colo.c 23.1 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13
/*
 * COarse-grain LOck-stepping Virtual Machines for Non-stop Service (COLO)
 * (a.k.a. Fault Tolerance or Continuous Replication)
 *
 * Copyright (c) 2016 HUAWEI TECHNOLOGIES CO., LTD.
 * Copyright (c) 2016 FUJITSU LIMITED
 * Copyright (c) 2016 Intel Corporation
 *
 * This work is licensed under the terms of the GNU GPL, version 2 or
 * later.  See the COPYING file in the top-level directory.
 */

#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"
J
Juan Quintela 已提交
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 29
#include "net/colo-compare.h"
#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

35
static bool vmstate_loading;
36
static Notifier packets_compare_notifier;
37

38 39
#define COLO_BUFFER_BASE_SIZE (4 * 1024 * 1024)

40 41 42 43 44 45 46
bool migration_in_colo_state(void)
{
    MigrationState *s = migrate_get_current();

    return (s->state == MIGRATION_STATUS_COLO);
}

47 48 49 50 51 52 53
bool migration_incoming_in_colo_state(void)
{
    MigrationIncomingState *mis = migration_incoming_get_current();

    return mis && (mis->state == MIGRATION_STATUS_COLO);
}

54 55 56 57 58
static bool colo_runstate_is_stopped(void)
{
    return runstate_check(RUN_STATE_COLO) || !runstate_is_running();
}

59 60 61 62
static void secondary_vm_do_failover(void)
{
    int old_state;
    MigrationIncomingState *mis = migration_incoming_get_current();
63
    Error *local_err = NULL;
64

65 66 67 68 69 70 71 72
    /* Can not do failover during the process of VM's loading VMstate, Or
     * it will break the secondary VM.
     */
    if (vmstate_loading) {
        old_state = failover_set_state(FAILOVER_STATUS_ACTIVE,
                        FAILOVER_STATUS_RELAUNCH);
        if (old_state != FAILOVER_STATUS_ACTIVE) {
            error_report("Unknown error while do failover for secondary VM,"
73
                         "old_state: %s", FailoverStatus_str(old_state));
74 75 76 77
        }
        return;
    }

78 79 80
    migrate_set_state(&mis->state, MIGRATION_STATUS_COLO,
                      MIGRATION_STATUS_COMPLETED);

81 82 83 84 85
    replication_stop_all(true, &local_err);
    if (local_err) {
        error_report_err(local_err);
    }

86 87 88 89 90
    if (!autostart) {
        error_report("\"-S\" qemu option will be ignored in secondary side");
        /* recover runstate to normal migration finish state */
        autostart = true;
    }
91 92 93 94 95 96 97 98 99 100 101 102
    /*
     * Make sure COLO incoming thread not block in recv or send,
     * If mis->from_src_file and mis->to_src_file use the same fd,
     * The second shutdown() will return -1, we ignore this value,
     * It is harmless.
     */
    if (mis->from_src_file) {
        qemu_file_shutdown(mis->from_src_file);
    }
    if (mis->to_src_file) {
        qemu_file_shutdown(mis->to_src_file);
    }
103 104 105 106 107

    old_state = failover_set_state(FAILOVER_STATUS_ACTIVE,
                                   FAILOVER_STATUS_COMPLETED);
    if (old_state != FAILOVER_STATUS_ACTIVE) {
        error_report("Incorrect state (%s) while doing failover for "
108
                     "secondary VM", FailoverStatus_str(old_state));
109 110
        return;
    }
111 112
    /* Notify COLO incoming thread that failover work is finished */
    qemu_sem_post(&mis->colo_incoming_sem);
113 114 115 116 117 118
    /* For Secondary VM, jump to incoming co */
    if (mis->migration_incoming_co) {
        qemu_coroutine_enter(mis->migration_incoming_co);
    }
}

119 120 121 122
static void primary_vm_do_failover(void)
{
    MigrationState *s = migrate_get_current();
    int old_state;
123
    Error *local_err = NULL;
124 125 126 127

    migrate_set_state(&s->state, MIGRATION_STATUS_COLO,
                      MIGRATION_STATUS_COMPLETED);

128 129 130 131 132 133 134 135 136 137 138 139
    /*
     * Wake up COLO thread which may blocked in recv() or send(),
     * The s->rp_state.from_dst_file and s->to_dst_file may use the
     * same fd, but we still shutdown the fd for twice, it is harmless.
     */
    if (s->to_dst_file) {
        qemu_file_shutdown(s->to_dst_file);
    }
    if (s->rp_state.from_dst_file) {
        qemu_file_shutdown(s->rp_state.from_dst_file);
    }

140 141 142 143
    old_state = failover_set_state(FAILOVER_STATUS_ACTIVE,
                                   FAILOVER_STATUS_COMPLETED);
    if (old_state != FAILOVER_STATUS_ACTIVE) {
        error_report("Incorrect state (%s) while doing failover for Primary VM",
144
                     FailoverStatus_str(old_state));
145 146
        return;
    }
147 148 149 150 151 152 153

    replication_stop_all(true, &local_err);
    if (local_err) {
        error_report_err(local_err);
        local_err = NULL;
    }

154 155
    /* Notify COLO thread that failover work is finished */
    qemu_sem_post(&s->colo_exit_sem);
156 157
}

158 159 160 161 162 163 164
COLOMode get_colo_mode(void)
{
    if (migration_in_colo_state()) {
        return COLO_MODE_PRIMARY;
    } else if (migration_incoming_in_colo_state()) {
        return COLO_MODE_SECONDARY;
    } else {
165
        return COLO_MODE_NONE;
166 167 168
    }
}

169 170 171 172 173 174 175 176 177
void colo_do_failover(MigrationState *s)
{
    /* Make sure VM stopped while failover happened. */
    if (!colo_runstate_is_stopped()) {
        vm_stop_force_state(RUN_STATE_COLO);
    }

    if (get_colo_mode() == COLO_MODE_PRIMARY) {
        primary_vm_do_failover();
178 179
    } else {
        secondary_vm_do_failover();
180 181 182
    }
}

183 184 185 186
void qmp_xen_set_replication(bool enable, bool primary,
                             bool has_failover, bool failover,
                             Error **errp)
{
187
#ifdef CONFIG_REPLICATION
188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205
    ReplicationMode mode = primary ?
                           REPLICATION_MODE_PRIMARY :
                           REPLICATION_MODE_SECONDARY;

    if (has_failover && enable) {
        error_setg(errp, "Parameter 'failover' is only for"
                   " stopping replication");
        return;
    }

    if (enable) {
        replication_start_all(mode, errp);
    } else {
        if (!has_failover) {
            failover = NULL;
        }
        replication_stop_all(failover, failover ? NULL : errp);
    }
206 207 208
#else
    abort();
#endif
209 210
}

211 212
ReplicationStatus *qmp_query_xen_replication_status(Error **errp)
{
213
#ifdef CONFIG_REPLICATION
214 215 216 217 218 219 220 221 222 223 224 225 226 227
    Error *err = NULL;
    ReplicationStatus *s = g_new0(ReplicationStatus, 1);

    replication_get_error_all(&err);
    if (err) {
        s->error = true;
        s->has_desc = true;
        s->desc = g_strdup(error_get_pretty(err));
    } else {
        s->error = false;
    }

    error_free(err);
    return s;
228 229 230
#else
    abort();
#endif
231 232 233 234
}

void qmp_xen_colo_do_checkpoint(Error **errp)
{
235
#ifdef CONFIG_REPLICATION
236
    replication_do_checkpoint_all(errp);
237 238 239
#else
    abort();
#endif
240 241
}

242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261
COLOStatus *qmp_query_colo_status(Error **errp)
{
    COLOStatus *s = g_new0(COLOStatus, 1);

    s->mode = get_colo_mode();

    switch (failover_get_state()) {
    case FAILOVER_STATUS_NONE:
        s->reason = COLO_EXIT_REASON_NONE;
        break;
    case FAILOVER_STATUS_REQUIRE:
        s->reason = COLO_EXIT_REASON_REQUEST;
        break;
    default:
        s->reason = COLO_EXIT_REASON_ERROR;
    }

    return s;
}

262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277
static void colo_send_message(QEMUFile *f, COLOMessage msg,
                              Error **errp)
{
    int ret;

    if (msg >= COLO_MESSAGE__MAX) {
        error_setg(errp, "%s: Invalid message", __func__);
        return;
    }
    qemu_put_be32(f, msg);
    qemu_fflush(f);

    ret = qemu_file_get_error(f);
    if (ret < 0) {
        error_setg_errno(errp, -ret, "Can't send COLO message");
    }
278
    trace_colo_send_message(COLOMessage_str(msg));
279 280
}

281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297
static void colo_send_message_value(QEMUFile *f, COLOMessage msg,
                                    uint64_t value, Error **errp)
{
    Error *local_err = NULL;
    int ret;

    colo_send_message(f, msg, &local_err);
    if (local_err) {
        error_propagate(errp, local_err);
        return;
    }
    qemu_put_be64(f, value);
    qemu_fflush(f);

    ret = qemu_file_get_error(f);
    if (ret < 0) {
        error_setg_errno(errp, -ret, "Failed to send value for message:%s",
298
                         COLOMessage_str(msg));
299 300 301
    }
}

302 303 304 305 306 307 308 309 310 311 312 313 314 315 316
static COLOMessage colo_receive_message(QEMUFile *f, Error **errp)
{
    COLOMessage msg;
    int ret;

    msg = qemu_get_be32(f);
    ret = qemu_file_get_error(f);
    if (ret < 0) {
        error_setg_errno(errp, -ret, "Can't receive COLO message");
        return msg;
    }
    if (msg >= COLO_MESSAGE__MAX) {
        error_setg(errp, "%s: Invalid message", __func__);
        return msg;
    }
317
    trace_colo_receive_message(COLOMessage_str(msg));
318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337
    return msg;
}

static void colo_receive_check_message(QEMUFile *f, COLOMessage expect_msg,
                                       Error **errp)
{
    COLOMessage msg;
    Error *local_err = NULL;

    msg = colo_receive_message(f, &local_err);
    if (local_err) {
        error_propagate(errp, local_err);
        return;
    }
    if (msg != expect_msg) {
        error_setg(errp, "Unexpected COLO message %d, expected %d",
                          msg, expect_msg);
    }
}

338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354
static uint64_t colo_receive_message_value(QEMUFile *f, uint32_t expect_msg,
                                           Error **errp)
{
    Error *local_err = NULL;
    uint64_t value;
    int ret;

    colo_receive_check_message(f, expect_msg, &local_err);
    if (local_err) {
        error_propagate(errp, local_err);
        return 0;
    }

    value = qemu_get_be64(f);
    ret = qemu_file_get_error(f);
    if (ret < 0) {
        error_setg_errno(errp, -ret, "Failed to get value for COLO message: %s",
355
                         COLOMessage_str(expect_msg));
356 357 358 359
    }
    return value;
}

360 361 362
static int colo_do_checkpoint_transaction(MigrationState *s,
                                          QIOChannelBuffer *bioc,
                                          QEMUFile *fb)
363 364
{
    Error *local_err = NULL;
365
    int ret = -1;
366 367 368 369 370 371 372 373 374 375 376 377

    colo_send_message(s->to_dst_file, COLO_MESSAGE_CHECKPOINT_REQUEST,
                      &local_err);
    if (local_err) {
        goto out;
    }

    colo_receive_check_message(s->rp_state.from_dst_file,
                    COLO_MESSAGE_CHECKPOINT_REPLY, &local_err);
    if (local_err) {
        goto out;
    }
378 379 380
    /* Reset channel-buffer directly */
    qio_channel_io_seek(QIO_CHANNEL(bioc), 0, 0, NULL);
    bioc->usage = 0;
381

382
    qemu_mutex_lock_iothread();
383 384 385 386
    if (failover_get_state() != FAILOVER_STATUS_NONE) {
        qemu_mutex_unlock_iothread();
        goto out;
    }
387 388 389
    vm_stop_force_state(RUN_STATE_COLO);
    qemu_mutex_unlock_iothread();
    trace_colo_vm_state_change("run", "stop");
390 391 392 393 394 395 396
    /*
     * Failover request bh could be called after vm_stop_force_state(),
     * So we need check failover_request_is_active() again.
     */
    if (failover_get_state() != FAILOVER_STATUS_NONE) {
        goto out;
    }
397

398 399 400 401 402
    colo_notify_compares_event(NULL, COLO_EVENT_CHECKPOINT, &local_err);
    if (local_err) {
        goto out;
    }

403
    /* Disable block migration */
404
    migrate_set_block_enabled(false, &local_err);
405
    qemu_mutex_lock_iothread();
406 407 408 409 410
    replication_do_checkpoint_all(&local_err);
    if (local_err) {
        qemu_mutex_unlock_iothread();
        goto out;
    }
411 412 413

    colo_send_message(s->to_dst_file, COLO_MESSAGE_VMSTATE_SEND, &local_err);
    if (local_err) {
414 415 416 417 418 419 420 421
        qemu_mutex_unlock_iothread();
        goto out;
    }
    /* Note: device state is saved into buffer */
    ret = qemu_save_device_state(fb);

    qemu_mutex_unlock_iothread();
    if (ret < 0) {
422 423
        goto out;
    }
424 425 426 427 428 429 430 431 432
    /*
     * Only save VM's live state, which not including device state.
     * TODO: We may need a timeout mechanism to prevent COLO process
     * to be blocked here.
     */
    qemu_savevm_live_state(s->to_dst_file);

    qemu_fflush(fb);

433 434 435 436 437 438 439 440 441
    /*
     * We need the size of the VMstate data in Secondary side,
     * With which we can decide how much data should be read.
     */
    colo_send_message_value(s->to_dst_file, COLO_MESSAGE_VMSTATE_SIZE,
                            bioc->usage, &local_err);
    if (local_err) {
        goto out;
    }
442

443 444 445 446 447 448
    qemu_put_buffer(s->to_dst_file, bioc->data, bioc->usage);
    qemu_fflush(s->to_dst_file);
    ret = qemu_file_get_error(s->to_dst_file);
    if (ret < 0) {
        goto out;
    }
449 450 451 452 453 454 455 456 457 458 459 460 461

    colo_receive_check_message(s->rp_state.from_dst_file,
                       COLO_MESSAGE_VMSTATE_RECEIVED, &local_err);
    if (local_err) {
        goto out;
    }

    colo_receive_check_message(s->rp_state.from_dst_file,
                       COLO_MESSAGE_VMSTATE_LOADED, &local_err);
    if (local_err) {
        goto out;
    }

462 463 464 465 466 467
    ret = 0;

    qemu_mutex_lock_iothread();
    vm_start();
    qemu_mutex_unlock_iothread();
    trace_colo_vm_state_change("stop", "run");
468 469 470 471 472

out:
    if (local_err) {
        error_report_err(local_err);
    }
473
    return ret;
474 475
}

476 477 478 479 480
static void colo_compare_notify_checkpoint(Notifier *notifier, void *data)
{
    colo_checkpoint_notify(data);
}

481 482
static void colo_process_checkpoint(MigrationState *s)
{
483 484
    QIOChannelBuffer *bioc;
    QEMUFile *fb = NULL;
485
    int64_t current_time = qemu_clock_get_ms(QEMU_CLOCK_HOST);
486 487 488
    Error *local_err = NULL;
    int ret;

489 490
    failover_init_state();

491 492 493 494 495 496
    s->rp_state.from_dst_file = qemu_file_get_return_path(s->to_dst_file);
    if (!s->rp_state.from_dst_file) {
        error_report("Open QEMUFile from_dst_file failed");
        goto out;
    }

497 498 499
    packets_compare_notifier.notify = colo_compare_notify_checkpoint;
    colo_compare_register_notifier(&packets_compare_notifier);

500 501 502 503 504 505 506 507 508
    /*
     * Wait for Secondary finish loading VM states and enter COLO
     * restore.
     */
    colo_receive_check_message(s->rp_state.from_dst_file,
                       COLO_MESSAGE_CHECKPOINT_READY, &local_err);
    if (local_err) {
        goto out;
    }
509 510 511
    bioc = qio_channel_buffer_new(COLO_BUFFER_BASE_SIZE);
    fb = qemu_fopen_channel_output(QIO_CHANNEL(bioc));
    object_unref(OBJECT(bioc));
512

513
    qemu_mutex_lock_iothread();
514 515 516 517 518 519
    replication_start_all(REPLICATION_MODE_PRIMARY, &local_err);
    if (local_err) {
        qemu_mutex_unlock_iothread();
        goto out;
    }

520 521 522 523
    vm_start();
    qemu_mutex_unlock_iothread();
    trace_colo_vm_state_change("stop", "run");

524 525 526
    timer_mod(s->colo_delay_timer,
            current_time + s->parameters.x_checkpoint_delay);

527
    while (s->state == MIGRATION_STATUS_COLO) {
528 529 530 531 532
        if (failover_get_state() != FAILOVER_STATUS_NONE) {
            error_report("failover request");
            goto out;
        }

533
        qemu_sem_wait(&s->colo_checkpoint_sem);
534

535
        ret = colo_do_checkpoint_transaction(s, bioc, fb);
536 537 538 539
        if (ret < 0) {
            goto out;
        }
    }
540

541
out:
542 543 544 545 546
    /* Throw the unreported error message after exited from loop */
    if (local_err) {
        error_report_err(local_err);
    }

547 548 549 550
    if (fb) {
        qemu_fclose(fb);
    }

551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567
    /*
     * There are only two reasons we can get here, some error happened
     * or the user triggered failover.
     */
    switch (failover_get_state()) {
    case FAILOVER_STATUS_NONE:
        qapi_event_send_colo_exit(COLO_MODE_PRIMARY,
                                  COLO_EXIT_REASON_ERROR);
        break;
    case FAILOVER_STATUS_REQUIRE:
        qapi_event_send_colo_exit(COLO_MODE_PRIMARY,
                                  COLO_EXIT_REASON_REQUEST);
        break;
    default:
        abort();
    }

568 569 570
    /* Hope this not to be too long to wait here */
    qemu_sem_wait(&s->colo_exit_sem);
    qemu_sem_destroy(&s->colo_exit_sem);
571 572 573 574 575 576 577 578 579 580 581 582

    /*
     * It is safe to unregister notifier after failover finished.
     * Besides, colo_delay_timer and colo_checkpoint_sem can't be
     * released befor unregister notifier, or there will be use-after-free
     * error.
     */
    colo_compare_unregister_notifier(&packets_compare_notifier);
    timer_del(s->colo_delay_timer);
    timer_free(s->colo_delay_timer);
    qemu_sem_destroy(&s->colo_checkpoint_sem);

583 584 585 586 587
    /*
     * Must be called after failover BH is completed,
     * Or the failover BH may shutdown the wrong fd that
     * re-used by other threads after we release here.
     */
588 589 590
    if (s->rp_state.from_dst_file) {
        qemu_fclose(s->rp_state.from_dst_file);
    }
591 592
}

593 594 595 596 597 598 599 600 601 602 603 604
void colo_checkpoint_notify(void *opaque)
{
    MigrationState *s = opaque;
    int64_t next_notify_time;

    qemu_sem_post(&s->colo_checkpoint_sem);
    s->colo_checkpoint_time = qemu_clock_get_ms(QEMU_CLOCK_HOST);
    next_notify_time = s->colo_checkpoint_time +
                    s->parameters.x_checkpoint_delay;
    timer_mod(s->colo_delay_timer, next_notify_time);
}

605 606 607
void migrate_start_colo_process(MigrationState *s)
{
    qemu_mutex_unlock_iothread();
608 609 610 611
    qemu_sem_init(&s->colo_checkpoint_sem, 0);
    s->colo_delay_timer =  timer_new_ms(QEMU_CLOCK_HOST,
                                colo_checkpoint_notify, s);

612
    qemu_sem_init(&s->colo_exit_sem, 0);
613 614 615 616 617
    migrate_set_state(&s->state, MIGRATION_STATUS_ACTIVE,
                      MIGRATION_STATUS_COLO);
    colo_process_checkpoint(s);
    qemu_mutex_lock_iothread();
}
618

619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641
static void colo_wait_handle_message(QEMUFile *f, int *checkpoint_request,
                                     Error **errp)
{
    COLOMessage msg;
    Error *local_err = NULL;

    msg = colo_receive_message(f, &local_err);
    if (local_err) {
        error_propagate(errp, local_err);
        return;
    }

    switch (msg) {
    case COLO_MESSAGE_CHECKPOINT_REQUEST:
        *checkpoint_request = 1;
        break;
    default:
        *checkpoint_request = 0;
        error_setg(errp, "Got unknown COLO message: %d", msg);
        break;
    }
}

642 643 644
void *colo_process_incoming_thread(void *opaque)
{
    MigrationIncomingState *mis = opaque;
645 646 647 648
    QEMUFile *fb = NULL;
    QIOChannelBuffer *bioc = NULL; /* Cache incoming device state */
    uint64_t total_size;
    uint64_t value;
649
    Error *local_err = NULL;
650
    int ret;
651

652
    rcu_register_thread();
653 654
    qemu_sem_init(&mis->colo_incoming_sem, 0);

655 656 657
    migrate_set_state(&mis->state, MIGRATION_STATUS_ACTIVE,
                      MIGRATION_STATUS_COLO);

658 659
    failover_init_state();

660 661 662 663 664 665 666 667 668 669 670 671 672
    mis->to_src_file = qemu_file_get_return_path(mis->from_src_file);
    if (!mis->to_src_file) {
        error_report("COLO incoming thread: Open QEMUFile to_src_file failed");
        goto out;
    }
    /*
     * Note: the communication between Primary side and Secondary side
     * should be sequential, we set the fd to unblocked in migration incoming
     * coroutine, and here we are in the COLO incoming thread, so it is ok to
     * set the fd back to blocked.
     */
    qemu_file_set_blocking(mis->from_src_file, true);

673 674 675 676
    bioc = qio_channel_buffer_new(COLO_BUFFER_BASE_SIZE);
    fb = qemu_fopen_channel_input(QIO_CHANNEL(bioc));
    object_unref(OBJECT(bioc));

677
    qemu_mutex_lock_iothread();
678 679 680 681 682
    replication_start_all(REPLICATION_MODE_SECONDARY, &local_err);
    if (local_err) {
        qemu_mutex_unlock_iothread();
        goto out;
    }
683 684 685 686
    vm_start();
    trace_colo_vm_state_change("stop", "run");
    qemu_mutex_unlock_iothread();

687 688 689 690 691 692 693
    colo_send_message(mis->to_src_file, COLO_MESSAGE_CHECKPOINT_READY,
                      &local_err);
    if (local_err) {
        goto out;
    }

    while (mis->state == MIGRATION_STATUS_COLO) {
694
        int request = 0;
695 696 697 698 699 700

        colo_wait_handle_message(mis->from_src_file, &request, &local_err);
        if (local_err) {
            goto out;
        }
        assert(request);
701 702 703 704 705
        if (failover_get_state() != FAILOVER_STATUS_NONE) {
            error_report("failover request");
            goto out;
        }

706 707 708 709 710
        qemu_mutex_lock_iothread();
        vm_stop_force_state(RUN_STATE_COLO);
        trace_colo_vm_state_change("run", "stop");
        qemu_mutex_unlock_iothread();

711 712 713 714 715 716 717 718 719 720 721 722 723
        /* FIXME: This is unnecessary for periodic checkpoint mode */
        colo_send_message(mis->to_src_file, COLO_MESSAGE_CHECKPOINT_REPLY,
                     &local_err);
        if (local_err) {
            goto out;
        }

        colo_receive_check_message(mis->from_src_file,
                           COLO_MESSAGE_VMSTATE_SEND, &local_err);
        if (local_err) {
            goto out;
        }

724 725 726 727 728 729 730 731 732 733
        qemu_mutex_lock_iothread();
        cpu_synchronize_all_pre_loadvm();
        ret = qemu_loadvm_state_main(mis->from_src_file, mis);
        qemu_mutex_unlock_iothread();

        if (ret < 0) {
            error_report("Load VM's live state (ram) error");
            goto out;
        }

734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756
        value = colo_receive_message_value(mis->from_src_file,
                                 COLO_MESSAGE_VMSTATE_SIZE, &local_err);
        if (local_err) {
            goto out;
        }

        /*
         * Read VM device state data into channel buffer,
         * It's better to re-use the memory allocated.
         * Here we need to handle the channel buffer directly.
         */
        if (value > bioc->capacity) {
            bioc->capacity = value;
            bioc->data = g_realloc(bioc->data, bioc->capacity);
        }
        total_size = qemu_get_buffer(mis->from_src_file, bioc->data, value);
        if (total_size != value) {
            error_report("Got %" PRIu64 " VMState data, less than expected"
                        " %" PRIu64, total_size, value);
            goto out;
        }
        bioc->usage = total_size;
        qio_channel_io_seek(QIO_CHANNEL(bioc), 0, 0, NULL);
757 758 759 760 761 762 763

        colo_send_message(mis->to_src_file, COLO_MESSAGE_VMSTATE_RECEIVED,
                     &local_err);
        if (local_err) {
            goto out;
        }

764
        qemu_mutex_lock_iothread();
765
        vmstate_loading = true;
766 767 768
        ret = qemu_load_device_state(fb);
        if (ret < 0) {
            error_report("COLO: load device state failed");
769 770 771
            qemu_mutex_unlock_iothread();
            goto out;
        }
772

773 774 775 776 777 778 779 780 781 782 783 784
        replication_get_error_all(&local_err);
        if (local_err) {
            qemu_mutex_unlock_iothread();
            goto out;
        }
        /* discard colo disk buffer */
        replication_do_checkpoint_all(&local_err);
        if (local_err) {
            qemu_mutex_unlock_iothread();
            goto out;
        }

785
        vmstate_loading = false;
786 787
        vm_start();
        trace_colo_vm_state_change("stop", "run");
788
        qemu_mutex_unlock_iothread();
789

790 791 792 793 794 795 796
        if (failover_get_state() == FAILOVER_STATUS_RELAUNCH) {
            failover_set_state(FAILOVER_STATUS_RELAUNCH,
                            FAILOVER_STATUS_NONE);
            failover_request_active(NULL);
            goto out;
        }

797 798 799 800 801 802
        colo_send_message(mis->to_src_file, COLO_MESSAGE_VMSTATE_LOADED,
                     &local_err);
        if (local_err) {
            goto out;
        }
    }
803

804
out:
805
    vmstate_loading = false;
806 807 808 809 810
    /* Throw the unreported error message after exited from loop */
    if (local_err) {
        error_report_err(local_err);
    }

811 812 813 814 815 816 817 818 819 820 821 822 823
    switch (failover_get_state()) {
    case FAILOVER_STATUS_NONE:
        qapi_event_send_colo_exit(COLO_MODE_SECONDARY,
                                  COLO_EXIT_REASON_ERROR);
        break;
    case FAILOVER_STATUS_REQUIRE:
        qapi_event_send_colo_exit(COLO_MODE_SECONDARY,
                                  COLO_EXIT_REASON_REQUEST);
        break;
    default:
        abort();
    }

824 825 826 827
    if (fb) {
        qemu_fclose(fb);
    }

828 829 830 831
    /* Hope this not to be too long to loop here */
    qemu_sem_wait(&mis->colo_incoming_sem);
    qemu_sem_destroy(&mis->colo_incoming_sem);
    /* Must be called after failover BH is completed */
832 833 834
    if (mis->to_src_file) {
        qemu_fclose(mis->to_src_file);
    }
835
    migration_incoming_disable_colo();
836

837
    rcu_unregister_thread();
838 839
    return NULL;
}