blockdev.c 131.8 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)",
733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748
        },{
            .name = "cyls",
            .type = QEMU_OPT_NUMBER,
            .help = "number of cylinders (ide disk geometry)",
        },{
            .name = "heads",
            .type = QEMU_OPT_NUMBER,
            .help = "number of heads (ide disk geometry)",
        },{
            .name = "secs",
            .type = QEMU_OPT_NUMBER,
            .help = "number of sectors (ide disk geometry)",
        },{
            .name = "trans",
            .type = QEMU_OPT_STRING,
            .help = "chs translation (auto, lba, none)",
749 750 751 752
        },{
            .name = "addr",
            .type = QEMU_OPT_STRING,
            .help = "pci address (virtio only)",
753 754 755 756
        },{
            .name = "serial",
            .type = QEMU_OPT_STRING,
            .help = "disk serial number",
757 758 759 760
        },{
            .name = "file",
            .type = QEMU_OPT_STRING,
            .help = "file name",
761
        },
762 763 764

        /* Options that are passed on, but have special semantics with -drive */
        {
765
            .name = BDRV_OPT_READ_ONLY,
766 767
            .type = QEMU_OPT_BOOL,
            .help = "open drive file as read-only",
768 769 770 771 772 773 774 775
        },{
            .name = "rerror",
            .type = QEMU_OPT_STRING,
            .help = "read error action",
        },{
            .name = "werror",
            .type = QEMU_OPT_STRING,
            .help = "write error action",
776 777 778 779 780 781
        },{
            .name = "copy-on-read",
            .type = QEMU_OPT_BOOL,
            .help = "copy read data from backing file into image file",
        },

782 783 784 785
        { /* end of list */ }
    },
};

786
DriveInfo *drive_new(QemuOpts *all_opts, BlockInterfaceType block_default_type)
787
{
788
    const char *value;
789
    BlockBackend *blk;
790
    DriveInfo *dinfo = NULL;
791
    QDict *bs_opts;
792 793
    QemuOpts *legacy_opts;
    DriveMediaType media = MEDIA_DISK;
794
    BlockInterfaceType type;
795
    int cyls, heads, secs, translation;
796
    int max_devs, bus_id, unit_id, index;
797
    const char *devaddr;
798
    const char *werror, *rerror;
799 800
    bool read_only = false;
    bool copy_on_read;
801
    const char *serial;
802
    const char *filename;
803
    Error *local_err = NULL;
804
    int i;
805
    const char *deprecated[] = {
806
        "serial", "trans", "secs", "heads", "cyls", "addr"
807
    };
808

809
    /* Change legacy command line options into QMP ones */
810 811 812 813 814 815 816
    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" },
817

818 819 820
        { "bps",            "throttling.bps-total" },
        { "bps_rd",         "throttling.bps-read" },
        { "bps_wr",         "throttling.bps-write" },
821

822 823 824
        { "iops_max",       "throttling.iops-total-max" },
        { "iops_rd_max",    "throttling.iops-read-max" },
        { "iops_wr_max",    "throttling.iops-write-max" },
825

826 827 828
        { "bps_max",        "throttling.bps-total-max" },
        { "bps_rd_max",     "throttling.bps-read-max" },
        { "bps_wr_max",     "throttling.bps-write-max" },
829

830
        { "iops_size",      "throttling.iops-size" },
831

832 833
        { "group",          "throttling.group" },

834
        { "readonly",       BDRV_OPT_READ_ONLY },
835 836 837
    };

    for (i = 0; i < ARRAY_SIZE(opt_renames); i++) {
838 839 840
        qemu_opt_rename(all_opts, opt_renames[i].from, opt_renames[i].to,
                        &local_err);
        if (local_err) {
841
            error_report_err(local_err);
842 843
            return NULL;
        }
844
    }
845

846 847 848
    value = qemu_opt_get(all_opts, "cache");
    if (value) {
        int flags = 0;
849
        bool writethrough;
850

851
        if (bdrv_parse_cache_mode(value, &flags, &writethrough) != 0) {
852 853 854 855 856
            error_report("invalid cache option");
            return NULL;
        }

        /* Specific options take precedence */
857 858
        if (!qemu_opt_get(all_opts, BDRV_OPT_CACHE_WB)) {
            qemu_opt_set_bool(all_opts, BDRV_OPT_CACHE_WB,
859
                              !writethrough, &error_abort);
860
        }
861 862
        if (!qemu_opt_get(all_opts, BDRV_OPT_CACHE_DIRECT)) {
            qemu_opt_set_bool(all_opts, BDRV_OPT_CACHE_DIRECT,
863
                              !!(flags & BDRV_O_NOCACHE), &error_abort);
864
        }
865 866
        if (!qemu_opt_get(all_opts, BDRV_OPT_CACHE_NO_FLUSH)) {
            qemu_opt_set_bool(all_opts, BDRV_OPT_CACHE_NO_FLUSH,
867
                              !!(flags & BDRV_O_NO_FLUSH), &error_abort);
868 869 870 871
        }
        qemu_opt_unset(all_opts, "cache");
    }

872 873 874 875
    /* Get a QDict for processing the options */
    bs_opts = qdict_new();
    qemu_opts_to_qdict(all_opts, bs_opts);

876 877
    legacy_opts = qemu_opts_create(&qemu_legacy_drive_opts, NULL, 0,
                                   &error_abort);
878
    qemu_opts_absorb_qdict(legacy_opts, bs_opts, &local_err);
879
    if (local_err) {
880
        error_report_err(local_err);
881 882 883
        goto fail;
    }

884 885 886 887 888 889 890 891 892 893
    /* Other deprecated options */
    if (!qtest_enabled()) {
        for (i = 0; i < ARRAY_SIZE(deprecated); i++) {
            if (qemu_opt_get(legacy_opts, deprecated[i]) != NULL) {
                error_report("'%s' is deprecated, please use the corresponding "
                             "option of '-device' instead", deprecated[i]);
            }
        }
    }

894 895 896 897 898 899 900
    /* 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;
901
            read_only = true;
902 903 904 905 906 907
        } else {
            error_report("'%s' invalid media", value);
            goto fail;
        }
    }

908
    /* copy-on-read is disabled with a warning for read-only devices */
909
    read_only |= qemu_opt_get_bool(legacy_opts, BDRV_OPT_READ_ONLY, false);
910 911 912
    copy_on_read = qemu_opt_get_bool(legacy_opts, "copy-on-read", false);

    if (read_only && copy_on_read) {
913
        warn_report("disabling copy-on-read on read-only drive");
914 915 916
        copy_on_read = false;
    }

917 918
    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");
919

920 921 922 923 924 925 926 927 928 929 930 931 932 933 934
    /* 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;
    }

935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985
    /* Geometry */
    cyls  = qemu_opt_get_number(legacy_opts, "cyls", 0);
    heads = qemu_opt_get_number(legacy_opts, "heads", 0);
    secs  = qemu_opt_get_number(legacy_opts, "secs", 0);

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

    translation = BIOS_ATA_TRANSLATION_AUTO;
    value = qemu_opt_get(legacy_opts, "trans");
    if (value != NULL) {
        if (!cyls) {
            error_report("'%s' trans must be used with cyls, heads and secs",
                         value);
            goto fail;
        }
        if (!strcmp(value, "none")) {
            translation = BIOS_ATA_TRANSLATION_NONE;
        } else if (!strcmp(value, "lba")) {
            translation = BIOS_ATA_TRANSLATION_LBA;
        } else if (!strcmp(value, "large")) {
            translation = BIOS_ATA_TRANSLATION_LARGE;
        } else if (!strcmp(value, "rechs")) {
            translation = BIOS_ATA_TRANSLATION_RECHS;
        } else if (!strcmp(value, "auto")) {
            translation = BIOS_ATA_TRANSLATION_AUTO;
        } else {
            error_report("'%s' invalid translation type", value);
            goto fail;
        }
    }

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

986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024
    /* 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;
    }

1025 1026 1027
    /* Serial number */
    serial = qemu_opt_get(legacy_opts, "serial");

1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041
    /* 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);
        }
1042
        qdict_put_str(bs_opts, "id", new_id);
1043 1044 1045
        g_free(new_id);
    }

1046
    /* Add virtio block device */
1047 1048 1049 1050 1051 1052
    devaddr = qemu_opt_get(legacy_opts, "addr");
    if (devaddr && type != IF_VIRTIO) {
        error_report("addr is not supported by this bus type");
        goto fail;
    }

1053 1054
    if (type == IF_VIRTIO) {
        QemuOpts *devopts;
1055 1056
        devopts = qemu_opts_create(qemu_find_opts("device"), NULL, 0,
                                   &error_abort);
1057
        if (arch_type == QEMU_ARCH_S390X) {
1058
            qemu_opt_set(devopts, "driver", "virtio-blk-ccw", &error_abort);
1059
        } else {
1060
            qemu_opt_set(devopts, "driver", "virtio-blk-pci", &error_abort);
1061
        }
1062 1063
        qemu_opt_set(devopts, "drive", qdict_get_str(bs_opts, "id"),
                     &error_abort);
1064 1065 1066
        if (devaddr) {
            qemu_opt_set(devopts, "addr", devaddr, &error_abort);
        }
1067 1068
    }

1069 1070
    filename = qemu_opt_get(legacy_opts, "file");

1071 1072 1073 1074 1075 1076 1077 1078
    /* 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;
        }
1079
        qdict_put_str(bs_opts, "werror", werror);
1080 1081 1082 1083 1084 1085 1086 1087 1088
    }

    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;
        }
1089
        qdict_put_str(bs_opts, "rerror", rerror);
1090 1091
    }

1092
    /* Actual block device init: Functionality shared with blockdev-add */
1093
    blk = blockdev_init(filename, bs_opts, &local_err);
1094
    bs_opts = NULL;
1095
    if (!blk) {
1096
        if (local_err) {
1097
            error_report_err(local_err);
1098
        }
1099
        goto fail;
1100
    } else {
1101
        assert(!local_err);
1102 1103
    }

1104 1105
    /* Create legacy DriveInfo */
    dinfo = g_malloc0(sizeof(*dinfo));
1106
    dinfo->opts = all_opts;
1107

1108 1109 1110 1111 1112
    dinfo->cyls = cyls;
    dinfo->heads = heads;
    dinfo->secs = secs;
    dinfo->trans = translation;

1113
    dinfo->type = type;
1114 1115
    dinfo->bus = bus_id;
    dinfo->unit = unit_id;
1116
    dinfo->devaddr = devaddr;
1117
    dinfo->serial = g_strdup(serial);
1118

1119 1120
    blk_set_legacy_dinfo(blk, dinfo);

1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131
    switch(type) {
    case IF_IDE:
    case IF_SCSI:
    case IF_XEN:
    case IF_NONE:
        dinfo->media_cd = media == MEDIA_CDROM;
        break;
    default:
        break;
    }

1132
fail:
1133
    qemu_opts_del(legacy_opts);
1134
    qobject_unref(bs_opts);
1135
    return dinfo;
1136 1137
}

1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159
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;
}

1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182
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;
}

1183
void hmp_commit(Monitor *mon, const QDict *qdict)
1184 1185
{
    const char *device = qdict_get_str(qdict, "device");
1186
    BlockBackend *blk;
1187
    int ret;
1188

1189
    if (!strcmp(device, "all")) {
1190
        ret = blk_commit_all();
1191
    } else {
1192 1193 1194
        BlockDriverState *bs;
        AioContext *aio_context;

1195 1196
        blk = blk_by_name(device);
        if (!blk) {
1197
            monitor_printf(mon, "Device '%s' not found\n", device);
1198
            return;
1199
        }
M
Max Reitz 已提交
1200 1201 1202 1203
        if (!blk_is_available(blk)) {
            monitor_printf(mon, "Device '%s' has no medium\n", device);
            return;
        }
1204 1205 1206 1207 1208 1209 1210 1211

        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);
1212 1213 1214 1215
    }
    if (ret < 0) {
        monitor_printf(mon, "'commit' error for '%s': %s\n", device,
                       strerror(-ret));
1216 1217 1218
    }
}

1219
static void blockdev_do_action(TransactionAction *action, Error **errp)
1220
{
1221
    TransactionActionList list;
1222

1223
    list.value = action;
1224
    list.next = NULL;
1225
    qmp_transaction(&list, false, NULL, errp);
1226 1227
}

1228 1229 1230 1231 1232
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,
1233
                                bool has_format, const char *format,
1234
                                bool has_mode, NewImageMode mode, Error **errp)
1235
{
1236
    BlockdevSnapshotSync snapshot = {
1237
        .has_device = has_device,
1238
        .device = (char *) device,
1239 1240
        .has_node_name = has_node_name,
        .node_name = (char *) node_name,
1241
        .snapshot_file = (char *) snapshot_file,
1242 1243
        .has_snapshot_node_name = has_snapshot_node_name,
        .snapshot_node_name = (char *) snapshot_node_name,
1244 1245 1246 1247 1248
        .has_format = has_format,
        .format = (char *) format,
        .has_mode = has_mode,
        .mode = mode,
    };
1249 1250
    TransactionAction action = {
        .type = TRANSACTION_ACTION_KIND_BLOCKDEV_SNAPSHOT_SYNC,
1251
        .u.blockdev_snapshot_sync.data = &snapshot,
1252 1253
    };
    blockdev_do_action(&action, errp);
1254 1255
}

1256 1257 1258 1259 1260 1261 1262
void qmp_blockdev_snapshot(const char *node, const char *overlay,
                           Error **errp)
{
    BlockdevSnapshot snapshot_data = {
        .node = (char *) node,
        .overlay = (char *) overlay
    };
1263 1264
    TransactionAction action = {
        .type = TRANSACTION_ACTION_KIND_BLOCKDEV_SNAPSHOT,
1265
        .u.blockdev_snapshot.data = &snapshot_data,
1266 1267
    };
    blockdev_do_action(&action, errp);
1268 1269
}

1270 1271 1272 1273 1274 1275 1276 1277
void qmp_blockdev_snapshot_internal_sync(const char *device,
                                         const char *name,
                                         Error **errp)
{
    BlockdevSnapshotInternal snapshot = {
        .device = (char *) device,
        .name = (char *) name
    };
1278 1279
    TransactionAction action = {
        .type = TRANSACTION_ACTION_KIND_BLOCKDEV_SNAPSHOT_INTERNAL_SYNC,
1280
        .u.blockdev_snapshot_internal_sync.data = &snapshot,
1281 1282
    };
    blockdev_do_action(&action, errp);
1283 1284
}

1285 1286 1287 1288 1289 1290 1291
SnapshotInfo *qmp_blockdev_snapshot_delete_internal_sync(const char *device,
                                                         bool has_id,
                                                         const char *id,
                                                         bool has_name,
                                                         const char *name,
                                                         Error **errp)
{
1292
    BlockDriverState *bs;
1293
    AioContext *aio_context;
1294 1295 1296 1297 1298
    QEMUSnapshotInfo sn;
    Error *local_err = NULL;
    SnapshotInfo *info = NULL;
    int ret;

1299 1300
    bs = qmp_get_root_bs(device, errp);
    if (!bs) {
1301 1302
        return NULL;
    }
1303
    aio_context = bdrv_get_aio_context(bs);
M
Max Reitz 已提交
1304
    aio_context_acquire(aio_context);
1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315

    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
1316
        goto out_aio_context;
1317 1318
    }

1319 1320 1321 1322
    if (bdrv_op_is_blocked(bs, BLOCK_OP_TYPE_INTERNAL_SNAPSHOT_DELETE, errp)) {
        goto out_aio_context;
    }

1323
    ret = bdrv_snapshot_find_by_id_and_name(bs, id, name, &sn, &local_err);
1324
    if (local_err) {
1325
        error_propagate(errp, local_err);
1326
        goto out_aio_context;
1327 1328 1329 1330 1331 1332
    }
    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);
1333
        goto out_aio_context;
1334 1335 1336
    }

    bdrv_snapshot_delete(bs, id, name, &local_err);
1337
    if (local_err) {
1338
        error_propagate(errp, local_err);
1339
        goto out_aio_context;
1340 1341
    }

1342 1343
    aio_context_release(aio_context);

1344
    info = g_new0(SnapshotInfo, 1);
1345 1346 1347 1348 1349 1350 1351 1352 1353
    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;
1354 1355 1356 1357

out_aio_context:
    aio_context_release(aio_context);
    return NULL;
1358
}
1359

1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397
/**
 * 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);
1398
        return NULL;
1399 1400 1401 1402 1403 1404 1405 1406 1407
    }

    if (pbs) {
        *pbs = bs;
    }

    return bitmap;
}

1408
/* New and old BlockDriverState structs for atomic group operations */
1409

1410
typedef struct BlkActionState BlkActionState;
1411

1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426
/**
 * 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 {
1427
    size_t instance_size;
1428 1429 1430 1431 1432
    void (*prepare)(BlkActionState *common, Error **errp);
    void (*commit)(BlkActionState *common);
    void (*abort)(BlkActionState *common);
    void (*clean)(BlkActionState *common);
} BlkActionOps;
1433

1434 1435 1436 1437 1438 1439
/**
 * 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.
1440
 * @block_job_txn: Transaction which this action belongs to.
1441 1442 1443 1444 1445
 * @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.
1446
 */
1447
struct BlkActionState {
1448
    TransactionAction *action;
1449
    const BlkActionOps *ops;
1450
    JobTxn *block_job_txn;
1451
    TransactionProperties *txn_props;
1452
    QSIMPLEQ_ENTRY(BlkActionState) entry;
1453 1454
};

1455 1456
/* internal snapshot private data */
typedef struct InternalSnapshotState {
1457
    BlkActionState common;
1458 1459
    BlockDriverState *bs;
    QEMUSnapshotInfo sn;
1460
    bool created;
1461 1462
} InternalSnapshotState;

1463 1464 1465 1466 1467 1468 1469

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",
1470 1471
                   TransactionActionKind_str(s->action->type),
                   ActionCompletionMode_str(s->txn_props->completion_mode));
1472 1473 1474 1475 1476
        return -1;
    }
    return 0;
}

1477
static void internal_snapshot_prepare(BlkActionState *common,
1478 1479
                                      Error **errp)
{
1480
    Error *local_err = NULL;
1481 1482 1483 1484 1485 1486 1487 1488
    const char *device;
    const char *name;
    BlockDriverState *bs;
    QEMUSnapshotInfo old_sn, *sn;
    bool ret;
    qemu_timeval tv;
    BlockdevSnapshotInternal *internal;
    InternalSnapshotState *state;
1489
    AioContext *aio_context;
1490 1491
    int ret1;

1492
    g_assert(common->action->type ==
1493
             TRANSACTION_ACTION_KIND_BLOCKDEV_SNAPSHOT_INTERNAL_SYNC);
1494
    internal = common->action->u.blockdev_snapshot_internal_sync.data;
1495 1496 1497 1498 1499 1500 1501
    state = DO_UPCAST(InternalSnapshotState, common, common);

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

    /* 2. check for validation */
1502 1503 1504 1505
    if (action_check_completion_mode(common, errp) < 0) {
        return;
    }

1506 1507
    bs = qmp_get_root_bs(device, errp);
    if (!bs) {
1508 1509 1510
        return;
    }

1511 1512
    aio_context = bdrv_get_aio_context(bs);
    aio_context_acquire(aio_context);
1513

1514
    state->bs = bs;
1515 1516

    /* Paired with .clean() */
1517 1518
    bdrv_drained_begin(bs);

1519
    if (bdrv_op_is_blocked(bs, BLOCK_OP_TYPE_INTERNAL_SNAPSHOT, errp)) {
1520
        goto out;
1521 1522
    }

1523
    if (bdrv_is_read_only(bs)) {
1524
        error_setg(errp, "Device '%s' is read only", device);
1525
        goto out;
1526 1527 1528
    }

    if (!bdrv_can_snapshot(bs)) {
1529 1530 1531
        error_setg(errp, "Block format '%s' used by device '%s' "
                   "does not support internal snapshots",
                   bs->drv->format_name, device);
1532
        goto out;
1533 1534 1535 1536
    }

    if (!strlen(name)) {
        error_setg(errp, "Name is empty");
1537
        goto out;
1538 1539 1540
    }

    /* check whether a snapshot with name exist */
1541 1542 1543 1544
    ret = bdrv_snapshot_find_by_id_and_name(bs, NULL, name, &old_sn,
                                            &local_err);
    if (local_err) {
        error_propagate(errp, local_err);
1545
        goto out;
1546 1547 1548 1549
    } else if (ret) {
        error_setg(errp,
                   "Snapshot with name '%s' already exists on device '%s'",
                   name, device);
1550
        goto out;
1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565
    }

    /* 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);
1566
        goto out;
1567 1568 1569
    }

    /* 4. succeed, mark a snapshot is created */
1570
    state->created = true;
1571 1572 1573

out:
    aio_context_release(aio_context);
1574 1575
}

1576
static void internal_snapshot_abort(BlkActionState *common)
1577 1578 1579 1580 1581
{
    InternalSnapshotState *state =
                             DO_UPCAST(InternalSnapshotState, common, common);
    BlockDriverState *bs = state->bs;
    QEMUSnapshotInfo *sn = &state->sn;
1582
    AioContext *aio_context;
1583 1584
    Error *local_error = NULL;

1585
    if (!state->created) {
1586 1587 1588
        return;
    }

1589 1590 1591
    aio_context = bdrv_get_aio_context(state->bs);
    aio_context_acquire(aio_context);

1592
    if (bdrv_snapshot_delete(bs, sn->id_str, sn->name, &local_error) < 0) {
1593 1594 1595 1596 1597
        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));
1598
    }
1599 1600

    aio_context_release(aio_context);
1601 1602
}

1603
static void internal_snapshot_clean(BlkActionState *common)
1604 1605 1606
{
    InternalSnapshotState *state = DO_UPCAST(InternalSnapshotState,
                                             common, common);
1607
    AioContext *aio_context;
1608

1609 1610
    if (!state->bs) {
        return;
1611
    }
1612 1613 1614 1615 1616 1617 1618

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

    bdrv_drained_end(state->bs);

    aio_context_release(aio_context);
1619 1620
}

1621
/* external snapshot private data */
1622
typedef struct ExternalSnapshotState {
1623
    BlkActionState common;
1624 1625
    BlockDriverState *old_bs;
    BlockDriverState *new_bs;
1626
    bool overlay_appended;
1627
} ExternalSnapshotState;
1628

1629
static void external_snapshot_prepare(BlkActionState *common,
1630 1631
                                      Error **errp)
{
1632
    int flags = 0;
1633
    QDict *options = NULL;
1634
    Error *local_err = NULL;
1635
    /* Device and node name of the image to generate the snapshot from */
1636
    const char *device;
1637
    const char *node_name;
1638 1639 1640
    /* Reference to the new image (for 'blockdev-snapshot') */
    const char *snapshot_ref;
    /* File name of the new image (for 'blockdev-snapshot-sync') */
1641
    const char *new_image_file;
1642 1643
    ExternalSnapshotState *state =
                             DO_UPCAST(ExternalSnapshotState, common, common);
1644
    TransactionAction *action = common->action;
1645
    AioContext *aio_context;
1646

1647 1648 1649 1650 1651
    /* 'blockdev-snapshot' and 'blockdev-snapshot-sync' have similar
     * purpose but a different set of parameters */
    switch (action->type) {
    case TRANSACTION_ACTION_KIND_BLOCKDEV_SNAPSHOT:
        {
1652
            BlockdevSnapshot *s = action->u.blockdev_snapshot.data;
1653 1654 1655 1656 1657 1658 1659 1660
            device = s->node;
            node_name = s->node;
            new_image_file = NULL;
            snapshot_ref = s->overlay;
        }
        break;
    case TRANSACTION_ACTION_KIND_BLOCKDEV_SNAPSHOT_SYNC:
        {
1661
            BlockdevSnapshotSync *s = action->u.blockdev_snapshot_sync.data;
1662 1663 1664 1665 1666 1667 1668 1669
            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();
1670 1671 1672
    }

    /* start processing */
1673 1674 1675 1676
    if (action_check_completion_mode(common, errp) < 0) {
        return;
    }

1677 1678
    state->old_bs = bdrv_lookup_bs(device, node_name, errp);
    if (!state->old_bs) {
1679 1680 1681
        return;
    }

1682 1683 1684 1685
    aio_context = bdrv_get_aio_context(state->old_bs);
    aio_context_acquire(aio_context);

    /* Paired with .clean() */
1686
    bdrv_drained_begin(state->old_bs);
1687

1688
    if (!bdrv_is_inserted(state->old_bs)) {
1689
        error_setg(errp, QERR_DEVICE_HAS_NO_MEDIUM, device);
1690
        goto out;
1691 1692
    }

1693 1694
    if (bdrv_op_is_blocked(state->old_bs,
                           BLOCK_OP_TYPE_EXTERNAL_SNAPSHOT, errp)) {
1695
        goto out;
1696 1697
    }

1698 1699
    if (!bdrv_is_read_only(state->old_bs)) {
        if (bdrv_flush(state->old_bs)) {
1700
            error_setg(errp, QERR_IO_ERROR);
1701
            goto out;
1702 1703 1704
        }
    }

1705
    if (!bdrv_is_first_non_filter(state->old_bs)) {
1706
        error_setg(errp, QERR_FEATURE_DISABLED, "snapshot");
1707
        goto out;
1708 1709
    }

1710
    if (action->type == TRANSACTION_ACTION_KIND_BLOCKDEV_SNAPSHOT_SYNC) {
1711
        BlockdevSnapshotSync *s = action->u.blockdev_snapshot_sync.data;
1712 1713 1714 1715
        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;
1716

1717 1718
        if (node_name && !snapshot_node_name) {
            error_setg(errp, "New snapshot node name missing");
1719
            goto out;
1720 1721
        }

1722 1723 1724
        if (snapshot_node_name &&
            bdrv_lookup_bs(snapshot_node_name, snapshot_node_name, NULL)) {
            error_setg(errp, "New snapshot node name already in use");
1725
            goto out;
1726 1727 1728
        }

        flags = state->old_bs->open_flags;
1729 1730
        flags &= ~(BDRV_O_SNAPSHOT | BDRV_O_COPY_ON_READ);
        flags |= BDRV_O_NO_BACKING;
1731 1732 1733 1734

        /* create new image w/backing file */
        mode = s->has_mode ? s->mode : NEW_IMAGE_MODE_ABSOLUTE_PATHS;
        if (mode != NEW_IMAGE_MODE_EXISTING) {
1735 1736 1737
            int64_t size = bdrv_getlength(state->old_bs);
            if (size < 0) {
                error_setg_errno(errp, -size, "bdrv_getlength failed");
1738
                goto out;
1739
            }
1740 1741 1742
            bdrv_img_create(new_image_file, format,
                            state->old_bs->filename,
                            state->old_bs->drv->format_name,
1743
                            NULL, size, flags, false, &local_err);
1744 1745
            if (local_err) {
                error_propagate(errp, local_err);
1746
                goto out;
1747 1748 1749 1750 1751
            }
        }

        options = qdict_new();
        if (s->has_snapshot_node_name) {
1752
            qdict_put_str(options, "node-name", snapshot_node_name);
1753
        }
1754
        qdict_put_str(options, "driver", format);
1755 1756
    }

1757 1758
    state->new_bs = bdrv_open(new_image_file, snapshot_ref, options, flags,
                              errp);
1759
    /* We will manually add the backing_hd field to the bs later */
1760
    if (!state->new_bs) {
1761
        goto out;
1762 1763
    }

1764
    if (bdrv_has_blk(state->new_bs)) {
1765
        error_setg(errp, "The snapshot is already in use");
1766
        goto out;
1767 1768 1769 1770
    }

    if (bdrv_op_is_blocked(state->new_bs, BLOCK_OP_TYPE_EXTERNAL_SNAPSHOT,
                           errp)) {
1771
        goto out;
1772 1773 1774 1775
    }

    if (state->new_bs->backing != NULL) {
        error_setg(errp, "The snapshot already has a backing image");
1776
        goto out;
1777 1778 1779 1780
    }

    if (!state->new_bs->drv->supports_backing) {
        error_setg(errp, "The snapshot does not support backing images");
1781
        goto out;
1782 1783
    }

1784
    bdrv_set_aio_context(state->new_bs, aio_context);
1785

1786 1787 1788 1789 1790 1791 1792
    /* 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);
1793
        goto out;
1794
    }
1795
    state->overlay_appended = true;
1796 1797 1798

out:
    aio_context_release(aio_context);
1799 1800
}

1801
static void external_snapshot_commit(BlkActionState *common)
1802
{
1803 1804
    ExternalSnapshotState *state =
                             DO_UPCAST(ExternalSnapshotState, common, common);
1805 1806 1807 1808
    AioContext *aio_context;

    aio_context = bdrv_get_aio_context(state->old_bs);
    aio_context_acquire(aio_context);
1809

1810 1811 1812
    /* 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 */
1813
    if (!atomic_read(&state->old_bs->copy_on_read)) {
1814 1815 1816
        bdrv_reopen(state->old_bs, state->old_bs->open_flags & ~BDRV_O_RDWR,
                    NULL);
    }
1817 1818

    aio_context_release(aio_context);
1819 1820
}

1821
static void external_snapshot_abort(BlkActionState *common)
1822
{
1823 1824 1825
    ExternalSnapshotState *state =
                             DO_UPCAST(ExternalSnapshotState, common, common);
    if (state->new_bs) {
1826
        if (state->overlay_appended) {
1827 1828 1829 1830 1831
            AioContext *aio_context;

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

1832 1833 1834
            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);
1835
            bdrv_replace_node(state->new_bs, state->old_bs, &error_abort);
1836
            bdrv_unref(state->old_bs); /* bdrv_replace_node() ref'ed old_bs */
1837 1838

            aio_context_release(aio_context);
1839
        }
1840
    }
1841 1842
}

1843
static void external_snapshot_clean(BlkActionState *common)
1844 1845 1846
{
    ExternalSnapshotState *state =
                             DO_UPCAST(ExternalSnapshotState, common, common);
1847 1848 1849 1850
    AioContext *aio_context;

    if (!state->old_bs) {
        return;
1851
    }
1852 1853 1854 1855 1856 1857 1858 1859

    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);
1860 1861
}

1862
typedef struct DriveBackupState {
1863
    BlkActionState common;
1864 1865 1866 1867
    BlockDriverState *bs;
    BlockJob *job;
} DriveBackupState;

1868
static BlockJob *do_drive_backup(DriveBackup *backup, JobTxn *txn,
1869
                            Error **errp);
1870

1871
static void drive_backup_prepare(BlkActionState *common, Error **errp)
1872 1873
{
    DriveBackupState *state = DO_UPCAST(DriveBackupState, common, common);
1874
    BlockDriverState *bs;
1875
    DriveBackup *backup;
1876
    AioContext *aio_context;
1877 1878
    Error *local_err = NULL;

1879
    assert(common->action->type == TRANSACTION_ACTION_KIND_DRIVE_BACKUP);
1880
    backup = common->action->u.drive_backup.data;
1881

1882 1883
    bs = qmp_get_root_bs(backup->device, errp);
    if (!bs) {
1884 1885 1886
        return;
    }

1887 1888 1889 1890
    aio_context = bdrv_get_aio_context(bs);
    aio_context_acquire(aio_context);

    /* Paired with .clean() */
1891
    bdrv_drained_begin(bs);
1892

1893
    state->bs = bs;
1894

1895
    state->job = do_drive_backup(backup, common->block_job_txn, &local_err);
1896
    if (local_err) {
1897
        error_propagate(errp, local_err);
1898
        goto out;
1899
    }
1900 1901 1902

out:
    aio_context_release(aio_context);
1903
}
1904

1905 1906 1907
static void drive_backup_commit(BlkActionState *common)
{
    DriveBackupState *state = DO_UPCAST(DriveBackupState, common, common);
1908 1909 1910 1911 1912
    AioContext *aio_context;

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

1913
    assert(state->job);
1914
    job_start(&state->job->job);
1915 1916

    aio_context_release(aio_context);
1917 1918
}

1919
static void drive_backup_abort(BlkActionState *common)
1920 1921 1922
{
    DriveBackupState *state = DO_UPCAST(DriveBackupState, common, common);

1923
    if (state->job) {
1924 1925 1926 1927 1928
        AioContext *aio_context;

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

1929
        job_cancel_sync(&state->job->job);
1930 1931

        aio_context_release(aio_context);
1932 1933 1934
    }
}

1935
static void drive_backup_clean(BlkActionState *common)
1936 1937
{
    DriveBackupState *state = DO_UPCAST(DriveBackupState, common, common);
1938
    AioContext *aio_context;
1939

1940 1941
    if (!state->bs) {
        return;
1942
    }
1943 1944 1945 1946 1947 1948 1949

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

    bdrv_drained_end(state->bs);

    aio_context_release(aio_context);
1950 1951
}

1952
typedef struct BlockdevBackupState {
1953
    BlkActionState common;
1954 1955 1956 1957
    BlockDriverState *bs;
    BlockJob *job;
} BlockdevBackupState;

1958
static BlockJob *do_blockdev_backup(BlockdevBackup *backup, JobTxn *txn,
1959
                                    Error **errp);
1960

1961
static void blockdev_backup_prepare(BlkActionState *common, Error **errp)
1962 1963 1964
{
    BlockdevBackupState *state = DO_UPCAST(BlockdevBackupState, common, common);
    BlockdevBackup *backup;
1965
    BlockDriverState *bs, *target;
1966
    AioContext *aio_context;
1967 1968
    Error *local_err = NULL;

1969
    assert(common->action->type == TRANSACTION_ACTION_KIND_BLOCKDEV_BACKUP);
1970
    backup = common->action->u.blockdev_backup.data;
1971

1972
    bs = bdrv_lookup_bs(backup->device, backup->device, errp);
1973
    if (!bs) {
1974 1975 1976
        return;
    }

1977
    target = bdrv_lookup_bs(backup->target, backup->target, errp);
M
Max Reitz 已提交
1978
    if (!target) {
1979 1980 1981
        return;
    }

1982 1983
    aio_context = bdrv_get_aio_context(bs);
    if (aio_context != bdrv_get_aio_context(target)) {
1984 1985 1986
        error_setg(errp, "Backup between two IO threads is not implemented");
        return;
    }
1987
    aio_context_acquire(aio_context);
1988
    state->bs = bs;
1989 1990

    /* Paired with .clean() */
1991
    bdrv_drained_begin(state->bs);
1992

1993
    state->job = do_blockdev_backup(backup, common->block_job_txn, &local_err);
1994 1995
    if (local_err) {
        error_propagate(errp, local_err);
1996
        goto out;
1997
    }
1998 1999 2000

out:
    aio_context_release(aio_context);
2001
}
2002

2003 2004 2005
static void blockdev_backup_commit(BlkActionState *common)
{
    BlockdevBackupState *state = DO_UPCAST(BlockdevBackupState, common, common);
2006 2007 2008 2009 2010
    AioContext *aio_context;

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

2011
    assert(state->job);
2012
    job_start(&state->job->job);
2013 2014

    aio_context_release(aio_context);
2015 2016
}

2017
static void blockdev_backup_abort(BlkActionState *common)
2018 2019 2020
{
    BlockdevBackupState *state = DO_UPCAST(BlockdevBackupState, common, common);

2021
    if (state->job) {
2022 2023 2024 2025 2026
        AioContext *aio_context;

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

2027
        job_cancel_sync(&state->job->job);
2028 2029

        aio_context_release(aio_context);
2030 2031 2032
    }
}

2033
static void blockdev_backup_clean(BlkActionState *common)
2034 2035
{
    BlockdevBackupState *state = DO_UPCAST(BlockdevBackupState, common, common);
2036
    AioContext *aio_context;
2037

2038 2039
    if (!state->bs) {
        return;
2040
    }
2041 2042 2043 2044 2045 2046 2047

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

    bdrv_drained_end(state->bs);

    aio_context_release(aio_context);
2048 2049
}

F
Fam Zheng 已提交
2050
typedef struct BlockDirtyBitmapState {
2051
    BlkActionState common;
F
Fam Zheng 已提交
2052 2053 2054 2055
    BdrvDirtyBitmap *bitmap;
    BlockDriverState *bs;
    HBitmap *backup;
    bool prepared;
2056
    bool was_enabled;
F
Fam Zheng 已提交
2057 2058
} BlockDirtyBitmapState;

2059
static void block_dirty_bitmap_add_prepare(BlkActionState *common,
F
Fam Zheng 已提交
2060 2061 2062 2063 2064 2065 2066
                                           Error **errp)
{
    Error *local_err = NULL;
    BlockDirtyBitmapAdd *action;
    BlockDirtyBitmapState *state = DO_UPCAST(BlockDirtyBitmapState,
                                             common, common);

2067 2068 2069 2070
    if (action_check_completion_mode(common, errp) < 0) {
        return;
    }

2071
    action = common->action->u.block_dirty_bitmap_add.data;
F
Fam Zheng 已提交
2072 2073 2074
    /* 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,
2075
                               action->has_persistent, action->persistent,
2076
                               action->has_autoload, action->autoload,
2077
                               action->has_x_disabled, action->x_disabled,
F
Fam Zheng 已提交
2078 2079 2080 2081 2082 2083 2084 2085 2086
                               &local_err);

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

2087
static void block_dirty_bitmap_add_abort(BlkActionState *common)
F
Fam Zheng 已提交
2088 2089 2090 2091 2092
{
    BlockDirtyBitmapAdd *action;
    BlockDirtyBitmapState *state = DO_UPCAST(BlockDirtyBitmapState,
                                             common, common);

2093
    action = common->action->u.block_dirty_bitmap_add.data;
F
Fam Zheng 已提交
2094 2095 2096 2097 2098 2099 2100 2101
    /* 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);
    }
}

2102
static void block_dirty_bitmap_clear_prepare(BlkActionState *common,
F
Fam Zheng 已提交
2103 2104 2105 2106 2107 2108
                                             Error **errp)
{
    BlockDirtyBitmapState *state = DO_UPCAST(BlockDirtyBitmapState,
                                             common, common);
    BlockDirtyBitmap *action;

2109 2110 2111 2112
    if (action_check_completion_mode(common, errp) < 0) {
        return;
    }

2113
    action = common->action->u.block_dirty_bitmap_clear.data;
F
Fam Zheng 已提交
2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124
    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;
2125 2126 2127
    } else if (bdrv_dirty_bitmap_qmp_locked(state->bitmap)) {
        error_setg(errp, "Cannot modify a locked bitmap");
        return;
F
Fam Zheng 已提交
2128 2129 2130
    } else if (!bdrv_dirty_bitmap_enabled(state->bitmap)) {
        error_setg(errp, "Cannot clear a disabled bitmap");
        return;
2131 2132 2133
    } else if (bdrv_dirty_bitmap_readonly(state->bitmap)) {
        error_setg(errp, "Cannot clear a readonly bitmap");
        return;
F
Fam Zheng 已提交
2134 2135 2136 2137 2138
    }

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

2139
static void block_dirty_bitmap_clear_abort(BlkActionState *common)
F
Fam Zheng 已提交
2140 2141 2142 2143
{
    BlockDirtyBitmapState *state = DO_UPCAST(BlockDirtyBitmapState,
                                             common, common);

2144 2145 2146
    if (state->backup) {
        bdrv_undo_clear_dirty_bitmap(state->bitmap, state->backup);
    }
F
Fam Zheng 已提交
2147 2148
}

2149
static void block_dirty_bitmap_clear_commit(BlkActionState *common)
F
Fam Zheng 已提交
2150 2151 2152 2153 2154 2155 2156
{
    BlockDirtyBitmapState *state = DO_UPCAST(BlockDirtyBitmapState,
                                             common, common);

    hbitmap_free(state->backup);
}

2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 2224
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);
    }
}

2225
static void abort_prepare(BlkActionState *common, Error **errp)
2226 2227 2228 2229
{
    error_setg(errp, "Transaction aborted using Abort action");
}

2230
static void abort_commit(BlkActionState *common)
2231
{
2232
    g_assert_not_reached(); /* this action never succeeds */
2233 2234
}

2235
static const BlkActionOps actions[] = {
2236 2237 2238 2239 2240
    [TRANSACTION_ACTION_KIND_BLOCKDEV_SNAPSHOT] = {
        .instance_size = sizeof(ExternalSnapshotState),
        .prepare  = external_snapshot_prepare,
        .commit   = external_snapshot_commit,
        .abort = external_snapshot_abort,
2241
        .clean = external_snapshot_clean,
2242
    },
2243
    [TRANSACTION_ACTION_KIND_BLOCKDEV_SNAPSHOT_SYNC] = {
2244
        .instance_size = sizeof(ExternalSnapshotState),
2245 2246 2247
        .prepare  = external_snapshot_prepare,
        .commit   = external_snapshot_commit,
        .abort = external_snapshot_abort,
2248
        .clean = external_snapshot_clean,
2249
    },
2250 2251 2252
    [TRANSACTION_ACTION_KIND_DRIVE_BACKUP] = {
        .instance_size = sizeof(DriveBackupState),
        .prepare = drive_backup_prepare,
2253
        .commit = drive_backup_commit,
2254
        .abort = drive_backup_abort,
2255
        .clean = drive_backup_clean,
2256
    },
2257 2258 2259
    [TRANSACTION_ACTION_KIND_BLOCKDEV_BACKUP] = {
        .instance_size = sizeof(BlockdevBackupState),
        .prepare = blockdev_backup_prepare,
2260
        .commit = blockdev_backup_commit,
2261 2262 2263
        .abort = blockdev_backup_abort,
        .clean = blockdev_backup_clean,
    },
2264
    [TRANSACTION_ACTION_KIND_ABORT] = {
2265
        .instance_size = sizeof(BlkActionState),
2266 2267 2268
        .prepare = abort_prepare,
        .commit = abort_commit,
    },
2269 2270 2271 2272
    [TRANSACTION_ACTION_KIND_BLOCKDEV_SNAPSHOT_INTERNAL_SYNC] = {
        .instance_size = sizeof(InternalSnapshotState),
        .prepare  = internal_snapshot_prepare,
        .abort = internal_snapshot_abort,
2273
        .clean = internal_snapshot_clean,
2274
    },
F
Fam Zheng 已提交
2275 2276 2277 2278 2279 2280 2281 2282 2283 2284
    [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,
2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295
    },
    [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,
     }
2296 2297
};

2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316
/**
 * 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;
}

2317
/*
2318 2319
 * 'Atomic' group operations.  The operations are performed as a set, and if
 * any fail then we roll back all operations in the group.
2320
 */
2321 2322 2323 2324
void qmp_transaction(TransactionActionList *dev_list,
                     bool has_props,
                     struct TransactionProperties *props,
                     Error **errp)
2325
{
2326
    TransactionActionList *dev_entry = dev_list;
2327
    JobTxn *block_job_txn = NULL;
2328
    BlkActionState *state, *next;
2329
    Error *local_err = NULL;
2330

2331
    QSIMPLEQ_HEAD(snap_bdrv_states, BlkActionState) snap_bdrv_states;
2332 2333
    QSIMPLEQ_INIT(&snap_bdrv_states);

2334
    /* Does this transaction get canceled as a group on failure?
2335
     * If not, we don't really need to make a JobTxn.
2336 2337 2338
     */
    props = get_transaction_properties(props);
    if (props->completion_mode != ACTION_COMPLETION_MODE_INDIVIDUAL) {
2339
        block_job_txn = job_txn_new();
2340 2341
    }

2342
    /* drain all i/o before any operations */
2343 2344
    bdrv_drain_all();

2345
    /* We don't do anything in this loop that commits us to the operations */
2346
    while (NULL != dev_entry) {
2347
        TransactionAction *dev_info = NULL;
2348
        const BlkActionOps *ops;
2349

2350 2351 2352
        dev_info = dev_entry->value;
        dev_entry = dev_entry->next;

2353
        assert(dev_info->type < ARRAY_SIZE(actions));
2354

2355
        ops = &actions[dev_info->type];
2356 2357
        assert(ops->instance_size > 0);

2358 2359 2360
        state = g_malloc0(ops->instance_size);
        state->ops = ops;
        state->action = dev_info;
2361 2362
        state->block_job_txn = block_job_txn;
        state->txn_props = props;
2363
        QSIMPLEQ_INSERT_TAIL(&snap_bdrv_states, state, entry);
2364

2365
        state->ops->prepare(state, &local_err);
2366
        if (local_err) {
2367 2368
            error_propagate(errp, local_err);
            goto delete_and_fail;
2369 2370 2371
        }
    }

2372
    QSIMPLEQ_FOREACH(state, &snap_bdrv_states, entry) {
2373 2374 2375
        if (state->ops->commit) {
            state->ops->commit(state);
        }
2376 2377 2378 2379 2380 2381
    }

    /* success */
    goto exit;

delete_and_fail:
2382
    /* failure, and it is all-or-none; roll back all operations */
2383 2384 2385
    QSIMPLEQ_FOREACH(state, &snap_bdrv_states, entry) {
        if (state->ops->abort) {
            state->ops->abort(state);
2386
        }
2387 2388
    }
exit:
2389 2390 2391
    QSIMPLEQ_FOREACH_SAFE(state, &snap_bdrv_states, entry, next) {
        if (state->ops->clean) {
            state->ops->clean(state);
2392
        }
2393
        g_free(state);
2394
    }
2395 2396 2397
    if (!has_props) {
        qapi_free_TransactionProperties(props);
    }
2398
    job_txn_unref(block_job_txn);
2399 2400
}

2401 2402 2403
void qmp_eject(bool has_device, const char *device,
               bool has_id, const char *id,
               bool has_force, bool force, Error **errp)
2404
{
2405
    Error *local_err = NULL;
2406 2407 2408 2409 2410
    int rc;

    if (!has_force) {
        force = false;
    }
2411

2412 2413 2414
    rc = do_open_tray(has_device ? device : NULL,
                      has_id ? id : NULL,
                      force, &local_err);
2415
    if (rc && rc != -ENOSYS) {
2416
        error_propagate(errp, local_err);
L
Luiz Capitulino 已提交
2417
        return;
2418
    }
2419
    error_free(local_err);
2420

2421
    blockdev_remove_medium(has_device, device, has_id, id, errp);
2422 2423
}

2424 2425 2426
void qmp_block_passwd(bool has_device, const char *device,
                      bool has_node_name, const char *node_name,
                      const char *password, Error **errp)
2427
{
2428 2429
    error_setg(errp,
               "Setting block passwords directly is no longer supported");
2430 2431
}

2432 2433 2434 2435 2436 2437 2438 2439 2440 2441
/*
 * 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.
2442
 */
2443 2444
static int do_open_tray(const char *blk_name, const char *qdev_id,
                        bool force, Error **errp)
2445 2446
{
    BlockBackend *blk;
2447
    const char *device = qdev_id ?: blk_name;
2448 2449
    bool locked;

2450
    blk = qmp_get_blk(blk_name, qdev_id, errp);
2451
    if (!blk) {
2452
        return -ENODEV;
2453 2454 2455 2456
    }

    if (!blk_dev_has_removable_media(blk)) {
        error_setg(errp, "Device '%s' is not removable", device);
2457
        return -ENOTSUP;
2458 2459
    }

2460
    if (!blk_dev_has_tray(blk)) {
2461 2462
        error_setg(errp, "Device '%s' does not have a tray", device);
        return -ENOSYS;
2463 2464
    }

2465
    if (blk_dev_is_tray_open(blk)) {
2466
        return 0;
2467 2468 2469 2470 2471 2472 2473 2474
    }

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

    if (!locked || force) {
2475
        blk_dev_change_media_cb(blk, false, &error_abort);
2476
    }
2477 2478

    if (locked && !force) {
2479 2480 2481
        error_setg(errp, "Device '%s' is locked and force was not specified, "
                   "wait for tray to open and try again", device);
        return -EINPROGRESS;
2482 2483 2484 2485 2486
    }

    return 0;
}

2487 2488 2489
void qmp_blockdev_open_tray(bool has_device, const char *device,
                            bool has_id, const char *id,
                            bool has_force, bool force,
2490 2491
                            Error **errp)
{
2492 2493 2494
    Error *local_err = NULL;
    int rc;

2495 2496 2497
    if (!has_force) {
        force = false;
    }
2498 2499 2500
    rc = do_open_tray(has_device ? device : NULL,
                      has_id ? id : NULL,
                      force, &local_err);
2501 2502 2503 2504 2505
    if (rc && rc != -ENOSYS && rc != -EINPROGRESS) {
        error_propagate(errp, local_err);
        return;
    }
    error_free(local_err);
2506 2507
}

2508 2509 2510
void qmp_blockdev_close_tray(bool has_device, const char *device,
                             bool has_id, const char *id,
                             Error **errp)
2511 2512
{
    BlockBackend *blk;
2513
    Error *local_err = NULL;
2514

2515 2516 2517 2518
    device = has_device ? device : NULL;
    id = has_id ? id : NULL;

    blk = qmp_get_blk(device, id, errp);
2519 2520 2521 2522 2523
    if (!blk) {
        return;
    }

    if (!blk_dev_has_removable_media(blk)) {
2524
        error_setg(errp, "Device '%s' is not removable", device ?: id);
2525 2526 2527
        return;
    }

2528 2529 2530 2531 2532
    if (!blk_dev_has_tray(blk)) {
        /* Ignore this command on tray-less devices */
        return;
    }

2533 2534 2535 2536
    if (!blk_dev_is_tray_open(blk)) {
        return;
    }

2537 2538 2539 2540 2541
    blk_dev_change_media_cb(blk, true, &local_err);
    if (local_err) {
        error_propagate(errp, local_err);
        return;
    }
2542 2543
}

2544 2545
static void blockdev_remove_medium(bool has_device, const char *device,
                                   bool has_id, const char *id, Error **errp)
2546 2547 2548 2549
{
    BlockBackend *blk;
    BlockDriverState *bs;
    AioContext *aio_context;
2550
    bool has_attached_device;
2551

2552 2553 2554 2555
    device = has_device ? device : NULL;
    id = has_id ? id : NULL;

    blk = qmp_get_blk(device, id, errp);
2556 2557 2558 2559 2560
    if (!blk) {
        return;
    }

    /* For BBs without a device, we can exchange the BDS tree at will */
2561
    has_attached_device = blk_get_attached_dev(blk);
2562

2563 2564
    if (has_attached_device && !blk_dev_has_removable_media(blk)) {
        error_setg(errp, "Device '%s' is not removable", device ?: id);
2565 2566 2567
        return;
    }

2568 2569 2570 2571
    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);
2572 2573 2574 2575 2576 2577 2578 2579 2580 2581 2582 2583 2584 2585 2586 2587 2588
        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);

2589 2590 2591 2592 2593
    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). */
2594
        blk_dev_change_media_cb(blk, false, &error_abort);
2595 2596
    }

2597 2598 2599 2600
out:
    aio_context_release(aio_context);
}

2601
void qmp_blockdev_remove_medium(const char *id, Error **errp)
2602 2603 2604 2605
{
    blockdev_remove_medium(false, NULL, true, id, errp);
}

2606
static void qmp_blockdev_insert_anon_medium(BlockBackend *blk,
2607 2608
                                            BlockDriverState *bs, Error **errp)
{
2609
    Error *local_err = NULL;
2610
    bool has_device;
2611
    int ret;
2612 2613 2614 2615 2616

    /* 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)) {
2617
        error_setg(errp, "Device is not removable");
2618 2619 2620
        return;
    }

2621
    if (has_device && blk_dev_has_tray(blk) && !blk_dev_is_tray_open(blk)) {
2622
        error_setg(errp, "Tray of the device is not open");
2623 2624 2625 2626
        return;
    }

    if (blk_bs(blk)) {
2627
        error_setg(errp, "There already is a medium in the device");
2628 2629 2630
        return;
    }

2631 2632 2633 2634
    ret = blk_insert_bs(blk, bs, errp);
    if (ret < 0) {
        return;
    }
2635

2636 2637 2638 2639 2640 2641
    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). */
2642 2643 2644 2645 2646 2647
        blk_dev_change_media_cb(blk, true, &local_err);
        if (local_err) {
            error_propagate(errp, local_err);
            blk_remove_bs(blk);
            return;
        }
2648
    }
2649 2650
}

2651 2652 2653
static void blockdev_insert_medium(bool has_device, const char *device,
                                   bool has_id, const char *id,
                                   const char *node_name, Error **errp)
2654
{
2655
    BlockBackend *blk;
2656 2657
    BlockDriverState *bs;

2658 2659 2660 2661 2662 2663 2664
    blk = qmp_get_blk(has_device ? device : NULL,
                      has_id ? id : NULL,
                      errp);
    if (!blk) {
        return;
    }

2665 2666 2667 2668 2669 2670
    bs = bdrv_find_node(node_name);
    if (!bs) {
        error_setg(errp, "Node '%s' not found", node_name);
        return;
    }

2671
    if (bdrv_has_blk(bs)) {
2672
        error_setg(errp, "Node '%s' is already in use", node_name);
2673 2674 2675
        return;
    }

2676
    qmp_blockdev_insert_anon_medium(blk, bs, errp);
2677 2678
}

2679 2680
void qmp_blockdev_insert_medium(const char *id, const char *node_name,
                                Error **errp)
2681 2682 2683 2684
{
    blockdev_insert_medium(false, NULL, true, id, node_name, errp);
}

2685 2686 2687
void qmp_blockdev_change_medium(bool has_device, const char *device,
                                bool has_id, const char *id,
                                const char *filename,
2688
                                bool has_format, const char *format,
2689 2690
                                bool has_read_only,
                                BlockdevChangeReadOnlyMode read_only,
2691
                                Error **errp)
2692 2693 2694
{
    BlockBackend *blk;
    BlockDriverState *medium_bs = NULL;
2695
    int bdrv_flags;
2696
    bool detect_zeroes;
2697
    int rc;
2698 2699 2700
    QDict *options = NULL;
    Error *err = NULL;

2701 2702 2703
    blk = qmp_get_blk(has_device ? device : NULL,
                      has_id ? id : NULL,
                      errp);
2704 2705 2706 2707 2708 2709 2710 2711 2712
    if (!blk) {
        goto fail;
    }

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

    bdrv_flags = blk_get_open_flags_from_root_state(blk);
2713 2714
    bdrv_flags &= ~(BDRV_O_TEMPORARY | BDRV_O_SNAPSHOT | BDRV_O_NO_BACKING |
        BDRV_O_PROTOCOL);
2715

2716 2717 2718 2719 2720 2721 2722 2723 2724 2725 2726 2727 2728 2729 2730 2731 2732 2733 2734 2735
    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();
    }

2736 2737
    options = qdict_new();
    detect_zeroes = blk_get_detect_zeroes_from_root_state(blk);
2738
    qdict_put_str(options, "detect-zeroes", detect_zeroes ? "on" : "off");
2739

2740
    if (has_format) {
2741
        qdict_put_str(options, "driver", format);
2742 2743
    }

2744 2745
    medium_bs = bdrv_open(filename, NULL, options, bdrv_flags, errp);
    if (!medium_bs) {
2746 2747 2748
        goto fail;
    }

2749 2750 2751
    rc = do_open_tray(has_device ? device : NULL,
                      has_id ? id : NULL,
                      false, &err);
2752
    if (rc && rc != -ENOSYS) {
2753 2754 2755
        error_propagate(errp, err);
        goto fail;
    }
2756 2757
    error_free(err);
    err = NULL;
2758

2759
    blockdev_remove_medium(has_device, device, has_id, id, &err);
2760 2761 2762 2763 2764
    if (err) {
        error_propagate(errp, err);
        goto fail;
    }

2765
    qmp_blockdev_insert_anon_medium(blk, medium_bs, &err);
2766 2767 2768 2769 2770
    if (err) {
        error_propagate(errp, err);
        goto fail;
    }

2771
    qmp_blockdev_close_tray(has_device, device, has_id, id, errp);
2772 2773 2774 2775 2776 2777 2778 2779

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);
}

2780
/* throttling disk I/O limits */
2781
void qmp_block_set_io_throttle(BlockIOThrottle *arg, Error **errp)
2782
{
2783
    ThrottleConfig cfg;
2784
    BlockDriverState *bs;
2785
    BlockBackend *blk;
2786
    AioContext *aio_context;
2787

2788 2789 2790
    blk = qmp_get_blk(arg->has_device ? arg->device : NULL,
                      arg->has_id ? arg->id : NULL,
                      errp);
2791
    if (!blk) {
2792
        return;
2793
    }
M
Max Reitz 已提交
2794 2795 2796 2797

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

2798
    bs = blk_bs(blk);
M
Max Reitz 已提交
2799
    if (!bs) {
2800
        error_setg(errp, "Device has no medium");
M
Max Reitz 已提交
2801 2802
        goto out;
    }
2803

2804
    throttle_config_init(&cfg);
2805 2806 2807
    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;
2808

2809 2810 2811
    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;
2812

2813 2814
    if (arg->has_bps_max) {
        cfg.buckets[THROTTLE_BPS_TOTAL].max = arg->bps_max;
2815
    }
2816 2817
    if (arg->has_bps_rd_max) {
        cfg.buckets[THROTTLE_BPS_READ].max = arg->bps_rd_max;
2818
    }
2819 2820
    if (arg->has_bps_wr_max) {
        cfg.buckets[THROTTLE_BPS_WRITE].max = arg->bps_wr_max;
2821
    }
2822 2823
    if (arg->has_iops_max) {
        cfg.buckets[THROTTLE_OPS_TOTAL].max = arg->iops_max;
2824
    }
2825 2826
    if (arg->has_iops_rd_max) {
        cfg.buckets[THROTTLE_OPS_READ].max = arg->iops_rd_max;
2827
    }
2828 2829
    if (arg->has_iops_wr_max) {
        cfg.buckets[THROTTLE_OPS_WRITE].max = arg->iops_wr_max;
2830
    }
2831

2832 2833
    if (arg->has_bps_max_length) {
        cfg.buckets[THROTTLE_BPS_TOTAL].burst_length = arg->bps_max_length;
2834
    }
2835 2836
    if (arg->has_bps_rd_max_length) {
        cfg.buckets[THROTTLE_BPS_READ].burst_length = arg->bps_rd_max_length;
2837
    }
2838 2839
    if (arg->has_bps_wr_max_length) {
        cfg.buckets[THROTTLE_BPS_WRITE].burst_length = arg->bps_wr_max_length;
2840
    }
2841 2842
    if (arg->has_iops_max_length) {
        cfg.buckets[THROTTLE_OPS_TOTAL].burst_length = arg->iops_max_length;
2843
    }
2844 2845
    if (arg->has_iops_rd_max_length) {
        cfg.buckets[THROTTLE_OPS_READ].burst_length = arg->iops_rd_max_length;
2846
    }
2847 2848
    if (arg->has_iops_wr_max_length) {
        cfg.buckets[THROTTLE_OPS_WRITE].burst_length = arg->iops_wr_max_length;
2849 2850
    }

2851 2852
    if (arg->has_iops_size) {
        cfg.op_size = arg->iops_size;
2853
    }
2854

2855
    if (!throttle_is_valid(&cfg, errp)) {
M
Max Reitz 已提交
2856
        goto out;
2857 2858
    }

2859 2860 2861
    if (throttle_enabled(&cfg)) {
        /* Enable I/O limits if they're not enabled yet, otherwise
         * just update the throttling group. */
2862
        if (!blk_get_public(blk)->throttle_group_member.throttle_state) {
2863
            blk_io_limits_enable(blk,
2864 2865 2866
                                 arg->has_group ? arg->group :
                                 arg->has_device ? arg->device :
                                 arg->id);
2867 2868
        } else if (arg->has_group) {
            blk_io_limits_update_group(blk, arg->group);
2869 2870
        }
        /* Set the new throttling configuration */
2871
        blk_set_io_limits(blk, &cfg);
2872
    } else if (blk_get_public(blk)->throttle_group_member.throttle_state) {
2873
        /* If all throttling settings are set to 0, disable I/O limits */
2874
        blk_io_limits_disable(blk);
2875
    }
2876

M
Max Reitz 已提交
2877
out:
2878
    aio_context_release(aio_context);
2879 2880
}

2881 2882
void qmp_block_dirty_bitmap_add(const char *node, const char *name,
                                bool has_granularity, uint32_t granularity,
2883
                                bool has_persistent, bool persistent,
2884
                                bool has_autoload, bool autoload,
2885
                                bool has_disabled, bool disabled,
2886 2887 2888
                                Error **errp)
{
    BlockDriverState *bs;
2889
    BdrvDirtyBitmap *bitmap;
2890 2891 2892 2893 2894 2895 2896 2897 2898 2899 2900 2901 2902 2903 2904

    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");
2905
            return;
2906 2907 2908 2909 2910 2911
        }
    } else {
        /* Default to cluster size, if available: */
        granularity = bdrv_get_default_bitmap_granularity(bs);
    }

2912 2913 2914
    if (!has_persistent) {
        persistent = false;
    }
2915

2916 2917
    if (has_autoload) {
        warn_report("Autoload option is deprecated and its value is ignored");
2918
    }
2919

2920 2921 2922 2923
    if (!has_disabled) {
        disabled = false;
    }

2924 2925 2926 2927 2928 2929 2930
    if (persistent &&
        !bdrv_can_store_new_dirty_bitmap(bs, name, granularity, errp))
    {
        return;
    }

    bitmap = bdrv_create_dirty_bitmap(bs, granularity, name, errp);
2931 2932
    if (bitmap == NULL) {
        return;
2933
    }
2934

2935 2936 2937 2938
    if (disabled) {
        bdrv_disable_dirty_bitmap(bitmap);
    }

2939
    bdrv_dirty_bitmap_set_persistance(bitmap, persistent);
2940 2941 2942 2943 2944 2945 2946
}

void qmp_block_dirty_bitmap_remove(const char *node, const char *name,
                                   Error **errp)
{
    BlockDriverState *bs;
    BdrvDirtyBitmap *bitmap;
2947
    Error *local_err = NULL;
2948

2949
    bitmap = block_dirty_bitmap_lookup(node, name, &bs, errp);
2950 2951 2952 2953
    if (!bitmap || !bs) {
        return;
    }

2954 2955 2956 2957
    if (bdrv_dirty_bitmap_frozen(bitmap)) {
        error_setg(errp,
                   "Bitmap '%s' is currently frozen and cannot be removed",
                   name);
2958
        return;
2959 2960 2961 2962 2963
    } else if (bdrv_dirty_bitmap_qmp_locked(bitmap)) {
        error_setg(errp,
                   "Bitmap '%s' is currently locked and cannot be removed",
                   name);
        return;
2964
    }
2965 2966 2967 2968 2969 2970 2971 2972 2973

    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;
        }
    }

2974 2975 2976
    bdrv_release_dirty_bitmap(bs, bitmap);
}

2977 2978 2979 2980 2981 2982 2983 2984 2985 2986
/**
 * 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;

2987
    bitmap = block_dirty_bitmap_lookup(node, name, &bs, errp);
2988 2989 2990 2991 2992 2993 2994 2995
    if (!bitmap || !bs) {
        return;
    }

    if (bdrv_dirty_bitmap_frozen(bitmap)) {
        error_setg(errp,
                   "Bitmap '%s' is currently frozen and cannot be modified",
                   name);
2996
        return;
2997 2998 2999 3000 3001
    } else if (bdrv_dirty_bitmap_qmp_locked(bitmap)) {
        error_setg(errp,
                   "Bitmap '%s' is currently locked and cannot be modified",
                   name);
        return;
3002 3003 3004 3005
    } else if (!bdrv_dirty_bitmap_enabled(bitmap)) {
        error_setg(errp,
                   "Bitmap '%s' is currently disabled and cannot be cleared",
                   name);
3006
        return;
3007 3008 3009
    } else if (bdrv_dirty_bitmap_readonly(bitmap)) {
        error_setg(errp, "Bitmap '%s' is readonly and cannot be cleared", name);
        return;
3010 3011
    }

F
Fam Zheng 已提交
3012
    bdrv_clear_dirty_bitmap(bitmap, NULL);
3013 3014
}

3015 3016 3017 3018 3019 3020 3021 3022 3023 3024 3025 3026 3027 3028 3029 3030 3031 3032 3033 3034 3035 3036 3037 3038 3039 3040 3041 3042 3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054 3055 3056
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);
}

3057 3058 3059 3060 3061 3062 3063 3064 3065 3066 3067 3068 3069 3070 3071 3072 3073 3074 3075 3076 3077 3078 3079 3080 3081 3082 3083 3084 3085 3086
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);
}

3087 3088 3089 3090 3091 3092 3093 3094 3095 3096 3097 3098 3099 3100 3101 3102 3103 3104 3105 3106 3107 3108 3109 3110 3111
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;
}

3112
void hmp_drive_del(Monitor *mon, const QDict *qdict)
3113 3114
{
    const char *id = qdict_get_str(qdict, "id");
3115
    BlockBackend *blk;
3116
    BlockDriverState *bs;
3117
    AioContext *aio_context;
3118
    Error *local_err = NULL;
3119

3120 3121
    bs = bdrv_find_node(id);
    if (bs) {
3122
        qmp_blockdev_del(id, &local_err);
3123 3124 3125 3126 3127 3128
        if (local_err) {
            error_report_err(local_err);
        }
        return;
    }

3129 3130
    blk = blk_by_name(id);
    if (!blk) {
3131
        error_report("Device '%s' not found", id);
3132
        return;
3133
    }
3134

3135
    if (!blk_legacy_dinfo(blk)) {
3136 3137
        error_report("Deleting device added with blockdev-add"
                     " is not supported");
3138
        return;
3139 3140
    }

M
Max Reitz 已提交
3141
    aio_context = blk_get_aio_context(blk);
3142 3143
    aio_context_acquire(aio_context);

M
Max Reitz 已提交
3144 3145 3146 3147 3148 3149 3150
    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;
        }
3151

3152
        blk_remove_bs(blk);
M
Max Reitz 已提交
3153
    }
3154

3155
    /* Make the BlockBackend and the attached BlockDriverState anonymous */
3156 3157
    monitor_remove_blk(blk);

3158 3159
    /* 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.
3160
     */
3161
    if (blk_get_attached_dev(blk)) {
3162
        /* Further I/O must not pause the guest */
3163 3164
        blk_set_on_error(blk, BLOCKDEV_ON_ERROR_REPORT,
                         BLOCKDEV_ON_ERROR_REPORT);
3165
    } else {
3166
        blk_unref(blk);
3167 3168
    }

3169
    aio_context_release(aio_context);
3170
}
3171

3172 3173 3174
void qmp_block_resize(bool has_device, const char *device,
                      bool has_node_name, const char *node_name,
                      int64_t size, Error **errp)
3175
{
3176
    Error *local_err = NULL;
3177
    BlockBackend *blk = NULL;
3178
    BlockDriverState *bs;
3179
    AioContext *aio_context;
3180
    int ret;
3181

3182 3183 3184
    bs = bdrv_lookup_bs(has_device ? device : NULL,
                        has_node_name ? node_name : NULL,
                        &local_err);
3185
    if (local_err) {
3186 3187 3188 3189
        error_propagate(errp, local_err);
        return;
    }

3190 3191 3192
    aio_context = bdrv_get_aio_context(bs);
    aio_context_acquire(aio_context);

3193
    if (!bdrv_is_first_non_filter(bs)) {
3194
        error_setg(errp, QERR_FEATURE_DISABLED, "resize");
3195
        goto out;
3196 3197 3198
    }

    if (size < 0) {
3199
        error_setg(errp, QERR_INVALID_PARAMETER_VALUE, "size", "a >0 size");
3200
        goto out;
3201 3202
    }

3203
    if (bdrv_op_is_blocked(bs, BLOCK_OP_TYPE_RESIZE, NULL)) {
3204
        error_setg(errp, QERR_DEVICE_IN_USE, device);
3205
        goto out;
3206 3207
    }

3208
    blk = blk_new(BLK_PERM_RESIZE, BLK_PERM_ALL);
3209 3210 3211 3212
    ret = blk_insert_bs(blk, bs, errp);
    if (ret < 0) {
        goto out;
    }
3213

3214
    bdrv_drained_begin(bs);
3215
    ret = blk_truncate(blk, size, PREALLOC_MODE_OFF, errp);
3216
    bdrv_drained_end(bs);
3217 3218

out:
3219
    blk_unref(blk);
3220
    aio_context_release(aio_context);
3221
}
3222

3223
void qmp_block_stream(bool has_job_id, const char *job_id, const char *device,
3224
                      bool has_base, const char *base,
3225
                      bool has_base_node, const char *base_node,
3226 3227
                      bool has_backing_file, const char *backing_file,
                      bool has_speed, int64_t speed,
3228 3229
                      bool has_on_error, BlockdevOnError on_error,
                      Error **errp)
3230
{
3231
    BlockDriverState *bs, *iter;
3232
    BlockDriverState *base_bs = NULL;
3233
    AioContext *aio_context;
3234
    Error *local_err = NULL;
3235
    const char *base_name = NULL;
3236

3237 3238 3239 3240
    if (!has_on_error) {
        on_error = BLOCKDEV_ON_ERROR_REPORT;
    }

3241
    bs = bdrv_lookup_bs(device, device, errp);
3242
    if (!bs) {
3243 3244 3245
        return;
    }

3246
    aio_context = bdrv_get_aio_context(bs);
3247 3248
    aio_context_acquire(aio_context);

3249 3250 3251 3252 3253 3254
    if (has_base && has_base_node) {
        error_setg(errp, "'base' and 'base-node' cannot be specified "
                   "at the same time");
        goto out;
    }

3255
    if (has_base) {
3256 3257
        base_bs = bdrv_find_backing_image(bs, base);
        if (base_bs == NULL) {
3258
            error_setg(errp, QERR_BASE_NOT_FOUND, base);
3259
            goto out;
3260
        }
3261
        assert(bdrv_get_aio_context(base_bs) == aio_context);
3262
        base_name = base;
3263 3264
    }

3265 3266 3267 3268 3269 3270 3271 3272 3273 3274 3275 3276 3277 3278
    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;
    }

3279 3280 3281 3282 3283 3284 3285
    /* 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;
        }
    }

3286 3287 3288 3289 3290
    /* 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");
3291
        goto out;
3292 3293 3294 3295 3296
    }

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

3297
    stream_start(has_job_id ? job_id : NULL, bs, base_bs, base_name,
3298
                 has_speed ? speed : 0, on_error, &local_err);
3299
    if (local_err) {
3300
        error_propagate(errp, local_err);
3301
        goto out;
3302 3303 3304
    }

    trace_qmp_block_stream(bs, bs->job);
3305 3306 3307

out:
    aio_context_release(aio_context);
3308
}
3309

3310
void qmp_block_commit(bool has_job_id, const char *job_id, const char *device,
3311 3312
                      bool has_base, const char *base,
                      bool has_top, const char *top,
3313
                      bool has_backing_file, const char *backing_file,
3314
                      bool has_speed, int64_t speed,
3315
                      bool has_filter_node_name, const char *filter_node_name,
3316 3317 3318
                      Error **errp)
{
    BlockDriverState *bs;
3319
    BlockDriverState *iter;
3320
    BlockDriverState *base_bs, *top_bs;
3321
    AioContext *aio_context;
3322 3323 3324 3325
    Error *local_err = NULL;
    /* This will be part of the QMP command, if/when the
     * BlockdevOnError change for blkmirror makes it in
     */
3326
    BlockdevOnError on_error = BLOCKDEV_ON_ERROR_REPORT;
3327

3328 3329 3330
    if (!has_speed) {
        speed = 0;
    }
3331 3332 3333
    if (!has_filter_node_name) {
        filter_node_name = NULL;
    }
3334

3335 3336 3337 3338 3339
    /* 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. */
3340 3341 3342 3343 3344 3345 3346 3347 3348 3349
    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);
        }
3350
        return;
3351 3352
    }

3353
    aio_context = bdrv_get_aio_context(bs);
3354 3355
    aio_context_acquire(aio_context);

3356
    if (bdrv_op_is_blocked(bs, BLOCK_OP_TYPE_COMMIT_SOURCE, errp)) {
3357
        goto out;
3358 3359 3360 3361 3362
    }

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

3363
    if (has_top && top) {
3364 3365 3366 3367 3368 3369 3370
        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");
3371
        goto out;
3372 3373
    }

3374 3375
    assert(bdrv_get_aio_context(top_bs) == aio_context);

3376 3377 3378 3379 3380 3381 3382
    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) {
3383
        error_setg(errp, QERR_BASE_NOT_FOUND, base ? base : "NULL");
3384
        goto out;
3385 3386
    }

3387 3388
    assert(bdrv_get_aio_context(base_bs) == aio_context);

3389 3390 3391 3392
    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;
        }
3393 3394
    }

3395 3396 3397
    /* Do not allow attempts to commit an image into itself */
    if (top_bs == base_bs) {
        error_setg(errp, "cannot commit an image into itself");
3398
        goto out;
3399 3400
    }

3401
    if (top_bs == bs) {
3402 3403 3404
        if (has_backing_file) {
            error_setg(errp, "'backing-file' specified,"
                             " but 'top' is the active layer");
3405
            goto out;
3406
        }
3407
        commit_active_start(has_job_id ? job_id : NULL, bs, base_bs,
3408
                            JOB_DEFAULT, speed, on_error,
3409
                            filter_node_name, NULL, NULL, false, &local_err);
3410
    } else {
3411 3412 3413 3414
        BlockDriverState *overlay_bs = bdrv_find_overlay(bs, top_bs);
        if (bdrv_op_is_blocked(overlay_bs, BLOCK_OP_TYPE_COMMIT_TARGET, errp)) {
            goto out;
        }
3415
        commit_start(has_job_id ? job_id : NULL, bs, base_bs, top_bs, speed,
3416
                     on_error, has_backing_file ? backing_file : NULL,
3417
                     filter_node_name, &local_err);
3418
    }
3419 3420
    if (local_err != NULL) {
        error_propagate(errp, local_err);
3421
        goto out;
3422
    }
3423 3424 3425

out:
    aio_context_release(aio_context);
3426 3427
}

3428
static BlockJob *do_drive_backup(DriveBackup *backup, JobTxn *txn,
3429
                                 Error **errp)
3430 3431 3432
{
    BlockDriverState *bs;
    BlockDriverState *target_bs;
3433
    BlockDriverState *source = NULL;
3434
    BlockJob *job = NULL;
3435
    BdrvDirtyBitmap *bmap = NULL;
3436
    AioContext *aio_context;
3437
    QDict *options = NULL;
3438
    Error *local_err = NULL;
3439
    int flags, job_flags = JOB_DEFAULT;
3440
    int64_t size;
3441
    bool set_backing_hd = false;
3442

3443 3444
    if (!backup->has_speed) {
        backup->speed = 0;
3445
    }
3446 3447
    if (!backup->has_on_source_error) {
        backup->on_source_error = BLOCKDEV_ON_ERROR_REPORT;
3448
    }
3449 3450
    if (!backup->has_on_target_error) {
        backup->on_target_error = BLOCKDEV_ON_ERROR_REPORT;
3451
    }
3452 3453 3454 3455 3456
    if (!backup->has_mode) {
        backup->mode = NEW_IMAGE_MODE_ABSOLUTE_PATHS;
    }
    if (!backup->has_job_id) {
        backup->job_id = NULL;
3457
    }
3458 3459 3460 3461 3462 3463
    if (!backup->has_auto_finalize) {
        backup->auto_finalize = true;
    }
    if (!backup->has_auto_dismiss) {
        backup->auto_dismiss = true;
    }
3464 3465 3466
    if (!backup->has_compress) {
        backup->compress = false;
    }
3467

3468
    bs = qmp_get_root_bs(backup->device, errp);
3469
    if (!bs) {
3470
        return NULL;
3471 3472
    }

3473
    aio_context = bdrv_get_aio_context(bs);
3474 3475
    aio_context_acquire(aio_context);

3476 3477 3478
    if (!backup->has_format) {
        backup->format = backup->mode == NEW_IMAGE_MODE_EXISTING ?
                         NULL : (char*) bs->drv->format_name;
3479 3480
    }

3481
    /* Early check to avoid creating target */
3482
    if (bdrv_op_is_blocked(bs, BLOCK_OP_TYPE_BACKUP_SOURCE, errp)) {
3483
        goto out;
3484 3485
    }

3486
    flags = bs->open_flags | BDRV_O_RDWR;
3487

3488 3489
    /* See if we have a backing HD we can use to create our new image
     * on top of. */
3490
    if (backup->sync == MIRROR_SYNC_MODE_TOP) {
3491
        source = backing_bs(bs);
3492
        if (!source) {
3493
            backup->sync = MIRROR_SYNC_MODE_FULL;
3494 3495
        }
    }
3496
    if (backup->sync == MIRROR_SYNC_MODE_NONE) {
3497
        source = bs;
3498 3499
        flags |= BDRV_O_NO_BACKING;
        set_backing_hd = true;
3500 3501
    }

3502 3503 3504
    size = bdrv_getlength(bs);
    if (size < 0) {
        error_setg_errno(errp, -size, "bdrv_getlength failed");
3505
        goto out;
3506 3507
    }

3508 3509
    if (backup->mode != NEW_IMAGE_MODE_EXISTING) {
        assert(backup->format);
3510
        if (source) {
3511
            bdrv_img_create(backup->target, backup->format, source->filename,
3512
                            source->drv->format_name, NULL,
3513
                            size, flags, false, &local_err);
3514
        } else {
3515
            bdrv_img_create(backup->target, backup->format, NULL, NULL, NULL,
3516
                            size, flags, false, &local_err);
3517
        }
3518 3519
    }

3520
    if (local_err) {
3521
        error_propagate(errp, local_err);
3522
        goto out;
3523 3524
    }

3525
    if (backup->format) {
3526 3527 3528
        if (!options) {
            options = qdict_new();
        }
3529
        qdict_put_str(options, "driver", backup->format);
3530 3531
    }

3532
    target_bs = bdrv_open(backup->target, NULL, options, flags, errp);
3533
    if (!target_bs) {
3534
        goto out;
3535 3536
    }

3537 3538
    bdrv_set_aio_context(target_bs, aio_context);

3539 3540 3541 3542 3543 3544 3545 3546
    if (set_backing_hd) {
        bdrv_set_backing_hd(target_bs, source, &local_err);
        if (local_err) {
            bdrv_unref(target_bs);
            goto out;
        }
    }

3547 3548
    if (backup->has_bitmap) {
        bmap = bdrv_find_dirty_bitmap(bs, backup->bitmap);
3549
        if (!bmap) {
3550
            error_setg(errp, "Bitmap '%s' could not be found", backup->bitmap);
3551
            bdrv_unref(target_bs);
3552 3553
            goto out;
        }
3554 3555 3556 3557 3558 3559
        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;
        }
3560
    }
3561
    if (!backup->auto_finalize) {
3562
        job_flags |= JOB_MANUAL_FINALIZE;
3563 3564
    }
    if (!backup->auto_dismiss) {
3565
        job_flags |= JOB_MANUAL_DISMISS;
3566
    }
3567

3568 3569 3570
    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,
3571
                            job_flags, NULL, NULL, txn, &local_err);
3572
    bdrv_unref(target_bs);
3573 3574
    if (local_err != NULL) {
        error_propagate(errp, local_err);
3575
        goto out;
3576
    }
3577 3578 3579

out:
    aio_context_release(aio_context);
3580
    return job;
3581 3582
}

3583
void qmp_drive_backup(DriveBackup *arg, Error **errp)
3584
{
3585 3586 3587 3588

    BlockJob *job;
    job = do_drive_backup(arg, NULL, errp);
    if (job) {
3589
        job_start(&job->job);
3590
    }
3591 3592
}

3593 3594
BlockDeviceInfoList *qmp_query_named_block_nodes(Error **errp)
{
3595
    return bdrv_named_nodes_list(errp);
3596 3597
}

3598
BlockJob *do_blockdev_backup(BlockdevBackup *backup, JobTxn *txn,
3599
                             Error **errp)
3600 3601 3602 3603 3604
{
    BlockDriverState *bs;
    BlockDriverState *target_bs;
    Error *local_err = NULL;
    AioContext *aio_context;
3605
    BlockJob *job = NULL;
3606
    int job_flags = JOB_DEFAULT;
3607

3608 3609
    if (!backup->has_speed) {
        backup->speed = 0;
3610
    }
3611 3612 3613 3614 3615
    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;
3616
    }
3617 3618
    if (!backup->has_job_id) {
        backup->job_id = NULL;
3619
    }
3620 3621 3622 3623 3624 3625
    if (!backup->has_auto_finalize) {
        backup->auto_finalize = true;
    }
    if (!backup->has_auto_dismiss) {
        backup->auto_dismiss = true;
    }
3626 3627 3628
    if (!backup->has_compress) {
        backup->compress = false;
    }
3629

3630
    bs = bdrv_lookup_bs(backup->device, backup->device, errp);
3631
    if (!bs) {
3632
        return NULL;
3633 3634
    }

3635
    aio_context = bdrv_get_aio_context(bs);
3636 3637
    aio_context_acquire(aio_context);

3638
    target_bs = bdrv_lookup_bs(backup->target, backup->target, errp);
3639
    if (!target_bs) {
M
Max Reitz 已提交
3640 3641
        goto out;
    }
3642

3643 3644 3645 3646 3647 3648 3649 3650 3651 3652 3653
    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;
        }
    }
3654
    if (!backup->auto_finalize) {
3655
        job_flags |= JOB_MANUAL_FINALIZE;
3656 3657
    }
    if (!backup->auto_dismiss) {
3658
        job_flags |= JOB_MANUAL_DISMISS;
3659
    }
3660 3661 3662
    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,
3663
                            job_flags, NULL, NULL, txn, &local_err);
3664 3665 3666 3667 3668
    if (local_err != NULL) {
        error_propagate(errp, local_err);
    }
out:
    aio_context_release(aio_context);
3669
    return job;
3670 3671
}

3672 3673
void qmp_blockdev_backup(BlockdevBackup *arg, Error **errp)
{
3674 3675 3676
    BlockJob *job;
    job = do_blockdev_backup(arg, NULL, errp);
    if (job) {
3677
        job_start(&job->job);
3678
    }
3679 3680
}

3681 3682 3683
/* Parameter check and block job starting for drive mirroring.
 * Caller should hold @device and @target's aio context (must be the same).
 **/
3684
static void blockdev_mirror_common(const char *job_id, BlockDriverState *bs,
3685 3686 3687
                                   BlockDriverState *target,
                                   bool has_replaces, const char *replaces,
                                   enum MirrorSyncMode sync,
3688
                                   BlockMirrorBackingMode backing_mode,
3689 3690 3691 3692 3693 3694 3695 3696
                                   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,
3697 3698
                                   bool has_filter_node_name,
                                   const char *filter_node_name,
3699
                                   bool has_copy_mode, MirrorCopyMode copy_mode,
3700
                                   Error **errp)
3701 3702 3703 3704 3705
{

    if (!has_speed) {
        speed = 0;
    }
3706 3707 3708 3709 3710 3711
    if (!has_on_source_error) {
        on_source_error = BLOCKDEV_ON_ERROR_REPORT;
    }
    if (!has_on_target_error) {
        on_target_error = BLOCKDEV_ON_ERROR_REPORT;
    }
3712 3713 3714
    if (!has_granularity) {
        granularity = 0;
    }
3715
    if (!has_buf_size) {
W
Wen Congyang 已提交
3716
        buf_size = 0;
3717
    }
3718 3719 3720
    if (!has_unmap) {
        unmap = true;
    }
3721 3722 3723
    if (!has_filter_node_name) {
        filter_node_name = NULL;
    }
3724 3725 3726
    if (!has_copy_mode) {
        copy_mode = MIRROR_COPY_MODE_BACKGROUND;
    }
3727

3728
    if (granularity != 0 && (granularity < 512 || granularity > 1048576 * 64)) {
3729 3730
        error_setg(errp, QERR_INVALID_PARAMETER_VALUE, "granularity",
                   "a value in range [512B, 64MB]");
3731 3732 3733
        return;
    }
    if (granularity & (granularity - 1)) {
3734 3735
        error_setg(errp, QERR_INVALID_PARAMETER_VALUE, "granularity",
                   "power of 2");
3736 3737
        return;
    }
3738

3739 3740 3741
    if (bdrv_op_is_blocked(bs, BLOCK_OP_TYPE_MIRROR_SOURCE, errp)) {
        return;
    }
3742 3743 3744
    if (bdrv_op_is_blocked(target, BLOCK_OP_TYPE_MIRROR_TARGET, errp)) {
        return;
    }
3745 3746 3747 3748 3749 3750 3751 3752

    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
     */
3753
    mirror_start(job_id, bs, target,
3754
                 has_replaces ? replaces : NULL,
3755
                 speed, granularity, buf_size, sync, backing_mode,
3756
                 on_source_error, on_target_error, unmap, filter_node_name,
3757
                 copy_mode, errp);
3758 3759
}

3760
void qmp_drive_mirror(DriveMirror *arg, Error **errp)
3761 3762 3763 3764
{
    BlockDriverState *bs;
    BlockDriverState *source, *target_bs;
    AioContext *aio_context;
3765
    BlockMirrorBackingMode backing_mode;
3766 3767 3768 3769
    Error *local_err = NULL;
    QDict *options = NULL;
    int flags;
    int64_t size;
3770
    const char *format = arg->format;
3771

3772 3773
    bs = qmp_get_root_bs(arg->device, errp);
    if (!bs) {
3774 3775 3776
        return;
    }

3777 3778 3779 3780 3781
    /* Early check to avoid creating target */
    if (bdrv_op_is_blocked(bs, BLOCK_OP_TYPE_MIRROR_SOURCE, errp)) {
        return;
    }

3782
    aio_context = bdrv_get_aio_context(bs);
3783 3784
    aio_context_acquire(aio_context);

3785 3786
    if (!arg->has_mode) {
        arg->mode = NEW_IMAGE_MODE_ABSOLUTE_PATHS;
3787
    }
3788

3789 3790 3791
    if (!arg->has_format) {
        format = (arg->mode == NEW_IMAGE_MODE_EXISTING
                  ? NULL : bs->drv->format_name);
3792 3793
    }

3794
    flags = bs->open_flags | BDRV_O_RDWR;
3795
    source = backing_bs(bs);
3796 3797
    if (!source && arg->sync == MIRROR_SYNC_MODE_TOP) {
        arg->sync = MIRROR_SYNC_MODE_FULL;
3798
    }
3799
    if (arg->sync == MIRROR_SYNC_MODE_NONE) {
3800 3801
        source = bs;
    }
3802

3803 3804 3805
    size = bdrv_getlength(bs);
    if (size < 0) {
        error_setg_errno(errp, -size, "bdrv_getlength failed");
3806
        goto out;
3807 3808
    }

3809
    if (arg->has_replaces) {
3810
        BlockDriverState *to_replace_bs;
3811 3812
        AioContext *replace_aio_context;
        int64_t replace_size;
3813

3814
        if (!arg->has_node_name) {
3815 3816
            error_setg(errp, "a node-name must be provided when replacing a"
                             " named node of the graph");
3817
            goto out;
3818 3819
        }

3820
        to_replace_bs = check_to_replace_node(bs, arg->replaces, &local_err);
3821 3822 3823

        if (!to_replace_bs) {
            error_propagate(errp, local_err);
3824
            goto out;
3825 3826
        }

3827 3828 3829 3830 3831 3832
        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) {
3833 3834
            error_setg(errp, "cannot replace image with a mirror image of "
                             "different size");
3835
            goto out;
3836 3837 3838
        }
    }

3839
    if (arg->mode == NEW_IMAGE_MODE_ABSOLUTE_PATHS) {
3840 3841 3842 3843 3844
        backing_mode = MIRROR_SOURCE_BACKING_CHAIN;
    } else {
        backing_mode = MIRROR_OPEN_BACKING_CHAIN;
    }

3845 3846 3847
    /* Don't open backing image in create() */
    flags |= BDRV_O_NO_BACKING;

3848 3849
    if ((arg->sync == MIRROR_SYNC_MODE_FULL || !source)
        && arg->mode != NEW_IMAGE_MODE_EXISTING)
3850
    {
3851
        /* create new image w/o backing file */
3852
        assert(format);
3853
        bdrv_img_create(arg->target, format,
3854
                        NULL, NULL, NULL, size, flags, false, &local_err);
3855
    } else {
3856
        switch (arg->mode) {
3857 3858 3859 3860
        case NEW_IMAGE_MODE_EXISTING:
            break;
        case NEW_IMAGE_MODE_ABSOLUTE_PATHS:
            /* create new image with backing file */
3861
            bdrv_img_create(arg->target, format,
3862 3863
                            source->filename,
                            source->drv->format_name,
3864
                            NULL, size, flags, false, &local_err);
3865 3866 3867 3868 3869 3870
            break;
        default:
            abort();
        }
    }

3871
    if (local_err) {
3872
        error_propagate(errp, local_err);
3873
        goto out;
3874 3875
    }

3876
    options = qdict_new();
3877
    if (arg->has_node_name) {
3878
        qdict_put_str(options, "node-name", arg->node_name);
3879
    }
3880
    if (format) {
3881
        qdict_put_str(options, "driver", format);
3882
    }
3883

3884 3885 3886
    /* Mirroring takes care of copy-on-write using the source's backing
     * file.
     */
3887
    target_bs = bdrv_open(arg->target, NULL, options, flags, errp);
3888
    if (!target_bs) {
3889
        goto out;
3890 3891
    }

3892 3893
    bdrv_set_aio_context(target_bs, aio_context);

3894 3895 3896 3897 3898 3899 3900 3901
    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,
3902
                           false, NULL,
3903
                           arg->has_copy_mode, arg->copy_mode,
3904
                           &local_err);
3905
    bdrv_unref(target_bs);
3906
    error_propagate(errp, local_err);
3907 3908
out:
    aio_context_release(aio_context);
3909 3910
}

3911 3912
void qmp_blockdev_mirror(bool has_job_id, const char *job_id,
                         const char *device, const char *target,
3913 3914 3915 3916 3917 3918 3919 3920 3921
                         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,
3922 3923
                         bool has_filter_node_name,
                         const char *filter_node_name,
3924
                         bool has_copy_mode, MirrorCopyMode copy_mode,
3925 3926 3927 3928 3929
                         Error **errp)
{
    BlockDriverState *bs;
    BlockDriverState *target_bs;
    AioContext *aio_context;
3930
    BlockMirrorBackingMode backing_mode = MIRROR_LEAVE_BACKING_CHAIN;
3931 3932
    Error *local_err = NULL;

3933
    bs = qmp_get_root_bs(device, errp);
3934 3935 3936 3937 3938 3939 3940 3941 3942 3943 3944 3945 3946 3947
    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);

3948
    blockdev_mirror_common(has_job_id ? job_id : NULL, bs, target_bs,
3949
                           has_replaces, replaces, sync, backing_mode,
3950 3951 3952 3953 3954 3955
                           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,
3956
                           has_filter_node_name, filter_node_name,
3957
                           has_copy_mode, copy_mode,
3958
                           &local_err);
3959
    error_propagate(errp, local_err);
3960 3961 3962 3963

    aio_context_release(aio_context);
}

3964 3965
/* Get a block job using its ID and acquire its AioContext */
static BlockJob *find_block_job(const char *id, AioContext **aio_context,
3966
                                Error **errp)
3967
{
3968
    BlockJob *job;
3969

3970
    assert(id != NULL);
M
Max Reitz 已提交
3971

3972
    *aio_context = NULL;
3973

3974
    job = block_job_get(id);
M
Max Reitz 已提交
3975

3976 3977 3978 3979
    if (!job) {
        error_set(errp, ERROR_CLASS_DEVICE_NOT_ACTIVE,
                  "Block job '%s' not found", id);
        return NULL;
3980
    }
3981

3982 3983
    *aio_context = blk_get_aio_context(job->blk);
    aio_context_acquire(*aio_context);
3984

3985
    return job;
3986 3987
}

3988
void qmp_block_job_set_speed(const char *device, int64_t speed, Error **errp)
3989
{
3990
    AioContext *aio_context;
3991
    BlockJob *job = find_block_job(device, &aio_context, errp);
3992 3993 3994 3995 3996

    if (!job) {
        return;
    }

3997
    block_job_set_speed(job, speed, errp);
3998
    aio_context_release(aio_context);
3999
}
4000

4001 4002
void qmp_block_job_cancel(const char *device,
                          bool has_force, bool force, Error **errp)
4003
{
4004
    AioContext *aio_context;
4005
    BlockJob *job = find_block_job(device, &aio_context, errp);
4006

4007 4008 4009
    if (!job) {
        return;
    }
4010 4011 4012 4013 4014

    if (!has_force) {
        force = false;
    }

4015
    if (job_user_paused(&job->job) && !force) {
4016 4017
        error_setg(errp, "The block job for device '%s' is currently paused",
                   device);
4018
        goto out;
4019
    }
4020 4021

    trace_qmp_block_job_cancel(job);
4022
    job_user_cancel(&job->job, force, errp);
4023 4024
out:
    aio_context_release(aio_context);
4025
}
4026

4027 4028
void qmp_block_job_pause(const char *device, Error **errp)
{
4029
    AioContext *aio_context;
4030
    BlockJob *job = find_block_job(device, &aio_context, errp);
4031

4032
    if (!job) {
4033 4034 4035 4036
        return;
    }

    trace_qmp_block_job_pause(job);
4037
    job_user_pause(&job->job, errp);
4038
    aio_context_release(aio_context);
4039 4040 4041 4042
}

void qmp_block_job_resume(const char *device, Error **errp)
{
4043
    AioContext *aio_context;
4044
    BlockJob *job = find_block_job(device, &aio_context, errp);
4045

4046
    if (!job) {
4047 4048 4049 4050
        return;
    }

    trace_qmp_block_job_resume(job);
4051
    job_user_resume(&job->job, errp);
4052
    aio_context_release(aio_context);
4053 4054
}

4055 4056
void qmp_block_job_complete(const char *device, Error **errp)
{
4057
    AioContext *aio_context;
4058
    BlockJob *job = find_block_job(device, &aio_context, errp);
4059 4060 4061 4062 4063 4064

    if (!job) {
        return;
    }

    trace_qmp_block_job_complete(job);
4065
    job_complete(&job->job, errp);
4066
    aio_context_release(aio_context);
4067 4068
}

4069 4070 4071 4072 4073 4074 4075 4076 4077 4078
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);
4079
    job_finalize(&job->job, errp);
4080 4081 4082
    aio_context_release(aio_context);
}

4083 4084 4085
void qmp_block_job_dismiss(const char *id, Error **errp)
{
    AioContext *aio_context;
K
Kevin Wolf 已提交
4086 4087
    BlockJob *bjob = find_block_job(id, &aio_context, errp);
    Job *job;
4088

K
Kevin Wolf 已提交
4089
    if (!bjob) {
4090 4091 4092
        return;
    }

K
Kevin Wolf 已提交
4093 4094 4095
    trace_qmp_block_job_dismiss(bjob);
    job = &bjob->job;
    job_dismiss(&job, errp);
4096 4097 4098
    aio_context_release(aio_context);
}

4099 4100 4101 4102 4103 4104
void qmp_change_backing_file(const char *device,
                             const char *image_node_name,
                             const char *backing_file,
                             Error **errp)
{
    BlockDriverState *bs = NULL;
4105
    AioContext *aio_context;
4106 4107 4108 4109 4110 4111
    BlockDriverState *image_bs = NULL;
    Error *local_err = NULL;
    bool ro;
    int open_flags;
    int ret;

4112 4113
    bs = qmp_get_root_bs(device, errp);
    if (!bs) {
4114 4115 4116
        return;
    }

4117
    aio_context = bdrv_get_aio_context(bs);
4118 4119
    aio_context_acquire(aio_context);

4120 4121 4122
    image_bs = bdrv_lookup_bs(NULL, image_node_name, &local_err);
    if (local_err) {
        error_propagate(errp, local_err);
4123
        goto out;
4124 4125 4126 4127
    }

    if (!image_bs) {
        error_setg(errp, "image file not found");
4128
        goto out;
4129 4130 4131 4132 4133
    }

    if (bdrv_find_base(image_bs) == image_bs) {
        error_setg(errp, "not allowing backing file change on an image "
                         "without a backing file");
4134
        goto out;
4135 4136 4137 4138 4139
    }

    /* 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)) {
4140
        goto out;
4141 4142 4143 4144 4145 4146
    }

    /* final sanity check */
    if (!bdrv_chain_contains(bs, image_bs)) {
        error_setg(errp, "'%s' and image file are not in the same chain",
                   device);
4147
        goto out;
4148 4149 4150 4151 4152 4153 4154 4155 4156 4157
    }

    /* 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);
4158
            goto out;
4159 4160 4161 4162 4163 4164 4165 4166 4167 4168 4169 4170 4171 4172 4173
        }
    }

    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);
4174
        error_propagate(errp, local_err);
4175
    }
4176 4177 4178

out:
    aio_context_release(aio_context);
4179 4180
}

4181 4182 4183 4184 4185 4186 4187 4188 4189 4190 4191 4192 4193 4194
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")) {
4195
        qobject_unref(qdict);
4196 4197 4198 4199 4200 4201 4202 4203 4204 4205 4206 4207 4208 4209 4210 4211
        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);
}

4212 4213
void qmp_blockdev_add(BlockdevOptions *options, Error **errp)
{
4214
    BlockDriverState *bs;
4215
    QObject *obj;
4216
    Visitor *v = qobject_output_visitor_new(&obj);
4217 4218 4219
    QDict *qdict;
    Error *local_err = NULL;

4220
    visit_type_BlockdevOptions(v, NULL, &options, &local_err);
4221
    if (local_err) {
4222 4223 4224 4225
        error_propagate(errp, local_err);
        goto fail;
    }

4226
    visit_complete(v, &obj);
4227
    qdict = qobject_to(QDict, obj);
4228 4229 4230

    qdict_flatten(qdict);

4231 4232 4233 4234
    if (!qdict_get_try_str(qdict, "node-name")) {
        error_setg(errp, "'node-name' must be specified for the root node");
        goto fail;
    }
4235

4236 4237 4238
    bs = bds_tree_init(qdict, errp);
    if (!bs) {
        goto fail;
4239 4240
    }

4241 4242
    QTAILQ_INSERT_TAIL(&monitor_bdrv_states, bs, monitor_list);

4243
fail:
4244
    visit_free(v);
4245 4246
}

4247
void qmp_blockdev_del(const char *node_name, Error **errp)
4248 4249 4250 4251
{
    AioContext *aio_context;
    BlockDriverState *bs;

4252 4253 4254
    bs = bdrv_find_node(node_name);
    if (!bs) {
        error_setg(errp, "Cannot find node %s", node_name);
4255 4256
        return;
    }
4257 4258 4259
    if (bdrv_has_blk(bs)) {
        error_setg(errp, "Node %s is in use", node_name);
        return;
4260
    }
4261
    aio_context = bdrv_get_aio_context(bs);
4262 4263
    aio_context_acquire(aio_context);

4264 4265 4266
    if (bdrv_op_is_blocked(bs, BLOCK_OP_TYPE_DRIVE_DEL, errp)) {
        goto out;
    }
4267

4268 4269 4270 4271
    if (!bs->monitor_list.tqe_prev) {
        error_setg(errp, "Node %s is not owned by the monitor",
                   bs->node_name);
        goto out;
4272 4273
    }

4274 4275 4276 4277
    if (bs->refcnt > 1) {
        error_setg(errp, "Block device %s is in use",
                   bdrv_get_device_or_node_name(bs));
        goto out;
4278 4279
    }

4280 4281 4282
    QTAILQ_REMOVE(&monitor_bdrv_states, bs, monitor_list);
    bdrv_unref(bs);

4283 4284 4285 4286
out:
    aio_context_release(aio_context);
}

4287 4288 4289 4290 4291 4292 4293 4294 4295 4296 4297 4298 4299 4300 4301 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
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);
    }
}

4342
BlockJobInfoList *qmp_query_block_jobs(Error **errp)
4343
{
4344
    BlockJobInfoList *head = NULL, **p_next = &head;
4345
    BlockJob *job;
4346

4347
    for (job = block_job_next(NULL); job; job = block_job_next(job)) {
4348 4349
        BlockJobInfoList *elem;
        AioContext *aio_context;
4350

4351 4352 4353 4354 4355
        if (block_job_is_internal(job)) {
            continue;
        }
        elem = g_new0(BlockJobInfoList, 1);
        aio_context = blk_get_aio_context(job->blk);
4356
        aio_context_acquire(aio_context);
4357
        elem->value = block_job_query(job, errp);
4358
        aio_context_release(aio_context);
4359 4360 4361 4362 4363
        if (!elem->value) {
            g_free(elem);
            qapi_free_BlockJobInfoList(head);
            return NULL;
        }
4364 4365
        *p_next = elem;
        p_next = &elem->next;
4366 4367
    }

4368
    return head;
4369
}
4370

4371
void qmp_x_blockdev_set_iothread(const char *node_name, StrOrNull *iothread,
4372
                                 bool has_force, bool force, Error **errp)
4373 4374 4375 4376 4377 4378 4379 4380 4381 4382 4383
{
    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;
    }

4384 4385 4386 4387 4388
    /* 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);
4389 4390 4391 4392 4393 4394 4395 4396 4397 4398 4399 4400 4401 4402 4403 4404 4405 4406 4407 4408 4409 4410 4411
        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);
}

4412 4413 4414 4415 4416 4417 4418 4419 4420 4421 4422 4423 4424 4425 4426 4427 4428 4429 4430 4431 4432 4433 4434 4435 4436 4437 4438 4439 4440 4441 4442 4443 4444 4445 4446 4447 4448 4449 4450 4451 4452 4453 4454
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);
    }
}

4455
QemuOptsList qemu_common_drive_opts = {
4456
    .name = "drive",
4457
    .head = QTAILQ_HEAD_INITIALIZER(qemu_common_drive_opts.head),
4458 4459 4460 4461 4462 4463 4464 4465 4466
    .desc = {
        {
            .name = "snapshot",
            .type = QEMU_OPT_BOOL,
            .help = "enable/disable snapshot mode",
        },{
            .name = "aio",
            .type = QEMU_OPT_STRING,
            .help = "host AIO implementation (threads, native)",
4467 4468 4469 4470
        },{
            .name = BDRV_OPT_CACHE_WB,
            .type = QEMU_OPT_BOOL,
            .help = "Enable writeback mode",
4471 4472 4473 4474 4475 4476 4477 4478 4479 4480 4481 4482 4483
        },{
            .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",
        },{
4484
            .name = BDRV_OPT_READ_ONLY,
4485 4486
            .type = QEMU_OPT_BOOL,
            .help = "open drive file as read-only",
4487 4488 4489 4490 4491
        },

        THROTTLE_OPTS,

        {
4492 4493 4494
            .name = "throttling.group",
            .type = QEMU_OPT_STRING,
            .help = "name of the block throttling group",
4495 4496 4497 4498
        },{
            .name = "copy-on-read",
            .type = QEMU_OPT_BOOL,
            .help = "copy read data from backing file into image file",
4499 4500 4501 4502
        },{
            .name = "detect-zeroes",
            .type = QEMU_OPT_STRING,
            .help = "try to optimize zero writes (off, on, unmap)",
4503 4504 4505 4506 4507 4508 4509 4510 4511 4512
        },{
            .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",
4513 4514 4515 4516
        },
        { /* end of list */ }
    },
};
4517 4518 4519 4520 4521

QemuOptsList qemu_drive_opts = {
    .name = "drive",
    .head = QTAILQ_HEAD_INITIALIZER(qemu_drive_opts.head),
    .desc = {
4522 4523 4524 4525
        /*
         * no elements => accept any params
         * validation will happen later
         */
4526 4527 4528
        { /* end of list */ }
    },
};