X-Git-Url: https://git.lttng.org/?a=blobdiff_plain;f=src%2Fbin%2Flttng-sessiond%2Fmain.c;h=0e7de6e5e4786d27383eec0481a48504a84b5c04;hb=fb27f84c28e9d8ec7c8ce5ac1b541b0671a4569e;hp=d284debd695e76787189f5f5b7cae256a67bf3e2;hpb=47b7e2f9d1199c32f87138f4c226e9c5153c7968;p=lttng-tools.git

diff --git a/src/bin/lttng-sessiond/main.c b/src/bin/lttng-sessiond/main.c
index d284debd6..0e7de6e5e 100644
--- a/src/bin/lttng-sessiond/main.c
+++ b/src/bin/lttng-sessiond/main.c
@@ -586,6 +586,34 @@ static int generate_lock_file_path(char *path, size_t len)
 	return ret;
 }
 
+/*
+ * Wait on consumer process termination.
+ *
+ * Need to be called with the consumer data lock held or from a context
+ * ensuring no concurrent access to data (e.g: cleanup).
+ */
+static void wait_consumer(struct consumer_data *consumer_data)
+{
+	pid_t ret;
+	int status;
+
+	if (consumer_data->pid <= 0) {
+		return;
+	}
+
+	DBG("Waiting for complete teardown of consumerd (PID: %d)",
+			consumer_data->pid);
+	ret = waitpid(consumer_data->pid, &status, 0);
+	if (ret == -1) {
+		PERROR("consumerd waitpid pid: %d", consumer_data->pid)
+	}
+	if (!WIFEXITED(status)) {
+		ERR("consumerd termination with error: %d",
+				WEXITSTATUS(ret));
+	}
+	consumer_data->pid = 0;
+}
+
 /*
  * Cleanup the session daemon's data structures.
  */
@@ -680,6 +708,10 @@ static void sessiond_cleanup(void)
 		}
 	}
 
+	wait_consumer(&kconsumer_data);
+	wait_consumer(&ustconsumer64_data);
+	wait_consumer(&ustconsumer32_data);
+
 	DBG("Cleaning up all agent apps");
 	agent_app_ht_clean();
 
@@ -1216,6 +1248,9 @@ static void *thread_manage_consumer(void *data)
 
 	DBG("[thread] Manage consumer started");
 
+	rcu_register_thread();
+	rcu_thread_online();
+
 	health_register(health_sessiond, HEALTH_SESSIOND_TYPE_CONSUMER);
 
 	health_code_update();
@@ -1497,7 +1532,6 @@ error:
 
 	unlink(consumer_data->err_unix_sock_path);
 	unlink(consumer_data->cmd_unix_sock_path);
-	consumer_data->pid = 0;
 	pthread_mutex_unlock(&consumer_data->lock);
 
 	/* Cleanup metadata socket mutex. */
@@ -1514,6 +1548,9 @@ error_poll:
 	health_unregister(health_sessiond);
 	DBG("consumer thread cleanup completed");
 
+	rcu_thread_offline();
+	rcu_unregister_thread();
+
 	return NULL;
 }
 
@@ -1781,6 +1818,12 @@ static void sanitize_wait_queue(struct ust_reg_wait_queue *wait_queue)
 				wait_queue->count--;
 				ust_app_destroy(wait_node->app);
 				free(wait_node);
+				/*
+				 * Silence warning of use-after-free in
+				 * cds_list_for_each_entry_safe which uses
+				 * __typeof__(*wait_node).
+				 */
+				wait_node = NULL;
 				break;
 			}
 		}
@@ -1983,7 +2026,7 @@ static void *thread_dispatch_ust_registration(void *data)
 				 * Don't care about return value. Let the manage apps threads
 				 * handle app unregistration upon socket close.
 				 */
-				(void) ust_app_register_done(app->sock);
+				(void) ust_app_register_done(app);
 
 				/*
 				 * Even if the application socket has been closed, send the app
@@ -2025,6 +2068,22 @@ error:
 		free(wait_node);
 	}
 
+	/* Empty command queue. */
+	for (;;) {
+		/* Dequeue command for registration */
+		node = cds_wfcq_dequeue_blocking(&ust_cmd_queue.head, &ust_cmd_queue.tail);
+		if (node == NULL) {
+			break;
+		}
+		ust_cmd = caa_container_of(node, struct ust_command, node);
+		ret = close(ust_cmd->sock);
+		if (ret < 0) {
+			PERROR("close ust sock exit dispatch %d", ust_cmd->sock);
+		}
+		lttng_fd_put(LTTNG_FD_APPS, 1);
+		free(ust_cmd);
+	}
+
 error_testpoint:
 	DBG("Dispatch thread dying");
 	if (err) {
@@ -2157,6 +2216,10 @@ static void *thread_registration_apps(void *data)
 					ust_cmd = zmalloc(sizeof(struct ust_command));
 					if (ust_cmd == NULL) {
 						PERROR("ust command zmalloc");
+						ret = close(sock);
+						if (ret) {
+							PERROR("close");
+						}
 						goto error;
 					}
 
@@ -2713,7 +2776,7 @@ static int copy_session_consumer(int domain, struct ltt_session *session)
 		 * domain.
 		 */
 		if (session->kernel_session->consumer) {
-			consumer_destroy_output(session->kernel_session->consumer);
+			consumer_output_put(session->kernel_session->consumer);
 		}
 		session->kernel_session->consumer =
 			consumer_copy_output(session->consumer);
@@ -2727,7 +2790,7 @@ static int copy_session_consumer(int domain, struct ltt_session *session)
 	case LTTNG_DOMAIN_UST:
 		DBG3("Copying tracing session consumer output in UST session");
 		if (session->ust_session->consumer) {
-			consumer_destroy_output(session->ust_session->consumer);
+			consumer_output_put(session->ust_session->consumer);
 		}
 		session->ust_session->consumer =
 			consumer_copy_output(session->consumer);
@@ -2847,7 +2910,7 @@ static int create_kernel_session(struct ltt_session *session)
 				session->kernel_session->consumer->dst.trace_path,
 				S_IRWXU | S_IRWXG, session->uid, session->gid);
 		if (ret < 0) {
-			if (ret != -EEXIST) {
+			if (errno != EEXIST) {
 				ERR("Trace directory creation error");
 				goto error;
 			}
@@ -2909,6 +2972,8 @@ static int process_client_msg(struct command_ctx *cmd_ctx, int sock,
 
 	DBG("Processing client command %d", cmd_ctx->lsm->cmd_type);
 
+	assert(!rcu_read_ongoing());
+
 	*sock_error = 0;
 
 	switch (cmd_ctx->lsm->cmd_type) {
@@ -3996,6 +4061,7 @@ setup_error:
 		session_unlock_list();
 	}
 init_setup_error:
+	assert(!rcu_read_ongoing());
 	return ret;
 }
 
@@ -5963,6 +6029,10 @@ exit_apps:
 	}
 exit_reg_apps:
 
+	/*
+	 * Join dispatch thread after joining reg_apps_thread to ensure
+	 * we don't leak applications in the queue.
+	 */
 	ret = pthread_join(dispatch_thread, &status);
 	if (ret) {
 		errno = ret;
@@ -6028,6 +6098,9 @@ exit_options:
 	sessiond_cleanup_options();
 
 exit_set_signal_handler:
+	/* Ensure all prior call_rcu are done. */
+	rcu_barrier();
+
 	if (!retval) {
 		exit(EXIT_SUCCESS);
 	} else {