-off-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Signed-off-by: Steven Dickson <SteveD@redhat.com>
Acked-by: Ingo Molnar <mingo@elte.hu>
Cc: Trond Myklebust <trond.myklebust@fys.uio.no>
Acked-by: Neil Brown <neilb@suse.de>
Cc: "David S. Miller" <davem@davemloft.net>
Signed-off-by: Andrew Morton <akpm@osdl.org>
[ Fixed patch corruption by quilt, pointed out by Peter Zijlstra ]
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

include/net/sock.h

@@ -746,6 +746,25 @@ static inline int sk_stream_wmem_schedule(struct sock *sk, int size)
  */
 #define sock_owned_by_user(sk)	((sk)->sk_lock.owner)
 
+/*
+ * Macro so as to not evaluate some arguments when
+ * lockdep is not enabled.
+ *
+ * Mark both the sk_lock and the sk_lock.slock as a
+ * per-address-family lock class.
+ */
+#define sock_lock_init_class_and_name(sk, sname, skey, name, key) 	\
+do {									\
+	sk->sk_lock.owner = NULL;					\
+	init_waitqueue_head(&sk->sk_lock.wq);				\
+	spin_lock_init(&(sk)->sk_lock.slock);				\
+	debug_check_no_locks_freed((void *)&(sk)->sk_lock,		\
+			sizeof((sk)->sk_lock));				\
+	lockdep_set_class_and_name(&(sk)->sk_lock.slock,		\
+		       	(skey), (sname));				\
+	lockdep_init_map(&(sk)->sk_lock.dep_map, (name), (key), 0);	\
+} while (0)
+
 extern void FASTCALL(lock_sock_nested(struct sock *sk, int subclass));
 
 static inline void lock_sock(struct sock *sk)

kernel/lockdep.c

@@ -2645,6 +2645,7 @@ void debug_check_no_locks_freed(const void *mem_from, unsigned long mem_len)
 	}
 	local_irq_restore(flags);
 }
+EXPORT_SYMBOL_GPL(debug_check_no_locks_freed);
 
 static void print_held_locks_bug(struct task_struct *curr)
 {

net/core/sock.c

@@ -810,24 +810,11 @@ lenout:
  */
 static void inline sock_lock_init(struct sock *sk)
 {
-	spin_lock_init(&sk->sk_lock.slock);
-	sk->sk_lock.owner = NULL;
-	init_waitqueue_head(&sk->sk_lock.wq);
-	/*
-	 * Make sure we are not reinitializing a held lock:
-	 */
-	debug_check_no_locks_freed((void *)&sk->sk_lock, sizeof(sk->sk_lock));
-
-	/*
-	 * Mark both the sk_lock and the sk_lock.slock as a
-	 * per-address-family lock class:
-	 */
-	lockdep_set_class_and_name(&sk->sk_lock.slock,
-				   af_family_slock_keys + sk->sk_family,
-				   af_family_slock_key_strings[sk->sk_family]);
-	lockdep_init_map(&sk->sk_lock.dep_map,
-			 af_family_key_strings[sk->sk_family],
-			 af_family_keys + sk->sk_family, 0);
+	sock_lock_init_class_and_name(sk,
+			af_family_slock_key_strings[sk->sk_family],
+			af_family_slock_keys + sk->sk_family,
+			af_family_key_strings[sk->sk_family],
+			af_family_keys + sk->sk_family);
 }
 
 /**

net/sunrpc/svcsock.c

@@ -85,6 +85,35 @@ static struct cache_deferred_req *svc_defer(struct cache_req *req);
  */
 static int svc_conn_age_period = 6*60;
 
+#ifdef CONFIG_DEBUG_LOCK_ALLOC
+static struct lock_class_key svc_key[2];
+static struct lock_class_key svc_slock_key[2];
+
+static inline void svc_reclassify_socket(struct socket *sock)
+{
+	struct sock *sk = sock->sk;
+	BUG_ON(sk->sk_lock.owner != NULL);
+	switch (sk->sk_family) {
+	case AF_INET:
+		sock_lock_init_class_and_name(sk, "slock-AF_INET-NFSD",
+		    &svc_slock_key[0], "sk_lock-AF_INET-NFSD", &svc_key[0]);
+		break;
+
+	case AF_INET6:
+		sock_lock_init_class_and_name(sk, "slock-AF_INET6-NFSD",
+		    &svc_slock_key[1], "sk_lock-AF_INET6-NFSD", &svc_key[1]);
+		break;
+
+	default:
+		BUG();
+	}
+}
+#else
+static inline void svc_reclassify_socket(struct socket *sock)
+{
+}
+#endif
+
 /*
  * Queue up an idle server thread.  Must have pool->sp_lock held.
  * Note: this is really a stack rather than a queue, so that we only
@@ -1557,6 +1586,8 @@ svc_create_socket(struct svc_serv *serv, int protocol, struct sockaddr_in *sin)
 	if ((error = sock_create_kern(PF_INET, type, protocol, &sock)) < 0)
 		return error;
 
+	svc_reclassify_socket(sock);
+
 	if (type == SOCK_STREAM)
 		sock->sk->sk_reuse = 1; /* allow address reuse */
 	error = kernel_bind(sock, (struct sockaddr *) sin,

net/sunrpc/xprtsock.c

@@ -1058,6 +1058,35 @@ static int xs_bindresvport(struct rpc_xprt *xprt, struct socket *sock)
 	return err;
 }
 
+#ifdef CONFIG_DEBUG_LOCK_ALLOC
+static struct lock_class_key xs_key[2];
+static struct lock_class_key xs_slock_key[2];
+
+static inline void xs_reclassify_socket(struct socket *sock)
+{
+	struct sock *sk = sock->sk;
+	BUG_ON(sk->sk_lock.owner != NULL);
+	switch (sk->sk_family) {
+	case AF_INET:
+		sock_lock_init_class_and_name(sk, "slock-AF_INET-NFS",
+			&xs_slock_key[0], "sk_lock-AF_INET-NFS", &xs_key[0]);
+		break;
+
+	case AF_INET6:
+		sock_lock_init_class_and_name(sk, "slock-AF_INET6-NFS",
+			&xs_slock_key[1], "sk_lock-AF_INET6-NFS", &xs_key[1]);
+		break;
+
+	default:
+		BUG();
+	}
+}
+#else
+static inline void xs_reclassify_socket(struct socket *sock)
+{
+}
+#endif
+
 /**
  * xs_udp_connect_worker - set up a UDP socket
  * @work: RPC transport to connect
@@ -1081,6 +1110,7 @@ static void xs_udp_connect_worker(struct work_struct *work)
 		dprintk("RPC:      can't create UDP transport socket (%d).\n", -err);
 		goto out;
 	}
+	xs_reclassify_socket(sock);
 
 	if (xprt->resvport && xs_bindresvport(xprt, sock) < 0) {
 		sock_release(sock);
@@ -1165,6 +1195,7 @@ static void xs_tcp_connect_worker(struct work_struct *work)
 			dprintk("RPC:      can't create TCP transport socket (%d).\n", -err);
 			goto out;
 		}
+		xs_reclassify_socket(sock);
 
 		if (xprt->resvport && xs_bindresvport(xprt, sock) < 0) {
 			sock_release(sock);