+ * GPL HEADER END
+ */
+/*
+ * Copyright (c) 2003, 2010, Oracle and/or its affiliates. All rights reserved.
+ * Use is subject to license terms.
+ *
+ * Copyright (c) 2011, 2016, Intel Corporation.
+ */
+/*
+ * This file is part of Lustre, http://www.lustre.org/
+ * Lustre is a trademark of Sun Microsystems, Inc.
+ */
+
+#define DEBUG_SUBSYSTEM S_LNET
+#include <linux/log2.h>
+#include <linux/ktime.h>
+
+#include <lnet/lib-lnet.h>
+
+#define D_LNI D_CONSOLE
+
+lnet_t the_lnet; /* THE state of the network */
+EXPORT_SYMBOL(the_lnet);
+
+static char *ip2nets = "";
+module_param(ip2nets, charp, 0444);
+MODULE_PARM_DESC(ip2nets, "LNET network <- IP table");
+
+static char *networks = "";
+module_param(networks, charp, 0444);
+MODULE_PARM_DESC(networks, "local networks");
+
+static char *routes = "";
+module_param(routes, charp, 0444);
+MODULE_PARM_DESC(routes, "routes to non-local networks");
+
+static int rnet_htable_size = LNET_REMOTE_NETS_HASH_DEFAULT;
+module_param(rnet_htable_size, int, 0444);
+MODULE_PARM_DESC(rnet_htable_size, "size of remote network hash table");
+
+static int use_tcp_bonding = false;
+module_param(use_tcp_bonding, int, 0444);
+MODULE_PARM_DESC(use_tcp_bonding,
+ "Set to 1 to use socklnd bonding. 0 to use Multi-Rail");
+
+/*
+ * This sequence number keeps track of how many times DLC was used to
+ * update the configuration. It is incremented on any DLC update and
+ * checked when sending a message to determine if there is a need to
+ * re-run the selection algorithm to handle configuration change.
+ * Look at lnet_select_pathway() for more details on its usage.
+ */
+static atomic_t lnet_dlc_seq_no = ATOMIC_INIT(0);
+
+static int lnet_ping(lnet_process_id_t id, signed long timeout,
+ lnet_process_id_t __user *ids, int n_ids);
+
+static char *
+lnet_get_routes(void)
+{
+ return routes;
+}
+
+static char *
+lnet_get_networks(void)
+{
+ char *nets;
+ int rc;
+
+ if (*networks != 0 && *ip2nets != 0) {
+ LCONSOLE_ERROR_MSG(0x101, "Please specify EITHER 'networks' or "
+ "'ip2nets' but not both at once\n");
+ return NULL;
+ }
+
+ if (*ip2nets != 0) {
+ rc = lnet_parse_ip2nets(&nets, ip2nets);
+ return (rc == 0) ? nets : NULL;
+ }
+
+ if (*networks != 0)
+ return networks;
+
+ return "tcp";
+}
+
+static void
+lnet_init_locks(void)
+{
+ spin_lock_init(&the_lnet.ln_eq_wait_lock);
+ init_waitqueue_head(&the_lnet.ln_eq_waitq);
+ init_waitqueue_head(&the_lnet.ln_rc_waitq);
+ mutex_init(&the_lnet.ln_lnd_mutex);
+ mutex_init(&the_lnet.ln_api_mutex);
+}
+
+static void
+lnet_fini_locks(void)
+{
+}
+
+struct kmem_cache *lnet_mes_cachep; /* MEs kmem_cache */
+struct kmem_cache *lnet_small_mds_cachep; /* <= LNET_SMALL_MD_SIZE bytes
+ * MDs kmem_cache */
+
+static int
+lnet_descriptor_setup(void)
+{
+ /* create specific kmem_cache for MEs and small MDs (i.e., originally
+ * allocated in <size-xxx> kmem_cache).
+ */
+ lnet_mes_cachep = kmem_cache_create("lnet_MEs", sizeof(lnet_me_t),
+ 0, 0, NULL);
+ if (!lnet_mes_cachep)
+ return -ENOMEM;
+
+ lnet_small_mds_cachep = kmem_cache_create("lnet_small_MDs",
+ LNET_SMALL_MD_SIZE, 0, 0,
+ NULL);
+ if (!lnet_small_mds_cachep)
+ return -ENOMEM;
+
+ return 0;
+}
+
+static void
+lnet_descriptor_cleanup(void)
+{
+
+ if (lnet_small_mds_cachep) {
+ kmem_cache_destroy(lnet_small_mds_cachep);
+ lnet_small_mds_cachep = NULL;
+ }
+
+ if (lnet_mes_cachep) {
+ kmem_cache_destroy(lnet_mes_cachep);
+ lnet_mes_cachep = NULL;
+ }
+}
+
+static int
+lnet_create_remote_nets_table(void)
+{
+ int i;
+ struct list_head *hash;
+
+ LASSERT(the_lnet.ln_remote_nets_hash == NULL);
+ LASSERT(the_lnet.ln_remote_nets_hbits > 0);
+ LIBCFS_ALLOC(hash, LNET_REMOTE_NETS_HASH_SIZE * sizeof(*hash));
+ if (hash == NULL) {
+ CERROR("Failed to create remote nets hash table\n");
+ return -ENOMEM;
+ }
+
+ for (i = 0; i < LNET_REMOTE_NETS_HASH_SIZE; i++)
+ INIT_LIST_HEAD(&hash[i]);
+ the_lnet.ln_remote_nets_hash = hash;
+ return 0;
+}
+
+static void
+lnet_destroy_remote_nets_table(void)
+{
+ int i;
+
+ if (the_lnet.ln_remote_nets_hash == NULL)
+ return;
+
+ for (i = 0; i < LNET_REMOTE_NETS_HASH_SIZE; i++)
+ LASSERT(list_empty(&the_lnet.ln_remote_nets_hash[i]));
+
+ LIBCFS_FREE(the_lnet.ln_remote_nets_hash,
+ LNET_REMOTE_NETS_HASH_SIZE *
+ sizeof(the_lnet.ln_remote_nets_hash[0]));
+ the_lnet.ln_remote_nets_hash = NULL;
+}
+
+static void
+lnet_destroy_locks(void)
+{
+ if (the_lnet.ln_res_lock != NULL) {
+ cfs_percpt_lock_free(the_lnet.ln_res_lock);
+ the_lnet.ln_res_lock = NULL;
+ }
+
+ if (the_lnet.ln_net_lock != NULL) {
+ cfs_percpt_lock_free(the_lnet.ln_net_lock);
+ the_lnet.ln_net_lock = NULL;
+ }
+
+ lnet_fini_locks();
+}
+
+static int
+lnet_create_locks(void)
+{
+ lnet_init_locks();
+
+ the_lnet.ln_res_lock = cfs_percpt_lock_alloc(lnet_cpt_table());
+ if (the_lnet.ln_res_lock == NULL)
+ goto failed;
+
+ the_lnet.ln_net_lock = cfs_percpt_lock_alloc(lnet_cpt_table());
+ if (the_lnet.ln_net_lock == NULL)
+ goto failed;
+
+ return 0;
+
+ failed:
+ lnet_destroy_locks();
+ return -ENOMEM;
+}
+
+static void lnet_assert_wire_constants(void)
+{
+ /* Wire protocol assertions generated by 'wirecheck'
+ * running on Linux robert.bartonsoftware.com 2.6.8-1.521
+ * #1 Mon Aug 16 09:01:18 EDT 2004 i686 athlon i386 GNU/Linux
+ * with gcc version 3.3.3 20040412 (Red Hat Linux 3.3.3-7) */
+
+ /* Constants... */
+ CLASSERT(LNET_PROTO_TCP_MAGIC == 0xeebc0ded);
+ CLASSERT(LNET_PROTO_TCP_VERSION_MAJOR == 1);
+ CLASSERT(LNET_PROTO_TCP_VERSION_MINOR == 0);
+ CLASSERT(LNET_MSG_ACK == 0);
+ CLASSERT(LNET_MSG_PUT == 1);
+ CLASSERT(LNET_MSG_GET == 2);
+ CLASSERT(LNET_MSG_REPLY == 3);
+ CLASSERT(LNET_MSG_HELLO == 4);
+
+ /* Checks for struct lnet_handle_wire */
+ CLASSERT((int)sizeof(struct lnet_handle_wire) == 16);
+ CLASSERT((int)offsetof(struct lnet_handle_wire, wh_interface_cookie) == 0);
+ CLASSERT((int)sizeof(((struct lnet_handle_wire *)0)->wh_interface_cookie) == 8);
+ CLASSERT((int)offsetof(struct lnet_handle_wire, wh_object_cookie) == 8);
+ CLASSERT((int)sizeof(((struct lnet_handle_wire *)0)->wh_object_cookie) == 8);
+
+ /* Checks for struct lnet_magicversion_t */
+ CLASSERT((int)sizeof(lnet_magicversion_t) == 8);
+ CLASSERT((int)offsetof(lnet_magicversion_t, magic) == 0);
+ CLASSERT((int)sizeof(((lnet_magicversion_t *)0)->magic) == 4);
+ CLASSERT((int)offsetof(lnet_magicversion_t, version_major) == 4);
+ CLASSERT((int)sizeof(((lnet_magicversion_t *)0)->version_major) == 2);
+ CLASSERT((int)offsetof(lnet_magicversion_t, version_minor) == 6);
+ CLASSERT((int)sizeof(((lnet_magicversion_t *)0)->version_minor) == 2);
+
+ /* Checks for struct lnet_hdr_t */
+ CLASSERT((int)sizeof(lnet_hdr_t) == 72);
+ CLASSERT((int)offsetof(lnet_hdr_t, dest_nid) == 0);
+ CLASSERT((int)sizeof(((lnet_hdr_t *)0)->dest_nid) == 8);
+ CLASSERT((int)offsetof(lnet_hdr_t, src_nid) == 8);
+ CLASSERT((int)sizeof(((lnet_hdr_t *)0)->src_nid) == 8);
+ CLASSERT((int)offsetof(lnet_hdr_t, dest_pid) == 16);
+ CLASSERT((int)sizeof(((lnet_hdr_t *)0)->dest_pid) == 4);
+ CLASSERT((int)offsetof(lnet_hdr_t, src_pid) == 20);
+ CLASSERT((int)sizeof(((lnet_hdr_t *)0)->src_pid) == 4);
+ CLASSERT((int)offsetof(lnet_hdr_t, type) == 24);
+ CLASSERT((int)sizeof(((lnet_hdr_t *)0)->type) == 4);
+ CLASSERT((int)offsetof(lnet_hdr_t, payload_length) == 28);
+ CLASSERT((int)sizeof(((lnet_hdr_t *)0)->payload_length) == 4);
+ CLASSERT((int)offsetof(lnet_hdr_t, msg) == 32);
+ CLASSERT((int)sizeof(((lnet_hdr_t *)0)->msg) == 40);
+
+ /* Ack */
+ CLASSERT((int)offsetof(lnet_hdr_t, msg.ack.dst_wmd) == 32);
+ CLASSERT((int)sizeof(((lnet_hdr_t *)0)->msg.ack.dst_wmd) == 16);
+ CLASSERT((int)offsetof(lnet_hdr_t, msg.ack.match_bits) == 48);
+ CLASSERT((int)sizeof(((lnet_hdr_t *)0)->msg.ack.match_bits) == 8);
+ CLASSERT((int)offsetof(lnet_hdr_t, msg.ack.mlength) == 56);
+ CLASSERT((int)sizeof(((lnet_hdr_t *)0)->msg.ack.mlength) == 4);
+
+ /* Put */
+ CLASSERT((int)offsetof(lnet_hdr_t, msg.put.ack_wmd) == 32);
+ CLASSERT((int)sizeof(((lnet_hdr_t *)0)->msg.put.ack_wmd) == 16);
+ CLASSERT((int)offsetof(lnet_hdr_t, msg.put.match_bits) == 48);
+ CLASSERT((int)sizeof(((lnet_hdr_t *)0)->msg.put.match_bits) == 8);
+ CLASSERT((int)offsetof(lnet_hdr_t, msg.put.hdr_data) == 56);
+ CLASSERT((int)sizeof(((lnet_hdr_t *)0)->msg.put.hdr_data) == 8);
+ CLASSERT((int)offsetof(lnet_hdr_t, msg.put.ptl_index) == 64);
+ CLASSERT((int)sizeof(((lnet_hdr_t *)0)->msg.put.ptl_index) == 4);
+ CLASSERT((int)offsetof(lnet_hdr_t, msg.put.offset) == 68);
+ CLASSERT((int)sizeof(((lnet_hdr_t *)0)->msg.put.offset) == 4);
+
+ /* Get */
+ CLASSERT((int)offsetof(lnet_hdr_t, msg.get.return_wmd) == 32);
+ CLASSERT((int)sizeof(((lnet_hdr_t *)0)->msg.get.return_wmd) == 16);
+ CLASSERT((int)offsetof(lnet_hdr_t, msg.get.match_bits) == 48);
+ CLASSERT((int)sizeof(((lnet_hdr_t *)0)->msg.get.match_bits) == 8);
+ CLASSERT((int)offsetof(lnet_hdr_t, msg.get.ptl_index) == 56);
+ CLASSERT((int)sizeof(((lnet_hdr_t *)0)->msg.get.ptl_index) == 4);
+ CLASSERT((int)offsetof(lnet_hdr_t, msg.get.src_offset) == 60);
+ CLASSERT((int)sizeof(((lnet_hdr_t *)0)->msg.get.src_offset) == 4);
+ CLASSERT((int)offsetof(lnet_hdr_t, msg.get.sink_length) == 64);
+ CLASSERT((int)sizeof(((lnet_hdr_t *)0)->msg.get.sink_length) == 4);
+
+ /* Reply */
+ CLASSERT((int)offsetof(lnet_hdr_t, msg.reply.dst_wmd) == 32);
+ CLASSERT((int)sizeof(((lnet_hdr_t *)0)->msg.reply.dst_wmd) == 16);
+
+ /* Hello */
+ CLASSERT((int)offsetof(lnet_hdr_t, msg.hello.incarnation) == 32);
+ CLASSERT((int)sizeof(((lnet_hdr_t *)0)->msg.hello.incarnation) == 8);
+ CLASSERT((int)offsetof(lnet_hdr_t, msg.hello.type) == 40);
+ CLASSERT((int)sizeof(((lnet_hdr_t *)0)->msg.hello.type) == 4);
+}
+
+static lnd_t *lnet_find_lnd_by_type(__u32 type)
+{
+ lnd_t *lnd;
+ struct list_head *tmp;
+
+ /* holding lnd mutex */
+ list_for_each(tmp, &the_lnet.ln_lnds) {
+ lnd = list_entry(tmp, lnd_t, lnd_list);
+
+ if (lnd->lnd_type == type)
+ return lnd;
+ }
+ return NULL;
+}
+
+void
+lnet_register_lnd (lnd_t *lnd)
+{
+ mutex_lock(&the_lnet.ln_lnd_mutex);
+
+ LASSERT(libcfs_isknown_lnd(lnd->lnd_type));
+ LASSERT(lnet_find_lnd_by_type(lnd->lnd_type) == NULL);
+
+ list_add_tail(&lnd->lnd_list, &the_lnet.ln_lnds);
+ lnd->lnd_refcount = 0;
+
+ CDEBUG(D_NET, "%s LND registered\n", libcfs_lnd2str(lnd->lnd_type));
+
+ mutex_unlock(&the_lnet.ln_lnd_mutex);
+}
+EXPORT_SYMBOL(lnet_register_lnd);
+
+void
+lnet_unregister_lnd (lnd_t *lnd)
+{
+ mutex_lock(&the_lnet.ln_lnd_mutex);
+
+ LASSERT(lnet_find_lnd_by_type(lnd->lnd_type) == lnd);
+ LASSERT(lnd->lnd_refcount == 0);
+
+ list_del(&lnd->lnd_list);
+ CDEBUG(D_NET, "%s LND unregistered\n", libcfs_lnd2str(lnd->lnd_type));
+
+ mutex_unlock(&the_lnet.ln_lnd_mutex);
+}
+EXPORT_SYMBOL(lnet_unregister_lnd);
+
+void
+lnet_counters_get(lnet_counters_t *counters)
+{
+ lnet_counters_t *ctr;
+ int i;
+
+ memset(counters, 0, sizeof(*counters));
+
+ lnet_net_lock(LNET_LOCK_EX);
+
+ cfs_percpt_for_each(ctr, i, the_lnet.ln_counters) {
+ counters->msgs_max += ctr->msgs_max;
+ counters->msgs_alloc += ctr->msgs_alloc;
+ counters->errors += ctr->errors;
+ counters->send_count += ctr->send_count;
+ counters->recv_count += ctr->recv_count;
+ counters->route_count += ctr->route_count;
+ counters->drop_count += ctr->drop_count;
+ counters->send_length += ctr->send_length;
+ counters->recv_length += ctr->recv_length;
+ counters->route_length += ctr->route_length;
+ counters->drop_length += ctr->drop_length;
+
+ }
+ lnet_net_unlock(LNET_LOCK_EX);
+}
+EXPORT_SYMBOL(lnet_counters_get);
+
+void
+lnet_counters_reset(void)
+{
+ lnet_counters_t *counters;
+ int i;
+
+ lnet_net_lock(LNET_LOCK_EX);
+
+ cfs_percpt_for_each(counters, i, the_lnet.ln_counters)
+ memset(counters, 0, sizeof(lnet_counters_t));
+
+ lnet_net_unlock(LNET_LOCK_EX);
+}
+
+static char *
+lnet_res_type2str(int type)
+{
+ switch (type) {
+ default:
+ LBUG();
+ case LNET_COOKIE_TYPE_MD:
+ return "MD";
+ case LNET_COOKIE_TYPE_ME:
+ return "ME";
+ case LNET_COOKIE_TYPE_EQ:
+ return "EQ";
+ }
+}
+
+static void
+lnet_res_container_cleanup(struct lnet_res_container *rec)
+{
+ int count = 0;
+
+ if (rec->rec_type == 0) /* not set yet, it's uninitialized */
+ return;
+
+ while (!list_empty(&rec->rec_active)) {
+ struct list_head *e = rec->rec_active.next;
+
+ list_del_init(e);
+ if (rec->rec_type == LNET_COOKIE_TYPE_EQ) {
+ lnet_eq_free(list_entry(e, lnet_eq_t, eq_list));
+
+ } else if (rec->rec_type == LNET_COOKIE_TYPE_MD) {
+ lnet_md_free(list_entry(e, lnet_libmd_t, md_list));
+
+ } else { /* NB: Active MEs should be attached on portals */
+ LBUG();
+ }
+ count++;
+ }
+
+ if (count > 0) {
+ /* Found alive MD/ME/EQ, user really should unlink/free
+ * all of them before finalize LNet, but if someone didn't,
+ * we have to recycle garbage for him */
+ CERROR("%d active elements on exit of %s container\n",
+ count, lnet_res_type2str(rec->rec_type));
+ }
+
+ if (rec->rec_lh_hash != NULL) {
+ LIBCFS_FREE(rec->rec_lh_hash,
+ LNET_LH_HASH_SIZE * sizeof(rec->rec_lh_hash[0]));
+ rec->rec_lh_hash = NULL;
+ }
+
+ rec->rec_type = 0; /* mark it as finalized */
+}
+
+static int
+lnet_res_container_setup(struct lnet_res_container *rec, int cpt, int type)
+{
+ int rc = 0;
+ int i;
+
+ LASSERT(rec->rec_type == 0);
+
+ rec->rec_type = type;
+ INIT_LIST_HEAD(&rec->rec_active);
+
+ rec->rec_lh_cookie = (cpt << LNET_COOKIE_TYPE_BITS) | type;
+
+ /* Arbitrary choice of hash table size */
+ LIBCFS_CPT_ALLOC(rec->rec_lh_hash, lnet_cpt_table(), cpt,
+ LNET_LH_HASH_SIZE * sizeof(rec->rec_lh_hash[0]));
+ if (rec->rec_lh_hash == NULL) {
+ rc = -ENOMEM;
+ goto out;
+ }
+
+ for (i = 0; i < LNET_LH_HASH_SIZE; i++)
+ INIT_LIST_HEAD(&rec->rec_lh_hash[i]);
+
+ return 0;
+
+out:
+ CERROR("Failed to setup %s resource container\n",
+ lnet_res_type2str(type));
+ lnet_res_container_cleanup(rec);
+ return rc;
+}
+
+static void
+lnet_res_containers_destroy(struct lnet_res_container **recs)
+{
+ struct lnet_res_container *rec;
+ int i;
+
+ cfs_percpt_for_each(rec, i, recs)
+ lnet_res_container_cleanup(rec);
+
+ cfs_percpt_free(recs);
+}
+
+static struct lnet_res_container **
+lnet_res_containers_create(int type)
+{
+ struct lnet_res_container **recs;
+ struct lnet_res_container *rec;
+ int rc;
+ int i;
+
+ recs = cfs_percpt_alloc(lnet_cpt_table(), sizeof(*rec));
+ if (recs == NULL) {
+ CERROR("Failed to allocate %s resource containers\n",
+ lnet_res_type2str(type));
+ return NULL;
+ }
+
+ cfs_percpt_for_each(rec, i, recs) {
+ rc = lnet_res_container_setup(rec, i, type);
+ if (rc != 0) {
+ lnet_res_containers_destroy(recs);
+ return NULL;
+ }
+ }
+
+ return recs;
+}
+
+lnet_libhandle_t *
+lnet_res_lh_lookup(struct lnet_res_container *rec, __u64 cookie)
+{
+ /* ALWAYS called with lnet_res_lock held */
+ struct list_head *head;
+ lnet_libhandle_t *lh;
+ unsigned int hash;
+
+ if ((cookie & LNET_COOKIE_MASK) != rec->rec_type)
+ return NULL;
+
+ hash = cookie >> (LNET_COOKIE_TYPE_BITS + LNET_CPT_BITS);
+ head = &rec->rec_lh_hash[hash & LNET_LH_HASH_MASK];
+
+ list_for_each_entry(lh, head, lh_hash_chain) {
+ if (lh->lh_cookie == cookie)
+ return lh;
+ }
+
+ return NULL;
+}
+
+void
+lnet_res_lh_initialize(struct lnet_res_container *rec, lnet_libhandle_t *lh)
+{
+ /* ALWAYS called with lnet_res_lock held */
+ unsigned int ibits = LNET_COOKIE_TYPE_BITS + LNET_CPT_BITS;
+ unsigned int hash;
+
+ lh->lh_cookie = rec->rec_lh_cookie;
+ rec->rec_lh_cookie += 1 << ibits;
+
+ hash = (lh->lh_cookie >> ibits) & LNET_LH_HASH_MASK;
+
+ list_add(&lh->lh_hash_chain, &rec->rec_lh_hash[hash]);
+}
+
+static int lnet_unprepare(void);
+
+static int
+lnet_prepare(lnet_pid_t requested_pid)
+{
+ /* Prepare to bring up the network */
+ struct lnet_res_container **recs;
+ int rc = 0;
+
+ if (requested_pid == LNET_PID_ANY) {
+ /* Don't instantiate LNET just for me */
+ return -ENETDOWN;
+ }
+
+ LASSERT(the_lnet.ln_refcount == 0);
+
+ the_lnet.ln_routing = 0;
+
+ LASSERT((requested_pid & LNET_PID_USERFLAG) == 0);
+ the_lnet.ln_pid = requested_pid;
+
+ INIT_LIST_HEAD(&the_lnet.ln_test_peers);
+ INIT_LIST_HEAD(&the_lnet.ln_peers);
+ INIT_LIST_HEAD(&the_lnet.ln_nets);
+ INIT_LIST_HEAD(&the_lnet.ln_routers);
+ INIT_LIST_HEAD(&the_lnet.ln_drop_rules);
+ INIT_LIST_HEAD(&the_lnet.ln_delay_rules);
+
+ rc = lnet_descriptor_setup();
+ if (rc != 0)
+ goto failed;
+
+ rc = lnet_create_remote_nets_table();
+ if (rc != 0)
+ goto failed;
+
+ /*
+ * NB the interface cookie in wire handles guards against delayed
+ * replies and ACKs appearing valid after reboot.
+ */
+ the_lnet.ln_interface_cookie = ktime_get_real_ns();
+
+ the_lnet.ln_counters = cfs_percpt_alloc(lnet_cpt_table(),
+ sizeof(lnet_counters_t));
+ if (the_lnet.ln_counters == NULL) {
+ CERROR("Failed to allocate counters for LNet\n");
+ rc = -ENOMEM;
+ goto failed;
+ }
+
+ rc = lnet_peer_tables_create();
+ if (rc != 0)
+ goto failed;
+
+ rc = lnet_msg_containers_create();
+ if (rc != 0)
+ goto failed;
+
+ rc = lnet_res_container_setup(&the_lnet.ln_eq_container, 0,
+ LNET_COOKIE_TYPE_EQ);
+ if (rc != 0)
+ goto failed;
+
+ recs = lnet_res_containers_create(LNET_COOKIE_TYPE_ME);
+ if (recs == NULL) {
+ rc = -ENOMEM;
+ goto failed;
+ }
+
+ the_lnet.ln_me_containers = recs;
+
+ recs = lnet_res_containers_create(LNET_COOKIE_TYPE_MD);
+ if (recs == NULL) {
+ rc = -ENOMEM;
+ goto failed;
+ }
+
+ the_lnet.ln_md_containers = recs;
+
+ rc = lnet_portals_create();
+ if (rc != 0) {
+ CERROR("Failed to create portals for LNet: %d\n", rc);
+ goto failed;
+ }
+
+ return 0;
+
+ failed:
+ lnet_unprepare();
+ return rc;
+}
+
+static int
+lnet_unprepare (void)
+{
+ /* NB no LNET_LOCK since this is the last reference. All LND instances
+ * have shut down already, so it is safe to unlink and free all
+ * descriptors, even those that appear committed to a network op (eg MD
+ * with non-zero pending count) */
+
+ lnet_fail_nid(LNET_NID_ANY, 0);
+
+ LASSERT(the_lnet.ln_refcount == 0);
+ LASSERT(list_empty(&the_lnet.ln_test_peers));
+ LASSERT(list_empty(&the_lnet.ln_nets));
+
+ lnet_portals_destroy();
+
+ if (the_lnet.ln_md_containers != NULL) {
+ lnet_res_containers_destroy(the_lnet.ln_md_containers);
+ the_lnet.ln_md_containers = NULL;
+ }
+
+ if (the_lnet.ln_me_containers != NULL) {
+ lnet_res_containers_destroy(the_lnet.ln_me_containers);
+ the_lnet.ln_me_containers = NULL;
+ }
+
+ lnet_res_container_cleanup(&the_lnet.ln_eq_container);
+
+ lnet_msg_containers_destroy();
+ lnet_peer_tables_destroy();
+ lnet_rtrpools_free(0);
+
+ if (the_lnet.ln_counters != NULL) {
+ cfs_percpt_free(the_lnet.ln_counters);
+ the_lnet.ln_counters = NULL;
+ }
+ lnet_destroy_remote_nets_table();
+ lnet_descriptor_cleanup();
+
+ return 0;
+}
+
+lnet_ni_t *
+lnet_net2ni_locked(__u32 net_id, int cpt)
+{
+ struct lnet_ni *ni;
+ struct lnet_net *net;
+
+ LASSERT(cpt != LNET_LOCK_EX);
+
+ list_for_each_entry(net, &the_lnet.ln_nets, net_list) {
+ if (net->net_id == net_id) {
+ ni = list_entry(net->net_ni_list.next, struct lnet_ni,
+ ni_netlist);
+ return ni;
+ }
+ }
+
+ return NULL;
+}
+
+lnet_ni_t *
+lnet_net2ni(__u32 net)
+{
+ lnet_ni_t *ni;
+
+ lnet_net_lock(0);
+ ni = lnet_net2ni_locked(net, 0);
+ lnet_net_unlock(0);
+
+ return ni;
+}
+EXPORT_SYMBOL(lnet_net2ni);
+
+struct lnet_net *
+lnet_get_net_locked(__u32 net_id)
+{
+ struct lnet_net *net;
+
+ list_for_each_entry(net, &the_lnet.ln_nets, net_list) {
+ if (net->net_id == net_id)
+ return net;
+ }
+
+ return NULL;
+}
+
+unsigned int
+lnet_nid_cpt_hash(lnet_nid_t nid, unsigned int number)
+{
+ __u64 key = nid;
+ unsigned int val;
+
+ LASSERT(number >= 1 && number <= LNET_CPT_NUMBER);
+
+ if (number == 1)
+ return 0;
+
+ val = hash_long(key, LNET_CPT_BITS);
+ /* NB: LNET_CP_NUMBER doesn't have to be PO2 */
+ if (val < number)
+ return val;
+
+ return (unsigned int)(key + val + (val >> 1)) % number;
+}
+
+int
+lnet_cpt_of_nid_locked(lnet_nid_t nid, struct lnet_ni *ni)
+{
+ struct lnet_net *net;
+
+ /* must called with hold of lnet_net_lock */
+ if (LNET_CPT_NUMBER == 1)
+ return 0; /* the only one */
+
+ /*
+ * If NI is provided then use the CPT identified in the NI cpt
+ * list if one exists. If one doesn't exist, then that NI is
+ * associated with all CPTs and it follows that the net it belongs
+ * to is implicitly associated with all CPTs, so just hash the nid
+ * and return that.
+ */
+ if (ni != NULL) {
+ if (ni->ni_cpts != NULL)
+ return ni->ni_cpts[lnet_nid_cpt_hash(nid,
+ ni->ni_ncpts)];
+ else
+ return lnet_nid_cpt_hash(nid, LNET_CPT_NUMBER);
+ }
+
+ /* no NI provided so look at the net */
+ net = lnet_get_net_locked(LNET_NIDNET(nid));
+
+ if (net != NULL && net->net_cpts != NULL) {
+ return net->net_cpts[lnet_nid_cpt_hash(nid, net->net_ncpts)];
+ }
+
+ return lnet_nid_cpt_hash(nid, LNET_CPT_NUMBER);
+}
+
+int
+lnet_cpt_of_nid(lnet_nid_t nid, struct lnet_ni *ni)
+{
+ int cpt;
+ int cpt2;
+
+ if (LNET_CPT_NUMBER == 1)
+ return 0; /* the only one */
+
+ cpt = lnet_net_lock_current();
+
+ cpt2 = lnet_cpt_of_nid_locked(nid, ni);
+
+ lnet_net_unlock(cpt);
+
+ return cpt2;
+}
+EXPORT_SYMBOL(lnet_cpt_of_nid);
+
+int
+lnet_islocalnet(__u32 net_id)
+{
+ struct lnet_net *net;
+ int cpt;
+ bool local;
+
+ cpt = lnet_net_lock_current();
+
+ net = lnet_get_net_locked(net_id);
+
+ local = net != NULL;
+
+ lnet_net_unlock(cpt);
+
+ return local;
+}
+
+bool
+lnet_is_ni_healthy_locked(struct lnet_ni *ni)
+{
+ if (ni->ni_state == LNET_NI_STATE_ACTIVE ||
+ ni->ni_state == LNET_NI_STATE_DEGRADED)
+ return true;
+
+ return false;
+}
+
+lnet_ni_t *
+lnet_nid2ni_locked(lnet_nid_t nid, int cpt)
+{
+ struct lnet_net *net;
+ struct lnet_ni *ni;
+
+ LASSERT(cpt != LNET_LOCK_EX);
+
+ list_for_each_entry(net, &the_lnet.ln_nets, net_list) {
+ list_for_each_entry(ni, &net->net_ni_list, ni_netlist) {
+ if (ni->ni_nid == nid)
+ return ni;
+ }
+ }
+
+ return NULL;
+}
+
+lnet_ni_t *
+lnet_nid2ni_addref(lnet_nid_t nid)
+{
+ lnet_ni_t *ni;
+
+ lnet_net_lock(0);
+ ni = lnet_nid2ni_locked(nid, 0);
+ if (ni)
+ lnet_ni_addref_locked(ni, 0);
+ lnet_net_unlock(0);
+
+ return ni;
+}
+EXPORT_SYMBOL(lnet_nid2ni_addref);
+
+int
+lnet_islocalnid(lnet_nid_t nid)
+{
+ struct lnet_ni *ni;
+ int cpt;
+
+ cpt = lnet_net_lock_current();
+ ni = lnet_nid2ni_locked(nid, cpt);
+ lnet_net_unlock(cpt);
+
+ return ni != NULL;
+}
+
+int
+lnet_count_acceptor_nets(void)
+{
+ /* Return the # of NIs that need the acceptor. */
+ int count = 0;
+ struct lnet_net *net;
+ int cpt;
+
+ cpt = lnet_net_lock_current();
+ list_for_each_entry(net, &the_lnet.ln_nets, net_list) {
+ /* all socklnd type networks should have the acceptor
+ * thread started */
+ if (net->net_lnd->lnd_accept != NULL)
+ count++;
+ }
+
+ lnet_net_unlock(cpt);
+
+ return count;
+}
+
+static struct lnet_ping_info *
+lnet_ping_info_create(int num_ni)
+{
+ struct lnet_ping_info *ping_info;
+ unsigned int infosz;
+
+ infosz = offsetof(struct lnet_ping_info, pi_ni[num_ni]);
+ LIBCFS_ALLOC(ping_info, infosz);
+ if (ping_info == NULL) {
+ CERROR("Can't allocate ping info[%d]\n", num_ni);
+ return NULL;
+ }
+
+ ping_info->pi_nnis = num_ni;
+ ping_info->pi_pid = the_lnet.ln_pid;
+ ping_info->pi_magic = LNET_PROTO_PING_MAGIC;
+ ping_info->pi_features = LNET_PING_FEAT_NI_STATUS;
+
+ return ping_info;
+}
+
+static inline int
+lnet_get_net_ni_count_locked(struct lnet_net *net)
+{
+ struct lnet_ni *ni;
+ int count = 0;
+
+ list_for_each_entry(ni, &net->net_ni_list, ni_netlist)
+ count++;
+
+ return count;
+}
+
+static inline int
+lnet_get_ni_count(void)
+{
+ struct lnet_ni *ni;
+ struct lnet_net *net;
+ int count = 0;
+
+ lnet_net_lock(0);
+
+ list_for_each_entry(net, &the_lnet.ln_nets, net_list) {
+ list_for_each_entry(ni, &net->net_ni_list, ni_netlist)
+ count++;
+ }
+
+ lnet_net_unlock(0);
+
+ return count;
+}
+
+static inline void
+lnet_ping_info_free(struct lnet_ping_info *pinfo)
+{
+ LIBCFS_FREE(pinfo,
+ offsetof(struct lnet_ping_info,
+ pi_ni[pinfo->pi_nnis]));
+}
+
+static void
+lnet_ping_info_destroy(void)
+{
+ struct lnet_net *net;
+ struct lnet_ni *ni;
+
+ lnet_net_lock(LNET_LOCK_EX);
+
+ list_for_each_entry(net, &the_lnet.ln_nets, net_list) {
+ list_for_each_entry(ni, &net->net_ni_list, ni_netlist) {
+ lnet_ni_lock(ni);
+ ni->ni_status = NULL;
+ lnet_ni_unlock(ni);
+ }
+ }
+
+ lnet_ping_info_free(the_lnet.ln_ping_info);
+ the_lnet.ln_ping_info = NULL;
+
+ lnet_net_unlock(LNET_LOCK_EX);
+}
+
+static void
+lnet_ping_event_handler(lnet_event_t *event)
+{
+ struct lnet_ping_info *pinfo = event->md.user_ptr;
+
+ if (event->unlinked)
+ pinfo->pi_features = LNET_PING_FEAT_INVAL;
+}
+
+static int
+lnet_ping_info_setup(struct lnet_ping_info **ppinfo, lnet_handle_md_t *md_handle,
+ int ni_count, bool set_eq)
+{
+ lnet_handle_me_t me_handle;
+ lnet_process_id_t id = {LNET_NID_ANY, LNET_PID_ANY};
+ lnet_md_t md = {NULL};
+ int rc, rc2;
+
+ if (set_eq) {
+ rc = LNetEQAlloc(0, lnet_ping_event_handler,
+ &the_lnet.ln_ping_target_eq);
+ if (rc != 0) {
+ CERROR("Can't allocate ping EQ: %d\n", rc);
+ return rc;
+ }
+ }
+
+ *ppinfo = lnet_ping_info_create(ni_count);
+ if (*ppinfo == NULL) {
+ rc = -ENOMEM;
+ goto failed_0;
+ }
+
+ rc = LNetMEAttach(LNET_RESERVED_PORTAL, id,
+ LNET_PROTO_PING_MATCHBITS, 0,
+ LNET_UNLINK, LNET_INS_AFTER,
+ &me_handle);
+ if (rc != 0) {
+ CERROR("Can't create ping ME: %d\n", rc);
+ goto failed_1;
+ }
+
+ /* initialize md content */
+ md.start = *ppinfo;
+ md.length = offsetof(struct lnet_ping_info,
+ pi_ni[(*ppinfo)->pi_nnis]);
+ md.threshold = LNET_MD_THRESH_INF;
+ md.max_size = 0;
+ md.options = LNET_MD_OP_GET | LNET_MD_TRUNCATE |
+ LNET_MD_MANAGE_REMOTE;
+ md.user_ptr = NULL;
+ md.eq_handle = the_lnet.ln_ping_target_eq;
+ md.user_ptr = *ppinfo;
+
+ rc = LNetMDAttach(me_handle, md, LNET_RETAIN, md_handle);
+ if (rc != 0) {
+ CERROR("Can't attach ping MD: %d\n", rc);
+ goto failed_2;
+ }
+
+ return 0;
+
+failed_2:
+ rc2 = LNetMEUnlink(me_handle);
+ LASSERT(rc2 == 0);
+failed_1:
+ lnet_ping_info_free(*ppinfo);
+ *ppinfo = NULL;
+failed_0:
+ if (set_eq)
+ LNetEQFree(the_lnet.ln_ping_target_eq);
+ return rc;
+}
+
+static void
+lnet_ping_md_unlink(struct lnet_ping_info *pinfo, lnet_handle_md_t *md_handle)
+{
+ sigset_t blocked = cfs_block_allsigs();
+
+ LNetMDUnlink(*md_handle);
+ LNetInvalidateHandle(md_handle);
+
+ /* NB md could be busy; this just starts the unlink */
+ while (pinfo->pi_features != LNET_PING_FEAT_INVAL) {
+ CDEBUG(D_NET, "Still waiting for ping MD to unlink\n");
+ set_current_state(TASK_UNINTERRUPTIBLE);
+ schedule_timeout(cfs_time_seconds(1));
+ }
+
+ cfs_restore_sigs(blocked);
+}
+
+static void
+lnet_ping_info_install_locked(struct lnet_ping_info *ping_info)
+{
+ int i;
+ struct lnet_ni *ni;
+ struct lnet_net *net;
+ struct lnet_ni_status *ns;
+
+ i = 0;
+ list_for_each_entry(net, &the_lnet.ln_nets, net_list) {
+ list_for_each_entry(ni, &net->net_ni_list, ni_netlist) {
+ LASSERT(i < ping_info->pi_nnis);
+
+ ns = &ping_info->pi_ni[i];
+
+ ns->ns_nid = ni->ni_nid;
+
+ lnet_ni_lock(ni);
+ ns->ns_status = (ni->ni_status != NULL) ?
+ ni->ni_status->ns_status :
+ LNET_NI_STATUS_UP;
+ ni->ni_status = ns;
+ lnet_ni_unlock(ni);
+
+ i++;
+ }
+
+ }
+}
+
+static void
+lnet_ping_target_update(struct lnet_ping_info *pinfo, lnet_handle_md_t md_handle)
+{
+ struct lnet_ping_info *old_pinfo = NULL;
+ lnet_handle_md_t old_md;
+
+ /* switch the NIs to point to the new ping info created */
+ lnet_net_lock(LNET_LOCK_EX);
+
+ if (!the_lnet.ln_routing)
+ pinfo->pi_features |= LNET_PING_FEAT_RTE_DISABLED;
+ lnet_ping_info_install_locked(pinfo);
+
+ if (the_lnet.ln_ping_info != NULL) {
+ old_pinfo = the_lnet.ln_ping_info;
+ old_md = the_lnet.ln_ping_target_md;
+ }
+ the_lnet.ln_ping_target_md = md_handle;
+ the_lnet.ln_ping_info = pinfo;
+
+ lnet_net_unlock(LNET_LOCK_EX);
+
+ if (old_pinfo != NULL) {
+ /* unlink the old ping info */
+ lnet_ping_md_unlink(old_pinfo, &old_md);
+ lnet_ping_info_free(old_pinfo);
+ }
+}
+
+static void
+lnet_ping_target_fini(void)
+{
+ int rc;
+
+ lnet_ping_md_unlink(the_lnet.ln_ping_info,
+ &the_lnet.ln_ping_target_md);
+
+ rc = LNetEQFree(the_lnet.ln_ping_target_eq);
+ LASSERT(rc == 0);
+
+ lnet_ping_info_destroy();
+}
+
+static int
+lnet_ni_tq_credits(lnet_ni_t *ni)
+{
+ int credits;
+
+ LASSERT(ni->ni_ncpts >= 1);
+
+ if (ni->ni_ncpts == 1)
+ return ni->ni_net->net_tunables.lct_max_tx_credits;
+
+ credits = ni->ni_net->net_tunables.lct_max_tx_credits / ni->ni_ncpts;
+ credits = max(credits, 8 * ni->ni_net->net_tunables.lct_peer_tx_credits);
+ credits = min(credits, ni->ni_net->net_tunables.lct_max_tx_credits);
+
+ return credits;
+}
+
+static void
+lnet_ni_unlink_locked(lnet_ni_t *ni)
+{
+ if (!list_empty(&ni->ni_cptlist)) {
+ list_del_init(&ni->ni_cptlist);
+ lnet_ni_decref_locked(ni, 0);
+ }
+
+ /* move it to zombie list and nobody can find it anymore */
+ LASSERT(!list_empty(&ni->ni_netlist));
+ list_move(&ni->ni_netlist, &ni->ni_net->net_ni_zombie);
+ lnet_ni_decref_locked(ni, 0);
+}
+
+static void
+lnet_clear_zombies_nis_locked(struct lnet_net *net)
+{
+ int i;
+ int islo;
+ lnet_ni_t *ni;
+ struct list_head *zombie_list = &net->net_ni_zombie;
+
+ /*
+ * Now wait for the NIs I just nuked to show up on the zombie
+ * list and shut them down in guaranteed thread context
+ */
+ i = 2;
+ while (!list_empty(zombie_list)) {
+ int *ref;
+ int j;
+
+ ni = list_entry(zombie_list->next,
+ lnet_ni_t, ni_netlist);
+ list_del_init(&ni->ni_netlist);
+ /* the ni should be in deleting state. If it's not it's
+ * a bug */
+ LASSERT(ni->ni_state == LNET_NI_STATE_DELETING);
+ cfs_percpt_for_each(ref, j, ni->ni_refs) {
+ if (*ref == 0)
+ continue;
+ /* still busy, add it back to zombie list */
+ list_add(&ni->ni_netlist, zombie_list);
+ break;
+ }
+
+ if (!list_empty(&ni->ni_netlist)) {
+ lnet_net_unlock(LNET_LOCK_EX);
+ ++i;
+ if ((i & (-i)) == i) {
+ CDEBUG(D_WARNING,
+ "Waiting for zombie LNI %s\n",
+ libcfs_nid2str(ni->ni_nid));
+ }
+ set_current_state(TASK_UNINTERRUPTIBLE);
+ schedule_timeout(cfs_time_seconds(1));
+ lnet_net_lock(LNET_LOCK_EX);
+ continue;
+ }
+
+ lnet_net_unlock(LNET_LOCK_EX);
+
+ islo = ni->ni_net->net_lnd->lnd_type == LOLND;
+
+ LASSERT(!in_interrupt());
+ (net->net_lnd->lnd_shutdown)(ni);
+
+ if (!islo)
+ CDEBUG(D_LNI, "Removed LNI %s\n",
+ libcfs_nid2str(ni->ni_nid));
+
+ lnet_ni_free(ni);
+ i = 2;
+ lnet_net_lock(LNET_LOCK_EX);
+ }
+}
+
+/* shutdown down the NI and release refcount */
+static void
+lnet_shutdown_lndni(struct lnet_ni *ni)
+{
+ int i;
+ struct lnet_net *net = ni->ni_net;
+
+ lnet_net_lock(LNET_LOCK_EX);
+ ni->ni_state = LNET_NI_STATE_DELETING;
+ lnet_ni_unlink_locked(ni);
+ lnet_net_unlock(LNET_LOCK_EX);
+
+ /* clear messages for this NI on the lazy portal */
+ for (i = 0; i < the_lnet.ln_nportals; i++)
+ lnet_clear_lazy_portal(ni, i, "Shutting down NI");
+
+ /* Do peer table cleanup for this ni */
+ lnet_peer_tables_cleanup(ni);
+
+ lnet_net_lock(LNET_LOCK_EX);
+ lnet_clear_zombies_nis_locked(net);
+ lnet_net_unlock(LNET_LOCK_EX);
+}
+
+static void
+lnet_shutdown_lndnet(struct lnet_net *net)
+{
+ struct lnet_ni *ni;
+
+ lnet_net_lock(LNET_LOCK_EX);
+
+ net->net_state = LNET_NET_STATE_DELETING;
+
+ list_del_init(&net->net_list);
+
+ while (!list_empty(&net->net_ni_list)) {
+ ni = list_entry(net->net_ni_list.next,
+ lnet_ni_t, ni_netlist);
+ lnet_net_unlock(LNET_LOCK_EX);
+ lnet_shutdown_lndni(ni);
+ lnet_net_lock(LNET_LOCK_EX);
+ }
+
+ /*
+ * decrement ref count on lnd only when the entire network goes
+ * away
+ */
+ net->net_lnd->lnd_refcount--;
+
+ lnet_net_unlock(LNET_LOCK_EX);
+
+ lnet_net_free(net);
+}
+
+static void
+lnet_shutdown_lndnets(void)
+{
+ struct lnet_net *net;
+
+ /* NB called holding the global mutex */
+
+ /* All quiet on the API front */
+ LASSERT(!the_lnet.ln_shutdown);
+ LASSERT(the_lnet.ln_refcount == 0);
+
+ lnet_net_lock(LNET_LOCK_EX);
+ the_lnet.ln_shutdown = 1; /* flag shutdown */
+
+ while (!list_empty(&the_lnet.ln_nets)) {
+ /*
+ * move the nets to the zombie list to avoid them being
+ * picked up for new work. LONET is also included in the
+ * Nets that will be moved to the zombie list
+ */
+ net = list_entry(the_lnet.ln_nets.next,
+ struct lnet_net, net_list);
+ list_move(&net->net_list, &the_lnet.ln_net_zombie);
+ }
+
+ /* Drop the cached loopback Net. */
+ if (the_lnet.ln_loni != NULL) {
+ lnet_ni_decref_locked(the_lnet.ln_loni, 0);
+ the_lnet.ln_loni = NULL;
+ }
+ lnet_net_unlock(LNET_LOCK_EX);
+
+ /* iterate through the net zombie list and delete each net */
+ while (!list_empty(&the_lnet.ln_net_zombie)) {
+ net = list_entry(the_lnet.ln_net_zombie.next,
+ struct lnet_net, net_list);
+ lnet_shutdown_lndnet(net);
+ }
+
+ lnet_net_lock(LNET_LOCK_EX);
+ the_lnet.ln_shutdown = 0;
+ lnet_net_unlock(LNET_LOCK_EX);
+}
+
+static int
+lnet_startup_lndni(struct lnet_ni *ni, struct lnet_lnd_tunables *tun)
+{
+ int rc = -EINVAL;
+ struct lnet_tx_queue *tq;
+ int i;
+ struct lnet_net *net = ni->ni_net;
+
+ mutex_lock(&the_lnet.ln_lnd_mutex);
+
+ if (tun) {
+ memcpy(&ni->ni_lnd_tunables, tun, sizeof(*tun));
+ ni->ni_lnd_tunables_set = true;
+ }
+
+ rc = (net->net_lnd->lnd_startup)(ni);
+
+ mutex_unlock(&the_lnet.ln_lnd_mutex);
+
+ if (rc != 0) {
+ LCONSOLE_ERROR_MSG(0x105, "Error %d starting up LNI %s\n",
+ rc, libcfs_lnd2str(net->net_lnd->lnd_type));
+ lnet_net_lock(LNET_LOCK_EX);
+ net->net_lnd->lnd_refcount--;
+ lnet_net_unlock(LNET_LOCK_EX);
+ goto failed0;
+ }
+
+ ni->ni_state = LNET_NI_STATE_ACTIVE;
+
+ /* We keep a reference on the loopback net through the loopback NI */
+ if (net->net_lnd->lnd_type == LOLND) {
+ lnet_ni_addref(ni);
+ LASSERT(the_lnet.ln_loni == NULL);
+ the_lnet.ln_loni = ni;
+ ni->ni_net->net_tunables.lct_peer_tx_credits = 0;
+ ni->ni_net->net_tunables.lct_peer_rtr_credits = 0;
+ ni->ni_net->net_tunables.lct_max_tx_credits = 0;
+ ni->ni_net->net_tunables.lct_peer_timeout = 0;
+ return 0;
+ }
+
+ if (ni->ni_net->net_tunables.lct_peer_tx_credits == 0 ||
+ ni->ni_net->net_tunables.lct_max_tx_credits == 0) {
+ LCONSOLE_ERROR_MSG(0x107, "LNI %s has no %scredits\n",
+ libcfs_lnd2str(net->net_lnd->lnd_type),
+ ni->ni_net->net_tunables.lct_peer_tx_credits == 0 ?
+ "" : "per-peer ");
+ /* shutdown the NI since if we get here then it must've already
+ * been started
+ */
+ lnet_shutdown_lndni(ni);
+ return -EINVAL;
+ }
+
+ cfs_percpt_for_each(tq, i, ni->ni_tx_queues) {
+ tq->tq_credits_min =
+ tq->tq_credits_max =
+ tq->tq_credits = lnet_ni_tq_credits(ni);
+ }
+
+ CDEBUG(D_LNI, "Added LNI %s [%d/%d/%d/%d]\n",
+ libcfs_nid2str(ni->ni_nid),
+ ni->ni_net->net_tunables.lct_peer_tx_credits,
+ lnet_ni_tq_credits(ni) * LNET_CPT_NUMBER,
+ ni->ni_net->net_tunables.lct_peer_rtr_credits,
+ ni->ni_net->net_tunables.lct_peer_timeout);
+
+ return 0;
+failed0:
+ lnet_ni_free(ni);
+ return rc;
+}
+
+static int
+lnet_startup_lndnet(struct lnet_net *net, struct lnet_lnd_tunables *tun)
+{
+ struct lnet_ni *ni;
+ struct lnet_net *net_l = NULL;
+ struct list_head local_ni_list;
+ int rc;
+ int ni_count = 0;
+ __u32 lnd_type;
+ lnd_t *lnd;
+ int peer_timeout =
+ net->net_tunables.lct_peer_timeout;
+ int maxtxcredits =
+ net->net_tunables.lct_max_tx_credits;
+ int peerrtrcredits =
+ net->net_tunables.lct_peer_rtr_credits;
+
+ INIT_LIST_HEAD(&local_ni_list);
+
+ /*
+ * make sure that this net is unique. If it isn't then
+ * we are adding interfaces to an already existing network, and
+ * 'net' is just a convenient way to pass in the list.
+ * if it is unique we need to find the LND and load it if
+ * necessary.
+ */
+ if (lnet_net_unique(net->net_id, &the_lnet.ln_nets, &net_l)) {
+ lnd_type = LNET_NETTYP(net->net_id);
+
+ LASSERT(libcfs_isknown_lnd(lnd_type));
+
+ if (lnd_type == CIBLND || lnd_type == OPENIBLND ||
+ lnd_type == IIBLND || lnd_type == VIBLND) {
+ CERROR("LND %s obsoleted\n", libcfs_lnd2str(lnd_type));
+ rc = -EINVAL;
+ goto failed0;
+ }
+
+ mutex_lock(&the_lnet.ln_lnd_mutex);
+ lnd = lnet_find_lnd_by_type(lnd_type);
+
+ if (lnd == NULL) {
+ mutex_unlock(&the_lnet.ln_lnd_mutex);
+ rc = request_module("%s", libcfs_lnd2modname(lnd_type));
+ mutex_lock(&the_lnet.ln_lnd_mutex);
+
+ lnd = lnet_find_lnd_by_type(lnd_type);
+ if (lnd == NULL) {
+ mutex_unlock(&the_lnet.ln_lnd_mutex);
+ CERROR("Can't load LND %s, module %s, rc=%d\n",
+ libcfs_lnd2str(lnd_type),
+ libcfs_lnd2modname(lnd_type), rc);
+#ifndef HAVE_MODULE_LOADING_SUPPORT
+ LCONSOLE_ERROR_MSG(0x104, "Your kernel must be "
+ "compiled with kernel module "
+ "loading support.");
+#endif
+ rc = -EINVAL;
+ goto failed0;
+ }
+ }
+
+ lnet_net_lock(LNET_LOCK_EX);
+ lnd->lnd_refcount++;
+ lnet_net_unlock(LNET_LOCK_EX);
+
+ net->net_lnd = lnd;
+
+ mutex_unlock(&the_lnet.ln_lnd_mutex);
+
+ net_l = net;
+ }
+
+ /*
+ * net_l: if the network being added is unique then net_l
+ * will point to that network
+ * if the network being added is not unique then
+ * net_l points to the existing network.
+ *
+ * When we enter the loop below, we'll pick NIs off he
+ * network beign added and start them up, then add them to
+ * a local ni list. Once we've successfully started all
+ * the NIs then we join the local NI list (of started up
+ * networks) with the net_l->net_ni_list, which should
+ * point to the correct network to add the new ni list to
+ *
+ * If any of the new NIs fail to start up, then we want to
+ * iterate through the local ni list, which should include
+ * any NIs which were successfully started up, and shut
+ * them down.
+ *
+ * After than we want to delete the network being added,
+ * to avoid a memory leak.
+ */
+
+ /*
+ * When a network uses TCP bonding then all its interfaces
+ * must be specified when the network is first defined: the
+ * TCP bonding code doesn't allow for interfaces to be added
+ * or removed.
+ */
+ if (net_l != net && net_l != NULL && use_tcp_bonding &&
+ LNET_NETTYP(net_l->net_id) == SOCKLND) {
+ rc = -EINVAL;
+ goto failed0;
+ }
+
+ while (!list_empty(&net->net_ni_added)) {
+ ni = list_entry(net->net_ni_added.next, struct lnet_ni,
+ ni_netlist);
+ list_del_init(&ni->ni_netlist);
+
+ /* make sure that the the NI we're about to start
+ * up is actually unique. if it's not fail. */
+ if (!lnet_ni_unique_net(&net_l->net_ni_list,
+ ni->ni_interfaces[0])) {
+ rc = -EINVAL;
+ goto failed1;
+ }
+
+ /* adjust the pointer the parent network, just in case it
+ * the net is a duplicate */
+ ni->ni_net = net_l;
+
+ rc = lnet_startup_lndni(ni, tun);
+
+ LASSERT(ni->ni_net->net_tunables.lct_peer_timeout <= 0 ||
+ ni->ni_net->net_lnd->lnd_query != NULL);
+
+ if (rc < 0)
+ goto failed1;
+
+ lnet_ni_addref(ni);
+ list_add_tail(&ni->ni_netlist, &local_ni_list);
+
+ ni_count++;
+ }
+
+ lnet_net_lock(LNET_LOCK_EX);
+ list_splice_tail(&local_ni_list, &net_l->net_ni_list);
+ lnet_incr_dlc_seq();
+ lnet_net_unlock(LNET_LOCK_EX);
+
+ /* if the network is not unique then we don't want to keep
+ * it around after we're done. Free it. Otherwise add that
+ * net to the global the_lnet.ln_nets */
+ if (net_l != net && net_l != NULL) {
+ /*
+ * TODO - note. currently the tunables can not be updated
+ * once added
+ */
+ lnet_net_free(net);
+ } else {
+ net->net_state = LNET_NET_STATE_ACTIVE;
+ /*
+ * restore tunables after it has been overwitten by the
+ * lnd
+ */
+ if (peer_timeout != -1)
+ net->net_tunables.lct_peer_timeout = peer_timeout;
+ if (maxtxcredits != -1)
+ net->net_tunables.lct_max_tx_credits = maxtxcredits;
+ if (peerrtrcredits != -1)
+ net->net_tunables.lct_peer_rtr_credits = peerrtrcredits;
+
+ lnet_net_lock(LNET_LOCK_EX);
+ list_add_tail(&net->net_list, &the_lnet.ln_nets);
+ lnet_net_unlock(LNET_LOCK_EX);
+ }
+
+ return ni_count;
+
+failed1:
+ /*
+ * shutdown the new NIs that are being started up
+ * free the NET being started
+ */
+ while (!list_empty(&local_ni_list)) {
+ ni = list_entry(local_ni_list.next, struct lnet_ni,
+ ni_netlist);
+
+ lnet_shutdown_lndni(ni);
+ }
+
+failed0:
+ lnet_net_free(net);
+
+ return rc;
+}
+
+static int
+lnet_startup_lndnets(struct list_head *netlist)
+{
+ struct lnet_net *net;
+ int rc;
+ int ni_count = 0;
+
+ while (!list_empty(netlist)) {
+ net = list_entry(netlist->next, struct lnet_net, net_list);
+ list_del_init(&net->net_list);
+
+ rc = lnet_startup_lndnet(net, NULL);
+
+ if (rc < 0)
+ goto failed;
+
+ ni_count += rc;
+ }
+
+ return ni_count;
+failed:
+ lnet_shutdown_lndnets();
+
+ return rc;
+}
+
+/**
+ * Initialize LNet library.
+ *
+ * Automatically called at module loading time. Caller has to call
+ * lnet_lib_exit() after a call to lnet_lib_init(), if and only if the
+ * latter returned 0. It must be called exactly once.