qemu_process.c 92.2 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 25 26 27
/*
 * qemu_process.h: QEMU process management
 *
 * 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 <fcntl.h>
#include <unistd.h>
#include <signal.h>
#include <sys/stat.h>
28 29
#include <sys/time.h>
#include <sys/resource.h>
30 31 32 33 34 35 36 37 38 39 40 41

#include "qemu_process.h"
#include "qemu_domain.h"
#include "qemu_cgroup.h"
#include "qemu_capabilities.h"
#include "qemu_monitor.h"
#include "qemu_command.h"
#include "qemu_audit.h"
#include "qemu_hostdev.h"
#include "qemu_hotplug.h"
#include "qemu_bridge_filter.h"

42
#if HAVE_NUMACTL
E
Eric Blake 已提交
43
# define NUMA_VERSION1_COMPATIBILITY 1
44 45 46
# include <numa.h>
#endif

47 48 49 50 51 52 53 54 55 56 57
#include "datatypes.h"
#include "logging.h"
#include "virterror_internal.h"
#include "memory.h"
#include "hooks.h"
#include "files.h"
#include "util.h"
#include "c-ctype.h"
#include "nodeinfo.h"
#include "processinfo.h"
#include "domain_nwfilter.h"
58
#include "locking/domain_lock.h"
59
#include "uuid.h"
60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109

#define VIR_FROM_THIS VIR_FROM_QEMU

#define START_POSTFIX ": starting up\n"
#define SHUTDOWN_POSTFIX ": shutting down\n"

/**
 * qemudRemoveDomainStatus
 *
 * remove all state files of a domain from statedir
 *
 * Returns 0 on success
 */
static int
qemuProcessRemoveDomainStatus(struct qemud_driver *driver,
                              virDomainObjPtr vm)
{
    char ebuf[1024];
    char *file = NULL;

    if (virAsprintf(&file, "%s/%s.xml", driver->stateDir, vm->def->name) < 0) {
        virReportOOMError();
        return(-1);
    }

    if (unlink(file) < 0 && errno != ENOENT && errno != ENOTDIR)
        VIR_WARN("Failed to remove domain XML for %s: %s",
                 vm->def->name, virStrerror(errno, ebuf, sizeof(ebuf)));
    VIR_FREE(file);

    if (virFileDeletePid(driver->stateDir, vm->def->name) != 0)
        VIR_WARN("Failed to remove PID file for %s: %s",
                 vm->def->name, virStrerror(errno, ebuf, sizeof(ebuf)));


    return 0;
}


/* XXX figure out how to remove this */
extern struct qemud_driver *qemu_driver;

/*
 * This is a callback registered with a qemuMonitorPtr  instance,
 * and to be invoked when the monitor console hits an end of file
 * condition, or error, thus indicating VM shutdown should be
 * performed
 */
static void
qemuProcessHandleMonitorEOF(qemuMonitorPtr mon ATTRIBUTE_UNUSED,
110
                            virDomainObjPtr vm)
111 112 113 114
{
    struct qemud_driver *driver = qemu_driver;
    virDomainEventPtr event = NULL;
    qemuDomainObjPrivatePtr priv;
115 116 117
    int eventReason = VIR_DOMAIN_EVENT_STOPPED_SHUTDOWN;
    int stopReason = VIR_DOMAIN_SHUTOFF_SHUTDOWN;
    const char *auditReason = "shutdown";
118 119 120

    VIR_DEBUG("Received EOF on %p '%s'", vm, vm->def->name);

121
    qemuDriverLock(driver);
122 123 124 125 126
    virDomainObjLock(vm);

    if (!virDomainObjIsActive(vm)) {
        VIR_DEBUG("Domain %p is not active, ignoring EOF", vm);
        virDomainObjUnlock(vm);
127
        qemuDriverUnlock(driver);
128 129 130 131
        return;
    }

    priv = vm->privateData;
132
    if (priv->monJSON && !priv->gotShutdown) {
133 134
        VIR_DEBUG("Monitor connection to '%s' closed without SHUTDOWN event; "
                  "assuming the domain crashed", vm->def->name);
135 136 137
        eventReason = VIR_DOMAIN_EVENT_STOPPED_FAILED;
        stopReason = VIR_DOMAIN_SHUTOFF_FAILED;
        auditReason = "failed";
138 139 140 141
    }

    event = virDomainEventNewFromObj(vm,
                                     VIR_DOMAIN_EVENT_STOPPED,
142 143 144
                                     eventReason);
    qemuProcessStop(driver, vm, 0, stopReason);
    qemuAuditDomainStop(vm, auditReason);
145 146 147 148 149 150

    if (!vm->persistent)
        virDomainRemoveInactive(&driver->domains, vm);
    else
        virDomainObjUnlock(vm);

151
    if (event)
152
        qemuDomainEventQueue(driver, event);
153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174
    qemuDriverUnlock(driver);
}


/*
 * This is invoked when there is some kind of error
 * parsing data to/from the monitor. The VM can continue
 * to run, but no further monitor commands will be
 * allowed
 */
static void
qemuProcessHandleMonitorError(qemuMonitorPtr mon ATTRIBUTE_UNUSED,
                              virDomainObjPtr vm)
{
    struct qemud_driver *driver = qemu_driver;
    virDomainEventPtr event = NULL;

    VIR_DEBUG("Received error on %p '%s'", vm, vm->def->name);

    qemuDriverLock(driver);
    virDomainObjLock(vm);

175
    ((qemuDomainObjPrivatePtr) vm->privateData)->monError = true;
176 177 178 179 180
    event = virDomainEventControlErrorNewFromObj(vm);
    if (event)
        qemuDomainEventQueue(driver, event);

    virDomainObjUnlock(vm);
181
    qemuDriverUnlock(driver);
182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357
}


static virDomainDiskDefPtr
qemuProcessFindDomainDiskByPath(virDomainObjPtr vm,
                                const char *path)
{
    int i;

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

        disk = vm->def->disks[i];
        if (disk->src != NULL && STREQ(disk->src, path))
            return disk;
    }

    qemuReportError(VIR_ERR_INTERNAL_ERROR,
                    _("no disk found with path %s"),
                    path);
    return NULL;
}

static virDomainDiskDefPtr
qemuProcessFindDomainDiskByAlias(virDomainObjPtr vm,
                                 const char *alias)
{
    int i;

    if (STRPREFIX(alias, QEMU_DRIVE_HOST_PREFIX))
        alias += strlen(QEMU_DRIVE_HOST_PREFIX);

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

        disk = vm->def->disks[i];
        if (disk->info.alias != NULL && STREQ(disk->info.alias, alias))
            return disk;
    }

    qemuReportError(VIR_ERR_INTERNAL_ERROR,
                    _("no disk found with alias %s"),
                    alias);
    return NULL;
}

static int
qemuProcessGetVolumeQcowPassphrase(virConnectPtr conn,
                                   virDomainDiskDefPtr disk,
                                   char **secretRet,
                                   size_t *secretLen)
{
    virSecretPtr secret;
    char *passphrase;
    unsigned char *data;
    size_t size;
    int ret = -1;
    virStorageEncryptionPtr enc;

    if (!disk->encryption) {
        qemuReportError(VIR_ERR_INTERNAL_ERROR,
                        _("disk %s does not have any encryption information"),
                        disk->src);
        return -1;
    }
    enc = disk->encryption;

    if (!conn) {
        qemuReportError(VIR_ERR_INTERNAL_ERROR,
                        "%s", _("cannot find secrets without a connection"));
        goto cleanup;
    }

    if (conn->secretDriver == NULL ||
        conn->secretDriver->lookupByUUID == NULL ||
        conn->secretDriver->getValue == NULL) {
        qemuReportError(VIR_ERR_NO_SUPPORT, "%s",
                        _("secret storage not supported"));
        goto cleanup;
    }

    if (enc->format != VIR_STORAGE_ENCRYPTION_FORMAT_QCOW ||
        enc->nsecrets != 1 ||
        enc->secrets[0]->type !=
        VIR_STORAGE_ENCRYPTION_SECRET_TYPE_PASSPHRASE) {
        qemuReportError(VIR_ERR_XML_ERROR,
                        _("invalid <encryption> for volume %s"), disk->src);
        goto cleanup;
    }

    secret = conn->secretDriver->lookupByUUID(conn,
                                              enc->secrets[0]->uuid);
    if (secret == NULL)
        goto cleanup;
    data = conn->secretDriver->getValue(secret, &size,
                                        VIR_SECRET_GET_VALUE_INTERNAL_CALL);
    virUnrefSecret(secret);
    if (data == NULL)
        goto cleanup;

    if (memchr(data, '\0', size) != NULL) {
        memset(data, 0, size);
        VIR_FREE(data);
        qemuReportError(VIR_ERR_XML_ERROR,
                        _("format='qcow' passphrase for %s must not contain a "
                          "'\\0'"), disk->src);
        goto cleanup;
    }

    if (VIR_ALLOC_N(passphrase, size + 1) < 0) {
        memset(data, 0, size);
        VIR_FREE(data);
        virReportOOMError();
        goto cleanup;
    }
    memcpy(passphrase, data, size);
    passphrase[size] = '\0';

    memset(data, 0, size);
    VIR_FREE(data);

    *secretRet = passphrase;
    *secretLen = size;

    ret = 0;

cleanup:
    return ret;
}

static int
qemuProcessFindVolumeQcowPassphrase(qemuMonitorPtr mon ATTRIBUTE_UNUSED,
                                    virConnectPtr conn,
                                    virDomainObjPtr vm,
                                    const char *path,
                                    char **secretRet,
                                    size_t *secretLen)
{
    virDomainDiskDefPtr disk;
    int ret = -1;

    virDomainObjLock(vm);
    disk = qemuProcessFindDomainDiskByPath(vm, path);

    if (!disk)
        goto cleanup;

    ret = qemuProcessGetVolumeQcowPassphrase(conn, disk, secretRet, secretLen);

cleanup:
    virDomainObjUnlock(vm);
    return ret;
}


static int
qemuProcessHandleReset(qemuMonitorPtr mon ATTRIBUTE_UNUSED,
                       virDomainObjPtr vm)
{
    struct qemud_driver *driver = qemu_driver;
    virDomainEventPtr event;

    virDomainObjLock(vm);
    event = virDomainEventRebootNewFromObj(vm);
    virDomainObjUnlock(vm);

    if (event) {
        qemuDriverLock(driver);
        qemuDomainEventQueue(driver, event);
        qemuDriverUnlock(driver);
    }

    return 0;
}


358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428
/*
 * Since we have the '-no-shutdown' flag set, the
 * QEMU process will currently have guest OS shutdown
 * and the CPUS stopped. To fake the reboot, we thus
 * want todo a reset of the virtual hardware, followed
 * by restart of the CPUs. This should result in the
 * guest OS booting up again
 */
static void
qemuProcessFakeReboot(void *opaque)
{
    struct qemud_driver *driver = qemu_driver;
    virDomainObjPtr vm = opaque;
    qemuDomainObjPrivatePtr priv = vm->privateData;
    virDomainEventPtr event = NULL;
    int ret = -1;
    VIR_DEBUG("vm=%p", vm);
    qemuDriverLock(driver);
    virDomainObjLock(vm);
    if (qemuDomainObjBeginJob(vm) < 0)
        goto cleanup;

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

    qemuDomainObjEnterMonitorWithDriver(driver, vm);
    if (qemuMonitorSystemReset(priv->mon) < 0) {
        qemuDomainObjExitMonitorWithDriver(driver, vm);
        goto endjob;
    }
    qemuDomainObjExitMonitorWithDriver(driver, vm);

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

    if (qemuProcessStartCPUs(driver, vm, NULL,
                             VIR_DOMAIN_RUNNING_BOOTED) < 0) {
        if (virGetLastError() == NULL)
            qemuReportError(VIR_ERR_INTERNAL_ERROR,
                            "%s", _("resume operation failed"));
        goto endjob;
    }
    event = virDomainEventNewFromObj(vm,
                                     VIR_DOMAIN_EVENT_RESUMED,
                                     VIR_DOMAIN_EVENT_RESUMED_UNPAUSED);

    ret = 0;

endjob:
    if (qemuDomainObjEndJob(vm) == 0)
        vm = NULL;

cleanup:
    if (vm) {
        if (ret == -1)
            qemuProcessKill(vm);
        if (virDomainObjUnref(vm) > 0)
            virDomainObjUnlock(vm);
    }
    if (event)
        qemuDomainEventQueue(driver, event);
    qemuDriverUnlock(driver);
}


429 430 431 432
static int
qemuProcessHandleShutdown(qemuMonitorPtr mon ATTRIBUTE_UNUSED,
                          virDomainObjPtr vm)
{
433 434 435
    qemuDomainObjPrivatePtr priv = vm->privateData;
    VIR_DEBUG("vm=%p", vm);

436
    virDomainObjLock(vm);
437 438 439 440 441 442 443 444
    priv->gotShutdown = true;
    if (priv->fakeReboot) {
        virDomainObjRef(vm);
        virThread th;
        if (virThreadCreate(&th,
                            false,
                            qemuProcessFakeReboot,
                            vm) < 0) {
445
            VIR_ERROR(_("Failed to create reboot thread, killing domain"));
446 447 448 449 450 451 452 453 454
            qemuProcessKill(vm);
            if (virDomainObjUnref(vm) == 0)
                vm = NULL;
        }
    } else {
        qemuProcessKill(vm);
    }
    if (vm)
        virDomainObjUnlock(vm);
455 456 457 458 459 460 461 462 463 464 465 466 467

    return 0;
}


static int
qemuProcessHandleStop(qemuMonitorPtr mon ATTRIBUTE_UNUSED,
                      virDomainObjPtr vm)
{
    struct qemud_driver *driver = qemu_driver;
    virDomainEventPtr event = NULL;

    virDomainObjLock(vm);
J
Jiri Denemark 已提交
468
    if (virDomainObjGetState(vm, NULL) == VIR_DOMAIN_RUNNING) {
469
        qemuDomainObjPrivatePtr priv = vm->privateData;
470 471
        VIR_DEBUG("Transitioned guest %s to paused state due to unknown event",
                  vm->def->name);
472

J
Jiri Denemark 已提交
473
        virDomainObjSetState(vm, VIR_DOMAIN_PAUSED, VIR_DOMAIN_PAUSED_UNKNOWN);
474 475 476 477
        event = virDomainEventNewFromObj(vm,
                                         VIR_DOMAIN_EVENT_SUSPENDED,
                                         VIR_DOMAIN_EVENT_SUSPENDED_PAUSED);

478 479 480 481 482
        VIR_FREE(priv->lockState);
        if (virDomainLockProcessPause(driver->lockManager, vm, &priv->lockState) < 0)
            VIR_WARN("Unable to release lease on %s", vm->def->name);
        VIR_DEBUG("Preserving lock state '%s'", NULLSTR(priv->lockState));

483 484 485 486
        if (virDomainSaveStatus(driver->caps, driver->stateDir, vm) < 0) {
            VIR_WARN("Unable to save status on vm %s after state change",
                     vm->def->name);
        }
487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515
    }
    virDomainObjUnlock(vm);

    if (event) {
        qemuDriverLock(driver);
        if (event)
            qemuDomainEventQueue(driver, event);
        qemuDriverUnlock(driver);
    }

    return 0;
}


static int
qemuProcessHandleRTCChange(qemuMonitorPtr mon ATTRIBUTE_UNUSED,
                           virDomainObjPtr vm,
                           long long offset)
{
    struct qemud_driver *driver = qemu_driver;
    virDomainEventPtr event;

    virDomainObjLock(vm);
    event = virDomainEventRTCChangeNewFromObj(vm, offset);

    if (vm->def->clock.offset == VIR_DOMAIN_CLOCK_OFFSET_VARIABLE)
        vm->def->clock.data.adjustment = offset;

    if (virDomainSaveStatus(driver->caps, driver->stateDir, vm) < 0)
516
        VIR_WARN("unable to save domain status with RTC change");
517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542

    virDomainObjUnlock(vm);

    if (event) {
        qemuDriverLock(driver);
        qemuDomainEventQueue(driver, event);
        qemuDriverUnlock(driver);
    }

    return 0;
}


static int
qemuProcessHandleWatchdog(qemuMonitorPtr mon ATTRIBUTE_UNUSED,
                          virDomainObjPtr vm,
                          int action)
{
    struct qemud_driver *driver = qemu_driver;
    virDomainEventPtr watchdogEvent = NULL;
    virDomainEventPtr lifecycleEvent = NULL;

    virDomainObjLock(vm);
    watchdogEvent = virDomainEventWatchdogNewFromObj(vm, action);

    if (action == VIR_DOMAIN_EVENT_WATCHDOG_PAUSE &&
J
Jiri Denemark 已提交
543
        virDomainObjGetState(vm, NULL) == VIR_DOMAIN_RUNNING) {
544
        qemuDomainObjPrivatePtr priv = vm->privateData;
545 546
        VIR_DEBUG("Transitioned guest %s to paused state due to watchdog", vm->def->name);

J
Jiri Denemark 已提交
547
        virDomainObjSetState(vm, VIR_DOMAIN_PAUSED, VIR_DOMAIN_PAUSED_WATCHDOG);
548 549 550 551
        lifecycleEvent = virDomainEventNewFromObj(vm,
                                                  VIR_DOMAIN_EVENT_SUSPENDED,
                                                  VIR_DOMAIN_EVENT_SUSPENDED_WATCHDOG);

552 553 554 555 556
        VIR_FREE(priv->lockState);
        if (virDomainLockProcessPause(driver->lockManager, vm, &priv->lockState) < 0)
            VIR_WARN("Unable to release lease on %s", vm->def->name);
        VIR_DEBUG("Preserving lock state '%s'", NULLSTR(priv->lockState));

557 558 559 560
        if (virDomainSaveStatus(driver->caps, driver->stateDir, vm) < 0) {
            VIR_WARN("Unable to save status on vm %s after watchdog event",
                     vm->def->name);
        }
561 562 563 564 565 566 567
    }

    if (vm->def->watchdog->action == VIR_DOMAIN_WATCHDOG_ACTION_DUMP) {
        struct qemuDomainWatchdogEvent *wdEvent;
        if (VIR_ALLOC(wdEvent) == 0) {
            wdEvent->action = VIR_DOMAIN_WATCHDOG_ACTION_DUMP;
            wdEvent->vm = vm;
W
Wen Congyang 已提交
568 569 570 571 572
            /* Hold an extra reference because we can't allow 'vm' to be
             * deleted before handling watchdog event is finished.
             */
            virDomainObjRef(vm);
            if (virThreadPoolSendJob(driver->workerPool, wdEvent) < 0) {
573
                if (virDomainObjUnref(vm) == 0)
574
                    vm = NULL;
W
Wen Congyang 已提交
575 576
                VIR_FREE(wdEvent);
            }
577
        } else {
578
            virReportOOMError();
579
        }
580 581
    }

582 583
    if (vm)
        virDomainObjUnlock(vm);
584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627

    if (watchdogEvent || lifecycleEvent) {
        qemuDriverLock(driver);
        if (watchdogEvent)
            qemuDomainEventQueue(driver, watchdogEvent);
        if (lifecycleEvent)
            qemuDomainEventQueue(driver, lifecycleEvent);
        qemuDriverUnlock(driver);
    }

    return 0;
}


static int
qemuProcessHandleIOError(qemuMonitorPtr mon ATTRIBUTE_UNUSED,
                         virDomainObjPtr vm,
                         const char *diskAlias,
                         int action,
                         const char *reason)
{
    struct qemud_driver *driver = qemu_driver;
    virDomainEventPtr ioErrorEvent = NULL;
    virDomainEventPtr ioErrorEvent2 = NULL;
    virDomainEventPtr lifecycleEvent = NULL;
    const char *srcPath;
    const char *devAlias;
    virDomainDiskDefPtr disk;

    virDomainObjLock(vm);
    disk = qemuProcessFindDomainDiskByAlias(vm, diskAlias);

    if (disk) {
        srcPath = disk->src;
        devAlias = disk->info.alias;
    } else {
        srcPath = "";
        devAlias = "";
    }

    ioErrorEvent = virDomainEventIOErrorNewFromObj(vm, srcPath, devAlias, action);
    ioErrorEvent2 = virDomainEventIOErrorReasonNewFromObj(vm, srcPath, devAlias, action, reason);

    if (action == VIR_DOMAIN_EVENT_IO_ERROR_PAUSE &&
J
Jiri Denemark 已提交
628
        virDomainObjGetState(vm, NULL) == VIR_DOMAIN_RUNNING) {
629
        qemuDomainObjPrivatePtr priv = vm->privateData;
630 631
        VIR_DEBUG("Transitioned guest %s to paused state due to IO error", vm->def->name);

J
Jiri Denemark 已提交
632
        virDomainObjSetState(vm, VIR_DOMAIN_PAUSED, VIR_DOMAIN_PAUSED_IOERROR);
633 634 635 636
        lifecycleEvent = virDomainEventNewFromObj(vm,
                                                  VIR_DOMAIN_EVENT_SUSPENDED,
                                                  VIR_DOMAIN_EVENT_SUSPENDED_IOERROR);

637 638 639 640 641
        VIR_FREE(priv->lockState);
        if (virDomainLockProcessPause(driver->lockManager, vm, &priv->lockState) < 0)
            VIR_WARN("Unable to release lease on %s", vm->def->name);
        VIR_DEBUG("Preserving lock state '%s'", NULLSTR(priv->lockState));

642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702
        if (virDomainSaveStatus(driver->caps, driver->stateDir, vm) < 0)
            VIR_WARN("Unable to save status on vm %s after IO error", vm->def->name);
    }
    virDomainObjUnlock(vm);

    if (ioErrorEvent || ioErrorEvent2 || lifecycleEvent) {
        qemuDriverLock(driver);
        if (ioErrorEvent)
            qemuDomainEventQueue(driver, ioErrorEvent);
        if (ioErrorEvent2)
            qemuDomainEventQueue(driver, ioErrorEvent2);
        if (lifecycleEvent)
            qemuDomainEventQueue(driver, lifecycleEvent);
        qemuDriverUnlock(driver);
    }

    return 0;
}


static int
qemuProcessHandleGraphics(qemuMonitorPtr mon ATTRIBUTE_UNUSED,
                          virDomainObjPtr vm,
                          int phase,
                          int localFamily,
                          const char *localNode,
                          const char *localService,
                          int remoteFamily,
                          const char *remoteNode,
                          const char *remoteService,
                          const char *authScheme,
                          const char *x509dname,
                          const char *saslUsername)
{
    struct qemud_driver *driver = qemu_driver;
    virDomainEventPtr event;
    virDomainEventGraphicsAddressPtr localAddr = NULL;
    virDomainEventGraphicsAddressPtr remoteAddr = NULL;
    virDomainEventGraphicsSubjectPtr subject = NULL;
    int i;

    if (VIR_ALLOC(localAddr) < 0)
        goto no_memory;
    localAddr->family = localFamily;
    if (!(localAddr->service = strdup(localService)) ||
        !(localAddr->node = strdup(localNode)))
        goto no_memory;

    if (VIR_ALLOC(remoteAddr) < 0)
        goto no_memory;
    remoteAddr->family = remoteFamily;
    if (!(remoteAddr->service = strdup(remoteService)) ||
        !(remoteAddr->node = strdup(remoteNode)))
        goto no_memory;

    if (VIR_ALLOC(subject) < 0)
        goto no_memory;
    if (x509dname) {
        if (VIR_REALLOC_N(subject->identities, subject->nidentity+1) < 0)
            goto no_memory;
        subject->nidentity++;
703 704 705
        if (!(subject->identities[subject->nidentity-1].type = strdup("x509dname")) ||
            !(subject->identities[subject->nidentity-1].name = strdup(x509dname)))
            goto no_memory;
706 707 708 709 710
    }
    if (saslUsername) {
        if (VIR_REALLOC_N(subject->identities, subject->nidentity+1) < 0)
            goto no_memory;
        subject->nidentity++;
711 712 713
        if (!(subject->identities[subject->nidentity-1].type = strdup("saslUsername")) ||
            !(subject->identities[subject->nidentity-1].name = strdup(saslUsername)))
            goto no_memory;
714 715
    }

716
    virDomainObjLock(vm);
717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755
    event = virDomainEventGraphicsNewFromObj(vm, phase, localAddr, remoteAddr, authScheme, subject);
    virDomainObjUnlock(vm);

    if (event) {
        qemuDriverLock(driver);
        qemuDomainEventQueue(driver, event);
        qemuDriverUnlock(driver);
    }

    return 0;

no_memory:
    virReportOOMError();
    if (localAddr) {
        VIR_FREE(localAddr->service);
        VIR_FREE(localAddr->node);
        VIR_FREE(localAddr);
    }
    if (remoteAddr) {
        VIR_FREE(remoteAddr->service);
        VIR_FREE(remoteAddr->node);
        VIR_FREE(remoteAddr);
    }
    if (subject) {
        for (i = 0 ; i < subject->nidentity ; i++) {
            VIR_FREE(subject->identities[i].type);
            VIR_FREE(subject->identities[i].name);
        }
        VIR_FREE(subject->identities);
        VIR_FREE(subject);
    }

    return -1;
}


static void qemuProcessHandleMonitorDestroy(qemuMonitorPtr mon,
                                            virDomainObjPtr vm)
{
756 757 758 759
    qemuDomainObjPrivatePtr priv;

    virDomainObjLock(vm);
    priv = vm->privateData;
760 761
    if (priv->mon == mon)
        priv->mon = NULL;
762 763
    if (virDomainObjUnref(vm) > 0)
        virDomainObjUnlock(vm);
764 765 766 767 768
}

static qemuMonitorCallbacks monitorCallbacks = {
    .destroy = qemuProcessHandleMonitorDestroy,
    .eofNotify = qemuProcessHandleMonitorEOF,
769
    .errorNotify = qemuProcessHandleMonitorError,
770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800
    .diskSecretLookup = qemuProcessFindVolumeQcowPassphrase,
    .domainShutdown = qemuProcessHandleShutdown,
    .domainStop = qemuProcessHandleStop,
    .domainReset = qemuProcessHandleReset,
    .domainRTCChange = qemuProcessHandleRTCChange,
    .domainWatchdog = qemuProcessHandleWatchdog,
    .domainIOError = qemuProcessHandleIOError,
    .domainGraphics = qemuProcessHandleGraphics,
};

static int
qemuConnectMonitor(struct qemud_driver *driver, virDomainObjPtr vm)
{
    qemuDomainObjPrivatePtr priv = vm->privateData;
    int ret = -1;

    if (virSecurityManagerSetSocketLabel(driver->securityManager, vm) < 0) {
        VIR_ERROR(_("Failed to set security context for monitor for %s"),
                  vm->def->name);
        goto error;
    }

    /* Hold an extra reference because we can't allow 'vm' to be
     * deleted while the monitor is active */
    virDomainObjRef(vm);

    priv->mon = qemuMonitorOpen(vm,
                                priv->monConfig,
                                priv->monJSON,
                                &monitorCallbacks);

801
    /* Safe to ignore value since ref count was incremented above */
802
    if (priv->mon == NULL)
803
        ignore_value(virDomainObjUnref(vm));
804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843

    if (virSecurityManagerClearSocketLabel(driver->securityManager, vm) < 0) {
        VIR_ERROR(_("Failed to clear security context for monitor for %s"),
                  vm->def->name);
        goto error;
    }

    if (priv->mon == NULL) {
        VIR_INFO("Failed to connect monitor for %s", vm->def->name);
        goto error;
    }


    qemuDomainObjEnterMonitorWithDriver(driver, vm);
    ret = qemuMonitorSetCapabilities(priv->mon);
    qemuDomainObjExitMonitorWithDriver(driver, vm);

error:

    return ret;
}

typedef int qemuProcessLogHandleOutput(virDomainObjPtr vm,
                                       const char *output,
                                       int fd);

/*
 * Returns -1 for error, 0 on success
 */
static int
qemuProcessReadLogOutput(virDomainObjPtr vm,
                         int fd,
                         char *buf,
                         size_t buflen,
                         qemuProcessLogHandleOutput func,
                         const char *what,
                         int timeout)
{
    int retries = (timeout*10);
    int got = 0;
844 845 846 847
    char *debug = NULL;
    int ret = -1;
    char *filter_next = buf;

848 849
    buf[0] = '\0';

850 851 852 853 854 855 856
    /* This relies on log message format generated by virLogFormatString() and
     * might need to be modified when message format changes. */
    if (virAsprintf(&debug, ": %d: debug : ", vm->pid) < 0) {
        virReportOOMError();
        return -1;
    }

857
    while (retries) {
858
        ssize_t func_ret, bytes;
859
        int isdead = 0;
860
        char *eol;
861 862 863 864 865 866 867 868

        func_ret = func(vm, buf, fd);

        if (kill(vm->pid, 0) == -1 && errno == ESRCH)
            isdead = 1;

        /* Any failures should be detected before we read the log, so we
         * always have something useful to report on failure. */
869 870
        bytes = saferead(fd, buf+got, buflen-got-1);
        if (bytes < 0) {
871 872 873
            virReportSystemError(errno,
                                 _("Failure while reading %s log output"),
                                 what);
874
            goto cleanup;
875 876
        }

877
        got += bytes;
878
        buf[got] = '\0';
879 880 881 882 883 884 885 886 887 888 889 890 891

        /* Filter out debug messages from intermediate libvirt process */
        while ((eol = strchr(filter_next, '\n'))) {
            *eol = '\0';
            if (strstr(filter_next, debug)) {
                memmove(filter_next, eol + 1, got - (eol - buf));
                got -= eol + 1 - filter_next;
            } else {
                filter_next = eol + 1;
                *eol = '\n';
            }
        }

892 893 894 895
        if (got == buflen-1) {
            qemuReportError(VIR_ERR_INTERNAL_ERROR,
                            _("Out of space while reading %s log output: %s"),
                            what, buf);
896
            goto cleanup;
897 898 899 900 901 902
        }

        if (isdead) {
            qemuReportError(VIR_ERR_INTERNAL_ERROR,
                            _("Process exited while reading %s log output: %s"),
                            what, buf);
903
            goto cleanup;
904 905
        }

906 907 908 909
        if (func_ret <= 0) {
            ret = func_ret;
            goto cleanup;
        }
910 911 912 913 914 915 916 917

        usleep(100*1000);
        retries--;
    }

    qemuReportError(VIR_ERR_INTERNAL_ERROR,
                    _("Timed out while reading %s log output: %s"),
                    what, buf);
918 919 920 921

cleanup:
    VIR_FREE(debug);
    return ret;
922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976
}


/*
 * Look at a chunk of data from the QEMU stdout logs and try to
 * find a TTY device, as indicated by a line like
 *
 * char device redirected to /dev/pts/3
 *
 * Returns -1 for error, 0 success, 1 continue reading
 */
static int
qemuProcessExtractTTYPath(const char *haystack,
                          size_t *offset,
                          char **path)
{
    static const char needle[] = "char device redirected to";
    char *tmp, *dev;

    VIR_FREE(*path);
    /* First look for our magic string */
    if (!(tmp = strstr(haystack + *offset, needle))) {
        return 1;
    }
    tmp += sizeof(needle);
    dev = tmp;

    /*
     * And look for first whitespace character and nul terminate
     * to mark end of the pty path
     */
    while (*tmp) {
        if (c_isspace(*tmp)) {
            *path = strndup(dev, tmp-dev);
            if (*path == NULL) {
                virReportOOMError();
                return -1;
            }

            /* ... now further update offset till we get EOL */
            *offset = tmp - haystack;
            return 0;
        }
        tmp++;
    }

    /*
     * We found a path, but didn't find any whitespace,
     * so it must be still incomplete - we should at
     * least see a \n - indicate that we want to carry
     * on trying again
     */
    return 1;
}

977 978 979
static int
qemuProcessLookupPTYs(virDomainChrDefPtr *devices,
                      int count,
C
Cole Robinson 已提交
980 981
                      virHashTablePtr paths,
                      bool chardevfmt)
982 983
{
    int i;
C
Cole Robinson 已提交
984
    const char *prefix = chardevfmt ? "char" : "";
985 986 987 988

    for (i = 0 ; i < count ; i++) {
        virDomainChrDefPtr chr = devices[i];
        if (chr->source.type == VIR_DOMAIN_CHR_TYPE_PTY) {
C
Cole Robinson 已提交
989
            char id[32];
990 991
            const char *path;

C
Cole Robinson 已提交
992 993
            if (snprintf(id, sizeof(id), "%s%s",
                         prefix, chr->info.alias) >= sizeof(id))
994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025
                return -1;

            path = (const char *) virHashLookup(paths, id);
            if (path == NULL) {
                if (chr->source.data.file.path == NULL) {
                    /* neither the log output nor 'info chardev' had a
                     * pty path for this chardev, report an error
                     */
                    qemuReportError(VIR_ERR_INTERNAL_ERROR,
                                    _("no assigned pty for device %s"), id);
                    return -1;
                } else {
                    /* 'info chardev' had no pty path for this chardev,
                     * but the log output had, so we're fine
                     */
                    continue;
                }
            }

            VIR_FREE(chr->source.data.file.path);
            chr->source.data.file.path = strdup(path);

            if (chr->source.data.file.path == NULL) {
                virReportOOMError();
                return -1;
            }
        }
    }

    return 0;
}

1026 1027
static int
qemuProcessFindCharDevicePTYsMonitor(virDomainObjPtr vm,
C
Cole Robinson 已提交
1028
                                     virBitmapPtr qemuCaps,
1029 1030
                                     virHashTablePtr paths)
{
C
Cole Robinson 已提交
1031 1032
    bool chardevfmt = qemuCapsGet(qemuCaps, QEMU_CAPS_CHARDEV);

1033
    if (qemuProcessLookupPTYs(vm->def->serials, vm->def->nserials,
C
Cole Robinson 已提交
1034
                              paths, chardevfmt) < 0)
1035 1036 1037
        return -1;

    if (qemuProcessLookupPTYs(vm->def->parallels, vm->def->nparallels,
C
Cole Robinson 已提交
1038
                              paths, chardevfmt) < 0)
1039
        return -1;
1040

1041
    if (qemuProcessLookupPTYs(vm->def->channels, vm->def->nchannels,
C
Cole Robinson 已提交
1042
                              paths, chardevfmt) < 0)
1043 1044 1045
        return -1;

    if (vm->def->console &&
C
Cole Robinson 已提交
1046
        qemuProcessLookupPTYs(&vm->def->console, 1, paths, chardevfmt) < 0)
1047
        return -1;
1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093

    return 0;
}

static int
qemuProcessFindCharDevicePTYs(virDomainObjPtr vm,
                              const char *output,
                              int fd ATTRIBUTE_UNUSED)
{
    size_t offset = 0;
    int ret, i;

    /* The order in which QEMU prints out the PTY paths is
       the order in which it procsses its serial and parallel
       device args. This code must match that ordering.... */

    /* first comes the serial devices */
    for (i = 0 ; i < vm->def->nserials ; i++) {
        virDomainChrDefPtr chr = vm->def->serials[i];
        if (chr->source.type == VIR_DOMAIN_CHR_TYPE_PTY) {
            if ((ret = qemuProcessExtractTTYPath(output, &offset,
                                                 &chr->source.data.file.path)) != 0)
                return ret;
        }
    }

    /* then the parallel devices */
    for (i = 0 ; i < vm->def->nparallels ; i++) {
        virDomainChrDefPtr chr = vm->def->parallels[i];
        if (chr->source.type == VIR_DOMAIN_CHR_TYPE_PTY) {
            if ((ret = qemuProcessExtractTTYPath(output, &offset,
                                                 &chr->source.data.file.path)) != 0)
                return ret;
        }
    }

    /* then the channel devices */
    for (i = 0 ; i < vm->def->nchannels ; i++) {
        virDomainChrDefPtr chr = vm->def->channels[i];
        if (chr->source.type == VIR_DOMAIN_CHR_TYPE_PTY) {
            if ((ret = qemuProcessExtractTTYPath(output, &offset,
                                                 &chr->source.data.file.path)) != 0)
                return ret;
        }
    }

1094 1095 1096 1097 1098 1099 1100 1101 1102 1103
    if (vm->def->console) {
        virDomainChrDefPtr chr = vm->def->console;
        if (chr->source.type == VIR_DOMAIN_CHR_TYPE_PTY &&
            chr->targetType == VIR_DOMAIN_CHR_CONSOLE_TARGET_TYPE_VIRTIO) {
            if ((ret = qemuProcessExtractTTYPath(output, &offset,
                                                 &chr->source.data.file.path)) != 0)
                return ret;
        }
    }

1104 1105 1106
    return 0;
}

1107
static void qemuProcessFreePtyPath(void *payload, const void *name ATTRIBUTE_UNUSED)
1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127
{
    VIR_FREE(payload);
}

static void
qemuProcessReadLogFD(int logfd, char *buf, int maxlen, int off)
{
    int ret;
    char *tmpbuf = buf + off;

    ret = saferead(logfd, tmpbuf, maxlen - off - 1);
    if (ret < 0) {
        ret = 0;
    }

    tmpbuf[ret] = '\0';
}

static int
qemuProcessWaitForMonitor(struct qemud_driver* driver,
C
Cole Robinson 已提交
1128 1129 1130
                          virDomainObjPtr vm,
                          virBitmapPtr qemuCaps,
                          off_t pos)
1131
{
1132 1133
    char *buf;
    size_t buf_size = 4096; /* Plenty of space to get startup greeting */
1134 1135 1136
    int logfd;
    int ret = -1;
    virHashTablePtr paths = NULL;
1137
    qemuDomainObjPrivatePtr priv;
1138

1139
    if ((logfd = qemuDomainOpenLog(driver, vm, pos)) < 0)
1140 1141
        return -1;

1142 1143 1144 1145 1146 1147
    if (VIR_ALLOC_N(buf, buf_size) < 0) {
        virReportOOMError();
        return -1;
    }

    if (qemuProcessReadLogOutput(vm, logfd, buf, buf_size,
1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160
                                 qemuProcessFindCharDevicePTYs,
                                 "console", 30) < 0)
        goto closelog;

    VIR_DEBUG("Connect monitor to %p '%s'", vm, vm->def->name);
    if (qemuConnectMonitor(driver, vm) < 0) {
        goto cleanup;
    }

    /* Try to get the pty path mappings again via the monitor. This is much more
     * reliable if it's available.
     * Note that the monitor itself can be on a pty, so we still need to try the
     * log output method. */
1161
    paths = virHashCreate(0, qemuProcessFreePtyPath);
1162
    if (paths == NULL)
1163 1164
        goto cleanup;

1165
    priv = vm->privateData;
E
Eric Blake 已提交
1166
    qemuDomainObjEnterMonitorWithDriver(driver, vm);
1167 1168 1169 1170 1171
    ret = qemuMonitorGetPtyPaths(priv->mon, paths);
    qemuDomainObjExitMonitorWithDriver(driver, vm);

    VIR_DEBUG("qemuMonitorGetPtyPaths returned %i", ret);
    if (ret == 0)
C
Cole Robinson 已提交
1172
        ret = qemuProcessFindCharDevicePTYsMonitor(vm, qemuCaps, paths);
1173 1174

cleanup:
1175
    virHashFree(paths);
1176 1177 1178 1179

    if (kill(vm->pid, 0) == -1 && errno == ESRCH) {
        /* VM is dead, any other error raised in the interim is probably
         * not as important as the qemu cmdline output */
1180
        qemuProcessReadLogFD(logfd, buf, buf_size, strlen(buf));
1181 1182 1183 1184 1185 1186
        qemuReportError(VIR_ERR_INTERNAL_ERROR,
                        _("process exited while connecting to monitor: %s"),
                        buf);
        ret = -1;
    }

1187
closelog:
1188 1189
    VIR_FREE(buf);

1190
    if (VIR_CLOSE(logfd) < 0) {
1191
        char ebuf[1024];
1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 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
        VIR_WARN("Unable to close logfile: %s",
                 virStrerror(errno, ebuf, sizeof ebuf));
    }

    return ret;
}

static int
qemuProcessDetectVcpuPIDs(struct qemud_driver *driver,
                          virDomainObjPtr vm)
{
    pid_t *cpupids = NULL;
    int ncpupids;
    qemuDomainObjPrivatePtr priv = vm->privateData;

    if (vm->def->virtType != VIR_DOMAIN_VIRT_KVM) {
        priv->nvcpupids = 1;
        if (VIR_ALLOC_N(priv->vcpupids, priv->nvcpupids) < 0) {
            virReportOOMError();
            return -1;
        }
        priv->vcpupids[0] = vm->pid;
        return 0;
    }

    /* What follows is now all KVM specific */

    qemuDomainObjEnterMonitorWithDriver(driver, vm);
    if ((ncpupids = qemuMonitorGetCPUInfo(priv->mon, &cpupids)) < 0) {
        qemuDomainObjExitMonitorWithDriver(driver, vm);
        return -1;
    }
    qemuDomainObjExitMonitorWithDriver(driver, vm);

    /* Treat failure to get VCPU<->PID mapping as non-fatal */
    if (ncpupids == 0)
        return 0;

    if (ncpupids != vm->def->vcpus) {
        qemuReportError(VIR_ERR_INTERNAL_ERROR,
                        _("got wrong number of vCPU pids from QEMU monitor. "
                          "got %d, wanted %d"),
                        ncpupids, vm->def->vcpus);
        VIR_FREE(cpupids);
        return -1;
    }

    priv->nvcpupids = ncpupids;
    priv->vcpupids = cpupids;
    return 0;
}

1244 1245 1246 1247 1248 1249 1250 1251 1252

/*
 * Set NUMA memory policy for qemu process, to be run between
 * fork/exec of QEMU only.
 */
#if HAVE_NUMACTL
static int
qemuProcessInitNumaMemoryPolicy(virDomainObjPtr vm)
{
E
Eric Blake 已提交
1253
    nodemask_t mask;
1254 1255 1256 1257 1258
    int mode = -1;
    int node = -1;
    int ret = -1;
    int i = 0;
    int maxnode = 0;
E
Eric Blake 已提交
1259
    bool warned = false;
1260

1261 1262 1263
    if (!vm->def->numatune.memory.nodemask)
        return 0;

1264 1265 1266 1267 1268 1269 1270 1271
    VIR_DEBUG("Setting NUMA memory policy");

    if (numa_available() < 0) {
        qemuReportError(VIR_ERR_INTERNAL_ERROR,
                        "%s", _("Host kernel is not aware of NUMA."));
        return -1;
    }

E
Eric Blake 已提交
1272
    maxnode = numa_max_node() + 1;
1273 1274

    /* Convert nodemask to NUMA bitmask. */
E
Eric Blake 已提交
1275
    nodemask_zero(&mask);
1276 1277
    for (i = 0; i < VIR_DOMAIN_CPUMASK_LEN; i++) {
        if (vm->def->numatune.memory.nodemask[i]) {
E
Eric Blake 已提交
1278 1279 1280 1281 1282 1283
            if (i > NUMA_NUM_NODES) {
                qemuReportError(VIR_ERR_INTERNAL_ERROR,
                                _("Host cannot support NUMA node %d"), i);
                return -1;
            }
            if (i > maxnode && !warned) {
1284 1285
                VIR_WARN("nodeset is out of range, there is only %d NUMA "
                         "nodes on host", maxnode);
E
Eric Blake 已提交
1286
                warned = true;
1287
             }
E
Eric Blake 已提交
1288
            nodemask_set(&mask, i);
1289 1290 1291 1292 1293 1294 1295
        }
    }

    mode = vm->def->numatune.memory.mode;

    if (mode == VIR_DOMAIN_NUMATUNE_MEM_STRICT) {
        numa_set_bind_policy(1);
E
Eric Blake 已提交
1296
        numa_set_membind(&mask);
1297 1298 1299
        numa_set_bind_policy(0);
    } else if (mode == VIR_DOMAIN_NUMATUNE_MEM_PREFERRED) {
        int nnodes = 0;
E
Eric Blake 已提交
1300 1301
        for (i = 0; i < NUMA_NUM_NODES; i++) {
            if (nodemask_isset(&mask, i)) {
1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316
                node = i;
                nnodes++;
            }
        }

        if (nnodes != 1) {
            qemuReportError(VIR_ERR_INTERNAL_ERROR,
                            "%s", _("NUMA memory tuning in 'preferred' mode "
                                    "only supports single node"));
            goto cleanup;
        }

        numa_set_bind_policy(0);
        numa_set_preferred(node);
    } else if (mode == VIR_DOMAIN_NUMATUNE_MEM_INTERLEAVE) {
E
Eric Blake 已提交
1317
        numa_set_interleave_mask(&mask);
1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346
    } else {
        /* XXX: Shouldn't go here, as we already do checking when
         * parsing domain XML.
         */
        qemuReportError(VIR_ERR_XML_ERROR,
                        "%s", _("Invalid mode for memory NUMA tuning."));
        goto cleanup;
    }

    ret = 0;

cleanup:
    return ret;
}
#else
static int
qemuProcessInitNumaMemoryPolicy(virDomainObjPtr vm)
{
    if (vm->def->numatune.memory.nodemask) {
        qemuReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                        _("libvirt is compiled without NUMA tuning support"));

        return -1;
    }

    return 0;
}
#endif

1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357
/*
 * To be run between fork/exec of QEMU only
 */
static int
qemuProcessInitCpuAffinity(virDomainObjPtr vm)
{
    int i, hostcpus, maxcpu = QEMUD_CPUMASK_LEN;
    virNodeInfo nodeinfo;
    unsigned char *cpumap;
    int cpumaplen;

1358
    VIR_DEBUG("Setting CPU affinity");
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 1400 1401 1402 1403

    if (nodeGetInfo(NULL, &nodeinfo) < 0)
        return -1;

    /* setaffinity fails if you set bits for CPUs which
     * aren't present, so we have to limit ourselves */
    hostcpus = VIR_NODEINFO_MAXCPUS(nodeinfo);
    if (maxcpu > hostcpus)
        maxcpu = hostcpus;

    cpumaplen = VIR_CPU_MAPLEN(maxcpu);
    if (VIR_ALLOC_N(cpumap, cpumaplen) < 0) {
        virReportOOMError();
        return -1;
    }

    if (vm->def->cpumask) {
        /* XXX why don't we keep 'cpumask' in the libvirt cpumap
         * format to start with ?!?! */
        for (i = 0 ; i < maxcpu && i < vm->def->cpumasklen ; i++)
            if (vm->def->cpumask[i])
                VIR_USE_CPU(cpumap, i);
    } else {
        /* You may think this is redundant, but we can't assume libvirtd
         * itself is running on all pCPUs, so we need to explicitly set
         * the spawned QEMU instance to all pCPUs if no map is given in
         * its config file */
        for (i = 0 ; i < maxcpu ; i++)
            VIR_USE_CPU(cpumap, i);
    }

    /* We are pressuming we are running between fork/exec of QEMU
     * so use '0' to indicate our own process ID. No threads are
     * running at this point
     */
    if (virProcessInfoSetAffinity(0, /* Self */
                                  cpumap, cpumaplen, maxcpu) < 0) {
        VIR_FREE(cpumap);
        return -1;
    }
    VIR_FREE(cpumap);

    return 0;
}

1404 1405 1406 1407 1408 1409 1410 1411 1412 1413
/* Set CPU affinites for vcpus if vcpupin xml provided. */
static int
qemuProcessSetVcpuAffinites(virConnectPtr conn,
                            virDomainObjPtr vm)
{
    qemuDomainObjPrivatePtr priv = vm->privateData;
    virDomainDefPtr def = vm->def;
    virNodeInfo nodeinfo;
    pid_t vcpupid;
    unsigned char *cpumask;
1414
    int vcpu, cpumaplen, hostcpus, maxcpu, n;
E
Eric Blake 已提交
1415 1416
    unsigned char *cpumap = NULL;
    int ret = -1;
1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437

    if (virNodeGetInfo(conn, &nodeinfo) != 0) {
        return  -1;
    }

    if (!def->cputune.nvcpupin)
        return 0;

    if (priv->vcpupids == NULL) {
        qemuReportError(VIR_ERR_NO_SUPPORT,
                        "%s", _("cpu affinity is not supported"));
        return -1;
    }

    hostcpus = VIR_NODEINFO_MAXCPUS(nodeinfo);
    cpumaplen = VIR_CPU_MAPLEN(hostcpus);
    maxcpu = cpumaplen * 8;

    if (maxcpu > hostcpus)
        maxcpu = hostcpus;

E
Eric Blake 已提交
1438 1439 1440 1441 1442
    if (VIR_ALLOC_N(cpumap, cpumaplen) < 0) {
        virReportOOMError();
        return -1;
    }

1443
    for (n = 0; n < def->cputune.nvcpupin; n++) {
1444
        int i;
1445
        vcpu = def->cputune.vcpupin[n]->vcpuid;
1446

E
Eric Blake 已提交
1447
        memset(cpumap, 0, cpumaplen);
1448
        cpumask = (unsigned char *)def->cputune.vcpupin[n]->cpumask;
1449 1450
        vcpupid = priv->vcpupids[vcpu];

1451
        for (i = 0 ; i < VIR_DOMAIN_CPUMASK_LEN ; i++) {
1452
            if (cpumask[i])
1453
                VIR_USE_CPU(cpumap, i);
1454 1455 1456 1457 1458 1459
        }

        if (virProcessInfoSetAffinity(vcpupid,
                                      cpumap,
                                      cpumaplen,
                                      maxcpu) < 0) {
E
Eric Blake 已提交
1460
            goto cleanup;
1461 1462 1463
        }
    }

E
Eric Blake 已提交
1464 1465 1466 1467
    ret = 0;
cleanup:
    VIR_FREE(cpumap);
    return ret;
1468
}
1469 1470 1471 1472

static int
qemuProcessInitPasswords(virConnectPtr conn,
                         struct qemud_driver *driver,
1473
                         virDomainObjPtr vm)
1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494
{
    int ret = 0;
    qemuDomainObjPrivatePtr priv = vm->privateData;

    if (vm->def->ngraphics == 1) {
        if (vm->def->graphics[0]->type == VIR_DOMAIN_GRAPHICS_TYPE_VNC) {
            ret = qemuDomainChangeGraphicsPasswords(driver, vm,
                                                    VIR_DOMAIN_GRAPHICS_TYPE_VNC,
                                                    &vm->def->graphics[0]->data.vnc.auth,
                                                    driver->vncPassword);
        } else if (vm->def->graphics[0]->type == VIR_DOMAIN_GRAPHICS_TYPE_SPICE) {
            ret = qemuDomainChangeGraphicsPasswords(driver, vm,
                                                    VIR_DOMAIN_GRAPHICS_TYPE_SPICE,
                                                    &vm->def->graphics[0]->data.spice.auth,
                                                    driver->spicePassword);
        }
    }

    if (ret < 0)
        goto cleanup;

1495
    if (qemuCapsGet(priv->qemuCaps, QEMU_CAPS_DEVICE)) {
1496 1497 1498 1499 1500
        int i;

        for (i = 0 ; i < vm->def->ndisks ; i++) {
            char *secret;
            size_t secretLen;
E
Eric Blake 已提交
1501
            const char *alias;
1502 1503 1504 1505 1506 1507 1508 1509 1510 1511

            if (!vm->def->disks[i]->encryption ||
                !vm->def->disks[i]->src)
                continue;

            if (qemuProcessGetVolumeQcowPassphrase(conn,
                                                   vm->def->disks[i],
                                                   &secret, &secretLen) < 0)
                goto cleanup;

E
Eric Blake 已提交
1512
            alias = vm->def->disks[i]->info.alias;
1513
            qemuDomainObjEnterMonitorWithDriver(driver, vm);
E
Eric Blake 已提交
1514
            ret = qemuMonitorSetDrivePassphrase(priv->mon, alias, secret);
1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 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 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 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 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 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 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006
            VIR_FREE(secret);
            qemuDomainObjExitMonitorWithDriver(driver, vm);
            if (ret < 0)
                goto cleanup;
        }
    }

cleanup:
    return ret;
}


#define QEMU_PCI_VENDOR_INTEL     0x8086
#define QEMU_PCI_VENDOR_LSI_LOGIC 0x1000
#define QEMU_PCI_VENDOR_REDHAT    0x1af4
#define QEMU_PCI_VENDOR_CIRRUS    0x1013
#define QEMU_PCI_VENDOR_REALTEK   0x10ec
#define QEMU_PCI_VENDOR_AMD       0x1022
#define QEMU_PCI_VENDOR_ENSONIQ   0x1274
#define QEMU_PCI_VENDOR_VMWARE    0x15ad
#define QEMU_PCI_VENDOR_QEMU      0x1234

#define QEMU_PCI_PRODUCT_DISK_VIRTIO 0x1001

#define QEMU_PCI_PRODUCT_BALLOON_VIRTIO 0x1002

#define QEMU_PCI_PRODUCT_NIC_NE2K     0x8029
#define QEMU_PCI_PRODUCT_NIC_PCNET    0x2000
#define QEMU_PCI_PRODUCT_NIC_RTL8139  0x8139
#define QEMU_PCI_PRODUCT_NIC_E1000    0x100E
#define QEMU_PCI_PRODUCT_NIC_VIRTIO   0x1000

#define QEMU_PCI_PRODUCT_VGA_CIRRUS 0x00b8
#define QEMU_PCI_PRODUCT_VGA_VMWARE 0x0405
#define QEMU_PCI_PRODUCT_VGA_STDVGA 0x1111

#define QEMU_PCI_PRODUCT_AUDIO_AC97    0x2415
#define QEMU_PCI_PRODUCT_AUDIO_ES1370  0x5000

#define QEMU_PCI_PRODUCT_CONTROLLER_PIIX 0x7010
#define QEMU_PCI_PRODUCT_CONTROLLER_LSI  0x0012

#define QEMU_PCI_PRODUCT_WATCHDOG_I63000ESB 0x25ab

static int
qemuProcessAssignNextPCIAddress(virDomainDeviceInfo *info,
                                int vendor,
                                int product,
                                qemuMonitorPCIAddress *addrs,
                                int naddrs)
{
    int found = 0;
    int i;

    VIR_DEBUG("Look for %x:%x out of %d", vendor, product, naddrs);

    for (i = 0 ; (i < naddrs) && !found; i++) {
        VIR_DEBUG("Maybe %x:%x", addrs[i].vendor, addrs[i].product);
        if (addrs[i].vendor == vendor &&
            addrs[i].product == product) {
            VIR_DEBUG("Match %d", i);
            found = 1;
            break;
        }
    }
    if (!found) {
        return -1;
    }

    /* Blank it out so this device isn't matched again */
    addrs[i].vendor = 0;
    addrs[i].product = 0;

    if (info->type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_NONE)
        info->type = VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI;

    if (info->type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI) {
        info->addr.pci.domain = addrs[i].addr.domain;
        info->addr.pci.bus = addrs[i].addr.bus;
        info->addr.pci.slot = addrs[i].addr.slot;
        info->addr.pci.function = addrs[i].addr.function;
    }

    return 0;
}

static int
qemuProcessGetPCIDiskVendorProduct(virDomainDiskDefPtr def,
                                   unsigned *vendor,
                                   unsigned *product)
{
    switch (def->bus) {
    case VIR_DOMAIN_DISK_BUS_VIRTIO:
        *vendor = QEMU_PCI_VENDOR_REDHAT;
        *product = QEMU_PCI_PRODUCT_DISK_VIRTIO;
        break;

    default:
        return -1;
    }

    return 0;
}

static int
qemuProcessGetPCINetVendorProduct(virDomainNetDefPtr def,
                                  unsigned *vendor,
                                  unsigned *product)
{
    if (!def->model)
        return -1;

    if (STREQ(def->model, "ne2k_pci")) {
        *vendor = QEMU_PCI_VENDOR_REALTEK;
        *product = QEMU_PCI_PRODUCT_NIC_NE2K;
    } else if (STREQ(def->model, "pcnet")) {
        *vendor = QEMU_PCI_VENDOR_AMD;
        *product = QEMU_PCI_PRODUCT_NIC_PCNET;
    } else if (STREQ(def->model, "rtl8139")) {
        *vendor = QEMU_PCI_VENDOR_REALTEK;
        *product = QEMU_PCI_PRODUCT_NIC_RTL8139;
    } else if (STREQ(def->model, "e1000")) {
        *vendor = QEMU_PCI_VENDOR_INTEL;
        *product = QEMU_PCI_PRODUCT_NIC_E1000;
    } else if (STREQ(def->model, "virtio")) {
        *vendor = QEMU_PCI_VENDOR_REDHAT;
        *product = QEMU_PCI_PRODUCT_NIC_VIRTIO;
    } else {
        VIR_INFO("Unexpected NIC model %s, cannot get PCI address",
                 def->model);
        return -1;
    }
    return 0;
}

static int
qemuProcessGetPCIControllerVendorProduct(virDomainControllerDefPtr def,
                                         unsigned *vendor,
                                         unsigned *product)
{
    switch (def->type) {
    case VIR_DOMAIN_CONTROLLER_TYPE_SCSI:
        *vendor = QEMU_PCI_VENDOR_LSI_LOGIC;
        *product = QEMU_PCI_PRODUCT_CONTROLLER_LSI;
        break;

    case VIR_DOMAIN_CONTROLLER_TYPE_FDC:
        /* XXX we could put in the ISA bridge address, but
           that's not technically the FDC's address */
        return -1;

    case VIR_DOMAIN_CONTROLLER_TYPE_IDE:
        *vendor = QEMU_PCI_VENDOR_INTEL;
        *product = QEMU_PCI_PRODUCT_CONTROLLER_PIIX;
        break;

    default:
        VIR_INFO("Unexpected controller type %s, cannot get PCI address",
                 virDomainControllerTypeToString(def->type));
        return -1;
    }

    return 0;
}

static int
qemuProcessGetPCIVideoVendorProduct(virDomainVideoDefPtr def,
                                    unsigned *vendor,
                                    unsigned *product)
{
    switch (def->type) {
    case VIR_DOMAIN_VIDEO_TYPE_CIRRUS:
        *vendor = QEMU_PCI_VENDOR_CIRRUS;
        *product = QEMU_PCI_PRODUCT_VGA_CIRRUS;
        break;

    case VIR_DOMAIN_VIDEO_TYPE_VGA:
        *vendor = QEMU_PCI_VENDOR_QEMU;
        *product = QEMU_PCI_PRODUCT_VGA_STDVGA;
        break;

    case VIR_DOMAIN_VIDEO_TYPE_VMVGA:
        *vendor = QEMU_PCI_VENDOR_VMWARE;
        *product = QEMU_PCI_PRODUCT_VGA_VMWARE;
        break;

    default:
        return -1;
    }
    return 0;
}

static int
qemuProcessGetPCISoundVendorProduct(virDomainSoundDefPtr def,
                                    unsigned *vendor,
                                    unsigned *product)
{
    switch (def->model) {
    case VIR_DOMAIN_SOUND_MODEL_ES1370:
        *vendor = QEMU_PCI_VENDOR_ENSONIQ;
        *product = QEMU_PCI_PRODUCT_AUDIO_ES1370;
        break;

    case VIR_DOMAIN_SOUND_MODEL_AC97:
        *vendor = QEMU_PCI_VENDOR_INTEL;
        *product = QEMU_PCI_PRODUCT_AUDIO_AC97;
        break;

    default:
        return -1;
    }

    return 0;
}

static int
qemuProcessGetPCIWatchdogVendorProduct(virDomainWatchdogDefPtr def,
                                       unsigned *vendor,
                                       unsigned *product)
{
    switch (def->model) {
    case VIR_DOMAIN_WATCHDOG_MODEL_I6300ESB:
        *vendor = QEMU_PCI_VENDOR_INTEL;
        *product = QEMU_PCI_PRODUCT_WATCHDOG_I63000ESB;
        break;

    default:
        return -1;
    }

    return 0;
}


static int
qemuProcessGetPCIMemballoonVendorProduct(virDomainMemballoonDefPtr def,
                                         unsigned *vendor,
                                         unsigned *product)
{
    switch (def->model) {
    case VIR_DOMAIN_MEMBALLOON_MODEL_VIRTIO:
        *vendor = QEMU_PCI_VENDOR_REDHAT;
        *product = QEMU_PCI_PRODUCT_BALLOON_VIRTIO;
        break;

    default:
        return -1;
    }

    return 0;
}


/*
 * This entire method assumes that PCI devices in 'info pci'
 * match ordering of devices specified on the command line
 * wrt to devices of matching vendor+product
 *
 * XXXX this might not be a valid assumption if we assign
 * some static addrs on CLI. Have to check that...
 */
static int
qemuProcessDetectPCIAddresses(virDomainObjPtr vm,
                              qemuMonitorPCIAddress *addrs,
                              int naddrs)
{
    unsigned int vendor = 0, product = 0;
    int i;

    /* XXX should all these vendor/product IDs be kept in the
     * actual device data structure instead ?
     */

    for (i = 0 ; i < vm->def->ndisks ; i++) {
        if (qemuProcessGetPCIDiskVendorProduct(vm->def->disks[i], &vendor, &product) < 0)
            continue;

        if (qemuProcessAssignNextPCIAddress(&(vm->def->disks[i]->info),
                                            vendor, product,
                                            addrs, naddrs) < 0) {
            qemuReportError(VIR_ERR_INTERNAL_ERROR,
                            _("cannot find PCI address for VirtIO disk %s"),
                            vm->def->disks[i]->dst);
            return -1;
        }
    }

    for (i = 0 ; i < vm->def->nnets ; i++) {
        if (qemuProcessGetPCINetVendorProduct(vm->def->nets[i], &vendor, &product) < 0)
            continue;

        if (qemuProcessAssignNextPCIAddress(&(vm->def->nets[i]->info),
                                            vendor, product,
                                            addrs,  naddrs) < 0) {
            qemuReportError(VIR_ERR_INTERNAL_ERROR,
                            _("cannot find PCI address for %s NIC"),
                            vm->def->nets[i]->model);
            return -1;
        }
    }

    for (i = 0 ; i < vm->def->ncontrollers ; i++) {
        if (qemuProcessGetPCIControllerVendorProduct(vm->def->controllers[i], &vendor, &product) < 0)
            continue;

        if (qemuProcessAssignNextPCIAddress(&(vm->def->controllers[i]->info),
                                            vendor, product,
                                            addrs,  naddrs) < 0) {
            qemuReportError(VIR_ERR_INTERNAL_ERROR,
                            _("cannot find PCI address for controller %s"),
                            virDomainControllerTypeToString(vm->def->controllers[i]->type));
            return -1;
        }
    }

    for (i = 0 ; i < vm->def->nvideos ; i++) {
        if (qemuProcessGetPCIVideoVendorProduct(vm->def->videos[i], &vendor, &product) < 0)
            continue;

        if (qemuProcessAssignNextPCIAddress(&(vm->def->videos[i]->info),
                                            vendor, product,
                                            addrs,  naddrs) < 0) {
            qemuReportError(VIR_ERR_INTERNAL_ERROR,
                            _("cannot find PCI address for video adapter %s"),
                            virDomainVideoTypeToString(vm->def->videos[i]->type));
            return -1;
        }
    }

    for (i = 0 ; i < vm->def->nsounds ; i++) {
        if (qemuProcessGetPCISoundVendorProduct(vm->def->sounds[i], &vendor, &product) < 0)
            continue;

        if (qemuProcessAssignNextPCIAddress(&(vm->def->sounds[i]->info),
                                    vendor, product,
                                     addrs,  naddrs) < 0) {
            qemuReportError(VIR_ERR_INTERNAL_ERROR,
                            _("cannot find PCI address for sound adapter %s"),
                            virDomainSoundModelTypeToString(vm->def->sounds[i]->model));
            return -1;
        }
    }


    if (vm->def->watchdog &&
        qemuProcessGetPCIWatchdogVendorProduct(vm->def->watchdog, &vendor, &product) == 0) {
        if (qemuProcessAssignNextPCIAddress(&(vm->def->watchdog->info),
                                            vendor, product,
                                            addrs,  naddrs) < 0) {
            qemuReportError(VIR_ERR_INTERNAL_ERROR,
                            _("cannot find PCI address for watchdog %s"),
                            virDomainWatchdogModelTypeToString(vm->def->watchdog->model));
            return -1;
        }
    }

    if (vm->def->memballoon &&
        qemuProcessGetPCIMemballoonVendorProduct(vm->def->memballoon, &vendor, &product) == 0) {
        if (qemuProcessAssignNextPCIAddress(&(vm->def->memballoon->info),
                                            vendor, product,
                                            addrs, naddrs) < 0) {
            qemuReportError(VIR_ERR_INTERNAL_ERROR,
                            _("cannot find PCI address for balloon %s"),
                            virDomainMemballoonModelTypeToString(vm->def->memballoon->model));
            return -1;
        }
    }

    /* XXX console (virtio) */


    /* ... and now things we don't have in our xml */

    /* XXX USB controller ? */

    /* XXX what about other PCI devices (ie bridges) */

    return 0;
}

static int
qemuProcessInitPCIAddresses(struct qemud_driver *driver,
                            virDomainObjPtr vm)
{
    qemuDomainObjPrivatePtr priv = vm->privateData;
    int naddrs;
    int ret;
    qemuMonitorPCIAddress *addrs = NULL;

    qemuDomainObjEnterMonitorWithDriver(driver, vm);
    naddrs = qemuMonitorGetAllPCIAddresses(priv->mon,
                                           &addrs);
    qemuDomainObjExitMonitorWithDriver(driver, vm);

    ret = qemuProcessDetectPCIAddresses(vm, addrs, naddrs);

    VIR_FREE(addrs);

    return ret;
}


static int qemuProcessNextFreePort(struct qemud_driver *driver,
                                   int startPort)
{
    int i;

    for (i = startPort ; i < QEMU_VNC_PORT_MAX; i++) {
        int fd;
        int reuse = 1;
        struct sockaddr_in addr;
        bool used = false;

        if (virBitmapGetBit(driver->reservedVNCPorts,
                            i - QEMU_VNC_PORT_MIN, &used) < 0)
            VIR_DEBUG("virBitmapGetBit failed on bit %d", i - QEMU_VNC_PORT_MIN);

        if (used)
            continue;

        addr.sin_family = AF_INET;
        addr.sin_port = htons(i);
        addr.sin_addr.s_addr = htonl(INADDR_ANY);
        fd = socket(PF_INET, SOCK_STREAM, 0);
        if (fd < 0)
            return -1;

        if (setsockopt(fd, SOL_SOCKET, SO_REUSEADDR, (void*)&reuse, sizeof(reuse)) < 0) {
            VIR_FORCE_CLOSE(fd);
            break;
        }

        if (bind(fd, (struct sockaddr*)&addr, sizeof(addr)) == 0) {
            /* Not in use, lets grab it */
            VIR_FORCE_CLOSE(fd);
            /* Add port to bitmap of reserved ports */
            if (virBitmapSetBit(driver->reservedVNCPorts,
                                i - QEMU_VNC_PORT_MIN) < 0) {
                VIR_DEBUG("virBitmapSetBit failed on bit %d",
                          i - QEMU_VNC_PORT_MIN);
            }
            return i;
        }
        VIR_FORCE_CLOSE(fd);

        if (errno == EADDRINUSE) {
            /* In use, try next */
            continue;
        }
        /* Some other bad failure, get out.. */
        break;
    }
    return -1;
}


static void
qemuProcessReturnPort(struct qemud_driver *driver,
                      int port)
{
    if (port < QEMU_VNC_PORT_MIN)
        return;

    if (virBitmapClearBit(driver->reservedVNCPorts,
                          port - QEMU_VNC_PORT_MIN) < 0)
        VIR_DEBUG("Could not mark port %d as unused", port);
}


static int
qemuProcessPrepareChardevDevice(virDomainDefPtr def ATTRIBUTE_UNUSED,
                                virDomainChrDefPtr dev,
                                void *opaque ATTRIBUTE_UNUSED)
{
    int fd;
    if (dev->source.type != VIR_DOMAIN_CHR_TYPE_FILE)
        return 0;

    if ((fd = open(dev->source.data.file.path,
                   O_CREAT | O_APPEND, S_IRUSR|S_IWUSR)) < 0) {
        virReportSystemError(errno,
                             _("Unable to pre-create chardev file '%s'"),
                             dev->source.data.file.path);
        return -1;
    }

    VIR_FORCE_CLOSE(fd);

    return 0;
}


2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025
static int
qemuProcessLimits(struct qemud_driver *driver)
{
    if (driver->maxProcesses > 0) {
        struct rlimit rlim;

        rlim.rlim_cur = rlim.rlim_max = driver->maxProcesses;
        if (setrlimit(RLIMIT_NPROC, &rlim) < 0) {
            virReportSystemError(errno,
                                 _("cannot limit number of processes to %d"),
                                 driver->maxProcesses);
            return -1;
        }
    }

    return 0;
}


2026 2027 2028 2029 2030 2031 2032 2033 2034
struct qemuProcessHookData {
    virConnectPtr conn;
    virDomainObjPtr vm;
    struct qemud_driver *driver;
};

static int qemuProcessHook(void *data)
{
    struct qemuProcessHookData *h = data;
2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045
    int ret = -1;

    /* Some later calls want pid present */
    h->vm->pid = getpid();

    VIR_DEBUG("Obtaining domain lock");
    if (virDomainLockProcessStart(h->driver->lockManager,
                                  h->vm,
                                  /* QEMU is always pased initially */
                                  true) < 0)
        goto cleanup;
2046

2047 2048 2049
    if (qemuProcessLimits(h->driver) < 0)
        return -1;

2050 2051 2052
    /* This must take place before exec(), so that all QEMU
     * memory allocation is on the correct NUMA node
     */
2053
    VIR_DEBUG("Moving procss to cgroup");
2054
    if (qemuAddToCgroup(h->driver, h->vm->def) < 0)
2055
        goto cleanup;
2056 2057 2058

    /* This must be done after cgroup placement to avoid resetting CPU
     * affinity */
2059
    VIR_DEBUG("Setup CPU affinity");
2060
    if (qemuProcessInitCpuAffinity(h->vm) < 0)
2061
        goto cleanup;
2062

2063 2064 2065
    if (qemuProcessInitNumaMemoryPolicy(h->vm) < 0)
        return -1;

2066
    VIR_DEBUG("Setting up security labeling");
2067
    if (virSecurityManagerSetProcessLabel(h->driver->securityManager, h->vm) < 0)
2068
        goto cleanup;
2069

2070 2071 2072 2073 2074
    ret = 0;

cleanup:
    VIR_DEBUG("Hook complete ret=%d", ret);
    return ret;
2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094
}

int
qemuProcessPrepareMonitorChr(struct qemud_driver *driver,
                             virDomainChrSourceDefPtr monConfig,
                             const char *vm)
{
    monConfig->type = VIR_DOMAIN_CHR_TYPE_UNIX;
    monConfig->data.nix.listen = true;

    if (virAsprintf(&monConfig->data.nix.path, "%s/%s.monitor",
                    driver->libDir, vm) < 0) {
        virReportOOMError();
        return -1;
    }

    return 0;
}


2095 2096 2097 2098 2099
/*
 * Precondition: Both driver and vm must be locked,
 * and a job must be active. This method will call
 * {Enter,Exit}MonitorWithDriver
 */
E
Eric Blake 已提交
2100 2101
int
qemuProcessStartCPUs(struct qemud_driver *driver, virDomainObjPtr vm,
J
Jiri Denemark 已提交
2102
                     virConnectPtr conn, virDomainRunningReason reason)
2103 2104 2105 2106
{
    int ret;
    qemuDomainObjPrivatePtr priv = vm->privateData;

2107 2108 2109 2110 2111 2112 2113 2114 2115 2116
    VIR_DEBUG("Using lock state '%s'", NULLSTR(priv->lockState));
    if (virDomainLockProcessResume(driver->lockManager, vm, priv->lockState) < 0) {
        /* Don't free priv->lockState on error, because we need
         * to make sure we have state still present if the user
         * tries to resume again
         */
        return -1;
    }
    VIR_FREE(priv->lockState);

2117 2118
    qemuDomainObjEnterMonitorWithDriver(driver, vm);
    ret = qemuMonitorStartCPUs(priv->mon, conn);
E
Eric Blake 已提交
2119
    qemuDomainObjExitMonitorWithDriver(driver, vm);
J
Jiri Denemark 已提交
2120

2121
    if (ret == 0) {
J
Jiri Denemark 已提交
2122
        virDomainObjSetState(vm, VIR_DOMAIN_RUNNING, reason);
2123 2124 2125 2126 2127
    } else {
        if (virDomainLockProcessPause(driver->lockManager, vm, &priv->lockState) < 0)
            VIR_WARN("Unable to release lease on %s", vm->def->name);
        VIR_DEBUG("Preserving lock state '%s'", NULLSTR(priv->lockState));
    }
2128 2129 2130 2131 2132

    return ret;
}


J
Jiri Denemark 已提交
2133 2134
int qemuProcessStopCPUs(struct qemud_driver *driver, virDomainObjPtr vm,
                        virDomainPausedReason reason)
2135 2136
{
    int ret;
J
Jiri Denemark 已提交
2137 2138
    int oldState;
    int oldReason;
2139 2140
    qemuDomainObjPrivatePtr priv = vm->privateData;

2141
    VIR_FREE(priv->lockState);
J
Jiri Denemark 已提交
2142 2143 2144
    oldState = virDomainObjGetState(vm, &oldReason);
    virDomainObjSetState(vm, VIR_DOMAIN_PAUSED, reason);

2145 2146 2147
    qemuDomainObjEnterMonitorWithDriver(driver, vm);
    ret = qemuMonitorStopCPUs(priv->mon);
    qemuDomainObjExitMonitorWithDriver(driver, vm);
J
Jiri Denemark 已提交
2148

2149 2150 2151 2152 2153
    if (ret == 0) {
        if (virDomainLockProcessPause(driver->lockManager, vm, &priv->lockState) < 0)
            VIR_WARN("Unable to release lease on %s", vm->def->name);
        VIR_DEBUG("Preserving lock state '%s'", NULLSTR(priv->lockState));
    } else {
J
Jiri Denemark 已提交
2154
        virDomainObjSetState(vm, oldState, oldReason);
2155
    }
J
Jiri Denemark 已提交
2156

2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184
    return ret;
}



static int
qemuProcessFiltersInstantiate(virConnectPtr conn,
                              virDomainDefPtr def)
{
    int err = 0;
    int i;

    if (!conn)
        return 1;

    for (i = 0 ; i < def->nnets ; i++) {
        virDomainNetDefPtr net = def->nets[i];
        if ((net->filter) && (net->ifname)) {
           if (virDomainConfNWFilterInstantiate(conn, net)) {
                err = 1;
                break;
            }
        }
    }

    return err;
}

2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215
static int
qemuProcessUpdateState(struct qemud_driver *driver, virDomainObjPtr vm)
{
    qemuDomainObjPrivatePtr priv = vm->privateData;
    virDomainState state;
    bool running;
    int ret;

    qemuDomainObjEnterMonitorWithDriver(driver, vm);
    ret = qemuMonitorGetStatus(priv->mon, &running);
    qemuDomainObjExitMonitorWithDriver(driver, vm);

    if (ret < 0 || !virDomainObjIsActive(vm))
        return -1;

    state = virDomainObjGetState(vm, NULL);

    if (state == VIR_DOMAIN_PAUSED && running) {
        VIR_DEBUG("Domain %s was unpaused while its monitor was disconnected;"
                  " changing state to running", vm->def->name);
        virDomainObjSetState(vm, VIR_DOMAIN_RUNNING,
                             VIR_DOMAIN_RUNNING_UNPAUSED);
    } else if (state == VIR_DOMAIN_RUNNING && !running) {
        VIR_DEBUG("Domain %s was paused while its monitor was disconnected;"
                  " changing state to paused", vm->def->name);
        virDomainObjSetState(vm, VIR_DOMAIN_PAUSED, VIR_DOMAIN_PAUSED_UNKNOWN);
    }

    return 0;
}

2216 2217 2218 2219 2220 2221 2222 2223 2224
struct qemuProcessReconnectData {
    virConnectPtr conn;
    struct qemud_driver *driver;
};
/*
 * Open an existing VM's monitor, re-detect VCPU threads
 * and re-reserve the security labels in use
 */
static void
2225
qemuProcessReconnect(void *payload, const void *name ATTRIBUTE_UNUSED, void *opaque)
2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250
{
    virDomainObjPtr obj = payload;
    struct qemuProcessReconnectData *data = opaque;
    struct qemud_driver *driver = data->driver;
    qemuDomainObjPrivatePtr priv;
    virConnectPtr conn = data->conn;

    virDomainObjLock(obj);

    VIR_DEBUG("Reconnect monitor to %p '%s'", obj, obj->def->name);

    priv = obj->privateData;

    /* Hold an extra reference because we can't allow 'vm' to be
     * deleted if qemuConnectMonitor() failed */
    virDomainObjRef(obj);

    /* XXX check PID liveliness & EXE path */
    if (qemuConnectMonitor(driver, obj) < 0)
        goto error;

    if (qemuUpdateActivePciHostdevs(driver, obj->def) < 0) {
        goto error;
    }

2251 2252 2253
    if (qemuProcessUpdateState(driver, obj) < 0)
        goto error;

2254 2255 2256 2257 2258
    /* If upgrading from old libvirtd we won't have found any
     * caps in the domain status, so re-query them
     */
    if (!priv->qemuCaps &&
        qemuCapsExtractVersionInfo(obj->def->emulator, obj->def->os.arch,
2259
                                   NULL,
2260 2261 2262 2263
                                   &priv->qemuCaps) < 0)
        goto error;

    if (qemuCapsGet(priv->qemuCaps, QEMU_CAPS_DEVICE)) {
2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276
        priv->persistentAddrs = 1;

        if (!(priv->pciaddrs = qemuDomainPCIAddressSetCreate(obj->def)) ||
            qemuAssignDevicePCISlots(obj->def, priv->pciaddrs) < 0)
            goto error;
    }

    if (virSecurityManagerReserveLabel(driver->securityManager, obj) < 0)
        goto error;

    if (qemuProcessFiltersInstantiate(conn, obj->def))
        goto error;

2277 2278 2279 2280
    /* update domain state XML with possibly updated state in virDomainObj */
    if (virDomainSaveStatus(driver->caps, driver->stateDir, obj) < 0)
        goto error;

2281 2282 2283 2284 2285
    if (obj->def->id >= driver->nextvmid)
        driver->nextvmid = obj->def->id + 1;

    if (virDomainObjUnref(obj) > 0)
        virDomainObjUnlock(obj);
2286

2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299
    return;

error:
    if (!virDomainObjIsActive(obj)) {
        if (virDomainObjUnref(obj) > 0)
            virDomainObjUnlock(obj);
        return;
    }

    if (virDomainObjUnref(obj) > 0) {
        /* We can't get the monitor back, so must kill the VM
         * to remove danger of it ending up running twice if
         * user tries to start it again later */
J
Jiri Denemark 已提交
2300
        qemuProcessStop(driver, obj, 0, VIR_DOMAIN_SHUTOFF_FAILED);
2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325
        if (!obj->persistent)
            virDomainRemoveInactive(&driver->domains, obj);
        else
            virDomainObjUnlock(obj);
    }
}

/**
 * qemuProcessReconnectAll
 *
 * Try to re-open the resources for live VMs that we care
 * about.
 */
void
qemuProcessReconnectAll(virConnectPtr conn, struct qemud_driver *driver)
{
    struct qemuProcessReconnectData data = {conn, driver};
    virHashForEach(driver->domains.objs, qemuProcessReconnect, &data);
}

int qemuProcessStart(virConnectPtr conn,
                     struct qemud_driver *driver,
                     virDomainObjPtr vm,
                     const char *migrateFrom,
                     bool start_paused,
2326
                     bool autodestroy,
2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339
                     int stdin_fd,
                     const char *stdin_path,
                     enum virVMOperationType vmop)
{
    int ret;
    off_t pos = -1;
    char ebuf[1024];
    char *pidfile = NULL;
    int logfile = -1;
    char *timestamp;
    qemuDomainObjPrivatePtr priv = vm->privateData;
    virCommandPtr cmd = NULL;
    struct qemuProcessHookData hookData;
E
Eric Blake 已提交
2340
    unsigned long cur_balloon;
2341 2342 2343 2344 2345

    hookData.conn = conn;
    hookData.vm = vm;
    hookData.driver = driver;

2346
    VIR_DEBUG("Beginning VM startup process");
2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357

    if (virDomainObjIsActive(vm)) {
        qemuReportError(VIR_ERR_OPERATION_INVALID,
                        "%s", _("VM is already active"));
        return -1;
    }

    /* Do this upfront, so any part of the startup process can add
     * runtime state to vm->def that won't be persisted. This let's us
     * report implicit runtime defaults in the XML, like vnc listen/socket
     */
2358
    VIR_DEBUG("Setting current domain def as transient");
2359 2360 2361 2362
    if (virDomainObjSetDefTransient(driver->caps, vm, true) < 0)
        goto cleanup;

    vm->def->id = driver->nextvmid++;
2363
    priv->fakeReboot = false;
2364

2365
    /* Run an early hook to set-up missing devices */
2366 2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380
    if (virHookPresent(VIR_HOOK_DRIVER_QEMU)) {
        char *xml = virDomainDefFormat(vm->def, 0);
        int hookret;

        hookret = virHookCall(VIR_HOOK_DRIVER_QEMU, vm->def->name,
                    VIR_HOOK_QEMU_OP_PREPARE, VIR_HOOK_SUBOP_BEGIN, NULL, xml);
        VIR_FREE(xml);

        /*
         * If the script raised an error abort the launch
         */
        if (hookret < 0)
            goto cleanup;
    }

2381
    /* Must be run before security labelling */
2382
    VIR_DEBUG("Preparing host devices");
2383 2384 2385
    if (qemuPrepareHostDevices(driver, vm->def) < 0)
        goto cleanup;

2386
    VIR_DEBUG("Preparing chr devices");
2387 2388 2389 2390 2391 2392 2393 2394
    if (virDomainChrDefForeach(vm->def,
                               true,
                               qemuProcessPrepareChardevDevice,
                               NULL) < 0)
        goto cleanup;

    /* If you are using a SecurityDriver with dynamic labelling,
       then generate a security label for isolation */
2395
    VIR_DEBUG("Generating domain security label (if required)");
2396
    if (virSecurityManagerGenLabel(driver->securityManager, vm) < 0) {
2397
        qemuAuditSecurityLabel(vm, false);
2398 2399
        goto cleanup;
    }
2400
    qemuAuditSecurityLabel(vm, true);
2401 2402 2403

    /* Ensure no historical cgroup for this VM is lying around bogus
     * settings */
2404
    VIR_DEBUG("Ensuring no historical cgroup is lying around");
2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449
    qemuRemoveCgroup(driver, vm, 1);

    if (vm->def->ngraphics == 1) {
        if (vm->def->graphics[0]->type == VIR_DOMAIN_GRAPHICS_TYPE_VNC &&
            !vm->def->graphics[0]->data.vnc.socket &&
            vm->def->graphics[0]->data.vnc.autoport) {
            int port = qemuProcessNextFreePort(driver, QEMU_VNC_PORT_MIN);
            if (port < 0) {
                qemuReportError(VIR_ERR_INTERNAL_ERROR,
                                "%s", _("Unable to find an unused VNC port"));
                goto cleanup;
            }
            vm->def->graphics[0]->data.vnc.port = port;
        } else if (vm->def->graphics[0]->type == VIR_DOMAIN_GRAPHICS_TYPE_SPICE &&
                   vm->def->graphics[0]->data.spice.autoport) {
            int port = qemuProcessNextFreePort(driver, QEMU_VNC_PORT_MIN);
            int tlsPort = -1;
            if (port < 0) {
                qemuReportError(VIR_ERR_INTERNAL_ERROR,
                                "%s", _("Unable to find an unused SPICE port"));
                goto cleanup;
            }

            if (driver->spiceTLS) {
                tlsPort = qemuProcessNextFreePort(driver, port + 1);
                if (tlsPort < 0) {
                    qemuReportError(VIR_ERR_INTERNAL_ERROR,
                                    "%s", _("Unable to find an unused SPICE TLS port"));
                    qemuProcessReturnPort(driver, port);
                    goto cleanup;
                }
            }

            vm->def->graphics[0]->data.spice.port = port;
            vm->def->graphics[0]->data.spice.tlsPort = tlsPort;
        }
    }

    if (virFileMakePath(driver->logDir) != 0) {
        virReportSystemError(errno,
                             _("cannot create log directory %s"),
                             driver->logDir);
        goto cleanup;
    }

2450
    VIR_DEBUG("Creating domain log file");
2451
    if ((logfile = qemuDomainCreateLog(driver, vm, false)) < 0)
2452 2453
        goto cleanup;

2454
    VIR_DEBUG("Determining emulator version");
2455 2456
    qemuCapsFree(priv->qemuCaps);
    priv->qemuCaps = NULL;
2457 2458
    if (qemuCapsExtractVersionInfo(vm->def->emulator, vm->def->os.arch,
                                   NULL,
2459
                                   &priv->qemuCaps) < 0)
2460 2461
        goto cleanup;

2462
    VIR_DEBUG("Setting up domain cgroup (if required)");
2463 2464 2465 2466 2467 2468 2469 2470
    if (qemuSetupCgroup(driver, vm) < 0)
        goto cleanup;

    if (VIR_ALLOC(priv->monConfig) < 0) {
        virReportOOMError();
        goto cleanup;
    }

2471
    VIR_DEBUG("Preparing monitor state");
2472 2473 2474 2475
    if (qemuProcessPrepareMonitorChr(driver, priv->monConfig, vm->def->name) < 0)
        goto cleanup;

#if HAVE_YAJL
2476
    if (qemuCapsGet(priv->qemuCaps, QEMU_CAPS_MONITOR_JSON))
2477 2478 2479 2480 2481
        priv->monJSON = 1;
    else
#endif
        priv->monJSON = 0;

2482 2483
    priv->monError = false;
    priv->monStart = 0;
2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 2504 2505
    priv->gotShutdown = false;

    if ((ret = virFileDeletePid(driver->stateDir, vm->def->name)) != 0) {
        virReportSystemError(ret,
                             _("Cannot remove stale PID file for %s"),
                             vm->def->name);
        goto cleanup;
    }

    if (!(pidfile = virFilePid(driver->stateDir, vm->def->name))) {
        virReportSystemError(errno,
                             "%s", _("Failed to build pidfile path."));
        goto cleanup;
    }

    /*
     * Normally PCI addresses are assigned in the virDomainCreate
     * or virDomainDefine methods. We might still need to assign
     * some here to cope with the question of upgrades. Regardless
     * we also need to populate the PCi address set cache for later
     * use in hotplug
     */
2506
    if (qemuCapsGet(priv->qemuCaps, QEMU_CAPS_DEVICE)) {
2507
        VIR_DEBUG("Assigning domain PCI addresses");
2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525
        /* Populate cache with current addresses */
        if (priv->pciaddrs) {
            qemuDomainPCIAddressSetFree(priv->pciaddrs);
            priv->pciaddrs = NULL;
        }
        if (!(priv->pciaddrs = qemuDomainPCIAddressSetCreate(vm->def)))
            goto cleanup;


        /* Assign any remaining addresses */
        if (qemuAssignDevicePCISlots(vm->def, priv->pciaddrs) < 0)
            goto cleanup;

        priv->persistentAddrs = 1;
    } else {
        priv->persistentAddrs = 0;
    }

2526
    VIR_DEBUG("Building emulator command line");
2527
    if (!(cmd = qemuBuildCommandLine(conn, driver, vm->def, priv->monConfig,
2528
                                     priv->monJSON != 0, priv->qemuCaps,
2529 2530 2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 2552 2553 2554 2555 2556 2557 2558 2559 2560 2561 2562 2563 2564 2565 2566 2567 2568 2569
                                     migrateFrom, stdin_fd,
                                     vm->current_snapshot, vmop)))
        goto cleanup;

#if 0
    /* XXX */
    if (qemuDomainSnapshotSetCurrentInactive(vm, driver->snapshotDir) < 0)
        goto cleanup;
#endif

    /* now that we know it is about to start call the hook if present */
    if (virHookPresent(VIR_HOOK_DRIVER_QEMU)) {
        char *xml = virDomainDefFormat(vm->def, 0);
        int hookret;

        hookret = virHookCall(VIR_HOOK_DRIVER_QEMU, vm->def->name,
                    VIR_HOOK_QEMU_OP_START, VIR_HOOK_SUBOP_BEGIN, NULL, xml);
        VIR_FREE(xml);

        /*
         * If the script raised an error abort the launch
         */
        if (hookret < 0)
            goto cleanup;
    }

    if ((timestamp = virTimestamp()) == NULL) {
        virReportOOMError();
        goto cleanup;
    } else {
        if (safewrite(logfile, timestamp, strlen(timestamp)) < 0 ||
            safewrite(logfile, START_POSTFIX, strlen(START_POSTFIX)) < 0) {
            VIR_WARN("Unable to write timestamp to logfile: %s",
                     virStrerror(errno, ebuf, sizeof ebuf));
        }

        VIR_FREE(timestamp);
    }

    virCommandWriteArgLog(cmd, logfile);

2570
    qemuDomainObjCheckTaint(driver, vm, logfile);
2571

2572 2573 2574 2575 2576 2577 2578 2579 2580 2581 2582 2583 2584 2585 2586 2587
    if ((pos = lseek(logfile, 0, SEEK_END)) < 0)
        VIR_WARN("Unable to seek to end of logfile: %s",
                 virStrerror(errno, ebuf, sizeof ebuf));

    VIR_DEBUG("Clear emulator capabilities: %d",
              driver->clearEmulatorCapabilities);
    if (driver->clearEmulatorCapabilities)
        virCommandClearCaps(cmd);

    virCommandSetPreExecHook(cmd, qemuProcessHook, &hookData);

    virCommandSetOutputFD(cmd, &logfile);
    virCommandSetErrorFD(cmd, &logfile);
    virCommandNonblockingFDs(cmd);
    virCommandSetPidFile(cmd, pidfile);
    virCommandDaemonize(cmd);
2588
    virCommandRequireHandshake(cmd);
2589 2590 2591 2592

    ret = virCommandRun(cmd, NULL);
    VIR_FREE(pidfile);

E
Eric Blake 已提交
2593
    /* wait for qemu process to show up */
2594 2595 2596 2597 2598 2599 2600 2601 2602 2603 2604 2605 2606
    if (ret == 0) {
        if (virFileReadPid(driver->stateDir, vm->def->name, &vm->pid)) {
            qemuReportError(VIR_ERR_INTERNAL_ERROR,
                            _("Domain %s didn't show up"), vm->def->name);
            ret = -1;
        }
#if 0
    } else if (ret == -2) {
        /*
         * XXX this is bogus. It isn't safe to set vm->pid = child
         * because the child no longer exists.
         */

2607
        /* The virCommand process that launches the daemon failed. Pending on
2608 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618
         * when it failed (we can't determine for sure), there may be
         * extra info in the domain log (if the hook failed for example).
         *
         * Pretend like things succeeded, and let 'WaitForMonitor' report
         * the log contents for us.
         */
        vm->pid = child;
        ret = 0;
#endif
    }

2619 2620 2621 2622 2623 2624 2625 2626 2627 2628 2629 2630 2631 2632 2633 2634 2635 2636 2637 2638 2639 2640 2641 2642 2643
    VIR_DEBUG("Waiting for handshake from child");
    if (virCommandHandshakeWait(cmd) < 0) {
        ret = -1;
        goto cleanup;
    }

    VIR_DEBUG("Setting domain security labels");
    if (virSecurityManagerSetAllLabel(driver->securityManager,
                                      vm, stdin_path) < 0)
        goto cleanup;

    if (stdin_fd != -1) {
        /* if there's an fd to migrate from, and it's a pipe, put the
         * proper security label on it
         */
        struct stat stdin_sb;

        VIR_DEBUG("setting security label on pipe used for migration");

        if (fstat(stdin_fd, &stdin_sb) < 0) {
            virReportSystemError(errno,
                                 _("cannot stat fd %d"), stdin_fd);
            goto cleanup;
        }
        if (S_ISFIFO(stdin_sb.st_mode) &&
2644
            virSecurityManagerSetImageFDLabel(driver->securityManager, vm, stdin_fd) < 0)
2645 2646 2647 2648 2649 2650 2651 2652 2653 2654
            goto cleanup;
    }

    VIR_DEBUG("Labelling done, completing handshake to child");
    if (virCommandHandshakeNotify(cmd) < 0) {
        ret = -1;
        goto cleanup;
    }
    VIR_DEBUG("Handshake complete, child running");

2655 2656 2657 2658 2659 2660 2661 2662 2663
    if (migrateFrom)
        start_paused = true;

    if (ret == -1) /* The VM failed to start; tear filters before taps */
        virDomainConfVMNWFilterTeardown(vm);

    if (ret == -1) /* The VM failed to start */
        goto cleanup;

2664
    VIR_DEBUG("Waiting for monitor to show up");
C
Cole Robinson 已提交
2665
    if (qemuProcessWaitForMonitor(driver, vm, priv->qemuCaps, pos) < 0)
2666 2667
        goto cleanup;

2668
    VIR_DEBUG("Detecting VCPU PIDs");
2669 2670 2671
    if (qemuProcessDetectVcpuPIDs(driver, vm) < 0)
        goto cleanup;

2672
    VIR_DEBUG("Setting VCPU affinities");
2673 2674 2675
    if (qemuProcessSetVcpuAffinites(conn, vm) < 0)
        goto cleanup;

2676
    VIR_DEBUG("Setting any required VM passwords");
2677
    if (qemuProcessInitPasswords(conn, driver, vm) < 0)
2678 2679 2680 2681
        goto cleanup;

    /* If we have -device, then addresses are assigned explicitly.
     * If not, then we have to detect dynamic ones here */
2682
    if (!qemuCapsGet(priv->qemuCaps, QEMU_CAPS_DEVICE)) {
2683
        VIR_DEBUG("Determining domain device PCI addresses");
2684 2685 2686 2687
        if (qemuProcessInitPCIAddresses(driver, vm) < 0)
            goto cleanup;
    }

2688
    VIR_DEBUG("Setting initial memory amount");
E
Eric Blake 已提交
2689
    cur_balloon = vm->def->mem.cur_balloon;
2690
    qemuDomainObjEnterMonitorWithDriver(driver, vm);
E
Eric Blake 已提交
2691
    if (qemuMonitorSetBalloon(priv->mon, cur_balloon) < 0) {
2692 2693 2694 2695 2696 2697
        qemuDomainObjExitMonitorWithDriver(driver, vm);
        goto cleanup;
    }
    qemuDomainObjExitMonitorWithDriver(driver, vm);

    if (!start_paused) {
2698
        VIR_DEBUG("Starting domain CPUs");
2699
        /* Allow the CPUS to start executing */
J
Jiri Denemark 已提交
2700 2701
        if (qemuProcessStartCPUs(driver, vm, conn,
                                 VIR_DOMAIN_RUNNING_BOOTED) < 0) {
2702 2703 2704 2705 2706
            if (virGetLastError() == NULL)
                qemuReportError(VIR_ERR_INTERNAL_ERROR,
                                "%s", _("resume operation failed"));
            goto cleanup;
        }
J
Jiri Denemark 已提交
2707 2708 2709 2710 2711
    } else {
        virDomainObjSetState(vm, VIR_DOMAIN_PAUSED,
                             migrateFrom ?
                             VIR_DOMAIN_PAUSED_MIGRATION :
                             VIR_DOMAIN_PAUSED_USER);
2712 2713
    }

2714 2715 2716 2717
    if (autodestroy &&
        qemuProcessAutoDestroyAdd(driver, vm, conn) < 0)
        goto cleanup;

2718
    VIR_DEBUG("Writing domain status to disk");
2719 2720 2721 2722 2723 2724 2725 2726 2727 2728 2729 2730 2731 2732
    if (virDomainSaveStatus(driver->caps, driver->stateDir, vm) < 0)
        goto cleanup;

    virCommandFree(cmd);
    VIR_FORCE_CLOSE(logfile);

    return 0;

cleanup:
    /* We jump here if we failed to start the VM for any reason, or
     * if we failed to initialize the now running VM. kill it off and
     * pretend we never started it */
    virCommandFree(cmd);
    VIR_FORCE_CLOSE(logfile);
J
Jiri Denemark 已提交
2733
    qemuProcessStop(driver, vm, 0, VIR_DOMAIN_SHUTOFF_FAILED);
2734 2735 2736 2737 2738

    return -1;
}


2739 2740 2741 2742 2743 2744 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762
void qemuProcessKill(virDomainObjPtr vm)
{
    int i;
    VIR_DEBUG("vm=%s pid=%d", vm->def->name, vm->pid);

    if (!virDomainObjIsActive(vm)) {
        VIR_DEBUG("VM '%s' not active", vm->def->name);
        return;
    }

    /* This loop sends SIGTERM, then waits a few iterations
     * (1.6 seconds) to see if it dies. If still alive then
     * it does SIGKILL, and waits a few more iterations (1.6
     * seconds more) to confirm that it has really gone.
     */
    for (i = 0 ; i < 15 ; i++) {
        int signum;
        if (i == 0)
            signum = SIGTERM;
        else if (i == 8)
            signum = SIGKILL;
        else
            signum = 0; /* Just check for existence */

2763 2764
        if (virKillProcess(vm->pid, signum) < 0) {
            if (errno != ESRCH) {
2765 2766 2767 2768 2769 2770 2771 2772 2773 2774 2775 2776
                char ebuf[1024];
                VIR_WARN("Failed to kill process %d %s",
                         vm->pid, virStrerror(errno, ebuf, sizeof ebuf));
            }
            break;
        }

        usleep(200 * 1000);
    }
}


2777 2778
void qemuProcessStop(struct qemud_driver *driver,
                     virDomainObjPtr vm,
J
Jiri Denemark 已提交
2779 2780
                     int migrated,
                     virDomainShutoffReason reason)
2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799
{
    int ret;
    int retries = 0;
    qemuDomainObjPrivatePtr priv = vm->privateData;
    virErrorPtr orig_err;
    virDomainDefPtr def;
    int i;
    int logfile = -1;
    char *timestamp;
    char ebuf[1024];

    VIR_DEBUG("Shutting down VM '%s' pid=%d migrated=%d",
              vm->def->name, vm->pid, migrated);

    if (!virDomainObjIsActive(vm)) {
        VIR_DEBUG("VM '%s' not active", vm->def->name);
        return;
    }

2800
    if ((logfile = qemuDomainCreateLog(driver, vm, true)) < 0) {
2801 2802 2803 2804 2805 2806 2807 2808 2809 2810 2811 2812 2813 2814 2815 2816 2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834 2835 2836 2837 2838 2839 2840 2841 2842 2843 2844 2845 2846 2847 2848 2849 2850 2851 2852 2853 2854 2855
        /* To not break the normal domain shutdown process, skip the
         * timestamp log writing if failed on opening log file. */
        VIR_WARN("Unable to open logfile: %s",
                  virStrerror(errno, ebuf, sizeof ebuf));
    } else {
        if ((timestamp = virTimestamp()) == NULL) {
            virReportOOMError();
        } else {
            if (safewrite(logfile, timestamp, strlen(timestamp)) < 0 ||
                safewrite(logfile, SHUTDOWN_POSTFIX,
                          strlen(SHUTDOWN_POSTFIX)) < 0) {
                VIR_WARN("Unable to write timestamp to logfile: %s",
                         virStrerror(errno, ebuf, sizeof ebuf));
            }

            VIR_FREE(timestamp);
        }

        if (VIR_CLOSE(logfile) < 0)
             VIR_WARN("Unable to close logfile: %s",
                      virStrerror(errno, ebuf, sizeof ebuf));
    }

    /* This method is routinely used in clean up paths. Disable error
     * reporting so we don't squash a legit error. */
    orig_err = virSaveLastError();

    virDomainConfVMNWFilterTeardown(vm);

    if (driver->macFilter) {
        def = vm->def;
        for (i = 0 ; i < def->nnets ; i++) {
            virDomainNetDefPtr net = def->nets[i];
            if (net->ifname == NULL)
                continue;
            if ((errno = networkDisallowMacOnPort(driver, net->ifname,
                                                  net->mac))) {
                virReportSystemError(errno,
             _("failed to remove ebtables rule to allow MAC address on  '%s'"),
                                     net->ifname);
            }
        }
    }

    if (priv->mon)
        qemuMonitorClose(priv->mon);

    if (priv->monConfig) {
        if (priv->monConfig->type == VIR_DOMAIN_CHR_TYPE_UNIX)
            unlink(priv->monConfig->data.nix.path);
        virDomainChrSourceDefFree(priv->monConfig);
        priv->monConfig = NULL;
    }

    /* shut it off for sure */
2856
    qemuProcessKill(vm);
2857

2858 2859 2860
    /* Stop autodestroy in case guest is restarted */
    qemuProcessAutoDestroyRemove(driver, vm);

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
    /* now that we know it's stopped call the hook if present */
    if (virHookPresent(VIR_HOOK_DRIVER_QEMU)) {
        char *xml = virDomainDefFormat(vm->def, 0);

        /* we can't stop the operation even if the script raised an error */
        virHookCall(VIR_HOOK_DRIVER_QEMU, vm->def->name,
                    VIR_HOOK_QEMU_OP_STOPPED, VIR_HOOK_SUBOP_END, NULL, xml);
        VIR_FREE(xml);
    }

    /* Reset Security Labels */
    virSecurityManagerRestoreAllLabel(driver->securityManager,
                                      vm, migrated);
    virSecurityManagerReleaseLabel(driver->securityManager, vm);

    /* Clear out dynamically assigned labels */
    if (vm->def->seclabel.type == VIR_DOMAIN_SECLABEL_DYNAMIC) {
        VIR_FREE(vm->def->seclabel.model);
        VIR_FREE(vm->def->seclabel.label);
        VIR_FREE(vm->def->seclabel.imagelabel);
    }

    virDomainDefClearDeviceAliases(vm->def);
    if (!priv->persistentAddrs) {
        virDomainDefClearPCIAddresses(vm->def);
        qemuDomainPCIAddressSetFree(priv->pciaddrs);
        priv->pciaddrs = NULL;
    }

    qemuDomainReAttachHostDevices(driver, vm->def);

#if WITH_MACVTAP
    def = vm->def;
    for (i = 0; i < def->nnets; i++) {
        virDomainNetDefPtr net = def->nets[i];
        if (net->type == VIR_DOMAIN_NET_TYPE_DIRECT) {
            delMacvtap(net->ifname, net->mac, net->data.direct.linkdev,
2898 2899
                       net->data.direct.mode,
                       &net->data.direct.virtPortProfile, driver->stateDir);
2900 2901 2902 2903 2904 2905 2906 2907 2908 2909 2910 2911 2912 2913 2914 2915 2916 2917 2918 2919 2920 2921 2922 2923 2924 2925 2926 2927 2928 2929 2930 2931
            VIR_FREE(net->ifname);
        }
    }
#endif

retry:
    if ((ret = qemuRemoveCgroup(driver, vm, 0)) < 0) {
        if (ret == -EBUSY && (retries++ < 5)) {
            usleep(200*1000);
            goto retry;
        }
        VIR_WARN("Failed to remove cgroup for %s",
                 vm->def->name);
    }

    qemuProcessRemoveDomainStatus(driver, vm);

    /* Remove VNC port from port reservation bitmap, but only if it was
       reserved by the driver (autoport=yes)
    */
    if ((vm->def->ngraphics == 1) &&
        vm->def->graphics[0]->type == VIR_DOMAIN_GRAPHICS_TYPE_VNC &&
        vm->def->graphics[0]->data.vnc.autoport) {
        qemuProcessReturnPort(driver, vm->def->graphics[0]->data.vnc.port);
    }
    if ((vm->def->ngraphics == 1) &&
        vm->def->graphics[0]->type == VIR_DOMAIN_GRAPHICS_TYPE_SPICE &&
        vm->def->graphics[0]->data.spice.autoport) {
        qemuProcessReturnPort(driver, vm->def->graphics[0]->data.spice.port);
        qemuProcessReturnPort(driver, vm->def->graphics[0]->data.spice.tlsPort);
    }

2932
    vm->taint = 0;
2933 2934
    vm->pid = -1;
    vm->def->id = -1;
J
Jiri Denemark 已提交
2935
    virDomainObjSetState(vm, VIR_DOMAIN_SHUTOFF, reason);
2936 2937
    VIR_FREE(priv->vcpupids);
    priv->nvcpupids = 0;
2938 2939
    qemuCapsFree(priv->qemuCaps);
    priv->qemuCaps = NULL;
2940

2941
    /* The "release" hook cleans up additional resources */
2942 2943 2944 2945 2946 2947 2948 2949 2950
    if (virHookPresent(VIR_HOOK_DRIVER_QEMU)) {
        char *xml = virDomainDefFormat(vm->def, 0);

        /* we can't stop the operation even if the script raised an error */
        virHookCall(VIR_HOOK_DRIVER_QEMU, vm->def->name,
                    VIR_HOOK_QEMU_OP_RELEASE, VIR_HOOK_SUBOP_END, NULL, xml);
        VIR_FREE(xml);
    }

2951 2952 2953 2954 2955 2956 2957 2958 2959 2960 2961 2962
    if (vm->newDef) {
        virDomainDefFree(vm->def);
        vm->def = vm->newDef;
        vm->def->id = -1;
        vm->newDef = NULL;
    }

    if (orig_err) {
        virSetError(orig_err);
        virFreeError(orig_err);
    }
}
2963 2964 2965 2966 2967 2968 2969 2970 2971 2972 2973 2974 2975 2976 2977 2978 2979 2980 2981 2982 2983 2984 2985 2986 2987 2988 2989 2990 2991 2992 2993 2994 2995 2996 2997 2998 2999 3000 3001 3002 3003 3004 3005 3006 3007 3008 3009 3010 3011 3012 3013 3014 3015 3016 3017 3018 3019 3020 3021 3022 3023 3024 3025 3026 3027 3028 3029 3030 3031 3032 3033 3034 3035 3036 3037 3038 3039 3040 3041 3042 3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054 3055 3056 3057 3058 3059 3060 3061 3062 3063 3064 3065 3066 3067 3068 3069 3070 3071 3072 3073 3074


int qemuProcessAutoDestroyInit(struct qemud_driver *driver)
{
    if (!(driver->autodestroy = virHashCreate(5, NULL)))
        return -1;

    return 0;
}

struct qemuProcessAutoDestroyData {
    struct qemud_driver *driver;
    virConnectPtr conn;
};

static void qemuProcessAutoDestroyDom(void *payload,
                                      const void *name,
                                      void *opaque)
{
    struct qemuProcessAutoDestroyData *data = opaque;
    virConnectPtr conn = payload;
    const char *uuidstr = name;
    unsigned char uuid[VIR_UUID_BUFLEN];
    virDomainObjPtr dom;
    qemuDomainObjPrivatePtr priv;
    virDomainEventPtr event = NULL;

    VIR_DEBUG("conn=%p uuidstr=%s thisconn=%p", conn, uuidstr, data->conn);

    if (data->conn != conn)
        return;

    if (virUUIDParse(uuidstr, uuid) < 0) {
        VIR_WARN("Failed to parse %s", uuidstr);
        return;
    }

    if (!(dom = virDomainFindByUUID(&data->driver->domains,
                                    uuid))) {
        VIR_DEBUG("No domain object to kill");
        return;
    }

    priv = dom->privateData;
    if (priv->jobActive == QEMU_JOB_MIGRATION_IN) {
        VIR_DEBUG("vm=%s has incoming migration active, cancelling",
                  dom->def->name);
        priv->jobActive = QEMU_JOB_NONE;
        memset(&priv->jobInfo, 0, sizeof(priv->jobInfo));
    }

    if (qemuDomainObjBeginJobWithDriver(data->driver, dom) < 0)
        goto cleanup;

    VIR_DEBUG("Killing domain");
    qemuProcessStop(data->driver, dom, 1, VIR_DOMAIN_SHUTOFF_DESTROYED);
    qemuAuditDomainStop(dom, "destroyed");
    event = virDomainEventNewFromObj(dom,
                                     VIR_DOMAIN_EVENT_STOPPED,
                                     VIR_DOMAIN_EVENT_STOPPED_DESTROYED);
    if (qemuDomainObjEndJob(dom) == 0)
        dom = NULL;
    if (dom && !dom->persistent)
        virDomainRemoveInactive(&data->driver->domains, dom);

cleanup:
    if (dom)
        virDomainObjUnlock(dom);
    if (event)
        qemuDomainEventQueue(data->driver, event);
    virHashRemoveEntry(data->driver->autodestroy, uuidstr);
}

/*
 * Precondition: driver is locked
 */
void qemuProcessAutoDestroyRun(struct qemud_driver *driver, virConnectPtr conn)
{
    struct qemuProcessAutoDestroyData data = {
        driver, conn
    };
    VIR_DEBUG("conn=%p", conn);
    virHashForEach(driver->autodestroy, qemuProcessAutoDestroyDom, &data);
}

void qemuProcessAutoDestroyShutdown(struct qemud_driver *driver)
{
    virHashFree(driver->autodestroy);
}

int qemuProcessAutoDestroyAdd(struct qemud_driver *driver,
                              virDomainObjPtr vm,
                              virConnectPtr conn)
{
    char uuidstr[VIR_UUID_STRING_BUFLEN];
    virUUIDFormat(vm->def->uuid, uuidstr);
    VIR_DEBUG("vm=%s uuid=%s conn=%p", vm->def->name, uuidstr, conn);
    if (virHashAddEntry(driver->autodestroy, uuidstr, conn) < 0)
        return -1;
    return 0;
}

int qemuProcessAutoDestroyRemove(struct qemud_driver *driver,
                                 virDomainObjPtr vm)
{
    char uuidstr[VIR_UUID_STRING_BUFLEN];
    virUUIDFormat(vm->def->uuid, uuidstr);
    VIR_DEBUG("vm=%s uuid=%s", vm->def->name, uuidstr);
    if (virHashRemoveEntry(driver->autodestroy, uuidstr) < 0)
        return -1;
    return 0;
}
3075 3076 3077 3078 3079 3080 3081 3082 3083 3084 3085

bool qemuProcessAutoDestroyActive(struct qemud_driver *driver,
                                  virDomainObjPtr vm)
{
    char uuidstr[VIR_UUID_STRING_BUFLEN];
    virUUIDFormat(vm->def->uuid, uuidstr);
    VIR_DEBUG("vm=%s uuid=%s", vm->def->name, uuidstr);
    if (virHashLookup(driver->autodestroy, uuidstr) != NULL)
        return true;
    return false;
}