X-Git-Url: https://git.whamcloud.com/?p=fs%2Flustre-release.git;a=blobdiff_plain;f=lnet%2Flnet%2Fapi-ni.c;h=322565a070a96fd5f72fa8c615154f53c4b3aa1a;hp=d10ff58e1e5f3c717af2d7e5fbd11f746ae214bc;hb=d123c47a18adbf5665ed63d99c53117b84db9ec8;hpb=d54afb86116c0640d7a201571b337042c87a3e40 diff --git a/lnet/lnet/api-ni.c b/lnet/lnet/api-ni.c index d10ff58..322565a 100644 --- a/lnet/lnet/api-ni.c +++ b/lnet/lnet/api-ni.c @@ -37,7 +37,11 @@ #include #include #include +#ifdef HAVE_SCHED_HEADERS +#include +#endif +#include #include #define D_LNI D_CONSOLE @@ -68,16 +72,64 @@ 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; +static int use_tcp_bonding; 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"); + "use_tcp_bonding parameter has been removed"); unsigned int lnet_numa_range = 0; module_param(lnet_numa_range, uint, 0444); MODULE_PARM_DESC(lnet_numa_range, "NUMA range to consider during Multi-Rail selection"); +/* + * lnet_health_sensitivity determines by how much we decrement the health + * value on sending error. The value defaults to 100, which means health + * interface health is decremented by 100 points every failure. + */ +unsigned int lnet_health_sensitivity = 100; +static int sensitivity_set(const char *val, cfs_kernel_param_arg_t *kp); +#ifdef HAVE_KERNEL_PARAM_OPS +static struct kernel_param_ops param_ops_health_sensitivity = { + .set = sensitivity_set, + .get = param_get_int, +}; +#define param_check_health_sensitivity(name, p) \ + __param_check(name, p, int) +module_param(lnet_health_sensitivity, health_sensitivity, S_IRUGO|S_IWUSR); +#else +module_param_call(lnet_health_sensitivity, sensitivity_set, param_get_int, + &lnet_health_sensitivity, S_IRUGO|S_IWUSR); +#endif +MODULE_PARM_DESC(lnet_health_sensitivity, + "Value to decrement the health value by on error"); + +/* + * lnet_recovery_interval determines how often we should perform recovery + * on unhealthy interfaces. + */ +unsigned int lnet_recovery_interval = 1; +static int recovery_interval_set(const char *val, cfs_kernel_param_arg_t *kp); +#ifdef HAVE_KERNEL_PARAM_OPS +static struct kernel_param_ops param_ops_recovery_interval = { + .set = recovery_interval_set, + .get = param_get_int, +}; +#define param_check_recovery_interval(name, p) \ + __param_check(name, p, int) +module_param(lnet_recovery_interval, recovery_interval, S_IRUGO|S_IWUSR); +#else +module_param_call(lnet_recovery_interval, recovery_interval_set, param_get_int, + &lnet_recovery_interval, S_IRUGO|S_IWUSR); +#endif +MODULE_PARM_DESC(lnet_recovery_interval, + "Interval to recover unhealthy interfaces in seconds"); + +unsigned int lnet_recovery_limit; +module_param(lnet_recovery_limit, uint, 0644); +MODULE_PARM_DESC(lnet_recovery_limit, + "How long to attempt recovery of unhealthy peer interfaces in seconds. Set to 0 to allow indefinite recovery"); + static int lnet_interfaces_max = LNET_INTERFACES_MAX_DEFAULT; static int intf_max_set(const char *val, cfs_kernel_param_arg_t *kp); @@ -117,10 +169,92 @@ module_param_call(lnet_peer_discovery_disabled, discovery_set, param_get_int, MODULE_PARM_DESC(lnet_peer_discovery_disabled, "Set to 1 to disable peer discovery on this node."); -unsigned lnet_transaction_timeout = 5; -module_param(lnet_transaction_timeout, uint, 0444); +unsigned int lnet_drop_asym_route; +static int drop_asym_route_set(const char *val, cfs_kernel_param_arg_t *kp); + +static struct kernel_param_ops param_ops_drop_asym_route = { + .set = drop_asym_route_set, + .get = param_get_int, +}; + +#define param_check_drop_asym_route(name, p) \ + __param_check(name, p, int) +#ifdef HAVE_KERNEL_PARAM_OPS +module_param(lnet_drop_asym_route, drop_asym_route, 0644); +#else +module_param_call(lnet_drop_asym_route, drop_asym_route_set, param_get_int, + ¶m_ops_drop_asym_route, 0644); +#endif +MODULE_PARM_DESC(lnet_drop_asym_route, + "Set to 1 to drop asymmetrical route messages."); + +#define LNET_TRANSACTION_TIMEOUT_DEFAULT 50 +unsigned int lnet_transaction_timeout = LNET_TRANSACTION_TIMEOUT_DEFAULT; +static int transaction_to_set(const char *val, cfs_kernel_param_arg_t *kp); +#ifdef HAVE_KERNEL_PARAM_OPS +static struct kernel_param_ops param_ops_transaction_timeout = { + .set = transaction_to_set, + .get = param_get_int, +}; + +#define param_check_transaction_timeout(name, p) \ + __param_check(name, p, int) +module_param(lnet_transaction_timeout, transaction_timeout, S_IRUGO|S_IWUSR); +#else +module_param_call(lnet_transaction_timeout, transaction_to_set, param_get_int, + &lnet_transaction_timeout, S_IRUGO|S_IWUSR); +#endif MODULE_PARM_DESC(lnet_transaction_timeout, - "Time in seconds to wait for a REPLY or an ACK"); + "Maximum number of seconds to wait for a peer response."); + +#define LNET_RETRY_COUNT_DEFAULT 2 +unsigned int lnet_retry_count = LNET_RETRY_COUNT_DEFAULT; +static int retry_count_set(const char *val, cfs_kernel_param_arg_t *kp); +#ifdef HAVE_KERNEL_PARAM_OPS +static struct kernel_param_ops param_ops_retry_count = { + .set = retry_count_set, + .get = param_get_int, +}; + +#define param_check_retry_count(name, p) \ + __param_check(name, p, int) +module_param(lnet_retry_count, retry_count, S_IRUGO|S_IWUSR); +#else +module_param_call(lnet_retry_count, retry_count_set, param_get_int, + &lnet_retry_count, S_IRUGO|S_IWUSR); +#endif +MODULE_PARM_DESC(lnet_retry_count, + "Maximum number of times to retry transmitting a message"); + +unsigned int lnet_response_tracking = 3; +static int response_tracking_set(const char *val, cfs_kernel_param_arg_t *kp); + +#ifdef HAVE_KERNEL_PARAM_OPS +static struct kernel_param_ops param_ops_response_tracking = { + .set = response_tracking_set, + .get = param_get_int, +}; + +#define param_check_response_tracking(name, p) \ + __param_check(name, p, int) +module_param(lnet_response_tracking, response_tracking, 0644); +#else +module_param_call(lnet_response_tracking, response_tracking_set, param_get_int, + &lnet_response_tracking, 0644); +#endif +MODULE_PARM_DESC(lnet_response_tracking, + "(0|1|2|3) LNet Internal Only|GET Reply only|PUT ACK only|Full Tracking (default)"); + +#define LNET_LND_TIMEOUT_DEFAULT ((LNET_TRANSACTION_TIMEOUT_DEFAULT - 1) / \ + (LNET_RETRY_COUNT_DEFAULT + 1)) +unsigned int lnet_lnd_timeout = LNET_LND_TIMEOUT_DEFAULT; +static void lnet_set_lnd_timeout(void) +{ + lnet_lnd_timeout = (lnet_transaction_timeout - 1) / + (lnet_retry_count + 1); +} + +unsigned int lnet_current_net_count; /* * This sequence number keeps track of how many times DLC was used to @@ -138,10 +272,79 @@ static int lnet_discover(struct lnet_process_id id, __u32 force, struct lnet_process_id __user *ids, int n_ids); static int +sensitivity_set(const char *val, cfs_kernel_param_arg_t *kp) +{ + int rc; + unsigned *sensitivity = (unsigned *)kp->arg; + unsigned long value; + + rc = kstrtoul(val, 0, &value); + if (rc) { + CERROR("Invalid module parameter value for 'lnet_health_sensitivity'\n"); + return rc; + } + + /* + * The purpose of locking the api_mutex here is to ensure that + * the correct value ends up stored properly. + */ + mutex_lock(&the_lnet.ln_api_mutex); + + if (value > LNET_MAX_HEALTH_VALUE) { + mutex_unlock(&the_lnet.ln_api_mutex); + CERROR("Invalid health value. Maximum: %d value = %lu\n", + LNET_MAX_HEALTH_VALUE, value); + return -EINVAL; + } + + if (*sensitivity != 0 && value == 0 && lnet_retry_count != 0) { + lnet_retry_count = 0; + lnet_set_lnd_timeout(); + } + + *sensitivity = value; + + mutex_unlock(&the_lnet.ln_api_mutex); + + return 0; +} + +static int +recovery_interval_set(const char *val, cfs_kernel_param_arg_t *kp) +{ + int rc; + unsigned *interval = (unsigned *)kp->arg; + unsigned long value; + + rc = kstrtoul(val, 0, &value); + if (rc) { + CERROR("Invalid module parameter value for 'lnet_recovery_interval'\n"); + return rc; + } + + if (value < 1) { + CERROR("lnet_recovery_interval must be at least 1 second\n"); + return -EINVAL; + } + + /* + * The purpose of locking the api_mutex here is to ensure that + * the correct value ends up stored properly. + */ + mutex_lock(&the_lnet.ln_api_mutex); + + *interval = value; + + mutex_unlock(&the_lnet.ln_api_mutex); + + return 0; +} + +static int discovery_set(const char *val, cfs_kernel_param_arg_t *kp) { int rc; - unsigned *discovery = (unsigned *)kp->arg; + unsigned *discovery_off = (unsigned *)kp->arg; unsigned long value; struct lnet_ping_buffer *pbuf; @@ -159,14 +362,20 @@ discovery_set(const char *val, cfs_kernel_param_arg_t *kp) */ mutex_lock(&the_lnet.ln_api_mutex); - if (value == *discovery) { + if (value == *discovery_off) { mutex_unlock(&the_lnet.ln_api_mutex); return 0; } - *discovery = value; - + /* + * We still want to set the discovery value even when LNet is not + * running. This is the case when LNet is being loaded and we want + * the module parameters to take effect. Otherwise if we're + * changing the value dynamically, we want to set it after + * updating the peers + */ if (the_lnet.ln_state != LNET_STATE_RUNNING) { + *discovery_off = value; mutex_unlock(&the_lnet.ln_api_mutex); return 0; } @@ -180,7 +389,130 @@ discovery_set(const char *val, cfs_kernel_param_arg_t *kp) pbuf->pb_info.pi_features |= LNET_PING_FEAT_DISCOVERY; lnet_net_unlock(LNET_LOCK_EX); - lnet_push_update_to_peers(1); + /* only send a push when we're turning off discovery */ + if (*discovery_off <= 0 && value > 0) + lnet_push_update_to_peers(1); + *discovery_off = value; + + mutex_unlock(&the_lnet.ln_api_mutex); + + return 0; +} + +static int +drop_asym_route_set(const char *val, cfs_kernel_param_arg_t *kp) +{ + int rc; + unsigned int *drop_asym_route = (unsigned int *)kp->arg; + unsigned long value; + + rc = kstrtoul(val, 0, &value); + if (rc) { + CERROR("Invalid module parameter value for " + "'lnet_drop_asym_route'\n"); + return rc; + } + + /* + * The purpose of locking the api_mutex here is to ensure that + * the correct value ends up stored properly. + */ + mutex_lock(&the_lnet.ln_api_mutex); + + if (value == *drop_asym_route) { + mutex_unlock(&the_lnet.ln_api_mutex); + return 0; + } + + *drop_asym_route = value; + + mutex_unlock(&the_lnet.ln_api_mutex); + + return 0; +} + +static int +transaction_to_set(const char *val, cfs_kernel_param_arg_t *kp) +{ + int rc; + unsigned *transaction_to = (unsigned *)kp->arg; + unsigned long value; + + rc = kstrtoul(val, 0, &value); + if (rc) { + CERROR("Invalid module parameter value for 'lnet_transaction_timeout'\n"); + return rc; + } + + /* + * The purpose of locking the api_mutex here is to ensure that + * the correct value ends up stored properly. + */ + mutex_lock(&the_lnet.ln_api_mutex); + + if (value <= lnet_retry_count || value == 0) { + mutex_unlock(&the_lnet.ln_api_mutex); + CERROR("Invalid value for lnet_transaction_timeout (%lu). " + "Has to be greater than lnet_retry_count (%u)\n", + value, lnet_retry_count); + return -EINVAL; + } + + if (value == *transaction_to) { + mutex_unlock(&the_lnet.ln_api_mutex); + return 0; + } + + *transaction_to = value; + /* Update the lnet_lnd_timeout now that we've modified the + * transaction timeout + */ + lnet_set_lnd_timeout(); + + mutex_unlock(&the_lnet.ln_api_mutex); + + return 0; +} + +static int +retry_count_set(const char *val, cfs_kernel_param_arg_t *kp) +{ + int rc; + unsigned *retry_count = (unsigned *)kp->arg; + unsigned long value; + + rc = kstrtoul(val, 0, &value); + if (rc) { + CERROR("Invalid module parameter value for 'lnet_retry_count'\n"); + return rc; + } + + /* + * The purpose of locking the api_mutex here is to ensure that + * the correct value ends up stored properly. + */ + mutex_lock(&the_lnet.ln_api_mutex); + + if (lnet_health_sensitivity == 0 && value > 0) { + mutex_unlock(&the_lnet.ln_api_mutex); + CERROR("Can not set lnet_retry_count when health feature is turned off\n"); + return -EINVAL; + } + + if (value > lnet_transaction_timeout) { + mutex_unlock(&the_lnet.ln_api_mutex); + CERROR("Invalid value for lnet_retry_count (%lu). " + "Has to be smaller than lnet_transaction_timeout (%u)\n", + value, lnet_transaction_timeout); + return -EINVAL; + } + + *retry_count = value; + + /* Update the lnet_lnd_timeout now that we've modified the + * retry count + */ + lnet_set_lnd_timeout(); mutex_unlock(&the_lnet.ln_api_mutex); @@ -209,17 +541,40 @@ intf_max_set(const char *val, cfs_kernel_param_arg_t *kp) return 0; } -static char * +static int +response_tracking_set(const char *val, cfs_kernel_param_arg_t *kp) +{ + int rc; + unsigned long new_value; + + rc = kstrtoul(val, 0, &new_value); + if (rc) { + CERROR("Invalid value for 'lnet_response_tracking'\n"); + return -EINVAL; + } + + if (new_value < 0 || new_value > 3) { + CWARN("Invalid value (%lu) for 'lnet_response_tracking'\n", + new_value); + return -EINVAL; + } + + lnet_response_tracking = new_value; + + return 0; +} + +static const char * lnet_get_routes(void) { return routes; } -static char * +static const char * lnet_get_networks(void) { - char *nets; - int rc; + const char *nets; + int rc; if (*networks != 0 && *ip2nets != 0) { LCONSOLE_ERROR_MSG(0x101, "Please specify EITHER 'networks' or " @@ -243,22 +598,19 @@ lnet_init_locks(void) { spin_lock_init(&the_lnet.ln_eq_wait_lock); spin_lock_init(&the_lnet.ln_msg_resend_lock); - init_waitqueue_head(&the_lnet.ln_eq_waitq); - init_waitqueue_head(&the_lnet.ln_rc_waitq); + init_completion(&the_lnet.ln_mt_wait_complete); mutex_init(&the_lnet.ln_lnd_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 */ +struct kmem_cache *lnet_udsp_cachep; /* udsp cache */ +struct kmem_cache *lnet_rspt_cachep; /* response tracker cache */ +struct kmem_cache *lnet_msg_cachep; static int -lnet_descriptor_setup(void) +lnet_slab_setup(void) { /* create specific kmem_cache for MEs and small MDs (i.e., originally * allocated in kmem_cache). @@ -274,12 +626,42 @@ lnet_descriptor_setup(void) if (!lnet_small_mds_cachep) return -ENOMEM; + lnet_udsp_cachep = kmem_cache_create("lnet_udsp", + sizeof(struct lnet_udsp), + 0, 0, NULL); + if (!lnet_udsp_cachep) + return -ENOMEM; + + lnet_rspt_cachep = kmem_cache_create("lnet_rspt", sizeof(struct lnet_rsp_tracker), + 0, 0, NULL); + if (!lnet_rspt_cachep) + return -ENOMEM; + + lnet_msg_cachep = kmem_cache_create("lnet_msg", sizeof(struct lnet_msg), + 0, 0, NULL); + if (!lnet_msg_cachep) + return -ENOMEM; + return 0; } static void -lnet_descriptor_cleanup(void) +lnet_slab_cleanup(void) { + if (lnet_msg_cachep) { + kmem_cache_destroy(lnet_msg_cachep); + lnet_msg_cachep = NULL; + } + + if (lnet_rspt_cachep) { + kmem_cache_destroy(lnet_rspt_cachep); + lnet_rspt_cachep = NULL; + } + + if (lnet_udsp_cachep) { + kmem_cache_destroy(lnet_udsp_cachep); + lnet_udsp_cachep = NULL; + } if (lnet_small_mds_cachep) { kmem_cache_destroy(lnet_small_mds_cachep); @@ -300,7 +682,7 @@ lnet_create_remote_nets_table(void) 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)); + CFS_ALLOC_PTR_ARRAY(hash, LNET_REMOTE_NETS_HASH_SIZE); if (hash == NULL) { CERROR("Failed to create remote nets hash table\n"); return -ENOMEM; @@ -323,9 +705,8 @@ lnet_destroy_remote_nets_table(void) 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])); + CFS_FREE_PTR_ARRAY(the_lnet.ln_remote_nets_hash, + LNET_REMOTE_NETS_HASH_SIZE); the_lnet.ln_remote_nets_hash = NULL; } @@ -341,8 +722,6 @@ lnet_destroy_locks(void) cfs_percpt_lock_free(the_lnet.ln_net_lock); the_lnet.ln_net_lock = NULL; } - - lnet_fini_locks(); } static int @@ -370,156 +749,211 @@ 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) */ + * 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); + BUILD_BUG_ON(LNET_PROTO_TCP_MAGIC != 0xeebc0ded); + BUILD_BUG_ON(LNET_PROTO_TCP_VERSION_MAJOR != 1); + BUILD_BUG_ON(LNET_PROTO_TCP_VERSION_MINOR != 0); + BUILD_BUG_ON(LNET_MSG_ACK != 0); + BUILD_BUG_ON(LNET_MSG_PUT != 1); + BUILD_BUG_ON(LNET_MSG_GET != 2); + BUILD_BUG_ON(LNET_MSG_REPLY != 3); + BUILD_BUG_ON(LNET_MSG_HELLO != 4); + + BUILD_BUG_ON((int)sizeof(lnet_nid_t) != 8); + BUILD_BUG_ON((int)sizeof(lnet_pid_t) != 4); + + /* Checks for struct lnet_process_id_packed */ + BUILD_BUG_ON((int)sizeof(struct lnet_process_id_packed) != 12); + BUILD_BUG_ON((int)offsetof(struct lnet_process_id_packed, nid) != 0); + BUILD_BUG_ON((int)sizeof(((struct lnet_process_id_packed *)0)->nid) != 8); + BUILD_BUG_ON((int)offsetof(struct lnet_process_id_packed, pid) != 8); + BUILD_BUG_ON((int)sizeof(((struct lnet_process_id_packed *)0)->pid) != 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); + BUILD_BUG_ON((int)sizeof(struct lnet_handle_wire) != 16); + BUILD_BUG_ON((int)offsetof(struct lnet_handle_wire, + wh_interface_cookie) != 0); + BUILD_BUG_ON((int)sizeof(((struct lnet_handle_wire *)0)->wh_interface_cookie) != 8); + BUILD_BUG_ON((int)offsetof(struct lnet_handle_wire, + wh_object_cookie) != 8); + BUILD_BUG_ON((int)sizeof(((struct lnet_handle_wire *)0)->wh_object_cookie) != 8); /* Checks for struct struct lnet_magicversion */ - CLASSERT((int)sizeof(struct lnet_magicversion) == 8); - CLASSERT((int)offsetof(struct lnet_magicversion, magic) == 0); - CLASSERT((int)sizeof(((struct lnet_magicversion *)0)->magic) == 4); - CLASSERT((int)offsetof(struct lnet_magicversion, version_major) == 4); - CLASSERT((int)sizeof(((struct lnet_magicversion *)0)->version_major) == 2); - CLASSERT((int)offsetof(struct lnet_magicversion, version_minor) == 6); - CLASSERT((int)sizeof(((struct lnet_magicversion *)0)->version_minor) == 2); + BUILD_BUG_ON((int)sizeof(struct lnet_magicversion) != 8); + BUILD_BUG_ON((int)offsetof(struct lnet_magicversion, magic) != 0); + BUILD_BUG_ON((int)sizeof(((struct lnet_magicversion *)0)->magic) != 4); + BUILD_BUG_ON((int)offsetof(struct lnet_magicversion, version_major) != 4); + BUILD_BUG_ON((int)sizeof(((struct lnet_magicversion *)0)->version_major) != 2); + BUILD_BUG_ON((int)offsetof(struct lnet_magicversion, + version_minor) != 6); + BUILD_BUG_ON((int)sizeof(((struct lnet_magicversion *)0)->version_minor) != 2); /* Checks for struct struct lnet_hdr */ - CLASSERT((int)sizeof(struct lnet_hdr) == 72); - CLASSERT((int)offsetof(struct lnet_hdr, dest_nid) == 0); - CLASSERT((int)sizeof(((struct lnet_hdr *)0)->dest_nid) == 8); - CLASSERT((int)offsetof(struct lnet_hdr, src_nid) == 8); - CLASSERT((int)sizeof(((struct lnet_hdr *)0)->src_nid) == 8); - CLASSERT((int)offsetof(struct lnet_hdr, dest_pid) == 16); - CLASSERT((int)sizeof(((struct lnet_hdr *)0)->dest_pid) == 4); - CLASSERT((int)offsetof(struct lnet_hdr, src_pid) == 20); - CLASSERT((int)sizeof(((struct lnet_hdr *)0)->src_pid) == 4); - CLASSERT((int)offsetof(struct lnet_hdr, type) == 24); - CLASSERT((int)sizeof(((struct lnet_hdr *)0)->type) == 4); - CLASSERT((int)offsetof(struct lnet_hdr, payload_length) == 28); - CLASSERT((int)sizeof(((struct lnet_hdr *)0)->payload_length) == 4); - CLASSERT((int)offsetof(struct lnet_hdr, msg) == 32); - CLASSERT((int)sizeof(((struct lnet_hdr *)0)->msg) == 40); + BUILD_BUG_ON((int)sizeof(struct lnet_hdr) != 72); + BUILD_BUG_ON((int)offsetof(struct lnet_hdr, dest_nid) != 0); + BUILD_BUG_ON((int)sizeof(((struct lnet_hdr *)0)->dest_nid) != 8); + BUILD_BUG_ON((int)offsetof(struct lnet_hdr, src_nid) != 8); + BUILD_BUG_ON((int)sizeof(((struct lnet_hdr *)0)->src_nid) != 8); + BUILD_BUG_ON((int)offsetof(struct lnet_hdr, dest_pid) != 16); + BUILD_BUG_ON((int)sizeof(((struct lnet_hdr *)0)->dest_pid) != 4); + BUILD_BUG_ON((int)offsetof(struct lnet_hdr, src_pid) != 20); + BUILD_BUG_ON((int)sizeof(((struct lnet_hdr *)0)->src_pid) != 4); + BUILD_BUG_ON((int)offsetof(struct lnet_hdr, type) != 24); + BUILD_BUG_ON((int)sizeof(((struct lnet_hdr *)0)->type) != 4); + BUILD_BUG_ON((int)offsetof(struct lnet_hdr, payload_length) != 28); + BUILD_BUG_ON((int)sizeof(((struct lnet_hdr *)0)->payload_length) != 4); + BUILD_BUG_ON((int)offsetof(struct lnet_hdr, msg) != 32); + BUILD_BUG_ON((int)sizeof(((struct lnet_hdr *)0)->msg) != 40); /* Ack */ - CLASSERT((int)offsetof(struct lnet_hdr, msg.ack.dst_wmd) == 32); - CLASSERT((int)sizeof(((struct lnet_hdr *)0)->msg.ack.dst_wmd) == 16); - CLASSERT((int)offsetof(struct lnet_hdr, msg.ack.match_bits) == 48); - CLASSERT((int)sizeof(((struct lnet_hdr *)0)->msg.ack.match_bits) == 8); - CLASSERT((int)offsetof(struct lnet_hdr, msg.ack.mlength) == 56); - CLASSERT((int)sizeof(((struct lnet_hdr *)0)->msg.ack.mlength) == 4); + BUILD_BUG_ON((int)offsetof(struct lnet_hdr, msg.ack.dst_wmd) != 32); + BUILD_BUG_ON((int)sizeof(((struct lnet_hdr *)0)->msg.ack.dst_wmd) != 16); + BUILD_BUG_ON((int)offsetof(struct lnet_hdr, msg.ack.match_bits) != 48); + BUILD_BUG_ON((int)sizeof(((struct lnet_hdr *)0)->msg.ack.match_bits) != 8); + BUILD_BUG_ON((int)offsetof(struct lnet_hdr, msg.ack.mlength) != 56); + BUILD_BUG_ON((int)sizeof(((struct lnet_hdr *)0)->msg.ack.mlength) != 4); /* Put */ - CLASSERT((int)offsetof(struct lnet_hdr, msg.put.ack_wmd) == 32); - CLASSERT((int)sizeof(((struct lnet_hdr *)0)->msg.put.ack_wmd) == 16); - CLASSERT((int)offsetof(struct lnet_hdr, msg.put.match_bits) == 48); - CLASSERT((int)sizeof(((struct lnet_hdr *)0)->msg.put.match_bits) == 8); - CLASSERT((int)offsetof(struct lnet_hdr, msg.put.hdr_data) == 56); - CLASSERT((int)sizeof(((struct lnet_hdr *)0)->msg.put.hdr_data) == 8); - CLASSERT((int)offsetof(struct lnet_hdr, msg.put.ptl_index) == 64); - CLASSERT((int)sizeof(((struct lnet_hdr *)0)->msg.put.ptl_index) == 4); - CLASSERT((int)offsetof(struct lnet_hdr, msg.put.offset) == 68); - CLASSERT((int)sizeof(((struct lnet_hdr *)0)->msg.put.offset) == 4); + BUILD_BUG_ON((int)offsetof(struct lnet_hdr, msg.put.ack_wmd) != 32); + BUILD_BUG_ON((int)sizeof(((struct lnet_hdr *)0)->msg.put.ack_wmd) != 16); + BUILD_BUG_ON((int)offsetof(struct lnet_hdr, msg.put.match_bits) != 48); + BUILD_BUG_ON((int)sizeof(((struct lnet_hdr *)0)->msg.put.match_bits) != 8); + BUILD_BUG_ON((int)offsetof(struct lnet_hdr, msg.put.hdr_data) != 56); + BUILD_BUG_ON((int)sizeof(((struct lnet_hdr *)0)->msg.put.hdr_data) != 8); + BUILD_BUG_ON((int)offsetof(struct lnet_hdr, msg.put.ptl_index) != 64); + BUILD_BUG_ON((int)sizeof(((struct lnet_hdr *)0)->msg.put.ptl_index) != 4); + BUILD_BUG_ON((int)offsetof(struct lnet_hdr, msg.put.offset) != 68); + BUILD_BUG_ON((int)sizeof(((struct lnet_hdr *)0)->msg.put.offset) != 4); /* Get */ - CLASSERT((int)offsetof(struct lnet_hdr, msg.get.return_wmd) == 32); - CLASSERT((int)sizeof(((struct lnet_hdr *)0)->msg.get.return_wmd) == 16); - CLASSERT((int)offsetof(struct lnet_hdr, msg.get.match_bits) == 48); - CLASSERT((int)sizeof(((struct lnet_hdr *)0)->msg.get.match_bits) == 8); - CLASSERT((int)offsetof(struct lnet_hdr, msg.get.ptl_index) == 56); - CLASSERT((int)sizeof(((struct lnet_hdr *)0)->msg.get.ptl_index) == 4); - CLASSERT((int)offsetof(struct lnet_hdr, msg.get.src_offset) == 60); - CLASSERT((int)sizeof(((struct lnet_hdr *)0)->msg.get.src_offset) == 4); - CLASSERT((int)offsetof(struct lnet_hdr, msg.get.sink_length) == 64); - CLASSERT((int)sizeof(((struct lnet_hdr *)0)->msg.get.sink_length) == 4); + BUILD_BUG_ON((int)offsetof(struct lnet_hdr, msg.get.return_wmd) != 32); + BUILD_BUG_ON((int)sizeof(((struct lnet_hdr *)0)->msg.get.return_wmd) != 16); + BUILD_BUG_ON((int)offsetof(struct lnet_hdr, msg.get.match_bits) != 48); + BUILD_BUG_ON((int)sizeof(((struct lnet_hdr *)0)->msg.get.match_bits) != 8); + BUILD_BUG_ON((int)offsetof(struct lnet_hdr, msg.get.ptl_index) != 56); + BUILD_BUG_ON((int)sizeof(((struct lnet_hdr *)0)->msg.get.ptl_index) != 4); + BUILD_BUG_ON((int)offsetof(struct lnet_hdr, msg.get.src_offset) != 60); + BUILD_BUG_ON((int)sizeof(((struct lnet_hdr *)0)->msg.get.src_offset) != 4); + BUILD_BUG_ON((int)offsetof(struct lnet_hdr, msg.get.sink_length) != 64); + BUILD_BUG_ON((int)sizeof(((struct lnet_hdr *)0)->msg.get.sink_length) != 4); /* Reply */ - CLASSERT((int)offsetof(struct lnet_hdr, msg.reply.dst_wmd) == 32); - CLASSERT((int)sizeof(((struct lnet_hdr *)0)->msg.reply.dst_wmd) == 16); + BUILD_BUG_ON((int)offsetof(struct lnet_hdr, msg.reply.dst_wmd) != 32); + BUILD_BUG_ON((int)sizeof(((struct lnet_hdr *)0)->msg.reply.dst_wmd) != 16); /* Hello */ - CLASSERT((int)offsetof(struct lnet_hdr, msg.hello.incarnation) == 32); - CLASSERT((int)sizeof(((struct lnet_hdr *)0)->msg.hello.incarnation) == 8); - CLASSERT((int)offsetof(struct lnet_hdr, msg.hello.type) == 40); - CLASSERT((int)sizeof(((struct lnet_hdr *)0)->msg.hello.type) == 4); + BUILD_BUG_ON((int)offsetof(struct lnet_hdr, msg.hello.incarnation) != 32); + BUILD_BUG_ON((int)sizeof(((struct lnet_hdr *)0)->msg.hello.incarnation) != 8); + BUILD_BUG_ON((int)offsetof(struct lnet_hdr, msg.hello.type) != 40); + BUILD_BUG_ON((int)sizeof(((struct lnet_hdr *)0)->msg.hello.type) != 4); /* Checks for struct lnet_ni_status and related constants */ - CLASSERT(LNET_NI_STATUS_INVALID == 0x00000000); - CLASSERT(LNET_NI_STATUS_UP == 0x15aac0de); - CLASSERT(LNET_NI_STATUS_DOWN == 0xdeadface); + BUILD_BUG_ON(LNET_NI_STATUS_INVALID != 0x00000000); + BUILD_BUG_ON(LNET_NI_STATUS_UP != 0x15aac0de); + BUILD_BUG_ON(LNET_NI_STATUS_DOWN != 0xdeadface); /* Checks for struct lnet_ni_status */ - CLASSERT((int)sizeof(struct lnet_ni_status) == 16); - CLASSERT((int)offsetof(struct lnet_ni_status, ns_nid) == 0); - CLASSERT((int)sizeof(((struct lnet_ni_status *)0)->ns_nid) == 8); - CLASSERT((int)offsetof(struct lnet_ni_status, ns_status) == 8); - CLASSERT((int)sizeof(((struct lnet_ni_status *)0)->ns_status) == 4); - CLASSERT((int)offsetof(struct lnet_ni_status, ns_unused) == 12); - CLASSERT((int)sizeof(((struct lnet_ni_status *)0)->ns_unused) == 4); + BUILD_BUG_ON((int)sizeof(struct lnet_ni_status) != 16); + BUILD_BUG_ON((int)offsetof(struct lnet_ni_status, ns_nid) != 0); + BUILD_BUG_ON((int)sizeof(((struct lnet_ni_status *)0)->ns_nid) != 8); + BUILD_BUG_ON((int)offsetof(struct lnet_ni_status, ns_status) != 8); + BUILD_BUG_ON((int)sizeof(((struct lnet_ni_status *)0)->ns_status) != 4); + BUILD_BUG_ON((int)offsetof(struct lnet_ni_status, ns_unused) != 12); + BUILD_BUG_ON((int)sizeof(((struct lnet_ni_status *)0)->ns_unused) != 4); /* Checks for struct lnet_ping_info and related constants */ - CLASSERT(LNET_PROTO_PING_MAGIC == 0x70696E67); - CLASSERT(LNET_PING_FEAT_INVAL == 0); - CLASSERT(LNET_PING_FEAT_BASE == 1); - CLASSERT(LNET_PING_FEAT_NI_STATUS == 2); - CLASSERT(LNET_PING_FEAT_RTE_DISABLED == 4); - CLASSERT(LNET_PING_FEAT_MULTI_RAIL == 8); - CLASSERT(LNET_PING_FEAT_DISCOVERY == 16); - CLASSERT(LNET_PING_FEAT_BITS == 31); + BUILD_BUG_ON(LNET_PROTO_PING_MAGIC != 0x70696E67); + BUILD_BUG_ON(LNET_PING_FEAT_INVAL != 0); + BUILD_BUG_ON(LNET_PING_FEAT_BASE != 1); + BUILD_BUG_ON(LNET_PING_FEAT_NI_STATUS != 2); + BUILD_BUG_ON(LNET_PING_FEAT_RTE_DISABLED != 4); + BUILD_BUG_ON(LNET_PING_FEAT_MULTI_RAIL != 8); + BUILD_BUG_ON(LNET_PING_FEAT_DISCOVERY != 16); + BUILD_BUG_ON(LNET_PING_FEAT_BITS != 31); /* Checks for struct lnet_ping_info */ - CLASSERT((int)sizeof(struct lnet_ping_info) == 16); - CLASSERT((int)offsetof(struct lnet_ping_info, pi_magic) == 0); - CLASSERT((int)sizeof(((struct lnet_ping_info *)0)->pi_magic) == 4); - CLASSERT((int)offsetof(struct lnet_ping_info, pi_features) == 4); - CLASSERT((int)sizeof(((struct lnet_ping_info *)0)->pi_features) == 4); - CLASSERT((int)offsetof(struct lnet_ping_info, pi_pid) == 8); - CLASSERT((int)sizeof(((struct lnet_ping_info *)0)->pi_pid) == 4); - CLASSERT((int)offsetof(struct lnet_ping_info, pi_nnis) == 12); - CLASSERT((int)sizeof(((struct lnet_ping_info *)0)->pi_nnis) == 4); - CLASSERT((int)offsetof(struct lnet_ping_info, pi_ni) == 16); - CLASSERT((int)sizeof(((struct lnet_ping_info *)0)->pi_ni) == 0); + BUILD_BUG_ON((int)sizeof(struct lnet_ping_info) != 16); + BUILD_BUG_ON((int)offsetof(struct lnet_ping_info, pi_magic) != 0); + BUILD_BUG_ON((int)sizeof(((struct lnet_ping_info *)0)->pi_magic) != 4); + BUILD_BUG_ON((int)offsetof(struct lnet_ping_info, pi_features) != 4); + BUILD_BUG_ON((int)sizeof(((struct lnet_ping_info *)0)->pi_features) != 4); + BUILD_BUG_ON((int)offsetof(struct lnet_ping_info, pi_pid) != 8); + BUILD_BUG_ON((int)sizeof(((struct lnet_ping_info *)0)->pi_pid) != 4); + BUILD_BUG_ON((int)offsetof(struct lnet_ping_info, pi_nnis) != 12); + BUILD_BUG_ON((int)sizeof(((struct lnet_ping_info *)0)->pi_nnis) != 4); + BUILD_BUG_ON((int)offsetof(struct lnet_ping_info, pi_ni) != 16); + BUILD_BUG_ON((int)sizeof(((struct lnet_ping_info *)0)->pi_ni) != 0); + + /* Acceptor connection request */ + BUILD_BUG_ON(LNET_PROTO_ACCEPTOR_VERSION != 1); + + /* Checks for struct lnet_acceptor_connreq */ + BUILD_BUG_ON((int)sizeof(struct lnet_acceptor_connreq) != 16); + BUILD_BUG_ON((int)offsetof(struct lnet_acceptor_connreq, acr_magic) != 0); + BUILD_BUG_ON((int)sizeof(((struct lnet_acceptor_connreq *)0)->acr_magic) != 4); + BUILD_BUG_ON((int)offsetof(struct lnet_acceptor_connreq, acr_version) != 4); + BUILD_BUG_ON((int)sizeof(((struct lnet_acceptor_connreq *)0)->acr_version) != 4); + BUILD_BUG_ON((int)offsetof(struct lnet_acceptor_connreq, acr_nid) != 8); + BUILD_BUG_ON((int)sizeof(((struct lnet_acceptor_connreq *)0)->acr_nid) != 8); + + /* Checks for struct lnet_counters_common */ + BUILD_BUG_ON((int)sizeof(struct lnet_counters_common) != 60); + BUILD_BUG_ON((int)offsetof(struct lnet_counters_common, lcc_msgs_alloc) != 0); + BUILD_BUG_ON((int)sizeof(((struct lnet_counters_common *)0)->lcc_msgs_alloc) != 4); + BUILD_BUG_ON((int)offsetof(struct lnet_counters_common, lcc_msgs_max) != 4); + BUILD_BUG_ON((int)sizeof(((struct lnet_counters_common *)0)->lcc_msgs_max) != 4); + BUILD_BUG_ON((int)offsetof(struct lnet_counters_common, lcc_errors) != 8); + BUILD_BUG_ON((int)sizeof(((struct lnet_counters_common *)0)->lcc_errors) != 4); + BUILD_BUG_ON((int)offsetof(struct lnet_counters_common, lcc_send_count) != 12); + BUILD_BUG_ON((int)sizeof(((struct lnet_counters_common *)0)->lcc_send_count) != 4); + BUILD_BUG_ON((int)offsetof(struct lnet_counters_common, lcc_recv_count) != 16); + BUILD_BUG_ON((int)sizeof(((struct lnet_counters_common *)0)->lcc_recv_count) != 4); + BUILD_BUG_ON((int)offsetof(struct lnet_counters_common, lcc_route_count) != 20); + BUILD_BUG_ON((int)sizeof(((struct lnet_counters_common *)0)->lcc_route_count) != 4); + BUILD_BUG_ON((int)offsetof(struct lnet_counters_common, lcc_drop_count) != 24); + BUILD_BUG_ON((int)sizeof(((struct lnet_counters_common *)0)->lcc_drop_count) != 4); + BUILD_BUG_ON((int)offsetof(struct lnet_counters_common, lcc_send_length) != 28); + BUILD_BUG_ON((int)sizeof(((struct lnet_counters_common *)0)->lcc_send_length) != 8); + BUILD_BUG_ON((int)offsetof(struct lnet_counters_common, lcc_recv_length) != 36); + BUILD_BUG_ON((int)sizeof(((struct lnet_counters_common *)0)->lcc_recv_length) != 8); + BUILD_BUG_ON((int)offsetof(struct lnet_counters_common, lcc_route_length) != 44); + BUILD_BUG_ON((int)sizeof(((struct lnet_counters_common *)0)->lcc_route_length) != 8); + BUILD_BUG_ON((int)offsetof(struct lnet_counters_common, lcc_drop_length) != 52); + BUILD_BUG_ON((int)sizeof(((struct lnet_counters_common *)0)->lcc_drop_length) != 8); } -static struct lnet_lnd *lnet_find_lnd_by_type(__u32 type) +static const struct lnet_lnd *lnet_find_lnd_by_type(__u32 type) { - struct lnet_lnd *lnd; - struct list_head *tmp; + const struct lnet_lnd *lnd; /* holding lnd mutex */ - list_for_each(tmp, &the_lnet.ln_lnds) { - lnd = list_entry(tmp, struct lnet_lnd, lnd_list); + if (type >= NUM_LNDS) + return NULL; + lnd = the_lnet.ln_lnds[type]; + LASSERT(!lnd || lnd->lnd_type == type); - if (lnd->lnd_type == type) - return lnd; - } - return NULL; + return lnd; +} + +unsigned int +lnet_get_lnd_timeout(void) +{ + return lnet_lnd_timeout; } +EXPORT_SYMBOL(lnet_get_lnd_timeout); void -lnet_register_lnd(struct lnet_lnd *lnd) +lnet_register_lnd(const struct lnet_lnd *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; + the_lnet.ln_lnds[lnd->lnd_type] = lnd; CDEBUG(D_NET, "%s LND registered\n", libcfs_lnd2str(lnd->lnd_type)); @@ -528,45 +962,99 @@ lnet_register_lnd(struct lnet_lnd *lnd) EXPORT_SYMBOL(lnet_register_lnd); void -lnet_unregister_lnd(struct lnet_lnd *lnd) +lnet_unregister_lnd(const struct lnet_lnd *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); + the_lnet.ln_lnds[lnd->lnd_type] = NULL; CDEBUG(D_NET, "%s LND unregistered\n", libcfs_lnd2str(lnd->lnd_type)); mutex_unlock(&the_lnet.ln_lnd_mutex); } EXPORT_SYMBOL(lnet_unregister_lnd); +static void +lnet_counters_get_common_locked(struct lnet_counters_common *common) +{ + struct lnet_counters *ctr; + int i; + + /* FIXME !!! Their is no assert_lnet_net_locked() to ensure this + * actually called under the protection of the lnet_net_lock. + */ + memset(common, 0, sizeof(*common)); + + cfs_percpt_for_each(ctr, i, the_lnet.ln_counters) { + common->lcc_msgs_max += ctr->lct_common.lcc_msgs_max; + common->lcc_msgs_alloc += ctr->lct_common.lcc_msgs_alloc; + common->lcc_errors += ctr->lct_common.lcc_errors; + common->lcc_send_count += ctr->lct_common.lcc_send_count; + common->lcc_recv_count += ctr->lct_common.lcc_recv_count; + common->lcc_route_count += ctr->lct_common.lcc_route_count; + common->lcc_drop_count += ctr->lct_common.lcc_drop_count; + common->lcc_send_length += ctr->lct_common.lcc_send_length; + common->lcc_recv_length += ctr->lct_common.lcc_recv_length; + common->lcc_route_length += ctr->lct_common.lcc_route_length; + common->lcc_drop_length += ctr->lct_common.lcc_drop_length; + } +} + void +lnet_counters_get_common(struct lnet_counters_common *common) +{ + lnet_net_lock(LNET_LOCK_EX); + lnet_counters_get_common_locked(common); + lnet_net_unlock(LNET_LOCK_EX); +} +EXPORT_SYMBOL(lnet_counters_get_common); + +int lnet_counters_get(struct lnet_counters *counters) { struct lnet_counters *ctr; - int i; + struct lnet_counters_health *health = &counters->lct_health; + int i, rc = 0; 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; + if (the_lnet.ln_state != LNET_STATE_RUNNING) + GOTO(out_unlock, rc = -ENODEV); - } + lnet_counters_get_common_locked(&counters->lct_common); + + cfs_percpt_for_each(ctr, i, the_lnet.ln_counters) { + health->lch_rst_alloc += ctr->lct_health.lch_rst_alloc; + health->lch_resend_count += ctr->lct_health.lch_resend_count; + health->lch_response_timeout_count += + ctr->lct_health.lch_response_timeout_count; + health->lch_local_interrupt_count += + ctr->lct_health.lch_local_interrupt_count; + health->lch_local_dropped_count += + ctr->lct_health.lch_local_dropped_count; + health->lch_local_aborted_count += + ctr->lct_health.lch_local_aborted_count; + health->lch_local_no_route_count += + ctr->lct_health.lch_local_no_route_count; + health->lch_local_timeout_count += + ctr->lct_health.lch_local_timeout_count; + health->lch_local_error_count += + ctr->lct_health.lch_local_error_count; + health->lch_remote_dropped_count += + ctr->lct_health.lch_remote_dropped_count; + health->lch_remote_error_count += + ctr->lct_health.lch_remote_error_count; + health->lch_remote_timeout_count += + ctr->lct_health.lch_remote_timeout_count; + health->lch_network_timeout_count += + ctr->lct_health.lch_network_timeout_count; + } +out_unlock: lnet_net_unlock(LNET_LOCK_EX); + return rc; } EXPORT_SYMBOL(lnet_counters_get); @@ -578,9 +1066,12 @@ lnet_counters_reset(void) lnet_net_lock(LNET_LOCK_EX); + if (the_lnet.ln_state != LNET_STATE_RUNNING) + goto avoid_reset; + cfs_percpt_for_each(counters, i, the_lnet.ln_counters) memset(counters, 0, sizeof(struct lnet_counters)); - +avoid_reset: lnet_net_unlock(LNET_LOCK_EX); } @@ -611,10 +1102,7 @@ lnet_res_container_cleanup(struct lnet_res_container *rec) 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, struct lnet_eq, eq_list)); - - } else if (rec->rec_type == LNET_COOKIE_TYPE_MD) { + if (rec->rec_type == LNET_COOKIE_TYPE_MD) { lnet_md_free(list_entry(e, struct lnet_libmd, md_list)); } else { /* NB: Active MEs should be attached on portals */ @@ -632,8 +1120,7 @@ lnet_res_container_cleanup(struct lnet_res_container *rec) } if (rec->rec_lh_hash != NULL) { - LIBCFS_FREE(rec->rec_lh_hash, - LNET_LH_HASH_SIZE * sizeof(rec->rec_lh_hash[0])); + CFS_FREE_PTR_ARRAY(rec->rec_lh_hash, LNET_LH_HASH_SIZE); rec->rec_lh_hash = NULL; } @@ -749,6 +1236,26 @@ lnet_res_lh_initialize(struct lnet_res_container *rec, list_add(&lh->lh_hash_chain, &rec->rec_lh_hash[hash]); } +struct list_head ** +lnet_create_array_of_queues(void) +{ + struct list_head **qs; + struct list_head *q; + int i; + + qs = cfs_percpt_alloc(lnet_cpt_table(), + sizeof(struct list_head)); + if (!qs) { + CERROR("Failed to allocate queues\n"); + return NULL; + } + + cfs_percpt_for_each(q, i, qs) + INIT_LIST_HEAD(q); + + return qs; +} + static int lnet_unprepare(void); static int @@ -779,9 +1286,14 @@ lnet_prepare(lnet_pid_t requested_pid) INIT_LIST_HEAD(&the_lnet.ln_dc_request); INIT_LIST_HEAD(&the_lnet.ln_dc_working); INIT_LIST_HEAD(&the_lnet.ln_dc_expired); + INIT_LIST_HEAD(&the_lnet.ln_mt_localNIRecovq); + INIT_LIST_HEAD(&the_lnet.ln_mt_peerNIRecovq); + INIT_LIST_HEAD(&the_lnet.ln_udsp_list); init_waitqueue_head(&the_lnet.ln_dc_waitq); + the_lnet.ln_mt_handler = NULL; + init_completion(&the_lnet.ln_started); - rc = lnet_descriptor_setup(); + rc = lnet_slab_setup(); if (rc != 0) goto failed; @@ -816,14 +1328,6 @@ lnet_prepare(lnet_pid_t requested_pid) 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; @@ -838,6 +1342,12 @@ lnet_prepare(lnet_pid_t requested_pid) goto failed; } + the_lnet.ln_mt_zombie_rstqs = lnet_create_array_of_queues(); + if (!the_lnet.ln_mt_zombie_rstqs) { + rc = -ENOMEM; + goto failed; + } + return 0; failed: @@ -859,6 +1369,14 @@ lnet_unprepare (void) LASSERT(list_empty(&the_lnet.ln_test_peers)); LASSERT(list_empty(&the_lnet.ln_nets)); + if (the_lnet.ln_mt_zombie_rstqs) { + lnet_clean_zombie_rstqs(); + the_lnet.ln_mt_zombie_rstqs = NULL; + } + + lnet_assert_handler_unused(the_lnet.ln_mt_handler); + the_lnet.ln_mt_handler = NULL; + lnet_portals_destroy(); if (the_lnet.ln_md_containers != NULL) { @@ -866,11 +1384,6 @@ lnet_unprepare (void) 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(); @@ -882,7 +1395,8 @@ lnet_unprepare (void) the_lnet.ln_counters = NULL; } lnet_destroy_remote_nets_table(); - lnet_descriptor_cleanup(); + lnet_udsp_destroy(true); + lnet_slab_cleanup(); return 0; } @@ -934,6 +1448,81 @@ lnet_get_net_locked(__u32 net_id) return NULL; } +void +lnet_net_clr_pref_rtrs(struct lnet_net *net) +{ + struct list_head zombies; + struct lnet_nid_list *ne; + struct lnet_nid_list *tmp; + + INIT_LIST_HEAD(&zombies); + + lnet_net_lock(LNET_LOCK_EX); + list_splice_init(&net->net_rtr_pref_nids, &zombies); + lnet_net_unlock(LNET_LOCK_EX); + + list_for_each_entry_safe(ne, tmp, &zombies, nl_list) { + list_del_init(&ne->nl_list); + LIBCFS_FREE(ne, sizeof(*ne)); + } +} + +int +lnet_net_add_pref_rtr(struct lnet_net *net, + lnet_nid_t gw_nid) +__must_hold(&the_lnet.ln_api_mutex) +{ + struct lnet_nid_list *ne; + + /* This function is called with api_mutex held. When the api_mutex + * is held the list can not be modified, as it is only modified as + * a result of applying a UDSP and that happens under api_mutex + * lock. + */ + list_for_each_entry(ne, &net->net_rtr_pref_nids, nl_list) { + if (ne->nl_nid == gw_nid) + return -EEXIST; + } + + LIBCFS_ALLOC(ne, sizeof(*ne)); + if (!ne) + return -ENOMEM; + + ne->nl_nid = gw_nid; + + /* Lock the cpt to protect against addition and checks in the + * selection algorithm + */ + lnet_net_lock(LNET_LOCK_EX); + list_add(&ne->nl_list, &net->net_rtr_pref_nids); + lnet_net_unlock(LNET_LOCK_EX); + + return 0; +} + +bool +lnet_net_is_pref_rtr_locked(struct lnet_net *net, lnet_nid_t rtr_nid) +{ + struct lnet_nid_list *ne; + + CDEBUG(D_NET, "%s: rtr pref emtpy: %d\n", + libcfs_net2str(net->net_id), + list_empty(&net->net_rtr_pref_nids)); + + if (list_empty(&net->net_rtr_pref_nids)) + return false; + + list_for_each_entry(ne, &net->net_rtr_pref_nids, nl_list) { + CDEBUG(D_NET, "Comparing pref %s with gw %s\n", + libcfs_nid2str(ne->nl_nid), + libcfs_nid2str(rtr_nid)); + if (rtr_nid == ne->nl_nid) + return true; + } + + return false; +} + unsigned int lnet_nid_cpt_hash(lnet_nid_t nid, unsigned int number) { @@ -1007,31 +1596,31 @@ lnet_cpt_of_nid(lnet_nid_t nid, struct lnet_ni *ni) EXPORT_SYMBOL(lnet_cpt_of_nid); int -lnet_islocalnet(__u32 net_id) +lnet_islocalnet_locked(__u32 net_id) { struct lnet_net *net; - int cpt; - bool local; - - cpt = lnet_net_lock_current(); + bool local; 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) +int +lnet_islocalnet(__u32 net_id) { - if (ni->ni_state == LNET_NI_STATE_ACTIVE || - ni->ni_state == LNET_NI_STATE_DEGRADED) - return true; + int cpt; + bool local; - return false; + cpt = lnet_net_lock_current(); + + local = lnet_islocalnet_locked(net_id); + + lnet_net_unlock(cpt); + + return local; } struct lnet_ni * @@ -1109,6 +1698,7 @@ lnet_ping_buffer_alloc(int nnis, gfp_t gfp) LIBCFS_ALLOC_GFP(pbuf, LNET_PING_BUFFER_SIZE(nnis), gfp); if (pbuf) { pbuf->pb_nnis = nnis; + pbuf->pb_needs_post = false; atomic_set(&pbuf->pb_refcnt, 1); } @@ -1118,7 +1708,7 @@ lnet_ping_buffer_alloc(int nnis, gfp_t gfp) void lnet_ping_buffer_free(struct lnet_ping_buffer *pbuf) { - LASSERT(lnet_ping_buffer_numref(pbuf) == 0); + LASSERT(atomic_read(&pbuf->pb_refcnt) == 0); LIBCFS_FREE(pbuf, LNET_PING_BUFFER_SIZE(pbuf->pb_nnis)); } @@ -1186,6 +1776,44 @@ lnet_get_ni_count(void) } int +lnet_get_net_count(void) +{ + struct lnet_net *net; + int count = 0; + + lnet_net_lock(0); + + list_for_each_entry(net, &the_lnet.ln_nets, net_list) { + count++; + } + + lnet_net_unlock(0); + + return count; +} + +void +lnet_swap_pinginfo(struct lnet_ping_buffer *pbuf) +{ + struct lnet_ni_status *stat; + int nnis; + int i; + + __swab32s(&pbuf->pb_info.pi_magic); + __swab32s(&pbuf->pb_info.pi_features); + __swab32s(&pbuf->pb_info.pi_pid); + __swab32s(&pbuf->pb_info.pi_nnis); + nnis = pbuf->pb_info.pi_nnis; + if (nnis > pbuf->pb_nnis) + nnis = pbuf->pb_nnis; + for (i = 0; i < nnis; i++) { + stat = &pbuf->pb_info.pi_ni[i]; + __swab64s(&stat->ns_nid); + __swab32s(&stat->ns_status); + } +} + +int lnet_ping_info_validate(struct lnet_ping_info *pinfo) { if (!pinfo) @@ -1197,7 +1825,7 @@ lnet_ping_info_validate(struct lnet_ping_info *pinfo) /* Loopback is guaranteed to be present */ if (pinfo->pi_nnis < 1 || pinfo->pi_nnis > lnet_interfaces_max) return -ERANGE; - if (LNET_NETTYP(LNET_NIDNET(LNET_PING_INFO_LONI(pinfo))) != LOLND) + if (LNET_PING_INFO_LONI(pinfo) != LNET_NID_LO_0) return -EPROTO; return 0; } @@ -1227,7 +1855,7 @@ lnet_ping_target_destroy(void) static void lnet_ping_target_event_handler(struct lnet_event *event) { - struct lnet_ping_buffer *pbuf = event->md.user_ptr; + struct lnet_ping_buffer *pbuf = event->md_user_ptr; if (event->unlinked) lnet_ping_buffer_decref(pbuf); @@ -1242,18 +1870,13 @@ lnet_ping_target_setup(struct lnet_ping_buffer **ppbuf, .nid = LNET_NID_ANY, .pid = LNET_PID_ANY }; - struct lnet_handle_me me_handle; + struct lnet_me *me; struct lnet_md md = { NULL }; - int rc, rc2; + int rc; - if (set_eq) { - rc = LNetEQAlloc(0, lnet_ping_target_event_handler, - &the_lnet.ln_ping_target_eq); - if (rc != 0) { - CERROR("Can't allocate ping buffer EQ: %d\n", rc); - return rc; - } - } + if (set_eq) + the_lnet.ln_ping_target_handler = + lnet_ping_target_event_handler; *ppbuf = lnet_ping_target_create(ni_count); if (*ppbuf == NULL) { @@ -1262,11 +1885,11 @@ lnet_ping_target_setup(struct lnet_ping_buffer **ppbuf, } /* Ping target ME/MD */ - rc = LNetMEAttach(LNET_RESERVED_PORTAL, id, + me = LNetMEAttach(LNET_RESERVED_PORTAL, id, LNET_PROTO_PING_MATCHBITS, 0, - LNET_UNLINK, LNET_INS_AFTER, - &me_handle); - if (rc != 0) { + LNET_UNLINK, LNET_INS_AFTER); + if (IS_ERR(me)) { + rc = PTR_ERR(me); CERROR("Can't create ping target ME: %d\n", rc); goto fail_decref_ping_buffer; } @@ -1278,30 +1901,23 @@ lnet_ping_target_setup(struct lnet_ping_buffer **ppbuf, md.max_size = 0; md.options = LNET_MD_OP_GET | LNET_MD_TRUNCATE | LNET_MD_MANAGE_REMOTE; - md.eq_handle = the_lnet.ln_ping_target_eq; + md.handler = the_lnet.ln_ping_target_handler; md.user_ptr = *ppbuf; - rc = LNetMDAttach(me_handle, md, LNET_RETAIN, ping_mdh); + rc = LNetMDAttach(me, &md, LNET_RETAIN, ping_mdh); if (rc != 0) { CERROR("Can't attach ping target MD: %d\n", rc); - goto fail_unlink_ping_me; + goto fail_decref_ping_buffer; } lnet_ping_buffer_addref(*ppbuf); return 0; -fail_unlink_ping_me: - rc2 = LNetMEUnlink(me_handle); - LASSERT(rc2 == 0); fail_decref_ping_buffer: - LASSERT(lnet_ping_buffer_numref(*ppbuf) == 1); + LASSERT(atomic_read(&(*ppbuf)->pb_refcnt) == 1); lnet_ping_buffer_decref(*ppbuf); *ppbuf = NULL; fail_free_eq: - if (set_eq) { - rc2 = LNetEQFree(the_lnet.ln_ping_target_eq); - LASSERT(rc2 == 0); - } return rc; } @@ -1309,19 +1925,13 @@ static void lnet_ping_md_unlink(struct lnet_ping_buffer *pbuf, struct lnet_handle_md *ping_mdh) { - sigset_t blocked = cfs_block_allsigs(); - LNetMDUnlink(*ping_mdh); LNetInvalidateMDHandle(ping_mdh); /* NB the MD could be busy; this just starts the unlink */ - while (lnet_ping_buffer_numref(pbuf) > 1) { - CDEBUG(D_NET, "Still waiting for ping data MD to unlink\n"); - set_current_state(TASK_UNINTERRUPTIBLE); - schedule_timeout(cfs_time_seconds(1)); - } - - cfs_restore_sigs(blocked); + wait_var_event_warning(&pbuf->pb_refcnt, + atomic_read(&pbuf->pb_refcnt) <= 1, + "Still waiting for ping data MD to unlink\n"); } static void @@ -1408,66 +2018,46 @@ lnet_ping_target_update(struct lnet_ping_buffer *pbuf, static void lnet_ping_target_fini(void) { - int rc; - lnet_ping_md_unlink(the_lnet.ln_ping_target, &the_lnet.ln_ping_target_md); - rc = LNetEQFree(the_lnet.ln_ping_target_eq); - LASSERT(rc == 0); - + lnet_assert_handler_unused(the_lnet.ln_ping_target_handler); lnet_ping_target_destroy(); } /* Resize the push target. */ int lnet_push_target_resize(void) { - struct lnet_process_id id = { LNET_NID_ANY, LNET_PID_ANY }; - struct lnet_md md = { NULL }; - struct lnet_handle_me meh; struct lnet_handle_md mdh; struct lnet_handle_md old_mdh; struct lnet_ping_buffer *pbuf; struct lnet_ping_buffer *old_pbuf; - int nnis = the_lnet.ln_push_target_nnis; + int nnis; int rc; +again: + nnis = the_lnet.ln_push_target_nnis; if (nnis <= 0) { - rc = -EINVAL; - goto fail_return; + CDEBUG(D_NET, "Invalid nnis %d\n", nnis); + return -EINVAL; } -again: + + /* NB: lnet_ping_buffer_alloc() sets pbuf refcount to 1. That ref is + * dropped when we need to resize again (see "old_pbuf" below) or when + * LNet is shutdown (see lnet_push_target_fini()) + */ pbuf = lnet_ping_buffer_alloc(nnis, GFP_NOFS); if (!pbuf) { - rc = -ENOMEM; - goto fail_return; - } - - rc = LNetMEAttach(LNET_RESERVED_PORTAL, id, - LNET_PROTO_PING_MATCHBITS, 0, - LNET_UNLINK, LNET_INS_AFTER, - &meh); - if (rc) { - CERROR("Can't create push target ME: %d\n", rc); - goto fail_decref_pbuf; + CDEBUG(D_NET, "Can't allocate pbuf for nnis %d\n", nnis); + return -ENOMEM; } - /* initialize md content */ - md.start = &pbuf->pb_info; - md.length = LNET_PING_INFO_SIZE(nnis); - md.threshold = LNET_MD_THRESH_INF; - md.max_size = 0; - md.options = LNET_MD_OP_PUT | LNET_MD_TRUNCATE | - LNET_MD_MANAGE_REMOTE; - md.user_ptr = pbuf; - md.eq_handle = the_lnet.ln_push_target_eq; - - rc = LNetMDAttach(meh, md, LNET_RETAIN, &mdh); + rc = lnet_push_target_post(pbuf, &mdh); if (rc) { - CERROR("Can't attach push MD: %d\n", rc); - goto fail_unlink_meh; + CDEBUG(D_NET, "Failed to post push target: %d\n", rc); + lnet_ping_buffer_decref(pbuf); + return rc; } - lnet_ping_buffer_addref(pbuf); lnet_net_lock(LNET_LOCK_EX); old_pbuf = the_lnet.ln_push_target; @@ -1478,34 +2068,81 @@ again: if (old_pbuf) { LNetMDUnlink(old_mdh); + /* Drop ref set by lnet_ping_buffer_alloc() */ lnet_ping_buffer_decref(old_pbuf); } + /* Received another push or reply that requires a larger buffer */ if (nnis < the_lnet.ln_push_target_nnis) goto again; CDEBUG(D_NET, "nnis %d success\n", nnis); - return 0; +} -fail_unlink_meh: - LNetMEUnlink(meh); -fail_decref_pbuf: - lnet_ping_buffer_decref(pbuf); -fail_return: - CDEBUG(D_NET, "nnis %d error %d\n", nnis, rc); - return rc; +int lnet_push_target_post(struct lnet_ping_buffer *pbuf, + struct lnet_handle_md *mdhp) +{ + struct lnet_process_id id = { LNET_NID_ANY, LNET_PID_ANY }; + struct lnet_md md = { NULL }; + struct lnet_me *me; + int rc; + + me = LNetMEAttach(LNET_RESERVED_PORTAL, id, + LNET_PROTO_PING_MATCHBITS, 0, + LNET_UNLINK, LNET_INS_AFTER); + if (IS_ERR(me)) { + rc = PTR_ERR(me); + CERROR("Can't create push target ME: %d\n", rc); + return rc; + } + + pbuf->pb_needs_post = false; + + /* This reference is dropped by lnet_push_target_event_handler() */ + lnet_ping_buffer_addref(pbuf); + + /* initialize md content */ + md.start = &pbuf->pb_info; + md.length = LNET_PING_INFO_SIZE(pbuf->pb_nnis); + md.threshold = 1; + md.max_size = 0; + md.options = LNET_MD_OP_PUT | LNET_MD_TRUNCATE; + md.user_ptr = pbuf; + md.handler = the_lnet.ln_push_target_handler; + + rc = LNetMDAttach(me, &md, LNET_UNLINK, mdhp); + if (rc) { + CERROR("Can't attach push MD: %d\n", rc); + lnet_ping_buffer_decref(pbuf); + pbuf->pb_needs_post = true; + return rc; + } + + CDEBUG(D_NET, "posted push target %p\n", pbuf); + + return 0; } static void lnet_push_target_event_handler(struct lnet_event *ev) { - struct lnet_ping_buffer *pbuf = ev->md.user_ptr; + struct lnet_ping_buffer *pbuf = ev->md_user_ptr; + + CDEBUG(D_NET, "type %d status %d unlinked %d\n", ev->type, ev->status, + ev->unlinked); if (pbuf->pb_info.pi_magic == __swab32(LNET_PROTO_PING_MAGIC)) lnet_swap_pinginfo(pbuf); + if (ev->type == LNET_EVENT_UNLINK) { + /* Drop ref added by lnet_push_target_post() */ + lnet_ping_buffer_decref(pbuf); + return; + } + lnet_peer_push_event(ev); if (ev->unlinked) + /* Drop ref added by lnet_push_target_post */ lnet_ping_buffer_decref(pbuf); } @@ -1517,12 +2154,11 @@ static int lnet_push_target_init(void) if (the_lnet.ln_push_target) return -EALREADY; - rc = LNetEQAlloc(0, lnet_push_target_event_handler, - &the_lnet.ln_push_target_eq); - if (rc) { - CERROR("Can't allocated push target EQ: %d\n", rc); - return rc; - } + the_lnet.ln_push_target_handler = + lnet_push_target_event_handler; + + rc = LNetSetLazyPortal(LNET_RESERVED_PORTAL); + LASSERT(rc == 0); /* Start at the required minimum, we'll enlarge if required. */ the_lnet.ln_push_target_nnis = LNET_INTERFACES_MIN; @@ -1530,8 +2166,8 @@ static int lnet_push_target_init(void) rc = lnet_push_target_resize(); if (rc) { - LNetEQFree(the_lnet.ln_push_target_eq); - LNetInvalidateEQHandle(&the_lnet.ln_push_target_eq); + LNetClearLazyPortal(LNET_RESERVED_PORTAL); + the_lnet.ln_push_target_handler = NULL; } return rc; @@ -1548,18 +2184,18 @@ static void lnet_push_target_fini(void) LNetInvalidateMDHandle(&the_lnet.ln_push_target_md); /* Wait for the unlink to complete. */ - while (lnet_ping_buffer_numref(the_lnet.ln_push_target) > 1) { - CDEBUG(D_NET, "Still waiting for ping data MD to unlink\n"); - set_current_state(TASK_UNINTERRUPTIBLE); - schedule_timeout(cfs_time_seconds(1)); - } + wait_var_event_warning(&the_lnet.ln_push_target->pb_refcnt, + atomic_read(&the_lnet.ln_push_target->pb_refcnt) <= 1, + "Still waiting for ping data MD to unlink\n"); + /* Drop ref set by lnet_ping_buffer_alloc() */ lnet_ping_buffer_decref(the_lnet.ln_push_target); the_lnet.ln_push_target = NULL; the_lnet.ln_push_target_nnis = 0; - LNetEQFree(the_lnet.ln_push_target_eq); - LNetInvalidateEQHandle(&the_lnet.ln_push_target_eq); + LNetClearLazyPortal(LNET_RESERVED_PORTAL); + lnet_assert_handler_unused(the_lnet.ln_push_target_handler); + the_lnet.ln_push_target_handler = NULL; } static int @@ -1582,11 +2218,6 @@ lnet_ni_tq_credits(struct lnet_ni *ni) static void lnet_ni_unlink_locked(struct lnet_ni *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); @@ -1625,15 +2256,22 @@ lnet_clear_zombies_nis_locked(struct lnet_net *net) } if (!list_empty(&ni->ni_netlist)) { + /* Unlock mutex while waiting to allow other + * threads to read the LNet state and fall through + * to avoid deadlock + */ lnet_net_unlock(LNET_LOCK_EX); + mutex_unlock(&the_lnet.ln_api_mutex); + ++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)); + schedule_timeout_uninterruptible(cfs_time_seconds(1)); + + mutex_lock(&the_lnet.ln_api_mutex); lnet_net_lock(LNET_LOCK_EX); continue; } @@ -1643,7 +2281,14 @@ lnet_clear_zombies_nis_locked(struct lnet_net *net) islo = ni->ni_net->net_lnd->lnd_type == LOLND; LASSERT(!in_interrupt()); + /* Holding the mutex makes it safe for lnd_shutdown + * to call module_put(). Module unload cannot finish + * until lnet_unregister_lnd() completes, and that + * requires the mutex. + */ + mutex_lock(&the_lnet.ln_lnd_mutex); (net->net_lnd->lnd_shutdown)(ni); + mutex_unlock(&the_lnet.ln_lnd_mutex); if (!islo) CDEBUG(D_LNI, "Removed LNI %s\n", @@ -1663,7 +2308,9 @@ lnet_shutdown_lndni(struct lnet_ni *ni) struct lnet_net *net = ni->ni_net; lnet_net_lock(LNET_LOCK_EX); + lnet_ni_lock(ni); ni->ni_state = LNET_NI_STATE_DELETING; + lnet_ni_unlock(ni); lnet_ni_unlink_locked(ni); lnet_incr_dlc_seq(); lnet_net_unlock(LNET_LOCK_EX); @@ -1684,8 +2331,6 @@ lnet_shutdown_lndnet(struct lnet_net *net) 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)) { @@ -1701,15 +2346,6 @@ lnet_shutdown_lndnet(struct lnet_net *net) /* Do peer table cleanup for this net */ lnet_peer_tables_cleanup(net); - 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); } @@ -1717,11 +2353,9 @@ static void lnet_shutdown_lndnets(void) { struct lnet_net *net; - struct list_head resend; + LIST_HEAD(resend); struct lnet_msg *msg, *tmp; - INIT_LIST_HEAD(&resend); - /* NB called holding the global mutex */ /* All quiet on the API front */ @@ -1731,16 +2365,12 @@ lnet_shutdown_lndnets(void) lnet_net_lock(LNET_LOCK_EX); the_lnet.ln_state = LNET_STATE_STOPPING; - 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); - } + /* + * 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 + */ + list_splice_init(&the_lnet.ln_nets, &the_lnet.ln_net_zombie); /* Drop the cached loopback Net. */ if (the_lnet.ln_loni != NULL) { @@ -1762,6 +2392,7 @@ lnet_shutdown_lndnets(void) list_for_each_entry_safe(msg, tmp, &resend, msg_list) { list_del_init(&msg->msg_list); + msg->msg_no_resend = true; lnet_finalize(msg, -ECANCELED); } @@ -1792,13 +2423,12 @@ lnet_startup_lndni(struct lnet_ni *ni, struct lnet_lnd_tunables *tun) 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; } + lnet_ni_lock(ni); ni->ni_state = LNET_NI_STATE_ACTIVE; + lnet_ni_unlock(ni); /* We keep a reference on the loopback net through the loopback NI */ if (net->net_lnd->lnd_type == LOLND) { @@ -1852,21 +2482,19 @@ 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; - struct lnet_lnd *lnd; - int peer_timeout = + struct lnet_net *net_l = NULL; + LIST_HEAD(local_ni_list); + int rc; + int ni_count = 0; + __u32 lnd_type; + const struct lnet_lnd *lnd; + int peer_timeout = net->net_tunables.lct_peer_timeout; - int maxtxcredits = + int maxtxcredits = net->net_tunables.lct_max_tx_credits; - int peerrtrcredits = + 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 @@ -1901,10 +2529,6 @@ lnet_startup_lndnet(struct lnet_net *net, struct lnet_lnd_tunables *tun) } } - 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); @@ -1933,19 +2557,6 @@ lnet_startup_lndnet(struct lnet_net *net, struct lnet_lnd_tunables *tun) * 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); @@ -1954,8 +2565,8 @@ lnet_startup_lndnet(struct lnet_net *net, struct lnet_lnd_tunables *tun) /* 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; + ni->ni_interface)) { + rc = -EEXIST; goto failed1; } @@ -1965,9 +2576,6 @@ lnet_startup_lndnet(struct lnet_net *net, struct lnet_lnd_tunables *tun) 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; @@ -1992,7 +2600,6 @@ lnet_startup_lndnet(struct lnet_net *net, struct lnet_lnd_tunables *tun) */ lnet_net_free(net); } else { - net->net_state = LNET_NET_STATE_ACTIVE; /* * restore tunables after it has been overwitten by the * lnd @@ -2009,6 +2616,9 @@ lnet_startup_lndnet(struct lnet_net *net, struct lnet_lnd_tunables *tun) lnet_net_unlock(LNET_LOCK_EX); } + /* update net count */ + lnet_current_net_count = lnet_get_net_count(); + return ni_count; failed1: @@ -2081,8 +2691,8 @@ int lnet_lib_init(void) lnet_assert_wire_constants(); /* refer to global cfs_cpt_table for now */ - the_lnet.ln_cpt_table = cfs_cpt_table; - the_lnet.ln_cpt_number = cfs_cpt_number(cfs_cpt_table); + the_lnet.ln_cpt_table = cfs_cpt_tab; + the_lnet.ln_cpt_number = cfs_cpt_number(cfs_cpt_tab); LASSERT(the_lnet.ln_cpt_number > 0); if (the_lnet.ln_cpt_number > LNET_CPT_MAX) { @@ -2103,12 +2713,8 @@ int lnet_lib_init(void) } the_lnet.ln_refcount = 0; - LNetInvalidateEQHandle(&the_lnet.ln_rc_eqh); - INIT_LIST_HEAD(&the_lnet.ln_lnds); INIT_LIST_HEAD(&the_lnet.ln_net_zombie); - INIT_LIST_HEAD(&the_lnet.ln_rcd_zombie); INIT_LIST_HEAD(&the_lnet.ln_msg_resend); - INIT_LIST_HEAD(&the_lnet.ln_rcd_deathrow); /* The hash table size is the number of bits it takes to express the set * ln_num_routes, minus 1 (better to under estimate than over so we @@ -2132,14 +2738,18 @@ int lnet_lib_init(void) * * \pre lnet_lib_init() called with success. * \pre All LNet users called LNetNIFini() for matching LNetNIInit() calls. + * + * As this happens at module-unload, all lnds must already be unloaded, + * so they must already be unregistered. */ void lnet_lib_exit(void) { - LASSERT(the_lnet.ln_refcount == 0); + int i; - while (!list_empty(&the_lnet.ln_lnds)) - lnet_unregister_lnd(list_entry(the_lnet.ln_lnds.next, - struct lnet_lnd, lnd_list)); + LASSERT(the_lnet.ln_refcount == 0); + lnet_unregister_lnd(&the_lolnd); + for (i = 0; i < NUM_LNDS; i++) + LASSERT(!the_lnet.ln_lnds[i]); lnet_destroy_locks(); } @@ -2166,11 +2776,9 @@ LNetNIInit(lnet_pid_t requested_pid) int ni_count; struct lnet_ping_buffer *pbuf; struct lnet_handle_md ping_mdh; - struct list_head net_head; + LIST_HEAD(net_head); struct lnet_net *net; - INIT_LIST_HEAD(&net_head); - mutex_lock(&the_lnet.ln_api_mutex); CDEBUG(D_OTHER, "refs %d\n", the_lnet.ln_refcount); @@ -2200,6 +2808,9 @@ LNetNIInit(lnet_pid_t requested_pid) goto err_empty_list; } + if (use_tcp_bonding) + CWARN("use_tcp_bonding has been removed. Use Multi-Rail and Dynamic Discovery instead, see LU-13641\n"); + /* If LNet is being initialized via DLC it is possible * that the user requests not to load module parameters (ones which * are supported by DLC) on initialization. Therefore, make sure not @@ -2207,8 +2818,7 @@ LNetNIInit(lnet_pid_t requested_pid) * in this case. On cleanup in case of failure only clean up * routes if it has been loaded */ if (!the_lnet.ln_nis_from_mod_params) { - rc = lnet_parse_networks(&net_head, lnet_get_networks(), - use_tcp_bonding); + rc = lnet_parse_networks(&net_head, lnet_get_networks()); if (rc < 0) goto err_empty_list; } @@ -2224,10 +2834,6 @@ LNetNIInit(lnet_pid_t requested_pid) if (rc != 0) goto err_shutdown_lndnis; - rc = lnet_check_routes(); - if (rc != 0) - goto err_destroy_routes; - rc = lnet_rtrpools_alloc(im_a_router); if (rc != 0) goto err_destroy_routes; @@ -2246,29 +2852,36 @@ LNetNIInit(lnet_pid_t requested_pid) lnet_ping_target_update(pbuf, ping_mdh); - rc = lnet_router_checker_start(); - if (rc != 0) - goto err_stop_ping; + the_lnet.ln_mt_handler = lnet_mt_event_handler; rc = lnet_push_target_init(); if (rc != 0) - goto err_stop_router_checker; + goto err_stop_ping; rc = lnet_peer_discovery_start(); if (rc != 0) - goto err_destroy_push_target; + goto err_destroy_push_target; + + rc = lnet_monitor_thr_start(); + if (rc != 0) + goto err_stop_discovery_thr; lnet_fault_init(); lnet_router_debugfs_init(); mutex_unlock(&the_lnet.ln_api_mutex); + complete_all(&the_lnet.ln_started); + + /* wait for all routers to start */ + lnet_wait_router_start(); + return 0; +err_stop_discovery_thr: + lnet_peer_discovery_stop(); err_destroy_push_target: lnet_push_target_fini(); -err_stop_router_checker: - lnet_router_checker_stop(); err_stop_ping: lnet_ping_target_fini(); err_acceptor_stop: @@ -2304,7 +2917,7 @@ EXPORT_SYMBOL(LNetNIInit); * \return always 0 for current implementation. */ int -LNetNIFini() +LNetNIFini(void) { mutex_lock(&the_lnet.ln_api_mutex); @@ -2317,10 +2930,10 @@ LNetNIFini() lnet_fault_fini(); - lnet_router_debugfs_init(); + lnet_router_debugfs_fini(); + lnet_monitor_thr_stop(); lnet_peer_discovery_stop(); lnet_push_target_fini(); - lnet_router_checker_stop(); lnet_ping_target_fini(); /* Teardown fns that use my own API functions BEFORE here */ @@ -2357,22 +2970,17 @@ lnet_fill_ni_info(struct lnet_ni *ni, struct lnet_ioctl_config_ni *cfg_ni, if (!ni || !cfg_ni || !tun) return; - if (ni->ni_interfaces[0] != NULL) { - for (i = 0; i < ARRAY_SIZE(ni->ni_interfaces); i++) { - if (ni->ni_interfaces[i] != NULL) { - strncpy(cfg_ni->lic_ni_intf[i], - ni->ni_interfaces[i], - sizeof(cfg_ni->lic_ni_intf[i])); - } - } + if (ni->ni_interface != NULL) { + strncpy(cfg_ni->lic_ni_intf, + ni->ni_interface, + sizeof(cfg_ni->lic_ni_intf)); } cfg_ni->lic_nid = ni->ni_nid; - if (LNET_NETTYP(LNET_NIDNET(ni->ni_nid)) == LOLND) + if (ni->ni_nid == LNET_NID_LO_0) cfg_ni->lic_status = LNET_NI_STATUS_UP; else cfg_ni->lic_status = ni->ni_status->ns_status; - cfg_ni->lic_tcp_bonding = use_tcp_bonding; cfg_ni->lic_dev_cpt = ni->ni_dev_cpt; memcpy(&tun->lt_cmn, &ni->ni_net->net_tunables, sizeof(tun->lt_cmn)); @@ -2438,17 +3046,12 @@ lnet_fill_ni_info_legacy(struct lnet_ni *ni, if (!net_config) return; - BUILD_BUG_ON(ARRAY_SIZE(ni->ni_interfaces) != - ARRAY_SIZE(net_config->ni_interfaces)); - - for (i = 0; i < ARRAY_SIZE(ni->ni_interfaces); i++) { - if (!ni->ni_interfaces[i]) - break; + if (!ni->ni_interface) + return; - strncpy(net_config->ni_interfaces[i], - ni->ni_interfaces[i], - sizeof(net_config->ni_interfaces[i])); - } + strncpy(net_config->ni_interface, + ni->ni_interface, + sizeof(net_config->ni_interface)); config->cfg_nid = ni->ni_nid; config->cfg_config_u.cfg_net.net_peer_timeout = @@ -2460,7 +3063,7 @@ lnet_fill_ni_info_legacy(struct lnet_ni *ni, config->cfg_config_u.cfg_net.net_peer_rtr_credits = ni->ni_net->net_tunables.lct_peer_rtr_credits; - if (LNET_NETTYP(LNET_NIDNET(ni->ni_nid)) == LOLND) + if (ni->ni_nid == LNET_NID_LO_0) net_config->ni_status = LNET_NI_STATUS_UP; else net_config->ni_status = ni->ni_status->ns_status; @@ -2515,16 +3118,38 @@ lnet_get_ni_idx_locked(int idx) return NULL; } +int lnet_get_net_healthv_locked(struct lnet_net *net) +{ + struct lnet_ni *ni; + int best_healthv = 0; + int healthv; + + list_for_each_entry(ni, &net->net_ni_list, ni_netlist) { + healthv = atomic_read(&ni->ni_healthv); + if (healthv > best_healthv) + best_healthv = healthv; + } + + return best_healthv; +} + struct lnet_ni * lnet_get_next_ni_locked(struct lnet_net *mynet, struct lnet_ni *prev) { struct lnet_ni *ni; struct lnet_net *net = mynet; + /* + * It is possible that the net has been cleaned out while there is + * a message being sent. This function accessed the net without + * checking if the list is empty + */ if (prev == NULL) { if (net == NULL) net = list_entry(the_lnet.ln_nets.next, struct lnet_net, net_list); + if (list_empty(&net->net_ni_list)) + return NULL; ni = list_entry(net->net_ni_list.next, struct lnet_ni, ni_netlist); @@ -2546,6 +3171,8 @@ lnet_get_next_ni_locked(struct lnet_net *mynet, struct lnet_ni *prev) /* get the next net */ net = list_entry(prev->ni_net->net_list.next, struct lnet_net, net_list); + if (list_empty(&net->net_ni_list)) + return NULL; /* get the ni on it */ ni = list_entry(net->net_ni_list.next, struct lnet_ni, ni_netlist); @@ -2553,6 +3180,9 @@ lnet_get_next_ni_locked(struct lnet_net *mynet, struct lnet_ni *prev) return ni; } + if (list_empty(&prev->ni_netlist)) + return NULL; + /* there are more nis left */ ni = list_entry(prev->ni_netlist.next, struct lnet_ni, ni_netlist); @@ -2636,13 +3266,13 @@ int lnet_get_ni_stats(struct lnet_ioctl_element_msg_stats *msg_stats) static int lnet_add_net_common(struct lnet_net *net, struct lnet_ioctl_config_lnd_tunables *tun) { - __u32 net_id; + struct lnet_handle_md ping_mdh; struct lnet_ping_buffer *pbuf; - struct lnet_handle_md ping_mdh; - int rc; struct lnet_remotenet *rnet; - int net_ni_count; - int num_acceptor_nets; + struct lnet_ni *ni; + int net_ni_count; + __u32 net_id; + int rc; lnet_net_lock(LNET_LOCK_EX); rnet = lnet_find_rnet_locked(net->net_id); @@ -2683,14 +3313,6 @@ static int lnet_add_net_common(struct lnet_net *net, else memset(&net->net_tunables, -1, sizeof(net->net_tunables)); - /* - * before starting this network get a count of the current TCP - * networks which require the acceptor thread running. If that - * count is == 0 before we start up this network, then we'd want to - * start up the acceptor thread after starting up this network - */ - num_acceptor_nets = lnet_count_acceptor_nets(); - net_id = net->net_id; rc = lnet_startup_lndnet(net, @@ -2700,15 +3322,30 @@ static int lnet_add_net_common(struct lnet_net *net, lnet_net_lock(LNET_LOCK_EX); net = lnet_get_net_locked(net_id); - lnet_net_unlock(LNET_LOCK_EX); - LASSERT(net); + /* apply the UDSPs */ + rc = lnet_udsp_apply_policies_on_net(net); + if (rc) + CERROR("Failed to apply UDSPs on local net %s\n", + libcfs_net2str(net->net_id)); + + /* At this point we lost track of which NI was just added, so we + * just re-apply the policies on all of the NIs on this net + */ + list_for_each_entry(ni, &net->net_ni_list, ni_netlist) { + rc = lnet_udsp_apply_policies_on_ni(ni); + if (rc) + CERROR("Failed to apply UDSPs on ni %s\n", + libcfs_nid2str(ni->ni_nid)); + } + lnet_net_unlock(LNET_LOCK_EX); + /* * Start the acceptor thread if this is the first network * being added that requires the thread. */ - if (net->net_lnd->lnd_accept && num_acceptor_nets == 0) { + if (net->net_lnd->lnd_accept) { rc = lnet_acceptor_start(); if (rc < 0) { /* shutdown the net that we just started */ @@ -2732,24 +3369,37 @@ failed: return rc; } +static void +lnet_set_tune_defaults(struct lnet_ioctl_config_lnd_tunables *tun) +{ + if (tun) { + if (!tun->lt_cmn.lct_peer_timeout) + tun->lt_cmn.lct_peer_timeout = DEFAULT_PEER_TIMEOUT; + if (!tun->lt_cmn.lct_peer_tx_credits) + tun->lt_cmn.lct_peer_tx_credits = DEFAULT_PEER_CREDITS; + if (!tun->lt_cmn.lct_max_tx_credits) + tun->lt_cmn.lct_max_tx_credits = DEFAULT_CREDITS; + } +} + static int lnet_handle_legacy_ip2nets(char *ip2nets, struct lnet_ioctl_config_lnd_tunables *tun) { struct lnet_net *net; - char *nets; + const char *nets; int rc; - struct list_head net_head; - - INIT_LIST_HEAD(&net_head); + LIST_HEAD(net_head); rc = lnet_parse_ip2nets(&nets, ip2nets); if (rc < 0) return rc; - rc = lnet_parse_networks(&net_head, nets, use_tcp_bonding); + rc = lnet_parse_networks(&net_head, nets); if (rc < 0) return rc; + lnet_set_tune_defaults(tun); + mutex_lock(&the_lnet.ln_api_mutex); while (!list_empty(&net_head)) { net = list_entry(net_head.next, struct lnet_net, net_list); @@ -2807,10 +3457,12 @@ int lnet_dyn_add_ni(struct lnet_ioctl_config_ni *conf) } ni = lnet_ni_alloc_w_cpt_array(net, conf->lic_cpts, conf->lic_ncpts, - conf->lic_ni_intf[0]); + conf->lic_ni_intf); if (!ni) return -ENOMEM; + lnet_set_tune_defaults(tun); + mutex_lock(&the_lnet.ln_api_mutex); rc = lnet_add_net_common(net, tun); @@ -2863,8 +3515,7 @@ int lnet_dyn_del_ni(struct lnet_ioctl_config_ni *conf) lnet_shutdown_lndnet(net); - if (lnet_count_acceptor_nets() == 0) - lnet_acceptor_stop(); + lnet_acceptor_stop(); lnet_ping_target_update(pbuf, ping_mdh); @@ -2891,8 +3542,7 @@ int lnet_dyn_del_ni(struct lnet_ioctl_config_ni *conf) lnet_shutdown_lndni(ni); - if (lnet_count_acceptor_nets() == 0) - lnet_acceptor_stop(); + lnet_acceptor_stop(); lnet_ping_target_update(pbuf, ping_mdh); @@ -2919,16 +3569,14 @@ unlock_api_mutex: int lnet_dyn_add_net(struct lnet_ioctl_config_data *conf) { - struct lnet_net *net; - struct list_head net_head; - int rc; + struct lnet_net *net; + LIST_HEAD(net_head); + int rc; struct lnet_ioctl_config_lnd_tunables tun; - char *nets = conf->cfg_config_u.cfg_net.net_intf; - - INIT_LIST_HEAD(&net_head); + const char *nets = conf->cfg_config_u.cfg_net.net_intf; /* Create a net/ni structures for the network string */ - rc = lnet_parse_networks(&net_head, nets, use_tcp_bonding); + rc = lnet_parse_networks(&net_head, nets); if (rc <= 0) return rc == 0 ? -EINVAL : rc; @@ -2947,13 +3595,16 @@ lnet_dyn_add_net(struct lnet_ioctl_config_data *conf) memset(&tun, 0, sizeof(tun)); tun.lt_cmn.lct_peer_timeout = - conf->cfg_config_u.cfg_net.net_peer_timeout; + (!conf->cfg_config_u.cfg_net.net_peer_timeout) ? DEFAULT_PEER_TIMEOUT : + conf->cfg_config_u.cfg_net.net_peer_timeout; tun.lt_cmn.lct_peer_tx_credits = - conf->cfg_config_u.cfg_net.net_peer_tx_credits; + (!conf->cfg_config_u.cfg_net.net_peer_tx_credits) ? DEFAULT_PEER_CREDITS : + conf->cfg_config_u.cfg_net.net_peer_tx_credits; tun.lt_cmn.lct_peer_rtr_credits = conf->cfg_config_u.cfg_net.net_peer_rtr_credits; tun.lt_cmn.lct_max_tx_credits = - conf->cfg_config_u.cfg_net.net_max_tx_credits; + (!conf->cfg_config_u.cfg_net.net_max_tx_credits) ? DEFAULT_CREDITS : + conf->cfg_config_u.cfg_net.net_max_tx_credits; rc = lnet_add_net_common(net, &tun); @@ -3004,8 +3655,7 @@ lnet_dyn_del_net(__u32 net_id) lnet_shutdown_lndnet(net); - if (lnet_count_acceptor_nets() == 0) - lnet_acceptor_stop(); + lnet_acceptor_stop(); lnet_ping_target_update(pbuf, ping_mdh); @@ -3025,6 +3675,102 @@ __u32 lnet_get_dlc_seq_locked(void) return atomic_read(&lnet_dlc_seq_no); } +static void +lnet_ni_set_healthv(lnet_nid_t nid, int value, bool all) +{ + 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) { + if (ni->ni_nid == nid || all) { + atomic_set(&ni->ni_healthv, value); + if (list_empty(&ni->ni_recovery) && + value < LNET_MAX_HEALTH_VALUE) { + CERROR("manually adding local NI %s to recovery\n", + libcfs_nid2str(ni->ni_nid)); + list_add_tail(&ni->ni_recovery, + &the_lnet.ln_mt_localNIRecovq); + lnet_ni_addref_locked(ni, 0); + } + if (!all) { + lnet_net_unlock(LNET_LOCK_EX); + return; + } + } + } + } + lnet_net_unlock(LNET_LOCK_EX); +} + +static int +lnet_get_local_ni_hstats(struct lnet_ioctl_local_ni_hstats *stats) +{ + int cpt, rc = 0; + struct lnet_ni *ni; + lnet_nid_t nid = stats->hlni_nid; + + cpt = lnet_net_lock_current(); + ni = lnet_nid2ni_locked(nid, cpt); + + if (!ni) { + rc = -ENOENT; + goto unlock; + } + + stats->hlni_local_interrupt = atomic_read(&ni->ni_hstats.hlt_local_interrupt); + stats->hlni_local_dropped = atomic_read(&ni->ni_hstats.hlt_local_dropped); + stats->hlni_local_aborted = atomic_read(&ni->ni_hstats.hlt_local_aborted); + stats->hlni_local_no_route = atomic_read(&ni->ni_hstats.hlt_local_no_route); + stats->hlni_local_timeout = atomic_read(&ni->ni_hstats.hlt_local_timeout); + stats->hlni_local_error = atomic_read(&ni->ni_hstats.hlt_local_error); + stats->hlni_health_value = atomic_read(&ni->ni_healthv); + +unlock: + lnet_net_unlock(cpt); + + return rc; +} + +static int +lnet_get_local_ni_recovery_list(struct lnet_ioctl_recovery_list *list) +{ + struct lnet_ni *ni; + int i = 0; + + lnet_net_lock(LNET_LOCK_EX); + list_for_each_entry(ni, &the_lnet.ln_mt_localNIRecovq, ni_recovery) { + list->rlst_nid_array[i] = ni->ni_nid; + i++; + if (i >= LNET_MAX_SHOW_NUM_NID) + break; + } + lnet_net_unlock(LNET_LOCK_EX); + list->rlst_num_nids = i; + + return 0; +} + +static int +lnet_get_peer_ni_recovery_list(struct lnet_ioctl_recovery_list *list) +{ + struct lnet_peer_ni *lpni; + int i = 0; + + lnet_net_lock(LNET_LOCK_EX); + list_for_each_entry(lpni, &the_lnet.ln_mt_peerNIRecovq, lpni_recovery) { + list->rlst_nid_array[i] = lpni->lpni_nid; + i++; + if (i >= LNET_MAX_SHOW_NUM_NID) + break; + } + lnet_net_unlock(LNET_LOCK_EX); + list->rlst_num_nids = i; + + return 0; +} + /** * LNet ioctl handler. * @@ -3050,26 +3796,28 @@ LNetCtl(unsigned int cmd, void *arg) case IOC_LIBCFS_FAIL_NID: return lnet_fail_nid(data->ioc_nid, data->ioc_count); - case IOC_LIBCFS_ADD_ROUTE: + case IOC_LIBCFS_ADD_ROUTE: { + /* default router sensitivity to 1 */ + unsigned int sensitivity = 1; config = arg; if (config->cfg_hdr.ioc_len < sizeof(*config)) return -EINVAL; + if (config->cfg_config_u.cfg_route.rtr_sensitivity) { + sensitivity = + config->cfg_config_u.cfg_route.rtr_sensitivity; + } + mutex_lock(&the_lnet.ln_api_mutex); rc = lnet_add_route(config->cfg_net, config->cfg_config_u.cfg_route.rtr_hop, config->cfg_nid, config->cfg_config_u.cfg_route. - rtr_priority); - if (rc == 0) { - rc = lnet_check_routes(); - if (rc != 0) - lnet_del_route(config->cfg_net, - config->cfg_nid); - } + rtr_priority, sensitivity); mutex_unlock(&the_lnet.ln_api_mutex); return rc; + } case IOC_LIBCFS_DEL_ROUTE: config = arg; @@ -3095,7 +3843,9 @@ LNetCtl(unsigned int cmd, void *arg) &config->cfg_nid, &config->cfg_config_u.cfg_route.rtr_flags, &config->cfg_config_u.cfg_route. - rtr_priority); + rtr_priority, + &config->cfg_config_u.cfg_route. + rtr_sensitivity); mutex_unlock(&the_lnet.ln_api_mutex); return rc; @@ -3161,9 +3911,9 @@ LNetCtl(unsigned int cmd, void *arg) return -EINVAL; mutex_lock(&the_lnet.ln_api_mutex); - lnet_counters_get(&lnet_stats->st_cntrs); + rc = lnet_counters_get(&lnet_stats->st_cntrs); mutex_unlock(&the_lnet.ln_api_mutex); - return 0; + return rc; } case IOC_LIBCFS_CONFIG_RTR: @@ -3235,6 +3985,33 @@ LNetCtl(unsigned int cmd, void *arg) return rc; } + case IOC_LIBCFS_GET_LOCAL_HSTATS: { + struct lnet_ioctl_local_ni_hstats *stats = arg; + + if (stats->hlni_hdr.ioc_len < sizeof(*stats)) + return -EINVAL; + + mutex_lock(&the_lnet.ln_api_mutex); + rc = lnet_get_local_ni_hstats(stats); + mutex_unlock(&the_lnet.ln_api_mutex); + + return rc; + } + + case IOC_LIBCFS_GET_RECOVERY_QUEUE: { + struct lnet_ioctl_recovery_list *list = arg; + if (list->rlst_hdr.ioc_len < sizeof(*list)) + return -EINVAL; + + mutex_lock(&the_lnet.ln_api_mutex); + if (list->rlst_type == LNET_HEALTH_TYPE_LOCAL_NI) + rc = lnet_get_local_ni_recovery_list(list); + else + rc = lnet_get_peer_ni_recovery_list(list); + mutex_unlock(&the_lnet.ln_api_mutex); + return rc; + } + case IOC_LIBCFS_ADD_PEER_NI: { struct lnet_ioctl_peer_cfg *cfg = arg; @@ -3310,6 +4087,30 @@ LNetCtl(unsigned int cmd, void *arg) return rc; } + case IOC_LIBCFS_SET_HEALHV: { + struct lnet_ioctl_reset_health_cfg *cfg = arg; + int value; + if (cfg->rh_hdr.ioc_len < sizeof(*cfg)) + return -EINVAL; + if (cfg->rh_value < 0 || + cfg->rh_value > LNET_MAX_HEALTH_VALUE) + value = LNET_MAX_HEALTH_VALUE; + else + value = cfg->rh_value; + CDEBUG(D_NET, "Manually setting healthv to %d for %s:%s. all = %d\n", + value, (cfg->rh_type == LNET_HEALTH_TYPE_LOCAL_NI) ? + "local" : "peer", libcfs_nid2str(cfg->rh_nid), cfg->rh_all); + mutex_lock(&the_lnet.ln_api_mutex); + if (cfg->rh_type == LNET_HEALTH_TYPE_LOCAL_NI) + lnet_ni_set_healthv(cfg->rh_nid, value, + cfg->rh_all); + else + lnet_peer_ni_set_healthv(cfg->rh_nid, value, + cfg->rh_all); + mutex_unlock(&the_lnet.ln_api_mutex); + return 0; + } + case IOC_LIBCFS_NOTIFY_ROUTER: { time64_t deadline = ktime_get_real_seconds() - data->ioc_u64[0]; @@ -3318,7 +4119,7 @@ LNetCtl(unsigned int cmd, void *arg) * that deadline to the wall clock. */ deadline += ktime_get_seconds(); - return lnet_notify(NULL, data->ioc_nid, data->ioc_flags, + return lnet_notify(NULL, data->ioc_nid, data->ioc_flags, false, deadline); } @@ -3331,9 +4132,7 @@ LNetCtl(unsigned int cmd, void *arg) return 0; case IOC_LIBCFS_TESTPROTOCOMPAT: - lnet_net_lock(LNET_LOCK_EX); the_lnet.ln_testprotocompat = data->ioc_flags; - lnet_net_unlock(LNET_LOCK_EX); return 0; case IOC_LIBCFS_LNET_FAULT: @@ -3348,9 +4147,9 @@ LNetCtl(unsigned int cmd, void *arg) /* If timeout is negative then set default of 3 minutes */ if (((s32)data->ioc_u32[1] <= 0) || data->ioc_u32[1] > (DEFAULT_PEER_TIMEOUT * MSEC_PER_SEC)) - timeout = msecs_to_jiffies(DEFAULT_PEER_TIMEOUT * MSEC_PER_SEC); + timeout = cfs_time_seconds(DEFAULT_PEER_TIMEOUT); else - timeout = msecs_to_jiffies(data->ioc_u32[1]); + timeout = nsecs_to_jiffies(data->ioc_u32[1] * NSEC_PER_MSEC); rc = lnet_ping(id, timeout, data->ioc_pbuf1, data->ioc_plen1 / sizeof(struct lnet_process_id)); @@ -3370,9 +4169,9 @@ LNetCtl(unsigned int cmd, void *arg) /* If timeout is negative then set default of 3 minutes */ if (((s32)ping->op_param) <= 0 || ping->op_param > (DEFAULT_PEER_TIMEOUT * MSEC_PER_SEC)) - timeout = msecs_to_jiffies(DEFAULT_PEER_TIMEOUT * MSEC_PER_SEC); + timeout = cfs_time_seconds(DEFAULT_PEER_TIMEOUT); else - timeout = msecs_to_jiffies(ping->op_param); + timeout = nsecs_to_jiffies(ping->op_param * NSEC_PER_MSEC); rc = lnet_ping(ping->ping_id, timeout, ping->ping_buf, @@ -3416,6 +4215,106 @@ LNetCtl(unsigned int cmd, void *arg) return 0; } + case IOC_LIBCFS_ADD_UDSP: { + struct lnet_ioctl_udsp *ioc_udsp = arg; + __u32 bulk_size = ioc_udsp->iou_hdr.ioc_len; + + mutex_lock(&the_lnet.ln_api_mutex); + rc = lnet_udsp_demarshal_add(arg, bulk_size); + if (!rc) { + rc = lnet_udsp_apply_policies(NULL, false); + CDEBUG(D_NET, "policy application returned %d\n", rc); + rc = 0; + } + mutex_unlock(&the_lnet.ln_api_mutex); + + return rc; + } + + case IOC_LIBCFS_DEL_UDSP: { + struct lnet_ioctl_udsp *ioc_udsp = arg; + int idx = ioc_udsp->iou_idx; + + if (ioc_udsp->iou_hdr.ioc_len < sizeof(*ioc_udsp)) + return -EINVAL; + + mutex_lock(&the_lnet.ln_api_mutex); + rc = lnet_udsp_del_policy(idx); + if (!rc) { + rc = lnet_udsp_apply_policies(NULL, false); + CDEBUG(D_NET, "policy re-application returned %d\n", + rc); + rc = 0; + } + mutex_unlock(&the_lnet.ln_api_mutex); + + return rc; + } + + case IOC_LIBCFS_GET_UDSP_SIZE: { + struct lnet_ioctl_udsp *ioc_udsp = arg; + struct lnet_udsp *udsp; + + if (ioc_udsp->iou_hdr.ioc_len < sizeof(*ioc_udsp)) + return -EINVAL; + + rc = 0; + + mutex_lock(&the_lnet.ln_api_mutex); + udsp = lnet_udsp_get_policy(ioc_udsp->iou_idx); + if (!udsp) { + rc = -ENOENT; + } else { + /* coming in iou_idx will hold the idx of the udsp + * to get the size of. going out the iou_idx will + * hold the size of the UDSP found at the passed + * in index. + */ + ioc_udsp->iou_idx = lnet_get_udsp_size(udsp); + if (ioc_udsp->iou_idx < 0) + rc = -EINVAL; + } + mutex_unlock(&the_lnet.ln_api_mutex); + + return rc; + } + + case IOC_LIBCFS_GET_UDSP: { + struct lnet_ioctl_udsp *ioc_udsp = arg; + struct lnet_udsp *udsp; + + if (ioc_udsp->iou_hdr.ioc_len < sizeof(*ioc_udsp)) + return -EINVAL; + + rc = 0; + + mutex_lock(&the_lnet.ln_api_mutex); + udsp = lnet_udsp_get_policy(ioc_udsp->iou_idx); + if (!udsp) + rc = -ENOENT; + else + rc = lnet_udsp_marshal(udsp, ioc_udsp); + mutex_unlock(&the_lnet.ln_api_mutex); + + return rc; + } + + case IOC_LIBCFS_GET_CONST_UDSP_INFO: { + struct lnet_ioctl_construct_udsp_info *info = arg; + + if (info->cud_hdr.ioc_len < sizeof(*info)) + return -EINVAL; + + CDEBUG(D_NET, "GET_UDSP_INFO for %s\n", + libcfs_nid2str(info->cud_nid)); + + mutex_lock(&the_lnet.ln_api_mutex); + lnet_udsp_get_construct_info(info); + mutex_unlock(&the_lnet.ln_api_mutex); + + return 0; + } + default: ni = lnet_net2ni_addref(data->ioc_net); if (ni == NULL) @@ -3508,24 +4407,44 @@ LNetGetId(unsigned int index, struct lnet_process_id *id) } EXPORT_SYMBOL(LNetGetId); +struct ping_data { + int rc; + int replied; + struct lnet_handle_md mdh; + struct completion completion; +}; + +static void +lnet_ping_event_handler(struct lnet_event *event) +{ + struct ping_data *pd = event->md_user_ptr; + + CDEBUG(D_NET, "ping event (%d %d)%s\n", + event->type, event->status, + event->unlinked ? " unlinked" : ""); + + if (event->status) { + if (!pd->rc) + pd->rc = event->status; + } else if (event->type == LNET_EVENT_REPLY) { + pd->replied = 1; + pd->rc = event->mlength; + } + if (event->unlinked) + complete(&pd->completion); +} + static int lnet_ping(struct lnet_process_id id, signed long timeout, struct lnet_process_id __user *ids, int n_ids) { - struct lnet_handle_eq eqh; - struct lnet_handle_md mdh; - struct lnet_event event; struct lnet_md md = { NULL }; - int which; - int unlinked = 0; - int replied = 0; - const signed long a_long_time = msecs_to_jiffies(60 * MSEC_PER_SEC); + struct ping_data pd = { 0 }; struct lnet_ping_buffer *pbuf; struct lnet_process_id tmpid; int i; int nob; int rc; int rc2; - sigset_t blocked; /* n_ids limit is arbitrary */ if (n_ids <= 0 || id.nid == LNET_NID_ANY) @@ -3545,92 +4464,46 @@ static int lnet_ping(struct lnet_process_id id, signed long timeout, if (!pbuf) return -ENOMEM; - /* NB 2 events max (including any unlink event) */ - rc = LNetEQAlloc(2, LNET_EQ_HANDLER_NONE, &eqh); - if (rc != 0) { - CERROR("Can't allocate EQ: %d\n", rc); - goto fail_ping_buffer_decref; - } - /* initialize md content */ md.start = &pbuf->pb_info; md.length = LNET_PING_INFO_SIZE(n_ids); md.threshold = 2; /* GET/REPLY */ md.max_size = 0; md.options = LNET_MD_TRUNCATE; - md.user_ptr = NULL; - md.eq_handle = eqh; + md.user_ptr = &pd; + md.handler = lnet_ping_event_handler; + + init_completion(&pd.completion); - rc = LNetMDBind(md, LNET_UNLINK, &mdh); + rc = LNetMDBind(&md, LNET_UNLINK, &pd.mdh); if (rc != 0) { CERROR("Can't bind MD: %d\n", rc); - goto fail_free_eq; + goto fail_ping_buffer_decref; } - rc = LNetGet(LNET_NID_ANY, mdh, id, + rc = LNetGet(LNET_NID_ANY, pd.mdh, id, LNET_RESERVED_PORTAL, - LNET_PROTO_PING_MATCHBITS, 0); + LNET_PROTO_PING_MATCHBITS, 0, false); if (rc != 0) { /* Don't CERROR; this could be deliberate! */ - rc2 = LNetMDUnlink(mdh); + rc2 = LNetMDUnlink(pd.mdh); LASSERT(rc2 == 0); /* NB must wait for the UNLINK event below... */ - unlinked = 1; - timeout = a_long_time; - } - - do { - /* MUST block for unlink to complete */ - if (unlinked) - blocked = cfs_block_allsigs(); - - rc2 = LNetEQPoll(&eqh, 1, timeout, &event, &which); - - if (unlinked) - cfs_restore_sigs(blocked); - - CDEBUG(D_NET, "poll %d(%d %d)%s\n", rc2, - (rc2 <= 0) ? -1 : event.type, - (rc2 <= 0) ? -1 : event.status, - (rc2 > 0 && event.unlinked) ? " unlinked" : ""); - - LASSERT(rc2 != -EOVERFLOW); /* can't miss anything */ - - if (rc2 <= 0 || event.status != 0) { - /* timeout or error */ - if (!replied && rc == 0) - rc = (rc2 < 0) ? rc2 : - (rc2 == 0) ? -ETIMEDOUT : - event.status; - - if (!unlinked) { - /* Ensure completion in finite time... */ - LNetMDUnlink(mdh); - /* No assertion (racing with network) */ - unlinked = 1; - timeout = a_long_time; - } else if (rc2 == 0) { - /* timed out waiting for unlink */ - CWARN("ping %s: late network completion\n", - libcfs_id2str(id)); - } - } else if (event.type == LNET_EVENT_REPLY) { - replied = 1; - rc = event.mlength; - } - } while (rc2 <= 0 || !event.unlinked); + } - if (!replied) { - if (rc >= 0) - CWARN("%s: Unexpected rc >= 0 but no reply!\n", - libcfs_id2str(id)); + if (wait_for_completion_timeout(&pd.completion, timeout) == 0) { + /* Ensure completion in finite time... */ + LNetMDUnlink(pd.mdh); + wait_for_completion(&pd.completion); + } + if (!pd.replied) { rc = -EIO; - goto fail_free_eq; + goto fail_ping_buffer_decref; } - nob = rc; + nob = pd.rc; LASSERT(nob >= 0 && nob <= LNET_PING_INFO_SIZE(n_ids)); rc = -EPROTO; /* if I can't parse... */ @@ -3638,7 +4511,7 @@ static int lnet_ping(struct lnet_process_id id, signed long timeout, if (nob < 8) { CERROR("%s: ping info too short %d\n", libcfs_id2str(id), nob); - goto fail_free_eq; + goto fail_ping_buffer_decref; } if (pbuf->pb_info.pi_magic == __swab32(LNET_PROTO_PING_MAGIC)) { @@ -3646,20 +4519,20 @@ static int lnet_ping(struct lnet_process_id id, signed long timeout, } else if (pbuf->pb_info.pi_magic != LNET_PROTO_PING_MAGIC) { CERROR("%s: Unexpected magic %08x\n", libcfs_id2str(id), pbuf->pb_info.pi_magic); - goto fail_free_eq; + goto fail_ping_buffer_decref; } if ((pbuf->pb_info.pi_features & LNET_PING_FEAT_NI_STATUS) == 0) { CERROR("%s: ping w/o NI status: 0x%x\n", libcfs_id2str(id), pbuf->pb_info.pi_features); - goto fail_free_eq; + goto fail_ping_buffer_decref; } if (nob < LNET_PING_INFO_SIZE(0)) { CERROR("%s: Short reply %d(%d min)\n", libcfs_id2str(id), nob, (int)LNET_PING_INFO_SIZE(0)); - goto fail_free_eq; + goto fail_ping_buffer_decref; } if (pbuf->pb_info.pi_nnis < n_ids) @@ -3669,7 +4542,7 @@ static int lnet_ping(struct lnet_process_id id, signed long timeout, CERROR("%s: Short reply %d(%d expected)\n", libcfs_id2str(id), nob, (int)LNET_PING_INFO_SIZE(n_ids)); - goto fail_free_eq; + goto fail_ping_buffer_decref; } rc = -EFAULT; /* if I segv in copy_to_user()... */ @@ -3679,16 +4552,10 @@ static int lnet_ping(struct lnet_process_id id, signed long timeout, tmpid.pid = pbuf->pb_info.pi_pid; tmpid.nid = pbuf->pb_info.pi_ni[i].ns_nid; if (copy_to_user(&ids[i], &tmpid, sizeof(tmpid))) - goto fail_free_eq; + goto fail_ping_buffer_decref; } rc = pbuf->pb_info.pi_nnis; - fail_free_eq: - rc2 = LNetEQFree(eqh); - if (rc2 != 0) - CERROR("rc2 %d\n", rc2); - LASSERT(rc2 == 0); - fail_ping_buffer_decref: lnet_ping_buffer_decref(pbuf); return rc; @@ -3705,8 +4572,6 @@ lnet_discover(struct lnet_process_id id, __u32 force, int cpt; int i; int rc; - int max_intf = lnet_interfaces_max; - size_t buf_size; if (n_ids <= 0 || id.nid == LNET_NID_ANY) @@ -3716,15 +4581,13 @@ lnet_discover(struct lnet_process_id id, __u32 force, id.pid = LNET_PID_LUSTRE; /* - * if the user buffer has more space than the max_intf - * then only fill it up to max_intf + * If the user buffer has more space than the lnet_interfaces_max, + * then only fill it up to lnet_interfaces_max. */ - if (n_ids > max_intf) - n_ids = max_intf; - - buf_size = n_ids * sizeof(*buf); + if (n_ids > lnet_interfaces_max) + n_ids = lnet_interfaces_max; - LIBCFS_ALLOC(buf, buf_size); + CFS_ALLOC_PTR_ARRAY(buf, n_ids); if (!buf) return -ENOMEM; @@ -3750,10 +4613,17 @@ lnet_discover(struct lnet_process_id id, __u32 force, if (rc) goto out_decref; - /* Peer may have changed. */ + /* The lpni (or lp) for this NID may have changed and our ref is + * the only thing keeping the old one around. Release the ref + * and lookup the lpni again + */ + lnet_peer_ni_decref_locked(lpni); + lpni = lnet_find_peer_ni_locked(id.nid); + if (!lpni) { + rc = -ENOENT; + goto out; + } lp = lpni->lpni_peer_net->lpn_peer; - if (lp->lp_nnis < n_ids) - n_ids = lp->lp_nnis; i = 0; p = NULL; @@ -3763,21 +4633,30 @@ lnet_discover(struct lnet_process_id id, __u32 force, if (++i >= n_ids) break; } + rc = i; - lnet_net_unlock(cpt); - - rc = -EFAULT; - if (copy_to_user(ids, buf, n_ids * sizeof(*buf))) - goto out_relock; - rc = n_ids; -out_relock: - lnet_net_lock(cpt); out_decref: lnet_peer_ni_decref_locked(lpni); out: lnet_net_unlock(cpt); - LIBCFS_FREE(buf, buf_size); + if (rc >= 0) + if (copy_to_user(ids, buf, rc * sizeof(*buf))) + rc = -EFAULT; + CFS_FREE_PTR_ARRAY(buf, n_ids); return rc; } + +/** + * Retrieve peer discovery status. + * + * \retval 1 if lnet_peer_discovery_disabled is 0 + * \retval 0 if lnet_peer_discovery_disabled is 1 + */ +int +LNetGetPeerDiscoveryStatus(void) +{ + return !lnet_peer_discovery_disabled; +} +EXPORT_SYMBOL(LNetGetPeerDiscoveryStatus);