[NET_SCHED]: Eliminate qdisc_tree_lock

Since we're now holding the rtnl during the entire dump operation, we
can remove qdisc_tree_lock, whose only purpose is to protect dump
callbacks from concurrent changes to the qdisc tree.

Signed-off-by: Patrick McHardy <kaber@trash.net>
Signed-off-by: David S. Miller <davem@davemloft.net>

3 files changed, 10 insertions, 43 deletions

diff --git a/net/sched/cls_api.c b/net/sched/cls_api.c
index ca3da5013b7a..ebf94edf0478 100644
--- a/net/sched/cls_api.c
+++ b/net/sched/cls_api.c
@@ -400,7 +400,6 @@ static int tc_dump_tfilter(struct sk_buff *skb, struct netlink_callback *cb)
         if ((dev = dev_get_by_index(tcm->tcm_ifindex)) == NULL)
                 return skb->len;
 
-        read_lock(&qdisc_tree_lock);
         if (!tcm->tcm_parent)
                 q = dev->qdisc_sleeping;
         else
@@ -457,7 +456,6 @@ errout:
         if (cl)
                 cops->put(q, cl);
 out:
-        read_unlock(&qdisc_tree_lock);
         dev_put(dev);
         return skb->len;
 }
diff --git a/net/sched/sch_api.c b/net/sched/sch_api.c
index 2e863bdaa9a1..0ce6914f5981 100644
--- a/net/sched/sch_api.c
+++ b/net/sched/sch_api.c
@@ -191,7 +191,7 @@ int unregister_qdisc(struct Qdisc_ops *qops)
    (root qdisc, all its children, children of children etc.)
  */
 
-static struct Qdisc *__qdisc_lookup(struct net_device *dev, u32 handle)
+struct Qdisc *qdisc_lookup(struct net_device *dev, u32 handle)
 {
         struct Qdisc *q;
 
@@ -202,16 +202,6 @@ static struct Qdisc *__qdisc_lookup(struct net_device *dev, u32 handle)
         return NULL;
 }
 
-struct Qdisc *qdisc_lookup(struct net_device *dev, u32 handle)
-{
-        struct Qdisc *q;
-
-        read_lock(&qdisc_tree_lock);
-        q = __qdisc_lookup(dev, handle);
-        read_unlock(&qdisc_tree_lock);
-        return q;
-}
-
 static struct Qdisc *qdisc_leaf(struct Qdisc *p, u32 classid)
 {
         unsigned long cl;
@@ -405,7 +395,7 @@ void qdisc_tree_decrease_qlen(struct Qdisc *sch, unsigned int n)
         if (n == 0)
                 return;
         while ((parentid = sch->parent)) {
-                sch = __qdisc_lookup(sch->dev, TC_H_MAJ(parentid));
+                sch = qdisc_lookup(sch->dev, TC_H_MAJ(parentid));
                 cops = sch->ops->cl_ops;
                 if (cops->qlen_notify) {
                         cl = cops->get(sch, parentid);
@@ -905,7 +895,6 @@ static int tc_dump_qdisc(struct sk_buff *skb, struct netlink_callback *cb)
                         continue;
                 if (idx > s_idx)
                         s_q_idx = 0;
-                read_lock(&qdisc_tree_lock);
                 q_idx = 0;
                 list_for_each_entry(q, &dev->qdisc_list, list) {
                         if (q_idx < s_q_idx) {
@@ -913,13 +902,10 @@ static int tc_dump_qdisc(struct sk_buff *skb, struct netlink_callback *cb)
                                 continue;
                         }
                         if (tc_fill_qdisc(skb, q, q->parent, NETLINK_CB(cb->skb).pid,
-                                          cb->nlh->nlmsg_seq, NLM_F_MULTI, RTM_NEWQDISC) <= 0) {
-                                read_unlock(&qdisc_tree_lock);
+                                          cb->nlh->nlmsg_seq, NLM_F_MULTI, RTM_NEWQDISC) <= 0)
                                 goto done;
-                        }
                         q_idx++;
                 }
-                read_unlock(&qdisc_tree_lock);
         }
 
 done:
@@ -1142,7 +1128,6 @@ static int tc_dump_tclass(struct sk_buff *skb, struct netlink_callback *cb)
         s_t = cb->args[0];
         t = 0;
 
-        read_lock(&qdisc_tree_lock);
         list_for_each_entry(q, &dev->qdisc_list, list) {
                 if (t < s_t || !q->ops->cl_ops ||
                     (tcm->tcm_parent &&
@@ -1164,7 +1149,6 @@ static int tc_dump_tclass(struct sk_buff *skb, struct netlink_callback *cb)
                         break;
                 t++;
         }
-        read_unlock(&qdisc_tree_lock);
 
         cb->args[0] = t;
 
diff --git a/net/sched/sch_generic.c b/net/sched/sch_generic.c
index 52eb3439d7c6..1894eb72f6cf 100644
--- a/net/sched/sch_generic.c
+++ b/net/sched/sch_generic.c
@@ -36,34 +36,23 @@
 
 /* Main transmission queue. */
 
-/* Main qdisc structure lock.
-
-   However, modifications
-   to data, participating in scheduling must be additionally
-   protected with dev->queue_lock spinlock.
-
-   The idea is the following:
-   - enqueue, dequeue are serialized via top level device
-     spinlock dev->queue_lock.
-   - tree walking is protected by read_lock(qdisc_tree_lock)
-     and this lock is used only in process context.
-   - updates to tree are made only under rtnl semaphore,
-     hence this lock may be made without local bh disabling.
-
-   qdisc_tree_lock must be grabbed BEFORE dev->queue_lock!
+/* Modifications to data participating in scheduling must be protected with
+ * dev->queue_lock spinlock.
+ *
+ * The idea is the following:
+ * - enqueue, dequeue are serialized via top level device
+ *   spinlock dev->queue_lock.
+ * - updates to tree and tree walking are only done under the rtnl mutex.
  */
-DEFINE_RWLOCK(qdisc_tree_lock);
 
 void qdisc_lock_tree(struct net_device *dev)
 {
-        write_lock(&qdisc_tree_lock);
         spin_lock_bh(&dev->queue_lock);
 }
 
 void qdisc_unlock_tree(struct net_device *dev)
 {
         spin_unlock_bh(&dev->queue_lock);
-        write_unlock(&qdisc_tree_lock);
 }
 
 /*
@@ -528,15 +517,11 @@ void dev_activate(struct net_device *dev)
                                 printk(KERN_INFO "%s: activation failed\n", dev->name);
                                 return;
                         }
-                        write_lock(&qdisc_tree_lock);
                         list_add_tail(&qdisc->list, &dev->qdisc_list);
-                        write_unlock(&qdisc_tree_lock);
                 } else {
                         qdisc =  &noqueue_qdisc;
                 }
-                write_lock(&qdisc_tree_lock);
                 dev->qdisc_sleeping = qdisc;
-                write_unlock(&qdisc_tree_lock);
         }
 
         if (!netif_carrier_ok(dev))