X-Git-Url: http://git.efficios.com/?a=blobdiff_plain;f=src%2Fbin%2Flttng-sessiond%2Fust-app.c;h=839a7143752b4cd340b826298a5daf4c431982b2;hb=refs%2Fheads%2Fsow-2020-0002-rev2;hp=856de63a3f24152d03944c003bfe28c890f4847a;hpb=a70ac2f49aa47a30abc2b50ec57b582eab4b1028;p=lttng-tools.git diff --git a/src/bin/lttng-sessiond/ust-app.c b/src/bin/lttng-sessiond/ust-app.c index 856de63a3..839a71437 100644 --- a/src/bin/lttng-sessiond/ust-app.c +++ b/src/bin/lttng-sessiond/ust-app.c @@ -1,38 +1,48 @@ /* - * Copyright (C) 2011 - David Goulet - * Copyright (C) 2016 - Jérémie Galarneau + * Copyright (C) 2011 David Goulet + * Copyright (C) 2016 Jérémie Galarneau * - * This program is free software; you can redistribute it and/or modify - * it under the terms of the GNU General Public License, version 2 only, - * as published by the Free Software Foundation. + * SPDX-License-Identifier: GPL-2.0-only * - * This program is distributed in the hope that it will be useful, - * but WITHOUT ANY WARRANTY; without even the implied warranty of - * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the - * GNU General Public License for more details. - * - * You should have received a copy of the GNU General Public License along - * with this program; if not, write to the Free Software Foundation, Inc., - * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. */ #define _LGPL_SOURCE #include +#include #include #include #include #include #include +#include #include #include #include #include #include +#include +#include #include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include #include #include "buffer-registry.h" +#include "condition-internal.h" #include "fd-limit.h" #include "health-sessiond.h" #include "ust-app.h" @@ -44,6 +54,14 @@ #include "lttng-sessiond.h" #include "notification-thread-commands.h" #include "rotate.h" +#include "event.h" +#include "event-notifier-error-accounting.h" +#include "map.h" + + +struct lttng_ht *ust_app_ht; +struct lttng_ht *ust_app_ht_by_sock; +struct lttng_ht *ust_app_ht_by_notify_sock; static int ust_app_flush_app_session(struct ust_app *app, struct ust_app_session *ua_sess); @@ -52,6 +70,10 @@ int ust_app_flush_app_session(struct ust_app *app, struct ust_app_session *ua_se static uint64_t _next_channel_key; static pthread_mutex_t next_channel_key_lock = PTHREAD_MUTEX_INITIALIZER; +/* Next available map key. Access under next_map_key_lock. */ +static uint64_t _next_map_key; +static pthread_mutex_t next_map_key_lock = PTHREAD_MUTEX_INITIALIZER; + /* Next available session ID. Access under next_session_id_lock. */ static uint64_t _next_session_id; static pthread_mutex_t next_session_id_lock = PTHREAD_MUTEX_INITIALIZER; @@ -69,6 +91,19 @@ static uint64_t get_next_channel_key(void) return ret; } +/* + * Return the incremented value of next_map_key. + */ +static uint64_t get_next_map_key(void) +{ + uint64_t ret; + + pthread_mutex_lock(&next_map_key_lock); + ret = ++_next_map_key; + pthread_mutex_unlock(&next_map_key_lock); + return ret; +} + /* * Return the atomically incremented value of next_session_id. */ @@ -115,7 +150,14 @@ static int ht_match_ust_app_event(struct cds_lfht_node *node, const void *_key) key = _key; ev_loglevel_value = event->attr.loglevel; - /* Match the 4 elements of the key: name, filter, loglevel, exclusions */ + /* + * Match the 5 elements of the key: + * tracer token, name, filter, loglevel, exclusions + */ + + if (event->attr.token != key->tracer_token) { + goto no_match; + } /* Event name */ if (strncmp(event->attr.name, key->name, sizeof(event->attr.name)) != 0) { @@ -178,25 +220,23 @@ no_match: * Unique add of an ust app event in the given ht. This uses the custom * ht_match_ust_app_event match function and the event name as hash. */ -static void add_unique_ust_app_event(struct ust_app_channel *ua_chan, +static void add_unique_ust_app_event(struct lttng_ht *events_ht, struct ust_app_event *event) { struct cds_lfht_node *node_ptr; struct ust_app_ht_key key; - struct lttng_ht *ht; - assert(ua_chan); - assert(ua_chan->events); + assert(events_ht); assert(event); - ht = ua_chan->events; key.name = event->attr.name; key.filter = event->filter; key.loglevel_type = event->attr.loglevel; key.exclusion = event->exclusion; + key.tracer_token = event->attr.token; - node_ptr = cds_lfht_add_unique(ht->ht, - ht->hash_fct(event->node.key, lttng_ht_seed), + node_ptr = cds_lfht_add_unique(events_ht->ht, + events_ht->hash_fct(event->node.key, lttng_ht_seed), ht_match_ust_app_event, &key, &event->node.node); assert(node_ptr == &event->node.node); } @@ -251,7 +291,7 @@ static struct ust_registry_session *get_session_registry( { struct buffer_reg_uid *reg_uid = buffer_reg_uid_find( ua_sess->tracing_id, ua_sess->bits_per_long, - ua_sess->real_credentials.uid); + lttng_credentials_get_uid(&ua_sess->real_credentials)); if (!reg_uid) { goto error; } @@ -319,6 +359,51 @@ void delete_ust_app_event(int sock, struct ust_app_event *ua_event, free(ua_event); } +/* + * Delayed reclaim of a ust_app_event_notifier_rule object. This MUST be called + * through a call_rcu(). + */ +static +void free_ust_app_event_notifier_rule_rcu(struct rcu_head *head) +{ + struct ust_app_event_notifier_rule *obj = caa_container_of( + head, struct ust_app_event_notifier_rule, rcu_head); + + free(obj); +} + +/* + * Delete ust app event notifier rule safely. + */ +static void delete_ust_app_event_notifier_rule(int sock, + struct ust_app_event_notifier_rule *ua_event_notifier_rule, + struct ust_app *app) +{ + int ret; + + assert(ua_event_notifier_rule); + + if (ua_event_notifier_rule->exclusion != NULL) { + free(ua_event_notifier_rule->exclusion); + } + + if (ua_event_notifier_rule->obj != NULL) { + pthread_mutex_lock(&app->sock_lock); + ret = ustctl_release_object(sock, ua_event_notifier_rule->obj); + pthread_mutex_unlock(&app->sock_lock); + if (ret < 0 && ret != -EPIPE && ret != -LTTNG_UST_ERR_EXITING) { + ERR("Failed to release event notifier object: app = '%s' (ppid %d), ret = %d", + app->name, (int) app->ppid, ret); + } + + free(ua_event_notifier_rule->obj); + } + + lttng_trigger_put(ua_event_notifier_rule->trigger); + call_rcu(&ua_event_notifier_rule->rcu_head, + free_ust_app_event_notifier_rule_rcu); +} + /* * Release ust data object of the given stream. * @@ -346,6 +431,33 @@ static int release_ust_app_stream(int sock, struct ust_app_stream *stream, return ret; } +/* + * Release ust data object of the given map_counter. + * + * Return 0 on success or else a negative value. + */ +static int release_ust_app_map_counter(int sock, struct ust_app_map_counter *map_counter, + struct ust_app *app) +{ + int ret = 0; + + assert(map_counter); + + if (map_counter->obj) { + pthread_mutex_lock(&app->sock_lock); + ret = ustctl_release_object(sock, map_counter->obj); + pthread_mutex_unlock(&app->sock_lock); + if (ret < 0 && ret != -EPIPE && ret != -LTTNG_UST_ERR_EXITING) { + ERR("UST app sock %d release map_counter obj failed with ret %d", + sock, ret); + } + lttng_fd_put(LTTNG_FD_APPS, 2); + free(map_counter->obj); + } + + return ret; +} + /* * Delete ust app stream safely. RCU read lock must be held before calling * this function. @@ -360,6 +472,20 @@ void delete_ust_app_stream(int sock, struct ust_app_stream *stream, free(stream); } +/* + * Delete ust app map_counter safely. RCU read lock must be held before calling + * this function. + */ +static +void delete_ust_app_map_counter(int sock, struct ust_app_map_counter *map_counter, + struct ust_app *app) +{ + assert(map_counter); + + (void) release_ust_app_map_counter(sock, map_counter, app); + free(map_counter); +} + /* * We need to execute ht_destroy outside of RCU read-side critical * section and outside of call_rcu thread, so we postpone its execution @@ -377,6 +503,22 @@ void delete_ust_app_channel_rcu(struct rcu_head *head) free(ua_chan); } +/* + * We need to execute ht_destroy outside of RCU read-side critical + * section and outside of call_rcu thread, so we postpone its execution + * using ht_cleanup_push. It is simpler than to change the semantic of + * the many callers of delete_ust_app_session(). + */ +static +void delete_ust_app_map_rcu(struct rcu_head *head) +{ + struct ust_app_map *ua_map = + caa_container_of(head, struct ust_app_map, rcu_head); + + ht_cleanup_push(ua_map->events); + free(ua_map); +} + /* * Extract the lost packet or discarded events counter when the channel is * being deleted and store the value in the parent channel so we can @@ -506,7 +648,7 @@ void delete_ust_app_channel(int sock, struct ust_app_channel *ua_chan, if (ua_chan->obj != NULL) { /* Remove channel from application UST object descriptor. */ iter.iter.node = &ua_chan->ust_objd_node.node; - ret = lttng_ht_del(app->ust_objd, &iter); + ret = lttng_ht_del(app->ust_chan_objd, &iter); assert(!ret); pthread_mutex_lock(&app->sock_lock); ret = ustctl_release_object(sock, ua_chan->obj); @@ -521,6 +663,123 @@ void delete_ust_app_channel(int sock, struct ust_app_channel *ua_chan, call_rcu(&ua_chan->rcu_head, delete_ust_app_channel_rcu); } +static +void copy_ust_app_map_values(int sock, struct ust_app_map *ua_map, + struct ust_app *app) +{ + struct ltt_ust_map_dead_pid_kv_values *kv_pair_list = ua_map->dead_app_kv_values; + struct ust_registry_session *ust_reg_sess; + struct ust_registry_map *ust_reg_map; + struct ust_registry_map_index_ht_entry *map_index_entry; + struct lttng_ht_iter key_iter; + struct lttng_ht *dead_app_kv_values; + + assert(app->buffer_type == LTTNG_BUFFER_PER_PID); + ust_reg_sess = get_session_registry(ua_map->session); + + pthread_mutex_lock(&ust_reg_sess->lock); + ust_reg_map = ust_registry_map_find(ust_reg_sess, ua_map->key); + pthread_mutex_unlock(&ust_reg_sess->lock); + assert(ust_reg_map); + + DBG("Aggregating dead map values"); + + pthread_mutex_lock(&kv_pair_list->lock); + + if (app->bits_per_long == 32) { + dead_app_kv_values = kv_pair_list->dead_app_kv_values_32bits; + } else { + dead_app_kv_values = kv_pair_list->dead_app_kv_values_64bits; + } + + /* Iterate over all the formated_key -> counter index */ + cds_lfht_for_each_entry(ust_reg_map->key_string_to_bucket_index_ht->ht, + &key_iter.iter, map_index_entry, node.node) { + bool overflow = 0, underflow = 0; + int64_t local_value = 0; + int ret; + size_t dimension_indexes[1] = {map_index_entry->index}; + + ret = ustctl_counter_aggregate(ua_map->map_handle, + dimension_indexes, &local_value, &overflow, + &underflow); + if (ret) { + ERR("Error getting counter value from the tracer: key = '%s'", + map_index_entry->formated_key); + ret = -1; + goto end; + } + + map_add_or_increment_map_values(dead_app_kv_values, + map_index_entry->formated_key, local_value, + underflow, overflow); + + } + +end: + pthread_mutex_unlock(&kv_pair_list->lock); + return; +} +/* + * Delete ust app map safely. RCU read lock must be held before calling + * this function. + * + * The session list lock must be held by the caller. + */ +static +void delete_ust_app_map(int sock, struct ust_app_map *ua_map, + struct ust_app *app) +{ + int ret; + struct lttng_ht_iter iter; + struct ust_app_event *ua_event; + struct ust_app_map_counter *map_counter, *ctmp; + struct ust_registry_session *registry; + + assert(ua_map); + + DBG3("UST app deleting map %s", ua_map->name); + + /* Wipe stream */ + cds_list_for_each_entry_safe(map_counter, ctmp, &ua_map->counters.head, list) { + cds_list_del(&map_counter->list); + delete_ust_app_map_counter(sock, map_counter, app); + } + + /* Wipe events */ + cds_lfht_for_each_entry(ua_map->events->ht, &iter.iter, ua_event, + node.node) { + ret = lttng_ht_del(ua_map->events, &iter); + assert(!ret); + delete_ust_app_event(sock, ua_event, app); + } + + if (ua_map->session->buffer_type == LTTNG_BUFFER_PER_PID) { + /* Wipe and free registry from session registry. */ + registry = get_session_registry(ua_map->session); + if (registry) { + ust_registry_map_del_free(registry, ua_map->key); + } + } + + if (ua_map->obj != NULL) { + /* Remove map from application UST object descriptor. */ + iter.iter.node = &ua_map->ust_objd_node.node; + ret = lttng_ht_del(app->ust_map_objd, &iter); + assert(!ret); + pthread_mutex_lock(&app->sock_lock); + ret = ustctl_release_object(sock, ua_map->obj); + pthread_mutex_unlock(&app->sock_lock); + if (ret < 0 && ret != -EPIPE && ret != -LTTNG_UST_ERR_EXITING) { + ERR("UST app sock %d release map obj failed with ret %d", + sock, ret); + } + lttng_fd_put(LTTNG_FD_APPS, 1); + free(ua_map->obj); + } + call_rcu(&ua_map->rcu_head, delete_ust_app_map_rcu); +} + int ust_app_register_done(struct ust_app *app) { int ret; @@ -809,6 +1068,7 @@ void delete_ust_app_session_rcu(struct rcu_head *head) caa_container_of(head, struct ust_app_session, rcu_head); ht_cleanup_push(ua_sess->channels); + ht_cleanup_push(ua_sess->maps); free(ua_sess); } @@ -825,6 +1085,7 @@ void delete_ust_app_session(int sock, struct ust_app_session *ua_sess, int ret; struct lttng_ht_iter iter; struct ust_app_channel *ua_chan; + struct ust_app_map *ua_map; struct ust_registry_session *registry; assert(ua_sess); @@ -859,6 +1120,16 @@ void delete_ust_app_session(int sock, struct ust_app_session *ua_sess, delete_ust_app_channel(sock, ua_chan, app); } + cds_lfht_for_each_entry(ua_sess->maps->ht, &iter.iter, ua_map, + node.node) { + if (ua_sess->buffer_type == LTTNG_BUFFER_PER_PID) { + copy_ust_app_map_values(sock, ua_map, app); + } + ret = lttng_ht_del(ua_sess->maps, &iter); + assert(!ret); + delete_ust_app_map(sock, ua_map, app); + } + /* In case of per PID, the registry is kept in the session. */ if (ua_sess->buffer_type == LTTNG_BUFFER_PER_PID) { struct buffer_reg_pid *reg_pid = buffer_reg_pid_find(ua_sess->id); @@ -904,6 +1175,9 @@ void delete_ust_app(struct ust_app *app) { int ret, sock; struct ust_app_session *ua_sess, *tmp_ua_sess; + struct lttng_ht_iter iter; + struct ust_app_event_notifier_rule *event_notifier_rule; + bool event_notifier_write_fd_is_open; /* * The session list lock must be held during this function to guarantee @@ -923,9 +1197,62 @@ void delete_ust_app(struct ust_app *app) rcu_read_unlock(); } + /* Remove the event notifier rules associated with this app. */ + rcu_read_lock(); + cds_lfht_for_each_entry (app->token_to_event_notifier_rule_ht->ht, + &iter.iter, event_notifier_rule, node.node) { + ret = lttng_ht_del(app->token_to_event_notifier_rule_ht, &iter); + assert(!ret); + + delete_ust_app_event_notifier_rule( + app->sock, event_notifier_rule, app); + } + + rcu_read_unlock(); + ht_cleanup_push(app->sessions); ht_cleanup_push(app->ust_sessions_objd); - ht_cleanup_push(app->ust_objd); + ht_cleanup_push(app->ust_chan_objd); + ht_cleanup_push(app->ust_map_objd); + ht_cleanup_push(app->token_to_event_notifier_rule_ht); + + /* + * This could be NULL if the event notifier setup failed (e.g the app + * was killed or the tracer does not support this feature). + */ + if (app->event_notifier_group.object) { + enum lttng_error_code ret_code; + enum event_notifier_error_accounting_status status; + + const int event_notifier_read_fd = lttng_pipe_get_readfd( + app->event_notifier_group.event_pipe); + + ret_code = notification_thread_command_remove_tracer_event_source( + notification_thread_handle, + event_notifier_read_fd); + if (ret_code != LTTNG_OK) { + ERR("Failed to remove application tracer event source from notification thread"); + } + + status = event_notifier_error_accounting_unregister_app(app); + if (status != EVENT_NOTIFIER_ERROR_ACCOUNTING_STATUS_OK) { + ERR("Error unregistering app from event notifier error accounting"); + } + + ustctl_release_object(sock, app->event_notifier_group.object); + free(app->event_notifier_group.object); + } + + event_notifier_write_fd_is_open = lttng_pipe_is_write_open( + app->event_notifier_group.event_pipe); + lttng_pipe_destroy(app->event_notifier_group.event_pipe); + /* + * Release the file descriptors reserved for the event notifier pipe. + * The app could be destroyed before the write end of the pipe could be + * passed to the application (and closed). In that case, both file + * descriptors must be released. + */ + lttng_fd_put(LTTNG_FD_APPS, event_notifier_write_fd_is_open ? 2 : 1); /* * Wait until we have deleted the application from the sock hash table @@ -1013,6 +1340,7 @@ struct ust_app_session *alloc_ust_app_session(void) ua_sess->handle = -1; ua_sess->channels = lttng_ht_new(0, LTTNG_HT_TYPE_STRING); ua_sess->metadata_attr.type = LTTNG_UST_CHAN_METADATA; + ua_sess->maps = lttng_ht_new(0, LTTNG_HT_TYPE_STRING); pthread_mutex_init(&ua_sess->lock, NULL); return ua_sess; @@ -1025,7 +1353,7 @@ error_free: * Alloc new UST app channel. */ static -struct ust_app_channel *alloc_ust_app_channel(char *name, +struct ust_app_channel *alloc_ust_app_channel(const char *name, struct ust_app_session *ua_sess, struct lttng_ust_channel_attr *attr) { @@ -1075,6 +1403,43 @@ error: return NULL; } +/* + * Alloc new UST app map. + */ +static +struct ust_app_map *alloc_ust_app_map(const char *name, + struct ust_app_session *ua_sess) +{ + struct ust_app_map *ua_map; + + /* Init most of the default value by allocating and zeroing */ + ua_map = zmalloc(sizeof(struct ust_app_map)); + if (ua_map == NULL) { + PERROR("malloc"); + goto error; + } + + /* Setup map name */ + strncpy(ua_map->name, name, sizeof(ua_map->name)); + ua_map->name[sizeof(ua_map->name) - 1] = '\0'; + + ua_map->enabled = 1; + ua_map->handle = -1; + ua_map->session = ua_sess; + ua_map->key = get_next_map_key(); + ua_map->events = lttng_ht_new(0, LTTNG_HT_TYPE_STRING); + lttng_ht_node_init_str(&ua_map->node, ua_map->name); + + CDS_INIT_LIST_HEAD(&ua_map->counters.head); + + DBG3("UST app map %s allocated", ua_map->name); + + return ua_map; + +error: + return NULL; +} + /* * Allocate and initialize a UST app stream. * @@ -1097,6 +1462,28 @@ error: return stream; } +/* + * Allocate and initialize a UST app map_counter. + * + * Return newly allocated map_counter pointer or NULL on error. + */ +struct ust_app_map_counter *ust_app_alloc_map_counter(void) +{ + struct ust_app_map_counter *map_counter = NULL; + + map_counter = zmalloc(sizeof(*map_counter)); + if (map_counter == NULL) { + PERROR("zmalloc ust app map_counter"); + goto error; + } + + /* Zero could be a valid value for a handle so flag it to -1. */ + map_counter->handle = -1; + +error: + return map_counter; +} + /* * Alloc new UST app event. */ @@ -1131,6 +1518,66 @@ error: return NULL; } + +/* + * Allocate a new UST app event notifier rule. + */ +static struct ust_app_event_notifier_rule *alloc_ust_app_event_notifier_rule( + struct lttng_trigger *trigger) +{ + enum lttng_event_rule_generate_exclusions_status + generate_exclusion_status; + struct ust_app_event_notifier_rule *ua_event_notifier_rule; + struct lttng_condition *condition = NULL; + const struct lttng_event_rule *event_rule = NULL; + + ua_event_notifier_rule = zmalloc(sizeof(struct ust_app_event_notifier_rule)); + if (ua_event_notifier_rule == NULL) { + PERROR("Failed to allocate ust_app_event_notifier_rule structure"); + goto error; + } + + ua_event_notifier_rule->enabled = 1; + ua_event_notifier_rule->token = lttng_trigger_get_tracer_token(trigger); + lttng_ht_node_init_u64(&ua_event_notifier_rule->node, + ua_event_notifier_rule->token); + + condition = lttng_trigger_get_condition(trigger); + assert(condition); + assert(lttng_condition_get_type(condition) == LTTNG_CONDITION_TYPE_ON_EVENT); + + assert(LTTNG_CONDITION_STATUS_OK == lttng_condition_on_event_get_rule(condition, &event_rule)); + assert(event_rule); + + /* Acquire the event notifier's reference to the trigger. */ + lttng_trigger_get(trigger); + + ua_event_notifier_rule->trigger = trigger; + ua_event_notifier_rule->filter = lttng_event_rule_get_filter_bytecode(event_rule); + generate_exclusion_status = lttng_event_rule_generate_exclusions( + event_rule, &ua_event_notifier_rule->exclusion); + switch (generate_exclusion_status) { + case LTTNG_EVENT_RULE_GENERATE_EXCLUSIONS_STATUS_OK: + case LTTNG_EVENT_RULE_GENERATE_EXCLUSIONS_STATUS_NONE: + break; + default: + /* Error occured. */ + ERR("Failed to generate exclusions from trigger while allocating an event notifier rule"); + goto error_put_trigger; + } + + DBG3("UST app event notifier rule allocated: token = %" PRIu64, + ua_event_notifier_rule->token); + + return ua_event_notifier_rule; + +error_put_trigger: + lttng_trigger_put(trigger); +error: + free(ua_event_notifier_rule); + return NULL; +} + /* * Alloc new UST app context. */ @@ -1149,7 +1596,7 @@ struct ust_app_ctx *alloc_ust_app_ctx(struct lttng_ust_context_attr *uctx) if (uctx) { memcpy(&ua_ctx->ctx, uctx, sizeof(ua_ctx->ctx)); if (uctx->ctx == LTTNG_UST_CONTEXT_APP_CONTEXT) { - char *provider_name = NULL, *ctx_name = NULL; + char *provider_name = NULL, *ctx_name = NULL; provider_name = strdup(uctx->u.app_ctx.provider_name); ctx_name = strdup(uctx->u.app_ctx.ctx_name); @@ -1172,50 +1619,51 @@ error: } /* - * Allocate a filter and copy the given original filter. + * Create a liblttng-ust filter bytecode from given bytecode. * * Return allocated filter or NULL on error. */ -static struct lttng_filter_bytecode *copy_filter_bytecode( - struct lttng_filter_bytecode *orig_f) +static struct lttng_ust_filter_bytecode * +create_ust_filter_bytecode_from_bytecode(const struct lttng_bytecode *orig_f) { - struct lttng_filter_bytecode *filter = NULL; + struct lttng_ust_filter_bytecode *filter = NULL; - /* Copy filter bytecode */ + /* Copy filter bytecode. */ filter = zmalloc(sizeof(*filter) + orig_f->len); if (!filter) { - PERROR("zmalloc alloc filter bytecode"); + PERROR("Failed to allocate lttng_ust_filter_bytecode: bytecode len = %" PRIu32 " bytes", orig_f->len); goto error; } + assert(sizeof(struct lttng_bytecode) == + sizeof(struct lttng_ust_filter_bytecode)); memcpy(filter, orig_f, sizeof(*filter) + orig_f->len); - error: return filter; } /* - * Create a liblttng-ust filter bytecode from given bytecode. + * Create a liblttng-ust capture bytecode from given bytecode. * * Return allocated filter or NULL on error. */ -static struct lttng_ust_filter_bytecode *create_ust_bytecode_from_bytecode( - struct lttng_filter_bytecode *orig_f) +static struct lttng_ust_capture_bytecode * +create_ust_capture_bytecode_from_bytecode(const struct lttng_bytecode *orig_f) { - struct lttng_ust_filter_bytecode *filter = NULL; + struct lttng_ust_capture_bytecode *capture = NULL; - /* Copy filter bytecode */ - filter = zmalloc(sizeof(*filter) + orig_f->len); - if (!filter) { - PERROR("zmalloc alloc ust filter bytecode"); + /* Copy capture bytecode. */ + capture = zmalloc(sizeof(*capture) + orig_f->len); + if (!capture) { + PERROR("Failed to allocate lttng_ust_capture_bytecode: bytecode len = %" PRIu32 " bytes", orig_f->len); goto error; } - assert(sizeof(struct lttng_filter_bytecode) == - sizeof(struct lttng_ust_filter_bytecode)); - memcpy(filter, orig_f, sizeof(*filter) + orig_f->len); + assert(sizeof(struct lttng_bytecode) == + sizeof(struct lttng_ust_capture_bytecode)); + memcpy(capture, orig_f, sizeof(*capture) + orig_f->len); error: - return filter; + return capture; } /* @@ -1270,9 +1718,10 @@ error: * Return an ust_app_event object or NULL on error. */ static struct ust_app_event *find_ust_app_event(struct lttng_ht *ht, - const char *name, const struct lttng_filter_bytecode *filter, + const char *name, const struct lttng_bytecode *filter, int loglevel_value, - const struct lttng_event_exclusion *exclusion) + const struct lttng_event_exclusion *exclusion, + uint64_t tracer_token) { struct lttng_ht_iter iter; struct lttng_ht_node_str *node; @@ -1288,6 +1737,7 @@ static struct ust_app_event *find_ust_app_event(struct lttng_ht *ht, key.loglevel_type = loglevel_value; /* lttng_event_exclusion and lttng_ust_event_exclusion structures are similar */ key.exclusion = exclusion; + key.tracer_token = tracer_token; /* Lookup using the event name as hash and a custom match fct. */ cds_lfht_lookup(ht->ht, ht->hash_fct((void *) name, lttng_ht_seed), @@ -1304,23 +1754,52 @@ end: } /* - * Create the channel context on the tracer. + * Look-up an event notifier rule based on its token id. * - * Called with UST app session lock held. + * Must be called with the RCU read lock held. + * Return an ust_app_event_notifier_rule object or NULL on error. */ -static -int create_ust_channel_context(struct ust_app_channel *ua_chan, - struct ust_app_ctx *ua_ctx, struct ust_app *app) +static struct ust_app_event_notifier_rule *find_ust_app_event_notifier_rule( + struct lttng_ht *ht, uint64_t token) { - int ret; + struct lttng_ht_iter iter; + struct lttng_ht_node_u64 *node; + struct ust_app_event_notifier_rule *event_notifier_rule = NULL; - health_code_update(); + assert(ht); - pthread_mutex_lock(&app->sock_lock); - ret = ustctl_add_context(app->sock, &ua_ctx->ctx, - ua_chan->obj, &ua_ctx->obj); - pthread_mutex_unlock(&app->sock_lock); - if (ret < 0) { + lttng_ht_lookup(ht, &token, &iter); + node = lttng_ht_iter_get_node_u64(&iter); + if (node == NULL) { + DBG2("UST app event notifier rule token not found: token = %" PRIu64, + token); + goto end; + } + + event_notifier_rule = caa_container_of( + node, struct ust_app_event_notifier_rule, node); +end: + return event_notifier_rule; +} + +/* + * Create the channel context on the tracer. + * + * Called with UST app session lock held. + */ +static +int create_ust_channel_context(struct ust_app_channel *ua_chan, + struct ust_app_ctx *ua_ctx, struct ust_app *app) +{ + int ret; + + health_code_update(); + + pthread_mutex_lock(&app->sock_lock); + ret = ustctl_add_context(app->sock, &ua_ctx->ctx, + ua_chan->obj, &ua_ctx->obj); + pthread_mutex_unlock(&app->sock_lock); + if (ret < 0) { if (ret != -EPIPE && ret != -LTTNG_UST_ERR_EXITING) { ERR("UST app create channel context failed for app (pid: %d) " "with ret %d", app->pid, ret); @@ -1349,33 +1828,82 @@ error: /* * Set the filter on the tracer. */ -static -int set_ust_event_filter(struct ust_app_event *ua_event, - struct ust_app *app) +static int set_ust_object_filter(struct ust_app *app, + const struct lttng_bytecode *bytecode, + struct lttng_ust_object_data *ust_object) { int ret; struct lttng_ust_filter_bytecode *ust_bytecode = NULL; health_code_update(); - if (!ua_event->filter) { - ret = 0; + ust_bytecode = create_ust_filter_bytecode_from_bytecode(bytecode); + if (!ust_bytecode) { + ret = -LTTNG_ERR_NOMEM; + goto error; + } + pthread_mutex_lock(&app->sock_lock); + ret = ustctl_set_filter(app->sock, ust_bytecode, + ust_object); + pthread_mutex_unlock(&app->sock_lock); + if (ret < 0) { + if (ret != -EPIPE && ret != -LTTNG_UST_ERR_EXITING) { + ERR("UST app set object filter failed: object = %p of app pid = %d, ret = %d", + ust_object, app->pid, ret); + } else { + /* + * This is normal behavior, an application can die during the + * creation process. Don't report an error so the execution can + * continue normally. + */ + ret = 0; + DBG3("Failed to set UST app object filter. Application is dead."); + } goto error; } - ust_bytecode = create_ust_bytecode_from_bytecode(ua_event->filter); + DBG2("UST filter successfully set: object = %p", ust_object); + +error: + health_code_update(); + free(ust_bytecode); + return ret; +} + +/* + * Set a capture bytecode for the passed object. + * The sequence number enforces the ordering at runtime and on reception of + * the captured payloads. + */ +static int set_ust_capture(struct ust_app *app, + const struct lttng_bytecode *bytecode, + unsigned int capture_seqnum, + struct lttng_ust_object_data *ust_object) +{ + int ret; + struct lttng_ust_capture_bytecode *ust_bytecode = NULL; + + health_code_update(); + + ust_bytecode = create_ust_capture_bytecode_from_bytecode(bytecode); if (!ust_bytecode) { ret = -LTTNG_ERR_NOMEM; goto error; } + + /* + * Set the sequence number to ensure the capture of fields is ordered. + */ + ust_bytecode->seqnum = capture_seqnum; + pthread_mutex_lock(&app->sock_lock); - ret = ustctl_set_filter(app->sock, ust_bytecode, - ua_event->obj); + ret = ustctl_set_capture(app->sock, ust_bytecode, + ust_object); pthread_mutex_unlock(&app->sock_lock); if (ret < 0) { if (ret != -EPIPE && ret != -LTTNG_UST_ERR_EXITING) { - ERR("UST app event %s filter failed for app (pid: %d) " - "with ret %d", ua_event->attr.name, app->pid, ret); + ERR("UST app set object capture failed: object = %p of app pid = %d, ret = %d", + ust_object, app->pid, ret); } else { /* * This is normal behavior, an application can die during the @@ -1383,12 +1911,13 @@ int set_ust_event_filter(struct ust_app_event *ua_event, * continue normally. */ ret = 0; - DBG3("UST app filter event failed. Application is dead."); + DBG3("Failed to set UST app object capture. Application is dead."); } + goto error; } - DBG2("UST filter set successfully for event %s", ua_event->name); + DBG2("UST capture successfully set: object = %p", ust_object); error: health_code_update(); @@ -1398,7 +1927,7 @@ error: static struct lttng_ust_event_exclusion *create_ust_exclusion_from_exclusion( - struct lttng_event_exclusion *exclusion) + const struct lttng_event_exclusion *exclusion) { struct lttng_ust_event_exclusion *ust_exclusion = NULL; size_t exclusion_alloc_size = sizeof(struct lttng_ust_event_exclusion) + @@ -1420,33 +1949,30 @@ end: /* * Set event exclusions on the tracer. */ -static -int set_ust_event_exclusion(struct ust_app_event *ua_event, - struct ust_app *app) +static int set_ust_object_exclusions(struct ust_app *app, + const struct lttng_event_exclusion *exclusions, + struct lttng_ust_object_data *ust_object) { int ret; - struct lttng_ust_event_exclusion *ust_exclusion = NULL; + struct lttng_ust_event_exclusion *ust_exclusions = NULL; - health_code_update(); + assert(exclusions && exclusions->count > 0); - if (!ua_event->exclusion || !ua_event->exclusion->count) { - ret = 0; - goto error; - } + health_code_update(); - ust_exclusion = create_ust_exclusion_from_exclusion( - ua_event->exclusion); - if (!ust_exclusion) { + ust_exclusions = create_ust_exclusion_from_exclusion( + exclusions); + if (!ust_exclusions) { ret = -LTTNG_ERR_NOMEM; goto error; } pthread_mutex_lock(&app->sock_lock); - ret = ustctl_set_exclusion(app->sock, ust_exclusion, ua_event->obj); + ret = ustctl_set_exclusion(app->sock, ust_exclusions, ust_object); pthread_mutex_unlock(&app->sock_lock); if (ret < 0) { if (ret != -EPIPE && ret != -LTTNG_UST_ERR_EXITING) { - ERR("UST app event %s exclusions failed for app (pid: %d) " - "with ret %d", ua_event->attr.name, app->pid, ret); + ERR("Failed to set UST app exclusions for object %p of app (pid: %d) " + "with ret %d", ust_object, app->pid, ret); } else { /* * This is normal behavior, an application can die during the @@ -1454,37 +1980,36 @@ int set_ust_event_exclusion(struct ust_app_event *ua_event, * continue normally. */ ret = 0; - DBG3("UST app event exclusion failed. Application is dead."); + DBG3("Failed to set UST app object exclusions. Application is dead."); } goto error; } - DBG2("UST exclusion set successfully for event %s", ua_event->name); + DBG2("UST exclusions set successfully for object %p", ust_object); error: health_code_update(); - free(ust_exclusion); + free(ust_exclusions); return ret; } /* * Disable the specified event on to UST tracer for the UST session. */ -static int disable_ust_event(struct ust_app *app, - struct ust_app_session *ua_sess, struct ust_app_event *ua_event) +static int disable_ust_object(struct ust_app *app, + struct lttng_ust_object_data *object) { int ret; health_code_update(); pthread_mutex_lock(&app->sock_lock); - ret = ustctl_disable(app->sock, ua_event->obj); + ret = ustctl_disable(app->sock, object); pthread_mutex_unlock(&app->sock_lock); if (ret < 0) { if (ret != -EPIPE && ret != -LTTNG_UST_ERR_EXITING) { - ERR("UST app event %s disable failed for app (pid: %d) " - "and session handle %d with ret %d", - ua_event->attr.name, app->pid, ua_sess->handle, ret); + ERR("Failed to disable UST app object %p app (pid: %d) with ret %d", + object, app->pid, ret); } else { /* * This is normal behavior, an application can die during the @@ -1492,13 +2017,13 @@ static int disable_ust_event(struct ust_app *app, * continue normally. */ ret = 0; - DBG3("UST app disable event failed. Application is dead."); + DBG3("Failed to disable UST app object. Application is dead."); } goto error; } - DBG2("UST app event %s disabled successfully for app (pid: %d)", - ua_event->attr.name, app->pid); + DBG2("UST app object %p disabled successfully for app (pid: %d)", + object, app->pid); error: health_code_update(); @@ -1543,6 +2068,44 @@ error: return ret; } +/* + * Disable the specified map on to UST tracer for the UST session. + */ +static int disable_ust_map(struct ust_app *app, + struct ust_app_session *ua_sess, struct ust_app_map *ua_map) +{ + int ret; + + health_code_update(); + + pthread_mutex_lock(&app->sock_lock); + ret = ustctl_disable(app->sock, ua_map->obj); + pthread_mutex_unlock(&app->sock_lock); + if (ret < 0) { + if (ret != -EPIPE && ret != -LTTNG_UST_ERR_EXITING) { + ERR("UST app map %s disable failed for app (pid: %d) " + "and session handle %d with ret %d", + ua_map->name, app->pid, ua_sess->handle, ret); + } else { + /* + * This is normal behavior, an application can die during the + * creation process. Don't report an error so the execution can + * continue normally. + */ + ret = 0; + DBG3("UST app disable map failed. Application is dead."); + } + goto error; + } + + DBG2("UST app map %s disabled successfully for app (pid: %d)", + ua_map->name, app->pid); + +error: + health_code_update(); + return ret; +} + /* * Enable the specified channel on to UST tracer for the UST session. */ @@ -1584,23 +2147,62 @@ error: } /* - * Enable the specified event on to UST tracer for the UST session. + * Enable the specified map on to UST tracer for the UST session. */ -static int enable_ust_event(struct ust_app *app, - struct ust_app_session *ua_sess, struct ust_app_event *ua_event) +static int enable_ust_map(struct ust_app *app, + struct ust_app_session *ua_sess, struct ust_app_map *ua_map) { int ret; health_code_update(); pthread_mutex_lock(&app->sock_lock); - ret = ustctl_enable(app->sock, ua_event->obj); + ret = ustctl_enable(app->sock, ua_map->obj); pthread_mutex_unlock(&app->sock_lock); if (ret < 0) { if (ret != -EPIPE && ret != -LTTNG_UST_ERR_EXITING) { - ERR("UST app event %s enable failed for app (pid: %d) " + ERR("UST app map %s enable failed for app (pid: %d) " "and session handle %d with ret %d", - ua_event->attr.name, app->pid, ua_sess->handle, ret); + ua_map->name, app->pid, ua_sess->handle, ret); + } else { + /* + * This is normal behavior, an application can die during the + * creation process. Don't report an error so the execution can + * continue normally. + */ + ret = 0; + DBG3("UST app enable map failed. Application is dead."); + } + goto error; + } + + ua_map->enabled = 1; + + DBG2("UST app map %s enabled successfully for app (pid: %d)", + ua_map->name, app->pid); + +error: + health_code_update(); + return ret; +} + +/* + * Enable the specified event on to UST tracer for the UST session. + */ +static int enable_ust_object( + struct ust_app *app, struct lttng_ust_object_data *ust_object) +{ + int ret; + + health_code_update(); + + pthread_mutex_lock(&app->sock_lock); + ret = ustctl_enable(app->sock, ust_object); + pthread_mutex_unlock(&app->sock_lock); + if (ret < 0) { + if (ret != -EPIPE && ret != -LTTNG_UST_ERR_EXITING) { + ERR("UST app enable failed for object %p app (pid: %d) with ret %d", + ust_object, app->pid, ret); } else { /* * This is normal behavior, an application can die during the @@ -1608,13 +2210,13 @@ static int enable_ust_event(struct ust_app *app, * continue normally. */ ret = 0; - DBG3("UST app enable event failed. Application is dead."); + DBG3("Failed to enable UST app object. Application is dead."); } goto error; } - DBG2("UST app event %s enabled successfully for app (pid: %d)", - ua_event->attr.name, app->pid); + DBG2("UST app object %p enabled successfully for app (pid: %d)", + ust_object, app->pid); error: health_code_update(); @@ -1673,13 +2275,64 @@ error: return ret; } +/* + * Send map and stream buffer to application. + * + * Return 0 on success. On error, a negative value is returned. + */ +static int send_map_pid_to_ust(struct ust_app *app, + struct ust_app_session *ua_sess, struct ust_app_map *ua_map) +{ + int ret; + struct ust_app_map_counter *counter, *ctmp; + + assert(app); + assert(ua_sess); + assert(ua_map); + + health_code_update(); + + DBG("UST app sending map %s to UST app sock %d", ua_map->name, + app->sock); + + /* Send map to the application. */ + pthread_mutex_lock(&app->sock_lock); + ret = ustctl_send_counter_data_to_ust(app->sock, + ua_sess->handle, ua_map->obj); + pthread_mutex_unlock(&app->sock_lock); + assert(ret == 0); + + ua_map->handle = ua_map->obj->handle; + + health_code_update(); + + /* Send all streams to application. */ + cds_list_for_each_entry_safe(counter, ctmp, &ua_map->counters.head, list) { + pthread_mutex_lock(&app->sock_lock); + // Do send the per cpu counter here + ret = ustctl_send_counter_cpu_data_to_ust(app->sock, + ua_map->obj, counter->obj); + pthread_mutex_unlock(&app->sock_lock); + assert(ret == 0); + + /* We don't need the stream anymore once sent to the tracer. */ + cds_list_del(&counter->list); + delete_ust_app_map_counter(-1, counter, app); + } + /* Flag the map that it is sent to the application. */ + ua_map->is_sent = 1; + + health_code_update(); + return ret; +} + /* * Create the specified event onto the UST tracer for a UST session. * * Should be called with session mutex held. */ static -int create_ust_event(struct ust_app *app, struct ust_app_session *ua_sess, +int create_ust_channel_event(struct ust_app *app, struct ust_app_session *ua_sess, struct ust_app_channel *ua_chan, struct ust_app_event *ua_event) { int ret = 0; @@ -1710,14 +2363,14 @@ int create_ust_event(struct ust_app *app, struct ust_app_session *ua_sess, ua_event->handle = ua_event->obj->handle; - DBG2("UST app event %s created successfully for pid:%d", - ua_event->attr.name, app->pid); + DBG2("UST app event %s created successfully for pid:%d object: %p", + ua_event->attr.name, app->pid, ua_event->obj); health_code_update(); /* Set filter if one is present. */ if (ua_event->filter) { - ret = set_ust_event_filter(ua_event, app); + ret = set_ust_object_filter(app, ua_event->filter, ua_event->obj); if (ret < 0) { goto error; } @@ -1725,7 +2378,7 @@ int create_ust_event(struct ust_app *app, struct ust_app_session *ua_sess, /* Set exclusions for the event */ if (ua_event->exclusion) { - ret = set_ust_event_exclusion(ua_event, app); + ret = set_ust_object_exclusions(app, ua_event->exclusion, ua_event->obj); if (ret < 0) { goto error; } @@ -1737,7 +2390,7 @@ int create_ust_event(struct ust_app *app, struct ust_app_session *ua_sess, * We now need to explicitly enable the event, since it * is now disabled at creation. */ - ret = enable_ust_event(app, ua_sess, ua_event); + ret = enable_ust_object(app, ua_event->obj); if (ret < 0) { /* * If we hit an EPERM, something is wrong with our enable call. If @@ -1764,99 +2417,467 @@ error: return ret; } -/* - * Copy data between an UST app event and a LTT event. - */ -static void shadow_copy_event(struct ust_app_event *ua_event, - struct ltt_ust_event *uevent) +static +void add_key_token(struct lttng_ust_key_token *ust_key_token, + const struct lttng_map_key_token *key_token) { - size_t exclusion_alloc_size; - - strncpy(ua_event->name, uevent->attr.name, sizeof(ua_event->name)); - ua_event->name[sizeof(ua_event->name) - 1] = '\0'; - - ua_event->enabled = uevent->enabled; + switch (key_token->type) { + case LTTNG_MAP_KEY_TOKEN_TYPE_STRING: + { + const struct lttng_map_key_token_string *str_token; + str_token = (typeof(str_token)) key_token; - /* Copy event attributes */ - memcpy(&ua_event->attr, &uevent->attr, sizeof(ua_event->attr)); + ust_key_token->type = LTTNG_UST_KEY_TOKEN_STRING; + strncpy(ust_key_token->arg.string, str_token->string, + LTTNG_UST_KEY_TOKEN_STRING_LEN_MAX); - /* Copy filter bytecode */ - if (uevent->filter) { - ua_event->filter = copy_filter_bytecode(uevent->filter); - /* Filter might be NULL here in case of ENONEM. */ + break; } - - /* Copy exclusion data */ - if (uevent->exclusion) { - exclusion_alloc_size = sizeof(struct lttng_event_exclusion) + - LTTNG_UST_SYM_NAME_LEN * uevent->exclusion->count; - ua_event->exclusion = zmalloc(exclusion_alloc_size); - if (ua_event->exclusion == NULL) { - PERROR("malloc"); - } else { - memcpy(ua_event->exclusion, uevent->exclusion, - exclusion_alloc_size); + case LTTNG_MAP_KEY_TOKEN_TYPE_VARIABLE: + { + const struct lttng_map_key_token_variable *var_token; + var_token = (typeof(var_token)) key_token; + switch (var_token->type) { + case LTTNG_MAP_KEY_TOKEN_VARIABLE_TYPE_EVENT_NAME: + ust_key_token->type = LTTNG_UST_KEY_TOKEN_EVENT_NAME; + break; + case LTTNG_MAP_KEY_TOKEN_VARIABLE_TYPE_PROVIDER_NAME: + ust_key_token->type = LTTNG_UST_KEY_TOKEN_PROVIDER_NAME; + break; + default: + abort(); } + + break; + } + default: + abort(); } } /* - * Copy data between an UST app channel and a LTT channel. + * Create the specified event onto the UST tracer for a UST session. + * + * Should be called with session mutex held. */ -static void shadow_copy_channel(struct ust_app_channel *ua_chan, - struct ltt_ust_channel *uchan) +static +int create_ust_map_event(struct ust_app *app, struct ust_app_session *ua_sess, + struct ust_app_map *ua_map, const struct lttng_map_key *key, + struct ust_app_event *ua_event) { - DBG2("UST app shadow copy of channel %s started", ua_chan->name); + int ret = 0; + unsigned int i, key_token_count; + enum lttng_map_key_status status; + struct lttng_ust_counter_event counter_event = {0}; - strncpy(ua_chan->name, uchan->name, sizeof(ua_chan->name)); - ua_chan->name[sizeof(ua_chan->name) - 1] = '\0'; + health_code_update(); - ua_chan->tracefile_size = uchan->tracefile_size; - ua_chan->tracefile_count = uchan->tracefile_count; + memcpy(&counter_event.event, &ua_event->attr, sizeof(struct lttng_ust_event)); - /* Copy event attributes since the layout is different. */ - ua_chan->attr.subbuf_size = uchan->attr.subbuf_size; - ua_chan->attr.num_subbuf = uchan->attr.num_subbuf; - ua_chan->attr.overwrite = uchan->attr.overwrite; - ua_chan->attr.switch_timer_interval = uchan->attr.switch_timer_interval; - ua_chan->attr.read_timer_interval = uchan->attr.read_timer_interval; - ua_chan->monitor_timer_interval = uchan->monitor_timer_interval; - ua_chan->attr.output = uchan->attr.output; - ua_chan->attr.blocking_timeout = uchan->attr.u.s.blocking_timeout; + status = lttng_map_key_get_token_count(key, &key_token_count); + if (status != LTTNG_MAP_KEY_STATUS_OK) { + ret = LTTNG_ERR_UNK; + goto error; + } - /* - * Note that the attribute channel type is not set since the channel on the - * tracing registry side does not have this information. - */ + assert(key_token_count > 0); - ua_chan->enabled = uchan->enabled; - ua_chan->tracing_channel_id = uchan->id; + counter_event.key.nr_dimensions = 1; + counter_event.key.key_dimensions[0].nr_key_tokens = key_token_count; - DBG3("UST app shadow copy of channel %s done", ua_chan->name); -} + if (key_token_count > LTTNG_UST_NR_KEY_TOKEN) { + ERR("Too many key tokens for UST tracer: token count = %u token count max =%u", + key_token_count, LTTNG_UST_NR_KEY_TOKEN); + ret = LTTNG_ERR_INVALID; + goto error; + } -/* - * Copy data between a UST app session and a regular LTT session. - */ -static void shadow_copy_session(struct ust_app_session *ua_sess, - struct ltt_ust_session *usess, struct ust_app *app) -{ - struct tm *timeinfo; - char datetime[16]; - int ret; - char tmp_shm_path[PATH_MAX]; + for (i = 0; i < key_token_count; i++) { + const struct lttng_map_key_token *token = + lttng_map_key_get_token_at_index(key, i); - timeinfo = localtime(&app->registration_time); - strftime(datetime, sizeof(datetime), "%Y%m%d-%H%M%S", timeinfo); + add_key_token(&counter_event.key.key_dimensions[0].key_tokens[i], + token); + } + + /* Create UST event on tracer */ + pthread_mutex_lock(&app->sock_lock); + ret = ustctl_counter_create_event(app->sock, &counter_event, ua_map->obj, + &ua_event->obj); + pthread_mutex_unlock(&app->sock_lock); + if (ret < 0) { + if (ret != -EPIPE && ret != -LTTNG_UST_ERR_EXITING) { + abort(); + ERR("Error ustctl counter create event %s for app pid: %d with ret %d", + ua_event->attr.name, app->pid, ret); + } else { + /* + * This is normal behavior, an application can die during the + * creation process. Don't report an error so the execution can + * continue normally. + */ + ret = 0; + DBG3("UST app counter create event failed. Application is dead."); + } + goto error; + } + + ua_event->handle = ua_event->obj->handle; + + DBG2("UST app map event %s created successfully for pid:%d object: %p", + ua_event->attr.name, app->pid, ua_event->obj); + + health_code_update(); + + /* Set filter if one is present. */ + if (ua_event->filter) { + ret = set_ust_object_filter(app, ua_event->filter, ua_event->obj); + if (ret < 0) { + goto error; + } + } + + /* Set exclusions for the event */ + if (ua_event->exclusion) { + ret = set_ust_object_exclusions(app, ua_event->exclusion, ua_event->obj); + if (ret < 0) { + goto error; + } + } + + /* If event not enabled, disable it on the tracer */ + if (ua_event->enabled) { + /* + * We now need to explicitly enable the event, since it + * is now disabled at creation. + */ + ret = enable_ust_object(app, ua_event->obj); + if (ret < 0) { + /* + * If we hit an EPERM, something is wrong with our enable call. If + * we get an EEXIST, there is a problem on the tracer side since we + * just created it. + */ + switch (ret) { + case -LTTNG_UST_ERR_PERM: + /* Code flow problem */ + assert(0); + case -LTTNG_UST_ERR_EXIST: + /* It's OK for our use case. */ + ret = 0; + break; + default: + break; + } + goto error; + } + } + +error: + health_code_update(); + return ret; +} + +static int init_ust_event_from_event_rule( + const struct lttng_event_rule *rule, + struct lttng_ust_event *event) +{ + enum lttng_event_rule_status status; + enum lttng_ust_loglevel_type ust_loglevel_type = LTTNG_UST_LOGLEVEL_ALL; + int loglevel = -1, ret = 0; + const char *pattern; + + /* For now only LTTNG_EVENT_RULE_TYPE_TRACEPOINT are supported. */ + assert(lttng_event_rule_get_type(rule) == + LTTNG_EVENT_RULE_TYPE_TRACEPOINT); + + if (lttng_event_rule_targets_agent_domain(rule)) { + /* + * Special event for agents + * The actual meat of the event is in the filter that will be + * attached later on. + * Set the default values for the agent event. + */ + pattern = event_get_default_agent_ust_name( + lttng_event_rule_get_domain_type(rule)); + loglevel = 0; + ust_loglevel_type = LTTNG_UST_LOGLEVEL_ALL; + } else { + const struct lttng_log_level_rule *log_level_rule; + + status = lttng_event_rule_tracepoint_get_pattern(rule, &pattern); + if (status != LTTNG_EVENT_RULE_STATUS_OK) { + /* At this point, this is a fatal error. */ + abort(); + } + + status = lttng_event_rule_tracepoint_get_log_level_rule( + rule, &log_level_rule); + if (status == LTTNG_EVENT_RULE_STATUS_UNSET) { + ust_loglevel_type = LTTNG_UST_LOGLEVEL_ALL; + } else if (status == LTTNG_EVENT_RULE_STATUS_OK) { + enum lttng_log_level_rule_status llr_status; + + switch (lttng_log_level_rule_get_type(log_level_rule)) { + case LTTNG_LOG_LEVEL_RULE_TYPE_EXACTLY: + ust_loglevel_type = LTTNG_UST_LOGLEVEL_SINGLE; + llr_status = lttng_log_level_rule_exactly_get_level(log_level_rule, &loglevel); + break; + case LTTNG_LOG_LEVEL_RULE_TYPE_AT_LEAST_AS_SEVERE_AS: + ust_loglevel_type = LTTNG_UST_LOGLEVEL_RANGE; + llr_status = lttng_log_level_rule_at_least_as_severe_as_get_level(log_level_rule, &loglevel); + break; + default: + abort(); + } + + assert(llr_status == LTTNG_LOG_LEVEL_RULE_STATUS_OK); + } else { + /* At this point this is a fatal error */ + assert(0); + } + } + + event->instrumentation = LTTNG_UST_TRACEPOINT; + ret = lttng_strncpy(event->name, pattern, + LTTNG_UST_SYM_NAME_LEN - 1); + if (ret) { + ERR("Failed to copy event rule pattern to notifier: pattern = '%s' ", + pattern); + goto end; + } + + event->loglevel_type = ust_loglevel_type; + event->loglevel = loglevel; +end: + return ret; +} + +/* + * Create the specified event notifier against the user space tracer of a + * given application. + */ +static int create_ust_event_notifier(struct ust_app *app, + struct ust_app_event_notifier_rule *ua_event_notifier_rule) +{ + int ret = 0; + struct lttng_ust_event_notifier event_notifier = {0}; + const struct lttng_condition *condition = NULL; + const struct lttng_event_rule *event_rule = NULL; + unsigned int capture_bytecode_count = 0, i; + enum lttng_condition_status cond_status; + + health_code_update(); + assert(app->event_notifier_group.object); + + condition = lttng_trigger_get_const_condition( + ua_event_notifier_rule->trigger); + assert(condition); + assert(lttng_condition_get_type(condition) == LTTNG_CONDITION_TYPE_ON_EVENT); + + lttng_condition_on_event_get_rule(condition, &event_rule); + assert(event_rule); + assert(lttng_event_rule_get_type(event_rule) == LTTNG_EVENT_RULE_TYPE_TRACEPOINT); + + init_ust_event_from_event_rule(event_rule, &event_notifier.event); + event_notifier.event.token = ua_event_notifier_rule->token; + event_notifier.error_counter_index = ua_event_notifier_rule->error_counter_index; + + /* Create UST event notifier against the tracer. */ + pthread_mutex_lock(&app->sock_lock); + ret = ustctl_create_event_notifier(app->sock, &event_notifier, + app->event_notifier_group.object, + &ua_event_notifier_rule->obj); + pthread_mutex_unlock(&app->sock_lock); + if (ret < 0) { + if (ret != -EPIPE && ret != -LTTNG_UST_ERR_EXITING) { + ERR("Error ustctl create event notifier: name = '%s', app = '%s' (ppid: %d), ret = %d", + event_notifier.event.name, app->name, + app->ppid, ret); + } else { + /* + * This is normal behavior, an application can die + * during the creation process. Don't report an error so + * the execution can continue normally. + */ + ret = 0; + DBG3("UST app create event notifier failed (application is dead): app = '%s' (ppid = %d)", + app->name, app->ppid); + } + + goto error; + } + + ua_event_notifier_rule->handle = ua_event_notifier_rule->obj->handle; + + DBG2("UST app event notifier %s created successfully: app = '%s' (ppid: %d), object: %p", + event_notifier.event.name, app->name, app->ppid, + ua_event_notifier_rule->obj); + + health_code_update(); + + /* Set filter if one is present. */ + if (ua_event_notifier_rule->filter) { + ret = set_ust_object_filter(app, ua_event_notifier_rule->filter, + ua_event_notifier_rule->obj); + if (ret < 0) { + goto error; + } + } + + /* Set exclusions for the event. */ + if (ua_event_notifier_rule->exclusion) { + ret = set_ust_object_exclusions(app, + ua_event_notifier_rule->exclusion, + ua_event_notifier_rule->obj); + if (ret < 0) { + goto error; + } + } + + /* Set the capture bytecodes. */ + cond_status = lttng_condition_on_event_get_capture_descriptor_count( + condition, &capture_bytecode_count); + assert(cond_status == LTTNG_CONDITION_STATUS_OK); + + for (i = 0; i < capture_bytecode_count; i++) { + const struct lttng_bytecode *capture_bytecode = + lttng_condition_on_event_get_capture_bytecode_at_index( + condition, i); + + ret = set_ust_capture(app, capture_bytecode, i, + ua_event_notifier_rule->obj); + if (ret < 0) { + goto error; + } + } + + /* + * We now need to explicitly enable the event, since it + * is disabled at creation. + */ + ret = enable_ust_object(app, ua_event_notifier_rule->obj); + if (ret < 0) { + /* + * If we hit an EPERM, something is wrong with our enable call. + * If we get an EEXIST, there is a problem on the tracer side + * since we just created it. + */ + switch (ret) { + case -LTTNG_UST_ERR_PERM: + /* Code flow problem. */ + abort(); + case -LTTNG_UST_ERR_EXIST: + /* It's OK for our use case. */ + ret = 0; + break; + default: + break; + } + + goto error; + } + + ua_event_notifier_rule->enabled = true; + +error: + health_code_update(); + return ret; +} + +/* + * Copy data between an UST app event and a LTT event. + */ +static void shadow_copy_event(struct ust_app_event *ua_event, + struct ltt_ust_event *uevent) +{ + size_t exclusion_alloc_size; + + strncpy(ua_event->name, uevent->attr.name, sizeof(ua_event->name)); + ua_event->name[sizeof(ua_event->name) - 1] = '\0'; + + ua_event->enabled = uevent->enabled; + + /* Copy event attributes */ + memcpy(&ua_event->attr, &uevent->attr, sizeof(ua_event->attr)); + + /* Copy filter bytecode */ + if (uevent->filter) { + ua_event->filter = lttng_bytecode_copy(uevent->filter); + /* Filter might be NULL here in case of ENONEM. */ + } + + /* Copy exclusion data */ + if (uevent->exclusion) { + exclusion_alloc_size = sizeof(struct lttng_event_exclusion) + + LTTNG_UST_SYM_NAME_LEN * uevent->exclusion->count; + ua_event->exclusion = zmalloc(exclusion_alloc_size); + if (ua_event->exclusion == NULL) { + PERROR("malloc"); + } else { + memcpy(ua_event->exclusion, uevent->exclusion, + exclusion_alloc_size); + } + } +} + +/* + * Copy data between an UST app channel and a LTT channel. + */ +static void shadow_copy_channel(struct ust_app_channel *ua_chan, + struct ltt_ust_channel *uchan) +{ + DBG2("UST app shadow copy of channel %s started", ua_chan->name); + + strncpy(ua_chan->name, uchan->name, sizeof(ua_chan->name)); + ua_chan->name[sizeof(ua_chan->name) - 1] = '\0'; + + ua_chan->tracefile_size = uchan->tracefile_size; + ua_chan->tracefile_count = uchan->tracefile_count; + + /* Copy event attributes since the layout is different. */ + ua_chan->attr.subbuf_size = uchan->attr.subbuf_size; + ua_chan->attr.num_subbuf = uchan->attr.num_subbuf; + ua_chan->attr.overwrite = uchan->attr.overwrite; + ua_chan->attr.switch_timer_interval = uchan->attr.switch_timer_interval; + ua_chan->attr.read_timer_interval = uchan->attr.read_timer_interval; + ua_chan->monitor_timer_interval = uchan->monitor_timer_interval; + ua_chan->attr.output = uchan->attr.output; + ua_chan->attr.blocking_timeout = uchan->attr.u.s.blocking_timeout; + + /* + * Note that the attribute channel type is not set since the channel on the + * tracing registry side does not have this information. + */ + + ua_chan->enabled = uchan->enabled; + ua_chan->tracing_channel_id = uchan->id; + + DBG3("UST app shadow copy of channel %s done", ua_chan->name); +} + +/* + * Copy data between a UST app session and a regular LTT session. + */ +static void shadow_copy_session(struct ust_app_session *ua_sess, + struct ltt_ust_session *usess, struct ust_app *app) +{ + struct tm *timeinfo; + char datetime[16]; + int ret; + char tmp_shm_path[PATH_MAX]; + + timeinfo = localtime(&app->registration_time); + strftime(datetime, sizeof(datetime), "%Y%m%d-%H%M%S", timeinfo); DBG2("Shadow copy of session handle %d", ua_sess->handle); ua_sess->tracing_id = usess->id; ua_sess->id = get_next_session_id(); - ua_sess->real_credentials.uid = app->uid; - ua_sess->real_credentials.gid = app->gid; - ua_sess->effective_credentials.uid = usess->uid; - ua_sess->effective_credentials.gid = usess->gid; + LTTNG_OPTIONAL_SET(&ua_sess->real_credentials.uid, app->uid); + LTTNG_OPTIONAL_SET(&ua_sess->real_credentials.gid, app->gid); + LTTNG_OPTIONAL_SET(&ua_sess->effective_credentials.uid, usess->uid); + LTTNG_OPTIONAL_SET(&ua_sess->effective_credentials.gid, usess->gid); ua_sess->buffer_type = usess->buffer_type; ua_sess->bits_per_long = app->bits_per_long; @@ -1878,7 +2899,7 @@ static void shadow_copy_session(struct ust_app_session *ua_sess, case LTTNG_BUFFER_PER_UID: ret = snprintf(ua_sess->path, sizeof(ua_sess->path), DEFAULT_UST_TRACE_UID_PATH, - ua_sess->real_credentials.uid, + lttng_credentials_get_uid(&ua_sess->real_credentials), app->bits_per_long); break; default: @@ -1901,12 +2922,12 @@ static void shadow_copy_session(struct ust_app_session *ua_sess, switch (ua_sess->buffer_type) { case LTTNG_BUFFER_PER_PID: ret = snprintf(tmp_shm_path, sizeof(tmp_shm_path), - DEFAULT_UST_TRACE_PID_PATH "/%s-%d-%s", + "/" DEFAULT_UST_TRACE_PID_PATH "/%s-%d-%s", app->name, app->pid, datetime); break; case LTTNG_BUFFER_PER_UID: ret = snprintf(tmp_shm_path, sizeof(tmp_shm_path), - DEFAULT_UST_TRACE_UID_PATH, + "/" DEFAULT_UST_TRACE_UID_PATH, app->uid, app->bits_per_long); break; default: @@ -2001,8 +3022,9 @@ static int setup_buffer_reg_pid(struct ust_app_session *ua_sess, app->uint64_t_alignment, app->long_alignment, app->byte_order, app->version.major, app->version.minor, reg_pid->root_shm_path, reg_pid->shm_path, - ua_sess->effective_credentials.uid, - ua_sess->effective_credentials.gid, ua_sess->tracing_id, + lttng_credentials_get_uid(&ua_sess->effective_credentials), + lttng_credentials_get_gid(&ua_sess->effective_credentials), + ua_sess->tracing_id, app->uid); if (ret < 0) { /* @@ -2302,7 +3324,7 @@ end: */ static int create_ust_app_channel_context(struct ust_app_channel *ua_chan, - struct lttng_ust_context_attr *uctx, + struct lttng_ust_context_attr *uctx, struct ust_app *app) { int ret = 0; @@ -2347,7 +3369,7 @@ int enable_ust_app_event(struct ust_app_session *ua_sess, { int ret; - ret = enable_ust_event(app, ua_sess, ua_event); + ret = enable_ust_object(app, ua_event->obj); if (ret < 0) { goto error; } @@ -2366,7 +3388,7 @@ static int disable_ust_app_event(struct ust_app_session *ua_sess, { int ret; - ret = disable_ust_event(app, ua_sess, ua_event); + ret = disable_ust_object(app, ua_event->obj); if (ret < 0) { goto error; } @@ -2398,13 +3420,33 @@ error: } /* - * Lookup ust app channel for session and enable it on the tracer side. This - * MUST be called with a RCU read side lock acquired. + * Lookup ust app map for session and disable it on the tracer side. */ -static int enable_ust_app_channel(struct ust_app_session *ua_sess, - struct ltt_ust_channel *uchan, struct ust_app *app) +static +int disable_ust_app_map(struct ust_app_session *ua_sess, + struct ust_app_map *ua_map, struct ust_app *app) { - int ret = 0; + int ret; + + ret = disable_ust_map(app, ua_sess, ua_map); + if (ret < 0) { + goto error; + } + + ua_map->enabled = 0; + +error: + return ret; +} + +/* + * Lookup ust app channel for session and enable it on the tracer side. This + * MUST be called with a RCU read side lock acquired. + */ +static int enable_ust_app_channel(struct ust_app_session *ua_sess, + struct ltt_ust_channel *uchan, struct ust_app *app) +{ + int ret = 0; struct lttng_ht_iter iter; struct lttng_ht_node_str *ua_chan_node; struct ust_app_channel *ua_chan; @@ -2428,6 +3470,37 @@ error: return ret; } +/* + * Lookup ust app map for session and enable it on the tracer side. This + * MUST be called with a RCU read side lock acquired. + */ +static int enable_ust_app_map(struct ust_app_session *ua_sess, + struct ltt_ust_map *umap, struct ust_app *app) +{ + int ret = 0; + struct lttng_ht_iter iter; + struct lttng_ht_node_str *ua_map_node; + struct ust_app_map *ua_map; + + lttng_ht_lookup(ua_sess->maps, (void *)umap->name, &iter); + ua_map_node = lttng_ht_iter_get_node_str(&iter); + if (ua_map_node == NULL) { + DBG2("Unable to find map %s in ust session id %" PRIu64, + umap->name, ua_sess->tracing_id); + goto error; + } + + ua_map = caa_container_of(ua_map_node, struct ust_app_map, node); + + ret = enable_ust_map(app, ua_sess, ua_map); + if (ret < 0) { + goto error; + } + +error: + return ret; +} + /* * Ask the consumer to create a channel and get it if successful. * @@ -2525,6 +3598,118 @@ error: return ret; } +static int create_map_object(struct ltt_ust_session *usess, + struct ust_app_session *ua_sess, struct ust_app_map *ua_map) +{ + int i, ret, nr_counter_cpu; + struct ustctl_counter_dimension dimension[1] = {0}; + struct ustctl_daemon_counter *daemon_counter; + struct lttng_ust_object_data **counter_cpus; + enum ustctl_counter_bitness bitness; + int *counter_cpu_fds; + + assert(usess); + assert(ua_sess); + assert(ua_map); + assert(ua_map->bucket_count > 0); + + DBG("Creating UST map \"%s\"", ua_map->name); + + if (ua_map->bitness == LTTNG_MAP_BITNESS_32BITS) { + bitness = USTCTL_COUNTER_BITNESS_32; + } else { + bitness = USTCTL_COUNTER_BITNESS_64; + } + + nr_counter_cpu = ustctl_get_nr_cpu_per_counter(); + counter_cpu_fds = zmalloc(nr_counter_cpu * sizeof(*counter_cpu_fds)); + if (!counter_cpu_fds) { + ret = -1; + goto end; + } + + counter_cpus = zmalloc(nr_counter_cpu * sizeof(**counter_cpus)); + if (!counter_cpus) { + ret = -1; + goto free_cpu_fds; + } + + /* Need one fd for each cpu counter of the map. */ + ret = lttng_fd_get(LTTNG_FD_APPS, nr_counter_cpu); + if (ret < 0) { + ERR("Exhausted number of available FD upon create map"); + goto free_cpu_counters; + } + + for (i = 0; i < nr_counter_cpu; i++) { + counter_cpu_fds[i] = shm_create_anonymous("ust-map-counter"); + if (counter_cpu_fds[i] < 0) { + ERR("Error creating anonymous shared memory object"); + ret = -1; + goto error; + } + } + + dimension[0].size = ua_map->bucket_count; + dimension[0].has_underflow = false; + dimension[0].has_overflow = false; + + daemon_counter = ustctl_create_counter(1, dimension, 0, -1, + nr_counter_cpu, counter_cpu_fds, + bitness, + USTCTL_COUNTER_ARITHMETIC_MODULAR, + USTCTL_COUNTER_ALLOC_PER_CPU, + ua_map->coalesce_hits); + assert(daemon_counter); + + DBG("Created daemon counter succesfully"); + + ua_map->map_handle = daemon_counter; + + ret = ustctl_create_counter_data(daemon_counter, &ua_map->obj); + assert(ret == 0); + DBG("Created counter data succesfully"); + + for (i = 0; i < nr_counter_cpu; i++) { + struct ust_app_map_counter *counter; + + /* Create UST counter */ + counter = ust_app_alloc_map_counter(); + if (counter == NULL) { + ret = -ENOMEM; + goto release_counters; + } + + ret = ustctl_create_counter_cpu_data(daemon_counter, i, + &counter->obj); + if (ret < 0) { + ERR("Creating map counter cpu data"); + free(counter); + goto error; + } + + cds_list_add_tail(&counter->list, &ua_map->counters.head); + ua_map->counters.count++; + + DBG2("UST app map counter %d created successfully", + ua_map->counters.count); + } + + ret = 0; + goto end; + +error: +release_counters: + //TODO +free_cpu_counters: + free(counter_cpus); + +free_cpu_fds: + free(counter_cpu_fds); +end: + return ret; +} + /* * Duplicate the ust data object of the ust app stream and save it in the * buffer registry stream. @@ -2539,7 +3724,7 @@ static int duplicate_stream_object(struct buffer_reg_stream *reg_stream, assert(reg_stream); assert(stream); - /* Reserve the amount of file descriptor we need. */ + /* Duplicating a stream requires 2 new fds. Reserve them. */ ret = lttng_fd_get(LTTNG_FD_APPS, 2); if (ret < 0) { ERR("Exhausted number of available FD upon duplicate stream"); @@ -2561,21 +3746,57 @@ error: return ret; } +/* + * Duplicate the ust data object of the ust app map_counter and save it in the + * buffer registry map_counter. + * + * Return 0 on success or else a negative value. + */ +static int duplicate_map_counter_object(struct buffer_reg_map_counter *reg_map_counter, + struct ust_app_map_counter *map_counter) +{ + int ret; + + assert(reg_map_counter); + assert(map_counter); + + /* Duplicating a map_counter requires 2 new fds. Reserve them. */ + ret = lttng_fd_get(LTTNG_FD_APPS, 2); + if (ret < 0) { + ERR("Exhausted number of available FD upon duplicate map_counter"); + goto error; + } + + /* Duplicate object for map_counter once the original is in the registry. */ + ret = ustctl_duplicate_ust_object_data(&map_counter->obj, + reg_map_counter->obj.ust); + if (ret < 0) { + ERR("Duplicate map_counter obj from %p to %p failed with ret %d", + reg_map_counter->obj.ust, map_counter->obj, ret); + lttng_fd_put(LTTNG_FD_APPS, 2); + goto error; + } + map_counter->handle = map_counter->obj->handle; + +error: + return ret; +} + /* * Duplicate the ust data object of the ust app. channel and save it in the * buffer registry channel. * * Return 0 on success or else a negative value. */ -static int duplicate_channel_object(struct buffer_reg_channel *reg_chan, +static int duplicate_channel_object(struct buffer_reg_channel *buf_reg_chan, struct ust_app_channel *ua_chan) { int ret; - assert(reg_chan); + assert(buf_reg_chan); assert(ua_chan); - /* Need two fds for the channel. */ + /* Duplicating a channel requires 1 new fd. Reserve it. */ ret = lttng_fd_get(LTTNG_FD_APPS, 1); if (ret < 0) { ERR("Exhausted number of available FD upon duplicate channel"); @@ -2583,10 +3804,10 @@ static int duplicate_channel_object(struct buffer_reg_channel *reg_chan, } /* Duplicate object for stream once the original is in the registry. */ - ret = ustctl_duplicate_ust_object_data(&ua_chan->obj, reg_chan->obj.ust); + ret = ustctl_duplicate_ust_object_data(&ua_chan->obj, buf_reg_chan->obj.ust); if (ret < 0) { ERR("Duplicate channel obj from %p to %p failed with ret: %d", - reg_chan->obj.ust, ua_chan->obj, ret); + buf_reg_chan->obj.ust, ua_chan->obj, ret); goto error; } ua_chan->handle = ua_chan->obj->handle; @@ -2599,20 +3820,58 @@ error_fd_get: return ret; } +/* + * Duplicate the ust data object of the ust app. map and save it in the + * buffer registry map. + * + * Return 0 on success or else a negative value. + */ +static int duplicate_map_object(struct buffer_reg_map *buf_reg_map, + struct ust_app_map *ua_map) +{ + int ret; + + assert(buf_reg_map); + assert(ua_map); + + /* Duplicating a map requires 1 new fd. Reserve it. */ + ret = lttng_fd_get(LTTNG_FD_APPS, 1); + if (ret < 0) { + ERR("Exhausted number of available FD upon duplicate map"); + goto error_fd_get; + } + + /* Duplicate object for stream once the original is in the registry. */ + ret = ustctl_duplicate_ust_object_data(&ua_map->obj, buf_reg_map->obj.ust); + if (ret < 0) { + ERR("Duplicate map obj from %p to %p failed with ret: %d", + buf_reg_map->obj.ust, ua_map->obj, ret); + goto error; + } + ua_map->handle = ua_map->obj->handle; + + return 0; + +error: + lttng_fd_put(LTTNG_FD_APPS, 1); +error_fd_get: + return ret; +} + /* * For a given channel buffer registry, setup all streams of the given ust * application channel. * * Return 0 on success or else a negative value. */ -static int setup_buffer_reg_streams(struct buffer_reg_channel *reg_chan, +static int setup_buffer_reg_streams(struct buffer_reg_channel *buf_reg_chan, struct ust_app_channel *ua_chan, struct ust_app *app) { int ret = 0; struct ust_app_stream *stream, *stmp; - assert(reg_chan); + assert(buf_reg_chan); assert(ua_chan); DBG2("UST app setup buffer registry stream"); @@ -2632,7 +3891,7 @@ static int setup_buffer_reg_streams(struct buffer_reg_channel *reg_chan, */ reg_stream->obj.ust = stream->obj; stream->obj = NULL; - buffer_reg_stream_add(reg_stream, reg_chan); + buffer_reg_stream_add(reg_stream, buf_reg_chan); /* We don't need the streams anymore. */ cds_list_del(&stream->list); @@ -2643,6 +3902,50 @@ error: return ret; } +/* + * For a given map buffer registry, setup all counters of the given ust + * application map. + * + * Return 0 on success or else a negative value. + */ +static int setup_buffer_reg_map_counters(struct buffer_reg_map *buf_reg_map, + struct ust_app_map *ua_map, + struct ust_app *app) +{ + int ret = 0; + struct ust_app_map_counter *counter, *stmp; + + assert(buf_reg_map); + assert(ua_map); + + DBG2("UST app setup buffer registry counter"); + + /* Send all counters to application. */ + cds_list_for_each_entry_safe(counter, stmp, &ua_map->counters.head, list) { + struct buffer_reg_map_counter *reg_counter; + + ret = buffer_reg_map_counter_create(®_counter); + if (ret < 0) { + goto error; + } + + /* + * Keep original pointer and nullify it in the counter so the delete + * counter call does not release the object. + */ + reg_counter->obj.ust = counter->obj; + counter->obj = NULL; + buffer_reg_map_counter_add(reg_counter, buf_reg_map); + + /* We don't need the counters anymore. */ + cds_list_del(&counter->list); + delete_ust_app_map_counter(-1, counter, app); + } + +error: + return ret; +} + /* * Create a buffer registry channel for the given session registry and * application channel object. If regp pointer is valid, it's set with the @@ -2655,7 +3958,7 @@ static int create_buffer_reg_channel(struct buffer_reg_session *reg_sess, struct ust_app_channel *ua_chan, struct buffer_reg_channel **regp) { int ret; - struct buffer_reg_channel *reg_chan = NULL; + struct buffer_reg_channel *buf_reg_chan = NULL; assert(reg_sess); assert(ua_chan); @@ -2663,14 +3966,14 @@ static int create_buffer_reg_channel(struct buffer_reg_session *reg_sess, DBG2("UST app creating buffer registry channel for %s", ua_chan->name); /* Create buffer registry channel. */ - ret = buffer_reg_channel_create(ua_chan->tracing_channel_id, ®_chan); + ret = buffer_reg_channel_create(ua_chan->tracing_channel_id, &buf_reg_chan); if (ret < 0) { goto error_create; } - assert(reg_chan); - reg_chan->consumer_key = ua_chan->key; - reg_chan->subbuf_size = ua_chan->attr.subbuf_size; - reg_chan->num_subbuf = ua_chan->attr.num_subbuf; + assert(buf_reg_chan); + buf_reg_chan->consumer_key = ua_chan->key; + buf_reg_chan->subbuf_size = ua_chan->attr.subbuf_size; + buf_reg_chan->num_subbuf = ua_chan->attr.num_subbuf; /* Create and add a channel registry to session. */ ret = ust_registry_channel_add(reg_sess->reg.ust, @@ -2678,17 +3981,63 @@ static int create_buffer_reg_channel(struct buffer_reg_session *reg_sess, if (ret < 0) { goto error; } - buffer_reg_channel_add(reg_sess, reg_chan); + buffer_reg_channel_add(reg_sess, buf_reg_chan); if (regp) { - *regp = reg_chan; + *regp = buf_reg_chan; } return 0; error: /* Safe because the registry channel object was not added to any HT. */ - buffer_reg_channel_destroy(reg_chan, LTTNG_DOMAIN_UST); + buffer_reg_channel_destroy(buf_reg_chan, LTTNG_DOMAIN_UST); +error_create: + return ret; +} + +/* + * Create a buffer registry map for the given session registry and + * application map object. If regp pointer is valid, it's set with the + * created object. Important, the created object is NOT added to the session + * registry hash table. + * + * Return 0 on success else a negative value. + */ +static int create_buffer_reg_map(struct buffer_reg_session *reg_sess, + struct ust_app_map *ua_map, struct buffer_reg_map **regp) +{ + int ret; + struct buffer_reg_map *buf_reg_map = NULL; + + assert(reg_sess); + assert(ua_map); + + DBG2("UST app creating buffer registry map for %s", ua_map->name); + + /* Create buffer registry map. */ + ret = buffer_reg_map_create(ua_map->tracing_map_id, &buf_reg_map); + if (ret < 0) { + goto error_create; + } + assert(buf_reg_map); + + /* Create and add a map registry to session. */ + ret = ust_registry_map_add(reg_sess->reg.ust, ua_map->tracing_map_id); + if (ret < 0) { + goto error; + } + buffer_reg_map_add(reg_sess, buf_reg_map); + + if (regp) { + *regp = buf_reg_map; + } + + return 0; + +error: + /* Safe because the registry map object was not added to any HT. */ + buffer_reg_map_destroy(buf_reg_map, LTTNG_DOMAIN_UST); error_create: return ret; } @@ -2700,32 +4049,70 @@ error_create: * Return 0 on success else a negative value. */ static int setup_buffer_reg_channel(struct buffer_reg_session *reg_sess, - struct ust_app_channel *ua_chan, struct buffer_reg_channel *reg_chan, + struct ust_app_channel *ua_chan, struct buffer_reg_channel *buf_reg_chan, struct ust_app *app) { int ret; assert(reg_sess); - assert(reg_chan); + assert(buf_reg_chan); assert(ua_chan); assert(ua_chan->obj); DBG2("UST app setup buffer registry channel for %s", ua_chan->name); /* Setup all streams for the registry. */ - ret = setup_buffer_reg_streams(reg_chan, ua_chan, app); + ret = setup_buffer_reg_streams(buf_reg_chan, ua_chan, app); if (ret < 0) { goto error; } - reg_chan->obj.ust = ua_chan->obj; + buf_reg_chan->obj.ust = ua_chan->obj; ua_chan->obj = NULL; return 0; error: - buffer_reg_channel_remove(reg_sess, reg_chan); - buffer_reg_channel_destroy(reg_chan, LTTNG_DOMAIN_UST); + buffer_reg_channel_remove(reg_sess, buf_reg_chan); + buffer_reg_channel_destroy(buf_reg_chan, LTTNG_DOMAIN_UST); + return ret; +} + +/* + * Setup buffer registry map for the given session registry and application + * map object. If regp pointer is valid, it's set with the created object. + * + * Return 0 on success else a negative value. + */ +static int setup_buffer_reg_map(struct buffer_reg_session *reg_sess, + struct ust_app_map *ua_map, struct buffer_reg_map *buf_reg_map, + struct ust_app *app) +{ + int ret; + + assert(reg_sess); + assert(buf_reg_map); + assert(ua_map); + assert(ua_map->obj); + + DBG2("UST app setup buffer registry map for %s", ua_map->name); + + /* Setup all counters for the registry. */ + ret = setup_buffer_reg_map_counters(buf_reg_map, ua_map, app); + if (ret < 0) { + goto error; + } + + buf_reg_map->obj.ust = ua_map->obj; + ua_map->obj = NULL; + buf_reg_map->daemon_counter = ua_map->map_handle; + ua_map->map_handle = NULL; + + return 0; + +error: + buffer_reg_map_remove(reg_sess, buf_reg_map); + buffer_reg_map_destroy(buf_reg_map, LTTNG_DOMAIN_UST); return ret; } @@ -2734,21 +4121,21 @@ error: * * Return 0 on success else a negative value. */ -static int send_channel_uid_to_ust(struct buffer_reg_channel *reg_chan, +static int send_channel_uid_to_ust(struct buffer_reg_channel *buf_reg_chan, struct ust_app *app, struct ust_app_session *ua_sess, struct ust_app_channel *ua_chan) { int ret; struct buffer_reg_stream *reg_stream; - assert(reg_chan); + assert(buf_reg_chan); assert(app); assert(ua_sess); assert(ua_chan); DBG("UST app sending buffer registry channel to ust sock %d", app->sock); - ret = duplicate_channel_object(reg_chan, ua_chan); + ret = duplicate_channel_object(buf_reg_chan, ua_chan); if (ret < 0) { goto error; } @@ -2765,8 +4152,8 @@ static int send_channel_uid_to_ust(struct buffer_reg_channel *reg_chan, health_code_update(); /* Send all streams to application. */ - pthread_mutex_lock(®_chan->stream_list_lock); - cds_list_for_each_entry(reg_stream, ®_chan->streams, lnode) { + pthread_mutex_lock(&buf_reg_chan->stream_list_lock); + cds_list_for_each_entry(reg_stream, &buf_reg_chan->streams, lnode) { struct ust_app_stream stream; ret = duplicate_stream_object(reg_stream, &stream); @@ -2792,7 +4179,87 @@ static int send_channel_uid_to_ust(struct buffer_reg_channel *reg_chan, ua_chan->is_sent = 1; error_stream_unlock: - pthread_mutex_unlock(®_chan->stream_list_lock); + pthread_mutex_unlock(&buf_reg_chan->stream_list_lock); +error: + return ret; +} + +/* + * Send buffer registry map to the application. + * + * Return 0 on success else a negative value. + */ +static int send_map_uid_to_ust(struct buffer_reg_map *buf_reg_map, + struct ust_app *app, struct ust_app_session *ua_sess, + struct ust_app_map *ua_map) +{ + int ret; + struct buffer_reg_map_counter *reg_map_counter; + + assert(buf_reg_map); + assert(app); + assert(ua_sess); + assert(ua_map); + + DBG("UST app sending buffer registry map to ust sock %d", app->sock); + + ret = duplicate_map_object(buf_reg_map, ua_map); + if (ret < 0) { + goto error; + } + + /* Send map to the application. */ + + pthread_mutex_lock(&app->sock_lock); + ret = ustctl_send_counter_data_to_ust(app->sock, + ua_sess->handle, ua_map->obj); + pthread_mutex_unlock(&app->sock_lock); + assert(ret == 0); + if (ret < 0) { + if (ret == -EPIPE || ret == -LTTNG_UST_ERR_EXITING) { + ret = -ENOTCONN; /* Caused by app exiting. */ + } + goto error_map_counter_unlock; + } + + ua_map->handle = ua_map->obj->handle; + + health_code_update(); + + /* Send all map_counters to application. */ + pthread_mutex_lock(&buf_reg_map->counter_list_lock); + cds_list_for_each_entry(reg_map_counter, &buf_reg_map->counters, lnode) { + struct ust_app_map_counter map_counter; + + ret = duplicate_map_counter_object(reg_map_counter, &map_counter); + if (ret < 0) { + goto error_map_counter_unlock; + } + + pthread_mutex_lock(&app->sock_lock); + // Do send the per cpu counter here + ret = ustctl_send_counter_cpu_data_to_ust(app->sock, + ua_map->obj, map_counter.obj); + pthread_mutex_unlock(&app->sock_lock); + assert(ret == 0); + if (ret < 0) { + (void) release_ust_app_map_counter(-1, &map_counter, app); + if (ret == -EPIPE || ret == -LTTNG_UST_ERR_EXITING) { + ret = -ENOTCONN; /* Caused by app exiting. */ + } + goto error_map_counter_unlock; + } + + /* + * The return value is not important here. This function will + * output an error if needed. + */ + (void) release_ust_app_map_counter(-1, &map_counter, app); + } + ua_map->is_sent = 1; + +error_map_counter_unlock: + pthread_mutex_unlock(&buf_reg_map->counter_list_lock); error: return ret; } @@ -2811,10 +4278,10 @@ static int create_channel_per_uid(struct ust_app *app, { int ret; struct buffer_reg_uid *reg_uid; - struct buffer_reg_channel *reg_chan; + struct buffer_reg_channel *buf_reg_chan; struct ltt_session *session = NULL; enum lttng_error_code notification_ret; - struct ust_registry_channel *chan_reg; + struct ust_registry_channel *ust_reg_chan; assert(app); assert(usess); @@ -2831,14 +4298,14 @@ static int create_channel_per_uid(struct ust_app *app, */ assert(reg_uid); - reg_chan = buffer_reg_channel_find(ua_chan->tracing_channel_id, + buf_reg_chan = buffer_reg_channel_find(ua_chan->tracing_channel_id, reg_uid); - if (reg_chan) { + if (buf_reg_chan) { goto send_channel; } /* Create the buffer registry channel object. */ - ret = create_buffer_reg_channel(reg_uid->registry, ua_chan, ®_chan); + ret = create_buffer_reg_channel(reg_uid->registry, ua_chan, &buf_reg_chan); if (ret < 0) { ERR("Error creating the UST channel \"%s\" registry instance", ua_chan->name); @@ -2867,8 +4334,8 @@ static int create_channel_per_uid(struct ust_app *app, */ ust_registry_channel_del_free(reg_uid->registry->reg.ust, ua_chan->tracing_channel_id, false); - buffer_reg_channel_remove(reg_uid->registry, reg_chan); - buffer_reg_channel_destroy(reg_chan, LTTNG_DOMAIN_UST); + buffer_reg_channel_remove(reg_uid->registry, buf_reg_chan); + buffer_reg_channel_destroy(buf_reg_chan, LTTNG_DOMAIN_UST); goto error; } @@ -2876,7 +4343,7 @@ static int create_channel_per_uid(struct ust_app *app, * Setup the streams and add it to the session registry. */ ret = setup_buffer_reg_channel(reg_uid->registry, - ua_chan, reg_chan, app); + ua_chan, buf_reg_chan, app); if (ret < 0) { ERR("Error setting up UST channel \"%s\"", ua_chan->name); goto error; @@ -2884,17 +4351,18 @@ static int create_channel_per_uid(struct ust_app *app, /* Notify the notification subsystem of the channel's creation. */ pthread_mutex_lock(®_uid->registry->reg.ust->lock); - chan_reg = ust_registry_channel_find(reg_uid->registry->reg.ust, + ust_reg_chan = ust_registry_channel_find(reg_uid->registry->reg.ust, ua_chan->tracing_channel_id); - assert(chan_reg); - chan_reg->consumer_key = ua_chan->key; - chan_reg = NULL; + assert(ust_reg_chan); + ust_reg_chan->consumer_key = ua_chan->key; + ust_reg_chan = NULL; pthread_mutex_unlock(®_uid->registry->reg.ust->lock); notification_ret = notification_thread_command_add_channel( notification_thread_handle, session->name, - ua_sess->effective_credentials.uid, - ua_sess->effective_credentials.gid, ua_chan->name, + lttng_credentials_get_uid(&ua_sess->effective_credentials), + lttng_credentials_get_gid(&ua_sess->effective_credentials), + ua_chan->name, ua_chan->key, LTTNG_DOMAIN_UST, ua_chan->attr.subbuf_size * ua_chan->attr.num_subbuf); if (notification_ret != LTTNG_OK) { @@ -2905,7 +4373,7 @@ static int create_channel_per_uid(struct ust_app *app, send_channel: /* Send buffers to the application. */ - ret = send_channel_uid_to_ust(reg_chan, app, ua_sess, ua_chan); + ret = send_channel_uid_to_ust(buf_reg_chan, app, ua_sess, ua_chan); if (ret < 0) { if (ret != -ENOTCONN) { ERR("Error sending channel to application"); @@ -2920,6 +4388,142 @@ error: return ret; } +/* + * Create and send to the application the created buffers with per UID buffers. + * + * This MUST be called with a RCU read side lock acquired. + * The session list lock and the session's lock must be acquired. + * + * Return 0 on success else a negative value. + */ +static int create_map_per_uid(struct ust_app *app, + struct ltt_ust_session *usess, struct ust_app_session *ua_sess, + struct ust_app_map *ua_map) +{ + int ret; + struct buffer_reg_uid *buffer_reg_uid; + struct buffer_reg_map *buffer_reg_map; + struct ltt_session *session = NULL; + struct ust_registry_map *ust_reg_map; + + assert(app); + assert(usess); + assert(ua_sess); + assert(ua_map); + + DBG("UST app creating map %s with per UID buffers", ua_map->name); + + buffer_reg_uid = buffer_reg_uid_find(usess->id, app->bits_per_long, app->uid); + /* + * The session creation handles the creation of this global registry + * object. If none can be find, there is a code flow problem or a + * teardown race. + */ + assert(buffer_reg_uid); + + buffer_reg_map = buffer_reg_map_find(ua_map->tracing_map_id, + buffer_reg_uid); + if (buffer_reg_map) { + goto send_map; + } + + /* Create the buffer registry map object. */ + ret = create_buffer_reg_map(buffer_reg_uid->registry, ua_map, + &buffer_reg_map); + if (ret < 0) { + ERR("Error creating the UST map \"%s\" registry instance", + ua_map->name); + goto error; + } + + session = session_find_by_id(ua_sess->tracing_id); + assert(session); + assert(pthread_mutex_trylock(&session->lock)); + assert(session_trylock_list()); + + /* + */ + ret = create_map_object(usess, ua_sess, ua_map); + assert(ret == 0); + if (ret < 0) { + ERR("Error creating UST map object: map_name = \"%s\"", ua_map->name); + goto error; + } + + /* + * Setup the streams and add it to the session registry. + */ + ret = setup_buffer_reg_map(buffer_reg_uid->registry, ua_map, + buffer_reg_map, app); + if (ret < 0) { + ERR("Error setting up UST map \"%s\"", ua_map->name); + goto error; + } + + /* Notify the notification subsystem of the map's creation. */ + pthread_mutex_lock(&buffer_reg_uid->registry->reg.ust->lock); + ust_reg_map = ust_registry_map_find(buffer_reg_uid->registry->reg.ust, + ua_map->tracing_map_id); + assert(ust_reg_map); + ust_reg_map = NULL; + pthread_mutex_unlock(&buffer_reg_uid->registry->reg.ust->lock); + +send_map: + /* Send buffers to the application. */ + ret = send_map_uid_to_ust(buffer_reg_map, app, ua_sess, ua_map); + if (ret < 0) { + if (ret != -ENOTCONN) { + ERR("Error sending map to application"); + } + goto error; + } + +error: + if (session) { + session_put(session); + } + return ret; +} + +//static int destroy_map_per_uid(struct ust_app *app, +// struct ltt_ust_session *usess, struct ust_app_session *ua_sess, +// struct ust_app_map *ua_map) +//{ +// int ret; +// struct buffer_reg_uid *buffer_reg_uid; +// struct buffer_reg_map *buffer_reg_map; +// +// assert(app); +// assert(usess); +// assert(ua_sess); +// assert(ua_map); +// +// DBG("UST app destroy map %s with per UID buffers", ua_map->name); +// +// buffer_reg_uid = buffer_reg_uid_find(usess->id, app->bits_per_long, app->uid); +// /* +// * The session creation handles the creation of this global registry +// * object. If none can be find, there is a code flow problem or a +// * teardown race. +// */ +// assert(buffer_reg_uid); +// +// buffer_reg_map = buffer_reg_map_find(ua_map->tracing_map_id, +// buffer_reg_uid); +// if (!buffer_reg_map) { +// ERR("Can't find map in buffer registry: map-name = '%s', uid = %d", +// ua_map->name, app->uid); +// ret = -1; +// goto end; +// } +// +// buffer_reg_map_destroy(buffer_reg_map, LTTNG_DOMAIN_UST); +// +// ret = 0; +//end: +// return ret; +//} + /* * Create and send to the application the created buffers with per PID buffers. * @@ -2937,7 +4541,7 @@ static int create_channel_per_pid(struct ust_app *app, enum lttng_error_code cmd_ret; struct ltt_session *session = NULL; uint64_t chan_reg_key; - struct ust_registry_channel *chan_reg; + struct ust_registry_channel *ust_reg_chan; assert(app); assert(usess); @@ -2986,15 +4590,16 @@ static int create_channel_per_pid(struct ust_app *app, chan_reg_key = ua_chan->key; pthread_mutex_lock(®istry->lock); - chan_reg = ust_registry_channel_find(registry, chan_reg_key); - assert(chan_reg); - chan_reg->consumer_key = ua_chan->key; + ust_reg_chan = ust_registry_channel_find(registry, chan_reg_key); + assert(ust_reg_chan); + ust_reg_chan->consumer_key = ua_chan->key; pthread_mutex_unlock(®istry->lock); cmd_ret = notification_thread_command_add_channel( notification_thread_handle, session->name, - ua_sess->effective_credentials.uid, - ua_sess->effective_credentials.gid, ua_chan->name, + lttng_credentials_get_uid(&ua_sess->effective_credentials), + lttng_credentials_get_gid(&ua_sess->effective_credentials), + ua_chan->name, ua_chan->key, LTTNG_DOMAIN_UST, ua_chan->attr.subbuf_size * ua_chan->attr.num_subbuf); if (cmd_ret != LTTNG_OK) { @@ -3015,6 +4620,85 @@ error: return ret; } +/* + * Create and send to the application the created buffers with per PID buffers. + * + * Called with UST app session lock held. + * The session list lock and the session's lock must be acquired. + * + * Return 0 on success else a negative value. + */ +static int create_map_per_pid(struct ust_app *app, + struct ltt_ust_session *usess, struct ust_app_session *ua_sess, + struct ust_app_map *ua_map) +{ + int ret; + struct ust_registry_session *registry; + struct ltt_session *session = NULL; + uint64_t map_reg_key; + struct ust_registry_map *ust_reg_map; + + assert(app); + assert(usess); + assert(ua_sess); + assert(ua_map); + + DBG("UST app creating map %s with per PID buffers", ua_map->name); + + rcu_read_lock(); + + registry = get_session_registry(ua_sess); + /* The UST app session lock is held, registry shall not be null. */ + assert(registry); + + /* Create and add a new map registry to session. */ + ret = ust_registry_map_add(registry, ua_map->key); + if (ret < 0) { + ERR("Error creating the UST map \"%s\" registry instance", + ua_map->name); + goto error; + } + + session = session_find_by_id(ua_sess->tracing_id); + assert(session); + + assert(pthread_mutex_trylock(&session->lock)); + assert(session_trylock_list()); + + /* Create and get map. */ + ret = create_map_object(usess, ua_sess, ua_map); + if (ret < 0) { + ERR("Error creating UST map object: map_name = \"%s\" ", + ua_map->name); + goto error_remove_from_registry; + } + + ret = send_map_pid_to_ust(app, ua_sess, ua_map); + if (ret < 0) { + if (ret != -ENOTCONN) { + ERR("Error sending map to application"); + } + goto error_remove_from_registry; + } + + map_reg_key = ua_map->key; + pthread_mutex_lock(®istry->lock); + ust_reg_map = ust_registry_map_find(registry, map_reg_key); + assert(ust_reg_map); + pthread_mutex_unlock(®istry->lock); + +error_remove_from_registry: + if (ret) { + ust_registry_map_del_free(registry, ua_map->key); + } +error: + rcu_read_unlock(); + if (session) { + session_put(session); + } + return ret; +} + /* * From an already allocated ust app channel, create the channel buffers if * needed and send them to the application. This MUST be called with a RCU read @@ -3063,7 +4747,7 @@ static int ust_app_channel_send(struct ust_app *app, /* Initialize ust objd object using the received handle and add it. */ lttng_ht_node_init_ulong(&ua_chan->ust_objd_node, ua_chan->handle); - lttng_ht_add_unique_ulong(app->ust_objd, &ua_chan->ust_objd_node); + lttng_ht_add_unique_ulong(app->ust_chan_objd, &ua_chan->ust_objd_node); /* If channel is not enabled, disable it on the tracer */ if (!ua_chan->enabled) { @@ -3077,6 +4761,68 @@ error: return ret; } +/* + * From an already allocated ust app map, create the map buffers if + * needed and send them to the application. This MUST be called with a RCU read + * side lock acquired. + * + * Called with UST app session lock held. + * + * Return 0 on success or else a negative value. Returns -ENOTCONN if + * the application exited concurrently. + */ +static int ust_app_map_send(struct ust_app *app, + struct ltt_ust_session *usess, struct ust_app_session *ua_sess, + struct ust_app_map *ua_map) +{ + int ret; + + assert(app); + assert(usess); + assert(usess->active); + assert(ua_sess); + assert(ua_map); + + /* Handle buffer type before sending the map to the application. */ + switch (usess->buffer_type) { + case LTTNG_BUFFER_PER_UID: + { + ret = create_map_per_uid(app, usess, ua_sess, ua_map); + if (ret < 0) { + goto error; + } + break; + } + case LTTNG_BUFFER_PER_PID: + { + ret = create_map_per_pid(app, usess, ua_sess, ua_map); + if (ret < 0) { + goto error; + } + break; + } + default: + assert(0); + ret = -EINVAL; + goto error; + } + + /* Initialize ust objd object using the received handle and add it. */ + lttng_ht_node_init_ulong(&ua_map->ust_objd_node, ua_map->handle); + lttng_ht_add_unique_ulong(app->ust_map_objd, &ua_map->ust_objd_node); + + /* If map is not enabled, disable it on the tracer */ + if (!ua_map->enabled) { + ret = disable_ust_map(app, ua_sess, ua_map); + if (ret < 0) { + goto error; + } + } + +error: + return ret; +} + /* * Create UST app channel and return it through ua_chanp if not NULL. * @@ -3127,13 +4873,71 @@ error: return ret; } +/* + * Create UST app map and return it through ua_mapp if not NULL. + * + * Called with UST app session lock and RCU read-side lock held. + * + * Return 0 on success or else a negative value. + */ +static int ust_app_map_allocate(struct ust_app_session *ua_sess, + struct ltt_ust_map *umap, + enum lttng_ust_chan_type type, struct ltt_ust_session *usess, + struct ust_app_map **ua_mapp) +{ + int ret = 0; + struct lttng_ht_iter iter; + struct lttng_ht_node_str *ua_map_node; + struct ust_app_map *ua_map; + + DBG("Allocating map id = %"PRIu64, umap->id); + + /* Lookup map in the ust app session */ + lttng_ht_lookup(ua_sess->maps, (void *)umap->name, &iter); + ua_map_node = lttng_ht_iter_get_node_str(&iter); + if (ua_map_node != NULL) { + ua_map = caa_container_of(ua_map_node, struct ust_app_map, node); + goto end; + } + + ua_map = alloc_ust_app_map(umap->name, ua_sess); + if (ua_map == NULL) { + /* Only malloc can fail here */ + ret = -ENOMEM; + goto error; + } + //shadow_copy_map(ua_map, umap); + ua_map->tracing_map_id = umap->id; + ua_map->coalesce_hits = umap->coalesce_hits; + ua_map->dead_app_kv_values = &umap->dead_app_kv_values; + ua_map->bitness = umap->bitness; + + /* Set map type. */ + //ua_map->attr.type = type; + ua_map->bucket_count = umap->bucket_count; + + /* Only add the map if successful on the tracer side. */ + lttng_ht_add_unique_str(ua_sess->maps, &ua_map->node); +end: + if (ua_mapp) { + *ua_mapp = ua_map; + } + + /* Everything went well. */ + return 0; + +error: + return ret; +} + /* * Create UST app event and create it on the tracer side. * + * Must be called with the RCU read side lock held. * Called with ust app session mutex held. */ static -int create_ust_app_event(struct ust_app_session *ua_sess, +int create_ust_app_channel_event(struct ust_app_session *ua_sess, struct ust_app_channel *ua_chan, struct ltt_ust_event *uevent, struct ust_app *app) { @@ -3149,7 +4953,7 @@ int create_ust_app_event(struct ust_app_session *ua_sess, shadow_copy_event(ua_event, uevent); /* Create it on the tracer side */ - ret = create_ust_event(app, ua_sess, ua_chan, ua_event); + ret = create_ust_channel_event(app, ua_sess, ua_chan, ua_event); if (ret < 0) { /* * Not found previously means that it does not exist on the @@ -3167,10 +4971,10 @@ int create_ust_app_event(struct ust_app_session *ua_sess, goto error; } - add_unique_ust_app_event(ua_chan, ua_event); + add_unique_ust_app_event(ua_chan->events, ua_event); - DBG2("UST app create event %s for PID %d completed", ua_event->name, - app->pid); + DBG2("UST app create event completed: app = '%s' (ppid: %d)", + app->name, app->ppid); end: return ret; @@ -3182,55 +4986,188 @@ error: } /* - * Create UST metadata and open it on the tracer side. + * Create UST app event and create it on the tracer side. * - * Called with UST app session lock held and RCU read side lock. + * Must be called with the RCU read side lock held. + * Called with ust app session mutex held. */ -static int create_ust_app_metadata(struct ust_app_session *ua_sess, - struct ust_app *app, struct consumer_output *consumer) +static +int create_ust_app_map_event(struct ust_app_session *ua_sess, + struct ust_app_map *ua_map, struct ltt_ust_event *uevent, + struct ust_app *app) { int ret = 0; - struct ust_app_channel *metadata; - struct consumer_socket *socket; + uint64_t map_reg_key; + struct ust_app_event *ua_event; struct ust_registry_session *registry; - struct ltt_session *session = NULL; - assert(ua_sess); - assert(app); - assert(consumer); + ua_event = alloc_ust_app_event(uevent->attr.name, &uevent->attr); + if (ua_event == NULL) { + /* Only failure mode of alloc_ust_app_event(). */ + ret = -ENOMEM; + goto end; + } + shadow_copy_event(ua_event, uevent); registry = get_session_registry(ua_sess); - /* The UST app session is held registry shall not be null. */ - assert(registry); - - pthread_mutex_lock(®istry->lock); - - /* Metadata already exists for this registry or it was closed previously */ - if (registry->metadata_key || registry->metadata_closed) { + if (!registry) { + DBG("Application session is being torn down. Abort event notify"); ret = 0; goto error; } - /* Allocate UST metadata */ - metadata = alloc_ust_app_channel(DEFAULT_METADATA_NAME, ua_sess, NULL); - if (!metadata) { - /* malloc() failed */ - ret = -ENOMEM; - goto error; + if (ua_sess->buffer_type == LTTNG_BUFFER_PER_UID) { + map_reg_key = ua_map->tracing_map_id; + } else { + map_reg_key = ua_map->key; } - memcpy(&metadata->attr, &ua_sess->metadata_attr, sizeof(metadata->attr)); + pthread_mutex_lock(®istry->lock); + ret = ust_registry_map_add_token_key_mapping(registry, map_reg_key, + uevent->attr.token, uevent->key); + assert(ret == 0); + pthread_mutex_unlock(®istry->lock); - /* Need one fd for the channel. */ - ret = lttng_fd_get(LTTNG_FD_APPS, 1); + /* Create it on the tracer side */ + ret = create_ust_map_event(app, ua_sess, ua_map, uevent->key, ua_event); if (ret < 0) { - ERR("Exhausted number of available FD upon create metadata"); - goto error; - } - - /* Get the right consumer socket for the application. */ - socket = consumer_find_socket_by_bitness(app->bits_per_long, consumer); - if (!socket) { + /* + * Not found previously means that it does not exist on the + * tracer. If the application reports that the event existed, + * it means there is a bug in the sessiond or lttng-ust + * (or corruption, etc.) + */ + if (ret == -LTTNG_UST_ERR_EXIST) { + ERR("Tracer for application reported that an event being created already existed: " + "event_name = \"%s\", pid = %d, ppid = %d, uid = %d, gid = %d", + uevent->attr.name, + app->pid, app->ppid, app->uid, + app->gid); + } + + /* + * FIXME: frdeso: remove key from tokey->key mapping. + */ + goto error; + } + + add_unique_ust_app_event(ua_map->events, ua_event); + + DBG2("UST app create event completed: app = '%s', tracer token = %"PRIu64" (ppid: %d)", + app->name, uevent->attr.token, app->ppid); + +end: + return ret; + +error: + /* Valid. Calling here is already in a read side lock */ + delete_ust_app_event(-1, ua_event, app); + return ret; +} + +/* + * Create UST app event notifier rule and create it on the tracer side. + * + * Must be called with the RCU read side lock held. + * Called with ust app session mutex held. + */ +static +int create_ust_app_event_notifier_rule(struct lttng_trigger *trigger, + struct ust_app *app) +{ + int ret = 0; + struct ust_app_event_notifier_rule *ua_event_notifier_rule; + + ua_event_notifier_rule = alloc_ust_app_event_notifier_rule(trigger); + if (ua_event_notifier_rule == NULL) { + ret = -ENOMEM; + goto end; + } + + /* Create it on the tracer side. */ + ret = create_ust_event_notifier(app, ua_event_notifier_rule); + if (ret < 0) { + /* + * Not found previously means that it does not exist on the + * tracer. If the application reports that the event existed, + * it means there is a bug in the sessiond or lttng-ust + * (or corruption, etc.) + */ + if (ret == -LTTNG_UST_ERR_EXIST) { + ERR("Tracer for application reported that an event notifier being created already exists: " + "token = \"%" PRIu64 "\", pid = %d, ppid = %d, uid = %d, gid = %d", + lttng_trigger_get_tracer_token(trigger), + app->pid, app->ppid, app->uid, + app->gid); + } + goto error; + } + + lttng_ht_add_unique_u64(app->token_to_event_notifier_rule_ht, + &ua_event_notifier_rule->node); + + DBG2("UST app create token event rule completed: app = '%s' (ppid: %d), token = %" PRIu64, + app->name, app->ppid, lttng_trigger_get_tracer_token(trigger)); + + goto end; + +error: + /* The RCU read side lock is already being held by the caller. */ + delete_ust_app_event_notifier_rule(-1, ua_event_notifier_rule, app); +end: + return ret; +} + +/* + * Create UST metadata and open it on the tracer side. + * + * Called with UST app session lock held and RCU read side lock. + */ +static int create_ust_app_metadata(struct ust_app_session *ua_sess, + struct ust_app *app, struct consumer_output *consumer) +{ + int ret = 0; + struct ust_app_channel *metadata; + struct consumer_socket *socket; + struct ust_registry_session *registry; + struct ltt_session *session = NULL; + + assert(ua_sess); + assert(app); + assert(consumer); + + registry = get_session_registry(ua_sess); + /* The UST app session is held registry shall not be null. */ + assert(registry); + + pthread_mutex_lock(®istry->lock); + + /* Metadata already exists for this registry or it was closed previously */ + if (registry->metadata_key || registry->metadata_closed) { + ret = 0; + goto error; + } + + /* Allocate UST metadata */ + metadata = alloc_ust_app_channel(DEFAULT_METADATA_NAME, ua_sess, NULL); + if (!metadata) { + /* malloc() failed */ + ret = -ENOMEM; + goto error; + } + + memcpy(&metadata->attr, &ua_sess->metadata_attr, sizeof(metadata->attr)); + + /* Need one fd for the channel. */ + ret = lttng_fd_get(LTTNG_FD_APPS, 1); + if (ret < 0) { + ERR("Exhausted number of available FD upon create metadata"); + goto error; + } + + /* Get the right consumer socket for the application. */ + socket = consumer_find_socket_by_bitness(app->bits_per_long, consumer); + if (!socket) { ret = -EINVAL; goto error_consumer; } @@ -3324,7 +5261,9 @@ error: */ struct ust_app *ust_app_create(struct ust_register_msg *msg, int sock) { + int ret; struct ust_app *lta = NULL; + struct lttng_pipe *event_notifier_event_source_pipe = NULL; assert(msg); assert(sock >= 0); @@ -3341,12 +5280,33 @@ struct ust_app *ust_app_create(struct ust_register_msg *msg, int sock) goto error; } + /* + * Reserve the two file descriptors of the event source pipe. The write + * end will be closed once it is passed to the application, at which + * point a single 'put' will be performed. + */ + ret = lttng_fd_get(LTTNG_FD_APPS, 2); + if (ret) { + ERR("Failed to reserve two file descriptors for the event source pipe while creating a new application instance: app = '%s' (ppid: %d)", + msg->name, (int) msg->ppid); + goto error; + } + + event_notifier_event_source_pipe = lttng_pipe_open(FD_CLOEXEC); + if (!event_notifier_event_source_pipe) { + PERROR("Failed to open application event source pipe: '%s' (ppid = %d)", + msg->name, msg->ppid); + goto error; + } + lta = zmalloc(sizeof(struct ust_app)); if (lta == NULL) { PERROR("malloc"); - goto error; + goto error_free_pipe; } + lta->event_notifier_group.event_pipe = event_notifier_event_source_pipe; + lta->ppid = msg->ppid; lta->uid = msg->uid; lta->gid = msg->gid; @@ -3362,9 +5322,11 @@ struct ust_app *ust_app_create(struct ust_register_msg *msg, int sock) lta->v_major = msg->major; lta->v_minor = msg->minor; lta->sessions = lttng_ht_new(0, LTTNG_HT_TYPE_U64); - lta->ust_objd = lttng_ht_new(0, LTTNG_HT_TYPE_ULONG); + lta->ust_chan_objd = lttng_ht_new(0, LTTNG_HT_TYPE_ULONG); + lta->ust_map_objd = lttng_ht_new(0, LTTNG_HT_TYPE_ULONG); lta->ust_sessions_objd = lttng_ht_new(0, LTTNG_HT_TYPE_ULONG); lta->notify_sock = -1; + lta->token_to_event_notifier_rule_ht = lttng_ht_new(0, LTTNG_HT_TYPE_U64); /* Copy name and make sure it's NULL terminated. */ strncpy(lta->name, msg->name, sizeof(lta->name)); @@ -3384,8 +5346,13 @@ struct ust_app *ust_app_create(struct ust_register_msg *msg, int sock) lttng_ht_node_init_ulong(<a->sock_n, (unsigned long) lta->sock); CDS_INIT_LIST_HEAD(<a->teardown_head); -error: return lta; + +error_free_pipe: + lttng_pipe_destroy(event_notifier_event_source_pipe); + lttng_fd_put(LTTNG_FD_APPS, 2); +error: + return NULL; } /* @@ -3451,6 +5418,83 @@ int ust_app_version(struct ust_app *app) return ret; } +/* + * Setup the base event notifier group. + * + * Return 0 on success else a negative value either an errno code or a + * LTTng-UST error code. + */ +int ust_app_setup_event_notifier_group(struct ust_app *app) +{ + int ret; + int event_pipe_write_fd; + struct lttng_ust_object_data *event_notifier_group = NULL; + enum lttng_error_code lttng_ret; + enum event_notifier_error_accounting_status event_notifier_error_accounting_status; + + assert(app); + + /* Get the write side of the pipe. */ + event_pipe_write_fd = lttng_pipe_get_writefd( + app->event_notifier_group.event_pipe); + + pthread_mutex_lock(&app->sock_lock); + ret = ustctl_create_event_notifier_group(app->sock, + event_pipe_write_fd, &event_notifier_group); + pthread_mutex_unlock(&app->sock_lock); + if (ret < 0) { + if (ret != -LTTNG_UST_ERR_EXITING && ret != -EPIPE) { + ERR("Failed to create application event notifier group: ret = %d, app socket fd = %d, event_pipe_write_fd = %d", + ret, app->sock, event_pipe_write_fd); + } else { + DBG("Failed to create application event notifier group (application is dead): app socket fd = %d", + app->sock); + } + + goto error; + } + + ret = lttng_pipe_write_close(app->event_notifier_group.event_pipe); + if (ret) { + ERR("Failed to close write end of the application's event source pipe: app = '%s' (ppid = %d)", + app->name, app->ppid); + goto error; + } + + /* + * Release the file descriptor that was reserved for the write-end of + * the pipe. + */ + lttng_fd_put(LTTNG_FD_APPS, 1); + + lttng_ret = notification_thread_command_add_tracer_event_source( + notification_thread_handle, + lttng_pipe_get_readfd(app->event_notifier_group.event_pipe), + LTTNG_DOMAIN_UST); + if (lttng_ret != LTTNG_OK) { + ERR("Failed to add tracer event source to notification thread"); + ret = - 1; + goto error; + } + + /* Assign handle only when the complete setup is valid. */ + app->event_notifier_group.object = event_notifier_group; + + event_notifier_error_accounting_status = event_notifier_error_accounting_register_app(app); + if (event_notifier_error_accounting_status != EVENT_NOTIFIER_ERROR_ACCOUNTING_STATUS_OK) { + ERR("Failed to setup event notifier error accounting for app"); + ret = -1; + goto error; + } + + return ret; + +error: + ustctl_release_object(app->sock, app->event_notifier_group.object); + free(app->event_notifier_group.object); + return ret; +} + /* * Unregister app by removing it from the global traceable app list and freeing * the data struct. @@ -3860,16 +5904,12 @@ void ust_app_clean_list(void) if (ust_app_ht_by_notify_sock) { cds_lfht_for_each_entry(ust_app_ht_by_notify_sock->ht, &iter.iter, app, notify_sock_n.node) { - struct cds_lfht_node *node; - struct ust_app *app; - - node = cds_lfht_iter_get_node(&iter.iter); - if (!node) { - continue; - } + /* + * Assert that all notifiers are gone as all triggers + * are unregistered prior to this clean-up. + */ + assert(lttng_ht_get_count(app->token_to_event_notifier_rule_ht) == 0); - app = container_of(node, struct ust_app, - notify_sock_n.node); ust_app_notify_sock_unregister(app->notify_sock); } } @@ -3981,6 +6021,62 @@ int ust_app_disable_channel_glb(struct ltt_ust_session *usess, return ret; } +/* + * For a specific UST session, disable the channel for all registered apps. + */ +int ust_app_disable_map_glb(struct ltt_ust_session *usess, + struct ltt_ust_map *umap) +{ + int ret = 0; + struct lttng_ht_iter iter; + struct lttng_ht_node_str *ua_map_node; + struct ust_app *app; + struct ust_app_session *ua_sess; + struct ust_app_map *ua_map; + + assert(usess->active); + DBG2("UST app disabling map %s from global domain for session id %" PRIu64, + umap->name, usess->id); + + rcu_read_lock(); + + /* For every registered applications */ + cds_lfht_for_each_entry(ust_app_ht->ht, &iter.iter, app, pid_n.node) { + struct lttng_ht_iter uiter; + if (!app->compatible) { + /* + * TODO: In time, we should notice the caller of this error by + * telling him that this is a version error. + */ + continue; + } + ua_sess = lookup_session_by_app(usess, app); + if (ua_sess == NULL) { + continue; + } + + /* Get map */ + lttng_ht_lookup(ua_sess->maps, (void *)umap->name, &uiter); + ua_map_node = lttng_ht_iter_get_node_str(&uiter); + /* If the session if found for the app, the map must be there */ + assert(ua_map_node); + + ua_map = caa_container_of(ua_map_node, struct ust_app_map, node); + /* The map must not be already disabled */ + assert(ua_map->enabled == 1); + + /* Disable map onto application */ + ret = disable_ust_app_map(ua_sess, ua_map, app); + if (ret < 0) { + /* XXX: We might want to report this error at some point... */ + continue; + } + } + + rcu_read_unlock(); + return ret; +} + /* * For a specific UST session, enable the channel for all registered apps. */ @@ -4024,10 +6120,53 @@ int ust_app_enable_channel_glb(struct ltt_ust_session *usess, return ret; } +/* + * For a specific UST session, enable the map for all registered apps. + */ +int ust_app_enable_map_glb(struct ltt_ust_session *usess, + struct ltt_ust_map *umap) +{ + int ret = 0; + struct lttng_ht_iter iter; + struct ust_app *app; + struct ust_app_session *ua_sess; + + assert(usess->active); + DBG2("UST app enabling map %s to global domain for session id %" PRIu64, + umap->name, usess->id); + + rcu_read_lock(); + + /* For every registered applications */ + cds_lfht_for_each_entry(ust_app_ht->ht, &iter.iter, app, pid_n.node) { + if (!app->compatible) { + /* + * TODO: In time, we should notice the caller of this error by + * telling him that this is a version error. + */ + continue; + } + ua_sess = lookup_session_by_app(usess, app); + if (ua_sess == NULL) { + continue; + } + + /* Enable map onto application */ + ret = enable_ust_app_map(ua_sess, umap, app); + if (ret < 0) { + /* XXX: We might want to report this error at some point... */ + continue; + } + } + + rcu_read_unlock(); + return ret; +} + /* * Disable an event in a channel and for a specific session. */ -int ust_app_disable_event_glb(struct ltt_ust_session *usess, +int ust_app_disable_channel_event_glb(struct ltt_ust_session *usess, struct ltt_ust_channel *uchan, struct ltt_ust_event *uevent) { int ret = 0; @@ -4072,7 +6211,7 @@ int ust_app_disable_event_glb(struct ltt_ust_session *usess, ua_event = find_ust_app_event(ua_chan->events, uevent->attr.name, uevent->filter, uevent->attr.loglevel, - uevent->exclusion); + uevent->exclusion, uevent->attr.token); if (ua_event == NULL) { DBG2("Event %s not found in channel %s for app pid %d." "Skipping", uevent->attr.name, uchan->name, app->pid); @@ -4090,6 +6229,72 @@ int ust_app_disable_event_glb(struct ltt_ust_session *usess, return ret; } +/* + * Disable an event in a map and for a specific session. + */ +int ust_app_disable_map_event_glb(struct ltt_ust_session *usess, + struct ltt_ust_map *umap, struct ltt_ust_event *uevent) +{ + int ret = 0; + struct lttng_ht_iter iter, uiter; + struct lttng_ht_node_str *ua_map_node; + struct ust_app *app; + struct ust_app_session *ua_sess; + struct ust_app_map *ua_map; + struct ust_app_event *ua_event; + + assert(usess->active); + DBG("UST app disabling event %s for all apps in map " + "%s for session id %" PRIu64, + uevent->attr.name, umap->name, usess->id); + + rcu_read_lock(); + + /* For all registered applications */ + cds_lfht_for_each_entry(ust_app_ht->ht, &iter.iter, app, pid_n.node) { + if (!app->compatible) { + /* + * TODO: In time, we should notice the caller of this error by + * telling him that this is a version error. + */ + continue; + } + ua_sess = lookup_session_by_app(usess, app); + if (ua_sess == NULL) { + /* Next app */ + continue; + } + + /* Lookup map in the ust app session */ + lttng_ht_lookup(ua_sess->maps, (void *)umap->name, &uiter); + ua_map_node = lttng_ht_iter_get_node_str(&uiter); + if (ua_map_node == NULL) { + DBG2("map %s not found in session id %" PRIu64 " for app pid %d." + "Skipping", umap->name, usess->id, app->pid); + continue; + } + ua_map = caa_container_of(ua_map_node, struct ust_app_map, node); + + ua_event = find_ust_app_event(ua_map->events, uevent->attr.name, + uevent->filter, uevent->attr.loglevel, + uevent->exclusion, uevent->attr.token); + if (ua_event == NULL) { + DBG2("Event %s not found in map %s for app pid %d." + "Skipping", uevent->attr.name, umap->name, app->pid); + continue; + } + + ret = disable_ust_app_event(ua_sess, ua_event, app); + if (ret < 0) { + /* XXX: Report error someday... */ + continue; + } + } + + rcu_read_unlock(); + return ret; +} + /* The ua_sess lock must be held by the caller. */ static int ust_app_channel_create(struct ltt_ust_session *usess, @@ -4118,11 +6323,14 @@ int ust_app_channel_create(struct ltt_ust_session *usess, ret = ust_app_channel_allocate(ua_sess, uchan, LTTNG_UST_CHAN_PER_CPU, usess, &ua_chan); - if (ret == 0) { - ret = ust_app_channel_send(app, usess, - ua_sess, ua_chan); - } else { - goto end; + if (ret < 0) { + goto error; + } + + ret = ust_app_channel_send(app, usess, + ua_sess, ua_chan); + if (ret) { + goto error; } /* Add contexts. */ @@ -4130,10 +6338,12 @@ int ust_app_channel_create(struct ltt_ust_session *usess, ret = create_ust_app_channel_context(ua_chan, &uctx->ctx, app); if (ret) { - goto end; + goto error; } } } + +error: if (ret < 0) { switch (ret) { case -ENOTCONN: @@ -4149,7 +6359,7 @@ int ust_app_channel_create(struct ltt_ust_session *usess, break; } } -end: + if (ret == 0 && _ua_chan) { /* * Only return the application's channel on success. Note @@ -4161,79 +6371,67 @@ end: return ret; } -/* - * For a specific UST session, create the channel for all registered apps. - */ -int ust_app_create_channel_glb(struct ltt_ust_session *usess, - struct ltt_ust_channel *uchan) +/* The ua_sess lock must be held by the caller. */ +static +int ust_app_map_create(struct ltt_ust_session *usess, + struct ust_app_session *ua_sess, + struct ltt_ust_map *umap, struct ust_app *app, + struct ust_app_map **_ua_map) { int ret = 0; - struct cds_lfht_iter iter; - struct ust_app *app; + struct ust_app_map *ua_map = NULL; - assert(usess); - assert(usess->active); - assert(uchan); + assert(ua_sess); + ASSERT_LOCKED(ua_sess->lock); - DBG2("UST app adding channel %s to UST domain for session id %" PRIu64, - uchan->name, usess->id); + /* + * Create map onto application and synchronize its + * configuration. + */ + ret = ust_app_map_allocate(ua_sess, umap, + LTTNG_UST_CHAN_PER_CPU, usess, + &ua_map); + if (ret < 0) { + goto error; + } - rcu_read_lock(); - /* For every registered applications */ - cds_lfht_for_each_entry(ust_app_ht->ht, &iter, app, pid_n.node) { - struct ust_app_session *ua_sess; - int session_was_created = 0; + ret = ust_app_map_send(app, usess, ua_sess, ua_map); + if (ret) { + goto error; + } - if (!app->compatible || - !trace_ust_pid_tracker_lookup(usess, app->pid)) { - goto error_rcu_unlock; +error: + if (ret < 0) { + switch (ret) { + case -ENOTCONN: + /* + * The application's socket is not valid. Either a bad socket + * or a timeout on it. We can't inform the caller that for a + * specific app, the session failed so lets continue here. + */ + ret = 0; /* Not an error. */ + break; + case -ENOMEM: + default: + break; } + } + if (ret == 0 && _ua_map) { /* - * Create session on the tracer side and add it to app session HT. Note - * that if session exist, it will simply return a pointer to the ust - * app session. + * Only return the application's map on success. Note + * that the map can still be part of the application's + * map hashtable on error. */ - ret = find_or_create_ust_app_session(usess, app, &ua_sess, - &session_was_created); - if (ret < 0) { - switch (ret) { - case -ENOTCONN: - /* - * The application's socket is not valid. Either a bad - * socket or a timeout on it. We can't inform the caller - * that for a specific app, the session failed so lets - * continue here; it is not an error. - */ - ret = 0; - goto error_rcu_unlock; - case -ENOMEM: - default: - goto error_rcu_unlock; - } - } - - if (ua_sess->deleted) { - continue; - } - ret = ust_app_channel_create(usess, ua_sess, uchan, app, NULL); - if (ret) { - if (session_was_created) { - destroy_app_session(app, ua_sess); - } - /* Continue to the next application. */ - } + *_ua_map = ua_map; } - -error_rcu_unlock: - rcu_read_unlock(); return ret; } /* * Enable event for a specific session and channel on the tracer. */ -int ust_app_enable_event_glb(struct ltt_ust_session *usess, +int ust_app_enable_channel_event_glb(struct ltt_ust_session *usess, struct ltt_ust_channel *uchan, struct ltt_ust_event *uevent) { int ret = 0; @@ -4295,7 +6493,95 @@ int ust_app_enable_event_glb(struct ltt_ust_session *usess, /* Get event node */ ua_event = find_ust_app_event(ua_chan->events, uevent->attr.name, - uevent->filter, uevent->attr.loglevel, uevent->exclusion); + uevent->filter, uevent->attr.loglevel, uevent->exclusion, + uevent->attr.token); + if (ua_event == NULL) { + DBG3("UST app enable event %s not found for app PID %d." + "Skipping app", uevent->attr.name, app->pid); + goto next_app; + } + + ret = enable_ust_app_event(ua_sess, ua_event, app); + if (ret < 0) { + pthread_mutex_unlock(&ua_sess->lock); + goto error; + } + next_app: + pthread_mutex_unlock(&ua_sess->lock); + } + +error: + rcu_read_unlock(); + return ret; +} + +/* + * Enable event for a specific session and map on the tracer. + */ +int ust_app_enable_map_event_glb(struct ltt_ust_session *usess, + struct ltt_ust_map *umap, struct ltt_ust_event *uevent) +{ + int ret = 0; + struct lttng_ht_iter iter, uiter; + struct lttng_ht_node_str *ua_map_node; + struct ust_app *app; + struct ust_app_session *ua_sess; + struct ust_app_map *ua_map; + struct ust_app_event *ua_event; + + assert(usess->active); + DBG("UST app enabling event %s for all apps for session id %" PRIu64, + uevent->attr.name, usess->id); + + /* + * NOTE: At this point, this function is called only if the session and + * map passed are already created for all apps. and enabled on the + * tracer also. + */ + + rcu_read_lock(); + + /* For all registered applications */ + cds_lfht_for_each_entry(ust_app_ht->ht, &iter.iter, app, pid_n.node) { + if (!app->compatible) { + /* + * TODO: In time, we should notice the caller of this error by + * telling him that this is a version error. + */ + continue; + } + ua_sess = lookup_session_by_app(usess, app); + if (!ua_sess) { + /* The application has problem or is probably dead. */ + continue; + } + + pthread_mutex_lock(&ua_sess->lock); + + if (ua_sess->deleted) { + pthread_mutex_unlock(&ua_sess->lock); + continue; + } + + /* Lookup map in the ust app session */ + lttng_ht_lookup(ua_sess->maps, (void *)umap->name, &uiter); + ua_map_node = lttng_ht_iter_get_node_str(&uiter); + /* + * It is possible that the map cannot be found is + * the map/event creation occurs concurrently with + * an application exit. + */ + if (!ua_map_node) { + pthread_mutex_unlock(&ua_sess->lock); + continue; + } + + ua_map = caa_container_of(ua_map_node, struct ust_app_map, node); + + /* Get event node */ + ua_event = find_ust_app_event(ua_map->events, uevent->attr.name, + uevent->filter, uevent->attr.loglevel, uevent->exclusion, + uevent->attr.token); if (ua_event == NULL) { DBG3("UST app enable event %s not found for app PID %d." "Skipping app", uevent->attr.name, app->pid); @@ -4320,7 +6606,7 @@ error: * For a specific existing UST session and UST channel, creates the event for * all registered apps. */ -int ust_app_create_event_glb(struct ltt_ust_session *usess, +int ust_app_create_channel_event_glb(struct ltt_ust_session *usess, struct ltt_ust_channel *uchan, struct ltt_ust_event *uevent) { int ret = 0; @@ -4366,7 +6652,7 @@ int ust_app_create_event_glb(struct ltt_ust_session *usess, ua_chan = caa_container_of(ua_chan_node, struct ust_app_channel, node); - ret = create_ust_app_event(ua_sess, ua_chan, uevent, app); + ret = create_ust_app_channel_event(ua_sess, ua_chan, uevent, app); pthread_mutex_unlock(&ua_sess->lock); if (ret < 0) { if (ret != -LTTNG_UST_ERR_EXIST) { @@ -4383,57 +6669,645 @@ int ust_app_create_event_glb(struct ltt_ust_session *usess, return ret; } -/* - * Start tracing for a specific UST session and app. - * - * Called with UST app session lock held. - * - */ static -int ust_app_start_trace(struct ltt_ust_session *usess, struct ust_app *app) +int snapshot_key_values(struct ustctl_daemon_counter *map_handle, + struct lttng_ht *key_to_bucket_index_ht, int cpu, + const char *key_filter, struct lttng_ht *values) { - int ret = 0; - struct ust_app_session *ua_sess; + int ret; + struct lttng_ht_iter key_iter; + struct ust_registry_map_index_ht_entry *map_index_entry; + + /* Iterate over all the formated_key -> counter index */ + cds_lfht_for_each_entry(key_to_bucket_index_ht->ht, + &key_iter.iter, map_index_entry, node.node) { + bool overflow = 0, underflow = 0; + int64_t local_value = 0; + size_t dimension_indexes[1] = {map_index_entry->index}; + + if (key_filter && strcmp(key_filter, + map_index_entry->formated_key) != 0) { + continue; + } - DBG("Starting tracing for ust app pid %d", app->pid); + ret = ustctl_counter_read(map_handle, + dimension_indexes, cpu, &local_value, + &overflow, &underflow); + if (ret) { + ERR("Error getting counter value from the tracer: key = '%s'", + map_index_entry->formated_key); + ret = -1; + goto end; + } - rcu_read_lock(); + map_add_or_increment_map_values(values, + map_index_entry->formated_key, local_value, + underflow, overflow); + } + ret = 0; +end: + return ret; +} - if (!app->compatible) { +static +int ust_app_map_list_values_per_uid_with_bitness_and_cpu( + const struct ltt_ust_session *usess, + const struct ltt_ust_map *umap, + uint32_t app_bitness, + uint32_t cpu, + const char *key_filter, + struct lttng_ht *values) +{ + int ret = 0; + struct lttng_ht_iter iter; + struct buffer_reg_uid *buf_reg_uid; + struct buffer_reg_map *buf_reg_map; + struct ust_registry_session *ust_reg_sess; + struct lttng_ht_node_u64 *ust_reg_map_node; + struct ust_registry_map *ust_reg_map; + + buf_reg_uid = buffer_reg_uid_find(usess->id, app_bitness, usess->uid); + if (!buf_reg_uid) { + /* + * Buffer registry entry for uid not found. Probably no app for + * this UID at the moment. + */ + DBG("No buffer registry entry found for uid: ust-sess-id = %"PRIu64", bitness = %"PRIu32", uid = %d", + usess->id, app_bitness, usess->uid); + /* + * Not an error. Leave the key value pair unchanged and return. + */ + ret = 0; goto end; } - ua_sess = lookup_session_by_app(usess, app); - if (ua_sess == NULL) { - /* The session is in teardown process. Ignore and continue. */ + buf_reg_map = buffer_reg_map_find(umap->id, buf_reg_uid); + if (!buf_reg_uid) { + ERR("Error getting per-uid map buffer registry entry: map-id = %"PRIu64, + umap->id); + ret = -1; goto end; } - pthread_mutex_lock(&ua_sess->lock); + ust_reg_sess = buf_reg_uid->registry->reg.ust; - if (ua_sess->deleted) { - pthread_mutex_unlock(&ua_sess->lock); + /* Get the ust_reg map object from the registry */ + // FIXME: frdeso: This can be changed to ust_registry_map_find() right? + + lttng_ht_lookup(ust_reg_sess->maps, (void *) &umap->id, &iter); + ust_reg_map_node = lttng_ht_iter_get_node_u64(&iter); + if (!ust_reg_map_node) { + ERR("Error getting per-uid map buffer registry entry: map-id = %"PRIu64, + umap->id); + ret = -1; goto end; } + ust_reg_map = caa_container_of(ust_reg_map_node, + struct ust_registry_map, node); - /* Upon restart, we skip the setup, already done */ - if (ua_sess->started) { - goto skip_setup; + ret = snapshot_key_values(buf_reg_map->daemon_counter, + ust_reg_map->key_string_to_bucket_index_ht, + cpu, key_filter, values); + if (ret) { + abort(); } - /* - * Create the metadata for the application. This returns gracefully if a - * metadata was already set for the session. - */ - ret = create_ust_app_metadata(ua_sess, app, usess->consumer); - if (ret < 0) { - goto error_unlock; + + ret = 0; +end: + return ret; +} + +static +int ust_app_map_list_values_per_uid(const struct ltt_ust_session *usess, + const struct ltt_ust_map *umap, + const struct lttng_map_query *query, + struct lttng_map_content *map_content) +{ + int i, ret = 0; + enum lttng_map_query_status map_query_status; + const char *key_filter; + struct lttng_ht *values = NULL; + bool sum_cpus = lttng_map_query_get_config_sum_by_cpu(query); + enum lttng_map_query_config_buffer config_buffer; + enum lttng_map_query_config_cpu config_cpu; + int selected_cpu; + + map_query_status = lttng_map_query_get_key_filter(query, &key_filter); + if (map_query_status == LTTNG_MAP_QUERY_STATUS_NONE) { + key_filter = NULL; + } else if (map_query_status != LTTNG_MAP_QUERY_STATUS_OK) { + ret = -1; + goto end; } - health_code_update(); + config_cpu = lttng_map_query_get_config_cpu(query); + if (config_cpu == LTTNG_MAP_QUERY_CONFIG_CPU_SUBSET) { + unsigned int count; + map_query_status = lttng_map_query_get_cpu_count(query, &count); + assert(map_query_status == LTTNG_MAP_QUERY_STATUS_OK); + assert(count == 1); -skip_setup: - /* This starts the UST tracing */ + map_query_status = lttng_map_query_get_cpu_at_index(query, 0, + &selected_cpu); + assert(map_query_status == LTTNG_MAP_QUERY_STATUS_OK); + } + + config_buffer = lttng_map_query_get_config_buffer(query); + if (config_buffer == LTTNG_MAP_QUERY_CONFIG_BUFFER_UST_UID_SUBSET) { + unsigned int count; + uid_t selected_uid; + + map_query_status = lttng_map_query_get_uid_count(query, &count); + assert(map_query_status == LTTNG_MAP_QUERY_STATUS_OK); + assert(count == 1); + + map_query_status = lttng_map_query_get_uid_at_index(query, 0, + &selected_uid); + assert(map_query_status == LTTNG_MAP_QUERY_STATUS_OK); + + if (selected_uid != usess->uid) { + ret = 0; + goto end; + } + } + + if (sum_cpus) { + values = lttng_ht_new(0, LTTNG_HT_TYPE_STRING); + } + + for (i = 0; i < umap->nr_cpu; i++) { + if (config_cpu == LTTNG_MAP_QUERY_CONFIG_CPU_SUBSET) { + if (selected_cpu != i) { + continue; + } + } + + if (!sum_cpus) { + values = lttng_ht_new(0, LTTNG_HT_TYPE_STRING); + } + + ret = ust_app_map_list_values_per_uid_with_bitness_and_cpu( + usess, umap, 32, i, key_filter, + values); + if (ret) { + abort(); + } + + ret = ust_app_map_list_values_per_uid_with_bitness_and_cpu( + usess, umap, 64, i, key_filter, + values); + if (ret) { + abort(); + } + if (!sum_cpus) { + ret = map_new_content_section(map_content, + LTTNG_MAP_KEY_VALUE_PAIR_LIST_TYPE_UST_PER_UID, + sum_cpus, usess->uid, i, values); + if (ret) { + abort(); + } + + lttng_ht_destroy(values); + } + } + + if (sum_cpus) { + ret = map_new_content_section(map_content, + LTTNG_MAP_KEY_VALUE_PAIR_LIST_TYPE_UST_PER_UID, + sum_cpus, usess->uid, 0, values); + if (ret) { + abort(); + } + lttng_ht_destroy(values); + } + +end: + return ret; +} + +static +int append_dead_app_kv(struct ltt_ust_map *umap, + const char *key_filter, + struct lttng_map_content *map_content) +{ + int ret; + struct lttng_ht *dead_app_kv_ht; + struct map_kv_ht_entry *kv_entry; + struct lttng_ht_iter key_iter; + + struct lttng_ht *values = lttng_ht_new(0, LTTNG_HT_TYPE_STRING); + + pthread_mutex_lock(&(umap->dead_app_kv_values.lock)); + + assert(umap->dead_app_kv_values.dead_app_kv_values_64bits); + dead_app_kv_ht = umap->dead_app_kv_values.dead_app_kv_values_64bits; + + cds_lfht_for_each_entry(dead_app_kv_ht->ht, &key_iter.iter, kv_entry, + node.node) { + if (key_filter && strcmp(key_filter, kv_entry->key) != 0) { + continue; + } + map_add_or_increment_map_values(values, kv_entry->key, + kv_entry->value, kv_entry->has_underflowed, + kv_entry->has_overflowed); + } + + assert(umap->dead_app_kv_values.dead_app_kv_values_32bits); + + dead_app_kv_ht = umap->dead_app_kv_values.dead_app_kv_values_32bits; + cds_lfht_for_each_entry(dead_app_kv_ht->ht, &key_iter.iter, kv_entry, + node.node) { + if (key_filter && strcmp(key_filter, kv_entry->key) != 0) { + continue; + } + map_add_or_increment_map_values(values, kv_entry->key, + kv_entry->value, kv_entry->has_underflowed, + kv_entry->has_overflowed); + } + + pthread_mutex_unlock(&umap->dead_app_kv_values.lock); + + ret = map_new_content_section(map_content, + LTTNG_MAP_KEY_VALUE_PAIR_LIST_TYPE_UST_PER_PID_AGGREGATED, + true, 0, 0, values); + + lttng_ht_destroy(values); + if (ret) { + ERR("Error appending deadapp kv"); + goto end; + } + + + ret = 0; + +end: + return ret; +} + +static +int ust_app_map_list_values_per_pid_with_bitness_and_cpu( + const struct ltt_ust_session *usess, + struct ust_app *app, + struct ltt_ust_map *umap, + uint32_t app_bitness, + uint32_t cpu, + const char *key_filter, + struct lttng_ht *values) +{ + int ret = 0; + + struct lttng_ht_iter map_iter; + struct lttng_ht_node_str *ua_map_node; + struct ust_app_map *ua_map; + struct ust_app_session *ua_sess; + struct ust_registry_session *ust_reg_sess; + struct ust_registry_map *ust_reg_map; + + if (app->bits_per_long != app_bitness) { + ret = 0; + goto end;; + } + + ua_sess = lookup_session_by_app(usess, app); + if (!ua_sess) { + /* Session not associated with this app. */ + ret = 0; + goto end;; + } + + ust_reg_sess = get_session_registry(ua_sess); + if (!ust_reg_sess) { + DBG("Application session is being torn down. Skip application."); + ret = 0; + goto end;; + } + + /* Lookup map in the ust app session */ + lttng_ht_lookup(ua_sess->maps, (void *)umap->name, &map_iter); + ua_map_node = lttng_ht_iter_get_node_str(&map_iter); + + assert(ua_map_node != NULL); + ua_map = caa_container_of(ua_map_node, struct ust_app_map, node); + + pthread_mutex_lock(&ust_reg_sess->lock); + ust_reg_map = ust_registry_map_find(ust_reg_sess, ua_map->key); + pthread_mutex_unlock(&ust_reg_sess->lock); + assert(ust_reg_map); + + ret = snapshot_key_values(ua_map->map_handle, + ust_reg_map->key_string_to_bucket_index_ht, + cpu, key_filter, values); + if (ret) { + ERR("Error snapshoting the content of map"); + goto end; + } + +end: + return ret; +} + +static +int ust_app_map_list_values_per_pid(const struct ltt_ust_session *usess, + struct ltt_ust_map *umap, + const struct lttng_map_query *query, + struct lttng_map_content *map_content) +{ + enum lttng_map_query_status map_query_status; + const char *key_filter; + struct lttng_ht *values; + bool sum_cpus = lttng_map_query_get_config_sum_by_cpu(query); + bool sum_pids = lttng_map_query_get_config_sum_by_pid(query); + enum lttng_map_query_config_cpu config_cpu; + int selected_cpu, i, ret = 0; + struct lttng_ht_iter app_iter; + struct ust_app *app; + + values = lttng_ht_new(0, LTTNG_HT_TYPE_STRING); + + map_query_status = lttng_map_query_get_key_filter(query, &key_filter); + if (map_query_status == LTTNG_MAP_QUERY_STATUS_NONE) { + key_filter = NULL; + } else if (map_query_status != LTTNG_MAP_QUERY_STATUS_OK) { + ret = -1; + goto end; + } + + config_cpu = lttng_map_query_get_config_cpu(query); + if (config_cpu == LTTNG_MAP_QUERY_CONFIG_CPU_SUBSET) { + unsigned int count; + map_query_status = lttng_map_query_get_cpu_count(query, &count); + assert(map_query_status == LTTNG_MAP_QUERY_STATUS_OK); + assert(count == 1); + + map_query_status = lttng_map_query_get_cpu_at_index(query, 0, + &selected_cpu); + assert(map_query_status == LTTNG_MAP_QUERY_STATUS_OK); + } + + /* Sum all cpus and pids on the same table. */ + if (sum_cpus && sum_pids) { + values = lttng_ht_new(0, LTTNG_HT_TYPE_STRING); + } + + if (!sum_cpus && sum_pids) { + /* Iterate over all currently registered apps. */ + for (i = 0; i < umap->nr_cpu; i++) { + values = lttng_ht_new(0, LTTNG_HT_TYPE_STRING); + cds_lfht_for_each_entry(ust_app_ht->ht, &app_iter.iter, app, pid_n.node) { + ret = ust_app_map_list_values_per_pid_with_bitness_and_cpu( + usess, app, umap, 32, i, key_filter, values); + if (ret) { + abort(); + } + ret = ust_app_map_list_values_per_pid_with_bitness_and_cpu( + usess, app, umap, 64, i, key_filter, values); + if (ret) { + abort(); + } + } + ret = map_new_content_section(map_content, + LTTNG_MAP_KEY_VALUE_PAIR_LIST_TYPE_UST_PER_PID, + sum_cpus, app->pid, i, values); + if (ret) { + abort(); + } + + lttng_ht_destroy(values); + } + } else { + /* Iterate over all currently registered apps. */ + cds_lfht_for_each_entry(ust_app_ht->ht, &app_iter.iter, app, pid_n.node) { + + if (sum_cpus && !sum_pids) { + values = lttng_ht_new(0, LTTNG_HT_TYPE_STRING); + } + + for (i = 0; i < umap->nr_cpu; i++) { + + if (config_cpu == LTTNG_MAP_QUERY_CONFIG_CPU_SUBSET) { + if (selected_cpu != i) { + continue; + } + } + + if (!sum_cpus && !sum_pids) { + values = lttng_ht_new(0, LTTNG_HT_TYPE_STRING); + } + + ret = ust_app_map_list_values_per_pid_with_bitness_and_cpu( + usess, app, umap, 32, i, key_filter, values); + if (ret) { + abort(); + } + ret = ust_app_map_list_values_per_pid_with_bitness_and_cpu( + usess, app, umap, 64, i, key_filter, values); + if (ret) { + abort(); + } + + if (!sum_cpus && !sum_pids) { + ret = map_new_content_section(map_content, + LTTNG_MAP_KEY_VALUE_PAIR_LIST_TYPE_UST_PER_PID, + sum_cpus, app->pid, i, values); + if (ret) { + abort(); + } + + lttng_ht_destroy(values); + } + } + if (sum_cpus && !sum_pids) { + ret = map_new_content_section(map_content, + LTTNG_MAP_KEY_VALUE_PAIR_LIST_TYPE_UST_PER_PID, + sum_cpus, app->pid, i, values); + if (ret) { + abort(); + } + + lttng_ht_destroy(values); + } + } + } + + if (sum_cpus && sum_pids) { + ret = map_new_content_section(map_content, + LTTNG_MAP_KEY_VALUE_PAIR_LIST_TYPE_UST_PER_PID, + sum_cpus, 0, 0, values); + if (ret) { + abort(); + } + lttng_ht_destroy(values); + } + + /* Append dead app aggregated key-value pairs. */ + ret = append_dead_app_kv(umap, key_filter, map_content); + if (ret) { + ERR("Error appending values from dead apps map"); + goto end; + } + +end: + return ret; +} + +int ust_app_map_list_values(const struct ltt_ust_session *usess, + struct ltt_ust_map *umap, + const struct lttng_map_query *query, + struct lttng_map_content **map_content) +{ + int ret; + struct lttng_map_content *local_map_content = NULL; + + local_map_content = lttng_map_content_create(usess->buffer_type); + if (!local_map_content) { + ERR("Error creating a map content list"); + ret = -1; + goto end; + } + rcu_read_lock(); + if (usess->buffer_type == LTTNG_BUFFER_PER_UID) { + ret = ust_app_map_list_values_per_uid(usess, umap, query, + local_map_content); + if (ret) { + ERR("Error adding per-uid map value"); + ret = -1; + goto end; + } + } else { + ret = ust_app_map_list_values_per_pid(usess, umap, query, + local_map_content); + if (ret) { + ERR("Error adding per-pid map value"); + ret = -1; + goto end; + } + } + *map_content = local_map_content; + local_map_content = NULL; + ret = 0; +end: + rcu_read_unlock(); + + lttng_map_content_destroy(local_map_content); + return ret; +} + +/* + * For a specific existing UST session and UST map, creates the event for + * all registered apps. + */ +int ust_app_create_map_event_glb(struct ltt_ust_session *usess, + struct ltt_ust_map *umap, struct ltt_ust_event *uevent) +{ + int ret = 0; + struct lttng_ht_iter iter, uiter; + struct lttng_ht_node_str *ua_map_node; + struct ust_app *app; + struct ust_app_session *ua_sess; + struct ust_app_map *ua_map; + + assert(usess->active); + DBG("UST app creating event %s in map %s for all apps for session id %" PRIu64, + uevent->attr.name, umap->name, usess->id); + + rcu_read_lock(); + + /* For all registered applications */ + cds_lfht_for_each_entry(ust_app_ht->ht, &iter.iter, app, pid_n.node) { + if (!app->compatible) { + /* + * TODO: In time, we should notice the caller of this error by + * telling him that this is a version error. + */ + continue; + } + ua_sess = lookup_session_by_app(usess, app); + if (!ua_sess) { + /* The application has problem or is probably dead. */ + continue; + } + + pthread_mutex_lock(&ua_sess->lock); + + if (ua_sess->deleted) { + pthread_mutex_unlock(&ua_sess->lock); + continue; + } + + /* Lookup map in the ust app session */ + lttng_ht_lookup(ua_sess->maps, (void *)umap->name, &uiter); + ua_map_node = lttng_ht_iter_get_node_str(&uiter); + /* If the map is not found, there is a code flow error */ + assert(ua_map_node); + + ua_map = caa_container_of(ua_map_node, struct ust_app_map, node); + + ret = create_ust_app_map_event(ua_sess, ua_map, uevent, app); + assert(!ret); + pthread_mutex_unlock(&ua_sess->lock); + if (ret < 0) { + if (ret != -LTTNG_UST_ERR_EXIST) { + /* Possible value at this point: -ENOMEM. If so, we stop! */ + break; + } + DBG2("UST app event %s already exist on app PID %d", + uevent->attr.name, app->pid); + continue; + } + } + + rcu_read_unlock(); + return ret; +} + +/* + * Start tracing for a specific UST session and app. + * + * Called with UST app session lock held. + * + */ +static +int ust_app_start_trace(struct ltt_ust_session *usess, struct ust_app *app) +{ + int ret = 0; + struct ust_app_session *ua_sess; + + DBG("Starting tracing for ust app pid %d", app->pid); + + rcu_read_lock(); + + if (!app->compatible) { + goto end; + } + + ua_sess = lookup_session_by_app(usess, app); + if (ua_sess == NULL) { + /* The session is in teardown process. Ignore and continue. */ + goto end; + } + + pthread_mutex_lock(&ua_sess->lock); + + if (ua_sess->deleted) { + pthread_mutex_unlock(&ua_sess->lock); + goto end; + } + + if (ua_sess->enabled) { + pthread_mutex_unlock(&ua_sess->lock); + goto end; + } + + /* Upon restart, we skip the setup, already done */ + if (ua_sess->started) { + goto skip_setup; + } + + health_code_update(); + +skip_setup: + /* This starts the UST tracing */ pthread_mutex_lock(&app->sock_lock); ret = ustctl_start_session(app->sock, ua_sess->handle); pthread_mutex_unlock(&app->sock_lock); @@ -4456,6 +7330,7 @@ skip_setup: /* Indicate that the session has been started once */ ua_sess->started = 1; + ua_sess->enabled = 1; pthread_mutex_unlock(&ua_sess->lock); @@ -4545,6 +7420,7 @@ int ust_app_stop_trace(struct ltt_ust_session *usess, struct ust_app *app) } health_code_update(); + ua_sess->enabled = 0; /* Quiescent wait after stopping trace */ pthread_mutex_lock(&app->sock_lock); @@ -4663,7 +7539,7 @@ int ust_app_flush_session(struct ltt_ust_session *usess) /* Flush all per UID buffers associated to that session. */ cds_list_for_each_entry(reg, &usess->buffer_reg_uid_list, lnode) { struct ust_registry_session *ust_session_reg; - struct buffer_reg_channel *reg_chan; + struct buffer_reg_channel *buf_reg_chan; struct consumer_socket *socket; /* Get consumer socket to use to push the metadata.*/ @@ -4675,13 +7551,13 @@ int ust_app_flush_session(struct ltt_ust_session *usess) } cds_lfht_for_each_entry(reg->registry->channels->ht, &iter.iter, - reg_chan, node.node) { + buf_reg_chan, node.node) { /* * The following call will print error values so the return * code is of little importance because whatever happens, we * have to try them all. */ - (void) consumer_flush_channel(socket, reg_chan->consumer_key); + (void) consumer_flush_channel(socket, buf_reg_chan->consumer_key); } ust_session_reg = reg->registry->reg.ust; @@ -4810,7 +7686,7 @@ int ust_app_clear_quiescent_session(struct ltt_ust_session *usess) */ cds_list_for_each_entry(reg, &usess->buffer_reg_uid_list, lnode) { struct consumer_socket *socket; - struct buffer_reg_channel *reg_chan; + struct buffer_reg_channel *buf_reg_chan; /* Get associated consumer socket.*/ socket = consumer_find_socket_by_bitness( @@ -4824,7 +7700,7 @@ int ust_app_clear_quiescent_session(struct ltt_ust_session *usess) } cds_lfht_for_each_entry(reg->registry->channels->ht, - &iter.iter, reg_chan, node.node) { + &iter.iter, buf_reg_chan, node.node) { /* * The following call will print error values so * the return code is of little importance @@ -4832,7 +7708,7 @@ int ust_app_clear_quiescent_session(struct ltt_ust_session *usess) * all. */ (void) consumer_clear_quiescent_channel(socket, - reg_chan->consumer_key); + buf_reg_chan->consumer_key); } } break; @@ -5035,6 +7911,36 @@ end: return ret; } +/* The ua_sess lock must be held by the caller. */ +static +int find_or_create_ust_app_map( + struct ltt_ust_session *usess, + struct ust_app_session *ua_sess, + struct ust_app *app, + struct ltt_ust_map *umap, + struct ust_app_map **ua_map) +{ + int ret = 0; + struct lttng_ht_iter iter; + struct lttng_ht_node_str *ua_map_node; + + lttng_ht_lookup(ua_sess->maps, (void *) umap->name, &iter); + ua_map_node = lttng_ht_iter_get_node_str(&iter); + if (ua_map_node) { + *ua_map = caa_container_of(ua_map_node, + struct ust_app_map, node); + goto end; + } + + DBG("UST map id = %"PRIu64" not found. Creating it.", umap->id); + ret = ust_app_map_create(usess, ua_sess, umap, app, ua_map); + if (ret) { + goto end; + } +end: + return ret; +} + static int ust_app_channel_synchronize_event(struct ust_app_channel *ua_chan, struct ltt_ust_event *uevent, struct ust_app_session *ua_sess, @@ -5044,9 +7950,11 @@ int ust_app_channel_synchronize_event(struct ust_app_channel *ua_chan, struct ust_app_event *ua_event = NULL; ua_event = find_ust_app_event(ua_chan->events, uevent->attr.name, - uevent->filter, uevent->attr.loglevel, uevent->exclusion); + uevent->filter, uevent->attr.loglevel, uevent->exclusion, + uevent->attr.token); if (!ua_event) { - ret = create_ust_app_event(ua_sess, ua_chan, uevent, app); + ret = create_ust_app_channel_event(ua_sess, ua_chan, uevent, app); + if (ret < 0) { goto end; } @@ -5057,22 +7965,311 @@ int ust_app_channel_synchronize_event(struct ust_app_channel *ua_chan, disable_ust_app_event(ua_sess, ua_event, app); } } - + +end: + return ret; +} + +/* Called with RCU read-side lock held. */ +static +int ust_app_map_synchronize_event(struct ust_app_map *ua_map, + struct ltt_ust_event *uevent, struct ust_app_session *ua_sess, + struct ust_app *app) +{ + int ret = 0; + struct ust_app_event *ua_event = NULL; + + ua_event = find_ust_app_event(ua_map->events, uevent->attr.name, + uevent->filter, uevent->attr.loglevel, uevent->exclusion, + uevent->attr.token); + if (!ua_event) { + ret = create_ust_app_map_event(ua_sess, ua_map, uevent, app); + if (ret < 0) { + goto end; + } + } else { + if (ua_event->enabled != uevent->enabled) { + ret = uevent->enabled ? + enable_ust_app_event(ua_sess, ua_event, app) : + disable_ust_app_event(ua_sess, ua_event, app); + } + } + +end: + return ret; +} + +static +void ust_app_synchronize_event_notifier_rules(struct ust_app *app) +{ + int ret = 0; + enum lttng_error_code ret_code; + enum lttng_trigger_status t_status; + struct lttng_ht_iter app_trigger_iter; + struct lttng_triggers *triggers = NULL; + struct ust_app_event_notifier_rule *event_notifier_rule; + unsigned int count, i; + + /* + * Currrently, registering or unregistering a trigger with an + * event rule condition causes a full synchronization of the event + * notifiers. + * + * The first step attempts to add an event notifier for all registered + * triggers that apply to the user space tracers. Then, the + * application's event notifiers rules are all checked against the list + * of registered triggers. Any event notifier that doesn't have a + * matching trigger can be assumed to have been disabled. + * + * All of this is inefficient, but is put in place to get the feature + * rolling as it is simpler at this moment. It will be optimized Soon™ + * to allow the state of enabled + * event notifiers to be synchronized in a piece-wise way. + */ + + /* Get all triggers using uid 0 (root) */ + ret_code = notification_thread_command_list_triggers( + notification_thread_handle, 0, &triggers); + if (ret_code != LTTNG_OK) { + ret = -1; + goto end; + } + + assert(triggers); + + t_status = lttng_triggers_get_count(triggers, &count); + if (t_status != LTTNG_TRIGGER_STATUS_OK) { + ret = -1; + goto end; + } + + for (i = 0; i < count; i++) { + const struct lttng_condition *condition; + const struct lttng_event_rule *event_rule; + struct lttng_trigger *trigger; + const struct ust_app_event_notifier_rule *looked_up_event_notifier_rule; + enum lttng_condition_status condition_status; + uint64_t token; + + trigger = lttng_triggers_borrow_mutable_at_index(triggers, i); + assert(trigger); + + token = lttng_trigger_get_tracer_token(trigger); + condition = lttng_trigger_get_const_condition(trigger); + + if (!lttng_trigger_needs_tracer_notifier(trigger)) { + continue; + } + + condition_status = lttng_condition_on_event_get_rule(condition, &event_rule); + assert(condition_status == LTTNG_CONDITION_STATUS_OK); + + if (lttng_event_rule_get_domain_type(event_rule) == LTTNG_DOMAIN_KERNEL) { + /* Skip kernel related triggers. */ + continue; + } + + /* + * Find or create the associated token event rule. The caller + * holds the RCU read lock, so this is safe to call without + * explicitly acquiring it here. + */ + looked_up_event_notifier_rule = find_ust_app_event_notifier_rule( + app->token_to_event_notifier_rule_ht, token); + if (!looked_up_event_notifier_rule) { + ret = create_ust_app_event_notifier_rule(trigger, app); + if (ret < 0) { + goto end; + } + } + } + + rcu_read_lock(); + /* Remove all unknown event sources from the app. */ + cds_lfht_for_each_entry (app->token_to_event_notifier_rule_ht->ht, + &app_trigger_iter.iter, event_notifier_rule, + node.node) { + const uint64_t app_token = event_notifier_rule->token; + bool found = false; + + /* + * Check if the app event trigger still exists on the + * notification side. + */ + for (i = 0; i < count; i++) { + uint64_t notification_thread_token; + const struct lttng_trigger *trigger = + lttng_triggers_get_at_index( + triggers, i); + + assert(trigger); + + notification_thread_token = + lttng_trigger_get_tracer_token(trigger); + + if (notification_thread_token == app_token) { + found = true; + break; + } + } + + if (found) { + /* Still valid. */ + continue; + } + + /* + * This trigger was unregistered, disable it on the tracer's + * side. + */ + ret = lttng_ht_del(app->token_to_event_notifier_rule_ht, + &app_trigger_iter); + assert(ret == 0); + + /* Callee logs errors. */ + (void) disable_ust_object(app, event_notifier_rule->obj); + + delete_ust_app_event_notifier_rule( + app->sock, event_notifier_rule, app); + } + + rcu_read_unlock(); + +end: + lttng_triggers_destroy(triggers); + return; +} + +/* + * Called with RCU read-side lock held. + */ +static +void ust_app_synchronize_all_channels(struct ltt_ust_session *usess, + struct ust_app_session *ua_sess, + struct ust_app *app) +{ + int ret = 0; + struct cds_lfht_iter uchan_iter; + struct ltt_ust_channel *uchan; + + assert(usess); + assert(ua_sess); + assert(app); + + cds_lfht_for_each_entry(usess->domain_global.channels->ht, &uchan_iter, + uchan, node.node) { + struct ust_app_channel *ua_chan; + struct cds_lfht_iter uevent_iter; + struct ltt_ust_event *uevent; + + /* + * Search for a matching ust_app_channel. If none is found, + * create it. Creating the channel will cause the ua_chan + * structure to be allocated, the channel buffers to be + * allocated (if necessary) and sent to the application, and + * all enabled contexts will be added to the channel. + */ + ret = find_or_create_ust_app_channel(usess, ua_sess, + app, uchan, &ua_chan); + if (ret) { + /* Tracer is probably gone or ENOMEM. */ + goto end; + } + + if (!ua_chan) { + /* ua_chan will be NULL for the metadata channel */ + continue; + } + + cds_lfht_for_each_entry(uchan->events->ht, &uevent_iter, uevent, + node.node) { + ret = ust_app_channel_synchronize_event(ua_chan, + uevent, ua_sess, app); + if (ret) { + goto end; + } + } + + if (ua_chan->enabled != uchan->enabled) { + ret = uchan->enabled ? + enable_ust_app_channel(ua_sess, uchan, app) : + disable_ust_app_channel(ua_sess, ua_chan, app); + if (ret) { + goto end; + } + } + } +end: + return; +} + +/* + * Called with RCU read-side lock held. + */ +static +void ust_app_synchronize_all_maps(struct ltt_ust_session *usess, + struct ust_app_session *ua_sess, + struct ust_app *app) +{ + int ret = 0; + struct cds_lfht_iter umap_iter; + struct ltt_ust_map *umap; + + assert(usess); + assert(ua_sess); + assert(app); + + cds_lfht_for_each_entry(usess->domain_global.maps->ht, &umap_iter, + umap, node.node) { + struct ust_app_map *ua_map; + struct cds_lfht_iter uevent_iter; + struct ltt_ust_event *uevent; + + DBG("Synchronizing UST map id = %"PRIu64, umap->id); + + ret = find_or_create_ust_app_map(usess, ua_sess, + app, umap, &ua_map); + if (ret) { + /* Tracer is probably gone or ENOMEM. */ + goto end; + } + + DBG("Synchronizing all events of UST map id = %"PRIu64, umap->id); + cds_lfht_for_each_entry(umap->events->ht, &uevent_iter, uevent, + node.node) { + ret = ust_app_map_synchronize_event(ua_map, + uevent, ua_sess, app); + if (ret) { + goto end; + } + } + + if (ua_map->enabled != umap->enabled) { + if (umap->enabled) { + DBG("Map disabled on the tracer side but shouldn't"); + ret = enable_ust_app_map(ua_sess, umap, app); + } else { + DBG("Map enabled on the tracer side but shouldn't"); + ret = disable_ust_app_map(ua_sess, ua_map, app); + } + if (ret) { + goto end; + } + } + } end: - return ret; + return; } /* * The caller must ensure that the application is compatible and is tracked - * by the PID tracker. + * by the process attribute trackers. */ static void ust_app_synchronize(struct ltt_ust_session *usess, struct ust_app *app) { int ret = 0; - struct cds_lfht_iter uchan_iter; - struct ltt_ust_channel *uchan; struct ust_app_session *ua_sess = NULL; /* @@ -5088,56 +8285,31 @@ void ust_app_synchronize(struct ltt_ust_session *usess, } assert(ua_sess); + + rcu_read_lock(); + pthread_mutex_lock(&ua_sess->lock); if (ua_sess->deleted) { pthread_mutex_unlock(&ua_sess->lock); goto end; } + ust_app_synchronize_all_channels(usess, ua_sess, app); + ust_app_synchronize_all_maps(usess, ua_sess, app); - rcu_read_lock(); - cds_lfht_for_each_entry(usess->domain_global.channels->ht, &uchan_iter, - uchan, node.node) { - struct ust_app_channel *ua_chan; - struct cds_lfht_iter uevent_iter; - struct ltt_ust_event *uevent; - - /* - * Search for a matching ust_app_channel. If none is found, - * create it. Creating the channel will cause the ua_chan - * structure to be allocated, the channel buffers to be - * allocated (if necessary) and sent to the application, and - * all enabled contexts will be added to the channel. - */ - ret = find_or_create_ust_app_channel(usess, ua_sess, - app, uchan, &ua_chan); - if (ret) { - /* Tracer is probably gone or ENOMEM. */ - goto error_unlock; - } - - if (!ua_chan) { - /* ua_chan will be NULL for the metadata channel */ - continue; - } - - cds_lfht_for_each_entry(uchan->events->ht, &uevent_iter, uevent, - node.node) { - ret = ust_app_channel_synchronize_event(ua_chan, - uevent, ua_sess, app); - if (ret) { - goto error_unlock; - } - } - - if (ua_chan->enabled != uchan->enabled) { - ret = uchan->enabled ? - enable_ust_app_channel(ua_sess, uchan, app) : - disable_ust_app_channel(ua_sess, ua_chan, app); - if (ret) { - goto error_unlock; - } - } + /* + * Create the metadata for the application. This returns gracefully if a + * metadata was already set for the session. + * + * The metadata channel must be created after the data channels as the + * consumer daemon assumes this ordering. When interacting with a relay + * daemon, the consumer will use this assumption to send the + * "STREAMS_SENT" message to the relay daemon. + */ + ret = create_ust_app_metadata(ua_sess, app, usess->consumer); + if (ret < 0) { + goto error_unlock; } + rcu_read_unlock(); end: @@ -5184,7 +8356,14 @@ void ust_app_global_update(struct ltt_ust_session *usess, struct ust_app *app) if (!app->compatible) { return; } - if (trace_ust_pid_tracker_lookup(usess, app->pid)) { + if (trace_ust_id_tracker_lookup(LTTNG_PROCESS_ATTR_VIRTUAL_PROCESS_ID, + usess, app->pid) && + trace_ust_id_tracker_lookup( + LTTNG_PROCESS_ATTR_VIRTUAL_USER_ID, + usess, app->uid) && + trace_ust_id_tracker_lookup( + LTTNG_PROCESS_ATTR_VIRTUAL_GROUP_ID, + usess, app->gid)) { /* * Synchronize the application's internal tracing configuration * and start tracing. @@ -5196,6 +8375,30 @@ void ust_app_global_update(struct ltt_ust_session *usess, struct ust_app *app) } } +/* + * Add all event notifiers to an application. + * + * Called with session lock held. + * Called with RCU read-side lock held. + */ +void ust_app_global_update_event_notifier_rules(struct ust_app *app) +{ + DBG2("UST application global event notifier rules update: app = '%s' (ppid: %d)", + app->name, app->ppid); + + if (!app->compatible) { + return; + } + + if (app->event_notifier_group.object == NULL) { + WARN("UST app global update of event notifiers for app skipped since communication handle is null: app = '%s' (ppid: %d)", + app->name, app->ppid); + return; + } + + ust_app_synchronize_event_notifier_rules(app); +} + /* * Called with session lock held. */ @@ -5211,6 +8414,33 @@ void ust_app_global_update_all(struct ltt_ust_session *usess) rcu_read_unlock(); } +void ust_app_global_update_all_event_notifier_rules(void) +{ + struct lttng_ht_iter iter; + struct ust_app *app; + + rcu_read_lock(); + cds_lfht_for_each_entry(ust_app_ht->ht, &iter.iter, app, pid_n.node) { + ust_app_global_update_event_notifier_rules(app); + } + + rcu_read_unlock(); +} + +void ust_app_update_event_notifier_error_count(struct lttng_trigger *trigger) +{ + uint64_t error_count = 0; + enum event_notifier_error_accounting_status status; + struct lttng_condition *condition = lttng_trigger_get_condition(trigger); + + status = event_notifier_error_accounting_get_count(trigger, &error_count); + if (status != EVENT_NOTIFIER_ERROR_ACCOUNTING_STATUS_OK) { + ERR("Error getting trigger error count."); + } + + lttng_condition_on_event_set_error_count(condition, error_count); +} + /* * Add context to a specific channel for global UST domain. */ @@ -5357,7 +8587,7 @@ static struct ust_app_channel *find_channel_by_objd(struct ust_app *app, assert(app); - lttng_ht_lookup(app->ust_objd, (void *)((unsigned long) objd), &iter); + lttng_ht_lookup(app->ust_chan_objd, (void *)((unsigned long) objd), &iter); node = lttng_ht_iter_get_node_ulong(&iter); if (node == NULL) { DBG2("UST app channel find by objd %d not found", objd); @@ -5370,6 +8600,33 @@ error: return ua_chan; } +/* + * Return a ust app map object using the application object and the map + * object descriptor has a key. If not found, NULL is returned. A RCU read side + * lock MUST be acquired before calling this function. + */ +static struct ust_app_map *find_map_by_objd(struct ust_app *app, + int objd) +{ + struct lttng_ht_node_ulong *node; + struct lttng_ht_iter iter; + struct ust_app_map *ua_map = NULL; + + assert(app); + + lttng_ht_lookup(app->ust_map_objd, (void *)((unsigned long) objd), &iter); + node = lttng_ht_iter_get_node_ulong(&iter); + if (node == NULL) { + DBG2("UST app map find by objd %d not found", objd); + goto error; + } + + ua_map = caa_container_of(node, struct ust_app_map, ust_objd_node); + +error: + return ua_map; +} + /* * Reply to a register channel notification from an application on the notify * socket. The channel metadata is also created. @@ -5389,7 +8646,7 @@ static int reply_ust_register_channel(int sock, int cobjd, struct ust_app_channel *ua_chan; struct ust_app_session *ua_sess; struct ust_registry_session *registry; - struct ust_registry_channel *chan_reg; + struct ust_registry_channel *ust_reg_chan; rcu_read_lock(); @@ -5430,30 +8687,30 @@ static int reply_ust_register_channel(int sock, int cobjd, pthread_mutex_lock(®istry->lock); - chan_reg = ust_registry_channel_find(registry, chan_reg_key); - assert(chan_reg); + ust_reg_chan = ust_registry_channel_find(registry, chan_reg_key); + assert(ust_reg_chan); - if (!chan_reg->register_done) { + if (!ust_reg_chan->register_done) { /* * TODO: eventually use the registry event count for * this channel to better guess header type for per-pid * buffers. */ type = USTCTL_CHANNEL_HEADER_LARGE; - chan_reg->nr_ctx_fields = nr_fields; - chan_reg->ctx_fields = fields; + ust_reg_chan->nr_ctx_fields = nr_fields; + ust_reg_chan->ctx_fields = fields; fields = NULL; - chan_reg->header_type = type; + ust_reg_chan->header_type = type; } else { /* Get current already assigned values. */ - type = chan_reg->header_type; + type = ust_reg_chan->header_type; } /* Channel id is set during the object creation. */ - chan_id = chan_reg->chan_id; + chan_id = ust_reg_chan->chan_id; /* Append to metadata */ - if (!chan_reg->metadata_dumped) { - ret_code = ust_metadata_channel_statedump(registry, chan_reg); + if (!ust_reg_chan->metadata_dumped) { + ret_code = ust_metadata_channel_statedump(registry, ust_reg_chan); if (ret_code) { ERR("Error appending channel metadata (errno = %d)", ret_code); goto reply; @@ -5476,7 +8733,7 @@ reply: } /* This channel registry registration is completed. */ - chan_reg->register_done = 1; + ust_reg_chan->register_done = 1; error: pthread_mutex_unlock(®istry->lock); @@ -5486,60 +8743,104 @@ error_rcu_unlock: return ret; } -/* - * Add event to the UST channel registry. When the event is added to the - * registry, the metadata is also created. Once done, this replies to the - * application with the appropriate error code. - * - * The session UST registry lock is acquired in the function. - * - * On success 0 is returned else a negative value. - */ -static int add_event_ust_registry(int sock, int sobjd, int cobjd, char *name, +static int add_event_ust_chan_registry(int sock, struct ust_app *ua, + struct ust_app_channel *ua_chan, int sobjd, int cobjd, char *name, char *sig, size_t nr_fields, struct ustctl_field *fields, int loglevel_value, char *model_emf_uri) { int ret, ret_code; uint32_t event_id = 0; uint64_t chan_reg_key; - struct ust_app *app; - struct ust_app_channel *ua_chan; struct ust_app_session *ua_sess; struct ust_registry_session *registry; + /* + * The counter index is unused for channel events. It's only used for + * map events. + */ + uint64_t counter_index = 0; - rcu_read_lock(); + assert(ua_chan->session); + ua_sess = ua_chan->session; - /* Lookup application. If not found, there is a code flow error. */ - app = find_app_by_notify_sock(sock); - if (!app) { - DBG("Application socket %d is being torn down. Abort event notify", - sock); + registry = get_session_registry(ua_sess); + if (!registry) { + DBG("Application session is being torn down. Abort event notify"); ret = 0; - goto error_rcu_unlock; + goto error; } - /* Lookup channel by UST object descriptor. */ - ua_chan = find_channel_by_objd(app, cobjd); - if (!ua_chan) { - DBG("Application channel is being torn down. Abort event notify"); - ret = 0; - goto error_rcu_unlock; + if (ua_sess->buffer_type == LTTNG_BUFFER_PER_UID) { + chan_reg_key = ua_chan->tracing_channel_id; + } else { + chan_reg_key = ua_chan->key; } - assert(ua_chan->session); - ua_sess = ua_chan->session; + pthread_mutex_lock(®istry->lock); + + /* + * From this point on, this call acquires the ownership of the sig, fields + * and model_emf_uri meaning any free are done inside it if needed. These + * three variables MUST NOT be read/write after this. + */ + ret_code = ust_registry_chan_create_event(registry, chan_reg_key, + sobjd, cobjd, name, sig, nr_fields, fields, + loglevel_value, model_emf_uri, ua_sess->buffer_type, + &event_id, ua); + sig = NULL; + fields = NULL; + model_emf_uri = NULL; + + /* + * The return value is returned to ustctl so in case of an error, the + * application can be notified. In case of an error, it's important not to + * return a negative error or else the application will get closed. + */ + ret = ustctl_reply_register_event(sock, event_id, counter_index, ret_code); + if (ret < 0) { + if (ret != -EPIPE && ret != -LTTNG_UST_ERR_EXITING) { + ERR("UST app reply event failed with ret %d", ret); + } else { + DBG3("UST app reply event failed. Application died"); + } + /* + * No need to wipe the create event since the application socket will + * get close on error hence cleaning up everything by itself. + */ + goto error; + } + + DBG3("UST registry event %s with id %" PRId32 " added successfully", + name, event_id); + +error: + pthread_mutex_unlock(®istry->lock); + return ret; +} + +static int add_event_ust_map_registry(int sock, struct ust_app *ua, + struct ust_app_map *ua_map, int sobjd, int cobjd, char *name, + char *sig, size_t nr_fields, struct ustctl_field *fields, + int loglevel_value, char *model_emf_uri, uint64_t tracer_token) +{ + int ret, ret_code; + uint64_t map_reg_key, counter_index; + struct ust_app_session *ua_sess; + struct ust_registry_session *registry; + + assert(ua_map->session); + ua_sess = ua_map->session; registry = get_session_registry(ua_sess); if (!registry) { DBG("Application session is being torn down. Abort event notify"); ret = 0; - goto error_rcu_unlock; + goto error; } if (ua_sess->buffer_type == LTTNG_BUFFER_PER_UID) { - chan_reg_key = ua_chan->tracing_channel_id; + map_reg_key = ua_map->tracing_map_id; } else { - chan_reg_key = ua_chan->key; + map_reg_key = ua_map->key; } pthread_mutex_lock(®istry->lock); @@ -5549,43 +8850,112 @@ static int add_event_ust_registry(int sock, int sobjd, int cobjd, char *name, * and model_emf_uri meaning any free are done inside it if needed. These * three variables MUST NOT be read/write after this. */ - ret_code = ust_registry_create_event(registry, chan_reg_key, + DBG("Registry_map_create_event on map=%"PRIu64" with token=%"PRIu64, + map_reg_key, tracer_token); + ret_code = ust_registry_map_create_event(registry, map_reg_key, sobjd, cobjd, name, sig, nr_fields, fields, loglevel_value, model_emf_uri, ua_sess->buffer_type, - &event_id, app); + tracer_token, &counter_index, ua); + assert(!ret_code); + sig = NULL; fields = NULL; model_emf_uri = NULL; - /* - * The return value is returned to ustctl so in case of an error, the - * application can be notified. In case of an error, it's important not to - * return a negative error or else the application will get closed. - */ - ret = ustctl_reply_register_event(sock, event_id, ret_code); - if (ret < 0) { - if (ret != -EPIPE && ret != -LTTNG_UST_ERR_EXITING) { - ERR("UST app reply event failed with ret %d", ret); - } else { - DBG3("UST app reply event failed. Application died"); + /* + * The return value is returned to ustctl so in case of an error, the + * application can be notified. In case of an error, it's important not to + * return a negative error or else the application will get closed. + */ + ret = ustctl_reply_register_event(sock, counter_index, counter_index, + ret_code); + if (ret < 0) { + if (ret != -EPIPE && ret != -LTTNG_UST_ERR_EXITING) { + ERR("UST app reply event failed with ret %d", ret); + } else { + DBG3("UST app reply event failed. Application died"); + } + /* + * No need to wipe the create event since the application socket will + * get close on error hence cleaning up everything by itself. + */ + goto error; + } + + DBG3("UST registry map event %s with counter index %" PRIu64 " added successfully", + name, counter_index); + +error: + pthread_mutex_unlock(®istry->lock); + return ret; +} + + +/* + * Add event to the UST channel registry. When the event is added to the + * registry, the metadata is also created. Once done, this replies to the + * application with the appropriate error code. + * + * The session UST registry lock is acquired in the function. + * + * On success 0 is returned else a negative value. + */ +static int add_event_ust_registry(int sock, int sobjd, int cobjd, char *name, + char *sig, size_t nr_fields, struct ustctl_field *fields, + int loglevel_value, char *model_emf_uri, uint64_t tracer_token) +{ + int ret; + struct ust_app *app; + struct ust_app_channel *ua_chan = NULL; + struct ust_app_map *ua_map = NULL; + + rcu_read_lock(); + + /* Lookup application. If not found, there is a code flow error. */ + app = find_app_by_notify_sock(sock); + if (!app) { + DBG("Application socket %d is being torn down. Abort event notify", + sock); + ret = 0; + goto end; + } + + /* Lookup channel by UST object descriptor. */ + ua_chan = find_channel_by_objd(app, cobjd); + if (ua_chan) { + ret = add_event_ust_chan_registry(sock, app, ua_chan, sobjd, cobjd, + name, sig, nr_fields, fields, loglevel_value, + model_emf_uri); + if (ret) { + ERR("Error adding channel event to registry: event_name = '%s'", name); } - /* - * No need to wipe the create event since the application socket will - * get close on error hence cleaning up everything by itself. - */ - goto error; + goto found; } - DBG3("UST registry event %s with id %" PRId32 " added successfully", - name, event_id); + /* Lookup map by UST object descriptor. */ + ua_map = find_map_by_objd(app, cobjd); + if (ua_map) { + ret = add_event_ust_map_registry(sock, app, ua_map, sobjd, cobjd, + name, sig, nr_fields, fields, loglevel_value, + model_emf_uri, tracer_token); + if (ret) { + ERR("Error adding map event to registry: event_name = '%s'", name); + goto end; + } + goto found; + } -error: - pthread_mutex_unlock(®istry->lock); -error_rcu_unlock: + if (!ua_chan && !ua_map) { + DBG("Application event container is being torn down. Abort event notify"); + ret = 0; + goto end; + } + +found: + ret = 0; + +end: rcu_read_unlock(); - free(sig); - free(fields); - free(model_emf_uri); return ret; } @@ -5701,13 +9071,14 @@ int ust_app_recv_notify(int sock) int sobjd, cobjd, loglevel_value; char name[LTTNG_UST_SYM_NAME_LEN], *sig, *model_emf_uri; size_t nr_fields; + uint64_t tracer_token = 0; struct ustctl_field *fields; DBG2("UST app ustctl register event received"); ret = ustctl_recv_register_event(sock, &sobjd, &cobjd, name, &loglevel_value, &sig, &nr_fields, &fields, - &model_emf_uri); + &model_emf_uri, &tracer_token); if (ret < 0) { if (ret != -EPIPE && ret != -LTTNG_UST_ERR_EXITING) { ERR("UST app recv event failed with ret %d", ret); @@ -5724,7 +9095,7 @@ int ust_app_recv_notify(int sock) * to the this function. */ ret = add_event_ust_registry(sock, sobjd, cobjd, name, sig, nr_fields, - fields, loglevel_value, model_emf_uri); + fields, loglevel_value, model_emf_uri, tracer_token); if (ret < 0) { goto error; } @@ -5918,9 +9289,10 @@ enum lttng_error_code ust_app_snapshot_record( struct buffer_reg_uid *reg; cds_list_for_each_entry(reg, &usess->buffer_reg_uid_list, lnode) { - struct buffer_reg_channel *reg_chan; + struct buffer_reg_channel *buf_reg_chan; struct consumer_socket *socket; char pathname[PATH_MAX]; + size_t consumer_path_offset = 0; if (!reg->registry->reg.ust->metadata_key) { /* Skip since no metadata is present */ @@ -5936,12 +9308,8 @@ enum lttng_error_code ust_app_snapshot_record( } memset(pathname, 0, sizeof(pathname)); - /* - * DEFAULT_UST_TRACE_UID_PATH already contains a path - * separator. - */ ret = snprintf(pathname, sizeof(pathname), - DEFAULT_UST_TRACE_DIR DEFAULT_UST_TRACE_UID_PATH, + DEFAULT_UST_TRACE_DIR "/" DEFAULT_UST_TRACE_UID_PATH, reg->uid, reg->bits_per_long); if (ret < 0) { PERROR("snprintf snapshot path"); @@ -5950,18 +9318,19 @@ enum lttng_error_code ust_app_snapshot_record( } /* Free path allowed on previous iteration. */ free(trace_path); - trace_path = setup_channel_trace_path(usess->consumer, pathname); + trace_path = setup_channel_trace_path(usess->consumer, pathname, + &consumer_path_offset); if (!trace_path) { status = LTTNG_ERR_INVALID; goto error; } - /* Add the UST default trace dir to path. */ + /* Add the UST default trace dir to path. */ cds_lfht_for_each_entry(reg->registry->channels->ht, &iter.iter, - reg_chan, node.node) { + buf_reg_chan, node.node) { status = consumer_snapshot_channel(socket, - reg_chan->consumer_key, + buf_reg_chan->consumer_key, output, 0, usess->uid, - usess->gid, trace_path, wait, + usess->gid, &trace_path[consumer_path_offset], wait, nb_packets_per_stream); if (status != LTTNG_OK) { goto error; @@ -5969,7 +9338,8 @@ enum lttng_error_code ust_app_snapshot_record( } status = consumer_snapshot_channel(socket, reg->registry->reg.ust->metadata_key, output, 1, - usess->uid, usess->gid, trace_path, wait, 0); + usess->uid, usess->gid, &trace_path[consumer_path_offset], + wait, 0); if (status != LTTNG_OK) { goto error; } @@ -5985,6 +9355,7 @@ enum lttng_error_code ust_app_snapshot_record( struct ust_app_session *ua_sess; struct ust_registry_session *registry; char pathname[PATH_MAX]; + size_t consumer_path_offset = 0; ua_sess = lookup_session_by_app(usess, app); if (!ua_sess) { @@ -6002,7 +9373,7 @@ enum lttng_error_code ust_app_snapshot_record( /* Add the UST default trace dir to path. */ memset(pathname, 0, sizeof(pathname)); - ret = snprintf(pathname, sizeof(pathname), DEFAULT_UST_TRACE_DIR "%s", + ret = snprintf(pathname, sizeof(pathname), DEFAULT_UST_TRACE_DIR "/%s", ua_sess->path); if (ret < 0) { status = LTTNG_ERR_INVALID; @@ -6011,20 +9382,19 @@ enum lttng_error_code ust_app_snapshot_record( } /* Free path allowed on previous iteration. */ free(trace_path); - trace_path = setup_channel_trace_path(usess->consumer, pathname); + trace_path = setup_channel_trace_path(usess->consumer, pathname, + &consumer_path_offset); if (!trace_path) { status = LTTNG_ERR_INVALID; goto error; } - cds_lfht_for_each_entry(ua_sess->channels->ht, &chan_iter.iter, + cds_lfht_for_each_entry(ua_sess->channels->ht, &chan_iter.iter, ua_chan, node.node) { status = consumer_snapshot_channel(socket, ua_chan->key, output, 0, - ua_sess->effective_credentials - .uid, - ua_sess->effective_credentials - .gid, - trace_path, wait, + lttng_credentials_get_uid(&ua_sess->effective_credentials), + lttng_credentials_get_gid(&ua_sess->effective_credentials), + &trace_path[consumer_path_offset], wait, nb_packets_per_stream); switch (status) { case LTTNG_OK: @@ -6043,9 +9413,9 @@ enum lttng_error_code ust_app_snapshot_record( } status = consumer_snapshot_channel(socket, registry->metadata_key, output, 1, - ua_sess->effective_credentials.uid, - ua_sess->effective_credentials.gid, - trace_path, wait, 0); + lttng_credentials_get_uid(&ua_sess->effective_credentials), + lttng_credentials_get_gid(&ua_sess->effective_credentials), + &trace_path[consumer_path_offset], wait, 0); switch (status) { case LTTNG_OK: break; @@ -6086,19 +9456,19 @@ uint64_t ust_app_get_size_one_more_packet_per_stream( struct buffer_reg_uid *reg; cds_list_for_each_entry(reg, &usess->buffer_reg_uid_list, lnode) { - struct buffer_reg_channel *reg_chan; + struct buffer_reg_channel *buf_reg_chan; rcu_read_lock(); cds_lfht_for_each_entry(reg->registry->channels->ht, &iter.iter, - reg_chan, node.node) { - if (cur_nr_packets >= reg_chan->num_subbuf) { + buf_reg_chan, node.node) { + if (cur_nr_packets >= buf_reg_chan->num_subbuf) { /* * Don't take channel into account if we * already grab all its packets. */ continue; } - tot_size += reg_chan->subbuf_size * reg_chan->stream_count; + tot_size += buf_reg_chan->subbuf_size * buf_reg_chan->stream_count; } rcu_read_unlock(); } @@ -6287,24 +9657,276 @@ int ust_app_regenerate_statedump_all(struct ltt_ust_session *usess) continue; } - ret = ust_app_regenerate_statedump(usess, app); - if (ret < 0) { - /* Continue to the next app even on error */ - continue; + ret = ust_app_regenerate_statedump(usess, app); + if (ret < 0) { + /* Continue to the next app even on error */ + continue; + } + } + + rcu_read_unlock(); + + return 0; +} + +/* + * Rotate all the channels of a session. + * + * Return LTTNG_OK on success or else an LTTng error code. + */ +enum lttng_error_code ust_app_rotate_session(struct ltt_session *session) +{ + int ret; + enum lttng_error_code cmd_ret = LTTNG_OK; + struct lttng_ht_iter iter; + struct ust_app *app; + struct ltt_ust_session *usess = session->ust_session; + + assert(usess); + + rcu_read_lock(); + + switch (usess->buffer_type) { + case LTTNG_BUFFER_PER_UID: + { + struct buffer_reg_uid *reg; + + cds_list_for_each_entry(reg, &usess->buffer_reg_uid_list, lnode) { + struct buffer_reg_channel *buf_reg_chan; + struct consumer_socket *socket; + + if (!reg->registry->reg.ust->metadata_key) { + /* Skip since no metadata is present */ + continue; + } + + /* Get consumer socket to use to push the metadata.*/ + socket = consumer_find_socket_by_bitness(reg->bits_per_long, + usess->consumer); + if (!socket) { + cmd_ret = LTTNG_ERR_INVALID; + goto error; + } + + /* Rotate the data channels. */ + cds_lfht_for_each_entry(reg->registry->channels->ht, &iter.iter, + buf_reg_chan, node.node) { + ret = consumer_rotate_channel(socket, + buf_reg_chan->consumer_key, + usess->uid, usess->gid, + usess->consumer, + /* is_metadata_channel */ false); + if (ret < 0) { + cmd_ret = LTTNG_ERR_ROTATION_FAIL_CONSUMER; + goto error; + } + } + + (void) push_metadata(reg->registry->reg.ust, usess->consumer); + + ret = consumer_rotate_channel(socket, + reg->registry->reg.ust->metadata_key, + usess->uid, usess->gid, + usess->consumer, + /* is_metadata_channel */ true); + if (ret < 0) { + cmd_ret = LTTNG_ERR_ROTATION_FAIL_CONSUMER; + goto error; + } + } + break; + } + case LTTNG_BUFFER_PER_PID: + { + cds_lfht_for_each_entry(ust_app_ht->ht, &iter.iter, app, pid_n.node) { + struct consumer_socket *socket; + struct lttng_ht_iter chan_iter; + struct ust_app_channel *ua_chan; + struct ust_app_session *ua_sess; + struct ust_registry_session *registry; + + ua_sess = lookup_session_by_app(usess, app); + if (!ua_sess) { + /* Session not associated with this app. */ + continue; + } + + /* Get the right consumer socket for the application. */ + socket = consumer_find_socket_by_bitness(app->bits_per_long, + usess->consumer); + if (!socket) { + cmd_ret = LTTNG_ERR_INVALID; + goto error; + } + + registry = get_session_registry(ua_sess); + if (!registry) { + DBG("Application session is being torn down. Skip application."); + continue; + } + + /* Rotate the data channels. */ + cds_lfht_for_each_entry(ua_sess->channels->ht, &chan_iter.iter, + ua_chan, node.node) { + ret = consumer_rotate_channel(socket, + ua_chan->key, + lttng_credentials_get_uid(&ua_sess->effective_credentials), + lttng_credentials_get_gid(&ua_sess->effective_credentials), + ua_sess->consumer, + /* is_metadata_channel */ false); + if (ret < 0) { + /* Per-PID buffer and application going away. */ + if (ret == -LTTNG_ERR_CHAN_NOT_FOUND) + continue; + cmd_ret = LTTNG_ERR_ROTATION_FAIL_CONSUMER; + goto error; + } + } + + /* Rotate the metadata channel. */ + (void) push_metadata(registry, usess->consumer); + ret = consumer_rotate_channel(socket, + registry->metadata_key, + lttng_credentials_get_uid(&ua_sess->effective_credentials), + lttng_credentials_get_gid(&ua_sess->effective_credentials), + ua_sess->consumer, + /* is_metadata_channel */ true); + if (ret < 0) { + /* Per-PID buffer and application going away. */ + if (ret == -LTTNG_ERR_CHAN_NOT_FOUND) + continue; + cmd_ret = LTTNG_ERR_ROTATION_FAIL_CONSUMER; + goto error; + } + } + break; + } + default: + assert(0); + break; + } + + cmd_ret = LTTNG_OK; + +error: + rcu_read_unlock(); + return cmd_ret; +} + +enum lttng_error_code ust_app_create_channel_subdirectories( + const struct ltt_ust_session *usess) +{ + enum lttng_error_code ret = LTTNG_OK; + struct lttng_ht_iter iter; + enum lttng_trace_chunk_status chunk_status; + char *pathname_index; + int fmt_ret; + + assert(usess->current_trace_chunk); + rcu_read_lock(); + + switch (usess->buffer_type) { + case LTTNG_BUFFER_PER_UID: + { + struct buffer_reg_uid *reg; + + cds_list_for_each_entry(reg, &usess->buffer_reg_uid_list, lnode) { + fmt_ret = asprintf(&pathname_index, + DEFAULT_UST_TRACE_DIR "/" DEFAULT_UST_TRACE_UID_PATH "/" DEFAULT_INDEX_DIR, + reg->uid, reg->bits_per_long); + if (fmt_ret < 0) { + ERR("Failed to format channel index directory"); + ret = LTTNG_ERR_CREATE_DIR_FAIL; + goto error; + } + + /* + * Create the index subdirectory which will take care + * of implicitly creating the channel's path. + */ + chunk_status = lttng_trace_chunk_create_subdirectory( + usess->current_trace_chunk, + pathname_index); + free(pathname_index); + if (chunk_status != LTTNG_TRACE_CHUNK_STATUS_OK) { + ret = LTTNG_ERR_CREATE_DIR_FAIL; + goto error; + } + } + break; + } + case LTTNG_BUFFER_PER_PID: + { + struct ust_app *app; + + /* + * Create the toplevel ust/ directory in case no apps are running. + */ + chunk_status = lttng_trace_chunk_create_subdirectory( + usess->current_trace_chunk, + DEFAULT_UST_TRACE_DIR); + if (chunk_status != LTTNG_TRACE_CHUNK_STATUS_OK) { + ret = LTTNG_ERR_CREATE_DIR_FAIL; + goto error; + } + + cds_lfht_for_each_entry(ust_app_ht->ht, &iter.iter, app, + pid_n.node) { + struct ust_app_session *ua_sess; + struct ust_registry_session *registry; + + ua_sess = lookup_session_by_app(usess, app); + if (!ua_sess) { + /* Session not associated with this app. */ + continue; + } + + registry = get_session_registry(ua_sess); + if (!registry) { + DBG("Application session is being torn down. Skip application."); + continue; + } + + fmt_ret = asprintf(&pathname_index, + DEFAULT_UST_TRACE_DIR "/%s/" DEFAULT_INDEX_DIR, + ua_sess->path); + if (fmt_ret < 0) { + ERR("Failed to format channel index directory"); + ret = LTTNG_ERR_CREATE_DIR_FAIL; + goto error; + } + /* + * Create the index subdirectory which will take care + * of implicitly creating the channel's path. + */ + chunk_status = lttng_trace_chunk_create_subdirectory( + usess->current_trace_chunk, + pathname_index); + free(pathname_index); + if (chunk_status != LTTNG_TRACE_CHUNK_STATUS_OK) { + ret = LTTNG_ERR_CREATE_DIR_FAIL; + goto error; + } } + break; + } + default: + abort(); } + ret = LTTNG_OK; +error: rcu_read_unlock(); - - return 0; + return ret; } /* - * Rotate all the channels of a session. + * Clear all the channels of a session. * * Return LTTNG_OK on success or else an LTTng error code. */ -enum lttng_error_code ust_app_rotate_session(struct ltt_session *session) +static +enum lttng_error_code ust_app_clear_session_channels(struct ltt_session *session) { int ret; enum lttng_error_code cmd_ret = LTTNG_OK; @@ -6316,13 +9938,19 @@ enum lttng_error_code ust_app_rotate_session(struct ltt_session *session) rcu_read_lock(); + if (usess->active) { + ERR("Expecting inactive session %s (%" PRIu64 ")", session->name, session->id); + cmd_ret = LTTNG_ERR_FATAL; + goto end; + } + switch (usess->buffer_type) { case LTTNG_BUFFER_PER_UID: { struct buffer_reg_uid *reg; cds_list_for_each_entry(reg, &usess->buffer_reg_uid_list, lnode) { - struct buffer_reg_channel *reg_chan; + struct buffer_reg_channel *buf_reg_chan; struct consumer_socket *socket; /* Get consumer socket to use to push the metadata.*/ @@ -6330,32 +9958,29 @@ enum lttng_error_code ust_app_rotate_session(struct ltt_session *session) usess->consumer); if (!socket) { cmd_ret = LTTNG_ERR_INVALID; - goto error; + goto error_socket; } - /* Rotate the data channels. */ + /* Clear the data channels. */ cds_lfht_for_each_entry(reg->registry->channels->ht, &iter.iter, - reg_chan, node.node) { - ret = consumer_rotate_channel(socket, - reg_chan->consumer_key, - usess->uid, usess->gid, - usess->consumer, - /* is_metadata_channel */ false); + buf_reg_chan, node.node) { + ret = consumer_clear_channel(socket, + buf_reg_chan->consumer_key); if (ret < 0) { - cmd_ret = LTTNG_ERR_ROTATION_FAIL_CONSUMER; goto error; } } (void) push_metadata(reg->registry->reg.ust, usess->consumer); - ret = consumer_rotate_channel(socket, - reg->registry->reg.ust->metadata_key, - usess->uid, usess->gid, - usess->consumer, - /* is_metadata_channel */ true); + /* + * Clear the metadata channel. + * Metadata channel is not cleared per se but we still need to + * perform a rotation operation on it behind the scene. + */ + ret = consumer_clear_channel(socket, + reg->registry->reg.ust->metadata_key); if (ret < 0) { - cmd_ret = LTTNG_ERR_ROTATION_FAIL_CONSUMER; goto error; } } @@ -6381,7 +10006,7 @@ enum lttng_error_code ust_app_rotate_session(struct ltt_session *session) usess->consumer); if (!socket) { cmd_ret = LTTNG_ERR_INVALID; - goto error; + goto error_socket; } registry = get_session_registry(ua_sess); @@ -6390,39 +10015,32 @@ enum lttng_error_code ust_app_rotate_session(struct ltt_session *session) continue; } - /* Rotate the data channels. */ + /* Clear the data channels. */ cds_lfht_for_each_entry(ua_sess->channels->ht, &chan_iter.iter, ua_chan, node.node) { - ret = consumer_rotate_channel(socket, - ua_chan->key, - ua_sess->effective_credentials - .uid, - ua_sess->effective_credentials - .gid, - ua_sess->consumer, - /* is_metadata_channel */ false); + ret = consumer_clear_channel(socket, ua_chan->key); if (ret < 0) { /* Per-PID buffer and application going away. */ - if (ret == -LTTNG_ERR_CHAN_NOT_FOUND) + if (ret == -LTTNG_ERR_CHAN_NOT_FOUND) { continue; - cmd_ret = LTTNG_ERR_ROTATION_FAIL_CONSUMER; + } goto error; } } - /* Rotate the metadata channel. */ (void) push_metadata(registry, usess->consumer); - ret = consumer_rotate_channel(socket, - registry->metadata_key, - ua_sess->effective_credentials.uid, - ua_sess->effective_credentials.gid, - ua_sess->consumer, - /* is_metadata_channel */ true); + + /* + * Clear the metadata channel. + * Metadata channel is not cleared per se but we still need to + * perform rotation operation on it behind the scene. + */ + ret = consumer_clear_channel(socket, registry->metadata_key); if (ret < 0) { /* Per-PID buffer and application going away. */ - if (ret == -LTTNG_ERR_CHAN_NOT_FOUND) + if (ret == -LTTNG_ERR_CHAN_NOT_FOUND) { continue; - cmd_ret = LTTNG_ERR_ROTATION_FAIL_CONSUMER; + } goto error; } } @@ -6434,22 +10052,269 @@ enum lttng_error_code ust_app_rotate_session(struct ltt_session *session) } cmd_ret = LTTNG_OK; + goto end; error: + switch (-ret) { + case LTTCOMM_CONSUMERD_RELAYD_CLEAR_DISALLOWED: + cmd_ret = LTTNG_ERR_CLEAR_RELAY_DISALLOWED; + break; + default: + cmd_ret = LTTNG_ERR_CLEAR_FAIL_CONSUMER; + } + +error_socket: +end: rcu_read_unlock(); return cmd_ret; } -enum lttng_error_code ust_app_create_channel_subdirectories( - const struct ltt_ust_session *usess) +static +enum lttng_error_code ust_app_clear_session_maps_per_uid( + struct ltt_ust_session *usess, struct ltt_ust_map *umap, + uint32_t app_bitness) +{ + struct lttng_ht_iter iter; + struct buffer_reg_uid *buf_reg_uid; + struct buffer_reg_map *buf_reg_map; + struct ust_registry_session *ust_reg_sess; + struct lttng_ht_node_u64 *ust_reg_map_node; + struct ust_registry_map *ust_reg_map; + struct ust_registry_map_index_ht_entry *map_index_entry; + enum lttng_error_code status; + + buf_reg_uid = buffer_reg_uid_find(usess->id, app_bitness, usess->uid); + if (!buf_reg_uid) { + /* + * Buffer registry entry for uid not found. Probably no app for + * this UID at the moment. + */ + DBG("No buffer registry entry found for uid: ust-sess-id = %"PRIu64", bitness = %"PRIu32", uid = %d", + usess->id, app_bitness, usess->uid); + /* + * Not an error. Leave the key value pair unchanged and return. + */ + status = LTTNG_OK; + goto end; + } + + buf_reg_map = buffer_reg_map_find(umap->id, buf_reg_uid); + if (!buf_reg_uid) { + ERR("Error getting per-uid map buffer registry entry: map-id = %"PRIu64, + umap->id); + status = LTTNG_ERR_UNK; + goto end; + } + + ust_reg_sess = buf_reg_uid->registry->reg.ust; + + /* Get the ust_reg map object from the registry */ + // FIXME: frdeso: This can be changed to ust_registry_map_find() right? + + lttng_ht_lookup(ust_reg_sess->maps, (void *) &umap->id, &iter); + ust_reg_map_node = lttng_ht_iter_get_node_u64(&iter); + if (!ust_reg_map_node) { + ERR("Error getting per-uid map buffer registry entry: map-id = %"PRIu64, + umap->id); + status = LTTNG_ERR_UNK; + goto end; + } + ust_reg_map = caa_container_of(ust_reg_map_node, + struct ust_registry_map, node); + + cds_lfht_for_each_entry(ust_reg_map->key_string_to_bucket_index_ht->ht, + &iter.iter, map_index_entry, node.node) { + int ret; + size_t dimension_indexes[1] = {map_index_entry->index}; + + ret = ustctl_counter_clear(buf_reg_map->daemon_counter, dimension_indexes); + if (ret) { + ERR("clearing counter index %"PRIu64, map_index_entry->index); + //fixme: frdeso: convert ust errors to tools errors + status = LTTNG_ERR_UNK; + goto end; + } + } + + status = LTTNG_OK; + +end: + return status; +} + +static +enum lttng_error_code ust_app_clear_session_maps_per_pid( + struct ltt_ust_session *usess, struct ltt_ust_map *umap, + uint32_t app_bitness) +{ + struct lttng_ht_iter app_iter; + enum lttng_error_code status; + struct ust_app *app; + struct map_kv_ht_entry *kv_entry; + struct lttng_ht_iter iter; + + cds_lfht_for_each_entry(ust_app_ht->ht, &app_iter.iter, app, pid_n.node) { + struct lttng_ht_iter map_iter, key_iter; + struct lttng_ht_node_str *ua_map_node; + struct ust_app_map *ua_map; + struct ust_app_session *ua_sess; + struct ust_registry_session *ust_reg_sess; + struct ust_registry_map *ust_reg_map; + struct ust_registry_map_index_ht_entry *map_index_entry; + + if (app->bits_per_long != app_bitness) { + continue; + } + + ua_sess = lookup_session_by_app(usess, app); + if (!ua_sess) { + /* Session not associated with this app. */ + continue; + } + + ust_reg_sess = get_session_registry(ua_sess); + if (!ust_reg_sess) { + DBG("Application session is being torn down. Skip application."); + continue; + } + + /* Lookup map in the ust app session */ + lttng_ht_lookup(ua_sess->maps, (void *)umap->name, &map_iter); + ua_map_node = lttng_ht_iter_get_node_str(&map_iter); + + assert(ua_map_node != NULL); + ua_map = caa_container_of(ua_map_node, struct ust_app_map, node); + + pthread_mutex_lock(&ust_reg_sess->lock); + ust_reg_map = ust_registry_map_find(ust_reg_sess, ua_map->key); + pthread_mutex_unlock(&ust_reg_sess->lock); + assert(ust_reg_map); + + /* Iterate over all the formated_key -> counter index */ + cds_lfht_for_each_entry(ust_reg_map->key_string_to_bucket_index_ht->ht, + &key_iter.iter, map_index_entry, node.node) { + + int ret; + size_t dimension_indexes[1] = {map_index_entry->index}; + + ret = ustctl_counter_clear(ua_map->map_handle, + dimension_indexes); + if (ret) { + ERR("clearing counter index %"PRIu64, map_index_entry->index); + //fixme: frdeso: convert ust errors to tools errors + status = LTTNG_ERR_UNK; + goto end; + } + } + } + + /* + * Emptying the dead app key values. + */ + pthread_mutex_lock(&umap->dead_app_kv_values.lock); + + if (app_bitness == 32) { + cds_lfht_for_each_entry(umap->dead_app_kv_values.dead_app_kv_values_32bits->ht, + &iter.iter, kv_entry, node.node) { + kv_entry->value = 0; + } + } else { + + cds_lfht_for_each_entry(umap->dead_app_kv_values.dead_app_kv_values_64bits->ht, + &iter.iter, kv_entry, node.node) { + kv_entry->value = 0; + } + } + + pthread_mutex_unlock(&umap->dead_app_kv_values.lock); + + status = LTTNG_OK; +end: + return status; +} + +static +enum lttng_error_code ust_app_clear_session_maps(struct ltt_session *session) +{ + struct ltt_ust_session *usess = session->ust_session; + enum lttng_error_code status; + struct lttng_ht_iter iter; + struct ltt_ust_map *umap; + + cds_lfht_for_each_entry(usess->domain_global.maps->ht, &iter.iter, + umap, node.node) { + + if (usess->buffer_type == LTTNG_BUFFER_PER_UID) { + status = ust_app_clear_session_maps_per_uid(session->ust_session, + umap, 32); + assert(status == LTTNG_OK); + + status = ust_app_clear_session_maps_per_uid(session->ust_session, + umap, 64); + assert(status == LTTNG_OK); + //fixme:frdeso:error handling + } else { + status = ust_app_clear_session_maps_per_pid(session->ust_session, + umap, 32); + assert(status == LTTNG_OK); + + status = ust_app_clear_session_maps_per_pid(session->ust_session, + umap, 64); + assert(status == LTTNG_OK); + //fixme:frdeso:error handling + // + // + } + + } + + return LTTNG_OK; +} + +enum lttng_error_code ust_app_clear_session(struct ltt_session *session) +{ + enum lttng_error_code cmd_ret; + + + cmd_ret = ust_app_clear_session_channels(session); + if (cmd_ret != LTTNG_OK) { + ERR("Clearing session's channels"); + goto end; + } + + cmd_ret = ust_app_clear_session_maps(session); + if (cmd_ret != LTTNG_OK) { + ERR("Clearing session's maps"); + goto end; + } +end: + return cmd_ret; +} + +/* + * This function skips the metadata channel as the begin/end timestamps of a + * metadata packet are useless. + * + * Moreover, opening a packet after a "clear" will cause problems for live + * sessions as it will introduce padding that was not part of the first trace + * chunk. The relay daemon expects the content of the metadata stream of + * successive metadata trace chunks to be strict supersets of one another. + * + * For example, flushing a packet at the beginning of the metadata stream of + * a trace chunk resulting from a "clear" session command will cause the + * size of the metadata stream of the new trace chunk to not match the size of + * the metadata stream of the original chunk. This will confuse the relay + * daemon as the same "offset" in a metadata stream will no longer point + * to the same content. + */ +enum lttng_error_code ust_app_open_packets(struct ltt_session *session) { enum lttng_error_code ret = LTTNG_OK; struct lttng_ht_iter iter; - enum lttng_trace_chunk_status chunk_status; - char *pathname_index; - int fmt_ret; + struct ltt_ust_session *usess = session->ust_session; + + assert(usess); - assert(usess->current_trace_chunk); rcu_read_lock(); switch (usess->buffer_type) { @@ -6457,27 +10322,29 @@ enum lttng_error_code ust_app_create_channel_subdirectories( { struct buffer_reg_uid *reg; - cds_list_for_each_entry(reg, &usess->buffer_reg_uid_list, lnode) { - fmt_ret = asprintf(&pathname_index, - DEFAULT_UST_TRACE_DIR DEFAULT_UST_TRACE_UID_PATH "/" DEFAULT_INDEX_DIR, - reg->uid, reg->bits_per_long); - if (fmt_ret < 0) { - ERR("Failed to format channel index directory"); - ret = LTTNG_ERR_CREATE_DIR_FAIL; + cds_list_for_each_entry ( + reg, &usess->buffer_reg_uid_list, lnode) { + struct buffer_reg_channel *buf_reg_chan; + struct consumer_socket *socket; + + socket = consumer_find_socket_by_bitness( + reg->bits_per_long, usess->consumer); + if (!socket) { + ret = LTTNG_ERR_FATAL; goto error; } - /* - * Create the index subdirectory which will take care - * of implicitly creating the channel's path. - */ - chunk_status = lttng_trace_chunk_create_subdirectory( - usess->current_trace_chunk, - pathname_index); - free(pathname_index); - if (chunk_status != LTTNG_TRACE_CHUNK_STATUS_OK) { - ret = LTTNG_ERR_CREATE_DIR_FAIL; - goto error; + cds_lfht_for_each_entry(reg->registry->channels->ht, + &iter.iter, buf_reg_chan, node.node) { + const int open_ret = + consumer_open_channel_packets( + socket, + buf_reg_chan->consumer_key); + + if (open_ret < 0) { + ret = LTTNG_ERR_UNK; + goto error; + } } } break; @@ -6486,8 +10353,11 @@ enum lttng_error_code ust_app_create_channel_subdirectories( { struct ust_app *app; - cds_lfht_for_each_entry(ust_app_ht->ht, &iter.iter, app, - pid_n.node) { + cds_lfht_for_each_entry ( + ust_app_ht->ht, &iter.iter, app, pid_n.node) { + struct consumer_socket *socket; + struct lttng_ht_iter chan_iter; + struct ust_app_channel *ua_chan; struct ust_app_session *ua_sess; struct ust_registry_session *registry; @@ -6497,40 +10367,48 @@ enum lttng_error_code ust_app_create_channel_subdirectories( continue; } + /* Get the right consumer socket for the application. */ + socket = consumer_find_socket_by_bitness( + app->bits_per_long, usess->consumer); + if (!socket) { + ret = LTTNG_ERR_FATAL; + goto error; + } + registry = get_session_registry(ua_sess); if (!registry) { DBG("Application session is being torn down. Skip application."); continue; } - fmt_ret = asprintf(&pathname_index, - DEFAULT_UST_TRACE_DIR "%s/" DEFAULT_INDEX_DIR, - ua_sess->path); - if (fmt_ret < 0) { - ERR("Failed to format channel index directory"); - ret = LTTNG_ERR_CREATE_DIR_FAIL; - goto error; - } - /* - * Create the index subdirectory which will take care - * of implicitly creating the channel's path. - */ - chunk_status = lttng_trace_chunk_create_subdirectory( - usess->current_trace_chunk, - pathname_index); - free(pathname_index); - if (chunk_status != LTTNG_TRACE_CHUNK_STATUS_OK) { - ret = LTTNG_ERR_CREATE_DIR_FAIL; - goto error; + cds_lfht_for_each_entry(ua_sess->channels->ht, + &chan_iter.iter, ua_chan, node.node) { + const int open_ret = + consumer_open_channel_packets( + socket, + ua_chan->key); + + if (open_ret < 0) { + /* + * Per-PID buffer and application going + * away. + */ + if (open_ret == -LTTNG_ERR_CHAN_NOT_FOUND) { + continue; + } + + ret = LTTNG_ERR_UNK; + goto error; + } } } break; } default: abort(); + break; } - ret = LTTNG_OK; error: rcu_read_unlock(); return ret;