block-qcow2.c 71.4 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"
B
bellard 已提交
25 26 27 28 29 30 31 32 33 34 35 36
#include "block_int.h"
#include <zlib.h>
#include "aes.h"
#include <assert.h>

/*
  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.
37
  - Size of compressed clusters is stored in sectors to reduce bit usage
B
bellard 已提交
38 39
    in the cluster offsets.
  - Support for storing additional data (such as the VM state) in the
40
    snapshots.
B
bellard 已提交
41 42 43 44 45 46 47
  - 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.
*/

//#define DEBUG_ALLOC
//#define DEBUG_ALLOC2
48

B
bellard 已提交
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
#define QCOW_MAGIC (('Q' << 24) | ('F' << 16) | ('I' << 8) | 0xfb)
#define QCOW_VERSION 2

#define QCOW_CRYPT_NONE 0
#define QCOW_CRYPT_AES  1

/* indicate that the refcount of the referenced cluster is exactly one. */
#define QCOW_OFLAG_COPIED     (1LL << 63)
/* indicate that the cluster is compressed (they never have the copied flag) */
#define QCOW_OFLAG_COMPRESSED (1LL << 62)

#define REFCOUNT_SHIFT 1 /* refcount size is 2 bytes */

#ifndef offsetof
#define offsetof(type, field) ((size_t) &((type *)0)->field)
#endif

typedef struct QCowHeader {
    uint32_t magic;
    uint32_t version;
    uint64_t backing_file_offset;
    uint32_t backing_file_size;
    uint32_t cluster_bits;
    uint64_t size; /* in bytes */
    uint32_t crypt_method;
    uint32_t l1_size; /* XXX: save number of clusters instead ? */
    uint64_t l1_table_offset;
    uint64_t refcount_table_offset;
    uint32_t refcount_table_clusters;
    uint32_t nb_snapshots;
    uint64_t snapshots_offset;
} QCowHeader;

typedef struct __attribute__((packed)) QCowSnapshotHeader {
    /* header is 8 byte aligned */
    uint64_t l1_table_offset;

    uint32_t l1_size;
    uint16_t id_str_size;
    uint16_t name_size;

    uint32_t date_sec;
    uint32_t date_nsec;

    uint64_t vm_clock_nsec;

    uint32_t vm_state_size;
    uint32_t extra_data_size; /* for extension */
    /* extra data follows */
    /* id_str follows */
    /* name follows  */
} QCowSnapshotHeader;

#define L2_CACHE_SIZE 16

typedef struct QCowSnapshot {
    uint64_t l1_table_offset;
    uint32_t l1_size;
    char *id_str;
    char *name;
    uint32_t vm_state_size;
    uint32_t date_sec;
    uint32_t date_nsec;
    uint64_t vm_clock_nsec;
} QCowSnapshot;

typedef struct BDRVQcowState {
    BlockDriverState *hd;
    int cluster_bits;
    int cluster_size;
    int cluster_sectors;
    int l2_bits;
    int l2_size;
    int l1_size;
    int l1_vm_state_index;
    int csize_shift;
    int csize_mask;
    uint64_t cluster_offset_mask;
    uint64_t l1_table_offset;
    uint64_t *l1_table;
    uint64_t *l2_cache;
    uint64_t l2_cache_offsets[L2_CACHE_SIZE];
    uint32_t l2_cache_counts[L2_CACHE_SIZE];
    uint8_t *cluster_cache;
    uint8_t *cluster_data;
    uint64_t cluster_cache_offset;

    uint64_t *refcount_table;
    uint64_t refcount_table_offset;
    uint32_t refcount_table_size;
    uint64_t refcount_block_cache_offset;
    uint16_t *refcount_block_cache;
    int64_t free_cluster_index;
    int64_t free_byte_offset;

    uint32_t crypt_method; /* current crypt method, 0 if no key yet */
    uint32_t crypt_method_header;
    AES_KEY aes_encrypt_key;
    AES_KEY aes_decrypt_key;
    uint64_t snapshots_offset;
    int snapshots_size;
    int nb_snapshots;
    QCowSnapshot *snapshots;
} BDRVQcowState;

static int decompress_cluster(BDRVQcowState *s, uint64_t cluster_offset);
155
static int qcow_read(BlockDriverState *bs, int64_t sector_num,
B
bellard 已提交
156 157 158 159 160 161
                     uint8_t *buf, int nb_sectors);
static int qcow_read_snapshots(BlockDriverState *bs);
static void qcow_free_snapshots(BlockDriverState *bs);
static int refcount_init(BlockDriverState *bs);
static void refcount_close(BlockDriverState *bs);
static int get_refcount(BlockDriverState *bs, int64_t cluster_index);
162
static int update_cluster_refcount(BlockDriverState *bs,
B
bellard 已提交
163 164
                                   int64_t cluster_index,
                                   int addend);
165 166
static void update_refcount(BlockDriverState *bs,
                            int64_t offset, int64_t length,
B
bellard 已提交
167 168 169
                            int addend);
static int64_t alloc_clusters(BlockDriverState *bs, int64_t size);
static int64_t alloc_bytes(BlockDriverState *bs, int size);
170
static void free_clusters(BlockDriverState *bs,
B
bellard 已提交
171 172 173 174 175 176 177 178
                          int64_t offset, int64_t size);
#ifdef DEBUG_ALLOC
static void check_refcounts(BlockDriverState *bs);
#endif

static int qcow_probe(const uint8_t *buf, int buf_size, const char *filename)
{
    const QCowHeader *cow_header = (const void *)buf;
179

B
bellard 已提交
180 181
    if (buf_size >= sizeof(QCowHeader) &&
        be32_to_cpu(cow_header->magic) == QCOW_MAGIC &&
182
        be32_to_cpu(cow_header->version) == QCOW_VERSION)
B
bellard 已提交
183 184 185 186 187 188 189 190 191 192 193
        return 100;
    else
        return 0;
}

static int qcow_open(BlockDriverState *bs, const char *filename, int flags)
{
    BDRVQcowState *s = bs->opaque;
    int len, i, shift, ret;
    QCowHeader header;

194
    ret = bdrv_file_open(&s->hd, filename, flags);
B
bellard 已提交
195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211
    if (ret < 0)
        return ret;
    if (bdrv_pread(s->hd, 0, &header, sizeof(header)) != sizeof(header))
        goto fail;
    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);
212

B
bellard 已提交
213 214
    if (header.magic != QCOW_MAGIC || header.version != QCOW_VERSION)
        goto fail;
215 216
    if (header.size <= 1 ||
        header.cluster_bits < 9 ||
B
bellard 已提交
217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233
        header.cluster_bits > 16)
        goto fail;
    if (header.crypt_method > QCOW_CRYPT_AES)
        goto fail;
    s->crypt_method_header = header.crypt_method;
    if (s->crypt_method_header)
        bs->encrypted = 1;
    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;
234
    s->refcount_table_size =
B
bellard 已提交
235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251
        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;
    shift = s->cluster_bits + s->l2_bits;
    s->l1_vm_state_index = (header.size + (1LL << shift) - 1) >> shift;
    /* the L1 table must contain at least enough entries to put
       header.size bytes */
    if (s->l1_size < s->l1_vm_state_index)
        goto fail;
    s->l1_table_offset = header.l1_table_offset;
    s->l1_table = qemu_malloc(s->l1_size * sizeof(uint64_t));
    if (!s->l1_table)
        goto fail;
252
    if (bdrv_pread(s->hd, s->l1_table_offset, s->l1_table, s->l1_size * sizeof(uint64_t)) !=
B
bellard 已提交
253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269
        s->l1_size * sizeof(uint64_t))
        goto fail;
    for(i = 0;i < s->l1_size; i++) {
        be64_to_cpus(&s->l1_table[i]);
    }
    /* alloc L2 cache */
    s->l2_cache = qemu_malloc(s->l2_size * L2_CACHE_SIZE * sizeof(uint64_t));
    if (!s->l2_cache)
        goto fail;
    s->cluster_cache = qemu_malloc(s->cluster_size);
    if (!s->cluster_cache)
        goto fail;
    /* one more sector for decompressed data alignment */
    s->cluster_data = qemu_malloc(s->cluster_size + 512);
    if (!s->cluster_data)
        goto fail;
    s->cluster_cache_offset = -1;
270

B
bellard 已提交
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
    if (refcount_init(bs) < 0)
        goto fail;

    /* read the backing file name */
    if (header.backing_file_offset != 0) {
        len = header.backing_file_size;
        if (len > 1023)
            len = 1023;
        if (bdrv_pread(s->hd, header.backing_file_offset, bs->backing_file, len) != len)
            goto fail;
        bs->backing_file[len] = '\0';
    }
    if (qcow_read_snapshots(bs) < 0)
        goto fail;

#ifdef DEBUG_ALLOC
    check_refcounts(bs);
#endif
    return 0;

 fail:
    qcow_free_snapshots(bs);
    refcount_close(bs);
    qemu_free(s->l1_table);
    qemu_free(s->l2_cache);
    qemu_free(s->cluster_cache);
    qemu_free(s->cluster_data);
    bdrv_delete(s->hd);
    return -1;
}

static int qcow_set_key(BlockDriverState *bs, const char *key)
{
    BDRVQcowState *s = bs->opaque;
    uint8_t keybuf[16];
    int len, i;
307

B
bellard 已提交
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
    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;
}

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

    for(i = 0; i < nb_sectors; i++) {
        ivec.ll[0] = cpu_to_le64(sector_num);
        ivec.ll[1] = 0;
361
        AES_cbc_encrypt(in_buf, out_buf, 512, key,
B
bellard 已提交
362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381
                        ivec.b, enc);
        sector_num++;
        in_buf += 512;
        out_buf += 512;
    }
}

static int copy_sectors(BlockDriverState *bs, uint64_t start_sect,
                        uint64_t cluster_offset, int n_start, int n_end)
{
    BDRVQcowState *s = bs->opaque;
    int n, ret;

    n = n_end - n_start;
    if (n <= 0)
        return 0;
    ret = qcow_read(bs, start_sect + n_start, s->cluster_data, n);
    if (ret < 0)
        return ret;
    if (s->crypt_method) {
382 383
        encrypt_sectors(s, start_sect + n_start,
                        s->cluster_data,
B
bellard 已提交
384 385 386
                        s->cluster_data, n, 1,
                        &s->aes_encrypt_key);
    }
387
    ret = bdrv_write(s->hd, (cluster_offset >> 9) + n_start,
B
bellard 已提交
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
                     s->cluster_data, n);
    if (ret < 0)
        return ret;
    return 0;
}

static void l2_cache_reset(BlockDriverState *bs)
{
    BDRVQcowState *s = bs->opaque;

    memset(s->l2_cache, 0, s->l2_size * L2_CACHE_SIZE * sizeof(uint64_t));
    memset(s->l2_cache_offsets, 0, L2_CACHE_SIZE * sizeof(uint64_t));
    memset(s->l2_cache_counts, 0, L2_CACHE_SIZE * sizeof(uint32_t));
}

static inline int l2_cache_new_entry(BlockDriverState *bs)
{
    BDRVQcowState *s = bs->opaque;
    uint32_t min_count;
    int min_index, i;

    /* find a new entry in the least used one */
    min_index = 0;
    min_count = 0xffffffff;
    for(i = 0; i < L2_CACHE_SIZE; i++) {
        if (s->l2_cache_counts[i] < min_count) {
            min_count = s->l2_cache_counts[i];
            min_index = i;
        }
    }
    return min_index;
}

static int64_t align_offset(int64_t offset, int n)
{
    offset = (offset + n - 1) & ~(n - 1);
    return offset;
}

static int grow_l1_table(BlockDriverState *bs, int min_size)
{
    BDRVQcowState *s = bs->opaque;
    int new_l1_size, new_l1_size2, ret, i;
    uint64_t *new_l1_table;
    uint64_t new_l1_table_offset;
    uint64_t data64;
    uint32_t data32;

    new_l1_size = s->l1_size;
    if (min_size <= new_l1_size)
        return 0;
    while (min_size > new_l1_size) {
        new_l1_size = (new_l1_size * 3 + 1) / 2;
    }
#ifdef DEBUG_ALLOC2
    printf("grow l1_table from %d to %d\n", s->l1_size, new_l1_size);
#endif

    new_l1_size2 = sizeof(uint64_t) * new_l1_size;
    new_l1_table = qemu_mallocz(new_l1_size2);
    if (!new_l1_table)
        return -ENOMEM;
    memcpy(new_l1_table, s->l1_table, s->l1_size * sizeof(uint64_t));

    /* write new table (align to cluster) */
    new_l1_table_offset = alloc_clusters(bs, new_l1_size2);
454

B
bellard 已提交
455 456 457 458 459 460 461
    for(i = 0; i < s->l1_size; i++)
        new_l1_table[i] = cpu_to_be64(new_l1_table[i]);
    ret = bdrv_pwrite(s->hd, new_l1_table_offset, new_l1_table, new_l1_size2);
    if (ret != new_l1_size2)
        goto fail;
    for(i = 0; i < s->l1_size; i++)
        new_l1_table[i] = be64_to_cpu(new_l1_table[i]);
462

B
bellard 已提交
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
    /* set new table */
    data64 = cpu_to_be64(new_l1_table_offset);
    if (bdrv_pwrite(s->hd, offsetof(QCowHeader, l1_table_offset),
                    &data64, sizeof(data64)) != sizeof(data64))
        goto fail;
    data32 = cpu_to_be32(new_l1_size);
    if (bdrv_pwrite(s->hd, offsetof(QCowHeader, l1_size),
                    &data32, sizeof(data32)) != sizeof(data32))
        goto fail;
    qemu_free(s->l1_table);
    free_clusters(bs, s->l1_table_offset, s->l1_size * sizeof(uint64_t));
    s->l1_table_offset = new_l1_table_offset;
    s->l1_table = new_l1_table;
    s->l1_size = new_l1_size;
    return 0;
 fail:
    qemu_free(s->l1_table);
    return -EIO;
}

/* 'allocate' is:
 *
 * 0 not to allocate.
 *
 * 1 to allocate a normal cluster (for sector indexes 'n_start' to
 * 'n_end')
 *
 * 2 to allocate a compressed cluster of size
 * 'compressed_size'. 'compressed_size' must be > 0 and <
492
 * cluster_size
B
bellard 已提交
493 494 495 496 497 498 499 500 501 502 503
 *
 * return 0 if not allocated.
 */
static uint64_t get_cluster_offset(BlockDriverState *bs,
                                   uint64_t offset, int allocate,
                                   int compressed_size,
                                   int n_start, int n_end)
{
    BDRVQcowState *s = bs->opaque;
    int min_index, i, j, l1_index, l2_index, ret;
    uint64_t l2_offset, *l2_table, cluster_offset, tmp, old_l2_offset;
504

B
bellard 已提交
505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523
    l1_index = offset >> (s->l2_bits + s->cluster_bits);
    if (l1_index >= s->l1_size) {
        /* outside l1 table is allowed: we grow the table if needed */
        if (!allocate)
            return 0;
        if (grow_l1_table(bs, l1_index + 1) < 0)
            return 0;
    }
    l2_offset = s->l1_table[l1_index];
    if (!l2_offset) {
        if (!allocate)
            return 0;
    l2_allocate:
        old_l2_offset = l2_offset;
        /* allocate a new l2 entry */
        l2_offset = alloc_clusters(bs, s->l2_size * sizeof(uint64_t));
        /* update the L1 entry */
        s->l1_table[l1_index] = l2_offset | QCOW_OFLAG_COPIED;
        tmp = cpu_to_be64(l2_offset | QCOW_OFLAG_COPIED);
524
        if (bdrv_pwrite(s->hd, s->l1_table_offset + l1_index * sizeof(tmp),
B
bellard 已提交
525 526 527 528 529 530 531 532
                        &tmp, sizeof(tmp)) != sizeof(tmp))
            return 0;
        min_index = l2_cache_new_entry(bs);
        l2_table = s->l2_cache + (min_index << s->l2_bits);

        if (old_l2_offset == 0) {
            memset(l2_table, 0, s->l2_size * sizeof(uint64_t));
        } else {
533
            if (bdrv_pread(s->hd, old_l2_offset,
B
bellard 已提交
534 535 536 537
                           l2_table, s->l2_size * sizeof(uint64_t)) !=
                s->l2_size * sizeof(uint64_t))
                return 0;
        }
538
        if (bdrv_pwrite(s->hd, l2_offset,
B
bellard 已提交
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
                        l2_table, s->l2_size * sizeof(uint64_t)) !=
            s->l2_size * sizeof(uint64_t))
            return 0;
    } else {
        if (!(l2_offset & QCOW_OFLAG_COPIED)) {
            if (allocate) {
                free_clusters(bs, l2_offset, s->l2_size * sizeof(uint64_t));
                goto l2_allocate;
            }
        } else {
            l2_offset &= ~QCOW_OFLAG_COPIED;
        }
        for(i = 0; i < L2_CACHE_SIZE; i++) {
            if (l2_offset == s->l2_cache_offsets[i]) {
                /* increment the hit count */
                if (++s->l2_cache_counts[i] == 0xffffffff) {
                    for(j = 0; j < L2_CACHE_SIZE; j++) {
                        s->l2_cache_counts[j] >>= 1;
                    }
                }
                l2_table = s->l2_cache + (i << s->l2_bits);
                goto found;
            }
        }
        /* not found: load a new entry in the least used one */
        min_index = l2_cache_new_entry(bs);
        l2_table = s->l2_cache + (min_index << s->l2_bits);
566
        if (bdrv_pread(s->hd, l2_offset, l2_table, s->l2_size * sizeof(uint64_t)) !=
B
bellard 已提交
567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583
            s->l2_size * sizeof(uint64_t))
            return 0;
    }
    s->l2_cache_offsets[min_index] = l2_offset;
    s->l2_cache_counts[min_index] = 1;
 found:
    l2_index = (offset >> s->cluster_bits) & (s->l2_size - 1);
    cluster_offset = be64_to_cpu(l2_table[l2_index]);
    if (!cluster_offset) {
        if (!allocate)
            return cluster_offset;
    } else if (!(cluster_offset & QCOW_OFLAG_COPIED)) {
        if (!allocate)
            return cluster_offset;
        /* free the cluster */
        if (cluster_offset & QCOW_OFLAG_COMPRESSED) {
            int nb_csectors;
584
            nb_csectors = ((cluster_offset >> s->csize_shift) &
B
bellard 已提交
585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602
                           s->csize_mask) + 1;
            free_clusters(bs, (cluster_offset & s->cluster_offset_mask) & ~511,
                          nb_csectors * 512);
        } else {
            free_clusters(bs, cluster_offset, s->cluster_size);
        }
    } else {
        cluster_offset &= ~QCOW_OFLAG_COPIED;
        return cluster_offset;
    }
    if (allocate == 1) {
        /* allocate a new cluster */
        cluster_offset = alloc_clusters(bs, s->cluster_size);

        /* we must initialize the cluster content which won't be
           written */
        if ((n_end - n_start) < s->cluster_sectors) {
            uint64_t start_sect;
603

B
bellard 已提交
604 605 606 607 608 609 610 611 612 613 614 615 616 617
            start_sect = (offset & ~(s->cluster_size - 1)) >> 9;
            ret = copy_sectors(bs, start_sect,
                               cluster_offset, 0, n_start);
            if (ret < 0)
                return 0;
            ret = copy_sectors(bs, start_sect,
                               cluster_offset, n_end, s->cluster_sectors);
            if (ret < 0)
                return 0;
        }
        tmp = cpu_to_be64(cluster_offset | QCOW_OFLAG_COPIED);
    } else {
        int nb_csectors;
        cluster_offset = alloc_bytes(bs, compressed_size);
618
        nb_csectors = ((cluster_offset + compressed_size - 1) >> 9) -
B
bellard 已提交
619
            (cluster_offset >> 9);
620
        cluster_offset |= QCOW_OFLAG_COMPRESSED |
B
bellard 已提交
621 622 623 624 625 626
            ((uint64_t)nb_csectors << s->csize_shift);
        /* compressed clusters never have the copied flag */
        tmp = cpu_to_be64(cluster_offset);
    }
    /* update L2 table */
    l2_table[l2_index] = tmp;
627
    if (bdrv_pwrite(s->hd,
B
bellard 已提交
628 629 630 631 632
                    l2_offset + l2_index * sizeof(tmp), &tmp, sizeof(tmp)) != sizeof(tmp))
        return 0;
    return cluster_offset;
}

633
static int qcow_is_allocated(BlockDriverState *bs, int64_t sector_num,
B
bellard 已提交
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
                             int nb_sectors, int *pnum)
{
    BDRVQcowState *s = bs->opaque;
    int index_in_cluster, n;
    uint64_t cluster_offset;

    cluster_offset = get_cluster_offset(bs, sector_num << 9, 0, 0, 0, 0);
    index_in_cluster = sector_num & (s->cluster_sectors - 1);
    n = s->cluster_sectors - index_in_cluster;
    if (n > nb_sectors)
        n = nb_sectors;
    *pnum = n;
    return (cluster_offset != 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;
}
675

B
bellard 已提交
676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698
static int decompress_cluster(BDRVQcowState *s, uint64_t cluster_offset)
{
    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;
        ret = bdrv_read(s->hd, coffset >> 9, s->cluster_data, nb_csectors);
        if (ret < 0) {
            return -1;
        }
        if (decompress_buffer(s->cluster_cache, s->cluster_size,
                              s->cluster_data + sector_offset, csize) < 0) {
            return -1;
        }
        s->cluster_cache_offset = coffset;
    }
    return 0;
}

B
bellard 已提交
699
/* handle reading after the end of the backing file */
700
static int backing_read1(BlockDriverState *bs,
B
bellard 已提交
701 702 703 704 705 706 707 708 709 710 711 712 713
                         int64_t sector_num, uint8_t *buf, int nb_sectors)
{
    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;
    memset(buf + n1 * 512, 0, 512 * (nb_sectors - n1));
    return n1;
}

714
static int qcow_read(BlockDriverState *bs, int64_t sector_num,
B
bellard 已提交
715 716 717
                     uint8_t *buf, int nb_sectors)
{
    BDRVQcowState *s = bs->opaque;
B
bellard 已提交
718
    int ret, index_in_cluster, n, n1;
B
bellard 已提交
719
    uint64_t cluster_offset;
720

B
bellard 已提交
721 722 723 724 725 726 727 728 729
    while (nb_sectors > 0) {
        cluster_offset = get_cluster_offset(bs, sector_num << 9, 0, 0, 0, 0);
        index_in_cluster = sector_num & (s->cluster_sectors - 1);
        n = s->cluster_sectors - index_in_cluster;
        if (n > nb_sectors)
            n = nb_sectors;
        if (!cluster_offset) {
            if (bs->backing_hd) {
                /* read from the base image */
B
bellard 已提交
730 731 732 733 734 735
                n1 = backing_read1(bs->backing_hd, sector_num, buf, n);
                if (n1 > 0) {
                    ret = bdrv_read(bs->backing_hd, sector_num, buf, n1);
                    if (ret < 0)
                        return -1;
                }
B
bellard 已提交
736 737 738 739 740 741 742 743 744
            } else {
                memset(buf, 0, 512 * n);
            }
        } else if (cluster_offset & QCOW_OFLAG_COMPRESSED) {
            if (decompress_cluster(s, cluster_offset) < 0)
                return -1;
            memcpy(buf, s->cluster_cache + index_in_cluster * 512, 512 * n);
        } else {
            ret = bdrv_pread(s->hd, cluster_offset + index_in_cluster * 512, buf, n * 512);
745
            if (ret != n * 512)
B
bellard 已提交
746 747
                return -1;
            if (s->crypt_method) {
748
                encrypt_sectors(s, sector_num, buf, buf, n, 0,
B
bellard 已提交
749 750 751 752 753 754 755 756 757 758
                                &s->aes_decrypt_key);
            }
        }
        nb_sectors -= n;
        sector_num += n;
        buf += n * 512;
    }
    return 0;
}

759
static int qcow_write(BlockDriverState *bs, int64_t sector_num,
B
bellard 已提交
760 761 762 763 764
                     const uint8_t *buf, int nb_sectors)
{
    BDRVQcowState *s = bs->opaque;
    int ret, index_in_cluster, n;
    uint64_t cluster_offset;
765

B
bellard 已提交
766 767 768 769 770
    while (nb_sectors > 0) {
        index_in_cluster = sector_num & (s->cluster_sectors - 1);
        n = s->cluster_sectors - index_in_cluster;
        if (n > nb_sectors)
            n = nb_sectors;
771 772
        cluster_offset = get_cluster_offset(bs, sector_num << 9, 1, 0,
                                            index_in_cluster,
B
bellard 已提交
773 774 775 776 777 778
                                            index_in_cluster + n);
        if (!cluster_offset)
            return -1;
        if (s->crypt_method) {
            encrypt_sectors(s, sector_num, s->cluster_data, buf, n, 1,
                            &s->aes_encrypt_key);
779
            ret = bdrv_pwrite(s->hd, cluster_offset + index_in_cluster * 512,
B
bellard 已提交
780 781 782 783
                              s->cluster_data, n * 512);
        } else {
            ret = bdrv_pwrite(s->hd, cluster_offset + index_in_cluster * 512, buf, n * 512);
        }
784
        if (ret != n * 512)
B
bellard 已提交
785 786 787 788 789 790 791 792 793
            return -1;
        nb_sectors -= n;
        sector_num += n;
        buf += n * 512;
    }
    s->cluster_cache_offset = -1; /* disable compressed cache */
    return 0;
}

794 795
typedef struct QCowAIOCB {
    BlockDriverAIOCB common;
B
bellard 已提交
796 797 798 799 800
    int64_t sector_num;
    uint8_t *buf;
    int nb_sectors;
    int n;
    uint64_t cluster_offset;
801
    uint8_t *cluster_data;
B
bellard 已提交
802 803 804 805 806
    BlockDriverAIOCB *hd_aiocb;
} QCowAIOCB;

static void qcow_aio_read_cb(void *opaque, int ret)
{
807 808
    QCowAIOCB *acb = opaque;
    BlockDriverState *bs = acb->common.bs;
B
bellard 已提交
809
    BDRVQcowState *s = bs->opaque;
B
bellard 已提交
810
    int index_in_cluster, n1;
B
bellard 已提交
811

812
    acb->hd_aiocb = NULL;
B
bellard 已提交
813 814
    if (ret < 0) {
    fail:
815 816
        acb->common.cb(acb->common.opaque, ret);
        qemu_aio_release(acb);
B
bellard 已提交
817 818 819 820 821
        return;
    }

 redo:
    /* post process the read buffer */
822
    if (!acb->cluster_offset) {
B
bellard 已提交
823
        /* nothing to do */
824
    } else if (acb->cluster_offset & QCOW_OFLAG_COMPRESSED) {
B
bellard 已提交
825 826 827
        /* nothing to do */
    } else {
        if (s->crypt_method) {
828 829
            encrypt_sectors(s, acb->sector_num, acb->buf, acb->buf,
                            acb->n, 0,
B
bellard 已提交
830 831 832 833
                            &s->aes_decrypt_key);
        }
    }

834 835 836
    acb->nb_sectors -= acb->n;
    acb->sector_num += acb->n;
    acb->buf += acb->n * 512;
B
bellard 已提交
837

838
    if (acb->nb_sectors == 0) {
B
bellard 已提交
839
        /* request completed */
840 841
        acb->common.cb(acb->common.opaque, 0);
        qemu_aio_release(acb);
B
bellard 已提交
842 843
        return;
    }
844

B
bellard 已提交
845
    /* prepare next AIO request */
846
    acb->cluster_offset = get_cluster_offset(bs, acb->sector_num << 9,
847 848 849 850 851 852 853
                                             0, 0, 0, 0);
    index_in_cluster = acb->sector_num & (s->cluster_sectors - 1);
    acb->n = s->cluster_sectors - index_in_cluster;
    if (acb->n > acb->nb_sectors)
        acb->n = acb->nb_sectors;

    if (!acb->cluster_offset) {
B
bellard 已提交
854 855
        if (bs->backing_hd) {
            /* read from the base image */
856
            n1 = backing_read1(bs->backing_hd, acb->sector_num,
857
                               acb->buf, acb->n);
B
bellard 已提交
858
            if (n1 > 0) {
859
                acb->hd_aiocb = bdrv_aio_read(bs->backing_hd, acb->sector_num,
860 861
                                    acb->buf, acb->n, qcow_aio_read_cb, acb);
                if (acb->hd_aiocb == NULL)
B
bellard 已提交
862 863 864 865
                    goto fail;
            } else {
                goto redo;
            }
B
bellard 已提交
866 867
        } else {
            /* Note: in this case, no need to wait */
868
            memset(acb->buf, 0, 512 * acb->n);
B
bellard 已提交
869 870
            goto redo;
        }
871
    } else if (acb->cluster_offset & QCOW_OFLAG_COMPRESSED) {
B
bellard 已提交
872
        /* add AIO support for compressed blocks ? */
873
        if (decompress_cluster(s, acb->cluster_offset) < 0)
B
bellard 已提交
874
            goto fail;
875
        memcpy(acb->buf,
876
               s->cluster_cache + index_in_cluster * 512, 512 * acb->n);
B
bellard 已提交
877 878
        goto redo;
    } else {
879
        if ((acb->cluster_offset & 511) != 0) {
B
bellard 已提交
880 881 882
            ret = -EIO;
            goto fail;
        }
883
        acb->hd_aiocb = bdrv_aio_read(s->hd,
884
                            (acb->cluster_offset >> 9) + index_in_cluster,
885 886
                            acb->buf, acb->n, qcow_aio_read_cb, acb);
        if (acb->hd_aiocb == NULL)
B
bellard 已提交
887 888 889 890
            goto fail;
    }
}

891 892 893
static QCowAIOCB *qcow_aio_setup(BlockDriverState *bs,
        int64_t sector_num, uint8_t *buf, int nb_sectors,
        BlockDriverCompletionFunc *cb, void *opaque)
B
bellard 已提交
894
{
895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917
    QCowAIOCB *acb;

    acb = qemu_aio_get(bs, cb, opaque);
    if (!acb)
        return NULL;
    acb->hd_aiocb = NULL;
    acb->sector_num = sector_num;
    acb->buf = buf;
    acb->nb_sectors = nb_sectors;
    acb->n = 0;
    acb->cluster_offset = 0;
    return acb;
}

static BlockDriverAIOCB *qcow_aio_read(BlockDriverState *bs,
        int64_t sector_num, uint8_t *buf, int nb_sectors,
        BlockDriverCompletionFunc *cb, void *opaque)
{
    QCowAIOCB *acb;

    acb = qcow_aio_setup(bs, sector_num, buf, nb_sectors, cb, opaque);
    if (!acb)
        return NULL;
B
bellard 已提交
918 919

    qcow_aio_read_cb(acb, 0);
920
    return &acb->common;
B
bellard 已提交
921 922 923 924
}

static void qcow_aio_write_cb(void *opaque, int ret)
{
925 926
    QCowAIOCB *acb = opaque;
    BlockDriverState *bs = acb->common.bs;
B
bellard 已提交
927 928 929 930
    BDRVQcowState *s = bs->opaque;
    int index_in_cluster;
    uint64_t cluster_offset;
    const uint8_t *src_buf;
931 932 933

    acb->hd_aiocb = NULL;

B
bellard 已提交
934 935
    if (ret < 0) {
    fail:
936 937
        acb->common.cb(acb->common.opaque, ret);
        qemu_aio_release(acb);
B
bellard 已提交
938 939 940
        return;
    }

941 942 943
    acb->nb_sectors -= acb->n;
    acb->sector_num += acb->n;
    acb->buf += acb->n * 512;
B
bellard 已提交
944

945
    if (acb->nb_sectors == 0) {
B
bellard 已提交
946
        /* request completed */
947 948
        acb->common.cb(acb->common.opaque, 0);
        qemu_aio_release(acb);
B
bellard 已提交
949 950
        return;
    }
951

952 953 954 955
    index_in_cluster = acb->sector_num & (s->cluster_sectors - 1);
    acb->n = s->cluster_sectors - index_in_cluster;
    if (acb->n > acb->nb_sectors)
        acb->n = acb->nb_sectors;
956 957
    cluster_offset = get_cluster_offset(bs, acb->sector_num << 9, 1, 0,
                                        index_in_cluster,
958
                                        index_in_cluster + acb->n);
B
bellard 已提交
959 960 961 962 963
    if (!cluster_offset || (cluster_offset & 511) != 0) {
        ret = -EIO;
        goto fail;
    }
    if (s->crypt_method) {
964 965 966
        if (!acb->cluster_data) {
            acb->cluster_data = qemu_mallocz(s->cluster_size);
            if (!acb->cluster_data) {
B
bellard 已提交
967 968 969 970
                ret = -ENOMEM;
                goto fail;
            }
        }
971
        encrypt_sectors(s, acb->sector_num, acb->cluster_data, acb->buf,
972 973
                        acb->n, 1, &s->aes_encrypt_key);
        src_buf = acb->cluster_data;
B
bellard 已提交
974
    } else {
975
        src_buf = acb->buf;
B
bellard 已提交
976
    }
977
    acb->hd_aiocb = bdrv_aio_write(s->hd,
978 979
                                   (cluster_offset >> 9) + index_in_cluster,
                                   src_buf, acb->n,
980 981
                                   qcow_aio_write_cb, acb);
    if (acb->hd_aiocb == NULL)
B
bellard 已提交
982 983 984
        goto fail;
}

985 986 987
static BlockDriverAIOCB *qcow_aio_write(BlockDriverState *bs,
        int64_t sector_num, const uint8_t *buf, int nb_sectors,
        BlockDriverCompletionFunc *cb, void *opaque)
B
bellard 已提交
988 989
{
    BDRVQcowState *s = bs->opaque;
990
    QCowAIOCB *acb;
991

B
bellard 已提交
992 993
    s->cluster_cache_offset = -1; /* disable compressed cache */

994 995 996
    acb = qcow_aio_setup(bs, sector_num, (uint8_t*)buf, nb_sectors, cb, opaque);
    if (!acb)
        return NULL;
997

B
bellard 已提交
998
    qcow_aio_write_cb(acb, 0);
999
    return &acb->common;
B
bellard 已提交
1000 1001
}

1002
static void qcow_aio_cancel(BlockDriverAIOCB *blockacb)
B
bellard 已提交
1003
{
1004 1005 1006 1007
    QCowAIOCB *acb = (QCowAIOCB *)blockacb;
    if (acb->hd_aiocb)
        bdrv_aio_cancel(acb->hd_aiocb);
    qemu_aio_release(acb);
B
bellard 已提交
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
}

static void qcow_close(BlockDriverState *bs)
{
    BDRVQcowState *s = bs->opaque;
    qemu_free(s->l1_table);
    qemu_free(s->l2_cache);
    qemu_free(s->cluster_cache);
    qemu_free(s->cluster_data);
    refcount_close(bs);
    bdrv_delete(s->hd);
}

/* XXX: use std qcow open function ? */
typedef struct QCowCreateState {
    int cluster_size;
    int cluster_bits;
    uint16_t *refcount_block;
    uint64_t *refcount_table;
    int64_t l1_table_offset;
    int64_t refcount_table_offset;
    int64_t refcount_block_offset;
} QCowCreateState;

static void create_refcount_update(QCowCreateState *s,
                                   int64_t offset, int64_t size)
{
    int refcount;
    int64_t start, last, cluster_offset;
    uint16_t *p;

    start = offset & ~(s->cluster_size - 1);
    last = (offset + size - 1)  & ~(s->cluster_size - 1);
1041
    for(cluster_offset = start; cluster_offset <= last;
B
bellard 已提交
1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056
        cluster_offset += s->cluster_size) {
        p = &s->refcount_block[cluster_offset >> s->cluster_bits];
        refcount = be16_to_cpu(*p);
        refcount++;
        *p = cpu_to_be16(refcount);
    }
}

static int qcow_create(const char *filename, int64_t total_size,
                      const char *backing_file, int flags)
{
    int fd, header_size, backing_filename_len, l1_size, i, shift, l2_bits;
    QCowHeader header;
    uint64_t tmp, offset;
    QCowCreateState s1, *s = &s1;
1057

B
bellard 已提交
1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078
    memset(s, 0, sizeof(*s));

    fd = open(filename, O_WRONLY | O_CREAT | O_TRUNC | O_BINARY, 0644);
    if (fd < 0)
        return -1;
    memset(&header, 0, sizeof(header));
    header.magic = cpu_to_be32(QCOW_MAGIC);
    header.version = cpu_to_be32(QCOW_VERSION);
    header.size = cpu_to_be64(total_size * 512);
    header_size = sizeof(header);
    backing_filename_len = 0;
    if (backing_file) {
        header.backing_file_offset = cpu_to_be64(header_size);
        backing_filename_len = strlen(backing_file);
        header.backing_file_size = cpu_to_be32(backing_filename_len);
        header_size += backing_filename_len;
    }
    s->cluster_bits = 12;  /* 4 KB clusters */
    s->cluster_size = 1 << s->cluster_bits;
    header.cluster_bits = cpu_to_be32(s->cluster_bits);
    header_size = (header_size + 7) & ~7;
1079
    if (flags & BLOCK_FLAG_ENCRYPT) {
B
bellard 已提交
1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090
        header.crypt_method = cpu_to_be32(QCOW_CRYPT_AES);
    } else {
        header.crypt_method = cpu_to_be32(QCOW_CRYPT_NONE);
    }
    l2_bits = s->cluster_bits - 3;
    shift = s->cluster_bits + l2_bits;
    l1_size = (((total_size * 512) + (1LL << shift) - 1) >> shift);
    offset = align_offset(header_size, s->cluster_size);
    s->l1_table_offset = offset;
    header.l1_table_offset = cpu_to_be64(s->l1_table_offset);
    header.l1_size = cpu_to_be32(l1_size);
B
bellard 已提交
1091
    offset += align_offset(l1_size * sizeof(uint64_t), s->cluster_size);
B
bellard 已提交
1092 1093 1094 1095 1096 1097 1098

    s->refcount_table = qemu_mallocz(s->cluster_size);
    if (!s->refcount_table)
        goto fail;
    s->refcount_block = qemu_mallocz(s->cluster_size);
    if (!s->refcount_block)
        goto fail;
1099

B
bellard 已提交
1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110
    s->refcount_table_offset = offset;
    header.refcount_table_offset = cpu_to_be64(offset);
    header.refcount_table_clusters = cpu_to_be32(1);
    offset += s->cluster_size;

    s->refcount_table[0] = cpu_to_be64(offset);
    s->refcount_block_offset = offset;
    offset += s->cluster_size;

    /* update refcounts */
    create_refcount_update(s, 0, header_size);
B
bellard 已提交
1111
    create_refcount_update(s, s->l1_table_offset, l1_size * sizeof(uint64_t));
B
bellard 已提交
1112 1113
    create_refcount_update(s, s->refcount_table_offset, s->cluster_size);
    create_refcount_update(s, s->refcount_block_offset, s->cluster_size);
1114

B
bellard 已提交
1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126
    /* write all the data */
    write(fd, &header, sizeof(header));
    if (backing_file) {
        write(fd, backing_file, backing_filename_len);
    }
    lseek(fd, s->l1_table_offset, SEEK_SET);
    tmp = 0;
    for(i = 0;i < l1_size; i++) {
        write(fd, &tmp, sizeof(tmp));
    }
    lseek(fd, s->refcount_table_offset, SEEK_SET);
    write(fd, s->refcount_table, s->cluster_size);
1127

B
bellard 已提交
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
    lseek(fd, s->refcount_block_offset, SEEK_SET);
    write(fd, s->refcount_block, s->cluster_size);

    qemu_free(s->refcount_table);
    qemu_free(s->refcount_block);
    close(fd);
    return 0;
 fail:
    qemu_free(s->refcount_table);
    qemu_free(s->refcount_block);
    close(fd);
    return -ENOMEM;
}

static int qcow_make_empty(BlockDriverState *bs)
{
#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);
    if (bdrv_pwrite(s->hd, s->l1_table_offset, s->l1_table, l1_length) < 0)
	return -1;
    ret = bdrv_truncate(s->hd, s->l1_table_offset + l1_length);
    if (ret < 0)
        return ret;
1156

B
bellard 已提交
1157 1158 1159 1160 1161 1162 1163
    l2_cache_reset(bs);
#endif
    return 0;
}

/* XXX: put compressed sectors first, then all the cluster aligned
   tables to avoid losing bytes in alignment */
1164
static int qcow_write_compressed(BlockDriverState *bs, int64_t sector_num,
B
bellard 已提交
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
                                 const uint8_t *buf, int nb_sectors)
{
    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 */
        cluster_offset = bdrv_getlength(s->hd);
        cluster_offset = (cluster_offset + 511) & ~511;
        bdrv_truncate(s->hd, cluster_offset);
        return 0;
    }

    if (nb_sectors != s->cluster_sectors)
        return -EINVAL;

    out_buf = qemu_malloc(s->cluster_size + (s->cluster_size / 1000) + 128);
    if (!out_buf)
        return -ENOMEM;

    /* best compression, small window, no zlib header */
    memset(&strm, 0, sizeof(strm));
    ret = deflateInit2(&strm, Z_DEFAULT_COMPRESSION,
1192
                       Z_DEFLATED, -12,
B
bellard 已提交
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
                       9, Z_DEFAULT_STRATEGY);
    if (ret != 0) {
        qemu_free(out_buf);
        return -1;
    }

    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) {
        qemu_free(out_buf);
        deflateEnd(&strm);
        return -1;
    }
    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 */
        qcow_write(bs, sector_num, buf, s->cluster_sectors);
    } else {
1218
        cluster_offset = get_cluster_offset(bs, sector_num << 9, 2,
B
bellard 已提交
1219 1220 1221 1222 1223 1224 1225
                                            out_len, 0, 0);
        cluster_offset &= s->cluster_offset_mask;
        if (bdrv_pwrite(s->hd, cluster_offset, out_buf, out_len) != out_len) {
            qemu_free(out_buf);
            return -1;
        }
    }
1226

B
bellard 已提交
1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240
    qemu_free(out_buf);
    return 0;
}

static void qcow_flush(BlockDriverState *bs)
{
    BDRVQcowState *s = bs->opaque;
    bdrv_flush(s->hd);
}

static int qcow_get_info(BlockDriverState *bs, BlockDriverInfo *bdi)
{
    BDRVQcowState *s = bs->opaque;
    bdi->cluster_size = s->cluster_size;
1241
    bdi->vm_state_offset = (int64_t)s->l1_vm_state_index <<
B
bellard 已提交
1242 1243 1244 1245 1246 1247 1248 1249
        (s->cluster_bits + s->l2_bits);
    return 0;
}

/*********************************************************/
/* snapshot support */

/* update the refcounts of snapshots and the copied flag */
1250
static int update_snapshot_refcount(BlockDriverState *bs,
B
bellard 已提交
1251 1252 1253 1254 1255 1256 1257 1258
                                    int64_t l1_table_offset,
                                    int l1_size,
                                    int addend)
{
    BDRVQcowState *s = bs->opaque;
    uint64_t *l1_table, *l2_table, l2_offset, offset, l1_size2, l1_allocated;
    int64_t old_offset, old_l2_offset;
    int l2_size, i, j, l1_modified, l2_modified, nb_csectors, refcount;
1259

B
bellard 已提交
1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270
    l2_cache_reset(bs);

    l2_table = NULL;
    l1_table = NULL;
    l1_size2 = l1_size * sizeof(uint64_t);
    l1_allocated = 0;
    if (l1_table_offset != s->l1_table_offset) {
        l1_table = qemu_malloc(l1_size2);
        if (!l1_table)
            goto fail;
        l1_allocated = 1;
1271
        if (bdrv_pread(s->hd, l1_table_offset,
B
bellard 已提交
1272 1273 1274 1275 1276 1277 1278 1279 1280
                       l1_table, l1_size2) != l1_size2)
            goto fail;
        for(i = 0;i < l1_size; i++)
            be64_to_cpus(&l1_table[i]);
    } else {
        assert(l1_size == s->l1_size);
        l1_table = s->l1_table;
        l1_allocated = 0;
    }
1281

B
bellard 已提交
1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300
    l2_size = s->l2_size * sizeof(uint64_t);
    l2_table = qemu_malloc(l2_size);
    if (!l2_table)
        goto fail;
    l1_modified = 0;
    for(i = 0; i < l1_size; i++) {
        l2_offset = l1_table[i];
        if (l2_offset) {
            old_l2_offset = l2_offset;
            l2_offset &= ~QCOW_OFLAG_COPIED;
            l2_modified = 0;
            if (bdrv_pread(s->hd, l2_offset, l2_table, l2_size) != l2_size)
                goto fail;
            for(j = 0; j < s->l2_size; j++) {
                offset = be64_to_cpu(l2_table[j]);
                if (offset != 0) {
                    old_offset = offset;
                    offset &= ~QCOW_OFLAG_COPIED;
                    if (offset & QCOW_OFLAG_COMPRESSED) {
1301
                        nb_csectors = ((offset >> s->csize_shift) &
B
bellard 已提交
1302 1303 1304 1305 1306
                                       s->csize_mask) + 1;
                        if (addend != 0)
                            update_refcount(bs, (offset & s->cluster_offset_mask) & ~511,
                                            nb_csectors * 512, addend);
                        /* compressed clusters are never modified */
1307
                        refcount = 2;
B
bellard 已提交
1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325
                    } else {
                        if (addend != 0) {
                            refcount = update_cluster_refcount(bs, offset >> s->cluster_bits, addend);
                        } else {
                            refcount = get_refcount(bs, offset >> s->cluster_bits);
                        }
                    }

                    if (refcount == 1) {
                        offset |= QCOW_OFLAG_COPIED;
                    }
                    if (offset != old_offset) {
                        l2_table[j] = cpu_to_be64(offset);
                        l2_modified = 1;
                    }
                }
            }
            if (l2_modified) {
1326
                if (bdrv_pwrite(s->hd,
B
bellard 已提交
1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347
                                l2_offset, l2_table, l2_size) != l2_size)
                    goto fail;
            }

            if (addend != 0) {
                refcount = update_cluster_refcount(bs, l2_offset >> s->cluster_bits, addend);
            } else {
                refcount = get_refcount(bs, l2_offset >> s->cluster_bits);
            }
            if (refcount == 1) {
                l2_offset |= QCOW_OFLAG_COPIED;
            }
            if (l2_offset != old_l2_offset) {
                l1_table[i] = l2_offset;
                l1_modified = 1;
            }
        }
    }
    if (l1_modified) {
        for(i = 0; i < l1_size; i++)
            cpu_to_be64s(&l1_table[i]);
1348
        if (bdrv_pwrite(s->hd, l1_table_offset, l1_table,
B
bellard 已提交
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
                        l1_size2) != l1_size2)
            goto fail;
        for(i = 0; i < l1_size; i++)
            be64_to_cpus(&l1_table[i]);
    }
    if (l1_allocated)
        qemu_free(l1_table);
    qemu_free(l2_table);
    return 0;
 fail:
    if (l1_allocated)
        qemu_free(l1_table);
    qemu_free(l2_table);
    return -EIO;
}

static void qcow_free_snapshots(BlockDriverState *bs)
{
    BDRVQcowState *s = bs->opaque;
    int i;

    for(i = 0; i < s->nb_snapshots; i++) {
        qemu_free(s->snapshots[i].name);
        qemu_free(s->snapshots[i].id_str);
    }
    qemu_free(s->snapshots);
    s->snapshots = NULL;
    s->nb_snapshots = 0;
}

static int qcow_read_snapshots(BlockDriverState *bs)
{
    BDRVQcowState *s = bs->opaque;
    QCowSnapshotHeader h;
    QCowSnapshot *sn;
    int i, id_str_size, name_size;
    int64_t offset;
    uint32_t extra_data_size;

    offset = s->snapshots_offset;
    s->snapshots = qemu_mallocz(s->nb_snapshots * sizeof(QCowSnapshot));
    if (!s->snapshots)
        goto fail;
    for(i = 0; i < s->nb_snapshots; i++) {
        offset = align_offset(offset, 8);
        if (bdrv_pread(s->hd, offset, &h, sizeof(h)) != sizeof(h))
            goto fail;
        offset += sizeof(h);
        sn = s->snapshots + i;
        sn->l1_table_offset = be64_to_cpu(h.l1_table_offset);
        sn->l1_size = be32_to_cpu(h.l1_size);
        sn->vm_state_size = be32_to_cpu(h.vm_state_size);
        sn->date_sec = be32_to_cpu(h.date_sec);
        sn->date_nsec = be32_to_cpu(h.date_nsec);
        sn->vm_clock_nsec = be64_to_cpu(h.vm_clock_nsec);
        extra_data_size = be32_to_cpu(h.extra_data_size);

        id_str_size = be16_to_cpu(h.id_str_size);
        name_size = be16_to_cpu(h.name_size);

        offset += extra_data_size;

        sn->id_str = qemu_malloc(id_str_size + 1);
        if (!sn->id_str)
            goto fail;
        if (bdrv_pread(s->hd, offset, sn->id_str, id_str_size) != id_str_size)
            goto fail;
        offset += id_str_size;
        sn->id_str[id_str_size] = '\0';

        sn->name = qemu_malloc(name_size + 1);
        if (!sn->name)
            goto fail;
        if (bdrv_pread(s->hd, offset, sn->name, name_size) != name_size)
            goto fail;
        offset += name_size;
        sn->name[name_size] = '\0';
    }
    s->snapshots_size = offset - s->snapshots_offset;
    return 0;
 fail:
    qcow_free_snapshots(bs);
    return -1;
}

/* add at the end of the file a new list of snapshots */
static int qcow_write_snapshots(BlockDriverState *bs)
{
    BDRVQcowState *s = bs->opaque;
    QCowSnapshot *sn;
    QCowSnapshotHeader h;
    int i, name_size, id_str_size, snapshots_size;
    uint64_t data64;
    uint32_t data32;
    int64_t offset, snapshots_offset;

    /* compute the size of the snapshots */
    offset = 0;
    for(i = 0; i < s->nb_snapshots; i++) {
        sn = s->snapshots + i;
        offset = align_offset(offset, 8);
        offset += sizeof(h);
        offset += strlen(sn->id_str);
        offset += strlen(sn->name);
    }
    snapshots_size = offset;

    snapshots_offset = alloc_clusters(bs, snapshots_size);
    offset = snapshots_offset;
1458

B
bellard 已提交
1459 1460 1461 1462 1463 1464 1465 1466 1467
    for(i = 0; i < s->nb_snapshots; i++) {
        sn = s->snapshots + i;
        memset(&h, 0, sizeof(h));
        h.l1_table_offset = cpu_to_be64(sn->l1_table_offset);
        h.l1_size = cpu_to_be32(sn->l1_size);
        h.vm_state_size = cpu_to_be32(sn->vm_state_size);
        h.date_sec = cpu_to_be32(sn->date_sec);
        h.date_nsec = cpu_to_be32(sn->date_nsec);
        h.vm_clock_nsec = cpu_to_be64(sn->vm_clock_nsec);
1468

B
bellard 已提交
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
        id_str_size = strlen(sn->id_str);
        name_size = strlen(sn->name);
        h.id_str_size = cpu_to_be16(id_str_size);
        h.name_size = cpu_to_be16(name_size);
        offset = align_offset(offset, 8);
        if (bdrv_pwrite(s->hd, offset, &h, sizeof(h)) != sizeof(h))
            goto fail;
        offset += sizeof(h);
        if (bdrv_pwrite(s->hd, offset, sn->id_str, id_str_size) != id_str_size)
            goto fail;
        offset += id_str_size;
        if (bdrv_pwrite(s->hd, offset, sn->name, name_size) != name_size)
            goto fail;
        offset += name_size;
    }

    /* update the various header fields */
    data64 = cpu_to_be64(snapshots_offset);
    if (bdrv_pwrite(s->hd, offsetof(QCowHeader, snapshots_offset),
                    &data64, sizeof(data64)) != sizeof(data64))
        goto fail;
    data32 = cpu_to_be32(s->nb_snapshots);
    if (bdrv_pwrite(s->hd, offsetof(QCowHeader, nb_snapshots),
                    &data32, sizeof(data32)) != sizeof(data32))
        goto fail;

    /* free the old snapshot table */
    free_clusters(bs, s->snapshots_offset, s->snapshots_size);
    s->snapshots_offset = snapshots_offset;
    s->snapshots_size = snapshots_size;
    return 0;
 fail:
    return -1;
}

static void find_new_snapshot_id(BlockDriverState *bs,
                                 char *id_str, int id_str_size)
{
    BDRVQcowState *s = bs->opaque;
    QCowSnapshot *sn;
    int i, id, id_max = 0;

    for(i = 0; i < s->nb_snapshots; i++) {
        sn = s->snapshots + i;
        id = strtoul(sn->id_str, NULL, 10);
        if (id > id_max)
            id_max = id;
    }
    snprintf(id_str, id_str_size, "%d", id_max + 1);
}

static int find_snapshot_by_id(BlockDriverState *bs, const char *id_str)
{
    BDRVQcowState *s = bs->opaque;
    int i;

    for(i = 0; i < s->nb_snapshots; i++) {
        if (!strcmp(s->snapshots[i].id_str, id_str))
            return i;
    }
    return -1;
}

static int find_snapshot_by_id_or_name(BlockDriverState *bs, const char *name)
{
    BDRVQcowState *s = bs->opaque;
    int i, ret;
1536

B
bellard 已提交
1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547
    ret = find_snapshot_by_id(bs, name);
    if (ret >= 0)
        return ret;
    for(i = 0; i < s->nb_snapshots; i++) {
        if (!strcmp(s->snapshots[i].name, name))
            return i;
    }
    return -1;
}

/* if no id is provided, a new one is constructed */
1548
static int qcow_snapshot_create(BlockDriverState *bs,
B
bellard 已提交
1549 1550 1551 1552 1553 1554
                                QEMUSnapshotInfo *sn_info)
{
    BDRVQcowState *s = bs->opaque;
    QCowSnapshot *snapshots1, sn1, *sn = &sn1;
    int i, ret;
    uint64_t *l1_table = NULL;
1555

B
bellard 已提交
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
    memset(sn, 0, sizeof(*sn));

    if (sn_info->id_str[0] == '\0') {
        /* compute a new id */
        find_new_snapshot_id(bs, sn_info->id_str, sizeof(sn_info->id_str));
    }

    /* check that the ID is unique */
    if (find_snapshot_by_id(bs, sn_info->id_str) >= 0)
        return -ENOENT;

    sn->id_str = qemu_strdup(sn_info->id_str);
    if (!sn->id_str)
        goto fail;
    sn->name = qemu_strdup(sn_info->name);
    if (!sn->name)
        goto fail;
    sn->vm_state_size = sn_info->vm_state_size;
    sn->date_sec = sn_info->date_sec;
    sn->date_nsec = sn_info->date_nsec;
    sn->vm_clock_nsec = sn_info->vm_clock_nsec;

    ret = update_snapshot_refcount(bs, s->l1_table_offset, s->l1_size, 1);
    if (ret < 0)
        goto fail;

    /* create the L1 table of the snapshot */
    sn->l1_table_offset = alloc_clusters(bs, s->l1_size * sizeof(uint64_t));
    sn->l1_size = s->l1_size;

    l1_table = qemu_malloc(s->l1_size * sizeof(uint64_t));
    if (!l1_table)
        goto fail;
    for(i = 0; i < s->l1_size; i++) {
        l1_table[i] = cpu_to_be64(s->l1_table[i]);
    }
    if (bdrv_pwrite(s->hd, sn->l1_table_offset,
1593
                    l1_table, s->l1_size * sizeof(uint64_t)) !=
B
bellard 已提交
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
        (s->l1_size * sizeof(uint64_t)))
        goto fail;
    qemu_free(l1_table);
    l1_table = NULL;

    snapshots1 = qemu_malloc((s->nb_snapshots + 1) * sizeof(QCowSnapshot));
    if (!snapshots1)
        goto fail;
    memcpy(snapshots1, s->snapshots, s->nb_snapshots * sizeof(QCowSnapshot));
    s->snapshots = snapshots1;
    s->snapshots[s->nb_snapshots++] = *sn;

    if (qcow_write_snapshots(bs) < 0)
        goto fail;
#ifdef DEBUG_ALLOC
    check_refcounts(bs);
#endif
    return 0;
 fail:
    qemu_free(sn->name);
    qemu_free(l1_table);
    return -1;
}

/* copy the snapshot 'snapshot_name' into the current disk image */
1619
static int qcow_snapshot_goto(BlockDriverState *bs,
B
bellard 已提交
1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639
                              const char *snapshot_id)
{
    BDRVQcowState *s = bs->opaque;
    QCowSnapshot *sn;
    int i, snapshot_index, l1_size2;

    snapshot_index = find_snapshot_by_id_or_name(bs, snapshot_id);
    if (snapshot_index < 0)
        return -ENOENT;
    sn = &s->snapshots[snapshot_index];

    if (update_snapshot_refcount(bs, s->l1_table_offset, s->l1_size, -1) < 0)
        goto fail;

    if (grow_l1_table(bs, sn->l1_size) < 0)
        goto fail;

    s->l1_size = sn->l1_size;
    l1_size2 = s->l1_size * sizeof(uint64_t);
    /* copy the snapshot l1 table to the current l1 table */
1640
    if (bdrv_pread(s->hd, sn->l1_table_offset,
B
bellard 已提交
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
                   s->l1_table, l1_size2) != l1_size2)
        goto fail;
    if (bdrv_pwrite(s->hd, s->l1_table_offset,
                    s->l1_table, l1_size2) != l1_size2)
        goto fail;
    for(i = 0;i < s->l1_size; i++) {
        be64_to_cpus(&s->l1_table[i]);
    }

    if (update_snapshot_refcount(bs, s->l1_table_offset, s->l1_size, 1) < 0)
        goto fail;

#ifdef DEBUG_ALLOC
    check_refcounts(bs);
#endif
    return 0;
 fail:
    return -EIO;
}

static int qcow_snapshot_delete(BlockDriverState *bs, const char *snapshot_id)
{
    BDRVQcowState *s = bs->opaque;
    QCowSnapshot *sn;
    int snapshot_index, ret;
1666

B
bellard 已提交
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
    snapshot_index = find_snapshot_by_id_or_name(bs, snapshot_id);
    if (snapshot_index < 0)
        return -ENOENT;
    sn = &s->snapshots[snapshot_index];

    ret = update_snapshot_refcount(bs, sn->l1_table_offset, sn->l1_size, -1);
    if (ret < 0)
        return ret;
    /* must update the copied flag on the current cluster offsets */
    ret = update_snapshot_refcount(bs, s->l1_table_offset, s->l1_size, 0);
    if (ret < 0)
        return ret;
    free_clusters(bs, sn->l1_table_offset, sn->l1_size * sizeof(uint64_t));

    qemu_free(sn->id_str);
    qemu_free(sn->name);
    memmove(sn, sn + 1, (s->nb_snapshots - snapshot_index - 1) * sizeof(*sn));
    s->nb_snapshots--;
    ret = qcow_write_snapshots(bs);
    if (ret < 0) {
        /* XXX: restore snapshot if error ? */
        return ret;
    }
#ifdef DEBUG_ALLOC
    check_refcounts(bs);
#endif
    return 0;
}

1696
static int qcow_snapshot_list(BlockDriverState *bs,
B
bellard 已提交
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
                              QEMUSnapshotInfo **psn_tab)
{
    BDRVQcowState *s = bs->opaque;
    QEMUSnapshotInfo *sn_tab, *sn_info;
    QCowSnapshot *sn;
    int i;

    sn_tab = qemu_mallocz(s->nb_snapshots * sizeof(QEMUSnapshotInfo));
    if (!sn_tab)
        goto fail;
    for(i = 0; i < s->nb_snapshots; i++) {
        sn_info = sn_tab + i;
        sn = s->snapshots + i;
        pstrcpy(sn_info->id_str, sizeof(sn_info->id_str),
                sn->id_str);
        pstrcpy(sn_info->name, sizeof(sn_info->name),
                sn->name);
        sn_info->vm_state_size = sn->vm_state_size;
        sn_info->date_sec = sn->date_sec;
        sn_info->date_nsec = sn->date_nsec;
        sn_info->vm_clock_nsec = sn->vm_clock_nsec;
    }
    *psn_tab = sn_tab;
    return s->nb_snapshots;
 fail:
    qemu_free(sn_tab);
    *psn_tab = NULL;
    return -ENOMEM;
}

/*********************************************************/
/* refcount handling */

static int refcount_init(BlockDriverState *bs)
{
    BDRVQcowState *s = bs->opaque;
    int ret, refcount_table_size2, i;
1734

B
bellard 已提交
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
    s->refcount_block_cache = qemu_malloc(s->cluster_size);
    if (!s->refcount_block_cache)
        goto fail;
    refcount_table_size2 = s->refcount_table_size * sizeof(uint64_t);
    s->refcount_table = qemu_malloc(refcount_table_size2);
    if (!s->refcount_table)
        goto fail;
    if (s->refcount_table_size > 0) {
        ret = bdrv_pread(s->hd, s->refcount_table_offset,
                         s->refcount_table, refcount_table_size2);
        if (ret != refcount_table_size2)
            goto fail;
        for(i = 0; i < s->refcount_table_size; i++)
            be64_to_cpus(&s->refcount_table[i]);
    }
    return 0;
 fail:
    return -ENOMEM;
}

static void refcount_close(BlockDriverState *bs)
{
    BDRVQcowState *s = bs->opaque;
    qemu_free(s->refcount_block_cache);
    qemu_free(s->refcount_table);
}


1763
static int load_refcount_block(BlockDriverState *bs,
B
bellard 已提交
1764 1765 1766 1767
                               int64_t refcount_block_offset)
{
    BDRVQcowState *s = bs->opaque;
    int ret;
1768
    ret = bdrv_pread(s->hd, refcount_block_offset, s->refcount_block_cache,
B
bellard 已提交
1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792
                     s->cluster_size);
    if (ret != s->cluster_size)
        return -EIO;
    s->refcount_block_cache_offset = refcount_block_offset;
    return 0;
}

static int get_refcount(BlockDriverState *bs, int64_t cluster_index)
{
    BDRVQcowState *s = bs->opaque;
    int refcount_table_index, block_index;
    int64_t refcount_block_offset;

    refcount_table_index = cluster_index >> (s->cluster_bits - REFCOUNT_SHIFT);
    if (refcount_table_index >= s->refcount_table_size)
        return 0;
    refcount_block_offset = s->refcount_table[refcount_table_index];
    if (!refcount_block_offset)
        return 0;
    if (refcount_block_offset != s->refcount_block_cache_offset) {
        /* better than nothing: return allocated if read error */
        if (load_refcount_block(bs, refcount_block_offset) < 0)
            return 1;
    }
1793
    block_index = cluster_index &
B
bellard 已提交
1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814
        ((1 << (s->cluster_bits - REFCOUNT_SHIFT)) - 1);
    return be16_to_cpu(s->refcount_block_cache[block_index]);
}

/* return < 0 if error */
static int64_t alloc_clusters_noref(BlockDriverState *bs, int64_t size)
{
    BDRVQcowState *s = bs->opaque;
    int i, nb_clusters;

    nb_clusters = (size + s->cluster_size - 1) >> s->cluster_bits;
    for(;;) {
        if (get_refcount(bs, s->free_cluster_index) == 0) {
            s->free_cluster_index++;
            for(i = 1; i < nb_clusters; i++) {
                if (get_refcount(bs, s->free_cluster_index) != 0)
                    goto not_found;
                s->free_cluster_index++;
            }
#ifdef DEBUG_ALLOC2
            printf("alloc_clusters: size=%lld -> %lld\n",
1815
                   size,
B
bellard 已提交
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
                   (s->free_cluster_index - nb_clusters) << s->cluster_bits);
#endif
            return (s->free_cluster_index - nb_clusters) << s->cluster_bits;
        } else {
        not_found:
            s->free_cluster_index++;
        }
    }
}

static int64_t alloc_clusters(BlockDriverState *bs, int64_t size)
{
    int64_t offset;

    offset = alloc_clusters_noref(bs, size);
    update_refcount(bs, offset, size, 1);
    return offset;
}

/* only used to allocate compressed sectors. We try to allocate
   contiguous sectors. size must be <= cluster_size */
static int64_t alloc_bytes(BlockDriverState *bs, int size)
{
    BDRVQcowState *s = bs->opaque;
    int64_t offset, cluster_offset;
    int free_in_cluster;
1842

B
bellard 已提交
1843 1844 1845 1846 1847
    assert(size > 0 && size <= s->cluster_size);
    if (s->free_byte_offset == 0) {
        s->free_byte_offset = alloc_clusters(bs, s->cluster_size);
    }
 redo:
1848
    free_in_cluster = s->cluster_size -
B
bellard 已提交
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
        (s->free_byte_offset & (s->cluster_size - 1));
    if (size <= free_in_cluster) {
        /* enough space in current cluster */
        offset = s->free_byte_offset;
        s->free_byte_offset += size;
        free_in_cluster -= size;
        if (free_in_cluster == 0)
            s->free_byte_offset = 0;
        if ((offset & (s->cluster_size - 1)) != 0)
            update_cluster_refcount(bs, offset >> s->cluster_bits, 1);
    } else {
        offset = alloc_clusters(bs, s->cluster_size);
        cluster_offset = s->free_byte_offset & ~(s->cluster_size - 1);
        if ((cluster_offset + s->cluster_size) == offset) {
            /* we are lucky: contiguous data */
            offset = s->free_byte_offset;
            update_cluster_refcount(bs, offset >> s->cluster_bits, 1);
            s->free_byte_offset += size;
        } else {
            s->free_byte_offset = offset;
            goto redo;
        }
    }
    return offset;
}

1875
static void free_clusters(BlockDriverState *bs,
B
bellard 已提交
1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888
                          int64_t offset, int64_t size)
{
    update_refcount(bs, offset, size, -1);
}

static int grow_refcount_table(BlockDriverState *bs, int min_size)
{
    BDRVQcowState *s = bs->opaque;
    int new_table_size, new_table_size2, refcount_table_clusters, i, ret;
    uint64_t *new_table;
    int64_t table_offset;
    uint64_t data64;
    uint32_t data32;
1889 1890
    int old_table_size;
    int64_t old_table_offset;
B
bellard 已提交
1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905

    if (min_size <= s->refcount_table_size)
        return 0;
    /* compute new table size */
    refcount_table_clusters = s->refcount_table_size >> (s->cluster_bits - 3);
    for(;;) {
        if (refcount_table_clusters == 0) {
            refcount_table_clusters = 1;
        } else {
            refcount_table_clusters = (refcount_table_clusters * 3 + 1) / 2;
        }
        new_table_size = refcount_table_clusters << (s->cluster_bits - 3);
        if (min_size <= new_table_size)
            break;
    }
B
bellard 已提交
1906 1907 1908 1909 1910
#ifdef DEBUG_ALLOC2
    printf("grow_refcount_table from %d to %d\n",
           s->refcount_table_size,
           new_table_size);
#endif
B
bellard 已提交
1911 1912 1913 1914
    new_table_size2 = new_table_size * sizeof(uint64_t);
    new_table = qemu_mallocz(new_table_size2);
    if (!new_table)
        return -ENOMEM;
1915
    memcpy(new_table, s->refcount_table,
B
bellard 已提交
1916 1917 1918 1919 1920 1921
           s->refcount_table_size * sizeof(uint64_t));
    for(i = 0; i < s->refcount_table_size; i++)
        cpu_to_be64s(&new_table[i]);
    /* Note: we cannot update the refcount now to avoid recursion */
    table_offset = alloc_clusters_noref(bs, new_table_size2);
    ret = bdrv_pwrite(s->hd, table_offset, new_table, new_table_size2);
1922
    if (ret != new_table_size2)
B
bellard 已提交
1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935
        goto fail;
    for(i = 0; i < s->refcount_table_size; i++)
        be64_to_cpus(&new_table[i]);

    data64 = cpu_to_be64(table_offset);
    if (bdrv_pwrite(s->hd, offsetof(QCowHeader, refcount_table_offset),
                    &data64, sizeof(data64)) != sizeof(data64))
        goto fail;
    data32 = cpu_to_be32(refcount_table_clusters);
    if (bdrv_pwrite(s->hd, offsetof(QCowHeader, refcount_table_clusters),
                    &data32, sizeof(data32)) != sizeof(data32))
        goto fail;
    qemu_free(s->refcount_table);
1936 1937
    old_table_offset = s->refcount_table_offset;
    old_table_size = s->refcount_table_size;
B
bellard 已提交
1938 1939
    s->refcount_table = new_table;
    s->refcount_table_size = new_table_size;
T
ths 已提交
1940
    s->refcount_table_offset = table_offset;
B
bellard 已提交
1941 1942

    update_refcount(bs, table_offset, new_table_size2, 1);
1943
    free_clusters(bs, old_table_offset, old_table_size * sizeof(uint64_t));
B
bellard 已提交
1944 1945 1946 1947 1948 1949 1950 1951 1952
    return 0;
 fail:
    free_clusters(bs, table_offset, new_table_size2);
    qemu_free(new_table);
    return -EIO;
}

/* addend must be 1 or -1 */
/* XXX: cache several refcount block clusters ? */
1953
static int update_cluster_refcount(BlockDriverState *bs,
B
bellard 已提交
1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982
                                   int64_t cluster_index,
                                   int addend)
{
    BDRVQcowState *s = bs->opaque;
    int64_t offset, refcount_block_offset;
    int ret, refcount_table_index, block_index, refcount;
    uint64_t data64;

    refcount_table_index = cluster_index >> (s->cluster_bits - REFCOUNT_SHIFT);
    if (refcount_table_index >= s->refcount_table_size) {
        if (addend < 0)
            return -EINVAL;
        ret = grow_refcount_table(bs, refcount_table_index + 1);
        if (ret < 0)
            return ret;
    }
    refcount_block_offset = s->refcount_table[refcount_table_index];
    if (!refcount_block_offset) {
        if (addend < 0)
            return -EINVAL;
        /* create a new refcount block */
        /* Note: we cannot update the refcount now to avoid recursion */
        offset = alloc_clusters_noref(bs, s->cluster_size);
        memset(s->refcount_block_cache, 0, s->cluster_size);
        ret = bdrv_pwrite(s->hd, offset, s->refcount_block_cache, s->cluster_size);
        if (ret != s->cluster_size)
            return -EINVAL;
        s->refcount_table[refcount_table_index] = offset;
        data64 = cpu_to_be64(offset);
1983 1984
        ret = bdrv_pwrite(s->hd, s->refcount_table_offset +
                          refcount_table_index * sizeof(uint64_t),
B
bellard 已提交
1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998
                          &data64, sizeof(data64));
        if (ret != sizeof(data64))
            return -EINVAL;

        refcount_block_offset = offset;
        s->refcount_block_cache_offset = offset;
        update_refcount(bs, offset, s->cluster_size, 1);
    } else {
        if (refcount_block_offset != s->refcount_block_cache_offset) {
            if (load_refcount_block(bs, refcount_block_offset) < 0)
                return -EIO;
        }
    }
    /* we can update the count and save it */
1999
    block_index = cluster_index &
B
bellard 已提交
2000 2001 2002 2003 2004 2005 2006 2007 2008
        ((1 << (s->cluster_bits - REFCOUNT_SHIFT)) - 1);
    refcount = be16_to_cpu(s->refcount_block_cache[block_index]);
    refcount += addend;
    if (refcount < 0 || refcount > 0xffff)
        return -EINVAL;
    if (refcount == 0 && cluster_index < s->free_cluster_index) {
        s->free_cluster_index = cluster_index;
    }
    s->refcount_block_cache[block_index] = cpu_to_be16(refcount);
2009 2010
    if (bdrv_pwrite(s->hd,
                    refcount_block_offset + (block_index << REFCOUNT_SHIFT),
B
bellard 已提交
2011 2012 2013 2014 2015
                    &s->refcount_block_cache[block_index], 2) != 2)
        return -EIO;
    return refcount;
}

2016 2017
static void update_refcount(BlockDriverState *bs,
                            int64_t offset, int64_t length,
B
bellard 已提交
2018 2019 2020 2021 2022 2023
                            int addend)
{
    BDRVQcowState *s = bs->opaque;
    int64_t start, last, cluster_offset;

#ifdef DEBUG_ALLOC2
2024
    printf("update_refcount: offset=%lld size=%lld addend=%d\n",
B
bellard 已提交
2025 2026 2027 2028 2029 2030
           offset, length, addend);
#endif
    if (length <= 0)
        return;
    start = offset & ~(s->cluster_size - 1);
    last = (offset + length - 1) & ~(s->cluster_size - 1);
2031
    for(cluster_offset = start; cluster_offset <= last;
B
bellard 已提交
2032 2033 2034 2035 2036 2037
        cluster_offset += s->cluster_size) {
        update_cluster_refcount(bs, cluster_offset >> s->cluster_bits, addend);
    }
}

#ifdef DEBUG_ALLOC
2038 2039
static void inc_refcounts(BlockDriverState *bs,
                          uint16_t *refcount_table,
B
bellard 已提交
2040 2041 2042 2043 2044 2045
                          int refcount_table_size,
                          int64_t offset, int64_t size)
{
    BDRVQcowState *s = bs->opaque;
    int64_t start, last, cluster_offset;
    int k;
2046

B
bellard 已提交
2047 2048 2049 2050 2051
    if (size <= 0)
        return;

    start = offset & ~(s->cluster_size - 1);
    last = (offset + size - 1) & ~(s->cluster_size - 1);
2052
    for(cluster_offset = start; cluster_offset <= last;
B
bellard 已提交
2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064
        cluster_offset += s->cluster_size) {
        k = cluster_offset >> s->cluster_bits;
        if (k < 0 || k >= refcount_table_size) {
            printf("ERROR: invalid cluster offset=0x%llx\n", cluster_offset);
        } else {
            if (++refcount_table[k] == 0) {
                printf("ERROR: overflow cluster offset=0x%llx\n", cluster_offset);
            }
        }
    }
}

2065 2066
static int check_refcounts_l1(BlockDriverState *bs,
                              uint16_t *refcount_table,
B
bellard 已提交
2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083
                              int refcount_table_size,
                              int64_t l1_table_offset, int l1_size,
                              int check_copied)
{
    BDRVQcowState *s = bs->opaque;
    uint64_t *l1_table, *l2_table, l2_offset, offset, l1_size2;
    int l2_size, i, j, nb_csectors, refcount;

    l2_table = NULL;
    l1_size2 = l1_size * sizeof(uint64_t);

    inc_refcounts(bs, refcount_table, refcount_table_size,
                  l1_table_offset, l1_size2);

    l1_table = qemu_malloc(l1_size2);
    if (!l1_table)
        goto fail;
2084
    if (bdrv_pread(s->hd, l1_table_offset,
B
bellard 已提交
2085 2086 2087 2088
                   l1_table, l1_size2) != l1_size2)
        goto fail;
    for(i = 0;i < l1_size; i++)
        be64_to_cpus(&l1_table[i]);
2089

B
bellard 已提交
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
    l2_size = s->l2_size * sizeof(uint64_t);
    l2_table = qemu_malloc(l2_size);
    if (!l2_table)
        goto fail;
    for(i = 0; i < l1_size; i++) {
        l2_offset = l1_table[i];
        if (l2_offset) {
            if (check_copied) {
                refcount = get_refcount(bs, (l2_offset & ~QCOW_OFLAG_COPIED) >> s->cluster_bits);
                if ((refcount == 1) != ((l2_offset & QCOW_OFLAG_COPIED) != 0)) {
                    printf("ERROR OFLAG_COPIED: l2_offset=%llx refcount=%d\n",
                           l2_offset, refcount);
                }
            }
            l2_offset &= ~QCOW_OFLAG_COPIED;
            if (bdrv_pread(s->hd, l2_offset, l2_table, l2_size) != l2_size)
                goto fail;
            for(j = 0; j < s->l2_size; j++) {
                offset = be64_to_cpu(l2_table[j]);
                if (offset != 0) {
                    if (offset & QCOW_OFLAG_COMPRESSED) {
                        if (offset & QCOW_OFLAG_COPIED) {
                            printf("ERROR: cluster %lld: copied flag must never be set for compressed clusters\n",
                                   offset >> s->cluster_bits);
                            offset &= ~QCOW_OFLAG_COPIED;
                        }
2116
                        nb_csectors = ((offset >> s->csize_shift) &
B
bellard 已提交
2117 2118
                                       s->csize_mask) + 1;
                        offset &= s->cluster_offset_mask;
2119
                        inc_refcounts(bs, refcount_table,
B
bellard 已提交
2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130
                                      refcount_table_size,
                                      offset & ~511, nb_csectors * 512);
                    } else {
                        if (check_copied) {
                            refcount = get_refcount(bs, (offset & ~QCOW_OFLAG_COPIED) >> s->cluster_bits);
                            if ((refcount == 1) != ((offset & QCOW_OFLAG_COPIED) != 0)) {
                                printf("ERROR OFLAG_COPIED: offset=%llx refcount=%d\n",
                                       offset, refcount);
                            }
                        }
                        offset &= ~QCOW_OFLAG_COPIED;
2131
                        inc_refcounts(bs, refcount_table,
B
bellard 已提交
2132 2133 2134 2135 2136
                                      refcount_table_size,
                                      offset, s->cluster_size);
                    }
                }
            }
2137
            inc_refcounts(bs, refcount_table,
B
bellard 已提交
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
                          refcount_table_size,
                          l2_offset,
                          s->cluster_size);
        }
    }
    qemu_free(l1_table);
    qemu_free(l2_table);
    return 0;
 fail:
    printf("ERROR: I/O error in check_refcounts_l1\n");
    qemu_free(l1_table);
    qemu_free(l2_table);
    return -EIO;
}

static void check_refcounts(BlockDriverState *bs)
{
    BDRVQcowState *s = bs->opaque;
    int64_t size;
    int nb_clusters, refcount1, refcount2, i;
    QCowSnapshot *sn;
    uint16_t *refcount_table;

    size = bdrv_getlength(s->hd);
    nb_clusters = (size + s->cluster_size - 1) >> s->cluster_bits;
    refcount_table = qemu_mallocz(nb_clusters * sizeof(uint16_t));
B
bellard 已提交
2164

B
bellard 已提交
2165 2166 2167
    /* header */
    inc_refcounts(bs, refcount_table, nb_clusters,
                  0, s->cluster_size);
2168

B
bellard 已提交
2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182
    check_refcounts_l1(bs, refcount_table, nb_clusters,
                       s->l1_table_offset, s->l1_size, 1);

    /* snapshots */
    for(i = 0; i < s->nb_snapshots; i++) {
        sn = s->snapshots + i;
        check_refcounts_l1(bs, refcount_table, nb_clusters,
                           sn->l1_table_offset, sn->l1_size, 0);
    }
    inc_refcounts(bs, refcount_table, nb_clusters,
                  s->snapshots_offset, s->snapshots_size);

    /* refcount data */
    inc_refcounts(bs, refcount_table, nb_clusters,
2183
                  s->refcount_table_offset,
B
bellard 已提交
2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243
                  s->refcount_table_size * sizeof(uint64_t));
    for(i = 0; i < s->refcount_table_size; i++) {
        int64_t offset;
        offset = s->refcount_table[i];
        if (offset != 0) {
            inc_refcounts(bs, refcount_table, nb_clusters,
                          offset, s->cluster_size);
        }
    }

    /* compare ref counts */
    for(i = 0; i < nb_clusters; i++) {
        refcount1 = get_refcount(bs, i);
        refcount2 = refcount_table[i];
        if (refcount1 != refcount2)
            printf("ERROR cluster %d refcount=%d reference=%d\n",
                   i, refcount1, refcount2);
    }

    qemu_free(refcount_table);
}

#if 0
static void dump_refcounts(BlockDriverState *bs)
{
    BDRVQcowState *s = bs->opaque;
    int64_t nb_clusters, k, k1, size;
    int refcount;

    size = bdrv_getlength(s->hd);
    nb_clusters = (size + s->cluster_size - 1) >> s->cluster_bits;
    for(k = 0; k < nb_clusters;) {
        k1 = k;
        refcount = get_refcount(bs, k);
        k++;
        while (k < nb_clusters && get_refcount(bs, k) == refcount)
            k++;
        printf("%lld: refcount=%d nb=%lld\n", k, refcount, k - k1);
    }
}
#endif
#endif

BlockDriver bdrv_qcow2 = {
    "qcow2",
    sizeof(BDRVQcowState),
    qcow_probe,
    qcow_open,
    NULL,
    NULL,
    qcow_close,
    qcow_create,
    qcow_flush,
    qcow_is_allocated,
    qcow_set_key,
    qcow_make_empty,

    .bdrv_aio_read = qcow_aio_read,
    .bdrv_aio_write = qcow_aio_write,
    .bdrv_aio_cancel = qcow_aio_cancel,
2244
    .aiocb_size = sizeof(QCowAIOCB),
B
bellard 已提交
2245 2246 2247 2248 2249 2250 2251 2252
    .bdrv_write_compressed = qcow_write_compressed,

    .bdrv_snapshot_create = qcow_snapshot_create,
    .bdrv_snapshot_goto = qcow_snapshot_goto,
    .bdrv_snapshot_delete = qcow_snapshot_delete,
    .bdrv_snapshot_list = qcow_snapshot_list,
    .bdrv_get_info = qcow_get_info,
};