vmdk.c 27.0 KB
Newer Older
B
bellard 已提交
1 2
/*
 * Block driver for the VMDK format
3
 *
B
bellard 已提交
4
 * Copyright (c) 2004 Fabrice Bellard
5
 * Copyright (c) 2005 Filip Navara
6
 *
B
bellard 已提交
7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
 * 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.
 */
25

P
pbrook 已提交
26
#include "qemu-common.h"
B
bellard 已提交
27
#include "block_int.h"
28
#include "module.h"
B
bellard 已提交
29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58

#define VMDK3_MAGIC (('C' << 24) | ('O' << 16) | ('W' << 8) | 'D')
#define VMDK4_MAGIC (('K' << 24) | ('D' << 16) | ('M' << 8) | 'V')

typedef struct {
    uint32_t version;
    uint32_t flags;
    uint32_t disk_sectors;
    uint32_t granularity;
    uint32_t l1dir_offset;
    uint32_t l1dir_size;
    uint32_t file_sectors;
    uint32_t cylinders;
    uint32_t heads;
    uint32_t sectors_per_track;
} VMDK3Header;

typedef struct {
    uint32_t version;
    uint32_t flags;
    int64_t capacity;
    int64_t granularity;
    int64_t desc_offset;
    int64_t desc_size;
    int32_t num_gtes_per_gte;
    int64_t rgd_offset;
    int64_t gd_offset;
    int64_t grain_offset;
    char filler[1];
    char check_bytes[4];
59
} __attribute__((packed)) VMDK4Header;
B
bellard 已提交
60 61 62 63

#define L2_CACHE_SIZE 16

typedef struct BDRVVmdkState {
64
    BlockDriverState *hd;
B
bellard 已提交
65
    int64_t l1_table_offset;
66
    int64_t l1_backup_table_offset;
B
bellard 已提交
67
    uint32_t *l1_table;
68
    uint32_t *l1_backup_table;
B
bellard 已提交
69 70 71 72 73 74 75 76 77
    unsigned int l1_size;
    uint32_t l1_entry_sectors;

    unsigned int l2_size;
    uint32_t *l2_cache;
    uint32_t l2_cache_offsets[L2_CACHE_SIZE];
    uint32_t l2_cache_counts[L2_CACHE_SIZE];

    unsigned int cluster_sectors;
78
    uint32_t parent_cid;
79
    int is_parent;
B
bellard 已提交
80 81
} BDRVVmdkState;

82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97
typedef struct VmdkMetaData {
    uint32_t offset;
    unsigned int l1_index;
    unsigned int l2_index;
    unsigned int l2_offset;
    int valid;
} VmdkMetaData;

typedef struct ActiveBDRVState{
    BlockDriverState *hd;            // active image handler
    uint64_t cluster_offset;         // current write offset
}ActiveBDRVState;

static ActiveBDRVState activeBDRV;


B
bellard 已提交
98 99 100 101 102 103 104 105 106 107 108 109 110 111
static int vmdk_probe(const uint8_t *buf, int buf_size, const char *filename)
{
    uint32_t magic;

    if (buf_size < 4)
        return 0;
    magic = be32_to_cpu(*(uint32_t *)buf);
    if (magic == VMDK3_MAGIC ||
        magic == VMDK4_MAGIC)
        return 100;
    else
        return 0;
}

112 113
#define CHECK_CID 1

114
#define SECTOR_SIZE 512
115
#define DESC_SIZE 20*SECTOR_SIZE	// 20 sectors of 512 bytes each
116
#define HEADER_SIZE 512   			// first sector of 512 bytes
117 118

static uint32_t vmdk_read_cid(BlockDriverState *bs, int parent)
B
bellard 已提交
119 120
{
    BDRVVmdkState *s = bs->opaque;
121 122
    char desc[DESC_SIZE];
    uint32_t cid;
123
    const char *p_name, *cid_str;
124 125 126 127 128 129 130 131 132 133 134 135 136 137
    size_t cid_str_size;

    /* the descriptor offset = 0x200 */
    if (bdrv_pread(s->hd, 0x200, desc, DESC_SIZE) != DESC_SIZE)
        return 0;

    if (parent) {
        cid_str = "parentCID";
        cid_str_size = sizeof("parentCID");
    } else {
        cid_str = "CID";
        cid_str_size = sizeof("CID");
    }

138
    if ((p_name = strstr(desc,cid_str)) != NULL) {
139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156
        p_name += cid_str_size;
        sscanf(p_name,"%x",&cid);
    }

    return cid;
}

static int vmdk_write_cid(BlockDriverState *bs, uint32_t cid)
{
    BDRVVmdkState *s = bs->opaque;
    char desc[DESC_SIZE], tmp_desc[DESC_SIZE];
    char *p_name, *tmp_str;

    /* the descriptor offset = 0x200 */
    if (bdrv_pread(s->hd, 0x200, desc, DESC_SIZE) != DESC_SIZE)
        return -1;

    tmp_str = strstr(desc,"parentCID");
B
blueswir1 已提交
157
    pstrcpy(tmp_desc, sizeof(tmp_desc), tmp_str);
158
    if ((p_name = strstr(desc,"CID")) != NULL) {
159
        p_name += sizeof("CID");
B
blueswir1 已提交
160 161
        snprintf(p_name, sizeof(desc) - (p_name - desc), "%x\n", cid);
        pstrcat(desc, sizeof(desc), tmp_desc);
162 163 164 165 166 167 168 169 170 171 172
    }

    if (bdrv_pwrite(s->hd, 0x200, desc, DESC_SIZE) != DESC_SIZE)
        return -1;
    return 0;
}

static int vmdk_is_cid_valid(BlockDriverState *bs)
{
#ifdef CHECK_CID
    BDRVVmdkState *s = bs->opaque;
K
Kevin Wolf 已提交
173
    BlockDriverState *p_bs = bs->backing_hd;
174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190
    uint32_t cur_pcid;

    if (p_bs) {
        cur_pcid = vmdk_read_cid(p_bs,0);
        if (s->parent_cid != cur_pcid)
            // CID not valid
            return 0;
    }
#endif
    // CID valid
    return 1;
}

static int vmdk_snapshot_create(const char *filename, const char *backing_file)
{
    int snp_fd, p_fd;
    uint32_t p_cid;
191
    char *p_name, *gd_buf, *rgd_buf;
192 193 194 195 196
    const char *real_filename, *temp_str;
    VMDK4Header header;
    uint32_t gde_entries, gd_size;
    int64_t gd_offset, rgd_offset, capacity, gt_size;
    char p_desc[DESC_SIZE], s_desc[DESC_SIZE], hdr[HEADER_SIZE];
197
    static const char desc_template[] =
198 199 200 201 202 203 204 205
    "# Disk DescriptorFile\n"
    "version=1\n"
    "CID=%x\n"
    "parentCID=%x\n"
    "createType=\"monolithicSparse\"\n"
    "parentFileNameHint=\"%s\"\n"
    "\n"
    "# Extent description\n"
206
    "RW %u SPARSE \"%s\"\n"
207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235
    "\n"
    "# The Disk Data Base \n"
    "#DDB\n"
    "\n";

    snp_fd = open(filename, O_RDWR | O_CREAT | O_TRUNC | O_BINARY | O_LARGEFILE, 0644);
    if (snp_fd < 0)
        return -1;
    p_fd = open(backing_file, O_RDONLY | O_BINARY | O_LARGEFILE);
    if (p_fd < 0) {
        close(snp_fd);
        return -1;
    }

    /* read the header */
    if (lseek(p_fd, 0x0, SEEK_SET) == -1)
        goto fail;
    if (read(p_fd, hdr, HEADER_SIZE) != HEADER_SIZE)
        goto fail;

    /* write the header */
    if (lseek(snp_fd, 0x0, SEEK_SET) == -1)
        goto fail;
    if (write(snp_fd, hdr, HEADER_SIZE) == -1)
        goto fail;

    memset(&header, 0, sizeof(header));
    memcpy(&header,&hdr[4], sizeof(header)); // skip the VMDK4_MAGIC

236 237
    if (ftruncate(snp_fd, header.grain_offset << 9))
        goto fail;
238 239 240 241 242 243
    /* the descriptor offset = 0x200 */
    if (lseek(p_fd, 0x200, SEEK_SET) == -1)
        goto fail;
    if (read(p_fd, p_desc, DESC_SIZE) != DESC_SIZE)
        goto fail;

244
    if ((p_name = strstr(p_desc,"CID")) != NULL) {
245 246 247 248 249 250 251 252 253 254 255 256
        p_name += sizeof("CID");
        sscanf(p_name,"%x",&p_cid);
    }

    real_filename = filename;
    if ((temp_str = strrchr(real_filename, '\\')) != NULL)
        real_filename = temp_str + 1;
    if ((temp_str = strrchr(real_filename, '/')) != NULL)
        real_filename = temp_str + 1;
    if ((temp_str = strrchr(real_filename, ':')) != NULL)
        real_filename = temp_str + 1;

B
blueswir1 已提交
257 258
    snprintf(s_desc, sizeof(s_desc), desc_template, p_cid, p_cid, backing_file,
             (uint32_t)header.capacity, real_filename);
259 260 261 262 263 264

    /* write the descriptor */
    if (lseek(snp_fd, 0x200, SEEK_SET) == -1)
        goto fail;
    if (write(snp_fd, s_desc, strlen(s_desc)) == -1)
        goto fail;
B
bellard 已提交
265

266 267 268 269 270 271 272 273 274 275
    gd_offset = header.gd_offset * SECTOR_SIZE;     // offset of GD table
    rgd_offset = header.rgd_offset * SECTOR_SIZE;   // offset of RGD table
    capacity = header.capacity * SECTOR_SIZE;       // Extent size
    /*
     * Each GDE span 32M disk, means:
     * 512 GTE per GT, each GTE points to grain
     */
    gt_size = (int64_t)header.num_gtes_per_gte * header.granularity * SECTOR_SIZE;
    if (!gt_size)
        goto fail;
276
    gde_entries = (uint32_t)(capacity / gt_size);  // number of gde/rgde
277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308
    gd_size = gde_entries * sizeof(uint32_t);

    /* write RGD */
    rgd_buf = qemu_malloc(gd_size);
    if (lseek(p_fd, rgd_offset, SEEK_SET) == -1)
        goto fail_rgd;
    if (read(p_fd, rgd_buf, gd_size) != gd_size)
        goto fail_rgd;
    if (lseek(snp_fd, rgd_offset, SEEK_SET) == -1)
        goto fail_rgd;
    if (write(snp_fd, rgd_buf, gd_size) == -1)
        goto fail_rgd;
    qemu_free(rgd_buf);

    /* write GD */
    gd_buf = qemu_malloc(gd_size);
    if (lseek(p_fd, gd_offset, SEEK_SET) == -1)
        goto fail_gd;
    if (read(p_fd, gd_buf, gd_size) != gd_size)
        goto fail_gd;
    if (lseek(snp_fd, gd_offset, SEEK_SET) == -1)
        goto fail_gd;
    if (write(snp_fd, gd_buf, gd_size) == -1)
        goto fail_gd;
    qemu_free(gd_buf);

    close(p_fd);
    close(snp_fd);
    return 0;

    fail_gd:
    qemu_free(gd_buf);
309
    fail_rgd:
310 311 312 313 314 315 316 317 318 319 320 321 322
    qemu_free(rgd_buf);
    fail:
    close(p_fd);
    close(snp_fd);
    return -1;
}

static void vmdk_parent_close(BlockDriverState *bs)
{
    if (bs->backing_hd)
        bdrv_close(bs->backing_hd);
}

323
static int parent_open = 0;
324 325 326
static int vmdk_parent_open(BlockDriverState *bs, const char * filename)
{
    BDRVVmdkState *s = bs->opaque;
327
    char *p_name;
328 329 330 331 332 333 334
    char desc[DESC_SIZE];
    char parent_img_name[1024];

    /* the descriptor offset = 0x200 */
    if (bdrv_pread(s->hd, 0x200, desc, DESC_SIZE) != DESC_SIZE)
        return -1;

335
    if ((p_name = strstr(desc,"parentFileNameHint")) != NULL) {
336 337 338 339
        char *end_name;
        struct stat file_buf;

        p_name += sizeof("parentFileNameHint") + 1;
340
        if ((end_name = strchr(p_name,'\"')) == NULL)
341
            return -1;
K
Kevin Wolf 已提交
342
        if ((end_name - p_name) > sizeof (bs->backing_file) - 1)
343
            return -1;
344

K
Kevin Wolf 已提交
345 346
        pstrcpy(bs->backing_file, end_name - p_name + 1, p_name);
        if (stat(bs->backing_file, &file_buf) != 0) {
347
            path_combine(parent_img_name, sizeof(parent_img_name),
K
Kevin Wolf 已提交
348
                         filename, bs->backing_file);
349
        } else {
B
blueswir1 已提交
350
            pstrcpy(parent_img_name, sizeof(parent_img_name),
K
Kevin Wolf 已提交
351
                    bs->backing_file);
352 353
        }

K
Kevin Wolf 已提交
354 355
        bs->backing_hd = bdrv_new("");
        if (!bs->backing_hd) {
356 357
            failure:
            bdrv_close(s->hd);
358
            return -1;
359
        }
360
        parent_open = 1;
361
        if (bdrv_open(bs->backing_hd, parent_img_name, 0) < 0)
362
            goto failure;
363
        parent_open = 0;
364
    }
365 366 367 368 369 370 371 372 373 374

    return 0;
}

static int vmdk_open(BlockDriverState *bs, const char *filename, int flags)
{
    BDRVVmdkState *s = bs->opaque;
    uint32_t magic;
    int l1_size, i, ret;

375 376 377 378
    if (parent_open) {
        /* Parent must be opened as RO, no RDWR. */
        flags = 0;
    }
379

380
    ret = bdrv_file_open(&s->hd, filename, flags);
381 382 383
    if (ret < 0)
        return ret;
    if (bdrv_pread(s->hd, 0, &magic, sizeof(magic)) != sizeof(magic))
B
bellard 已提交
384
        goto fail;
385

B
bellard 已提交
386
    magic = be32_to_cpu(magic);
B
bellard 已提交
387 388
    if (magic == VMDK3_MAGIC) {
        VMDK3Header header;
389 390

        if (bdrv_pread(s->hd, sizeof(magic), &header, sizeof(header)) != sizeof(header))
B
bellard 已提交
391 392 393 394 395
            goto fail;
        s->cluster_sectors = le32_to_cpu(header.granularity);
        s->l2_size = 1 << 9;
        s->l1_size = 1 << 6;
        bs->total_sectors = le32_to_cpu(header.disk_sectors);
396 397
        s->l1_table_offset = le32_to_cpu(header.l1dir_offset) << 9;
        s->l1_backup_table_offset = 0;
B
bellard 已提交
398 399 400
        s->l1_entry_sectors = s->l2_size * s->cluster_sectors;
    } else if (magic == VMDK4_MAGIC) {
        VMDK4Header header;
401 402

        if (bdrv_pread(s->hd, sizeof(magic), &header, sizeof(header)) != sizeof(header))
B
bellard 已提交
403
            goto fail;
404 405
        bs->total_sectors = le64_to_cpu(header.capacity);
        s->cluster_sectors = le64_to_cpu(header.granularity);
B
bellard 已提交
406 407 408 409
        s->l2_size = le32_to_cpu(header.num_gtes_per_gte);
        s->l1_entry_sectors = s->l2_size * s->cluster_sectors;
        if (s->l1_entry_sectors <= 0)
            goto fail;
410
        s->l1_size = (bs->total_sectors + s->l1_entry_sectors - 1)
B
bellard 已提交
411
            / s->l1_entry_sectors;
412 413
        s->l1_table_offset = le64_to_cpu(header.rgd_offset) << 9;
        s->l1_backup_table_offset = le64_to_cpu(header.gd_offset) << 9;
414

415 416 417 418 419
        if (parent_open)
            s->is_parent = 1;
        else
            s->is_parent = 0;

420 421 422 423 424
        // try to open parent images, if exist
        if (vmdk_parent_open(bs, filename) != 0)
            goto fail;
        // write the CID once after the image creation
        s->parent_cid = vmdk_read_cid(bs,1);
B
bellard 已提交
425 426 427
    } else {
        goto fail;
    }
428

B
bellard 已提交
429 430 431
    /* read the L1 table */
    l1_size = s->l1_size * sizeof(uint32_t);
    s->l1_table = qemu_malloc(l1_size);
432
    if (bdrv_pread(s->hd, s->l1_table_offset, s->l1_table, l1_size) != l1_size)
B
bellard 已提交
433 434 435 436 437
        goto fail;
    for(i = 0; i < s->l1_size; i++) {
        le32_to_cpus(&s->l1_table[i]);
    }

438 439
    if (s->l1_backup_table_offset) {
        s->l1_backup_table = qemu_malloc(l1_size);
440
        if (bdrv_pread(s->hd, s->l1_backup_table_offset, s->l1_backup_table, l1_size) != l1_size)
441 442 443 444 445 446
            goto fail;
        for(i = 0; i < s->l1_size; i++) {
            le32_to_cpus(&s->l1_backup_table[i]);
        }
    }

B
bellard 已提交
447 448 449
    s->l2_cache = qemu_malloc(s->l2_size * L2_CACHE_SIZE * sizeof(uint32_t));
    return 0;
 fail:
450
    qemu_free(s->l1_backup_table);
B
bellard 已提交
451 452
    qemu_free(s->l1_table);
    qemu_free(s->l2_cache);
453
    bdrv_delete(s->hd);
B
bellard 已提交
454 455 456
    return -1;
}

457 458
static uint64_t get_cluster_offset(BlockDriverState *bs, VmdkMetaData *m_data,
                                   uint64_t offset, int allocate);
459 460 461 462 463 464 465 466 467 468

static int get_whole_cluster(BlockDriverState *bs, uint64_t cluster_offset,
                             uint64_t offset, int allocate)
{
    uint64_t parent_cluster_offset;
    BDRVVmdkState *s = bs->opaque;
    uint8_t  whole_grain[s->cluster_sectors*512];        // 128 sectors * 512 bytes each = grain size 64KB

    // we will be here if it's first write on non-exist grain(cluster).
    // try to read from parent image, if exist
K
Kevin Wolf 已提交
469 470
    if (bs->backing_hd) {
        BDRVVmdkState *ps = bs->backing_hd->opaque;
471 472 473 474

        if (!vmdk_is_cid_valid(bs))
            return -1;

K
Kevin Wolf 已提交
475 476
        parent_cluster_offset = get_cluster_offset(bs->backing_hd, NULL,
            offset, allocate);
477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504

        if (parent_cluster_offset) {
            BDRVVmdkState *act_s = activeBDRV.hd->opaque;

            if (bdrv_pread(ps->hd, parent_cluster_offset, whole_grain, ps->cluster_sectors*512) != ps->cluster_sectors*512)
                return -1;

            //Write grain only into the active image
            if (bdrv_pwrite(act_s->hd, activeBDRV.cluster_offset << 9, whole_grain, sizeof(whole_grain)) != sizeof(whole_grain))
                return -1;
        }
    }
    return 0;
}

static int vmdk_L2update(BlockDriverState *bs, VmdkMetaData *m_data)
{
    BDRVVmdkState *s = bs->opaque;

    /* update L2 table */
    if (bdrv_pwrite(s->hd, ((int64_t)m_data->l2_offset * 512) + (m_data->l2_index * sizeof(m_data->offset)),
                    &(m_data->offset), sizeof(m_data->offset)) != sizeof(m_data->offset))
        return -1;
    /* update backup L2 table */
    if (s->l1_backup_table_offset != 0) {
        m_data->l2_offset = s->l1_backup_table[m_data->l1_index];
        if (bdrv_pwrite(s->hd, ((int64_t)m_data->l2_offset * 512) + (m_data->l2_index * sizeof(m_data->offset)),
                        &(m_data->offset), sizeof(m_data->offset)) != sizeof(m_data->offset))
505 506
            return -1;
    }
507

508 509 510
    return 0;
}

511
static uint64_t get_cluster_offset(BlockDriverState *bs, VmdkMetaData *m_data,
512
                                   uint64_t offset, int allocate)
B
bellard 已提交
513 514 515 516
{
    BDRVVmdkState *s = bs->opaque;
    unsigned int l1_index, l2_offset, l2_index;
    int min_index, i, j;
517
    uint32_t min_count, *l2_table, tmp = 0;
B
bellard 已提交
518
    uint64_t cluster_offset;
519 520 521 522

    if (m_data)
        m_data->valid = 0;

B
bellard 已提交
523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550
    l1_index = (offset >> 9) / s->l1_entry_sectors;
    if (l1_index >= s->l1_size)
        return 0;
    l2_offset = s->l1_table[l1_index];
    if (!l2_offset)
        return 0;
    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_size);
            goto found;
        }
    }
    /* not found: load 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;
        }
    }
    l2_table = s->l2_cache + (min_index * s->l2_size);
551
    if (bdrv_pread(s->hd, (int64_t)l2_offset * 512, l2_table, s->l2_size * sizeof(uint32_t)) !=
552
                                                                        s->l2_size * sizeof(uint32_t))
B
bellard 已提交
553
        return 0;
554

B
bellard 已提交
555 556 557 558 559
    s->l2_cache_offsets[min_index] = l2_offset;
    s->l2_cache_counts[min_index] = 1;
 found:
    l2_index = ((offset >> 9) / s->cluster_sectors) % s->l2_size;
    cluster_offset = le32_to_cpu(l2_table[l2_index]);
560

561 562 563
    if (!cluster_offset) {
        if (!allocate)
            return 0;
564 565
        // Avoid the L2 tables update for the images that have snapshots.
        if (!s->is_parent) {
T
ths 已提交
566
            cluster_offset = bdrv_getlength(s->hd);
567 568 569 570 571 572 573 574
            bdrv_truncate(s->hd, cluster_offset + (s->cluster_sectors << 9));

            cluster_offset >>= 9;
            tmp = cpu_to_le32(cluster_offset);
            l2_table[l2_index] = tmp;
            // Save the active image state
            activeBDRV.cluster_offset = cluster_offset;
            activeBDRV.hd = bs;
575
        }
576 577 578 579 580
        /* First of all we write grain itself, to avoid race condition
         * that may to corrupt the image.
         * This problem may occur because of insufficient space on host disk
         * or inappropriate VM shutdown.
         */
581 582
        if (get_whole_cluster(bs, cluster_offset, offset, allocate) == -1)
            return 0;
583 584 585 586 587 588 589 590

        if (m_data) {
            m_data->offset = tmp;
            m_data->l1_index = l1_index;
            m_data->l2_index = l2_index;
            m_data->l2_offset = l2_offset;
            m_data->valid = 1;
        }
591
    }
B
bellard 已提交
592 593 594 595
    cluster_offset <<= 9;
    return cluster_offset;
}

596
static int vmdk_is_allocated(BlockDriverState *bs, int64_t sector_num,
B
bellard 已提交
597 598 599 600 601 602
                             int nb_sectors, int *pnum)
{
    BDRVVmdkState *s = bs->opaque;
    int index_in_cluster, n;
    uint64_t cluster_offset;

603
    cluster_offset = get_cluster_offset(bs, NULL, sector_num << 9, 0);
B
bellard 已提交
604 605 606 607 608 609 610 611
    index_in_cluster = sector_num % s->cluster_sectors;
    n = s->cluster_sectors - index_in_cluster;
    if (n > nb_sectors)
        n = nb_sectors;
    *pnum = n;
    return (cluster_offset != 0);
}

612
static int vmdk_read(BlockDriverState *bs, int64_t sector_num,
B
bellard 已提交
613 614 615
                    uint8_t *buf, int nb_sectors)
{
    BDRVVmdkState *s = bs->opaque;
616
    int index_in_cluster, n, ret;
B
bellard 已提交
617
    uint64_t cluster_offset;
618

B
bellard 已提交
619
    while (nb_sectors > 0) {
620
        cluster_offset = get_cluster_offset(bs, NULL, sector_num << 9, 0);
B
bellard 已提交
621 622 623 624 625
        index_in_cluster = sector_num % s->cluster_sectors;
        n = s->cluster_sectors - index_in_cluster;
        if (n > nb_sectors)
            n = nb_sectors;
        if (!cluster_offset) {
626
            // try to read from parent image, if exist
K
Kevin Wolf 已提交
627
            if (bs->backing_hd) {
628 629
                if (!vmdk_is_cid_valid(bs))
                    return -1;
K
Kevin Wolf 已提交
630
                ret = bdrv_read(bs->backing_hd, sector_num, buf, n);
631 632 633 634 635
                if (ret < 0)
                    return -1;
            } else {
                memset(buf, 0, 512 * n);
            }
B
bellard 已提交
636
        } else {
637
            if(bdrv_pread(s->hd, cluster_offset + index_in_cluster * 512, buf, n * 512) != n * 512)
B
bellard 已提交
638 639 640 641 642 643 644 645 646
                return -1;
        }
        nb_sectors -= n;
        sector_num += n;
        buf += n * 512;
    }
    return 0;
}

647
static int vmdk_write(BlockDriverState *bs, int64_t sector_num,
B
bellard 已提交
648 649
                     const uint8_t *buf, int nb_sectors)
{
650
    BDRVVmdkState *s = bs->opaque;
651
    VmdkMetaData m_data;
652
    int index_in_cluster, n;
653
    uint64_t cluster_offset;
654
    static int cid_update = 0;
655

656 657
    if (sector_num > bs->total_sectors) {
        fprintf(stderr,
658 659
                "(VMDK) Wrong offset: sector_num=0x%" PRIx64
                " total_sectors=0x%" PRIx64 "\n",
660 661 662 663
                sector_num, bs->total_sectors);
        return -1;
    }

664 665 666 667 668
    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;
669
        cluster_offset = get_cluster_offset(bs, &m_data, sector_num << 9, 1);
670 671
        if (!cluster_offset)
            return -1;
672

673
        if (bdrv_pwrite(s->hd, cluster_offset + index_in_cluster * 512, buf, n * 512) != n * 512)
674
            return -1;
675 676 677 678 679
        if (m_data.valid) {
            /* update L2 tables */
            if (vmdk_L2update(bs, &m_data) == -1)
                return -1;
        }
680 681 682
        nb_sectors -= n;
        sector_num += n;
        buf += n * 512;
683 684 685 686 687 688

        // update CID on the first write every time the virtual disk is opened
        if (!cid_update) {
            vmdk_write_cid(bs, time(NULL));
            cid_update++;
        }
689 690
    }
    return 0;
B
bellard 已提交
691 692
}

693
static int vmdk_create(const char *filename, QEMUOptionParameter *options)
694 695 696 697
{
    int fd, i;
    VMDK4Header header;
    uint32_t tmp, magic, grains, gd_size, gt_size, gt_count;
698
    static const char desc_template[] =
699 700 701 702 703 704 705
        "# Disk DescriptorFile\n"
        "version=1\n"
        "CID=%x\n"
        "parentCID=ffffffff\n"
        "createType=\"monolithicSparse\"\n"
        "\n"
        "# Extent description\n"
706
        "RW %" PRId64 " SPARSE \"%s\"\n"
707 708 709 710
        "\n"
        "# The Disk Data Base \n"
        "#DDB\n"
        "\n"
711
        "ddb.virtualHWVersion = \"%d\"\n"
712
        "ddb.geometry.cylinders = \"%" PRId64 "\"\n"
713 714 715 716 717
        "ddb.geometry.heads = \"16\"\n"
        "ddb.geometry.sectors = \"63\"\n"
        "ddb.adapterType = \"ide\"\n";
    char desc[1024];
    const char *real_filename, *temp_str;
718 719 720
    int64_t total_size = 0;
    const char *backing_file = NULL;
    int flags = 0;
721
    int ret;
722 723 724 725 726 727 728 729 730 731 732 733

    // Read out options
    while (options && options->name) {
        if (!strcmp(options->name, BLOCK_OPT_SIZE)) {
            total_size = 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_COMPAT6)) {
            flags |= options->value.n ? BLOCK_FLAG_COMPAT6: 0;
        }
        options++;
    }
734 735

    /* XXX: add support for backing file */
736 737 738
    if (backing_file) {
        return vmdk_snapshot_create(filename, backing_file);
    }
739 740 741 742

    fd = open(filename, O_WRONLY | O_CREAT | O_TRUNC | O_BINARY | O_LARGEFILE,
              0644);
    if (fd < 0)
J
Juan Quintela 已提交
743
        return -errno;
744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775
    magic = cpu_to_be32(VMDK4_MAGIC);
    memset(&header, 0, sizeof(header));
    header.version = cpu_to_le32(1);
    header.flags = cpu_to_le32(3); /* ?? */
    header.capacity = cpu_to_le64(total_size);
    header.granularity = cpu_to_le64(128);
    header.num_gtes_per_gte = cpu_to_le32(512);

    grains = (total_size + header.granularity - 1) / header.granularity;
    gt_size = ((header.num_gtes_per_gte * sizeof(uint32_t)) + 511) >> 9;
    gt_count = (grains + header.num_gtes_per_gte - 1) / header.num_gtes_per_gte;
    gd_size = (gt_count * sizeof(uint32_t) + 511) >> 9;

    header.desc_offset = 1;
    header.desc_size = 20;
    header.rgd_offset = header.desc_offset + header.desc_size;
    header.gd_offset = header.rgd_offset + gd_size + (gt_size * gt_count);
    header.grain_offset =
       ((header.gd_offset + gd_size + (gt_size * gt_count) +
         header.granularity - 1) / header.granularity) *
        header.granularity;

    header.desc_offset = cpu_to_le64(header.desc_offset);
    header.desc_size = cpu_to_le64(header.desc_size);
    header.rgd_offset = cpu_to_le64(header.rgd_offset);
    header.gd_offset = cpu_to_le64(header.gd_offset);
    header.grain_offset = cpu_to_le64(header.grain_offset);

    header.check_bytes[0] = 0xa;
    header.check_bytes[1] = 0x20;
    header.check_bytes[2] = 0xd;
    header.check_bytes[3] = 0xa;
776 777

    /* write all the data */
778 779
    ret = qemu_write_full(fd, &magic, sizeof(magic));
    if (ret != sizeof(magic)) {
J
Juan Quintela 已提交
780
        ret = -errno;
781 782 783 784
        goto exit;
    }
    ret = qemu_write_full(fd, &header, sizeof(header));
    if (ret != sizeof(header)) {
J
Juan Quintela 已提交
785
        ret = -errno;
786 787
        goto exit;
    }
788

789 790
    ret = ftruncate(fd, header.grain_offset << 9);
    if (ret < 0) {
J
Juan Quintela 已提交
791
        ret = -errno;
792 793
        goto exit;
    }
794 795 796 797

    /* write grain directory */
    lseek(fd, le64_to_cpu(header.rgd_offset) << 9, SEEK_SET);
    for (i = 0, tmp = header.rgd_offset + gd_size;
798 799 800
         i < gt_count; i++, tmp += gt_size) {
        ret = qemu_write_full(fd, &tmp, sizeof(tmp));
        if (ret != sizeof(tmp)) {
J
Juan Quintela 已提交
801
            ret = -errno;
802 803 804
            goto exit;
        }
    }
805

806 807 808
    /* write backup grain directory */
    lseek(fd, le64_to_cpu(header.gd_offset) << 9, SEEK_SET);
    for (i = 0, tmp = header.gd_offset + gd_size;
809 810 811
         i < gt_count; i++, tmp += gt_size) {
        ret = qemu_write_full(fd, &tmp, sizeof(tmp));
        if (ret != sizeof(tmp)) {
J
Juan Quintela 已提交
812
            ret = -errno;
813 814 815
            goto exit;
        }
    }
816 817 818 819 820 821 822 823 824

    /* compose the descriptor */
    real_filename = filename;
    if ((temp_str = strrchr(real_filename, '\\')) != NULL)
        real_filename = temp_str + 1;
    if ((temp_str = strrchr(real_filename, '/')) != NULL)
        real_filename = temp_str + 1;
    if ((temp_str = strrchr(real_filename, ':')) != NULL)
        real_filename = temp_str + 1;
825
    snprintf(desc, sizeof(desc), desc_template, (unsigned int)time(NULL),
826 827 828
             total_size, real_filename,
             (flags & BLOCK_FLAG_COMPAT6 ? 6 : 4),
             total_size / (int64_t)(63 * 16));
829 830 831

    /* write the descriptor */
    lseek(fd, le64_to_cpu(header.desc_offset) << 9, SEEK_SET);
832 833
    ret = qemu_write_full(fd, desc, strlen(desc));
    if (ret != strlen(desc)) {
J
Juan Quintela 已提交
834
        ret = -errno;
835 836
        goto exit;
    }
837

838 839
    ret = 0;
exit:
840
    close(fd);
841
    return ret;
842 843
}

B
bellard 已提交
844
static void vmdk_close(BlockDriverState *bs)
B
bellard 已提交
845 846
{
    BDRVVmdkState *s = bs->opaque;
847

B
bellard 已提交
848 849
    qemu_free(s->l1_table);
    qemu_free(s->l2_cache);
850 851
    // try to close parent image, if exist
    vmdk_parent_close(s->hd);
852
    bdrv_delete(s->hd);
B
bellard 已提交
853 854
}

P
pbrook 已提交
855 856 857
static void vmdk_flush(BlockDriverState *bs)
{
    BDRVVmdkState *s = bs->opaque;
858
    bdrv_flush(s->hd);
P
pbrook 已提交
859 860
}

861 862

static QEMUOptionParameter vmdk_create_options[] = {
863 864 865 866 867 868 869 870 871 872 873 874 875 876 877
    {
        .name = BLOCK_OPT_SIZE,
        .type = OPT_SIZE,
        .help = "Virtual disk size"
    },
    {
        .name = BLOCK_OPT_BACKING_FILE,
        .type = OPT_STRING,
        .help = "File name of a base image"
    },
    {
        .name = BLOCK_OPT_COMPAT6,
        .type = OPT_FLAG,
        .help = "VMDK version 6 image"
    },
878 879 880
    { NULL }
};

881
static BlockDriver bdrv_vmdk = {
882 883 884 885 886 887 888 889 890 891
    .format_name	= "vmdk",
    .instance_size	= sizeof(BDRVVmdkState),
    .bdrv_probe		= vmdk_probe,
    .bdrv_open		= vmdk_open,
    .bdrv_read		= vmdk_read,
    .bdrv_write		= vmdk_write,
    .bdrv_close		= vmdk_close,
    .bdrv_create	= vmdk_create,
    .bdrv_flush		= vmdk_flush,
    .bdrv_is_allocated	= vmdk_is_allocated,
892 893

    .create_options = vmdk_create_options,
B
bellard 已提交
894
};
895 896 897 898 899 900 901

static void bdrv_vmdk_init(void)
{
    bdrv_register(&bdrv_vmdk);
}

block_init(bdrv_vmdk_init);