Move to kernel style SPDX license identifiers
[lttng-tools.git] / src / bin / lttng-sessiond / dispatch.c
CommitLineData
5d1b0219 1/*
ab5be9fa
MJ
2 * Copyright (C) 2011 David Goulet <david.goulet@polymtl.ca>
3 * Copyright (C) 2011 Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
4 * Copyright (C) 2013 Jérémie Galarneau <jeremie.galarneau@efficios.com>
5d1b0219 5 *
ab5be9fa 6 * SPDX-License-Identifier: GPL-2.0-only
5d1b0219 7 *
5d1b0219
JG
8 */
9
10#include <stddef.h>
11#include <stdlib.h>
12#include <urcu.h>
13#include <common/futex.h>
14#include <common/macros.h>
15
16#include "dispatch.h"
17#include "ust-app.h"
18#include "testpoint.h"
19#include "fd-limit.h"
20#include "health-sessiond.h"
21#include "lttng-sessiond.h"
22#include "thread.h"
23
24struct thread_notifiers {
25 struct ust_cmd_queue *ust_cmd_queue;
26 int apps_cmd_pipe_write_fd;
27 int apps_cmd_notify_pipe_write_fd;
28 int dispatch_thread_exit;
29};
30
31/*
32 * For each tracing session, update newly registered apps. The session list
33 * lock MUST be acquired before calling this.
34 */
35static void update_ust_app(int app_sock)
36{
37 struct ltt_session *sess, *stmp;
38 const struct ltt_session_list *session_list = session_get_list();
39
40 /* Consumer is in an ERROR state. Stop any application update. */
41 if (uatomic_read(&ust_consumerd_state) == CONSUMER_ERROR) {
42 /* Stop the update process since the consumer is dead. */
43 return;
44 }
45
46 /* For all tracing session(s) */
47 cds_list_for_each_entry_safe(sess, stmp, &session_list->head, list) {
48 struct ust_app *app;
49
50 if (!session_get(sess)) {
51 continue;
52 }
53 session_lock(sess);
88e3c2f5 54 if (!sess->active || !sess->ust_session) {
5d1b0219
JG
55 goto unlock_session;
56 }
57
58 rcu_read_lock();
59 assert(app_sock >= 0);
60 app = ust_app_find_by_sock(app_sock);
61 if (app == NULL) {
62 /*
63 * Application can be unregistered before so
64 * this is possible hence simply stopping the
65 * update.
66 */
67 DBG3("UST app update failed to find app sock %d",
68 app_sock);
69 goto unlock_rcu;
70 }
71 ust_app_global_update(sess->ust_session, app);
72 unlock_rcu:
73 rcu_read_unlock();
74 unlock_session:
75 session_unlock(sess);
76 session_put(sess);
77 }
78}
79
80/*
81 * Sanitize the wait queue of the dispatch registration thread meaning removing
82 * invalid nodes from it. This is to avoid memory leaks for the case the UST
83 * notify socket is never received.
84 */
85static void sanitize_wait_queue(struct ust_reg_wait_queue *wait_queue)
86{
87 int ret, nb_fd = 0, i;
88 unsigned int fd_added = 0;
89 struct lttng_poll_event events;
90 struct ust_reg_wait_node *wait_node = NULL, *tmp_wait_node;
91
92 assert(wait_queue);
93
94 lttng_poll_init(&events);
95
96 /* Just skip everything for an empty queue. */
97 if (!wait_queue->count) {
98 goto end;
99 }
100
101 ret = lttng_poll_create(&events, wait_queue->count, LTTNG_CLOEXEC);
102 if (ret < 0) {
103 goto error_create;
104 }
105
106 cds_list_for_each_entry_safe(wait_node, tmp_wait_node,
107 &wait_queue->head, head) {
108 assert(wait_node->app);
109 ret = lttng_poll_add(&events, wait_node->app->sock,
110 LPOLLHUP | LPOLLERR);
111 if (ret < 0) {
112 goto error;
113 }
114
115 fd_added = 1;
116 }
117
118 if (!fd_added) {
119 goto end;
120 }
121
122 /*
123 * Poll but don't block so we can quickly identify the faulty events and
124 * clean them afterwards from the wait queue.
125 */
126 ret = lttng_poll_wait(&events, 0);
127 if (ret < 0) {
128 goto error;
129 }
130 nb_fd = ret;
131
132 for (i = 0; i < nb_fd; i++) {
133 /* Get faulty FD. */
134 uint32_t revents = LTTNG_POLL_GETEV(&events, i);
135 int pollfd = LTTNG_POLL_GETFD(&events, i);
136
5d1b0219
JG
137 cds_list_for_each_entry_safe(wait_node, tmp_wait_node,
138 &wait_queue->head, head) {
139 if (pollfd == wait_node->app->sock &&
140 (revents & (LPOLLHUP | LPOLLERR))) {
141 cds_list_del(&wait_node->head);
142 wait_queue->count--;
143 ust_app_destroy(wait_node->app);
144 free(wait_node);
145 /*
146 * Silence warning of use-after-free in
147 * cds_list_for_each_entry_safe which uses
148 * __typeof__(*wait_node).
149 */
150 wait_node = NULL;
151 break;
152 } else {
153 ERR("Unexpected poll events %u for sock %d", revents, pollfd);
154 goto error;
155 }
156 }
157 }
158
159 if (nb_fd > 0) {
160 DBG("Wait queue sanitized, %d node were cleaned up", nb_fd);
161 }
162
163end:
164 lttng_poll_clean(&events);
165 return;
166
167error:
168 lttng_poll_clean(&events);
169error_create:
170 ERR("Unable to sanitize wait queue");
171 return;
172}
173
174/*
175 * Send a socket to a thread This is called from the dispatch UST registration
176 * thread once all sockets are set for the application.
177 *
178 * The sock value can be invalid, we don't really care, the thread will handle
179 * it and make the necessary cleanup if so.
180 *
181 * On success, return 0 else a negative value being the errno message of the
182 * write().
183 */
184static int send_socket_to_thread(int fd, int sock)
185{
186 ssize_t ret;
187
188 /*
189 * It's possible that the FD is set as invalid with -1 concurrently just
190 * before calling this function being a shutdown state of the thread.
191 */
192 if (fd < 0) {
193 ret = -EBADF;
194 goto error;
195 }
196
197 ret = lttng_write(fd, &sock, sizeof(sock));
198 if (ret < sizeof(sock)) {
199 PERROR("write apps pipe %d", fd);
200 if (ret < 0) {
201 ret = -errno;
202 }
203 goto error;
204 }
205
206 /* All good. Don't send back the write positive ret value. */
207 ret = 0;
208error:
209 return (int) ret;
210}
211
212static void cleanup_ust_dispatch_thread(void *data)
213{
214 free(data);
215}
216
217/*
218 * Dispatch request from the registration threads to the application
219 * communication thread.
220 */
221static void *thread_dispatch_ust_registration(void *data)
222{
223 int ret, err = -1;
224 struct cds_wfcq_node *node;
225 struct ust_command *ust_cmd = NULL;
226 struct ust_reg_wait_node *wait_node = NULL, *tmp_wait_node;
227 struct ust_reg_wait_queue wait_queue = {
228 .count = 0,
229 };
230 struct thread_notifiers *notifiers = data;
231
232 rcu_register_thread();
233
234 health_register(health_sessiond, HEALTH_SESSIOND_TYPE_APP_REG_DISPATCH);
235
236 if (testpoint(sessiond_thread_app_reg_dispatch)) {
237 goto error_testpoint;
238 }
239
240 health_code_update();
241
242 CDS_INIT_LIST_HEAD(&wait_queue.head);
243
244 DBG("[thread] Dispatch UST command started");
245
246 for (;;) {
247 health_code_update();
248
249 /* Atomically prepare the queue futex */
250 futex_nto1_prepare(&notifiers->ust_cmd_queue->futex);
251
252 if (CMM_LOAD_SHARED(notifiers->dispatch_thread_exit)) {
253 break;
254 }
255
256 do {
257 struct ust_app *app = NULL;
258 ust_cmd = NULL;
259
260 /*
261 * Make sure we don't have node(s) that have hung up before receiving
262 * the notify socket. This is to clean the list in order to avoid
263 * memory leaks from notify socket that are never seen.
264 */
265 sanitize_wait_queue(&wait_queue);
266
267 health_code_update();
268 /* Dequeue command for registration */
269 node = cds_wfcq_dequeue_blocking(
270 &notifiers->ust_cmd_queue->head,
271 &notifiers->ust_cmd_queue->tail);
272 if (node == NULL) {
273 DBG("Woken up but nothing in the UST command queue");
274 /* Continue thread execution */
275 break;
276 }
277
278 ust_cmd = caa_container_of(node, struct ust_command, node);
279
280 DBG("Dispatching UST registration pid:%d ppid:%d uid:%d"
281 " gid:%d sock:%d name:%s (version %d.%d)",
282 ust_cmd->reg_msg.pid, ust_cmd->reg_msg.ppid,
283 ust_cmd->reg_msg.uid, ust_cmd->reg_msg.gid,
284 ust_cmd->sock, ust_cmd->reg_msg.name,
285 ust_cmd->reg_msg.major, ust_cmd->reg_msg.minor);
286
287 if (ust_cmd->reg_msg.type == USTCTL_SOCKET_CMD) {
288 wait_node = zmalloc(sizeof(*wait_node));
289 if (!wait_node) {
290 PERROR("zmalloc wait_node dispatch");
291 ret = close(ust_cmd->sock);
292 if (ret < 0) {
293 PERROR("close ust sock dispatch %d", ust_cmd->sock);
294 }
295 lttng_fd_put(LTTNG_FD_APPS, 1);
296 free(ust_cmd);
58f835e1 297 ust_cmd = NULL;
5d1b0219
JG
298 goto error;
299 }
300 CDS_INIT_LIST_HEAD(&wait_node->head);
301
302 /* Create application object if socket is CMD. */
303 wait_node->app = ust_app_create(&ust_cmd->reg_msg,
304 ust_cmd->sock);
305 if (!wait_node->app) {
306 ret = close(ust_cmd->sock);
307 if (ret < 0) {
308 PERROR("close ust sock dispatch %d", ust_cmd->sock);
309 }
310 lttng_fd_put(LTTNG_FD_APPS, 1);
311 free(wait_node);
58f835e1 312 wait_node = NULL;
5d1b0219 313 free(ust_cmd);
58f835e1 314 ust_cmd = NULL;
5d1b0219
JG
315 continue;
316 }
317 /*
318 * Add application to the wait queue so we can set the notify
319 * socket before putting this object in the global ht.
320 */
321 cds_list_add(&wait_node->head, &wait_queue.head);
322 wait_queue.count++;
323
324 free(ust_cmd);
58f835e1 325 ust_cmd = NULL;
5d1b0219
JG
326 /*
327 * We have to continue here since we don't have the notify
328 * socket and the application MUST be added to the hash table
329 * only at that moment.
330 */
331 continue;
332 } else {
333 /*
334 * Look for the application in the local wait queue and set the
335 * notify socket if found.
336 */
337 cds_list_for_each_entry_safe(wait_node, tmp_wait_node,
338 &wait_queue.head, head) {
339 health_code_update();
340 if (wait_node->app->pid == ust_cmd->reg_msg.pid) {
341 wait_node->app->notify_sock = ust_cmd->sock;
342 cds_list_del(&wait_node->head);
343 wait_queue.count--;
344 app = wait_node->app;
345 free(wait_node);
58f835e1 346 wait_node = NULL;
5d1b0219
JG
347 DBG3("UST app notify socket %d is set", ust_cmd->sock);
348 break;
349 }
350 }
351
352 /*
353 * With no application at this stage the received socket is
354 * basically useless so close it before we free the cmd data
355 * structure for good.
356 */
357 if (!app) {
358 ret = close(ust_cmd->sock);
359 if (ret < 0) {
360 PERROR("close ust sock dispatch %d", ust_cmd->sock);
361 }
362 lttng_fd_put(LTTNG_FD_APPS, 1);
363 }
364 free(ust_cmd);
58f835e1 365 ust_cmd = NULL;
5d1b0219
JG
366 }
367
368 if (app) {
369 /*
370 * @session_lock_list
371 *
372 * Lock the global session list so from the register up to the
373 * registration done message, no thread can see the application
374 * and change its state.
375 */
376 session_lock_list();
377 rcu_read_lock();
378
379 /*
380 * Add application to the global hash table. This needs to be
381 * done before the update to the UST registry can locate the
382 * application.
383 */
384 ust_app_add(app);
385
386 /* Set app version. This call will print an error if needed. */
387 (void) ust_app_version(app);
388
389 /* Send notify socket through the notify pipe. */
390 ret = send_socket_to_thread(
391 notifiers->apps_cmd_notify_pipe_write_fd,
392 app->notify_sock);
393 if (ret < 0) {
394 rcu_read_unlock();
395 session_unlock_list();
396 /*
397 * No notify thread, stop the UST tracing. However, this is
398 * not an internal error of the this thread thus setting
399 * the health error code to a normal exit.
400 */
401 err = 0;
402 goto error;
403 }
404
405 /*
406 * Update newly registered application with the tracing
407 * registry info already enabled information.
408 */
409 update_ust_app(app->sock);
410
411 /*
412 * Don't care about return value. Let the manage apps threads
413 * handle app unregistration upon socket close.
414 */
415 (void) ust_app_register_done(app);
416
417 /*
418 * Even if the application socket has been closed, send the app
419 * to the thread and unregistration will take place at that
420 * place.
421 */
422 ret = send_socket_to_thread(
423 notifiers->apps_cmd_pipe_write_fd,
424 app->sock);
425 if (ret < 0) {
426 rcu_read_unlock();
427 session_unlock_list();
428 /*
429 * No apps. thread, stop the UST tracing. However, this is
430 * not an internal error of the this thread thus setting
431 * the health error code to a normal exit.
432 */
433 err = 0;
434 goto error;
435 }
436
437 rcu_read_unlock();
438 session_unlock_list();
439 }
440 } while (node != NULL);
441
442 health_poll_entry();
443 /* Futex wait on queue. Blocking call on futex() */
444 futex_nto1_wait(&notifiers->ust_cmd_queue->futex);
445 health_poll_exit();
446 }
447 /* Normal exit, no error */
448 err = 0;
449
450error:
451 /* Clean up wait queue. */
452 cds_list_for_each_entry_safe(wait_node, tmp_wait_node,
453 &wait_queue.head, head) {
454 cds_list_del(&wait_node->head);
455 wait_queue.count--;
456 free(wait_node);
457 }
458
459 /* Empty command queue. */
460 for (;;) {
461 /* Dequeue command for registration */
462 node = cds_wfcq_dequeue_blocking(
463 &notifiers->ust_cmd_queue->head,
464 &notifiers->ust_cmd_queue->tail);
465 if (node == NULL) {
466 break;
467 }
468 ust_cmd = caa_container_of(node, struct ust_command, node);
469 ret = close(ust_cmd->sock);
470 if (ret < 0) {
471 PERROR("close ust sock exit dispatch %d", ust_cmd->sock);
472 }
473 lttng_fd_put(LTTNG_FD_APPS, 1);
474 free(ust_cmd);
475 }
476
477error_testpoint:
478 DBG("Dispatch thread dying");
479 if (err) {
480 health_error();
481 ERR("Health error occurred in %s", __func__);
482 }
483 health_unregister(health_sessiond);
484 rcu_unregister_thread();
485 return NULL;
486}
487
488static bool shutdown_ust_dispatch_thread(void *data)
489{
490 struct thread_notifiers *notifiers = data;
491
492 CMM_STORE_SHARED(notifiers->dispatch_thread_exit, 1);
493 futex_nto1_wake(&notifiers->ust_cmd_queue->futex);
494 return true;
495}
496
497bool launch_ust_dispatch_thread(struct ust_cmd_queue *cmd_queue,
498 int apps_cmd_pipe_write_fd,
499 int apps_cmd_notify_pipe_write_fd)
500{
501 struct lttng_thread *thread;
502 struct thread_notifiers *notifiers;
503
504 notifiers = zmalloc(sizeof(*notifiers));
505 if (!notifiers) {
506 goto error;
507 }
508 notifiers->ust_cmd_queue = cmd_queue;
509 notifiers->apps_cmd_pipe_write_fd = apps_cmd_pipe_write_fd;
510 notifiers->apps_cmd_notify_pipe_write_fd = apps_cmd_notify_pipe_write_fd;
511
512 thread = lttng_thread_create("UST registration dispatch",
513 thread_dispatch_ust_registration,
514 shutdown_ust_dispatch_thread,
515 cleanup_ust_dispatch_thread,
516 notifiers);
517 if (!thread) {
518 goto error;
519 }
520 lttng_thread_put(thread);
521 return true;
522error:
523 free(notifiers);
524 return false;
525}
This page took 0.045945 seconds and 4 git commands to generate.