block.c 112.5 KB
Newer Older
B
bellard 已提交
1 2
/*
 * QEMU System Emulator block driver
3
 *
B
bellard 已提交
4
 * Copyright (c) 2003 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.
 */
24
#include "config-host.h"
P
pbrook 已提交
25
#include "qemu-common.h"
26
#include "trace.h"
27 28
#include "block/block_int.h"
#include "block/blockjob.h"
29
#include "qemu/module.h"
30
#include "qapi/qmp/qjson.h"
31
#include "sysemu/block-backend.h"
32
#include "sysemu/sysemu.h"
33
#include "qemu/notify.h"
34
#include "block/coroutine.h"
35
#include "block/qapi.h"
L
Luiz Capitulino 已提交
36
#include "qmp-commands.h"
37
#include "qemu/timer.h"
38
#include "qapi-event.h"
B
bellard 已提交
39

J
Juan Quintela 已提交
40
#ifdef CONFIG_BSD
B
bellard 已提交
41 42 43
#include <sys/types.h>
#include <sys/stat.h>
#include <sys/ioctl.h>
B
Blue Swirl 已提交
44
#include <sys/queue.h>
45
#ifndef __DragonFly__
B
bellard 已提交
46 47
#include <sys/disk.h>
#endif
48
#endif
B
bellard 已提交
49

50 51 52 53
#ifdef _WIN32
#include <windows.h>
#endif

J
John Snow 已提交
54 55 56 57 58 59 60 61
/**
 * A BdrvDirtyBitmap can be in three possible states:
 * (1) successor is NULL and disabled is false: full r/w mode
 * (2) successor is NULL and disabled is true: read only mode ("disabled")
 * (3) successor is set: frozen mode.
 *     A frozen bitmap cannot be renamed, deleted, anonymized, cleared, set,
 *     or enabled. A frozen bitmap can only abdicate() or reclaim().
 */
F
Fam Zheng 已提交
62
struct BdrvDirtyBitmap {
63 64 65 66 67
    HBitmap *bitmap;            /* Dirty sector bitmap implementation */
    BdrvDirtyBitmap *successor; /* Anonymous child; implies frozen status */
    char *name;                 /* Optional non-empty unique ID */
    int64_t size;               /* Size of the bitmap (Number of sectors) */
    bool disabled;              /* Bitmap is read-only */
F
Fam Zheng 已提交
68 69 70
    QLIST_ENTRY(BdrvDirtyBitmap) list;
};

71 72
#define NOT_DONE 0x7fffffff /* used while emulated sync operation in progress */

73 74
static QTAILQ_HEAD(, BlockDriverState) bdrv_states =
    QTAILQ_HEAD_INITIALIZER(bdrv_states);
75

76 77 78
static QTAILQ_HEAD(, BlockDriverState) graph_bdrv_states =
    QTAILQ_HEAD_INITIALIZER(graph_bdrv_states);

79 80
static QLIST_HEAD(, BlockDriver) bdrv_drivers =
    QLIST_HEAD_INITIALIZER(bdrv_drivers);
B
bellard 已提交
81

82
static void bdrv_dirty_bitmap_truncate(BlockDriverState *bs);
83 84 85
/* If non-zero, use only whitelisted block drivers */
static int use_bdrv_whitelist;

86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105
#ifdef _WIN32
static int is_windows_drive_prefix(const char *filename)
{
    return (((filename[0] >= 'a' && filename[0] <= 'z') ||
             (filename[0] >= 'A' && filename[0] <= 'Z')) &&
            filename[1] == ':');
}

int is_windows_drive(const char *filename)
{
    if (is_windows_drive_prefix(filename) &&
        filename[2] == '\0')
        return 1;
    if (strstart(filename, "\\\\.\\", NULL) ||
        strstart(filename, "//./", NULL))
        return 1;
    return 0;
}
#endif

106 107 108
size_t bdrv_opt_mem_align(BlockDriverState *bs)
{
    if (!bs || !bs->drv) {
109 110
        /* page size or 4k (hdd sector size) should be on the safe side */
        return MAX(4096, getpagesize());
111 112 113 114 115
    }

    return bs->bl.opt_mem_alignment;
}

116 117 118
size_t bdrv_min_mem_align(BlockDriverState *bs)
{
    if (!bs || !bs->drv) {
119 120
        /* page size or 4k (hdd sector size) should be on the safe side */
        return MAX(4096, getpagesize());
121 122 123 124 125
    }

    return bs->bl.min_mem_alignment;
}

126
/* check if the path starts with "<protocol>:" */
127
int path_has_protocol(const char *path)
128
{
129 130
    const char *p;

131 132 133 134 135
#ifdef _WIN32
    if (is_windows_drive(path) ||
        is_windows_drive_prefix(path)) {
        return 0;
    }
136 137 138
    p = path + strcspn(path, ":/\\");
#else
    p = path + strcspn(path, ":/");
139 140
#endif

141
    return *p == ':';
142 143
}

B
bellard 已提交
144
int path_is_absolute(const char *path)
145
{
B
bellard 已提交
146 147
#ifdef _WIN32
    /* specific case for names like: "\\.\d:" */
P
Paolo Bonzini 已提交
148
    if (is_windows_drive(path) || is_windows_drive_prefix(path)) {
B
bellard 已提交
149
        return 1;
P
Paolo Bonzini 已提交
150 151
    }
    return (*path == '/' || *path == '\\');
152
#else
P
Paolo Bonzini 已提交
153
    return (*path == '/');
154
#endif
155 156
}

B
bellard 已提交
157 158 159 160 161 162
/* if filename is absolute, just copy it to dest. Otherwise, build a
   path to it by considering it is relative to base_path. URL are
   supported. */
void path_combine(char *dest, int dest_size,
                  const char *base_path,
                  const char *filename)
163
{
B
bellard 已提交
164 165 166 167 168 169 170 171 172 173 174 175 176
    const char *p, *p1;
    int len;

    if (dest_size <= 0)
        return;
    if (path_is_absolute(filename)) {
        pstrcpy(dest, dest_size, filename);
    } else {
        p = strchr(base_path, ':');
        if (p)
            p++;
        else
            p = base_path;
177 178 179 180 181 182 183 184 185
        p1 = strrchr(base_path, '/');
#ifdef _WIN32
        {
            const char *p2;
            p2 = strrchr(base_path, '\\');
            if (!p1 || p2 > p1)
                p1 = p2;
        }
#endif
B
bellard 已提交
186 187 188 189 190 191 192 193 194 195 196 197
        if (p1)
            p1++;
        else
            p1 = base_path;
        if (p1 > p)
            p = p1;
        len = p - base_path;
        if (len > dest_size - 1)
            len = dest_size - 1;
        memcpy(dest, base_path, len);
        dest[len] = '\0';
        pstrcat(dest, dest_size, filename);
198 199 200
    }
}

201 202
void bdrv_get_full_backing_filename_from_filename(const char *backed,
                                                  const char *backing,
203 204
                                                  char *dest, size_t sz,
                                                  Error **errp)
205
{
206 207 208
    if (backing[0] == '\0' || path_has_protocol(backing) ||
        path_is_absolute(backing))
    {
209
        pstrcpy(dest, sz, backing);
210 211 212
    } else if (backed[0] == '\0' || strstart(backed, "json:", NULL)) {
        error_setg(errp, "Cannot use relative backing file names for '%s'",
                   backed);
213
    } else {
214
        path_combine(dest, sz, backed, backing);
215 216 217
    }
}

218 219
void bdrv_get_full_backing_filename(BlockDriverState *bs, char *dest, size_t sz,
                                    Error **errp)
220
{
221 222 223 224
    char *backed = bs->exact_filename[0] ? bs->exact_filename : bs->filename;

    bdrv_get_full_backing_filename_from_filename(backed, bs->backing_file,
                                                 dest, sz, errp);
225 226
}

227 228 229
void bdrv_register(BlockDriver *bdrv)
{
    bdrv_setup_io_funcs(bdrv);
230

231
    QLIST_INSERT_HEAD(&bdrv_drivers, bdrv, list);
B
bellard 已提交
232
}
B
bellard 已提交
233

234
BlockDriverState *bdrv_new_root(void)
B
bellard 已提交
235
{
236
    BlockDriverState *bs = bdrv_new();
237 238 239 240 241 242 243 244 245 246

    QTAILQ_INSERT_TAIL(&bdrv_states, bs, device_list);
    return bs;
}

BlockDriverState *bdrv_new(void)
{
    BlockDriverState *bs;
    int i;

247
    bs = g_new0(BlockDriverState, 1);
F
Fam Zheng 已提交
248
    QLIST_INIT(&bs->dirty_bitmaps);
249 250 251
    for (i = 0; i < BLOCK_OP_TYPE_MAX; i++) {
        QLIST_INIT(&bs->op_blockers[i]);
    }
252
    bdrv_iostatus_disable(bs);
P
Paolo Bonzini 已提交
253
    notifier_list_init(&bs->close_notifiers);
254
    notifier_with_return_list_init(&bs->before_write_notifiers);
255 256
    qemu_co_queue_init(&bs->throttled_reqs[0]);
    qemu_co_queue_init(&bs->throttled_reqs[1]);
257
    bs->refcnt = 1;
258
    bs->aio_context = qemu_get_aio_context();
P
Paolo Bonzini 已提交
259

B
bellard 已提交
260 261 262
    return bs;
}

P
Paolo Bonzini 已提交
263 264 265 266 267
void bdrv_add_close_notifier(BlockDriverState *bs, Notifier *notify)
{
    notifier_list_add(&bs->close_notifiers, notify);
}

B
bellard 已提交
268 269 270
BlockDriver *bdrv_find_format(const char *format_name)
{
    BlockDriver *drv1;
271 272
    QLIST_FOREACH(drv1, &bdrv_drivers, list) {
        if (!strcmp(drv1->format_name, format_name)) {
B
bellard 已提交
273
            return drv1;
274
        }
B
bellard 已提交
275 276 277 278
    }
    return NULL;
}

279
static int bdrv_is_whitelisted(BlockDriver *drv, bool read_only)
280
{
281 282 283 284 285
    static const char *whitelist_rw[] = {
        CONFIG_BDRV_RW_WHITELIST
    };
    static const char *whitelist_ro[] = {
        CONFIG_BDRV_RO_WHITELIST
286 287 288
    };
    const char **p;

289
    if (!whitelist_rw[0] && !whitelist_ro[0]) {
290
        return 1;               /* no whitelist, anything goes */
291
    }
292

293
    for (p = whitelist_rw; *p; p++) {
294 295 296 297
        if (!strcmp(drv->format_name, *p)) {
            return 1;
        }
    }
298 299 300 301 302 303 304
    if (read_only) {
        for (p = whitelist_ro; *p; p++) {
            if (!strcmp(drv->format_name, *p)) {
                return 1;
            }
        }
    }
305 306 307
    return 0;
}

308 309
BlockDriver *bdrv_find_whitelisted_format(const char *format_name,
                                          bool read_only)
310 311
{
    BlockDriver *drv = bdrv_find_format(format_name);
312
    return drv && bdrv_is_whitelisted(drv, read_only) ? drv : NULL;
313 314
}

315 316 317
typedef struct CreateCo {
    BlockDriver *drv;
    char *filename;
318
    QemuOpts *opts;
319
    int ret;
320
    Error *err;
321 322 323 324
} CreateCo;

static void coroutine_fn bdrv_create_co_entry(void *opaque)
{
325 326 327
    Error *local_err = NULL;
    int ret;

328 329 330
    CreateCo *cco = opaque;
    assert(cco->drv);

C
Chunyan Liu 已提交
331
    ret = cco->drv->bdrv_create(cco->filename, cco->opts, &local_err);
332
    if (local_err) {
333 334 335
        error_propagate(&cco->err, local_err);
    }
    cco->ret = ret;
336 337
}

338
int bdrv_create(BlockDriver *drv, const char* filename,
339
                QemuOpts *opts, Error **errp)
B
bellard 已提交
340
{
341 342 343 344 345 346
    int ret;

    Coroutine *co;
    CreateCo cco = {
        .drv = drv,
        .filename = g_strdup(filename),
347
        .opts = opts,
348
        .ret = NOT_DONE,
349
        .err = NULL,
350 351
    };

C
Chunyan Liu 已提交
352
    if (!drv->bdrv_create) {
353
        error_setg(errp, "Driver '%s' does not support image creation", drv->format_name);
354 355
        ret = -ENOTSUP;
        goto out;
356 357 358 359 360 361 362 363 364
    }

    if (qemu_in_coroutine()) {
        /* Fast-path if already in coroutine context */
        bdrv_create_co_entry(&cco);
    } else {
        co = qemu_coroutine_create(bdrv_create_co_entry);
        qemu_coroutine_enter(co, &cco);
        while (cco.ret == NOT_DONE) {
365
            aio_poll(qemu_get_aio_context(), true);
366 367 368 369
        }
    }

    ret = cco.ret;
370
    if (ret < 0) {
371
        if (cco.err) {
372 373 374 375 376
            error_propagate(errp, cco.err);
        } else {
            error_setg_errno(errp, -ret, "Could not create image");
        }
    }
377

378 379
out:
    g_free(cco.filename);
380
    return ret;
B
bellard 已提交
381 382
}

C
Chunyan Liu 已提交
383
int bdrv_create_file(const char *filename, QemuOpts *opts, Error **errp)
384 385
{
    BlockDriver *drv;
386 387
    Error *local_err = NULL;
    int ret;
388

389
    drv = bdrv_find_protocol(filename, true, errp);
390
    if (drv == NULL) {
391
        return -ENOENT;
392 393
    }

C
Chunyan Liu 已提交
394
    ret = bdrv_create(drv, filename, opts, &local_err);
395
    if (local_err) {
396 397 398
        error_propagate(errp, local_err);
    }
    return ret;
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
/**
 * Try to get @bs's logical and physical block size.
 * On success, store them in @bsz struct and return 0.
 * On failure return -errno.
 * @bs must not be empty.
 */
int bdrv_probe_blocksizes(BlockDriverState *bs, BlockSizes *bsz)
{
    BlockDriver *drv = bs->drv;

    if (drv && drv->bdrv_probe_blocksizes) {
        return drv->bdrv_probe_blocksizes(bs, bsz);
    }

    return -ENOTSUP;
}

/**
 * Try to get @bs's geometry (cyls, heads, sectors).
 * On success, store them in @geo struct and return 0.
 * On failure return -errno.
 * @bs must not be empty.
 */
int bdrv_probe_geometry(BlockDriverState *bs, HDGeometry *geo)
{
    BlockDriver *drv = bs->drv;

    if (drv && drv->bdrv_probe_geometry) {
        return drv->bdrv_probe_geometry(bs, geo);
    }

    return -ENOTSUP;
}

435 436 437 438 439
/*
 * Create a uniquely-named empty temporary file.
 * Return 0 upon success, otherwise a negative errno value.
 */
int get_tmp_filename(char *filename, int size)
B
bellard 已提交
440
{
441
#ifdef _WIN32
442
    char temp_dir[MAX_PATH];
443 444 445 446 447 448
    /* GetTempFileName requires that its output buffer (4th param)
       have length MAX_PATH or greater.  */
    assert(size >= MAX_PATH);
    return (GetTempPath(MAX_PATH, temp_dir)
            && GetTempFileName(temp_dir, "qem", 0, filename)
            ? 0 : -GetLastError());
B
bellard 已提交
449
#else
B
bellard 已提交
450
    int fd;
451
    const char *tmpdir;
A
aurel32 已提交
452
    tmpdir = getenv("TMPDIR");
453 454 455
    if (!tmpdir) {
        tmpdir = "/var/tmp";
    }
456 457 458
    if (snprintf(filename, size, "%s/vl.XXXXXX", tmpdir) >= size) {
        return -EOVERFLOW;
    }
B
bellard 已提交
459
    fd = mkstemp(filename);
460 461 462 463 464
    if (fd < 0) {
        return -errno;
    }
    if (close(fd) != 0) {
        unlink(filename);
465 466 467
        return -errno;
    }
    return 0;
B
bellard 已提交
468
#endif
469
}
B
bellard 已提交
470

471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492
/*
 * Detect host devices. By convention, /dev/cdrom[N] is always
 * recognized as a host CDROM.
 */
static BlockDriver *find_hdev_driver(const char *filename)
{
    int score_max = 0, score;
    BlockDriver *drv = NULL, *d;

    QLIST_FOREACH(d, &bdrv_drivers, list) {
        if (d->bdrv_probe_device) {
            score = d->bdrv_probe_device(filename);
            if (score > score_max) {
                score_max = score;
                drv = d;
            }
        }
    }

    return drv;
}

493
BlockDriver *bdrv_find_protocol(const char *filename,
494 495
                                bool allow_protocol_prefix,
                                Error **errp)
B
bellard 已提交
496 497 498
{
    BlockDriver *drv1;
    char protocol[128];
499
    int len;
B
bellard 已提交
500
    const char *p;
B
bellard 已提交
501

502 503
    /* TODO Drivers without bdrv_file_open must be specified explicitly */

504 505 506 507 508 509 510 511 512 513 514 515
    /*
     * XXX(hch): we really should not let host device detection
     * override an explicit protocol specification, but moving this
     * later breaks access to device names with colons in them.
     * Thanks to the brain-dead persistent naming schemes on udev-
     * based Linux systems those actually are quite common.
     */
    drv1 = find_hdev_driver(filename);
    if (drv1) {
        return drv1;
    }

516
    if (!path_has_protocol(filename) || !allow_protocol_prefix) {
517
        return &bdrv_file;
518
    }
519

520 521
    p = strchr(filename, ':');
    assert(p != NULL);
522 523 524 525 526
    len = p - filename;
    if (len > sizeof(protocol) - 1)
        len = sizeof(protocol) - 1;
    memcpy(protocol, filename, len);
    protocol[len] = '\0';
527
    QLIST_FOREACH(drv1, &bdrv_drivers, list) {
528
        if (drv1->protocol_name &&
529
            !strcmp(drv1->protocol_name, protocol)) {
B
bellard 已提交
530
            return drv1;
531
        }
B
bellard 已提交
532
    }
533 534

    error_setg(errp, "Unknown protocol '%s'", protocol);
B
bellard 已提交
535 536 537
    return NULL;
}

538 539 540 541 542 543
/*
 * Guess image format by probing its contents.
 * This is not a good idea when your image is raw (CVE-2008-2004), but
 * we do it anyway for backward compatibility.
 *
 * @buf         contains the image's first @buf_size bytes.
544 545
 * @buf_size    is the buffer size in bytes (generally BLOCK_PROBE_BUF_SIZE,
 *              but can be smaller if the image file is smaller)
546 547 548 549 550 551
 * @filename    is its filename.
 *
 * For all block drivers, call the bdrv_probe() method to get its
 * probing score.
 * Return the first block driver with the highest probing score.
 */
552 553
BlockDriver *bdrv_probe_all(const uint8_t *buf, int buf_size,
                            const char *filename)
554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570
{
    int score_max = 0, score;
    BlockDriver *drv = NULL, *d;

    QLIST_FOREACH(d, &bdrv_drivers, list) {
        if (d->bdrv_probe) {
            score = d->bdrv_probe(buf, buf_size, filename);
            if (score > score_max) {
                score_max = score;
                drv = d;
            }
        }
    }

    return drv;
}

571
static int find_image_format(BlockDriverState *bs, const char *filename,
572
                             BlockDriver **pdrv, Error **errp)
573
{
574
    BlockDriver *drv;
575
    uint8_t buf[BLOCK_PROBE_BUF_SIZE];
576
    int ret = 0;
577

578
    /* Return the raw BlockDriver * to scsi-generic devices or empty drives */
579
    if (bs->sg || !bdrv_is_inserted(bs) || bdrv_getlength(bs) == 0) {
580
        *pdrv = &bdrv_raw;
581
        return ret;
582
    }
583

B
bellard 已提交
584 585
    ret = bdrv_pread(bs, 0, buf, sizeof(buf));
    if (ret < 0) {
586 587
        error_setg_errno(errp, -ret, "Could not read image for determining its "
                         "format");
588 589
        *pdrv = NULL;
        return ret;
B
bellard 已提交
590 591
    }

592
    drv = bdrv_probe_all(buf, ret, filename);
593
    if (!drv) {
594 595
        error_setg(errp, "Could not determine image format: No compatible "
                   "driver found");
596 597 598 599
        ret = -ENOENT;
    }
    *pdrv = drv;
    return ret;
B
bellard 已提交
600 601
}

602 603
/**
 * Set the current 'total_sectors' value
M
Markus Armbruster 已提交
604
 * Return 0 on success, -errno on error.
605 606 607 608 609
 */
static int refresh_total_sectors(BlockDriverState *bs, int64_t hint)
{
    BlockDriver *drv = bs->drv;

610 611 612 613
    /* Do not attempt drv->bdrv_getlength() on scsi-generic devices */
    if (bs->sg)
        return 0;

614 615 616 617 618 619
    /* query actual device if possible, otherwise just trust the hint */
    if (drv->bdrv_getlength) {
        int64_t length = drv->bdrv_getlength(bs);
        if (length < 0) {
            return length;
        }
F
Fam Zheng 已提交
620
        hint = DIV_ROUND_UP(length, BDRV_SECTOR_SIZE);
621 622 623 624 625 626
    }

    bs->total_sectors = hint;
    return 0;
}

P
Paolo Bonzini 已提交
627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646
/**
 * Set open flags for a given discard mode
 *
 * Return 0 on success, -1 if the discard mode was invalid.
 */
int bdrv_parse_discard_flags(const char *mode, int *flags)
{
    *flags &= ~BDRV_O_UNMAP;

    if (!strcmp(mode, "off") || !strcmp(mode, "ignore")) {
        /* do nothing */
    } else if (!strcmp(mode, "on") || !strcmp(mode, "unmap")) {
        *flags |= BDRV_O_UNMAP;
    } else {
        return -1;
    }

    return 0;
}

647 648 649 650 651 652 653 654 655 656 657
/**
 * Set open flags for a given cache mode
 *
 * Return 0 on success, -1 if the cache mode was invalid.
 */
int bdrv_parse_cache_flags(const char *mode, int *flags)
{
    *flags &= ~BDRV_O_CACHE_MASK;

    if (!strcmp(mode, "off") || !strcmp(mode, "none")) {
        *flags |= BDRV_O_NOCACHE | BDRV_O_CACHE_WB;
658 659
    } else if (!strcmp(mode, "directsync")) {
        *flags |= BDRV_O_NOCACHE;
660 661 662 663 664 665 666 667 668 669 670 671 672 673
    } else if (!strcmp(mode, "writeback")) {
        *flags |= BDRV_O_CACHE_WB;
    } else if (!strcmp(mode, "unsafe")) {
        *flags |= BDRV_O_CACHE_WB;
        *flags |= BDRV_O_NO_FLUSH;
    } else if (!strcmp(mode, "writethrough")) {
        /* this is the default */
    } else {
        return -1;
    }

    return 0;
}

674 675 676 677 678 679 680 681 682 683
/*
 * Returns the flags that a temporary snapshot should get, based on the
 * originally requested flags (the originally requested image will have flags
 * like a backing file)
 */
static int bdrv_temp_snapshot_flags(int flags)
{
    return (flags & ~BDRV_O_SNAPSHOT) | BDRV_O_TEMPORARY;
}

684 685 686 687 688 689 690 691 692 693 694 695 696 697
/*
 * Returns the flags that bs->file should get, based on the given flags for
 * the parent BDS
 */
static int bdrv_inherited_flags(int flags)
{
    /* Enable protocol handling, disable format probing for bs->file */
    flags |= BDRV_O_PROTOCOL;

    /* Our block drivers take care to send flushes and respect unmap policy,
     * so we can enable both unconditionally on lower layers. */
    flags |= BDRV_O_CACHE_WB | BDRV_O_UNMAP;

    /* Clear flags that only apply to the top layer */
698
    flags &= ~(BDRV_O_SNAPSHOT | BDRV_O_NO_BACKING | BDRV_O_COPY_ON_READ);
699 700 701 702

    return flags;
}

K
Kevin Wolf 已提交
703 704 705 706 707 708 709 710 711 712
/*
 * Returns the flags that bs->backing_hd should get, based on the given flags
 * for the parent BDS
 */
static int bdrv_backing_flags(int flags)
{
    /* backing files always opened read-only */
    flags &= ~(BDRV_O_RDWR | BDRV_O_COPY_ON_READ);

    /* snapshot=on is handled on the top layer */
713
    flags &= ~(BDRV_O_SNAPSHOT | BDRV_O_TEMPORARY);
K
Kevin Wolf 已提交
714 715 716 717

    return flags;
}

K
Kevin Wolf 已提交
718 719 720 721 722 723 724 725
static int bdrv_open_flags(BlockDriverState *bs, int flags)
{
    int open_flags = flags | BDRV_O_CACHE_WB;

    /*
     * Clear flags that are internal to the block layer before opening the
     * image.
     */
726
    open_flags &= ~(BDRV_O_SNAPSHOT | BDRV_O_NO_BACKING | BDRV_O_PROTOCOL);
K
Kevin Wolf 已提交
727 728 729 730

    /*
     * Snapshots should be writable.
     */
731
    if (flags & BDRV_O_TEMPORARY) {
K
Kevin Wolf 已提交
732 733 734 735 736 737
        open_flags |= BDRV_O_RDWR;
    }

    return open_flags;
}

738 739 740
static void bdrv_assign_node_name(BlockDriverState *bs,
                                  const char *node_name,
                                  Error **errp)
741 742
{
    if (!node_name) {
743
        return;
744 745
    }

K
Kevin Wolf 已提交
746
    /* Check for empty string or invalid characters */
747
    if (!id_wellformed(node_name)) {
K
Kevin Wolf 已提交
748
        error_setg(errp, "Invalid node name");
749
        return;
750 751
    }

752
    /* takes care of avoiding namespaces collisions */
753
    if (blk_by_name(node_name)) {
754 755
        error_setg(errp, "node-name=%s is conflicting with a device id",
                   node_name);
756
        return;
757 758
    }

759 760 761
    /* takes care of avoiding duplicates node names */
    if (bdrv_find_node(node_name)) {
        error_setg(errp, "Duplicate node name");
762
        return;
763 764 765 766 767 768 769
    }

    /* copy node name into the bs and insert it into the graph list */
    pstrcpy(bs->node_name, sizeof(bs->node_name), node_name);
    QTAILQ_INSERT_TAIL(&graph_bdrv_states, bs, node_list);
}

770 771
/*
 * Common part for opening disk images and files
772 773
 *
 * Removes all processed options from *options.
774
 */
775
static int bdrv_open_common(BlockDriverState *bs, BlockDriverState *file,
776
    QDict *options, int flags, BlockDriver *drv, Error **errp)
777 778
{
    int ret, open_flags;
K
Kevin Wolf 已提交
779
    const char *filename;
780
    const char *node_name = NULL;
781
    Error *local_err = NULL;
782 783

    assert(drv != NULL);
784
    assert(bs->file == NULL);
785
    assert(options != NULL && bs->options != options);
786

787 788 789 790 791 792
    if (file != NULL) {
        filename = file->filename;
    } else {
        filename = qdict_get_try_str(options, "filename");
    }

793 794 795 796 797 798
    if (drv->bdrv_needs_filename && !filename) {
        error_setg(errp, "The '%s' block driver requires a file name",
                   drv->format_name);
        return -EINVAL;
    }

799
    trace_bdrv_open_common(bs, filename ?: "", flags, drv->format_name);
S
Stefan Hajnoczi 已提交
800

801
    node_name = qdict_get_try_str(options, "node-name");
802
    bdrv_assign_node_name(bs, node_name, &local_err);
803
    if (local_err) {
804 805
        error_propagate(errp, local_err);
        return -EINVAL;
806 807 808
    }
    qdict_del(options, "node-name");

809 810 811 812 813 814 815 816
    /* bdrv_open() with directly using a protocol as drv. This layer is already
     * opened, so assign it to bs (while file becomes a closed BlockDriverState)
     * and return immediately. */
    if (file != NULL && drv->bdrv_file_open) {
        bdrv_swap(file, bs);
        return 0;
    }

817
    bs->open_flags = flags;
818
    bs->guest_block_size = 512;
819
    bs->request_alignment = 512;
A
Asias He 已提交
820
    bs->zero_beyond_eof = true;
821 822 823 824
    open_flags = bdrv_open_flags(bs, flags);
    bs->read_only = !(open_flags & BDRV_O_RDWR);

    if (use_bdrv_whitelist && !bdrv_is_whitelisted(drv, bs->read_only)) {
825 826 827 828 829
        error_setg(errp,
                   !bs->read_only && bdrv_is_whitelisted(drv, true)
                        ? "Driver '%s' can only be used for read-only devices"
                        : "Driver '%s' is not whitelisted",
                   drv->format_name);
830 831
        return -ENOTSUP;
    }
832

833
    assert(bs->copy_on_read == 0); /* bdrv_new() and bdrv_close() make it so */
834 835 836 837 838 839 840
    if (flags & BDRV_O_COPY_ON_READ) {
        if (!bs->read_only) {
            bdrv_enable_copy_on_read(bs);
        } else {
            error_setg(errp, "Can't use copy-on-read on read-only device");
            return -EINVAL;
        }
841 842
    }

843 844 845 846 847
    if (filename != NULL) {
        pstrcpy(bs->filename, sizeof(bs->filename), filename);
    } else {
        bs->filename[0] = '\0';
    }
M
Max Reitz 已提交
848
    pstrcpy(bs->exact_filename, sizeof(bs->exact_filename), bs->filename);
849 850

    bs->drv = drv;
851
    bs->opaque = g_malloc0(drv->instance_size);
852

853
    bs->enable_write_cache = !!(flags & BDRV_O_CACHE_WB);
854

855 856
    /* Open the image, either directly or using a protocol */
    if (drv->bdrv_file_open) {
857
        assert(file == NULL);
858
        assert(!drv->bdrv_needs_filename || filename != NULL);
859
        ret = drv->bdrv_file_open(bs, options, open_flags, &local_err);
860
    } else {
861
        if (file == NULL) {
862 863
            error_setg(errp, "Can't use '%s' as a block driver for the "
                       "protocol level", drv->format_name);
864 865 866
            ret = -EINVAL;
            goto free_and_fail;
        }
867
        bs->file = file;
868
        ret = drv->bdrv_open(bs, options, open_flags, &local_err);
869 870
    }

871
    if (ret < 0) {
872
        if (local_err) {
873
            error_propagate(errp, local_err);
874 875
        } else if (bs->filename[0]) {
            error_setg_errno(errp, -ret, "Could not open '%s'", bs->filename);
876 877 878
        } else {
            error_setg_errno(errp, -ret, "Could not open image");
        }
879 880 881
        goto free_and_fail;
    }

882 883 884 885 886 887 888
    if (bs->encrypted) {
        error_report("Encrypted images are deprecated");
        error_printf("Support for them will be removed in a future release.\n"
                     "You can use 'qemu-img convert' to convert your image"
                     " to an unencrypted one.\n");
    }

889 890
    ret = refresh_total_sectors(bs, bs->total_sectors);
    if (ret < 0) {
891
        error_setg_errno(errp, -ret, "Could not refresh total sector count");
892
        goto free_and_fail;
893
    }
894

895 896 897 898 899 900 901
    bdrv_refresh_limits(bs, &local_err);
    if (local_err) {
        error_propagate(errp, local_err);
        ret = -EINVAL;
        goto free_and_fail;
    }

902
    assert(bdrv_opt_mem_align(bs) != 0);
903
    assert(bdrv_min_mem_align(bs) != 0);
904
    assert((bs->request_alignment != 0) || bs->sg);
905 906 907
    return 0;

free_and_fail:
908
    bs->file = NULL;
909
    g_free(bs->opaque);
910 911 912 913 914
    bs->opaque = NULL;
    bs->drv = NULL;
    return ret;
}

915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941
static QDict *parse_json_filename(const char *filename, Error **errp)
{
    QObject *options_obj;
    QDict *options;
    int ret;

    ret = strstart(filename, "json:", &filename);
    assert(ret);

    options_obj = qobject_from_json(filename);
    if (!options_obj) {
        error_setg(errp, "Could not parse the JSON options");
        return NULL;
    }

    if (qobject_type(options_obj) != QTYPE_QDICT) {
        qobject_decref(options_obj);
        error_setg(errp, "Invalid JSON object given");
        return NULL;
    }

    options = qobject_to_qdict(options_obj);
    qdict_flatten(options);

    return options;
}

K
Kevin Wolf 已提交
942
/*
K
Kevin Wolf 已提交
943 944
 * Fills in default options for opening images and converts the legacy
 * filename/flags pair to option QDict entries.
K
Kevin Wolf 已提交
945
 */
946
static int bdrv_fill_options(QDict **options, const char **pfilename, int flags,
947
                             BlockDriver *drv, Error **errp)
B
bellard 已提交
948
{
949
    const char *filename = *pfilename;
950
    const char *drvname;
951
    bool protocol = flags & BDRV_O_PROTOCOL;
952
    bool parse_filename = false;
953
    Error *local_err = NULL;
B
bellard 已提交
954

955 956 957 958 959 960 961 962 963 964 965 966 967 968 969
    /* Parse json: pseudo-protocol */
    if (filename && g_str_has_prefix(filename, "json:")) {
        QDict *json_options = parse_json_filename(filename, &local_err);
        if (local_err) {
            error_propagate(errp, local_err);
            return -EINVAL;
        }

        /* Options given in the filename have lower priority than options
         * specified directly */
        qdict_join(*options, json_options, false);
        QDECREF(json_options);
        *pfilename = filename = NULL;
    }

K
Kevin Wolf 已提交
970
    /* Fetch the file name from the options QDict if necessary */
971
    if (protocol && filename) {
K
Kevin Wolf 已提交
972 973 974 975 976 977 978 979
        if (!qdict_haskey(*options, "filename")) {
            qdict_put(*options, "filename", qstring_from_str(filename));
            parse_filename = true;
        } else {
            error_setg(errp, "Can't specify 'file' and 'filename' options at "
                             "the same time");
            return -EINVAL;
        }
K
Kevin Wolf 已提交
980 981
    }

982
    /* Find the right block driver */
K
Kevin Wolf 已提交
983
    filename = qdict_get_try_str(*options, "filename");
984
    drvname = qdict_get_try_str(*options, "driver");
K
Kevin Wolf 已提交
985

986 987 988 989 990 991 992 993 994 995
    if (drv) {
        if (drvname) {
            error_setg(errp, "Driver specified twice");
            return -EINVAL;
        }
        drvname = drv->format_name;
        qdict_put(*options, "driver", qstring_from_str(drvname));
    } else {
        if (!drvname && protocol) {
            if (filename) {
996
                drv = bdrv_find_protocol(filename, parse_filename, errp);
997 998 999 1000 1001 1002 1003 1004
                if (!drv) {
                    return -EINVAL;
                }

                drvname = drv->format_name;
                qdict_put(*options, "driver", qstring_from_str(drvname));
            } else {
                error_setg(errp, "Must specify either driver or file");
K
Kevin Wolf 已提交
1005 1006
                return -EINVAL;
            }
1007 1008 1009 1010 1011 1012
        } else if (drvname) {
            drv = bdrv_find_format(drvname);
            if (!drv) {
                error_setg(errp, "Unknown driver '%s'", drvname);
                return -ENOENT;
            }
1013
        }
1014 1015
    }

1016
    assert(drv || !protocol);
1017

K
Kevin Wolf 已提交
1018
    /* Driver-specific filename parsing */
1019
    if (drv && drv->bdrv_parse_filename && parse_filename) {
1020
        drv->bdrv_parse_filename(filename, *options, &local_err);
1021
        if (local_err) {
1022
            error_propagate(errp, local_err);
K
Kevin Wolf 已提交
1023
            return -EINVAL;
1024
        }
1025 1026 1027 1028

        if (!drv->bdrv_needs_filename) {
            qdict_del(*options, "filename");
        }
1029 1030
    }

K
Kevin Wolf 已提交
1031 1032 1033
    return 0;
}

F
Fam Zheng 已提交
1034 1035 1036
void bdrv_set_backing_hd(BlockDriverState *bs, BlockDriverState *backing_hd)
{

1037 1038 1039 1040 1041
    if (bs->backing_hd) {
        assert(bs->backing_blocker);
        bdrv_op_unblock_all(bs->backing_hd, bs->backing_blocker);
    } else if (backing_hd) {
        error_setg(&bs->backing_blocker,
1042 1043
                   "node is used as backing hd of '%s'",
                   bdrv_get_device_or_node_name(bs));
1044 1045
    }

F
Fam Zheng 已提交
1046 1047
    bs->backing_hd = backing_hd;
    if (!backing_hd) {
1048 1049
        error_free(bs->backing_blocker);
        bs->backing_blocker = NULL;
F
Fam Zheng 已提交
1050 1051 1052 1053 1054 1055
        goto out;
    }
    bs->open_flags &= ~BDRV_O_NO_BACKING;
    pstrcpy(bs->backing_file, sizeof(bs->backing_file), backing_hd->filename);
    pstrcpy(bs->backing_format, sizeof(bs->backing_format),
            backing_hd->drv ? backing_hd->drv->format_name : "");
1056 1057 1058

    bdrv_op_block_all(bs->backing_hd, bs->backing_blocker);
    /* Otherwise we won't be able to commit due to check in bdrv_commit */
1059
    bdrv_op_unblock(bs->backing_hd, BLOCK_OP_TYPE_COMMIT_TARGET,
1060
                    bs->backing_blocker);
F
Fam Zheng 已提交
1061
out:
1062
    bdrv_refresh_limits(bs, NULL);
F
Fam Zheng 已提交
1063 1064
}

1065 1066 1067 1068 1069 1070 1071 1072
/*
 * Opens the backing file for a BlockDriverState if not yet open
 *
 * options is a QDict of options to pass to the block drivers, or NULL for an
 * empty set of options. The reference to the QDict is transferred to this
 * function (even on failure), so if the caller intends to reuse the dictionary,
 * it needs to use QINCREF() before calling bdrv_file_open.
 */
1073
int bdrv_open_backing_file(BlockDriverState *bs, QDict *options, Error **errp)
P
Paolo Bonzini 已提交
1074
{
1075
    char *backing_filename = g_malloc0(PATH_MAX);
K
Kevin Wolf 已提交
1076
    int ret = 0;
F
Fam Zheng 已提交
1077
    BlockDriverState *backing_hd;
1078
    Error *local_err = NULL;
P
Paolo Bonzini 已提交
1079 1080

    if (bs->backing_hd != NULL) {
1081
        QDECREF(options);
1082
        goto free_exit;
P
Paolo Bonzini 已提交
1083 1084
    }

1085 1086 1087 1088 1089
    /* NULL means an empty set of options */
    if (options == NULL) {
        options = qdict_new();
    }

P
Paolo Bonzini 已提交
1090
    bs->open_flags &= ~BDRV_O_NO_BACKING;
1091 1092 1093
    if (qdict_haskey(options, "file.filename")) {
        backing_filename[0] = '\0';
    } else if (bs->backing_file[0] == '\0' && qdict_size(options) == 0) {
1094
        QDECREF(options);
1095
        goto free_exit;
F
Fam Zheng 已提交
1096
    } else {
1097 1098 1099 1100 1101 1102 1103 1104
        bdrv_get_full_backing_filename(bs, backing_filename, PATH_MAX,
                                       &local_err);
        if (local_err) {
            ret = -EINVAL;
            error_propagate(errp, local_err);
            QDECREF(options);
            goto free_exit;
        }
P
Paolo Bonzini 已提交
1105 1106
    }

1107 1108 1109 1110 1111 1112 1113
    if (!bs->drv || !bs->drv->supports_backing) {
        ret = -EINVAL;
        error_setg(errp, "Driver doesn't support backing files");
        QDECREF(options);
        goto free_exit;
    }

1114
    backing_hd = bdrv_new();
F
Fam Zheng 已提交
1115

1116 1117
    if (bs->backing_format[0] != '\0' && !qdict_haskey(options, "driver")) {
        qdict_put(options, "driver", qstring_from_str(bs->backing_format));
P
Paolo Bonzini 已提交
1118 1119
    }

1120
    assert(bs->backing_hd == NULL);
F
Fam Zheng 已提交
1121
    ret = bdrv_open(&backing_hd,
1122
                    *backing_filename ? backing_filename : NULL, NULL, options,
1123
                    bdrv_backing_flags(bs->open_flags), NULL, &local_err);
P
Paolo Bonzini 已提交
1124
    if (ret < 0) {
F
Fam Zheng 已提交
1125 1126
        bdrv_unref(backing_hd);
        backing_hd = NULL;
P
Paolo Bonzini 已提交
1127
        bs->open_flags |= BDRV_O_NO_BACKING;
1128 1129 1130
        error_setg(errp, "Could not open backing file: %s",
                   error_get_pretty(local_err));
        error_free(local_err);
1131
        goto free_exit;
P
Paolo Bonzini 已提交
1132
    }
F
Fam Zheng 已提交
1133
    bdrv_set_backing_hd(bs, backing_hd);
P
Peter Feiner 已提交
1134

1135 1136 1137
free_exit:
    g_free(backing_filename);
    return ret;
P
Paolo Bonzini 已提交
1138 1139
}

M
Max Reitz 已提交
1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152
/*
 * Opens a disk image whose options are given as BlockdevRef in another block
 * device's options.
 *
 * If allow_none is true, no image will be opened if filename is false and no
 * BlockdevRef is given. *pbs will remain unchanged and 0 will be returned.
 *
 * bdrev_key specifies the key for the image's BlockdevRef in the options QDict.
 * That QDict has to be flattened; therefore, if the BlockdevRef is a QDict
 * itself, all options starting with "${bdref_key}." are considered part of the
 * BlockdevRef.
 *
 * The BlockdevRef will be removed from the options QDict.
1153 1154
 *
 * To conform with the behavior of bdrv_open(), *pbs has to be NULL.
M
Max Reitz 已提交
1155 1156 1157
 */
int bdrv_open_image(BlockDriverState **pbs, const char *filename,
                    QDict *options, const char *bdref_key, int flags,
1158
                    bool allow_none, Error **errp)
M
Max Reitz 已提交
1159 1160 1161 1162 1163 1164
{
    QDict *image_options;
    int ret;
    char *bdref_key_dot;
    const char *reference;

1165 1166 1167
    assert(pbs);
    assert(*pbs == NULL);

M
Max Reitz 已提交
1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180
    bdref_key_dot = g_strdup_printf("%s.", bdref_key);
    qdict_extract_subqdict(options, &image_options, bdref_key_dot);
    g_free(bdref_key_dot);

    reference = qdict_get_try_str(options, bdref_key);
    if (!filename && !reference && !qdict_size(image_options)) {
        if (allow_none) {
            ret = 0;
        } else {
            error_setg(errp, "A block device must be specified for \"%s\"",
                       bdref_key);
            ret = -EINVAL;
        }
1181
        QDECREF(image_options);
M
Max Reitz 已提交
1182 1183 1184
        goto done;
    }

1185
    ret = bdrv_open(pbs, filename, reference, image_options, flags, NULL, errp);
M
Max Reitz 已提交
1186 1187 1188 1189 1190 1191

done:
    qdict_del(options, bdref_key);
    return ret;
}

1192
int bdrv_append_temp_snapshot(BlockDriverState *bs, int flags, Error **errp)
1193 1194
{
    /* TODO: extra byte is a hack to ensure MAX_PATH space on Windows. */
1195
    char *tmp_filename = g_malloc0(PATH_MAX + 1);
1196
    int64_t total_size;
1197
    QemuOpts *opts = NULL;
1198 1199 1200 1201 1202 1203 1204 1205 1206
    QDict *snapshot_options;
    BlockDriverState *bs_snapshot;
    Error *local_err;
    int ret;

    /* if snapshot, we create a temporary backing file and open it
       instead of opening 'filename' directly */

    /* Get the required size from the image */
1207 1208
    total_size = bdrv_getlength(bs);
    if (total_size < 0) {
1209
        ret = total_size;
1210
        error_setg_errno(errp, -total_size, "Could not get image size");
1211
        goto out;
1212
    }
1213 1214

    /* Create the temporary image */
1215
    ret = get_tmp_filename(tmp_filename, PATH_MAX + 1);
1216 1217
    if (ret < 0) {
        error_setg_errno(errp, -ret, "Could not get temporary filename");
1218
        goto out;
1219 1220
    }

1221
    opts = qemu_opts_create(bdrv_qcow2.create_opts, NULL, 0,
C
Chunyan Liu 已提交
1222
                            &error_abort);
1223
    qemu_opt_set_number(opts, BLOCK_OPT_SIZE, total_size, &error_abort);
1224
    ret = bdrv_create(&bdrv_qcow2, tmp_filename, opts, &local_err);
1225
    qemu_opts_del(opts);
1226 1227 1228 1229 1230
    if (ret < 0) {
        error_setg_errno(errp, -ret, "Could not create temporary overlay "
                         "'%s': %s", tmp_filename,
                         error_get_pretty(local_err));
        error_free(local_err);
1231
        goto out;
1232 1233 1234 1235 1236 1237 1238 1239 1240
    }

    /* Prepare a new options QDict for the temporary file */
    snapshot_options = qdict_new();
    qdict_put(snapshot_options, "file.driver",
              qstring_from_str("file"));
    qdict_put(snapshot_options, "file.filename",
              qstring_from_str(tmp_filename));

1241
    bs_snapshot = bdrv_new();
1242 1243

    ret = bdrv_open(&bs_snapshot, NULL, NULL, snapshot_options,
1244
                    flags, &bdrv_qcow2, &local_err);
1245 1246
    if (ret < 0) {
        error_propagate(errp, local_err);
1247
        goto out;
1248 1249 1250
    }

    bdrv_append(bs_snapshot, bs);
1251 1252 1253

out:
    g_free(tmp_filename);
1254
    return ret;
1255 1256
}

K
Kevin Wolf 已提交
1257 1258
/*
 * Opens a disk image (raw, qcow2, vmdk, ...)
1259 1260 1261 1262 1263
 *
 * options is a QDict of options to pass to the block drivers, or NULL for an
 * empty set of options. The reference to the QDict belongs to the block layer
 * after the call (even on failure), so if the caller intends to reuse the
 * dictionary, it needs to use QINCREF() before calling bdrv_open.
1264 1265 1266
 *
 * If *pbs is NULL, a new BDS will be created with a pointer to it stored there.
 * If it is not NULL, the referenced BDS will be reused.
1267 1268 1269 1270
 *
 * The reference parameter may be used to specify an existing block device which
 * should be opened. If specified, neither options nor a filename may be given,
 * nor can an existing BDS be reused (that is, *pbs has to be NULL).
K
Kevin Wolf 已提交
1271
 */
1272 1273 1274
int bdrv_open(BlockDriverState **pbs, const char *filename,
              const char *reference, QDict *options, int flags,
              BlockDriver *drv, Error **errp)
B
bellard 已提交
1275
{
K
Kevin Wolf 已提交
1276
    int ret;
1277
    BlockDriverState *file = NULL, *bs;
1278
    const char *drvname;
1279
    Error *local_err = NULL;
1280
    int snapshot_flags = 0;
B
bellard 已提交
1281

1282 1283
    assert(pbs);

1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308
    if (reference) {
        bool options_non_empty = options ? qdict_size(options) : false;
        QDECREF(options);

        if (*pbs) {
            error_setg(errp, "Cannot reuse an existing BDS when referencing "
                       "another block device");
            return -EINVAL;
        }

        if (filename || options_non_empty) {
            error_setg(errp, "Cannot reference an existing block device with "
                       "additional options or a new filename");
            return -EINVAL;
        }

        bs = bdrv_lookup_bs(reference, reference, errp);
        if (!bs) {
            return -ENODEV;
        }
        bdrv_ref(bs);
        *pbs = bs;
        return 0;
    }

1309 1310 1311
    if (*pbs) {
        bs = *pbs;
    } else {
1312
        bs = bdrv_new();
1313 1314
    }

1315 1316 1317 1318 1319
    /* NULL means an empty set of options */
    if (options == NULL) {
        options = qdict_new();
    }

1320
    ret = bdrv_fill_options(&options, &filename, flags, drv, &local_err);
1321 1322 1323 1324
    if (local_err) {
        goto fail;
    }

1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344
    /* Find the right image format driver */
    drv = NULL;
    drvname = qdict_get_try_str(options, "driver");
    if (drvname) {
        drv = bdrv_find_format(drvname);
        qdict_del(options, "driver");
        if (!drv) {
            error_setg(errp, "Unknown driver: '%s'", drvname);
            ret = -EINVAL;
            goto fail;
        }
    }

    assert(drvname || !(flags & BDRV_O_PROTOCOL));
    if (drv && !drv->bdrv_file_open) {
        /* If the user explicitly wants a format driver here, we'll need to add
         * another layer for the protocol in bs->file */
        flags &= ~BDRV_O_PROTOCOL;
    }

1345
    bs->options = options;
1346
    options = qdict_clone_shallow(options);
1347

1348
    /* Open image file without format layer */
1349 1350 1351 1352 1353 1354 1355 1356
    if ((flags & BDRV_O_PROTOCOL) == 0) {
        if (flags & BDRV_O_RDWR) {
            flags |= BDRV_O_ALLOW_RDWR;
        }
        if (flags & BDRV_O_SNAPSHOT) {
            snapshot_flags = bdrv_temp_snapshot_flags(flags);
            flags = bdrv_backing_flags(flags);
        }
1357

1358 1359 1360 1361 1362 1363 1364
        assert(file == NULL);
        ret = bdrv_open_image(&file, filename, options, "file",
                              bdrv_inherited_flags(flags),
                              true, &local_err);
        if (ret < 0) {
            goto fail;
        }
1365 1366
    }

1367
    /* Image format probing */
1368
    bs->probed = !drv;
1369
    if (!drv && file) {
1370 1371
        ret = find_image_format(file, filename, &drv, &local_err);
        if (ret < 0) {
1372
            goto fail;
1373
        }
1374
    } else if (!drv) {
1375 1376
        error_setg(errp, "Must specify either driver or file");
        ret = -EINVAL;
1377
        goto fail;
B
bellard 已提交
1378
    }
K
Kevin Wolf 已提交
1379 1380

    /* Open the image */
1381
    ret = bdrv_open_common(bs, file, options, flags, drv, &local_err);
K
Kevin Wolf 已提交
1382
    if (ret < 0) {
1383
        goto fail;
1384 1385
    }

1386
    if (file && (bs->file != file)) {
F
Fam Zheng 已提交
1387
        bdrv_unref(file);
1388 1389 1390
        file = NULL;
    }

K
Kevin Wolf 已提交
1391
    /* If there is a backing file, use it */
P
Paolo Bonzini 已提交
1392
    if ((flags & BDRV_O_NO_BACKING) == 0) {
1393 1394
        QDict *backing_options;

1395
        qdict_extract_subqdict(options, &backing_options, "backing.");
1396
        ret = bdrv_open_backing_file(bs, backing_options, &local_err);
K
Kevin Wolf 已提交
1397
        if (ret < 0) {
1398
            goto close_and_fail;
K
Kevin Wolf 已提交
1399 1400 1401
        }
    }

M
Max Reitz 已提交
1402 1403
    bdrv_refresh_filename(bs);

1404 1405
    /* For snapshot=on, create a temporary qcow2 overlay. bs points to the
     * temporary snapshot afterwards. */
1406
    if (snapshot_flags) {
1407
        ret = bdrv_append_temp_snapshot(bs, snapshot_flags, &local_err);
1408 1409 1410 1411 1412
        if (local_err) {
            goto close_and_fail;
        }
    }

1413
    /* Check if any unknown options were used */
1414
    if (options && (qdict_size(options) != 0)) {
1415
        const QDictEntry *entry = qdict_first(options);
1416 1417 1418 1419 1420 1421
        if (flags & BDRV_O_PROTOCOL) {
            error_setg(errp, "Block protocol '%s' doesn't support the option "
                       "'%s'", drv->format_name, entry->key);
        } else {
            error_setg(errp, "Block format '%s' used by device '%s' doesn't "
                       "support the option '%s'", drv->format_name,
1422
                       bdrv_get_device_name(bs), entry->key);
1423
        }
1424 1425 1426 1427 1428

        ret = -EINVAL;
        goto close_and_fail;
    }

K
Kevin Wolf 已提交
1429
    if (!bdrv_key_required(bs)) {
1430 1431 1432
        if (bs->blk) {
            blk_dev_change_media_cb(bs->blk, true);
        }
1433 1434 1435 1436 1437 1438 1439
    } else if (!runstate_check(RUN_STATE_PRELAUNCH)
               && !runstate_check(RUN_STATE_INMIGRATE)
               && !runstate_check(RUN_STATE_PAUSED)) { /* HACK */
        error_setg(errp,
                   "Guest must be stopped for opening of encrypted image");
        ret = -EBUSY;
        goto close_and_fail;
K
Kevin Wolf 已提交
1440 1441
    }

1442
    QDECREF(options);
1443
    *pbs = bs;
K
Kevin Wolf 已提交
1444 1445
    return 0;

1446
fail:
1447
    if (file != NULL) {
F
Fam Zheng 已提交
1448
        bdrv_unref(file);
1449
    }
1450
    QDECREF(bs->options);
1451
    QDECREF(options);
1452
    bs->options = NULL;
1453 1454 1455 1456 1457 1458
    if (!*pbs) {
        /* If *pbs is NULL, a new BDS has been created in this function and
           needs to be freed now. Otherwise, it does not need to be closed,
           since it has not really been opened yet. */
        bdrv_unref(bs);
    }
1459
    if (local_err) {
1460 1461
        error_propagate(errp, local_err);
    }
1462
    return ret;
1463

1464
close_and_fail:
1465 1466 1467 1468 1469 1470
    /* See fail path, but now the BDS has to be always closed */
    if (*pbs) {
        bdrv_close(bs);
    } else {
        bdrv_unref(bs);
    }
1471
    QDECREF(options);
1472
    if (local_err) {
1473 1474
        error_propagate(errp, local_err);
    }
K
Kevin Wolf 已提交
1475 1476 1477
    return ret;
}

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
typedef struct BlockReopenQueueEntry {
     bool prepared;
     BDRVReopenState state;
     QSIMPLEQ_ENTRY(BlockReopenQueueEntry) entry;
} BlockReopenQueueEntry;

/*
 * Adds a BlockDriverState to a simple queue for an atomic, transactional
 * reopen of multiple devices.
 *
 * bs_queue can either be an existing BlockReopenQueue that has had QSIMPLE_INIT
 * already performed, or alternatively may be NULL a new BlockReopenQueue will
 * be created and initialized. This newly created BlockReopenQueue should be
 * passed back in for subsequent calls that are intended to be of the same
 * atomic 'set'.
 *
 * bs is the BlockDriverState to add to the reopen queue.
 *
 * flags contains the open flags for the associated bs
 *
 * returns a pointer to bs_queue, which is either the newly allocated
 * bs_queue, or the existing bs_queue being used.
 *
 */
BlockReopenQueue *bdrv_reopen_queue(BlockReopenQueue *bs_queue,
                                    BlockDriverState *bs, int flags)
{
    assert(bs != NULL);

    BlockReopenQueueEntry *bs_entry;
    if (bs_queue == NULL) {
        bs_queue = g_new0(BlockReopenQueue, 1);
        QSIMPLEQ_INIT(bs_queue);
    }

1513 1514 1515
    /* bdrv_open() masks this flag out */
    flags &= ~BDRV_O_PROTOCOL;

1516
    if (bs->file) {
1517
        bdrv_reopen_queue(bs_queue, bs->file, bdrv_inherited_flags(flags));
1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629
    }

    bs_entry = g_new0(BlockReopenQueueEntry, 1);
    QSIMPLEQ_INSERT_TAIL(bs_queue, bs_entry, entry);

    bs_entry->state.bs = bs;
    bs_entry->state.flags = flags;

    return bs_queue;
}

/*
 * Reopen multiple BlockDriverStates atomically & transactionally.
 *
 * The queue passed in (bs_queue) must have been built up previous
 * via bdrv_reopen_queue().
 *
 * Reopens all BDS specified in the queue, with the appropriate
 * flags.  All devices are prepared for reopen, and failure of any
 * device will cause all device changes to be abandonded, and intermediate
 * data cleaned up.
 *
 * If all devices prepare successfully, then the changes are committed
 * to all devices.
 *
 */
int bdrv_reopen_multiple(BlockReopenQueue *bs_queue, Error **errp)
{
    int ret = -1;
    BlockReopenQueueEntry *bs_entry, *next;
    Error *local_err = NULL;

    assert(bs_queue != NULL);

    bdrv_drain_all();

    QSIMPLEQ_FOREACH(bs_entry, bs_queue, entry) {
        if (bdrv_reopen_prepare(&bs_entry->state, bs_queue, &local_err)) {
            error_propagate(errp, local_err);
            goto cleanup;
        }
        bs_entry->prepared = true;
    }

    /* If we reach this point, we have success and just need to apply the
     * changes
     */
    QSIMPLEQ_FOREACH(bs_entry, bs_queue, entry) {
        bdrv_reopen_commit(&bs_entry->state);
    }

    ret = 0;

cleanup:
    QSIMPLEQ_FOREACH_SAFE(bs_entry, bs_queue, entry, next) {
        if (ret && bs_entry->prepared) {
            bdrv_reopen_abort(&bs_entry->state);
        }
        g_free(bs_entry);
    }
    g_free(bs_queue);
    return ret;
}


/* Reopen a single BlockDriverState with the specified flags. */
int bdrv_reopen(BlockDriverState *bs, int bdrv_flags, Error **errp)
{
    int ret = -1;
    Error *local_err = NULL;
    BlockReopenQueue *queue = bdrv_reopen_queue(NULL, bs, bdrv_flags);

    ret = bdrv_reopen_multiple(queue, &local_err);
    if (local_err != NULL) {
        error_propagate(errp, local_err);
    }
    return ret;
}


/*
 * Prepares a BlockDriverState for reopen. All changes are staged in the
 * 'opaque' field of the BDRVReopenState, which is used and allocated by
 * the block driver layer .bdrv_reopen_prepare()
 *
 * bs is the BlockDriverState to reopen
 * flags are the new open flags
 * queue is the reopen queue
 *
 * Returns 0 on success, non-zero on error.  On error errp will be set
 * as well.
 *
 * On failure, bdrv_reopen_abort() will be called to clean up any data.
 * It is the responsibility of the caller to then call the abort() or
 * commit() for any other BDS that have been left in a prepare() state
 *
 */
int bdrv_reopen_prepare(BDRVReopenState *reopen_state, BlockReopenQueue *queue,
                        Error **errp)
{
    int ret = -1;
    Error *local_err = NULL;
    BlockDriver *drv;

    assert(reopen_state != NULL);
    assert(reopen_state->bs->drv != NULL);
    drv = reopen_state->bs->drv;

    /* if we are to stay read-only, do not allow permission change
     * to r/w */
    if (!(reopen_state->bs->open_flags & BDRV_O_ALLOW_RDWR) &&
        reopen_state->flags & BDRV_O_RDWR) {
1630 1631
        error_setg(errp, "Node '%s' is read only",
                   bdrv_get_device_or_node_name(reopen_state->bs));
1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648
        goto error;
    }


    ret = bdrv_flush(reopen_state->bs);
    if (ret) {
        error_set(errp, ERROR_CLASS_GENERIC_ERROR, "Error (%s) flushing drive",
                  strerror(-ret));
        goto error;
    }

    if (drv->bdrv_reopen_prepare) {
        ret = drv->bdrv_reopen_prepare(reopen_state, queue, &local_err);
        if (ret) {
            if (local_err != NULL) {
                error_propagate(errp, local_err);
            } else {
1649 1650
                error_setg(errp, "failed while preparing to reopen image '%s'",
                           reopen_state->bs->filename);
1651 1652 1653 1654 1655 1656
            }
            goto error;
        }
    } else {
        /* It is currently mandatory to have a bdrv_reopen_prepare()
         * handler for each supported drv. */
1657 1658 1659
        error_setg(errp, "Block format '%s' used by node '%s' "
                   "does not support reopening files", drv->format_name,
                   bdrv_get_device_or_node_name(reopen_state->bs));
1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692
        ret = -1;
        goto error;
    }

    ret = 0;

error:
    return ret;
}

/*
 * Takes the staged changes for the reopen from bdrv_reopen_prepare(), and
 * makes them final by swapping the staging BlockDriverState contents into
 * the active BlockDriverState contents.
 */
void bdrv_reopen_commit(BDRVReopenState *reopen_state)
{
    BlockDriver *drv;

    assert(reopen_state != NULL);
    drv = reopen_state->bs->drv;
    assert(drv != NULL);

    /* If there are any driver level actions to take */
    if (drv->bdrv_reopen_commit) {
        drv->bdrv_reopen_commit(reopen_state);
    }

    /* set BDS specific flags now */
    reopen_state->bs->open_flags         = reopen_state->flags;
    reopen_state->bs->enable_write_cache = !!(reopen_state->flags &
                                              BDRV_O_CACHE_WB);
    reopen_state->bs->read_only = !(reopen_state->flags & BDRV_O_RDWR);
1693

1694
    bdrv_refresh_limits(reopen_state->bs, NULL);
1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714
}

/*
 * Abort the reopen, and delete and free the staged changes in
 * reopen_state
 */
void bdrv_reopen_abort(BDRVReopenState *reopen_state)
{
    BlockDriver *drv;

    assert(reopen_state != NULL);
    drv = reopen_state->bs->drv;
    assert(drv != NULL);

    if (drv->bdrv_reopen_abort) {
        drv->bdrv_reopen_abort(reopen_state);
    }
}


B
bellard 已提交
1715 1716
void bdrv_close(BlockDriverState *bs)
{
M
Max Reitz 已提交
1717 1718
    BdrvAioNotifier *ban, *ban_next;

1719 1720 1721
    if (bs->job) {
        block_job_cancel_sync(bs->job);
    }
1722 1723 1724
    bdrv_drain_all(); /* complete I/O */
    bdrv_flush(bs);
    bdrv_drain_all(); /* in case flush left pending I/O */
P
Paolo Bonzini 已提交
1725
    notifier_list_notify(&bs->close_notifiers, bs);
K
Kevin Wolf 已提交
1726

1727
    if (bs->drv) {
1728
        if (bs->backing_hd) {
1729 1730 1731
            BlockDriverState *backing_hd = bs->backing_hd;
            bdrv_set_backing_hd(bs, NULL);
            bdrv_unref(backing_hd);
1732
        }
B
bellard 已提交
1733
        bs->drv->bdrv_close(bs);
1734
        g_free(bs->opaque);
B
bellard 已提交
1735 1736
        bs->opaque = NULL;
        bs->drv = NULL;
1737
        bs->copy_on_read = 0;
1738 1739
        bs->backing_file[0] = '\0';
        bs->backing_format[0] = '\0';
1740 1741 1742 1743
        bs->total_sectors = 0;
        bs->encrypted = 0;
        bs->valid_key = 0;
        bs->sg = 0;
A
Asias He 已提交
1744
        bs->zero_beyond_eof = false;
1745 1746
        QDECREF(bs->options);
        bs->options = NULL;
M
Max Reitz 已提交
1747 1748
        QDECREF(bs->full_open_options);
        bs->full_open_options = NULL;
B
bellard 已提交
1749

1750
        if (bs->file != NULL) {
F
Fam Zheng 已提交
1751
            bdrv_unref(bs->file);
1752
            bs->file = NULL;
1753
        }
B
bellard 已提交
1754
    }
Z
Zhi Yong Wu 已提交
1755

1756 1757 1758
    if (bs->blk) {
        blk_dev_change_media_cb(bs->blk, false);
    }
P
Pavel Hrdina 已提交
1759

Z
Zhi Yong Wu 已提交
1760 1761 1762 1763
    /*throttling disk I/O limits*/
    if (bs->io_limits_enabled) {
        bdrv_io_limits_disable(bs);
    }
M
Max Reitz 已提交
1764 1765 1766 1767 1768

    QLIST_FOREACH_SAFE(ban, &bs->aio_notifiers, list, ban_next) {
        g_free(ban);
    }
    QLIST_INIT(&bs->aio_notifiers);
B
bellard 已提交
1769 1770
}

1771 1772 1773 1774
void bdrv_close_all(void)
{
    BlockDriverState *bs;

1775
    QTAILQ_FOREACH(bs, &bdrv_states, device_list) {
1776 1777 1778
        AioContext *aio_context = bdrv_get_aio_context(bs);

        aio_context_acquire(aio_context);
1779
        bdrv_close(bs);
1780
        aio_context_release(aio_context);
1781 1782 1783
    }
}

1784 1785
/* make a BlockDriverState anonymous by removing from bdrv_state and
 * graph_bdrv_state list.
1786 1787 1788
   Also, NULL terminate the device_name to prevent double remove */
void bdrv_make_anon(BlockDriverState *bs)
{
1789 1790 1791 1792 1793 1794 1795 1796
    /*
     * Take care to remove bs from bdrv_states only when it's actually
     * in it.  Note that bs->device_list.tqe_prev is initially null,
     * and gets set to non-null by QTAILQ_INSERT_TAIL().  Establish
     * the useful invariant "bs in bdrv_states iff bs->tqe_prev" by
     * resetting it to null on remove.
     */
    if (bs->device_list.tqe_prev) {
1797
        QTAILQ_REMOVE(&bdrv_states, bs, device_list);
1798
        bs->device_list.tqe_prev = NULL;
1799
    }
1800 1801 1802 1803
    if (bs->node_name[0] != '\0') {
        QTAILQ_REMOVE(&graph_bdrv_states, bs, node_list);
    }
    bs->node_name[0] = '\0';
1804 1805
}

P
Paolo Bonzini 已提交
1806 1807 1808 1809 1810 1811 1812
static void bdrv_rebind(BlockDriverState *bs)
{
    if (bs->drv && bs->drv->bdrv_rebind) {
        bs->drv->bdrv_rebind(bs);
    }
}

1813 1814
static void bdrv_move_feature_fields(BlockDriverState *bs_dest,
                                     BlockDriverState *bs_src)
1815
{
1816
    /* move some fields that need to stay attached to the device */
1817 1818

    /* dev info */
1819
    bs_dest->guest_block_size   = bs_src->guest_block_size;
1820
    bs_dest->copy_on_read       = bs_src->copy_on_read;
1821

1822
    bs_dest->enable_write_cache = bs_src->enable_write_cache;
1823

1824 1825 1826 1827 1828 1829
    /* i/o throttled req */
    memcpy(&bs_dest->throttle_state,
           &bs_src->throttle_state,
           sizeof(ThrottleState));
    bs_dest->throttled_reqs[0]  = bs_src->throttled_reqs[0];
    bs_dest->throttled_reqs[1]  = bs_src->throttled_reqs[1];
1830
    bs_dest->io_limits_enabled  = bs_src->io_limits_enabled;
1831 1832

    /* r/w error */
1833 1834
    bs_dest->on_read_error      = bs_src->on_read_error;
    bs_dest->on_write_error     = bs_src->on_write_error;
1835 1836

    /* i/o status */
1837 1838
    bs_dest->iostatus_enabled   = bs_src->iostatus_enabled;
    bs_dest->iostatus           = bs_src->iostatus;
1839

1840
    /* dirty bitmap */
F
Fam Zheng 已提交
1841
    bs_dest->dirty_bitmaps      = bs_src->dirty_bitmaps;
1842

1843 1844 1845
    /* reference count */
    bs_dest->refcnt             = bs_src->refcnt;

1846
    /* job */
1847
    bs_dest->job                = bs_src->job;
1848

1849
    /* keep the same entry in bdrv_states */
1850
    bs_dest->device_list = bs_src->device_list;
1851 1852
    bs_dest->blk = bs_src->blk;

1853 1854
    memcpy(bs_dest->op_blockers, bs_src->op_blockers,
           sizeof(bs_dest->op_blockers));
1855
}
1856

1857 1858 1859 1860 1861 1862 1863 1864
/*
 * Swap bs contents for two image chains while they are live,
 * while keeping required fields on the BlockDriverState that is
 * actually attached to a device.
 *
 * This will modify the BlockDriverState fields, and swap contents
 * between bs_new and bs_old. Both bs_new and bs_old are modified.
 *
1865
 * bs_new must not be attached to a BlockBackend.
1866 1867 1868 1869 1870 1871
 *
 * This function does not create any image files.
 */
void bdrv_swap(BlockDriverState *bs_new, BlockDriverState *bs_old)
{
    BlockDriverState tmp;
J
Jeff Cody 已提交
1872

1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883
    /* The code needs to swap the node_name but simply swapping node_list won't
     * work so first remove the nodes from the graph list, do the swap then
     * insert them back if needed.
     */
    if (bs_new->node_name[0] != '\0') {
        QTAILQ_REMOVE(&graph_bdrv_states, bs_new, node_list);
    }
    if (bs_old->node_name[0] != '\0') {
        QTAILQ_REMOVE(&graph_bdrv_states, bs_old, node_list);
    }

1884
    /* bs_new must be unattached and shouldn't have anything fancy enabled */
1885
    assert(!bs_new->blk);
F
Fam Zheng 已提交
1886
    assert(QLIST_EMPTY(&bs_new->dirty_bitmaps));
1887 1888
    assert(bs_new->job == NULL);
    assert(bs_new->io_limits_enabled == false);
1889
    assert(!throttle_have_timer(&bs_new->throttle_state));
1890

1891 1892 1893
    tmp = *bs_new;
    *bs_new = *bs_old;
    *bs_old = tmp;
1894

1895 1896 1897 1898
    /* there are some fields that should not be swapped, move them back */
    bdrv_move_feature_fields(&tmp, bs_old);
    bdrv_move_feature_fields(bs_old, bs_new);
    bdrv_move_feature_fields(bs_new, &tmp);
1899

1900
    /* bs_new must remain unattached */
1901
    assert(!bs_new->blk);
1902 1903 1904 1905

    /* Check a few fields that should remain attached to the device */
    assert(bs_new->job == NULL);
    assert(bs_new->io_limits_enabled == false);
1906
    assert(!throttle_have_timer(&bs_new->throttle_state));
P
Paolo Bonzini 已提交
1907

1908 1909 1910 1911 1912 1913 1914 1915
    /* insert the nodes back into the graph node list if needed */
    if (bs_new->node_name[0] != '\0') {
        QTAILQ_INSERT_TAIL(&graph_bdrv_states, bs_new, node_list);
    }
    if (bs_old->node_name[0] != '\0') {
        QTAILQ_INSERT_TAIL(&graph_bdrv_states, bs_old, node_list);
    }

P
Paolo Bonzini 已提交
1916
    bdrv_rebind(bs_new);
1917 1918 1919 1920 1921 1922 1923 1924 1925 1926
    bdrv_rebind(bs_old);
}

/*
 * Add new bs contents at the top of an image chain while the chain is
 * live, while keeping required fields on the top layer.
 *
 * This will modify the BlockDriverState fields, and swap contents
 * between bs_new and bs_top. Both bs_new and bs_top are modified.
 *
1927
 * bs_new must not be attached to a BlockBackend.
1928 1929 1930 1931 1932 1933 1934 1935 1936
 *
 * This function does not create any image files.
 */
void bdrv_append(BlockDriverState *bs_new, BlockDriverState *bs_top)
{
    bdrv_swap(bs_new, bs_top);

    /* The contents of 'tmp' will become bs_top, as we are
     * swapping bs_new and bs_top contents. */
F
Fam Zheng 已提交
1937
    bdrv_set_backing_hd(bs_top, bs_new);
1938 1939
}

F
Fam Zheng 已提交
1940
static void bdrv_delete(BlockDriverState *bs)
B
bellard 已提交
1941
{
1942
    assert(!bs->job);
1943
    assert(bdrv_op_blocker_is_empty(bs));
F
Fam Zheng 已提交
1944
    assert(!bs->refcnt);
F
Fam Zheng 已提交
1945
    assert(QLIST_EMPTY(&bs->dirty_bitmaps));
1946

1947 1948
    bdrv_close(bs);

1949
    /* remove from list, if necessary */
1950
    bdrv_make_anon(bs);
1951

1952
    g_free(bs);
B
bellard 已提交
1953 1954
}

A
aliguori 已提交
1955 1956 1957
/*
 * Run consistency checks on an image
 *
1958
 * Returns 0 if the check could be completed (it doesn't mean that the image is
1959
 * free of errors) or -errno when an internal error occurred. The results of the
1960
 * check are stored in res.
A
aliguori 已提交
1961
 */
1962
int bdrv_check(BlockDriverState *bs, BdrvCheckResult *res, BdrvCheckMode fix)
A
aliguori 已提交
1963
{
1964 1965 1966
    if (bs->drv == NULL) {
        return -ENOMEDIUM;
    }
A
aliguori 已提交
1967 1968 1969 1970
    if (bs->drv->bdrv_check == NULL) {
        return -ENOTSUP;
    }

1971
    memset(res, 0, sizeof(*res));
1972
    return bs->drv->bdrv_check(bs, res, fix);
A
aliguori 已提交
1973 1974
}

1975 1976
#define COMMIT_BUF_SECTORS 2048

1977 1978 1979
/* commit COW file into the raw image */
int bdrv_commit(BlockDriverState *bs)
{
B
bellard 已提交
1980
    BlockDriver *drv = bs->drv;
1981
    int64_t sector, total_sectors, length, backing_length;
1982
    int n, ro, open_flags;
1983
    int ret = 0;
1984
    uint8_t *buf = NULL;
1985

B
bellard 已提交
1986 1987
    if (!drv)
        return -ENOMEDIUM;
1988

1989 1990
    if (!bs->backing_hd) {
        return -ENOTSUP;
1991 1992
    }

1993 1994
    if (bdrv_op_is_blocked(bs, BLOCK_OP_TYPE_COMMIT_SOURCE, NULL) ||
        bdrv_op_is_blocked(bs->backing_hd, BLOCK_OP_TYPE_COMMIT_TARGET, NULL)) {
1995 1996 1997
        return -EBUSY;
    }

1998 1999 2000 2001
    ro = bs->backing_hd->read_only;
    open_flags =  bs->backing_hd->open_flags;

    if (ro) {
2002 2003
        if (bdrv_reopen(bs->backing_hd, open_flags | BDRV_O_RDWR, NULL)) {
            return -EACCES;
2004
        }
B
bellard 已提交
2005
    }
2006

2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029
    length = bdrv_getlength(bs);
    if (length < 0) {
        ret = length;
        goto ro_cleanup;
    }

    backing_length = bdrv_getlength(bs->backing_hd);
    if (backing_length < 0) {
        ret = backing_length;
        goto ro_cleanup;
    }

    /* If our top snapshot is larger than the backing file image,
     * grow the backing file image if possible.  If not possible,
     * we must return an error */
    if (length > backing_length) {
        ret = bdrv_truncate(bs->backing_hd, length);
        if (ret < 0) {
            goto ro_cleanup;
        }
    }

    total_sectors = length >> BDRV_SECTOR_BITS;
2030 2031 2032 2033 2034 2035 2036 2037

    /* qemu_try_blockalign() for bs will choose an alignment that works for
     * bs->backing_hd as well, so no need to compare the alignment manually. */
    buf = qemu_try_blockalign(bs, COMMIT_BUF_SECTORS * BDRV_SECTOR_SIZE);
    if (buf == NULL) {
        ret = -ENOMEM;
        goto ro_cleanup;
    }
2038 2039

    for (sector = 0; sector < total_sectors; sector += n) {
2040 2041 2042 2043 2044
        ret = bdrv_is_allocated(bs, sector, COMMIT_BUF_SECTORS, &n);
        if (ret < 0) {
            goto ro_cleanup;
        }
        if (ret) {
K
Kevin Wolf 已提交
2045 2046
            ret = bdrv_read(bs, sector, buf, n);
            if (ret < 0) {
2047 2048 2049
                goto ro_cleanup;
            }

K
Kevin Wolf 已提交
2050 2051
            ret = bdrv_write(bs->backing_hd, sector, buf, n);
            if (ret < 0) {
2052 2053
                goto ro_cleanup;
            }
B
bellard 已提交
2054
        }
2055
    }
2056

2057 2058
    if (drv->bdrv_make_empty) {
        ret = drv->bdrv_make_empty(bs);
K
Kevin Wolf 已提交
2059 2060 2061
        if (ret < 0) {
            goto ro_cleanup;
        }
2062 2063
        bdrv_flush(bs);
    }
2064

2065 2066 2067 2068
    /*
     * Make sure all data we wrote to the backing device is actually
     * stable on disk.
     */
K
Kevin Wolf 已提交
2069
    if (bs->backing_hd) {
2070
        bdrv_flush(bs->backing_hd);
K
Kevin Wolf 已提交
2071
    }
2072

K
Kevin Wolf 已提交
2073
    ret = 0;
2074
ro_cleanup:
2075
    qemu_vfree(buf);
2076 2077

    if (ro) {
2078 2079
        /* ignoring error return here */
        bdrv_reopen(bs->backing_hd, open_flags & ~BDRV_O_RDWR, NULL);
2080 2081
    }

2082
    return ret;
2083 2084
}

2085
int bdrv_commit_all(void)
2086 2087 2088
{
    BlockDriverState *bs;

2089
    QTAILQ_FOREACH(bs, &bdrv_states, device_list) {
2090 2091 2092
        AioContext *aio_context = bdrv_get_aio_context(bs);

        aio_context_acquire(aio_context);
2093 2094 2095
        if (bs->drv && bs->backing_hd) {
            int ret = bdrv_commit(bs);
            if (ret < 0) {
2096
                aio_context_release(aio_context);
2097 2098
                return ret;
            }
2099
        }
2100
        aio_context_release(aio_context);
2101
    }
2102
    return 0;
2103 2104
}

K
Kevin Wolf 已提交
2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116
/*
 * Return values:
 * 0        - success
 * -EINVAL  - backing format specified, but no file
 * -ENOSPC  - can't update the backing file because no space is left in the
 *            image file header
 * -ENOTSUP - format driver doesn't support changing the backing file
 */
int bdrv_change_backing_file(BlockDriverState *bs,
    const char *backing_file, const char *backing_fmt)
{
    BlockDriver *drv = bs->drv;
2117
    int ret;
K
Kevin Wolf 已提交
2118

2119 2120 2121 2122 2123
    /* Backing file format doesn't make sense without a backing file */
    if (backing_fmt && !backing_file) {
        return -EINVAL;
    }

K
Kevin Wolf 已提交
2124
    if (drv->bdrv_change_backing_file != NULL) {
2125
        ret = drv->bdrv_change_backing_file(bs, backing_file, backing_fmt);
K
Kevin Wolf 已提交
2126
    } else {
2127
        ret = -ENOTSUP;
K
Kevin Wolf 已提交
2128
    }
2129 2130 2131 2132 2133 2134

    if (ret == 0) {
        pstrcpy(bs->backing_file, sizeof(bs->backing_file), backing_file ?: "");
        pstrcpy(bs->backing_format, sizeof(bs->backing_format), backing_fmt ?: "");
    }
    return ret;
K
Kevin Wolf 已提交
2135 2136
}

2137 2138 2139 2140 2141 2142 2143
/*
 * Finds the image layer in the chain that has 'bs' as its backing file.
 *
 * active is the current topmost image.
 *
 * Returns NULL if bs is not found in active's image chain,
 * or if active == bs.
2144 2145
 *
 * Returns the bottommost base image if bs == NULL.
2146 2147 2148 2149
 */
BlockDriverState *bdrv_find_overlay(BlockDriverState *active,
                                    BlockDriverState *bs)
{
2150 2151
    while (active && bs != active->backing_hd) {
        active = active->backing_hd;
2152 2153
    }

2154 2155
    return active;
}
2156

2157 2158 2159 2160
/* Given a BDS, searches for the base layer. */
BlockDriverState *bdrv_find_base(BlockDriverState *bs)
{
    return bdrv_find_overlay(bs, NULL);
2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190
}

typedef struct BlkIntermediateStates {
    BlockDriverState *bs;
    QSIMPLEQ_ENTRY(BlkIntermediateStates) entry;
} BlkIntermediateStates;


/*
 * Drops images above 'base' up to and including 'top', and sets the image
 * above 'top' to have base as its backing file.
 *
 * Requires that the overlay to 'top' is opened r/w, so that the backing file
 * information in 'bs' can be properly updated.
 *
 * E.g., this will convert the following chain:
 * bottom <- base <- intermediate <- top <- active
 *
 * to
 *
 * bottom <- base <- active
 *
 * It is allowed for bottom==base, in which case it converts:
 *
 * base <- intermediate <- top <- active
 *
 * to
 *
 * base <- active
 *
2191 2192 2193
 * If backing_file_str is non-NULL, it will be used when modifying top's
 * overlay image metadata.
 *
2194 2195 2196 2197 2198
 * Error conditions:
 *  if active == top, that is considered an error
 *
 */
int bdrv_drop_intermediate(BlockDriverState *active, BlockDriverState *top,
2199
                           BlockDriverState *base, const char *backing_file_str)
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
{
    BlockDriverState *intermediate;
    BlockDriverState *base_bs = NULL;
    BlockDriverState *new_top_bs = NULL;
    BlkIntermediateStates *intermediate_state, *next;
    int ret = -EIO;

    QSIMPLEQ_HEAD(states_to_delete, BlkIntermediateStates) states_to_delete;
    QSIMPLEQ_INIT(&states_to_delete);

    if (!top->drv || !base->drv) {
        goto exit;
    }

    new_top_bs = bdrv_find_overlay(active, top);

    if (new_top_bs == NULL) {
        /* we could not find the image above 'top', this is an error */
        goto exit;
    }

    /* special case of new_top_bs->backing_hd already pointing to base - nothing
     * to do, no intermediate images */
    if (new_top_bs->backing_hd == base) {
        ret = 0;
        goto exit;
    }

    intermediate = top;

    /* now we will go down through the list, and add each BDS we find
     * into our deletion queue, until we hit the 'base'
     */
    while (intermediate) {
2234
        intermediate_state = g_new0(BlkIntermediateStates, 1);
2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250
        intermediate_state->bs = intermediate;
        QSIMPLEQ_INSERT_TAIL(&states_to_delete, intermediate_state, entry);

        if (intermediate->backing_hd == base) {
            base_bs = intermediate->backing_hd;
            break;
        }
        intermediate = intermediate->backing_hd;
    }
    if (base_bs == NULL) {
        /* something went wrong, we did not end at the base. safely
         * unravel everything, and exit with error */
        goto exit;
    }

    /* success - we can delete the intermediate states, and link top->base */
2251 2252
    backing_file_str = backing_file_str ? backing_file_str : base_bs->filename;
    ret = bdrv_change_backing_file(new_top_bs, backing_file_str,
2253 2254 2255 2256
                                   base_bs->drv ? base_bs->drv->format_name : "");
    if (ret) {
        goto exit;
    }
2257
    bdrv_set_backing_hd(new_top_bs, base_bs);
2258 2259 2260

    QSIMPLEQ_FOREACH_SAFE(intermediate_state, &states_to_delete, entry, next) {
        /* so that bdrv_close() does not recursively close the chain */
2261
        bdrv_set_backing_hd(intermediate_state->bs, NULL);
F
Fam Zheng 已提交
2262
        bdrv_unref(intermediate_state->bs);
2263 2264 2265 2266 2267 2268 2269 2270 2271 2272
    }
    ret = 0;

exit:
    QSIMPLEQ_FOREACH_SAFE(intermediate_state, &states_to_delete, entry, next) {
        g_free(intermediate_state);
    }
    return ret;
}

2273 2274 2275 2276
/**
 * Truncate file to 'offset' bytes (needed only for file protocols)
 */
int bdrv_truncate(BlockDriverState *bs, int64_t offset)
2277
{
2278 2279 2280
    BlockDriver *drv = bs->drv;
    int ret;
    if (!drv)
2281
        return -ENOMEDIUM;
2282 2283 2284 2285
    if (!drv->bdrv_truncate)
        return -ENOTSUP;
    if (bs->read_only)
        return -EACCES;
2286

2287 2288 2289 2290 2291 2292 2293
    ret = drv->bdrv_truncate(bs, offset);
    if (ret == 0) {
        ret = refresh_total_sectors(bs, offset >> BDRV_SECTOR_BITS);
        bdrv_dirty_bitmap_truncate(bs);
        if (bs->blk) {
            blk_dev_resize_cb(bs->blk);
        }
M
Max Reitz 已提交
2294
    }
2295
    return ret;
2296 2297
}

2298 2299 2300 2301 2302
/**
 * Length of a allocated file in bytes. Sparse files are counted by actual
 * allocated space. Return < 0 if error or unknown.
 */
int64_t bdrv_get_allocated_file_size(BlockDriverState *bs)
2303
{
2304 2305 2306
    BlockDriver *drv = bs->drv;
    if (!drv) {
        return -ENOMEDIUM;
2307
    }
2308 2309 2310 2311 2312
    if (drv->bdrv_get_allocated_file_size) {
        return drv->bdrv_get_allocated_file_size(bs);
    }
    if (bs->file) {
        return bdrv_get_allocated_file_size(bs->file);
2313
    }
2314
    return -ENOTSUP;
2315
}
2316

2317 2318
/**
 * Return number of sectors on success, -errno on error.
2319
 */
2320
int64_t bdrv_nb_sectors(BlockDriverState *bs)
2321
{
2322
    BlockDriver *drv = bs->drv;
2323

2324 2325
    if (!drv)
        return -ENOMEDIUM;
2326

2327 2328 2329 2330
    if (drv->has_variable_length) {
        int ret = refresh_total_sectors(bs, bs->total_sectors);
        if (ret < 0) {
            return ret;
2331 2332
        }
    }
2333
    return bs->total_sectors;
2334
}
B
bellard 已提交
2335

2336 2337 2338
/**
 * Return length in bytes on success, -errno on error.
 * The length is always a multiple of BDRV_SECTOR_SIZE.
2339
 */
2340
int64_t bdrv_getlength(BlockDriverState *bs)
2341
{
2342
    int64_t ret = bdrv_nb_sectors(bs);
2343

2344
    ret = ret > INT64_MAX / BDRV_SECTOR_SIZE ? -EFBIG : ret;
2345
    return ret < 0 ? ret : ret * BDRV_SECTOR_SIZE;
B
bellard 已提交
2346 2347
}

2348 2349
/* return 0 as number of sectors if no device present or error */
void bdrv_get_geometry(BlockDriverState *bs, uint64_t *nb_sectors_ptr)
2350
{
2351
    int64_t nb_sectors = bdrv_nb_sectors(bs);
2352

2353
    *nb_sectors_ptr = nb_sectors < 0 ? 0 : nb_sectors;
2354 2355
}

2356 2357
void bdrv_set_on_error(BlockDriverState *bs, BlockdevOnError on_read_error,
                       BlockdevOnError on_write_error)
B
bellard 已提交
2358
{
2359 2360
    bs->on_read_error = on_read_error;
    bs->on_write_error = on_write_error;
B
bellard 已提交
2361 2362
}

2363
BlockdevOnError bdrv_get_on_error(BlockDriverState *bs, bool is_read)
P
Peter Lieven 已提交
2364
{
2365
    return is_read ? bs->on_read_error : bs->on_write_error;
2366 2367
}

2368
BlockErrorAction bdrv_get_error_action(BlockDriverState *bs, bool is_read, int error)
P
Peter Lieven 已提交
2369
{
2370
    BlockdevOnError on_err = is_read ? bs->on_read_error : bs->on_write_error;
2371

2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383
    switch (on_err) {
    case BLOCKDEV_ON_ERROR_ENOSPC:
        return (error == ENOSPC) ?
               BLOCK_ERROR_ACTION_STOP : BLOCK_ERROR_ACTION_REPORT;
    case BLOCKDEV_ON_ERROR_STOP:
        return BLOCK_ERROR_ACTION_STOP;
    case BLOCKDEV_ON_ERROR_REPORT:
        return BLOCK_ERROR_ACTION_REPORT;
    case BLOCKDEV_ON_ERROR_IGNORE:
        return BLOCK_ERROR_ACTION_IGNORE;
    default:
        abort();
P
Peter Lieven 已提交
2384 2385 2386
    }
}

2387 2388 2389
static void send_qmp_error_event(BlockDriverState *bs,
                                 BlockErrorAction action,
                                 bool is_read, int error)
B
bellard 已提交
2390
{
2391
    IoOperationType optype;
2392

2393 2394 2395 2396 2397
    optype = is_read ? IO_OPERATION_TYPE_READ : IO_OPERATION_TYPE_WRITE;
    qapi_event_send_block_io_error(bdrv_get_device_name(bs), optype, action,
                                   bdrv_iostatus_is_enabled(bs),
                                   error == ENOSPC, strerror(error),
                                   &error_abort);
B
bellard 已提交
2398 2399
}

2400 2401 2402 2403 2404 2405
/* This is done by device models because, while the block layer knows
 * about the error, it does not know whether an operation comes from
 * the device or the block layer (from a job, for example).
 */
void bdrv_error_action(BlockDriverState *bs, BlockErrorAction action,
                       bool is_read, int error)
B
bellard 已提交
2406
{
2407
    assert(error >= 0);
B
bellard 已提交
2408

2409 2410 2411 2412 2413 2414
    if (action == BLOCK_ERROR_ACTION_STOP) {
        /* First set the iostatus, so that "info block" returns an iostatus
         * that matches the events raised so far (an additional error iostatus
         * is fine, but not a lost one).
         */
        bdrv_iostatus_set_err(bs, error);
B
bellard 已提交
2415

2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429
        /* Then raise the request to stop the VM and the event.
         * qemu_system_vmstop_request_prepare has two effects.  First,
         * it ensures that the STOP event always comes after the
         * BLOCK_IO_ERROR event.  Second, it ensures that even if management
         * can observe the STOP event and do a "cont" before the STOP
         * event is issued, the VM will not stop.  In this case, vm_start()
         * also ensures that the STOP/RESUME pair of events is emitted.
         */
        qemu_system_vmstop_request_prepare();
        send_qmp_error_event(bs, action, is_read, error);
        qemu_system_vmstop_request(RUN_STATE_IO_ERROR);
    } else {
        send_qmp_error_event(bs, action, is_read, error);
    }
2430 2431
}

2432
int bdrv_is_read_only(BlockDriverState *bs)
2433
{
2434
    return bs->read_only;
B
bellard 已提交
2435 2436
}

2437
int bdrv_is_sg(BlockDriverState *bs)
K
Kevin Wolf 已提交
2438
{
2439
    return bs->sg;
K
Kevin Wolf 已提交
2440 2441
}

2442
int bdrv_enable_write_cache(BlockDriverState *bs)
S
Stefan Hajnoczi 已提交
2443
{
2444
    return bs->enable_write_cache;
S
Stefan Hajnoczi 已提交
2445 2446
}

2447
void bdrv_set_enable_write_cache(BlockDriverState *bs, bool wce)
K
Kevin Wolf 已提交
2448
{
2449
    bs->enable_write_cache = wce;
S
Stefan Hajnoczi 已提交
2450

2451 2452 2453
    /* so a reopen() will preserve wce */
    if (wce) {
        bs->open_flags |= BDRV_O_CACHE_WB;
2454
    } else {
2455
        bs->open_flags &= ~BDRV_O_CACHE_WB;
2456
    }
K
Kevin Wolf 已提交
2457 2458
}

2459
int bdrv_is_encrypted(BlockDriverState *bs)
F
Fam Zheng 已提交
2460
{
2461 2462 2463
    if (bs->backing_hd && bs->backing_hd->encrypted)
        return 1;
    return bs->encrypted;
F
Fam Zheng 已提交
2464 2465
}

2466
int bdrv_key_required(BlockDriverState *bs)
F
Fam Zheng 已提交
2467
{
2468 2469 2470 2471 2472
    BlockDriverState *backing_hd = bs->backing_hd;

    if (backing_hd && backing_hd->encrypted && !backing_hd->valid_key)
        return 1;
    return (bs->encrypted && !bs->valid_key);
F
Fam Zheng 已提交
2473 2474
}

2475
int bdrv_set_key(BlockDriverState *bs, const char *key)
2476 2477
{
    int ret;
2478 2479 2480 2481 2482 2483 2484 2485 2486 2487
    if (bs->backing_hd && bs->backing_hd->encrypted) {
        ret = bdrv_set_key(bs->backing_hd, key);
        if (ret < 0)
            return ret;
        if (!bs->encrypted)
            return 0;
    }
    if (!bs->encrypted) {
        return -EINVAL;
    } else if (!bs->drv || !bs->drv->bdrv_set_key) {
2488 2489
        return -ENOMEDIUM;
    }
2490
    ret = bs->drv->bdrv_set_key(bs, key);
2491
    if (ret < 0) {
2492 2493 2494 2495 2496 2497 2498
        bs->valid_key = 0;
    } else if (!bs->valid_key) {
        bs->valid_key = 1;
        if (bs->blk) {
            /* call the change callback now, we skipped it on open */
            blk_dev_change_media_cb(bs->blk, true);
        }
2499
    }
2500 2501
    return ret;
}
2502

2503
/*
2504 2505 2506 2507 2508 2509 2510 2511 2512
 * Provide an encryption key for @bs.
 * If @key is non-null:
 *     If @bs is not encrypted, fail.
 *     Else if the key is invalid, fail.
 *     Else set @bs's key to @key, replacing the existing key, if any.
 * If @key is null:
 *     If @bs is encrypted and still lacks a key, fail.
 *     Else do nothing.
 * On failure, store an error object through @errp if non-null.
2513
 */
2514
void bdrv_add_key(BlockDriverState *bs, const char *key, Error **errp)
2515
{
2516 2517 2518 2519 2520 2521
    if (key) {
        if (!bdrv_is_encrypted(bs)) {
            error_setg(errp, "Node '%s' is not encrypted",
                      bdrv_get_device_or_node_name(bs));
        } else if (bdrv_set_key(bs, key) < 0) {
            error_set(errp, QERR_INVALID_PASSWORD);
2522 2523 2524
        }
    } else {
        if (bdrv_key_required(bs)) {
2525 2526
            error_set(errp, ERROR_CLASS_DEVICE_ENCRYPTED,
                      "'%s' (%s) is encrypted",
2527
                      bdrv_get_device_or_node_name(bs),
2528 2529 2530 2531 2532
                      bdrv_get_encrypted_filename(bs));
        }
    }
}

2533
const char *bdrv_get_format_name(BlockDriverState *bs)
K
Kevin Wolf 已提交
2534
{
2535
    return bs->drv ? bs->drv->format_name : NULL;
K
Kevin Wolf 已提交
2536 2537
}

2538
static int qsort_strcmp(const void *a, const void *b)
K
Kevin Wolf 已提交
2539
{
2540
    return strcmp(a, b);
K
Kevin Wolf 已提交
2541 2542
}

2543 2544
void bdrv_iterate_format(void (*it)(void *opaque, const char *name),
                         void *opaque)
K
Kevin Wolf 已提交
2545
{
2546 2547 2548 2549
    BlockDriver *drv;
    int count = 0;
    int i;
    const char **formats = NULL;
K
Kevin Wolf 已提交
2550

2551 2552 2553 2554 2555 2556 2557
    QLIST_FOREACH(drv, &bdrv_drivers, list) {
        if (drv->format_name) {
            bool found = false;
            int i = count;
            while (formats && i && !found) {
                found = !strcmp(formats[--i], drv->format_name);
            }
2558

2559 2560 2561 2562
            if (!found) {
                formats = g_renew(const char *, formats, count + 1);
                formats[count++] = drv->format_name;
            }
2563
        }
2564
    }
2565

2566
    qsort(formats, count, sizeof(formats[0]), qsort_strcmp);
K
Kevin Wolf 已提交
2567

2568 2569 2570
    for (i = 0; i < count; i++) {
        it(opaque, formats[i]);
    }
K
Kevin Wolf 已提交
2571

2572 2573
    g_free(formats);
}
K
Kevin Wolf 已提交
2574

2575 2576 2577 2578
/* This function is to find a node in the bs graph */
BlockDriverState *bdrv_find_node(const char *node_name)
{
    BlockDriverState *bs;
2579

2580
    assert(node_name);
K
Kevin Wolf 已提交
2581

2582 2583 2584
    QTAILQ_FOREACH(bs, &graph_bdrv_states, node_list) {
        if (!strcmp(node_name, bs->node_name)) {
            return bs;
K
Kevin Wolf 已提交
2585 2586
        }
    }
2587
    return NULL;
K
Kevin Wolf 已提交
2588 2589
}

2590 2591
/* Put this QMP function here so it can access the static graph_bdrv_states. */
BlockDeviceInfoList *bdrv_named_nodes_list(Error **errp)
K
Kevin Wolf 已提交
2592
{
2593 2594
    BlockDeviceInfoList *list, *entry;
    BlockDriverState *bs;
K
Kevin Wolf 已提交
2595

2596 2597 2598 2599 2600 2601
    list = NULL;
    QTAILQ_FOREACH(bs, &graph_bdrv_states, node_list) {
        BlockDeviceInfo *info = bdrv_block_device_info(bs, errp);
        if (!info) {
            qapi_free_BlockDeviceInfoList(list);
            return NULL;
2602
        }
2603 2604 2605 2606
        entry = g_malloc0(sizeof(*entry));
        entry->value = info;
        entry->next = list;
        list = entry;
2607 2608
    }

2609 2610
    return list;
}
K
Kevin Wolf 已提交
2611

2612 2613 2614 2615 2616 2617
BlockDriverState *bdrv_lookup_bs(const char *device,
                                 const char *node_name,
                                 Error **errp)
{
    BlockBackend *blk;
    BlockDriverState *bs;
K
Kevin Wolf 已提交
2618

2619 2620
    if (device) {
        blk = blk_by_name(device);
K
Kevin Wolf 已提交
2621

2622 2623 2624 2625
        if (blk) {
            return blk_bs(blk);
        }
    }
K
Kevin Wolf 已提交
2626

2627 2628
    if (node_name) {
        bs = bdrv_find_node(node_name);
2629

2630 2631 2632
        if (bs) {
            return bs;
        }
K
Kevin Wolf 已提交
2633 2634
    }

2635 2636 2637 2638
    error_setg(errp, "Cannot find device=%s nor node_name=%s",
                     device ? device : "",
                     node_name ? node_name : "");
    return NULL;
K
Kevin Wolf 已提交
2639 2640
}

2641 2642 2643
/* If 'base' is in the same chain as 'top', return true. Otherwise,
 * return false.  If either argument is NULL, return false. */
bool bdrv_chain_contains(BlockDriverState *top, BlockDriverState *base)
B
bellard 已提交
2644
{
2645 2646
    while (top && top != base) {
        top = top->backing_hd;
F
Fam Zheng 已提交
2647
    }
2648 2649

    return top != NULL;
F
Fam Zheng 已提交
2650 2651
}

2652
BlockDriverState *bdrv_next_node(BlockDriverState *bs)
F
Fam Zheng 已提交
2653
{
2654 2655
    if (!bs) {
        return QTAILQ_FIRST(&graph_bdrv_states);
F
Fam Zheng 已提交
2656
    }
2657
    return QTAILQ_NEXT(bs, node_list);
B
bellard 已提交
2658 2659
}

2660
BlockDriverState *bdrv_next(BlockDriverState *bs)
B
bellard 已提交
2661
{
2662 2663
    if (!bs) {
        return QTAILQ_FIRST(&bdrv_states);
2664
    }
2665
    return QTAILQ_NEXT(bs, device_list);
B
bellard 已提交
2666
}
2667

2668
const char *bdrv_get_node_name(const BlockDriverState *bs)
B
bellard 已提交
2669
{
2670
    return bs->node_name;
2671 2672
}

2673 2674
/* TODO check what callers really want: bs->node_name or blk_name() */
const char *bdrv_get_device_name(const BlockDriverState *bs)
2675
{
2676
    return bs->blk ? blk_name(bs->blk) : "";
2677
}
B
bellard 已提交
2678

2679 2680 2681 2682 2683
/* This can be used to identify nodes that might not have a device
 * name associated. Since node and device names live in the same
 * namespace, the result is unambiguous. The exception is if both are
 * absent, then this returns an empty (non-null) string. */
const char *bdrv_get_device_or_node_name(const BlockDriverState *bs)
2684
{
2685
    return bs->blk ? blk_name(bs->blk) : bs->node_name;
2686 2687
}

2688
int bdrv_get_flags(BlockDriverState *bs)
2689
{
2690
    return bs->open_flags;
2691 2692
}

2693
int bdrv_has_zero_init_1(BlockDriverState *bs)
2694
{
2695
    return 1;
2696 2697
}

2698
int bdrv_has_zero_init(BlockDriverState *bs)
2699
{
2700
    assert(bs->drv);
2701

2702 2703 2704 2705 2706 2707 2708
    /* If BS is a copy on write image, it is initialized to
       the contents of the base image, which may not be zeroes.  */
    if (bs->backing_hd) {
        return 0;
    }
    if (bs->drv->bdrv_has_zero_init) {
        return bs->drv->bdrv_has_zero_init(bs);
2709
    }
2710 2711 2712

    /* safe default */
    return 0;
2713 2714
}

2715
bool bdrv_unallocated_blocks_are_zero(BlockDriverState *bs)
2716
{
2717
    BlockDriverInfo bdi;
2718

2719 2720 2721 2722 2723 2724
    if (bs->backing_hd) {
        return false;
    }

    if (bdrv_get_info(bs, &bdi) == 0) {
        return bdi.unallocated_blocks_are_zero;
2725 2726
    }

2727
    return false;
2728 2729
}

2730
bool bdrv_can_write_zeroes_with_unmap(BlockDriverState *bs)
2731
{
2732
    BlockDriverInfo bdi;
2733

2734 2735 2736
    if (bs->backing_hd || !(bs->open_flags & BDRV_O_UNMAP)) {
        return false;
    }
2737

2738 2739 2740
    if (bdrv_get_info(bs, &bdi) == 0) {
        return bdi.can_write_zeroes_with_unmap;
    }
2741

2742
    return false;
2743 2744
}

2745
const char *bdrv_get_encrypted_filename(BlockDriverState *bs)
2746
{
2747 2748 2749 2750 2751 2752
    if (bs->backing_hd && bs->backing_hd->encrypted)
        return bs->backing_file;
    else if (bs->encrypted)
        return bs->filename;
    else
        return NULL;
2753 2754
}

2755 2756
void bdrv_get_backing_filename(BlockDriverState *bs,
                               char *filename, int filename_size)
2757
{
2758 2759
    pstrcpy(filename, filename_size, bs->backing_file);
}
K
Kevin Wolf 已提交
2760

2761 2762 2763 2764 2765 2766 2767 2768 2769 2770
int bdrv_get_info(BlockDriverState *bs, BlockDriverInfo *bdi)
{
    BlockDriver *drv = bs->drv;
    if (!drv)
        return -ENOMEDIUM;
    if (!drv->bdrv_get_info)
        return -ENOTSUP;
    memset(bdi, 0, sizeof(*bdi));
    return drv->bdrv_get_info(bs, bdi);
}
2771

2772 2773 2774 2775 2776 2777 2778
ImageInfoSpecific *bdrv_get_specific_info(BlockDriverState *bs)
{
    BlockDriver *drv = bs->drv;
    if (drv && drv->bdrv_get_specific_info) {
        return drv->bdrv_get_specific_info(bs);
    }
    return NULL;
2779 2780
}

2781
void bdrv_debug_event(BlockDriverState *bs, BlkDebugEvent event)
2782
{
2783 2784 2785
    if (!bs || !bs->drv || !bs->drv->bdrv_debug_event) {
        return;
    }
2786

2787
    bs->drv->bdrv_debug_event(bs, event);
2788 2789
}

2790 2791
int bdrv_debug_breakpoint(BlockDriverState *bs, const char *event,
                          const char *tag)
2792
{
2793 2794 2795
    while (bs && bs->drv && !bs->drv->bdrv_debug_breakpoint) {
        bs = bs->file;
    }
2796

2797 2798 2799
    if (bs && bs->drv && bs->drv->bdrv_debug_breakpoint) {
        return bs->drv->bdrv_debug_breakpoint(bs, event, tag);
    }
2800

2801
    return -ENOTSUP;
2802 2803
}

2804
int bdrv_debug_remove_breakpoint(BlockDriverState *bs, const char *tag)
B
bellard 已提交
2805
{
2806 2807 2808
    while (bs && bs->drv && !bs->drv->bdrv_debug_remove_breakpoint) {
        bs = bs->file;
    }
2809

2810 2811 2812 2813 2814
    if (bs && bs->drv && bs->drv->bdrv_debug_remove_breakpoint) {
        return bs->drv->bdrv_debug_remove_breakpoint(bs, tag);
    }

    return -ENOTSUP;
2815 2816
}

2817
int bdrv_debug_resume(BlockDriverState *bs, const char *tag)
2818
{
2819 2820 2821
    while (bs && (!bs->drv || !bs->drv->bdrv_debug_resume)) {
        bs = bs->file;
    }
2822

2823 2824 2825
    if (bs && bs->drv && bs->drv->bdrv_debug_resume) {
        return bs->drv->bdrv_debug_resume(bs, tag);
    }
2826

2827
    return -ENOTSUP;
2828 2829
}

2830
bool bdrv_debug_is_suspended(BlockDriverState *bs, const char *tag)
2831
{
2832 2833
    while (bs && bs->drv && !bs->drv->bdrv_debug_is_suspended) {
        bs = bs->file;
2834
    }
B
bellard 已提交
2835

2836 2837 2838
    if (bs && bs->drv && bs->drv->bdrv_debug_is_suspended) {
        return bs->drv->bdrv_debug_is_suspended(bs, tag);
    }
2839

2840 2841
    return false;
}
2842

2843
int bdrv_is_snapshot(BlockDriverState *bs)
2844
{
2845
    return !!(bs->open_flags & BDRV_O_SNAPSHOT);
2846 2847
}

2848 2849 2850 2851 2852 2853
/* backing_file can either be relative, or absolute, or a protocol.  If it is
 * relative, it must be relative to the chain.  So, passing in bs->filename
 * from a BDS as backing_file should not be done, as that may be relative to
 * the CWD rather than the chain. */
BlockDriverState *bdrv_find_backing_image(BlockDriverState *bs,
        const char *backing_file)
2854
{
2855 2856 2857 2858 2859 2860
    char *filename_full = NULL;
    char *backing_file_full = NULL;
    char *filename_tmp = NULL;
    int is_protocol = 0;
    BlockDriverState *curr_bs = NULL;
    BlockDriverState *retval = NULL;
2861

2862 2863
    if (!bs || !bs->drv || !backing_file) {
        return NULL;
2864 2865
    }

2866 2867 2868
    filename_full     = g_malloc(PATH_MAX);
    backing_file_full = g_malloc(PATH_MAX);
    filename_tmp      = g_malloc(PATH_MAX);
2869

2870
    is_protocol = path_has_protocol(backing_file);
2871

2872
    for (curr_bs = bs; curr_bs->backing_hd; curr_bs = curr_bs->backing_hd) {
2873

2874 2875 2876 2877 2878 2879 2880 2881 2882 2883 2884 2885
        /* If either of the filename paths is actually a protocol, then
         * compare unmodified paths; otherwise make paths relative */
        if (is_protocol || path_has_protocol(curr_bs->backing_file)) {
            if (strcmp(backing_file, curr_bs->backing_file) == 0) {
                retval = curr_bs->backing_hd;
                break;
            }
        } else {
            /* If not an absolute filename path, make it relative to the current
             * image's filename path */
            path_combine(filename_tmp, PATH_MAX, curr_bs->filename,
                         backing_file);
2886

2887 2888 2889 2890
            /* We are going to compare absolute pathnames */
            if (!realpath(filename_tmp, filename_full)) {
                continue;
            }
P
Paolo Bonzini 已提交
2891

2892 2893 2894 2895
            /* We need to make sure the backing filename we are comparing against
             * is relative to the current image filename (or absolute) */
            path_combine(filename_tmp, PATH_MAX, curr_bs->filename,
                         curr_bs->backing_file);
P
Paolo Bonzini 已提交
2896

2897 2898 2899
            if (!realpath(filename_tmp, backing_file_full)) {
                continue;
            }
K
Kevin Wolf 已提交
2900

2901 2902 2903 2904 2905
            if (strcmp(backing_file_full, filename_full) == 0) {
                retval = curr_bs->backing_hd;
                break;
            }
        }
K
Kevin Wolf 已提交
2906 2907
    }

2908 2909 2910 2911 2912 2913 2914 2915 2916 2917
    g_free(filename_full);
    g_free(backing_file_full);
    g_free(filename_tmp);
    return retval;
}

int bdrv_get_backing_file_depth(BlockDriverState *bs)
{
    if (!bs->drv) {
        return 0;
K
Kevin Wolf 已提交
2918 2919
    }

2920 2921
    if (!bs->backing_hd) {
        return 0;
2922 2923
    }

2924 2925
    return 1 + bdrv_get_backing_file_depth(bs->backing_hd);
}
P
Paolo Bonzini 已提交
2926

2927 2928 2929 2930
void bdrv_init(void)
{
    module_call_init(MODULE_INIT_BLOCK);
}
2931

2932 2933 2934 2935
void bdrv_init_with_whitelist(void)
{
    use_bdrv_whitelist = 1;
    bdrv_init();
P
Paolo Bonzini 已提交
2936 2937
}

2938
void bdrv_invalidate_cache(BlockDriverState *bs, Error **errp)
2939
{
2940 2941 2942
    Error *local_err = NULL;
    int ret;

2943 2944 2945 2946
    if (!bs->drv)  {
        return;
    }

2947 2948 2949 2950 2951
    if (!(bs->open_flags & BDRV_O_INCOMING)) {
        return;
    }
    bs->open_flags &= ~BDRV_O_INCOMING;

2952
    if (bs->drv->bdrv_invalidate_cache) {
2953
        bs->drv->bdrv_invalidate_cache(bs, &local_err);
2954
    } else if (bs->file) {
2955 2956 2957 2958 2959
        bdrv_invalidate_cache(bs->file, &local_err);
    }
    if (local_err) {
        error_propagate(errp, local_err);
        return;
2960
    }
2961

2962 2963 2964 2965 2966
    ret = refresh_total_sectors(bs, bs->total_sectors);
    if (ret < 0) {
        error_setg_errno(errp, -ret, "Could not refresh total sector count");
        return;
    }
2967 2968
}

2969
void bdrv_invalidate_cache_all(Error **errp)
2970 2971
{
    BlockDriverState *bs;
2972
    Error *local_err = NULL;
2973

2974
    QTAILQ_FOREACH(bs, &bdrv_states, device_list) {
2975 2976 2977
        AioContext *aio_context = bdrv_get_aio_context(bs);

        aio_context_acquire(aio_context);
2978
        bdrv_invalidate_cache(bs, &local_err);
2979
        aio_context_release(aio_context);
2980 2981 2982 2983
        if (local_err) {
            error_propagate(errp, local_err);
            return;
        }
2984 2985 2986
    }
}

B
bellard 已提交
2987 2988 2989 2990 2991 2992 2993 2994 2995
/**************************************************************/
/* removable device support */

/**
 * Return TRUE if the media is present
 */
int bdrv_is_inserted(BlockDriverState *bs)
{
    BlockDriver *drv = bs->drv;
2996

B
bellard 已提交
2997 2998 2999
    if (!drv)
        return 0;
    if (!drv->bdrv_is_inserted)
3000 3001
        return 1;
    return drv->bdrv_is_inserted(bs);
B
bellard 已提交
3002 3003 3004
}

/**
3005 3006
 * Return whether the media changed since the last call to this
 * function, or -ENOTSUP if we don't know.  Most drivers don't know.
B
bellard 已提交
3007 3008 3009 3010 3011
 */
int bdrv_media_changed(BlockDriverState *bs)
{
    BlockDriver *drv = bs->drv;

3012 3013 3014 3015
    if (drv && drv->bdrv_media_changed) {
        return drv->bdrv_media_changed(bs);
    }
    return -ENOTSUP;
B
bellard 已提交
3016 3017 3018 3019 3020
}

/**
 * If eject_flag is TRUE, eject the media. Otherwise, close the tray
 */
3021
void bdrv_eject(BlockDriverState *bs, bool eject_flag)
B
bellard 已提交
3022 3023
{
    BlockDriver *drv = bs->drv;
3024
    const char *device_name;
B
bellard 已提交
3025

3026 3027
    if (drv && drv->bdrv_eject) {
        drv->bdrv_eject(bs, eject_flag);
B
bellard 已提交
3028
    }
3029

3030 3031 3032
    device_name = bdrv_get_device_name(bs);
    if (device_name[0] != '\0') {
        qapi_event_send_device_tray_moved(device_name,
3033
                                          eject_flag, &error_abort);
3034
    }
B
bellard 已提交
3035 3036 3037 3038 3039 3040
}

/**
 * Lock or unlock the media (if it is locked, the user won't be able
 * to eject it manually).
 */
3041
void bdrv_lock_medium(BlockDriverState *bs, bool locked)
B
bellard 已提交
3042 3043 3044
{
    BlockDriver *drv = bs->drv;

3045
    trace_bdrv_lock_medium(bs, locked);
S
Stefan Hajnoczi 已提交
3046

3047 3048
    if (drv && drv->bdrv_lock_medium) {
        drv->bdrv_lock_medium(bs, locked);
B
bellard 已提交
3049 3050
    }
}
3051

3052
void bdrv_set_guest_block_size(BlockDriverState *bs, int align)
3053
{
3054
    bs->guest_block_size = align;
3055
}
3056

3057 3058 3059 3060 3061 3062 3063 3064 3065 3066 3067 3068 3069
BdrvDirtyBitmap *bdrv_find_dirty_bitmap(BlockDriverState *bs, const char *name)
{
    BdrvDirtyBitmap *bm;

    assert(name);
    QLIST_FOREACH(bm, &bs->dirty_bitmaps, list) {
        if (bm->name && !strcmp(name, bm->name)) {
            return bm;
        }
    }
    return NULL;
}

3070
void bdrv_dirty_bitmap_make_anon(BdrvDirtyBitmap *bitmap)
3071
{
J
John Snow 已提交
3072
    assert(!bdrv_dirty_bitmap_frozen(bitmap));
3073 3074 3075 3076 3077
    g_free(bitmap->name);
    bitmap->name = NULL;
}

BdrvDirtyBitmap *bdrv_create_dirty_bitmap(BlockDriverState *bs,
3078
                                          uint32_t granularity,
3079
                                          const char *name,
3080
                                          Error **errp)
3081 3082
{
    int64_t bitmap_size;
F
Fam Zheng 已提交
3083
    BdrvDirtyBitmap *bitmap;
3084
    uint32_t sector_granularity;
3085

3086 3087
    assert((granularity & (granularity - 1)) == 0);

3088 3089 3090 3091
    if (name && bdrv_find_dirty_bitmap(bs, name)) {
        error_setg(errp, "Bitmap already exists: %s", name);
        return NULL;
    }
3092 3093
    sector_granularity = granularity >> BDRV_SECTOR_BITS;
    assert(sector_granularity);
3094
    bitmap_size = bdrv_nb_sectors(bs);
3095 3096 3097 3098 3099
    if (bitmap_size < 0) {
        error_setg_errno(errp, -bitmap_size, "could not get length of device");
        errno = -bitmap_size;
        return NULL;
    }
3100
    bitmap = g_new0(BdrvDirtyBitmap, 1);
3101
    bitmap->bitmap = hbitmap_alloc(bitmap_size, ctz32(sector_granularity));
J
John Snow 已提交
3102
    bitmap->size = bitmap_size;
3103
    bitmap->name = g_strdup(name);
J
John Snow 已提交
3104
    bitmap->disabled = false;
F
Fam Zheng 已提交
3105 3106 3107 3108
    QLIST_INSERT_HEAD(&bs->dirty_bitmaps, bitmap, list);
    return bitmap;
}

J
John Snow 已提交
3109 3110 3111 3112 3113
bool bdrv_dirty_bitmap_frozen(BdrvDirtyBitmap *bitmap)
{
    return bitmap->successor;
}

J
John Snow 已提交
3114 3115
bool bdrv_dirty_bitmap_enabled(BdrvDirtyBitmap *bitmap)
{
J
John Snow 已提交
3116 3117 3118
    return !(bitmap->disabled || bitmap->successor);
}

J
John Snow 已提交
3119 3120 3121 3122 3123 3124 3125 3126 3127 3128 3129
DirtyBitmapStatus bdrv_dirty_bitmap_status(BdrvDirtyBitmap *bitmap)
{
    if (bdrv_dirty_bitmap_frozen(bitmap)) {
        return DIRTY_BITMAP_STATUS_FROZEN;
    } else if (!bdrv_dirty_bitmap_enabled(bitmap)) {
        return DIRTY_BITMAP_STATUS_DISABLED;
    } else {
        return DIRTY_BITMAP_STATUS_ACTIVE;
    }
}

J
John Snow 已提交
3130 3131 3132 3133 3134 3135 3136 3137 3138 3139 3140 3141 3142 3143 3144 3145 3146 3147 3148 3149 3150 3151 3152 3153 3154 3155 3156 3157 3158 3159 3160 3161 3162 3163 3164 3165 3166 3167 3168 3169 3170 3171 3172 3173 3174 3175 3176 3177 3178 3179 3180 3181 3182 3183 3184 3185 3186 3187 3188 3189 3190 3191 3192 3193 3194 3195 3196 3197 3198 3199 3200 3201 3202 3203 3204 3205 3206 3207 3208 3209 3210 3211
/**
 * Create a successor bitmap destined to replace this bitmap after an operation.
 * Requires that the bitmap is not frozen and has no successor.
 */
int bdrv_dirty_bitmap_create_successor(BlockDriverState *bs,
                                       BdrvDirtyBitmap *bitmap, Error **errp)
{
    uint64_t granularity;
    BdrvDirtyBitmap *child;

    if (bdrv_dirty_bitmap_frozen(bitmap)) {
        error_setg(errp, "Cannot create a successor for a bitmap that is "
                   "currently frozen");
        return -1;
    }
    assert(!bitmap->successor);

    /* Create an anonymous successor */
    granularity = bdrv_dirty_bitmap_granularity(bitmap);
    child = bdrv_create_dirty_bitmap(bs, granularity, NULL, errp);
    if (!child) {
        return -1;
    }

    /* Successor will be on or off based on our current state. */
    child->disabled = bitmap->disabled;

    /* Install the successor and freeze the parent */
    bitmap->successor = child;
    return 0;
}

/**
 * For a bitmap with a successor, yield our name to the successor,
 * delete the old bitmap, and return a handle to the new bitmap.
 */
BdrvDirtyBitmap *bdrv_dirty_bitmap_abdicate(BlockDriverState *bs,
                                            BdrvDirtyBitmap *bitmap,
                                            Error **errp)
{
    char *name;
    BdrvDirtyBitmap *successor = bitmap->successor;

    if (successor == NULL) {
        error_setg(errp, "Cannot relinquish control if "
                   "there's no successor present");
        return NULL;
    }

    name = bitmap->name;
    bitmap->name = NULL;
    successor->name = name;
    bitmap->successor = NULL;
    bdrv_release_dirty_bitmap(bs, bitmap);

    return successor;
}

/**
 * In cases of failure where we can no longer safely delete the parent,
 * we may wish to re-join the parent and child/successor.
 * The merged parent will be un-frozen, but not explicitly re-enabled.
 */
BdrvDirtyBitmap *bdrv_reclaim_dirty_bitmap(BlockDriverState *bs,
                                           BdrvDirtyBitmap *parent,
                                           Error **errp)
{
    BdrvDirtyBitmap *successor = parent->successor;

    if (!successor) {
        error_setg(errp, "Cannot reclaim a successor when none is present");
        return NULL;
    }

    if (!hbitmap_merge(parent->bitmap, successor->bitmap)) {
        error_setg(errp, "Merging of parent and successor bitmap failed");
        return NULL;
    }
    bdrv_release_dirty_bitmap(bs, successor);
    parent->successor = NULL;

    return parent;
J
John Snow 已提交
3212 3213
}

3214 3215 3216 3217 3218 3219 3220 3221 3222 3223 3224 3225 3226
/**
 * Truncates _all_ bitmaps attached to a BDS.
 */
static void bdrv_dirty_bitmap_truncate(BlockDriverState *bs)
{
    BdrvDirtyBitmap *bitmap;
    uint64_t size = bdrv_nb_sectors(bs);

    QLIST_FOREACH(bitmap, &bs->dirty_bitmaps, list) {
        if (bdrv_dirty_bitmap_frozen(bitmap)) {
            continue;
        }
        hbitmap_truncate(bitmap->bitmap, size);
3227
        bitmap->size = size;
3228 3229 3230
    }
}

F
Fam Zheng 已提交
3231 3232 3233 3234 3235
void bdrv_release_dirty_bitmap(BlockDriverState *bs, BdrvDirtyBitmap *bitmap)
{
    BdrvDirtyBitmap *bm, *next;
    QLIST_FOREACH_SAFE(bm, &bs->dirty_bitmaps, list, next) {
        if (bm == bitmap) {
J
John Snow 已提交
3236
            assert(!bdrv_dirty_bitmap_frozen(bm));
F
Fam Zheng 已提交
3237 3238
            QLIST_REMOVE(bitmap, list);
            hbitmap_free(bitmap->bitmap);
3239
            g_free(bitmap->name);
F
Fam Zheng 已提交
3240 3241
            g_free(bitmap);
            return;
3242
        }
3243 3244 3245
    }
}

J
John Snow 已提交
3246 3247
void bdrv_disable_dirty_bitmap(BdrvDirtyBitmap *bitmap)
{
J
John Snow 已提交
3248
    assert(!bdrv_dirty_bitmap_frozen(bitmap));
J
John Snow 已提交
3249 3250 3251 3252 3253
    bitmap->disabled = true;
}

void bdrv_enable_dirty_bitmap(BdrvDirtyBitmap *bitmap)
{
J
John Snow 已提交
3254
    assert(!bdrv_dirty_bitmap_frozen(bitmap));
J
John Snow 已提交
3255 3256 3257
    bitmap->disabled = false;
}

F
Fam Zheng 已提交
3258 3259 3260 3261 3262 3263 3264
BlockDirtyInfoList *bdrv_query_dirty_bitmaps(BlockDriverState *bs)
{
    BdrvDirtyBitmap *bm;
    BlockDirtyInfoList *list = NULL;
    BlockDirtyInfoList **plist = &list;

    QLIST_FOREACH(bm, &bs->dirty_bitmaps, list) {
3265 3266
        BlockDirtyInfo *info = g_new0(BlockDirtyInfo, 1);
        BlockDirtyInfoList *entry = g_new0(BlockDirtyInfoList, 1);
3267
        info->count = bdrv_get_dirty_count(bm);
3268
        info->granularity = bdrv_dirty_bitmap_granularity(bm);
3269 3270
        info->has_name = !!bm->name;
        info->name = g_strdup(bm->name);
J
John Snow 已提交
3271
        info->status = bdrv_dirty_bitmap_status(bm);
F
Fam Zheng 已提交
3272 3273 3274 3275 3276 3277 3278 3279
        entry->value = info;
        *plist = entry;
        plist = &entry->next;
    }

    return list;
}

F
Fam Zheng 已提交
3280
int bdrv_get_dirty(BlockDriverState *bs, BdrvDirtyBitmap *bitmap, int64_t sector)
3281
{
F
Fam Zheng 已提交
3282 3283
    if (bitmap) {
        return hbitmap_get(bitmap->bitmap, sector);
3284 3285 3286 3287 3288
    } else {
        return 0;
    }
}

3289 3290 3291 3292 3293 3294 3295 3296 3297 3298 3299 3300 3301 3302 3303 3304 3305 3306 3307 3308
/**
 * Chooses a default granularity based on the existing cluster size,
 * but clamped between [4K, 64K]. Defaults to 64K in the case that there
 * is no cluster size information available.
 */
uint32_t bdrv_get_default_bitmap_granularity(BlockDriverState *bs)
{
    BlockDriverInfo bdi;
    uint32_t granularity;

    if (bdrv_get_info(bs, &bdi) >= 0 && bdi.cluster_size > 0) {
        granularity = MAX(4096, bdi.cluster_size);
        granularity = MIN(65536, granularity);
    } else {
        granularity = 65536;
    }

    return granularity;
}

3309 3310 3311 3312 3313
uint32_t bdrv_dirty_bitmap_granularity(BdrvDirtyBitmap *bitmap)
{
    return BDRV_SECTOR_SIZE << hbitmap_granularity(bitmap->bitmap);
}

3314
void bdrv_dirty_iter_init(BdrvDirtyBitmap *bitmap, HBitmapIter *hbi)
3315
{
F
Fam Zheng 已提交
3316
    hbitmap_iter_init(hbi, bitmap->bitmap, 0);
3317 3318
}

3319
void bdrv_set_dirty_bitmap(BdrvDirtyBitmap *bitmap,
3320 3321
                           int64_t cur_sector, int nr_sectors)
{
J
John Snow 已提交
3322
    assert(bdrv_dirty_bitmap_enabled(bitmap));
3323 3324 3325
    hbitmap_set(bitmap->bitmap, cur_sector, nr_sectors);
}

3326
void bdrv_reset_dirty_bitmap(BdrvDirtyBitmap *bitmap,
3327 3328
                             int64_t cur_sector, int nr_sectors)
{
J
John Snow 已提交
3329
    assert(bdrv_dirty_bitmap_enabled(bitmap));
3330 3331 3332
    hbitmap_reset(bitmap->bitmap, cur_sector, nr_sectors);
}

J
John Snow 已提交
3333 3334 3335 3336 3337 3338
void bdrv_clear_dirty_bitmap(BdrvDirtyBitmap *bitmap)
{
    assert(bdrv_dirty_bitmap_enabled(bitmap));
    hbitmap_reset(bitmap->bitmap, 0, bitmap->size);
}

3339 3340
void bdrv_set_dirty(BlockDriverState *bs, int64_t cur_sector,
                    int nr_sectors)
3341
{
F
Fam Zheng 已提交
3342 3343
    BdrvDirtyBitmap *bitmap;
    QLIST_FOREACH(bitmap, &bs->dirty_bitmaps, list) {
J
John Snow 已提交
3344 3345 3346
        if (!bdrv_dirty_bitmap_enabled(bitmap)) {
            continue;
        }
F
Fam Zheng 已提交
3347 3348
        hbitmap_set(bitmap->bitmap, cur_sector, nr_sectors);
    }
3349 3350
}

3351 3352
void bdrv_reset_dirty(BlockDriverState *bs, int64_t cur_sector,
                      int nr_sectors)
3353
{
F
Fam Zheng 已提交
3354 3355
    BdrvDirtyBitmap *bitmap;
    QLIST_FOREACH(bitmap, &bs->dirty_bitmaps, list) {
J
John Snow 已提交
3356 3357 3358
        if (!bdrv_dirty_bitmap_enabled(bitmap)) {
            continue;
        }
F
Fam Zheng 已提交
3359 3360
        hbitmap_reset(bitmap->bitmap, cur_sector, nr_sectors);
    }
3361
}
3362

3363 3364 3365 3366 3367 3368 3369 3370 3371
/**
 * Advance an HBitmapIter to an arbitrary offset.
 */
void bdrv_set_dirty_iter(HBitmapIter *hbi, int64_t offset)
{
    assert(hbi->hb);
    hbitmap_iter_init(hbi, hbi->hb, offset);
}

3372
int64_t bdrv_get_dirty_count(BdrvDirtyBitmap *bitmap)
3373
{
F
Fam Zheng 已提交
3374
    return hbitmap_count(bitmap->bitmap);
3375
}
J
Jes Sorensen 已提交
3376

3377 3378 3379 3380 3381 3382 3383 3384 3385 3386 3387
/* Get a reference to bs */
void bdrv_ref(BlockDriverState *bs)
{
    bs->refcnt++;
}

/* Release a previously grabbed reference to bs.
 * If after releasing, reference count is zero, the BlockDriverState is
 * deleted. */
void bdrv_unref(BlockDriverState *bs)
{
3388 3389 3390
    if (!bs) {
        return;
    }
3391 3392 3393 3394 3395 3396
    assert(bs->refcnt > 0);
    if (--bs->refcnt == 0) {
        bdrv_delete(bs);
    }
}

3397 3398 3399 3400 3401 3402 3403 3404 3405 3406 3407 3408
struct BdrvOpBlocker {
    Error *reason;
    QLIST_ENTRY(BdrvOpBlocker) list;
};

bool bdrv_op_is_blocked(BlockDriverState *bs, BlockOpType op, Error **errp)
{
    BdrvOpBlocker *blocker;
    assert((int) op >= 0 && op < BLOCK_OP_TYPE_MAX);
    if (!QLIST_EMPTY(&bs->op_blockers[op])) {
        blocker = QLIST_FIRST(&bs->op_blockers[op]);
        if (errp) {
3409 3410
            error_setg(errp, "Node '%s' is busy: %s",
                       bdrv_get_device_or_node_name(bs),
3411
                       error_get_pretty(blocker->reason));
3412 3413 3414 3415 3416 3417 3418 3419 3420 3421 3422
        }
        return true;
    }
    return false;
}

void bdrv_op_block(BlockDriverState *bs, BlockOpType op, Error *reason)
{
    BdrvOpBlocker *blocker;
    assert((int) op >= 0 && op < BLOCK_OP_TYPE_MAX);

3423
    blocker = g_new0(BdrvOpBlocker, 1);
3424 3425 3426 3427 3428 3429 3430 3431 3432 3433 3434 3435 3436 3437 3438 3439 3440 3441 3442 3443 3444 3445 3446 3447 3448 3449 3450 3451 3452 3453 3454 3455 3456 3457 3458 3459 3460 3461 3462 3463 3464 3465 3466 3467
    blocker->reason = reason;
    QLIST_INSERT_HEAD(&bs->op_blockers[op], blocker, list);
}

void bdrv_op_unblock(BlockDriverState *bs, BlockOpType op, Error *reason)
{
    BdrvOpBlocker *blocker, *next;
    assert((int) op >= 0 && op < BLOCK_OP_TYPE_MAX);
    QLIST_FOREACH_SAFE(blocker, &bs->op_blockers[op], list, next) {
        if (blocker->reason == reason) {
            QLIST_REMOVE(blocker, list);
            g_free(blocker);
        }
    }
}

void bdrv_op_block_all(BlockDriverState *bs, Error *reason)
{
    int i;
    for (i = 0; i < BLOCK_OP_TYPE_MAX; i++) {
        bdrv_op_block(bs, i, reason);
    }
}

void bdrv_op_unblock_all(BlockDriverState *bs, Error *reason)
{
    int i;
    for (i = 0; i < BLOCK_OP_TYPE_MAX; i++) {
        bdrv_op_unblock(bs, i, reason);
    }
}

bool bdrv_op_blocker_is_empty(BlockDriverState *bs)
{
    int i;

    for (i = 0; i < BLOCK_OP_TYPE_MAX; i++) {
        if (!QLIST_EMPTY(&bs->op_blockers[i])) {
            return false;
        }
    }
    return true;
}

3468 3469
void bdrv_iostatus_enable(BlockDriverState *bs)
{
3470
    bs->iostatus_enabled = true;
3471
    bs->iostatus = BLOCK_DEVICE_IO_STATUS_OK;
3472 3473 3474 3475 3476 3477
}

/* The I/O status is only enabled if the drive explicitly
 * enables it _and_ the VM is configured to stop on errors */
bool bdrv_iostatus_is_enabled(const BlockDriverState *bs)
{
3478
    return (bs->iostatus_enabled &&
3479 3480 3481
           (bs->on_write_error == BLOCKDEV_ON_ERROR_ENOSPC ||
            bs->on_write_error == BLOCKDEV_ON_ERROR_STOP   ||
            bs->on_read_error == BLOCKDEV_ON_ERROR_STOP));
3482 3483 3484 3485
}

void bdrv_iostatus_disable(BlockDriverState *bs)
{
3486
    bs->iostatus_enabled = false;
3487 3488 3489 3490 3491
}

void bdrv_iostatus_reset(BlockDriverState *bs)
{
    if (bdrv_iostatus_is_enabled(bs)) {
3492
        bs->iostatus = BLOCK_DEVICE_IO_STATUS_OK;
3493 3494 3495
        if (bs->job) {
            block_job_iostatus_reset(bs->job);
        }
3496 3497 3498 3499 3500
    }
}

void bdrv_iostatus_set_err(BlockDriverState *bs, int error)
{
3501 3502
    assert(bdrv_iostatus_is_enabled(bs));
    if (bs->iostatus == BLOCK_DEVICE_IO_STATUS_OK) {
3503 3504
        bs->iostatus = error == ENOSPC ? BLOCK_DEVICE_IO_STATUS_NOSPACE :
                                         BLOCK_DEVICE_IO_STATUS_FAILED;
3505 3506 3507
    }
}

3508 3509
void bdrv_img_create(const char *filename, const char *fmt,
                     const char *base_filename, const char *base_fmt,
3510 3511
                     char *options, uint64_t img_size, int flags,
                     Error **errp, bool quiet)
J
Jes Sorensen 已提交
3512
{
3513 3514 3515 3516
    QemuOptsList *create_opts = NULL;
    QemuOpts *opts = NULL;
    const char *backing_fmt, *backing_file;
    int64_t size;
J
Jes Sorensen 已提交
3517
    BlockDriver *drv, *proto_drv;
3518
    BlockDriver *backing_drv = NULL;
3519
    Error *local_err = NULL;
J
Jes Sorensen 已提交
3520 3521 3522 3523 3524
    int ret = 0;

    /* Find driver and parse its options */
    drv = bdrv_find_format(fmt);
    if (!drv) {
3525
        error_setg(errp, "Unknown file format '%s'", fmt);
3526
        return;
J
Jes Sorensen 已提交
3527 3528
    }

3529
    proto_drv = bdrv_find_protocol(filename, true, errp);
J
Jes Sorensen 已提交
3530
    if (!proto_drv) {
3531
        return;
J
Jes Sorensen 已提交
3532 3533
    }

3534 3535 3536 3537 3538 3539 3540 3541 3542 3543 3544 3545
    if (!drv->create_opts) {
        error_setg(errp, "Format driver '%s' does not support image creation",
                   drv->format_name);
        return;
    }

    if (!proto_drv->create_opts) {
        error_setg(errp, "Protocol driver '%s' does not support image creation",
                   proto_drv->format_name);
        return;
    }

C
Chunyan Liu 已提交
3546 3547
    create_opts = qemu_opts_append(create_opts, drv->create_opts);
    create_opts = qemu_opts_append(create_opts, proto_drv->create_opts);
J
Jes Sorensen 已提交
3548 3549

    /* Create parameter list with default values */
3550
    opts = qemu_opts_create(create_opts, NULL, 0, &error_abort);
3551
    qemu_opt_set_number(opts, BLOCK_OPT_SIZE, img_size, &error_abort);
J
Jes Sorensen 已提交
3552 3553 3554

    /* Parse -o options */
    if (options) {
3555 3556 3557 3558
        qemu_opts_do_parse(opts, options, NULL, &local_err);
        if (local_err) {
            error_report_err(local_err);
            local_err = NULL;
3559
            error_setg(errp, "Invalid options for file format '%s'", fmt);
J
Jes Sorensen 已提交
3560 3561 3562 3563 3564
            goto out;
        }
    }

    if (base_filename) {
3565
        qemu_opt_set(opts, BLOCK_OPT_BACKING_FILE, base_filename, &local_err);
3566
        if (local_err) {
3567 3568
            error_setg(errp, "Backing file not supported for file format '%s'",
                       fmt);
J
Jes Sorensen 已提交
3569 3570 3571 3572 3573
            goto out;
        }
    }

    if (base_fmt) {
3574
        qemu_opt_set(opts, BLOCK_OPT_BACKING_FMT, base_fmt, &local_err);
3575
        if (local_err) {
3576 3577
            error_setg(errp, "Backing file format not supported for file "
                             "format '%s'", fmt);
J
Jes Sorensen 已提交
3578 3579 3580 3581
            goto out;
        }
    }

3582 3583 3584
    backing_file = qemu_opt_get(opts, BLOCK_OPT_BACKING_FILE);
    if (backing_file) {
        if (!strcmp(filename, backing_file)) {
3585 3586
            error_setg(errp, "Error: Trying to create an image with the "
                             "same filename as the backing file");
3587 3588 3589 3590
            goto out;
        }
    }

3591 3592 3593
    backing_fmt = qemu_opt_get(opts, BLOCK_OPT_BACKING_FMT);
    if (backing_fmt) {
        backing_drv = bdrv_find_format(backing_fmt);
3594
        if (!backing_drv) {
3595
            error_setg(errp, "Unknown backing file format '%s'",
3596
                       backing_fmt);
J
Jes Sorensen 已提交
3597 3598 3599 3600 3601 3602
            goto out;
        }
    }

    // The size for the image must always be specified, with one exception:
    // If we are using a backing file, we can obtain the size from there
3603 3604 3605
    size = qemu_opt_get_size(opts, BLOCK_OPT_SIZE, 0);
    if (size == -1) {
        if (backing_file) {
3606
            BlockDriverState *bs;
3607
            char *full_backing = g_new0(char, PATH_MAX);
3608
            int64_t size;
3609 3610
            int back_flags;

3611 3612 3613 3614 3615 3616 3617 3618
            bdrv_get_full_backing_filename_from_filename(filename, backing_file,
                                                         full_backing, PATH_MAX,
                                                         &local_err);
            if (local_err) {
                g_free(full_backing);
                goto out;
            }

3619 3620 3621
            /* backing files always opened read-only */
            back_flags =
                flags & ~(BDRV_O_RDWR | BDRV_O_SNAPSHOT | BDRV_O_NO_BACKING);
J
Jes Sorensen 已提交
3622

3623
            bs = NULL;
3624
            ret = bdrv_open(&bs, full_backing, NULL, NULL, back_flags,
3625
                            backing_drv, &local_err);
3626
            g_free(full_backing);
J
Jes Sorensen 已提交
3627 3628 3629
            if (ret < 0) {
                goto out;
            }
3630 3631 3632 3633 3634 3635 3636
            size = bdrv_getlength(bs);
            if (size < 0) {
                error_setg_errno(errp, -size, "Could not get size of '%s'",
                                 backing_file);
                bdrv_unref(bs);
                goto out;
            }
J
Jes Sorensen 已提交
3637

3638
            qemu_opt_set_number(opts, BLOCK_OPT_SIZE, size, &error_abort);
3639 3640

            bdrv_unref(bs);
J
Jes Sorensen 已提交
3641
        } else {
3642
            error_setg(errp, "Image creation needs a size parameter");
J
Jes Sorensen 已提交
3643 3644 3645 3646
            goto out;
        }
    }

3647
    if (!quiet) {
3648 3649
        printf("Formatting '%s', fmt=%s", filename, fmt);
        qemu_opts_print(opts, " ");
3650 3651
        puts("");
    }
3652

C
Chunyan Liu 已提交
3653
    ret = bdrv_create(drv, filename, opts, &local_err);
3654

3655 3656 3657 3658 3659
    if (ret == -EFBIG) {
        /* This is generally a better message than whatever the driver would
         * deliver (especially because of the cluster_size_hint), since that
         * is most probably not much different from "image too large". */
        const char *cluster_size_hint = "";
3660
        if (qemu_opt_get_size(opts, BLOCK_OPT_CLUSTER_SIZE, 0)) {
3661
            cluster_size_hint = " (try using a larger cluster size)";
J
Jes Sorensen 已提交
3662
        }
3663 3664 3665 3666
        error_setg(errp, "The image size is too large for file format '%s'"
                   "%s", fmt, cluster_size_hint);
        error_free(local_err);
        local_err = NULL;
J
Jes Sorensen 已提交
3667 3668 3669
    }

out:
3670 3671
    qemu_opts_del(opts);
    qemu_opts_free(create_opts);
3672
    if (local_err) {
3673 3674
        error_propagate(errp, local_err);
    }
J
Jes Sorensen 已提交
3675
}
3676 3677 3678

AioContext *bdrv_get_aio_context(BlockDriverState *bs)
{
3679 3680 3681 3682 3683
    return bs->aio_context;
}

void bdrv_detach_aio_context(BlockDriverState *bs)
{
M
Max Reitz 已提交
3684 3685
    BdrvAioNotifier *baf;

3686 3687 3688 3689
    if (!bs->drv) {
        return;
    }

M
Max Reitz 已提交
3690 3691 3692 3693
    QLIST_FOREACH(baf, &bs->aio_notifiers, list) {
        baf->detach_aio_context(baf->opaque);
    }

3694 3695 3696
    if (bs->io_limits_enabled) {
        throttle_detach_aio_context(&bs->throttle_state);
    }
3697 3698 3699 3700 3701 3702 3703 3704 3705 3706 3707 3708 3709 3710 3711 3712
    if (bs->drv->bdrv_detach_aio_context) {
        bs->drv->bdrv_detach_aio_context(bs);
    }
    if (bs->file) {
        bdrv_detach_aio_context(bs->file);
    }
    if (bs->backing_hd) {
        bdrv_detach_aio_context(bs->backing_hd);
    }

    bs->aio_context = NULL;
}

void bdrv_attach_aio_context(BlockDriverState *bs,
                             AioContext *new_context)
{
M
Max Reitz 已提交
3713 3714
    BdrvAioNotifier *ban;

3715 3716 3717 3718 3719 3720 3721 3722 3723 3724 3725 3726 3727 3728 3729
    if (!bs->drv) {
        return;
    }

    bs->aio_context = new_context;

    if (bs->backing_hd) {
        bdrv_attach_aio_context(bs->backing_hd, new_context);
    }
    if (bs->file) {
        bdrv_attach_aio_context(bs->file, new_context);
    }
    if (bs->drv->bdrv_attach_aio_context) {
        bs->drv->bdrv_attach_aio_context(bs, new_context);
    }
3730 3731 3732
    if (bs->io_limits_enabled) {
        throttle_attach_aio_context(&bs->throttle_state, new_context);
    }
M
Max Reitz 已提交
3733 3734 3735 3736

    QLIST_FOREACH(ban, &bs->aio_notifiers, list) {
        ban->attached_aio_context(new_context, ban->opaque);
    }
3737 3738 3739 3740 3741 3742 3743 3744 3745 3746 3747 3748 3749 3750
}

void bdrv_set_aio_context(BlockDriverState *bs, AioContext *new_context)
{
    bdrv_drain_all(); /* ensure there are no in-flight requests */

    bdrv_detach_aio_context(bs);

    /* This function executes in the old AioContext so acquire the new one in
     * case it runs in a different thread.
     */
    aio_context_acquire(new_context);
    bdrv_attach_aio_context(bs, new_context);
    aio_context_release(new_context);
3751
}
3752

M
Max Reitz 已提交
3753 3754 3755 3756 3757 3758 3759 3760 3761 3762 3763 3764 3765 3766 3767 3768 3769 3770 3771 3772 3773 3774 3775 3776 3777 3778 3779 3780 3781 3782 3783 3784 3785 3786 3787 3788 3789
void bdrv_add_aio_context_notifier(BlockDriverState *bs,
        void (*attached_aio_context)(AioContext *new_context, void *opaque),
        void (*detach_aio_context)(void *opaque), void *opaque)
{
    BdrvAioNotifier *ban = g_new(BdrvAioNotifier, 1);
    *ban = (BdrvAioNotifier){
        .attached_aio_context = attached_aio_context,
        .detach_aio_context   = detach_aio_context,
        .opaque               = opaque
    };

    QLIST_INSERT_HEAD(&bs->aio_notifiers, ban, list);
}

void bdrv_remove_aio_context_notifier(BlockDriverState *bs,
                                      void (*attached_aio_context)(AioContext *,
                                                                   void *),
                                      void (*detach_aio_context)(void *),
                                      void *opaque)
{
    BdrvAioNotifier *ban, *ban_next;

    QLIST_FOREACH_SAFE(ban, &bs->aio_notifiers, list, ban_next) {
        if (ban->attached_aio_context == attached_aio_context &&
            ban->detach_aio_context   == detach_aio_context   &&
            ban->opaque               == opaque)
        {
            QLIST_REMOVE(ban, list);
            g_free(ban);

            return;
        }
    }

    abort();
}

3790 3791
int bdrv_amend_options(BlockDriverState *bs, QemuOpts *opts,
                       BlockDriverAmendStatusCB *status_cb)
M
Max Reitz 已提交
3792
{
C
Chunyan Liu 已提交
3793
    if (!bs->drv->bdrv_amend_options) {
M
Max Reitz 已提交
3794 3795
        return -ENOTSUP;
    }
3796
    return bs->drv->bdrv_amend_options(bs, opts, status_cb);
M
Max Reitz 已提交
3797
}
3798

3799 3800 3801 3802
/* This function will be called by the bdrv_recurse_is_first_non_filter method
 * of block filter and by bdrv_is_first_non_filter.
 * It is used to test if the given bs is the candidate or recurse more in the
 * node graph.
3803
 */
3804
bool bdrv_recurse_is_first_non_filter(BlockDriverState *bs,
3805
                                      BlockDriverState *candidate)
3806
{
3807 3808
    /* return false if basic checks fails */
    if (!bs || !bs->drv) {
3809
        return false;
3810 3811
    }

3812 3813 3814 3815 3816
    /* the code reached a non block filter driver -> check if the bs is
     * the same as the candidate. It's the recursion termination condition.
     */
    if (!bs->drv->is_filter) {
        return bs == candidate;
3817
    }
3818
    /* Down this path the driver is a block filter driver */
3819

3820 3821 3822 3823
    /* If the block filter recursion method is defined use it to recurse down
     * the node graph.
     */
    if (bs->drv->bdrv_recurse_is_first_non_filter) {
3824
        return bs->drv->bdrv_recurse_is_first_non_filter(bs, candidate);
3825 3826
    }

3827 3828 3829
    /* the driver is a block filter but don't allow to recurse -> return false
     */
    return false;
3830 3831
}

3832 3833 3834 3835 3836
/* This function checks if the candidate is the first non filter bs down it's
 * bs chain. Since we don't have pointers to parents it explore all bs chains
 * from the top. Some filters can choose not to pass down the recursion.
 */
bool bdrv_is_first_non_filter(BlockDriverState *candidate)
3837
{
3838 3839 3840 3841 3842 3843
    BlockDriverState *bs;

    /* walk down the bs forest recursively */
    QTAILQ_FOREACH(bs, &bdrv_states, device_list) {
        bool perm;

3844
        /* try to recurse in this top level bs */
3845
        perm = bdrv_recurse_is_first_non_filter(bs, candidate);
3846 3847 3848 3849 3850 3851 3852 3853

        /* candidate is the first non filter */
        if (perm) {
            return true;
        }
    }

    return false;
3854
}
3855 3856 3857 3858

BlockDriverState *check_to_replace_node(const char *node_name, Error **errp)
{
    BlockDriverState *to_replace_bs = bdrv_find_node(node_name);
3859 3860
    AioContext *aio_context;

3861 3862 3863 3864 3865
    if (!to_replace_bs) {
        error_setg(errp, "Node name '%s' not found", node_name);
        return NULL;
    }

3866 3867 3868
    aio_context = bdrv_get_aio_context(to_replace_bs);
    aio_context_acquire(aio_context);

3869
    if (bdrv_op_is_blocked(to_replace_bs, BLOCK_OP_TYPE_REPLACE, errp)) {
3870 3871
        to_replace_bs = NULL;
        goto out;
3872 3873 3874 3875 3876 3877 3878 3879 3880
    }

    /* We don't want arbitrary node of the BDS chain to be replaced only the top
     * most non filter in order to prevent data corruption.
     * Another benefit is that this tests exclude backing files which are
     * blocked by the backing blockers.
     */
    if (!bdrv_is_first_non_filter(to_replace_bs)) {
        error_setg(errp, "Only top most non filter can be replaced");
3881 3882
        to_replace_bs = NULL;
        goto out;
3883 3884
    }

3885 3886
out:
    aio_context_release(aio_context);
3887 3888
    return to_replace_bs;
}
3889

M
Max Reitz 已提交
3890 3891 3892 3893 3894 3895 3896 3897 3898 3899 3900 3901 3902 3903 3904 3905 3906 3907 3908 3909 3910 3911 3912 3913 3914 3915 3916 3917 3918 3919 3920 3921 3922 3923 3924 3925 3926 3927 3928 3929 3930 3931 3932 3933 3934 3935 3936 3937 3938 3939 3940 3941 3942 3943 3944 3945 3946 3947 3948 3949 3950 3951 3952 3953 3954 3955 3956 3957 3958 3959 3960 3961 3962 3963 3964 3965 3966 3967 3968 3969 3970 3971 3972 3973 3974 3975 3976 3977 3978 3979 3980 3981 3982 3983 3984 3985 3986 3987 3988 3989 3990 3991 3992 3993 3994 3995 3996 3997 3998 3999 4000 4001 4002 4003 4004 4005 4006 4007 4008 4009 4010 4011 4012 4013 4014 4015 4016 4017 4018
static bool append_open_options(QDict *d, BlockDriverState *bs)
{
    const QDictEntry *entry;
    bool found_any = false;

    for (entry = qdict_first(bs->options); entry;
         entry = qdict_next(bs->options, entry))
    {
        /* Only take options for this level and exclude all non-driver-specific
         * options */
        if (!strchr(qdict_entry_key(entry), '.') &&
            strcmp(qdict_entry_key(entry), "node-name"))
        {
            qobject_incref(qdict_entry_value(entry));
            qdict_put_obj(d, qdict_entry_key(entry), qdict_entry_value(entry));
            found_any = true;
        }
    }

    return found_any;
}

/* Updates the following BDS fields:
 *  - exact_filename: A filename which may be used for opening a block device
 *                    which (mostly) equals the given BDS (even without any
 *                    other options; so reading and writing must return the same
 *                    results, but caching etc. may be different)
 *  - full_open_options: Options which, when given when opening a block device
 *                       (without a filename), result in a BDS (mostly)
 *                       equalling the given one
 *  - filename: If exact_filename is set, it is copied here. Otherwise,
 *              full_open_options is converted to a JSON object, prefixed with
 *              "json:" (for use through the JSON pseudo protocol) and put here.
 */
void bdrv_refresh_filename(BlockDriverState *bs)
{
    BlockDriver *drv = bs->drv;
    QDict *opts;

    if (!drv) {
        return;
    }

    /* This BDS's file name will most probably depend on its file's name, so
     * refresh that first */
    if (bs->file) {
        bdrv_refresh_filename(bs->file);
    }

    if (drv->bdrv_refresh_filename) {
        /* Obsolete information is of no use here, so drop the old file name
         * information before refreshing it */
        bs->exact_filename[0] = '\0';
        if (bs->full_open_options) {
            QDECREF(bs->full_open_options);
            bs->full_open_options = NULL;
        }

        drv->bdrv_refresh_filename(bs);
    } else if (bs->file) {
        /* Try to reconstruct valid information from the underlying file */
        bool has_open_options;

        bs->exact_filename[0] = '\0';
        if (bs->full_open_options) {
            QDECREF(bs->full_open_options);
            bs->full_open_options = NULL;
        }

        opts = qdict_new();
        has_open_options = append_open_options(opts, bs);

        /* If no specific options have been given for this BDS, the filename of
         * the underlying file should suffice for this one as well */
        if (bs->file->exact_filename[0] && !has_open_options) {
            strcpy(bs->exact_filename, bs->file->exact_filename);
        }
        /* Reconstructing the full options QDict is simple for most format block
         * drivers, as long as the full options are known for the underlying
         * file BDS. The full options QDict of that file BDS should somehow
         * contain a representation of the filename, therefore the following
         * suffices without querying the (exact_)filename of this BDS. */
        if (bs->file->full_open_options) {
            qdict_put_obj(opts, "driver",
                          QOBJECT(qstring_from_str(drv->format_name)));
            QINCREF(bs->file->full_open_options);
            qdict_put_obj(opts, "file", QOBJECT(bs->file->full_open_options));

            bs->full_open_options = opts;
        } else {
            QDECREF(opts);
        }
    } else if (!bs->full_open_options && qdict_size(bs->options)) {
        /* There is no underlying file BDS (at least referenced by BDS.file),
         * so the full options QDict should be equal to the options given
         * specifically for this block device when it was opened (plus the
         * driver specification).
         * Because those options don't change, there is no need to update
         * full_open_options when it's already set. */

        opts = qdict_new();
        append_open_options(opts, bs);
        qdict_put_obj(opts, "driver",
                      QOBJECT(qstring_from_str(drv->format_name)));

        if (bs->exact_filename[0]) {
            /* This may not work for all block protocol drivers (some may
             * require this filename to be parsed), but we have to find some
             * default solution here, so just include it. If some block driver
             * does not support pure options without any filename at all or
             * needs some special format of the options QDict, it needs to
             * implement the driver-specific bdrv_refresh_filename() function.
             */
            qdict_put_obj(opts, "filename",
                          QOBJECT(qstring_from_str(bs->exact_filename)));
        }

        bs->full_open_options = opts;
    }

    if (bs->exact_filename[0]) {
        pstrcpy(bs->filename, sizeof(bs->filename), bs->exact_filename);
    } else if (bs->full_open_options) {
        QString *json = qobject_to_json(QOBJECT(bs->full_open_options));
        snprintf(bs->filename, sizeof(bs->filename), "json:%s",
                 qstring_get_str(json));
        QDECREF(json);
    }
}
4019 4020 4021 4022 4023 4024 4025 4026 4027 4028 4029

/* This accessor function purpose is to allow the device models to access the
 * BlockAcctStats structure embedded inside a BlockDriverState without being
 * aware of the BlockDriverState structure layout.
 * It will go away when the BlockAcctStats structure will be moved inside
 * the device models.
 */
BlockAcctStats *bdrv_get_stats(BlockDriverState *bs)
{
    return &bs->stats;
}