block/replication.c

   1 /*
   2  * Replication Block filter
   3  *
   4  * Copyright (c) 2016 HUAWEI TECHNOLOGIES CO., LTD.
   5  * Copyright (c) 2016 Intel Corporation
   6  * Copyright (c) 2016 FUJITSU LIMITED
   7  *
   8  * Author:
   9  *   Wen Congyang <wency@cn.fujitsu.com>
  10  *
  11  * This work is licensed under the terms of the GNU GPL, version 2 or later.
  12  * See the COPYING file in the top-level directory.
  13  */
  14
  15 #include "qemu/osdep.h"
  16 #include "qemu/module.h"
  17 #include "qemu/option.h"
  18 #include "block/nbd.h"
  19 #include "block/blockjob.h"
  20 #include "block/block_int.h"
  21 #include "block/block_backup.h"
  22 #include "sysemu/block-backend.h"
  23 #include "qapi/error.h"
  24 #include "qapi/qmp/qdict.h"
  25 #include "block/replication.h"
  26
  27 typedef enum {
  28     BLOCK_REPLICATION_NONE,             /* block replication is not started */
  29     BLOCK_REPLICATION_RUNNING,          /* block replication is running */
  30     BLOCK_REPLICATION_FAILOVER,         /* failover is running in background */
  31     BLOCK_REPLICATION_FAILOVER_FAILED,  /* failover failed */
  32     BLOCK_REPLICATION_DONE,             /* block replication is done */
  33 } ReplicationStage;
  34
  35 typedef struct BDRVReplicationState {
  36     ReplicationMode mode;
  37     ReplicationStage stage;
  38     BlockJob *commit_job;
  39     BdrvChild *hidden_disk;
  40     BdrvChild *secondary_disk;
  41     BlockJob *backup_job;
  42     char *top_id;
  43     ReplicationState *rs;
  44     Error *blocker;
  45     bool orig_hidden_read_only;
  46     bool orig_secondary_read_only;
  47     int error;
  48 } BDRVReplicationState;
  49
  50 static void replication_start(ReplicationState *rs, ReplicationMode mode,
  51                               Error **errp);
  52 static void replication_do_checkpoint(ReplicationState *rs, Error **errp);
  53 static void replication_get_error(ReplicationState *rs, Error **errp);
  54 static void replication_stop(ReplicationState *rs, bool failover,
  55                              Error **errp);
  56
  57 #define REPLICATION_MODE        "mode"
  58 #define REPLICATION_TOP_ID      "top-id"
  59 static QemuOptsList replication_runtime_opts = {
  60     .name = "replication",
  61     .head = QTAILQ_HEAD_INITIALIZER(replication_runtime_opts.head),
  62     .desc = {
  63         {
  64             .name = REPLICATION_MODE,
  65             .type = QEMU_OPT_STRING,
  66         },
  67         {
  68             .name = REPLICATION_TOP_ID,
  69             .type = QEMU_OPT_STRING,
  70         },
  71         { /* end of list */ }
  72     },
  73 };
  74
  75 static ReplicationOps replication_ops = {
  76     .start = replication_start,
  77     .checkpoint = replication_do_checkpoint,
  78     .get_error = replication_get_error,
  79     .stop = replication_stop,
  80 };
  81
  82 static int replication_open(BlockDriverState *bs, QDict *options,
  83                             int flags, Error **errp)
  84 {
  85     int ret;
  86     BDRVReplicationState *s = bs->opaque;
  87     QemuOpts *opts = NULL;
  88     const char *mode;
  89     const char *top_id;
  90
  91     ret = bdrv_open_file_child(NULL, options, "file", bs, errp);
  92     if (ret < 0) {
  93         return ret;
  94     }
  95
  96     ret = -EINVAL;
  97     opts = qemu_opts_create(&replication_runtime_opts, NULL, 0, &error_abort);
  98     if (!qemu_opts_absorb_qdict(opts, options, errp)) {
  99         goto fail;
 100     }
 101
 102     mode = qemu_opt_get(opts, REPLICATION_MODE);
 103     if (!mode) {
 104         error_setg(errp, "Missing the option mode");
 105         goto fail;
 106     }
 107
 108     if (!strcmp(mode, "primary")) {
 109         s->mode = REPLICATION_MODE_PRIMARY;
 110         top_id = qemu_opt_get(opts, REPLICATION_TOP_ID);
 111         if (top_id) {
 112             error_setg(errp,
 113                        "The primary side does not support option top-id");
 114             goto fail;
 115         }
 116     } else if (!strcmp(mode, "secondary")) {
 117         s->mode = REPLICATION_MODE_SECONDARY;
 118         top_id = qemu_opt_get(opts, REPLICATION_TOP_ID);
 119         s->top_id = g_strdup(top_id);
 120         if (!s->top_id) {
 121             error_setg(errp, "Missing the option top-id");
 122             goto fail;
 123         }
 124     } else {
 125         error_setg(errp,
 126                    "The option mode's value should be primary or secondary");
 127         goto fail;
 128     }
 129
 130     s->rs = replication_new(bs, &replication_ops);
 131
 132     ret = 0;
 133
 134 fail:
 135     qemu_opts_del(opts);
 136     return ret;
 137 }
 138
 139 static void replication_close(BlockDriverState *bs)
 140 {
 141     BDRVReplicationState *s = bs->opaque;
 142     Job *commit_job;
 143     GLOBAL_STATE_CODE();
 144
 145     if (s->stage == BLOCK_REPLICATION_RUNNING) {
 146         replication_stop(s->rs, false, NULL);
 147     }
 148     if (s->stage == BLOCK_REPLICATION_FAILOVER) {
 149         commit_job = &s->commit_job->job;
 150         assert(commit_job->aio_context == qemu_get_current_aio_context());
 151         job_cancel_sync(commit_job, false);
 152     }
 153
 154     if (s->mode == REPLICATION_MODE_SECONDARY) {
 155         g_free(s->top_id);
 156     }
 157
 158     replication_remove(s->rs);
 159 }
 160
 161 static void replication_child_perm(BlockDriverState *bs, BdrvChild *c,
 162                                    BdrvChildRole role,
 163                                    BlockReopenQueue *reopen_queue,
 164                                    uint64_t perm, uint64_t shared,
 165                                    uint64_t *nperm, uint64_t *nshared)
 166 {
 167     if (role & BDRV_CHILD_PRIMARY) {
 168         *nperm = BLK_PERM_CONSISTENT_READ;
 169     } else {
 170         *nperm = 0;
 171     }
 172
 173     if ((bs->open_flags & (BDRV_O_INACTIVE | BDRV_O_RDWR)) == BDRV_O_RDWR) {
 174         *nperm |= BLK_PERM_WRITE;
 175     }
 176     *nshared = BLK_PERM_CONSISTENT_READ
 177                | BLK_PERM_WRITE
 178                | BLK_PERM_WRITE_UNCHANGED;
 179     return;
 180 }
 181
 182 static int64_t coroutine_fn GRAPH_RDLOCK
 183 replication_co_getlength(BlockDriverState *bs)
 184 {
 185     return bdrv_co_getlength(bs->file->bs);
 186 }
 187
 188 static int replication_get_io_status(BDRVReplicationState *s)
 189 {
 190     switch (s->stage) {
 191     case BLOCK_REPLICATION_NONE:
 192         return -EIO;
 193     case BLOCK_REPLICATION_RUNNING:
 194         return 0;
 195     case BLOCK_REPLICATION_FAILOVER:
 196         return s->mode == REPLICATION_MODE_PRIMARY ? -EIO : 0;
 197     case BLOCK_REPLICATION_FAILOVER_FAILED:
 198         return s->mode == REPLICATION_MODE_PRIMARY ? -EIO : 1;
 199     case BLOCK_REPLICATION_DONE:
 200         /*
 201          * active commit job completes, and active disk and secondary_disk
 202          * is swapped, so we can operate bs->file directly
 203          */
 204         return s->mode == REPLICATION_MODE_PRIMARY ? -EIO : 0;
 205     default:
 206         abort();
 207     }
 208 }
 209
 210 static int replication_return_value(BDRVReplicationState *s, int ret)
 211 {
 212     if (s->mode == REPLICATION_MODE_SECONDARY) {
 213         return ret;
 214     }
 215
 216     if (ret < 0) {
 217         s->error = ret;
 218         ret = 0;
 219     }
 220
 221     return ret;
 222 }
 223
 224 static int coroutine_fn GRAPH_RDLOCK
 225 replication_co_readv(BlockDriverState *bs, int64_t sector_num,
 226                      int remaining_sectors, QEMUIOVector *qiov)
 227 {
 228     BDRVReplicationState *s = bs->opaque;
 229     int ret;
 230
 231     if (s->mode == REPLICATION_MODE_PRIMARY) {
 232         /* We only use it to forward primary write requests */
 233         return -EIO;
 234     }
 235
 236     ret = replication_get_io_status(s);
 237     if (ret < 0) {
 238         return ret;
 239     }
 240
 241     ret = bdrv_co_preadv(bs->file, sector_num * BDRV_SECTOR_SIZE,
 242                          remaining_sectors * BDRV_SECTOR_SIZE, qiov, 0);
 243
 244     return replication_return_value(s, ret);
 245 }
 246
 247 static int coroutine_fn GRAPH_RDLOCK
 248 replication_co_writev(BlockDriverState *bs, int64_t sector_num,
 249                       int remaining_sectors, QEMUIOVector *qiov, int flags)
 250 {
 251     BDRVReplicationState *s = bs->opaque;
 252     QEMUIOVector hd_qiov;
 253     uint64_t bytes_done = 0;
 254     BdrvChild *top = bs->file;
 255     BdrvChild *base = s->secondary_disk;
 256     BdrvChild *target;
 257     int ret;
 258     int64_t n;
 259
 260     ret = replication_get_io_status(s);
 261     if (ret < 0) {
 262         goto out;
 263     }
 264
 265     if (ret == 0) {
 266         ret = bdrv_co_pwritev(top, sector_num * BDRV_SECTOR_SIZE,
 267                               remaining_sectors * BDRV_SECTOR_SIZE, qiov, 0);
 268         return replication_return_value(s, ret);
 269     }
 270
 271     /*
 272      * Failover failed, only write to active disk if the sectors
 273      * have already been allocated in active disk/hidden disk.
 274      */
 275     qemu_iovec_init(&hd_qiov, qiov->niov);
 276     while (remaining_sectors > 0) {
 277         int64_t count;
 278
 279         ret = bdrv_is_allocated_above(top->bs, base->bs, false,
 280                                       sector_num * BDRV_SECTOR_SIZE,
 281                                       remaining_sectors * BDRV_SECTOR_SIZE,
 282                                       &count);
 283         if (ret < 0) {
 284             goto out1;
 285         }
 286
 287         assert(QEMU_IS_ALIGNED(count, BDRV_SECTOR_SIZE));
 288         n = count >> BDRV_SECTOR_BITS;
 289         qemu_iovec_reset(&hd_qiov);
 290         qemu_iovec_concat(&hd_qiov, qiov, bytes_done, count);
 291
 292         target = ret ? top : base;
 293         ret = bdrv_co_pwritev(target, sector_num * BDRV_SECTOR_SIZE,
 294                               n * BDRV_SECTOR_SIZE, &hd_qiov, 0);
 295         if (ret < 0) {
 296             goto out1;
 297         }
 298
 299         remaining_sectors -= n;
 300         sector_num += n;
 301         bytes_done += count;
 302     }
 303
 304 out1:
 305     qemu_iovec_destroy(&hd_qiov);
 306 out:
 307     return ret;
 308 }
 309
 310 static void secondary_do_checkpoint(BlockDriverState *bs, Error **errp)
 311 {
 312     BDRVReplicationState *s = bs->opaque;
 313     BdrvChild *active_disk = bs->file;
 314     Error *local_err = NULL;
 315     int ret;
 316
 317     if (!s->backup_job) {
 318         error_setg(errp, "Backup job was cancelled unexpectedly");
 319         return;
 320     }
 321
 322     backup_do_checkpoint(s->backup_job, &local_err);
 323     if (local_err) {
 324         error_propagate(errp, local_err);
 325         return;
 326     }
 327
 328     if (!active_disk->bs->drv) {
 329         error_setg(errp, "Active disk %s is ejected",
 330                    active_disk->bs->node_name);
 331         return;
 332     }
 333
 334     ret = bdrv_make_empty(active_disk, errp);
 335     if (ret < 0) {
 336         return;
 337     }
 338
 339     if (!s->hidden_disk->bs->drv) {
 340         error_setg(errp, "Hidden disk %s is ejected",
 341                    s->hidden_disk->bs->node_name);
 342         return;
 343     }
 344
 345     ret = bdrv_make_empty(s->hidden_disk, errp);
 346     if (ret < 0) {
 347         return;
 348     }
 349 }
 350
 351 /* This function is supposed to be called twice:
 352  * first with writable = true, then with writable = false.
 353  * The first call puts s->hidden_disk and s->secondary_disk in
 354  * r/w mode, and the second puts them back in their original state.
 355  */
 356 static void reopen_backing_file(BlockDriverState *bs, bool writable,
 357                                 Error **errp)
 358 {
 359     BDRVReplicationState *s = bs->opaque;
 360     BdrvChild *hidden_disk, *secondary_disk;
 361     BlockReopenQueue *reopen_queue = NULL;
 362
 363     /*
 364      * s->hidden_disk and s->secondary_disk may not be set yet, as they will
 365      * only be set after the children are writable.
 366      */
 367     hidden_disk = bs->file->bs->backing;
 368     secondary_disk = hidden_disk->bs->backing;
 369
 370     if (writable) {
 371         s->orig_hidden_read_only = bdrv_is_read_only(hidden_disk->bs);
 372         s->orig_secondary_read_only = bdrv_is_read_only(secondary_disk->bs);
 373     }
 374
 375     if (s->orig_hidden_read_only) {
 376         QDict *opts = qdict_new();
 377         qdict_put_bool(opts, BDRV_OPT_READ_ONLY, !writable);
 378         reopen_queue = bdrv_reopen_queue(reopen_queue, hidden_disk->bs,
 379                                          opts, true);
 380     }
 381
 382     if (s->orig_secondary_read_only) {
 383         QDict *opts = qdict_new();
 384         qdict_put_bool(opts, BDRV_OPT_READ_ONLY, !writable);
 385         reopen_queue = bdrv_reopen_queue(reopen_queue, secondary_disk->bs,
 386                                          opts, true);
 387     }
 388
 389     if (reopen_queue) {
 390         AioContext *ctx = bdrv_get_aio_context(bs);
 391         if (ctx != qemu_get_aio_context()) {
 392             aio_context_release(ctx);
 393         }
 394         bdrv_reopen_multiple(reopen_queue, errp);
 395         if (ctx != qemu_get_aio_context()) {
 396             aio_context_acquire(ctx);
 397         }
 398     }
 399 }
 400
 401 static void backup_job_cleanup(BlockDriverState *bs)
 402 {
 403     BDRVReplicationState *s = bs->opaque;
 404     BlockDriverState *top_bs;
 405
 406     s->backup_job = NULL;
 407
 408     top_bs = bdrv_lookup_bs(s->top_id, s->top_id, NULL);
 409     if (!top_bs) {
 410         return;
 411     }
 412     bdrv_op_unblock_all(top_bs, s->blocker);
 413     error_free(s->blocker);
 414     reopen_backing_file(bs, false, NULL);
 415 }
 416
 417 static void backup_job_completed(void *opaque, int ret)
 418 {
 419     BlockDriverState *bs = opaque;
 420     BDRVReplicationState *s = bs->opaque;
 421
 422     if (s->stage != BLOCK_REPLICATION_FAILOVER) {
 423         /* The backup job is cancelled unexpectedly */
 424         s->error = -EIO;
 425     }
 426
 427     backup_job_cleanup(bs);
 428 }
 429
 430 static bool check_top_bs(BlockDriverState *top_bs, BlockDriverState *bs)
 431 {
 432     BdrvChild *child;
 433
 434     /* The bs itself is the top_bs */
 435     if (top_bs == bs) {
 436         return true;
 437     }
 438
 439     /* Iterate over top_bs's children */
 440     QLIST_FOREACH(child, &top_bs->children, next) {
 441         if (child->bs == bs || check_top_bs(child->bs, bs)) {
 442             return true;
 443         }
 444     }
 445
 446     return false;
 447 }
 448
 449 static void replication_start(ReplicationState *rs, ReplicationMode mode,
 450                               Error **errp)
 451 {
 452     BlockDriverState *bs = rs->opaque;
 453     BDRVReplicationState *s;
 454     BlockDriverState *top_bs;
 455     BdrvChild *active_disk, *hidden_disk, *secondary_disk;
 456     int64_t active_length, hidden_length, disk_length;
 457     AioContext *aio_context;
 458     Error *local_err = NULL;
 459     BackupPerf perf = { .use_copy_range = true, .max_workers = 1 };
 460
 461     aio_context = bdrv_get_aio_context(bs);
 462     aio_context_acquire(aio_context);
 463     s = bs->opaque;
 464
 465     if (s->stage == BLOCK_REPLICATION_DONE ||
 466         s->stage == BLOCK_REPLICATION_FAILOVER) {
 467         /*
 468          * This case happens when a secondary is promoted to primary.
 469          * Ignore the request because the secondary side of replication
 470          * doesn't have to do anything anymore.
 471          */
 472         aio_context_release(aio_context);
 473         return;
 474     }
 475
 476     if (s->stage != BLOCK_REPLICATION_NONE) {
 477         error_setg(errp, "Block replication is running or done");
 478         aio_context_release(aio_context);
 479         return;
 480     }
 481
 482     if (s->mode != mode) {
 483         error_setg(errp, "The parameter mode's value is invalid, needs %d,"
 484                    " but got %d", s->mode, mode);
 485         aio_context_release(aio_context);
 486         return;
 487     }
 488
 489     switch (s->mode) {
 490     case REPLICATION_MODE_PRIMARY:
 491         break;
 492     case REPLICATION_MODE_SECONDARY:
 493         active_disk = bs->file;
 494         if (!active_disk || !active_disk->bs || !active_disk->bs->backing) {
 495             error_setg(errp, "Active disk doesn't have backing file");
 496             aio_context_release(aio_context);
 497             return;
 498         }
 499
 500         hidden_disk = active_disk->bs->backing;
 501         if (!hidden_disk->bs || !hidden_disk->bs->backing) {
 502             error_setg(errp, "Hidden disk doesn't have backing file");
 503             aio_context_release(aio_context);
 504             return;
 505         }
 506
 507         secondary_disk = hidden_disk->bs->backing;
 508         if (!secondary_disk->bs || !bdrv_has_blk(secondary_disk->bs)) {
 509             error_setg(errp, "The secondary disk doesn't have block backend");
 510             aio_context_release(aio_context);
 511             return;
 512         }
 513
 514         /* verify the length */
 515         active_length = bdrv_getlength(active_disk->bs);
 516         hidden_length = bdrv_getlength(hidden_disk->bs);
 517         disk_length = bdrv_getlength(secondary_disk->bs);
 518         if (active_length < 0 || hidden_length < 0 || disk_length < 0 ||
 519             active_length != hidden_length || hidden_length != disk_length) {
 520             error_setg(errp, "Active disk, hidden disk, secondary disk's length"
 521                        " are not the same");
 522             aio_context_release(aio_context);
 523             return;
 524         }
 525
 526         /* Must be true, or the bdrv_getlength() calls would have failed */
 527         assert(active_disk->bs->drv && hidden_disk->bs->drv);
 528
 529         if (!active_disk->bs->drv->bdrv_make_empty ||
 530             !hidden_disk->bs->drv->bdrv_make_empty) {
 531             error_setg(errp,
 532                        "Active disk or hidden disk doesn't support make_empty");
 533             aio_context_release(aio_context);
 534             return;
 535         }
 536
 537         /* reopen the backing file in r/w mode */
 538         reopen_backing_file(bs, true, &local_err);
 539         if (local_err) {
 540             error_propagate(errp, local_err);
 541             aio_context_release(aio_context);
 542             return;
 543         }
 544
 545         bdrv_ref(hidden_disk->bs);
 546         s->hidden_disk = bdrv_attach_child(bs, hidden_disk->bs, "hidden disk",
 547                                            &child_of_bds, BDRV_CHILD_DATA,
 548                                            &local_err);
 549         if (local_err) {
 550             error_propagate(errp, local_err);
 551             aio_context_release(aio_context);
 552             return;
 553         }
 554
 555         bdrv_ref(secondary_disk->bs);
 556         s->secondary_disk = bdrv_attach_child(bs, secondary_disk->bs,
 557                                               "secondary disk", &child_of_bds,
 558                                               BDRV_CHILD_DATA, &local_err);
 559         if (local_err) {
 560             error_propagate(errp, local_err);
 561             aio_context_release(aio_context);
 562             return;
 563         }
 564
 565         /* start backup job now */
 566         error_setg(&s->blocker,
 567                    "Block device is in use by internal backup job");
 568
 569         top_bs = bdrv_lookup_bs(s->top_id, s->top_id, NULL);
 570         if (!top_bs || !bdrv_is_root_node(top_bs) ||
 571             !check_top_bs(top_bs, bs)) {
 572             error_setg(errp, "No top_bs or it is invalid");
 573             reopen_backing_file(bs, false, NULL);
 574             aio_context_release(aio_context);
 575             return;
 576         }
 577         bdrv_op_block_all(top_bs, s->blocker);
 578         bdrv_op_unblock(top_bs, BLOCK_OP_TYPE_DATAPLANE, s->blocker);
 579
 580         s->backup_job = backup_job_create(
 581                                 NULL, s->secondary_disk->bs, s->hidden_disk->bs,
 582                                 0, MIRROR_SYNC_MODE_NONE, NULL, 0, false, NULL,
 583                                 &perf,
 584                                 BLOCKDEV_ON_ERROR_REPORT,
 585                                 BLOCKDEV_ON_ERROR_REPORT, JOB_INTERNAL,
 586                                 backup_job_completed, bs, NULL, &local_err);
 587         if (local_err) {
 588             error_propagate(errp, local_err);
 589             backup_job_cleanup(bs);
 590             aio_context_release(aio_context);
 591             return;
 592         }
 593         job_start(&s->backup_job->job);
 594         break;
 595     default:
 596         aio_context_release(aio_context);
 597         abort();
 598     }
 599
 600     s->stage = BLOCK_REPLICATION_RUNNING;
 601
 602     if (s->mode == REPLICATION_MODE_SECONDARY) {
 603         secondary_do_checkpoint(bs, errp);
 604     }
 605
 606     s->error = 0;
 607     aio_context_release(aio_context);
 608 }
 609
 610 static void replication_do_checkpoint(ReplicationState *rs, Error **errp)
 611 {
 612     BlockDriverState *bs = rs->opaque;
 613     BDRVReplicationState *s;
 614     AioContext *aio_context;
 615
 616     aio_context = bdrv_get_aio_context(bs);
 617     aio_context_acquire(aio_context);
 618     s = bs->opaque;
 619
 620     if (s->stage == BLOCK_REPLICATION_DONE ||
 621         s->stage == BLOCK_REPLICATION_FAILOVER) {
 622         /*
 623          * This case happens when a secondary was promoted to primary.
 624          * Ignore the request because the secondary side of replication
 625          * doesn't have to do anything anymore.
 626          */
 627         aio_context_release(aio_context);
 628         return;
 629     }
 630
 631     if (s->mode == REPLICATION_MODE_SECONDARY) {
 632         secondary_do_checkpoint(bs, errp);
 633     }
 634     aio_context_release(aio_context);
 635 }
 636
 637 static void replication_get_error(ReplicationState *rs, Error **errp)
 638 {
 639     BlockDriverState *bs = rs->opaque;
 640     BDRVReplicationState *s;
 641     AioContext *aio_context;
 642
 643     aio_context = bdrv_get_aio_context(bs);
 644     aio_context_acquire(aio_context);
 645     s = bs->opaque;
 646
 647     if (s->stage == BLOCK_REPLICATION_NONE) {
 648         error_setg(errp, "Block replication is not running");
 649         aio_context_release(aio_context);
 650         return;
 651     }
 652
 653     if (s->error) {
 654         error_setg(errp, "I/O error occurred");
 655         aio_context_release(aio_context);
 656         return;
 657     }
 658     aio_context_release(aio_context);
 659 }
 660
 661 static void replication_done(void *opaque, int ret)
 662 {
 663     BlockDriverState *bs = opaque;
 664     BDRVReplicationState *s = bs->opaque;
 665
 666     if (ret == 0) {
 667         s->stage = BLOCK_REPLICATION_DONE;
 668
 669         bdrv_unref_child(bs, s->secondary_disk);
 670         s->secondary_disk = NULL;
 671         bdrv_unref_child(bs, s->hidden_disk);
 672         s->hidden_disk = NULL;
 673         s->error = 0;
 674     } else {
 675         s->stage = BLOCK_REPLICATION_FAILOVER_FAILED;
 676         s->error = -EIO;
 677     }
 678 }
 679
 680 static void replication_stop(ReplicationState *rs, bool failover, Error **errp)
 681 {
 682     BlockDriverState *bs = rs->opaque;
 683     BDRVReplicationState *s;
 684     AioContext *aio_context;
 685
 686     aio_context = bdrv_get_aio_context(bs);
 687     aio_context_acquire(aio_context);
 688     s = bs->opaque;
 689
 690     if (s->stage == BLOCK_REPLICATION_DONE ||
 691         s->stage == BLOCK_REPLICATION_FAILOVER) {
 692         /*
 693          * This case happens when a secondary was promoted to primary.
 694          * Ignore the request because the secondary side of replication
 695          * doesn't have to do anything anymore.
 696          */
 697         aio_context_release(aio_context);
 698         return;
 699     }
 700
 701     if (s->stage != BLOCK_REPLICATION_RUNNING) {
 702         error_setg(errp, "Block replication is not running");
 703         aio_context_release(aio_context);
 704         return;
 705     }
 706
 707     switch (s->mode) {
 708     case REPLICATION_MODE_PRIMARY:
 709         s->stage = BLOCK_REPLICATION_DONE;
 710         s->error = 0;
 711         break;
 712     case REPLICATION_MODE_SECONDARY:
 713         /*
 714          * This BDS will be closed, and the job should be completed
 715          * before the BDS is closed, because we will access hidden
 716          * disk, secondary disk in backup_job_completed().
 717          */
 718         if (s->backup_job) {
 719             aio_context_release(aio_context);
 720             job_cancel_sync(&s->backup_job->job, true);
 721             aio_context_acquire(aio_context);
 722         }
 723
 724         if (!failover) {
 725             secondary_do_checkpoint(bs, errp);
 726             s->stage = BLOCK_REPLICATION_DONE;
 727             aio_context_release(aio_context);
 728             return;
 729         }
 730
 731         s->stage = BLOCK_REPLICATION_FAILOVER;
 732         s->commit_job = commit_active_start(
 733                             NULL, bs->file->bs, s->secondary_disk->bs,
 734                             JOB_INTERNAL, 0, BLOCKDEV_ON_ERROR_REPORT,
 735                             NULL, replication_done, bs, true, errp);
 736         break;
 737     default:
 738         aio_context_release(aio_context);
 739         abort();
 740     }
 741     aio_context_release(aio_context);
 742 }
 743
 744 static const char *const replication_strong_runtime_opts[] = {
 745     REPLICATION_MODE,
 746     REPLICATION_TOP_ID,
 747
 748     NULL
 749 };
 750
 751 static BlockDriver bdrv_replication = {
 752     .format_name                = "replication",
 753     .instance_size              = sizeof(BDRVReplicationState),
 754
 755     .bdrv_open                  = replication_open,
 756     .bdrv_close                 = replication_close,
 757     .bdrv_child_perm            = replication_child_perm,
 758
 759     .bdrv_co_getlength          = replication_co_getlength,
 760     .bdrv_co_readv              = replication_co_readv,
 761     .bdrv_co_writev             = replication_co_writev,
 762
 763     .is_filter                  = true,
 764
 765     .has_variable_length        = true,
 766     .strong_runtime_opts        = replication_strong_runtime_opts,
 767 };
 768
 769 static void bdrv_replication_init(void)
 770 {
 771     bdrv_register(&bdrv_replication);
 772 }
 773
 774 block_init(bdrv_replication_init);