blockdev.c 128.3 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
 */

P
Peter Maydell 已提交
33
#include "qemu/osdep.h"
34
#include "sysemu/block-backend.h"
35
#include "sysemu/blockdev.h"
36
#include "hw/block/block.h"
37
#include "block/blockjob.h"
38
#include "block/qdict.h"
39
#include "block/throttle-groups.h"
40
#include "monitor/monitor.h"
41
#include "qemu/error-report.h"
42 43
#include "qemu/option.h"
#include "qemu/config-file.h"
44 45 46
#include "qapi/qapi-commands-block.h"
#include "qapi/qapi-commands-transaction.h"
#include "qapi/qapi-visit-block-core.h"
47
#include "qapi/qmp/qdict.h"
48
#include "qapi/qmp/qnum.h"
49
#include "qapi/qmp/qstring.h"
50
#include "qapi/error.h"
51
#include "qapi/qmp/qerror.h"
52
#include "qapi/qmp/qlist.h"
53
#include "qapi/qobject-output-visitor.h"
54
#include "sysemu/sysemu.h"
55
#include "sysemu/iothread.h"
56
#include "block/block_int.h"
57
#include "block/trace.h"
58
#include "sysemu/arch_init.h"
59
#include "sysemu/qtest.h"
60 61
#include "qemu/cutils.h"
#include "qemu/help_option.h"
62
#include "qemu/throttle-options.h"
63

64 65 66
static QTAILQ_HEAD(, BlockDriverState) monitor_bdrv_states =
    QTAILQ_HEAD_INITIALIZER(monitor_bdrv_states);

67 68
static int do_open_tray(const char *blk_name, const char *qdev_id,
                        bool force, Error **errp);
69 70 71 72 73
static void blockdev_remove_medium(bool has_device, const char *device,
                                   bool has_id, const char *id, Error **errp);
static void blockdev_insert_medium(bool has_device, const char *device,
                                   bool has_id, const char *id,
                                   const char *node_name, Error **errp);
74

75 76 77 78 79 80 81 82 83 84 85 86
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",
};

87
static int if_max_devs[IF_COUNT] = {
88 89 90 91
    /*
     * Do not change these numbers!  They govern how drive option
     * index maps to unit and bus.  That mapping is ABI.
     *
92
     * All controllers used to implement if=T drives need to support
93 94 95 96 97 98 99 100 101 102 103
     * 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,
104 105
};

106 107 108 109 110 111
/**
 * Boards may call this to offer board-by-board overrides
 * of the default, global values.
 */
void override_max_devs(BlockInterfaceType type, int max_devs)
{
112
    BlockBackend *blk;
113 114 115 116 117 118
    DriveInfo *dinfo;

    if (max_devs <= 0) {
        return;
    }

119 120
    for (blk = blk_next(NULL); blk; blk = blk_next(blk)) {
        dinfo = blk_legacy_dinfo(blk);
121 122 123 124 125 126 127 128 129 130 131
        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;
}

132 133 134 135 136 137 138
/*
 * 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.
 */
139
void blockdev_mark_auto_del(BlockBackend *blk)
140
{
141
    DriveInfo *dinfo = blk_legacy_dinfo(blk);
142
    BlockDriverState *bs = blk_bs(blk);
143
    AioContext *aio_context;
144

145
    if (!dinfo) {
146 147 148
        return;
    }

M
Max Reitz 已提交
149 150 151
    if (bs) {
        aio_context = bdrv_get_aio_context(bs);
        aio_context_acquire(aio_context);
152

M
Max Reitz 已提交
153
        if (bs->job) {
154
            job_cancel(&bs->job->job, false);
M
Max Reitz 已提交
155
        }
156

M
Max Reitz 已提交
157 158
        aio_context_release(aio_context);
    }
159

160
    dinfo->auto_del = 1;
161 162
}

163
void blockdev_auto_del(BlockBackend *blk)
164
{
165
    DriveInfo *dinfo = blk_legacy_dinfo(blk);
166

167
    if (dinfo && dinfo->auto_del) {
168
        monitor_remove_blk(blk);
169
        blk_unref(blk);
170 171 172
    }
}

173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189
/**
 * 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;
}

190 191 192 193 194 195 196 197 198 199 200 201
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;
}

202 203
QemuOpts *drive_def(const char *optstr)
{
204
    return qemu_opts_parse_noisily(qemu_find_opts("drive"), optstr, false);
205 206 207
}

QemuOpts *drive_add(BlockInterfaceType type, int index, const char *file,
208
                    const char *optstr)
209 210 211
{
    QemuOpts *opts;

212
    opts = drive_def(optstr);
213 214 215
    if (!opts) {
        return NULL;
    }
216
    if (type != IF_DEFAULT) {
217
        qemu_opt_set(opts, "if", if_name[type], &error_abort);
218 219
    }
    if (index >= 0) {
220
        qemu_opt_set_number(opts, "index", index, &error_abort);
221
    }
222
    if (file)
223
        qemu_opt_set(opts, "file", file, &error_abort);
224 225 226 227 228
    return opts;
}

DriveInfo *drive_get(BlockInterfaceType type, int bus, int unit)
{
229
    BlockBackend *blk;
230 231
    DriveInfo *dinfo;

232 233 234 235
    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) {
236
            return dinfo;
237
        }
238 239 240 241 242
    }

    return NULL;
}

243
void drive_check_orphaned(void)
244
{
245
    BlockBackend *blk;
246
    DriveInfo *dinfo;
247
    Location loc;
248
    bool orphans = false;
249

250 251
    for (blk = blk_next(NULL); blk; blk = blk_next(blk)) {
        dinfo = blk_legacy_dinfo(blk);
252
        if (!blk_get_attached_dev(blk) && !dinfo->is_default &&
253
            dinfo->type != IF_NONE) {
254 255
            loc_push_none(&loc);
            qemu_opts_loc_restore(dinfo->opts);
256
            error_report("machine type does not support"
257 258 259
                         " if=%s,bus=%d,unit=%d",
                         if_name[dinfo->type], dinfo->bus, dinfo->unit);
            loc_pop(&loc);
260
            orphans = true;
261 262 263
        }
    }

264 265 266
    if (orphans) {
        exit(1);
    }
267 268
}

269 270 271 272 273 274 275
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));
}

276 277 278
int drive_get_max_bus(BlockInterfaceType type)
{
    int max_bus;
279
    BlockBackend *blk;
280 281 282
    DriveInfo *dinfo;

    max_bus = -1;
283 284 285
    for (blk = blk_next(NULL); blk; blk = blk_next(blk)) {
        dinfo = blk_legacy_dinfo(blk);
        if (dinfo && dinfo->type == type && dinfo->bus > max_bus) {
286
            max_bus = dinfo->bus;
287
        }
288 289 290 291
    }
    return max_bus;
}

292 293 294 295 296 297 298 299 300 301
/* 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]++);
}

302 303
static void bdrv_format_print(void *opaque, const char *name)
{
304
    error_printf(" %s", name);
305 306
}

307 308
typedef struct {
    QEMUBH *bh;
309 310
    BlockDriverState *bs;
} BDRVPutRefBH;
311

312
static int parse_block_error_action(const char *buf, bool is_read, Error **errp)
313 314
{
    if (!strcmp(buf, "ignore")) {
315
        return BLOCKDEV_ON_ERROR_IGNORE;
316
    } else if (!is_read && !strcmp(buf, "enospc")) {
317
        return BLOCKDEV_ON_ERROR_ENOSPC;
318
    } else if (!strcmp(buf, "stop")) {
319
        return BLOCKDEV_ON_ERROR_STOP;
320
    } else if (!strcmp(buf, "report")) {
321
        return BLOCKDEV_ON_ERROR_REPORT;
322
    } else {
323 324
        error_setg(errp, "'%s' invalid %s error action",
                   buf, is_read ? "read" : "write");
325 326 327 328
        return -1;
    }
}

329 330 331 332 333 334 335 336 337
static bool parse_stats_intervals(BlockAcctStats *stats, QList *intervals,
                                  Error **errp)
{
    const QListEntry *entry;
    for (entry = qlist_first(intervals); entry; entry = qlist_next(entry)) {
        switch (qobject_type(entry->value)) {

        case QTYPE_QSTRING: {
            unsigned long long length;
338 339
            const char *str = qstring_get_str(qobject_to(QString,
                                                         entry->value));
340 341 342 343 344 345 346 347 348 349
            if (parse_uint_full(str, &length, 10) == 0 &&
                length > 0 && length <= UINT_MAX) {
                block_acct_add_interval(stats, (unsigned) length);
            } else {
                error_setg(errp, "Invalid interval length: %s", str);
                return false;
            }
            break;
        }

350
        case QTYPE_QNUM: {
351
            int64_t length = qnum_get_int(qobject_to(QNum, entry->value));
352

353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369
            if (length > 0 && length <= UINT_MAX) {
                block_acct_add_interval(stats, (unsigned) length);
            } else {
                error_setg(errp, "Invalid interval length: %" PRId64, length);
                return false;
            }
            break;
        }

        default:
            error_setg(errp, "The specification of stats-intervals is invalid");
            return false;
        }
    }
    return true;
}

370 371
typedef enum { MEDIA_DISK, MEDIA_CDROM } DriveMediaType;

372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402
/* All parameters but @opts are optional and may be set to NULL. */
static void extract_common_blockdev_options(QemuOpts *opts, int *bdrv_flags,
    const char **throttling_group, ThrottleConfig *throttle_cfg,
    BlockdevDetectZeroesOptions *detect_zeroes, Error **errp)
{
    Error *local_error = NULL;
    const char *aio;

    if (bdrv_flags) {
        if (qemu_opt_get_bool(opts, "copy-on-read", false)) {
            *bdrv_flags |= BDRV_O_COPY_ON_READ;
        }

        if ((aio = qemu_opt_get(opts, "aio")) != NULL) {
            if (!strcmp(aio, "native")) {
                *bdrv_flags |= BDRV_O_NATIVE_AIO;
            } else if (!strcmp(aio, "threads")) {
                /* this is the default */
            } else {
               error_setg(errp, "invalid aio option");
               return;
            }
        }
    }

    /* disk I/O throttling */
    if (throttling_group) {
        *throttling_group = qemu_opt_get(opts, "throttling.group");
    }

    if (throttle_cfg) {
403
        throttle_config_init(throttle_cfg);
404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429
        throttle_cfg->buckets[THROTTLE_BPS_TOTAL].avg =
            qemu_opt_get_number(opts, "throttling.bps-total", 0);
        throttle_cfg->buckets[THROTTLE_BPS_READ].avg  =
            qemu_opt_get_number(opts, "throttling.bps-read", 0);
        throttle_cfg->buckets[THROTTLE_BPS_WRITE].avg =
            qemu_opt_get_number(opts, "throttling.bps-write", 0);
        throttle_cfg->buckets[THROTTLE_OPS_TOTAL].avg =
            qemu_opt_get_number(opts, "throttling.iops-total", 0);
        throttle_cfg->buckets[THROTTLE_OPS_READ].avg =
            qemu_opt_get_number(opts, "throttling.iops-read", 0);
        throttle_cfg->buckets[THROTTLE_OPS_WRITE].avg =
            qemu_opt_get_number(opts, "throttling.iops-write", 0);

        throttle_cfg->buckets[THROTTLE_BPS_TOTAL].max =
            qemu_opt_get_number(opts, "throttling.bps-total-max", 0);
        throttle_cfg->buckets[THROTTLE_BPS_READ].max  =
            qemu_opt_get_number(opts, "throttling.bps-read-max", 0);
        throttle_cfg->buckets[THROTTLE_BPS_WRITE].max =
            qemu_opt_get_number(opts, "throttling.bps-write-max", 0);
        throttle_cfg->buckets[THROTTLE_OPS_TOTAL].max =
            qemu_opt_get_number(opts, "throttling.iops-total-max", 0);
        throttle_cfg->buckets[THROTTLE_OPS_READ].max =
            qemu_opt_get_number(opts, "throttling.iops-read-max", 0);
        throttle_cfg->buckets[THROTTLE_OPS_WRITE].max =
            qemu_opt_get_number(opts, "throttling.iops-write-max", 0);

430 431 432 433 434 435 436 437 438 439 440 441 442
        throttle_cfg->buckets[THROTTLE_BPS_TOTAL].burst_length =
            qemu_opt_get_number(opts, "throttling.bps-total-max-length", 1);
        throttle_cfg->buckets[THROTTLE_BPS_READ].burst_length  =
            qemu_opt_get_number(opts, "throttling.bps-read-max-length", 1);
        throttle_cfg->buckets[THROTTLE_BPS_WRITE].burst_length =
            qemu_opt_get_number(opts, "throttling.bps-write-max-length", 1);
        throttle_cfg->buckets[THROTTLE_OPS_TOTAL].burst_length =
            qemu_opt_get_number(opts, "throttling.iops-total-max-length", 1);
        throttle_cfg->buckets[THROTTLE_OPS_READ].burst_length =
            qemu_opt_get_number(opts, "throttling.iops-read-max-length", 1);
        throttle_cfg->buckets[THROTTLE_OPS_WRITE].burst_length =
            qemu_opt_get_number(opts, "throttling.iops-write-max-length", 1);

443 444 445
        throttle_cfg->op_size =
            qemu_opt_get_number(opts, "throttling.iops-size", 0);

446
        if (!throttle_is_valid(throttle_cfg, errp)) {
447 448 449 450 451 452
            return;
        }
    }

    if (detect_zeroes) {
        *detect_zeroes =
453
            qapi_enum_parse(&BlockdevDetectZeroesOptions_lookup,
454 455 456 457 458 459 460 461 462 463
                            qemu_opt_get(opts, "detect-zeroes"),
                            BLOCKDEV_DETECT_ZEROES_OPTIONS_OFF,
                            &local_error);
        if (local_error) {
            error_propagate(errp, local_error);
            return;
        }
    }
}

464
/* Takes the ownership of bs_opts */
465 466
static BlockBackend *blockdev_init(const char *file, QDict *bs_opts,
                                   Error **errp)
467 468 469 470
{
    const char *buf;
    int bdrv_flags = 0;
    int on_read_error, on_write_error;
471
    bool account_invalid, account_failed;
472
    bool writethrough, read_only;
473
    BlockBackend *blk;
474
    BlockDriverState *bs;
475
    ThrottleConfig cfg;
476
    int snapshot = 0;
477
    Error *error = NULL;
478
    QemuOpts *opts;
479 480
    QDict *interval_dict = NULL;
    QList *interval_list = NULL;
481
    const char *id;
482 483 484
    BlockdevDetectZeroesOptions detect_zeroes =
        BLOCKDEV_DETECT_ZEROES_OPTIONS_OFF;
    const char *throttling_group = NULL;
485

486 487 488
    /* 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");
489
    opts = qemu_opts_create(&qemu_common_drive_opts, id, 1, &error);
490
    if (error) {
491
        error_propagate(errp, error);
492
        goto err_no_opts;
493 494 495
    }

    qemu_opts_absorb_qdict(opts, bs_opts, &error);
496
    if (error) {
497
        error_propagate(errp, error);
498
        goto early_err;
499 500 501 502 503 504
    }

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

505 506 507
    /* extract parameters */
    snapshot = qemu_opt_get_bool(opts, "snapshot", 0);

508 509 510
    account_invalid = qemu_opt_get_bool(opts, "stats-account-invalid", true);
    account_failed = qemu_opt_get_bool(opts, "stats-account-failed", true);

511 512
    writethrough = !qemu_opt_get_bool(opts, BDRV_OPT_CACHE_WB, true);

513 514
    id = qemu_opts_id(opts);

515 516 517 518 519 520 521 522
    qdict_extract_subqdict(bs_opts, &interval_dict, "stats-intervals.");
    qdict_array_split(interval_dict, &interval_list);

    if (qdict_size(interval_dict) != 0) {
        error_setg(errp, "Invalid option stats-intervals.%s",
                   qdict_first(interval_dict)->key);
        goto early_err;
    }
523

524 525 526 527 528
    extract_common_blockdev_options(opts, &bdrv_flags, &throttling_group, &cfg,
                                    &detect_zeroes, &error);
    if (error) {
        error_propagate(errp, error);
        goto early_err;
529 530 531
    }

    if ((buf = qemu_opt_get(opts, "format")) != NULL) {
532 533 534 535
        if (is_help_option(buf)) {
            error_printf("Supported formats:");
            bdrv_iterate_format(bdrv_format_print, NULL);
            error_printf("\n");
536
            goto early_err;
537
        }
538

539 540
        if (qdict_haskey(bs_opts, "driver")) {
            error_setg(errp, "Cannot specify both 'driver' and 'format'");
541
            goto early_err;
542
        }
543
        qdict_put_str(bs_opts, "driver", buf);
544 545
    }

546
    on_write_error = BLOCKDEV_ON_ERROR_ENOSPC;
547
    if ((buf = qemu_opt_get(opts, "werror")) != NULL) {
548
        on_write_error = parse_block_error_action(buf, 0, &error);
549
        if (error) {
550
            error_propagate(errp, error);
551
            goto early_err;
552 553 554
        }
    }

555
    on_read_error = BLOCKDEV_ON_ERROR_REPORT;
556
    if ((buf = qemu_opt_get(opts, "rerror")) != NULL) {
557
        on_read_error = parse_block_error_action(buf, 1, &error);
558
        if (error) {
559
            error_propagate(errp, error);
560
            goto early_err;
561 562 563
        }
    }

564
    if (snapshot) {
565
        bdrv_flags |= BDRV_O_SNAPSHOT;
566 567
    }

568 569
    read_only = qemu_opt_get_bool(opts, BDRV_OPT_READ_ONLY, false);

570
    /* init */
571
    if ((!file || !*file) && !qdict_size(bs_opts)) {
572 573
        BlockBackendRootState *blk_rs;

574
        blk = blk_new(0, BLK_PERM_ALL);
575 576
        blk_rs = blk_get_root_state(blk);
        blk_rs->open_flags    = bdrv_flags;
577
        blk_rs->read_only     = read_only;
578 579
        blk_rs->detect_zeroes = detect_zeroes;

580
        qobject_unref(bs_opts);
581 582
    } else {
        if (file && !*file) {
583 584
            file = NULL;
        }
585

586 587 588 589 590
        /* bdrv_open() defaults to the values in bdrv_flags (for compatibility
         * with other callers) rather than what we want as the real defaults.
         * Apply the defaults here instead. */
        qdict_set_default_str(bs_opts, BDRV_OPT_CACHE_DIRECT, "off");
        qdict_set_default_str(bs_opts, BDRV_OPT_CACHE_NO_FLUSH, "off");
591 592
        qdict_set_default_str(bs_opts, BDRV_OPT_READ_ONLY,
                              read_only ? "on" : "off");
593
        assert((bdrv_flags & BDRV_O_CACHE_MASK) == 0);
594

595 596 597 598
        if (runstate_check(RUN_STATE_INMIGRATE)) {
            bdrv_flags |= BDRV_O_INACTIVE;
        }

599
        blk = blk_new_open(file, NULL, bs_opts, bdrv_flags, errp);
600 601 602 603
        if (!blk) {
            goto err_no_bs_opts;
        }
        bs = blk_bs(blk);
604

605
        bs->detect_zeroes = detect_zeroes;
606

607
        block_acct_setup(blk_get_stats(blk), account_invalid, account_failed);
608

609 610 611 612
        if (!parse_stats_intervals(blk_get_stats(blk), interval_list, errp)) {
            blk_unref(blk);
            blk = NULL;
            goto err_no_bs_opts;
613
        }
614 615
    }

616 617 618
    /* disk I/O throttling */
    if (throttle_enabled(&cfg)) {
        if (!throttling_group) {
619
            throttling_group = id;
620 621 622 623 624
        }
        blk_io_limits_enable(blk, throttling_group);
        blk_set_io_limits(blk, &cfg);
    }

625
    blk_set_enable_write_cache(blk, !writethrough);
626
    blk_set_on_error(blk, on_read_error, on_write_error);
627

628
    if (!monitor_add_blk(blk, id, errp)) {
629 630 631 632 633
        blk_unref(blk);
        blk = NULL;
        goto err_no_bs_opts;
    }

634
err_no_bs_opts:
635
    qemu_opts_del(opts);
636 637
    qobject_unref(interval_dict);
    qobject_unref(interval_list);
638
    return blk;
639

640 641
early_err:
    qemu_opts_del(opts);
642 643
    qobject_unref(interval_dict);
    qobject_unref(interval_list);
644
err_no_opts:
645
    qobject_unref(bs_opts);
646
    return NULL;
647 648
}

649 650 651 652 653
/* Takes the ownership of bs_opts */
static BlockDriverState *bds_tree_init(QDict *bs_opts, Error **errp)
{
    int bdrv_flags = 0;

654 655 656 657 658
    /* bdrv_open() defaults to the values in bdrv_flags (for compatibility
     * with other callers) rather than what we want as the real defaults.
     * Apply the defaults here instead. */
    qdict_set_default_str(bs_opts, BDRV_OPT_CACHE_DIRECT, "off");
    qdict_set_default_str(bs_opts, BDRV_OPT_CACHE_NO_FLUSH, "off");
659
    qdict_set_default_str(bs_opts, BDRV_OPT_READ_ONLY, "off");
660

661 662 663 664
    if (runstate_check(RUN_STATE_INMIGRATE)) {
        bdrv_flags |= BDRV_O_INACTIVE;
    }

665
    return bdrv_open(NULL, NULL, bs_opts, bdrv_flags, errp);
666 667
}

668 669 670 671 672 673 674 675 676 677 678 679 680
void blockdev_close_all_bdrv_states(void)
{
    BlockDriverState *bs, *next_bs;

    QTAILQ_FOREACH_SAFE(bs, &monitor_bdrv_states, monitor_list, next_bs) {
        AioContext *ctx = bdrv_get_aio_context(bs);

        aio_context_acquire(ctx);
        bdrv_unref(bs);
        aio_context_release(ctx);
    }
}

681 682 683 684 685 686 687
/* Iterates over the list of monitor-owned BlockDriverStates */
BlockDriverState *bdrv_next_monitor_owned(BlockDriverState *bs)
{
    return bs ? QTAILQ_NEXT(bs, monitor_list)
              : QTAILQ_FIRST(&monitor_bdrv_states);
}

688 689
static void qemu_opt_rename(QemuOpts *opts, const char *from, const char *to,
                            Error **errp)
690 691 692 693 694
{
    const char *value;

    value = qemu_opt_get(opts, from);
    if (value) {
695 696 697 698 699
        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;
        }
700 701 702 703
    }

    /* rename all items in opts */
    while ((value = qemu_opt_get(opts, from))) {
704
        qemu_opt_set(opts, to, value, &error_abort);
705 706 707 708
        qemu_opt_unset(opts, from);
    }
}

709 710 711 712 713
QemuOptsList qemu_legacy_drive_opts = {
    .name = "drive",
    .head = QTAILQ_HEAD_INITIALIZER(qemu_legacy_drive_opts.head),
    .desc = {
        {
714 715 716 717 718 719 720 721 722 723 724 725
            .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",
        },{
726 727 728
            .name = "media",
            .type = QEMU_OPT_STRING,
            .help = "media type (disk, cdrom)",
729 730 731 732
        },{
            .name = "if",
            .type = QEMU_OPT_STRING,
            .help = "interface (ide, scsi, sd, mtd, floppy, pflash, virtio)",
733 734 735 736
        },{
            .name = "file",
            .type = QEMU_OPT_STRING,
            .help = "file name",
737
        },
738 739 740

        /* Options that are passed on, but have special semantics with -drive */
        {
741
            .name = BDRV_OPT_READ_ONLY,
742 743
            .type = QEMU_OPT_BOOL,
            .help = "open drive file as read-only",
744 745 746 747 748 749 750 751
        },{
            .name = "rerror",
            .type = QEMU_OPT_STRING,
            .help = "read error action",
        },{
            .name = "werror",
            .type = QEMU_OPT_STRING,
            .help = "write error action",
752 753 754 755 756 757
        },{
            .name = "copy-on-read",
            .type = QEMU_OPT_BOOL,
            .help = "copy read data from backing file into image file",
        },

758 759 760 761
        { /* end of list */ }
    },
};

762
DriveInfo *drive_new(QemuOpts *all_opts, BlockInterfaceType block_default_type)
763
{
764
    const char *value;
765
    BlockBackend *blk;
766
    DriveInfo *dinfo = NULL;
767
    QDict *bs_opts;
768 769
    QemuOpts *legacy_opts;
    DriveMediaType media = MEDIA_DISK;
770
    BlockInterfaceType type;
771
    int max_devs, bus_id, unit_id, index;
772
    const char *werror, *rerror;
773 774
    bool read_only = false;
    bool copy_on_read;
775
    const char *filename;
776
    Error *local_err = NULL;
777
    int i;
778

779
    /* Change legacy command line options into QMP ones */
780 781 782 783 784 785 786
    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" },
787

788 789 790
        { "bps",            "throttling.bps-total" },
        { "bps_rd",         "throttling.bps-read" },
        { "bps_wr",         "throttling.bps-write" },
791

792 793 794
        { "iops_max",       "throttling.iops-total-max" },
        { "iops_rd_max",    "throttling.iops-read-max" },
        { "iops_wr_max",    "throttling.iops-write-max" },
795

796 797 798
        { "bps_max",        "throttling.bps-total-max" },
        { "bps_rd_max",     "throttling.bps-read-max" },
        { "bps_wr_max",     "throttling.bps-write-max" },
799

800
        { "iops_size",      "throttling.iops-size" },
801

802 803
        { "group",          "throttling.group" },

804
        { "readonly",       BDRV_OPT_READ_ONLY },
805 806 807
    };

    for (i = 0; i < ARRAY_SIZE(opt_renames); i++) {
808 809 810
        qemu_opt_rename(all_opts, opt_renames[i].from, opt_renames[i].to,
                        &local_err);
        if (local_err) {
811
            error_report_err(local_err);
812 813
            return NULL;
        }
814
    }
815

816 817 818
    value = qemu_opt_get(all_opts, "cache");
    if (value) {
        int flags = 0;
819
        bool writethrough;
820

821
        if (bdrv_parse_cache_mode(value, &flags, &writethrough) != 0) {
822 823 824 825 826
            error_report("invalid cache option");
            return NULL;
        }

        /* Specific options take precedence */
827 828
        if (!qemu_opt_get(all_opts, BDRV_OPT_CACHE_WB)) {
            qemu_opt_set_bool(all_opts, BDRV_OPT_CACHE_WB,
829
                              !writethrough, &error_abort);
830
        }
831 832
        if (!qemu_opt_get(all_opts, BDRV_OPT_CACHE_DIRECT)) {
            qemu_opt_set_bool(all_opts, BDRV_OPT_CACHE_DIRECT,
833
                              !!(flags & BDRV_O_NOCACHE), &error_abort);
834
        }
835 836
        if (!qemu_opt_get(all_opts, BDRV_OPT_CACHE_NO_FLUSH)) {
            qemu_opt_set_bool(all_opts, BDRV_OPT_CACHE_NO_FLUSH,
837
                              !!(flags & BDRV_O_NO_FLUSH), &error_abort);
838 839 840 841
        }
        qemu_opt_unset(all_opts, "cache");
    }

842 843 844 845
    /* Get a QDict for processing the options */
    bs_opts = qdict_new();
    qemu_opts_to_qdict(all_opts, bs_opts);

846 847
    legacy_opts = qemu_opts_create(&qemu_legacy_drive_opts, NULL, 0,
                                   &error_abort);
848
    qemu_opts_absorb_qdict(legacy_opts, bs_opts, &local_err);
849
    if (local_err) {
850
        error_report_err(local_err);
851 852 853 854 855 856 857 858 859 860
        goto fail;
    }

    /* 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;
861
            read_only = true;
862 863 864 865 866 867
        } else {
            error_report("'%s' invalid media", value);
            goto fail;
        }
    }

868
    /* copy-on-read is disabled with a warning for read-only devices */
869
    read_only |= qemu_opt_get_bool(legacy_opts, BDRV_OPT_READ_ONLY, false);
870 871 872
    copy_on_read = qemu_opt_get_bool(legacy_opts, "copy-on-read", false);

    if (read_only && copy_on_read) {
873
        warn_report("disabling copy-on-read on read-only drive");
874 875 876
        copy_on_read = false;
    }

877 878
    qdict_put_str(bs_opts, BDRV_OPT_READ_ONLY, read_only ? "on" : "off");
    qdict_put_str(bs_opts, "copy-on-read", copy_on_read ? "on" : "off");
879

880 881 882 883 884 885 886 887 888 889 890 891 892 893 894
    /* 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;
    }

895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947
    /* 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;
    }

    /* 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);
        }
948
        qdict_put_str(bs_opts, "id", new_id);
949 950 951
        g_free(new_id);
    }

952 953 954
    /* Add virtio block device */
    if (type == IF_VIRTIO) {
        QemuOpts *devopts;
955 956
        devopts = qemu_opts_create(qemu_find_opts("device"), NULL, 0,
                                   &error_abort);
957
        if (arch_type == QEMU_ARCH_S390X) {
958
            qemu_opt_set(devopts, "driver", "virtio-blk-ccw", &error_abort);
959
        } else {
960
            qemu_opt_set(devopts, "driver", "virtio-blk-pci", &error_abort);
961
        }
962 963
        qemu_opt_set(devopts, "drive", qdict_get_str(bs_opts, "id"),
                     &error_abort);
964 965
    }

966 967
    filename = qemu_opt_get(legacy_opts, "file");

968 969 970 971 972 973 974 975
    /* 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;
        }
976
        qdict_put_str(bs_opts, "werror", werror);
977 978 979 980 981 982 983 984 985
    }

    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;
        }
986
        qdict_put_str(bs_opts, "rerror", rerror);
987 988
    }

989
    /* Actual block device init: Functionality shared with blockdev-add */
990
    blk = blockdev_init(filename, bs_opts, &local_err);
991
    bs_opts = NULL;
992
    if (!blk) {
993
        if (local_err) {
994
            error_report_err(local_err);
995
        }
996
        goto fail;
997
    } else {
998
        assert(!local_err);
999 1000
    }

1001 1002
    /* Create legacy DriveInfo */
    dinfo = g_malloc0(sizeof(*dinfo));
1003
    dinfo->opts = all_opts;
1004

1005
    dinfo->type = type;
1006 1007
    dinfo->bus = bus_id;
    dinfo->unit = unit_id;
1008

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
    qobject_unref(bs_opts);
1025
    return dinfo;
1026 1027
}

1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049
static BlockDriverState *qmp_get_root_bs(const char *name, Error **errp)
{
    BlockDriverState *bs;

    bs = bdrv_lookup_bs(name, name, errp);
    if (bs == NULL) {
        return NULL;
    }

    if (!bdrv_is_root_node(bs)) {
        error_setg(errp, "Need a root block node");
        return NULL;
    }

    if (!bdrv_is_inserted(bs)) {
        error_setg(errp, "Device has no medium");
        return NULL;
    }

    return bs;
}

1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072
static BlockBackend *qmp_get_blk(const char *blk_name, const char *qdev_id,
                                 Error **errp)
{
    BlockBackend *blk;

    if (!blk_name == !qdev_id) {
        error_setg(errp, "Need exactly one of 'device' and 'id'");
        return NULL;
    }

    if (qdev_id) {
        blk = blk_by_qdev_id(qdev_id, errp);
    } else {
        blk = blk_by_name(blk_name);
        if (blk == NULL) {
            error_set(errp, ERROR_CLASS_DEVICE_NOT_FOUND,
                      "Device '%s' not found", blk_name);
        }
    }

    return blk;
}

1073
void hmp_commit(Monitor *mon, const QDict *qdict)
1074 1075
{
    const char *device = qdict_get_str(qdict, "device");
1076
    BlockBackend *blk;
1077
    int ret;
1078

1079
    if (!strcmp(device, "all")) {
1080
        ret = blk_commit_all();
1081
    } else {
1082 1083 1084
        BlockDriverState *bs;
        AioContext *aio_context;

1085 1086
        blk = blk_by_name(device);
        if (!blk) {
1087
            monitor_printf(mon, "Device '%s' not found\n", device);
1088
            return;
1089
        }
M
Max Reitz 已提交
1090 1091 1092 1093
        if (!blk_is_available(blk)) {
            monitor_printf(mon, "Device '%s' has no medium\n", device);
            return;
        }
1094 1095 1096 1097 1098 1099 1100 1101

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

        ret = bdrv_commit(bs);

        aio_context_release(aio_context);
1102 1103 1104 1105
    }
    if (ret < 0) {
        monitor_printf(mon, "'commit' error for '%s': %s\n", device,
                       strerror(-ret));
1106 1107 1108
    }
}

1109
static void blockdev_do_action(TransactionAction *action, Error **errp)
1110
{
1111
    TransactionActionList list;
1112

1113
    list.value = action;
1114
    list.next = NULL;
1115
    qmp_transaction(&list, false, NULL, errp);
1116 1117
}

1118 1119 1120 1121 1122
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,
1123
                                bool has_format, const char *format,
1124
                                bool has_mode, NewImageMode mode, Error **errp)
1125
{
1126
    BlockdevSnapshotSync snapshot = {
1127
        .has_device = has_device,
1128
        .device = (char *) device,
1129 1130
        .has_node_name = has_node_name,
        .node_name = (char *) node_name,
1131
        .snapshot_file = (char *) snapshot_file,
1132 1133
        .has_snapshot_node_name = has_snapshot_node_name,
        .snapshot_node_name = (char *) snapshot_node_name,
1134 1135 1136 1137 1138
        .has_format = has_format,
        .format = (char *) format,
        .has_mode = has_mode,
        .mode = mode,
    };
1139 1140
    TransactionAction action = {
        .type = TRANSACTION_ACTION_KIND_BLOCKDEV_SNAPSHOT_SYNC,
1141
        .u.blockdev_snapshot_sync.data = &snapshot,
1142 1143
    };
    blockdev_do_action(&action, errp);
1144 1145
}

1146 1147 1148 1149 1150 1151 1152
void qmp_blockdev_snapshot(const char *node, const char *overlay,
                           Error **errp)
{
    BlockdevSnapshot snapshot_data = {
        .node = (char *) node,
        .overlay = (char *) overlay
    };
1153 1154
    TransactionAction action = {
        .type = TRANSACTION_ACTION_KIND_BLOCKDEV_SNAPSHOT,
1155
        .u.blockdev_snapshot.data = &snapshot_data,
1156 1157
    };
    blockdev_do_action(&action, errp);
1158 1159
}

1160 1161 1162 1163 1164 1165 1166 1167
void qmp_blockdev_snapshot_internal_sync(const char *device,
                                         const char *name,
                                         Error **errp)
{
    BlockdevSnapshotInternal snapshot = {
        .device = (char *) device,
        .name = (char *) name
    };
1168 1169
    TransactionAction action = {
        .type = TRANSACTION_ACTION_KIND_BLOCKDEV_SNAPSHOT_INTERNAL_SYNC,
1170
        .u.blockdev_snapshot_internal_sync.data = &snapshot,
1171 1172
    };
    blockdev_do_action(&action, errp);
1173 1174
}

1175 1176 1177 1178 1179 1180 1181
SnapshotInfo *qmp_blockdev_snapshot_delete_internal_sync(const char *device,
                                                         bool has_id,
                                                         const char *id,
                                                         bool has_name,
                                                         const char *name,
                                                         Error **errp)
{
1182
    BlockDriverState *bs;
1183
    AioContext *aio_context;
1184 1185 1186 1187 1188
    QEMUSnapshotInfo sn;
    Error *local_err = NULL;
    SnapshotInfo *info = NULL;
    int ret;

1189 1190
    bs = qmp_get_root_bs(device, errp);
    if (!bs) {
1191 1192
        return NULL;
    }
1193
    aio_context = bdrv_get_aio_context(bs);
M
Max Reitz 已提交
1194
    aio_context_acquire(aio_context);
1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205

    if (!has_id) {
        id = NULL;
    }

    if (!has_name) {
        name = NULL;
    }

    if (!id && !name) {
        error_setg(errp, "Name or id must be provided");
M
Max Reitz 已提交
1206
        goto out_aio_context;
1207 1208
    }

1209 1210 1211 1212
    if (bdrv_op_is_blocked(bs, BLOCK_OP_TYPE_INTERNAL_SNAPSHOT_DELETE, errp)) {
        goto out_aio_context;
    }

1213
    ret = bdrv_snapshot_find_by_id_and_name(bs, id, name, &sn, &local_err);
1214
    if (local_err) {
1215
        error_propagate(errp, local_err);
1216
        goto out_aio_context;
1217 1218 1219 1220 1221 1222
    }
    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);
1223
        goto out_aio_context;
1224 1225 1226
    }

    bdrv_snapshot_delete(bs, id, name, &local_err);
1227
    if (local_err) {
1228
        error_propagate(errp, local_err);
1229
        goto out_aio_context;
1230 1231
    }

1232 1233
    aio_context_release(aio_context);

1234
    info = g_new0(SnapshotInfo, 1);
1235 1236 1237 1238 1239 1240 1241 1242 1243
    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;
1244 1245 1246 1247

out_aio_context:
    aio_context_release(aio_context);
    return NULL;
1248
}
1249

1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287
/**
 * block_dirty_bitmap_lookup:
 * Return a dirty bitmap (if present), after validating
 * the node reference and bitmap names.
 *
 * @node: The name of the BDS node to search for bitmaps
 * @name: The name of the bitmap to search for
 * @pbs: Output pointer for BDS lookup, if desired. Can be NULL.
 * @paio: Output pointer for aio_context acquisition, if desired. Can be NULL.
 * @errp: Output pointer for error information. Can be NULL.
 *
 * @return: A bitmap object on success, or NULL on failure.
 */
static BdrvDirtyBitmap *block_dirty_bitmap_lookup(const char *node,
                                                  const char *name,
                                                  BlockDriverState **pbs,
                                                  Error **errp)
{
    BlockDriverState *bs;
    BdrvDirtyBitmap *bitmap;

    if (!node) {
        error_setg(errp, "Node cannot be NULL");
        return NULL;
    }
    if (!name) {
        error_setg(errp, "Bitmap name cannot be NULL");
        return NULL;
    }
    bs = bdrv_lookup_bs(node, node, NULL);
    if (!bs) {
        error_setg(errp, "Node '%s' not found", node);
        return NULL;
    }

    bitmap = bdrv_find_dirty_bitmap(bs, name);
    if (!bitmap) {
        error_setg(errp, "Dirty bitmap '%s' not found", name);
1288
        return NULL;
1289 1290 1291 1292 1293 1294 1295 1296 1297
    }

    if (pbs) {
        *pbs = bs;
    }

    return bitmap;
}

1298
/* New and old BlockDriverState structs for atomic group operations */
1299

1300
typedef struct BlkActionState BlkActionState;
1301

1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316
/**
 * BlkActionOps:
 * Table of operations that define an Action.
 *
 * @instance_size: Size of state struct, in bytes.
 * @prepare: Prepare the work, must NOT be NULL.
 * @commit: Commit the changes, can be NULL.
 * @abort: Abort the changes on fail, can be NULL.
 * @clean: Clean up resources after all transaction actions have called
 *         commit() or abort(). Can be NULL.
 *
 * Only prepare() may fail. In a single transaction, only one of commit() or
 * abort() will be called. clean() will always be called if it is present.
 */
typedef struct BlkActionOps {
1317
    size_t instance_size;
1318 1319 1320 1321 1322
    void (*prepare)(BlkActionState *common, Error **errp);
    void (*commit)(BlkActionState *common);
    void (*abort)(BlkActionState *common);
    void (*clean)(BlkActionState *common);
} BlkActionOps;
1323

1324 1325 1326 1327 1328 1329
/**
 * BlkActionState:
 * Describes one Action's state within a Transaction.
 *
 * @action: QAPI-defined enum identifying which Action to perform.
 * @ops: Table of ActionOps this Action can perform.
1330
 * @block_job_txn: Transaction which this action belongs to.
1331 1332 1333 1334 1335
 * @entry: List membership for all Actions in this Transaction.
 *
 * This structure must be arranged as first member in a subclassed type,
 * assuming that the compiler will also arrange it to the same offsets as the
 * base class.
1336
 */
1337
struct BlkActionState {
1338
    TransactionAction *action;
1339
    const BlkActionOps *ops;
1340
    JobTxn *block_job_txn;
1341
    TransactionProperties *txn_props;
1342
    QSIMPLEQ_ENTRY(BlkActionState) entry;
1343 1344
};

1345 1346
/* internal snapshot private data */
typedef struct InternalSnapshotState {
1347
    BlkActionState common;
1348 1349
    BlockDriverState *bs;
    QEMUSnapshotInfo sn;
1350
    bool created;
1351 1352
} InternalSnapshotState;

1353 1354 1355 1356 1357 1358 1359

static int action_check_completion_mode(BlkActionState *s, Error **errp)
{
    if (s->txn_props->completion_mode != ACTION_COMPLETION_MODE_INDIVIDUAL) {
        error_setg(errp,
                   "Action '%s' does not support Transaction property "
                   "completion-mode = %s",
1360 1361
                   TransactionActionKind_str(s->action->type),
                   ActionCompletionMode_str(s->txn_props->completion_mode));
1362 1363 1364 1365 1366
        return -1;
    }
    return 0;
}

1367
static void internal_snapshot_prepare(BlkActionState *common,
1368 1369
                                      Error **errp)
{
1370
    Error *local_err = NULL;
1371 1372 1373 1374 1375 1376 1377 1378
    const char *device;
    const char *name;
    BlockDriverState *bs;
    QEMUSnapshotInfo old_sn, *sn;
    bool ret;
    qemu_timeval tv;
    BlockdevSnapshotInternal *internal;
    InternalSnapshotState *state;
1379
    AioContext *aio_context;
1380 1381
    int ret1;

1382
    g_assert(common->action->type ==
1383
             TRANSACTION_ACTION_KIND_BLOCKDEV_SNAPSHOT_INTERNAL_SYNC);
1384
    internal = common->action->u.blockdev_snapshot_internal_sync.data;
1385 1386 1387 1388 1389 1390 1391
    state = DO_UPCAST(InternalSnapshotState, common, common);

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

    /* 2. check for validation */
1392 1393 1394 1395
    if (action_check_completion_mode(common, errp) < 0) {
        return;
    }

1396 1397
    bs = qmp_get_root_bs(device, errp);
    if (!bs) {
1398 1399 1400
        return;
    }

1401 1402
    aio_context = bdrv_get_aio_context(bs);
    aio_context_acquire(aio_context);
1403

1404
    state->bs = bs;
1405 1406

    /* Paired with .clean() */
1407 1408
    bdrv_drained_begin(bs);

1409
    if (bdrv_op_is_blocked(bs, BLOCK_OP_TYPE_INTERNAL_SNAPSHOT, errp)) {
1410
        goto out;
1411 1412
    }

1413
    if (bdrv_is_read_only(bs)) {
1414
        error_setg(errp, "Device '%s' is read only", device);
1415
        goto out;
1416 1417 1418
    }

    if (!bdrv_can_snapshot(bs)) {
1419 1420 1421
        error_setg(errp, "Block format '%s' used by device '%s' "
                   "does not support internal snapshots",
                   bs->drv->format_name, device);
1422
        goto out;
1423 1424 1425 1426
    }

    if (!strlen(name)) {
        error_setg(errp, "Name is empty");
1427
        goto out;
1428 1429 1430
    }

    /* check whether a snapshot with name exist */
1431 1432 1433 1434
    ret = bdrv_snapshot_find_by_id_and_name(bs, NULL, name, &old_sn,
                                            &local_err);
    if (local_err) {
        error_propagate(errp, local_err);
1435
        goto out;
1436 1437 1438 1439
    } else if (ret) {
        error_setg(errp,
                   "Snapshot with name '%s' already exists on device '%s'",
                   name, device);
1440
        goto out;
1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455
    }

    /* 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);
1456
        goto out;
1457 1458 1459
    }

    /* 4. succeed, mark a snapshot is created */
1460
    state->created = true;
1461 1462 1463

out:
    aio_context_release(aio_context);
1464 1465
}

1466
static void internal_snapshot_abort(BlkActionState *common)
1467 1468 1469 1470 1471
{
    InternalSnapshotState *state =
                             DO_UPCAST(InternalSnapshotState, common, common);
    BlockDriverState *bs = state->bs;
    QEMUSnapshotInfo *sn = &state->sn;
1472
    AioContext *aio_context;
1473 1474
    Error *local_error = NULL;

1475
    if (!state->created) {
1476 1477 1478
        return;
    }

1479 1480 1481
    aio_context = bdrv_get_aio_context(state->bs);
    aio_context_acquire(aio_context);

1482
    if (bdrv_snapshot_delete(bs, sn->id_str, sn->name, &local_error) < 0) {
1483 1484 1485 1486 1487
        error_reportf_err(local_error,
                          "Failed to delete snapshot with id '%s' and "
                          "name '%s' on device '%s' in abort: ",
                          sn->id_str, sn->name,
                          bdrv_get_device_name(bs));
1488
    }
1489 1490

    aio_context_release(aio_context);
1491 1492
}

1493
static void internal_snapshot_clean(BlkActionState *common)
1494 1495 1496
{
    InternalSnapshotState *state = DO_UPCAST(InternalSnapshotState,
                                             common, common);
1497
    AioContext *aio_context;
1498

1499 1500
    if (!state->bs) {
        return;
1501
    }
1502 1503 1504 1505 1506 1507 1508

    aio_context = bdrv_get_aio_context(state->bs);
    aio_context_acquire(aio_context);

    bdrv_drained_end(state->bs);

    aio_context_release(aio_context);
1509 1510
}

1511
/* external snapshot private data */
1512
typedef struct ExternalSnapshotState {
1513
    BlkActionState common;
1514 1515
    BlockDriverState *old_bs;
    BlockDriverState *new_bs;
1516
    bool overlay_appended;
1517
} ExternalSnapshotState;
1518

1519
static void external_snapshot_prepare(BlkActionState *common,
1520 1521
                                      Error **errp)
{
1522
    int flags = 0;
1523
    QDict *options = NULL;
1524
    Error *local_err = NULL;
1525
    /* Device and node name of the image to generate the snapshot from */
1526
    const char *device;
1527
    const char *node_name;
1528 1529 1530
    /* Reference to the new image (for 'blockdev-snapshot') */
    const char *snapshot_ref;
    /* File name of the new image (for 'blockdev-snapshot-sync') */
1531
    const char *new_image_file;
1532 1533
    ExternalSnapshotState *state =
                             DO_UPCAST(ExternalSnapshotState, common, common);
1534
    TransactionAction *action = common->action;
1535
    AioContext *aio_context;
1536

1537 1538 1539 1540 1541
    /* 'blockdev-snapshot' and 'blockdev-snapshot-sync' have similar
     * purpose but a different set of parameters */
    switch (action->type) {
    case TRANSACTION_ACTION_KIND_BLOCKDEV_SNAPSHOT:
        {
1542
            BlockdevSnapshot *s = action->u.blockdev_snapshot.data;
1543 1544 1545 1546 1547 1548 1549 1550
            device = s->node;
            node_name = s->node;
            new_image_file = NULL;
            snapshot_ref = s->overlay;
        }
        break;
    case TRANSACTION_ACTION_KIND_BLOCKDEV_SNAPSHOT_SYNC:
        {
1551
            BlockdevSnapshotSync *s = action->u.blockdev_snapshot_sync.data;
1552 1553 1554 1555 1556 1557 1558 1559
            device = s->has_device ? s->device : NULL;
            node_name = s->has_node_name ? s->node_name : NULL;
            new_image_file = s->snapshot_file;
            snapshot_ref = NULL;
        }
        break;
    default:
        g_assert_not_reached();
1560 1561 1562
    }

    /* start processing */
1563 1564 1565 1566
    if (action_check_completion_mode(common, errp) < 0) {
        return;
    }

1567 1568
    state->old_bs = bdrv_lookup_bs(device, node_name, errp);
    if (!state->old_bs) {
1569 1570 1571
        return;
    }

1572 1573 1574 1575
    aio_context = bdrv_get_aio_context(state->old_bs);
    aio_context_acquire(aio_context);

    /* Paired with .clean() */
1576
    bdrv_drained_begin(state->old_bs);
1577

1578
    if (!bdrv_is_inserted(state->old_bs)) {
1579
        error_setg(errp, QERR_DEVICE_HAS_NO_MEDIUM, device);
1580
        goto out;
1581 1582
    }

1583 1584
    if (bdrv_op_is_blocked(state->old_bs,
                           BLOCK_OP_TYPE_EXTERNAL_SNAPSHOT, errp)) {
1585
        goto out;
1586 1587
    }

1588 1589
    if (!bdrv_is_read_only(state->old_bs)) {
        if (bdrv_flush(state->old_bs)) {
1590
            error_setg(errp, QERR_IO_ERROR);
1591
            goto out;
1592 1593 1594
        }
    }

1595
    if (!bdrv_is_first_non_filter(state->old_bs)) {
1596
        error_setg(errp, QERR_FEATURE_DISABLED, "snapshot");
1597
        goto out;
1598 1599
    }

1600
    if (action->type == TRANSACTION_ACTION_KIND_BLOCKDEV_SNAPSHOT_SYNC) {
1601
        BlockdevSnapshotSync *s = action->u.blockdev_snapshot_sync.data;
1602 1603 1604 1605
        const char *format = s->has_format ? s->format : "qcow2";
        enum NewImageMode mode;
        const char *snapshot_node_name =
            s->has_snapshot_node_name ? s->snapshot_node_name : NULL;
1606

1607 1608
        if (node_name && !snapshot_node_name) {
            error_setg(errp, "New snapshot node name missing");
1609
            goto out;
1610 1611
        }

1612 1613 1614
        if (snapshot_node_name &&
            bdrv_lookup_bs(snapshot_node_name, snapshot_node_name, NULL)) {
            error_setg(errp, "New snapshot node name already in use");
1615
            goto out;
1616 1617 1618
        }

        flags = state->old_bs->open_flags;
1619 1620
        flags &= ~(BDRV_O_SNAPSHOT | BDRV_O_COPY_ON_READ);
        flags |= BDRV_O_NO_BACKING;
1621 1622 1623 1624

        /* create new image w/backing file */
        mode = s->has_mode ? s->mode : NEW_IMAGE_MODE_ABSOLUTE_PATHS;
        if (mode != NEW_IMAGE_MODE_EXISTING) {
1625 1626 1627
            int64_t size = bdrv_getlength(state->old_bs);
            if (size < 0) {
                error_setg_errno(errp, -size, "bdrv_getlength failed");
1628
                goto out;
1629
            }
1630 1631 1632
            bdrv_img_create(new_image_file, format,
                            state->old_bs->filename,
                            state->old_bs->drv->format_name,
1633
                            NULL, size, flags, false, &local_err);
1634 1635
            if (local_err) {
                error_propagate(errp, local_err);
1636
                goto out;
1637 1638 1639 1640 1641
            }
        }

        options = qdict_new();
        if (s->has_snapshot_node_name) {
1642
            qdict_put_str(options, "node-name", snapshot_node_name);
1643
        }
1644
        qdict_put_str(options, "driver", format);
1645 1646
    }

1647 1648
    state->new_bs = bdrv_open(new_image_file, snapshot_ref, options, flags,
                              errp);
1649
    /* We will manually add the backing_hd field to the bs later */
1650
    if (!state->new_bs) {
1651
        goto out;
1652 1653
    }

1654
    if (bdrv_has_blk(state->new_bs)) {
1655
        error_setg(errp, "The snapshot is already in use");
1656
        goto out;
1657 1658 1659 1660
    }

    if (bdrv_op_is_blocked(state->new_bs, BLOCK_OP_TYPE_EXTERNAL_SNAPSHOT,
                           errp)) {
1661
        goto out;
1662 1663 1664 1665
    }

    if (state->new_bs->backing != NULL) {
        error_setg(errp, "The snapshot already has a backing image");
1666
        goto out;
1667 1668 1669 1670
    }

    if (!state->new_bs->drv->supports_backing) {
        error_setg(errp, "The snapshot does not support backing images");
1671
        goto out;
1672 1673
    }

1674
    bdrv_set_aio_context(state->new_bs, aio_context);
1675

1676 1677 1678 1679 1680 1681 1682
    /* This removes our old bs and adds the new bs. This is an operation that
     * can fail, so we need to do it in .prepare; undoing it for abort is
     * always possible. */
    bdrv_ref(state->new_bs);
    bdrv_append(state->new_bs, state->old_bs, &local_err);
    if (local_err) {
        error_propagate(errp, local_err);
1683
        goto out;
1684
    }
1685
    state->overlay_appended = true;
1686 1687 1688

out:
    aio_context_release(aio_context);
1689 1690
}

1691
static void external_snapshot_commit(BlkActionState *common)
1692
{
1693 1694
    ExternalSnapshotState *state =
                             DO_UPCAST(ExternalSnapshotState, common, common);
1695 1696 1697 1698
    AioContext *aio_context;

    aio_context = bdrv_get_aio_context(state->old_bs);
    aio_context_acquire(aio_context);
1699

1700 1701 1702
    /* 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 */
1703
    if (!atomic_read(&state->old_bs->copy_on_read)) {
1704 1705 1706
        bdrv_reopen(state->old_bs, state->old_bs->open_flags & ~BDRV_O_RDWR,
                    NULL);
    }
1707 1708

    aio_context_release(aio_context);
1709 1710
}

1711
static void external_snapshot_abort(BlkActionState *common)
1712
{
1713 1714 1715
    ExternalSnapshotState *state =
                             DO_UPCAST(ExternalSnapshotState, common, common);
    if (state->new_bs) {
1716
        if (state->overlay_appended) {
1717 1718 1719 1720 1721
            AioContext *aio_context;

            aio_context = bdrv_get_aio_context(state->old_bs);
            aio_context_acquire(aio_context);

1722 1723 1724
            bdrv_ref(state->old_bs);   /* we can't let bdrv_set_backind_hd()
                                          close state->old_bs; we need it */
            bdrv_set_backing_hd(state->new_bs, NULL, &error_abort);
1725
            bdrv_replace_node(state->new_bs, state->old_bs, &error_abort);
1726
            bdrv_unref(state->old_bs); /* bdrv_replace_node() ref'ed old_bs */
1727 1728

            aio_context_release(aio_context);
1729
        }
1730
    }
1731 1732
}

1733
static void external_snapshot_clean(BlkActionState *common)
1734 1735 1736
{
    ExternalSnapshotState *state =
                             DO_UPCAST(ExternalSnapshotState, common, common);
1737 1738 1739 1740
    AioContext *aio_context;

    if (!state->old_bs) {
        return;
1741
    }
1742 1743 1744 1745 1746 1747 1748 1749

    aio_context = bdrv_get_aio_context(state->old_bs);
    aio_context_acquire(aio_context);

    bdrv_drained_end(state->old_bs);
    bdrv_unref(state->new_bs);

    aio_context_release(aio_context);
1750 1751
}

1752
typedef struct DriveBackupState {
1753
    BlkActionState common;
1754 1755 1756 1757
    BlockDriverState *bs;
    BlockJob *job;
} DriveBackupState;

1758
static BlockJob *do_drive_backup(DriveBackup *backup, JobTxn *txn,
1759
                            Error **errp);
1760

1761
static void drive_backup_prepare(BlkActionState *common, Error **errp)
1762 1763
{
    DriveBackupState *state = DO_UPCAST(DriveBackupState, common, common);
1764
    BlockDriverState *bs;
1765
    DriveBackup *backup;
1766
    AioContext *aio_context;
1767 1768
    Error *local_err = NULL;

1769
    assert(common->action->type == TRANSACTION_ACTION_KIND_DRIVE_BACKUP);
1770
    backup = common->action->u.drive_backup.data;
1771

1772 1773
    bs = qmp_get_root_bs(backup->device, errp);
    if (!bs) {
1774 1775 1776
        return;
    }

1777 1778 1779 1780
    aio_context = bdrv_get_aio_context(bs);
    aio_context_acquire(aio_context);

    /* Paired with .clean() */
1781
    bdrv_drained_begin(bs);
1782

1783
    state->bs = bs;
1784

1785
    state->job = do_drive_backup(backup, common->block_job_txn, &local_err);
1786
    if (local_err) {
1787
        error_propagate(errp, local_err);
1788
        goto out;
1789
    }
1790 1791 1792

out:
    aio_context_release(aio_context);
1793
}
1794

1795 1796 1797
static void drive_backup_commit(BlkActionState *common)
{
    DriveBackupState *state = DO_UPCAST(DriveBackupState, common, common);
1798 1799 1800 1801 1802
    AioContext *aio_context;

    aio_context = bdrv_get_aio_context(state->bs);
    aio_context_acquire(aio_context);

1803
    assert(state->job);
1804
    job_start(&state->job->job);
1805 1806

    aio_context_release(aio_context);
1807 1808
}

1809
static void drive_backup_abort(BlkActionState *common)
1810 1811 1812
{
    DriveBackupState *state = DO_UPCAST(DriveBackupState, common, common);

1813
    if (state->job) {
1814 1815 1816 1817 1818
        AioContext *aio_context;

        aio_context = bdrv_get_aio_context(state->bs);
        aio_context_acquire(aio_context);

1819
        job_cancel_sync(&state->job->job);
1820 1821

        aio_context_release(aio_context);
1822 1823 1824
    }
}

1825
static void drive_backup_clean(BlkActionState *common)
1826 1827
{
    DriveBackupState *state = DO_UPCAST(DriveBackupState, common, common);
1828
    AioContext *aio_context;
1829

1830 1831
    if (!state->bs) {
        return;
1832
    }
1833 1834 1835 1836 1837 1838 1839

    aio_context = bdrv_get_aio_context(state->bs);
    aio_context_acquire(aio_context);

    bdrv_drained_end(state->bs);

    aio_context_release(aio_context);
1840 1841
}

1842
typedef struct BlockdevBackupState {
1843
    BlkActionState common;
1844 1845 1846 1847
    BlockDriverState *bs;
    BlockJob *job;
} BlockdevBackupState;

1848
static BlockJob *do_blockdev_backup(BlockdevBackup *backup, JobTxn *txn,
1849
                                    Error **errp);
1850

1851
static void blockdev_backup_prepare(BlkActionState *common, Error **errp)
1852 1853 1854
{
    BlockdevBackupState *state = DO_UPCAST(BlockdevBackupState, common, common);
    BlockdevBackup *backup;
1855
    BlockDriverState *bs, *target;
1856
    AioContext *aio_context;
1857 1858
    Error *local_err = NULL;

1859
    assert(common->action->type == TRANSACTION_ACTION_KIND_BLOCKDEV_BACKUP);
1860
    backup = common->action->u.blockdev_backup.data;
1861

1862
    bs = bdrv_lookup_bs(backup->device, backup->device, errp);
1863
    if (!bs) {
1864 1865 1866
        return;
    }

1867
    target = bdrv_lookup_bs(backup->target, backup->target, errp);
M
Max Reitz 已提交
1868
    if (!target) {
1869 1870 1871
        return;
    }

1872 1873
    aio_context = bdrv_get_aio_context(bs);
    if (aio_context != bdrv_get_aio_context(target)) {
1874 1875 1876
        error_setg(errp, "Backup between two IO threads is not implemented");
        return;
    }
1877
    aio_context_acquire(aio_context);
1878
    state->bs = bs;
1879 1880

    /* Paired with .clean() */
1881
    bdrv_drained_begin(state->bs);
1882

1883
    state->job = do_blockdev_backup(backup, common->block_job_txn, &local_err);
1884 1885
    if (local_err) {
        error_propagate(errp, local_err);
1886
        goto out;
1887
    }
1888 1889 1890

out:
    aio_context_release(aio_context);
1891
}
1892

1893 1894 1895
static void blockdev_backup_commit(BlkActionState *common)
{
    BlockdevBackupState *state = DO_UPCAST(BlockdevBackupState, common, common);
1896 1897 1898 1899 1900
    AioContext *aio_context;

    aio_context = bdrv_get_aio_context(state->bs);
    aio_context_acquire(aio_context);

1901
    assert(state->job);
1902
    job_start(&state->job->job);
1903 1904

    aio_context_release(aio_context);
1905 1906
}

1907
static void blockdev_backup_abort(BlkActionState *common)
1908 1909 1910
{
    BlockdevBackupState *state = DO_UPCAST(BlockdevBackupState, common, common);

1911
    if (state->job) {
1912 1913 1914 1915 1916
        AioContext *aio_context;

        aio_context = bdrv_get_aio_context(state->bs);
        aio_context_acquire(aio_context);

1917
        job_cancel_sync(&state->job->job);
1918 1919

        aio_context_release(aio_context);
1920 1921 1922
    }
}

1923
static void blockdev_backup_clean(BlkActionState *common)
1924 1925
{
    BlockdevBackupState *state = DO_UPCAST(BlockdevBackupState, common, common);
1926
    AioContext *aio_context;
1927

1928 1929
    if (!state->bs) {
        return;
1930
    }
1931 1932 1933 1934 1935 1936 1937

    aio_context = bdrv_get_aio_context(state->bs);
    aio_context_acquire(aio_context);

    bdrv_drained_end(state->bs);

    aio_context_release(aio_context);
1938 1939
}

F
Fam Zheng 已提交
1940
typedef struct BlockDirtyBitmapState {
1941
    BlkActionState common;
F
Fam Zheng 已提交
1942 1943 1944 1945
    BdrvDirtyBitmap *bitmap;
    BlockDriverState *bs;
    HBitmap *backup;
    bool prepared;
1946
    bool was_enabled;
F
Fam Zheng 已提交
1947 1948
} BlockDirtyBitmapState;

1949
static void block_dirty_bitmap_add_prepare(BlkActionState *common,
F
Fam Zheng 已提交
1950 1951 1952 1953 1954 1955 1956
                                           Error **errp)
{
    Error *local_err = NULL;
    BlockDirtyBitmapAdd *action;
    BlockDirtyBitmapState *state = DO_UPCAST(BlockDirtyBitmapState,
                                             common, common);

1957 1958 1959 1960
    if (action_check_completion_mode(common, errp) < 0) {
        return;
    }

1961
    action = common->action->u.block_dirty_bitmap_add.data;
F
Fam Zheng 已提交
1962 1963 1964
    /* AIO context taken and released within qmp_block_dirty_bitmap_add */
    qmp_block_dirty_bitmap_add(action->node, action->name,
                               action->has_granularity, action->granularity,
1965
                               action->has_persistent, action->persistent,
1966
                               action->has_autoload, action->autoload,
1967
                               action->has_x_disabled, action->x_disabled,
F
Fam Zheng 已提交
1968 1969 1970 1971 1972 1973 1974 1975 1976
                               &local_err);

    if (!local_err) {
        state->prepared = true;
    } else {
        error_propagate(errp, local_err);
    }
}

1977
static void block_dirty_bitmap_add_abort(BlkActionState *common)
F
Fam Zheng 已提交
1978 1979 1980 1981 1982
{
    BlockDirtyBitmapAdd *action;
    BlockDirtyBitmapState *state = DO_UPCAST(BlockDirtyBitmapState,
                                             common, common);

1983
    action = common->action->u.block_dirty_bitmap_add.data;
F
Fam Zheng 已提交
1984 1985 1986 1987 1988 1989 1990 1991
    /* Should not be able to fail: IF the bitmap was added via .prepare(),
     * then the node reference and bitmap name must have been valid.
     */
    if (state->prepared) {
        qmp_block_dirty_bitmap_remove(action->node, action->name, &error_abort);
    }
}

1992
static void block_dirty_bitmap_clear_prepare(BlkActionState *common,
F
Fam Zheng 已提交
1993 1994 1995 1996 1997 1998
                                             Error **errp)
{
    BlockDirtyBitmapState *state = DO_UPCAST(BlockDirtyBitmapState,
                                             common, common);
    BlockDirtyBitmap *action;

1999 2000 2001 2002
    if (action_check_completion_mode(common, errp) < 0) {
        return;
    }

2003
    action = common->action->u.block_dirty_bitmap_clear.data;
F
Fam Zheng 已提交
2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
    state->bitmap = block_dirty_bitmap_lookup(action->node,
                                              action->name,
                                              &state->bs,
                                              errp);
    if (!state->bitmap) {
        return;
    }

    if (bdrv_dirty_bitmap_frozen(state->bitmap)) {
        error_setg(errp, "Cannot modify a frozen bitmap");
        return;
2015 2016 2017
    } else if (bdrv_dirty_bitmap_qmp_locked(state->bitmap)) {
        error_setg(errp, "Cannot modify a locked bitmap");
        return;
F
Fam Zheng 已提交
2018 2019 2020
    } else if (!bdrv_dirty_bitmap_enabled(state->bitmap)) {
        error_setg(errp, "Cannot clear a disabled bitmap");
        return;
2021 2022 2023
    } else if (bdrv_dirty_bitmap_readonly(state->bitmap)) {
        error_setg(errp, "Cannot clear a readonly bitmap");
        return;
F
Fam Zheng 已提交
2024 2025 2026 2027 2028
    }

    bdrv_clear_dirty_bitmap(state->bitmap, &state->backup);
}

2029
static void block_dirty_bitmap_clear_abort(BlkActionState *common)
F
Fam Zheng 已提交
2030 2031 2032 2033
{
    BlockDirtyBitmapState *state = DO_UPCAST(BlockDirtyBitmapState,
                                             common, common);

2034 2035 2036
    if (state->backup) {
        bdrv_undo_clear_dirty_bitmap(state->bitmap, state->backup);
    }
F
Fam Zheng 已提交
2037 2038
}

2039
static void block_dirty_bitmap_clear_commit(BlkActionState *common)
F
Fam Zheng 已提交
2040 2041 2042 2043 2044 2045 2046
{
    BlockDirtyBitmapState *state = DO_UPCAST(BlockDirtyBitmapState,
                                             common, common);

    hbitmap_free(state->backup);
}

2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114
static void block_dirty_bitmap_enable_prepare(BlkActionState *common,
                                              Error **errp)
{
    BlockDirtyBitmap *action;
    BlockDirtyBitmapState *state = DO_UPCAST(BlockDirtyBitmapState,
                                             common, common);

    if (action_check_completion_mode(common, errp) < 0) {
        return;
    }

    action = common->action->u.x_block_dirty_bitmap_enable.data;
    state->bitmap = block_dirty_bitmap_lookup(action->node,
                                              action->name,
                                              NULL,
                                              errp);
    if (!state->bitmap) {
        return;
    }

    state->was_enabled = bdrv_dirty_bitmap_enabled(state->bitmap);
    bdrv_enable_dirty_bitmap(state->bitmap);
}

static void block_dirty_bitmap_enable_abort(BlkActionState *common)
{
    BlockDirtyBitmapState *state = DO_UPCAST(BlockDirtyBitmapState,
                                             common, common);

    if (!state->was_enabled) {
        bdrv_disable_dirty_bitmap(state->bitmap);
    }
}

static void block_dirty_bitmap_disable_prepare(BlkActionState *common,
                                               Error **errp)
{
    BlockDirtyBitmap *action;
    BlockDirtyBitmapState *state = DO_UPCAST(BlockDirtyBitmapState,
                                             common, common);

    if (action_check_completion_mode(common, errp) < 0) {
        return;
    }

    action = common->action->u.x_block_dirty_bitmap_disable.data;
    state->bitmap = block_dirty_bitmap_lookup(action->node,
                                              action->name,
                                              NULL,
                                              errp);
    if (!state->bitmap) {
        return;
    }

    state->was_enabled = bdrv_dirty_bitmap_enabled(state->bitmap);
    bdrv_disable_dirty_bitmap(state->bitmap);
}

static void block_dirty_bitmap_disable_abort(BlkActionState *common)
{
    BlockDirtyBitmapState *state = DO_UPCAST(BlockDirtyBitmapState,
                                             common, common);

    if (state->was_enabled) {
        bdrv_enable_dirty_bitmap(state->bitmap);
    }
}

2115
static void abort_prepare(BlkActionState *common, Error **errp)
2116 2117 2118 2119
{
    error_setg(errp, "Transaction aborted using Abort action");
}

2120
static void abort_commit(BlkActionState *common)
2121
{
2122
    g_assert_not_reached(); /* this action never succeeds */
2123 2124
}

2125
static const BlkActionOps actions[] = {
2126 2127 2128 2129 2130
    [TRANSACTION_ACTION_KIND_BLOCKDEV_SNAPSHOT] = {
        .instance_size = sizeof(ExternalSnapshotState),
        .prepare  = external_snapshot_prepare,
        .commit   = external_snapshot_commit,
        .abort = external_snapshot_abort,
2131
        .clean = external_snapshot_clean,
2132
    },
2133
    [TRANSACTION_ACTION_KIND_BLOCKDEV_SNAPSHOT_SYNC] = {
2134
        .instance_size = sizeof(ExternalSnapshotState),
2135 2136 2137
        .prepare  = external_snapshot_prepare,
        .commit   = external_snapshot_commit,
        .abort = external_snapshot_abort,
2138
        .clean = external_snapshot_clean,
2139
    },
2140 2141 2142
    [TRANSACTION_ACTION_KIND_DRIVE_BACKUP] = {
        .instance_size = sizeof(DriveBackupState),
        .prepare = drive_backup_prepare,
2143
        .commit = drive_backup_commit,
2144
        .abort = drive_backup_abort,
2145
        .clean = drive_backup_clean,
2146
    },
2147 2148 2149
    [TRANSACTION_ACTION_KIND_BLOCKDEV_BACKUP] = {
        .instance_size = sizeof(BlockdevBackupState),
        .prepare = blockdev_backup_prepare,
2150
        .commit = blockdev_backup_commit,
2151 2152 2153
        .abort = blockdev_backup_abort,
        .clean = blockdev_backup_clean,
    },
2154
    [TRANSACTION_ACTION_KIND_ABORT] = {
2155
        .instance_size = sizeof(BlkActionState),
2156 2157 2158
        .prepare = abort_prepare,
        .commit = abort_commit,
    },
2159 2160 2161 2162
    [TRANSACTION_ACTION_KIND_BLOCKDEV_SNAPSHOT_INTERNAL_SYNC] = {
        .instance_size = sizeof(InternalSnapshotState),
        .prepare  = internal_snapshot_prepare,
        .abort = internal_snapshot_abort,
2163
        .clean = internal_snapshot_clean,
2164
    },
F
Fam Zheng 已提交
2165 2166 2167 2168 2169 2170 2171 2172 2173 2174
    [TRANSACTION_ACTION_KIND_BLOCK_DIRTY_BITMAP_ADD] = {
        .instance_size = sizeof(BlockDirtyBitmapState),
        .prepare = block_dirty_bitmap_add_prepare,
        .abort = block_dirty_bitmap_add_abort,
    },
    [TRANSACTION_ACTION_KIND_BLOCK_DIRTY_BITMAP_CLEAR] = {
        .instance_size = sizeof(BlockDirtyBitmapState),
        .prepare = block_dirty_bitmap_clear_prepare,
        .commit = block_dirty_bitmap_clear_commit,
        .abort = block_dirty_bitmap_clear_abort,
2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185
    },
    [TRANSACTION_ACTION_KIND_X_BLOCK_DIRTY_BITMAP_ENABLE] = {
        .instance_size = sizeof(BlockDirtyBitmapState),
        .prepare = block_dirty_bitmap_enable_prepare,
        .abort = block_dirty_bitmap_enable_abort,
    },
    [TRANSACTION_ACTION_KIND_X_BLOCK_DIRTY_BITMAP_DISABLE] = {
        .instance_size = sizeof(BlockDirtyBitmapState),
        .prepare = block_dirty_bitmap_disable_prepare,
        .abort = block_dirty_bitmap_disable_abort,
     }
2186 2187
};

2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206
/**
 * Allocate a TransactionProperties structure if necessary, and fill
 * that structure with desired defaults if they are unset.
 */
static TransactionProperties *get_transaction_properties(
    TransactionProperties *props)
{
    if (!props) {
        props = g_new0(TransactionProperties, 1);
    }

    if (!props->has_completion_mode) {
        props->has_completion_mode = true;
        props->completion_mode = ACTION_COMPLETION_MODE_INDIVIDUAL;
    }

    return props;
}

2207
/*
2208 2209
 * 'Atomic' group operations.  The operations are performed as a set, and if
 * any fail then we roll back all operations in the group.
2210
 */
2211 2212 2213 2214
void qmp_transaction(TransactionActionList *dev_list,
                     bool has_props,
                     struct TransactionProperties *props,
                     Error **errp)
2215
{
2216
    TransactionActionList *dev_entry = dev_list;
2217
    JobTxn *block_job_txn = NULL;
2218
    BlkActionState *state, *next;
2219
    Error *local_err = NULL;
2220

2221
    QSIMPLEQ_HEAD(snap_bdrv_states, BlkActionState) snap_bdrv_states;
2222 2223
    QSIMPLEQ_INIT(&snap_bdrv_states);

2224
    /* Does this transaction get canceled as a group on failure?
2225
     * If not, we don't really need to make a JobTxn.
2226 2227 2228
     */
    props = get_transaction_properties(props);
    if (props->completion_mode != ACTION_COMPLETION_MODE_INDIVIDUAL) {
2229
        block_job_txn = job_txn_new();
2230 2231
    }

2232
    /* drain all i/o before any operations */
2233 2234
    bdrv_drain_all();

2235
    /* We don't do anything in this loop that commits us to the operations */
2236
    while (NULL != dev_entry) {
2237
        TransactionAction *dev_info = NULL;
2238
        const BlkActionOps *ops;
2239

2240 2241 2242
        dev_info = dev_entry->value;
        dev_entry = dev_entry->next;

2243
        assert(dev_info->type < ARRAY_SIZE(actions));
2244

2245
        ops = &actions[dev_info->type];
2246 2247
        assert(ops->instance_size > 0);

2248 2249 2250
        state = g_malloc0(ops->instance_size);
        state->ops = ops;
        state->action = dev_info;
2251 2252
        state->block_job_txn = block_job_txn;
        state->txn_props = props;
2253
        QSIMPLEQ_INSERT_TAIL(&snap_bdrv_states, state, entry);
2254

2255
        state->ops->prepare(state, &local_err);
2256
        if (local_err) {
2257 2258
            error_propagate(errp, local_err);
            goto delete_and_fail;
2259 2260 2261
        }
    }

2262
    QSIMPLEQ_FOREACH(state, &snap_bdrv_states, entry) {
2263 2264 2265
        if (state->ops->commit) {
            state->ops->commit(state);
        }
2266 2267 2268 2269 2270 2271
    }

    /* success */
    goto exit;

delete_and_fail:
2272
    /* failure, and it is all-or-none; roll back all operations */
2273 2274 2275
    QSIMPLEQ_FOREACH(state, &snap_bdrv_states, entry) {
        if (state->ops->abort) {
            state->ops->abort(state);
2276
        }
2277 2278
    }
exit:
2279 2280 2281
    QSIMPLEQ_FOREACH_SAFE(state, &snap_bdrv_states, entry, next) {
        if (state->ops->clean) {
            state->ops->clean(state);
2282
        }
2283
        g_free(state);
2284
    }
2285 2286 2287
    if (!has_props) {
        qapi_free_TransactionProperties(props);
    }
2288
    job_txn_unref(block_job_txn);
2289 2290
}

2291 2292 2293
void qmp_eject(bool has_device, const char *device,
               bool has_id, const char *id,
               bool has_force, bool force, Error **errp)
2294
{
2295
    Error *local_err = NULL;
2296 2297 2298 2299 2300
    int rc;

    if (!has_force) {
        force = false;
    }
2301

2302 2303 2304
    rc = do_open_tray(has_device ? device : NULL,
                      has_id ? id : NULL,
                      force, &local_err);
2305
    if (rc && rc != -ENOSYS) {
2306
        error_propagate(errp, local_err);
L
Luiz Capitulino 已提交
2307
        return;
2308
    }
2309
    error_free(local_err);
2310

2311
    blockdev_remove_medium(has_device, device, has_id, id, errp);
2312 2313
}

2314 2315 2316
void qmp_block_passwd(bool has_device, const char *device,
                      bool has_node_name, const char *node_name,
                      const char *password, Error **errp)
2317
{
2318 2319
    error_setg(errp,
               "Setting block passwords directly is no longer supported");
2320 2321
}

2322 2323 2324 2325 2326 2327 2328 2329 2330 2331
/*
 * Attempt to open the tray of @device.
 * If @force, ignore its tray lock.
 * Else, if the tray is locked, don't open it, but ask the guest to open it.
 * On error, store an error through @errp and return -errno.
 * If @device does not exist, return -ENODEV.
 * If it has no removable media, return -ENOTSUP.
 * If it has no tray, return -ENOSYS.
 * If the guest was asked to open the tray, return -EINPROGRESS.
 * Else, return 0.
2332
 */
2333 2334
static int do_open_tray(const char *blk_name, const char *qdev_id,
                        bool force, Error **errp)
2335 2336
{
    BlockBackend *blk;
2337
    const char *device = qdev_id ?: blk_name;
2338 2339
    bool locked;

2340
    blk = qmp_get_blk(blk_name, qdev_id, errp);
2341
    if (!blk) {
2342
        return -ENODEV;
2343 2344 2345 2346
    }

    if (!blk_dev_has_removable_media(blk)) {
        error_setg(errp, "Device '%s' is not removable", device);
2347
        return -ENOTSUP;
2348 2349
    }

2350
    if (!blk_dev_has_tray(blk)) {
2351 2352
        error_setg(errp, "Device '%s' does not have a tray", device);
        return -ENOSYS;
2353 2354
    }

2355
    if (blk_dev_is_tray_open(blk)) {
2356
        return 0;
2357 2358 2359 2360 2361 2362 2363 2364
    }

    locked = blk_dev_is_medium_locked(blk);
    if (locked) {
        blk_dev_eject_request(blk, force);
    }

    if (!locked || force) {
2365
        blk_dev_change_media_cb(blk, false, &error_abort);
2366
    }
2367 2368

    if (locked && !force) {
2369 2370 2371
        error_setg(errp, "Device '%s' is locked and force was not specified, "
                   "wait for tray to open and try again", device);
        return -EINPROGRESS;
2372 2373 2374 2375 2376
    }

    return 0;
}

2377 2378 2379
void qmp_blockdev_open_tray(bool has_device, const char *device,
                            bool has_id, const char *id,
                            bool has_force, bool force,
2380 2381
                            Error **errp)
{
2382 2383 2384
    Error *local_err = NULL;
    int rc;

2385 2386 2387
    if (!has_force) {
        force = false;
    }
2388 2389 2390
    rc = do_open_tray(has_device ? device : NULL,
                      has_id ? id : NULL,
                      force, &local_err);
2391 2392 2393 2394 2395
    if (rc && rc != -ENOSYS && rc != -EINPROGRESS) {
        error_propagate(errp, local_err);
        return;
    }
    error_free(local_err);
2396 2397
}

2398 2399 2400
void qmp_blockdev_close_tray(bool has_device, const char *device,
                             bool has_id, const char *id,
                             Error **errp)
2401 2402
{
    BlockBackend *blk;
2403
    Error *local_err = NULL;
2404

2405 2406 2407 2408
    device = has_device ? device : NULL;
    id = has_id ? id : NULL;

    blk = qmp_get_blk(device, id, errp);
2409 2410 2411 2412 2413
    if (!blk) {
        return;
    }

    if (!blk_dev_has_removable_media(blk)) {
2414
        error_setg(errp, "Device '%s' is not removable", device ?: id);
2415 2416 2417
        return;
    }

2418 2419 2420 2421 2422
    if (!blk_dev_has_tray(blk)) {
        /* Ignore this command on tray-less devices */
        return;
    }

2423 2424 2425 2426
    if (!blk_dev_is_tray_open(blk)) {
        return;
    }

2427 2428 2429 2430 2431
    blk_dev_change_media_cb(blk, true, &local_err);
    if (local_err) {
        error_propagate(errp, local_err);
        return;
    }
2432 2433
}

2434 2435
static void blockdev_remove_medium(bool has_device, const char *device,
                                   bool has_id, const char *id, Error **errp)
2436 2437 2438 2439
{
    BlockBackend *blk;
    BlockDriverState *bs;
    AioContext *aio_context;
2440
    bool has_attached_device;
2441

2442 2443 2444 2445
    device = has_device ? device : NULL;
    id = has_id ? id : NULL;

    blk = qmp_get_blk(device, id, errp);
2446 2447 2448 2449 2450
    if (!blk) {
        return;
    }

    /* For BBs without a device, we can exchange the BDS tree at will */
2451
    has_attached_device = blk_get_attached_dev(blk);
2452

2453 2454
    if (has_attached_device && !blk_dev_has_removable_media(blk)) {
        error_setg(errp, "Device '%s' is not removable", device ?: id);
2455 2456 2457
        return;
    }

2458 2459 2460 2461
    if (has_attached_device && blk_dev_has_tray(blk) &&
        !blk_dev_is_tray_open(blk))
    {
        error_setg(errp, "Tray of device '%s' is not open", device ?: id);
2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478
        return;
    }

    bs = blk_bs(blk);
    if (!bs) {
        return;
    }

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

    if (bdrv_op_is_blocked(bs, BLOCK_OP_TYPE_EJECT, errp)) {
        goto out;
    }

    blk_remove_bs(blk);

2479 2480 2481 2482 2483
    if (!blk_dev_has_tray(blk)) {
        /* For tray-less devices, blockdev-open-tray is a no-op (or may not be
         * called at all); therefore, the medium needs to be ejected here.
         * Do it after blk_remove_bs() so blk_is_inserted(blk) returns the @load
         * value passed here (i.e. false). */
2484
        blk_dev_change_media_cb(blk, false, &error_abort);
2485 2486
    }

2487 2488 2489 2490
out:
    aio_context_release(aio_context);
}

2491
void qmp_blockdev_remove_medium(const char *id, Error **errp)
2492 2493 2494 2495
{
    blockdev_remove_medium(false, NULL, true, id, errp);
}

2496
static void qmp_blockdev_insert_anon_medium(BlockBackend *blk,
2497 2498
                                            BlockDriverState *bs, Error **errp)
{
2499
    Error *local_err = NULL;
2500
    bool has_device;
2501
    int ret;
2502 2503 2504 2505 2506

    /* For BBs without a device, we can exchange the BDS tree at will */
    has_device = blk_get_attached_dev(blk);

    if (has_device && !blk_dev_has_removable_media(blk)) {
2507
        error_setg(errp, "Device is not removable");
2508 2509 2510
        return;
    }

2511
    if (has_device && blk_dev_has_tray(blk) && !blk_dev_is_tray_open(blk)) {
2512
        error_setg(errp, "Tray of the device is not open");
2513 2514 2515 2516
        return;
    }

    if (blk_bs(blk)) {
2517
        error_setg(errp, "There already is a medium in the device");
2518 2519 2520
        return;
    }

2521 2522 2523 2524
    ret = blk_insert_bs(blk, bs, errp);
    if (ret < 0) {
        return;
    }
2525

2526 2527 2528 2529 2530 2531
    if (!blk_dev_has_tray(blk)) {
        /* For tray-less devices, blockdev-close-tray is a no-op (or may not be
         * called at all); therefore, the medium needs to be pushed into the
         * slot here.
         * Do it after blk_insert_bs() so blk_is_inserted(blk) returns the @load
         * value passed here (i.e. true). */
2532 2533 2534 2535 2536 2537
        blk_dev_change_media_cb(blk, true, &local_err);
        if (local_err) {
            error_propagate(errp, local_err);
            blk_remove_bs(blk);
            return;
        }
2538
    }
2539 2540
}

2541 2542 2543
static void blockdev_insert_medium(bool has_device, const char *device,
                                   bool has_id, const char *id,
                                   const char *node_name, Error **errp)
2544
{
2545
    BlockBackend *blk;
2546 2547
    BlockDriverState *bs;

2548 2549 2550 2551 2552 2553 2554
    blk = qmp_get_blk(has_device ? device : NULL,
                      has_id ? id : NULL,
                      errp);
    if (!blk) {
        return;
    }

2555 2556 2557 2558 2559 2560
    bs = bdrv_find_node(node_name);
    if (!bs) {
        error_setg(errp, "Node '%s' not found", node_name);
        return;
    }

2561
    if (bdrv_has_blk(bs)) {
2562
        error_setg(errp, "Node '%s' is already in use", node_name);
2563 2564 2565
        return;
    }

2566
    qmp_blockdev_insert_anon_medium(blk, bs, errp);
2567 2568
}

2569 2570
void qmp_blockdev_insert_medium(const char *id, const char *node_name,
                                Error **errp)
2571 2572 2573 2574
{
    blockdev_insert_medium(false, NULL, true, id, node_name, errp);
}

2575 2576 2577
void qmp_blockdev_change_medium(bool has_device, const char *device,
                                bool has_id, const char *id,
                                const char *filename,
2578
                                bool has_format, const char *format,
2579 2580
                                bool has_read_only,
                                BlockdevChangeReadOnlyMode read_only,
2581
                                Error **errp)
2582 2583 2584
{
    BlockBackend *blk;
    BlockDriverState *medium_bs = NULL;
2585
    int bdrv_flags;
2586
    bool detect_zeroes;
2587
    int rc;
2588 2589 2590
    QDict *options = NULL;
    Error *err = NULL;

2591 2592 2593
    blk = qmp_get_blk(has_device ? device : NULL,
                      has_id ? id : NULL,
                      errp);
2594 2595 2596 2597 2598 2599 2600 2601 2602
    if (!blk) {
        goto fail;
    }

    if (blk_bs(blk)) {
        blk_update_root_state(blk);
    }

    bdrv_flags = blk_get_open_flags_from_root_state(blk);
2603 2604
    bdrv_flags &= ~(BDRV_O_TEMPORARY | BDRV_O_SNAPSHOT | BDRV_O_NO_BACKING |
        BDRV_O_PROTOCOL);
2605

2606 2607 2608 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 2624 2625
    if (!has_read_only) {
        read_only = BLOCKDEV_CHANGE_READ_ONLY_MODE_RETAIN;
    }

    switch (read_only) {
    case BLOCKDEV_CHANGE_READ_ONLY_MODE_RETAIN:
        break;

    case BLOCKDEV_CHANGE_READ_ONLY_MODE_READ_ONLY:
        bdrv_flags &= ~BDRV_O_RDWR;
        break;

    case BLOCKDEV_CHANGE_READ_ONLY_MODE_READ_WRITE:
        bdrv_flags |= BDRV_O_RDWR;
        break;

    default:
        abort();
    }

2626 2627
    options = qdict_new();
    detect_zeroes = blk_get_detect_zeroes_from_root_state(blk);
2628
    qdict_put_str(options, "detect-zeroes", detect_zeroes ? "on" : "off");
2629

2630
    if (has_format) {
2631
        qdict_put_str(options, "driver", format);
2632 2633
    }

2634 2635
    medium_bs = bdrv_open(filename, NULL, options, bdrv_flags, errp);
    if (!medium_bs) {
2636 2637 2638
        goto fail;
    }

2639 2640 2641
    rc = do_open_tray(has_device ? device : NULL,
                      has_id ? id : NULL,
                      false, &err);
2642
    if (rc && rc != -ENOSYS) {
2643 2644 2645
        error_propagate(errp, err);
        goto fail;
    }
2646 2647
    error_free(err);
    err = NULL;
2648

2649
    blockdev_remove_medium(has_device, device, has_id, id, &err);
2650 2651 2652 2653 2654
    if (err) {
        error_propagate(errp, err);
        goto fail;
    }

2655
    qmp_blockdev_insert_anon_medium(blk, medium_bs, &err);
2656 2657 2658 2659 2660
    if (err) {
        error_propagate(errp, err);
        goto fail;
    }

2661
    qmp_blockdev_close_tray(has_device, device, has_id, id, errp);
2662 2663 2664 2665 2666 2667 2668 2669

fail:
    /* If the medium has been inserted, the device has its own reference, so
     * ours must be relinquished; and if it has not been inserted successfully,
     * the reference must be relinquished anyway */
    bdrv_unref(medium_bs);
}

2670
/* throttling disk I/O limits */
2671
void qmp_block_set_io_throttle(BlockIOThrottle *arg, Error **errp)
2672
{
2673
    ThrottleConfig cfg;
2674
    BlockDriverState *bs;
2675
    BlockBackend *blk;
2676
    AioContext *aio_context;
2677

2678 2679 2680
    blk = qmp_get_blk(arg->has_device ? arg->device : NULL,
                      arg->has_id ? arg->id : NULL,
                      errp);
2681
    if (!blk) {
2682
        return;
2683
    }
M
Max Reitz 已提交
2684 2685 2686 2687

    aio_context = blk_get_aio_context(blk);
    aio_context_acquire(aio_context);

2688
    bs = blk_bs(blk);
M
Max Reitz 已提交
2689
    if (!bs) {
2690
        error_setg(errp, "Device has no medium");
M
Max Reitz 已提交
2691 2692
        goto out;
    }
2693

2694
    throttle_config_init(&cfg);
2695 2696 2697
    cfg.buckets[THROTTLE_BPS_TOTAL].avg = arg->bps;
    cfg.buckets[THROTTLE_BPS_READ].avg  = arg->bps_rd;
    cfg.buckets[THROTTLE_BPS_WRITE].avg = arg->bps_wr;
2698

2699 2700 2701
    cfg.buckets[THROTTLE_OPS_TOTAL].avg = arg->iops;
    cfg.buckets[THROTTLE_OPS_READ].avg  = arg->iops_rd;
    cfg.buckets[THROTTLE_OPS_WRITE].avg = arg->iops_wr;
2702

2703 2704
    if (arg->has_bps_max) {
        cfg.buckets[THROTTLE_BPS_TOTAL].max = arg->bps_max;
2705
    }
2706 2707
    if (arg->has_bps_rd_max) {
        cfg.buckets[THROTTLE_BPS_READ].max = arg->bps_rd_max;
2708
    }
2709 2710
    if (arg->has_bps_wr_max) {
        cfg.buckets[THROTTLE_BPS_WRITE].max = arg->bps_wr_max;
2711
    }
2712 2713
    if (arg->has_iops_max) {
        cfg.buckets[THROTTLE_OPS_TOTAL].max = arg->iops_max;
2714
    }
2715 2716
    if (arg->has_iops_rd_max) {
        cfg.buckets[THROTTLE_OPS_READ].max = arg->iops_rd_max;
2717
    }
2718 2719
    if (arg->has_iops_wr_max) {
        cfg.buckets[THROTTLE_OPS_WRITE].max = arg->iops_wr_max;
2720
    }
2721

2722 2723
    if (arg->has_bps_max_length) {
        cfg.buckets[THROTTLE_BPS_TOTAL].burst_length = arg->bps_max_length;
2724
    }
2725 2726
    if (arg->has_bps_rd_max_length) {
        cfg.buckets[THROTTLE_BPS_READ].burst_length = arg->bps_rd_max_length;
2727
    }
2728 2729
    if (arg->has_bps_wr_max_length) {
        cfg.buckets[THROTTLE_BPS_WRITE].burst_length = arg->bps_wr_max_length;
2730
    }
2731 2732
    if (arg->has_iops_max_length) {
        cfg.buckets[THROTTLE_OPS_TOTAL].burst_length = arg->iops_max_length;
2733
    }
2734 2735
    if (arg->has_iops_rd_max_length) {
        cfg.buckets[THROTTLE_OPS_READ].burst_length = arg->iops_rd_max_length;
2736
    }
2737 2738
    if (arg->has_iops_wr_max_length) {
        cfg.buckets[THROTTLE_OPS_WRITE].burst_length = arg->iops_wr_max_length;
2739 2740
    }

2741 2742
    if (arg->has_iops_size) {
        cfg.op_size = arg->iops_size;
2743
    }
2744

2745
    if (!throttle_is_valid(&cfg, errp)) {
M
Max Reitz 已提交
2746
        goto out;
2747 2748
    }

2749 2750 2751
    if (throttle_enabled(&cfg)) {
        /* Enable I/O limits if they're not enabled yet, otherwise
         * just update the throttling group. */
2752
        if (!blk_get_public(blk)->throttle_group_member.throttle_state) {
2753
            blk_io_limits_enable(blk,
2754 2755 2756
                                 arg->has_group ? arg->group :
                                 arg->has_device ? arg->device :
                                 arg->id);
2757 2758
        } else if (arg->has_group) {
            blk_io_limits_update_group(blk, arg->group);
2759 2760
        }
        /* Set the new throttling configuration */
2761
        blk_set_io_limits(blk, &cfg);
2762
    } else if (blk_get_public(blk)->throttle_group_member.throttle_state) {
2763
        /* If all throttling settings are set to 0, disable I/O limits */
2764
        blk_io_limits_disable(blk);
2765
    }
2766

M
Max Reitz 已提交
2767
out:
2768
    aio_context_release(aio_context);
2769 2770
}

2771 2772
void qmp_block_dirty_bitmap_add(const char *node, const char *name,
                                bool has_granularity, uint32_t granularity,
2773
                                bool has_persistent, bool persistent,
2774
                                bool has_autoload, bool autoload,
2775
                                bool has_disabled, bool disabled,
2776 2777 2778
                                Error **errp)
{
    BlockDriverState *bs;
2779
    BdrvDirtyBitmap *bitmap;
2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793 2794

    if (!name || name[0] == '\0') {
        error_setg(errp, "Bitmap name cannot be empty");
        return;
    }

    bs = bdrv_lookup_bs(node, node, errp);
    if (!bs) {
        return;
    }

    if (has_granularity) {
        if (granularity < 512 || !is_power_of_2(granularity)) {
            error_setg(errp, "Granularity must be power of 2 "
                             "and at least 512");
2795
            return;
2796 2797 2798 2799 2800 2801
        }
    } else {
        /* Default to cluster size, if available: */
        granularity = bdrv_get_default_bitmap_granularity(bs);
    }

2802 2803 2804
    if (!has_persistent) {
        persistent = false;
    }
2805

2806 2807
    if (has_autoload) {
        warn_report("Autoload option is deprecated and its value is ignored");
2808
    }
2809

2810 2811 2812 2813
    if (!has_disabled) {
        disabled = false;
    }

2814 2815 2816 2817 2818 2819 2820
    if (persistent &&
        !bdrv_can_store_new_dirty_bitmap(bs, name, granularity, errp))
    {
        return;
    }

    bitmap = bdrv_create_dirty_bitmap(bs, granularity, name, errp);
2821 2822
    if (bitmap == NULL) {
        return;
2823
    }
2824

2825 2826 2827 2828
    if (disabled) {
        bdrv_disable_dirty_bitmap(bitmap);
    }

2829
    bdrv_dirty_bitmap_set_persistance(bitmap, persistent);
2830 2831 2832 2833 2834 2835 2836
}

void qmp_block_dirty_bitmap_remove(const char *node, const char *name,
                                   Error **errp)
{
    BlockDriverState *bs;
    BdrvDirtyBitmap *bitmap;
2837
    Error *local_err = NULL;
2838

2839
    bitmap = block_dirty_bitmap_lookup(node, name, &bs, errp);
2840 2841 2842 2843
    if (!bitmap || !bs) {
        return;
    }

2844 2845 2846 2847
    if (bdrv_dirty_bitmap_frozen(bitmap)) {
        error_setg(errp,
                   "Bitmap '%s' is currently frozen and cannot be removed",
                   name);
2848
        return;
2849 2850 2851 2852 2853
    } else if (bdrv_dirty_bitmap_qmp_locked(bitmap)) {
        error_setg(errp,
                   "Bitmap '%s' is currently locked and cannot be removed",
                   name);
        return;
2854
    }
2855 2856 2857 2858 2859 2860 2861 2862 2863

    if (bdrv_dirty_bitmap_get_persistance(bitmap)) {
        bdrv_remove_persistent_dirty_bitmap(bs, name, &local_err);
        if (local_err != NULL) {
            error_propagate(errp, local_err);
            return;
        }
    }

2864 2865 2866
    bdrv_release_dirty_bitmap(bs, bitmap);
}

2867 2868 2869 2870 2871 2872 2873 2874 2875 2876
/**
 * Completely clear a bitmap, for the purposes of synchronizing a bitmap
 * immediately after a full backup operation.
 */
void qmp_block_dirty_bitmap_clear(const char *node, const char *name,
                                  Error **errp)
{
    BdrvDirtyBitmap *bitmap;
    BlockDriverState *bs;

2877
    bitmap = block_dirty_bitmap_lookup(node, name, &bs, errp);
2878 2879 2880 2881 2882 2883 2884 2885
    if (!bitmap || !bs) {
        return;
    }

    if (bdrv_dirty_bitmap_frozen(bitmap)) {
        error_setg(errp,
                   "Bitmap '%s' is currently frozen and cannot be modified",
                   name);
2886
        return;
2887 2888 2889 2890 2891
    } else if (bdrv_dirty_bitmap_qmp_locked(bitmap)) {
        error_setg(errp,
                   "Bitmap '%s' is currently locked and cannot be modified",
                   name);
        return;
2892 2893 2894 2895
    } else if (!bdrv_dirty_bitmap_enabled(bitmap)) {
        error_setg(errp,
                   "Bitmap '%s' is currently disabled and cannot be cleared",
                   name);
2896
        return;
2897 2898 2899
    } else if (bdrv_dirty_bitmap_readonly(bitmap)) {
        error_setg(errp, "Bitmap '%s' is readonly and cannot be cleared", name);
        return;
2900 2901
    }

F
Fam Zheng 已提交
2902
    bdrv_clear_dirty_bitmap(bitmap, NULL);
2903 2904
}

2905 2906 2907 2908 2909 2910 2911 2912 2913 2914 2915 2916 2917 2918 2919 2920 2921 2922 2923 2924 2925 2926 2927 2928 2929 2930 2931 2932 2933 2934 2935 2936 2937 2938 2939 2940 2941 2942 2943 2944 2945 2946
void qmp_x_block_dirty_bitmap_enable(const char *node, const char *name,
                                   Error **errp)
{
    BlockDriverState *bs;
    BdrvDirtyBitmap *bitmap;

    bitmap = block_dirty_bitmap_lookup(node, name, &bs, errp);
    if (!bitmap) {
        return;
    }

    if (bdrv_dirty_bitmap_frozen(bitmap)) {
        error_setg(errp,
                   "Bitmap '%s' is currently frozen and cannot be enabled",
                   name);
        return;
    }

    bdrv_enable_dirty_bitmap(bitmap);
}

void qmp_x_block_dirty_bitmap_disable(const char *node, const char *name,
                                    Error **errp)
{
    BlockDriverState *bs;
    BdrvDirtyBitmap *bitmap;

    bitmap = block_dirty_bitmap_lookup(node, name, &bs, errp);
    if (!bitmap) {
        return;
    }

    if (bdrv_dirty_bitmap_frozen(bitmap)) {
        error_setg(errp,
                   "Bitmap '%s' is currently frozen and cannot be disabled",
                   name);
        return;
    }

    bdrv_disable_dirty_bitmap(bitmap);
}

2947 2948 2949 2950 2951 2952 2953 2954 2955 2956 2957 2958 2959 2960 2961 2962 2963 2964 2965 2966 2967 2968 2969 2970 2971 2972 2973 2974 2975 2976
void qmp_x_block_dirty_bitmap_merge(const char *node, const char *dst_name,
                                    const char *src_name, Error **errp)
{
    BlockDriverState *bs;
    BdrvDirtyBitmap *dst, *src;

    dst = block_dirty_bitmap_lookup(node, dst_name, &bs, errp);
    if (!dst) {
        return;
    }

    if (bdrv_dirty_bitmap_frozen(dst)) {
        error_setg(errp, "Bitmap '%s' is frozen and cannot be modified",
                   dst_name);
        return;
    } else if (bdrv_dirty_bitmap_readonly(dst)) {
        error_setg(errp, "Bitmap '%s' is readonly and cannot be modified",
                   dst_name);
        return;
    }

    src = bdrv_find_dirty_bitmap(bs, src_name);
    if (!src) {
        error_setg(errp, "Dirty bitmap '%s' not found", src_name);
        return;
    }

    bdrv_merge_dirty_bitmap(dst, src, errp);
}

2977 2978 2979 2980 2981 2982 2983 2984 2985 2986 2987 2988 2989 2990 2991 2992 2993 2994 2995 2996 2997 2998 2999 3000 3001
BlockDirtyBitmapSha256 *qmp_x_debug_block_dirty_bitmap_sha256(const char *node,
                                                              const char *name,
                                                              Error **errp)
{
    BdrvDirtyBitmap *bitmap;
    BlockDriverState *bs;
    BlockDirtyBitmapSha256 *ret = NULL;
    char *sha256;

    bitmap = block_dirty_bitmap_lookup(node, name, &bs, errp);
    if (!bitmap || !bs) {
        return NULL;
    }

    sha256 = bdrv_dirty_bitmap_sha256(bitmap, errp);
    if (sha256 == NULL) {
        return NULL;
    }

    ret = g_new(BlockDirtyBitmapSha256, 1);
    ret->sha256 = sha256;

    return ret;
}

3002
void hmp_drive_del(Monitor *mon, const QDict *qdict)
3003 3004
{
    const char *id = qdict_get_str(qdict, "id");
3005
    BlockBackend *blk;
3006
    BlockDriverState *bs;
3007
    AioContext *aio_context;
3008
    Error *local_err = NULL;
3009

3010 3011
    bs = bdrv_find_node(id);
    if (bs) {
3012
        qmp_blockdev_del(id, &local_err);
3013 3014 3015 3016 3017 3018
        if (local_err) {
            error_report_err(local_err);
        }
        return;
    }

3019 3020
    blk = blk_by_name(id);
    if (!blk) {
3021
        error_report("Device '%s' not found", id);
3022
        return;
3023
    }
3024

3025
    if (!blk_legacy_dinfo(blk)) {
3026 3027
        error_report("Deleting device added with blockdev-add"
                     " is not supported");
3028
        return;
3029 3030
    }

M
Max Reitz 已提交
3031
    aio_context = blk_get_aio_context(blk);
3032 3033
    aio_context_acquire(aio_context);

M
Max Reitz 已提交
3034 3035 3036 3037 3038 3039 3040
    bs = blk_bs(blk);
    if (bs) {
        if (bdrv_op_is_blocked(bs, BLOCK_OP_TYPE_DRIVE_DEL, &local_err)) {
            error_report_err(local_err);
            aio_context_release(aio_context);
            return;
        }
3041

3042
        blk_remove_bs(blk);
M
Max Reitz 已提交
3043
    }
3044

3045
    /* Make the BlockBackend and the attached BlockDriverState anonymous */
3046 3047
    monitor_remove_blk(blk);

3048 3049
    /* If this BlockBackend has a device attached to it, its refcount will be
     * decremented when the device is removed; otherwise we have to do so here.
3050
     */
3051
    if (blk_get_attached_dev(blk)) {
3052
        /* Further I/O must not pause the guest */
3053 3054
        blk_set_on_error(blk, BLOCKDEV_ON_ERROR_REPORT,
                         BLOCKDEV_ON_ERROR_REPORT);
3055
    } else {
3056
        blk_unref(blk);
3057 3058
    }

3059
    aio_context_release(aio_context);
3060
}
3061

3062 3063 3064
void qmp_block_resize(bool has_device, const char *device,
                      bool has_node_name, const char *node_name,
                      int64_t size, Error **errp)
3065
{
3066
    Error *local_err = NULL;
3067
    BlockBackend *blk = NULL;
3068
    BlockDriverState *bs;
3069
    AioContext *aio_context;
3070
    int ret;
3071

3072 3073 3074
    bs = bdrv_lookup_bs(has_device ? device : NULL,
                        has_node_name ? node_name : NULL,
                        &local_err);
3075
    if (local_err) {
3076 3077 3078 3079
        error_propagate(errp, local_err);
        return;
    }

3080 3081 3082
    aio_context = bdrv_get_aio_context(bs);
    aio_context_acquire(aio_context);

3083
    if (!bdrv_is_first_non_filter(bs)) {
3084
        error_setg(errp, QERR_FEATURE_DISABLED, "resize");
3085
        goto out;
3086 3087 3088
    }

    if (size < 0) {
3089
        error_setg(errp, QERR_INVALID_PARAMETER_VALUE, "size", "a >0 size");
3090
        goto out;
3091 3092
    }

3093
    if (bdrv_op_is_blocked(bs, BLOCK_OP_TYPE_RESIZE, NULL)) {
3094
        error_setg(errp, QERR_DEVICE_IN_USE, device);
3095
        goto out;
3096 3097
    }

3098
    blk = blk_new(BLK_PERM_RESIZE, BLK_PERM_ALL);
3099 3100 3101 3102
    ret = blk_insert_bs(blk, bs, errp);
    if (ret < 0) {
        goto out;
    }
3103

3104
    bdrv_drained_begin(bs);
3105
    ret = blk_truncate(blk, size, PREALLOC_MODE_OFF, errp);
3106
    bdrv_drained_end(bs);
3107 3108

out:
3109
    blk_unref(blk);
3110
    aio_context_release(aio_context);
3111
}
3112

3113
void qmp_block_stream(bool has_job_id, const char *job_id, const char *device,
3114
                      bool has_base, const char *base,
3115
                      bool has_base_node, const char *base_node,
3116 3117
                      bool has_backing_file, const char *backing_file,
                      bool has_speed, int64_t speed,
3118 3119
                      bool has_on_error, BlockdevOnError on_error,
                      Error **errp)
3120
{
3121
    BlockDriverState *bs, *iter;
3122
    BlockDriverState *base_bs = NULL;
3123
    AioContext *aio_context;
3124
    Error *local_err = NULL;
3125
    const char *base_name = NULL;
3126

3127 3128 3129 3130
    if (!has_on_error) {
        on_error = BLOCKDEV_ON_ERROR_REPORT;
    }

3131
    bs = bdrv_lookup_bs(device, device, errp);
3132
    if (!bs) {
3133 3134 3135
        return;
    }

3136
    aio_context = bdrv_get_aio_context(bs);
3137 3138
    aio_context_acquire(aio_context);

3139 3140 3141 3142 3143 3144
    if (has_base && has_base_node) {
        error_setg(errp, "'base' and 'base-node' cannot be specified "
                   "at the same time");
        goto out;
    }

3145
    if (has_base) {
3146 3147
        base_bs = bdrv_find_backing_image(bs, base);
        if (base_bs == NULL) {
3148
            error_setg(errp, QERR_BASE_NOT_FOUND, base);
3149
            goto out;
3150
        }
3151
        assert(bdrv_get_aio_context(base_bs) == aio_context);
3152
        base_name = base;
3153 3154
    }

3155 3156 3157 3158 3159 3160 3161 3162 3163 3164 3165 3166 3167 3168
    if (has_base_node) {
        base_bs = bdrv_lookup_bs(NULL, base_node, errp);
        if (!base_bs) {
            goto out;
        }
        if (bs == base_bs || !bdrv_chain_contains(bs, base_bs)) {
            error_setg(errp, "Node '%s' is not a backing image of '%s'",
                       base_node, device);
            goto out;
        }
        assert(bdrv_get_aio_context(base_bs) == aio_context);
        base_name = base_bs->filename;
    }

3169 3170 3171 3172 3173 3174 3175
    /* Check for op blockers in the whole chain between bs and base */
    for (iter = bs; iter && iter != base_bs; iter = backing_bs(iter)) {
        if (bdrv_op_is_blocked(iter, BLOCK_OP_TYPE_STREAM, errp)) {
            goto out;
        }
    }

3176 3177 3178 3179 3180
    /* 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");
3181
        goto out;
3182 3183 3184 3185 3186
    }

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

3187
    stream_start(has_job_id ? job_id : NULL, bs, base_bs, base_name,
3188
                 has_speed ? speed : 0, on_error, &local_err);
3189
    if (local_err) {
3190
        error_propagate(errp, local_err);
3191
        goto out;
3192 3193 3194
    }

    trace_qmp_block_stream(bs, bs->job);
3195 3196 3197

out:
    aio_context_release(aio_context);
3198
}
3199

3200
void qmp_block_commit(bool has_job_id, const char *job_id, const char *device,
3201 3202
                      bool has_base, const char *base,
                      bool has_top, const char *top,
3203
                      bool has_backing_file, const char *backing_file,
3204
                      bool has_speed, int64_t speed,
3205
                      bool has_filter_node_name, const char *filter_node_name,
3206 3207 3208
                      Error **errp)
{
    BlockDriverState *bs;
3209
    BlockDriverState *iter;
3210
    BlockDriverState *base_bs, *top_bs;
3211
    AioContext *aio_context;
3212 3213 3214 3215
    Error *local_err = NULL;
    /* This will be part of the QMP command, if/when the
     * BlockdevOnError change for blkmirror makes it in
     */
3216
    BlockdevOnError on_error = BLOCKDEV_ON_ERROR_REPORT;
3217

3218 3219 3220
    if (!has_speed) {
        speed = 0;
    }
3221 3222 3223
    if (!has_filter_node_name) {
        filter_node_name = NULL;
    }
3224

3225 3226 3227 3228 3229
    /* 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. */
3230 3231 3232 3233 3234 3235 3236 3237 3238 3239
    bs = qmp_get_root_bs(device, &local_err);
    if (!bs) {
        bs = bdrv_lookup_bs(device, device, NULL);
        if (!bs) {
            error_free(local_err);
            error_set(errp, ERROR_CLASS_DEVICE_NOT_FOUND,
                      "Device '%s' not found", device);
        } else {
            error_propagate(errp, local_err);
        }
3240
        return;
3241 3242
    }

3243
    aio_context = bdrv_get_aio_context(bs);
3244 3245
    aio_context_acquire(aio_context);

3246
    if (bdrv_op_is_blocked(bs, BLOCK_OP_TYPE_COMMIT_SOURCE, errp)) {
3247
        goto out;
3248 3249 3250 3251 3252
    }

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

3253
    if (has_top && top) {
3254 3255 3256 3257 3258 3259 3260
        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");
3261
        goto out;
3262 3263
    }

3264 3265
    assert(bdrv_get_aio_context(top_bs) == aio_context);

3266 3267 3268 3269 3270 3271 3272
    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) {
3273
        error_setg(errp, QERR_BASE_NOT_FOUND, base ? base : "NULL");
3274
        goto out;
3275 3276
    }

3277 3278
    assert(bdrv_get_aio_context(base_bs) == aio_context);

3279 3280 3281 3282
    for (iter = top_bs; iter != backing_bs(base_bs); iter = backing_bs(iter)) {
        if (bdrv_op_is_blocked(iter, BLOCK_OP_TYPE_COMMIT_TARGET, errp)) {
            goto out;
        }
3283 3284
    }

3285 3286 3287
    /* Do not allow attempts to commit an image into itself */
    if (top_bs == base_bs) {
        error_setg(errp, "cannot commit an image into itself");
3288
        goto out;
3289 3290
    }

3291
    if (top_bs == bs) {
3292 3293 3294
        if (has_backing_file) {
            error_setg(errp, "'backing-file' specified,"
                             " but 'top' is the active layer");
3295
            goto out;
3296
        }
3297
        commit_active_start(has_job_id ? job_id : NULL, bs, base_bs,
3298
                            JOB_DEFAULT, speed, on_error,
3299
                            filter_node_name, NULL, NULL, false, &local_err);
3300
    } else {
3301 3302 3303 3304
        BlockDriverState *overlay_bs = bdrv_find_overlay(bs, top_bs);
        if (bdrv_op_is_blocked(overlay_bs, BLOCK_OP_TYPE_COMMIT_TARGET, errp)) {
            goto out;
        }
3305
        commit_start(has_job_id ? job_id : NULL, bs, base_bs, top_bs, speed,
3306
                     on_error, has_backing_file ? backing_file : NULL,
3307
                     filter_node_name, &local_err);
3308
    }
3309 3310
    if (local_err != NULL) {
        error_propagate(errp, local_err);
3311
        goto out;
3312
    }
3313 3314 3315

out:
    aio_context_release(aio_context);
3316 3317
}

3318
static BlockJob *do_drive_backup(DriveBackup *backup, JobTxn *txn,
3319
                                 Error **errp)
3320 3321 3322
{
    BlockDriverState *bs;
    BlockDriverState *target_bs;
3323
    BlockDriverState *source = NULL;
3324
    BlockJob *job = NULL;
3325
    BdrvDirtyBitmap *bmap = NULL;
3326
    AioContext *aio_context;
3327
    QDict *options = NULL;
3328
    Error *local_err = NULL;
3329
    int flags, job_flags = JOB_DEFAULT;
3330
    int64_t size;
3331
    bool set_backing_hd = false;
3332

3333 3334
    if (!backup->has_speed) {
        backup->speed = 0;
3335
    }
3336 3337
    if (!backup->has_on_source_error) {
        backup->on_source_error = BLOCKDEV_ON_ERROR_REPORT;
3338
    }
3339 3340
    if (!backup->has_on_target_error) {
        backup->on_target_error = BLOCKDEV_ON_ERROR_REPORT;
3341
    }
3342 3343 3344 3345 3346
    if (!backup->has_mode) {
        backup->mode = NEW_IMAGE_MODE_ABSOLUTE_PATHS;
    }
    if (!backup->has_job_id) {
        backup->job_id = NULL;
3347
    }
3348 3349 3350 3351 3352 3353
    if (!backup->has_auto_finalize) {
        backup->auto_finalize = true;
    }
    if (!backup->has_auto_dismiss) {
        backup->auto_dismiss = true;
    }
3354 3355 3356
    if (!backup->has_compress) {
        backup->compress = false;
    }
3357

3358
    bs = qmp_get_root_bs(backup->device, errp);
3359
    if (!bs) {
3360
        return NULL;
3361 3362
    }

3363
    aio_context = bdrv_get_aio_context(bs);
3364 3365
    aio_context_acquire(aio_context);

3366 3367 3368
    if (!backup->has_format) {
        backup->format = backup->mode == NEW_IMAGE_MODE_EXISTING ?
                         NULL : (char*) bs->drv->format_name;
3369 3370
    }

3371
    /* Early check to avoid creating target */
3372
    if (bdrv_op_is_blocked(bs, BLOCK_OP_TYPE_BACKUP_SOURCE, errp)) {
3373
        goto out;
3374 3375
    }

3376
    flags = bs->open_flags | BDRV_O_RDWR;
3377

3378 3379
    /* See if we have a backing HD we can use to create our new image
     * on top of. */
3380
    if (backup->sync == MIRROR_SYNC_MODE_TOP) {
3381
        source = backing_bs(bs);
3382
        if (!source) {
3383
            backup->sync = MIRROR_SYNC_MODE_FULL;
3384 3385
        }
    }
3386
    if (backup->sync == MIRROR_SYNC_MODE_NONE) {
3387
        source = bs;
3388 3389
        flags |= BDRV_O_NO_BACKING;
        set_backing_hd = true;
3390 3391
    }

3392 3393 3394
    size = bdrv_getlength(bs);
    if (size < 0) {
        error_setg_errno(errp, -size, "bdrv_getlength failed");
3395
        goto out;
3396 3397
    }

3398 3399
    if (backup->mode != NEW_IMAGE_MODE_EXISTING) {
        assert(backup->format);
3400
        if (source) {
3401
            bdrv_img_create(backup->target, backup->format, source->filename,
3402
                            source->drv->format_name, NULL,
3403
                            size, flags, false, &local_err);
3404
        } else {
3405
            bdrv_img_create(backup->target, backup->format, NULL, NULL, NULL,
3406
                            size, flags, false, &local_err);
3407
        }
3408 3409
    }

3410
    if (local_err) {
3411
        error_propagate(errp, local_err);
3412
        goto out;
3413 3414
    }

3415
    if (backup->format) {
3416 3417 3418
        if (!options) {
            options = qdict_new();
        }
3419
        qdict_put_str(options, "driver", backup->format);
3420 3421
    }

3422
    target_bs = bdrv_open(backup->target, NULL, options, flags, errp);
3423
    if (!target_bs) {
3424
        goto out;
3425 3426
    }

3427 3428
    bdrv_set_aio_context(target_bs, aio_context);

3429 3430 3431 3432 3433 3434 3435 3436
    if (set_backing_hd) {
        bdrv_set_backing_hd(target_bs, source, &local_err);
        if (local_err) {
            bdrv_unref(target_bs);
            goto out;
        }
    }

3437 3438
    if (backup->has_bitmap) {
        bmap = bdrv_find_dirty_bitmap(bs, backup->bitmap);
3439
        if (!bmap) {
3440
            error_setg(errp, "Bitmap '%s' could not be found", backup->bitmap);
3441
            bdrv_unref(target_bs);
3442 3443
            goto out;
        }
3444 3445 3446 3447 3448 3449
        if (bdrv_dirty_bitmap_qmp_locked(bmap)) {
            error_setg(errp,
                       "Bitmap '%s' is currently locked and cannot be used for "
                       "backup", backup->bitmap);
            goto out;
        }
3450
    }
3451
    if (!backup->auto_finalize) {
3452
        job_flags |= JOB_MANUAL_FINALIZE;
3453 3454
    }
    if (!backup->auto_dismiss) {
3455
        job_flags |= JOB_MANUAL_DISMISS;
3456
    }
3457

3458 3459 3460
    job = backup_job_create(backup->job_id, bs, target_bs, backup->speed,
                            backup->sync, bmap, backup->compress,
                            backup->on_source_error, backup->on_target_error,
3461
                            job_flags, NULL, NULL, txn, &local_err);
3462
    bdrv_unref(target_bs);
3463 3464
    if (local_err != NULL) {
        error_propagate(errp, local_err);
3465
        goto out;
3466
    }
3467 3468 3469

out:
    aio_context_release(aio_context);
3470
    return job;
3471 3472
}

3473
void qmp_drive_backup(DriveBackup *arg, Error **errp)
3474
{
3475 3476 3477 3478

    BlockJob *job;
    job = do_drive_backup(arg, NULL, errp);
    if (job) {
3479
        job_start(&job->job);
3480
    }
3481 3482
}

3483 3484
BlockDeviceInfoList *qmp_query_named_block_nodes(Error **errp)
{
3485
    return bdrv_named_nodes_list(errp);
3486 3487
}

3488
BlockJob *do_blockdev_backup(BlockdevBackup *backup, JobTxn *txn,
3489
                             Error **errp)
3490 3491 3492 3493 3494
{
    BlockDriverState *bs;
    BlockDriverState *target_bs;
    Error *local_err = NULL;
    AioContext *aio_context;
3495
    BlockJob *job = NULL;
3496
    int job_flags = JOB_DEFAULT;
3497

3498 3499
    if (!backup->has_speed) {
        backup->speed = 0;
3500
    }
3501 3502 3503 3504 3505
    if (!backup->has_on_source_error) {
        backup->on_source_error = BLOCKDEV_ON_ERROR_REPORT;
    }
    if (!backup->has_on_target_error) {
        backup->on_target_error = BLOCKDEV_ON_ERROR_REPORT;
3506
    }
3507 3508
    if (!backup->has_job_id) {
        backup->job_id = NULL;
3509
    }
3510 3511 3512 3513 3514 3515
    if (!backup->has_auto_finalize) {
        backup->auto_finalize = true;
    }
    if (!backup->has_auto_dismiss) {
        backup->auto_dismiss = true;
    }
3516 3517 3518
    if (!backup->has_compress) {
        backup->compress = false;
    }
3519

3520
    bs = bdrv_lookup_bs(backup->device, backup->device, errp);
3521
    if (!bs) {
3522
        return NULL;
3523 3524
    }

3525
    aio_context = bdrv_get_aio_context(bs);
3526 3527
    aio_context_acquire(aio_context);

3528
    target_bs = bdrv_lookup_bs(backup->target, backup->target, errp);
3529
    if (!target_bs) {
M
Max Reitz 已提交
3530 3531
        goto out;
    }
3532

3533 3534 3535 3536 3537 3538 3539 3540 3541 3542 3543
    if (bdrv_get_aio_context(target_bs) != aio_context) {
        if (!bdrv_has_blk(target_bs)) {
            /* The target BDS is not attached, we can safely move it to another
             * AioContext. */
            bdrv_set_aio_context(target_bs, aio_context);
        } else {
            error_setg(errp, "Target is attached to a different thread from "
                             "source.");
            goto out;
        }
    }
3544
    if (!backup->auto_finalize) {
3545
        job_flags |= JOB_MANUAL_FINALIZE;
3546 3547
    }
    if (!backup->auto_dismiss) {
3548
        job_flags |= JOB_MANUAL_DISMISS;
3549
    }
3550 3551 3552
    job = backup_job_create(backup->job_id, bs, target_bs, backup->speed,
                            backup->sync, NULL, backup->compress,
                            backup->on_source_error, backup->on_target_error,
3553
                            job_flags, NULL, NULL, txn, &local_err);
3554 3555 3556 3557 3558
    if (local_err != NULL) {
        error_propagate(errp, local_err);
    }
out:
    aio_context_release(aio_context);
3559
    return job;
3560 3561
}

3562 3563
void qmp_blockdev_backup(BlockdevBackup *arg, Error **errp)
{
3564 3565 3566
    BlockJob *job;
    job = do_blockdev_backup(arg, NULL, errp);
    if (job) {
3567
        job_start(&job->job);
3568
    }
3569 3570
}

3571 3572 3573
/* Parameter check and block job starting for drive mirroring.
 * Caller should hold @device and @target's aio context (must be the same).
 **/
3574
static void blockdev_mirror_common(const char *job_id, BlockDriverState *bs,
3575 3576 3577
                                   BlockDriverState *target,
                                   bool has_replaces, const char *replaces,
                                   enum MirrorSyncMode sync,
3578
                                   BlockMirrorBackingMode backing_mode,
3579 3580 3581 3582 3583 3584 3585 3586
                                   bool has_speed, int64_t speed,
                                   bool has_granularity, uint32_t granularity,
                                   bool has_buf_size, int64_t buf_size,
                                   bool has_on_source_error,
                                   BlockdevOnError on_source_error,
                                   bool has_on_target_error,
                                   BlockdevOnError on_target_error,
                                   bool has_unmap, bool unmap,
3587 3588
                                   bool has_filter_node_name,
                                   const char *filter_node_name,
3589
                                   bool has_copy_mode, MirrorCopyMode copy_mode,
3590
                                   Error **errp)
3591 3592 3593 3594 3595
{

    if (!has_speed) {
        speed = 0;
    }
3596 3597 3598 3599 3600 3601
    if (!has_on_source_error) {
        on_source_error = BLOCKDEV_ON_ERROR_REPORT;
    }
    if (!has_on_target_error) {
        on_target_error = BLOCKDEV_ON_ERROR_REPORT;
    }
3602 3603 3604
    if (!has_granularity) {
        granularity = 0;
    }
3605
    if (!has_buf_size) {
W
Wen Congyang 已提交
3606
        buf_size = 0;
3607
    }
3608 3609 3610
    if (!has_unmap) {
        unmap = true;
    }
3611 3612 3613
    if (!has_filter_node_name) {
        filter_node_name = NULL;
    }
3614 3615 3616
    if (!has_copy_mode) {
        copy_mode = MIRROR_COPY_MODE_BACKGROUND;
    }
3617

3618
    if (granularity != 0 && (granularity < 512 || granularity > 1048576 * 64)) {
3619 3620
        error_setg(errp, QERR_INVALID_PARAMETER_VALUE, "granularity",
                   "a value in range [512B, 64MB]");
3621 3622 3623
        return;
    }
    if (granularity & (granularity - 1)) {
3624 3625
        error_setg(errp, QERR_INVALID_PARAMETER_VALUE, "granularity",
                   "power of 2");
3626 3627
        return;
    }
3628

3629 3630 3631
    if (bdrv_op_is_blocked(bs, BLOCK_OP_TYPE_MIRROR_SOURCE, errp)) {
        return;
    }
3632 3633 3634
    if (bdrv_op_is_blocked(target, BLOCK_OP_TYPE_MIRROR_TARGET, errp)) {
        return;
    }
3635 3636 3637 3638 3639 3640 3641 3642

    if (!bs->backing && sync == MIRROR_SYNC_MODE_TOP) {
        sync = MIRROR_SYNC_MODE_FULL;
    }

    /* 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
     */
3643
    mirror_start(job_id, bs, target,
3644
                 has_replaces ? replaces : NULL,
3645
                 speed, granularity, buf_size, sync, backing_mode,
3646
                 on_source_error, on_target_error, unmap, filter_node_name,
3647
                 copy_mode, errp);
3648 3649
}

3650
void qmp_drive_mirror(DriveMirror *arg, Error **errp)
3651 3652 3653 3654
{
    BlockDriverState *bs;
    BlockDriverState *source, *target_bs;
    AioContext *aio_context;
3655
    BlockMirrorBackingMode backing_mode;
3656 3657 3658 3659
    Error *local_err = NULL;
    QDict *options = NULL;
    int flags;
    int64_t size;
3660
    const char *format = arg->format;
3661

3662 3663
    bs = qmp_get_root_bs(arg->device, errp);
    if (!bs) {
3664 3665 3666
        return;
    }

3667 3668 3669 3670 3671
    /* Early check to avoid creating target */
    if (bdrv_op_is_blocked(bs, BLOCK_OP_TYPE_MIRROR_SOURCE, errp)) {
        return;
    }

3672
    aio_context = bdrv_get_aio_context(bs);
3673 3674
    aio_context_acquire(aio_context);

3675 3676
    if (!arg->has_mode) {
        arg->mode = NEW_IMAGE_MODE_ABSOLUTE_PATHS;
3677
    }
3678

3679 3680 3681
    if (!arg->has_format) {
        format = (arg->mode == NEW_IMAGE_MODE_EXISTING
                  ? NULL : bs->drv->format_name);
3682 3683
    }

3684
    flags = bs->open_flags | BDRV_O_RDWR;
3685
    source = backing_bs(bs);
3686 3687
    if (!source && arg->sync == MIRROR_SYNC_MODE_TOP) {
        arg->sync = MIRROR_SYNC_MODE_FULL;
3688
    }
3689
    if (arg->sync == MIRROR_SYNC_MODE_NONE) {
3690 3691
        source = bs;
    }
3692

3693 3694 3695
    size = bdrv_getlength(bs);
    if (size < 0) {
        error_setg_errno(errp, -size, "bdrv_getlength failed");
3696
        goto out;
3697 3698
    }

3699
    if (arg->has_replaces) {
3700
        BlockDriverState *to_replace_bs;
3701 3702
        AioContext *replace_aio_context;
        int64_t replace_size;
3703

3704
        if (!arg->has_node_name) {
3705 3706
            error_setg(errp, "a node-name must be provided when replacing a"
                             " named node of the graph");
3707
            goto out;
3708 3709
        }

3710
        to_replace_bs = check_to_replace_node(bs, arg->replaces, &local_err);
3711 3712 3713

        if (!to_replace_bs) {
            error_propagate(errp, local_err);
3714
            goto out;
3715 3716
        }

3717 3718 3719 3720 3721 3722
        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) {
3723 3724
            error_setg(errp, "cannot replace image with a mirror image of "
                             "different size");
3725
            goto out;
3726 3727 3728
        }
    }

3729
    if (arg->mode == NEW_IMAGE_MODE_ABSOLUTE_PATHS) {
3730 3731 3732 3733 3734
        backing_mode = MIRROR_SOURCE_BACKING_CHAIN;
    } else {
        backing_mode = MIRROR_OPEN_BACKING_CHAIN;
    }

3735 3736 3737
    /* Don't open backing image in create() */
    flags |= BDRV_O_NO_BACKING;

3738 3739
    if ((arg->sync == MIRROR_SYNC_MODE_FULL || !source)
        && arg->mode != NEW_IMAGE_MODE_EXISTING)
3740
    {
3741
        /* create new image w/o backing file */
3742
        assert(format);
3743
        bdrv_img_create(arg->target, format,
3744
                        NULL, NULL, NULL, size, flags, false, &local_err);
3745
    } else {
3746
        switch (arg->mode) {
3747 3748 3749 3750
        case NEW_IMAGE_MODE_EXISTING:
            break;
        case NEW_IMAGE_MODE_ABSOLUTE_PATHS:
            /* create new image with backing file */
3751
            bdrv_img_create(arg->target, format,
3752 3753
                            source->filename,
                            source->drv->format_name,
3754
                            NULL, size, flags, false, &local_err);
3755 3756 3757 3758 3759 3760
            break;
        default:
            abort();
        }
    }

3761
    if (local_err) {
3762
        error_propagate(errp, local_err);
3763
        goto out;
3764 3765
    }

3766
    options = qdict_new();
3767
    if (arg->has_node_name) {
3768
        qdict_put_str(options, "node-name", arg->node_name);
3769
    }
3770
    if (format) {
3771
        qdict_put_str(options, "driver", format);
3772
    }
3773

3774 3775 3776
    /* Mirroring takes care of copy-on-write using the source's backing
     * file.
     */
3777
    target_bs = bdrv_open(arg->target, NULL, options, flags, errp);
3778
    if (!target_bs) {
3779
        goto out;
3780 3781
    }

3782 3783
    bdrv_set_aio_context(target_bs, aio_context);

3784 3785 3786 3787 3788 3789 3790 3791
    blockdev_mirror_common(arg->has_job_id ? arg->job_id : NULL, bs, target_bs,
                           arg->has_replaces, arg->replaces, arg->sync,
                           backing_mode, arg->has_speed, arg->speed,
                           arg->has_granularity, arg->granularity,
                           arg->has_buf_size, arg->buf_size,
                           arg->has_on_source_error, arg->on_source_error,
                           arg->has_on_target_error, arg->on_target_error,
                           arg->has_unmap, arg->unmap,
3792
                           false, NULL,
3793
                           arg->has_copy_mode, arg->copy_mode,
3794
                           &local_err);
3795
    bdrv_unref(target_bs);
3796
    error_propagate(errp, local_err);
3797 3798
out:
    aio_context_release(aio_context);
3799 3800
}

3801 3802
void qmp_blockdev_mirror(bool has_job_id, const char *job_id,
                         const char *device, const char *target,
3803 3804 3805 3806 3807 3808 3809 3810 3811
                         bool has_replaces, const char *replaces,
                         MirrorSyncMode sync,
                         bool has_speed, int64_t speed,
                         bool has_granularity, uint32_t granularity,
                         bool has_buf_size, int64_t buf_size,
                         bool has_on_source_error,
                         BlockdevOnError on_source_error,
                         bool has_on_target_error,
                         BlockdevOnError on_target_error,
3812 3813
                         bool has_filter_node_name,
                         const char *filter_node_name,
3814
                         bool has_copy_mode, MirrorCopyMode copy_mode,
3815 3816 3817 3818 3819
                         Error **errp)
{
    BlockDriverState *bs;
    BlockDriverState *target_bs;
    AioContext *aio_context;
3820
    BlockMirrorBackingMode backing_mode = MIRROR_LEAVE_BACKING_CHAIN;
3821 3822
    Error *local_err = NULL;

3823
    bs = qmp_get_root_bs(device, errp);
3824 3825 3826 3827 3828 3829 3830 3831 3832 3833 3834 3835 3836 3837
    if (!bs) {
        return;
    }

    target_bs = bdrv_lookup_bs(target, target, errp);
    if (!target_bs) {
        return;
    }

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

    bdrv_set_aio_context(target_bs, aio_context);

3838
    blockdev_mirror_common(has_job_id ? job_id : NULL, bs, target_bs,
3839
                           has_replaces, replaces, sync, backing_mode,
3840 3841 3842 3843 3844 3845
                           has_speed, speed,
                           has_granularity, granularity,
                           has_buf_size, buf_size,
                           has_on_source_error, on_source_error,
                           has_on_target_error, on_target_error,
                           true, true,
3846
                           has_filter_node_name, filter_node_name,
3847
                           has_copy_mode, copy_mode,
3848
                           &local_err);
3849
    error_propagate(errp, local_err);
3850 3851 3852 3853

    aio_context_release(aio_context);
}

3854 3855
/* Get a block job using its ID and acquire its AioContext */
static BlockJob *find_block_job(const char *id, AioContext **aio_context,
3856
                                Error **errp)
3857
{
3858
    BlockJob *job;
3859

3860
    assert(id != NULL);
M
Max Reitz 已提交
3861

3862
    *aio_context = NULL;
3863

3864
    job = block_job_get(id);
M
Max Reitz 已提交
3865

3866 3867 3868 3869
    if (!job) {
        error_set(errp, ERROR_CLASS_DEVICE_NOT_ACTIVE,
                  "Block job '%s' not found", id);
        return NULL;
3870
    }
3871

3872 3873
    *aio_context = blk_get_aio_context(job->blk);
    aio_context_acquire(*aio_context);
3874

3875
    return job;
3876 3877
}

3878
void qmp_block_job_set_speed(const char *device, int64_t speed, Error **errp)
3879
{
3880
    AioContext *aio_context;
3881
    BlockJob *job = find_block_job(device, &aio_context, errp);
3882 3883 3884 3885 3886

    if (!job) {
        return;
    }

3887
    block_job_set_speed(job, speed, errp);
3888
    aio_context_release(aio_context);
3889
}
3890

3891 3892
void qmp_block_job_cancel(const char *device,
                          bool has_force, bool force, Error **errp)
3893
{
3894
    AioContext *aio_context;
3895
    BlockJob *job = find_block_job(device, &aio_context, errp);
3896

3897 3898 3899
    if (!job) {
        return;
    }
3900 3901 3902 3903 3904

    if (!has_force) {
        force = false;
    }

3905
    if (job_user_paused(&job->job) && !force) {
3906 3907
        error_setg(errp, "The block job for device '%s' is currently paused",
                   device);
3908
        goto out;
3909
    }
3910 3911

    trace_qmp_block_job_cancel(job);
3912
    job_user_cancel(&job->job, force, errp);
3913 3914
out:
    aio_context_release(aio_context);
3915
}
3916

3917 3918
void qmp_block_job_pause(const char *device, Error **errp)
{
3919
    AioContext *aio_context;
3920
    BlockJob *job = find_block_job(device, &aio_context, errp);
3921

3922
    if (!job) {
3923 3924 3925 3926
        return;
    }

    trace_qmp_block_job_pause(job);
3927
    job_user_pause(&job->job, errp);
3928
    aio_context_release(aio_context);
3929 3930 3931 3932
}

void qmp_block_job_resume(const char *device, Error **errp)
{
3933
    AioContext *aio_context;
3934
    BlockJob *job = find_block_job(device, &aio_context, errp);
3935

3936
    if (!job) {
3937 3938 3939 3940
        return;
    }

    trace_qmp_block_job_resume(job);
3941
    job_user_resume(&job->job, errp);
3942
    aio_context_release(aio_context);
3943 3944
}

3945 3946
void qmp_block_job_complete(const char *device, Error **errp)
{
3947
    AioContext *aio_context;
3948
    BlockJob *job = find_block_job(device, &aio_context, errp);
3949 3950 3951 3952 3953 3954

    if (!job) {
        return;
    }

    trace_qmp_block_job_complete(job);
3955
    job_complete(&job->job, errp);
3956
    aio_context_release(aio_context);
3957 3958
}

3959 3960 3961 3962 3963 3964 3965 3966 3967 3968
void qmp_block_job_finalize(const char *id, Error **errp)
{
    AioContext *aio_context;
    BlockJob *job = find_block_job(id, &aio_context, errp);

    if (!job) {
        return;
    }

    trace_qmp_block_job_finalize(job);
3969
    job_finalize(&job->job, errp);
3970 3971 3972
    aio_context_release(aio_context);
}

3973 3974 3975
void qmp_block_job_dismiss(const char *id, Error **errp)
{
    AioContext *aio_context;
K
Kevin Wolf 已提交
3976 3977
    BlockJob *bjob = find_block_job(id, &aio_context, errp);
    Job *job;
3978

K
Kevin Wolf 已提交
3979
    if (!bjob) {
3980 3981 3982
        return;
    }

K
Kevin Wolf 已提交
3983 3984 3985
    trace_qmp_block_job_dismiss(bjob);
    job = &bjob->job;
    job_dismiss(&job, errp);
3986 3987 3988
    aio_context_release(aio_context);
}

3989 3990 3991 3992 3993 3994
void qmp_change_backing_file(const char *device,
                             const char *image_node_name,
                             const char *backing_file,
                             Error **errp)
{
    BlockDriverState *bs = NULL;
3995
    AioContext *aio_context;
3996 3997 3998 3999 4000 4001
    BlockDriverState *image_bs = NULL;
    Error *local_err = NULL;
    bool ro;
    int open_flags;
    int ret;

4002 4003
    bs = qmp_get_root_bs(device, errp);
    if (!bs) {
4004 4005 4006
        return;
    }

4007
    aio_context = bdrv_get_aio_context(bs);
4008 4009
    aio_context_acquire(aio_context);

4010 4011 4012
    image_bs = bdrv_lookup_bs(NULL, image_node_name, &local_err);
    if (local_err) {
        error_propagate(errp, local_err);
4013
        goto out;
4014 4015 4016 4017
    }

    if (!image_bs) {
        error_setg(errp, "image file not found");
4018
        goto out;
4019 4020 4021 4022 4023
    }

    if (bdrv_find_base(image_bs) == image_bs) {
        error_setg(errp, "not allowing backing file change on an image "
                         "without a backing file");
4024
        goto out;
4025 4026 4027 4028 4029
    }

    /* 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)) {
4030
        goto out;
4031 4032 4033 4034 4035 4036
    }

    /* final sanity check */
    if (!bdrv_chain_contains(bs, image_bs)) {
        error_setg(errp, "'%s' and image file are not in the same chain",
                   device);
4037
        goto out;
4038 4039 4040 4041 4042 4043 4044 4045 4046 4047
    }

    /* 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);
4048
            goto out;
4049 4050 4051 4052 4053 4054 4055 4056 4057 4058 4059 4060 4061 4062 4063
        }
    }

    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);
4064
        error_propagate(errp, local_err);
4065
    }
4066 4067 4068

out:
    aio_context_release(aio_context);
4069 4070
}

4071 4072 4073 4074 4075 4076 4077 4078 4079 4080 4081 4082 4083 4084
void hmp_drive_add_node(Monitor *mon, const char *optstr)
{
    QemuOpts *opts;
    QDict *qdict;
    Error *local_err = NULL;

    opts = qemu_opts_parse_noisily(&qemu_drive_opts, optstr, false);
    if (!opts) {
        return;
    }

    qdict = qemu_opts_to_qdict(opts, NULL);

    if (!qdict_get_try_str(qdict, "node-name")) {
4085
        qobject_unref(qdict);
4086 4087 4088 4089 4090 4091 4092 4093 4094 4095 4096 4097 4098 4099 4100 4101
        error_report("'node-name' needs to be specified");
        goto out;
    }

    BlockDriverState *bs = bds_tree_init(qdict, &local_err);
    if (!bs) {
        error_report_err(local_err);
        goto out;
    }

    QTAILQ_INSERT_TAIL(&monitor_bdrv_states, bs, monitor_list);

out:
    qemu_opts_del(opts);
}

4102 4103
void qmp_blockdev_add(BlockdevOptions *options, Error **errp)
{
4104
    BlockDriverState *bs;
4105
    QObject *obj;
4106
    Visitor *v = qobject_output_visitor_new(&obj);
4107 4108 4109
    QDict *qdict;
    Error *local_err = NULL;

4110
    visit_type_BlockdevOptions(v, NULL, &options, &local_err);
4111
    if (local_err) {
4112 4113 4114 4115
        error_propagate(errp, local_err);
        goto fail;
    }

4116
    visit_complete(v, &obj);
4117
    qdict = qobject_to(QDict, obj);
4118 4119 4120

    qdict_flatten(qdict);

4121 4122 4123 4124
    if (!qdict_get_try_str(qdict, "node-name")) {
        error_setg(errp, "'node-name' must be specified for the root node");
        goto fail;
    }
4125

4126 4127 4128
    bs = bds_tree_init(qdict, errp);
    if (!bs) {
        goto fail;
4129 4130
    }

4131 4132
    QTAILQ_INSERT_TAIL(&monitor_bdrv_states, bs, monitor_list);

4133
fail:
4134
    visit_free(v);
4135 4136
}

4137
void qmp_blockdev_del(const char *node_name, Error **errp)
4138 4139 4140 4141
{
    AioContext *aio_context;
    BlockDriverState *bs;

4142 4143 4144
    bs = bdrv_find_node(node_name);
    if (!bs) {
        error_setg(errp, "Cannot find node %s", node_name);
4145 4146
        return;
    }
4147 4148 4149
    if (bdrv_has_blk(bs)) {
        error_setg(errp, "Node %s is in use", node_name);
        return;
4150
    }
4151
    aio_context = bdrv_get_aio_context(bs);
4152 4153
    aio_context_acquire(aio_context);

4154 4155 4156
    if (bdrv_op_is_blocked(bs, BLOCK_OP_TYPE_DRIVE_DEL, errp)) {
        goto out;
    }
4157

4158 4159 4160 4161
    if (!bs->monitor_list.tqe_prev) {
        error_setg(errp, "Node %s is not owned by the monitor",
                   bs->node_name);
        goto out;
4162 4163
    }

4164 4165 4166 4167
    if (bs->refcnt > 1) {
        error_setg(errp, "Block device %s is in use",
                   bdrv_get_device_or_node_name(bs));
        goto out;
4168 4169
    }

4170 4171 4172
    QTAILQ_REMOVE(&monitor_bdrv_states, bs, monitor_list);
    bdrv_unref(bs);

4173 4174 4175 4176
out:
    aio_context_release(aio_context);
}

4177 4178 4179 4180 4181 4182 4183 4184 4185 4186 4187 4188 4189 4190 4191 4192 4193 4194 4195 4196 4197 4198 4199 4200 4201 4202 4203 4204 4205 4206 4207 4208 4209 4210 4211 4212 4213 4214 4215 4216 4217 4218 4219 4220 4221 4222 4223 4224 4225 4226 4227 4228 4229 4230 4231
static BdrvChild *bdrv_find_child(BlockDriverState *parent_bs,
                                  const char *child_name)
{
    BdrvChild *child;

    QLIST_FOREACH(child, &parent_bs->children, next) {
        if (strcmp(child->name, child_name) == 0) {
            return child;
        }
    }

    return NULL;
}

void qmp_x_blockdev_change(const char *parent, bool has_child,
                           const char *child, bool has_node,
                           const char *node, Error **errp)
{
    BlockDriverState *parent_bs, *new_bs = NULL;
    BdrvChild *p_child;

    parent_bs = bdrv_lookup_bs(parent, parent, errp);
    if (!parent_bs) {
        return;
    }

    if (has_child == has_node) {
        if (has_child) {
            error_setg(errp, "The parameters child and node are in conflict");
        } else {
            error_setg(errp, "Either child or node must be specified");
        }
        return;
    }

    if (has_child) {
        p_child = bdrv_find_child(parent_bs, child);
        if (!p_child) {
            error_setg(errp, "Node '%s' does not have child '%s'",
                       parent, child);
            return;
        }
        bdrv_del_child(parent_bs, p_child, errp);
    }

    if (has_node) {
        new_bs = bdrv_find_node(node);
        if (!new_bs) {
            error_setg(errp, "Node '%s' not found", node);
            return;
        }
        bdrv_add_child(parent_bs, new_bs, errp);
    }
}

4232
BlockJobInfoList *qmp_query_block_jobs(Error **errp)
4233
{
4234
    BlockJobInfoList *head = NULL, **p_next = &head;
4235
    BlockJob *job;
4236

4237
    for (job = block_job_next(NULL); job; job = block_job_next(job)) {
4238 4239
        BlockJobInfoList *elem;
        AioContext *aio_context;
4240

4241 4242 4243 4244 4245
        if (block_job_is_internal(job)) {
            continue;
        }
        elem = g_new0(BlockJobInfoList, 1);
        aio_context = blk_get_aio_context(job->blk);
4246
        aio_context_acquire(aio_context);
4247
        elem->value = block_job_query(job, errp);
4248
        aio_context_release(aio_context);
4249 4250 4251 4252 4253
        if (!elem->value) {
            g_free(elem);
            qapi_free_BlockJobInfoList(head);
            return NULL;
        }
4254 4255
        *p_next = elem;
        p_next = &elem->next;
4256 4257
    }

4258
    return head;
4259
}
4260

4261
void qmp_x_blockdev_set_iothread(const char *node_name, StrOrNull *iothread,
4262
                                 bool has_force, bool force, Error **errp)
4263 4264 4265 4266 4267 4268 4269 4270 4271 4272 4273
{
    AioContext *old_context;
    AioContext *new_context;
    BlockDriverState *bs;

    bs = bdrv_find_node(node_name);
    if (!bs) {
        error_setg(errp, "Cannot find node %s", node_name);
        return;
    }

4274 4275 4276 4277 4278
    /* Protects against accidents. */
    if (!(has_force && force) && bdrv_has_blk(bs)) {
        error_setg(errp, "Node %s is associated with a BlockBackend and could "
                         "be in use (use force=true to override this check)",
                         node_name);
4279 4280 4281 4282 4283 4284 4285 4286 4287 4288 4289 4290 4291 4292 4293 4294 4295 4296 4297 4298 4299 4300 4301
        return;
    }

    if (iothread->type == QTYPE_QSTRING) {
        IOThread *obj = iothread_by_id(iothread->u.s);
        if (!obj) {
            error_setg(errp, "Cannot find iothread %s", iothread->u.s);
            return;
        }

        new_context = iothread_get_aio_context(obj);
    } else {
        new_context = qemu_get_aio_context();
    }

    old_context = bdrv_get_aio_context(bs);
    aio_context_acquire(old_context);

    bdrv_set_aio_context(bs, new_context);

    aio_context_release(old_context);
}

4302 4303 4304 4305 4306 4307 4308 4309 4310 4311 4312 4313 4314 4315 4316 4317 4318 4319 4320 4321 4322 4323 4324 4325 4326 4327 4328 4329 4330 4331 4332 4333 4334 4335 4336 4337 4338 4339 4340 4341 4342 4343 4344
void qmp_x_block_latency_histogram_set(
    const char *device,
    bool has_boundaries, uint64List *boundaries,
    bool has_boundaries_read, uint64List *boundaries_read,
    bool has_boundaries_write, uint64List *boundaries_write,
    bool has_boundaries_flush, uint64List *boundaries_flush,
    Error **errp)
{
    BlockBackend *blk = blk_by_name(device);
    BlockAcctStats *stats;

    if (!blk) {
        error_setg(errp, "Device '%s' not found", device);
        return;
    }
    stats = blk_get_stats(blk);

    if (!has_boundaries && !has_boundaries_read && !has_boundaries_write &&
        !has_boundaries_flush)
    {
        block_latency_histograms_clear(stats);
        return;
    }

    if (has_boundaries || has_boundaries_read) {
        block_latency_histogram_set(
            stats, BLOCK_ACCT_READ,
            has_boundaries_read ? boundaries_read : boundaries);
    }

    if (has_boundaries || has_boundaries_write) {
        block_latency_histogram_set(
            stats, BLOCK_ACCT_WRITE,
            has_boundaries_write ? boundaries_write : boundaries);
    }

    if (has_boundaries || has_boundaries_flush) {
        block_latency_histogram_set(
            stats, BLOCK_ACCT_FLUSH,
            has_boundaries_flush ? boundaries_flush : boundaries);
    }
}

4345
QemuOptsList qemu_common_drive_opts = {
4346
    .name = "drive",
4347
    .head = QTAILQ_HEAD_INITIALIZER(qemu_common_drive_opts.head),
4348 4349 4350 4351 4352 4353 4354 4355 4356
    .desc = {
        {
            .name = "snapshot",
            .type = QEMU_OPT_BOOL,
            .help = "enable/disable snapshot mode",
        },{
            .name = "aio",
            .type = QEMU_OPT_STRING,
            .help = "host AIO implementation (threads, native)",
4357 4358 4359 4360
        },{
            .name = BDRV_OPT_CACHE_WB,
            .type = QEMU_OPT_BOOL,
            .help = "Enable writeback mode",
4361 4362 4363 4364 4365 4366 4367 4368 4369 4370 4371 4372 4373
        },{
            .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",
        },{
4374
            .name = BDRV_OPT_READ_ONLY,
4375 4376
            .type = QEMU_OPT_BOOL,
            .help = "open drive file as read-only",
4377 4378 4379 4380 4381
        },

        THROTTLE_OPTS,

        {
4382 4383 4384
            .name = "throttling.group",
            .type = QEMU_OPT_STRING,
            .help = "name of the block throttling group",
4385 4386 4387 4388
        },{
            .name = "copy-on-read",
            .type = QEMU_OPT_BOOL,
            .help = "copy read data from backing file into image file",
4389 4390 4391 4392
        },{
            .name = "detect-zeroes",
            .type = QEMU_OPT_STRING,
            .help = "try to optimize zero writes (off, on, unmap)",
4393 4394 4395 4396 4397 4398 4399 4400 4401 4402
        },{
            .name = "stats-account-invalid",
            .type = QEMU_OPT_BOOL,
            .help = "whether to account for invalid I/O operations "
                    "in the statistics",
        },{
            .name = "stats-account-failed",
            .type = QEMU_OPT_BOOL,
            .help = "whether to account for failed I/O operations "
                    "in the statistics",
4403 4404 4405 4406
        },
        { /* end of list */ }
    },
};
4407 4408 4409 4410 4411

QemuOptsList qemu_drive_opts = {
    .name = "drive",
    .head = QTAILQ_HEAD_INITIALIZER(qemu_drive_opts.head),
    .desc = {
4412 4413 4414 4415
        /*
         * no elements => accept any params
         * validation will happen later
         */
4416 4417 4418
        { /* end of list */ }
    },
};
新手
引导
客服 返回
顶部