rtnetlink: unlock on error path in netlink_dump()
authorDan Carpenter <error27@gmail.com>
Wed, 15 Jun 2011 03:11:42 +0000 (03:11 +0000)
committerDavid S. Miller <davem@conan.davemloft.net>
Fri, 17 Jun 2011 03:51:35 +0000 (23:51 -0400)
In c7ac8679bec939 "rtnetlink: Compute and store minimum ifinfo dump
size", we moved the allocation under the lock so we need to unlock
on error path.

Signed-off-by: Dan Carpenter <error27@gmail.com>
Signed-off-by: David S. Miller <davem@conan.davemloft.net>
net/netlink/af_netlink.c

index 0b92f75491b19342cc30fae6e41c29f726ec7688..ca5276c5180446e9bd001643ff22e1d945a9bd24 100644 (file)
@@ -1676,7 +1676,7 @@ static int netlink_dump(struct sock *sk)
 
        skb = sock_rmalloc(sk, alloc_size, 0, GFP_KERNEL);
        if (!skb)
-               goto errout;
+               goto errout_skb;
 
        len = cb->dump(skb, cb);
 
@@ -1716,7 +1716,6 @@ static int netlink_dump(struct sock *sk)
 errout_skb:
        mutex_unlock(nlk->cb_mutex);
        kfree_skb(skb);
-errout:
        return err;
 }