staging: lustre: lnet: Don't access NULL NI on failure path
authorDoug Oucharek <doug.s.oucharek@intel.com>
Thu, 9 Jun 2016 22:45:45 +0000 (18:45 -0400)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Fri, 10 Jun 2016 01:33:27 +0000 (18:33 -0700)
In kiblnd_passive_connect(), if we are failing the connection
attempt because we cannot find a valid NI (we have a NULL NI),
we were coring after the "goto fail" because the failure
path was assuming non-NULL NI.

This patch ensures we don't dereference a NULL NI on that
failure path.

Signed-off-by: Doug Oucharek <doug.s.oucharek@intel.com>
Intel-bug-id: https://jira.hpdd.intel.com/browse/LU-8022
Reviewed-on: http://review.whamcloud.com/19614
Reviewed-by: Dmitry Eremin <dmitry.eremin@intel.com>
Reviewed-by: James Simmons <uja.ornl@yahoo.com>
Reviewed-by: Matt Ezell <ezellma@ornl.gov>
Reviewed-by: Oleg Drokin <oleg.drokin@intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drivers/staging/lustre/lnet/klnds/o2iblnd/o2iblnd_cb.c

index bbfee53cfcf50296e52a7849b4f0964145db1b13..845e49a52430db2f091ed4ff7bbeeed8f52efad7 100644 (file)
@@ -2521,12 +2521,13 @@ kiblnd_passive_connect(struct rdma_cm_id *cmid, void *priv, int priv_nob)
        return 0;
 
  failed:
-       if (ni)
+       if (ni) {
                lnet_ni_decref(ni);
+               rej.ibr_cp.ibcp_queue_depth = kiblnd_msg_queue_size(version, ni);
+               rej.ibr_cp.ibcp_max_frags = kiblnd_rdma_frags(version, ni);
+       }
 
        rej.ibr_version             = version;
-       rej.ibr_cp.ibcp_queue_depth = kiblnd_msg_queue_size(version, ni);
-       rej.ibr_cp.ibcp_max_frags = kiblnd_rdma_frags(version, ni);
        kiblnd_reject(cmid, &rej);
 
        return -ECONNREFUSED;