libceph: fix fault locking; close socket on lossy fault

If we fault on a lossy connection, we should still close the socket
immediately, and do so under the con mutex.

We should also take the con mutex before printing out the state bits in
the debug output.

Signed-off-by: Sage Weil <sage@inktank.com>

diff --git a/net/ceph/messenger.c b/net/ceph/messenger.c
index 07204f1..9aaf539 100644 (file)
--- a/net/ceph/messenger.c
+++ b/net/ceph/messenger.c
@@ -2330,22 +2330,23 @@ fault:
  */
 static void ceph_fault(struct ceph_connection *con)
 {
+       mutex_lock(&con->mutex);
+
        pr_err("%s%lld %s %s\n", ENTITY_NAME(con->peer_name),
               ceph_pr_addr(&con->peer_addr.in_addr), con->error_msg);
        dout("fault %p state %lu to peer %s\n",
             con, con->state, ceph_pr_addr(&con->peer_addr.in_addr));
 
-       if (test_bit(LOSSYTX, &con->flags)) {
-               dout("fault on LOSSYTX channel\n");
-               goto out;
-       }
-
-       mutex_lock(&con->mutex);
        if (test_bit(CLOSED, &con->state))
                goto out_unlock;
 
        con_close_socket(con);
 
+       if (test_bit(LOSSYTX, &con->flags)) {
+               dout("fault on LOSSYTX channel\n");
+               goto out_unlock;
+       }
+
        if (con->in_msg) {
                BUG_ON(con->in_msg->con != con);
                con->in_msg->con = NULL;
@@ -2392,7 +2393,6 @@ static void ceph_fault(struct ceph_connection *con)
 
 out_unlock:
        mutex_unlock(&con->mutex);
-out:
        /*
         * in case we faulted due to authentication, invalidate our
         * current tickets so that we can get new ones.