block.c 124.1 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"
A
aliguori 已提交
27
#include "monitor.h"
B
bellard 已提交
28
#include "block_int.h"
29
#include "blockjob.h"
30
#include "module.h"
31
#include "qjson.h"
32
#include "sysemu.h"
P
Paolo Bonzini 已提交
33
#include "notify.h"
34
#include "qemu-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));
}

147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175
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->slice_time  = 5 * BLOCK_IO_SLICE_TIME;
    bs->slice_start = qemu_get_clock_ns(vm_clock);
    bs->slice_end   = bs->slice_start + bs->slice_time;
    memset(&bs->io_base, 0, sizeof(bs->io_base));
    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
176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200
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);
}

201 202 203
/* check if the path starts with "<protocol>:" */
static int path_has_protocol(const char *path)
{
204 205
    const char *p;

206 207 208 209 210
#ifdef _WIN32
    if (is_windows_drive(path) ||
        is_windows_drive_prefix(path)) {
        return 0;
    }
211 212 213
    p = path + strcspn(path, ":/\\");
#else
    p = path + strcspn(path, ":/");
214 215
#endif

216
    return *p == ':';
217 218
}

B
bellard 已提交
219
int path_is_absolute(const char *path)
220
{
B
bellard 已提交
221 222
#ifdef _WIN32
    /* specific case for names like: "\\.\d:" */
P
Paolo Bonzini 已提交
223
    if (is_windows_drive(path) || is_windows_drive_prefix(path)) {
B
bellard 已提交
224
        return 1;
P
Paolo Bonzini 已提交
225 226
    }
    return (*path == '/' || *path == '\\');
227
#else
P
Paolo Bonzini 已提交
228
    return (*path == '/');
229
#endif
230 231
}

B
bellard 已提交
232 233 234 235 236 237
/* 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)
238
{
B
bellard 已提交
239 240 241 242 243 244 245 246 247 248 249 250 251
    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;
252 253 254 255 256 257 258 259 260
        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
261 262 263 264 265 266 267 268 269 270 271 272
        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);
273 274 275
    }
}

276 277 278 279 280 281 282 283 284
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);
    }
}

285
void bdrv_register(BlockDriver *bdrv)
B
bellard 已提交
286
{
287 288
    /* Block drivers without coroutine functions need emulation */
    if (!bdrv->bdrv_co_readv) {
289 290 291
        bdrv->bdrv_co_readv = bdrv_co_readv_em;
        bdrv->bdrv_co_writev = bdrv_co_writev_em;

292 293 294
        /* 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.
         */
295 296 297 298 299
        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
300
    }
301

302
    QLIST_INSERT_HEAD(&bdrv_drivers, bdrv, list);
B
bellard 已提交
303
}
B
bellard 已提交
304 305 306 307

/* create a new block device (by default it is empty) */
BlockDriverState *bdrv_new(const char *device_name)
{
308
    BlockDriverState *bs;
B
bellard 已提交
309

310
    bs = g_malloc0(sizeof(BlockDriverState));
B
bellard 已提交
311
    pstrcpy(bs->device_name, sizeof(bs->device_name), device_name);
B
bellard 已提交
312
    if (device_name[0] != '\0') {
313
        QTAILQ_INSERT_TAIL(&bdrv_states, bs, list);
B
bellard 已提交
314
    }
315
    bdrv_iostatus_disable(bs);
P
Paolo Bonzini 已提交
316 317
    notifier_list_init(&bs->close_notifiers);

B
bellard 已提交
318 319 320
    return bs;
}

P
Paolo Bonzini 已提交
321 322 323 324 325
void bdrv_add_close_notifier(BlockDriverState *bs, Notifier *notify)
{
    notifier_list_add(&bs->close_notifiers, notify);
}

B
bellard 已提交
326 327 328
BlockDriver *bdrv_find_format(const char *format_name)
{
    BlockDriver *drv1;
329 330
    QLIST_FOREACH(drv1, &bdrv_drivers, list) {
        if (!strcmp(drv1->format_name, format_name)) {
B
bellard 已提交
331
            return drv1;
332
        }
B
bellard 已提交
333 334 335 336
    }
    return NULL;
}

337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360
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;
}

361 362 363 364 365 366 367 368 369 370 371 372 373 374 375
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);
}

376 377
int bdrv_create(BlockDriver *drv, const char* filename,
    QEMUOptionParameter *options)
B
bellard 已提交
378
{
379 380 381 382 383 384 385 386 387 388 389
    int ret;

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

    if (!drv->bdrv_create) {
390 391
        ret = -ENOTSUP;
        goto out;
392 393 394 395 396 397 398 399 400 401 402 403 404 405
    }

    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;
406

407 408
out:
    g_free(cco.filename);
409
    return ret;
B
bellard 已提交
410 411
}

412 413 414 415
int bdrv_create_file(const char* filename, QEMUOptionParameter *options)
{
    BlockDriver *drv;

416
    drv = bdrv_find_protocol(filename);
417
    if (drv == NULL) {
418
        return -ENOENT;
419 420 421 422 423
    }

    return bdrv_create(drv, filename, options);
}

424 425 426 427 428
/*
 * 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
429
{
430
#ifdef _WIN32
431
    char temp_dir[MAX_PATH];
432 433 434 435 436 437
    /* 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
438
#else
B
bellard 已提交
439
    int fd;
440
    const char *tmpdir;
A
aurel32 已提交
441 442 443
    tmpdir = getenv("TMPDIR");
    if (!tmpdir)
        tmpdir = "/tmp";
444 445 446
    if (snprintf(filename, size, "%s/vl.XXXXXX", tmpdir) >= size) {
        return -EOVERFLOW;
    }
B
bellard 已提交
447
    fd = mkstemp(filename);
448 449 450 451 452
    if (fd < 0) {
        return -errno;
    }
    if (close(fd) != 0) {
        unlink(filename);
453 454 455
        return -errno;
    }
    return 0;
B
bellard 已提交
456
#endif
457
}
B
bellard 已提交
458

459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480
/*
 * 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;
}

481
BlockDriver *bdrv_find_protocol(const char *filename)
B
bellard 已提交
482 483 484
{
    BlockDriver *drv1;
    char protocol[128];
485
    int len;
B
bellard 已提交
486
    const char *p;
B
bellard 已提交
487

488 489
    /* TODO Drivers without bdrv_file_open must be specified explicitly */

490 491 492 493 494 495 496 497 498 499 500 501
    /*
     * 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;
    }

502
    if (!path_has_protocol(filename)) {
503
        return bdrv_find_format("file");
504
    }
505 506
    p = strchr(filename, ':');
    assert(p != NULL);
507 508 509 510 511
    len = p - filename;
    if (len > sizeof(protocol) - 1)
        len = sizeof(protocol) - 1;
    memcpy(protocol, filename, len);
    protocol[len] = '\0';
512
    QLIST_FOREACH(drv1, &bdrv_drivers, list) {
513
        if (drv1->protocol_name &&
514
            !strcmp(drv1->protocol_name, protocol)) {
B
bellard 已提交
515
            return drv1;
516
        }
B
bellard 已提交
517 518 519 520
    }
    return NULL;
}

521
static int find_image_format(const char *filename, BlockDriver **pdrv)
522 523 524 525 526 527
{
    int ret, score, score_max;
    BlockDriver *drv1, *drv;
    uint8_t buf[2048];
    BlockDriverState *bs;

528
    ret = bdrv_file_open(&bs, filename, 0);
529 530 531 532
    if (ret < 0) {
        *pdrv = NULL;
        return ret;
    }
533

534 535
    /* Return the raw BlockDriver * to scsi-generic devices or empty drives */
    if (bs->sg || !bdrv_is_inserted(bs)) {
536
        bdrv_delete(bs);
537 538 539 540 541 542
        drv = bdrv_find_format("raw");
        if (!drv) {
            ret = -ENOENT;
        }
        *pdrv = drv;
        return ret;
543
    }
544

B
bellard 已提交
545 546 547
    ret = bdrv_pread(bs, 0, buf, sizeof(buf));
    bdrv_delete(bs);
    if (ret < 0) {
548 549
        *pdrv = NULL;
        return ret;
B
bellard 已提交
550 551
    }

B
bellard 已提交
552
    score_max = 0;
553
    drv = NULL;
554
    QLIST_FOREACH(drv1, &bdrv_drivers, list) {
B
bellard 已提交
555 556 557 558 559 560
        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
561
        }
B
bellard 已提交
562
    }
563 564 565 566 567
    if (!drv) {
        ret = -ENOENT;
    }
    *pdrv = drv;
    return ret;
B
bellard 已提交
568 569
}

570 571 572 573 574 575 576
/**
 * Set the current 'total_sectors' value
 */
static int refresh_total_sectors(BlockDriverState *bs, int64_t hint)
{
    BlockDriver *drv = bs->drv;

577 578 579 580
    /* Do not attempt drv->bdrv_getlength() on scsi-generic devices */
    if (bs->sg)
        return 0;

581 582 583 584 585 586 587 588 589 590 591 592 593
    /* 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;
}

594 595 596 597 598 599 600 601 602 603 604
/**
 * 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;
605 606
    } else if (!strcmp(mode, "directsync")) {
        *flags |= BDRV_O_NOCACHE;
607 608 609 610 611 612 613 614 615 616 617 618 619 620
    } 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;
}

621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636
/**
 * 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--;
}

637 638 639 640 641 642 643 644 645
/*
 * Common part for opening disk images and files
 */
static int bdrv_open_common(BlockDriverState *bs, const char *filename,
    int flags, BlockDriver *drv)
{
    int ret, open_flags;

    assert(drv != NULL);
646
    assert(bs->file == NULL);
647

S
Stefan Hajnoczi 已提交
648 649
    trace_bdrv_open_common(bs, filename, flags, drv->format_name);

650 651 652
    bs->open_flags = flags;
    bs->buffer_alignment = 512;

653 654 655 656 657
    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);
    }

658 659 660 661 662 663 664
    pstrcpy(bs->filename, sizeof(bs->filename), filename);

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

    bs->drv = drv;
665
    bs->opaque = g_malloc0(drv->instance_size);
666

667
    bs->enable_write_cache = !!(flags & BDRV_O_CACHE_WB);
668
    open_flags = flags | BDRV_O_CACHE_WB;
669 670 671 672 673

    /*
     * Clear flags that are internal to the block layer before opening the
     * image.
     */
674
    open_flags &= ~(BDRV_O_SNAPSHOT | BDRV_O_NO_BACKING);
675 676

    /*
S
Stefan Weil 已提交
677
     * Snapshots should be writable.
678 679 680 681 682
     */
    if (bs->is_temporary) {
        open_flags |= BDRV_O_RDWR;
    }

683
    bs->read_only = !(open_flags & BDRV_O_RDWR);
684

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

695 696 697 698
    if (ret < 0) {
        goto free_and_fail;
    }

699 700 701
    ret = refresh_total_sectors(bs, bs->total_sectors);
    if (ret < 0) {
        goto free_and_fail;
702
    }
703

704 705 706 707 708 709 710 711
#ifndef _WIN32
    if (bs->is_temporary) {
        unlink(filename);
    }
#endif
    return 0;

free_and_fail:
712 713 714 715
    if (bs->file) {
        bdrv_delete(bs->file);
        bs->file = NULL;
    }
716
    g_free(bs->opaque);
717 718 719 720 721
    bs->opaque = NULL;
    bs->drv = NULL;
    return ret;
}

K
Kevin Wolf 已提交
722 723 724
/*
 * Opens a file using a protocol (file, host_device, nbd, ...)
 */
B
bellard 已提交
725
int bdrv_file_open(BlockDriverState **pbs, const char *filename, int flags)
B
bellard 已提交
726
{
B
bellard 已提交
727
    BlockDriverState *bs;
728
    BlockDriver *drv;
B
bellard 已提交
729 730
    int ret;

731
    drv = bdrv_find_protocol(filename);
732 733 734 735
    if (!drv) {
        return -ENOENT;
    }

B
bellard 已提交
736
    bs = bdrv_new("");
K
Kevin Wolf 已提交
737
    ret = bdrv_open_common(bs, filename, flags, drv);
B
bellard 已提交
738 739 740
    if (ret < 0) {
        bdrv_delete(bs);
        return ret;
741
    }
742
    bs->growable = 1;
B
bellard 已提交
743 744 745 746
    *pbs = bs;
    return 0;
}

P
Paolo Bonzini 已提交
747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782
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
783 784 785
/*
 * Opens a disk image (raw, qcow2, vmdk, ...)
 */
K
Kevin Wolf 已提交
786 787
int bdrv_open(BlockDriverState *bs, const char *filename, int flags,
              BlockDriver *drv)
B
bellard 已提交
788
{
K
Kevin Wolf 已提交
789
    int ret;
790 791
    /* TODO: extra byte is a hack to ensure MAX_PATH space on Windows. */
    char tmp_filename[PATH_MAX + 1];
B
bellard 已提交
792

B
bellard 已提交
793
    if (flags & BDRV_O_SNAPSHOT) {
B
bellard 已提交
794 795
        BlockDriverState *bs1;
        int64_t total_size;
A
aliguori 已提交
796
        int is_protocol = 0;
K
Kevin Wolf 已提交
797 798
        BlockDriver *bdrv_qcow2;
        QEMUOptionParameter *options;
K
Kevin Wolf 已提交
799
        char backing_filename[PATH_MAX];
800

B
bellard 已提交
801 802
        /* if snapshot, we create a temporary backing file and open it
           instead of opening 'filename' directly */
803

B
bellard 已提交
804 805
        /* if there is a backing file, use it */
        bs1 = bdrv_new("");
K
Kevin Wolf 已提交
806
        ret = bdrv_open(bs1, filename, 0, drv);
807
        if (ret < 0) {
B
bellard 已提交
808
            bdrv_delete(bs1);
809
            return ret;
B
bellard 已提交
810
        }
811
        total_size = bdrv_getlength(bs1) & BDRV_SECTOR_MASK;
A
aliguori 已提交
812 813 814 815

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

B
bellard 已提交
816
        bdrv_delete(bs1);
817

818 819 820 821
        ret = get_tmp_filename(tmp_filename, sizeof(tmp_filename));
        if (ret < 0) {
            return ret;
        }
A
aliguori 已提交
822 823 824 825 826

        /* Real path is meaningless for protocols */
        if (is_protocol)
            snprintf(backing_filename, sizeof(backing_filename),
                     "%s", filename);
827 828
        else if (!realpath(filename, backing_filename))
            return -errno;
A
aliguori 已提交
829

K
Kevin Wolf 已提交
830 831 832
        bdrv_qcow2 = bdrv_find_format("qcow2");
        options = parse_option_parameters("", bdrv_qcow2->create_options, NULL);

833
        set_option_parameter_int(options, BLOCK_OPT_SIZE, total_size);
K
Kevin Wolf 已提交
834 835 836 837 838 839 840
        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);
841
        free_option_parameters(options);
842 843
        if (ret < 0) {
            return ret;
B
bellard 已提交
844
        }
K
Kevin Wolf 已提交
845

B
bellard 已提交
846
        filename = tmp_filename;
K
Kevin Wolf 已提交
847
        drv = bdrv_qcow2;
B
bellard 已提交
848 849
        bs->is_temporary = 1;
    }
B
bellard 已提交
850

K
Kevin Wolf 已提交
851
    /* Find the right image format driver */
852
    if (!drv) {
853
        ret = find_image_format(filename, &drv);
854
    }
855

856 857
    if (!drv) {
        goto unlink_and_fail;
B
bellard 已提交
858
    }
K
Kevin Wolf 已提交
859

860 861 862 863
    if (flags & BDRV_O_RDWR) {
        flags |= BDRV_O_ALLOW_RDWR;
    }

K
Kevin Wolf 已提交
864 865 866
    /* Open the image */
    ret = bdrv_open_common(bs, filename, flags, drv);
    if (ret < 0) {
867 868 869
        goto unlink_and_fail;
    }

K
Kevin Wolf 已提交
870
    /* If there is a backing file, use it */
P
Paolo Bonzini 已提交
871 872
    if ((flags & BDRV_O_NO_BACKING) == 0) {
        ret = bdrv_open_backing_file(bs);
K
Kevin Wolf 已提交
873 874 875 876 877 878 879
        if (ret < 0) {
            bdrv_close(bs);
            return ret;
        }
    }

    if (!bdrv_key_required(bs)) {
880
        bdrv_dev_change_media_cb(bs, true);
K
Kevin Wolf 已提交
881 882
    }

Z
Zhi Yong Wu 已提交
883 884 885 886 887
    /* throttling disk I/O limits */
    if (bs->io_limits_enabled) {
        bdrv_io_limits_enable(bs);
    }

K
Kevin Wolf 已提交
888 889 890 891 892 893 894 895 896
    return 0;

unlink_and_fail:
    if (bs->is_temporary) {
        unlink(filename);
    }
    return ret;
}

897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 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
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
1129 1130
void bdrv_close(BlockDriverState *bs)
{
1131
    bdrv_flush(bs);
1132 1133 1134 1135
    if (bs->job) {
        block_job_cancel_sync(bs->job);
    }
    bdrv_drain_all();
P
Paolo Bonzini 已提交
1136
    notifier_list_notify(&bs->close_notifiers, bs);
K
Kevin Wolf 已提交
1137

1138
    if (bs->drv) {
1139 1140 1141
        if (bs == bs_snapshots) {
            bs_snapshots = NULL;
        }
1142
        if (bs->backing_hd) {
B
bellard 已提交
1143
            bdrv_delete(bs->backing_hd);
1144 1145
            bs->backing_hd = NULL;
        }
B
bellard 已提交
1146
        bs->drv->bdrv_close(bs);
1147
        g_free(bs->opaque);
B
bellard 已提交
1148 1149 1150 1151
#ifdef _WIN32
        if (bs->is_temporary) {
            unlink(bs->filename);
        }
B
bellard 已提交
1152
#endif
B
bellard 已提交
1153 1154
        bs->opaque = NULL;
        bs->drv = NULL;
1155
        bs->copy_on_read = 0;
1156 1157
        bs->backing_file[0] = '\0';
        bs->backing_format[0] = '\0';
1158 1159 1160 1161 1162
        bs->total_sectors = 0;
        bs->encrypted = 0;
        bs->valid_key = 0;
        bs->sg = 0;
        bs->growable = 0;
B
bellard 已提交
1163

1164
        if (bs->file != NULL) {
1165 1166
            bdrv_delete(bs->file);
            bs->file = NULL;
1167
        }
B
bellard 已提交
1168
    }
Z
Zhi Yong Wu 已提交
1169

P
Pavel Hrdina 已提交
1170 1171
    bdrv_dev_change_media_cb(bs, false);

Z
Zhi Yong Wu 已提交
1172 1173 1174 1175
    /*throttling disk I/O limits*/
    if (bs->io_limits_enabled) {
        bdrv_io_limits_disable(bs);
    }
B
bellard 已提交
1176 1177
}

1178 1179 1180 1181 1182 1183 1184 1185 1186
void bdrv_close_all(void)
{
    BlockDriverState *bs;

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

1187 1188 1189 1190 1191
/*
 * 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.
1192 1193 1194 1195 1196 1197
 *
 * 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.
1198 1199 1200 1201
 */
void bdrv_drain_all(void)
{
    BlockDriverState *bs;
1202 1203 1204 1205
    bool busy;

    do {
        busy = qemu_aio_wait();
1206

1207 1208 1209 1210 1211 1212 1213 1214 1215 1216
        /* 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);
1217 1218 1219 1220 1221 1222 1223 1224

    /* 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));
    }
}

1225 1226 1227 1228 1229 1230 1231 1232 1233 1234
/* 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
1235 1236 1237 1238 1239 1240 1241
static void bdrv_rebind(BlockDriverState *bs)
{
    if (bs->drv && bs->drv->bdrv_rebind) {
        bs->drv->bdrv_rebind(bs);
    }
}

1242 1243
static void bdrv_move_feature_fields(BlockDriverState *bs_dest,
                                     BlockDriverState *bs_src)
1244
{
1245 1246
    /* move some fields that need to stay attached to the device */
    bs_dest->open_flags         = bs_src->open_flags;
1247 1248

    /* dev info */
1249 1250 1251 1252 1253
    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;
1254

1255
    bs_dest->enable_write_cache = bs_src->enable_write_cache;
1256

1257
    /* i/o timing parameters */
1258 1259 1260 1261 1262 1263 1264 1265
    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;
1266 1267

    /* r/w error */
1268 1269
    bs_dest->on_read_error      = bs_src->on_read_error;
    bs_dest->on_write_error     = bs_src->on_write_error;
1270 1271

    /* i/o status */
1272 1273
    bs_dest->iostatus_enabled   = bs_src->iostatus_enabled;
    bs_dest->iostatus           = bs_src->iostatus;
1274

1275
    /* dirty bitmap */
1276 1277
    bs_dest->dirty_count        = bs_src->dirty_count;
    bs_dest->dirty_bitmap       = bs_src->dirty_bitmap;
1278 1279

    /* job */
1280 1281
    bs_dest->in_use             = bs_src->in_use;
    bs_dest->job                = bs_src->job;
1282

1283
    /* keep the same entry in bdrv_states */
1284 1285 1286 1287
    pstrcpy(bs_dest->device_name, sizeof(bs_dest->device_name),
            bs_src->device_name);
    bs_dest->list = bs_src->list;
}
1288

1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303
/*
 * 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
1304

1305 1306 1307 1308 1309 1310 1311 1312
    /* 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);
1313

1314 1315 1316
    tmp = *bs_new;
    *bs_new = *bs_old;
    *bs_old = tmp;
1317

1318 1319 1320 1321
    /* 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);
1322

1323 1324 1325 1326 1327 1328 1329 1330 1331
    /* 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
1332 1333

    bdrv_rebind(bs_new);
1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359
    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 : "");
1360 1361
}

B
bellard 已提交
1362 1363
void bdrv_delete(BlockDriverState *bs)
{
1364
    assert(!bs->dev);
1365 1366
    assert(!bs->job);
    assert(!bs->in_use);
1367

1368
    /* remove from list, if necessary */
1369
    bdrv_make_anon(bs);
1370

B
bellard 已提交
1371
    bdrv_close(bs);
1372

1373
    assert(bs != bs_snapshots);
1374
    g_free(bs);
B
bellard 已提交
1375 1376
}

1377 1378
int bdrv_attach_dev(BlockDriverState *bs, void *dev)
/* TODO change to DeviceState *dev when all users are qdevified */
1379
{
1380
    if (bs->dev) {
1381 1382
        return -EBUSY;
    }
1383
    bs->dev = dev;
1384
    bdrv_iostatus_reset(bs);
1385 1386 1387
    return 0;
}

1388 1389
/* TODO qdevified devices don't use this, remove when devices are qdevified */
void bdrv_attach_dev_nofail(BlockDriverState *bs, void *dev)
1390
{
1391 1392 1393 1394 1395 1396 1397 1398 1399 1400
    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;
1401 1402
    bs->dev_ops = NULL;
    bs->dev_opaque = NULL;
1403
    bs->buffer_alignment = 512;
1404 1405
}

1406 1407
/* TODO change to return DeviceState * when all users are qdevified */
void *bdrv_get_attached_dev(BlockDriverState *bs)
1408
{
1409
    return bs->dev;
1410 1411
}

1412 1413 1414 1415 1416
void bdrv_set_dev_ops(BlockDriverState *bs, const BlockDevOps *ops,
                      void *opaque)
{
    bs->dev_ops = ops;
    bs->dev_opaque = opaque;
1417 1418 1419
    if (bdrv_dev_has_removable_media(bs) && bs == bs_snapshots) {
        bs_snapshots = NULL;
    }
1420 1421
}

P
Paolo Bonzini 已提交
1422 1423 1424
void bdrv_emit_qmp_error_event(const BlockDriverState *bdrv,
                               enum MonitorEvent ev,
                               BlockErrorAction action, bool is_read)
1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446
{
    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
1447
    monitor_protocol_event(ev, data);
1448 1449 1450 1451

    qobject_decref(data);
}

1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462
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);
}

1463
static void bdrv_dev_change_media_cb(BlockDriverState *bs, bool load)
1464
{
1465
    if (bs->dev_ops && bs->dev_ops->change_media_cb) {
1466
        bool tray_was_closed = !bdrv_dev_is_tray_open(bs);
1467
        bs->dev_ops->change_media_cb(bs->dev_opaque, load);
1468 1469 1470 1471 1472 1473 1474 1475
        if (tray_was_closed) {
            /* tray open */
            bdrv_emit_qmp_eject_event(bs, true);
        }
        if (load) {
            /* tray close */
            bdrv_emit_qmp_eject_event(bs, false);
        }
1476 1477 1478
    }
}

1479 1480 1481 1482 1483
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
1484 1485 1486 1487 1488 1489 1490
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);
    }
}

1491 1492 1493 1494 1495 1496 1497 1498
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;
}

1499 1500 1501 1502
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);
1503 1504 1505
    }
}

1506 1507 1508 1509 1510 1511 1512 1513
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
1514 1515 1516
/*
 * Run consistency checks on an image
 *
1517
 * Returns 0 if the check could be completed (it doesn't mean that the image is
1518
 * free of errors) or -errno when an internal error occurred. The results of the
1519
 * check are stored in res.
A
aliguori 已提交
1520
 */
1521
int bdrv_check(BlockDriverState *bs, BdrvCheckResult *res, BdrvCheckMode fix)
A
aliguori 已提交
1522 1523 1524 1525 1526
{
    if (bs->drv->bdrv_check == NULL) {
        return -ENOTSUP;
    }

1527
    memset(res, 0, sizeof(*res));
1528
    return bs->drv->bdrv_check(bs, res, fix);
A
aliguori 已提交
1529 1530
}

1531 1532
#define COMMIT_BUF_SECTORS 2048

1533 1534 1535
/* commit COW file into the raw image */
int bdrv_commit(BlockDriverState *bs)
{
B
bellard 已提交
1536
    BlockDriver *drv = bs->drv;
1537 1538
    int64_t sector, total_sectors;
    int n, ro, open_flags;
1539
    int ret = 0;
1540
    uint8_t *buf;
1541
    char filename[PATH_MAX];
1542

B
bellard 已提交
1543 1544
    if (!drv)
        return -ENOMEDIUM;
1545 1546 1547
    
    if (!bs->backing_hd) {
        return -ENOTSUP;
1548 1549
    }

1550 1551 1552 1553
    if (bdrv_in_use(bs) || bdrv_in_use(bs->backing_hd)) {
        return -EBUSY;
    }

1554
    ro = bs->backing_hd->read_only;
1555 1556
    /* Use pstrcpy (not strncpy): filename must be NUL-terminated. */
    pstrcpy(filename, sizeof(filename), bs->backing_hd->filename);
1557 1558 1559
    open_flags =  bs->backing_hd->open_flags;

    if (ro) {
1560 1561
        if (bdrv_reopen(bs->backing_hd, open_flags | BDRV_O_RDWR, NULL)) {
            return -EACCES;
1562
        }
B
bellard 已提交
1563
    }
1564

J
Jan Kiszka 已提交
1565
    total_sectors = bdrv_getlength(bs) >> BDRV_SECTOR_BITS;
1566
    buf = g_malloc(COMMIT_BUF_SECTORS * BDRV_SECTOR_SIZE);
1567 1568

    for (sector = 0; sector < total_sectors; sector += n) {
1569
        if (bdrv_is_allocated(bs, sector, COMMIT_BUF_SECTORS, &n)) {
1570 1571 1572 1573 1574 1575 1576 1577 1578 1579

            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
1580
        }
1581
    }
1582

1583 1584 1585 1586
    if (drv->bdrv_make_empty) {
        ret = drv->bdrv_make_empty(bs);
        bdrv_flush(bs);
    }
1587

1588 1589 1590 1591 1592 1593
    /*
     * Make sure all data we wrote to the backing device is actually
     * stable on disk.
     */
    if (bs->backing_hd)
        bdrv_flush(bs->backing_hd);
1594 1595

ro_cleanup:
1596
    g_free(buf);
1597 1598

    if (ro) {
1599 1600
        /* ignoring error return here */
        bdrv_reopen(bs->backing_hd, open_flags & ~BDRV_O_RDWR, NULL);
1601 1602
    }

1603
    return ret;
1604 1605
}

1606
int bdrv_commit_all(void)
1607 1608 1609 1610
{
    BlockDriverState *bs;

    QTAILQ_FOREACH(bs, &bdrv_states, list) {
1611 1612 1613 1614
        int ret = bdrv_commit(bs);
        if (ret < 0) {
            return ret;
        }
1615
    }
1616
    return 0;
1617 1618
}

S
Stefan Hajnoczi 已提交
1619 1620 1621 1622 1623 1624
struct BdrvTrackedRequest {
    BlockDriverState *bs;
    int64_t sector_num;
    int nb_sectors;
    bool is_write;
    QLIST_ENTRY(BdrvTrackedRequest) list;
1625
    Coroutine *co; /* owner, used for deadlock detection */
1626
    CoQueue wait_queue; /* coroutines blocked on this request */
S
Stefan Hajnoczi 已提交
1627 1628 1629 1630 1631 1632 1633 1634 1635 1636
};

/**
 * 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);
1637
    qemu_co_queue_restart_all(&req->wait_queue);
S
Stefan Hajnoczi 已提交
1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652
}

/**
 * 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,
1653
        .co = qemu_coroutine_self(),
S
Stefan Hajnoczi 已提交
1654 1655
    };

1656 1657
    qemu_co_queue_init(&req->wait_queue);

S
Stefan Hajnoczi 已提交
1658 1659 1660
    QLIST_INSERT_HEAD(&bs->tracked_requests, req, list);
}

S
Stefan Hajnoczi 已提交
1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681
/**
 * Round a region to cluster boundaries
 */
static void round_to_clusters(BlockDriverState *bs,
                              int64_t sector_num, int nb_sectors,
                              int64_t *cluster_sector_num,
                              int *cluster_nb_sectors)
{
    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);
    }
}

1682 1683
static bool tracked_request_overlaps(BdrvTrackedRequest *req,
                                     int64_t sector_num, int nb_sectors) {
S
Stefan Hajnoczi 已提交
1684 1685 1686 1687 1688 1689 1690 1691 1692
    /*        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;
1693 1694 1695 1696 1697 1698
}

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
1699 1700
    int64_t cluster_sector_num;
    int cluster_nb_sectors;
1701 1702
    bool retry;

S
Stefan Hajnoczi 已提交
1703 1704 1705 1706 1707 1708 1709 1710 1711
    /* 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.
     */
    round_to_clusters(bs, sector_num, nb_sectors,
                      &cluster_sector_num, &cluster_nb_sectors);

1712 1713 1714
    do {
        retry = false;
        QLIST_FOREACH(req, &bs->tracked_requests, list) {
S
Stefan Hajnoczi 已提交
1715 1716
            if (tracked_request_overlaps(req, cluster_sector_num,
                                         cluster_nb_sectors)) {
1717 1718 1719 1720 1721 1722
                /* 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);

1723 1724 1725 1726 1727 1728 1729 1730
                qemu_co_queue_wait(&req->wait_queue);
                retry = true;
                break;
            }
        }
    } while (retry);
}

K
Kevin Wolf 已提交
1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742
/*
 * 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;
1743
    int ret;
K
Kevin Wolf 已提交
1744

1745 1746 1747 1748 1749
    /* Backing file format doesn't make sense without a backing file */
    if (backing_fmt && !backing_file) {
        return -EINVAL;
    }

K
Kevin Wolf 已提交
1750
    if (drv->bdrv_change_backing_file != NULL) {
1751
        ret = drv->bdrv_change_backing_file(bs, backing_file, backing_fmt);
K
Kevin Wolf 已提交
1752
    } else {
1753
        ret = -ENOTSUP;
K
Kevin Wolf 已提交
1754
    }
1755 1756 1757 1758 1759 1760

    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
1761 1762
}

1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 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
/*
 * 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;
}


1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918
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
1919 1920 1921 1922
    if (offset < 0)
        return -EIO;

    if ((offset > len) || (len - offset < size))
1923 1924 1925 1926 1927 1928 1929 1930
        return -EIO;

    return 0;
}

static int bdrv_check_request(BlockDriverState *bs, int64_t sector_num,
                              int nb_sectors)
{
1931 1932
    return bdrv_check_byte_request(bs, sector_num * BDRV_SECTOR_SIZE,
                                   nb_sectors * BDRV_SECTOR_SIZE);
1933 1934
}

1935 1936 1937 1938 1939 1940 1941 1942 1943 1944
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
1945
{
1946
    RwCo *rwco = opaque;
B
bellard 已提交
1947

1948 1949
    if (!rwco->is_write) {
        rwco->ret = bdrv_co_do_readv(rwco->bs, rwco->sector_num,
1950
                                     rwco->nb_sectors, rwco->qiov, 0);
1951 1952
    } else {
        rwco->ret = bdrv_co_do_writev(rwco->bs, rwco->sector_num,
1953
                                      rwco->nb_sectors, rwco->qiov, 0);
1954 1955
    }
}
1956

1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976
/*
 * 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,
    };
1977

1978
    qemu_iovec_init_external(&qiov, &iov, 1);
1979

1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990
    /**
     * 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);
    }

1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002
    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
2003

2004 2005 2006 2007 2008
/* 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
2009 2010
}

2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024
/* 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;
}

2025 2026
#define BITS_PER_LONG  (sizeof(unsigned long) * 8)

2027
static void set_dirty_bitmap(BlockDriverState *bs, int64_t sector_num,
2028
                             int nb_sectors, int dirty)
2029 2030
{
    int64_t start, end;
2031
    unsigned long val, idx, bit;
2032

J
Jan Kiszka 已提交
2033
    start = sector_num / BDRV_SECTORS_PER_DIRTY_CHUNK;
2034
    end = (sector_num + nb_sectors - 1) / BDRV_SECTORS_PER_DIRTY_CHUNK;
2035 2036

    for (; start <= end; start++) {
2037 2038
        idx = start / BITS_PER_LONG;
        bit = start % BITS_PER_LONG;
2039 2040
        val = bs->dirty_bitmap[idx];
        if (dirty) {
2041
            if (!(val & (1UL << bit))) {
2042
                bs->dirty_count++;
2043
                val |= 1UL << bit;
2044
            }
2045
        } else {
2046
            if (val & (1UL << bit)) {
2047
                bs->dirty_count--;
2048
                val &= ~(1UL << bit);
2049
            }
2050 2051
        }
        bs->dirty_bitmap[idx] = val;
2052 2053 2054
    }
}

2055
/* Return < 0 if error. Important errors are:
B
bellard 已提交
2056 2057 2058 2059 2060
  -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
*/
2061
int bdrv_write(BlockDriverState *bs, int64_t sector_num,
B
bellard 已提交
2062 2063
               const uint8_t *buf, int nb_sectors)
{
2064
    return bdrv_rw_co(bs, sector_num, (uint8_t *)buf, nb_sectors, true);
B
bellard 已提交
2065 2066
}

2067 2068
int bdrv_pread(BlockDriverState *bs, int64_t offset,
               void *buf, int count1)
B
bellard 已提交
2069
{
J
Jan Kiszka 已提交
2070
    uint8_t tmp_buf[BDRV_SECTOR_SIZE];
B
bellard 已提交
2071 2072
    int len, nb_sectors, count;
    int64_t sector_num;
2073
    int ret;
B
bellard 已提交
2074 2075 2076

    count = count1;
    /* first read to align to sector start */
J
Jan Kiszka 已提交
2077
    len = (BDRV_SECTOR_SIZE - offset) & (BDRV_SECTOR_SIZE - 1);
B
bellard 已提交
2078 2079
    if (len > count)
        len = count;
J
Jan Kiszka 已提交
2080
    sector_num = offset >> BDRV_SECTOR_BITS;
B
bellard 已提交
2081
    if (len > 0) {
2082 2083
        if ((ret = bdrv_read(bs, sector_num, tmp_buf, 1)) < 0)
            return ret;
J
Jan Kiszka 已提交
2084
        memcpy(buf, tmp_buf + (offset & (BDRV_SECTOR_SIZE - 1)), len);
B
bellard 已提交
2085 2086 2087 2088 2089 2090 2091 2092
        count -= len;
        if (count == 0)
            return count1;
        sector_num++;
        buf += len;
    }

    /* read the sectors "in place" */
J
Jan Kiszka 已提交
2093
    nb_sectors = count >> BDRV_SECTOR_BITS;
B
bellard 已提交
2094
    if (nb_sectors > 0) {
2095 2096
        if ((ret = bdrv_read(bs, sector_num, buf, nb_sectors)) < 0)
            return ret;
B
bellard 已提交
2097
        sector_num += nb_sectors;
J
Jan Kiszka 已提交
2098
        len = nb_sectors << BDRV_SECTOR_BITS;
B
bellard 已提交
2099 2100 2101 2102 2103 2104
        buf += len;
        count -= len;
    }

    /* add data from the last sector */
    if (count > 0) {
2105 2106
        if ((ret = bdrv_read(bs, sector_num, tmp_buf, 1)) < 0)
            return ret;
B
bellard 已提交
2107 2108 2109 2110 2111
        memcpy(buf, tmp_buf, count);
    }
    return count1;
}

2112 2113
int bdrv_pwrite(BlockDriverState *bs, int64_t offset,
                const void *buf, int count1)
B
bellard 已提交
2114
{
J
Jan Kiszka 已提交
2115
    uint8_t tmp_buf[BDRV_SECTOR_SIZE];
B
bellard 已提交
2116 2117
    int len, nb_sectors, count;
    int64_t sector_num;
2118
    int ret;
B
bellard 已提交
2119 2120 2121

    count = count1;
    /* first write to align to sector start */
J
Jan Kiszka 已提交
2122
    len = (BDRV_SECTOR_SIZE - offset) & (BDRV_SECTOR_SIZE - 1);
B
bellard 已提交
2123 2124
    if (len > count)
        len = count;
J
Jan Kiszka 已提交
2125
    sector_num = offset >> BDRV_SECTOR_BITS;
B
bellard 已提交
2126
    if (len > 0) {
2127 2128
        if ((ret = bdrv_read(bs, sector_num, tmp_buf, 1)) < 0)
            return ret;
J
Jan Kiszka 已提交
2129
        memcpy(tmp_buf + (offset & (BDRV_SECTOR_SIZE - 1)), buf, len);
2130 2131
        if ((ret = bdrv_write(bs, sector_num, tmp_buf, 1)) < 0)
            return ret;
B
bellard 已提交
2132 2133 2134 2135 2136 2137 2138 2139
        count -= len;
        if (count == 0)
            return count1;
        sector_num++;
        buf += len;
    }

    /* write the sectors "in place" */
J
Jan Kiszka 已提交
2140
    nb_sectors = count >> BDRV_SECTOR_BITS;
B
bellard 已提交
2141
    if (nb_sectors > 0) {
2142 2143
        if ((ret = bdrv_write(bs, sector_num, buf, nb_sectors)) < 0)
            return ret;
B
bellard 已提交
2144
        sector_num += nb_sectors;
J
Jan Kiszka 已提交
2145
        len = nb_sectors << BDRV_SECTOR_BITS;
B
bellard 已提交
2146 2147 2148 2149 2150 2151
        buf += len;
        count -= len;
    }

    /* add data from the last sector */
    if (count > 0) {
2152 2153
        if ((ret = bdrv_read(bs, sector_num, tmp_buf, 1)) < 0)
            return ret;
B
bellard 已提交
2154
        memcpy(tmp_buf, buf, count);
2155 2156
        if ((ret = bdrv_write(bs, sector_num, tmp_buf, 1)) < 0)
            return ret;
B
bellard 已提交
2157 2158 2159 2160
    }
    return count1;
}

K
Kevin Wolf 已提交
2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176
/*
 * 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;
    }

2177 2178
    /* No flush needed for cache modes that already do it */
    if (bs->enable_write_cache) {
K
Kevin Wolf 已提交
2179 2180 2181 2182 2183 2184
        bdrv_flush(bs);
    }

    return 0;
}

2185
static int coroutine_fn bdrv_co_do_copy_on_readv(BlockDriverState *bs,
S
Stefan Hajnoczi 已提交
2186 2187 2188 2189 2190 2191 2192 2193 2194
        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;

2195
    BlockDriver *drv = bs->drv;
S
Stefan Hajnoczi 已提交
2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208
    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.
     */
    round_to_clusters(bs, sector_num, nb_sectors,
                      &cluster_sector_num, &cluster_nb_sectors);

2209 2210
    trace_bdrv_co_do_copy_on_readv(bs, sector_num, nb_sectors,
                                   cluster_sector_num, cluster_nb_sectors);
S
Stefan Hajnoczi 已提交
2211 2212 2213 2214 2215

    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);

2216 2217
    ret = drv->bdrv_co_readv(bs, cluster_sector_num, cluster_nb_sectors,
                             &bounce_qiov);
S
Stefan Hajnoczi 已提交
2218 2219 2220 2221
    if (ret < 0) {
        goto err;
    }

2222 2223
    if (drv->bdrv_co_write_zeroes &&
        buffer_is_zero(bounce_buffer, iov.iov_len)) {
K
Kevin Wolf 已提交
2224 2225
        ret = bdrv_co_do_write_zeroes(bs, cluster_sector_num,
                                      cluster_nb_sectors);
2226
    } else {
2227 2228 2229
        /* This does not change the data on the disk, it is not necessary
         * to flush even in cache=writethrough mode.
         */
2230
        ret = drv->bdrv_co_writev(bs, cluster_sector_num, cluster_nb_sectors,
S
Stefan Hajnoczi 已提交
2231
                                  &bounce_qiov);
2232 2233
    }

S
Stefan Hajnoczi 已提交
2234 2235 2236 2237 2238 2239 2240 2241 2242
    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;
2243 2244
    qemu_iovec_from_buf(qiov, 0, bounce_buffer + skip_bytes,
                        nb_sectors * BDRV_SECTOR_SIZE);
S
Stefan Hajnoczi 已提交
2245 2246 2247 2248 2249 2250

err:
    qemu_vfree(bounce_buffer);
    return ret;
}

2251 2252 2253 2254
/*
 * Handle a read request in coroutine context
 */
static int coroutine_fn bdrv_co_do_readv(BlockDriverState *bs,
2255 2256
    int64_t sector_num, int nb_sectors, QEMUIOVector *qiov,
    BdrvRequestFlags flags)
K
Kevin Wolf 已提交
2257 2258
{
    BlockDriver *drv = bs->drv;
S
Stefan Hajnoczi 已提交
2259 2260
    BdrvTrackedRequest req;
    int ret;
K
Kevin Wolf 已提交
2261 2262 2263 2264 2265 2266 2267 2268

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

Z
Zhi Yong Wu 已提交
2269 2270 2271 2272 2273
    /* throttling disk read I/O */
    if (bs->io_limits_enabled) {
        bdrv_io_limits_intercept(bs, false, nb_sectors);
    }

2274
    if (bs->copy_on_read) {
2275 2276 2277 2278 2279 2280 2281
        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) {
2282 2283 2284
        wait_for_overlapping_requests(bs, sector_num, nb_sectors);
    }

S
Stefan Hajnoczi 已提交
2285
    tracked_request_begin(&req, bs, sector_num, nb_sectors, false);
S
Stefan Hajnoczi 已提交
2286

2287
    if (flags & BDRV_REQ_COPY_ON_READ) {
S
Stefan Hajnoczi 已提交
2288 2289 2290 2291 2292 2293 2294 2295
        int pnum;

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

        if (!ret || pnum != nb_sectors) {
2296
            ret = bdrv_co_do_copy_on_readv(bs, sector_num, nb_sectors, qiov);
S
Stefan Hajnoczi 已提交
2297 2298 2299 2300
            goto out;
        }
    }

S
Stefan Hajnoczi 已提交
2301
    ret = drv->bdrv_co_readv(bs, sector_num, nb_sectors, qiov);
S
Stefan Hajnoczi 已提交
2302 2303

out:
S
Stefan Hajnoczi 已提交
2304
    tracked_request_end(&req);
2305 2306 2307 2308 2309

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

S
Stefan Hajnoczi 已提交
2310
    return ret;
K
Kevin Wolf 已提交
2311 2312
}

2313
int coroutine_fn bdrv_co_readv(BlockDriverState *bs, int64_t sector_num,
K
Kevin Wolf 已提交
2314 2315
    int nb_sectors, QEMUIOVector *qiov)
{
2316
    trace_bdrv_co_readv(bs, sector_num, nb_sectors);
K
Kevin Wolf 已提交
2317

2318 2319 2320 2321 2322 2323 2324 2325 2326 2327
    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);
2328 2329
}

2330 2331 2332 2333 2334 2335 2336 2337
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
2338 2339 2340
    /* TODO Emulate only part of misaligned requests instead of letting block
     * drivers return -ENOTSUP and emulate everything */

2341 2342
    /* First try the efficient write zeroes operation */
    if (drv->bdrv_co_write_zeroes) {
K
Kevin Wolf 已提交
2343 2344 2345 2346
        ret = drv->bdrv_co_write_zeroes(bs, sector_num, nb_sectors);
        if (ret != -ENOTSUP) {
            return ret;
        }
2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360
    }

    /* 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;
}

2361 2362 2363 2364
/*
 * Handle a write request in coroutine context
 */
static int coroutine_fn bdrv_co_do_writev(BlockDriverState *bs,
2365 2366
    int64_t sector_num, int nb_sectors, QEMUIOVector *qiov,
    BdrvRequestFlags flags)
2367 2368
{
    BlockDriver *drv = bs->drv;
S
Stefan Hajnoczi 已提交
2369
    BdrvTrackedRequest req;
2370
    int ret;
K
Kevin Wolf 已提交
2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381

    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
2382 2383 2384 2385 2386
    /* throttling disk write I/O */
    if (bs->io_limits_enabled) {
        bdrv_io_limits_intercept(bs, true, nb_sectors);
    }

2387
    if (bs->copy_on_read_in_flight) {
2388 2389 2390
        wait_for_overlapping_requests(bs, sector_num, nb_sectors);
    }

S
Stefan Hajnoczi 已提交
2391 2392
    tracked_request_begin(&req, bs, sector_num, nb_sectors, true);

2393 2394 2395 2396 2397
    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);
    }
2398

2399 2400 2401 2402
    if (ret == 0 && !bs->enable_write_cache) {
        ret = bdrv_co_flush(bs);
    }

K
Kevin Wolf 已提交
2403
    if (bs->dirty_bitmap) {
2404
        bdrv_set_dirty(bs, sector_num, nb_sectors);
K
Kevin Wolf 已提交
2405 2406 2407 2408 2409 2410
    }

    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
2411 2412
    tracked_request_end(&req);

2413
    return ret;
K
Kevin Wolf 已提交
2414 2415
}

2416 2417 2418 2419 2420
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);

2421 2422 2423 2424 2425 2426 2427 2428 2429 2430
    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);
2431 2432
}

B
bellard 已提交
2433 2434 2435 2436 2437 2438
/**
 * Truncate file to 'offset' bytes (needed only for file protocols)
 */
int bdrv_truncate(BlockDriverState *bs, int64_t offset)
{
    BlockDriver *drv = bs->drv;
2439
    int ret;
B
bellard 已提交
2440
    if (!drv)
B
bellard 已提交
2441
        return -ENOMEDIUM;
B
bellard 已提交
2442 2443
    if (!drv->bdrv_truncate)
        return -ENOTSUP;
2444 2445
    if (bs->read_only)
        return -EACCES;
M
Marcelo Tosatti 已提交
2446 2447
    if (bdrv_in_use(bs))
        return -EBUSY;
2448 2449 2450
    ret = drv->bdrv_truncate(bs, offset);
    if (ret == 0) {
        ret = refresh_total_sectors(bs, offset >> BDRV_SECTOR_BITS);
2451
        bdrv_dev_resize_cb(bs);
2452 2453
    }
    return ret;
B
bellard 已提交
2454 2455
}

2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474
/**
 * 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
2475 2476 2477 2478 2479 2480 2481
/**
 * 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
2482
        return -ENOMEDIUM;
2483

2484
    if (bs->growable || bdrv_dev_has_removable_media(bs)) {
2485 2486 2487
        if (drv->bdrv_getlength) {
            return drv->bdrv_getlength(bs);
        }
B
bellard 已提交
2488
    }
2489
    return bs->total_sectors * BDRV_SECTOR_SIZE;
B
bellard 已提交
2490 2491
}

B
bellard 已提交
2492
/* return 0 as number of sectors if no device present or error */
2493
void bdrv_get_geometry(BlockDriverState *bs, uint64_t *nb_sectors_ptr)
B
bellard 已提交
2494
{
B
bellard 已提交
2495 2496 2497 2498 2499
    int64_t length;
    length = bdrv_getlength(bs);
    if (length < 0)
        length = 0;
    else
J
Jan Kiszka 已提交
2500
        length = length >> BDRV_SECTOR_BITS;
B
bellard 已提交
2501
    *nb_sectors_ptr = length;
B
bellard 已提交
2502
}
B
bellard 已提交
2503

2504 2505 2506 2507 2508 2509 2510 2511
/* 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);
}

2512 2513
void bdrv_set_on_error(BlockDriverState *bs, BlockdevOnError on_read_error,
                       BlockdevOnError on_write_error)
2514 2515 2516 2517 2518
{
    bs->on_read_error = on_read_error;
    bs->on_write_error = on_write_error;
}

2519
BlockdevOnError bdrv_get_on_error(BlockDriverState *bs, bool is_read)
2520 2521 2522 2523
{
    return is_read ? bs->on_read_error : bs->on_write_error;
}

2524 2525 2526 2527 2528 2529 2530 2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549
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
2550
    bdrv_emit_qmp_error_event(bs, QEVENT_BLOCK_IO_ERROR, action, is_read);
2551 2552 2553 2554 2555 2556
    if (action == BDRV_ACTION_STOP) {
        vm_stop(RUN_STATE_IO_ERROR);
        bdrv_iostatus_set_err(bs, error);
    }
}

B
bellard 已提交
2557 2558 2559 2560 2561
int bdrv_is_read_only(BlockDriverState *bs)
{
    return bs->read_only;
}

2562 2563 2564 2565 2566
int bdrv_is_sg(BlockDriverState *bs)
{
    return bs->sg;
}

2567 2568 2569 2570 2571
int bdrv_enable_write_cache(BlockDriverState *bs)
{
    return bs->enable_write_cache;
}

2572 2573 2574
void bdrv_set_enable_write_cache(BlockDriverState *bs, bool wce)
{
    bs->enable_write_cache = wce;
2575 2576 2577 2578 2579 2580 2581

    /* so a reopen() will preserve wce */
    if (wce) {
        bs->open_flags |= BDRV_O_CACHE_WB;
    } else {
        bs->open_flags &= ~BDRV_O_CACHE_WB;
    }
2582 2583
}

B
bellard 已提交
2584 2585 2586 2587 2588 2589 2590
int bdrv_is_encrypted(BlockDriverState *bs)
{
    if (bs->backing_hd && bs->backing_hd->encrypted)
        return 1;
    return bs->encrypted;
}

2591 2592 2593 2594 2595 2596 2597 2598 2599
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
2600 2601 2602 2603 2604 2605 2606 2607 2608 2609
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;
    }
2610 2611 2612 2613 2614
    if (!bs->encrypted) {
        return -EINVAL;
    } else if (!bs->drv || !bs->drv->bdrv_set_key) {
        return -ENOMEDIUM;
    }
2615
    ret = bs->drv->bdrv_set_key(bs, key);
2616 2617 2618 2619 2620
    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 */
2621
        bdrv_dev_change_media_cb(bs, true);
2622
    }
2623
    return ret;
B
bellard 已提交
2624 2625
}

2626
const char *bdrv_get_format_name(BlockDriverState *bs)
B
bellard 已提交
2627
{
2628
    return bs->drv ? bs->drv->format_name : NULL;
B
bellard 已提交
2629 2630
}

2631
void bdrv_iterate_format(void (*it)(void *opaque, const char *name),
B
bellard 已提交
2632 2633 2634 2635
                         void *opaque)
{
    BlockDriver *drv;

2636
    QLIST_FOREACH(drv, &bdrv_drivers, list) {
B
bellard 已提交
2637 2638 2639 2640
        it(opaque, drv->format_name);
    }
}

B
bellard 已提交
2641 2642 2643 2644
BlockDriverState *bdrv_find(const char *name)
{
    BlockDriverState *bs;

2645 2646
    QTAILQ_FOREACH(bs, &bdrv_states, list) {
        if (!strcmp(name, bs->device_name)) {
B
bellard 已提交
2647
            return bs;
2648
        }
B
bellard 已提交
2649 2650 2651 2652
    }
    return NULL;
}

M
Markus Armbruster 已提交
2653 2654 2655 2656 2657 2658 2659 2660
BlockDriverState *bdrv_next(BlockDriverState *bs)
{
    if (!bs) {
        return QTAILQ_FIRST(&bdrv_states);
    }
    return QTAILQ_NEXT(bs, list);
}

2661
void bdrv_iterate(void (*it)(void *opaque, BlockDriverState *bs), void *opaque)
B
bellard 已提交
2662 2663 2664
{
    BlockDriverState *bs;

2665
    QTAILQ_FOREACH(bs, &bdrv_states, list) {
2666
        it(opaque, bs);
B
bellard 已提交
2667 2668 2669
    }
}

B
bellard 已提交
2670 2671 2672 2673 2674
const char *bdrv_get_device_name(BlockDriverState *bs)
{
    return bs->device_name;
}

M
Markus Armbruster 已提交
2675 2676 2677 2678 2679
int bdrv_get_flags(BlockDriverState *bs)
{
    return bs->open_flags;
}

A
aliguori 已提交
2680 2681 2682 2683
void bdrv_flush_all(void)
{
    BlockDriverState *bs;

2684
    QTAILQ_FOREACH(bs, &bdrv_states, list) {
2685
        bdrv_flush(bs);
2686
    }
A
aliguori 已提交
2687 2688
}

K
Kevin Wolf 已提交
2689 2690 2691 2692
int bdrv_has_zero_init(BlockDriverState *bs)
{
    assert(bs->drv);

K
Kevin Wolf 已提交
2693 2694
    if (bs->drv->bdrv_has_zero_init) {
        return bs->drv->bdrv_has_zero_init(bs);
K
Kevin Wolf 已提交
2695 2696 2697 2698 2699
    }

    return 1;
}

2700 2701 2702 2703 2704 2705 2706 2707 2708
typedef struct BdrvCoIsAllocatedData {
    BlockDriverState *bs;
    int64_t sector_num;
    int nb_sectors;
    int *pnum;
    int ret;
    bool done;
} BdrvCoIsAllocatedData;

2709 2710 2711 2712 2713
/*
 * 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.
 *
2714 2715 2716
 * If 'sector_num' is beyond the end of the disk image the return value is 0
 * and 'pnum' is set to 0.
 *
2717 2718 2719 2720
 * '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.
 *
2721 2722
 * '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.
2723
 */
2724 2725
int coroutine_fn bdrv_co_is_allocated(BlockDriverState *bs, int64_t sector_num,
                                      int nb_sectors, int *pnum)
2726
{
2727 2728 2729 2730 2731 2732 2733 2734 2735 2736 2737 2738
    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;
    }

2739
    if (!bs->drv->bdrv_co_is_allocated) {
2740
        *pnum = nb_sectors;
2741 2742
        return 1;
    }
2743

2744 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765
    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)
{
2766 2767 2768 2769 2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780
    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;
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 2814 2815 2816 2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827 2828 2829 2830 2831
/*
 * 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
2832
BlockInfo *bdrv_query_info(BlockDriverState *bs)
B
bellard 已提交
2833
{
P
Paolo Bonzini 已提交
2834 2835 2836 2837 2838
    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
2839

P
Paolo Bonzini 已提交
2840 2841 2842 2843
    if (bdrv_dev_has_removable_media(bs)) {
        info->has_tray_open = true;
        info->tray_open = bdrv_dev_is_tray_open(bs);
    }
2844

P
Paolo Bonzini 已提交
2845 2846 2847 2848
    if (bdrv_iostatus_is_enabled(bs)) {
        info->has_io_status = true;
        info->io_status = bs->iostatus;
    }
2849

2850 2851 2852 2853 2854 2855
    if (bs->dirty_bitmap) {
        info->has_dirty = true;
        info->dirty = g_malloc0(sizeof(*info->dirty));
        info->dirty->count = bdrv_get_dirty_count(bs) *
            BDRV_SECTORS_PER_DIRTY_CHUNK * BDRV_SECTOR_SIZE;
    }
2856

P
Paolo Bonzini 已提交
2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 2867 2868
    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);
2869
        }
2870

P
Paolo Bonzini 已提交
2871 2872 2873 2874 2875 2876 2877 2878 2879 2880 2881 2882 2883 2884 2885
        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];
2886
        }
P
Paolo Bonzini 已提交
2887 2888 2889
    }
    return info;
}
2890

P
Paolo Bonzini 已提交
2891 2892 2893 2894
BlockInfoList *qmp_query_block(Error **errp)
{
    BlockInfoList *head = NULL, **p_next = &head;
    BlockDriverState *bs;
Z
Zhi Yong Wu 已提交
2895

P
Paolo Bonzini 已提交
2896 2897 2898
    QTAILQ_FOREACH(bs, &bdrv_states, list) {
        BlockInfoList *info = g_malloc0(sizeof(*info));
        info->value = bdrv_query_info(bs);
2899

P
Paolo Bonzini 已提交
2900 2901
        *p_next = info;
        p_next = &info->next;
B
bellard 已提交
2902
    }
2903

L
Luiz Capitulino 已提交
2904
    return head;
B
bellard 已提交
2905
}
2906

P
Paolo Bonzini 已提交
2907
BlockStats *bdrv_query_stats(const BlockDriverState *bs)
L
Luiz Capitulino 已提交
2908 2909 2910 2911 2912 2913 2914 2915
{
    BlockStats *s;

    s = g_malloc0(sizeof(*s));

    if (bs->device_name[0]) {
        s->has_device = true;
        s->device = g_strdup(bs->device_name);
2916 2917
    }

L
Luiz Capitulino 已提交
2918 2919 2920 2921 2922 2923 2924 2925 2926 2927 2928
    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];

2929
    if (bs->file) {
L
Luiz Capitulino 已提交
2930
        s->has_parent = true;
P
Paolo Bonzini 已提交
2931
        s->parent = bdrv_query_stats(bs->file);
2932 2933
    }

L
Luiz Capitulino 已提交
2934
    return s;
2935 2936
}

L
Luiz Capitulino 已提交
2937
BlockStatsList *qmp_query_blockstats(Error **errp)
2938
{
P
Paolo Bonzini 已提交
2939
    BlockStatsList *head = NULL, **p_next = &head;
2940 2941
    BlockDriverState *bs;

2942
    QTAILQ_FOREACH(bs, &bdrv_states, list) {
L
Luiz Capitulino 已提交
2943
        BlockStatsList *info = g_malloc0(sizeof(*info));
P
Paolo Bonzini 已提交
2944
        info->value = bdrv_query_stats(bs);
L
Luiz Capitulino 已提交
2945

P
Paolo Bonzini 已提交
2946 2947
        *p_next = info;
        p_next = &info->next;
2948
    }
2949

L
Luiz Capitulino 已提交
2950
    return head;
2951
}
B
bellard 已提交
2952

2953 2954 2955 2956 2957 2958 2959 2960 2961 2962
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;
}

2963
void bdrv_get_backing_filename(BlockDriverState *bs,
B
bellard 已提交
2964 2965
                               char *filename, int filename_size)
{
K
Kevin Wolf 已提交
2966
    pstrcpy(filename, filename_size, bs->backing_file);
B
bellard 已提交
2967 2968
}

2969
int bdrv_write_compressed(BlockDriverState *bs, int64_t sector_num,
B
bellard 已提交
2970 2971 2972 2973
                          const uint8_t *buf, int nb_sectors)
{
    BlockDriver *drv = bs->drv;
    if (!drv)
B
bellard 已提交
2974
        return -ENOMEDIUM;
B
bellard 已提交
2975 2976
    if (!drv->bdrv_write_compressed)
        return -ENOTSUP;
K
Kevin Wolf 已提交
2977 2978
    if (bdrv_check_request(bs, sector_num, nb_sectors))
        return -EIO;
2979

2980
    assert(!bs->dirty_bitmap);
2981

B
bellard 已提交
2982 2983
    return drv->bdrv_write_compressed(bs, sector_num, buf, nb_sectors);
}
2984

B
bellard 已提交
2985 2986 2987 2988
int bdrv_get_info(BlockDriverState *bs, BlockDriverInfo *bdi)
{
    BlockDriver *drv = bs->drv;
    if (!drv)
B
bellard 已提交
2989
        return -ENOMEDIUM;
B
bellard 已提交
2990 2991 2992 2993 2994 2995
    if (!drv->bdrv_get_info)
        return -ENOTSUP;
    memset(bdi, 0, sizeof(*bdi));
    return drv->bdrv_get_info(bs, bdi);
}

2996 2997
int bdrv_save_vmstate(BlockDriverState *bs, const uint8_t *buf,
                      int64_t pos, int size)
2998 2999 3000 3001
{
    BlockDriver *drv = bs->drv;
    if (!drv)
        return -ENOMEDIUM;
3002 3003 3004 3005 3006
    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;
3007 3008
}

3009 3010
int bdrv_load_vmstate(BlockDriverState *bs, uint8_t *buf,
                      int64_t pos, int size)
3011 3012 3013 3014
{
    BlockDriver *drv = bs->drv;
    if (!drv)
        return -ENOMEDIUM;
3015 3016 3017 3018 3019
    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;
3020 3021
}

K
Kevin Wolf 已提交
3022 3023 3024 3025 3026 3027 3028 3029
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
3030
    drv->bdrv_debug_event(bs, event);
K
Kevin Wolf 已提交
3031 3032 3033

}

B
bellard 已提交
3034 3035 3036
/**************************************************************/
/* handling of snapshots */

3037 3038 3039
int bdrv_can_snapshot(BlockDriverState *bs)
{
    BlockDriver *drv = bs->drv;
3040
    if (!drv || !bdrv_is_inserted(bs) || bdrv_is_read_only(bs)) {
3041 3042 3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053
        return 0;
    }

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

    return 1;
}

3054 3055 3056 3057 3058
int bdrv_is_snapshot(BlockDriverState *bs)
{
    return !!(bs->open_flags & BDRV_O_SNAPSHOT);
}

3059 3060 3061 3062
BlockDriverState *bdrv_snapshots(void)
{
    BlockDriverState *bs;

3063
    if (bs_snapshots) {
3064
        return bs_snapshots;
3065
    }
3066 3067 3068 3069

    bs = NULL;
    while ((bs = bdrv_next(bs))) {
        if (bdrv_can_snapshot(bs)) {
3070 3071
            bs_snapshots = bs;
            return bs;
3072 3073 3074 3075 3076
        }
    }
    return NULL;
}

3077
int bdrv_snapshot_create(BlockDriverState *bs,
B
bellard 已提交
3078 3079 3080 3081
                         QEMUSnapshotInfo *sn_info)
{
    BlockDriver *drv = bs->drv;
    if (!drv)
B
bellard 已提交
3082
        return -ENOMEDIUM;
3083 3084 3085 3086 3087
    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
3088 3089
}

3090
int bdrv_snapshot_goto(BlockDriverState *bs,
B
bellard 已提交
3091 3092 3093
                       const char *snapshot_id)
{
    BlockDriver *drv = bs->drv;
3094 3095
    int ret, open_ret;

B
bellard 已提交
3096
    if (!drv)
B
bellard 已提交
3097
        return -ENOMEDIUM;
3098 3099 3100 3101 3102 3103 3104 3105 3106 3107 3108 3109 3110 3111 3112 3113
    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
3114 3115 3116 3117 3118 3119
}

int bdrv_snapshot_delete(BlockDriverState *bs, const char *snapshot_id)
{
    BlockDriver *drv = bs->drv;
    if (!drv)
B
bellard 已提交
3120
        return -ENOMEDIUM;
3121 3122 3123 3124 3125
    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
3126 3127
}

3128
int bdrv_snapshot_list(BlockDriverState *bs,
B
bellard 已提交
3129 3130 3131 3132
                       QEMUSnapshotInfo **psn_info)
{
    BlockDriver *drv = bs->drv;
    if (!drv)
B
bellard 已提交
3133
        return -ENOMEDIUM;
3134 3135 3136 3137 3138
    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
3139 3140
}

E
edison 已提交
3141 3142 3143 3144 3145 3146 3147 3148 3149 3150 3151 3152 3153 3154 3155 3156
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;
}

3157 3158 3159 3160
/* 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. */
3161 3162 3163
BlockDriverState *bdrv_find_backing_image(BlockDriverState *bs,
        const char *backing_file)
{
3164 3165 3166 3167 3168 3169 3170 3171
    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) {
3172 3173 3174
        return NULL;
    }

3175 3176 3177 3178 3179 3180 3181 3182 3183 3184 3185 3186 3187 3188 3189
    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;
            }
3190
        } else {
3191 3192 3193 3194 3195 3196 3197 3198 3199 3200 3201 3202 3203 3204 3205 3206 3207 3208 3209 3210 3211 3212 3213
            /* 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;
            }
3214 3215 3216
        }
    }

3217 3218 3219 3220
    g_free(filename_full);
    g_free(backing_file_full);
    g_free(filename_tmp);
    return retval;
3221 3222
}

3223 3224 3225 3226 3227 3228 3229 3230 3231 3232 3233 3234 3235
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);
}

3236 3237 3238 3239 3240 3241 3242 3243 3244 3245 3246 3247 3248 3249 3250 3251
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
3252 3253 3254 3255 3256 3257 3258 3259 3260 3261 3262 3263 3264 3265
#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)) {
3266
                snprintf(buf, buf_size, "%0.1f%c",
B
bellard 已提交
3267 3268 3269 3270
                         (double)size / base,
                         suffixes[i]);
                break;
            } else if (size < (1000 * base) || i == (NB_SUFFIXES - 1)) {
3271
                snprintf(buf, buf_size, "%" PRId64 "%c",
B
bellard 已提交
3272 3273 3274 3275 3276 3277 3278 3279 3280 3281 3282 3283 3284
                         ((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];
3285 3286 3287
#ifdef _WIN32
    struct tm *ptm;
#else
B
bellard 已提交
3288
    struct tm tm;
3289
#endif
B
bellard 已提交
3290 3291 3292 3293
    time_t ti;
    int64_t secs;

    if (!sn) {
3294 3295
        snprintf(buf, buf_size,
                 "%-10s%-20s%7s%20s%15s",
B
bellard 已提交
3296 3297 3298
                 "ID", "TAG", "VM SIZE", "DATE", "VM CLOCK");
    } else {
        ti = sn->date_sec;
3299 3300 3301 3302 3303
#ifdef _WIN32
        ptm = localtime(&ti);
        strftime(date_buf, sizeof(date_buf),
                 "%Y-%m-%d %H:%M:%S", ptm);
#else
B
bellard 已提交
3304 3305 3306
        localtime_r(&ti, &tm);
        strftime(date_buf, sizeof(date_buf),
                 "%Y-%m-%d %H:%M:%S", &tm);
3307
#endif
B
bellard 已提交
3308 3309 3310 3311 3312
        secs = sn->vm_clock_nsec / 1000000000;
        snprintf(clock_buf, sizeof(clock_buf),
                 "%02d:%02d:%02d.%03d",
                 (int)(secs / 3600),
                 (int)((secs / 60) % 60),
3313
                 (int)(secs % 60),
B
bellard 已提交
3314 3315
                 (int)((sn->vm_clock_nsec / 1000000) % 1000));
        snprintf(buf, buf_size,
3316
                 "%-10s%-20s%7s%20s%15s",
B
bellard 已提交
3317 3318 3319 3320 3321 3322 3323 3324
                 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
3325
/**************************************************************/
B
bellard 已提交
3326
/* async I/Os */
B
bellard 已提交
3327

3328
BlockDriverAIOCB *bdrv_aio_readv(BlockDriverState *bs, int64_t sector_num,
3329
                                 QEMUIOVector *qiov, int nb_sectors,
3330
                                 BlockDriverCompletionFunc *cb, void *opaque)
B
bellard 已提交
3331
{
3332 3333
    trace_bdrv_aio_readv(bs, sector_num, nb_sectors, opaque);

3334
    return bdrv_co_aio_rw_vector(bs, sector_num, qiov, nb_sectors,
3335
                                 cb, opaque, false);
B
bellard 已提交
3336 3337
}

3338 3339 3340
BlockDriverAIOCB *bdrv_aio_writev(BlockDriverState *bs, int64_t sector_num,
                                  QEMUIOVector *qiov, int nb_sectors,
                                  BlockDriverCompletionFunc *cb, void *opaque)
B
bellard 已提交
3341
{
3342 3343
    trace_bdrv_aio_writev(bs, sector_num, nb_sectors, opaque);

3344
    return bdrv_co_aio_rw_vector(bs, sector_num, qiov, nb_sectors,
3345
                                 cb, opaque, true);
B
bellard 已提交
3346 3347
}

K
Kevin Wolf 已提交
3348 3349 3350 3351 3352 3353 3354 3355 3356 3357 3358 3359 3360 3361 3362 3363 3364 3365

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);
3366 3367 3368
        if (mcb->callbacks[i].free_qiov) {
            qemu_iovec_destroy(mcb->callbacks[i].free_qiov);
        }
3369
        g_free(mcb->callbacks[i].free_qiov);
K
Kevin Wolf 已提交
3370 3371 3372 3373 3374 3375 3376
    }
}

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

3377 3378
    trace_multiwrite_cb(mcb, ret);

3379
    if (ret < 0 && !mcb->error) {
K
Kevin Wolf 已提交
3380 3381 3382 3383 3384
        mcb->error = ret;
    }

    mcb->num_requests--;
    if (mcb->num_requests == 0) {
3385
        multiwrite_user_cb(mcb);
3386
        g_free(mcb);
K
Kevin Wolf 已提交
3387 3388 3389 3390 3391
    }
}

static int multiwrite_req_compare(const void *a, const void *b)
{
3392 3393 3394 3395 3396 3397 3398 3399 3400 3401 3402 3403 3404
    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
3405 3406 3407 3408 3409 3410 3411 3412 3413 3414 3415 3416 3417 3418 3419 3420 3421 3422 3423 3424 3425
}

/*
 * 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;

3426
        // Handle exactly sequential writes and overlapping writes.
K
Kevin Wolf 已提交
3427 3428 3429 3430
        if (reqs[i].sector <= oldreq_last) {
            merge = 1;
        }

3431 3432 3433 3434
        if (reqs[outidx].qiov->niov + reqs[i].qiov->niov + 1 > IOV_MAX) {
            merge = 0;
        }

K
Kevin Wolf 已提交
3435 3436
        if (merge) {
            size_t size;
3437
            QEMUIOVector *qiov = g_malloc0(sizeof(*qiov));
K
Kevin Wolf 已提交
3438 3439 3440 3441 3442 3443
            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;
3444
            qemu_iovec_concat(qiov, reqs[outidx].qiov, 0, size);
K
Kevin Wolf 已提交
3445

3446 3447
            // We should need to add any zeros between the two requests
            assert (reqs[i].sector <= oldreq_last);
K
Kevin Wolf 已提交
3448 3449

            // Add the second request
3450
            qemu_iovec_concat(qiov, reqs[i].qiov, 0, reqs[i].qiov->size);
K
Kevin Wolf 已提交
3451

3452
            reqs[outidx].nb_sectors = qiov->size >> 9;
K
Kevin Wolf 已提交
3453 3454 3455 3456 3457 3458 3459 3460 3461 3462 3463 3464 3465 3466 3467 3468 3469 3470 3471 3472 3473 3474 3475 3476 3477 3478 3479 3480 3481 3482 3483 3484 3485
            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;

3486 3487 3488 3489 3490 3491 3492 3493
    /* 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
3494 3495 3496 3497 3498
    if (num_reqs == 0) {
        return 0;
    }

    // Create MultiwriteCB structure
3499
    mcb = g_malloc0(sizeof(*mcb) + num_reqs * sizeof(*mcb->callbacks));
K
Kevin Wolf 已提交
3500 3501 3502 3503 3504 3505 3506 3507 3508 3509 3510
    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);

3511 3512
    trace_bdrv_aio_multiwrite(mcb, mcb->num_callbacks, num_reqs);

3513 3514
    /* Run the aio requests. */
    mcb->num_requests = num_reqs;
K
Kevin Wolf 已提交
3515
    for (i = 0; i < num_reqs; i++) {
3516
        bdrv_aio_writev(bs, reqs[i].sector, reqs[i].qiov,
K
Kevin Wolf 已提交
3517 3518 3519 3520 3521 3522
            reqs[i].nb_sectors, multiwrite_cb, mcb);
    }

    return 0;
}

B
bellard 已提交
3523 3524
void bdrv_aio_cancel(BlockDriverAIOCB *acb)
{
S
Stefan Hajnoczi 已提交
3525
    acb->aiocb_info->cancel(acb);
B
bellard 已提交
3526 3527
}

Z
Zhi Yong Wu 已提交
3528 3529 3530 3531 3532 3533 3534 3535 3536 3537 3538 3539 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
/* 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;
}
3692

B
bellard 已提交
3693 3694 3695
/**************************************************************/
/* async block device emulation */

3696 3697 3698 3699 3700 3701 3702 3703 3704 3705 3706 3707
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)
{
3708 3709
    BlockDriverAIOCBSync *acb =
        container_of(blockacb, BlockDriverAIOCBSync, common);
3710
    qemu_bh_delete(acb->bh);
A
Avi Kivity 已提交
3711
    acb->bh = NULL;
3712 3713 3714
    qemu_aio_release(acb);
}

S
Stefan Hajnoczi 已提交
3715
static const AIOCBInfo bdrv_em_aiocb_info = {
3716 3717 3718 3719
    .aiocb_size         = sizeof(BlockDriverAIOCBSync),
    .cancel             = bdrv_aio_cancel_em,
};

3720
static void bdrv_aio_bh_cb(void *opaque)
B
bellard 已提交
3721
{
3722
    BlockDriverAIOCBSync *acb = opaque;
3723 3724

    if (!acb->is_write)
3725
        qemu_iovec_from_buf(acb->qiov, 0, acb->bounce, acb->qiov->size);
3726
    qemu_vfree(acb->bounce);
3727
    acb->common.cb(acb->common.opaque, acb->ret);
3728
    qemu_bh_delete(acb->bh);
A
Avi Kivity 已提交
3729
    acb->bh = NULL;
3730
    qemu_aio_release(acb);
B
bellard 已提交
3731
}
3732

3733 3734 3735 3736 3737 3738 3739 3740
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
3741
{
3742 3743
    BlockDriverAIOCBSync *acb;

S
Stefan Hajnoczi 已提交
3744
    acb = qemu_aio_get(&bdrv_em_aiocb_info, bs, cb, opaque);
3745 3746
    acb->is_write = is_write;
    acb->qiov = qiov;
3747
    acb->bounce = qemu_blockalign(bs, qiov->size);
3748
    acb->bh = qemu_bh_new(bdrv_aio_bh_cb, acb);
3749 3750

    if (is_write) {
3751
        qemu_iovec_to_buf(acb->qiov, 0, acb->bounce, qiov->size);
3752
        acb->ret = bs->drv->bdrv_write(bs, sector_num, acb->bounce, nb_sectors);
3753
    } else {
3754
        acb->ret = bs->drv->bdrv_read(bs, sector_num, acb->bounce, nb_sectors);
3755 3756
    }

3757
    qemu_bh_schedule(acb->bh);
3758

3759
    return &acb->common;
3760 3761
}

3762 3763
static BlockDriverAIOCB *bdrv_aio_readv_em(BlockDriverState *bs,
        int64_t sector_num, QEMUIOVector *qiov, int nb_sectors,
3764
        BlockDriverCompletionFunc *cb, void *opaque)
3765
{
3766 3767
    return bdrv_aio_rw_vector(bs, sector_num, qiov, nb_sectors, cb, opaque, 0);
}
B
bellard 已提交
3768

3769 3770 3771 3772 3773
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);
3774 3775
}

3776 3777 3778 3779 3780 3781 3782 3783 3784 3785 3786 3787 3788

typedef struct BlockDriverAIOCBCoroutine {
    BlockDriverAIOCB common;
    BlockRequest req;
    bool is_write;
    QEMUBH* bh;
} BlockDriverAIOCBCoroutine;

static void bdrv_aio_co_cancel_em(BlockDriverAIOCB *blockacb)
{
    qemu_aio_flush();
}

S
Stefan Hajnoczi 已提交
3789
static const AIOCBInfo bdrv_em_co_aiocb_info = {
3790 3791 3792 3793
    .aiocb_size         = sizeof(BlockDriverAIOCBCoroutine),
    .cancel             = bdrv_aio_co_cancel_em,
};

P
Paolo Bonzini 已提交
3794
static void bdrv_co_em_bh(void *opaque)
3795 3796 3797 3798 3799 3800 3801 3802
{
    BlockDriverAIOCBCoroutine *acb = opaque;

    acb->common.cb(acb->common.opaque, acb->req.error);
    qemu_bh_delete(acb->bh);
    qemu_aio_release(acb);
}

3803 3804 3805 3806 3807 3808 3809 3810
/* 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,
3811
            acb->req.nb_sectors, acb->req.qiov, 0);
3812 3813
    } else {
        acb->req.error = bdrv_co_do_writev(bs, acb->req.sector,
3814
            acb->req.nb_sectors, acb->req.qiov, 0);
3815 3816
    }

P
Paolo Bonzini 已提交
3817
    acb->bh = qemu_bh_new(bdrv_co_em_bh, acb);
3818 3819 3820
    qemu_bh_schedule(acb->bh);
}

3821 3822 3823 3824 3825 3826
static BlockDriverAIOCB *bdrv_co_aio_rw_vector(BlockDriverState *bs,
                                               int64_t sector_num,
                                               QEMUIOVector *qiov,
                                               int nb_sectors,
                                               BlockDriverCompletionFunc *cb,
                                               void *opaque,
3827
                                               bool is_write)
3828 3829 3830 3831
{
    Coroutine *co;
    BlockDriverAIOCBCoroutine *acb;

S
Stefan Hajnoczi 已提交
3832
    acb = qemu_aio_get(&bdrv_em_co_aiocb_info, bs, cb, opaque);
3833 3834 3835 3836 3837
    acb->req.sector = sector_num;
    acb->req.nb_sectors = nb_sectors;
    acb->req.qiov = qiov;
    acb->is_write = is_write;

3838
    co = qemu_coroutine_create(bdrv_co_do_rw);
3839 3840 3841 3842 3843
    qemu_coroutine_enter(co, acb);

    return &acb->common;
}

P
Paolo Bonzini 已提交
3844
static void coroutine_fn bdrv_aio_flush_co_entry(void *opaque)
3845
{
P
Paolo Bonzini 已提交
3846 3847
    BlockDriverAIOCBCoroutine *acb = opaque;
    BlockDriverState *bs = acb->common.bs;
3848

P
Paolo Bonzini 已提交
3849 3850
    acb->req.error = bdrv_co_flush(bs);
    acb->bh = qemu_bh_new(bdrv_co_em_bh, acb);
3851 3852 3853
    qemu_bh_schedule(acb->bh);
}

P
Paolo Bonzini 已提交
3854
BlockDriverAIOCB *bdrv_aio_flush(BlockDriverState *bs,
3855 3856
        BlockDriverCompletionFunc *cb, void *opaque)
{
P
Paolo Bonzini 已提交
3857
    trace_bdrv_aio_flush(bs, opaque);
3858

P
Paolo Bonzini 已提交
3859 3860
    Coroutine *co;
    BlockDriverAIOCBCoroutine *acb;
3861

S
Stefan Hajnoczi 已提交
3862
    acb = qemu_aio_get(&bdrv_em_co_aiocb_info, bs, cb, opaque);
P
Paolo Bonzini 已提交
3863 3864
    co = qemu_coroutine_create(bdrv_aio_flush_co_entry);
    qemu_coroutine_enter(co, acb);
3865 3866 3867 3868

    return &acb->common;
}

3869 3870 3871 3872 3873 3874 3875 3876 3877 3878 3879 3880 3881 3882 3883 3884 3885 3886 3887
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
3888
    acb = qemu_aio_get(&bdrv_em_co_aiocb_info, bs, cb, opaque);
3889 3890 3891 3892 3893 3894 3895 3896
    acb->req.sector = sector_num;
    acb->req.nb_sectors = nb_sectors;
    co = qemu_coroutine_create(bdrv_aio_discard_co_entry);
    qemu_coroutine_enter(co, acb);

    return &acb->common;
}

B
bellard 已提交
3897 3898
void bdrv_init(void)
{
3899
    module_call_init(MODULE_INIT_BLOCK);
B
bellard 已提交
3900
}
3901

3902 3903 3904 3905 3906 3907
void bdrv_init_with_whitelist(void)
{
    use_bdrv_whitelist = 1;
    bdrv_init();
}

S
Stefan Hajnoczi 已提交
3908
void *qemu_aio_get(const AIOCBInfo *aiocb_info, BlockDriverState *bs,
3909
                   BlockDriverCompletionFunc *cb, void *opaque)
3910 3911 3912
{
    BlockDriverAIOCB *acb;

S
Stefan Hajnoczi 已提交
3913 3914
    acb = g_slice_alloc(aiocb_info->aiocb_size);
    acb->aiocb_info = aiocb_info;
3915 3916 3917 3918 3919 3920 3921 3922
    acb->bs = bs;
    acb->cb = cb;
    acb->opaque = opaque;
    return acb;
}

void qemu_aio_release(void *p)
{
3923
    BlockDriverAIOCB *acb = p;
S
Stefan Hajnoczi 已提交
3924
    g_slice_free1(acb->aiocb_info->aiocb_size, acb);
3925
}
B
bellard 已提交
3926

3927 3928 3929 3930 3931 3932 3933 3934 3935 3936 3937 3938 3939 3940 3941 3942 3943 3944 3945 3946 3947 3948 3949 3950 3951 3952
/**************************************************************/
/* 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) {
3953 3954
        acb = bs->drv->bdrv_aio_writev(bs, sector_num, iov, nb_sectors,
                                       bdrv_co_io_em_complete, &co);
3955
    } else {
3956 3957
        acb = bs->drv->bdrv_aio_readv(bs, sector_num, iov, nb_sectors,
                                      bdrv_co_io_em_complete, &co);
3958 3959
    }

3960
    trace_bdrv_co_io_em(bs, sector_num, nb_sectors, is_write, acb);
3961 3962 3963 3964 3965 3966 3967 3968 3969 3970 3971 3972 3973 3974 3975 3976 3977 3978 3979 3980 3981 3982
    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
3983
static void coroutine_fn bdrv_flush_co_entry(void *opaque)
3984
{
P
Paolo Bonzini 已提交
3985 3986 3987 3988 3989 3990 3991
    RwCo *rwco = opaque;

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

int coroutine_fn bdrv_co_flush(BlockDriverState *bs)
{
K
Kevin Wolf 已提交
3992 3993
    int ret;

3994
    if (!bs || !bdrv_is_inserted(bs) || bdrv_is_read_only(bs)) {
P
Paolo Bonzini 已提交
3995
        return 0;
K
Kevin Wolf 已提交
3996 3997
    }

3998
    /* Write back cached data to the OS even with cache=unsafe */
K
Kevin Wolf 已提交
3999 4000 4001 4002 4003 4004 4005
    if (bs->drv->bdrv_co_flush_to_os) {
        ret = bs->drv->bdrv_co_flush_to_os(bs);
        if (ret < 0) {
            return ret;
        }
    }

4006 4007
    /* But don't actually force it to the disk with cache=unsafe */
    if (bs->open_flags & BDRV_O_NO_FLUSH) {
4008
        goto flush_parent;
4009 4010
    }

K
Kevin Wolf 已提交
4011
    if (bs->drv->bdrv_co_flush_to_disk) {
4012
        ret = bs->drv->bdrv_co_flush_to_disk(bs);
P
Paolo Bonzini 已提交
4013 4014 4015 4016 4017 4018 4019 4020
    } 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) {
4021
            ret = -EIO;
P
Paolo Bonzini 已提交
4022 4023
        } else {
            qemu_coroutine_yield();
4024
            ret = co.ret;
P
Paolo Bonzini 已提交
4025 4026 4027 4028 4029 4030 4031 4032 4033 4034 4035 4036 4037
        }
    } 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.
         */
4038
        ret = 0;
P
Paolo Bonzini 已提交
4039
    }
4040 4041 4042 4043 4044 4045 4046
    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.
     */
4047
flush_parent:
4048
    return bdrv_co_flush(bs->file);
P
Paolo Bonzini 已提交
4049 4050
}

4051 4052 4053 4054 4055 4056 4057 4058 4059 4060 4061 4062 4063 4064 4065 4066
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);
    }
}

4067 4068 4069 4070 4071 4072 4073 4074 4075
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
4076 4077 4078 4079 4080 4081
int bdrv_flush(BlockDriverState *bs)
{
    Coroutine *co;
    RwCo rwco = {
        .bs = bs,
        .ret = NOT_DONE,
4082 4083
    };

P
Paolo Bonzini 已提交
4084 4085 4086 4087 4088 4089 4090 4091 4092
    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();
        }
4093
    }
P
Paolo Bonzini 已提交
4094 4095

    return rwco.ret;
4096 4097
}

4098 4099 4100 4101 4102 4103 4104 4105 4106 4107 4108 4109 4110 4111 4112 4113 4114 4115 4116 4117 4118 4119 4120 4121 4122 4123 4124 4125 4126 4127 4128 4129 4130 4131 4132 4133 4134 4135 4136 4137 4138 4139 4140 4141 4142 4143 4144 4145 4146 4147 4148 4149 4150 4151 4152 4153 4154 4155 4156 4157 4158
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;
    } else if (bs->drv->bdrv_co_discard) {
        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
4159 4160 4161 4162 4163 4164 4165 4166 4167
/**************************************************************/
/* removable device support */

/**
 * Return TRUE if the media is present
 */
int bdrv_is_inserted(BlockDriverState *bs)
{
    BlockDriver *drv = bs->drv;
4168

B
bellard 已提交
4169 4170 4171
    if (!drv)
        return 0;
    if (!drv->bdrv_is_inserted)
4172 4173
        return 1;
    return drv->bdrv_is_inserted(bs);
B
bellard 已提交
4174 4175 4176
}

/**
4177 4178
 * 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
4179 4180 4181 4182 4183
 */
int bdrv_media_changed(BlockDriverState *bs)
{
    BlockDriver *drv = bs->drv;

4184 4185 4186 4187
    if (drv && drv->bdrv_media_changed) {
        return drv->bdrv_media_changed(bs);
    }
    return -ENOTSUP;
B
bellard 已提交
4188 4189 4190 4191 4192
}

/**
 * If eject_flag is TRUE, eject the media. Otherwise, close the tray
 */
4193
void bdrv_eject(BlockDriverState *bs, bool eject_flag)
B
bellard 已提交
4194 4195 4196
{
    BlockDriver *drv = bs->drv;

4197 4198
    if (drv && drv->bdrv_eject) {
        drv->bdrv_eject(bs, eject_flag);
B
bellard 已提交
4199
    }
4200 4201 4202 4203

    if (bs->device_name[0] != '\0') {
        bdrv_emit_qmp_eject_event(bs, eject_flag);
    }
B
bellard 已提交
4204 4205 4206 4207 4208 4209
}

/**
 * Lock or unlock the media (if it is locked, the user won't be able
 * to eject it manually).
 */
4210
void bdrv_lock_medium(BlockDriverState *bs, bool locked)
B
bellard 已提交
4211 4212 4213
{
    BlockDriver *drv = bs->drv;

4214
    trace_bdrv_lock_medium(bs, locked);
S
Stefan Hajnoczi 已提交
4215

4216 4217
    if (drv && drv->bdrv_lock_medium) {
        drv->bdrv_lock_medium(bs, locked);
B
bellard 已提交
4218 4219
    }
}
4220 4221 4222 4223 4224 4225 4226 4227 4228 4229 4230

/* 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;
}
4231

4232 4233 4234
BlockDriverAIOCB *bdrv_aio_ioctl(BlockDriverState *bs,
        unsigned long int req, void *buf,
        BlockDriverCompletionFunc *cb, void *opaque)
4235
{
4236
    BlockDriver *drv = bs->drv;
4237

4238 4239 4240
    if (drv && drv->bdrv_aio_ioctl)
        return drv->bdrv_aio_ioctl(bs, req, buf, cb, opaque);
    return NULL;
4241
}
4242

4243 4244 4245 4246
void bdrv_set_buffer_alignment(BlockDriverState *bs, int align)
{
    bs->buffer_alignment = align;
}
4247

4248 4249 4250 4251
void *qemu_blockalign(BlockDriverState *bs, size_t size)
{
    return qemu_memalign((bs && bs->buffer_alignment) ? bs->buffer_alignment : 512, size);
}
4252 4253 4254 4255

void bdrv_set_dirty_tracking(BlockDriverState *bs, int enable)
{
    int64_t bitmap_size;
4256

4257
    bs->dirty_count = 0;
4258
    if (enable) {
4259 4260
        if (!bs->dirty_bitmap) {
            bitmap_size = (bdrv_getlength(bs) >> BDRV_SECTOR_BITS) +
4261 4262
                    BDRV_SECTORS_PER_DIRTY_CHUNK * BITS_PER_LONG - 1;
            bitmap_size /= BDRV_SECTORS_PER_DIRTY_CHUNK * BITS_PER_LONG;
4263

4264
            bs->dirty_bitmap = g_new0(unsigned long, bitmap_size);
4265
        }
4266
    } else {
4267
        if (bs->dirty_bitmap) {
4268
            g_free(bs->dirty_bitmap);
4269
            bs->dirty_bitmap = NULL;
4270
        }
4271 4272 4273 4274 4275
    }
}

int bdrv_get_dirty(BlockDriverState *bs, int64_t sector)
{
J
Jan Kiszka 已提交
4276
    int64_t chunk = sector / (int64_t)BDRV_SECTORS_PER_DIRTY_CHUNK;
4277

4278 4279
    if (bs->dirty_bitmap &&
        (sector << BDRV_SECTOR_BITS) < bdrv_getlength(bs)) {
4280 4281
        return !!(bs->dirty_bitmap[chunk / BITS_PER_LONG] &
            (1UL << (chunk % BITS_PER_LONG)));
4282 4283 4284 4285 4286
    } else {
        return 0;
    }
}

4287 4288 4289 4290 4291 4292 4293 4294 4295 4296 4297 4298 4299 4300 4301 4302 4303 4304 4305 4306 4307 4308 4309 4310 4311 4312 4313 4314 4315 4316 4317 4318 4319 4320 4321 4322 4323 4324 4325 4326 4327
int64_t bdrv_get_next_dirty(BlockDriverState *bs, int64_t sector)
{
    int64_t chunk;
    int bit, elem;

    /* Avoid an infinite loop.  */
    assert(bs->dirty_count > 0);

    sector = (sector | (BDRV_SECTORS_PER_DIRTY_CHUNK - 1)) + 1;
    chunk = sector / (int64_t)BDRV_SECTORS_PER_DIRTY_CHUNK;

    QEMU_BUILD_BUG_ON(sizeof(bs->dirty_bitmap[0]) * 8 != BITS_PER_LONG);
    elem = chunk / BITS_PER_LONG;
    bit = chunk % BITS_PER_LONG;
    for (;;) {
        if (sector >= bs->total_sectors) {
            sector = 0;
            bit = elem = 0;
        }
        if (bit == 0 && bs->dirty_bitmap[elem] == 0) {
            sector += BDRV_SECTORS_PER_DIRTY_CHUNK * BITS_PER_LONG;
            elem++;
        } else {
            if (bs->dirty_bitmap[elem] & (1UL << bit)) {
                return sector;
            }
            sector += BDRV_SECTORS_PER_DIRTY_CHUNK;
            if (++bit == BITS_PER_LONG) {
                bit = 0;
                elem++;
            }
        }
    }
}

void bdrv_set_dirty(BlockDriverState *bs, int64_t cur_sector,
                    int nr_sectors)
{
    set_dirty_bitmap(bs, cur_sector, nr_sectors, 1);
}

4328 4329
void bdrv_reset_dirty(BlockDriverState *bs, int64_t cur_sector,
                      int nr_sectors)
4330 4331 4332
{
    set_dirty_bitmap(bs, cur_sector, nr_sectors, 0);
}
4333 4334 4335 4336 4337

int64_t bdrv_get_dirty_count(BlockDriverState *bs)
{
    return bs->dirty_count;
}
J
Jes Sorensen 已提交
4338

4339 4340 4341 4342 4343 4344 4345 4346 4347 4348 4349
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;
}

4350 4351
void bdrv_iostatus_enable(BlockDriverState *bs)
{
4352
    bs->iostatus_enabled = true;
4353
    bs->iostatus = BLOCK_DEVICE_IO_STATUS_OK;
4354 4355 4356 4357 4358 4359
}

/* 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)
{
4360
    return (bs->iostatus_enabled &&
4361 4362 4363
           (bs->on_write_error == BLOCKDEV_ON_ERROR_ENOSPC ||
            bs->on_write_error == BLOCKDEV_ON_ERROR_STOP   ||
            bs->on_read_error == BLOCKDEV_ON_ERROR_STOP));
4364 4365 4366 4367
}

void bdrv_iostatus_disable(BlockDriverState *bs)
{
4368
    bs->iostatus_enabled = false;
4369 4370 4371 4372 4373
}

void bdrv_iostatus_reset(BlockDriverState *bs)
{
    if (bdrv_iostatus_is_enabled(bs)) {
4374
        bs->iostatus = BLOCK_DEVICE_IO_STATUS_OK;
4375 4376 4377
        if (bs->job) {
            block_job_iostatus_reset(bs->job);
        }
4378 4379 4380 4381 4382
    }
}

void bdrv_iostatus_set_err(BlockDriverState *bs, int error)
{
4383 4384
    assert(bdrv_iostatus_is_enabled(bs));
    if (bs->iostatus == BLOCK_DEVICE_IO_STATUS_OK) {
4385 4386
        bs->iostatus = error == ENOSPC ? BLOCK_DEVICE_IO_STATUS_NOSPACE :
                                         BLOCK_DEVICE_IO_STATUS_FAILED;
4387 4388 4389
    }
}

4390 4391 4392 4393 4394 4395 4396
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
4397
    cookie->start_time_ns = get_clock();
4398 4399 4400 4401 4402 4403 4404 4405 4406 4407
    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
4408
    bs->total_time_ns[cookie->type] += get_clock() - cookie->start_time_ns;
4409 4410
}

J
Jes Sorensen 已提交
4411 4412 4413 4414 4415
int 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)
{
    QEMUOptionParameter *param = NULL, *create_options = NULL;
K
Kevin Wolf 已提交
4416
    QEMUOptionParameter *backing_fmt, *backing_file, *size;
J
Jes Sorensen 已提交
4417 4418
    BlockDriverState *bs = NULL;
    BlockDriver *drv, *proto_drv;
4419
    BlockDriver *backing_drv = NULL;
J
Jes Sorensen 已提交
4420 4421 4422 4423 4424 4425
    int ret = 0;

    /* Find driver and parse its options */
    drv = bdrv_find_format(fmt);
    if (!drv) {
        error_report("Unknown file format '%s'", fmt);
4426
        ret = -EINVAL;
J
Jes Sorensen 已提交
4427 4428 4429 4430 4431 4432
        goto out;
    }

    proto_drv = bdrv_find_protocol(filename);
    if (!proto_drv) {
        error_report("Unknown protocol '%s'", filename);
4433
        ret = -EINVAL;
J
Jes Sorensen 已提交
4434 4435 4436 4437 4438 4439 4440 4441 4442 4443 4444 4445 4446 4447 4448 4449 4450 4451
        goto out;
    }

    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) {
            error_report("Invalid options for file format '%s'.", fmt);
4452
            ret = -EINVAL;
J
Jes Sorensen 已提交
4453 4454 4455 4456 4457 4458 4459 4460 4461
            goto out;
        }
    }

    if (base_filename) {
        if (set_option_parameter(param, BLOCK_OPT_BACKING_FILE,
                                 base_filename)) {
            error_report("Backing file not supported for file format '%s'",
                         fmt);
4462
            ret = -EINVAL;
J
Jes Sorensen 已提交
4463 4464 4465 4466 4467 4468 4469 4470
            goto out;
        }
    }

    if (base_fmt) {
        if (set_option_parameter(param, BLOCK_OPT_BACKING_FMT, base_fmt)) {
            error_report("Backing file format not supported for file "
                         "format '%s'", fmt);
4471
            ret = -EINVAL;
J
Jes Sorensen 已提交
4472 4473 4474 4475
            goto out;
        }
    }

4476 4477 4478 4479 4480
    backing_file = get_option_parameter(param, BLOCK_OPT_BACKING_FILE);
    if (backing_file && backing_file->value.s) {
        if (!strcmp(filename, backing_file->value.s)) {
            error_report("Error: Trying to create an image with the "
                         "same filename as the backing file");
4481
            ret = -EINVAL;
4482 4483 4484 4485
            goto out;
        }
    }

J
Jes Sorensen 已提交
4486 4487
    backing_fmt = get_option_parameter(param, BLOCK_OPT_BACKING_FMT);
    if (backing_fmt && backing_fmt->value.s) {
4488 4489
        backing_drv = bdrv_find_format(backing_fmt->value.s);
        if (!backing_drv) {
J
Jes Sorensen 已提交
4490 4491
            error_report("Unknown backing file format '%s'",
                         backing_fmt->value.s);
4492
            ret = -EINVAL;
J
Jes Sorensen 已提交
4493 4494 4495 4496 4497 4498
            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
4499 4500
    size = get_option_parameter(param, BLOCK_OPT_SIZE);
    if (size && size->value.n == -1) {
J
Jes Sorensen 已提交
4501 4502 4503
        if (backing_file && backing_file->value.s) {
            uint64_t size;
            char buf[32];
4504 4505 4506 4507 4508
            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
4509 4510 4511

            bs = bdrv_new("");

4512
            ret = bdrv_open(bs, backing_file->value.s, back_flags, backing_drv);
J
Jes Sorensen 已提交
4513
            if (ret < 0) {
4514
                error_report("Could not open '%s'", backing_file->value.s);
J
Jes Sorensen 已提交
4515 4516 4517 4518 4519 4520 4521 4522 4523
                goto out;
            }
            bdrv_get_geometry(bs, &size);
            size *= 512;

            snprintf(buf, sizeof(buf), "%" PRId64, size);
            set_option_parameter(param, BLOCK_OPT_SIZE, buf);
        } else {
            error_report("Image creation needs a size parameter");
4524
            ret = -EINVAL;
J
Jes Sorensen 已提交
4525 4526 4527 4528 4529 4530 4531 4532 4533 4534 4535 4536 4537 4538 4539 4540 4541 4542 4543 4544 4545 4546 4547 4548 4549 4550 4551 4552 4553 4554
            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) {
            error_report("Formatting or formatting option not supported for "
                         "file format '%s'", fmt);
        } else if (ret == -EFBIG) {
            error_report("The image size is too large for file format '%s'",
                         fmt);
        } else {
            error_report("%s: error while creating %s: %s", filename, fmt,
                         strerror(-ret));
        }
    }

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

    if (bs) {
        bdrv_delete(bs);
    }
4555 4556

    return ret;
J
Jes Sorensen 已提交
4557
}