blockdev.c 87.3 KB
Newer Older
1 2 3 4 5 6 7
/*
 * 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.
8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
 *
 * This file incorporates work covered by the following copyright and
 * permission notice:
 *
 * Copyright (c) 2003-2008 Fabrice Bellard
 *
 * 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.
31 32
 */

M
Markus Armbruster 已提交
33
#include "sysemu/block-backend.h"
34
#include "sysemu/blockdev.h"
P
Paolo Bonzini 已提交
35
#include "hw/block/block.h"
36
#include "block/blockjob.h"
37
#include "monitor/monitor.h"
38 39
#include "qemu/option.h"
#include "qemu/config-file.h"
40
#include "qapi/qmp/types.h"
K
Kevin Wolf 已提交
41 42
#include "qapi-visit.h"
#include "qapi/qmp-output-visitor.h"
43
#include "qapi/util.h"
44
#include "sysemu/sysemu.h"
45
#include "block/block_int.h"
L
Luiz Capitulino 已提交
46
#include "qmp-commands.h"
S
Stefan Hajnoczi 已提交
47
#include "trace.h"
48
#include "sysemu/arch_init.h"
49

50 51 52 53 54 55 56 57 58 59 60 61
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",
};

62
static int if_max_devs[IF_COUNT] = {
63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78
    /*
     * 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,
79 80
};

81 82 83 84 85 86
/**
 * Boards may call this to offer board-by-board overrides
 * of the default, global values.
 */
void override_max_devs(BlockInterfaceType type, int max_devs)
{
87
    BlockBackend *blk;
88 89 90 91 92 93
    DriveInfo *dinfo;

    if (max_devs <= 0) {
        return;
    }

94 95
    for (blk = blk_next(NULL); blk; blk = blk_next(blk)) {
        dinfo = blk_legacy_dinfo(blk);
96 97 98 99 100 101 102 103 104 105 106
        if (dinfo->type == type) {
            fprintf(stderr, "Cannot override units-per-bus property of"
                    " the %s interface, because a drive of that type has"
                    " already been added.\n", if_name[type]);
            g_assert_not_reached();
        }
    }

    if_max_devs[type] = max_devs;
}

107 108 109 110 111 112 113
/*
 * 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.
 */
114
void blockdev_mark_auto_del(BlockBackend *blk)
115
{
116
    DriveInfo *dinfo = blk_legacy_dinfo(blk);
117
    BlockDriverState *bs = blk_bs(blk);
118
    AioContext *aio_context;
119

120
    if (!dinfo) {
121 122 123
        return;
    }

124 125 126
    aio_context = bdrv_get_aio_context(bs);
    aio_context_acquire(aio_context);

127 128 129
    if (bs->job) {
        block_job_cancel(bs->job);
    }
130 131 132

    aio_context_release(aio_context);

133
    dinfo->auto_del = 1;
134 135
}

136
void blockdev_auto_del(BlockBackend *blk)
137
{
138
    DriveInfo *dinfo = blk_legacy_dinfo(blk);
139

140
    if (dinfo && dinfo->auto_del) {
141
        blk_unref(blk);
142 143 144
    }
}

145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161
/**
 * Returns the current mapping of how many units per bus
 * a particular interface can support.
 *
 *  A positive integer indicates n units per bus.
 *  0 implies the mapping has not been established.
 * -1 indicates an invalid BlockInterfaceType was given.
 */
int drive_get_max_devs(BlockInterfaceType type)
{
    if (type >= IF_IDE && type < IF_COUNT) {
        return if_max_devs[type];
    }

    return -1;
}

162 163 164 165 166 167 168 169 170 171 172 173
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;
}

174 175 176 177 178 179
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,
180
                    const char *optstr)
181 182
{
    QemuOpts *opts;
183
    char buf[32];
184

185
    opts = drive_def(optstr);
186 187 188
    if (!opts) {
        return NULL;
    }
189 190 191 192 193 194 195
    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);
    }
196 197 198 199 200 201 202
    if (file)
        qemu_opt_set(opts, "file", file);
    return opts;
}

DriveInfo *drive_get(BlockInterfaceType type, int bus, int unit)
{
203
    BlockBackend *blk;
204 205
    DriveInfo *dinfo;

206 207 208 209
    for (blk = blk_next(NULL); blk; blk = blk_next(blk)) {
        dinfo = blk_legacy_dinfo(blk);
        if (dinfo && dinfo->type == type
            && dinfo->bus == bus && dinfo->unit == unit) {
210
            return dinfo;
211
        }
212 213 214 215 216
    }

    return NULL;
}

J
John Snow 已提交
217 218
bool drive_check_orphaned(void)
{
219
    BlockBackend *blk;
J
John Snow 已提交
220 221 222
    DriveInfo *dinfo;
    bool rs = false;

223 224
    for (blk = blk_next(NULL); blk; blk = blk_next(blk)) {
        dinfo = blk_legacy_dinfo(blk);
J
John Snow 已提交
225 226
        /* If dinfo->bdrv->dev is NULL, it has no device attached. */
        /* Unless this is a default drive, this may be an oversight. */
227
        if (!blk_get_attached_dev(blk) && !dinfo->is_default &&
J
John Snow 已提交
228 229 230
            dinfo->type != IF_NONE) {
            fprintf(stderr, "Warning: Orphaned drive without device: "
                    "id=%s,file=%s,if=%s,bus=%d,unit=%d\n",
231
                    blk_name(blk), blk_bs(blk)->filename, if_name[dinfo->type],
J
John Snow 已提交
232 233 234 235 236 237 238 239
                    dinfo->bus, dinfo->unit);
            rs = true;
        }
    }

    return rs;
}

240 241 242 243 244 245 246
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));
}

247 248 249
int drive_get_max_bus(BlockInterfaceType type)
{
    int max_bus;
250
    BlockBackend *blk;
251 252 253
    DriveInfo *dinfo;

    max_bus = -1;
254 255 256
    for (blk = blk_next(NULL); blk; blk = blk_next(blk)) {
        dinfo = blk_legacy_dinfo(blk);
        if (dinfo && dinfo->type == type && dinfo->bus > max_bus) {
257
            max_bus = dinfo->bus;
258
        }
259 260 261 262
    }
    return max_bus;
}

263 264 265 266 267 268 269 270 271 272
/* 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]++);
}

273 274
static void bdrv_format_print(void *opaque, const char *name)
{
275
    error_printf(" %s", name);
276 277
}

278 279
typedef struct {
    QEMUBH *bh;
F
Fam Zheng 已提交
280 281
    BlockDriverState *bs;
} BDRVPutRefBH;
282

F
Fam Zheng 已提交
283
static void bdrv_put_ref_bh(void *opaque)
284
{
F
Fam Zheng 已提交
285
    BDRVPutRefBH *s = opaque;
286

F
Fam Zheng 已提交
287
    bdrv_unref(s->bs);
288 289 290 291 292
    qemu_bh_delete(s->bh);
    g_free(s);
}

/*
F
Fam Zheng 已提交
293
 * Release a BDS reference in a BH
294
 *
F
Fam Zheng 已提交
295 296 297
 * It is not safe to use bdrv_unref() from a callback function when the callers
 * still need the BlockDriverState.  In such cases we schedule a BH to release
 * the reference.
298
 */
F
Fam Zheng 已提交
299
static void bdrv_put_ref_bh_schedule(BlockDriverState *bs)
300
{
F
Fam Zheng 已提交
301
    BDRVPutRefBH *s;
302

F
Fam Zheng 已提交
303 304 305
    s = g_new(BDRVPutRefBH, 1);
    s->bh = qemu_bh_new(bdrv_put_ref_bh, s);
    s->bs = bs;
306 307 308
    qemu_bh_schedule(s->bh);
}

309
static int parse_block_error_action(const char *buf, bool is_read, Error **errp)
310 311
{
    if (!strcmp(buf, "ignore")) {
312
        return BLOCKDEV_ON_ERROR_IGNORE;
313
    } else if (!is_read && !strcmp(buf, "enospc")) {
314
        return BLOCKDEV_ON_ERROR_ENOSPC;
315
    } else if (!strcmp(buf, "stop")) {
316
        return BLOCKDEV_ON_ERROR_STOP;
317
    } else if (!strcmp(buf, "report")) {
318
        return BLOCKDEV_ON_ERROR_REPORT;
319
    } else {
320 321
        error_setg(errp, "'%s' invalid %s error action",
                   buf, is_read ? "read" : "write");
322 323 324 325
        return -1;
    }
}

326
static bool check_throttle_config(ThrottleConfig *cfg, Error **errp)
327
{
328 329 330
    if (throttle_conflicting(cfg)) {
        error_setg(errp, "bps/iops/max total values and read/write values"
                         " cannot be used at the same time");
331 332 333
        return false;
    }

334 335
    if (!throttle_is_valid(cfg)) {
        error_setg(errp, "bps/iops/maxs values must be 0 or greater");
336 337 338
        return false;
    }

339 340 341
    return true;
}

342 343
typedef enum { MEDIA_DISK, MEDIA_CDROM } DriveMediaType;

344
/* Takes the ownership of bs_opts */
345 346
static BlockBackend *blockdev_init(const char *file, QDict *bs_opts,
                                   Error **errp)
347 348 349 350 351
{
    const char *buf;
    int ro = 0;
    int bdrv_flags = 0;
    int on_read_error, on_write_error;
M
Markus Armbruster 已提交
352
    BlockBackend *blk;
353
    BlockDriverState *bs;
354
    ThrottleConfig cfg;
355
    int snapshot = 0;
356
    bool copy_on_read;
357
    int ret;
358
    Error *error = NULL;
359 360
    QemuOpts *opts;
    const char *id;
361
    bool has_driver_specific_opts;
362
    BlockdevDetectZeroesOptions detect_zeroes;
363
    BlockDriver *drv = NULL;
364

365 366 367
    /* Check common options by copying from bs_opts to opts, all other options
     * stay in bs_opts for processing by bdrv_open(). */
    id = qdict_get_try_str(bs_opts, "id");
368
    opts = qemu_opts_create(&qemu_common_drive_opts, id, 1, &error);
369
    if (error) {
370
        error_propagate(errp, error);
371
        goto err_no_opts;
372 373 374
    }

    qemu_opts_absorb_qdict(opts, bs_opts, &error);
375
    if (error) {
376
        error_propagate(errp, error);
377
        goto early_err;
378 379 380 381 382 383
    }

    if (id) {
        qdict_del(bs_opts, "id");
    }

384 385
    has_driver_specific_opts = !!qdict_size(bs_opts);

386 387
    /* extract parameters */
    snapshot = qemu_opt_get_bool(opts, "snapshot", 0);
388
    ro = qemu_opt_get_bool(opts, "read-only", 0);
389
    copy_on_read = qemu_opt_get_bool(opts, "copy-on-read", false);
390

391 392
    if ((buf = qemu_opt_get(opts, "discard")) != NULL) {
        if (bdrv_parse_discard_flags(buf, &bdrv_flags) != 0) {
393
            error_setg(errp, "invalid discard option");
394
            goto early_err;
395 396 397
        }
    }

K
Kevin Wolf 已提交
398 399 400 401 402 403
    if (qemu_opt_get_bool(opts, "cache.writeback", true)) {
        bdrv_flags |= BDRV_O_CACHE_WB;
    }
    if (qemu_opt_get_bool(opts, "cache.direct", false)) {
        bdrv_flags |= BDRV_O_NOCACHE;
    }
404
    if (qemu_opt_get_bool(opts, "cache.no-flush", false)) {
K
Kevin Wolf 已提交
405
        bdrv_flags |= BDRV_O_NO_FLUSH;
406 407 408 409 410 411 412 413 414
    }

#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 {
415
           error_setg(errp, "invalid aio option");
416
           goto early_err;
417 418 419 420 421
        }
    }
#endif

    if ((buf = qemu_opt_get(opts, "format")) != NULL) {
422 423 424 425
        if (is_help_option(buf)) {
            error_printf("Supported formats:");
            bdrv_iterate_format(bdrv_format_print, NULL);
            error_printf("\n");
426
            goto early_err;
427
        }
428

429
        drv = bdrv_find_format(buf);
430
        if (!drv) {
431
            error_setg(errp, "'%s' invalid format", buf);
432
            goto early_err;
433
        }
434 435
    }

436
    /* disk I/O throttling */
437 438
    memset(&cfg, 0, sizeof(cfg));
    cfg.buckets[THROTTLE_BPS_TOTAL].avg =
439
        qemu_opt_get_number(opts, "throttling.bps-total", 0);
440
    cfg.buckets[THROTTLE_BPS_READ].avg  =
441
        qemu_opt_get_number(opts, "throttling.bps-read", 0);
442
    cfg.buckets[THROTTLE_BPS_WRITE].avg =
443
        qemu_opt_get_number(opts, "throttling.bps-write", 0);
444
    cfg.buckets[THROTTLE_OPS_TOTAL].avg =
445
        qemu_opt_get_number(opts, "throttling.iops-total", 0);
446
    cfg.buckets[THROTTLE_OPS_READ].avg =
447
        qemu_opt_get_number(opts, "throttling.iops-read", 0);
448
    cfg.buckets[THROTTLE_OPS_WRITE].avg =
449
        qemu_opt_get_number(opts, "throttling.iops-write", 0);
450

451 452 453 454 455 456 457 458 459 460 461 462
    cfg.buckets[THROTTLE_BPS_TOTAL].max =
        qemu_opt_get_number(opts, "throttling.bps-total-max", 0);
    cfg.buckets[THROTTLE_BPS_READ].max  =
        qemu_opt_get_number(opts, "throttling.bps-read-max", 0);
    cfg.buckets[THROTTLE_BPS_WRITE].max =
        qemu_opt_get_number(opts, "throttling.bps-write-max", 0);
    cfg.buckets[THROTTLE_OPS_TOTAL].max =
        qemu_opt_get_number(opts, "throttling.iops-total-max", 0);
    cfg.buckets[THROTTLE_OPS_READ].max =
        qemu_opt_get_number(opts, "throttling.iops-read-max", 0);
    cfg.buckets[THROTTLE_OPS_WRITE].max =
        qemu_opt_get_number(opts, "throttling.iops-write-max", 0);
463

464
    cfg.op_size = qemu_opt_get_number(opts, "throttling.iops-size", 0);
465 466

    if (!check_throttle_config(&cfg, &error)) {
467
        error_propagate(errp, error);
468
        goto early_err;
469 470
    }

471
    on_write_error = BLOCKDEV_ON_ERROR_ENOSPC;
472
    if ((buf = qemu_opt_get(opts, "werror")) != NULL) {
473
        on_write_error = parse_block_error_action(buf, 0, &error);
474
        if (error) {
475
            error_propagate(errp, error);
476
            goto early_err;
477 478 479
        }
    }

480
    on_read_error = BLOCKDEV_ON_ERROR_REPORT;
481
    if ((buf = qemu_opt_get(opts, "rerror")) != NULL) {
482
        on_read_error = parse_block_error_action(buf, 1, &error);
483
        if (error) {
484
            error_propagate(errp, error);
485
            goto early_err;
486 487 488
        }
    }

489
    detect_zeroes =
490 491 492 493 494
        qapi_enum_parse(BlockdevDetectZeroesOptions_lookup,
                        qemu_opt_get(opts, "detect-zeroes"),
                        BLOCKDEV_DETECT_ZEROES_OPTIONS_MAX,
                        BLOCKDEV_DETECT_ZEROES_OPTIONS_OFF,
                        &error);
495 496 497 498 499 500 501 502 503 504 505 506
    if (error) {
        error_propagate(errp, error);
        goto early_err;
    }

    if (detect_zeroes == BLOCKDEV_DETECT_ZEROES_OPTIONS_UNMAP &&
        !(bdrv_flags & BDRV_O_UNMAP)) {
        error_setg(errp, "setting detect-zeroes to unmap is not allowed "
                         "without setting discard operation to unmap");
        goto early_err;
    }

507
    /* init */
508
    blk = blk_new_with_bs(qemu_opts_id(opts), errp);
M
Markus Armbruster 已提交
509 510 511
    if (!blk) {
        goto early_err;
    }
512
    bs = blk_bs(blk);
513 514 515
    bs->open_flags = snapshot ? BDRV_O_SNAPSHOT : 0;
    bs->read_only = ro;
    bs->detect_zeroes = detect_zeroes;
516

517
    bdrv_set_on_error(bs, on_read_error, on_write_error);
518

519
    /* disk I/O throttling */
520
    if (throttle_enabled(&cfg)) {
521 522
        bdrv_io_limits_enable(bs);
        bdrv_set_io_limits(bs, &cfg);
523
    }
524

525
    if (!file || !*file) {
526
        if (has_driver_specific_opts) {
527 528
            file = NULL;
        } else {
529 530
            QDECREF(bs_opts);
            qemu_opts_del(opts);
531
            return blk;
532
        }
533 534 535 536 537 538 539
    }
    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);
    }

540 541 542 543
    if (copy_on_read) {
        bdrv_flags |= BDRV_O_COPY_ON_READ;
    }

544 545 546 547
    if (runstate_check(RUN_STATE_INMIGRATE)) {
        bdrv_flags |= BDRV_O_INCOMING;
    }

548 549
    bdrv_flags |= ro ? 0 : BDRV_O_RDWR;

550
    QINCREF(bs_opts);
551
    ret = bdrv_open(&bs, file, NULL, bs_opts, bdrv_flags, drv, &error);
552
    assert(bs == blk_bs(blk));
553

554
    if (ret < 0) {
555
        error_setg(errp, "could not open disk image %s: %s",
556
                   file ?: blk_name(blk), error_get_pretty(error));
557
        error_free(error);
558
        goto err;
559 560
    }

561
    if (bdrv_key_required(bs)) {
562
        autostart = 0;
563
    }
564

565
    QDECREF(bs_opts);
566 567
    qemu_opts_del(opts);

568
    return blk;
569 570

err:
M
Markus Armbruster 已提交
571
    blk_unref(blk);
572 573
early_err:
    qemu_opts_del(opts);
574 575
err_no_opts:
    QDECREF(bs_opts);
576
    return NULL;
577 578
}

579 580
static void qemu_opt_rename(QemuOpts *opts, const char *from, const char *to,
                            Error **errp)
581 582 583 584 585
{
    const char *value;

    value = qemu_opt_get(opts, from);
    if (value) {
586 587 588 589 590
        if (qemu_opt_find(opts, to)) {
            error_setg(errp, "'%s' and its alias '%s' can't be used at the "
                       "same time", to, from);
            return;
        }
591 592 593 594
    }

    /* rename all items in opts */
    while ((value = qemu_opt_get(opts, from))) {
595 596 597 598 599
        qemu_opt_set(opts, to, value);
        qemu_opt_unset(opts, from);
    }
}

600 601 602 603 604
QemuOptsList qemu_legacy_drive_opts = {
    .name = "drive",
    .head = QTAILQ_HEAD_INITIALIZER(qemu_legacy_drive_opts.head),
    .desc = {
        {
605 606 607 608 609 610 611 612 613 614 615 616
            .name = "bus",
            .type = QEMU_OPT_NUMBER,
            .help = "bus number",
        },{
            .name = "unit",
            .type = QEMU_OPT_NUMBER,
            .help = "unit number (i.e. lun for scsi)",
        },{
            .name = "index",
            .type = QEMU_OPT_NUMBER,
            .help = "index number",
        },{
617 618 619
            .name = "media",
            .type = QEMU_OPT_STRING,
            .help = "media type (disk, cdrom)",
620 621 622 623
        },{
            .name = "if",
            .type = QEMU_OPT_STRING,
            .help = "interface (ide, scsi, sd, mtd, floppy, pflash, virtio)",
624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639
        },{
            .name = "cyls",
            .type = QEMU_OPT_NUMBER,
            .help = "number of cylinders (ide disk geometry)",
        },{
            .name = "heads",
            .type = QEMU_OPT_NUMBER,
            .help = "number of heads (ide disk geometry)",
        },{
            .name = "secs",
            .type = QEMU_OPT_NUMBER,
            .help = "number of sectors (ide disk geometry)",
        },{
            .name = "trans",
            .type = QEMU_OPT_STRING,
            .help = "chs translation (auto, lba, none)",
640 641 642 643
        },{
            .name = "boot",
            .type = QEMU_OPT_BOOL,
            .help = "(deprecated, ignored)",
644 645 646 647
        },{
            .name = "addr",
            .type = QEMU_OPT_STRING,
            .help = "pci address (virtio only)",
648 649 650 651
        },{
            .name = "serial",
            .type = QEMU_OPT_STRING,
            .help = "disk serial number",
M
Max Reitz 已提交
652 653 654 655
        },{
            .name = "file",
            .type = QEMU_OPT_STRING,
            .help = "file name",
656
        },
657 658 659 660 661 662

        /* Options that are passed on, but have special semantics with -drive */
        {
            .name = "read-only",
            .type = QEMU_OPT_BOOL,
            .help = "open drive file as read-only",
663 664 665 666 667 668 669 670
        },{
            .name = "rerror",
            .type = QEMU_OPT_STRING,
            .help = "read error action",
        },{
            .name = "werror",
            .type = QEMU_OPT_STRING,
            .help = "write error action",
671 672 673 674 675 676
        },{
            .name = "copy-on-read",
            .type = QEMU_OPT_BOOL,
            .help = "copy read data from backing file into image file",
        },

677 678 679 680
        { /* end of list */ }
    },
};

681
DriveInfo *drive_new(QemuOpts *all_opts, BlockInterfaceType block_default_type)
682
{
K
Kevin Wolf 已提交
683
    const char *value;
684
    BlockBackend *blk;
685
    DriveInfo *dinfo = NULL;
686
    QDict *bs_opts;
687 688
    QemuOpts *legacy_opts;
    DriveMediaType media = MEDIA_DISK;
689
    BlockInterfaceType type;
690
    int cyls, heads, secs, translation;
691
    int max_devs, bus_id, unit_id, index;
692
    const char *devaddr;
693
    const char *werror, *rerror;
F
Fam Zheng 已提交
694 695
    bool read_only = false;
    bool copy_on_read;
696
    const char *serial;
M
Max Reitz 已提交
697
    const char *filename;
698
    Error *local_err = NULL;
699
    int i;
K
Kevin Wolf 已提交
700

701
    /* Change legacy command line options into QMP ones */
702 703 704 705 706 707 708
    static const struct {
        const char *from;
        const char *to;
    } opt_renames[] = {
        { "iops",           "throttling.iops-total" },
        { "iops_rd",        "throttling.iops-read" },
        { "iops_wr",        "throttling.iops-write" },
709

710 711 712
        { "bps",            "throttling.bps-total" },
        { "bps_rd",         "throttling.bps-read" },
        { "bps_wr",         "throttling.bps-write" },
713

714 715 716
        { "iops_max",       "throttling.iops-total-max" },
        { "iops_rd_max",    "throttling.iops-read-max" },
        { "iops_wr_max",    "throttling.iops-write-max" },
717

718 719 720
        { "bps_max",        "throttling.bps-total-max" },
        { "bps_rd_max",     "throttling.bps-read-max" },
        { "bps_wr_max",     "throttling.bps-write-max" },
721

722
        { "iops_size",      "throttling.iops-size" },
723

724 725 726 727
        { "readonly",       "read-only" },
    };

    for (i = 0; i < ARRAY_SIZE(opt_renames); i++) {
728 729 730 731 732 733 734
        qemu_opt_rename(all_opts, opt_renames[i].from, opt_renames[i].to,
                        &local_err);
        if (local_err) {
            error_report("%s", error_get_pretty(local_err));
            error_free(local_err);
            return NULL;
        }
735
    }
736

K
Kevin Wolf 已提交
737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761
    value = qemu_opt_get(all_opts, "cache");
    if (value) {
        int flags = 0;

        if (bdrv_parse_cache_flags(value, &flags) != 0) {
            error_report("invalid cache option");
            return NULL;
        }

        /* Specific options take precedence */
        if (!qemu_opt_get(all_opts, "cache.writeback")) {
            qemu_opt_set_bool(all_opts, "cache.writeback",
                              !!(flags & BDRV_O_CACHE_WB));
        }
        if (!qemu_opt_get(all_opts, "cache.direct")) {
            qemu_opt_set_bool(all_opts, "cache.direct",
                              !!(flags & BDRV_O_NOCACHE));
        }
        if (!qemu_opt_get(all_opts, "cache.no-flush")) {
            qemu_opt_set_bool(all_opts, "cache.no-flush",
                              !!(flags & BDRV_O_NO_FLUSH));
        }
        qemu_opt_unset(all_opts, "cache");
    }

762 763 764 765
    /* Get a QDict for processing the options */
    bs_opts = qdict_new();
    qemu_opts_to_qdict(all_opts, bs_opts);

766 767
    legacy_opts = qemu_opts_create(&qemu_legacy_drive_opts, NULL, 0,
                                   &error_abort);
768
    qemu_opts_absorb_qdict(legacy_opts, bs_opts, &local_err);
769
    if (local_err) {
770
        error_report("%s", error_get_pretty(local_err));
771 772 773 774
        error_free(local_err);
        goto fail;
    }

775 776 777 778 779 780 781
    /* Deprecated option boot=[on|off] */
    if (qemu_opt_get(legacy_opts, "boot") != NULL) {
        fprintf(stderr, "qemu-kvm: boot=on|off is deprecated and will be "
                "ignored. Future versions will reject this parameter. Please "
                "update your scripts.\n");
    }

782 783 784 785 786 787 788
    /* Media type */
    value = qemu_opt_get(legacy_opts, "media");
    if (value) {
        if (!strcmp(value, "disk")) {
            media = MEDIA_DISK;
        } else if (!strcmp(value, "cdrom")) {
            media = MEDIA_CDROM;
F
Fam Zheng 已提交
789
            read_only = true;
790 791 792 793 794 795
        } else {
            error_report("'%s' invalid media", value);
            goto fail;
        }
    }

796
    /* copy-on-read is disabled with a warning for read-only devices */
F
Fam Zheng 已提交
797
    read_only |= qemu_opt_get_bool(legacy_opts, "read-only", false);
798 799 800 801 802 803 804 805 806 807 808 809
    copy_on_read = qemu_opt_get_bool(legacy_opts, "copy-on-read", false);

    if (read_only && copy_on_read) {
        error_report("warning: disabling copy-on-read on read-only drive");
        copy_on_read = false;
    }

    qdict_put(bs_opts, "read-only",
              qstring_from_str(read_only ? "on" : "off"));
    qdict_put(bs_opts, "copy-on-read",
              qstring_from_str(copy_on_read ? "on" :"off"));

810 811 812 813 814 815 816 817 818 819 820 821 822 823 824
    /* Controller type */
    value = qemu_opt_get(legacy_opts, "if");
    if (value) {
        for (type = 0;
             type < IF_COUNT && strcmp(value, if_name[type]);
             type++) {
        }
        if (type == IF_COUNT) {
            error_report("unsupported bus type '%s'", value);
            goto fail;
        }
    } else {
        type = block_default_type;
    }

825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856
    /* Geometry */
    cyls  = qemu_opt_get_number(legacy_opts, "cyls", 0);
    heads = qemu_opt_get_number(legacy_opts, "heads", 0);
    secs  = qemu_opt_get_number(legacy_opts, "secs", 0);

    if (cyls || heads || secs) {
        if (cyls < 1) {
            error_report("invalid physical cyls number");
            goto fail;
        }
        if (heads < 1) {
            error_report("invalid physical heads number");
            goto fail;
        }
        if (secs < 1) {
            error_report("invalid physical secs number");
            goto fail;
        }
    }

    translation = BIOS_ATA_TRANSLATION_AUTO;
    value = qemu_opt_get(legacy_opts, "trans");
    if (value != NULL) {
        if (!cyls) {
            error_report("'%s' trans must be used with cyls, heads and secs",
                         value);
            goto fail;
        }
        if (!strcmp(value, "none")) {
            translation = BIOS_ATA_TRANSLATION_NONE;
        } else if (!strcmp(value, "lba")) {
            translation = BIOS_ATA_TRANSLATION_LBA;
857 858 859 860
        } else if (!strcmp(value, "large")) {
            translation = BIOS_ATA_TRANSLATION_LARGE;
        } else if (!strcmp(value, "rechs")) {
            translation = BIOS_ATA_TRANSLATION_RECHS;
861 862 863 864 865 866 867 868 869 870 871 872 873 874 875
        } else if (!strcmp(value, "auto")) {
            translation = BIOS_ATA_TRANSLATION_AUTO;
        } else {
            error_report("'%s' invalid translation type", value);
            goto fail;
        }
    }

    if (media == MEDIA_CDROM) {
        if (cyls || secs || heads) {
            error_report("CHS can't be set with media=cdrom");
            goto fail;
        }
    }

876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914
    /* Device address specified by bus/unit or index.
     * If none was specified, try to find the first free one. */
    bus_id  = qemu_opt_get_number(legacy_opts, "bus", 0);
    unit_id = qemu_opt_get_number(legacy_opts, "unit", -1);
    index   = qemu_opt_get_number(legacy_opts, "index", -1);

    max_devs = if_max_devs[type];

    if (index != -1) {
        if (bus_id != 0 || unit_id != -1) {
            error_report("index cannot be used with bus and unit");
            goto fail;
        }
        bus_id = drive_index_to_bus_id(type, index);
        unit_id = drive_index_to_unit_id(type, index);
    }

    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++;
           }
       }
    }

    if (max_devs && unit_id >= max_devs) {
        error_report("unit %d too big (max is %d)", unit_id, max_devs - 1);
        goto fail;
    }

    if (drive_get(type, bus_id, unit_id) != NULL) {
        error_report("drive with bus=%d, unit=%d (index=%d) exists",
                     bus_id, unit_id, index);
        goto fail;
    }

915 916 917
    /* Serial number */
    serial = qemu_opt_get(legacy_opts, "serial");

918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935
    /* no id supplied -> create one */
    if (qemu_opts_id(all_opts) == NULL) {
        char *new_id;
        const char *mediastr = "";
        if (type == IF_IDE || type == IF_SCSI) {
            mediastr = (media == MEDIA_CDROM) ? "-cd" : "-hd";
        }
        if (max_devs) {
            new_id = g_strdup_printf("%s%i%s%i", if_name[type], bus_id,
                                     mediastr, unit_id);
        } else {
            new_id = g_strdup_printf("%s%s%i", if_name[type],
                                     mediastr, unit_id);
        }
        qdict_put(bs_opts, "id", qstring_from_str(new_id));
        g_free(new_id);
    }

936 937 938 939 940 941 942 943 944
    /* Add virtio block device */
    devaddr = qemu_opt_get(legacy_opts, "addr");
    if (devaddr && type != IF_VIRTIO) {
        error_report("addr is not supported by this bus type");
        goto fail;
    }

    if (type == IF_VIRTIO) {
        QemuOpts *devopts;
945 946
        devopts = qemu_opts_create(qemu_find_opts("device"), NULL, 0,
                                   &error_abort);
947 948 949 950 951 952 953 954 955 956 957
        if (arch_type == QEMU_ARCH_S390X) {
            qemu_opt_set(devopts, "driver", "virtio-blk-s390");
        } else {
            qemu_opt_set(devopts, "driver", "virtio-blk-pci");
        }
        qemu_opt_set(devopts, "drive", qdict_get_str(bs_opts, "id"));
        if (devaddr) {
            qemu_opt_set(devopts, "addr", devaddr);
        }
    }

M
Max Reitz 已提交
958 959
    filename = qemu_opt_get(legacy_opts, "file");

960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980
    /* Check werror/rerror compatibility with if=... */
    werror = qemu_opt_get(legacy_opts, "werror");
    if (werror != NULL) {
        if (type != IF_IDE && type != IF_SCSI && type != IF_VIRTIO &&
            type != IF_NONE) {
            error_report("werror is not supported by this bus type");
            goto fail;
        }
        qdict_put(bs_opts, "werror", qstring_from_str(werror));
    }

    rerror = qemu_opt_get(legacy_opts, "rerror");
    if (rerror != NULL) {
        if (type != IF_IDE && type != IF_VIRTIO && type != IF_SCSI &&
            type != IF_NONE) {
            error_report("rerror is not supported by this bus type");
            goto fail;
        }
        qdict_put(bs_opts, "rerror", qstring_from_str(rerror));
    }

981
    /* Actual block device init: Functionality shared with blockdev-add */
982
    blk = blockdev_init(filename, bs_opts, &local_err);
983
    bs_opts = NULL;
984
    if (!blk) {
985
        if (local_err) {
986
            error_report("%s", error_get_pretty(local_err));
987 988
            error_free(local_err);
        }
989
        goto fail;
990
    } else {
991
        assert(!local_err);
992 993
    }

994 995
    /* Create legacy DriveInfo */
    dinfo = g_malloc0(sizeof(*dinfo));
996
    dinfo->opts = all_opts;
997

998 999 1000 1001 1002
    dinfo->cyls = cyls;
    dinfo->heads = heads;
    dinfo->secs = secs;
    dinfo->trans = translation;

1003
    dinfo->type = type;
1004 1005
    dinfo->bus = bus_id;
    dinfo->unit = unit_id;
1006
    dinfo->devaddr = devaddr;
1007 1008
    dinfo->serial = g_strdup(serial);

1009 1010
    blk_set_legacy_dinfo(blk, dinfo);

1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021
    switch(type) {
    case IF_IDE:
    case IF_SCSI:
    case IF_XEN:
    case IF_NONE:
        dinfo->media_cd = media == MEDIA_CDROM;
        break;
    default:
        break;
    }

1022
fail:
1023
    qemu_opts_del(legacy_opts);
1024
    QDECREF(bs_opts);
1025
    return dinfo;
1026 1027
}

1028 1029 1030
void do_commit(Monitor *mon, const QDict *qdict)
{
    const char *device = qdict_get_str(qdict, "device");
1031
    BlockDriverState *bs;
1032
    int ret;
1033

1034
    if (!strcmp(device, "all")) {
1035
        ret = bdrv_commit_all();
1036 1037
    } else {
        bs = bdrv_find(device);
1038
        if (!bs) {
1039
            monitor_printf(mon, "Device '%s' not found\n", device);
1040
            return;
1041
        }
1042
        ret = bdrv_commit(bs);
1043 1044 1045 1046
    }
    if (ret < 0) {
        monitor_printf(mon, "'commit' error for '%s': %s\n", device,
                       strerror(-ret));
1047 1048 1049
    }
}

1050 1051
static void blockdev_do_action(int kind, void *data, Error **errp)
{
1052 1053
    TransactionAction action;
    TransactionActionList list;
1054 1055 1056 1057 1058 1059 1060 1061

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

1062 1063 1064 1065 1066
void qmp_blockdev_snapshot_sync(bool has_device, const char *device,
                                bool has_node_name, const char *node_name,
                                const char *snapshot_file,
                                bool has_snapshot_node_name,
                                const char *snapshot_node_name,
1067
                                bool has_format, const char *format,
1068
                                bool has_mode, NewImageMode mode, Error **errp)
1069
{
1070
    BlockdevSnapshot snapshot = {
1071
        .has_device = has_device,
1072
        .device = (char *) device,
1073 1074
        .has_node_name = has_node_name,
        .node_name = (char *) node_name,
1075
        .snapshot_file = (char *) snapshot_file,
1076 1077
        .has_snapshot_node_name = has_snapshot_node_name,
        .snapshot_node_name = (char *) snapshot_node_name,
1078 1079 1080 1081 1082
        .has_format = has_format,
        .format = (char *) format,
        .has_mode = has_mode,
        .mode = mode,
    };
1083 1084
    blockdev_do_action(TRANSACTION_ACTION_KIND_BLOCKDEV_SNAPSHOT_SYNC,
                       &snapshot, errp);
1085 1086
}

1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099
void qmp_blockdev_snapshot_internal_sync(const char *device,
                                         const char *name,
                                         Error **errp)
{
    BlockdevSnapshotInternal snapshot = {
        .device = (char *) device,
        .name = (char *) name
    };

    blockdev_do_action(TRANSACTION_ACTION_KIND_BLOCKDEV_SNAPSHOT_INTERNAL_SYNC,
                       &snapshot, errp);
}

1100 1101 1102 1103 1104 1105 1106 1107
SnapshotInfo *qmp_blockdev_snapshot_delete_internal_sync(const char *device,
                                                         bool has_id,
                                                         const char *id,
                                                         bool has_name,
                                                         const char *name,
                                                         Error **errp)
{
    BlockDriverState *bs = bdrv_find(device);
1108
    AioContext *aio_context;
1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131
    QEMUSnapshotInfo sn;
    Error *local_err = NULL;
    SnapshotInfo *info = NULL;
    int ret;

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

    if (!has_id) {
        id = NULL;
    }

    if (!has_name) {
        name = NULL;
    }

    if (!id && !name) {
        error_setg(errp, "Name or id must be provided");
        return NULL;
    }

1132 1133 1134
    aio_context = bdrv_get_aio_context(bs);
    aio_context_acquire(aio_context);

1135 1136 1137 1138
    if (bdrv_op_is_blocked(bs, BLOCK_OP_TYPE_INTERNAL_SNAPSHOT_DELETE, errp)) {
        goto out_aio_context;
    }

1139
    ret = bdrv_snapshot_find_by_id_and_name(bs, id, name, &sn, &local_err);
1140
    if (local_err) {
1141
        error_propagate(errp, local_err);
1142
        goto out_aio_context;
1143 1144 1145 1146 1147 1148
    }
    if (!ret) {
        error_setg(errp,
                   "Snapshot with id '%s' and name '%s' does not exist on "
                   "device '%s'",
                   STR_OR_NULL(id), STR_OR_NULL(name), device);
1149
        goto out_aio_context;
1150 1151 1152
    }

    bdrv_snapshot_delete(bs, id, name, &local_err);
1153
    if (local_err) {
1154
        error_propagate(errp, local_err);
1155
        goto out_aio_context;
1156 1157
    }

1158 1159
    aio_context_release(aio_context);

1160
    info = g_new0(SnapshotInfo, 1);
1161 1162 1163 1164 1165 1166 1167 1168 1169
    info->id = g_strdup(sn.id_str);
    info->name = g_strdup(sn.name);
    info->date_nsec = sn.date_nsec;
    info->date_sec = sn.date_sec;
    info->vm_state_size = sn.vm_state_size;
    info->vm_clock_nsec = sn.vm_clock_nsec % 1000000000;
    info->vm_clock_sec = sn.vm_clock_nsec / 1000000000;

    return info;
1170 1171 1172 1173

out_aio_context:
    aio_context_release(aio_context);
    return NULL;
1174
}
1175

1176
/* New and old BlockDriverState structs for atomic group operations */
1177

1178
typedef struct BlkTransactionState BlkTransactionState;
1179 1180 1181 1182 1183 1184 1185

/* Only prepare() may fail. In a single transaction, only one of commit() or
   abort() will be called, clean() will always be called if it present. */
typedef struct BdrvActionOps {
    /* Size of state struct, in bytes. */
    size_t instance_size;
    /* Prepare the work, must NOT be NULL. */
1186
    void (*prepare)(BlkTransactionState *common, Error **errp);
1187
    /* Commit the changes, can be NULL. */
1188
    void (*commit)(BlkTransactionState *common);
1189
    /* Abort the changes on fail, can be NULL. */
1190
    void (*abort)(BlkTransactionState *common);
1191
    /* Clean up resource in the end, can be NULL. */
1192
    void (*clean)(BlkTransactionState *common);
1193 1194 1195 1196 1197 1198 1199
} BdrvActionOps;

/*
 * This structure must be arranged as first member in child type, assuming
 * that compiler will also arrange it to the same address with parent instance.
 * Later it will be used in free().
 */
1200
struct BlkTransactionState {
1201
    TransactionAction *action;
1202
    const BdrvActionOps *ops;
1203
    QSIMPLEQ_ENTRY(BlkTransactionState) entry;
1204 1205
};

1206 1207 1208 1209
/* internal snapshot private data */
typedef struct InternalSnapshotState {
    BlkTransactionState common;
    BlockDriverState *bs;
1210
    AioContext *aio_context;
1211 1212 1213 1214 1215 1216
    QEMUSnapshotInfo sn;
} InternalSnapshotState;

static void internal_snapshot_prepare(BlkTransactionState *common,
                                      Error **errp)
{
1217
    Error *local_err = NULL;
1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243
    const char *device;
    const char *name;
    BlockDriverState *bs;
    QEMUSnapshotInfo old_sn, *sn;
    bool ret;
    qemu_timeval tv;
    BlockdevSnapshotInternal *internal;
    InternalSnapshotState *state;
    int ret1;

    g_assert(common->action->kind ==
             TRANSACTION_ACTION_KIND_BLOCKDEV_SNAPSHOT_INTERNAL_SYNC);
    internal = common->action->blockdev_snapshot_internal_sync;
    state = DO_UPCAST(InternalSnapshotState, common, common);

    /* 1. parse input */
    device = internal->device;
    name = internal->name;

    /* 2. check for validation */
    bs = bdrv_find(device);
    if (!bs) {
        error_set(errp, QERR_DEVICE_NOT_FOUND, device);
        return;
    }

1244 1245 1246 1247
    /* AioContext is released in .clean() */
    state->aio_context = bdrv_get_aio_context(bs);
    aio_context_acquire(state->aio_context);

1248 1249 1250 1251 1252
    if (!bdrv_is_inserted(bs)) {
        error_set(errp, QERR_DEVICE_HAS_NO_MEDIUM, device);
        return;
    }

1253 1254 1255 1256
    if (bdrv_op_is_blocked(bs, BLOCK_OP_TYPE_INTERNAL_SNAPSHOT, errp)) {
        return;
    }

1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273
    if (bdrv_is_read_only(bs)) {
        error_set(errp, QERR_DEVICE_IS_READ_ONLY, device);
        return;
    }

    if (!bdrv_can_snapshot(bs)) {
        error_set(errp, QERR_BLOCK_FORMAT_FEATURE_NOT_SUPPORTED,
                  bs->drv->format_name, device, "internal snapshot");
        return;
    }

    if (!strlen(name)) {
        error_setg(errp, "Name is empty");
        return;
    }

    /* check whether a snapshot with name exist */
1274 1275 1276 1277
    ret = bdrv_snapshot_find_by_id_and_name(bs, NULL, name, &old_sn,
                                            &local_err);
    if (local_err) {
        error_propagate(errp, local_err);
1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328
        return;
    } else if (ret) {
        error_setg(errp,
                   "Snapshot with name '%s' already exists on device '%s'",
                   name, device);
        return;
    }

    /* 3. take the snapshot */
    sn = &state->sn;
    pstrcpy(sn->name, sizeof(sn->name), name);
    qemu_gettimeofday(&tv);
    sn->date_sec = tv.tv_sec;
    sn->date_nsec = tv.tv_usec * 1000;
    sn->vm_clock_nsec = qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL);

    ret1 = bdrv_snapshot_create(bs, sn);
    if (ret1 < 0) {
        error_setg_errno(errp, -ret1,
                         "Failed to create snapshot '%s' on device '%s'",
                         name, device);
        return;
    }

    /* 4. succeed, mark a snapshot is created */
    state->bs = bs;
}

static void internal_snapshot_abort(BlkTransactionState *common)
{
    InternalSnapshotState *state =
                             DO_UPCAST(InternalSnapshotState, common, common);
    BlockDriverState *bs = state->bs;
    QEMUSnapshotInfo *sn = &state->sn;
    Error *local_error = NULL;

    if (!bs) {
        return;
    }

    if (bdrv_snapshot_delete(bs, sn->id_str, sn->name, &local_error) < 0) {
        error_report("Failed to delete snapshot with id '%s' and name '%s' on "
                     "device '%s' in abort: %s",
                     sn->id_str,
                     sn->name,
                     bdrv_get_device_name(bs),
                     error_get_pretty(local_error));
        error_free(local_error);
    }
}

1329 1330 1331 1332 1333 1334 1335 1336 1337 1338
static void internal_snapshot_clean(BlkTransactionState *common)
{
    InternalSnapshotState *state = DO_UPCAST(InternalSnapshotState,
                                             common, common);

    if (state->aio_context) {
        aio_context_release(state->aio_context);
    }
}

1339
/* external snapshot private data */
1340 1341
typedef struct ExternalSnapshotState {
    BlkTransactionState common;
1342 1343
    BlockDriverState *old_bs;
    BlockDriverState *new_bs;
1344
    AioContext *aio_context;
1345
} ExternalSnapshotState;
1346

1347
static void external_snapshot_prepare(BlkTransactionState *common,
1348 1349 1350 1351
                                      Error **errp)
{
    BlockDriver *drv;
    int flags, ret;
1352
    QDict *options = NULL;
1353
    Error *local_err = NULL;
1354
    bool has_device = false;
1355
    const char *device;
1356 1357 1358 1359
    bool has_node_name = false;
    const char *node_name;
    bool has_snapshot_node_name = false;
    const char *snapshot_node_name;
1360 1361 1362
    const char *new_image_file;
    const char *format = "qcow2";
    enum NewImageMode mode = NEW_IMAGE_MODE_ABSOLUTE_PATHS;
1363 1364
    ExternalSnapshotState *state =
                             DO_UPCAST(ExternalSnapshotState, common, common);
1365
    TransactionAction *action = common->action;
1366

1367
    /* get parameters */
1368
    g_assert(action->kind == TRANSACTION_ACTION_KIND_BLOCKDEV_SNAPSHOT_SYNC);
1369

1370
    has_device = action->blockdev_snapshot_sync->has_device;
1371
    device = action->blockdev_snapshot_sync->device;
1372 1373 1374 1375 1376 1377
    has_node_name = action->blockdev_snapshot_sync->has_node_name;
    node_name = action->blockdev_snapshot_sync->node_name;
    has_snapshot_node_name =
        action->blockdev_snapshot_sync->has_snapshot_node_name;
    snapshot_node_name = action->blockdev_snapshot_sync->snapshot_node_name;

1378 1379 1380 1381 1382 1383 1384 1385 1386
    new_image_file = action->blockdev_snapshot_sync->snapshot_file;
    if (action->blockdev_snapshot_sync->has_format) {
        format = action->blockdev_snapshot_sync->format;
    }
    if (action->blockdev_snapshot_sync->has_mode) {
        mode = action->blockdev_snapshot_sync->mode;
    }

    /* start processing */
1387 1388 1389 1390 1391 1392
    drv = bdrv_find_format(format);
    if (!drv) {
        error_set(errp, QERR_INVALID_BLOCK_FORMAT, format);
        return;
    }

1393 1394 1395
    state->old_bs = bdrv_lookup_bs(has_device ? device : NULL,
                                   has_node_name ? node_name : NULL,
                                   &local_err);
1396
    if (local_err) {
1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407
        error_propagate(errp, local_err);
        return;
    }

    if (has_node_name && !has_snapshot_node_name) {
        error_setg(errp, "New snapshot node name missing");
        return;
    }

    if (has_snapshot_node_name && bdrv_find_node(snapshot_node_name)) {
        error_setg(errp, "New snapshot node name already existing");
1408 1409 1410
        return;
    }

1411 1412 1413 1414
    /* Acquire AioContext now so any threads operating on old_bs stop */
    state->aio_context = bdrv_get_aio_context(state->old_bs);
    aio_context_acquire(state->aio_context);

1415
    if (!bdrv_is_inserted(state->old_bs)) {
1416 1417 1418 1419
        error_set(errp, QERR_DEVICE_HAS_NO_MEDIUM, device);
        return;
    }

1420 1421
    if (bdrv_op_is_blocked(state->old_bs,
                           BLOCK_OP_TYPE_EXTERNAL_SNAPSHOT, errp)) {
1422 1423 1424
        return;
    }

1425 1426
    if (!bdrv_is_read_only(state->old_bs)) {
        if (bdrv_flush(state->old_bs)) {
1427 1428 1429 1430 1431
            error_set(errp, QERR_IO_ERROR);
            return;
        }
    }

1432
    if (!bdrv_is_first_non_filter(state->old_bs)) {
1433 1434 1435 1436
        error_set(errp, QERR_FEATURE_DISABLED, "snapshot");
        return;
    }

1437
    flags = state->old_bs->open_flags;
1438 1439 1440 1441

    /* create new image w/backing file */
    if (mode != NEW_IMAGE_MODE_EXISTING) {
        bdrv_img_create(new_image_file, format,
1442 1443
                        state->old_bs->filename,
                        state->old_bs->drv->format_name,
1444
                        NULL, -1, flags, &local_err, false);
1445
        if (local_err) {
1446 1447 1448 1449 1450
            error_propagate(errp, local_err);
            return;
        }
    }

1451 1452 1453 1454 1455 1456
    if (has_snapshot_node_name) {
        options = qdict_new();
        qdict_put(options, "node-name",
                  qstring_from_str(snapshot_node_name));
    }

1457 1458
    /* TODO Inherit bs->options or only take explicit options with an
     * extended QMP command? */
1459
    assert(state->new_bs == NULL);
1460
    ret = bdrv_open(&state->new_bs, new_image_file, NULL, options,
1461
                    flags | BDRV_O_NO_BACKING, drv, &local_err);
1462
    /* We will manually add the backing_hd field to the bs later */
1463
    if (ret != 0) {
1464
        error_propagate(errp, local_err);
1465 1466 1467
    }
}

1468
static void external_snapshot_commit(BlkTransactionState *common)
1469
{
1470 1471
    ExternalSnapshotState *state =
                             DO_UPCAST(ExternalSnapshotState, common, common);
1472

1473 1474
    bdrv_set_aio_context(state->new_bs, state->aio_context);

1475 1476
    /* This removes our old bs and adds the new bs */
    bdrv_append(state->new_bs, state->old_bs);
1477 1478 1479
    /* We don't need (or want) to use the transactional
     * bdrv_reopen_multiple() across all the entries at once, because we
     * don't want to abort all of them if one of them fails the reopen */
1480
    bdrv_reopen(state->new_bs, state->new_bs->open_flags & ~BDRV_O_RDWR,
1481
                NULL);
1482 1483

    aio_context_release(state->aio_context);
1484 1485
}

1486
static void external_snapshot_abort(BlkTransactionState *common)
1487
{
1488 1489 1490
    ExternalSnapshotState *state =
                             DO_UPCAST(ExternalSnapshotState, common, common);
    if (state->new_bs) {
F
Fam Zheng 已提交
1491
        bdrv_unref(state->new_bs);
1492
    }
1493 1494 1495
    if (state->aio_context) {
        aio_context_release(state->aio_context);
    }
1496 1497
}

1498 1499 1500
typedef struct DriveBackupState {
    BlkTransactionState common;
    BlockDriverState *bs;
1501
    AioContext *aio_context;
1502 1503 1504 1505 1506 1507
    BlockJob *job;
} DriveBackupState;

static void drive_backup_prepare(BlkTransactionState *common, Error **errp)
{
    DriveBackupState *state = DO_UPCAST(DriveBackupState, common, common);
1508
    BlockDriverState *bs;
1509 1510 1511 1512 1513 1514
    DriveBackup *backup;
    Error *local_err = NULL;

    assert(common->action->kind == TRANSACTION_ACTION_KIND_DRIVE_BACKUP);
    backup = common->action->drive_backup;

1515 1516 1517 1518 1519 1520 1521 1522 1523 1524
    bs = bdrv_find(backup->device);
    if (!bs) {
        error_set(errp, QERR_DEVICE_NOT_FOUND, backup->device);
        return;
    }

    /* AioContext is released in .clean() */
    state->aio_context = bdrv_get_aio_context(bs);
    aio_context_acquire(state->aio_context);

1525 1526
    qmp_drive_backup(backup->device, backup->target,
                     backup->has_format, backup->format,
1527
                     backup->sync,
1528 1529 1530 1531 1532
                     backup->has_mode, backup->mode,
                     backup->has_speed, backup->speed,
                     backup->has_on_source_error, backup->on_source_error,
                     backup->has_on_target_error, backup->on_target_error,
                     &local_err);
1533
    if (local_err) {
1534 1535 1536 1537
        error_propagate(errp, local_err);
        return;
    }

1538
    state->bs = bs;
1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552
    state->job = state->bs->job;
}

static void drive_backup_abort(BlkTransactionState *common)
{
    DriveBackupState *state = DO_UPCAST(DriveBackupState, common, common);
    BlockDriverState *bs = state->bs;

    /* Only cancel if it's the job we started */
    if (bs && bs->job && bs->job == state->job) {
        block_job_cancel_sync(bs->job);
    }
}

1553 1554 1555 1556 1557 1558 1559 1560 1561
static void drive_backup_clean(BlkTransactionState *common)
{
    DriveBackupState *state = DO_UPCAST(DriveBackupState, common, common);

    if (state->aio_context) {
        aio_context_release(state->aio_context);
    }
}

1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634
typedef struct BlockdevBackupState {
    BlkTransactionState common;
    BlockDriverState *bs;
    BlockJob *job;
    AioContext *aio_context;
} BlockdevBackupState;

static void blockdev_backup_prepare(BlkTransactionState *common, Error **errp)
{
    BlockdevBackupState *state = DO_UPCAST(BlockdevBackupState, common, common);
    BlockdevBackup *backup;
    BlockDriverState *bs, *target;
    Error *local_err = NULL;

    assert(common->action->kind == TRANSACTION_ACTION_KIND_BLOCKDEV_BACKUP);
    backup = common->action->blockdev_backup;

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

    target = bdrv_find(backup->target);
    if (!target) {
        error_set(errp, QERR_DEVICE_NOT_FOUND, backup->target);
        return;
    }

    /* AioContext is released in .clean() */
    state->aio_context = bdrv_get_aio_context(bs);
    if (state->aio_context != bdrv_get_aio_context(target)) {
        state->aio_context = NULL;
        error_setg(errp, "Backup between two IO threads is not implemented");
        return;
    }
    aio_context_acquire(state->aio_context);

    qmp_blockdev_backup(backup->device, backup->target,
                        backup->sync,
                        backup->has_speed, backup->speed,
                        backup->has_on_source_error, backup->on_source_error,
                        backup->has_on_target_error, backup->on_target_error,
                        &local_err);
    if (local_err) {
        error_propagate(errp, local_err);
        return;
    }

    state->bs = bs;
    state->job = state->bs->job;
}

static void blockdev_backup_abort(BlkTransactionState *common)
{
    BlockdevBackupState *state = DO_UPCAST(BlockdevBackupState, common, common);
    BlockDriverState *bs = state->bs;

    /* Only cancel if it's the job we started */
    if (bs && bs->job && bs->job == state->job) {
        block_job_cancel_sync(bs->job);
    }
}

static void blockdev_backup_clean(BlkTransactionState *common)
{
    BlockdevBackupState *state = DO_UPCAST(BlockdevBackupState, common, common);

    if (state->aio_context) {
        aio_context_release(state->aio_context);
    }
}

1635 1636 1637 1638 1639 1640 1641
static void abort_prepare(BlkTransactionState *common, Error **errp)
{
    error_setg(errp, "Transaction aborted using Abort action");
}

static void abort_commit(BlkTransactionState *common)
{
1642
    g_assert_not_reached(); /* this action never succeeds */
1643 1644
}

1645
static const BdrvActionOps actions[] = {
1646
    [TRANSACTION_ACTION_KIND_BLOCKDEV_SNAPSHOT_SYNC] = {
1647
        .instance_size = sizeof(ExternalSnapshotState),
1648 1649 1650 1651
        .prepare  = external_snapshot_prepare,
        .commit   = external_snapshot_commit,
        .abort = external_snapshot_abort,
    },
1652 1653 1654 1655
    [TRANSACTION_ACTION_KIND_DRIVE_BACKUP] = {
        .instance_size = sizeof(DriveBackupState),
        .prepare = drive_backup_prepare,
        .abort = drive_backup_abort,
1656
        .clean = drive_backup_clean,
1657
    },
1658 1659 1660 1661 1662 1663
    [TRANSACTION_ACTION_KIND_BLOCKDEV_BACKUP] = {
        .instance_size = sizeof(BlockdevBackupState),
        .prepare = blockdev_backup_prepare,
        .abort = blockdev_backup_abort,
        .clean = blockdev_backup_clean,
    },
1664 1665 1666 1667 1668
    [TRANSACTION_ACTION_KIND_ABORT] = {
        .instance_size = sizeof(BlkTransactionState),
        .prepare = abort_prepare,
        .commit = abort_commit,
    },
1669 1670 1671 1672
    [TRANSACTION_ACTION_KIND_BLOCKDEV_SNAPSHOT_INTERNAL_SYNC] = {
        .instance_size = sizeof(InternalSnapshotState),
        .prepare  = internal_snapshot_prepare,
        .abort = internal_snapshot_abort,
1673
        .clean = internal_snapshot_clean,
1674
    },
1675 1676
};

1677
/*
1678 1679
 * 'Atomic' group operations.  The operations are performed as a set, and if
 * any fail then we roll back all operations in the group.
1680
 */
1681
void qmp_transaction(TransactionActionList *dev_list, Error **errp)
1682
{
1683
    TransactionActionList *dev_entry = dev_list;
1684
    BlkTransactionState *state, *next;
1685
    Error *local_err = NULL;
1686

1687
    QSIMPLEQ_HEAD(snap_bdrv_states, BlkTransactionState) snap_bdrv_states;
1688 1689
    QSIMPLEQ_INIT(&snap_bdrv_states);

1690
    /* drain all i/o before any operations */
1691 1692
    bdrv_drain_all();

1693
    /* We don't do anything in this loop that commits us to the operations */
1694
    while (NULL != dev_entry) {
1695
        TransactionAction *dev_info = NULL;
1696
        const BdrvActionOps *ops;
1697

1698 1699 1700
        dev_info = dev_entry->value;
        dev_entry = dev_entry->next;

1701 1702 1703
        assert(dev_info->kind < ARRAY_SIZE(actions));

        ops = &actions[dev_info->kind];
1704 1705
        assert(ops->instance_size > 0);

1706 1707 1708 1709
        state = g_malloc0(ops->instance_size);
        state->ops = ops;
        state->action = dev_info;
        QSIMPLEQ_INSERT_TAIL(&snap_bdrv_states, state, entry);
1710

1711
        state->ops->prepare(state, &local_err);
1712
        if (local_err) {
1713 1714
            error_propagate(errp, local_err);
            goto delete_and_fail;
1715 1716 1717
        }
    }

1718
    QSIMPLEQ_FOREACH(state, &snap_bdrv_states, entry) {
1719 1720 1721
        if (state->ops->commit) {
            state->ops->commit(state);
        }
1722 1723 1724 1725 1726 1727
    }

    /* success */
    goto exit;

delete_and_fail:
1728
    /* failure, and it is all-or-none; roll back all operations */
1729 1730 1731
    QSIMPLEQ_FOREACH(state, &snap_bdrv_states, entry) {
        if (state->ops->abort) {
            state->ops->abort(state);
1732
        }
1733 1734
    }
exit:
1735 1736 1737
    QSIMPLEQ_FOREACH_SAFE(state, &snap_bdrv_states, entry, next) {
        if (state->ops->clean) {
            state->ops->clean(state);
1738
        }
1739
        g_free(state);
1740 1741 1742 1743
    }
}


1744
static void eject_device(BlockBackend *blk, int force, Error **errp)
1745
{
1746
    BlockDriverState *bs = blk_bs(blk);
1747 1748 1749 1750
    AioContext *aio_context;

    aio_context = bdrv_get_aio_context(bs);
    aio_context_acquire(aio_context);
1751

1752
    if (bdrv_op_is_blocked(bs, BLOCK_OP_TYPE_EJECT, errp)) {
1753
        goto out;
1754
    }
1755
    if (!blk_dev_has_removable_media(blk)) {
1756 1757
        error_setg(errp, "Device '%s' is not removable",
                   bdrv_get_device_name(bs));
1758
        goto out;
1759
    }
1760

1761 1762
    if (blk_dev_is_medium_locked(blk) && !blk_dev_is_tray_open(blk)) {
        blk_dev_eject_request(blk, force);
P
Paolo Bonzini 已提交
1763
        if (!force) {
1764 1765
            error_setg(errp, "Device '%s' is locked",
                       bdrv_get_device_name(bs));
1766
            goto out;
P
Paolo Bonzini 已提交
1767
        }
1768
    }
1769

1770
    bdrv_close(bs);
1771 1772 1773

out:
    aio_context_release(aio_context);
1774 1775
}

L
Luiz Capitulino 已提交
1776
void qmp_eject(const char *device, bool has_force, bool force, Error **errp)
1777
{
1778
    BlockBackend *blk;
1779

1780 1781
    blk = blk_by_name(device);
    if (!blk) {
L
Luiz Capitulino 已提交
1782 1783
        error_set(errp, QERR_DEVICE_NOT_FOUND, device);
        return;
1784 1785
    }

1786
    eject_device(blk, force, errp);
1787 1788
}

1789 1790 1791
void qmp_block_passwd(bool has_device, const char *device,
                      bool has_node_name, const char *node_name,
                      const char *password, Error **errp)
1792
{
1793
    Error *local_err = NULL;
1794
    BlockDriverState *bs;
1795
    AioContext *aio_context;
1796

1797 1798 1799
    bs = bdrv_lookup_bs(has_device ? device : NULL,
                        has_node_name ? node_name : NULL,
                        &local_err);
1800
    if (local_err) {
1801
        error_propagate(errp, local_err);
L
Luiz Capitulino 已提交
1802
        return;
1803 1804
    }

1805 1806 1807
    aio_context = bdrv_get_aio_context(bs);
    aio_context_acquire(aio_context);

1808
    bdrv_add_key(bs, password, errp);
1809 1810

    aio_context_release(aio_context);
1811 1812
}

1813
/* Assumes AioContext is held */
L
Luiz Capitulino 已提交
1814 1815 1816 1817
static void qmp_bdrv_open_encrypted(BlockDriverState *bs, const char *filename,
                                    int bdrv_flags, BlockDriver *drv,
                                    const char *password, Error **errp)
{
1818
    Error *local_err = NULL;
1819 1820
    int ret;

1821
    ret = bdrv_open(&bs, filename, NULL, NULL, bdrv_flags, drv, &local_err);
1822
    if (ret < 0) {
1823
        error_propagate(errp, local_err);
L
Luiz Capitulino 已提交
1824 1825 1826
        return;
    }

1827
    bdrv_add_key(bs, password, errp);
L
Luiz Capitulino 已提交
1828 1829 1830
}

void qmp_change_blockdev(const char *device, const char *filename,
1831
                         const char *format, Error **errp)
1832
{
1833
    BlockBackend *blk;
1834
    BlockDriverState *bs;
1835
    AioContext *aio_context;
1836 1837
    BlockDriver *drv = NULL;
    int bdrv_flags;
1838
    Error *err = NULL;
1839

1840 1841
    blk = blk_by_name(device);
    if (!blk) {
L
Luiz Capitulino 已提交
1842 1843
        error_set(errp, QERR_DEVICE_NOT_FOUND, device);
        return;
1844
    }
1845
    bs = blk_bs(blk);
L
Luiz Capitulino 已提交
1846

1847 1848 1849
    aio_context = bdrv_get_aio_context(bs);
    aio_context_acquire(aio_context);

L
Luiz Capitulino 已提交
1850
    if (format) {
1851
        drv = bdrv_find_whitelisted_format(format, bs->read_only);
1852
        if (!drv) {
L
Luiz Capitulino 已提交
1853
            error_set(errp, QERR_INVALID_BLOCK_FORMAT, format);
1854
            goto out;
1855 1856
        }
    }
L
Luiz Capitulino 已提交
1857

1858
    eject_device(blk, 0, &err);
1859
    if (err) {
L
Luiz Capitulino 已提交
1860
        error_propagate(errp, err);
1861
        goto out;
1862
    }
L
Luiz Capitulino 已提交
1863

1864
    bdrv_flags = bdrv_is_read_only(bs) ? 0 : BDRV_O_RDWR;
1865
    bdrv_flags |= bdrv_is_snapshot(bs) ? BDRV_O_SNAPSHOT : 0;
L
Luiz Capitulino 已提交
1866 1867

    qmp_bdrv_open_encrypted(bs, filename, bdrv_flags, drv, NULL, errp);
1868 1869 1870

out:
    aio_context_release(aio_context);
1871
}
1872

Z
Zhi Yong Wu 已提交
1873
/* throttling disk I/O limits */
1874
void qmp_block_set_io_throttle(const char *device, int64_t bps, int64_t bps_rd,
1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889
                               int64_t bps_wr,
                               int64_t iops,
                               int64_t iops_rd,
                               int64_t iops_wr,
                               bool has_bps_max,
                               int64_t bps_max,
                               bool has_bps_rd_max,
                               int64_t bps_rd_max,
                               bool has_bps_wr_max,
                               int64_t bps_wr_max,
                               bool has_iops_max,
                               int64_t iops_max,
                               bool has_iops_rd_max,
                               int64_t iops_rd_max,
                               bool has_iops_wr_max,
1890 1891 1892
                               int64_t iops_wr_max,
                               bool has_iops_size,
                               int64_t iops_size, Error **errp)
Z
Zhi Yong Wu 已提交
1893
{
1894
    ThrottleConfig cfg;
Z
Zhi Yong Wu 已提交
1895
    BlockDriverState *bs;
1896
    AioContext *aio_context;
Z
Zhi Yong Wu 已提交
1897

1898
    bs = bdrv_find(device);
Z
Zhi Yong Wu 已提交
1899
    if (!bs) {
1900 1901
        error_set(errp, QERR_DEVICE_NOT_FOUND, device);
        return;
Z
Zhi Yong Wu 已提交
1902 1903
    }

1904 1905 1906 1907 1908 1909 1910 1911 1912
    memset(&cfg, 0, sizeof(cfg));
    cfg.buckets[THROTTLE_BPS_TOTAL].avg = bps;
    cfg.buckets[THROTTLE_BPS_READ].avg  = bps_rd;
    cfg.buckets[THROTTLE_BPS_WRITE].avg = bps_wr;

    cfg.buckets[THROTTLE_OPS_TOTAL].avg = iops;
    cfg.buckets[THROTTLE_OPS_READ].avg  = iops_rd;
    cfg.buckets[THROTTLE_OPS_WRITE].avg = iops_wr;

1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930
    if (has_bps_max) {
        cfg.buckets[THROTTLE_BPS_TOTAL].max = bps_max;
    }
    if (has_bps_rd_max) {
        cfg.buckets[THROTTLE_BPS_READ].max = bps_rd_max;
    }
    if (has_bps_wr_max) {
        cfg.buckets[THROTTLE_BPS_WRITE].max = bps_wr_max;
    }
    if (has_iops_max) {
        cfg.buckets[THROTTLE_OPS_TOTAL].max = iops_max;
    }
    if (has_iops_rd_max) {
        cfg.buckets[THROTTLE_OPS_READ].max = iops_rd_max;
    }
    if (has_iops_wr_max) {
        cfg.buckets[THROTTLE_OPS_WRITE].max = iops_wr_max;
    }
Z
Zhi Yong Wu 已提交
1931

1932 1933 1934
    if (has_iops_size) {
        cfg.op_size = iops_size;
    }
1935 1936

    if (!check_throttle_config(&cfg, errp)) {
1937
        return;
Z
Zhi Yong Wu 已提交
1938 1939
    }

1940 1941 1942
    aio_context = bdrv_get_aio_context(bs);
    aio_context_acquire(aio_context);

1943
    if (!bs->io_limits_enabled && throttle_enabled(&cfg)) {
Z
Zhi Yong Wu 已提交
1944
        bdrv_io_limits_enable(bs);
1945
    } else if (bs->io_limits_enabled && !throttle_enabled(&cfg)) {
Z
Zhi Yong Wu 已提交
1946
        bdrv_io_limits_disable(bs);
1947 1948 1949 1950
    }

    if (bs->io_limits_enabled) {
        bdrv_set_io_limits(bs, &cfg);
Z
Zhi Yong Wu 已提交
1951
    }
1952 1953

    aio_context_release(aio_context);
Z
Zhi Yong Wu 已提交
1954 1955
}

1956 1957 1958
int do_drive_del(Monitor *mon, const QDict *qdict, QObject **ret_data)
{
    const char *id = qdict_get_str(qdict, "id");
1959
    BlockBackend *blk;
1960
    BlockDriverState *bs;
1961
    AioContext *aio_context;
1962
    Error *local_err = NULL;
1963

1964 1965
    blk = blk_by_name(id);
    if (!blk) {
1966
        error_report("Device '%s' not found", id);
1967 1968
        return -1;
    }
1969
    bs = blk_bs(blk);
1970

1971
    if (!blk_legacy_dinfo(blk)) {
1972 1973 1974 1975 1976
        error_report("Deleting device added with blockdev-add"
                     " is not supported");
        return -1;
    }

1977 1978 1979
    aio_context = bdrv_get_aio_context(bs);
    aio_context_acquire(aio_context);

1980 1981 1982
    if (bdrv_op_is_blocked(bs, BLOCK_OP_TYPE_DRIVE_DEL, &local_err)) {
        error_report("%s", error_get_pretty(local_err));
        error_free(local_err);
1983
        aio_context_release(aio_context);
M
Marcelo Tosatti 已提交
1984 1985
        return -1;
    }
1986 1987

    /* quiesce block driver; prevent further io */
1988
    bdrv_drain_all();
1989 1990 1991
    bdrv_flush(bs);
    bdrv_close(bs);

1992
    /* if we have a device attached to this BlockDriverState
1993 1994 1995 1996
     * 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.
     */
1997
    if (blk_get_attached_dev(blk)) {
1998
        blk_hide_on_behalf_of_do_drive_del(blk);
1999 2000 2001
        /* Further I/O must not pause the guest */
        bdrv_set_on_error(bs, BLOCKDEV_ON_ERROR_REPORT,
                          BLOCKDEV_ON_ERROR_REPORT);
2002
    } else {
2003
        blk_unref(blk);
2004 2005
    }

2006
    aio_context_release(aio_context);
2007 2008
    return 0;
}
2009

2010 2011 2012
void qmp_block_resize(bool has_device, const char *device,
                      bool has_node_name, const char *node_name,
                      int64_t size, Error **errp)
2013
{
2014
    Error *local_err = NULL;
2015
    BlockDriverState *bs;
2016
    AioContext *aio_context;
2017
    int ret;
2018

2019 2020 2021
    bs = bdrv_lookup_bs(has_device ? device : NULL,
                        has_node_name ? node_name : NULL,
                        &local_err);
2022
    if (local_err) {
2023 2024 2025 2026
        error_propagate(errp, local_err);
        return;
    }

2027 2028 2029
    aio_context = bdrv_get_aio_context(bs);
    aio_context_acquire(aio_context);

2030 2031
    if (!bdrv_is_first_non_filter(bs)) {
        error_set(errp, QERR_FEATURE_DISABLED, "resize");
2032
        goto out;
2033 2034 2035
    }

    if (size < 0) {
2036
        error_set(errp, QERR_INVALID_PARAMETER_VALUE, "size", "a >0 size");
2037
        goto out;
2038 2039
    }

2040 2041
    if (bdrv_op_is_blocked(bs, BLOCK_OP_TYPE_RESIZE, NULL)) {
        error_set(errp, QERR_DEVICE_IN_USE, device);
2042
        goto out;
2043 2044
    }

2045 2046 2047
    /* complete all in-flight operations before resizing the device */
    bdrv_drain_all();

2048 2049
    ret = bdrv_truncate(bs, size);
    switch (ret) {
2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064
    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:
2065
        error_setg_errno(errp, -ret, "Could not resize");
2066
        break;
2067
    }
2068 2069 2070

out:
    aio_context_release(aio_context);
2071
}
S
Stefan Hajnoczi 已提交
2072

2073
static void block_job_cb(void *opaque, int ret)
S
Stefan Hajnoczi 已提交
2074
{
2075 2076 2077 2078 2079
    /* Note that this function may be executed from another AioContext besides
     * the QEMU main loop.  If you need to access anything that assumes the
     * QEMU global mutex, use a BH or introduce a mutex.
     */

S
Stefan Hajnoczi 已提交
2080
    BlockDriverState *bs = opaque;
2081
    const char *msg = NULL;
S
Stefan Hajnoczi 已提交
2082

2083
    trace_block_job_cb(bs, bs->job, ret);
S
Stefan Hajnoczi 已提交
2084 2085

    assert(bs->job);
2086

S
Stefan Hajnoczi 已提交
2087
    if (ret < 0) {
2088
        msg = strerror(-ret);
S
Stefan Hajnoczi 已提交
2089 2090
    }

2091
    if (block_job_is_cancelled(bs->job)) {
2092
        block_job_event_cancelled(bs->job);
2093
    } else {
2094
        block_job_event_completed(bs->job, msg);
2095
    }
2096

F
Fam Zheng 已提交
2097
    bdrv_put_ref_bh_schedule(bs);
S
Stefan Hajnoczi 已提交
2098 2099
}

2100 2101 2102 2103
void qmp_block_stream(const char *device,
                      bool has_base, const char *base,
                      bool has_backing_file, const char *backing_file,
                      bool has_speed, int64_t speed,
P
Paolo Bonzini 已提交
2104 2105
                      bool has_on_error, BlockdevOnError on_error,
                      Error **errp)
S
Stefan Hajnoczi 已提交
2106 2107
{
    BlockDriverState *bs;
2108
    BlockDriverState *base_bs = NULL;
2109
    AioContext *aio_context;
2110
    Error *local_err = NULL;
2111
    const char *base_name = NULL;
S
Stefan Hajnoczi 已提交
2112

P
Paolo Bonzini 已提交
2113 2114 2115 2116
    if (!has_on_error) {
        on_error = BLOCKDEV_ON_ERROR_REPORT;
    }

S
Stefan Hajnoczi 已提交
2117 2118 2119 2120 2121 2122
    bs = bdrv_find(device);
    if (!bs) {
        error_set(errp, QERR_DEVICE_NOT_FOUND, device);
        return;
    }

2123 2124 2125
    aio_context = bdrv_get_aio_context(bs);
    aio_context_acquire(aio_context);

2126
    if (bdrv_op_is_blocked(bs, BLOCK_OP_TYPE_STREAM, errp)) {
2127
        goto out;
2128 2129
    }

2130
    if (has_base) {
2131 2132 2133
        base_bs = bdrv_find_backing_image(bs, base);
        if (base_bs == NULL) {
            error_set(errp, QERR_BASE_NOT_FOUND, base);
2134
            goto out;
2135
        }
2136
        assert(bdrv_get_aio_context(base_bs) == aio_context);
2137
        base_name = base;
S
Stefan Hajnoczi 已提交
2138 2139
    }

2140 2141 2142 2143 2144
    /* if we are streaming the entire chain, the result will have no backing
     * file, and specifying one is therefore an error */
    if (base_bs == NULL && has_backing_file) {
        error_setg(errp, "backing file specified, but streaming the "
                         "entire chain");
2145
        goto out;
2146 2147 2148 2149 2150 2151
    }

    /* backing_file string overrides base bs filename */
    base_name = has_backing_file ? backing_file : base_name;

    stream_start(bs, base_bs, base_name, has_speed ? speed : 0,
P
Paolo Bonzini 已提交
2152
                 on_error, block_job_cb, bs, &local_err);
2153
    if (local_err) {
2154
        error_propagate(errp, local_err);
2155
        goto out;
S
Stefan Hajnoczi 已提交
2156 2157 2158
    }

    trace_qmp_block_stream(bs, bs->job);
2159 2160 2161

out:
    aio_context_release(aio_context);
S
Stefan Hajnoczi 已提交
2162
}
2163

2164
void qmp_block_commit(const char *device,
2165 2166
                      bool has_base, const char *base,
                      bool has_top, const char *top,
2167
                      bool has_backing_file, const char *backing_file,
2168 2169 2170 2171 2172
                      bool has_speed, int64_t speed,
                      Error **errp)
{
    BlockDriverState *bs;
    BlockDriverState *base_bs, *top_bs;
2173
    AioContext *aio_context;
2174 2175 2176 2177
    Error *local_err = NULL;
    /* This will be part of the QMP command, if/when the
     * BlockdevOnError change for blkmirror makes it in
     */
2178
    BlockdevOnError on_error = BLOCKDEV_ON_ERROR_REPORT;
2179

2180 2181 2182 2183
    if (!has_speed) {
        speed = 0;
    }

2184 2185 2186 2187 2188
    /* Important Note:
     *  libvirt relies on the DeviceNotFound error class in order to probe for
     *  live commit feature versions; for this to work, we must make sure to
     *  perform the device lookup before any generic errors that may occur in a
     *  scenario in which all optional arguments are omitted. */
2189 2190 2191 2192
    bs = bdrv_find(device);
    if (!bs) {
        error_set(errp, QERR_DEVICE_NOT_FOUND, device);
        return;
2193 2194
    }

2195 2196 2197 2198 2199 2200
    aio_context = bdrv_get_aio_context(bs);
    aio_context_acquire(aio_context);

    /* drain all i/o before commits */
    bdrv_drain_all();

2201
    if (bdrv_op_is_blocked(bs, BLOCK_OP_TYPE_COMMIT_SOURCE, errp)) {
2202
        goto out;
2203 2204 2205 2206 2207
    }

    /* default top_bs is the active layer */
    top_bs = bs;

2208
    if (has_top && top) {
2209 2210 2211 2212 2213 2214 2215
        if (strcmp(bs->filename, top) != 0) {
            top_bs = bdrv_find_backing_image(bs, top);
        }
    }

    if (top_bs == NULL) {
        error_setg(errp, "Top image file %s not found", top ? top : "NULL");
2216
        goto out;
2217 2218
    }

2219 2220
    assert(bdrv_get_aio_context(top_bs) == aio_context);

2221 2222 2223 2224 2225 2226 2227 2228
    if (has_base && base) {
        base_bs = bdrv_find_backing_image(top_bs, base);
    } else {
        base_bs = bdrv_find_base(top_bs);
    }

    if (base_bs == NULL) {
        error_set(errp, QERR_BASE_NOT_FOUND, base ? base : "NULL");
2229
        goto out;
2230 2231
    }

2232 2233
    assert(bdrv_get_aio_context(base_bs) == aio_context);

2234 2235 2236 2237
    if (bdrv_op_is_blocked(base_bs, BLOCK_OP_TYPE_COMMIT_TARGET, errp)) {
        goto out;
    }

2238 2239 2240
    /* Do not allow attempts to commit an image into itself */
    if (top_bs == base_bs) {
        error_setg(errp, "cannot commit an image into itself");
2241
        goto out;
2242 2243
    }

F
Fam Zheng 已提交
2244
    if (top_bs == bs) {
2245 2246 2247
        if (has_backing_file) {
            error_setg(errp, "'backing-file' specified,"
                             " but 'top' is the active layer");
2248
            goto out;
2249
        }
F
Fam Zheng 已提交
2250 2251 2252 2253
        commit_active_start(bs, base_bs, speed, on_error, block_job_cb,
                            bs, &local_err);
    } else {
        commit_start(bs, base_bs, top_bs, speed, on_error, block_job_cb, bs,
2254
                     has_backing_file ? backing_file : NULL, &local_err);
F
Fam Zheng 已提交
2255
    }
2256 2257
    if (local_err != NULL) {
        error_propagate(errp, local_err);
2258
        goto out;
2259
    }
2260 2261 2262

out:
    aio_context_release(aio_context);
2263 2264
}

2265 2266
void qmp_drive_backup(const char *device, const char *target,
                      bool has_format, const char *format,
2267
                      enum MirrorSyncMode sync,
2268 2269 2270 2271 2272 2273 2274 2275
                      bool has_mode, enum NewImageMode mode,
                      bool has_speed, int64_t speed,
                      bool has_on_source_error, BlockdevOnError on_source_error,
                      bool has_on_target_error, BlockdevOnError on_target_error,
                      Error **errp)
{
    BlockDriverState *bs;
    BlockDriverState *target_bs;
2276
    BlockDriverState *source = NULL;
2277
    AioContext *aio_context;
2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302
    BlockDriver *drv = NULL;
    Error *local_err = NULL;
    int flags;
    int64_t size;
    int ret;

    if (!has_speed) {
        speed = 0;
    }
    if (!has_on_source_error) {
        on_source_error = BLOCKDEV_ON_ERROR_REPORT;
    }
    if (!has_on_target_error) {
        on_target_error = BLOCKDEV_ON_ERROR_REPORT;
    }
    if (!has_mode) {
        mode = NEW_IMAGE_MODE_ABSOLUTE_PATHS;
    }

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

2303 2304 2305
    aio_context = bdrv_get_aio_context(bs);
    aio_context_acquire(aio_context);

F
Fam Zheng 已提交
2306 2307
    /* Although backup_run has this check too, we need to use bs->drv below, so
     * do an early check redundantly. */
2308 2309
    if (!bdrv_is_inserted(bs)) {
        error_set(errp, QERR_DEVICE_HAS_NO_MEDIUM, device);
2310
        goto out;
2311 2312 2313 2314 2315 2316 2317 2318 2319
    }

    if (!has_format) {
        format = mode == NEW_IMAGE_MODE_EXISTING ? NULL : bs->drv->format_name;
    }
    if (format) {
        drv = bdrv_find_format(format);
        if (!drv) {
            error_set(errp, QERR_INVALID_BLOCK_FORMAT, format);
2320
            goto out;
2321 2322 2323
        }
    }

F
Fam Zheng 已提交
2324
    /* Early check to avoid creating target */
2325
    if (bdrv_op_is_blocked(bs, BLOCK_OP_TYPE_BACKUP_SOURCE, errp)) {
2326
        goto out;
2327 2328 2329 2330
    }

    flags = bs->open_flags | BDRV_O_RDWR;

2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342
    /* See if we have a backing HD we can use to create our new image
     * on top of. */
    if (sync == MIRROR_SYNC_MODE_TOP) {
        source = bs->backing_hd;
        if (!source) {
            sync = MIRROR_SYNC_MODE_FULL;
        }
    }
    if (sync == MIRROR_SYNC_MODE_NONE) {
        source = bs;
    }

2343 2344 2345
    size = bdrv_getlength(bs);
    if (size < 0) {
        error_setg_errno(errp, -size, "bdrv_getlength failed");
2346
        goto out;
2347 2348 2349 2350
    }

    if (mode != NEW_IMAGE_MODE_EXISTING) {
        assert(format && drv);
2351 2352 2353 2354 2355 2356 2357 2358
        if (source) {
            bdrv_img_create(target, format, source->filename,
                            source->drv->format_name, NULL,
                            size, flags, &local_err, false);
        } else {
            bdrv_img_create(target, format, NULL, NULL, NULL,
                            size, flags, &local_err, false);
        }
2359 2360
    }

2361
    if (local_err) {
2362
        error_propagate(errp, local_err);
2363
        goto out;
2364 2365
    }

2366
    target_bs = NULL;
2367
    ret = bdrv_open(&target_bs, target, NULL, NULL, flags, drv, &local_err);
2368
    if (ret < 0) {
2369
        error_propagate(errp, local_err);
2370
        goto out;
2371 2372
    }

2373 2374
    bdrv_set_aio_context(target_bs, aio_context);

2375
    backup_start(bs, target_bs, speed, sync, on_source_error, on_target_error,
2376 2377
                 block_job_cb, bs, &local_err);
    if (local_err != NULL) {
F
Fam Zheng 已提交
2378
        bdrv_unref(target_bs);
2379
        error_propagate(errp, local_err);
2380
        goto out;
2381
    }
2382 2383 2384

out:
    aio_context_release(aio_context);
2385 2386
}

2387 2388 2389 2390 2391
BlockDeviceInfoList *qmp_query_named_block_nodes(Error **errp)
{
    return bdrv_named_nodes_list();
}

F
Fam Zheng 已提交
2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442
void qmp_blockdev_backup(const char *device, const char *target,
                         enum MirrorSyncMode sync,
                         bool has_speed, int64_t speed,
                         bool has_on_source_error,
                         BlockdevOnError on_source_error,
                         bool has_on_target_error,
                         BlockdevOnError on_target_error,
                         Error **errp)
{
    BlockDriverState *bs;
    BlockDriverState *target_bs;
    Error *local_err = NULL;
    AioContext *aio_context;

    if (!has_speed) {
        speed = 0;
    }
    if (!has_on_source_error) {
        on_source_error = BLOCKDEV_ON_ERROR_REPORT;
    }
    if (!has_on_target_error) {
        on_target_error = BLOCKDEV_ON_ERROR_REPORT;
    }

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

    aio_context = bdrv_get_aio_context(bs);
    aio_context_acquire(aio_context);

    target_bs = bdrv_find(target);
    if (!target_bs) {
        error_set(errp, QERR_DEVICE_NOT_FOUND, target);
        goto out;
    }

    bdrv_ref(target_bs);
    bdrv_set_aio_context(target_bs, aio_context);
    backup_start(bs, target_bs, speed, sync, on_source_error, on_target_error,
                 block_job_cb, bs, &local_err);
    if (local_err != NULL) {
        bdrv_unref(target_bs);
        error_propagate(errp, local_err);
    }
out:
    aio_context_release(aio_context);
}

2443 2444
#define DEFAULT_MIRROR_BUF_SIZE   (10 << 20)

P
Paolo Bonzini 已提交
2445 2446
void qmp_drive_mirror(const char *device, const char *target,
                      bool has_format, const char *format,
2447
                      bool has_node_name, const char *node_name,
2448
                      bool has_replaces, const char *replaces,
P
Paolo Bonzini 已提交
2449 2450
                      enum MirrorSyncMode sync,
                      bool has_mode, enum NewImageMode mode,
2451
                      bool has_speed, int64_t speed,
2452
                      bool has_granularity, uint32_t granularity,
2453
                      bool has_buf_size, int64_t buf_size,
2454 2455 2456
                      bool has_on_source_error, BlockdevOnError on_source_error,
                      bool has_on_target_error, BlockdevOnError on_target_error,
                      Error **errp)
P
Paolo Bonzini 已提交
2457 2458 2459
{
    BlockDriverState *bs;
    BlockDriverState *source, *target_bs;
2460
    AioContext *aio_context;
P
Paolo Bonzini 已提交
2461 2462
    BlockDriver *drv = NULL;
    Error *local_err = NULL;
2463
    QDict *options = NULL;
P
Paolo Bonzini 已提交
2464
    int flags;
2465
    int64_t size;
P
Paolo Bonzini 已提交
2466 2467 2468 2469 2470
    int ret;

    if (!has_speed) {
        speed = 0;
    }
2471 2472 2473 2474 2475 2476
    if (!has_on_source_error) {
        on_source_error = BLOCKDEV_ON_ERROR_REPORT;
    }
    if (!has_on_target_error) {
        on_target_error = BLOCKDEV_ON_ERROR_REPORT;
    }
P
Paolo Bonzini 已提交
2477 2478 2479
    if (!has_mode) {
        mode = NEW_IMAGE_MODE_ABSOLUTE_PATHS;
    }
2480 2481 2482
    if (!has_granularity) {
        granularity = 0;
    }
2483 2484 2485 2486
    if (!has_buf_size) {
        buf_size = DEFAULT_MIRROR_BUF_SIZE;
    }

2487
    if (granularity != 0 && (granularity < 512 || granularity > 1048576 * 64)) {
2488 2489
        error_set(errp, QERR_INVALID_PARAMETER_VALUE, "granularity",
                  "a value in range [512B, 64MB]");
2490 2491 2492
        return;
    }
    if (granularity & (granularity - 1)) {
2493
        error_set(errp, QERR_INVALID_PARAMETER_VALUE, "granularity", "power of 2");
2494 2495
        return;
    }
P
Paolo Bonzini 已提交
2496 2497 2498 2499 2500 2501 2502

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

2503 2504 2505
    aio_context = bdrv_get_aio_context(bs);
    aio_context_acquire(aio_context);

P
Paolo Bonzini 已提交
2506 2507
    if (!bdrv_is_inserted(bs)) {
        error_set(errp, QERR_DEVICE_HAS_NO_MEDIUM, device);
2508
        goto out;
P
Paolo Bonzini 已提交
2509 2510 2511 2512 2513 2514 2515 2516 2517
    }

    if (!has_format) {
        format = mode == NEW_IMAGE_MODE_EXISTING ? NULL : bs->drv->format_name;
    }
    if (format) {
        drv = bdrv_find_format(format);
        if (!drv) {
            error_set(errp, QERR_INVALID_BLOCK_FORMAT, format);
2518
            goto out;
P
Paolo Bonzini 已提交
2519 2520 2521
        }
    }

2522
    if (bdrv_op_is_blocked(bs, BLOCK_OP_TYPE_MIRROR, errp)) {
2523
        goto out;
P
Paolo Bonzini 已提交
2524 2525 2526 2527 2528 2529 2530
    }

    flags = bs->open_flags | BDRV_O_RDWR;
    source = bs->backing_hd;
    if (!source && sync == MIRROR_SYNC_MODE_TOP) {
        sync = MIRROR_SYNC_MODE_FULL;
    }
2531 2532 2533
    if (sync == MIRROR_SYNC_MODE_NONE) {
        source = bs;
    }
P
Paolo Bonzini 已提交
2534

2535 2536 2537
    size = bdrv_getlength(bs);
    if (size < 0) {
        error_setg_errno(errp, -size, "bdrv_getlength failed");
2538
        goto out;
2539 2540
    }

2541 2542
    if (has_replaces) {
        BlockDriverState *to_replace_bs;
2543 2544
        AioContext *replace_aio_context;
        int64_t replace_size;
2545 2546 2547 2548

        if (!has_node_name) {
            error_setg(errp, "a node-name must be provided when replacing a"
                             " named node of the graph");
2549
            goto out;
2550 2551 2552 2553 2554 2555
        }

        to_replace_bs = check_to_replace_node(replaces, &local_err);

        if (!to_replace_bs) {
            error_propagate(errp, local_err);
2556
            goto out;
2557 2558
        }

2559 2560 2561 2562 2563 2564
        replace_aio_context = bdrv_get_aio_context(to_replace_bs);
        aio_context_acquire(replace_aio_context);
        replace_size = bdrv_getlength(to_replace_bs);
        aio_context_release(replace_aio_context);

        if (size != replace_size) {
2565 2566
            error_setg(errp, "cannot replace image with a mirror image of "
                             "different size");
2567
            goto out;
2568 2569 2570
        }
    }

2571 2572 2573
    if ((sync == MIRROR_SYNC_MODE_FULL || !source)
        && mode != NEW_IMAGE_MODE_EXISTING)
    {
P
Paolo Bonzini 已提交
2574 2575
        /* create new image w/o backing file */
        assert(format && drv);
2576
        bdrv_img_create(target, format,
2577
                        NULL, NULL, NULL, size, flags, &local_err, false);
P
Paolo Bonzini 已提交
2578 2579 2580 2581 2582 2583
    } else {
        switch (mode) {
        case NEW_IMAGE_MODE_EXISTING:
            break;
        case NEW_IMAGE_MODE_ABSOLUTE_PATHS:
            /* create new image with backing file */
2584 2585 2586
            bdrv_img_create(target, format,
                            source->filename,
                            source->drv->format_name,
2587
                            NULL, size, flags, &local_err, false);
P
Paolo Bonzini 已提交
2588 2589 2590 2591 2592 2593
            break;
        default:
            abort();
        }
    }

2594
    if (local_err) {
2595
        error_propagate(errp, local_err);
2596
        goto out;
P
Paolo Bonzini 已提交
2597 2598
    }

2599 2600 2601 2602 2603
    if (has_node_name) {
        options = qdict_new();
        qdict_put(options, "node-name", qstring_from_str(node_name));
    }

2604 2605 2606
    /* Mirroring takes care of copy-on-write using the source's backing
     * file.
     */
2607
    target_bs = NULL;
2608 2609
    ret = bdrv_open(&target_bs, target, NULL, options,
                    flags | BDRV_O_NO_BACKING, drv, &local_err);
P
Paolo Bonzini 已提交
2610
    if (ret < 0) {
2611
        error_propagate(errp, local_err);
2612
        goto out;
P
Paolo Bonzini 已提交
2613 2614
    }

2615 2616
    bdrv_set_aio_context(target_bs, aio_context);

2617 2618 2619 2620 2621 2622
    /* pass the node name to replace to mirror start since it's loose coupling
     * and will allow to check whether the node still exist at mirror completion
     */
    mirror_start(bs, target_bs,
                 has_replaces ? replaces : NULL,
                 speed, granularity, buf_size, sync,
2623
                 on_source_error, on_target_error,
2624
                 block_job_cb, bs, &local_err);
P
Paolo Bonzini 已提交
2625
    if (local_err != NULL) {
F
Fam Zheng 已提交
2626
        bdrv_unref(target_bs);
P
Paolo Bonzini 已提交
2627
        error_propagate(errp, local_err);
2628
        goto out;
P
Paolo Bonzini 已提交
2629
    }
2630 2631 2632

out:
    aio_context_release(aio_context);
P
Paolo Bonzini 已提交
2633 2634
}

2635
/* Get the block job for a given device name and acquire its AioContext */
2636 2637
static BlockJob *find_block_job(const char *device, AioContext **aio_context,
                                Error **errp)
2638 2639 2640 2641
{
    BlockDriverState *bs;

    bs = bdrv_find(device);
2642 2643 2644 2645 2646 2647 2648 2649 2650 2651
    if (!bs) {
        goto notfound;
    }

    *aio_context = bdrv_get_aio_context(bs);
    aio_context_acquire(*aio_context);

    if (!bs->job) {
        aio_context_release(*aio_context);
        goto notfound;
2652
    }
2653

2654
    return bs->job;
2655 2656

notfound:
2657 2658
    error_set(errp, ERROR_CLASS_DEVICE_NOT_ACTIVE,
              "No active block job on device '%s'", device);
2659 2660
    *aio_context = NULL;
    return NULL;
2661 2662
}

2663
void qmp_block_job_set_speed(const char *device, int64_t speed, Error **errp)
2664
{
2665
    AioContext *aio_context;
2666
    BlockJob *job = find_block_job(device, &aio_context, errp);
2667 2668 2669 2670 2671

    if (!job) {
        return;
    }

2672
    block_job_set_speed(job, speed, errp);
2673
    aio_context_release(aio_context);
2674
}
2675

2676 2677
void qmp_block_job_cancel(const char *device,
                          bool has_force, bool force, Error **errp)
2678
{
2679
    AioContext *aio_context;
2680
    BlockJob *job = find_block_job(device, &aio_context, errp);
2681

2682 2683 2684
    if (!job) {
        return;
    }
2685 2686 2687 2688 2689

    if (!has_force) {
        force = false;
    }

2690
    if (job->paused && !force) {
2691 2692
        error_setg(errp, "The block job for device '%s' is currently paused",
                   device);
2693
        goto out;
2694
    }
2695 2696 2697

    trace_qmp_block_job_cancel(job);
    block_job_cancel(job);
2698 2699
out:
    aio_context_release(aio_context);
2700
}
S
Stefan Hajnoczi 已提交
2701

2702 2703
void qmp_block_job_pause(const char *device, Error **errp)
{
2704
    AioContext *aio_context;
2705
    BlockJob *job = find_block_job(device, &aio_context, errp);
2706 2707 2708 2709 2710 2711 2712

    if (!job) {
        return;
    }

    trace_qmp_block_job_pause(job);
    block_job_pause(job);
2713
    aio_context_release(aio_context);
2714 2715 2716 2717
}

void qmp_block_job_resume(const char *device, Error **errp)
{
2718
    AioContext *aio_context;
2719
    BlockJob *job = find_block_job(device, &aio_context, errp);
2720 2721 2722 2723 2724 2725 2726

    if (!job) {
        return;
    }

    trace_qmp_block_job_resume(job);
    block_job_resume(job);
2727
    aio_context_release(aio_context);
2728 2729
}

P
Paolo Bonzini 已提交
2730 2731
void qmp_block_job_complete(const char *device, Error **errp)
{
2732
    AioContext *aio_context;
2733
    BlockJob *job = find_block_job(device, &aio_context, errp);
P
Paolo Bonzini 已提交
2734 2735 2736 2737 2738 2739 2740

    if (!job) {
        return;
    }

    trace_qmp_block_job_complete(job);
    block_job_complete(job, errp);
2741
    aio_context_release(aio_context);
P
Paolo Bonzini 已提交
2742 2743
}

2744 2745 2746 2747 2748 2749
void qmp_change_backing_file(const char *device,
                             const char *image_node_name,
                             const char *backing_file,
                             Error **errp)
{
    BlockDriverState *bs = NULL;
2750
    AioContext *aio_context;
2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763
    BlockDriverState *image_bs = NULL;
    Error *local_err = NULL;
    bool ro;
    int open_flags;
    int ret;

    /* find the top layer BDS of the chain */
    bs = bdrv_find(device);
    if (!bs) {
        error_set(errp, QERR_DEVICE_NOT_FOUND, device);
        return;
    }

2764 2765 2766
    aio_context = bdrv_get_aio_context(bs);
    aio_context_acquire(aio_context);

2767 2768 2769
    image_bs = bdrv_lookup_bs(NULL, image_node_name, &local_err);
    if (local_err) {
        error_propagate(errp, local_err);
2770
        goto out;
2771 2772 2773 2774
    }

    if (!image_bs) {
        error_setg(errp, "image file not found");
2775
        goto out;
2776 2777 2778 2779 2780
    }

    if (bdrv_find_base(image_bs) == image_bs) {
        error_setg(errp, "not allowing backing file change on an image "
                         "without a backing file");
2781
        goto out;
2782 2783 2784 2785 2786
    }

    /* even though we are not necessarily operating on bs, we need it to
     * determine if block ops are currently prohibited on the chain */
    if (bdrv_op_is_blocked(bs, BLOCK_OP_TYPE_CHANGE, errp)) {
2787
        goto out;
2788 2789 2790 2791 2792 2793
    }

    /* final sanity check */
    if (!bdrv_chain_contains(bs, image_bs)) {
        error_setg(errp, "'%s' and image file are not in the same chain",
                   device);
2794
        goto out;
2795 2796 2797 2798 2799 2800 2801 2802 2803 2804
    }

    /* if not r/w, reopen to make r/w */
    open_flags = image_bs->open_flags;
    ro = bdrv_is_read_only(image_bs);

    if (ro) {
        bdrv_reopen(image_bs, open_flags | BDRV_O_RDWR, &local_err);
        if (local_err) {
            error_propagate(errp, local_err);
2805
            goto out;
2806 2807 2808 2809 2810 2811 2812 2813 2814 2815 2816 2817 2818 2819 2820 2821 2822 2823 2824
        }
    }

    ret = bdrv_change_backing_file(image_bs, backing_file,
                               image_bs->drv ? image_bs->drv->format_name : "");

    if (ret < 0) {
        error_setg_errno(errp, -ret, "Could not change backing file to '%s'",
                         backing_file);
        /* don't exit here, so we can try to restore open flags if
         * appropriate */
    }

    if (ro) {
        bdrv_reopen(image_bs, open_flags, &local_err);
        if (local_err) {
            error_propagate(errp, local_err); /* will preserve prior errp */
        }
    }
2825 2826 2827

out:
    aio_context_release(aio_context);
2828 2829
}

K
Kevin Wolf 已提交
2830 2831 2832
void qmp_blockdev_add(BlockdevOptions *options, Error **errp)
{
    QmpOutputVisitor *ov = qmp_output_visitor_new();
2833
    BlockBackend *blk;
K
Kevin Wolf 已提交
2834 2835 2836 2837 2838 2839 2840 2841 2842 2843
    QObject *obj;
    QDict *qdict;
    Error *local_err = NULL;

    /* Require an ID in the top level */
    if (!options->has_id) {
        error_setg(errp, "Block device needs an ID");
        goto fail;
    }

2844
    /* TODO Sort it out in raw-posix and drive_new(): Reject aio=native with
K
Kevin Wolf 已提交
2845
     * cache.direct=false instead of silently switching to aio=threads, except
2846
     * when called from drive_new().
K
Kevin Wolf 已提交
2847 2848 2849
     *
     * For now, simply forbidding the combination for all drivers will do. */
    if (options->has_aio && options->aio == BLOCKDEV_AIO_OPTIONS_NATIVE) {
2850 2851 2852 2853
        bool direct = options->has_cache &&
                      options->cache->has_direct &&
                      options->cache->direct;
        if (!direct) {
K
Kevin Wolf 已提交
2854 2855 2856 2857 2858 2859 2860
            error_setg(errp, "aio=native requires cache.direct=true");
            goto fail;
        }
    }

    visit_type_BlockdevOptions(qmp_output_get_visitor(ov),
                               &options, NULL, &local_err);
2861
    if (local_err) {
K
Kevin Wolf 已提交
2862 2863 2864 2865 2866 2867 2868 2869 2870
        error_propagate(errp, local_err);
        goto fail;
    }

    obj = qmp_output_get_qobject(ov);
    qdict = qobject_to_qdict(obj);

    qdict_flatten(qdict);

2871
    blk = blockdev_init(NULL, qdict, &local_err);
2872
    if (local_err) {
2873
        error_propagate(errp, local_err);
K
Kevin Wolf 已提交
2874 2875 2876
        goto fail;
    }

2877 2878
    if (bdrv_key_required(blk_bs(blk))) {
        blk_unref(blk);
2879 2880 2881 2882
        error_setg(errp, "blockdev-add doesn't support encrypted devices");
        goto fail;
    }

K
Kevin Wolf 已提交
2883 2884 2885 2886
fail:
    qmp_output_visitor_cleanup(ov);
}

2887
BlockJobInfoList *qmp_query_block_jobs(Error **errp)
S
Stefan Hajnoczi 已提交
2888
{
2889 2890
    BlockJobInfoList *head = NULL, **p_next = &head;
    BlockDriverState *bs;
S
Stefan Hajnoczi 已提交
2891

2892
    for (bs = bdrv_next(NULL); bs; bs = bdrv_next(bs)) {
2893 2894 2895 2896
        AioContext *aio_context = bdrv_get_aio_context(bs);

        aio_context_acquire(aio_context);

2897 2898 2899 2900 2901 2902
        if (bs->job) {
            BlockJobInfoList *elem = g_new0(BlockJobInfoList, 1);
            elem->value = block_job_query(bs->job);
            *p_next = elem;
            p_next = &elem->next;
        }
2903 2904

        aio_context_release(aio_context);
S
Stefan Hajnoczi 已提交
2905 2906
    }

2907
    return head;
S
Stefan Hajnoczi 已提交
2908
}
2909

2910
QemuOptsList qemu_common_drive_opts = {
2911
    .name = "drive",
2912
    .head = QTAILQ_HEAD_INITIALIZER(qemu_common_drive_opts.head),
2913 2914 2915 2916 2917
    .desc = {
        {
            .name = "snapshot",
            .type = QEMU_OPT_BOOL,
            .help = "enable/disable snapshot mode",
2918 2919 2920 2921
        },{
            .name = "discard",
            .type = QEMU_OPT_STRING,
            .help = "discard operation (ignore/off, unmap/on)",
2922
        },{
K
Kevin Wolf 已提交
2923 2924 2925 2926 2927 2928 2929 2930 2931 2932 2933
            .name = "cache.writeback",
            .type = QEMU_OPT_BOOL,
            .help = "enables writeback mode for any caches",
        },{
            .name = "cache.direct",
            .type = QEMU_OPT_BOOL,
            .help = "enables use of O_DIRECT (bypass the host page cache)",
        },{
            .name = "cache.no-flush",
            .type = QEMU_OPT_BOOL,
            .help = "ignore any flush requests for the device",
2934 2935 2936 2937 2938 2939 2940 2941 2942 2943 2944 2945 2946 2947 2948 2949 2950
        },{
            .name = "aio",
            .type = QEMU_OPT_STRING,
            .help = "host AIO implementation (threads, native)",
        },{
            .name = "format",
            .type = QEMU_OPT_STRING,
            .help = "disk format (raw, qcow2, ...)",
        },{
            .name = "rerror",
            .type = QEMU_OPT_STRING,
            .help = "read error action",
        },{
            .name = "werror",
            .type = QEMU_OPT_STRING,
            .help = "write error action",
        },{
2951
            .name = "read-only",
2952 2953 2954
            .type = QEMU_OPT_BOOL,
            .help = "open drive file as read-only",
        },{
2955
            .name = "throttling.iops-total",
2956 2957 2958
            .type = QEMU_OPT_NUMBER,
            .help = "limit total I/O operations per second",
        },{
2959
            .name = "throttling.iops-read",
2960 2961 2962
            .type = QEMU_OPT_NUMBER,
            .help = "limit read operations per second",
        },{
2963
            .name = "throttling.iops-write",
2964 2965 2966
            .type = QEMU_OPT_NUMBER,
            .help = "limit write operations per second",
        },{
2967
            .name = "throttling.bps-total",
2968 2969 2970
            .type = QEMU_OPT_NUMBER,
            .help = "limit total bytes per second",
        },{
2971
            .name = "throttling.bps-read",
2972 2973 2974
            .type = QEMU_OPT_NUMBER,
            .help = "limit read bytes per second",
        },{
2975
            .name = "throttling.bps-write",
2976 2977
            .type = QEMU_OPT_NUMBER,
            .help = "limit write bytes per second",
2978 2979 2980 2981 2982 2983 2984 2985 2986 2987 2988 2989 2990 2991 2992 2993 2994 2995 2996 2997 2998 2999 3000 3001
        },{
            .name = "throttling.iops-total-max",
            .type = QEMU_OPT_NUMBER,
            .help = "I/O operations burst",
        },{
            .name = "throttling.iops-read-max",
            .type = QEMU_OPT_NUMBER,
            .help = "I/O operations read burst",
        },{
            .name = "throttling.iops-write-max",
            .type = QEMU_OPT_NUMBER,
            .help = "I/O operations write burst",
        },{
            .name = "throttling.bps-total-max",
            .type = QEMU_OPT_NUMBER,
            .help = "total bytes burst",
        },{
            .name = "throttling.bps-read-max",
            .type = QEMU_OPT_NUMBER,
            .help = "total bytes read burst",
        },{
            .name = "throttling.bps-write-max",
            .type = QEMU_OPT_NUMBER,
            .help = "total bytes write burst",
3002 3003 3004 3005
        },{
            .name = "throttling.iops-size",
            .type = QEMU_OPT_NUMBER,
            .help = "when limiting by iops max size of an I/O in bytes",
3006 3007 3008 3009
        },{
            .name = "copy-on-read",
            .type = QEMU_OPT_BOOL,
            .help = "copy read data from backing file into image file",
3010 3011 3012 3013
        },{
            .name = "detect-zeroes",
            .type = QEMU_OPT_STRING,
            .help = "try to optimize zero writes (off, on, unmap)",
3014 3015 3016 3017
        },
        { /* end of list */ }
    },
};
3018 3019 3020 3021 3022

QemuOptsList qemu_drive_opts = {
    .name = "drive",
    .head = QTAILQ_HEAD_INITIALIZER(qemu_drive_opts.head),
    .desc = {
3023 3024 3025 3026
        /*
         * no elements => accept any params
         * validation will happen later
         */
3027 3028 3029
        { /* end of list */ }
    },
};