block.c 123.9 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
#include "monitor/monitor.h"
28 29
#include "block/block_int.h"
#include "block/blockjob.h"
30
#include "qemu/module.h"
31
#include "qapi/qmp/qjson.h"
32
#include "sysemu/sysemu.h"
33
#include "qemu/notify.h"
34
#include "block/coroutine.h"
L
Luiz Capitulino 已提交
35
#include "qmp-commands.h"
36
#include "qemu/timer.h"
B
bellard 已提交
37

J
Juan Quintela 已提交
38
#ifdef CONFIG_BSD
B
bellard 已提交
39 40 41
#include <sys/types.h>
#include <sys/stat.h>
#include <sys/ioctl.h>
B
Blue Swirl 已提交
42
#include <sys/queue.h>
43
#ifndef __DragonFly__
B
bellard 已提交
44 45
#include <sys/disk.h>
#endif
46
#endif
B
bellard 已提交
47

48 49 50 51
#ifdef _WIN32
#include <windows.h>
#endif

52 53
#define NOT_DONE 0x7fffffff /* used while emulated sync operation in progress */

54 55
typedef enum {
    BDRV_REQ_COPY_ON_READ = 0x1,
56
    BDRV_REQ_ZERO_WRITE   = 0x2,
57 58
} BdrvRequestFlags;

59
static void bdrv_dev_change_media_cb(BlockDriverState *bs, bool load);
60 61
static BlockDriverAIOCB *bdrv_aio_readv_em(BlockDriverState *bs,
        int64_t sector_num, QEMUIOVector *qiov, int nb_sectors,
62
        BlockDriverCompletionFunc *cb, void *opaque);
63 64
static BlockDriverAIOCB *bdrv_aio_writev_em(BlockDriverState *bs,
        int64_t sector_num, QEMUIOVector *qiov, int nb_sectors,
65
        BlockDriverCompletionFunc *cb, void *opaque);
66 67 68 69 70 71
static int coroutine_fn bdrv_co_readv_em(BlockDriverState *bs,
                                         int64_t sector_num, int nb_sectors,
                                         QEMUIOVector *iov);
static int coroutine_fn bdrv_co_writev_em(BlockDriverState *bs,
                                         int64_t sector_num, int nb_sectors,
                                         QEMUIOVector *iov);
72
static int coroutine_fn bdrv_co_do_readv(BlockDriverState *bs,
73 74
    int64_t sector_num, int nb_sectors, QEMUIOVector *qiov,
    BdrvRequestFlags flags);
75
static int coroutine_fn bdrv_co_do_writev(BlockDriverState *bs,
76 77
    int64_t sector_num, int nb_sectors, QEMUIOVector *qiov,
    BdrvRequestFlags flags);
78 79 80 81 82 83
static BlockDriverAIOCB *bdrv_co_aio_rw_vector(BlockDriverState *bs,
                                               int64_t sector_num,
                                               QEMUIOVector *qiov,
                                               int nb_sectors,
                                               BlockDriverCompletionFunc *cb,
                                               void *opaque,
84
                                               bool is_write);
85
static void coroutine_fn bdrv_co_do_rw(void *opaque);
K
Kevin Wolf 已提交
86 87
static int coroutine_fn bdrv_co_do_write_zeroes(BlockDriverState *bs,
    int64_t sector_num, int nb_sectors);
B
bellard 已提交
88

Z
Zhi Yong Wu 已提交
89 90 91 92 93 94 95
static bool bdrv_exceed_bps_limits(BlockDriverState *bs, int nb_sectors,
        bool is_write, double elapsed_time, uint64_t *wait);
static bool bdrv_exceed_iops_limits(BlockDriverState *bs, bool is_write,
        double elapsed_time, uint64_t *wait);
static bool bdrv_exceed_io_limits(BlockDriverState *bs, int nb_sectors,
        bool is_write, int64_t *wait);

96 97
static QTAILQ_HEAD(, BlockDriverState) bdrv_states =
    QTAILQ_HEAD_INITIALIZER(bdrv_states);
98

99 100
static QLIST_HEAD(, BlockDriver) bdrv_drivers =
    QLIST_HEAD_INITIALIZER(bdrv_drivers);
B
bellard 已提交
101

102 103 104
/* The device to use for VM snapshots */
static BlockDriverState *bs_snapshots;

105 106 107
/* If non-zero, use only whitelisted block drivers */
static int use_bdrv_whitelist;

108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127
#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

128
/* throttling disk I/O limits */
Z
Zhi Yong Wu 已提交
129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146
void bdrv_io_limits_disable(BlockDriverState *bs)
{
    bs->io_limits_enabled = false;

    while (qemu_co_queue_next(&bs->throttled_reqs));

    if (bs->block_timer) {
        qemu_del_timer(bs->block_timer);
        qemu_free_timer(bs->block_timer);
        bs->block_timer = NULL;
    }

    bs->slice_start = 0;
    bs->slice_end   = 0;
    bs->slice_time  = 0;
    memset(&bs->io_base, 0, sizeof(bs->io_base));
}

147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171
static void bdrv_block_timer(void *opaque)
{
    BlockDriverState *bs = opaque;

    qemu_co_queue_next(&bs->throttled_reqs);
}

void bdrv_io_limits_enable(BlockDriverState *bs)
{
    qemu_co_queue_init(&bs->throttled_reqs);
    bs->block_timer = qemu_new_timer_ns(vm_clock, bdrv_block_timer, bs);
    bs->io_limits_enabled = true;
}

bool bdrv_io_limits_enabled(BlockDriverState *bs)
{
    BlockIOLimit *io_limits = &bs->io_limits;
    return io_limits->bps[BLOCK_IO_LIMIT_READ]
         || io_limits->bps[BLOCK_IO_LIMIT_WRITE]
         || io_limits->bps[BLOCK_IO_LIMIT_TOTAL]
         || io_limits->iops[BLOCK_IO_LIMIT_READ]
         || io_limits->iops[BLOCK_IO_LIMIT_WRITE]
         || io_limits->iops[BLOCK_IO_LIMIT_TOTAL];
}

Z
Zhi Yong Wu 已提交
172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196
static void bdrv_io_limits_intercept(BlockDriverState *bs,
                                     bool is_write, int nb_sectors)
{
    int64_t wait_time = -1;

    if (!qemu_co_queue_empty(&bs->throttled_reqs)) {
        qemu_co_queue_wait(&bs->throttled_reqs);
    }

    /* In fact, we hope to keep each request's timing, in FIFO mode. The next
     * throttled requests will not be dequeued until the current request is
     * allowed to be serviced. So if the current request still exceeds the
     * limits, it will be inserted to the head. All requests followed it will
     * be still in throttled_reqs queue.
     */

    while (bdrv_exceed_io_limits(bs, nb_sectors, is_write, &wait_time)) {
        qemu_mod_timer(bs->block_timer,
                       wait_time + qemu_get_clock_ns(vm_clock));
        qemu_co_queue_wait_insert_head(&bs->throttled_reqs);
    }

    qemu_co_queue_next(&bs->throttled_reqs);
}

197 198 199
/* check if the path starts with "<protocol>:" */
static int path_has_protocol(const char *path)
{
200 201
    const char *p;

202 203 204 205 206
#ifdef _WIN32
    if (is_windows_drive(path) ||
        is_windows_drive_prefix(path)) {
        return 0;
    }
207 208 209
    p = path + strcspn(path, ":/\\");
#else
    p = path + strcspn(path, ":/");
210 211
#endif

212
    return *p == ':';
213 214
}

B
bellard 已提交
215
int path_is_absolute(const char *path)
216
{
B
bellard 已提交
217 218
#ifdef _WIN32
    /* specific case for names like: "\\.\d:" */
P
Paolo Bonzini 已提交
219
    if (is_windows_drive(path) || is_windows_drive_prefix(path)) {
B
bellard 已提交
220
        return 1;
P
Paolo Bonzini 已提交
221 222
    }
    return (*path == '/' || *path == '\\');
223
#else
P
Paolo Bonzini 已提交
224
    return (*path == '/');
225
#endif
226 227
}

B
bellard 已提交
228 229 230 231 232 233
/* 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)
234
{
B
bellard 已提交
235 236 237 238 239 240 241 242 243 244 245 246 247
    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;
248 249 250 251 252 253 254 255 256
        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
257 258 259 260 261 262 263 264 265 266 267 268
        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);
269 270 271
    }
}

272 273 274 275 276 277 278 279 280
void bdrv_get_full_backing_filename(BlockDriverState *bs, char *dest, size_t sz)
{
    if (bs->backing_file[0] == '\0' || path_has_protocol(bs->backing_file)) {
        pstrcpy(dest, sz, bs->backing_file);
    } else {
        path_combine(dest, sz, bs->filename, bs->backing_file);
    }
}

281
void bdrv_register(BlockDriver *bdrv)
B
bellard 已提交
282
{
283 284
    /* Block drivers without coroutine functions need emulation */
    if (!bdrv->bdrv_co_readv) {
285 286 287
        bdrv->bdrv_co_readv = bdrv_co_readv_em;
        bdrv->bdrv_co_writev = bdrv_co_writev_em;

288 289 290
        /* bdrv_co_readv_em()/brdv_co_writev_em() work in terms of aio, so if
         * the block driver lacks aio we need to emulate that too.
         */
291 292 293 294 295
        if (!bdrv->bdrv_aio_readv) {
            /* add AIO emulation layer */
            bdrv->bdrv_aio_readv = bdrv_aio_readv_em;
            bdrv->bdrv_aio_writev = bdrv_aio_writev_em;
        }
B
bellard 已提交
296
    }
297

298
    QLIST_INSERT_HEAD(&bdrv_drivers, bdrv, list);
B
bellard 已提交
299
}
B
bellard 已提交
300 301 302 303

/* create a new block device (by default it is empty) */
BlockDriverState *bdrv_new(const char *device_name)
{
304
    BlockDriverState *bs;
B
bellard 已提交
305

306
    bs = g_malloc0(sizeof(BlockDriverState));
B
bellard 已提交
307
    pstrcpy(bs->device_name, sizeof(bs->device_name), device_name);
B
bellard 已提交
308
    if (device_name[0] != '\0') {
309
        QTAILQ_INSERT_TAIL(&bdrv_states, bs, list);
B
bellard 已提交
310
    }
311
    bdrv_iostatus_disable(bs);
P
Paolo Bonzini 已提交
312 313
    notifier_list_init(&bs->close_notifiers);

B
bellard 已提交
314 315 316
    return bs;
}

P
Paolo Bonzini 已提交
317 318 319 320 321
void bdrv_add_close_notifier(BlockDriverState *bs, Notifier *notify)
{
    notifier_list_add(&bs->close_notifiers, notify);
}

B
bellard 已提交
322 323 324
BlockDriver *bdrv_find_format(const char *format_name)
{
    BlockDriver *drv1;
325 326
    QLIST_FOREACH(drv1, &bdrv_drivers, list) {
        if (!strcmp(drv1->format_name, format_name)) {
B
bellard 已提交
327
            return drv1;
328
        }
B
bellard 已提交
329 330 331 332
    }
    return NULL;
}

333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356
static int bdrv_is_whitelisted(BlockDriver *drv)
{
    static const char *whitelist[] = {
        CONFIG_BDRV_WHITELIST
    };
    const char **p;

    if (!whitelist[0])
        return 1;               /* no whitelist, anything goes */

    for (p = whitelist; *p; p++) {
        if (!strcmp(drv->format_name, *p)) {
            return 1;
        }
    }
    return 0;
}

BlockDriver *bdrv_find_whitelisted_format(const char *format_name)
{
    BlockDriver *drv = bdrv_find_format(format_name);
    return drv && bdrv_is_whitelisted(drv) ? drv : NULL;
}

357 358 359 360 361 362 363 364 365 366 367 368 369 370 371
typedef struct CreateCo {
    BlockDriver *drv;
    char *filename;
    QEMUOptionParameter *options;
    int ret;
} CreateCo;

static void coroutine_fn bdrv_create_co_entry(void *opaque)
{
    CreateCo *cco = opaque;
    assert(cco->drv);

    cco->ret = cco->drv->bdrv_create(cco->filename, cco->options);
}

372 373
int bdrv_create(BlockDriver *drv, const char* filename,
    QEMUOptionParameter *options)
B
bellard 已提交
374
{
375 376 377 378 379 380 381 382 383 384 385
    int ret;

    Coroutine *co;
    CreateCo cco = {
        .drv = drv,
        .filename = g_strdup(filename),
        .options = options,
        .ret = NOT_DONE,
    };

    if (!drv->bdrv_create) {
386 387
        ret = -ENOTSUP;
        goto out;
388 389 390 391 392 393 394 395 396 397 398 399 400 401
    }

    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) {
            qemu_aio_wait();
        }
    }

    ret = cco.ret;
402

403 404
out:
    g_free(cco.filename);
405
    return ret;
B
bellard 已提交
406 407
}

408 409 410 411
int bdrv_create_file(const char* filename, QEMUOptionParameter *options)
{
    BlockDriver *drv;

412
    drv = bdrv_find_protocol(filename);
413
    if (drv == NULL) {
414
        return -ENOENT;
415 416 417 418 419
    }

    return bdrv_create(drv, filename, options);
}

420 421 422 423 424
/*
 * 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
425
{
426
#ifdef _WIN32
427
    char temp_dir[MAX_PATH];
428 429 430 431 432 433
    /* 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
434
#else
B
bellard 已提交
435
    int fd;
436
    const char *tmpdir;
A
aurel32 已提交
437 438 439
    tmpdir = getenv("TMPDIR");
    if (!tmpdir)
        tmpdir = "/tmp";
440 441 442
    if (snprintf(filename, size, "%s/vl.XXXXXX", tmpdir) >= size) {
        return -EOVERFLOW;
    }
B
bellard 已提交
443
    fd = mkstemp(filename);
444 445 446 447 448
    if (fd < 0) {
        return -errno;
    }
    if (close(fd) != 0) {
        unlink(filename);
449 450 451
        return -errno;
    }
    return 0;
B
bellard 已提交
452
#endif
453
}
B
bellard 已提交
454

455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476
/*
 * 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;
}

477
BlockDriver *bdrv_find_protocol(const char *filename)
B
bellard 已提交
478 479 480
{
    BlockDriver *drv1;
    char protocol[128];
481
    int len;
B
bellard 已提交
482
    const char *p;
B
bellard 已提交
483

484 485
    /* TODO Drivers without bdrv_file_open must be specified explicitly */

486 487 488 489 490 491 492 493 494 495 496 497
    /*
     * 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;
    }

498
    if (!path_has_protocol(filename)) {
499
        return bdrv_find_format("file");
500
    }
501 502
    p = strchr(filename, ':');
    assert(p != NULL);
503 504 505 506 507
    len = p - filename;
    if (len > sizeof(protocol) - 1)
        len = sizeof(protocol) - 1;
    memcpy(protocol, filename, len);
    protocol[len] = '\0';
508
    QLIST_FOREACH(drv1, &bdrv_drivers, list) {
509
        if (drv1->protocol_name &&
510
            !strcmp(drv1->protocol_name, protocol)) {
B
bellard 已提交
511
            return drv1;
512
        }
B
bellard 已提交
513 514 515 516
    }
    return NULL;
}

517 518
static int find_image_format(BlockDriverState *bs, const char *filename,
                             BlockDriver **pdrv)
519
{
520
    int score, score_max;
521 522
    BlockDriver *drv1, *drv;
    uint8_t buf[2048];
523
    int ret = 0;
524

525
    /* Return the raw BlockDriver * to scsi-generic devices or empty drives */
526
    if (bs->sg || !bdrv_is_inserted(bs) || bdrv_getlength(bs) == 0) {
527 528 529 530 531 532
        drv = bdrv_find_format("raw");
        if (!drv) {
            ret = -ENOENT;
        }
        *pdrv = drv;
        return ret;
533
    }
534

B
bellard 已提交
535 536
    ret = bdrv_pread(bs, 0, buf, sizeof(buf));
    if (ret < 0) {
537 538
        *pdrv = NULL;
        return ret;
B
bellard 已提交
539 540
    }

B
bellard 已提交
541
    score_max = 0;
542
    drv = NULL;
543
    QLIST_FOREACH(drv1, &bdrv_drivers, list) {
B
bellard 已提交
544 545 546 547 548 549
        if (drv1->bdrv_probe) {
            score = drv1->bdrv_probe(buf, ret, filename);
            if (score > score_max) {
                score_max = score;
                drv = drv1;
            }
B
bellard 已提交
550
        }
B
bellard 已提交
551
    }
552 553 554 555 556
    if (!drv) {
        ret = -ENOENT;
    }
    *pdrv = drv;
    return ret;
B
bellard 已提交
557 558
}

559 560 561 562 563 564 565
/**
 * Set the current 'total_sectors' value
 */
static int refresh_total_sectors(BlockDriverState *bs, int64_t hint)
{
    BlockDriver *drv = bs->drv;

566 567 568 569
    /* Do not attempt drv->bdrv_getlength() on scsi-generic devices */
    if (bs->sg)
        return 0;

570 571 572 573 574 575 576 577 578 579 580 581 582
    /* 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;
        }
        hint = length >> BDRV_SECTOR_BITS;
    }

    bs->total_sectors = hint;
    return 0;
}

583 584 585 586 587 588 589 590 591 592 593
/**
 * 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;
594 595
    } else if (!strcmp(mode, "directsync")) {
        *flags |= BDRV_O_NOCACHE;
596 597 598 599 600 601 602 603 604 605 606 607 608 609
    } 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;
}

610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625
/**
 * The copy-on-read flag is actually a reference count so multiple users may
 * use the feature without worrying about clobbering its previous state.
 * Copy-on-read stays enabled until all users have called to disable it.
 */
void bdrv_enable_copy_on_read(BlockDriverState *bs)
{
    bs->copy_on_read++;
}

void bdrv_disable_copy_on_read(BlockDriverState *bs)
{
    assert(bs->copy_on_read > 0);
    bs->copy_on_read--;
}

K
Kevin Wolf 已提交
626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645
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.
     */
    open_flags &= ~(BDRV_O_SNAPSHOT | BDRV_O_NO_BACKING);

    /*
     * Snapshots should be writable.
     */
    if (bs->is_temporary) {
        open_flags |= BDRV_O_RDWR;
    }

    return open_flags;
}

646 647 648
/*
 * Common part for opening disk images and files
 */
649 650
static int bdrv_open_common(BlockDriverState *bs, BlockDriverState *file,
    const char *filename,
651 652 653 654 655
    int flags, BlockDriver *drv)
{
    int ret, open_flags;

    assert(drv != NULL);
656
    assert(bs->file == NULL);
657

S
Stefan Hajnoczi 已提交
658 659
    trace_bdrv_open_common(bs, filename, flags, drv->format_name);

660 661 662
    bs->open_flags = flags;
    bs->buffer_alignment = 512;

663 664 665 666 667
    assert(bs->copy_on_read == 0); /* bdrv_new() and bdrv_close() make it so */
    if ((flags & BDRV_O_RDWR) && (flags & BDRV_O_COPY_ON_READ)) {
        bdrv_enable_copy_on_read(bs);
    }

668 669 670 671 672 673 674
    pstrcpy(bs->filename, sizeof(bs->filename), filename);

    if (use_bdrv_whitelist && !bdrv_is_whitelisted(drv)) {
        return -ENOTSUP;
    }

    bs->drv = drv;
675
    bs->opaque = g_malloc0(drv->instance_size);
676

677
    bs->enable_write_cache = !!(flags & BDRV_O_CACHE_WB);
K
Kevin Wolf 已提交
678
    open_flags = bdrv_open_flags(bs, flags);
679

680
    bs->read_only = !(open_flags & BDRV_O_RDWR);
681

682 683
    /* Open the image, either directly or using a protocol */
    if (drv->bdrv_file_open) {
684 685 686 687 688
        if (file != NULL) {
            bdrv_swap(file, bs);
            ret = 0;
        } else {
            ret = drv->bdrv_file_open(bs, filename, open_flags);
689
        }
690 691 692 693
    } else {
        assert(file != NULL);
        bs->file = file;
        ret = drv->bdrv_open(bs, open_flags);
694 695
    }

696 697 698 699
    if (ret < 0) {
        goto free_and_fail;
    }

700 701 702
    ret = refresh_total_sectors(bs, bs->total_sectors);
    if (ret < 0) {
        goto free_and_fail;
703
    }
704

705 706 707 708 709 710 711 712
#ifndef _WIN32
    if (bs->is_temporary) {
        unlink(filename);
    }
#endif
    return 0;

free_and_fail:
713
    bs->file = NULL;
714
    g_free(bs->opaque);
715 716 717 718 719
    bs->opaque = NULL;
    bs->drv = NULL;
    return ret;
}

K
Kevin Wolf 已提交
720 721 722
/*
 * Opens a file using a protocol (file, host_device, nbd, ...)
 */
B
bellard 已提交
723
int bdrv_file_open(BlockDriverState **pbs, const char *filename, int flags)
B
bellard 已提交
724
{
B
bellard 已提交
725
    BlockDriverState *bs;
726
    BlockDriver *drv;
B
bellard 已提交
727 728
    int ret;

729
    drv = bdrv_find_protocol(filename);
730 731 732 733
    if (!drv) {
        return -ENOENT;
    }

B
bellard 已提交
734
    bs = bdrv_new("");
735
    ret = bdrv_open_common(bs, NULL, filename, flags, drv);
B
bellard 已提交
736 737 738
    if (ret < 0) {
        bdrv_delete(bs);
        return ret;
739
    }
740
    bs->growable = 1;
B
bellard 已提交
741 742 743 744
    *pbs = bs;
    return 0;
}

P
Paolo Bonzini 已提交
745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780
int bdrv_open_backing_file(BlockDriverState *bs)
{
    char backing_filename[PATH_MAX];
    int back_flags, ret;
    BlockDriver *back_drv = NULL;

    if (bs->backing_hd != NULL) {
        return 0;
    }

    bs->open_flags &= ~BDRV_O_NO_BACKING;
    if (bs->backing_file[0] == '\0') {
        return 0;
    }

    bs->backing_hd = bdrv_new("");
    bdrv_get_full_backing_filename(bs, backing_filename,
                                   sizeof(backing_filename));

    if (bs->backing_format[0] != '\0') {
        back_drv = bdrv_find_format(bs->backing_format);
    }

    /* backing files always opened read-only */
    back_flags = bs->open_flags & ~(BDRV_O_RDWR | BDRV_O_SNAPSHOT);

    ret = bdrv_open(bs->backing_hd, backing_filename, back_flags, back_drv);
    if (ret < 0) {
        bdrv_delete(bs->backing_hd);
        bs->backing_hd = NULL;
        bs->open_flags |= BDRV_O_NO_BACKING;
        return ret;
    }
    return 0;
}

K
Kevin Wolf 已提交
781 782 783
/*
 * Opens a disk image (raw, qcow2, vmdk, ...)
 */
K
Kevin Wolf 已提交
784 785
int bdrv_open(BlockDriverState *bs, const char *filename, int flags,
              BlockDriver *drv)
B
bellard 已提交
786
{
K
Kevin Wolf 已提交
787
    int ret;
788 789
    /* TODO: extra byte is a hack to ensure MAX_PATH space on Windows. */
    char tmp_filename[PATH_MAX + 1];
790
    BlockDriverState *file = NULL;
B
bellard 已提交
791

B
bellard 已提交
792
    if (flags & BDRV_O_SNAPSHOT) {
B
bellard 已提交
793 794
        BlockDriverState *bs1;
        int64_t total_size;
A
aliguori 已提交
795
        int is_protocol = 0;
K
Kevin Wolf 已提交
796 797
        BlockDriver *bdrv_qcow2;
        QEMUOptionParameter *options;
K
Kevin Wolf 已提交
798
        char backing_filename[PATH_MAX];
799

B
bellard 已提交
800 801
        /* if snapshot, we create a temporary backing file and open it
           instead of opening 'filename' directly */
802

B
bellard 已提交
803 804
        /* if there is a backing file, use it */
        bs1 = bdrv_new("");
K
Kevin Wolf 已提交
805
        ret = bdrv_open(bs1, filename, 0, drv);
806
        if (ret < 0) {
B
bellard 已提交
807
            bdrv_delete(bs1);
808
            return ret;
B
bellard 已提交
809
        }
810
        total_size = bdrv_getlength(bs1) & BDRV_SECTOR_MASK;
A
aliguori 已提交
811 812 813 814

        if (bs1->drv && bs1->drv->protocol_name)
            is_protocol = 1;

B
bellard 已提交
815
        bdrv_delete(bs1);
816

817 818 819 820
        ret = get_tmp_filename(tmp_filename, sizeof(tmp_filename));
        if (ret < 0) {
            return ret;
        }
A
aliguori 已提交
821 822 823 824 825

        /* Real path is meaningless for protocols */
        if (is_protocol)
            snprintf(backing_filename, sizeof(backing_filename),
                     "%s", filename);
826 827
        else if (!realpath(filename, backing_filename))
            return -errno;
A
aliguori 已提交
828

K
Kevin Wolf 已提交
829 830 831
        bdrv_qcow2 = bdrv_find_format("qcow2");
        options = parse_option_parameters("", bdrv_qcow2->create_options, NULL);

832
        set_option_parameter_int(options, BLOCK_OPT_SIZE, total_size);
K
Kevin Wolf 已提交
833 834 835 836 837 838 839
        set_option_parameter(options, BLOCK_OPT_BACKING_FILE, backing_filename);
        if (drv) {
            set_option_parameter(options, BLOCK_OPT_BACKING_FMT,
                drv->format_name);
        }

        ret = bdrv_create(bdrv_qcow2, tmp_filename, options);
840
        free_option_parameters(options);
841 842
        if (ret < 0) {
            return ret;
B
bellard 已提交
843
        }
K
Kevin Wolf 已提交
844

B
bellard 已提交
845
        filename = tmp_filename;
K
Kevin Wolf 已提交
846
        drv = bdrv_qcow2;
B
bellard 已提交
847 848
        bs->is_temporary = 1;
    }
B
bellard 已提交
849

850 851 852 853 854 855 856 857 858 859
    /* Open image file without format layer */
    if (flags & BDRV_O_RDWR) {
        flags |= BDRV_O_ALLOW_RDWR;
    }

    ret = bdrv_file_open(&file, filename, bdrv_open_flags(bs, flags));
    if (ret < 0) {
        return ret;
    }

K
Kevin Wolf 已提交
860
    /* Find the right image format driver */
861
    if (!drv) {
862
        ret = find_image_format(file, filename, &drv);
863
    }
864

865 866
    if (!drv) {
        goto unlink_and_fail;
B
bellard 已提交
867
    }
K
Kevin Wolf 已提交
868 869

    /* Open the image */
870
    ret = bdrv_open_common(bs, file, filename, flags, drv);
K
Kevin Wolf 已提交
871
    if (ret < 0) {
872 873 874
        goto unlink_and_fail;
    }

875 876 877 878 879
    if (bs->file != file) {
        bdrv_delete(file);
        file = NULL;
    }

K
Kevin Wolf 已提交
880
    /* If there is a backing file, use it */
P
Paolo Bonzini 已提交
881 882
    if ((flags & BDRV_O_NO_BACKING) == 0) {
        ret = bdrv_open_backing_file(bs);
K
Kevin Wolf 已提交
883 884 885 886 887 888 889
        if (ret < 0) {
            bdrv_close(bs);
            return ret;
        }
    }

    if (!bdrv_key_required(bs)) {
890
        bdrv_dev_change_media_cb(bs, true);
K
Kevin Wolf 已提交
891 892
    }

Z
Zhi Yong Wu 已提交
893 894 895 896 897
    /* throttling disk I/O limits */
    if (bs->io_limits_enabled) {
        bdrv_io_limits_enable(bs);
    }

K
Kevin Wolf 已提交
898 899 900
    return 0;

unlink_and_fail:
901 902 903
    if (file != NULL) {
        bdrv_delete(file);
    }
K
Kevin Wolf 已提交
904 905 906 907 908 909
    if (bs->is_temporary) {
        unlink(filename);
    }
    return ret;
}

910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141
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);
    }

    if (bs->file) {
        bdrv_reopen_queue(bs_queue, bs->file, flags);
    }

    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) {
        error_set(errp, QERR_DEVICE_IS_READ_ONLY,
                  reopen_state->bs->device_name);
        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 {
                error_set(errp, QERR_OPEN_FILE_FAILED,
                          reopen_state->bs->filename);
            }
            goto error;
        }
    } else {
        /* It is currently mandatory to have a bdrv_reopen_prepare()
         * handler for each supported drv. */
        error_set(errp, QERR_BLOCK_FORMAT_FEATURE_NOT_SUPPORTED,
                  drv->format_name, reopen_state->bs->device_name,
                 "reopening of file");
        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);
}

/*
 * 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
1142 1143
void bdrv_close(BlockDriverState *bs)
{
1144
    bdrv_flush(bs);
1145 1146 1147 1148
    if (bs->job) {
        block_job_cancel_sync(bs->job);
    }
    bdrv_drain_all();
P
Paolo Bonzini 已提交
1149
    notifier_list_notify(&bs->close_notifiers, bs);
K
Kevin Wolf 已提交
1150

1151
    if (bs->drv) {
1152 1153 1154
        if (bs == bs_snapshots) {
            bs_snapshots = NULL;
        }
1155
        if (bs->backing_hd) {
B
bellard 已提交
1156
            bdrv_delete(bs->backing_hd);
1157 1158
            bs->backing_hd = NULL;
        }
B
bellard 已提交
1159
        bs->drv->bdrv_close(bs);
1160
        g_free(bs->opaque);
B
bellard 已提交
1161 1162 1163 1164
#ifdef _WIN32
        if (bs->is_temporary) {
            unlink(bs->filename);
        }
B
bellard 已提交
1165
#endif
B
bellard 已提交
1166 1167
        bs->opaque = NULL;
        bs->drv = NULL;
1168
        bs->copy_on_read = 0;
1169 1170
        bs->backing_file[0] = '\0';
        bs->backing_format[0] = '\0';
1171 1172 1173 1174 1175
        bs->total_sectors = 0;
        bs->encrypted = 0;
        bs->valid_key = 0;
        bs->sg = 0;
        bs->growable = 0;
B
bellard 已提交
1176

1177
        if (bs->file != NULL) {
1178 1179
            bdrv_delete(bs->file);
            bs->file = NULL;
1180
        }
B
bellard 已提交
1181
    }
Z
Zhi Yong Wu 已提交
1182

P
Pavel Hrdina 已提交
1183 1184
    bdrv_dev_change_media_cb(bs, false);

Z
Zhi Yong Wu 已提交
1185 1186 1187 1188
    /*throttling disk I/O limits*/
    if (bs->io_limits_enabled) {
        bdrv_io_limits_disable(bs);
    }
B
bellard 已提交
1189 1190
}

1191 1192 1193 1194 1195 1196 1197 1198 1199
void bdrv_close_all(void)
{
    BlockDriverState *bs;

    QTAILQ_FOREACH(bs, &bdrv_states, list) {
        bdrv_close(bs);
    }
}

1200 1201 1202 1203 1204
/*
 * Wait for pending requests to complete across all BlockDriverStates
 *
 * This function does not flush data to disk, use bdrv_flush_all() for that
 * after calling this function.
1205 1206 1207 1208 1209 1210
 *
 * Note that completion of an asynchronous I/O operation can trigger any
 * number of other I/O operations on other devices---for example a coroutine
 * can be arbitrarily complex and a constant flow of I/O can come until the
 * coroutine is complete.  Because of this, it is not possible to have a
 * function to drain a single device's I/O queue.
1211 1212 1213 1214
 */
void bdrv_drain_all(void)
{
    BlockDriverState *bs;
1215 1216 1217 1218
    bool busy;

    do {
        busy = qemu_aio_wait();
1219

1220 1221 1222 1223 1224 1225 1226 1227 1228 1229
        /* FIXME: We do not have timer support here, so this is effectively
         * a busy wait.
         */
        QTAILQ_FOREACH(bs, &bdrv_states, list) {
            if (!qemu_co_queue_empty(&bs->throttled_reqs)) {
                qemu_co_queue_restart_all(&bs->throttled_reqs);
                busy = true;
            }
        }
    } while (busy);
1230 1231 1232 1233 1234 1235 1236 1237

    /* If requests are still pending there is a bug somewhere */
    QTAILQ_FOREACH(bs, &bdrv_states, list) {
        assert(QLIST_EMPTY(&bs->tracked_requests));
        assert(qemu_co_queue_empty(&bs->throttled_reqs));
    }
}

1238 1239 1240 1241 1242 1243 1244 1245 1246 1247
/* make a BlockDriverState anonymous by removing from bdrv_state list.
   Also, NULL terminate the device_name to prevent double remove */
void bdrv_make_anon(BlockDriverState *bs)
{
    if (bs->device_name[0] != '\0') {
        QTAILQ_REMOVE(&bdrv_states, bs, list);
    }
    bs->device_name[0] = '\0';
}

P
Paolo Bonzini 已提交
1248 1249 1250 1251 1252 1253 1254
static void bdrv_rebind(BlockDriverState *bs)
{
    if (bs->drv && bs->drv->bdrv_rebind) {
        bs->drv->bdrv_rebind(bs);
    }
}

1255 1256
static void bdrv_move_feature_fields(BlockDriverState *bs_dest,
                                     BlockDriverState *bs_src)
1257
{
1258 1259
    /* move some fields that need to stay attached to the device */
    bs_dest->open_flags         = bs_src->open_flags;
1260 1261

    /* dev info */
1262 1263 1264 1265 1266
    bs_dest->dev_ops            = bs_src->dev_ops;
    bs_dest->dev_opaque         = bs_src->dev_opaque;
    bs_dest->dev                = bs_src->dev;
    bs_dest->buffer_alignment   = bs_src->buffer_alignment;
    bs_dest->copy_on_read       = bs_src->copy_on_read;
1267

1268
    bs_dest->enable_write_cache = bs_src->enable_write_cache;
1269

1270
    /* i/o timing parameters */
1271 1272 1273 1274 1275 1276 1277 1278
    bs_dest->slice_time         = bs_src->slice_time;
    bs_dest->slice_start        = bs_src->slice_start;
    bs_dest->slice_end          = bs_src->slice_end;
    bs_dest->io_limits          = bs_src->io_limits;
    bs_dest->io_base            = bs_src->io_base;
    bs_dest->throttled_reqs     = bs_src->throttled_reqs;
    bs_dest->block_timer        = bs_src->block_timer;
    bs_dest->io_limits_enabled  = bs_src->io_limits_enabled;
1279 1280

    /* r/w error */
1281 1282
    bs_dest->on_read_error      = bs_src->on_read_error;
    bs_dest->on_write_error     = bs_src->on_write_error;
1283 1284

    /* i/o status */
1285 1286
    bs_dest->iostatus_enabled   = bs_src->iostatus_enabled;
    bs_dest->iostatus           = bs_src->iostatus;
1287

1288
    /* dirty bitmap */
1289
    bs_dest->dirty_bitmap       = bs_src->dirty_bitmap;
1290 1291

    /* job */
1292 1293
    bs_dest->in_use             = bs_src->in_use;
    bs_dest->job                = bs_src->job;
1294

1295
    /* keep the same entry in bdrv_states */
1296 1297 1298 1299
    pstrcpy(bs_dest->device_name, sizeof(bs_dest->device_name),
            bs_src->device_name);
    bs_dest->list = bs_src->list;
}
1300

1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315
/*
 * 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.
 *
 * bs_new is required to be anonymous.
 *
 * 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
1316

1317 1318 1319 1320 1321 1322 1323 1324
    /* bs_new must be anonymous and shouldn't have anything fancy enabled */
    assert(bs_new->device_name[0] == '\0');
    assert(bs_new->dirty_bitmap == NULL);
    assert(bs_new->job == NULL);
    assert(bs_new->dev == NULL);
    assert(bs_new->in_use == 0);
    assert(bs_new->io_limits_enabled == false);
    assert(bs_new->block_timer == NULL);
1325

1326 1327 1328
    tmp = *bs_new;
    *bs_new = *bs_old;
    *bs_old = tmp;
1329

1330 1331 1332 1333
    /* 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);
1334

1335 1336 1337 1338 1339 1340 1341 1342 1343
    /* bs_new shouldn't be in bdrv_states even after the swap!  */
    assert(bs_new->device_name[0] == '\0');

    /* Check a few fields that should remain attached to the device */
    assert(bs_new->dev == NULL);
    assert(bs_new->job == NULL);
    assert(bs_new->in_use == 0);
    assert(bs_new->io_limits_enabled == false);
    assert(bs_new->block_timer == NULL);
P
Paolo Bonzini 已提交
1344 1345

    bdrv_rebind(bs_new);
1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371
    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.
 *
 * bs_new is required to be anonymous.
 *
 * 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. */
    bs_top->backing_hd = bs_new;
    bs_top->open_flags &= ~BDRV_O_NO_BACKING;
    pstrcpy(bs_top->backing_file, sizeof(bs_top->backing_file),
            bs_new->filename);
    pstrcpy(bs_top->backing_format, sizeof(bs_top->backing_format),
            bs_new->drv ? bs_new->drv->format_name : "");
1372 1373
}

B
bellard 已提交
1374 1375
void bdrv_delete(BlockDriverState *bs)
{
1376
    assert(!bs->dev);
1377 1378
    assert(!bs->job);
    assert(!bs->in_use);
1379

1380
    /* remove from list, if necessary */
1381
    bdrv_make_anon(bs);
1382

B
bellard 已提交
1383
    bdrv_close(bs);
1384

1385
    assert(bs != bs_snapshots);
1386
    g_free(bs);
B
bellard 已提交
1387 1388
}

1389 1390
int bdrv_attach_dev(BlockDriverState *bs, void *dev)
/* TODO change to DeviceState *dev when all users are qdevified */
1391
{
1392
    if (bs->dev) {
1393 1394
        return -EBUSY;
    }
1395
    bs->dev = dev;
1396
    bdrv_iostatus_reset(bs);
1397 1398 1399
    return 0;
}

1400 1401
/* TODO qdevified devices don't use this, remove when devices are qdevified */
void bdrv_attach_dev_nofail(BlockDriverState *bs, void *dev)
1402
{
1403 1404 1405 1406 1407 1408 1409 1410 1411 1412
    if (bdrv_attach_dev(bs, dev) < 0) {
        abort();
    }
}

void bdrv_detach_dev(BlockDriverState *bs, void *dev)
/* TODO change to DeviceState *dev when all users are qdevified */
{
    assert(bs->dev == dev);
    bs->dev = NULL;
1413 1414
    bs->dev_ops = NULL;
    bs->dev_opaque = NULL;
1415
    bs->buffer_alignment = 512;
1416 1417
}

1418 1419
/* TODO change to return DeviceState * when all users are qdevified */
void *bdrv_get_attached_dev(BlockDriverState *bs)
1420
{
1421
    return bs->dev;
1422 1423
}

1424 1425 1426 1427 1428
void bdrv_set_dev_ops(BlockDriverState *bs, const BlockDevOps *ops,
                      void *opaque)
{
    bs->dev_ops = ops;
    bs->dev_opaque = opaque;
1429 1430 1431
    if (bdrv_dev_has_removable_media(bs) && bs == bs_snapshots) {
        bs_snapshots = NULL;
    }
1432 1433
}

P
Paolo Bonzini 已提交
1434 1435 1436
void bdrv_emit_qmp_error_event(const BlockDriverState *bdrv,
                               enum MonitorEvent ev,
                               BlockErrorAction action, bool is_read)
1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458
{
    QObject *data;
    const char *action_str;

    switch (action) {
    case BDRV_ACTION_REPORT:
        action_str = "report";
        break;
    case BDRV_ACTION_IGNORE:
        action_str = "ignore";
        break;
    case BDRV_ACTION_STOP:
        action_str = "stop";
        break;
    default:
        abort();
    }

    data = qobject_from_jsonf("{ 'device': %s, 'action': %s, 'operation': %s }",
                              bdrv->device_name,
                              action_str,
                              is_read ? "read" : "write");
P
Paolo Bonzini 已提交
1459
    monitor_protocol_event(ev, data);
1460 1461 1462 1463

    qobject_decref(data);
}

1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474
static void bdrv_emit_qmp_eject_event(BlockDriverState *bs, bool ejected)
{
    QObject *data;

    data = qobject_from_jsonf("{ 'device': %s, 'tray-open': %i }",
                              bdrv_get_device_name(bs), ejected);
    monitor_protocol_event(QEVENT_DEVICE_TRAY_MOVED, data);

    qobject_decref(data);
}

1475
static void bdrv_dev_change_media_cb(BlockDriverState *bs, bool load)
1476
{
1477
    if (bs->dev_ops && bs->dev_ops->change_media_cb) {
1478
        bool tray_was_closed = !bdrv_dev_is_tray_open(bs);
1479
        bs->dev_ops->change_media_cb(bs->dev_opaque, load);
1480 1481 1482 1483 1484 1485 1486 1487
        if (tray_was_closed) {
            /* tray open */
            bdrv_emit_qmp_eject_event(bs, true);
        }
        if (load) {
            /* tray close */
            bdrv_emit_qmp_eject_event(bs, false);
        }
1488 1489 1490
    }
}

1491 1492 1493 1494 1495
bool bdrv_dev_has_removable_media(BlockDriverState *bs)
{
    return !bs->dev || (bs->dev_ops && bs->dev_ops->change_media_cb);
}

P
Paolo Bonzini 已提交
1496 1497 1498 1499 1500 1501 1502
void bdrv_dev_eject_request(BlockDriverState *bs, bool force)
{
    if (bs->dev_ops && bs->dev_ops->eject_request_cb) {
        bs->dev_ops->eject_request_cb(bs->dev_opaque, force);
    }
}

1503 1504 1505 1506 1507 1508 1509 1510
bool bdrv_dev_is_tray_open(BlockDriverState *bs)
{
    if (bs->dev_ops && bs->dev_ops->is_tray_open) {
        return bs->dev_ops->is_tray_open(bs->dev_opaque);
    }
    return false;
}

1511 1512 1513 1514
static void bdrv_dev_resize_cb(BlockDriverState *bs)
{
    if (bs->dev_ops && bs->dev_ops->resize_cb) {
        bs->dev_ops->resize_cb(bs->dev_opaque);
1515 1516 1517
    }
}

1518 1519 1520 1521 1522 1523 1524 1525
bool bdrv_dev_is_medium_locked(BlockDriverState *bs)
{
    if (bs->dev_ops && bs->dev_ops->is_medium_locked) {
        return bs->dev_ops->is_medium_locked(bs->dev_opaque);
    }
    return false;
}

A
aliguori 已提交
1526 1527 1528
/*
 * Run consistency checks on an image
 *
1529
 * Returns 0 if the check could be completed (it doesn't mean that the image is
1530
 * free of errors) or -errno when an internal error occurred. The results of the
1531
 * check are stored in res.
A
aliguori 已提交
1532
 */
1533
int bdrv_check(BlockDriverState *bs, BdrvCheckResult *res, BdrvCheckMode fix)
A
aliguori 已提交
1534 1535 1536 1537 1538
{
    if (bs->drv->bdrv_check == NULL) {
        return -ENOTSUP;
    }

1539
    memset(res, 0, sizeof(*res));
1540
    return bs->drv->bdrv_check(bs, res, fix);
A
aliguori 已提交
1541 1542
}

1543 1544
#define COMMIT_BUF_SECTORS 2048

1545 1546 1547
/* commit COW file into the raw image */
int bdrv_commit(BlockDriverState *bs)
{
B
bellard 已提交
1548
    BlockDriver *drv = bs->drv;
1549 1550
    int64_t sector, total_sectors;
    int n, ro, open_flags;
1551
    int ret = 0;
1552
    uint8_t *buf;
1553
    char filename[PATH_MAX];
1554

B
bellard 已提交
1555 1556
    if (!drv)
        return -ENOMEDIUM;
1557 1558 1559
    
    if (!bs->backing_hd) {
        return -ENOTSUP;
1560 1561
    }

1562 1563 1564 1565
    if (bdrv_in_use(bs) || bdrv_in_use(bs->backing_hd)) {
        return -EBUSY;
    }

1566
    ro = bs->backing_hd->read_only;
1567 1568
    /* Use pstrcpy (not strncpy): filename must be NUL-terminated. */
    pstrcpy(filename, sizeof(filename), bs->backing_hd->filename);
1569 1570 1571
    open_flags =  bs->backing_hd->open_flags;

    if (ro) {
1572 1573
        if (bdrv_reopen(bs->backing_hd, open_flags | BDRV_O_RDWR, NULL)) {
            return -EACCES;
1574
        }
B
bellard 已提交
1575
    }
1576

J
Jan Kiszka 已提交
1577
    total_sectors = bdrv_getlength(bs) >> BDRV_SECTOR_BITS;
1578
    buf = g_malloc(COMMIT_BUF_SECTORS * BDRV_SECTOR_SIZE);
1579 1580

    for (sector = 0; sector < total_sectors; sector += n) {
1581
        if (bdrv_is_allocated(bs, sector, COMMIT_BUF_SECTORS, &n)) {
1582 1583 1584 1585 1586 1587 1588 1589 1590 1591

            if (bdrv_read(bs, sector, buf, n) != 0) {
                ret = -EIO;
                goto ro_cleanup;
            }

            if (bdrv_write(bs->backing_hd, sector, buf, n) != 0) {
                ret = -EIO;
                goto ro_cleanup;
            }
B
bellard 已提交
1592
        }
1593
    }
1594

1595 1596 1597 1598
    if (drv->bdrv_make_empty) {
        ret = drv->bdrv_make_empty(bs);
        bdrv_flush(bs);
    }
1599

1600 1601 1602 1603 1604 1605
    /*
     * Make sure all data we wrote to the backing device is actually
     * stable on disk.
     */
    if (bs->backing_hd)
        bdrv_flush(bs->backing_hd);
1606 1607

ro_cleanup:
1608
    g_free(buf);
1609 1610

    if (ro) {
1611 1612
        /* ignoring error return here */
        bdrv_reopen(bs->backing_hd, open_flags & ~BDRV_O_RDWR, NULL);
1613 1614
    }

1615
    return ret;
1616 1617
}

1618
int bdrv_commit_all(void)
1619 1620 1621 1622
{
    BlockDriverState *bs;

    QTAILQ_FOREACH(bs, &bdrv_states, list) {
1623 1624 1625 1626
        int ret = bdrv_commit(bs);
        if (ret < 0) {
            return ret;
        }
1627
    }
1628
    return 0;
1629 1630
}

S
Stefan Hajnoczi 已提交
1631 1632 1633 1634 1635 1636
struct BdrvTrackedRequest {
    BlockDriverState *bs;
    int64_t sector_num;
    int nb_sectors;
    bool is_write;
    QLIST_ENTRY(BdrvTrackedRequest) list;
1637
    Coroutine *co; /* owner, used for deadlock detection */
1638
    CoQueue wait_queue; /* coroutines blocked on this request */
S
Stefan Hajnoczi 已提交
1639 1640 1641 1642 1643 1644 1645 1646 1647 1648
};

/**
 * Remove an active request from the tracked requests list
 *
 * This function should be called when a tracked request is completing.
 */
static void tracked_request_end(BdrvTrackedRequest *req)
{
    QLIST_REMOVE(req, list);
1649
    qemu_co_queue_restart_all(&req->wait_queue);
S
Stefan Hajnoczi 已提交
1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664
}

/**
 * Add an active request to the tracked requests list
 */
static void tracked_request_begin(BdrvTrackedRequest *req,
                                  BlockDriverState *bs,
                                  int64_t sector_num,
                                  int nb_sectors, bool is_write)
{
    *req = (BdrvTrackedRequest){
        .bs = bs,
        .sector_num = sector_num,
        .nb_sectors = nb_sectors,
        .is_write = is_write,
1665
        .co = qemu_coroutine_self(),
S
Stefan Hajnoczi 已提交
1666 1667
    };

1668 1669
    qemu_co_queue_init(&req->wait_queue);

S
Stefan Hajnoczi 已提交
1670 1671 1672
    QLIST_INSERT_HEAD(&bs->tracked_requests, req, list);
}

S
Stefan Hajnoczi 已提交
1673 1674 1675
/**
 * Round a region to cluster boundaries
 */
1676 1677 1678 1679
void bdrv_round_to_clusters(BlockDriverState *bs,
                            int64_t sector_num, int nb_sectors,
                            int64_t *cluster_sector_num,
                            int *cluster_nb_sectors)
S
Stefan Hajnoczi 已提交
1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693
{
    BlockDriverInfo bdi;

    if (bdrv_get_info(bs, &bdi) < 0 || bdi.cluster_size == 0) {
        *cluster_sector_num = sector_num;
        *cluster_nb_sectors = nb_sectors;
    } else {
        int64_t c = bdi.cluster_size / BDRV_SECTOR_SIZE;
        *cluster_sector_num = QEMU_ALIGN_DOWN(sector_num, c);
        *cluster_nb_sectors = QEMU_ALIGN_UP(sector_num - *cluster_sector_num +
                                            nb_sectors, c);
    }
}

1694 1695
static bool tracked_request_overlaps(BdrvTrackedRequest *req,
                                     int64_t sector_num, int nb_sectors) {
S
Stefan Hajnoczi 已提交
1696 1697 1698 1699 1700 1701 1702 1703 1704
    /*        aaaa   bbbb */
    if (sector_num >= req->sector_num + req->nb_sectors) {
        return false;
    }
    /* bbbb   aaaa        */
    if (req->sector_num >= sector_num + nb_sectors) {
        return false;
    }
    return true;
1705 1706 1707 1708 1709 1710
}

static void coroutine_fn wait_for_overlapping_requests(BlockDriverState *bs,
        int64_t sector_num, int nb_sectors)
{
    BdrvTrackedRequest *req;
S
Stefan Hajnoczi 已提交
1711 1712
    int64_t cluster_sector_num;
    int cluster_nb_sectors;
1713 1714
    bool retry;

S
Stefan Hajnoczi 已提交
1715 1716 1717 1718 1719 1720
    /* If we touch the same cluster it counts as an overlap.  This guarantees
     * that allocating writes will be serialized and not race with each other
     * for the same cluster.  For example, in copy-on-read it ensures that the
     * CoR read and write operations are atomic and guest writes cannot
     * interleave between them.
     */
1721 1722
    bdrv_round_to_clusters(bs, sector_num, nb_sectors,
                           &cluster_sector_num, &cluster_nb_sectors);
S
Stefan Hajnoczi 已提交
1723

1724 1725 1726
    do {
        retry = false;
        QLIST_FOREACH(req, &bs->tracked_requests, list) {
S
Stefan Hajnoczi 已提交
1727 1728
            if (tracked_request_overlaps(req, cluster_sector_num,
                                         cluster_nb_sectors)) {
1729 1730 1731 1732 1733 1734
                /* Hitting this means there was a reentrant request, for
                 * example, a block driver issuing nested requests.  This must
                 * never happen since it means deadlock.
                 */
                assert(qemu_coroutine_self() != req->co);

1735 1736 1737 1738 1739 1740 1741 1742
                qemu_co_queue_wait(&req->wait_queue);
                retry = true;
                break;
            }
        }
    } while (retry);
}

K
Kevin Wolf 已提交
1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754
/*
 * 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;
1755
    int ret;
K
Kevin Wolf 已提交
1756

1757 1758 1759 1760 1761
    /* Backing file format doesn't make sense without a backing file */
    if (backing_fmt && !backing_file) {
        return -EINVAL;
    }

K
Kevin Wolf 已提交
1762
    if (drv->bdrv_change_backing_file != NULL) {
1763
        ret = drv->bdrv_change_backing_file(bs, backing_file, backing_fmt);
K
Kevin Wolf 已提交
1764
    } else {
1765
        ret = -ENOTSUP;
K
Kevin Wolf 已提交
1766
    }
1767 1768 1769 1770 1771 1772

    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
1773 1774
}

1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917
/*
 * 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.
 */
BlockDriverState *bdrv_find_overlay(BlockDriverState *active,
                                    BlockDriverState *bs)
{
    BlockDriverState *overlay = NULL;
    BlockDriverState *intermediate;

    assert(active != NULL);
    assert(bs != NULL);

    /* if bs is the same as active, then by definition it has no overlay
     */
    if (active == bs) {
        return NULL;
    }

    intermediate = active;
    while (intermediate->backing_hd) {
        if (intermediate->backing_hd == bs) {
            overlay = intermediate;
            break;
        }
        intermediate = intermediate->backing_hd;
    }

    return overlay;
}

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
 *
 * Error conditions:
 *  if active == top, that is considered an error
 *
 */
int bdrv_drop_intermediate(BlockDriverState *active, BlockDriverState *top,
                           BlockDriverState *base)
{
    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) {
        intermediate_state = g_malloc0(sizeof(BlkIntermediateStates));
        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 */
    ret = bdrv_change_backing_file(new_top_bs, base_bs->filename,
                                   base_bs->drv ? base_bs->drv->format_name : "");
    if (ret) {
        goto exit;
    }
    new_top_bs->backing_hd = base_bs;


    QSIMPLEQ_FOREACH_SAFE(intermediate_state, &states_to_delete, entry, next) {
        /* so that bdrv_close() does not recursively close the chain */
        intermediate_state->bs->backing_hd = NULL;
        bdrv_delete(intermediate_state->bs);
    }
    ret = 0;

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


1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930
static int bdrv_check_byte_request(BlockDriverState *bs, int64_t offset,
                                   size_t size)
{
    int64_t len;

    if (!bdrv_is_inserted(bs))
        return -ENOMEDIUM;

    if (bs->growable)
        return 0;

    len = bdrv_getlength(bs);

K
Kevin Wolf 已提交
1931 1932 1933 1934
    if (offset < 0)
        return -EIO;

    if ((offset > len) || (len - offset < size))
1935 1936 1937 1938 1939 1940 1941 1942
        return -EIO;

    return 0;
}

static int bdrv_check_request(BlockDriverState *bs, int64_t sector_num,
                              int nb_sectors)
{
1943 1944
    return bdrv_check_byte_request(bs, sector_num * BDRV_SECTOR_SIZE,
                                   nb_sectors * BDRV_SECTOR_SIZE);
1945 1946
}

1947 1948 1949 1950 1951 1952 1953 1954 1955 1956
typedef struct RwCo {
    BlockDriverState *bs;
    int64_t sector_num;
    int nb_sectors;
    QEMUIOVector *qiov;
    bool is_write;
    int ret;
} RwCo;

static void coroutine_fn bdrv_rw_co_entry(void *opaque)
B
bellard 已提交
1957
{
1958
    RwCo *rwco = opaque;
B
bellard 已提交
1959

1960 1961
    if (!rwco->is_write) {
        rwco->ret = bdrv_co_do_readv(rwco->bs, rwco->sector_num,
1962
                                     rwco->nb_sectors, rwco->qiov, 0);
1963 1964
    } else {
        rwco->ret = bdrv_co_do_writev(rwco->bs, rwco->sector_num,
1965
                                      rwco->nb_sectors, rwco->qiov, 0);
1966 1967
    }
}
1968

1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988
/*
 * Process a synchronous request using coroutines
 */
static int bdrv_rw_co(BlockDriverState *bs, int64_t sector_num, uint8_t *buf,
                      int nb_sectors, bool is_write)
{
    QEMUIOVector qiov;
    struct iovec iov = {
        .iov_base = (void *)buf,
        .iov_len = nb_sectors * BDRV_SECTOR_SIZE,
    };
    Coroutine *co;
    RwCo rwco = {
        .bs = bs,
        .sector_num = sector_num,
        .nb_sectors = nb_sectors,
        .qiov = &qiov,
        .is_write = is_write,
        .ret = NOT_DONE,
    };
1989

1990
    qemu_iovec_init_external(&qiov, &iov, 1);
1991

1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
    /**
     * In sync call context, when the vcpu is blocked, this throttling timer
     * will not fire; so the I/O throttling function has to be disabled here
     * if it has been enabled.
     */
    if (bs->io_limits_enabled) {
        fprintf(stderr, "Disabling I/O throttling on '%s' due "
                        "to synchronous I/O.\n", bdrv_get_device_name(bs));
        bdrv_io_limits_disable(bs);
    }

2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
    if (qemu_in_coroutine()) {
        /* Fast-path if already in coroutine context */
        bdrv_rw_co_entry(&rwco);
    } else {
        co = qemu_coroutine_create(bdrv_rw_co_entry);
        qemu_coroutine_enter(co, &rwco);
        while (rwco.ret == NOT_DONE) {
            qemu_aio_wait();
        }
    }
    return rwco.ret;
}
B
bellard 已提交
2015

2016 2017 2018 2019 2020
/* return < 0 if error. See bdrv_write() for the return codes */
int bdrv_read(BlockDriverState *bs, int64_t sector_num,
              uint8_t *buf, int nb_sectors)
{
    return bdrv_rw_co(bs, sector_num, buf, nb_sectors, false);
B
bellard 已提交
2021 2022
}

2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036
/* Just like bdrv_read(), but with I/O throttling temporarily disabled */
int bdrv_read_unthrottled(BlockDriverState *bs, int64_t sector_num,
                          uint8_t *buf, int nb_sectors)
{
    bool enabled;
    int ret;

    enabled = bs->io_limits_enabled;
    bs->io_limits_enabled = false;
    ret = bdrv_read(bs, 0, buf, 1);
    bs->io_limits_enabled = enabled;
    return ret;
}

2037
/* Return < 0 if error. Important errors are:
B
bellard 已提交
2038 2039 2040 2041 2042
  -EIO         generic I/O error (may happen for all errors)
  -ENOMEDIUM   No media inserted.
  -EINVAL      Invalid sector number or nb_sectors
  -EACCES      Trying to write a read-only device
*/
2043
int bdrv_write(BlockDriverState *bs, int64_t sector_num,
B
bellard 已提交
2044 2045
               const uint8_t *buf, int nb_sectors)
{
2046
    return bdrv_rw_co(bs, sector_num, (uint8_t *)buf, nb_sectors, true);
B
bellard 已提交
2047 2048
}

2049 2050
int bdrv_pread(BlockDriverState *bs, int64_t offset,
               void *buf, int count1)
B
bellard 已提交
2051
{
J
Jan Kiszka 已提交
2052
    uint8_t tmp_buf[BDRV_SECTOR_SIZE];
B
bellard 已提交
2053 2054
    int len, nb_sectors, count;
    int64_t sector_num;
2055
    int ret;
B
bellard 已提交
2056 2057 2058

    count = count1;
    /* first read to align to sector start */
J
Jan Kiszka 已提交
2059
    len = (BDRV_SECTOR_SIZE - offset) & (BDRV_SECTOR_SIZE - 1);
B
bellard 已提交
2060 2061
    if (len > count)
        len = count;
J
Jan Kiszka 已提交
2062
    sector_num = offset >> BDRV_SECTOR_BITS;
B
bellard 已提交
2063
    if (len > 0) {
2064 2065
        if ((ret = bdrv_read(bs, sector_num, tmp_buf, 1)) < 0)
            return ret;
J
Jan Kiszka 已提交
2066
        memcpy(buf, tmp_buf + (offset & (BDRV_SECTOR_SIZE - 1)), len);
B
bellard 已提交
2067 2068 2069 2070 2071 2072 2073 2074
        count -= len;
        if (count == 0)
            return count1;
        sector_num++;
        buf += len;
    }

    /* read the sectors "in place" */
J
Jan Kiszka 已提交
2075
    nb_sectors = count >> BDRV_SECTOR_BITS;
B
bellard 已提交
2076
    if (nb_sectors > 0) {
2077 2078
        if ((ret = bdrv_read(bs, sector_num, buf, nb_sectors)) < 0)
            return ret;
B
bellard 已提交
2079
        sector_num += nb_sectors;
J
Jan Kiszka 已提交
2080
        len = nb_sectors << BDRV_SECTOR_BITS;
B
bellard 已提交
2081 2082 2083 2084 2085 2086
        buf += len;
        count -= len;
    }

    /* add data from the last sector */
    if (count > 0) {
2087 2088
        if ((ret = bdrv_read(bs, sector_num, tmp_buf, 1)) < 0)
            return ret;
B
bellard 已提交
2089 2090 2091 2092 2093
        memcpy(buf, tmp_buf, count);
    }
    return count1;
}

2094 2095
int bdrv_pwrite(BlockDriverState *bs, int64_t offset,
                const void *buf, int count1)
B
bellard 已提交
2096
{
J
Jan Kiszka 已提交
2097
    uint8_t tmp_buf[BDRV_SECTOR_SIZE];
B
bellard 已提交
2098 2099
    int len, nb_sectors, count;
    int64_t sector_num;
2100
    int ret;
B
bellard 已提交
2101 2102 2103

    count = count1;
    /* first write to align to sector start */
J
Jan Kiszka 已提交
2104
    len = (BDRV_SECTOR_SIZE - offset) & (BDRV_SECTOR_SIZE - 1);
B
bellard 已提交
2105 2106
    if (len > count)
        len = count;
J
Jan Kiszka 已提交
2107
    sector_num = offset >> BDRV_SECTOR_BITS;
B
bellard 已提交
2108
    if (len > 0) {
2109 2110
        if ((ret = bdrv_read(bs, sector_num, tmp_buf, 1)) < 0)
            return ret;
J
Jan Kiszka 已提交
2111
        memcpy(tmp_buf + (offset & (BDRV_SECTOR_SIZE - 1)), buf, len);
2112 2113
        if ((ret = bdrv_write(bs, sector_num, tmp_buf, 1)) < 0)
            return ret;
B
bellard 已提交
2114 2115 2116 2117 2118 2119 2120 2121
        count -= len;
        if (count == 0)
            return count1;
        sector_num++;
        buf += len;
    }

    /* write the sectors "in place" */
J
Jan Kiszka 已提交
2122
    nb_sectors = count >> BDRV_SECTOR_BITS;
B
bellard 已提交
2123
    if (nb_sectors > 0) {
2124 2125
        if ((ret = bdrv_write(bs, sector_num, buf, nb_sectors)) < 0)
            return ret;
B
bellard 已提交
2126
        sector_num += nb_sectors;
J
Jan Kiszka 已提交
2127
        len = nb_sectors << BDRV_SECTOR_BITS;
B
bellard 已提交
2128 2129 2130 2131 2132 2133
        buf += len;
        count -= len;
    }

    /* add data from the last sector */
    if (count > 0) {
2134 2135
        if ((ret = bdrv_read(bs, sector_num, tmp_buf, 1)) < 0)
            return ret;
B
bellard 已提交
2136
        memcpy(tmp_buf, buf, count);
2137 2138
        if ((ret = bdrv_write(bs, sector_num, tmp_buf, 1)) < 0)
            return ret;
B
bellard 已提交
2139 2140 2141 2142
    }
    return count1;
}

K
Kevin Wolf 已提交
2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158
/*
 * Writes to the file and ensures that no writes are reordered across this
 * request (acts as a barrier)
 *
 * Returns 0 on success, -errno in error cases.
 */
int bdrv_pwrite_sync(BlockDriverState *bs, int64_t offset,
    const void *buf, int count)
{
    int ret;

    ret = bdrv_pwrite(bs, offset, buf, count);
    if (ret < 0) {
        return ret;
    }

2159 2160
    /* No flush needed for cache modes that already do it */
    if (bs->enable_write_cache) {
K
Kevin Wolf 已提交
2161 2162 2163 2164 2165 2166
        bdrv_flush(bs);
    }

    return 0;
}

2167
static int coroutine_fn bdrv_co_do_copy_on_readv(BlockDriverState *bs,
S
Stefan Hajnoczi 已提交
2168 2169 2170 2171 2172 2173 2174 2175 2176
        int64_t sector_num, int nb_sectors, QEMUIOVector *qiov)
{
    /* Perform I/O through a temporary buffer so that users who scribble over
     * their read buffer while the operation is in progress do not end up
     * modifying the image file.  This is critical for zero-copy guest I/O
     * where anything might happen inside guest memory.
     */
    void *bounce_buffer;

2177
    BlockDriver *drv = bs->drv;
S
Stefan Hajnoczi 已提交
2178 2179 2180 2181 2182 2183 2184 2185 2186 2187
    struct iovec iov;
    QEMUIOVector bounce_qiov;
    int64_t cluster_sector_num;
    int cluster_nb_sectors;
    size_t skip_bytes;
    int ret;

    /* Cover entire cluster so no additional backing file I/O is required when
     * allocating cluster in the image file.
     */
2188 2189
    bdrv_round_to_clusters(bs, sector_num, nb_sectors,
                           &cluster_sector_num, &cluster_nb_sectors);
S
Stefan Hajnoczi 已提交
2190

2191 2192
    trace_bdrv_co_do_copy_on_readv(bs, sector_num, nb_sectors,
                                   cluster_sector_num, cluster_nb_sectors);
S
Stefan Hajnoczi 已提交
2193 2194 2195 2196 2197

    iov.iov_len = cluster_nb_sectors * BDRV_SECTOR_SIZE;
    iov.iov_base = bounce_buffer = qemu_blockalign(bs, iov.iov_len);
    qemu_iovec_init_external(&bounce_qiov, &iov, 1);

2198 2199
    ret = drv->bdrv_co_readv(bs, cluster_sector_num, cluster_nb_sectors,
                             &bounce_qiov);
S
Stefan Hajnoczi 已提交
2200 2201 2202 2203
    if (ret < 0) {
        goto err;
    }

2204 2205
    if (drv->bdrv_co_write_zeroes &&
        buffer_is_zero(bounce_buffer, iov.iov_len)) {
K
Kevin Wolf 已提交
2206 2207
        ret = bdrv_co_do_write_zeroes(bs, cluster_sector_num,
                                      cluster_nb_sectors);
2208
    } else {
2209 2210 2211
        /* This does not change the data on the disk, it is not necessary
         * to flush even in cache=writethrough mode.
         */
2212
        ret = drv->bdrv_co_writev(bs, cluster_sector_num, cluster_nb_sectors,
S
Stefan Hajnoczi 已提交
2213
                                  &bounce_qiov);
2214 2215
    }

S
Stefan Hajnoczi 已提交
2216 2217 2218 2219 2220 2221 2222 2223 2224
    if (ret < 0) {
        /* It might be okay to ignore write errors for guest requests.  If this
         * is a deliberate copy-on-read then we don't want to ignore the error.
         * Simply report it in all cases.
         */
        goto err;
    }

    skip_bytes = (sector_num - cluster_sector_num) * BDRV_SECTOR_SIZE;
2225 2226
    qemu_iovec_from_buf(qiov, 0, bounce_buffer + skip_bytes,
                        nb_sectors * BDRV_SECTOR_SIZE);
S
Stefan Hajnoczi 已提交
2227 2228 2229 2230 2231 2232

err:
    qemu_vfree(bounce_buffer);
    return ret;
}

2233 2234 2235 2236
/*
 * Handle a read request in coroutine context
 */
static int coroutine_fn bdrv_co_do_readv(BlockDriverState *bs,
2237 2238
    int64_t sector_num, int nb_sectors, QEMUIOVector *qiov,
    BdrvRequestFlags flags)
K
Kevin Wolf 已提交
2239 2240
{
    BlockDriver *drv = bs->drv;
S
Stefan Hajnoczi 已提交
2241 2242
    BdrvTrackedRequest req;
    int ret;
K
Kevin Wolf 已提交
2243 2244 2245 2246 2247 2248 2249 2250

    if (!drv) {
        return -ENOMEDIUM;
    }
    if (bdrv_check_request(bs, sector_num, nb_sectors)) {
        return -EIO;
    }

Z
Zhi Yong Wu 已提交
2251 2252 2253 2254 2255
    /* throttling disk read I/O */
    if (bs->io_limits_enabled) {
        bdrv_io_limits_intercept(bs, false, nb_sectors);
    }

2256
    if (bs->copy_on_read) {
2257 2258 2259 2260 2261 2262 2263
        flags |= BDRV_REQ_COPY_ON_READ;
    }
    if (flags & BDRV_REQ_COPY_ON_READ) {
        bs->copy_on_read_in_flight++;
    }

    if (bs->copy_on_read_in_flight) {
2264 2265 2266
        wait_for_overlapping_requests(bs, sector_num, nb_sectors);
    }

S
Stefan Hajnoczi 已提交
2267
    tracked_request_begin(&req, bs, sector_num, nb_sectors, false);
S
Stefan Hajnoczi 已提交
2268

2269
    if (flags & BDRV_REQ_COPY_ON_READ) {
S
Stefan Hajnoczi 已提交
2270 2271 2272 2273 2274 2275 2276 2277
        int pnum;

        ret = bdrv_co_is_allocated(bs, sector_num, nb_sectors, &pnum);
        if (ret < 0) {
            goto out;
        }

        if (!ret || pnum != nb_sectors) {
2278
            ret = bdrv_co_do_copy_on_readv(bs, sector_num, nb_sectors, qiov);
S
Stefan Hajnoczi 已提交
2279 2280 2281 2282
            goto out;
        }
    }

S
Stefan Hajnoczi 已提交
2283
    ret = drv->bdrv_co_readv(bs, sector_num, nb_sectors, qiov);
S
Stefan Hajnoczi 已提交
2284 2285

out:
S
Stefan Hajnoczi 已提交
2286
    tracked_request_end(&req);
2287 2288 2289 2290 2291

    if (flags & BDRV_REQ_COPY_ON_READ) {
        bs->copy_on_read_in_flight--;
    }

S
Stefan Hajnoczi 已提交
2292
    return ret;
K
Kevin Wolf 已提交
2293 2294
}

2295
int coroutine_fn bdrv_co_readv(BlockDriverState *bs, int64_t sector_num,
K
Kevin Wolf 已提交
2296 2297
    int nb_sectors, QEMUIOVector *qiov)
{
2298
    trace_bdrv_co_readv(bs, sector_num, nb_sectors);
K
Kevin Wolf 已提交
2299

2300 2301 2302 2303 2304 2305 2306 2307 2308 2309
    return bdrv_co_do_readv(bs, sector_num, nb_sectors, qiov, 0);
}

int coroutine_fn bdrv_co_copy_on_readv(BlockDriverState *bs,
    int64_t sector_num, int nb_sectors, QEMUIOVector *qiov)
{
    trace_bdrv_co_copy_on_readv(bs, sector_num, nb_sectors);

    return bdrv_co_do_readv(bs, sector_num, nb_sectors, qiov,
                            BDRV_REQ_COPY_ON_READ);
2310 2311
}

2312 2313 2314 2315 2316 2317 2318 2319
static int coroutine_fn bdrv_co_do_write_zeroes(BlockDriverState *bs,
    int64_t sector_num, int nb_sectors)
{
    BlockDriver *drv = bs->drv;
    QEMUIOVector qiov;
    struct iovec iov;
    int ret;

K
Kevin Wolf 已提交
2320 2321 2322
    /* TODO Emulate only part of misaligned requests instead of letting block
     * drivers return -ENOTSUP and emulate everything */

2323 2324
    /* First try the efficient write zeroes operation */
    if (drv->bdrv_co_write_zeroes) {
K
Kevin Wolf 已提交
2325 2326 2327 2328
        ret = drv->bdrv_co_write_zeroes(bs, sector_num, nb_sectors);
        if (ret != -ENOTSUP) {
            return ret;
        }
2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342
    }

    /* Fall back to bounce buffer if write zeroes is unsupported */
    iov.iov_len  = nb_sectors * BDRV_SECTOR_SIZE;
    iov.iov_base = qemu_blockalign(bs, iov.iov_len);
    memset(iov.iov_base, 0, iov.iov_len);
    qemu_iovec_init_external(&qiov, &iov, 1);

    ret = drv->bdrv_co_writev(bs, sector_num, nb_sectors, &qiov);

    qemu_vfree(iov.iov_base);
    return ret;
}

2343 2344 2345 2346
/*
 * Handle a write request in coroutine context
 */
static int coroutine_fn bdrv_co_do_writev(BlockDriverState *bs,
2347 2348
    int64_t sector_num, int nb_sectors, QEMUIOVector *qiov,
    BdrvRequestFlags flags)
2349 2350
{
    BlockDriver *drv = bs->drv;
S
Stefan Hajnoczi 已提交
2351
    BdrvTrackedRequest req;
2352
    int ret;
K
Kevin Wolf 已提交
2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363

    if (!bs->drv) {
        return -ENOMEDIUM;
    }
    if (bs->read_only) {
        return -EACCES;
    }
    if (bdrv_check_request(bs, sector_num, nb_sectors)) {
        return -EIO;
    }

Z
Zhi Yong Wu 已提交
2364 2365 2366 2367 2368
    /* throttling disk write I/O */
    if (bs->io_limits_enabled) {
        bdrv_io_limits_intercept(bs, true, nb_sectors);
    }

2369
    if (bs->copy_on_read_in_flight) {
2370 2371 2372
        wait_for_overlapping_requests(bs, sector_num, nb_sectors);
    }

S
Stefan Hajnoczi 已提交
2373 2374
    tracked_request_begin(&req, bs, sector_num, nb_sectors, true);

2375 2376 2377 2378 2379
    if (flags & BDRV_REQ_ZERO_WRITE) {
        ret = bdrv_co_do_write_zeroes(bs, sector_num, nb_sectors);
    } else {
        ret = drv->bdrv_co_writev(bs, sector_num, nb_sectors, qiov);
    }
2380

2381 2382 2383 2384
    if (ret == 0 && !bs->enable_write_cache) {
        ret = bdrv_co_flush(bs);
    }

K
Kevin Wolf 已提交
2385
    if (bs->dirty_bitmap) {
2386
        bdrv_set_dirty(bs, sector_num, nb_sectors);
K
Kevin Wolf 已提交
2387 2388 2389 2390 2391 2392
    }

    if (bs->wr_highest_sector < sector_num + nb_sectors - 1) {
        bs->wr_highest_sector = sector_num + nb_sectors - 1;
    }

S
Stefan Hajnoczi 已提交
2393 2394
    tracked_request_end(&req);

2395
    return ret;
K
Kevin Wolf 已提交
2396 2397
}

2398 2399 2400 2401 2402
int coroutine_fn bdrv_co_writev(BlockDriverState *bs, int64_t sector_num,
    int nb_sectors, QEMUIOVector *qiov)
{
    trace_bdrv_co_writev(bs, sector_num, nb_sectors);

2403 2404 2405 2406 2407 2408 2409 2410 2411 2412
    return bdrv_co_do_writev(bs, sector_num, nb_sectors, qiov, 0);
}

int coroutine_fn bdrv_co_write_zeroes(BlockDriverState *bs,
                                      int64_t sector_num, int nb_sectors)
{
    trace_bdrv_co_write_zeroes(bs, sector_num, nb_sectors);

    return bdrv_co_do_writev(bs, sector_num, nb_sectors, NULL,
                             BDRV_REQ_ZERO_WRITE);
2413 2414
}

B
bellard 已提交
2415 2416 2417 2418 2419 2420
/**
 * Truncate file to 'offset' bytes (needed only for file protocols)
 */
int bdrv_truncate(BlockDriverState *bs, int64_t offset)
{
    BlockDriver *drv = bs->drv;
2421
    int ret;
B
bellard 已提交
2422
    if (!drv)
B
bellard 已提交
2423
        return -ENOMEDIUM;
B
bellard 已提交
2424 2425
    if (!drv->bdrv_truncate)
        return -ENOTSUP;
2426 2427
    if (bs->read_only)
        return -EACCES;
M
Marcelo Tosatti 已提交
2428 2429
    if (bdrv_in_use(bs))
        return -EBUSY;
2430 2431 2432
    ret = drv->bdrv_truncate(bs, offset);
    if (ret == 0) {
        ret = refresh_total_sectors(bs, offset >> BDRV_SECTOR_BITS);
2433
        bdrv_dev_resize_cb(bs);
2434 2435
    }
    return ret;
B
bellard 已提交
2436 2437
}

2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456
/**
 * 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)
{
    BlockDriver *drv = bs->drv;
    if (!drv) {
        return -ENOMEDIUM;
    }
    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);
    }
    return -ENOTSUP;
}

B
bellard 已提交
2457 2458 2459 2460 2461 2462 2463
/**
 * Length of a file in bytes. Return < 0 if error or unknown.
 */
int64_t bdrv_getlength(BlockDriverState *bs)
{
    BlockDriver *drv = bs->drv;
    if (!drv)
B
bellard 已提交
2464
        return -ENOMEDIUM;
2465

2466
    if (bs->growable || bdrv_dev_has_removable_media(bs)) {
2467 2468 2469
        if (drv->bdrv_getlength) {
            return drv->bdrv_getlength(bs);
        }
B
bellard 已提交
2470
    }
2471
    return bs->total_sectors * BDRV_SECTOR_SIZE;
B
bellard 已提交
2472 2473
}

B
bellard 已提交
2474
/* return 0 as number of sectors if no device present or error */
2475
void bdrv_get_geometry(BlockDriverState *bs, uint64_t *nb_sectors_ptr)
B
bellard 已提交
2476
{
B
bellard 已提交
2477 2478 2479 2480 2481
    int64_t length;
    length = bdrv_getlength(bs);
    if (length < 0)
        length = 0;
    else
J
Jan Kiszka 已提交
2482
        length = length >> BDRV_SECTOR_BITS;
B
bellard 已提交
2483
    *nb_sectors_ptr = length;
B
bellard 已提交
2484
}
B
bellard 已提交
2485

2486 2487 2488 2489 2490 2491 2492 2493
/* throttling disk io limits */
void bdrv_set_io_limits(BlockDriverState *bs,
                        BlockIOLimit *io_limits)
{
    bs->io_limits = *io_limits;
    bs->io_limits_enabled = bdrv_io_limits_enabled(bs);
}

2494 2495
void bdrv_set_on_error(BlockDriverState *bs, BlockdevOnError on_read_error,
                       BlockdevOnError on_write_error)
2496 2497 2498 2499 2500
{
    bs->on_read_error = on_read_error;
    bs->on_write_error = on_write_error;
}

2501
BlockdevOnError bdrv_get_on_error(BlockDriverState *bs, bool is_read)
2502 2503 2504 2505
{
    return is_read ? bs->on_read_error : bs->on_write_error;
}

2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 2527 2528 2529 2530 2531
BlockErrorAction bdrv_get_error_action(BlockDriverState *bs, bool is_read, int error)
{
    BlockdevOnError on_err = is_read ? bs->on_read_error : bs->on_write_error;

    switch (on_err) {
    case BLOCKDEV_ON_ERROR_ENOSPC:
        return (error == ENOSPC) ? BDRV_ACTION_STOP : BDRV_ACTION_REPORT;
    case BLOCKDEV_ON_ERROR_STOP:
        return BDRV_ACTION_STOP;
    case BLOCKDEV_ON_ERROR_REPORT:
        return BDRV_ACTION_REPORT;
    case BLOCKDEV_ON_ERROR_IGNORE:
        return BDRV_ACTION_IGNORE;
    default:
        abort();
    }
}

/* 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)
{
    assert(error >= 0);
P
Paolo Bonzini 已提交
2532
    bdrv_emit_qmp_error_event(bs, QEVENT_BLOCK_IO_ERROR, action, is_read);
2533 2534 2535 2536 2537 2538
    if (action == BDRV_ACTION_STOP) {
        vm_stop(RUN_STATE_IO_ERROR);
        bdrv_iostatus_set_err(bs, error);
    }
}

B
bellard 已提交
2539 2540 2541 2542 2543
int bdrv_is_read_only(BlockDriverState *bs)
{
    return bs->read_only;
}

2544 2545 2546 2547 2548
int bdrv_is_sg(BlockDriverState *bs)
{
    return bs->sg;
}

2549 2550 2551 2552 2553
int bdrv_enable_write_cache(BlockDriverState *bs)
{
    return bs->enable_write_cache;
}

2554 2555 2556
void bdrv_set_enable_write_cache(BlockDriverState *bs, bool wce)
{
    bs->enable_write_cache = wce;
2557 2558 2559 2560 2561 2562 2563

    /* so a reopen() will preserve wce */
    if (wce) {
        bs->open_flags |= BDRV_O_CACHE_WB;
    } else {
        bs->open_flags &= ~BDRV_O_CACHE_WB;
    }
2564 2565
}

B
bellard 已提交
2566 2567 2568 2569 2570 2571 2572
int bdrv_is_encrypted(BlockDriverState *bs)
{
    if (bs->backing_hd && bs->backing_hd->encrypted)
        return 1;
    return bs->encrypted;
}

2573 2574 2575 2576 2577 2578 2579 2580 2581
int bdrv_key_required(BlockDriverState *bs)
{
    BlockDriverState *backing_hd = bs->backing_hd;

    if (backing_hd && backing_hd->encrypted && !backing_hd->valid_key)
        return 1;
    return (bs->encrypted && !bs->valid_key);
}

B
bellard 已提交
2582 2583 2584 2585 2586 2587 2588 2589 2590 2591
int bdrv_set_key(BlockDriverState *bs, const char *key)
{
    int ret;
    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;
    }
2592 2593 2594 2595 2596
    if (!bs->encrypted) {
        return -EINVAL;
    } else if (!bs->drv || !bs->drv->bdrv_set_key) {
        return -ENOMEDIUM;
    }
2597
    ret = bs->drv->bdrv_set_key(bs, key);
2598 2599 2600 2601 2602
    if (ret < 0) {
        bs->valid_key = 0;
    } else if (!bs->valid_key) {
        bs->valid_key = 1;
        /* call the change callback now, we skipped it on open */
2603
        bdrv_dev_change_media_cb(bs, true);
2604
    }
2605
    return ret;
B
bellard 已提交
2606 2607
}

2608
const char *bdrv_get_format_name(BlockDriverState *bs)
B
bellard 已提交
2609
{
2610
    return bs->drv ? bs->drv->format_name : NULL;
B
bellard 已提交
2611 2612
}

2613
void bdrv_iterate_format(void (*it)(void *opaque, const char *name),
B
bellard 已提交
2614 2615 2616 2617
                         void *opaque)
{
    BlockDriver *drv;

2618
    QLIST_FOREACH(drv, &bdrv_drivers, list) {
B
bellard 已提交
2619 2620 2621 2622
        it(opaque, drv->format_name);
    }
}

B
bellard 已提交
2623 2624 2625 2626
BlockDriverState *bdrv_find(const char *name)
{
    BlockDriverState *bs;

2627 2628
    QTAILQ_FOREACH(bs, &bdrv_states, list) {
        if (!strcmp(name, bs->device_name)) {
B
bellard 已提交
2629
            return bs;
2630
        }
B
bellard 已提交
2631 2632 2633 2634
    }
    return NULL;
}

M
Markus Armbruster 已提交
2635 2636 2637 2638 2639 2640 2641 2642
BlockDriverState *bdrv_next(BlockDriverState *bs)
{
    if (!bs) {
        return QTAILQ_FIRST(&bdrv_states);
    }
    return QTAILQ_NEXT(bs, list);
}

2643
void bdrv_iterate(void (*it)(void *opaque, BlockDriverState *bs), void *opaque)
B
bellard 已提交
2644 2645 2646
{
    BlockDriverState *bs;

2647
    QTAILQ_FOREACH(bs, &bdrv_states, list) {
2648
        it(opaque, bs);
B
bellard 已提交
2649 2650 2651
    }
}

B
bellard 已提交
2652 2653 2654 2655 2656
const char *bdrv_get_device_name(BlockDriverState *bs)
{
    return bs->device_name;
}

M
Markus Armbruster 已提交
2657 2658 2659 2660 2661
int bdrv_get_flags(BlockDriverState *bs)
{
    return bs->open_flags;
}

A
aliguori 已提交
2662 2663 2664 2665
void bdrv_flush_all(void)
{
    BlockDriverState *bs;

2666
    QTAILQ_FOREACH(bs, &bdrv_states, list) {
2667
        bdrv_flush(bs);
2668
    }
A
aliguori 已提交
2669 2670
}

K
Kevin Wolf 已提交
2671 2672 2673 2674
int bdrv_has_zero_init(BlockDriverState *bs)
{
    assert(bs->drv);

K
Kevin Wolf 已提交
2675 2676
    if (bs->drv->bdrv_has_zero_init) {
        return bs->drv->bdrv_has_zero_init(bs);
K
Kevin Wolf 已提交
2677 2678 2679 2680 2681
    }

    return 1;
}

2682 2683 2684 2685 2686 2687 2688 2689 2690
typedef struct BdrvCoIsAllocatedData {
    BlockDriverState *bs;
    int64_t sector_num;
    int nb_sectors;
    int *pnum;
    int ret;
    bool done;
} BdrvCoIsAllocatedData;

2691 2692 2693 2694 2695
/*
 * Returns true iff the specified sector is present in the disk image. Drivers
 * not implementing the functionality are assumed to not support backing files,
 * hence all their sectors are reported as allocated.
 *
2696 2697 2698
 * If 'sector_num' is beyond the end of the disk image the return value is 0
 * and 'pnum' is set to 0.
 *
2699 2700 2701 2702
 * 'pnum' is set to the number of sectors (including and immediately following
 * the specified sector) that are known to be in the same
 * allocated/unallocated state.
 *
2703 2704
 * 'nb_sectors' is the max value 'pnum' should be set to.  If nb_sectors goes
 * beyond the end of the disk image it will be clamped.
2705
 */
2706 2707
int coroutine_fn bdrv_co_is_allocated(BlockDriverState *bs, int64_t sector_num,
                                      int nb_sectors, int *pnum)
2708
{
2709 2710 2711 2712 2713 2714 2715 2716 2717 2718 2719 2720
    int64_t n;

    if (sector_num >= bs->total_sectors) {
        *pnum = 0;
        return 0;
    }

    n = bs->total_sectors - sector_num;
    if (n < nb_sectors) {
        nb_sectors = n;
    }

2721
    if (!bs->drv->bdrv_co_is_allocated) {
2722
        *pnum = nb_sectors;
2723 2724
        return 1;
    }
2725

2726 2727 2728 2729 2730 2731 2732 2733 2734 2735 2736 2737 2738 2739 2740 2741 2742 2743 2744 2745 2746 2747
    return bs->drv->bdrv_co_is_allocated(bs, sector_num, nb_sectors, pnum);
}

/* Coroutine wrapper for bdrv_is_allocated() */
static void coroutine_fn bdrv_is_allocated_co_entry(void *opaque)
{
    BdrvCoIsAllocatedData *data = opaque;
    BlockDriverState *bs = data->bs;

    data->ret = bdrv_co_is_allocated(bs, data->sector_num, data->nb_sectors,
                                     data->pnum);
    data->done = true;
}

/*
 * Synchronous wrapper around bdrv_co_is_allocated().
 *
 * See bdrv_co_is_allocated() for details.
 */
int bdrv_is_allocated(BlockDriverState *bs, int64_t sector_num, int nb_sectors,
                      int *pnum)
{
2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762
    Coroutine *co;
    BdrvCoIsAllocatedData data = {
        .bs = bs,
        .sector_num = sector_num,
        .nb_sectors = nb_sectors,
        .pnum = pnum,
        .done = false,
    };

    co = qemu_coroutine_create(bdrv_is_allocated_co_entry);
    qemu_coroutine_enter(co, &data);
    while (!data.done) {
        qemu_aio_wait();
    }
    return data.ret;
2763 2764
}

2765 2766 2767 2768 2769 2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799 2800 2801 2802 2803 2804 2805 2806 2807 2808 2809 2810 2811 2812 2813
/*
 * Given an image chain: ... -> [BASE] -> [INTER1] -> [INTER2] -> [TOP]
 *
 * Return true if the given sector is allocated in any image between
 * BASE and TOP (inclusive).  BASE can be NULL to check if the given
 * sector is allocated in any image of the chain.  Return false otherwise.
 *
 * 'pnum' is set to the number of sectors (including and immediately following
 *  the specified sector) that are known to be in the same
 *  allocated/unallocated state.
 *
 */
int coroutine_fn bdrv_co_is_allocated_above(BlockDriverState *top,
                                            BlockDriverState *base,
                                            int64_t sector_num,
                                            int nb_sectors, int *pnum)
{
    BlockDriverState *intermediate;
    int ret, n = nb_sectors;

    intermediate = top;
    while (intermediate && intermediate != base) {
        int pnum_inter;
        ret = bdrv_co_is_allocated(intermediate, sector_num, nb_sectors,
                                   &pnum_inter);
        if (ret < 0) {
            return ret;
        } else if (ret) {
            *pnum = pnum_inter;
            return 1;
        }

        /*
         * [sector_num, nb_sectors] is unallocated on top but intermediate
         * might have
         *
         * [sector_num+x, nr_sectors] allocated.
         */
        if (n > pnum_inter) {
            n = pnum_inter;
        }

        intermediate = intermediate->backing_hd;
    }

    *pnum = n;
    return 0;
}

P
Paolo Bonzini 已提交
2814
BlockInfo *bdrv_query_info(BlockDriverState *bs)
B
bellard 已提交
2815
{
P
Paolo Bonzini 已提交
2816 2817 2818 2819 2820
    BlockInfo *info = g_malloc0(sizeof(*info));
    info->device = g_strdup(bs->device_name);
    info->type = g_strdup("unknown");
    info->locked = bdrv_dev_is_medium_locked(bs);
    info->removable = bdrv_dev_has_removable_media(bs);
B
bellard 已提交
2821

P
Paolo Bonzini 已提交
2822 2823 2824 2825
    if (bdrv_dev_has_removable_media(bs)) {
        info->has_tray_open = true;
        info->tray_open = bdrv_dev_is_tray_open(bs);
    }
2826

P
Paolo Bonzini 已提交
2827 2828 2829 2830
    if (bdrv_iostatus_is_enabled(bs)) {
        info->has_io_status = true;
        info->io_status = bs->iostatus;
    }
2831

2832 2833 2834
    if (bs->dirty_bitmap) {
        info->has_dirty = true;
        info->dirty = g_malloc0(sizeof(*info->dirty));
2835
        info->dirty->count = bdrv_get_dirty_count(bs) * BDRV_SECTOR_SIZE;
2836 2837
        info->dirty->granularity =
            ((int64_t) BDRV_SECTOR_SIZE << hbitmap_granularity(bs->dirty_bitmap));
2838
    }
2839

P
Paolo Bonzini 已提交
2840 2841 2842 2843 2844 2845 2846 2847 2848 2849 2850 2851
    if (bs->drv) {
        info->has_inserted = true;
        info->inserted = g_malloc0(sizeof(*info->inserted));
        info->inserted->file = g_strdup(bs->filename);
        info->inserted->ro = bs->read_only;
        info->inserted->drv = g_strdup(bs->drv->format_name);
        info->inserted->encrypted = bs->encrypted;
        info->inserted->encryption_key_missing = bdrv_key_required(bs);

        if (bs->backing_file[0]) {
            info->inserted->has_backing_file = true;
            info->inserted->backing_file = g_strdup(bs->backing_file);
2852
        }
2853

P
Paolo Bonzini 已提交
2854 2855 2856 2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 2867 2868
        info->inserted->backing_file_depth = bdrv_get_backing_file_depth(bs);

        if (bs->io_limits_enabled) {
            info->inserted->bps =
                           bs->io_limits.bps[BLOCK_IO_LIMIT_TOTAL];
            info->inserted->bps_rd =
                           bs->io_limits.bps[BLOCK_IO_LIMIT_READ];
            info->inserted->bps_wr =
                           bs->io_limits.bps[BLOCK_IO_LIMIT_WRITE];
            info->inserted->iops =
                           bs->io_limits.iops[BLOCK_IO_LIMIT_TOTAL];
            info->inserted->iops_rd =
                           bs->io_limits.iops[BLOCK_IO_LIMIT_READ];
            info->inserted->iops_wr =
                           bs->io_limits.iops[BLOCK_IO_LIMIT_WRITE];
2869
        }
P
Paolo Bonzini 已提交
2870 2871 2872
    }
    return info;
}
2873

P
Paolo Bonzini 已提交
2874 2875 2876 2877
BlockInfoList *qmp_query_block(Error **errp)
{
    BlockInfoList *head = NULL, **p_next = &head;
    BlockDriverState *bs;
Z
Zhi Yong Wu 已提交
2878

P
Paolo Bonzini 已提交
2879 2880 2881
    QTAILQ_FOREACH(bs, &bdrv_states, list) {
        BlockInfoList *info = g_malloc0(sizeof(*info));
        info->value = bdrv_query_info(bs);
2882

P
Paolo Bonzini 已提交
2883 2884
        *p_next = info;
        p_next = &info->next;
B
bellard 已提交
2885
    }
2886

L
Luiz Capitulino 已提交
2887
    return head;
B
bellard 已提交
2888
}
2889

P
Paolo Bonzini 已提交
2890
BlockStats *bdrv_query_stats(const BlockDriverState *bs)
L
Luiz Capitulino 已提交
2891 2892 2893 2894 2895 2896 2897 2898
{
    BlockStats *s;

    s = g_malloc0(sizeof(*s));

    if (bs->device_name[0]) {
        s->has_device = true;
        s->device = g_strdup(bs->device_name);
2899 2900
    }

L
Luiz Capitulino 已提交
2901 2902 2903 2904 2905 2906 2907 2908 2909 2910 2911
    s->stats = g_malloc0(sizeof(*s->stats));
    s->stats->rd_bytes = bs->nr_bytes[BDRV_ACCT_READ];
    s->stats->wr_bytes = bs->nr_bytes[BDRV_ACCT_WRITE];
    s->stats->rd_operations = bs->nr_ops[BDRV_ACCT_READ];
    s->stats->wr_operations = bs->nr_ops[BDRV_ACCT_WRITE];
    s->stats->wr_highest_offset = bs->wr_highest_sector * BDRV_SECTOR_SIZE;
    s->stats->flush_operations = bs->nr_ops[BDRV_ACCT_FLUSH];
    s->stats->wr_total_time_ns = bs->total_time_ns[BDRV_ACCT_WRITE];
    s->stats->rd_total_time_ns = bs->total_time_ns[BDRV_ACCT_READ];
    s->stats->flush_total_time_ns = bs->total_time_ns[BDRV_ACCT_FLUSH];

2912
    if (bs->file) {
L
Luiz Capitulino 已提交
2913
        s->has_parent = true;
P
Paolo Bonzini 已提交
2914
        s->parent = bdrv_query_stats(bs->file);
2915 2916
    }

L
Luiz Capitulino 已提交
2917
    return s;
2918 2919
}

L
Luiz Capitulino 已提交
2920
BlockStatsList *qmp_query_blockstats(Error **errp)
2921
{
P
Paolo Bonzini 已提交
2922
    BlockStatsList *head = NULL, **p_next = &head;
2923 2924
    BlockDriverState *bs;

2925
    QTAILQ_FOREACH(bs, &bdrv_states, list) {
L
Luiz Capitulino 已提交
2926
        BlockStatsList *info = g_malloc0(sizeof(*info));
P
Paolo Bonzini 已提交
2927
        info->value = bdrv_query_stats(bs);
L
Luiz Capitulino 已提交
2928

P
Paolo Bonzini 已提交
2929 2930
        *p_next = info;
        p_next = &info->next;
2931
    }
2932

L
Luiz Capitulino 已提交
2933
    return head;
2934
}
B
bellard 已提交
2935

2936 2937 2938 2939 2940 2941 2942 2943 2944 2945
const char *bdrv_get_encrypted_filename(BlockDriverState *bs)
{
    if (bs->backing_hd && bs->backing_hd->encrypted)
        return bs->backing_file;
    else if (bs->encrypted)
        return bs->filename;
    else
        return NULL;
}

2946
void bdrv_get_backing_filename(BlockDriverState *bs,
B
bellard 已提交
2947 2948
                               char *filename, int filename_size)
{
K
Kevin Wolf 已提交
2949
    pstrcpy(filename, filename_size, bs->backing_file);
B
bellard 已提交
2950 2951
}

2952
int bdrv_write_compressed(BlockDriverState *bs, int64_t sector_num,
B
bellard 已提交
2953 2954 2955 2956
                          const uint8_t *buf, int nb_sectors)
{
    BlockDriver *drv = bs->drv;
    if (!drv)
B
bellard 已提交
2957
        return -ENOMEDIUM;
B
bellard 已提交
2958 2959
    if (!drv->bdrv_write_compressed)
        return -ENOTSUP;
K
Kevin Wolf 已提交
2960 2961
    if (bdrv_check_request(bs, sector_num, nb_sectors))
        return -EIO;
2962

2963
    assert(!bs->dirty_bitmap);
2964

B
bellard 已提交
2965 2966
    return drv->bdrv_write_compressed(bs, sector_num, buf, nb_sectors);
}
2967

B
bellard 已提交
2968 2969 2970 2971
int bdrv_get_info(BlockDriverState *bs, BlockDriverInfo *bdi)
{
    BlockDriver *drv = bs->drv;
    if (!drv)
B
bellard 已提交
2972
        return -ENOMEDIUM;
B
bellard 已提交
2973 2974 2975 2976 2977 2978
    if (!drv->bdrv_get_info)
        return -ENOTSUP;
    memset(bdi, 0, sizeof(*bdi));
    return drv->bdrv_get_info(bs, bdi);
}

2979 2980
int bdrv_save_vmstate(BlockDriverState *bs, const uint8_t *buf,
                      int64_t pos, int size)
2981 2982 2983 2984
{
    BlockDriver *drv = bs->drv;
    if (!drv)
        return -ENOMEDIUM;
2985 2986 2987 2988 2989
    if (drv->bdrv_save_vmstate)
        return drv->bdrv_save_vmstate(bs, buf, pos, size);
    if (bs->file)
        return bdrv_save_vmstate(bs->file, buf, pos, size);
    return -ENOTSUP;
2990 2991
}

2992 2993
int bdrv_load_vmstate(BlockDriverState *bs, uint8_t *buf,
                      int64_t pos, int size)
2994 2995 2996 2997
{
    BlockDriver *drv = bs->drv;
    if (!drv)
        return -ENOMEDIUM;
2998 2999 3000 3001 3002
    if (drv->bdrv_load_vmstate)
        return drv->bdrv_load_vmstate(bs, buf, pos, size);
    if (bs->file)
        return bdrv_load_vmstate(bs->file, buf, pos, size);
    return -ENOTSUP;
3003 3004
}

K
Kevin Wolf 已提交
3005 3006 3007 3008 3009 3010 3011 3012
void bdrv_debug_event(BlockDriverState *bs, BlkDebugEvent event)
{
    BlockDriver *drv = bs->drv;

    if (!drv || !drv->bdrv_debug_event) {
        return;
    }

B
Blue Swirl 已提交
3013
    drv->bdrv_debug_event(bs, event);
K
Kevin Wolf 已提交
3014 3015 3016 3017 3018 3019 3020 3021 3022 3023 3024 3025 3026 3027 3028 3029 3030 3031 3032 3033 3034
}

int bdrv_debug_breakpoint(BlockDriverState *bs, const char *event,
                          const char *tag)
{
    while (bs && bs->drv && !bs->drv->bdrv_debug_breakpoint) {
        bs = bs->file;
    }

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

    return -ENOTSUP;
}

int bdrv_debug_resume(BlockDriverState *bs, const char *tag)
{
    while (bs && bs->drv && !bs->drv->bdrv_debug_resume) {
        bs = bs->file;
    }
K
Kevin Wolf 已提交
3035

K
Kevin Wolf 已提交
3036 3037 3038 3039 3040 3041 3042 3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053
    if (bs && bs->drv && bs->drv->bdrv_debug_resume) {
        return bs->drv->bdrv_debug_resume(bs, tag);
    }

    return -ENOTSUP;
}

bool bdrv_debug_is_suspended(BlockDriverState *bs, const char *tag)
{
    while (bs && bs->drv && !bs->drv->bdrv_debug_is_suspended) {
        bs = bs->file;
    }

    if (bs && bs->drv && bs->drv->bdrv_debug_is_suspended) {
        return bs->drv->bdrv_debug_is_suspended(bs, tag);
    }

    return false;
K
Kevin Wolf 已提交
3054 3055
}

B
bellard 已提交
3056 3057 3058
/**************************************************************/
/* handling of snapshots */

3059 3060 3061
int bdrv_can_snapshot(BlockDriverState *bs)
{
    BlockDriver *drv = bs->drv;
3062
    if (!drv || !bdrv_is_inserted(bs) || bdrv_is_read_only(bs)) {
3063 3064 3065 3066 3067 3068 3069 3070 3071 3072 3073 3074 3075
        return 0;
    }

    if (!drv->bdrv_snapshot_create) {
        if (bs->file != NULL) {
            return bdrv_can_snapshot(bs->file);
        }
        return 0;
    }

    return 1;
}

3076 3077 3078 3079 3080
int bdrv_is_snapshot(BlockDriverState *bs)
{
    return !!(bs->open_flags & BDRV_O_SNAPSHOT);
}

3081 3082 3083 3084
BlockDriverState *bdrv_snapshots(void)
{
    BlockDriverState *bs;

3085
    if (bs_snapshots) {
3086
        return bs_snapshots;
3087
    }
3088 3089 3090 3091

    bs = NULL;
    while ((bs = bdrv_next(bs))) {
        if (bdrv_can_snapshot(bs)) {
3092 3093
            bs_snapshots = bs;
            return bs;
3094 3095 3096 3097 3098
        }
    }
    return NULL;
}

3099
int bdrv_snapshot_create(BlockDriverState *bs,
B
bellard 已提交
3100 3101 3102 3103
                         QEMUSnapshotInfo *sn_info)
{
    BlockDriver *drv = bs->drv;
    if (!drv)
B
bellard 已提交
3104
        return -ENOMEDIUM;
3105 3106 3107 3108 3109
    if (drv->bdrv_snapshot_create)
        return drv->bdrv_snapshot_create(bs, sn_info);
    if (bs->file)
        return bdrv_snapshot_create(bs->file, sn_info);
    return -ENOTSUP;
B
bellard 已提交
3110 3111
}

3112
int bdrv_snapshot_goto(BlockDriverState *bs,
B
bellard 已提交
3113 3114 3115
                       const char *snapshot_id)
{
    BlockDriver *drv = bs->drv;
3116 3117
    int ret, open_ret;

B
bellard 已提交
3118
    if (!drv)
B
bellard 已提交
3119
        return -ENOMEDIUM;
3120 3121 3122 3123 3124 3125 3126 3127 3128 3129 3130 3131 3132 3133 3134 3135
    if (drv->bdrv_snapshot_goto)
        return drv->bdrv_snapshot_goto(bs, snapshot_id);

    if (bs->file) {
        drv->bdrv_close(bs);
        ret = bdrv_snapshot_goto(bs->file, snapshot_id);
        open_ret = drv->bdrv_open(bs, bs->open_flags);
        if (open_ret < 0) {
            bdrv_delete(bs->file);
            bs->drv = NULL;
            return open_ret;
        }
        return ret;
    }

    return -ENOTSUP;
B
bellard 已提交
3136 3137 3138 3139 3140 3141
}

int bdrv_snapshot_delete(BlockDriverState *bs, const char *snapshot_id)
{
    BlockDriver *drv = bs->drv;
    if (!drv)
B
bellard 已提交
3142
        return -ENOMEDIUM;
3143 3144 3145 3146 3147
    if (drv->bdrv_snapshot_delete)
        return drv->bdrv_snapshot_delete(bs, snapshot_id);
    if (bs->file)
        return bdrv_snapshot_delete(bs->file, snapshot_id);
    return -ENOTSUP;
B
bellard 已提交
3148 3149
}

3150
int bdrv_snapshot_list(BlockDriverState *bs,
B
bellard 已提交
3151 3152 3153 3154
                       QEMUSnapshotInfo **psn_info)
{
    BlockDriver *drv = bs->drv;
    if (!drv)
B
bellard 已提交
3155
        return -ENOMEDIUM;
3156 3157 3158 3159 3160
    if (drv->bdrv_snapshot_list)
        return drv->bdrv_snapshot_list(bs, psn_info);
    if (bs->file)
        return bdrv_snapshot_list(bs->file, psn_info);
    return -ENOTSUP;
B
bellard 已提交
3161 3162
}

E
edison 已提交
3163 3164 3165 3166 3167 3168 3169 3170 3171 3172 3173 3174 3175 3176 3177 3178
int bdrv_snapshot_load_tmp(BlockDriverState *bs,
        const char *snapshot_name)
{
    BlockDriver *drv = bs->drv;
    if (!drv) {
        return -ENOMEDIUM;
    }
    if (!bs->read_only) {
        return -EINVAL;
    }
    if (drv->bdrv_snapshot_load_tmp) {
        return drv->bdrv_snapshot_load_tmp(bs, snapshot_name);
    }
    return -ENOTSUP;
}

3179 3180 3181 3182
/* 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. */
3183 3184 3185
BlockDriverState *bdrv_find_backing_image(BlockDriverState *bs,
        const char *backing_file)
{
3186 3187 3188 3189 3190 3191 3192 3193
    char *filename_full = NULL;
    char *backing_file_full = NULL;
    char *filename_tmp = NULL;
    int is_protocol = 0;
    BlockDriverState *curr_bs = NULL;
    BlockDriverState *retval = NULL;

    if (!bs || !bs->drv || !backing_file) {
3194 3195 3196
        return NULL;
    }

3197 3198 3199 3200 3201 3202 3203 3204 3205 3206 3207 3208 3209 3210 3211
    filename_full     = g_malloc(PATH_MAX);
    backing_file_full = g_malloc(PATH_MAX);
    filename_tmp      = g_malloc(PATH_MAX);

    is_protocol = path_has_protocol(backing_file);

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

        /* 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;
            }
3212
        } else {
3213 3214 3215 3216 3217 3218 3219 3220 3221 3222 3223 3224 3225 3226 3227 3228 3229 3230 3231 3232 3233 3234 3235
            /* 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);

            /* We are going to compare absolute pathnames */
            if (!realpath(filename_tmp, filename_full)) {
                continue;
            }

            /* 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);

            if (!realpath(filename_tmp, backing_file_full)) {
                continue;
            }

            if (strcmp(backing_file_full, filename_full) == 0) {
                retval = curr_bs->backing_hd;
                break;
            }
3236 3237 3238
        }
    }

3239 3240 3241 3242
    g_free(filename_full);
    g_free(backing_file_full);
    g_free(filename_tmp);
    return retval;
3243 3244
}

3245 3246 3247 3248 3249 3250 3251 3252 3253 3254 3255 3256 3257
int bdrv_get_backing_file_depth(BlockDriverState *bs)
{
    if (!bs->drv) {
        return 0;
    }

    if (!bs->backing_hd) {
        return 0;
    }

    return 1 + bdrv_get_backing_file_depth(bs->backing_hd);
}

3258 3259 3260 3261 3262 3263 3264 3265 3266 3267 3268 3269 3270 3271 3272 3273
BlockDriverState *bdrv_find_base(BlockDriverState *bs)
{
    BlockDriverState *curr_bs = NULL;

    if (!bs) {
        return NULL;
    }

    curr_bs = bs;

    while (curr_bs->backing_hd) {
        curr_bs = curr_bs->backing_hd;
    }
    return curr_bs;
}

B
bellard 已提交
3274 3275 3276 3277 3278 3279 3280 3281 3282 3283 3284 3285 3286 3287
#define NB_SUFFIXES 4

char *get_human_readable_size(char *buf, int buf_size, int64_t size)
{
    static const char suffixes[NB_SUFFIXES] = "KMGT";
    int64_t base;
    int i;

    if (size <= 999) {
        snprintf(buf, buf_size, "%" PRId64, size);
    } else {
        base = 1024;
        for(i = 0; i < NB_SUFFIXES; i++) {
            if (size < (10 * base)) {
3288
                snprintf(buf, buf_size, "%0.1f%c",
B
bellard 已提交
3289 3290 3291 3292
                         (double)size / base,
                         suffixes[i]);
                break;
            } else if (size < (1000 * base) || i == (NB_SUFFIXES - 1)) {
3293
                snprintf(buf, buf_size, "%" PRId64 "%c",
B
bellard 已提交
3294 3295 3296 3297 3298 3299 3300 3301 3302 3303 3304 3305 3306 3307 3308 3309 3310 3311
                         ((size + (base >> 1)) / base),
                         suffixes[i]);
                break;
            }
            base = base * 1024;
        }
    }
    return buf;
}

char *bdrv_snapshot_dump(char *buf, int buf_size, QEMUSnapshotInfo *sn)
{
    char buf1[128], date_buf[128], clock_buf[128];
    struct tm tm;
    time_t ti;
    int64_t secs;

    if (!sn) {
3312 3313
        snprintf(buf, buf_size,
                 "%-10s%-20s%7s%20s%15s",
B
bellard 已提交
3314 3315 3316 3317 3318 3319 3320 3321 3322 3323 3324
                 "ID", "TAG", "VM SIZE", "DATE", "VM CLOCK");
    } else {
        ti = sn->date_sec;
        localtime_r(&ti, &tm);
        strftime(date_buf, sizeof(date_buf),
                 "%Y-%m-%d %H:%M:%S", &tm);
        secs = sn->vm_clock_nsec / 1000000000;
        snprintf(clock_buf, sizeof(clock_buf),
                 "%02d:%02d:%02d.%03d",
                 (int)(secs / 3600),
                 (int)((secs / 60) % 60),
3325
                 (int)(secs % 60),
B
bellard 已提交
3326 3327
                 (int)((sn->vm_clock_nsec / 1000000) % 1000));
        snprintf(buf, buf_size,
3328
                 "%-10s%-20s%7s%20s%15s",
B
bellard 已提交
3329 3330 3331 3332 3333 3334 3335 3336
                 sn->id_str, sn->name,
                 get_human_readable_size(buf1, sizeof(buf1), sn->vm_state_size),
                 date_buf,
                 clock_buf);
    }
    return buf;
}

B
bellard 已提交
3337
/**************************************************************/
B
bellard 已提交
3338
/* async I/Os */
B
bellard 已提交
3339

3340
BlockDriverAIOCB *bdrv_aio_readv(BlockDriverState *bs, int64_t sector_num,
3341
                                 QEMUIOVector *qiov, int nb_sectors,
3342
                                 BlockDriverCompletionFunc *cb, void *opaque)
B
bellard 已提交
3343
{
3344 3345
    trace_bdrv_aio_readv(bs, sector_num, nb_sectors, opaque);

3346
    return bdrv_co_aio_rw_vector(bs, sector_num, qiov, nb_sectors,
3347
                                 cb, opaque, false);
B
bellard 已提交
3348 3349
}

3350 3351 3352
BlockDriverAIOCB *bdrv_aio_writev(BlockDriverState *bs, int64_t sector_num,
                                  QEMUIOVector *qiov, int nb_sectors,
                                  BlockDriverCompletionFunc *cb, void *opaque)
B
bellard 已提交
3353
{
3354 3355
    trace_bdrv_aio_writev(bs, sector_num, nb_sectors, opaque);

3356
    return bdrv_co_aio_rw_vector(bs, sector_num, qiov, nb_sectors,
3357
                                 cb, opaque, true);
B
bellard 已提交
3358 3359
}

K
Kevin Wolf 已提交
3360 3361 3362 3363 3364 3365 3366 3367 3368 3369 3370 3371 3372 3373 3374 3375 3376 3377

typedef struct MultiwriteCB {
    int error;
    int num_requests;
    int num_callbacks;
    struct {
        BlockDriverCompletionFunc *cb;
        void *opaque;
        QEMUIOVector *free_qiov;
    } callbacks[];
} MultiwriteCB;

static void multiwrite_user_cb(MultiwriteCB *mcb)
{
    int i;

    for (i = 0; i < mcb->num_callbacks; i++) {
        mcb->callbacks[i].cb(mcb->callbacks[i].opaque, mcb->error);
3378 3379 3380
        if (mcb->callbacks[i].free_qiov) {
            qemu_iovec_destroy(mcb->callbacks[i].free_qiov);
        }
3381
        g_free(mcb->callbacks[i].free_qiov);
K
Kevin Wolf 已提交
3382 3383 3384 3385 3386 3387 3388
    }
}

static void multiwrite_cb(void *opaque, int ret)
{
    MultiwriteCB *mcb = opaque;

3389 3390
    trace_multiwrite_cb(mcb, ret);

3391
    if (ret < 0 && !mcb->error) {
K
Kevin Wolf 已提交
3392 3393 3394 3395 3396
        mcb->error = ret;
    }

    mcb->num_requests--;
    if (mcb->num_requests == 0) {
3397
        multiwrite_user_cb(mcb);
3398
        g_free(mcb);
K
Kevin Wolf 已提交
3399 3400 3401 3402 3403
    }
}

static int multiwrite_req_compare(const void *a, const void *b)
{
3404 3405 3406 3407 3408 3409 3410 3411 3412 3413 3414 3415 3416
    const BlockRequest *req1 = a, *req2 = b;

    /*
     * Note that we can't simply subtract req2->sector from req1->sector
     * here as that could overflow the return value.
     */
    if (req1->sector > req2->sector) {
        return 1;
    } else if (req1->sector < req2->sector) {
        return -1;
    } else {
        return 0;
    }
K
Kevin Wolf 已提交
3417 3418 3419 3420 3421 3422 3423 3424 3425 3426 3427 3428 3429 3430 3431 3432 3433 3434 3435 3436 3437
}

/*
 * Takes a bunch of requests and tries to merge them. Returns the number of
 * requests that remain after merging.
 */
static int multiwrite_merge(BlockDriverState *bs, BlockRequest *reqs,
    int num_reqs, MultiwriteCB *mcb)
{
    int i, outidx;

    // Sort requests by start sector
    qsort(reqs, num_reqs, sizeof(*reqs), &multiwrite_req_compare);

    // Check if adjacent requests touch the same clusters. If so, combine them,
    // filling up gaps with zero sectors.
    outidx = 0;
    for (i = 1; i < num_reqs; i++) {
        int merge = 0;
        int64_t oldreq_last = reqs[outidx].sector + reqs[outidx].nb_sectors;

3438
        // Handle exactly sequential writes and overlapping writes.
K
Kevin Wolf 已提交
3439 3440 3441 3442
        if (reqs[i].sector <= oldreq_last) {
            merge = 1;
        }

3443 3444 3445 3446
        if (reqs[outidx].qiov->niov + reqs[i].qiov->niov + 1 > IOV_MAX) {
            merge = 0;
        }

K
Kevin Wolf 已提交
3447 3448
        if (merge) {
            size_t size;
3449
            QEMUIOVector *qiov = g_malloc0(sizeof(*qiov));
K
Kevin Wolf 已提交
3450 3451 3452 3453 3454 3455
            qemu_iovec_init(qiov,
                reqs[outidx].qiov->niov + reqs[i].qiov->niov + 1);

            // Add the first request to the merged one. If the requests are
            // overlapping, drop the last sectors of the first request.
            size = (reqs[i].sector - reqs[outidx].sector) << 9;
3456
            qemu_iovec_concat(qiov, reqs[outidx].qiov, 0, size);
K
Kevin Wolf 已提交
3457

3458 3459
            // We should need to add any zeros between the two requests
            assert (reqs[i].sector <= oldreq_last);
K
Kevin Wolf 已提交
3460 3461

            // Add the second request
3462
            qemu_iovec_concat(qiov, reqs[i].qiov, 0, reqs[i].qiov->size);
K
Kevin Wolf 已提交
3463

3464
            reqs[outidx].nb_sectors = qiov->size >> 9;
K
Kevin Wolf 已提交
3465 3466 3467 3468 3469 3470 3471 3472 3473 3474 3475 3476 3477 3478 3479 3480 3481 3482 3483 3484 3485 3486 3487 3488 3489 3490 3491 3492 3493 3494 3495 3496 3497
            reqs[outidx].qiov = qiov;

            mcb->callbacks[i].free_qiov = reqs[outidx].qiov;
        } else {
            outidx++;
            reqs[outidx].sector     = reqs[i].sector;
            reqs[outidx].nb_sectors = reqs[i].nb_sectors;
            reqs[outidx].qiov       = reqs[i].qiov;
        }
    }

    return outidx + 1;
}

/*
 * Submit multiple AIO write requests at once.
 *
 * On success, the function returns 0 and all requests in the reqs array have
 * been submitted. In error case this function returns -1, and any of the
 * requests may or may not be submitted yet. In particular, this means that the
 * callback will be called for some of the requests, for others it won't. The
 * caller must check the error field of the BlockRequest to wait for the right
 * callbacks (if error != 0, no callback will be called).
 *
 * The implementation may modify the contents of the reqs array, e.g. to merge
 * requests. However, the fields opaque and error are left unmodified as they
 * are used to signal failure for a single request to the caller.
 */
int bdrv_aio_multiwrite(BlockDriverState *bs, BlockRequest *reqs, int num_reqs)
{
    MultiwriteCB *mcb;
    int i;

3498 3499 3500 3501 3502 3503 3504 3505
    /* don't submit writes if we don't have a medium */
    if (bs->drv == NULL) {
        for (i = 0; i < num_reqs; i++) {
            reqs[i].error = -ENOMEDIUM;
        }
        return -1;
    }

K
Kevin Wolf 已提交
3506 3507 3508 3509 3510
    if (num_reqs == 0) {
        return 0;
    }

    // Create MultiwriteCB structure
3511
    mcb = g_malloc0(sizeof(*mcb) + num_reqs * sizeof(*mcb->callbacks));
K
Kevin Wolf 已提交
3512 3513 3514 3515 3516 3517 3518 3519 3520 3521 3522
    mcb->num_requests = 0;
    mcb->num_callbacks = num_reqs;

    for (i = 0; i < num_reqs; i++) {
        mcb->callbacks[i].cb = reqs[i].cb;
        mcb->callbacks[i].opaque = reqs[i].opaque;
    }

    // Check for mergable requests
    num_reqs = multiwrite_merge(bs, reqs, num_reqs, mcb);

3523 3524
    trace_bdrv_aio_multiwrite(mcb, mcb->num_callbacks, num_reqs);

3525 3526
    /* Run the aio requests. */
    mcb->num_requests = num_reqs;
K
Kevin Wolf 已提交
3527
    for (i = 0; i < num_reqs; i++) {
3528
        bdrv_aio_writev(bs, reqs[i].sector, reqs[i].qiov,
K
Kevin Wolf 已提交
3529 3530 3531 3532 3533 3534
            reqs[i].nb_sectors, multiwrite_cb, mcb);
    }

    return 0;
}

B
bellard 已提交
3535 3536
void bdrv_aio_cancel(BlockDriverAIOCB *acb)
{
S
Stefan Hajnoczi 已提交
3537
    acb->aiocb_info->cancel(acb);
B
bellard 已提交
3538 3539
}

Z
Zhi Yong Wu 已提交
3540 3541 3542 3543 3544 3545 3546 3547 3548 3549 3550 3551 3552 3553 3554 3555 3556 3557 3558 3559 3560 3561 3562 3563 3564 3565 3566 3567 3568 3569 3570 3571 3572 3573 3574 3575 3576 3577 3578 3579 3580 3581 3582 3583 3584 3585 3586 3587 3588 3589 3590 3591 3592 3593 3594 3595 3596 3597 3598 3599 3600 3601 3602 3603 3604 3605 3606 3607 3608 3609 3610 3611 3612 3613 3614 3615 3616 3617 3618 3619 3620 3621 3622 3623 3624 3625 3626 3627 3628 3629 3630 3631 3632 3633 3634 3635 3636 3637 3638 3639 3640 3641 3642 3643 3644 3645 3646 3647 3648 3649 3650 3651 3652 3653 3654 3655 3656 3657 3658 3659 3660 3661 3662 3663 3664 3665 3666 3667 3668 3669 3670 3671 3672 3673 3674 3675 3676 3677 3678 3679 3680 3681 3682 3683 3684 3685 3686 3687 3688 3689 3690 3691 3692 3693 3694 3695 3696 3697 3698 3699 3700 3701 3702 3703
/* block I/O throttling */
static bool bdrv_exceed_bps_limits(BlockDriverState *bs, int nb_sectors,
                 bool is_write, double elapsed_time, uint64_t *wait)
{
    uint64_t bps_limit = 0;
    double   bytes_limit, bytes_base, bytes_res;
    double   slice_time, wait_time;

    if (bs->io_limits.bps[BLOCK_IO_LIMIT_TOTAL]) {
        bps_limit = bs->io_limits.bps[BLOCK_IO_LIMIT_TOTAL];
    } else if (bs->io_limits.bps[is_write]) {
        bps_limit = bs->io_limits.bps[is_write];
    } else {
        if (wait) {
            *wait = 0;
        }

        return false;
    }

    slice_time = bs->slice_end - bs->slice_start;
    slice_time /= (NANOSECONDS_PER_SECOND);
    bytes_limit = bps_limit * slice_time;
    bytes_base  = bs->nr_bytes[is_write] - bs->io_base.bytes[is_write];
    if (bs->io_limits.bps[BLOCK_IO_LIMIT_TOTAL]) {
        bytes_base += bs->nr_bytes[!is_write] - bs->io_base.bytes[!is_write];
    }

    /* bytes_base: the bytes of data which have been read/written; and
     *             it is obtained from the history statistic info.
     * bytes_res: the remaining bytes of data which need to be read/written.
     * (bytes_base + bytes_res) / bps_limit: used to calcuate
     *             the total time for completing reading/writting all data.
     */
    bytes_res   = (unsigned) nb_sectors * BDRV_SECTOR_SIZE;

    if (bytes_base + bytes_res <= bytes_limit) {
        if (wait) {
            *wait = 0;
        }

        return false;
    }

    /* Calc approx time to dispatch */
    wait_time = (bytes_base + bytes_res) / bps_limit - elapsed_time;

    /* When the I/O rate at runtime exceeds the limits,
     * bs->slice_end need to be extended in order that the current statistic
     * info can be kept until the timer fire, so it is increased and tuned
     * based on the result of experiment.
     */
    bs->slice_time = wait_time * BLOCK_IO_SLICE_TIME * 10;
    bs->slice_end += bs->slice_time - 3 * BLOCK_IO_SLICE_TIME;
    if (wait) {
        *wait = wait_time * BLOCK_IO_SLICE_TIME * 10;
    }

    return true;
}

static bool bdrv_exceed_iops_limits(BlockDriverState *bs, bool is_write,
                             double elapsed_time, uint64_t *wait)
{
    uint64_t iops_limit = 0;
    double   ios_limit, ios_base;
    double   slice_time, wait_time;

    if (bs->io_limits.iops[BLOCK_IO_LIMIT_TOTAL]) {
        iops_limit = bs->io_limits.iops[BLOCK_IO_LIMIT_TOTAL];
    } else if (bs->io_limits.iops[is_write]) {
        iops_limit = bs->io_limits.iops[is_write];
    } else {
        if (wait) {
            *wait = 0;
        }

        return false;
    }

    slice_time = bs->slice_end - bs->slice_start;
    slice_time /= (NANOSECONDS_PER_SECOND);
    ios_limit  = iops_limit * slice_time;
    ios_base   = bs->nr_ops[is_write] - bs->io_base.ios[is_write];
    if (bs->io_limits.iops[BLOCK_IO_LIMIT_TOTAL]) {
        ios_base += bs->nr_ops[!is_write] - bs->io_base.ios[!is_write];
    }

    if (ios_base + 1 <= ios_limit) {
        if (wait) {
            *wait = 0;
        }

        return false;
    }

    /* Calc approx time to dispatch */
    wait_time = (ios_base + 1) / iops_limit;
    if (wait_time > elapsed_time) {
        wait_time = wait_time - elapsed_time;
    } else {
        wait_time = 0;
    }

    bs->slice_time = wait_time * BLOCK_IO_SLICE_TIME * 10;
    bs->slice_end += bs->slice_time - 3 * BLOCK_IO_SLICE_TIME;
    if (wait) {
        *wait = wait_time * BLOCK_IO_SLICE_TIME * 10;
    }

    return true;
}

static bool bdrv_exceed_io_limits(BlockDriverState *bs, int nb_sectors,
                           bool is_write, int64_t *wait)
{
    int64_t  now, max_wait;
    uint64_t bps_wait = 0, iops_wait = 0;
    double   elapsed_time;
    int      bps_ret, iops_ret;

    now = qemu_get_clock_ns(vm_clock);
    if ((bs->slice_start < now)
        && (bs->slice_end > now)) {
        bs->slice_end = now + bs->slice_time;
    } else {
        bs->slice_time  =  5 * BLOCK_IO_SLICE_TIME;
        bs->slice_start = now;
        bs->slice_end   = now + bs->slice_time;

        bs->io_base.bytes[is_write]  = bs->nr_bytes[is_write];
        bs->io_base.bytes[!is_write] = bs->nr_bytes[!is_write];

        bs->io_base.ios[is_write]    = bs->nr_ops[is_write];
        bs->io_base.ios[!is_write]   = bs->nr_ops[!is_write];
    }

    elapsed_time  = now - bs->slice_start;
    elapsed_time  /= (NANOSECONDS_PER_SECOND);

    bps_ret  = bdrv_exceed_bps_limits(bs, nb_sectors,
                                      is_write, elapsed_time, &bps_wait);
    iops_ret = bdrv_exceed_iops_limits(bs, is_write,
                                      elapsed_time, &iops_wait);
    if (bps_ret || iops_ret) {
        max_wait = bps_wait > iops_wait ? bps_wait : iops_wait;
        if (wait) {
            *wait = max_wait;
        }

        now = qemu_get_clock_ns(vm_clock);
        if (bs->slice_end < now + max_wait) {
            bs->slice_end = now + max_wait;
        }

        return true;
    }

    if (wait) {
        *wait = 0;
    }

    return false;
}
3704

B
bellard 已提交
3705 3706 3707
/**************************************************************/
/* async block device emulation */

3708 3709 3710 3711 3712 3713 3714 3715 3716 3717 3718 3719
typedef struct BlockDriverAIOCBSync {
    BlockDriverAIOCB common;
    QEMUBH *bh;
    int ret;
    /* vector translation state */
    QEMUIOVector *qiov;
    uint8_t *bounce;
    int is_write;
} BlockDriverAIOCBSync;

static void bdrv_aio_cancel_em(BlockDriverAIOCB *blockacb)
{
3720 3721
    BlockDriverAIOCBSync *acb =
        container_of(blockacb, BlockDriverAIOCBSync, common);
3722
    qemu_bh_delete(acb->bh);
A
Avi Kivity 已提交
3723
    acb->bh = NULL;
3724 3725 3726
    qemu_aio_release(acb);
}

S
Stefan Hajnoczi 已提交
3727
static const AIOCBInfo bdrv_em_aiocb_info = {
3728 3729 3730 3731
    .aiocb_size         = sizeof(BlockDriverAIOCBSync),
    .cancel             = bdrv_aio_cancel_em,
};

3732
static void bdrv_aio_bh_cb(void *opaque)
B
bellard 已提交
3733
{
3734
    BlockDriverAIOCBSync *acb = opaque;
3735 3736

    if (!acb->is_write)
3737
        qemu_iovec_from_buf(acb->qiov, 0, acb->bounce, acb->qiov->size);
3738
    qemu_vfree(acb->bounce);
3739
    acb->common.cb(acb->common.opaque, acb->ret);
3740
    qemu_bh_delete(acb->bh);
A
Avi Kivity 已提交
3741
    acb->bh = NULL;
3742
    qemu_aio_release(acb);
B
bellard 已提交
3743
}
3744

3745 3746 3747 3748 3749 3750 3751 3752
static BlockDriverAIOCB *bdrv_aio_rw_vector(BlockDriverState *bs,
                                            int64_t sector_num,
                                            QEMUIOVector *qiov,
                                            int nb_sectors,
                                            BlockDriverCompletionFunc *cb,
                                            void *opaque,
                                            int is_write)

B
bellard 已提交
3753
{
3754 3755
    BlockDriverAIOCBSync *acb;

S
Stefan Hajnoczi 已提交
3756
    acb = qemu_aio_get(&bdrv_em_aiocb_info, bs, cb, opaque);
3757 3758
    acb->is_write = is_write;
    acb->qiov = qiov;
3759
    acb->bounce = qemu_blockalign(bs, qiov->size);
3760
    acb->bh = qemu_bh_new(bdrv_aio_bh_cb, acb);
3761 3762

    if (is_write) {
3763
        qemu_iovec_to_buf(acb->qiov, 0, acb->bounce, qiov->size);
3764
        acb->ret = bs->drv->bdrv_write(bs, sector_num, acb->bounce, nb_sectors);
3765
    } else {
3766
        acb->ret = bs->drv->bdrv_read(bs, sector_num, acb->bounce, nb_sectors);
3767 3768
    }

3769
    qemu_bh_schedule(acb->bh);
3770

3771
    return &acb->common;
3772 3773
}

3774 3775
static BlockDriverAIOCB *bdrv_aio_readv_em(BlockDriverState *bs,
        int64_t sector_num, QEMUIOVector *qiov, int nb_sectors,
3776
        BlockDriverCompletionFunc *cb, void *opaque)
3777
{
3778 3779
    return bdrv_aio_rw_vector(bs, sector_num, qiov, nb_sectors, cb, opaque, 0);
}
B
bellard 已提交
3780

3781 3782 3783 3784 3785
static BlockDriverAIOCB *bdrv_aio_writev_em(BlockDriverState *bs,
        int64_t sector_num, QEMUIOVector *qiov, int nb_sectors,
        BlockDriverCompletionFunc *cb, void *opaque)
{
    return bdrv_aio_rw_vector(bs, sector_num, qiov, nb_sectors, cb, opaque, 1);
3786 3787
}

3788 3789 3790 3791 3792

typedef struct BlockDriverAIOCBCoroutine {
    BlockDriverAIOCB common;
    BlockRequest req;
    bool is_write;
K
Kevin Wolf 已提交
3793
    bool *done;
3794 3795 3796 3797 3798
    QEMUBH* bh;
} BlockDriverAIOCBCoroutine;

static void bdrv_aio_co_cancel_em(BlockDriverAIOCB *blockacb)
{
K
Kevin Wolf 已提交
3799 3800 3801 3802 3803 3804 3805 3806
    BlockDriverAIOCBCoroutine *acb =
        container_of(blockacb, BlockDriverAIOCBCoroutine, common);
    bool done = false;

    acb->done = &done;
    while (!done) {
        qemu_aio_wait();
    }
3807 3808
}

S
Stefan Hajnoczi 已提交
3809
static const AIOCBInfo bdrv_em_co_aiocb_info = {
3810 3811 3812 3813
    .aiocb_size         = sizeof(BlockDriverAIOCBCoroutine),
    .cancel             = bdrv_aio_co_cancel_em,
};

P
Paolo Bonzini 已提交
3814
static void bdrv_co_em_bh(void *opaque)
3815 3816 3817 3818
{
    BlockDriverAIOCBCoroutine *acb = opaque;

    acb->common.cb(acb->common.opaque, acb->req.error);
K
Kevin Wolf 已提交
3819 3820 3821 3822 3823

    if (acb->done) {
        *acb->done = true;
    }

3824 3825 3826 3827
    qemu_bh_delete(acb->bh);
    qemu_aio_release(acb);
}

3828 3829 3830 3831 3832 3833 3834 3835
/* Invoke bdrv_co_do_readv/bdrv_co_do_writev */
static void coroutine_fn bdrv_co_do_rw(void *opaque)
{
    BlockDriverAIOCBCoroutine *acb = opaque;
    BlockDriverState *bs = acb->common.bs;

    if (!acb->is_write) {
        acb->req.error = bdrv_co_do_readv(bs, acb->req.sector,
3836
            acb->req.nb_sectors, acb->req.qiov, 0);
3837 3838
    } else {
        acb->req.error = bdrv_co_do_writev(bs, acb->req.sector,
3839
            acb->req.nb_sectors, acb->req.qiov, 0);
3840 3841
    }

P
Paolo Bonzini 已提交
3842
    acb->bh = qemu_bh_new(bdrv_co_em_bh, acb);
3843 3844 3845
    qemu_bh_schedule(acb->bh);
}

3846 3847 3848 3849 3850 3851
static BlockDriverAIOCB *bdrv_co_aio_rw_vector(BlockDriverState *bs,
                                               int64_t sector_num,
                                               QEMUIOVector *qiov,
                                               int nb_sectors,
                                               BlockDriverCompletionFunc *cb,
                                               void *opaque,
3852
                                               bool is_write)
3853 3854 3855 3856
{
    Coroutine *co;
    BlockDriverAIOCBCoroutine *acb;

S
Stefan Hajnoczi 已提交
3857
    acb = qemu_aio_get(&bdrv_em_co_aiocb_info, bs, cb, opaque);
3858 3859 3860 3861
    acb->req.sector = sector_num;
    acb->req.nb_sectors = nb_sectors;
    acb->req.qiov = qiov;
    acb->is_write = is_write;
K
Kevin Wolf 已提交
3862
    acb->done = NULL;
3863

3864
    co = qemu_coroutine_create(bdrv_co_do_rw);
3865 3866 3867 3868 3869
    qemu_coroutine_enter(co, acb);

    return &acb->common;
}

P
Paolo Bonzini 已提交
3870
static void coroutine_fn bdrv_aio_flush_co_entry(void *opaque)
3871
{
P
Paolo Bonzini 已提交
3872 3873
    BlockDriverAIOCBCoroutine *acb = opaque;
    BlockDriverState *bs = acb->common.bs;
3874

P
Paolo Bonzini 已提交
3875 3876
    acb->req.error = bdrv_co_flush(bs);
    acb->bh = qemu_bh_new(bdrv_co_em_bh, acb);
3877 3878 3879
    qemu_bh_schedule(acb->bh);
}

P
Paolo Bonzini 已提交
3880
BlockDriverAIOCB *bdrv_aio_flush(BlockDriverState *bs,
3881 3882
        BlockDriverCompletionFunc *cb, void *opaque)
{
P
Paolo Bonzini 已提交
3883
    trace_bdrv_aio_flush(bs, opaque);
3884

P
Paolo Bonzini 已提交
3885 3886
    Coroutine *co;
    BlockDriverAIOCBCoroutine *acb;
3887

S
Stefan Hajnoczi 已提交
3888
    acb = qemu_aio_get(&bdrv_em_co_aiocb_info, bs, cb, opaque);
K
Kevin Wolf 已提交
3889 3890
    acb->done = NULL;

P
Paolo Bonzini 已提交
3891 3892
    co = qemu_coroutine_create(bdrv_aio_flush_co_entry);
    qemu_coroutine_enter(co, acb);
3893 3894 3895 3896

    return &acb->common;
}

3897 3898 3899 3900 3901 3902 3903 3904 3905 3906 3907 3908 3909 3910 3911 3912 3913 3914 3915
static void coroutine_fn bdrv_aio_discard_co_entry(void *opaque)
{
    BlockDriverAIOCBCoroutine *acb = opaque;
    BlockDriverState *bs = acb->common.bs;

    acb->req.error = bdrv_co_discard(bs, acb->req.sector, acb->req.nb_sectors);
    acb->bh = qemu_bh_new(bdrv_co_em_bh, acb);
    qemu_bh_schedule(acb->bh);
}

BlockDriverAIOCB *bdrv_aio_discard(BlockDriverState *bs,
        int64_t sector_num, int nb_sectors,
        BlockDriverCompletionFunc *cb, void *opaque)
{
    Coroutine *co;
    BlockDriverAIOCBCoroutine *acb;

    trace_bdrv_aio_discard(bs, sector_num, nb_sectors, opaque);

S
Stefan Hajnoczi 已提交
3916
    acb = qemu_aio_get(&bdrv_em_co_aiocb_info, bs, cb, opaque);
3917 3918
    acb->req.sector = sector_num;
    acb->req.nb_sectors = nb_sectors;
K
Kevin Wolf 已提交
3919
    acb->done = NULL;
3920 3921 3922 3923 3924 3925
    co = qemu_coroutine_create(bdrv_aio_discard_co_entry);
    qemu_coroutine_enter(co, acb);

    return &acb->common;
}

B
bellard 已提交
3926 3927
void bdrv_init(void)
{
3928
    module_call_init(MODULE_INIT_BLOCK);
B
bellard 已提交
3929
}
3930

3931 3932 3933 3934 3935 3936
void bdrv_init_with_whitelist(void)
{
    use_bdrv_whitelist = 1;
    bdrv_init();
}

S
Stefan Hajnoczi 已提交
3937
void *qemu_aio_get(const AIOCBInfo *aiocb_info, BlockDriverState *bs,
3938
                   BlockDriverCompletionFunc *cb, void *opaque)
3939 3940 3941
{
    BlockDriverAIOCB *acb;

S
Stefan Hajnoczi 已提交
3942 3943
    acb = g_slice_alloc(aiocb_info->aiocb_size);
    acb->aiocb_info = aiocb_info;
3944 3945 3946 3947 3948 3949 3950 3951
    acb->bs = bs;
    acb->cb = cb;
    acb->opaque = opaque;
    return acb;
}

void qemu_aio_release(void *p)
{
3952
    BlockDriverAIOCB *acb = p;
S
Stefan Hajnoczi 已提交
3953
    g_slice_free1(acb->aiocb_info->aiocb_size, acb);
3954
}
B
bellard 已提交
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
/**************************************************************/
/* Coroutine block device emulation */

typedef struct CoroutineIOCompletion {
    Coroutine *coroutine;
    int ret;
} CoroutineIOCompletion;

static void bdrv_co_io_em_complete(void *opaque, int ret)
{
    CoroutineIOCompletion *co = opaque;

    co->ret = ret;
    qemu_coroutine_enter(co->coroutine, NULL);
}

static int coroutine_fn bdrv_co_io_em(BlockDriverState *bs, int64_t sector_num,
                                      int nb_sectors, QEMUIOVector *iov,
                                      bool is_write)
{
    CoroutineIOCompletion co = {
        .coroutine = qemu_coroutine_self(),
    };
    BlockDriverAIOCB *acb;

    if (is_write) {
3982 3983
        acb = bs->drv->bdrv_aio_writev(bs, sector_num, iov, nb_sectors,
                                       bdrv_co_io_em_complete, &co);
3984
    } else {
3985 3986
        acb = bs->drv->bdrv_aio_readv(bs, sector_num, iov, nb_sectors,
                                      bdrv_co_io_em_complete, &co);
3987 3988
    }

3989
    trace_bdrv_co_io_em(bs, sector_num, nb_sectors, is_write, acb);
3990 3991 3992 3993 3994 3995 3996 3997 3998 3999 4000 4001 4002 4003 4004 4005 4006 4007 4008 4009 4010 4011
    if (!acb) {
        return -EIO;
    }
    qemu_coroutine_yield();

    return co.ret;
}

static int coroutine_fn bdrv_co_readv_em(BlockDriverState *bs,
                                         int64_t sector_num, int nb_sectors,
                                         QEMUIOVector *iov)
{
    return bdrv_co_io_em(bs, sector_num, nb_sectors, iov, false);
}

static int coroutine_fn bdrv_co_writev_em(BlockDriverState *bs,
                                         int64_t sector_num, int nb_sectors,
                                         QEMUIOVector *iov)
{
    return bdrv_co_io_em(bs, sector_num, nb_sectors, iov, true);
}

P
Paolo Bonzini 已提交
4012
static void coroutine_fn bdrv_flush_co_entry(void *opaque)
4013
{
P
Paolo Bonzini 已提交
4014 4015 4016 4017 4018 4019 4020
    RwCo *rwco = opaque;

    rwco->ret = bdrv_co_flush(rwco->bs);
}

int coroutine_fn bdrv_co_flush(BlockDriverState *bs)
{
K
Kevin Wolf 已提交
4021 4022
    int ret;

4023
    if (!bs || !bdrv_is_inserted(bs) || bdrv_is_read_only(bs)) {
P
Paolo Bonzini 已提交
4024
        return 0;
K
Kevin Wolf 已提交
4025 4026
    }

4027
    /* Write back cached data to the OS even with cache=unsafe */
K
Kevin Wolf 已提交
4028 4029 4030 4031 4032 4033 4034
    if (bs->drv->bdrv_co_flush_to_os) {
        ret = bs->drv->bdrv_co_flush_to_os(bs);
        if (ret < 0) {
            return ret;
        }
    }

4035 4036
    /* But don't actually force it to the disk with cache=unsafe */
    if (bs->open_flags & BDRV_O_NO_FLUSH) {
4037
        goto flush_parent;
4038 4039
    }

K
Kevin Wolf 已提交
4040
    if (bs->drv->bdrv_co_flush_to_disk) {
4041
        ret = bs->drv->bdrv_co_flush_to_disk(bs);
P
Paolo Bonzini 已提交
4042 4043 4044 4045 4046 4047 4048 4049
    } else if (bs->drv->bdrv_aio_flush) {
        BlockDriverAIOCB *acb;
        CoroutineIOCompletion co = {
            .coroutine = qemu_coroutine_self(),
        };

        acb = bs->drv->bdrv_aio_flush(bs, bdrv_co_io_em_complete, &co);
        if (acb == NULL) {
4050
            ret = -EIO;
P
Paolo Bonzini 已提交
4051 4052
        } else {
            qemu_coroutine_yield();
4053
            ret = co.ret;
P
Paolo Bonzini 已提交
4054 4055 4056 4057 4058 4059 4060 4061 4062 4063 4064 4065 4066
        }
    } else {
        /*
         * Some block drivers always operate in either writethrough or unsafe
         * mode and don't support bdrv_flush therefore. Usually qemu doesn't
         * know how the server works (because the behaviour is hardcoded or
         * depends on server-side configuration), so we can't ensure that
         * everything is safe on disk. Returning an error doesn't work because
         * that would break guests even if the server operates in writethrough
         * mode.
         *
         * Let's hope the user knows what he's doing.
         */
4067
        ret = 0;
P
Paolo Bonzini 已提交
4068
    }
4069 4070 4071 4072 4073 4074 4075
    if (ret < 0) {
        return ret;
    }

    /* Now flush the underlying protocol.  It will also have BDRV_O_NO_FLUSH
     * in the case of cache=unsafe, so there are no useless flushes.
     */
4076
flush_parent:
4077
    return bdrv_co_flush(bs->file);
P
Paolo Bonzini 已提交
4078 4079
}

4080 4081 4082 4083 4084 4085 4086 4087 4088 4089 4090 4091 4092 4093 4094 4095
void bdrv_invalidate_cache(BlockDriverState *bs)
{
    if (bs->drv && bs->drv->bdrv_invalidate_cache) {
        bs->drv->bdrv_invalidate_cache(bs);
    }
}

void bdrv_invalidate_cache_all(void)
{
    BlockDriverState *bs;

    QTAILQ_FOREACH(bs, &bdrv_states, list) {
        bdrv_invalidate_cache(bs);
    }
}

4096 4097 4098 4099 4100 4101 4102 4103 4104
void bdrv_clear_incoming_migration_all(void)
{
    BlockDriverState *bs;

    QTAILQ_FOREACH(bs, &bdrv_states, list) {
        bs->open_flags = bs->open_flags & ~(BDRV_O_INCOMING);
    }
}

P
Paolo Bonzini 已提交
4105 4106 4107 4108 4109 4110
int bdrv_flush(BlockDriverState *bs)
{
    Coroutine *co;
    RwCo rwco = {
        .bs = bs,
        .ret = NOT_DONE,
4111 4112
    };

P
Paolo Bonzini 已提交
4113 4114 4115 4116 4117 4118 4119 4120 4121
    if (qemu_in_coroutine()) {
        /* Fast-path if already in coroutine context */
        bdrv_flush_co_entry(&rwco);
    } else {
        co = qemu_coroutine_create(bdrv_flush_co_entry);
        qemu_coroutine_enter(co, &rwco);
        while (rwco.ret == NOT_DONE) {
            qemu_aio_wait();
        }
4122
    }
P
Paolo Bonzini 已提交
4123 4124

    return rwco.ret;
4125 4126
}

4127 4128 4129 4130 4131 4132 4133 4134 4135 4136 4137 4138 4139 4140 4141 4142
static void coroutine_fn bdrv_discard_co_entry(void *opaque)
{
    RwCo *rwco = opaque;

    rwco->ret = bdrv_co_discard(rwco->bs, rwco->sector_num, rwco->nb_sectors);
}

int coroutine_fn bdrv_co_discard(BlockDriverState *bs, int64_t sector_num,
                                 int nb_sectors)
{
    if (!bs->drv) {
        return -ENOMEDIUM;
    } else if (bdrv_check_request(bs, sector_num, nb_sectors)) {
        return -EIO;
    } else if (bs->read_only) {
        return -EROFS;
4143 4144 4145
    }

    if (bs->dirty_bitmap) {
4146
        bdrv_reset_dirty(bs, sector_num, nb_sectors);
4147 4148 4149
    }

    if (bs->drv->bdrv_co_discard) {
4150 4151 4152 4153 4154 4155 4156 4157 4158 4159 4160 4161 4162 4163 4164 4165 4166 4167 4168 4169 4170 4171 4172 4173 4174 4175 4176 4177 4178 4179 4180 4181 4182 4183 4184 4185 4186 4187 4188 4189 4190 4191 4192 4193
        return bs->drv->bdrv_co_discard(bs, sector_num, nb_sectors);
    } else if (bs->drv->bdrv_aio_discard) {
        BlockDriverAIOCB *acb;
        CoroutineIOCompletion co = {
            .coroutine = qemu_coroutine_self(),
        };

        acb = bs->drv->bdrv_aio_discard(bs, sector_num, nb_sectors,
                                        bdrv_co_io_em_complete, &co);
        if (acb == NULL) {
            return -EIO;
        } else {
            qemu_coroutine_yield();
            return co.ret;
        }
    } else {
        return 0;
    }
}

int bdrv_discard(BlockDriverState *bs, int64_t sector_num, int nb_sectors)
{
    Coroutine *co;
    RwCo rwco = {
        .bs = bs,
        .sector_num = sector_num,
        .nb_sectors = nb_sectors,
        .ret = NOT_DONE,
    };

    if (qemu_in_coroutine()) {
        /* Fast-path if already in coroutine context */
        bdrv_discard_co_entry(&rwco);
    } else {
        co = qemu_coroutine_create(bdrv_discard_co_entry);
        qemu_coroutine_enter(co, &rwco);
        while (rwco.ret == NOT_DONE) {
            qemu_aio_wait();
        }
    }

    return rwco.ret;
}

B
bellard 已提交
4194 4195 4196 4197 4198 4199 4200 4201 4202
/**************************************************************/
/* removable device support */

/**
 * Return TRUE if the media is present
 */
int bdrv_is_inserted(BlockDriverState *bs)
{
    BlockDriver *drv = bs->drv;
4203

B
bellard 已提交
4204 4205 4206
    if (!drv)
        return 0;
    if (!drv->bdrv_is_inserted)
4207 4208
        return 1;
    return drv->bdrv_is_inserted(bs);
B
bellard 已提交
4209 4210 4211
}

/**
4212 4213
 * 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
4214 4215 4216 4217 4218
 */
int bdrv_media_changed(BlockDriverState *bs)
{
    BlockDriver *drv = bs->drv;

4219 4220 4221 4222
    if (drv && drv->bdrv_media_changed) {
        return drv->bdrv_media_changed(bs);
    }
    return -ENOTSUP;
B
bellard 已提交
4223 4224 4225 4226 4227
}

/**
 * If eject_flag is TRUE, eject the media. Otherwise, close the tray
 */
4228
void bdrv_eject(BlockDriverState *bs, bool eject_flag)
B
bellard 已提交
4229 4230 4231
{
    BlockDriver *drv = bs->drv;

4232 4233
    if (drv && drv->bdrv_eject) {
        drv->bdrv_eject(bs, eject_flag);
B
bellard 已提交
4234
    }
4235 4236 4237 4238

    if (bs->device_name[0] != '\0') {
        bdrv_emit_qmp_eject_event(bs, eject_flag);
    }
B
bellard 已提交
4239 4240 4241 4242 4243 4244
}

/**
 * Lock or unlock the media (if it is locked, the user won't be able
 * to eject it manually).
 */
4245
void bdrv_lock_medium(BlockDriverState *bs, bool locked)
B
bellard 已提交
4246 4247 4248
{
    BlockDriver *drv = bs->drv;

4249
    trace_bdrv_lock_medium(bs, locked);
S
Stefan Hajnoczi 已提交
4250

4251 4252
    if (drv && drv->bdrv_lock_medium) {
        drv->bdrv_lock_medium(bs, locked);
B
bellard 已提交
4253 4254
    }
}
4255 4256 4257 4258 4259 4260 4261 4262 4263 4264 4265

/* needed for generic scsi interface */

int bdrv_ioctl(BlockDriverState *bs, unsigned long int req, void *buf)
{
    BlockDriver *drv = bs->drv;

    if (drv && drv->bdrv_ioctl)
        return drv->bdrv_ioctl(bs, req, buf);
    return -ENOTSUP;
}
4266

4267 4268 4269
BlockDriverAIOCB *bdrv_aio_ioctl(BlockDriverState *bs,
        unsigned long int req, void *buf,
        BlockDriverCompletionFunc *cb, void *opaque)
4270
{
4271
    BlockDriver *drv = bs->drv;
4272

4273 4274 4275
    if (drv && drv->bdrv_aio_ioctl)
        return drv->bdrv_aio_ioctl(bs, req, buf, cb, opaque);
    return NULL;
4276
}
4277

4278 4279 4280 4281
void bdrv_set_buffer_alignment(BlockDriverState *bs, int align)
{
    bs->buffer_alignment = align;
}
4282

4283 4284 4285 4286
void *qemu_blockalign(BlockDriverState *bs, size_t size)
{
    return qemu_memalign((bs && bs->buffer_alignment) ? bs->buffer_alignment : 512, size);
}
4287

4288 4289 4290 4291 4292 4293 4294 4295 4296 4297 4298 4299 4300 4301 4302 4303
/*
 * Check if all memory in this vector is sector aligned.
 */
bool bdrv_qiov_is_aligned(BlockDriverState *bs, QEMUIOVector *qiov)
{
    int i;

    for (i = 0; i < qiov->niov; i++) {
        if ((uintptr_t) qiov->iov[i].iov_base % bs->buffer_alignment) {
            return false;
        }
    }

    return true;
}

4304
void bdrv_set_dirty_tracking(BlockDriverState *bs, int granularity)
4305 4306
{
    int64_t bitmap_size;
4307

4308 4309 4310 4311 4312 4313 4314
    assert((granularity & (granularity - 1)) == 0);

    if (granularity) {
        granularity >>= BDRV_SECTOR_BITS;
        assert(!bs->dirty_bitmap);
        bitmap_size = (bdrv_getlength(bs) >> BDRV_SECTOR_BITS);
        bs->dirty_bitmap = hbitmap_alloc(bitmap_size, ffs(granularity) - 1);
4315
    } else {
4316
        if (bs->dirty_bitmap) {
4317
            hbitmap_free(bs->dirty_bitmap);
4318
            bs->dirty_bitmap = NULL;
4319
        }
4320 4321 4322 4323 4324
    }
}

int bdrv_get_dirty(BlockDriverState *bs, int64_t sector)
{
4325 4326
    if (bs->dirty_bitmap) {
        return hbitmap_get(bs->dirty_bitmap, sector);
4327 4328 4329 4330 4331
    } else {
        return 0;
    }
}

4332
void bdrv_dirty_iter_init(BlockDriverState *bs, HBitmapIter *hbi)
4333
{
4334
    hbitmap_iter_init(hbi, bs->dirty_bitmap, 0);
4335 4336 4337 4338 4339
}

void bdrv_set_dirty(BlockDriverState *bs, int64_t cur_sector,
                    int nr_sectors)
{
4340
    hbitmap_set(bs->dirty_bitmap, cur_sector, nr_sectors);
4341 4342
}

4343 4344
void bdrv_reset_dirty(BlockDriverState *bs, int64_t cur_sector,
                      int nr_sectors)
4345
{
4346
    hbitmap_reset(bs->dirty_bitmap, cur_sector, nr_sectors);
4347
}
4348 4349 4350

int64_t bdrv_get_dirty_count(BlockDriverState *bs)
{
4351
    if (bs->dirty_bitmap) {
4352
        return hbitmap_count(bs->dirty_bitmap);
4353 4354 4355
    } else {
        return 0;
    }
4356
}
J
Jes Sorensen 已提交
4357

4358 4359 4360 4361 4362 4363 4364 4365 4366 4367 4368
void bdrv_set_in_use(BlockDriverState *bs, int in_use)
{
    assert(bs->in_use != in_use);
    bs->in_use = in_use;
}

int bdrv_in_use(BlockDriverState *bs)
{
    return bs->in_use;
}

4369 4370
void bdrv_iostatus_enable(BlockDriverState *bs)
{
4371
    bs->iostatus_enabled = true;
4372
    bs->iostatus = BLOCK_DEVICE_IO_STATUS_OK;
4373 4374 4375 4376 4377 4378
}

/* 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)
{
4379
    return (bs->iostatus_enabled &&
4380 4381 4382
           (bs->on_write_error == BLOCKDEV_ON_ERROR_ENOSPC ||
            bs->on_write_error == BLOCKDEV_ON_ERROR_STOP   ||
            bs->on_read_error == BLOCKDEV_ON_ERROR_STOP));
4383 4384 4385 4386
}

void bdrv_iostatus_disable(BlockDriverState *bs)
{
4387
    bs->iostatus_enabled = false;
4388 4389 4390 4391 4392
}

void bdrv_iostatus_reset(BlockDriverState *bs)
{
    if (bdrv_iostatus_is_enabled(bs)) {
4393
        bs->iostatus = BLOCK_DEVICE_IO_STATUS_OK;
4394 4395 4396
        if (bs->job) {
            block_job_iostatus_reset(bs->job);
        }
4397 4398 4399 4400 4401
    }
}

void bdrv_iostatus_set_err(BlockDriverState *bs, int error)
{
4402 4403
    assert(bdrv_iostatus_is_enabled(bs));
    if (bs->iostatus == BLOCK_DEVICE_IO_STATUS_OK) {
4404 4405
        bs->iostatus = error == ENOSPC ? BLOCK_DEVICE_IO_STATUS_NOSPACE :
                                         BLOCK_DEVICE_IO_STATUS_FAILED;
4406 4407 4408
    }
}

4409 4410 4411 4412 4413 4414 4415
void
bdrv_acct_start(BlockDriverState *bs, BlockAcctCookie *cookie, int64_t bytes,
        enum BlockAcctType type)
{
    assert(type < BDRV_MAX_IOTYPE);

    cookie->bytes = bytes;
C
Christoph Hellwig 已提交
4416
    cookie->start_time_ns = get_clock();
4417 4418 4419 4420 4421 4422 4423 4424 4425 4426
    cookie->type = type;
}

void
bdrv_acct_done(BlockDriverState *bs, BlockAcctCookie *cookie)
{
    assert(cookie->type < BDRV_MAX_IOTYPE);

    bs->nr_bytes[cookie->type] += cookie->bytes;
    bs->nr_ops[cookie->type]++;
C
Christoph Hellwig 已提交
4427
    bs->total_time_ns[cookie->type] += get_clock() - cookie->start_time_ns;
4428 4429
}

4430 4431 4432
void bdrv_img_create(const char *filename, const char *fmt,
                     const char *base_filename, const char *base_fmt,
                     char *options, uint64_t img_size, int flags, Error **errp)
J
Jes Sorensen 已提交
4433 4434
{
    QEMUOptionParameter *param = NULL, *create_options = NULL;
K
Kevin Wolf 已提交
4435
    QEMUOptionParameter *backing_fmt, *backing_file, *size;
J
Jes Sorensen 已提交
4436 4437
    BlockDriverState *bs = NULL;
    BlockDriver *drv, *proto_drv;
4438
    BlockDriver *backing_drv = NULL;
J
Jes Sorensen 已提交
4439 4440 4441 4442 4443
    int ret = 0;

    /* Find driver and parse its options */
    drv = bdrv_find_format(fmt);
    if (!drv) {
4444
        error_setg(errp, "Unknown file format '%s'", fmt);
4445
        return;
J
Jes Sorensen 已提交
4446 4447 4448 4449
    }

    proto_drv = bdrv_find_protocol(filename);
    if (!proto_drv) {
4450
        error_setg(errp, "Unknown protocol '%s'", filename);
4451
        return;
J
Jes Sorensen 已提交
4452 4453 4454 4455 4456 4457 4458 4459 4460 4461 4462 4463 4464 4465 4466 4467
    }

    create_options = append_option_parameters(create_options,
                                              drv->create_options);
    create_options = append_option_parameters(create_options,
                                              proto_drv->create_options);

    /* Create parameter list with default values */
    param = parse_option_parameters("", create_options, param);

    set_option_parameter_int(param, BLOCK_OPT_SIZE, img_size);

    /* Parse -o options */
    if (options) {
        param = parse_option_parameters(options, create_options, param);
        if (param == NULL) {
4468
            error_setg(errp, "Invalid options for file format '%s'.", fmt);
J
Jes Sorensen 已提交
4469 4470 4471 4472 4473 4474 4475
            goto out;
        }
    }

    if (base_filename) {
        if (set_option_parameter(param, BLOCK_OPT_BACKING_FILE,
                                 base_filename)) {
4476 4477
            error_setg(errp, "Backing file not supported for file format '%s'",
                       fmt);
J
Jes Sorensen 已提交
4478 4479 4480 4481 4482 4483
            goto out;
        }
    }

    if (base_fmt) {
        if (set_option_parameter(param, BLOCK_OPT_BACKING_FMT, base_fmt)) {
4484 4485
            error_setg(errp, "Backing file format not supported for file "
                             "format '%s'", fmt);
J
Jes Sorensen 已提交
4486 4487 4488 4489
            goto out;
        }
    }

4490 4491 4492
    backing_file = get_option_parameter(param, BLOCK_OPT_BACKING_FILE);
    if (backing_file && backing_file->value.s) {
        if (!strcmp(filename, backing_file->value.s)) {
4493 4494
            error_setg(errp, "Error: Trying to create an image with the "
                             "same filename as the backing file");
4495 4496 4497 4498
            goto out;
        }
    }

J
Jes Sorensen 已提交
4499 4500
    backing_fmt = get_option_parameter(param, BLOCK_OPT_BACKING_FMT);
    if (backing_fmt && backing_fmt->value.s) {
4501 4502
        backing_drv = bdrv_find_format(backing_fmt->value.s);
        if (!backing_drv) {
4503 4504
            error_setg(errp, "Unknown backing file format '%s'",
                       backing_fmt->value.s);
J
Jes Sorensen 已提交
4505 4506 4507 4508 4509 4510
            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
K
Kevin Wolf 已提交
4511 4512
    size = get_option_parameter(param, BLOCK_OPT_SIZE);
    if (size && size->value.n == -1) {
J
Jes Sorensen 已提交
4513 4514 4515
        if (backing_file && backing_file->value.s) {
            uint64_t size;
            char buf[32];
4516 4517 4518 4519 4520
            int back_flags;

            /* 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
4521 4522 4523

            bs = bdrv_new("");

4524
            ret = bdrv_open(bs, backing_file->value.s, back_flags, backing_drv);
J
Jes Sorensen 已提交
4525
            if (ret < 0) {
4526 4527
                error_setg_errno(errp, -ret, "Could not open '%s'",
                                 backing_file->value.s);
J
Jes Sorensen 已提交
4528 4529 4530 4531 4532 4533 4534 4535
                goto out;
            }
            bdrv_get_geometry(bs, &size);
            size *= 512;

            snprintf(buf, sizeof(buf), "%" PRId64, size);
            set_option_parameter(param, BLOCK_OPT_SIZE, buf);
        } else {
4536
            error_setg(errp, "Image creation needs a size parameter");
J
Jes Sorensen 已提交
4537 4538 4539 4540 4541 4542 4543 4544 4545 4546 4547
            goto out;
        }
    }

    printf("Formatting '%s', fmt=%s ", filename, fmt);
    print_option_parameters(param);
    puts("");

    ret = bdrv_create(drv, filename, param);
    if (ret < 0) {
        if (ret == -ENOTSUP) {
4548 4549
            error_setg(errp,"Formatting or formatting option not supported for "
                            "file format '%s'", fmt);
J
Jes Sorensen 已提交
4550
        } else if (ret == -EFBIG) {
4551 4552
            error_setg(errp, "The image size is too large for file format '%s'",
                       fmt);
J
Jes Sorensen 已提交
4553
        } else {
4554 4555
            error_setg(errp, "%s: error while creating %s: %s", filename, fmt,
                       strerror(-ret));
J
Jes Sorensen 已提交
4556 4557 4558 4559 4560 4561 4562 4563 4564 4565 4566
        }
    }

out:
    free_option_parameters(create_options);
    free_option_parameters(param);

    if (bs) {
        bdrv_delete(bs);
    }
}