src/common/kernel-consumer/kernel-consumer.cpp

   1 /*
   2  * Copyright (C) 2011 EfficiOS Inc.
   3  * Copyright (C) 2011 Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
   4  * Copyright (C) 2017 Jérémie Galarneau <jeremie.galarneau@efficios.com>
   5  *
   6  * SPDX-License-Identifier: GPL-2.0-only
   7  *
   8  */
   9
  10 #define _LGPL_SOURCE
  11 #include <poll.h>
  12 #include <pthread.h>
  13 #include <stdlib.h>
  14 #include <string.h>
  15 #include <sys/mman.h>
  16 #include <sys/socket.h>
  17 #include <sys/types.h>
  18 #include <inttypes.h>
  19 #include <unistd.h>
  20 #include <sys/stat.h>
  21 #include <stdint.h>
  22
  23 #include <bin/lttng-consumerd/health-consumerd.hpp>
  24 #include <common/common.hpp>
  25 #include <common/kernel-ctl/kernel-ctl.hpp>
  26 #include <common/sessiond-comm/sessiond-comm.hpp>
  27 #include <common/sessiond-comm/relayd.hpp>
  28 #include <common/compat/fcntl.hpp>
  29 #include <common/compat/endian.hpp>
  30 #include <common/pipe.hpp>
  31 #include <common/relayd/relayd.hpp>
  32 #include <common/utils.hpp>
  33 #include <common/consumer/consumer-stream.hpp>
  34 #include <common/index/index.hpp>
  35 #include <common/consumer/consumer-timer.hpp>
  36 #include <common/optional.hpp>
  37 #include <common/buffer-view.hpp>
  38 #include <common/consumer/consumer.hpp>
  39 #include <common/consumer/metadata-bucket.hpp>
  40
  41 #include "kernel-consumer.hpp"
  42
  43 extern struct lttng_consumer_global_data the_consumer_data;
  44 extern int consumer_poll_timeout;
  45
  46 /*
  47  * Take a snapshot for a specific fd
  48  *
  49  * Returns 0 on success, < 0 on error
  50  */
  51 int lttng_kconsumer_take_snapshot(struct lttng_consumer_stream *stream)
  52 {
  53         int ret = 0;
  54         int infd = stream->wait_fd;
  55
  56         ret = kernctl_snapshot(infd);
  57         /*
  58          * -EAGAIN is not an error, it just means that there is no data to
  59          *  be read.
  60          */
  61         if (ret != 0 && ret != -EAGAIN) {
  62                 PERROR("Getting sub-buffer snapshot.");
  63         }
  64
  65         return ret;
  66 }
  67
  68 /*
  69  * Sample consumed and produced positions for a specific fd.
  70  *
  71  * Returns 0 on success, < 0 on error.
  72  */
  73 int lttng_kconsumer_sample_snapshot_positions(
  74                 struct lttng_consumer_stream *stream)
  75 {
  76         LTTNG_ASSERT(stream);
  77
  78         return kernctl_snapshot_sample_positions(stream->wait_fd);
  79 }
  80
  81 /*
  82  * Get the produced position
  83  *
  84  * Returns 0 on success, < 0 on error
  85  */
  86 int lttng_kconsumer_get_produced_snapshot(struct lttng_consumer_stream *stream,
  87                 unsigned long *pos)
  88 {
  89         int ret;
  90         int infd = stream->wait_fd;
  91
  92         ret = kernctl_snapshot_get_produced(infd, pos);
  93         if (ret != 0) {
  94                 PERROR("kernctl_snapshot_get_produced");
  95         }
  96
  97         return ret;
  98 }
  99
 100 /*
 101  * Get the consumerd position
 102  *
 103  * Returns 0 on success, < 0 on error
 104  */
 105 int lttng_kconsumer_get_consumed_snapshot(struct lttng_consumer_stream *stream,
 106                 unsigned long *pos)
 107 {
 108         int ret;
 109         int infd = stream->wait_fd;
 110
 111         ret = kernctl_snapshot_get_consumed(infd, pos);
 112         if (ret != 0) {
 113                 PERROR("kernctl_snapshot_get_consumed");
 114         }
 115
 116         return ret;
 117 }
 118
 119 static
 120 int get_current_subbuf_addr(struct lttng_consumer_stream *stream,
 121                 const char **addr)
 122 {
 123         int ret;
 124         unsigned long mmap_offset;
 125         const char *mmap_base = (const char *) stream->mmap_base;
 126
 127         ret = kernctl_get_mmap_read_offset(stream->wait_fd, &mmap_offset);
 128         if (ret < 0) {
 129                 PERROR("Failed to get mmap read offset");
 130                 goto error;
 131         }
 132
 133         *addr = mmap_base + mmap_offset;
 134 error:
 135         return ret;
 136 }
 137
 138 /*
 139  * Take a snapshot of all the stream of a channel
 140  * RCU read-side lock must be held across this function to ensure existence of
 141  * channel.
 142  *
 143  * Returns 0 on success, < 0 on error
 144  */
 145 static int lttng_kconsumer_snapshot_channel(
 146                 struct lttng_consumer_channel *channel,
 147                 uint64_t key, char *path, uint64_t relayd_id,
 148                 uint64_t nb_packets_per_stream)
 149 {
 150         int ret;
 151         struct lttng_consumer_stream *stream;
 152
 153         DBG("Kernel consumer snapshot channel %" PRIu64, key);
 154
 155         /* Prevent channel modifications while we perform the snapshot.*/
 156         pthread_mutex_lock(&channel->lock);
 157
 158         rcu_read_lock();
 159
 160         /* Splice is not supported yet for channel snapshot. */
 161         if (channel->output != CONSUMER_CHANNEL_MMAP) {
 162                 ERR("Unsupported output type for channel \"%s\": mmap output is required to record a snapshot",
 163                                 channel->name);
 164                 ret = -1;
 165                 goto end;
 166         }
 167
 168         cds_list_for_each_entry(stream, &channel->streams.head, send_node) {
 169                 unsigned long consumed_pos, produced_pos;
 170
 171                 health_code_update();
 172
 173                 /*
 174                  * Lock stream because we are about to change its state.
 175                  */
 176                 pthread_mutex_lock(&stream->lock);
 177
 178                 LTTNG_ASSERT(channel->trace_chunk);
 179                 if (!lttng_trace_chunk_get(channel->trace_chunk)) {
 180                         /*
 181                          * Can't happen barring an internal error as the channel
 182                          * holds a reference to the trace chunk.
 183                          */
 184                         ERR("Failed to acquire reference to channel's trace chunk");
 185                         ret = -1;
 186                         goto end_unlock;
 187                 }
 188                 LTTNG_ASSERT(!stream->trace_chunk);
 189                 stream->trace_chunk = channel->trace_chunk;
 190
 191                 /*
 192                  * Assign the received relayd ID so we can use it for streaming. The streams
 193                  * are not visible to anyone so this is OK to change it.
 194                  */
 195                 stream->net_seq_idx = relayd_id;
 196                 channel->relayd_id = relayd_id;
 197                 if (relayd_id != (uint64_t) -1ULL) {
 198                         ret = consumer_send_relayd_stream(stream, path);
 199                         if (ret < 0) {
 200                                 ERR("sending stream to relayd");
 201                                 goto error_close_stream_output;
 202                         }
 203                 } else {
 204                         ret = consumer_stream_create_output_files(stream,
 205                                         false);
 206                         if (ret < 0) {
 207                                 goto error_close_stream_output;
 208                         }
 209                         DBG("Kernel consumer snapshot stream (%" PRIu64 ")",
 210                                         stream->key);
 211                 }
 212
 213                 ret = kernctl_buffer_flush_empty(stream->wait_fd);
 214                 if (ret < 0) {
 215                         /*
 216                          * Doing a buffer flush which does not take into
 217                          * account empty packets. This is not perfect
 218                          * for stream intersection, but required as a
 219                          * fall-back when "flush_empty" is not
 220                          * implemented by lttng-modules.
 221                          */
 222                         ret = kernctl_buffer_flush(stream->wait_fd);
 223                         if (ret < 0) {
 224                                 ERR("Failed to flush kernel stream");
 225                                 goto error_close_stream_output;
 226                         }
 227                         goto end_unlock;
 228                 }
 229
 230                 ret = lttng_kconsumer_take_snapshot(stream);
 231                 if (ret < 0) {
 232                         ERR("Taking kernel snapshot");
 233                         goto error_close_stream_output;
 234                 }
 235
 236                 ret = lttng_kconsumer_get_produced_snapshot(stream, &produced_pos);
 237                 if (ret < 0) {
 238                         ERR("Produced kernel snapshot position");
 239                         goto error_close_stream_output;
 240                 }
 241
 242                 ret = lttng_kconsumer_get_consumed_snapshot(stream, &consumed_pos);
 243                 if (ret < 0) {
 244                         ERR("Consumerd kernel snapshot position");
 245                         goto error_close_stream_output;
 246                 }
 247
 248                 consumed_pos = consumer_get_consume_start_pos(consumed_pos,
 249                                 produced_pos, nb_packets_per_stream,
 250                                 stream->max_sb_size);
 251
 252                 while ((long) (consumed_pos - produced_pos) < 0) {
 253                         ssize_t read_len;
 254                         unsigned long len, padded_len;
 255                         const char *subbuf_addr;
 256                         struct lttng_buffer_view subbuf_view;
 257
 258                         health_code_update();
 259                         DBG("Kernel consumer taking snapshot at pos %lu", consumed_pos);
 260
 261                         ret = kernctl_get_subbuf(stream->wait_fd, &consumed_pos);
 262                         if (ret < 0) {
 263                                 if (ret != -EAGAIN) {
 264                                         PERROR("kernctl_get_subbuf snapshot");
 265                                         goto error_close_stream_output;
 266                                 }
 267                                 DBG("Kernel consumer get subbuf failed. Skipping it.");
 268                                 consumed_pos += stream->max_sb_size;
 269                                 stream->chan->lost_packets++;
 270                                 continue;
 271                         }
 272
 273                         ret = kernctl_get_subbuf_size(stream->wait_fd, &len);
 274                         if (ret < 0) {
 275                                 ERR("Snapshot kernctl_get_subbuf_size");
 276                                 goto error_put_subbuf;
 277                         }
 278
 279                         ret = kernctl_get_padded_subbuf_size(stream->wait_fd, &padded_len);
 280                         if (ret < 0) {
 281                                 ERR("Snapshot kernctl_get_padded_subbuf_size");
 282                                 goto error_put_subbuf;
 283                         }
 284
 285                         ret = get_current_subbuf_addr(stream, &subbuf_addr);
 286                         if (ret) {
 287                                 goto error_put_subbuf;
 288                         }
 289
 290                         subbuf_view = lttng_buffer_view_init(
 291                                         subbuf_addr, 0, padded_len);
 292                         read_len = lttng_consumer_on_read_subbuffer_mmap(
 293                                         stream, &subbuf_view,
 294                                         padded_len - len);
 295                         /*
 296                          * We write the padded len in local tracefiles but the data len
 297                          * when using a relay. Display the error but continue processing
 298                          * to try to release the subbuffer.
 299                          */
 300                         if (relayd_id != (uint64_t) -1ULL) {
 301                                 if (read_len != len) {
 302                                         ERR("Error sending to the relay (ret: %zd != len: %lu)",
 303                                                         read_len, len);
 304                                 }
 305                         } else {
 306                                 if (read_len != padded_len) {
 307                                         ERR("Error writing to tracefile (ret: %zd != len: %lu)",
 308                                                         read_len, padded_len);
 309                                 }
 310                         }
 311
 312                         ret = kernctl_put_subbuf(stream->wait_fd);
 313                         if (ret < 0) {
 314                                 ERR("Snapshot kernctl_put_subbuf");
 315                                 goto error_close_stream_output;
 316                         }
 317                         consumed_pos += stream->max_sb_size;
 318                 }
 319
 320                 consumer_stream_close_output(stream);
 321                 pthread_mutex_unlock(&stream->lock);
 322         }
 323
 324         /* All good! */
 325         ret = 0;
 326         goto end;
 327
 328 error_put_subbuf:
 329         ret = kernctl_put_subbuf(stream->wait_fd);
 330         if (ret < 0) {
 331                 ERR("Snapshot kernctl_put_subbuf error path");
 332         }
 333 error_close_stream_output:
 334         consumer_stream_close_output(stream);
 335 end_unlock:
 336         pthread_mutex_unlock(&stream->lock);
 337 end:
 338         rcu_read_unlock();
 339         pthread_mutex_unlock(&channel->lock);
 340         return ret;
 341 }
 342
 343 /*
 344  * Read the whole metadata available for a snapshot.
 345  * RCU read-side lock must be held across this function to ensure existence of
 346  * metadata_channel.
 347  *
 348  * Returns 0 on success, < 0 on error
 349  */
 350 static int lttng_kconsumer_snapshot_metadata(
 351                 struct lttng_consumer_channel *metadata_channel,
 352                 uint64_t key, char *path, uint64_t relayd_id,
 353                 struct lttng_consumer_local_data *ctx)
 354 {
 355         int ret, use_relayd = 0;
 356         ssize_t ret_read;
 357         struct lttng_consumer_stream *metadata_stream;
 358
 359         LTTNG_ASSERT(ctx);
 360
 361         DBG("Kernel consumer snapshot metadata with key %" PRIu64 " at path %s",
 362                         key, path);
 363
 364         rcu_read_lock();
 365
 366         metadata_stream = metadata_channel->metadata_stream;
 367         LTTNG_ASSERT(metadata_stream);
 368
 369         metadata_stream->read_subbuffer_ops.lock(metadata_stream);
 370         LTTNG_ASSERT(metadata_channel->trace_chunk);
 371         LTTNG_ASSERT(metadata_stream->trace_chunk);
 372
 373         /* Flag once that we have a valid relayd for the stream. */
 374         if (relayd_id != (uint64_t) -1ULL) {
 375                 use_relayd = 1;
 376         }
 377
 378         if (use_relayd) {
 379                 ret = consumer_send_relayd_stream(metadata_stream, path);
 380                 if (ret < 0) {
 381                         goto error_snapshot;
 382                 }
 383         } else {
 384                 ret = consumer_stream_create_output_files(metadata_stream,
 385                                 false);
 386                 if (ret < 0) {
 387                         goto error_snapshot;
 388                 }
 389         }
 390
 391         do {
 392                 health_code_update();
 393
 394                 ret_read = lttng_consumer_read_subbuffer(metadata_stream, ctx, true);
 395                 if (ret_read < 0) {
 396                         ERR("Kernel snapshot reading metadata subbuffer (ret: %zd)",
 397                                         ret_read);
 398                         ret = ret_read;
 399                         goto error_snapshot;
 400                 }
 401         } while (ret_read > 0);
 402
 403         if (use_relayd) {
 404                 close_relayd_stream(metadata_stream);
 405                 metadata_stream->net_seq_idx = (uint64_t) -1ULL;
 406         } else {
 407                 if (metadata_stream->out_fd >= 0) {
 408                         ret = close(metadata_stream->out_fd);
 409                         if (ret < 0) {
 410                                 PERROR("Kernel consumer snapshot metadata close out_fd");
 411                                 /*
 412                                  * Don't go on error here since the snapshot was successful at this
 413                                  * point but somehow the close failed.
 414                                  */
 415                         }
 416                         metadata_stream->out_fd = -1;
 417                         lttng_trace_chunk_put(metadata_stream->trace_chunk);
 418                         metadata_stream->trace_chunk = NULL;
 419                 }
 420         }
 421
 422         ret = 0;
 423 error_snapshot:
 424         metadata_stream->read_subbuffer_ops.unlock(metadata_stream);
 425         consumer_stream_destroy(metadata_stream, NULL);
 426         metadata_channel->metadata_stream = NULL;
 427         rcu_read_unlock();
 428         return ret;
 429 }
 430
 431 /*
 432  * Receive command from session daemon and process it.
 433  *
 434  * Return 1 on success else a negative value or 0.
 435  */
 436 int lttng_kconsumer_recv_cmd(struct lttng_consumer_local_data *ctx,
 437                 int sock, struct pollfd *consumer_sockpoll)
 438 {
 439         int ret_func;
 440         enum lttcomm_return_code ret_code = LTTCOMM_CONSUMERD_SUCCESS;
 441         struct lttcomm_consumer_msg msg;
 442
 443         health_code_update();
 444
 445         {
 446                 ssize_t ret_recv;
 447
 448                 ret_recv = lttcomm_recv_unix_sock(sock, &msg, sizeof(msg));
 449                 if (ret_recv != sizeof(msg)) {
 450                         if (ret_recv > 0) {
 451                                 lttng_consumer_send_error(ctx,
 452                                                 LTTCOMM_CONSUMERD_ERROR_RECV_CMD);
 453                                 ret_recv = -1;
 454                         }
 455                         return ret_recv;
 456                 }
 457         }
 458
 459         health_code_update();
 460
 461         /* Deprecated command */
 462         LTTNG_ASSERT(msg.cmd_type != LTTNG_CONSUMER_STOP);
 463
 464         health_code_update();
 465
 466         /* relayd needs RCU read-side protection */
 467         rcu_read_lock();
 468
 469         switch (msg.cmd_type) {
 470         case LTTNG_CONSUMER_ADD_RELAYD_SOCKET:
 471         {
 472                 uint32_t major = msg.u.relayd_sock.major;
 473                 uint32_t minor = msg.u.relayd_sock.minor;
 474                 enum lttcomm_sock_proto protocol = (enum lttcomm_sock_proto)
 475                                 msg.u.relayd_sock.relayd_socket_protocol;
 476
 477                 /* Session daemon status message are handled in the following call. */
 478                 consumer_add_relayd_socket(msg.u.relayd_sock.net_index,
 479                                 msg.u.relayd_sock.type, ctx, sock,
 480                                 consumer_sockpoll, msg.u.relayd_sock.session_id,
 481                                 msg.u.relayd_sock.relayd_session_id, major,
 482                                 minor, protocol);
 483                 goto end_nosignal;
 484         }
 485         case LTTNG_CONSUMER_ADD_CHANNEL:
 486         {
 487                 struct lttng_consumer_channel *new_channel;
 488                 int ret_send_status, ret_add_channel = 0;
 489                 const uint64_t chunk_id = msg.u.channel.chunk_id.value;
 490
 491                 health_code_update();
 492
 493                 /* First send a status message before receiving the fds. */
 494                 ret_send_status = consumer_send_status_msg(sock, ret_code);
 495                 if (ret_send_status < 0) {
 496                         /* Somehow, the session daemon is not responding anymore. */
 497                         goto error_fatal;
 498                 }
 499
 500                 health_code_update();
 501
 502                 DBG("consumer_add_channel %" PRIu64, msg.u.channel.channel_key);
 503                 new_channel = consumer_allocate_channel(msg.u.channel.channel_key,
 504                                 msg.u.channel.session_id,
 505                                 msg.u.channel.chunk_id.is_set ?
 506                                                 &chunk_id : NULL,
 507                                 msg.u.channel.pathname,
 508                                 msg.u.channel.name,
 509                                 msg.u.channel.relayd_id, msg.u.channel.output,
 510                                 msg.u.channel.tracefile_size,
 511                                 msg.u.channel.tracefile_count, 0,
 512                                 msg.u.channel.monitor,
 513                                 msg.u.channel.live_timer_interval,
 514                                 msg.u.channel.is_live,
 515                                 NULL, NULL);
 516                 if (new_channel == NULL) {
 517                         lttng_consumer_send_error(ctx, LTTCOMM_CONSUMERD_OUTFD_ERROR);
 518                         goto end_nosignal;
 519                 }
 520                 new_channel->nb_init_stream_left = msg.u.channel.nb_init_streams;
 521                 switch (msg.u.channel.output) {
 522                 case LTTNG_EVENT_SPLICE:
 523                         new_channel->output = CONSUMER_CHANNEL_SPLICE;
 524                         break;
 525                 case LTTNG_EVENT_MMAP:
 526                         new_channel->output = CONSUMER_CHANNEL_MMAP;
 527                         break;
 528                 default:
 529                         ERR("Channel output unknown %d", msg.u.channel.output);
 530                         goto end_nosignal;
 531                 }
 532
 533                 /* Translate and save channel type. */
 534                 switch (msg.u.channel.type) {
 535                 case CONSUMER_CHANNEL_TYPE_DATA:
 536                 case CONSUMER_CHANNEL_TYPE_METADATA:
 537                         new_channel->type = (consumer_channel_type) msg.u.channel.type;
 538                         break;
 539                 default:
 540                         abort();
 541                         goto end_nosignal;
 542                 };
 543
 544                 health_code_update();
 545
 546                 if (ctx->on_recv_channel != NULL) {
 547                         int ret_recv_channel =
 548                                         ctx->on_recv_channel(new_channel);
 549                         if (ret_recv_channel == 0) {
 550                                 ret_add_channel = consumer_add_channel(
 551                                                 new_channel, ctx);
 552                         } else if (ret_recv_channel < 0) {
 553                                 goto end_nosignal;
 554                         }
 555                 } else {
 556                         ret_add_channel =
 557                                         consumer_add_channel(new_channel, ctx);
 558                 }
 559                 if (msg.u.channel.type == CONSUMER_CHANNEL_TYPE_DATA &&
 560                                 !ret_add_channel) {
 561                         int monitor_start_ret;
 562
 563                         DBG("Consumer starting monitor timer");
 564                         consumer_timer_live_start(new_channel,
 565                                         msg.u.channel.live_timer_interval);
 566                         monitor_start_ret = consumer_timer_monitor_start(
 567                                         new_channel,
 568                                         msg.u.channel.monitor_timer_interval);
 569                         if (monitor_start_ret < 0) {
 570                                 ERR("Starting channel monitoring timer failed");
 571                                 goto end_nosignal;
 572                         }
 573                 }
 574
 575                 health_code_update();
 576
 577                 /* If we received an error in add_channel, we need to report it. */
 578                 if (ret_add_channel < 0) {
 579                         ret_send_status = consumer_send_status_msg(
 580                                         sock, ret_add_channel);
 581                         if (ret_send_status < 0) {
 582                                 goto error_fatal;
 583                         }
 584                         goto end_nosignal;
 585                 }
 586
 587                 goto end_nosignal;
 588         }
 589         case LTTNG_CONSUMER_ADD_STREAM:
 590         {
 591                 int fd;
 592                 struct lttng_pipe *stream_pipe;
 593                 struct lttng_consumer_stream *new_stream;
 594                 struct lttng_consumer_channel *channel;
 595                 int alloc_ret = 0;
 596                 int ret_send_status, ret_poll, ret_get_max_subbuf_size;
 597                 ssize_t ret_pipe_write, ret_recv;
 598
 599                 /*
 600                  * Get stream's channel reference. Needed when adding the stream to the
 601                  * global hash table.
 602                  */
 603                 channel = consumer_find_channel(msg.u.stream.channel_key);
 604                 if (!channel) {
 605                         /*
 606                          * We could not find the channel. Can happen if cpu hotplug
 607                          * happens while tearing down.
 608                          */
 609                         ERR("Unable to find channel key %" PRIu64, msg.u.stream.channel_key);
 610                         ret_code = LTTCOMM_CONSUMERD_CHAN_NOT_FOUND;
 611                 }
 612
 613                 health_code_update();
 614
 615                 /* First send a status message before receiving the fds. */
 616                 ret_send_status = consumer_send_status_msg(sock, ret_code);
 617                 if (ret_send_status < 0) {
 618                         /* Somehow, the session daemon is not responding anymore. */
 619                         goto error_add_stream_fatal;
 620                 }
 621
 622                 health_code_update();
 623
 624                 if (ret_code != LTTCOMM_CONSUMERD_SUCCESS) {
 625                         /* Channel was not found. */
 626                         goto error_add_stream_nosignal;
 627                 }
 628
 629                 /* Blocking call */
 630                 health_poll_entry();
 631                 ret_poll = lttng_consumer_poll_socket(consumer_sockpoll);
 632                 health_poll_exit();
 633                 if (ret_poll) {
 634                         goto error_add_stream_fatal;
 635                 }
 636
 637                 health_code_update();
 638
 639                 /* Get stream file descriptor from socket */
 640                 ret_recv = lttcomm_recv_fds_unix_sock(sock, &fd, 1);
 641                 if (ret_recv != sizeof(fd)) {
 642                         lttng_consumer_send_error(ctx, LTTCOMM_CONSUMERD_ERROR_RECV_FD);
 643                         ret_func = ret_recv;
 644                         goto end;
 645                 }
 646
 647                 health_code_update();
 648
 649                 /*
 650                  * Send status code to session daemon only if the recv works. If the
 651                  * above recv() failed, the session daemon is notified through the
 652                  * error socket and the teardown is eventually done.
 653                  */
 654                 ret_send_status = consumer_send_status_msg(sock, ret_code);
 655                 if (ret_send_status < 0) {
 656                         /* Somehow, the session daemon is not responding anymore. */
 657                         goto error_add_stream_nosignal;
 658                 }
 659
 660                 health_code_update();
 661
 662                 pthread_mutex_lock(&channel->lock);
 663                 new_stream = consumer_stream_create(
 664                                 channel,
 665                                 channel->key,
 666                                 fd,
 667                                 channel->name,
 668                                 channel->relayd_id,
 669                                 channel->session_id,
 670                                 channel->trace_chunk,
 671                                 msg.u.stream.cpu,
 672                                 &alloc_ret,
 673                                 channel->type,
 674                                 channel->monitor);
 675                 if (new_stream == NULL) {
 676                         switch (alloc_ret) {
 677                         case -ENOMEM:
 678                         case -EINVAL:
 679                         default:
 680                                 lttng_consumer_send_error(ctx, LTTCOMM_CONSUMERD_OUTFD_ERROR);
 681                                 break;
 682                         }
 683                         pthread_mutex_unlock(&channel->lock);
 684                         goto error_add_stream_nosignal;
 685                 }
 686
 687                 new_stream->wait_fd = fd;
 688                 ret_get_max_subbuf_size = kernctl_get_max_subbuf_size(
 689                                 new_stream->wait_fd, &new_stream->max_sb_size);
 690                 if (ret_get_max_subbuf_size < 0) {
 691                         pthread_mutex_unlock(&channel->lock);
 692                         ERR("Failed to get kernel maximal subbuffer size");
 693                         goto error_add_stream_nosignal;
 694                 }
 695
 696                 consumer_stream_update_channel_attributes(new_stream,
 697                                 channel);
 698
 699                 /*
 700                  * We've just assigned the channel to the stream so increment the
 701                  * refcount right now. We don't need to increment the refcount for
 702                  * streams in no monitor because we handle manually the cleanup of
 703                  * those. It is very important to make sure there is NO prior
 704                  * consumer_del_stream() calls or else the refcount will be unbalanced.
 705                  */
 706                 if (channel->monitor) {
 707                         uatomic_inc(&new_stream->chan->refcount);
 708                 }
 709
 710                 /*
 711                  * The buffer flush is done on the session daemon side for the kernel
 712                  * so no need for the stream "hangup_flush_done" variable to be
 713                  * tracked. This is important for a kernel stream since we don't rely
 714                  * on the flush state of the stream to read data. It's not the case for
 715                  * user space tracing.
 716                  */
 717                 new_stream->hangup_flush_done = 0;
 718
 719                 health_code_update();
 720
 721                 pthread_mutex_lock(&new_stream->lock);
 722                 if (ctx->on_recv_stream) {
 723                         int ret_recv_stream = ctx->on_recv_stream(new_stream);
 724                         if (ret_recv_stream < 0) {
 725                                 pthread_mutex_unlock(&new_stream->lock);
 726                                 pthread_mutex_unlock(&channel->lock);
 727                                 consumer_stream_free(new_stream);
 728                                 goto error_add_stream_nosignal;
 729                         }
 730                 }
 731                 health_code_update();
 732
 733                 if (new_stream->metadata_flag) {
 734                         channel->metadata_stream = new_stream;
 735                 }
 736
 737                 /* Do not monitor this stream. */
 738                 if (!channel->monitor) {
 739                         DBG("Kernel consumer add stream %s in no monitor mode with "
 740                                         "relayd id %" PRIu64, new_stream->name,
 741                                         new_stream->net_seq_idx);
 742                         cds_list_add(&new_stream->send_node, &channel->streams.head);
 743                         pthread_mutex_unlock(&new_stream->lock);
 744                         pthread_mutex_unlock(&channel->lock);
 745                         goto end_add_stream;
 746                 }
 747
 748                 /* Send stream to relayd if the stream has an ID. */
 749                 if (new_stream->net_seq_idx != (uint64_t) -1ULL) {
 750                         int ret_send_relayd_stream;
 751
 752                         ret_send_relayd_stream = consumer_send_relayd_stream(
 753                                         new_stream, new_stream->chan->pathname);
 754                         if (ret_send_relayd_stream < 0) {
 755                                 pthread_mutex_unlock(&new_stream->lock);
 756                                 pthread_mutex_unlock(&channel->lock);
 757                                 consumer_stream_free(new_stream);
 758                                 goto error_add_stream_nosignal;
 759                         }
 760
 761                         /*
 762                          * If adding an extra stream to an already
 763                          * existing channel (e.g. cpu hotplug), we need
 764                          * to send the "streams_sent" command to relayd.
 765                          */
 766                         if (channel->streams_sent_to_relayd) {
 767                                 int ret_send_relayd_streams_sent;
 768
 769                                 ret_send_relayd_streams_sent =
 770                                                 consumer_send_relayd_streams_sent(
 771                                                                 new_stream->net_seq_idx);
 772                                 if (ret_send_relayd_streams_sent < 0) {
 773                                         pthread_mutex_unlock(&new_stream->lock);
 774                                         pthread_mutex_unlock(&channel->lock);
 775                                         goto error_add_stream_nosignal;
 776                                 }
 777                         }
 778                 }
 779                 pthread_mutex_unlock(&new_stream->lock);
 780                 pthread_mutex_unlock(&channel->lock);
 781
 782                 /* Get the right pipe where the stream will be sent. */
 783                 if (new_stream->metadata_flag) {
 784                         consumer_add_metadata_stream(new_stream);
 785                         stream_pipe = ctx->consumer_metadata_pipe;
 786                 } else {
 787                         consumer_add_data_stream(new_stream);
 788                         stream_pipe = ctx->consumer_data_pipe;
 789                 }
 790
 791                 /* Visible to other threads */
 792                 new_stream->globally_visible = 1;
 793
 794                 health_code_update();
 795
 796                 ret_pipe_write = lttng_pipe_write(
 797                                 stream_pipe, &new_stream, sizeof(new_stream));
 798                 if (ret_pipe_write < 0) {
 799                         ERR("Consumer write %s stream to pipe %d",
 800                                         new_stream->metadata_flag ? "metadata" : "data",
 801                                         lttng_pipe_get_writefd(stream_pipe));
 802                         if (new_stream->metadata_flag) {
 803                                 consumer_del_stream_for_metadata(new_stream);
 804                         } else {
 805                                 consumer_del_stream_for_data(new_stream);
 806                         }
 807                         goto error_add_stream_nosignal;
 808                 }
 809
 810                 DBG("Kernel consumer ADD_STREAM %s (fd: %d) %s with relayd id %" PRIu64,
 811                                 new_stream->name, fd, new_stream->chan->pathname, new_stream->relayd_stream_id);
 812 end_add_stream:
 813                 break;
 814 error_add_stream_nosignal:
 815                 goto end_nosignal;
 816 error_add_stream_fatal:
 817                 goto error_fatal;
 818         }
 819         case LTTNG_CONSUMER_STREAMS_SENT:
 820         {
 821                 struct lttng_consumer_channel *channel;
 822                 int ret_send_status;
 823
 824                 /*
 825                  * Get stream's channel reference. Needed when adding the stream to the
 826                  * global hash table.
 827                  */
 828                 channel = consumer_find_channel(msg.u.sent_streams.channel_key);
 829                 if (!channel) {
 830                         /*
 831                          * We could not find the channel. Can happen if cpu hotplug
 832                          * happens while tearing down.
 833                          */
 834                         ERR("Unable to find channel key %" PRIu64,
 835                                         msg.u.sent_streams.channel_key);
 836                         ret_code = LTTCOMM_CONSUMERD_CHAN_NOT_FOUND;
 837                 }
 838
 839                 health_code_update();
 840
 841                 /*
 842                  * Send status code to session daemon.
 843                  */
 844                 ret_send_status = consumer_send_status_msg(sock, ret_code);
 845                 if (ret_send_status < 0 ||
 846                                 ret_code != LTTCOMM_CONSUMERD_SUCCESS) {
 847                         /* Somehow, the session daemon is not responding anymore. */
 848                         goto error_streams_sent_nosignal;
 849                 }
 850
 851                 health_code_update();
 852
 853                 /*
 854                  * We should not send this message if we don't monitor the
 855                  * streams in this channel.
 856                  */
 857                 if (!channel->monitor) {
 858                         goto end_error_streams_sent;
 859                 }
 860
 861                 health_code_update();
 862                 /* Send stream to relayd if the stream has an ID. */
 863                 if (msg.u.sent_streams.net_seq_idx != (uint64_t) -1ULL) {
 864                         int ret_send_relay_streams;
 865
 866                         ret_send_relay_streams = consumer_send_relayd_streams_sent(
 867                                         msg.u.sent_streams.net_seq_idx);
 868                         if (ret_send_relay_streams < 0) {
 869                                 goto error_streams_sent_nosignal;
 870                         }
 871                         channel->streams_sent_to_relayd = true;
 872                 }
 873 end_error_streams_sent:
 874                 break;
 875 error_streams_sent_nosignal:
 876                 goto end_nosignal;
 877         }
 878         case LTTNG_CONSUMER_UPDATE_STREAM:
 879         {
 880                 rcu_read_unlock();
 881                 return -ENOSYS;
 882         }
 883         case LTTNG_CONSUMER_DESTROY_RELAYD:
 884         {
 885                 uint64_t index = msg.u.destroy_relayd.net_seq_idx;
 886                 struct consumer_relayd_sock_pair *relayd;
 887                 int ret_send_status;
 888
 889                 DBG("Kernel consumer destroying relayd %" PRIu64, index);
 890
 891                 /* Get relayd reference if exists. */
 892                 relayd = consumer_find_relayd(index);
 893                 if (relayd == NULL) {
 894                         DBG("Unable to find relayd %" PRIu64, index);
 895                         ret_code = LTTCOMM_CONSUMERD_RELAYD_FAIL;
 896                 }
 897
 898                 /*
 899                  * Each relayd socket pair has a refcount of stream attached to it
 900                  * which tells if the relayd is still active or not depending on the
 901                  * refcount value.
 902                  *
 903                  * This will set the destroy flag of the relayd object and destroy it
 904                  * if the refcount reaches zero when called.
 905                  *
 906                  * The destroy can happen either here or when a stream fd hangs up.
 907                  */
 908                 if (relayd) {
 909                         consumer_flag_relayd_for_destroy(relayd);
 910                 }
 911
 912                 health_code_update();
 913
 914                 ret_send_status = consumer_send_status_msg(sock, ret_code);
 915                 if (ret_send_status < 0) {
 916                         /* Somehow, the session daemon is not responding anymore. */
 917                         goto error_fatal;
 918                 }
 919
 920                 goto end_nosignal;
 921         }
 922         case LTTNG_CONSUMER_DATA_PENDING:
 923         {
 924                 int32_t ret_data_pending;
 925                 uint64_t id = msg.u.data_pending.session_id;
 926                 ssize_t ret_send;
 927
 928                 DBG("Kernel consumer data pending command for id %" PRIu64, id);
 929
 930                 ret_data_pending = consumer_data_pending(id);
 931
 932                 health_code_update();
 933
 934                 /* Send back returned value to session daemon */
 935                 ret_send = lttcomm_send_unix_sock(sock, &ret_data_pending,
 936                                 sizeof(ret_data_pending));
 937                 if (ret_send < 0) {
 938                         PERROR("send data pending ret code");
 939                         goto error_fatal;
 940                 }
 941
 942                 /*
 943                  * No need to send back a status message since the data pending
 944                  * returned value is the response.
 945                  */
 946                 break;
 947         }
 948         case LTTNG_CONSUMER_SNAPSHOT_CHANNEL:
 949         {
 950                 struct lttng_consumer_channel *channel;
 951                 uint64_t key = msg.u.snapshot_channel.key;
 952                 int ret_send_status;
 953
 954                 channel = consumer_find_channel(key);
 955                 if (!channel) {
 956                         ERR("Channel %" PRIu64 " not found", key);
 957                         ret_code = LTTCOMM_CONSUMERD_CHAN_NOT_FOUND;
 958                 } else {
 959                         if (msg.u.snapshot_channel.metadata == 1) {
 960                                 int ret_snapshot;
 961
 962                                 ret_snapshot = lttng_kconsumer_snapshot_metadata(
 963                                                 channel, key,
 964                                                 msg.u.snapshot_channel.pathname,
 965                                                 msg.u.snapshot_channel.relayd_id,
 966                                                 ctx);
 967                                 if (ret_snapshot < 0) {
 968                                         ERR("Snapshot metadata failed");
 969                                         ret_code = LTTCOMM_CONSUMERD_SNAPSHOT_FAILED;
 970                                 }
 971                         } else {
 972                                 int ret_snapshot;
 973
 974                                 ret_snapshot = lttng_kconsumer_snapshot_channel(
 975                                                 channel, key,
 976                                                 msg.u.snapshot_channel.pathname,
 977                                                 msg.u.snapshot_channel.relayd_id,
 978                                                 msg.u.snapshot_channel
 979                                                                 .nb_packets_per_stream);
 980                                 if (ret_snapshot < 0) {
 981                                         ERR("Snapshot channel failed");
 982                                         ret_code = LTTCOMM_CONSUMERD_SNAPSHOT_FAILED;
 983                                 }
 984                         }
 985                 }
 986                 health_code_update();
 987
 988                 ret_send_status = consumer_send_status_msg(sock, ret_code);
 989                 if (ret_send_status < 0) {
 990                         /* Somehow, the session daemon is not responding anymore. */
 991                         goto end_nosignal;
 992                 }
 993                 break;
 994         }
 995         case LTTNG_CONSUMER_DESTROY_CHANNEL:
 996         {
 997                 uint64_t key = msg.u.destroy_channel.key;
 998                 struct lttng_consumer_channel *channel;
 999                 int ret_send_status;
1000
1001                 channel = consumer_find_channel(key);
1002                 if (!channel) {
1003                         ERR("Kernel consumer destroy channel %" PRIu64 " not found", key);
1004                         ret_code = LTTCOMM_CONSUMERD_CHAN_NOT_FOUND;
1005                 }
1006
1007                 health_code_update();
1008
1009                 ret_send_status = consumer_send_status_msg(sock, ret_code);
1010                 if (ret_send_status < 0) {
1011                         /* Somehow, the session daemon is not responding anymore. */
1012                         goto end_destroy_channel;
1013                 }
1014
1015                 health_code_update();
1016
1017                 /* Stop right now if no channel was found. */
1018                 if (!channel) {
1019                         goto end_destroy_channel;
1020                 }
1021
1022                 /*
1023                  * This command should ONLY be issued for channel with streams set in
1024                  * no monitor mode.
1025                  */
1026                 LTTNG_ASSERT(!channel->monitor);
1027
1028                 /*
1029                  * The refcount should ALWAYS be 0 in the case of a channel in no
1030                  * monitor mode.
1031                  */
1032                 LTTNG_ASSERT(!uatomic_sub_return(&channel->refcount, 1));
1033
1034                 consumer_del_channel(channel);
1035 end_destroy_channel:
1036                 goto end_nosignal;
1037         }
1038         case LTTNG_CONSUMER_DISCARDED_EVENTS:
1039         {
1040                 ssize_t ret;
1041                 uint64_t count;
1042                 struct lttng_consumer_channel *channel;
1043                 uint64_t id = msg.u.discarded_events.session_id;
1044                 uint64_t key = msg.u.discarded_events.channel_key;
1045
1046                 DBG("Kernel consumer discarded events command for session id %"
1047                                 PRIu64 ", channel key %" PRIu64, id, key);
1048
1049                 channel = consumer_find_channel(key);
1050                 if (!channel) {
1051                         ERR("Kernel consumer discarded events channel %"
1052                                         PRIu64 " not found", key);
1053                         count = 0;
1054                 } else {
1055                         count = channel->discarded_events;
1056                 }
1057
1058                 health_code_update();
1059
1060                 /* Send back returned value to session daemon */
1061                 ret = lttcomm_send_unix_sock(sock, &count, sizeof(count));
1062                 if (ret < 0) {
1063                         PERROR("send discarded events");
1064                         goto error_fatal;
1065                 }
1066
1067                 break;
1068         }
1069         case LTTNG_CONSUMER_LOST_PACKETS:
1070         {
1071                 ssize_t ret;
1072                 uint64_t count;
1073                 struct lttng_consumer_channel *channel;
1074                 uint64_t id = msg.u.lost_packets.session_id;
1075                 uint64_t key = msg.u.lost_packets.channel_key;
1076
1077                 DBG("Kernel consumer lost packets command for session id %"
1078                                 PRIu64 ", channel key %" PRIu64, id, key);
1079
1080                 channel = consumer_find_channel(key);
1081                 if (!channel) {
1082                         ERR("Kernel consumer lost packets channel %"
1083                                         PRIu64 " not found", key);
1084                         count = 0;
1085                 } else {
1086                         count = channel->lost_packets;
1087                 }
1088
1089                 health_code_update();
1090
1091                 /* Send back returned value to session daemon */
1092                 ret = lttcomm_send_unix_sock(sock, &count, sizeof(count));
1093                 if (ret < 0) {
1094                         PERROR("send lost packets");
1095                         goto error_fatal;
1096                 }
1097
1098                 break;
1099         }
1100         case LTTNG_CONSUMER_SET_CHANNEL_MONITOR_PIPE:
1101         {
1102                 int channel_monitor_pipe;
1103                 int ret_send_status, ret_set_channel_monitor_pipe;
1104                 ssize_t ret_recv;
1105
1106                 ret_code = LTTCOMM_CONSUMERD_SUCCESS;
1107                 /* Successfully received the command's type. */
1108                 ret_send_status = consumer_send_status_msg(sock, ret_code);
1109                 if (ret_send_status < 0) {
1110                         goto error_fatal;
1111                 }
1112
1113                 ret_recv = lttcomm_recv_fds_unix_sock(
1114                                 sock, &channel_monitor_pipe, 1);
1115                 if (ret_recv != sizeof(channel_monitor_pipe)) {
1116                         ERR("Failed to receive channel monitor pipe");
1117                         goto error_fatal;
1118                 }
1119
1120                 DBG("Received channel monitor pipe (%d)", channel_monitor_pipe);
1121                 ret_set_channel_monitor_pipe =
1122                                 consumer_timer_thread_set_channel_monitor_pipe(
1123                                                 channel_monitor_pipe);
1124                 if (!ret_set_channel_monitor_pipe) {
1125                         int flags;
1126                         int ret_fcntl;
1127
1128                         ret_code = LTTCOMM_CONSUMERD_SUCCESS;
1129                         /* Set the pipe as non-blocking. */
1130                         ret_fcntl = fcntl(channel_monitor_pipe, F_GETFL, 0);
1131                         if (ret_fcntl == -1) {
1132                                 PERROR("fcntl get flags of the channel monitoring pipe");
1133                                 goto error_fatal;
1134                         }
1135                         flags = ret_fcntl;
1136
1137                         ret_fcntl = fcntl(channel_monitor_pipe, F_SETFL,
1138                                         flags | O_NONBLOCK);
1139                         if (ret_fcntl == -1) {
1140                                 PERROR("fcntl set O_NONBLOCK flag of the channel monitoring pipe");
1141                                 goto error_fatal;
1142                         }
1143                         DBG("Channel monitor pipe set as non-blocking");
1144                 } else {
1145                         ret_code = LTTCOMM_CONSUMERD_ALREADY_SET;
1146                 }
1147                 ret_send_status = consumer_send_status_msg(sock, ret_code);
1148                 if (ret_send_status < 0) {
1149                         goto error_fatal;
1150                 }
1151                 break;
1152         }
1153         case LTTNG_CONSUMER_ROTATE_CHANNEL:
1154         {
1155                 struct lttng_consumer_channel *channel;
1156                 uint64_t key = msg.u.rotate_channel.key;
1157                 int ret_send_status;
1158
1159                 DBG("Consumer rotate channel %" PRIu64, key);
1160
1161                 channel = consumer_find_channel(key);
1162                 if (!channel) {
1163                         ERR("Channel %" PRIu64 " not found", key);
1164                         ret_code = LTTCOMM_CONSUMERD_CHAN_NOT_FOUND;
1165                 } else {
1166                         /*
1167                          * Sample the rotate position of all the streams in this channel.
1168                          */
1169                         int ret_rotate_channel;
1170
1171                         ret_rotate_channel = lttng_consumer_rotate_channel(
1172                                         channel, key,
1173                                         msg.u.rotate_channel.relayd_id);
1174                         if (ret_rotate_channel < 0) {
1175                                 ERR("Rotate channel failed");
1176                                 ret_code = LTTCOMM_CONSUMERD_ROTATION_FAIL;
1177                         }
1178
1179                         health_code_update();
1180                 }
1181
1182                 ret_send_status = consumer_send_status_msg(sock, ret_code);
1183                 if (ret_send_status < 0) {
1184                         /* Somehow, the session daemon is not responding anymore. */
1185                         goto error_rotate_channel;
1186                 }
1187                 if (channel) {
1188                         /* Rotate the streams that are ready right now. */
1189                         int ret_rotate;
1190
1191                         ret_rotate = lttng_consumer_rotate_ready_streams(
1192                                         channel, key);
1193                         if (ret_rotate < 0) {
1194                                 ERR("Rotate ready streams failed");
1195                         }
1196                 }
1197                 break;
1198 error_rotate_channel:
1199                 goto end_nosignal;
1200         }
1201         case LTTNG_CONSUMER_CLEAR_CHANNEL:
1202         {
1203                 struct lttng_consumer_channel *channel;
1204                 uint64_t key = msg.u.clear_channel.key;
1205                 int ret_send_status;
1206
1207                 channel = consumer_find_channel(key);
1208                 if (!channel) {
1209                         DBG("Channel %" PRIu64 " not found", key);
1210                         ret_code = LTTCOMM_CONSUMERD_CHAN_NOT_FOUND;
1211                 } else {
1212                         int ret_clear_channel;
1213
1214                         ret_clear_channel =
1215                                         lttng_consumer_clear_channel(channel);
1216                         if (ret_clear_channel) {
1217                                 ERR("Clear channel failed");
1218                                 ret_code = (lttcomm_return_code) ret_clear_channel;
1219                         }
1220
1221                         health_code_update();
1222                 }
1223
1224                 ret_send_status = consumer_send_status_msg(sock, ret_code);
1225                 if (ret_send_status < 0) {
1226                         /* Somehow, the session daemon is not responding anymore. */
1227                         goto end_nosignal;
1228                 }
1229
1230                 break;
1231         }
1232         case LTTNG_CONSUMER_INIT:
1233         {
1234                 int ret_send_status;
1235                 lttng_uuid sessiond_uuid;
1236
1237                 std::copy(std::begin(msg.u.init.sessiond_uuid), std::end(msg.u.init.sessiond_uuid),
1238                                 sessiond_uuid.begin());
1239
1240                 ret_code = lttng_consumer_init_command(ctx,
1241                                 sessiond_uuid);
1242                 health_code_update();
1243                 ret_send_status = consumer_send_status_msg(sock, ret_code);
1244                 if (ret_send_status < 0) {
1245                         /* Somehow, the session daemon is not responding anymore. */
1246                         goto end_nosignal;
1247                 }
1248                 break;
1249         }
1250         case LTTNG_CONSUMER_CREATE_TRACE_CHUNK:
1251         {
1252                 const struct lttng_credentials credentials = {
1253                         .uid = LTTNG_OPTIONAL_INIT_VALUE(msg.u.create_trace_chunk.credentials.value.uid),
1254                         .gid = LTTNG_OPTIONAL_INIT_VALUE(msg.u.create_trace_chunk.credentials.value.gid),
1255                 };
1256                 const bool is_local_trace =
1257                                 !msg.u.create_trace_chunk.relayd_id.is_set;
1258                 const uint64_t relayd_id =
1259                                 msg.u.create_trace_chunk.relayd_id.value;
1260                 const char *chunk_override_name =
1261                                 *msg.u.create_trace_chunk.override_name ?
1262                                         msg.u.create_trace_chunk.override_name :
1263                                         NULL;
1264                 struct lttng_directory_handle *chunk_directory_handle = NULL;
1265
1266                 /*
1267                  * The session daemon will only provide a chunk directory file
1268                  * descriptor for local traces.
1269                  */
1270                 if (is_local_trace) {
1271                         int chunk_dirfd;
1272                         int ret_send_status;
1273                         ssize_t ret_recv;
1274
1275                         /* Acnowledge the reception of the command. */
1276                         ret_send_status = consumer_send_status_msg(
1277                                         sock, LTTCOMM_CONSUMERD_SUCCESS);
1278                         if (ret_send_status < 0) {
1279                                 /* Somehow, the session daemon is not responding anymore. */
1280                                 goto end_nosignal;
1281                         }
1282
1283                         ret_recv = lttcomm_recv_fds_unix_sock(
1284                                         sock, &chunk_dirfd, 1);
1285                         if (ret_recv != sizeof(chunk_dirfd)) {
1286                                 ERR("Failed to receive trace chunk directory file descriptor");
1287                                 goto error_fatal;
1288                         }
1289
1290                         DBG("Received trace chunk directory fd (%d)",
1291                                         chunk_dirfd);
1292                         chunk_directory_handle = lttng_directory_handle_create_from_dirfd(
1293                                         chunk_dirfd);
1294                         if (!chunk_directory_handle) {
1295                                 ERR("Failed to initialize chunk directory handle from directory file descriptor");
1296                                 if (close(chunk_dirfd)) {
1297                                         PERROR("Failed to close chunk directory file descriptor");
1298                                 }
1299                                 goto error_fatal;
1300                         }
1301                 }
1302
1303                 ret_code = lttng_consumer_create_trace_chunk(
1304                                 !is_local_trace ? &relayd_id : NULL,
1305                                 msg.u.create_trace_chunk.session_id,
1306                                 msg.u.create_trace_chunk.chunk_id,
1307                                 (time_t) msg.u.create_trace_chunk
1308                                                 .creation_timestamp,
1309                                 chunk_override_name,
1310                                 msg.u.create_trace_chunk.credentials.is_set ?
1311                                                 &credentials :
1312                                                 NULL,
1313                                 chunk_directory_handle);
1314                 lttng_directory_handle_put(chunk_directory_handle);
1315                 goto end_msg_sessiond;
1316         }
1317         case LTTNG_CONSUMER_CLOSE_TRACE_CHUNK:
1318         {
1319                 enum lttng_trace_chunk_command_type close_command =
1320                                 (lttng_trace_chunk_command_type) msg.u.close_trace_chunk.close_command.value;
1321                 const uint64_t relayd_id =
1322                                 msg.u.close_trace_chunk.relayd_id.value;
1323                 struct lttcomm_consumer_close_trace_chunk_reply reply;
1324                 char path[LTTNG_PATH_MAX];
1325                 ssize_t ret_send;
1326
1327                 ret_code = lttng_consumer_close_trace_chunk(
1328                                 msg.u.close_trace_chunk.relayd_id.is_set ?
1329                                                 &relayd_id :
1330                                                 NULL,
1331                                 msg.u.close_trace_chunk.session_id,
1332                                 msg.u.close_trace_chunk.chunk_id,
1333                                 (time_t) msg.u.close_trace_chunk.close_timestamp,
1334                                 msg.u.close_trace_chunk.close_command.is_set ?
1335                                                 &close_command :
1336                                                 NULL, path);
1337                 reply.ret_code = ret_code;
1338                 reply.path_length = strlen(path) + 1;
1339                 ret_send = lttcomm_send_unix_sock(sock, &reply, sizeof(reply));
1340                 if (ret_send != sizeof(reply)) {
1341                         goto error_fatal;
1342                 }
1343                 ret_send = lttcomm_send_unix_sock(
1344                                 sock, path, reply.path_length);
1345                 if (ret_send != reply.path_length) {
1346                         goto error_fatal;
1347                 }
1348                 goto end_nosignal;
1349         }
1350         case LTTNG_CONSUMER_TRACE_CHUNK_EXISTS:
1351         {
1352                 const uint64_t relayd_id =
1353                                 msg.u.trace_chunk_exists.relayd_id.value;
1354
1355                 ret_code = lttng_consumer_trace_chunk_exists(
1356                                 msg.u.trace_chunk_exists.relayd_id.is_set ?
1357                                                 &relayd_id : NULL,
1358                                 msg.u.trace_chunk_exists.session_id,
1359                                 msg.u.trace_chunk_exists.chunk_id);
1360                 goto end_msg_sessiond;
1361         }
1362         case LTTNG_CONSUMER_OPEN_CHANNEL_PACKETS:
1363         {
1364                 const uint64_t key = msg.u.open_channel_packets.key;
1365                 struct lttng_consumer_channel *channel =
1366                                 consumer_find_channel(key);
1367
1368                 if (channel) {
1369                         pthread_mutex_lock(&channel->lock);
1370                         ret_code = lttng_consumer_open_channel_packets(channel);
1371                         pthread_mutex_unlock(&channel->lock);
1372                 } else {
1373                         WARN("Channel %" PRIu64 " not found", key);
1374                         ret_code = LTTCOMM_CONSUMERD_CHAN_NOT_FOUND;
1375                 }
1376
1377                 health_code_update();
1378                 goto end_msg_sessiond;
1379         }
1380         default:
1381                 goto end_nosignal;
1382         }
1383
1384 end_nosignal:
1385         /*
1386          * Return 1 to indicate success since the 0 value can be a socket
1387          * shutdown during the recv() or send() call.
1388          */
1389         ret_func = 1;
1390         goto end;
1391 error_fatal:
1392         /* This will issue a consumer stop. */
1393         ret_func = -1;
1394         goto end;
1395 end_msg_sessiond:
1396         /*
1397          * The returned value here is not useful since either way we'll return 1 to
1398          * the caller because the session daemon socket management is done
1399          * elsewhere. Returning a negative code or 0 will shutdown the consumer.
1400          */
1401         {
1402                 int ret_send_status;
1403
1404                 ret_send_status = consumer_send_status_msg(sock, ret_code);
1405                 if (ret_send_status < 0) {
1406                         goto error_fatal;
1407                 }
1408         }
1409
1410         ret_func = 1;
1411
1412 end:
1413         health_code_update();
1414         rcu_read_unlock();
1415         return ret_func;
1416 }
1417
1418 /*
1419  * Sync metadata meaning request them to the session daemon and snapshot to the
1420  * metadata thread can consumer them.
1421  *
1422  * Metadata stream lock MUST be acquired.
1423  */
1424 enum sync_metadata_status lttng_kconsumer_sync_metadata(
1425                 struct lttng_consumer_stream *metadata)
1426 {
1427         int ret;
1428         enum sync_metadata_status status;
1429
1430         LTTNG_ASSERT(metadata);
1431
1432         ret = kernctl_buffer_flush(metadata->wait_fd);
1433         if (ret < 0) {
1434                 ERR("Failed to flush kernel stream");
1435                 status = SYNC_METADATA_STATUS_ERROR;
1436                 goto end;
1437         }
1438
1439         ret = kernctl_snapshot(metadata->wait_fd);
1440         if (ret < 0) {
1441                 if (errno == EAGAIN) {
1442                         /* No new metadata, exit. */
1443                         DBG("Sync metadata, no new kernel metadata");
1444                         status = SYNC_METADATA_STATUS_NO_DATA;
1445                 } else {
1446                         ERR("Sync metadata, taking kernel snapshot failed.");
1447                         status = SYNC_METADATA_STATUS_ERROR;
1448                 }
1449         } else {
1450                 status = SYNC_METADATA_STATUS_NEW_DATA;
1451         }
1452
1453 end:
1454         return status;
1455 }
1456
1457 static
1458 int extract_common_subbuffer_info(struct lttng_consumer_stream *stream,
1459                 struct stream_subbuffer *subbuf)
1460 {
1461         int ret;
1462
1463         ret = kernctl_get_subbuf_size(
1464                         stream->wait_fd, &subbuf->info.data.subbuf_size);
1465         if (ret) {
1466                 goto end;
1467         }
1468
1469         ret = kernctl_get_padded_subbuf_size(
1470                         stream->wait_fd, &subbuf->info.data.padded_subbuf_size);
1471         if (ret) {
1472                 goto end;
1473         }
1474
1475 end:
1476         return ret;
1477 }
1478
1479 static
1480 int extract_metadata_subbuffer_info(struct lttng_consumer_stream *stream,
1481                 struct stream_subbuffer *subbuf)
1482 {
1483         int ret;
1484
1485         ret = extract_common_subbuffer_info(stream, subbuf);
1486         if (ret) {
1487                 goto end;
1488         }
1489
1490         ret = kernctl_get_metadata_version(
1491                         stream->wait_fd, &subbuf->info.metadata.version);
1492         if (ret) {
1493                 goto end;
1494         }
1495
1496 end:
1497         return ret;
1498 }
1499
1500 static
1501 int extract_data_subbuffer_info(struct lttng_consumer_stream *stream,
1502                 struct stream_subbuffer *subbuf)
1503 {
1504         int ret;
1505
1506         ret = extract_common_subbuffer_info(stream, subbuf);
1507         if (ret) {
1508                 goto end;
1509         }
1510
1511         ret = kernctl_get_packet_size(
1512                         stream->wait_fd, &subbuf->info.data.packet_size);
1513         if (ret < 0) {
1514                 PERROR("Failed to get sub-buffer packet size");
1515                 goto end;
1516         }
1517
1518         ret = kernctl_get_content_size(
1519                         stream->wait_fd, &subbuf->info.data.content_size);
1520         if (ret < 0) {
1521                 PERROR("Failed to get sub-buffer content size");
1522                 goto end;
1523         }
1524
1525         ret = kernctl_get_timestamp_begin(
1526                         stream->wait_fd, &subbuf->info.data.timestamp_begin);
1527         if (ret < 0) {
1528                 PERROR("Failed to get sub-buffer begin timestamp");
1529                 goto end;
1530         }
1531
1532         ret = kernctl_get_timestamp_end(
1533                         stream->wait_fd, &subbuf->info.data.timestamp_end);
1534         if (ret < 0) {
1535                 PERROR("Failed to get sub-buffer end timestamp");
1536                 goto end;
1537         }
1538
1539         ret = kernctl_get_events_discarded(
1540                         stream->wait_fd, &subbuf->info.data.events_discarded);
1541         if (ret) {
1542                 PERROR("Failed to get sub-buffer events discarded count");
1543                 goto end;
1544         }
1545
1546         ret = kernctl_get_sequence_number(stream->wait_fd,
1547                         &subbuf->info.data.sequence_number.value);
1548         if (ret) {
1549                 /* May not be supported by older LTTng-modules. */
1550                 if (ret != -ENOTTY) {
1551                         PERROR("Failed to get sub-buffer sequence number");
1552                         goto end;
1553                 }
1554         } else {
1555                 subbuf->info.data.sequence_number.is_set = true;
1556         }
1557
1558         ret = kernctl_get_stream_id(
1559                         stream->wait_fd, &subbuf->info.data.stream_id);
1560         if (ret < 0) {
1561                 PERROR("Failed to get stream id");
1562                 goto end;
1563         }
1564
1565         ret = kernctl_get_instance_id(stream->wait_fd,
1566                         &subbuf->info.data.stream_instance_id.value);
1567         if (ret) {
1568                 /* May not be supported by older LTTng-modules. */
1569                 if (ret != -ENOTTY) {
1570                         PERROR("Failed to get stream instance id");
1571                         goto end;
1572                 }
1573         } else {
1574                 subbuf->info.data.stream_instance_id.is_set = true;
1575         }
1576 end:
1577         return ret;
1578 }
1579
1580 static
1581 enum get_next_subbuffer_status get_subbuffer_common(
1582                 struct lttng_consumer_stream *stream,
1583                 struct stream_subbuffer *subbuffer)
1584 {
1585         int ret;
1586         enum get_next_subbuffer_status status;
1587
1588         ret = kernctl_get_next_subbuf(stream->wait_fd);
1589         switch (ret) {
1590         case 0:
1591                 status = GET_NEXT_SUBBUFFER_STATUS_OK;
1592                 break;
1593         case -ENODATA:
1594                 case -EAGAIN:
1595                 /*
1596                  * The caller only expects -ENODATA when there is no data to
1597                  * read, but the kernel tracer returns -EAGAIN when there is
1598                  * currently no data for a non-finalized stream, and -ENODATA
1599                  * when there is no data for a finalized stream. Those can be
1600                  * combined into a -ENODATA return value.
1601                  */
1602                 status = GET_NEXT_SUBBUFFER_STATUS_NO_DATA;
1603                 goto end;
1604         default:
1605                 status = GET_NEXT_SUBBUFFER_STATUS_ERROR;
1606                 goto end;
1607         }
1608
1609         ret = stream->read_subbuffer_ops.extract_subbuffer_info(
1610                 stream, subbuffer);
1611         if (ret) {
1612                 status = GET_NEXT_SUBBUFFER_STATUS_ERROR;
1613         }
1614 end:
1615         return status;
1616 }
1617
1618 static
1619 enum get_next_subbuffer_status get_next_subbuffer_splice(
1620                 struct lttng_consumer_stream *stream,
1621                 struct stream_subbuffer *subbuffer)
1622 {
1623         const enum get_next_subbuffer_status status =
1624                         get_subbuffer_common(stream, subbuffer);
1625
1626         if (status != GET_NEXT_SUBBUFFER_STATUS_OK) {
1627                 goto end;
1628         }
1629
1630         subbuffer->buffer.fd = stream->wait_fd;
1631 end:
1632         return status;
1633 }
1634
1635 static
1636 enum get_next_subbuffer_status get_next_subbuffer_mmap(
1637                 struct lttng_consumer_stream *stream,
1638                 struct stream_subbuffer *subbuffer)
1639 {
1640         int ret;
1641         enum get_next_subbuffer_status status;
1642         const char *addr;
1643
1644         status = get_subbuffer_common(stream, subbuffer);
1645         if (status != GET_NEXT_SUBBUFFER_STATUS_OK) {
1646                 goto end;
1647         }
1648
1649         ret = get_current_subbuf_addr(stream, &addr);
1650         if (ret) {
1651                 status = GET_NEXT_SUBBUFFER_STATUS_ERROR;
1652                 goto end;
1653         }
1654
1655         subbuffer->buffer.buffer = lttng_buffer_view_init(
1656                         addr, 0, subbuffer->info.data.padded_subbuf_size);
1657 end:
1658         return status;
1659 }
1660
1661 static
1662 enum get_next_subbuffer_status get_next_subbuffer_metadata_check(struct lttng_consumer_stream *stream,
1663                 struct stream_subbuffer *subbuffer)
1664 {
1665         int ret;
1666         const char *addr;
1667         bool coherent;
1668         enum get_next_subbuffer_status status;
1669
1670         ret = kernctl_get_next_subbuf_metadata_check(stream->wait_fd,
1671                         &coherent);
1672         if (ret) {
1673                 goto end;
1674         }
1675
1676         ret = stream->read_subbuffer_ops.extract_subbuffer_info(
1677                         stream, subbuffer);
1678         if (ret) {
1679                 goto end;
1680         }
1681
1682         LTTNG_OPTIONAL_SET(&subbuffer->info.metadata.coherent, coherent);
1683
1684         ret = get_current_subbuf_addr(stream, &addr);
1685         if (ret) {
1686                 goto end;
1687         }
1688
1689         subbuffer->buffer.buffer = lttng_buffer_view_init(
1690                         addr, 0, subbuffer->info.data.padded_subbuf_size);
1691         DBG("Got metadata packet with padded_subbuf_size = %lu, coherent = %s",
1692                         subbuffer->info.metadata.padded_subbuf_size,
1693                         coherent ? "true" : "false");
1694 end:
1695         /*
1696          * The caller only expects -ENODATA when there is no data to read, but
1697          * the kernel tracer returns -EAGAIN when there is currently no data
1698          * for a non-finalized stream, and -ENODATA when there is no data for a
1699          * finalized stream. Those can be combined into a -ENODATA return value.
1700          */
1701         switch (ret) {
1702         case 0:
1703                 status = GET_NEXT_SUBBUFFER_STATUS_OK;
1704                 break;
1705         case -ENODATA:
1706         case -EAGAIN:
1707                 /*
1708                  * The caller only expects -ENODATA when there is no data to
1709                  * read, but the kernel tracer returns -EAGAIN when there is
1710                  * currently no data for a non-finalized stream, and -ENODATA
1711                  * when there is no data for a finalized stream. Those can be
1712                  * combined into a -ENODATA return value.
1713                  */
1714                 status = GET_NEXT_SUBBUFFER_STATUS_NO_DATA;
1715                 break;
1716         default:
1717                 status = GET_NEXT_SUBBUFFER_STATUS_ERROR;
1718                 break;
1719         }
1720
1721         return status;
1722 }
1723
1724 static
1725 int put_next_subbuffer(struct lttng_consumer_stream *stream,
1726                 struct stream_subbuffer *subbuffer __attribute__((unused)))
1727 {
1728         const int ret = kernctl_put_next_subbuf(stream->wait_fd);
1729
1730         if (ret) {
1731                 if (ret == -EFAULT) {
1732                         PERROR("Error in unreserving sub buffer");
1733                 } else if (ret == -EIO) {
1734                         /* Should never happen with newer LTTng versions */
1735                         PERROR("Reader has been pushed by the writer, last sub-buffer corrupted");
1736                 }
1737         }
1738
1739         return ret;
1740 }
1741
1742 static
1743 bool is_get_next_check_metadata_available(int tracer_fd)
1744 {
1745         const int ret = kernctl_get_next_subbuf_metadata_check(tracer_fd, NULL);
1746         const bool available = ret != -ENOTTY;
1747
1748         if (ret == 0) {
1749                 /* get succeeded, make sure to put the subbuffer. */
1750                 kernctl_put_subbuf(tracer_fd);
1751         }
1752
1753         return available;
1754 }
1755
1756 static
1757 int signal_metadata(struct lttng_consumer_stream *stream,
1758                 struct lttng_consumer_local_data *ctx __attribute__((unused)))
1759 {
1760         ASSERT_LOCKED(stream->metadata_rdv_lock);
1761         return pthread_cond_broadcast(&stream->metadata_rdv) ? -errno : 0;
1762 }
1763
1764 static
1765 int lttng_kconsumer_set_stream_ops(
1766                 struct lttng_consumer_stream *stream)
1767 {
1768         int ret = 0;
1769
1770         if (stream->metadata_flag && stream->chan->is_live) {
1771                 DBG("Attempting to enable metadata bucketization for live consumers");
1772                 if (is_get_next_check_metadata_available(stream->wait_fd)) {
1773                         DBG("Kernel tracer supports get_next_subbuffer_metadata_check, metadata will be accumulated until a coherent state is reached");
1774                         stream->read_subbuffer_ops.get_next_subbuffer =
1775                                         get_next_subbuffer_metadata_check;
1776                         ret = consumer_stream_enable_metadata_bucketization(
1777                                         stream);
1778                         if (ret) {
1779                                 goto end;
1780                         }
1781                 } else {
1782                         /*
1783                          * The kernel tracer version is too old to indicate
1784                          * when the metadata stream has reached a "coherent"
1785                          * (parseable) point.
1786                          *
1787                          * This means that a live viewer may see an incoherent
1788                          * sequence of metadata and fail to parse it.
1789                          */
1790                         WARN("Kernel tracer does not support get_next_subbuffer_metadata_check which may cause live clients to fail to parse the metadata stream");
1791                         metadata_bucket_destroy(stream->metadata_bucket);
1792                         stream->metadata_bucket = NULL;
1793                 }
1794
1795                 stream->read_subbuffer_ops.on_sleep = signal_metadata;
1796         }
1797
1798         if (!stream->read_subbuffer_ops.get_next_subbuffer) {
1799                 if (stream->chan->output == CONSUMER_CHANNEL_MMAP) {
1800                         stream->read_subbuffer_ops.get_next_subbuffer =
1801                                         get_next_subbuffer_mmap;
1802                 } else {
1803                         stream->read_subbuffer_ops.get_next_subbuffer =
1804                                         get_next_subbuffer_splice;
1805                 }
1806         }
1807
1808         if (stream->metadata_flag) {
1809                 stream->read_subbuffer_ops.extract_subbuffer_info =
1810                                 extract_metadata_subbuffer_info;
1811         } else {
1812                 stream->read_subbuffer_ops.extract_subbuffer_info =
1813                                 extract_data_subbuffer_info;
1814                 if (stream->chan->is_live) {
1815                         stream->read_subbuffer_ops.send_live_beacon =
1816                                         consumer_flush_kernel_index;
1817                 }
1818         }
1819
1820         stream->read_subbuffer_ops.put_next_subbuffer = put_next_subbuffer;
1821 end:
1822         return ret;
1823 }
1824
1825 int lttng_kconsumer_on_recv_stream(struct lttng_consumer_stream *stream)
1826 {
1827         int ret;
1828
1829         LTTNG_ASSERT(stream);
1830
1831         /*
1832          * Don't create anything if this is set for streaming or if there is
1833          * no current trace chunk on the parent channel.
1834          */
1835         if (stream->net_seq_idx == (uint64_t) -1ULL && stream->chan->monitor &&
1836                         stream->chan->trace_chunk) {
1837                 ret = consumer_stream_create_output_files(stream, true);
1838                 if (ret) {
1839                         goto error;
1840                 }
1841         }
1842
1843         if (stream->output == LTTNG_EVENT_MMAP) {
1844                 /* get the len of the mmap region */
1845                 unsigned long mmap_len;
1846
1847                 ret = kernctl_get_mmap_len(stream->wait_fd, &mmap_len);
1848                 if (ret != 0) {
1849                         PERROR("kernctl_get_mmap_len");
1850                         goto error_close_fd;
1851                 }
1852                 stream->mmap_len = (size_t) mmap_len;
1853
1854                 stream->mmap_base = mmap(NULL, stream->mmap_len, PROT_READ,
1855                                 MAP_PRIVATE, stream->wait_fd, 0);
1856                 if (stream->mmap_base == MAP_FAILED) {
1857                         PERROR("Error mmaping");
1858                         ret = -1;
1859                         goto error_close_fd;
1860                 }
1861         }
1862
1863         ret = lttng_kconsumer_set_stream_ops(stream);
1864         if (ret) {
1865                 goto error_close_fd;
1866         }
1867
1868         /* we return 0 to let the library handle the FD internally */
1869         return 0;
1870
1871 error_close_fd:
1872         if (stream->out_fd >= 0) {
1873                 int err;
1874
1875                 err = close(stream->out_fd);
1876                 LTTNG_ASSERT(!err);
1877                 stream->out_fd = -1;
1878         }
1879 error:
1880         return ret;
1881 }
1882
1883 /*
1884  * Check if data is still being extracted from the buffers for a specific
1885  * stream. Consumer data lock MUST be acquired before calling this function
1886  * and the stream lock.
1887  *
1888  * Return 1 if the traced data are still getting read else 0 meaning that the
1889  * data is available for trace viewer reading.
1890  */
1891 int lttng_kconsumer_data_pending(struct lttng_consumer_stream *stream)
1892 {
1893         int ret;
1894
1895         LTTNG_ASSERT(stream);
1896
1897         if (stream->endpoint_status != CONSUMER_ENDPOINT_ACTIVE) {
1898                 ret = 0;
1899                 goto end;
1900         }
1901
1902         ret = kernctl_get_next_subbuf(stream->wait_fd);
1903         if (ret == 0) {
1904                 /* There is still data so let's put back this subbuffer. */
1905                 ret = kernctl_put_subbuf(stream->wait_fd);
1906                 LTTNG_ASSERT(ret == 0);
1907                 ret = 1;   /* Data is pending */
1908                 goto end;
1909         }
1910
1911         /* Data is NOT pending and ready to be read. */
1912         ret = 0;
1913
1914 end:
1915         return ret;
1916 }