Merge pull request open-mpi#6010 from ICLDisco/export/orte_crashfini

Export/orte crashfini

Author: abouteiller

orte/mca/errmgr/default_app/errmgr_default_app.c

@@ -4,7 +4,7 @@
  *
  * Copyright (c) 2010      Cisco Systems, Inc.  All rights reserved.
  *
- * Copyright (c) 2004-2006 The University of Tennessee and The University
+ * Copyright (c) 2004-2018 The University of Tennessee and The University
  *                         of Tennessee Research Foundation.  All rights
  *                         reserved.
  * Copyright (c) 2011-2013 Los Alamos National Security, LLC.
@@ -101,14 +101,14 @@ static void notify_cbfunc(int status,
             state = ORTE_PROC_STATE_TERMINATED;
     }
 
+    /* push it into our event base */
+    ORTE_ACTIVATE_PROC_STATE((orte_process_name_t*)source, state);
+
     /* let the caller know we processed this, but allow the
      * chain to continue */
     if (NULL != cbfunc) {
         cbfunc(ORTE_SUCCESS, NULL, NULL, NULL, cbdata);
     }
-
-    /* push it into our event base */
-    ORTE_ACTIVATE_PROC_STATE((orte_process_name_t*)source, state);
 }
 
 /************************
@@ -143,6 +143,7 @@ static int finalize(void)
 {
     if (SIZE_MAX != myerrhandle) {
         opal_pmix.deregister_evhandler(myerrhandle, NULL, NULL);
+        myerrhandle = SIZE_MAX;
     }
     return ORTE_SUCCESS;
 }

orte/mca/ess/base/ess_base_std_orted.c

@@ -2,7 +2,7 @@
  * Copyright (c) 2004-2010 The Trustees of Indiana University and Indiana
  *                         University Research and Technology
  *                         Corporation.  All rights reserved.
- * Copyright (c) 2004-2011 The University of Tennessee and The University
+ * Copyright (c) 2004-2018 The University of Tennessee and The University
  *                         of Tennessee Research Foundation.  All rights
  *                         reserved.
  * Copyright (c) 2004-2005 High Performance Computing Center Stuttgart,
@@ -669,7 +669,9 @@ int orte_ess_base_orted_finalize(void)
     (void) mca_base_framework_close(&orte_filem_base_framework);
     (void) mca_base_framework_close(&orte_grpcomm_base_framework);
     (void) mca_base_framework_close(&orte_iof_base_framework);
-    (void) mca_base_framework_close(&orte_errmgr_base_framework);
+    /* first stage shutdown of the errmgr, deregister the handler but keep
+     * the required facilities until the rml and oob are offline */
+    orte_errmgr.finalize();
     (void) mca_base_framework_close(&orte_plm_base_framework);
     /* make sure our local procs are dead */
     orte_odls.kill_local_procs(NULL);
@@ -678,6 +680,7 @@ int orte_ess_base_orted_finalize(void)
     (void) mca_base_framework_close(&orte_routed_base_framework);
     (void) mca_base_framework_close(&orte_rml_base_framework);
     (void) mca_base_framework_close(&orte_oob_base_framework);
+    (void) mca_base_framework_close(&orte_errmgr_base_framework);
     (void) mca_base_framework_close(&orte_state_base_framework);
     /* remove our use of the session directory tree */
     orte_session_dir_finalize(ORTE_PROC_MY_NAME);

orte/mca/ess/hnp/ess_hnp_module.c

@@ -3,7 +3,7 @@
  * Copyright (c) 2004-2010 The Trustees of Indiana University and Indiana
  *                         University Research and Technology
  *                         Corporation.  All rights reserved.
- * Copyright (c) 2004-2011 The University of Tennessee and The University
+ * Copyright (c) 2004-2018 The University of Tennessee and The University
  *                         of Tennessee Research Foundation.  All rights
  *                         reserved.
  * Copyright (c) 2004-2005 High Performance Computing Center Stuttgart,
@@ -799,8 +799,9 @@ static int rte_finalize(void)
     (void) mca_base_framework_close(&orte_grpcomm_base_framework);
     (void) mca_base_framework_close(&orte_routed_base_framework);
     (void) mca_base_framework_close(&orte_plm_base_framework);
-    (void) mca_base_framework_close(&orte_errmgr_base_framework);
-    (void) mca_base_framework_close(&orte_state_base_framework);
+    /* first stage shutdown of the errmgr, deregister the handler but keep
+     * the required facilities until the rml and oob are offline */
+    orte_errmgr.finalize();
 
     /* cleanup the pstat stuff */
     (void) mca_base_framework_close(&opal_pstat_base_framework);
@@ -816,6 +817,8 @@ static int rte_finalize(void)
     /* shutdown the messaging frameworks */
     (void) mca_base_framework_close(&orte_rml_base_framework);
     (void) mca_base_framework_close(&orte_oob_base_framework);
+    (void) mca_base_framework_close(&orte_errmgr_base_framework);
+    (void) mca_base_framework_close(&orte_state_base_framework);
 
     /* remove our use of the session directory tree */
     orte_session_dir_finalize(ORTE_PROC_MY_NAME);

orte/mca/iof/hnp/iof_hnp_read.c

@@ -2,7 +2,7 @@
  * Copyright (c) 2004-2007 The Trustees of Indiana University and Indiana
  *                         University Research and Technology
  *                         Corporation.  All rights reserved.
- * Copyright (c) 2004-2011 The University of Tennessee and The University
+ * Copyright (c) 2004-2018 The University of Tennessee and The University
  *                         of Tennessee Research Foundation.  All rights
  *                         reserved.
  * Copyright (c) 2004-2005 High Performance Computing Center Stuttgart,
@@ -284,6 +284,8 @@ void orte_iof_hnp_read_local_handler(int fd, short event, void *cbdata)
         /* if we read 0 bytes from the stdout/err/diag, there is
          * nothing to output - release the appropriate event.
          * This will delete the read event and close the file descriptor */
+        /* make sure we don't do recursive delete on the proct */
+        OBJ_RETAIN(proct);
         if (rev->tag & ORTE_IOF_STDOUT) {
             orte_iof_base_static_dump_output(proct->revstdout);
             OBJ_RELEASE(proct->revstdout);
@@ -305,6 +307,7 @@ void orte_iof_hnp_read_local_handler(int fd, short event, void *cbdata)
             /* this proc's iof is complete */
             ORTE_ACTIVATE_PROC_STATE(&proct->name, ORTE_PROC_STATE_IOF_COMPLETE);
         }
+        OBJ_RELEASE(proct);
         return;
     }