open-mpi · bwbarrett · Jun 1, 2017 · May 5, 2017 · May 7, 2017 · May 8, 2017
diff --git a/.gitignore b/.gitignore
@@ -413,6 +413,7 @@ orte/test/mpi/memcached-dummy
 orte/test/mpi/coll_test
 orte/test/mpi/badcoll
 orte/test/mpi/iof
+orte/test/mpi/no-disconnect
 
 orte/test/system/radix
 orte/test/system/sigusr_trap

diff --git a/config/orte_check_slurm.m4 b/config/orte_check_slurm.m4
@@ -13,6 +13,7 @@
 # Copyright (c) 2009-2016 Cisco Systems, Inc.  All rights reserved.
 # Copyright (c) 2016      Los Alamos National Security, LLC. All rights
 #                         reserved.
+# Copyright (c) 2017      Intel, Inc. All rights reserved.
 # $COPYRIGHT$
 #
 # Additional copyrights may follow
@@ -68,6 +69,15 @@ AC_DEFUN([ORTE_CHECK_SLURM],[
                              [orte_check_slurm_happy="yes"],
                              [orte_check_slurm_happy="no"])])
 
+        # check to see if this is a Cray nativized slurm env.
+
+        slurm_cray_env=0
+        OPAL_CHECK_ALPS([orte_slurm_cray],
+                        [slurm_cray_env=1])
+
+        AC_DEFINE_UNQUOTED([SLURM_CRAY_ENV],[$slurm_cray_env],
+                           [defined to 1 if slurm cray env, 0 otherwise])
+
         OPAL_SUMMARY_ADD([[Resource Managers]],[[Slurm]],[$1],[$orte_check_slurm_happy])
     fi
 

diff --git a/ompi/dpm/dpm.c b/ompi/dpm/dpm.c
@@ -15,7 +15,7 @@
  * Copyright (c) 2009      Sun Microsystems, Inc.  All rights reserved.
  * Copyright (c) 2011-2015 Los Alamos National Security, LLC.  All rights
  *                         reserved.
- * Copyright (c) 2013-2016 Intel, Inc.  All rights reserved.
+ * Copyright (c) 2013-2017 Intel, Inc. All rights reserved.
  * Copyright (c) 2014-2017 Research Organization for Information Science
  *                         and Technology (RIST). All rights reserved.
  * $COPYRIGHT$
@@ -40,6 +40,7 @@
 #include "opal/util/argv.h"
 #include "opal/util/opal_getcwd.h"
 #include "opal/util/proc.h"
+#include "opal/util/show_help.h"
 #include "opal/dss/dss.h"
 #include "opal/mca/hwloc/base/base.h"
 #include "opal/mca/pmix/pmix.h"
@@ -112,6 +113,12 @@ int ompi_dpm_connect_accept(ompi_communicator_t *comm, int root,
     if (NULL == opal_pmix.publish || NULL == opal_pmix.connect ||
         NULL == opal_pmix.unpublish ||
        (NULL == opal_pmix.lookup && NULL == opal_pmix.lookup_nb)) {
+        /* print a nice message explaining we don't have support */
+        opal_show_help("help-mpi-runtime.txt", "noconxcpt", true);
+        return OMPI_ERR_NOT_SUPPORTED;
+    }
+    if (!ompi_rte_connect_accept_support(port_string)) {
+        /* they will have printed the help message */
         return OMPI_ERR_NOT_SUPPORTED;
     }
 

diff --git a/ompi/mca/rte/orte/rte_orte.h b/ompi/mca/rte/orte/rte_orte.h
@@ -1,7 +1,7 @@
 /*
  * Copyright (c) 2012-2013 Los Alamos National Security, LLC.
  *                         All rights reserved.
- * Copyright (c) 2013-2015 Intel, Inc. All rights reserved
+ * Copyright (c) 2013-2017 Intel, Inc. All rights reserved.
  * Copyright (c) 2014      Cisco Systems, Inc.  All rights reserved.
  * Copyright (c) 2014-2016 Research Organization for Information Science
  *                         and Technology (RIST). All rights reserved.
@@ -116,6 +116,9 @@ static inline orte_process_name_t * OMPI_CAST_RTE_NAME(opal_process_name_t * nam
 }
 #endif
 
+/* check dynamics support */
+OMPI_DECLSPEC bool ompi_rte_connect_accept_support(const char *port);
+
 END_C_DECLS
 
 #endif /* MCA_OMPI_RTE_ORTE_H */
diff --git a/ompi/mca/rte/orte/rte_orte_module.c b/ompi/mca/rte/orte/rte_orte_module.c
@@ -39,6 +39,7 @@
 #include "orte/mca/routed/routed.h"
 #include "orte/util/name_fns.h"
 #include "orte/util/session_dir.h"
+#include "orte/util/show_help.h"
 #include "orte/runtime/orte_globals.h"
 #include "orte/runtime/orte_wait.h"
 #include "orte/runtime/orte_data_server.h"
@@ -190,3 +191,47 @@ void ompi_rte_wait_for_debugger(void)
         opal_pmix.deregister_evhandler(handler, NULL, NULL);
     }
 }
+
+bool ompi_rte_connect_accept_support(const char *port)
+{
+    char *ptr, *tmp;
+    orte_process_name_t name;
+
+    /* were we launched by mpirun, or are we calling
+     * without a defined port? */
+    if (NULL == orte_process_info.my_hnp_uri ||
+        NULL == port || 0 == strlen(port)) {
+        return true;
+    }
+
+    /* is the job family in the port different than my own? */
+    tmp = strdup(port);  // protect input
+    if (NULL == (ptr = strchr(tmp, ':'))) {
+        /* this port didn't come from us! */
+        orte_show_help("help-orterun.txt", "orterun:malformedport", true);
+        free(tmp);
+        return false;
+    }
+    *ptr = '\0';
+    if (ORTE_SUCCESS != orte_util_convert_string_to_process_name(&name, tmp)) {
+        free(tmp);
+        orte_show_help("help-orterun.txt", "orterun:malformedport", true);
+        return false;
+    }
+    free(tmp);
+    if (ORTE_JOB_FAMILY(ORTE_PROC_MY_NAME->jobid) == ORTE_JOB_FAMILY(name.jobid)) {
+        /* same job family, so our infrastructure is adequate */
+        return true;
+    }
+
+    /* if the job family of the port is different than our own
+     * and we were launched by mpirun, then we require ompi-server
+     * support */
+    if (NULL == orte_data_server_uri) {
+        /* print a pretty help message */
+        orte_show_help("help-orterun.txt", "orterun:server-unavailable", true);
+        return false;
+    }
+
+    return true;
+}
diff --git a/ompi/runtime/help-mpi-runtime.txt b/ompi/runtime/help-mpi-runtime.txt
@@ -12,6 +12,7 @@
 #                         All rights reserved.
 # Copyright (c) 2007-2015 Cisco Systems, Inc.  All rights reserved.
 # Copyright (c) 2013      NVIDIA Corporation.  All rights reserved.
+# Copyright (c) 2017      Intel, Inc. All rights reserved.
 # $COPYRIGHT$
 #
 # Additional copyrights may follow
@@ -93,3 +94,13 @@ Open MPI with --enable-heterogeneous.
 [no cuda support]
 The user requested CUDA support with the --mca mpi_cuda_support 1 flag
 but the library was not compiled with any support.
+#
+[noconxcpt]
+The user has called an operation involving MPI_Connect and/or MPI_Accept,
+but this environment lacks the necessary infrastructure support for
+that operation. Open MPI relies on the PMIx_Publish/Lookup (or one of
+its predecessors) APIs for this operation.
+
+This typically happens when launching outside of mpirun where the underlying
+resource manager does not provide publish/lookup support. One way of solving
+the problem is to simply use mpirun to start the application.
diff --git a/opal/include/opal/constants.h b/opal/include/opal/constants.h
@@ -10,7 +10,7 @@
  * Copyright (c) 2004-2005 The Regents of the University of California.
  *                         All rights reserved.
  * Copyright (c) 2010-2012 Cisco Systems, Inc.  All rights reserved.
- * Copyright (c) 2014-2016 Intel, Inc. All rights reserved.
+ * Copyright (c) 2014-2017 Intel, Inc. All rights reserved.
  * $COPYRIGHT$
  *
  * Additional copyrights may follow
@@ -94,10 +94,11 @@ enum {
     OPAL_ERR_PROC_RESTART                   = (OPAL_ERR_BASE - 63),
     OPAL_ERR_PROC_CHECKPOINT                = (OPAL_ERR_BASE - 64),
     OPAL_ERR_PROC_MIGRATE                   = (OPAL_ERR_BASE - 65),
-    OPAL_ERR_EVENT_REGISTRATION             = (OPAL_ERR_BASE - 66)
+    OPAL_ERR_EVENT_REGISTRATION             = (OPAL_ERR_BASE - 66),
+    OPAL_ERR_HEARTBEAT_ALERT                = (OPAL_ERR_BASE - 67),
+    OPAL_ERR_FILE_ALERT                     = (OPAL_ERR_BASE - 68)
 };
 
 #define OPAL_ERR_MAX                (OPAL_ERR_BASE - 100)
 
 #endif /* OPAL_CONSTANTS_H */
-
diff --git a/opal/mca/pmix/base/base.h b/opal/mca/pmix/base/base.h
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2014-2016 Intel, Inc. All rights reserved.
+ * Copyright (c) 2014-2017 Intel, Inc. All rights reserved.
  * $COPYRIGHT$
  *
  * Additional copyrights may follow
@@ -57,6 +57,7 @@ OPAL_DECLSPEC void opal_pmix_base_set_evbase(opal_event_base_t *evbase);
 
 typedef struct {
     opal_event_base_t *evbase;
+    int timeout;
 } opal_pmix_base_t;
 
 extern opal_pmix_base_t opal_pmix_base;

diff --git a/opal/mca/pmix/base/pmix_base_fns.c b/opal/mca/pmix/base/pmix_base_fns.c
@@ -2,7 +2,7 @@
 /*
  * Copyright (c) 2012-2015 Los Alamos National Security, LLC.  All rights
  *                         reserved.
- * Copyright (c) 2014-2015 Intel, Inc. All rights reserved.
+ * Copyright (c) 2014-2017 Intel, Inc. All rights reserved.
  * Copyright (c) 2014-2017 Research Organization for Information Science
  *                         and Technology (RIST). All rights reserved.
  * Copyright (c) 2016      Mellanox Technologies, Inc.
@@ -118,6 +118,13 @@ static void lookup_cbfunc(int status, opal_list_t *data, void *cbdata)
     cd->active = false;
 }
 
+static void opcbfunc(int status, void *cbdata)
+{
+    struct lookup_caddy_t *cd = (struct lookup_caddy_t*)cbdata;
+    cd->status = status;
+    cd->active = false;
+}
+
 int opal_pmix_base_exchange(opal_value_t *indat,
                             opal_pmix_pdata_t *outdat,
                             int timeout)
@@ -141,17 +148,37 @@ int opal_pmix_base_exchange(opal_value_t *indat,
     opal_list_append(&ilist, &info->super);
 
     /* publish it with "session" scope */
-    rc = opal_pmix.publish(&ilist);
-    OPAL_LIST_DESTRUCT(&ilist);
-    if (OPAL_SUCCESS != rc) {
-        OPAL_ERROR_LOG(rc);
-        return rc;
+    if (NULL == opal_pmix.publish_nb) {
+        rc = opal_pmix.publish(&ilist);
+        OPAL_LIST_DESTRUCT(&ilist);
+        if (OPAL_SUCCESS != rc) {
+            OPAL_ERROR_LOG(rc);
+            return rc;
+        }
+    } else {
+        caddy.status = -1;
+        caddy.active = true;
+        caddy.pdat = NULL;
+        rc = opal_pmix.publish_nb(&ilist, opcbfunc, &caddy);
+        if (OPAL_SUCCESS != rc) {
+            OPAL_ERROR_LOG(rc);
+            OPAL_LIST_DESTRUCT(&ilist);
+            return rc;
+        }
+        while (caddy.active) {
+            usleep(10);
+        }
+        OPAL_LIST_DESTRUCT(&ilist);
+        if (OPAL_SUCCESS != caddy.status) {
+            OPAL_ERROR_LOG(caddy.status);
+            return caddy.status;
+        }
     }
 
-   /* lookup the other side's info - if a non-blocking form
-    * of lookup isn't available, then we use the blocking
-    * form and trust that the underlying system will WAIT
-    * until the other side publishes its data */
+    /* lookup the other side's info - if a non-blocking form
+     * of lookup isn't available, then we use the blocking
+     * form and trust that the underlying system will WAIT
+     * until the other side publishes its data */
     pdat = OBJ_NEW(opal_pmix_pdata_t);
     pdat->value.key = strdup(outdat->value.key);
     pdat->value.type = outdat->value.type;
@@ -169,7 +196,12 @@ int opal_pmix_base_exchange(opal_value_t *indat,
     info = OBJ_NEW(opal_value_t);
     info->key = strdup(OPAL_PMIX_TIMEOUT);
     info->type = OPAL_INT;
-    info->data.integer = timeout;
+    if (0 < opal_pmix_base.timeout) {
+        /* the user has overridden the default */
+        info->data.integer = opal_pmix_base.timeout;
+    } else {
+        info->data.integer = timeout;
+    }
     opal_list_append(&mlist, &info->super);
 
     /* if a non-blocking version of lookup isn't
@@ -185,6 +217,7 @@ int opal_pmix_base_exchange(opal_value_t *indat,
             return rc;
         }
     } else {
+        caddy.status = -1;
         caddy.active = true;
         caddy.pdat = pdat;
         keys = NULL;

diff --git a/opal/mca/pmix/base/pmix_base_frame.c b/opal/mca/pmix/base/pmix_base_frame.c
@@ -1,5 +1,5 @@
 /*
- * Copyright (c) 2014-2016 Intel, Inc.  All rights reserved.
+ * Copyright (c) 2014-2017 Intel, Inc. All rights reserved.
  * Copyright (c) 2015-2016 Cisco Systems, Inc.  All rights reserved.
  * $COPYRIGHT$
  *
@@ -47,6 +47,12 @@ static int opal_pmix_base_frame_register(mca_base_register_flag_t flags)
     (void) mca_base_var_register("opal", "pmix", "base", "collect_data", "Collect all data during modex",
                                  MCA_BASE_VAR_TYPE_BOOL, NULL, 0, 0, OPAL_INFO_LVL_9,
                                  MCA_BASE_VAR_SCOPE_READONLY, &opal_pmix_collect_all_data);
+
+    opal_pmix_base.timeout = -1;
+    (void) mca_base_var_register("opal", "pmix", "base", "exchange_timeout",
+                                 "Time (in seconds) to wait for a data exchange to complete",
+                                 MCA_BASE_VAR_TYPE_INT, NULL, 0, 0, OPAL_INFO_LVL_3,
+                                 MCA_BASE_VAR_SCOPE_READONLY, &opal_pmix_base.timeout);
     return OPAL_SUCCESS;
 }
 

diff --git a/opal/mca/pmix/ext2x/pmix2x.c b/opal/mca/pmix/ext2x/pmix2x.c
@@ -1,6 +1,6 @@
 /* -*- Mode: C; c-basic-offset:4 ; indent-tabs-mode:nil -*- */
 /*
- * Copyright (c) 2014-2016 Intel, Inc.  All rights reserved.
+ * Copyright (c) 2014-2017 Intel, Inc. All rights reserved.
  * Copyright (c) 2014-2017 Research Organization for Information Science
  *                         and Technology (RIST). All rights reserved.
  * Copyright (c) 2014-2015 Mellanox Technologies, Inc.
@@ -352,7 +352,7 @@ static void _event_hdlr(int sd, short args, void *cbdata)
     if (NULL != chain->final_cbfunc) {
         chain->final_cbfunc(PMIX_SUCCESS, chain->final_cbdata);
     }
-    
+
     OBJ_RELEASE(chain);
 
     return;

diff --git a/opal/mca/pmix/pmix2x/configure.m4 b/opal/mca/pmix/pmix2x/configure.m4
@@ -49,15 +49,27 @@ AC_DEFUN([MCA_opal_pmix_pmix2x_CONFIG],[
         opal_pmix_pmix2x_sm_flag=--disable-dstore
     fi
 
-    opal_pmix_pmix2x_args="--with-pmix-symbol-rename=OPAL_MCA_PMIX2X_ $opal_pmix_pmix2x_sm_flag --without-tests-examples --disable-visibility --enable-embedded-libevent --with-libevent-header=\\\"opal/mca/event/$opal_event_base_include\\\""
+    AC_ARG_ENABLE([pmix-timing],
+                  [AC_HELP_STRING([--enable-pmix-timing],
+                                  [Enable PMIx timing measurements (default: disabled)])])
+    AC_MSG_CHECKING([if PMIx timing is enabled])
+    if test "$enable_pmix_timing" == "yes"; then
+        AC_MSG_RESULT([yes])
+        opal_pmix_pmix2x_timing_flag=--enable-pmix-timing
+    else
+        AC_MSG_RESULT([no (disabled)])
+        opal_pmix_pmix2x_timing_flag=--disable-pmix-timing
+    fi
+
+    opal_pmix_pmix2x_args="--with-pmix-symbol-rename=OPAL_MCA_PMIX2X_ $opal_pmix_pmix2x_sm_flag $opal_pmix_pmix2x_timing_flag --without-tests-examples --disable-pmix-backward-compatibility --disable-visibility --enable-embedded-libevent --with-libevent-header=\\\"opal/mca/event/$opal_event_base_include\\\" --enable-embedded-mode"
     AS_IF([test "$enable_debug" = "yes"],
           [opal_pmix_pmix2x_args="--enable-debug $opal_pmix_pmix2x_args"
            CFLAGS="$OPAL_CFLAGS_BEFORE_PICKY $OPAL_VISIBILITY_CFLAGS -g"],
           [opal_pmix_pmix2x_args="--disable-debug $opal_pmix_pmix2x_args"
            CFLAGS="$OPAL_CFLAGS_BEFORE_PICKY $OPAL_VISIBILITY_CFLAGS"])
     AS_IF([test "$with_devel_headers" = "yes"],
           [opal_pmix_pmix2x_args="--with-devel-headers $opal_pmix_pmix2x_args"],
-          [opal_pmix_pmix2x_args="--enable-embedded-mode $opal_pmix_pmix2x_args"])
+          [opal_pmix_pmix2x_args=$opal_pmix_pmix2x_args])
     CPPFLAGS="-I$OPAL_TOP_SRCDIR -I$OPAL_TOP_BUILDDIR -I$OPAL_TOP_SRCDIR/opal/include -I$OPAL_TOP_BUILDDIR/opal/include $CPPFLAGS"
 
     OPAL_CONFIG_SUBDIR([$opal_pmix_pmix2x_basedir/pmix],

diff --git a/opal/mca/pmix/pmix2x/pmix/NEWS b/opal/mca/pmix/pmix2x/pmix/NEWS
@@ -1,4 +1,5 @@
-Copyright (c) 2015-2016 Intel, Inc.  All rights reserved.
+Copyright (c) 2015-2017 Intel, Inc. All rights reserved.
+Copyright (c) 2017      IBM Corporation.  All rights reserved.
 $COPYRIGHT$
 
 Additional copyrights may follow
@@ -23,6 +24,32 @@ current release as well as the "stable" bug fix release branch.
 Master (not on release branches yet)
 ------------------------------------
 
+1.2.2 -- 21 March 2017
+----------------------
+- Compiler fix for Sun/Oracle CC (PR #322)
+- Fix missing include (PR #326)
+- Improve error checking around posix_fallocate (PR #329)
+- Fix possible memory corruption (PR #331)
+
+
+1.2.1 -- 21 Feb. 2017
+----------------------
+- dstore: Fix data corruption bug in key overwrite cases
+- dstore: Performance and scalability fixes
+- sm: Use posix_fallocate() before mmap
+- pmi1/pmi2: Restore support
+- dstore: Fix extension slot size allocation (Issue #280)
+
+
+1.2.0 -- 14 Dec. 2016
+----------------------
+- Add shared memory data storage (dstore) option. Default: enabled
+  Configure option: --disable-dstore
+- PMIx_Commit performance improvements
+- Disable errhandler support
+- Keep job info in the shared memory dstore
+- PMIx_Get performance and memory improvements
+
 1.1.5
 -----
 - Add pmix_version.h to support direct detection of PMIx library version