RPC: Ensure that we disconnect TCP sockets if there has been no

NFS traffic for the last 5 minutes. This code also affects
NFSv2/v3.

include/linux/sunrpc/xprt.h

@@ -163,6 +163,12 @@ struct rpc_xprt {
 				tcp_offset;	/* fragment offset */
 	unsigned long		tcp_copied,	/* copied to request */
 				tcp_flags;
+	/*
+	 * Disconnection of idle sockets
+	 */
+	struct work_struct	task_cleanup;
+	struct timer_list	timer;
+	unsigned long		last_used;
 
 	/*
 	 * Send stuff
@@ -202,6 +208,7 @@ int			xprt_clear_backlog(struct rpc_xprt *);
 void			xprt_sock_setbufsize(struct rpc_xprt *);
 
 #define XPRT_CONNECT	0
+#define XPRT_LOCKED	1
 
 #define xprt_connected(xp)		(test_bit(XPRT_CONNECT, &(xp)->sockstate))
 #define xprt_set_connected(xp)		(set_bit(XPRT_CONNECT, &(xp)->sockstate))

net/sunrpc/xprt.c

@@ -59,6 +59,7 @@
 #include <linux/unistd.h>
 #include <linux/sunrpc/clnt.h>
 #include <linux/file.h>
+#include <linux/workqueue.h>
 
 #include <net/sock.h>
 #include <net/checksum.h>
@@ -75,6 +76,7 @@
 #endif
 
 #define XPRT_MAX_BACKOFF	(8)
+#define XPRT_IDLE_TIMEOUT	(5*60*HZ)
 
 /*
  * Local functions
@@ -139,25 +141,33 @@ __xprt_lock_write(struct rpc_xprt *xprt, struct rpc_task *task)
 {
 	struct rpc_rqst *req = task->tk_rqstp;
 
-	if (!xprt->snd_task) {
-		if (xprt->nocong || __xprt_get_cong(xprt, task)) {
-			xprt->snd_task = task;
-			if (req) {
-				req->rq_bytes_sent = 0;
-				req->rq_ntrans++;
-			}
-		}
+	if (test_and_set_bit(XPRT_LOCKED, &xprt->sockstate)) {
+		if (task == xprt->snd_task)
+			return 1;
+		if (task == NULL)
+			return 0;
+		goto out_sleep;
 	}
-	if (xprt->snd_task != task) {
-		dprintk("RPC: %4d TCP write queue full\n", task->tk_pid);
-		task->tk_timeout = 0;
-		task->tk_status = -EAGAIN;
-		if (req && req->rq_ntrans)
-			rpc_sleep_on(&xprt->resend, task, NULL, NULL);
-		else
-			rpc_sleep_on(&xprt->sending, task, NULL, NULL);
+	if (xprt->nocong || __xprt_get_cong(xprt, task)) {
+		xprt->snd_task = task;
+		if (req) {
+			req->rq_bytes_sent = 0;
+			req->rq_ntrans++;
+		}
+		return 1;
 	}
-	return xprt->snd_task == task;
+	smp_mb__before_clear_bit();
+	clear_bit(XPRT_LOCKED, &xprt->sockstate);
+	smp_mb__after_clear_bit();
+out_sleep:
+	dprintk("RPC: %4d failed to lock socket %p\n", task->tk_pid, xprt);
+	task->tk_timeout = 0;
+	task->tk_status = -EAGAIN;
+	if (req && req->rq_ntrans)
+		rpc_sleep_on(&xprt->resend, task, NULL, NULL);
+	else
+		rpc_sleep_on(&xprt->sending, task, NULL, NULL);
+	return 0;
 }
 
 static inline int
@@ -177,15 +187,15 @@ __xprt_lock_write_next(struct rpc_xprt *xprt)
 {
 	struct rpc_task *task;
 
-	if (xprt->snd_task)
+	if (test_and_set_bit(XPRT_LOCKED, &xprt->sockstate))
 		return;
+	if (!xprt->nocong && RPCXPRT_CONGESTED(xprt))
+		goto out_unlock;
 	task = rpc_wake_up_next(&xprt->resend);
 	if (!task) {
-		if (!xprt->nocong && RPCXPRT_CONGESTED(xprt))
-			return;
 		task = rpc_wake_up_next(&xprt->sending);
 		if (!task)
-			return;
+			goto out_unlock;
 	}
 	if (xprt->nocong || __xprt_get_cong(xprt, task)) {
 		struct rpc_rqst *req = task->tk_rqstp;
@@ -194,7 +204,12 @@ __xprt_lock_write_next(struct rpc_xprt *xprt)
 			req->rq_bytes_sent = 0;
 			req->rq_ntrans++;
 		}
+		return;
 	}
+out_unlock:
+	smp_mb__before_clear_bit();
+	clear_bit(XPRT_LOCKED, &xprt->sockstate);
+	smp_mb__after_clear_bit();
 }
 
 /*
@@ -203,9 +218,13 @@ __xprt_lock_write_next(struct rpc_xprt *xprt)
 static void
 __xprt_release_write(struct rpc_xprt *xprt, struct rpc_task *task)
 {
-	if (xprt->snd_task == task)
+	if (xprt->snd_task == task) {
 		xprt->snd_task = NULL;
-	__xprt_lock_write_next(xprt);
+		smp_mb__before_clear_bit();
+		clear_bit(XPRT_LOCKED, &xprt->sockstate);
+		smp_mb__after_clear_bit();
+		__xprt_lock_write_next(xprt);
+	}
 }
 
 static inline void
@@ -393,6 +412,15 @@ xprt_close(struct rpc_xprt *xprt)
 	sock_release(sock);
 }
 
+static void
+xprt_socket_autoclose(void *args)
+{
+	struct rpc_xprt *xprt = (struct rpc_xprt *)args;
+
+	xprt_close(xprt);
+	xprt_release_write(xprt, NULL);
+}
+
 /*
  * Mark a transport as disconnected
  */
@@ -406,6 +434,27 @@ xprt_disconnect(struct rpc_xprt *xprt)
 	spin_unlock_bh(&xprt->sock_lock);
 }
 
+/*
+ * Used to allow disconnection when we've been idle
+ */
+static void
+xprt_init_autodisconnect(unsigned long data)
+{
+	struct rpc_xprt *xprt = (struct rpc_xprt *)data;
+
+	spin_lock(&xprt->sock_lock);
+	if (!list_empty(&xprt->recv) || xprt->shutdown)
+		goto out_abort;
+	if (test_and_set_bit(XPRT_LOCKED, &xprt->sockstate))
+		goto out_abort;
+	spin_unlock(&xprt->sock_lock);
+	/* Let keventd close the socket */
+	schedule_work(&xprt->task_cleanup);
+	return;
+out_abort:
+	spin_unlock(&xprt->sock_lock);
+}
+
 /*
  * Attempt to connect a TCP socket.
  *
@@ -1254,6 +1303,8 @@ xprt_reserve(struct rpc_task *task)
 		spin_lock(&xprt->xprt_lock);
 		do_xprt_reserve(task);
 		spin_unlock(&xprt->xprt_lock);
+		if (task->tk_rqstp)
+			del_timer_sync(&xprt->timer);
 	}
 }
 
@@ -1333,6 +1384,9 @@ xprt_release(struct rpc_task *task)
 	__xprt_put_cong(xprt, req);
 	if (!list_empty(&req->rq_list))
 		list_del(&req->rq_list);
+	xprt->last_used = jiffies;
+	if (list_empty(&xprt->recv) && !xprt->shutdown)
+		mod_timer(&xprt->timer, xprt->last_used + XPRT_IDLE_TIMEOUT);
 	spin_unlock_bh(&xprt->sock_lock);
 	task->tk_rqstp = NULL;
 	memset(req, 0, sizeof(*req));	/* mark unused */
@@ -1403,6 +1457,11 @@ xprt_setup(int proto, struct sockaddr_in *ap, struct rpc_timeout *to)
 	init_waitqueue_head(&xprt->cong_wait);
 
 	INIT_LIST_HEAD(&xprt->recv);
+	INIT_WORK(&xprt->task_cleanup, xprt_socket_autoclose, xprt);
+	init_timer(&xprt->timer);
+	xprt->timer.function = xprt_init_autodisconnect;
+	xprt->timer.data = (unsigned long) xprt;
+	xprt->last_used = jiffies;
 
 	/* Set timeout parameters */
 	if (to) {
@@ -1583,6 +1642,7 @@ xprt_shutdown(struct rpc_xprt *xprt)
 	rpc_wake_up(&xprt->backlog);
 	if (waitqueue_active(&xprt->cong_wait))
 		wake_up(&xprt->cong_wait);
+	del_timer_sync(&xprt->timer);
 }
 
 /*