qemu_driver.c 200.5 KB
Newer Older
D
Daniel P. Berrange 已提交
1 2 3
/*
 * driver.c: core driver methods for managing qemu guests
 *
4
 * Copyright (C) 2006, 2007, 2008, 2009 Red Hat, Inc.
D
Daniel P. Berrange 已提交
5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
 * Copyright (C) 2006 Daniel P. Berrange
 *
 * 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
 *
 * Author: Daniel P. Berrange <berrange@redhat.com>
 */

24
#include <config.h>
25

D
Daniel P. Berrange 已提交
26 27 28 29 30 31 32 33 34 35 36
#include <sys/types.h>
#include <sys/poll.h>
#include <dirent.h>
#include <limits.h>
#include <string.h>
#include <stdio.h>
#include <strings.h>
#include <stdarg.h>
#include <stdlib.h>
#include <unistd.h>
#include <errno.h>
37
#include <sys/utsname.h>
38 39 40 41
#include <sys/stat.h>
#include <fcntl.h>
#include <signal.h>
#include <paths.h>
42 43
#include <pwd.h>
#include <stdio.h>
44
#include <sys/wait.h>
45
#include <sys/ioctl.h>
D
Daniel P. Berrange 已提交
46

47 48 49 50
#if HAVE_SCHED_H
#include <sched.h>
#endif

51
#include "virterror_internal.h"
52
#include "logging.h"
53
#include "datatypes.h"
54 55
#include "qemu_driver.h"
#include "qemu_conf.h"
J
Jim Meyering 已提交
56
#include "c-ctype.h"
57
#include "event.h"
58
#include "buf.h"
59
#include "util.h"
60
#include "nodeinfo.h"
61
#include "stats_linux.h"
62
#include "capabilities.h"
63
#include "memory.h"
64
#include "uuid.h"
65
#include "domain_conf.h"
66 67
#include "node_device_conf.h"
#include "pci.h"
68
#include "security.h"
69
#include "cgroup.h"
70

71

72 73
#define VIR_FROM_THIS VIR_FROM_QEMU

R
Richard W.M. Jones 已提交
74
/* For storing short-lived temporary files. */
75
#define TEMPDIR LOCAL_STATE_DIR "/cache/libvirt/qemu"
R
Richard W.M. Jones 已提交
76

77 78 79
#define QEMU_CMD_PROMPT "\n(qemu) "
#define QEMU_PASSWD_PROMPT "Password: "

80 81
static int qemudShutdown(void);

82 83
static void qemuDriverLock(struct qemud_driver *driver)
{
84
    virMutexLock(&driver->lock);
85 86 87
}
static void qemuDriverUnlock(struct qemud_driver *driver)
{
88
    virMutexUnlock(&driver->lock);
89 90
}

91 92 93
static void qemuDomainEventFlush(int timer, void *opaque);
static void qemuDomainEventQueue(struct qemud_driver *driver,
                                 virDomainEventPtr event);
94

95 96
static void qemudDispatchVMEvent(int watch,
                                 int fd,
97 98 99
                                 int events,
                                 void *opaque);

100 101
static int qemudStartVMDaemon(virConnectPtr conn,
                              struct qemud_driver *driver,
102
                              virDomainObjPtr vm,
103 104
                              const char *migrateFrom,
                              int stdin_fd);
105

106 107
static void qemudShutdownVMDaemon(virConnectPtr conn,
                                  struct qemud_driver *driver,
108
                                  virDomainObjPtr vm);
109

110
static int qemudDomainGetMaxVcpus(virDomainPtr dom);
111

112 113 114 115 116 117 118 119
static int qemudMonitorCommand(const virDomainObjPtr vm,
                               const char *cmd,
                               char **reply);
static int qemudMonitorCommandExtra(const virDomainObjPtr vm,
                                    const char *cmd,
                                    const char *extra,
                                    const char *extraPrompt,
                                    char **reply);
120 121 122
static int qemudDomainSetMemoryBalloon(virConnectPtr conn,
                                       virDomainObjPtr vm,
                                       unsigned long newmem);
123 124
static int qemudDetectVcpuPIDs(virConnectPtr conn,
                               virDomainObjPtr vm);
125

J
Jim Meyering 已提交
126
static struct qemud_driver *qemu_driver = NULL;
127

128 129 130 131 132 133 134 135 136
static int qemuCgroupControllerActive(struct qemud_driver *driver,
                                      int controller)
{
    if (driver->cgroup == NULL)
        return 0;
    if (driver->cgroupControllers & (1 << controller))
        return 1;
    return 0;
}
137

138
static int
139
qemudLogFD(virConnectPtr conn, struct qemud_driver *driver, const char* name)
140 141 142
{
    char logfile[PATH_MAX];
    mode_t logmode;
G
Guido Günther 已提交
143
    int ret, fd = -1;
144

145 146
    if ((ret = snprintf(logfile, sizeof(logfile), "%s/%s.log",
                        driver->logDir, name))
G
Guido Günther 已提交
147
        < 0 || ret >= sizeof(logfile)) {
148
        virReportOOMError(conn);
149 150 151 152
        return -1;
    }

    logmode = O_CREAT | O_WRONLY;
153 154
    /* Only logrotate files in /var/log, so only append if running privileged */
    if (driver->privileged)
155
        logmode |= O_APPEND;
156 157 158
    else
        logmode |= O_TRUNC;

159
    if ((fd = open(logfile, logmode, S_IRUSR | S_IWUSR)) < 0) {
160 161 162
        virReportSystemError(conn, errno,
                             _("failed to create logfile %s"),
                             logfile);
163 164
        return -1;
    }
165
    if (virSetCloseExec(fd) < 0) {
166 167
        virReportSystemError(conn, errno, "%s",
                             _("Unable to set VM logfile close-on-exec flag"));
168 169 170 171 172 173 174
        close(fd);
        return -1;
    }
    return fd;
}


175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191
static int
qemudLogReadFD(virConnectPtr conn, const char* logDir, const char* name, off_t pos)
{
    char logfile[PATH_MAX];
    mode_t logmode = O_RDONLY;
    int ret, fd = -1;

    if ((ret = snprintf(logfile, sizeof(logfile), "%s/%s.log", logDir, name))
        < 0 || ret >= sizeof(logfile)) {
        qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                         _("failed to build logfile name %s/%s.log"),
                         logDir, name);
        return -1;
    }


    if ((fd = open(logfile, logmode)) < 0) {
192 193 194
        virReportSystemError(conn, errno,
                             _("failed to create logfile %s"),
                             logfile);
195 196
        return -1;
    }
197
    if (virSetCloseExec(fd) < 0) {
198 199
        virReportSystemError(conn, errno, "%s",
                             _("Unable to set VM logfile close-on-exec flag"));
200 201 202 203
        close(fd);
        return -1;
    }
    if (lseek(fd, pos, SEEK_SET) < 0) {
204 205 206
        virReportSystemError(conn, errno,
                             _("Unable to seek to %lld in %s"),
                             (long long) pos, logfile);
207 208 209 210 211 212
        close(fd);
    }
    return fd;
}


213 214 215
static void
qemudAutostartConfigs(struct qemud_driver *driver) {
    unsigned int i;
216 217 218 219 220
    /* XXX: Figure out a better way todo this. The domain
     * startup code needs a connection handle in order
     * to lookup the bridge associated with a virtual
     * network
     */
221 222 223
    virConnectPtr conn = virConnectOpen(driver->privileged ?
                                        "qemu:///system" :
                                        "qemu:///session");
224
    /* Ignoring NULL conn which is mostly harmless here */
225

226
    qemuDriverLock(driver);
227
    for (i = 0 ; i < driver->domains.count ; i++) {
228
        virDomainObjPtr vm = driver->domains.objs[i];
229
        virDomainObjLock(vm);
230 231
        if (vm->autostart &&
            !virDomainIsActive(vm)) {
232
            int ret = qemudStartVMDaemon(conn, driver, vm, NULL, -1);
233 234
            if (ret < 0) {
                virErrorPtr err = virGetLastError();
235 236
                VIR_ERROR(_("Failed to autostart VM '%s': %s\n"),
                          vm->def->name,
237
                          err ? err->message : "");
238
            } else {
239 240 241 242 243 244
                virDomainEventPtr event =
                    virDomainEventNewFromObj(vm,
                                             VIR_DOMAIN_EVENT_STARTED,
                                             VIR_DOMAIN_EVENT_STARTED_BOOTED);
                if (event)
                    qemuDomainEventQueue(driver, event);
245
            }
246
        }
247
        virDomainObjUnlock(vm);
248
    }
249
    qemuDriverUnlock(driver);
250

251 252
    if (conn)
        virConnectClose(conn);
253 254
}

255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271

/**
 * qemudRemoveDomainStatus
 *
 * remove all state files of a domain from statedir
 *
 * Returns 0 on success
 */
static int
qemudRemoveDomainStatus(virConnectPtr conn,
                        struct qemud_driver *driver,
                        virDomainObjPtr vm)
{
    int rc = -1;
    char *file = NULL;

    if (virAsprintf(&file, "%s/%s.xml", driver->stateDir, vm->def->name) < 0) {
272
        virReportOOMError(conn);
273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291
        goto cleanup;
    }

    if (unlink(file) < 0 && errno != ENOENT && errno != ENOTDIR) {
        qemudReportError(conn, vm, NULL, VIR_ERR_INTERNAL_ERROR,
                         _("Failed to unlink status file %s"), file);
        goto cleanup;
    }

    if(virFileDeletePid(driver->stateDir, vm->def->name))
        goto cleanup;

    rc = 0;
cleanup:
    VIR_FREE(file);
    return rc;
}


292 293 294 295 296
static int qemudOpenMonitor(virConnectPtr conn,
                            struct qemud_driver* driver,
                            virDomainObjPtr vm,
                            int reconnect);

297 298 299 300

/*
 * Open an existing VM's monitor, re-detect VCPU threads
 * and re-reserve the security labels in use
301 302
 */
static int
303 304
qemuReconnectDomain(struct qemud_driver *driver,
                    virDomainObjPtr obj)
305
{
306
    int rc;
307

308
    if ((rc = qemudOpenMonitor(NULL, driver, obj, 1)) != 0) {
309 310 311 312
        VIR_ERROR(_("Failed to reconnect monitor for %s: %d\n"),
                  obj->def->name, rc);
        goto error;
    }
313

314 315 316
    if (qemudDetectVcpuPIDs(NULL, obj) < 0) {
        goto error;
    }
317

318 319 320 321 322
    if (obj->def->seclabel.type == VIR_DOMAIN_SECLABEL_DYNAMIC &&
        driver->securityDriver &&
        driver->securityDriver->domainReserveSecurityLabel &&
        driver->securityDriver->domainReserveSecurityLabel(NULL, obj) < 0)
        return -1;
323

324 325
    if (obj->def->id >= driver->nextvmid)
        driver->nextvmid = obj->def->id + 1;
326

327
    return 0;
328

329 330 331
error:
    return -1;
}
332

333 334 335 336 337 338 339 340 341 342
/**
 * qemudReconnectVMs
 *
 * Try to re-open the resources for live VMs that we care
 * about.
 */
static void
qemuReconnectDomains(struct qemud_driver *driver)
{
    int i;
343

344 345 346 347 348 349 350 351 352
    for (i = 0 ; i < driver->domains.count ; i++) {
        virDomainObjPtr obj = driver->domains.objs[i];

        virDomainObjLock(obj);
        if (qemuReconnectDomain(driver, obj) < 0) {
            /* If we can't get the monitor back, then kill the VM
             * so user has ability to start it again later without
             * danger of ending up running twice */
            qemudShutdownVMDaemon(NULL, driver, obj);
353
        }
354
        virDomainObjUnlock(obj);
355 356 357
    }
}

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

static int
qemudSecurityCapsInit(virSecurityDriverPtr secdrv,
                      virCapsPtr caps)
{
    const char *doi, *model;

    doi = virSecurityDriverGetDOI(secdrv);
    model = virSecurityDriverGetModel(secdrv);

    caps->host.secModel.model = strdup(model);
    if (!caps->host.secModel.model) {
        char ebuf[1024];
        VIR_ERROR(_("Failed to copy secModel model: %s"),
                  virStrerror(errno, ebuf, sizeof ebuf));
        return -1;
    }

    caps->host.secModel.doi = strdup(doi);
    if (!caps->host.secModel.doi) {
        char ebuf[1024];
        VIR_ERROR(_("Failed to copy secModel DOI: %s"),
                  virStrerror(errno, ebuf, sizeof ebuf));
        return -1;
    }

    VIR_DEBUG("Initialized caps for security driver \"%s\" with "
              "DOI \"%s\"", model, doi);

    return 0;
}


391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410
static int
qemudSecurityInit(struct qemud_driver *qemud_drv)
{
    int ret;
    virSecurityDriverPtr security_drv;

    ret = virSecurityDriverStartup(&security_drv,
                                   qemud_drv->securityDriverName);
    if (ret == -1) {
        VIR_ERROR0(_("Failed to start security driver"));
        return -1;
    }
    /* No security driver wanted to be enabled: just return */
    if (ret == -2) {
        VIR_INFO0(_("No security driver available"));
        return 0;
    }

    qemud_drv->securityDriver = security_drv;

411
    VIR_INFO("Initialized security driver %s", security_drv->name);
412 413 414 415 416

    /*
     * Add security policy host caps now that the security driver is
     * initialized.
     */
417 418
    return qemudSecurityCapsInit(security_drv, qemud_drv->caps);
}
419 420


421

422 423 424 425 426 427
/**
 * qemudStartup:
 *
 * Initialization function for the QEmu daemon
 */
static int
428
qemudStartup(int privileged) {
429
    char *base = NULL;
D
Daniel P. Berrange 已提交
430
    char driverConf[PATH_MAX];
431
    int rc;
432

433
    if (VIR_ALLOC(qemu_driver) < 0)
434 435
        return -1;

436
    if (virMutexInit(&qemu_driver->lock) < 0) {
437
        VIR_ERROR("%s", _("cannot initialize mutex"));
438 439 440
        VIR_FREE(qemu_driver);
        return -1;
    }
441
    qemuDriverLock(qemu_driver);
442
    qemu_driver->privileged = privileged;
443

444 445 446
    /* Don't have a dom0 so start from 1 */
    qemu_driver->nextvmid = 1;

447 448
    /* Init callback list */
    if(VIR_ALLOC(qemu_driver->domainEventCallbacks) < 0)
449
        goto out_of_memory;
450 451 452 453 454 455
    if (!(qemu_driver->domainEventQueue = virDomainEventQueueNew()))
        goto out_of_memory;

    if ((qemu_driver->domainEventTimer =
         virEventAddTimeout(-1, qemuDomainEventFlush, qemu_driver, NULL)) < 0)
        goto error;
456

457
    if (privileged) {
458 459
        if (virAsprintf(&qemu_driver->logDir,
                        "%s/log/libvirt/qemu", LOCAL_STATE_DIR) == -1)
460
            goto out_of_memory;
461

D
Daniel P. Berrange 已提交
462
        if ((base = strdup (SYSCONF_DIR "/libvirt")) == NULL)
463
            goto out_of_memory;
464 465 466 467

        if (virAsprintf(&qemu_driver->stateDir,
                      "%s/run/libvirt/qemu/", LOCAL_STATE_DIR) == -1)
            goto out_of_memory;
468
    } else {
469
        uid_t uid = geteuid();
470 471
        char *userdir = virGetUserDirectory(NULL, uid);
        if (!userdir)
472
            goto error;
473

474
        if (virAsprintf(&qemu_driver->logDir,
475 476
                        "%s/.libvirt/qemu/log", userdir) == -1) {
            VIR_FREE(userdir);
477
            goto out_of_memory;
478
        }
479

480 481
        if (virAsprintf(&base, "%s/.libvirt", userdir) == -1) {
            VIR_FREE(userdir);
482
            goto out_of_memory;
483 484
        }
        VIR_FREE(userdir);
485 486 487 488 489 490

        if (virAsprintf(&qemu_driver->stateDir, "%s/qemu/run", base) == -1)
            goto out_of_memory;
    }

    if (virFileMakePath(qemu_driver->stateDir) < 0) {
491
        char ebuf[1024];
492 493
        VIR_ERROR(_("Failed to create state dir '%s': %s\n"),
                  qemu_driver->stateDir, virStrerror(errno, ebuf, sizeof ebuf));
494
        goto error;
495 496 497 498 499
    }

    /* Configuration paths are either ~/.libvirt/qemu/... (session) or
     * /etc/libvirt/qemu/... (system).
     */
D
Daniel P. Berrange 已提交
500
    if (snprintf (driverConf, sizeof(driverConf), "%s/qemu.conf", base) == -1)
501
        goto out_of_memory;
D
Daniel P. Berrange 已提交
502
    driverConf[sizeof(driverConf)-1] = '\0';
503

504
    if (virAsprintf(&qemu_driver->configDir, "%s/qemu", base) == -1)
505 506
        goto out_of_memory;

507
    if (virAsprintf(&qemu_driver->autostartDir, "%s/qemu/autostart", base) == -1)
508 509
        goto out_of_memory;

510
    VIR_FREE(base);
511

512 513 514 515 516 517 518
    rc = virCgroupForDriver("qemu", &qemu_driver->cgroup, privileged, 1);
    if (rc < 0) {
        char buf[1024];
        VIR_WARN("Unable to create cgroup for driver: %s",
                 virStrerror(-rc, buf, sizeof(buf)));
    }

519
    if ((qemu_driver->caps = qemudCapsInit(NULL)) == NULL)
520
        goto out_of_memory;
D
Daniel P. Berrange 已提交
521

522
    if (qemudLoadDriverConfig(qemu_driver, driverConf) < 0) {
523 524 525
        goto error;
    }

526
    if (qemudSecurityInit(qemu_driver) < 0) {
527
        goto error;
D
Daniel P. Berrange 已提交
528 529
    }

530 531 532 533 534 535 536 537 538 539 540 541
    /* Get all the running persistent or transient configs first */
    if (virDomainLoadAllConfigs(NULL,
                                qemu_driver->caps,
                                &qemu_driver->domains,
                                qemu_driver->stateDir,
                                NULL,
                                1, NULL, NULL) < 0)
        goto error;

    qemuReconnectDomains(qemu_driver);

    /* Then inactive persistent configs */
542 543 544 545
    if (virDomainLoadAllConfigs(NULL,
                                qemu_driver->caps,
                                &qemu_driver->domains,
                                qemu_driver->configDir,
546
                                qemu_driver->autostartDir,
547
                                0, NULL, NULL) < 0)
548
        goto error;
549 550
    qemuDriverUnlock(qemu_driver);

551 552
    qemudAutostartConfigs(qemu_driver);

553

554 555
    return 0;

556
out_of_memory:
557
    virReportOOMError(NULL);
558 559 560
error:
    if (qemu_driver)
        qemuDriverUnlock(qemu_driver);
561
    VIR_FREE(base);
562
    qemudShutdown();
563 564 565
    return -1;
}

566 567 568 569
static void qemudNotifyLoadDomain(virDomainObjPtr vm, int newVM, void *opaque)
{
    struct qemud_driver *driver = opaque;

570 571 572 573 574 575 576 577
    if (newVM) {
        virDomainEventPtr event =
            virDomainEventNewFromObj(vm,
                                     VIR_DOMAIN_EVENT_DEFINED,
                                     VIR_DOMAIN_EVENT_DEFINED_ADDED);
        if (event)
            qemuDomainEventQueue(driver, event);
    }
578 579
}

580 581 582 583 584 585 586 587
/**
 * qemudReload:
 *
 * Function to restart the QEmu daemon, it will recheck the configuration
 * files and update its state and the networking
 */
static int
qemudReload(void) {
588 589 590
    if (!qemu_driver)
        return 0;

591
    qemuDriverLock(qemu_driver);
592 593 594 595
    virDomainLoadAllConfigs(NULL,
                            qemu_driver->caps,
                            &qemu_driver->domains,
                            qemu_driver->configDir,
596
                            qemu_driver->autostartDir,
597
                            0, qemudNotifyLoadDomain, qemu_driver);
598
    qemuDriverUnlock(qemu_driver);
599

600
    qemudAutostartConfigs(qemu_driver);
601 602

    return 0;
603 604
}

605 606 607 608 609 610 611 612 613 614
/**
 * qemudActive:
 *
 * Checks if the QEmu daemon is active, i.e. has an active domain or
 * an active network
 *
 * Returns 1 if active, 0 otherwise
 */
static int
qemudActive(void) {
615
    unsigned int i;
616
    int active = 0;
617

618 619 620
    if (!qemu_driver)
        return 0;

621 622 623 624 625 626 627 628
    qemuDriverLock(qemu_driver);
    for (i = 0 ; i < qemu_driver->domains.count ; i++) {
        virDomainObjPtr vm = qemu_driver->domains.objs[i];
        virDomainObjLock(vm);
        if (virDomainIsActive(vm))
            active = 1;
        virDomainObjUnlock(vm);
    }
629

630 631
    qemuDriverUnlock(qemu_driver);
    return active;
632 633
}

634 635 636 637 638 639 640
/**
 * qemudShutdown:
 *
 * Shutdown the QEmu daemon, it will stop all active domains and networks
 */
static int
qemudShutdown(void) {
641

642
    if (!qemu_driver)
643
        return -1;
644

645
    qemuDriverLock(qemu_driver);
646 647
    virCapabilitiesFree(qemu_driver->caps);

648
    virDomainObjListFree(&qemu_driver->domains);
649

650
    VIR_FREE(qemu_driver->securityDriverName);
651
    VIR_FREE(qemu_driver->logDir);
652 653
    VIR_FREE(qemu_driver->configDir);
    VIR_FREE(qemu_driver->autostartDir);
654
    VIR_FREE(qemu_driver->stateDir);
655
    VIR_FREE(qemu_driver->vncTLSx509certdir);
J
Jim Meyering 已提交
656
    VIR_FREE(qemu_driver->vncListen);
657
    VIR_FREE(qemu_driver->vncPassword);
658
    VIR_FREE(qemu_driver->vncSASLdir);
D
Daniel P. Berrange 已提交
659

660 661
    /* Free domain callback list */
    virDomainEventCallbackListFree(qemu_driver->domainEventCallbacks);
662 663 664 665
    virDomainEventQueueFree(qemu_driver->domainEventQueue);

    if (qemu_driver->domainEventTimer != -1)
        virEventRemoveTimeout(qemu_driver->domainEventTimer);
666

667 668 669
    if (qemu_driver->brctl)
        brShutdown(qemu_driver->brctl);

670 671
    virCgroupFree(&qemu_driver->cgroup);

672
    qemuDriverUnlock(qemu_driver);
673
    virMutexDestroy(&qemu_driver->lock);
674
    VIR_FREE(qemu_driver);
675 676

    return 0;
677 678 679
}

/* Return -1 for error, 1 to continue reading and 0 for success */
680
typedef int qemudHandlerMonitorOutput(virConnectPtr conn,
681
                                      virDomainObjPtr vm,
682 683 684
                                      const char *output,
                                      int fd);

685 686 687
/*
 * Returns -1 for error, 0 on end-of-file, 1 for success
 */
688
static int
689
qemudReadMonitorOutput(virConnectPtr conn,
690
                       virDomainObjPtr vm,
691 692
                       int fd,
                       char *buf,
G
Guido Günther 已提交
693
                       size_t buflen,
694
                       qemudHandlerMonitorOutput func,
695 696
                       const char *what,
                       int timeout)
697
{
G
Guido Günther 已提交
698
    size_t got = 0;
699
    buf[0] = '\0';
700
    timeout *= 1000; /* poll wants milli seconds */
701

702
    /* Consume & discard the initial greeting */
703
    while (got < (buflen-1)) {
G
Guido Günther 已提交
704
        ssize_t ret;
705 706

        ret = read(fd, buf+got, buflen-got-1);
707

708 709 710 711 712 713
        if (ret < 0) {
            struct pollfd pfd = { .fd = fd, .events = POLLIN };
            if (errno == EINTR)
                continue;

            if (errno != EAGAIN) {
714 715 716
                virReportSystemError(conn, errno,
                                     _("Failure while reading %s startup output"),
                                     what);
717 718 719
                return -1;
            }

720
            ret = poll(&pfd, 1, timeout);
721
            if (ret == 0) {
722
                qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
723
                                 _("Timed out while reading %s startup output"), what);
724 725 726
                return -1;
            } else if (ret == -1) {
                if (errno != EINTR) {
727 728 729
                    virReportSystemError(conn, errno,
                                         _("Failure while reading %s startup output"),
                                         what);
730 731 732 733 734 735 736 737
                    return -1;
                }
            } else {
                /* Make sure we continue loop & read any further data
                   available before dealing with EOF */
                if (pfd.revents & (POLLIN | POLLHUP))
                    continue;

738
                qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
739
                                 _("Failure while reading %s startup output"), what);
740 741
                return -1;
            }
742 743
        } else if (ret == 0) {
            return 0;
744 745 746
        } else {
            got += ret;
            buf[got] = '\0';
747 748 749 750 751 752
            ret = func(conn, vm, buf, fd);
            if (ret == -1)
                return -1;
            if (ret == 1)
                continue;
            return 1;
753 754 755
        }
    }

756
    qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
757
                     _("Out of space while reading %s startup output"), what);
758 759 760 761
    return -1;

}

762 763 764 765 766 767 768 769 770

/*
 * Returns -1 for error, 0 on success
 */
static int
qemudReadLogOutput(virConnectPtr conn,
                   virDomainObjPtr vm,
                   int fd,
                   char *buf,
G
Guido Günther 已提交
771
                   size_t buflen,
772 773 774 775
                   qemudHandlerMonitorOutput func,
                   const char *what,
                   int timeout)
{
776
    int retries = (timeout*10);
777
    int got = 0;
778 779 780
    buf[0] = '\0';

    while (retries) {
781
        ssize_t func_ret, ret;
782
        int isdead = 0;
G
Guido Günther 已提交
783

784 785
        func_ret = func(conn, vm, buf, fd);

786 787
        if (kill(vm->pid, 0) == -1 && errno == ESRCH)
            isdead = 1;
788

789 790
        /* Any failures should be detected before we read the log, so we
         * always have something useful to report on failure. */
791 792
        ret = saferead(fd, buf+got, buflen-got-1);
        if (ret < 0) {
793 794 795 796 797 798
            virReportSystemError(conn, errno,
                                 _("Failure while reading %s log output"),
                                 what);
            return -1;
        }

799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814
        got += ret;
        buf[got] = '\0';
        if (got == buflen-1) {
            qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                             _("Out of space while reading %s log output"),
                             what);
            return -1;
        }

        if (isdead) {
            qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                             _("Process exited while reading %s log output"),
                             what);
            return -1;
        }

815 816
        if (func_ret <= 0)
            return func_ret;
817 818 819 820 821 822 823 824 825 826

        usleep(100*1000);
        retries--;
    }
    if (retries == 0)
        qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                         _("Timed out while reading %s log output"), what);
    return -1;
}

827
static int
828
qemudCheckMonitorPrompt(virConnectPtr conn ATTRIBUTE_UNUSED,
829
                        virDomainObjPtr vm,
830 831 832 833 834 835 836 837 838 839 840
                        const char *output,
                        int fd)
{
    if (strstr(output, "(qemu) ") == NULL)
        return 1; /* keep reading */

    vm->monitor = fd;

    return 0;
}

841
static int
842 843 844 845 846
qemudOpenMonitorCommon(virConnectPtr conn,
                       struct qemud_driver* driver,
                       virDomainObjPtr vm,
                       int monfd,
                       int reconnect)
847
{
848
    char buf[1024];
849
    int ret;
850

851
    if (virSetCloseExec(monfd) < 0) {
852
        qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
853
                         "%s", _("Unable to set monitor close-on-exec flag"));
854
        return -1;
855
    }
856
    if (virSetNonBlock(monfd) < 0) {
857
        qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
858
                         "%s", _("Unable to put monitor into non-blocking mode"));
859
        return -1;
860 861
    }

862
    if (!reconnect) {
863 864 865 866
        if (qemudReadMonitorOutput(conn,
                                   vm, monfd,
                                   buf, sizeof(buf),
                                   qemudCheckMonitorPrompt,
867
                                   "monitor", 10) <= 0)
868 869 870
            ret = -1;
        else
            ret = 0;
871 872 873 874 875 876
    } else {
        vm->monitor = monfd;
        ret = 0;
    }

    if (ret != 0)
877
        return ret;
878

879 880 881
    if ((vm->monitorWatch = virEventAddHandle(vm->monitor, 0,
                                              qemudDispatchVMEvent,
                                              driver, NULL)) < 0)
882
        return -1;
883

884 885
    return 0;
}
886

887 888 889 890 891 892 893 894 895
static int
qemudOpenMonitorUnix(virConnectPtr conn,
                     struct qemud_driver* driver,
                     virDomainObjPtr vm,
                     const char *monitor,
                     int reconnect)
{
    struct sockaddr_un addr;
    int monfd;
896
    int timeout = 3; /* In seconds */
897
    int ret, i = 0;
898 899 900 901 902 903 904 905 906 907 908

    if ((monfd = socket(AF_UNIX, SOCK_STREAM, 0)) < 0) {
        virReportSystemError(conn, errno,
                             "%s", _("failed to create socket"));
        return -1;
    }

    memset(&addr, 0, sizeof(addr));
    addr.sun_family = AF_UNIX;
    strncpy(addr.sun_path, monitor, sizeof(addr.sun_path));

909 910 911 912 913 914 915 916 917 918 919 920
    do {
        ret = connect(monfd, (struct sockaddr *) &addr, sizeof(addr));

        if (ret == 0)
            break;

        if (errno == EACCES || errno == ECONNREFUSED) {
            /* EACCES       : Socket may not have shown up yet
             * ECONNREFUSED : Leftover socket hasn't been removed yet */
            continue;
        }

921 922 923
        virReportSystemError(conn, errno, "%s",
                             _("failed to connect to monitor socket"));
        goto error;
924 925 926 927 928 929 930

    } while ((++i <= timeout*5) && (usleep(.2 * 1000000) <= 0));

    if (ret != 0) {
        virReportSystemError(conn, errno, "%s",
                             _("monitor socket did not show up."));
        goto error;
931 932 933 934 935 936 937 938 939 940 941 942
    }

    if (qemudOpenMonitorCommon(conn, driver, vm, monfd, reconnect) < 0)
        goto error;

    return 0;

error:
    close(monfd);
    return -1;
}

943 944 945 946 947 948 949 950
static int
qemudOpenMonitorPty(virConnectPtr conn,
                    struct qemud_driver* driver,
                    virDomainObjPtr vm,
                    const char *monitor,
                    int reconnect)
{
    int monfd;
951

952 953 954 955 956 957 958 959 960 961 962 963
    if ((monfd = open(monitor, O_RDWR)) < 0) {
        qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                         _("Unable to open monitor path %s"), monitor);
        return -1;
    }

    if (qemudOpenMonitorCommon(conn, driver, vm, monfd, reconnect) < 0)
        goto error;

    return 0;

error:
964
    close(monfd);
965 966 967 968 969 970 971 972 973 974
    return -1;
}

static int
qemudOpenMonitor(virConnectPtr conn,
                 struct qemud_driver *driver,
                 virDomainObjPtr vm,
                 int reconnect)
{
    switch (vm->monitor_chr->type) {
975 976 977 978
    case VIR_DOMAIN_CHR_TYPE_UNIX:
        return qemudOpenMonitorUnix(conn, driver, vm,
                                    vm->monitor_chr->data.nix.path,
                                    reconnect);
979 980 981 982 983 984 985 986 987 988
    case VIR_DOMAIN_CHR_TYPE_PTY:
        return qemudOpenMonitorPty(conn, driver, vm,
                                   vm->monitor_chr->data.file.path,
                                   reconnect);
    default:
        qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                         _("unable to handle monitor type: %s"),
                         virDomainChrTypeToString(vm->monitor_chr->type));
        return -1;
    }
989 990
}

991
/* Returns -1 for error, 0 success, 1 continue reading */
992 993 994 995 996 997
static int
qemudExtractMonitorPath(virConnectPtr conn,
                        const char *haystack,
                        size_t *offset,
                        char **path)
{
998
    static const char needle[] = "char device redirected to";
999
    char *tmp, *dev;
1000

1001
    VIR_FREE(*path);
1002
    /* First look for our magic string */
1003 1004 1005 1006 1007
    if (!(tmp = strstr(haystack + *offset, needle))) {
        return 1;
    }
    tmp += sizeof(needle);
    dev = tmp;
1008

1009 1010 1011 1012 1013
    /*
     * And look for first whitespace character and nul terminate
     * to mark end of the pty path
     */
    while (*tmp) {
1014
        if (c_isspace(*tmp)) {
1015
            if (VIR_ALLOC_N(*path, (tmp-dev)+1) < 0) {
1016
                virReportOOMError(conn);
1017 1018 1019 1020
                return -1;
            }
            strncpy(*path, dev, (tmp-dev));
            (*path)[(tmp-dev)] = '\0';
1021
            /* ... now further update offset till we get EOL */
1022
            *offset = tmp - haystack;
1023 1024
            return 0;
        }
1025
        tmp++;
1026 1027 1028 1029 1030
    }

    /*
     * We found a path, but didn't find any whitespace,
     * so it must be still incomplete - we should at
1031 1032
     * least see a \n - indicate that we want to carry
     * on trying again
1033
     */
1034
    return 1;
1035 1036 1037
}

static int
1038
qemudFindCharDevicePTYs(virConnectPtr conn,
1039
                        virDomainObjPtr vm,
1040 1041
                        const char *output,
                        int fd ATTRIBUTE_UNUSED)
1042
{
1043
    size_t offset = 0;
1044
    int ret, i;
1045 1046

    /* The order in which QEMU prints out the PTY paths is
1047 1048
       the order in which it procsses its serial and parallel
       device args. This code must match that ordering.... */
1049

1050
    /* first comes the serial devices */
1051 1052
    for (i = 0 ; i < vm->def->nserials ; i++) {
        virDomainChrDefPtr chr = vm->def->serials[i];
1053 1054 1055
        if (chr->type == VIR_DOMAIN_CHR_TYPE_PTY) {
            if ((ret = qemudExtractMonitorPath(conn, output, &offset,
                                               &chr->data.file.path)) != 0)
1056
                return ret;
1057 1058 1059
        }
    }

1060
    /* then the parallel devices */
1061 1062
    for (i = 0 ; i < vm->def->nparallels ; i++) {
        virDomainChrDefPtr chr = vm->def->parallels[i];
1063 1064 1065
        if (chr->type == VIR_DOMAIN_CHR_TYPE_PTY) {
            if ((ret = qemudExtractMonitorPath(conn, output, &offset,
                                               &chr->data.file.path)) != 0)
1066
                return ret;
1067 1068 1069
        }
    }

1070
    return 0;
1071 1072
}

1073 1074 1075 1076
static int
qemudWaitForMonitor(virConnectPtr conn,
                    struct qemud_driver* driver,
                    virDomainObjPtr vm, off_t pos)
1077
{
1078
    char buf[4096]; /* Plenty of space to get startup greeting */
1079 1080 1081 1082 1083
    int logfd;
    int ret;

    if ((logfd = qemudLogReadFD(conn, driver->logDir, vm->def->name, pos))
        < 0)
1084
        return -1;
1085

1086 1087 1088
    ret = qemudReadLogOutput(conn, vm, logfd, buf, sizeof(buf),
                             qemudFindCharDevicePTYs,
                             "console", 3);
1089
    if (close(logfd) < 0) {
1090
        char ebuf[4096];
1091
        VIR_WARN(_("Unable to close logfile: %s\n"),
1092 1093
                 virStrerror(errno, ebuf, sizeof ebuf));
    }
1094

1095 1096 1097 1098 1099 1100
    if (ret < 0) {
        /* Unexpected end of file - inform user of QEMU log data */
        qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                         _("unable to start guest: %s"), buf);
        return -1;
    }
1101

1102 1103 1104 1105
    if (qemudOpenMonitor(conn, driver, vm, 0) < 0)
        return -1;

    return 0;
1106 1107
}

1108 1109
static int
qemudDetectVcpuPIDs(virConnectPtr conn,
1110
                    virDomainObjPtr vm) {
1111 1112 1113 1114 1115 1116
    char *qemucpus = NULL;
    char *line;
    int lastVcpu = -1;

    /* Only KVM has seperate threads for CPUs,
       others just use main QEMU process for CPU */
1117
    if (vm->def->virtType != VIR_DOMAIN_VIRT_KVM)
1118 1119 1120 1121 1122
        vm->nvcpupids = 1;
    else
        vm->nvcpupids = vm->def->vcpus;

    if (VIR_ALLOC_N(vm->vcpupids, vm->nvcpupids) < 0) {
1123
        virReportOOMError(conn);
1124 1125 1126
        return -1;
    }

1127
    if (vm->def->virtType != VIR_DOMAIN_VIRT_KVM) {
1128 1129 1130 1131
        vm->vcpupids[0] = vm->pid;
        return 0;
    }

1132
    if (qemudMonitorCommand(vm, "info cpus", &qemucpus) < 0) {
1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190
        qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                         "%s", _("cannot run monitor command to fetch CPU thread info"));
        VIR_FREE(vm->vcpupids);
        vm->nvcpupids = 0;
        return -1;
    }

    /*
     * This is the gross format we're about to parse :-{
     *
     * (qemu) info cpus
     * * CPU #0: pc=0x00000000000f0c4a thread_id=30019
     *   CPU #1: pc=0x00000000fffffff0 thread_id=30020
     *   CPU #2: pc=0x00000000fffffff0 thread_id=30021
     *
     */
    line = qemucpus;
    do {
        char *offset = strchr(line, '#');
        char *end = NULL;
        int vcpu = 0, tid = 0;

        /* See if we're all done */
        if (offset == NULL)
            break;

        /* Extract VCPU number */
        if (virStrToLong_i(offset + 1, &end, 10, &vcpu) < 0)
            goto error;
        if (end == NULL || *end != ':')
            goto error;

        /* Extract host Thread ID */
        if ((offset = strstr(line, "thread_id=")) == NULL)
            goto error;
        if (virStrToLong_i(offset + strlen("thread_id="), &end, 10, &tid) < 0)
            goto error;
        if (end == NULL || !c_isspace(*end))
            goto error;

        /* Validate the VCPU is in expected range & order */
        if (vcpu > vm->nvcpupids ||
            vcpu != (lastVcpu + 1))
            goto error;

        lastVcpu = vcpu;
        vm->vcpupids[vcpu] = tid;

        /* Skip to next data line */
        line = strchr(offset, '\r');
        if (line == NULL)
            line = strchr(offset, '\n');
    } while (line != NULL);

    /* Validate we got data for all VCPUs we expected */
    if (lastVcpu != (vm->def->vcpus - 1))
        goto error;

1191
    VIR_FREE(qemucpus);
1192 1193 1194 1195
    return 0;

error:
    VIR_FREE(vm->vcpupids);
1196 1197
    vm->nvcpupids = 0;
    VIR_FREE(qemucpus);
1198 1199 1200 1201 1202 1203 1204 1205

    /* Explicitly return success, not error. Older KVM does
       not have vCPU -> Thread mapping info and we don't
       want to break its use. This merely disables ability
       to pin vCPUS with libvirt */
    return 0;
}

1206 1207
static int
qemudInitCpus(virConnectPtr conn,
D
Daniel Veillard 已提交
1208 1209
              virDomainObjPtr vm,
              const char *migrateFrom) {
1210 1211 1212 1213 1214 1215
    char *info = NULL;
#if HAVE_SCHED_GETAFFINITY
    cpu_set_t mask;
    int i, maxcpu = QEMUD_CPUMASK_LEN;
    virNodeInfo nodeinfo;

1216
    if (nodeGetInfo(conn, &nodeinfo) < 0)
1217 1218 1219 1220 1221 1222 1223 1224
        return -1;

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

    CPU_ZERO(&mask);
D
Daniel P. Berrange 已提交
1225 1226 1227 1228 1229 1230
    if (vm->def->cpumask) {
        for (i = 0 ; i < maxcpu ; i++)
            if (vm->def->cpumask[i])
                CPU_SET(i, &mask);
    } else {
        for (i = 0 ; i < maxcpu ; i++)
1231
            CPU_SET(i, &mask);
D
Daniel P. Berrange 已提交
1232
    }
1233 1234 1235 1236

    for (i = 0 ; i < vm->nvcpupids ; i++) {
        if (sched_setaffinity(vm->vcpupids[i],
                              sizeof(mask), &mask) < 0) {
1237 1238
            virReportSystemError(conn, errno, "%s",
                                 _("failed to set CPU affinity"));
1239 1240 1241 1242 1243
            return -1;
        }
    }
#endif /* HAVE_SCHED_GETAFFINITY */

D
Daniel Veillard 已提交
1244 1245
    if (migrateFrom == NULL) {
        /* Allow the CPUS to start executing */
1246
        if (qemudMonitorCommand(vm, "cont", &info) < 0) {
D
Daniel Veillard 已提交
1247 1248 1249 1250 1251
            qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                             "%s", _("resume operation failed"));
            return -1;
        }
        VIR_FREE(info);
1252 1253 1254 1255 1256 1257
    }

    return 0;
}


1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269
static int
qemudInitPasswords(virConnectPtr conn,
                   struct qemud_driver *driver,
                   virDomainObjPtr vm) {
    char *info = NULL;

    /*
     * NB: Might have more passwords to set in the future. eg a qcow
     * disk decryption password, but there's no monitor command
     * for that yet...
     */

1270 1271 1272
    if ((vm->def->ngraphics == 1) &&
        vm->def->graphics[0]->type == VIR_DOMAIN_GRAPHICS_TYPE_VNC &&
        (vm->def->graphics[0]->data.vnc.passwd || driver->vncPassword)) {
1273 1274

        if (qemudMonitorCommandExtra(vm, "change vnc password",
1275 1276
                                     vm->def->graphics[0]->data.vnc.passwd ?
                                     vm->def->graphics[0]->data.vnc.passwd :
1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290
                                     driver->vncPassword,
                                     QEMU_PASSWD_PROMPT,
                                     &info) < 0) {
            qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                             "%s", _("setting VNC password failed"));
            return -1;
        }
        VIR_FREE(info);
    }

    return 0;
}


1291
static int qemudNextFreeVNCPort(struct qemud_driver *driver ATTRIBUTE_UNUSED) {
1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326
    int i;

    for (i = 5900 ; i < 6000 ; i++) {
        int fd;
        int reuse = 1;
        struct sockaddr_in addr;
        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) {
            close(fd);
            break;
        }

        if (bind(fd, (struct sockaddr*)&addr, sizeof(addr)) == 0) {
            /* Not in use, lets grab it */
            close(fd);
            return i;
        }
        close(fd);

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

1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344
static int qemuPrepareHostDevices(virConnectPtr conn,
                                  virDomainDefPtr def) {
    int i;

    /* We have to use 2 loops here. *All* devices must
     * be detached before we reset any of them, because
     * in some cases you have to reset the whole PCI,
     * which impacts all devices on it
     */

    for (i = 0 ; i < def->nhostdevs ; i++) {
        virDomainHostdevDefPtr hostdev = def->hostdevs[i];

        if (hostdev->mode != VIR_DOMAIN_HOSTDEV_MODE_SUBSYS)
            continue;
        if (hostdev->source.subsys.type != VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_PCI)
            continue;

1345
        if (hostdev->managed) {
1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399
            pciDevice *dev = pciGetDevice(conn,
                                          hostdev->source.subsys.u.pci.domain,
                                          hostdev->source.subsys.u.pci.bus,
                                          hostdev->source.subsys.u.pci.slot,
                                          hostdev->source.subsys.u.pci.function);
            if (!dev)
                goto error;

            if (pciDettachDevice(conn, dev) < 0) {
                pciFreeDevice(conn, dev);
                goto error;
            }

            pciFreeDevice(conn, dev);
        } /* else {
             XXX validate that non-managed device isn't in use, eg
             by checking that device is either un-bound, or bound
             to pci-stub.ko
        } */
    }

    /* Now that all the PCI hostdevs have be dettached, we can safely
     * reset them */
    for (i = 0 ; i < def->nhostdevs ; i++) {
        virDomainHostdevDefPtr hostdev = def->hostdevs[i];
        pciDevice *dev;

        if (hostdev->mode != VIR_DOMAIN_HOSTDEV_MODE_SUBSYS)
            continue;
        if (hostdev->source.subsys.type != VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_PCI)
            continue;

        dev = pciGetDevice(conn,
                           hostdev->source.subsys.u.pci.domain,
                           hostdev->source.subsys.u.pci.bus,
                           hostdev->source.subsys.u.pci.slot,
                           hostdev->source.subsys.u.pci.function);
        if (!dev)
            goto error;

        if (pciResetDevice(conn, dev) < 0) {
            pciFreeDevice(conn, dev);
            goto error;
        }

        pciFreeDevice(conn, dev);
    }

    return 0;

error:
    return -1;
}

1400 1401 1402 1403 1404 1405 1406 1407 1408 1409
static const char *const defaultDeviceACL[] = {
    "/dev/null", "/dev/full", "/dev/zero",
    "/dev/random", "/dev/urandom",
    "/dev/ptmx", "/dev/kvm", "/dev/kqemu",
    "/dev/rtc", "/dev/hpet", "/dev/net/tun",
    NULL,
};
#define DEVICE_PTY_MAJOR 136
#define DEVICE_SND_MAJOR 116

1410 1411 1412 1413 1414 1415
static int qemuSetupCgroup(virConnectPtr conn,
                           struct qemud_driver *driver,
                           virDomainObjPtr vm)
{
    virCgroupPtr cgroup = NULL;
    int rc;
1416
    unsigned int i;
1417 1418 1419 1420
    const char *const *deviceACL =
        driver->cgroupDeviceACL ?
        (const char *const *)driver->cgroupDeviceACL :
        defaultDeviceACL;
1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432

    if (driver->cgroup == NULL)
        return 0; /* Not supported, so claim success */

    rc = virCgroupForDomain(driver->cgroup, vm->def->name, &cgroup, 1);
    if (rc != 0) {
        virReportSystemError(conn, -rc,
                             _("Unable to create cgroup for %s"),
                             vm->def->name);
        goto cleanup;
    }

1433 1434
    if (qemuCgroupControllerActive(driver, VIR_CGROUP_CONTROLLER_DEVICES)) {
        rc = virCgroupDenyAllDevices(cgroup);
1435
        if (rc != 0) {
1436 1437 1438 1439 1440
            if (rc == -EPERM) {
                VIR_WARN0("Group devices ACL is not accessible, disabling whitelisting");
                goto done;
            }

1441
            virReportSystemError(conn, -rc,
1442
                                 _("Unable to deny all devices for %s"), vm->def->name);
1443 1444 1445
            goto cleanup;
        }

1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459
        for (i = 0; i < vm->def->ndisks ; i++) {
            if (vm->def->disks[i]->type != VIR_DOMAIN_DISK_TYPE_BLOCK ||
                vm->def->disks[i]->src == NULL)
                continue;

            rc = virCgroupAllowDevicePath(cgroup,
                                          vm->def->disks[i]->src);
            if (rc != 0) {
                virReportSystemError(conn, -rc,
                                     _("Unable to allow device %s for %s"),
                                     vm->def->disks[i]->src, vm->def->name);
                goto cleanup;
            }
        }
1460

1461
        rc = virCgroupAllowDeviceMajor(cgroup, 'c', DEVICE_PTY_MAJOR);
1462 1463
        if (rc != 0) {
            virReportSystemError(conn, -rc, "%s",
1464
                                 _("unable to allow /dev/pts/ devices"));
1465 1466 1467
            goto cleanup;
        }

1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486
        if (vm->def->nsounds) {
            rc = virCgroupAllowDeviceMajor(cgroup, 'c', DEVICE_SND_MAJOR);
            if (rc != 0) {
                virReportSystemError(conn, -rc, "%s",
                                     _("unable to allow /dev/snd/ devices"));
                goto cleanup;
            }
        }

        for (i = 0; deviceACL[i] != NULL ; i++) {
            rc = virCgroupAllowDevicePath(cgroup,
                                          deviceACL[i]);
            if (rc < 0 &&
                rc != -ENOENT) {
                virReportSystemError(conn, -rc,
                                     _("unable to allow device %s"),
                                     deviceACL[i]);
                goto cleanup;
            }
1487 1488 1489 1490
        }
    }

done:
1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 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
    virCgroupFree(&cgroup);
    return 0;

cleanup:
    if (cgroup) {
        virCgroupRemove(cgroup);
        virCgroupFree(&cgroup);
    }
    return -1;
}


static int qemuRemoveCgroup(virConnectPtr conn,
                            struct qemud_driver *driver,
                            virDomainObjPtr vm)
{
    virCgroupPtr cgroup;
    int rc;

    if (driver->cgroup == NULL)
        return 0; /* Not supported, so claim success */

    rc = virCgroupForDomain(driver->cgroup, vm->def->name, &cgroup, 0);
    if (rc != 0) {
        qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                         _("Unable to find cgroup for %s\n"),
                         vm->def->name);
        return rc;
    }

    rc = virCgroupRemove(cgroup);
    virCgroupFree(&cgroup);
    return rc;
}

static int qemuAddToCgroup(struct qemud_driver *driver,
                           virDomainDefPtr def)
{
    virCgroupPtr cgroup = NULL;
    int ret = -1;
    int rc;

    if (driver->cgroup == NULL)
        return 0; /* Not supported, so claim success */

    rc = virCgroupForDomain(driver->cgroup, def->name, &cgroup, 0);
    if (rc != 0) {
        virReportSystemError(NULL, -rc,
                             _("unable to find cgroup for domain %s"),
                             def->name);
        goto cleanup;
    }

    rc = virCgroupAddTask(cgroup, getpid());
    if (rc != 0) {
        virReportSystemError(NULL, -rc,
                             _("unable to add domain %s task %d to cgroup"),
                             def->name, getpid());
        goto cleanup;
    }

    ret = 0;

cleanup:
    virCgroupFree(&cgroup);
    return ret;
}


1560 1561 1562 1563 1564 1565 1566 1567 1568
static int qemudDomainSetSecurityLabel(virConnectPtr conn, struct qemud_driver *driver, virDomainObjPtr vm)
{
    if (vm->def->seclabel.label != NULL)
        if (driver->securityDriver && driver->securityDriver->domainSetSecurityLabel)
            return driver->securityDriver->domainSetSecurityLabel(conn, driver->securityDriver,
                                                                 vm);
    return 0;
}

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

#ifdef __linux__
static int qemuDomainSetHostdevUSBOwnership(virConnectPtr conn,
                                            virDomainHostdevDefPtr def,
                                            uid_t uid, gid_t gid)
{
    char *usbpath = NULL;

    /* XXX what todo for USB devs assigned based on product/vendor ? Doom :-( */
    if (!def->source.subsys.u.usb.bus ||
        !def->source.subsys.u.usb.device)
        return 0;

    if (virAsprintf(&usbpath, "/dev/bus/usb/%03d/%03d",
                    def->source.subsys.u.usb.bus,
                    def->source.subsys.u.usb.device) < 0) {
        virReportOOMError(conn);
        return -1;
    }

    VIR_DEBUG("Setting ownership on %s to %d:%d", usbpath, uid, gid);
    if (chown(usbpath, uid, gid) < 0) {
        virReportSystemError(conn, errno, _("cannot set ownership on %s"), usbpath);
        VIR_FREE(usbpath);
        return -1;
    }
    VIR_FREE(usbpath);

    return 0;
}

static int qemuDomainSetHostdevPCIOwnership(virConnectPtr conn,
                                            virDomainHostdevDefPtr def,
                                            uid_t uid, gid_t gid)
{
    char *pcidir = NULL;
    char *file = NULL;
    DIR *dir = NULL;
    int ret = -1;
    struct dirent *ent;

    if (virAsprintf(&pcidir, "/sys/bus/pci/devices/%04x:%02x:%02x.%x",
                    def->source.subsys.u.pci.domain,
                    def->source.subsys.u.pci.bus,
                    def->source.subsys.u.pci.slot,
                    def->source.subsys.u.pci.function) < 0) {
        virReportOOMError(conn);
        goto cleanup;
    }

    if (!(dir = opendir(pcidir))) {
        virReportSystemError(conn, errno,
                             _("cannot open %s"), pcidir);
        goto cleanup;
    }

    while ((ent = readdir(dir)) != NULL) {
        /* QEMU device assignment requires:
         *   $PCIDIR/config, $PCIDIR/resource, $PCIDIR/resourceNNN, $PCIDIR/rom
         */
        if (STREQ(ent->d_name, "config") ||
            STRPREFIX(ent->d_name, "resource") ||
            STREQ(ent->d_name, "rom")) {
            if (virAsprintf(&file, "%s/%s", pcidir, ent->d_name) < 0) {
                virReportOOMError(conn);
                goto cleanup;
            }
            VIR_DEBUG("Setting ownership on %s to %d:%d", file, uid, gid);
            if (chown(file, uid, gid) < 0) {
                virReportSystemError(conn, errno, _("cannot set ownership on %s"), file);
                goto cleanup;
            }
            VIR_FREE(file);
        }
    }

    ret = 0;

cleanup:
    if (dir)
        closedir(dir);
    VIR_FREE(file);
    VIR_FREE(pcidir);
    return ret;
}
#endif


static int qemuDomainSetHostdevOwnership(virConnectPtr conn,
                                         virDomainHostdevDefPtr def,
                                         uid_t uid, gid_t gid)
{
    if (def->mode != VIR_DOMAIN_HOSTDEV_MODE_SUBSYS)
        return 0;

#ifdef __linux__
    switch (def->source.subsys.type) {
    case VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_USB:
        return qemuDomainSetHostdevUSBOwnership(conn, def, uid, gid);

    case VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_PCI:
        return qemuDomainSetHostdevPCIOwnership(conn, def, uid, gid);

    }
    return 0;
#else
    qemudReportError(conn, NULL, NULL, "%s",
                     _("unable to set host device ownership on this platform"));
    return -1;
#endif

}

static int qemuDomainSetDiskOwnership(virConnectPtr conn,
                                      virDomainDiskDefPtr def,
                                      uid_t uid, gid_t gid)
{

    if (!def->src)
        return 0;

    VIR_DEBUG("Setting ownership on %s to %d:%d", def->src, uid, gid);
    if (chown(def->src, uid, gid) < 0) {
        virReportSystemError(conn, errno, _("cannot set ownership on %s"),
                             def->src);
        return -1;
    }
    return 0;
}

static int qemuDomainSetDeviceOwnership(virConnectPtr conn,
                                        struct qemud_driver *driver,
                                        virDomainDeviceDefPtr def,
                                        int restore)
{
    uid_t uid;
    gid_t gid;

    if (!driver->privileged)
        return 0;

    /* short circuit case of root:root */
    if (!driver->user && !driver->group)
        return 0;

    uid = restore ? 0 : driver->user;
    gid = restore ? 0 : driver->group;

    switch (def->type) {
    case VIR_DOMAIN_DEVICE_DISK:
        if (restore &&
            (def->data.disk->readonly || def->data.disk->shared))
            return 0;

        return qemuDomainSetDiskOwnership(conn, def->data.disk, uid, gid);

    case VIR_DOMAIN_DEVICE_HOSTDEV:
        return qemuDomainSetHostdevOwnership(conn, def->data.hostdev, uid, gid);
    }

    return 0;
}

static int qemuDomainSetAllDeviceOwnership(virConnectPtr conn,
                                           struct qemud_driver *driver,
                                           virDomainDefPtr def,
                                           int restore)
{
    int i;
    uid_t uid;
    gid_t gid;

    if (!driver->privileged)
        return 0;

    /* short circuit case of root:root */
    if (!driver->user && !driver->group)
        return 0;

    uid = restore ? 0 : driver->user;
    gid = restore ? 0 : driver->group;

    for (i = 0 ; i < def->ndisks ; i++) {
        if (restore &&
            (def->disks[i]->readonly || def->disks[i]->shared))
            continue;

        if (qemuDomainSetDiskOwnership(conn, def->disks[i], uid, gid) < 0)
            return -1;
    }

    for (i = 0 ; i < def->nhostdevs ; i++) {
        if (qemuDomainSetHostdevOwnership(conn, def->hostdevs[i], uid, gid) < 0)
            return -1;
    }

    return 0;
}

1768 1769 1770
static virDomainPtr qemudDomainLookupByName(virConnectPtr conn,
                                            const char *name);

1771 1772 1773 1774
struct qemudHookData {
    virConnectPtr conn;
    virDomainObjPtr vm;
    struct qemud_driver *driver;
1775 1776 1777
};

static int qemudSecurityHook(void *data) {
1778 1779 1780 1781
    struct qemudHookData *h = data;

    if (qemuAddToCgroup(h->driver, h->vm->def) < 0)
        return -1;
1782 1783 1784

    if (qemudDomainSetSecurityLabel(h->conn, h->driver, h->vm) < 0) {
        qemudReportError(h->conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
1785
                         "%s", _("Failed to set security label"));
1786 1787 1788 1789 1790 1791 1792 1793
        return -1;
    }

    if (h->driver->privileged) {
        DEBUG("Dropping privileges of VM to %d:%d", h->driver->user, h->driver->group);

        if (qemuDomainSetAllDeviceOwnership(h->conn, h->driver, h->vm->def, 0) < 0)
            return -1;
1794

1795 1796 1797 1798 1799
        if (h->driver->group) {
            if (setregid(h->driver->group, h->driver->group) < 0) {
                virReportSystemError(NULL, errno,
                                     _("cannot change to '%d' group"),
                                     h->driver->group);
1800
                return -1;
1801
            }
1802
        }
1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813
        if (h->driver->user) {
            if (setreuid(h->driver->user, h->driver->user) < 0) {
                virReportSystemError(NULL, errno,
                                     _("cannot change to '%d' user"),
                                     h->driver->user);
                return -1;
            }
        }
    }

    return 0;
1814 1815
}

1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833
static int
qemuPrepareMonitorChr(virConnectPtr conn,
                      struct qemud_driver *driver,
                      virDomainChrDefPtr monitor_chr,
                      const char *vm)
{
    monitor_chr->type = VIR_DOMAIN_CHR_TYPE_UNIX;
    monitor_chr->data.nix.listen = 1;

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

    return 0;
}

1834 1835
static int qemudStartVMDaemon(virConnectPtr conn,
                              struct qemud_driver *driver,
1836
                              virDomainObjPtr vm,
1837 1838
                              const char *migrateFrom,
                              int stdin_fd) {
1839
    const char **argv = NULL, **tmp;
1840
    const char **progenv = NULL;
1841
    int i, ret;
1842
    struct stat sb;
1843 1844
    int *tapfds = NULL;
    int ntapfds = 0;
1845
    unsigned int qemuCmdFlags;
1846
    fd_set keepfd;
1847
    const char *emulator;
G
Guido Günther 已提交
1848
    pid_t child;
1849
    int pos = -1;
1850
    char ebuf[1024];
1851
    char *pidfile = NULL;
1852
    int logfile;
1853

1854
    struct qemudHookData hookData;
1855 1856 1857 1858
    hookData.conn = conn;
    hookData.vm = vm;
    hookData.driver = driver;

1859
    FD_ZERO(&keepfd);
1860

1861
    if (virDomainIsActive(vm)) {
1862
        qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_INVALID,
1863
                         "%s", _("VM is already active"));
1864 1865 1866
        return -1;
    }

1867 1868 1869 1870 1871 1872 1873 1874
    /* If you are using a SecurityDriver with dynamic labelling,
       then generate a security label for isolation */
    if (vm->def->seclabel.type == VIR_DOMAIN_SECLABEL_DYNAMIC &&
        driver->securityDriver &&
        driver->securityDriver->domainGenSecurityLabel &&
        driver->securityDriver->domainGenSecurityLabel(conn, vm) < 0)
        return -1;

1875 1876 1877
    /* Ensure no historical cgroup for this VM is lieing around bogus settings */
    qemuRemoveCgroup(conn, driver, vm);

1878 1879 1880
    if ((vm->def->ngraphics == 1) &&
        vm->def->graphics[0]->type == VIR_DOMAIN_GRAPHICS_TYPE_VNC &&
        vm->def->graphics[0]->data.vnc.autoport) {
1881
        int port = qemudNextFreeVNCPort(driver);
1882
        if (port < 0) {
1883
            qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
1884
                             "%s", _("Unable to find an unused VNC port"));
1885
            goto cleanup;
1886
        }
1887
        vm->def->graphics[0]->data.vnc.port = port;
1888
    }
1889

1890
    if (virFileMakePath(driver->logDir) < 0) {
1891 1892 1893
        virReportSystemError(conn, errno,
                             _("cannot create log directory %s"),
                             driver->logDir);
1894
        goto cleanup;
1895 1896
    }

1897
    if ((logfile = qemudLogFD(conn, driver, vm->def->name)) < 0)
1898
        goto cleanup;
1899

1900 1901
    emulator = vm->def->emulator;

1902 1903 1904 1905
    /* Make sure the binary we are about to try exec'ing exists.
     * Technically we could catch the exec() failure, but that's
     * in a sub-process so its hard to feed back a useful error
     */
1906
    if (stat(emulator, &sb) < 0) {
1907 1908 1909
        virReportSystemError(conn, errno,
                             _("Cannot find QEMU binary %s"),
                             emulator);
1910
        goto cleanup;
1911 1912
    }

1913
    if (qemudExtractVersionInfo(emulator,
1914
                                NULL,
1915
                                &qemuCmdFlags) < 0)
1916
        goto cleanup;
1917

1918 1919 1920
    if (qemuSetupCgroup(conn, driver, vm) < 0)
        goto cleanup;

1921
    if (qemuPrepareHostDevices(conn, vm->def) < 0)
1922
        goto cleanup;
1923

1924 1925 1926 1927 1928
    if (VIR_ALLOC(vm->monitor_chr) < 0) {
        virReportOOMError(conn);
        goto cleanup;
    }

1929 1930
    if (qemuPrepareMonitorChr(conn, driver, vm->monitor_chr, vm->def->name) < 0)
        goto cleanup;
1931

D
Daniel P. Berrange 已提交
1932 1933 1934 1935 1936 1937 1938
    if ((ret = virFileDeletePid(driver->stateDir, vm->def->name)) != 0) {
        virReportSystemError(conn, ret,
                             _("Cannot remove stale PID file for %s"),
                             vm->def->name);
        goto cleanup;
    }

1939 1940 1941 1942 1943
    if (!(pidfile = virFilePid(driver->stateDir, vm->def->name))) {
        virReportSystemError(conn, errno,
                             "%s", _("Failed to build pidfile path."));
        goto cleanup;
    }
D
Daniel P. Berrange 已提交
1944

1945
    vm->def->id = driver->nextvmid++;
1946
    if (qemudBuildCommandLine(conn, driver, vm->def, vm->monitor_chr,
1947
                              qemuCmdFlags, &argv, &progenv,
1948 1949
                              &tapfds, &ntapfds, migrateFrom) < 0)
        goto cleanup;
1950

1951 1952
    tmp = progenv;
    while (*tmp) {
1953
        if (safewrite(logfile, *tmp, strlen(*tmp)) < 0)
1954
            VIR_WARN(_("Unable to write envv to logfile: %s\n"),
1955
                     virStrerror(errno, ebuf, sizeof ebuf));
1956
        if (safewrite(logfile, " ", 1) < 0)
1957
            VIR_WARN(_("Unable to write envv to logfile: %s\n"),
1958
                     virStrerror(errno, ebuf, sizeof ebuf));
1959 1960
        tmp++;
    }
1961 1962
    tmp = argv;
    while (*tmp) {
1963
        if (safewrite(logfile, *tmp, strlen(*tmp)) < 0)
1964
            VIR_WARN(_("Unable to write argv to logfile: %s\n"),
1965
                     virStrerror(errno, ebuf, sizeof ebuf));
1966
        if (safewrite(logfile, " ", 1) < 0)
1967
            VIR_WARN(_("Unable to write argv to logfile: %s\n"),
1968
                     virStrerror(errno, ebuf, sizeof ebuf));
1969 1970
        tmp++;
    }
1971
    if (safewrite(logfile, "\n", 1) < 0)
1972
        VIR_WARN(_("Unable to write argv to logfile: %s\n"),
1973
                 virStrerror(errno, ebuf, sizeof ebuf));
1974

1975
    if ((pos = lseek(logfile, 0, SEEK_END)) < 0)
1976
        VIR_WARN(_("Unable to seek to end of logfile: %s\n"),
1977
                 virStrerror(errno, ebuf, sizeof ebuf));
1978

1979 1980 1981
    for (i = 0 ; i < ntapfds ; i++)
        FD_SET(tapfds[i], &keepfd);

1982
    ret = virExecDaemonize(conn, argv, progenv, &keepfd, &child,
1983
                           stdin_fd, &logfile, &logfile,
1984
                           VIR_EXEC_NONBLOCK | VIR_EXEC_CLEAR_CAPS,
1985 1986 1987
                           qemudSecurityHook, &hookData,
                           pidfile);
    VIR_FREE(pidfile);
G
Guido Günther 已提交
1988 1989 1990

    /* wait for qemu process to to show up */
    if (ret == 0) {
1991
        if (virFileReadPid(driver->stateDir, vm->def->name, &vm->pid)) {
1992
            qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
1993
                             _("Domain %s didn't show up\n"), vm->def->name);
1994
            ret = -1;
G
Guido Günther 已提交
1995
        }
1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006
    } else if (ret == -2) {
        /* The virExec process that launches the daemon failed. Pending on
         * 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;
    }
2007 2008

    vm->state = migrateFrom ? VIR_DOMAIN_PAUSED : VIR_DOMAIN_RUNNING;
2009

2010
    for (i = 0 ; argv[i] ; i++)
2011 2012
        VIR_FREE(argv[i]);
    VIR_FREE(argv);
2013

2014 2015 2016 2017
    for (i = 0 ; progenv[i] ; i++)
        VIR_FREE(progenv[i]);
    VIR_FREE(progenv);

2018 2019 2020
    if (tapfds) {
        for (i = 0 ; i < ntapfds ; i++) {
            close(tapfds[i]);
2021
        }
2022
        VIR_FREE(tapfds);
2023 2024
    }

2025 2026 2027 2028 2029 2030 2031 2032
    if (ret == -1)
        goto cleanup;

    if ((qemudWaitForMonitor(conn, driver, vm, pos) < 0) ||
        (qemudDetectVcpuPIDs(conn, vm) < 0) ||
        (qemudInitCpus(conn, vm, migrateFrom) < 0) ||
        (qemudInitPasswords(conn, driver, vm) < 0) ||
        (qemudDomainSetMemoryBalloon(conn, vm, vm->def->memory) < 0) ||
2033
        (virDomainSaveStatus(conn, driver->stateDir, vm) < 0)) {
2034 2035 2036
        qemudShutdownVMDaemon(conn, driver, vm);
        ret = -1;
        /* No need for 'goto cleanup' now since qemudShutdownVMDaemon does enough */
2037 2038
    }

2039
    return ret;
2040 2041 2042 2043 2044 2045 2046

cleanup:
    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);
    }
2047
    qemuRemoveCgroup(conn, driver, vm);
2048 2049 2050 2051
    if ((vm->def->ngraphics == 1) &&
        vm->def->graphics[0]->type == VIR_DOMAIN_GRAPHICS_TYPE_VNC &&
        vm->def->graphics[0]->data.vnc.autoport)
        vm->def->graphics[0]->data.vnc.port = -1;
2052 2053
    if (logfile != -1)
        close(logfile);
2054 2055
    vm->def->id = -1;
    return -1;
2056 2057 2058
}


2059
static void qemudShutdownVMDaemon(virConnectPtr conn,
2060 2061
                                  struct qemud_driver *driver,
                                  virDomainObjPtr vm) {
D
Daniel P. Berrange 已提交
2062
    int ret;
2063
    int retries = 0;
D
Daniel P. Berrange 已提交
2064

2065
    if (!virDomainIsActive(vm))
2066
        return;
2067

2068
    VIR_DEBUG(_("Shutting down VM '%s'\n"), vm->def->name);
2069

G
Guido Günther 已提交
2070 2071
    if (virKillProcess(vm->pid, 0) == 0 &&
        virKillProcess(vm->pid, SIGTERM) < 0)
2072 2073 2074
        virReportSystemError(conn, errno,
                             _("Failed to send SIGTERM to %s (%d)"),
                             vm->def->name, vm->pid);
2075

2076 2077 2078
    if (vm->monitorWatch != -1) {
        virEventRemoveHandle(vm->monitorWatch);
        vm->monitorWatch = -1;
2079
    }
2080 2081 2082 2083 2084

    if (vm->monitor != -1)
        close(vm->monitor);
    vm->monitor = -1;

G
Guido Günther 已提交
2085 2086
    /* shut it off for sure */
    virKillProcess(vm->pid, SIGKILL);
2087

2088 2089 2090 2091
    /* Reset Security Labels */
    if (driver->securityDriver)
        driver->securityDriver->domainRestoreSecurityLabel(conn, vm);

2092 2093 2094 2095 2096 2097 2098
    /* 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);
    }

2099 2100 2101 2102
    if (qemuDomainSetAllDeviceOwnership(conn, driver, vm->def, 1) < 0)
        VIR_WARN("Failed to restore all device ownership for %s",
                 vm->def->name);

2103 2104 2105 2106 2107 2108 2109 2110 2111 2112
retry:
    if ((ret = qemuRemoveCgroup(conn, driver, vm)) < 0) {
        if (ret == -EBUSY && (retries++ < 5)) {
            usleep(200*1000);
            goto retry;
        }
        VIR_WARN("Failed to remove cgroup for %s",
                 vm->def->name);
    }

2113
    if (qemudRemoveDomainStatus(conn, driver, vm) < 0) {
2114
        VIR_WARN(_("Failed to remove domain status for %s"),
2115 2116
                 vm->def->name);
    }
D
Daniel P. Berrange 已提交
2117 2118 2119 2120 2121 2122
    if ((ret = virFileDeletePid(driver->stateDir, vm->def->name)) != 0) {
        char ebuf[1024];
        VIR_WARN(_("Failed to remove PID file for %s: %s"),
                 vm->def->name, virStrerror(errno, ebuf, sizeof ebuf));
    }

2123
    vm->pid = -1;
2124
    vm->def->id = -1;
2125
    vm->state = VIR_DOMAIN_SHUTOFF;
2126
    VIR_FREE(vm->vcpupids);
2127
    vm->nvcpupids = 0;
2128 2129

    if (vm->newDef) {
2130
        virDomainDefFree(vm->def);
2131
        vm->def = vm->newDef;
2132
        vm->def->id = -1;
2133 2134 2135 2136 2137
        vm->newDef = NULL;
    }
}


2138
static void
2139
qemudDispatchVMEvent(int watch, int fd, int events, void *opaque) {
2140
    struct qemud_driver *driver = opaque;
2141
    virDomainObjPtr vm = NULL;
2142
    virDomainEventPtr event = NULL;
2143
    unsigned int i;
2144
    int quit = 0, failed = 0;
2145

2146
    qemuDriverLock(driver);
2147
    for (i = 0 ; i < driver->domains.count ; i++) {
2148 2149 2150
        virDomainObjPtr tmpvm = driver->domains.objs[i];
        virDomainObjLock(tmpvm);
        if (virDomainIsActive(tmpvm) &&
2151
            tmpvm->monitorWatch == watch) {
2152
            vm = tmpvm;
2153
            break;
2154
        }
2155
        virDomainObjUnlock(tmpvm);
2156 2157 2158
    }

    if (!vm)
2159
        goto cleanup;
2160

2161
    if (vm->monitor != fd) {
2162 2163
        failed = 1;
    } else {
2164
        if (events & (VIR_EVENT_HANDLE_HANGUP | VIR_EVENT_HANDLE_ERROR))
2165
            quit = 1;
2166
        else {
2167 2168
            VIR_ERROR(_("unhandled fd event %d for %s"),
                      events, vm->def->name);
2169
            failed = 1;
2170
        }
2171 2172
    }

2173
    if (failed || quit) {
2174 2175 2176 2177 2178
        event = virDomainEventNewFromObj(vm,
                                         VIR_DOMAIN_EVENT_STOPPED,
                                         quit ?
                                         VIR_DOMAIN_EVENT_STOPPED_SHUTDOWN :
                                         VIR_DOMAIN_EVENT_STOPPED_FAILED);
2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189
        qemudShutdownVMDaemon(NULL, driver, vm);
        if (!vm->persistent) {
            virDomainRemoveInactive(&driver->domains,
                                    vm);
            vm = NULL;
        }
    }

cleanup:
    if (vm)
        virDomainObjUnlock(vm);
2190 2191
    if (event)
        qemuDomainEventQueue(driver, event);
2192
    qemuDriverUnlock(driver);
2193 2194
}

2195 2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216

/* Throw away any data available on the monitor
 * This is done before executing a command, in order
 * to allow re-synchronization if something went badly
 * wrong in the past. it also deals with problem of
 * QEMU *sometimes* re-printing its initial greeting
 * when we reconnect to the monitor after restarts.
 */
static void
qemuMonitorDiscardPendingData(virDomainObjPtr vm) {
    char buf[1024];
    int ret = 0;

    /* Monitor is non-blocking, so just loop till we
     * get -1 or 0. Don't bother with detecting
     * errors, since we'll deal with that better later */
    do {
        ret = read(vm->monitor, buf, sizeof (buf)-1);
    } while (ret > 0);
}


2217
static int
2218 2219 2220 2221 2222
qemudMonitorCommandExtra(const virDomainObjPtr vm,
                         const char *cmd,
                         const char *extra,
                         const char *extraPrompt,
                         char **reply) {
D
Daniel P. Berrange 已提交
2223 2224
    int size = 0;
    char *buf = NULL;
2225
    size_t cmdlen = strlen(cmd);
2226
    size_t extralen = extra ? strlen(extra) : 0;
D
Daniel P. Berrange 已提交
2227

2228 2229
    qemuMonitorDiscardPendingData(vm);

2230 2231 2232
    if (safewrite(vm->monitor, cmd, cmdlen) != cmdlen)
        return -1;
    if (safewrite(vm->monitor, "\r", 1) != 1)
D
Daniel P. Berrange 已提交
2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244
        return -1;

    *reply = NULL;

    for (;;) {
        struct pollfd fd = { vm->monitor, POLLIN | POLLERR | POLLHUP, 0 };
        char *tmp;

        /* Read all the data QEMU has sent thus far */
        for (;;) {
            char data[1024];
            int got = read(vm->monitor, data, sizeof(data));
D
Daniel P. Berrange 已提交
2245

2246 2247
            if (got == 0)
                goto error;
D
Daniel P. Berrange 已提交
2248 2249 2250 2251 2252
            if (got < 0) {
                if (errno == EINTR)
                    continue;
                if (errno == EAGAIN)
                    break;
2253
                goto error;
2254
            }
2255
            if (VIR_REALLOC_N(buf, size+got+1) < 0)
2256 2257
                goto error;

D
Daniel P. Berrange 已提交
2258 2259 2260 2261
            memmove(buf+size, data, got);
            buf[size+got] = '\0';
            size += got;
        }
2262

D
Daniel P. Berrange 已提交
2263
        /* Look for QEMU prompt to indicate completion */
2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288
        if (buf) {
            if (extra) {
                if (strstr(buf, extraPrompt) != NULL) {
                    if (safewrite(vm->monitor, extra, extralen) != extralen)
                        return -1;
                    if (safewrite(vm->monitor, "\r", 1) != 1)
                        return -1;
                    extra = NULL;
                }
            } else if ((tmp = strstr(buf, QEMU_CMD_PROMPT)) != NULL) {
                char *commptr = NULL, *nlptr = NULL;
                /* Preserve the newline */
                tmp[1] = '\0';

                /* The monitor doesn't dump clean output after we have written to
                 * it. Every character we write dumps a bunch of useless stuff,
                 * so the result looks like "cXcoXcomXcommXcommaXcommanXcommand"
                 * Try to throw away everything before the first full command
                 * occurence, and inbetween the command and the newline starting
                 * the response
                 */
                if ((commptr = strstr(buf, cmd)))
                    memmove(buf, commptr, strlen(commptr)+1);
                if ((nlptr = strchr(buf, '\n')))
                    memmove(buf+strlen(cmd), nlptr, strlen(nlptr)+1);
C
Cole Robinson 已提交
2289

2290 2291
                break;
            }
D
Daniel P. Berrange 已提交
2292 2293 2294 2295 2296 2297
        }
    pollagain:
        /* Need to wait for more data */
        if (poll(&fd, 1, -1) < 0) {
            if (errno == EINTR)
                goto pollagain;
2298
            goto error;
D
Daniel P. Berrange 已提交
2299 2300 2301 2302
        }
    }
    *reply = buf;
    return 0;
2303 2304

 error:
2305
    VIR_FREE(buf);
2306
    return -1;
D
Daniel P. Berrange 已提交
2307 2308
}

2309 2310 2311 2312 2313 2314 2315 2316
static int
qemudMonitorCommand(const virDomainObjPtr vm,
                    const char *cmd,
                    char **reply) {
    return qemudMonitorCommandExtra(vm, cmd, NULL, NULL, reply);
}


2317

2318
static virDrvOpenStatus qemudOpen(virConnectPtr conn,
2319
                                  virConnectAuthPtr auth ATTRIBUTE_UNUSED,
2320
                                  int flags ATTRIBUTE_UNUSED) {
2321
    if (conn->uri == NULL) {
2322 2323 2324
        if (qemu_driver == NULL)
            return VIR_DRV_OPEN_DECLINED;

2325
        conn->uri = xmlParseURI(qemu_driver->privileged ?
2326 2327
                                "qemu:///system" :
                                "qemu:///session");
2328
        if (!conn->uri) {
2329
            virReportOOMError(conn);
2330 2331
            return VIR_DRV_OPEN_ERROR;
        }
2332 2333 2334 2335 2336 2337 2338 2339 2340 2341
    } else {
        /* If URI isn't 'qemu' its definitely not for us */
        if (conn->uri->scheme == NULL ||
            STRNEQ(conn->uri->scheme, "qemu"))
            return VIR_DRV_OPEN_DECLINED;

        /* Allow remote driver to deal with URIs with hostname server */
        if (conn->uri->server != NULL)
            return VIR_DRV_OPEN_DECLINED;

2342 2343 2344 2345 2346 2347
        if (qemu_driver == NULL) {
            qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR, "%s",
                             _("qemu state driver is not active"));
            return VIR_DRV_OPEN_ERROR;
        }

2348
        if (qemu_driver->privileged) {
2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363
            if (STRNEQ (conn->uri->path, "/system") &&
                STRNEQ (conn->uri->path, "/session")) {
                qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                                 _("unexpected QEMU URI path '%s', try qemu:///system"),
                                 conn->uri->path);
                return VIR_DRV_OPEN_ERROR;
            }
        } else {
            if (STRNEQ (conn->uri->path, "/session")) {
                qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                                 _("unexpected QEMU URI path '%s', try qemu:///session"),
                                 conn->uri->path);
                return VIR_DRV_OPEN_ERROR;
            }
        }
2364 2365 2366 2367 2368 2369 2370
    }
    conn->privateData = qemu_driver;

    return VIR_DRV_OPEN_SUCCESS;
}

static int qemudClose(virConnectPtr conn) {
2371
    struct qemud_driver *driver = conn->privateData;
2372 2373

    /* Get rid of callbacks registered for this conn */
2374
    qemuDriverLock(driver);
2375
    virDomainEventCallbackListRemoveConn(conn, driver->domainEventCallbacks);
2376
    qemuDriverUnlock(driver);
2377 2378 2379 2380 2381 2382

    conn->privateData = NULL;

    return 0;
}

D
Daniel Veillard 已提交
2383 2384 2385 2386 2387 2388 2389 2390 2391 2392
/* Which features are supported by this driver? */
static int
qemudSupportsFeature (virConnectPtr conn ATTRIBUTE_UNUSED, int feature)
{
    switch (feature) {
    case VIR_DRV_FEATURE_MIGRATION_V2: return 1;
    default: return 0;
    }
}

2393
static const char *qemudGetType(virConnectPtr conn ATTRIBUTE_UNUSED) {
2394
    return "QEMU";
2395 2396
}

2397 2398 2399 2400 2401

static int kvmGetMaxVCPUs(void) {
    int maxvcpus = 1;

    int r, fd;
2402

2403 2404
    fd = open(KVM_DEVICE, O_RDONLY);
    if (fd < 0) {
2405 2406
        virReportSystemError(NULL, errno, _("Unable to open %s"), KVM_DEVICE);
        return -1;
2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417
    }

    r = ioctl(fd, KVM_CHECK_EXTENSION, KVM_CAP_NR_VCPUS);
    if (r > 0)
        maxvcpus = r;

    close(fd);
    return maxvcpus;
}


2418
static int qemudGetMaxVCPUs(virConnectPtr conn, const char *type) {
2419 2420 2421
    if (!type)
        return 16;

2422
    if (STRCASEEQ(type, "qemu"))
2423 2424
        return 16;

2425
    if (STRCASEEQ(type, "kvm"))
2426
        return kvmGetMaxVCPUs();
2427

2428
    if (STRCASEEQ(type, "kqemu"))
2429
        return 1;
2430 2431 2432

    qemudReportError(conn, NULL, NULL, VIR_ERR_INVALID_ARG,
                     _("unknown type '%s'"), type);
2433 2434 2435
    return -1;
}

2436

2437
static char *qemudGetCapabilities(virConnectPtr conn) {
2438
    struct qemud_driver *driver = conn->privateData;
2439
    virCapsPtr caps;
2440
    char *xml = NULL;
2441

2442
    qemuDriverLock(driver);
2443
    if ((caps = qemudCapsInit(qemu_driver->caps)) == NULL) {
2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454
        virReportOOMError(conn);
        goto cleanup;
    }

    if (qemu_driver->securityDriver &&
        qemudSecurityCapsInit(qemu_driver->securityDriver, caps) < 0) {
        virCapabilitiesFree(caps);
        virReportOOMError(conn);
        goto cleanup;
    }

2455
    virCapabilitiesFree(qemu_driver->caps);
2456 2457 2458
    qemu_driver->caps = caps;

    if ((xml = virCapabilitiesFormatXML(driver->caps)) == NULL)
2459
        virReportOOMError(conn);
2460 2461

cleanup:
2462
    qemuDriverUnlock(driver);
2463

2464
    return xml;
2465 2466 2467
}


D
Daniel P. Berrange 已提交
2468 2469 2470
static int qemudGetProcessInfo(unsigned long long *cpuTime, int pid) {
    char proc[PATH_MAX];
    FILE *pidinfo;
2471
    unsigned long long usertime, systime;
D
Daniel P. Berrange 已提交
2472 2473 2474 2475 2476 2477

    if (snprintf(proc, sizeof(proc), "/proc/%d/stat", pid) >= (int)sizeof(proc)) {
        return -1;
    }

    if (!(pidinfo = fopen(proc, "r"))) {
2478
        /*printf("cannot read pid info");*/
D
Daniel P. Berrange 已提交
2479 2480 2481 2482 2483
        /* VM probably shut down, so fake 0 */
        *cpuTime = 0;
        return 0;
    }

2484
    if (fscanf(pidinfo, "%*d %*s %*c %*d %*d %*d %*d %*d %*u %*u %*u %*u %*u %llu %llu", &usertime, &systime) != 2) {
2485
        qemudDebug("not enough arg");
D
Daniel P. Berrange 已提交
2486 2487 2488 2489 2490 2491 2492 2493
        return -1;
    }

    /* We got jiffies
     * We want nanoseconds
     * _SC_CLK_TCK is jiffies per second
     * So calulate thus....
     */
2494
    *cpuTime = 1000ull * 1000ull * 1000ull * (usertime + systime) / (unsigned long long)sysconf(_SC_CLK_TCK);
D
Daniel P. Berrange 已提交
2495

2496
    qemudDebug("Got %llu %llu %llu", usertime, systime, *cpuTime);
D
Daniel P. Berrange 已提交
2497 2498 2499 2500 2501 2502 2503

    fclose(pidinfo);

    return 0;
}


2504
static virDomainPtr qemudDomainLookupByID(virConnectPtr conn,
2505
                                          int id) {
2506 2507 2508 2509
    struct qemud_driver *driver = conn->privateData;
    virDomainObjPtr vm;
    virDomainPtr dom = NULL;

2510
    qemuDriverLock(driver);
2511
    vm  = virDomainFindByID(&driver->domains, id);
2512
    qemuDriverUnlock(driver);
2513 2514

    if (!vm) {
2515 2516
        qemudReportError(conn, NULL, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching id %d"), id);
2517
        goto cleanup;
2518 2519
    }

2520
    dom = virGetDomain(conn, vm->def->name, vm->def->uuid);
2521
    if (dom) dom->id = vm->def->id;
2522 2523

cleanup:
2524 2525
    if (vm)
        virDomainObjUnlock(vm);
2526 2527
    return dom;
}
2528

2529
static virDomainPtr qemudDomainLookupByUUID(virConnectPtr conn,
2530
                                            const unsigned char *uuid) {
2531 2532 2533
    struct qemud_driver *driver = conn->privateData;
    virDomainObjPtr vm;
    virDomainPtr dom = NULL;
2534

2535
    qemuDriverLock(driver);
2536
    vm = virDomainFindByUUID(&driver->domains, uuid);
2537 2538
    qemuDriverUnlock(driver);

2539
    if (!vm) {
2540 2541 2542
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(uuid, uuidstr);
        qemudReportError(conn, NULL, NULL, VIR_ERR_NO_DOMAIN,
2543
                         _("no domain with matching uuid '%s'"), uuidstr);
2544
        goto cleanup;
2545 2546
    }

2547
    dom = virGetDomain(conn, vm->def->name, vm->def->uuid);
2548
    if (dom) dom->id = vm->def->id;
2549 2550

cleanup:
2551 2552
    if (vm)
        virDomainObjUnlock(vm);
2553 2554
    return dom;
}
2555

2556
static virDomainPtr qemudDomainLookupByName(virConnectPtr conn,
2557
                                            const char *name) {
2558 2559 2560
    struct qemud_driver *driver = conn->privateData;
    virDomainObjPtr vm;
    virDomainPtr dom = NULL;
2561

2562
    qemuDriverLock(driver);
2563
    vm = virDomainFindByName(&driver->domains, name);
2564 2565
    qemuDriverUnlock(driver);

2566
    if (!vm) {
2567 2568
        qemudReportError(conn, NULL, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching name '%s'"), name);
2569
        goto cleanup;
2570 2571
    }

2572
    dom = virGetDomain(conn, vm->def->name, vm->def->uuid);
2573
    if (dom) dom->id = vm->def->id;
2574 2575

cleanup:
2576 2577
    if (vm)
        virDomainObjUnlock(vm);
2578 2579 2580
    return dom;
}

2581
static int qemudGetVersion(virConnectPtr conn, unsigned long *version) {
2582 2583 2584
    struct qemud_driver *driver = conn->privateData;
    int ret = -1;

2585
    qemuDriverLock(driver);
2586
    if (qemudExtractVersion(conn, driver) < 0)
2587
        goto cleanup;
2588

2589
    *version = qemu_driver->qemuVersion;
2590 2591 2592
    ret = 0;

cleanup:
2593
    qemuDriverUnlock(driver);
2594
    return ret;
D
Daniel P. Berrange 已提交
2595 2596
}

2597 2598 2599
static char *
qemudGetHostname (virConnectPtr conn)
{
2600
    char *result;
2601

2602 2603
    result = virGetHostname();
    if (result == NULL) {
2604 2605
        virReportSystemError (conn, errno,
                              "%s", _("failed to determine host name"));
2606 2607 2608
        return NULL;
    }
    /* Caller frees this string. */
2609
    return result;
2610 2611
}

2612
static int qemudListDomains(virConnectPtr conn, int *ids, int nids) {
2613
    struct qemud_driver *driver = conn->privateData;
2614 2615
    int got = 0, i;

2616 2617 2618
    qemuDriverLock(driver);
    for (i = 0 ; i < driver->domains.count && got < nids ; i++) {
        virDomainObjLock(driver->domains.objs[i]);
2619 2620
        if (virDomainIsActive(driver->domains.objs[i]))
            ids[got++] = driver->domains.objs[i]->def->id;
2621 2622 2623
        virDomainObjUnlock(driver->domains.objs[i]);
    }
    qemuDriverUnlock(driver);
2624

D
Daniel P. Berrange 已提交
2625 2626
    return got;
}
2627

2628
static int qemudNumDomains(virConnectPtr conn) {
2629
    struct qemud_driver *driver = conn->privateData;
2630 2631
    int n = 0, i;

2632 2633 2634
    qemuDriverLock(driver);
    for (i = 0 ; i < driver->domains.count ; i++) {
        virDomainObjLock(driver->domains.objs[i]);
2635
        if (virDomainIsActive(driver->domains.objs[i]))
2636
            n++;
2637 2638 2639
        virDomainObjUnlock(driver->domains.objs[i]);
    }
    qemuDriverUnlock(driver);
2640

2641
    return n;
D
Daniel P. Berrange 已提交
2642
}
2643

2644
static virDomainPtr qemudDomainCreate(virConnectPtr conn, const char *xml,
2645
                                      unsigned int flags ATTRIBUTE_UNUSED) {
2646
    struct qemud_driver *driver = conn->privateData;
2647
    virDomainDefPtr def;
2648
    virDomainObjPtr vm = NULL;
2649
    virDomainPtr dom = NULL;
2650
    virDomainEventPtr event = NULL;
D
Daniel P. Berrange 已提交
2651

2652
    qemuDriverLock(driver);
2653 2654
    if (!(def = virDomainDefParseString(conn, driver->caps, xml,
                                        VIR_DOMAIN_XML_INACTIVE)))
2655
        goto cleanup;
2656

2657 2658 2659
    if (virSecurityDriverVerify(conn, def) < 0)
        goto cleanup;

2660
    /* See if a VM with matching UUID already exists */
2661
    vm = virDomainFindByUUID(&driver->domains, def->uuid);
2662
    if (vm) {
2663 2664 2665 2666 2667 2668 2669 2670 2671
        /* UUID matches, but if names don't match, refuse it */
        if (STRNEQ(vm->def->name, def->name)) {
            char uuidstr[VIR_UUID_STRING_BUFLEN];
            virUUIDFormat(vm->def->uuid, uuidstr);
            qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
                             _("domain '%s' is already defined with uuid %s"),
                             vm->def->name, uuidstr);
            goto cleanup;
        }
2672

2673 2674 2675 2676 2677 2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688 2689 2690
        /* UUID & name match, but if VM is already active, refuse it */
        if (virDomainIsActive(vm)) {
            qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
                             _("domain is already active as '%s'"), vm->def->name);
            goto cleanup;
        }
        virDomainObjUnlock(vm);
    } else {
        /* UUID does not match, but if a name matches, refuse it */
        vm = virDomainFindByName(&driver->domains, def->name);
        if (vm) {
            char uuidstr[VIR_UUID_STRING_BUFLEN];
            virUUIDFormat(vm->def->uuid, uuidstr);
            qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
                             _("domain '%s' is already defined with uuid %s"),
                             def->name, uuidstr);
            goto cleanup;
        }
2691
    }
2692

2693 2694
    if (!(vm = virDomainAssignDef(conn,
                                  &driver->domains,
2695 2696 2697 2698
                                  def)))
        goto cleanup;

    def = NULL;
D
Daniel P. Berrange 已提交
2699

2700
    if (qemudStartVMDaemon(conn, driver, vm, NULL, -1) < 0) {
2701 2702
        virDomainRemoveInactive(&driver->domains,
                                vm);
2703
        vm = NULL;
2704
        goto cleanup;
D
Daniel P. Berrange 已提交
2705
    }
2706 2707 2708 2709

    event = virDomainEventNewFromObj(vm,
                                     VIR_DOMAIN_EVENT_STARTED,
                                     VIR_DOMAIN_EVENT_STARTED_BOOTED);
D
Daniel P. Berrange 已提交
2710

2711
    dom = virGetDomain(conn, vm->def->name, vm->def->uuid);
2712
    if (dom) dom->id = vm->def->id;
2713 2714 2715

cleanup:
    virDomainDefFree(def);
2716 2717
    if (vm)
        virDomainObjUnlock(vm);
2718 2719
    if (event)
        qemuDomainEventQueue(driver, event);
2720
    qemuDriverUnlock(driver);
2721
    return dom;
D
Daniel P. Berrange 已提交
2722 2723 2724
}


2725
static int qemudDomainSuspend(virDomainPtr dom) {
2726
    struct qemud_driver *driver = dom->conn->privateData;
D
Daniel P. Berrange 已提交
2727
    char *info;
2728 2729
    virDomainObjPtr vm;
    int ret = -1;
2730
    virDomainEventPtr event = NULL;
2731

2732
    qemuDriverLock(driver);
2733
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
2734

D
Daniel P. Berrange 已提交
2735
    if (!vm) {
2736 2737 2738 2739
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
2740
        goto cleanup;
D
Daniel P. Berrange 已提交
2741
    }
2742
    if (!virDomainIsActive(vm)) {
2743
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_INVALID,
2744
                         "%s", _("domain is not running"));
2745
        goto cleanup;
D
Daniel P. Berrange 已提交
2746
    }
2747
    if (vm->state != VIR_DOMAIN_PAUSED) {
2748
        if (qemudMonitorCommand(vm, "stop", &info) < 0) {
2749 2750 2751 2752 2753 2754
            qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
                             "%s", _("suspend operation failed"));
            goto cleanup;
        }
        vm->state = VIR_DOMAIN_PAUSED;
        qemudDebug("Reply %s", info);
2755 2756 2757
        event = virDomainEventNewFromObj(vm,
                                         VIR_DOMAIN_EVENT_SUSPENDED,
                                         VIR_DOMAIN_EVENT_SUSPENDED_PAUSED);
2758
        VIR_FREE(info);
D
Daniel P. Berrange 已提交
2759
    }
2760
    if (virDomainSaveStatus(dom->conn, driver->stateDir, vm) < 0)
2761
        goto cleanup;
2762 2763 2764
    ret = 0;

cleanup:
2765 2766
    if (vm)
        virDomainObjUnlock(vm);
2767

2768
    if (event)
2769
        qemuDomainEventQueue(driver, event);
2770
    qemuDriverUnlock(driver);
2771
    return ret;
D
Daniel P. Berrange 已提交
2772 2773 2774
}


2775
static int qemudDomainResume(virDomainPtr dom) {
2776
    struct qemud_driver *driver = dom->conn->privateData;
D
Daniel P. Berrange 已提交
2777
    char *info;
2778 2779
    virDomainObjPtr vm;
    int ret = -1;
2780
    virDomainEventPtr event = NULL;
2781

2782
    qemuDriverLock(driver);
2783
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
2784

D
Daniel P. Berrange 已提交
2785
    if (!vm) {
2786 2787 2788 2789
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
2790
        goto cleanup;
D
Daniel P. Berrange 已提交
2791
    }
2792
    if (!virDomainIsActive(vm)) {
2793
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_INVALID,
2794
                         "%s", _("domain is not running"));
2795
        goto cleanup;
D
Daniel P. Berrange 已提交
2796
    }
2797
    if (vm->state == VIR_DOMAIN_PAUSED) {
2798
        if (qemudMonitorCommand(vm, "cont", &info) < 0) {
2799 2800 2801 2802 2803 2804
            qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
                             "%s", _("resume operation failed"));
            goto cleanup;
        }
        vm->state = VIR_DOMAIN_RUNNING;
        qemudDebug("Reply %s", info);
2805 2806 2807
        event = virDomainEventNewFromObj(vm,
                                         VIR_DOMAIN_EVENT_RESUMED,
                                         VIR_DOMAIN_EVENT_RESUMED_UNPAUSED);
2808
        VIR_FREE(info);
D
Daniel P. Berrange 已提交
2809
    }
2810
    if (virDomainSaveStatus(dom->conn, driver->stateDir, vm) < 0)
2811
        goto cleanup;
2812 2813 2814
    ret = 0;

cleanup:
2815 2816
    if (vm)
        virDomainObjUnlock(vm);
2817
    if (event)
2818
        qemuDomainEventQueue(driver, event);
2819
    qemuDriverUnlock(driver);
2820
    return ret;
D
Daniel P. Berrange 已提交
2821 2822 2823
}


2824
static int qemudDomainShutdown(virDomainPtr dom) {
2825 2826
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
2827
    char* info;
2828
    int ret = -1;
2829

2830
    qemuDriverLock(driver);
2831
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
2832 2833
    qemuDriverUnlock(driver);

2834
    if (!vm) {
2835 2836 2837 2838
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
2839
        goto cleanup;
2840 2841
    }

2842
    if (qemudMonitorCommand(vm, "system_powerdown", &info) < 0) {
2843
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
2844
                         "%s", _("shutdown operation failed"));
2845
        goto cleanup;
2846
    }
D
Daniel Veillard 已提交
2847
    VIR_FREE(info);
2848 2849 2850
    ret = 0;

cleanup:
2851 2852
    if (vm)
        virDomainObjUnlock(vm);
2853
    return ret;
2854 2855 2856
}


2857
static int qemudDomainDestroy(virDomainPtr dom) {
2858 2859 2860
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
    int ret = -1;
2861
    virDomainEventPtr event = NULL;
2862

2863
    qemuDriverLock(driver);
2864
    vm  = virDomainFindByUUID(&driver->domains, dom->uuid);
D
Daniel P. Berrange 已提交
2865
    if (!vm) {
2866 2867 2868 2869
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
2870
        goto cleanup;
D
Daniel P. Berrange 已提交
2871
    }
2872 2873 2874 2875 2876
    if (!virDomainIsActive(vm)) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
                         "%s", _("domain is not running"));
        goto cleanup;
    }
2877

2878
    qemudShutdownVMDaemon(dom->conn, driver, vm);
2879 2880 2881
    event = virDomainEventNewFromObj(vm,
                                     VIR_DOMAIN_EVENT_STOPPED,
                                     VIR_DOMAIN_EVENT_STOPPED_DESTROYED);
2882
    if (!vm->persistent) {
2883 2884
        virDomainRemoveInactive(&driver->domains,
                                vm);
2885 2886
        vm = NULL;
    }
2887 2888 2889
    ret = 0;

cleanup:
2890 2891
    if (vm)
        virDomainObjUnlock(vm);
2892 2893
    if (event)
        qemuDomainEventQueue(driver, event);
2894
    qemuDriverUnlock(driver);
2895
    return ret;
D
Daniel P. Berrange 已提交
2896 2897 2898
}


2899
static char *qemudDomainGetOSType(virDomainPtr dom) {
2900 2901 2902
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
    char *type = NULL;
2903

2904
    qemuDriverLock(driver);
2905
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
2906
    qemuDriverUnlock(driver);
2907
    if (!vm) {
2908 2909 2910 2911
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
2912
        goto cleanup;
2913 2914
    }

2915
    if (!(type = strdup(vm->def->os.type)))
2916
        virReportOOMError(dom->conn);
2917 2918

cleanup:
2919 2920
    if (vm)
        virDomainObjUnlock(vm);
2921 2922 2923
    return type;
}

2924 2925
/* Returns max memory in kb, 0 if error */
static unsigned long qemudDomainGetMaxMemory(virDomainPtr dom) {
2926 2927 2928
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
    unsigned long ret = 0;
2929

2930
    qemuDriverLock(driver);
2931
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
2932 2933
    qemuDriverUnlock(driver);

2934
    if (!vm) {
2935 2936
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
2937
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
2938
                         _("no domain with matching uuid '%s'"), uuidstr);
2939
        goto cleanup;
2940 2941
    }

2942 2943 2944
    ret = vm->def->maxmem;

cleanup:
2945 2946
    if (vm)
        virDomainObjUnlock(vm);
2947
    return ret;
2948 2949 2950
}

static int qemudDomainSetMaxMemory(virDomainPtr dom, unsigned long newmax) {
2951 2952 2953
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
    int ret = -1;
2954

2955
    qemuDriverLock(driver);
2956
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
2957 2958
    qemuDriverUnlock(driver);

2959
    if (!vm) {
2960 2961
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
2962
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
2963
                         _("no domain with matching uuid '%s'"), uuidstr);
2964
        goto cleanup;
2965 2966 2967 2968
    }

    if (newmax < vm->def->memory) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_INVALID_ARG,
2969
                         "%s", _("cannot set max memory lower than current memory"));
2970
        goto cleanup;;
2971 2972 2973
    }

    vm->def->maxmem = newmax;
2974 2975 2976
    ret = 0;

cleanup:
2977 2978
    if (vm)
        virDomainObjUnlock(vm);
2979
    return ret;
2980 2981
}

2982 2983 2984 2985 2986 2987 2988 2989 2990 2991 2992 2993 2994 2995 2996

/* The reply from QEMU contains 'ballon: actual=421' where value is in MB */
#define BALLOON_PREFIX "balloon: actual="

/*
 * Returns: 0 if balloon not supported, +1 if balloon query worked
 * or -1 on failure
 */
static int qemudDomainGetMemoryBalloon(virConnectPtr conn,
                                       virDomainObjPtr vm,
                                       unsigned long *currmem) {
    char *reply = NULL;
    int ret = -1;
    char *offset;

2997 2998 2999
    if (!virDomainIsActive(vm))
        return 0;

3000 3001 3002 3003 3004 3005
    if (qemudMonitorCommand(vm, "info balloon", &reply) < 0) {
        qemudReportError(conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
                         "%s", _("could not query memory balloon allocation"));
        goto cleanup;
    }

3006
    DEBUG ("%s: balloon reply: '%s'", vm->def->name, reply);
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
    if ((offset = strstr(reply, BALLOON_PREFIX)) != NULL) {
        unsigned int memMB;
        char *end;
        offset += strlen(BALLOON_PREFIX);
        if (virStrToLong_ui(offset, &end, 10, &memMB) < 0) {
            qemudReportError(conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
                             "%s", _("could not parse memory balloon allocation"));
            goto cleanup;
        }
        *currmem = memMB * 1024;
        ret = 1;
    } else {
        /* We don't raise an error here, since its to be expected that
         * many QEMU's don't support ballooning
         */
        ret = 0;
    }

cleanup:
    VIR_FREE(reply);
    return ret;
}

/*
 * Returns: 0 if balloon not supported, +1 if balloon query worked
 * or -1 on failure
 */
static int qemudDomainSetMemoryBalloon(virConnectPtr conn,
                                       virDomainObjPtr vm,
                                       unsigned long newmem) {
    char *cmd;
    char *reply = NULL;
    int ret = -1;

    /*
     * 'newmem' is in KB, QEMU monitor works in MB, and we all wish
     * we just worked in bytes with unsigned long long everywhere.
     */
    if (virAsprintf(&cmd, "balloon %lu", (newmem / 1024)) < 0) {
        virReportOOMError(conn);
        goto cleanup;
    }

    if (qemudMonitorCommand(vm, cmd, &reply) < 0) {
        qemudReportError(conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
                         "%s", _("could not balloon memory allocation"));
        VIR_FREE(cmd);
        goto cleanup;
    }
    VIR_FREE(cmd);

    /* If the command failed qemu prints: 'unknown command'
     * No message is printed on success it seems */
3060
    DEBUG ("%s: balloon reply: %s",vm->def->name,  reply);
3061 3062 3063 3064 3065 3066 3067 3068 3069 3070 3071 3072 3073
    if (strstr(reply, "\nunknown command:")) {
        /* Don't set error - it is expected memory balloon fails on many qemu */
        ret = 0;
    } else {
        ret = 1;
    }

cleanup:
    VIR_FREE(reply);
    return ret;
}


3074
static int qemudDomainSetMemory(virDomainPtr dom, unsigned long newmem) {
3075 3076 3077
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
    int ret = -1;
3078

3079
    qemuDriverLock(driver);
3080
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
3081
    qemuDriverUnlock(driver);
3082
    if (!vm) {
3083 3084
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
3085
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
3086
                         _("no domain with matching uuid '%s'"), uuidstr);
3087
        goto cleanup;
3088 3089 3090 3091
    }

    if (newmem > vm->def->maxmem) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_INVALID_ARG,
3092
                         "%s", _("cannot set memory higher than max memory"));
3093
        goto cleanup;
3094 3095
    }

3096 3097 3098 3099 3100 3101 3102 3103 3104
    if (virDomainIsActive(vm)) {
        ret = qemudDomainSetMemoryBalloon(dom->conn, vm, newmem);
        if (ret == 0)
            qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_SUPPORT,
                             "%s", _("cannot set memory of an active domain"));
    } else {
        vm->def->memory = newmem;
        ret = 0;
    }
3105 3106

cleanup:
3107 3108
    if (vm)
        virDomainObjUnlock(vm);
3109
    return ret;
3110 3111
}

3112
static int qemudDomainGetInfo(virDomainPtr dom,
3113
                              virDomainInfoPtr info) {
3114 3115 3116
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
    int ret = -1;
3117 3118
    int err;
    unsigned long balloon;
3119

3120
    qemuDriverLock(driver);
3121
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
3122
    qemuDriverUnlock(driver);
D
Daniel P. Berrange 已提交
3123
    if (!vm) {
3124 3125 3126 3127
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
3128
        goto cleanup;
D
Daniel P. Berrange 已提交
3129 3130
    }

3131
    info->state = vm->state;
D
Daniel P. Berrange 已提交
3132

3133
    if (!virDomainIsActive(vm)) {
3134
        info->cpuTime = 0;
D
Daniel P. Berrange 已提交
3135
    } else {
3136
        if (qemudGetProcessInfo(&(info->cpuTime), vm->pid) < 0) {
3137
            qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED, ("cannot read cputime for domain"));
3138
            goto cleanup;
D
Daniel P. Berrange 已提交
3139 3140 3141
        }
    }

3142
    info->maxMem = vm->def->maxmem;
3143 3144 3145 3146 3147 3148 3149 3150 3151 3152 3153 3154 3155 3156 3157

    if (virDomainIsActive(vm)) {
        err = qemudDomainGetMemoryBalloon(dom->conn, vm, &balloon);
        if (err < 0)
            goto cleanup;

        if (err == 0)
            /* Balloon not supported, so maxmem is always the allocation */
            info->memory = vm->def->maxmem;
        else
            info->memory = balloon;
    } else {
        info->memory = vm->def->memory;
    }

3158
    info->nrVirtCpu = vm->def->vcpus;
3159 3160 3161
    ret = 0;

cleanup:
3162 3163
    if (vm)
        virDomainObjUnlock(vm);
3164
    return ret;
D
Daniel P. Berrange 已提交
3165 3166 3167
}


D
Daniel P. Berrange 已提交
3168
static char *qemudEscape(const char *in, int shell)
3169 3170 3171 3172 3173 3174 3175 3176 3177 3178 3179 3180 3181 3182 3183 3184 3185 3186 3187 3188 3189
{
    int len = 0;
    int i, j;
    char *out;

    /* To pass through the QEMU monitor, we need to use escape
       sequences: \r, \n, \", \\

       To pass through both QEMU + the shell, we need to escape
       the single character ' as the five characters '\\''
    */

    for (i = 0; in[i] != '\0'; i++) {
        switch(in[i]) {
        case '\r':
        case '\n':
        case '"':
        case '\\':
            len += 2;
            break;
        case '\'':
D
Daniel P. Berrange 已提交
3190 3191 3192 3193
            if (shell)
                len += 5;
            else
                len += 1;
3194 3195 3196 3197 3198 3199 3200
            break;
        default:
            len += 1;
            break;
        }
    }

3201
    if (VIR_ALLOC_N(out, len + 1) < 0)
3202 3203 3204 3205 3206 3207 3208 3209 3210 3211 3212 3213 3214 3215 3216 3217 3218 3219
        return NULL;

    for (i = j = 0; in[i] != '\0'; i++) {
        switch(in[i]) {
        case '\r':
            out[j++] = '\\';
            out[j++] = 'r';
            break;
        case '\n':
            out[j++] = '\\';
            out[j++] = 'n';
            break;
        case '"':
        case '\\':
            out[j++] = '\\';
            out[j++] = in[i];
            break;
        case '\'':
D
Daniel P. Berrange 已提交
3220 3221 3222 3223 3224 3225 3226 3227 3228
            if (shell) {
                out[j++] = '\'';
                out[j++] = '\\';
                out[j++] = '\\';
                out[j++] = '\'';
                out[j++] = '\'';
            } else {
                out[j++] = in[i];
            }
3229 3230 3231 3232 3233 3234 3235 3236 3237 3238 3239
            break;
        default:
            out[j++] = in[i];
            break;
        }
    }
    out[j] = '\0';

    return out;
}

3240 3241 3242 3243 3244
static char *qemudEscapeMonitorArg(const char *in)
{
    return qemudEscape(in, 0);
}

D
Daniel P. Berrange 已提交
3245 3246 3247 3248
static char *qemudEscapeShellArg(const char *in)
{
    return qemudEscape(in, 1);
}
3249

3250 3251 3252 3253 3254 3255 3256 3257 3258 3259 3260
#define QEMUD_SAVE_MAGIC "LibvirtQemudSave"
#define QEMUD_SAVE_VERSION 1

struct qemud_save_header {
    char magic[sizeof(QEMUD_SAVE_MAGIC)-1];
    int version;
    int xml_len;
    int was_running;
    int unused[16];
};

3261
static int qemudDomainSave(virDomainPtr dom,
3262
                           const char *path) {
3263 3264 3265 3266 3267 3268 3269
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
    char *command = NULL;
    char *info = NULL;
    int fd = -1;
    char *safe_path = NULL;
    char *xml = NULL;
3270
    struct qemud_save_header header;
3271
    int ret = -1;
3272
    virDomainEventPtr event = NULL;
3273 3274 3275 3276 3277

    memset(&header, 0, sizeof(header));
    memcpy(header.magic, QEMUD_SAVE_MAGIC, sizeof(header.magic));
    header.version = QEMUD_SAVE_VERSION;

3278
    qemuDriverLock(driver);
3279
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
3280

D
Daniel P. Berrange 已提交
3281
    if (!vm) {
3282 3283 3284 3285
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
3286
        goto cleanup;
D
Daniel P. Berrange 已提交
3287
    }
3288

3289
    if (!virDomainIsActive(vm)) {
3290
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_INVALID,
3291
                         "%s", _("domain is not running"));
3292
        goto cleanup;
D
Daniel P. Berrange 已提交
3293
    }
3294 3295 3296 3297

    /* Pause */
    if (vm->state == VIR_DOMAIN_RUNNING) {
        header.was_running = 1;
3298
        if (qemudMonitorCommand(vm, "stop", &info) < 0) {
3299
            qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
3300
                             "%s", _("suspend operation failed"));
3301
            goto cleanup;
3302
        }
3303 3304 3305
        vm->state = VIR_DOMAIN_PAUSED;
        qemudDebug("Reply %s", info);
        VIR_FREE(info);
3306 3307 3308
    }

    /* Get XML for the domain */
3309
    xml = virDomainDefFormat(dom->conn, vm->def, VIR_DOMAIN_XML_SECURE);
3310 3311
    if (!xml) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
3312
                         "%s", _("failed to get domain xml"));
3313
        goto cleanup;
3314 3315 3316 3317 3318 3319
    }
    header.xml_len = strlen(xml) + 1;

    /* Write header to file, followed by XML */
    if ((fd = open(path, O_CREAT|O_TRUNC|O_WRONLY, S_IRUSR|S_IWUSR)) < 0) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
3320
                         _("failed to create '%s'"), path);
3321
        goto cleanup;
3322 3323 3324 3325
    }

    if (safewrite(fd, &header, sizeof(header)) != sizeof(header)) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
3326
                         "%s", _("failed to write save header"));
3327
        goto cleanup;
3328 3329 3330 3331
    }

    if (safewrite(fd, xml, header.xml_len) != header.xml_len) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
3332
                         "%s", _("failed to write xml"));
3333
        goto cleanup;
3334 3335
    }

3336
    if (close(fd) < 0) {
3337 3338 3339
        virReportSystemError(dom->conn, errno,
                             _("unable to save file %s"),
                             path);
3340 3341 3342
        goto cleanup;
    }
    fd = -1;
3343 3344 3345 3346

    /* Migrate to file */
    safe_path = qemudEscapeShellArg(path);
    if (!safe_path) {
3347
        virReportOOMError(dom->conn);
3348
        goto cleanup;
3349
    }
3350
    if (virAsprintf(&command, "migrate \"exec:"
3351
                  "dd of='%s' oflag=append conv=notrunc 2>/dev/null"
3352
                  "\"", safe_path) == -1) {
3353
        virReportOOMError(dom->conn);
3354 3355
        command = NULL;
        goto cleanup;
3356 3357
    }

3358
    if (qemudMonitorCommand(vm, command, &info) < 0) {
3359
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
3360
                         "%s", _("migrate operation failed"));
3361
        goto cleanup;
3362 3363
    }

3364
    DEBUG ("%s: migrate reply: %s", vm->def->name, info);
3365 3366 3367 3368 3369 3370 3371

    /* If the command isn't supported then qemu prints:
     * unknown command: migrate" */
    if (strstr(info, "unknown command:")) {
        qemudReportError (dom->conn, dom, NULL, VIR_ERR_NO_SUPPORT,
                          "%s",
                          _("'migrate' not supported by this qemu"));
3372
        goto cleanup;
3373 3374
    }

3375 3376
    /* Shut it down */
    qemudShutdownVMDaemon(dom->conn, driver, vm);
3377 3378 3379
    event = virDomainEventNewFromObj(vm,
                                     VIR_DOMAIN_EVENT_STOPPED,
                                     VIR_DOMAIN_EVENT_STOPPED_SAVED);
3380
    if (!vm->persistent) {
3381 3382
        virDomainRemoveInactive(&driver->domains,
                                vm);
3383 3384
        vm = NULL;
    }
3385 3386 3387 3388 3389 3390 3391 3392 3393 3394 3395
    ret = 0;

cleanup:
    if (fd != -1)
        close(fd);
    VIR_FREE(xml);
    VIR_FREE(safe_path);
    VIR_FREE(command);
    VIR_FREE(info);
    if (ret != 0)
        unlink(path);
3396 3397
    if (vm)
        virDomainObjUnlock(vm);
3398 3399
    if (event)
        qemuDomainEventQueue(driver, event);
3400
    qemuDriverUnlock(driver);
3401
    return ret;
D
Daniel P. Berrange 已提交
3402 3403 3404
}


P
Paolo Bonzini 已提交
3405 3406 3407 3408 3409 3410 3411 3412 3413 3414 3415 3416 3417 3418 3419 3420 3421 3422 3423 3424 3425 3426 3427 3428 3429 3430 3431 3432 3433 3434 3435 3436 3437 3438 3439 3440 3441 3442 3443 3444 3445 3446 3447 3448 3449 3450 3451 3452 3453 3454 3455 3456 3457 3458 3459 3460 3461 3462 3463 3464 3465 3466 3467 3468 3469 3470 3471 3472 3473 3474 3475 3476 3477 3478 3479 3480 3481 3482 3483 3484 3485 3486 3487 3488 3489 3490 3491 3492 3493 3494 3495 3496 3497 3498 3499 3500 3501 3502 3503 3504 3505 3506
static int qemudDomainCoreDump(virDomainPtr dom,
                               const char *path,
                               int flags ATTRIBUTE_UNUSED) {
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
    char *command = NULL;
    char *info = NULL;
    char *safe_path = NULL;
    int resume = 0, paused = 0;
    int ret = -1;

    qemuDriverLock(driver);
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
    qemuDriverUnlock(driver);

    if (!vm) {
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
        goto cleanup;
    }

    if (!virDomainIsActive(vm)) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_INVALID,
                         "%s", _("domain is not running"));
        goto cleanup;
    }

    /* Migrate will always stop the VM, so once we support live dumping
       the resume condition will stay the same, independent of whether
       the stop command is issued.  */
    resume = (vm->state == VIR_DOMAIN_RUNNING);

    /* Pause domain for non-live dump */
    if (vm->state == VIR_DOMAIN_RUNNING) {
        if (qemudMonitorCommand (vm, "stop", &info) < 0) {
            qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
                             "%s", _("suspending before dump failed"));
            goto cleanup;
        }
        DEBUG ("%s: stop reply: %s", vm->def->name, info);
        VIR_FREE(info);
        paused = 1;
    }

    /* Migrate to file */
    safe_path = qemudEscapeShellArg(path);
    if (!safe_path) {
        virReportOOMError(dom->conn);
        goto cleanup;
    }
    if (virAsprintf(&command, "migrate \"exec:"
                  "dd of='%s' 2>/dev/null"
                  "\"", safe_path) == -1) {
        virReportOOMError(dom->conn);
        command = NULL;
        goto cleanup;
    }

    if (qemudMonitorCommand(vm, command, &info) < 0) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
                         "%s", _("migrate operation failed"));
        goto cleanup;
    }

    DEBUG ("%s: migrate reply: %s", vm->def->name, info);

    /* If the command isn't supported then qemu prints:
     * unknown command: migrate" */
    if (strstr(info, "unknown command:")) {
        qemudReportError (dom->conn, dom, NULL, VIR_ERR_NO_SUPPORT,
                          "%s",
                          _("'migrate' not supported by this qemu"));
        goto cleanup;
    }

    paused = 1;
    ret = 0;
cleanup:
    VIR_FREE(safe_path);
    VIR_FREE(command);
    VIR_FREE(info);

    /* Since the monitor is always attached to a pty for libvirt, it
       will support synchronous operations so we always get here after
       the migration is complete.  */
    if (resume && paused) {
        if (qemudMonitorCommand(vm, "cont", &info) < 0) {
            qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
                             "%s", _("resuming after dump failed"));
            goto cleanup;
        }
        DEBUG ("%s: cont reply: %s", vm->def->name, info);
        VIR_FREE(info);
    }
    if (vm)
        virDomainObjUnlock(vm);
    return ret;
}


3507
static int qemudDomainSetVcpus(virDomainPtr dom, unsigned int nvcpus) {
3508 3509
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
3510
    int max;
3511
    int ret = -1;
3512
    const char *type;
3513

3514
    qemuDriverLock(driver);
3515
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
3516 3517
    qemuDriverUnlock(driver);

3518
    if (!vm) {
3519 3520
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
3521
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
3522
                         _("no domain with matching uuid '%s'"), uuidstr);
3523
        goto cleanup;
3524 3525
    }

3526
    if (virDomainIsActive(vm)) {
3527
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_INVALID, "%s",
3528
                         _("cannot change vcpu count of an active domain"));
3529
        goto cleanup;
3530 3531
    }

3532 3533 3534 3535 3536 3537 3538 3539
    if (!(type = virDomainVirtTypeToString(vm->def->virtType))) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_INTERNAL_ERROR,
                         _("unknown virt type in domain definition '%d'"),
                         vm->def->virtType);
        goto cleanup;
    }

    if ((max = qemudGetMaxVCPUs(dom->conn, type)) < 0) {
3540 3541
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_INTERNAL_ERROR, "%s",
                         _("could not determine max vcpus for the domain"));
3542
        goto cleanup;
3543 3544 3545 3546 3547 3548
    }

    if (nvcpus > max) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_INVALID_ARG,
                         _("requested vcpus is greater than max allowable"
                           " vcpus for the domain: %d > %d"), nvcpus, max);
3549
        goto cleanup;
3550 3551 3552
    }

    vm->def->vcpus = nvcpus;
3553 3554 3555
    ret = 0;

cleanup:
3556 3557
    if (vm)
        virDomainObjUnlock(vm);
3558
    return ret;
3559 3560
}

3561 3562 3563 3564 3565 3566 3567

#if HAVE_SCHED_GETAFFINITY
static int
qemudDomainPinVcpu(virDomainPtr dom,
                   unsigned int vcpu,
                   unsigned char *cpumap,
                   int maplen) {
3568 3569
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
3570 3571 3572
    cpu_set_t mask;
    int i, maxcpu;
    virNodeInfo nodeinfo;
3573
    int ret = -1;
3574

3575
    qemuDriverLock(driver);
3576
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
3577 3578
    qemuDriverUnlock(driver);

3579 3580 3581 3582 3583 3584 3585 3586
    if (!vm) {
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
        goto cleanup;
    }

3587
    if (!virDomainIsActive(vm)) {
3588
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_INVALID,
3589
                         "%s",_("cannot pin vcpus on an inactive domain"));
3590
        goto cleanup;
3591 3592 3593 3594 3595 3596
    }

    if (vcpu > (vm->nvcpupids-1)) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_INVALID_ARG,
                         _("vcpu number out of range %d > %d"),
                         vcpu, vm->nvcpupids);
3597
        goto cleanup;
3598 3599
    }

3600
    if (nodeGetInfo(dom->conn, &nodeinfo) < 0)
3601
        goto cleanup;
3602 3603 3604 3605 3606 3607 3608 3609 3610 3611 3612 3613 3614

    maxcpu = maplen * 8;
    if (maxcpu > nodeinfo.cpus)
        maxcpu = nodeinfo.cpus;

    CPU_ZERO(&mask);
    for (i = 0 ; i < maxcpu ; i++) {
        if ((cpumap[i/8] >> (i % 8)) & 1)
            CPU_SET(i, &mask);
    }

    if (vm->vcpupids != NULL) {
        if (sched_setaffinity(vm->vcpupids[vcpu], sizeof(mask), &mask) < 0) {
3615 3616
            virReportSystemError(dom->conn, errno, "%s",
                                 _("cannot set affinity"));
3617
            goto cleanup;
3618 3619 3620 3621
        }
    } else {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_SUPPORT,
                         "%s", _("cpu affinity is not supported"));
3622
        goto cleanup;
3623
    }
3624
    ret = 0;
3625

3626
cleanup:
3627 3628
    if (vm)
        virDomainObjUnlock(vm);
3629
    return ret;
3630 3631 3632 3633 3634 3635 3636 3637
}

static int
qemudDomainGetVcpus(virDomainPtr dom,
                    virVcpuInfoPtr info,
                    int maxinfo,
                    unsigned char *cpumaps,
                    int maplen) {
3638 3639
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
3640 3641
    virNodeInfo nodeinfo;
    int i, v, maxcpu;
3642
    int ret = -1;
3643

3644
    qemuDriverLock(driver);
3645
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
3646 3647
    qemuDriverUnlock(driver);

3648 3649 3650 3651 3652 3653 3654 3655
    if (!vm) {
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
        goto cleanup;
    }

3656
    if (!virDomainIsActive(vm)) {
3657
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_INVALID,
3658
                         "%s",_("cannot pin vcpus on an inactive domain"));
3659
        goto cleanup;
3660 3661
    }

3662
    if (nodeGetInfo(dom->conn, &nodeinfo) < 0)
3663
        goto cleanup;
3664 3665 3666 3667 3668 3669 3670 3671 3672

    maxcpu = maplen * 8;
    if (maxcpu > nodeinfo.cpus)
        maxcpu = nodeinfo.cpus;

    /* Clamp to actual number of vcpus */
    if (maxinfo > vm->nvcpupids)
        maxinfo = vm->nvcpupids;

3673 3674 3675 3676 3677 3678 3679 3680
    if (maxinfo >= 1) {
        if (info != NULL) {
            memset(info, 0, sizeof(*info) * maxinfo);
            for (i = 0 ; i < maxinfo ; i++) {
                info[i].number = i;
                info[i].state = VIR_VCPU_RUNNING;
                /* XXX cpu time, current pCPU mapping */
            }
3681 3682
        }

3683 3684 3685 3686 3687 3688 3689 3690 3691
        if (cpumaps != NULL) {
            memset(cpumaps, 0, maplen * maxinfo);
            if (vm->vcpupids != NULL) {
                for (v = 0 ; v < maxinfo ; v++) {
                    cpu_set_t mask;
                    unsigned char *cpumap = VIR_GET_CPUMAP(cpumaps, maplen, v);
                    CPU_ZERO(&mask);

                    if (sched_getaffinity(vm->vcpupids[v], sizeof(mask), &mask) < 0) {
3692 3693
                        virReportSystemError(dom->conn, errno, "%s",
                                             _("cannot get affinity"));
3694 3695 3696 3697 3698 3699
                        goto cleanup;
                    }

                    for (i = 0 ; i < maxcpu ; i++)
                        if (CPU_ISSET(i, &mask))
                            VIR_USE_CPU(cpumap, i);
3700
                }
3701 3702 3703 3704
            } else {
                qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_SUPPORT,
                                 "%s", _("cpu affinity is not available"));
                goto cleanup;
3705 3706 3707
            }
        }
    }
3708
    ret = maxinfo;
3709

3710
cleanup:
3711 3712
    if (vm)
        virDomainObjUnlock(vm);
3713
    return ret;
3714 3715 3716 3717
}
#endif /* HAVE_SCHED_GETAFFINITY */


3718
static int qemudDomainGetMaxVcpus(virDomainPtr dom) {
3719 3720
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
3721
    const char *type;
3722
    int ret = -1;
3723

3724
    qemuDriverLock(driver);
3725
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
3726 3727
    qemuDriverUnlock(driver);

3728
    if (!vm) {
3729 3730
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
3731
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
3732
                         _("no domain with matching uuid '%s'"), uuidstr);
3733
        goto cleanup;
3734 3735
    }

3736
    if (!(type = virDomainVirtTypeToString(vm->def->virtType))) {
3737 3738 3739
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_INTERNAL_ERROR,
                         _("unknown virt type in domain definition '%d'"),
                         vm->def->virtType);
3740
        goto cleanup;
3741 3742
    }

3743
    ret = qemudGetMaxVCPUs(dom->conn, type);
3744

3745
cleanup:
3746 3747
    if (vm)
        virDomainObjUnlock(vm);
3748 3749 3750
    return ret;
}

3751 3752 3753 3754 3755 3756 3757 3758 3759 3760
static int qemudDomainGetSecurityLabel(virDomainPtr dom, virSecurityLabelPtr seclabel)
{
    struct qemud_driver *driver = (struct qemud_driver *)dom->conn->privateData;
    virDomainObjPtr vm;
    const char *type;
    int ret = -1;

    qemuDriverLock(driver);
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);

3761 3762
    memset(seclabel, 0, sizeof(*seclabel));

3763 3764 3765
    if (!vm) {
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
3766
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
3767 3768 3769 3770 3771 3772 3773 3774 3775 3776 3777 3778 3779 3780 3781 3782 3783 3784 3785 3786 3787 3788 3789 3790 3791 3792 3793 3794 3795
                         _("no domain with matching uuid '%s'"), uuidstr);
        goto cleanup;
    }

    if (!(type = virDomainVirtTypeToString(vm->def->virtType))) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_INTERNAL_ERROR,
                         _("unknown virt type in domain definition '%d'"),
                         vm->def->virtType);
        goto cleanup;
    }

    /*
     * Theoretically, the pid can be replaced during this operation and
     * return the label of a different process.  If atomicity is needed,
     * further validation will be required.
     *
     * Comment from Dan Berrange:
     *
     *   Well the PID as stored in the virDomainObjPtr can't be changed
     *   because you've got a locked object.  The OS level PID could have
     *   exited, though and in extreme circumstances have cycled through all
     *   PIDs back to ours. We could sanity check that our PID still exists
     *   after reading the label, by checking that our FD connecting to the
     *   QEMU monitor hasn't seen SIGHUP/ERR on poll().
     */
    if (virDomainIsActive(vm)) {
        if (driver->securityDriver && driver->securityDriver->domainGetSecurityLabel) {
            if (driver->securityDriver->domainGetSecurityLabel(dom->conn, vm, seclabel) == -1) {
                qemudReportError(dom->conn, dom, NULL, VIR_ERR_INTERNAL_ERROR,
3796
                                 "%s", _("Failed to get security label"));
3797 3798 3799 3800 3801 3802 3803 3804 3805 3806
                goto cleanup;
            }
        }
    }

    ret = 0;

cleanup:
    if (vm)
        virDomainObjUnlock(vm);
3807
    qemuDriverUnlock(driver);
3808 3809 3810
    return ret;
}

3811 3812
static int qemudNodeGetSecurityModel(virConnectPtr conn,
                                     virSecurityModelPtr secmodel)
3813 3814 3815
{
    struct qemud_driver *driver = (struct qemud_driver *)conn->privateData;
    char *p;
3816
    int ret = 0;
3817

3818 3819
    qemuDriverLock(driver);
    if (!driver->securityDriver) {
3820
        memset(secmodel, 0, sizeof (*secmodel));
3821 3822
        goto cleanup;
    }
3823

3824 3825 3826 3827 3828
    p = driver->caps->host.secModel.model;
    if (strlen(p) >= VIR_SECURITY_MODEL_BUFLEN-1) {
        qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                         _("security model string exceeds max %d bytes"),
                         VIR_SECURITY_MODEL_BUFLEN-1);
3829 3830
        ret = -1;
        goto cleanup;
3831 3832 3833 3834 3835 3836 3837 3838
    }
    strcpy(secmodel->model, p);

    p = driver->caps->host.secModel.doi;
    if (strlen(p) >= VIR_SECURITY_DOI_BUFLEN-1) {
        qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                         _("security DOI string exceeds max %d bytes"),
                         VIR_SECURITY_DOI_BUFLEN-1);
3839 3840
        ret = -1;
        goto cleanup;
3841 3842
    }
    strcpy(secmodel->doi, p);
3843 3844 3845 3846

cleanup:
    qemuDriverUnlock(driver);
    return ret;
3847 3848 3849
}

/* TODO: check seclabel restore */
3850
static int qemudDomainRestore(virConnectPtr conn,
3851 3852 3853
                              const char *path) {
    struct qemud_driver *driver = conn->privateData;
    virDomainDefPtr def = NULL;
3854
    virDomainObjPtr vm = NULL;
3855 3856 3857
    int fd = -1;
    int ret = -1;
    char *xml = NULL;
3858
    struct qemud_save_header header;
3859
    virDomainEventPtr event = NULL;
3860

3861
    qemuDriverLock(driver);
3862 3863 3864
    /* Verify the header and read the XML */
    if ((fd = open(path, O_RDONLY)) < 0) {
        qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
3865
                         "%s", _("cannot read domain image"));
3866
        goto cleanup;
3867 3868 3869 3870
    }

    if (saferead(fd, &header, sizeof(header)) != sizeof(header)) {
        qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
3871
                         "%s", _("failed to read qemu header"));
3872
        goto cleanup;
3873 3874 3875 3876
    }

    if (memcmp(header.magic, QEMUD_SAVE_MAGIC, sizeof(header.magic)) != 0) {
        qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
3877
                         "%s", _("image magic is incorrect"));
3878
        goto cleanup;
3879 3880 3881 3882
    }

    if (header.version > QEMUD_SAVE_VERSION) {
        qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
3883
                         _("image version is not supported (%d > %d)"),
3884
                         header.version, QEMUD_SAVE_VERSION);
3885
        goto cleanup;
3886 3887
    }

3888
    if (VIR_ALLOC_N(xml, header.xml_len) < 0) {
3889
        virReportOOMError(conn);
3890
        goto cleanup;
3891 3892 3893 3894
    }

    if (saferead(fd, xml, header.xml_len) != header.xml_len) {
        qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
3895
                         "%s", _("failed to read XML"));
3896
        goto cleanup;
3897 3898 3899
    }

    /* Create a domain from this XML */
3900 3901
    if (!(def = virDomainDefParseString(conn, driver->caps, xml,
                                        VIR_DOMAIN_XML_INACTIVE))) {
3902
        qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
3903
                         "%s", _("failed to parse XML"));
3904
        goto cleanup;
3905 3906
    }

3907
    /* See if a VM with matching UUID already exists */
3908
    vm = virDomainFindByUUID(&driver->domains, def->uuid);
3909
    if (vm) {
3910 3911 3912 3913 3914 3915 3916 3917 3918 3919 3920
        /* UUID matches, but if names don't match, refuse it */
        if (STRNEQ(vm->def->name, def->name)) {
            char uuidstr[VIR_UUID_STRING_BUFLEN];
            virUUIDFormat(vm->def->uuid, uuidstr);
            qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
                             _("domain '%s' is already defined with uuid %s"),
                             vm->def->name, uuidstr);
            goto cleanup;
        }

        /* UUID & name match, but if VM is already active, refuse it */
3921
        if (virDomainIsActive(vm)) {
3922
            qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_INVALID,
3923 3924
                             _("domain is already active as '%s'"), vm->def->name);
            goto cleanup;
3925 3926 3927 3928 3929 3930 3931 3932 3933 3934 3935 3936
        }
        virDomainObjUnlock(vm);
    } else {
        /* UUID does not match, but if a name matches, refuse it */
        vm = virDomainFindByName(&driver->domains, def->name);
        if (vm) {
            char uuidstr[VIR_UUID_STRING_BUFLEN];
            virUUIDFormat(vm->def->uuid, uuidstr);
            qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
                             _("domain '%s' is already defined with uuid %s"),
                             def->name, uuidstr);
            goto cleanup;
3937
        }
3938 3939
    }

3940 3941 3942
    if (!(vm = virDomainAssignDef(conn,
                                  &driver->domains,
                                  def))) {
3943
        qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
3944
                         "%s", _("failed to assign new VM"));
3945
        goto cleanup;
3946
    }
3947
    def = NULL;
3948 3949

    /* Set the migration source and start it up. */
3950
    ret = qemudStartVMDaemon(conn, driver, vm, "stdio", fd);
3951
    close(fd);
3952
    fd = -1;
3953
    if (ret < 0) {
3954
        if (!vm->persistent) {
3955 3956
            virDomainRemoveInactive(&driver->domains,
                                    vm);
3957 3958
            vm = NULL;
        }
3959
        goto cleanup;
3960 3961
    }

3962 3963 3964
    event = virDomainEventNewFromObj(vm,
                                     VIR_DOMAIN_EVENT_STARTED,
                                     VIR_DOMAIN_EVENT_STARTED_RESTORED);
3965

3966 3967 3968
    /* If it was running before, resume it now. */
    if (header.was_running) {
        char *info;
3969
        if (qemudMonitorCommand(vm, "cont", &info) < 0) {
3970
            qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
3971
                             "%s", _("failed to resume domain"));
3972
            goto cleanup;
3973
        }
3974
        VIR_FREE(info);
3975 3976
        vm->state = VIR_DOMAIN_RUNNING;
    }
3977
    ret = 0;
3978

3979 3980 3981 3982 3983
cleanup:
    virDomainDefFree(def);
    VIR_FREE(xml);
    if (fd != -1)
        close(fd);
3984 3985
    if (vm)
        virDomainObjUnlock(vm);
3986 3987
    if (event)
        qemuDomainEventQueue(driver, event);
3988
    qemuDriverUnlock(driver);
3989
    return ret;
D
Daniel P. Berrange 已提交
3990 3991 3992
}


3993
static char *qemudDomainDumpXML(virDomainPtr dom,
3994
                                int flags) {
3995 3996 3997
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
    char *ret = NULL;
3998 3999
    unsigned long balloon;
    int err;
4000

4001
    qemuDriverLock(driver);
4002
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
4003 4004
    qemuDriverUnlock(driver);

D
Daniel P. Berrange 已提交
4005
    if (!vm) {
4006 4007 4008 4009
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
4010
        goto cleanup;
D
Daniel P. Berrange 已提交
4011 4012
    }

4013 4014 4015 4016 4017 4018 4019
    /* Refresh current memory based on balloon info */
    err = qemudDomainGetMemoryBalloon(dom->conn, vm, &balloon);
    if (err < 0)
        goto cleanup;
    if (err > 0)
        vm->def->memory = balloon;

4020 4021 4022 4023 4024 4025
    ret = virDomainDefFormat(dom->conn,
                             (flags & VIR_DOMAIN_XML_INACTIVE) && vm->newDef ?
                             vm->newDef : vm->def,
                             flags);

cleanup:
4026 4027
    if (vm)
        virDomainObjUnlock(vm);
4028
    return ret;
D
Daniel P. Berrange 已提交
4029 4030 4031
}


4032 4033 4034 4035
static char *qemuDomainXMLFromNative(virConnectPtr conn,
                                     const char *format,
                                     const char *config,
                                     unsigned int flags ATTRIBUTE_UNUSED) {
4036
    struct qemud_driver *driver = conn->privateData;
4037 4038 4039 4040 4041 4042 4043 4044 4045
    virDomainDefPtr def = NULL;
    char *xml = NULL;

    if (STRNEQ(format, QEMU_CONFIG_FORMAT_ARGV)) {
        qemudReportError(conn, NULL, NULL, VIR_ERR_INVALID_ARG,
                         _("unsupported config type %s"), format);
        goto cleanup;
    }

4046
    def = qemuParseCommandLineString(conn, driver->caps, config);
4047 4048 4049 4050 4051 4052 4053 4054 4055 4056
    if (!def)
        goto cleanup;

    xml = virDomainDefFormat(conn, def, VIR_DOMAIN_XML_INACTIVE);

cleanup:
    virDomainDefFree(def);
    return xml;
}

4057 4058 4059 4060 4061 4062
static char *qemuDomainXMLToNative(virConnectPtr conn,
                                   const char *format,
                                   const char *xmlData,
                                   unsigned int flags ATTRIBUTE_UNUSED) {
    struct qemud_driver *driver = conn->privateData;
    virDomainDefPtr def = NULL;
4063
    virDomainChrDef monitor_chr;
4064 4065 4066 4067 4068 4069 4070 4071 4072 4073
    const char *emulator;
    unsigned int qemuCmdFlags;
    struct stat sb;
    const char **retargv = NULL;
    const char **retenv = NULL;
    const char **tmp;
    virBuffer buf = VIR_BUFFER_INITIALIZER;
    char *ret = NULL;
    int i;

4074 4075
    qemuDriverLock(driver);

4076 4077 4078 4079 4080 4081 4082 4083 4084 4085 4086 4087 4088 4089 4090 4091 4092 4093 4094 4095 4096 4097 4098 4099 4100 4101 4102 4103 4104 4105 4106 4107 4108 4109 4110 4111 4112 4113 4114 4115 4116 4117 4118 4119 4120 4121 4122 4123 4124 4125 4126 4127 4128 4129 4130 4131 4132 4133 4134 4135 4136 4137 4138 4139 4140 4141
    if (STRNEQ(format, QEMU_CONFIG_FORMAT_ARGV)) {
        qemudReportError(conn, NULL, NULL, VIR_ERR_INVALID_ARG,
                         _("unsupported config type %s"), format);
        goto cleanup;
    }

    def = virDomainDefParseString(conn, driver->caps, xmlData, 0);
    if (!def)
        goto cleanup;

    /* Since we're just exporting args, we can't do bridge/network
     * setups, since libvirt will normally create TAP devices
     * directly. We convert those configs into generic 'ethernet'
     * config and assume the user has suitable 'ifup-qemu' scripts
     */
    for (i = 0 ; i < def->nnets ; i++) {
        virDomainNetDefPtr net = def->nets[i];
        if (net->type == VIR_DOMAIN_NET_TYPE_NETWORK) {
            VIR_FREE(net->data.network.name);

            memset(net, 0, sizeof *net);

            net->type = VIR_DOMAIN_NET_TYPE_ETHERNET;
            net->data.ethernet.dev = NULL;
            net->data.ethernet.script = NULL;
            net->data.ethernet.ipaddr = NULL;
        } else if (net->type == VIR_DOMAIN_NET_TYPE_BRIDGE) {
            char *brname = net->data.bridge.brname;
            char *script = net->data.bridge.script;
            char *ipaddr = net->data.bridge.ipaddr;

            memset(net, 0, sizeof *net);

            net->type = VIR_DOMAIN_NET_TYPE_ETHERNET;
            net->data.ethernet.dev = brname;
            net->data.ethernet.script = script;
            net->data.ethernet.ipaddr = ipaddr;
        }
    }
    for (i = 0 ; i < def->ngraphics ; i++) {
        if (def->graphics[i]->type == VIR_DOMAIN_GRAPHICS_TYPE_VNC &&
            def->graphics[i]->data.vnc.autoport)
            def->graphics[i]->data.vnc.port = 5900;
    }
    emulator = def->emulator;

    /* Make sure the binary we are about to try exec'ing exists.
     * Technically we could catch the exec() failure, but that's
     * in a sub-process so its hard to feed back a useful error
     */
    if (stat(emulator, &sb) < 0) {
        virReportSystemError(conn, errno,
                             _("Cannot find QEMU binary %s"),
                             emulator);
        goto cleanup;
    }

    if (qemudExtractVersionInfo(emulator,
                                NULL,
                                &qemuCmdFlags) < 0) {
        qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                         _("Cannot determine QEMU argv syntax %s"),
                         emulator);
        goto cleanup;
    }

4142 4143
    if (qemuPrepareMonitorChr(conn, driver, &monitor_chr, def->name) < 0)
        goto cleanup;
4144 4145

    if (qemudBuildCommandLine(conn, driver, def,
4146
                              &monitor_chr, qemuCmdFlags,
4147 4148 4149 4150 4151 4152 4153 4154 4155 4156 4157 4158 4159 4160 4161 4162 4163 4164 4165 4166 4167 4168 4169 4170 4171
                              &retargv, &retenv,
                              NULL, NULL, /* Don't want it to create TAP devices */
                              NULL) < 0) {
        goto cleanup;
    }

    tmp = retenv;
    while (*tmp) {
        virBufferAdd(&buf, *tmp, strlen(*tmp));
        virBufferAddLit(&buf, " ");
        tmp++;
    }
    tmp = retargv;
    while (*tmp) {
        virBufferAdd(&buf, *tmp, strlen(*tmp));
        virBufferAddLit(&buf, " ");
        tmp++;
    }

    if (virBufferError(&buf))
        goto cleanup;

    ret = virBufferContentAndReset(&buf);

cleanup:
4172
    qemuDriverUnlock(driver);
4173 4174 4175 4176 4177 4178 4179 4180 4181 4182 4183 4184 4185
    for (tmp = retargv ; tmp && *tmp ; tmp++)
        VIR_FREE(*tmp);
    VIR_FREE(retargv);

    for (tmp = retenv ; tmp && *tmp ; tmp++)
        VIR_FREE(*tmp);
    VIR_FREE(retenv);

    virDomainDefFree(def);
    return ret;
}


4186
static int qemudListDefinedDomains(virConnectPtr conn,
4187
                            char **const names, int nnames) {
4188
    struct qemud_driver *driver = conn->privateData;
4189
    int got = 0, i;
4190

4191
    qemuDriverLock(driver);
4192
    for (i = 0 ; i < driver->domains.count && got < nnames ; i++) {
4193
        virDomainObjLock(driver->domains.objs[i]);
4194 4195
        if (!virDomainIsActive(driver->domains.objs[i])) {
            if (!(names[got++] = strdup(driver->domains.objs[i]->def->name))) {
4196
                virReportOOMError(conn);
4197
                virDomainObjUnlock(driver->domains.objs[i]);
4198 4199
                goto cleanup;
            }
4200
        }
4201
        virDomainObjUnlock(driver->domains.objs[i]);
D
Daniel P. Berrange 已提交
4202
    }
4203

4204
    qemuDriverUnlock(driver);
D
Daniel P. Berrange 已提交
4205
    return got;
4206 4207 4208

 cleanup:
    for (i = 0 ; i < got ; i++)
4209
        VIR_FREE(names[i]);
4210
    qemuDriverUnlock(driver);
4211
    return -1;
D
Daniel P. Berrange 已提交
4212 4213
}

4214
static int qemudNumDefinedDomains(virConnectPtr conn) {
4215
    struct qemud_driver *driver = conn->privateData;
4216 4217
    int n = 0, i;

4218
    qemuDriverLock(driver);
4219 4220
    for (i = 0 ; i < driver->domains.count ; i++)
        if (!virDomainIsActive(driver->domains.objs[i]))
4221
            n++;
4222
    qemuDriverUnlock(driver);
4223

4224
    return n;
D
Daniel P. Berrange 已提交
4225 4226 4227
}


4228
static int qemudDomainStart(virDomainPtr dom) {
4229 4230 4231
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
    int ret = -1;
4232
    virDomainEventPtr event = NULL;
4233

4234
    qemuDriverLock(driver);
4235
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
4236

4237
    if (!vm) {
4238 4239 4240 4241
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
4242
        goto cleanup;
4243 4244
    }

4245
    ret = qemudStartVMDaemon(dom->conn, driver, vm, NULL, -1);
4246
    if (ret != -1)
4247 4248 4249
        event = virDomainEventNewFromObj(vm,
                                         VIR_DOMAIN_EVENT_STARTED,
                                         VIR_DOMAIN_EVENT_STARTED_BOOTED);
4250 4251

cleanup:
4252 4253
    if (vm)
        virDomainObjUnlock(vm);
4254
    if (event)
4255
        qemuDomainEventQueue(driver, event);
4256
    qemuDriverUnlock(driver);
4257
    return ret;
D
Daniel P. Berrange 已提交
4258 4259
}

4260 4261 4262 4263 4264 4265 4266 4267 4268 4269 4270 4271 4272 4273 4274 4275 4276 4277 4278 4279 4280 4281 4282 4283 4284 4285 4286 4287 4288
static int
qemudCanonicalizeMachineFromInfo(virDomainDefPtr def,
                                 virCapsGuestDomainInfoPtr info,
                                 char **canonical)
{
    int i;

    *canonical = NULL;

    for (i = 0; i < info->nmachines; i++) {
        virCapsGuestMachinePtr machine = info->machines[i];

        if (!machine->canonical)
            continue;

        if (strcmp(def->os.machine, machine->name) != 0)
            continue;

        if (!(*canonical = strdup(machine->canonical))) {
            virReportOOMError(NULL);
            return -1;
        }

        break;
    }

    return 0;
}

4289 4290 4291 4292 4293 4294 4295 4296 4297 4298 4299 4300 4301 4302 4303 4304 4305 4306 4307 4308 4309 4310 4311 4312 4313 4314 4315 4316
static int
qemudCanonicalizeMachineDirect(virDomainDefPtr def, char **canonical)
{
    virCapsGuestMachinePtr *machines = NULL;
    int i, nmachines = 0;

    if (qemudProbeMachineTypes(def->emulator, &machines, &nmachines) < 0) {
        virReportOOMError(NULL);
        return -1;
    }

    for (i = 0; i < nmachines; i++) {
        if (!machines[i]->canonical)
            continue;

        if (strcmp(def->os.machine, machines[i]->name) != 0)
            continue;

        *canonical = machines[i]->canonical;
        machines[i]->canonical = NULL;
        break;
    }

    virCapabilitiesFreeMachines(machines, nmachines);

    return 0;
}

4317 4318 4319 4320 4321 4322 4323 4324 4325 4326 4327 4328 4329 4330 4331 4332 4333 4334 4335 4336 4337 4338 4339 4340 4341 4342 4343 4344 4345 4346 4347 4348 4349 4350 4351 4352 4353
static int
qemudCanonicalizeMachine(virConnectPtr conn, virDomainDefPtr def)
{
    struct qemud_driver *driver = conn->privateData;
    char *canonical = NULL;
    int i;

    for (i = 0; i < driver->caps->nguests; i++) {
        virCapsGuestPtr guest = driver->caps->guests[i];
        int j;

        for (j = 0; j < guest->arch.ndomains; j++) {
            virCapsGuestDomainPtr dom = guest->arch.domains[j];

            if (dom->info.emulator &&
                STREQ(dom->info.emulator, def->emulator)) {
                if (qemudCanonicalizeMachineFromInfo(def, &dom->info,
                                                     &canonical) < 0)
                    return -1;
                if (canonical)
                    goto out;
                break;
            }
        }

        /* if we matched one of the domain's emulators, or if
         * we match the default emulator
         */
        if (j < guest->arch.ndomains ||
            (guest->arch.defaultInfo.emulator &&
             STREQ(guest->arch.defaultInfo.emulator, def->emulator))) {
            if (qemudCanonicalizeMachineFromInfo(def, &guest->arch.defaultInfo,
                                                 &canonical) < 0)
                return -1;
            goto out;
        }
    }
4354 4355 4356 4357

    if (qemudCanonicalizeMachineDirect(def, &canonical) < 0)
        return -1;

4358 4359 4360 4361 4362 4363 4364
out:
    if (canonical) {
        VIR_FREE(def->os.machine);
        def->os.machine = canonical;
    }
    return 0;
}
D
Daniel P. Berrange 已提交
4365

4366
static virDomainPtr qemudDomainDefine(virConnectPtr conn, const char *xml) {
4367
    struct qemud_driver *driver = conn->privateData;
4368
    virDomainDefPtr def;
4369
    virDomainObjPtr vm = NULL;
4370
    virDomainPtr dom = NULL;
4371
    virDomainEventPtr event = NULL;
4372
    int newVM = 1;
4373

4374
    qemuDriverLock(driver);
4375 4376
    if (!(def = virDomainDefParseString(conn, driver->caps, xml,
                                        VIR_DOMAIN_XML_INACTIVE)))
4377
        goto cleanup;
4378

4379 4380 4381
    if (virSecurityDriverVerify(conn, def) < 0)
        goto cleanup;

4382 4383
    /* See if a VM with matching UUID already exists */
    vm = virDomainFindByUUID(&driver->domains, def->uuid);
4384
    if (vm) {
4385 4386 4387 4388 4389 4390 4391 4392 4393 4394 4395
        /* UUID matches, but if names don't match, refuse it */
        if (STRNEQ(vm->def->name, def->name)) {
            char uuidstr[VIR_UUID_STRING_BUFLEN];
            virUUIDFormat(vm->def->uuid, uuidstr);
            qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
                             _("domain '%s' is already defined with uuid %s"),
                             vm->def->name, uuidstr);
            goto cleanup;
        }

        /* UUID & name match */
4396
        virDomainObjUnlock(vm);
4397
        newVM = 0;
4398 4399 4400 4401 4402 4403 4404 4405 4406 4407 4408
    } else {
        /* UUID does not match, but if a name matches, refuse it */
        vm = virDomainFindByName(&driver->domains, def->name);
        if (vm) {
            char uuidstr[VIR_UUID_STRING_BUFLEN];
            virUUIDFormat(vm->def->uuid, uuidstr);
            qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
                             _("domain '%s' is already defined with uuid %s"),
                             def->name, uuidstr);
            goto cleanup;
        }
4409
    }
4410

4411 4412 4413
    if (qemudCanonicalizeMachine(conn, def) < 0)
        goto cleanup;

4414 4415 4416
    if (!(vm = virDomainAssignDef(conn,
                                  &driver->domains,
                                  def))) {
4417
        goto cleanup;
4418
    }
4419
    def = NULL;
4420
    vm->persistent = 1;
4421

4422 4423
    if (virDomainSaveConfig(conn,
                            driver->configDir,
4424
                            vm->newDef ? vm->newDef : vm->def) < 0) {
4425 4426
        virDomainRemoveInactive(&driver->domains,
                                vm);
4427
        vm = NULL;
4428
        goto cleanup;
4429 4430
    }

4431 4432 4433 4434 4435
    event = virDomainEventNewFromObj(vm,
                                     VIR_DOMAIN_EVENT_DEFINED,
                                     newVM ?
                                     VIR_DOMAIN_EVENT_DEFINED_ADDED :
                                     VIR_DOMAIN_EVENT_DEFINED_UPDATED);
4436

4437
    dom = virGetDomain(conn, vm->def->name, vm->def->uuid);
4438
    if (dom) dom->id = vm->def->id;
4439 4440

cleanup:
4441
    virDomainDefFree(def);
4442 4443
    if (vm)
        virDomainObjUnlock(vm);
4444 4445
    if (event)
        qemuDomainEventQueue(driver, event);
4446
    qemuDriverUnlock(driver);
4447
    return dom;
D
Daniel P. Berrange 已提交
4448 4449
}

4450
static int qemudDomainUndefine(virDomainPtr dom) {
4451 4452
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
4453
    virDomainEventPtr event = NULL;
4454
    int ret = -1;
D
Daniel P. Berrange 已提交
4455

4456
    qemuDriverLock(driver);
4457
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
4458

D
Daniel P. Berrange 已提交
4459
    if (!vm) {
4460 4461 4462 4463
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
4464
        goto cleanup;
D
Daniel P. Berrange 已提交
4465 4466
    }

4467
    if (virDomainIsActive(vm)) {
4468
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_INVALID,
4469
                         "%s", _("cannot delete active domain"));
4470
        goto cleanup;
D
Daniel P. Berrange 已提交
4471 4472
    }

4473 4474 4475
    if (!vm->persistent) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_INTERNAL_ERROR,
                         "%s", _("cannot undefine transient domain"));
4476
        goto cleanup;
4477 4478 4479
    }

    if (virDomainDeleteConfig(dom->conn, driver->configDir, driver->autostartDir, vm) < 0)
4480
        goto cleanup;
D
Daniel P. Berrange 已提交
4481

4482 4483 4484
    event = virDomainEventNewFromObj(vm,
                                     VIR_DOMAIN_EVENT_UNDEFINED,
                                     VIR_DOMAIN_EVENT_UNDEFINED_REMOVED);
4485

4486 4487
    virDomainRemoveInactive(&driver->domains,
                            vm);
4488
    vm = NULL;
4489
    ret = 0;
D
Daniel P. Berrange 已提交
4490

4491
cleanup:
4492 4493
    if (vm)
        virDomainObjUnlock(vm);
4494 4495
    if (event)
        qemuDomainEventQueue(driver, event);
4496
    qemuDriverUnlock(driver);
4497
    return ret;
D
Daniel P. Berrange 已提交
4498 4499
}

4500
/* Return the disks name for use in monitor commands */
4501
static char *qemudDiskDeviceName(const virConnectPtr conn,
4502
                                 const virDomainDiskDefPtr disk) {
4503 4504 4505 4506 4507 4508

    int busid, devid;
    int ret;
    char *devname;

    if (virDiskNameToBusDeviceIndex(disk, &busid, &devid) < 0) {
4509
        qemudReportError(conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
4510 4511 4512 4513 4514 4515 4516
                         _("cannot convert disk '%s' to bus/device index"),
                         disk->dst);
        return NULL;
    }

    switch (disk->bus) {
        case VIR_DOMAIN_DISK_BUS_IDE:
4517
            if (disk->device== VIR_DOMAIN_DISK_DEVICE_DISK)
4518
                ret = virAsprintf(&devname, "ide%d-hd%d", busid, devid);
4519
            else
4520
                ret = virAsprintf(&devname, "ide%d-cd%d", busid, devid);
4521 4522
            break;
        case VIR_DOMAIN_DISK_BUS_SCSI:
4523
            if (disk->device == VIR_DOMAIN_DISK_DEVICE_DISK)
4524
                ret = virAsprintf(&devname, "scsi%d-hd%d", busid, devid);
4525
            else
4526
                ret = virAsprintf(&devname, "scsi%d-cd%d", busid, devid);
4527 4528
            break;
        case VIR_DOMAIN_DISK_BUS_FDC:
4529
            ret = virAsprintf(&devname, "floppy%d", devid);
4530 4531
            break;
        case VIR_DOMAIN_DISK_BUS_VIRTIO:
4532
            ret = virAsprintf(&devname, "virtio%d", devid);
4533 4534
            break;
        default:
4535
            qemudReportError(conn, NULL, NULL, VIR_ERR_NO_SUPPORT,
4536 4537 4538 4539 4540 4541
                             _("Unsupported disk name mapping for bus '%s'"),
                             virDomainDiskBusTypeToString(disk->bus));
            return NULL;
    }

    if (ret == -1) {
4542
        virReportOOMError(conn);
4543 4544 4545 4546 4547 4548
        return NULL;
    }

    return devname;
}

4549 4550
static int qemudDomainChangeEjectableMedia(virConnectPtr conn,
                                           virDomainObjPtr vm,
4551 4552
                                           virDomainDeviceDefPtr dev,
                                           unsigned int qemuCmdFlags)
4553
{
4554
    virDomainDiskDefPtr origdisk = NULL, newdisk;
4555
    char *cmd, *reply, *safe_path;
4556
    char *devname = NULL;
4557
    int i;
4558

4559
    origdisk = NULL;
4560
    newdisk = dev->data.disk;
4561 4562 4563 4564
    for (i = 0 ; i < vm->def->ndisks ; i++) {
        if (vm->def->disks[i]->bus == newdisk->bus &&
            STREQ(vm->def->disks[i]->dst, newdisk->dst)) {
            origdisk = vm->def->disks[i];
4565
            break;
4566
        }
4567 4568 4569
    }

    if (!origdisk) {
4570
        qemudReportError(conn, dom, NULL, VIR_ERR_INTERNAL_ERROR,
4571 4572 4573 4574 4575 4576 4577
                         _("No device with bus '%s' and target '%s'"),
                         virDomainDiskBusTypeToString(newdisk->bus),
                         newdisk->dst);
        return -1;
    }

    if (qemuCmdFlags & QEMUD_CMD_FLAG_DRIVE) {
4578
        if (!(devname = qemudDiskDeviceName(conn, newdisk)))
4579 4580 4581 4582 4583 4584 4585 4586 4587
            return -1;
    } else {
        /* Back compat for no -drive option */
        if (newdisk->device == VIR_DOMAIN_DISK_DEVICE_FLOPPY)
            devname = strdup(newdisk->dst);
        else if (newdisk->device == VIR_DOMAIN_DISK_DEVICE_CDROM &&
                 STREQ(newdisk->dst, "hdc"))
            devname = strdup("cdrom");
        else {
4588
            qemudReportError(conn, dom, NULL, VIR_ERR_INTERNAL_ERROR,
4589 4590 4591 4592 4593 4594 4595 4596
                             _("Emulator version does not support removable "
                               "media for device '%s' and target '%s'"),
                               virDomainDiskDeviceTypeToString(newdisk->device),
                               newdisk->dst);
            return -1;
        }

        if (!devname) {
4597
            virReportOOMError(conn);
4598 4599 4600
            return -1;
        }
    }
4601

4602
    if (newdisk->src) {
4603 4604
        safe_path = qemudEscapeMonitorArg(newdisk->src);
        if (!safe_path) {
4605
            virReportOOMError(conn);
4606
            VIR_FREE(devname);
4607 4608
            return -1;
        }
4609
        if (virAsprintf(&cmd, "change %s \"%s\"", devname, safe_path) == -1) {
4610
            virReportOOMError(conn);
4611
            VIR_FREE(safe_path);
4612
            VIR_FREE(devname);
4613 4614
            return -1;
        }
4615
        VIR_FREE(safe_path);
4616

4617
    } else if (virAsprintf(&cmd, "eject %s", devname) == -1) {
4618
        virReportOOMError(conn);
4619
        VIR_FREE(devname);
4620 4621
        return -1;
    }
4622
    VIR_FREE(devname);
4623

4624
    if (qemudMonitorCommand(vm, cmd, &reply) < 0) {
4625
        qemudReportError(conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
C
Cole Robinson 已提交
4626
                         "%s", _("could not change cdrom media"));
4627
        VIR_FREE(cmd);
4628 4629
        return -1;
    }
4630 4631 4632 4633

    /* If the command failed qemu prints:
     * device not found, device is locked ...
     * No message is printed on success it seems */
4634
    DEBUG ("%s: ejectable media change reply: %s", vm->def->name, reply);
4635
    if (strstr(reply, "\ndevice ")) {
4636
        qemudReportError (conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
C
Cole Robinson 已提交
4637
                          _("changing cdrom media failed: %s"), reply);
4638 4639 4640 4641
        VIR_FREE(reply);
        VIR_FREE(cmd);
        return -1;
    }
4642 4643
    VIR_FREE(reply);
    VIR_FREE(cmd);
4644

4645 4646
    VIR_FREE(origdisk->src);
    origdisk->src = newdisk->src;
4647
    newdisk->src = NULL;
4648
    origdisk->type = newdisk->type;
4649 4650 4651
    return 0;
}

4652 4653 4654 4655 4656 4657 4658 4659 4660 4661 4662 4663 4664 4665 4666 4667 4668 4669 4670 4671 4672 4673 4674 4675 4676 4677 4678 4679 4680 4681 4682 4683 4684 4685 4686 4687 4688 4689 4690 4691 4692 4693 4694 4695 4696 4697 4698 4699 4700 4701 4702 4703 4704 4705 4706 4707 4708 4709 4710 4711 4712 4713 4714 4715 4716 4717 4718
static int
qemudParsePciAddReply(virDomainObjPtr vm,
                      const char *reply,
                      unsigned *domain,
                      unsigned *bus,
                      unsigned *slot)
{
    char *s, *e;

    DEBUG("%s: pci_add reply: %s", vm->def->name, reply);

    /* If the command succeeds qemu prints:
     * OK bus 0, slot XXX...
     * or
     * OK domain 0, bus 0, slot XXX
     */
    if (!(s = strstr(reply, "OK ")))
        return -1;

    s += 3;

    if (STRPREFIX(s, "domain ")) {
        s += strlen("domain ");

        if (virStrToLong_ui(s, &e, 10, domain) == -1) {
            VIR_WARN(_("Unable to parse domain number '%s'\n"), s);
            return -1;
        }

        if (!STRPREFIX(e, ", ")) {
            VIR_WARN(_("Expected ', ' parsing pci_add reply '%s'\n"), s);
            return -1;
        }
        s = e + 2;
    }

    if (!STRPREFIX(s, "bus ")) {
        VIR_WARN(_("Expected 'bus ' parsing pci_add reply '%s'\n"), s);
        return -1;
    }
    s += strlen("bus ");

    if (virStrToLong_ui(s, &e, 10, bus) == -1) {
        VIR_WARN(_("Unable to parse bus number '%s'\n"), s);
        return -1;
    }

    if (!STRPREFIX(e, ", ")) {
        VIR_WARN(_("Expected ', ' parsing pci_add reply '%s'\n"), s);
        return -1;
    }
    s = e + 2;

    if (!STRPREFIX(s, "slot ")) {
        VIR_WARN(_("Expected 'slot ' parsing pci_add reply '%s'\n"), s);
        return -1;
    }
    s += strlen("slot ");

    if (virStrToLong_ui(s, &e, 10, slot) == -1) {
        VIR_WARN(_("Unable to parse slot number '%s'\n"), s);
        return -1;
    }

    return 0;
}

4719 4720 4721
static int qemudDomainAttachPciDiskDevice(virConnectPtr conn,
                                          virDomainObjPtr vm,
                                          virDomainDeviceDefPtr dev)
4722 4723
{
    int ret, i;
4724
    char *cmd, *reply;
4725 4726
    char *safe_path;
    const char* type = virDomainDiskBusTypeToString(dev->data.disk->bus);
4727
    int tryOldSyntax = 0;
4728
    unsigned domain, bus, slot;
4729 4730 4731

    for (i = 0 ; i < vm->def->ndisks ; i++) {
        if (STREQ(vm->def->disks[i]->dst, dev->data.disk->dst)) {
4732
            qemudReportError(conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
4733 4734 4735 4736 4737 4738
                           _("target %s already exists"), dev->data.disk->dst);
            return -1;
        }
    }

    if (VIR_REALLOC_N(vm->def->disks, vm->def->ndisks+1) < 0) {
4739
        virReportOOMError(conn);
4740 4741 4742
        return -1;
    }

4743
try_command:
4744 4745
    safe_path = qemudEscapeMonitorArg(dev->data.disk->src);
    if (!safe_path) {
4746
        virReportOOMError(conn);
4747 4748 4749
        return -1;
    }

4750 4751
    ret = virAsprintf(&cmd, "pci_add %s storage file=%s,if=%s",
                      (tryOldSyntax ? "0": "pci_addr=auto"), safe_path, type);
4752 4753
    VIR_FREE(safe_path);
    if (ret == -1) {
4754
        virReportOOMError(conn);
4755 4756 4757
        return ret;
    }

4758
    if (qemudMonitorCommand(vm, cmd, &reply) < 0) {
4759
        qemudReportError(conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
4760 4761 4762 4763 4764
                         _("cannot attach %s disk"), type);
        VIR_FREE(cmd);
        return -1;
    }

4765 4766
    VIR_FREE(cmd);

4767 4768 4769 4770 4771 4772
    if (qemudParsePciAddReply(vm, reply, &domain, &bus, &slot) < 0) {
        if (!tryOldSyntax && strstr(reply, "invalid char in expression")) {
            VIR_FREE(reply);
            tryOldSyntax = 1;
            goto try_command;
        }
4773

4774
        qemudReportError (conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
4775
                          _("adding %s disk failed: %s"), type, reply);
4776 4777 4778 4779
        VIR_FREE(reply);
        return -1;
    }

4780 4781 4782 4783 4784 4785
    VIR_FREE(reply);

    dev->data.disk->pci_addr.domain = domain;
    dev->data.disk->pci_addr.bus    = bus;
    dev->data.disk->pci_addr.slot   = slot;

4786 4787 4788 4789 4790 4791
    vm->def->disks[vm->def->ndisks++] = dev->data.disk;
    qsort(vm->def->disks, vm->def->ndisks, sizeof(*vm->def->disks),
          virDomainDiskQSort);

    return 0;
}
4792

4793 4794 4795
static int qemudDomainAttachUsbMassstorageDevice(virConnectPtr conn,
                                                 virDomainObjPtr vm,
                                                 virDomainDeviceDefPtr dev)
4796
{
4797 4798
    int ret, i;
    char *safe_path;
4799 4800
    char *cmd, *reply;

4801 4802
    for (i = 0 ; i < vm->def->ndisks ; i++) {
        if (STREQ(vm->def->disks[i]->dst, dev->data.disk->dst)) {
4803
            qemudReportError(conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
4804 4805 4806 4807 4808
                           _("target %s already exists"), dev->data.disk->dst);
            return -1;
        }
    }

4809
    if (VIR_REALLOC_N(vm->def->disks, vm->def->ndisks+1) < 0) {
4810
        virReportOOMError(conn);
4811
        return -1;
4812 4813
    }

4814 4815
    safe_path = qemudEscapeMonitorArg(dev->data.disk->src);
    if (!safe_path) {
4816
        virReportOOMError(conn);
4817 4818 4819
        return -1;
    }

4820
    ret = virAsprintf(&cmd, "usb_add disk:%s", safe_path);
4821
    VIR_FREE(safe_path);
4822
    if (ret == -1) {
4823
        virReportOOMError(conn);
4824 4825 4826
        return ret;
    }

4827
    if (qemudMonitorCommand(vm, cmd, &reply) < 0) {
4828
        qemudReportError(conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
4829
                         "%s", _("cannot attach usb disk"));
4830 4831 4832 4833
        VIR_FREE(cmd);
        return -1;
    }

4834
    DEBUG ("%s: attach_usb reply: %s",vm->def->name,  reply);
4835 4836 4837
    /* If the command failed qemu prints:
     * Could not add ... */
    if (strstr(reply, "Could not add ")) {
4838
        qemudReportError (conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
4839
                          "%s",
4840
                          _("adding usb disk failed"));
4841 4842 4843 4844
        VIR_FREE(reply);
        VIR_FREE(cmd);
        return -1;
    }
4845

4846 4847 4848
    vm->def->disks[vm->def->ndisks++] = dev->data.disk;
    qsort(vm->def->disks, vm->def->ndisks, sizeof(*vm->def->disks),
          virDomainDiskQSort);
4849

4850 4851 4852 4853 4854
    VIR_FREE(reply);
    VIR_FREE(cmd);
    return 0;
}

M
Mark McLoughlin 已提交
4855 4856 4857 4858 4859 4860
static int qemudDomainAttachNetDevice(virConnectPtr conn,
                                      virDomainObjPtr vm,
                                      virDomainDeviceDefPtr dev,
                                      unsigned int qemuCmdFlags)
{
    virDomainNetDefPtr net = dev->data.net;
4861
    char *cmd, *reply, *remove_cmd;
M
Mark McLoughlin 已提交
4862
    int i;
4863
    unsigned domain, bus, slot;
M
Mark McLoughlin 已提交
4864 4865 4866 4867 4868 4869 4870 4871 4872 4873 4874 4875 4876 4877 4878 4879 4880 4881 4882 4883 4884 4885 4886 4887 4888 4889 4890 4891 4892 4893 4894 4895 4896 4897 4898 4899 4900 4901

    if (!(qemuCmdFlags & QEMUD_CMD_FLAG_HOST_NET_ADD)) {
        qemudReportError(conn, dom, NULL, VIR_ERR_NO_SUPPORT, "%s",
                         _("installed qemu version does not support host_net_add"));
        return -1;
    }

    if (net->type == VIR_DOMAIN_NET_TYPE_BRIDGE ||
        net->type == VIR_DOMAIN_NET_TYPE_NETWORK) {
        qemudReportError(conn, dom, NULL, VIR_ERR_NO_SUPPORT,
                         _("network device type '%s' cannot be attached"),
                         virDomainNetTypeToString(net->type));
        return -1;
    }

    if (VIR_REALLOC_N(vm->def->nets, vm->def->nnets+1) < 0) {
        virReportOOMError(conn);
        return -1;
    }

    if ((qemuCmdFlags & QEMUD_CMD_FLAG_NET_NAME) &&
        qemuAssignNetNames(vm->def, net) < 0) {
        virReportOOMError(conn);
        return -1;
    }

    /* Choose a vlan value greater than all other values since
     * older versions did not store the value in the state file.
     */
    net->vlan = vm->def->nnets;
    for (i = 0; i < vm->def->nnets; i++)
        if (vm->def->nets[i]->vlan >= net->vlan)
            net->vlan = vm->def->nets[i]->vlan;

    if (qemuBuildHostNetStr(conn, net,
                            "host_net_add ", ' ', net->vlan, -1, &cmd) < 0)
        return -1;

4902 4903 4904 4905 4906 4907 4908 4909 4910
    remove_cmd = NULL;
    if (net->vlan >= 0 && net->hostnet_name &&
        virAsprintf(&remove_cmd, "host_net_remove %d %s",
                    net->vlan, net->hostnet_name) < 0) {
        VIR_FREE(cmd);
        virReportOOMError(conn);
        return -1;
    }

M
Mark McLoughlin 已提交
4911 4912 4913
    if (qemudMonitorCommand(vm, cmd, &reply) < 0) {
        qemudReportError(conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
                         _("failed to add network backend with '%s'"), cmd);
4914
        VIR_FREE(remove_cmd);
M
Mark McLoughlin 已提交
4915 4916 4917 4918 4919 4920 4921 4922 4923 4924
        VIR_FREE(cmd);
        return -1;
    }

    DEBUG("%s: host_net_add reply: %s", vm->def->name, reply);

    VIR_FREE(reply);
    VIR_FREE(cmd);

    if (qemuBuildNicStr(conn, net,
4925 4926
                        "pci_add pci_addr=auto ", ' ', net->vlan, &cmd) < 0)
        goto try_remove;
M
Mark McLoughlin 已提交
4927 4928 4929 4930 4931

    if (qemudMonitorCommand(vm, cmd, &reply) < 0) {
        qemudReportError(conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
                         _("failed to add NIC with '%s'"), cmd);
        VIR_FREE(cmd);
4932
        goto try_remove;
M
Mark McLoughlin 已提交
4933 4934 4935
    }

    VIR_FREE(cmd);
4936 4937 4938 4939 4940 4941 4942 4943 4944

    if (qemudParsePciAddReply(vm, reply, &domain, &bus, &slot) < 0) {
        qemudReportError(conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
                         _("parsing pci_add reply failed: %s"), reply);
        VIR_FREE(reply);
        goto try_remove;
    }

    VIR_FREE(reply);
4945
    VIR_FREE(remove_cmd);
M
Mark McLoughlin 已提交
4946

4947 4948 4949 4950
    net->pci_addr.domain = domain;
    net->pci_addr.bus    = bus;
    net->pci_addr.slot   = slot;

M
Mark McLoughlin 已提交
4951 4952 4953
    vm->def->nets[vm->def->nnets++] = net;

    return 0;
4954 4955 4956 4957 4958 4959 4960 4961 4962 4963 4964 4965 4966 4967 4968

try_remove:
    reply = NULL;

    if (!remove_cmd)
        VIR_WARN0(_("Unable to remove network backend\n"));
    else if (qemudMonitorCommand(vm, remove_cmd, &reply) < 0)
        VIR_WARN(_("Failed to remove network backend with '%s'\n"), remove_cmd);
    else
        VIR_DEBUG("%s: host_net_remove reply: %s\n", vm->def->name, reply);

    VIR_FREE(reply);
    VIR_FREE(remove_cmd);

    return -1;
M
Mark McLoughlin 已提交
4969 4970
}

4971 4972 4973
static int qemudDomainAttachHostDevice(virConnectPtr conn,
                                       virDomainObjPtr vm,
                                       virDomainDeviceDefPtr dev)
4974 4975 4976 4977
{
    int ret;
    char *cmd, *reply;

4978
    if (VIR_REALLOC_N(vm->def->hostdevs, vm->def->nhostdevs+1) < 0) {
4979
        virReportOOMError(conn);
4980 4981
        return -1;
    }
4982

4983
    if (dev->data.hostdev->source.subsys.u.usb.vendor) {
4984 4985 4986
        ret = virAsprintf(&cmd, "usb_add host:%.4x:%.4x",
                          dev->data.hostdev->source.subsys.u.usb.vendor,
                          dev->data.hostdev->source.subsys.u.usb.product);
4987
    } else {
4988 4989 4990
        ret = virAsprintf(&cmd, "usb_add host:%.3d.%.3d",
                          dev->data.hostdev->source.subsys.u.usb.bus,
                          dev->data.hostdev->source.subsys.u.usb.device);
4991 4992
    }
    if (ret == -1) {
4993
        virReportOOMError(conn);
4994 4995 4996
        return -1;
    }

4997
    if (qemudMonitorCommand(vm, cmd, &reply) < 0) {
4998
        qemudReportError(conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
4999 5000 5001 5002 5003
                         "%s", _("cannot attach usb device"));
        VIR_FREE(cmd);
        return -1;
    }

5004
    DEBUG ("%s: attach_usb reply: %s", vm->def->name, reply);
5005 5006 5007
    /* If the command failed qemu prints:
     * Could not add ... */
    if (strstr(reply, "Could not add ")) {
5008
        qemudReportError (conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
5009 5010 5011 5012 5013 5014
                          "%s",
                          _("adding usb device failed"));
        VIR_FREE(reply);
        VIR_FREE(cmd);
        return -1;
    }
5015

5016
    vm->def->hostdevs[vm->def->nhostdevs++] = dev->data.hostdev;
5017

5018 5019
    VIR_FREE(reply);
    VIR_FREE(cmd);
5020 5021 5022
    return 0;
}

5023 5024
static int qemudDomainAttachDevice(virDomainPtr dom,
                                   const char *xml) {
5025 5026 5027
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
    virDomainDeviceDefPtr dev = NULL;
5028
    unsigned int qemuCmdFlags;
5029
    virCgroupPtr cgroup = NULL;
5030
    int ret = -1;
5031

5032
    qemuDriverLock(driver);
5033
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
5034
    if (!vm) {
5035 5036 5037 5038
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
5039
        goto cleanup;
5040 5041 5042
    }

    if (!virDomainIsActive(vm)) {
5043
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_INVALID,
5044
                         "%s", _("cannot attach device on inactive domain"));
5045
        goto cleanup;
5046 5047
    }

5048 5049
    dev = virDomainDeviceDefParse(dom->conn, driver->caps, vm->def, xml,
                                  VIR_DOMAIN_XML_INACTIVE);
5050 5051 5052
    if (dev == NULL)
        goto cleanup;

5053 5054 5055 5056
    if (qemudExtractVersionInfo(vm->def->emulator,
                                NULL,
                                &qemuCmdFlags) < 0)
        goto cleanup;
5057

5058
    if (dev->type == VIR_DOMAIN_DEVICE_DISK) {
5059
        if (qemuCgroupControllerActive(driver, VIR_CGROUP_CONTROLLER_DEVICES)) {
5060 5061 5062 5063 5064 5065 5066 5067 5068 5069 5070 5071 5072 5073 5074 5075 5076 5077 5078 5079
            if (virCgroupForDomain(driver->cgroup, vm->def->name, &cgroup, 0) !=0 ) {
                qemudReportError(dom->conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                                 _("Unable to find cgroup for %s\n"),
                                 vm->def->name);
                goto cleanup;
            }
            if (dev->data.disk->src != NULL &&
                dev->data.disk->type == VIR_DOMAIN_DISK_TYPE_BLOCK &&
                virCgroupAllowDevicePath(cgroup,
                                         dev->data.disk->src) < 0) {
                qemudReportError(dom->conn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                                 _("unable to allow device %s"),
                                 dev->data.disk->src);
                goto cleanup;
            }
        }

        if (driver->securityDriver)
            driver->securityDriver->domainSetSecurityImageLabel(dom->conn, vm, dev->data.disk);

5080
        switch (dev->data.disk->device) {
5081 5082
        case VIR_DOMAIN_DISK_DEVICE_CDROM:
        case VIR_DOMAIN_DISK_DEVICE_FLOPPY:
5083 5084
            if (driver->securityDriver)
                driver->securityDriver->domainSetSecurityImageLabel(dom->conn, vm, dev->data.disk);
5085 5086 5087 5088

            if (qemuDomainSetDeviceOwnership(dom->conn, driver, dev, 0) < 0)
                goto cleanup;

5089
            ret = qemudDomainChangeEjectableMedia(dom->conn, vm, dev, qemuCmdFlags);
5090
            break;
5091

5092
        case VIR_DOMAIN_DISK_DEVICE_DISK:
5093 5094
            if (driver->securityDriver)
                driver->securityDriver->domainSetSecurityImageLabel(dom->conn, vm, dev->data.disk);
5095 5096 5097 5098

            if (qemuDomainSetDeviceOwnership(dom->conn, driver, dev, 0) < 0)
                goto cleanup;

5099
            if (dev->data.disk->bus == VIR_DOMAIN_DISK_BUS_USB) {
5100
                ret = qemudDomainAttachUsbMassstorageDevice(dom->conn, vm, dev);
5101 5102
            } else if (dev->data.disk->bus == VIR_DOMAIN_DISK_BUS_SCSI ||
                       dev->data.disk->bus == VIR_DOMAIN_DISK_BUS_VIRTIO) {
5103
                ret = qemudDomainAttachPciDiskDevice(dom->conn, vm, dev);
5104 5105 5106 5107
            } else {
                qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_SUPPORT,
                                 _("disk bus '%s' cannot be hotplugged."),
                                 virDomainDiskBusTypeToString(dev->data.disk->bus));
5108
                /* fallthrough */
5109 5110
            }
            break;
5111

5112 5113
        default:
            qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_SUPPORT,
5114 5115
                             _("disk device type '%s' cannot be hotplugged"),
                             virDomainDiskDeviceTypeToString(dev->data.disk->device));
5116 5117 5118 5119 5120
            /* Fallthrough */
        }
        if (ret != 0) {
            virCgroupDenyDevicePath(cgroup,
                                    dev->data.disk->src);
5121
        }
M
Mark McLoughlin 已提交
5122 5123
    } else if (dev->type == VIR_DOMAIN_DEVICE_NET) {
        ret = qemudDomainAttachNetDevice(dom->conn, vm, dev, qemuCmdFlags);
5124
    } else if (dev->type == VIR_DOMAIN_DEVICE_HOSTDEV &&
5125 5126
               dev->data.hostdev->mode == VIR_DOMAIN_HOSTDEV_MODE_SUBSYS &&
               dev->data.hostdev->source.subsys.type == VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_USB) {
5127 5128 5129
        if (qemuDomainSetDeviceOwnership(dom->conn, driver, dev, 0) < 0)
            goto cleanup;

5130
        ret = qemudDomainAttachHostDevice(dom->conn, vm, dev);
5131
    } else {
5132
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_SUPPORT,
5133 5134
                         _("device type '%s' cannot be attached"),
                         virDomainDeviceTypeToString(dev->type));
5135
        goto cleanup;
5136 5137
    }

5138
    if (!ret && virDomainSaveStatus(dom->conn, driver->stateDir, vm) < 0)
5139 5140
        ret = -1;

5141
cleanup:
5142 5143 5144
    if (cgroup)
        virCgroupFree(&cgroup);

5145 5146 5147
    if (ret < 0) {
        if (qemuDomainSetDeviceOwnership(dom->conn, driver, dev, 1) < 0)
            VIR_WARN0("Fail to restore disk device ownership");
G
Guido Günther 已提交
5148
        virDomainDeviceDefFree(dev);
5149
    }
5150 5151
    if (vm)
        virDomainObjUnlock(vm);
5152
    qemuDriverUnlock(driver);
5153 5154 5155
    return ret;
}

5156 5157
static int qemudDomainDetachPciDiskDevice(virConnectPtr conn,
                                          virDomainObjPtr vm, virDomainDeviceDefPtr dev)
5158 5159
{
    int i, ret = -1;
5160 5161
    char *cmd = NULL;
    char *reply = NULL;
5162
    virDomainDiskDefPtr detach = NULL;
5163
    int tryOldSyntax = 0;
5164 5165 5166 5167 5168 5169 5170 5171 5172

    for (i = 0 ; i < vm->def->ndisks ; i++) {
        if (STREQ(vm->def->disks[i]->dst, dev->data.disk->dst)) {
            detach = vm->def->disks[i];
            break;
        }
    }

    if (!detach) {
5173
        qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
5174
                         _("disk %s not found"), dev->data.disk->dst);
5175
        goto cleanup;
5176 5177
    }

5178
    if (!virDiskHasValidPciAddr(detach)) {
5179
        qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
5180 5181
                         _("disk %s cannot be detached - no PCI address for device"),
                           detach->dst);
5182
        goto cleanup;
5183 5184
    }

5185 5186
try_command:
    if (tryOldSyntax) {
5187
        if (virAsprintf(&cmd, "pci_del 0 %.2x", detach->pci_addr.slot) < 0) {
5188 5189 5190 5191
            virReportOOMError(conn);
            goto cleanup;
        }
    } else {
5192 5193 5194 5195
        if (virAsprintf(&cmd, "pci_del pci_addr=%.4x:%.2x:%.2x",
                        detach->pci_addr.domain,
                        detach->pci_addr.bus,
                        detach->pci_addr.slot) < 0) {
5196 5197 5198
            virReportOOMError(conn);
            goto cleanup;
        }
5199 5200
    }

5201
    if (qemudMonitorCommand(vm, cmd, &reply) < 0) {
5202
        qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
5203
                          _("failed to execute detach disk %s command"), detach->dst);
5204
        goto cleanup;
5205 5206
    }

5207
    DEBUG ("%s: pci_del reply: %s",vm->def->name,  reply);
5208 5209 5210 5211 5212 5213

    if (!tryOldSyntax &&
        strstr(reply, "extraneous characters")) {
        tryOldSyntax = 1;
        goto try_command;
    }
5214 5215
    /* If the command fails due to a wrong slot qemu prints: invalid slot,
     * nothing is printed on success */
5216 5217
    if (strstr(reply, "invalid slot") ||
        strstr(reply, "Invalid pci address")) {
5218
        qemudReportError (conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
5219 5220 5221 5222 5223 5224
                          _("failed to detach disk %s: invalid PCI address %.4x:%.2x:%.2x: %s"),
                          detach->dst,
                          detach->pci_addr.domain,
                          detach->pci_addr.bus,
                          detach->pci_addr.slot,
                          reply);
5225
        goto cleanup;
5226 5227 5228 5229 5230
    }

    if (vm->def->ndisks > 1) {
        vm->def->disks[i] = vm->def->disks[--vm->def->ndisks];
        if (VIR_REALLOC_N(vm->def->disks, vm->def->ndisks) < 0) {
5231
            virReportOOMError(conn);
5232
            goto cleanup;
5233 5234 5235 5236 5237 5238 5239 5240
        }
        qsort(vm->def->disks, vm->def->ndisks, sizeof(*vm->def->disks),
              virDomainDiskQSort);
    } else {
        VIR_FREE(vm->def->disks[0]);
        vm->def->ndisks = 0;
    }
    ret = 0;
5241 5242

cleanup:
5243 5244 5245 5246 5247
    VIR_FREE(reply);
    VIR_FREE(cmd);
    return ret;
}

5248 5249 5250 5251 5252 5253 5254 5255 5256 5257 5258 5259 5260 5261 5262 5263 5264 5265 5266 5267 5268 5269 5270 5271 5272 5273 5274 5275 5276 5277 5278 5279 5280 5281 5282 5283 5284 5285 5286 5287 5288 5289 5290 5291 5292 5293 5294 5295 5296 5297 5298 5299 5300 5301 5302 5303 5304 5305 5306 5307 5308 5309 5310 5311 5312 5313 5314 5315 5316 5317 5318 5319 5320 5321 5322 5323 5324 5325 5326 5327 5328 5329 5330 5331 5332 5333 5334 5335 5336 5337 5338 5339 5340 5341 5342 5343 5344
static int
qemudDomainDetachNetDevice(virConnectPtr conn,
                           virDomainObjPtr vm,
                           virDomainDeviceDefPtr dev)
{
    int i, ret = -1;
    char *cmd = NULL;
    char *reply = NULL;
    virDomainNetDefPtr detach = NULL;

    for (i = 0 ; i < vm->def->nnets ; i++) {
        virDomainNetDefPtr net = vm->def->nets[i];

        if (!memcmp(net->mac, dev->data.net->mac,  sizeof(net->mac))) {
            detach = net;
            break;
        }
    }

    if (!detach) {
        qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
                         _("network device %02x:%02x:%02x:%02x:%02x:%02x not found"),
                         dev->data.net->mac[0], dev->data.net->mac[1],
                         dev->data.net->mac[2], dev->data.net->mac[3],
                         dev->data.net->mac[4], dev->data.net->mac[5]);
        goto cleanup;
    }

    if (!virNetHasValidPciAddr(detach) || detach->vlan < 0 || !detach->hostnet_name) {
        qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
                         "%s", _("network device cannot be detached - device state missing"));
        goto cleanup;
    }

    if (virAsprintf(&cmd, "pci_del pci_addr=%.4x:%.2x:%.2x",
                    detach->pci_addr.domain,
                    detach->pci_addr.bus,
                    detach->pci_addr.slot) < 0) {
        virReportOOMError(conn);
        goto cleanup;
    }

    if (qemudMonitorCommand(vm, cmd, &reply) < 0) {
        qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
                          _("network device dettach command '%s' failed"), cmd);
        goto cleanup;
    }

    DEBUG("%s: pci_del reply: %s", vm->def->name,  reply);

    /* If the command fails due to a wrong PCI address qemu prints
     * 'invalid pci address'; nothing is printed on success */
    if (strstr(reply, "Invalid pci address")) {
        qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
                         _("failed to detach network device: invalid PCI address %.4x:%.2x:%.2x: %s"),
                         detach->pci_addr.domain,
                         detach->pci_addr.bus,
                         detach->pci_addr.slot,
                         reply);
        goto cleanup;
    }

    VIR_FREE(reply);
    VIR_FREE(cmd);

    if (virAsprintf(&cmd, "host_net_remove %d %s",
                    detach->vlan, detach->hostnet_name) < 0) {
        virReportOOMError(conn);
        goto cleanup;
    }

    if (qemudMonitorCommand(vm, cmd, &reply) < 0) {
        qemudReportError(conn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
                          _("network device dettach command '%s' failed"), cmd);
        goto cleanup;
    }

    DEBUG("%s: host_net_remove reply: %s", vm->def->name,  reply);

    if (vm->def->nnets > 1) {
        vm->def->nets[i] = vm->def->nets[--vm->def->nnets];
        if (VIR_REALLOC_N(vm->def->nets, vm->def->nnets) < 0) {
            virReportOOMError(conn);
            goto cleanup;
        }
    } else {
        VIR_FREE(vm->def->nets[0]);
        vm->def->nnets = 0;
    }
    ret = 0;

cleanup:
    VIR_FREE(reply);
    VIR_FREE(cmd);
    return ret;
}

5345 5346
static int qemudDomainDetachDevice(virDomainPtr dom,
                                   const char *xml) {
5347 5348 5349 5350
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
    virDomainDeviceDefPtr dev = NULL;
    int ret = -1;
5351

5352
    qemuDriverLock(driver);
5353
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
5354
    if (!vm) {
5355 5356 5357 5358
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
5359
        goto cleanup;
5360 5361 5362
    }

    if (!virDomainIsActive(vm)) {
5363
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_INVALID,
G
Guido Günther 已提交
5364
                         "%s", _("cannot detach device on inactive domain"));
5365
        goto cleanup;
5366 5367
    }

5368 5369
    dev = virDomainDeviceDefParse(dom->conn, driver->caps, vm->def, xml,
                                  VIR_DOMAIN_XML_INACTIVE);
5370 5371 5372
    if (dev == NULL)
        goto cleanup;

5373 5374 5375 5376

    if (dev->type == VIR_DOMAIN_DEVICE_DISK &&
        dev->data.disk->device == VIR_DOMAIN_DISK_DEVICE_DISK &&
        (dev->data.disk->bus == VIR_DOMAIN_DISK_BUS_SCSI ||
5377
         dev->data.disk->bus == VIR_DOMAIN_DISK_BUS_VIRTIO)) {
5378
        ret = qemudDomainDetachPciDiskDevice(dom->conn, vm, dev);
5379
        if (driver->securityDriver)
5380
            driver->securityDriver->domainRestoreSecurityImageLabel(dom->conn, dev->data.disk);
5381 5382
        if (qemuDomainSetDeviceOwnership(dom->conn, driver, dev, 1) < 0)
            VIR_WARN0("Fail to restore disk device ownership");
5383 5384 5385
    } else if (dev->type == VIR_DOMAIN_DEVICE_NET) {
        ret = qemudDomainDetachNetDevice(dom->conn, vm, dev);
    } else
5386 5387 5388
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_SUPPORT,
                         "%s", _("only SCSI or virtio disk device can be detached dynamically"));

5389
    if (!ret && virDomainSaveStatus(dom->conn, driver->stateDir, vm) < 0)
5390 5391
        ret = -1;

5392 5393
cleanup:
    virDomainDeviceDefFree(dev);
5394 5395
    if (vm)
        virDomainObjUnlock(vm);
5396
    qemuDriverUnlock(driver);
5397 5398 5399
    return ret;
}

5400
static int qemudDomainGetAutostart(virDomainPtr dom,
5401
                                   int *autostart) {
5402 5403 5404
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
    int ret = -1;
5405

5406
    qemuDriverLock(driver);
5407
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
5408 5409
    qemuDriverUnlock(driver);

5410
    if (!vm) {
5411 5412 5413 5414
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
5415
        goto cleanup;
5416 5417 5418
    }

    *autostart = vm->autostart;
5419
    ret = 0;
5420

5421
cleanup:
5422 5423
    if (vm)
        virDomainObjUnlock(vm);
5424
    return ret;
5425 5426
}

5427
static int qemudDomainSetAutostart(virDomainPtr dom,
5428
                                   int autostart) {
5429 5430
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
5431 5432
    char *configFile = NULL, *autostartLink = NULL;
    int ret = -1;
5433

5434
    qemuDriverLock(driver);
5435
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
5436

5437
    if (!vm) {
5438 5439 5440 5441
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
5442
        goto cleanup;
5443 5444
    }

5445 5446 5447
    if (!vm->persistent) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_INTERNAL_ERROR,
                         "%s", _("cannot set autostart for transient domain"));
5448
        goto cleanup;
5449 5450
    }

5451 5452
    autostart = (autostart != 0);

5453 5454 5455 5456 5457
    if (vm->autostart != autostart) {
        if ((configFile = virDomainConfigFile(dom->conn, driver->configDir, vm->def->name)) == NULL)
            goto cleanup;
        if ((autostartLink = virDomainConfigFile(dom->conn, driver->autostartDir, vm->def->name)) == NULL)
            goto cleanup;
5458

5459 5460
        if (autostart) {
            int err;
5461

5462
            if ((err = virFileMakePath(driver->autostartDir))) {
5463 5464 5465
                virReportSystemError(dom->conn, err,
                                     _("cannot create autostart directory %s"),
                                     driver->autostartDir);
5466 5467
                goto cleanup;
            }
5468

5469
            if (symlink(configFile, autostartLink) < 0) {
5470 5471 5472
                virReportSystemError(dom->conn, errno,
                                     _("Failed to create symlink '%s to '%s'"),
                                     autostartLink, configFile);
5473 5474 5475 5476
                goto cleanup;
            }
        } else {
            if (unlink(autostartLink) < 0 && errno != ENOENT && errno != ENOTDIR) {
5477 5478 5479
                virReportSystemError(dom->conn, errno,
                                     _("Failed to delete symlink '%s'"),
                                     autostartLink);
5480 5481
                goto cleanup;
            }
5482 5483
        }

5484
        vm->autostart = autostart;
5485
    }
5486
    ret = 0;
5487

5488 5489 5490
cleanup:
    VIR_FREE(configFile);
    VIR_FREE(autostartLink);
5491 5492
    if (vm)
        virDomainObjUnlock(vm);
5493
    qemuDriverUnlock(driver);
5494
    return ret;
5495 5496
}

5497 5498 5499 5500 5501 5502 5503

static char *qemuGetSchedulerType(virDomainPtr dom,
                                  int *nparams)
{
    struct qemud_driver *driver = dom->conn->privateData;
    char *ret;

5504
    if (!qemuCgroupControllerActive(driver, VIR_CGROUP_CONTROLLER_CPU)) {
5505 5506 5507 5508 5509 5510 5511 5512 5513 5514 5515 5516 5517 5518 5519 5520 5521 5522 5523 5524 5525 5526 5527 5528
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_SUPPORT,
                         __FUNCTION__);
        return NULL;
    }

    if (nparams)
        *nparams = 1;

    ret = strdup("posix");
    if (!ret)
        virReportOOMError(dom->conn);
    return ret;
}

static int qemuSetSchedulerParameters(virDomainPtr dom,
                                      virSchedParameterPtr params,
                                      int nparams)
{
    struct qemud_driver *driver = dom->conn->privateData;
    int i;
    virCgroupPtr group = NULL;
    virDomainObjPtr vm = NULL;
    int ret = -1;

5529
    if (!qemuCgroupControllerActive(driver, VIR_CGROUP_CONTROLLER_CPU)) {
5530 5531 5532 5533 5534 5535 5536 5537 5538 5539 5540 5541 5542 5543 5544 5545 5546 5547 5548 5549 5550 5551 5552 5553 5554 5555 5556 5557 5558 5559 5560 5561 5562 5563 5564 5565 5566 5567 5568 5569 5570 5571 5572 5573 5574 5575 5576 5577 5578 5579 5580 5581 5582 5583 5584 5585 5586 5587 5588 5589 5590 5591 5592 5593
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_SUPPORT,
                         __FUNCTION__);
        return -1;
    }

    qemuDriverLock(driver);
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
    qemuDriverUnlock(driver);

    if (vm == NULL) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_INTERNAL_ERROR,
                         _("No such domain %s"), dom->uuid);
        goto cleanup;
    }

    if (virCgroupForDomain(driver->cgroup, vm->def->name, &group, 0) != 0) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_INTERNAL_ERROR,
                         _("cannot find cgroup for domain %s"), vm->def->name);
        goto cleanup;
    }

    for (i = 0; i < nparams; i++) {
        virSchedParameterPtr param = &params[i];

        if (STREQ(param->field, "cpu_shares")) {
            int rc;
            if (param->type != VIR_DOMAIN_SCHED_FIELD_ULLONG) {
                qemudReportError(dom->conn, dom, NULL, VIR_ERR_INVALID_ARG,
                                 _("invalid type for cpu_shares tunable, expected a 'ullong'"));
                goto cleanup;
            }

            rc = virCgroupSetCpuShares(group, params[i].value.ul);
            if (rc != 0) {
                virReportSystemError(dom->conn, -rc, "%s",
                                     _("unable to set cpu shares tunable"));
                goto cleanup;
            }
        } else {
            qemudReportError(dom->conn, domain, NULL, VIR_ERR_INVALID_ARG,
                             _("Invalid parameter `%s'"), param->field);
            goto cleanup;
        }
    }
    ret = 0;

cleanup:
    virCgroupFree(&group);
    if (vm)
        virDomainObjUnlock(vm);
    return ret;
}

static int qemuGetSchedulerParameters(virDomainPtr dom,
                                      virSchedParameterPtr params,
                                      int *nparams)
{
    struct qemud_driver *driver = dom->conn->privateData;
    virCgroupPtr group = NULL;
    virDomainObjPtr vm = NULL;
    unsigned long long val;
    int ret = -1;
    int rc;

5594
    if (!qemuCgroupControllerActive(driver, VIR_CGROUP_CONTROLLER_CPU)) {
5595 5596 5597 5598 5599 5600 5601 5602 5603 5604 5605 5606 5607 5608 5609 5610 5611 5612 5613 5614 5615 5616 5617 5618 5619 5620 5621 5622 5623 5624 5625 5626 5627 5628 5629 5630 5631 5632 5633 5634 5635 5636 5637 5638 5639 5640 5641
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_SUPPORT,
                         __FUNCTION__);
        return -1;
    }

    if ((*nparams) != 1) {
        qemudReportError(dom->conn, domain, NULL, VIR_ERR_INVALID_ARG,
                         "%s", _("Invalid parameter count"));
        return -1;
    }

    qemuDriverLock(driver);
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
    qemuDriverUnlock(driver);

    if (vm == NULL) {
        qemudReportError(dom->conn, domain, NULL, VIR_ERR_INTERNAL_ERROR,
                         _("No such domain %s"), dom->uuid);
        goto cleanup;
    }

    if (virCgroupForDomain(driver->cgroup, vm->def->name, &group, 0) != 0) {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_INTERNAL_ERROR,
                         _("cannot find cgroup for domain %s"), vm->def->name);
        goto cleanup;
    }

    rc = virCgroupGetCpuShares(group, &val);
    if (rc != 0) {
        virReportSystemError(dom->conn, -rc, "%s",
                             _("unable to get cpu shares tunable"));
        goto cleanup;
    }
    params[0].value.ul = val;
    strncpy(params[0].field, "cpu_shares", sizeof(params[0].field));
    params[0].type = VIR_DOMAIN_SCHED_FIELD_ULLONG;

    ret = 0;

cleanup:
    virCgroupFree(&group);
    if (vm)
        virDomainObjUnlock(vm);
    return ret;
}


5642 5643 5644 5645 5646 5647 5648 5649 5650
/* This uses the 'info blockstats' monitor command which was
 * integrated into both qemu & kvm in late 2007.  If the command is
 * not supported we detect this and return the appropriate error.
 */
static int
qemudDomainBlockStats (virDomainPtr dom,
                       const char *path,
                       struct _virDomainBlockStats *stats)
{
5651
    struct qemud_driver *driver = dom->conn->privateData;
5652
    char *dummy, *info = NULL;
5653
    const char *p, *eol;
5654
    const char *qemu_dev_name = NULL;
5655
    size_t len;
5656
    int i, ret = -1;
5657
    virDomainObjPtr vm;
5658
    virDomainDiskDefPtr disk = NULL;
5659

5660
    qemuDriverLock(driver);
5661
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
5662
    qemuDriverUnlock(driver);
5663
    if (!vm) {
5664 5665 5666 5667
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
5668
        goto cleanup;
5669
    }
5670
    if (!virDomainIsActive (vm)) {
5671
        qemudReportError (dom->conn, dom, NULL, VIR_ERR_OPERATION_INVALID,
5672
                          "%s", _("domain is not running"));
5673
        goto cleanup;
5674 5675
    }

5676 5677 5678 5679 5680 5681 5682 5683
    for (i = 0 ; i < vm->def->ndisks ; i++) {
        if (STREQ(path, vm->def->disks[i]->dst)) {
            disk = vm->def->disks[i];
            break;
        }
    }

    if (!disk) {
5684 5685
        qemudReportError (dom->conn, dom, NULL, VIR_ERR_INVALID_ARG,
                          _("invalid path: %s"), path);
5686
        goto cleanup;
5687 5688
    }

5689
    qemu_dev_name = qemudDiskDeviceName(dom->conn, disk);
5690
    if (!qemu_dev_name)
5691
        goto cleanup;
5692 5693
    len = strlen (qemu_dev_name);

5694
    if (qemudMonitorCommand (vm, "info blockstats", &info) < 0) {
5695 5696
        qemudReportError (dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
                          "%s", _("'info blockstats' command failed"));
5697
        goto cleanup;
5698
    }
5699
    DEBUG ("%s: info blockstats reply: %s", vm->def->name, info);
5700 5701 5702 5703 5704 5705

    /* If the command isn't supported then qemu prints the supported
     * info commands, so the output starts "info ".  Since this is
     * unlikely to be the name of a block device, we can use this
     * to detect if qemu supports the command.
     */
C
Cole Robinson 已提交
5706
    if (strstr(info, "\ninfo ")) {
5707 5708 5709
        qemudReportError (dom->conn, dom, NULL, VIR_ERR_NO_SUPPORT,
                          "%s",
                          _("'info blockstats' not supported by this qemu"));
5710
        goto cleanup;
5711 5712 5713 5714 5715 5716 5717 5718 5719 5720 5721 5722 5723 5724 5725 5726 5727 5728 5729 5730 5731 5732 5733 5734 5735 5736
    }

    stats->rd_req = -1;
    stats->rd_bytes = -1;
    stats->wr_req = -1;
    stats->wr_bytes = -1;
    stats->errs = -1;

    /* The output format for both qemu & KVM is:
     *   blockdevice: rd_bytes=% wr_bytes=% rd_operations=% wr_operations=%
     *   (repeated for each block device)
     * where '%' is a 64 bit number.
     */
    p = info;

    while (*p) {
        if (STREQLEN (p, qemu_dev_name, len)
            && p[len] == ':' && p[len+1] == ' ') {

            eol = strchr (p, '\n');
            if (!eol)
                eol = p + strlen (p);

            p += len+2;         /* Skip to first label. */

            while (*p) {
5737
                if (STRPREFIX (p, "rd_bytes=")) {
5738 5739
                    p += 9;
                    if (virStrToLong_ll (p, &dummy, 10, &stats->rd_bytes) == -1)
5740 5741
                        DEBUG ("%s: error reading rd_bytes: %s",
                               vm->def->name, p);
5742
                } else if (STRPREFIX (p, "wr_bytes=")) {
5743 5744
                    p += 9;
                    if (virStrToLong_ll (p, &dummy, 10, &stats->wr_bytes) == -1)
5745 5746
                        DEBUG ("%s: error reading wr_bytes: %s",
                               vm->def->name, p);
5747
                } else if (STRPREFIX (p, "rd_operations=")) {
5748 5749
                    p += 14;
                    if (virStrToLong_ll (p, &dummy, 10, &stats->rd_req) == -1)
5750 5751
                        DEBUG ("%s: error reading rd_req: %s",
                               vm->def->name, p);
5752
                } else if (STRPREFIX (p, "wr_operations=")) {
5753 5754
                    p += 14;
                    if (virStrToLong_ll (p, &dummy, 10, &stats->wr_req) == -1)
5755 5756
                        DEBUG ("%s: error reading wr_req: %s",
                               vm->def->name, p);
5757
                } else
5758
                    DEBUG ("%s: unknown block stat near %s", vm->def->name, p);
5759 5760 5761 5762 5763 5764

                /* Skip to next label. */
                p = strchr (p, ' ');
                if (!p || p >= eol) break;
                p++;
            }
5765
            ret = 0;
5766
            goto cleanup;
5767 5768 5769 5770 5771 5772 5773 5774 5775 5776 5777
        }

        /* Skip to next line. */
        p = strchr (p, '\n');
        if (!p) break;
        p++;
    }

    /* If we reach here then the device was not found. */
    qemudReportError (dom->conn, dom, NULL, VIR_ERR_INVALID_ARG,
                      _("device not found: %s (%s)"), path, qemu_dev_name);
5778
 cleanup:
5779 5780
    VIR_FREE(qemu_dev_name);
    VIR_FREE(info);
5781 5782
    if (vm)
        virDomainObjUnlock(vm);
5783
    return ret;
5784 5785
}

5786
#ifdef __linux__
5787 5788 5789 5790 5791
static int
qemudDomainInterfaceStats (virDomainPtr dom,
                           const char *path,
                           struct _virDomainInterfaceStats *stats)
{
5792 5793
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
5794
    int i;
5795
    int ret = -1;
5796

5797
    qemuDriverLock(driver);
5798
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
5799 5800
    qemuDriverUnlock(driver);

5801
    if (!vm) {
5802 5803 5804 5805
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
5806
        goto cleanup;
5807 5808
    }

5809
    if (!virDomainIsActive(vm)) {
5810
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_INVALID,
5811
                         "%s", _("domain is not running"));
5812
        goto cleanup;
5813 5814 5815
    }

    /* Check the path is one of the domain's network interfaces. */
5816 5817
    for (i = 0 ; i < vm->def->nnets ; i++) {
        if (vm->def->nets[i]->ifname &&
5818 5819 5820 5821
            STREQ (vm->def->nets[i]->ifname, path)) {
            ret = 0;
            break;
        }
5822 5823
    }

5824 5825 5826 5827 5828
    if (ret == 0)
        ret = linuxDomainInterfaceStats (dom->conn, path, stats);
    else
        qemudReportError (dom->conn, dom, NULL, VIR_ERR_INVALID_ARG,
                          _("invalid path, '%s' is not a known interface"), path);
5829

5830
cleanup:
5831 5832
    if (vm)
        virDomainObjUnlock(vm);
5833 5834
    return ret;
}
5835
#else
5836 5837 5838 5839
static int
qemudDomainInterfaceStats (virDomainPtr dom,
                           const char *path ATTRIBUTE_UNUSED,
                           struct _virDomainInterfaceStats *stats ATTRIBUTE_UNUSED)
5840 5841 5842 5843
    qemudReportError (dom->conn, dom, NULL, VIR_ERR_NO_SUPPORT,
                      "%s", __FUNCTION__);
    return -1;
}
5844
#endif
5845

5846 5847 5848 5849 5850 5851 5852
static int
qemudDomainBlockPeek (virDomainPtr dom,
                      const char *path,
                      unsigned long long offset, size_t size,
                      void *buffer,
                      unsigned int flags ATTRIBUTE_UNUSED)
{
5853 5854 5855
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
    int fd = -1, ret = -1, i;
5856

5857
    qemuDriverLock(driver);
5858
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
5859 5860
    qemuDriverUnlock(driver);

5861
    if (!vm) {
5862 5863 5864 5865
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
5866
        goto cleanup;
5867 5868 5869 5870
    }

    if (!path || path[0] == '\0') {
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_INVALID_ARG,
J
Jim Meyering 已提交
5871
                         "%s", _("NULL or empty path"));
5872
        goto cleanup;
5873 5874 5875
    }

    /* Check the path belongs to this domain. */
5876 5877
    for (i = 0 ; i < vm->def->ndisks ; i++) {
        if (vm->def->disks[i]->src != NULL &&
5878 5879 5880 5881
            STREQ (vm->def->disks[i]->src, path)) {
            ret = 0;
            break;
        }
5882 5883
    }

5884 5885 5886 5887 5888
    if (ret == 0) {
        ret = -1;
        /* The path is correct, now try to open it and get its size. */
        fd = open (path, O_RDONLY);
        if (fd == -1) {
5889 5890
            virReportSystemError (dom->conn, errno,
                                  _("%s: failed to open"), path);
5891 5892
            goto cleanup;
        }
5893

5894 5895 5896 5897 5898 5899
        /* Seek and read. */
        /* NB. Because we configure with AC_SYS_LARGEFILE, off_t should
         * be 64 bits on all platforms.
         */
        if (lseek (fd, offset, SEEK_SET) == (off_t) -1 ||
            saferead (fd, buffer, size) == (ssize_t) -1) {
5900 5901
            virReportSystemError (dom->conn, errno,
                                  _("%s: failed to seek or read"), path);
5902 5903 5904 5905 5906 5907 5908
            goto cleanup;
        }

        ret = 0;
    } else {
        qemudReportError (dom->conn, dom, NULL, VIR_ERR_INVALID_ARG,
                          "%s", _("invalid path"));
5909 5910
    }

5911 5912 5913
cleanup:
    if (fd >= 0)
        close (fd);
5914 5915
    if (vm)
        virDomainObjUnlock(vm);
5916 5917 5918
    return ret;
}

R
Richard W.M. Jones 已提交
5919 5920 5921 5922 5923 5924
static int
qemudDomainMemoryPeek (virDomainPtr dom,
                       unsigned long long offset, size_t size,
                       void *buffer,
                       unsigned int flags)
{
5925 5926 5927
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
    char cmd[256], *info = NULL;
R
Richard W.M. Jones 已提交
5928 5929 5930
    char tmp[] = TEMPDIR "/qemu.mem.XXXXXX";
    int fd = -1, ret = -1;

5931
    qemuDriverLock(driver);
5932
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
5933
    qemuDriverUnlock(driver);
R
Richard W.M. Jones 已提交
5934 5935

    if (!vm) {
5936 5937 5938 5939
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
5940 5941 5942
        goto cleanup;
    }

5943
    if (flags != VIR_MEMORY_VIRTUAL && flags != VIR_MEMORY_PHYSICAL) {
5944
        qemudReportError (dom->conn, dom, NULL, VIR_ERR_INVALID_ARG,
5945
                     "%s", _("flags parameter must be VIR_MEMORY_VIRTUAL or VIR_MEMORY_PHYSICAL"));
5946
        goto cleanup;
R
Richard W.M. Jones 已提交
5947 5948
    }

5949
    if (!virDomainIsActive(vm)) {
5950
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_INVALID,
R
Richard W.M. Jones 已提交
5951
                         "%s", _("domain is not running"));
5952
        goto cleanup;
R
Richard W.M. Jones 已提交
5953 5954 5955 5956
    }

    /* Create a temporary filename. */
    if ((fd = mkstemp (tmp)) == -1) {
5957 5958
        virReportSystemError (dom->conn, errno,
                              _("mkstemp(\"%s\") failed"), tmp);
5959
        goto cleanup;
R
Richard W.M. Jones 已提交
5960 5961
    }

5962 5963 5964 5965 5966 5967 5968
    if (flags == VIR_MEMORY_VIRTUAL)
        /* Issue the memsave command. */
        snprintf (cmd, sizeof cmd, "memsave %llu %zi \"%s\"", offset, size, tmp);
    else
        /* Issue the pmemsave command. */
        snprintf (cmd, sizeof cmd, "pmemsave %llu %zi \"%s\"", offset, size, tmp);

5969
    if (qemudMonitorCommand (vm, cmd, &info) < 0) {
R
Richard W.M. Jones 已提交
5970
        qemudReportError (dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
5971
                          "%s", _("'memsave' command failed"));
5972
        goto cleanup;
R
Richard W.M. Jones 已提交
5973 5974
    }

5975
    DEBUG ("%s: (p)memsave reply: %s", vm->def->name, info);
R
Richard W.M. Jones 已提交
5976 5977 5978

    /* Read the memory file into buffer. */
    if (saferead (fd, buffer, size) == (ssize_t) -1) {
5979 5980 5981
        virReportSystemError (dom->conn, errno,
                              _("failed to read temporary file "
                                "created with template %s"), tmp);
5982
        goto cleanup;
R
Richard W.M. Jones 已提交
5983 5984 5985
    }

    ret = 0;
5986 5987 5988

cleanup:
    VIR_FREE(info);
R
Richard W.M. Jones 已提交
5989 5990
    if (fd >= 0) close (fd);
    unlink (tmp);
5991 5992
    if (vm)
        virDomainObjUnlock(vm);
R
Richard W.M. Jones 已提交
5993 5994 5995
    return ret;
}

5996

5997 5998
static int
qemudDomainEventRegister (virConnectPtr conn,
5999
                          virConnectDomainEventCallback callback,
6000 6001
                          void *opaque,
                          virFreeCallback freecb)
6002
{
6003 6004 6005
    struct qemud_driver *driver = conn->privateData;
    int ret;

6006
    qemuDriverLock(driver);
6007 6008
    ret = virDomainEventCallbackListAdd(conn, driver->domainEventCallbacks,
                                        callback, opaque, freecb);
6009
    qemuDriverUnlock(driver);
6010

6011
    return ret;
6012 6013 6014 6015
}

static int
qemudDomainEventDeregister (virConnectPtr conn,
6016
                            virConnectDomainEventCallback callback)
6017
{
6018 6019 6020
    struct qemud_driver *driver = conn->privateData;
    int ret;

6021
    qemuDriverLock(driver);
6022 6023 6024 6025 6026 6027
    if (driver->domainEventDispatching)
        ret = virDomainEventCallbackListMarkDelete(conn, driver->domainEventCallbacks,
                                                   callback);
    else
        ret = virDomainEventCallbackListRemove(conn, driver->domainEventCallbacks,
                                               callback);
6028
    qemuDriverUnlock(driver);
6029

6030
    return ret;
6031 6032
}

6033 6034 6035 6036 6037
static void qemuDomainEventDispatchFunc(virConnectPtr conn,
                                        virDomainEventPtr event,
                                        virConnectDomainEventCallback cb,
                                        void *cbopaque,
                                        void *opaque)
6038
{
6039
    struct qemud_driver *driver = opaque;
6040

6041 6042 6043 6044 6045 6046 6047 6048 6049 6050 6051 6052 6053 6054 6055 6056 6057 6058 6059 6060 6061 6062 6063 6064 6065 6066 6067 6068 6069 6070 6071 6072 6073 6074 6075 6076 6077 6078 6079 6080 6081 6082 6083 6084
    /* Drop the lock whle dispatching, for sake of re-entrancy */
    qemuDriverUnlock(driver);
    virDomainEventDispatchDefaultFunc(conn, event, cb, cbopaque, NULL);
    qemuDriverLock(driver);
}

static void qemuDomainEventFlush(int timer ATTRIBUTE_UNUSED, void *opaque)
{
    struct qemud_driver *driver = opaque;
    virDomainEventQueue tempQueue;

    qemuDriverLock(driver);

    driver->domainEventDispatching = 1;

    /* Copy the queue, so we're reentrant safe */
    tempQueue.count = driver->domainEventQueue->count;
    tempQueue.events = driver->domainEventQueue->events;
    driver->domainEventQueue->count = 0;
    driver->domainEventQueue->events = NULL;

    virEventUpdateTimeout(driver->domainEventTimer, -1);
    virDomainEventQueueDispatch(&tempQueue,
                                driver->domainEventCallbacks,
                                qemuDomainEventDispatchFunc,
                                driver);

    /* Purge any deleted callbacks */
    virDomainEventCallbackListPurgeMarked(driver->domainEventCallbacks);

    driver->domainEventDispatching = 0;
    qemuDriverUnlock(driver);
}


/* driver must be locked before calling */
static void qemuDomainEventQueue(struct qemud_driver *driver,
                                 virDomainEventPtr event)
{
    if (virDomainEventQueuePush(driver->domainEventQueue,
                                event) < 0)
        virDomainEventFree(event);
    if (qemu_driver->domainEventQueue->count == 1)
        virEventUpdateTimeout(driver->domainEventTimer, 0);
6085 6086
}

D
Daniel Veillard 已提交
6087 6088 6089 6090 6091 6092 6093 6094 6095 6096 6097 6098 6099 6100 6101 6102 6103 6104
/* Migration support. */

/* Prepare is the first step, and it runs on the destination host.
 *
 * This starts an empty VM listening on a TCP port.
 */
static int
qemudDomainMigratePrepare2 (virConnectPtr dconn,
                            char **cookie ATTRIBUTE_UNUSED,
                            int *cookielen ATTRIBUTE_UNUSED,
                            const char *uri_in,
                            char **uri_out,
                            unsigned long flags ATTRIBUTE_UNUSED,
                            const char *dname,
                            unsigned long resource ATTRIBUTE_UNUSED,
                            const char *dom_xml)
{
    static int port = 0;
6105 6106
    struct qemud_driver *driver = dconn->privateData;
    virDomainDefPtr def = NULL;
D
Daniel Veillard 已提交
6107 6108 6109 6110 6111
    virDomainObjPtr vm = NULL;
    int this_port;
    char hostname [HOST_NAME_MAX+1];
    char migrateFrom [64];
    const char *p;
6112
    virDomainEventPtr event = NULL;
6113 6114 6115
    int ret = -1;;

    *uri_out = NULL;
D
Daniel Veillard 已提交
6116

6117
    qemuDriverLock(driver);
D
Daniel Veillard 已提交
6118 6119 6120
    if (!dom_xml) {
        qemudReportError (dconn, NULL, NULL, VIR_ERR_INTERNAL_ERROR,
                          "%s", _("no domain XML passed"));
6121
        goto cleanup;
D
Daniel Veillard 已提交
6122 6123 6124 6125 6126 6127 6128 6129 6130 6131 6132 6133 6134 6135 6136 6137 6138 6139
    }

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

        /* Get hostname */
        if (gethostname (hostname, HOST_NAME_MAX+1) == -1) {
6140 6141
            virReportSystemError (dconn, errno,
                                  "%s", _("failed to determine host name"));
6142
            goto cleanup;
D
Daniel Veillard 已提交
6143 6144 6145
        }

        /* Caller frees */
6146
        if (virAsprintf(uri_out, "tcp:%s:%d", hostname, this_port) < 0) {
6147
            virReportOOMError (dconn);
6148
            goto cleanup;
D
Daniel Veillard 已提交
6149 6150 6151 6152 6153 6154
        }
    } else {
        /* Check the URI starts with "tcp:".  We will escape the
         * URI when passing it to the qemu monitor, so bad
         * characters in hostname part don't matter.
         */
6155
        if (!STRPREFIX (uri_in, "tcp:")) {
D
Daniel Veillard 已提交
6156 6157
            qemudReportError (dconn, NULL, NULL, VIR_ERR_INVALID_ARG,
                  "%s", _("only tcp URIs are supported for KVM migrations"));
6158
            goto cleanup;
D
Daniel Veillard 已提交
6159 6160 6161 6162 6163 6164 6165 6166 6167
        }

        /* Get the port number. */
        p = strrchr (uri_in, ':');
        p++; /* definitely has a ':' in it, see above */
        this_port = virParseNumber (&p);
        if (this_port == -1 || p-uri_in != strlen (uri_in)) {
            qemudReportError (dconn, NULL, NULL, VIR_ERR_INVALID_ARG,
                              "%s", _("URI did not have ':port' at the end"));
6168
            goto cleanup;
D
Daniel Veillard 已提交
6169 6170 6171 6172
        }
    }

    /* Parse the domain XML. */
6173 6174
    if (!(def = virDomainDefParseString(dconn, driver->caps, dom_xml,
                                        VIR_DOMAIN_XML_INACTIVE))) {
D
Daniel Veillard 已提交
6175 6176
        qemudReportError (dconn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
                          "%s", _("failed to parse XML"));
6177
        goto cleanup;
D
Daniel Veillard 已提交
6178 6179 6180 6181 6182 6183 6184 6185 6186 6187 6188 6189 6190 6191 6192 6193
    }

    /* Target domain name, maybe renamed. */
    dname = dname ? dname : def->name;

#if 1
    /* Ensure the name and UUID don't already exist in an active VM */
    vm = virDomainFindByUUID(&driver->domains, def->uuid);
#else
    /* For TESTING ONLY you can change #if 1 -> #if 0 above and use
     * this code which lets you do localhost migrations.  You must still
     * supply a fresh 'dname' but this code assigns a random UUID.
     */
    if (virUUIDGenerate (def->uuid) == -1) {
        qemudReportError (dconn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
            _("could not generate random UUID"));
6194
        goto cleanup;
D
Daniel Veillard 已提交
6195 6196 6197 6198 6199 6200 6201 6202 6203
    }
#endif

    if (!vm) vm = virDomainFindByName(&driver->domains, dname);
    if (vm) {
        if (virDomainIsActive(vm)) {
            qemudReportError (dconn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
                              _("domain with the same name or UUID already exists as '%s'"),
                              vm->def->name);
6204
            goto cleanup;
D
Daniel Veillard 已提交
6205
        }
6206
        virDomainObjUnlock(vm);
D
Daniel Veillard 已提交
6207 6208 6209 6210 6211 6212 6213
    }

    if (!(vm = virDomainAssignDef(dconn,
                                  &driver->domains,
                                  def))) {
        qemudReportError (dconn, NULL, NULL, VIR_ERR_OPERATION_FAILED,
                          "%s", _("failed to assign new VM"));
6214
        goto cleanup;
D
Daniel Veillard 已提交
6215
    }
6216
    def = NULL;
D
Daniel Veillard 已提交
6217 6218 6219 6220 6221 6222 6223 6224

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

    /* Start the QEMU daemon, with the same command-line arguments plus
     * -incoming tcp:0.0.0.0:port
     */
    snprintf (migrateFrom, sizeof (migrateFrom), "tcp:0.0.0.0:%d", this_port);
6225
    if (qemudStartVMDaemon (dconn, driver, vm, migrateFrom, -1) < 0) {
6226
        if (!vm->persistent) {
D
Daniel Veillard 已提交
6227
            virDomainRemoveInactive(&driver->domains, vm);
6228 6229
            vm = NULL;
        }
6230
        goto cleanup;
D
Daniel Veillard 已提交
6231
    }
6232 6233 6234 6235

    event = virDomainEventNewFromObj(vm,
                                     VIR_DOMAIN_EVENT_STARTED,
                                     VIR_DOMAIN_EVENT_STARTED_MIGRATED);
6236
    ret = 0;
D
Daniel Veillard 已提交
6237

6238 6239 6240 6241 6242
cleanup:
    virDomainDefFree(def);
    if (ret != 0) {
        VIR_FREE(*uri_out);
    }
6243 6244
    if (vm)
        virDomainObjUnlock(vm);
6245 6246
    if (event)
        qemuDomainEventQueue(driver, event);
6247
    qemuDriverUnlock(driver);
6248
    return ret;
D
Daniel Veillard 已提交
6249 6250 6251 6252 6253 6254 6255 6256 6257 6258 6259 6260
}

/* Perform is the second step, and it runs on the source host. */
static int
qemudDomainMigratePerform (virDomainPtr dom,
                           const char *cookie ATTRIBUTE_UNUSED,
                           int cookielen ATTRIBUTE_UNUSED,
                           const char *uri,
                           unsigned long flags ATTRIBUTE_UNUSED,
                           const char *dname ATTRIBUTE_UNUSED,
                           unsigned long resource)
{
6261 6262
    struct qemud_driver *driver = dom->conn->privateData;
    virDomainObjPtr vm;
6263
    virDomainEventPtr event = NULL;
D
Daniel Veillard 已提交
6264 6265
    char *safe_uri;
    char cmd[HOST_NAME_MAX+50];
6266 6267
    char *info = NULL;
    int ret = -1;
6268
    int paused = 0;
D
Daniel Veillard 已提交
6269

6270
    qemuDriverLock(driver);
6271
    vm = virDomainFindByUUID(&driver->domains, dom->uuid);
D
Daniel Veillard 已提交
6272
    if (!vm) {
6273 6274 6275 6276
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(dom->uuid, uuidstr);
        qemudReportError(dom->conn, dom, NULL, VIR_ERR_NO_DOMAIN,
                         _("no domain with matching uuid '%s'"), uuidstr);
6277
        goto cleanup;
D
Daniel Veillard 已提交
6278 6279 6280
    }

    if (!virDomainIsActive(vm)) {
6281
        qemudReportError (dom->conn, dom, NULL, VIR_ERR_OPERATION_INVALID,
D
Daniel Veillard 已提交
6282
                          "%s", _("domain is not running"));
6283
        goto cleanup;
D
Daniel Veillard 已提交
6284 6285
    }

6286 6287
    if (!(flags & VIR_MIGRATE_LIVE)) {
        /* Pause domain for non-live migration */
6288 6289 6290 6291 6292
        if (qemudMonitorCommand (vm, "stop", &info) < 0) {
            qemudReportError(dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
                             "%s", _("off-line migration specified, but suspend operation failed"));
            goto cleanup;
        }
6293
        DEBUG ("%s: stop reply: %s", vm->def->name, info);
6294
        VIR_FREE(info);
6295
        paused = 1;
6296

6297 6298 6299 6300 6301 6302
        event = virDomainEventNewFromObj(vm,
                                         VIR_DOMAIN_EVENT_SUSPENDED,
                                         VIR_DOMAIN_EVENT_SUSPENDED_MIGRATED);
        if (event)
            qemuDomainEventQueue(driver, event);
        event = NULL;
6303 6304
    }

D
Daniel Veillard 已提交
6305 6306 6307
    if (resource > 0) {
        /* Issue migrate_set_speed command.  Don't worry if it fails. */
        snprintf (cmd, sizeof cmd, "migrate_set_speed %lum", resource);
6308
        qemudMonitorCommand (vm, cmd, &info);
D
Daniel Veillard 已提交
6309

6310
        DEBUG ("%s: migrate_set_speed reply: %s", vm->def->name, info);
D
Daniel Veillard 已提交
6311 6312 6313 6314 6315 6316
        VIR_FREE (info);
    }

    /* Issue the migrate command. */
    safe_uri = qemudEscapeMonitorArg (uri);
    if (!safe_uri) {
6317
        virReportOOMError (dom->conn);
6318
        goto cleanup;
D
Daniel Veillard 已提交
6319 6320 6321 6322
    }
    snprintf (cmd, sizeof cmd, "migrate \"%s\"", safe_uri);
    VIR_FREE (safe_uri);

6323
    if (qemudMonitorCommand (vm, cmd, &info) < 0) {
D
Daniel Veillard 已提交
6324 6325
        qemudReportError (dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
                          "%s", _("migrate operation failed"));
6326
        goto cleanup;
D
Daniel Veillard 已提交
6327 6328
    }

6329
    DEBUG ("%s: migrate reply: %s", vm->def->name, info);
D
Daniel Veillard 已提交
6330 6331 6332 6333 6334

    /* Now check for "fail" in the output string */
    if (strstr(info, "fail") != NULL) {
        qemudReportError (dom->conn, dom, NULL, VIR_ERR_OPERATION_FAILED,
                          _("migrate failed: %s"), info);
6335
        goto cleanup;
D
Daniel Veillard 已提交
6336 6337 6338 6339
    }

    /* Clean up the source domain. */
    qemudShutdownVMDaemon (dom->conn, driver, vm);
6340
    paused = 0;
6341 6342 6343 6344

    event = virDomainEventNewFromObj(vm,
                                     VIR_DOMAIN_EVENT_STOPPED,
                                     VIR_DOMAIN_EVENT_STOPPED_MIGRATED);
6345
    if (!vm->persistent) {
D
Daniel Veillard 已提交
6346
        virDomainRemoveInactive(&driver->domains, vm);
6347 6348
        vm = NULL;
    }
6349
    ret = 0;
D
Daniel Veillard 已提交
6350

6351
cleanup:
6352 6353 6354
    /* Note that we have to free info *first*, since we are re-using the
     * variable below (and otherwise might cause a memory leak)
     */
6355
    VIR_FREE(info);
6356 6357 6358 6359 6360 6361 6362 6363

    if (paused) {
        /* we got here through some sort of failure; start the domain again */
        if (qemudMonitorCommand (vm, "cont", &info) < 0) {
            /* Hm, we already know we are in error here.  We don't want to
             * overwrite the previous error, though, so we just throw something
             * to the logs and hope for the best
             */
6364 6365
            VIR_ERROR(_("Failed to resume guest %s after failure\n"),
                      vm->def->name);
6366 6367
        }
        else {
6368
            DEBUG ("%s: cont reply: %s", vm->def->name, info);
6369 6370 6371 6372 6373 6374 6375 6376
            VIR_FREE(info);
        }

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

6377 6378
    if (vm)
        virDomainObjUnlock(vm);
6379 6380
    if (event)
        qemuDomainEventQueue(driver, event);
6381
    qemuDriverUnlock(driver);
6382
    return ret;
D
Daniel Veillard 已提交
6383 6384 6385 6386 6387 6388 6389 6390 6391 6392 6393 6394
}

/* Finish is the third and final step, and it runs on the destination host. */
static virDomainPtr
qemudDomainMigrateFinish2 (virConnectPtr dconn,
                           const char *dname,
                           const char *cookie ATTRIBUTE_UNUSED,
                           int cookielen ATTRIBUTE_UNUSED,
                           const char *uri ATTRIBUTE_UNUSED,
                           unsigned long flags ATTRIBUTE_UNUSED,
                           int retcode)
{
6395 6396 6397
    struct qemud_driver *driver = dconn->privateData;
    virDomainObjPtr vm;
    virDomainPtr dom = NULL;
6398
    virDomainEventPtr event = NULL;
D
Daniel Veillard 已提交
6399 6400
    char *info = NULL;

6401
    qemuDriverLock(driver);
6402
    vm = virDomainFindByName(&driver->domains, dname);
D
Daniel Veillard 已提交
6403
    if (!vm) {
6404 6405
        qemudReportError (dconn, NULL, NULL, VIR_ERR_NO_DOMAIN,
                          _("no domain with matching name '%s'"), dname);
6406
        goto cleanup;
D
Daniel Veillard 已提交
6407 6408 6409 6410 6411 6412 6413 6414 6415
    }

    /* Did the migration go as planned?  If yes, return the domain
     * object, but if no, clean up the empty qemu process.
     */
    if (retcode == 0) {
        dom = virGetDomain (dconn, vm->def->name, vm->def->uuid);
        VIR_FREE(info);
        vm->state = VIR_DOMAIN_RUNNING;
6416 6417 6418
        event = virDomainEventNewFromObj(vm,
                                         VIR_DOMAIN_EVENT_RESUMED,
                                         VIR_DOMAIN_EVENT_RESUMED_MIGRATED);
D
Daniel Veillard 已提交
6419
        virDomainSaveStatus(dconn, driver->stateDir, vm);
D
Daniel Veillard 已提交
6420 6421
    } else {
        qemudShutdownVMDaemon (dconn, driver, vm);
6422 6423 6424
        event = virDomainEventNewFromObj(vm,
                                         VIR_DOMAIN_EVENT_STOPPED,
                                         VIR_DOMAIN_EVENT_STOPPED_FAILED);
6425
        if (!vm->persistent) {
D
Daniel Veillard 已提交
6426
            virDomainRemoveInactive(&driver->domains, vm);
6427 6428
            vm = NULL;
        }
D
Daniel Veillard 已提交
6429
    }
6430 6431

cleanup:
6432 6433
    if (vm)
        virDomainObjUnlock(vm);
6434 6435
    if (event)
        qemuDomainEventQueue(driver, event);
6436
    qemuDriverUnlock(driver);
6437
    return dom;
D
Daniel Veillard 已提交
6438 6439
}

6440 6441 6442 6443 6444 6445 6446 6447 6448 6449 6450 6451 6452 6453 6454 6455
static int
qemudNodeDeviceGetPciInfo (virNodeDevicePtr dev,
                           unsigned *domain,
                           unsigned *bus,
                           unsigned *slot,
                           unsigned *function)
{
    virNodeDeviceDefPtr def = NULL;
    virNodeDevCapsDefPtr cap;
    char *xml = NULL;
    int ret = -1;

    xml = virNodeDeviceGetXMLDesc(dev, 0);
    if (!xml)
        goto out;

6456
    def = virNodeDeviceDefParseString(dev->conn, xml, EXISTING_DEVICE);
6457 6458 6459 6460 6461 6462 6463 6464 6465 6466 6467 6468 6469 6470 6471 6472 6473 6474 6475 6476 6477 6478 6479 6480 6481 6482 6483 6484 6485 6486 6487 6488 6489 6490 6491 6492 6493 6494 6495 6496 6497 6498 6499 6500 6501 6502 6503 6504 6505 6506 6507 6508 6509 6510 6511 6512 6513 6514 6515 6516 6517 6518 6519 6520 6521 6522 6523 6524 6525 6526 6527 6528 6529 6530 6531 6532 6533 6534 6535 6536 6537 6538 6539 6540 6541 6542 6543 6544 6545 6546 6547 6548 6549 6550 6551 6552 6553 6554
    if (!def)
        goto out;

    cap = def->caps;
    while (cap) {
        if (cap->type == VIR_NODE_DEV_CAP_PCI_DEV) {
            *domain   = cap->data.pci_dev.domain;
            *bus      = cap->data.pci_dev.bus;
            *slot     = cap->data.pci_dev.slot;
            *function = cap->data.pci_dev.function;
            break;
        }

        cap = cap->next;
    }

    if (!cap) {
        qemudReportError(dev->conn, NULL, NULL, VIR_ERR_INVALID_ARG,
                         _("device %s is not a PCI device"), dev->name);
        goto out;
    }

    ret = 0;
out:
    virNodeDeviceDefFree(def);
    VIR_FREE(xml);
    return ret;
}

static int
qemudNodeDeviceDettach (virNodeDevicePtr dev)
{
    pciDevice *pci;
    unsigned domain, bus, slot, function;
    int ret = -1;

    if (qemudNodeDeviceGetPciInfo(dev, &domain, &bus, &slot, &function) < 0)
        return -1;

    pci = pciGetDevice(dev->conn, domain, bus, slot, function);
    if (!pci)
        return -1;

    if (pciDettachDevice(dev->conn, pci) < 0)
        goto out;

    ret = 0;
out:
    pciFreeDevice(dev->conn, pci);
    return ret;
}

static int
qemudNodeDeviceReAttach (virNodeDevicePtr dev)
{
    pciDevice *pci;
    unsigned domain, bus, slot, function;
    int ret = -1;

    if (qemudNodeDeviceGetPciInfo(dev, &domain, &bus, &slot, &function) < 0)
        return -1;

    pci = pciGetDevice(dev->conn, domain, bus, slot, function);
    if (!pci)
        return -1;

    if (pciReAttachDevice(dev->conn, pci) < 0)
        goto out;

    ret = 0;
out:
    pciFreeDevice(dev->conn, pci);
    return ret;
}

static int
qemudNodeDeviceReset (virNodeDevicePtr dev)
{
    pciDevice *pci;
    unsigned domain, bus, slot, function;
    int ret = -1;

    if (qemudNodeDeviceGetPciInfo(dev, &domain, &bus, &slot, &function) < 0)
        return -1;

    pci = pciGetDevice(dev->conn, domain, bus, slot, function);
    if (!pci)
        return -1;

    if (pciResetDevice(dev->conn, pci) < 0)
        goto out;

    ret = 0;
out:
    pciFreeDevice(dev->conn, pci);
    return ret;
}

6555 6556 6557 6558 6559
static virDriver qemuDriver = {
    VIR_DRV_QEMU,
    "QEMU",
    qemudOpen, /* open */
    qemudClose, /* close */
D
Daniel Veillard 已提交
6560
    qemudSupportsFeature, /* supports_feature */
6561 6562
    qemudGetType, /* type */
    qemudGetVersion, /* version */
6563
    qemudGetHostname, /* getHostname */
6564
    qemudGetMaxVCPUs, /* getMaxVcpus */
6565
    nodeGetInfo, /* nodeGetInfo */
6566 6567 6568
    qemudGetCapabilities, /* getCapabilities */
    qemudListDomains, /* listDomains */
    qemudNumDomains, /* numOfDomains */
6569
    qemudDomainCreate, /* domainCreateXML */
6570 6571 6572 6573 6574
    qemudDomainLookupByID, /* domainLookupByID */
    qemudDomainLookupByUUID, /* domainLookupByUUID */
    qemudDomainLookupByName, /* domainLookupByName */
    qemudDomainSuspend, /* domainSuspend */
    qemudDomainResume, /* domainResume */
6575
    qemudDomainShutdown, /* domainShutdown */
6576 6577 6578
    NULL, /* domainReboot */
    qemudDomainDestroy, /* domainDestroy */
    qemudDomainGetOSType, /* domainGetOSType */
6579 6580 6581
    qemudDomainGetMaxMemory, /* domainGetMaxMemory */
    qemudDomainSetMaxMemory, /* domainSetMaxMemory */
    qemudDomainSetMemory, /* domainSetMemory */
6582 6583 6584
    qemudDomainGetInfo, /* domainGetInfo */
    qemudDomainSave, /* domainSave */
    qemudDomainRestore, /* domainRestore */
P
Paolo Bonzini 已提交
6585
    qemudDomainCoreDump, /* domainCoreDump */
6586
    qemudDomainSetVcpus, /* domainSetVcpus */
6587 6588 6589 6590
#if HAVE_SCHED_GETAFFINITY
    qemudDomainPinVcpu, /* domainPinVcpu */
    qemudDomainGetVcpus, /* domainGetVcpus */
#else
6591 6592
    NULL, /* domainPinVcpu */
    NULL, /* domainGetVcpus */
6593
#endif
6594
    qemudDomainGetMaxVcpus, /* domainGetMaxVcpus */
6595 6596
    qemudDomainGetSecurityLabel, /* domainGetSecurityLabel */
    qemudNodeGetSecurityModel, /* nodeGetSecurityModel */
6597
    qemudDomainDumpXML, /* domainDumpXML */
6598
    qemuDomainXMLFromNative, /* domainXmlFromNative */
6599
    qemuDomainXMLToNative, /* domainXMLToNative */
6600 6601
    qemudListDefinedDomains, /* listDefinedDomains */
    qemudNumDefinedDomains, /* numOfDefinedDomains */
6602 6603 6604
    qemudDomainStart, /* domainCreate */
    qemudDomainDefine, /* domainDefineXML */
    qemudDomainUndefine, /* domainUndefine */
6605
    qemudDomainAttachDevice, /* domainAttachDevice */
6606
    qemudDomainDetachDevice, /* domainDetachDevice */
6607 6608
    qemudDomainGetAutostart, /* domainGetAutostart */
    qemudDomainSetAutostart, /* domainSetAutostart */
6609 6610 6611
    qemuGetSchedulerType, /* domainGetSchedulerType */
    qemuGetSchedulerParameters, /* domainGetSchedulerParameters */
    qemuSetSchedulerParameters, /* domainSetSchedulerParameters */
D
Daniel Veillard 已提交
6612 6613
    NULL, /* domainMigratePrepare (v1) */
    qemudDomainMigratePerform, /* domainMigratePerform */
6614
    NULL, /* domainMigrateFinish */
6615
    qemudDomainBlockStats, /* domainBlockStats */
6616
    qemudDomainInterfaceStats, /* domainInterfaceStats */
6617
    qemudDomainBlockPeek, /* domainBlockPeek */
R
Richard W.M. Jones 已提交
6618
    qemudDomainMemoryPeek, /* domainMemoryPeek */
6619 6620
    nodeGetCellsFreeMemory, /* nodeGetCellsFreeMemory */
    nodeGetFreeMemory,  /* getFreeMemory */
6621 6622
    qemudDomainEventRegister, /* domainEventRegister */
    qemudDomainEventDeregister, /* domainEventDeregister */
D
Daniel Veillard 已提交
6623 6624
    qemudDomainMigratePrepare2, /* domainMigratePrepare2 */
    qemudDomainMigrateFinish2, /* domainMigrateFinish2 */
6625 6626 6627
    qemudNodeDeviceDettach, /* nodeDeviceDettach */
    qemudNodeDeviceReAttach, /* nodeDeviceReAttach */
    qemudNodeDeviceReset, /* nodeDeviceReset */
6628 6629 6630
};


6631
static virStateDriver qemuStateDriver = {
6632 6633 6634 6635
    .initialize = qemudStartup,
    .cleanup = qemudShutdown,
    .reload = qemudReload,
    .active = qemudActive,
6636
};
6637

6638
int qemuRegister(void) {
6639 6640 6641 6642
    virRegisterDriver(&qemuDriver);
    virRegisterStateDriver(&qemuStateDriver);
    return 0;
}