qcow2.c 70.2 KB
Newer Older
B
bellard 已提交
1 2
/*
 * Block driver for the QCOW version 2 format
3
 *
B
bellard 已提交
4
 * Copyright (c) 2004-2006 Fabrice Bellard
5
 *
B
bellard 已提交
6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
 * 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.
 */
P
pbrook 已提交
24
#include "qemu-common.h"
25
#include "block/block_int.h"
26
#include "qemu/module.h"
B
bellard 已提交
27
#include <zlib.h>
28
#include "qemu/aes.h"
K
Kevin Wolf 已提交
29
#include "block/qcow2.h"
30
#include "qemu/error-report.h"
31
#include "qapi/qmp/qerror.h"
32
#include "qapi/qmp/qbool.h"
K
Kevin Wolf 已提交
33
#include "trace.h"
B
bellard 已提交
34 35 36 37 38 39 40 41

/*
  Differences with QCOW:

  - Support for multiple incremental snapshots.
  - Memory management by reference counts.
  - Clusters which have a reference count of one have the bit
    QCOW_OFLAG_COPIED to optimize write performance.
42
  - Size of compressed clusters is stored in sectors to reduce bit usage
B
bellard 已提交
43 44
    in the cluster offsets.
  - Support for storing additional data (such as the VM state) in the
45
    snapshots.
B
bellard 已提交
46 47 48 49 50
  - If a backing store is used, the cluster size is not constrained
    (could be backported to QCOW).
  - L2 tables have always a size of one cluster.
*/

51 52 53 54

typedef struct {
    uint32_t magic;
    uint32_t len;
55
} QEMU_PACKED QCowExtension;
J
Jeff Cody 已提交
56

57 58
#define  QCOW2_EXT_MAGIC_END 0
#define  QCOW2_EXT_MAGIC_BACKING_FORMAT 0xE2792ACA
59
#define  QCOW2_EXT_MAGIC_FEATURE_TABLE 0x6803f857
60

61
static int qcow2_probe(const uint8_t *buf, int buf_size, const char *filename)
B
bellard 已提交
62 63
{
    const QCowHeader *cow_header = (const void *)buf;
64

B
bellard 已提交
65 66
    if (buf_size >= sizeof(QCowHeader) &&
        be32_to_cpu(cow_header->magic) == QCOW_MAGIC &&
K
Kevin Wolf 已提交
67
        be32_to_cpu(cow_header->version) >= 2)
B
bellard 已提交
68 69 70 71 72
        return 100;
    else
        return 0;
}

73 74 75 76 77 78 79 80

/* 
 * read qcow2 extension and fill bs
 * start reading from start_offset
 * finish reading upon magic of value 0 or when end_offset reached
 * unknown magic is skipped (future extension this version knows nothing about)
 * return 0 upon success, non-0 otherwise
 */
81
static int qcow2_read_extensions(BlockDriverState *bs, uint64_t start_offset,
M
Max Reitz 已提交
82 83
                                 uint64_t end_offset, void **p_feature_table,
                                 Error **errp)
84
{
85
    BDRVQcowState *s = bs->opaque;
86 87
    QCowExtension ext;
    uint64_t offset;
88
    int ret;
89 90

#ifdef DEBUG_EXT
91
    printf("qcow2_read_extensions: start=%ld end=%ld\n", start_offset, end_offset);
92 93 94 95 96 97 98
#endif
    offset = start_offset;
    while (offset < end_offset) {

#ifdef DEBUG_EXT
        /* Sanity check */
        if (offset > s->cluster_size)
99
            printf("qcow2_read_extension: suspicious offset %lu\n", offset);
100

D
Dong Xu Wang 已提交
101
        printf("attempting to read extended header in offset %lu\n", offset);
102 103
#endif

M
Max Reitz 已提交
104 105 106 107
        ret = bdrv_pread(bs->file, offset, &ext, sizeof(ext));
        if (ret < 0) {
            error_setg_errno(errp, -ret, "qcow2_read_extension: ERROR: "
                             "pread fail from offset %" PRIu64, offset);
108 109 110 111 112 113 114 115
            return 1;
        }
        be32_to_cpus(&ext.magic);
        be32_to_cpus(&ext.len);
        offset += sizeof(ext);
#ifdef DEBUG_EXT
        printf("ext.magic = 0x%x\n", ext.magic);
#endif
116
        if (ext.len > end_offset - offset) {
M
Max Reitz 已提交
117
            error_setg(errp, "Header extension too large");
118 119 120
            return -EINVAL;
        }

121
        switch (ext.magic) {
122
        case QCOW2_EXT_MAGIC_END:
123
            return 0;
124

125
        case QCOW2_EXT_MAGIC_BACKING_FORMAT:
126
            if (ext.len >= sizeof(bs->backing_format)) {
M
Max Reitz 已提交
127 128
                error_setg(errp, "ERROR: ext_backing_format: len=%u too large"
                           " (>=%zu)", ext.len, sizeof(bs->backing_format));
129 130
                return 2;
            }
M
Max Reitz 已提交
131 132 133 134
            ret = bdrv_pread(bs->file, offset, bs->backing_format, ext.len);
            if (ret < 0) {
                error_setg_errno(errp, -ret, "ERROR: ext_backing_format: "
                                 "Could not read format name");
135
                return 3;
M
Max Reitz 已提交
136
            }
137 138 139 140 141 142
            bs->backing_format[ext.len] = '\0';
#ifdef DEBUG_EXT
            printf("Qcow2: Got format extension %s\n", bs->backing_format);
#endif
            break;

143 144 145 146 147
        case QCOW2_EXT_MAGIC_FEATURE_TABLE:
            if (p_feature_table != NULL) {
                void* feature_table = g_malloc0(ext.len + 2 * sizeof(Qcow2Feature));
                ret = bdrv_pread(bs->file, offset , feature_table, ext.len);
                if (ret < 0) {
M
Max Reitz 已提交
148 149
                    error_setg_errno(errp, -ret, "ERROR: ext_feature_table: "
                                     "Could not read table");
150 151 152 153 154 155 156
                    return ret;
                }

                *p_feature_table = feature_table;
            }
            break;

157
        default:
158 159 160 161 162 163 164 165 166 167 168
            /* unknown magic - save it in case we need to rewrite the header */
            {
                Qcow2UnknownHeaderExtension *uext;

                uext = g_malloc0(sizeof(*uext)  + ext.len);
                uext->magic = ext.magic;
                uext->len = ext.len;
                QLIST_INSERT_HEAD(&s->unknown_header_ext, uext, next);

                ret = bdrv_pread(bs->file, offset , uext->data, uext->len);
                if (ret < 0) {
M
Max Reitz 已提交
169 170
                    error_setg_errno(errp, -ret, "ERROR: unknown extension: "
                                     "Could not read data");
171 172 173
                    return ret;
                }
            }
174 175
            break;
        }
176 177

        offset += ((ext.len + 7) & ~7);
178 179 180 181 182
    }

    return 0;
}

183 184 185 186 187 188 189 190 191 192
static void cleanup_unknown_header_ext(BlockDriverState *bs)
{
    BDRVQcowState *s = bs->opaque;
    Qcow2UnknownHeaderExtension *uext, *next;

    QLIST_FOREACH_SAFE(uext, &s->unknown_header_ext, next, next) {
        QLIST_REMOVE(uext, next);
        g_free(uext);
    }
}
193

M
Max Reitz 已提交
194 195
static void GCC_FMT_ATTR(3, 4) report_unsupported(BlockDriverState *bs,
    Error **errp, const char *fmt, ...)
K
Kevin Wolf 已提交
196 197 198 199 200 201 202 203
{
    char msg[64];
    va_list ap;

    va_start(ap, fmt);
    vsnprintf(msg, sizeof(msg), fmt, ap);
    va_end(ap);

M
Max Reitz 已提交
204 205
    error_set(errp, QERR_UNKNOWN_BLOCK_FORMAT_FEATURE, bs->device_name, "qcow2",
              msg);
K
Kevin Wolf 已提交
206 207
}

208
static void report_unsupported_feature(BlockDriverState *bs,
M
Max Reitz 已提交
209
    Error **errp, Qcow2Feature *table, uint64_t mask)
210 211 212 213
{
    while (table && table->name[0] != '\0') {
        if (table->type == QCOW2_FEAT_TYPE_INCOMPATIBLE) {
            if (mask & (1 << table->bit)) {
M
Max Reitz 已提交
214
                report_unsupported(bs, errp, "%.46s", table->name);
215 216 217 218 219 220 221
                mask &= ~(1 << table->bit);
            }
        }
        table++;
    }

    if (mask) {
M
Max Reitz 已提交
222 223
        report_unsupported(bs, errp, "Unknown incompatible feature: %" PRIx64,
                           mask);
224 225 226
    }
}

227 228 229 230 231 232 233
/*
 * Sets the dirty bit and flushes afterwards if necessary.
 *
 * The incompatible_features bit is only set if the image file header was
 * updated successfully.  Therefore it is not required to check the return
 * value of this function.
 */
234
int qcow2_mark_dirty(BlockDriverState *bs)
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
{
    BDRVQcowState *s = bs->opaque;
    uint64_t val;
    int ret;

    assert(s->qcow_version >= 3);

    if (s->incompatible_features & QCOW2_INCOMPAT_DIRTY) {
        return 0; /* already dirty */
    }

    val = cpu_to_be64(s->incompatible_features | QCOW2_INCOMPAT_DIRTY);
    ret = bdrv_pwrite(bs->file, offsetof(QCowHeader, incompatible_features),
                      &val, sizeof(val));
    if (ret < 0) {
        return ret;
    }
    ret = bdrv_flush(bs->file);
    if (ret < 0) {
        return ret;
    }

    /* Only treat image as dirty if the header was updated successfully */
    s->incompatible_features |= QCOW2_INCOMPAT_DIRTY;
    return 0;
}

S
Stefan Hajnoczi 已提交
262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282
/*
 * Clears the dirty bit and flushes before if necessary.  Only call this
 * function when there are no pending requests, it does not guard against
 * concurrent requests dirtying the image.
 */
static int qcow2_mark_clean(BlockDriverState *bs)
{
    BDRVQcowState *s = bs->opaque;

    if (s->incompatible_features & QCOW2_INCOMPAT_DIRTY) {
        int ret = bdrv_flush(bs);
        if (ret < 0) {
            return ret;
        }

        s->incompatible_features &= ~QCOW2_INCOMPAT_DIRTY;
        return qcow2_update_header(bs);
    }
    return 0;
}

M
Max Reitz 已提交
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
/*
 * Marks the image as corrupt.
 */
int qcow2_mark_corrupt(BlockDriverState *bs)
{
    BDRVQcowState *s = bs->opaque;

    s->incompatible_features |= QCOW2_INCOMPAT_CORRUPT;
    return qcow2_update_header(bs);
}

/*
 * Marks the image as consistent, i.e., unsets the corrupt bit, and flushes
 * before if necessary.
 */
int qcow2_mark_consistent(BlockDriverState *bs)
{
    BDRVQcowState *s = bs->opaque;

    if (s->incompatible_features & QCOW2_INCOMPAT_CORRUPT) {
        int ret = bdrv_flush(bs);
        if (ret < 0) {
            return ret;
        }

        s->incompatible_features &= ~QCOW2_INCOMPAT_CORRUPT;
        return qcow2_update_header(bs);
    }
    return 0;
}

314 315 316 317 318 319 320 321 322
static int qcow2_check(BlockDriverState *bs, BdrvCheckResult *result,
                       BdrvCheckMode fix)
{
    int ret = qcow2_check_refcounts(bs, result, fix);
    if (ret < 0) {
        return ret;
    }

    if (fix && result->check_errors == 0 && result->corruptions == 0) {
M
Max Reitz 已提交
323 324 325 326 327
        ret = qcow2_mark_clean(bs);
        if (ret < 0) {
            return ret;
        }
        return qcow2_mark_consistent(bs);
328 329 330 331
    }
    return ret;
}

332 333 334 335 336
static QemuOptsList qcow2_runtime_opts = {
    .name = "qcow2",
    .head = QTAILQ_HEAD_INITIALIZER(qcow2_runtime_opts.head),
    .desc = {
        {
337
            .name = QCOW2_OPT_LAZY_REFCOUNTS,
338 339 340
            .type = QEMU_OPT_BOOL,
            .help = "Postpone refcount updates",
        },
341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356
        {
            .name = QCOW2_OPT_DISCARD_REQUEST,
            .type = QEMU_OPT_BOOL,
            .help = "Pass guest discard requests to the layer below",
        },
        {
            .name = QCOW2_OPT_DISCARD_SNAPSHOT,
            .type = QEMU_OPT_BOOL,
            .help = "Generate discard requests when snapshot related space "
                    "is freed",
        },
        {
            .name = QCOW2_OPT_DISCARD_OTHER,
            .type = QEMU_OPT_BOOL,
            .help = "Generate discard requests when other clusters are freed",
        },
M
Max Reitz 已提交
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
        {
            .name = QCOW2_OPT_OVERLAP,
            .type = QEMU_OPT_STRING,
            .help = "Selects which overlap checks to perform from a range of "
                    "templates (none, constant, cached, all)",
        },
        {
            .name = QCOW2_OPT_OVERLAP_MAIN_HEADER,
            .type = QEMU_OPT_BOOL,
            .help = "Check for unintended writes into the main qcow2 header",
        },
        {
            .name = QCOW2_OPT_OVERLAP_ACTIVE_L1,
            .type = QEMU_OPT_BOOL,
            .help = "Check for unintended writes into the active L1 table",
        },
        {
            .name = QCOW2_OPT_OVERLAP_ACTIVE_L2,
            .type = QEMU_OPT_BOOL,
            .help = "Check for unintended writes into an active L2 table",
        },
        {
            .name = QCOW2_OPT_OVERLAP_REFCOUNT_TABLE,
            .type = QEMU_OPT_BOOL,
            .help = "Check for unintended writes into the refcount table",
        },
        {
            .name = QCOW2_OPT_OVERLAP_REFCOUNT_BLOCK,
            .type = QEMU_OPT_BOOL,
            .help = "Check for unintended writes into a refcount block",
        },
        {
            .name = QCOW2_OPT_OVERLAP_SNAPSHOT_TABLE,
            .type = QEMU_OPT_BOOL,
            .help = "Check for unintended writes into the snapshot table",
        },
        {
            .name = QCOW2_OPT_OVERLAP_INACTIVE_L1,
            .type = QEMU_OPT_BOOL,
            .help = "Check for unintended writes into an inactive L1 table",
        },
        {
            .name = QCOW2_OPT_OVERLAP_INACTIVE_L2,
            .type = QEMU_OPT_BOOL,
            .help = "Check for unintended writes into an inactive L2 table",
        },
403 404 405 406
        { /* end of list */ }
    },
};

407 408 409 410 411 412 413 414 415 416 417
static const char *overlap_bool_option_names[QCOW2_OL_MAX_BITNR] = {
    [QCOW2_OL_MAIN_HEADER_BITNR]    = QCOW2_OPT_OVERLAP_MAIN_HEADER,
    [QCOW2_OL_ACTIVE_L1_BITNR]      = QCOW2_OPT_OVERLAP_ACTIVE_L1,
    [QCOW2_OL_ACTIVE_L2_BITNR]      = QCOW2_OPT_OVERLAP_ACTIVE_L2,
    [QCOW2_OL_REFCOUNT_TABLE_BITNR] = QCOW2_OPT_OVERLAP_REFCOUNT_TABLE,
    [QCOW2_OL_REFCOUNT_BLOCK_BITNR] = QCOW2_OPT_OVERLAP_REFCOUNT_BLOCK,
    [QCOW2_OL_SNAPSHOT_TABLE_BITNR] = QCOW2_OPT_OVERLAP_SNAPSHOT_TABLE,
    [QCOW2_OL_INACTIVE_L1_BITNR]    = QCOW2_OPT_OVERLAP_INACTIVE_L1,
    [QCOW2_OL_INACTIVE_L2_BITNR]    = QCOW2_OPT_OVERLAP_INACTIVE_L2,
};

M
Max Reitz 已提交
418 419
static int qcow2_open(BlockDriverState *bs, QDict *options, int flags,
                      Error **errp)
B
bellard 已提交
420 421
{
    BDRVQcowState *s = bs->opaque;
422
    int len, i, ret = 0;
B
bellard 已提交
423
    QCowHeader header;
424 425
    QemuOpts *opts;
    Error *local_err = NULL;
426
    uint64_t ext_end;
427
    uint64_t l1_vm_state_index;
M
Max Reitz 已提交
428 429
    const char *opt_overlap_check;
    int overlap_check_template = 0;
B
bellard 已提交
430

431 432
    ret = bdrv_pread(bs->file, 0, &header, sizeof(header));
    if (ret < 0) {
M
Max Reitz 已提交
433
        error_setg_errno(errp, -ret, "Could not read qcow2 header");
B
bellard 已提交
434
        goto fail;
435
    }
B
bellard 已提交
436 437 438 439 440 441 442 443 444 445 446 447 448
    be32_to_cpus(&header.magic);
    be32_to_cpus(&header.version);
    be64_to_cpus(&header.backing_file_offset);
    be32_to_cpus(&header.backing_file_size);
    be64_to_cpus(&header.size);
    be32_to_cpus(&header.cluster_bits);
    be32_to_cpus(&header.crypt_method);
    be64_to_cpus(&header.l1_table_offset);
    be32_to_cpus(&header.l1_size);
    be64_to_cpus(&header.refcount_table_offset);
    be32_to_cpus(&header.refcount_table_clusters);
    be64_to_cpus(&header.snapshots_offset);
    be32_to_cpus(&header.nb_snapshots);
449

K
Kevin Wolf 已提交
450
    if (header.magic != QCOW_MAGIC) {
M
Max Reitz 已提交
451
        error_setg(errp, "Image is not in qcow2 format");
452
        ret = -EMEDIUMTYPE;
B
bellard 已提交
453
        goto fail;
454
    }
K
Kevin Wolf 已提交
455
    if (header.version < 2 || header.version > 3) {
M
Max Reitz 已提交
456
        report_unsupported(bs, errp, "QCOW version %d", header.version);
K
Kevin Wolf 已提交
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
        ret = -ENOTSUP;
        goto fail;
    }

    s->qcow_version = header.version;

    /* Initialise version 3 header fields */
    if (header.version == 2) {
        header.incompatible_features    = 0;
        header.compatible_features      = 0;
        header.autoclear_features       = 0;
        header.refcount_order           = 4;
        header.header_length            = 72;
    } else {
        be64_to_cpus(&header.incompatible_features);
        be64_to_cpus(&header.compatible_features);
        be64_to_cpus(&header.autoclear_features);
        be32_to_cpus(&header.refcount_order);
        be32_to_cpus(&header.header_length);
    }

    if (header.header_length > sizeof(header)) {
        s->unknown_header_fields_size = header.header_length - sizeof(header);
        s->unknown_header_fields = g_malloc(s->unknown_header_fields_size);
        ret = bdrv_pread(bs->file, sizeof(header), s->unknown_header_fields,
                         s->unknown_header_fields_size);
        if (ret < 0) {
M
Max Reitz 已提交
484 485
            error_setg_errno(errp, -ret, "Could not read unknown qcow2 header "
                             "fields");
K
Kevin Wolf 已提交
486 487 488 489
            goto fail;
        }
    }

490 491 492 493 494 495
    if (header.backing_file_offset) {
        ext_end = header.backing_file_offset;
    } else {
        ext_end = 1 << header.cluster_bits;
    }

K
Kevin Wolf 已提交
496 497 498 499 500
    /* Handle feature bits */
    s->incompatible_features    = header.incompatible_features;
    s->compatible_features      = header.compatible_features;
    s->autoclear_features       = header.autoclear_features;

S
Stefan Hajnoczi 已提交
501
    if (s->incompatible_features & ~QCOW2_INCOMPAT_MASK) {
502 503
        void *feature_table = NULL;
        qcow2_read_extensions(bs, header.header_length, ext_end,
M
Max Reitz 已提交
504 505
                              &feature_table, NULL);
        report_unsupported_feature(bs, errp, feature_table,
S
Stefan Hajnoczi 已提交
506 507
                                   s->incompatible_features &
                                   ~QCOW2_INCOMPAT_MASK);
K
Kevin Wolf 已提交
508 509 510 511
        ret = -ENOTSUP;
        goto fail;
    }

M
Max Reitz 已提交
512 513 514 515
    if (s->incompatible_features & QCOW2_INCOMPAT_CORRUPT) {
        /* Corrupt images may not be written to unless they are being repaired
         */
        if ((flags & BDRV_O_RDWR) && !(flags & BDRV_O_CHECK)) {
M
Max Reitz 已提交
516 517
            error_setg(errp, "qcow2: Image is corrupt; cannot be opened "
                       "read/write");
M
Max Reitz 已提交
518 519 520 521 522
            ret = -EACCES;
            goto fail;
        }
    }

K
Kevin Wolf 已提交
523 524
    /* Check support for various header values */
    if (header.refcount_order != 4) {
M
Max Reitz 已提交
525
        report_unsupported(bs, errp, "%d bit reference counts",
K
Kevin Wolf 已提交
526
                           1 << header.refcount_order);
K
Kevin Wolf 已提交
527 528 529
        ret = -ENOTSUP;
        goto fail;
    }
530
    s->refcount_order = header.refcount_order;
K
Kevin Wolf 已提交
531

532
    if (header.cluster_bits < MIN_CLUSTER_BITS ||
533
        header.cluster_bits > MAX_CLUSTER_BITS) {
M
Max Reitz 已提交
534
        error_setg(errp, "Unsupported cluster size: 2^%i", header.cluster_bits);
535
        ret = -EINVAL;
B
bellard 已提交
536
        goto fail;
537 538
    }
    if (header.crypt_method > QCOW_CRYPT_AES) {
M
Max Reitz 已提交
539 540
        error_setg(errp, "Unsupported encryption method: %i",
                   header.crypt_method);
541
        ret = -EINVAL;
B
bellard 已提交
542
        goto fail;
543
    }
B
bellard 已提交
544
    s->crypt_method_header = header.crypt_method;
545
    if (s->crypt_method_header) {
B
bellard 已提交
546
        bs->encrypted = 1;
547
    }
B
bellard 已提交
548 549 550 551 552 553 554 555 556 557
    s->cluster_bits = header.cluster_bits;
    s->cluster_size = 1 << s->cluster_bits;
    s->cluster_sectors = 1 << (s->cluster_bits - 9);
    s->l2_bits = s->cluster_bits - 3; /* L2 is always one cluster */
    s->l2_size = 1 << s->l2_bits;
    bs->total_sectors = header.size / 512;
    s->csize_shift = (62 - (s->cluster_bits - 8));
    s->csize_mask = (1 << (s->cluster_bits - 8)) - 1;
    s->cluster_offset_mask = (1LL << s->csize_shift) - 1;
    s->refcount_table_offset = header.refcount_table_offset;
558
    s->refcount_table_size =
B
bellard 已提交
559 560 561 562 563 564 565
        header.refcount_table_clusters << (s->cluster_bits - 3);

    s->snapshots_offset = header.snapshots_offset;
    s->nb_snapshots = header.nb_snapshots;

    /* read the level 1 table */
    s->l1_size = header.l1_size;
566 567 568

    l1_vm_state_index = size_to_l1(s, header.size);
    if (l1_vm_state_index > INT_MAX) {
M
Max Reitz 已提交
569
        error_setg(errp, "Image is too big");
570 571 572 573 574
        ret = -EFBIG;
        goto fail;
    }
    s->l1_vm_state_index = l1_vm_state_index;

B
bellard 已提交
575 576
    /* the L1 table must contain at least enough entries to put
       header.size bytes */
577
    if (s->l1_size < s->l1_vm_state_index) {
M
Max Reitz 已提交
578
        error_setg(errp, "L1 table is too small");
579
        ret = -EINVAL;
B
bellard 已提交
580
        goto fail;
581
    }
B
bellard 已提交
582
    s->l1_table_offset = header.l1_table_offset;
583
    if (s->l1_size > 0) {
584
        s->l1_table = g_malloc0(
585
            align_offset(s->l1_size * sizeof(uint64_t), 512));
586 587 588
        ret = bdrv_pread(bs->file, s->l1_table_offset, s->l1_table,
                         s->l1_size * sizeof(uint64_t));
        if (ret < 0) {
M
Max Reitz 已提交
589
            error_setg_errno(errp, -ret, "Could not read L1 table");
590
            goto fail;
591
        }
592 593 594
        for(i = 0;i < s->l1_size; i++) {
            be64_to_cpus(&s->l1_table[i]);
        }
B
bellard 已提交
595
    }
K
Kevin Wolf 已提交
596 597

    /* alloc L2 table/refcount block cache */
598 599
    s->l2_table_cache = qcow2_cache_create(bs, L2_CACHE_SIZE);
    s->refcount_block_cache = qcow2_cache_create(bs, REFCOUNT_CACHE_SIZE);
K
Kevin Wolf 已提交
600

601
    s->cluster_cache = g_malloc(s->cluster_size);
B
bellard 已提交
602
    /* one more sector for decompressed data alignment */
603
    s->cluster_data = qemu_blockalign(bs, QCOW_MAX_CRYPT_CLUSTERS * s->cluster_size
604
                                  + 512);
B
bellard 已提交
605
    s->cluster_cache_offset = -1;
606
    s->flags = flags;
607

608 609
    ret = qcow2_refcount_init(bs);
    if (ret != 0) {
M
Max Reitz 已提交
610
        error_setg_errno(errp, -ret, "Could not initialize refcount handling");
B
bellard 已提交
611
        goto fail;
612
    }
B
bellard 已提交
613

B
Blue Swirl 已提交
614
    QLIST_INIT(&s->cluster_allocs);
K
Kevin Wolf 已提交
615
    QTAILQ_INIT(&s->discards);
616

617
    /* read qcow2 extensions */
M
Max Reitz 已提交
618 619 620
    if (qcow2_read_extensions(bs, header.header_length, ext_end, NULL,
        &local_err)) {
        error_propagate(errp, local_err);
621
        ret = -EINVAL;
622
        goto fail;
623
    }
624

B
bellard 已提交
625 626 627
    /* read the backing file name */
    if (header.backing_file_offset != 0) {
        len = header.backing_file_size;
628
        if (len > 1023) {
B
bellard 已提交
629
            len = 1023;
630 631 632 633
        }
        ret = bdrv_pread(bs->file, header.backing_file_offset,
                         bs->backing_file, len);
        if (ret < 0) {
M
Max Reitz 已提交
634
            error_setg_errno(errp, -ret, "Could not read backing file name");
B
bellard 已提交
635
            goto fail;
636
        }
B
bellard 已提交
637 638
        bs->backing_file[len] = '\0';
    }
639 640 641

    ret = qcow2_read_snapshots(bs);
    if (ret < 0) {
M
Max Reitz 已提交
642
        error_setg_errno(errp, -ret, "Could not read snapshots");
B
bellard 已提交
643
        goto fail;
644
    }
B
bellard 已提交
645

646 647 648 649 650
    /* Clear unknown autoclear feature bits */
    if (!bs->read_only && s->autoclear_features != 0) {
        s->autoclear_features = 0;
        ret = qcow2_update_header(bs);
        if (ret < 0) {
M
Max Reitz 已提交
651
            error_setg_errno(errp, -ret, "Could not update qcow2 header");
652 653 654 655
            goto fail;
        }
    }

K
Kevin Wolf 已提交
656 657 658
    /* Initialise locks */
    qemu_co_mutex_init(&s->lock);

S
Stefan Hajnoczi 已提交
659
    /* Repair image if dirty */
660 661
    if (!(flags & BDRV_O_CHECK) && !bs->read_only &&
        (s->incompatible_features & QCOW2_INCOMPAT_DIRTY)) {
S
Stefan Hajnoczi 已提交
662 663
        BdrvCheckResult result = {0};

664
        ret = qcow2_check(bs, &result, BDRV_FIX_ERRORS);
S
Stefan Hajnoczi 已提交
665
        if (ret < 0) {
M
Max Reitz 已提交
666
            error_setg_errno(errp, -ret, "Could not repair dirty image");
S
Stefan Hajnoczi 已提交
667 668 669 670
            goto fail;
        }
    }

671 672 673 674
    /* Enable lazy_refcounts according to image and command line options */
    opts = qemu_opts_create_nofail(&qcow2_runtime_opts);
    qemu_opts_absorb_qdict(opts, options, &local_err);
    if (error_is_set(&local_err)) {
M
Max Reitz 已提交
675
        error_propagate(errp, local_err);
676 677 678 679
        ret = -EINVAL;
        goto fail;
    }

680
    s->use_lazy_refcounts = qemu_opt_get_bool(opts, QCOW2_OPT_LAZY_REFCOUNTS,
681 682
        (s->compatible_features & QCOW2_COMPAT_LAZY_REFCOUNTS));

683 684 685 686 687 688 689 690 691 692
    s->discard_passthrough[QCOW2_DISCARD_NEVER] = false;
    s->discard_passthrough[QCOW2_DISCARD_ALWAYS] = true;
    s->discard_passthrough[QCOW2_DISCARD_REQUEST] =
        qemu_opt_get_bool(opts, QCOW2_OPT_DISCARD_REQUEST,
                          flags & BDRV_O_UNMAP);
    s->discard_passthrough[QCOW2_DISCARD_SNAPSHOT] =
        qemu_opt_get_bool(opts, QCOW2_OPT_DISCARD_SNAPSHOT, true);
    s->discard_passthrough[QCOW2_DISCARD_OTHER] =
        qemu_opt_get_bool(opts, QCOW2_OPT_DISCARD_OTHER, false);

M
Max Reitz 已提交
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
    opt_overlap_check = qemu_opt_get(opts, "overlap-check") ?: "cached";
    if (!strcmp(opt_overlap_check, "none")) {
        overlap_check_template = 0;
    } else if (!strcmp(opt_overlap_check, "constant")) {
        overlap_check_template = QCOW2_OL_CONSTANT;
    } else if (!strcmp(opt_overlap_check, "cached")) {
        overlap_check_template = QCOW2_OL_CACHED;
    } else if (!strcmp(opt_overlap_check, "all")) {
        overlap_check_template = QCOW2_OL_ALL;
    } else {
        error_setg(errp, "Unsupported value '%s' for qcow2 option "
                   "'overlap-check'. Allowed are either of the following: "
                   "none, constant, cached, all", opt_overlap_check);
        qemu_opts_del(opts);
        ret = -EINVAL;
        goto fail;
    }

    s->overlap_check = 0;
    for (i = 0; i < QCOW2_OL_MAX_BITNR; i++) {
        /* overlap-check defines a template bitmask, but every flag may be
         * overwritten through the associated boolean option */
        s->overlap_check |=
            qemu_opt_get_bool(opts, overlap_bool_option_names[i],
                              overlap_check_template & (1 << i)) << i;
    }
719

720 721 722
    qemu_opts_del(opts);

    if (s->use_lazy_refcounts && s->qcow_version < 3) {
M
Max Reitz 已提交
723 724
        error_setg(errp, "Lazy refcounts require a qcow2 image with at least "
                   "qemu 1.1 compatibility level");
725 726 727 728
        ret = -EINVAL;
        goto fail;
    }

B
bellard 已提交
729
#ifdef DEBUG_ALLOC
P
Philipp Hahn 已提交
730 731
    {
        BdrvCheckResult result = {0};
732
        qcow2_check_refcounts(bs, &result, 0);
P
Philipp Hahn 已提交
733
    }
B
bellard 已提交
734
#endif
735
    return ret;
B
bellard 已提交
736 737

 fail:
K
Kevin Wolf 已提交
738
    g_free(s->unknown_header_fields);
739
    cleanup_unknown_header_ext(bs);
K
Kevin Wolf 已提交
740 741
    qcow2_free_snapshots(bs);
    qcow2_refcount_close(bs);
742
    g_free(s->l1_table);
743 744
    /* else pre-write overlap checks in cache_destroy may crash */
    s->l1_table = NULL;
K
Kevin Wolf 已提交
745 746 747
    if (s->l2_table_cache) {
        qcow2_cache_destroy(bs, s->l2_table_cache);
    }
748
    g_free(s->cluster_cache);
749
    qemu_vfree(s->cluster_data);
750
    return ret;
B
bellard 已提交
751 752
}

753
static int qcow2_set_key(BlockDriverState *bs, const char *key)
B
bellard 已提交
754 755 756 757
{
    BDRVQcowState *s = bs->opaque;
    uint8_t keybuf[16];
    int len, i;
758

B
bellard 已提交
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
    memset(keybuf, 0, 16);
    len = strlen(key);
    if (len > 16)
        len = 16;
    /* XXX: we could compress the chars to 7 bits to increase
       entropy */
    for(i = 0;i < len;i++) {
        keybuf[i] = key[i];
    }
    s->crypt_method = s->crypt_method_header;

    if (AES_set_encrypt_key(keybuf, 128, &s->aes_encrypt_key) != 0)
        return -1;
    if (AES_set_decrypt_key(keybuf, 128, &s->aes_decrypt_key) != 0)
        return -1;
#if 0
    /* test */
    {
        uint8_t in[16];
        uint8_t out[16];
        uint8_t tmp[16];
        for(i=0;i<16;i++)
            in[i] = i;
        AES_encrypt(in, tmp, &s->aes_encrypt_key);
        AES_decrypt(tmp, out, &s->aes_decrypt_key);
        for(i = 0; i < 16; i++)
            printf(" %02x", tmp[i]);
        printf("\n");
        for(i = 0; i < 16; i++)
            printf(" %02x", out[i]);
        printf("\n");
    }
#endif
    return 0;
}

J
Jeff Cody 已提交
795 796 797 798 799 800 801 802
/* We have nothing to do for QCOW2 reopen, stubs just return
 * success */
static int qcow2_reopen_prepare(BDRVReopenState *state,
                                BlockReopenQueue *queue, Error **errp)
{
    return 0;
}

803
static int64_t coroutine_fn qcow2_co_get_block_status(BlockDriverState *bs,
804
        int64_t sector_num, int nb_sectors, int *pnum)
B
bellard 已提交
805
{
806
    BDRVQcowState *s = bs->opaque;
B
bellard 已提交
807
    uint64_t cluster_offset;
808 809
    int index_in_cluster, ret;
    int64_t status = 0;
B
bellard 已提交
810

811
    *pnum = nb_sectors;
812
    qemu_co_mutex_lock(&s->lock);
813
    ret = qcow2_get_cluster_offset(bs, sector_num << 9, pnum, &cluster_offset);
814
    qemu_co_mutex_unlock(&s->lock);
815
    if (ret < 0) {
816
        return ret;
817
    }
818

819 820 821 822 823 824 825 826 827 828 829 830
    if (cluster_offset != 0 && ret != QCOW2_CLUSTER_COMPRESSED &&
        !s->crypt_method) {
        index_in_cluster = sector_num & (s->cluster_sectors - 1);
        cluster_offset |= (index_in_cluster << BDRV_SECTOR_BITS);
        status |= BDRV_BLOCK_OFFSET_VALID | cluster_offset;
    }
    if (ret == QCOW2_CLUSTER_ZERO) {
        status |= BDRV_BLOCK_ZERO;
    } else if (ret != QCOW2_CLUSTER_UNALLOCATED) {
        status |= BDRV_BLOCK_DATA;
    }
    return status;
B
bellard 已提交
831 832
}

B
bellard 已提交
833
/* handle reading after the end of the backing file */
834 835
int qcow2_backing_read1(BlockDriverState *bs, QEMUIOVector *qiov,
                  int64_t sector_num, int nb_sectors)
B
bellard 已提交
836 837 838 839 840 841 842 843
{
    int n1;
    if ((sector_num + nb_sectors) <= bs->total_sectors)
        return nb_sectors;
    if (sector_num >= bs->total_sectors)
        n1 = 0;
    else
        n1 = bs->total_sectors - sector_num;
844

845
    qemu_iovec_memset(qiov, 512 * n1, 0, 512 * (nb_sectors - n1));
846

B
bellard 已提交
847 848 849
    return n1;
}

850
static coroutine_fn int qcow2_co_readv(BlockDriverState *bs, int64_t sector_num,
851
                          int remaining_sectors, QEMUIOVector *qiov)
B
bellard 已提交
852 853
{
    BDRVQcowState *s = bs->opaque;
B
bellard 已提交
854
    int index_in_cluster, n1;
K
Kevin Wolf 已提交
855
    int ret;
856
    int cur_nr_sectors; /* number of sectors in current iteration */
857
    uint64_t cluster_offset = 0;
858 859 860
    uint64_t bytes_done = 0;
    QEMUIOVector hd_qiov;
    uint8_t *cluster_data = NULL;
B
bellard 已提交
861

862 863 864 865 866
    qemu_iovec_init(&hd_qiov, qiov->niov);

    qemu_co_mutex_lock(&s->lock);

    while (remaining_sectors != 0) {
867

868
        /* prepare next request */
869
        cur_nr_sectors = remaining_sectors;
870 871 872
        if (s->crypt_method) {
            cur_nr_sectors = MIN(cur_nr_sectors,
                QCOW_MAX_CRYPT_CLUSTERS * s->cluster_sectors);
B
bellard 已提交
873
        }
874

875
        ret = qcow2_get_cluster_offset(bs, sector_num << 9,
876
            &cur_nr_sectors, &cluster_offset);
877
        if (ret < 0) {
878
            goto fail;
879
        }
880

881
        index_in_cluster = sector_num & (s->cluster_sectors - 1);
882

883
        qemu_iovec_reset(&hd_qiov);
884
        qemu_iovec_concat(&hd_qiov, qiov, bytes_done,
885 886
            cur_nr_sectors * 512);

887 888
        switch (ret) {
        case QCOW2_CLUSTER_UNALLOCATED:
889 890 891

            if (bs->backing_hd) {
                /* read from the base image */
892 893
                n1 = qcow2_backing_read1(bs->backing_hd, &hd_qiov,
                    sector_num, cur_nr_sectors);
894 895 896
                if (n1 > 0) {
                    BLKDBG_EVENT(bs->file, BLKDBG_READ_BACKING_AIO);
                    qemu_co_mutex_unlock(&s->lock);
897 898
                    ret = bdrv_co_readv(bs->backing_hd, sector_num,
                                        n1, &hd_qiov);
899 900
                    qemu_co_mutex_lock(&s->lock);
                    if (ret < 0) {
901
                        goto fail;
902 903 904 905
                    }
                }
            } else {
                /* Note: in this case, no need to wait */
906
                qemu_iovec_memset(&hd_qiov, 0, 0, 512 * cur_nr_sectors);
907
            }
908 909
            break;

910
        case QCOW2_CLUSTER_ZERO:
911
            qemu_iovec_memset(&hd_qiov, 0, 0, 512 * cur_nr_sectors);
912 913
            break;

914
        case QCOW2_CLUSTER_COMPRESSED:
915 916 917
            /* add AIO support for compressed blocks ? */
            ret = qcow2_decompress_cluster(bs, cluster_offset);
            if (ret < 0) {
918
                goto fail;
919 920
            }

921
            qemu_iovec_from_buf(&hd_qiov, 0,
922
                s->cluster_cache + index_in_cluster * 512,
923
                512 * cur_nr_sectors);
924 925 926
            break;

        case QCOW2_CLUSTER_NORMAL:
927
            if ((cluster_offset & 511) != 0) {
928 929
                ret = -EIO;
                goto fail;
930
            }
931

932 933 934 935 936
            if (s->crypt_method) {
                /*
                 * For encrypted images, read everything into a temporary
                 * contiguous buffer on which the AES functions can work.
                 */
937 938
                if (!cluster_data) {
                    cluster_data =
939
                        qemu_blockalign(bs, QCOW_MAX_CRYPT_CLUSTERS * s->cluster_size);
940 941 942 943
                }

                assert(cur_nr_sectors <=
                    QCOW_MAX_CRYPT_CLUSTERS * s->cluster_sectors);
944 945
                qemu_iovec_reset(&hd_qiov);
                qemu_iovec_add(&hd_qiov, cluster_data,
946 947 948 949 950 951 952
                    512 * cur_nr_sectors);
            }

            BLKDBG_EVENT(bs->file, BLKDBG_READ_AIO);
            qemu_co_mutex_unlock(&s->lock);
            ret = bdrv_co_readv(bs->file,
                                (cluster_offset >> 9) + index_in_cluster,
953
                                cur_nr_sectors, &hd_qiov);
954 955
            qemu_co_mutex_lock(&s->lock);
            if (ret < 0) {
956
                goto fail;
957 958
            }
            if (s->crypt_method) {
959 960
                qcow2_encrypt_sectors(s, sector_num,  cluster_data,
                    cluster_data, cur_nr_sectors, 0, &s->aes_decrypt_key);
961 962
                qemu_iovec_from_buf(qiov, bytes_done,
                    cluster_data, 512 * cur_nr_sectors);
963
            }
964 965 966 967 968 969
            break;

        default:
            g_assert_not_reached();
            ret = -EIO;
            goto fail;
970
        }
971

972 973 974
        remaining_sectors -= cur_nr_sectors;
        sector_num += cur_nr_sectors;
        bytes_done += cur_nr_sectors * 512;
975
    }
976
    ret = 0;
977

978
fail:
K
Kevin Wolf 已提交
979
    qemu_co_mutex_unlock(&s->lock);
K
Kevin Wolf 已提交
980

981
    qemu_iovec_destroy(&hd_qiov);
982
    qemu_vfree(cluster_data);
K
Kevin Wolf 已提交
983 984

    return ret;
B
bellard 已提交
985 986
}

987
static coroutine_fn int qcow2_co_writev(BlockDriverState *bs,
988 989 990
                           int64_t sector_num,
                           int remaining_sectors,
                           QEMUIOVector *qiov)
B
bellard 已提交
991 992 993
{
    BDRVQcowState *s = bs->opaque;
    int index_in_cluster;
994
    int n_end;
K
Kevin Wolf 已提交
995
    int ret;
996
    int cur_nr_sectors; /* number of sectors in current iteration */
997
    uint64_t cluster_offset;
998 999 1000
    QEMUIOVector hd_qiov;
    uint64_t bytes_done = 0;
    uint8_t *cluster_data = NULL;
1001
    QCowL2Meta *l2meta = NULL;
1002

K
Kevin Wolf 已提交
1003 1004 1005
    trace_qcow2_writev_start_req(qemu_coroutine_self(), sector_num,
                                 remaining_sectors);

1006 1007 1008
    qemu_iovec_init(&hd_qiov, qiov->niov);

    s->cluster_cache_offset = -1; /* disable compressed cache */
1009

1010 1011 1012 1013
    qemu_co_mutex_lock(&s->lock);

    while (remaining_sectors != 0) {

1014
        l2meta = NULL;
K
Kevin Wolf 已提交
1015

K
Kevin Wolf 已提交
1016
        trace_qcow2_writev_start_part(qemu_coroutine_self());
1017 1018
        index_in_cluster = sector_num & (s->cluster_sectors - 1);
        n_end = index_in_cluster + remaining_sectors;
1019 1020 1021 1022
        if (s->crypt_method &&
            n_end > QCOW_MAX_CRYPT_CLUSTERS * s->cluster_sectors) {
            n_end = QCOW_MAX_CRYPT_CLUSTERS * s->cluster_sectors;
        }
1023

1024
        ret = qcow2_alloc_cluster_offset(bs, sector_num << 9,
1025
            index_in_cluster, n_end, &cur_nr_sectors, &cluster_offset, &l2meta);
1026
        if (ret < 0) {
1027
            goto fail;
1028
        }
1029

1030
        assert((cluster_offset & 511) == 0);
1031

1032
        qemu_iovec_reset(&hd_qiov);
1033
        qemu_iovec_concat(&hd_qiov, qiov, bytes_done,
1034
            cur_nr_sectors * 512);
1035

1036
        if (s->crypt_method) {
1037
            if (!cluster_data) {
1038
                cluster_data = qemu_blockalign(bs, QCOW_MAX_CRYPT_CLUSTERS *
1039 1040
                                                 s->cluster_size);
            }
1041

1042
            assert(hd_qiov.size <=
1043
                   QCOW_MAX_CRYPT_CLUSTERS * s->cluster_size);
1044
            qemu_iovec_to_buf(&hd_qiov, 0, cluster_data, hd_qiov.size);
1045

1046 1047
            qcow2_encrypt_sectors(s, sector_num, cluster_data,
                cluster_data, cur_nr_sectors, 1, &s->aes_encrypt_key);
1048

1049 1050
            qemu_iovec_reset(&hd_qiov);
            qemu_iovec_add(&hd_qiov, cluster_data,
1051 1052
                cur_nr_sectors * 512);
        }
1053

M
Max Reitz 已提交
1054
        ret = qcow2_pre_write_overlap_check(bs, 0,
1055 1056 1057 1058 1059 1060
                cluster_offset + index_in_cluster * BDRV_SECTOR_SIZE,
                cur_nr_sectors * BDRV_SECTOR_SIZE);
        if (ret < 0) {
            goto fail;
        }

1061
        qemu_co_mutex_unlock(&s->lock);
1062
        BLKDBG_EVENT(bs->file, BLKDBG_WRITE_AIO);
K
Kevin Wolf 已提交
1063 1064
        trace_qcow2_writev_data(qemu_coroutine_self(),
                                (cluster_offset >> 9) + index_in_cluster);
1065 1066
        ret = bdrv_co_writev(bs->file,
                             (cluster_offset >> 9) + index_in_cluster,
1067
                             cur_nr_sectors, &hd_qiov);
1068 1069
        qemu_co_mutex_lock(&s->lock);
        if (ret < 0) {
1070
            goto fail;
1071
        }
1072

1073 1074 1075
        while (l2meta != NULL) {
            QCowL2Meta *next;

1076 1077 1078 1079
            ret = qcow2_alloc_cluster_link_l2(bs, l2meta);
            if (ret < 0) {
                goto fail;
            }
1080

1081 1082 1083 1084 1085 1086 1087
            /* Take the request off the list of running requests */
            if (l2meta->nb_clusters != 0) {
                QLIST_REMOVE(l2meta, next_in_flight);
            }

            qemu_co_queue_restart_all(&l2meta->dependent_requests);

1088
            next = l2meta->next;
1089
            g_free(l2meta);
1090
            l2meta = next;
1091
        }
1092

1093 1094 1095
        remaining_sectors -= cur_nr_sectors;
        sector_num += cur_nr_sectors;
        bytes_done += cur_nr_sectors * 512;
K
Kevin Wolf 已提交
1096
        trace_qcow2_writev_done_part(qemu_coroutine_self(), cur_nr_sectors);
1097
    }
1098
    ret = 0;
1099

1100
fail:
1101 1102
    qemu_co_mutex_unlock(&s->lock);

1103 1104 1105
    while (l2meta != NULL) {
        QCowL2Meta *next;

1106 1107 1108 1109
        if (l2meta->nb_clusters != 0) {
            QLIST_REMOVE(l2meta, next_in_flight);
        }
        qemu_co_queue_restart_all(&l2meta->dependent_requests);
1110 1111

        next = l2meta->next;
K
Kevin Wolf 已提交
1112
        g_free(l2meta);
1113
        l2meta = next;
K
Kevin Wolf 已提交
1114
    }
1115

1116
    qemu_iovec_destroy(&hd_qiov);
1117
    qemu_vfree(cluster_data);
K
Kevin Wolf 已提交
1118
    trace_qcow2_writev_done_req(qemu_coroutine_self(), ret);
K
Kevin Wolf 已提交
1119

K
Kevin Wolf 已提交
1120
    return ret;
B
bellard 已提交
1121 1122
}

1123
static void qcow2_close(BlockDriverState *bs)
B
bellard 已提交
1124 1125
{
    BDRVQcowState *s = bs->opaque;
1126
    g_free(s->l1_table);
1127 1128
    /* else pre-write overlap checks in cache_destroy may crash */
    s->l1_table = NULL;
K
Kevin Wolf 已提交
1129 1130 1131 1132

    qcow2_cache_flush(bs, s->l2_table_cache);
    qcow2_cache_flush(bs, s->refcount_block_cache);

S
Stefan Hajnoczi 已提交
1133 1134
    qcow2_mark_clean(bs);

K
Kevin Wolf 已提交
1135 1136 1137
    qcow2_cache_destroy(bs, s->l2_table_cache);
    qcow2_cache_destroy(bs, s->refcount_block_cache);

K
Kevin Wolf 已提交
1138
    g_free(s->unknown_header_fields);
1139
    cleanup_unknown_header_ext(bs);
K
Kevin Wolf 已提交
1140

1141
    g_free(s->cluster_cache);
1142
    qemu_vfree(s->cluster_data);
K
Kevin Wolf 已提交
1143
    qcow2_refcount_close(bs);
1144
    qcow2_free_snapshots(bs);
B
bellard 已提交
1145 1146
}

1147 1148 1149 1150 1151 1152 1153
static void qcow2_invalidate_cache(BlockDriverState *bs)
{
    BDRVQcowState *s = bs->opaque;
    int flags = s->flags;
    AES_KEY aes_encrypt_key;
    AES_KEY aes_decrypt_key;
    uint32_t crypt_method = 0;
1154
    QDict *options;
1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168

    /*
     * Backing files are read-only which makes all of their metadata immutable,
     * that means we don't have to worry about reopening them here.
     */

    if (s->crypt_method) {
        crypt_method = s->crypt_method;
        memcpy(&aes_encrypt_key, &s->aes_encrypt_key, sizeof(aes_encrypt_key));
        memcpy(&aes_decrypt_key, &s->aes_decrypt_key, sizeof(aes_decrypt_key));
    }

    qcow2_close(bs);

1169 1170 1171 1172
    options = qdict_new();
    qdict_put(options, QCOW2_OPT_LAZY_REFCOUNTS,
              qbool_from_int(s->use_lazy_refcounts));

1173
    memset(s, 0, sizeof(BDRVQcowState));
M
Max Reitz 已提交
1174
    qcow2_open(bs, options, flags, NULL);
1175 1176

    QDECREF(options);
1177 1178 1179 1180 1181 1182 1183 1184

    if (crypt_method) {
        s->crypt_method = crypt_method;
        memcpy(&s->aes_encrypt_key, &aes_encrypt_key, sizeof(aes_encrypt_key));
        memcpy(&s->aes_decrypt_key, &aes_decrypt_key, sizeof(aes_decrypt_key));
    }
}

K
Kevin Wolf 已提交
1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203
static size_t header_ext_add(char *buf, uint32_t magic, const void *s,
    size_t len, size_t buflen)
{
    QCowExtension *ext_backing_fmt = (QCowExtension*) buf;
    size_t ext_len = sizeof(QCowExtension) + ((len + 7) & ~7);

    if (buflen < ext_len) {
        return -ENOSPC;
    }

    *ext_backing_fmt = (QCowExtension) {
        .magic  = cpu_to_be32(magic),
        .len    = cpu_to_be32(len),
    };
    memcpy(buf + sizeof(QCowExtension), s, len);

    return ext_len;
}

K
Kevin Wolf 已提交
1204
/*
K
Kevin Wolf 已提交
1205 1206 1207 1208
 * Updates the qcow2 header, including the variable length parts of it, i.e.
 * the backing file name and all extensions. qcow2 was not designed to allow
 * such changes, so if we run out of space (we can only use the first cluster)
 * this function may fail.
K
Kevin Wolf 已提交
1209 1210 1211
 *
 * Returns 0 on success, -errno in error cases.
 */
K
Kevin Wolf 已提交
1212
int qcow2_update_header(BlockDriverState *bs)
K
Kevin Wolf 已提交
1213 1214
{
    BDRVQcowState *s = bs->opaque;
K
Kevin Wolf 已提交
1215 1216 1217
    QCowHeader *header;
    char *buf;
    size_t buflen = s->cluster_size;
K
Kevin Wolf 已提交
1218
    int ret;
K
Kevin Wolf 已提交
1219 1220
    uint64_t total_size;
    uint32_t refcount_table_clusters;
K
Kevin Wolf 已提交
1221
    size_t header_length;
1222
    Qcow2UnknownHeaderExtension *uext;
K
Kevin Wolf 已提交
1223

K
Kevin Wolf 已提交
1224
    buf = qemu_blockalign(bs, buflen);
K
Kevin Wolf 已提交
1225

K
Kevin Wolf 已提交
1226 1227
    /* Header structure */
    header = (QCowHeader*) buf;
K
Kevin Wolf 已提交
1228

K
Kevin Wolf 已提交
1229 1230 1231
    if (buflen < sizeof(*header)) {
        ret = -ENOSPC;
        goto fail;
K
Kevin Wolf 已提交
1232 1233
    }

K
Kevin Wolf 已提交
1234
    header_length = sizeof(*header) + s->unknown_header_fields_size;
K
Kevin Wolf 已提交
1235 1236 1237 1238
    total_size = bs->total_sectors * BDRV_SECTOR_SIZE;
    refcount_table_clusters = s->refcount_table_size >> (s->cluster_bits - 3);

    *header = (QCowHeader) {
K
Kevin Wolf 已提交
1239
        /* Version 2 fields */
K
Kevin Wolf 已提交
1240
        .magic                  = cpu_to_be32(QCOW_MAGIC),
K
Kevin Wolf 已提交
1241
        .version                = cpu_to_be32(s->qcow_version),
K
Kevin Wolf 已提交
1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252
        .backing_file_offset    = 0,
        .backing_file_size      = 0,
        .cluster_bits           = cpu_to_be32(s->cluster_bits),
        .size                   = cpu_to_be64(total_size),
        .crypt_method           = cpu_to_be32(s->crypt_method_header),
        .l1_size                = cpu_to_be32(s->l1_size),
        .l1_table_offset        = cpu_to_be64(s->l1_table_offset),
        .refcount_table_offset  = cpu_to_be64(s->refcount_table_offset),
        .refcount_table_clusters = cpu_to_be32(refcount_table_clusters),
        .nb_snapshots           = cpu_to_be32(s->nb_snapshots),
        .snapshots_offset       = cpu_to_be64(s->snapshots_offset),
K
Kevin Wolf 已提交
1253 1254 1255 1256 1257

        /* Version 3 fields */
        .incompatible_features  = cpu_to_be64(s->incompatible_features),
        .compatible_features    = cpu_to_be64(s->compatible_features),
        .autoclear_features     = cpu_to_be64(s->autoclear_features),
1258
        .refcount_order         = cpu_to_be32(s->refcount_order),
K
Kevin Wolf 已提交
1259
        .header_length          = cpu_to_be32(header_length),
K
Kevin Wolf 已提交
1260
    };
K
Kevin Wolf 已提交
1261

K
Kevin Wolf 已提交
1262 1263 1264 1265 1266 1267 1268 1269 1270
    /* For older versions, write a shorter header */
    switch (s->qcow_version) {
    case 2:
        ret = offsetof(QCowHeader, incompatible_features);
        break;
    case 3:
        ret = sizeof(*header);
        break;
    default:
1271 1272
        ret = -EINVAL;
        goto fail;
K
Kevin Wolf 已提交
1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289
    }

    buf += ret;
    buflen -= ret;
    memset(buf, 0, buflen);

    /* Preserve any unknown field in the header */
    if (s->unknown_header_fields_size) {
        if (buflen < s->unknown_header_fields_size) {
            ret = -ENOSPC;
            goto fail;
        }

        memcpy(buf, s->unknown_header_fields, s->unknown_header_fields_size);
        buf += s->unknown_header_fields_size;
        buflen -= s->unknown_header_fields_size;
    }
K
Kevin Wolf 已提交
1290

K
Kevin Wolf 已提交
1291 1292 1293 1294 1295 1296 1297
    /* Backing file format header extension */
    if (*bs->backing_format) {
        ret = header_ext_add(buf, QCOW2_EXT_MAGIC_BACKING_FORMAT,
                             bs->backing_format, strlen(bs->backing_format),
                             buflen);
        if (ret < 0) {
            goto fail;
K
Kevin Wolf 已提交
1298 1299
        }

K
Kevin Wolf 已提交
1300 1301
        buf += ret;
        buflen -= ret;
K
Kevin Wolf 已提交
1302 1303
    }

1304 1305
    /* Feature table */
    Qcow2Feature features[] = {
S
Stefan Hajnoczi 已提交
1306 1307 1308 1309 1310
        {
            .type = QCOW2_FEAT_TYPE_INCOMPATIBLE,
            .bit  = QCOW2_INCOMPAT_DIRTY_BITNR,
            .name = "dirty bit",
        },
M
Max Reitz 已提交
1311 1312 1313 1314 1315
        {
            .type = QCOW2_FEAT_TYPE_INCOMPATIBLE,
            .bit  = QCOW2_INCOMPAT_CORRUPT_BITNR,
            .name = "corrupt bit",
        },
1316 1317 1318 1319 1320
        {
            .type = QCOW2_FEAT_TYPE_COMPATIBLE,
            .bit  = QCOW2_COMPAT_LAZY_REFCOUNTS_BITNR,
            .name = "lazy refcounts",
        },
1321 1322 1323 1324 1325 1326 1327 1328 1329 1330
    };

    ret = header_ext_add(buf, QCOW2_EXT_MAGIC_FEATURE_TABLE,
                         features, sizeof(features), buflen);
    if (ret < 0) {
        goto fail;
    }
    buf += ret;
    buflen -= ret;

1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341
    /* Keep unknown header extensions */
    QLIST_FOREACH(uext, &s->unknown_header_ext, next) {
        ret = header_ext_add(buf, uext->magic, uext->data, uext->len, buflen);
        if (ret < 0) {
            goto fail;
        }

        buf += ret;
        buflen -= ret;
    }

K
Kevin Wolf 已提交
1342 1343
    /* End of header extensions */
    ret = header_ext_add(buf, QCOW2_EXT_MAGIC_END, NULL, 0, buflen);
K
Kevin Wolf 已提交
1344 1345 1346 1347
    if (ret < 0) {
        goto fail;
    }

K
Kevin Wolf 已提交
1348 1349
    buf += ret;
    buflen -= ret;
K
Kevin Wolf 已提交
1350

K
Kevin Wolf 已提交
1351 1352 1353 1354 1355 1356 1357 1358 1359
    /* Backing file name */
    if (*bs->backing_file) {
        size_t backing_file_len = strlen(bs->backing_file);

        if (buflen < backing_file_len) {
            ret = -ENOSPC;
            goto fail;
        }

1360
        /* Using strncpy is ok here, since buf is not NUL-terminated. */
K
Kevin Wolf 已提交
1361 1362 1363 1364
        strncpy(buf, bs->backing_file, buflen);

        header->backing_file_offset = cpu_to_be64(buf - ((char*) header));
        header->backing_file_size   = cpu_to_be32(backing_file_len);
K
Kevin Wolf 已提交
1365 1366
    }

K
Kevin Wolf 已提交
1367 1368
    /* Write the new header */
    ret = bdrv_pwrite(bs->file, 0, header, s->cluster_size);
K
Kevin Wolf 已提交
1369 1370 1371 1372 1373 1374
    if (ret < 0) {
        goto fail;
    }

    ret = 0;
fail:
K
Kevin Wolf 已提交
1375
    qemu_vfree(header);
K
Kevin Wolf 已提交
1376 1377 1378 1379 1380 1381
    return ret;
}

static int qcow2_change_backing_file(BlockDriverState *bs,
    const char *backing_file, const char *backing_fmt)
{
K
Kevin Wolf 已提交
1382 1383 1384 1385
    pstrcpy(bs->backing_file, sizeof(bs->backing_file), backing_file ?: "");
    pstrcpy(bs->backing_format, sizeof(bs->backing_format), backing_fmt ?: "");

    return qcow2_update_header(bs);
K
Kevin Wolf 已提交
1386 1387
}

K
Kevin Wolf 已提交
1388 1389 1390 1391
static int preallocate(BlockDriverState *bs)
{
    uint64_t nb_sectors;
    uint64_t offset;
K
Kevin Wolf 已提交
1392
    uint64_t host_offset = 0;
K
Kevin Wolf 已提交
1393
    int num;
1394
    int ret;
1395
    QCowL2Meta *meta;
K
Kevin Wolf 已提交
1396 1397 1398 1399 1400 1401

    nb_sectors = bdrv_getlength(bs) >> 9;
    offset = 0;

    while (nb_sectors) {
        num = MIN(nb_sectors, INT_MAX >> 9);
K
Kevin Wolf 已提交
1402 1403
        ret = qcow2_alloc_cluster_offset(bs, offset, 0, num, &num,
                                         &host_offset, &meta);
1404
        if (ret < 0) {
1405
            return ret;
K
Kevin Wolf 已提交
1406 1407
        }

1408
        ret = qcow2_alloc_cluster_link_l2(bs, meta);
1409
        if (ret < 0) {
1410 1411
            qcow2_free_any_clusters(bs, meta->alloc_offset, meta->nb_clusters,
                                    QCOW2_DISCARD_NEVER);
1412
            return ret;
K
Kevin Wolf 已提交
1413 1414
        }

1415 1416
        /* There are no dependent requests, but we need to remove our request
         * from the list of in-flight requests */
1417
        if (meta != NULL) {
1418
            QLIST_REMOVE(meta, next_in_flight);
1419
        }
1420

K
Kevin Wolf 已提交
1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431
        /* TODO Preallocate data if requested */

        nb_sectors -= num;
        offset += num << 9;
    }

    /*
     * It is expected that the image file is large enough to actually contain
     * all of the allocated clusters (otherwise we get failing reads after
     * EOF). Extend the image to the last allocated sector.
     */
K
Kevin Wolf 已提交
1432
    if (host_offset != 0) {
K
Kevin Wolf 已提交
1433 1434
        uint8_t buf[512];
        memset(buf, 0, 512);
K
Kevin Wolf 已提交
1435
        ret = bdrv_write(bs->file, (host_offset >> 9) + num - 1, buf, 1);
1436 1437 1438
        if (ret < 0) {
            return ret;
        }
K
Kevin Wolf 已提交
1439 1440 1441 1442 1443
    }

    return 0;
}

1444 1445 1446
static int qcow2_create2(const char *filename, int64_t total_size,
                         const char *backing_file, const char *backing_format,
                         int flags, size_t cluster_size, int prealloc,
M
Max Reitz 已提交
1447 1448
                         QEMUOptionParameter *options, int version,
                         Error **errp)
K
Kevin Wolf 已提交
1449
{
D
Dong Xu Wang 已提交
1450
    /* Calculate cluster_bits */
K
Kevin Wolf 已提交
1451 1452 1453 1454 1455
    int cluster_bits;
    cluster_bits = ffs(cluster_size) - 1;
    if (cluster_bits < MIN_CLUSTER_BITS || cluster_bits > MAX_CLUSTER_BITS ||
        (1 << cluster_bits) != cluster_size)
    {
M
Max Reitz 已提交
1456 1457
        error_setg(errp, "Cluster size must be a power of two between %d and "
                   "%dk", 1 << MIN_CLUSTER_BITS, 1 << (MAX_CLUSTER_BITS - 10));
K
Kevin Wolf 已提交
1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475
        return -EINVAL;
    }

    /*
     * Open the image file and write a minimal qcow2 header.
     *
     * We keep things simple and start with a zero-sized image. We also
     * do without refcount blocks or a L1 table for now. We'll fix the
     * inconsistency later.
     *
     * We do need a refcount table because growing the refcount table means
     * allocating two new refcount blocks - the seconds of which would be at
     * 2 GB for 64k clusters, and we don't want to have a 2 GB initial file
     * size for any qcow2 image.
     */
    BlockDriverState* bs;
    QCowHeader header;
    uint8_t* refcount_table;
M
Max Reitz 已提交
1476
    Error *local_err = NULL;
K
Kevin Wolf 已提交
1477 1478
    int ret;

M
Max Reitz 已提交
1479
    ret = bdrv_create_file(filename, options, &local_err);
K
Kevin Wolf 已提交
1480
    if (ret < 0) {
M
Max Reitz 已提交
1481
        error_propagate(errp, local_err);
K
Kevin Wolf 已提交
1482 1483 1484
        return ret;
    }

M
Max Reitz 已提交
1485
    ret = bdrv_file_open(&bs, filename, NULL, BDRV_O_RDWR, &local_err);
K
Kevin Wolf 已提交
1486
    if (ret < 0) {
M
Max Reitz 已提交
1487
        error_propagate(errp, local_err);
K
Kevin Wolf 已提交
1488 1489 1490 1491 1492 1493
        return ret;
    }

    /* Write the header */
    memset(&header, 0, sizeof(header));
    header.magic = cpu_to_be32(QCOW_MAGIC);
K
Kevin Wolf 已提交
1494
    header.version = cpu_to_be32(version);
K
Kevin Wolf 已提交
1495 1496 1497 1498 1499 1500
    header.cluster_bits = cpu_to_be32(cluster_bits);
    header.size = cpu_to_be64(0);
    header.l1_table_offset = cpu_to_be64(0);
    header.l1_size = cpu_to_be32(0);
    header.refcount_table_offset = cpu_to_be64(cluster_size);
    header.refcount_table_clusters = cpu_to_be32(1);
K
Kevin Wolf 已提交
1501 1502
    header.refcount_order = cpu_to_be32(3 + REFCOUNT_SHIFT);
    header.header_length = cpu_to_be32(sizeof(header));
K
Kevin Wolf 已提交
1503 1504 1505 1506 1507 1508 1509

    if (flags & BLOCK_FLAG_ENCRYPT) {
        header.crypt_method = cpu_to_be32(QCOW_CRYPT_AES);
    } else {
        header.crypt_method = cpu_to_be32(QCOW_CRYPT_NONE);
    }

1510 1511 1512 1513 1514
    if (flags & BLOCK_FLAG_LAZY_REFCOUNTS) {
        header.compatible_features |=
            cpu_to_be64(QCOW2_COMPAT_LAZY_REFCOUNTS);
    }

K
Kevin Wolf 已提交
1515 1516
    ret = bdrv_pwrite(bs, 0, &header, sizeof(header));
    if (ret < 0) {
M
Max Reitz 已提交
1517
        error_setg_errno(errp, -ret, "Could not write qcow2 header");
K
Kevin Wolf 已提交
1518 1519 1520 1521
        goto out;
    }

    /* Write an empty refcount table */
1522
    refcount_table = g_malloc0(cluster_size);
K
Kevin Wolf 已提交
1523
    ret = bdrv_pwrite(bs, cluster_size, refcount_table, cluster_size);
1524
    g_free(refcount_table);
K
Kevin Wolf 已提交
1525 1526

    if (ret < 0) {
M
Max Reitz 已提交
1527
        error_setg_errno(errp, -ret, "Could not write refcount table");
K
Kevin Wolf 已提交
1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539
        goto out;
    }

    bdrv_close(bs);

    /*
     * And now open the image and make it consistent first (i.e. increase the
     * refcount of the cluster that is occupied by the header and the refcount
     * table)
     */
    BlockDriver* drv = bdrv_find_format("qcow2");
    assert(drv != NULL);
1540
    ret = bdrv_open(bs, filename, NULL,
M
Max Reitz 已提交
1541
        BDRV_O_RDWR | BDRV_O_CACHE_WB | BDRV_O_NO_FLUSH, drv, &local_err);
K
Kevin Wolf 已提交
1542
    if (ret < 0) {
M
Max Reitz 已提交
1543
        error_propagate(errp, local_err);
K
Kevin Wolf 已提交
1544 1545 1546 1547 1548
        goto out;
    }

    ret = qcow2_alloc_clusters(bs, 2 * cluster_size);
    if (ret < 0) {
M
Max Reitz 已提交
1549 1550
        error_setg_errno(errp, -ret, "Could not allocate clusters for qcow2 "
                         "header and refcount table");
K
Kevin Wolf 已提交
1551 1552 1553 1554 1555 1556 1557 1558 1559 1560
        goto out;

    } else if (ret != 0) {
        error_report("Huh, first cluster in empty image is already in use?");
        abort();
    }

    /* Okay, now that we have a valid image, let's give it the right size */
    ret = bdrv_truncate(bs, total_size * BDRV_SECTOR_SIZE);
    if (ret < 0) {
M
Max Reitz 已提交
1561
        error_setg_errno(errp, -ret, "Could not resize image");
K
Kevin Wolf 已提交
1562 1563 1564 1565 1566 1567 1568
        goto out;
    }

    /* Want a backing file? There you go.*/
    if (backing_file) {
        ret = bdrv_change_backing_file(bs, backing_file, backing_format);
        if (ret < 0) {
M
Max Reitz 已提交
1569 1570
            error_setg_errno(errp, -ret, "Could not assign backing file '%s' "
                             "with format '%s'", backing_file, backing_format);
K
Kevin Wolf 已提交
1571 1572 1573 1574 1575 1576
            goto out;
        }
    }

    /* And if we're supposed to preallocate metadata, do that now */
    if (prealloc) {
Z
Zhi Yong Wu 已提交
1577 1578
        BDRVQcowState *s = bs->opaque;
        qemu_co_mutex_lock(&s->lock);
K
Kevin Wolf 已提交
1579
        ret = preallocate(bs);
Z
Zhi Yong Wu 已提交
1580
        qemu_co_mutex_unlock(&s->lock);
K
Kevin Wolf 已提交
1581
        if (ret < 0) {
M
Max Reitz 已提交
1582
            error_setg_errno(errp, -ret, "Could not preallocate metadata");
K
Kevin Wolf 已提交
1583 1584 1585 1586
            goto out;
        }
    }

M
Max Reitz 已提交
1587 1588 1589 1590
    bdrv_close(bs);

    /* Reopen the image without BDRV_O_NO_FLUSH to flush it before returning */
    ret = bdrv_open(bs, filename, NULL,
1591 1592
                    BDRV_O_RDWR | BDRV_O_CACHE_WB | BDRV_O_NO_BACKING,
                    drv, &local_err);
M
Max Reitz 已提交
1593 1594 1595 1596 1597
    if (error_is_set(&local_err)) {
        error_propagate(errp, local_err);
        goto out;
    }

K
Kevin Wolf 已提交
1598 1599
    ret = 0;
out:
F
Fam Zheng 已提交
1600
    bdrv_unref(bs);
K
Kevin Wolf 已提交
1601 1602
    return ret;
}
K
Kevin Wolf 已提交
1603

1604 1605
static int qcow2_create(const char *filename, QEMUOptionParameter *options,
                        Error **errp)
K
Kevin Wolf 已提交
1606 1607 1608 1609 1610
{
    const char *backing_file = NULL;
    const char *backing_fmt = NULL;
    uint64_t sectors = 0;
    int flags = 0;
1611
    size_t cluster_size = DEFAULT_CLUSTER_SIZE;
K
Kevin Wolf 已提交
1612
    int prealloc = 0;
1613
    int version = 3;
M
Max Reitz 已提交
1614 1615
    Error *local_err = NULL;
    int ret;
K
Kevin Wolf 已提交
1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636

    /* Read out options */
    while (options && options->name) {
        if (!strcmp(options->name, BLOCK_OPT_SIZE)) {
            sectors = options->value.n / 512;
        } else if (!strcmp(options->name, BLOCK_OPT_BACKING_FILE)) {
            backing_file = options->value.s;
        } else if (!strcmp(options->name, BLOCK_OPT_BACKING_FMT)) {
            backing_fmt = options->value.s;
        } else if (!strcmp(options->name, BLOCK_OPT_ENCRYPT)) {
            flags |= options->value.n ? BLOCK_FLAG_ENCRYPT : 0;
        } else if (!strcmp(options->name, BLOCK_OPT_CLUSTER_SIZE)) {
            if (options->value.n) {
                cluster_size = options->value.n;
            }
        } else if (!strcmp(options->name, BLOCK_OPT_PREALLOC)) {
            if (!options->value.s || !strcmp(options->value.s, "off")) {
                prealloc = 0;
            } else if (!strcmp(options->value.s, "metadata")) {
                prealloc = 1;
            } else {
M
Max Reitz 已提交
1637 1638
                error_setg(errp, "Invalid preallocation mode: '%s'",
                           options->value.s);
K
Kevin Wolf 已提交
1639 1640
                return -EINVAL;
            }
K
Kevin Wolf 已提交
1641
        } else if (!strcmp(options->name, BLOCK_OPT_COMPAT_LEVEL)) {
1642 1643 1644
            if (!options->value.s) {
                /* keep the default */
            } else if (!strcmp(options->value.s, "0.10")) {
K
Kevin Wolf 已提交
1645 1646 1647 1648
                version = 2;
            } else if (!strcmp(options->value.s, "1.1")) {
                version = 3;
            } else {
M
Max Reitz 已提交
1649 1650
                error_setg(errp, "Invalid compatibility level: '%s'",
                           options->value.s);
K
Kevin Wolf 已提交
1651 1652
                return -EINVAL;
            }
1653 1654
        } else if (!strcmp(options->name, BLOCK_OPT_LAZY_REFCOUNTS)) {
            flags |= options->value.n ? BLOCK_FLAG_LAZY_REFCOUNTS : 0;
K
Kevin Wolf 已提交
1655 1656 1657 1658 1659
        }
        options++;
    }

    if (backing_file && prealloc) {
M
Max Reitz 已提交
1660 1661
        error_setg(errp, "Backing file and preallocation cannot be used at "
                   "the same time");
K
Kevin Wolf 已提交
1662 1663 1664
        return -EINVAL;
    }

1665
    if (version < 3 && (flags & BLOCK_FLAG_LAZY_REFCOUNTS)) {
M
Max Reitz 已提交
1666 1667
        error_setg(errp, "Lazy refcounts only supported with compatibility "
                   "level 1.1 and above (use compat=1.1 or greater)");
1668 1669 1670
        return -EINVAL;
    }

M
Max Reitz 已提交
1671 1672 1673 1674 1675 1676
    ret = qcow2_create2(filename, sectors, backing_file, backing_fmt, flags,
                        cluster_size, prealloc, options, version, &local_err);
    if (error_is_set(&local_err)) {
        error_propagate(errp, local_err);
    }
    return ret;
K
Kevin Wolf 已提交
1677 1678
}

1679
static int qcow2_make_empty(BlockDriverState *bs)
B
Blue Swirl 已提交
1680 1681 1682 1683 1684 1685 1686 1687
{
#if 0
    /* XXX: not correct */
    BDRVQcowState *s = bs->opaque;
    uint32_t l1_length = s->l1_size * sizeof(uint64_t);
    int ret;

    memset(s->l1_table, 0, l1_length);
1688
    if (bdrv_pwrite(bs->file, s->l1_table_offset, s->l1_table, l1_length) < 0)
B
Blue Swirl 已提交
1689
        return -1;
1690
    ret = bdrv_truncate(bs->file, s->l1_table_offset + l1_length);
B
Blue Swirl 已提交
1691 1692 1693 1694 1695 1696 1697 1698
    if (ret < 0)
        return ret;

    l2_cache_reset(bs);
#endif
    return 0;
}

K
Kevin Wolf 已提交
1699
static coroutine_fn int qcow2_co_write_zeroes(BlockDriverState *bs,
1700
    int64_t sector_num, int nb_sectors, BdrvRequestFlags flags)
K
Kevin Wolf 已提交
1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718
{
    int ret;
    BDRVQcowState *s = bs->opaque;

    /* Emulate misaligned zero writes */
    if (sector_num % s->cluster_sectors || nb_sectors % s->cluster_sectors) {
        return -ENOTSUP;
    }

    /* Whatever is left can use real zero clusters */
    qemu_co_mutex_lock(&s->lock);
    ret = qcow2_zero_clusters(bs, sector_num << BDRV_SECTOR_BITS,
        nb_sectors);
    qemu_co_mutex_unlock(&s->lock);

    return ret;
}

1719 1720
static coroutine_fn int qcow2_co_discard(BlockDriverState *bs,
    int64_t sector_num, int nb_sectors)
K
Kevin Wolf 已提交
1721
{
1722 1723 1724 1725 1726
    int ret;
    BDRVQcowState *s = bs->opaque;

    qemu_co_mutex_lock(&s->lock);
    ret = qcow2_discard_clusters(bs, sector_num << BDRV_SECTOR_BITS,
1727
        nb_sectors, QCOW2_DISCARD_REQUEST);
1728 1729
    qemu_co_mutex_unlock(&s->lock);
    return ret;
K
Kevin Wolf 已提交
1730 1731
}

1732 1733 1734
static int qcow2_truncate(BlockDriverState *bs, int64_t offset)
{
    BDRVQcowState *s = bs->opaque;
1735 1736
    int64_t new_l1_size;
    int ret;
1737 1738

    if (offset & 511) {
1739
        error_report("The new size must be a multiple of 512");
1740 1741 1742 1743 1744
        return -EINVAL;
    }

    /* cannot proceed if image has snapshots */
    if (s->nb_snapshots) {
1745
        error_report("Can't resize an image which has snapshots");
1746 1747 1748 1749 1750
        return -ENOTSUP;
    }

    /* shrinking is currently not supported */
    if (offset < bs->total_sectors * 512) {
1751
        error_report("qcow2 doesn't support shrinking images yet");
1752 1753 1754 1755
        return -ENOTSUP;
    }

    new_l1_size = size_to_l1(s, offset);
1756
    ret = qcow2_grow_l1_table(bs, new_l1_size, true);
1757 1758 1759 1760 1761 1762
    if (ret < 0) {
        return ret;
    }

    /* write updated header.size */
    offset = cpu_to_be64(offset);
1763 1764
    ret = bdrv_pwrite_sync(bs->file, offsetof(QCowHeader, size),
                           &offset, sizeof(uint64_t));
1765 1766 1767 1768 1769 1770 1771 1772
    if (ret < 0) {
        return ret;
    }

    s->l1_vm_state_index = new_l1_size;
    return 0;
}

B
Blue Swirl 已提交
1773 1774
/* XXX: put compressed sectors first, then all the cluster aligned
   tables to avoid losing bytes in alignment */
1775 1776
static int qcow2_write_compressed(BlockDriverState *bs, int64_t sector_num,
                                  const uint8_t *buf, int nb_sectors)
B
Blue Swirl 已提交
1777 1778 1779 1780 1781 1782 1783 1784 1785 1786
{
    BDRVQcowState *s = bs->opaque;
    z_stream strm;
    int ret, out_len;
    uint8_t *out_buf;
    uint64_t cluster_offset;

    if (nb_sectors == 0) {
        /* align end of file to a sector boundary to ease reading with
           sector based I/Os */
1787
        cluster_offset = bdrv_getlength(bs->file);
B
Blue Swirl 已提交
1788
        cluster_offset = (cluster_offset + 511) & ~511;
1789
        bdrv_truncate(bs->file, cluster_offset);
B
Blue Swirl 已提交
1790 1791 1792
        return 0;
    }

1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807
    if (nb_sectors != s->cluster_sectors) {
        ret = -EINVAL;

        /* Zero-pad last write if image size is not cluster aligned */
        if (sector_num + nb_sectors == bs->total_sectors &&
            nb_sectors < s->cluster_sectors) {
            uint8_t *pad_buf = qemu_blockalign(bs, s->cluster_size);
            memset(pad_buf, 0, s->cluster_size);
            memcpy(pad_buf, buf, nb_sectors * BDRV_SECTOR_SIZE);
            ret = qcow2_write_compressed(bs, sector_num,
                                         pad_buf, s->cluster_sectors);
            qemu_vfree(pad_buf);
        }
        return ret;
    }
B
Blue Swirl 已提交
1808

1809
    out_buf = g_malloc(s->cluster_size + (s->cluster_size / 1000) + 128);
B
Blue Swirl 已提交
1810 1811 1812 1813 1814 1815 1816

    /* best compression, small window, no zlib header */
    memset(&strm, 0, sizeof(strm));
    ret = deflateInit2(&strm, Z_DEFAULT_COMPRESSION,
                       Z_DEFLATED, -12,
                       9, Z_DEFAULT_STRATEGY);
    if (ret != 0) {
1817 1818
        ret = -EINVAL;
        goto fail;
B
Blue Swirl 已提交
1819 1820 1821 1822 1823 1824 1825 1826 1827 1828
    }

    strm.avail_in = s->cluster_size;
    strm.next_in = (uint8_t *)buf;
    strm.avail_out = s->cluster_size;
    strm.next_out = out_buf;

    ret = deflate(&strm, Z_FINISH);
    if (ret != Z_STREAM_END && ret != Z_OK) {
        deflateEnd(&strm);
1829 1830
        ret = -EINVAL;
        goto fail;
B
Blue Swirl 已提交
1831 1832 1833 1834 1835 1836 1837
    }
    out_len = strm.next_out - out_buf;

    deflateEnd(&strm);

    if (ret != Z_STREAM_END || out_len >= s->cluster_size) {
        /* could not compress: write normal cluster */
1838 1839 1840 1841
        ret = bdrv_write(bs, sector_num, buf, s->cluster_sectors);
        if (ret < 0) {
            goto fail;
        }
B
Blue Swirl 已提交
1842 1843 1844
    } else {
        cluster_offset = qcow2_alloc_compressed_cluster_offset(bs,
            sector_num << 9, out_len);
1845 1846 1847 1848
        if (!cluster_offset) {
            ret = -EIO;
            goto fail;
        }
B
Blue Swirl 已提交
1849
        cluster_offset &= s->cluster_offset_mask;
1850

M
Max Reitz 已提交
1851
        ret = qcow2_pre_write_overlap_check(bs, 0, cluster_offset, out_len);
1852 1853 1854 1855
        if (ret < 0) {
            goto fail;
        }

1856
        BLKDBG_EVENT(bs->file, BLKDBG_WRITE_COMPRESSED);
1857 1858 1859
        ret = bdrv_pwrite(bs->file, cluster_offset, out_buf, out_len);
        if (ret < 0) {
            goto fail;
B
Blue Swirl 已提交
1860 1861 1862
        }
    }

1863 1864
    ret = 0;
fail:
1865
    g_free(out_buf);
1866
    return ret;
B
Blue Swirl 已提交
1867 1868
}

1869
static coroutine_fn int qcow2_co_flush_to_os(BlockDriverState *bs)
B
Blue Swirl 已提交
1870
{
K
Kevin Wolf 已提交
1871 1872 1873
    BDRVQcowState *s = bs->opaque;
    int ret;

P
Paolo Bonzini 已提交
1874
    qemu_co_mutex_lock(&s->lock);
K
Kevin Wolf 已提交
1875 1876
    ret = qcow2_cache_flush(bs, s->l2_table_cache);
    if (ret < 0) {
D
Dong Xu Wang 已提交
1877
        qemu_co_mutex_unlock(&s->lock);
P
Paolo Bonzini 已提交
1878
        return ret;
K
Kevin Wolf 已提交
1879 1880
    }

1881 1882 1883 1884 1885 1886
    if (qcow2_need_accurate_refcounts(s)) {
        ret = qcow2_cache_flush(bs, s->refcount_block_cache);
        if (ret < 0) {
            qemu_co_mutex_unlock(&s->lock);
            return ret;
        }
K
Kevin Wolf 已提交
1887
    }
P
Paolo Bonzini 已提交
1888
    qemu_co_mutex_unlock(&s->lock);
K
Kevin Wolf 已提交
1889

K
Kevin Wolf 已提交
1890 1891 1892
    return 0;
}

1893
static int qcow2_get_info(BlockDriverState *bs, BlockDriverInfo *bdi)
B
Blue Swirl 已提交
1894 1895 1896
{
    BDRVQcowState *s = bs->opaque;
    bdi->cluster_size = s->cluster_size;
1897
    bdi->vm_state_offset = qcow2_vm_state_offset(s);
B
Blue Swirl 已提交
1898 1899 1900
    return 0;
}

1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927
static ImageInfoSpecific *qcow2_get_specific_info(BlockDriverState *bs)
{
    BDRVQcowState *s = bs->opaque;
    ImageInfoSpecific *spec_info = g_new(ImageInfoSpecific, 1);

    *spec_info = (ImageInfoSpecific){
        .kind  = IMAGE_INFO_SPECIFIC_KIND_QCOW2,
        {
            .qcow2 = g_new(ImageInfoSpecificQCow2, 1),
        },
    };
    if (s->qcow_version == 2) {
        *spec_info->qcow2 = (ImageInfoSpecificQCow2){
            .compat = g_strdup("0.10"),
        };
    } else if (s->qcow_version == 3) {
        *spec_info->qcow2 = (ImageInfoSpecificQCow2){
            .compat             = g_strdup("1.1"),
            .lazy_refcounts     = s->compatible_features &
                                  QCOW2_COMPAT_LAZY_REFCOUNTS,
            .has_lazy_refcounts = true,
        };
    }

    return spec_info;
}

B
Blue Swirl 已提交
1928 1929 1930 1931 1932 1933 1934
#if 0
static void dump_refcounts(BlockDriverState *bs)
{
    BDRVQcowState *s = bs->opaque;
    int64_t nb_clusters, k, k1, size;
    int refcount;

1935
    size = bdrv_getlength(bs->file);
B
Blue Swirl 已提交
1936 1937 1938 1939 1940 1941 1942
    nb_clusters = size_to_clusters(s, size);
    for(k = 0; k < nb_clusters;) {
        k1 = k;
        refcount = get_refcount(bs, k);
        k++;
        while (k < nb_clusters && get_refcount(bs, k) == refcount)
            k++;
B
Blue Swirl 已提交
1943 1944
        printf("%" PRId64 ": refcount=%d nb=%" PRId64 "\n", k, refcount,
               k - k1);
B
Blue Swirl 已提交
1945 1946 1947 1948
    }
}
#endif

1949 1950
static int qcow2_save_vmstate(BlockDriverState *bs, QEMUIOVector *qiov,
                              int64_t pos)
B
Blue Swirl 已提交
1951 1952
{
    BDRVQcowState *s = bs->opaque;
1953
    int64_t total_sectors = bs->total_sectors;
B
Blue Swirl 已提交
1954
    int growable = bs->growable;
1955
    bool zero_beyond_eof = bs->zero_beyond_eof;
B
Blue Swirl 已提交
1956 1957
    int ret;

1958
    BLKDBG_EVENT(bs->file, BLKDBG_VMSTATE_SAVE);
B
Blue Swirl 已提交
1959
    bs->growable = 1;
1960
    bs->zero_beyond_eof = false;
1961
    ret = bdrv_pwritev(bs, qcow2_vm_state_offset(s) + pos, qiov);
B
Blue Swirl 已提交
1962
    bs->growable = growable;
1963
    bs->zero_beyond_eof = zero_beyond_eof;
B
Blue Swirl 已提交
1964

1965 1966 1967 1968 1969
    /* bdrv_co_do_writev will have increased the total_sectors value to include
     * the VM state - the VM state is however not an actual part of the block
     * device, therefore, we need to restore the old value. */
    bs->total_sectors = total_sectors;

B
Blue Swirl 已提交
1970 1971 1972
    return ret;
}

1973 1974
static int qcow2_load_vmstate(BlockDriverState *bs, uint8_t *buf,
                              int64_t pos, int size)
B
Blue Swirl 已提交
1975 1976 1977
{
    BDRVQcowState *s = bs->opaque;
    int growable = bs->growable;
A
Asias He 已提交
1978
    bool zero_beyond_eof = bs->zero_beyond_eof;
B
Blue Swirl 已提交
1979 1980
    int ret;

1981
    BLKDBG_EVENT(bs->file, BLKDBG_VMSTATE_LOAD);
B
Blue Swirl 已提交
1982
    bs->growable = 1;
A
Asias He 已提交
1983
    bs->zero_beyond_eof = false;
1984
    ret = bdrv_pread(bs, qcow2_vm_state_offset(s) + pos, buf, size);
B
Blue Swirl 已提交
1985
    bs->growable = growable;
A
Asias He 已提交
1986
    bs->zero_beyond_eof = zero_beyond_eof;
B
Blue Swirl 已提交
1987 1988 1989 1990

    return ret;
}

M
Max Reitz 已提交
1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
/*
 * Downgrades an image's version. To achieve this, any incompatible features
 * have to be removed.
 */
static int qcow2_downgrade(BlockDriverState *bs, int target_version)
{
    BDRVQcowState *s = bs->opaque;
    int current_version = s->qcow_version;
    int ret;

    if (target_version == current_version) {
        return 0;
    } else if (target_version > current_version) {
        return -EINVAL;
    } else if (target_version != 2) {
        return -EINVAL;
    }

    if (s->refcount_order != 4) {
        /* we would have to convert the image to a refcount_order == 4 image
         * here; however, since qemu (at the time of writing this) does not
         * support anything different than 4 anyway, there is no point in doing
         * so right now; however, we should error out (if qemu supports this in
         * the future and this code has not been adapted) */
2015
        error_report("qcow2_downgrade: Image refcount orders other than 4 are "
M
Max Reitz 已提交
2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183
                     "currently not supported.");
        return -ENOTSUP;
    }

    /* clear incompatible features */
    if (s->incompatible_features & QCOW2_INCOMPAT_DIRTY) {
        ret = qcow2_mark_clean(bs);
        if (ret < 0) {
            return ret;
        }
    }

    /* with QCOW2_INCOMPAT_CORRUPT, it is pretty much impossible to get here in
     * the first place; if that happens nonetheless, returning -ENOTSUP is the
     * best thing to do anyway */

    if (s->incompatible_features) {
        return -ENOTSUP;
    }

    /* since we can ignore compatible features, we can set them to 0 as well */
    s->compatible_features = 0;
    /* if lazy refcounts have been used, they have already been fixed through
     * clearing the dirty flag */

    /* clearing autoclear features is trivial */
    s->autoclear_features = 0;

    ret = qcow2_expand_zero_clusters(bs);
    if (ret < 0) {
        return ret;
    }

    s->qcow_version = target_version;
    ret = qcow2_update_header(bs);
    if (ret < 0) {
        s->qcow_version = current_version;
        return ret;
    }
    return 0;
}

static int qcow2_amend_options(BlockDriverState *bs,
                               QEMUOptionParameter *options)
{
    BDRVQcowState *s = bs->opaque;
    int old_version = s->qcow_version, new_version = old_version;
    uint64_t new_size = 0;
    const char *backing_file = NULL, *backing_format = NULL;
    bool lazy_refcounts = s->use_lazy_refcounts;
    int ret;
    int i;

    for (i = 0; options[i].name; i++)
    {
        if (!options[i].assigned) {
            /* only change explicitly defined options */
            continue;
        }

        if (!strcmp(options[i].name, "compat")) {
            if (!options[i].value.s) {
                /* preserve default */
            } else if (!strcmp(options[i].value.s, "0.10")) {
                new_version = 2;
            } else if (!strcmp(options[i].value.s, "1.1")) {
                new_version = 3;
            } else {
                fprintf(stderr, "Unknown compatibility level %s.\n",
                        options[i].value.s);
                return -EINVAL;
            }
        } else if (!strcmp(options[i].name, "preallocation")) {
            fprintf(stderr, "Cannot change preallocation mode.\n");
            return -ENOTSUP;
        } else if (!strcmp(options[i].name, "size")) {
            new_size = options[i].value.n;
        } else if (!strcmp(options[i].name, "backing_file")) {
            backing_file = options[i].value.s;
        } else if (!strcmp(options[i].name, "backing_fmt")) {
            backing_format = options[i].value.s;
        } else if (!strcmp(options[i].name, "encryption")) {
            if ((options[i].value.n != !!s->crypt_method)) {
                fprintf(stderr, "Changing the encryption flag is not "
                        "supported.\n");
                return -ENOTSUP;
            }
        } else if (!strcmp(options[i].name, "cluster_size")) {
            if (options[i].value.n != s->cluster_size) {
                fprintf(stderr, "Changing the cluster size is not "
                        "supported.\n");
                return -ENOTSUP;
            }
        } else if (!strcmp(options[i].name, "lazy_refcounts")) {
            lazy_refcounts = options[i].value.n;
        } else {
            /* if this assertion fails, this probably means a new option was
             * added without having it covered here */
            assert(false);
        }
    }

    if (new_version != old_version) {
        if (new_version > old_version) {
            /* Upgrade */
            s->qcow_version = new_version;
            ret = qcow2_update_header(bs);
            if (ret < 0) {
                s->qcow_version = old_version;
                return ret;
            }
        } else {
            ret = qcow2_downgrade(bs, new_version);
            if (ret < 0) {
                return ret;
            }
        }
    }

    if (backing_file || backing_format) {
        ret = qcow2_change_backing_file(bs, backing_file ?: bs->backing_file,
                                        backing_format ?: bs->backing_format);
        if (ret < 0) {
            return ret;
        }
    }

    if (s->use_lazy_refcounts != lazy_refcounts) {
        if (lazy_refcounts) {
            if (s->qcow_version < 3) {
                fprintf(stderr, "Lazy refcounts only supported with compatibility "
                        "level 1.1 and above (use compat=1.1 or greater)\n");
                return -EINVAL;
            }
            s->compatible_features |= QCOW2_COMPAT_LAZY_REFCOUNTS;
            ret = qcow2_update_header(bs);
            if (ret < 0) {
                s->compatible_features &= ~QCOW2_COMPAT_LAZY_REFCOUNTS;
                return ret;
            }
            s->use_lazy_refcounts = true;
        } else {
            /* make image clean first */
            ret = qcow2_mark_clean(bs);
            if (ret < 0) {
                return ret;
            }
            /* now disallow lazy refcounts */
            s->compatible_features &= ~QCOW2_COMPAT_LAZY_REFCOUNTS;
            ret = qcow2_update_header(bs);
            if (ret < 0) {
                s->compatible_features |= QCOW2_COMPAT_LAZY_REFCOUNTS;
                return ret;
            }
            s->use_lazy_refcounts = false;
        }
    }

    if (new_size) {
        ret = bdrv_truncate(bs, new_size);
        if (ret < 0) {
            return ret;
        }
    }

    return 0;
}

2184
static QEMUOptionParameter qcow2_create_options[] = {
B
Blue Swirl 已提交
2185 2186 2187 2188 2189
    {
        .name = BLOCK_OPT_SIZE,
        .type = OPT_SIZE,
        .help = "Virtual disk size"
    },
K
Kevin Wolf 已提交
2190 2191 2192 2193 2194
    {
        .name = BLOCK_OPT_COMPAT_LEVEL,
        .type = OPT_STRING,
        .help = "Compatibility level (0.10 or 1.1)"
    },
B
Blue Swirl 已提交
2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212
    {
        .name = BLOCK_OPT_BACKING_FILE,
        .type = OPT_STRING,
        .help = "File name of a base image"
    },
    {
        .name = BLOCK_OPT_BACKING_FMT,
        .type = OPT_STRING,
        .help = "Image format of the base image"
    },
    {
        .name = BLOCK_OPT_ENCRYPT,
        .type = OPT_FLAG,
        .help = "Encrypt the image"
    },
    {
        .name = BLOCK_OPT_CLUSTER_SIZE,
        .type = OPT_SIZE,
2213 2214
        .help = "qcow2 cluster size",
        .value = { .n = DEFAULT_CLUSTER_SIZE },
B
Blue Swirl 已提交
2215 2216 2217 2218 2219 2220
    },
    {
        .name = BLOCK_OPT_PREALLOC,
        .type = OPT_STRING,
        .help = "Preallocation mode (allowed values: off, metadata)"
    },
2221 2222 2223 2224 2225
    {
        .name = BLOCK_OPT_LAZY_REFCOUNTS,
        .type = OPT_FLAG,
        .help = "Postpone refcount updates",
    },
B
Blue Swirl 已提交
2226 2227 2228 2229
    { NULL }
};

static BlockDriver bdrv_qcow2 = {
2230 2231 2232 2233 2234
    .format_name        = "qcow2",
    .instance_size      = sizeof(BDRVQcowState),
    .bdrv_probe         = qcow2_probe,
    .bdrv_open          = qcow2_open,
    .bdrv_close         = qcow2_close,
J
Jeff Cody 已提交
2235
    .bdrv_reopen_prepare  = qcow2_reopen_prepare,
2236
    .bdrv_create        = qcow2_create,
2237
    .bdrv_has_zero_init = bdrv_has_zero_init_1,
2238
    .bdrv_co_get_block_status = qcow2_co_get_block_status,
2239 2240 2241
    .bdrv_set_key       = qcow2_set_key,
    .bdrv_make_empty    = qcow2_make_empty,

2242 2243
    .bdrv_co_readv          = qcow2_co_readv,
    .bdrv_co_writev         = qcow2_co_writev,
K
Kevin Wolf 已提交
2244
    .bdrv_co_flush_to_os    = qcow2_co_flush_to_os,
2245

K
Kevin Wolf 已提交
2246
    .bdrv_co_write_zeroes   = qcow2_co_write_zeroes,
2247
    .bdrv_co_discard        = qcow2_co_discard,
2248
    .bdrv_truncate          = qcow2_truncate,
2249
    .bdrv_write_compressed  = qcow2_write_compressed,
B
Blue Swirl 已提交
2250 2251 2252 2253 2254

    .bdrv_snapshot_create   = qcow2_snapshot_create,
    .bdrv_snapshot_goto     = qcow2_snapshot_goto,
    .bdrv_snapshot_delete   = qcow2_snapshot_delete,
    .bdrv_snapshot_list     = qcow2_snapshot_list,
E
edison 已提交
2255
    .bdrv_snapshot_load_tmp     = qcow2_snapshot_load_tmp,
2256
    .bdrv_get_info      = qcow2_get_info,
2257
    .bdrv_get_specific_info = qcow2_get_specific_info,
B
Blue Swirl 已提交
2258

2259 2260
    .bdrv_save_vmstate    = qcow2_save_vmstate,
    .bdrv_load_vmstate    = qcow2_load_vmstate,
B
Blue Swirl 已提交
2261 2262 2263

    .bdrv_change_backing_file   = qcow2_change_backing_file,

2264 2265
    .bdrv_invalidate_cache      = qcow2_invalidate_cache,

2266 2267
    .create_options = qcow2_create_options,
    .bdrv_check = qcow2_check,
M
Max Reitz 已提交
2268
    .bdrv_amend_options = qcow2_amend_options,
B
Blue Swirl 已提交
2269 2270
};

2271 2272 2273 2274 2275 2276
static void bdrv_qcow2_init(void)
{
    bdrv_register(&bdrv_qcow2);
}

block_init(bdrv_qcow2_init);