[PATCH net 3/3] net: make sure devices go through netdev_wait_all_refs

[Jakub Kicinski <kuba@kernel.org>]

If register_netdevice() fails at the very last stage - the
notifier call - some subsystems may have already seen it and
grabbed a reference. struct net_device can't be freed right
away without calling netdev_wait_all_refs().

Now that we have a clean interface in form of dev->needs_free_netdev
and lenient free_netdev() we can undo what commit 93ee31f14f6f ("[NET]:
Fix free_netdev on register_netdev failure.") has done and complete
the unregistration path by bringing the net_set_todo() call back.

After registration fails user is still expected to explicitly
free the net_device, so make sure ->needs_free_netdev is cleared,
otherwise rolling back the registration will cause the old double
free for callers who release rtnl_lock before the free.

This also solves the problem of priv_destructor not being called
on notifier error.

net_set_todo() will be moved back into unregister_netdevice_queue()
in a follow up.

Reported-by: Hulk Robot <hulkci@huawei.com>
Reported-by: Yang Yingliang <yangyingliang@huawei.com>
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
---
 net/core/dev.c | 14 ++++----------
 1 file changed, 4 insertions(+), 10 deletions(-)

diff --git a/net/core/dev.c b/net/core/dev.c
index adde93cbca9f..0071a11a6dc3 100644
--- a/net/core/dev.c
+++ b/net/core/dev.c
@@ -10077,17 +10077,11 @@ int register_netdevice(struct net_device *dev)
 	ret = call_netdevice_notifiers(NETDEV_REGISTER, dev);
 	ret = notifier_to_errno(ret);
 	if (ret) {
+		/* Expect explicit free_netdev() on failure */
+		dev->needs_free_netdev = false;
 		rollback_registered(dev);
-		rcu_barrier();
-
-		dev->reg_state = NETREG_UNREGISTERED;
-		/* We should put the kobject that hold in
-		 * netdev_unregister_kobject(), otherwise
-		 * the net device cannot be freed when
-		 * driver calls free_netdev(), because the
-		 * kobject is being hold.
-		 */
-		kobject_put(&dev->dev.kobj);
+		net_set_todo(dev);
+		goto out;
 	}
 	/*
 	 *	Prevent userspace races by waiting until the network
-- 
2.26.2