aboutsummaryrefslogtreecommitdiff
path: root/migration/block.c
blob: 91f98ef44a1ae09ac77737640aa4aa0134094866 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
/*
 * QEMU live block migration
 *
 * Copyright IBM, Corp. 2009
 *
 * Authors:
 *  Liran Schour   <lirans@il.ibm.com>
 *
 * This work is licensed under the terms of the GNU GPL, version 2.  See
 * the COPYING file in the top-level directory.
 *
 * Contributions after 2012-01-13 are licensed under the terms of the
 * GNU GPL, version 2 or (at your option) any later version.
 */

#include "qemu/osdep.h"
#include "qapi/error.h"
#include "qemu/error-report.h"
#include "qemu/cutils.h"
#include "qemu/queue.h"
#include "block.h"
#include "migration/misc.h"
#include "migration.h"
#include "migration/register.h"
#include "qemu-file.h"
#include "migration/vmstate.h"
#include "sysemu/block-backend.h"

#define BLOCK_SIZE                       (1 << 20)
#define BDRV_SECTORS_PER_DIRTY_CHUNK     (BLOCK_SIZE >> BDRV_SECTOR_BITS)

#define BLK_MIG_FLAG_DEVICE_BLOCK       0x01
#define BLK_MIG_FLAG_EOS                0x02
#define BLK_MIG_FLAG_PROGRESS           0x04
#define BLK_MIG_FLAG_ZERO_BLOCK         0x08

#define MAX_IS_ALLOCATED_SEARCH (65536 * BDRV_SECTOR_SIZE)

#define MAX_IO_BUFFERS 512
#define MAX_PARALLEL_IO 16

//#define DEBUG_BLK_MIGRATION

#ifdef DEBUG_BLK_MIGRATION
#define DPRINTF(fmt, ...) \
    do { printf("blk_migration: " fmt, ## __VA_ARGS__); } while (0)
#else
#define DPRINTF(fmt, ...) \
    do { } while (0)
#endif

typedef struct BlkMigDevState {
    /* Written during setup phase.  Can be read without a lock.  */
    BlockBackend *blk;
    char *blk_name;
    int shared_base;
    int64_t total_sectors;
    QSIMPLEQ_ENTRY(BlkMigDevState) entry;
    Error *blocker;

    /* Only used by migration thread.  Does not need a lock.  */
    int bulk_completed;
    int64_t cur_sector;
    int64_t cur_dirty;

    /* Data in the aio_bitmap is protected by block migration lock.
     * Allocation and free happen during setup and cleanup respectively.
     */
    unsigned long *aio_bitmap;

    /* Protected by block migration lock.  */
    int64_t completed_sectors;

    /* During migration this is protected by iothread lock / AioContext.
     * Allocation and free happen during setup and cleanup respectively.
     */
    BdrvDirtyBitmap *dirty_bitmap;
} BlkMigDevState;

typedef struct BlkMigBlock {
    /* Only used by migration thread.  */
    uint8_t *buf;
    BlkMigDevState *bmds;
    int64_t sector;
    int nr_sectors;
    QEMUIOVector qiov;
    BlockAIOCB *aiocb;

    /* Protected by block migration lock.  */
    int ret;
    QSIMPLEQ_ENTRY(BlkMigBlock) entry;
} BlkMigBlock;

typedef struct BlkMigState {
    QSIMPLEQ_HEAD(, BlkMigDevState) bmds_list;
    int64_t total_sector_sum;
    bool zero_blocks;

    /* Protected by lock.  */
    QSIMPLEQ_HEAD(, BlkMigBlock) blk_list;
    int submitted;
    int read_done;

    /* Only used by migration thread.  Does not need a lock.  */
    int transferred;
    int prev_progress;
    int bulk_completed;

    /* Lock must be taken _inside_ the iothread lock and any AioContexts.  */
    QemuMutex lock;
} BlkMigState;

static BlkMigState block_mig_state;

static void blk_mig_lock(void)
{
    qemu_mutex_lock(&block_mig_state.lock);
}

static void blk_mig_unlock(void)
{
    qemu_mutex_unlock(&block_mig_state.lock);
}

/* Must run outside of the iothread lock during the bulk phase,
 * or the VM will stall.
 */

static void blk_send(QEMUFile *f, BlkMigBlock * blk)
{
    int len;
    uint64_t flags = BLK_MIG_FLAG_DEVICE_BLOCK;

    if (block_mig_state.zero_blocks &&
        buffer_is_zero(blk->buf, BLOCK_SIZE)) {
        flags |= BLK_MIG_FLAG_ZERO_BLOCK;
    }

    /* sector number and flags */
    qemu_put_be64(f, (blk->sector << BDRV_SECTOR_BITS)
                     | flags);

    /* device name */
    len = strlen(blk->bmds->blk_name);
    qemu_put_byte(f, len);
    qemu_put_buffer(f, (uint8_t *) blk->bmds->blk_name, len);

    /* if a block is zero we need to flush here since the network
     * bandwidth is now a lot higher than the storage device bandwidth.
     * thus if we queue zero blocks we slow down the migration */
    if (flags & BLK_MIG_FLAG_ZERO_BLOCK) {
        qemu_fflush(f);
        return;
    }

    qemu_put_buffer(f, blk->buf, BLOCK_SIZE);
}

int blk_mig_active(void)
{
    return !QSIMPLEQ_EMPTY(&block_mig_state.bmds_list);
}

int blk_mig_bulk_active(void)
{
    return blk_mig_active() && !block_mig_state.bulk_completed;
}

uint64_t blk_mig_bytes_transferred(void)
{
    BlkMigDevState *bmds;
    uint64_t sum = 0;

    blk_mig_lock();
    QSIMPLEQ_FOREACH(bmds, &block_mig_state.bmds_list, entry) {
        sum += bmds->completed_sectors;
    }
    blk_mig_unlock();
    return sum << BDRV_SECTOR_BITS;
}

uint64_t blk_mig_bytes_remaining(void)
{
    return blk_mig_bytes_total() - blk_mig_bytes_transferred();
}

uint64_t blk_mig_bytes_total(void)
{
    BlkMigDevState *bmds;
    uint64_t sum = 0;

    QSIMPLEQ_FOREACH(bmds, &block_mig_state.bmds_list, entry) {
        sum += bmds->total_sectors;
    }
    return sum << BDRV_SECTOR_BITS;
}


/* Called with migration lock held.  */

static int bmds_aio_inflight(BlkMigDevState *bmds, int64_t sector)
{
    int64_t chunk = sector / (int64_t)BDRV_SECTORS_PER_DIRTY_CHUNK;

    if (sector < blk_nb_sectors(bmds->blk)) {
        return !!(bmds->aio_bitmap[chunk / (sizeof(unsigned long) * 8)] &
            (1UL << (chunk % (sizeof(unsigned long) * 8))));
    } else {
        return 0;
    }
}

/* Called with migration lock held.  */

static void bmds_set_aio_inflight(BlkMigDevState *bmds, int64_t sector_num,
                             int nb_sectors, int set)
{
    int64_t start, end;
    unsigned long val, idx, bit;

    start = sector_num / BDRV_SECTORS_PER_DIRTY_CHUNK;
    end = (sector_num + nb_sectors - 1) / BDRV_SECTORS_PER_DIRTY_CHUNK;

    for (; start <= end; start++) {
        idx = start / (sizeof(unsigned long) * 8);
        bit = start % (sizeof(unsigned long) * 8);
        val = bmds->aio_bitmap[idx];
        if (set) {
            val |= 1UL << bit;
        } else {
            val &= ~(1UL << bit);
        }
        bmds->aio_bitmap[idx] = val;
    }
}

static void alloc_aio_bitmap(BlkMigDevState *bmds)
{
    BlockBackend *bb = bmds->blk;
    int64_t bitmap_size;

    bitmap_size = blk_nb_sectors(bb) + BDRV_SECTORS_PER_DIRTY_CHUNK * 8 - 1;
    bitmap_size /= BDRV_SECTORS_PER_DIRTY_CHUNK * 8;

    bmds->aio_bitmap = g_malloc0(bitmap_size);
}

/* Never hold migration lock when yielding to the main loop!  */

static void blk_mig_read_cb(void *opaque, int ret)
{
    BlkMigBlock *blk = opaque;

    blk_mig_lock();
    blk->ret = ret;

    QSIMPLEQ_INSERT_TAIL(&block_mig_state.blk_list, blk, entry);
    bmds_set_aio_inflight(blk->bmds, blk->sector, blk->nr_sectors, 0);

    block_mig_state.submitted--;
    block_mig_state.read_done++;
    assert(block_mig_state.submitted >= 0);
    blk_mig_unlock();
}

/* Called with no lock taken.  */

static int mig_save_device_bulk(QEMUFile *f, BlkMigDevState *bmds)
{
    int64_t total_sectors = bmds->total_sectors;
    int64_t cur_sector = bmds->cur_sector;
    BlockBackend *bb = bmds->blk;
    BlkMigBlock *blk;
    int nr_sectors;
    int64_t count;

    if (bmds->shared_base) {
        qemu_mutex_lock_iothread();
        aio_context_acquire(blk_get_aio_context(bb));
        /* Skip unallocated sectors; intentionally treats failure or
         * partial sector as an allocated sector */
        while (cur_sector < total_sectors &&
               !bdrv_is_allocated(blk_bs(bb), cur_sector * BDRV_SECTOR_SIZE,
                                  MAX_IS_ALLOCATED_SEARCH, &count)) {
            if (count < BDRV_SECTOR_SIZE) {
                break;
            }
            cur_sector += count >> BDRV_SECTOR_BITS;
        }
        aio_context_release(blk_get_aio_context(bb));
        qemu_mutex_unlock_iothread();
    }

    if (cur_sector >= total_sectors) {
        bmds->cur_sector = bmds->completed_sectors = total_sectors;
        return 1;
    }

    bmds->completed_sectors = cur_sector;

    cur_sector &= ~((int64_t)BDRV_SECTORS_PER_DIRTY_CHUNK - 1);

    /* we are going to transfer a full block even if it is not allocated */
    nr_sectors = BDRV_SECTORS_PER_DIRTY_CHUNK;

    if (total_sectors - cur_sector < BDRV_SECTORS_PER_DIRTY_CHUNK) {
        nr_sectors = total_sectors - cur_sector;
    }

    blk = g_new(BlkMigBlock, 1);
    blk->buf = g_malloc(BLOCK_SIZE);
    blk->bmds = bmds;
    blk->sector = cur_sector;
    blk->nr_sectors = nr_sectors;

    qemu_iovec_init_buf(&blk->qiov, blk->buf, nr_sectors * BDRV_SECTOR_SIZE);

    blk_mig_lock();
    block_mig_state.submitted++;
    blk_mig_unlock();

    /* We do not know if bs is under the main thread (and thus does
     * not acquire the AioContext when doing AIO) or rather under
     * dataplane.  Thus acquire both the iothread mutex and the
     * AioContext.
     *
     * This is ugly and will disappear when we make bdrv_* thread-safe,
     * without the need to acquire the AioContext.
     */
    qemu_mutex_lock_iothread();
    aio_context_acquire(blk_get_aio_context(bmds->blk));
    bdrv_reset_dirty_bitmap(bmds->dirty_bitmap, cur_sector * BDRV_SECTOR_SIZE,
                            nr_sectors * BDRV_SECTOR_SIZE);
    blk->aiocb = blk_aio_preadv(bb, cur_sector * BDRV_SECTOR_SIZE, &blk->qiov,
                                0, blk_mig_read_cb, blk);
    aio_context_release(blk_get_aio_context(bmds->blk));
    qemu_mutex_unlock_iothread();

    bmds->cur_sector = cur_sector + nr_sectors;
    return (bmds->cur_sector >= total_sectors);
}

/* Called with iothread lock taken.  */

static int set_dirty_tracking(void)
{
    BlkMigDevState *bmds;
    int ret;

    QSIMPLEQ_FOREACH(bmds, &block_mig_state.bmds_list, entry) {
        bmds->dirty_bitmap = bdrv_create_dirty_bitmap(blk_bs(bmds->blk),
                                                      BLOCK_SIZE, NULL, NULL);
        if (!bmds->dirty_bitmap) {
            ret = -errno;
            goto fail;
        }
    }
    return 0;

fail:
    QSIMPLEQ_FOREACH(bmds, &block_mig_state.bmds_list, entry) {
        if (bmds->dirty_bitmap) {
            bdrv_release_dirty_bitmap(blk_bs(bmds->blk), bmds->dirty_bitmap);
        }
    }
    return ret;
}

/* Called with iothread lock taken.  */

static void unset_dirty_tracking(void)
{
    BlkMigDevState *bmds;

    QSIMPLEQ_FOREACH(bmds, &block_mig_state.bmds_list, entry) {
        bdrv_release_dirty_bitmap(blk_bs(bmds->blk), bmds->dirty_bitmap);
    }
}

static int init_blk_migration(QEMUFile *f)
{
    BlockDriverState *bs;
    BlkMigDevState *bmds;
    int64_t sectors;
    BdrvNextIterator it;
    int i, num_bs = 0;
    struct {
        BlkMigDevState *bmds;
        BlockDriverState *bs;
    } *bmds_bs;
    Error *local_err = NULL;
    int ret;

    block_mig_state.submitted = 0;
    block_mig_state.read_done = 0;
    block_mig_state.transferred = 0;
    block_mig_state.total_sector_sum = 0;
    block_mig_state.prev_progress = -1;
    block_mig_state.bulk_completed = 0;
    block_mig_state.zero_blocks = migrate_zero_blocks();

    for (bs = bdrv_first(&it); bs; bs = bdrv_next(&it)) {
        num_bs++;
    }
    bmds_bs = g_malloc0(num_bs * sizeof(*bmds_bs));

    for (i = 0, bs = bdrv_first(&it); bs; bs = bdrv_next(&it), i++) {
        if (bdrv_is_read_only(bs)) {
            continue;
        }

        sectors = bdrv_nb_sectors(bs);
        if (sectors <= 0) {
            ret = sectors;
            bdrv_next_cleanup(&it);
            goto out;
        }

        bmds = g_new0(BlkMigDevState, 1);
        bmds->blk = blk_new(qemu_get_aio_context(),
                            BLK_PERM_CONSISTENT_READ, BLK_PERM_ALL);
        bmds->blk_name = g_strdup(bdrv_get_device_name(bs));
        bmds->bulk_completed = 0;
        bmds->total_sectors = sectors;
        bmds->completed_sectors = 0;
        bmds->shared_base = migrate_use_block_incremental();

        assert(i < num_bs);
        bmds_bs[i].bmds = bmds;
        bmds_bs[i].bs = bs;

        block_mig_state.total_sector_sum += sectors;

        if (bmds->shared_base) {
            DPRINTF("Start migration for %s with shared base image\n",
                    bdrv_get_device_name(bs));
        } else {
            DPRINTF("Start full migration for %s\n", bdrv_get_device_name(bs));
        }

        QSIMPLEQ_INSERT_TAIL(&block_mig_state.bmds_list, bmds, entry);
    }

    /* Can only insert new BDSes now because doing so while iterating block
     * devices may end up in a deadlock (iterating the new BDSes, too). */
    for (i = 0; i < num_bs; i++) {
        BlkMigDevState *bmds = bmds_bs[i].bmds;
        BlockDriverState *bs = bmds_bs[i].bs;

        if (bmds) {
            ret = blk_insert_bs(bmds->blk, bs, &local_err);
            if (ret < 0) {
                error_report_err(local_err);
                goto out;
            }

            alloc_aio_bitmap(bmds);
            error_setg(&bmds->blocker, "block device is in use by migration");
            bdrv_op_block_all(bs, bmds->blocker);
        }
    }

    ret = 0;
out:
    g_free(bmds_bs);
    return ret;
}

/* Called with no lock taken.  */

static int blk_mig_save_bulked_block(QEMUFile *f)
{
    int64_t completed_sector_sum = 0;
    BlkMigDevState *bmds;
    int progress;
    int ret = 0;

    QSIMPLEQ_FOREACH(bmds, &block_mig_state.bmds_list, entry) {
        if (bmds->bulk_completed == 0) {
            if (mig_save_device_bulk(f, bmds) == 1) {
                /* completed bulk section for this device */
                bmds->bulk_completed = 1;
            }
            completed_sector_sum += bmds->completed_sectors;
            ret = 1;
            break;
        } else {
            completed_sector_sum += bmds->completed_sectors;
        }
    }

    if (block_mig_state.total_sector_sum != 0) {
        progress = completed_sector_sum * 100 /
                   block_mig_state.total_sector_sum;
    } else {
        progress = 100;
    }
    if (progress != block_mig_state.prev_progress) {
        block_mig_state.prev_progress = progress;
        qemu_put_be64(f, (progress << BDRV_SECTOR_BITS)
                         | BLK_MIG_FLAG_PROGRESS);
        DPRINTF("Completed %d %%\r", progress);
    }

    return ret;
}

static void blk_mig_reset_dirty_cursor(void)
{
    BlkMigDevState *bmds;

    QSIMPLEQ_FOREACH(bmds, &block_mig_state.bmds_list, entry) {
        bmds->cur_dirty = 0;
    }
}

/* Called with iothread lock and AioContext taken.  */

static int mig_save_device_dirty(QEMUFile *f, BlkMigDevState *bmds,
                                 int is_async)
{
    BlkMigBlock *blk;
    BlockDriverState *bs = blk_bs(bmds->blk);
    int64_t total_sectors = bmds->total_sectors;
    int64_t sector;
    int nr_sectors;
    int ret = -EIO;

    for (sector = bmds->cur_dirty; sector < bmds->total_sectors;) {
        blk_mig_lock();
        if (bmds_aio_inflight(bmds, sector)) {
            blk_mig_unlock();
            blk_drain(bmds->blk);
        } else {
            blk_mig_unlock();
        }
        bdrv_dirty_bitmap_lock(bmds->dirty_bitmap);
        if (bdrv_get_dirty_locked(bs, bmds->dirty_bitmap,
                                  sector * BDRV_SECTOR_SIZE)) {
            if (total_sectors - sector < BDRV_SECTORS_PER_DIRTY_CHUNK) {
                nr_sectors = total_sectors - sector;
            } else {
                nr_sectors = BDRV_SECTORS_PER_DIRTY_CHUNK;
            }
            bdrv_reset_dirty_bitmap_locked(bmds->dirty_bitmap,
                                           sector * BDRV_SECTOR_SIZE,
                                           nr_sectors * BDRV_SECTOR_SIZE);
            bdrv_dirty_bitmap_unlock(bmds->dirty_bitmap);

            blk = g_new(BlkMigBlock, 1);
            blk->buf = g_malloc(BLOCK_SIZE);
            blk->bmds = bmds;
            blk->sector = sector;
            blk->nr_sectors = nr_sectors;

            if (is_async) {
                qemu_iovec_init_buf(&blk->qiov, blk->buf,
                                    nr_sectors * BDRV_SECTOR_SIZE);

                blk->aiocb = blk_aio_preadv(bmds->blk,
                                            sector * BDRV_SECTOR_SIZE,
                                            &blk->qiov, 0, blk_mig_read_cb,
                                            blk);

                blk_mig_lock();
                block_mig_state.submitted++;
                bmds_set_aio_inflight(bmds, sector, nr_sectors, 1);
                blk_mig_unlock();
            } else {
                ret = blk_pread(bmds->blk, sector * BDRV_SECTOR_SIZE, blk->buf,
                                nr_sectors * BDRV_SECTOR_SIZE);
                if (ret < 0) {
                    goto error;
                }
                blk_send(f, blk);

                g_free(blk->buf);
                g_free(blk);
            }

            sector += nr_sectors;
            bmds->cur_dirty = sector;
            break;
        }

        bdrv_dirty_bitmap_unlock(bmds->dirty_bitmap);
        sector += BDRV_SECTORS_PER_DIRTY_CHUNK;
        bmds->cur_dirty = sector;
    }

    return (bmds->cur_dirty >= bmds->total_sectors);

error:
    DPRINTF("Error reading sector %" PRId64 "\n", sector);
    g_free(blk->buf);
    g_free(blk);
    return ret;
}

/* Called with iothread lock taken.
 *
 * return value:
 * 0: too much data for max_downtime
 * 1: few enough data for max_downtime
*/
static int blk_mig_save_dirty_block(QEMUFile *f, int is_async)
{
    BlkMigDevState *bmds;
    int ret = 1;

    QSIMPLEQ_FOREACH(bmds, &block_mig_state.bmds_list, entry) {
        aio_context_acquire(blk_get_aio_context(bmds->blk));
        ret = mig_save_device_dirty(f, bmds, is_async);
        aio_context_release(blk_get_aio_context(bmds->blk));
        if (ret <= 0) {
            break;
        }
    }

    return ret;
}

/* Called with no locks taken.  */

static int flush_blks(QEMUFile *f)
{
    BlkMigBlock *blk;
    int ret = 0;

    DPRINTF("%s Enter submitted %d read_done %d transferred %d\n",
            __func__, block_mig_state.submitted, block_mig_state.read_done,
            block_mig_state.transferred);

    blk_mig_lock();
    while ((blk = QSIMPLEQ_FIRST(&block_mig_state.blk_list)) != NULL) {
        if (qemu_file_rate_limit(f)) {
            break;
        }
        if (blk->ret < 0) {
            ret = blk->ret;
            break;
        }

        QSIMPLEQ_REMOVE_HEAD(&block_mig_state.blk_list, entry);
        blk_mig_unlock();
        blk_send(f, blk);
        blk_mig_lock();

        g_free(blk->buf);
        g_free(blk);

        block_mig_state.read_done--;
        block_mig_state.transferred++;
        assert(block_mig_state.read_done >= 0);
    }
    blk_mig_unlock();

    DPRINTF("%s Exit submitted %d read_done %d transferred %d\n", __func__,
            block_mig_state.submitted, block_mig_state.read_done,
            block_mig_state.transferred);
    return ret;
}

/* Called with iothread lock taken.  */

static int64_t get_remaining_dirty(void)
{
    BlkMigDevState *bmds;
    int64_t dirty = 0;

    QSIMPLEQ_FOREACH(bmds, &block_mig_state.bmds_list, entry) {
        aio_context_acquire(blk_get_aio_context(bmds->blk));
        dirty += bdrv_get_dirty_count(bmds->dirty_bitmap);
        aio_context_release(blk_get_aio_context(bmds->blk));
    }

    return dirty;
}



/* Called with iothread lock taken.  */
static void block_migration_cleanup_bmds(void)
{
    BlkMigDevState *bmds;
    AioContext *ctx;

    unset_dirty_tracking();

    while ((bmds = QSIMPLEQ_FIRST(&block_mig_state.bmds_list)) != NULL) {
        QSIMPLEQ_REMOVE_HEAD(&block_mig_state.bmds_list, entry);
        bdrv_op_unblock_all(blk_bs(bmds->blk), bmds->blocker);
        error_free(bmds->blocker);

        /* Save ctx, because bmds->blk can disappear during blk_unref.  */
        ctx = blk_get_aio_context(bmds->blk);
        aio_context_acquire(ctx);
        blk_unref(bmds->blk);
        aio_context_release(ctx);

        g_free(bmds->blk_name);
        g_free(bmds->aio_bitmap);
        g_free(bmds);
    }
}

/* Called with iothread lock taken.  */
static void block_migration_cleanup(void *opaque)
{
    BlkMigBlock *blk;

    bdrv_drain_all();

    block_migration_cleanup_bmds();

    blk_mig_lock();
    while ((blk = QSIMPLEQ_FIRST(&block_mig_state.blk_list)) != NULL) {
        QSIMPLEQ_REMOVE_HEAD(&block_mig_state.blk_list, entry);
        g_free(blk->buf);
        g_free(blk);
    }
    blk_mig_unlock();
}

static int block_save_setup(QEMUFile *f, void *opaque)
{
    int ret;

    DPRINTF("Enter save live setup submitted %d transferred %d\n",
            block_mig_state.submitted, block_mig_state.transferred);

    qemu_mutex_lock_iothread();
    ret = init_blk_migration(f);
    if (ret < 0) {
        qemu_mutex_unlock_iothread();
        return ret;
    }

    /* start track dirty blocks */
    ret = set_dirty_tracking();

    qemu_mutex_unlock_iothread();

    if (ret) {
        return ret;
    }

    ret = flush_blks(f);
    blk_mig_reset_dirty_cursor();
    qemu_put_be64(f, BLK_MIG_FLAG_EOS);

    return ret;
}

static int block_save_iterate(QEMUFile *f, void *opaque)
{
    int ret;
    int64_t last_ftell = qemu_ftell(f);
    int64_t delta_ftell;

    DPRINTF("Enter save live iterate submitted %d transferred %d\n",
            block_mig_state.submitted, block_mig_state.transferred);

    ret = flush_blks(f);
    if (ret) {
        return ret;
    }

    blk_mig_reset_dirty_cursor();

    /* control the rate of transfer */
    blk_mig_lock();
    while (block_mig_state.read_done * BLOCK_SIZE <
           qemu_file_get_rate_limit(f) &&
           block_mig_state.submitted < MAX_PARALLEL_IO &&
           (block_mig_state.submitted + block_mig_state.read_done) <
           MAX_IO_BUFFERS) {
        blk_mig_unlock();
        if (block_mig_state.bulk_completed == 0) {
            /* first finish the bulk phase */
            if (blk_mig_save_bulked_block(f) == 0) {
                /* finished saving bulk on all devices */
                block_mig_state.bulk_completed = 1;
            }
            ret = 0;
        } else {
            /* Always called with iothread lock taken for
             * simplicity, block_save_complete also calls it.
             */
            qemu_mutex_lock_iothread();
            ret = blk_mig_save_dirty_block(f, 1);
            qemu_mutex_unlock_iothread();
        }
        if (ret < 0) {
            return ret;
        }
        blk_mig_lock();
        if (ret != 0) {
            /* no more dirty blocks */
            break;
        }
    }
    blk_mig_unlock();

    ret = flush_blks(f);
    if (ret) {
        return ret;
    }

    qemu_put_be64(f, BLK_MIG_FLAG_EOS);
    delta_ftell = qemu_ftell(f) - last_ftell;
    if (delta_ftell > 0) {
        return 1;
    } else if (delta_ftell < 0) {
        return -1;
    } else {
        return 0;
    }
}

/* Called with iothread lock taken.  */

static int block_save_complete(QEMUFile *f, void *opaque)
{
    int ret;

    DPRINTF("Enter save live complete submitted %d transferred %d\n",
            block_mig_state.submitted, block_mig_state.transferred);

    ret = flush_blks(f);
    if (ret) {
        return ret;
    }

    blk_mig_reset_dirty_cursor();

    /* we know for sure that save bulk is completed and
       all async read completed */
    blk_mig_lock();
    assert(block_mig_state.submitted == 0);
    blk_mig_unlock();

    do {
        ret = blk_mig_save_dirty_block(f, 0);
        if (ret < 0) {
            return ret;
        }
    } while (ret == 0);

    /* report completion */
    qemu_put_be64(f, (100 << BDRV_SECTOR_BITS) | BLK_MIG_FLAG_PROGRESS);

    DPRINTF("Block migration completed\n");

    qemu_put_be64(f, BLK_MIG_FLAG_EOS);

    /* Make sure that our BlockBackends are gone, so that the block driver
     * nodes can be inactivated. */
    block_migration_cleanup_bmds();

    return 0;
}

static void block_save_pending(QEMUFile *f, void *opaque, uint64_t max_size,
                               uint64_t *res_precopy_only,
                               uint64_t *res_compatible,
                               uint64_t *res_postcopy_only)
{
    /* Estimate pending number of bytes to send */
    uint64_t pending;

    qemu_mutex_lock_iothread();
    pending = get_remaining_dirty();
    qemu_mutex_unlock_iothread();

    blk_mig_lock();
    pending += block_mig_state.submitted * BLOCK_SIZE +
               block_mig_state.read_done * BLOCK_SIZE;
    blk_mig_unlock();

    /* Report at least one block pending during bulk phase */
    if (pending <= max_size && !block_mig_state.bulk_completed) {
        pending = max_size + BLOCK_SIZE;
    }

    DPRINTF("Enter save live pending  %" PRIu64 "\n", pending);
    /* We don't do postcopy */
    *res_precopy_only += pending;
}

static int block_load(QEMUFile *f, void *opaque, int version_id)
{
    static int banner_printed;
    int len, flags;
    char device_name[256];
    int64_t addr;
    BlockBackend *blk, *blk_prev = NULL;
    Error *local_err = NULL;
    uint8_t *buf;
    int64_t total_sectors = 0;
    int nr_sectors;
    int ret;
    BlockDriverInfo bdi;
    int cluster_size = BLOCK_SIZE;

    do {
        addr = qemu_get_be64(f);

        flags = addr & ~BDRV_SECTOR_MASK;
        addr >>= BDRV_SECTOR_BITS;

        if (flags & BLK_MIG_FLAG_DEVICE_BLOCK) {
            /* get device name */
            len = qemu_get_byte(f);
            qemu_get_buffer(f, (uint8_t *)device_name, len);
            device_name[len] = '\0';

            blk = blk_by_name(device_name);
            if (!blk) {
                fprintf(stderr, "Error unknown block device %s\n",
                        device_name);
                return -EINVAL;
            }

            if (blk != blk_prev) {
                blk_prev = blk;
                total_sectors = blk_nb_sectors(blk);
                if (total_sectors <= 0) {
                    error_report("Error getting length of block device %s",
                                 device_name);
                    return -EINVAL;
                }

                blk_invalidate_cache(blk, &local_err);
                if (local_err) {
                    error_report_err(local_err);
                    return -EINVAL;
                }

                ret = bdrv_get_info(blk_bs(blk), &bdi);
                if (ret == 0 && bdi.cluster_size > 0 &&
                    bdi.cluster_size <= BLOCK_SIZE &&
                    BLOCK_SIZE % bdi.cluster_size == 0) {
                    cluster_size = bdi.cluster_size;
                } else {
                    cluster_size = BLOCK_SIZE;
                }
            }

            if (total_sectors - addr < BDRV_SECTORS_PER_DIRTY_CHUNK) {
                nr_sectors = total_sectors - addr;
            } else {
                nr_sectors = BDRV_SECTORS_PER_DIRTY_CHUNK;
            }

            if (flags & BLK_MIG_FLAG_ZERO_BLOCK) {
                ret = blk_pwrite_zeroes(blk, addr * BDRV_SECTOR_SIZE,
                                        nr_sectors * BDRV_SECTOR_SIZE,
                                        BDRV_REQ_MAY_UNMAP);
            } else {
                int i;
                int64_t cur_addr;
                uint8_t *cur_buf;

                buf = g_malloc(BLOCK_SIZE);
                qemu_get_buffer(f, buf, BLOCK_SIZE);
                for (i = 0; i < BLOCK_SIZE / cluster_size; i++) {
                    cur_addr = addr * BDRV_SECTOR_SIZE + i * cluster_size;
                    cur_buf = buf + i * cluster_size;

                    if ((!block_mig_state.zero_blocks ||
                        cluster_size < BLOCK_SIZE) &&
                        buffer_is_zero(cur_buf, cluster_size)) {
                        ret = blk_pwrite_zeroes(blk, cur_addr,
                                                cluster_size,
                                                BDRV_REQ_MAY_UNMAP);
                    } else {
                        ret = blk_pwrite(blk, cur_addr, cur_buf,
                                         cluster_size, 0);
                    }
                    if (ret < 0) {
                        break;
                    }
                }
                g_free(buf);
            }

            if (ret < 0) {
                return ret;
            }
        } else if (flags & BLK_MIG_FLAG_PROGRESS) {
            if (!banner_printed) {
                printf("Receiving block device images\n");
                banner_printed = 1;
            }
            printf("Completed %d %%%c", (int)addr,
                   (addr == 100) ? '\n' : '\r');
            fflush(stdout);
        } else if (!(flags & BLK_MIG_FLAG_EOS)) {
            fprintf(stderr, "Unknown block migration flags: %#x\n", flags);
            return -EINVAL;
        }
        ret = qemu_file_get_error(f);
        if (ret != 0) {
            return ret;
        }
    } while (!(flags & BLK_MIG_FLAG_EOS));

    return 0;
}

static bool block_is_active(void *opaque)
{
    return migrate_use_block();
}

static SaveVMHandlers savevm_block_handlers = {
    .save_setup = block_save_setup,
    .save_live_iterate = block_save_iterate,
    .save_live_complete_precopy = block_save_complete,
    .save_live_pending = block_save_pending,
    .load_state = block_load,
    .save_cleanup = block_migration_cleanup,
    .is_active = block_is_active,
};

void blk_mig_init(void)
{
    QSIMPLEQ_INIT(&block_mig_state.bmds_list);
    QSIMPLEQ_INIT(&block_mig_state.blk_list);
    qemu_mutex_init(&block_mig_state.lock);

    register_savevm_live(NULL, "block", 0, 1, &savevm_block_handlers,
                         &block_mig_state);
}