savevm.c 48.1 KB
Newer Older
A
aliguori 已提交
1 2 3 4
/*
 * QEMU System Emulator
 *
 * Copyright (c) 2003-2008 Fabrice Bellard
5 6 7 8
 * Copyright (c) 2009-2015 Red Hat Inc
 *
 * Authors:
 *  Juan Quintela <quintela@redhat.com>
A
aliguori 已提交
9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28
 *
 * 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.
 */

B
blueswir1 已提交
29
#include "config-host.h"
30
#include "qemu-common.h"
31
#include "hw/boards.h"
32
#include "hw/hw.h"
33
#include "hw/qdev.h"
P
Paolo Bonzini 已提交
34
#include "net/net.h"
35
#include "monitor/monitor.h"
36
#include "sysemu/sysemu.h"
37
#include "qemu/timer.h"
38
#include "audio/audio.h"
39
#include "migration/migration.h"
40
#include "qapi/qmp/qerror.h"
41
#include "qemu/error-report.h"
42 43
#include "qemu/sockets.h"
#include "qemu/queue.h"
44
#include "sysemu/cpus.h"
45
#include "exec/memory.h"
46
#include "qmp-commands.h"
47
#include "trace.h"
48
#include "qemu/iov.h"
49
#include "block/snapshot.h"
50
#include "block/qapi.h"
51

A
aliguori 已提交
52

N
Nolan 已提交
53
#ifndef ETH_P_RARP
S
Stefan Berger 已提交
54
#define ETH_P_RARP 0x8035
N
Nolan 已提交
55 56 57 58 59
#endif
#define ARP_HTYPE_ETH 0x0001
#define ARP_PTYPE_IP 0x0800
#define ARP_OP_REQUEST_REV 0x3

60 61
static bool skip_section_footers;

D
Dr. David Alan Gilbert 已提交
62 63 64 65 66
static struct mig_cmd_args {
    ssize_t     len; /* -1 = variable */
    const char *name;
} mig_cmd_args[] = {
    [MIG_CMD_INVALID]          = { .len = -1, .name = "INVALID" },
67 68
    [MIG_CMD_OPEN_RETURN_PATH] = { .len =  0, .name = "OPEN_RETURN_PATH" },
    [MIG_CMD_PING]             = { .len = sizeof(uint32_t), .name = "PING" },
D
Dr. David Alan Gilbert 已提交
69 70 71
    [MIG_CMD_MAX]              = { .len = -1, .name = "MAX" },
};

N
Nolan 已提交
72
static int announce_self_create(uint8_t *buf,
73
                                uint8_t *mac_addr)
A
aliguori 已提交
74
{
N
Nolan 已提交
75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94
    /* Ethernet header. */
    memset(buf, 0xff, 6);         /* destination MAC addr */
    memcpy(buf + 6, mac_addr, 6); /* source MAC addr */
    *(uint16_t *)(buf + 12) = htons(ETH_P_RARP); /* ethertype */

    /* RARP header. */
    *(uint16_t *)(buf + 14) = htons(ARP_HTYPE_ETH); /* hardware addr space */
    *(uint16_t *)(buf + 16) = htons(ARP_PTYPE_IP); /* protocol addr space */
    *(buf + 18) = 6; /* hardware addr length (ethernet) */
    *(buf + 19) = 4; /* protocol addr length (IPv4) */
    *(uint16_t *)(buf + 20) = htons(ARP_OP_REQUEST_REV); /* opcode */
    memcpy(buf + 22, mac_addr, 6); /* source hw addr */
    memset(buf + 28, 0x00, 4);     /* source protocol addr */
    memcpy(buf + 32, mac_addr, 6); /* target hw addr */
    memset(buf + 38, 0x00, 4);     /* target protocol addr */

    /* Padding to get up to 60 bytes (ethernet min packet size, minus FCS). */
    memset(buf + 42, 0x00, 18);

    return 60; /* len (FCS will be added by hardware) */
A
aliguori 已提交
95 96
}

M
Mark McLoughlin 已提交
97
static void qemu_announce_self_iter(NICState *nic, void *opaque)
A
aliguori 已提交
98
{
N
Nolan 已提交
99
    uint8_t buf[60];
M
Mark McLoughlin 已提交
100 101
    int len;

102
    trace_qemu_announce_self_iter(qemu_ether_ntoa(&nic->conf->macaddr));
M
Mark McLoughlin 已提交
103 104
    len = announce_self_create(buf, nic->conf->macaddr.a);

J
Jason Wang 已提交
105
    qemu_send_packet_raw(qemu_get_queue(nic), buf, len);
M
Mark McLoughlin 已提交
106 107 108 109 110
}


static void qemu_announce_self_once(void *opaque)
{
111 112
    static int count = SELF_ANNOUNCE_ROUNDS;
    QEMUTimer *timer = *(QEMUTimer **)opaque;
A
aliguori 已提交
113

M
Mark McLoughlin 已提交
114 115
    qemu_foreach_nic(qemu_announce_self_iter, NULL);

N
Nolan 已提交
116 117
    if (--count) {
        /* delay 50ms, 150ms, 250ms, ... */
118
        timer_mod(timer, qemu_clock_get_ms(QEMU_CLOCK_REALTIME) +
119
                  self_announce_delay(count));
120
    } else {
121 122
            timer_del(timer);
            timer_free(timer);
123 124 125 126 127
    }
}

void qemu_announce_self(void)
{
128 129 130
    static QEMUTimer *timer;
    timer = timer_new_ms(QEMU_CLOCK_REALTIME, qemu_announce_self_once, &timer);
    qemu_announce_self_once(&timer);
A
aliguori 已提交
131 132 133 134 135
}

/***********************************************************/
/* savevm/loadvm support */

136 137 138 139 140 141 142 143 144 145 146 147 148 149 150
static ssize_t block_writev_buffer(void *opaque, struct iovec *iov, int iovcnt,
                                   int64_t pos)
{
    int ret;
    QEMUIOVector qiov;

    qemu_iovec_init_external(&qiov, iov, iovcnt);
    ret = bdrv_writev_vmstate(opaque, &qiov, pos);
    if (ret < 0) {
        return ret;
    }

    return qiov.size;
}

151 152
static ssize_t block_put_buffer(void *opaque, const uint8_t *buf,
                                int64_t pos, size_t size)
A
aliguori 已提交
153
{
154
    bdrv_save_vmstate(opaque, buf, pos, size);
A
aliguori 已提交
155 156 157
    return size;
}

158 159
static ssize_t block_get_buffer(void *opaque, uint8_t *buf, int64_t pos,
                                size_t size)
A
aliguori 已提交
160
{
161
    return bdrv_load_vmstate(opaque, buf, pos, size);
A
aliguori 已提交
162 163 164 165
}

static int bdrv_fclose(void *opaque)
{
166
    return bdrv_flush(opaque);
A
aliguori 已提交
167 168
}

169 170 171 172 173 174
static const QEMUFileOps bdrv_read_ops = {
    .get_buffer = block_get_buffer,
    .close =      bdrv_fclose
};

static const QEMUFileOps bdrv_write_ops = {
175 176 177
    .put_buffer     = block_put_buffer,
    .writev_buffer  = block_writev_buffer,
    .close          = bdrv_fclose
178 179
};

180
static QEMUFile *qemu_fopen_bdrv(BlockDriverState *bs, int is_writable)
A
aliguori 已提交
181
{
E
Eduardo Habkost 已提交
182
    if (is_writable) {
183
        return qemu_fopen_ops(bs, &bdrv_write_ops);
E
Eduardo Habkost 已提交
184
    }
185
    return qemu_fopen_ops(bs, &bdrv_read_ops);
A
aliguori 已提交
186 187
}

188

189 190 191
/* QEMUFile timer support.
 * Not in qemu-file.c to not add qemu-timer.c as dependency to qemu-file.c
 */
192

193
void timer_put(QEMUFile *f, QEMUTimer *ts)
194 195 196
{
    uint64_t expire_time;

197
    expire_time = timer_expire_time_ns(ts);
198 199 200
    qemu_put_be64(f, expire_time);
}

201
void timer_get(QEMUFile *f, QEMUTimer *ts)
202 203 204 205 206
{
    uint64_t expire_time;

    expire_time = qemu_get_be64(f);
    if (expire_time != -1) {
207
        timer_mod_ns(ts, expire_time);
208
    } else {
209
        timer_del(ts);
210 211 212 213
    }
}


214 215 216
/* VMState timer support.
 * Not in vmstate.c to not add qemu-timer.c as dependency to vmstate.c
 */
J
Juan Quintela 已提交
217 218 219 220

static int get_timer(QEMUFile *f, void *pv, size_t size)
{
    QEMUTimer *v = pv;
221
    timer_get(f, v);
J
Juan Quintela 已提交
222 223 224
    return 0;
}

225
static void put_timer(QEMUFile *f, void *pv, size_t size)
J
Juan Quintela 已提交
226
{
227
    QEMUTimer *v = pv;
228
    timer_put(f, v);
J
Juan Quintela 已提交
229 230 231 232 233 234 235 236
}

const VMStateInfo vmstate_info_timer = {
    .name = "timer",
    .get  = get_timer,
    .put  = put_timer,
};

237

238 239 240 241 242
typedef struct CompatEntry {
    char idstr[256];
    int instance_id;
} CompatEntry;

A
aliguori 已提交
243
typedef struct SaveStateEntry {
B
Blue Swirl 已提交
244
    QTAILQ_ENTRY(SaveStateEntry) entry;
A
aliguori 已提交
245 246
    char idstr[256];
    int instance_id;
J
Jan Kiszka 已提交
247
    int alias_id;
A
aliguori 已提交
248 249
    int version_id;
    int section_id;
250
    SaveVMHandlers *ops;
251
    const VMStateDescription *vmsd;
A
aliguori 已提交
252
    void *opaque;
253
    CompatEntry *compat;
254
    int is_ram;
A
aliguori 已提交
255 256
} SaveStateEntry;

J
Juan Quintela 已提交
257 258 259
typedef struct SaveState {
    QTAILQ_HEAD(, SaveStateEntry) handlers;
    int global_section_id;
260 261 262
    bool skip_configuration;
    uint32_t len;
    const char *name;
J
Juan Quintela 已提交
263 264 265 266 267
} SaveState;

static SaveState savevm_state = {
    .handlers = QTAILQ_HEAD_INITIALIZER(savevm_state.handlers),
    .global_section_id = 0,
268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308
    .skip_configuration = false,
};

void savevm_skip_configuration(void)
{
    savevm_state.skip_configuration = true;
}


static void configuration_pre_save(void *opaque)
{
    SaveState *state = opaque;
    const char *current_name = MACHINE_GET_CLASS(current_machine)->name;

    state->len = strlen(current_name);
    state->name = current_name;
}

static int configuration_post_load(void *opaque, int version_id)
{
    SaveState *state = opaque;
    const char *current_name = MACHINE_GET_CLASS(current_machine)->name;

    if (strncmp(state->name, current_name, state->len) != 0) {
        error_report("Machine type received is '%s' and local is '%s'",
                     state->name, current_name);
        return -EINVAL;
    }
    return 0;
}

static const VMStateDescription vmstate_configuration = {
    .name = "configuration",
    .version_id = 1,
    .post_load = configuration_post_load,
    .pre_save = configuration_pre_save,
    .fields = (VMStateField[]) {
        VMSTATE_UINT32(len, SaveState),
        VMSTATE_VBUFFER_ALLOC_UINT32(name, SaveState, 0, NULL, 0, len),
        VMSTATE_END_OF_LIST()
    },
J
Juan Quintela 已提交
309
};
A
aliguori 已提交
310

311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333
static void dump_vmstate_vmsd(FILE *out_file,
                              const VMStateDescription *vmsd, int indent,
                              bool is_subsection);

static void dump_vmstate_vmsf(FILE *out_file, const VMStateField *field,
                              int indent)
{
    fprintf(out_file, "%*s{\n", indent, "");
    indent += 2;
    fprintf(out_file, "%*s\"field\": \"%s\",\n", indent, "", field->name);
    fprintf(out_file, "%*s\"version_id\": %d,\n", indent, "",
            field->version_id);
    fprintf(out_file, "%*s\"field_exists\": %s,\n", indent, "",
            field->field_exists ? "true" : "false");
    fprintf(out_file, "%*s\"size\": %zu", indent, "", field->size);
    if (field->vmsd != NULL) {
        fprintf(out_file, ",\n");
        dump_vmstate_vmsd(out_file, field->vmsd, indent, false);
    }
    fprintf(out_file, "\n%*s}", indent - 2, "");
}

static void dump_vmstate_vmss(FILE *out_file,
334
                              const VMStateDescription **subsection,
335 336
                              int indent)
{
337 338
    if (*subsection != NULL) {
        dump_vmstate_vmsd(out_file, *subsection, indent, true);
339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378
    }
}

static void dump_vmstate_vmsd(FILE *out_file,
                              const VMStateDescription *vmsd, int indent,
                              bool is_subsection)
{
    if (is_subsection) {
        fprintf(out_file, "%*s{\n", indent, "");
    } else {
        fprintf(out_file, "%*s\"%s\": {\n", indent, "", "Description");
    }
    indent += 2;
    fprintf(out_file, "%*s\"name\": \"%s\",\n", indent, "", vmsd->name);
    fprintf(out_file, "%*s\"version_id\": %d,\n", indent, "",
            vmsd->version_id);
    fprintf(out_file, "%*s\"minimum_version_id\": %d", indent, "",
            vmsd->minimum_version_id);
    if (vmsd->fields != NULL) {
        const VMStateField *field = vmsd->fields;
        bool first;

        fprintf(out_file, ",\n%*s\"Fields\": [\n", indent, "");
        first = true;
        while (field->name != NULL) {
            if (field->flags & VMS_MUST_EXIST) {
                /* Ignore VMSTATE_VALIDATE bits; these don't get migrated */
                field++;
                continue;
            }
            if (!first) {
                fprintf(out_file, ",\n");
            }
            dump_vmstate_vmsf(out_file, field, indent + 2);
            field++;
            first = false;
        }
        fprintf(out_file, "\n%*s]", indent, "");
    }
    if (vmsd->subsections != NULL) {
379
        const VMStateDescription **subsection = vmsd->subsections;
380 381 382 383
        bool first;

        fprintf(out_file, ",\n%*s\"Subsections\": [\n", indent, "");
        first = true;
384
        while (*subsection != NULL) {
385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448
            if (!first) {
                fprintf(out_file, ",\n");
            }
            dump_vmstate_vmss(out_file, subsection, indent + 2);
            subsection++;
            first = false;
        }
        fprintf(out_file, "\n%*s]", indent, "");
    }
    fprintf(out_file, "\n%*s}", indent - 2, "");
}

static void dump_machine_type(FILE *out_file)
{
    MachineClass *mc;

    mc = MACHINE_GET_CLASS(current_machine);

    fprintf(out_file, "  \"vmschkmachine\": {\n");
    fprintf(out_file, "    \"Name\": \"%s\"\n", mc->name);
    fprintf(out_file, "  },\n");
}

void dump_vmstate_json_to_file(FILE *out_file)
{
    GSList *list, *elt;
    bool first;

    fprintf(out_file, "{\n");
    dump_machine_type(out_file);

    first = true;
    list = object_class_get_list(TYPE_DEVICE, true);
    for (elt = list; elt; elt = elt->next) {
        DeviceClass *dc = OBJECT_CLASS_CHECK(DeviceClass, elt->data,
                                             TYPE_DEVICE);
        const char *name;
        int indent = 2;

        if (!dc->vmsd) {
            continue;
        }

        if (!first) {
            fprintf(out_file, ",\n");
        }
        name = object_class_get_name(OBJECT_CLASS(dc));
        fprintf(out_file, "%*s\"%s\": {\n", indent, "", name);
        indent += 2;
        fprintf(out_file, "%*s\"Name\": \"%s\",\n", indent, "", name);
        fprintf(out_file, "%*s\"version_id\": %d,\n", indent, "",
                dc->vmsd->version_id);
        fprintf(out_file, "%*s\"minimum_version_id\": %d,\n", indent, "",
                dc->vmsd->minimum_version_id);

        dump_vmstate_vmsd(out_file, dc->vmsd, indent, false);

        fprintf(out_file, "\n%*s}", indent - 2, "");
        first = false;
    }
    fprintf(out_file, "\n}\n");
    fclose(out_file);
}

449 450 451 452 453
static int calculate_new_instance_id(const char *idstr)
{
    SaveStateEntry *se;
    int instance_id = 0;

J
Juan Quintela 已提交
454
    QTAILQ_FOREACH(se, &savevm_state.handlers, entry) {
455 456 457 458 459 460 461 462
        if (strcmp(idstr, se->idstr) == 0
            && instance_id <= se->instance_id) {
            instance_id = se->instance_id + 1;
        }
    }
    return instance_id;
}

463 464 465 466 467
static int calculate_compat_instance_id(const char *idstr)
{
    SaveStateEntry *se;
    int instance_id = 0;

J
Juan Quintela 已提交
468
    QTAILQ_FOREACH(se, &savevm_state.handlers, entry) {
E
Eduardo Habkost 已提交
469
        if (!se->compat) {
470
            continue;
E
Eduardo Habkost 已提交
471
        }
472 473 474 475 476 477 478 479 480

        if (strcmp(idstr, se->compat->idstr) == 0
            && instance_id <= se->compat->instance_id) {
            instance_id = se->compat->instance_id + 1;
        }
    }
    return instance_id;
}

A
aliguori 已提交
481 482 483 484
/* TODO: Individual devices generally have very little idea about the rest
   of the system, so instance_id should be removed/replaced.
   Meanwhile pass -1 as instance_id if you do not already have a clearly
   distinguishing id for all instances of your device class. */
A
Alex Williamson 已提交
485 486
int register_savevm_live(DeviceState *dev,
                         const char *idstr,
A
aliguori 已提交
487 488
                         int instance_id,
                         int version_id,
489
                         SaveVMHandlers *ops,
A
aliguori 已提交
490 491
                         void *opaque)
{
492
    SaveStateEntry *se;
A
aliguori 已提交
493

494
    se = g_new0(SaveStateEntry, 1);
A
aliguori 已提交
495
    se->version_id = version_id;
J
Juan Quintela 已提交
496
    se->section_id = savevm_state.global_section_id++;
497
    se->ops = ops;
A
aliguori 已提交
498
    se->opaque = opaque;
499
    se->vmsd = NULL;
500
    /* if this is a live_savem then set is_ram */
501
    if (ops->save_live_setup != NULL) {
502 503
        se->is_ram = 1;
    }
A
aliguori 已提交
504

505 506
    if (dev) {
        char *id = qdev_get_dev_path(dev);
507 508 509
        if (id) {
            pstrcpy(se->idstr, sizeof(se->idstr), id);
            pstrcat(se->idstr, sizeof(se->idstr), "/");
510
            g_free(id);
511

512
            se->compat = g_new0(CompatEntry, 1);
513 514 515 516 517 518 519 520
            pstrcpy(se->compat->idstr, sizeof(se->compat->idstr), idstr);
            se->compat->instance_id = instance_id == -1 ?
                         calculate_compat_instance_id(idstr) : instance_id;
            instance_id = -1;
        }
    }
    pstrcat(se->idstr, sizeof(se->idstr), idstr);

521
    if (instance_id == -1) {
522
        se->instance_id = calculate_new_instance_id(se->idstr);
523 524
    } else {
        se->instance_id = instance_id;
A
aliguori 已提交
525
    }
526
    assert(!se->compat || se->instance_id == 0);
527
    /* add at the end of list */
J
Juan Quintela 已提交
528
    QTAILQ_INSERT_TAIL(&savevm_state.handlers, se, entry);
A
aliguori 已提交
529 530 531
    return 0;
}

A
Alex Williamson 已提交
532 533
int register_savevm(DeviceState *dev,
                    const char *idstr,
A
aliguori 已提交
534 535 536 537 538 539
                    int instance_id,
                    int version_id,
                    SaveStateHandler *save_state,
                    LoadStateHandler *load_state,
                    void *opaque)
{
540
    SaveVMHandlers *ops = g_new0(SaveVMHandlers, 1);
541 542
    ops->save_state = save_state;
    ops->load_state = load_state;
A
Alex Williamson 已提交
543
    return register_savevm_live(dev, idstr, instance_id, version_id,
544
                                ops, opaque);
A
aliguori 已提交
545 546
}

A
Alex Williamson 已提交
547
void unregister_savevm(DeviceState *dev, const char *idstr, void *opaque)
548
{
549
    SaveStateEntry *se, *new_se;
550 551
    char id[256] = "";

552 553
    if (dev) {
        char *path = qdev_get_dev_path(dev);
554 555 556
        if (path) {
            pstrcpy(id, sizeof(id), path);
            pstrcat(id, sizeof(id), "/");
557
            g_free(path);
558 559 560
        }
    }
    pstrcat(id, sizeof(id), idstr);
561

J
Juan Quintela 已提交
562
    QTAILQ_FOREACH_SAFE(se, &savevm_state.handlers, entry, new_se) {
563
        if (strcmp(se->idstr, id) == 0 && se->opaque == opaque) {
J
Juan Quintela 已提交
564
            QTAILQ_REMOVE(&savevm_state.handlers, se, entry);
565
            g_free(se->compat);
566
            g_free(se->ops);
567
            g_free(se);
568 569 570 571
        }
    }
}

A
Alex Williamson 已提交
572
int vmstate_register_with_alias_id(DeviceState *dev, int instance_id,
J
Jan Kiszka 已提交
573 574 575
                                   const VMStateDescription *vmsd,
                                   void *opaque, int alias_id,
                                   int required_for_version)
576
{
577
    SaveStateEntry *se;
578

J
Jan Kiszka 已提交
579 580 581
    /* If this triggers, alias support can be dropped for the vmsd. */
    assert(alias_id == -1 || required_for_version >= vmsd->minimum_version_id);

582
    se = g_new0(SaveStateEntry, 1);
583
    se->version_id = vmsd->version_id;
J
Juan Quintela 已提交
584
    se->section_id = savevm_state.global_section_id++;
585 586
    se->opaque = opaque;
    se->vmsd = vmsd;
J
Jan Kiszka 已提交
587
    se->alias_id = alias_id;
588

589 590
    if (dev) {
        char *id = qdev_get_dev_path(dev);
591 592 593
        if (id) {
            pstrcpy(se->idstr, sizeof(se->idstr), id);
            pstrcat(se->idstr, sizeof(se->idstr), "/");
594
            g_free(id);
595

596
            se->compat = g_new0(CompatEntry, 1);
597 598 599 600 601 602 603 604
            pstrcpy(se->compat->idstr, sizeof(se->compat->idstr), vmsd->name);
            se->compat->instance_id = instance_id == -1 ?
                         calculate_compat_instance_id(vmsd->name) : instance_id;
            instance_id = -1;
        }
    }
    pstrcat(se->idstr, sizeof(se->idstr), vmsd->name);

605
    if (instance_id == -1) {
606
        se->instance_id = calculate_new_instance_id(se->idstr);
607 608
    } else {
        se->instance_id = instance_id;
609
    }
610
    assert(!se->compat || se->instance_id == 0);
611
    /* add at the end of list */
J
Juan Quintela 已提交
612
    QTAILQ_INSERT_TAIL(&savevm_state.handlers, se, entry);
613 614 615
    return 0;
}

A
Alex Williamson 已提交
616 617
void vmstate_unregister(DeviceState *dev, const VMStateDescription *vmsd,
                        void *opaque)
618
{
619 620
    SaveStateEntry *se, *new_se;

J
Juan Quintela 已提交
621
    QTAILQ_FOREACH_SAFE(se, &savevm_state.handlers, entry, new_se) {
622
        if (se->vmsd == vmsd && se->opaque == opaque) {
J
Juan Quintela 已提交
623
            QTAILQ_REMOVE(&savevm_state.handlers, se, entry);
624
            g_free(se->compat);
625
            g_free(se);
626 627
        }
    }
628 629
}

630 631
static int vmstate_load(QEMUFile *f, SaveStateEntry *se, int version_id)
{
632
    trace_vmstate_load(se->idstr, se->vmsd ? se->vmsd->name : "(old)");
633
    if (!se->vmsd) {         /* Old style */
634
        return se->ops->load_state(f, se->opaque, version_id);
635 636
    }
    return vmstate_load_state(f, se->vmsd, se->opaque, version_id);
637 638
}

639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659
static void vmstate_save_old_style(QEMUFile *f, SaveStateEntry *se, QJSON *vmdesc)
{
    int64_t old_offset, size;

    old_offset = qemu_ftell_fast(f);
    se->ops->save_state(f, se->opaque);
    size = qemu_ftell_fast(f) - old_offset;

    if (vmdesc) {
        json_prop_int(vmdesc, "size", size);
        json_start_array(vmdesc, "fields");
        json_start_object(vmdesc, NULL);
        json_prop_str(vmdesc, "name", "data");
        json_prop_int(vmdesc, "size", size);
        json_prop_str(vmdesc, "type", "buffer");
        json_end_object(vmdesc);
        json_end_array(vmdesc);
    }
}

static void vmstate_save(QEMUFile *f, SaveStateEntry *se, QJSON *vmdesc)
660
{
661
    trace_vmstate_save(se->idstr, se->vmsd ? se->vmsd->name : "(old)");
662 663
    if (!se->vmsd) {
        vmstate_save_old_style(f, se, vmdesc);
A
Alex Williamson 已提交
664
        return;
665
    }
666
    vmstate_save_state(f, se->vmsd, se->opaque, vmdesc);
667 668
}

669 670 671 672 673
void savevm_skip_section_footers(void)
{
    skip_section_footers = true;
}

674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694
/*
 * Write the header for device section (QEMU_VM_SECTION START/END/PART/FULL)
 */
static void save_section_header(QEMUFile *f, SaveStateEntry *se,
                                uint8_t section_type)
{
    qemu_put_byte(f, section_type);
    qemu_put_be32(f, se->section_id);

    if (section_type == QEMU_VM_SECTION_FULL ||
        section_type == QEMU_VM_SECTION_START) {
        /* ID string */
        size_t len = strlen(se->idstr);
        qemu_put_byte(f, len);
        qemu_put_buffer(f, (uint8_t *)se->idstr, len);

        qemu_put_be32(f, se->instance_id);
        qemu_put_be32(f, se->version_id);
    }
}

695 696 697 698 699 700 701 702 703 704 705 706
/*
 * Write a footer onto device sections that catches cases misformatted device
 * sections.
 */
static void save_section_footer(QEMUFile *f, SaveStateEntry *se)
{
    if (!skip_section_footers) {
        qemu_put_byte(f, QEMU_VM_SECTION_FOOTER);
        qemu_put_be32(f, se->section_id);
    }
}

D
Dr. David Alan Gilbert 已提交
707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728
/**
 * qemu_savevm_command_send: Send a 'QEMU_VM_COMMAND' type element with the
 *                           command and associated data.
 *
 * @f: File to send command on
 * @command: Command type to send
 * @len: Length of associated data
 * @data: Data associated with command.
 */
void qemu_savevm_command_send(QEMUFile *f,
                              enum qemu_vm_cmd command,
                              uint16_t len,
                              uint8_t *data)
{
    trace_savevm_command_send(command, len);
    qemu_put_byte(f, QEMU_VM_COMMAND);
    qemu_put_be16(f, (uint16_t)command);
    qemu_put_be16(f, len);
    qemu_put_buffer(f, data, len);
    qemu_fflush(f);
}

729 730 731 732 733 734 735 736 737 738 739 740 741 742 743
void qemu_savevm_send_ping(QEMUFile *f, uint32_t value)
{
    uint32_t buf;

    trace_savevm_send_ping(value);
    buf = cpu_to_be32(value);
    qemu_savevm_command_send(f, MIG_CMD_PING, sizeof(value), (uint8_t *)&buf);
}

void qemu_savevm_send_open_return_path(QEMUFile *f)
{
    trace_savevm_send_open_return_path();
    qemu_savevm_command_send(f, MIG_CMD_OPEN_RETURN_PATH, 0, NULL);
}

L
Luiz Capitulino 已提交
744
bool qemu_savevm_state_blocked(Error **errp)
A
Alex Williamson 已提交
745 746 747
{
    SaveStateEntry *se;

J
Juan Quintela 已提交
748
    QTAILQ_FOREACH(se, &savevm_state.handlers, entry) {
749
        if (se->vmsd && se->vmsd->unmigratable) {
750 751
            error_setg(errp, "State blocked by non-migratable device '%s'",
                       se->idstr);
A
Alex Williamson 已提交
752 753 754 755 756 757
            return true;
        }
    }
    return false;
}

758 759 760 761 762
void qemu_savevm_state_header(QEMUFile *f)
{
    trace_savevm_state_header();
    qemu_put_be32(f, QEMU_VM_FILE_MAGIC);
    qemu_put_be32(f, QEMU_VM_FILE_VERSION);
763 764 765 766 767 768

    if (!savevm_state.skip_configuration) {
        qemu_put_byte(f, QEMU_VM_CONFIGURATION);
        vmstate_save_state(f, &vmstate_configuration, &savevm_state, 0);
    }

769 770
}

771 772
void qemu_savevm_state_begin(QEMUFile *f,
                             const MigrationParams *params)
A
aliguori 已提交
773 774
{
    SaveStateEntry *se;
J
Juan Quintela 已提交
775
    int ret;
A
aliguori 已提交
776

777
    trace_savevm_state_begin();
J
Juan Quintela 已提交
778
    QTAILQ_FOREACH(se, &savevm_state.handlers, entry) {
779
        if (!se->ops || !se->ops->set_params) {
L
lirans@il.ibm.com 已提交
780
            continue;
I
Isaku Yamahata 已提交
781
        }
782
        se->ops->set_params(params, se->opaque);
L
lirans@il.ibm.com 已提交
783
    }
E
Eduardo Habkost 已提交
784

J
Juan Quintela 已提交
785
    QTAILQ_FOREACH(se, &savevm_state.handlers, entry) {
786
        if (!se->ops || !se->ops->save_live_setup) {
A
aliguori 已提交
787
            continue;
788
        }
789 790 791 792 793
        if (se->ops && se->ops->is_active) {
            if (!se->ops->is_active(se->opaque)) {
                continue;
            }
        }
794
        save_section_header(f, se, QEMU_VM_SECTION_START);
A
aliguori 已提交
795

796
        ret = se->ops->save_live_setup(f, se->opaque);
797
        save_section_footer(f, se);
798
        if (ret < 0) {
799 800
            qemu_file_set_error(f, ret);
            break;
801
        }
A
aliguori 已提交
802 803 804
    }
}

J
Juan Quintela 已提交
805
/*
D
Dong Xu Wang 已提交
806
 * this function has three return values:
J
Juan Quintela 已提交
807 808 809 810
 *   negative: there was one error, and we have -errno.
 *   0 : We haven't finished, caller have to go again
 *   1 : We have finished, we can go to complete phase
 */
811
int qemu_savevm_state_iterate(QEMUFile *f)
A
aliguori 已提交
812 813 814 815
{
    SaveStateEntry *se;
    int ret = 1;

816
    trace_savevm_state_iterate();
J
Juan Quintela 已提交
817
    QTAILQ_FOREACH(se, &savevm_state.handlers, entry) {
818
        if (!se->ops || !se->ops->save_live_iterate) {
A
aliguori 已提交
819
            continue;
820
        }
821 822 823 824 825
        if (se->ops && se->ops->is_active) {
            if (!se->ops->is_active(se->opaque)) {
                continue;
            }
        }
826 827 828
        if (qemu_file_rate_limit(f)) {
            return 0;
        }
829
        trace_savevm_section_start(se->idstr, se->section_id);
830 831

        save_section_header(f, se, QEMU_VM_SECTION_PART);
A
aliguori 已提交
832

833
        ret = se->ops->save_live_iterate(f, se->opaque);
834
        trace_savevm_section_end(se->idstr, se->section_id, ret);
835
        save_section_footer(f, se);
836

837 838 839
        if (ret < 0) {
            qemu_file_set_error(f, ret);
        }
840
        if (ret <= 0) {
841 842 843 844 845 846
            /* Do not proceed to the next vmstate before this one reported
               completion of the current stage. This serializes the migration
               and reduces the probability that a faster changing state is
               synchronized over and over again. */
            break;
        }
A
aliguori 已提交
847
    }
J
Juan Quintela 已提交
848
    return ret;
A
aliguori 已提交
849 850
}

851 852 853 854 855 856
static bool should_send_vmdesc(void)
{
    MachineState *machine = MACHINE(qdev_get_machine());
    return !machine->suppress_vmdesc;
}

857
void qemu_savevm_state_complete_precopy(QEMUFile *f)
A
aliguori 已提交
858
{
859 860
    QJSON *vmdesc;
    int vmdesc_len;
A
aliguori 已提交
861
    SaveStateEntry *se;
862
    int ret;
A
aliguori 已提交
863

864
    trace_savevm_state_complete_precopy();
865

866 867
    cpu_synchronize_all_states();

J
Juan Quintela 已提交
868
    QTAILQ_FOREACH(se, &savevm_state.handlers, entry) {
869
        if (!se->ops || !se->ops->save_live_complete_precopy) {
A
aliguori 已提交
870
            continue;
871
        }
872 873 874 875 876
        if (se->ops && se->ops->is_active) {
            if (!se->ops->is_active(se->opaque)) {
                continue;
            }
        }
877
        trace_savevm_section_start(se->idstr, se->section_id);
878 879

        save_section_header(f, se, QEMU_VM_SECTION_END);
A
aliguori 已提交
880

881
        ret = se->ops->save_live_complete_precopy(f, se->opaque);
882
        trace_savevm_section_end(se->idstr, se->section_id, ret);
883
        save_section_footer(f, se);
884
        if (ret < 0) {
885 886
            qemu_file_set_error(f, ret);
            return;
887
        }
A
aliguori 已提交
888 889
    }

890 891 892
    vmdesc = qjson_new();
    json_prop_int(vmdesc, "page_size", TARGET_PAGE_SIZE);
    json_start_array(vmdesc, "devices");
J
Juan Quintela 已提交
893
    QTAILQ_FOREACH(se, &savevm_state.handlers, entry) {
A
aliguori 已提交
894

895
        if ((!se->ops || !se->ops->save_state) && !se->vmsd) {
896
            continue;
897
        }
898 899 900 901 902
        if (se->vmsd && !vmstate_save_needed(se->vmsd, se->opaque)) {
            trace_savevm_section_skip(se->idstr, se->section_id);
            continue;
        }

903
        trace_savevm_section_start(se->idstr, se->section_id);
904 905 906 907 908

        json_start_object(vmdesc, NULL);
        json_prop_str(vmdesc, "name", se->idstr);
        json_prop_int(vmdesc, "instance_id", se->instance_id);

909
        save_section_header(f, se, QEMU_VM_SECTION_FULL);
A
aliguori 已提交
910

911 912 913
        vmstate_save(f, se, vmdesc);

        json_end_object(vmdesc);
914
        trace_savevm_section_end(se->idstr, se->section_id, 0);
915
        save_section_footer(f, se);
A
aliguori 已提交
916 917 918
    }

    qemu_put_byte(f, QEMU_VM_EOF);
919 920 921 922 923

    json_end_array(vmdesc);
    qjson_finish(vmdesc);
    vmdesc_len = strlen(qjson_get_str(vmdesc));

924 925 926 927 928
    if (should_send_vmdesc()) {
        qemu_put_byte(f, QEMU_VM_VMDESCRIPTION);
        qemu_put_be32(f, vmdesc_len);
        qemu_put_buffer(f, (uint8_t *)qjson_get_str(vmdesc), vmdesc_len);
    }
929 930
    object_unref(OBJECT(vmdesc));

931
    qemu_fflush(f);
A
aliguori 已提交
932 933
}

934 935 936 937 938
uint64_t qemu_savevm_state_pending(QEMUFile *f, uint64_t max_size)
{
    SaveStateEntry *se;
    uint64_t ret = 0;

J
Juan Quintela 已提交
939
    QTAILQ_FOREACH(se, &savevm_state.handlers, entry) {
940 941 942 943 944 945 946 947 948 949 950 951 952
        if (!se->ops || !se->ops->save_live_pending) {
            continue;
        }
        if (se->ops && se->ops->is_active) {
            if (!se->ops->is_active(se->opaque)) {
                continue;
            }
        }
        ret += se->ops->save_live_pending(f, se->opaque, max_size);
    }
    return ret;
}

953
void qemu_savevm_state_cleanup(void)
954 955 956
{
    SaveStateEntry *se;

957
    trace_savevm_state_cleanup();
J
Juan Quintela 已提交
958
    QTAILQ_FOREACH(se, &savevm_state.handlers, entry) {
959 960
        if (se->ops && se->ops->cleanup) {
            se->ops->cleanup(se->opaque);
961 962 963 964
        }
    }
}

965
static int qemu_savevm_state(QEMUFile *f, Error **errp)
A
aliguori 已提交
966 967
{
    int ret;
I
Isaku Yamahata 已提交
968 969 970 971
    MigrationParams params = {
        .blk = 0,
        .shared = 0
    };
972 973
    MigrationState *ms = migrate_init(&params);
    ms->file = f;
A
aliguori 已提交
974

975
    if (qemu_savevm_state_blocked(errp)) {
976
        return -EINVAL;
A
Alex Williamson 已提交
977 978
    }

979
    qemu_mutex_unlock_iothread();
980
    qemu_savevm_state_header(f);
981
    qemu_savevm_state_begin(f, &params);
982 983
    qemu_mutex_lock_iothread();

984 985 986 987 988
    while (qemu_file_get_error(f) == 0) {
        if (qemu_savevm_state_iterate(f) > 0) {
            break;
        }
    }
A
aliguori 已提交
989

990
    ret = qemu_file_get_error(f);
J
Juan Quintela 已提交
991
    if (ret == 0) {
992
        qemu_savevm_state_complete_precopy(f);
993
        ret = qemu_file_get_error(f);
J
Juan Quintela 已提交
994
    }
995
    if (ret != 0) {
996
        qemu_savevm_state_cleanup();
997
        error_setg_errno(errp, -ret, "Error while writing VM state");
998
    }
A
aliguori 已提交
999 1000 1001
    return ret;
}

1002 1003 1004 1005 1006 1007 1008 1009 1010
static int qemu_save_device_state(QEMUFile *f)
{
    SaveStateEntry *se;

    qemu_put_be32(f, QEMU_VM_FILE_MAGIC);
    qemu_put_be32(f, QEMU_VM_FILE_VERSION);

    cpu_synchronize_all_states();

J
Juan Quintela 已提交
1011
    QTAILQ_FOREACH(se, &savevm_state.handlers, entry) {
1012 1013 1014
        if (se->is_ram) {
            continue;
        }
1015
        if ((!se->ops || !se->ops->save_state) && !se->vmsd) {
1016 1017
            continue;
        }
1018 1019 1020
        if (se->vmsd && !vmstate_save_needed(se->vmsd, se->opaque)) {
            continue;
        }
1021

1022
        save_section_header(f, se, QEMU_VM_SECTION_FULL);
1023

1024
        vmstate_save(f, se, NULL);
1025 1026

        save_section_footer(f, se);
1027 1028 1029 1030 1031 1032 1033
    }

    qemu_put_byte(f, QEMU_VM_EOF);

    return qemu_file_get_error(f);
}

A
aliguori 已提交
1034 1035 1036 1037
static SaveStateEntry *find_se(const char *idstr, int instance_id)
{
    SaveStateEntry *se;

J
Juan Quintela 已提交
1038
    QTAILQ_FOREACH(se, &savevm_state.handlers, entry) {
A
aliguori 已提交
1039
        if (!strcmp(se->idstr, idstr) &&
J
Jan Kiszka 已提交
1040 1041
            (instance_id == se->instance_id ||
             instance_id == se->alias_id))
A
aliguori 已提交
1042
            return se;
1043 1044 1045 1046 1047 1048 1049
        /* Migrating from an older version? */
        if (strstr(se->idstr, idstr) && se->compat) {
            if (!strcmp(se->compat->idstr, idstr) &&
                (instance_id == se->compat->instance_id ||
                 instance_id == se->alias_id))
                return se;
        }
A
aliguori 已提交
1050 1051 1052 1053
    }
    return NULL;
}

D
Dr. David Alan Gilbert 已提交
1054 1055 1056 1057 1058 1059 1060 1061 1062
/**
 * loadvm_process_command: Process an incoming 'QEMU_VM_COMMAND'
 *
 * Returns: 0 on success, negative on error (in which case it will issue an
 *          error message).
 * @f: The stream to read the command data from.
 */
static int loadvm_process_command(QEMUFile *f)
{
1063
    MigrationIncomingState *mis = migration_incoming_get_current();
D
Dr. David Alan Gilbert 已提交
1064 1065
    uint16_t cmd;
    uint16_t len;
1066
    uint32_t tmp32;
D
Dr. David Alan Gilbert 已提交
1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084

    cmd = qemu_get_be16(f);
    len = qemu_get_be16(f);

    trace_loadvm_process_command(cmd, len);
    if (cmd >= MIG_CMD_MAX || cmd == MIG_CMD_INVALID) {
        error_report("MIG_CMD 0x%x unknown (len 0x%x)", cmd, len);
        return -EINVAL;
    }

    if (mig_cmd_args[cmd].len != -1 && mig_cmd_args[cmd].len != len) {
        error_report("%s received with bad length - expecting %zu, got %d",
                     mig_cmd_args[cmd].name,
                     (size_t)mig_cmd_args[cmd].len, len);
        return -ERANGE;
    }

    switch (cmd) {
1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107
    case MIG_CMD_OPEN_RETURN_PATH:
        if (mis->to_src_file) {
            error_report("CMD_OPEN_RETURN_PATH called when RP already open");
            /* Not really a problem, so don't give up */
            return 0;
        }
        mis->to_src_file = qemu_file_get_return_path(f);
        if (!mis->to_src_file) {
            error_report("CMD_OPEN_RETURN_PATH failed");
            return -1;
        }
        break;

    case MIG_CMD_PING:
        tmp32 = qemu_get_be32(f);
        trace_loadvm_process_command_ping(tmp32);
        if (!mis->to_src_file) {
            error_report("CMD_PING (0x%x) received with no return path",
                         tmp32);
            return -1;
        }
        /* migrate_send_rp_pong(mis, tmp32); TODO: gets added later */
        break;
D
Dr. David Alan Gilbert 已提交
1108 1109 1110 1111 1112
    }

    return 0;
}

1113
struct LoadStateEntry {
B
Blue Swirl 已提交
1114
    QLIST_ENTRY(LoadStateEntry) entry;
A
aliguori 已提交
1115 1116 1117
    SaveStateEntry *se;
    int section_id;
    int version_id;
1118
};
A
aliguori 已提交
1119

1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154
/*
 * Read a footer off the wire and check that it matches the expected section
 *
 * Returns: true if the footer was good
 *          false if there is a problem (and calls error_report to say why)
 */
static bool check_section_footer(QEMUFile *f, LoadStateEntry *le)
{
    uint8_t read_mark;
    uint32_t read_section_id;

    if (skip_section_footers) {
        /* No footer to check */
        return true;
    }

    read_mark = qemu_get_byte(f);

    if (read_mark != QEMU_VM_SECTION_FOOTER) {
        error_report("Missing section footer for %s", le->se->idstr);
        return false;
    }

    read_section_id = qemu_get_be32(f);
    if (read_section_id != le->section_id) {
        error_report("Mismatched section id in footer for %s -"
                     " read 0x%x expected 0x%x",
                     le->se->idstr, read_section_id, le->section_id);
        return false;
    }

    /* All good */
    return true;
}

1155
void loadvm_free_handlers(MigrationIncomingState *mis)
A
aliguori 已提交
1156
{
1157
    LoadStateEntry *le, *new_le;
1158 1159 1160 1161 1162 1163 1164 1165 1166 1167

    QLIST_FOREACH_SAFE(le, &mis->loadvm_handlers, entry, new_le) {
        QLIST_REMOVE(le, entry);
        g_free(le);
    }
}

int qemu_loadvm_state(QEMUFile *f)
{
    MigrationIncomingState *mis = migration_incoming_get_current();
1168
    Error *local_err = NULL;
A
aliguori 已提交
1169 1170 1171
    uint8_t section_type;
    unsigned int v;
    int ret;
1172
    int file_error_after_eof = -1;
A
aliguori 已提交
1173

1174
    if (qemu_savevm_state_blocked(&local_err)) {
1175
        error_report_err(local_err);
A
Alex Williamson 已提交
1176 1177 1178
        return -EINVAL;
    }

A
aliguori 已提交
1179
    v = qemu_get_be32(f);
E
Eduardo Habkost 已提交
1180
    if (v != QEMU_VM_FILE_MAGIC) {
1181
        error_report("Not a migration stream");
A
aliguori 已提交
1182
        return -EINVAL;
E
Eduardo Habkost 已提交
1183
    }
A
aliguori 已提交
1184 1185

    v = qemu_get_be32(f);
J
Juan Quintela 已提交
1186
    if (v == QEMU_VM_FILE_VERSION_COMPAT) {
1187
        error_report("SaveVM v2 format is obsolete and don't work anymore");
J
Juan Quintela 已提交
1188 1189
        return -ENOTSUP;
    }
E
Eduardo Habkost 已提交
1190
    if (v != QEMU_VM_FILE_VERSION) {
1191
        error_report("Unsupported migration stream version");
A
aliguori 已提交
1192
        return -ENOTSUP;
E
Eduardo Habkost 已提交
1193
    }
A
aliguori 已提交
1194

1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206
    if (!savevm_state.skip_configuration) {
        if (qemu_get_byte(f) != QEMU_VM_CONFIGURATION) {
            error_report("Configuration section missing");
            return -EINVAL;
        }
        ret = vmstate_load_state(f, &vmstate_configuration, &savevm_state, 0);

        if (ret) {
            return ret;
        }
    }

A
aliguori 已提交
1207 1208 1209
    while ((section_type = qemu_get_byte(f)) != QEMU_VM_EOF) {
        uint32_t instance_id, version_id, section_id;
        SaveStateEntry *se;
1210
        LoadStateEntry *le;
1211
        char idstr[256];
A
aliguori 已提交
1212

1213
        trace_qemu_loadvm_state_section(section_type);
A
aliguori 已提交
1214 1215 1216 1217 1218
        switch (section_type) {
        case QEMU_VM_SECTION_START:
        case QEMU_VM_SECTION_FULL:
            /* Read section start */
            section_id = qemu_get_be32(f);
1219 1220 1221 1222 1223
            if (!qemu_get_counted_string(f, idstr)) {
                error_report("Unable to read ID string for section %u",
                            section_id);
                return -EINVAL;
            }
A
aliguori 已提交
1224 1225 1226
            instance_id = qemu_get_be32(f);
            version_id = qemu_get_be32(f);

1227 1228
            trace_qemu_loadvm_state_section_startfull(section_id, idstr,
                                                      instance_id, version_id);
A
aliguori 已提交
1229 1230 1231
            /* Find savevm section */
            se = find_se(idstr, instance_id);
            if (se == NULL) {
1232 1233
                error_report("Unknown savevm section or instance '%s' %d",
                             idstr, instance_id);
A
aliguori 已提交
1234 1235 1236 1237 1238 1239
                ret = -EINVAL;
                goto out;
            }

            /* Validate version */
            if (version_id > se->version_id) {
1240 1241
                error_report("savevm: unsupported version %d for '%s' v%d",
                             version_id, idstr, se->version_id);
A
aliguori 已提交
1242 1243 1244 1245 1246
                ret = -EINVAL;
                goto out;
            }

            /* Add entry */
1247
            le = g_malloc0(sizeof(*le));
A
aliguori 已提交
1248 1249 1250 1251

            le->se = se;
            le->section_id = section_id;
            le->version_id = version_id;
1252
            QLIST_INSERT_HEAD(&mis->loadvm_handlers, le, entry);
A
aliguori 已提交
1253

1254
            ret = vmstate_load(f, le->se, le->version_id);
1255
            if (ret < 0) {
1256 1257
                error_report("error while loading state for instance 0x%x of"
                             " device '%s'", instance_id, idstr);
1258 1259
                goto out;
            }
1260
            if (!check_section_footer(f, le)) {
1261 1262 1263
                ret = -EINVAL;
                goto out;
            }
A
aliguori 已提交
1264 1265 1266 1267 1268
            break;
        case QEMU_VM_SECTION_PART:
        case QEMU_VM_SECTION_END:
            section_id = qemu_get_be32(f);

1269
            trace_qemu_loadvm_state_section_partend(section_id);
1270
            QLIST_FOREACH(le, &mis->loadvm_handlers, entry) {
1271 1272 1273 1274
                if (le->section_id == section_id) {
                    break;
                }
            }
A
aliguori 已提交
1275
            if (le == NULL) {
1276
                error_report("Unknown savevm section %d", section_id);
A
aliguori 已提交
1277 1278 1279 1280
                ret = -EINVAL;
                goto out;
            }

1281
            ret = vmstate_load(f, le->se, le->version_id);
1282
            if (ret < 0) {
1283 1284
                error_report("error while loading state section id %d(%s)",
                             section_id, le->se->idstr);
1285 1286
                goto out;
            }
1287
            if (!check_section_footer(f, le)) {
1288 1289 1290
                ret = -EINVAL;
                goto out;
            }
A
aliguori 已提交
1291
            break;
D
Dr. David Alan Gilbert 已提交
1292 1293 1294 1295 1296 1297
        case QEMU_VM_COMMAND:
            ret = loadvm_process_command(f);
            if (ret < 0) {
                goto out;
            }
            break;
A
aliguori 已提交
1298
        default:
1299
            error_report("Unknown savevm section type %d", section_type);
A
aliguori 已提交
1300 1301 1302 1303 1304
            ret = -EINVAL;
            goto out;
        }
    }

1305 1306 1307 1308 1309 1310
    file_error_after_eof = qemu_file_get_error(f);

    /*
     * Try to read in the VMDESC section as well, so that dumping tools that
     * intercept our migration stream have the chance to see it.
     */
1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338

    /* We've got to be careful; if we don't read the data and just shut the fd
     * then the sender can error if we close while it's still sending.
     * We also mustn't read data that isn't there; some transports (RDMA)
     * will stall waiting for that data when the source has already closed.
     */
    if (should_send_vmdesc()) {
        uint8_t *buf;
        uint32_t size;
        section_type = qemu_get_byte(f);

        if (section_type != QEMU_VM_VMDESCRIPTION) {
            error_report("Expected vmdescription section, but got %d",
                         section_type);
            /*
             * It doesn't seem worth failing at this point since
             * we apparently have an otherwise valid VM state
             */
        } else {
            buf = g_malloc(0x1000);
            size = qemu_get_be32(f);

            while (size > 0) {
                uint32_t read_chunk = MIN(size, 0x1000);
                qemu_get_buffer(f, buf, read_chunk);
                size -= read_chunk;
            }
            g_free(buf);
1339 1340 1341
        }
    }

1342 1343
    cpu_synchronize_all_post_init();

A
aliguori 已提交
1344 1345 1346
    ret = 0;

out:
1347
    if (ret == 0) {
1348 1349
        /* We may not have a VMDESC section, so ignore relative errors */
        ret = file_error_after_eof;
1350
    }
A
aliguori 已提交
1351 1352 1353 1354

    return ret;
}

1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365
static BlockDriverState *find_vmstate_bs(void)
{
    BlockDriverState *bs = NULL;
    while ((bs = bdrv_next(bs))) {
        if (bdrv_can_snapshot(bs)) {
            return bs;
        }
    }
    return NULL;
}

1366 1367 1368 1369 1370 1371 1372
/*
 * Deletes snapshots of a given name in all opened images.
 */
static int del_existing_snapshots(Monitor *mon, const char *name)
{
    BlockDriverState *bs;
    QEMUSnapshotInfo sn1, *snapshot = &sn1;
1373
    Error *err = NULL;
1374

1375 1376
    bs = NULL;
    while ((bs = bdrv_next(bs))) {
1377
        if (bdrv_can_snapshot(bs) &&
E
Eduardo Habkost 已提交
1378
            bdrv_snapshot_find(bs, snapshot, name) >= 0) {
1379
            bdrv_snapshot_delete_by_id_or_name(bs, name, &err);
1380
            if (err) {
1381
                monitor_printf(mon,
1382 1383 1384 1385 1386
                               "Error while deleting snapshot on device '%s':"
                               " %s\n",
                               bdrv_get_device_name(bs),
                               error_get_pretty(err));
                error_free(err);
1387 1388 1389 1390 1391 1392 1393 1394
                return -1;
            }
        }
    }

    return 0;
}

1395
void hmp_savevm(Monitor *mon, const QDict *qdict)
A
aliguori 已提交
1396 1397 1398
{
    BlockDriverState *bs, *bs1;
    QEMUSnapshotInfo sn1, *sn = &sn1, old_sn1, *old_sn = &old_sn1;
1399
    int ret;
A
aliguori 已提交
1400 1401
    QEMUFile *f;
    int saved_vm_running;
K
Kevin Wolf 已提交
1402
    uint64_t vm_state_size;
1403
    qemu_timeval tv;
1404
    struct tm tm;
1405
    const char *name = qdict_get_try_str(qdict, "name");
1406
    Error *local_err = NULL;
A
aliguori 已提交
1407

1408
    /* Verify if there is a device that doesn't support snapshots and is writable */
1409 1410
    bs = NULL;
    while ((bs = bdrv_next(bs))) {
1411

1412
        if (!bdrv_is_inserted(bs) || bdrv_is_read_only(bs)) {
1413 1414 1415 1416 1417 1418 1419 1420 1421 1422
            continue;
        }

        if (!bdrv_can_snapshot(bs)) {
            monitor_printf(mon, "Device '%s' is writable but does not support snapshots.\n",
                               bdrv_get_device_name(bs));
            return;
        }
    }

1423
    bs = find_vmstate_bs();
A
aliguori 已提交
1424
    if (!bs) {
A
aliguori 已提交
1425
        monitor_printf(mon, "No block device can accept snapshots\n");
A
aliguori 已提交
1426 1427 1428
        return;
    }

1429
    saved_vm_running = runstate_is_running();
1430 1431 1432 1433 1434 1435

    ret = global_state_store();
    if (ret) {
        monitor_printf(mon, "Error saving global state\n");
        return;
    }
1436
    vm_stop(RUN_STATE_SAVE_VM);
A
aliguori 已提交
1437

1438
    memset(sn, 0, sizeof(*sn));
A
aliguori 已提交
1439 1440

    /* fill auxiliary fields */
1441
    qemu_gettimeofday(&tv);
A
aliguori 已提交
1442 1443
    sn->date_sec = tv.tv_sec;
    sn->date_nsec = tv.tv_usec * 1000;
1444
    sn->vm_clock_nsec = qemu_clock_get_ns(QEMU_CLOCK_VIRTUAL);
A
aliguori 已提交
1445

1446 1447 1448 1449 1450 1451 1452 1453 1454
    if (name) {
        ret = bdrv_snapshot_find(bs, old_sn, name);
        if (ret >= 0) {
            pstrcpy(sn->name, sizeof(sn->name), old_sn->name);
            pstrcpy(sn->id_str, sizeof(sn->id_str), old_sn->id_str);
        } else {
            pstrcpy(sn->name, sizeof(sn->name), name);
        }
    } else {
B
Blue Swirl 已提交
1455 1456
        /* cast below needed for OpenBSD where tv_sec is still 'long' */
        localtime_r((const time_t *)&tv.tv_sec, &tm);
1457 1458 1459
        strftime(sn->name, sizeof(sn->name), "vm-%Y%m%d%H%M%S", &tm);
    }

1460
    /* Delete old snapshots of the same name */
1461
    if (name && del_existing_snapshots(mon, name) < 0) {
1462 1463 1464
        goto the_end;
    }

A
aliguori 已提交
1465
    /* save the VM state */
1466
    f = qemu_fopen_bdrv(bs, 1);
A
aliguori 已提交
1467
    if (!f) {
A
aliguori 已提交
1468
        monitor_printf(mon, "Could not open VM state file\n");
A
aliguori 已提交
1469 1470
        goto the_end;
    }
1471
    ret = qemu_savevm_state(f, &local_err);
1472
    vm_state_size = qemu_ftell(f);
A
aliguori 已提交
1473 1474
    qemu_fclose(f);
    if (ret < 0) {
1475 1476
        monitor_printf(mon, "%s\n", error_get_pretty(local_err));
        error_free(local_err);
A
aliguori 已提交
1477 1478 1479 1480 1481
        goto the_end;
    }

    /* create the snapshots */

1482 1483
    bs1 = NULL;
    while ((bs1 = bdrv_next(bs1))) {
1484
        if (bdrv_can_snapshot(bs1)) {
1485 1486
            /* Write VM state size only to the image that contains the state */
            sn->vm_state_size = (bs == bs1 ? vm_state_size : 0);
A
aliguori 已提交
1487 1488
            ret = bdrv_snapshot_create(bs1, sn);
            if (ret < 0) {
A
aliguori 已提交
1489 1490
                monitor_printf(mon, "Error while creating snapshot on '%s'\n",
                               bdrv_get_device_name(bs1));
A
aliguori 已提交
1491 1492 1493 1494 1495
            }
        }
    }

 the_end:
E
Eduardo Habkost 已提交
1496
    if (saved_vm_running) {
A
aliguori 已提交
1497
        vm_start();
E
Eduardo Habkost 已提交
1498
    }
A
aliguori 已提交
1499 1500
}

1501 1502 1503 1504 1505 1506 1507 1508
void qmp_xen_save_devices_state(const char *filename, Error **errp)
{
    QEMUFile *f;
    int saved_vm_running;
    int ret;

    saved_vm_running = runstate_is_running();
    vm_stop(RUN_STATE_SAVE_VM);
1509
    global_state_store_running();
1510 1511 1512

    f = qemu_fopen(filename, "wb");
    if (!f) {
1513
        error_setg_file_open(errp, errno, filename);
1514 1515 1516 1517 1518
        goto the_end;
    }
    ret = qemu_save_device_state(f);
    qemu_fclose(f);
    if (ret < 0) {
1519
        error_setg(errp, QERR_IO_ERROR);
1520 1521 1522
    }

 the_end:
E
Eduardo Habkost 已提交
1523
    if (saved_vm_running) {
1524
        vm_start();
E
Eduardo Habkost 已提交
1525
    }
1526 1527
}

1528
int load_vmstate(const char *name)
A
aliguori 已提交
1529
{
1530
    BlockDriverState *bs, *bs_vm_state;
1531
    QEMUSnapshotInfo sn;
A
aliguori 已提交
1532
    QEMUFile *f;
G
Gerd Hoffmann 已提交
1533
    int ret;
A
aliguori 已提交
1534

1535
    bs_vm_state = find_vmstate_bs();
1536 1537 1538 1539 1540 1541 1542 1543 1544 1545
    if (!bs_vm_state) {
        error_report("No block device supports snapshots");
        return -ENOTSUP;
    }

    /* Don't even try to load empty VM states */
    ret = bdrv_snapshot_find(bs_vm_state, &sn, name);
    if (ret < 0) {
        return ret;
    } else if (sn.vm_state_size == 0) {
1546 1547
        error_report("This is a disk-only snapshot. Revert to it offline "
            "using qemu-img.");
1548 1549 1550 1551 1552
        return -EINVAL;
    }

    /* Verify if there is any device that doesn't support snapshots and is
    writable and check if the requested snapshot is available too. */
1553 1554
    bs = NULL;
    while ((bs = bdrv_next(bs))) {
1555

1556
        if (!bdrv_is_inserted(bs) || bdrv_is_read_only(bs)) {
1557 1558 1559 1560 1561 1562 1563 1564 1565
            continue;
        }

        if (!bdrv_can_snapshot(bs)) {
            error_report("Device '%s' is writable but does not support snapshots.",
                               bdrv_get_device_name(bs));
            return -ENOTSUP;
        }

1566 1567 1568 1569 1570 1571
        ret = bdrv_snapshot_find(bs, &sn, name);
        if (ret < 0) {
            error_report("Device '%s' does not have the requested snapshot '%s'",
                           bdrv_get_device_name(bs), name);
            return ret;
        }
A
aliguori 已提交
1572 1573 1574
    }

    /* Flush all IO requests so they don't interfere with the new state.  */
1575
    bdrv_drain_all();
A
aliguori 已提交
1576

1577 1578 1579 1580
    bs = NULL;
    while ((bs = bdrv_next(bs))) {
        if (bdrv_can_snapshot(bs)) {
            ret = bdrv_snapshot_goto(bs, name);
A
aliguori 已提交
1581
            if (ret < 0) {
1582 1583 1584
                error_report("Error %d while activating snapshot '%s' on '%s'",
                             ret, name, bdrv_get_device_name(bs));
                return ret;
A
aliguori 已提交
1585 1586 1587 1588 1589
            }
        }
    }

    /* restore the VM state */
1590
    f = qemu_fopen_bdrv(bs_vm_state, 0);
A
aliguori 已提交
1591
    if (!f) {
1592
        error_report("Could not open VM state file");
1593
        return -EINVAL;
A
aliguori 已提交
1594
    }
1595

J
Jan Kiszka 已提交
1596
    qemu_system_reset(VMRESET_SILENT);
1597
    migration_incoming_state_new(f);
A
aliguori 已提交
1598
    ret = qemu_loadvm_state(f);
1599

A
aliguori 已提交
1600
    qemu_fclose(f);
1601
    migration_incoming_state_destroy();
A
aliguori 已提交
1602
    if (ret < 0) {
1603
        error_report("Error %d while loading VM state", ret);
1604
        return ret;
A
aliguori 已提交
1605
    }
1606

1607
    return 0;
1608 1609
}

1610
void hmp_delvm(Monitor *mon, const QDict *qdict)
A
aliguori 已提交
1611
{
1612
    BlockDriverState *bs;
1613
    Error *err;
1614
    const char *name = qdict_get_str(qdict, "name");
A
aliguori 已提交
1615

1616
    if (!find_vmstate_bs()) {
A
aliguori 已提交
1617
        monitor_printf(mon, "No block device supports snapshots\n");
A
aliguori 已提交
1618 1619 1620
        return;
    }

1621 1622 1623
    bs = NULL;
    while ((bs = bdrv_next(bs))) {
        if (bdrv_can_snapshot(bs)) {
1624
            err = NULL;
1625
            bdrv_snapshot_delete_by_id_or_name(bs, name, &err);
1626
            if (err) {
1627 1628 1629 1630 1631 1632
                monitor_printf(mon,
                               "Error while deleting snapshot on device '%s':"
                               " %s\n",
                               bdrv_get_device_name(bs),
                               error_get_pretty(err));
                error_free(err);
A
aliguori 已提交
1633 1634 1635 1636 1637
            }
        }
    }
}

1638
void hmp_info_snapshots(Monitor *mon, const QDict *qdict)
A
aliguori 已提交
1639 1640
{
    BlockDriverState *bs, *bs1;
1641 1642 1643 1644
    QEMUSnapshotInfo *sn_tab, *sn, s, *sn_info = &s;
    int nb_sns, i, ret, available;
    int total;
    int *available_snapshots;
A
aliguori 已提交
1645

1646
    bs = find_vmstate_bs();
A
aliguori 已提交
1647
    if (!bs) {
A
aliguori 已提交
1648
        monitor_printf(mon, "No available block device supports snapshots\n");
A
aliguori 已提交
1649 1650 1651 1652 1653
        return;
    }

    nb_sns = bdrv_snapshot_list(bs, &sn_tab);
    if (nb_sns < 0) {
A
aliguori 已提交
1654
        monitor_printf(mon, "bdrv_snapshot_list: error %d\n", nb_sns);
A
aliguori 已提交
1655 1656
        return;
    }
1657 1658 1659 1660 1661 1662

    if (nb_sns == 0) {
        monitor_printf(mon, "There is no snapshot available.\n");
        return;
    }

1663
    available_snapshots = g_new0(int, nb_sns);
1664 1665
    total = 0;
    for (i = 0; i < nb_sns; i++) {
A
aliguori 已提交
1666
        sn = &sn_tab[i];
1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683
        available = 1;
        bs1 = NULL;

        while ((bs1 = bdrv_next(bs1))) {
            if (bdrv_can_snapshot(bs1) && bs1 != bs) {
                ret = bdrv_snapshot_find(bs1, sn_info, sn->id_str);
                if (ret < 0) {
                    available = 0;
                    break;
                }
            }
        }

        if (available) {
            available_snapshots[total] = i;
            total++;
        }
A
aliguori 已提交
1684
    }
1685 1686

    if (total > 0) {
1687 1688
        bdrv_snapshot_dump((fprintf_function)monitor_printf, mon, NULL);
        monitor_printf(mon, "\n");
1689 1690
        for (i = 0; i < total; i++) {
            sn = &sn_tab[available_snapshots[i]];
1691 1692
            bdrv_snapshot_dump((fprintf_function)monitor_printf, mon, sn);
            monitor_printf(mon, "\n");
1693 1694 1695 1696 1697
        }
    } else {
        monitor_printf(mon, "There is no suitable snapshot available\n");
    }

1698 1699
    g_free(sn_tab);
    g_free(available_snapshots);
1700

A
aliguori 已提交
1701
}
1702 1703 1704

void vmstate_register_ram(MemoryRegion *mr, DeviceState *dev)
{
1705
    qemu_ram_set_idstr(memory_region_get_ram_addr(mr) & TARGET_PAGE_MASK,
1706 1707 1708 1709 1710
                       memory_region_name(mr), dev);
}

void vmstate_unregister_ram(MemoryRegion *mr, DeviceState *dev)
{
1711
    qemu_ram_unset_idstr(memory_region_get_ram_addr(mr) & TARGET_PAGE_MASK);
1712 1713 1714 1715 1716 1717
}

void vmstate_register_ram_global(MemoryRegion *mr)
{
    vmstate_register_ram(mr, NULL);
}