X-Git-Url: http://git.efficios.com/?a=blobdiff_plain;f=src%2Fbin%2Flttng-sessiond%2Fust-app.c;h=dcb6e7c00ff5129cb9ec8930a46fac39d65bcc9b;hb=c4b884061bd2f095631a115c70d33d13fb1e6fb6;hp=e029b42d8f93827ac70da99203608abd80214e19;hpb=5368d366c6c7f80ad941172a6bc9c6a5174cbdf0;p=lttng-tools.git diff --git a/src/bin/lttng-sessiond/ust-app.c b/src/bin/lttng-sessiond/ust-app.c index e029b42d8..dcb6e7c00 100644 --- a/src/bin/lttng-sessiond/ust-app.c +++ b/src/bin/lttng-sessiond/ust-app.c @@ -16,6 +16,7 @@ */ #define _GNU_SOURCE +#define _LGPL_SOURCE #include #include #include @@ -40,6 +41,9 @@ #include "ust-ctl.h" #include "utils.h" +static +int ust_app_flush_app_session(struct ust_app *app, struct ust_app_session *ua_sess); + /* Next available channel key. Access under next_channel_key_lock. */ static uint64_t _next_channel_key; static pthread_mutex_t next_channel_key_lock = PTHREAD_MUTEX_INITIALIZER; @@ -1446,11 +1450,15 @@ int create_ust_event(struct ust_app *app, struct ust_app_session *ua_sess, } /* If event not enabled, disable it on the tracer */ - if (ua_event->enabled == 0) { - ret = disable_ust_event(app, ua_sess, ua_event); + if (ua_event->enabled) { + /* + * We now need to explicitly enable the event, since it + * is now disabled at creation. + */ + ret = enable_ust_event(app, ua_sess, ua_event); if (ret < 0) { /* - * If we hit an EPERM, something is wrong with our disable call. If + * 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. */ @@ -1552,7 +1560,7 @@ static void shadow_copy_channel(struct ust_app_channel *ua_chan, } lttng_ht_node_init_ulong(&ua_ctx->node, (unsigned long) ua_ctx->ctx.ctx); - lttng_ht_add_unique_ulong(ua_chan->ctx, &ua_ctx->node); + lttng_ht_add_ulong(ua_chan->ctx, &ua_ctx->node); cds_list_add_tail(&ua_ctx->list, &ua_chan->ctx_list); } @@ -1727,7 +1735,6 @@ static int setup_buffer_reg_pid(struct ust_app_session *ua_sess, if (ret < 0) { goto error; } - buffer_reg_pid_add(reg_pid); } else { goto end; } @@ -1740,9 +1747,18 @@ static int setup_buffer_reg_pid(struct ust_app_session *ua_sess, app->byte_order, app->version.major, app->version.minor); if (ret < 0) { + /* + * reg_pid->registry->reg.ust is NULL upon error, so we need to + * destroy the buffer registry, because it is always expected + * that if the buffer registry can be found, its ust registry is + * non-NULL. + */ + buffer_reg_pid_destroy(reg_pid); goto error; } + buffer_reg_pid_add(reg_pid); + DBG3("UST app buffer registry per PID created successfully"); end: @@ -1783,7 +1799,6 @@ static int setup_buffer_reg_uid(struct ltt_ust_session *usess, if (ret < 0) { goto error; } - buffer_reg_uid_add(reg_uid); } else { goto end; } @@ -1796,13 +1811,21 @@ static int setup_buffer_reg_uid(struct ltt_ust_session *usess, app->byte_order, app->version.major, app->version.minor); if (ret < 0) { + /* + * reg_uid->registry->reg.ust is NULL upon error, so we need to + * destroy the buffer registry, because it is always expected + * that if the buffer registry can be found, its ust registry is + * non-NULL. + */ + buffer_reg_uid_destroy(reg_uid, NULL); goto error; } /* Add node to teardown list of the session. */ cds_list_add(®_uid->lnode, &usess->buffer_reg_uid_list); - DBG3("UST app buffer registry per UID created successfully"); + buffer_reg_uid_add(reg_uid); + DBG3("UST app buffer registry per UID created successfully"); end: if (regp) { *regp = reg_uid; @@ -1855,6 +1878,7 @@ static int create_ust_app_session(struct ltt_ust_session *usess, /* Init local registry. */ ret = setup_buffer_reg_pid(ua_sess, app, NULL); if (ret < 0) { + delete_ust_app_session(-1, ua_sess, app); goto error; } break; @@ -1862,6 +1886,7 @@ static int create_ust_app_session(struct ltt_ust_session *usess, /* Look for a global registry. If none exists, create one. */ ret = setup_buffer_reg_uid(usess, app, NULL); if (ret < 0) { + delete_ust_app_session(-1, ua_sess, app); goto error; } break; @@ -1923,6 +1948,75 @@ error: return ret; } +/* + * Match function for a hash table lookup of ust_app_ctx. + * + * It matches an ust app context based on the context type and, in the case + * of perf counters, their name. + */ +static int ht_match_ust_app_ctx(struct cds_lfht_node *node, const void *_key) +{ + struct ust_app_ctx *ctx; + const struct lttng_ust_context *key; + + assert(node); + assert(_key); + + ctx = caa_container_of(node, struct ust_app_ctx, node.node); + key = _key; + + /* Context type */ + if (ctx->ctx.ctx != key->ctx) { + goto no_match; + } + + /* Check the name in the case of perf thread counters. */ + if (key->ctx == LTTNG_UST_CONTEXT_PERF_THREAD_COUNTER) { + if (strncmp(key->u.perf_counter.name, + ctx->ctx.u.perf_counter.name, + sizeof(key->u.perf_counter.name))) { + goto no_match; + } + } + + /* Match. */ + return 1; + +no_match: + return 0; +} + +/* + * Lookup for an ust app context from an lttng_ust_context. + * + * Must be called while holding RCU read side lock. + * Return an ust_app_ctx object or NULL on error. + */ +static +struct ust_app_ctx *find_ust_app_context(struct lttng_ht *ht, + struct lttng_ust_context *uctx) +{ + struct lttng_ht_iter iter; + struct lttng_ht_node_ulong *node; + struct ust_app_ctx *app_ctx = NULL; + + assert(uctx); + assert(ht); + + /* Lookup using the lttng_ust_context_type and a custom match fct. */ + cds_lfht_lookup(ht->ht, ht->hash_fct((void *) uctx->ctx, lttng_ht_seed), + ht_match_ust_app_ctx, uctx, &iter.iter); + node = lttng_ht_iter_get_node_ulong(&iter); + if (!node) { + goto end; + } + + app_ctx = caa_container_of(node, struct ust_app_ctx, node); + +end: + return app_ctx; +} + /* * Create a context for the channel on the tracer. * @@ -1934,15 +2028,12 @@ int create_ust_app_channel_context(struct ust_app_session *ua_sess, struct ust_app *app) { int ret = 0; - struct lttng_ht_iter iter; - struct lttng_ht_node_ulong *node; struct ust_app_ctx *ua_ctx; DBG2("UST app adding context to channel %s", ua_chan->name); - lttng_ht_lookup(ua_chan->ctx, (void *)((unsigned long)uctx->ctx), &iter); - node = lttng_ht_iter_get_node_ulong(&iter); - if (node != NULL) { + ua_ctx = find_ust_app_context(ua_chan->ctx, uctx); + if (ua_ctx) { ret = -EEXIST; goto error; } @@ -1955,7 +2046,7 @@ int create_ust_app_channel_context(struct ust_app_session *ua_sess, } lttng_ht_node_init_ulong(&ua_ctx->node, (unsigned long) ua_ctx->ctx.ctx); - lttng_ht_add_unique_ulong(ua_chan->ctx, &ua_ctx->node); + lttng_ht_add_ulong(ua_chan->ctx, &ua_ctx->node); cds_list_add_tail(&ua_ctx->list, &ua_chan->ctx_list); ret = create_ust_channel_context(ua_chan, ua_ctx, app); @@ -2450,6 +2541,8 @@ static int create_channel_per_uid(struct ust_app *app, /* Create the buffer registry channel object. */ ret = create_buffer_reg_channel(reg_uid->registry, ua_chan, ®_chan); if (ret < 0) { + ERR("Error creating the UST channel \"%s\" registry instance", + ua_chan->name); goto error; } assert(reg_chan); @@ -2461,6 +2554,9 @@ static int create_channel_per_uid(struct ust_app *app, ret = do_consumer_create_channel(usess, ua_sess, ua_chan, app->bits_per_long, reg_uid->registry->reg.ust); if (ret < 0) { + ERR("Error creating UST channel \"%s\" on the consumer daemon", + ua_chan->name); + /* * Let's remove the previously created buffer registry channel so * it's not visible anymore in the session registry. @@ -2477,6 +2573,8 @@ static int create_channel_per_uid(struct ust_app *app, */ ret = setup_buffer_reg_channel(reg_uid->registry, ua_chan, reg_chan); if (ret < 0) { + ERR("Error setting up UST channel \"%s\"", + ua_chan->name); goto error; } @@ -2485,6 +2583,10 @@ static int create_channel_per_uid(struct ust_app *app, /* Send buffers to the application. */ ret = send_channel_uid_to_ust(reg_chan, app, ua_sess, ua_chan); if (ret < 0) { + /* + * Don't report error to the console, since it may be + * caused by application concurrently exiting. + */ goto error; } @@ -2519,6 +2621,8 @@ static int create_channel_per_pid(struct ust_app *app, /* Create and add a new channel registry to session. */ ret = ust_registry_channel_add(registry, ua_chan->key); if (ret < 0) { + ERR("Error creating the UST channel \"%s\" registry instance", + ua_chan->name); goto error; } @@ -2526,11 +2630,17 @@ static int create_channel_per_pid(struct ust_app *app, ret = do_consumer_create_channel(usess, ua_sess, ua_chan, app->bits_per_long, registry); if (ret < 0) { + ERR("Error creating UST channel \"%s\" on the consumer daemon", + ua_chan->name); goto error; } ret = send_channel_pid_to_ust(app, ua_sess, ua_chan); if (ret < 0) { + /* + * Don't report error to the console, since it may be + * caused by application concurrently exiting. + */ goto error; } @@ -2972,6 +3082,7 @@ void ust_app_unregister(int sock) { struct ust_app *lta; struct lttng_ht_node_ulong *node; + struct lttng_ht_iter ust_app_sock_iter; struct lttng_ht_iter iter; struct ust_app_session *ua_sess; int ret; @@ -2979,39 +3090,19 @@ void ust_app_unregister(int sock) rcu_read_lock(); /* Get the node reference for a call_rcu */ - lttng_ht_lookup(ust_app_ht_by_sock, (void *)((unsigned long) sock), &iter); - node = lttng_ht_iter_get_node_ulong(&iter); + lttng_ht_lookup(ust_app_ht_by_sock, (void *)((unsigned long) sock), &ust_app_sock_iter); + node = lttng_ht_iter_get_node_ulong(&ust_app_sock_iter); assert(node); lta = caa_container_of(node, struct ust_app, sock_n); DBG("PID %d unregistering with sock %d", lta->pid, sock); - /* Remove application from PID hash table */ - ret = lttng_ht_del(ust_app_ht_by_sock, &iter); - assert(!ret); - /* - * Remove application from notify hash table. The thread handling the - * notify socket could have deleted the node so ignore on error because - * either way it's valid. The close of that socket is handled by the other - * thread. + * Perform "push metadata" and flush all application streams + * before removing app from hash tables, ensuring proper + * behavior of data_pending check. + * Remove sessions so they are not visible during deletion. */ - iter.iter.node = <a->notify_sock_n.node; - (void) lttng_ht_del(ust_app_ht_by_notify_sock, &iter); - - /* - * Ignore return value since the node might have been removed before by an - * add replace during app registration because the PID can be reassigned by - * the OS. - */ - iter.iter.node = <a->pid_n.node; - ret = lttng_ht_del(ust_app_ht, &iter); - if (ret) { - DBG3("Unregister app by PID %d failed. This can happen on pid reuse", - lta->pid); - } - - /* Remove sessions so they are not visible during deletion.*/ cds_lfht_for_each_entry(lta->sessions->ht, &iter.iter, ua_sess, node.node) { struct ust_registry_session *registry; @@ -3022,6 +3113,8 @@ void ust_app_unregister(int sock) continue; } + (void) ust_app_flush_app_session(lta, ua_sess); + /* * Add session to list for teardown. This is safe since at this point we * are the only one using this list. @@ -3056,11 +3149,36 @@ void ust_app_unregister(int sock) (void) close_metadata(registry, ua_sess->consumer); } } - cds_list_add(&ua_sess->teardown_node, <a->teardown_head); + pthread_mutex_unlock(&ua_sess->lock); } + /* Remove application from PID hash table */ + ret = lttng_ht_del(ust_app_ht_by_sock, &ust_app_sock_iter); + assert(!ret); + + /* + * Remove application from notify hash table. The thread handling the + * notify socket could have deleted the node so ignore on error because + * either way it's valid. The close of that socket is handled by the other + * thread. + */ + iter.iter.node = <a->notify_sock_n.node; + (void) lttng_ht_del(ust_app_ht_by_notify_sock, &iter); + + /* + * Ignore return value since the node might have been removed before by an + * add replace during app registration because the PID can be reassigned by + * the OS. + */ + iter.iter.node = <a->pid_n.node; + ret = lttng_ht_del(ust_app_ht, &iter); + if (ret) { + DBG3("Unregister app by PID %d failed. This can happen on pid reuse", + lta->pid); + } + /* Free memory */ call_rcu(<a->pid_n.head, delete_ust_app_rcu); @@ -3301,41 +3419,63 @@ void ust_app_clean_list(void) rcu_read_lock(); - cds_lfht_for_each_entry(ust_app_ht->ht, &iter.iter, app, pid_n.node) { - ret = lttng_ht_del(ust_app_ht, &iter); - assert(!ret); - call_rcu(&app->pid_n.head, delete_ust_app_rcu); + if (ust_app_ht) { + cds_lfht_for_each_entry(ust_app_ht->ht, &iter.iter, app, pid_n.node) { + ret = lttng_ht_del(ust_app_ht, &iter); + assert(!ret); + call_rcu(&app->pid_n.head, delete_ust_app_rcu); + } } /* Cleanup socket hash table */ - cds_lfht_for_each_entry(ust_app_ht_by_sock->ht, &iter.iter, app, - sock_n.node) { - ret = lttng_ht_del(ust_app_ht_by_sock, &iter); - assert(!ret); + if (ust_app_ht_by_sock) { + cds_lfht_for_each_entry(ust_app_ht_by_sock->ht, &iter.iter, app, + sock_n.node) { + ret = lttng_ht_del(ust_app_ht_by_sock, &iter); + assert(!ret); + } } /* Cleanup notify socket hash table */ - cds_lfht_for_each_entry(ust_app_ht_by_notify_sock->ht, &iter.iter, app, - notify_sock_n.node) { - ret = lttng_ht_del(ust_app_ht_by_notify_sock, &iter); - assert(!ret); + 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) { + ret = lttng_ht_del(ust_app_ht_by_notify_sock, &iter); + assert(!ret); + } } rcu_read_unlock(); /* Destroy is done only when the ht is empty */ - ht_cleanup_push(ust_app_ht); - ht_cleanup_push(ust_app_ht_by_sock); - ht_cleanup_push(ust_app_ht_by_notify_sock); + if (ust_app_ht) { + ht_cleanup_push(ust_app_ht); + } + if (ust_app_ht_by_sock) { + ht_cleanup_push(ust_app_ht_by_sock); + } + if (ust_app_ht_by_notify_sock) { + ht_cleanup_push(ust_app_ht_by_notify_sock); + } } /* * Init UST app hash table. */ -void ust_app_ht_alloc(void) +int ust_app_ht_alloc(void) { ust_app_ht = lttng_ht_new(0, LTTNG_HT_TYPE_ULONG); + if (!ust_app_ht) { + return -1; + } ust_app_ht_by_sock = lttng_ht_new(0, LTTNG_HT_TYPE_ULONG); + if (!ust_app_ht_by_sock) { + return -1; + } ust_app_ht_by_notify_sock = lttng_ht_new(0, LTTNG_HT_TYPE_ULONG); + if (!ust_app_ht_by_notify_sock) { + return -1; + } + return 0; } /* @@ -3918,28 +4058,21 @@ error_rcu_unlock: return -1; } -/* - * Flush buffers for a specific UST session and app. - */ static -int ust_app_flush_trace(struct ltt_ust_session *usess, struct ust_app *app) +int ust_app_flush_app_session(struct ust_app *app, + struct ust_app_session *ua_sess) { - int ret = 0; + int ret, retval = 0; struct lttng_ht_iter iter; - struct ust_app_session *ua_sess; struct ust_app_channel *ua_chan; + struct consumer_socket *socket; - DBG("Flushing buffers for ust app pid %d", app->pid); + DBG("Flushing app session buffers for ust app pid %d", app->pid); rcu_read_lock(); if (!app->compatible) { - goto end_no_session; - } - - ua_sess = lookup_session_by_app(usess, app); - if (ua_sess == NULL) { - goto end_no_session; + goto end_not_compatible; } pthread_mutex_lock(&ua_sess->lock); @@ -3947,25 +4080,16 @@ int ust_app_flush_trace(struct ltt_ust_session *usess, struct ust_app *app) health_code_update(); /* Flushing buffers */ + socket = consumer_find_socket_by_bitness(app->bits_per_long, + ua_sess->consumer); cds_lfht_for_each_entry(ua_sess->channels->ht, &iter.iter, ua_chan, node.node) { health_code_update(); assert(ua_chan->is_sent); - ret = ustctl_sock_flush_buffer(app->sock, ua_chan->obj); - if (ret < 0) { - if (ret != -EPIPE && ret != -LTTNG_UST_ERR_EXITING) { - ERR("UST app PID %d channel %s flush failed with ret %d", - app->pid, ua_chan->name, ret); - } else { - DBG3("UST app failed to flush %s. Application is dead.", - ua_chan->name); - /* - * This is normal behavior, an application can die during the - * creation process. Don't report an error so the execution can - * continue normally. - */ - } - /* Continuing flushing all buffers */ + ret = consumer_flush_channel(socket, ua_chan->key); + if (ret) { + ERR("Error flushing consumer channel"); + retval = -1; continue; } } @@ -3973,10 +4097,37 @@ int ust_app_flush_trace(struct ltt_ust_session *usess, struct ust_app *app) health_code_update(); pthread_mutex_unlock(&ua_sess->lock); +end_not_compatible: + rcu_read_unlock(); + health_code_update(); + return retval; +} + +/* + * Flush buffers for a specific UST session and app. + */ +static +int ust_app_flush_session(struct ust_app *app, struct ltt_ust_session *usess) + +{ + int ret; + struct ust_app_session *ua_sess; + + DBG("Flushing session buffers for ust app pid %d", app->pid); + + rcu_read_lock(); + + ua_sess = lookup_session_by_app(usess, app); + if (ua_sess == NULL) { + ret = -1; + goto end_no_session; + } + ret = ust_app_flush_app_session(app, ua_sess); + end_no_session: rcu_read_unlock(); health_code_update(); - return 0; + return ret; } /* @@ -4110,7 +4261,7 @@ int ust_app_stop_trace_all(struct ltt_ust_session *usess) } case LTTNG_BUFFER_PER_PID: cds_lfht_for_each_entry(ust_app_ht->ht, &iter.iter, app, pid_n.node) { - ret = ust_app_flush_trace(usess, app); + ret = ust_app_flush_session(app, usess); if (ret < 0) { /* Continue to next apps even on error */ continue; @@ -4911,28 +5062,19 @@ void ust_app_destroy(struct ust_app *app) * Return 0 on success or else a negative value. */ int ust_app_snapshot_record(struct ltt_ust_session *usess, - struct snapshot_output *output, int wait, unsigned int nb_streams) + struct snapshot_output *output, int wait, uint64_t max_stream_size) { int ret = 0; unsigned int snapshot_done = 0; struct lttng_ht_iter iter; struct ust_app *app; char pathname[PATH_MAX]; - uint64_t max_stream_size = 0; assert(usess); assert(output); rcu_read_lock(); - /* - * Compute the maximum size of a single stream if a max size is asked by - * the caller. - */ - if (output->max_size > 0 && nb_streams > 0) { - max_stream_size = output->max_size / nb_streams; - } - switch (usess->buffer_type) { case LTTNG_BUFFER_PER_UID: { @@ -4962,30 +5104,16 @@ int ust_app_snapshot_record(struct ltt_ust_session *usess, /* Add the UST default trace dir to path. */ cds_lfht_for_each_entry(reg->registry->channels->ht, &iter.iter, reg_chan, node.node) { - - /* - * Make sure the maximum stream size is not lower than the - * subbuffer size or else it's an error since we won't be able to - * snapshot anything. - */ - if (max_stream_size && - reg_chan->subbuf_size > max_stream_size) { - ret = -EINVAL; - DBG3("UST app snapshot record maximum stream size %" PRIu64 - " is smaller than subbuffer size of %zu", - max_stream_size, reg_chan->subbuf_size); - goto error; - } - ret = consumer_snapshot_channel(socket, reg_chan->consumer_key, output, 0, - usess->uid, usess->gid, pathname, wait, + ret = consumer_snapshot_channel(socket, reg_chan->consumer_key, + output, 0, usess->uid, usess->gid, pathname, wait, max_stream_size); if (ret < 0) { goto error; } } - ret = consumer_snapshot_channel(socket, reg->registry->reg.ust->metadata_key, output, - 1, usess->uid, usess->gid, pathname, wait, - max_stream_size); + ret = consumer_snapshot_channel(socket, + reg->registry->reg.ust->metadata_key, output, 1, + usess->uid, usess->gid, pathname, wait, max_stream_size); if (ret < 0) { goto error; } @@ -5027,22 +5155,8 @@ int ust_app_snapshot_record(struct ltt_ust_session *usess, cds_lfht_for_each_entry(ua_sess->channels->ht, &chan_iter.iter, ua_chan, node.node) { - /* - * Make sure the maximum stream size is not lower than the - * subbuffer size or else it's an error since we won't be able to - * snapshot anything. - */ - if (max_stream_size && - ua_chan->attr.subbuf_size > max_stream_size) { - ret = -EINVAL; - DBG3("UST app snapshot record maximum stream size %" PRIu64 - " is smaller than subbuffer size of %" PRIu64, - max_stream_size, ua_chan->attr.subbuf_size); - goto error; - } - - ret = consumer_snapshot_channel(socket, ua_chan->key, output, 0, - ua_sess->euid, ua_sess->egid, pathname, wait, + ret = consumer_snapshot_channel(socket, ua_chan->key, output, + 0, ua_sess->euid, ua_sess->egid, pathname, wait, max_stream_size); if (ret < 0) { goto error; @@ -5099,10 +5213,12 @@ unsigned int ust_app_get_nb_stream(struct ltt_ust_session *usess) cds_list_for_each_entry(reg, &usess->buffer_reg_uid_list, lnode) { struct buffer_reg_channel *reg_chan; + rcu_read_lock(); cds_lfht_for_each_entry(reg->registry->channels->ht, &iter.iter, reg_chan, node.node) { ret += reg_chan->stream_count; } + rcu_read_unlock(); } break; }