net: dev: Change the order of the arguments for the contended condition.

[linux-2.6-microblaze.git] / net / core / dev.c

diff --git a/net/core/dev.c b/net/core/dev.c
index 92c9258..a855e41 100644 (file)
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -365,12 +365,12 @@ static void list_netdevice(struct net_device *dev)
 
        ASSERT_RTNL();
 
-       write_lock_bh(&dev_base_lock);
+       write_lock(&dev_base_lock);
        list_add_tail_rcu(&dev->dev_list, &net->dev_base_head);
        netdev_name_node_add(net, dev->name_node);
        hlist_add_head_rcu(&dev->index_hlist,
                           dev_index_hash(net, dev->ifindex));
-       write_unlock_bh(&dev_base_lock);
+       write_unlock(&dev_base_lock);
 
        dev_base_seq_inc(net);
 }
@@ -383,11 +383,11 @@ static void unlist_netdevice(struct net_device *dev)
        ASSERT_RTNL();
 
        /* Unlink dev from the device chain */
-       write_lock_bh(&dev_base_lock);
+       write_lock(&dev_base_lock);
        list_del_rcu(&dev->dev_list);
        netdev_name_node_del(dev->name_node);
        hlist_del_rcu(&dev->index_hlist);
-       write_unlock_bh(&dev_base_lock);
+       write_unlock(&dev_base_lock);
 
        dev_base_seq_inc(dev_net(dev));
 }
@@ -1188,15 +1188,15 @@ rollback:
 
        netdev_adjacent_rename_links(dev, oldname);
 
-       write_lock_bh(&dev_base_lock);
+       write_lock(&dev_base_lock);
        netdev_name_node_del(dev->name_node);
-       write_unlock_bh(&dev_base_lock);
+       write_unlock(&dev_base_lock);
 
        synchronize_rcu();
 
-       write_lock_bh(&dev_base_lock);
+       write_lock(&dev_base_lock);
        netdev_name_node_add(net, dev->name_node);
-       write_unlock_bh(&dev_base_lock);
+       write_unlock(&dev_base_lock);
 
        ret = call_netdevice_notifiers(NETDEV_CHANGENAME, dev);
        ret = notifier_to_errno(ret);
@@ -1377,6 +1377,7 @@ static int __dev_open(struct net_device *dev, struct netlink_ext_ack *extack)
        int ret;
 
        ASSERT_RTNL();
+       dev_addr_check(dev);
 
        if (!netif_device_present(dev)) {
                /* may be detached because parent is runtime-suspended */
@@ -3395,7 +3396,7 @@ static netdev_features_t gso_features_check(const struct sk_buff *skb,
 {
        u16 gso_segs = skb_shinfo(skb)->gso_segs;
 
-       if (gso_segs > dev->gso_max_segs)
+       if (gso_segs > READ_ONCE(dev->gso_max_segs))
                return features & ~NETIF_F_GSO_MASK;
 
        if (!skb_shinfo(skb)->gso_type) {
@@ -3718,8 +3719,12 @@ no_lock_out:
         * separate lock before trying to get qdisc main lock.
         * This permits qdisc->running owner to get the lock more
         * often and dequeue packets faster.
+        * On PREEMPT_RT it is possible to preempt the qdisc owner during xmit
+        * and then other tasks will only enqueue packets. The packets will be
+        * sent after the qdisc owner is scheduled again. To prevent this
+        * scenario the task always serialize on the lock.
         */
-       contended = qdisc_is_running(q);
+       contended = qdisc_is_running(q) || IS_ENABLED(CONFIG_PREEMPT_RT);
        if (unlikely(contended))
                spin_lock(&q->busylock);
 
@@ -4092,7 +4097,10 @@ static int __dev_queue_xmit(struct sk_buff *skb, struct net_device *sb_dev)
        if (dev->flags & IFF_UP) {
                int cpu = smp_processor_id(); /* ok because BHs are off */
 
-               if (txq->xmit_lock_owner != cpu) {
+               /* Other cpus might concurrently change txq->xmit_lock_owner
+                * to -1 or to their cpu id, but not to our id.
+                */
+               if (READ_ONCE(txq->xmit_lock_owner) != cpu) {
                        if (dev_xmit_recursion())
                                goto recursion_alert;
 
@@ -6533,6 +6541,7 @@ static __latent_entropy void net_rx_action(struct softirq_action *h)
 
 struct netdev_adjacent {
        struct net_device *dev;
+       netdevice_tracker dev_tracker;
 
        /* upper master flag, there can only be one master device per list */
        bool master;
@@ -7297,7 +7306,7 @@ static int __netdev_adjacent_dev_insert(struct net_device *dev,
        adj->ref_nr = 1;
        adj->private = private;
        adj->ignore = false;
-       dev_hold(adj_dev);
+       dev_hold_track(adj_dev, &adj->dev_tracker, GFP_KERNEL);
 
        pr_debug("Insert adjacency: dev %s adj_dev %s adj->ref_nr %d; dev_hold on %s\n",
                 dev->name, adj_dev->name, adj->ref_nr, adj_dev->name);
@@ -7326,8 +7335,8 @@ remove_symlinks:
        if (netdev_adjacent_is_neigh_list(dev, adj_dev, dev_list))
                netdev_adjacent_sysfs_del(dev, adj_dev->name, dev_list);
 free_adj:
+       dev_put_track(adj_dev, &adj->dev_tracker);
        kfree(adj);
-       dev_put(adj_dev);
 
        return ret;
 }
@@ -7368,7 +7377,7 @@ static void __netdev_adjacent_dev_remove(struct net_device *dev,
        list_del_rcu(&adj->list);
        pr_debug("adjacency: dev_put for %s, because link removed from %s to %s\n",
                 adj_dev->name, dev->name, adj_dev->name);
-       dev_put(adj_dev);
+       dev_put_track(adj_dev, &adj->dev_tracker);
        kfree_rcu(adj, rcu);
 }
 
@@ -8557,35 +8566,17 @@ bool netdev_port_same_parent_id(struct net_device *a, struct net_device *b)
 EXPORT_SYMBOL(netdev_port_same_parent_id);
 
 /**
- *     dev_change_proto_down - update protocol port state information
+ *     dev_change_proto_down - set carrier according to proto_down.
+ *
  *     @dev: device
  *     @proto_down: new value
- *
- *     This info can be used by switch drivers to set the phys state of the
- *     port.
  */
 int dev_change_proto_down(struct net_device *dev, bool proto_down)
 {
-       const struct net_device_ops *ops = dev->netdev_ops;
-
-       if (!ops->ndo_change_proto_down)
+       if (!(dev->priv_flags & IFF_CHANGE_PROTO_DOWN))
                return -EOPNOTSUPP;
        if (!netif_device_present(dev))
                return -ENODEV;
-       return ops->ndo_change_proto_down(dev, proto_down);
-}
-EXPORT_SYMBOL(dev_change_proto_down);
-
-/**
- *     dev_change_proto_down_generic - generic implementation for
- *     ndo_change_proto_down that sets carrier according to
- *     proto_down.
- *
- *     @dev: device
- *     @proto_down: new value
- */
-int dev_change_proto_down_generic(struct net_device *dev, bool proto_down)
-{
        if (proto_down)
                netif_carrier_off(dev);
        else
@@ -8593,7 +8584,7 @@ int dev_change_proto_down_generic(struct net_device *dev, bool proto_down)
        dev->proto_down = proto_down;
        return 0;
 }
-EXPORT_SYMBOL(dev_change_proto_down_generic);
+EXPORT_SYMBOL(dev_change_proto_down);
 
 /**
  *     dev_change_proto_down_reason - proto down reason
@@ -9878,6 +9869,7 @@ static void netdev_wait_allrefs(struct net_device *dev)
                               netdev_unregister_timeout_secs * HZ)) {
                        pr_emerg("unregister_netdevice: waiting for %s to become free. Usage count = %d\n",
                                 dev->name, refcnt);
+                       ref_tracker_dir_print(&dev->refcnt_tracker, 10);
                        warning_time = jiffies;
                }
        }
@@ -10168,6 +10160,7 @@ struct net_device *alloc_netdev_mqs(int sizeof_priv, const char *name,
        dev = PTR_ALIGN(p, NETDEV_ALIGN);
        dev->padded = (char *)dev - (char *)p;
 
+       ref_tracker_dir_init(&dev->refcnt_tracker, 128);
 #ifdef CONFIG_PCPU_DEV_REFCNT
        dev->pcpu_refcnt = alloc_percpu(int);
        if (!dev->pcpu_refcnt)
@@ -10284,6 +10277,7 @@ void free_netdev(struct net_device *dev)
        list_for_each_entry_safe(p, n, &dev->napi_list, dev_list)
                netif_napi_del(p);
 
+       ref_tracker_dir_exit(&dev->refcnt_tracker);
 #ifdef CONFIG_PCPU_DEV_REFCNT
        free_percpu(dev->pcpu_refcnt);
        dev->pcpu_refcnt = NULL;