block.c 109.4 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 "module.h"
30
#include "qjson.h"
31
#include "qemu-coroutine.h"
L
Luiz Capitulino 已提交
32
#include "qmp-commands.h"
33
#include "qemu-timer.h"
B
bellard 已提交
34

J
Juan Quintela 已提交
35
#ifdef CONFIG_BSD
B
bellard 已提交
36 37 38
#include <sys/types.h>
#include <sys/stat.h>
#include <sys/ioctl.h>
B
Blue Swirl 已提交
39
#include <sys/queue.h>
40
#ifndef __DragonFly__
B
bellard 已提交
41 42
#include <sys/disk.h>
#endif
43
#endif
B
bellard 已提交
44

45 46 47 48
#ifdef _WIN32
#include <windows.h>
#endif

49 50
#define NOT_DONE 0x7fffffff /* used while emulated sync operation in progress */

51 52
typedef enum {
    BDRV_REQ_COPY_ON_READ = 0x1,
53
    BDRV_REQ_ZERO_WRITE   = 0x2,
54 55
} BdrvRequestFlags;

56
static void bdrv_dev_change_media_cb(BlockDriverState *bs, bool load);
57 58
static BlockDriverAIOCB *bdrv_aio_readv_em(BlockDriverState *bs,
        int64_t sector_num, QEMUIOVector *qiov, int nb_sectors,
59
        BlockDriverCompletionFunc *cb, void *opaque);
60 61
static BlockDriverAIOCB *bdrv_aio_writev_em(BlockDriverState *bs,
        int64_t sector_num, QEMUIOVector *qiov, int nb_sectors,
62
        BlockDriverCompletionFunc *cb, void *opaque);
63 64 65 66 67 68
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);
69
static int coroutine_fn bdrv_co_do_readv(BlockDriverState *bs,
70 71
    int64_t sector_num, int nb_sectors, QEMUIOVector *qiov,
    BdrvRequestFlags flags);
72
static int coroutine_fn bdrv_co_do_writev(BlockDriverState *bs,
73 74
    int64_t sector_num, int nb_sectors, QEMUIOVector *qiov,
    BdrvRequestFlags flags);
75 76 77 78 79 80
static BlockDriverAIOCB *bdrv_co_aio_rw_vector(BlockDriverState *bs,
                                               int64_t sector_num,
                                               QEMUIOVector *qiov,
                                               int nb_sectors,
                                               BlockDriverCompletionFunc *cb,
                                               void *opaque,
81
                                               bool is_write);
82
static void coroutine_fn bdrv_co_do_rw(void *opaque);
B
bellard 已提交
83

Z
Zhi Yong Wu 已提交
84 85 86 87 88 89 90
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);

91 92
static QTAILQ_HEAD(, BlockDriverState) bdrv_states =
    QTAILQ_HEAD_INITIALIZER(bdrv_states);
93

94 95
static QLIST_HEAD(, BlockDriver) bdrv_drivers =
    QLIST_HEAD_INITIALIZER(bdrv_drivers);
B
bellard 已提交
96

97 98 99
/* The device to use for VM snapshots */
static BlockDriverState *bs_snapshots;

100 101 102
/* If non-zero, use only whitelisted block drivers */
static int use_bdrv_whitelist;

103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122
#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

123
/* throttling disk I/O limits */
Z
Zhi Yong Wu 已提交
124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141
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));
}

142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170
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
171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195
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);
}

196 197 198 199 200 201 202 203 204 205 206 207 208
/* check if the path starts with "<protocol>:" */
static int path_has_protocol(const char *path)
{
#ifdef _WIN32
    if (is_windows_drive(path) ||
        is_windows_drive_prefix(path)) {
        return 0;
    }
#endif

    return strchr(path, ':') != NULL;
}

B
bellard 已提交
209
int path_is_absolute(const char *path)
210
{
B
bellard 已提交
211
    const char *p;
B
bellard 已提交
212 213 214 215 216
#ifdef _WIN32
    /* specific case for names like: "\\.\d:" */
    if (*path == '/' || *path == '\\')
        return 1;
#endif
B
bellard 已提交
217 218 219 220 221
    p = strchr(path, ':');
    if (p)
        p++;
    else
        p = path;
222 223 224 225 226
#ifdef _WIN32
    return (*p == '/' || *p == '\\');
#else
    return (*p == '/');
#endif
227 228
}

B
bellard 已提交
229 230 231 232 233 234
/* 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)
235
{
B
bellard 已提交
236 237 238 239 240 241 242 243 244 245 246 247 248
    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;
249 250 251 252 253 254 255 256 257
        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
258 259 260 261 262 263 264 265 266 267 268 269
        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);
270 271 272
    }
}

273
void bdrv_register(BlockDriver *bdrv)
B
bellard 已提交
274
{
275 276
    /* Block drivers without coroutine functions need emulation */
    if (!bdrv->bdrv_co_readv) {
277 278 279
        bdrv->bdrv_co_readv = bdrv_co_readv_em;
        bdrv->bdrv_co_writev = bdrv_co_writev_em;

280 281 282
        /* 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.
         */
283 284 285 286 287
        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
288
    }
289

290
    QLIST_INSERT_HEAD(&bdrv_drivers, bdrv, list);
B
bellard 已提交
291
}
B
bellard 已提交
292 293 294 295

/* create a new block device (by default it is empty) */
BlockDriverState *bdrv_new(const char *device_name)
{
296
    BlockDriverState *bs;
B
bellard 已提交
297

298
    bs = g_malloc0(sizeof(BlockDriverState));
B
bellard 已提交
299
    pstrcpy(bs->device_name, sizeof(bs->device_name), device_name);
B
bellard 已提交
300
    if (device_name[0] != '\0') {
301
        QTAILQ_INSERT_TAIL(&bdrv_states, bs, list);
B
bellard 已提交
302
    }
303
    bdrv_iostatus_disable(bs);
B
bellard 已提交
304 305 306
    return bs;
}

B
bellard 已提交
307 308 309
BlockDriver *bdrv_find_format(const char *format_name)
{
    BlockDriver *drv1;
310 311
    QLIST_FOREACH(drv1, &bdrv_drivers, list) {
        if (!strcmp(drv1->format_name, format_name)) {
B
bellard 已提交
312
            return drv1;
313
        }
B
bellard 已提交
314 315 316 317
    }
    return NULL;
}

318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341
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;
}

342 343
int bdrv_create(BlockDriver *drv, const char* filename,
    QEMUOptionParameter *options)
B
bellard 已提交
344 345 346
{
    if (!drv->bdrv_create)
        return -ENOTSUP;
347 348

    return drv->bdrv_create(filename, options);
B
bellard 已提交
349 350
}

351 352 353 354
int bdrv_create_file(const char* filename, QEMUOptionParameter *options)
{
    BlockDriver *drv;

355
    drv = bdrv_find_protocol(filename);
356
    if (drv == NULL) {
357
        return -ENOENT;
358 359 360 361 362
    }

    return bdrv_create(drv, filename, options);
}

B
bellard 已提交
363
#ifdef _WIN32
364
void get_tmp_filename(char *filename, int size)
B
bellard 已提交
365
{
366
    char temp_dir[MAX_PATH];
367

368 369
    GetTempPath(MAX_PATH, temp_dir);
    GetTempFileName(temp_dir, "qem", 0, filename);
B
bellard 已提交
370 371
}
#else
372
void get_tmp_filename(char *filename, int size)
B
bellard 已提交
373
{
B
bellard 已提交
374
    int fd;
375
    const char *tmpdir;
B
bellard 已提交
376
    /* XXX: race condition possible */
A
aurel32 已提交
377 378 379 380
    tmpdir = getenv("TMPDIR");
    if (!tmpdir)
        tmpdir = "/tmp";
    snprintf(filename, size, "%s/vl.XXXXXX", tmpdir);
B
bellard 已提交
381 382 383
    fd = mkstemp(filename);
    close(fd);
}
B
bellard 已提交
384
#endif
B
bellard 已提交
385

386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407
/*
 * 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;
}

408
BlockDriver *bdrv_find_protocol(const char *filename)
B
bellard 已提交
409 410 411
{
    BlockDriver *drv1;
    char protocol[128];
412
    int len;
B
bellard 已提交
413
    const char *p;
B
bellard 已提交
414

415 416
    /* TODO Drivers without bdrv_file_open must be specified explicitly */

417 418 419 420 421 422 423 424 425 426 427 428
    /*
     * 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;
    }

429
    if (!path_has_protocol(filename)) {
430
        return bdrv_find_format("file");
431
    }
432 433
    p = strchr(filename, ':');
    assert(p != NULL);
434 435 436 437 438
    len = p - filename;
    if (len > sizeof(protocol) - 1)
        len = sizeof(protocol) - 1;
    memcpy(protocol, filename, len);
    protocol[len] = '\0';
439
    QLIST_FOREACH(drv1, &bdrv_drivers, list) {
440
        if (drv1->protocol_name &&
441
            !strcmp(drv1->protocol_name, protocol)) {
B
bellard 已提交
442
            return drv1;
443
        }
B
bellard 已提交
444 445 446 447
    }
    return NULL;
}

448
static int find_image_format(const char *filename, BlockDriver **pdrv)
449 450 451 452 453 454
{
    int ret, score, score_max;
    BlockDriver *drv1, *drv;
    uint8_t buf[2048];
    BlockDriverState *bs;

455
    ret = bdrv_file_open(&bs, filename, 0);
456 457 458 459
    if (ret < 0) {
        *pdrv = NULL;
        return ret;
    }
460

461 462
    /* Return the raw BlockDriver * to scsi-generic devices or empty drives */
    if (bs->sg || !bdrv_is_inserted(bs)) {
463
        bdrv_delete(bs);
464 465 466 467 468 469
        drv = bdrv_find_format("raw");
        if (!drv) {
            ret = -ENOENT;
        }
        *pdrv = drv;
        return ret;
470
    }
471

B
bellard 已提交
472 473 474
    ret = bdrv_pread(bs, 0, buf, sizeof(buf));
    bdrv_delete(bs);
    if (ret < 0) {
475 476
        *pdrv = NULL;
        return ret;
B
bellard 已提交
477 478
    }

B
bellard 已提交
479
    score_max = 0;
480
    drv = NULL;
481
    QLIST_FOREACH(drv1, &bdrv_drivers, list) {
B
bellard 已提交
482 483 484 485 486 487
        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
488
        }
B
bellard 已提交
489
    }
490 491 492 493 494
    if (!drv) {
        ret = -ENOENT;
    }
    *pdrv = drv;
    return ret;
B
bellard 已提交
495 496
}

497 498 499 500 501 502 503
/**
 * Set the current 'total_sectors' value
 */
static int refresh_total_sectors(BlockDriverState *bs, int64_t hint)
{
    BlockDriver *drv = bs->drv;

504 505 506 507
    /* Do not attempt drv->bdrv_getlength() on scsi-generic devices */
    if (bs->sg)
        return 0;

508 509 510 511 512 513 514 515 516 517 518 519 520
    /* 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;
}

521 522 523 524 525 526 527 528 529 530 531
/**
 * 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;
532 533
    } else if (!strcmp(mode, "directsync")) {
        *flags |= BDRV_O_NOCACHE;
534 535 536 537 538 539 540 541 542 543 544 545 546 547
    } 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;
}

548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563
/**
 * 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--;
}

564 565 566 567 568 569 570 571 572 573
/*
 * 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);

S
Stefan Hajnoczi 已提交
574 575
    trace_bdrv_open_common(bs, filename, flags, drv->format_name);

576
    bs->file = NULL;
577
    bs->total_sectors = 0;
578 579
    bs->encrypted = 0;
    bs->valid_key = 0;
580
    bs->sg = 0;
581
    bs->open_flags = flags;
582
    bs->growable = 0;
583 584
    bs->buffer_alignment = 512;

585 586 587 588 589
    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);
    }

590
    pstrcpy(bs->filename, sizeof(bs->filename), filename);
591
    bs->backing_file[0] = '\0';
592 593 594 595 596 597

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

    bs->drv = drv;
598
    bs->opaque = g_malloc0(drv->instance_size);
599

600
    bs->enable_write_cache = !!(flags & BDRV_O_CACHE_WB);
601 602 603 604 605 606 607 608

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

    /*
S
Stefan Weil 已提交
609
     * Snapshots should be writable.
610 611 612 613 614
     */
    if (bs->is_temporary) {
        open_flags |= BDRV_O_RDWR;
    }

615 616
    bs->keep_read_only = bs->read_only = !(open_flags & BDRV_O_RDWR);

617 618 619 620 621 622 623 624 625 626
    /* 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);
        }
    }

627 628 629 630
    if (ret < 0) {
        goto free_and_fail;
    }

631 632 633
    ret = refresh_total_sectors(bs, bs->total_sectors);
    if (ret < 0) {
        goto free_and_fail;
634
    }
635

636 637 638 639 640 641 642 643
#ifndef _WIN32
    if (bs->is_temporary) {
        unlink(filename);
    }
#endif
    return 0;

free_and_fail:
644 645 646 647
    if (bs->file) {
        bdrv_delete(bs->file);
        bs->file = NULL;
    }
648
    g_free(bs->opaque);
649 650 651 652 653
    bs->opaque = NULL;
    bs->drv = NULL;
    return ret;
}

K
Kevin Wolf 已提交
654 655 656
/*
 * Opens a file using a protocol (file, host_device, nbd, ...)
 */
B
bellard 已提交
657
int bdrv_file_open(BlockDriverState **pbs, const char *filename, int flags)
B
bellard 已提交
658
{
B
bellard 已提交
659
    BlockDriverState *bs;
660
    BlockDriver *drv;
B
bellard 已提交
661 662
    int ret;

663
    drv = bdrv_find_protocol(filename);
664 665 666 667
    if (!drv) {
        return -ENOENT;
    }

B
bellard 已提交
668
    bs = bdrv_new("");
K
Kevin Wolf 已提交
669
    ret = bdrv_open_common(bs, filename, flags, drv);
B
bellard 已提交
670 671 672
    if (ret < 0) {
        bdrv_delete(bs);
        return ret;
673
    }
674
    bs->growable = 1;
B
bellard 已提交
675 676 677 678
    *pbs = bs;
    return 0;
}

K
Kevin Wolf 已提交
679 680 681
/*
 * Opens a disk image (raw, qcow2, vmdk, ...)
 */
K
Kevin Wolf 已提交
682 683
int bdrv_open(BlockDriverState *bs, const char *filename, int flags,
              BlockDriver *drv)
B
bellard 已提交
684
{
K
Kevin Wolf 已提交
685
    int ret;
K
Kevin Wolf 已提交
686
    char tmp_filename[PATH_MAX];
B
bellard 已提交
687

B
bellard 已提交
688
    if (flags & BDRV_O_SNAPSHOT) {
B
bellard 已提交
689 690
        BlockDriverState *bs1;
        int64_t total_size;
A
aliguori 已提交
691
        int is_protocol = 0;
K
Kevin Wolf 已提交
692 693
        BlockDriver *bdrv_qcow2;
        QEMUOptionParameter *options;
K
Kevin Wolf 已提交
694
        char backing_filename[PATH_MAX];
695

B
bellard 已提交
696 697
        /* if snapshot, we create a temporary backing file and open it
           instead of opening 'filename' directly */
698

B
bellard 已提交
699 700
        /* if there is a backing file, use it */
        bs1 = bdrv_new("");
K
Kevin Wolf 已提交
701
        ret = bdrv_open(bs1, filename, 0, drv);
702
        if (ret < 0) {
B
bellard 已提交
703
            bdrv_delete(bs1);
704
            return ret;
B
bellard 已提交
705
        }
706
        total_size = bdrv_getlength(bs1) & BDRV_SECTOR_MASK;
A
aliguori 已提交
707 708 709 710

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

B
bellard 已提交
711
        bdrv_delete(bs1);
712

B
bellard 已提交
713
        get_tmp_filename(tmp_filename, sizeof(tmp_filename));
A
aliguori 已提交
714 715 716 717 718

        /* Real path is meaningless for protocols */
        if (is_protocol)
            snprintf(backing_filename, sizeof(backing_filename),
                     "%s", filename);
719 720
        else if (!realpath(filename, backing_filename))
            return -errno;
A
aliguori 已提交
721

K
Kevin Wolf 已提交
722 723 724
        bdrv_qcow2 = bdrv_find_format("qcow2");
        options = parse_option_parameters("", bdrv_qcow2->create_options, NULL);

725
        set_option_parameter_int(options, BLOCK_OPT_SIZE, total_size);
K
Kevin Wolf 已提交
726 727 728 729 730 731 732
        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);
733
        free_option_parameters(options);
734 735
        if (ret < 0) {
            return ret;
B
bellard 已提交
736
        }
K
Kevin Wolf 已提交
737

B
bellard 已提交
738
        filename = tmp_filename;
K
Kevin Wolf 已提交
739
        drv = bdrv_qcow2;
B
bellard 已提交
740 741
        bs->is_temporary = 1;
    }
B
bellard 已提交
742

K
Kevin Wolf 已提交
743
    /* Find the right image format driver */
744
    if (!drv) {
745
        ret = find_image_format(filename, &drv);
746
    }
747

748 749
    if (!drv) {
        goto unlink_and_fail;
B
bellard 已提交
750
    }
K
Kevin Wolf 已提交
751 752 753 754

    /* Open the image */
    ret = bdrv_open_common(bs, filename, flags, drv);
    if (ret < 0) {
755 756 757
        goto unlink_and_fail;
    }

K
Kevin Wolf 已提交
758 759 760 761 762 763 764
    /* If there is a backing file, use it */
    if ((flags & BDRV_O_NO_BACKING) == 0 && bs->backing_file[0] != '\0') {
        char backing_filename[PATH_MAX];
        int back_flags;
        BlockDriver *back_drv = NULL;

        bs->backing_hd = bdrv_new("");
765 766 767 768 769 770 771 772 773 774

        if (path_has_protocol(bs->backing_file)) {
            pstrcpy(backing_filename, sizeof(backing_filename),
                    bs->backing_file);
        } else {
            path_combine(backing_filename, sizeof(backing_filename),
                         filename, bs->backing_file);
        }

        if (bs->backing_format[0] != '\0') {
K
Kevin Wolf 已提交
775
            back_drv = bdrv_find_format(bs->backing_format);
776
        }
K
Kevin Wolf 已提交
777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795

        /* backing files always opened read-only */
        back_flags =
            flags & ~(BDRV_O_RDWR | BDRV_O_SNAPSHOT | BDRV_O_NO_BACKING);

        ret = bdrv_open(bs->backing_hd, backing_filename, back_flags, back_drv);
        if (ret < 0) {
            bdrv_close(bs);
            return ret;
        }
        if (bs->is_temporary) {
            bs->backing_hd->keep_read_only = !(flags & BDRV_O_RDWR);
        } else {
            /* base image inherits from "parent" */
            bs->backing_hd->keep_read_only = bs->keep_read_only;
        }
    }

    if (!bdrv_key_required(bs)) {
796
        bdrv_dev_change_media_cb(bs, true);
K
Kevin Wolf 已提交
797 798
    }

Z
Zhi Yong Wu 已提交
799 800 801 802 803
    /* throttling disk I/O limits */
    if (bs->io_limits_enabled) {
        bdrv_io_limits_enable(bs);
    }

K
Kevin Wolf 已提交
804 805 806 807 808 809 810 811 812
    return 0;

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

B
bellard 已提交
813 814
void bdrv_close(BlockDriverState *bs)
{
B
bellard 已提交
815
    if (bs->drv) {
816 817 818
        if (bs == bs_snapshots) {
            bs_snapshots = NULL;
        }
819
        if (bs->backing_hd) {
B
bellard 已提交
820
            bdrv_delete(bs->backing_hd);
821 822
            bs->backing_hd = NULL;
        }
B
bellard 已提交
823
        bs->drv->bdrv_close(bs);
824
        g_free(bs->opaque);
B
bellard 已提交
825 826 827 828
#ifdef _WIN32
        if (bs->is_temporary) {
            unlink(bs->filename);
        }
B
bellard 已提交
829
#endif
B
bellard 已提交
830 831
        bs->opaque = NULL;
        bs->drv = NULL;
832
        bs->copy_on_read = 0;
B
bellard 已提交
833

834 835 836 837
        if (bs->file != NULL) {
            bdrv_close(bs->file);
        }

838
        bdrv_dev_change_media_cb(bs, false);
B
bellard 已提交
839
    }
Z
Zhi Yong Wu 已提交
840 841 842 843 844

    /*throttling disk I/O limits*/
    if (bs->io_limits_enabled) {
        bdrv_io_limits_disable(bs);
    }
B
bellard 已提交
845 846
}

847 848 849 850 851 852 853 854 855
void bdrv_close_all(void)
{
    BlockDriverState *bs;

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

856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874
/*
 * 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.
 */
void bdrv_drain_all(void)
{
    BlockDriverState *bs;

    qemu_aio_flush();

    /* 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));
    }
}

875 876 877 878 879 880 881 882 883 884
/* 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';
}

B
bellard 已提交
885 886
void bdrv_delete(BlockDriverState *bs)
{
887
    assert(!bs->dev);
888

889
    /* remove from list, if necessary */
890
    bdrv_make_anon(bs);
891

B
bellard 已提交
892
    bdrv_close(bs);
893 894 895 896
    if (bs->file != NULL) {
        bdrv_delete(bs->file);
    }

897
    assert(bs != bs_snapshots);
898
    g_free(bs);
B
bellard 已提交
899 900
}

901 902
int bdrv_attach_dev(BlockDriverState *bs, void *dev)
/* TODO change to DeviceState *dev when all users are qdevified */
903
{
904
    if (bs->dev) {
905 906
        return -EBUSY;
    }
907
    bs->dev = dev;
908
    bdrv_iostatus_reset(bs);
909 910 911
    return 0;
}

912 913
/* TODO qdevified devices don't use this, remove when devices are qdevified */
void bdrv_attach_dev_nofail(BlockDriverState *bs, void *dev)
914
{
915 916 917 918 919 920 921 922 923 924
    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;
925 926
    bs->dev_ops = NULL;
    bs->dev_opaque = NULL;
927
    bs->buffer_alignment = 512;
928 929
}

930 931
/* TODO change to return DeviceState * when all users are qdevified */
void *bdrv_get_attached_dev(BlockDriverState *bs)
932
{
933
    return bs->dev;
934 935
}

936 937 938 939 940
void bdrv_set_dev_ops(BlockDriverState *bs, const BlockDevOps *ops,
                      void *opaque)
{
    bs->dev_ops = ops;
    bs->dev_opaque = opaque;
941 942 943
    if (bdrv_dev_has_removable_media(bs) && bs == bs_snapshots) {
        bs_snapshots = NULL;
    }
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
void bdrv_emit_qmp_error_event(const BlockDriverState *bdrv,
                               BlockQMPEventAction action, int is_read)
{
    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");
    monitor_protocol_event(QEVENT_BLOCK_IO_ERROR, data);

    qobject_decref(data);
}

975 976 977 978 979 980 981 982 983 984 985
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);
}

986
static void bdrv_dev_change_media_cb(BlockDriverState *bs, bool load)
987
{
988
    if (bs->dev_ops && bs->dev_ops->change_media_cb) {
989
        bool tray_was_closed = !bdrv_dev_is_tray_open(bs);
990
        bs->dev_ops->change_media_cb(bs->dev_opaque, load);
991 992 993 994 995 996 997 998
        if (tray_was_closed) {
            /* tray open */
            bdrv_emit_qmp_eject_event(bs, true);
        }
        if (load) {
            /* tray close */
            bdrv_emit_qmp_eject_event(bs, false);
        }
999 1000 1001
    }
}

1002 1003 1004 1005 1006
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
1007 1008 1009 1010 1011 1012 1013
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);
    }
}

1014 1015 1016 1017 1018 1019 1020 1021
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;
}

1022 1023 1024 1025
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);
1026 1027 1028
    }
}

1029 1030 1031 1032 1033 1034 1035 1036
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
1037 1038 1039
/*
 * Run consistency checks on an image
 *
1040
 * Returns 0 if the check could be completed (it doesn't mean that the image is
1041
 * free of errors) or -errno when an internal error occurred. The results of the
1042
 * check are stored in res.
A
aliguori 已提交
1043
 */
1044
int bdrv_check(BlockDriverState *bs, BdrvCheckResult *res)
A
aliguori 已提交
1045 1046 1047 1048 1049
{
    if (bs->drv->bdrv_check == NULL) {
        return -ENOTSUP;
    }

1050
    memset(res, 0, sizeof(*res));
1051
    return bs->drv->bdrv_check(bs, res);
A
aliguori 已提交
1052 1053
}

1054 1055
#define COMMIT_BUF_SECTORS 2048

1056 1057 1058
/* commit COW file into the raw image */
int bdrv_commit(BlockDriverState *bs)
{
B
bellard 已提交
1059
    BlockDriver *drv = bs->drv;
1060
    BlockDriver *backing_drv;
1061 1062
    int64_t sector, total_sectors;
    int n, ro, open_flags;
1063
    int ret = 0, rw_ret = 0;
1064
    uint8_t *buf;
1065 1066
    char filename[1024];
    BlockDriverState *bs_rw, *bs_ro;
1067

B
bellard 已提交
1068 1069
    if (!drv)
        return -ENOMEDIUM;
1070 1071 1072
    
    if (!bs->backing_hd) {
        return -ENOTSUP;
1073 1074
    }

1075 1076 1077
    if (bs->backing_hd->keep_read_only) {
        return -EACCES;
    }
1078

1079 1080 1081 1082
    if (bdrv_in_use(bs) || bdrv_in_use(bs->backing_hd)) {
        return -EBUSY;
    }

1083
    backing_drv = bs->backing_hd->drv;
1084 1085 1086 1087 1088 1089 1090 1091 1092
    ro = bs->backing_hd->read_only;
    strncpy(filename, bs->backing_hd->filename, sizeof(filename));
    open_flags =  bs->backing_hd->open_flags;

    if (ro) {
        /* re-open as RW */
        bdrv_delete(bs->backing_hd);
        bs->backing_hd = NULL;
        bs_rw = bdrv_new("");
1093 1094
        rw_ret = bdrv_open(bs_rw, filename, open_flags | BDRV_O_RDWR,
            backing_drv);
1095 1096 1097 1098
        if (rw_ret < 0) {
            bdrv_delete(bs_rw);
            /* try to re-open read-only */
            bs_ro = bdrv_new("");
1099 1100
            ret = bdrv_open(bs_ro, filename, open_flags & ~BDRV_O_RDWR,
                backing_drv);
1101 1102 1103 1104 1105 1106 1107 1108 1109 1110
            if (ret < 0) {
                bdrv_delete(bs_ro);
                /* drive not functional anymore */
                bs->drv = NULL;
                return ret;
            }
            bs->backing_hd = bs_ro;
            return rw_ret;
        }
        bs->backing_hd = bs_rw;
B
bellard 已提交
1111
    }
1112

J
Jan Kiszka 已提交
1113
    total_sectors = bdrv_getlength(bs) >> BDRV_SECTOR_BITS;
1114
    buf = g_malloc(COMMIT_BUF_SECTORS * BDRV_SECTOR_SIZE);
1115 1116

    for (sector = 0; sector < total_sectors; sector += n) {
1117
        if (bdrv_is_allocated(bs, sector, COMMIT_BUF_SECTORS, &n)) {
1118 1119 1120 1121 1122 1123 1124 1125 1126 1127

            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
1128
        }
1129
    }
1130

1131 1132 1133 1134
    if (drv->bdrv_make_empty) {
        ret = drv->bdrv_make_empty(bs);
        bdrv_flush(bs);
    }
1135

1136 1137 1138 1139 1140 1141
    /*
     * Make sure all data we wrote to the backing device is actually
     * stable on disk.
     */
    if (bs->backing_hd)
        bdrv_flush(bs->backing_hd);
1142 1143

ro_cleanup:
1144
    g_free(buf);
1145 1146 1147 1148 1149 1150

    if (ro) {
        /* re-open as RO */
        bdrv_delete(bs->backing_hd);
        bs->backing_hd = NULL;
        bs_ro = bdrv_new("");
1151 1152
        ret = bdrv_open(bs_ro, filename, open_flags & ~BDRV_O_RDWR,
            backing_drv);
1153 1154 1155 1156 1157 1158 1159 1160 1161 1162
        if (ret < 0) {
            bdrv_delete(bs_ro);
            /* drive not functional anymore */
            bs->drv = NULL;
            return ret;
        }
        bs->backing_hd = bs_ro;
        bs->backing_hd->keep_read_only = 0;
    }

1163
    return ret;
1164 1165
}

1166 1167 1168 1169 1170 1171 1172 1173 1174
void bdrv_commit_all(void)
{
    BlockDriverState *bs;

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

S
Stefan Hajnoczi 已提交
1175 1176 1177 1178 1179 1180
struct BdrvTrackedRequest {
    BlockDriverState *bs;
    int64_t sector_num;
    int nb_sectors;
    bool is_write;
    QLIST_ENTRY(BdrvTrackedRequest) list;
1181
    Coroutine *co; /* owner, used for deadlock detection */
1182
    CoQueue wait_queue; /* coroutines blocked on this request */
S
Stefan Hajnoczi 已提交
1183 1184 1185 1186 1187 1188 1189 1190 1191 1192
};

/**
 * 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);
1193
    qemu_co_queue_restart_all(&req->wait_queue);
S
Stefan Hajnoczi 已提交
1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208
}

/**
 * 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,
1209
        .co = qemu_coroutine_self(),
S
Stefan Hajnoczi 已提交
1210 1211
    };

1212 1213
    qemu_co_queue_init(&req->wait_queue);

S
Stefan Hajnoczi 已提交
1214 1215 1216
    QLIST_INSERT_HEAD(&bs->tracked_requests, req, list);
}

S
Stefan Hajnoczi 已提交
1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237
/**
 * 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);
    }
}

1238 1239
static bool tracked_request_overlaps(BdrvTrackedRequest *req,
                                     int64_t sector_num, int nb_sectors) {
S
Stefan Hajnoczi 已提交
1240 1241 1242 1243 1244 1245 1246 1247 1248
    /*        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;
1249 1250 1251 1252 1253 1254
}

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
1255 1256
    int64_t cluster_sector_num;
    int cluster_nb_sectors;
1257 1258
    bool retry;

S
Stefan Hajnoczi 已提交
1259 1260 1261 1262 1263 1264 1265 1266 1267
    /* 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);

1268 1269 1270
    do {
        retry = false;
        QLIST_FOREACH(req, &bs->tracked_requests, list) {
S
Stefan Hajnoczi 已提交
1271 1272
            if (tracked_request_overlaps(req, cluster_sector_num,
                                         cluster_nb_sectors)) {
1273 1274 1275 1276 1277 1278
                /* 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);

1279 1280 1281 1282 1283 1284 1285 1286
                qemu_co_queue_wait(&req->wait_queue);
                retry = true;
                break;
            }
        }
    } while (retry);
}

K
Kevin Wolf 已提交
1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306
/*
 * 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;

    if (drv->bdrv_change_backing_file != NULL) {
        return drv->bdrv_change_backing_file(bs, backing_file, backing_fmt);
    } else {
        return -ENOTSUP;
    }
}

1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319
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
1320 1321 1322 1323
    if (offset < 0)
        return -EIO;

    if ((offset > len) || (len - offset < size))
1324 1325 1326 1327 1328 1329 1330 1331
        return -EIO;

    return 0;
}

static int bdrv_check_request(BlockDriverState *bs, int64_t sector_num,
                              int nb_sectors)
{
1332 1333
    return bdrv_check_byte_request(bs, sector_num * BDRV_SECTOR_SIZE,
                                   nb_sectors * BDRV_SECTOR_SIZE);
1334 1335
}

1336 1337 1338 1339 1340 1341 1342 1343 1344 1345
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
1346
{
1347
    RwCo *rwco = opaque;
B
bellard 已提交
1348

1349 1350
    if (!rwco->is_write) {
        rwco->ret = bdrv_co_do_readv(rwco->bs, rwco->sector_num,
1351
                                     rwco->nb_sectors, rwco->qiov, 0);
1352 1353
    } else {
        rwco->ret = bdrv_co_do_writev(rwco->bs, rwco->sector_num,
1354
                                      rwco->nb_sectors, rwco->qiov, 0);
1355 1356
    }
}
1357

1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377
/*
 * 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,
    };
1378

1379
    qemu_iovec_init_external(&qiov, &iov, 1);
1380

1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392
    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
1393

1394 1395 1396 1397 1398
/* 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
1399 1400
}

1401
static void set_dirty_bitmap(BlockDriverState *bs, int64_t sector_num,
1402
                             int nb_sectors, int dirty)
1403 1404
{
    int64_t start, end;
1405
    unsigned long val, idx, bit;
1406

J
Jan Kiszka 已提交
1407
    start = sector_num / BDRV_SECTORS_PER_DIRTY_CHUNK;
1408
    end = (sector_num + nb_sectors - 1) / BDRV_SECTORS_PER_DIRTY_CHUNK;
1409 1410

    for (; start <= end; start++) {
1411 1412 1413 1414
        idx = start / (sizeof(unsigned long) * 8);
        bit = start % (sizeof(unsigned long) * 8);
        val = bs->dirty_bitmap[idx];
        if (dirty) {
1415
            if (!(val & (1UL << bit))) {
1416
                bs->dirty_count++;
1417
                val |= 1UL << bit;
1418
            }
1419
        } else {
1420
            if (val & (1UL << bit)) {
1421
                bs->dirty_count--;
1422
                val &= ~(1UL << bit);
1423
            }
1424 1425
        }
        bs->dirty_bitmap[idx] = val;
1426 1427 1428
    }
}

1429
/* Return < 0 if error. Important errors are:
B
bellard 已提交
1430 1431 1432 1433 1434
  -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
*/
1435
int bdrv_write(BlockDriverState *bs, int64_t sector_num,
B
bellard 已提交
1436 1437
               const uint8_t *buf, int nb_sectors)
{
1438
    return bdrv_rw_co(bs, sector_num, (uint8_t *)buf, nb_sectors, true);
B
bellard 已提交
1439 1440
}

1441 1442
int bdrv_pread(BlockDriverState *bs, int64_t offset,
               void *buf, int count1)
B
bellard 已提交
1443
{
J
Jan Kiszka 已提交
1444
    uint8_t tmp_buf[BDRV_SECTOR_SIZE];
B
bellard 已提交
1445 1446
    int len, nb_sectors, count;
    int64_t sector_num;
1447
    int ret;
B
bellard 已提交
1448 1449 1450

    count = count1;
    /* first read to align to sector start */
J
Jan Kiszka 已提交
1451
    len = (BDRV_SECTOR_SIZE - offset) & (BDRV_SECTOR_SIZE - 1);
B
bellard 已提交
1452 1453
    if (len > count)
        len = count;
J
Jan Kiszka 已提交
1454
    sector_num = offset >> BDRV_SECTOR_BITS;
B
bellard 已提交
1455
    if (len > 0) {
1456 1457
        if ((ret = bdrv_read(bs, sector_num, tmp_buf, 1)) < 0)
            return ret;
J
Jan Kiszka 已提交
1458
        memcpy(buf, tmp_buf + (offset & (BDRV_SECTOR_SIZE - 1)), len);
B
bellard 已提交
1459 1460 1461 1462 1463 1464 1465 1466
        count -= len;
        if (count == 0)
            return count1;
        sector_num++;
        buf += len;
    }

    /* read the sectors "in place" */
J
Jan Kiszka 已提交
1467
    nb_sectors = count >> BDRV_SECTOR_BITS;
B
bellard 已提交
1468
    if (nb_sectors > 0) {
1469 1470
        if ((ret = bdrv_read(bs, sector_num, buf, nb_sectors)) < 0)
            return ret;
B
bellard 已提交
1471
        sector_num += nb_sectors;
J
Jan Kiszka 已提交
1472
        len = nb_sectors << BDRV_SECTOR_BITS;
B
bellard 已提交
1473 1474 1475 1476 1477 1478
        buf += len;
        count -= len;
    }

    /* add data from the last sector */
    if (count > 0) {
1479 1480
        if ((ret = bdrv_read(bs, sector_num, tmp_buf, 1)) < 0)
            return ret;
B
bellard 已提交
1481 1482 1483 1484 1485
        memcpy(buf, tmp_buf, count);
    }
    return count1;
}

1486 1487
int bdrv_pwrite(BlockDriverState *bs, int64_t offset,
                const void *buf, int count1)
B
bellard 已提交
1488
{
J
Jan Kiszka 已提交
1489
    uint8_t tmp_buf[BDRV_SECTOR_SIZE];
B
bellard 已提交
1490 1491
    int len, nb_sectors, count;
    int64_t sector_num;
1492
    int ret;
B
bellard 已提交
1493 1494 1495

    count = count1;
    /* first write to align to sector start */
J
Jan Kiszka 已提交
1496
    len = (BDRV_SECTOR_SIZE - offset) & (BDRV_SECTOR_SIZE - 1);
B
bellard 已提交
1497 1498
    if (len > count)
        len = count;
J
Jan Kiszka 已提交
1499
    sector_num = offset >> BDRV_SECTOR_BITS;
B
bellard 已提交
1500
    if (len > 0) {
1501 1502
        if ((ret = bdrv_read(bs, sector_num, tmp_buf, 1)) < 0)
            return ret;
J
Jan Kiszka 已提交
1503
        memcpy(tmp_buf + (offset & (BDRV_SECTOR_SIZE - 1)), buf, len);
1504 1505
        if ((ret = bdrv_write(bs, sector_num, tmp_buf, 1)) < 0)
            return ret;
B
bellard 已提交
1506 1507 1508 1509 1510 1511 1512 1513
        count -= len;
        if (count == 0)
            return count1;
        sector_num++;
        buf += len;
    }

    /* write the sectors "in place" */
J
Jan Kiszka 已提交
1514
    nb_sectors = count >> BDRV_SECTOR_BITS;
B
bellard 已提交
1515
    if (nb_sectors > 0) {
1516 1517
        if ((ret = bdrv_write(bs, sector_num, buf, nb_sectors)) < 0)
            return ret;
B
bellard 已提交
1518
        sector_num += nb_sectors;
J
Jan Kiszka 已提交
1519
        len = nb_sectors << BDRV_SECTOR_BITS;
B
bellard 已提交
1520 1521 1522 1523 1524 1525
        buf += len;
        count -= len;
    }

    /* add data from the last sector */
    if (count > 0) {
1526 1527
        if ((ret = bdrv_read(bs, sector_num, tmp_buf, 1)) < 0)
            return ret;
B
bellard 已提交
1528
        memcpy(tmp_buf, buf, count);
1529 1530
        if ((ret = bdrv_write(bs, sector_num, tmp_buf, 1)) < 0)
            return ret;
B
bellard 已提交
1531 1532 1533 1534
    }
    return count1;
}

K
Kevin Wolf 已提交
1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550
/*
 * 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;
    }

1551 1552
    /* No flush needed for cache modes that use O_DSYNC */
    if ((bs->open_flags & BDRV_O_CACHE_WB) != 0) {
K
Kevin Wolf 已提交
1553 1554 1555 1556 1557 1558
        bdrv_flush(bs);
    }

    return 0;
}

1559
static int coroutine_fn bdrv_co_do_copy_on_readv(BlockDriverState *bs,
S
Stefan Hajnoczi 已提交
1560 1561 1562 1563 1564 1565 1566 1567 1568
        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;

1569
    BlockDriver *drv = bs->drv;
S
Stefan Hajnoczi 已提交
1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582
    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);

1583 1584
    trace_bdrv_co_do_copy_on_readv(bs, sector_num, nb_sectors,
                                   cluster_sector_num, cluster_nb_sectors);
S
Stefan Hajnoczi 已提交
1585 1586 1587 1588 1589

    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);

1590 1591
    ret = drv->bdrv_co_readv(bs, cluster_sector_num, cluster_nb_sectors,
                             &bounce_qiov);
S
Stefan Hajnoczi 已提交
1592 1593 1594 1595
    if (ret < 0) {
        goto err;
    }

1596 1597 1598 1599 1600 1601
    if (drv->bdrv_co_write_zeroes &&
        buffer_is_zero(bounce_buffer, iov.iov_len)) {
        ret = drv->bdrv_co_write_zeroes(bs, cluster_sector_num,
                                        cluster_nb_sectors);
    } else {
        ret = drv->bdrv_co_writev(bs, cluster_sector_num, cluster_nb_sectors,
S
Stefan Hajnoczi 已提交
1602
                                  &bounce_qiov);
1603 1604
    }

S
Stefan Hajnoczi 已提交
1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621
    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;
    qemu_iovec_from_buffer(qiov, bounce_buffer + skip_bytes,
                           nb_sectors * BDRV_SECTOR_SIZE);

err:
    qemu_vfree(bounce_buffer);
    return ret;
}

1622 1623 1624 1625
/*
 * Handle a read request in coroutine context
 */
static int coroutine_fn bdrv_co_do_readv(BlockDriverState *bs,
1626 1627
    int64_t sector_num, int nb_sectors, QEMUIOVector *qiov,
    BdrvRequestFlags flags)
K
Kevin Wolf 已提交
1628 1629
{
    BlockDriver *drv = bs->drv;
S
Stefan Hajnoczi 已提交
1630 1631
    BdrvTrackedRequest req;
    int ret;
K
Kevin Wolf 已提交
1632 1633 1634 1635 1636 1637 1638 1639

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

Z
Zhi Yong Wu 已提交
1640 1641 1642 1643 1644
    /* throttling disk read I/O */
    if (bs->io_limits_enabled) {
        bdrv_io_limits_intercept(bs, false, nb_sectors);
    }

1645
    if (bs->copy_on_read) {
1646 1647 1648 1649 1650 1651 1652
        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) {
1653 1654 1655
        wait_for_overlapping_requests(bs, sector_num, nb_sectors);
    }

S
Stefan Hajnoczi 已提交
1656
    tracked_request_begin(&req, bs, sector_num, nb_sectors, false);
S
Stefan Hajnoczi 已提交
1657

1658
    if (flags & BDRV_REQ_COPY_ON_READ) {
S
Stefan Hajnoczi 已提交
1659 1660 1661 1662 1663 1664 1665 1666
        int pnum;

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

        if (!ret || pnum != nb_sectors) {
1667
            ret = bdrv_co_do_copy_on_readv(bs, sector_num, nb_sectors, qiov);
S
Stefan Hajnoczi 已提交
1668 1669 1670 1671
            goto out;
        }
    }

S
Stefan Hajnoczi 已提交
1672
    ret = drv->bdrv_co_readv(bs, sector_num, nb_sectors, qiov);
S
Stefan Hajnoczi 已提交
1673 1674

out:
S
Stefan Hajnoczi 已提交
1675
    tracked_request_end(&req);
1676 1677 1678 1679 1680

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

S
Stefan Hajnoczi 已提交
1681
    return ret;
K
Kevin Wolf 已提交
1682 1683
}

1684
int coroutine_fn bdrv_co_readv(BlockDriverState *bs, int64_t sector_num,
K
Kevin Wolf 已提交
1685 1686
    int nb_sectors, QEMUIOVector *qiov)
{
1687
    trace_bdrv_co_readv(bs, sector_num, nb_sectors);
K
Kevin Wolf 已提交
1688

1689 1690 1691 1692 1693 1694 1695 1696 1697 1698
    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);
1699 1700
}

1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725
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;

    /* First try the efficient write zeroes operation */
    if (drv->bdrv_co_write_zeroes) {
        return drv->bdrv_co_write_zeroes(bs, sector_num, nb_sectors);
    }

    /* 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;
}

1726 1727 1728 1729
/*
 * Handle a write request in coroutine context
 */
static int coroutine_fn bdrv_co_do_writev(BlockDriverState *bs,
1730 1731
    int64_t sector_num, int nb_sectors, QEMUIOVector *qiov,
    BdrvRequestFlags flags)
1732 1733
{
    BlockDriver *drv = bs->drv;
S
Stefan Hajnoczi 已提交
1734
    BdrvTrackedRequest req;
1735
    int ret;
K
Kevin Wolf 已提交
1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746

    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
1747 1748 1749 1750 1751
    /* throttling disk write I/O */
    if (bs->io_limits_enabled) {
        bdrv_io_limits_intercept(bs, true, nb_sectors);
    }

1752
    if (bs->copy_on_read_in_flight) {
1753 1754 1755
        wait_for_overlapping_requests(bs, sector_num, nb_sectors);
    }

S
Stefan Hajnoczi 已提交
1756 1757
    tracked_request_begin(&req, bs, sector_num, nb_sectors, true);

1758 1759 1760 1761 1762
    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);
    }
1763

K
Kevin Wolf 已提交
1764 1765 1766 1767 1768 1769 1770 1771
    if (bs->dirty_bitmap) {
        set_dirty_bitmap(bs, sector_num, nb_sectors, 1);
    }

    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
1772 1773
    tracked_request_end(&req);

1774
    return ret;
K
Kevin Wolf 已提交
1775 1776
}

1777 1778 1779 1780 1781
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);

1782 1783 1784 1785 1786 1787 1788 1789 1790 1791
    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);
1792 1793
}

B
bellard 已提交
1794 1795 1796 1797 1798 1799
/**
 * Truncate file to 'offset' bytes (needed only for file protocols)
 */
int bdrv_truncate(BlockDriverState *bs, int64_t offset)
{
    BlockDriver *drv = bs->drv;
1800
    int ret;
B
bellard 已提交
1801
    if (!drv)
B
bellard 已提交
1802
        return -ENOMEDIUM;
B
bellard 已提交
1803 1804
    if (!drv->bdrv_truncate)
        return -ENOTSUP;
1805 1806
    if (bs->read_only)
        return -EACCES;
M
Marcelo Tosatti 已提交
1807 1808
    if (bdrv_in_use(bs))
        return -EBUSY;
1809 1810 1811
    ret = drv->bdrv_truncate(bs, offset);
    if (ret == 0) {
        ret = refresh_total_sectors(bs, offset >> BDRV_SECTOR_BITS);
1812
        bdrv_dev_resize_cb(bs);
1813 1814
    }
    return ret;
B
bellard 已提交
1815 1816
}

1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835
/**
 * 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
1836 1837 1838 1839 1840 1841 1842
/**
 * 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
1843
        return -ENOMEDIUM;
1844

1845
    if (bs->growable || bdrv_dev_has_removable_media(bs)) {
1846 1847 1848
        if (drv->bdrv_getlength) {
            return drv->bdrv_getlength(bs);
        }
B
bellard 已提交
1849
    }
1850
    return bs->total_sectors * BDRV_SECTOR_SIZE;
B
bellard 已提交
1851 1852
}

B
bellard 已提交
1853
/* return 0 as number of sectors if no device present or error */
1854
void bdrv_get_geometry(BlockDriverState *bs, uint64_t *nb_sectors_ptr)
B
bellard 已提交
1855
{
B
bellard 已提交
1856 1857 1858 1859 1860
    int64_t length;
    length = bdrv_getlength(bs);
    if (length < 0)
        length = 0;
    else
J
Jan Kiszka 已提交
1861
        length = length >> BDRV_SECTOR_BITS;
B
bellard 已提交
1862
    *nb_sectors_ptr = length;
B
bellard 已提交
1863
}
B
bellard 已提交
1864

1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875
struct partition {
        uint8_t boot_ind;           /* 0x80 - active */
        uint8_t head;               /* starting head */
        uint8_t sector;             /* starting sector */
        uint8_t cyl;                /* starting cylinder */
        uint8_t sys_ind;            /* What partition type */
        uint8_t end_head;           /* end head */
        uint8_t end_sector;         /* end sector */
        uint8_t end_cyl;            /* end cylinder */
        uint32_t start_sect;        /* starting sector counting from 0 */
        uint32_t nr_sects;          /* nr of sectors in partition */
1876
} QEMU_PACKED;
1877 1878 1879 1880 1881

/* try to guess the disk logical geometry from the MSDOS partition table. Return 0 if OK, -1 if could not guess */
static int guess_disk_lchs(BlockDriverState *bs,
                           int *pcylinders, int *pheads, int *psectors)
{
1882
    uint8_t buf[BDRV_SECTOR_SIZE];
1883 1884 1885
    int ret, i, heads, sectors, cylinders;
    struct partition *p;
    uint32_t nr_sects;
B
blueswir1 已提交
1886
    uint64_t nb_sectors;
1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925

    bdrv_get_geometry(bs, &nb_sectors);

    ret = bdrv_read(bs, 0, buf, 1);
    if (ret < 0)
        return -1;
    /* test msdos magic */
    if (buf[510] != 0x55 || buf[511] != 0xaa)
        return -1;
    for(i = 0; i < 4; i++) {
        p = ((struct partition *)(buf + 0x1be)) + i;
        nr_sects = le32_to_cpu(p->nr_sects);
        if (nr_sects && p->end_head) {
            /* We make the assumption that the partition terminates on
               a cylinder boundary */
            heads = p->end_head + 1;
            sectors = p->end_sector & 63;
            if (sectors == 0)
                continue;
            cylinders = nb_sectors / (heads * sectors);
            if (cylinders < 1 || cylinders > 16383)
                continue;
            *pheads = heads;
            *psectors = sectors;
            *pcylinders = cylinders;
#if 0
            printf("guessed geometry: LCHS=%d %d %d\n",
                   cylinders, heads, sectors);
#endif
            return 0;
        }
    }
    return -1;
}

void bdrv_guess_geometry(BlockDriverState *bs, int *pcyls, int *pheads, int *psecs)
{
    int translation, lba_detected = 0;
    int cylinders, heads, secs;
B
blueswir1 已提交
1926
    uint64_t nb_sectors;
1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980

    /* if a geometry hint is available, use it */
    bdrv_get_geometry(bs, &nb_sectors);
    bdrv_get_geometry_hint(bs, &cylinders, &heads, &secs);
    translation = bdrv_get_translation_hint(bs);
    if (cylinders != 0) {
        *pcyls = cylinders;
        *pheads = heads;
        *psecs = secs;
    } else {
        if (guess_disk_lchs(bs, &cylinders, &heads, &secs) == 0) {
            if (heads > 16) {
                /* if heads > 16, it means that a BIOS LBA
                   translation was active, so the default
                   hardware geometry is OK */
                lba_detected = 1;
                goto default_geometry;
            } else {
                *pcyls = cylinders;
                *pheads = heads;
                *psecs = secs;
                /* disable any translation to be in sync with
                   the logical geometry */
                if (translation == BIOS_ATA_TRANSLATION_AUTO) {
                    bdrv_set_translation_hint(bs,
                                              BIOS_ATA_TRANSLATION_NONE);
                }
            }
        } else {
        default_geometry:
            /* if no geometry, use a standard physical disk geometry */
            cylinders = nb_sectors / (16 * 63);

            if (cylinders > 16383)
                cylinders = 16383;
            else if (cylinders < 2)
                cylinders = 2;
            *pcyls = cylinders;
            *pheads = 16;
            *psecs = 63;
            if ((lba_detected == 1) && (translation == BIOS_ATA_TRANSLATION_AUTO)) {
                if ((*pcyls * *pheads) <= 131072) {
                    bdrv_set_translation_hint(bs,
                                              BIOS_ATA_TRANSLATION_LARGE);
                } else {
                    bdrv_set_translation_hint(bs,
                                              BIOS_ATA_TRANSLATION_LBA);
                }
            }
        }
        bdrv_set_geometry_hint(bs, *pcyls, *pheads, *psecs);
    }
}

1981
void bdrv_set_geometry_hint(BlockDriverState *bs,
B
bellard 已提交
1982 1983 1984 1985 1986 1987 1988
                            int cyls, int heads, int secs)
{
    bs->cyls = cyls;
    bs->heads = heads;
    bs->secs = secs;
}

B
bellard 已提交
1989 1990 1991 1992 1993
void bdrv_set_translation_hint(BlockDriverState *bs, int translation)
{
    bs->translation = translation;
}

1994
void bdrv_get_geometry_hint(BlockDriverState *bs,
B
bellard 已提交
1995 1996 1997 1998 1999 2000 2001
                            int *pcyls, int *pheads, int *psecs)
{
    *pcyls = bs->cyls;
    *pheads = bs->heads;
    *psecs = bs->secs;
}

2002 2003 2004 2005 2006 2007 2008 2009
/* 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);
}

2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112
/* Recognize floppy formats */
typedef struct FDFormat {
    FDriveType drive;
    uint8_t last_sect;
    uint8_t max_track;
    uint8_t max_head;
} FDFormat;

static const FDFormat fd_formats[] = {
    /* First entry is default format */
    /* 1.44 MB 3"1/2 floppy disks */
    { FDRIVE_DRV_144, 18, 80, 1, },
    { FDRIVE_DRV_144, 20, 80, 1, },
    { FDRIVE_DRV_144, 21, 80, 1, },
    { FDRIVE_DRV_144, 21, 82, 1, },
    { FDRIVE_DRV_144, 21, 83, 1, },
    { FDRIVE_DRV_144, 22, 80, 1, },
    { FDRIVE_DRV_144, 23, 80, 1, },
    { FDRIVE_DRV_144, 24, 80, 1, },
    /* 2.88 MB 3"1/2 floppy disks */
    { FDRIVE_DRV_288, 36, 80, 1, },
    { FDRIVE_DRV_288, 39, 80, 1, },
    { FDRIVE_DRV_288, 40, 80, 1, },
    { FDRIVE_DRV_288, 44, 80, 1, },
    { FDRIVE_DRV_288, 48, 80, 1, },
    /* 720 kB 3"1/2 floppy disks */
    { FDRIVE_DRV_144,  9, 80, 1, },
    { FDRIVE_DRV_144, 10, 80, 1, },
    { FDRIVE_DRV_144, 10, 82, 1, },
    { FDRIVE_DRV_144, 10, 83, 1, },
    { FDRIVE_DRV_144, 13, 80, 1, },
    { FDRIVE_DRV_144, 14, 80, 1, },
    /* 1.2 MB 5"1/4 floppy disks */
    { FDRIVE_DRV_120, 15, 80, 1, },
    { FDRIVE_DRV_120, 18, 80, 1, },
    { FDRIVE_DRV_120, 18, 82, 1, },
    { FDRIVE_DRV_120, 18, 83, 1, },
    { FDRIVE_DRV_120, 20, 80, 1, },
    /* 720 kB 5"1/4 floppy disks */
    { FDRIVE_DRV_120,  9, 80, 1, },
    { FDRIVE_DRV_120, 11, 80, 1, },
    /* 360 kB 5"1/4 floppy disks */
    { FDRIVE_DRV_120,  9, 40, 1, },
    { FDRIVE_DRV_120,  9, 40, 0, },
    { FDRIVE_DRV_120, 10, 41, 1, },
    { FDRIVE_DRV_120, 10, 42, 1, },
    /* 320 kB 5"1/4 floppy disks */
    { FDRIVE_DRV_120,  8, 40, 1, },
    { FDRIVE_DRV_120,  8, 40, 0, },
    /* 360 kB must match 5"1/4 better than 3"1/2... */
    { FDRIVE_DRV_144,  9, 80, 0, },
    /* end */
    { FDRIVE_DRV_NONE, -1, -1, 0, },
};

void bdrv_get_floppy_geometry_hint(BlockDriverState *bs, int *nb_heads,
                                   int *max_track, int *last_sect,
                                   FDriveType drive_in, FDriveType *drive)
{
    const FDFormat *parse;
    uint64_t nb_sectors, size;
    int i, first_match, match;

    bdrv_get_geometry_hint(bs, nb_heads, max_track, last_sect);
    if (*nb_heads != 0 && *max_track != 0 && *last_sect != 0) {
        /* User defined disk */
    } else {
        bdrv_get_geometry(bs, &nb_sectors);
        match = -1;
        first_match = -1;
        for (i = 0; ; i++) {
            parse = &fd_formats[i];
            if (parse->drive == FDRIVE_DRV_NONE) {
                break;
            }
            if (drive_in == parse->drive ||
                drive_in == FDRIVE_DRV_NONE) {
                size = (parse->max_head + 1) * parse->max_track *
                    parse->last_sect;
                if (nb_sectors == size) {
                    match = i;
                    break;
                }
                if (first_match == -1) {
                    first_match = i;
                }
            }
        }
        if (match == -1) {
            if (first_match == -1) {
                match = 1;
            } else {
                match = first_match;
            }
            parse = &fd_formats[match];
        }
        *nb_heads = parse->max_head + 1;
        *max_track = parse->max_track;
        *last_sect = parse->last_sect;
        *drive = parse->drive;
    }
}

B
bellard 已提交
2113 2114 2115 2116 2117
int bdrv_get_translation_hint(BlockDriverState *bs)
{
    return bs->translation;
}

2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129
void bdrv_set_on_error(BlockDriverState *bs, BlockErrorAction on_read_error,
                       BlockErrorAction on_write_error)
{
    bs->on_read_error = on_read_error;
    bs->on_write_error = on_write_error;
}

BlockErrorAction bdrv_get_on_error(BlockDriverState *bs, int is_read)
{
    return is_read ? bs->on_read_error : bs->on_write_error;
}

B
bellard 已提交
2130 2131 2132 2133 2134
int bdrv_is_read_only(BlockDriverState *bs)
{
    return bs->read_only;
}

2135 2136 2137 2138 2139
int bdrv_is_sg(BlockDriverState *bs)
{
    return bs->sg;
}

2140 2141 2142 2143 2144
int bdrv_enable_write_cache(BlockDriverState *bs)
{
    return bs->enable_write_cache;
}

B
bellard 已提交
2145 2146 2147 2148 2149 2150 2151
int bdrv_is_encrypted(BlockDriverState *bs)
{
    if (bs->backing_hd && bs->backing_hd->encrypted)
        return 1;
    return bs->encrypted;
}

2152 2153 2154 2155 2156 2157 2158 2159 2160
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
2161 2162 2163 2164 2165 2166 2167 2168 2169 2170
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;
    }
2171 2172 2173 2174 2175
    if (!bs->encrypted) {
        return -EINVAL;
    } else if (!bs->drv || !bs->drv->bdrv_set_key) {
        return -ENOMEDIUM;
    }
2176
    ret = bs->drv->bdrv_set_key(bs, key);
2177 2178 2179 2180 2181
    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 */
2182
        bdrv_dev_change_media_cb(bs, true);
2183
    }
2184
    return ret;
B
bellard 已提交
2185 2186 2187 2188
}

void bdrv_get_format(BlockDriverState *bs, char *buf, int buf_size)
{
B
bellard 已提交
2189
    if (!bs->drv) {
B
bellard 已提交
2190 2191 2192 2193 2194 2195
        buf[0] = '\0';
    } else {
        pstrcpy(buf, buf_size, bs->drv->format_name);
    }
}

2196
void bdrv_iterate_format(void (*it)(void *opaque, const char *name),
B
bellard 已提交
2197 2198 2199 2200
                         void *opaque)
{
    BlockDriver *drv;

2201
    QLIST_FOREACH(drv, &bdrv_drivers, list) {
B
bellard 已提交
2202 2203 2204 2205
        it(opaque, drv->format_name);
    }
}

B
bellard 已提交
2206 2207 2208 2209
BlockDriverState *bdrv_find(const char *name)
{
    BlockDriverState *bs;

2210 2211
    QTAILQ_FOREACH(bs, &bdrv_states, list) {
        if (!strcmp(name, bs->device_name)) {
B
bellard 已提交
2212
            return bs;
2213
        }
B
bellard 已提交
2214 2215 2216 2217
    }
    return NULL;
}

M
Markus Armbruster 已提交
2218 2219 2220 2221 2222 2223 2224 2225
BlockDriverState *bdrv_next(BlockDriverState *bs)
{
    if (!bs) {
        return QTAILQ_FIRST(&bdrv_states);
    }
    return QTAILQ_NEXT(bs, list);
}

2226
void bdrv_iterate(void (*it)(void *opaque, BlockDriverState *bs), void *opaque)
B
bellard 已提交
2227 2228 2229
{
    BlockDriverState *bs;

2230
    QTAILQ_FOREACH(bs, &bdrv_states, list) {
2231
        it(opaque, bs);
B
bellard 已提交
2232 2233 2234
    }
}

B
bellard 已提交
2235 2236 2237 2238 2239
const char *bdrv_get_device_name(BlockDriverState *bs)
{
    return bs->device_name;
}

A
aliguori 已提交
2240 2241 2242 2243
void bdrv_flush_all(void)
{
    BlockDriverState *bs;

2244
    QTAILQ_FOREACH(bs, &bdrv_states, list) {
2245
        if (!bdrv_is_read_only(bs) && bdrv_is_inserted(bs)) {
A
aliguori 已提交
2246
            bdrv_flush(bs);
2247 2248
        }
    }
A
aliguori 已提交
2249 2250
}

K
Kevin Wolf 已提交
2251 2252 2253 2254
int bdrv_has_zero_init(BlockDriverState *bs)
{
    assert(bs->drv);

K
Kevin Wolf 已提交
2255 2256
    if (bs->drv->bdrv_has_zero_init) {
        return bs->drv->bdrv_has_zero_init(bs);
K
Kevin Wolf 已提交
2257 2258 2259 2260 2261
    }

    return 1;
}

2262 2263 2264 2265 2266 2267 2268 2269 2270
typedef struct BdrvCoIsAllocatedData {
    BlockDriverState *bs;
    int64_t sector_num;
    int nb_sectors;
    int *pnum;
    int ret;
    bool done;
} BdrvCoIsAllocatedData;

2271 2272 2273 2274 2275
/*
 * 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.
 *
2276 2277 2278
 * If 'sector_num' is beyond the end of the disk image the return value is 0
 * and 'pnum' is set to 0.
 *
2279 2280 2281 2282
 * '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.
 *
2283 2284
 * '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.
2285
 */
2286 2287
int coroutine_fn bdrv_co_is_allocated(BlockDriverState *bs, int64_t sector_num,
                                      int nb_sectors, int *pnum)
2288
{
2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300
    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;
    }

2301
    if (!bs->drv->bdrv_co_is_allocated) {
2302
        *pnum = nb_sectors;
2303 2304
        return 1;
    }
2305

2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327
    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)
{
2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342
    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;
2343 2344
}

L
Luiz Capitulino 已提交
2345
BlockInfoList *qmp_query_block(Error **errp)
B
bellard 已提交
2346
{
L
Luiz Capitulino 已提交
2347
    BlockInfoList *head = NULL, *cur_item = NULL;
B
bellard 已提交
2348 2349
    BlockDriverState *bs;

2350
    QTAILQ_FOREACH(bs, &bdrv_states, list) {
L
Luiz Capitulino 已提交
2351
        BlockInfoList *info = g_malloc0(sizeof(*info));
2352

L
Luiz Capitulino 已提交
2353 2354 2355 2356 2357
        info->value = g_malloc0(sizeof(*info->value));
        info->value->device = g_strdup(bs->device_name);
        info->value->type = g_strdup("unknown");
        info->value->locked = bdrv_dev_is_medium_locked(bs);
        info->value->removable = bdrv_dev_has_removable_media(bs);
2358

2359
        if (bdrv_dev_has_removable_media(bs)) {
L
Luiz Capitulino 已提交
2360 2361
            info->value->has_tray_open = true;
            info->value->tray_open = bdrv_dev_is_tray_open(bs);
2362
        }
2363 2364

        if (bdrv_iostatus_is_enabled(bs)) {
L
Luiz Capitulino 已提交
2365 2366
            info->value->has_io_status = true;
            info->value->io_status = bs->iostatus;
2367 2368
        }

B
bellard 已提交
2369
        if (bs->drv) {
L
Luiz Capitulino 已提交
2370 2371 2372 2373 2374 2375 2376 2377 2378
            info->value->has_inserted = true;
            info->value->inserted = g_malloc0(sizeof(*info->value->inserted));
            info->value->inserted->file = g_strdup(bs->filename);
            info->value->inserted->ro = bs->read_only;
            info->value->inserted->drv = g_strdup(bs->drv->format_name);
            info->value->inserted->encrypted = bs->encrypted;
            if (bs->backing_file[0]) {
                info->value->inserted->has_backing_file = true;
                info->value->inserted->backing_file = g_strdup(bs->backing_file);
A
aliguori 已提交
2379
            }
Z
Zhi Yong Wu 已提交
2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394

            if (bs->io_limits_enabled) {
                info->value->inserted->bps =
                               bs->io_limits.bps[BLOCK_IO_LIMIT_TOTAL];
                info->value->inserted->bps_rd =
                               bs->io_limits.bps[BLOCK_IO_LIMIT_READ];
                info->value->inserted->bps_wr =
                               bs->io_limits.bps[BLOCK_IO_LIMIT_WRITE];
                info->value->inserted->iops =
                               bs->io_limits.iops[BLOCK_IO_LIMIT_TOTAL];
                info->value->inserted->iops_rd =
                               bs->io_limits.iops[BLOCK_IO_LIMIT_READ];
                info->value->inserted->iops_wr =
                               bs->io_limits.iops[BLOCK_IO_LIMIT_WRITE];
            }
L
Luiz Capitulino 已提交
2395
        }
2396

L
Luiz Capitulino 已提交
2397 2398 2399 2400 2401 2402
        /* XXX: waiting for the qapi to support GSList */
        if (!cur_item) {
            head = cur_item = info;
        } else {
            cur_item->next = info;
            cur_item = info;
B
bellard 已提交
2403 2404
        }
    }
2405

L
Luiz Capitulino 已提交
2406
    return head;
B
bellard 已提交
2407
}
2408

L
Luiz Capitulino 已提交
2409 2410 2411 2412 2413 2414 2415 2416 2417 2418
/* Consider exposing this as a full fledged QMP command */
static BlockStats *qmp_query_blockstat(const BlockDriverState *bs, Error **errp)
{
    BlockStats *s;

    s = g_malloc0(sizeof(*s));

    if (bs->device_name[0]) {
        s->has_device = true;
        s->device = g_strdup(bs->device_name);
2419 2420
    }

L
Luiz Capitulino 已提交
2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431
    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];

2432
    if (bs->file) {
L
Luiz Capitulino 已提交
2433 2434
        s->has_parent = true;
        s->parent = qmp_query_blockstat(bs->file, NULL);
2435 2436
    }

L
Luiz Capitulino 已提交
2437
    return s;
2438 2439
}

L
Luiz Capitulino 已提交
2440
BlockStatsList *qmp_query_blockstats(Error **errp)
2441
{
L
Luiz Capitulino 已提交
2442
    BlockStatsList *head = NULL, *cur_item = NULL;
2443 2444
    BlockDriverState *bs;

2445
    QTAILQ_FOREACH(bs, &bdrv_states, list) {
L
Luiz Capitulino 已提交
2446 2447 2448 2449 2450 2451 2452 2453 2454 2455
        BlockStatsList *info = g_malloc0(sizeof(*info));
        info->value = qmp_query_blockstat(bs, NULL);

        /* XXX: waiting for the qapi to support GSList */
        if (!cur_item) {
            head = cur_item = info;
        } else {
            cur_item->next = info;
            cur_item = info;
        }
2456
    }
2457

L
Luiz Capitulino 已提交
2458
    return head;
2459
}
B
bellard 已提交
2460

2461 2462 2463 2464 2465 2466 2467 2468 2469 2470
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;
}

2471
void bdrv_get_backing_filename(BlockDriverState *bs,
B
bellard 已提交
2472 2473
                               char *filename, int filename_size)
{
K
Kevin Wolf 已提交
2474
    pstrcpy(filename, filename_size, bs->backing_file);
B
bellard 已提交
2475 2476
}

2477
int bdrv_write_compressed(BlockDriverState *bs, int64_t sector_num,
B
bellard 已提交
2478 2479 2480 2481
                          const uint8_t *buf, int nb_sectors)
{
    BlockDriver *drv = bs->drv;
    if (!drv)
B
bellard 已提交
2482
        return -ENOMEDIUM;
B
bellard 已提交
2483 2484
    if (!drv->bdrv_write_compressed)
        return -ENOTSUP;
K
Kevin Wolf 已提交
2485 2486
    if (bdrv_check_request(bs, sector_num, nb_sectors))
        return -EIO;
2487

2488
    if (bs->dirty_bitmap) {
2489 2490
        set_dirty_bitmap(bs, sector_num, nb_sectors, 1);
    }
2491

B
bellard 已提交
2492 2493
    return drv->bdrv_write_compressed(bs, sector_num, buf, nb_sectors);
}
2494

B
bellard 已提交
2495 2496 2497 2498
int bdrv_get_info(BlockDriverState *bs, BlockDriverInfo *bdi)
{
    BlockDriver *drv = bs->drv;
    if (!drv)
B
bellard 已提交
2499
        return -ENOMEDIUM;
B
bellard 已提交
2500 2501 2502 2503 2504 2505
    if (!drv->bdrv_get_info)
        return -ENOTSUP;
    memset(bdi, 0, sizeof(*bdi));
    return drv->bdrv_get_info(bs, bdi);
}

2506 2507
int bdrv_save_vmstate(BlockDriverState *bs, const uint8_t *buf,
                      int64_t pos, int size)
2508 2509 2510 2511
{
    BlockDriver *drv = bs->drv;
    if (!drv)
        return -ENOMEDIUM;
2512 2513 2514 2515 2516
    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;
2517 2518
}

2519 2520
int bdrv_load_vmstate(BlockDriverState *bs, uint8_t *buf,
                      int64_t pos, int size)
2521 2522 2523 2524
{
    BlockDriver *drv = bs->drv;
    if (!drv)
        return -ENOMEDIUM;
2525 2526 2527 2528 2529
    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;
2530 2531
}

K
Kevin Wolf 已提交
2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543
void bdrv_debug_event(BlockDriverState *bs, BlkDebugEvent event)
{
    BlockDriver *drv = bs->drv;

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

    return drv->bdrv_debug_event(bs, event);

}

B
bellard 已提交
2544 2545 2546
/**************************************************************/
/* handling of snapshots */

2547 2548 2549
int bdrv_can_snapshot(BlockDriverState *bs)
{
    BlockDriver *drv = bs->drv;
2550
    if (!drv || !bdrv_is_inserted(bs) || bdrv_is_read_only(bs)) {
2551 2552 2553 2554 2555 2556 2557 2558 2559 2560 2561 2562 2563
        return 0;
    }

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

    return 1;
}

2564 2565 2566 2567 2568
int bdrv_is_snapshot(BlockDriverState *bs)
{
    return !!(bs->open_flags & BDRV_O_SNAPSHOT);
}

2569 2570 2571 2572
BlockDriverState *bdrv_snapshots(void)
{
    BlockDriverState *bs;

2573
    if (bs_snapshots) {
2574
        return bs_snapshots;
2575
    }
2576 2577 2578 2579

    bs = NULL;
    while ((bs = bdrv_next(bs))) {
        if (bdrv_can_snapshot(bs)) {
2580 2581
            bs_snapshots = bs;
            return bs;
2582 2583 2584 2585 2586
        }
    }
    return NULL;
}

2587
int bdrv_snapshot_create(BlockDriverState *bs,
B
bellard 已提交
2588 2589 2590 2591
                         QEMUSnapshotInfo *sn_info)
{
    BlockDriver *drv = bs->drv;
    if (!drv)
B
bellard 已提交
2592
        return -ENOMEDIUM;
2593 2594 2595 2596 2597
    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
2598 2599
}

2600
int bdrv_snapshot_goto(BlockDriverState *bs,
B
bellard 已提交
2601 2602 2603
                       const char *snapshot_id)
{
    BlockDriver *drv = bs->drv;
2604 2605
    int ret, open_ret;

B
bellard 已提交
2606
    if (!drv)
B
bellard 已提交
2607
        return -ENOMEDIUM;
2608 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622 2623
    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
2624 2625 2626 2627 2628 2629
}

int bdrv_snapshot_delete(BlockDriverState *bs, const char *snapshot_id)
{
    BlockDriver *drv = bs->drv;
    if (!drv)
B
bellard 已提交
2630
        return -ENOMEDIUM;
2631 2632 2633 2634 2635
    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
2636 2637
}

2638
int bdrv_snapshot_list(BlockDriverState *bs,
B
bellard 已提交
2639 2640 2641 2642
                       QEMUSnapshotInfo **psn_info)
{
    BlockDriver *drv = bs->drv;
    if (!drv)
B
bellard 已提交
2643
        return -ENOMEDIUM;
2644 2645 2646 2647 2648
    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
2649 2650
}

E
edison 已提交
2651 2652 2653 2654 2655 2656 2657 2658 2659 2660 2661 2662 2663 2664 2665 2666
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;
}

2667 2668 2669 2670 2671 2672 2673 2674 2675 2676 2677 2678 2679 2680 2681 2682 2683 2684
BlockDriverState *bdrv_find_backing_image(BlockDriverState *bs,
        const char *backing_file)
{
    if (!bs->drv) {
        return NULL;
    }

    if (bs->backing_hd) {
        if (strcmp(bs->backing_file, backing_file) == 0) {
            return bs->backing_hd;
        } else {
            return bdrv_find_backing_image(bs->backing_hd, backing_file);
        }
    }

    return NULL;
}

B
bellard 已提交
2685 2686 2687 2688 2689 2690 2691 2692 2693 2694 2695 2696 2697 2698
#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)) {
2699
                snprintf(buf, buf_size, "%0.1f%c",
B
bellard 已提交
2700 2701 2702 2703
                         (double)size / base,
                         suffixes[i]);
                break;
            } else if (size < (1000 * base) || i == (NB_SUFFIXES - 1)) {
2704
                snprintf(buf, buf_size, "%" PRId64 "%c",
B
bellard 已提交
2705 2706 2707 2708 2709 2710 2711 2712 2713 2714 2715 2716 2717
                         ((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];
2718 2719 2720
#ifdef _WIN32
    struct tm *ptm;
#else
B
bellard 已提交
2721
    struct tm tm;
2722
#endif
B
bellard 已提交
2723 2724 2725 2726
    time_t ti;
    int64_t secs;

    if (!sn) {
2727 2728
        snprintf(buf, buf_size,
                 "%-10s%-20s%7s%20s%15s",
B
bellard 已提交
2729 2730 2731
                 "ID", "TAG", "VM SIZE", "DATE", "VM CLOCK");
    } else {
        ti = sn->date_sec;
2732 2733 2734 2735 2736
#ifdef _WIN32
        ptm = localtime(&ti);
        strftime(date_buf, sizeof(date_buf),
                 "%Y-%m-%d %H:%M:%S", ptm);
#else
B
bellard 已提交
2737 2738 2739
        localtime_r(&ti, &tm);
        strftime(date_buf, sizeof(date_buf),
                 "%Y-%m-%d %H:%M:%S", &tm);
2740
#endif
B
bellard 已提交
2741 2742 2743 2744 2745
        secs = sn->vm_clock_nsec / 1000000000;
        snprintf(clock_buf, sizeof(clock_buf),
                 "%02d:%02d:%02d.%03d",
                 (int)(secs / 3600),
                 (int)((secs / 60) % 60),
2746
                 (int)(secs % 60),
B
bellard 已提交
2747 2748
                 (int)((sn->vm_clock_nsec / 1000000) % 1000));
        snprintf(buf, buf_size,
2749
                 "%-10s%-20s%7s%20s%15s",
B
bellard 已提交
2750 2751 2752 2753 2754 2755 2756 2757
                 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
2758
/**************************************************************/
B
bellard 已提交
2759
/* async I/Os */
B
bellard 已提交
2760

2761
BlockDriverAIOCB *bdrv_aio_readv(BlockDriverState *bs, int64_t sector_num,
2762
                                 QEMUIOVector *qiov, int nb_sectors,
2763
                                 BlockDriverCompletionFunc *cb, void *opaque)
B
bellard 已提交
2764
{
2765 2766
    trace_bdrv_aio_readv(bs, sector_num, nb_sectors, opaque);

2767
    return bdrv_co_aio_rw_vector(bs, sector_num, qiov, nb_sectors,
2768
                                 cb, opaque, false);
B
bellard 已提交
2769 2770
}

2771 2772 2773
BlockDriverAIOCB *bdrv_aio_writev(BlockDriverState *bs, int64_t sector_num,
                                  QEMUIOVector *qiov, int nb_sectors,
                                  BlockDriverCompletionFunc *cb, void *opaque)
B
bellard 已提交
2774
{
2775 2776
    trace_bdrv_aio_writev(bs, sector_num, nb_sectors, opaque);

2777
    return bdrv_co_aio_rw_vector(bs, sector_num, qiov, nb_sectors,
2778
                                 cb, opaque, true);
B
bellard 已提交
2779 2780
}

K
Kevin Wolf 已提交
2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799

typedef struct MultiwriteCB {
    int error;
    int num_requests;
    int num_callbacks;
    struct {
        BlockDriverCompletionFunc *cb;
        void *opaque;
        QEMUIOVector *free_qiov;
        void *free_buf;
    } 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);
2800 2801 2802
        if (mcb->callbacks[i].free_qiov) {
            qemu_iovec_destroy(mcb->callbacks[i].free_qiov);
        }
2803
        g_free(mcb->callbacks[i].free_qiov);
2804
        qemu_vfree(mcb->callbacks[i].free_buf);
K
Kevin Wolf 已提交
2805 2806 2807 2808 2809 2810 2811
    }
}

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

2812 2813
    trace_multiwrite_cb(mcb, ret);

2814
    if (ret < 0 && !mcb->error) {
K
Kevin Wolf 已提交
2815 2816 2817 2818 2819
        mcb->error = ret;
    }

    mcb->num_requests--;
    if (mcb->num_requests == 0) {
2820
        multiwrite_user_cb(mcb);
2821
        g_free(mcb);
K
Kevin Wolf 已提交
2822 2823 2824 2825 2826
    }
}

static int multiwrite_req_compare(const void *a, const void *b)
{
2827 2828 2829 2830 2831 2832 2833 2834 2835 2836 2837 2838 2839
    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
2840 2841 2842 2843 2844 2845 2846 2847 2848 2849 2850 2851 2852 2853 2854 2855 2856 2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 2867 2868 2869 2870 2871 2872 2873 2874
}

/*
 * 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;

        // This handles the cases that are valid for all block drivers, namely
        // exactly sequential writes and overlapping writes.
        if (reqs[i].sector <= oldreq_last) {
            merge = 1;
        }

        // The block driver may decide that it makes sense to combine requests
        // even if there is a gap of some sectors between them. In this case,
        // the gap is filled with zeros (therefore only applicable for yet
        // unused space in format like qcow2).
        if (!merge && bs->drv->bdrv_merge_requests) {
            merge = bs->drv->bdrv_merge_requests(bs, &reqs[outidx], &reqs[i]);
        }

2875 2876 2877 2878
        if (reqs[outidx].qiov->niov + reqs[i].qiov->niov + 1 > IOV_MAX) {
            merge = 0;
        }

K
Kevin Wolf 已提交
2879 2880
        if (merge) {
            size_t size;
2881
            QEMUIOVector *qiov = g_malloc0(sizeof(*qiov));
K
Kevin Wolf 已提交
2882 2883 2884 2885 2886 2887 2888 2889 2890 2891 2892 2893 2894 2895 2896 2897 2898 2899 2900 2901
            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;
            qemu_iovec_concat(qiov, reqs[outidx].qiov, size);

            // We might need to add some zeros between the two requests
            if (reqs[i].sector > oldreq_last) {
                size_t zero_bytes = (reqs[i].sector - oldreq_last) << 9;
                uint8_t *buf = qemu_blockalign(bs, zero_bytes);
                memset(buf, 0, zero_bytes);
                qemu_iovec_add(qiov, buf, zero_bytes);
                mcb->callbacks[i].free_buf = buf;
            }

            // Add the second request
            qemu_iovec_concat(qiov, reqs[i].qiov, reqs[i].qiov->size);

2902
            reqs[outidx].nb_sectors = qiov->size >> 9;
K
Kevin Wolf 已提交
2903 2904 2905 2906 2907 2908 2909 2910 2911 2912 2913 2914 2915 2916 2917 2918 2919 2920 2921 2922 2923 2924 2925 2926 2927 2928 2929 2930 2931 2932 2933 2934 2935
            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;

2936 2937 2938 2939 2940 2941 2942 2943
    /* 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
2944 2945 2946 2947 2948
    if (num_reqs == 0) {
        return 0;
    }

    // Create MultiwriteCB structure
2949
    mcb = g_malloc0(sizeof(*mcb) + num_reqs * sizeof(*mcb->callbacks));
K
Kevin Wolf 已提交
2950 2951 2952 2953 2954 2955 2956 2957 2958 2959 2960
    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);

2961 2962
    trace_bdrv_aio_multiwrite(mcb, mcb->num_callbacks, num_reqs);

2963 2964
    /* Run the aio requests. */
    mcb->num_requests = num_reqs;
K
Kevin Wolf 已提交
2965
    for (i = 0; i < num_reqs; i++) {
2966
        bdrv_aio_writev(bs, reqs[i].sector, reqs[i].qiov,
K
Kevin Wolf 已提交
2967 2968 2969 2970 2971 2972
            reqs[i].nb_sectors, multiwrite_cb, mcb);
    }

    return 0;
}

B
bellard 已提交
2973 2974
void bdrv_aio_cancel(BlockDriverAIOCB *acb)
{
2975
    acb->pool->cancel(acb);
B
bellard 已提交
2976 2977
}

Z
Zhi Yong Wu 已提交
2978 2979 2980 2981 2982 2983 2984 2985 2986 2987 2988 2989 2990 2991 2992 2993 2994 2995 2996 2997 2998 2999 3000 3001 3002 3003 3004 3005 3006 3007 3008 3009 3010 3011 3012 3013 3014 3015 3016 3017 3018 3019 3020 3021 3022 3023 3024 3025 3026 3027 3028 3029 3030 3031 3032 3033 3034 3035 3036 3037 3038 3039 3040 3041 3042 3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054 3055 3056 3057 3058 3059 3060 3061 3062 3063 3064 3065 3066 3067 3068 3069 3070 3071 3072 3073 3074 3075 3076 3077 3078 3079 3080 3081 3082 3083 3084 3085 3086 3087 3088 3089 3090 3091 3092 3093 3094 3095 3096 3097 3098 3099 3100 3101 3102 3103 3104 3105 3106 3107 3108 3109 3110 3111 3112 3113 3114 3115 3116 3117 3118 3119 3120 3121 3122 3123 3124 3125 3126 3127 3128 3129 3130 3131 3132 3133 3134 3135 3136 3137 3138 3139 3140 3141
/* 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;
}
3142

B
bellard 已提交
3143 3144 3145
/**************************************************************/
/* async block device emulation */

3146 3147 3148 3149 3150 3151 3152 3153 3154 3155 3156 3157
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)
{
3158 3159
    BlockDriverAIOCBSync *acb =
        container_of(blockacb, BlockDriverAIOCBSync, common);
3160
    qemu_bh_delete(acb->bh);
A
Avi Kivity 已提交
3161
    acb->bh = NULL;
3162 3163 3164 3165 3166 3167 3168 3169
    qemu_aio_release(acb);
}

static AIOPool bdrv_em_aio_pool = {
    .aiocb_size         = sizeof(BlockDriverAIOCBSync),
    .cancel             = bdrv_aio_cancel_em,
};

3170
static void bdrv_aio_bh_cb(void *opaque)
B
bellard 已提交
3171
{
3172
    BlockDriverAIOCBSync *acb = opaque;
3173 3174 3175

    if (!acb->is_write)
        qemu_iovec_from_buffer(acb->qiov, acb->bounce, acb->qiov->size);
3176
    qemu_vfree(acb->bounce);
3177
    acb->common.cb(acb->common.opaque, acb->ret);
3178
    qemu_bh_delete(acb->bh);
A
Avi Kivity 已提交
3179
    acb->bh = NULL;
3180
    qemu_aio_release(acb);
B
bellard 已提交
3181
}
3182

3183 3184 3185 3186 3187 3188 3189 3190
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
3191
{
3192 3193
    BlockDriverAIOCBSync *acb;

3194
    acb = qemu_aio_get(&bdrv_em_aio_pool, bs, cb, opaque);
3195 3196
    acb->is_write = is_write;
    acb->qiov = qiov;
3197
    acb->bounce = qemu_blockalign(bs, qiov->size);
3198
    acb->bh = qemu_bh_new(bdrv_aio_bh_cb, acb);
3199 3200 3201

    if (is_write) {
        qemu_iovec_to_buffer(acb->qiov, acb->bounce);
3202
        acb->ret = bs->drv->bdrv_write(bs, sector_num, acb->bounce, nb_sectors);
3203
    } else {
3204
        acb->ret = bs->drv->bdrv_read(bs, sector_num, acb->bounce, nb_sectors);
3205 3206
    }

3207
    qemu_bh_schedule(acb->bh);
3208

3209
    return &acb->common;
3210 3211
}

3212 3213
static BlockDriverAIOCB *bdrv_aio_readv_em(BlockDriverState *bs,
        int64_t sector_num, QEMUIOVector *qiov, int nb_sectors,
3214
        BlockDriverCompletionFunc *cb, void *opaque)
3215
{
3216 3217
    return bdrv_aio_rw_vector(bs, sector_num, qiov, nb_sectors, cb, opaque, 0);
}
B
bellard 已提交
3218

3219 3220 3221 3222 3223
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);
3224 3225
}

3226 3227 3228 3229 3230 3231 3232 3233 3234 3235 3236 3237 3238 3239 3240 3241 3242 3243

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();
}

static AIOPool bdrv_em_co_aio_pool = {
    .aiocb_size         = sizeof(BlockDriverAIOCBCoroutine),
    .cancel             = bdrv_aio_co_cancel_em,
};

P
Paolo Bonzini 已提交
3244
static void bdrv_co_em_bh(void *opaque)
3245 3246 3247 3248 3249 3250 3251 3252
{
    BlockDriverAIOCBCoroutine *acb = opaque;

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

3253 3254 3255 3256 3257 3258 3259 3260
/* 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,
3261
            acb->req.nb_sectors, acb->req.qiov, 0);
3262 3263
    } else {
        acb->req.error = bdrv_co_do_writev(bs, acb->req.sector,
3264
            acb->req.nb_sectors, acb->req.qiov, 0);
3265 3266
    }

P
Paolo Bonzini 已提交
3267
    acb->bh = qemu_bh_new(bdrv_co_em_bh, acb);
3268 3269 3270
    qemu_bh_schedule(acb->bh);
}

3271 3272 3273 3274 3275 3276
static BlockDriverAIOCB *bdrv_co_aio_rw_vector(BlockDriverState *bs,
                                               int64_t sector_num,
                                               QEMUIOVector *qiov,
                                               int nb_sectors,
                                               BlockDriverCompletionFunc *cb,
                                               void *opaque,
3277
                                               bool is_write)
3278 3279 3280 3281 3282 3283 3284 3285 3286 3287
{
    Coroutine *co;
    BlockDriverAIOCBCoroutine *acb;

    acb = qemu_aio_get(&bdrv_em_co_aio_pool, bs, cb, opaque);
    acb->req.sector = sector_num;
    acb->req.nb_sectors = nb_sectors;
    acb->req.qiov = qiov;
    acb->is_write = is_write;

3288
    co = qemu_coroutine_create(bdrv_co_do_rw);
3289 3290 3291 3292 3293
    qemu_coroutine_enter(co, acb);

    return &acb->common;
}

P
Paolo Bonzini 已提交
3294
static void coroutine_fn bdrv_aio_flush_co_entry(void *opaque)
3295
{
P
Paolo Bonzini 已提交
3296 3297
    BlockDriverAIOCBCoroutine *acb = opaque;
    BlockDriverState *bs = acb->common.bs;
3298

P
Paolo Bonzini 已提交
3299 3300
    acb->req.error = bdrv_co_flush(bs);
    acb->bh = qemu_bh_new(bdrv_co_em_bh, acb);
3301 3302 3303
    qemu_bh_schedule(acb->bh);
}

P
Paolo Bonzini 已提交
3304
BlockDriverAIOCB *bdrv_aio_flush(BlockDriverState *bs,
3305 3306
        BlockDriverCompletionFunc *cb, void *opaque)
{
P
Paolo Bonzini 已提交
3307
    trace_bdrv_aio_flush(bs, opaque);
3308

P
Paolo Bonzini 已提交
3309 3310
    Coroutine *co;
    BlockDriverAIOCBCoroutine *acb;
3311

P
Paolo Bonzini 已提交
3312 3313 3314
    acb = qemu_aio_get(&bdrv_em_co_aio_pool, bs, cb, opaque);
    co = qemu_coroutine_create(bdrv_aio_flush_co_entry);
    qemu_coroutine_enter(co, acb);
3315 3316 3317 3318

    return &acb->common;
}

3319 3320 3321 3322 3323 3324 3325 3326 3327 3328 3329 3330 3331 3332 3333 3334 3335 3336 3337 3338 3339 3340 3341 3342 3343 3344 3345 3346
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);

    acb = qemu_aio_get(&bdrv_em_co_aio_pool, bs, cb, opaque);
    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
3347 3348
void bdrv_init(void)
{
3349
    module_call_init(MODULE_INIT_BLOCK);
B
bellard 已提交
3350
}
3351

3352 3353 3354 3355 3356 3357
void bdrv_init_with_whitelist(void)
{
    use_bdrv_whitelist = 1;
    bdrv_init();
}

3358 3359
void *qemu_aio_get(AIOPool *pool, BlockDriverState *bs,
                   BlockDriverCompletionFunc *cb, void *opaque)
3360 3361 3362
{
    BlockDriverAIOCB *acb;

3363 3364 3365
    if (pool->free_aiocb) {
        acb = pool->free_aiocb;
        pool->free_aiocb = acb->next;
3366
    } else {
3367
        acb = g_malloc0(pool->aiocb_size);
3368
        acb->pool = pool;
3369 3370 3371 3372 3373 3374 3375 3376 3377
    }
    acb->bs = bs;
    acb->cb = cb;
    acb->opaque = opaque;
    return acb;
}

void qemu_aio_release(void *p)
{
3378 3379 3380 3381
    BlockDriverAIOCB *acb = (BlockDriverAIOCB *)p;
    AIOPool *pool = acb->pool;
    acb->next = pool->free_aiocb;
    pool->free_aiocb = acb;
3382
}
B
bellard 已提交
3383

3384 3385 3386 3387 3388 3389 3390 3391 3392 3393 3394 3395 3396 3397 3398 3399 3400 3401 3402 3403 3404 3405 3406 3407 3408 3409
/**************************************************************/
/* 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) {
3410 3411
        acb = bs->drv->bdrv_aio_writev(bs, sector_num, iov, nb_sectors,
                                       bdrv_co_io_em_complete, &co);
3412
    } else {
3413 3414
        acb = bs->drv->bdrv_aio_readv(bs, sector_num, iov, nb_sectors,
                                      bdrv_co_io_em_complete, &co);
3415 3416
    }

3417
    trace_bdrv_co_io_em(bs, sector_num, nb_sectors, is_write, acb);
3418 3419 3420 3421 3422 3423 3424 3425 3426 3427 3428 3429 3430 3431 3432 3433 3434 3435 3436 3437 3438 3439
    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
3440
static void coroutine_fn bdrv_flush_co_entry(void *opaque)
3441
{
P
Paolo Bonzini 已提交
3442 3443 3444 3445 3446 3447 3448
    RwCo *rwco = opaque;

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

int coroutine_fn bdrv_co_flush(BlockDriverState *bs)
{
K
Kevin Wolf 已提交
3449 3450
    int ret;

3451
    if (!bs->drv) {
P
Paolo Bonzini 已提交
3452
        return 0;
K
Kevin Wolf 已提交
3453 3454
    }

3455
    /* Write back cached data to the OS even with cache=unsafe */
K
Kevin Wolf 已提交
3456 3457 3458 3459 3460 3461 3462
    if (bs->drv->bdrv_co_flush_to_os) {
        ret = bs->drv->bdrv_co_flush_to_os(bs);
        if (ret < 0) {
            return ret;
        }
    }

3463 3464 3465 3466 3467
    /* But don't actually force it to the disk with cache=unsafe */
    if (bs->open_flags & BDRV_O_NO_FLUSH) {
        return 0;
    }

K
Kevin Wolf 已提交
3468
    if (bs->drv->bdrv_co_flush_to_disk) {
3469
        return bs->drv->bdrv_co_flush_to_disk(bs);
P
Paolo Bonzini 已提交
3470 3471 3472 3473 3474 3475 3476 3477 3478 3479 3480 3481 3482 3483 3484 3485 3486 3487 3488 3489 3490 3491 3492 3493 3494 3495 3496 3497 3498
    } 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) {
            return -EIO;
        } else {
            qemu_coroutine_yield();
            return co.ret;
        }
    } 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.
         */
        return 0;
    }
}

3499 3500 3501 3502 3503 3504 3505 3506 3507 3508 3509 3510 3511 3512 3513 3514
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);
    }
}

P
Paolo Bonzini 已提交
3515 3516 3517 3518 3519 3520
int bdrv_flush(BlockDriverState *bs)
{
    Coroutine *co;
    RwCo rwco = {
        .bs = bs,
        .ret = NOT_DONE,
3521 3522
    };

P
Paolo Bonzini 已提交
3523 3524 3525 3526 3527 3528 3529 3530 3531
    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();
        }
3532
    }
P
Paolo Bonzini 已提交
3533 3534

    return rwco.ret;
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
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
3598 3599 3600 3601 3602 3603 3604 3605 3606
/**************************************************************/
/* removable device support */

/**
 * Return TRUE if the media is present
 */
int bdrv_is_inserted(BlockDriverState *bs)
{
    BlockDriver *drv = bs->drv;
3607

B
bellard 已提交
3608 3609 3610
    if (!drv)
        return 0;
    if (!drv->bdrv_is_inserted)
3611 3612
        return 1;
    return drv->bdrv_is_inserted(bs);
B
bellard 已提交
3613 3614 3615
}

/**
3616 3617
 * 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
3618 3619 3620 3621 3622
 */
int bdrv_media_changed(BlockDriverState *bs)
{
    BlockDriver *drv = bs->drv;

3623 3624 3625 3626
    if (drv && drv->bdrv_media_changed) {
        return drv->bdrv_media_changed(bs);
    }
    return -ENOTSUP;
B
bellard 已提交
3627 3628 3629 3630 3631
}

/**
 * If eject_flag is TRUE, eject the media. Otherwise, close the tray
 */
3632
void bdrv_eject(BlockDriverState *bs, bool eject_flag)
B
bellard 已提交
3633 3634 3635
{
    BlockDriver *drv = bs->drv;

3636 3637
    if (drv && drv->bdrv_eject) {
        drv->bdrv_eject(bs, eject_flag);
B
bellard 已提交
3638
    }
3639 3640 3641 3642

    if (bs->device_name[0] != '\0') {
        bdrv_emit_qmp_eject_event(bs, eject_flag);
    }
B
bellard 已提交
3643 3644 3645 3646 3647 3648
}

/**
 * Lock or unlock the media (if it is locked, the user won't be able
 * to eject it manually).
 */
3649
void bdrv_lock_medium(BlockDriverState *bs, bool locked)
B
bellard 已提交
3650 3651 3652
{
    BlockDriver *drv = bs->drv;

3653
    trace_bdrv_lock_medium(bs, locked);
S
Stefan Hajnoczi 已提交
3654

3655 3656
    if (drv && drv->bdrv_lock_medium) {
        drv->bdrv_lock_medium(bs, locked);
B
bellard 已提交
3657 3658
    }
}
3659 3660 3661 3662 3663 3664 3665 3666 3667 3668 3669

/* 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;
}
3670

3671 3672 3673
BlockDriverAIOCB *bdrv_aio_ioctl(BlockDriverState *bs,
        unsigned long int req, void *buf,
        BlockDriverCompletionFunc *cb, void *opaque)
3674
{
3675
    BlockDriver *drv = bs->drv;
3676

3677 3678 3679
    if (drv && drv->bdrv_aio_ioctl)
        return drv->bdrv_aio_ioctl(bs, req, buf, cb, opaque);
    return NULL;
3680
}
3681

3682 3683 3684 3685
void bdrv_set_buffer_alignment(BlockDriverState *bs, int align)
{
    bs->buffer_alignment = align;
}
3686

3687 3688 3689 3690
void *qemu_blockalign(BlockDriverState *bs, size_t size)
{
    return qemu_memalign((bs && bs->buffer_alignment) ? bs->buffer_alignment : 512, size);
}
3691 3692 3693 3694

void bdrv_set_dirty_tracking(BlockDriverState *bs, int enable)
{
    int64_t bitmap_size;
3695

3696
    bs->dirty_count = 0;
3697
    if (enable) {
3698 3699 3700 3701
        if (!bs->dirty_bitmap) {
            bitmap_size = (bdrv_getlength(bs) >> BDRV_SECTOR_BITS) +
                    BDRV_SECTORS_PER_DIRTY_CHUNK * 8 - 1;
            bitmap_size /= BDRV_SECTORS_PER_DIRTY_CHUNK * 8;
3702

3703
            bs->dirty_bitmap = g_malloc0(bitmap_size);
3704
        }
3705
    } else {
3706
        if (bs->dirty_bitmap) {
3707
            g_free(bs->dirty_bitmap);
3708
            bs->dirty_bitmap = NULL;
3709
        }
3710 3711 3712 3713 3714
    }
}

int bdrv_get_dirty(BlockDriverState *bs, int64_t sector)
{
J
Jan Kiszka 已提交
3715
    int64_t chunk = sector / (int64_t)BDRV_SECTORS_PER_DIRTY_CHUNK;
3716

3717 3718
    if (bs->dirty_bitmap &&
        (sector << BDRV_SECTOR_BITS) < bdrv_getlength(bs)) {
3719 3720
        return !!(bs->dirty_bitmap[chunk / (sizeof(unsigned long) * 8)] &
            (1UL << (chunk % (sizeof(unsigned long) * 8))));
3721 3722 3723 3724 3725
    } else {
        return 0;
    }
}

3726 3727
void bdrv_reset_dirty(BlockDriverState *bs, int64_t cur_sector,
                      int nr_sectors)
3728 3729 3730
{
    set_dirty_bitmap(bs, cur_sector, nr_sectors, 0);
}
3731 3732 3733 3734 3735

int64_t bdrv_get_dirty_count(BlockDriverState *bs)
{
    return bs->dirty_count;
}
J
Jes Sorensen 已提交
3736

3737 3738 3739 3740 3741 3742 3743 3744 3745 3746 3747
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;
}

3748 3749
void bdrv_iostatus_enable(BlockDriverState *bs)
{
3750
    bs->iostatus_enabled = true;
3751
    bs->iostatus = BLOCK_DEVICE_IO_STATUS_OK;
3752 3753 3754 3755 3756 3757
}

/* 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)
{
3758
    return (bs->iostatus_enabled &&
3759 3760 3761 3762 3763 3764 3765
           (bs->on_write_error == BLOCK_ERR_STOP_ENOSPC ||
            bs->on_write_error == BLOCK_ERR_STOP_ANY    ||
            bs->on_read_error == BLOCK_ERR_STOP_ANY));
}

void bdrv_iostatus_disable(BlockDriverState *bs)
{
3766
    bs->iostatus_enabled = false;
3767 3768 3769 3770 3771
}

void bdrv_iostatus_reset(BlockDriverState *bs)
{
    if (bdrv_iostatus_is_enabled(bs)) {
3772
        bs->iostatus = BLOCK_DEVICE_IO_STATUS_OK;
3773 3774 3775 3776 3777 3778 3779 3780
    }
}

/* XXX: Today this is set by device models because it makes the implementation
   quite simple. However, the block layer knows about the error, so it's
   possible to implement this without device models being involved */
void bdrv_iostatus_set_err(BlockDriverState *bs, int error)
{
3781 3782
    if (bdrv_iostatus_is_enabled(bs) &&
        bs->iostatus == BLOCK_DEVICE_IO_STATUS_OK) {
3783
        assert(error >= 0);
3784 3785
        bs->iostatus = error == ENOSPC ? BLOCK_DEVICE_IO_STATUS_NOSPACE :
                                         BLOCK_DEVICE_IO_STATUS_FAILED;
3786 3787 3788
    }
}

3789 3790 3791 3792 3793 3794 3795
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
3796
    cookie->start_time_ns = get_clock();
3797 3798 3799 3800 3801 3802 3803 3804 3805 3806
    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
3807
    bs->total_time_ns[cookie->type] += get_clock() - cookie->start_time_ns;
3808 3809
}

J
Jes Sorensen 已提交
3810 3811 3812 3813 3814
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
3815
    QEMUOptionParameter *backing_fmt, *backing_file, *size;
J
Jes Sorensen 已提交
3816 3817
    BlockDriverState *bs = NULL;
    BlockDriver *drv, *proto_drv;
3818
    BlockDriver *backing_drv = NULL;
J
Jes Sorensen 已提交
3819 3820 3821 3822 3823 3824
    int ret = 0;

    /* Find driver and parse its options */
    drv = bdrv_find_format(fmt);
    if (!drv) {
        error_report("Unknown file format '%s'", fmt);
3825
        ret = -EINVAL;
J
Jes Sorensen 已提交
3826 3827 3828 3829 3830 3831
        goto out;
    }

    proto_drv = bdrv_find_protocol(filename);
    if (!proto_drv) {
        error_report("Unknown protocol '%s'", filename);
3832
        ret = -EINVAL;
J
Jes Sorensen 已提交
3833 3834 3835 3836 3837 3838 3839 3840 3841 3842 3843 3844 3845 3846 3847 3848 3849 3850
        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);
3851
            ret = -EINVAL;
J
Jes Sorensen 已提交
3852 3853 3854 3855 3856 3857 3858 3859 3860
            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);
3861
            ret = -EINVAL;
J
Jes Sorensen 已提交
3862 3863 3864 3865 3866 3867 3868 3869
            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);
3870
            ret = -EINVAL;
J
Jes Sorensen 已提交
3871 3872 3873 3874
            goto out;
        }
    }

3875 3876 3877 3878 3879
    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");
3880
            ret = -EINVAL;
3881 3882 3883 3884
            goto out;
        }
    }

J
Jes Sorensen 已提交
3885 3886
    backing_fmt = get_option_parameter(param, BLOCK_OPT_BACKING_FMT);
    if (backing_fmt && backing_fmt->value.s) {
3887 3888
        backing_drv = bdrv_find_format(backing_fmt->value.s);
        if (!backing_drv) {
J
Jes Sorensen 已提交
3889 3890
            error_report("Unknown backing file format '%s'",
                         backing_fmt->value.s);
3891
            ret = -EINVAL;
J
Jes Sorensen 已提交
3892 3893 3894 3895 3896 3897
            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
3898 3899
    size = get_option_parameter(param, BLOCK_OPT_SIZE);
    if (size && size->value.n == -1) {
J
Jes Sorensen 已提交
3900 3901 3902 3903 3904 3905
        if (backing_file && backing_file->value.s) {
            uint64_t size;
            char buf[32];

            bs = bdrv_new("");

3906
            ret = bdrv_open(bs, backing_file->value.s, flags, backing_drv);
J
Jes Sorensen 已提交
3907
            if (ret < 0) {
3908
                error_report("Could not open '%s'", backing_file->value.s);
J
Jes Sorensen 已提交
3909 3910 3911 3912 3913 3914 3915 3916 3917
                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");
3918
            ret = -EINVAL;
J
Jes Sorensen 已提交
3919 3920 3921 3922 3923 3924 3925 3926 3927 3928 3929 3930 3931 3932 3933 3934 3935 3936 3937 3938 3939 3940 3941 3942 3943 3944 3945 3946 3947 3948
            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);
    }
3949 3950

    return ret;
J
Jes Sorensen 已提交
3951
}
3952 3953 3954 3955 3956 3957 3958 3959 3960 3961 3962 3963 3964 3965 3966 3967 3968 3969 3970 3971 3972 3973 3974 3975 3976 3977 3978 3979 3980 3981 3982 3983 3984 3985 3986 3987 3988 3989 3990 3991 3992 3993 3994 3995 3996 3997 3998 3999

void *block_job_create(const BlockJobType *job_type, BlockDriverState *bs,
                       BlockDriverCompletionFunc *cb, void *opaque)
{
    BlockJob *job;

    if (bs->job || bdrv_in_use(bs)) {
        return NULL;
    }
    bdrv_set_in_use(bs, 1);

    job = g_malloc0(job_type->instance_size);
    job->job_type      = job_type;
    job->bs            = bs;
    job->cb            = cb;
    job->opaque        = opaque;
    bs->job = job;
    return job;
}

void block_job_complete(BlockJob *job, int ret)
{
    BlockDriverState *bs = job->bs;

    assert(bs->job == job);
    job->cb(job->opaque, ret);
    bs->job = NULL;
    g_free(job);
    bdrv_set_in_use(bs, 0);
}

int block_job_set_speed(BlockJob *job, int64_t value)
{
    if (!job->job_type->set_speed) {
        return -ENOTSUP;
    }
    return job->job_type->set_speed(job, value);
}

void block_job_cancel(BlockJob *job)
{
    job->cancelled = true;
}

bool block_job_is_cancelled(BlockJob *job)
{
    return job->cancelled;
}