failover: return error when SRV lookup returned only duplicates

https://fedorahosted.org/sssd/ticket/1947

Otherwise we risk that the meta server is removed from the server list,
but without a chance to return, because there may be no fo_server with
srv_data = meta.

Also if state->meta->next is NULL (it is still orphaned because we try
to errornously expand it without invoking collapse first), state->out
will be NULL and SSSD will crash.

New error code: ERR_SRV_DUPLICATES

3 files changed, 23 insertions, 2 deletions

diff --git a/src/providers/fail_over.c b/src/providers/fail_over.c
index c4dd1ea01..c5db70242 100644
--- a/src/providers/fail_over.c
+++ b/src/providers/fail_over.c
@@ -1288,21 +1288,40 @@ resolve_srv_done(struct tevent_req *subreq)
                                      backup_servers, num_backup_servers,
                                      state->meta->srv_data,
                                      state->meta->user_data,
-                                     false, NULL);
+                                     false, &last_server);
             if (ret != EOK) {
                 goto done;
             }
         }
 
+        if (last_server == state->meta) {
+            /* SRV lookup returned only those servers
+             * that are already present. */
+            DEBUG(SSSDBG_TRACE_FUNC, ("SRV lookup did not return "
+                                      "any new server.\n"));
+            ret = ERR_SRV_DUPLICATES;
+            goto done;
+        }
+
+        /* At least one new server was inserted.
+         * We will return the first new server. */
+        if (state->meta->next == NULL) {
+            DEBUG(SSSDBG_CRIT_FAILURE,
+                 ("BUG: state->meta->next is NULL\n"));
+            ret = ERR_INTERNAL;
+            goto done;
+        }
+
         state->out = state->meta->next;
 
+        /* And remove meta server from the server list. It will be
+         * inserted again during srv collapse. */
         DLIST_REMOVE(state->service->server_list, state->meta);
         if (state->service->last_tried_server == state->meta) {
             state->service->last_tried_server = state->out;
         }
 
         set_srv_data_status(state->meta->srv_data, SRV_RESOLVED);
-
         ret = EOK;
         break;
     case ERR_SRV_NOT_FOUND:
diff --git a/src/util/util_errors.c b/src/util/util_errors.c
index 22a3045a6..015c5acaf 100644
--- a/src/util/util_errors.c
+++ b/src/util/util_errors.c
@@ -44,6 +44,7 @@ struct err_string error_to_str[] = {
     { "Host Access Denied" }, /* ERR_ACCESS_DENIED */
     { "SRV record not found" }, /* ERR_SRV_NOT_FOUND */
     { "SRV lookup error" }, /* ERR_SRV_LOOKUP_ERROR */
+    { "SRV lookup did not return any new server "}, /* ERR_SRV_DUPLICATES */
     { "Dynamic DNS update failed" }, /* ERR_DYNDNS_FAILED */
     { "Dynamic DNS update timed out" }, /* ERR_DYNDNS_TIMEOUT */
     { "Dynamic DNS update not possible while offline" }, /* ERR_DYNDNS_OFFLINE */
diff --git a/src/util/util_errors.h b/src/util/util_errors.h
index 65d37aedb..ca4472823 100644
--- a/src/util/util_errors.h
+++ b/src/util/util_errors.h
@@ -66,6 +66,7 @@ enum sssd_errors {
     ERR_ACCESS_DENIED,
     ERR_SRV_NOT_FOUND,
     ERR_SRV_LOOKUP_ERROR,
+    ERR_SRV_DUPLICATES,
     ERR_DYNDNS_FAILED,
     ERR_DYNDNS_TIMEOUT,
     ERR_DYNDNS_OFFLINE,