qemu_migration.c 112.5 KB
Newer Older
1 2 3
/*
 * qemu_migration.c: QEMU migration handling
 *
M
Martin Kletzander 已提交
4
 * Copyright (C) 2006-2012 Red Hat, Inc.
5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
 *
 * This library is free software; you can redistribute it and/or
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
 * version 2.1 of the License, or (at your option) any later version.
 *
 * This library is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * Lesser General Public License for more details.
 *
 * You should have received a copy of the GNU Lesser General Public
 * License along with this library; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307  USA
 *
 */

#include <config.h>

#include <sys/time.h>
25 26
#include <gnutls/gnutls.h>
#include <gnutls/x509.h>
27
#include <fcntl.h>
28
#include <poll.h>
29 30 31 32 33 34

#include "qemu_migration.h"
#include "qemu_monitor.h"
#include "qemu_domain.h"
#include "qemu_process.h"
#include "qemu_capabilities.h"
35
#include "qemu_cgroup.h"
36

37
#include "domain_audit.h"
38 39 40 41
#include "logging.h"
#include "virterror_internal.h"
#include "memory.h"
#include "util.h"
E
Eric Blake 已提交
42
#include "virfile.h"
43 44
#include "datatypes.h"
#include "fdstream.h"
45
#include "uuid.h"
46
#include "virtime.h"
47
#include "locking/domain_lock.h"
48
#include "rpc/virnetsocket.h"
49
#include "storage_file.h"
M
Martin Kletzander 已提交
50
#include "viruri.h"
J
Jiri Denemark 已提交
51
#include "hooks.h"
52

53 54 55

#define VIR_FROM_THIS VIR_FROM_QEMU

56 57 58 59 60 61 62 63 64 65 66 67 68
VIR_ENUM_IMPL(qemuMigrationJobPhase, QEMU_MIGRATION_PHASE_LAST,
              "none",
              "perform2",
              "begin3",
              "perform3",
              "perform3_done",
              "confirm3_cancelled",
              "confirm3",
              "prepare",
              "finish2",
              "finish3",
);

69
enum qemuMigrationCookieFlags {
70
    QEMU_MIGRATION_COOKIE_FLAG_GRAPHICS,
71
    QEMU_MIGRATION_COOKIE_FLAG_LOCKSTATE,
72
    QEMU_MIGRATION_COOKIE_FLAG_PERSISTENT,
73 74 75 76 77 78 79

    QEMU_MIGRATION_COOKIE_FLAG_LAST
};

VIR_ENUM_DECL(qemuMigrationCookieFlag);
VIR_ENUM_IMPL(qemuMigrationCookieFlag,
              QEMU_MIGRATION_COOKIE_FLAG_LAST,
80
              "graphics", "lockstate", "persistent");
81 82 83

enum qemuMigrationCookieFeatures {
    QEMU_MIGRATION_COOKIE_GRAPHICS  = (1 << QEMU_MIGRATION_COOKIE_FLAG_GRAPHICS),
84
    QEMU_MIGRATION_COOKIE_LOCKSTATE = (1 << QEMU_MIGRATION_COOKIE_FLAG_LOCKSTATE),
85
    QEMU_MIGRATION_COOKIE_PERSISTENT = (1 << QEMU_MIGRATION_COOKIE_FLAG_PERSISTENT),
86 87 88 89 90 91 92 93 94 95 96 97
};

typedef struct _qemuMigrationCookieGraphics qemuMigrationCookieGraphics;
typedef qemuMigrationCookieGraphics *qemuMigrationCookieGraphicsPtr;
struct _qemuMigrationCookieGraphics {
    int type;
    int port;
    int tlsPort;
    char *listen;
    char *tlsSubject;
};

98 99 100
typedef struct _qemuMigrationCookie qemuMigrationCookie;
typedef qemuMigrationCookie *qemuMigrationCookiePtr;
struct _qemuMigrationCookie {
E
Eric Blake 已提交
101 102
    unsigned int flags;
    unsigned int flagsMandatory;
103 104

    /* Host properties */
105 106 107 108
    unsigned char localHostuuid[VIR_UUID_BUFLEN];
    unsigned char remoteHostuuid[VIR_UUID_BUFLEN];
    char *localHostname;
    char *remoteHostname;
109 110 111 112

    /* Guest properties */
    unsigned char uuid[VIR_UUID_BUFLEN];
    char *name;
113

114 115 116 117
    /* If (flags & QEMU_MIGRATION_COOKIE_LOCKSTATE) */
    char *lockState;
    char *lockDriver;

118 119
    /* If (flags & QEMU_MIGRATION_COOKIE_GRAPHICS) */
    qemuMigrationCookieGraphicsPtr graphics;
120 121 122

    /* If (flags & QEMU_MIGRATION_COOKIE_PERSISTENT) */
    virDomainDefPtr persistent;
123 124
};

125 126 127 128 129 130 131 132 133
static void qemuMigrationCookieGraphicsFree(qemuMigrationCookieGraphicsPtr grap)
{
    if (!grap)
        return;
    VIR_FREE(grap->listen);
    VIR_FREE(grap->tlsSubject);
    VIR_FREE(grap);
}

134 135 136 137 138 139

static void qemuMigrationCookieFree(qemuMigrationCookiePtr mig)
{
    if (!mig)
        return;

140 141 142
    if (mig->flags & QEMU_MIGRATION_COOKIE_GRAPHICS)
        qemuMigrationCookieGraphicsFree(mig->graphics);

143 144
    VIR_FREE(mig->localHostname);
    VIR_FREE(mig->remoteHostname);
145
    VIR_FREE(mig->name);
146 147
    VIR_FREE(mig->lockState);
    VIR_FREE(mig->lockDriver);
148 149 150 151
    VIR_FREE(mig);
}


152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224
static char *
qemuDomainExtractTLSSubject(const char *certdir)
{
    char *certfile = NULL;
    char *subject = NULL;
    char *pemdata = NULL;
    gnutls_datum_t pemdatum;
    gnutls_x509_crt_t cert;
    int ret;
    size_t subjectlen;

    if (virAsprintf(&certfile, "%s/server-cert.pem", certdir) < 0)
        goto no_memory;

    if (virFileReadAll(certfile, 8192, &pemdata) < 0) {
        qemuReportError(VIR_ERR_INTERNAL_ERROR,
                        _("unable to read server cert %s"), certfile);
        goto error;
    }

    ret = gnutls_x509_crt_init(&cert);
    if (ret < 0) {
        qemuReportError(VIR_ERR_INTERNAL_ERROR,
                        _("cannot initialize cert object: %s"),
                        gnutls_strerror(ret));
        goto error;
    }

    pemdatum.data = (unsigned char *)pemdata;
    pemdatum.size = strlen(pemdata);

    ret = gnutls_x509_crt_import(cert, &pemdatum, GNUTLS_X509_FMT_PEM);
    if (ret < 0) {
        qemuReportError(VIR_ERR_INTERNAL_ERROR,
                        _("cannot load cert data from %s: %s"),
                        certfile, gnutls_strerror(ret));
        goto error;
    }

    subjectlen = 1024;
    if (VIR_ALLOC_N(subject, subjectlen+1) < 0)
        goto no_memory;

    gnutls_x509_crt_get_dn(cert, subject, &subjectlen);
    subject[subjectlen] = '\0';

    VIR_FREE(certfile);
    VIR_FREE(pemdata);

    return subject;

no_memory:
    virReportOOMError();
error:
    VIR_FREE(certfile);
    VIR_FREE(pemdata);
    return NULL;
}


static qemuMigrationCookieGraphicsPtr
qemuMigrationCookieGraphicsAlloc(struct qemud_driver *driver,
                                 virDomainGraphicsDefPtr def)
{
    qemuMigrationCookieGraphicsPtr mig = NULL;
    const char *listenAddr;

    if (VIR_ALLOC(mig) < 0)
        goto no_memory;

    mig->type = def->type;
    if (mig->type == VIR_DOMAIN_GRAPHICS_TYPE_VNC) {
        mig->port = def->data.vnc.port;
225
        listenAddr = virDomainGraphicsListenGetAddress(def, 0);
226 227 228 229 230 231 232 233 234 235 236 237
        if (!listenAddr)
            listenAddr = driver->vncListen;

        if (driver->vncTLS &&
            !(mig->tlsSubject = qemuDomainExtractTLSSubject(driver->vncTLSx509certdir)))
            goto error;
    } else {
        mig->port = def->data.spice.port;
        if (driver->spiceTLS)
            mig->tlsPort = def->data.spice.tlsPort;
        else
            mig->tlsPort = -1;
238
        listenAddr = virDomainGraphicsListenGetAddress(def, 0);
239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258
        if (!listenAddr)
            listenAddr = driver->spiceListen;

        if (driver->spiceTLS &&
            !(mig->tlsSubject = qemuDomainExtractTLSSubject(driver->spiceTLSx509certdir)))
            goto error;
    }
    if (!(mig->listen = strdup(listenAddr)))
        goto no_memory;

    return mig;

no_memory:
    virReportOOMError();
error:
    qemuMigrationCookieGraphicsFree(mig);
    return NULL;
}


259 260 261
static qemuMigrationCookiePtr
qemuMigrationCookieNew(virDomainObjPtr dom)
{
J
Jiri Denemark 已提交
262
    qemuDomainObjPrivatePtr priv = dom->privateData;
263
    qemuMigrationCookiePtr mig = NULL;
J
Jiri Denemark 已提交
264
    const char *name;
265 266 267 268

    if (VIR_ALLOC(mig) < 0)
        goto no_memory;

J
Jiri Denemark 已提交
269 270 271 272 273
    if (priv->origname)
        name = priv->origname;
    else
        name = dom->def->name;
    if (!(mig->name = strdup(name)))
274 275 276
        goto no_memory;
    memcpy(mig->uuid, dom->def->uuid, VIR_UUID_BUFLEN);

277
    if (!(mig->localHostname = virGetHostname(NULL)))
278
        goto error;
279
    if (virGetHostUUID(mig->localHostuuid) < 0) {
280 281 282 283 284 285 286 287 288 289 290 291 292 293 294
        qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                        _("Unable to obtain host UUID"));
        goto error;
    }

    return mig;

no_memory:
    virReportOOMError();
error:
    qemuMigrationCookieFree(mig);
    return NULL;
}


295 296 297 298 299 300 301 302 303 304 305 306 307
static int
qemuMigrationCookieAddGraphics(qemuMigrationCookiePtr mig,
                               struct qemud_driver *driver,
                               virDomainObjPtr dom)
{
    if (mig->flags & QEMU_MIGRATION_COOKIE_GRAPHICS) {
        qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                        _("Migration graphics data already present"));
        return -1;
    }

    if (dom->def->ngraphics == 1 &&
        (dom->def->graphics[0]->type == VIR_DOMAIN_GRAPHICS_TYPE_VNC ||
308 309 310 311 312 313
         dom->def->graphics[0]->type == VIR_DOMAIN_GRAPHICS_TYPE_SPICE)) {
        if (!(mig->graphics =
              qemuMigrationCookieGraphicsAlloc(driver, dom->def->graphics[0])))
            return -1;
        mig->flags |= QEMU_MIGRATION_COOKIE_GRAPHICS;
    }
314 315 316 317 318

    return 0;
}


319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352
static int
qemuMigrationCookieAddLockstate(qemuMigrationCookiePtr mig,
                                struct qemud_driver *driver,
                                virDomainObjPtr dom)
{
    qemuDomainObjPrivatePtr priv = dom->privateData;

    if (mig->flags & QEMU_MIGRATION_COOKIE_LOCKSTATE) {
        qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                        _("Migration lockstate data already present"));
        return -1;
    }

    if (virDomainObjGetState(dom, NULL) == VIR_DOMAIN_PAUSED) {
        if (priv->lockState &&
            !(mig->lockState = strdup(priv->lockState)))
            return -1;
    } else {
        if (virDomainLockProcessInquire(driver->lockManager, dom, &mig->lockState) < 0)
            return -1;
    }

    if (!(mig->lockDriver = strdup(virLockManagerPluginGetName(driver->lockManager)))) {
        VIR_FREE(mig->lockState);
        return -1;
    }

    mig->flags |= QEMU_MIGRATION_COOKIE_LOCKSTATE;
    mig->flagsMandatory |= QEMU_MIGRATION_COOKIE_LOCKSTATE;

    return 0;
}


353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372
static int
qemuMigrationCookieAddPersistent(qemuMigrationCookiePtr mig,
                                 virDomainObjPtr dom)
{
    if (mig->flags & QEMU_MIGRATION_COOKIE_PERSISTENT) {
        qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                        _("Migration persistent data already present"));
        return -1;
    }

    if (!dom->newDef)
        return 0;

    mig->persistent = dom->newDef;
    mig->flags |= QEMU_MIGRATION_COOKIE_PERSISTENT;
    mig->flagsMandatory |= QEMU_MIGRATION_COOKIE_PERSISTENT;
    return 0;
}


373

374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391
static void qemuMigrationCookieGraphicsXMLFormat(virBufferPtr buf,
                                                 qemuMigrationCookieGraphicsPtr grap)
{
    virBufferAsprintf(buf, "  <graphics type='%s' port='%d' listen='%s'",
                      virDomainGraphicsTypeToString(grap->type),
                      grap->port, grap->listen);
    if (grap->type == VIR_DOMAIN_GRAPHICS_TYPE_SPICE)
        virBufferAsprintf(buf, " tlsPort='%d'", grap->tlsPort);
    if (grap->tlsSubject) {
        virBufferAddLit(buf, ">\n");
        virBufferEscapeString(buf, "    <cert info='subject' value='%s'/>\n", grap->tlsSubject);
        virBufferAddLit(buf, "  </graphics>\n");
    } else {
        virBufferAddLit(buf, "/>\n");
    }
}


392
static int
393 394
qemuMigrationCookieXMLFormat(struct qemud_driver *driver,
                             virBufferPtr buf,
395
                             qemuMigrationCookiePtr mig)
396 397 398
{
    char uuidstr[VIR_UUID_STRING_BUFLEN];
    char hostuuidstr[VIR_UUID_STRING_BUFLEN];
399
    int i;
400 401

    virUUIDFormat(mig->uuid, uuidstr);
402
    virUUIDFormat(mig->localHostuuid, hostuuidstr);
403 404 405 406

    virBufferAsprintf(buf, "<qemu-migration>\n");
    virBufferEscapeString(buf, "  <name>%s</name>\n", mig->name);
    virBufferAsprintf(buf, "  <uuid>%s</uuid>\n", uuidstr);
407
    virBufferEscapeString(buf, "  <hostname>%s</hostname>\n", mig->localHostname);
408
    virBufferAsprintf(buf, "  <hostuuid>%s</hostuuid>\n", hostuuidstr);
409

410 411 412 413 414 415
    for (i = 0 ; i < QEMU_MIGRATION_COOKIE_FLAG_LAST ; i++) {
        if (mig->flagsMandatory & (1 << i))
            virBufferAsprintf(buf, "  <feature name='%s'/>\n",
                              qemuMigrationCookieFlagTypeToString(i));
    }

416 417
    if ((mig->flags & QEMU_MIGRATION_COOKIE_GRAPHICS) &&
        mig->graphics)
418 419
        qemuMigrationCookieGraphicsXMLFormat(buf, mig->graphics);

420 421 422 423 424 425 426 427 428
    if ((mig->flags & QEMU_MIGRATION_COOKIE_LOCKSTATE) &&
        mig->lockState) {
        virBufferAsprintf(buf, "  <lockstate driver='%s'>\n",
                          mig->lockDriver);
        virBufferAsprintf(buf, "    <leases>%s</leases>\n",
                          mig->lockState);
        virBufferAddLit(buf, "  </lockstate>\n");
    }

429 430
    if ((mig->flags & QEMU_MIGRATION_COOKIE_PERSISTENT) &&
        mig->persistent) {
431
        virBufferAdjustIndent(buf, 2);
432 433 434 435
        if (qemuDomainDefFormatBuf(driver,
                                   mig->persistent,
                                   VIR_DOMAIN_XML_INACTIVE |
                                   VIR_DOMAIN_XML_SECURE,
436
                                   true,
437
                                   buf) < 0)
438 439
            return -1;
        virBufferAdjustIndent(buf, -2);
440 441
    }

442
    virBufferAddLit(buf, "</qemu-migration>\n");
443
    return 0;
444 445 446
}


447 448
static char *qemuMigrationCookieXMLFormatStr(struct qemud_driver *driver,
                                             qemuMigrationCookiePtr mig)
449 450 451
{
    virBuffer buf = VIR_BUFFER_INITIALIZER;

452
    if (qemuMigrationCookieXMLFormat(driver, &buf, mig) < 0) {
453 454 455
        virBufferFreeAndReset(&buf);
        return NULL;
    }
456 457 458

    if (virBufferError(&buf)) {
        virReportOOMError();
459
        virBufferFreeAndReset(&buf);
460 461 462 463 464 465 466
        return NULL;
    }

    return virBufferContentAndReset(&buf);
}


467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486
static qemuMigrationCookieGraphicsPtr
qemuMigrationCookieGraphicsXMLParse(xmlXPathContextPtr ctxt)
{
    qemuMigrationCookieGraphicsPtr grap;
    char *tmp;

    if (VIR_ALLOC(grap) < 0)
        goto no_memory;

    if (!(tmp = virXPathString("string(./graphics/@type)", ctxt))) {
        qemuReportError(VIR_ERR_INTERNAL_ERROR,
                        "%s", _("missing type attribute in migration data"));
        goto error;
    }
    if ((grap->type = virDomainGraphicsTypeFromString(tmp)) < 0) {
        qemuReportError(VIR_ERR_INTERNAL_ERROR,
                        _("unknown graphics type %s"), tmp);
        VIR_FREE(tmp);
        goto error;
    }
E
Eric Blake 已提交
487
    VIR_FREE(tmp);
488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505
    if (virXPathInt("string(./graphics/@port)", ctxt, &grap->port) < 0) {
        qemuReportError(VIR_ERR_INTERNAL_ERROR,
                        "%s", _("missing port attribute in migration data"));
        goto error;
    }
    if (grap->type == VIR_DOMAIN_GRAPHICS_TYPE_SPICE) {
        if (virXPathInt("string(./graphics/@tlsPort)", ctxt, &grap->tlsPort) < 0) {
            qemuReportError(VIR_ERR_INTERNAL_ERROR,
                            "%s", _("missing tlsPort attribute in migration data"));
            goto error;
        }
    }
    if (!(grap->listen = virXPathString("string(./graphics/@listen)", ctxt))) {
        qemuReportError(VIR_ERR_INTERNAL_ERROR,
                        "%s", _("missing listen attribute in migration data"));
        goto error;
    }
    /* Optional */
506
    grap->tlsSubject = virXPathString("string(./graphics/cert[@info='subject']/@value)", ctxt);
507 508 509 510 511 512 513 514 515 516 517 518


    return grap;

no_memory:
    virReportOOMError();
error:
    qemuMigrationCookieGraphicsFree(grap);
    return NULL;
}


519 520
static int
qemuMigrationCookieXMLParse(qemuMigrationCookiePtr mig,
521 522
                            struct qemud_driver *driver,
                            xmlDocPtr doc,
523
                            xmlXPathContextPtr ctxt,
E
Eric Blake 已提交
524
                            unsigned int flags)
525 526 527
{
    char uuidstr[VIR_UUID_STRING_BUFLEN];
    char *tmp;
528 529
    xmlNodePtr *nodes = NULL;
    int i, n;
530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565

    /* We don't store the uuid, name, hostname, or hostuuid
     * values. We just compare them to local data to do some
     * sanity checking on migration operation
     */

    /* Extract domain name */
    if (!(tmp = virXPathString("string(./name[1])", ctxt))) {
        qemuReportError(VIR_ERR_INTERNAL_ERROR,
                        "%s", _("missing name element in migration data"));
        goto error;
    }
    if (STRNEQ(tmp, mig->name)) {
        qemuReportError(VIR_ERR_INTERNAL_ERROR,
                        _("Incoming cookie data had unexpected name %s vs %s"),
                        tmp, mig->name);
        goto error;
    }
    VIR_FREE(tmp);

    /* Extract domain uuid */
    tmp = virXPathString("string(./uuid[1])", ctxt);
    if (!tmp) {
        qemuReportError(VIR_ERR_INTERNAL_ERROR,
                        "%s", _("missing uuid element in migration data"));
        goto error;
    }
    virUUIDFormat(mig->uuid, uuidstr);
    if (STRNEQ(tmp, uuidstr)) {
        qemuReportError(VIR_ERR_INTERNAL_ERROR,
                        _("Incoming cookie data had unexpected UUID %s vs %s"),
                        tmp, uuidstr);
    }
    VIR_FREE(tmp);

    /* Check & forbid "localhost" migration */
566
    if (!(mig->remoteHostname = virXPathString("string(./hostname[1])", ctxt))) {
567 568 569 570
        qemuReportError(VIR_ERR_INTERNAL_ERROR,
                        "%s", _("missing hostname element in migration data"));
        goto error;
    }
571
    if (STREQ(mig->remoteHostname, mig->localHostname)) {
572 573
        qemuReportError(VIR_ERR_INTERNAL_ERROR,
                        _("Attempt to migrate guest to the same host %s"),
574
                        mig->remoteHostname);
575 576 577 578 579 580 581 582
        goto error;
    }

    if (!(tmp = virXPathString("string(./hostuuid[1])", ctxt))) {
        qemuReportError(VIR_ERR_INTERNAL_ERROR,
                        "%s", _("missing hostuuid element in migration data"));
        goto error;
    }
583 584 585 586 587 588
    if (virUUIDParse(tmp, mig->remoteHostuuid) < 0) {
        qemuReportError(VIR_ERR_INTERNAL_ERROR,
                        "%s", _("malformed hostuuid element in migration data"));
        goto error;
    }
    if (memcmp(mig->remoteHostuuid, mig->localHostuuid, VIR_UUID_BUFLEN) == 0) {
589 590 591 592 593 594 595
        qemuReportError(VIR_ERR_INTERNAL_ERROR,
                        _("Attempt to migrate guest to the same host %s"),
                        tmp);
        goto error;
    }
    VIR_FREE(tmp);

596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627
    /* Check to ensure all mandatory features from XML are also
     * present in 'flags' */
    if ((n = virXPathNodeSet("./features", ctxt, &nodes)) < 0)
        goto error;

    for (i = 0 ; i < n ; i++) {
        int val;
        char *str = virXMLPropString(nodes[i], "name");
        if (!str) {
            qemuReportError(VIR_ERR_INTERNAL_ERROR,
                            "%s", _("missing feature name"));
            goto error;
        }

        if ((val = qemuMigrationCookieFlagTypeFromString(str)) < 0) {
            qemuReportError(VIR_ERR_INTERNAL_ERROR,
                            _("Unknown migration cookie feature %s"),
                            str);
            VIR_FREE(str);
            goto error;
        }

        if ((flags & (1 << val)) == 0) {
            qemuReportError(VIR_ERR_INTERNAL_ERROR,
                            _("Unsupported migration cookie feature %s"),
                            str);
            VIR_FREE(str);
        }
        VIR_FREE(str);
    }
    VIR_FREE(nodes);

628 629 630 631 632
    if ((flags & QEMU_MIGRATION_COOKIE_GRAPHICS) &&
        virXPathBoolean("count(./graphics) > 0", ctxt) &&
        (!(mig->graphics = qemuMigrationCookieGraphicsXMLParse(ctxt))))
        goto error;

633 634 635 636 637 638 639 640 641 642 643 644 645
    if ((flags & QEMU_MIGRATION_COOKIE_LOCKSTATE) &&
        virXPathBoolean("count(./lockstate) > 0", ctxt)) {
        mig->lockDriver = virXPathString("string(./lockstate[1]/@driver)", ctxt);
        if (!mig->lockDriver) {
            qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                            _("Missing lock driver name in migration cookie"));
            goto error;
        }
        mig->lockState = virXPathString("string(./lockstate[1]/leases[1])", ctxt);
        if (mig->lockState && STREQ(mig->lockState, ""))
            VIR_FREE(mig->lockState);
    }

646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664
    if ((flags & QEMU_MIGRATION_COOKIE_PERSISTENT) &&
        virXPathBoolean("count(./domain) > 0", ctxt)) {
        if ((n = virXPathNodeSet("./domain", ctxt, &nodes)) > 1) {
            qemuReportError(VIR_ERR_INTERNAL_ERROR,
                            _("Too many domain elements in "
                              "migration cookie: %d"),
                            n);
            goto error;
        }
        mig->persistent = virDomainDefParseNode(driver->caps, doc, nodes[0],
                                                -1, VIR_DOMAIN_XML_INACTIVE);
        if (!mig->persistent) {
            /* virDomainDefParseNode already reported
             * an error for us */
            goto error;
        }
        VIR_FREE(nodes);
    }

665 666 667 668
    return 0;

error:
    VIR_FREE(tmp);
669
    VIR_FREE(nodes);
670 671 672 673 674 675
    return -1;
}


static int
qemuMigrationCookieXMLParseStr(qemuMigrationCookiePtr mig,
676
                               struct qemud_driver *driver,
677
                               const char *xml,
E
Eric Blake 已提交
678
                               unsigned int flags)
679 680 681
{
    xmlDocPtr doc = NULL;
    xmlXPathContextPtr ctxt = NULL;
682
    int ret = -1;
683 684 685

    VIR_DEBUG("xml=%s", NULLSTR(xml));

686
    if (!(doc = virXMLParseStringCtxt(xml, _("(qemu_migration_cookie)"), &ctxt)))
687 688
        goto cleanup;

689
    ret = qemuMigrationCookieXMLParse(mig, driver, doc, ctxt, flags);
690 691 692 693 694 695 696 697 698 699 700

cleanup:
    xmlXPathFreeContext(ctxt);
    xmlFreeDoc(doc);

    return ret;
}


static int
qemuMigrationBakeCookie(qemuMigrationCookiePtr mig,
701 702
                        struct qemud_driver *driver,
                        virDomainObjPtr dom,
703 704
                        char **cookieout,
                        int *cookieoutlen,
E
Eric Blake 已提交
705
                        unsigned int flags)
706
{
707 708
    if (!cookieout || !cookieoutlen)
        return 0;
709 710 711

    *cookieoutlen = 0;

712 713 714 715
    if (flags & QEMU_MIGRATION_COOKIE_GRAPHICS &&
        qemuMigrationCookieAddGraphics(mig, driver, dom) < 0)
        return -1;

716 717 718 719
    if (flags & QEMU_MIGRATION_COOKIE_LOCKSTATE &&
        qemuMigrationCookieAddLockstate(mig, driver, dom) < 0)
        return -1;

720 721 722 723
    if (flags & QEMU_MIGRATION_COOKIE_PERSISTENT &&
        qemuMigrationCookieAddPersistent(mig, dom) < 0)
        return -1;

724
    if (!(*cookieout = qemuMigrationCookieXMLFormatStr(driver, mig)))
725 726 727 728 729 730 731 732 733 734 735
        return -1;

    *cookieoutlen = strlen(*cookieout) + 1;

    VIR_DEBUG("cookielen=%d cookie=%s", *cookieoutlen, *cookieout);

    return 0;
}


static qemuMigrationCookiePtr
736 737
qemuMigrationEatCookie(struct qemud_driver *driver,
                       virDomainObjPtr dom,
738 739
                       const char *cookiein,
                       int cookieinlen,
E
Eric Blake 已提交
740
                       unsigned int flags)
741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758
{
    qemuMigrationCookiePtr mig = NULL;

    /* Parse & validate incoming cookie (if any) */
    if (cookiein && cookieinlen &&
        cookiein[cookieinlen-1] != '\0') {
        qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                        _("Migration cookie was not NULL terminated"));
        goto error;
    }

    VIR_DEBUG("cookielen=%d cookie='%s'", cookieinlen, NULLSTR(cookiein));

    if (!(mig = qemuMigrationCookieNew(dom)))
        return NULL;

    if (cookiein && cookieinlen &&
        qemuMigrationCookieXMLParseStr(mig,
759
                                       driver,
760 761 762 763
                                       cookiein,
                                       flags) < 0)
        goto error;

764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781
    if (mig->flags & QEMU_MIGRATION_COOKIE_LOCKSTATE) {
        if (!mig->lockDriver) {
            if (virLockManagerPluginUsesState(driver->lockManager)) {
                qemuReportError(VIR_ERR_INTERNAL_ERROR,
                                _("Missing %s lock state for migration cookie"),
                                virLockManagerPluginGetName(driver->lockManager));
                goto error;
            }
        } else if (STRNEQ(mig->lockDriver,
                          virLockManagerPluginGetName(driver->lockManager))) {
            qemuReportError(VIR_ERR_INTERNAL_ERROR,
                            _("Source host lock driver %s different from target %s"),
                            mig->lockDriver,
                            virLockManagerPluginGetName(driver->lockManager));
            goto error;
        }
    }

782 783 784 785 786 787
    return mig;

error:
    qemuMigrationCookieFree(mig);
    return NULL;
}
788

789 790 791 792 793 794 795 796 797 798 799 800
/* Validate whether the domain is safe to migrate.  If vm is NULL,
 * then this is being run in the v2 Prepare stage on the destination
 * (where we only have the target xml); if vm is provided, then this
 * is being run in either v2 Perform or v3 Begin (where we also have
 * access to all of the domain's metadata, such as whether it is
 * marked autodestroy or has snapshots).  While it would be nice to
 * assume that checking on source is sufficient to prevent ever
 * talking to the destination in the first place, we are stuck with
 * the fact that older servers did not do checks on the source. */
static bool
qemuMigrationIsAllowed(struct qemud_driver *driver, virDomainObjPtr vm,
                       virDomainDefPtr def)
801
{
802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818
    int nsnapshots;

    if (vm) {
        if (qemuProcessAutoDestroyActive(driver, vm)) {
            qemuReportError(VIR_ERR_OPERATION_INVALID,
                            "%s", _("domain is marked for auto destroy"));
            return false;
        }
        if ((nsnapshots = virDomainSnapshotObjListNum(&vm->snapshots, 0))) {
            qemuReportError(VIR_ERR_OPERATION_INVALID,
                            _("cannot migrate domain with %d snapshots"),
                            nsnapshots);
            return false;
        }

        def = vm->def;
    }
819 820 821 822 823 824 825 826 827
    if (def->nhostdevs > 0) {
        qemuReportError(VIR_ERR_OPERATION_INVALID,
            "%s", _("Domain with assigned host devices cannot be migrated"));
        return false;
    }

    return true;
}

828 829 830 831 832 833 834 835
static bool
qemuMigrationIsSafe(virDomainDefPtr def)
{
    int i;

    for (i = 0 ; i < def->ndisks ; i++) {
        virDomainDiskDefPtr disk = def->disks[i];

836 837
        /* Our code elsewhere guarantees shared disks are either readonly (in
         * which case cache mode doesn't matter) or used with cache=none */
838
        if (disk->src &&
839 840 841
            !disk->shared &&
            !disk->readonly &&
            disk->cachemode != VIR_DOMAIN_DISK_CACHE_DISABLE) {
842
            int cfs;
843 844 845 846 847 848 849

            if (disk->type == VIR_DOMAIN_DISK_TYPE_FILE) {
                if ((cfs = virStorageFileIsClusterFS(disk->src)) == 1)
                    continue;
                else if (cfs < 0)
                    return false;
            }
850 851 852 853 854 855 856 857 858 859 860

            qemuReportError(VIR_ERR_MIGRATE_UNSAFE, "%s",
                            _("Migration may lead to data corruption if disks"
                              " use cache != none"));
            return false;
        }
    }

    return true;
}

861 862 863 864 865 866 867 868
/** qemuMigrationSetOffline
 * Pause domain for non-live migration.
 */
int
qemuMigrationSetOffline(struct qemud_driver *driver,
                        virDomainObjPtr vm)
{
    int ret;
869
    VIR_DEBUG("driver=%p vm=%p", driver, vm);
870 871
    ret = qemuProcessStopCPUs(driver, vm, VIR_DOMAIN_PAUSED_MIGRATION,
                              QEMU_ASYNC_JOB_MIGRATION_OUT);
872 873 874 875 876 877 878 879 880 881 882 883 884 885
    if (ret == 0) {
        virDomainEventPtr event;

        event = virDomainEventNewFromObj(vm,
                                         VIR_DOMAIN_EVENT_SUSPENDED,
                                         VIR_DOMAIN_EVENT_SUSPENDED_MIGRATED);
        if (event)
            qemuDomainEventQueue(driver, event);
    }

    return ret;
}


886 887 888
static int
qemuMigrationUpdateJobStatus(struct qemud_driver *driver,
                             virDomainObjPtr vm,
889 890
                             const char *job,
                             enum qemuDomainAsyncJob asyncJob)
891
{
892
    qemuDomainObjPrivatePtr priv = vm->privateData;
893
    int ret;
894 895 896 897
    int status;
    unsigned long long memProcessed;
    unsigned long long memRemaining;
    unsigned long long memTotal;
898

899 900 901
    ret = qemuDomainObjEnterMonitorAsync(driver, vm, asyncJob);
    if (ret < 0) {
        /* Guest already exited; nothing further to update.  */
902 903
        return -1;
    }
904 905 906 907 908 909
    ret = qemuMonitorGetMigrationStatus(priv->mon,
                                        &status,
                                        &memProcessed,
                                        &memRemaining,
                                        &memTotal);
    qemuDomainObjExitMonitorWithDriver(driver, vm);
910

911
    if (ret < 0 || virTimeMillisNow(&priv->job.info.timeElapsed) < 0) {
912
        priv->job.info.type = VIR_DOMAIN_JOB_FAILED;
913 914
        return -1;
    }
915
    priv->job.info.timeElapsed -= priv->job.start;
916

917
    ret = -1;
918 919
    switch (status) {
    case QEMU_MONITOR_MIGRATION_STATUS_INACTIVE:
920
        priv->job.info.type = VIR_DOMAIN_JOB_NONE;
921 922 923 924 925
        qemuReportError(VIR_ERR_OPERATION_FAILED,
                        _("%s: %s"), job, _("is not active"));
        break;

    case QEMU_MONITOR_MIGRATION_STATUS_ACTIVE:
926 927 928
        priv->job.info.dataTotal = memTotal;
        priv->job.info.dataRemaining = memRemaining;
        priv->job.info.dataProcessed = memProcessed;
929

930 931 932
        priv->job.info.memTotal = memTotal;
        priv->job.info.memRemaining = memRemaining;
        priv->job.info.memProcessed = memProcessed;
933 934 935 936 937

        ret = 0;
        break;

    case QEMU_MONITOR_MIGRATION_STATUS_COMPLETED:
938
        priv->job.info.type = VIR_DOMAIN_JOB_COMPLETED;
939 940 941 942
        ret = 0;
        break;

    case QEMU_MONITOR_MIGRATION_STATUS_ERROR:
943
        priv->job.info.type = VIR_DOMAIN_JOB_FAILED;
944 945 946 947 948
        qemuReportError(VIR_ERR_OPERATION_FAILED,
                        _("%s: %s"), job, _("unexpectedly failed"));
        break;

    case QEMU_MONITOR_MIGRATION_STATUS_CANCELLED:
949
        priv->job.info.type = VIR_DOMAIN_JOB_CANCELLED;
950
        qemuReportError(VIR_ERR_OPERATION_ABORTED,
951 952 953 954 955 956 957 958
                        _("%s: %s"), job, _("canceled by client"));
        break;
    }

    return ret;
}


959 960
static int
qemuMigrationWaitForCompletion(struct qemud_driver *driver, virDomainObjPtr vm,
961 962
                               enum qemuDomainAsyncJob asyncJob,
                               virConnectPtr dconn)
963
{
964
    qemuDomainObjPrivatePtr priv = vm->privateData;
965 966
    const char *job;

967 968
    switch (priv->job.asyncJob) {
    case QEMU_ASYNC_JOB_MIGRATION_OUT:
969 970
        job = _("migration job");
        break;
971
    case QEMU_ASYNC_JOB_SAVE:
972 973
        job = _("domain save job");
        break;
974
    case QEMU_ASYNC_JOB_DUMP:
975 976 977 978 979
        job = _("domain core dump job");
        break;
    default:
        job = _("job");
    }
980

981
    priv->job.info.type = VIR_DOMAIN_JOB_UNBOUNDED;
982

983
    while (priv->job.info.type == VIR_DOMAIN_JOB_UNBOUNDED) {
984 985 986
        /* Poll every 50ms for progress & to allow cancellation */
        struct timespec ts = { .tv_sec = 0, .tv_nsec = 50 * 1000 * 1000ull };

987
        if (qemuMigrationUpdateJobStatus(driver, vm, job, asyncJob) < 0)
988 989
            goto cleanup;

990 991 992 993 994 995
        if (dconn && virConnectIsAlive(dconn) <= 0) {
            qemuReportError(VIR_ERR_OPERATION_FAILED, "%s",
                            _("Lost connection to destination host"));
            goto cleanup;
        }

996 997 998 999 1000 1001 1002 1003 1004 1005
        virDomainObjUnlock(vm);
        qemuDriverUnlock(driver);

        nanosleep(&ts, NULL);

        qemuDriverLock(driver);
        virDomainObjLock(vm);
    }

cleanup:
1006
    if (priv->job.info.type == VIR_DOMAIN_JOB_COMPLETED)
1007 1008 1009
        return 0;
    else
        return -1;
1010 1011 1012
}


1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032
static int
qemuDomainMigrateGraphicsRelocate(struct qemud_driver *driver,
                                  virDomainObjPtr vm,
                                  qemuMigrationCookiePtr cookie)
{
    qemuDomainObjPrivatePtr priv = vm->privateData;
    int ret;

    if (!cookie)
        return 0;

    if (!cookie->graphics)
        return 0;

    /* QEMU doesn't support VNC relocation yet, so
     * skip it to avoid generating an error
     */
    if (cookie->graphics->type != VIR_DOMAIN_GRAPHICS_TYPE_SPICE)
        return 0;

1033 1034
    ret = qemuDomainObjEnterMonitorAsync(driver, vm,
                                         QEMU_ASYNC_JOB_MIGRATION_OUT);
1035 1036 1037 1038 1039 1040 1041 1042 1043
    if (ret == 0) {
        ret = qemuMonitorGraphicsRelocate(priv->mon,
                                          cookie->graphics->type,
                                          cookie->remoteHostname,
                                          cookie->graphics->port,
                                          cookie->graphics->tlsPort,
                                          cookie->graphics->tlsSubject);
        qemuDomainObjExitMonitorWithDriver(driver, vm);
    }
1044 1045 1046 1047 1048

    return ret;
}


1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109
/* This is called for outgoing non-p2p migrations when a connection to the
 * client which initiated the migration was closed but we were waiting for it
 * to follow up with the next phase, that is, in between
 * qemuDomainMigrateBegin3 and qemuDomainMigratePerform3 or
 * qemuDomainMigratePerform3 and qemuDomainMigrateConfirm3.
 */
virDomainObjPtr
qemuMigrationCleanup(struct qemud_driver *driver,
                     virDomainObjPtr vm,
                     virConnectPtr conn)
{
    qemuDomainObjPrivatePtr priv = vm->privateData;

    VIR_DEBUG("vm=%s, conn=%p, asyncJob=%s, phase=%s",
              vm->def->name, conn,
              qemuDomainAsyncJobTypeToString(priv->job.asyncJob),
              qemuDomainAsyncJobPhaseToString(priv->job.asyncJob,
                                              priv->job.phase));

    if (!qemuMigrationJobIsActive(vm, QEMU_ASYNC_JOB_MIGRATION_OUT))
        goto cleanup;

    VIR_DEBUG("The connection which started outgoing migration of domain %s"
              " was closed; canceling the migration",
              vm->def->name);

    switch ((enum qemuMigrationJobPhase) priv->job.phase) {
    case QEMU_MIGRATION_PHASE_BEGIN3:
        /* just forget we were about to migrate */
        qemuDomainObjDiscardAsyncJob(driver, vm);
        break;

    case QEMU_MIGRATION_PHASE_PERFORM3_DONE:
        VIR_WARN("Migration of domain %s finished but we don't know if the"
                 " domain was successfully started on destination or not",
                 vm->def->name);
        /* clear the job and let higher levels decide what to do */
        qemuDomainObjDiscardAsyncJob(driver, vm);
        break;

    case QEMU_MIGRATION_PHASE_PERFORM3:
        /* cannot be seen without an active migration API; unreachable */
    case QEMU_MIGRATION_PHASE_CONFIRM3:
    case QEMU_MIGRATION_PHASE_CONFIRM3_CANCELLED:
        /* all done; unreachable */
    case QEMU_MIGRATION_PHASE_PREPARE:
    case QEMU_MIGRATION_PHASE_FINISH2:
    case QEMU_MIGRATION_PHASE_FINISH3:
        /* incoming migration; unreachable */
    case QEMU_MIGRATION_PHASE_PERFORM2:
        /* single phase outgoing migration; unreachable */
    case QEMU_MIGRATION_PHASE_NONE:
    case QEMU_MIGRATION_PHASE_LAST:
        /* unreachable */
        ;
    }

cleanup:
    return vm;
}

1110
/* The caller is supposed to lock the vm and start a migration job. */
1111 1112
char *qemuMigrationBegin(struct qemud_driver *driver,
                         virDomainObjPtr vm,
1113
                         const char *xmlin,
1114
                         const char *dname,
1115
                         char **cookieout,
1116 1117
                         int *cookieoutlen,
                         unsigned long flags)
1118 1119 1120
{
    char *rv = NULL;
    qemuMigrationCookiePtr mig = NULL;
1121
    virDomainDefPtr def = NULL;
1122 1123
    qemuDomainObjPrivatePtr priv = vm->privateData;

1124
    VIR_DEBUG("driver=%p, vm=%p, xmlin=%s, dname=%s,"
1125
              " cookieout=%p, cookieoutlen=%p, flags=%lx",
1126
              driver, vm, NULLSTR(xmlin), NULLSTR(dname),
1127
              cookieout, cookieoutlen, flags);
1128

1129 1130 1131 1132 1133 1134
    /* Only set the phase if we are inside QEMU_ASYNC_JOB_MIGRATION_OUT.
     * Otherwise we will start the async job later in the perform phase losing
     * change protection.
     */
    if (priv->job.asyncJob == QEMU_ASYNC_JOB_MIGRATION_OUT)
        qemuMigrationJobSetPhase(driver, vm, QEMU_MIGRATION_PHASE_BEGIN3);
1135

1136
    if (!qemuMigrationIsAllowed(driver, vm, NULL))
1137 1138
        goto cleanup;

1139 1140 1141
    if (!(flags & VIR_MIGRATE_UNSAFE) && !qemuMigrationIsSafe(vm->def))
        goto cleanup;

1142
    if (!(mig = qemuMigrationEatCookie(driver, vm, NULL, 0, 0)))
1143 1144 1145 1146
        goto cleanup;

    if (qemuMigrationBakeCookie(mig, driver, vm,
                                cookieout, cookieoutlen,
1147
                                QEMU_MIGRATION_COOKIE_LOCKSTATE) < 0)
1148 1149
        goto cleanup;

1150 1151
    if (xmlin) {
        if (!(def = virDomainDefParseString(driver->caps, xmlin,
M
Matthias Bolte 已提交
1152
                                            QEMU_EXPECTED_VIRT_TYPES,
1153 1154 1155
                                            VIR_DOMAIN_XML_INACTIVE)))
            goto cleanup;

1156
        if (STRNEQ(def->name, vm->def->name)) {
1157
            qemuReportError(VIR_ERR_INVALID_ARG, "%s",
1158
                            _("target domain name doesn't match source name"));
1159 1160 1161
            goto cleanup;
        }

1162
        if (!virDomainDefCheckABIStability(vm->def, def))
1163 1164
            goto cleanup;

1165
        rv = qemuDomainDefFormatLive(driver, def, false, true);
1166
    } else {
1167
        rv = qemuDomainDefFormatLive(driver, vm->def, false, true);
1168
    }
1169 1170 1171

cleanup:
    qemuMigrationCookieFree(mig);
1172
    virDomainDefFree(def);
1173 1174 1175 1176
    return rv;
}


1177 1178
/* Prepare is the first step, and it runs on the destination host.
 */
1179

1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196
static void
qemuMigrationPrepareCleanup(struct qemud_driver *driver,
                            virDomainObjPtr vm)
{
    qemuDomainObjPrivatePtr priv = vm->privateData;

    VIR_DEBUG("driver=%p, vm=%s, job=%s, asyncJob=%s",
              driver,
              vm->def->name,
              qemuDomainJobTypeToString(priv->job.active),
              qemuDomainAsyncJobTypeToString(priv->job.asyncJob));

    if (!qemuMigrationJobIsActive(vm, QEMU_ASYNC_JOB_MIGRATION_IN))
        return;
    qemuDomainObjDiscardAsyncJob(driver, vm);
}

1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207
static int
qemuMigrationPrepareAny(struct qemud_driver *driver,
                        virConnectPtr dconn,
                        const char *cookiein,
                        int cookieinlen,
                        char **cookieout,
                        int *cookieoutlen,
                        const char *dname,
                        const char *dom_xml,
                        const char *migrateFrom,
                        virStreamPtr st)
1208 1209 1210 1211 1212
{
    virDomainDefPtr def = NULL;
    virDomainObjPtr vm = NULL;
    virDomainEventPtr event = NULL;
    int ret = -1;
1213
    int dataFD[2] = { -1, -1 };
1214
    qemuDomainObjPrivatePtr priv = NULL;
J
Jiri Denemark 已提交
1215
    unsigned long long now;
1216
    qemuMigrationCookiePtr mig = NULL;
1217
    bool tunnel = !!st;
J
Jiri Denemark 已提交
1218
    char *origname = NULL;
J
Jiri Denemark 已提交
1219
    char *xmlout = NULL;
1220

1221
    if (virTimeMillisNow(&now) < 0)
1222 1223 1224
        return -1;

    if (!(def = virDomainDefParseString(driver->caps, dom_xml,
M
Matthias Bolte 已提交
1225
                                        QEMU_EXPECTED_VIRT_TYPES,
1226 1227 1228
                                        VIR_DOMAIN_XML_INACTIVE)))
        goto cleanup;

1229
    if (!qemuMigrationIsAllowed(driver, NULL, def))
1230 1231 1232 1233
        goto cleanup;

    /* Target domain name, maybe renamed. */
    if (dname) {
J
Jiri Denemark 已提交
1234
        origname = def->name;
1235 1236 1237 1238 1239
        def->name = strdup(dname);
        if (def->name == NULL)
            goto cleanup;
    }

J
Jiri Denemark 已提交
1240 1241 1242 1243 1244
    /* Let migration hook filter domain XML */
    if (virHookPresent(VIR_HOOK_DRIVER_QEMU)) {
        char *xml;
        int hookret;

1245
        if (!(xml = qemuDomainDefFormatXML(driver, def,
1246
                                           VIR_DOMAIN_XML_SECURE, false)))
J
Jiri Denemark 已提交
1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280
            goto cleanup;

        hookret = virHookCall(VIR_HOOK_DRIVER_QEMU, def->name,
                              VIR_HOOK_QEMU_OP_MIGRATE, VIR_HOOK_SUBOP_BEGIN,
                              NULL, xml, &xmlout);
        VIR_FREE(xml);

        if (hookret < 0) {
            goto cleanup;
        } else if (hookret == 0) {
            if (!*xmlout) {
                VIR_DEBUG("Migrate hook filter returned nothing; using the"
                          " original XML");
            } else {
                virDomainDefPtr newdef;

                VIR_DEBUG("Using hook-filtered domain XML: %s", xmlout);
                newdef = virDomainDefParseString(driver->caps, xmlout,
                                                 QEMU_EXPECTED_VIRT_TYPES,
                                                 VIR_DOMAIN_XML_INACTIVE);
                if (!newdef)
                    goto cleanup;

                if (!virDomainDefCheckABIStability(def, newdef)) {
                    virDomainDefFree(newdef);
                    goto cleanup;
                }

                virDomainDefFree(def);
                def = newdef;
            }
        }
    }

1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291
    if (virDomainObjIsDuplicate(&driver->domains, def, 1) < 0)
        goto cleanup;

    if (!(vm = virDomainAssignDef(driver->caps,
                                  &driver->domains,
                                  def, true))) {
        /* virDomainAssignDef already set the error */
        goto cleanup;
    }
    def = NULL;
    priv = vm->privateData;
J
Jiri Denemark 已提交
1292 1293
    priv->origname = origname;
    origname = NULL;
1294

1295 1296
    if (!(mig = qemuMigrationEatCookie(driver, vm, cookiein, cookieinlen,
                                       QEMU_MIGRATION_COOKIE_LOCKSTATE)))
1297 1298
        goto cleanup;

1299
    if (qemuMigrationJobStart(driver, vm, QEMU_ASYNC_JOB_MIGRATION_IN) < 0)
1300
        goto cleanup;
1301
    qemuMigrationJobSetPhase(driver, vm, QEMU_MIGRATION_PHASE_PREPARE);
1302 1303 1304 1305

    /* Domain starts inactive, even if the domain XML had an id field. */
    vm->def->id = -1;

1306 1307
    if (tunnel &&
        (pipe(dataFD) < 0 || virSetCloseExec(dataFD[1]) < 0)) {
1308 1309
        virReportSystemError(errno, "%s",
                             _("cannot create pipe for tunnelled migration"));
1310 1311 1312 1313
        goto endjob;
    }

    /* Start the QEMU daemon, with the same command-line arguments plus
1314
     * -incoming $migrateFrom
1315
     */
1316 1317 1318 1319
    if (qemuProcessStart(dconn, driver, vm, migrateFrom, dataFD[0], NULL, NULL,
                         VIR_NETDEV_VPORT_PROFILE_OP_MIGRATE_IN_START,
                         VIR_QEMU_PROCESS_START_PAUSED |
                         VIR_QEMU_PROCESS_START_AUTODESROY) < 0) {
1320
        virDomainAuditStart(vm, "migrated", false);
1321 1322 1323 1324 1325 1326
        /* Note that we don't set an error here because qemuProcessStart
         * should have already done that.
         */
        goto endjob;
    }

1327 1328 1329 1330 1331
    if (tunnel) {
        if (virFDStreamOpen(st, dataFD[1]) < 0) {
            virReportSystemError(errno, "%s",
                                 _("cannot pass pipe for tunnelled migration"));
            virDomainAuditStart(vm, "migrated", false);
1332
            qemuProcessStop(driver, vm, VIR_DOMAIN_SHUTOFF_FAILED, 0);
1333
            goto endjob;
1334
        }
1335
        dataFD[1] = -1; /* 'st' owns the FD now & will close it */
1336 1337
    }

1338 1339 1340 1341 1342 1343 1344 1345
    if (mig->lockState) {
        VIR_DEBUG("Received lockstate %s", mig->lockState);
        VIR_FREE(priv->lockState);
        priv->lockState = mig->lockState;
        mig->lockState = NULL;
    } else {
        VIR_DEBUG("Received no lockstate");
    }
1346

1347 1348
    if (qemuMigrationBakeCookie(mig, driver, vm, cookieout, cookieoutlen,
                                QEMU_MIGRATION_COOKIE_GRAPHICS) < 0) {
1349 1350 1351 1352 1353 1354 1355
        /* We could tear down the whole guest here, but
         * cookie data is (so far) non-critical, so that
         * seems a little harsh. We'll just warn for now.
         */
        VIR_WARN("Unable to encode migration cookie");
    }

1356 1357 1358
    if (qemuDomainCleanupAdd(vm, qemuMigrationPrepareCleanup) < 0)
        goto endjob;

1359 1360 1361 1362
    virDomainAuditStart(vm, "migrated", true);
    event = virDomainEventNewFromObj(vm,
                                     VIR_DOMAIN_EVENT_STARTED,
                                     VIR_DOMAIN_EVENT_STARTED_MIGRATED);
1363

1364 1365 1366 1367 1368
    /* We keep the job active across API calls until the finish() call.
     * This prevents any other APIs being invoked while incoming
     * migration is taking place.
     */
    if (qemuMigrationJobContinue(vm) == 0) {
1369
        vm = NULL;
1370 1371 1372
        qemuReportError(VIR_ERR_OPERATION_FAILED,
                        "%s", _("domain disappeared"));
        goto cleanup;
1373
    }
1374

1375
    ret = 0;
1376 1377

cleanup:
J
Jiri Denemark 已提交
1378
    VIR_FREE(origname);
J
Jiri Denemark 已提交
1379
    VIR_FREE(xmlout);
1380
    virDomainDefFree(def);
1381 1382
    VIR_FORCE_CLOSE(dataFD[0]);
    VIR_FORCE_CLOSE(dataFD[1]);
1383 1384 1385 1386 1387 1388
    if (vm) {
        if (ret >= 0 || vm->persistent)
            virDomainObjUnlock(vm);
        else
            qemuDomainRemoveInactive(driver, vm);
    }
1389 1390
    if (event)
        qemuDomainEventQueue(driver, event);
1391
    qemuMigrationCookieFree(mig);
1392
    return ret;
1393 1394 1395 1396 1397 1398

endjob:
    if (qemuMigrationJobFinish(driver, vm) == 0) {
        vm = NULL;
    }
    goto cleanup;
1399 1400 1401
}


1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433
/*
 * This version starts an empty VM listening on a localhost TCP port, and
 * sets up the corresponding virStream to handle the incoming data.
 */
int
qemuMigrationPrepareTunnel(struct qemud_driver *driver,
                           virConnectPtr dconn,
                           const char *cookiein,
                           int cookieinlen,
                           char **cookieout,
                           int *cookieoutlen,
                           virStreamPtr st,
                           const char *dname,
                           const char *dom_xml)
{
    int ret;

    VIR_DEBUG("driver=%p, dconn=%p, cookiein=%s, cookieinlen=%d, "
              "cookieout=%p, cookieoutlen=%p, st=%p, dname=%s, dom_xml=%s",
              driver, dconn, NULLSTR(cookiein), cookieinlen,
              cookieout, cookieoutlen, st, NULLSTR(dname), dom_xml);

    /* QEMU will be started with -incoming stdio (which qemu_command might
     * convert to exec:cat or fd:n)
     */
    ret = qemuMigrationPrepareAny(driver, dconn, cookiein, cookieinlen,
                                  cookieout, cookieoutlen, dname, dom_xml,
                                  "stdio", st);
    return ret;
}


1434 1435 1436
int
qemuMigrationPrepareDirect(struct qemud_driver *driver,
                           virConnectPtr dconn,
1437 1438 1439 1440
                           const char *cookiein,
                           int cookieinlen,
                           char **cookieout,
                           int *cookieoutlen,
1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451
                           const char *uri_in,
                           char **uri_out,
                           const char *dname,
                           const char *dom_xml)
{
    static int port = 0;
    int this_port;
    char *hostname = NULL;
    char migrateFrom [64];
    const char *p;
    int ret = -1;
J
Jiri Denemark 已提交
1452

1453 1454 1455 1456 1457 1458
    VIR_DEBUG("driver=%p, dconn=%p, cookiein=%s, cookieinlen=%d, "
              "cookieout=%p, cookieoutlen=%p, uri_in=%s, uri_out=%p, "
              "dname=%s, dom_xml=%s",
              driver, dconn, NULLSTR(cookiein), cookieinlen,
              cookieout, cookieoutlen, NULLSTR(uri_in), uri_out,
              NULLSTR(dname), dom_xml);
1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479

    /* The URI passed in may be NULL or a string "tcp://somehostname:port".
     *
     * If the URI passed in is NULL then we allocate a port number
     * from our pool of port numbers and return a URI of
     * "tcp://ourhostname:port".
     *
     * If the URI passed in is not NULL then we try to parse out the
     * port number and use that (note that the hostname is assumed
     * to be a correct hostname which refers to the target machine).
     */
    if (uri_in == NULL) {
        this_port = QEMUD_MIGRATION_FIRST_PORT + port++;
        if (port == QEMUD_MIGRATION_NUM_PORTS) port = 0;

        /* Get hostname */
        if ((hostname = virGetHostname(NULL)) == NULL)
            goto cleanup;

        if (STRPREFIX(hostname, "localhost")) {
            qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s",
1480 1481
                            _("hostname on destination resolved to localhost,"
                              " but migration requires an FQDN"));
1482 1483 1484 1485 1486
            goto cleanup;
        }

        /* XXX this really should have been a properly well-formed
         * URI, but we can't add in tcp:// now without breaking
1487
         * compatibility with old targets. We at least make the
1488 1489 1490
         * new targets accept both syntaxes though.
         */
        /* Caller frees */
1491
        if (virAsprintf(uri_out, "tcp:%s:%d", hostname, this_port) < 0) {
1492 1493 1494 1495 1496 1497 1498 1499 1500
            virReportOOMError();
            goto cleanup;
        }
    } else {
        /* Check the URI starts with "tcp:".  We will escape the
         * URI when passing it to the qemu monitor, so bad
         * characters in hostname part don't matter.
         */
        if (!STRPREFIX (uri_in, "tcp:")) {
1501 1502 1503
            qemuReportError(VIR_ERR_INVALID_ARG, "%s",
                            _("only tcp URIs are supported for KVM/QEMU"
                              " migrations"));
1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534
            goto cleanup;
        }

        /* Get the port number. */
        p = strrchr (uri_in, ':');
        if (p == strchr(uri_in, ':')) {
            /* Generate a port */
            this_port = QEMUD_MIGRATION_FIRST_PORT + port++;
            if (port == QEMUD_MIGRATION_NUM_PORTS)
                port = 0;

            /* Caller frees */
            if (virAsprintf(uri_out, "%s:%d", uri_in, this_port) < 0) {
                virReportOOMError();
                goto cleanup;
            }

        } else {
            p++; /* definitely has a ':' in it, see above */
            this_port = virParseNumber (&p);
            if (this_port == -1 || p-uri_in != strlen (uri_in)) {
                qemuReportError(VIR_ERR_INVALID_ARG,
                                "%s", _("URI ended with incorrect ':port'"));
                goto cleanup;
            }
        }
    }

    if (*uri_out)
        VIR_DEBUG("Generated uri_out=%s", *uri_out);

1535
    /* QEMU will be started with -incoming tcp:0.0.0.0:port */
1536
    snprintf(migrateFrom, sizeof(migrateFrom), "tcp:0.0.0.0:%d", this_port);
1537

1538 1539 1540
    ret = qemuMigrationPrepareAny(driver, dconn, cookiein, cookieinlen,
                                  cookieout, cookieoutlen, dname, dom_xml,
                                  migrateFrom, NULL);
1541 1542 1543 1544 1545 1546 1547 1548
cleanup:
    VIR_FREE(hostname);
    if (ret != 0)
        VIR_FREE(*uri_out);
    return ret;
}


1549 1550
enum qemuMigrationDestinationType {
    MIGRATION_DEST_HOST,
1551
    MIGRATION_DEST_CONNECT_HOST,
1552
    MIGRATION_DEST_UNIX,
1553
    MIGRATION_DEST_FD,
1554
};
1555

1556 1557 1558 1559
enum qemuMigrationForwardType {
    MIGRATION_FWD_DIRECT,
    MIGRATION_FWD_STREAM,
};
1560

1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571
typedef struct _qemuMigrationSpec qemuMigrationSpec;
typedef qemuMigrationSpec *qemuMigrationSpecPtr;
struct _qemuMigrationSpec {
    enum qemuMigrationDestinationType destType;
    union {
        struct {
            const char *name;
            int port;
        } host;

        struct {
1572
            char *file;
1573 1574
            int sock;
        } unix_socket;
1575 1576 1577 1578 1579

        struct {
            int qemu;
            int local;
        } fd;
1580 1581 1582 1583 1584 1585 1586
    } dest;

    enum qemuMigrationForwardType fwdType;
    union {
        virStreamPtr stream;
    } fwd;
};
1587 1588 1589

#define TUNNEL_SEND_BUF_SIZE 65536

1590 1591 1592 1593 1594 1595 1596
typedef struct _qemuMigrationIOThread qemuMigrationIOThread;
typedef qemuMigrationIOThread *qemuMigrationIOThreadPtr;
struct _qemuMigrationIOThread {
    virThread thread;
    virStreamPtr st;
    int sock;
    virError err;
1597 1598
    int wakeupRecvFD;
    int wakeupSendFD;
1599 1600 1601
};

static void qemuMigrationIOFunc(void *arg)
1602
{
1603
    qemuMigrationIOThreadPtr data = arg;
1604 1605 1606 1607 1608 1609 1610
    char *buffer = NULL;
    struct pollfd fds[2];
    int timeout = -1;
    virErrorPtr err = NULL;

    VIR_DEBUG("Running migration tunnel; stream=%p, sock=%d",
              data->st, data->sock);
1611 1612 1613

    if (VIR_ALLOC_N(buffer, TUNNEL_SEND_BUF_SIZE) < 0) {
        virReportOOMError();
1614
        goto abrt;
1615 1616
    }

1617 1618 1619
    fds[0].fd = data->sock;
    fds[1].fd = data->wakeupRecvFD;

1620
    for (;;) {
1621 1622 1623 1624 1625 1626 1627 1628 1629 1630
        int ret;

        fds[0].events = fds[1].events = POLLIN;
        fds[0].revents = fds[1].revents = 0;

        ret = poll(fds, ARRAY_CARDINALITY(fds), timeout);

        if (ret < 0) {
            if (errno == EAGAIN || errno == EINTR)
                continue;
1631
            virReportSystemError(errno, "%s",
1632 1633
                                 _("poll failed in migration tunnel"));
            goto abrt;
1634
        }
1635 1636 1637 1638 1639 1640 1641

        if (ret == 0) {
            /* We were asked to gracefully stop but reading would block. This
             * can only happen if qemu told us migration finished but didn't
             * close the migration fd. We handle this in the same way as EOF.
             */
            VIR_DEBUG("QEMU forgot to close migration fd");
1642
            break;
1643
        }
1644

1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660
        if (fds[1].revents & (POLLIN | POLLERR | POLLHUP)) {
            char stop = 0;

            if (saferead(data->wakeupRecvFD, &stop, 1) != 1) {
                virReportSystemError(errno, "%s",
                                     _("failed to read from wakeup fd"));
                goto abrt;
            }

            VIR_DEBUG("Migration tunnel was asked to %s",
                      stop ? "abort" : "finish");
            if (stop) {
                goto abrt;
            } else {
                timeout = 0;
            }
1661 1662
        }

1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679
        if (fds[0].revents & (POLLIN | POLLERR | POLLHUP)) {
            int nbytes;

            nbytes = saferead(data->sock, buffer, TUNNEL_SEND_BUF_SIZE);
            if (nbytes > 0) {
                if (virStreamSend(data->st, buffer, nbytes) < 0)
                    goto error;
            } else if (nbytes < 0) {
                virReportSystemError(errno, "%s",
                        _("tunnelled migration failed to read from qemu"));
                goto abrt;
            } else {
                /* EOF; get out of here */
                break;
            }
        }
    }
1680

1681 1682
    if (virStreamFinish(data->st) < 0)
        goto error;
1683

1684 1685
    VIR_FREE(buffer);

1686 1687
    return;

1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699
abrt:
    err = virSaveLastError();
    if (err && err->code == VIR_ERR_OK) {
        virFreeError(err);
        err = NULL;
    }
    virStreamAbort(data->st);
    if (err) {
        virSetError(err);
        virFreeError(err);
    }

1700 1701 1702
error:
    virCopyLastError(&data->err);
    virResetLastError();
1703
    VIR_FREE(buffer);
1704 1705 1706 1707 1708 1709 1710
}


static qemuMigrationIOThreadPtr
qemuMigrationStartTunnel(virStreamPtr st,
                         int sock)
{
1711 1712
    qemuMigrationIOThreadPtr io = NULL;
    int wakeupFD[2] = { -1, -1 };
1713

1714 1715 1716 1717
    if (pipe2(wakeupFD, O_CLOEXEC) < 0) {
        virReportSystemError(errno, "%s",
                             _("Unable to make pipe"));
        goto error;
1718 1719
    }

1720 1721 1722
    if (VIR_ALLOC(io) < 0)
        goto no_memory;

1723 1724
    io->st = st;
    io->sock = sock;
1725 1726
    io->wakeupRecvFD = wakeupFD[0];
    io->wakeupSendFD = wakeupFD[1];
1727 1728 1729 1730 1731 1732

    if (virThreadCreate(&io->thread, true,
                        qemuMigrationIOFunc,
                        io) < 0) {
        virReportSystemError(errno, "%s",
                             _("Unable to create migration thread"));
1733
        goto error;
1734 1735 1736
    }

    return io;
1737 1738 1739 1740 1741 1742 1743 1744

no_memory:
    virReportOOMError();
error:
    VIR_FORCE_CLOSE(wakeupFD[0]);
    VIR_FORCE_CLOSE(wakeupFD[1]);
    VIR_FREE(io);
    return NULL;
1745 1746 1747
}

static int
1748
qemuMigrationStopTunnel(qemuMigrationIOThreadPtr io, bool error)
1749 1750
{
    int rv = -1;
1751 1752 1753 1754 1755 1756 1757 1758 1759
    char stop = error ? 1 : 0;

    /* make sure the thread finishes its job and is joinable */
    if (safewrite(io->wakeupSendFD, &stop, 1) != 1) {
        virReportSystemError(errno, "%s",
                             _("failed to wakeup migration tunnel"));
        goto cleanup;
    }

1760 1761 1762 1763
    virThreadJoin(&io->thread);

    /* Forward error from the IO thread, to this thread */
    if (io->err.code != VIR_ERR_OK) {
1764 1765 1766 1767
        if (error)
            rv = 0;
        else
            virSetError(&io->err);
1768 1769 1770 1771 1772 1773 1774
        virResetError(&io->err);
        goto cleanup;
    }

    rv = 0;

cleanup:
1775 1776
    VIR_FORCE_CLOSE(io->wakeupSendFD);
    VIR_FORCE_CLOSE(io->wakeupRecvFD);
1777 1778
    VIR_FREE(io);
    return rv;
1779 1780
}

1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818
static int
qemuMigrationConnect(struct qemud_driver *driver,
                     virDomainObjPtr vm,
                     qemuMigrationSpecPtr spec)
{
    virNetSocketPtr sock;
    const char *host;
    char *port = NULL;
    int ret = -1;

    host = spec->dest.host.name;
    if (virAsprintf(&port, "%d", spec->dest.host.port) < 0) {
        virReportOOMError();
        return -1;
    }

    spec->destType = MIGRATION_DEST_FD;
    spec->dest.fd.qemu = -1;

    if (virSecurityManagerSetSocketLabel(driver->securityManager, vm->def) < 0)
        goto cleanup;
    if (virNetSocketNewConnectTCP(host, port, &sock) == 0) {
        spec->dest.fd.qemu = virNetSocketDupFD(sock, true);
        virNetSocketFree(sock);
    }
    if (virSecurityManagerClearSocketLabel(driver->securityManager, vm->def) < 0 ||
        spec->dest.fd.qemu == -1)
        goto cleanup;

    ret = 0;

cleanup:
    VIR_FREE(port);
    if (ret < 0)
        VIR_FORCE_CLOSE(spec->dest.fd.qemu);
    return ret;
}

1819 1820 1821 1822 1823 1824 1825 1826 1827
static int
qemuMigrationRun(struct qemud_driver *driver,
                 virDomainObjPtr vm,
                 const char *cookiein,
                 int cookieinlen,
                 char **cookieout,
                 int *cookieoutlen,
                 unsigned long flags,
                 unsigned long resource,
1828 1829
                 qemuMigrationSpecPtr spec,
                 virConnectPtr dconn)
1830
{
1831
    int ret = -1;
1832 1833
    unsigned int migrate_flags = QEMU_MONITOR_MIGRATE_BACKGROUND;
    qemuDomainObjPrivatePtr priv = vm->privateData;
1834
    qemuMigrationCookiePtr mig = NULL;
1835
    qemuMigrationIOThreadPtr iothread = NULL;
1836
    int fd = -1;
1837
    unsigned long migrate_speed = resource ? resource : priv->migMaxBandwidth;
1838
    virErrorPtr orig_err = NULL;
1839 1840 1841 1842 1843 1844 1845

    VIR_DEBUG("driver=%p, vm=%p, cookiein=%s, cookieinlen=%d, "
              "cookieout=%p, cookieoutlen=%p, flags=%lx, resource=%lu, "
              "spec=%p (dest=%d, fwd=%d)",
              driver, vm, NULLSTR(cookiein), cookieinlen,
              cookieout, cookieoutlen, flags, resource,
              spec, spec->destType, spec->fwdType);
1846

1847 1848 1849
    if (virLockManagerPluginUsesState(driver->lockManager) &&
        !cookieout) {
        qemuReportError(VIR_ERR_INTERNAL_ERROR,
1850 1851
                        _("Migration with lock driver %s requires"
                          " cookie support"),
1852 1853 1854 1855 1856
                        virLockManagerPluginGetName(driver->lockManager));
        return -1;
    }

    if (!(mig = qemuMigrationEatCookie(driver, vm, cookiein, cookieinlen,
1857 1858 1859 1860 1861 1862
                                       QEMU_MIGRATION_COOKIE_GRAPHICS)))
        goto cleanup;

    if (qemuDomainMigrateGraphicsRelocate(driver, vm, mig) < 0)
        VIR_WARN("unable to provide data for graphics client relocation");

1863
    /* Before EnterMonitor, since qemuMigrationSetOffline already does that */
1864 1865 1866 1867 1868 1869
    if (!(flags & VIR_MIGRATE_LIVE) &&
        virDomainObjGetState(vm, NULL) == VIR_DOMAIN_RUNNING) {
        if (qemuMigrationSetOffline(driver, vm) < 0)
            goto cleanup;
    }

1870 1871
    if (qemuDomainObjEnterMonitorAsync(driver, vm,
                                       QEMU_ASYNC_JOB_MIGRATION_OUT) < 0)
1872 1873
        goto cleanup;

1874
    if (qemuMonitorSetMigrationSpeed(priv->mon, migrate_speed) < 0) {
1875 1876 1877
        qemuDomainObjExitMonitorWithDriver(driver, vm);
        goto cleanup;
    }
1878

1879
    if (flags & VIR_MIGRATE_NON_SHARED_DISK)
1880 1881
        migrate_flags |= QEMU_MONITOR_MIGRATE_NON_SHARED_DISK;

1882
    if (flags & VIR_MIGRATE_NON_SHARED_INC)
1883 1884
        migrate_flags |= QEMU_MONITOR_MIGRATE_NON_SHARED_INC;

1885 1886
    /* connect to the destination qemu if needed */
    if (spec->destType == MIGRATION_DEST_CONNECT_HOST &&
1887 1888
        qemuMigrationConnect(driver, vm, spec) < 0) {
        qemuDomainObjExitMonitorWithDriver(driver, vm);
1889
        goto cleanup;
1890
    }
1891

1892 1893 1894 1895 1896 1897 1898
    switch (spec->destType) {
    case MIGRATION_DEST_HOST:
        ret = qemuMonitorMigrateToHost(priv->mon, migrate_flags,
                                       spec->dest.host.name,
                                       spec->dest.host.port);
        break;

1899 1900 1901 1902
    case MIGRATION_DEST_CONNECT_HOST:
        /* handled above and transformed into MIGRATION_DEST_FD */
        break;

1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913
    case MIGRATION_DEST_UNIX:
        if (qemuCapsGet(priv->qemuCaps, QEMU_CAPS_MIGRATE_QEMU_UNIX)) {
            ret = qemuMonitorMigrateToUnix(priv->mon, migrate_flags,
                                           spec->dest.unix_socket.file);
        } else {
            const char *args[] = {
                "nc", "-U", spec->dest.unix_socket.file, NULL
            };
            ret = qemuMonitorMigrateToCommand(priv->mon, migrate_flags, args);
        }
        break;
1914 1915

    case MIGRATION_DEST_FD:
1916
        if (spec->fwdType != MIGRATION_FWD_DIRECT) {
1917
            fd = spec->dest.fd.local;
1918 1919
            spec->dest.fd.local = -1;
        }
1920 1921 1922 1923
        ret = qemuMonitorMigrateToFd(priv->mon, migrate_flags,
                                     spec->dest.fd.qemu);
        VIR_FORCE_CLOSE(spec->dest.fd.qemu);
        break;
1924 1925
    }
    qemuDomainObjExitMonitorWithDriver(driver, vm);
1926
    if (ret < 0)
1927 1928
        goto cleanup;
    ret = -1;
1929 1930 1931 1932 1933 1934 1935 1936 1937 1938

    if (!virDomainObjIsActive(vm)) {
        qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                        _("guest unexpectedly quit"));
        goto cleanup;
    }

    /* From this point onwards we *must* call cancel to abort the
     * migration on source if anything goes wrong */

1939 1940 1941 1942 1943 1944 1945 1946
    if (spec->destType == MIGRATION_DEST_UNIX) {
        /* It is also possible that the migrate didn't fail initially, but
         * rather failed later on.  Check its status before waiting for a
         * connection from qemu which may never be initiated.
         */
        if (qemuMigrationUpdateJobStatus(driver, vm, _("migration job"),
                                         QEMU_ASYNC_JOB_MIGRATION_OUT) < 0)
            goto cancel;
1947

1948 1949 1950 1951 1952 1953 1954
        while ((fd = accept(spec->dest.unix_socket.sock, NULL, NULL)) < 0) {
            if (errno == EAGAIN || errno == EINTR)
                continue;
            virReportSystemError(errno, "%s",
                                 _("failed to accept connection from qemu"));
            goto cancel;
        }
1955 1956
    }

1957 1958
    if (spec->fwdType != MIGRATION_FWD_DIRECT &&
        !(iothread = qemuMigrationStartTunnel(spec->fwd.stream, fd)))
1959 1960
        goto cancel;

1961
    if (qemuMigrationWaitForCompletion(driver, vm,
1962 1963
                                       QEMU_ASYNC_JOB_MIGRATION_OUT,
                                       dconn) < 0)
1964
        goto cleanup;
1965

1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978
    /* When migration completed, QEMU will have paused the
     * CPUs for us, but unless we're using the JSON monitor
     * we won't have been notified of this, so might still
     * think we're running. For v2 protocol this doesn't
     * matter because we'll kill the VM soon, but for v3
     * this is important because we stay paused until the
     * confirm3 step, but need to release the lock state
     */
    if (virDomainObjGetState(vm, NULL) == VIR_DOMAIN_RUNNING) {
        if (qemuMigrationSetOffline(driver, vm) < 0)
            goto cleanup;
    }

1979
    ret = 0;
1980

1981
cleanup:
1982 1983 1984
    if (ret < 0 && !orig_err)
        orig_err = virSaveLastError();

1985
    if (spec->fwdType != MIGRATION_FWD_DIRECT) {
1986
        if (iothread && qemuMigrationStopTunnel(iothread, ret < 0) < 0)
1987
            ret = -1;
1988
        VIR_FORCE_CLOSE(fd);
1989
    }
1990

1991
    if (ret == 0 &&
1992 1993
        qemuMigrationBakeCookie(mig, driver, vm, cookieout, cookieoutlen,
                                QEMU_MIGRATION_COOKIE_PERSISTENT ) < 0)
1994 1995
        VIR_WARN("Unable to encode migration cookie");

1996 1997
    qemuMigrationCookieFree(mig);

1998 1999 2000 2001 2002
    if (orig_err) {
        virSetError(orig_err);
        virFreeError(orig_err);
    }

2003 2004
    return ret;

2005
cancel:
2006 2007
    orig_err = virSaveLastError();

2008
    if (virDomainObjIsActive(vm)) {
2009 2010
        if (qemuDomainObjEnterMonitorAsync(driver, vm,
                                           QEMU_ASYNC_JOB_MIGRATION_OUT) == 0) {
2011 2012 2013
            qemuMonitorMigrateCancel(priv->mon);
            qemuDomainObjExitMonitorWithDriver(driver, vm);
        }
2014
    }
2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028
    goto cleanup;
}

/* Perform migration using QEMU's native TCP migrate support,
 * not encrypted obviously
 */
static int doNativeMigrate(struct qemud_driver *driver,
                           virDomainObjPtr vm,
                           const char *uri,
                           const char *cookiein,
                           int cookieinlen,
                           char **cookieout,
                           int *cookieoutlen,
                           unsigned long flags,
2029 2030
                           unsigned long resource,
                           virConnectPtr dconn)
2031
{
2032
    qemuDomainObjPrivatePtr priv = vm->privateData;
M
Martin Kletzander 已提交
2033
    virURIPtr uribits = NULL;
2034
    int ret = -1;
2035 2036 2037 2038 2039 2040 2041 2042
    qemuMigrationSpec spec;

    VIR_DEBUG("driver=%p, vm=%p, uri=%s, cookiein=%s, cookieinlen=%d, "
              "cookieout=%p, cookieoutlen=%p, flags=%lx, resource=%lu",
              driver, vm, uri, NULLSTR(cookiein), cookieinlen,
              cookieout, cookieoutlen, flags, resource);

    if (STRPREFIX(uri, "tcp:") && !STRPREFIX(uri, "tcp://")) {
2043
        char *tmp;
2044
        /* HACK: source host generates bogus URIs, so fix them up */
2045
        if (virAsprintf(&tmp, "tcp://%s", uri + strlen("tcp:")) < 0) {
2046 2047 2048
            virReportOOMError();
            return -1;
        }
M
Martin Kletzander 已提交
2049
        uribits = virURIParse(tmp);
2050
        VIR_FREE(tmp);
2051
    } else {
M
Martin Kletzander 已提交
2052
        uribits = virURIParse(uri);
2053
    }
2054
    if (!uribits)
2055 2056
        return -1;

2057 2058 2059
    if (qemuCapsGet(priv->qemuCaps, QEMU_CAPS_MIGRATE_QEMU_FD))
        spec.destType = MIGRATION_DEST_CONNECT_HOST;
    else
2060
        spec.destType = MIGRATION_DEST_HOST;
2061 2062 2063
    spec.dest.host.name = uribits->server;
    spec.dest.host.port = uribits->port;
    spec.fwdType = MIGRATION_FWD_DIRECT;
2064

2065
    ret = qemuMigrationRun(driver, vm, cookiein, cookieinlen, cookieout,
2066
                           cookieoutlen, flags, resource, &spec, dconn);
2067 2068 2069 2070

    if (spec.destType == MIGRATION_DEST_FD)
        VIR_FORCE_CLOSE(spec.dest.fd.qemu);

2071
    virURIFree(uribits);
2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084

    return ret;
}


static int doTunnelMigrate(struct qemud_driver *driver,
                           virDomainObjPtr vm,
                           virStreamPtr st,
                           const char *cookiein,
                           int cookieinlen,
                           char **cookieout,
                           int *cookieoutlen,
                           unsigned long flags,
2085 2086
                           unsigned long resource,
                           virConnectPtr dconn)
2087 2088
{
    qemuDomainObjPrivatePtr priv = vm->privateData;
2089
    virNetSocketPtr sock = NULL;
2090 2091 2092 2093 2094 2095 2096 2097
    int ret = -1;
    qemuMigrationSpec spec;

    VIR_DEBUG("driver=%p, vm=%p, st=%p, cookiein=%s, cookieinlen=%d, "
              "cookieout=%p, cookieoutlen=%p, flags=%lx, resource=%lu",
              driver, vm, st, NULLSTR(cookiein), cookieinlen,
              cookieout, cookieoutlen, flags, resource);

2098 2099
    if (!qemuCapsGet(priv->qemuCaps, QEMU_CAPS_MIGRATE_QEMU_FD) &&
        !qemuCapsGet(priv->qemuCaps, QEMU_CAPS_MIGRATE_QEMU_UNIX) &&
2100
        !qemuCapsGet(priv->qemuCaps, QEMU_CAPS_MIGRATE_QEMU_EXEC)) {
2101 2102 2103
        qemuReportError(VIR_ERR_OPERATION_FAILED, "%s",
                        _("Source qemu is too old to support tunnelled migration"));
        return -1;
2104 2105 2106 2107 2108
    }

    spec.fwdType = MIGRATION_FWD_STREAM;
    spec.fwd.stream = st;

2109 2110 2111 2112 2113 2114 2115
    if (qemuCapsGet(priv->qemuCaps, QEMU_CAPS_MIGRATE_QEMU_FD)) {
        int fds[2];

        spec.destType = MIGRATION_DEST_FD;
        spec.dest.fd.qemu = -1;
        spec.dest.fd.local = -1;

2116
        if (pipe2(fds, O_CLOEXEC) == 0) {
2117 2118 2119 2120
            spec.dest.fd.qemu = fds[1];
            spec.dest.fd.local = fds[0];
        }
        if (spec.dest.fd.qemu == -1 ||
2121
            virSecurityManagerSetImageFDLabel(driver->securityManager, vm->def,
2122
                                              spec.dest.fd.qemu) < 0) {
2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147
            virReportSystemError(errno, "%s",
                        _("cannot create pipe for tunnelled migration"));
            goto cleanup;
        }
    } else {
        spec.destType = MIGRATION_DEST_UNIX;
        spec.dest.unix_socket.sock = -1;
        spec.dest.unix_socket.file = NULL;

        if (virAsprintf(&spec.dest.unix_socket.file,
                        "%s/qemu.tunnelmigrate.src.%s",
                        driver->libDir, vm->def->name) < 0) {
            virReportOOMError();
            goto cleanup;
        }

        if (virNetSocketNewListenUNIX(spec.dest.unix_socket.file, 0700,
                                      driver->user, driver->group,
                                      &sock) < 0 ||
            virNetSocketListen(sock, 1) < 0)
            goto cleanup;

        spec.dest.unix_socket.sock = virNetSocketGetFD(sock);
    }

2148
    ret = qemuMigrationRun(driver, vm, cookiein, cookieinlen, cookieout,
2149
                           cookieoutlen, flags, resource, &spec, dconn);
2150

2151
cleanup:
2152 2153 2154 2155 2156 2157 2158
    if (spec.destType == MIGRATION_DEST_FD) {
        VIR_FORCE_CLOSE(spec.dest.fd.qemu);
        VIR_FORCE_CLOSE(spec.dest.fd.local);
    } else {
        virNetSocketFree(sock);
        VIR_FREE(spec.dest.unix_socket.file);
    }
2159 2160 2161 2162 2163

    return ret;
}


2164 2165 2166 2167 2168
/* This is essentially a re-impl of virDomainMigrateVersion2
 * from libvirt.c, but running in source libvirtd context,
 * instead of client app context & also adding in tunnel
 * handling */
static int doPeer2PeerMigrate2(struct qemud_driver *driver,
2169
                               virConnectPtr sconn ATTRIBUTE_UNUSED,
2170 2171
                               virConnectPtr dconn,
                               virDomainObjPtr vm,
2172
                               const char *dconnuri,
2173 2174 2175
                               unsigned long flags,
                               const char *dname,
                               unsigned long resource)
2176 2177 2178
{
    virDomainPtr ddomain = NULL;
    char *uri_out = NULL;
2179
    char *cookie = NULL;
2180 2181 2182 2183 2184
    char *dom_xml = NULL;
    int cookielen = 0, ret;
    virErrorPtr orig_err = NULL;
    int cancelled;
    virStreamPtr st = NULL;
2185
    VIR_DEBUG("driver=%p, sconn=%p, dconn=%p, vm=%p, dconnuri=%s, "
2186
              "flags=%lx, dname=%s, resource=%lu",
2187 2188
              driver, sconn, dconn, vm, NULLSTR(dconnuri),
              flags, NULLSTR(dname), resource);
2189

2190 2191 2192 2193 2194 2195
    /* In version 2 of the protocol, the prepare step is slightly
     * different.  We fetch the domain XML of the source domain
     * and pass it to Prepare2.
     */
    if (!(dom_xml = qemuDomainFormatXML(driver, vm,
                                        VIR_DOMAIN_XML_SECURE |
2196 2197
                                        VIR_DOMAIN_XML_UPDATE_CPU,
                                        true)))
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 2225
        return -1;

    if (virDomainObjGetState(vm, NULL) == VIR_DOMAIN_PAUSED)
        flags |= VIR_MIGRATE_PAUSED;

    VIR_DEBUG("Prepare2 %p", dconn);
    if (flags & VIR_MIGRATE_TUNNELLED) {
        /*
         * Tunnelled Migrate Version 2 does not support cookies
         * due to missing parameters in the prepareTunnel() API.
         */

        if (!(st = virStreamNew(dconn, 0)))
            goto cleanup;

        qemuDomainObjEnterRemoteWithDriver(driver, vm);
        ret = dconn->driver->domainMigratePrepareTunnel
            (dconn, st, flags, dname, resource, dom_xml);
        qemuDomainObjExitRemoteWithDriver(driver, vm);
    } else {
        qemuDomainObjEnterRemoteWithDriver(driver, vm);
        ret = dconn->driver->domainMigratePrepare2
            (dconn, &cookie, &cookielen, NULL, &uri_out,
             flags, dname, resource, dom_xml);
        qemuDomainObjExitRemoteWithDriver(driver, vm);
    }
    VIR_FREE(dom_xml);
    if (ret == -1)
2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236
        goto cleanup;

    /* the domain may have shutdown or crashed while we had the locks dropped
     * in qemuDomainObjEnterRemoteWithDriver, so check again
     */
    if (!virDomainObjIsActive(vm)) {
        qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                        _("guest unexpectedly quit"));
        goto cleanup;
    }

2237 2238 2239
    if (!(flags & VIR_MIGRATE_TUNNELLED) &&
        (uri_out == NULL)) {
        qemuReportError(VIR_ERR_INTERNAL_ERROR,
2240
                        _("domainMigratePrepare2 did not set uri"));
2241 2242
        cancelled = 1;
        goto finish;
2243 2244
    }

2245 2246 2247 2248
    /* Perform the migration.  The driver isn't supposed to return
     * until the migration is complete.
     */
    VIR_DEBUG("Perform %p", sconn);
2249
    qemuMigrationJobSetPhase(driver, vm, QEMU_MIGRATION_PHASE_PERFORM2);
2250
    if (flags & VIR_MIGRATE_TUNNELLED)
2251 2252
        ret = doTunnelMigrate(driver, vm, st,
                              NULL, 0, NULL, NULL,
2253
                              flags, resource, dconn);
2254 2255 2256 2257
    else
        ret = doNativeMigrate(driver, vm, uri_out,
                              cookie, cookielen,
                              NULL, NULL, /* No out cookie with v2 migration */
2258
                              flags, resource, dconn);
2259 2260 2261 2262

    /* Perform failed. Make sure Finish doesn't overwrite the error */
    if (ret < 0)
        orig_err = virSaveLastError();
2263

2264 2265 2266 2267
    /* If Perform returns < 0, then we need to cancel the VM
     * startup on the destination
     */
    cancelled = ret < 0 ? 1 : 0;
2268 2269

finish:
2270 2271 2272 2273
    /* In version 2 of the migration protocol, we pass the
     * status code from the sender to the destination host,
     * so it can do any cleanup if the migration failed.
     */
2274
    dname = dname ? dname : vm->def->name;
2275
    VIR_DEBUG("Finish2 %p ret=%d", dconn, ret);
2276 2277
    qemuDomainObjEnterRemoteWithDriver(driver, vm);
    ddomain = dconn->driver->domainMigrateFinish2
2278
        (dconn, dname, cookie, cookielen,
2279
         uri_out ? uri_out : dconnuri, flags, cancelled);
2280 2281
    qemuDomainObjExitRemoteWithDriver(driver, vm);

2282 2283
cleanup:
    if (ddomain) {
2284
        virUnrefDomain(ddomain);
2285 2286 2287 2288
        ret = 0;
    } else {
        ret = -1;
    }
2289

2290 2291 2292 2293 2294 2295 2296 2297
    if (st)
        virUnrefStream(st);

    if (orig_err) {
        virSetError(orig_err);
        virFreeError(orig_err);
    }
    VIR_FREE(uri_out);
2298
    VIR_FREE(cookie);
2299 2300

    return ret;
2301 2302 2303
}


2304 2305 2306 2307 2308 2309 2310 2311
/* This is essentially a re-impl of virDomainMigrateVersion3
 * from libvirt.c, but running in source libvirtd context,
 * instead of client app context & also adding in tunnel
 * handling */
static int doPeer2PeerMigrate3(struct qemud_driver *driver,
                               virConnectPtr sconn,
                               virConnectPtr dconn,
                               virDomainObjPtr vm,
2312
                               const char *xmlin,
2313
                               const char *dconnuri,
2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329
                               const char *uri,
                               unsigned long flags,
                               const char *dname,
                               unsigned long resource)
{
    virDomainPtr ddomain = NULL;
    char *uri_out = NULL;
    char *cookiein = NULL;
    char *cookieout = NULL;
    char *dom_xml = NULL;
    int cookieinlen = 0;
    int cookieoutlen = 0;
    int ret = -1;
    virErrorPtr orig_err = NULL;
    int cancelled;
    virStreamPtr st = NULL;
2330
    VIR_DEBUG("driver=%p, sconn=%p, dconn=%p, vm=%p, xmlin=%s, "
2331
              "dconnuri=%s, uri=%s, flags=%lx, dname=%s, resource=%lu",
2332 2333 2334
              driver, sconn, dconn, vm, NULLSTR(xmlin),
              NULLSTR(dconnuri), NULLSTR(uri), flags,
              NULLSTR(dname), resource);
2335

2336 2337 2338 2339 2340
    /* Unlike the virDomainMigrateVersion3 counterpart, we don't need
     * to worry about auto-setting the VIR_MIGRATE_CHANGE_PROTECTION
     * bit here, because we are already running inside the context of
     * a single job.  */

2341
    dom_xml = qemuMigrationBegin(driver, vm, xmlin, dname,
2342
                                 &cookieout, &cookieoutlen, flags);
2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367
    if (!dom_xml)
        goto cleanup;

    if (virDomainObjGetState(vm, NULL) == VIR_DOMAIN_PAUSED)
        flags |= VIR_MIGRATE_PAUSED;

    VIR_DEBUG("Prepare3 %p", dconn);
    cookiein = cookieout;
    cookieinlen = cookieoutlen;
    cookieout = NULL;
    cookieoutlen = 0;
    if (flags & VIR_MIGRATE_TUNNELLED) {
        if (!(st = virStreamNew(dconn, 0)))
            goto cleanup;

        qemuDomainObjEnterRemoteWithDriver(driver, vm);
        ret = dconn->driver->domainMigratePrepareTunnel3
            (dconn, st, cookiein, cookieinlen,
             &cookieout, &cookieoutlen,
             flags, dname, resource, dom_xml);
        qemuDomainObjExitRemoteWithDriver(driver, vm);
    } else {
        qemuDomainObjEnterRemoteWithDriver(driver, vm);
        ret = dconn->driver->domainMigratePrepare3
            (dconn, cookiein, cookieinlen, &cookieout, &cookieoutlen,
2368
             uri, &uri_out, flags, dname, resource, dom_xml);
2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387
        qemuDomainObjExitRemoteWithDriver(driver, vm);
    }
    VIR_FREE(dom_xml);
    if (ret == -1)
        goto cleanup;

    if (!(flags & VIR_MIGRATE_TUNNELLED) &&
        (uri_out == NULL)) {
        qemuReportError(VIR_ERR_INTERNAL_ERROR,
                        _("domainMigratePrepare3 did not set uri"));
        cancelled = 1;
        goto finish;
    }

    /* Perform the migration.  The driver isn't supposed to return
     * until the migration is complete. The src VM should remain
     * running, but in paused state until the destination can
     * confirm migration completion.
     */
2388
    VIR_DEBUG("Perform3 %p uri=%s uri_out=%s", sconn, uri, uri_out);
2389
    qemuMigrationJobSetPhase(driver, vm, QEMU_MIGRATION_PHASE_PERFORM3);
2390 2391 2392 2393 2394 2395 2396 2397 2398
    VIR_FREE(cookiein);
    cookiein = cookieout;
    cookieinlen = cookieoutlen;
    cookieout = NULL;
    cookieoutlen = 0;
    if (flags & VIR_MIGRATE_TUNNELLED)
        ret = doTunnelMigrate(driver, vm, st,
                              cookiein, cookieinlen,
                              &cookieout, &cookieoutlen,
2399
                              flags, resource, dconn);
2400 2401 2402 2403
    else
        ret = doNativeMigrate(driver, vm, uri_out,
                              cookiein, cookieinlen,
                              &cookieout, &cookieoutlen,
2404
                              flags, resource, dconn);
2405 2406

    /* Perform failed. Make sure Finish doesn't overwrite the error */
2407
    if (ret < 0) {
2408
        orig_err = virSaveLastError();
2409 2410 2411 2412
    } else {
        qemuMigrationJobSetPhase(driver, vm,
                                 QEMU_MIGRATION_PHASE_PERFORM3_DONE);
    }
2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433

    /* If Perform returns < 0, then we need to cancel the VM
     * startup on the destination
     */
    cancelled = ret < 0 ? 1 : 0;

finish:
    /*
     * The status code from the source is passed to the destination.
     * The dest can cleanup in the source indicated it failed to
     * send all migration data. Returns NULL for ddomain if
     * the dest was unable to complete migration.
     */
    VIR_DEBUG("Finish3 %p ret=%d", dconn, ret);
    VIR_FREE(cookiein);
    cookiein = cookieout;
    cookieinlen = cookieoutlen;
    cookieout = NULL;
    cookieoutlen = 0;
    dname = dname ? dname : vm->def->name;
    qemuDomainObjEnterRemoteWithDriver(driver, vm);
2434
    ddomain = dconn->driver->domainMigrateFinish3
2435
        (dconn, dname, cookiein, cookieinlen, &cookieout, &cookieoutlen,
2436
         dconnuri, uri_out ? uri_out : uri, flags, cancelled);
2437 2438
    qemuDomainObjExitRemoteWithDriver(driver, vm);

2439 2440 2441 2442 2443 2444 2445
    /* If ddomain is NULL, then we were unable to start
     * the guest on the target, and must restart on the
     * source. There is a small chance that the ddomain
     * is NULL due to an RPC failure, in which case
     * ddomain could in fact be running on the dest.
     * The lock manager plugins should take care of
     * safety in this scenario.
2446
     */
2447
    cancelled = ddomain == NULL ? 1 : 0;
2448

2449 2450 2451 2452 2453 2454
    /* If finish3 set an error, and we don't have an earlier
     * one we need to preserve it in case confirm3 overwrites
     */
    if (!orig_err)
        orig_err = virSaveLastError();

2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466
    /*
     * If cancelled, then src VM will be restarted, else
     * it will be killed
     */
    VIR_DEBUG("Confirm3 %p ret=%d vm=%p", sconn, ret, vm);
    VIR_FREE(cookiein);
    cookiein = cookieout;
    cookieinlen = cookieoutlen;
    cookieout = NULL;
    cookieoutlen = 0;
    ret = qemuMigrationConfirm(driver, sconn, vm,
                               cookiein, cookieinlen,
2467
                               flags, cancelled);
2468 2469 2470 2471
    /* If Confirm3 returns -1, there's nothing more we can
     * do, but fortunately worst case is that there is a
     * domain left in 'paused' state on source.
     */
2472 2473 2474
    if (ret < 0)
        VIR_WARN("Guest %s probably left in 'paused' state on source",
                 vm->def->name);
2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 2496 2497 2498

 cleanup:
    if (ddomain) {
        virUnrefDomain(ddomain);
        ret = 0;
    } else {
        ret = -1;
    }

    if (st)
        virUnrefStream(st);

    if (orig_err) {
        virSetError(orig_err);
        virFreeError(orig_err);
    }
    VIR_FREE(uri_out);
    VIR_FREE(cookiein);
    VIR_FREE(cookieout);

    return ret;
}


2499
static int doPeer2PeerMigrate(struct qemud_driver *driver,
2500
                              virConnectPtr sconn,
2501
                              virDomainObjPtr vm,
2502
                              const char *xmlin,
2503
                              const char *dconnuri,
2504 2505 2506
                              const char *uri,
                              unsigned long flags,
                              const char *dname,
2507 2508
                              unsigned long resource,
                              bool *v3proto)
2509 2510 2511 2512
{
    int ret = -1;
    virConnectPtr dconn = NULL;
    bool p2p;
2513 2514
    virErrorPtr orig_err = NULL;

2515
    VIR_DEBUG("driver=%p, sconn=%p, vm=%p, xmlin=%s, dconnuri=%s, "
2516
              "uri=%s, flags=%lx, dname=%s, resource=%lu",
2517 2518
              driver, sconn, vm, NULLSTR(xmlin), NULLSTR(dconnuri),
              NULLSTR(uri), flags, NULLSTR(dname), resource);
2519 2520 2521 2522 2523 2524

    /* the order of operations is important here; we make sure the
     * destination side is completely setup before we touch the source
     */

    qemuDomainObjEnterRemoteWithDriver(driver, vm);
2525
    dconn = virConnectOpen(dconnuri);
2526 2527 2528
    qemuDomainObjExitRemoteWithDriver(driver, vm);
    if (dconn == NULL) {
        qemuReportError(VIR_ERR_OPERATION_FAILED,
2529
                        _("Failed to connect to remote libvirt URI %s"), dconnuri);
2530 2531 2532
        return -1;
    }

2533 2534 2535 2536
    if (virConnectSetKeepAlive(dconn, driver->keepAliveInterval,
                               driver->keepAliveCount) < 0)
        goto cleanup;

2537 2538 2539
    qemuDomainObjEnterRemoteWithDriver(driver, vm);
    p2p = VIR_DRV_SUPPORTS_FEATURE(dconn->driver, dconn,
                                   VIR_DRV_FEATURE_MIGRATION_P2P);
2540
        /* v3proto reflects whether the caller used Perform3, but with
2541
         * p2p migrate, regardless of whether Perform2 or Perform3
2542 2543 2544 2545
         * were used, we decide protocol based on what target supports
         */
    *v3proto = VIR_DRV_SUPPORTS_FEATURE(dconn->driver, dconn,
                                        VIR_DRV_FEATURE_MIGRATION_V3);
2546
    qemuDomainObjExitRemoteWithDriver(driver, vm);
2547

2548 2549 2550 2551 2552 2553 2554 2555 2556 2557 2558 2559 2560
    if (!p2p) {
        qemuReportError(VIR_ERR_OPERATION_FAILED, "%s",
                        _("Destination libvirt does not support peer-to-peer migration protocol"));
        goto cleanup;
    }

    /* domain may have been stopped while we were talking to remote daemon */
    if (!virDomainObjIsActive(vm)) {
        qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                        _("guest unexpectedly quit"));
        goto cleanup;
    }

2561 2562 2563 2564 2565 2566 2567
    /* Change protection is only required on the source side (us), and
     * only for v3 migration when begin and perform are separate jobs.
     * But peer-2-peer is already a single job, and we still want to
     * talk to older destinations that would reject the flag.
     * Therefore it is safe to clear the bit here.  */
    flags &= ~VIR_MIGRATE_CHANGE_PROTECTION;

2568
    if (*v3proto)
2569
        ret = doPeer2PeerMigrate3(driver, sconn, dconn, vm, xmlin,
2570
                                  dconnuri, uri, flags, dname, resource);
2571 2572
    else
        ret = doPeer2PeerMigrate2(driver, sconn, dconn, vm,
2573
                                  dconnuri, flags, dname, resource);
2574 2575

cleanup:
2576
    orig_err = virSaveLastError();
2577
    qemuDomainObjEnterRemoteWithDriver(driver, vm);
2578
    virConnectClose(dconn);
2579
    qemuDomainObjExitRemoteWithDriver(driver, vm);
2580 2581 2582 2583
    if (orig_err) {
        virSetError(orig_err);
        virFreeError(orig_err);
    }
2584 2585 2586 2587 2588

    return ret;
}


2589 2590 2591 2592 2593 2594 2595 2596 2597 2598 2599 2600 2601 2602 2603 2604 2605 2606 2607 2608
/*
 * This implements perform part of the migration protocol when migration job
 * does not need to be active across several APIs, i.e., peer2peer migration or
 * perform phase of v2 non-peer2peer migration.
 */
static int
qemuMigrationPerformJob(struct qemud_driver *driver,
                        virConnectPtr conn,
                        virDomainObjPtr vm,
                        const char *xmlin,
                        const char *dconnuri,
                        const char *uri,
                        const char *cookiein,
                        int cookieinlen,
                        char **cookieout,
                        int *cookieoutlen,
                        unsigned long flags,
                        const char *dname,
                        unsigned long resource,
                        bool v3proto)
2609 2610 2611 2612
{
    virDomainEventPtr event = NULL;
    int ret = -1;
    int resume = 0;
2613
    virErrorPtr orig_err = NULL;
2614

2615
    if (qemuMigrationJobStart(driver, vm, QEMU_ASYNC_JOB_MIGRATION_OUT) < 0)
2616 2617 2618 2619 2620 2621 2622 2623
        goto cleanup;

    if (!virDomainObjIsActive(vm)) {
        qemuReportError(VIR_ERR_OPERATION_INVALID,
                        "%s", _("domain is not running"));
        goto endjob;
    }

2624 2625
    if (!qemuMigrationIsAllowed(driver, vm, NULL))
        goto cleanup;
2626

2627 2628 2629
    if (!(flags & VIR_MIGRATE_UNSAFE) && !qemuMigrationIsSafe(vm->def))
        goto cleanup;

J
Jiri Denemark 已提交
2630
    resume = virDomainObjGetState(vm, NULL) == VIR_DOMAIN_RUNNING;
2631 2632

    if ((flags & (VIR_MIGRATE_TUNNELLED | VIR_MIGRATE_PEER2PEER))) {
2633 2634 2635
        ret = doPeer2PeerMigrate(driver, conn, vm, xmlin,
                                 dconnuri, uri, flags, dname,
                                 resource, &v3proto);
2636
    } else {
2637 2638 2639
        qemuMigrationJobSetPhase(driver, vm, QEMU_MIGRATION_PHASE_PERFORM2);
        ret = doNativeMigrate(driver, vm, uri, cookiein, cookieinlen,
                              cookieout, cookieoutlen,
2640
                              flags, resource, NULL);
2641
    }
2642 2643
    if (ret < 0)
        goto endjob;
2644

2645 2646 2647 2648
    /*
     * In v3 protocol, the source VM is not killed off until the
     * confirm step.
     */
2649
    if (!v3proto) {
2650 2651
        qemuProcessStop(driver, vm, VIR_DOMAIN_SHUTOFF_MIGRATED,
                        VIR_QEMU_PROCESS_STOP_MIGRATED);
2652
        virDomainAuditStop(vm, "migrated");
2653 2654 2655
        event = virDomainEventNewFromObj(vm,
                                         VIR_DOMAIN_EVENT_STOPPED,
                                         VIR_DOMAIN_EVENT_STOPPED_MIGRATED);
2656
    }
2657
    resume = 0;
2658 2659

endjob:
2660 2661 2662
    if (ret < 0)
        orig_err = virSaveLastError();

J
Jiri Denemark 已提交
2663
    if (resume && virDomainObjGetState(vm, NULL) == VIR_DOMAIN_PAUSED) {
2664
        /* we got here through some sort of failure; start the domain again */
J
Jiri Denemark 已提交
2665
        if (qemuProcessStartCPUs(driver, vm, conn,
2666 2667
                                 VIR_DOMAIN_RUNNING_MIGRATION_CANCELED,
                                 QEMU_ASYNC_JOB_MIGRATION_OUT) < 0) {
2668 2669 2670 2671 2672 2673 2674 2675 2676 2677 2678 2679
            /* Hm, we already know we are in error here.  We don't want to
             * overwrite the previous error, though, so we just throw something
             * to the logs and hope for the best
             */
            VIR_ERROR(_("Failed to resume guest %s after failure"),
                      vm->def->name);
        }

        event = virDomainEventNewFromObj(vm,
                                         VIR_DOMAIN_EVENT_RESUMED,
                                         VIR_DOMAIN_EVENT_RESUMED_MIGRATED);
    }
2680 2681 2682 2683 2684 2685 2686 2687

    if (qemuMigrationJobFinish(driver, vm) == 0) {
        vm = NULL;
    } else if (!virDomainObjIsActive(vm) &&
               (!vm->persistent ||
                (ret == 0 && (flags & VIR_MIGRATE_UNDEFINE_SOURCE)))) {
        if (flags & VIR_MIGRATE_UNDEFINE_SOURCE)
            virDomainDeleteConfig(driver->configDir, driver->autostartDir, vm);
2688
        qemuDomainRemoveInactive(driver, vm);
2689 2690 2691
        vm = NULL;
    }

2692 2693 2694 2695 2696
    if (orig_err) {
        virSetError(orig_err);
        virFreeError(orig_err);
    }

2697 2698 2699 2700 2701 2702 2703 2704 2705 2706 2707 2708 2709 2710 2711 2712 2713 2714 2715 2716 2717 2718 2719 2720 2721 2722 2723 2724 2725 2726 2727 2728 2729 2730 2731 2732 2733
cleanup:
    if (vm)
        virDomainObjUnlock(vm);
    if (event)
        qemuDomainEventQueue(driver, event);
    return ret;
}

/*
 * This implements perform phase of v3 migration protocol.
 */
static int
qemuMigrationPerformPhase(struct qemud_driver *driver,
                          virConnectPtr conn,
                          virDomainObjPtr vm,
                          const char *uri,
                          const char *cookiein,
                          int cookieinlen,
                          char **cookieout,
                          int *cookieoutlen,
                          unsigned long flags,
                          unsigned long resource)
{
    virDomainEventPtr event = NULL;
    int ret = -1;
    bool resume;
    int refs;

    /* If we didn't start the job in the begin phase, start it now. */
    if (!(flags & VIR_MIGRATE_CHANGE_PROTECTION)) {
        if (qemuMigrationJobStart(driver, vm, QEMU_ASYNC_JOB_MIGRATION_OUT) < 0)
            goto cleanup;
    } else if (!qemuMigrationJobIsActive(vm, QEMU_ASYNC_JOB_MIGRATION_OUT)) {
        goto cleanup;
    }

    qemuMigrationJobStartPhase(driver, vm, QEMU_MIGRATION_PHASE_PERFORM3);
2734
    qemuDriverCloseCallbackUnset(driver, vm, qemuMigrationCleanup);
2735 2736 2737 2738

    resume = virDomainObjGetState(vm, NULL) == VIR_DOMAIN_RUNNING;
    ret = doNativeMigrate(driver, vm, uri, cookiein, cookieinlen,
                          cookieout, cookieoutlen,
2739
                          flags, resource, NULL);
2740 2741 2742 2743 2744

    if (ret < 0 && resume &&
        virDomainObjGetState(vm, NULL) == VIR_DOMAIN_PAUSED) {
        /* we got here through some sort of failure; start the domain again */
        if (qemuProcessStartCPUs(driver, vm, conn,
2745 2746
                                 VIR_DOMAIN_RUNNING_MIGRATION_CANCELED,
                                 QEMU_ASYNC_JOB_MIGRATION_OUT) < 0) {
2747 2748 2749 2750 2751 2752
            /* Hm, we already know we are in error here.  We don't want to
             * overwrite the previous error, though, so we just throw something
             * to the logs and hope for the best
             */
            VIR_ERROR(_("Failed to resume guest %s after failure"),
                      vm->def->name);
2753
        }
2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764

        event = virDomainEventNewFromObj(vm,
                                         VIR_DOMAIN_EVENT_RESUMED,
                                         VIR_DOMAIN_EVENT_RESUMED_MIGRATED);
    }

    if (ret < 0)
        goto endjob;

    qemuMigrationJobSetPhase(driver, vm, QEMU_MIGRATION_PHASE_PERFORM3_DONE);

2765 2766 2767 2768
    if (qemuDriverCloseCallbackSet(driver, vm, conn,
                                   qemuMigrationCleanup) < 0)
        goto endjob;

2769 2770 2771 2772 2773 2774 2775 2776
endjob:
    if (ret < 0)
        refs = qemuMigrationJobFinish(driver, vm);
    else
        refs = qemuMigrationJobContinue(vm);
    if (refs == 0) {
        vm = NULL;
    } else if (!virDomainObjIsActive(vm) && !vm->persistent) {
2777
        qemuDomainRemoveInactive(driver, vm);
2778
        vm = NULL;
2779
    }
2780 2781 2782 2783 2784 2785 2786 2787 2788

cleanup:
    if (vm)
        virDomainObjUnlock(vm);
    if (event)
        qemuDomainEventQueue(driver, event);
    return ret;
}

2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799 2800 2801 2802 2803 2804 2805 2806 2807 2808 2809 2810 2811 2812 2813 2814 2815 2816 2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834
int
qemuMigrationPerform(struct qemud_driver *driver,
                     virConnectPtr conn,
                     virDomainObjPtr vm,
                     const char *xmlin,
                     const char *dconnuri,
                     const char *uri,
                     const char *cookiein,
                     int cookieinlen,
                     char **cookieout,
                     int *cookieoutlen,
                     unsigned long flags,
                     const char *dname,
                     unsigned long resource,
                     bool v3proto)
{
    VIR_DEBUG("driver=%p, conn=%p, vm=%p, xmlin=%s, dconnuri=%s, "
              "uri=%s, cookiein=%s, cookieinlen=%d, cookieout=%p, "
              "cookieoutlen=%p, flags=%lx, dname=%s, resource=%lu, v3proto=%d",
              driver, conn, vm, NULLSTR(xmlin), NULLSTR(dconnuri),
              NULLSTR(uri), NULLSTR(cookiein), cookieinlen,
              cookieout, cookieoutlen, flags, NULLSTR(dname),
              resource, v3proto);

    if ((flags & (VIR_MIGRATE_TUNNELLED | VIR_MIGRATE_PEER2PEER))) {
        if (cookieinlen) {
            qemuReportError(VIR_ERR_OPERATION_INVALID,
                            "%s", _("received unexpected cookie with P2P migration"));
            return -1;
        }

        return qemuMigrationPerformJob(driver, conn, vm, xmlin, dconnuri, uri,
                                       cookiein, cookieinlen, cookieout,
                                       cookieoutlen, flags, dname, resource,
                                       v3proto);
    } else {
        if (dconnuri) {
            qemuReportError(VIR_ERR_INTERNAL_ERROR,
                            "%s", _("Unexpected dconnuri parameter with non-peer2peer migration"));
            return -1;
        }

        if (v3proto) {
            return qemuMigrationPerformPhase(driver, conn, vm, uri,
                                             cookiein, cookieinlen,
                                             cookieout, cookieoutlen,
2835
                                             flags, resource);
2836 2837 2838 2839 2840 2841 2842 2843
        } else {
            return qemuMigrationPerformJob(driver, conn, vm, xmlin, dconnuri,
                                           uri, cookiein, cookieinlen,
                                           cookieout, cookieoutlen, flags,
                                           dname, resource, v3proto);
        }
    }
}
2844

2845
static int
2846 2847 2848 2849 2850 2851 2852
qemuMigrationVPAssociatePortProfiles(virDomainDefPtr def) {
    int i;
    int last_good_net = -1;
    virDomainNetDefPtr net;

    for (i = 0; i < def->nnets; i++) {
        net = def->nets[i];
2853
        if (virDomainNetGetActualType(net) == VIR_DOMAIN_NET_TYPE_DIRECT) {
2854
            if (virNetDevVPortProfileAssociate(net->ifname,
2855
                                               virDomainNetGetActualVirtPortProfile(net),
2856 2857
                                               net->mac,
                                               virDomainNetGetActualDirectDev(net),
2858
                                               -1,
2859
                                               def->uuid,
2860 2861 2862 2863 2864
                                               VIR_NETDEV_VPORT_PROFILE_OP_MIGRATE_IN_FINISH,
                                               false) < 0) {
                qemuReportError(VIR_ERR_OPERATION_FAILED,
                                _("Port profile Associate failed for %s"),
                                net->ifname);
2865
                goto err_exit;
2866 2867
            }
            VIR_DEBUG("Port profile Associate succeeded for %s", net->ifname);
2868 2869 2870 2871 2872 2873

            if (virNetDevMacVLanVPortProfileRegisterCallback(net->ifname, net->mac,
                                                             virDomainNetGetActualDirectDev(net), def->uuid,
                                                             virDomainNetGetActualVirtPortProfile(net),
                                                             VIR_NETDEV_VPORT_PROFILE_OP_CREATE))
                goto err_exit;
2874 2875 2876 2877
        }
        last_good_net = i;
    }

2878
    return 0;
2879 2880 2881 2882

err_exit:
    for (i = 0; i < last_good_net; i++) {
        net = def->nets[i];
2883
        if (virDomainNetGetActualType(net) == VIR_DOMAIN_NET_TYPE_DIRECT) {
2884
            ignore_value(virNetDevVPortProfileDisassociate(net->ifname,
2885
                                                           virDomainNetGetActualVirtPortProfile(net),
2886 2887
                                                           net->mac,
                                                           virDomainNetGetActualDirectDev(net),
2888
                                                           -1,
2889
                                                           VIR_NETDEV_VPORT_PROFILE_OP_MIGRATE_IN_FINISH));
2890 2891
        }
    }
2892
    return -1;
2893 2894 2895 2896 2897 2898 2899
}


virDomainPtr
qemuMigrationFinish(struct qemud_driver *driver,
                    virConnectPtr dconn,
                    virDomainObjPtr vm,
2900 2901 2902 2903
                    const char *cookiein,
                    int cookieinlen,
                    char **cookieout,
                    int *cookieoutlen,
2904
                    unsigned long flags,
2905 2906
                    int retcode,
                    bool v3proto)
2907 2908 2909 2910
{
    virDomainPtr dom = NULL;
    virDomainEventPtr event = NULL;
    int newVM = 1;
2911
    qemuMigrationCookiePtr mig = NULL;
2912
    virErrorPtr orig_err = NULL;
2913
    int cookie_flags = 0;
J
Jiri Denemark 已提交
2914
    qemuDomainObjPrivatePtr priv = vm->privateData;
2915

2916
    VIR_DEBUG("driver=%p, dconn=%p, vm=%p, cookiein=%s, cookieinlen=%d, "
2917
              "cookieout=%p, cookieoutlen=%p, flags=%lx, retcode=%d",
2918 2919
              driver, dconn, vm, NULLSTR(cookiein), cookieinlen,
              cookieout, cookieoutlen, flags, retcode);
2920

2921
    if (!qemuMigrationJobIsActive(vm, QEMU_ASYNC_JOB_MIGRATION_IN))
2922 2923
        goto cleanup;

2924 2925 2926
    qemuMigrationJobStartPhase(driver, vm,
                               v3proto ? QEMU_MIGRATION_PHASE_FINISH3
                                       : QEMU_MIGRATION_PHASE_FINISH2);
2927

2928 2929
    qemuDomainCleanupRemove(vm, qemuMigrationPrepareCleanup);

2930 2931 2932 2933 2934
    if (flags & VIR_MIGRATE_PERSIST_DEST)
        cookie_flags |= QEMU_MIGRATION_COOKIE_PERSISTENT;

    if (!(mig = qemuMigrationEatCookie(driver, vm, cookiein,
                                       cookieinlen, cookie_flags)))
2935
        goto endjob;
2936 2937 2938 2939 2940 2941 2942 2943

    /* Did the migration go as planned?  If yes, return the domain
     * object, but if no, clean up the empty qemu process.
     */
    if (retcode == 0) {
        if (!virDomainObjIsActive(vm)) {
            qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                            _("guest unexpectedly quit"));
2944
            goto endjob;
2945 2946
        }

2947
        if (qemuMigrationVPAssociatePortProfiles(vm->def) < 0) {
2948 2949
            qemuProcessStop(driver, vm, VIR_DOMAIN_SHUTOFF_FAILED,
                            VIR_QEMU_PROCESS_STOP_MIGRATED);
2950 2951 2952 2953 2954 2955
            virDomainAuditStop(vm, "failed");
            event = virDomainEventNewFromObj(vm,
                                             VIR_DOMAIN_EVENT_STOPPED,
                                             VIR_DOMAIN_EVENT_STOPPED_FAILED);
            goto endjob;
        }
2956 2957

        if (flags & VIR_MIGRATE_PERSIST_DEST) {
2958
            virDomainDefPtr vmdef;
2959 2960 2961
            if (vm->persistent)
                newVM = 0;
            vm->persistent = 1;
2962 2963 2964 2965
            if (mig->persistent)
                vm->newDef = vmdef = mig->persistent;
            else
                vmdef = virDomainObjGetPersistentDef(driver->caps, vm);
A
Alex Jia 已提交
2966
            if (!vmdef || virDomainSaveConfig(driver->configDir, vmdef) < 0) {
2967 2968 2969 2970 2971 2972 2973 2974 2975
                /* Hmpf.  Migration was successful, but making it persistent
                 * was not.  If we report successful, then when this domain
                 * shuts down, management tools are in for a surprise.  On the
                 * other hand, if we report failure, then the management tools
                 * might try to restart the domain on the source side, even
                 * though the domain is actually running on the destination.
                 * Return a NULL dom pointer, and hope that this is a rare
                 * situation and management tools are smart.
                 */
2976 2977

                /*
2978 2979
                 * However, in v3 protocol, the source VM is still available
                 * to restart during confirm() step, so we kill it off now.
2980 2981
                 */
                if (v3proto) {
2982 2983
                    qemuProcessStop(driver, vm, VIR_DOMAIN_SHUTOFF_FAILED,
                                    VIR_QEMU_PROCESS_STOP_MIGRATED);
2984
                    virDomainAuditStop(vm, "failed");
2985 2986
                    if (newVM)
                        vm->persistent = 0;
2987
                }
A
Alex Jia 已提交
2988 2989 2990
                if (!vmdef)
                    qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                                    _("can't get vmdef"));
2991 2992 2993 2994 2995 2996 2997 2998 2999 3000 3001 3002 3003 3004 3005 3006 3007 3008
                goto endjob;
            }

            event = virDomainEventNewFromObj(vm,
                                             VIR_DOMAIN_EVENT_DEFINED,
                                             newVM ?
                                             VIR_DOMAIN_EVENT_DEFINED_ADDED :
                                             VIR_DOMAIN_EVENT_DEFINED_UPDATED);
            if (event)
                qemuDomainEventQueue(driver, event);
            event = NULL;
        }

        if (!(flags & VIR_MIGRATE_PAUSED)) {
            /* run 'cont' on the destination, which allows migration on qemu
             * >= 0.10.6 to work properly.  This isn't strictly necessary on
             * older qemu's, but it also doesn't hurt anything there
             */
J
Jiri Denemark 已提交
3009
            if (qemuProcessStartCPUs(driver, vm, dconn,
3010 3011
                                     VIR_DOMAIN_RUNNING_MIGRATED,
                                     QEMU_ASYNC_JOB_MIGRATION_IN) < 0) {
3012 3013 3014 3015 3016 3017 3018 3019
                if (virGetLastError() == NULL)
                    qemuReportError(VIR_ERR_INTERNAL_ERROR,
                                    "%s", _("resume operation failed"));
                /* Need to save the current error, in case shutting
                 * down the process overwrites it
                 */
                orig_err = virSaveLastError();

3020 3021 3022 3023 3024 3025 3026 3027 3028
                /*
                 * In v3 protocol, the source VM is still available to
                 * restart during confirm() step, so we kill it off
                 * now.
                 * In v2 protocol, the source is dead, so we leave
                 * target in paused state, in case admin can fix
                 * things up
                 */
                if (v3proto) {
3029 3030
                    qemuProcessStop(driver, vm, VIR_DOMAIN_SHUTOFF_FAILED,
                                    VIR_QEMU_PROCESS_STOP_MIGRATED);
3031
                    virDomainAuditStop(vm, "failed");
3032 3033 3034 3035
                    event = virDomainEventNewFromObj(vm,
                                                     VIR_DOMAIN_EVENT_STOPPED,
                                                     VIR_DOMAIN_EVENT_STOPPED_FAILED);
                }
3036 3037 3038 3039
                goto endjob;
            }
        }

3040 3041
        dom = virGetDomain (dconn, vm->def->name, vm->def->uuid);

3042 3043 3044
        event = virDomainEventNewFromObj(vm,
                                         VIR_DOMAIN_EVENT_RESUMED,
                                         VIR_DOMAIN_EVENT_RESUMED_MIGRATED);
J
Jiri Denemark 已提交
3045 3046
        if (virDomainObjGetState(vm, NULL) == VIR_DOMAIN_PAUSED) {
            virDomainObjSetState(vm, VIR_DOMAIN_PAUSED, VIR_DOMAIN_PAUSED_USER);
3047 3048
            if (event)
                qemuDomainEventQueue(driver, event);
3049 3050 3051 3052 3053 3054 3055 3056
            event = virDomainEventNewFromObj(vm,
                                             VIR_DOMAIN_EVENT_SUSPENDED,
                                             VIR_DOMAIN_EVENT_SUSPENDED_PAUSED);
        }
        if (virDomainSaveStatus(driver->caps, driver->stateDir, vm) < 0) {
            VIR_WARN("Failed to save status on vm %s", vm->def->name);
            goto endjob;
        }
3057 3058 3059

        /* Guest is successfully running, so cancel previous auto destroy */
        qemuProcessAutoDestroyRemove(driver, vm);
3060
    } else {
3061 3062
        qemuProcessStop(driver, vm, VIR_DOMAIN_SHUTOFF_FAILED,
                        VIR_QEMU_PROCESS_STOP_MIGRATED);
3063
        virDomainAuditStop(vm, "failed");
3064 3065 3066 3067 3068
        event = virDomainEventNewFromObj(vm,
                                         VIR_DOMAIN_EVENT_STOPPED,
                                         VIR_DOMAIN_EVENT_STOPPED_FAILED);
    }

3069 3070 3071
    if (qemuMigrationBakeCookie(mig, driver, vm, cookieout, cookieoutlen, 0) < 0)
        VIR_WARN("Unable to encode migration cookie");

3072
endjob:
E
Eric Blake 已提交
3073 3074 3075
    if (qemuMigrationJobFinish(driver, vm) == 0) {
        vm = NULL;
    } else if (!vm->persistent && !virDomainObjIsActive(vm)) {
3076
        qemuDomainRemoveInactive(driver, vm);
E
Eric Blake 已提交
3077
        vm = NULL;
3078
    }
3079 3080

cleanup:
J
Jiri Denemark 已提交
3081 3082
    if (vm) {
        VIR_FREE(priv->origname);
3083
        virDomainObjUnlock(vm);
J
Jiri Denemark 已提交
3084
    }
3085 3086
    if (event)
        qemuDomainEventQueue(driver, event);
3087
    qemuMigrationCookieFree(mig);
3088 3089 3090 3091
    if (orig_err) {
        virSetError(orig_err);
        virFreeError(orig_err);
    }
3092 3093
    return dom;
}
3094

3095 3096 3097 3098 3099 3100

int qemuMigrationConfirm(struct qemud_driver *driver,
                         virConnectPtr conn,
                         virDomainObjPtr vm,
                         const char *cookiein,
                         int cookieinlen,
E
Eric Blake 已提交
3101
                         unsigned int flags,
3102
                         int retcode)
3103 3104 3105 3106
{
    qemuMigrationCookiePtr mig;
    virDomainEventPtr event = NULL;
    int rv = -1;
3107
    VIR_DEBUG("driver=%p, conn=%p, vm=%p, cookiein=%s, cookieinlen=%d, "
3108
              "flags=%x, retcode=%d",
3109 3110
              driver, conn, vm, NULLSTR(cookiein), cookieinlen,
              flags, retcode);
3111

3112
    virCheckFlags(QEMU_MIGRATION_FLAGS, -1);
E
Eric Blake 已提交
3113

3114 3115 3116 3117 3118
    qemuMigrationJobSetPhase(driver, vm,
                             retcode == 0
                             ? QEMU_MIGRATION_PHASE_CONFIRM3
                             : QEMU_MIGRATION_PHASE_CONFIRM3_CANCELLED);

3119
    if (!(mig = qemuMigrationEatCookie(driver, vm, cookiein, cookieinlen, 0)))
3120 3121 3122 3123 3124 3125
        return -1;

    /* Did the migration go as planned?  If yes, kill off the
     * domain object, but if no, resume CPUs
     */
    if (retcode == 0) {
3126 3127
        qemuProcessStop(driver, vm, VIR_DOMAIN_SHUTOFF_MIGRATED,
                        VIR_QEMU_PROCESS_STOP_MIGRATED);
3128
        virDomainAuditStop(vm, "migrated");
3129 3130 3131 3132 3133 3134 3135 3136 3137 3138 3139

        event = virDomainEventNewFromObj(vm,
                                         VIR_DOMAIN_EVENT_STOPPED,
                                         VIR_DOMAIN_EVENT_STOPPED_MIGRATED);
    } else {

        /* run 'cont' on the destination, which allows migration on qemu
         * >= 0.10.6 to work properly.  This isn't strictly necessary on
         * older qemu's, but it also doesn't hurt anything there
         */
        if (qemuProcessStartCPUs(driver, vm, conn,
3140 3141
                                 VIR_DOMAIN_RUNNING_MIGRATED,
                                 QEMU_ASYNC_JOB_MIGRATION_OUT) < 0) {
3142 3143 3144
            if (virGetLastError() == NULL)
                qemuReportError(VIR_ERR_INTERNAL_ERROR,
                                "%s", _("resume operation failed"));
3145
            goto cleanup;
3146 3147 3148 3149 3150 3151 3152
        }

        event = virDomainEventNewFromObj(vm,
                                         VIR_DOMAIN_EVENT_RESUMED,
                                         VIR_DOMAIN_EVENT_RESUMED_MIGRATED);
        if (virDomainSaveStatus(driver->caps, driver->stateDir, vm) < 0) {
            VIR_WARN("Failed to save status on vm %s", vm->def->name);
3153
            goto cleanup;
3154 3155 3156 3157 3158 3159 3160 3161 3162 3163 3164 3165 3166
        }
    }

    qemuMigrationCookieFree(mig);
    rv = 0;

cleanup:
    if (event)
        qemuDomainEventQueue(driver, event);
    return rv;
}


3167 3168 3169 3170 3171
/* Helper function called while driver lock is held and vm is active.  */
int
qemuMigrationToFile(struct qemud_driver *driver, virDomainObjPtr vm,
                    int fd, off_t offset, const char *path,
                    const char *compressor,
E
Eric Blake 已提交
3172
                    bool bypassSecurityDriver,
3173
                    enum qemuDomainAsyncJob asyncJob)
3174 3175 3176 3177 3178 3179
{
    qemuDomainObjPrivatePtr priv = vm->privateData;
    virCgroupPtr cgroup = NULL;
    int ret = -1;
    int rc;
    bool restoreLabel = false;
3180 3181
    virCommandPtr cmd = NULL;
    int pipeFD[2] = { -1, -1 };
3182 3183 3184 3185 3186 3187 3188 3189 3190 3191
    unsigned long saveMigBandwidth = priv->migMaxBandwidth;

    /* Increase migration bandwidth to unlimited since target is a file.
     * Failure to change migration speed is not fatal. */
    if (qemuDomainObjEnterMonitorAsync(driver, vm, asyncJob) == 0) {
        qemuMonitorSetMigrationSpeed(priv->mon,
                                     QEMU_DOMAIN_FILE_MIG_BANDWIDTH_MAX);
        priv->migMaxBandwidth = QEMU_DOMAIN_FILE_MIG_BANDWIDTH_MAX;
        qemuDomainObjExitMonitorWithDriver(driver, vm);
    }
3192

3193
    if (qemuCapsGet(priv->qemuCaps, QEMU_CAPS_MIGRATE_QEMU_FD) &&
3194
        (!compressor || pipe(pipeFD) == 0)) {
3195
        /* All right! We can use fd migration, which means that qemu
3196 3197 3198
         * doesn't have to open() the file, so while we still have to
         * grant SELinux access, we can do it on fd and avoid cleanup
         * later, as well as skip futzing with cgroup.  */
3199
        if (virSecurityManagerSetImageFDLabel(driver->securityManager, vm->def,
3200
                                              compressor ? pipeFD[1] : fd) < 0)
3201
            goto cleanup;
3202 3203 3204
        bypassSecurityDriver = true;
    } else {
        /* Phooey - we have to fall back on exec migration, where qemu
E
Eric Blake 已提交
3205 3206
         * has to popen() the file by name, and block devices have to be
         * given cgroup ACL permission.  We might also stumble on
3207 3208
         * a race present in some qemu versions where it does a wait()
         * that botches pclose.  */
E
Eric Blake 已提交
3209
        if (qemuCgroupControllerActive(driver,
3210 3211 3212 3213 3214 3215 3216 3217 3218 3219
                                       VIR_CGROUP_CONTROLLER_DEVICES)) {
            if (virCgroupForDomain(driver->cgroup, vm->def->name,
                                   &cgroup, 0) != 0) {
                qemuReportError(VIR_ERR_INTERNAL_ERROR,
                                _("Unable to find cgroup for %s"),
                                vm->def->name);
                goto cleanup;
            }
            rc = virCgroupAllowDevicePath(cgroup, path,
                                          VIR_CGROUP_DEVICE_RW);
3220
            virDomainAuditCgroupPath(vm, cgroup, "allow", path, "rw", rc);
E
Eric Blake 已提交
3221 3222 3223 3224
            if (rc == 1) {
                /* path was not a device, no further need for cgroup */
                virCgroupFree(&cgroup);
            } else if (rc < 0) {
3225 3226 3227 3228 3229
                virReportSystemError(-rc,
                                     _("Unable to allow device %s for %s"),
                                     path, vm->def->name);
                goto cleanup;
            }
3230
        }
3231 3232
        if ((!bypassSecurityDriver) &&
            virSecurityManagerSetSavedStateLabel(driver->securityManager,
3233
                                                 vm->def, path) < 0)
3234
            goto cleanup;
3235
        restoreLabel = true;
3236 3237
    }

3238
    if (qemuDomainObjEnterMonitorAsync(driver, vm, asyncJob) < 0)
3239 3240
        goto cleanup;

3241 3242 3243
    if (!compressor) {
        const char *args[] = { "cat", NULL };

3244
        if (qemuCapsGet(priv->qemuCaps, QEMU_CAPS_MIGRATE_QEMU_FD) &&
3245 3246 3247 3248 3249 3250 3251 3252 3253 3254 3255 3256 3257 3258 3259 3260
            priv->monConfig->type == VIR_DOMAIN_CHR_TYPE_UNIX) {
            rc = qemuMonitorMigrateToFd(priv->mon,
                                        QEMU_MONITOR_MIGRATE_BACKGROUND,
                                        fd);
        } else {
            rc = qemuMonitorMigrateToFile(priv->mon,
                                          QEMU_MONITOR_MIGRATE_BACKGROUND,
                                          args, path, offset);
        }
    } else {
        const char *prog = compressor;
        const char *args[] = {
            prog,
            "-c",
            NULL
        };
3261 3262 3263 3264 3265 3266 3267 3268 3269 3270 3271 3272 3273 3274 3275 3276 3277 3278 3279
        if (pipeFD[0] != -1) {
            cmd = virCommandNewArgs(args);
            virCommandSetInputFD(cmd, pipeFD[0]);
            virCommandSetOutputFD(cmd, &fd);
            if (virSetCloseExec(pipeFD[1]) < 0) {
                virReportSystemError(errno, "%s",
                                     _("Unable to set cloexec flag"));
                qemuDomainObjExitMonitorWithDriver(driver, vm);
                goto cleanup;
            }
            if (virCommandRunAsync(cmd, NULL) < 0) {
                qemuDomainObjExitMonitorWithDriver(driver, vm);
                goto cleanup;
            }
            rc = qemuMonitorMigrateToFd(priv->mon,
                                        QEMU_MONITOR_MIGRATE_BACKGROUND,
                                        pipeFD[1]);
            if (VIR_CLOSE(pipeFD[0]) < 0 ||
                VIR_CLOSE(pipeFD[1]) < 0)
3280
                VIR_WARN("failed to close intermediate pipe");
3281 3282 3283 3284 3285
        } else {
            rc = qemuMonitorMigrateToFile(priv->mon,
                                          QEMU_MONITOR_MIGRATE_BACKGROUND,
                                          args, path, offset);
        }
3286
    }
3287
    qemuDomainObjExitMonitorWithDriver(driver, vm);
3288 3289 3290 3291

    if (rc < 0)
        goto cleanup;

3292
    rc = qemuMigrationWaitForCompletion(driver, vm, asyncJob, NULL);
3293 3294 3295 3296

    if (rc < 0)
        goto cleanup;

3297 3298 3299
    if (cmd && virCommandWait(cmd, NULL) < 0)
        goto cleanup;

3300 3301 3302
    ret = 0;

cleanup:
3303 3304 3305 3306 3307 3308 3309
    /* Restore max migration bandwidth */
    if (qemuDomainObjEnterMonitorAsync(driver, vm, asyncJob) == 0) {
        qemuMonitorSetMigrationSpeed(priv->mon, saveMigBandwidth);
        priv->migMaxBandwidth = saveMigBandwidth;
        qemuDomainObjExitMonitorWithDriver(driver, vm);
    }

3310 3311 3312
    VIR_FORCE_CLOSE(pipeFD[0]);
    VIR_FORCE_CLOSE(pipeFD[1]);
    virCommandFree(cmd);
3313 3314
    if (restoreLabel && (!bypassSecurityDriver) &&
        virSecurityManagerRestoreSavedStateLabel(driver->securityManager,
3315
                                                 vm->def, path) < 0)
3316 3317 3318 3319 3320
        VIR_WARN("failed to restore save state label on %s", path);

    if (cgroup != NULL) {
        rc = virCgroupDenyDevicePath(cgroup, path,
                                     VIR_CGROUP_DEVICE_RWM);
3321
        virDomainAuditCgroupPath(vm, cgroup, "deny", path, "rwm", rc);
3322 3323 3324 3325 3326 3327 3328
        if (rc < 0)
            VIR_WARN("Unable to deny device %s for %s %d",
                     path, vm->def->name, rc);
        virCgroupFree(&cgroup);
    }
    return ret;
}
3329 3330 3331 3332 3333 3334 3335 3336 3337 3338 3339

int
qemuMigrationJobStart(struct qemud_driver *driver,
                      virDomainObjPtr vm,
                      enum qemuDomainAsyncJob job)
{
    qemuDomainObjPrivatePtr priv = vm->privateData;

    if (qemuDomainObjBeginAsyncJobWithDriver(driver, vm, job) < 0)
        return -1;

3340
    if (job == QEMU_ASYNC_JOB_MIGRATION_IN) {
3341
        qemuDomainObjSetAsyncJobMask(vm, QEMU_JOB_NONE);
3342 3343
    } else {
        qemuDomainObjSetAsyncJobMask(vm, DEFAULT_JOB_MASK |
3344
                                     JOB_MASK(QEMU_JOB_SUSPEND) |
3345 3346
                                     JOB_MASK(QEMU_JOB_MIGRATION_OP));
    }
3347 3348 3349 3350 3351 3352 3353 3354 3355 3356 3357 3358 3359 3360 3361 3362 3363 3364 3365 3366 3367 3368 3369 3370 3371 3372 3373 3374 3375 3376 3377 3378 3379 3380 3381

    priv->job.info.type = VIR_DOMAIN_JOB_UNBOUNDED;

    return 0;
}

void
qemuMigrationJobSetPhase(struct qemud_driver *driver,
                         virDomainObjPtr vm,
                         enum qemuMigrationJobPhase phase)
{
    qemuDomainObjPrivatePtr priv = vm->privateData;

    if (phase < priv->job.phase) {
        VIR_ERROR(_("migration protocol going backwards %s => %s"),
                  qemuMigrationJobPhaseTypeToString(priv->job.phase),
                  qemuMigrationJobPhaseTypeToString(phase));
        return;
    }

    qemuDomainObjSetJobPhase(driver, vm, phase);
}

void
qemuMigrationJobStartPhase(struct qemud_driver *driver,
                           virDomainObjPtr vm,
                           enum qemuMigrationJobPhase phase)
{
    virDomainObjRef(vm);
    qemuMigrationJobSetPhase(driver, vm, phase);
}

int
qemuMigrationJobContinue(virDomainObjPtr vm)
{
3382
    qemuDomainObjReleaseAsyncJob(vm);
3383 3384 3385 3386 3387 3388 3389 3390 3391 3392 3393 3394 3395 3396 3397 3398 3399 3400 3401 3402 3403 3404 3405 3406 3407 3408 3409 3410
    return virDomainObjUnref(vm);
}

bool
qemuMigrationJobIsActive(virDomainObjPtr vm,
                         enum qemuDomainAsyncJob job)
{
    qemuDomainObjPrivatePtr priv = vm->privateData;

    if (priv->job.asyncJob != job) {
        const char *msg;

        if (job == QEMU_ASYNC_JOB_MIGRATION_IN)
            msg = _("domain '%s' is not processing incoming migration");
        else
            msg = _("domain '%s' is not being migrated");

        qemuReportError(VIR_ERR_OPERATION_INVALID, msg, vm->def->name);
        return false;
    }
    return true;
}

int
qemuMigrationJobFinish(struct qemud_driver *driver, virDomainObjPtr vm)
{
    return qemuDomainObjEndAsyncJob(driver, vm);
}