netns: foreach_netdev_safe is insufficient in default_device_exit

During network namespace teardown we either move or delete
all of the network devices associated with a network namespace.
In the case of veth devices deleting one will also delete it's
pair device.  If both devices are in the same network namespace
then for_each_netdev_safe is insufficient as next may point
to the second veth device we have deleted.

To avoid problems I do what we do in __rtnl_kill_links and
restart the scan of the device list, after we have deleted
a device.

Currently dev_change_netnamespace does not appear to suffer from
this problem, but wireless devices are also paired and likely
should be moved between network namespaces together.  So I have
errored on the side of caution and restart the scan of the network
devices in that case as well.

Signed-off-by: Eric W. Biederman <ebiederm@aristanetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
diff --git a/net/core/dev.c b/net/core/dev.c
index 4464240..09c66a4 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -5066,13 +5066,14 @@
 
 static void __net_exit default_device_exit(struct net *net)
 {
-	struct net_device *dev, *next;
+	struct net_device *dev;
 	/*
 	 * Push all migratable of the network devices back to the
 	 * initial network namespace
 	 */
 	rtnl_lock();
-	for_each_netdev_safe(net, dev, next) {
+restart:
+	for_each_netdev(net, dev) {
 		int err;
 		char fb_name[IFNAMSIZ];
 
@@ -5083,7 +5084,7 @@
 		/* Delete virtual devices */
 		if (dev->rtnl_link_ops && dev->rtnl_link_ops->dellink) {
 			dev->rtnl_link_ops->dellink(dev);
-			continue;
+			goto restart;
 		}
 
 		/* Push remaing network devices to init_net */
@@ -5094,6 +5095,7 @@
 				__func__, dev->name, err);
 			BUG();
 		}
+		goto restart;
 	}
 	rtnl_unlock();
 }