summaryrefslogtreecommitdiff
path: root/block/qcow2-cluster.c
blob: 67be0ce2c9458856910bb34185339d70285330e6 (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
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
1085
1086
1087
1088
1089
1090
1091
1092
1093
1094
1095
1096
1097
1098
1099
1100
1101
1102
1103
1104
1105
1106
1107
1108
1109
1110
1111
1112
1113
1114
1115
1116
1117
1118
1119
1120
1121
1122
1123
1124
1125
1126
1127
1128
1129
1130
1131
1132
1133
1134
1135
1136
1137
1138
1139
1140
1141
1142
1143
1144
1145
1146
1147
1148
1149
1150
1151
1152
1153
1154
1155
1156
1157
1158
1159
1160
1161
1162
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
1181
1182
1183
1184
1185
1186
1187
1188
1189
1190
1191
1192
1193
1194
1195
1196
1197
1198
1199
1200
1201
1202
1203
1204
1205
1206
1207
1208
1209
1210
1211
1212
1213
1214
1215
1216
1217
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
1228
1229
1230
1231
1232
1233
1234
1235
1236
1237
1238
1239
1240
1241
1242
1243
1244
1245
1246
1247
1248
1249
1250
1251
1252
1253
1254
1255
1256
1257
1258
1259
1260
1261
1262
1263
1264
1265
1266
1267
1268
1269
1270
1271
1272
1273
1274
1275
1276
1277
1278
1279
1280
1281
1282
1283
1284
1285
1286
1287
1288
1289
1290
1291
1292
1293
1294
1295
1296
1297
1298
1299
1300
1301
1302
1303
1304
1305
1306
1307
1308
1309
1310
1311
1312
1313
1314
1315
1316
1317
1318
1319
1320
1321
1322
1323
1324
1325
1326
1327
1328
1329
1330
1331
1332
1333
1334
1335
1336
1337
1338
1339
1340
1341
1342
1343
1344
1345
1346
1347
1348
1349
1350
1351
1352
1353
1354
1355
1356
1357
1358
1359
1360
1361
1362
1363
1364
1365
1366
1367
1368
1369
1370
1371
1372
1373
1374
1375
1376
1377
1378
1379
1380
1381
1382
1383
1384
1385
1386
1387
1388
1389
1390
1391
1392
1393
1394
1395
1396
1397
1398
1399
1400
1401
1402
1403
1404
1405
1406
1407
1408
1409
1410
1411
1412
1413
1414
1415
1416
1417
1418
1419
1420
1421
1422
1423
1424
1425
1426
1427
1428
1429
1430
1431
1432
1433
1434
1435
1436
1437
1438
1439
1440
1441
1442
1443
1444
1445
1446
1447
1448
1449
1450
1451
1452
1453
1454
1455
1456
1457
1458
1459
1460
1461
1462
1463
1464
1465
1466
1467
1468
1469
1470
1471
1472
1473
1474
1475
1476
1477
1478
1479
1480
1481
1482
1483
1484
1485
1486
1487
1488
1489
1490
1491
1492
1493
1494
1495
1496
1497
1498
1499
1500
1501
1502
1503
1504
1505
1506
1507
1508
1509
1510
1511
1512
1513
1514
1515
1516
1517
1518
1519
1520
1521
1522
1523
1524
1525
1526
1527
1528
1529
1530
1531
1532
1533
1534
1535
1536
1537
1538
1539
1540
1541
1542
1543
1544
1545
1546
1547
1548
1549
1550
1551
1552
1553
1554
1555
1556
1557
1558
1559
1560
1561
1562
1563
1564
1565
1566
1567
1568
1569
1570
1571
1572
1573
1574
1575
1576
1577
1578
1579
1580
1581
1582
1583
1584
1585
1586
1587
1588
1589
1590
1591
1592
1593
1594
1595
1596
1597
1598
1599
1600
1601
1602
1603
1604
1605
1606
1607
1608
1609
1610
1611
1612
1613
1614
1615
1616
1617
1618
1619
1620
1621
1622
1623
1624
1625
1626
1627
1628
1629
1630
1631
1632
1633
1634
1635
1636
1637
1638
1639
1640
1641
1642
1643
1644
1645
1646
1647
1648
1649
1650
1651
1652
1653
1654
1655
1656
1657
1658
1659
1660
1661
1662
1663
1664
1665
1666
1667
1668
1669
1670
1671
1672
1673
1674
1675
1676
1677
1678
1679
1680
1681
1682
1683
1684
1685
1686
1687
1688
1689
1690
1691
1692
1693
1694
1695
1696
1697
1698
1699
1700
1701
1702
1703
1704
1705
1706
1707
1708
1709
1710
1711
1712
1713
1714
1715
1716
1717
1718
1719
1720
1721
1722
1723
1724
1725
1726
1727
1728
1729
1730
1731
1732
1733
1734
1735
1736
1737
1738
1739
1740
1741
1742
1743
1744
1745
1746
1747
1748
1749
1750
1751
1752
1753
1754
1755
1756
1757
1758
1759
1760
1761
1762
1763
1764
1765
1766
1767
1768
1769
1770
1771
1772
1773
1774
1775
1776
1777
1778
1779
1780
1781
1782
1783
1784
1785
1786
1787
1788
1789
1790
1791
1792
1793
1794
1795
1796
1797
1798
1799
1800
1801
1802
1803
1804
1805
1806
1807
1808
1809
1810
1811
1812
1813
1814
1815
1816
1817
1818
1819
1820
1821
1822
1823
1824
1825
1826
1827
1828
1829
1830
1831
1832
1833
1834
1835
1836
1837
1838
1839
1840
1841
1842
1843
1844
1845
1846
1847
1848
1849
1850
1851
1852
1853
1854
1855
1856
1857
1858
1859
1860
1861
1862
1863
1864
1865
1866
1867
1868
1869
1870
1871
1872
1873
1874
1875
1876
1877
1878
1879
1880
1881
1882
1883
1884
1885
1886
1887
1888
1889
1890
1891
1892
/*
 * Block driver for the QCOW version 2 format
 *
 * Copyright (c) 2004-2006 Fabrice Bellard
 *
 * Permission is hereby granted, free of charge, to any person obtaining a copy
 * of this software and associated documentation files (the "Software"), to deal
 * in the Software without restriction, including without limitation the rights
 * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
 * copies of the Software, and to permit persons to whom the Software is
 * furnished to do so, subject to the following conditions:
 *
 * The above copyright notice and this permission notice shall be included in
 * all copies or substantial portions of the Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
 * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
 * OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
 * THE SOFTWARE.
 */

#include <zlib.h>

#include "qemu-common.h"
#include "block/block_int.h"
#include "block/qcow2.h"
#include "trace.h"

int qcow2_grow_l1_table(BlockDriverState *bs, uint64_t min_size,
                        bool exact_size)
{
    BDRVQcow2State *s = bs->opaque;
    int new_l1_size2, ret, i;
    uint64_t *new_l1_table;
    int64_t old_l1_table_offset, old_l1_size;
    int64_t new_l1_table_offset, new_l1_size;
    uint8_t data[12];

    if (min_size <= s->l1_size)
        return 0;

    /* Do a sanity check on min_size before trying to calculate new_l1_size
     * (this prevents overflows during the while loop for the calculation of
     * new_l1_size) */
    if (min_size > INT_MAX / sizeof(uint64_t)) {
        return -EFBIG;
    }

    if (exact_size) {
        new_l1_size = min_size;
    } else {
        /* Bump size up to reduce the number of times we have to grow */
        new_l1_size = s->l1_size;
        if (new_l1_size == 0) {
            new_l1_size = 1;
        }
        while (min_size > new_l1_size) {
            new_l1_size = (new_l1_size * 3 + 1) / 2;
        }
    }

    if (new_l1_size > INT_MAX / sizeof(uint64_t)) {
        return -EFBIG;
    }

#ifdef DEBUG_ALLOC2
    fprintf(stderr, "grow l1_table from %d to %" PRId64 "\n",
            s->l1_size, new_l1_size);
#endif

    new_l1_size2 = sizeof(uint64_t) * new_l1_size;
    new_l1_table = qemu_try_blockalign(bs->file->bs,
                                       align_offset(new_l1_size2, 512));
    if (new_l1_table == NULL) {
        return -ENOMEM;
    }
    memset(new_l1_table, 0, align_offset(new_l1_size2, 512));

    memcpy(new_l1_table, s->l1_table, s->l1_size * sizeof(uint64_t));

    /* write new table (align to cluster) */
    BLKDBG_EVENT(bs->file, BLKDBG_L1_GROW_ALLOC_TABLE);
    new_l1_table_offset = qcow2_alloc_clusters(bs, new_l1_size2);
    if (new_l1_table_offset < 0) {
        qemu_vfree(new_l1_table);
        return new_l1_table_offset;
    }

    ret = qcow2_cache_flush(bs, s->refcount_block_cache);
    if (ret < 0) {
        goto fail;
    }

    /* the L1 position has not yet been updated, so these clusters must
     * indeed be completely free */
    ret = qcow2_pre_write_overlap_check(bs, 0, new_l1_table_offset,
                                        new_l1_size2);
    if (ret < 0) {
        goto fail;
    }

    BLKDBG_EVENT(bs->file, BLKDBG_L1_GROW_WRITE_TABLE);
    for(i = 0; i < s->l1_size; i++)
        new_l1_table[i] = cpu_to_be64(new_l1_table[i]);
    ret = bdrv_pwrite_sync(bs->file->bs, new_l1_table_offset,
                           new_l1_table, new_l1_size2);
    if (ret < 0)
        goto fail;
    for(i = 0; i < s->l1_size; i++)
        new_l1_table[i] = be64_to_cpu(new_l1_table[i]);

    /* set new table */
    BLKDBG_EVENT(bs->file, BLKDBG_L1_GROW_ACTIVATE_TABLE);
    cpu_to_be32w((uint32_t*)data, new_l1_size);
    stq_be_p(data + 4, new_l1_table_offset);
    ret = bdrv_pwrite_sync(bs->file->bs, offsetof(QCowHeader, l1_size),
                           data, sizeof(data));
    if (ret < 0) {
        goto fail;
    }
    qemu_vfree(s->l1_table);
    old_l1_table_offset = s->l1_table_offset;
    s->l1_table_offset = new_l1_table_offset;
    s->l1_table = new_l1_table;
    old_l1_size = s->l1_size;
    s->l1_size = new_l1_size;
    qcow2_free_clusters(bs, old_l1_table_offset, old_l1_size * sizeof(uint64_t),
                        QCOW2_DISCARD_OTHER);
    return 0;
 fail:
    qemu_vfree(new_l1_table);
    qcow2_free_clusters(bs, new_l1_table_offset, new_l1_size2,
                        QCOW2_DISCARD_OTHER);
    return ret;
}

/*
 * l2_load
 *
 * Loads a L2 table into memory. If the table is in the cache, the cache
 * is used; otherwise the L2 table is loaded from the image file.
 *
 * Returns a pointer to the L2 table on success, or NULL if the read from
 * the image file failed.
 */

static int l2_load(BlockDriverState *bs, uint64_t l2_offset,
    uint64_t **l2_table)
{
    BDRVQcow2State *s = bs->opaque;
    int ret;

    ret = qcow2_cache_get(bs, s->l2_table_cache, l2_offset, (void**) l2_table);

    return ret;
}

/*
 * Writes one sector of the L1 table to the disk (can't update single entries
 * and we really don't want bdrv_pread to perform a read-modify-write)
 */
#define L1_ENTRIES_PER_SECTOR (512 / 8)
int qcow2_write_l1_entry(BlockDriverState *bs, int l1_index)
{
    BDRVQcow2State *s = bs->opaque;
    uint64_t buf[L1_ENTRIES_PER_SECTOR] = { 0 };
    int l1_start_index;
    int i, ret;

    l1_start_index = l1_index & ~(L1_ENTRIES_PER_SECTOR - 1);
    for (i = 0; i < L1_ENTRIES_PER_SECTOR && l1_start_index + i < s->l1_size;
         i++)
    {
        buf[i] = cpu_to_be64(s->l1_table[l1_start_index + i]);
    }

    ret = qcow2_pre_write_overlap_check(bs, QCOW2_OL_ACTIVE_L1,
            s->l1_table_offset + 8 * l1_start_index, sizeof(buf));
    if (ret < 0) {
        return ret;
    }

    BLKDBG_EVENT(bs->file, BLKDBG_L1_UPDATE);
    ret = bdrv_pwrite_sync(bs->file->bs,
                           s->l1_table_offset + 8 * l1_start_index,
                           buf, sizeof(buf));
    if (ret < 0) {
        return ret;
    }

    return 0;
}

/*
 * l2_allocate
 *
 * Allocate a new l2 entry in the file. If l1_index points to an already
 * used entry in the L2 table (i.e. we are doing a copy on write for the L2
 * table) copy the contents of the old L2 table into the newly allocated one.
 * Otherwise the new table is initialized with zeros.
 *
 */

static int l2_allocate(BlockDriverState *bs, int l1_index, uint64_t **table)
{
    BDRVQcow2State *s = bs->opaque;
    uint64_t old_l2_offset;
    uint64_t *l2_table = NULL;
    int64_t l2_offset;
    int ret;

    old_l2_offset = s->l1_table[l1_index];

    trace_qcow2_l2_allocate(bs, l1_index);

    /* allocate a new l2 entry */

    l2_offset = qcow2_alloc_clusters(bs, s->l2_size * sizeof(uint64_t));
    if (l2_offset < 0) {
        ret = l2_offset;
        goto fail;
    }

    ret = qcow2_cache_flush(bs, s->refcount_block_cache);
    if (ret < 0) {
        goto fail;
    }

    /* allocate a new entry in the l2 cache */

    trace_qcow2_l2_allocate_get_empty(bs, l1_index);
    ret = qcow2_cache_get_empty(bs, s->l2_table_cache, l2_offset, (void**) table);
    if (ret < 0) {
        goto fail;
    }

    l2_table = *table;

    if ((old_l2_offset & L1E_OFFSET_MASK) == 0) {
        /* if there was no old l2 table, clear the new table */
        memset(l2_table, 0, s->l2_size * sizeof(uint64_t));
    } else {
        uint64_t* old_table;

        /* if there was an old l2 table, read it from the disk */
        BLKDBG_EVENT(bs->file, BLKDBG_L2_ALLOC_COW_READ);
        ret = qcow2_cache_get(bs, s->l2_table_cache,
            old_l2_offset & L1E_OFFSET_MASK,
            (void**) &old_table);
        if (ret < 0) {
            goto fail;
        }

        memcpy(l2_table, old_table, s->cluster_size);

        qcow2_cache_put(bs, s->l2_table_cache, (void **) &old_table);
    }

    /* write the l2 table to the file */
    BLKDBG_EVENT(bs->file, BLKDBG_L2_ALLOC_WRITE);

    trace_qcow2_l2_allocate_write_l2(bs, l1_index);
    qcow2_cache_entry_mark_dirty(bs, s->l2_table_cache, l2_table);
    ret = qcow2_cache_flush(bs, s->l2_table_cache);
    if (ret < 0) {
        goto fail;
    }

    /* update the L1 entry */
    trace_qcow2_l2_allocate_write_l1(bs, l1_index);
    s->l1_table[l1_index] = l2_offset | QCOW_OFLAG_COPIED;
    ret = qcow2_write_l1_entry(bs, l1_index);
    if (ret < 0) {
        goto fail;
    }

    *table = l2_table;
    trace_qcow2_l2_allocate_done(bs, l1_index, 0);
    return 0;

fail:
    trace_qcow2_l2_allocate_done(bs, l1_index, ret);
    if (l2_table != NULL) {
        qcow2_cache_put(bs, s->l2_table_cache, (void**) table);
    }
    s->l1_table[l1_index] = old_l2_offset;
    if (l2_offset > 0) {
        qcow2_free_clusters(bs, l2_offset, s->l2_size * sizeof(uint64_t),
                            QCOW2_DISCARD_ALWAYS);
    }
    return ret;
}

/*
 * Checks how many clusters in a given L2 table are contiguous in the image
 * file. As soon as one of the flags in the bitmask stop_flags changes compared
 * to the first cluster, the search is stopped and the cluster is not counted
 * as contiguous. (This allows it, for example, to stop at the first compressed
 * cluster which may require a different handling)
 */
static int count_contiguous_clusters(int nb_clusters, int cluster_size,
        uint64_t *l2_table, uint64_t stop_flags)
{
    int i;
    uint64_t mask = stop_flags | L2E_OFFSET_MASK | QCOW_OFLAG_COMPRESSED;
    uint64_t first_entry = be64_to_cpu(l2_table[0]);
    uint64_t offset = first_entry & mask;

    if (!offset)
        return 0;

    assert(qcow2_get_cluster_type(first_entry) != QCOW2_CLUSTER_COMPRESSED);

    for (i = 0; i < nb_clusters; i++) {
        uint64_t l2_entry = be64_to_cpu(l2_table[i]) & mask;
        if (offset + (uint64_t) i * cluster_size != l2_entry) {
            break;
        }
    }

	return i;
}

static int count_contiguous_free_clusters(int nb_clusters, uint64_t *l2_table)
{
    int i;

    for (i = 0; i < nb_clusters; i++) {
        int type = qcow2_get_cluster_type(be64_to_cpu(l2_table[i]));

        if (type != QCOW2_CLUSTER_UNALLOCATED) {
            break;
        }
    }

    return i;
}

/* The crypt function is compatible with the linux cryptoloop
   algorithm for < 4 GB images. NOTE: out_buf == in_buf is
   supported */
int qcow2_encrypt_sectors(BDRVQcow2State *s, int64_t sector_num,
                          uint8_t *out_buf, const uint8_t *in_buf,
                          int nb_sectors, bool enc,
                          Error **errp)
{
    union {
        uint64_t ll[2];
        uint8_t b[16];
    } ivec;
    int i;
    int ret;

    for(i = 0; i < nb_sectors; i++) {
        ivec.ll[0] = cpu_to_le64(sector_num);
        ivec.ll[1] = 0;
        if (qcrypto_cipher_setiv(s->cipher,
                                 ivec.b, G_N_ELEMENTS(ivec.b),
                                 errp) < 0) {
            return -1;
        }
        if (enc) {
            ret = qcrypto_cipher_encrypt(s->cipher,
                                         in_buf,
                                         out_buf,
                                         512,
                                         errp);
        } else {
            ret = qcrypto_cipher_decrypt(s->cipher,
                                         in_buf,
                                         out_buf,
                                         512,
                                         errp);
        }
        if (ret < 0) {
            return -1;
        }
        sector_num++;
        in_buf += 512;
        out_buf += 512;
    }
    return 0;
}

static int coroutine_fn copy_sectors(BlockDriverState *bs,
                                     uint64_t start_sect,
                                     uint64_t cluster_offset,
                                     int n_start, int n_end)
{
    BDRVQcow2State *s = bs->opaque;
    QEMUIOVector qiov;
    struct iovec iov;
    int n, ret;

    n = n_end - n_start;
    if (n <= 0) {
        return 0;
    }

    iov.iov_len = n * BDRV_SECTOR_SIZE;
    iov.iov_base = qemu_try_blockalign(bs, iov.iov_len);
    if (iov.iov_base == NULL) {
        return -ENOMEM;
    }

    qemu_iovec_init_external(&qiov, &iov, 1);

    BLKDBG_EVENT(bs->file, BLKDBG_COW_READ);

    if (!bs->drv) {
        ret = -ENOMEDIUM;
        goto out;
    }

    /* Call .bdrv_co_readv() directly instead of using the public block-layer
     * interface.  This avoids double I/O throttling and request tracking,
     * which can lead to deadlock when block layer copy-on-read is enabled.
     */
    ret = bs->drv->bdrv_co_readv(bs, start_sect + n_start, n, &qiov);
    if (ret < 0) {
        goto out;
    }

    if (bs->encrypted) {
        Error *err = NULL;
        assert(s->cipher);
        if (qcow2_encrypt_sectors(s, start_sect + n_start,
                                  iov.iov_base, iov.iov_base, n,
                                  true, &err) < 0) {
            ret = -EIO;
            error_free(err);
            goto out;
        }
    }

    ret = qcow2_pre_write_overlap_check(bs, 0,
            cluster_offset + n_start * BDRV_SECTOR_SIZE, n * BDRV_SECTOR_SIZE);
    if (ret < 0) {
        goto out;
    }

    BLKDBG_EVENT(bs->file, BLKDBG_COW_WRITE);
    ret = bdrv_co_writev(bs->file->bs, (cluster_offset >> 9) + n_start, n,
                         &qiov);
    if (ret < 0) {
        goto out;
    }

    ret = 0;
out:
    qemu_vfree(iov.iov_base);
    return ret;
}


/*
 * get_cluster_offset
 *
 * For a given offset of the disk image, find the cluster offset in
 * qcow2 file. The offset is stored in *cluster_offset.
 *
 * on entry, *num is the number of contiguous sectors we'd like to
 * access following offset.
 *
 * on exit, *num is the number of contiguous sectors we can read.
 *
 * Returns the cluster type (QCOW2_CLUSTER_*) on success, -errno in error
 * cases.
 */
int qcow2_get_cluster_offset(BlockDriverState *bs, uint64_t offset,
    int *num, uint64_t *cluster_offset)
{
    BDRVQcow2State *s = bs->opaque;
    unsigned int l2_index;
    uint64_t l1_index, l2_offset, *l2_table;
    int l1_bits, c;
    unsigned int index_in_cluster, nb_clusters;
    uint64_t nb_available, nb_needed;
    int ret;

    index_in_cluster = (offset >> 9) & (s->cluster_sectors - 1);
    nb_needed = *num + index_in_cluster;

    l1_bits = s->l2_bits + s->cluster_bits;

    /* compute how many bytes there are between the offset and
     * the end of the l1 entry
     */

    nb_available = (1ULL << l1_bits) - (offset & ((1ULL << l1_bits) - 1));

    /* compute the number of available sectors */

    nb_available = (nb_available >> 9) + index_in_cluster;

    if (nb_needed > nb_available) {
        nb_needed = nb_available;
    }
    assert(nb_needed <= INT_MAX);

    *cluster_offset = 0;

    /* seek to the l2 offset in the l1 table */

    l1_index = offset >> l1_bits;
    if (l1_index >= s->l1_size) {
        ret = QCOW2_CLUSTER_UNALLOCATED;
        goto out;
    }

    l2_offset = s->l1_table[l1_index] & L1E_OFFSET_MASK;
    if (!l2_offset) {
        ret = QCOW2_CLUSTER_UNALLOCATED;
        goto out;
    }

    if (offset_into_cluster(s, l2_offset)) {
        qcow2_signal_corruption(bs, true, -1, -1, "L2 table offset %#" PRIx64
                                " unaligned (L1 index: %#" PRIx64 ")",
                                l2_offset, l1_index);
        return -EIO;
    }

    /* load the l2 table in memory */

    ret = l2_load(bs, l2_offset, &l2_table);
    if (ret < 0) {
        return ret;
    }

    /* find the cluster offset for the given disk offset */

    l2_index = (offset >> s->cluster_bits) & (s->l2_size - 1);
    *cluster_offset = be64_to_cpu(l2_table[l2_index]);

    /* nb_needed <= INT_MAX, thus nb_clusters <= INT_MAX, too */
    nb_clusters = size_to_clusters(s, nb_needed << 9);

    ret = qcow2_get_cluster_type(*cluster_offset);
    switch (ret) {
    case QCOW2_CLUSTER_COMPRESSED:
        /* Compressed clusters can only be processed one by one */
        c = 1;
        *cluster_offset &= L2E_COMPRESSED_OFFSET_SIZE_MASK;
        break;
    case QCOW2_CLUSTER_ZERO:
        if (s->qcow_version < 3) {
            qcow2_signal_corruption(bs, true, -1, -1, "Zero cluster entry found"
                                    " in pre-v3 image (L2 offset: %#" PRIx64
                                    ", L2 index: %#x)", l2_offset, l2_index);
            ret = -EIO;
            goto fail;
        }
        c = count_contiguous_clusters(nb_clusters, s->cluster_size,
                &l2_table[l2_index], QCOW_OFLAG_ZERO);
        *cluster_offset = 0;
        break;
    case QCOW2_CLUSTER_UNALLOCATED:
        /* how many empty clusters ? */
        c = count_contiguous_free_clusters(nb_clusters, &l2_table[l2_index]);
        *cluster_offset = 0;
        break;
    case QCOW2_CLUSTER_NORMAL:
        /* how many allocated clusters ? */
        c = count_contiguous_clusters(nb_clusters, s->cluster_size,
                &l2_table[l2_index], QCOW_OFLAG_ZERO);
        *cluster_offset &= L2E_OFFSET_MASK;
        if (offset_into_cluster(s, *cluster_offset)) {
            qcow2_signal_corruption(bs, true, -1, -1, "Data cluster offset %#"
                                    PRIx64 " unaligned (L2 offset: %#" PRIx64
                                    ", L2 index: %#x)", *cluster_offset,
                                    l2_offset, l2_index);
            ret = -EIO;
            goto fail;
        }
        break;
    default:
        abort();
    }

    qcow2_cache_put(bs, s->l2_table_cache, (void**) &l2_table);

    nb_available = (c * s->cluster_sectors);

out:
    if (nb_available > nb_needed)
        nb_available = nb_needed;

    *num = nb_available - index_in_cluster;

    return ret;

fail:
    qcow2_cache_put(bs, s->l2_table_cache, (void **)&l2_table);
    return ret;
}

/*
 * get_cluster_table
 *
 * for a given disk offset, load (and allocate if needed)
 * the l2 table.
 *
 * the l2 table offset in the qcow2 file and the cluster index
 * in the l2 table are given to the caller.
 *
 * Returns 0 on success, -errno in failure case
 */
static int get_cluster_table(BlockDriverState *bs, uint64_t offset,
                             uint64_t **new_l2_table,
                             int *new_l2_index)
{
    BDRVQcow2State *s = bs->opaque;
    unsigned int l2_index;
    uint64_t l1_index, l2_offset;
    uint64_t *l2_table = NULL;
    int ret;

    /* seek to the l2 offset in the l1 table */

    l1_index = offset >> (s->l2_bits + s->cluster_bits);
    if (l1_index >= s->l1_size) {
        ret = qcow2_grow_l1_table(bs, l1_index + 1, false);
        if (ret < 0) {
            return ret;
        }
    }

    assert(l1_index < s->l1_size);
    l2_offset = s->l1_table[l1_index] & L1E_OFFSET_MASK;
    if (offset_into_cluster(s, l2_offset)) {
        qcow2_signal_corruption(bs, true, -1, -1, "L2 table offset %#" PRIx64
                                " unaligned (L1 index: %#" PRIx64 ")",
                                l2_offset, l1_index);
        return -EIO;
    }

    /* seek the l2 table of the given l2 offset */

    if (s->l1_table[l1_index] & QCOW_OFLAG_COPIED) {
        /* load the l2 table in memory */
        ret = l2_load(bs, l2_offset, &l2_table);
        if (ret < 0) {
            return ret;
        }
    } else {
        /* First allocate a new L2 table (and do COW if needed) */
        ret = l2_allocate(bs, l1_index, &l2_table);
        if (ret < 0) {
            return ret;
        }

        /* Then decrease the refcount of the old table */
        if (l2_offset) {
            qcow2_free_clusters(bs, l2_offset, s->l2_size * sizeof(uint64_t),
                                QCOW2_DISCARD_OTHER);
        }
    }

    /* find the cluster offset for the given disk offset */

    l2_index = (offset >> s->cluster_bits) & (s->l2_size - 1);

    *new_l2_table = l2_table;
    *new_l2_index = l2_index;

    return 0;
}

/*
 * alloc_compressed_cluster_offset
 *
 * For a given offset of the disk image, return cluster offset in
 * qcow2 file.
 *
 * If the offset is not found, allocate a new compressed cluster.
 *
 * Return the cluster offset if successful,
 * Return 0, otherwise.
 *
 */

uint64_t qcow2_alloc_compressed_cluster_offset(BlockDriverState *bs,
                                               uint64_t offset,
                                               int compressed_size)
{
    BDRVQcow2State *s = bs->opaque;
    int l2_index, ret;
    uint64_t *l2_table;
    int64_t cluster_offset;
    int nb_csectors;

    ret = get_cluster_table(bs, offset, &l2_table, &l2_index);
    if (ret < 0) {
        return 0;
    }

    /* Compression can't overwrite anything. Fail if the cluster was already
     * allocated. */
    cluster_offset = be64_to_cpu(l2_table[l2_index]);
    if (cluster_offset & L2E_OFFSET_MASK) {
        qcow2_cache_put(bs, s->l2_table_cache, (void**) &l2_table);
        return 0;
    }

    cluster_offset = qcow2_alloc_bytes(bs, compressed_size);
    if (cluster_offset < 0) {
        qcow2_cache_put(bs, s->l2_table_cache, (void**) &l2_table);
        return 0;
    }

    nb_csectors = ((cluster_offset + compressed_size - 1) >> 9) -
                  (cluster_offset >> 9);

    cluster_offset |= QCOW_OFLAG_COMPRESSED |
                      ((uint64_t)nb_csectors << s->csize_shift);

    /* update L2 table */

    /* compressed clusters never have the copied flag */

    BLKDBG_EVENT(bs->file, BLKDBG_L2_UPDATE_COMPRESSED);
    qcow2_cache_entry_mark_dirty(bs, s->l2_table_cache, l2_table);
    l2_table[l2_index] = cpu_to_be64(cluster_offset);
    qcow2_cache_put(bs, s->l2_table_cache, (void **) &l2_table);

    return cluster_offset;
}

static int perform_cow(BlockDriverState *bs, QCowL2Meta *m, Qcow2COWRegion *r)
{
    BDRVQcow2State *s = bs->opaque;
    int ret;

    if (r->nb_sectors == 0) {
        return 0;
    }

    qemu_co_mutex_unlock(&s->lock);
    ret = copy_sectors(bs, m->offset / BDRV_SECTOR_SIZE, m->alloc_offset,
                       r->offset / BDRV_SECTOR_SIZE,
                       r->offset / BDRV_SECTOR_SIZE + r->nb_sectors);
    qemu_co_mutex_lock(&s->lock);

    if (ret < 0) {
        return ret;
    }

    /*
     * Before we update the L2 table to actually point to the new cluster, we
     * need to be sure that the refcounts have been increased and COW was
     * handled.
     */
    qcow2_cache_depends_on_flush(s->l2_table_cache);

    return 0;
}

int qcow2_alloc_cluster_link_l2(BlockDriverState *bs, QCowL2Meta *m)
{
    BDRVQcow2State *s = bs->opaque;
    int i, j = 0, l2_index, ret;
    uint64_t *old_cluster, *l2_table;
    uint64_t cluster_offset = m->alloc_offset;

    trace_qcow2_cluster_link_l2(qemu_coroutine_self(), m->nb_clusters);
    assert(m->nb_clusters > 0);

    old_cluster = g_try_new(uint64_t, m->nb_clusters);
    if (old_cluster == NULL) {
        ret = -ENOMEM;
        goto err;
    }

    /* copy content of unmodified sectors */
    ret = perform_cow(bs, m, &m->cow_start);
    if (ret < 0) {
        goto err;
    }

    ret = perform_cow(bs, m, &m->cow_end);
    if (ret < 0) {
        goto err;
    }

    /* Update L2 table. */
    if (s->use_lazy_refcounts) {
        qcow2_mark_dirty(bs);
    }
    if (qcow2_need_accurate_refcounts(s)) {
        qcow2_cache_set_dependency(bs, s->l2_table_cache,
                                   s->refcount_block_cache);
    }

    ret = get_cluster_table(bs, m->offset, &l2_table, &l2_index);
    if (ret < 0) {
        goto err;
    }
    qcow2_cache_entry_mark_dirty(bs, s->l2_table_cache, l2_table);

    assert(l2_index + m->nb_clusters <= s->l2_size);
    for (i = 0; i < m->nb_clusters; i++) {
        /* if two concurrent writes happen to the same unallocated cluster
	 * each write allocates separate cluster and writes data concurrently.
	 * The first one to complete updates l2 table with pointer to its
	 * cluster the second one has to do RMW (which is done above by
	 * copy_sectors()), update l2 table with its cluster pointer and free
	 * old cluster. This is what this loop does */
        if(l2_table[l2_index + i] != 0)
            old_cluster[j++] = l2_table[l2_index + i];

        l2_table[l2_index + i] = cpu_to_be64((cluster_offset +
                    (i << s->cluster_bits)) | QCOW_OFLAG_COPIED);
     }


    qcow2_cache_put(bs, s->l2_table_cache, (void **) &l2_table);

    /*
     * If this was a COW, we need to decrease the refcount of the old cluster.
     *
     * Don't discard clusters that reach a refcount of 0 (e.g. compressed
     * clusters), the next write will reuse them anyway.
     */
    if (j != 0) {
        for (i = 0; i < j; i++) {
            qcow2_free_any_clusters(bs, be64_to_cpu(old_cluster[i]), 1,
                                    QCOW2_DISCARD_NEVER);
        }
    }

    ret = 0;
err:
    g_free(old_cluster);
    return ret;
 }

/*
 * Returns the number of contiguous clusters that can be used for an allocating
 * write, but require COW to be performed (this includes yet unallocated space,
 * which must copy from the backing file)
 */
static int count_cow_clusters(BDRVQcow2State *s, int nb_clusters,
    uint64_t *l2_table, int l2_index)
{
    int i;

    for (i = 0; i < nb_clusters; i++) {
        uint64_t l2_entry = be64_to_cpu(l2_table[l2_index + i]);
        int cluster_type = qcow2_get_cluster_type(l2_entry);

        switch(cluster_type) {
        case QCOW2_CLUSTER_NORMAL:
            if (l2_entry & QCOW_OFLAG_COPIED) {
                goto out;
            }
            break;
        case QCOW2_CLUSTER_UNALLOCATED:
        case QCOW2_CLUSTER_COMPRESSED:
        case QCOW2_CLUSTER_ZERO:
            break;
        default:
            abort();
        }
    }

out:
    assert(i <= nb_clusters);
    return i;
}

/*
 * Check if there already is an AIO write request in flight which allocates
 * the same cluster. In this case we need to wait until the previous
 * request has completed and updated the L2 table accordingly.
 *
 * Returns:
 *   0       if there was no dependency. *cur_bytes indicates the number of
 *           bytes from guest_offset that can be read before the next
 *           dependency must be processed (or the request is complete)
 *
 *   -EAGAIN if we had to wait for another request, previously gathered
 *           information on cluster allocation may be invalid now. The caller
 *           must start over anyway, so consider *cur_bytes undefined.
 */
static int handle_dependencies(BlockDriverState *bs, uint64_t guest_offset,
    uint64_t *cur_bytes, QCowL2Meta **m)
{
    BDRVQcow2State *s = bs->opaque;
    QCowL2Meta *old_alloc;
    uint64_t bytes = *cur_bytes;

    QLIST_FOREACH(old_alloc, &s->cluster_allocs, next_in_flight) {

        uint64_t start = guest_offset;
        uint64_t end = start + bytes;
        uint64_t old_start = l2meta_cow_start(old_alloc);
        uint64_t old_end = l2meta_cow_end(old_alloc);

        if (end <= old_start || start >= old_end) {
            /* No intersection */
        } else {
            if (start < old_start) {
                /* Stop at the start of a running allocation */
                bytes = old_start - start;
            } else {
                bytes = 0;
            }

            /* Stop if already an l2meta exists. After yielding, it wouldn't
             * be valid any more, so we'd have to clean up the old L2Metas
             * and deal with requests depending on them before starting to
             * gather new ones. Not worth the trouble. */
            if (bytes == 0 && *m) {
                *cur_bytes = 0;
                return 0;
            }

            if (bytes == 0) {
                /* Wait for the dependency to complete. We need to recheck
                 * the free/allocated clusters when we continue. */
                qemu_co_mutex_unlock(&s->lock);
                qemu_co_queue_wait(&old_alloc->dependent_requests);
                qemu_co_mutex_lock(&s->lock);
                return -EAGAIN;
            }
        }
    }

    /* Make sure that existing clusters and new allocations are only used up to
     * the next dependency if we shortened the request above */
    *cur_bytes = bytes;

    return 0;
}

/*
 * Checks how many already allocated clusters that don't require a copy on
 * write there are at the given guest_offset (up to *bytes). If
 * *host_offset is not zero, only physically contiguous clusters beginning at
 * this host offset are counted.
 *
 * Note that guest_offset may not be cluster aligned. In this case, the
 * returned *host_offset points to exact byte referenced by guest_offset and
 * therefore isn't cluster aligned as well.
 *
 * Returns:
 *   0:     if no allocated clusters are available at the given offset.
 *          *bytes is normally unchanged. It is set to 0 if the cluster
 *          is allocated and doesn't need COW, but doesn't have the right
 *          physical offset.
 *
 *   1:     if allocated clusters that don't require a COW are available at
 *          the requested offset. *bytes may have decreased and describes
 *          the length of the area that can be written to.
 *
 *  -errno: in error cases
 */
static int handle_copied(BlockDriverState *bs, uint64_t guest_offset,
    uint64_t *host_offset, uint64_t *bytes, QCowL2Meta **m)
{
    BDRVQcow2State *s = bs->opaque;
    int l2_index;
    uint64_t cluster_offset;
    uint64_t *l2_table;
    uint64_t nb_clusters;
    unsigned int keep_clusters;
    int ret;

    trace_qcow2_handle_copied(qemu_coroutine_self(), guest_offset, *host_offset,
                              *bytes);

    assert(*host_offset == 0 ||    offset_into_cluster(s, guest_offset)
                                == offset_into_cluster(s, *host_offset));

    /*
     * Calculate the number of clusters to look for. We stop at L2 table
     * boundaries to keep things simple.
     */
    nb_clusters =
        size_to_clusters(s, offset_into_cluster(s, guest_offset) + *bytes);

    l2_index = offset_to_l2_index(s, guest_offset);
    nb_clusters = MIN(nb_clusters, s->l2_size - l2_index);
    assert(nb_clusters <= INT_MAX);

    /* Find L2 entry for the first involved cluster */
    ret = get_cluster_table(bs, guest_offset, &l2_table, &l2_index);
    if (ret < 0) {
        return ret;
    }

    cluster_offset = be64_to_cpu(l2_table[l2_index]);

    /* Check how many clusters are already allocated and don't need COW */
    if (qcow2_get_cluster_type(cluster_offset) == QCOW2_CLUSTER_NORMAL
        && (cluster_offset & QCOW_OFLAG_COPIED))
    {
        /* If a specific host_offset is required, check it */
        bool offset_matches =
            (cluster_offset & L2E_OFFSET_MASK) == *host_offset;

        if (offset_into_cluster(s, cluster_offset & L2E_OFFSET_MASK)) {
            qcow2_signal_corruption(bs, true, -1, -1, "Data cluster offset "
                                    "%#llx unaligned (guest offset: %#" PRIx64
                                    ")", cluster_offset & L2E_OFFSET_MASK,
                                    guest_offset);
            ret = -EIO;
            goto out;
        }

        if (*host_offset != 0 && !offset_matches) {
            *bytes = 0;
            ret = 0;
            goto out;
        }

        /* We keep all QCOW_OFLAG_COPIED clusters */
        keep_clusters =
            count_contiguous_clusters(nb_clusters, s->cluster_size,
                                      &l2_table[l2_index],
                                      QCOW_OFLAG_COPIED | QCOW_OFLAG_ZERO);
        assert(keep_clusters <= nb_clusters);

        *bytes = MIN(*bytes,
                 keep_clusters * s->cluster_size
                 - offset_into_cluster(s, guest_offset));

        ret = 1;
    } else {
        ret = 0;
    }

    /* Cleanup */
out:
    qcow2_cache_put(bs, s->l2_table_cache, (void **) &l2_table);

    /* Only return a host offset if we actually made progress. Otherwise we
     * would make requirements for handle_alloc() that it can't fulfill */
    if (ret > 0) {
        *host_offset = (cluster_offset & L2E_OFFSET_MASK)
                     + offset_into_cluster(s, guest_offset);
    }

    return ret;
}

/*
 * Allocates new clusters for the given guest_offset.
 *
 * At most *nb_clusters are allocated, and on return *nb_clusters is updated to
 * contain the number of clusters that have been allocated and are contiguous
 * in the image file.
 *
 * If *host_offset is non-zero, it specifies the offset in the image file at
 * which the new clusters must start. *nb_clusters can be 0 on return in this
 * case if the cluster at host_offset is already in use. If *host_offset is
 * zero, the clusters can be allocated anywhere in the image file.
 *
 * *host_offset is updated to contain the offset into the image file at which
 * the first allocated cluster starts.
 *
 * Return 0 on success and -errno in error cases. -EAGAIN means that the
 * function has been waiting for another request and the allocation must be
 * restarted, but the whole request should not be failed.
 */
static int do_alloc_cluster_offset(BlockDriverState *bs, uint64_t guest_offset,
                                   uint64_t *host_offset, uint64_t *nb_clusters)
{
    BDRVQcow2State *s = bs->opaque;

    trace_qcow2_do_alloc_clusters_offset(qemu_coroutine_self(), guest_offset,
                                         *host_offset, *nb_clusters);

    /* Allocate new clusters */
    trace_qcow2_cluster_alloc_phys(qemu_coroutine_self());
    if (*host_offset == 0) {
        int64_t cluster_offset =
            qcow2_alloc_clusters(bs, *nb_clusters * s->cluster_size);
        if (cluster_offset < 0) {
            return cluster_offset;
        }
        *host_offset = cluster_offset;
        return 0;
    } else {
        int64_t ret = qcow2_alloc_clusters_at(bs, *host_offset, *nb_clusters);
        if (ret < 0) {
            return ret;
        }
        *nb_clusters = ret;
        return 0;
    }
}

/*
 * Allocates new clusters for an area that either is yet unallocated or needs a
 * copy on write. If *host_offset is non-zero, clusters are only allocated if
 * the new allocation can match the specified host offset.
 *
 * Note that guest_offset may not be cluster aligned. In this case, the
 * returned *host_offset points to exact byte referenced by guest_offset and
 * therefore isn't cluster aligned as well.
 *
 * Returns:
 *   0:     if no clusters could be allocated. *bytes is set to 0,
 *          *host_offset is left unchanged.
 *
 *   1:     if new clusters were allocated. *bytes may be decreased if the
 *          new allocation doesn't cover all of the requested area.
 *          *host_offset is updated to contain the host offset of the first
 *          newly allocated cluster.
 *
 *  -errno: in error cases
 */
static int handle_alloc(BlockDriverState *bs, uint64_t guest_offset,
    uint64_t *host_offset, uint64_t *bytes, QCowL2Meta **m)
{
    BDRVQcow2State *s = bs->opaque;
    int l2_index;
    uint64_t *l2_table;
    uint64_t entry;
    uint64_t nb_clusters;
    int ret;

    uint64_t alloc_cluster_offset;

    trace_qcow2_handle_alloc(qemu_coroutine_self(), guest_offset, *host_offset,
                             *bytes);
    assert(*bytes > 0);

    /*
     * Calculate the number of clusters to look for. We stop at L2 table
     * boundaries to keep things simple.
     */
    nb_clusters =
        size_to_clusters(s, offset_into_cluster(s, guest_offset) + *bytes);

    l2_index = offset_to_l2_index(s, guest_offset);
    nb_clusters = MIN(nb_clusters, s->l2_size - l2_index);
    assert(nb_clusters <= INT_MAX);

    /* Find L2 entry for the first involved cluster */
    ret = get_cluster_table(bs, guest_offset, &l2_table, &l2_index);
    if (ret < 0) {
        return ret;
    }

    entry = be64_to_cpu(l2_table[l2_index]);

    /* For the moment, overwrite compressed clusters one by one */
    if (entry & QCOW_OFLAG_COMPRESSED) {
        nb_clusters = 1;
    } else {
        nb_clusters = count_cow_clusters(s, nb_clusters, l2_table, l2_index);
    }

    /* This function is only called when there were no non-COW clusters, so if
     * we can't find any unallocated or COW clusters either, something is
     * wrong with our code. */
    assert(nb_clusters > 0);

    qcow2_cache_put(bs, s->l2_table_cache, (void **) &l2_table);

    /* Allocate, if necessary at a given offset in the image file */
    alloc_cluster_offset = start_of_cluster(s, *host_offset);
    ret = do_alloc_cluster_offset(bs, guest_offset, &alloc_cluster_offset,
                                  &nb_clusters);
    if (ret < 0) {
        goto fail;
    }

    /* Can't extend contiguous allocation */
    if (nb_clusters == 0) {
        *bytes = 0;
        return 0;
    }

    /* !*host_offset would overwrite the image header and is reserved for "no
     * host offset preferred". If 0 was a valid host offset, it'd trigger the
     * following overlap check; do that now to avoid having an invalid value in
     * *host_offset. */
    if (!alloc_cluster_offset) {
        ret = qcow2_pre_write_overlap_check(bs, 0, alloc_cluster_offset,
                                            nb_clusters * s->cluster_size);
        assert(ret < 0);
        goto fail;
    }

    /*
     * Save info needed for meta data update.
     *
     * requested_sectors: Number of sectors from the start of the first
     * newly allocated cluster to the end of the (possibly shortened
     * before) write request.
     *
     * avail_sectors: Number of sectors from the start of the first
     * newly allocated to the end of the last newly allocated cluster.
     *
     * nb_sectors: The number of sectors from the start of the first
     * newly allocated cluster to the end of the area that the write
     * request actually writes to (excluding COW at the end)
     */
    int requested_sectors =
        (*bytes + offset_into_cluster(s, guest_offset))
        >> BDRV_SECTOR_BITS;
    int avail_sectors = nb_clusters
                        << (s->cluster_bits - BDRV_SECTOR_BITS);
    int alloc_n_start = offset_into_cluster(s, guest_offset)
                        >> BDRV_SECTOR_BITS;
    int nb_sectors = MIN(requested_sectors, avail_sectors);
    QCowL2Meta *old_m = *m;

    *m = g_malloc0(sizeof(**m));

    **m = (QCowL2Meta) {
        .next           = old_m,

        .alloc_offset   = alloc_cluster_offset,
        .offset         = start_of_cluster(s, guest_offset),
        .nb_clusters    = nb_clusters,
        .nb_available   = nb_sectors,

        .cow_start = {
            .offset     = 0,
            .nb_sectors = alloc_n_start,
        },
        .cow_end = {
            .offset     = nb_sectors * BDRV_SECTOR_SIZE,
            .nb_sectors = avail_sectors - nb_sectors,
        },
    };
    qemu_co_queue_init(&(*m)->dependent_requests);
    QLIST_INSERT_HEAD(&s->cluster_allocs, *m, next_in_flight);

    *host_offset = alloc_cluster_offset + offset_into_cluster(s, guest_offset);
    *bytes = MIN(*bytes, (nb_sectors * BDRV_SECTOR_SIZE)
                         - offset_into_cluster(s, guest_offset));
    assert(*bytes != 0);

    return 1;

fail:
    if (*m && (*m)->nb_clusters > 0) {
        QLIST_REMOVE(*m, next_in_flight);
    }
    return ret;
}

/*
 * alloc_cluster_offset
 *
 * For a given offset on the virtual disk, find the cluster offset in qcow2
 * file. If the offset is not found, allocate a new cluster.
 *
 * If the cluster was already allocated, m->nb_clusters is set to 0 and
 * other fields in m are meaningless.
 *
 * If the cluster is newly allocated, m->nb_clusters is set to the number of
 * contiguous clusters that have been allocated. In this case, the other
 * fields of m are valid and contain information about the first allocated
 * cluster.
 *
 * If the request conflicts with another write request in flight, the coroutine
 * is queued and will be reentered when the dependency has completed.
 *
 * Return 0 on success and -errno in error cases
 */
int qcow2_alloc_cluster_offset(BlockDriverState *bs, uint64_t offset,
    int *num, uint64_t *host_offset, QCowL2Meta **m)
{
    BDRVQcow2State *s = bs->opaque;
    uint64_t start, remaining;
    uint64_t cluster_offset;
    uint64_t cur_bytes;
    int ret;

    trace_qcow2_alloc_clusters_offset(qemu_coroutine_self(), offset, *num);

    assert((offset & ~BDRV_SECTOR_MASK) == 0);

again:
    start = offset;
    remaining = (uint64_t)*num << BDRV_SECTOR_BITS;
    cluster_offset = 0;
    *host_offset = 0;
    cur_bytes = 0;
    *m = NULL;

    while (true) {

        if (!*host_offset) {
            *host_offset = start_of_cluster(s, cluster_offset);
        }

        assert(remaining >= cur_bytes);

        start           += cur_bytes;
        remaining       -= cur_bytes;
        cluster_offset  += cur_bytes;

        if (remaining == 0) {
            break;
        }

        cur_bytes = remaining;

        /*
         * Now start gathering as many contiguous clusters as possible:
         *
         * 1. Check for overlaps with in-flight allocations
         *
         *      a) Overlap not in the first cluster -> shorten this request and
         *         let the caller handle the rest in its next loop iteration.
         *
         *      b) Real overlaps of two requests. Yield and restart the search
         *         for contiguous clusters (the situation could have changed
         *         while we were sleeping)
         *
         *      c) TODO: Request starts in the same cluster as the in-flight
         *         allocation ends. Shorten the COW of the in-fight allocation,
         *         set cluster_offset to write to the same cluster and set up
         *         the right synchronisation between the in-flight request and
         *         the new one.
         */
        ret = handle_dependencies(bs, start, &cur_bytes, m);
        if (ret == -EAGAIN) {
            /* Currently handle_dependencies() doesn't yield if we already had
             * an allocation. If it did, we would have to clean up the L2Meta
             * structs before starting over. */
            assert(*m == NULL);
            goto again;
        } else if (ret < 0) {
            return ret;
        } else if (cur_bytes == 0) {
            break;
        } else {
            /* handle_dependencies() may have decreased cur_bytes (shortened
             * the allocations below) so that the next dependency is processed
             * correctly during the next loop iteration. */
        }

        /*
         * 2. Count contiguous COPIED clusters.
         */
        ret = handle_copied(bs, start, &cluster_offset, &cur_bytes, m);
        if (ret < 0) {
            return ret;
        } else if (ret) {
            continue;
        } else if (cur_bytes == 0) {
            break;
        }

        /*
         * 3. If the request still hasn't completed, allocate new clusters,
         *    considering any cluster_offset of steps 1c or 2.
         */
        ret = handle_alloc(bs, start, &cluster_offset, &cur_bytes, m);
        if (ret < 0) {
            return ret;
        } else if (ret) {
            continue;
        } else {
            assert(cur_bytes == 0);
            break;
        }
    }

    *num -= remaining >> BDRV_SECTOR_BITS;
    assert(*num > 0);
    assert(*host_offset != 0);

    return 0;
}

static int decompress_buffer(uint8_t *out_buf, int out_buf_size,
                             const uint8_t *buf, int buf_size)
{
    z_stream strm1, *strm = &strm1;
    int ret, out_len;

    memset(strm, 0, sizeof(*strm));

    strm->next_in = (uint8_t *)buf;
    strm->avail_in = buf_size;
    strm->next_out = out_buf;
    strm->avail_out = out_buf_size;

    ret = inflateInit2(strm, -12);
    if (ret != Z_OK)
        return -1;
    ret = inflate(strm, Z_FINISH);
    out_len = strm->next_out - out_buf;
    if ((ret != Z_STREAM_END && ret != Z_BUF_ERROR) ||
        out_len != out_buf_size) {
        inflateEnd(strm);
        return -1;
    }
    inflateEnd(strm);
    return 0;
}

int qcow2_decompress_cluster(BlockDriverState *bs, uint64_t cluster_offset)
{
    BDRVQcow2State *s = bs->opaque;
    int ret, csize, nb_csectors, sector_offset;
    uint64_t coffset;

    coffset = cluster_offset & s->cluster_offset_mask;
    if (s->cluster_cache_offset != coffset) {
        nb_csectors = ((cluster_offset >> s->csize_shift) & s->csize_mask) + 1;
        sector_offset = coffset & 511;
        csize = nb_csectors * 512 - sector_offset;
        BLKDBG_EVENT(bs->file, BLKDBG_READ_COMPRESSED);
        ret = bdrv_read(bs->file->bs, coffset >> 9, s->cluster_data,
                        nb_csectors);
        if (ret < 0) {
            return ret;
        }
        if (decompress_buffer(s->cluster_cache, s->cluster_size,
                              s->cluster_data + sector_offset, csize) < 0) {
            return -EIO;
        }
        s->cluster_cache_offset = coffset;
    }
    return 0;
}

/*
 * This discards as many clusters of nb_clusters as possible at once (i.e.
 * all clusters in the same L2 table) and returns the number of discarded
 * clusters.
 */
static int discard_single_l2(BlockDriverState *bs, uint64_t offset,
                             uint64_t nb_clusters, enum qcow2_discard_type type,
                             bool full_discard)
{
    BDRVQcow2State *s = bs->opaque;
    uint64_t *l2_table;
    int l2_index;
    int ret;
    int i;

    ret = get_cluster_table(bs, offset, &l2_table, &l2_index);
    if (ret < 0) {
        return ret;
    }

    /* Limit nb_clusters to one L2 table */
    nb_clusters = MIN(nb_clusters, s->l2_size - l2_index);
    assert(nb_clusters <= INT_MAX);

    for (i = 0; i < nb_clusters; i++) {
        uint64_t old_l2_entry;

        old_l2_entry = be64_to_cpu(l2_table[l2_index + i]);

        /*
         * If full_discard is false, make sure that a discarded area reads back
         * as zeroes for v3 images (we cannot do it for v2 without actually
         * writing a zero-filled buffer). We can skip the operation if the
         * cluster is already marked as zero, or if it's unallocated and we
         * don't have a backing file.
         *
         * TODO We might want to use bdrv_get_block_status(bs) here, but we're
         * holding s->lock, so that doesn't work today.
         *
         * If full_discard is true, the sector should not read back as zeroes,
         * but rather fall through to the backing file.
         */
        switch (qcow2_get_cluster_type(old_l2_entry)) {
            case QCOW2_CLUSTER_UNALLOCATED:
                if (full_discard || !bs->backing) {
                    continue;
                }
                break;

            case QCOW2_CLUSTER_ZERO:
                if (!full_discard) {
                    continue;
                }
                break;

            case QCOW2_CLUSTER_NORMAL:
            case QCOW2_CLUSTER_COMPRESSED:
                break;

            default:
                abort();
        }

        /* First remove L2 entries */
        qcow2_cache_entry_mark_dirty(bs, s->l2_table_cache, l2_table);
        if (!full_discard && s->qcow_version >= 3) {
            l2_table[l2_index + i] = cpu_to_be64(QCOW_OFLAG_ZERO);
        } else {
            l2_table[l2_index + i] = cpu_to_be64(0);
        }

        /* Then decrease the refcount */
        qcow2_free_any_clusters(bs, old_l2_entry, 1, type);
    }

    qcow2_cache_put(bs, s->l2_table_cache, (void **) &l2_table);

    return nb_clusters;
}

int qcow2_discard_clusters(BlockDriverState *bs, uint64_t offset,
    int nb_sectors, enum qcow2_discard_type type, bool full_discard)
{
    BDRVQcow2State *s = bs->opaque;
    uint64_t end_offset;
    uint64_t nb_clusters;
    int ret;

    end_offset = offset + (nb_sectors << BDRV_SECTOR_BITS);

    /* Round start up and end down */
    offset = align_offset(offset, s->cluster_size);
    end_offset = start_of_cluster(s, end_offset);

    if (offset > end_offset) {
        return 0;
    }

    nb_clusters = size_to_clusters(s, end_offset - offset);

    s->cache_discards = true;

    /* Each L2 table is handled by its own loop iteration */
    while (nb_clusters > 0) {
        ret = discard_single_l2(bs, offset, nb_clusters, type, full_discard);
        if (ret < 0) {
            goto fail;
        }

        nb_clusters -= ret;
        offset += (ret * s->cluster_size);
    }

    ret = 0;
fail:
    s->cache_discards = false;
    qcow2_process_discards(bs, ret);

    return ret;
}

/*
 * This zeroes as many clusters of nb_clusters as possible at once (i.e.
 * all clusters in the same L2 table) and returns the number of zeroed
 * clusters.
 */
static int zero_single_l2(BlockDriverState *bs, uint64_t offset,
                          uint64_t nb_clusters)
{
    BDRVQcow2State *s = bs->opaque;
    uint64_t *l2_table;
    int l2_index;
    int ret;
    int i;

    ret = get_cluster_table(bs, offset, &l2_table, &l2_index);
    if (ret < 0) {
        return ret;
    }

    /* Limit nb_clusters to one L2 table */
    nb_clusters = MIN(nb_clusters, s->l2_size - l2_index);
    assert(nb_clusters <= INT_MAX);

    for (i = 0; i < nb_clusters; i++) {
        uint64_t old_offset;

        old_offset = be64_to_cpu(l2_table[l2_index + i]);

        /* Update L2 entries */
        qcow2_cache_entry_mark_dirty(bs, s->l2_table_cache, l2_table);
        if (old_offset & QCOW_OFLAG_COMPRESSED) {
            l2_table[l2_index + i] = cpu_to_be64(QCOW_OFLAG_ZERO);
            qcow2_free_any_clusters(bs, old_offset, 1, QCOW2_DISCARD_REQUEST);
        } else {
            l2_table[l2_index + i] |= cpu_to_be64(QCOW_OFLAG_ZERO);
        }
    }

    qcow2_cache_put(bs, s->l2_table_cache, (void **) &l2_table);

    return nb_clusters;
}

int qcow2_zero_clusters(BlockDriverState *bs, uint64_t offset, int nb_sectors)
{
    BDRVQcow2State *s = bs->opaque;
    uint64_t nb_clusters;
    int ret;

    /* The zero flag is only supported by version 3 and newer */
    if (s->qcow_version < 3) {
        return -ENOTSUP;
    }

    /* Each L2 table is handled by its own loop iteration */
    nb_clusters = size_to_clusters(s, nb_sectors << BDRV_SECTOR_BITS);

    s->cache_discards = true;

    while (nb_clusters > 0) {
        ret = zero_single_l2(bs, offset, nb_clusters);
        if (ret < 0) {
            goto fail;
        }

        nb_clusters -= ret;
        offset += (ret * s->cluster_size);
    }

    ret = 0;
fail:
    s->cache_discards = false;
    qcow2_process_discards(bs, ret);

    return ret;
}

/*
 * Expands all zero clusters in a specific L1 table (or deallocates them, for
 * non-backed non-pre-allocated zero clusters).
 *
 * l1_entries and *visited_l1_entries are used to keep track of progress for
 * status_cb(). l1_entries contains the total number of L1 entries and
 * *visited_l1_entries counts all visited L1 entries.
 */
static int expand_zero_clusters_in_l1(BlockDriverState *bs, uint64_t *l1_table,
                                      int l1_size, int64_t *visited_l1_entries,
                                      int64_t l1_entries,
                                      BlockDriverAmendStatusCB *status_cb)
{
    BDRVQcow2State *s = bs->opaque;
    bool is_active_l1 = (l1_table == s->l1_table);
    uint64_t *l2_table = NULL;
    int ret;
    int i, j;

    if (!is_active_l1) {
        /* inactive L2 tables require a buffer to be stored in when loading
         * them from disk */
        l2_table = qemu_try_blockalign(bs->file->bs, s->cluster_size);
        if (l2_table == NULL) {
            return -ENOMEM;
        }
    }

    for (i = 0; i < l1_size; i++) {
        uint64_t l2_offset = l1_table[i] & L1E_OFFSET_MASK;
        bool l2_dirty = false;
        uint64_t l2_refcount;

        if (!l2_offset) {
            /* unallocated */
            (*visited_l1_entries)++;
            if (status_cb) {
                status_cb(bs, *visited_l1_entries, l1_entries);
            }
            continue;
        }

        if (offset_into_cluster(s, l2_offset)) {
            qcow2_signal_corruption(bs, true, -1, -1, "L2 table offset %#"
                                    PRIx64 " unaligned (L1 index: %#x)",
                                    l2_offset, i);
            ret = -EIO;
            goto fail;
        }

        if (is_active_l1) {
            /* get active L2 tables from cache */
            ret = qcow2_cache_get(bs, s->l2_table_cache, l2_offset,
                    (void **)&l2_table);
        } else {
            /* load inactive L2 tables from disk */
            ret = bdrv_read(bs->file->bs, l2_offset / BDRV_SECTOR_SIZE,
                            (void *)l2_table, s->cluster_sectors);
        }
        if (ret < 0) {
            goto fail;
        }

        ret = qcow2_get_refcount(bs, l2_offset >> s->cluster_bits,
                                 &l2_refcount);
        if (ret < 0) {
            goto fail;
        }

        for (j = 0; j < s->l2_size; j++) {
            uint64_t l2_entry = be64_to_cpu(l2_table[j]);
            int64_t offset = l2_entry & L2E_OFFSET_MASK;
            int cluster_type = qcow2_get_cluster_type(l2_entry);
            bool preallocated = offset != 0;

            if (cluster_type != QCOW2_CLUSTER_ZERO) {
                continue;
            }

            if (!preallocated) {
                if (!bs->backing) {
                    /* not backed; therefore we can simply deallocate the
                     * cluster */
                    l2_table[j] = 0;
                    l2_dirty = true;
                    continue;
                }

                offset = qcow2_alloc_clusters(bs, s->cluster_size);
                if (offset < 0) {
                    ret = offset;
                    goto fail;
                }

                if (l2_refcount > 1) {
                    /* For shared L2 tables, set the refcount accordingly (it is
                     * already 1 and needs to be l2_refcount) */
                    ret = qcow2_update_cluster_refcount(bs,
                            offset >> s->cluster_bits,
                            refcount_diff(1, l2_refcount), false,
                            QCOW2_DISCARD_OTHER);
                    if (ret < 0) {
                        qcow2_free_clusters(bs, offset, s->cluster_size,
                                            QCOW2_DISCARD_OTHER);
                        goto fail;
                    }
                }
            }

            if (offset_into_cluster(s, offset)) {
                qcow2_signal_corruption(bs, true, -1, -1, "Data cluster offset "
                                        "%#" PRIx64 " unaligned (L2 offset: %#"
                                        PRIx64 ", L2 index: %#x)", offset,
                                        l2_offset, j);
                if (!preallocated) {
                    qcow2_free_clusters(bs, offset, s->cluster_size,
                                        QCOW2_DISCARD_ALWAYS);
                }
                ret = -EIO;
                goto fail;
            }

            ret = qcow2_pre_write_overlap_check(bs, 0, offset, s->cluster_size);
            if (ret < 0) {
                if (!preallocated) {
                    qcow2_free_clusters(bs, offset, s->cluster_size,
                                        QCOW2_DISCARD_ALWAYS);
                }
                goto fail;
            }

            ret = bdrv_write_zeroes(bs->file->bs, offset / BDRV_SECTOR_SIZE,
                                    s->cluster_sectors, 0);
            if (ret < 0) {
                if (!preallocated) {
                    qcow2_free_clusters(bs, offset, s->cluster_size,
                                        QCOW2_DISCARD_ALWAYS);
                }
                goto fail;
            }

            if (l2_refcount == 1) {
                l2_table[j] = cpu_to_be64(offset | QCOW_OFLAG_COPIED);
            } else {
                l2_table[j] = cpu_to_be64(offset);
            }
            l2_dirty = true;
        }

        if (is_active_l1) {
            if (l2_dirty) {
                qcow2_cache_entry_mark_dirty(bs, s->l2_table_cache, l2_table);
                qcow2_cache_depends_on_flush(s->l2_table_cache);
            }
            qcow2_cache_put(bs, s->l2_table_cache, (void **) &l2_table);
        } else {
            if (l2_dirty) {
                ret = qcow2_pre_write_overlap_check(bs,
                        QCOW2_OL_INACTIVE_L2 | QCOW2_OL_ACTIVE_L2, l2_offset,
                        s->cluster_size);
                if (ret < 0) {
                    goto fail;
                }

                ret = bdrv_write(bs->file->bs, l2_offset / BDRV_SECTOR_SIZE,
                                 (void *)l2_table, s->cluster_sectors);
                if (ret < 0) {
                    goto fail;
                }
            }
        }

        (*visited_l1_entries)++;
        if (status_cb) {
            status_cb(bs, *visited_l1_entries, l1_entries);
        }
    }

    ret = 0;

fail:
    if (l2_table) {
        if (!is_active_l1) {
            qemu_vfree(l2_table);
        } else {
            qcow2_cache_put(bs, s->l2_table_cache, (void **) &l2_table);
        }
    }
    return ret;
}

/*
 * For backed images, expands all zero clusters on the image. For non-backed
 * images, deallocates all non-pre-allocated zero clusters (and claims the
 * allocation for pre-allocated ones). This is important for downgrading to a
 * qcow2 version which doesn't yet support metadata zero clusters.
 */
int qcow2_expand_zero_clusters(BlockDriverState *bs,
                               BlockDriverAmendStatusCB *status_cb)
{
    BDRVQcow2State *s = bs->opaque;
    uint64_t *l1_table = NULL;
    int64_t l1_entries = 0, visited_l1_entries = 0;
    int ret;
    int i, j;

    if (status_cb) {
        l1_entries = s->l1_size;
        for (i = 0; i < s->nb_snapshots; i++) {
            l1_entries += s->snapshots[i].l1_size;
        }
    }

    ret = expand_zero_clusters_in_l1(bs, s->l1_table, s->l1_size,
                                     &visited_l1_entries, l1_entries,
                                     status_cb);
    if (ret < 0) {
        goto fail;
    }

    /* Inactive L1 tables may point to active L2 tables - therefore it is
     * necessary to flush the L2 table cache before trying to access the L2
     * tables pointed to by inactive L1 entries (else we might try to expand
     * zero clusters that have already been expanded); furthermore, it is also
     * necessary to empty the L2 table cache, since it may contain tables which
     * are now going to be modified directly on disk, bypassing the cache.
     * qcow2_cache_empty() does both for us. */
    ret = qcow2_cache_empty(bs, s->l2_table_cache);
    if (ret < 0) {
        goto fail;
    }

    for (i = 0; i < s->nb_snapshots; i++) {
        int l1_sectors = (s->snapshots[i].l1_size * sizeof(uint64_t) +
                BDRV_SECTOR_SIZE - 1) / BDRV_SECTOR_SIZE;

        l1_table = g_realloc(l1_table, l1_sectors * BDRV_SECTOR_SIZE);

        ret = bdrv_read(bs->file->bs,
                        s->snapshots[i].l1_table_offset / BDRV_SECTOR_SIZE,
                        (void *)l1_table, l1_sectors);
        if (ret < 0) {
            goto fail;
        }

        for (j = 0; j < s->snapshots[i].l1_size; j++) {
            be64_to_cpus(&l1_table[j]);
        }

        ret = expand_zero_clusters_in_l1(bs, l1_table, s->snapshots[i].l1_size,
                                         &visited_l1_entries, l1_entries,
                                         status_cb);
        if (ret < 0) {
            goto fail;
        }
    }

    ret = 0;

fail:
    g_free(l1_table);
    return ret;
}