]> git.hungrycats.org Git - linux/commitdiff
netlink: put module reference if dump start fails
authorJason A. Donenfeld <Jason@zx2c4.com>
Wed, 21 Feb 2018 03:41:59 +0000 (04:41 +0100)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Fri, 9 Mar 2018 06:47:40 +0000 (22:47 -0800)
[ Upstream commit b87b6194be631c94785fe93398651e804ed43e28 ]

Before, if cb->start() failed, the module reference would never be put,
because cb->cb_running is intentionally false at this point. Users are
generally annoyed by this because they can no longer unload modules that
leak references. Also, it may be possible to tediously wrap a reference
counter back to zero, especially since module.c still uses atomic_inc
instead of refcount_inc.

This patch expands the error path to simply call module_put if
cb->start() fails.

Fixes: 41c87425a1ac ("netlink: do not set cb_running if dump's start() errs")
Signed-off-by: Jason A. Donenfeld <Jason@zx2c4.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
net/netlink/af_netlink.c

index 84a4e4c3be4b9738ed9277a09d217327137588f8..ca9c0544c856ccc60ddcc8c1074ffb0acf6b8f67 100644 (file)
@@ -2275,7 +2275,7 @@ int __netlink_dump_start(struct sock *ssk, struct sk_buff *skb,
        if (cb->start) {
                ret = cb->start(cb);
                if (ret)
-                       goto error_unlock;
+                       goto error_put;
        }
 
        nlk->cb_running = true;
@@ -2295,6 +2295,8 @@ int __netlink_dump_start(struct sock *ssk, struct sk_buff *skb,
         */
        return -EINTR;
 
+error_put:
+       module_put(control->module);
 error_unlock:
        sock_put(sk);
        mutex_unlock(nlk->cb_mutex);