vpc.c 24.4 KB
Newer Older
1
/*
S
Stefan Weil 已提交
2
 * Block driver for Connectix / Microsoft Virtual PC images
3
 *
4
 * Copyright (c) 2005 Alex Beregszaszi
A
aliguori 已提交
5
 * Copyright (c) 2009 Kevin Wolf <kwolf@suse.de>
6
 *
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.
 */
P
pbrook 已提交
25
#include "qemu-common.h"
26
#include "block/block_int.h"
27
#include "qemu/module.h"
28
#include "migration/migration.h"
29 30 31
#if defined(CONFIG_UUID)
#include <uuid/uuid.h>
#endif
32 33 34 35 36 37 38

/**************************************************************/

#define HEADER_SIZE 512

//#define CACHE

39 40 41 42 43 44
enum vhd_type {
    VHD_FIXED           = 2,
    VHD_DYNAMIC         = 3,
    VHD_DIFFERENCING    = 4,
};

A
aliguori 已提交
45 46 47
// Seconds since Jan 1, 2000 0:00:00 (UTC)
#define VHD_TIMESTAMP_BASE 946684800

48
// always big-endian
49
typedef struct vhd_footer {
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
    char        creator[8]; // "conectix"
    uint32_t    features;
    uint32_t    version;

    // Offset of next header structure, 0xFFFFFFFF if none
    uint64_t    data_offset;

    // Seconds since Jan 1, 2000 0:00:00 (UTC)
    uint32_t    timestamp;

    char        creator_app[4]; // "vpc "
    uint16_t    major;
    uint16_t    minor;
    char        creator_os[4]; // "Wi2k"

    uint64_t    orig_size;
    uint64_t    size;

    uint16_t    cyls;
    uint8_t     heads;
    uint8_t     secs_per_cyl;

    uint32_t    type;

    // Checksum of the Hard Disk Footer ("one's complement of the sum of all
    // the bytes in the footer without the checksum field")
    uint32_t    checksum;

    // UUID used to identify a parent hard disk (backing file)
    uint8_t     uuid[16];

    uint8_t     in_saved_state;
82
} QEMU_PACKED VHDFooter;
83

84
typedef struct vhd_dyndisk_header {
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
    char        magic[8]; // "cxsparse"

    // Offset of next header structure, 0xFFFFFFFF if none
    uint64_t    data_offset;

    // Offset of the Block Allocation Table (BAT)
    uint64_t    table_offset;

    uint32_t    version;
    uint32_t    max_table_entries; // 32bit/entry

    // 2 MB by default, must be a power of two
    uint32_t    block_size;

    uint32_t    checksum;
    uint8_t     parent_uuid[16];
    uint32_t    parent_timestamp;
    uint32_t    reserved;

    // Backing file name (in UTF-16)
    uint8_t     parent_name[512];

    struct {
        uint32_t    platform;
        uint32_t    data_space;
        uint32_t    data_length;
        uint32_t    reserved;
        uint64_t    data_offset;
    } parent_locator[8];
114
} QEMU_PACKED VHDDynDiskHeader;
115 116

typedef struct BDRVVPCState {
117
    CoMutex lock;
A
aliguori 已提交
118 119
    uint8_t footer_buf[HEADER_SIZE];
    uint64_t free_data_block_offset;
120
    int max_table_entries;
121
    uint32_t *pagetable;
A
aliguori 已提交
122 123
    uint64_t bat_offset;
    uint64_t last_bitmap_offset;
124

125
    uint32_t block_size;
A
aliguori 已提交
126 127
    uint32_t bitmap_size;

128 129 130 131
#ifdef CACHE
    uint8_t *pageentry_u8;
    uint32_t *pageentry_u32;
    uint16_t *pageentry_u16;
132

133 134
    uint64_t last_bitmap;
#endif
K
Kevin Wolf 已提交
135 136

    Error *migration_blocker;
137 138
} BDRVVPCState;

A
aliguori 已提交
139 140 141 142 143 144 145 146 147 148 149 150
static uint32_t vpc_checksum(uint8_t* buf, size_t size)
{
    uint32_t res = 0;
    int i;

    for (i = 0; i < size; i++)
        res += buf[i];

    return ~res;
}


151 152
static int vpc_probe(const uint8_t *buf, int buf_size, const char *filename)
{
T
ths 已提交
153
    if (buf_size >= 8 && !strncmp((char *)buf, "conectix", 8))
154 155 156 157
	return 100;
    return 0;
}

M
Max Reitz 已提交
158 159
static int vpc_open(BlockDriverState *bs, QDict *options, int flags,
                    Error **errp)
160 161
{
    BDRVVPCState *s = bs->opaque;
162
    int i;
163 164
    VHDFooter *footer;
    VHDDynDiskHeader *dyndisk_header;
165
    uint8_t buf[HEADER_SIZE];
A
aliguori 已提交
166
    uint32_t checksum;
167
    int disk_type = VHD_DYNAMIC;
K
Kevin Wolf 已提交
168
    int ret;
169

K
Kevin Wolf 已提交
170 171
    ret = bdrv_pread(bs->file, 0, s->footer_buf, HEADER_SIZE);
    if (ret < 0) {
172
        goto fail;
K
Kevin Wolf 已提交
173
    }
174

175
    footer = (VHDFooter *) s->footer_buf;
176 177
    if (strncmp(footer->creator, "conectix", 8)) {
        int64_t offset = bdrv_getlength(bs->file);
K
Kevin Wolf 已提交
178 179 180 181 182
        if (offset < 0) {
            ret = offset;
            goto fail;
        } else if (offset < HEADER_SIZE) {
            ret = -EINVAL;
183 184
            goto fail;
        }
K
Kevin Wolf 已提交
185

186
        /* If a fixed disk, the footer is found only at the end of the file */
K
Kevin Wolf 已提交
187 188 189
        ret = bdrv_pread(bs->file, offset-HEADER_SIZE, s->footer_buf,
                         HEADER_SIZE);
        if (ret < 0) {
190 191 192
            goto fail;
        }
        if (strncmp(footer->creator, "conectix", 8)) {
K
Kevin Wolf 已提交
193
            ret = -EMEDIUMTYPE;
194 195 196 197
            goto fail;
        }
        disk_type = VHD_FIXED;
    }
198

A
aliguori 已提交
199 200 201 202
    checksum = be32_to_cpu(footer->checksum);
    footer->checksum = 0;
    if (vpc_checksum(s->footer_buf, HEADER_SIZE) != checksum)
        fprintf(stderr, "block-vpc: The header checksum of '%s' is "
203
            "incorrect.\n", bs->filename);
A
aliguori 已提交
204

205 206 207
    /* Write 'checksum' back to footer, or else will leave it with zero. */
    footer->checksum = be32_to_cpu(checksum);

208 209 210 211 212
    // The visible size of a image in Virtual PC depends on the geometry
    // rather than on the size stored in the footer (the size in the footer
    // is too large usually)
    bs->total_sectors = (int64_t)
        be16_to_cpu(footer->cyls) * footer->heads * footer->secs_per_cyl;
A
aliguori 已提交
213

214 215 216 217 218 219 220 221 222
    /* images created with disk2vhd report a far higher virtual size
     * than expected with the cyls * heads * sectors_per_cyl formula.
     * use the footer->size instead if the image was created with
     * disk2vhd.
     */
    if (!strncmp(footer->creator_app, "d2v", 4)) {
        bs->total_sectors = be64_to_cpu(footer->size) / BDRV_SECTOR_SIZE;
    }

223 224
    /* Allow a maximum disk size of approximately 2 TB */
    if (bs->total_sectors >= 65535LL * 255 * 255) {
K
Kevin Wolf 已提交
225
        ret = -EFBIG;
226 227 228
        goto fail;
    }

229
    if (disk_type == VHD_DYNAMIC) {
K
Kevin Wolf 已提交
230 231 232
        ret = bdrv_pread(bs->file, be64_to_cpu(footer->data_offset), buf,
                         HEADER_SIZE);
        if (ret < 0) {
233 234
            goto fail;
        }
235

236
        dyndisk_header = (VHDDynDiskHeader *) buf;
237

238
        if (strncmp(dyndisk_header->magic, "cxsparse", 8)) {
K
Kevin Wolf 已提交
239
            ret = -EINVAL;
240 241
            goto fail;
        }
242

243 244
        s->block_size = be32_to_cpu(dyndisk_header->block_size);
        s->bitmap_size = ((s->block_size / (8 * 512)) + 511) & ~511;
A
aliguori 已提交
245

246 247
        s->max_table_entries = be32_to_cpu(dyndisk_header->max_table_entries);
        s->pagetable = g_malloc(s->max_table_entries * 4);
248

249
        s->bat_offset = be64_to_cpu(dyndisk_header->table_offset);
K
Kevin Wolf 已提交
250 251 252 253

        ret = bdrv_pread(bs->file, s->bat_offset, s->pagetable,
                         s->max_table_entries * 4);
        if (ret < 0) {
254 255
            goto fail;
        }
256

257 258
        s->free_data_block_offset =
            (s->bat_offset + (s->max_table_entries * 4) + 511) & ~511;
A
aliguori 已提交
259

260 261 262 263 264
        for (i = 0; i < s->max_table_entries; i++) {
            be32_to_cpus(&s->pagetable[i]);
            if (s->pagetable[i] != 0xFFFFFFFF) {
                int64_t next = (512 * (int64_t) s->pagetable[i]) +
                    s->bitmap_size + s->block_size;
A
aliguori 已提交
265

266 267 268 269
                if (next > s->free_data_block_offset) {
                    s->free_data_block_offset = next;
                }
            }
A
aliguori 已提交
270 271
        }

272 273 274 275 276 277 278
        if (s->free_data_block_offset > bdrv_getlength(bs->file)) {
            error_setg(errp, "block-vpc: free_data_block_offset points after "
                             "the end of file. The image has been truncated.");
            ret = -EINVAL;
            goto fail;
        }

279
        s->last_bitmap_offset = (int64_t) -1;
280 281

#ifdef CACHE
282 283 284 285
        s->pageentry_u8 = g_malloc(512);
        s->pageentry_u32 = s->pageentry_u8;
        s->pageentry_u16 = s->pageentry_u8;
        s->last_pagetable = -1;
286
#endif
287
    }
288

289
    qemu_co_mutex_init(&s->lock);
K
Kevin Wolf 已提交
290 291 292 293 294 295 296

    /* Disable migration when VHD images are used */
    error_set(&s->migration_blocker,
              QERR_BLOCK_FORMAT_FEATURE_NOT_SUPPORTED,
              "vpc", bs->device_name, "live migration");
    migrate_add_blocker(s->migration_blocker);

297
    return 0;
K
Kevin Wolf 已提交
298 299 300 301 302 303 304

fail:
    g_free(s->pagetable);
#ifdef CACHE
    g_free(s->pageentry_u8);
#endif
    return ret;
305 306
}

J
Jeff Cody 已提交
307 308 309 310 311 312
static int vpc_reopen_prepare(BDRVReopenState *state,
                              BlockReopenQueue *queue, Error **errp)
{
    return 0;
}

313 314 315
/*
 * Returns the absolute byte offset of the given sector in the image file.
 * If the sector is not allocated, -1 is returned instead.
A
aliguori 已提交
316 317 318
 *
 * The parameter write must be 1 if the offset will be used for a write
 * operation (the block bitmaps is updated then), 0 otherwise.
319
 */
A
aliguori 已提交
320 321
static inline int64_t get_sector_offset(BlockDriverState *bs,
    int64_t sector_num, int write)
322 323 324 325 326 327
{
    BDRVVPCState *s = bs->opaque;
    uint64_t offset = sector_num * 512;
    uint64_t bitmap_offset, block_offset;
    uint32_t pagetable_index, pageentry_index;

328 329
    pagetable_index = offset / s->block_size;
    pageentry_index = (offset % s->block_size) / 512;
330

A
aliguori 已提交
331 332
    if (pagetable_index >= s->max_table_entries || s->pagetable[pagetable_index] == 0xffffffff)
        return -1; // not allocated
333

334
    bitmap_offset = 512 * (uint64_t) s->pagetable[pagetable_index];
A
aliguori 已提交
335 336 337 338 339 340 341 342 343 344 345 346
    block_offset = bitmap_offset + s->bitmap_size + (512 * pageentry_index);

    // We must ensure that we don't write to any sectors which are marked as
    // unused in the bitmap. We get away with setting all bits in the block
    // bitmap each time we write to a new block. This might cause Virtual PC to
    // miss sparse read optimization, but it's not a problem in terms of
    // correctness.
    if (write && (s->last_bitmap_offset != bitmap_offset)) {
        uint8_t bitmap[s->bitmap_size];

        s->last_bitmap_offset = bitmap_offset;
        memset(bitmap, 0xff, s->bitmap_size);
347
        bdrv_pwrite_sync(bs->file, bitmap_offset, bitmap, s->bitmap_size);
A
aliguori 已提交
348
    }
349

B
bellard 已提交
350
//    printf("sector: %" PRIx64 ", index: %x, offset: %x, bioff: %" PRIx64 ", bloff: %" PRIx64 "\n",
351 352 353 354 355 356 357 358 359 360 361
//	sector_num, pagetable_index, pageentry_index,
//	bitmap_offset, block_offset);

// disabled by reason
#if 0
#ifdef CACHE
    if (bitmap_offset != s->last_bitmap)
    {
	lseek(s->fd, bitmap_offset, SEEK_SET);

	s->last_bitmap = bitmap_offset;
362

363 364 365 366 367 368 369 370 371 372 373
	// Scary! Bitmap is stored as big endian 32bit entries,
	// while we used to look it up byte by byte
	read(s->fd, s->pageentry_u8, 512);
	for (i = 0; i < 128; i++)
	    be32_to_cpus(&s->pageentry_u32[i]);
    }

    if ((s->pageentry_u8[pageentry_index / 8] >> (pageentry_index % 8)) & 1)
	return -1;
#else
    lseek(s->fd, bitmap_offset + (pageentry_index / 8), SEEK_SET);
374

375 376 377 378 379 380 381
    read(s->fd, &bitmap_entry, 1);

    if ((bitmap_entry >> (pageentry_index % 8)) & 1)
	return -1; // not allocated
#endif
#endif

382
    return block_offset;
383 384
}

A
aliguori 已提交
385 386 387 388 389 390 391 392 393 394 395 396
/*
 * Writes the footer to the end of the image file. This is needed when the
 * file grows as it overwrites the old footer
 *
 * Returns 0 on success and < 0 on error
 */
static int rewrite_footer(BlockDriverState* bs)
{
    int ret;
    BDRVVPCState *s = bs->opaque;
    int64_t offset = s->free_data_block_offset;

397
    ret = bdrv_pwrite_sync(bs->file, offset, s->footer_buf, HEADER_SIZE);
A
aliguori 已提交
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
    if (ret < 0)
        return ret;

    return 0;
}

/*
 * Allocates a new block. This involves writing a new footer and updating
 * the Block Allocation Table to use the space at the old end of the image
 * file (overwriting the old footer)
 *
 * Returns the sectors' offset in the image file on success and < 0 on error
 */
static int64_t alloc_block(BlockDriverState* bs, int64_t sector_num)
{
    BDRVVPCState *s = bs->opaque;
    int64_t bat_offset;
    uint32_t index, bat_value;
    int ret;
    uint8_t bitmap[s->bitmap_size];

    // Check if sector_num is valid
    if ((sector_num < 0) || (sector_num > bs->total_sectors))
        return -1;

    // Write entry into in-memory BAT
    index = (sector_num * 512) / s->block_size;
    if (s->pagetable[index] != 0xFFFFFFFF)
        return -1;

    s->pagetable[index] = s->free_data_block_offset / 512;

    // Initialize the block's bitmap
    memset(bitmap, 0xff, s->bitmap_size);
432
    ret = bdrv_pwrite_sync(bs->file, s->free_data_block_offset, bitmap,
433
        s->bitmap_size);
434 435 436
    if (ret < 0) {
        return ret;
    }
A
aliguori 已提交
437 438 439 440 441 442 443 444 445 446

    // Write new footer (the old one will be overwritten)
    s->free_data_block_offset += s->block_size + s->bitmap_size;
    ret = rewrite_footer(bs);
    if (ret < 0)
        goto fail;

    // Write BAT entry to disk
    bat_offset = s->bat_offset + (4 * index);
    bat_value = be32_to_cpu(s->pagetable[index]);
447
    ret = bdrv_pwrite_sync(bs->file, bat_offset, &bat_value, 4);
A
aliguori 已提交
448 449 450 451 452 453 454 455 456 457
    if (ret < 0)
        goto fail;

    return get_sector_offset(bs, sector_num, 0);

fail:
    s->free_data_block_offset -= (s->block_size + s->bitmap_size);
    return -1;
}

458
static int vpc_read(BlockDriverState *bs, int64_t sector_num,
459 460
                    uint8_t *buf, int nb_sectors)
{
461
    BDRVVPCState *s = bs->opaque;
462
    int ret;
463
    int64_t offset;
464
    int64_t sectors, sectors_per_block;
465
    VHDFooter *footer = (VHDFooter *) s->footer_buf;
466

467 468 469
    if (cpu_to_be32(footer->type) == VHD_FIXED) {
        return bdrv_read(bs->file, sector_num, buf, nb_sectors);
    }
470
    while (nb_sectors > 0) {
A
aliguori 已提交
471
        offset = get_sector_offset(bs, sector_num, 0);
472

473 474 475 476 477 478
        sectors_per_block = s->block_size >> BDRV_SECTOR_BITS;
        sectors = sectors_per_block - (sector_num % sectors_per_block);
        if (sectors > nb_sectors) {
            sectors = nb_sectors;
        }

479
        if (offset == -1) {
480
            memset(buf, 0, sectors * BDRV_SECTOR_SIZE);
481
        } else {
482 483 484
            ret = bdrv_pread(bs->file, offset, buf,
                sectors * BDRV_SECTOR_SIZE);
            if (ret != sectors * BDRV_SECTOR_SIZE) {
485
                return -1;
486
            }
487 488
        }

489 490 491
        nb_sectors -= sectors;
        sector_num += sectors;
        buf += sectors * BDRV_SECTOR_SIZE;
492 493 494 495
    }
    return 0;
}

496 497 498 499 500 501 502 503 504 505 506
static coroutine_fn int vpc_co_read(BlockDriverState *bs, int64_t sector_num,
                                    uint8_t *buf, int nb_sectors)
{
    int ret;
    BDRVVPCState *s = bs->opaque;
    qemu_co_mutex_lock(&s->lock);
    ret = vpc_read(bs, sector_num, buf, nb_sectors);
    qemu_co_mutex_unlock(&s->lock);
    return ret;
}

A
aliguori 已提交
507 508 509
static int vpc_write(BlockDriverState *bs, int64_t sector_num,
    const uint8_t *buf, int nb_sectors)
{
510
    BDRVVPCState *s = bs->opaque;
A
aliguori 已提交
511
    int64_t offset;
512
    int64_t sectors, sectors_per_block;
A
aliguori 已提交
513
    int ret;
514
    VHDFooter *footer =  (VHDFooter *) s->footer_buf;
A
aliguori 已提交
515

516 517 518
    if (cpu_to_be32(footer->type) == VHD_FIXED) {
        return bdrv_write(bs->file, sector_num, buf, nb_sectors);
    }
A
aliguori 已提交
519 520 521
    while (nb_sectors > 0) {
        offset = get_sector_offset(bs, sector_num, 1);

522 523 524 525 526 527
        sectors_per_block = s->block_size >> BDRV_SECTOR_BITS;
        sectors = sectors_per_block - (sector_num % sectors_per_block);
        if (sectors > nb_sectors) {
            sectors = nb_sectors;
        }

A
aliguori 已提交
528 529 530 531 532 533
        if (offset == -1) {
            offset = alloc_block(bs, sector_num);
            if (offset < 0)
                return -1;
        }

534 535
        ret = bdrv_pwrite(bs->file, offset, buf, sectors * BDRV_SECTOR_SIZE);
        if (ret != sectors * BDRV_SECTOR_SIZE) {
A
aliguori 已提交
536
            return -1;
537
        }
A
aliguori 已提交
538

539 540 541
        nb_sectors -= sectors;
        sector_num += sectors;
        buf += sectors * BDRV_SECTOR_SIZE;
A
aliguori 已提交
542 543 544 545 546
    }

    return 0;
}

547 548 549 550 551 552 553 554 555 556 557
static coroutine_fn int vpc_co_write(BlockDriverState *bs, int64_t sector_num,
                                     const uint8_t *buf, int nb_sectors)
{
    int ret;
    BDRVVPCState *s = bs->opaque;
    qemu_co_mutex_lock(&s->lock);
    ret = vpc_write(bs, sector_num, buf, nb_sectors);
    qemu_co_mutex_unlock(&s->lock);
    return ret;
}

A
aliguori 已提交
558 559 560 561 562 563 564
/*
 * Calculates the number of cylinders, heads and sectors per cylinder
 * based on a given number of sectors. This is the algorithm described
 * in the VHD specification.
 *
 * Note that the geometry doesn't always exactly match total_sectors but
 * may round it down.
565
 *
566 567 568
 * Returns 0 on success, -EFBIG if the size is larger than ~2 TB. Override
 * the hardware EIDE and ATA-2 limit of 16 heads (max disk size of 127 GB)
 * and instead allow up to 255 heads.
A
aliguori 已提交
569
 */
570
static int calculate_geometry(int64_t total_sectors, uint16_t* cyls,
A
aliguori 已提交
571 572 573 574
    uint8_t* heads, uint8_t* secs_per_cyl)
{
    uint32_t cyls_times_heads;

575 576
    /* Allow a maximum disk size of approximately 2 TB */
    if (total_sectors > 65535LL * 255 * 255) {
577
        return -EFBIG;
578
    }
A
aliguori 已提交
579 580 581

    if (total_sectors > 65535 * 16 * 63) {
        *secs_per_cyl = 255;
582 583 584 585 586
        if (total_sectors > 65535 * 16 * 255) {
            *heads = 255;
        } else {
            *heads = 16;
        }
A
aliguori 已提交
587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608
        cyls_times_heads = total_sectors / *secs_per_cyl;
    } else {
        *secs_per_cyl = 17;
        cyls_times_heads = total_sectors / *secs_per_cyl;
        *heads = (cyls_times_heads + 1023) / 1024;

        if (*heads < 4)
            *heads = 4;

        if (cyls_times_heads >= (*heads * 1024) || *heads > 16) {
            *secs_per_cyl = 31;
            *heads = 16;
            cyls_times_heads = total_sectors / *secs_per_cyl;
        }

        if (cyls_times_heads >= (*heads * 1024)) {
            *secs_per_cyl = 63;
            *heads = 16;
            cyls_times_heads = total_sectors / *secs_per_cyl;
        }
    }

609
    *cyls = cyls_times_heads / *heads;
610 611

    return 0;
A
aliguori 已提交
612 613
}

614
static int create_dynamic_disk(int fd, uint8_t *buf, int64_t total_sectors)
A
aliguori 已提交
615
{
616 617
    VHDDynDiskHeader *dyndisk_header =
        (VHDDynDiskHeader *) buf;
A
aliguori 已提交
618
    size_t block_size, num_bat_entries;
619
    int i;
B
Blue Swirl 已提交
620
    int ret = -EIO;
A
aliguori 已提交
621 622 623 624 625

    // Write the footer (twice: at the beginning and at the end)
    block_size = 0x200000;
    num_bat_entries = (total_sectors + block_size / 512) / (block_size / 512);

B
Blue Swirl 已提交
626 627 628
    if (write(fd, buf, HEADER_SIZE) != HEADER_SIZE) {
        goto fail;
    }
A
aliguori 已提交
629

B
Blue Swirl 已提交
630 631 632 633 634 635
    if (lseek(fd, 1536 + ((num_bat_entries * 4 + 511) & ~511), SEEK_SET) < 0) {
        goto fail;
    }
    if (write(fd, buf, HEADER_SIZE) != HEADER_SIZE) {
        goto fail;
    }
A
aliguori 已提交
636 637

    // Write the initial BAT
B
Blue Swirl 已提交
638 639 640
    if (lseek(fd, 3 * 512, SEEK_SET) < 0) {
        goto fail;
    }
A
aliguori 已提交
641 642

    memset(buf, 0xFF, 512);
B
Blue Swirl 已提交
643 644 645 646 647
    for (i = 0; i < (num_bat_entries * 4 + 511) / 512; i++) {
        if (write(fd, buf, 512) != 512) {
            goto fail;
        }
    }
A
aliguori 已提交
648 649 650 651

    // Prepare the Dynamic Disk Header
    memset(buf, 0, 1024);

652
    memcpy(dyndisk_header->magic, "cxsparse", 8);
A
aliguori 已提交
653

654 655 656 657 658
    /*
     * Note: The spec is actually wrong here for data_offset, it says
     * 0xFFFFFFFF, but MS tools expect all 64 bits to be set.
     */
    dyndisk_header->data_offset = be64_to_cpu(0xFFFFFFFFFFFFFFFFULL);
A
aliguori 已提交
659 660 661 662 663 664 665 666
    dyndisk_header->table_offset = be64_to_cpu(3 * 512);
    dyndisk_header->version = be32_to_cpu(0x00010000);
    dyndisk_header->block_size = be32_to_cpu(block_size);
    dyndisk_header->max_table_entries = be32_to_cpu(num_bat_entries);

    dyndisk_header->checksum = be32_to_cpu(vpc_checksum(buf, 1024));

    // Write the header
B
Blue Swirl 已提交
667 668 669
    if (lseek(fd, 512, SEEK_SET) < 0) {
        goto fail;
    }
A
aliguori 已提交
670

B
Blue Swirl 已提交
671 672 673 674 675
    if (write(fd, buf, 1024) != 1024) {
        goto fail;
    }
    ret = 0;

676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702
 fail:
    return ret;
}

static int create_fixed_disk(int fd, uint8_t *buf, int64_t total_size)
{
    int ret = -EIO;

    /* Add footer to total size */
    total_size += 512;
    if (ftruncate(fd, total_size) != 0) {
        ret = -errno;
        goto fail;
    }
    if (lseek(fd, -512, SEEK_END) < 0) {
        goto fail;
    }
    if (write(fd, buf, HEADER_SIZE) != HEADER_SIZE) {
        goto fail;
    }

    ret = 0;

 fail:
    return ret;
}

703 704
static int vpc_create(const char *filename, QEMUOptionParameter *options,
                      Error **errp)
705 706
{
    uint8_t buf[1024];
707
    VHDFooter *footer = (VHDFooter *) buf;
708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734
    QEMUOptionParameter *disk_type_param;
    int fd, i;
    uint16_t cyls = 0;
    uint8_t heads = 0;
    uint8_t secs_per_cyl = 0;
    int64_t total_sectors;
    int64_t total_size;
    int disk_type;
    int ret = -EIO;

    /* Read out options */
    total_size = get_option_parameter(options, BLOCK_OPT_SIZE)->value.n;

    disk_type_param = get_option_parameter(options, BLOCK_OPT_SUBFMT);
    if (disk_type_param && disk_type_param->value.s) {
        if (!strcmp(disk_type_param->value.s, "dynamic")) {
            disk_type = VHD_DYNAMIC;
        } else if (!strcmp(disk_type_param->value.s, "fixed")) {
            disk_type = VHD_FIXED;
        } else {
            return -EINVAL;
        }
    } else {
        disk_type = VHD_DYNAMIC;
    }

    /* Create the file */
735
    fd = qemu_open(filename, O_WRONLY | O_CREAT | O_TRUNC | O_BINARY, 0644);
736 737 738 739
    if (fd < 0) {
        return -EIO;
    }

740 741 742 743 744
    /*
     * Calculate matching total_size and geometry. Increase the number of
     * sectors requested until we get enough (or fail). This ensures that
     * qemu-img convert doesn't truncate images, but rather rounds up.
     */
745
    total_sectors = total_size / BDRV_SECTOR_SIZE;
746 747 748 749
    for (i = 0; total_sectors > (int64_t)cyls * heads * secs_per_cyl; i++) {
        if (calculate_geometry(total_sectors + i, &cyls, &heads,
                               &secs_per_cyl))
        {
750 751 752 753
            ret = -EFBIG;
            goto fail;
        }
    }
754

755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789
    total_sectors = (int64_t) cyls * heads * secs_per_cyl;

    /* Prepare the Hard Disk Footer */
    memset(buf, 0, 1024);

    memcpy(footer->creator, "conectix", 8);
    /* TODO Check if "qemu" creator_app is ok for VPC */
    memcpy(footer->creator_app, "qemu", 4);
    memcpy(footer->creator_os, "Wi2k", 4);

    footer->features = be32_to_cpu(0x02);
    footer->version = be32_to_cpu(0x00010000);
    if (disk_type == VHD_DYNAMIC) {
        footer->data_offset = be64_to_cpu(HEADER_SIZE);
    } else {
        footer->data_offset = be64_to_cpu(0xFFFFFFFFFFFFFFFFULL);
    }
    footer->timestamp = be32_to_cpu(time(NULL) - VHD_TIMESTAMP_BASE);

    /* Version of Virtual PC 2007 */
    footer->major = be16_to_cpu(0x0005);
    footer->minor = be16_to_cpu(0x0003);
    if (disk_type == VHD_DYNAMIC) {
        footer->orig_size = be64_to_cpu(total_sectors * 512);
        footer->size = be64_to_cpu(total_sectors * 512);
    } else {
        footer->orig_size = be64_to_cpu(total_size);
        footer->size = be64_to_cpu(total_size);
    }
    footer->cyls = be16_to_cpu(cyls);
    footer->heads = heads;
    footer->secs_per_cyl = secs_per_cyl;

    footer->type = be32_to_cpu(disk_type);

790 791 792
#if defined(CONFIG_UUID)
    uuid_generate(footer->uuid);
#endif
793 794 795 796 797 798 799 800 801

    footer->checksum = be32_to_cpu(vpc_checksum(buf, HEADER_SIZE));

    if (disk_type == VHD_DYNAMIC) {
        ret = create_dynamic_disk(fd, buf, total_sectors);
    } else {
        ret = create_fixed_disk(fd, buf, total_size);
    }

B
Blue Swirl 已提交
802
 fail:
803
    qemu_close(fd);
B
Blue Swirl 已提交
804
    return ret;
A
aliguori 已提交
805 806
}

K
Kevin Wolf 已提交
807 808 809
static int vpc_has_zero_init(BlockDriverState *bs)
{
    BDRVVPCState *s = bs->opaque;
810
    VHDFooter *footer =  (VHDFooter *) s->footer_buf;
K
Kevin Wolf 已提交
811 812 813 814 815 816 817 818

    if (cpu_to_be32(footer->type) == VHD_FIXED) {
        return bdrv_has_zero_init(bs->file);
    } else {
        return 1;
    }
}

819 820 821
static void vpc_close(BlockDriverState *bs)
{
    BDRVVPCState *s = bs->opaque;
822
    g_free(s->pagetable);
823
#ifdef CACHE
824
    g_free(s->pageentry_u8);
825
#endif
K
Kevin Wolf 已提交
826 827 828

    migrate_del_blocker(s->migration_blocker);
    error_free(s->migration_blocker);
829 830
}

831
static QEMUOptionParameter vpc_create_options[] = {
832 833 834 835 836
    {
        .name = BLOCK_OPT_SIZE,
        .type = OPT_SIZE,
        .help = "Virtual disk size"
    },
837 838 839 840 841 842 843
    {
        .name = BLOCK_OPT_SUBFMT,
        .type = OPT_STRING,
        .help =
            "Type of virtual hard disk format. Supported formats are "
            "{dynamic (default) | fixed} "
    },
844 845 846
    { NULL }
};

847
static BlockDriver bdrv_vpc = {
K
Kevin Wolf 已提交
848 849
    .format_name    = "vpc",
    .instance_size  = sizeof(BDRVVPCState),
850

K
Kevin Wolf 已提交
851 852 853 854 855
    .bdrv_probe             = vpc_probe,
    .bdrv_open              = vpc_open,
    .bdrv_close             = vpc_close,
    .bdrv_reopen_prepare    = vpc_reopen_prepare,
    .bdrv_create            = vpc_create,
856

857 858 859
    .bdrv_read              = vpc_co_read,
    .bdrv_write             = vpc_co_write,

K
Kevin Wolf 已提交
860 861
    .create_options         = vpc_create_options,
    .bdrv_has_zero_init     = vpc_has_zero_init,
862
};
863 864 865 866 867 868 869

static void bdrv_vpc_init(void)
{
    bdrv_register(&bdrv_vpc);
}

block_init(bdrv_vpc_init);