netdev: Fix lockdep warnings in multiqueue configurations.

When support for multiple TX queues were added, the
netif_tx_lock() routines we converted to iterate over
all TX queues and grab each queue's spinlock.

This causes heartburn for lockdep and it's not a healthy
thing to do with lots of TX queues anyways.

So modify this to use a top-level lock and a "frozen"
state for the individual TX queues.

Signed-off-by: David S. Miller <davem@davemloft.net>
diff --git a/include/linux/netdevice.h b/include/linux/netdevice.h
index b4d056c..ee583f6 100644
--- a/include/linux/netdevice.h
+++ b/include/linux/netdevice.h
@@ -440,6 +440,7 @@
 enum netdev_queue_state_t
 {
 	__QUEUE_STATE_XOFF,
+	__QUEUE_STATE_FROZEN,
 };
 
 struct netdev_queue {
@@ -636,7 +637,7 @@
 	unsigned int		real_num_tx_queues;
 
 	unsigned long		tx_queue_len;	/* Max frames per queue allowed */
-
+	spinlock_t		tx_global_lock;
 /*
  * One part is mostly used on xmit path (device)
  */
@@ -1099,6 +1100,11 @@
 	return netif_tx_queue_stopped(netdev_get_tx_queue(dev, 0));
 }
 
+static inline int netif_tx_queue_frozen(const struct netdev_queue *dev_queue)
+{
+	return test_bit(__QUEUE_STATE_FROZEN, &dev_queue->state);
+}
+
 /**
  *	netif_running - test if up
  *	@dev: network device
@@ -1475,30 +1481,6 @@
 	txq->xmit_lock_owner = smp_processor_id();
 }
 
-/**
- *	netif_tx_lock - grab network device transmit lock
- *	@dev: network device
- *	@cpu: cpu number of lock owner
- *
- * Get network device transmit lock
- */
-static inline void netif_tx_lock(struct net_device *dev)
-{
-	int cpu = smp_processor_id();
-	unsigned int i;
-
-	for (i = 0; i < dev->num_tx_queues; i++) {
-		struct netdev_queue *txq = netdev_get_tx_queue(dev, i);
-		__netif_tx_lock(txq, cpu);
-	}
-}
-
-static inline void netif_tx_lock_bh(struct net_device *dev)
-{
-	local_bh_disable();
-	netif_tx_lock(dev);
-}
-
 static inline int __netif_tx_trylock(struct netdev_queue *txq)
 {
 	int ok = spin_trylock(&txq->_xmit_lock);
@@ -1507,11 +1489,6 @@
 	return ok;
 }
 
-static inline int netif_tx_trylock(struct net_device *dev)
-{
-	return __netif_tx_trylock(netdev_get_tx_queue(dev, 0));
-}
-
 static inline void __netif_tx_unlock(struct netdev_queue *txq)
 {
 	txq->xmit_lock_owner = -1;
@@ -1524,15 +1501,57 @@
 	spin_unlock_bh(&txq->_xmit_lock);
 }
 
+/**
+ *	netif_tx_lock - grab network device transmit lock
+ *	@dev: network device
+ *	@cpu: cpu number of lock owner
+ *
+ * Get network device transmit lock
+ */
+static inline void netif_tx_lock(struct net_device *dev)
+{
+	unsigned int i;
+	int cpu;
+
+	spin_lock(&dev->tx_global_lock);
+	cpu = smp_processor_id();
+	for (i = 0; i < dev->num_tx_queues; i++) {
+		struct netdev_queue *txq = netdev_get_tx_queue(dev, i);
+
+		/* We are the only thread of execution doing a
+		 * freeze, but we have to grab the _xmit_lock in
+		 * order to synchronize with threads which are in
+		 * the ->hard_start_xmit() handler and already
+		 * checked the frozen bit.
+		 */
+		__netif_tx_lock(txq, cpu);
+		set_bit(__QUEUE_STATE_FROZEN, &txq->state);
+		__netif_tx_unlock(txq);
+	}
+}
+
+static inline void netif_tx_lock_bh(struct net_device *dev)
+{
+	local_bh_disable();
+	netif_tx_lock(dev);
+}
+
 static inline void netif_tx_unlock(struct net_device *dev)
 {
 	unsigned int i;
 
 	for (i = 0; i < dev->num_tx_queues; i++) {
 		struct netdev_queue *txq = netdev_get_tx_queue(dev, i);
-		__netif_tx_unlock(txq);
-	}
 
+		/* No need to grab the _xmit_lock here.  If the
+		 * queue is not stopped for another reason, we
+		 * force a schedule.
+		 */
+		clear_bit(__QUEUE_STATE_FROZEN, &txq->state);
+		if (!test_bit(__QUEUE_STATE_XOFF, &txq->state))
+			__netif_schedule(txq->qdisc);
+	}
+	spin_unlock(&dev->tx_global_lock);
 }
 
 static inline void netif_tx_unlock_bh(struct net_device *dev)
@@ -1556,13 +1575,18 @@
 static inline void netif_tx_disable(struct net_device *dev)
 {
 	unsigned int i;
+	int cpu;
 
-	netif_tx_lock_bh(dev);
+	local_bh_disable();
+	cpu = smp_processor_id();
 	for (i = 0; i < dev->num_tx_queues; i++) {
 		struct netdev_queue *txq = netdev_get_tx_queue(dev, i);
+
+		__netif_tx_lock(txq, cpu);
 		netif_tx_stop_queue(txq);
+		__netif_tx_unlock(txq);
 	}
-	netif_tx_unlock_bh(dev);
+	local_bh_enable();
 }
 
 static inline void netif_addr_lock(struct net_device *dev)