Cleanup the sensor framework close - existing code was using incorrect object type. Don't start sensors if sample rate is zero. Don't add zero-byte data from resusage as it means nothing was measured.

cmr=v1.7.4:reviewer=hjelmn

This commit was SVN r30150.

orte/mca/sensor/base/sensor_base_fns.c

@@ -26,31 +26,32 @@ void orte_sensor_base_start(orte_jobid_t job)
     orte_sensor_active_module_t *i_module;
     int i;
 
-    opal_output_verbose(5, orte_sensor_base_framework.framework_output,
-                        "%s sensor:base: starting sensors",
-                        ORTE_NAME_PRINT(ORTE_PROC_MY_NAME));
-
-    /* call the start function of all modules in priority order */
-    for (i=0; i < orte_sensor_base.modules.size; i++) {
-        if (NULL == (i_module = (orte_sensor_active_module_t*)opal_pointer_array_get_item(&orte_sensor_base.modules, i))) {
-            continue;
+    if (0 < orte_sensor_base.rate.tv_sec) {
+        opal_output_verbose(5, orte_sensor_base_framework.framework_output,
+                            "%s sensor:base: starting sensors",
+                            ORTE_NAME_PRINT(ORTE_PROC_MY_NAME));
+        /* call the start function of all modules in priority order */
+        for (i=0; i < orte_sensor_base.modules.size; i++) {
+            if (NULL == (i_module = (orte_sensor_active_module_t*)opal_pointer_array_get_item(&orte_sensor_base.modules, i))) {
+                continue;
+            }
+            mods_active = true;
+            if (NULL != i_module->module->start) {
+                i_module->module->start(job);
+            }
         }
-        mods_active = true;
-        if (NULL != i_module->module->start) {
-            i_module->module->start(job);
-        }
-    }
 
-    if (mods_active && !orte_sensor_base.active) {
-        /* setup a buffer to collect samples */
-        orte_sensor_base.samples = OBJ_NEW(opal_buffer_t);
-        /* startup a timer to wake us up periodically
-         * for a data sample
-         */
-        orte_sensor_base.active = true;
-        opal_event_evtimer_set(orte_event_base, &orte_sensor_base.sample_ev,
-                               orte_sensor_base_sample, NULL);
-        opal_event_evtimer_add(&orte_sensor_base.sample_ev, &orte_sensor_base.rate);
+        if (mods_active && !orte_sensor_base.active) {
+            /* setup a buffer to collect samples */
+            orte_sensor_base.samples = OBJ_NEW(opal_buffer_t);
+            /* startup a timer to wake us up periodically
+             * for a data sample
+             */
+            orte_sensor_base.active = true;
+            opal_event_evtimer_set(orte_event_base, &orte_sensor_base.sample_ev,
+                                   orte_sensor_base_sample, NULL);
+            opal_event_evtimer_add(&orte_sensor_base.sample_ev, &orte_sensor_base.rate);
+        }
     }
     return;    
 }

orte/mca/sensor/base/sensor_base_frame.c

@@ -79,15 +79,15 @@ static int orte_sensor_base_register(mca_base_register_flag_t flags)
 
 static int orte_sensor_base_close(void)
 {
-    orte_sensor_base_module_t *i_module;
+    orte_sensor_active_module_t *i_module;
     int i;
     
     for (i=0; i < orte_sensor_base.modules.size; i++) {
-        if (NULL == (i_module = (orte_sensor_base_module_t*)opal_pointer_array_get_item(&orte_sensor_base.modules, i))) {
+        if (NULL == (i_module = (orte_sensor_active_module_t*)opal_pointer_array_get_item(&orte_sensor_base.modules, i))) {
             continue;
         }
-        if (NULL != i_module->finalize) {
-            i_module->finalize();
+        if (NULL != i_module->module->finalize) {
+            i_module->module->finalize();
         }
     }
     OBJ_DESTRUCT(&orte_sensor_base.modules);

orte/mca/sensor/resusage/sensor_resusage.c

@@ -70,6 +70,7 @@ static int init(void)
 
 static void finalize(void)
 {
+    return;
 }
 
 static void sample(void)
@@ -150,8 +151,10 @@ static void sample(void)
                 continue;
             }
             stats = OBJ_NEW(opal_pstats_t);
-            if (ORTE_SUCCESS != (rc = opal_pstat.query(child->pid, stats, NULL))) {
-                ORTE_ERROR_LOG(rc);
+            if (ORTE_SUCCESS != opal_pstat.query(child->pid, stats, NULL)) {
+                /* may hit a race condition where the process has
+                 * terminated, so just ignore any error
+                 */
                 OBJ_RELEASE(stats);
                 continue;
             }
@@ -171,12 +174,14 @@ static void sample(void)
         }
     }
 
-    /* xfer the data for transmission */
-    bptr = &buf;
-    if (OPAL_SUCCESS != (rc = opal_dss.pack(orte_sensor_base.samples, &bptr, 1, OPAL_BUFFER))) {
-        ORTE_ERROR_LOG(rc);
-        OBJ_DESTRUCT(&buf);
-        return;
+    /* xfer any data for transmission */
+    if (0 < buf.bytes_used) {
+        bptr = &buf;
+        if (OPAL_SUCCESS != (rc = opal_dss.pack(orte_sensor_base.samples, &bptr, 1, OPAL_BUFFER))) {
+            ORTE_ERROR_LOG(rc);
+            OBJ_DESTRUCT(&buf);
+            return;
+        }
     }
     OBJ_DESTRUCT(&buf);