blockjob.c

   1 /*
   2  * QEMU System Emulator block driver
   3  *
   4  * Copyright (c) 2011 IBM Corp.
   5  * Copyright (c) 2012 Red Hat, Inc.
   6  *
   7  * Permission is hereby granted, free of charge, to any person obtaining a copy
   8  * of this software and associated documentation files (the "Software"), to deal
   9  * in the Software without restriction, including without limitation the rights
  10  * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
  11  * copies of the Software, and to permit persons to whom the Software is
  12  * furnished to do so, subject to the following conditions:
  13  *
  14  * The above copyright notice and this permission notice shall be included in
  15  * all copies or substantial portions of the Software.
  16  *
  17  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
  18  * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
  19  * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
  20  * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
  21  * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
  22  * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
  23  * THE SOFTWARE.
  24  */
  25
  26 #include "qemu/osdep.h"
  27 #include "qemu-common.h"
  28 #include "block/block.h"
  29 #include "block/blockjob_int.h"
  30 #include "block/block_int.h"
  31 #include "sysemu/block-backend.h"
  32 #include "qapi/qmp/qerror.h"
  33 #include "qapi/qmp/qjson.h"
  34 #include "qemu/coroutine.h"
  35 #include "qemu/id.h"
  36 #include "qmp-commands.h"
  37 #include "qemu/timer.h"
  38 #include "qapi-event.h"
  39
  40 /* Right now, this mutex is only needed to synchronize accesses to job->busy
  41  * and job->sleep_timer, such as concurrent calls to block_job_do_yield and
  42  * block_job_enter. */
  43 static QemuMutex block_job_mutex;
  44
  45 static void block_job_lock(void)
  46 {
  47     qemu_mutex_lock(&block_job_mutex);
  48 }
  49
  50 static void block_job_unlock(void)
  51 {
  52     qemu_mutex_unlock(&block_job_mutex);
  53 }
  54
  55 static void __attribute__((__constructor__)) block_job_init(void)
  56 {
  57     qemu_mutex_init(&block_job_mutex);
  58 }
  59
  60 static void block_job_event_cancelled(BlockJob *job);
  61 static void block_job_event_completed(BlockJob *job, const char *msg);
  62
  63 /* Transactional group of block jobs */
  64 struct BlockJobTxn {
  65
  66     /* Is this txn being cancelled? */
  67     bool aborting;
  68
  69     /* List of jobs */
  70     QLIST_HEAD(, BlockJob) jobs;
  71
  72     /* Reference count */
  73     int refcnt;
  74 };
  75
  76 static QLIST_HEAD(, BlockJob) block_jobs = QLIST_HEAD_INITIALIZER(block_jobs);
  77
  78 /*
  79  * The block job API is composed of two categories of functions.
  80  *
  81  * The first includes functions used by the monitor.  The monitor is
  82  * peculiar in that it accesses the block job list with block_job_get, and
  83  * therefore needs consistency across block_job_get and the actual operation
  84  * (e.g. block_job_set_speed).  The consistency is achieved with
  85  * aio_context_acquire/release.  These functions are declared in blockjob.h.
  86  *
  87  * The second includes functions used by the block job drivers and sometimes
  88  * by the core block layer.  These do not care about locking, because the
  89  * whole coroutine runs under the AioContext lock, and are declared in
  90  * blockjob_int.h.
  91  */
  92
  93 BlockJob *block_job_next(BlockJob *job)
  94 {
  95     if (!job) {
  96         return QLIST_FIRST(&block_jobs);
  97     }
  98     return QLIST_NEXT(job, job_list);
  99 }
 100
 101 BlockJob *block_job_get(const char *id)
 102 {
 103     BlockJob *job;
 104
 105     QLIST_FOREACH(job, &block_jobs, job_list) {
 106         if (job->id && !strcmp(id, job->id)) {
 107             return job;
 108         }
 109     }
 110
 111     return NULL;
 112 }
 113
 114 BlockJobTxn *block_job_txn_new(void)
 115 {
 116     BlockJobTxn *txn = g_new0(BlockJobTxn, 1);
 117     QLIST_INIT(&txn->jobs);
 118     txn->refcnt = 1;
 119     return txn;
 120 }
 121
 122 static void block_job_txn_ref(BlockJobTxn *txn)
 123 {
 124     txn->refcnt++;
 125 }
 126
 127 void block_job_txn_unref(BlockJobTxn *txn)
 128 {
 129     if (txn && --txn->refcnt == 0) {
 130         g_free(txn);
 131     }
 132 }
 133
 134 void block_job_txn_add_job(BlockJobTxn *txn, BlockJob *job)
 135 {
 136     if (!txn) {
 137         return;
 138     }
 139
 140     assert(!job->txn);
 141     job->txn = txn;
 142
 143     QLIST_INSERT_HEAD(&txn->jobs, job, txn_list);
 144     block_job_txn_ref(txn);
 145 }
 146
 147 static void block_job_pause(BlockJob *job)
 148 {
 149     job->pause_count++;
 150 }
 151
 152 static void block_job_resume(BlockJob *job)
 153 {
 154     assert(job->pause_count > 0);
 155     job->pause_count--;
 156     if (job->pause_count) {
 157         return;
 158     }
 159     block_job_enter(job);
 160 }
 161
 162 void block_job_ref(BlockJob *job)
 163 {
 164     ++job->refcnt;
 165 }
 166
 167 static void block_job_attached_aio_context(AioContext *new_context,
 168                                            void *opaque);
 169 static void block_job_detach_aio_context(void *opaque);
 170
 171 void block_job_unref(BlockJob *job)
 172 {
 173     if (--job->refcnt == 0) {
 174         BlockDriverState *bs = blk_bs(job->blk);
 175         QLIST_REMOVE(job, job_list);
 176         bs->job = NULL;
 177         block_job_remove_all_bdrv(job);
 178         blk_remove_aio_context_notifier(job->blk,
 179                                         block_job_attached_aio_context,
 180                                         block_job_detach_aio_context, job);
 181         blk_unref(job->blk);
 182         error_free(job->blocker);
 183         g_free(job->id);
 184         assert(!timer_pending(&job->sleep_timer));
 185         g_free(job);
 186     }
 187 }
 188
 189 static void block_job_attached_aio_context(AioContext *new_context,
 190                                            void *opaque)
 191 {
 192     BlockJob *job = opaque;
 193
 194     if (job->driver->attached_aio_context) {
 195         job->driver->attached_aio_context(job, new_context);
 196     }
 197
 198     block_job_resume(job);
 199 }
 200
 201 static void block_job_drain(BlockJob *job)
 202 {
 203     /* If job is !job->busy this kicks it into the next pause point. */
 204     block_job_enter(job);
 205
 206     blk_drain(job->blk);
 207     if (job->driver->drain) {
 208         job->driver->drain(job);
 209     }
 210 }
 211
 212 static void block_job_detach_aio_context(void *opaque)
 213 {
 214     BlockJob *job = opaque;
 215
 216     /* In case the job terminates during aio_poll()... */
 217     block_job_ref(job);
 218
 219     block_job_pause(job);
 220
 221     while (!job->paused && !job->completed) {
 222         block_job_drain(job);
 223     }
 224
 225     block_job_unref(job);
 226 }
 227
 228 static char *child_job_get_parent_desc(BdrvChild *c)
 229 {
 230     BlockJob *job = c->opaque;
 231     return g_strdup_printf("%s job '%s'",
 232                            BlockJobType_str(job->driver->job_type),
 233                            job->id);
 234 }
 235
 236 static const BdrvChildRole child_job = {
 237     .get_parent_desc    = child_job_get_parent_desc,
 238     .stay_at_node       = true,
 239 };
 240
 241 static void block_job_drained_begin(void *opaque)
 242 {
 243     BlockJob *job = opaque;
 244     block_job_pause(job);
 245 }
 246
 247 static void block_job_drained_end(void *opaque)
 248 {
 249     BlockJob *job = opaque;
 250     block_job_resume(job);
 251 }
 252
 253 static const BlockDevOps block_job_dev_ops = {
 254     .drained_begin = block_job_drained_begin,
 255     .drained_end = block_job_drained_end,
 256 };
 257
 258 void block_job_remove_all_bdrv(BlockJob *job)
 259 {
 260     GSList *l;
 261     for (l = job->nodes; l; l = l->next) {
 262         BdrvChild *c = l->data;
 263         bdrv_op_unblock_all(c->bs, job->blocker);
 264         bdrv_root_unref_child(c);
 265     }
 266     g_slist_free(job->nodes);
 267     job->nodes = NULL;
 268 }
 269
 270 int block_job_add_bdrv(BlockJob *job, const char *name, BlockDriverState *bs,
 271                        uint64_t perm, uint64_t shared_perm, Error **errp)
 272 {
 273     BdrvChild *c;
 274
 275     c = bdrv_root_attach_child(bs, name, &child_job, perm, shared_perm,
 276                                job, errp);
 277     if (c == NULL) {
 278         return -EPERM;
 279     }
 280
 281     job->nodes = g_slist_prepend(job->nodes, c);
 282     bdrv_ref(bs);
 283     bdrv_op_block_all(bs, job->blocker);
 284
 285     return 0;
 286 }
 287
 288 bool block_job_is_internal(BlockJob *job)
 289 {
 290     return (job->id == NULL);
 291 }
 292
 293 static bool block_job_started(BlockJob *job)
 294 {
 295     return job->co;
 296 }
 297
 298 /**
 299  * All jobs must allow a pause point before entering their job proper. This
 300  * ensures that jobs can be paused prior to being started, then resumed later.
 301  */
 302 static void coroutine_fn block_job_co_entry(void *opaque)
 303 {
 304     BlockJob *job = opaque;
 305
 306     assert(job && job->driver && job->driver->start);
 307     block_job_pause_point(job);
 308     job->driver->start(job);
 309 }
 310
 311 static void block_job_sleep_timer_cb(void *opaque)
 312 {
 313     BlockJob *job = opaque;
 314
 315     block_job_enter(job);
 316 }
 317
 318 void block_job_start(BlockJob *job)
 319 {
 320     assert(job && !block_job_started(job) && job->paused &&
 321            job->driver && job->driver->start);
 322     job->co = qemu_coroutine_create(block_job_co_entry, job);
 323     job->pause_count--;
 324     job->busy = true;
 325     job->paused = false;
 326     bdrv_coroutine_enter(blk_bs(job->blk), job->co);
 327 }
 328
 329 static void block_job_completed_single(BlockJob *job)
 330 {
 331     assert(job->completed);
 332
 333     if (!job->ret) {
 334         if (job->driver->commit) {
 335             job->driver->commit(job);
 336         }
 337     } else {
 338         if (job->driver->abort) {
 339             job->driver->abort(job);
 340         }
 341     }
 342     if (job->driver->clean) {
 343         job->driver->clean(job);
 344     }
 345
 346     if (job->cb) {
 347         job->cb(job->opaque, job->ret);
 348     }
 349
 350     /* Emit events only if we actually started */
 351     if (block_job_started(job)) {
 352         if (block_job_is_cancelled(job)) {
 353             block_job_event_cancelled(job);
 354         } else {
 355             const char *msg = NULL;
 356             if (job->ret < 0) {
 357                 msg = strerror(-job->ret);
 358             }
 359             block_job_event_completed(job, msg);
 360         }
 361     }
 362
 363     if (job->txn) {
 364         QLIST_REMOVE(job, txn_list);
 365         block_job_txn_unref(job->txn);
 366     }
 367     block_job_unref(job);
 368 }
 369
 370 static void block_job_cancel_async(BlockJob *job)
 371 {
 372     if (job->iostatus != BLOCK_DEVICE_IO_STATUS_OK) {
 373         block_job_iostatus_reset(job);
 374     }
 375     if (job->user_paused) {
 376         /* Do not call block_job_enter here, the caller will handle it.  */
 377         job->user_paused = false;
 378         job->pause_count--;
 379     }
 380     job->cancelled = true;
 381 }
 382
 383 static int block_job_finish_sync(BlockJob *job,
 384                                  void (*finish)(BlockJob *, Error **errp),
 385                                  Error **errp)
 386 {
 387     Error *local_err = NULL;
 388     int ret;
 389
 390     assert(blk_bs(job->blk)->job == job);
 391
 392     block_job_ref(job);
 393
 394     if (finish) {
 395         finish(job, &local_err);
 396     }
 397     if (local_err) {
 398         error_propagate(errp, local_err);
 399         block_job_unref(job);
 400         return -EBUSY;
 401     }
 402     /* block_job_drain calls block_job_enter, and it should be enough to
 403      * induce progress until the job completes or moves to the main thread.
 404     */
 405     while (!job->deferred_to_main_loop && !job->completed) {
 406         block_job_drain(job);
 407     }
 408     while (!job->completed) {
 409         aio_poll(qemu_get_aio_context(), true);
 410     }
 411     ret = (job->cancelled && job->ret == 0) ? -ECANCELED : job->ret;
 412     block_job_unref(job);
 413     return ret;
 414 }
 415
 416 static void block_job_completed_txn_abort(BlockJob *job)
 417 {
 418     AioContext *ctx;
 419     BlockJobTxn *txn = job->txn;
 420     BlockJob *other_job;
 421
 422     if (txn->aborting) {
 423         /*
 424          * We are cancelled by another job, which will handle everything.
 425          */
 426         return;
 427     }
 428     txn->aborting = true;
 429     block_job_txn_ref(txn);
 430
 431     /* We are the first failed job. Cancel other jobs. */
 432     QLIST_FOREACH(other_job, &txn->jobs, txn_list) {
 433         ctx = blk_get_aio_context(other_job->blk);
 434         aio_context_acquire(ctx);
 435     }
 436
 437     /* Other jobs are effectively cancelled by us, set the status for
 438      * them; this job, however, may or may not be cancelled, depending
 439      * on the caller, so leave it. */
 440     QLIST_FOREACH(other_job, &txn->jobs, txn_list) {
 441         if (other_job != job) {
 442             block_job_cancel_async(other_job);
 443         }
 444     }
 445     while (!QLIST_EMPTY(&txn->jobs)) {
 446         other_job = QLIST_FIRST(&txn->jobs);
 447         ctx = blk_get_aio_context(other_job->blk);
 448         if (!other_job->completed) {
 449             assert(other_job->cancelled);
 450             block_job_finish_sync(other_job, NULL, NULL);
 451         }
 452         block_job_completed_single(other_job);
 453         aio_context_release(ctx);
 454     }
 455
 456     block_job_txn_unref(txn);
 457 }
 458
 459 static void block_job_completed_txn_success(BlockJob *job)
 460 {
 461     AioContext *ctx;
 462     BlockJobTxn *txn = job->txn;
 463     BlockJob *other_job, *next;
 464     /*
 465      * Successful completion, see if there are other running jobs in this
 466      * txn.
 467      */
 468     QLIST_FOREACH(other_job, &txn->jobs, txn_list) {
 469         if (!other_job->completed) {
 470             return;
 471         }
 472     }
 473     /* We are the last completed job, commit the transaction. */
 474     QLIST_FOREACH_SAFE(other_job, &txn->jobs, txn_list, next) {
 475         ctx = blk_get_aio_context(other_job->blk);
 476         aio_context_acquire(ctx);
 477         assert(other_job->ret == 0);
 478         block_job_completed_single(other_job);
 479         aio_context_release(ctx);
 480     }
 481 }
 482
 483 void block_job_set_speed(BlockJob *job, int64_t speed, Error **errp)
 484 {
 485     Error *local_err = NULL;
 486
 487     if (!job->driver->set_speed) {
 488         error_setg(errp, QERR_UNSUPPORTED);
 489         return;
 490     }
 491     job->driver->set_speed(job, speed, &local_err);
 492     if (local_err) {
 493         error_propagate(errp, local_err);
 494         return;
 495     }
 496
 497     job->speed = speed;
 498 }
 499
 500 void block_job_complete(BlockJob *job, Error **errp)
 501 {
 502     /* Should not be reachable via external interface for internal jobs */
 503     assert(job->id);
 504     if (job->pause_count || job->cancelled ||
 505         !block_job_started(job) || !job->driver->complete) {
 506         error_setg(errp, "The active block job '%s' cannot be completed",
 507                    job->id);
 508         return;
 509     }
 510
 511     job->driver->complete(job, errp);
 512 }
 513
 514 void block_job_user_pause(BlockJob *job)
 515 {
 516     job->user_paused = true;
 517     block_job_pause(job);
 518 }
 519
 520 bool block_job_user_paused(BlockJob *job)
 521 {
 522     return job->user_paused;
 523 }
 524
 525 void block_job_user_resume(BlockJob *job)
 526 {
 527     if (job && job->user_paused && job->pause_count > 0) {
 528         block_job_iostatus_reset(job);
 529         job->user_paused = false;
 530         block_job_resume(job);
 531     }
 532 }
 533
 534 void block_job_cancel(BlockJob *job)
 535 {
 536     if (block_job_started(job)) {
 537         block_job_cancel_async(job);
 538         block_job_enter(job);
 539     } else {
 540         block_job_completed(job, -ECANCELED);
 541     }
 542 }
 543
 544 /* A wrapper around block_job_cancel() taking an Error ** parameter so it may be
 545  * used with block_job_finish_sync() without the need for (rather nasty)
 546  * function pointer casts there. */
 547 static void block_job_cancel_err(BlockJob *job, Error **errp)
 548 {
 549     block_job_cancel(job);
 550 }
 551
 552 int block_job_cancel_sync(BlockJob *job)
 553 {
 554     return block_job_finish_sync(job, &block_job_cancel_err, NULL);
 555 }
 556
 557 void block_job_cancel_sync_all(void)
 558 {
 559     BlockJob *job;
 560     AioContext *aio_context;
 561
 562     while ((job = QLIST_FIRST(&block_jobs))) {
 563         aio_context = blk_get_aio_context(job->blk);
 564         aio_context_acquire(aio_context);
 565         block_job_cancel_sync(job);
 566         aio_context_release(aio_context);
 567     }
 568 }
 569
 570 int block_job_complete_sync(BlockJob *job, Error **errp)
 571 {
 572     return block_job_finish_sync(job, &block_job_complete, errp);
 573 }
 574
 575 BlockJobInfo *block_job_query(BlockJob *job, Error **errp)
 576 {
 577     BlockJobInfo *info;
 578
 579     if (block_job_is_internal(job)) {
 580         error_setg(errp, "Cannot query QEMU internal jobs");
 581         return NULL;
 582     }
 583     info = g_new0(BlockJobInfo, 1);
 584     info->type      = g_strdup(BlockJobType_str(job->driver->job_type));
 585     info->device    = g_strdup(job->id);
 586     info->len       = job->len;
 587     info->busy      = atomic_read(&job->busy);
 588     info->paused    = job->pause_count > 0;
 589     info->offset    = job->offset;
 590     info->speed     = job->speed;
 591     info->io_status = job->iostatus;
 592     info->ready     = job->ready;
 593     return info;
 594 }
 595
 596 static void block_job_iostatus_set_err(BlockJob *job, int error)
 597 {
 598     if (job->iostatus == BLOCK_DEVICE_IO_STATUS_OK) {
 599         job->iostatus = error == ENOSPC ? BLOCK_DEVICE_IO_STATUS_NOSPACE :
 600                                           BLOCK_DEVICE_IO_STATUS_FAILED;
 601     }
 602 }
 603
 604 static void block_job_event_cancelled(BlockJob *job)
 605 {
 606     if (block_job_is_internal(job)) {
 607         return;
 608     }
 609
 610     qapi_event_send_block_job_cancelled(job->driver->job_type,
 611                                         job->id,
 612                                         job->len,
 613                                         job->offset,
 614                                         job->speed,
 615                                         &error_abort);
 616 }
 617
 618 static void block_job_event_completed(BlockJob *job, const char *msg)
 619 {
 620     if (block_job_is_internal(job)) {
 621         return;
 622     }
 623
 624     qapi_event_send_block_job_completed(job->driver->job_type,
 625                                         job->id,
 626                                         job->len,
 627                                         job->offset,
 628                                         job->speed,
 629                                         !!msg,
 630                                         msg,
 631                                         &error_abort);
 632 }
 633
 634 /*
 635  * API for block job drivers and the block layer.  These functions are
 636  * declared in blockjob_int.h.
 637  */
 638
 639 void *block_job_create(const char *job_id, const BlockJobDriver *driver,
 640                        BlockDriverState *bs, uint64_t perm,
 641                        uint64_t shared_perm, int64_t speed, int flags,
 642                        BlockCompletionFunc *cb, void *opaque, Error **errp)
 643 {
 644     BlockBackend *blk;
 645     BlockJob *job;
 646     int ret;
 647
 648     if (bs->job) {
 649         error_setg(errp, QERR_DEVICE_IN_USE, bdrv_get_device_name(bs));
 650         return NULL;
 651     }
 652
 653     if (job_id == NULL && !(flags & BLOCK_JOB_INTERNAL)) {
 654         job_id = bdrv_get_device_name(bs);
 655         if (!*job_id) {
 656             error_setg(errp, "An explicit job ID is required for this node");
 657             return NULL;
 658         }
 659     }
 660
 661     if (job_id) {
 662         if (flags & BLOCK_JOB_INTERNAL) {
 663             error_setg(errp, "Cannot specify job ID for internal block job");
 664             return NULL;
 665         }
 666
 667         if (!id_wellformed(job_id)) {
 668             error_setg(errp, "Invalid job ID '%s'", job_id);
 669             return NULL;
 670         }
 671
 672         if (block_job_get(job_id)) {
 673             error_setg(errp, "Job ID '%s' already in use", job_id);
 674             return NULL;
 675         }
 676     }
 677
 678     blk = blk_new(perm, shared_perm);
 679     ret = blk_insert_bs(blk, bs, errp);
 680     if (ret < 0) {
 681         blk_unref(blk);
 682         return NULL;
 683     }
 684
 685     job = g_malloc0(driver->instance_size);
 686     job->driver        = driver;
 687     job->id            = g_strdup(job_id);
 688     job->blk           = blk;
 689     job->cb            = cb;
 690     job->opaque        = opaque;
 691     job->busy          = false;
 692     job->paused        = true;
 693     job->pause_count   = 1;
 694     job->refcnt        = 1;
 695     aio_timer_init(qemu_get_aio_context(), &job->sleep_timer,
 696                    QEMU_CLOCK_REALTIME, SCALE_NS,
 697                    block_job_sleep_timer_cb, job);
 698
 699     error_setg(&job->blocker, "block device is in use by block job: %s",
 700                BlockJobType_str(driver->job_type));
 701     block_job_add_bdrv(job, "main node", bs, 0, BLK_PERM_ALL, &error_abort);
 702     bs->job = job;
 703
 704     blk_set_dev_ops(blk, &block_job_dev_ops, job);
 705     bdrv_op_unblock(bs, BLOCK_OP_TYPE_DATAPLANE, job->blocker);
 706
 707     QLIST_INSERT_HEAD(&block_jobs, job, job_list);
 708
 709     blk_add_aio_context_notifier(blk, block_job_attached_aio_context,
 710                                  block_job_detach_aio_context, job);
 711
 712     /* Only set speed when necessary to avoid NotSupported error */
 713     if (speed != 0) {
 714         Error *local_err = NULL;
 715
 716         block_job_set_speed(job, speed, &local_err);
 717         if (local_err) {
 718             block_job_unref(job);
 719             error_propagate(errp, local_err);
 720             return NULL;
 721         }
 722     }
 723     return job;
 724 }
 725
 726 void block_job_pause_all(void)
 727 {
 728     BlockJob *job = NULL;
 729     while ((job = block_job_next(job))) {
 730         AioContext *aio_context = blk_get_aio_context(job->blk);
 731
 732         aio_context_acquire(aio_context);
 733         block_job_ref(job);
 734         block_job_pause(job);
 735         aio_context_release(aio_context);
 736     }
 737 }
 738
 739 void block_job_early_fail(BlockJob *job)
 740 {
 741     block_job_unref(job);
 742 }
 743
 744 void block_job_completed(BlockJob *job, int ret)
 745 {
 746     assert(blk_bs(job->blk)->job == job);
 747     assert(!job->completed);
 748     job->completed = true;
 749     job->ret = ret;
 750     if (!job->txn) {
 751         block_job_completed_single(job);
 752     } else if (ret < 0 || block_job_is_cancelled(job)) {
 753         block_job_completed_txn_abort(job);
 754     } else {
 755         block_job_completed_txn_success(job);
 756     }
 757 }
 758
 759 static bool block_job_should_pause(BlockJob *job)
 760 {
 761     return job->pause_count > 0;
 762 }
 763
 764 /* Yield, and schedule a timer to reenter the coroutine after @ns nanoseconds.
 765  * Reentering the job coroutine with block_job_enter() before the timer has
 766  * expired is allowed and cancels the timer.
 767  *
 768  * If @ns is (uint64_t) -1, no timer is scheduled and block_job_enter() must be
 769  * called explicitly. */
 770 static void block_job_do_yield(BlockJob *job, uint64_t ns)
 771 {
 772     block_job_lock();
 773     if (ns != -1) {
 774         timer_mod(&job->sleep_timer, ns);
 775     }
 776     job->busy = false;
 777     block_job_unlock();
 778     qemu_coroutine_yield();
 779
 780     /* Set by block_job_enter before re-entering the coroutine.  */
 781     assert(job->busy);
 782 }
 783
 784 void coroutine_fn block_job_pause_point(BlockJob *job)
 785 {
 786     assert(job && block_job_started(job));
 787
 788     if (!block_job_should_pause(job)) {
 789         return;
 790     }
 791     if (block_job_is_cancelled(job)) {
 792         return;
 793     }
 794
 795     if (job->driver->pause) {
 796         job->driver->pause(job);
 797     }
 798
 799     if (block_job_should_pause(job) && !block_job_is_cancelled(job)) {
 800         job->paused = true;
 801         block_job_do_yield(job, -1);
 802         job->paused = false;
 803     }
 804
 805     if (job->driver->resume) {
 806         job->driver->resume(job);
 807     }
 808 }
 809
 810 void block_job_resume_all(void)
 811 {
 812     BlockJob *job, *next;
 813
 814     QLIST_FOREACH_SAFE(job, &block_jobs, job_list, next) {
 815         AioContext *aio_context = blk_get_aio_context(job->blk);
 816
 817         aio_context_acquire(aio_context);
 818         block_job_resume(job);
 819         block_job_unref(job);
 820         aio_context_release(aio_context);
 821     }
 822 }
 823
 824 void block_job_enter(BlockJob *job)
 825 {
 826     if (!block_job_started(job)) {
 827         return;
 828     }
 829     if (job->deferred_to_main_loop) {
 830         return;
 831     }
 832
 833     block_job_lock();
 834     if (job->busy) {
 835         block_job_unlock();
 836         return;
 837     }
 838
 839     assert(!job->deferred_to_main_loop);
 840     timer_del(&job->sleep_timer);
 841     job->busy = true;
 842     block_job_unlock();
 843     aio_co_wake(job->co);
 844 }
 845
 846 bool block_job_is_cancelled(BlockJob *job)
 847 {
 848     return job->cancelled;
 849 }
 850
 851 void block_job_sleep_ns(BlockJob *job, int64_t ns)
 852 {
 853     assert(job->busy);
 854
 855     /* Check cancellation *before* setting busy = false, too!  */
 856     if (block_job_is_cancelled(job)) {
 857         return;
 858     }
 859
 860     if (!block_job_should_pause(job)) {
 861         block_job_do_yield(job, qemu_clock_get_ns(QEMU_CLOCK_REALTIME) + ns);
 862     }
 863
 864     block_job_pause_point(job);
 865 }
 866
 867 void block_job_yield(BlockJob *job)
 868 {
 869     assert(job->busy);
 870
 871     /* Check cancellation *before* setting busy = false, too!  */
 872     if (block_job_is_cancelled(job)) {
 873         return;
 874     }
 875
 876     if (!block_job_should_pause(job)) {
 877         block_job_do_yield(job, -1);
 878     }
 879
 880     block_job_pause_point(job);
 881 }
 882
 883 void block_job_iostatus_reset(BlockJob *job)
 884 {
 885     if (job->iostatus == BLOCK_DEVICE_IO_STATUS_OK) {
 886         return;
 887     }
 888     assert(job->user_paused && job->pause_count > 0);
 889     job->iostatus = BLOCK_DEVICE_IO_STATUS_OK;
 890 }
 891
 892 void block_job_event_ready(BlockJob *job)
 893 {
 894     job->ready = true;
 895
 896     if (block_job_is_internal(job)) {
 897         return;
 898     }
 899
 900     qapi_event_send_block_job_ready(job->driver->job_type,
 901                                     job->id,
 902                                     job->len,
 903                                     job->offset,
 904                                     job->speed, &error_abort);
 905 }
 906
 907 BlockErrorAction block_job_error_action(BlockJob *job, BlockdevOnError on_err,
 908                                         int is_read, int error)
 909 {
 910     BlockErrorAction action;
 911
 912     switch (on_err) {
 913     case BLOCKDEV_ON_ERROR_ENOSPC:
 914     case BLOCKDEV_ON_ERROR_AUTO:
 915         action = (error == ENOSPC) ?
 916                  BLOCK_ERROR_ACTION_STOP : BLOCK_ERROR_ACTION_REPORT;
 917         break;
 918     case BLOCKDEV_ON_ERROR_STOP:
 919         action = BLOCK_ERROR_ACTION_STOP;
 920         break;
 921     case BLOCKDEV_ON_ERROR_REPORT:
 922         action = BLOCK_ERROR_ACTION_REPORT;
 923         break;
 924     case BLOCKDEV_ON_ERROR_IGNORE:
 925         action = BLOCK_ERROR_ACTION_IGNORE;
 926         break;
 927     default:
 928         abort();
 929     }
 930     if (!block_job_is_internal(job)) {
 931         qapi_event_send_block_job_error(job->id,
 932                                         is_read ? IO_OPERATION_TYPE_READ :
 933                                         IO_OPERATION_TYPE_WRITE,
 934                                         action, &error_abort);
 935     }
 936     if (action == BLOCK_ERROR_ACTION_STOP) {
 937         /* make the pause user visible, which will be resumed from QMP. */
 938         block_job_user_pause(job);
 939         block_job_iostatus_set_err(job, error);
 940     }
 941     return action;
 942 }
 943
 944 typedef struct {
 945     BlockJob *job;
 946     AioContext *aio_context;
 947     BlockJobDeferToMainLoopFn *fn;
 948     void *opaque;
 949 } BlockJobDeferToMainLoopData;
 950
 951 static void block_job_defer_to_main_loop_bh(void *opaque)
 952 {
 953     BlockJobDeferToMainLoopData *data = opaque;
 954     AioContext *aio_context;
 955
 956     /* Prevent race with block_job_defer_to_main_loop() */
 957     aio_context_acquire(data->aio_context);
 958
 959     /* Fetch BDS AioContext again, in case it has changed */
 960     aio_context = blk_get_aio_context(data->job->blk);
 961     if (aio_context != data->aio_context) {
 962         aio_context_acquire(aio_context);
 963     }
 964
 965     data->fn(data->job, data->opaque);
 966
 967     if (aio_context != data->aio_context) {
 968         aio_context_release(aio_context);
 969     }
 970
 971     aio_context_release(data->aio_context);
 972
 973     g_free(data);
 974 }
 975
 976 void block_job_defer_to_main_loop(BlockJob *job,
 977                                   BlockJobDeferToMainLoopFn *fn,
 978                                   void *opaque)
 979 {
 980     BlockJobDeferToMainLoopData *data = g_malloc(sizeof(*data));
 981     data->job = job;
 982     data->aio_context = blk_get_aio_context(job->blk);
 983     data->fn = fn;
 984     data->opaque = opaque;
 985     job->deferred_to_main_loop = true;
 986
 987     aio_bh_schedule_oneshot(qemu_get_aio_context(),
 988                             block_job_defer_to_main_loop_bh, data);
 989 }