blockdev.c 32.3 KB
Newer Older
1 2 3 4 5 6 7 8 9 10
/*
 * QEMU host block devices
 *
 * Copyright (c) 2003-2008 Fabrice Bellard
 *
 * This work is licensed under the terms of the GNU GPL, version 2 or
 * later.  See the COPYING file in the top-level directory.
 */

#include "blockdev.h"
11
#include "hw/block-common.h"
12 13 14 15
#include "monitor.h"
#include "qerror.h"
#include "qemu-option.h"
#include "qemu-config.h"
S
Stefan Hajnoczi 已提交
16
#include "qemu-objects.h"
17
#include "sysemu.h"
18
#include "block_int.h"
L
Luiz Capitulino 已提交
19
#include "qmp-commands.h"
S
Stefan Hajnoczi 已提交
20
#include "trace.h"
21
#include "arch_init.h"
22

23
static QTAILQ_HEAD(drivelist, DriveInfo) drives = QTAILQ_HEAD_INITIALIZER(drives);
24

25 26 27 28 29 30 31 32 33 34 35 36 37
static const char *const if_name[IF_COUNT] = {
    [IF_NONE] = "none",
    [IF_IDE] = "ide",
    [IF_SCSI] = "scsi",
    [IF_FLOPPY] = "floppy",
    [IF_PFLASH] = "pflash",
    [IF_MTD] = "mtd",
    [IF_SD] = "sd",
    [IF_VIRTIO] = "virtio",
    [IF_XEN] = "xen",
};

static const int if_max_devs[IF_COUNT] = {
38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53
    /*
     * Do not change these numbers!  They govern how drive option
     * index maps to unit and bus.  That mapping is ABI.
     *
     * All controllers used to imlement if=T drives need to support
     * if_max_devs[T] units, for any T with if_max_devs[T] != 0.
     * Otherwise, some index values map to "impossible" bus, unit
     * values.
     *
     * For instance, if you change [IF_SCSI] to 255, -drive
     * if=scsi,index=12 no longer means bus=1,unit=5, but
     * bus=0,unit=12.  With an lsi53c895a controller (7 units max),
     * the drive can't be set up.  Regression.
     */
    [IF_IDE] = 2,
    [IF_SCSI] = 7,
54 55
};

56 57 58 59 60 61 62 63 64 65 66
/*
 * We automatically delete the drive when a device using it gets
 * unplugged.  Questionable feature, but we can't just drop it.
 * Device models call blockdev_mark_auto_del() to schedule the
 * automatic deletion, and generic qdev code calls blockdev_auto_del()
 * when deletion is actually safe.
 */
void blockdev_mark_auto_del(BlockDriverState *bs)
{
    DriveInfo *dinfo = drive_get_by_blockdev(bs);

67 68 69
    if (bs->job) {
        block_job_cancel(bs->job);
    }
70 71 72
    if (dinfo) {
        dinfo->auto_del = 1;
    }
73 74 75 76 77 78
}

void blockdev_auto_del(BlockDriverState *bs)
{
    DriveInfo *dinfo = drive_get_by_blockdev(bs);

79
    if (dinfo && dinfo->auto_del) {
80
        drive_put_ref(dinfo);
81 82 83
    }
}

84 85 86 87 88 89 90 91 92 93 94 95
static int drive_index_to_bus_id(BlockInterfaceType type, int index)
{
    int max_devs = if_max_devs[type];
    return max_devs ? index / max_devs : 0;
}

static int drive_index_to_unit_id(BlockInterfaceType type, int index)
{
    int max_devs = if_max_devs[type];
    return max_devs ? index % max_devs : index;
}

96 97 98 99 100 101
QemuOpts *drive_def(const char *optstr)
{
    return qemu_opts_parse(qemu_find_opts("drive"), optstr, 0);
}

QemuOpts *drive_add(BlockInterfaceType type, int index, const char *file,
102
                    const char *optstr)
103 104
{
    QemuOpts *opts;
105
    char buf[32];
106

107
    opts = drive_def(optstr);
108 109 110
    if (!opts) {
        return NULL;
    }
111 112 113 114 115 116 117
    if (type != IF_DEFAULT) {
        qemu_opt_set(opts, "if", if_name[type]);
    }
    if (index >= 0) {
        snprintf(buf, sizeof(buf), "%d", index);
        qemu_opt_set(opts, "index", buf);
    }
118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138
    if (file)
        qemu_opt_set(opts, "file", file);
    return opts;
}

DriveInfo *drive_get(BlockInterfaceType type, int bus, int unit)
{
    DriveInfo *dinfo;

    /* seek interface, bus and unit */

    QTAILQ_FOREACH(dinfo, &drives, next) {
        if (dinfo->type == type &&
	    dinfo->bus == bus &&
	    dinfo->unit == unit)
            return dinfo;
    }

    return NULL;
}

139 140 141 142 143 144 145
DriveInfo *drive_get_by_index(BlockInterfaceType type, int index)
{
    return drive_get(type,
                     drive_index_to_bus_id(type, index),
                     drive_index_to_unit_id(type, index));
}

146 147 148 149 150 151 152 153 154 155 156 157 158 159
int drive_get_max_bus(BlockInterfaceType type)
{
    int max_bus;
    DriveInfo *dinfo;

    max_bus = -1;
    QTAILQ_FOREACH(dinfo, &drives, next) {
        if(dinfo->type == type &&
           dinfo->bus > max_bus)
            max_bus = dinfo->bus;
    }
    return max_bus;
}

160 161 162 163 164 165 166 167 168 169
/* Get a block device.  This should only be used for single-drive devices
   (e.g. SD/Floppy/MTD).  Multi-disk devices (scsi/ide) should use the
   appropriate bus.  */
DriveInfo *drive_get_next(BlockInterfaceType type)
{
    static int next_block_unit[IF_COUNT];

    return drive_get(type, 0, next_block_unit[type]++);
}

170
DriveInfo *drive_get_by_blockdev(BlockDriverState *bs)
171 172 173 174
{
    DriveInfo *dinfo;

    QTAILQ_FOREACH(dinfo, &drives, next) {
175 176 177
        if (dinfo->bdrv == bs) {
            return dinfo;
        }
178
    }
179
    return NULL;
180 181 182 183
}

static void bdrv_format_print(void *opaque, const char *name)
{
184
    error_printf(" %s", name);
185 186
}

187
static void drive_uninit(DriveInfo *dinfo)
188 189 190
{
    qemu_opts_del(dinfo->opts);
    bdrv_delete(dinfo->bdrv);
191
    g_free(dinfo->id);
192
    QTAILQ_REMOVE(&drives, dinfo, next);
193
    g_free(dinfo);
194 195
}

196 197 198 199 200 201 202 203 204 205 206 207 208
void drive_put_ref(DriveInfo *dinfo)
{
    assert(dinfo->refcount);
    if (--dinfo->refcount == 0) {
        drive_uninit(dinfo);
    }
}

void drive_get_ref(DriveInfo *dinfo)
{
    dinfo->refcount++;
}

209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239
typedef struct {
    QEMUBH *bh;
    DriveInfo *dinfo;
} DrivePutRefBH;

static void drive_put_ref_bh(void *opaque)
{
    DrivePutRefBH *s = opaque;

    drive_put_ref(s->dinfo);
    qemu_bh_delete(s->bh);
    g_free(s);
}

/*
 * Release a drive reference in a BH
 *
 * It is not possible to use drive_put_ref() from a callback function when the
 * callers still need the drive.  In such cases we schedule a BH to release the
 * reference.
 */
static void drive_put_ref_bh_schedule(DriveInfo *dinfo)
{
    DrivePutRefBH *s;

    s = g_new(DrivePutRefBH, 1);
    s->bh = qemu_bh_new(drive_put_ref_bh, s);
    s->dinfo = dinfo;
    qemu_bh_schedule(s->bh);
}

240 241 242 243 244 245 246 247 248 249 250
static int parse_block_error_action(const char *buf, int is_read)
{
    if (!strcmp(buf, "ignore")) {
        return BLOCK_ERR_IGNORE;
    } else if (!is_read && !strcmp(buf, "enospc")) {
        return BLOCK_ERR_STOP_ENOSPC;
    } else if (!strcmp(buf, "stop")) {
        return BLOCK_ERR_STOP_ANY;
    } else if (!strcmp(buf, "report")) {
        return BLOCK_ERR_REPORT;
    } else {
251 252
        error_report("'%s' invalid %s error action",
                     buf, is_read ? "read" : "write");
253 254 255 256
        return -1;
    }
}

257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276
static bool do_check_io_limits(BlockIOLimit *io_limits)
{
    bool bps_flag;
    bool iops_flag;

    assert(io_limits);

    bps_flag  = (io_limits->bps[BLOCK_IO_LIMIT_TOTAL] != 0)
                 && ((io_limits->bps[BLOCK_IO_LIMIT_READ] != 0)
                 || (io_limits->bps[BLOCK_IO_LIMIT_WRITE] != 0));
    iops_flag = (io_limits->iops[BLOCK_IO_LIMIT_TOTAL] != 0)
                 && ((io_limits->iops[BLOCK_IO_LIMIT_READ] != 0)
                 || (io_limits->iops[BLOCK_IO_LIMIT_WRITE] != 0));
    if (bps_flag || iops_flag) {
        return false;
    }

    return true;
}

277
DriveInfo *drive_init(QemuOpts *opts, int default_to_scsi)
278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294
{
    const char *buf;
    const char *file = NULL;
    const char *serial;
    const char *mediastr = "";
    BlockInterfaceType type;
    enum { MEDIA_DISK, MEDIA_CDROM } media;
    int bus_id, unit_id;
    int cyls, heads, secs, translation;
    BlockDriver *drv = NULL;
    int max_devs;
    int index;
    int ro = 0;
    int bdrv_flags = 0;
    int on_read_error, on_write_error;
    const char *devaddr;
    DriveInfo *dinfo;
295
    BlockIOLimit io_limits;
296
    int snapshot = 0;
297
    bool copy_on_read;
298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313
    int ret;

    translation = BIOS_ATA_TRANSLATION_AUTO;
    media = MEDIA_DISK;

    /* extract parameters */
    bus_id  = qemu_opt_get_number(opts, "bus", 0);
    unit_id = qemu_opt_get_number(opts, "unit", -1);
    index   = qemu_opt_get_number(opts, "index", -1);

    cyls  = qemu_opt_get_number(opts, "cyls", 0);
    heads = qemu_opt_get_number(opts, "heads", 0);
    secs  = qemu_opt_get_number(opts, "secs", 0);

    snapshot = qemu_opt_get_bool(opts, "snapshot", 0);
    ro = qemu_opt_get_bool(opts, "readonly", 0);
314
    copy_on_read = qemu_opt_get_bool(opts, "copy-on-read", false);
315 316 317 318 319

    file = qemu_opt_get(opts, "file");
    serial = qemu_opt_get(opts, "serial");

    if ((buf = qemu_opt_get(opts, "if")) != NULL) {
320 321 322
        for (type = 0; type < IF_COUNT && strcmp(buf, if_name[type]); type++)
            ;
        if (type == IF_COUNT) {
323
            error_report("unsupported bus type '%s'", buf);
324 325
            return NULL;
	}
326 327
    } else {
        type = default_to_scsi ? IF_SCSI : IF_IDE;
328
    }
329

330
    max_devs = if_max_devs[type];
331 332

    if (cyls || heads || secs) {
333
        if (cyls < 1) {
334
            error_report("invalid physical cyls number");
335 336
	    return NULL;
	}
337
        if (heads < 1) {
338
            error_report("invalid physical heads number");
339 340
	    return NULL;
	}
341
        if (secs < 1) {
342
            error_report("invalid physical secs number");
343 344 345 346 347 348
	    return NULL;
	}
    }

    if ((buf = qemu_opt_get(opts, "trans")) != NULL) {
        if (!cyls) {
349
            error_report("'%s' trans must be used with cyls, heads and secs",
350
                         buf);
351 352 353 354 355 356 357 358 359
            return NULL;
        }
        if (!strcmp(buf, "none"))
            translation = BIOS_ATA_TRANSLATION_NONE;
        else if (!strcmp(buf, "lba"))
            translation = BIOS_ATA_TRANSLATION_LBA;
        else if (!strcmp(buf, "auto"))
            translation = BIOS_ATA_TRANSLATION_AUTO;
	else {
360
            error_report("'%s' invalid translation type", buf);
361 362 363 364 365 366 367 368 369
	    return NULL;
	}
    }

    if ((buf = qemu_opt_get(opts, "media")) != NULL) {
        if (!strcmp(buf, "disk")) {
	    media = MEDIA_DISK;
	} else if (!strcmp(buf, "cdrom")) {
            if (cyls || secs || heads) {
370
                error_report("CHS can't be set with media=%s", buf);
371 372 373 374
	        return NULL;
            }
	    media = MEDIA_CDROM;
	} else {
375
	    error_report("'%s' invalid media", buf);
376 377 378 379 380
	    return NULL;
	}
    }

    if ((buf = qemu_opt_get(opts, "cache")) != NULL) {
381 382 383
        if (bdrv_parse_cache_flags(buf, &bdrv_flags) != 0) {
            error_report("invalid cache option");
            return NULL;
384 385 386 387 388 389 390 391 392 393
        }
    }

#ifdef CONFIG_LINUX_AIO
    if ((buf = qemu_opt_get(opts, "aio")) != NULL) {
        if (!strcmp(buf, "native")) {
            bdrv_flags |= BDRV_O_NATIVE_AIO;
        } else if (!strcmp(buf, "threads")) {
            /* this is the default */
        } else {
394
           error_report("invalid aio option");
395 396 397 398 399 400 401
           return NULL;
        }
    }
#endif

    if ((buf = qemu_opt_get(opts, "format")) != NULL) {
       if (strcmp(buf, "?") == 0) {
402 403 404 405
           error_printf("Supported formats:");
           bdrv_iterate_format(bdrv_format_print, NULL);
           error_printf("\n");
           return NULL;
406 407 408
        }
        drv = bdrv_find_whitelisted_format(buf);
        if (!drv) {
409
            error_report("'%s' invalid format", buf);
410 411 412 413
            return NULL;
        }
    }

414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433
    /* disk I/O throttling */
    io_limits.bps[BLOCK_IO_LIMIT_TOTAL]  =
                           qemu_opt_get_number(opts, "bps", 0);
    io_limits.bps[BLOCK_IO_LIMIT_READ]   =
                           qemu_opt_get_number(opts, "bps_rd", 0);
    io_limits.bps[BLOCK_IO_LIMIT_WRITE]  =
                           qemu_opt_get_number(opts, "bps_wr", 0);
    io_limits.iops[BLOCK_IO_LIMIT_TOTAL] =
                           qemu_opt_get_number(opts, "iops", 0);
    io_limits.iops[BLOCK_IO_LIMIT_READ]  =
                           qemu_opt_get_number(opts, "iops_rd", 0);
    io_limits.iops[BLOCK_IO_LIMIT_WRITE] =
                           qemu_opt_get_number(opts, "iops_wr", 0);

    if (!do_check_io_limits(&io_limits)) {
        error_report("bps(iops) and bps_rd/bps_wr(iops_rd/iops_wr) "
                     "cannot be used at the same time");
        return NULL;
    }

434 435 436
    on_write_error = BLOCK_ERR_STOP_ENOSPC;
    if ((buf = qemu_opt_get(opts, "werror")) != NULL) {
        if (type != IF_IDE && type != IF_SCSI && type != IF_VIRTIO && type != IF_NONE) {
437
            error_report("werror is not supported by this bus type");
438 439 440 441 442 443 444 445 446 447 448
            return NULL;
        }

        on_write_error = parse_block_error_action(buf, 0);
        if (on_write_error < 0) {
            return NULL;
        }
    }

    on_read_error = BLOCK_ERR_REPORT;
    if ((buf = qemu_opt_get(opts, "rerror")) != NULL) {
K
Kevin Wolf 已提交
449
        if (type != IF_IDE && type != IF_VIRTIO && type != IF_SCSI && type != IF_NONE) {
450
            error_report("rerror is not supported by this bus type");
451 452 453 454 455 456 457 458 459 460 461
            return NULL;
        }

        on_read_error = parse_block_error_action(buf, 1);
        if (on_read_error < 0) {
            return NULL;
        }
    }

    if ((devaddr = qemu_opt_get(opts, "addr")) != NULL) {
        if (type != IF_VIRTIO) {
462
            error_report("addr is not supported by this bus type");
463 464 465 466 467 468 469 470
            return NULL;
        }
    }

    /* compute bus and unit according index */

    if (index != -1) {
        if (bus_id != 0 || unit_id != -1) {
471
            error_report("index cannot be used with bus and unit");
472 473
            return NULL;
        }
474 475
        bus_id = drive_index_to_bus_id(type, index);
        unit_id = drive_index_to_unit_id(type, index);
476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495
    }

    /* if user doesn't specify a unit_id,
     * try to find the first free
     */

    if (unit_id == -1) {
       unit_id = 0;
       while (drive_get(type, bus_id, unit_id) != NULL) {
           unit_id++;
           if (max_devs && unit_id >= max_devs) {
               unit_id -= max_devs;
               bus_id++;
           }
       }
    }

    /* check unit id */

    if (max_devs && unit_id >= max_devs) {
496 497
        error_report("unit %d too big (max is %d)",
                     unit_id, max_devs - 1);
498 499 500 501
        return NULL;
    }

    /*
502
     * catch multiple definitions
503 504 505
     */

    if (drive_get(type, bus_id, unit_id) != NULL) {
506 507
        error_report("drive with bus=%d, unit=%d (index=%d) exists",
                     bus_id, unit_id, index);
508 509 510 511 512
        return NULL;
    }

    /* init */

513
    dinfo = g_malloc0(sizeof(*dinfo));
514
    if ((buf = qemu_opts_id(opts)) != NULL) {
515
        dinfo->id = g_strdup(buf);
516 517
    } else {
        /* no id supplied -> create one */
518
        dinfo->id = g_malloc0(32);
519 520 521 522
        if (type == IF_IDE || type == IF_SCSI)
            mediastr = (media == MEDIA_CDROM) ? "-cd" : "-hd";
        if (max_devs)
            snprintf(dinfo->id, 32, "%s%i%s%i",
523
                     if_name[type], bus_id, mediastr, unit_id);
524 525
        else
            snprintf(dinfo->id, 32, "%s%s%i",
526
                     if_name[type], mediastr, unit_id);
527 528 529 530 531 532
    }
    dinfo->bdrv = bdrv_new(dinfo->id);
    dinfo->devaddr = devaddr;
    dinfo->type = type;
    dinfo->bus = bus_id;
    dinfo->unit = unit_id;
533 534 535 536
    dinfo->cyls = cyls;
    dinfo->heads = heads;
    dinfo->secs = secs;
    dinfo->trans = translation;
537
    dinfo->opts = opts;
538
    dinfo->refcount = 1;
539
    dinfo->serial = serial;
540 541
    QTAILQ_INSERT_TAIL(&drives, dinfo, next);

542 543
    bdrv_set_on_error(dinfo->bdrv, on_read_error, on_write_error);

544 545 546
    /* disk I/O throttling */
    bdrv_set_io_limits(dinfo->bdrv, &io_limits);

547 548 549 550 551
    switch(type) {
    case IF_IDE:
    case IF_SCSI:
    case IF_XEN:
    case IF_NONE:
552
        dinfo->media_cd = media == MEDIA_CDROM;
553 554 555 556 557 558 559 560
        break;
    case IF_SD:
    case IF_FLOPPY:
    case IF_PFLASH:
    case IF_MTD:
        break;
    case IF_VIRTIO:
        /* add virtio block device */
561
        opts = qemu_opts_create(qemu_find_opts("device"), NULL, 0, NULL);
562 563 564 565 566
        if (arch_type == QEMU_ARCH_S390X) {
            qemu_opt_set(opts, "driver", "virtio-blk-s390");
        } else {
            qemu_opt_set(opts, "driver", "virtio-blk-pci");
        }
567 568 569 570
        qemu_opt_set(opts, "drive", dinfo->id);
        if (devaddr)
            qemu_opt_set(opts, "addr", devaddr);
        break;
571
    default:
572 573
        abort();
    }
574
    if (!file || !*file) {
575
        return dinfo;
576 577 578 579 580 581 582
    }
    if (snapshot) {
        /* always use cache=unsafe with snapshot */
        bdrv_flags &= ~BDRV_O_CACHE_MASK;
        bdrv_flags |= (BDRV_O_SNAPSHOT|BDRV_O_CACHE_WB|BDRV_O_NO_FLUSH);
    }

583 584 585 586
    if (copy_on_read) {
        bdrv_flags |= BDRV_O_COPY_ON_READ;
    }

587 588 589 590
    if (runstate_check(RUN_STATE_INMIGRATE)) {
        bdrv_flags |= BDRV_O_INCOMING;
    }

591 592 593 594
    if (media == MEDIA_CDROM) {
        /* CDROM is fine for any interface, don't check.  */
        ro = 1;
    } else if (ro == 1) {
595 596
        if (type != IF_SCSI && type != IF_VIRTIO && type != IF_FLOPPY &&
            type != IF_NONE && type != IF_PFLASH) {
597
            error_report("readonly not supported by this bus type");
598
            goto err;
599 600 601 602 603
        }
    }

    bdrv_flags |= ro ? 0 : BDRV_O_RDWR;

604 605 606 607
    if (ro && copy_on_read) {
        error_report("warning: disabling copy_on_read on readonly drive");
    }

608 609
    ret = bdrv_open(dinfo->bdrv, file, bdrv_flags, drv);
    if (ret < 0) {
610 611
        error_report("could not open disk image %s: %s",
                     file, strerror(-ret));
612
        goto err;
613 614 615 616 617
    }

    if (bdrv_key_required(dinfo->bdrv))
        autostart = 0;
    return dinfo;
618 619 620

err:
    bdrv_delete(dinfo->bdrv);
621
    g_free(dinfo->id);
622
    QTAILQ_REMOVE(&drives, dinfo, next);
623
    g_free(dinfo);
624
    return NULL;
625 626 627 628 629
}

void do_commit(Monitor *mon, const QDict *qdict)
{
    const char *device = qdict_get_str(qdict, "device");
630
    BlockDriverState *bs;
631
    int ret;
632

633
    if (!strcmp(device, "all")) {
634 635 636 637 638
        ret = bdrv_commit_all();
        if (ret == -EBUSY) {
            qerror_report(QERR_DEVICE_IN_USE, device);
            return;
        }
639 640
    } else {
        bs = bdrv_find(device);
641 642 643
        if (!bs) {
            qerror_report(QERR_DEVICE_NOT_FOUND, device);
            return;
644
        }
645 646 647 648 649
        ret = bdrv_commit(bs);
        if (ret == -EBUSY) {
            qerror_report(QERR_DEVICE_IN_USE, device);
            return;
        }
650 651 652
    }
}

653 654 655 656 657 658 659 660 661 662 663 664
static void blockdev_do_action(int kind, void *data, Error **errp)
{
    BlockdevAction action;
    BlockdevActionList list;

    action.kind = kind;
    action.data = data;
    list.value = &action;
    list.next = NULL;
    qmp_transaction(&list, errp);
}

665 666
void qmp_blockdev_snapshot_sync(const char *device, const char *snapshot_file,
                                bool has_format, const char *format,
667
                                bool has_mode, enum NewImageMode mode,
668
                                Error **errp)
669
{
670 671 672 673 674 675 676 677 678 679
    BlockdevSnapshot snapshot = {
        .device = (char *) device,
        .snapshot_file = (char *) snapshot_file,
        .has_format = has_format,
        .format = (char *) format,
        .has_mode = has_mode,
        .mode = mode,
    };
    blockdev_do_action(BLOCKDEV_ACTION_KIND_BLOCKDEV_SNAPSHOT_SYNC, &snapshot,
                       errp);
680 681
}

682 683

/* New and old BlockDriverState structs for group snapshots */
684
typedef struct BlkTransactionStates {
685 686
    BlockDriverState *old_bs;
    BlockDriverState *new_bs;
687 688
    QSIMPLEQ_ENTRY(BlkTransactionStates) entry;
} BlkTransactionStates;
689 690 691 692 693 694

/*
 * 'Atomic' group snapshots.  The snapshots are taken as a set, and if any fail
 *  then we do not pivot any of the devices in the group, and abandon the
 *  snapshots
 */
695
void qmp_transaction(BlockdevActionList *dev_list, Error **errp)
696 697
{
    int ret = 0;
698 699
    BlockdevActionList *dev_entry = dev_list;
    BlkTransactionStates *states, *next;
700

701
    QSIMPLEQ_HEAD(snap_bdrv_states, BlkTransactionStates) snap_bdrv_states;
702 703 704 705 706 707 708
    QSIMPLEQ_INIT(&snap_bdrv_states);

    /* drain all i/o before any snapshots */
    bdrv_drain_all();

    /* We don't do anything in this loop that commits us to the snapshot */
    while (NULL != dev_entry) {
709 710 711 712
        BlockdevAction *dev_info = NULL;
        BlockDriver *proto_drv;
        BlockDriver *drv;
        int flags;
713 714
        enum NewImageMode mode;
        const char *new_image_file;
715 716 717
        const char *device;
        const char *format = "qcow2";

718 719 720
        dev_info = dev_entry->value;
        dev_entry = dev_entry->next;

721
        states = g_malloc0(sizeof(BlkTransactionStates));
722 723
        QSIMPLEQ_INSERT_TAIL(&snap_bdrv_states, states, entry);

724 725 726
        switch (dev_info->kind) {
        case BLOCKDEV_ACTION_KIND_BLOCKDEV_SNAPSHOT_SYNC:
            device = dev_info->blockdev_snapshot_sync->device;
727 728 729 730
            if (!dev_info->blockdev_snapshot_sync->has_mode) {
                dev_info->blockdev_snapshot_sync->mode = NEW_IMAGE_MODE_ABSOLUTE_PATHS;
            }
            new_image_file = dev_info->blockdev_snapshot_sync->snapshot_file;
731 732 733
            if (dev_info->blockdev_snapshot_sync->has_format) {
                format = dev_info->blockdev_snapshot_sync->format;
            }
734
            mode = dev_info->blockdev_snapshot_sync->mode;
735 736 737 738
            break;
        default:
            abort();
        }
739

740 741 742 743 744 745 746
        drv = bdrv_find_format(format);
        if (!drv) {
            error_set(errp, QERR_INVALID_BLOCK_FORMAT, format);
            goto delete_and_fail;
        }

        states->old_bs = bdrv_find(device);
747
        if (!states->old_bs) {
748
            error_set(errp, QERR_DEVICE_NOT_FOUND, device);
749 750 751
            goto delete_and_fail;
        }

752 753 754 755 756
        if (!bdrv_is_inserted(states->old_bs)) {
            error_set(errp, QERR_DEVICE_HAS_NO_MEDIUM, device);
            goto delete_and_fail;
        }

757
        if (bdrv_in_use(states->old_bs)) {
758
            error_set(errp, QERR_DEVICE_IN_USE, device);
759 760 761
            goto delete_and_fail;
        }

762
        if (!bdrv_is_read_only(states->old_bs)) {
763 764 765 766 767 768 769 770
            if (bdrv_flush(states->old_bs)) {
                error_set(errp, QERR_IO_ERROR);
                goto delete_and_fail;
            }
        }

        flags = states->old_bs->open_flags;

771
        proto_drv = bdrv_find_protocol(new_image_file);
772 773 774 775 776 777
        if (!proto_drv) {
            error_set(errp, QERR_INVALID_BLOCK_FORMAT, format);
            goto delete_and_fail;
        }

        /* create new image w/backing file */
778 779 780 781 782 783 784 785 786
        if (mode != NEW_IMAGE_MODE_EXISTING) {
            ret = bdrv_img_create(new_image_file, format,
                                  states->old_bs->filename,
                                  states->old_bs->drv->format_name,
                                  NULL, -1, flags);
            if (ret) {
                error_set(errp, QERR_OPEN_FILE_FAILED, new_image_file);
                goto delete_and_fail;
            }
787 788 789 790
        }

        /* We will manually add the backing_hd field to the bs later */
        states->new_bs = bdrv_new("");
791
        ret = bdrv_open(states->new_bs, new_image_file,
792 793
                        flags | BDRV_O_NO_BACKING, drv);
        if (ret != 0) {
794
            error_set(errp, QERR_OPEN_FILE_FAILED, new_image_file);
795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820
            goto delete_and_fail;
        }
    }


    /* Now we are going to do the actual pivot.  Everything up to this point
     * is reversible, but we are committed at this point */
    QSIMPLEQ_FOREACH(states, &snap_bdrv_states, entry) {
        /* This removes our old bs from the bdrv_states, and adds the new bs */
        bdrv_append(states->new_bs, states->old_bs);
    }

    /* success */
    goto exit;

delete_and_fail:
    /*
    * failure, and it is all-or-none; abandon each new bs, and keep using
    * the original bs for all images
    */
    QSIMPLEQ_FOREACH(states, &snap_bdrv_states, entry) {
        if (states->new_bs) {
             bdrv_delete(states->new_bs);
        }
    }
exit:
821
    QSIMPLEQ_FOREACH_SAFE(states, &snap_bdrv_states, entry, next) {
822 823 824 825 826 827
        g_free(states);
    }
    return;
}


828
static void eject_device(BlockDriverState *bs, int force, Error **errp)
829
{
830 831 832 833
    if (bdrv_in_use(bs)) {
        error_set(errp, QERR_DEVICE_IN_USE, bdrv_get_device_name(bs));
        return;
    }
834
    if (!bdrv_dev_has_removable_media(bs)) {
835 836
        error_set(errp, QERR_DEVICE_NOT_REMOVABLE, bdrv_get_device_name(bs));
        return;
837
    }
838

P
Paolo Bonzini 已提交
839 840 841
    if (bdrv_dev_is_medium_locked(bs) && !bdrv_dev_is_tray_open(bs)) {
        bdrv_dev_eject_request(bs, force);
        if (!force) {
842 843
            error_set(errp, QERR_DEVICE_LOCKED, bdrv_get_device_name(bs));
            return;
P
Paolo Bonzini 已提交
844
        }
845
    }
846

847
    bdrv_close(bs);
848 849
}

L
Luiz Capitulino 已提交
850
void qmp_eject(const char *device, bool has_force, bool force, Error **errp)
851 852 853
{
    BlockDriverState *bs;

L
Luiz Capitulino 已提交
854
    bs = bdrv_find(device);
855
    if (!bs) {
L
Luiz Capitulino 已提交
856 857
        error_set(errp, QERR_DEVICE_NOT_FOUND, device);
        return;
858 859
    }

L
Luiz Capitulino 已提交
860
    eject_device(bs, force, errp);
861 862
}

L
Luiz Capitulino 已提交
863
void qmp_block_passwd(const char *device, const char *password, Error **errp)
864 865 866 867
{
    BlockDriverState *bs;
    int err;

L
Luiz Capitulino 已提交
868
    bs = bdrv_find(device);
869
    if (!bs) {
L
Luiz Capitulino 已提交
870 871
        error_set(errp, QERR_DEVICE_NOT_FOUND, device);
        return;
872 873
    }

L
Luiz Capitulino 已提交
874
    err = bdrv_set_key(bs, password);
875
    if (err == -EINVAL) {
L
Luiz Capitulino 已提交
876 877
        error_set(errp, QERR_DEVICE_NOT_ENCRYPTED, bdrv_get_device_name(bs));
        return;
878
    } else if (err < 0) {
L
Luiz Capitulino 已提交
879 880
        error_set(errp, QERR_INVALID_PASSWORD);
        return;
881 882 883
    }
}

L
Luiz Capitulino 已提交
884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908
static void qmp_bdrv_open_encrypted(BlockDriverState *bs, const char *filename,
                                    int bdrv_flags, BlockDriver *drv,
                                    const char *password, Error **errp)
{
    if (bdrv_open(bs, filename, bdrv_flags, drv) < 0) {
        error_set(errp, QERR_OPEN_FILE_FAILED, filename);
        return;
    }

    if (bdrv_key_required(bs)) {
        if (password) {
            if (bdrv_set_key(bs, password) < 0) {
                error_set(errp, QERR_INVALID_PASSWORD);
            }
        } else {
            error_set(errp, QERR_DEVICE_ENCRYPTED, bdrv_get_device_name(bs),
                      bdrv_get_encrypted_filename(bs));
        }
    } else if (password) {
        error_set(errp, QERR_DEVICE_NOT_ENCRYPTED, bdrv_get_device_name(bs));
    }
}

void qmp_change_blockdev(const char *device, const char *filename,
                         bool has_format, const char *format, Error **errp)
909 910 911 912
{
    BlockDriverState *bs;
    BlockDriver *drv = NULL;
    int bdrv_flags;
913
    Error *err = NULL;
914 915 916

    bs = bdrv_find(device);
    if (!bs) {
L
Luiz Capitulino 已提交
917 918
        error_set(errp, QERR_DEVICE_NOT_FOUND, device);
        return;
919
    }
L
Luiz Capitulino 已提交
920 921 922

    if (format) {
        drv = bdrv_find_whitelisted_format(format);
923
        if (!drv) {
L
Luiz Capitulino 已提交
924 925
            error_set(errp, QERR_INVALID_BLOCK_FORMAT, format);
            return;
926 927
        }
    }
L
Luiz Capitulino 已提交
928

929 930
    eject_device(bs, 0, &err);
    if (error_is_set(&err)) {
L
Luiz Capitulino 已提交
931 932
        error_propagate(errp, err);
        return;
933
    }
L
Luiz Capitulino 已提交
934

935
    bdrv_flags = bdrv_is_read_only(bs) ? 0 : BDRV_O_RDWR;
936
    bdrv_flags |= bdrv_is_snapshot(bs) ? BDRV_O_SNAPSHOT : 0;
L
Luiz Capitulino 已提交
937 938

    qmp_bdrv_open_encrypted(bs, filename, bdrv_flags, drv, NULL, errp);
939
}
940

Z
Zhi Yong Wu 已提交
941
/* throttling disk I/O limits */
942 943 944
void qmp_block_set_io_throttle(const char *device, int64_t bps, int64_t bps_rd,
                               int64_t bps_wr, int64_t iops, int64_t iops_rd,
                               int64_t iops_wr, Error **errp)
Z
Zhi Yong Wu 已提交
945 946 947 948
{
    BlockIOLimit io_limits;
    BlockDriverState *bs;

949
    bs = bdrv_find(device);
Z
Zhi Yong Wu 已提交
950
    if (!bs) {
951 952
        error_set(errp, QERR_DEVICE_NOT_FOUND, device);
        return;
Z
Zhi Yong Wu 已提交
953 954
    }

955 956 957 958 959 960
    io_limits.bps[BLOCK_IO_LIMIT_TOTAL] = bps;
    io_limits.bps[BLOCK_IO_LIMIT_READ]  = bps_rd;
    io_limits.bps[BLOCK_IO_LIMIT_WRITE] = bps_wr;
    io_limits.iops[BLOCK_IO_LIMIT_TOTAL]= iops;
    io_limits.iops[BLOCK_IO_LIMIT_READ] = iops_rd;
    io_limits.iops[BLOCK_IO_LIMIT_WRITE]= iops_wr;
Z
Zhi Yong Wu 已提交
961 962

    if (!do_check_io_limits(&io_limits)) {
963 964
        error_set(errp, QERR_INVALID_PARAMETER_COMBINATION);
        return;
Z
Zhi Yong Wu 已提交
965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980
    }

    bs->io_limits = io_limits;
    bs->slice_time = BLOCK_IO_SLICE_TIME;

    if (!bs->io_limits_enabled && bdrv_io_limits_enabled(bs)) {
        bdrv_io_limits_enable(bs);
    } else if (bs->io_limits_enabled && !bdrv_io_limits_enabled(bs)) {
        bdrv_io_limits_disable(bs);
    } else {
        if (bs->block_timer) {
            qemu_mod_timer(bs->block_timer, qemu_get_clock_ns(vm_clock));
        }
    }
}

981 982 983 984 985 986 987 988 989 990
int do_drive_del(Monitor *mon, const QDict *qdict, QObject **ret_data)
{
    const char *id = qdict_get_str(qdict, "id");
    BlockDriverState *bs;

    bs = bdrv_find(id);
    if (!bs) {
        qerror_report(QERR_DEVICE_NOT_FOUND, id);
        return -1;
    }
M
Marcelo Tosatti 已提交
991 992 993 994
    if (bdrv_in_use(bs)) {
        qerror_report(QERR_DEVICE_IN_USE, id);
        return -1;
    }
995 996

    /* quiesce block driver; prevent further io */
997
    bdrv_drain_all();
998 999 1000
    bdrv_flush(bs);
    bdrv_close(bs);

1001
    /* if we have a device attached to this BlockDriverState
1002 1003 1004 1005
     * then we need to make the drive anonymous until the device
     * can be removed.  If this is a drive with no device backing
     * then we can just get rid of the block driver state right here.
     */
1006
    if (bdrv_get_attached_dev(bs)) {
1007 1008 1009
        bdrv_make_anon(bs);
    } else {
        drive_uninit(drive_get_by_blockdev(bs));
1010 1011 1012 1013
    }

    return 0;
}
1014

L
Luiz Capitulino 已提交
1015
void qmp_block_resize(const char *device, int64_t size, Error **errp)
1016 1017 1018 1019 1020
{
    BlockDriverState *bs;

    bs = bdrv_find(device);
    if (!bs) {
L
Luiz Capitulino 已提交
1021 1022
        error_set(errp, QERR_DEVICE_NOT_FOUND, device);
        return;
1023 1024 1025
    }

    if (size < 0) {
1026
        error_set(errp, QERR_INVALID_PARAMETER_VALUE, "size", "a >0 size");
L
Luiz Capitulino 已提交
1027
        return;
1028 1029
    }

1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045
    switch (bdrv_truncate(bs, size)) {
    case 0:
        break;
    case -ENOMEDIUM:
        error_set(errp, QERR_DEVICE_HAS_NO_MEDIUM, device);
        break;
    case -ENOTSUP:
        error_set(errp, QERR_UNSUPPORTED);
        break;
    case -EACCES:
        error_set(errp, QERR_DEVICE_IS_READ_ONLY, device);
        break;
    case -EBUSY:
        error_set(errp, QERR_DEVICE_IN_USE, device);
        break;
    default:
L
Luiz Capitulino 已提交
1046
        error_set(errp, QERR_UNDEFINED_ERROR);
1047
        break;
1048 1049
    }
}
S
Stefan Hajnoczi 已提交
1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078

static QObject *qobject_from_block_job(BlockJob *job)
{
    return qobject_from_jsonf("{ 'type': %s,"
                              "'device': %s,"
                              "'len': %" PRId64 ","
                              "'offset': %" PRId64 ","
                              "'speed': %" PRId64 " }",
                              job->job_type->job_type,
                              bdrv_get_device_name(job->bs),
                              job->len,
                              job->offset,
                              job->speed);
}

static void block_stream_cb(void *opaque, int ret)
{
    BlockDriverState *bs = opaque;
    QObject *obj;

    trace_block_stream_cb(bs, bs->job, ret);

    assert(bs->job);
    obj = qobject_from_block_job(bs->job);
    if (ret < 0) {
        QDict *dict = qobject_to_qdict(obj);
        qdict_put(dict, "error", qstring_from_str(strerror(-ret)));
    }

1079 1080 1081 1082 1083
    if (block_job_is_cancelled(bs->job)) {
        monitor_protocol_event(QEVENT_BLOCK_JOB_CANCELLED, obj);
    } else {
        monitor_protocol_event(QEVENT_BLOCK_JOB_COMPLETED, obj);
    }
S
Stefan Hajnoczi 已提交
1084
    qobject_decref(obj);
1085 1086

    drive_put_ref_bh_schedule(drive_get_by_blockdev(bs));
S
Stefan Hajnoczi 已提交
1087 1088 1089
}

void qmp_block_stream(const char *device, bool has_base,
1090 1091
                      const char *base, bool has_speed,
                      int64_t speed, Error **errp)
S
Stefan Hajnoczi 已提交
1092 1093
{
    BlockDriverState *bs;
1094
    BlockDriverState *base_bs = NULL;
1095
    Error *local_err = NULL;
S
Stefan Hajnoczi 已提交
1096 1097 1098 1099 1100 1101 1102 1103

    bs = bdrv_find(device);
    if (!bs) {
        error_set(errp, QERR_DEVICE_NOT_FOUND, device);
        return;
    }

    if (base) {
1104 1105 1106 1107 1108
        base_bs = bdrv_find_backing_image(bs, base);
        if (base_bs == NULL) {
            error_set(errp, QERR_BASE_NOT_FOUND, base);
            return;
        }
S
Stefan Hajnoczi 已提交
1109 1110
    }

1111 1112
    stream_start(bs, base_bs, base, has_speed ? speed : 0,
                 block_stream_cb, bs, &local_err);
1113 1114 1115
    if (error_is_set(&local_err)) {
        error_propagate(errp, local_err);
        return;
S
Stefan Hajnoczi 已提交
1116 1117
    }

1118 1119 1120 1121 1122
    /* Grab a reference so hotplug does not delete the BlockDriverState from
     * underneath us.
     */
    drive_get_ref(drive_get_by_blockdev(bs));

S
Stefan Hajnoczi 已提交
1123 1124
    trace_qmp_block_stream(bs, bs->job);
}
1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136

static BlockJob *find_block_job(const char *device)
{
    BlockDriverState *bs;

    bs = bdrv_find(device);
    if (!bs || !bs->job) {
        return NULL;
    }
    return bs->job;
}

1137
void qmp_block_job_set_speed(const char *device, int64_t speed, Error **errp)
1138 1139 1140 1141 1142 1143 1144 1145
{
    BlockJob *job = find_block_job(device);

    if (!job) {
        error_set(errp, QERR_DEVICE_NOT_ACTIVE, device);
        return;
    }

1146
    block_job_set_speed(job, speed, errp);
1147
}
1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160

void qmp_block_job_cancel(const char *device, Error **errp)
{
    BlockJob *job = find_block_job(device);

    if (!job) {
        error_set(errp, QERR_DEVICE_NOT_ACTIVE, device);
        return;
    }

    trace_qmp_block_job_cancel(job);
    block_job_cancel(job);
}
S
Stefan Hajnoczi 已提交
1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193

static void do_qmp_query_block_jobs_one(void *opaque, BlockDriverState *bs)
{
    BlockJobInfoList **prev = opaque;
    BlockJob *job = bs->job;

    if (job) {
        BlockJobInfoList *elem;
        BlockJobInfo *info = g_new(BlockJobInfo, 1);
        *info = (BlockJobInfo){
            .type   = g_strdup(job->job_type->job_type),
            .device = g_strdup(bdrv_get_device_name(bs)),
            .len    = job->len,
            .offset = job->offset,
            .speed  = job->speed,
        };

        elem = g_new0(BlockJobInfoList, 1);
        elem->value = info;

        (*prev)->next = elem;
        *prev = elem;
    }
}

BlockJobInfoList *qmp_query_block_jobs(Error **errp)
{
    /* Dummy is a fake list element for holding the head pointer */
    BlockJobInfoList dummy = {};
    BlockJobInfoList *prev = &dummy;
    bdrv_iterate(do_qmp_query_block_jobs_one, &prev);
    return dummy.next;
}