mctp: make __mctp_dev_get() take a refcount hold

Previously there was a race that could allow the mctp_dev refcount
to hit zero:

rcu_read_lock();
mdev = __mctp_dev_get(dev);
// mctp_unregister() happens here, mdev->refs hits zero
mctp_dev_hold(dev);
rcu_read_unlock();

Now we make __mctp_dev_get() take the hold itself. It is safe to test
against the zero refcount because __mctp_dev_get() is called holding
rcu_read_lock and mctp_dev uses kfree_rcu().

Reported-by: Jakub Kicinski <[e-mail]>
Signed-off-by: Matt Johnston <[e-mail]>
Signed-off-by: David S. Miller <[e-mail]>

net/mctp/device.c

···
 	size_t a_idx;
 };
 
-/* unlocked: caller must hold rcu_read_lock */
+/* unlocked: caller must hold rcu_read_lock.
+ * Returned mctp_dev has its refcount incremented, or NULL if unset.
+ */
 struct mctp_dev *__mctp_dev_get(const struct net_device *dev)
 {
-	return rcu_dereference(dev->mctp_ptr);
+	struct mctp_dev *mdev = rcu_dereference(dev->mctp_ptr);
+
+	/* RCU guarantees that any mdev is still live.
+	 * Zero refcount implies a pending free, return NULL.
+	 */
+	if (mdev)
+		if (!refcount_inc_not_zero(&mdev->refs))
+			return NULL;
+	return mdev;
 }
 
+/* Returned mctp_dev does not have refcount incremented. The returned pointer
+ * remains live while rtnl_lock is held, as that prevents mctp_unregister()
+ */
 struct mctp_dev *mctp_dev_get_rtnl(const struct net_device *dev)
 {
 	return rtnl_dereference(dev->mctp_ptr);
···
 				if (mdev) {
 					rc = mctp_dump_dev_addrinfo(mdev,
 								    skb, cb);
+					mctp_dev_put(mdev);
 					// Error indicates full buffer, this
 					// callback will get retried.
 					if (rc < 0)
···
 
 void mctp_dev_put(struct mctp_dev *mdev)
 {
-	if (refcount_dec_and_test(&mdev->refs)) {
+	if (mdev && refcount_dec_and_test(&mdev->refs)) {
 		dev_put(mdev->dev);
 		kfree_rcu(mdev, rcu);
 	}
···
 	if (!mdev)
 		return 0;
 	ret = nla_total_size(4); /* IFLA_MCTP_NET */
+	mctp_dev_put(mdev);
 	return ret;
 }
 

net/mctp/route.c

···
 {
 	struct mctp_sock *msk = container_of(sk, struct mctp_sock, sk);
 	struct mctp_skb_cb *cb = mctp_cb(skb);
-	struct mctp_route tmp_rt;
+	struct mctp_route tmp_rt = {0};
 	struct mctp_sk_key *key;
 	struct net_device *dev;
 	struct mctp_hdr *hdr;
···
 		mctp_route_release(rt);
 
 	dev_put(dev);
+	mctp_dev_put(tmp_rt.dev);
 
 	return rc;
 
···
 
 	rt->output(rt, skb);
 	mctp_route_release(rt);
+	mctp_dev_put(mdev);
 
 	return NET_RX_SUCCESS;
 
 err_drop:
 	kfree_skb(skb);
+	mctp_dev_put(mdev);
 	return NET_RX_DROP;
 }
 

net/mctp/test/utils.c

···
 
 	rcu_read_lock();
 	dev->mdev = __mctp_dev_get(ndev);
-	mctp_dev_hold(dev->mdev);
 	rcu_read_unlock();
 
 	return dev;