diff --git a/net/core/dev.c b/net/core/dev.c
index 7f31d0f..c8f5ea9 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -1747,10 +1747,10 @@
 
 		skb->tc_verd = SET_TC_AT(skb->tc_verd,AT_INGRESS);
 
-		spin_lock(&dev->queue_lock);
+		spin_lock(&dev->ingress_lock);
 		if ((q = dev->qdisc_ingress) != NULL)
 			result = q->enqueue(skb, q);
-		spin_unlock(&dev->queue_lock);
+		spin_unlock(&dev->ingress_lock);
 
 	}
 
diff --git a/net/sched/cls_route.c b/net/sched/cls_route.c
index e92d716..cc941d0 100644
--- a/net/sched/cls_route.c
+++ b/net/sched/cls_route.c
@@ -89,9 +89,9 @@
 static inline
 void route4_reset_fastmap(struct net_device *dev, struct route4_head *head, u32 id)
 {
-	spin_lock_bh(&dev->queue_lock);
+	qdisc_lock_tree(dev);
 	memset(head->fastmap, 0, sizeof(head->fastmap));
-	spin_unlock_bh(&dev->queue_lock);
+	qdisc_unlock_tree(dev);
 }
 
 static inline void
diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c
index 0ce6914..8699e70 100644
--- a/net/sched/sch_api.c
+++ b/net/sched/sch_api.c
@@ -500,12 +500,16 @@
 
 	if (handle == TC_H_INGRESS) {
 		sch->flags |= TCQ_F_INGRESS;
+		sch->stats_lock = &dev->ingress_lock;
 		handle = TC_H_MAKE(TC_H_INGRESS, 0);
-	} else if (handle == 0) {
-		handle = qdisc_alloc_handle(dev);
-		err = -ENOMEM;
-		if (handle == 0)
-			goto err_out3;
+	} else {
+		sch->stats_lock = &dev->queue_lock;
+		if (handle == 0) {
+			handle = qdisc_alloc_handle(dev);
+			err = -ENOMEM;
+			if (handle == 0)
+				goto err_out3;
+		}
 	}
 
 	sch->handle = handle;
@@ -654,9 +658,9 @@
 			return err;
 		if (q) {
 			qdisc_notify(skb, n, clid, q, NULL);
-			spin_lock_bh(&dev->queue_lock);
+			qdisc_lock_tree(dev);
 			qdisc_destroy(q);
-			spin_unlock_bh(&dev->queue_lock);
+			qdisc_unlock_tree(dev);
 		}
 	} else {
 		qdisc_notify(skb, n, clid, NULL, q);
@@ -789,17 +793,17 @@
 		err = qdisc_graft(dev, p, clid, q, &old_q);
 		if (err) {
 			if (q) {
-				spin_lock_bh(&dev->queue_lock);
+				qdisc_lock_tree(dev);
 				qdisc_destroy(q);
-				spin_unlock_bh(&dev->queue_lock);
+				qdisc_unlock_tree(dev);
 			}
 			return err;
 		}
 		qdisc_notify(skb, n, clid, old_q, q);
 		if (old_q) {
-			spin_lock_bh(&dev->queue_lock);
+			qdisc_lock_tree(dev);
 			qdisc_destroy(old_q);
-			spin_unlock_bh(&dev->queue_lock);
+			qdisc_unlock_tree(dev);
 		}
 	}
 	return 0;
diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c
index 1894eb7..3385ee5 100644
--- a/net/sched/sch_generic.c
+++ b/net/sched/sch_generic.c
@@ -42,16 +42,20 @@
  * The idea is the following:
  * - enqueue, dequeue are serialized via top level device
  *   spinlock dev->queue_lock.
+ * - ingress filtering is serialized via top level device
+ *   spinlock dev->ingress_lock.
  * - updates to tree and tree walking are only done under the rtnl mutex.
  */
 
 void qdisc_lock_tree(struct net_device *dev)
 {
 	spin_lock_bh(&dev->queue_lock);
+	spin_lock(&dev->ingress_lock);
 }
 
 void qdisc_unlock_tree(struct net_device *dev)
 {
+	spin_unlock(&dev->ingress_lock);
 	spin_unlock_bh(&dev->queue_lock);
 }
 
@@ -431,7 +435,6 @@
 	sch->dequeue = ops->dequeue;
 	sch->dev = dev;
 	dev_hold(dev);
-	sch->stats_lock = &dev->queue_lock;
 	atomic_set(&sch->refcnt, 1);
 
 	return sch;
@@ -447,6 +450,7 @@
 	sch = qdisc_alloc(dev, ops);
 	if (IS_ERR(sch))
 		goto errout;
+	sch->stats_lock = &dev->queue_lock;
 	sch->parent = parentid;
 
 	if (!ops->init || ops->init(sch, NULL) == 0)
diff --git a/net/sched/sch_ingress.c b/net/sched/sch_ingress.c
index 1fb60ab..ad22dc6 100644
--- a/net/sched/sch_ingress.c
+++ b/net/sched/sch_ingress.c
@@ -248,16 +248,11 @@
 		skb->dev ? (*pskb)->dev->name : "(no dev)",
 		skb->len);
 
-/*
-revisit later: Use a private since lock dev->queue_lock is also
-used on the egress (might slow things for an iota)
-*/
-
 	if (dev->qdisc_ingress) {
-		spin_lock(&dev->queue_lock);
+		spin_lock(&dev->ingress_lock);
 		if ((q = dev->qdisc_ingress) != NULL)
 			fwres = q->enqueue(skb, q);
-		spin_unlock(&dev->queue_lock);
+		spin_unlock(&dev->ingress_lock);
 	}
 
 	return fwres;