X-Git-Url: https://git.lttng.org/?p=lttng-tools.git;a=blobdiff_plain;f=src%2Fbin%2Flttng-sessiond%2Fsession.c;h=f8134f8ad966a950b0c97f13bbb18f2ba12d89b9;hp=491de61759825f0485326cae5b2d7e72db72c6ad;hb=d7b377ed1acd4043bde353d99122e0e56fa4e975;hpb=b178f53e90c376dd44b020535c32649edef8f80e

diff --git a/src/bin/lttng-sessiond/session.c b/src/bin/lttng-sessiond/session.c
index 491de6175..f8134f8ad 100644
--- a/src/bin/lttng-sessiond/session.c
+++ b/src/bin/lttng-sessiond/session.c
@@ -1,18 +1,8 @@
 /*
- * Copyright (C)  2011 - David Goulet <david.goulet@polymtl.ca>
+ * Copyright (C) 2011 David Goulet <david.goulet@polymtl.ca>
  *
- * 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
@@ -28,6 +18,8 @@
 #include <pthread.h>
 
 #include <common/common.h>
+#include <common/utils.h>
+#include <common/trace-chunk.h>
 #include <common/sessiond-comm/sessiond-comm.h>
 #include <lttng/location-internal.h>
 #include "lttng-sessiond.h"
@@ -37,6 +29,17 @@
 #include "utils.h"
 #include "trace-ust.h"
 #include "timer.h"
+#include "cmd.h"
+
+struct ltt_session_destroy_notifier_element {
+	ltt_session_destroy_notifier notifier;
+	void *user_data;
+};
+
+struct ltt_session_clear_notifier_element {
+	ltt_session_clear_notifier notifier;
+	void *user_data;
+};
 
 /*
  * NOTES:
@@ -243,18 +246,28 @@ void session_get_net_consumer_ports(const struct ltt_session *session,
  * The caller must hold the session lock.
  */
 struct lttng_trace_archive_location *session_get_trace_archive_location(
-		struct ltt_session *session)
+		const struct ltt_session *session)
 {
+	int ret;
 	struct lttng_trace_archive_location *location = NULL;
+	char *chunk_path = NULL;
 
-	if (session->rotation_state != LTTNG_ROTATION_STATE_COMPLETED) {
+	if (session->rotation_state != LTTNG_ROTATION_STATE_COMPLETED ||
+			!session->last_archived_chunk_name) {
 		goto end;
 	}
 
 	switch (session_get_consumer_destination_type(session)) {
 	case CONSUMER_DST_LOCAL:
+		ret = asprintf(&chunk_path,
+				"%s/" DEFAULT_ARCHIVED_TRACE_CHUNKS_DIRECTORY "/%s",
+				session_get_base_path(session),
+				session->last_archived_chunk_name);
+		if (ret == -1) {
+			goto end;
+		}
 		location = lttng_trace_archive_location_local_create(
-				session->rotation_chunk.current_rotate_path);
+				chunk_path);
 		break;
 	case CONSUMER_DST_NET:
 	{
@@ -268,14 +281,14 @@ struct lttng_trace_archive_location *session_get_trace_archive_location(
 		location = lttng_trace_archive_location_relay_create(
 				hostname,
 				LTTNG_TRACE_ARCHIVE_LOCATION_RELAY_PROTOCOL_TYPE_TCP,
-				control_port, data_port,
-				session->rotation_chunk.current_rotate_path);
+				control_port, data_port, session->last_chunk_path);
 		break;
 	}
 	default:
 		abort();
 	}
 end:
+	free(chunk_path);
 	return location;
 }
 
@@ -284,7 +297,7 @@ end:
  *
  * The session list lock must be held.
  */
-int ltt_sessions_ht_alloc(void)
+static int ltt_sessions_ht_alloc(void)
 {
 	int ret = 0;
 
@@ -400,6 +413,511 @@ void session_unlock(struct ltt_session *session)
 	pthread_mutex_unlock(&session->lock);
 }
 
+static
+int _session_set_trace_chunk_no_lock_check(struct ltt_session *session,
+		struct lttng_trace_chunk *new_trace_chunk,
+		struct lttng_trace_chunk **_current_trace_chunk)
+{
+	int ret = 0;
+	unsigned int i, refs_to_acquire = 0, refs_acquired = 0, refs_to_release = 0;
+	struct cds_lfht_iter iter;
+	struct consumer_socket *socket;
+	struct lttng_trace_chunk *current_trace_chunk;
+	uint64_t chunk_id;
+	enum lttng_trace_chunk_status chunk_status;
+
+	rcu_read_lock();
+	/*
+	 * Ownership of current trace chunk is transferred to
+	 * `current_trace_chunk`.
+	 */
+	current_trace_chunk = session->current_trace_chunk;
+	session->current_trace_chunk = NULL;
+	if (session->ust_session) {
+		lttng_trace_chunk_put(
+				session->ust_session->current_trace_chunk);
+		session->ust_session->current_trace_chunk = NULL;
+	}
+	if (session->kernel_session) {
+		lttng_trace_chunk_put(
+				session->kernel_session->current_trace_chunk);
+		session->kernel_session->current_trace_chunk = NULL;
+	}
+	if (!new_trace_chunk) {
+		ret = 0;
+		goto end;
+	}
+	chunk_status = lttng_trace_chunk_get_id(new_trace_chunk, &chunk_id);
+	assert(chunk_status == LTTNG_TRACE_CHUNK_STATUS_OK);
+
+	refs_to_acquire = 1;
+	refs_to_acquire += !!session->ust_session;
+	refs_to_acquire += !!session->kernel_session;
+
+	for (refs_acquired = 0; refs_acquired < refs_to_acquire;
+			refs_acquired++) {
+		if (!lttng_trace_chunk_get(new_trace_chunk)) {
+			ERR("Failed to acquire reference to new trace chunk of session \"%s\"",
+					session->name);
+			goto error;
+		}
+	}
+
+	if (session->ust_session) {
+		const uint64_t relayd_id =
+				session->ust_session->consumer->net_seq_index;
+		const bool is_local_trace =
+				session->ust_session->consumer->type ==
+				CONSUMER_DST_LOCAL;
+
+		session->ust_session->current_trace_chunk = new_trace_chunk;
+		if (is_local_trace) {
+			enum lttng_error_code ret_error_code;
+
+			ret_error_code = ust_app_create_channel_subdirectories(
+					session->ust_session);
+			if (ret_error_code != LTTNG_OK) {
+				goto error;
+			}
+		}
+		cds_lfht_for_each_entry(
+				session->ust_session->consumer->socks->ht,
+				&iter, socket, node.node) {
+			pthread_mutex_lock(socket->lock);
+			ret = consumer_create_trace_chunk(socket,
+					relayd_id,
+					session->id, new_trace_chunk,
+					DEFAULT_UST_TRACE_DIR);
+			pthread_mutex_unlock(socket->lock);
+			if (ret) {
+				goto error;
+			}
+		}
+	}
+	if (session->kernel_session) {
+		const uint64_t relayd_id =
+				session->kernel_session->consumer->net_seq_index;
+		const bool is_local_trace =
+				session->kernel_session->consumer->type ==
+				CONSUMER_DST_LOCAL;
+
+		session->kernel_session->current_trace_chunk = new_trace_chunk;
+		if (is_local_trace) {
+			enum lttng_error_code ret_error_code;
+
+			ret_error_code = kernel_create_channel_subdirectories(
+					session->kernel_session);
+			if (ret_error_code != LTTNG_OK) {
+				goto error;
+			}
+		}
+		cds_lfht_for_each_entry(
+				session->kernel_session->consumer->socks->ht,
+				&iter, socket, node.node) {
+			pthread_mutex_lock(socket->lock);
+			ret = consumer_create_trace_chunk(socket,
+					relayd_id,
+					session->id, new_trace_chunk,
+					DEFAULT_KERNEL_TRACE_DIR);
+			pthread_mutex_unlock(socket->lock);
+			if (ret) {
+				goto error;
+			}
+		}
+	}
+
+	/*
+	 * Update local current trace chunk state last, only if all remote
+	 * creations succeeded.
+	 */
+	session->current_trace_chunk = new_trace_chunk;
+	LTTNG_OPTIONAL_SET(&session->most_recent_chunk_id, chunk_id);
+end:
+	if (_current_trace_chunk) {
+		*_current_trace_chunk = current_trace_chunk;
+		current_trace_chunk = NULL;
+	}
+end_no_move:
+	rcu_read_unlock();
+	lttng_trace_chunk_put(current_trace_chunk);
+	return ret;
+error:
+	if (session->ust_session) {
+		session->ust_session->current_trace_chunk = NULL;
+	}
+	if (session->kernel_session) {
+		session->kernel_session->current_trace_chunk = NULL;
+	}
+	/*
+	 * Release references taken in the case where all references could not
+	 * be acquired.
+	 */
+	refs_to_release = refs_to_acquire - refs_acquired;
+	for (i = 0; i < refs_to_release; i++) {
+		lttng_trace_chunk_put(new_trace_chunk);
+	}
+	ret = -1;
+	goto end_no_move;
+}
+
+struct lttng_trace_chunk *session_create_new_trace_chunk(
+		const struct ltt_session *session,
+		const struct consumer_output *consumer_output_override,
+		const char *session_base_path_override,
+		const char *chunk_name_override)
+{
+	int ret;
+	struct lttng_trace_chunk *trace_chunk = NULL;
+	enum lttng_trace_chunk_status chunk_status;
+	const time_t chunk_creation_ts = time(NULL);
+	bool is_local_trace;
+	const char *base_path;
+	struct lttng_directory_handle *session_output_directory = NULL;
+	const struct lttng_credentials session_credentials = {
+		.uid = session->uid,
+		.gid = session->gid,
+	};
+	uint64_t next_chunk_id;
+	const struct consumer_output *output;
+	const char *new_path;
+
+	if (consumer_output_override) {
+		output = consumer_output_override;
+	} else {
+		assert(session->ust_session || session->kernel_session);
+		output = session->ust_session ?
+					 session->ust_session->consumer :
+					 session->kernel_session->consumer;
+	}
+
+	is_local_trace = output->type == CONSUMER_DST_LOCAL;
+	base_path = session_base_path_override ? :
+			consumer_output_get_base_path(output);
+
+	if (chunk_creation_ts == (time_t) -1) {
+		PERROR("Failed to sample time while creation session \"%s\" trace chunk",
+				session->name);
+		goto error;
+	}
+
+	next_chunk_id = session->most_recent_chunk_id.is_set ?
+			session->most_recent_chunk_id.value + 1 : 0;
+
+	if (session->current_trace_chunk &&
+			!lttng_trace_chunk_get_name_overridden(session->current_trace_chunk)) {
+		chunk_status = lttng_trace_chunk_rename_path(session->current_trace_chunk,
+					DEFAULT_CHUNK_TMP_OLD_DIRECTORY);
+		if (chunk_status != LTTNG_TRACE_CHUNK_STATUS_OK) {
+			goto error;
+		}
+	}
+	if (!session->current_trace_chunk) {
+		if (!session->rotated) {
+			new_path = "";
+		} else {
+			new_path = NULL;
+		}
+	} else {
+		new_path = DEFAULT_CHUNK_TMP_NEW_DIRECTORY;
+	}
+
+	trace_chunk = lttng_trace_chunk_create(next_chunk_id,
+			chunk_creation_ts, new_path);
+	if (!trace_chunk) {
+		goto error;
+	}
+
+	if (chunk_name_override) {
+		chunk_status = lttng_trace_chunk_override_name(trace_chunk,
+				chunk_name_override);
+		if (chunk_status != LTTNG_TRACE_CHUNK_STATUS_OK) {
+			goto error;
+		}
+	}
+
+	if (!is_local_trace) {
+		/*
+		 * No need to set crendentials and output directory
+		 * for remote trace chunks.
+		 */
+		goto end;
+	}
+
+	chunk_status = lttng_trace_chunk_set_credentials(trace_chunk,
+			&session_credentials);
+	if (chunk_status != LTTNG_TRACE_CHUNK_STATUS_OK) {
+		goto error;
+	}
+
+	DBG("Creating base output directory of session \"%s\" at %s",
+			session->name, base_path);
+	ret = utils_mkdir_recursive(base_path, S_IRWXU | S_IRWXG,
+			session->uid, session->gid);
+	if (ret) {
+		goto error;
+	}
+	session_output_directory = lttng_directory_handle_create(base_path);
+	if (!session_output_directory) {
+		goto error;
+	}
+	chunk_status = lttng_trace_chunk_set_as_owner(trace_chunk,
+			session_output_directory);
+	lttng_directory_handle_put(session_output_directory);
+	session_output_directory = NULL;
+	if (chunk_status != LTTNG_TRACE_CHUNK_STATUS_OK) {
+		goto error;
+	}
+end:
+	return trace_chunk;
+error:
+	lttng_directory_handle_put(session_output_directory);
+	lttng_trace_chunk_put(trace_chunk);
+	trace_chunk = NULL;
+	goto end;
+}
+
+int session_close_trace_chunk(struct ltt_session *session,
+		struct lttng_trace_chunk *trace_chunk,
+		enum lttng_trace_chunk_command_type close_command,
+		char *closed_trace_chunk_path)
+{
+	int ret = 0;
+	bool error_occurred = false;
+	struct cds_lfht_iter iter;
+	struct consumer_socket *socket;
+	enum lttng_trace_chunk_status chunk_status;
+	const time_t chunk_close_timestamp = time(NULL);
+	const char *new_path;
+
+	chunk_status = lttng_trace_chunk_set_close_command(
+			trace_chunk, close_command);
+	if (chunk_status != LTTNG_TRACE_CHUNK_STATUS_OK) {
+		ret = -1;
+		goto end;
+	}
+
+	if (chunk_close_timestamp == (time_t) -1) {
+		ERR("Failed to sample the close timestamp of the current trace chunk of session \"%s\"",
+				session->name);
+		ret = -1;
+		goto end;
+	}
+
+	if (close_command == LTTNG_TRACE_CHUNK_COMMAND_TYPE_DELETE && !session->rotated) {
+		/* New chunk stays in session output directory. */
+		new_path = "";
+	} else {
+		/* Use chunk name for new chunk. */
+		new_path = NULL;
+	}
+	if (session->current_trace_chunk &&
+			!lttng_trace_chunk_get_name_overridden(session->current_trace_chunk)) {
+		/* Rename new chunk path. */
+		chunk_status = lttng_trace_chunk_rename_path(session->current_trace_chunk,
+					new_path);
+		if (chunk_status != LTTNG_TRACE_CHUNK_STATUS_OK) {
+			ret = -1;
+			goto end;
+		}
+	}
+	if (!lttng_trace_chunk_get_name_overridden(trace_chunk) &&
+			close_command == LTTNG_TRACE_CHUNK_COMMAND_TYPE_NO_OPERATION) {
+		const char *old_path;
+
+		if (!session->rotated) {
+			old_path = "";
+		} else {
+			old_path = NULL;
+		}
+		/* We need to move back the .tmp_old_chunk to its rightful place. */
+		chunk_status = lttng_trace_chunk_rename_path(trace_chunk,
+					old_path);
+		if (chunk_status != LTTNG_TRACE_CHUNK_STATUS_OK) {
+			ret = -1;
+			goto end;
+		}
+	}
+	if (close_command == LTTNG_TRACE_CHUNK_COMMAND_TYPE_MOVE_TO_COMPLETED) {
+		session->rotated = true;
+	}
+	chunk_status = lttng_trace_chunk_set_close_timestamp(trace_chunk,
+			chunk_close_timestamp);
+	if (chunk_status != LTTNG_TRACE_CHUNK_STATUS_OK) {
+		ERR("Failed to set the close timestamp of the current trace chunk of session \"%s\"",
+				session->name);
+		ret = -1;
+		goto end;
+	}
+
+	if (session->ust_session) {
+		const uint64_t relayd_id =
+				session->ust_session->consumer->net_seq_index;
+
+		cds_lfht_for_each_entry(
+				session->ust_session->consumer->socks->ht,
+				&iter, socket, node.node) {
+			pthread_mutex_lock(socket->lock);
+			ret = consumer_close_trace_chunk(socket,
+					relayd_id,
+					session->id,
+					trace_chunk, closed_trace_chunk_path);
+			pthread_mutex_unlock(socket->lock);
+			if (ret) {
+				ERR("Failed to close trace chunk on user space consumer");
+				error_occurred = true;
+			}
+		}
+	}
+	if (session->kernel_session) {
+		const uint64_t relayd_id =
+				session->kernel_session->consumer->net_seq_index;
+
+		cds_lfht_for_each_entry(
+				session->kernel_session->consumer->socks->ht,
+				&iter, socket, node.node) {
+			pthread_mutex_lock(socket->lock);
+			ret = consumer_close_trace_chunk(socket,
+					relayd_id,
+					session->id,
+					trace_chunk, closed_trace_chunk_path);
+			pthread_mutex_unlock(socket->lock);
+			if (ret) {
+				ERR("Failed to close trace chunk on kernel consumer");
+				error_occurred = true;
+			}
+		}
+	}
+	ret = error_occurred ? -1 : 0;
+end:
+	return 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.
+ */
+static
+enum lttng_error_code session_kernel_open_packets(struct ltt_session *session)
+{
+	enum lttng_error_code ret = LTTNG_OK;
+	struct consumer_socket *socket;
+	struct lttng_ht_iter iter;
+	struct cds_lfht_node *node;
+	struct ltt_kernel_channel *chan;
+
+	rcu_read_lock();
+
+	cds_lfht_first(session->kernel_session->consumer->socks->ht, &iter.iter);
+	node = cds_lfht_iter_get_node(&iter.iter);
+	socket = container_of(node, typeof(*socket), node.node);
+
+	cds_list_for_each_entry(chan,
+			&session->kernel_session->channel_list.head, list) {
+		int open_ret;
+
+		DBG("Open packet of kernel channel: channel key = %" PRIu64
+				", session name = %s, session_id = %" PRIu64,
+				chan->key, session->name, session->id);
+
+		open_ret = consumer_open_channel_packets(socket, chan->key);
+		if (open_ret < 0) {
+			/* General error (no known error expected). */
+			ret = LTTNG_ERR_UNK;
+			goto end;
+		}
+	}
+
+end:
+	rcu_read_unlock();
+	return ret;
+}
+
+enum lttng_error_code session_open_packets(struct ltt_session *session)
+{
+	enum lttng_error_code ret = LTTNG_OK;
+
+	DBG("Opening packets of session channels: session name = %s, session id = %" PRIu64,
+			session->name, session->id);
+
+	if (session->ust_session) {
+		ret = ust_app_open_packets(session);
+		if (ret != LTTNG_OK) {
+			goto end;
+		}
+	}
+
+	if (session->kernel_session) {
+		ret = session_kernel_open_packets(session);
+		if (ret != LTTNG_OK) {
+			goto end;
+		}
+	}
+
+end:
+	return ret;
+}
+
+/*
+ * Set a session's current trace chunk.
+ *
+ * Must be called with the session lock held.
+ */
+int session_set_trace_chunk(struct ltt_session *session,
+		struct lttng_trace_chunk *new_trace_chunk,
+		struct lttng_trace_chunk **current_trace_chunk)
+{
+	ASSERT_LOCKED(session->lock);
+	return _session_set_trace_chunk_no_lock_check(session, new_trace_chunk,
+			current_trace_chunk);
+}
+
+static
+void session_notify_destruction(const struct ltt_session *session)
+{
+	size_t i;
+	const size_t count = lttng_dynamic_array_get_count(
+			&session->destroy_notifiers);
+
+	for (i = 0; i < count; i++) {
+		const struct ltt_session_destroy_notifier_element *element =
+			lttng_dynamic_array_get_element(
+					&session->destroy_notifiers, i);
+
+		element->notifier(session, element->user_data);
+	}
+}
+
+/*
+ * Fire each clear notifier once, and remove them from the array.
+ */
+void session_notify_clear(struct ltt_session *session)
+{
+	size_t i;
+	const size_t count = lttng_dynamic_array_get_count(
+			&session->clear_notifiers);
+
+	for (i = 0; i < count; i++) {
+		const struct ltt_session_clear_notifier_element *element =
+			lttng_dynamic_array_get_element(
+					&session->clear_notifiers, i);
+
+		element->notifier(session, element->user_data);
+	}
+	lttng_dynamic_array_clear(&session->clear_notifiers);
+}
+
 static
 void session_release(struct urcu_ref *ref)
 {
@@ -407,14 +925,17 @@ void session_release(struct urcu_ref *ref)
 	struct ltt_ust_session *usess;
 	struct ltt_kernel_session *ksess;
 	struct ltt_session *session = container_of(ref, typeof(*session), ref);
+	const bool session_published = session->published;
+
+	assert(!session->chunk_being_archived);
 
 	usess = session->ust_session;
 	ksess = session->kernel_session;
 
-	/* Clean kernel session teardown */
+	/* Clean kernel session teardown, keeping data for destroy notifier. */
 	kernel_destroy_session(ksess);
 
-	/* UST session teardown */
+	/* UST session teardown, keeping data for destroy notifier. */
 	if (usess) {
 		/* Close any relayd session */
 		consumer_output_send_destroy_relayd(usess->consumer);
@@ -425,7 +946,7 @@ void session_release(struct urcu_ref *ref)
 			ERR("Error in ust_app_destroy_trace_all");
 		}
 
-		/* Clean up the rest. */
+		/* Clean up the rest, keeping destroy notifier data. */
 		trace_ust_destroy_session(usess);
 	}
 
@@ -439,18 +960,37 @@ void session_release(struct urcu_ref *ref)
 	}
 
 	DBG("Destroying session %s (id %" PRIu64 ")", session->name, session->id);
-	pthread_mutex_destroy(&session->lock);
 
-	consumer_output_put(session->consumer);
 	snapshot_destroy(&session->snapshot);
 
-	if (session->published) {
+	pthread_mutex_destroy(&session->lock);
+
+	if (session_published) {
 		ASSERT_LOCKED(ltt_session_list.lock);
 		del_session_list(session);
 		del_session_ht(session);
-		pthread_cond_broadcast(&ltt_session_list.removal_cond);
 	}
+	session_notify_destruction(session);
+
+	consumer_output_put(session->consumer);
+	kernel_free_session(ksess);
+	session->kernel_session = NULL;
+	if (usess) {
+		trace_ust_free_session(usess);
+		session->ust_session = NULL;
+	}
+	lttng_dynamic_array_reset(&session->destroy_notifiers);
+	lttng_dynamic_array_reset(&session->clear_notifiers);
+	free(session->last_archived_chunk_name);
+	free(session->base_path);
 	free(session);
+	if (session_published) {
+		/*
+		 * Broadcast after free-ing to ensure the memory is
+		 * reclaimed before the main thread exits.
+		 */
+		pthread_cond_broadcast(&ltt_session_list.removal_cond);
+	}
 }
 
 /*
@@ -497,6 +1037,30 @@ void session_destroy(struct ltt_session *session)
 	session_put(session);
 }
 
+int session_add_destroy_notifier(struct ltt_session *session,
+		ltt_session_destroy_notifier notifier, void *user_data)
+{
+	const struct ltt_session_destroy_notifier_element element = {
+		.notifier = notifier,
+		.user_data = user_data
+	};
+
+	return lttng_dynamic_array_add_element(&session->destroy_notifiers,
+			&element);
+}
+
+int session_add_clear_notifier(struct ltt_session *session,
+		ltt_session_clear_notifier notifier, void *user_data)
+{
+	const struct ltt_session_clear_notifier_element element = {
+		.notifier = notifier,
+		.user_data = user_data
+	};
+
+	return lttng_dynamic_array_add_element(&session->clear_notifiers,
+			&element);
+}
+
 /*
  * Return a ltt_session structure ptr that matches name. If no session found,
  * NULL is returned. This must be called with the session list lock held using
@@ -585,6 +1149,12 @@ enum lttng_error_code session_create(const char *name, uid_t uid, gid_t gid,
 		goto error;
 	}
 
+	lttng_dynamic_array_init(&new_session->destroy_notifiers,
+			sizeof(struct ltt_session_destroy_notifier_element),
+			NULL);
+	lttng_dynamic_array_init(&new_session->clear_notifiers,
+			sizeof(struct ltt_session_clear_notifier_element),
+			NULL);
 	urcu_ref_init(&new_session->ref);
 	pthread_mutex_init(&new_session->lock, NULL);
 
@@ -641,6 +1211,7 @@ enum lttng_error_code session_create(const char *name, uid_t uid, gid_t gid,
 						DEFAULT_SESSION_NAME, i,
 						datetime);
 			}
+			new_session->name_contains_creation_time = true;
 			if (ret == -1 || ret >= sizeof(new_session->name)) {
 				/*
 				 * Null-terminate in case the name is used
@@ -725,18 +1296,13 @@ error:
 }
 
 /*
- * Check if the UID or GID match the session. Root user has access to all
+ * Check if the UID matches the session. Root user has access to all
  * sessions.
  */
-int session_access_ok(struct ltt_session *session, uid_t uid, gid_t gid)
+bool session_access_ok(struct ltt_session *session, uid_t uid)
 {
 	assert(session);
-
-	if (uid != session->uid && gid != session->gid && uid != 0) {
-		return 0;
-	} else {
-		return 1;
-	}
+	return (uid == session->uid) || uid == 0;
 }
 
 /*
@@ -748,7 +1314,7 @@ int session_access_ok(struct ltt_session *session, uid_t uid, gid_t gid)
  *
  * Be careful of the result passed to this function. For instance,
  * on failure to launch a rotation, a client will expect the rotation
- * state to be set to "NO_ROTATION". If an error occured while the
+ * state to be set to "NO_ROTATION". If an error occurred while the
  * rotation was "ONGOING", result should be set to "ERROR", which will
  * allow a client to report it.
  *
@@ -762,12 +1328,27 @@ int session_reset_rotation_state(struct ltt_session *session,
 	ASSERT_LOCKED(ltt_session_list.lock);
 	ASSERT_LOCKED(session->lock);
 
-	session->rotation_pending_local = false;
-	session->rotation_pending_relay = false;
-	session->rotated_after_last_stop = false;
 	session->rotation_state = result;
 	if (session->rotation_pending_check_timer_enabled) {
 		ret = timer_session_rotation_pending_check_stop(session);
 	}
+	if (session->chunk_being_archived) {
+		uint64_t chunk_id;
+		enum lttng_trace_chunk_status chunk_status;
+
+		chunk_status = lttng_trace_chunk_get_id(
+				session->chunk_being_archived,
+				&chunk_id);
+		assert(chunk_status == LTTNG_TRACE_CHUNK_STATUS_OK);
+		LTTNG_OPTIONAL_SET(&session->last_archived_chunk_id,
+				chunk_id);
+		lttng_trace_chunk_put(session->chunk_being_archived);
+		session->chunk_being_archived = NULL;
+		/*
+		 * Fire the clear reply notifiers if we are completing a clear
+		 * rotation.
+		 */
+		session_notify_clear(session);
+	}
 	return ret;
 }