rbd.c 22.5 KB
Newer Older
1 2 3
/*
 * QEMU Block driver for RADOS (Ceph)
 *
4 5
 * Copyright (C) 2010-2011 Christian Brunner <chb@muc.de>,
 *                         Josh Durgin <josh.durgin@dreamhost.com>
6 7 8 9
 *
 * This work is licensed under the terms of the GNU GPL, version 2.  See
 * the COPYING file in the top-level directory.
 *
10 11
 * Contributions after 2012-01-13 are licensed under the terms of the
 * GNU GPL, version 2 or (at your option) any later version.
12 13
 */

14 15
#include <inttypes.h>

16 17 18 19
#include "qemu-common.h"
#include "qemu-error.h"
#include "block_int.h"

20
#include <rbd/librbd.h>
21 22 23 24

/*
 * When specifying the image filename use:
 *
25
 * rbd:poolname/devicename[@snapshotname][:option1=value1[:option2=value2...]]
26
 *
S
Sage Weil 已提交
27
 * poolname must be the name of an existing rados pool.
28
 *
S
Sage Weil 已提交
29
 * devicename is the name of the rbd image.
30
 *
S
Sage Weil 已提交
31 32
 * Each option given is used to configure rados, and may be any valid
 * Ceph option, "id", or "conf".
33
 *
S
Sage Weil 已提交
34 35 36
 * The "id" option indicates what user we should authenticate as to
 * the Ceph cluster.  If it is excluded we will use the Ceph default
 * (normally 'admin').
37
 *
S
Sage Weil 已提交
38 39 40 41
 * The "conf" option specifies a Ceph configuration file to read.  If
 * it is not specified, we will read from the default Ceph locations
 * (e.g., /etc/ceph/ceph.conf).  To avoid reading _any_ configuration
 * file, specify conf=/dev/null.
42
 *
S
Sage Weil 已提交
43 44
 * Configuration values containing :, @, or = can be escaped with a
 * leading "\".
45 46 47 48
 */

#define OBJ_MAX_SIZE (1UL << OBJ_DEFAULT_OBJ_ORDER)

49 50 51 52 53 54 55
#define RBD_MAX_CONF_NAME_SIZE 128
#define RBD_MAX_CONF_VAL_SIZE 512
#define RBD_MAX_CONF_SIZE 1024
#define RBD_MAX_POOL_NAME_SIZE 128
#define RBD_MAX_SNAP_NAME_SIZE 128
#define RBD_MAX_SNAPS 100

56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73
typedef struct RBDAIOCB {
    BlockDriverAIOCB common;
    QEMUBH *bh;
    int ret;
    QEMUIOVector *qiov;
    char *bounce;
    int write;
    int64_t sector_num;
    int error;
    struct BDRVRBDState *s;
    int cancelled;
} RBDAIOCB;

typedef struct RADOSCB {
    int rcbid;
    RBDAIOCB *acb;
    struct BDRVRBDState *s;
    int done;
74
    int64_t size;
75 76 77 78 79 80 81 82 83
    char *buf;
    int ret;
} RADOSCB;

#define RBD_FD_READ 0
#define RBD_FD_WRITE 1

typedef struct BDRVRBDState {
    int fds[2];
84 85 86 87
    rados_t cluster;
    rados_ioctx_t io_ctx;
    rbd_image_t image;
    char name[RBD_MAX_IMAGE_NAME_SIZE];
88
    int qemu_aio_count;
89
    char *snap;
90 91 92 93 94 95
    int event_reader_pos;
    RADOSCB *event_rcb;
} BDRVRBDState;

static void rbd_aio_bh_cb(void *opaque);

96 97 98 99
static int qemu_rbd_next_tok(char *dst, int dst_len,
                             char *src, char delim,
                             const char *name,
                             char **p)
100 101 102 103 104 105 106
{
    int l;
    char *end;

    *p = NULL;

    if (delim != '\0') {
S
Sage Weil 已提交
107 108 109 110 111 112 113 114 115
        for (end = src; *end; ++end) {
            if (*end == delim) {
                break;
            }
            if (*end == '\\' && end[1] != '\0') {
                end++;
            }
        }
        if (*end == delim) {
116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133
            *p = end + 1;
            *end = '\0';
        }
    }
    l = strlen(src);
    if (l >= dst_len) {
        error_report("%s too long", name);
        return -EINVAL;
    } else if (l == 0) {
        error_report("%s too short", name);
        return -EINVAL;
    }

    pstrcpy(dst, dst_len, src);

    return 0;
}

S
Sage Weil 已提交
134 135 136 137 138 139 140 141 142 143 144 145 146
static void qemu_rbd_unescape(char *src)
{
    char *p;

    for (p = src; *src; ++src, ++p) {
        if (*src == '\\' && src[1] != '\0') {
            src++;
        }
        *p = *src;
    }
    *p = '\0';
}

147 148 149
static int qemu_rbd_parsename(const char *filename,
                              char *pool, int pool_len,
                              char *snap, int snap_len,
150 151
                              char *name, int name_len,
                              char *conf, int conf_len)
152 153 154 155 156 157 158 159 160
{
    const char *start;
    char *p, *buf;
    int ret;

    if (!strstart(filename, "rbd:", &start)) {
        return -EINVAL;
    }

161
    buf = g_strdup(start);
162
    p = buf;
163 164
    *snap = '\0';
    *conf = '\0';
165

166
    ret = qemu_rbd_next_tok(pool, pool_len, p, '/', "pool name", &p);
167 168 169 170
    if (ret < 0 || !p) {
        ret = -EINVAL;
        goto done;
    }
S
Sage Weil 已提交
171
    qemu_rbd_unescape(pool);
172 173 174 175 176 177 178

    if (strchr(p, '@')) {
        ret = qemu_rbd_next_tok(name, name_len, p, '@', "object name", &p);
        if (ret < 0) {
            goto done;
        }
        ret = qemu_rbd_next_tok(snap, snap_len, p, ':', "snap name", &p);
S
Sage Weil 已提交
179
        qemu_rbd_unescape(snap);
180 181
    } else {
        ret = qemu_rbd_next_tok(name, name_len, p, ':', "object name", &p);
182
    }
S
Sage Weil 已提交
183
    qemu_rbd_unescape(name);
184
    if (ret < 0 || !p) {
185 186 187
        goto done;
    }

188
    ret = qemu_rbd_next_tok(conf, conf_len, p, '\0', "configuration", &p);
189 190

done:
191
    g_free(buf);
192 193 194
    return ret;
}

195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222
static char *qemu_rbd_parse_clientname(const char *conf, char *clientname)
{
    const char *p = conf;

    while (*p) {
        int len;
        const char *end = strchr(p, ':');

        if (end) {
            len = end - p;
        } else {
            len = strlen(p);
        }

        if (strncmp(p, "id=", 3) == 0) {
            len -= 3;
            strncpy(clientname, p + 3, len);
            clientname[len] = '\0';
            return clientname;
        }
        if (end == NULL) {
            break;
        }
        p = end + 1;
    }
    return NULL;
}

223 224 225 226 227 228 229
static int qemu_rbd_set_conf(rados_t cluster, const char *conf)
{
    char *p, *buf;
    char name[RBD_MAX_CONF_NAME_SIZE];
    char value[RBD_MAX_CONF_VAL_SIZE];
    int ret = 0;

230
    buf = g_strdup(conf);
231 232 233 234 235 236 237 238
    p = buf;

    while (p) {
        ret = qemu_rbd_next_tok(name, sizeof(name), p,
                                '=', "conf option name", &p);
        if (ret < 0) {
            break;
        }
S
Sage Weil 已提交
239
        qemu_rbd_unescape(name);
240 241 242 243 244 245 246 247 248 249 250 251

        if (!p) {
            error_report("conf option %s has no value", name);
            ret = -EINVAL;
            break;
        }

        ret = qemu_rbd_next_tok(value, sizeof(value), p,
                                ':', "conf option value", &p);
        if (ret < 0) {
            break;
        }
S
Sage Weil 已提交
252
        qemu_rbd_unescape(value);
253

254 255
        if (strcmp(name, "conf") == 0) {
            ret = rados_conf_read_file(cluster, value);
256
            if (ret < 0) {
257
                error_report("error reading conf file %s", value);
258 259
                break;
            }
260 261
        } else if (strcmp(name, "id") == 0) {
            /* ignore, this is parsed by qemu_rbd_parse_clientname() */
262
        } else {
263
            ret = rados_conf_set(cluster, name, value);
264
            if (ret < 0) {
265 266
                error_report("invalid conf option %s", name);
                ret = -EINVAL;
267 268 269 270 271
                break;
            }
        }
    }

272
    g_free(buf);
273 274 275
    return ret;
}

276
static int qemu_rbd_create(const char *filename, QEMUOptionParameter *options)
277 278 279
{
    int64_t bytes = 0;
    int64_t objsize;
280 281 282 283
    int obj_order = 0;
    char pool[RBD_MAX_POOL_NAME_SIZE];
    char name[RBD_MAX_IMAGE_NAME_SIZE];
    char snap_buf[RBD_MAX_SNAP_NAME_SIZE];
284
    char conf[RBD_MAX_CONF_SIZE];
285 286
    char clientname_buf[RBD_MAX_CONF_SIZE];
    char *clientname;
287 288
    rados_t cluster;
    rados_ioctx_t io_ctx;
289 290
    int ret;

291 292
    if (qemu_rbd_parsename(filename, pool, sizeof(pool),
                           snap_buf, sizeof(snap_buf),
293 294
                           name, sizeof(name),
                           conf, sizeof(conf)) < 0) {
295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312
        return -EINVAL;
    }

    /* Read out options */
    while (options && options->name) {
        if (!strcmp(options->name, BLOCK_OPT_SIZE)) {
            bytes = options->value.n;
        } else if (!strcmp(options->name, BLOCK_OPT_CLUSTER_SIZE)) {
            if (options->value.n) {
                objsize = options->value.n;
                if ((objsize - 1) & objsize) {    /* not a power of 2? */
                    error_report("obj size needs to be power of 2");
                    return -EINVAL;
                }
                if (objsize < 4096) {
                    error_report("obj size too small");
                    return -EINVAL;
                }
313
                obj_order = ffs(objsize) - 1;
314 315 316 317 318
            }
        }
        options++;
    }

319 320
    clientname = qemu_rbd_parse_clientname(conf, clientname_buf);
    if (rados_create(&cluster, clientname) < 0) {
321 322 323 324
        error_report("error initializing");
        return -EIO;
    }

325
    if (strstr(conf, "conf=") == NULL) {
326 327
        /* try default location, but ignore failure */
        rados_conf_read_file(cluster, NULL);
328 329 330 331 332
    }

    if (conf[0] != '\0' &&
        qemu_rbd_set_conf(cluster, conf) < 0) {
        error_report("error setting config options");
333
        rados_shutdown(cluster);
334 335 336
        return -EIO;
    }

337 338 339
    if (rados_connect(cluster) < 0) {
        error_report("error connecting");
        rados_shutdown(cluster);
340 341 342
        return -EIO;
    }

343 344 345 346
    if (rados_ioctx_create(cluster, pool, &io_ctx) < 0) {
        error_report("error opening pool %s", pool);
        rados_shutdown(cluster);
        return -EIO;
347 348
    }

349 350 351
    ret = rbd_create(io_ctx, name, bytes, &obj_order);
    rados_ioctx_destroy(io_ctx);
    rados_shutdown(cluster);
352 353 354 355 356

    return ret;
}

/*
357 358
 * This aio completion is being called from qemu_rbd_aio_event_reader()
 * and runs in qemu context. It schedules a bh, but just in case the aio
359 360
 * was not cancelled before.
 */
361
static void qemu_rbd_complete_aio(RADOSCB *rcb)
362 363 364 365 366
{
    RBDAIOCB *acb = rcb->acb;
    int64_t r;

    if (acb->cancelled) {
367 368
        qemu_vfree(acb->bounce);
        qemu_aio_release(acb);
369 370 371 372 373 374 375 376 377 378
        goto done;
    }

    r = rcb->ret;

    if (acb->write) {
        if (r < 0) {
            acb->ret = r;
            acb->error = 1;
        } else if (!acb->error) {
379
            acb->ret = rcb->size;
380 381
        }
    } else {
382 383
        if (r < 0) {
            memset(rcb->buf, 0, rcb->size);
384 385
            acb->ret = r;
            acb->error = 1;
386 387
        } else if (r < rcb->size) {
            memset(rcb->buf + r, 0, rcb->size - r);
388
            if (!acb->error) {
389
                acb->ret = rcb->size;
390 391
            }
        } else if (!acb->error) {
392
            acb->ret = r;
393 394 395
        }
    }
    /* Note that acb->bh can be NULL in case where the aio was cancelled */
396 397
    acb->bh = qemu_bh_new(rbd_aio_bh_cb, acb);
    qemu_bh_schedule(acb->bh);
398
done:
399
    g_free(rcb);
400 401 402 403 404 405
}

/*
 * aio fd read handler. It runs in the qemu context and calls the
 * completion handling of completed rados aio operations.
 */
406
static void qemu_rbd_aio_event_reader(void *opaque)
407 408 409 410 411 412 413 414 415
{
    BDRVRBDState *s = opaque;

    ssize_t ret;

    do {
        char *p = (char *)&s->event_rcb;

        /* now read the rcb pointer that was sent from a non qemu thread */
S
Sage Weil 已提交
416 417 418 419 420 421 422 423
        ret = read(s->fds[RBD_FD_READ], p + s->event_reader_pos,
                   sizeof(s->event_rcb) - s->event_reader_pos);
        if (ret > 0) {
            s->event_reader_pos += ret;
            if (s->event_reader_pos == sizeof(s->event_rcb)) {
                s->event_reader_pos = 0;
                qemu_rbd_complete_aio(s->event_rcb);
                s->qemu_aio_count--;
424 425 426 427 428
            }
        }
    } while (ret < 0 && errno == EINTR);
}

429
static int qemu_rbd_aio_flush_cb(void *opaque)
430 431 432 433 434 435
{
    BDRVRBDState *s = opaque;

    return (s->qemu_aio_count > 0);
}

436
static int qemu_rbd_open(BlockDriverState *bs, const char *filename, int flags)
437 438
{
    BDRVRBDState *s = bs->opaque;
439 440
    char pool[RBD_MAX_POOL_NAME_SIZE];
    char snap_buf[RBD_MAX_SNAP_NAME_SIZE];
441
    char conf[RBD_MAX_CONF_SIZE];
442 443
    char clientname_buf[RBD_MAX_CONF_SIZE];
    char *clientname;
444 445
    int r;

446 447
    if (qemu_rbd_parsename(filename, pool, sizeof(pool),
                           snap_buf, sizeof(snap_buf),
448 449
                           s->name, sizeof(s->name),
                           conf, sizeof(conf)) < 0) {
450 451 452
        return -EINVAL;
    }

453 454
    clientname = qemu_rbd_parse_clientname(conf, clientname_buf);
    r = rados_create(&s->cluster, clientname);
455
    if (r < 0) {
456 457 458 459
        error_report("error initializing");
        return r;
    }

460 461 462 463 464
    s->snap = NULL;
    if (snap_buf[0] != '\0') {
        s->snap = g_strdup(snap_buf);
    }

465
    if (strstr(conf, "conf=") == NULL) {
466 467
        /* try default location, but ignore failure */
        rados_conf_read_file(s->cluster, NULL);
468 469 470 471 472 473
    }

    if (conf[0] != '\0') {
        r = qemu_rbd_set_conf(s->cluster, conf);
        if (r < 0) {
            error_report("error setting config options");
474
            goto failed_shutdown;
475
        }
476 477
    }

478 479 480
    r = rados_connect(s->cluster);
    if (r < 0) {
        error_report("error connecting");
481
        goto failed_shutdown;
482 483
    }

484 485 486
    r = rados_ioctx_create(s->cluster, pool, &s->io_ctx);
    if (r < 0) {
        error_report("error opening pool %s", pool);
487
        goto failed_shutdown;
488 489
    }

490
    r = rbd_open(s->io_ctx, s->name, &s->image, s->snap);
491
    if (r < 0) {
492
        error_report("error reading header from %s", s->name);
493
        goto failed_open;
494 495
    }

496
    bs->read_only = (s->snap != NULL);
497 498 499 500 501 502 503 504 505

    s->event_reader_pos = 0;
    r = qemu_pipe(s->fds);
    if (r < 0) {
        error_report("error opening eventfd");
        goto failed;
    }
    fcntl(s->fds[0], F_SETFL, O_NONBLOCK);
    fcntl(s->fds[1], F_SETFL, O_NONBLOCK);
506 507
    qemu_aio_set_fd_handler(s->fds[RBD_FD_READ], qemu_rbd_aio_event_reader,
                            NULL, qemu_rbd_aio_flush_cb, NULL, s);
508 509 510 511 512


    return 0;

failed:
513
    rbd_close(s->image);
514
failed_open:
515
    rados_ioctx_destroy(s->io_ctx);
516
failed_shutdown:
517
    rados_shutdown(s->cluster);
518
    g_free(s->snap);
519 520 521
    return r;
}

522
static void qemu_rbd_close(BlockDriverState *bs)
523 524 525 526 527 528 529 530
{
    BDRVRBDState *s = bs->opaque;

    close(s->fds[0]);
    close(s->fds[1]);
    qemu_aio_set_fd_handler(s->fds[RBD_FD_READ], NULL , NULL, NULL, NULL,
        NULL);

531 532
    rbd_close(s->image);
    rados_ioctx_destroy(s->io_ctx);
533
    g_free(s->snap);
534
    rados_shutdown(s->cluster);
535 536 537 538 539 540
}

/*
 * Cancel aio. Since we don't reference acb in a non qemu threads,
 * it is safe to access it here.
 */
541
static void qemu_rbd_aio_cancel(BlockDriverAIOCB *blockacb)
542 543 544 545 546 547 548
{
    RBDAIOCB *acb = (RBDAIOCB *) blockacb;
    acb->cancelled = 1;
}

static AIOPool rbd_aio_pool = {
    .aiocb_size = sizeof(RBDAIOCB),
549
    .cancel = qemu_rbd_aio_cancel,
550 551
};

552
static int qemu_rbd_send_pipe(BDRVRBDState *s, RADOSCB *rcb)
553
{
554
    int ret = 0;
555 556
    while (1) {
        fd_set wfd;
557
        int fd = s->fds[RBD_FD_WRITE];
558

559 560
        /* send the op pointer to the qemu thread that is responsible
           for the aio/op completion. Must do it in a qemu thread context */
561 562 563 564 565 566
        ret = write(fd, (void *)&rcb, sizeof(rcb));
        if (ret >= 0) {
            break;
        }
        if (errno == EINTR) {
            continue;
567
        }
568 569
        if (errno != EAGAIN) {
            break;
570
        }
571 572 573 574 575 576 577 578

        FD_ZERO(&wfd);
        FD_SET(fd, &wfd);
        do {
            ret = select(fd + 1, NULL, &wfd, NULL, NULL);
        } while (ret < 0 && errno == EINTR);
    }

579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596
    return ret;
}

/*
 * This is the callback function for rbd_aio_read and _write
 *
 * Note: this function is being called from a non qemu thread so
 * we need to be careful about what we do here. Generally we only
 * write to the block notification pipe, and do the rest of the
 * io completion handling from qemu_rbd_aio_event_reader() which
 * runs in a qemu context.
 */
static void rbd_finish_aiocb(rbd_completion_t c, RADOSCB *rcb)
{
    int ret;
    rcb->ret = rbd_aio_get_return_value(c);
    rbd_aio_release(c);
    ret = qemu_rbd_send_pipe(rcb->s, rcb);
597
    if (ret < 0) {
598
        error_report("failed writing to acb->s->fds");
599
        g_free(rcb);
600 601 602
    }
}

603
/* Callback when all queued rbd_aio requests are complete */
604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628

static void rbd_aio_bh_cb(void *opaque)
{
    RBDAIOCB *acb = opaque;

    if (!acb->write) {
        qemu_iovec_from_buffer(acb->qiov, acb->bounce, acb->qiov->size);
    }
    qemu_vfree(acb->bounce);
    acb->common.cb(acb->common.opaque, (acb->ret > 0 ? 0 : acb->ret));
    qemu_bh_delete(acb->bh);
    acb->bh = NULL;

    qemu_aio_release(acb);
}

static BlockDriverAIOCB *rbd_aio_rw_vector(BlockDriverState *bs,
                                           int64_t sector_num,
                                           QEMUIOVector *qiov,
                                           int nb_sectors,
                                           BlockDriverCompletionFunc *cb,
                                           void *opaque, int write)
{
    RBDAIOCB *acb;
    RADOSCB *rcb;
629
    rbd_completion_t c;
630 631
    int64_t off, size;
    char *buf;
632
    int r;
633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654

    BDRVRBDState *s = bs->opaque;

    acb = qemu_aio_get(&rbd_aio_pool, bs, cb, opaque);
    acb->write = write;
    acb->qiov = qiov;
    acb->bounce = qemu_blockalign(bs, qiov->size);
    acb->ret = 0;
    acb->error = 0;
    acb->s = s;
    acb->cancelled = 0;
    acb->bh = NULL;

    if (write) {
        qemu_iovec_to_buffer(acb->qiov, acb->bounce);
    }

    buf = acb->bounce;

    off = sector_num * BDRV_SECTOR_SIZE;
    size = nb_sectors * BDRV_SECTOR_SIZE;

655
    s->qemu_aio_count++; /* All the RADOSCB */
656

657
    rcb = g_malloc(sizeof(RADOSCB));
658 659 660 661 662
    rcb->done = 0;
    rcb->acb = acb;
    rcb->buf = buf;
    rcb->s = acb->s;
    rcb->size = size;
663 664 665 666
    r = rbd_aio_create_completion(rcb, (rbd_callback_t) rbd_finish_aiocb, &c);
    if (r < 0) {
        goto failed;
    }
667

668
    if (write) {
669
        r = rbd_aio_write(s->image, off, size, buf, c);
670
    } else {
671 672 673 674 675
        r = rbd_aio_read(s->image, off, size, buf, c);
    }

    if (r < 0) {
        goto failed;
676 677 678
    }

    return &acb->common;
679 680

failed:
681
    g_free(rcb);
682 683 684
    s->qemu_aio_count--;
    qemu_aio_release(acb);
    return NULL;
685 686
}

687 688 689 690 691 692
static BlockDriverAIOCB *qemu_rbd_aio_readv(BlockDriverState *bs,
                                            int64_t sector_num,
                                            QEMUIOVector *qiov,
                                            int nb_sectors,
                                            BlockDriverCompletionFunc *cb,
                                            void *opaque)
693 694 695 696
{
    return rbd_aio_rw_vector(bs, sector_num, qiov, nb_sectors, cb, opaque, 0);
}

697 698 699 700 701 702
static BlockDriverAIOCB *qemu_rbd_aio_writev(BlockDriverState *bs,
                                             int64_t sector_num,
                                             QEMUIOVector *qiov,
                                             int nb_sectors,
                                             BlockDriverCompletionFunc *cb,
                                             void *opaque)
703 704 705 706
{
    return rbd_aio_rw_vector(bs, sector_num, qiov, nb_sectors, cb, opaque, 1);
}

P
Paolo Bonzini 已提交
707
static int qemu_rbd_co_flush(BlockDriverState *bs)
S
Sage Weil 已提交
708 709 710 711 712 713 714 715 716 717
{
#if LIBRBD_VERSION_CODE >= LIBRBD_VERSION(0, 1, 1)
    /* rbd_flush added in 0.1.1 */
    BDRVRBDState *s = bs->opaque;
    return rbd_flush(s->image);
#else
    return 0;
#endif
}

718
static int qemu_rbd_getinfo(BlockDriverState *bs, BlockDriverInfo *bdi)
719 720
{
    BDRVRBDState *s = bs->opaque;
721 722 723 724 725 726 727 728 729
    rbd_image_info_t info;
    int r;

    r = rbd_stat(s->image, &info, sizeof(info));
    if (r < 0) {
        return r;
    }

    bdi->cluster_size = info.obj_size;
730 731 732
    return 0;
}

733
static int64_t qemu_rbd_getlength(BlockDriverState *bs)
734 735
{
    BDRVRBDState *s = bs->opaque;
736 737
    rbd_image_info_t info;
    int r;
738

739 740 741 742 743 744
    r = rbd_stat(s->image, &info, sizeof(info));
    if (r < 0) {
        return r;
    }

    return info.size;
745 746
}

747 748 749 750 751 752 753 754 755 756 757 758 759
static int qemu_rbd_truncate(BlockDriverState *bs, int64_t offset)
{
    BDRVRBDState *s = bs->opaque;
    int r;

    r = rbd_resize(s->image, offset);
    if (r < 0) {
        return r;
    }

    return 0;
}

760 761
static int qemu_rbd_snap_create(BlockDriverState *bs,
                                QEMUSnapshotInfo *sn_info)
762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782
{
    BDRVRBDState *s = bs->opaque;
    int r;

    if (sn_info->name[0] == '\0') {
        return -EINVAL; /* we need a name for rbd snapshots */
    }

    /*
     * rbd snapshots are using the name as the user controlled unique identifier
     * we can't use the rbd snapid for that purpose, as it can't be set
     */
    if (sn_info->id_str[0] != '\0' &&
        strcmp(sn_info->id_str, sn_info->name) != 0) {
        return -EINVAL;
    }

    if (strlen(sn_info->name) >= sizeof(sn_info->id_str)) {
        return -ERANGE;
    }

783
    r = rbd_snap_create(s->image, sn_info->name);
784
    if (r < 0) {
785
        error_report("failed to create snap: %s", strerror(-r));
786 787 788 789 790 791
        return r;
    }

    return 0;
}

792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811
static int qemu_rbd_snap_remove(BlockDriverState *bs,
                                const char *snapshot_name)
{
    BDRVRBDState *s = bs->opaque;
    int r;

    r = rbd_snap_remove(s->image, snapshot_name);
    return r;
}

static int qemu_rbd_snap_rollback(BlockDriverState *bs,
                                  const char *snapshot_name)
{
    BDRVRBDState *s = bs->opaque;
    int r;

    r = rbd_snap_rollback(s->image, snapshot_name);
    return r;
}

812 813
static int qemu_rbd_snap_list(BlockDriverState *bs,
                              QEMUSnapshotInfo **psn_tab)
814 815 816
{
    BDRVRBDState *s = bs->opaque;
    QEMUSnapshotInfo *sn_info, *sn_tab = NULL;
817 818 819
    int i, snap_count;
    rbd_snap_info_t *snaps;
    int max_snaps = RBD_MAX_SNAPS;
820

821
    do {
822
        snaps = g_malloc(sizeof(*snaps) * max_snaps);
823 824
        snap_count = rbd_snap_list(s->image, snaps, &max_snaps);
        if (snap_count < 0) {
825
            g_free(snaps);
826
        }
827
    } while (snap_count == -ERANGE);
828

829
    if (snap_count <= 0) {
830
        goto done;
831 832
    }

833
    sn_tab = g_malloc0(snap_count * sizeof(QEMUSnapshotInfo));
834

835 836
    for (i = 0; i < snap_count; i++) {
        const char *snap_name = snaps[i].name;
837 838 839 840 841

        sn_info = sn_tab + i;
        pstrcpy(sn_info->id_str, sizeof(sn_info->id_str), snap_name);
        pstrcpy(sn_info->name, sizeof(sn_info->name), snap_name);

842
        sn_info->vm_state_size = snaps[i].size;
843 844 845 846
        sn_info->date_sec = 0;
        sn_info->date_nsec = 0;
        sn_info->vm_clock_nsec = 0;
    }
847 848
    rbd_snap_list_end(snaps);

849
 done:
850 851 852 853
    *psn_tab = sn_tab;
    return snap_count;
}

854
static QEMUOptionParameter qemu_rbd_create_options[] = {
855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870
    {
     .name = BLOCK_OPT_SIZE,
     .type = OPT_SIZE,
     .help = "Virtual disk size"
    },
    {
     .name = BLOCK_OPT_CLUSTER_SIZE,
     .type = OPT_SIZE,
     .help = "RBD object size"
    },
    {NULL}
};

static BlockDriver bdrv_rbd = {
    .format_name        = "rbd",
    .instance_size      = sizeof(BDRVRBDState),
871 872 873 874 875 876
    .bdrv_file_open     = qemu_rbd_open,
    .bdrv_close         = qemu_rbd_close,
    .bdrv_create        = qemu_rbd_create,
    .bdrv_get_info      = qemu_rbd_getinfo,
    .create_options     = qemu_rbd_create_options,
    .bdrv_getlength     = qemu_rbd_getlength,
877
    .bdrv_truncate      = qemu_rbd_truncate,
878 879
    .protocol_name      = "rbd",

880 881 882
    .bdrv_aio_readv         = qemu_rbd_aio_readv,
    .bdrv_aio_writev        = qemu_rbd_aio_writev,
    .bdrv_co_flush_to_disk  = qemu_rbd_co_flush,
883

884
    .bdrv_snapshot_create   = qemu_rbd_snap_create,
885
    .bdrv_snapshot_delete   = qemu_rbd_snap_remove,
886
    .bdrv_snapshot_list     = qemu_rbd_snap_list,
887
    .bdrv_snapshot_goto     = qemu_rbd_snap_rollback,
888 889 890 891 892 893 894 895
};

static void bdrv_rbd_init(void)
{
    bdrv_register(&bdrv_rbd);
}

block_init(bdrv_rbd_init);