migration/colo.c

   1 /*
   2  * COarse-grain LOck-stepping Virtual Machines for Non-stop Service (COLO)
   3  * (a.k.a. Fault Tolerance or Continuous Replication)
   4  *
   5  * Copyright (c) 2016 HUAWEI TECHNOLOGIES CO., LTD.
   6  * Copyright (c) 2016 FUJITSU LIMITED
   7  * Copyright (c) 2016 Intel Corporation
   8  *
   9  * This work is licensed under the terms of the GNU GPL, version 2 or
  10  * later.  See the COPYING file in the top-level directory.
  11  */
  12
  13 #include "qemu/osdep.h"
  14 #include "sysemu/sysemu.h"
  15 #include "qapi/error.h"
  16 #include "qapi/qapi-commands-migration.h"
  17 #include "qemu-file-channel.h"
  18 #include "migration.h"
  19 #include "qemu-file.h"
  20 #include "savevm.h"
  21 #include "migration/colo.h"
  22 #include "block.h"
  23 #include "io/channel-buffer.h"
  24 #include "trace.h"
  25 #include "qemu/error-report.h"
  26 #include "migration/failover.h"
  27 #include "replication.h"
  28 #include "net/colo-compare.h"
  29 #include "net/colo.h"
  30 #include "block/block.h"
  31 #include "qapi/qapi-events-migration.h"
  32 #include "qapi/qmp/qerror.h"
  33 #include "sysemu/cpus.h"
  34 #include "net/filter.h"
  35
  36 static bool vmstate_loading;
  37 static Notifier packets_compare_notifier;
  38
  39 #define COLO_BUFFER_BASE_SIZE (4 * 1024 * 1024)
  40
  41 bool migration_in_colo_state(void)
  42 {
  43     MigrationState *s = migrate_get_current();
  44
  45     return (s->state == MIGRATION_STATUS_COLO);
  46 }
  47
  48 bool migration_incoming_in_colo_state(void)
  49 {
  50     MigrationIncomingState *mis = migration_incoming_get_current();
  51
  52     return mis && (mis->state == MIGRATION_STATUS_COLO);
  53 }
  54
  55 static bool colo_runstate_is_stopped(void)
  56 {
  57     return runstate_check(RUN_STATE_COLO) || !runstate_is_running();
  58 }
  59
  60 static void secondary_vm_do_failover(void)
  61 {
  62 /* COLO needs enable block-replication */
  63 #ifdef CONFIG_REPLICATION
  64     int old_state;
  65     MigrationIncomingState *mis = migration_incoming_get_current();
  66     Error *local_err = NULL;
  67
  68     /* Can not do failover during the process of VM's loading VMstate, Or
  69      * it will break the secondary VM.
  70      */
  71     if (vmstate_loading) {
  72         old_state = failover_set_state(FAILOVER_STATUS_ACTIVE,
  73                         FAILOVER_STATUS_RELAUNCH);
  74         if (old_state != FAILOVER_STATUS_ACTIVE) {
  75             error_report("Unknown error while do failover for secondary VM,"
  76                          "old_state: %s", FailoverStatus_str(old_state));
  77         }
  78         return;
  79     }
  80
  81     migrate_set_state(&mis->state, MIGRATION_STATUS_COLO,
  82                       MIGRATION_STATUS_COMPLETED);
  83
  84     replication_stop_all(true, &local_err);
  85     if (local_err) {
  86         error_report_err(local_err);
  87     }
  88
  89     /* Notify all filters of all NIC to do checkpoint */
  90     colo_notify_filters_event(COLO_EVENT_FAILOVER, &local_err);
  91     if (local_err) {
  92         error_report_err(local_err);
  93     }
  94
  95     if (!autostart) {
  96         error_report("\"-S\" qemu option will be ignored in secondary side");
  97         /* recover runstate to normal migration finish state */
  98         autostart = true;
  99     }
 100     /*
 101      * Make sure COLO incoming thread not block in recv or send,
 102      * If mis->from_src_file and mis->to_src_file use the same fd,
 103      * The second shutdown() will return -1, we ignore this value,
 104      * It is harmless.
 105      */
 106     if (mis->from_src_file) {
 107         qemu_file_shutdown(mis->from_src_file);
 108     }
 109     if (mis->to_src_file) {
 110         qemu_file_shutdown(mis->to_src_file);
 111     }
 112
 113     old_state = failover_set_state(FAILOVER_STATUS_ACTIVE,
 114                                    FAILOVER_STATUS_COMPLETED);
 115     if (old_state != FAILOVER_STATUS_ACTIVE) {
 116         error_report("Incorrect state (%s) while doing failover for "
 117                      "secondary VM", FailoverStatus_str(old_state));
 118         return;
 119     }
 120     /* Notify COLO incoming thread that failover work is finished */
 121     qemu_sem_post(&mis->colo_incoming_sem);
 122     /* For Secondary VM, jump to incoming co */
 123     if (mis->migration_incoming_co) {
 124         qemu_coroutine_enter(mis->migration_incoming_co);
 125     }
 126 #else
 127     abort();
 128 #endif
 129 }
 130
 131 static void primary_vm_do_failover(void)
 132 {
 133 #ifdef CONFIG_REPLICATION
 134     MigrationState *s = migrate_get_current();
 135     int old_state;
 136     Error *local_err = NULL;
 137
 138     migrate_set_state(&s->state, MIGRATION_STATUS_COLO,
 139                       MIGRATION_STATUS_COMPLETED);
 140     /*
 141      * kick COLO thread which might wait at
 142      * qemu_sem_wait(&s->colo_checkpoint_sem).
 143      */
 144     colo_checkpoint_notify(migrate_get_current());
 145
 146     /*
 147      * Wake up COLO thread which may blocked in recv() or send(),
 148      * The s->rp_state.from_dst_file and s->to_dst_file may use the
 149      * same fd, but we still shutdown the fd for twice, it is harmless.
 150      */
 151     if (s->to_dst_file) {
 152         qemu_file_shutdown(s->to_dst_file);
 153     }
 154     if (s->rp_state.from_dst_file) {
 155         qemu_file_shutdown(s->rp_state.from_dst_file);
 156     }
 157
 158     old_state = failover_set_state(FAILOVER_STATUS_ACTIVE,
 159                                    FAILOVER_STATUS_COMPLETED);
 160     if (old_state != FAILOVER_STATUS_ACTIVE) {
 161         error_report("Incorrect state (%s) while doing failover for Primary VM",
 162                      FailoverStatus_str(old_state));
 163         return;
 164     }
 165
 166     replication_stop_all(true, &local_err);
 167     if (local_err) {
 168         error_report_err(local_err);
 169         local_err = NULL;
 170     }
 171
 172     /* Notify COLO thread that failover work is finished */
 173     qemu_sem_post(&s->colo_exit_sem);
 174 #else
 175     abort();
 176 #endif
 177 }
 178
 179 COLOMode get_colo_mode(void)
 180 {
 181     if (migration_in_colo_state()) {
 182         return COLO_MODE_PRIMARY;
 183     } else if (migration_incoming_in_colo_state()) {
 184         return COLO_MODE_SECONDARY;
 185     } else {
 186         return COLO_MODE_NONE;
 187     }
 188 }
 189
 190 void colo_do_failover(MigrationState *s)
 191 {
 192     /* Make sure VM stopped while failover happened. */
 193     if (!colo_runstate_is_stopped()) {
 194         vm_stop_force_state(RUN_STATE_COLO);
 195     }
 196
 197     if (get_colo_mode() == COLO_MODE_PRIMARY) {
 198         primary_vm_do_failover();
 199     } else {
 200         secondary_vm_do_failover();
 201     }
 202 }
 203
 204 void qmp_xen_set_replication(bool enable, bool primary,
 205                              bool has_failover, bool failover,
 206                              Error **errp)
 207 {
 208 #ifdef CONFIG_REPLICATION
 209     ReplicationMode mode = primary ?
 210                            REPLICATION_MODE_PRIMARY :
 211                            REPLICATION_MODE_SECONDARY;
 212
 213     if (has_failover && enable) {
 214         error_setg(errp, "Parameter 'failover' is only for"
 215                    " stopping replication");
 216         return;
 217     }
 218
 219     if (enable) {
 220         replication_start_all(mode, errp);
 221     } else {
 222         if (!has_failover) {
 223             failover = NULL;
 224         }
 225         replication_stop_all(failover, failover ? NULL : errp);
 226     }
 227 #else
 228     abort();
 229 #endif
 230 }
 231
 232 ReplicationStatus *qmp_query_xen_replication_status(Error **errp)
 233 {
 234 #ifdef CONFIG_REPLICATION
 235     Error *err = NULL;
 236     ReplicationStatus *s = g_new0(ReplicationStatus, 1);
 237
 238     replication_get_error_all(&err);
 239     if (err) {
 240         s->error = true;
 241         s->has_desc = true;
 242         s->desc = g_strdup(error_get_pretty(err));
 243     } else {
 244         s->error = false;
 245     }
 246
 247     error_free(err);
 248     return s;
 249 #else
 250     abort();
 251 #endif
 252 }
 253
 254 void qmp_xen_colo_do_checkpoint(Error **errp)
 255 {
 256 #ifdef CONFIG_REPLICATION
 257     replication_do_checkpoint_all(errp);
 258 #else
 259     abort();
 260 #endif
 261 }
 262
 263 COLOStatus *qmp_query_colo_status(Error **errp)
 264 {
 265     COLOStatus *s = g_new0(COLOStatus, 1);
 266
 267     s->mode = get_colo_mode();
 268
 269     switch (failover_get_state()) {
 270     case FAILOVER_STATUS_NONE:
 271         s->reason = COLO_EXIT_REASON_NONE;
 272         break;
 273     case FAILOVER_STATUS_REQUIRE:
 274         s->reason = COLO_EXIT_REASON_REQUEST;
 275         break;
 276     default:
 277         s->reason = COLO_EXIT_REASON_ERROR;
 278     }
 279
 280     return s;
 281 }
 282
 283 static void colo_send_message(QEMUFile *f, COLOMessage msg,
 284                               Error **errp)
 285 {
 286     int ret;
 287
 288     if (msg >= COLO_MESSAGE__MAX) {
 289         error_setg(errp, "%s: Invalid message", __func__);
 290         return;
 291     }
 292     qemu_put_be32(f, msg);
 293     qemu_fflush(f);
 294
 295     ret = qemu_file_get_error(f);
 296     if (ret < 0) {
 297         error_setg_errno(errp, -ret, "Can't send COLO message");
 298     }
 299     trace_colo_send_message(COLOMessage_str(msg));
 300 }
 301
 302 static void colo_send_message_value(QEMUFile *f, COLOMessage msg,
 303                                     uint64_t value, Error **errp)
 304 {
 305     Error *local_err = NULL;
 306     int ret;
 307
 308     colo_send_message(f, msg, &local_err);
 309     if (local_err) {
 310         error_propagate(errp, local_err);
 311         return;
 312     }
 313     qemu_put_be64(f, value);
 314     qemu_fflush(f);
 315
 316     ret = qemu_file_get_error(f);
 317     if (ret < 0) {
 318         error_setg_errno(errp, -ret, "Failed to send value for message:%s",
 319                          COLOMessage_str(msg));
 320     }
 321 }
 322
 323 static COLOMessage colo_receive_message(QEMUFile *f, Error **errp)
 324 {
 325     COLOMessage msg;
 326     int ret;
 327
 328     msg = qemu_get_be32(f);
 329     ret = qemu_file_get_error(f);
 330     if (ret < 0) {
 331         error_setg_errno(errp, -ret, "Can't receive COLO message");
 332         return msg;
 333     }
 334     if (msg >= COLO_MESSAGE__MAX) {
 335         error_setg(errp, "%s: Invalid message", __func__);
 336         return msg;
 337     }
 338     trace_colo_receive_message(COLOMessage_str(msg));
 339     return msg;
 340 }
 341
 342 static void colo_receive_check_message(QEMUFile *f, COLOMessage expect_msg,
 343                                        Error **errp)
 344 {
 345     COLOMessage msg;
 346     Error *local_err = NULL;
 347
 348     msg = colo_receive_message(f, &local_err);
 349     if (local_err) {
 350         error_propagate(errp, local_err);
 351         return;
 352     }
 353     if (msg != expect_msg) {
 354         error_setg(errp, "Unexpected COLO message %d, expected %d",
 355                           msg, expect_msg);
 356     }
 357 }
 358
 359 static uint64_t colo_receive_message_value(QEMUFile *f, uint32_t expect_msg,
 360                                            Error **errp)
 361 {
 362     Error *local_err = NULL;
 363     uint64_t value;
 364     int ret;
 365
 366     colo_receive_check_message(f, expect_msg, &local_err);
 367     if (local_err) {
 368         error_propagate(errp, local_err);
 369         return 0;
 370     }
 371
 372     value = qemu_get_be64(f);
 373     ret = qemu_file_get_error(f);
 374     if (ret < 0) {
 375         error_setg_errno(errp, -ret, "Failed to get value for COLO message: %s",
 376                          COLOMessage_str(expect_msg));
 377     }
 378     return value;
 379 }
 380
 381 static int colo_do_checkpoint_transaction(MigrationState *s,
 382                                           QIOChannelBuffer *bioc,
 383                                           QEMUFile *fb)
 384 {
 385     Error *local_err = NULL;
 386     int ret = -1;
 387
 388     colo_send_message(s->to_dst_file, COLO_MESSAGE_CHECKPOINT_REQUEST,
 389                       &local_err);
 390     if (local_err) {
 391         goto out;
 392     }
 393
 394     colo_receive_check_message(s->rp_state.from_dst_file,
 395                     COLO_MESSAGE_CHECKPOINT_REPLY, &local_err);
 396     if (local_err) {
 397         goto out;
 398     }
 399     /* Reset channel-buffer directly */
 400     qio_channel_io_seek(QIO_CHANNEL(bioc), 0, 0, NULL);
 401     bioc->usage = 0;
 402
 403     qemu_mutex_lock_iothread();
 404     if (failover_get_state() != FAILOVER_STATUS_NONE) {
 405         qemu_mutex_unlock_iothread();
 406         goto out;
 407     }
 408     vm_stop_force_state(RUN_STATE_COLO);
 409     qemu_mutex_unlock_iothread();
 410     trace_colo_vm_state_change("run", "stop");
 411     /*
 412      * Failover request bh could be called after vm_stop_force_state(),
 413      * So we need check failover_request_is_active() again.
 414      */
 415     if (failover_get_state() != FAILOVER_STATUS_NONE) {
 416         goto out;
 417     }
 418
 419     colo_notify_compares_event(NULL, COLO_EVENT_CHECKPOINT, &local_err);
 420     if (local_err) {
 421         goto out;
 422     }
 423
 424     /* Disable block migration */
 425     migrate_set_block_enabled(false, &local_err);
 426     qemu_mutex_lock_iothread();
 427
 428 #ifdef CONFIG_REPLICATION
 429     replication_do_checkpoint_all(&local_err);
 430     if (local_err) {
 431         qemu_mutex_unlock_iothread();
 432         goto out;
 433     }
 434 #else
 435         abort();
 436 #endif
 437
 438     colo_send_message(s->to_dst_file, COLO_MESSAGE_VMSTATE_SEND, &local_err);
 439     if (local_err) {
 440         qemu_mutex_unlock_iothread();
 441         goto out;
 442     }
 443     /* Note: device state is saved into buffer */
 444     ret = qemu_save_device_state(fb);
 445
 446     qemu_mutex_unlock_iothread();
 447     if (ret < 0) {
 448         goto out;
 449     }
 450     /*
 451      * Only save VM's live state, which not including device state.
 452      * TODO: We may need a timeout mechanism to prevent COLO process
 453      * to be blocked here.
 454      */
 455     qemu_savevm_live_state(s->to_dst_file);
 456
 457     qemu_fflush(fb);
 458
 459     /*
 460      * We need the size of the VMstate data in Secondary side,
 461      * With which we can decide how much data should be read.
 462      */
 463     colo_send_message_value(s->to_dst_file, COLO_MESSAGE_VMSTATE_SIZE,
 464                             bioc->usage, &local_err);
 465     if (local_err) {
 466         goto out;
 467     }
 468
 469     qemu_put_buffer(s->to_dst_file, bioc->data, bioc->usage);
 470     qemu_fflush(s->to_dst_file);
 471     ret = qemu_file_get_error(s->to_dst_file);
 472     if (ret < 0) {
 473         goto out;
 474     }
 475
 476     colo_receive_check_message(s->rp_state.from_dst_file,
 477                        COLO_MESSAGE_VMSTATE_RECEIVED, &local_err);
 478     if (local_err) {
 479         goto out;
 480     }
 481
 482     colo_receive_check_message(s->rp_state.from_dst_file,
 483                        COLO_MESSAGE_VMSTATE_LOADED, &local_err);
 484     if (local_err) {
 485         goto out;
 486     }
 487
 488     ret = 0;
 489
 490     qemu_mutex_lock_iothread();
 491     vm_start();
 492     qemu_mutex_unlock_iothread();
 493     trace_colo_vm_state_change("stop", "run");
 494
 495 out:
 496     if (local_err) {
 497         error_report_err(local_err);
 498     }
 499     return ret;
 500 }
 501
 502 static void colo_compare_notify_checkpoint(Notifier *notifier, void *data)
 503 {
 504     colo_checkpoint_notify(data);
 505 }
 506
 507 static void colo_process_checkpoint(MigrationState *s)
 508 {
 509     QIOChannelBuffer *bioc;
 510     QEMUFile *fb = NULL;
 511     int64_t current_time = qemu_clock_get_ms(QEMU_CLOCK_HOST);
 512     Error *local_err = NULL;
 513     int ret;
 514
 515     failover_init_state();
 516
 517     s->rp_state.from_dst_file = qemu_file_get_return_path(s->to_dst_file);
 518     if (!s->rp_state.from_dst_file) {
 519         error_report("Open QEMUFile from_dst_file failed");
 520         goto out;
 521     }
 522
 523     packets_compare_notifier.notify = colo_compare_notify_checkpoint;
 524     colo_compare_register_notifier(&packets_compare_notifier);
 525
 526     /*
 527      * Wait for Secondary finish loading VM states and enter COLO
 528      * restore.
 529      */
 530     colo_receive_check_message(s->rp_state.from_dst_file,
 531                        COLO_MESSAGE_CHECKPOINT_READY, &local_err);
 532     if (local_err) {
 533         goto out;
 534     }
 535     bioc = qio_channel_buffer_new(COLO_BUFFER_BASE_SIZE);
 536     fb = qemu_fopen_channel_output(QIO_CHANNEL(bioc));
 537     object_unref(OBJECT(bioc));
 538
 539     qemu_mutex_lock_iothread();
 540 #ifdef CONFIG_REPLICATION
 541     replication_start_all(REPLICATION_MODE_PRIMARY, &local_err);
 542     if (local_err) {
 543         qemu_mutex_unlock_iothread();
 544         goto out;
 545     }
 546 #else
 547         abort();
 548 #endif
 549
 550     vm_start();
 551     qemu_mutex_unlock_iothread();
 552     trace_colo_vm_state_change("stop", "run");
 553
 554     timer_mod(s->colo_delay_timer,
 555             current_time + s->parameters.x_checkpoint_delay);
 556
 557     while (s->state == MIGRATION_STATUS_COLO) {
 558         if (failover_get_state() != FAILOVER_STATUS_NONE) {
 559             error_report("failover request");
 560             goto out;
 561         }
 562
 563         qemu_sem_wait(&s->colo_checkpoint_sem);
 564
 565         if (s->state != MIGRATION_STATUS_COLO) {
 566             goto out;
 567         }
 568         ret = colo_do_checkpoint_transaction(s, bioc, fb);
 569         if (ret < 0) {
 570             goto out;
 571         }
 572     }
 573
 574 out:
 575     /* Throw the unreported error message after exited from loop */
 576     if (local_err) {
 577         error_report_err(local_err);
 578     }
 579
 580     if (fb) {
 581         qemu_fclose(fb);
 582     }
 583
 584     /*
 585      * There are only two reasons we can get here, some error happened
 586      * or the user triggered failover.
 587      */
 588     switch (failover_get_state()) {
 589     case FAILOVER_STATUS_NONE:
 590         qapi_event_send_colo_exit(COLO_MODE_PRIMARY,
 591                                   COLO_EXIT_REASON_ERROR);
 592         break;
 593     case FAILOVER_STATUS_REQUIRE:
 594         qapi_event_send_colo_exit(COLO_MODE_PRIMARY,
 595                                   COLO_EXIT_REASON_REQUEST);
 596         break;
 597     default:
 598         abort();
 599     }
 600
 601     /* Hope this not to be too long to wait here */
 602     qemu_sem_wait(&s->colo_exit_sem);
 603     qemu_sem_destroy(&s->colo_exit_sem);
 604
 605     /*
 606      * It is safe to unregister notifier after failover finished.
 607      * Besides, colo_delay_timer and colo_checkpoint_sem can't be
 608      * released befor unregister notifier, or there will be use-after-free
 609      * error.
 610      */
 611     colo_compare_unregister_notifier(&packets_compare_notifier);
 612     timer_del(s->colo_delay_timer);
 613     timer_free(s->colo_delay_timer);
 614     qemu_sem_destroy(&s->colo_checkpoint_sem);
 615
 616     /*
 617      * Must be called after failover BH is completed,
 618      * Or the failover BH may shutdown the wrong fd that
 619      * re-used by other threads after we release here.
 620      */
 621     if (s->rp_state.from_dst_file) {
 622         qemu_fclose(s->rp_state.from_dst_file);
 623     }
 624 }
 625
 626 void colo_checkpoint_notify(void *opaque)
 627 {
 628     MigrationState *s = opaque;
 629     int64_t next_notify_time;
 630
 631     qemu_sem_post(&s->colo_checkpoint_sem);
 632     s->colo_checkpoint_time = qemu_clock_get_ms(QEMU_CLOCK_HOST);
 633     next_notify_time = s->colo_checkpoint_time +
 634                     s->parameters.x_checkpoint_delay;
 635     timer_mod(s->colo_delay_timer, next_notify_time);
 636 }
 637
 638 void migrate_start_colo_process(MigrationState *s)
 639 {
 640     qemu_mutex_unlock_iothread();
 641     qemu_sem_init(&s->colo_checkpoint_sem, 0);
 642     s->colo_delay_timer =  timer_new_ms(QEMU_CLOCK_HOST,
 643                                 colo_checkpoint_notify, s);
 644
 645     qemu_sem_init(&s->colo_exit_sem, 0);
 646     migrate_set_state(&s->state, MIGRATION_STATUS_ACTIVE,
 647                       MIGRATION_STATUS_COLO);
 648     colo_process_checkpoint(s);
 649     qemu_mutex_lock_iothread();
 650 }
 651
 652 static void colo_wait_handle_message(QEMUFile *f, int *checkpoint_request,
 653                                      Error **errp)
 654 {
 655     COLOMessage msg;
 656     Error *local_err = NULL;
 657
 658     msg = colo_receive_message(f, &local_err);
 659     if (local_err) {
 660         error_propagate(errp, local_err);
 661         return;
 662     }
 663
 664     switch (msg) {
 665     case COLO_MESSAGE_CHECKPOINT_REQUEST:
 666         *checkpoint_request = 1;
 667         break;
 668     default:
 669         *checkpoint_request = 0;
 670         error_setg(errp, "Got unknown COLO message: %d", msg);
 671         break;
 672     }
 673 }
 674
 675 void *colo_process_incoming_thread(void *opaque)
 676 {
 677     MigrationIncomingState *mis = opaque;
 678     QEMUFile *fb = NULL;
 679     QIOChannelBuffer *bioc = NULL; /* Cache incoming device state */
 680     uint64_t total_size;
 681     uint64_t value;
 682     Error *local_err = NULL;
 683     int ret;
 684
 685     rcu_register_thread();
 686     qemu_sem_init(&mis->colo_incoming_sem, 0);
 687
 688     migrate_set_state(&mis->state, MIGRATION_STATUS_ACTIVE,
 689                       MIGRATION_STATUS_COLO);
 690
 691     failover_init_state();
 692
 693     mis->to_src_file = qemu_file_get_return_path(mis->from_src_file);
 694     if (!mis->to_src_file) {
 695         error_report("COLO incoming thread: Open QEMUFile to_src_file failed");
 696         goto out;
 697     }
 698     /*
 699      * Note: the communication between Primary side and Secondary side
 700      * should be sequential, we set the fd to unblocked in migration incoming
 701      * coroutine, and here we are in the COLO incoming thread, so it is ok to
 702      * set the fd back to blocked.
 703      */
 704     qemu_file_set_blocking(mis->from_src_file, true);
 705
 706     bioc = qio_channel_buffer_new(COLO_BUFFER_BASE_SIZE);
 707     fb = qemu_fopen_channel_input(QIO_CHANNEL(bioc));
 708     object_unref(OBJECT(bioc));
 709
 710     qemu_mutex_lock_iothread();
 711 #ifdef CONFIG_REPLICATION
 712     replication_start_all(REPLICATION_MODE_SECONDARY, &local_err);
 713     if (local_err) {
 714         qemu_mutex_unlock_iothread();
 715         goto out;
 716     }
 717 #else
 718         abort();
 719 #endif
 720     vm_start();
 721     trace_colo_vm_state_change("stop", "run");
 722     qemu_mutex_unlock_iothread();
 723
 724     colo_send_message(mis->to_src_file, COLO_MESSAGE_CHECKPOINT_READY,
 725                       &local_err);
 726     if (local_err) {
 727         goto out;
 728     }
 729
 730     while (mis->state == MIGRATION_STATUS_COLO) {
 731         int request = 0;
 732
 733         colo_wait_handle_message(mis->from_src_file, &request, &local_err);
 734         if (local_err) {
 735             goto out;
 736         }
 737         assert(request);
 738         if (failover_get_state() != FAILOVER_STATUS_NONE) {
 739             error_report("failover request");
 740             goto out;
 741         }
 742
 743         qemu_mutex_lock_iothread();
 744         vm_stop_force_state(RUN_STATE_COLO);
 745         trace_colo_vm_state_change("run", "stop");
 746         qemu_mutex_unlock_iothread();
 747
 748         /* FIXME: This is unnecessary for periodic checkpoint mode */
 749         colo_send_message(mis->to_src_file, COLO_MESSAGE_CHECKPOINT_REPLY,
 750                      &local_err);
 751         if (local_err) {
 752             goto out;
 753         }
 754
 755         colo_receive_check_message(mis->from_src_file,
 756                            COLO_MESSAGE_VMSTATE_SEND, &local_err);
 757         if (local_err) {
 758             goto out;
 759         }
 760
 761         qemu_mutex_lock_iothread();
 762         cpu_synchronize_all_pre_loadvm();
 763         ret = qemu_loadvm_state_main(mis->from_src_file, mis);
 764         qemu_mutex_unlock_iothread();
 765
 766         if (ret < 0) {
 767             error_report("Load VM's live state (ram) error");
 768             goto out;
 769         }
 770
 771         value = colo_receive_message_value(mis->from_src_file,
 772                                  COLO_MESSAGE_VMSTATE_SIZE, &local_err);
 773         if (local_err) {
 774             goto out;
 775         }
 776
 777         /*
 778          * Read VM device state data into channel buffer,
 779          * It's better to re-use the memory allocated.
 780          * Here we need to handle the channel buffer directly.
 781          */
 782         if (value > bioc->capacity) {
 783             bioc->capacity = value;
 784             bioc->data = g_realloc(bioc->data, bioc->capacity);
 785         }
 786         total_size = qemu_get_buffer(mis->from_src_file, bioc->data, value);
 787         if (total_size != value) {
 788             error_report("Got %" PRIu64 " VMState data, less than expected"
 789                         " %" PRIu64, total_size, value);
 790             goto out;
 791         }
 792         bioc->usage = total_size;
 793         qio_channel_io_seek(QIO_CHANNEL(bioc), 0, 0, NULL);
 794
 795         colo_send_message(mis->to_src_file, COLO_MESSAGE_VMSTATE_RECEIVED,
 796                      &local_err);
 797         if (local_err) {
 798             goto out;
 799         }
 800
 801         qemu_mutex_lock_iothread();
 802         vmstate_loading = true;
 803         ret = qemu_load_device_state(fb);
 804         if (ret < 0) {
 805             error_report("COLO: load device state failed");
 806             qemu_mutex_unlock_iothread();
 807             goto out;
 808         }
 809
 810 #ifdef CONFIG_REPLICATION
 811         replication_get_error_all(&local_err);
 812         if (local_err) {
 813             qemu_mutex_unlock_iothread();
 814             goto out;
 815         }
 816
 817         /* discard colo disk buffer */
 818         replication_do_checkpoint_all(&local_err);
 819         if (local_err) {
 820             qemu_mutex_unlock_iothread();
 821             goto out;
 822         }
 823 #else
 824         abort();
 825 #endif
 826         /* Notify all filters of all NIC to do checkpoint */
 827         colo_notify_filters_event(COLO_EVENT_CHECKPOINT, &local_err);
 828
 829         if (local_err) {
 830             qemu_mutex_unlock_iothread();
 831             goto out;
 832         }
 833
 834         vmstate_loading = false;
 835         vm_start();
 836         trace_colo_vm_state_change("stop", "run");
 837         qemu_mutex_unlock_iothread();
 838
 839         if (failover_get_state() == FAILOVER_STATUS_RELAUNCH) {
 840             failover_set_state(FAILOVER_STATUS_RELAUNCH,
 841                             FAILOVER_STATUS_NONE);
 842             failover_request_active(NULL);
 843             goto out;
 844         }
 845
 846         colo_send_message(mis->to_src_file, COLO_MESSAGE_VMSTATE_LOADED,
 847                      &local_err);
 848         if (local_err) {
 849             goto out;
 850         }
 851     }
 852
 853 out:
 854     vmstate_loading = false;
 855     /* Throw the unreported error message after exited from loop */
 856     if (local_err) {
 857         error_report_err(local_err);
 858     }
 859
 860     switch (failover_get_state()) {
 861     case FAILOVER_STATUS_NONE:
 862         qapi_event_send_colo_exit(COLO_MODE_SECONDARY,
 863                                   COLO_EXIT_REASON_ERROR);
 864         break;
 865     case FAILOVER_STATUS_REQUIRE:
 866         qapi_event_send_colo_exit(COLO_MODE_SECONDARY,
 867                                   COLO_EXIT_REASON_REQUEST);
 868         break;
 869     default:
 870         abort();
 871     }
 872
 873     if (fb) {
 874         qemu_fclose(fb);
 875     }
 876
 877     /* Hope this not to be too long to loop here */
 878     qemu_sem_wait(&mis->colo_incoming_sem);
 879     qemu_sem_destroy(&mis->colo_incoming_sem);
 880     /* Must be called after failover BH is completed */
 881     if (mis->to_src_file) {
 882         qemu_fclose(mis->to_src_file);
 883     }
 884     migration_incoming_disable_colo();
 885
 886     rcu_unregister_thread();
 887     return NULL;
 888 }