int
opal_installdirs_base_open(void)
{
    int i, ret;
    mca_base_component_list_item_t *cli;

    OBJ_CONSTRUCT(&opal_installdirs_components, opal_list_t);
    for (i = 0 ; mca_installdirs_base_static_components[i] != NULL ; ++i) {
        opal_installdirs_base_component_t *component = 
            (opal_installdirs_base_component_t*) 
            mca_installdirs_base_static_components[i];

        /* Save it in a global list for ompi_info */
        cli = OBJ_NEW(mca_base_component_list_item_t);
        cli->cli_component = mca_installdirs_base_static_components[i];
        opal_list_append(&opal_installdirs_components, 
                         &cli->super);

        if (NULL != component->component.mca_open_component) {
            ret = component->component.mca_open_component();
            if (OPAL_SUCCESS != ret) continue;
        }

        /* copy over the data, if something isn't already there */
        CONDITIONAL_COPY(opal_install_dirs, component->install_dirs_data, 
                         prefix);
        CONDITIONAL_COPY(opal_install_dirs, component->install_dirs_data,
                         exec_prefix);
        CONDITIONAL_COPY(opal_install_dirs, component->install_dirs_data,
                         bindir);
        CONDITIONAL_COPY(opal_install_dirs, component->install_dirs_data,
                         sbindir);
        CONDITIONAL_COPY(opal_install_dirs, component->install_dirs_data,
                         libexecdir);
        CONDITIONAL_COPY(opal_install_dirs, component->install_dirs_data,
                         datarootdir);
        CONDITIONAL_COPY(opal_install_dirs, component->install_dirs_data,
                         datadir);
        CONDITIONAL_COPY(opal_install_dirs, component->install_dirs_data, 
                         sysconfdir);
        CONDITIONAL_COPY(opal_install_dirs, component->install_dirs_data, 
                         sharedstatedir);
        CONDITIONAL_COPY(opal_install_dirs, component->install_dirs_data, 
                         localstatedir);
        CONDITIONAL_COPY(opal_install_dirs, component->install_dirs_data, 
                         libdir);
        CONDITIONAL_COPY(opal_install_dirs, component->install_dirs_data, 
                         includedir);
        CONDITIONAL_COPY(opal_install_dirs, component->install_dirs_data, 
                         infodir);
        CONDITIONAL_COPY(opal_install_dirs, component->install_dirs_data, 
                         mandir);
        CONDITIONAL_COPY(opal_install_dirs, component->install_dirs_data,
                         pkgdatadir);
        CONDITIONAL_COPY(opal_install_dirs, component->install_dirs_data, 
                         pkglibdir);
        CONDITIONAL_COPY(opal_install_dirs, component->install_dirs_data, 
                         pkgincludedir);
    }

    /* expand out all the fields */
    opal_install_dirs.prefix = 
        opal_install_dirs_expand(opal_install_dirs.prefix);
    opal_install_dirs.exec_prefix = 
        opal_install_dirs_expand(opal_install_dirs.exec_prefix);
    opal_install_dirs.bindir = 
        opal_install_dirs_expand(opal_install_dirs.bindir);
    opal_install_dirs.sbindir = 
        opal_install_dirs_expand(opal_install_dirs.sbindir);
    opal_install_dirs.libexecdir = 
        opal_install_dirs_expand(opal_install_dirs.libexecdir);
    opal_install_dirs.datarootdir = 
        opal_install_dirs_expand(opal_install_dirs.datarootdir);
    opal_install_dirs.datadir = 
        opal_install_dirs_expand(opal_install_dirs.datadir);
    opal_install_dirs.sysconfdir = 
        opal_install_dirs_expand(opal_install_dirs.sysconfdir);
    opal_install_dirs.sharedstatedir = 
        opal_install_dirs_expand(opal_install_dirs.sharedstatedir);
    opal_install_dirs.localstatedir = 
        opal_install_dirs_expand(opal_install_dirs.localstatedir);
    opal_install_dirs.libdir = 
        opal_install_dirs_expand(opal_install_dirs.libdir);
    opal_install_dirs.includedir = 
        opal_install_dirs_expand(opal_install_dirs.includedir);
    opal_install_dirs.infodir = 
        opal_install_dirs_expand(opal_install_dirs.infodir);
    opal_install_dirs.mandir = 
        opal_install_dirs_expand(opal_install_dirs.mandir);
    opal_install_dirs.pkgdatadir = 
        opal_install_dirs_expand(opal_install_dirs.pkgdatadir);
    opal_install_dirs.pkglibdir = 
        opal_install_dirs_expand(opal_install_dirs.pkglibdir);
    opal_install_dirs.pkgincludedir = 
        opal_install_dirs_expand(opal_install_dirs.pkgincludedir);

#if 0
    fprintf(stderr, "prefix:         %s\n", opal_install_dirs.prefix);
    fprintf(stderr, "exec_prefix:    %s\n", opal_install_dirs.exec_prefix);
    fprintf(stderr, "bindir:         %s\n", opal_install_dirs.bindir);
//.........这里部分代码省略.........

//.........这里部分代码省略.........

    item  = opal_list_get_first(&mca_btl_base_components_opened);
    while(item != opal_list_get_end(&mca_btl_base_components_opened)) {
        opal_list_item_t *next = opal_list_get_next(item);
        cli = (mca_base_component_list_item_t *) item;

        component = (mca_btl_base_component_t *) cli->cli_component;

        /* if there is an include list - item must be in the list to be included */
        if ( NULL != include ) {
            char** argv = include; 
            bool found = false;
            while(argv && *argv) {
                if(strcmp(component->btl_version.mca_component_name,*argv) == 0) {
                    found = true;
                    break;
                }
                argv++;
            }
            if(found == false) {
                item = next;
                continue;
            }

            /* otherwise - check the exclude list to see if this item has been specifically excluded */
        } else if ( NULL != exclude ) {
            char** argv = exclude; 
            bool found = false;
            while(argv && *argv) {
                if(strcmp(component->btl_version.mca_component_name,*argv) == 0) {
                    found = true;
                    break;
                }
                argv++;
            }
            if(found == true) {
                item = next;
                continue;
            }
        }

        opal_output_verbose(10, mca_btl_base_output, 
                            "select: initializing %s component %s",
                            component->btl_version.mca_type_name,
                            component->btl_version.mca_component_name);
        if (NULL == component->btl_init) {
            opal_output_verbose(10, mca_btl_base_output,
                                "select: no init function; ignoring component %s",
                                component->btl_version.mca_component_name);
        } else {
            modules = component->btl_init(&num_btls, enable_progress_threads,
                                          enable_mpi_threads);

            /* If the component didn't initialize, remove it from the opened
               list and remove it from the component repository */

            if (NULL == modules) {
                opal_output_verbose(10, mca_btl_base_output,
                                    "select: init of component %s returned failure",
                                    component->btl_version.mca_component_name);
                opal_output_verbose(10, mca_btl_base_output,
                                    "select: module %s unloaded",
                                    component->btl_version.mca_component_name);

                mca_base_component_repository_release((mca_base_component_t *) component);
                opal_list_remove_item(&mca_btl_base_components_opened, item);
            } 

            /* Otherwise, it initialized properly.  Save it. */

            else {
                opal_output_verbose(10, mca_btl_base_output,
                                    "select: init of component %s returned success",
                                    component->btl_version.mca_component_name);

                for (i = 0; i < num_btls; ++i) {
                    sm = OBJ_NEW(mca_btl_base_selected_module_t);
                    if (NULL == sm) {
                        return OMPI_ERR_OUT_OF_RESOURCE;
                    }
                    sm->btl_component = component;
                    sm->btl_module = modules[i];
                    opal_list_append(&mca_btl_base_modules_initialized,
                                     (opal_list_item_t*) sm);
                }
                free(modules);
            }
        }
        item = next;
    }

    /* Finished querying all components.  Check for the bozo case. */

    if (0 == opal_list_get_size(&mca_btl_base_modules_initialized)) {
        orte_show_help("help-mca-base.txt", "find-available:none-found", true,
                       "btl");
        orte_errmgr.abort(1, NULL);
    }
    return OMPI_SUCCESS;
}

static int show_help(const char *filename, const char *topic,
                     const char *output, orte_process_name_t *sender)
{
    int rc;
    tuple_list_item_t *tli = NULL;
    orte_namelist_t *pnli;
    time_t now = time(NULL);

    /* If we're aggregating, check for duplicates.  Otherwise, don't
       track duplicates at all and always display the message. */
    if (orte_help_want_aggregate) {
        rc = get_tli(filename, topic, &tli);
    } else {
        rc = ORTE_ERR_NOT_FOUND;
    }

    /* If there's no output string (i.e., this is a control message
       asking us to suppress), then skip to the end. */
    if (NULL == output) {
        tli->tli_display = false;
        goto after_output;
    }

    /* Was it already displayed? */
    if (ORTE_SUCCESS == rc) {
        /* Yes.  But do we want to print anything?  That's complicated.

           We always show the first message of a given (filename,
           topic) tuple as soon as it arrives.  But we don't want to
           show duplicate notices often, because we could get overrun
           with them.  So we want to gather them up and say "We got N
           duplicates" every once in a while.

           And keep in mind that at termination, we'll unconditionally
           show all accumulated duplicate notices.

           A simple scheme is as follows:
           - when the first of a (filename, topic) tuple arrives
             - print the message
             - if a timer is not set, set T=now
           - when a duplicate (filename, topic) tuple arrives
             - if now>(T+5) and timer is not set (due to
               non-pre-emptiveness of our libevent, a timer *could* be
               set!)
               - print all accumulated duplicates
               - reset T=now
             - else if a timer was not set, set the timer for T+5
             - else if a timer was set, do nothing (just wait)
           - set T=now when the timer expires
        */           
        ++tli->tli_count_since_last_display;
        if (now > show_help_time_last_displayed + 5 && !show_help_timer_set) {
            show_accumulated_duplicates(0, 0, NULL);
        } else if (!show_help_timer_set) {
            opal_event_evtimer_set(orte_event_base, &show_help_timer_event,
                                   show_accumulated_duplicates, NULL);
            opal_event_evtimer_add(&show_help_timer_event, &show_help_interval);
            show_help_timer_set = true;
        }
    } 
    /* Not already displayed */
    else if (ORTE_ERR_NOT_FOUND == rc) {
        if (orte_xml_output) {
            char *tmp;
            tmp = xml_format((unsigned char*)output);
            fprintf(orte_xml_fp, "%s", tmp);
            fflush(orte_xml_fp);
            free(tmp);
        } else {
            opal_output(orte_clean_output, "%s", output);
        }
        if (!show_help_timer_set) {
            show_help_time_last_displayed = now;
        }
    }
    /* Some other error occurred */
    else {
        ORTE_ERROR_LOG(rc);
        return rc;
    }

 after_output:
    /* If we're aggregating, add this process name to the list */
    if (orte_help_want_aggregate) {
        pnli = OBJ_NEW(orte_namelist_t);
        if (NULL == pnli) {
            rc = ORTE_ERR_OUT_OF_RESOURCE;
            ORTE_ERROR_LOG(rc);
            return rc;
        }
        pnli->name = *sender;
        opal_list_append(&(tli->tli_processes), &(pnli->super));
    }
    return ORTE_SUCCESS;
}

static void mca_btl_tcp_event_construct(mca_btl_tcp_event_t* event)
{
    OPAL_THREAD_LOCK(&mca_btl_tcp_component.tcp_lock);
    opal_list_append(&mca_btl_tcp_component.tcp_events, &event->item);
    OPAL_THREAD_UNLOCK(&mca_btl_tcp_component.tcp_lock);
}

static int store(const opal_identifier_t *uid,
                 opal_scope_t scope,
                 const char *key, const void *data,
                 opal_data_type_t type)
{
    proc_data_t *proc_data;
    opal_value_t *kv;
    opal_byte_object_t *boptr;
    opal_identifier_t id;

    /* data must have an assigned scope */
    if (OPAL_SCOPE_UNDEF == scope) {
        return OPAL_ERR_BAD_PARAM;
    }

    /* to protect alignment, copy the data across */
    memcpy(&id, uid, sizeof(opal_identifier_t));

    /* we are at the bottom of the store priorities, so
     * if this fell to us, we store it
     */
    opal_output_verbose(1, opal_db_base_framework.framework_output,
                        "db:hash:store storing data for proc %" PRIu64 " for scope %d",
                        id, (int)scope);

    /* lookup the proc data object for this proc */
    if (NULL == (proc_data = lookup_opal_proc(&hash_data, id))) {
        /* unrecoverable error */
        OPAL_OUTPUT_VERBOSE((5, opal_db_base_framework.framework_output,
                             "db:hash:store: storing key %s[%s] for proc %" PRIu64 " unrecoverably failed",
                             key, opal_dss.lookup_data_type(type), id));
        return OPAL_ERR_OUT_OF_RESOURCE;
    }

    /* see if we already have this key in the data - means we are updating
     * a pre-existing value
     */
    kv = lookup_keyval(proc_data, key);
    OPAL_OUTPUT_VERBOSE((5, opal_db_base_framework.framework_output,
                         "db:hash:store: %s key %s[%s] for proc %" PRIu64 "",
                         (NULL == kv ? "storing" : "updating"),
                         key, opal_dss.lookup_data_type(type), id));

    if (NULL != kv) {
        opal_list_remove_item(&proc_data->data, &kv->super);
        OBJ_RELEASE(kv);
    }
    kv = OBJ_NEW(opal_value_t);
    kv->key = strdup(key);
    kv->scope = scope;
    opal_list_append(&proc_data->data, &kv->super);

    /* the type could come in as an OPAL one (e.g., OPAL_VPID). Since
     * the value is an OPAL definition, it cannot cover OPAL data
     * types, so convert to the underlying OPAL type
     */
    switch (type) {
    case OPAL_STRING:
        kv->type = OPAL_STRING;
        if (NULL != data) {
            kv->data.string = strdup( (const char *) data);
        } else {
            kv->data.string = NULL;
        }
        break;
    case OPAL_UINT64:
        if (NULL == data) {
            OPAL_ERROR_LOG(OPAL_ERR_BAD_PARAM);
            return OPAL_ERR_BAD_PARAM;
        }
        kv->type = OPAL_UINT64;
        kv->data.uint64 = *(uint64_t*)(data);
        break;
    case OPAL_UINT32:
        if (NULL == data) {
            OPAL_ERROR_LOG(OPAL_ERR_BAD_PARAM);
            return OPAL_ERR_BAD_PARAM;
        }
        kv->type = OPAL_UINT32;
        kv->data.uint32 = *(uint32_t*)data;
        break;
    case OPAL_UINT16:
        if (NULL == data) {
            OPAL_ERROR_LOG(OPAL_ERR_BAD_PARAM);
            return OPAL_ERR_BAD_PARAM;
        }
        kv->type = OPAL_UINT16;
        kv->data.uint16 = *(uint16_t*)(data);
        break;
    case OPAL_INT:
        if (NULL == data) {
            OPAL_ERROR_LOG(OPAL_ERR_BAD_PARAM);
            return OPAL_ERR_BAD_PARAM;
        }
        kv->type = OPAL_INT;
        kv->data.integer = *(int*)(data);
        break;
    case OPAL_UINT:
        if (NULL == data) {
            OPAL_ERROR_LOG(OPAL_ERR_BAD_PARAM);
//.........这里部分代码省略.........

//.........这里部分代码省略.........
    orte_node_t *node;

    /* Accumulate all of the host name mappings */
    for (j = 0; j < opal_argv_count(host_argv); ++j) {
        mini_map = opal_argv_split(host_argv[j], ',');
        
        if (mapped_nodes == NULL) {
            mapped_nodes = mini_map;
        } else {
            for (k = 0; NULL != mini_map[k]; ++k) {
                rc = opal_argv_append_nosize(&mapped_nodes, 
                                             mini_map[k]);
                if (OPAL_SUCCESS != rc) {
                    goto cleanup;
                }
            }
            opal_argv_free(mini_map);
        }
    }

    /* Did we find anything? If not, then do nothing */
    if (NULL == mapped_nodes) {
        return ORTE_SUCCESS;
    }
    
    /*  go through the names found and
       add them to the host list. If they're not unique, then
       bump the slots count for each duplicate */
    
    for (i = 0; NULL != mapped_nodes[i]; ++i) {
        /* if the specified node contains a relative node syntax,
         * this is an error
         */
        if ('+' == mapped_nodes[i][0]) {
            orte_show_help("help-dash-host.txt", "dash-host:relative-syntax",
                           true, mapped_nodes[i]);
            rc = ORTE_ERR_SILENT;
            goto cleanup;
        }
        
        /* see if the node is already on the list */
        for (item = opal_list_get_first(nodes); 
             item != opal_list_get_end(nodes);
             item = opal_list_get_next(item)) {
            node = (orte_node_t*) item;
            if (0 == strcmp(node->name, mapped_nodes[i]) ||
               (0 == strcmp(node->name, orte_process_info.nodename) &&
               (0 == strcmp(mapped_nodes[i], "localhost") || opal_ifislocal(mapped_nodes[i])))) {
                ++node->slots;
                break;
            }
        }
        
        /* If we didn't find it, add it to the list */
        
        if (item == opal_list_get_end(nodes)) {
            node = OBJ_NEW(orte_node_t);
            if (NULL == node) {
                return ORTE_ERR_OUT_OF_RESOURCE;
            }
            /* check to see if this is a local name */
            if (0 == strcmp(mapped_nodes[i], "localhost") ||
                opal_ifislocal(mapped_nodes[i])) {
                /* it is local, so use the local nodename to avoid
                 * later confusion
                 */
                if (orte_show_resolved_nodenames &&
                    0 != strcmp(mapped_nodes[i], orte_process_info.nodename)) {
                    /* add to list of aliases for this node - only add if unique */
                    opal_argv_append_unique_nosize(&node->alias, mapped_nodes[i]);
                }
                node->name = strdup(orte_process_info.nodename);
            } else {
                /* not local - use the given name */
                node->name = strdup(mapped_nodes[i]);
            }
            node->state = ORTE_NODE_STATE_UP;
            node->slots_inuse = 0;
            node->slots_max = 0;
            node->slots = 1;
            /* indicate that ORTE should override any oversubscribed conditions
             * based on local hardware limits since the user (a) might not have
             * provided us any info on the #slots for a node, and (b) the user
             * might have been wrong! If we don't check the number of local physical
             * processors, then we could be too aggressive on our sched_yield setting
             * and cause performance problems.
             */
            *override_oversubscribed = true;
            opal_list_append(nodes, &node->super);
        }
    }
    rc = ORTE_SUCCESS;

cleanup:
    if (NULL != mapped_nodes) {
        opal_argv_free(mapped_nodes);
    }

    return rc;
}

//.........这里部分代码省略.........
            asprintf( &hostname, "%d", apSlots[ix].nid );
            if (NULL == hostname) {
                ORTE_ERROR_LOG(ORTE_ERR_OUT_OF_RESOURCE);
                return ORTE_ERR_OUT_OF_RESOURCE;
            }

            /*          If this matches the prior nodename, just add to the slot count.   */
            if( NULL!=node && !strcmp(node->name, hostname) ) {

                free(hostname);             /* free hostname since not needed */
                ++node->slots;
            } else {                        /* must be new, so add to list    */

                opal_output_verbose(1, orte_ras_base.ras_output,
                                    "ras:alps:read_appinfo: added NID %d to list", apSlots[ix].nid);

                node = OBJ_NEW(orte_node_t);
                node->name = hostname;
                node->launch_id = apSlots[ix].nid;
                node->slots_inuse = 0;
                node->slots_max = 0;
                node->slots = 1;
                /* need to order these node ids so the regex generator
                 * can properly function
                 */
                added = false;
                for (item = opal_list_get_first(nodes);
                     item != opal_list_get_end(nodes);
                     item = opal_list_get_next(item)) {
                    n2 = (orte_node_t*)item;
                    if (node->launch_id < n2->launch_id) {
                        /* insert the new node before this one */
                        opal_list_insert_pos(nodes, item, &node->super);
                        added = true;
                        break;
                    }
                }
                if (!added) {
                    /* add it to the end */
                    opal_list_append(nodes, &node->super);
                }
                sNodes++;                   /* Increment the node count       */
            }
        }
#else
        /* in newer versions of alps, there is one entry for each node in the
         * allocation, and that struct directly carries the number of PEs
         * allocated on that node to this job.
         */
        apNodes=(placeNodeList_t *)(cpBuf+oNow+oInfo+oDet+oSlots);
        oEntry=sizeof(placeNodeList_t)*apInfo->numPlaces;

        oNow+=(oDet+oSlots+oEntry);         /* Target next entry               */

        if( apInfo->resId != *uMe ) continue; /* Filter to our reservation Id */

        for( ix=0; ix<apInfo->numPlaces; ix++ ) {
            opal_output_verbose(5, orte_ras_base.ras_output,
                                "ras:alps:read_appinfo(modern): processing NID %d with %d slots",
                                apNodes[ix].nid, apNodes[ix].numPEs);
            asprintf( &hostname, "%d", apNodes[ix].nid );
            if (NULL == hostname) {
                ORTE_ERROR_LOG(ORTE_ERR_OUT_OF_RESOURCE);
                return ORTE_ERR_OUT_OF_RESOURCE;
            }

            node = OBJ_NEW(orte_node_t);
            node->name = hostname;
            node->launch_id = apNodes[ix].nid;
            node->slots_inuse = 0;
            node->slots_max = 0;
            node->slots = apNodes[ix].numPEs;
            /* need to order these node ids so the regex generator
             * can properly function
             */
            added = false;
            for (item = opal_list_get_first(nodes);
                 item != opal_list_get_end(nodes);
                 item = opal_list_get_next(item)) {
                n2 = (orte_node_t*)item;
                if (node->launch_id < n2->launch_id) {
                    /* insert the new node before this one */
                    opal_list_insert_pos(nodes, item, &node->super);
                    added = true;
                    break;
                }
            }
            if (!added) {
                /* add it to the end */
                opal_list_append(nodes, &node->super);
            }
            sNodes++;                   /* Increment the node count       */
        }
#endif
        break;                              /* Extended details ignored       */
    }
    free(cpBuf);                            /* Free the buffer                */

    return ORTE_SUCCESS;
}

static void 
rml_oob_queued_progress(int fd, short event, void *arg)
{
    orte_rml_oob_queued_msg_t *qmsg;
    orte_rml_oob_msg_header_t *hdr;
    int real_tag;
    int ret;
    orte_process_name_t next, origin;

    while (true) {
        OPAL_THREAD_LOCK(&orte_rml_oob_module.queued_lock);
        qmsg = (orte_rml_oob_queued_msg_t*) opal_list_remove_first(&orte_rml_oob_module.queued_routing_messages);
        OPAL_THREAD_UNLOCK(&orte_rml_oob_module.queued_lock);
        if (NULL == qmsg) break;

        hdr = (orte_rml_oob_msg_header_t*) qmsg->payload[0].iov_base;
        origin = hdr->origin;

        next = orte_routed.get_route(&hdr->destination);
        if (next.vpid == ORTE_VPID_INVALID) {
            opal_output(0,
                        "%s:queued progress tried routing message from %s to %s:%d, can't find route",
                        ORTE_NAME_PRINT(ORTE_PROC_MY_NAME),
                        ORTE_NAME_PRINT(&hdr->origin),
                        ORTE_NAME_PRINT(&hdr->destination),
                        hdr->tag);
            opal_backtrace_print(stderr);
            orte_errmgr.abort(ORTE_ERROR_DEFAULT_EXIT_CODE, NULL);
        }

        if (OPAL_EQUAL == orte_util_compare_name_fields(ORTE_NS_CMP_ALL, &next, ORTE_PROC_MY_NAME)) {
            opal_output(0, "%s:queued progress trying to get message from %s to %s:%d, routing loop",
                        ORTE_NAME_PRINT(ORTE_PROC_MY_NAME),
                        ORTE_NAME_PRINT(&hdr->origin),
                        ORTE_NAME_PRINT(&hdr->destination),
                        hdr->tag);
            opal_backtrace_print(stderr);
            orte_errmgr.abort(ORTE_ERROR_DEFAULT_EXIT_CODE, NULL);
        }

        if (OPAL_EQUAL == orte_util_compare_name_fields(ORTE_NS_CMP_ALL, &next, &hdr->destination)) {
            real_tag = hdr->tag;
        } else {
            real_tag = ORTE_RML_TAG_RML_ROUTE;
        }

        OPAL_OUTPUT_VERBOSE((1, orte_rml_base_output,
                             "%s routing message from %s for %s to %s (tag: %d)",
                             ORTE_NAME_PRINT(ORTE_PROC_MY_NAME),
                             ORTE_NAME_PRINT(&hdr->origin),
                             ORTE_NAME_PRINT(&hdr->destination),
                             ORTE_NAME_PRINT(&next),
                             hdr->tag));

        ORTE_RML_OOB_MSG_HEADER_HTON(*hdr);

        ret = orte_rml_oob_module.active_oob->oob_send_nb(&next,
                                                          &origin,
                                                          qmsg->payload,
                                                          1,
                                                          real_tag,
                                                          0,
                                                          rml_oob_recv_route_queued_send_callback,
                                                          qmsg);

        if (ORTE_SUCCESS != ret) {
            if (ORTE_ERR_ADDRESSEE_UNKNOWN == ret) {
                /* still no route -- try again */
                ORTE_RML_OOB_MSG_HEADER_NTOH(*hdr);
                OPAL_THREAD_LOCK(&orte_rml_oob_module.queued_lock);
                opal_list_append(&orte_rml_oob_module.queued_routing_messages,
                                 &qmsg->super);
                if (1 == opal_list_get_size(&orte_rml_oob_module.queued_routing_messages)) {
                    opal_event_evtimer_add(orte_rml_oob_module.timer_event,
                                           &orte_rml_oob_module.timeout);
                }
                OPAL_THREAD_UNLOCK(&orte_rml_oob_module.queued_lock);
            } else {
                opal_output(0,
                            "%s failed to send message from %s to %s:%d %s (rc = %d)",
                            ORTE_NAME_PRINT(ORTE_PROC_MY_NAME),
                            ORTE_NAME_PRINT(&next),
                            ORTE_NAME_PRINT(&origin),
                            real_tag,
                            ORTE_ERROR_NAME(ret),
                            ret);
                abort();
            }
        }
    }
}

/*
 * If we have a valid section, see if we have a matching section
 * somewhere (i.e., same vendor ID and vendor part ID).  If we do,
 * update the values.  If not, save the values in a new instance and
 * add it to the list.
 */
static int save_section(parsed_section_values_t *s)
{
    int i, j;
    opal_list_item_t *item;
    device_values_t *h;
    bool found;

    /* Is the parsed section valid? */
    if (NULL == s->name || 0 == s->vendor_ids_len ||
            0 == s->vendor_part_ids_len) {
        return OMPI_ERR_BAD_PARAM;
    }

    /* Iterate over each of the vendor/part IDs in the parsed
       values */
    for (i = 0; i < s->vendor_ids_len; ++i) {
        for (j = 0; j < s->vendor_part_ids_len; ++j) {
            found = false;

            /* Iterate over all the saved devices */
            for (item = opal_list_get_first(&devices);
                    item != opal_list_get_end(&devices);
                    item = opal_list_get_next(item)) {
                h = (device_values_t*) item;
                if (s->vendor_ids[i] == h->vendor_id &&
                        s->vendor_part_ids[j] == h->vendor_part_id) {
                    /* Found a match.  Update any newly-set values. */
                    if (s->values.mtu_set) {
                        h->values.mtu = s->values.mtu;
                        h->values.mtu_set = true;
                    }

                    if (s->values.use_eager_rdma_set) {
                        h->values.use_eager_rdma = s->values.use_eager_rdma;
                        h->values.use_eager_rdma_set = true;
                    }

                    if (NULL != s->values.receive_queues) {
                        h->values.receive_queues =
                            strdup(s->values.receive_queues);
                    }

                    if (s->values.max_inline_data_set) {
                        h->values.max_inline_data = s->values.max_inline_data;
                        h->values.max_inline_data_set = true;
                    }

                    if (s->values.rdmacm_reject_causes_connect_error_set) {
                        h->values.rdmacm_reject_causes_connect_error =
                            s->values.rdmacm_reject_causes_connect_error;
                        h->values.rdmacm_reject_causes_connect_error_set =
                            true;
                    }

                    found = true;
                    break;
                }
            }

            /* Did we find/update it in the exising list?  If not,
               create a new one. */
            if (!found) {
                h = OBJ_NEW(device_values_t);
                h->section_name = strdup(s->name);
                h->vendor_id = s->vendor_ids[i];
                h->vendor_part_id = s->vendor_part_ids[j];
                /* NOTE: There is a bug in the PGI 6.2 series that
                   causes the compiler to choke when copying structs
                   containing bool members by value.  So do a memcpy
                   here instead. */
                memcpy(&h->values, &s->values, sizeof(s->values));
                /* Need to strdup the string, though */
                if (NULL != h->values.receive_queues) {
                    h->values.receive_queues = strdup(s->values.receive_queues);
                }
                opal_list_append(&devices, &h->super);
            }
        }
    }

    /* All done */

    return OMPI_SUCCESS;
}

//.........这里部分代码省略.........
        swap_dest.vpid  = errmgr_cmdline_sender.vpid;

        errmgr_cmdline_sender = *sender;
        orte_errmgr_base_migrate_update(ORTE_ERRMGR_MIGRATE_STATE_ERR_INPROGRESS);

        errmgr_cmdline_sender.jobid = swap_dest.jobid;
        errmgr_cmdline_sender.vpid  = swap_dest.vpid;

        return;
    }

    errmgr_cmdline_sender = *sender;

    count = 1;
    if (ORTE_SUCCESS != (ret = opal_dss.unpack(buffer, &command, &count, ORTE_ERRMGR_MIGRATE_TOOL_CMD))) {
        ORTE_ERROR_LOG(ret);
        return;
    }

    /*
     * orte-migrate has requested that a checkpoint be taken
     */
    if (ORTE_ERRMGR_MIGRATE_TOOL_INIT_CMD == command) {
        OPAL_OUTPUT_VERBOSE((10, orte_errmgr_base_framework.framework_output,
                             "errmgr:base:tool:recv() Command line requested process migration [command %d]\n",
                             command));

        /*
         * Unpack the buffer from the orte-migrate command
         */
        count = 1;
        if (ORTE_SUCCESS != (ret = opal_dss.unpack(buffer, &(off_procs), &count, OPAL_STRING))) {
            ORTE_ERROR_LOG(ret);
	    return;
        }

        if (ORTE_SUCCESS != (ret = opal_dss.unpack(buffer, &(off_nodes), &count, OPAL_STRING))) {
            ORTE_ERROR_LOG(ret);
            return;
        }

        if (ORTE_SUCCESS != (ret = opal_dss.unpack(buffer, &(onto_nodes), &count, OPAL_STRING))) {
            ORTE_ERROR_LOG(ret);
            return;
        }

        /*
         * Parse the comma separated list
         */
        proc_list = OBJ_NEW(opal_list_t);
        node_list = OBJ_NEW(opal_list_t);
        suggested_map_list = OBJ_NEW(opal_list_t);

        split_off_procs     = opal_argv_split(off_procs,  ',');
        cnt = opal_argv_count(split_off_procs);
        if( cnt > 0 ) {
            for(i = 0; i < cnt; ++i) {
                off_proc = OBJ_NEW(orte_errmgr_predicted_proc_t);
                off_proc->proc_name.vpid = atoi(split_off_procs[i]);
                opal_list_append(proc_list, &(off_proc->super));
            }
        }

        split_off_nodes     = opal_argv_split(off_nodes,  ',');
        cnt = opal_argv_count(split_off_nodes);
        if( cnt > 0 ) {
            for(i = 0; i < cnt; ++i) {
                off_node = OBJ_NEW(orte_errmgr_predicted_node_t);
                off_node->node_name = strdup(split_off_nodes[i]);
                opal_list_append(node_list, &(off_node->super));
            }
        }

        split_onto_nodes    = opal_argv_split(onto_nodes, ',');
        cnt = opal_argv_count(split_onto_nodes);
        if( cnt > 0 ) {
            for(i = 0; i < cnt; ++i) {
                onto_map = OBJ_NEW(orte_errmgr_predicted_map_t);
                onto_map->map_node_name = strdup(split_onto_nodes[i]);
                opal_list_append(suggested_map_list, &(onto_map->super));
            }
        }

        /*
         * Pass to the predicted fault function to see how they would like to progress
         */
        orte_errmgr.predicted_fault(proc_list, node_list, suggested_map_list);
    }
    /*
     * Unknown command
     */
    else {
        OPAL_OUTPUT_VERBOSE((10, orte_errmgr_base_framework.framework_output,
                             "errmgr:base:tool:recv() Command line sent an unknown command (command %d)\n",
                             command));
        ORTE_ERROR_LOG(ORTE_ERR_NOT_SUPPORTED);
    }

    return;
}

/* configure using getifaddrs(3) */
static int if_bsdx_open(void)
{
    struct ifaddrs **ifadd_list;
    struct ifaddrs *cur_ifaddrs;
    struct sockaddr_in* sin_addr;

    /* 
     * the manpage claims that getifaddrs() allocates the memory,
     * and freeifaddrs() is later used to release the allocated memory.
     * however, without this malloc the call to getifaddrs() segfaults
     */
    ifadd_list = (struct ifaddrs **) malloc(sizeof(struct ifaddrs*));

    /* create the linked list of ifaddrs structs */
    if (getifaddrs(ifadd_list) < 0) {
        opal_output(0, "opal_ifinit: getifaddrs() failed with error=%d\n",
                    errno);
        return OPAL_ERROR;
    }

    for (cur_ifaddrs = *ifadd_list; NULL != cur_ifaddrs; 
         cur_ifaddrs = cur_ifaddrs->ifa_next) {
        opal_if_t *intf;
        struct in_addr a4;

        /* skip non- af_inet interface addresses */
        if (AF_INET != cur_ifaddrs->ifa_addr->sa_family) {
            continue;
        }

        /* skip interface if it is down (IFF_UP not set) */
        if (0 == (cur_ifaddrs->ifa_flags & IFF_UP)) {
            continue;
        }

        /* skip interface if it is a loopback device (IFF_LOOPBACK set) */
        if (!opal_if_retain_loopback && 0 != (cur_ifaddrs->ifa_flags & IFF_LOOPBACK)) {
            continue;
        }

        /* or if it is a point-to-point interface */
        /* TODO: do we really skip p2p? */
        if (0 != (cur_ifaddrs->ifa_flags & IFF_POINTOPOINT)) {
            continue;
        }

        sin_addr = (struct sockaddr_in *) cur_ifaddrs->ifa_addr;

        intf = OBJ_NEW(opal_if_t);
        if (NULL == intf) {
            opal_output(0, "opal_ifinit: unable to allocate %d bytes\n",
                        (int) sizeof(opal_if_t));
            return OPAL_ERR_OUT_OF_RESOURCE;
        }

        /* fill values into the opal_if_t */
        memcpy(&a4, &(sin_addr->sin_addr), sizeof(struct in_addr));
            
        strncpy(intf->if_name, cur_ifaddrs->ifa_name, IF_NAMESIZE);
        intf->if_index = opal_list_get_size(&opal_if_list) + 1;
        ((struct sockaddr_in*) &intf->if_addr)->sin_addr = a4;
        ((struct sockaddr_in*) &intf->if_addr)->sin_family = AF_INET;
        ((struct sockaddr_in*) &intf->if_addr)->sin_len =  cur_ifaddrs->ifa_addr->sa_len;

        intf->if_mask = prefix( sin_addr->sin_addr.s_addr);
        intf->if_flags = cur_ifaddrs->ifa_flags;

        intf->if_kernel_index = 
            (uint16_t) if_nametoindex(cur_ifaddrs->ifa_name);

        opal_list_append(&opal_if_list, &(intf->super));
    }   /*  of for loop over ifaddrs list */

    return OPAL_SUCCESS;
}

static int orte_rds_hostfile_query(orte_jobid_t job)
{
    opal_list_t existing;
    opal_list_t updates, rds_updates;
    opal_list_item_t *item;
    orte_rds_cell_desc_t *rds_item;
    orte_rds_cell_attr_t *new_attr;
    orte_ras_node_t *ras_item;
    int rc;

    if (orte_rds_hostfile_queried) {
        /* if we have already been queried, then
         * our info is on the registry, so just
         * return. Note that this restriction
         * may eventually be lifted - ideally, 
         * we might check to see if this is a
         * new file name and go ahead with the
         * query if so.
         */
        return ORTE_SUCCESS;
    }
    orte_rds_hostfile_queried = true;
    
    OBJ_CONSTRUCT(&existing, opal_list_t);
    OBJ_CONSTRUCT(&updates, opal_list_t);
    OBJ_CONSTRUCT(&rds_updates, opal_list_t);
    rc = orte_ras_base_node_query(&existing);
    if(ORTE_SUCCESS != rc) {
        goto cleanup;
    }

    rc = mca_base_param_find("rds", "hostfile", "path");
    mca_base_param_lookup_string(rc, &mca_rds_hostfile_component.path);

    rc = orte_rds_hostfile_parse(mca_rds_hostfile_component.path, &existing, &updates);
    if (ORTE_ERR_NOT_FOUND == rc) {
        if(mca_rds_hostfile_component.default_hostfile) {
            rc = ORTE_SUCCESS;
        } else {
            opal_show_help("help-rds-hostfile.txt", "rds:no-hostfile",
                           true,
                           mca_rds_hostfile_component.path);
        }
        goto cleanup;
    } else if (ORTE_SUCCESS != rc) {
        goto cleanup;
    }

    if ( !opal_list_is_empty(&updates) ) {

        /* Convert RAS update list to RDS update list */
        for ( ras_item  = (orte_ras_node_t*)opal_list_get_first(&updates);
              ras_item != (orte_ras_node_t*)opal_list_get_end(&updates);
              ras_item  = (orte_ras_node_t*)opal_list_get_next(ras_item)) {

            rds_item = OBJ_NEW(orte_rds_cell_desc_t);
            if (NULL == rds_item) {
                ORTE_ERROR_LOG(ORTE_ERR_OUT_OF_RESOURCE);
                return ORTE_ERR_OUT_OF_RESOURCE;
            }

            rds_item->site  = strdup("Hostfile");
            rds_item->name  = strdup(ras_item->node_name);
            if (need_cellid) {
#if 0 /* JJH Repair when cellid's are fixed */
                /* Create a new cellid for this hostfile */
                rc = orte_ns.create_cellid(&local_cellid, rds_item->site, rds_item->name);
                if (ORTE_SUCCESS != rc) {
                    ORTE_ERROR_LOG(rc);
                    return rc;
                }
#endif
                local_cellid = 0;
                need_cellid = false;
            }

            rds_item->cellid      = local_cellid;
            ras_item->node_cellid = local_cellid;

            new_attr = OBJ_NEW(orte_rds_cell_attr_t);
            if (NULL == new_attr) {
                ORTE_ERROR_LOG(ORTE_ERR_OUT_OF_RESOURCE);
                return ORTE_ERR_OUT_OF_RESOURCE;
            }
            new_attr->keyval.key          = strdup(ORTE_RDS_NAME);
            new_attr->keyval.value = OBJ_NEW(orte_data_value_t);
            if (NULL == new_attr->keyval.value) {
                ORTE_ERROR_LOG(ORTE_ERR_OUT_OF_RESOURCE);
                return ORTE_ERR_OUT_OF_RESOURCE;
            }
            new_attr->keyval.value->type   = ORTE_STRING;
            new_attr->keyval.value->data   = strdup(ras_item->node_name);
            opal_list_append(&(rds_item->attributes), &new_attr->super);

            new_attr = OBJ_NEW(orte_rds_cell_attr_t);
            if (NULL == new_attr) {
                ORTE_ERROR_LOG(ORTE_ERR_OUT_OF_RESOURCE);
                return ORTE_ERR_OUT_OF_RESOURCE;
            }
            new_attr->keyval.key          = strdup(ORTE_CELLID_KEY);
//.........这里部分代码省略.........

//.........这里部分代码省略.........
               not found, then we have a new node instance that needs
               to be added to the updates list. */

            update = true;
        }
        else {
            /* If it was in the existing list, then we can use its cellid
             * to add the reset of the hosts in the file to. */
            local_cellid = node->node_cellid;
            need_cellid = false;
        }
    } else {
        orte_rds_hostfile_parse_error(token);
        return ORTE_ERROR;
    }

    got_count = false;
    while (!orte_rds_hostfile_done) {
        token = orte_rds_hostfile_lex();

        switch (token) {
        case ORTE_RDS_HOSTFILE_DONE:
            goto done;

        case ORTE_RDS_HOSTFILE_NEWLINE:
            goto done;

        case ORTE_RDS_HOSTFILE_USERNAME:
            node->node_username = orte_rds_hostfile_parse_string();
            break;

        case ORTE_RDS_HOSTFILE_COUNT:
        case ORTE_RDS_HOSTFILE_CPU:
        case ORTE_RDS_HOSTFILE_SLOTS:
            rc = orte_rds_hostfile_parse_int();
            if (rc < 0) {
                opal_show_help("help-rds-hostfile.txt", "rds:slots",
                               true,
                               cur_hostfile_name, rc);
                OBJ_RELEASE(node);
                return ORTE_ERROR;
            }
            node->node_slots += rc;
            update = true;
            got_count = true;

            /* Ensure that node_slots_max >= node_slots */
            if (node->node_slots_max != 0 && node->node_slots_max < node->node_slots) {
                node->node_slots_max = node->node_slots;
            }
            break;

        case ORTE_RDS_HOSTFILE_SLOTS_MAX:
            rc = orte_rds_hostfile_parse_int();
            if (rc < 0) {
                opal_show_help("help-rds-hostfile.txt", "rds:max_slots",
                               true,
                               cur_hostfile_name, ((size_t) rc));
                OBJ_RELEASE(node);
                return ORTE_ERROR;
            }
            /* Only take this update if it puts us >= node_slots */
            if (rc >= node->node_slots) {
                if (node->node_slots_max != rc) {
                    node->node_slots_max = rc;
                    update = true;
                    got_max = true;
                }
            } else {
                opal_show_help("help-rds-hostfile.txt", "rds:max_slots_lt",
                               true,
                               cur_hostfile_name, node->node_slots, rc);