qemu_migration.c 95.1 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
/*
 * qemu_migration.c: QEMU migration handling
 *
 * Copyright (C) 2006-2011 Red Hat, Inc.
 *
 * 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 29 30 31 32 33

#include "qemu_migration.h"
#include "qemu_monitor.h"
#include "qemu_domain.h"
#include "qemu_process.h"
#include "qemu_capabilities.h"
34
#include "qemu_cgroup.h"
35

36
#include "domain_audit.h"
37 38 39 40
#include "logging.h"
#include "virterror_internal.h"
#include "memory.h"
#include "util.h"
E
Eric Blake 已提交
41
#include "virfile.h"
42 43
#include "datatypes.h"
#include "fdstream.h"
44
#include "uuid.h"
45
#include "locking/domain_lock.h"
46
#include "rpc/virnetsocket.h"
47

48 49 50

#define VIR_FROM_THIS VIR_FROM_QEMU

51 52 53 54 55 56 57 58 59 60 61 62 63
VIR_ENUM_IMPL(qemuMigrationJobPhase, QEMU_MIGRATION_PHASE_LAST,
              "none",
              "perform2",
              "begin3",
              "perform3",
              "perform3_done",
              "confirm3_cancelled",
              "confirm3",
              "prepare",
              "finish2",
              "finish3",
);

64
enum qemuMigrationCookieFlags {
65
    QEMU_MIGRATION_COOKIE_FLAG_GRAPHICS,
66
    QEMU_MIGRATION_COOKIE_FLAG_LOCKSTATE,
67 68 69 70 71 72 73

    QEMU_MIGRATION_COOKIE_FLAG_LAST
};

VIR_ENUM_DECL(qemuMigrationCookieFlag);
VIR_ENUM_IMPL(qemuMigrationCookieFlag,
              QEMU_MIGRATION_COOKIE_FLAG_LAST,
74
              "graphics", "lockstate");
75 76 77

enum qemuMigrationCookieFeatures {
    QEMU_MIGRATION_COOKIE_GRAPHICS  = (1 << QEMU_MIGRATION_COOKIE_FLAG_GRAPHICS),
78
    QEMU_MIGRATION_COOKIE_LOCKSTATE = (1 << QEMU_MIGRATION_COOKIE_FLAG_LOCKSTATE),
79 80 81 82 83 84 85 86 87 88 89 90
};

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

91 92 93
typedef struct _qemuMigrationCookie qemuMigrationCookie;
typedef qemuMigrationCookie *qemuMigrationCookiePtr;
struct _qemuMigrationCookie {
E
Eric Blake 已提交
94 95
    unsigned int flags;
    unsigned int flagsMandatory;
96 97

    /* Host properties */
98 99 100 101
    unsigned char localHostuuid[VIR_UUID_BUFLEN];
    unsigned char remoteHostuuid[VIR_UUID_BUFLEN];
    char *localHostname;
    char *remoteHostname;
102 103 104 105

    /* Guest properties */
    unsigned char uuid[VIR_UUID_BUFLEN];
    char *name;
106

107 108 109 110
    /* If (flags & QEMU_MIGRATION_COOKIE_LOCKSTATE) */
    char *lockState;
    char *lockDriver;

111 112
    /* If (flags & QEMU_MIGRATION_COOKIE_GRAPHICS) */
    qemuMigrationCookieGraphicsPtr graphics;
113 114
};

115 116 117 118 119 120 121 122 123
static void qemuMigrationCookieGraphicsFree(qemuMigrationCookieGraphicsPtr grap)
{
    if (!grap)
        return;
    VIR_FREE(grap->listen);
    VIR_FREE(grap->tlsSubject);
    VIR_FREE(grap);
}

124 125 126 127 128 129

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

130 131 132
    if (mig->flags & QEMU_MIGRATION_COOKIE_GRAPHICS)
        qemuMigrationCookieGraphicsFree(mig->graphics);

133 134
    VIR_FREE(mig->localHostname);
    VIR_FREE(mig->remoteHostname);
135
    VIR_FREE(mig->name);
136 137
    VIR_FREE(mig->lockState);
    VIR_FREE(mig->lockDriver);
138 139 140 141
    VIR_FREE(mig);
}


142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214
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;
215
        listenAddr = virDomainGraphicsListenGetAddress(def, 0);
216 217 218 219 220 221 222 223 224 225 226 227
        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;
228
        listenAddr = virDomainGraphicsListenGetAddress(def, 0);
229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248
        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;
}


249 250 251 252 253 254 255 256 257 258 259 260
static qemuMigrationCookiePtr
qemuMigrationCookieNew(virDomainObjPtr dom)
{
    qemuMigrationCookiePtr mig = NULL;

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

    if (!(mig->name = strdup(dom->def->name)))
        goto no_memory;
    memcpy(mig->uuid, dom->def->uuid, VIR_UUID_BUFLEN);

261
    if (!(mig->localHostname = virGetHostname(NULL)))
262
        goto error;
263
    if (virGetHostUUID(mig->localHostuuid) < 0) {
264 265 266 267 268 269 270 271 272 273 274 275 276 277 278
        qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                        _("Unable to obtain host UUID"));
        goto error;
    }

    return mig;

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


279 280 281 282 283 284 285 286 287 288 289 290 291
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 ||
292 293 294 295 296 297
         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;
    }
298 299 300 301 302

    return 0;
}


303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337
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;
}



338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355
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");
    }
}


356 357 358 359 360
static void qemuMigrationCookieXMLFormat(virBufferPtr buf,
                                         qemuMigrationCookiePtr mig)
{
    char uuidstr[VIR_UUID_STRING_BUFLEN];
    char hostuuidstr[VIR_UUID_STRING_BUFLEN];
361
    int i;
362 363

    virUUIDFormat(mig->uuid, uuidstr);
364
    virUUIDFormat(mig->localHostuuid, hostuuidstr);
365 366 367 368

    virBufferAsprintf(buf, "<qemu-migration>\n");
    virBufferEscapeString(buf, "  <name>%s</name>\n", mig->name);
    virBufferAsprintf(buf, "  <uuid>%s</uuid>\n", uuidstr);
369
    virBufferEscapeString(buf, "  <hostname>%s</hostname>\n", mig->localHostname);
370
    virBufferAsprintf(buf, "  <hostuuid>%s</hostuuid>\n", hostuuidstr);
371

372 373 374 375 376 377
    for (i = 0 ; i < QEMU_MIGRATION_COOKIE_FLAG_LAST ; i++) {
        if (mig->flagsMandatory & (1 << i))
            virBufferAsprintf(buf, "  <feature name='%s'/>\n",
                              qemuMigrationCookieFlagTypeToString(i));
    }

378 379
    if ((mig->flags & QEMU_MIGRATION_COOKIE_GRAPHICS) &&
        mig->graphics)
380 381
        qemuMigrationCookieGraphicsXMLFormat(buf, mig->graphics);

382 383 384 385 386 387 388 389 390
    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");
    }

391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409
    virBufferAddLit(buf, "</qemu-migration>\n");
}


static char *qemuMigrationCookieXMLFormatStr(qemuMigrationCookiePtr mig)
{
    virBuffer buf = VIR_BUFFER_INITIALIZER;

    qemuMigrationCookieXMLFormat(&buf, mig);

    if (virBufferError(&buf)) {
        virReportOOMError();
        return NULL;
    }

    return virBufferContentAndReset(&buf);
}


410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447
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;
    }
    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 */
448
    grap->tlsSubject = virXPathString("string(./graphics/cert[@info='subject']/@value)", ctxt);
449 450 451 452 453 454 455 456 457 458 459 460


    return grap;

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


461 462 463
static int
qemuMigrationCookieXMLParse(qemuMigrationCookiePtr mig,
                            xmlXPathContextPtr ctxt,
E
Eric Blake 已提交
464
                            unsigned int flags)
465 466 467
{
    char uuidstr[VIR_UUID_STRING_BUFLEN];
    char *tmp;
468 469
    xmlNodePtr *nodes = NULL;
    int i, n;
470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505

    /* 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 */
506
    if (!(mig->remoteHostname = virXPathString("string(./hostname[1])", ctxt))) {
507 508 509 510
        qemuReportError(VIR_ERR_INTERNAL_ERROR,
                        "%s", _("missing hostname element in migration data"));
        goto error;
    }
511
    if (STREQ(mig->remoteHostname, mig->localHostname)) {
512 513
        qemuReportError(VIR_ERR_INTERNAL_ERROR,
                        _("Attempt to migrate guest to the same host %s"),
514
                        mig->remoteHostname);
515 516 517 518 519 520 521 522
        goto error;
    }

    if (!(tmp = virXPathString("string(./hostuuid[1])", ctxt))) {
        qemuReportError(VIR_ERR_INTERNAL_ERROR,
                        "%s", _("missing hostuuid element in migration data"));
        goto error;
    }
523 524 525 526 527 528
    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) {
529 530 531 532 533 534 535
        qemuReportError(VIR_ERR_INTERNAL_ERROR,
                        _("Attempt to migrate guest to the same host %s"),
                        tmp);
        goto error;
    }
    VIR_FREE(tmp);

536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567
    /* 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);

568 569 570 571 572
    if ((flags & QEMU_MIGRATION_COOKIE_GRAPHICS) &&
        virXPathBoolean("count(./graphics) > 0", ctxt) &&
        (!(mig->graphics = qemuMigrationCookieGraphicsXMLParse(ctxt))))
        goto error;

573 574 575 576 577 578 579 580 581 582 583 584 585
    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);
    }

586 587 588 589
    return 0;

error:
    VIR_FREE(tmp);
590
    VIR_FREE(nodes);
591 592 593 594 595 596 597
    return -1;
}


static int
qemuMigrationCookieXMLParseStr(qemuMigrationCookiePtr mig,
                               const char *xml,
E
Eric Blake 已提交
598
                               unsigned int flags)
599 600 601
{
    xmlDocPtr doc = NULL;
    xmlXPathContextPtr ctxt = NULL;
602
    int ret = -1;
603 604 605

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

606
    if (!(doc = virXMLParseStringCtxt(xml, "qemumigration.xml", &ctxt)))
607 608 609 610 611 612 613 614 615 616 617 618 619 620
        goto cleanup;

    ret = qemuMigrationCookieXMLParse(mig, ctxt, flags);

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

    return ret;
}


static int
qemuMigrationBakeCookie(qemuMigrationCookiePtr mig,
621 622
                        struct qemud_driver *driver,
                        virDomainObjPtr dom,
623 624
                        char **cookieout,
                        int *cookieoutlen,
E
Eric Blake 已提交
625
                        unsigned int flags)
626
{
627 628
    if (!cookieout || !cookieoutlen)
        return 0;
629 630 631

    *cookieoutlen = 0;

632 633 634 635
    if (flags & QEMU_MIGRATION_COOKIE_GRAPHICS &&
        qemuMigrationCookieAddGraphics(mig, driver, dom) < 0)
        return -1;

636 637 638 639
    if (flags & QEMU_MIGRATION_COOKIE_LOCKSTATE &&
        qemuMigrationCookieAddLockstate(mig, driver, dom) < 0)
        return -1;

640 641 642 643 644 645 646 647 648 649 650 651
    if (!(*cookieout = qemuMigrationCookieXMLFormatStr(mig)))
        return -1;

    *cookieoutlen = strlen(*cookieout) + 1;

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

    return 0;
}


static qemuMigrationCookiePtr
652 653
qemuMigrationEatCookie(struct qemud_driver *driver,
                       virDomainObjPtr dom,
654 655
                       const char *cookiein,
                       int cookieinlen,
E
Eric Blake 已提交
656
                       unsigned int flags)
657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678
{
    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,
                                       cookiein,
                                       flags) < 0)
        goto error;

679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696
    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;
        }
    }

697 698 699 700 701 702
    return mig;

error:
    qemuMigrationCookieFree(mig);
    return NULL;
}
703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723

bool
qemuMigrationIsAllowed(virDomainDefPtr def)
{
    if (def->nhostdevs > 0) {
        qemuReportError(VIR_ERR_OPERATION_INVALID,
            "%s", _("Domain with assigned host devices cannot be migrated"));
        return false;
    }

    return true;
}

/** qemuMigrationSetOffline
 * Pause domain for non-live migration.
 */
int
qemuMigrationSetOffline(struct qemud_driver *driver,
                        virDomainObjPtr vm)
{
    int ret;
724
    VIR_DEBUG("driver=%p vm=%p", driver, vm);
725 726
    ret = qemuProcessStopCPUs(driver, vm, VIR_DOMAIN_PAUSED_MIGRATION,
                              QEMU_ASYNC_JOB_MIGRATION_OUT);
727 728 729 730 731 732 733 734 735 736 737 738 739 740
    if (ret == 0) {
        virDomainEventPtr event;

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

    return ret;
}


741 742 743
static int
qemuMigrationUpdateJobStatus(struct qemud_driver *driver,
                             virDomainObjPtr vm,
744 745
                             const char *job,
                             enum qemuDomainAsyncJob asyncJob)
746
{
747
    qemuDomainObjPrivatePtr priv = vm->privateData;
748 749 750 751 752
    int ret = -1;
    int status;
    unsigned long long memProcessed;
    unsigned long long memRemaining;
    unsigned long long memTotal;
753

754 755 756
    ret = qemuDomainObjEnterMonitorAsync(driver, vm, asyncJob);
    if (ret < 0) {
        /* Guest already exited; nothing further to update.  */
757 758
        return -1;
    }
759 760 761 762 763 764
    ret = qemuMonitorGetMigrationStatus(priv->mon,
                                        &status,
                                        &memProcessed,
                                        &memRemaining,
                                        &memTotal);
    qemuDomainObjExitMonitorWithDriver(driver, vm);
765

766 767
    if (ret < 0 || virTimeMs(&priv->job.info.timeElapsed) < 0) {
        priv->job.info.type = VIR_DOMAIN_JOB_FAILED;
768 769
        return -1;
    }
770
    priv->job.info.timeElapsed -= priv->job.start;
771 772 773

    switch (status) {
    case QEMU_MONITOR_MIGRATION_STATUS_INACTIVE:
774
        priv->job.info.type = VIR_DOMAIN_JOB_NONE;
775 776 777 778 779
        qemuReportError(VIR_ERR_OPERATION_FAILED,
                        _("%s: %s"), job, _("is not active"));
        break;

    case QEMU_MONITOR_MIGRATION_STATUS_ACTIVE:
780 781 782
        priv->job.info.dataTotal = memTotal;
        priv->job.info.dataRemaining = memRemaining;
        priv->job.info.dataProcessed = memProcessed;
783

784 785 786
        priv->job.info.memTotal = memTotal;
        priv->job.info.memRemaining = memRemaining;
        priv->job.info.memProcessed = memProcessed;
787 788 789 790 791

        ret = 0;
        break;

    case QEMU_MONITOR_MIGRATION_STATUS_COMPLETED:
792
        priv->job.info.type = VIR_DOMAIN_JOB_COMPLETED;
793 794 795 796
        ret = 0;
        break;

    case QEMU_MONITOR_MIGRATION_STATUS_ERROR:
797
        priv->job.info.type = VIR_DOMAIN_JOB_FAILED;
798 799 800 801 802
        qemuReportError(VIR_ERR_OPERATION_FAILED,
                        _("%s: %s"), job, _("unexpectedly failed"));
        break;

    case QEMU_MONITOR_MIGRATION_STATUS_CANCELLED:
803
        priv->job.info.type = VIR_DOMAIN_JOB_CANCELLED;
804 805 806 807 808 809 810 811 812
        qemuReportError(VIR_ERR_OPERATION_FAILED,
                        _("%s: %s"), job, _("canceled by client"));
        break;
    }

    return ret;
}


813 814 815
static int
qemuMigrationWaitForCompletion(struct qemud_driver *driver, virDomainObjPtr vm,
                               enum qemuDomainAsyncJob asyncJob)
816
{
817
    qemuDomainObjPrivatePtr priv = vm->privateData;
818 819
    const char *job;

820 821
    switch (priv->job.asyncJob) {
    case QEMU_ASYNC_JOB_MIGRATION_OUT:
822 823
        job = _("migration job");
        break;
824
    case QEMU_ASYNC_JOB_SAVE:
825 826
        job = _("domain save job");
        break;
827
    case QEMU_ASYNC_JOB_DUMP:
828 829 830 831 832
        job = _("domain core dump job");
        break;
    default:
        job = _("job");
    }
833

834
    priv->job.info.type = VIR_DOMAIN_JOB_UNBOUNDED;
835

836
    while (priv->job.info.type == VIR_DOMAIN_JOB_UNBOUNDED) {
837 838 839
        /* Poll every 50ms for progress & to allow cancellation */
        struct timespec ts = { .tv_sec = 0, .tv_nsec = 50 * 1000 * 1000ull };

840
        if (qemuMigrationUpdateJobStatus(driver, vm, job, asyncJob) < 0)
841 842 843 844 845 846 847 848 849 850 851 852
            goto cleanup;

        virDomainObjUnlock(vm);
        qemuDriverUnlock(driver);

        nanosleep(&ts, NULL);

        qemuDriverLock(driver);
        virDomainObjLock(vm);
    }

cleanup:
853
    if (priv->job.info.type == VIR_DOMAIN_JOB_COMPLETED)
854 855 856
        return 0;
    else
        return -1;
857 858 859
}


860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879
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;

880 881
    ret = qemuDomainObjEnterMonitorAsync(driver, vm,
                                         QEMU_ASYNC_JOB_MIGRATION_OUT);
882 883 884 885 886 887 888 889 890
    if (ret == 0) {
        ret = qemuMonitorGraphicsRelocate(priv->mon,
                                          cookie->graphics->type,
                                          cookie->remoteHostname,
                                          cookie->graphics->port,
                                          cookie->graphics->tlsPort,
                                          cookie->graphics->tlsSubject);
        qemuDomainObjExitMonitorWithDriver(driver, vm);
    }
891 892 893 894 895

    return ret;
}


896
/* The caller is supposed to lock the vm and start a migration job. */
897 898
char *qemuMigrationBegin(struct qemud_driver *driver,
                         virDomainObjPtr vm,
899
                         const char *xmlin,
900 901 902 903 904
                         char **cookieout,
                         int *cookieoutlen)
{
    char *rv = NULL;
    qemuMigrationCookiePtr mig = NULL;
905
    virDomainDefPtr def = NULL;
906 907
    qemuDomainObjPrivatePtr priv = vm->privateData;

908 909
    VIR_DEBUG("driver=%p, vm=%p, xmlin=%s, cookieout=%p, cookieoutlen=%p",
              driver, vm, NULLSTR(xmlin), cookieout, cookieoutlen);
910

911 912 913 914 915 916
    /* 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);
917

918 919 920 921 922 923
    if (qemuProcessAutoDestroyActive(driver, vm)) {
        qemuReportError(VIR_ERR_OPERATION_INVALID,
                        "%s", _("domain is marked for auto destroy"));
        goto cleanup;
    }

924 925 926
    if (!qemuMigrationIsAllowed(vm->def))
        goto cleanup;

927
    if (!(mig = qemuMigrationEatCookie(driver, vm, NULL, 0, 0)))
928 929 930 931
        goto cleanup;

    if (qemuMigrationBakeCookie(mig, driver, vm,
                                cookieout, cookieoutlen,
932
                                QEMU_MIGRATION_COOKIE_LOCKSTATE) < 0)
933 934
        goto cleanup;

935 936
    if (xmlin) {
        if (!(def = virDomainDefParseString(driver->caps, xmlin,
M
Matthias Bolte 已提交
937
                                            QEMU_EXPECTED_VIRT_TYPES,
938 939 940
                                            VIR_DOMAIN_XML_INACTIVE)))
            goto cleanup;

941
        if (!virDomainDefCheckABIStability(vm->def, def))
942 943 944 945 946 947 948 949 950 951
            goto cleanup;

        rv = qemuDomainDefFormatXML(driver, def,
                                    VIR_DOMAIN_XML_SECURE |
                                    VIR_DOMAIN_XML_UPDATE_CPU);
    } else {
        rv = qemuDomainFormatXML(driver, vm,
                                 VIR_DOMAIN_XML_SECURE |
                                 VIR_DOMAIN_XML_UPDATE_CPU);
    }
952 953 954

cleanup:
    qemuMigrationCookieFree(mig);
955
    virDomainDefFree(def);
956 957 958 959
    return rv;
}


960 961
/* Prepare is the first step, and it runs on the destination host.
 */
962 963 964 965 966 967 968 969 970 971 972 973

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)
974 975 976 977 978
{
    virDomainDefPtr def = NULL;
    virDomainObjPtr vm = NULL;
    virDomainEventPtr event = NULL;
    int ret = -1;
979
    int dataFD[2] = { -1, -1 };
980
    qemuDomainObjPrivatePtr priv = NULL;
J
Jiri Denemark 已提交
981
    unsigned long long now;
982
    qemuMigrationCookiePtr mig = NULL;
983
    bool tunnel = !!st;
984

J
Jiri Denemark 已提交
985
    if (virTimeMs(&now) < 0)
986 987 988
        return -1;

    if (!(def = virDomainDefParseString(driver->caps, dom_xml,
M
Matthias Bolte 已提交
989
                                        QEMU_EXPECTED_VIRT_TYPES,
990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015
                                        VIR_DOMAIN_XML_INACTIVE)))
        goto cleanup;

    if (!qemuMigrationIsAllowed(def))
        goto cleanup;

    /* Target domain name, maybe renamed. */
    if (dname) {
        VIR_FREE(def->name);
        def->name = strdup(dname);
        if (def->name == NULL)
            goto cleanup;
    }

    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;

1016 1017
    if (!(mig = qemuMigrationEatCookie(driver, vm, cookiein, cookieinlen,
                                       QEMU_MIGRATION_COOKIE_LOCKSTATE)))
1018 1019
        goto cleanup;

1020
    if (qemuMigrationJobStart(driver, vm, QEMU_ASYNC_JOB_MIGRATION_IN) < 0)
1021
        goto cleanup;
1022
    qemuMigrationJobSetPhase(driver, vm, QEMU_MIGRATION_PHASE_PREPARE);
1023 1024 1025 1026

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

1027 1028
    if (tunnel &&
        (pipe(dataFD) < 0 || virSetCloseExec(dataFD[1]) < 0)) {
1029 1030
        virReportSystemError(errno, "%s",
                             _("cannot create pipe for tunnelled migration"));
1031 1032 1033 1034
        goto endjob;
    }

    /* Start the QEMU daemon, with the same command-line arguments plus
1035
     * -incoming $migrateFrom
1036
     */
1037
    if (qemuProcessStart(dconn, driver, vm, migrateFrom, true,
1038
                         true, dataFD[0], NULL, NULL,
1039
                         VIR_VM_OP_MIGRATE_IN_START) < 0) {
1040
        virDomainAuditStart(vm, "migrated", false);
1041 1042 1043 1044 1045 1046
        /* Note that we don't set an error here because qemuProcessStart
         * should have already done that.
         */
        goto endjob;
    }

1047 1048 1049 1050 1051 1052 1053
    if (tunnel) {
        if (virFDStreamOpen(st, dataFD[1]) < 0) {
            virReportSystemError(errno, "%s",
                                 _("cannot pass pipe for tunnelled migration"));
            virDomainAuditStart(vm, "migrated", false);
            qemuProcessStop(driver, vm, 0, VIR_DOMAIN_SHUTOFF_FAILED);
            goto endjob;
1054
        }
1055
        dataFD[1] = -1; /* 'st' owns the FD now & will close it */
1056 1057
    }

1058 1059 1060 1061 1062 1063 1064 1065
    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");
    }
1066

1067 1068
    if (qemuMigrationBakeCookie(mig, driver, vm, cookieout, cookieoutlen,
                                QEMU_MIGRATION_COOKIE_GRAPHICS) < 0) {
1069 1070 1071 1072 1073 1074 1075
        /* 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");
    }

1076 1077 1078 1079
    virDomainAuditStart(vm, "migrated", true);
    event = virDomainEventNewFromObj(vm,
                                     VIR_DOMAIN_EVENT_STARTED,
                                     VIR_DOMAIN_EVENT_STARTED_MIGRATED);
1080

1081 1082 1083 1084 1085
    /* 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) {
1086
        vm = NULL;
1087 1088 1089
        qemuReportError(VIR_ERR_OPERATION_FAILED,
                        "%s", _("domain disappeared"));
        goto cleanup;
1090
    }
1091

1092
    ret = 0;
1093 1094 1095

cleanup:
    virDomainDefFree(def);
1096 1097
    VIR_FORCE_CLOSE(dataFD[0]);
    VIR_FORCE_CLOSE(dataFD[1]);
1098 1099 1100 1101
    if (vm)
        virDomainObjUnlock(vm);
    if (event)
        qemuDomainEventQueue(driver, event);
1102
    qemuMigrationCookieFree(mig);
1103
    return ret;
1104 1105 1106 1107 1108 1109 1110 1111 1112

endjob:
    if (qemuMigrationJobFinish(driver, vm) == 0) {
        vm = NULL;
    } else if (!vm->persistent) {
        virDomainRemoveInactive(&driver->domains, vm);
        vm = NULL;
    }
    goto cleanup;
1113 1114 1115
}


1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147
/*
 * 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;
}


1148 1149 1150
int
qemuMigrationPrepareDirect(struct qemud_driver *driver,
                           virConnectPtr dconn,
1151 1152 1153 1154
                           const char *cookiein,
                           int cookieinlen,
                           char **cookieout,
                           int *cookieoutlen,
1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165
                           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
1166

1167 1168 1169 1170 1171 1172
    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);
1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193

    /* 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",
1194 1195
                            _("hostname on destination resolved to localhost,"
                              " but migration requires an FQDN"));
1196 1197 1198 1199 1200 1201 1202 1203 1204
            goto cleanup;
        }

        /* XXX this really should have been a properly well-formed
         * URI, but we can't add in tcp:// now without breaking
         * compatability with old targets. We at least make the
         * new targets accept both syntaxes though.
         */
        /* Caller frees */
1205
        if (virAsprintf(uri_out, "tcp:%s:%d", hostname, this_port) < 0) {
1206 1207 1208 1209 1210 1211 1212 1213 1214
            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:")) {
1215 1216 1217
            qemuReportError(VIR_ERR_INVALID_ARG, "%s",
                            _("only tcp URIs are supported for KVM/QEMU"
                              " migrations"));
1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248
            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);

1249 1250
    /* QEMU will be started with -incoming tcp:0.0.0.0:port */
    snprintf(migrateFrom, sizeof (migrateFrom), "tcp:0.0.0.0:%d", this_port);
1251

1252 1253 1254
    ret = qemuMigrationPrepareAny(driver, dconn, cookiein, cookieinlen,
                                  cookieout, cookieoutlen, dname, dom_xml,
                                  migrateFrom, NULL);
1255 1256 1257 1258 1259 1260 1261 1262
cleanup:
    VIR_FREE(hostname);
    if (ret != 0)
        VIR_FREE(*uri_out);
    return ret;
}


1263 1264 1265
enum qemuMigrationDestinationType {
    MIGRATION_DEST_HOST,
    MIGRATION_DEST_UNIX,
1266
    MIGRATION_DEST_FD,
1267
};
1268

1269 1270 1271 1272
enum qemuMigrationForwardType {
    MIGRATION_FWD_DIRECT,
    MIGRATION_FWD_STREAM,
};
1273

1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284
typedef struct _qemuMigrationSpec qemuMigrationSpec;
typedef qemuMigrationSpec *qemuMigrationSpecPtr;
struct _qemuMigrationSpec {
    enum qemuMigrationDestinationType destType;
    union {
        struct {
            const char *name;
            int port;
        } host;

        struct {
1285
            char *file;
1286 1287
            int sock;
        } unix_socket;
1288 1289 1290 1291 1292

        struct {
            int qemu;
            int local;
        } fd;
1293 1294 1295 1296 1297 1298 1299
    } dest;

    enum qemuMigrationForwardType fwdType;
    union {
        virStreamPtr stream;
    } fwd;
};
1300 1301 1302

#define TUNNEL_SEND_BUF_SIZE 65536

1303 1304 1305 1306 1307 1308 1309 1310 1311 1312
typedef struct _qemuMigrationIOThread qemuMigrationIOThread;
typedef qemuMigrationIOThread *qemuMigrationIOThreadPtr;
struct _qemuMigrationIOThread {
    virThread thread;
    virStreamPtr st;
    int sock;
    virError err;
};

static void qemuMigrationIOFunc(void *arg)
1313
{
1314
    qemuMigrationIOThreadPtr data = arg;
1315 1316 1317 1318 1319
    char *buffer;
    int nbytes = TUNNEL_SEND_BUF_SIZE;

    if (VIR_ALLOC_N(buffer, TUNNEL_SEND_BUF_SIZE) < 0) {
        virReportOOMError();
1320 1321
        virStreamAbort(data->st);
        goto error;
1322 1323 1324
    }

    for (;;) {
1325
        nbytes = saferead(data->sock, buffer, TUNNEL_SEND_BUF_SIZE);
1326 1327 1328
        if (nbytes < 0) {
            virReportSystemError(errno, "%s",
                                 _("tunnelled migration failed to read from qemu"));
1329
            virStreamAbort(data->st);
1330
            VIR_FREE(buffer);
1331
            goto error;
1332 1333 1334 1335 1336
        }
        else if (nbytes == 0)
            /* EOF; get out of here */
            break;

1337
        if (virStreamSend(data->st, buffer, nbytes) < 0) {
1338
            VIR_FREE(buffer);
1339
            goto error;
1340 1341 1342 1343 1344
        }
    }

    VIR_FREE(buffer);

1345 1346
    if (virStreamFinish(data->st) < 0)
        goto error;
1347

1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399
    return;

error:
    virCopyLastError(&data->err);
    virResetLastError();
}


static qemuMigrationIOThreadPtr
qemuMigrationStartTunnel(virStreamPtr st,
                         int sock)
{
    qemuMigrationIOThreadPtr io;

    if (VIR_ALLOC(io) < 0) {
        virReportOOMError();
        return NULL;
    }

    io->st = st;
    io->sock = sock;

    if (virThreadCreate(&io->thread, true,
                        qemuMigrationIOFunc,
                        io) < 0) {
        virReportSystemError(errno, "%s",
                             _("Unable to create migration thread"));
        VIR_FREE(io);
        return NULL;
    }

    return io;
}

static int
qemuMigrationStopTunnel(qemuMigrationIOThreadPtr io)
{
    int rv = -1;
    virThreadJoin(&io->thread);

    /* Forward error from the IO thread, to this thread */
    if (io->err.code != VIR_ERR_OK) {
        virSetError(&io->err);
        virResetError(&io->err);
        goto cleanup;
    }

    rv = 0;

cleanup:
    VIR_FREE(io);
    return rv;
1400 1401
}

1402 1403 1404 1405 1406 1407 1408 1409 1410 1411
static int
qemuMigrationRun(struct qemud_driver *driver,
                 virDomainObjPtr vm,
                 const char *cookiein,
                 int cookieinlen,
                 char **cookieout,
                 int *cookieoutlen,
                 unsigned long flags,
                 unsigned long resource,
                 qemuMigrationSpecPtr spec)
1412
{
1413
    int ret = -1;
1414 1415
    unsigned int migrate_flags = QEMU_MONITOR_MIGRATE_BACKGROUND;
    qemuDomainObjPrivatePtr priv = vm->privateData;
1416
    qemuMigrationCookiePtr mig = NULL;
1417
    qemuMigrationIOThreadPtr iothread = NULL;
1418 1419 1420 1421 1422 1423 1424 1425
    int fd = -1;

    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);
1426

1427 1428 1429
    if (virLockManagerPluginUsesState(driver->lockManager) &&
        !cookieout) {
        qemuReportError(VIR_ERR_INTERNAL_ERROR,
1430 1431
                        _("Migration with lock driver %s requires"
                          " cookie support"),
1432 1433 1434 1435 1436
                        virLockManagerPluginGetName(driver->lockManager));
        return -1;
    }

    if (!(mig = qemuMigrationEatCookie(driver, vm, cookiein, cookieinlen,
1437 1438 1439 1440 1441 1442
                                       QEMU_MIGRATION_COOKIE_GRAPHICS)))
        goto cleanup;

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

1443
    /* Before EnterMonitor, since qemuMigrationSetOffline already does that */
1444 1445 1446 1447 1448 1449
    if (!(flags & VIR_MIGRATE_LIVE) &&
        virDomainObjGetState(vm, NULL) == VIR_DOMAIN_RUNNING) {
        if (qemuMigrationSetOffline(driver, vm) < 0)
            goto cleanup;
    }

1450 1451
    if (qemuDomainObjEnterMonitorAsync(driver, vm,
                                       QEMU_ASYNC_JOB_MIGRATION_OUT) < 0)
1452 1453
        goto cleanup;

1454 1455 1456 1457 1458
    if (resource > 0 &&
        qemuMonitorSetMigrationSpeed(priv->mon, resource) < 0) {
        qemuDomainObjExitMonitorWithDriver(driver, vm);
        goto cleanup;
    }
1459

1460
    if (flags & VIR_MIGRATE_NON_SHARED_DISK)
1461 1462
        migrate_flags |= QEMU_MONITOR_MIGRATE_NON_SHARED_DISK;

1463
    if (flags & VIR_MIGRATE_NON_SHARED_INC)
1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483
        migrate_flags |= QEMU_MONITOR_MIGRATE_NON_SHARED_INC;

    switch (spec->destType) {
    case MIGRATION_DEST_HOST:
        ret = qemuMonitorMigrateToHost(priv->mon, migrate_flags,
                                       spec->dest.host.name,
                                       spec->dest.host.port);
        break;

    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;
1484 1485 1486 1487 1488 1489 1490 1491

    case MIGRATION_DEST_FD:
        if (spec->fwdType != MIGRATION_FWD_DIRECT)
            fd = spec->dest.fd.local;
        ret = qemuMonitorMigrateToFd(priv->mon, migrate_flags,
                                     spec->dest.fd.qemu);
        VIR_FORCE_CLOSE(spec->dest.fd.qemu);
        break;
1492 1493
    }
    qemuDomainObjExitMonitorWithDriver(driver, vm);
1494
    if (ret < 0)
1495 1496
        goto cleanup;
    ret = -1;
1497 1498 1499 1500 1501 1502 1503 1504 1505 1506

    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 */

1507 1508 1509 1510 1511 1512 1513 1514
    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;
1515

1516 1517 1518 1519 1520 1521 1522
        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;
        }
1523 1524
    }

1525 1526
    if (spec->fwdType != MIGRATION_FWD_DIRECT &&
        !(iothread = qemuMigrationStartTunnel(spec->fwd.stream, fd)))
1527 1528
        goto cancel;

1529 1530 1531
    if (qemuMigrationWaitForCompletion(driver, vm,
                                       QEMU_ASYNC_JOB_MIGRATION_OUT) < 0)
        goto cleanup;
1532

1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545
    /* 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;
    }

1546
    ret = 0;
1547

1548 1549 1550 1551 1552 1553 1554
cleanup:
    if (spec->fwdType != MIGRATION_FWD_DIRECT) {
        /* Close now to ensure the IO thread quits & is joinable */
        VIR_FORCE_CLOSE(fd);
        if (iothread && qemuMigrationStopTunnel(iothread) < 0)
            ret = -1;
    }
1555

1556 1557 1558 1559
    if (ret == 0 &&
        qemuMigrationBakeCookie(mig, driver, vm, cookieout, cookieoutlen, 0) < 0)
        VIR_WARN("Unable to encode migration cookie");

1560 1561 1562 1563
    qemuMigrationCookieFree(mig);

    return ret;

1564
cancel:
1565
    if (virDomainObjIsActive(vm)) {
1566 1567
        if (qemuDomainObjEnterMonitorAsync(driver, vm,
                                           QEMU_ASYNC_JOB_MIGRATION_OUT) == 0) {
1568 1569 1570
            qemuMonitorMigrateCancel(priv->mon);
            qemuDomainObjExitMonitorWithDriver(driver, vm);
        }
1571
    }
1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587
    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,
                           unsigned long resource)
{
1588
    qemuDomainObjPrivatePtr priv = vm->privateData;
1589
    xmlURIPtr uribits = NULL;
1590
    int ret = -1;
1591
    qemuMigrationSpec spec;
1592
    char *tmp = NULL;
1593 1594 1595 1596 1597 1598 1599 1600

    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://")) {
        /* HACK: source host generates bogus URIs, so fix them up */
1601
        if (virAsprintf(&tmp, "tcp://%s", uri + strlen("tcp:")) < 0) {
1602 1603 1604
            virReportOOMError();
            return -1;
        }
1605 1606
        uribits = xmlParseURI(tmp);
        VIR_FREE(tmp);
1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617
    } else {
        uribits = xmlParseURI(uri);
    }
    if (!uribits) {
        qemuReportError(VIR_ERR_INTERNAL_ERROR,
                        _("cannot parse URI %s"), uri);
        return -1;
    }

    spec.fwdType = MIGRATION_FWD_DIRECT;

1618 1619 1620 1621 1622 1623 1624 1625 1626 1627
    if (qemuCapsGet(priv->qemuCaps, QEMU_CAPS_MIGRATE_QEMU_FD)) {
        virNetSocketPtr sock;

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

        if (virAsprintf(&tmp, "%d", uribits->port) < 0) {
            virReportOOMError();
            goto cleanup;
        }
1628 1629
        if (virSecurityManagerSetSocketLabel(driver->securityManager, vm) < 0)
            goto cleanup;
1630 1631 1632 1633
        if (virNetSocketNewConnectTCP(uribits->server, tmp, &sock) == 0) {
            spec.dest.fd.qemu = virNetSocketDupFD(sock, true);
            virNetSocketFree(sock);
        }
1634 1635
        if (virSecurityManagerClearSocketLabel(driver->securityManager, vm) < 0 ||
            spec.dest.fd.qemu == -1)
1636 1637 1638 1639 1640 1641 1642
            goto cleanup;
    } else {
        spec.destType = MIGRATION_DEST_HOST;
        spec.dest.host.name = uribits->server;
        spec.dest.host.port = uribits->port;
    }

1643 1644
    ret = qemuMigrationRun(driver, vm, cookiein, cookieinlen, cookieout,
                           cookieoutlen, flags, resource, &spec);
1645 1646 1647 1648 1649 1650

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

    VIR_FREE(tmp);
1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667
    xmlFreeURI(uribits);

    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,
                           unsigned long resource)
{
    qemuDomainObjPrivatePtr priv = vm->privateData;
1668
    virNetSocketPtr sock = NULL;
1669 1670 1671 1672 1673 1674 1675 1676
    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);

1677 1678
    if (!qemuCapsGet(priv->qemuCaps, QEMU_CAPS_MIGRATE_QEMU_FD) &&
        !qemuCapsGet(priv->qemuCaps, QEMU_CAPS_MIGRATE_QEMU_UNIX) &&
1679
        !qemuCapsGet(priv->qemuCaps, QEMU_CAPS_MIGRATE_QEMU_EXEC)) {
1680 1681 1682
        qemuReportError(VIR_ERR_OPERATION_FAILED, "%s",
                        _("Source qemu is too old to support tunnelled migration"));
        return -1;
1683 1684 1685 1686 1687
    }

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

1688 1689 1690 1691 1692 1693 1694
    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;

1695
        if (pipe2(fds, O_CLOEXEC) == 0) {
1696 1697 1698 1699
            spec.dest.fd.qemu = fds[1];
            spec.dest.fd.local = fds[0];
        }
        if (spec.dest.fd.qemu == -1 ||
1700 1701
            virSecurityManagerSetImageFDLabel(driver->securityManager, vm,
                                              spec.dest.fd.qemu) < 0) {
1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726
            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);
    }

1727 1728
    ret = qemuMigrationRun(driver, vm, cookiein, cookieinlen, cookieout,
                           cookieoutlen, flags, resource, &spec);
1729

1730
cleanup:
1731 1732 1733 1734 1735 1736 1737
    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);
    }
1738 1739 1740 1741 1742

    return ret;
}


1743 1744 1745 1746 1747 1748
/* 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,
                               virConnectPtr sconn,
1749 1750
                               virConnectPtr dconn,
                               virDomainObjPtr vm,
1751
                               const char *dconnuri,
1752 1753 1754
                               unsigned long flags,
                               const char *dname,
                               unsigned long resource)
1755 1756 1757
{
    virDomainPtr ddomain = NULL;
    char *uri_out = NULL;
1758
    char *cookie = NULL;
1759 1760 1761 1762 1763
    char *dom_xml = NULL;
    int cookielen = 0, ret;
    virErrorPtr orig_err = NULL;
    int cancelled;
    virStreamPtr st = NULL;
1764
    VIR_DEBUG("driver=%p, sconn=%p, dconn=%p, vm=%p, dconnuri=%s, "
1765
              "flags=%lx, dname=%s, resource=%lu",
1766 1767
              driver, sconn, dconn, vm, NULLSTR(dconnuri),
              flags, NULLSTR(dname), resource);
1768

1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 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
    /* 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 |
                                        VIR_DOMAIN_XML_UPDATE_CPU)))
        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)
1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814
        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;
    }

1815 1816 1817
    if (!(flags & VIR_MIGRATE_TUNNELLED) &&
        (uri_out == NULL)) {
        qemuReportError(VIR_ERR_INTERNAL_ERROR,
1818
                        _("domainMigratePrepare2 did not set uri"));
1819 1820
        cancelled = 1;
        goto finish;
1821 1822
    }

1823 1824 1825 1826
    /* Perform the migration.  The driver isn't supposed to return
     * until the migration is complete.
     */
    VIR_DEBUG("Perform %p", sconn);
1827
    qemuMigrationJobSetPhase(driver, vm, QEMU_MIGRATION_PHASE_PERFORM2);
1828
    if (flags & VIR_MIGRATE_TUNNELLED)
1829 1830 1831
        ret = doTunnelMigrate(driver, vm, st,
                              NULL, 0, NULL, NULL,
                              flags, resource);
1832 1833 1834 1835
    else
        ret = doNativeMigrate(driver, vm, uri_out,
                              cookie, cookielen,
                              NULL, NULL, /* No out cookie with v2 migration */
1836
                              flags, resource);
1837 1838 1839 1840

    /* Perform failed. Make sure Finish doesn't overwrite the error */
    if (ret < 0)
        orig_err = virSaveLastError();
1841

1842 1843 1844 1845
    /* If Perform returns < 0, then we need to cancel the VM
     * startup on the destination
     */
    cancelled = ret < 0 ? 1 : 0;
1846 1847

finish:
1848 1849 1850 1851
    /* 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.
     */
1852
    dname = dname ? dname : vm->def->name;
1853
    VIR_DEBUG("Finish2 %p ret=%d", dconn, ret);
1854 1855
    qemuDomainObjEnterRemoteWithDriver(driver, vm);
    ddomain = dconn->driver->domainMigrateFinish2
1856
        (dconn, dname, cookie, cookielen,
1857
         uri_out ? uri_out : dconnuri, flags, cancelled);
1858 1859
    qemuDomainObjExitRemoteWithDriver(driver, vm);

1860 1861
cleanup:
    if (ddomain) {
1862
        virUnrefDomain(ddomain);
1863 1864 1865 1866
        ret = 0;
    } else {
        ret = -1;
    }
1867

1868 1869 1870 1871 1872 1873 1874 1875
    if (st)
        virUnrefStream(st);

    if (orig_err) {
        virSetError(orig_err);
        virFreeError(orig_err);
    }
    VIR_FREE(uri_out);
1876
    VIR_FREE(cookie);
1877 1878

    return ret;
1879 1880 1881
}


1882 1883 1884 1885 1886 1887 1888 1889
/* 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,
1890
                               const char *xmlin,
1891
                               const char *dconnuri,
1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907
                               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;
1908
    VIR_DEBUG("driver=%p, sconn=%p, dconn=%p, vm=%p, xmlin=%s, "
1909
              "dconnuri=%s, uri=%s, flags=%lx, dname=%s, resource=%lu",
1910 1911 1912
              driver, sconn, dconn, vm, NULLSTR(xmlin),
              NULLSTR(dconnuri), NULLSTR(uri), flags,
              NULLSTR(dname), resource);
1913

1914 1915 1916 1917 1918
    /* 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.  */

1919
    dom_xml = qemuMigrationBegin(driver, vm, xmlin,
1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945
                                 &cookieout, &cookieoutlen);
    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,
1946
             uri, &uri_out, flags, dname, resource, dom_xml);
1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965
        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.
     */
1966
    VIR_DEBUG("Perform3 %p uri=%s uri_out=%s", sconn, uri, uri_out);
1967
    qemuMigrationJobSetPhase(driver, vm, QEMU_MIGRATION_PHASE_PERFORM3);
1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981
    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,
                              flags, resource);
    else
        ret = doNativeMigrate(driver, vm, uri_out,
                              cookiein, cookieinlen,
                              &cookieout, &cookieoutlen,
1982
                              flags, resource);
1983 1984

    /* Perform failed. Make sure Finish doesn't overwrite the error */
1985
    if (ret < 0) {
1986
        orig_err = virSaveLastError();
1987 1988 1989 1990
    } else {
        qemuMigrationJobSetPhase(driver, vm,
                                 QEMU_MIGRATION_PHASE_PERFORM3_DONE);
    }
1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011

    /* 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);
2012
    ddomain = dconn->driver->domainMigrateFinish3
2013
        (dconn, dname, cookiein, cookieinlen, &cookieout, &cookieoutlen,
2014
         dconnuri, uri_out ? uri_out : uri, flags, cancelled);
2015 2016
    qemuDomainObjExitRemoteWithDriver(driver, vm);

2017 2018 2019 2020 2021 2022 2023
    /* 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.
2024
     */
2025
    cancelled = ddomain == NULL ? 1 : 0;
2026

2027 2028 2029 2030 2031 2032
    /* 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();

2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044
    /*
     * 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,
2045
                               flags, cancelled);
2046 2047 2048 2049
    /* 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.
     */
2050 2051 2052
    if (ret < 0)
        VIR_WARN("Guest %s probably left in 'paused' state on source",
                 vm->def->name);
2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076

 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;
}


2077
static int doPeer2PeerMigrate(struct qemud_driver *driver,
2078
                              virConnectPtr sconn,
2079
                              virDomainObjPtr vm,
2080
                              const char *xmlin,
2081
                              const char *dconnuri,
2082 2083 2084
                              const char *uri,
                              unsigned long flags,
                              const char *dname,
2085 2086
                              unsigned long resource,
                              bool *v3proto)
2087 2088 2089 2090
{
    int ret = -1;
    virConnectPtr dconn = NULL;
    bool p2p;
2091 2092
    virErrorPtr orig_err = NULL;

2093
    VIR_DEBUG("driver=%p, sconn=%p, vm=%p, xmlin=%s, dconnuri=%s, "
2094
              "uri=%s, flags=%lx, dname=%s, resource=%lu",
2095 2096
              driver, sconn, vm, NULLSTR(xmlin), NULLSTR(dconnuri),
              NULLSTR(uri), flags, NULLSTR(dname), resource);
2097 2098 2099 2100 2101 2102

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

    qemuDomainObjEnterRemoteWithDriver(driver, vm);
2103
    dconn = virConnectOpen(dconnuri);
2104 2105 2106
    qemuDomainObjExitRemoteWithDriver(driver, vm);
    if (dconn == NULL) {
        qemuReportError(VIR_ERR_OPERATION_FAILED,
2107
                        _("Failed to connect to remote libvirt URI %s"), dconnuri);
2108 2109 2110 2111 2112 2113
        return -1;
    }

    qemuDomainObjEnterRemoteWithDriver(driver, vm);
    p2p = VIR_DRV_SUPPORTS_FEATURE(dconn->driver, dconn,
                                   VIR_DRV_FEATURE_MIGRATION_P2P);
2114
        /* v3proto reflects whether the caller used Perform3, but with
2115
         * p2p migrate, regardless of whether Perform2 or Perform3
2116 2117 2118 2119
         * were used, we decide protocol based on what target supports
         */
    *v3proto = VIR_DRV_SUPPORTS_FEATURE(dconn->driver, dconn,
                                        VIR_DRV_FEATURE_MIGRATION_V3);
2120
    qemuDomainObjExitRemoteWithDriver(driver, vm);
2121

2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134
    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;
    }

2135 2136 2137 2138 2139 2140 2141
    /* 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;

2142
    if (*v3proto)
2143
        ret = doPeer2PeerMigrate3(driver, sconn, dconn, vm, xmlin,
2144
                                  dconnuri, uri, flags, dname, resource);
2145 2146
    else
        ret = doPeer2PeerMigrate2(driver, sconn, dconn, vm,
2147
                                  dconnuri, flags, dname, resource);
2148 2149

cleanup:
2150
    orig_err = virSaveLastError();
2151
    qemuDomainObjEnterRemoteWithDriver(driver, vm);
2152
    virConnectClose(dconn);
2153
    qemuDomainObjExitRemoteWithDriver(driver, vm);
2154 2155 2156 2157
    if (orig_err) {
        virSetError(orig_err);
        virFreeError(orig_err);
    }
2158 2159 2160 2161 2162

    return ret;
}


2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182
/*
 * 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)
2183 2184 2185 2186 2187
{
    virDomainEventPtr event = NULL;
    int ret = -1;
    int resume = 0;

2188
    if (qemuMigrationJobStart(driver, vm, QEMU_ASYNC_JOB_MIGRATION_OUT) < 0)
2189 2190 2191 2192 2193 2194 2195 2196
        goto cleanup;

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

2197 2198 2199 2200 2201 2202
    if (qemuProcessAutoDestroyActive(driver, vm)) {
        qemuReportError(VIR_ERR_OPERATION_INVALID,
                        "%s", _("domain is marked for auto destroy"));
        goto endjob;
    }

J
Jiri Denemark 已提交
2203
    resume = virDomainObjGetState(vm, NULL) == VIR_DOMAIN_RUNNING;
2204 2205

    if ((flags & (VIR_MIGRATE_TUNNELLED | VIR_MIGRATE_PEER2PEER))) {
2206 2207 2208
        ret = doPeer2PeerMigrate(driver, conn, vm, xmlin,
                                 dconnuri, uri, flags, dname,
                                 resource, &v3proto);
2209
    } else {
2210 2211 2212
        qemuMigrationJobSetPhase(driver, vm, QEMU_MIGRATION_PHASE_PERFORM2);
        ret = doNativeMigrate(driver, vm, uri, cookiein, cookieinlen,
                              cookieout, cookieoutlen,
2213
                              flags, resource);
2214
    }
2215 2216
    if (ret < 0)
        goto endjob;
2217

2218 2219 2220 2221
    /*
     * In v3 protocol, the source VM is not killed off until the
     * confirm step.
     */
2222
    if (!v3proto) {
2223
        qemuProcessStop(driver, vm, 1, VIR_DOMAIN_SHUTOFF_MIGRATED);
2224
        virDomainAuditStop(vm, "migrated");
2225 2226 2227
        event = virDomainEventNewFromObj(vm,
                                         VIR_DOMAIN_EVENT_STOPPED,
                                         VIR_DOMAIN_EVENT_STOPPED_MIGRATED);
2228
    }
2229
    resume = 0;
2230 2231

endjob:
J
Jiri Denemark 已提交
2232
    if (resume && virDomainObjGetState(vm, NULL) == VIR_DOMAIN_PAUSED) {
2233
        /* we got here through some sort of failure; start the domain again */
J
Jiri Denemark 已提交
2234
        if (qemuProcessStartCPUs(driver, vm, conn,
2235 2236
                                 VIR_DOMAIN_RUNNING_MIGRATION_CANCELED,
                                 QEMU_ASYNC_JOB_MIGRATION_OUT) < 0) {
2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248
            /* 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);
    }
2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301

    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);
        virDomainRemoveInactive(&driver->domains, vm);
        vm = NULL;
    }

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);

    resume = virDomainObjGetState(vm, NULL) == VIR_DOMAIN_RUNNING;
    ret = doNativeMigrate(driver, vm, uri, cookiein, cookieinlen,
                          cookieout, cookieoutlen,
2302
                          flags, resource);
2303 2304 2305 2306 2307

    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,
2308 2309
                                 VIR_DOMAIN_RUNNING_MIGRATION_CANCELED,
                                 QEMU_ASYNC_JOB_MIGRATION_OUT) < 0) {
2310 2311 2312 2313 2314 2315
            /* 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);
2316
        }
2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337

        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);

endjob:
    if (ret < 0)
        refs = qemuMigrationJobFinish(driver, vm);
    else
        refs = qemuMigrationJobContinue(vm);
    if (refs == 0) {
        vm = NULL;
    } else if (!virDomainObjIsActive(vm) && !vm->persistent) {
        virDomainRemoveInactive(&driver->domains, vm);
        vm = NULL;
2338
    }
2339 2340 2341 2342 2343 2344 2345 2346 2347

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

2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393
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,
2394
                                             flags, resource);
2395 2396 2397 2398 2399 2400 2401 2402
        } else {
            return qemuMigrationPerformJob(driver, conn, vm, xmlin, dconnuri,
                                           uri, cookiein, cookieinlen,
                                           cookieout, cookieoutlen, flags,
                                           dname, resource, v3proto);
        }
    }
}
2403 2404 2405 2406 2407 2408 2409 2410 2411 2412

#if WITH_MACVTAP
static void
qemuMigrationVPAssociatePortProfiles(virDomainDefPtr def) {
    int i;
    int last_good_net = -1;
    virDomainNetDefPtr net;

    for (i = 0; i < def->nnets; i++) {
        net = def->nets[i];
2413
        if (virDomainNetGetActualType(net) == VIR_DOMAIN_NET_TYPE_DIRECT) {
2414 2415
            if (vpAssociatePortProfileId(net->ifname,
                                         net->mac,
2416 2417
                                         virDomainNetGetActualDirectDev(net),
                                         virDomainNetGetActualDirectVirtPortProfile(net),
2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429
                                         def->uuid,
                                         VIR_VM_OP_MIGRATE_IN_FINISH) != 0)
                goto err_exit;
        }
        last_good_net = i;
    }

    return;

err_exit:
    for (i = 0; i < last_good_net; i++) {
        net = def->nets[i];
2430
        if (virDomainNetGetActualType(net) == VIR_DOMAIN_NET_TYPE_DIRECT) {
2431 2432
            vpDisassociatePortProfileId(net->ifname,
                                        net->mac,
2433 2434
                                        virDomainNetGetActualDirectDev(net),
                                        virDomainNetGetActualDirectVirtPortProfile(net),
2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448
                                        VIR_VM_OP_MIGRATE_IN_FINISH);
        }
    }
}
#else /* !WITH_MACVTAP */
static void
qemuMigrationVPAssociatePortProfiles(virDomainDefPtr def ATTRIBUTE_UNUSED) { }
#endif /* WITH_MACVTAP */


virDomainPtr
qemuMigrationFinish(struct qemud_driver *driver,
                    virConnectPtr dconn,
                    virDomainObjPtr vm,
2449 2450 2451 2452
                    const char *cookiein,
                    int cookieinlen,
                    char **cookieout,
                    int *cookieoutlen,
2453
                    unsigned long flags,
2454 2455
                    int retcode,
                    bool v3proto)
2456 2457 2458 2459
{
    virDomainPtr dom = NULL;
    virDomainEventPtr event = NULL;
    int newVM = 1;
2460
    qemuMigrationCookiePtr mig = NULL;
2461 2462
    virErrorPtr orig_err = NULL;

2463
    VIR_DEBUG("driver=%p, dconn=%p, vm=%p, cookiein=%s, cookieinlen=%d, "
2464
              "cookieout=%p, cookieoutlen=%p, flags=%lx, retcode=%d",
2465 2466
              driver, dconn, vm, NULLSTR(cookiein), cookieinlen,
              cookieout, cookieoutlen, flags, retcode);
2467

2468
    if (!qemuMigrationJobIsActive(vm, QEMU_ASYNC_JOB_MIGRATION_IN))
2469 2470
        goto cleanup;

2471 2472 2473
    qemuMigrationJobStartPhase(driver, vm,
                               v3proto ? QEMU_MIGRATION_PHASE_FINISH3
                                       : QEMU_MIGRATION_PHASE_FINISH2);
2474

2475 2476
    if (!(mig = qemuMigrationEatCookie(driver, vm, cookiein, cookieinlen, 0)))
        goto endjob;
2477 2478 2479 2480 2481 2482 2483 2484

    /* 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"));
2485
            goto endjob;
2486 2487 2488 2489 2490
        }

        qemuMigrationVPAssociatePortProfiles(vm->def);

        if (flags & VIR_MIGRATE_PERSIST_DEST) {
2491
            virDomainDefPtr vmdef;
2492 2493 2494
            if (vm->persistent)
                newVM = 0;
            vm->persistent = 1;
2495 2496
            vmdef = virDomainObjGetPersistentDef(driver->caps, vm);
            if (virDomainSaveConfig(driver->configDir, vmdef) < 0) {
2497 2498 2499 2500 2501 2502 2503 2504 2505
                /* 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.
                 */
2506 2507

                /*
2508 2509
                 * However, in v3 protocol, the source VM is still available
                 * to restart during confirm() step, so we kill it off now.
2510 2511 2512
                 */
                if (v3proto) {
                    qemuProcessStop(driver, vm, 1, VIR_DOMAIN_SHUTOFF_FAILED);
2513
                    virDomainAuditStop(vm, "failed");
2514 2515
                    if (newVM)
                        vm->persistent = 0;
2516
                }
2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 2527 2528 2529 2530 2531 2532 2533 2534
                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
2535
            if (qemuProcessStartCPUs(driver, vm, dconn,
2536 2537
                                     VIR_DOMAIN_RUNNING_MIGRATED,
                                     QEMU_ASYNC_JOB_MIGRATION_IN) < 0) {
2538 2539 2540 2541 2542 2543 2544 2545
                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();

2546 2547 2548 2549 2550 2551 2552 2553 2554 2555
                /*
                 * 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) {
                    qemuProcessStop(driver, vm, 1, VIR_DOMAIN_SHUTOFF_FAILED);
2556
                    virDomainAuditStop(vm, "failed");
2557 2558 2559 2560
                    event = virDomainEventNewFromObj(vm,
                                                     VIR_DOMAIN_EVENT_STOPPED,
                                                     VIR_DOMAIN_EVENT_STOPPED_FAILED);
                }
2561 2562 2563 2564
                goto endjob;
            }
        }

2565 2566
        dom = virGetDomain (dconn, vm->def->name, vm->def->uuid);

2567 2568 2569
        event = virDomainEventNewFromObj(vm,
                                         VIR_DOMAIN_EVENT_RESUMED,
                                         VIR_DOMAIN_EVENT_RESUMED_MIGRATED);
J
Jiri Denemark 已提交
2570 2571
        if (virDomainObjGetState(vm, NULL) == VIR_DOMAIN_PAUSED) {
            virDomainObjSetState(vm, VIR_DOMAIN_PAUSED, VIR_DOMAIN_PAUSED_USER);
2572 2573
            if (event)
                qemuDomainEventQueue(driver, event);
2574 2575 2576 2577 2578 2579 2580 2581
            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;
        }
2582 2583 2584

        /* Guest is successfully running, so cancel previous auto destroy */
        qemuProcessAutoDestroyRemove(driver, vm);
2585
    } else {
J
Jiri Denemark 已提交
2586
        qemuProcessStop(driver, vm, 1, VIR_DOMAIN_SHUTOFF_FAILED);
2587
        virDomainAuditStop(vm, "failed");
2588 2589 2590 2591 2592
        event = virDomainEventNewFromObj(vm,
                                         VIR_DOMAIN_EVENT_STOPPED,
                                         VIR_DOMAIN_EVENT_STOPPED_FAILED);
    }

2593 2594 2595
    if (qemuMigrationBakeCookie(mig, driver, vm, cookieout, cookieoutlen, 0) < 0)
        VIR_WARN("Unable to encode migration cookie");

2596
endjob:
E
Eric Blake 已提交
2597 2598 2599 2600 2601
    if (qemuMigrationJobFinish(driver, vm) == 0) {
        vm = NULL;
    } else if (!vm->persistent && !virDomainObjIsActive(vm)) {
        virDomainRemoveInactive(&driver->domains, vm);
        vm = NULL;
2602
    }
2603 2604 2605 2606 2607 2608

cleanup:
    if (vm)
        virDomainObjUnlock(vm);
    if (event)
        qemuDomainEventQueue(driver, event);
2609
    qemuMigrationCookieFree(mig);
2610 2611 2612 2613
    if (orig_err) {
        virSetError(orig_err);
        virFreeError(orig_err);
    }
2614 2615
    return dom;
}
2616

2617 2618 2619 2620 2621 2622

int qemuMigrationConfirm(struct qemud_driver *driver,
                         virConnectPtr conn,
                         virDomainObjPtr vm,
                         const char *cookiein,
                         int cookieinlen,
E
Eric Blake 已提交
2623
                         unsigned int flags,
2624
                         int retcode)
2625 2626 2627 2628
{
    qemuMigrationCookiePtr mig;
    virDomainEventPtr event = NULL;
    int rv = -1;
2629
    VIR_DEBUG("driver=%p, conn=%p, vm=%p, cookiein=%s, cookieinlen=%d, "
2630
              "flags=%x, retcode=%d",
2631 2632
              driver, conn, vm, NULLSTR(cookiein), cookieinlen,
              flags, retcode);
2633

2634
    virCheckFlags(QEMU_MIGRATION_FLAGS, -1);
E
Eric Blake 已提交
2635

2636 2637 2638 2639 2640
    qemuMigrationJobSetPhase(driver, vm,
                             retcode == 0
                             ? QEMU_MIGRATION_PHASE_CONFIRM3
                             : QEMU_MIGRATION_PHASE_CONFIRM3_CANCELLED);

2641
    if (!(mig = qemuMigrationEatCookie(driver, vm, cookiein, cookieinlen, 0)))
2642 2643 2644 2645 2646 2647 2648
        return -1;

    /* Did the migration go as planned?  If yes, kill off the
     * domain object, but if no, resume CPUs
     */
    if (retcode == 0) {
        qemuProcessStop(driver, vm, 1, VIR_DOMAIN_SHUTOFF_MIGRATED);
2649
        virDomainAuditStop(vm, "migrated");
2650 2651 2652 2653 2654 2655 2656 2657 2658 2659 2660

        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,
2661 2662
                                 VIR_DOMAIN_RUNNING_MIGRATED,
                                 QEMU_ASYNC_JOB_MIGRATION_OUT) < 0) {
2663 2664 2665
            if (virGetLastError() == NULL)
                qemuReportError(VIR_ERR_INTERNAL_ERROR,
                                "%s", _("resume operation failed"));
2666
            goto cleanup;
2667 2668 2669 2670 2671 2672 2673
        }

        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);
2674
            goto cleanup;
2675 2676 2677 2678 2679 2680 2681 2682 2683 2684 2685 2686 2687
        }
    }

    qemuMigrationCookieFree(mig);
    rv = 0;

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


2688 2689 2690 2691 2692
/* 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,
2693 2694
                    bool is_reg, bool bypassSecurityDriver,
                    enum qemuDomainAsyncJob asyncJob)
2695 2696 2697 2698 2699 2700
{
    qemuDomainObjPrivatePtr priv = vm->privateData;
    virCgroupPtr cgroup = NULL;
    int ret = -1;
    int rc;
    bool restoreLabel = false;
2701 2702
    virCommandPtr cmd = NULL;
    int pipeFD[2] = { -1, -1 };
2703

2704
    if (qemuCapsGet(priv->qemuCaps, QEMU_CAPS_MIGRATE_QEMU_FD) &&
2705
        (!compressor || pipe(pipeFD) == 0)) {
2706
        /* All right! We can use fd migration, which means that qemu
2707 2708 2709
         * 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.  */
2710 2711
        if (virSecurityManagerSetImageFDLabel(driver->securityManager, vm,
                                              compressor ? pipeFD[1] : fd) < 0)
2712
            goto cleanup;
2713 2714 2715 2716 2717 2718 2719 2720 2721 2722 2723 2724 2725 2726 2727 2728 2729 2730
        bypassSecurityDriver = true;
    } else {
        /* Phooey - we have to fall back on exec migration, where qemu
         * has to popen() the file by name.  We might also stumble on
         * a race present in some qemu versions where it does a wait()
         * that botches pclose.  */
        if (!is_reg &&
            qemuCgroupControllerActive(driver,
                                       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);
2731
            virDomainAuditCgroupPath(vm, cgroup, "allow", path, "rw", rc);
2732 2733 2734 2735 2736 2737
            if (rc < 0) {
                virReportSystemError(-rc,
                                     _("Unable to allow device %s for %s"),
                                     path, vm->def->name);
                goto cleanup;
            }
2738
        }
2739 2740 2741
        if ((!bypassSecurityDriver) &&
            virSecurityManagerSetSavedStateLabel(driver->securityManager,
                                                 vm, path) < 0)
2742
            goto cleanup;
2743
        restoreLabel = true;
2744 2745
    }

2746
    if (qemuDomainObjEnterMonitorAsync(driver, vm, asyncJob) < 0)
2747 2748
        goto cleanup;

2749 2750 2751
    if (!compressor) {
        const char *args[] = { "cat", NULL };

2752
        if (qemuCapsGet(priv->qemuCaps, QEMU_CAPS_MIGRATE_QEMU_FD) &&
2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766 2767 2768
            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
        };
2769 2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787
        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)
2788
                VIR_WARN("failed to close intermediate pipe");
2789 2790 2791 2792 2793
        } else {
            rc = qemuMonitorMigrateToFile(priv->mon,
                                          QEMU_MONITOR_MIGRATE_BACKGROUND,
                                          args, path, offset);
        }
2794
    }
2795
    qemuDomainObjExitMonitorWithDriver(driver, vm);
2796 2797 2798 2799

    if (rc < 0)
        goto cleanup;

2800
    rc = qemuMigrationWaitForCompletion(driver, vm, asyncJob);
2801 2802 2803 2804

    if (rc < 0)
        goto cleanup;

2805 2806 2807
    if (cmd && virCommandWait(cmd, NULL) < 0)
        goto cleanup;

2808 2809 2810
    ret = 0;

cleanup:
2811 2812 2813
    VIR_FORCE_CLOSE(pipeFD[0]);
    VIR_FORCE_CLOSE(pipeFD[1]);
    virCommandFree(cmd);
2814 2815 2816 2817 2818 2819 2820 2821
    if (restoreLabel && (!bypassSecurityDriver) &&
        virSecurityManagerRestoreSavedStateLabel(driver->securityManager,
                                                 vm, path) < 0)
        VIR_WARN("failed to restore save state label on %s", path);

    if (cgroup != NULL) {
        rc = virCgroupDenyDevicePath(cgroup, path,
                                     VIR_CGROUP_DEVICE_RWM);
2822
        virDomainAuditCgroupPath(vm, cgroup, "deny", path, "rwm", rc);
2823 2824 2825 2826 2827 2828 2829
        if (rc < 0)
            VIR_WARN("Unable to deny device %s for %s %d",
                     path, vm->def->name, rc);
        virCgroupFree(&cgroup);
    }
    return ret;
}
2830 2831 2832 2833 2834 2835 2836 2837 2838 2839 2840

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

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

2841
    if (job == QEMU_ASYNC_JOB_MIGRATION_IN) {
2842
        qemuDomainObjSetAsyncJobMask(vm, QEMU_JOB_NONE);
2843 2844
    } else {
        qemuDomainObjSetAsyncJobMask(vm, DEFAULT_JOB_MASK |
2845
                                     JOB_MASK(QEMU_JOB_SUSPEND) |
2846 2847
                                     JOB_MASK(QEMU_JOB_MIGRATION_OP));
    }
2848 2849 2850 2851 2852 2853 2854 2855 2856 2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 2867 2868 2869 2870 2871 2872 2873 2874 2875 2876 2877 2878 2879 2880 2881 2882 2883 2884 2885 2886 2887 2888 2889 2890 2891 2892 2893 2894 2895 2896 2897 2898 2899 2900 2901 2902 2903 2904 2905 2906 2907 2908 2909 2910

    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)
{
    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);
}