Update to PMIx v2.0alpha

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

opal/mca/pmix/pmix2x/pmix/VERSION

@@ -23,14 +23,14 @@ release=0
 # The only requirement is that it must be entirely printable ASCII
 # characters and have no white space.
 
-greek=
+greek=a1
 
 # If repo_rev is empty, then the repository version number will be
 # obtained during "make dist" via the "git describe --tags --always"
 # command, or with the date (if "git describe" fails) in the form of
 # "date<date>".
 
-repo_rev=git4cdd5e0
+repo_rev=gitc442ba8
 
 # If tarball_version is not empty, it is used as the version string in
 # the tarball filename, regardless of all other versions listed in
@@ -44,7 +44,7 @@ tarball_version=
 
 # The date when this release was created
 
-date="Mar 11, 2017"
+date="Apr 02, 2017"
 
 # The shared library version of each of PMIx's public libraries.
 # These versions are maintained in accordance with the "Library

opal/mca/pmix/pmix2x/pmix/config/pmix_functions.m4

@@ -13,7 +13,9 @@ dnl All rights reserved.
 dnl Copyright (c) 2007 Sun Microsystems, Inc. All rights reserved.
 dnl Copyright (c) 2009 Oak Ridge National Labs. All rights reserved.
 dnl Copyright (c) 2009-2016 Cisco Systems, Inc. All rights reserved.
-dnl Copyright (c) 2013-2016 Intel, Inc. All rights reserved.
+dnl Copyright (c) 2013-2017 Intel, Inc. All rights reserved.
+dnl Copyright (c) 2017 Research Organization for Information Science
+dnl and Technology (RIST). All rights reserved.
 dnl
 dnl $COPYRIGHT$
 dnl
@@ -278,7 +280,7 @@ for val in ${$1}; do
  # http://www.open-mpi.org/community/lists/devel/2012/08/11362.php).
 
  case $val in
- -Xclang)
+ -Xclang|-Xg)
  pmix_found=0
  pmix_i=`expr $pmix_count + 1`
  ;;
@@ -366,7 +368,7 @@ AC_DEFUN([PMIX_FLAGS_UNIQ],[
  # https://github.com/open-mpi/ompi/issues/324).
 
  case $val in
- -Xclang)
+ -Xclang|-Xg)
  pmix_found=0
  pmix_i=`expr $pmix_count + 1`
  ;;

opal/mca/pmix/pmix2x/pmix/examples/Makefile.am

@@ -21,7 +21,7 @@
 
 AM_CPPFLAGS = -I$(top_builddir)/src -I$(top_builddir)/src/include -I$(top_builddir)/include -I$(top_builddir)/include/pmix
 
-noinst_PROGRAMS = client dmodex dynamic fault pub tool debugger debuggerd alloc
+noinst_PROGRAMS = client dmodex dynamic fault pub tool debugger debuggerd alloc jctrl
 if !WANT_HIDDEN
 # these examples use internal symbols
 # use --disable-visibility
@@ -40,11 +40,14 @@ debuggerd_SOURCES = debuggerd.c
 debuggerd_LDFLAGS = $(PMIX_PKG_CONFIG_LDFLAGS)
 debuggerd_LDADD = $(top_builddir)/src/libpmix.la
 
-
 alloc_SOURCES = alloc.c
 alloc_LDFLAGS = $(PMIX_PKG_CONFIG_LDFLAGS)
 alloc_LDADD = $(top_builddir)/src/libpmix.la
 
+jctrl_SOURCES = jctrl.c
+jctrl_LDFLAGS = $(PMIX_PKG_CONFIG_LDFLAGS)
+jctrl_LDADD = $(top_builddir)/src/libpmix.la
+
 dmodex_SOURCES = dmodex.c
 dmodex_LDFLAGS = $(PMIX_PKG_CONFIG_LDFLAGS)
 dmodex_LDADD = $(top_builddir)/src/libpmix.la

opal/mca/pmix/pmix2x/pmix/examples/jctrl.c

@@ -0,0 +1,229 @@
+/*
+ * 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
+ * of Tennessee Research Foundation. All rights
+ * reserved.
+ * Copyright (c) 2004-2005 High Performance Computing Center Stuttgart,
+ * University of Stuttgart. All rights reserved.
+ * Copyright (c) 2004-2005 The Regents of the University of California.
+ * All rights reserved.
+ * Copyright (c) 2006-2013 Los Alamos National Security, LLC.
+ * All rights reserved.
+ * Copyright (c) 2009-2012 Cisco Systems, Inc. All rights reserved.
+ * Copyright (c) 2011 Oak Ridge National Labs. All rights reserved.
+ * Copyright (c) 2013-2017 Intel, Inc. All rights reserved.
+ * Copyright (c) 2015 Mellanox Technologies, Inc. All rights reserved.
+ * $COPYRIGHT$
+ *
+ * Additional copyrights may follow
+ *
+ * $HEADER$
+ *
+ */
+
+#define _GNU_SOURCE
+#include <stdbool.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <unistd.h>
+#include <time.h>
+#include <signal.h>
+
+#include <pmix.h>
+
+static pmix_proc_t myproc;
+
+/* this is the event notification function we pass down below
+ * when registering for general events - i.e.,, the default
+ * handler. We don't technically need to register one, but it
+ * is usually good practice to catch any events that occur */
+static void notification_fn(size_t evhdlr_registration_id,
+ pmix_status_t status,
+ const pmix_proc_t *source,
+ pmix_info_t info[], size_t ninfo,
+ pmix_info_t results[], size_t nresults,
+ pmix_event_notification_cbfunc_fn_t cbfunc,
+ void *cbdata)
+{
+ if (NULL != cbfunc) {
+ cbfunc(PMIX_EVENT_ACTION_COMPLETE, NULL, 0, NULL, NULL, cbdata);
+ }
+}
+
+/* event handler registration is done asynchronously because it
+ * may involve the PMIx server registering with the host RM for
+ * external events. So we provide a callback function that returns
+ * the status of the request (success or an error), plus a numerical index
+ * to the registered event. The index is used later on to deregister
+ * an event handler - if we don't explicitly deregister it, then the
+ * PMIx server will do so when it see us exit */
+static void evhandler_reg_callbk(pmix_status_t status,
+ size_t evhandler_ref,
+ void *cbdata)
+{
+ volatile int *active = (volatile int*)cbdata;
+
+ if (PMIX_SUCCESS != status) {
+ fprintf(stderr, "Client %s:%d EVENT HANDLER REGISTRATION FAILED WITH STATUS %d, ref=%lu\n",
+ myproc.nspace, myproc.rank, status, (unsigned long)evhandler_ref);
+ }
+ *active = status;
+}
+
+static void infocbfunc(pmix_status_t status,
+ pmix_info_t *info, size_t ninfo,
+ void *cbdata,
+ pmix_release_cbfunc_t release_fn,
+ void *release_cbdata)
+{
+ volatile int *active = (volatile int*)cbdata;
+
+ /* release the caller */
+ if (NULL != release_fn) {
+ release_fn(release_cbdata);
+ }
+
+ *active = status;
+}
+
+int main(int argc, char **argv)
+{
+ int rc;
+ pmix_value_t value;
+ pmix_value_t *val = &value;
+ pmix_proc_t proc;
+ uint32_t nprocs, n;
+ pmix_info_t *info, *iptr;
+ bool flag;
+ volatile int active;
+ pmix_data_array_t *dptr;
+
+ /* init us - note that the call to "init" includes the return of
+ * any job-related info provided by the RM. */
+ if (PMIX_SUCCESS != (rc = PMIx_Init(&myproc, NULL, 0))) {
+ fprintf(stderr, "Client ns %s rank %d: PMIx_Init failed: %d\n", myproc.nspace, myproc.rank, rc);
+ exit(0);
+ }
+ fprintf(stderr, "Client ns %s rank %d: Running\n", myproc.nspace, myproc.rank);
+
+
+ /* register our default event handler - again, this isn't strictly
+ * required, but is generally good practice */
+ active = -1;
+ PMIx_Register_event_handler(NULL, 0, NULL, 0,
+ notification_fn, evhandler_reg_callbk, (void*)&active);
+ while (-1 == active) {
+ sleep(1);
+ }
+ if (0 != active) {
+ fprintf(stderr, "[%s:%d] Default handler registration failed\n", myproc.nspace, myproc.rank);
+ exit(active);
+ }
+
+ /* job-related info is found in our nspace, assigned to the
+ * wildcard rank as it doesn't relate to a specific rank. Setup
+ * a name to retrieve such values */
+ PMIX_PROC_CONSTRUCT(&proc);
+ (void)strncpy(proc.nspace, myproc.nspace, PMIX_MAX_NSLEN);
+ proc.rank = PMIX_RANK_WILDCARD;
+
+ /* get our universe size */
+ if (PMIX_SUCCESS != (rc = PMIx_Get(&proc, PMIX_UNIV_SIZE, NULL, 0, &val))) {
+ fprintf(stderr, "Client ns %s rank %d: PMIx_Get universe size failed: %d\n", myproc.nspace, myproc.rank, rc);
+ goto done;
+ }
+ nprocs = val->data.uint32;
+ PMIX_VALUE_RELEASE(val);
+ fprintf(stderr, "Client %s:%d universe size %d\n", myproc.nspace, myproc.rank, nprocs);
+
+ /* inform the RM that we are preemptible, and that our checkpoint methods are
+ * "signal" on SIGUSR2 and event on PMIX_JCTRL_CHECKPOINT */
+ PMIX_INFO_CREATE(info, 2);
+ flag = true;
+ PMIX_INFO_LOAD(&info[0], PMIX_JOB_CTRL_PREEMPTIBLE, (void*)&flag, PMIX_BOOL);
+ /* can't use "load" to load a pmix_data_array_t */
+ (void)strncpy(info[1].key, PMIX_JOB_CTRL_CHECKPOINT_METHOD, PMIX_MAX_KEYLEN);
+ info[1].value.type = PMIX_DATA_ARRAY;
+ dptr = (pmix_data_array_t*)malloc(sizeof(pmix_data_array_t));
+ info[1].value.data.darray = dptr;
+ dptr->type = PMIX_INFO;
+ dptr->size = 2;
+ PMIX_INFO_CREATE(dptr->array, dptr->size);
+ rc = SIGUSR2;
+ iptr = (pmix_info_t*)dptr->array;
+ PMIX_INFO_LOAD(&iptr[0], PMIX_JOB_CTRL_CHECKPOINT_SIGNAL, &rc, PMIX_INT);
+ rc = PMIX_JCTRL_CHECKPOINT;
+ PMIX_INFO_LOAD(&iptr[1], PMIX_JOB_CTRL_CHECKPOINT_EVENT, &rc, PMIX_STATUS);
+
+ /* since this is informational and not a requested operation, the target parameter
+ * doesn't mean anything and can be ignored */
+ active = -1;
+ if (PMIX_SUCCESS != (rc = PMIx_Job_control_nb(NULL, 0, info, 2, infocbfunc, (void*)&active))) {
+ fprintf(stderr, "Client ns %s rank %d: PMIx_Job_control_nb failed: %d\n", myproc.nspace, myproc.rank, rc);
+ goto done;
+ }
+ while (-1 == active) {
+ sleep(1);
+ }
+ PMIX_INFO_FREE(info, 2);
+ if (0 != active) {
+ fprintf(stderr, "Client ns %s rank %d: PMIx_Job_control_nb failed: %d\n", myproc.nspace, myproc.rank, rc);
+ exit(active);
+ }
+
+ /* now request that this process be monitored using heartbeats */
+ PMIX_INFO_CREATE(iptr, 1);
+ PMIX_INFO_LOAD(&iptr[0], PMIX_MONITOR_HEARTBEAT, NULL, PMIX_POINTER);
+
+ PMIX_INFO_CREATE(info, 3);
+ PMIX_INFO_LOAD(&info[0], PMIX_MONITOR_ID, "MONITOR1", PMIX_STRING);
+ n = 5; // require a heartbeat every 5 seconds
+ PMIX_INFO_LOAD(&info[1], PMIX_MONITOR_HEARTBEAT_TIME, &n, PMIX_UINT32);
+ n = 2; // two heartbeats can be missed before declaring us "stalled"
+ PMIX_INFO_LOAD(&info[2], PMIX_MONITOR_HEARTBEAT_DROPS, &n, PMIX_UINT32);
+
+ /* make the request */
+ active = -1;
+ if (PMIX_SUCCESS != (rc = PMIx_Process_monitor_nb(iptr, PMIX_MONITOR_HEARTBEAT_ALERT,
+ info, 3, infocbfunc, (void*)&active))) {
+ fprintf(stderr, "Client ns %s rank %d: PMIx_Process_monitor_nb failed: %d\n", myproc.nspace, myproc.rank, rc);
+ goto done;
+ }
+ while (-1 == active) {
+ sleep(1);
+ }
+ PMIX_INFO_FREE(iptr, 1);
+ PMIX_INFO_FREE(info, 3);
+ if (0 != active) {
+ fprintf(stderr, "Client ns %s rank %d: PMIx_Process_monitor_nb failed: %d\n", myproc.nspace, myproc.rank, rc);
+ exit(active);
+ }
+
+ /* send a heartbeat */
+ PMIx_Heartbeat();
+
+ /* call fence to synchronize with our peers - no need to
+ * collect any info as we didn't "put" anything */
+ PMIX_INFO_CREATE(info, 1);
+ flag = false;
+ PMIX_INFO_LOAD(info, PMIX_COLLECT_DATA, &flag, PMIX_BOOL);
+ if (PMIX_SUCCESS != (rc = PMIx_Fence(&proc, 1, info, 1))) {
+ fprintf(stderr, "Client ns %s rank %d: PMIx_Fence failed: %d\n", myproc.nspace, myproc.rank, rc);
+ goto done;
+ }
+ PMIX_INFO_FREE(info, 1);
+
+
+ done:
+ /* finalize us */
+ fprintf(stderr, "Client ns %s rank %d: Finalizing\n", myproc.nspace, myproc.rank);
+ if (PMIX_SUCCESS != (rc = PMIx_Finalize(NULL, 0))) {
+ fprintf(stderr, "Client ns %s rank %d:PMIx_Finalize failed: %d\n", myproc.nspace, myproc.rank, rc);
+ } else {
+ fprintf(stderr, "Client ns %s rank %d:PMIx_Finalize successfully completed\n", myproc.nspace, myproc.rank);
+ }
+ fflush(stderr);
+ return(0);
+}

opal/mca/pmix/pmix2x/pmix/include/pmix.h

@@ -523,8 +523,14 @@ pmix_status_t PMIx_Process_monitor_nb(const pmix_info_t *monitor, pmix_status_t
  pmix_info_cbfunc_t cbfunc, void *cbdata);
 
 /* define a special macro to simplify sending of a heartbeat */
-#define PMIx_Heartbeat() \
- PMIx_Process_monitor_nb(PMIX_SEND_HEARTBEAT, NULL, 0, NULL, NULL)
+#define PMIx_Heartbeat() \
+ do { \
+ pmix_info_t _in; \
+ PMIX_INFO_CONSTRUCT(&_in); \
+ PMIX_INFO_LOAD(&_in, PMIX_SEND_HEARTBEAT, NULL, PMIX_POINTER); \
+ PMIx_Process_monitor_nb(&_in, PMIX_SUCCESS, NULL, 0, NULL, NULL); \
+ PMIX_INFO_DESTRUCT(&_in); \
+ } while(0)
 
 #if defined(c_plusplus) || defined(__cplusplus)
 }