qemu_domain.c 204.6 KB
Newer Older
1
/*
2
 * qemu_domain.c: QEMU domain private state
3
 *
4
 * Copyright (C) 2006-2016 Red Hat, Inc.
5 6 7 8 9 10 11 12 13 14 15 16 17
 * 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
18
 * License along with this library.  If not, see
O
Osier Yang 已提交
19
 * <http://www.gnu.org/licenses/>.
20 21 22 23 24 25 26
 *
 * Author: Daniel P. Berrange <berrange@redhat.com>
 */

#include <config.h>

#include "qemu_domain.h"
J
John Ferlan 已提交
27
#include "qemu_alias.h"
28
#include "qemu_cgroup.h"
29
#include "qemu_command.h"
30
#include "qemu_process.h"
31
#include "qemu_parse_command.h"
32
#include "qemu_capabilities.h"
33
#include "qemu_migration.h"
34
#include "viralloc.h"
35
#include "virlog.h"
36
#include "virerror.h"
37
#include "c-ctype.h"
38
#include "cpu/cpu.h"
39
#include "viruuid.h"
E
Eric Blake 已提交
40
#include "virfile.h"
41
#include "domain_addr.h"
42
#include "domain_event.h"
43
#include "virtime.h"
44
#include "virstoragefile.h"
45
#include "virstring.h"
46
#include "virthreadjob.h"
47
#include "viratomic.h"
48
#include "virprocess.h"
49
#include "vircrypto.h"
50
#include "secret_util.h"
51
#include "logging/log_manager.h"
52
#include "locking/domain_lock.h"
53

54 55
#include "storage/storage_driver.h"

56
#include <sys/time.h>
57
#include <fcntl.h>
58

59 60 61 62
#include <libxml/xpathInternals.h>

#define VIR_FROM_THIS VIR_FROM_QEMU

63 64
VIR_LOG_INIT("qemu.qemu_domain");

65 66
#define QEMU_NAMESPACE_HREF "http://libvirt.org/schemas/domain/qemu/1.0"

67 68 69 70 71 72
VIR_ENUM_IMPL(qemuDomainJob, QEMU_JOB_LAST,
              "none",
              "query",
              "destroy",
              "suspend",
              "modify",
73
              "abort",
74
              "migration operation",
75 76 77 78 79 80 81 82 83 84
              "none",   /* async job is never stored in job.active */
              "async nested",
);

VIR_ENUM_IMPL(qemuDomainAsyncJob, QEMU_ASYNC_JOB_LAST,
              "none",
              "migration out",
              "migration in",
              "save",
              "dump",
85
              "snapshot",
86
              "start",
87 88
);

89

90 91 92
struct _qemuDomainLogContext {
    int refs;
    int writefd;
93
    int readfd; /* Only used if manager == NULL */
94
    off_t pos;
95
    ino_t inode; /* Only used if manager != NULL */
96
    char *path;
97
    virLogManagerPtr manager;
98 99
};

J
Jiri Denemark 已提交
100
const char *
101
qemuDomainAsyncJobPhaseToString(qemuDomainAsyncJob job,
J
Jiri Denemark 已提交
102 103 104 105 106
                                int phase ATTRIBUTE_UNUSED)
{
    switch (job) {
    case QEMU_ASYNC_JOB_MIGRATION_OUT:
    case QEMU_ASYNC_JOB_MIGRATION_IN:
107 108
        return qemuMigrationJobPhaseTypeToString(phase);

J
Jiri Denemark 已提交
109 110
    case QEMU_ASYNC_JOB_SAVE:
    case QEMU_ASYNC_JOB_DUMP:
111
    case QEMU_ASYNC_JOB_SNAPSHOT:
112
    case QEMU_ASYNC_JOB_START:
J
Jiri Denemark 已提交
113 114 115 116 117 118 119 120 121
    case QEMU_ASYNC_JOB_NONE:
    case QEMU_ASYNC_JOB_LAST:
        ; /* fall through */
    }

    return "none";
}

int
122
qemuDomainAsyncJobPhaseFromString(qemuDomainAsyncJob job,
J
Jiri Denemark 已提交
123 124 125 126 127 128 129 130
                                  const char *phase)
{
    if (!phase)
        return 0;

    switch (job) {
    case QEMU_ASYNC_JOB_MIGRATION_OUT:
    case QEMU_ASYNC_JOB_MIGRATION_IN:
131 132
        return qemuMigrationJobPhaseTypeFromString(phase);

J
Jiri Denemark 已提交
133 134
    case QEMU_ASYNC_JOB_SAVE:
    case QEMU_ASYNC_JOB_DUMP:
135
    case QEMU_ASYNC_JOB_SNAPSHOT:
136
    case QEMU_ASYNC_JOB_START:
J
Jiri Denemark 已提交
137 138 139 140 141 142 143 144 145 146 147
    case QEMU_ASYNC_JOB_NONE:
    case QEMU_ASYNC_JOB_LAST:
        ; /* fall through */
    }

    if (STREQ(phase, "none"))
        return 0;
    else
        return -1;
}

148

149
void qemuDomainEventQueue(virQEMUDriverPtr driver,
150
                          virObjectEventPtr event)
151
{
152 153
    if (event)
        virObjectEventStateQueue(driver->domainEventState, event);
154 155 156
}


157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180
void
qemuDomainEventEmitJobCompleted(virQEMUDriverPtr driver,
                                virDomainObjPtr vm)
{
    qemuDomainObjPrivatePtr priv = vm->privateData;
    virObjectEventPtr event;
    virTypedParameterPtr params = NULL;
    int nparams = 0;
    int type;

    if (!priv->job.completed)
        return;

    if (qemuDomainJobInfoToParams(priv->job.completed, &type,
                                  &params, &nparams) < 0) {
        VIR_WARN("Could not get stats for completed job; domain %s",
                 vm->def->name);
    }

    event = virDomainEventJobCompletedNewFromObj(vm, params, nparams);
    qemuDomainEventQueue(driver, event);
}


181 182 183 184 185 186 187 188
static int
qemuDomainObjInitJob(qemuDomainObjPrivatePtr priv)
{
    memset(&priv->job, 0, sizeof(priv->job));

    if (virCondInit(&priv->job.cond) < 0)
        return -1;

189
    if (virCondInit(&priv->job.asyncCond) < 0) {
190
        virCondDestroy(&priv->job.cond);
191 192 193
        return -1;
    }

194 195 196 197 198 199 200 201 202
    return 0;
}

static void
qemuDomainObjResetJob(qemuDomainObjPrivatePtr priv)
{
    struct qemuDomainJobObj *job = &priv->job;

    job->active = QEMU_JOB_NONE;
203
    job->owner = 0;
204
    job->ownerAPI = NULL;
J
Jiri Denemark 已提交
205
    job->started = 0;
206 207 208 209 210 211 212 213
}

static void
qemuDomainObjResetAsyncJob(qemuDomainObjPrivatePtr priv)
{
    struct qemuDomainJobObj *job = &priv->job;

    job->asyncJob = QEMU_ASYNC_JOB_NONE;
214
    job->asyncOwner = 0;
215
    job->asyncOwnerAPI = NULL;
J
Jiri Denemark 已提交
216
    job->asyncStarted = 0;
J
Jiri Denemark 已提交
217
    job->phase = 0;
218
    job->mask = QEMU_JOB_DEFAULT_MASK;
219
    job->dump_memory_only = false;
220
    job->abortJob = false;
221
    job->spiceMigration = false;
222
    job->spiceMigrated = false;
223
    job->postcopyEnabled = false;
J
Jiri Denemark 已提交
224
    VIR_FREE(job->current);
225 226
}

227 228 229 230 231 232 233 234
void
qemuDomainObjRestoreJob(virDomainObjPtr obj,
                        struct qemuDomainJobObj *job)
{
    qemuDomainObjPrivatePtr priv = obj->privateData;

    memset(job, 0, sizeof(*job));
    job->active = priv->job.active;
235
    job->owner = priv->job.owner;
236
    job->asyncJob = priv->job.asyncJob;
237
    job->asyncOwner = priv->job.asyncOwner;
J
Jiri Denemark 已提交
238
    job->phase = priv->job.phase;
239 240 241 242 243

    qemuDomainObjResetJob(priv);
    qemuDomainObjResetAsyncJob(priv);
}

244 245 246
static void
qemuDomainObjFreeJob(qemuDomainObjPrivatePtr priv)
{
J
Jiri Denemark 已提交
247
    VIR_FREE(priv->job.current);
248
    VIR_FREE(priv->job.completed);
249 250
    virCondDestroy(&priv->job.cond);
    virCondDestroy(&priv->job.asyncCond);
251 252
}

253
static bool
254
qemuDomainTrackJob(qemuDomainJob job)
255 256 257 258
{
    return (QEMU_DOMAIN_TRACK_JOBS & JOB_MASK(job)) != 0;
}

259

J
Jiri Denemark 已提交
260 261 262 263 264 265 266 267 268 269 270
int
qemuDomainJobInfoUpdateTime(qemuDomainJobInfoPtr jobInfo)
{
    unsigned long long now;

    if (!jobInfo->started)
        return 0;

    if (virTimeMillisNow(&now) < 0)
        return -1;

271 272 273 274 275 276
    if (now < jobInfo->started) {
        VIR_WARN("Async job starts in the future");
        jobInfo->started = 0;
        return 0;
    }

J
Jiri Denemark 已提交
277 278 279 280
    jobInfo->timeElapsed = now - jobInfo->started;
    return 0;
}

281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297
int
qemuDomainJobInfoUpdateDowntime(qemuDomainJobInfoPtr jobInfo)
{
    unsigned long long now;

    if (!jobInfo->stopped)
        return 0;

    if (virTimeMillisNow(&now) < 0)
        return -1;

    if (now < jobInfo->stopped) {
        VIR_WARN("Guest's CPUs stopped in the future");
        jobInfo->stopped = 0;
        return 0;
    }

298 299
    jobInfo->stats.downtime = now - jobInfo->stopped;
    jobInfo->stats.downtime_set = true;
300 301 302
    return 0;
}

J
Jiri Denemark 已提交
303 304 305 306
int
qemuDomainJobInfoToInfo(qemuDomainJobInfoPtr jobInfo,
                        virDomainJobInfoPtr info)
{
307
    info->type = jobInfo->type;
J
Jiri Denemark 已提交
308 309 310
    info->timeElapsed = jobInfo->timeElapsed;
    info->timeRemaining = jobInfo->timeRemaining;

311 312 313
    info->memTotal = jobInfo->stats.ram_total;
    info->memRemaining = jobInfo->stats.ram_remaining;
    info->memProcessed = jobInfo->stats.ram_transferred;
J
Jiri Denemark 已提交
314

315 316 317
    info->fileTotal = jobInfo->stats.disk_total;
    info->fileRemaining = jobInfo->stats.disk_remaining;
    info->fileProcessed = jobInfo->stats.disk_transferred;
J
Jiri Denemark 已提交
318 319 320 321 322 323 324 325 326 327 328 329 330 331

    info->dataTotal = info->memTotal + info->fileTotal;
    info->dataRemaining = info->memRemaining + info->fileRemaining;
    info->dataProcessed = info->memProcessed + info->fileProcessed;

    return 0;
}

int
qemuDomainJobInfoToParams(qemuDomainJobInfoPtr jobInfo,
                          int *type,
                          virTypedParameterPtr *params,
                          int *nparams)
{
332
    qemuMonitorMigrationStats *stats = &jobInfo->stats;
J
Jiri Denemark 已提交
333 334 335 336 337 338 339 340 341
    virTypedParameterPtr par = NULL;
    int maxpar = 0;
    int npar = 0;

    if (virTypedParamsAddULLong(&par, &npar, &maxpar,
                                VIR_DOMAIN_JOB_TIME_ELAPSED,
                                jobInfo->timeElapsed) < 0)
        goto error;

342 343 344 345 346 347 348
    if (jobInfo->timeDeltaSet &&
        jobInfo->timeElapsed > jobInfo->timeDelta &&
        virTypedParamsAddULLong(&par, &npar, &maxpar,
                                VIR_DOMAIN_JOB_TIME_ELAPSED_NET,
                                jobInfo->timeElapsed - jobInfo->timeDelta) < 0)
        goto error;

J
Jiri Denemark 已提交
349 350 351 352 353 354
    if (jobInfo->type == VIR_DOMAIN_JOB_BOUNDED &&
        virTypedParamsAddULLong(&par, &npar, &maxpar,
                                VIR_DOMAIN_JOB_TIME_REMAINING,
                                jobInfo->timeRemaining) < 0)
        goto error;

355
    if (stats->downtime_set &&
J
Jiri Denemark 已提交
356 357
        virTypedParamsAddULLong(&par, &npar, &maxpar,
                                VIR_DOMAIN_JOB_DOWNTIME,
358
                                stats->downtime) < 0)
J
Jiri Denemark 已提交
359 360
        goto error;

361
    if (stats->downtime_set &&
362
        jobInfo->timeDeltaSet &&
363
        stats->downtime > jobInfo->timeDelta &&
364 365
        virTypedParamsAddULLong(&par, &npar, &maxpar,
                                VIR_DOMAIN_JOB_DOWNTIME_NET,
366
                                stats->downtime - jobInfo->timeDelta) < 0)
367 368
        goto error;

369
    if (stats->setup_time_set &&
370 371
        virTypedParamsAddULLong(&par, &npar, &maxpar,
                                VIR_DOMAIN_JOB_SETUP_TIME,
372
                                stats->setup_time) < 0)
373 374
        goto error;

J
Jiri Denemark 已提交
375 376
    if (virTypedParamsAddULLong(&par, &npar, &maxpar,
                                VIR_DOMAIN_JOB_DATA_TOTAL,
377 378
                                stats->ram_total +
                                stats->disk_total) < 0 ||
J
Jiri Denemark 已提交
379 380
        virTypedParamsAddULLong(&par, &npar, &maxpar,
                                VIR_DOMAIN_JOB_DATA_PROCESSED,
381 382
                                stats->ram_transferred +
                                stats->disk_transferred) < 0 ||
J
Jiri Denemark 已提交
383 384
        virTypedParamsAddULLong(&par, &npar, &maxpar,
                                VIR_DOMAIN_JOB_DATA_REMAINING,
385 386
                                stats->ram_remaining +
                                stats->disk_remaining) < 0)
J
Jiri Denemark 已提交
387 388 389 390
        goto error;

    if (virTypedParamsAddULLong(&par, &npar, &maxpar,
                                VIR_DOMAIN_JOB_MEMORY_TOTAL,
391
                                stats->ram_total) < 0 ||
J
Jiri Denemark 已提交
392 393
        virTypedParamsAddULLong(&par, &npar, &maxpar,
                                VIR_DOMAIN_JOB_MEMORY_PROCESSED,
394
                                stats->ram_transferred) < 0 ||
J
Jiri Denemark 已提交
395 396
        virTypedParamsAddULLong(&par, &npar, &maxpar,
                                VIR_DOMAIN_JOB_MEMORY_REMAINING,
397
                                stats->ram_remaining) < 0)
J
Jiri Denemark 已提交
398 399
        goto error;

400
    if (stats->ram_bps &&
401 402
        virTypedParamsAddULLong(&par, &npar, &maxpar,
                                VIR_DOMAIN_JOB_MEMORY_BPS,
403
                                stats->ram_bps) < 0)
404 405
        goto error;

406
    if (stats->ram_duplicate_set) {
J
Jiri Denemark 已提交
407 408
        if (virTypedParamsAddULLong(&par, &npar, &maxpar,
                                    VIR_DOMAIN_JOB_MEMORY_CONSTANT,
409
                                    stats->ram_duplicate) < 0 ||
J
Jiri Denemark 已提交
410 411
            virTypedParamsAddULLong(&par, &npar, &maxpar,
                                    VIR_DOMAIN_JOB_MEMORY_NORMAL,
412
                                    stats->ram_normal) < 0 ||
J
Jiri Denemark 已提交
413 414
            virTypedParamsAddULLong(&par, &npar, &maxpar,
                                    VIR_DOMAIN_JOB_MEMORY_NORMAL_BYTES,
415
                                    stats->ram_normal_bytes) < 0)
J
Jiri Denemark 已提交
416 417 418
            goto error;
    }

419 420 421 422 423 424 425 426
    if (virTypedParamsAddULLong(&par, &npar, &maxpar,
                                VIR_DOMAIN_JOB_MEMORY_DIRTY_RATE,
                                stats->ram_dirty_rate) < 0 ||
        virTypedParamsAddULLong(&par, &npar, &maxpar,
                                VIR_DOMAIN_JOB_MEMORY_ITERATION,
                                stats->ram_iteration) < 0)
        goto error;

J
Jiri Denemark 已提交
427 428
    if (virTypedParamsAddULLong(&par, &npar, &maxpar,
                                VIR_DOMAIN_JOB_DISK_TOTAL,
429
                                stats->disk_total) < 0 ||
J
Jiri Denemark 已提交
430 431
        virTypedParamsAddULLong(&par, &npar, &maxpar,
                                VIR_DOMAIN_JOB_DISK_PROCESSED,
432
                                stats->disk_transferred) < 0 ||
J
Jiri Denemark 已提交
433 434
        virTypedParamsAddULLong(&par, &npar, &maxpar,
                                VIR_DOMAIN_JOB_DISK_REMAINING,
435
                                stats->disk_remaining) < 0)
J
Jiri Denemark 已提交
436 437
        goto error;

438
    if (stats->disk_bps &&
439 440
        virTypedParamsAddULLong(&par, &npar, &maxpar,
                                VIR_DOMAIN_JOB_DISK_BPS,
441
                                stats->disk_bps) < 0)
442 443
        goto error;

444
    if (stats->xbzrle_set) {
J
Jiri Denemark 已提交
445 446
        if (virTypedParamsAddULLong(&par, &npar, &maxpar,
                                    VIR_DOMAIN_JOB_COMPRESSION_CACHE,
447
                                    stats->xbzrle_cache_size) < 0 ||
J
Jiri Denemark 已提交
448 449
            virTypedParamsAddULLong(&par, &npar, &maxpar,
                                    VIR_DOMAIN_JOB_COMPRESSION_BYTES,
450
                                    stats->xbzrle_bytes) < 0 ||
J
Jiri Denemark 已提交
451 452
            virTypedParamsAddULLong(&par, &npar, &maxpar,
                                    VIR_DOMAIN_JOB_COMPRESSION_PAGES,
453
                                    stats->xbzrle_pages) < 0 ||
J
Jiri Denemark 已提交
454 455
            virTypedParamsAddULLong(&par, &npar, &maxpar,
                                    VIR_DOMAIN_JOB_COMPRESSION_CACHE_MISSES,
456
                                    stats->xbzrle_cache_miss) < 0 ||
J
Jiri Denemark 已提交
457 458
            virTypedParamsAddULLong(&par, &npar, &maxpar,
                                    VIR_DOMAIN_JOB_COMPRESSION_OVERFLOW,
459
                                    stats->xbzrle_overflow) < 0)
J
Jiri Denemark 已提交
460 461 462
            goto error;
    }

463 464 465 466 467 468
    if (stats->cpu_throttle_percentage &&
        virTypedParamsAddInt(&par, &npar, &maxpar,
                             VIR_DOMAIN_JOB_AUTO_CONVERGE_THROTTLE,
                             stats->cpu_throttle_percentage) < 0)
        goto error;

J
Jiri Denemark 已提交
469 470 471 472 473 474 475 476 477 478 479
    *type = jobInfo->type;
    *params = par;
    *nparams = npar;
    return 0;

 error:
    virTypedParamsFree(par, npar);
    return -1;
}


J
John Ferlan 已提交
480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501
/* qemuDomainGetMasterKeyFilePath:
 * @libDir: Directory path to domain lib files
 *
 * Generate a path to the domain master key file for libDir.
 * It's up to the caller to handle checking if path exists.
 *
 * Returns path to memory containing the name of the file. It is up to the
 * caller to free; otherwise, NULL on failure.
 */
char *
qemuDomainGetMasterKeyFilePath(const char *libDir)
{
    if (!libDir) {
        virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                       _("invalid path for master key file"));
        return NULL;
    }
    return virFileBuildPath(libDir, "master-key.aes", NULL);
}


/* qemuDomainWriteMasterKeyFile:
502 503
 * @driver: qemu driver data
 * @vm: Pointer to the vm object
J
John Ferlan 已提交
504 505 506 507 508
 *
 * Get the desired path to the masterKey file and store it in the path.
 *
 * Returns 0 on success, -1 on failure with error message indicating failure
 */
509
int
M
Martin Kletzander 已提交
510 511
qemuDomainWriteMasterKeyFile(virQEMUDriverPtr driver,
                             virDomainObjPtr vm)
J
John Ferlan 已提交
512 513 514 515
{
    char *path;
    int fd = -1;
    int ret = -1;
M
Martin Kletzander 已提交
516
    qemuDomainObjPrivatePtr priv = vm->privateData;
J
John Ferlan 已提交
517

518 519 520 521
    /* Only gets filled in if we have the capability */
    if (!priv->masterKey)
        return 0;

J
John Ferlan 已提交
522 523 524 525 526 527 528 529 530 531 532 533 534 535 536
    if (!(path = qemuDomainGetMasterKeyFilePath(priv->libDir)))
        return -1;

    if ((fd = open(path, O_WRONLY|O_TRUNC|O_CREAT, 0600)) < 0) {
        virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                       _("failed to open domain master key file for write"));
        goto cleanup;
    }

    if (safewrite(fd, priv->masterKey, priv->masterKeyLen) < 0) {
        virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                       _("failed to write master key file for domain"));
        goto cleanup;
    }

537 538
    if (virSecurityManagerDomainSetPathLabel(driver->securityManager,
                                             vm->def, path) < 0)
M
Martin Kletzander 已提交
539 540
        goto cleanup;

J
John Ferlan 已提交
541 542 543 544 545 546 547 548 549 550
    ret = 0;

 cleanup:
    VIR_FORCE_CLOSE(fd);
    VIR_FREE(path);

    return ret;
}


551 552 553 554 555 556 557 558 559
static void
qemuDomainMasterKeyFree(qemuDomainObjPrivatePtr priv)
{
    if (!priv->masterKey)
        return;

    VIR_DISPOSE_N(priv->masterKey, priv->masterKeyLen);
}

J
John Ferlan 已提交
560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657
/* qemuDomainMasterKeyReadFile:
 * @priv: pointer to domain private object
 *
 * Expected to be called during qemuProcessReconnect once the domain
 * libDir has been generated through qemuStateInitialize calling
 * virDomainObjListLoadAllConfigs which will restore the libDir path
 * to the domain private object.
 *
 * This function will get the path to the master key file and if it
 * exists, it will read the contents of the file saving it in priv->masterKey.
 *
 * Once the file exists, the validity checks may cause failures; however,
 * if the file doesn't exist or the capability doesn't exist, we just
 * return (mostly) quietly.
 *
 * Returns 0 on success or lack of capability
 *        -1 on failure with error message indicating failure
 */
int
qemuDomainMasterKeyReadFile(qemuDomainObjPrivatePtr priv)
{
    char *path;
    int fd = -1;
    uint8_t *masterKey = NULL;
    ssize_t masterKeyLen = 0;

    /* If we don't have the capability, then do nothing. */
    if (!virQEMUCapsGet(priv->qemuCaps, QEMU_CAPS_OBJECT_SECRET))
        return 0;

    if (!(path = qemuDomainGetMasterKeyFilePath(priv->libDir)))
        return -1;

    if (!virFileExists(path)) {
        virReportError(VIR_ERR_INTERNAL_ERROR,
                       _("domain master key file doesn't exist in %s"),
                       priv->libDir);
        goto error;
    }

    if ((fd = open(path, O_RDONLY)) < 0) {
        virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                       _("failed to open domain master key file for read"));
        goto error;
    }

    if (VIR_ALLOC_N(masterKey, 1024) < 0)
        goto error;

    if ((masterKeyLen = saferead(fd, masterKey, 1024)) < 0) {
        virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                       _("unable to read domain master key file"));
        goto error;
    }

    if (masterKeyLen != QEMU_DOMAIN_MASTER_KEY_LEN) {
        virReportError(VIR_ERR_INTERNAL_ERROR,
                       _("invalid master key read, size=%zd"), masterKeyLen);
        goto error;
    }

    ignore_value(VIR_REALLOC_N_QUIET(masterKey, masterKeyLen));

    priv->masterKey = masterKey;
    priv->masterKeyLen = masterKeyLen;

    VIR_FORCE_CLOSE(fd);
    VIR_FREE(path);

    return 0;

 error:
    if (masterKeyLen > 0)
        memset(masterKey, 0, masterKeyLen);
    VIR_FREE(masterKey);

    VIR_FORCE_CLOSE(fd);
    VIR_FREE(path);

    return -1;
}


/* qemuDomainMasterKeyRemove:
 * @priv: Pointer to the domain private object
 *
 * Remove the traces of the master key, clear the heap, clear the file,
 * delete the file.
 */
void
qemuDomainMasterKeyRemove(qemuDomainObjPrivatePtr priv)
{
    char *path = NULL;

    if (!priv->masterKey)
        return;

    /* Clear the contents */
658
    qemuDomainMasterKeyFree(priv);
J
John Ferlan 已提交
659 660 661 662 663 664 665 666 667 668

    /* Delete the master key file */
    path = qemuDomainGetMasterKeyFilePath(priv->libDir);
    unlink(path);

    VIR_FREE(path);
}


/* qemuDomainMasterKeyCreate:
669
 * @vm: Pointer to the domain object
J
John Ferlan 已提交
670 671 672 673 674 675 676 677
 *
 * As long as the underlying qemu has the secret capability,
 * generate and store 'raw' in a file a random 32-byte key to
 * be used as a secret shared with qemu to share sensitive data.
 *
 * Returns: 0 on success, -1 w/ error message on failure
 */
int
678
qemuDomainMasterKeyCreate(virDomainObjPtr vm)
J
John Ferlan 已提交
679
{
M
Martin Kletzander 已提交
680 681
    qemuDomainObjPrivatePtr priv = vm->privateData;

J
John Ferlan 已提交
682 683 684 685 686
    /* If we don't have the capability, then do nothing. */
    if (!virQEMUCapsGet(priv->qemuCaps, QEMU_CAPS_OBJECT_SECRET))
        return 0;

    if (!(priv->masterKey =
687
          virCryptoGenerateRandom(QEMU_DOMAIN_MASTER_KEY_LEN)))
688
        return -1;
J
John Ferlan 已提交
689 690 691 692 693 694 695

    priv->masterKeyLen = QEMU_DOMAIN_MASTER_KEY_LEN;

    return 0;
}


696
static void
697
qemuDomainSecretPlainClear(qemuDomainSecretPlain secret)
698 699
{
    VIR_FREE(secret.username);
700
    VIR_DISPOSE_N(secret.secret, secret.secretlen);
701 702 703
}


J
John Ferlan 已提交
704
static void
705
qemuDomainSecretAESClear(qemuDomainSecretAES secret)
J
John Ferlan 已提交
706 707 708 709 710 711 712 713
{
    VIR_FREE(secret.username);
    VIR_FREE(secret.alias);
    VIR_FREE(secret.iv);
    VIR_FREE(secret.ciphertext);
}


714 715 716 717 718 719
static void
qemuDomainSecretInfoFree(qemuDomainSecretInfoPtr *secinfo)
{
    if (!*secinfo)
        return;

J
John Ferlan 已提交
720 721
    switch ((qemuDomainSecretInfoType) (*secinfo)->type) {
    case VIR_DOMAIN_SECRET_INFO_TYPE_PLAIN:
722
        qemuDomainSecretPlainClear((*secinfo)->s.plain);
J
John Ferlan 已提交
723 724
        break;

725
    case VIR_DOMAIN_SECRET_INFO_TYPE_AES:
726
        qemuDomainSecretAESClear((*secinfo)->s.aes);
J
John Ferlan 已提交
727 728 729 730 731 732
        break;

    case VIR_DOMAIN_SECRET_INFO_TYPE_LAST:
        break;
    }

733 734 735 736
    VIR_FREE(*secinfo);
}


737
static virClassPtr qemuDomainDiskPrivateClass;
738
static void qemuDomainDiskPrivateDispose(void *obj);
739 740 741 742 743 744 745

static int
qemuDomainDiskPrivateOnceInit(void)
{
    qemuDomainDiskPrivateClass = virClassNew(virClassForObject(),
                                             "qemuDomainDiskPrivate",
                                             sizeof(qemuDomainDiskPrivate),
746
                                             qemuDomainDiskPrivateDispose);
747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769
    if (!qemuDomainDiskPrivateClass)
        return -1;
    else
        return 0;
}

VIR_ONCE_GLOBAL_INIT(qemuDomainDiskPrivate)

static virObjectPtr
qemuDomainDiskPrivateNew(void)
{
    qemuDomainDiskPrivatePtr priv;

    if (qemuDomainDiskPrivateInitialize() < 0)
        return NULL;

    if (!(priv = virObjectNew(qemuDomainDiskPrivateClass)))
        return NULL;

    return (virObjectPtr) priv;
}


770 771 772 773 774 775
static void
qemuDomainDiskPrivateDispose(void *obj)
{
    qemuDomainDiskPrivatePtr priv = obj;

    qemuDomainSecretInfoFree(&priv->secinfo);
J
John Ferlan 已提交
776
    qemuDomainSecretInfoFree(&priv->encinfo);
777 778 779
}


780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822
static virClassPtr qemuDomainHostdevPrivateClass;
static void qemuDomainHostdevPrivateDispose(void *obj);

static int
qemuDomainHostdevPrivateOnceInit(void)
{
    qemuDomainHostdevPrivateClass =
        virClassNew(virClassForObject(),
                    "qemuDomainHostdevPrivate",
                    sizeof(qemuDomainHostdevPrivate),
                    qemuDomainHostdevPrivateDispose);
    if (!qemuDomainHostdevPrivateClass)
        return -1;
    else
        return 0;
}

VIR_ONCE_GLOBAL_INIT(qemuDomainHostdevPrivate)

static virObjectPtr
qemuDomainHostdevPrivateNew(void)
{
    qemuDomainHostdevPrivatePtr priv;

    if (qemuDomainHostdevPrivateInitialize() < 0)
        return NULL;

    if (!(priv = virObjectNew(qemuDomainHostdevPrivateClass)))
        return NULL;

    return (virObjectPtr) priv;
}


static void
qemuDomainHostdevPrivateDispose(void *obj)
{
    qemuDomainHostdevPrivatePtr priv = obj;

    qemuDomainSecretInfoFree(&priv->secinfo);
}


823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856
static virClassPtr qemuDomainVcpuPrivateClass;
static void qemuDomainVcpuPrivateDispose(void *obj);

static int
qemuDomainVcpuPrivateOnceInit(void)
{
    qemuDomainVcpuPrivateClass = virClassNew(virClassForObject(),
                                             "qemuDomainVcpuPrivate",
                                             sizeof(qemuDomainVcpuPrivate),
                                             qemuDomainVcpuPrivateDispose);
    if (!qemuDomainVcpuPrivateClass)
        return -1;
    else
        return 0;
}

VIR_ONCE_GLOBAL_INIT(qemuDomainVcpuPrivate)

static virObjectPtr
qemuDomainVcpuPrivateNew(void)
{
    qemuDomainVcpuPrivatePtr priv;

    if (qemuDomainVcpuPrivateInitialize() < 0)
        return NULL;

    if (!(priv = virObjectNew(qemuDomainVcpuPrivateClass)))
        return NULL;

    return (virObjectPtr) priv;
}


static void
857
qemuDomainVcpuPrivateDispose(void *obj)
858
{
859 860 861 862
    qemuDomainVcpuPrivatePtr priv = obj;

    VIR_FREE(priv->type);
    VIR_FREE(priv->alias);
863 864 865 866
    return;
}


867 868
static virClassPtr qemuDomainChrSourcePrivateClass;
static void qemuDomainChrSourcePrivateDispose(void *obj);
869 870

static int
871
qemuDomainChrSourcePrivateOnceInit(void)
872
{
873
    qemuDomainChrSourcePrivateClass =
874
        virClassNew(virClassForObject(),
875 876 877 878
                    "qemuDomainChrSourcePrivate",
                    sizeof(qemuDomainChrSourcePrivate),
                    qemuDomainChrSourcePrivateDispose);
    if (!qemuDomainChrSourcePrivateClass)
879 880 881 882 883
        return -1;
    else
        return 0;
}

884
VIR_ONCE_GLOBAL_INIT(qemuDomainChrSourcePrivate)
885 886

static virObjectPtr
887
qemuDomainChrSourcePrivateNew(void)
888
{
889
    qemuDomainChrSourcePrivatePtr priv;
890

891
    if (qemuDomainChrSourcePrivateInitialize() < 0)
892 893
        return NULL;

894
    if (!(priv = virObjectNew(qemuDomainChrSourcePrivateClass)))
895 896 897 898 899 900 901
        return NULL;

    return (virObjectPtr) priv;
}


static void
902
qemuDomainChrSourcePrivateDispose(void *obj)
903
{
904
    qemuDomainChrSourcePrivatePtr priv = obj;
905 906 907 908 909

    qemuDomainSecretInfoFree(&priv->secinfo);
}


910 911 912
/* qemuDomainSecretPlainSetup:
 * @conn: Pointer to connection
 * @secinfo: Pointer to secret info
913
 * @secretUsageType: The virSecretUsageType
914 915
 * @username: username to use for authentication (may be NULL)
 * @seclookupdef: Pointer to seclookupdef data
916 917 918 919 920 921 922 923
 *
 * Taking a secinfo, fill in the plaintext information
 *
 * Returns 0 on success, -1 on failure with error message
 */
static int
qemuDomainSecretPlainSetup(virConnectPtr conn,
                           qemuDomainSecretInfoPtr secinfo,
924
                           virSecretUsageType secretUsageType,
925 926
                           const char *username,
                           virSecretLookupTypeDefPtr seclookupdef)
927
{
928
    secinfo->type = VIR_DOMAIN_SECRET_INFO_TYPE_PLAIN;
929
    if (VIR_STRDUP(secinfo->s.plain.username, username) < 0)
930 931
        return -1;

932
    return virSecretGetSecretString(conn, seclookupdef, secretUsageType,
933 934
                                    &secinfo->s.plain.secret,
                                    &secinfo->s.plain.secretlen);
935 936 937
}


938 939 940 941 942
/* qemuDomainSecretAESSetup:
 * @conn: Pointer to connection
 * @priv: pointer to domain private object
 * @secinfo: Pointer to secret info
 * @srcalias: Alias of the disk/hostdev used to generate the secret alias
943
 * @secretUsageType: The virSecretUsageType
944 945
 * @username: username to use for authentication (may be NULL)
 * @seclookupdef: Pointer to seclookupdef data
946
 * @isLuks: True/False for is for luks (alias generation)
947 948 949 950 951 952 953 954 955 956
 *
 * Taking a secinfo, fill in the AES specific information using the
 *
 * Returns 0 on success, -1 on failure with error message
 */
static int
qemuDomainSecretAESSetup(virConnectPtr conn,
                         qemuDomainObjPrivatePtr priv,
                         qemuDomainSecretInfoPtr secinfo,
                         const char *srcalias,
957
                         virSecretUsageType secretUsageType,
958
                         const char *username,
959 960
                         virSecretLookupTypeDefPtr seclookupdef,
                         bool isLuks)
961 962 963 964 965 966 967 968 969 970
{
    int ret = -1;
    uint8_t *raw_iv = NULL;
    size_t ivlen = QEMU_DOMAIN_AES_IV_LEN;
    uint8_t *secret = NULL;
    size_t secretlen = 0;
    uint8_t *ciphertext = NULL;
    size_t ciphertextlen = 0;

    secinfo->type = VIR_DOMAIN_SECRET_INFO_TYPE_AES;
971
    if (VIR_STRDUP(secinfo->s.aes.username, username) < 0)
972 973
        return -1;

974
    if (!(secinfo->s.aes.alias = qemuDomainGetSecretAESAlias(srcalias, isLuks)))
975 976 977 978 979 980 981 982 983 984 985
        return -1;

    /* Create a random initialization vector */
    if (!(raw_iv = virCryptoGenerateRandom(ivlen)))
        return -1;

    /* Encode the IV and save that since qemu will need it */
    if (!(secinfo->s.aes.iv = virStringEncodeBase64(raw_iv, ivlen)))
        goto cleanup;

    /* Grab the unencoded secret */
986
    if (virSecretGetSecretString(conn, seclookupdef, secretUsageType,
987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014
                                 &secret, &secretlen) < 0)
        goto cleanup;

    if (virCryptoEncryptData(VIR_CRYPTO_CIPHER_AES256CBC,
                             priv->masterKey, QEMU_DOMAIN_MASTER_KEY_LEN,
                             raw_iv, ivlen, secret, secretlen,
                             &ciphertext, &ciphertextlen) < 0)
        goto cleanup;

    /* Clear out the secret */
    memset(secret, 0, secretlen);

    /* Now encode the ciphertext and store to be passed to qemu */
    if (!(secinfo->s.aes.ciphertext = virStringEncodeBase64(ciphertext,
                                                            ciphertextlen)))
        goto cleanup;

    ret = 0;

 cleanup:
    VIR_DISPOSE_N(raw_iv, ivlen);
    VIR_DISPOSE_N(secret, secretlen);
    VIR_DISPOSE_N(ciphertext, ciphertextlen);

    return ret;
}


1015 1016
/* qemuDomainSecretSetup:
 * @conn: Pointer to connection
1017
 * @priv: pointer to domain private object
1018
 * @secinfo: Pointer to secret info
1019
 * @srcalias: Alias of the disk/hostdev used to generate the secret alias
1020
 * @secretUsageType: The virSecretUsageType
1021 1022
 * @username: username to use for authentication (may be NULL)
 * @seclookupdef: Pointer to seclookupdef data
1023
 * @isLuks: True when is luks (generates different alias)
1024
 *
1025 1026 1027 1028
 * If we have the encryption API present and can support a secret object, then
 * build the AES secret; otherwise, build the Plain secret. This is the magic
 * decision point for utilizing the AES secrets for an RBD disk. For now iSCSI
 * disks and hostdevs will not be able to utilize this mechanism.
1029 1030 1031 1032 1033
 *
 * Returns 0 on success, -1 on failure
 */
static int
qemuDomainSecretSetup(virConnectPtr conn,
1034
                      qemuDomainObjPrivatePtr priv,
1035
                      qemuDomainSecretInfoPtr secinfo,
1036
                      const char *srcalias,
1037
                      virSecretUsageType secretUsageType,
1038
                      const char *username,
1039 1040
                      virSecretLookupTypeDefPtr seclookupdef,
                      bool isLuks)
1041
{
1042 1043
    if (virCryptoHaveCipher(VIR_CRYPTO_CIPHER_AES256CBC) &&
        virQEMUCapsGet(priv->qemuCaps, QEMU_CAPS_OBJECT_SECRET) &&
1044
        (secretUsageType == VIR_SECRET_USAGE_TYPE_CEPH ||
1045 1046
         secretUsageType == VIR_SECRET_USAGE_TYPE_VOLUME ||
         secretUsageType == VIR_SECRET_USAGE_TYPE_TLS)) {
1047
        if (qemuDomainSecretAESSetup(conn, priv, secinfo, srcalias,
1048
                                     secretUsageType, username,
1049
                                     seclookupdef, isLuks) < 0)
1050 1051
            return -1;
    } else {
1052
        if (qemuDomainSecretPlainSetup(conn, secinfo, secretUsageType,
1053
                                       username, seclookupdef) < 0)
1054 1055 1056
            return -1;
    }
    return 0;
1057 1058 1059
}


1060 1061 1062 1063 1064 1065 1066 1067 1068 1069
/* qemuDomainSecretDiskDestroy:
 * @disk: Pointer to a disk definition
 *
 * Clear and destroy memory associated with the secret
 */
void
qemuDomainSecretDiskDestroy(virDomainDiskDefPtr disk)
{
    qemuDomainDiskPrivatePtr diskPriv = QEMU_DOMAIN_DISK_PRIVATE(disk);

1070
    if (!diskPriv || !diskPriv->secinfo)
1071 1072 1073 1074 1075 1076
        return;

    qemuDomainSecretInfoFree(&diskPriv->secinfo);
}


1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090
bool
qemuDomainSecretDiskCapable(virStorageSourcePtr src)
{
    if (!virStorageSourceIsEmpty(src) &&
        virStorageSourceGetActualType(src) == VIR_STORAGE_TYPE_NETWORK &&
        src->auth &&
        (src->protocol == VIR_STORAGE_NET_PROTOCOL_ISCSI ||
         src->protocol == VIR_STORAGE_NET_PROTOCOL_RBD))
        return true;

    return false;
}


1091 1092
/* qemuDomainSecretDiskPrepare:
 * @conn: Pointer to connection
J
John Ferlan 已提交
1093
 * @priv: pointer to domain private object
1094 1095 1096 1097 1098 1099 1100 1101
 * @disk: Pointer to a disk definition
 *
 * For the right disk, generate the qemuDomainSecretInfo structure.
 *
 * Returns 0 on success, -1 on failure
 */
int
qemuDomainSecretDiskPrepare(virConnectPtr conn,
1102
                            qemuDomainObjPrivatePtr priv,
1103 1104 1105
                            virDomainDiskDefPtr disk)
{
    virStorageSourcePtr src = disk->src;
1106
    qemuDomainDiskPrivatePtr diskPriv = QEMU_DOMAIN_DISK_PRIVATE(disk);
1107 1108
    qemuDomainSecretInfoPtr secinfo = NULL;

1109
    if (qemuDomainSecretDiskCapable(src)) {
1110
        virSecretUsageType secretUsageType = VIR_SECRET_USAGE_TYPE_ISCSI;
1111 1112 1113 1114

        if (VIR_ALLOC(secinfo) < 0)
            return -1;

1115 1116 1117
        if (src->protocol == VIR_STORAGE_NET_PROTOCOL_RBD)
            secretUsageType = VIR_SECRET_USAGE_TYPE_CEPH;

1118
        if (qemuDomainSecretSetup(conn, priv, secinfo, disk->info.alias,
1119
                                  secretUsageType, src->auth->username,
1120
                                  &src->auth->seclookupdef, false) < 0)
1121 1122 1123 1124 1125
            goto error;

        diskPriv->secinfo = secinfo;
    }

1126
    if (!virStorageSourceIsEmpty(src) && src->encryption &&
1127
        src->encryption->format == VIR_STORAGE_ENCRYPTION_FORMAT_LUKS) {
1128 1129 1130 1131 1132 1133 1134 1135 1136 1137

        if (VIR_ALLOC(secinfo) < 0)
            return -1;

        if (qemuDomainSecretSetup(conn, priv, secinfo, disk->info.alias,
                                  VIR_SECRET_USAGE_TYPE_VOLUME, NULL,
                                  &src->encryption->secrets[0]->seclookupdef,
                                  true) < 0)
            goto error;

1138 1139 1140 1141 1142 1143 1144
        if (secinfo->type == VIR_DOMAIN_SECRET_INFO_TYPE_PLAIN) {
            virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                           _("luks encryption requires encrypted secrets "
                             "to be supported"));
            goto error;
        }

1145 1146 1147
        diskPriv->encinfo = secinfo;
    }

1148 1149 1150 1151 1152 1153 1154 1155
    return 0;

 error:
    qemuDomainSecretInfoFree(&secinfo);
    return -1;
}


1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166
/* qemuDomainSecretHostdevDestroy:
 * @disk: Pointer to a hostdev definition
 *
 * Clear and destroy memory associated with the secret
 */
void
qemuDomainSecretHostdevDestroy(virDomainHostdevDefPtr hostdev)
{
    qemuDomainHostdevPrivatePtr hostdevPriv =
        QEMU_DOMAIN_HOSTDEV_PRIVATE(hostdev);

1167
    if (!hostdevPriv || !hostdevPriv->secinfo)
1168 1169 1170 1171 1172 1173 1174 1175
        return;

    qemuDomainSecretInfoFree(&hostdevPriv->secinfo);
}


/* qemuDomainSecretHostdevPrepare:
 * @conn: Pointer to connection
J
John Ferlan 已提交
1176
 * @priv: pointer to domain private object
1177 1178 1179 1180 1181 1182 1183 1184
 * @hostdev: Pointer to a hostdev definition
 *
 * For the right host device, generate the qemuDomainSecretInfo structure.
 *
 * Returns 0 on success, -1 on failure
 */
int
qemuDomainSecretHostdevPrepare(virConnectPtr conn,
1185
                               qemuDomainObjPrivatePtr priv,
1186 1187 1188 1189 1190
                               virDomainHostdevDefPtr hostdev)
{
    virDomainHostdevSubsysPtr subsys = &hostdev->source.subsys;
    qemuDomainSecretInfoPtr secinfo = NULL;

1191
    if (hostdev->mode == VIR_DOMAIN_HOSTDEV_MODE_SUBSYS &&
1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205
        subsys->type == VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_SCSI) {

        virDomainHostdevSubsysSCSIPtr scsisrc = &hostdev->source.subsys.u.scsi;
        virDomainHostdevSubsysSCSIiSCSIPtr iscsisrc = &scsisrc->u.iscsi;

        if (scsisrc->protocol == VIR_DOMAIN_HOSTDEV_SCSI_PROTOCOL_TYPE_ISCSI &&
            iscsisrc->auth) {

            qemuDomainHostdevPrivatePtr hostdevPriv =
                QEMU_DOMAIN_HOSTDEV_PRIVATE(hostdev);

            if (VIR_ALLOC(secinfo) < 0)
                return -1;

1206
            if (qemuDomainSecretSetup(conn, priv, secinfo, hostdev->info->alias,
1207
                                      VIR_SECRET_USAGE_TYPE_ISCSI,
1208
                                      iscsisrc->auth->username,
1209
                                      &iscsisrc->auth->seclookupdef, false) < 0)
1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223
                goto error;

            hostdevPriv->secinfo = secinfo;
        }
    }

    return 0;

 error:
    qemuDomainSecretInfoFree(&secinfo);
    return -1;
}


1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310
/* qemuDomainSecretChardevDestroy:
 * @disk: Pointer to a chardev definition
 *
 * Clear and destroy memory associated with the secret
 */
void
qemuDomainSecretChardevDestroy(virDomainChrSourceDefPtr dev)
{
    qemuDomainChrSourcePrivatePtr chrSourcePriv =
        QEMU_DOMAIN_CHR_SOURCE_PRIVATE(dev);

    if (!chrSourcePriv || !chrSourcePriv->secinfo)
        return;

    qemuDomainSecretInfoFree(&chrSourcePriv->secinfo);
}


/* qemuDomainSecretChardevPrepare:
 * @conn: Pointer to connection
 * @cfg: Pointer to driver config object
 * @priv: pointer to domain private object
 * @chrAlias: Alias of the chr device
 * @dev: Pointer to a char source definition
 *
 * For a TCP character device, generate a qemuDomainSecretInfo to be used
 * by the command line code to generate the secret for the tls-creds to use.
 *
 * Returns 0 on success, -1 on failure
 */
int
qemuDomainSecretChardevPrepare(virConnectPtr conn,
                               virQEMUDriverConfigPtr cfg,
                               qemuDomainObjPrivatePtr priv,
                               const char *chrAlias,
                               virDomainChrSourceDefPtr dev)
{
    virSecretLookupTypeDef seclookupdef = {0};
    qemuDomainSecretInfoPtr secinfo = NULL;
    char *charAlias = NULL;

    if (dev->type != VIR_DOMAIN_CHR_TYPE_TCP)
        return 0;

    if (dev->data.tcp.haveTLS == VIR_TRISTATE_BOOL_YES &&
        cfg->chardevTLSx509secretUUID) {
        qemuDomainChrSourcePrivatePtr chrSourcePriv =
            QEMU_DOMAIN_CHR_SOURCE_PRIVATE(dev);

        if (virUUIDParse(cfg->chardevTLSx509secretUUID,
                         seclookupdef.u.uuid) < 0) {
            virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                           _("malformed chardev TLS secret uuid in qemu.conf"));
            goto error;
        }
        seclookupdef.type = VIR_SECRET_LOOKUP_TYPE_UUID;

        if (VIR_ALLOC(secinfo) < 0)
            goto error;

        if (!(charAlias = qemuAliasChardevFromDevAlias(chrAlias)))
            goto error;

        if (qemuDomainSecretSetup(conn, priv, secinfo, charAlias,
                                  VIR_SECRET_USAGE_TYPE_TLS, NULL,
                                  &seclookupdef, false) < 0)
            goto error;

        if (secinfo->type == VIR_DOMAIN_SECRET_INFO_TYPE_PLAIN) {
            virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                           _("TLS X.509 requires encrypted secrets "
                             "to be supported"));
            goto error;
        }

        chrSourcePriv->secinfo = secinfo;
    }

    VIR_FREE(charAlias);
    return 0;

 error:
    qemuDomainSecretInfoFree(&secinfo);
    return -1;
}


1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323
/* qemuDomainSecretDestroy:
 * @vm: Domain object
 *
 * Once completed with the generation of the command line it is
 * expect to remove the secrets
 */
void
qemuDomainSecretDestroy(virDomainObjPtr vm)
{
    size_t i;

    for (i = 0; i < vm->def->ndisks; i++)
        qemuDomainSecretDiskDestroy(vm->def->disks[i]);
1324 1325 1326

    for (i = 0; i < vm->def->nhostdevs; i++)
        qemuDomainSecretHostdevDestroy(vm->def->hostdevs[i]);
1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352

    for (i = 0; i < vm->def->nserials; i++)
        qemuDomainSecretChardevDestroy(vm->def->serials[i]->source);

    for (i = 0; i < vm->def->nparallels; i++)
        qemuDomainSecretChardevDestroy(vm->def->parallels[i]->source);

    for (i = 0; i < vm->def->nchannels; i++)
        qemuDomainSecretChardevDestroy(vm->def->channels[i]->source);

    for (i = 0; i < vm->def->nconsoles; i++)
        qemuDomainSecretChardevDestroy(vm->def->consoles[i]->source);

    for (i = 0; i < vm->def->nsmartcards; i++) {
        if (vm->def->smartcards[i]->type ==
            VIR_DOMAIN_SMARTCARD_TYPE_PASSTHROUGH)
            qemuDomainSecretChardevDestroy(vm->def->smartcards[i]->data.passthru);
    }

    for (i = 0; i < vm->def->nrngs; i++) {
        if (vm->def->rngs[i]->backend == VIR_DOMAIN_RNG_BACKEND_EGD)
            qemuDomainSecretChardevDestroy(vm->def->rngs[i]->source.chardev);
    }

    for (i = 0; i < vm->def->nredirdevs; i++)
        qemuDomainSecretChardevDestroy(vm->def->redirdevs[i]->source);
1353 1354 1355 1356 1357
}


/* qemuDomainSecretPrepare:
 * @conn: Pointer to connection
1358
 * @driver: Pointer to driver object
1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370
 * @vm: Domain object
 *
 * For any objects that may require an auth/secret setup, create a
 * qemuDomainSecretInfo and save it in the approriate place within
 * the private structures. This will be used by command line build
 * code in order to pass the secret along to qemu in order to provide
 * the necessary authentication data.
 *
 * Returns 0 on success, -1 on failure with error message set
 */
int
qemuDomainSecretPrepare(virConnectPtr conn,
1371
                        virQEMUDriverPtr driver,
1372 1373
                        virDomainObjPtr vm)
{
J
John Ferlan 已提交
1374
    qemuDomainObjPrivatePtr priv = vm->privateData;
1375
    virQEMUDriverConfigPtr cfg = virQEMUDriverGetConfig(driver);
1376
    size_t i;
1377
    int ret = -1;
1378 1379

    for (i = 0; i < vm->def->ndisks; i++) {
J
John Ferlan 已提交
1380
        if (qemuDomainSecretDiskPrepare(conn, priv, vm->def->disks[i]) < 0)
1381
            goto cleanup;
1382 1383
    }

1384
    for (i = 0; i < vm->def->nhostdevs; i++) {
J
John Ferlan 已提交
1385 1386
        if (qemuDomainSecretHostdevPrepare(conn, priv,
                                           vm->def->hostdevs[i]) < 0)
1387
            goto cleanup;
1388 1389
    }

1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445
    for (i = 0; i < vm->def->nserials; i++) {
        if (qemuDomainSecretChardevPrepare(conn, cfg, priv,
                                           vm->def->serials[i]->info.alias,
                                           vm->def->serials[i]->source) < 0)
            goto cleanup;
    }

    for (i = 0; i < vm->def->nparallels; i++) {
        if (qemuDomainSecretChardevPrepare(conn, cfg, priv,
                                           vm->def->parallels[i]->info.alias,
                                           vm->def->parallels[i]->source) < 0)
            goto cleanup;
    }

    for (i = 0; i < vm->def->nchannels; i++) {
        if (qemuDomainSecretChardevPrepare(conn, cfg, priv,
                                           vm->def->channels[i]->info.alias,
                                           vm->def->channels[i]->source) < 0)
            goto cleanup;
    }

    for (i = 0; i < vm->def->nconsoles; i++) {
        if (qemuDomainSecretChardevPrepare(conn, cfg, priv,
                                           vm->def->consoles[i]->info.alias,
                                           vm->def->consoles[i]->source) < 0)
            goto cleanup;
    }

    for (i = 0; i < vm->def->nsmartcards; i++)
        if (vm->def->smartcards[i]->type ==
            VIR_DOMAIN_SMARTCARD_TYPE_PASSTHROUGH &&
            qemuDomainSecretChardevPrepare(conn, cfg, priv,
                                           vm->def->smartcards[i]->info.alias,
                                           vm->def->smartcards[i]->data.passthru) < 0)
            goto cleanup;

    for (i = 0; i < vm->def->nrngs; i++) {
        if (vm->def->rngs[i]->backend == VIR_DOMAIN_RNG_BACKEND_EGD &&
            qemuDomainSecretChardevPrepare(conn, cfg, priv,
                                           vm->def->rngs[i]->info.alias,
                                           vm->def->rngs[i]->source.chardev) < 0)
            goto cleanup;
    }

    for (i = 0; i < vm->def->nredirdevs; i++) {
        if (qemuDomainSecretChardevPrepare(conn, cfg, priv,
                                           vm->def->redirdevs[i]->info.alias,
                                           vm->def->redirdevs[i]->source) < 0)
            goto cleanup;
    }

    ret = 0;

 cleanup:
    virObjectUnref(cfg);
    return ret;
1446 1447 1448
}


1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475
/* This is the old way of setting up per-domain directories */
static int
qemuDomainSetPrivatePathsOld(virQEMUDriverPtr driver,
                             virDomainObjPtr vm)
{
    qemuDomainObjPrivatePtr priv = vm->privateData;
    virQEMUDriverConfigPtr cfg = virQEMUDriverGetConfig(driver);
    int ret = -1;

    if (!priv->libDir &&
        virAsprintf(&priv->libDir, "%s/domain-%s",
                    cfg->libDir, vm->def->name) < 0)
        goto cleanup;

    if (!priv->channelTargetDir &&
        virAsprintf(&priv->channelTargetDir, "%s/domain-%s",
                    cfg->channelTargetDir, vm->def->name) < 0)
        goto cleanup;

    ret = 0;
 cleanup:
    virObjectUnref(cfg);
    return ret;
}


int
1476 1477
qemuDomainSetPrivatePaths(virQEMUDriverPtr driver,
                          virDomainObjPtr vm)
1478
{
1479 1480
    virQEMUDriverConfigPtr cfg = virQEMUDriverGetConfig(driver);
    qemuDomainObjPrivatePtr priv = vm->privateData;
M
Martin Kletzander 已提交
1481
    char *domname = virDomainObjGetShortName(vm);
1482
    int ret = -1;
1483

M
Martin Kletzander 已提交
1484 1485 1486
    if (!domname)
        goto cleanup;

1487
    if (!priv->libDir &&
M
Martin Kletzander 已提交
1488
        virAsprintf(&priv->libDir, "%s/domain-%s", cfg->libDir, domname) < 0)
1489
        goto cleanup;
1490

1491
    if (!priv->channelTargetDir &&
M
Martin Kletzander 已提交
1492 1493
        virAsprintf(&priv->channelTargetDir, "%s/domain-%s",
                    cfg->channelTargetDir, domname) < 0)
1494
        goto cleanup;
1495

1496 1497 1498
    ret = 0;
 cleanup:
    virObjectUnref(cfg);
M
Martin Kletzander 已提交
1499
    VIR_FREE(domname);
1500
    return ret;
1501 1502 1503
}


1504 1505 1506 1507 1508 1509 1510 1511 1512 1513
void
qemuDomainClearPrivatePaths(virDomainObjPtr vm)
{
    qemuDomainObjPrivatePtr priv = vm->privateData;

    VIR_FREE(priv->libDir);
    VIR_FREE(priv->channelTargetDir);
}


1514 1515
static void *
qemuDomainObjPrivateAlloc(void)
1516 1517 1518 1519 1520 1521
{
    qemuDomainObjPrivatePtr priv;

    if (VIR_ALLOC(priv) < 0)
        return NULL;

1522 1523 1524
    if (qemuDomainObjInitJob(priv) < 0) {
        virReportSystemError(errno, "%s",
                             _("Unable to init qemu driver mutexes"));
1525
        goto error;
1526
    }
1527

1528
    if (!(priv->devs = virChrdevAlloc()))
1529 1530
        goto error;

1531
    priv->migMaxBandwidth = QEMU_DOMAIN_MIG_BANDWIDTH_MAX;
1532

1533
    return priv;
1534

1535
 error:
1536 1537
    VIR_FREE(priv);
    return NULL;
1538 1539
}

1540 1541
static void
qemuDomainObjPrivateFree(void *data)
1542 1543 1544
{
    qemuDomainObjPrivatePtr priv = data;

1545
    virObjectUnref(priv->qemuCaps);
1546

1547
    virCgroupFree(&priv->cgroup);
1548
    virDomainPCIAddressSetFree(priv->pciaddrs);
J
Ján Tomko 已提交
1549
    virDomainUSBAddressSetFree(priv->usbaddrs);
1550
    virDomainChrSourceDefFree(priv->monConfig);
1551
    qemuDomainObjFreeJob(priv);
1552
    VIR_FREE(priv->lockState);
J
Jiri Denemark 已提交
1553
    VIR_FREE(priv->origname);
1554

1555
    virStringFreeList(priv->qemuDevices);
1556
    virChrdevFree(priv->devs);
1557

1558 1559
    /* This should never be non-NULL if we get here, but just in case... */
    if (priv->mon) {
1560
        VIR_ERROR(_("Unexpected QEMU monitor still active during domain deletion"));
1561 1562
        qemuMonitorClose(priv->mon);
    }
D
Daniel P. Berrange 已提交
1563 1564 1565 1566
    if (priv->agent) {
        VIR_ERROR(_("Unexpected QEMU agent still active during domain deletion"));
        qemuAgentClose(priv->agent);
    }
1567
    VIR_FREE(priv->cleanupCallbacks);
1568
    virBitmapFree(priv->autoNodeset);
1569
    virBitmapFree(priv->autoCpuset);
1570 1571 1572

    VIR_FREE(priv->libDir);
    VIR_FREE(priv->channelTargetDir);
1573
    qemuDomainMasterKeyFree(priv);
1574

1575 1576 1577 1578
    VIR_FREE(priv);
}


1579 1580
static void
qemuDomainObjPrivateXMLFormatVcpus(virBufferPtr buf,
1581
                                   virDomainDefPtr def)
1582 1583
{
    size_t i;
1584 1585 1586
    size_t maxvcpus = virDomainDefGetVcpusMax(def);
    virDomainVcpuDefPtr vcpu;
    pid_t tid;
1587 1588 1589 1590

    virBufferAddLit(buf, "<vcpus>\n");
    virBufferAdjustIndent(buf, 2);

1591 1592 1593 1594 1595 1596 1597 1598 1599
    for (i = 0; i < maxvcpus; i++) {
        vcpu = virDomainDefGetVcpu(def, i);
        tid = QEMU_DOMAIN_VCPU_PRIVATE(vcpu)->tid;

        if (!vcpu->online || tid == 0)
            continue;

        virBufferAsprintf(buf, "<vcpu id='%zu' pid='%d'/>\n", i, tid);
    }
1600 1601 1602 1603 1604 1605

    virBufferAdjustIndent(buf, -2);
    virBufferAddLit(buf, "</vcpus>\n");
}


1606
static int
1607 1608
qemuDomainObjPrivateXMLFormat(virBufferPtr buf,
                              virDomainObjPtr vm)
1609
{
1610
    qemuDomainObjPrivatePtr priv = vm->privateData;
1611
    const char *monitorpath;
1612
    qemuDomainJob job;
1613 1614 1615

    /* priv->monitor_chr is set only for qemu */
    if (priv->monConfig) {
1616
        switch (priv->monConfig->type) {
1617
        case VIR_DOMAIN_CHR_TYPE_UNIX:
1618
            monitorpath = priv->monConfig->data.nix.path;
1619 1620 1621
            break;
        default:
        case VIR_DOMAIN_CHR_TYPE_PTY:
1622
            monitorpath = priv->monConfig->data.file.path;
1623 1624 1625
            break;
        }

1626
        virBufferEscapeString(buf, "<monitor path='%s'", monitorpath);
1627 1628
        if (priv->monJSON)
            virBufferAddLit(buf, " json='1'");
1629
        virBufferAsprintf(buf, " type='%s'/>\n",
1630
                          virDomainChrTypeToString(priv->monConfig->type));
1631 1632
    }

1633
    qemuDomainObjPrivateXMLFormatVcpus(buf, vm->def);
1634

1635
    if (priv->qemuCaps) {
1636
        size_t i;
1637 1638
        virBufferAddLit(buf, "<qemuCaps>\n");
        virBufferAdjustIndent(buf, 2);
1639
        for (i = 0; i < QEMU_CAPS_LAST; i++) {
1640
            if (virQEMUCapsGet(priv->qemuCaps, i)) {
1641
                virBufferAsprintf(buf, "<flag name='%s'/>\n",
1642
                                  virQEMUCapsTypeToString(i));
1643 1644
            }
        }
1645 1646
        virBufferAdjustIndent(buf, -2);
        virBufferAddLit(buf, "</qemuCaps>\n");
1647 1648
    }

1649
    if (priv->lockState)
1650
        virBufferAsprintf(buf, "<lockstate>%s</lockstate>\n", priv->lockState);
1651

1652 1653 1654 1655
    job = priv->job.active;
    if (!qemuDomainTrackJob(job))
        priv->job.active = QEMU_JOB_NONE;

1656
    if (priv->job.active || priv->job.asyncJob) {
1657
        virBufferAsprintf(buf, "<job type='%s' async='%s'",
1658 1659
                          qemuDomainJobTypeToString(priv->job.active),
                          qemuDomainAsyncJobTypeToString(priv->job.asyncJob));
J
Jiri Denemark 已提交
1660 1661 1662 1663 1664
        if (priv->job.phase) {
            virBufferAsprintf(buf, " phase='%s'",
                              qemuDomainAsyncJobPhaseToString(
                                    priv->job.asyncJob, priv->job.phase));
        }
1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685
        if (priv->job.asyncJob != QEMU_ASYNC_JOB_MIGRATION_OUT) {
            virBufferAddLit(buf, "/>\n");
        } else {
            size_t i;
            virDomainDiskDefPtr disk;
            qemuDomainDiskPrivatePtr diskPriv;

            virBufferAddLit(buf, ">\n");
            virBufferAdjustIndent(buf, 2);

            for (i = 0; i < vm->def->ndisks; i++) {
                disk = vm->def->disks[i];
                diskPriv = QEMU_DOMAIN_DISK_PRIVATE(disk);
                virBufferAsprintf(buf, "<disk dev='%s' migrating='%s'/>\n",
                                  disk->dst,
                                  diskPriv->migrating ? "yes" : "no");
            }

            virBufferAdjustIndent(buf, -2);
            virBufferAddLit(buf, "</job>\n");
        }
1686
    }
1687
    priv->job.active = job;
1688

1689
    if (priv->fakeReboot)
1690
        virBufferAddLit(buf, "<fakereboot/>\n");
1691

1692 1693
    if (priv->qemuDevices && *priv->qemuDevices) {
        char **tmp = priv->qemuDevices;
1694 1695
        virBufferAddLit(buf, "<devices>\n");
        virBufferAdjustIndent(buf, 2);
1696
        while (*tmp) {
1697
            virBufferAsprintf(buf, "<device alias='%s'/>\n", *tmp);
1698 1699
            tmp++;
        }
1700 1701
        virBufferAdjustIndent(buf, -2);
        virBufferAddLit(buf, "</devices>\n");
1702 1703
    }

1704 1705 1706 1707 1708 1709 1710 1711 1712 1713
    if (priv->autoNodeset) {
        char *nodeset = virBitmapFormat(priv->autoNodeset);

        if (!nodeset)
            return -1;

        virBufferAsprintf(buf, "<numad nodeset='%s'/>\n", nodeset);
        VIR_FREE(nodeset);
    }

1714 1715 1716 1717 1718
    /* Various per-domain paths */
    virBufferEscapeString(buf, "<libDir path='%s'/>\n", priv->libDir);
    virBufferEscapeString(buf, "<channelTargetDir path='%s'/>\n",
                          priv->channelTargetDir);

1719 1720 1721
    return 0;
}

1722 1723 1724 1725

static int
qemuDomainObjPrivateXMLParseVcpu(xmlNodePtr node,
                                 unsigned int idx,
1726
                                 virDomainDefPtr def)
1727
{
1728
    virDomainVcpuDefPtr vcpu;
1729
    char *idstr;
1730
    char *pidstr;
1731
    unsigned int tmp;
1732 1733
    int ret = -1;

1734 1735
    idstr = virXMLPropString(node, "id");

1736 1737
    if (idstr &&
        (virStrToLong_uip(idstr, NULL, 10, &idx) < 0)) {
1738
        virReportError(VIR_ERR_INTERNAL_ERROR,
1739 1740 1741 1742 1743 1744
                       _("cannot parse vcpu index '%s'"), idstr);
        goto cleanup;
    }
    if (!(vcpu = virDomainDefGetVcpu(def, idx))) {
        virReportError(VIR_ERR_INTERNAL_ERROR,
                       _("invalid vcpu index '%u'"), idx);
1745
        goto cleanup;
1746 1747
    }

1748 1749 1750
    if (!(pidstr = virXMLPropString(node, "pid")))
        goto cleanup;

1751
    if (virStrToLong_uip(pidstr, NULL, 10, &tmp) < 0)
1752 1753
        goto cleanup;

1754 1755
    QEMU_DOMAIN_VCPU_PRIVATE(vcpu)->tid = tmp;

1756 1757 1758
    ret = 0;

 cleanup:
1759
    VIR_FREE(idstr);
1760 1761 1762 1763 1764
    VIR_FREE(pidstr);
    return ret;
}


1765
static int
1766
qemuDomainObjPrivateXMLParse(xmlXPathContextPtr ctxt,
1767
                             virDomainObjPtr vm,
1768
                             virDomainDefParserConfigPtr config)
1769
{
1770
    qemuDomainObjPrivatePtr priv = vm->privateData;
1771
    virQEMUDriverPtr driver = config->priv;
1772
    char *monitorpath;
1773
    char *tmp = NULL;
1774 1775
    int n;
    size_t i;
1776
    xmlNodePtr *nodes = NULL;
1777
    virQEMUCapsPtr qemuCaps = NULL;
1778
    virCapsPtr caps = NULL;
1779

1780
    if (VIR_ALLOC(priv->monConfig) < 0)
1781 1782 1783 1784
        goto error;

    if (!(monitorpath =
          virXPathString("string(./monitor[1]/@path)", ctxt))) {
1785 1786
        virReportError(VIR_ERR_INTERNAL_ERROR,
                       "%s", _("no monitor path"));
1787 1788 1789 1790 1791
        goto error;
    }

    tmp = virXPathString("string(./monitor[1]/@type)", ctxt);
    if (tmp)
1792
        priv->monConfig->type = virDomainChrTypeFromString(tmp);
1793
    else
1794
        priv->monConfig->type = VIR_DOMAIN_CHR_TYPE_PTY;
1795 1796
    VIR_FREE(tmp);

E
Eric Blake 已提交
1797 1798
    priv->monJSON = virXPathBoolean("count(./monitor[@json = '1']) > 0",
                                    ctxt) > 0;
1799

1800
    switch (priv->monConfig->type) {
1801
    case VIR_DOMAIN_CHR_TYPE_PTY:
1802
        priv->monConfig->data.file.path = monitorpath;
1803 1804
        break;
    case VIR_DOMAIN_CHR_TYPE_UNIX:
1805
        priv->monConfig->data.nix.path = monitorpath;
1806 1807 1808
        break;
    default:
        VIR_FREE(monitorpath);
1809 1810 1811
        virReportError(VIR_ERR_INTERNAL_ERROR,
                       _("unsupported monitor type '%s'"),
                       virDomainChrTypeToString(priv->monConfig->type));
1812 1813 1814
        goto error;
    }

1815
    if ((n = virXPathNodeSet("./vcpus/vcpu", ctxt, &nodes)) < 0)
1816 1817
        goto error;

1818
    for (i = 0; i < n; i++) {
1819
        if (qemuDomainObjPrivateXMLParseVcpu(nodes[i], i, vm->def) < 0)
1820
            goto error;
1821
    }
1822
    VIR_FREE(nodes);
1823

1824
    if ((n = virXPathNodeSet("./qemuCaps/flag", ctxt, &nodes)) < 0) {
1825 1826
        virReportError(VIR_ERR_INTERNAL_ERROR,
                       "%s", _("failed to parse qemu capabilities flags"));
1827 1828 1829
        goto error;
    }
    if (n > 0) {
1830
        if (!(qemuCaps = virQEMUCapsNew()))
1831 1832
            goto error;

1833
        for (i = 0; i < n; i++) {
1834 1835
            char *str = virXMLPropString(nodes[i], "name");
            if (str) {
1836
                int flag = virQEMUCapsTypeFromString(str);
1837
                if (flag < 0) {
1838 1839
                    virReportError(VIR_ERR_INTERNAL_ERROR,
                                   _("Unknown qemu capabilities flag %s"), str);
1840
                    VIR_FREE(str);
1841 1842
                    goto error;
                }
1843
                VIR_FREE(str);
1844
                virQEMUCapsSet(qemuCaps, flag);
1845 1846 1847
            }
        }

1848
        priv->qemuCaps = qemuCaps;
1849
        qemuCaps = NULL;
1850 1851 1852
    }
    VIR_FREE(nodes);

1853
    priv->lockState = virXPathString("string(./lockstate)", ctxt);
1854

1855 1856 1857 1858
    if ((tmp = virXPathString("string(./job[1]/@type)", ctxt))) {
        int type;

        if ((type = qemuDomainJobTypeFromString(tmp)) < 0) {
1859 1860
            virReportError(VIR_ERR_INTERNAL_ERROR,
                           _("Unknown job type %s"), tmp);
1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871
            VIR_FREE(tmp);
            goto error;
        }
        VIR_FREE(tmp);
        priv->job.active = type;
    }

    if ((tmp = virXPathString("string(./job[1]/@async)", ctxt))) {
        int async;

        if ((async = qemuDomainAsyncJobTypeFromString(tmp)) < 0) {
1872 1873
            virReportError(VIR_ERR_INTERNAL_ERROR,
                           _("Unknown async job type %s"), tmp);
1874 1875 1876 1877 1878
            VIR_FREE(tmp);
            goto error;
        }
        VIR_FREE(tmp);
        priv->job.asyncJob = async;
J
Jiri Denemark 已提交
1879 1880 1881 1882

        if ((tmp = virXPathString("string(./job[1]/@phase)", ctxt))) {
            priv->job.phase = qemuDomainAsyncJobPhaseFromString(async, tmp);
            if (priv->job.phase < 0) {
1883 1884
                virReportError(VIR_ERR_INTERNAL_ERROR,
                               _("Unknown job phase %s"), tmp);
J
Jiri Denemark 已提交
1885 1886 1887 1888 1889
                VIR_FREE(tmp);
                goto error;
            }
            VIR_FREE(tmp);
        }
1890 1891
    }

1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914
    if ((n = virXPathNodeSet("./job[1]/disk[@migrating='yes']",
                             ctxt, &nodes)) < 0) {
        virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                       _("failed to parse list of disks marked for migration"));
        goto error;
    }
    if (n > 0) {
        if (priv->job.asyncJob != QEMU_ASYNC_JOB_MIGRATION_OUT) {
            VIR_WARN("Found disks marked for migration but we were not "
                     "migrating");
            n = 0;
        }
        for (i = 0; i < n; i++) {
            char *dst = virXMLPropString(nodes[i], "dev");
            virDomainDiskDefPtr disk;

            if (dst && (disk = virDomainDiskByName(vm->def, dst, false)))
                QEMU_DOMAIN_DISK_PRIVATE(disk)->migrating = true;
            VIR_FREE(dst);
        }
    }
    VIR_FREE(nodes);

1915 1916
    priv->fakeReboot = virXPathBoolean("boolean(./fakereboot)", ctxt) == 1;

1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937
    if ((n = virXPathNodeSet("./devices/device", ctxt, &nodes)) < 0) {
        virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                       _("failed to parse qemu device list"));
        goto error;
    }
    if (n > 0) {
        /* NULL-terminated list */
        if (VIR_ALLOC_N(priv->qemuDevices, n + 1) < 0)
            goto error;

        for (i = 0; i < n; i++) {
            priv->qemuDevices[i] = virXMLPropString(nodes[i], "alias");
            if (!priv->qemuDevices[i]) {
                virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                               _("failed to parse qemu device list"));
                goto error;
            }
        }
    }
    VIR_FREE(nodes);

1938 1939 1940 1941
    if (!(caps = virQEMUDriverGetCapabilities(driver, false)))
        goto error;

    if ((tmp = virXPathString("string(./numad/@nodeset)", ctxt))) {
1942
        if (virBitmapParse(tmp, &priv->autoNodeset,
1943 1944 1945 1946 1947 1948 1949 1950 1951 1952
                           caps->host.nnumaCell_max) < 0)
            goto error;

        if (!(priv->autoCpuset = virCapabilitiesGetCpusForNodemask(caps,
                                                                   priv->autoNodeset)))
            goto error;
    }
    virObjectUnref(caps);
    VIR_FREE(tmp);

1953 1954 1955 1956 1957 1958 1959 1960 1961
    if ((tmp = virXPathString("string(./libDir/@path)", ctxt)))
        priv->libDir = tmp;
    if ((tmp = virXPathString("string(./channelTargetDir/@path)", ctxt)))
        priv->channelTargetDir = tmp;
    tmp = NULL;

    if (qemuDomainSetPrivatePathsOld(driver, vm) < 0)
        goto error;

1962 1963
    return 0;

1964
 error:
1965
    virDomainChrSourceDefFree(priv->monConfig);
1966 1967
    priv->monConfig = NULL;
    VIR_FREE(nodes);
1968
    VIR_FREE(tmp);
1969 1970
    virStringFreeList(priv->qemuDevices);
    priv->qemuDevices = NULL;
1971
    virObjectUnref(qemuCaps);
1972
    virObjectUnref(caps);
1973 1974 1975 1976
    return -1;
}


1977 1978 1979
virDomainXMLPrivateDataCallbacks virQEMUDriverPrivateDataCallbacks = {
    .alloc = qemuDomainObjPrivateAlloc,
    .free = qemuDomainObjPrivateFree,
1980
    .diskNew = qemuDomainDiskPrivateNew,
1981
    .vcpuNew = qemuDomainVcpuPrivateNew,
1982
    .hostdevNew = qemuDomainHostdevPrivateNew,
1983
    .chrSourceNew = qemuDomainChrSourcePrivateNew,
1984 1985 1986 1987 1988
    .parse = qemuDomainObjPrivateXMLParse,
    .format = qemuDomainObjPrivateXMLFormat,
};


1989 1990 1991 1992 1993
static void
qemuDomainDefNamespaceFree(void *nsdata)
{
    qemuDomainCmdlineDefPtr cmd = nsdata;

1994
    qemuDomainCmdlineDefFree(cmd);
1995 1996 1997
}

static int
P
Philipp Hahn 已提交
1998 1999
qemuDomainDefNamespaceParse(xmlDocPtr xml ATTRIBUTE_UNUSED,
                            xmlNodePtr root ATTRIBUTE_UNUSED,
2000 2001 2002 2003
                            xmlXPathContextPtr ctxt,
                            void **data)
{
    qemuDomainCmdlineDefPtr cmd = NULL;
P
Philipp Hahn 已提交
2004
    bool uses_qemu_ns = false;
2005
    xmlNodePtr *nodes = NULL;
2006 2007
    int n;
    size_t i;
2008

P
Philipp Hahn 已提交
2009
    if (xmlXPathRegisterNs(ctxt, BAD_CAST "qemu", BAD_CAST QEMU_NAMESPACE_HREF) < 0) {
2010 2011 2012
        virReportError(VIR_ERR_INTERNAL_ERROR,
                       _("Failed to register xml namespace '%s'"),
                       QEMU_NAMESPACE_HREF);
2013 2014 2015
        return -1;
    }

2016
    if (VIR_ALLOC(cmd) < 0)
2017 2018 2019 2020 2021 2022
        return -1;

    /* first handle the extra command-line arguments */
    n = virXPathNodeSet("./qemu:commandline/qemu:arg", ctxt, &nodes);
    if (n < 0)
        goto error;
P
Philipp Hahn 已提交
2023
    uses_qemu_ns |= n > 0;
2024 2025

    if (n && VIR_ALLOC_N(cmd->args, n) < 0)
2026
        goto error;
2027 2028 2029 2030

    for (i = 0; i < n; i++) {
        cmd->args[cmd->num_args] = virXMLPropString(nodes[i], "value");
        if (cmd->args[cmd->num_args] == NULL) {
2031 2032
            virReportError(VIR_ERR_INTERNAL_ERROR,
                           "%s", _("No qemu command-line argument specified"));
2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043
            goto error;
        }
        cmd->num_args++;
    }

    VIR_FREE(nodes);

    /* now handle the extra environment variables */
    n = virXPathNodeSet("./qemu:commandline/qemu:env", ctxt, &nodes);
    if (n < 0)
        goto error;
P
Philipp Hahn 已提交
2044
    uses_qemu_ns |= n > 0;
2045 2046

    if (n && VIR_ALLOC_N(cmd->env_name, n) < 0)
2047
        goto error;
2048 2049

    if (n && VIR_ALLOC_N(cmd->env_value, n) < 0)
2050
        goto error;
2051 2052 2053 2054 2055 2056

    for (i = 0; i < n; i++) {
        char *tmp;

        tmp = virXMLPropString(nodes[i], "name");
        if (tmp == NULL) {
2057 2058
            virReportError(VIR_ERR_INTERNAL_ERROR,
                           "%s", _("No qemu environment name specified"));
2059 2060 2061
            goto error;
        }
        if (tmp[0] == '\0') {
2062 2063
            virReportError(VIR_ERR_INTERNAL_ERROR,
                           "%s", _("Empty qemu environment name specified"));
2064 2065 2066
            goto error;
        }
        if (!c_isalpha(tmp[0]) && tmp[0] != '_') {
2067 2068
            virReportError(VIR_ERR_INTERNAL_ERROR,
                           "%s", _("Invalid environment name, it must begin with a letter or underscore"));
2069 2070 2071
            goto error;
        }
        if (strspn(tmp, "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789_") != strlen(tmp)) {
2072 2073
            virReportError(VIR_ERR_INTERNAL_ERROR,
                           "%s", _("Invalid environment name, it must contain only alphanumerics and underscore"));
2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085
            goto error;
        }

        cmd->env_name[cmd->num_env] = tmp;

        cmd->env_value[cmd->num_env] = virXMLPropString(nodes[i], "value");
        /* a NULL value for command is allowed, since it might be empty */
        cmd->num_env++;
    }

    VIR_FREE(nodes);

P
Philipp Hahn 已提交
2086 2087 2088 2089
    if (uses_qemu_ns)
        *data = cmd;
    else
        VIR_FREE(cmd);
2090 2091 2092

    return 0;

2093
 error:
2094 2095 2096 2097 2098 2099 2100 2101 2102 2103
    VIR_FREE(nodes);
    qemuDomainDefNamespaceFree(cmd);
    return -1;
}

static int
qemuDomainDefNamespaceFormatXML(virBufferPtr buf,
                                void *nsdata)
{
    qemuDomainCmdlineDefPtr cmd = nsdata;
2104
    size_t i;
2105 2106 2107 2108

    if (!cmd->num_args && !cmd->num_env)
        return 0;

2109 2110 2111
    virBufferAddLit(buf, "<qemu:commandline>\n");
    virBufferAdjustIndent(buf, 2);

2112
    for (i = 0; i < cmd->num_args; i++)
2113
        virBufferEscapeString(buf, "<qemu:arg value='%s'/>\n",
2114 2115
                              cmd->args[i]);
    for (i = 0; i < cmd->num_env; i++) {
2116
        virBufferAsprintf(buf, "<qemu:env name='%s'", cmd->env_name[i]);
2117 2118 2119 2120 2121
        if (cmd->env_value[i])
            virBufferEscapeString(buf, " value='%s'", cmd->env_value[i]);
        virBufferAddLit(buf, "/>\n");
    }

2122 2123
    virBufferAdjustIndent(buf, -2);
    virBufferAddLit(buf, "</qemu:commandline>\n");
2124 2125 2126 2127 2128 2129 2130 2131 2132 2133
    return 0;
}

static const char *
qemuDomainDefNamespaceHref(void)
{
    return "xmlns:qemu='" QEMU_NAMESPACE_HREF "'";
}


2134 2135 2136 2137 2138 2139
virDomainXMLNamespace virQEMUDriverDomainXMLNamespace = {
    .parse = qemuDomainDefNamespaceParse,
    .free = qemuDomainDefNamespaceFree,
    .format = qemuDomainDefNamespaceFormatXML,
    .href = qemuDomainDefNamespaceHref,
};
2140

2141

P
Pavel Hrdina 已提交
2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160
static int
qemuDomainDefAddImplicitInputDevice(virDomainDef *def)
{
    if (ARCH_IS_X86(def->os.arch)) {
        if (virDomainDefMaybeAddInput(def,
                                      VIR_DOMAIN_INPUT_TYPE_MOUSE,
                                      VIR_DOMAIN_INPUT_BUS_PS2) < 0)
            return -1;

        if (virDomainDefMaybeAddInput(def,
                                      VIR_DOMAIN_INPUT_TYPE_KBD,
                                      VIR_DOMAIN_INPUT_BUS_PS2) < 0)
            return -1;
    }

    return 0;
}


2161
static int
2162 2163
qemuDomainDefAddDefaultDevices(virDomainDefPtr def,
                               virQEMUCapsPtr qemuCaps)
2164
{
2165
    bool addDefaultUSB = true;
2166
    int usbModel = -1; /* "default for machinetype" */
2167
    int pciRoot;       /* index within def->controllers */
2168
    bool addImplicitSATA = false;
2169
    bool addPCIRoot = false;
L
Laine Stump 已提交
2170
    bool addPCIeRoot = false;
2171
    bool addDefaultMemballoon = true;
2172 2173
    bool addDefaultUSBKBD = false;
    bool addDefaultUSBMouse = false;
2174
    bool addPanicDevice = false;
2175
    int ret = -1;
2176

P
Pavel Hrdina 已提交
2177 2178 2179 2180
    /* add implicit input devices */
    if (qemuDomainDefAddImplicitInputDevice(def) < 0)
        goto cleanup;

2181 2182 2183 2184
    /* Add implicit PCI root controller if the machine has one */
    switch (def->os.arch) {
    case VIR_ARCH_I686:
    case VIR_ARCH_X86_64:
L
Laine Stump 已提交
2185
        if (STREQ(def->os.machine, "isapc")) {
2186
            addDefaultUSB = false;
2187
            break;
2188
        }
2189
        if (qemuDomainMachineIsQ35(def)) {
2190 2191
            addPCIeRoot = true;
            addImplicitSATA = true;
2192

2193 2194 2195 2196
            /* Prefer adding USB3 controller if supported
             * (nec-usb-xhci). Failing that, add a USB2 controller set
             * if the ich9-usb-ehci1 device is supported. Otherwise
             * don't add anything.
2197
             */
2198 2199 2200
            if (virQEMUCapsGet(qemuCaps, QEMU_CAPS_NEC_USB_XHCI))
                usbModel = VIR_DOMAIN_CONTROLLER_MODEL_USB_NEC_XHCI;
            else if (virQEMUCapsGet(qemuCaps, QEMU_CAPS_ICH9_USB_EHCI1))
2201 2202 2203
                usbModel = VIR_DOMAIN_CONTROLLER_MODEL_USB_ICH9_EHCI1;
            else
                addDefaultUSB = false;
2204
            break;
L
Laine Stump 已提交
2205
        }
2206 2207
        if (qemuDomainMachineIsI440FX(def))
            addPCIRoot = true;
2208 2209
        break;

2210
    case VIR_ARCH_ARMV7L:
2211
    case VIR_ARCH_AARCH64:
2212 2213
        addDefaultUSB = false;
        addDefaultMemballoon = false;
J
Ján Tomko 已提交
2214
        if (qemuDomainMachineIsVirt(def))
2215 2216
            addPCIeRoot = virQEMUCapsGet(qemuCaps, QEMU_CAPS_OBJECT_GPEX);
        break;
2217

2218
    case VIR_ARCH_PPC64:
2219
    case VIR_ARCH_PPC64LE:
2220 2221 2222
        addPCIRoot = true;
        addDefaultUSBKBD = true;
        addDefaultUSBMouse = true;
2223 2224 2225
        /* For pSeries guests, the firmware provides the same
         * functionality as the pvpanic device, so automatically
         * add the definition if not already present */
2226
        if (qemuDomainMachineIsPSeries(def))
2227
            addPanicDevice = true;
2228 2229
        break;

2230 2231 2232 2233 2234 2235 2236
    case VIR_ARCH_ALPHA:
    case VIR_ARCH_PPC:
    case VIR_ARCH_PPCEMB:
    case VIR_ARCH_SH4:
    case VIR_ARCH_SH4EB:
        addPCIRoot = true;
        break;
2237

2238 2239 2240
    case VIR_ARCH_S390:
    case VIR_ARCH_S390X:
        addDefaultUSB = false;
2241
        addPanicDevice = true;
2242
        break;
2243 2244 2245 2246 2247 2248

    case VIR_ARCH_SPARC:
    case VIR_ARCH_SPARC64:
        addPCIRoot = true;
        break;

2249 2250 2251 2252
    default:
        break;
    }

2253
    if (addDefaultUSB &&
2254 2255
        virDomainControllerFind(def, VIR_DOMAIN_CONTROLLER_TYPE_USB, 0) < 0 &&
        virDomainDefAddUSBController(def, 0, usbModel) < 0)
2256
        goto cleanup;
2257

2258 2259 2260
    if (addImplicitSATA &&
        virDomainDefMaybeAddController(
            def, VIR_DOMAIN_CONTROLLER_TYPE_SATA, 0, -1) < 0)
2261
        goto cleanup;
2262

2263 2264
    pciRoot = virDomainControllerFind(def, VIR_DOMAIN_CONTROLLER_TYPE_PCI, 0);

2265 2266 2267
    /* NB: any machine that sets addPCIRoot to true must also return
     * true from the function qemuDomainSupportsPCI().
     */
2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282
    if (addPCIRoot) {
        if (pciRoot >= 0) {
            if (def->controllers[pciRoot]->model != VIR_DOMAIN_CONTROLLER_MODEL_PCI_ROOT) {
                virReportError(VIR_ERR_XML_ERROR,
                               _("The PCI controller with index='0' must be "
                                 "model='pci-root' for this machine type, "
                                 "but model='%s' was found instead"),
                               virDomainControllerModelPCITypeToString(def->controllers[pciRoot]->model));
                goto cleanup;
            }
        } else if (!virDomainDefAddController(def, VIR_DOMAIN_CONTROLLER_TYPE_PCI, 0,
                                              VIR_DOMAIN_CONTROLLER_MODEL_PCI_ROOT)) {
            goto cleanup;
        }
    }
2283

2284 2285 2286
    /* When a machine has a pcie-root, make sure that there is always
     * a dmi-to-pci-bridge controller added as bus 1, and a pci-bridge
     * as bus 2, so that standard PCI devices can be connected
2287 2288 2289
     *
     * NB: any machine that sets addPCIeRoot to true must also return
     * true from the function qemuDomainSupportsPCI().
2290 2291
     */
    if (addPCIeRoot) {
2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304
        if (pciRoot >= 0) {
            if (def->controllers[pciRoot]->model != VIR_DOMAIN_CONTROLLER_MODEL_PCIE_ROOT) {
                virReportError(VIR_ERR_XML_ERROR,
                               _("The PCI controller with index='0' must be "
                                 "model='pcie-root' for this machine type, "
                                 "but model='%s' was found instead"),
                               virDomainControllerModelPCITypeToString(def->controllers[pciRoot]->model));
                goto cleanup;
            }
        } else if (!virDomainDefAddController(def, VIR_DOMAIN_CONTROLLER_TYPE_PCI, 0,
                                             VIR_DOMAIN_CONTROLLER_MODEL_PCIE_ROOT)) {
            goto cleanup;
        }
2305
    }
2306

2307
    if (addDefaultMemballoon && !def->memballoon) {
2308 2309
        virDomainMemballoonDefPtr memballoon;
        if (VIR_ALLOC(memballoon) < 0)
2310
            goto cleanup;
2311 2312 2313 2314 2315

        memballoon->model = VIR_DOMAIN_MEMBALLOON_MODEL_VIRTIO;
        def->memballoon = memballoon;
    }

2316 2317 2318 2319 2320
    if (addDefaultUSBKBD &&
        def->ngraphics > 0 &&
        virDomainDefMaybeAddInput(def,
                                  VIR_DOMAIN_INPUT_TYPE_KBD,
                                  VIR_DOMAIN_INPUT_BUS_USB) < 0)
2321
        goto cleanup;
2322 2323 2324 2325 2326 2327

    if (addDefaultUSBMouse &&
        def->ngraphics > 0 &&
        virDomainDefMaybeAddInput(def,
                                  VIR_DOMAIN_INPUT_TYPE_MOUSE,
                                  VIR_DOMAIN_INPUT_BUS_USB) < 0)
2328
        goto cleanup;
2329

D
Dmitry Andreev 已提交
2330 2331 2332 2333
    if (addPanicDevice) {
        size_t j;
        for (j = 0; j < def->npanics; j++) {
            if (def->panics[j]->model == VIR_DOMAIN_PANIC_MODEL_DEFAULT ||
2334 2335 2336 2337
                (ARCH_IS_PPC64(def->os.arch) &&
                     def->panics[j]->model == VIR_DOMAIN_PANIC_MODEL_PSERIES) ||
                (ARCH_IS_S390(def->os.arch) &&
                     def->panics[j]->model == VIR_DOMAIN_PANIC_MODEL_S390))
D
Dmitry Andreev 已提交
2338 2339
                break;
        }
2340

D
Dmitry Andreev 已提交
2341 2342 2343 2344 2345 2346 2347 2348 2349
        if (j == def->npanics) {
            virDomainPanicDefPtr panic;
            if (VIR_ALLOC(panic) < 0 ||
                VIR_APPEND_ELEMENT_COPY(def->panics,
                                        def->npanics, panic) < 0) {
                VIR_FREE(panic);
                goto cleanup;
            }
        }
2350 2351
    }

2352 2353 2354 2355 2356 2357
    ret = 0;
 cleanup:
    return ret;
}


A
Andrea Bolognani 已提交
2358 2359 2360
/**
 * qemuDomainDefEnableDefaultFeatures:
 * @def: domain definition
2361
 * @qemuCaps: QEMU capabilities
A
Andrea Bolognani 已提交
2362 2363 2364 2365 2366
 *
 * Make sure that features that should be enabled by default are actually
 * enabled and configure default values related to those features.
 */
static void
2367 2368
qemuDomainDefEnableDefaultFeatures(virDomainDefPtr def,
                                   virQEMUCapsPtr qemuCaps)
A
Andrea Bolognani 已提交
2369
{
2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389
    virGICVersion version;

    /* The virt machine type always uses GIC: if the relevant element
     * was not included in the domain XML, we need to choose a suitable
     * GIC version ourselves */
    if (def->features[VIR_DOMAIN_FEATURE_GIC] == VIR_TRISTATE_SWITCH_ABSENT &&
        qemuDomainMachineIsVirt(def)) {

        VIR_DEBUG("Looking for usable GIC version in domain capabilities");
        for (version = VIR_GIC_VERSION_LAST - 1;
             version > VIR_GIC_VERSION_NONE;
             version--) {
            if (virQEMUCapsSupportsGICVersion(qemuCaps,
                                              def->virtType,
                                              version)) {
                VIR_DEBUG("Using GIC version %s",
                          virGICVersionTypeToString(version));
                def->gic_version = version;
                break;
            }
2390 2391
        }

2392 2393 2394
        /* Even if we haven't found a usable GIC version in the domain
         * capabilities, we still want to enable this */
        def->features[VIR_DOMAIN_FEATURE_GIC] = VIR_TRISTATE_SWITCH_ON;
2395 2396
    }

2397
    /* Use the default GIC version if no version was specified */
A
Andrea Bolognani 已提交
2398 2399
    if (def->features[VIR_DOMAIN_FEATURE_GIC] == VIR_TRISTATE_SWITCH_ON &&
        def->gic_version == VIR_GIC_VERSION_NONE)
2400
        def->gic_version = VIR_GIC_VERSION_DEFAULT;
A
Andrea Bolognani 已提交
2401 2402 2403
}


2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423
static int
qemuCanonicalizeMachine(virDomainDefPtr def, virQEMUCapsPtr qemuCaps)
{
    const char *canon;

    if (!(canon = virQEMUCapsGetCanonicalMachine(qemuCaps, def->os.machine)))
        return 0;

    if (STRNEQ(canon, def->os.machine)) {
        char *tmp;
        if (VIR_STRDUP(tmp, canon) < 0)
            return -1;
        VIR_FREE(def->os.machine);
        def->os.machine = tmp;
    }

    return 0;
}


2424
static int
2425 2426 2427 2428 2429
qemuDomainRecheckInternalPaths(virDomainDefPtr def,
                               virQEMUDriverConfigPtr cfg,
                               unsigned int flags)
{
    size_t i = 0;
2430
    size_t j = 0;
2431 2432 2433 2434

    for (i = 0; i < def->ngraphics; ++i) {
        virDomainGraphicsDefPtr graphics = def->graphics[i];

2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452
        for (j = 0; j < graphics->nListens; ++j) {
            virDomainGraphicsListenDefPtr glisten =  &graphics->listens[j];

            /* This will happen only if we parse XML from old libvirts where
             * unix socket was available only for VNC graphics.  In this
             * particular case we should follow the behavior and if we remove
             * the auto-generated socket base on config option from qemu.conf
             * we need to change the listen type to address. */
            if (graphics->type == VIR_DOMAIN_GRAPHICS_TYPE_VNC &&
                glisten->type == VIR_DOMAIN_GRAPHICS_LISTEN_TYPE_SOCKET &&
                glisten->socket &&
                STRPREFIX(glisten->socket, cfg->libDir)) {
                if (flags & VIR_DOMAIN_DEF_PARSE_INACTIVE) {
                    VIR_FREE(glisten->socket);
                    glisten->type = VIR_DOMAIN_GRAPHICS_LISTEN_TYPE_ADDRESS;
                } else {
                    glisten->fromConfig = true;
                }
2453
            }
2454 2455
        }
    }
2456 2457

    return 0;
2458 2459 2460
}


2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 2527 2528 2529 2530
static int
qemuDomainDefVcpusPostParse(virDomainDefPtr def)
{
    unsigned int maxvcpus = virDomainDefGetVcpusMax(def);
    virDomainVcpuDefPtr vcpu;
    virDomainVcpuDefPtr prevvcpu;
    size_t i;
    bool has_order = false;

    /* vcpu 0 needs to be present, first, and non-hotpluggable */
    vcpu = virDomainDefGetVcpu(def, 0);
    if (!vcpu->online) {
        virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                       _("vcpu 0 can't be offline"));
        return -1;
    }
    if (vcpu->hotpluggable == VIR_TRISTATE_BOOL_YES) {
        virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                       _("vcpu0 can't be hotpluggable"));
        return -1;
    }
    if (vcpu->order != 0 && vcpu->order != 1) {
        virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                       _("vcpu0 must be enabled first"));
        return -1;
    }

    if (vcpu->order != 0)
        has_order = true;

    prevvcpu = vcpu;

    /* all online vcpus or non online vcpu need to have order set */
    for (i = 1; i < maxvcpus; i++) {
        vcpu = virDomainDefGetVcpu(def, i);

        if (vcpu->online &&
            (vcpu->order != 0) != has_order) {
            virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                           _("all vcpus must have either set or unset order"));
            return -1;
        }

        /* few conditions for non-hotpluggable (thus online) vcpus */
        if (vcpu->hotpluggable == VIR_TRISTATE_BOOL_NO) {
            /* they can be ordered only at the beginning */
            if (prevvcpu->hotpluggable == VIR_TRISTATE_BOOL_YES) {
                virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                               _("online non-hotpluggable vcpus need to be "
                                 "ordered prior to hotplugable vcpus"));
                return -1;
            }

            /* they need to be in order (qemu doesn't support any order yet).
             * Also note that multiple vcpus may share order on some platforms */
            if (prevvcpu->order > vcpu->order) {
                virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                               _("online non-hotpluggable vcpus must be ordered "
                                 "in ascending order"));
                return -1;
            }
        }

        prevvcpu = vcpu;
    }

    return 0;
}


2531 2532 2533
static int
qemuDomainDefPostParse(virDomainDefPtr def,
                       virCapsPtr caps,
2534
                       unsigned int parseFlags,
2535
                       void *opaque,
2536
                       void *parseOpaque)
2537 2538
{
    virQEMUDriverPtr driver = opaque;
2539
    virQEMUDriverConfigPtr cfg = virQEMUDriverGetConfig(driver);
2540
    virQEMUCapsPtr qemuCaps = parseOpaque;
2541 2542 2543 2544 2545
    int ret = -1;

    if (def->os.bootloader || def->os.bootloaderArgs) {
        virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                       _("bootloader is not supported by QEMU"));
2546
        goto cleanup;
2547 2548
    }

2549 2550 2551
    if (!def->os.machine) {
        virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                       _("missing machine type"));
2552
        goto cleanup;
2553 2554
    }

2555 2556 2557 2558 2559 2560 2561 2562 2563
    if (def->os.loader &&
        def->os.loader->type == VIR_DOMAIN_LOADER_TYPE_PFLASH &&
        def->os.loader->readonly == VIR_TRISTATE_SWITCH_ON &&
        !def->os.loader->nvram) {
        if (virAsprintf(&def->os.loader->nvram, "%s/%s_VARS.fd",
                        cfg->nvramDir, def->name) < 0)
            goto cleanup;
    }

2564 2565 2566
    /* check for emulator and create a default one if needed */
    if (!def->emulator &&
        !(def->emulator = virDomainDefGetDefaultEmulator(def, caps)))
2567
        goto cleanup;
2568

2569 2570 2571 2572 2573 2574 2575 2576
    if (qemuCaps) {
        virObjectRef(qemuCaps);
    } else {
        if (!(qemuCaps = virQEMUCapsCacheLookup(caps,
                                                driver->qemuCapsCache,
                                                def->emulator)))
            goto cleanup;
    }
2577 2578 2579 2580

    if (qemuDomainDefAddDefaultDevices(def, qemuCaps) < 0)
        goto cleanup;

2581 2582 2583
    if (qemuCanonicalizeMachine(def, qemuCaps) < 0)
        goto cleanup;

2584
    qemuDomainDefEnableDefaultFeatures(def, qemuCaps);
A
Andrea Bolognani 已提交
2585

2586 2587
    if (qemuDomainRecheckInternalPaths(def, cfg, parseFlags) < 0)
        goto cleanup;
2588

2589 2590 2591
    if (virSecurityManagerVerify(driver->securityManager, def) < 0)
        goto cleanup;

2592 2593 2594
    if (qemuDomainDefVcpusPostParse(def) < 0)
        goto cleanup;

2595 2596 2597
    ret = 0;
 cleanup:
    virObjectUnref(qemuCaps);
2598
    virObjectUnref(cfg);
2599
    return ret;
2600 2601
}

2602

2603 2604 2605 2606 2607 2608 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 2624 2625 2626 2627 2628 2629
static int
qemuDomainDefValidateVideo(const virDomainDef *def)
{
    size_t i;
    virDomainVideoDefPtr video;

    for (i = 0; i < def->nvideos; i++) {
        video = def->videos[i];

        switch (video->type) {
        case VIR_DOMAIN_VIDEO_TYPE_XEN:
        case VIR_DOMAIN_VIDEO_TYPE_VBOX:
        case VIR_DOMAIN_VIDEO_TYPE_PARALLELS:
            virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
                           _("video type '%s' is not supported with QEMU"),
                           virDomainVideoTypeToString(video->type));
            return -1;
        case VIR_DOMAIN_VIDEO_TYPE_VGA:
        case VIR_DOMAIN_VIDEO_TYPE_CIRRUS:
        case VIR_DOMAIN_VIDEO_TYPE_VMVGA:
        case VIR_DOMAIN_VIDEO_TYPE_QXL:
        case VIR_DOMAIN_VIDEO_TYPE_VIRTIO:
        case VIR_DOMAIN_VIDEO_TYPE_LAST:
            break;
        }

        if (!video->primary &&
2630 2631
            video->type != VIR_DOMAIN_VIDEO_TYPE_QXL &&
            video->type != VIR_DOMAIN_VIDEO_TYPE_VIRTIO) {
2632 2633 2634 2635 2636 2637 2638 2639 2640 2641 2642 2643 2644 2645 2646 2647 2648 2649 2650 2651 2652 2653 2654 2655 2656 2657 2658 2659 2660 2661 2662 2663 2664 2665 2666 2667 2668 2669 2670 2671 2672 2673 2674
            virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
                           _("video type '%s' is only valid as primary "
                             "video device"),
                           virDomainVideoTypeToString(video->type));
            return -1;
        }

        if (video->accel && video->accel->accel2d == VIR_TRISTATE_SWITCH_ON) {
            virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                           _("qemu does not support the accel2d setting"));
            return -1;
        }

        if (video->type == VIR_DOMAIN_VIDEO_TYPE_QXL) {
            if (video->vram > (UINT_MAX / 1024)) {
                virReportError(VIR_ERR_OVERFLOW,
                               _("value for 'vram' must be less than '%u'"),
                               UINT_MAX / 1024);
                return -1;
            }
            if (video->ram > (UINT_MAX / 1024)) {
                virReportError(VIR_ERR_OVERFLOW,
                               _("value for 'ram' must be less than '%u'"),
                               UINT_MAX / 1024);
                return -1;
            }
        }

        if (video->type == VIR_DOMAIN_VIDEO_TYPE_VGA ||
            video->type == VIR_DOMAIN_VIDEO_TYPE_VMVGA) {
            if (video->vram && video->vram < 1024) {
                virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
                               "%s", _("value for 'vram' must be at least "
                                       "1 MiB (1024 KiB)"));
                return -1;
            }
        }
    }

    return 0;
}


2675 2676
static int
qemuDomainDefValidate(const virDomainDef *def,
2677
                      virCapsPtr caps,
2678
                      void *opaque)
2679
{
2680 2681
    virQEMUDriverPtr driver = opaque;
    virQEMUCapsPtr qemuCaps = NULL;
2682
    unsigned int topologycpus;
2683 2684
    int ret = -1;

2685 2686
    if (!(qemuCaps = virQEMUCapsCacheLookup(caps,
                                            driver->qemuCapsCache,
2687 2688 2689
                                            def->emulator)))
        goto cleanup;

2690 2691 2692
    if (def->mem.min_guarantee) {
        virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                       _("Parameter 'min_guarantee' not supported by QEMU."));
2693
        goto cleanup;
2694 2695
    }

M
Michal Privoznik 已提交
2696 2697 2698 2699 2700 2701 2702 2703
    if (def->os.loader &&
        def->os.loader->secure == VIR_TRISTATE_BOOL_YES) {
        /* These are the QEMU implementation limitations. But we
         * have to live with them for now. */

        if (!qemuDomainMachineIsQ35(def)) {
            virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                           _("Secure boot is supported with q35 machine types only"));
2704
            goto cleanup;
M
Michal Privoznik 已提交
2705 2706 2707 2708 2709 2710 2711 2712
        }

        /* Now, technically it is possible to have secure boot on
         * 32bits too, but that would require some -cpu xxx magic
         * too. Not worth it unless we are explicitly asked. */
        if (def->os.arch != VIR_ARCH_X86_64) {
            virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                           _("Secure boot is supported for x86_64 architecture only"));
2713
            goto cleanup;
M
Michal Privoznik 已提交
2714 2715 2716 2717 2718
        }

        if (def->features[VIR_DOMAIN_FEATURE_SMM] != VIR_TRISTATE_SWITCH_ON) {
            virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                           _("Secure boot requires SMM feature enabled"));
2719
            goto cleanup;
M
Michal Privoznik 已提交
2720 2721 2722
        }
    }

2723
    /* qemu as of 2.5.0 rejects SMP topologies that don't match the cpu count */
2724 2725 2726 2727 2728 2729 2730
    if (virDomainDefGetVcpusTopology(def, &topologycpus) == 0 &&
        topologycpus != virDomainDefGetVcpusMax(def)) {
        /* presence of query-hotpluggable-cpus should be a good enough witness */
        if (virQEMUCapsGet(qemuCaps, QEMU_CAPS_QUERY_HOTPLUGGABLE_CPUS)) {
            virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                           _("CPU topology doesn't match maximum vcpu count"));
            goto cleanup;
2731 2732 2733
        }
    }

2734 2735 2736
    if (qemuDomainDefValidateVideo(def) < 0)
        goto cleanup;

2737 2738 2739 2740 2741
    ret = 0;

 cleanup:
    virObjectUnref(qemuCaps);
    return ret;
2742 2743 2744
}


2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763
static int
qemuDomainDeviceDefValidate(const virDomainDeviceDef *dev,
                            const virDomainDef *def ATTRIBUTE_UNUSED,
                            void *opaque)
{
    virQEMUDriverPtr driver = opaque;
    virQEMUDriverConfigPtr cfg = virQEMUDriverGetConfig(driver);
    int ret = -1;

    if (dev->type == VIR_DOMAIN_DEVICE_NET) {
        const virDomainNetDef *net = dev->data.net;

        if (net->guestIP.nroutes || net->guestIP.nips) {
            virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                           _("Invalid attempt to set network interface "
                             "guest-side IP route and/or address info, "
                             "not supported by QEMU"));
            goto cleanup;
        }
2764 2765 2766 2767 2768 2769 2770

        if (STREQ_NULLABLE(net->model, "virtio") &&
            net->driver.virtio.rx_queue_size & (net->driver.virtio.rx_queue_size - 1)) {
            virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                           _("rx_queue_size has to be a power of two"));
            goto cleanup;
        }
2771 2772 2773 2774 2775 2776 2777 2778 2779
    }

    ret = 0;
 cleanup:
    virObjectUnref(cfg);
    return ret;
}


2780
static const char *
2781 2782
qemuDomainDefaultNetModel(const virDomainDef *def,
                          virQEMUCapsPtr qemuCaps)
2783
{
2784
    if (ARCH_IS_S390(def->os.arch))
2785 2786
        return "virtio";

2787 2788
    if (def->os.arch == VIR_ARCH_ARMV7L ||
        def->os.arch == VIR_ARCH_AARCH64) {
2789 2790 2791
        if (STREQ(def->os.machine, "versatilepb"))
            return "smc91c111";

J
Ján Tomko 已提交
2792
        if (qemuDomainMachineIsVirt(def))
2793 2794
            return "virtio";

2795 2796 2797 2798 2799
        /* Incomplete. vexpress (and a few others) use this, but not all
         * arm boards */
        return "lan9118";
    }

2800 2801 2802 2803 2804 2805 2806 2807 2808 2809 2810 2811
    /* Try several network devices in turn; each of these devices is
     * less likely be supported out-of-the-box by the guest operating
     * system than the previous one */
    if (virQEMUCapsGet(qemuCaps, QEMU_CAPS_DEVICE_RTL8139))
        return "rtl8139";
    else if (virQEMUCapsGet(qemuCaps, QEMU_CAPS_DEVICE_E1000))
        return "e1000";
    else if (virQEMUCapsGet(qemuCaps, QEMU_CAPS_DEVICE_VIRTIO_NET))
        return "virtio";

    /* We've had no luck detecting support for any network device,
     * but we have to return something: might as well be rtl8139 */
2812 2813
    return "rtl8139";
}
2814

2815 2816 2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827

/*
 * Clear auto generated unix socket path, i.e., the one which starts with our
 * channel directory.
 */
static void
qemuDomainChrDefDropDefaultPath(virDomainChrDefPtr chr,
                                virQEMUDriverPtr driver)
{
    virQEMUDriverConfigPtr cfg = virQEMUDriverGetConfig(driver);

    if (chr->deviceType == VIR_DOMAIN_CHR_DEVICE_TYPE_CHANNEL &&
        chr->targetType == VIR_DOMAIN_CHR_CHANNEL_TARGET_TYPE_VIRTIO &&
2828 2829 2830 2831
        chr->source->type == VIR_DOMAIN_CHR_TYPE_UNIX &&
        chr->source->data.nix.path &&
        STRPREFIX(chr->source->data.nix.path, cfg->channelTargetDir)) {
        VIR_FREE(chr->source->data.nix.path);
2832 2833 2834 2835 2836 2837
    }

    virObjectUnref(cfg);
}


2838 2839 2840 2841 2842 2843 2844 2845 2846 2847 2848 2849 2850 2851 2852 2853 2854 2855 2856 2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 2867 2868 2869 2870 2871 2872 2873 2874 2875 2876 2877 2878 2879 2880 2881 2882 2883 2884 2885 2886 2887
static int
qemuDomainShmemDefPostParse(virDomainShmemDefPtr shm)
{
    /* This was the default since the introduction of this device. */
    if (shm->model != VIR_DOMAIN_SHMEM_MODEL_IVSHMEM_DOORBELL && !shm->size)
        shm->size = 4 << 20;

    /* Nothing more to check/change for IVSHMEM */
    if (shm->model == VIR_DOMAIN_SHMEM_MODEL_IVSHMEM)
        return 0;

    if (!shm->server.enabled) {
        if (shm->model == VIR_DOMAIN_SHMEM_MODEL_IVSHMEM_DOORBELL) {
            virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
                           _("shmem model '%s' is supported "
                             "only with server option enabled"),
                           virDomainShmemModelTypeToString(shm->model));
            return -1;
        }

        if (shm->msi.enabled) {
            virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
                           _("shmem model '%s' doesn't support "
                             "msi"),
                           virDomainShmemModelTypeToString(shm->model));
        }
    } else {
        if (shm->model == VIR_DOMAIN_SHMEM_MODEL_IVSHMEM_PLAIN) {
            virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
                           _("shmem model '%s' is supported "
                             "only with server option disabled"),
                           virDomainShmemModelTypeToString(shm->model));
            return -1;
        }

        if (shm->size) {
            virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
                           _("shmem model '%s' does not support size setting"),
                           virDomainShmemModelTypeToString(shm->model));
            return -1;
        }
        shm->msi.enabled = true;
        if (!shm->msi.ioeventfd)
            shm->msi.ioeventfd = VIR_TRISTATE_SWITCH_ON;
    }

    return 0;
}


2888 2889
static int
qemuDomainDeviceDefPostParse(virDomainDeviceDefPtr dev,
2890
                             const virDomainDef *def,
2891
                             virCapsPtr caps,
2892
                             unsigned int parseFlags,
2893
                             void *opaque,
2894
                             void *parseOpaque)
2895
{
2896
    virQEMUDriverPtr driver = opaque;
2897
    virQEMUCapsPtr qemuCaps = parseOpaque;
2898
    virQEMUDriverConfigPtr cfg = virQEMUDriverGetConfig(driver);
2899
    int ret = -1;
2900

2901 2902 2903 2904 2905 2906
    if (qemuCaps) {
        virObjectRef(qemuCaps);
    } else {
        qemuCaps = virQEMUCapsCacheLookup(caps, driver->qemuCapsCache,
                                          def->emulator);
    }
2907

2908
    if (dev->type == VIR_DOMAIN_DEVICE_NET &&
2909 2910
        dev->data.net->type != VIR_DOMAIN_NET_TYPE_HOSTDEV &&
        !dev->data.net->model) {
2911
        if (VIR_STRDUP(dev->data.net->model,
2912
                       qemuDomainDefaultNetModel(def, qemuCaps)) < 0)
2913
            goto cleanup;
2914
    }
2915

2916 2917 2918 2919
    /* set default disk types and drivers */
    if (dev->type == VIR_DOMAIN_DEVICE_DISK) {
        virDomainDiskDefPtr disk = dev->data.disk;

2920 2921 2922 2923 2924 2925 2926 2927 2928 2929 2930 2931 2932 2933 2934 2935 2936 2937 2938 2939 2940 2941 2942 2943 2944 2945 2946 2947
        /* assign default storage format and driver according to config */
        if (cfg->allowDiskFormatProbing) {
            /* default disk format for drives */
            if (virDomainDiskGetFormat(disk) == VIR_STORAGE_FILE_NONE &&
                (virDomainDiskGetType(disk) == VIR_STORAGE_TYPE_FILE ||
                 virDomainDiskGetType(disk) == VIR_STORAGE_TYPE_BLOCK))
                virDomainDiskSetFormat(disk, VIR_STORAGE_FILE_AUTO);

            /* default disk format for mirrored drive */
            if (disk->mirror &&
                disk->mirror->format == VIR_STORAGE_FILE_NONE)
                disk->mirror->format = VIR_STORAGE_FILE_AUTO;
        } else {
            /* default driver if probing is forbidden */
            if (!virDomainDiskGetDriver(disk) &&
                virDomainDiskSetDriver(disk, "qemu") < 0)
                goto cleanup;

            /* default disk format for drives */
            if (virDomainDiskGetFormat(disk) == VIR_STORAGE_FILE_NONE &&
                (virDomainDiskGetType(disk) == VIR_STORAGE_TYPE_FILE ||
                 virDomainDiskGetType(disk) == VIR_STORAGE_TYPE_BLOCK))
                virDomainDiskSetFormat(disk, VIR_STORAGE_FILE_RAW);

            /* default disk format for mirrored drive */
            if (disk->mirror &&
                disk->mirror->format == VIR_STORAGE_FILE_NONE)
                disk->mirror->format = VIR_STORAGE_FILE_RAW;
2948 2949 2950
        }
    }

2951 2952 2953 2954
    /* set the default console type for S390 arches */
    if (dev->type == VIR_DOMAIN_DEVICE_CHR &&
        dev->data.chr->deviceType == VIR_DOMAIN_CHR_DEVICE_TYPE_CONSOLE &&
        dev->data.chr->targetType == VIR_DOMAIN_CHR_CONSOLE_TARGET_TYPE_NONE &&
2955
        ARCH_IS_S390(def->os.arch))
2956 2957
        dev->data.chr->targetType = VIR_DOMAIN_CHR_CONSOLE_TARGET_TYPE_VIRTIO;

2958 2959 2960 2961 2962
    /* set the default USB model to none for s390 unless an address is found */
    if (dev->type == VIR_DOMAIN_DEVICE_CONTROLLER &&
        dev->data.controller->type == VIR_DOMAIN_CONTROLLER_TYPE_USB &&
        dev->data.controller->model == -1 &&
        dev->data.controller->info.type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_NONE &&
2963
        ARCH_IS_S390(def->os.arch))
2964 2965
        dev->data.controller->model = VIR_DOMAIN_CONTROLLER_MODEL_USB_NONE;

2966 2967 2968 2969 2970 2971 2972 2973 2974 2975 2976 2977 2978
    /* forbid usb model 'qusb1' and 'qusb2' in this kind of hyperviosr */
    if (dev->type == VIR_DOMAIN_DEVICE_CONTROLLER &&
        dev->data.controller->type == VIR_DOMAIN_CONTROLLER_TYPE_USB &&
        (dev->data.controller->model == VIR_DOMAIN_CONTROLLER_MODEL_USB_QUSB1 ||
         dev->data.controller->model == VIR_DOMAIN_CONTROLLER_MODEL_USB_QUSB2)) {
        virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
                       _("USB controller model type 'qusb1' or 'qusb2' "
                         "is not supported in %s"),
                       virDomainVirtTypeToString(def->virtType));
        goto cleanup;
    }


2979 2980 2981 2982 2983 2984 2985
    /* set the default SCSI controller model for S390 arches */
    if (dev->type == VIR_DOMAIN_DEVICE_CONTROLLER &&
        dev->data.controller->type == VIR_DOMAIN_CONTROLLER_TYPE_SCSI &&
        dev->data.controller->model == -1 &&
        ARCH_IS_S390(def->os.arch))
        dev->data.controller->model = VIR_DOMAIN_CONTROLLER_MODEL_SCSI_VIRTIO_SCSI;

2986 2987
    /* clear auto generated unix socket path for inactive definitions */
    if ((parseFlags & VIR_DOMAIN_DEF_PARSE_INACTIVE) &&
2988 2989
        dev->type == VIR_DOMAIN_DEVICE_CHR)
        qemuDomainChrDefDropDefaultPath(dev->data.chr, driver);
2990

2991 2992 2993 2994 2995 2996 2997 2998 2999 3000
    /* forbid capabilities mode hostdev in this kind of hypervisor */
    if (dev->type == VIR_DOMAIN_DEVICE_HOSTDEV &&
        dev->data.hostdev->mode == VIR_DOMAIN_HOSTDEV_MODE_CAPABILITIES) {
        virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
                       _("hostdev mode 'capabilities' is not "
                         "supported in %s"),
                       virDomainVirtTypeToString(def->virtType));
        goto cleanup;
    }

3001 3002 3003 3004 3005 3006 3007 3008 3009 3010 3011 3012 3013 3014 3015
    if (dev->type == VIR_DOMAIN_DEVICE_VIDEO &&
        dev->data.video->type == VIR_DOMAIN_VIDEO_TYPE_QXL) {
        if (dev->data.video->vgamem) {
            if (dev->data.video->vgamem < 1024) {
                virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                               _("value for 'vgamem' must be at least 1 MiB "
                                 "(1024 KiB)"));
                goto cleanup;
            }
            if (dev->data.video->vgamem != VIR_ROUND_UP_POWER_OF_TWO(dev->data.video->vgamem)) {
                virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                               _("value for 'vgamem' must be power of two"));
                goto cleanup;
            }
        } else {
3016
            dev->data.video->vgamem = QEMU_QXL_VGAMEM_DEFAULT;
3017 3018 3019
        }
    }

3020 3021
    if (dev->type == VIR_DOMAIN_DEVICE_PANIC &&
        dev->data.panic->model == VIR_DOMAIN_PANIC_MODEL_DEFAULT) {
3022
        if (qemuDomainMachineIsPSeries(def))
3023
            dev->data.panic->model = VIR_DOMAIN_PANIC_MODEL_PSERIES;
3024 3025
        else if (ARCH_IS_S390(def->os.arch))
            dev->data.panic->model = VIR_DOMAIN_PANIC_MODEL_S390;
3026 3027 3028 3029
        else
            dev->data.panic->model = VIR_DOMAIN_PANIC_MODEL_ISA;
    }

3030 3031 3032 3033

    if (dev->type == VIR_DOMAIN_DEVICE_CONTROLLER) {
        virDomainControllerDefPtr cont = dev->data.controller;

3034 3035 3036
        if (cont->type == VIR_DOMAIN_CONTROLLER_TYPE_PCI) {
            if (cont->model == VIR_DOMAIN_CONTROLLER_MODEL_PCI_EXPANDER_BUS &&
                !qemuDomainMachineIsI440FX(def)) {
3037 3038 3039 3040 3041
                virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                               _("pci-expander-bus controllers are only supported "
                                 "on 440fx-based machinetypes"));
                goto cleanup;
            }
3042 3043 3044 3045 3046 3047 3048
            if (cont->model == VIR_DOMAIN_CONTROLLER_MODEL_PCIE_EXPANDER_BUS &&
                !qemuDomainMachineIsQ35(def)) {
                virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                               _("pcie-expander-bus controllers are only supported "
                                 "on q35-based machinetypes"));
                goto cleanup;
            }
3049 3050 3051 3052 3053 3054

            /* if a PCI expander bus has a NUMA node set, make sure
             * that NUMA node is configured in the guest <cpu><numa>
             * array. NUMA cell id's in this array are numbered
             * from 0 .. size-1.
             */
3055 3056 3057
            if ((cont->model == VIR_DOMAIN_CONTROLLER_MODEL_PCI_EXPANDER_BUS ||
                 cont->model == VIR_DOMAIN_CONTROLLER_MODEL_PCIE_EXPANDER_BUS) &&
                (int) virDomainNumaGetNodeCount(def->numa)
3058 3059
                <= cont->opts.pciopts.numaNode) {
                virReportError(VIR_ERR_XML_ERROR,
3060
                               _("%s with index %d is "
3061 3062 3063
                                 "configured for a NUMA node (%d) "
                                 "not present in the domain's "
                                 "<cpu><numa> array (%zu)"),
3064
                               virDomainControllerModelPCITypeToString(cont->model),
3065 3066 3067 3068
                               cont->idx, cont->opts.pciopts.numaNode,
                               virDomainNumaGetNodeCount(def->numa));
                goto cleanup;
            }
3069 3070 3071 3072 3073 3074 3075 3076 3077 3078 3079 3080 3081 3082 3083 3084 3085 3086 3087 3088
        } else if (cont->type == VIR_DOMAIN_CONTROLLER_TYPE_USB &&
                   cont->model == -1) {
            /* Pick a suitable default model for the USB controller if none
             * has been selected by the user.
             *
             * We rely on device availability instead of setting the model
             * unconditionally because, for some machine types, there's a
             * chance we will get away with using the legacy USB controller
             * when the relevant device is not available.
             *
             * See qemuBuildControllerDevCommandLine() */
            if (ARCH_IS_PPC64(def->os.arch)) {
                /* Default USB controller for ppc64 is pci-ohci */
                if (virQEMUCapsGet(qemuCaps, QEMU_CAPS_PCI_OHCI))
                    cont->model = VIR_DOMAIN_CONTROLLER_MODEL_USB_PCI_OHCI;
            } else {
                /* Default USB controller for anything else is piix3-uhci */
                if (virQEMUCapsGet(qemuCaps, QEMU_CAPS_PIIX3_USB_UHCI))
                    cont->model = VIR_DOMAIN_CONTROLLER_MODEL_USB_PIIX3_UHCI;
            }
3089 3090 3091
        }
    }

3092 3093 3094 3095
    if (dev->type == VIR_DOMAIN_DEVICE_SHMEM &&
        qemuDomainShmemDefPostParse(dev->data.shmem) < 0)
        goto cleanup;

3096 3097
    ret = 0;

3098
 cleanup:
3099
    virObjectUnref(qemuCaps);
3100 3101
    virObjectUnref(cfg);
    return ret;
3102 3103 3104
}


3105 3106
static int
qemuDomainDefAssignAddresses(virDomainDef *def,
3107
                             virCapsPtr caps,
3108
                             unsigned int parseFlags ATTRIBUTE_UNUSED,
3109
                             void *opaque,
3110
                             void *parseOpaque)
3111 3112
{
    virQEMUDriverPtr driver = opaque;
3113
    virQEMUCapsPtr qemuCaps = parseOpaque;
3114
    int ret = -1;
3115
    bool newDomain = parseFlags & VIR_DOMAIN_DEF_PARSE_ABI_UPDATE;
3116

3117 3118 3119 3120 3121 3122 3123 3124
    if (qemuCaps) {
        virObjectRef(qemuCaps);
    } else {
        if (!(qemuCaps = virQEMUCapsCacheLookup(caps,
                                                driver->qemuCapsCache,
                                                def->emulator)))
            goto cleanup;
    }
3125

3126
    if (qemuDomainAssignAddresses(def, qemuCaps, NULL, newDomain) < 0)
3127 3128 3129 3130 3131 3132 3133 3134 3135
        goto cleanup;

    ret = 0;
 cleanup:
    virObjectUnref(qemuCaps);
    return ret;
}


3136 3137
virDomainDefParserConfig virQEMUDriverDomainDefParserConfig = {
    .devicesPostParseCallback = qemuDomainDeviceDefPostParse,
3138
    .domainPostParseCallback = qemuDomainDefPostParse,
3139
    .assignAddressesCallback = qemuDomainDefAssignAddresses,
3140
    .domainValidateCallback = qemuDomainDefValidate,
3141 3142
    .deviceValidateCallback = qemuDomainDeviceDefValidate,

3143
    .features = VIR_DOMAIN_DEF_FEATURE_MEMORY_HOTPLUG |
3144 3145
                VIR_DOMAIN_DEF_FEATURE_OFFLINE_VCPUPIN |
                VIR_DOMAIN_DEF_FEATURE_INDIVIDUAL_VCPUS,
3146 3147 3148
};


3149
static void
3150
qemuDomainObjSaveJob(virQEMUDriverPtr driver, virDomainObjPtr obj)
3151
{
3152 3153 3154
    virQEMUDriverConfigPtr cfg = virQEMUDriverGetConfig(driver);

    if (virDomainObjIsActive(obj)) {
3155
        if (virDomainSaveStatus(driver->xmlopt, cfg->stateDir, obj, driver->caps) < 0)
3156
            VIR_WARN("Failed to save status on vm %s", obj->def->name);
3157
    }
3158

3159
    virObjectUnref(cfg);
3160 3161
}

J
Jiri Denemark 已提交
3162
void
3163
qemuDomainObjSetJobPhase(virQEMUDriverPtr driver,
J
Jiri Denemark 已提交
3164 3165 3166 3167
                         virDomainObjPtr obj,
                         int phase)
{
    qemuDomainObjPrivatePtr priv = obj->privateData;
3168
    unsigned long long me = virThreadSelfID();
J
Jiri Denemark 已提交
3169 3170 3171 3172

    if (!priv->job.asyncJob)
        return;

3173 3174 3175 3176 3177
    VIR_DEBUG("Setting '%s' phase to '%s'",
              qemuDomainAsyncJobTypeToString(priv->job.asyncJob),
              qemuDomainAsyncJobPhaseToString(priv->job.asyncJob, phase));

    if (priv->job.asyncOwner && me != priv->job.asyncOwner) {
3178
        VIR_WARN("'%s' async job is owned by thread %llu",
3179 3180 3181 3182
                 qemuDomainAsyncJobTypeToString(priv->job.asyncJob),
                 priv->job.asyncOwner);
    }

J
Jiri Denemark 已提交
3183
    priv->job.phase = phase;
3184
    priv->job.asyncOwner = me;
J
Jiri Denemark 已提交
3185 3186 3187
    qemuDomainObjSaveJob(driver, obj);
}

3188
void
3189 3190
qemuDomainObjSetAsyncJobMask(virDomainObjPtr obj,
                             unsigned long long allowedJobs)
3191 3192 3193
{
    qemuDomainObjPrivatePtr priv = obj->privateData;

3194 3195 3196 3197 3198 3199 3200
    if (!priv->job.asyncJob)
        return;

    priv->job.mask = allowedJobs | JOB_MASK(QEMU_JOB_DESTROY);
}

void
3201
qemuDomainObjDiscardAsyncJob(virQEMUDriverPtr driver, virDomainObjPtr obj)
3202 3203 3204 3205 3206 3207
{
    qemuDomainObjPrivatePtr priv = obj->privateData;

    if (priv->job.active == QEMU_JOB_ASYNC_NESTED)
        qemuDomainObjResetJob(priv);
    qemuDomainObjResetAsyncJob(priv);
3208
    qemuDomainObjSaveJob(driver, obj);
3209 3210
}

3211 3212 3213 3214 3215 3216 3217 3218 3219
void
qemuDomainObjReleaseAsyncJob(virDomainObjPtr obj)
{
    qemuDomainObjPrivatePtr priv = obj->privateData;

    VIR_DEBUG("Releasing ownership of '%s' async job",
              qemuDomainAsyncJobTypeToString(priv->job.asyncJob));

    if (priv->job.asyncOwner != virThreadSelfID()) {
3220
        VIR_WARN("'%s' async job is owned by thread %llu",
3221 3222 3223 3224 3225 3226
                 qemuDomainAsyncJobTypeToString(priv->job.asyncJob),
                 priv->job.asyncOwner);
    }
    priv->job.asyncOwner = 0;
}

3227
static bool
3228
qemuDomainNestedJobAllowed(qemuDomainObjPrivatePtr priv, qemuDomainJob job)
3229 3230
{
    return !priv->job.asyncJob || (priv->job.mask & JOB_MASK(job)) != 0;
3231 3232
}

3233
bool
3234
qemuDomainJobAllowed(qemuDomainObjPrivatePtr priv, qemuDomainJob job)
3235 3236 3237 3238
{
    return !priv->job.active && qemuDomainNestedJobAllowed(priv, job);
}

3239 3240 3241
/* Give up waiting for mutex after 30 seconds */
#define QEMU_JOB_WAIT_TIME (1000ull * 30)

3242
/*
3243
 * obj must be locked before calling
3244
 */
3245
static int ATTRIBUTE_NONNULL(1)
3246
qemuDomainObjBeginJobInternal(virQEMUDriverPtr driver,
3247
                              virDomainObjPtr obj,
3248 3249
                              qemuDomainJob job,
                              qemuDomainAsyncJob asyncJob)
3250 3251
{
    qemuDomainObjPrivatePtr priv = obj->privateData;
J
Jiri Denemark 已提交
3252
    unsigned long long now;
3253
    unsigned long long then;
3254
    bool nested = job == QEMU_JOB_ASYNC_NESTED;
J
Jiri Denemark 已提交
3255
    bool async = job == QEMU_JOB_ASYNC;
3256
    virQEMUDriverConfigPtr cfg = virQEMUDriverGetConfig(driver);
3257
    const char *blocker = NULL;
J
Jiri Denemark 已提交
3258
    int ret = -1;
J
Jiri Denemark 已提交
3259 3260
    unsigned long long duration = 0;
    unsigned long long asyncDuration = 0;
J
Jiri Denemark 已提交
3261
    const char *jobStr;
3262

J
Jiri Denemark 已提交
3263 3264 3265 3266 3267 3268 3269 3270 3271
    if (async)
        jobStr = qemuDomainAsyncJobTypeToString(asyncJob);
    else
        jobStr = qemuDomainJobTypeToString(job);

    VIR_DEBUG("Starting %s: %s (vm=%p name=%s, current job=%s async=%s)",
              async ? "async job" : "job", jobStr, obj, obj->def->name,
              qemuDomainJobTypeToString(priv->job.active),
              qemuDomainAsyncJobTypeToString(priv->job.asyncJob));
3272

3273 3274
    if (virTimeMillisNow(&now) < 0) {
        virObjectUnref(cfg);
3275
        return -1;
3276 3277
    }

3278
    priv->jobs_queued++;
J
Jiri Denemark 已提交
3279
    then = now + QEMU_JOB_WAIT_TIME;
3280

3281
 retry:
3282 3283
    if (cfg->maxQueuedJobs &&
        priv->jobs_queued > cfg->maxQueuedJobs) {
3284 3285 3286
        goto error;
    }

3287
    while (!nested && !qemuDomainNestedJobAllowed(priv, job)) {
3288
        VIR_DEBUG("Waiting for async job (vm=%p name=%s)", obj, obj->def->name);
3289
        if (virCondWaitUntil(&priv->job.asyncCond, &obj->parent.lock, then) < 0)
3290 3291 3292
            goto error;
    }

3293
    while (priv->job.active) {
3294
        VIR_DEBUG("Waiting for job (vm=%p name=%s)", obj, obj->def->name);
3295
        if (virCondWaitUntil(&priv->job.cond, &obj->parent.lock, then) < 0)
3296
            goto error;
3297
    }
3298 3299 3300

    /* No job is active but a new async job could have been started while obj
     * was unlocked, so we need to recheck it. */
3301
    if (!nested && !qemuDomainNestedJobAllowed(priv, job))
3302 3303
        goto retry;

3304
    qemuDomainObjResetJob(priv);
3305

J
Jiri Denemark 已提交
3306 3307
    ignore_value(virTimeMillisNow(&now));

3308
    if (job != QEMU_JOB_ASYNC) {
3309
        VIR_DEBUG("Started job: %s (async=%s vm=%p name=%s)",
3310
                   qemuDomainJobTypeToString(job),
3311 3312
                  qemuDomainAsyncJobTypeToString(priv->job.asyncJob),
                  obj, obj->def->name);
3313
        priv->job.active = job;
3314
        priv->job.owner = virThreadSelfID();
3315
        priv->job.ownerAPI = virThreadJobGet();
J
Jiri Denemark 已提交
3316
        priv->job.started = now;
3317
    } else {
3318 3319 3320
        VIR_DEBUG("Started async job: %s (vm=%p name=%s)",
                  qemuDomainAsyncJobTypeToString(asyncJob),
                  obj, obj->def->name);
3321
        qemuDomainObjResetAsyncJob(priv);
J
Jiri Denemark 已提交
3322 3323
        if (VIR_ALLOC(priv->job.current) < 0)
            goto cleanup;
3324
        priv->job.asyncJob = asyncJob;
3325
        priv->job.asyncOwner = virThreadSelfID();
3326
        priv->job.asyncOwnerAPI = virThreadJobGet();
J
Jiri Denemark 已提交
3327
        priv->job.asyncStarted = now;
J
Jiri Denemark 已提交
3328
        priv->job.current->started = now;
3329
    }
3330

3331 3332
    if (qemuDomainTrackJob(job))
        qemuDomainObjSaveJob(driver, obj);
3333

3334
    virObjectUnref(cfg);
3335
    return 0;
3336

3337
 error:
J
Jiri Denemark 已提交
3338 3339 3340 3341 3342 3343
    ignore_value(virTimeMillisNow(&now));
    if (priv->job.active && priv->job.started)
        duration = now - priv->job.started;
    if (priv->job.asyncJob && priv->job.asyncStarted)
        asyncDuration = now - priv->job.asyncStarted;

3344
    VIR_WARN("Cannot start job (%s, %s) for domain %s; "
J
Jiri Denemark 已提交
3345 3346
             "current job is (%s, %s) owned by (%llu %s, %llu %s) "
             "for (%llus, %llus)",
3347 3348 3349 3350 3351
             qemuDomainJobTypeToString(job),
             qemuDomainAsyncJobTypeToString(asyncJob),
             obj->def->name,
             qemuDomainJobTypeToString(priv->job.active),
             qemuDomainAsyncJobTypeToString(priv->job.asyncJob),
3352
             priv->job.owner, NULLSTR(priv->job.ownerAPI),
J
Jiri Denemark 已提交
3353 3354
             priv->job.asyncOwner, NULLSTR(priv->job.asyncOwnerAPI),
             duration / 1000, asyncDuration / 1000);
3355 3356 3357 3358 3359

    if (nested || qemuDomainNestedJobAllowed(priv, job))
        blocker = priv->job.ownerAPI;
    else
        blocker = priv->job.asyncOwnerAPI;
3360

3361 3362
    ret = -1;
    if (errno == ETIMEDOUT) {
3363 3364 3365 3366 3367 3368 3369 3370
        if (blocker) {
            virReportError(VIR_ERR_OPERATION_TIMEOUT,
                           _("cannot acquire state change lock (held by %s)"),
                           blocker);
        } else {
            virReportError(VIR_ERR_OPERATION_TIMEOUT, "%s",
                           _("cannot acquire state change lock"));
        }
3371 3372 3373
        ret = -2;
    } else if (cfg->maxQueuedJobs &&
               priv->jobs_queued > cfg->maxQueuedJobs) {
3374 3375 3376 3377 3378 3379 3380 3381 3382 3383
        if (blocker) {
            virReportError(VIR_ERR_OPERATION_FAILED,
                           _("cannot acquire state change lock (held by %s) "
                             "due to max_queued limit"),
                           blocker);
        } else {
            virReportError(VIR_ERR_OPERATION_FAILED, "%s",
                           _("cannot acquire state change lock "
                             "due to max_queued limit"));
        }
3384 3385
        ret = -2;
    } else {
3386
        virReportSystemError(errno, "%s", _("cannot acquire job mutex"));
3387
    }
J
Jiri Denemark 已提交
3388 3389

 cleanup:
3390
    priv->jobs_queued--;
3391
    virObjectUnref(cfg);
3392
    return ret;
3393 3394 3395
}

/*
3396
 * obj must be locked before calling
3397 3398 3399 3400
 *
 * This must be called by anything that will change the VM state
 * in any way, or anything that will use the QEMU monitor.
 *
3401
 * Successful calls must be followed by EndJob eventually
3402
 */
3403
int qemuDomainObjBeginJob(virQEMUDriverPtr driver,
3404
                          virDomainObjPtr obj,
3405
                          qemuDomainJob job)
3406
{
3407 3408 3409 3410 3411
    if (qemuDomainObjBeginJobInternal(driver, obj, job,
                                      QEMU_ASYNC_JOB_NONE) < 0)
        return -1;
    else
        return 0;
3412 3413
}

3414
int qemuDomainObjBeginAsyncJob(virQEMUDriverPtr driver,
3415
                               virDomainObjPtr obj,
3416
                               qemuDomainAsyncJob asyncJob)
3417
{
3418 3419 3420 3421 3422
    if (qemuDomainObjBeginJobInternal(driver, obj, QEMU_JOB_ASYNC,
                                      asyncJob) < 0)
        return -1;
    else
        return 0;
3423 3424
}

3425
int
3426 3427
qemuDomainObjBeginNestedJob(virQEMUDriverPtr driver,
                            virDomainObjPtr obj,
3428
                            qemuDomainAsyncJob asyncJob)
3429 3430 3431 3432 3433 3434 3435 3436 3437 3438
{
    qemuDomainObjPrivatePtr priv = obj->privateData;

    if (asyncJob != priv->job.asyncJob) {
        virReportError(VIR_ERR_INTERNAL_ERROR,
                       _("unexpected async job %d"), asyncJob);
        return -1;
    }

    if (priv->job.asyncOwner != virThreadSelfID()) {
3439
        VIR_WARN("This thread doesn't seem to be the async job owner: %llu",
3440 3441 3442 3443 3444 3445 3446 3447
                 priv->job.asyncOwner);
    }

    return qemuDomainObjBeginJobInternal(driver, obj,
                                         QEMU_JOB_ASYNC_NESTED,
                                         QEMU_ASYNC_JOB_NONE);
}

3448

3449
/*
3450
 * obj must be locked and have a reference before calling
3451 3452 3453 3454
 *
 * To be called after completing the work associated with the
 * earlier qemuDomainBeginJob() call
 */
3455 3456
void
qemuDomainObjEndJob(virQEMUDriverPtr driver, virDomainObjPtr obj)
3457 3458
{
    qemuDomainObjPrivatePtr priv = obj->privateData;
3459
    qemuDomainJob job = priv->job.active;
3460

3461 3462
    priv->jobs_queued--;

3463
    VIR_DEBUG("Stopping job: %s (async=%s vm=%p name=%s)",
3464
              qemuDomainJobTypeToString(job),
3465 3466
              qemuDomainAsyncJobTypeToString(priv->job.asyncJob),
              obj, obj->def->name);
3467

3468
    qemuDomainObjResetJob(priv);
3469 3470
    if (qemuDomainTrackJob(job))
        qemuDomainObjSaveJob(driver, obj);
3471
    virCondSignal(&priv->job.cond);
3472 3473
}

3474
void
3475
qemuDomainObjEndAsyncJob(virQEMUDriverPtr driver, virDomainObjPtr obj)
3476 3477
{
    qemuDomainObjPrivatePtr priv = obj->privateData;
3478

3479 3480
    priv->jobs_queued--;

3481 3482 3483
    VIR_DEBUG("Stopping async job: %s (vm=%p name=%s)",
              qemuDomainAsyncJobTypeToString(priv->job.asyncJob),
              obj, obj->def->name);
3484

3485
    qemuDomainObjResetAsyncJob(priv);
3486
    qemuDomainObjSaveJob(driver, obj);
3487 3488 3489
    virCondBroadcast(&priv->job.asyncCond);
}

3490 3491 3492 3493 3494
void
qemuDomainObjAbortAsyncJob(virDomainObjPtr obj)
{
    qemuDomainObjPrivatePtr priv = obj->privateData;

3495 3496 3497
    VIR_DEBUG("Requesting abort of async job: %s (vm=%p name=%s)",
              qemuDomainAsyncJobTypeToString(priv->job.asyncJob),
              obj, obj->def->name);
3498

3499 3500
    priv->job.abortJob = true;
    virDomainObjBroadcast(obj);
3501 3502
}

3503 3504 3505 3506 3507 3508 3509 3510 3511
/*
 * obj must be locked before calling
 *
 * To be called immediately before any QEMU monitor API call
 * Must have already either called qemuDomainObjBeginJob() and checked
 * that the VM is still active; may not be used for nested async jobs.
 *
 * To be followed with qemuDomainObjExitMonitor() once complete
 */
3512
static int
3513
qemuDomainObjEnterMonitorInternal(virQEMUDriverPtr driver,
3514
                                  virDomainObjPtr obj,
3515
                                  qemuDomainAsyncJob asyncJob)
3516 3517 3518
{
    qemuDomainObjPrivatePtr priv = obj->privateData;

3519
    if (asyncJob != QEMU_ASYNC_JOB_NONE) {
3520 3521 3522
        int ret;
        if ((ret = qemuDomainObjBeginNestedJob(driver, obj, asyncJob)) < 0)
            return ret;
3523
        if (!virDomainObjIsActive(obj)) {
3524 3525
            virReportError(VIR_ERR_OPERATION_FAILED, "%s",
                           _("domain is no longer running"));
3526
            qemuDomainObjEndJob(driver, obj);
3527 3528
            return -1;
        }
3529 3530 3531
    } else if (priv->job.asyncOwner == virThreadSelfID()) {
        VIR_WARN("This thread seems to be the async job owner; entering"
                 " monitor without asking for a nested job is dangerous");
3532 3533
    }

3534 3535
    VIR_DEBUG("Entering monitor (mon=%p vm=%p name=%s)",
              priv->mon, obj, obj->def->name);
3536
    virObjectLock(priv->mon);
3537
    virObjectRef(priv->mon);
3538
    ignore_value(virTimeMillisNow(&priv->monStart));
3539
    virObjectUnlock(obj);
3540 3541

    return 0;
3542 3543
}

3544
static void ATTRIBUTE_NONNULL(1)
3545
qemuDomainObjExitMonitorInternal(virQEMUDriverPtr driver,
3546
                                 virDomainObjPtr obj)
3547 3548
{
    qemuDomainObjPrivatePtr priv = obj->privateData;
3549
    bool hasRefs;
3550

3551
    hasRefs = virObjectUnref(priv->mon);
3552

3553
    if (hasRefs)
3554
        virObjectUnlock(priv->mon);
3555

3556
    virObjectLock(obj);
3557 3558
    VIR_DEBUG("Exited monitor (mon=%p vm=%p name=%s)",
              priv->mon, obj, obj->def->name);
3559

3560
    priv->monStart = 0;
3561
    if (!hasRefs)
3562
        priv->mon = NULL;
3563

J
Jiri Denemark 已提交
3564 3565
    if (priv->job.active == QEMU_JOB_ASYNC_NESTED)
        qemuDomainObjEndJob(driver, obj);
3566 3567
}

3568
void qemuDomainObjEnterMonitor(virQEMUDriverPtr driver,
3569
                               virDomainObjPtr obj)
3570
{
3571
    ignore_value(qemuDomainObjEnterMonitorInternal(driver, obj,
3572
                                                   QEMU_ASYNC_JOB_NONE));
3573 3574
}

3575
/* obj must NOT be locked before calling
3576 3577
 *
 * Should be paired with an earlier qemuDomainObjEnterMonitor() call
3578 3579 3580 3581 3582 3583
 *
 * Returns -1 if the domain is no longer alive after exiting the monitor.
 * In that case, the caller should be careful when using obj's data,
 * e.g. the live definition in vm->def has been freed by qemuProcessStop
 * and replaced by the persistent definition, so pointers stolen
 * from the live definition could no longer be valid.
3584
 */
3585 3586
int qemuDomainObjExitMonitor(virQEMUDriverPtr driver,
                             virDomainObjPtr obj)
3587
{
3588
    qemuDomainObjExitMonitorInternal(driver, obj);
3589
    if (!virDomainObjIsActive(obj)) {
3590 3591 3592
        if (!virGetLastError())
            virReportError(VIR_ERR_OPERATION_FAILED, "%s",
                           _("domain is no longer running"));
3593 3594 3595
        return -1;
    }
    return 0;
3596
}
3597 3598

/*
3599
 * obj must be locked before calling
3600 3601
 *
 * To be called immediately before any QEMU monitor API call.
3602
 * Must have already either called qemuDomainObjBeginJob()
3603 3604 3605 3606 3607
 * and checked that the VM is still active, with asyncJob of
 * QEMU_ASYNC_JOB_NONE; or already called qemuDomainObjBeginAsyncJob,
 * with the same asyncJob.
 *
 * Returns 0 if job was started, in which case this must be followed with
3608 3609 3610
 * qemuDomainObjExitMonitor(); -2 if waiting for the nested job times out;
 * or -1 if the job could not be started (probably because the vm exited
 * in the meantime).
3611 3612
 */
int
3613
qemuDomainObjEnterMonitorAsync(virQEMUDriverPtr driver,
3614
                               virDomainObjPtr obj,
3615
                               qemuDomainAsyncJob asyncJob)
3616
{
3617
    return qemuDomainObjEnterMonitorInternal(driver, obj, asyncJob);
3618 3619
}

D
Daniel P. Berrange 已提交
3620

3621 3622 3623 3624 3625 3626 3627 3628 3629 3630 3631 3632
/**
 * qemuDomainGetAgent:
 * @vm: domain object
 *
 * Returns the agent pointer of @vm;
 */
qemuAgentPtr
qemuDomainGetAgent(virDomainObjPtr vm)
{
    return (((qemuDomainObjPrivatePtr)(vm->privateData))->agent);
}

D
Daniel P. Berrange 已提交
3633

3634 3635 3636 3637 3638 3639 3640 3641 3642 3643 3644
/*
 * obj must be locked before calling
 *
 * To be called immediately before any QEMU agent API call.
 * Must have already called qemuDomainObjBeginJob() and checked
 * that the VM is still active.
 *
 * To be followed with qemuDomainObjExitAgent() once complete
 */
void
qemuDomainObjEnterAgent(virDomainObjPtr obj)
D
Daniel P. Berrange 已提交
3645 3646 3647
{
    qemuDomainObjPrivatePtr priv = obj->privateData;

3648 3649
    VIR_DEBUG("Entering agent (agent=%p vm=%p name=%s)",
              priv->agent, obj, obj->def->name);
3650
    virObjectLock(priv->agent);
3651
    virObjectRef(priv->agent);
D
Daniel P. Berrange 已提交
3652
    ignore_value(virTimeMillisNow(&priv->agentStart));
3653
    virObjectUnlock(obj);
D
Daniel P. Berrange 已提交
3654 3655
}

3656 3657 3658 3659 3660 3661 3662

/* obj must NOT be locked before calling
 *
 * Should be paired with an earlier qemuDomainObjEnterAgent() call
 */
void
qemuDomainObjExitAgent(virDomainObjPtr obj)
D
Daniel P. Berrange 已提交
3663 3664
{
    qemuDomainObjPrivatePtr priv = obj->privateData;
3665
    bool hasRefs;
D
Daniel P. Berrange 已提交
3666

3667
    hasRefs = virObjectUnref(priv->agent);
D
Daniel P. Berrange 已提交
3668

3669
    if (hasRefs)
3670
        virObjectUnlock(priv->agent);
D
Daniel P. Berrange 已提交
3671

3672
    virObjectLock(obj);
3673 3674
    VIR_DEBUG("Exited agent (agent=%p vm=%p name=%s)",
              priv->agent, obj, obj->def->name);
D
Daniel P. Berrange 已提交
3675 3676

    priv->agentStart = 0;
3677
    if (!hasRefs)
D
Daniel P. Berrange 已提交
3678 3679 3680
        priv->agent = NULL;
}

3681
void qemuDomainObjEnterRemote(virDomainObjPtr obj)
3682
{
3683 3684
    VIR_DEBUG("Entering remote (vm=%p name=%s)",
              obj, obj->def->name);
3685
    virObjectUnlock(obj);
3686 3687
}

3688
void qemuDomainObjExitRemote(virDomainObjPtr obj)
3689
{
3690
    virObjectLock(obj);
3691 3692
    VIR_DEBUG("Exited remote (vm=%p name=%s)",
              obj, obj->def->name);
3693
}
3694 3695


3696 3697 3698 3699 3700 3701 3702 3703
virDomainDefPtr
qemuDomainDefCopy(virQEMUDriverPtr driver,
                  virDomainDefPtr src,
                  unsigned int flags)
{
    virBuffer buf = VIR_BUFFER_INITIALIZER;
    virDomainDefPtr ret = NULL;
    virCapsPtr caps = NULL;
3704
    char *xml = NULL;
3705 3706 3707 3708 3709 3710 3711 3712 3713

    if (qemuDomainDefFormatBuf(driver, src, flags, &buf) < 0)
        goto cleanup;

    xml = virBufferContentAndReset(&buf);

    if (!(caps = virQEMUDriverGetCapabilities(driver, false)))
        goto cleanup;

3714
    if (!(ret = virDomainDefParseString(xml, caps, driver->xmlopt, NULL,
3715
                                        VIR_DOMAIN_DEF_PARSE_INACTIVE |
3716
                                        VIR_DOMAIN_DEF_PARSE_SKIP_VALIDATE)))
3717 3718
        goto cleanup;

3719
 cleanup:
3720 3721 3722 3723 3724
    VIR_FREE(xml);
    virObjectUnref(caps);
    return ret;
}

3725
int
3726
qemuDomainDefFormatBuf(virQEMUDriverPtr driver,
3727 3728 3729
                       virDomainDefPtr def,
                       unsigned int flags,
                       virBuffer *buf)
3730
{
3731
    int ret = -1;
3732
    virDomainDefPtr copy = NULL;
3733 3734 3735 3736
    virCapsPtr caps = NULL;

    if (!(caps = virQEMUDriverGetCapabilities(driver, false)))
        goto cleanup;
3737

3738 3739 3740
    if (!(flags & (VIR_DOMAIN_XML_UPDATE_CPU | VIR_DOMAIN_XML_MIGRATABLE)))
        goto format;

3741
    if (!(copy = virDomainDefCopy(def, caps, driver->xmlopt, NULL,
3742 3743 3744 3745 3746
                                  flags & VIR_DOMAIN_XML_MIGRATABLE)))
        goto cleanup;

    def = copy;

3747
    /* Update guest CPU requirements according to host CPU */
3748
    if ((flags & VIR_DOMAIN_XML_UPDATE_CPU) &&
3749
        def->cpu &&
3750 3751
        (def->cpu->mode != VIR_CPU_MODE_CUSTOM ||
         def->cpu->model)) {
J
Jiri Denemark 已提交
3752
        if (virCPUUpdate(def->os.arch, def->cpu, caps->host.cpu) < 0)
3753 3754 3755
            goto cleanup;
    }

3756
    if ((flags & VIR_DOMAIN_XML_MIGRATABLE)) {
3757
        size_t i;
3758
        int toremove = 0;
3759
        virDomainControllerDefPtr usb = NULL, pci = NULL;
3760 3761 3762 3763 3764 3765 3766 3767 3768 3769 3770 3771 3772 3773 3774

        /* If only the default USB controller is present, we can remove it
         * and make the XML compatible with older versions of libvirt which
         * didn't support USB controllers in the XML but always added the
         * default one to qemu anyway.
         */
        for (i = 0; i < def->ncontrollers; i++) {
            if (def->controllers[i]->type == VIR_DOMAIN_CONTROLLER_TYPE_USB) {
                if (usb) {
                    usb = NULL;
                    break;
                }
                usb = def->controllers[i];
            }
        }
3775 3776 3777 3778 3779 3780 3781 3782 3783

        /* In order to maintain compatibility with version of libvirt that
         * didn't support <controller type='usb'/> (<= 0.9.4), we need to
         * drop the default USB controller, ie. a USB controller at index
         * zero with no model or with the default piix3-ohci model.
         *
         * However, we only need to do so for x86 i440fx machine types,
         * because other architectures and machine types were introduced
         * when libvirt already supported <controller type='usb'/>.
3784
         */
3785
        if (ARCH_IS_X86(def->os.arch) && qemuDomainMachineIsI440FX(def) &&
3786 3787 3788
            usb && usb->idx == 0 &&
            (usb->model == -1 ||
             usb->model == VIR_DOMAIN_CONTROLLER_MODEL_USB_PIIX3_UHCI)) {
3789 3790
            VIR_DEBUG("Removing default USB controller from domain '%s'"
                      " for migration compatibility", def->name);
3791
            toremove++;
3792 3793 3794 3795 3796 3797 3798 3799 3800 3801 3802 3803 3804 3805 3806 3807 3808 3809
        } else {
            usb = NULL;
        }

        /* Remove the default PCI controller if there is only one present
         * and its model is pci-root */
        for (i = 0; i < def->ncontrollers; i++) {
            if (def->controllers[i]->type == VIR_DOMAIN_CONTROLLER_TYPE_PCI) {
                if (pci) {
                    pci = NULL;
                    break;
                }
                pci = def->controllers[i];
            }
        }

        if (pci && pci->idx == 0 &&
            pci->model == VIR_DOMAIN_CONTROLLER_MODEL_PCI_ROOT) {
L
Laine Stump 已提交
3810
            VIR_DEBUG("Removing default pci-root from domain '%s'"
3811
                      " for migration compatibility", def->name);
3812
            toremove++;
3813 3814 3815 3816
        } else {
            pci = NULL;
        }

3817
        if (toremove) {
3818 3819 3820
            virDomainControllerDefPtr *controllers = def->controllers;
            int ncontrollers = def->ncontrollers;

3821
            if (VIR_ALLOC_N(def->controllers, ncontrollers - toremove) < 0) {
3822
                def->controllers = controllers;
3823 3824 3825 3826 3827
                goto cleanup;
            }

            def->ncontrollers = 0;
            for (i = 0; i < ncontrollers; i++) {
3828
                if (controllers[i] != usb && controllers[i] != pci)
3829 3830
                    def->controllers[def->ncontrollers++] = controllers[i];
            }
3831 3832 3833 3834

            VIR_FREE(controllers);
            virDomainControllerDefFree(pci);
            virDomainControllerDefFree(usb);
3835
        }
3836

3837 3838 3839 3840 3841 3842 3843 3844 3845
        /* Remove the panic device for selected models if present */
        for (i = 0; i < def->npanics; i++) {
            if (def->panics[i]->model == VIR_DOMAIN_PANIC_MODEL_S390 ||
                def->panics[i]->model == VIR_DOMAIN_PANIC_MODEL_PSERIES) {
                VIR_DELETE_ELEMENT(def->panics, i, def->npanics);
                break;
            }
        }

3846 3847
        for (i = 0; i < def->nchannels; i++)
            qemuDomainChrDefDropDefaultPath(def->channels[i], driver);
3848 3849
    }

3850 3851
 format:
    ret = virDomainDefFormatInternal(def, caps,
3852 3853
                                     virDomainDefFormatConvertXMLFlags(flags),
                                     buf);
3854

3855
 cleanup:
3856
    virDomainDefFree(copy);
3857
    virObjectUnref(caps);
3858 3859
    return ret;
}
3860

3861
char *qemuDomainDefFormatXML(virQEMUDriverPtr driver,
3862
                             virDomainDefPtr def,
3863
                             unsigned int flags)
3864 3865 3866
{
    virBuffer buf = VIR_BUFFER_INITIALIZER;

3867
    if (qemuDomainDefFormatBuf(driver, def, flags, &buf) < 0) {
3868 3869 3870 3871 3872 3873 3874 3875 3876 3877 3878 3879 3880
        virBufferFreeAndReset(&buf);
        return NULL;
    }

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

    return virBufferContentAndReset(&buf);
}

3881
char *qemuDomainFormatXML(virQEMUDriverPtr driver,
3882
                          virDomainObjPtr vm,
3883
                          unsigned int flags)
3884 3885 3886
{
    virDomainDefPtr def;

3887
    if ((flags & VIR_DOMAIN_XML_INACTIVE) && vm->newDef) {
3888
        def = vm->newDef;
3889
    } else {
3890
        def = vm->def;
3891 3892 3893
        if (virDomainObjIsActive(vm))
            flags &= ~VIR_DOMAIN_XML_UPDATE_CPU;
    }
3894

3895
    return qemuDomainDefFormatXML(driver, def, flags);
3896 3897
}

3898
char *
3899
qemuDomainDefFormatLive(virQEMUDriverPtr driver,
3900
                        virDomainDefPtr def,
3901 3902
                        bool inactive,
                        bool compatible)
3903 3904 3905 3906 3907
{
    unsigned int flags = QEMU_DOMAIN_FORMAT_LIVE_FLAGS;

    if (inactive)
        flags |= VIR_DOMAIN_XML_INACTIVE;
3908 3909
    if (compatible)
        flags |= VIR_DOMAIN_XML_MIGRATABLE;
3910

3911
    return qemuDomainDefFormatXML(driver, def, flags);
3912 3913
}

3914

3915
void qemuDomainObjTaint(virQEMUDriverPtr driver,
3916
                        virDomainObjPtr obj,
3917
                        virDomainTaintFlags taint,
3918
                        qemuDomainLogContextPtr logCtxt)
3919
{
3920
    virErrorPtr orig_err = NULL;
3921
    bool closeLog = false;
3922

3923 3924 3925 3926 3927 3928 3929 3930 3931
    if (virDomainObjTaint(obj, taint)) {
        char uuidstr[VIR_UUID_STRING_BUFLEN];
        virUUIDFormat(obj->def->uuid, uuidstr);

        VIR_WARN("Domain id=%d name='%s' uuid=%s is tainted: %s",
                 obj->def->id,
                 obj->def->name,
                 uuidstr,
                 virDomainTaintTypeToString(taint));
3932 3933 3934 3935 3936

        /* We don't care about errors logging taint info, so
         * preserve original error, and clear any error that
         * is raised */
        orig_err = virSaveLastError();
3937 3938 3939 3940 3941 3942 3943 3944 3945 3946 3947 3948 3949 3950 3951 3952 3953 3954
        if (logCtxt == NULL) {
            logCtxt = qemuDomainLogContextNew(driver, obj,
                                              QEMU_DOMAIN_LOG_CONTEXT_MODE_ATTACH);
            if (!logCtxt) {
                if (orig_err) {
                    virSetError(orig_err);
                    virFreeError(orig_err);
                }
                VIR_WARN("Unable to open domainlog");
                return;
            }
            closeLog = true;
        }

        if (qemuDomainLogContextWrite(logCtxt,
                                      "Domain id=%d is tainted: %s\n",
                                      obj->def->id,
                                      virDomainTaintTypeToString(taint)) < 0)
3955
            virResetLastError();
3956 3957
        if (closeLog)
            qemuDomainLogContextFree(logCtxt);
3958 3959 3960 3961
        if (orig_err) {
            virSetError(orig_err);
            virFreeError(orig_err);
        }
3962 3963 3964 3965
    }
}


3966
void qemuDomainObjCheckTaint(virQEMUDriverPtr driver,
3967
                             virDomainObjPtr obj,
3968
                             qemuDomainLogContextPtr logCtxt)
3969
{
3970
    size_t i;
3971
    virQEMUDriverConfigPtr cfg = virQEMUDriverGetConfig(driver);
3972
    qemuDomainObjPrivatePtr priv = obj->privateData;
3973

3974
    if (virQEMUDriverIsPrivileged(driver) &&
3975 3976 3977
        (!cfg->clearEmulatorCapabilities ||
         cfg->user == 0 ||
         cfg->group == 0))
3978
        qemuDomainObjTaint(driver, obj, VIR_DOMAIN_TAINT_HIGH_PRIVILEGES, logCtxt);
3979

3980
    if (priv->hookRun)
3981
        qemuDomainObjTaint(driver, obj, VIR_DOMAIN_TAINT_HOOK, logCtxt);
3982

3983 3984 3985
    if (obj->def->namespaceData) {
        qemuDomainCmdlineDefPtr qemucmd = obj->def->namespaceData;
        if (qemucmd->num_args || qemucmd->num_env)
3986
            qemuDomainObjTaint(driver, obj, VIR_DOMAIN_TAINT_CUSTOM_ARGV, logCtxt);
3987 3988
    }

3989
    if (obj->def->cpu && obj->def->cpu->mode == VIR_CPU_MODE_HOST_PASSTHROUGH)
3990
        qemuDomainObjTaint(driver, obj, VIR_DOMAIN_TAINT_HOST_CPU, logCtxt);
3991

3992
    for (i = 0; i < obj->def->ndisks; i++)
3993
        qemuDomainObjCheckDiskTaint(driver, obj, obj->def->disks[i], logCtxt);
3994

3995 3996
    for (i = 0; i < obj->def->nhostdevs; i++)
        qemuDomainObjCheckHostdevTaint(driver, obj, obj->def->hostdevs[i],
3997
                                       logCtxt);
3998

3999
    for (i = 0; i < obj->def->nnets; i++)
4000
        qemuDomainObjCheckNetTaint(driver, obj, obj->def->nets[i], logCtxt);
4001

4002
    if (obj->def->os.dtb)
4003
        qemuDomainObjTaint(driver, obj, VIR_DOMAIN_TAINT_CUSTOM_DTB, logCtxt);
4004

4005
    virObjectUnref(cfg);
4006 4007 4008
}


4009
void qemuDomainObjCheckDiskTaint(virQEMUDriverPtr driver,
4010
                                 virDomainObjPtr obj,
4011
                                 virDomainDiskDefPtr disk,
4012
                                 qemuDomainLogContextPtr logCtxt)
4013
{
4014
    virQEMUDriverConfigPtr cfg = virQEMUDriverGetConfig(driver);
4015
    int format = virDomainDiskGetFormat(disk);
4016

4017
    if ((!format || format == VIR_STORAGE_FILE_AUTO) &&
4018
        cfg->allowDiskFormatProbing)
4019
        qemuDomainObjTaint(driver, obj, VIR_DOMAIN_TAINT_DISK_PROBING, logCtxt);
4020

4021 4022
    if (disk->rawio == VIR_TRISTATE_BOOL_YES)
        qemuDomainObjTaint(driver, obj, VIR_DOMAIN_TAINT_HIGH_PRIVILEGES,
4023
                           logCtxt);
4024

4025 4026 4027 4028
    if (disk->device == VIR_DOMAIN_DISK_DEVICE_CDROM &&
        virStorageSourceGetActualType(disk->src) == VIR_STORAGE_TYPE_BLOCK &&
        disk->src->path)
        qemuDomainObjTaint(driver, obj, VIR_DOMAIN_TAINT_CDROM_PASSTHROUGH,
4029
                           logCtxt);
4030

4031
    virObjectUnref(cfg);
4032 4033 4034
}


4035 4036 4037
void qemuDomainObjCheckHostdevTaint(virQEMUDriverPtr driver,
                                    virDomainObjPtr obj,
                                    virDomainHostdevDefPtr hostdev,
4038
                                    qemuDomainLogContextPtr logCtxt)
4039 4040 4041 4042 4043 4044
{
    virDomainHostdevSubsysSCSIPtr scsisrc = &hostdev->source.subsys.u.scsi;

    if (hostdev->source.subsys.type == VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_SCSI &&
        scsisrc->rawio == VIR_TRISTATE_BOOL_YES)
            qemuDomainObjTaint(driver, obj, VIR_DOMAIN_TAINT_HIGH_PRIVILEGES,
4045
                               logCtxt);
4046 4047 4048
}


4049
void qemuDomainObjCheckNetTaint(virQEMUDriverPtr driver,
4050
                                virDomainObjPtr obj,
4051
                                virDomainNetDefPtr net,
4052
                                qemuDomainLogContextPtr logCtxt)
4053
{
4054 4055 4056 4057 4058 4059
    /* script is only useful for NET_TYPE_ETHERNET (qemu) and
     * NET_TYPE_BRIDGE (xen), but could be (incorrectly) specified for
     * any interface type. In any case, it's adding user sauce into
     * the soup, so it should taint the domain.
     */
    if (net->script != NULL)
4060
        qemuDomainObjTaint(driver, obj, VIR_DOMAIN_TAINT_SHELL_SCRIPTS, logCtxt);
4061
}
4062 4063


4064 4065 4066 4067 4068 4069 4070 4071 4072 4073
qemuDomainLogContextPtr qemuDomainLogContextNew(virQEMUDriverPtr driver,
                                                virDomainObjPtr vm,
                                                qemuDomainLogContextMode mode)
{
    virQEMUDriverConfigPtr cfg = virQEMUDriverGetConfig(driver);
    qemuDomainLogContextPtr ctxt = NULL;

    if (VIR_ALLOC(ctxt) < 0)
        goto error;

4074
    VIR_DEBUG("Context new %p stdioLogD=%d", ctxt, cfg->stdioLogD);
4075 4076 4077 4078
    ctxt->writefd = -1;
    ctxt->readfd = -1;
    virAtomicIntSet(&ctxt->refs, 1);

4079 4080 4081
    if (virAsprintf(&ctxt->path, "%s/%s.log", cfg->logDir, vm->def->name) < 0)
        goto error;

4082 4083 4084 4085
    if (cfg->stdioLogD) {
        ctxt->manager = virLogManagerNew(virQEMUDriverIsPrivileged(driver));
        if (!ctxt->manager)
            goto error;
4086

4087 4088 4089 4090
        ctxt->writefd = virLogManagerDomainOpenLogFile(ctxt->manager,
                                                       "qemu",
                                                       vm->def->uuid,
                                                       vm->def->name,
4091
                                                       ctxt->path,
4092 4093 4094 4095 4096 4097
                                                       0,
                                                       &ctxt->inode,
                                                       &ctxt->pos);
        if (ctxt->writefd < 0)
            goto error;
    } else {
4098
        if ((ctxt->writefd = open(ctxt->path, O_WRONLY | O_CREAT | O_APPEND, S_IRUSR | S_IWUSR)) < 0) {
4099
            virReportSystemError(errno, _("failed to create logfile %s"),
4100
                                 ctxt->path);
4101 4102
            goto error;
        }
4103
        if (virSetCloseExec(ctxt->writefd) < 0) {
4104
            virReportSystemError(errno, _("failed to set close-on-exec flag on %s"),
4105
                                 ctxt->path);
4106 4107 4108
            goto error;
        }

4109 4110 4111 4112 4113 4114 4115
        /* For unprivileged startup we must truncate the file since
         * we can't rely on logrotate. We don't use O_TRUNC since
         * it is better for SELinux policy if we truncate afterwards */
        if (mode == QEMU_DOMAIN_LOG_CONTEXT_MODE_START &&
            !virQEMUDriverIsPrivileged(driver) &&
            ftruncate(ctxt->writefd, 0) < 0) {
            virReportSystemError(errno, _("failed to truncate %s"),
4116
                                 ctxt->path);
4117 4118 4119 4120
            goto error;
        }

        if (mode == QEMU_DOMAIN_LOG_CONTEXT_MODE_START) {
4121
            if ((ctxt->readfd = open(ctxt->path, O_RDONLY, S_IRUSR | S_IWUSR)) < 0) {
4122
                virReportSystemError(errno, _("failed to open logfile %s"),
4123
                                     ctxt->path);
4124 4125 4126 4127
                goto error;
            }
            if (virSetCloseExec(ctxt->readfd) < 0) {
                virReportSystemError(errno, _("failed to set close-on-exec flag on %s"),
4128
                                     ctxt->path);
4129 4130 4131 4132
                goto error;
            }
        }

4133 4134
        if ((ctxt->pos = lseek(ctxt->writefd, 0, SEEK_END)) < 0) {
            virReportSystemError(errno, _("failed to seek in log file %s"),
4135
                                 ctxt->path);
4136 4137 4138 4139
            goto error;
        }
    }

4140
 cleanup:
4141 4142 4143 4144 4145
    virObjectUnref(cfg);
    return ctxt;

 error:
    qemuDomainLogContextFree(ctxt);
4146 4147
    ctxt = NULL;
    goto cleanup;
4148 4149 4150 4151 4152 4153 4154 4155 4156 4157 4158 4159 4160 4161
}


int qemuDomainLogContextWrite(qemuDomainLogContextPtr ctxt,
                              const char *fmt, ...)
{
    va_list argptr;
    char *message = NULL;
    int ret = -1;

    va_start(argptr, fmt);

    if (virVasprintf(&message, fmt, argptr) < 0)
        goto cleanup;
4162 4163
    if (!ctxt->manager &&
        lseek(ctxt->writefd, 0, SEEK_END) < 0) {
4164
        virReportSystemError(errno, "%s",
4165
                             _("Unable to seek to end of domain logfile"));
4166 4167 4168 4169 4170 4171 4172 4173 4174 4175 4176 4177 4178 4179 4180 4181 4182 4183 4184 4185
        goto cleanup;
    }
    if (safewrite(ctxt->writefd, message, strlen(message)) < 0) {
        virReportSystemError(errno, "%s",
                             _("Unable to write to domain logfile"));
        goto cleanup;
    }

    ret = 0;

 cleanup:
    va_end(argptr);
    VIR_FREE(message);
    return ret;
}


ssize_t qemuDomainLogContextRead(qemuDomainLogContextPtr ctxt,
                                 char **msg)
{
4186 4187 4188 4189
    VIR_DEBUG("Context read %p manager=%p inode=%llu pos=%llu",
              ctxt, ctxt->manager,
              (unsigned long long)ctxt->inode,
              (unsigned long long)ctxt->pos);
4190
    char *buf;
4191 4192 4193
    size_t buflen;
    if (ctxt->manager) {
        buf = virLogManagerDomainReadLogFile(ctxt->manager,
4194
                                             ctxt->path,
4195 4196 4197 4198 4199 4200 4201 4202 4203
                                             ctxt->inode,
                                             ctxt->pos,
                                             1024 * 128,
                                             0);
        if (!buf)
            return -1;
        buflen = strlen(buf);
    } else {
        ssize_t got;
4204

4205
        buflen = 1024 * 128;
4206

4207 4208
        /* Best effort jump to start of messages */
        ignore_value(lseek(ctxt->readfd, ctxt->pos, SEEK_SET));
4209

4210 4211
        if (VIR_ALLOC_N(buf, buflen) < 0)
            return -1;
4212

4213 4214 4215 4216 4217 4218 4219 4220 4221 4222 4223 4224 4225
        got = saferead(ctxt->readfd, buf, buflen - 1);
        if (got < 0) {
            VIR_FREE(buf);
            virReportSystemError(errno, "%s",
                                 _("Unable to read from log file"));
            return -1;
        }

        buf[got] = '\0';

        ignore_value(VIR_REALLOC_N_QUIET(buf, got + 1));
        buflen = got;
    }
4226 4227 4228

    *msg = buf;

4229
    return buflen;
4230 4231 4232
}


4233 4234 4235 4236 4237 4238 4239 4240 4241 4242 4243 4244 4245 4246 4247 4248 4249 4250 4251 4252 4253 4254 4255 4256 4257 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 4289 4290 4291 4292 4293 4294
/**
 * qemuDomainLogAppendMessage:
 *
 * This is a best-effort attempt to add a log message to the qemu log file
 * either by using virtlogd or the legacy approach */
int
qemuDomainLogAppendMessage(virQEMUDriverPtr driver,
                           virDomainObjPtr vm,
                           const char *fmt,
                           ...)
{
    virQEMUDriverConfigPtr cfg = virQEMUDriverGetConfig(driver);
    virLogManagerPtr manager = NULL;
    va_list ap;
    char *path = NULL;
    int writefd = -1;
    char *message = NULL;
    int ret = -1;

    va_start(ap, fmt);

    if (virVasprintf(&message, fmt, ap) < 0)
        goto cleanup;

    VIR_DEBUG("Append log message (vm='%s' message='%s) stdioLogD=%d",
              vm->def->name, message, cfg->stdioLogD);

    if (virAsprintf(&path, "%s/%s.log", cfg->logDir, vm->def->name) < 0)
        goto cleanup;

    if (cfg->stdioLogD) {
        if (!(manager = virLogManagerNew(virQEMUDriverIsPrivileged(driver))))
            goto cleanup;

        if (virLogManagerDomainAppendMessage(manager, "qemu", vm->def->uuid,
                                             vm->def->name, path, message, 0) < 0)
            goto cleanup;
    } else {
        if ((writefd = open(path, O_WRONLY | O_CREAT | O_APPEND, S_IRUSR | S_IWUSR)) < 0) {
            virReportSystemError(errno, _("failed to create logfile %s"),
                                 path);
            goto cleanup;
        }

        if (safewrite(writefd, message, strlen(message)) < 0)
            goto cleanup;
    }

    ret = 0;

 cleanup:
    va_end(ap);
    VIR_FREE(message);
    VIR_FORCE_CLOSE(writefd);
    virLogManagerFree(manager);
    virObjectUnref(cfg);
    VIR_FREE(path);

    return ret;
}


4295 4296 4297 4298 4299 4300 4301 4302
int qemuDomainLogContextGetWriteFD(qemuDomainLogContextPtr ctxt)
{
    return ctxt->writefd;
}


void qemuDomainLogContextMarkPosition(qemuDomainLogContextPtr ctxt)
{
4303 4304
    if (ctxt->manager)
        virLogManagerDomainGetLogFilePosition(ctxt->manager,
4305
                                              ctxt->path,
4306 4307 4308 4309 4310
                                              0,
                                              &ctxt->inode,
                                              &ctxt->pos);
    else
        ctxt->pos = lseek(ctxt->writefd, 0, SEEK_END);
4311 4312 4313 4314 4315
}


void qemuDomainLogContextRef(qemuDomainLogContextPtr ctxt)
{
4316
    VIR_DEBUG("Context ref %p", ctxt);
4317 4318 4319 4320
    virAtomicIntInc(&ctxt->refs);
}


4321 4322 4323 4324 4325 4326
virLogManagerPtr qemuDomainLogContextGetManager(qemuDomainLogContextPtr ctxt)
{
    return ctxt->manager;
}


4327 4328 4329 4330 4331 4332 4333 4334
void qemuDomainLogContextFree(qemuDomainLogContextPtr ctxt)
{
    bool lastRef;

    if (!ctxt)
        return;

    lastRef = virAtomicIntDecAndTest(&ctxt->refs);
4335
    VIR_DEBUG("Context free %p lastref=%d", ctxt, lastRef);
4336 4337 4338
    if (!lastRef)
        return;

4339
    virLogManagerFree(ctxt->manager);
4340
    VIR_FREE(ctxt->path);
4341 4342 4343 4344 4345 4346
    VIR_FORCE_CLOSE(ctxt->writefd);
    VIR_FORCE_CLOSE(ctxt->readfd);
    VIR_FREE(ctxt);
}


4347 4348
/* Locate an appropriate 'qemu-img' binary.  */
const char *
4349
qemuFindQemuImgBinary(virQEMUDriverPtr driver)
4350
{
4351 4352
    if (!driver->qemuImgBinary)
        virReportError(VIR_ERR_INTERNAL_ERROR,
4353
                       "%s", _("unable to find qemu-img"));
4354 4355 4356 4357 4358 4359 4360

    return driver->qemuImgBinary;
}

int
qemuDomainSnapshotWriteMetadata(virDomainObjPtr vm,
                                virDomainSnapshotObjPtr snapshot,
4361
                                virCapsPtr caps,
4362 4363 4364 4365 4366 4367 4368 4369 4370
                                char *snapshotDir)
{
    char *newxml = NULL;
    int ret = -1;
    char *snapDir = NULL;
    char *snapFile = NULL;
    char uuidstr[VIR_UUID_STRING_BUFLEN];

    virUUIDFormat(vm->def->uuid, uuidstr);
4371
    newxml = virDomainSnapshotDefFormat(
4372
        uuidstr, snapshot->def, caps,
4373 4374
        virDomainDefFormatConvertXMLFlags(QEMU_DOMAIN_FORMAT_LIVE_FLAGS),
        1);
4375
    if (newxml == NULL)
4376 4377
        return -1;

4378
    if (virAsprintf(&snapDir, "%s/%s", snapshotDir, vm->def->name) < 0)
4379 4380 4381 4382 4383 4384 4385
        goto cleanup;
    if (virFileMakePath(snapDir) < 0) {
        virReportSystemError(errno, _("cannot create snapshot directory '%s'"),
                             snapDir);
        goto cleanup;
    }

4386
    if (virAsprintf(&snapFile, "%s/%s.xml", snapDir, snapshot->def->name) < 0)
4387 4388
        goto cleanup;

J
Ján Tomko 已提交
4389
    ret = virXMLSaveFile(snapFile, NULL, "snapshot-edit", newxml);
4390

4391
 cleanup:
4392 4393 4394 4395 4396 4397 4398 4399
    VIR_FREE(snapFile);
    VIR_FREE(snapDir);
    VIR_FREE(newxml);
    return ret;
}

/* The domain is expected to be locked and inactive. Return -1 on normal
 * failure, 1 if we skipped a disk due to try_all.  */
4400
static int
4401
qemuDomainSnapshotForEachQcow2Raw(virQEMUDriverPtr driver,
4402 4403 4404 4405 4406
                                  virDomainDefPtr def,
                                  const char *name,
                                  const char *op,
                                  bool try_all,
                                  int ndisks)
4407 4408
{
    const char *qemuimgarg[] = { NULL, "snapshot", NULL, NULL, NULL, NULL };
4409
    size_t i;
4410 4411 4412 4413 4414 4415 4416 4417 4418
    bool skipped = false;

    qemuimgarg[0] = qemuFindQemuImgBinary(driver);
    if (qemuimgarg[0] == NULL) {
        /* qemuFindQemuImgBinary set the error */
        return -1;
    }

    qemuimgarg[2] = op;
4419
    qemuimgarg[3] = name;
4420

4421
    for (i = 0; i < ndisks; i++) {
4422
        /* FIXME: we also need to handle LVM here */
4423
        if (def->disks[i]->device == VIR_DOMAIN_DISK_DEVICE_DISK) {
4424 4425 4426
            int format = virDomainDiskGetFormat(def->disks[i]);

            if (format > 0 && format != VIR_STORAGE_FILE_QCOW2) {
4427 4428 4429 4430 4431
                if (try_all) {
                    /* Continue on even in the face of error, since other
                     * disks in this VM may have the same snapshot name.
                     */
                    VIR_WARN("skipping snapshot action on %s",
4432
                             def->disks[i]->dst);
4433 4434
                    skipped = true;
                    continue;
4435 4436 4437 4438 4439
                } else if (STREQ(op, "-c") && i) {
                    /* We must roll back partial creation by deleting
                     * all earlier snapshots.  */
                    qemuDomainSnapshotForEachQcow2Raw(driver, def, name,
                                                      "-d", false, i);
4440
                }
4441 4442 4443 4444
                virReportError(VIR_ERR_OPERATION_INVALID,
                               _("Disk device '%s' does not support"
                                 " snapshotting"),
                               def->disks[i]->dst);
4445 4446 4447
                return -1;
            }

4448
            qemuimgarg[4] = virDomainDiskGetSource(def->disks[i]);
4449 4450 4451 4452

            if (virRun(qemuimgarg, NULL) < 0) {
                if (try_all) {
                    VIR_WARN("skipping snapshot action on %s",
4453
                             def->disks[i]->dst);
4454 4455
                    skipped = true;
                    continue;
4456 4457 4458 4459 4460
                } else if (STREQ(op, "-c") && i) {
                    /* We must roll back partial creation by deleting
                     * all earlier snapshots.  */
                    qemuDomainSnapshotForEachQcow2Raw(driver, def, name,
                                                      "-d", false, i);
4461 4462 4463 4464 4465 4466 4467 4468 4469
                }
                return -1;
            }
        }
    }

    return skipped ? 1 : 0;
}

4470 4471 4472
/* The domain is expected to be locked and inactive. Return -1 on normal
 * failure, 1 if we skipped a disk due to try_all.  */
int
4473
qemuDomainSnapshotForEachQcow2(virQEMUDriverPtr driver,
4474 4475 4476 4477 4478 4479 4480 4481 4482 4483 4484 4485 4486 4487 4488 4489
                               virDomainObjPtr vm,
                               virDomainSnapshotObjPtr snap,
                               const char *op,
                               bool try_all)
{
    /* Prefer action on the disks in use at the time the snapshot was
     * created; but fall back to current definition if dealing with a
     * snapshot created prior to libvirt 0.9.5.  */
    virDomainDefPtr def = snap->def->dom;

    if (!def)
        def = vm->def;
    return qemuDomainSnapshotForEachQcow2Raw(driver, def, snap->def->name,
                                             op, try_all, def->ndisks);
}

4490 4491
/* Discard one snapshot (or its metadata), without reparenting any children.  */
int
4492
qemuDomainSnapshotDiscard(virQEMUDriverPtr driver,
4493 4494 4495 4496 4497 4498 4499 4500 4501
                          virDomainObjPtr vm,
                          virDomainSnapshotObjPtr snap,
                          bool update_current,
                          bool metadata_only)
{
    char *snapFile = NULL;
    int ret = -1;
    qemuDomainObjPrivatePtr priv;
    virDomainSnapshotObjPtr parentsnap = NULL;
4502
    virQEMUDriverConfigPtr cfg = virQEMUDriverGetConfig(driver);
4503 4504 4505 4506 4507 4508 4509 4510 4511

    if (!metadata_only) {
        if (!virDomainObjIsActive(vm)) {
            /* Ignore any skipped disks */
            if (qemuDomainSnapshotForEachQcow2(driver, vm, snap, "-d",
                                               true) < 0)
                goto cleanup;
        } else {
            priv = vm->privateData;
4512
            qemuDomainObjEnterMonitor(driver, vm);
4513 4514
            /* we continue on even in the face of error */
            qemuMonitorDeleteSnapshot(priv->mon, snap->def->name);
4515
            ignore_value(qemuDomainObjExitMonitor(driver, vm));
4516 4517 4518
        }
    }

4519
    if (virAsprintf(&snapFile, "%s/%s/%s.xml", cfg->snapshotDir,
4520
                    vm->def->name, snap->def->name) < 0)
4521 4522 4523 4524
        goto cleanup;

    if (snap == vm->current_snapshot) {
        if (update_current && snap->def->parent) {
4525
            parentsnap = virDomainSnapshotFindByName(vm->snapshots,
4526 4527 4528 4529 4530 4531
                                                     snap->def->parent);
            if (!parentsnap) {
                VIR_WARN("missing parent snapshot matching name '%s'",
                         snap->def->parent);
            } else {
                parentsnap->def->current = true;
4532
                if (qemuDomainSnapshotWriteMetadata(vm, parentsnap, driver->caps,
4533
                                                    cfg->snapshotDir) < 0) {
4534 4535 4536 4537 4538 4539 4540 4541 4542 4543 4544 4545
                    VIR_WARN("failed to set parent snapshot '%s' as current",
                             snap->def->parent);
                    parentsnap->def->current = false;
                    parentsnap = NULL;
                }
            }
        }
        vm->current_snapshot = parentsnap;
    }

    if (unlink(snapFile) < 0)
        VIR_WARN("Failed to unlink %s", snapFile);
4546
    virDomainSnapshotObjListRemove(vm->snapshots, snap);
4547 4548 4549

    ret = 0;

4550
 cleanup:
4551
    VIR_FREE(snapFile);
4552
    virObjectUnref(cfg);
4553 4554 4555 4556
    return ret;
}

/* Hash iterator callback to discard multiple snapshots.  */
4557 4558 4559
int qemuDomainSnapshotDiscardAll(void *payload,
                                 const void *name ATTRIBUTE_UNUSED,
                                 void *data)
4560 4561
{
    virDomainSnapshotObjPtr snap = payload;
4562
    virQEMUSnapRemovePtr curr = data;
4563 4564 4565 4566 4567 4568 4569 4570
    int err;

    if (snap->def->current)
        curr->current = true;
    err = qemuDomainSnapshotDiscard(curr->driver, curr->vm, snap, false,
                                    curr->metadata_only);
    if (err && !curr->err)
        curr->err = err;
4571
    return 0;
4572 4573 4574
}

int
4575
qemuDomainSnapshotDiscardAllMetadata(virQEMUDriverPtr driver,
4576 4577
                                     virDomainObjPtr vm)
{
4578
    virQEMUSnapRemove rem;
4579 4580 4581 4582 4583

    rem.driver = driver;
    rem.vm = vm;
    rem.metadata_only = true;
    rem.err = 0;
4584 4585
    virDomainSnapshotForEach(vm->snapshots, qemuDomainSnapshotDiscardAll,
                             &rem);
4586 4587 4588 4589 4590

    return rem.err;
}

/*
4591
 * The caller must hold a lock the vm.
4592 4593
 */
void
4594
qemuDomainRemoveInactive(virQEMUDriverPtr driver,
4595 4596
                         virDomainObjPtr vm)
{
4597
    bool haveJob = true;
4598
    char *snapDir;
4599 4600 4601 4602 4603 4604 4605 4606
    virQEMUDriverConfigPtr cfg;

    if (vm->persistent) {
        /* Short-circuit, we don't want to remove a persistent domain */
        return;
    }

    cfg = virQEMUDriverGetConfig(driver);
4607

4608 4609 4610
    if (qemuDomainObjBeginJob(driver, vm, QEMU_JOB_MODIFY) < 0)
        haveJob = false;

4611 4612 4613 4614 4615
    /* Remove any snapshot metadata prior to removing the domain */
    if (qemuDomainSnapshotDiscardAllMetadata(driver, vm) < 0) {
        VIR_WARN("unable to remove all snapshots for domain %s",
                 vm->def->name);
    }
4616
    else if (virAsprintf(&snapDir, "%s/%s", cfg->snapshotDir,
4617 4618
                         vm->def->name) < 0) {
        VIR_WARN("unable to remove snapshot directory %s/%s",
4619
                 cfg->snapshotDir, vm->def->name);
4620 4621 4622 4623 4624
    } else {
        if (rmdir(snapDir) < 0 && errno != ENOENT)
            VIR_WARN("unable to remove snapshot directory %s", snapDir);
        VIR_FREE(snapDir);
    }
4625 4626 4627

    virObjectRef(vm);

4628
    virDomainObjListRemove(driver->domains, vm);
4629 4630 4631 4632 4633 4634 4635 4636 4637 4638 4639 4640 4641
    /*
     * virDomainObjListRemove() leaves the domain unlocked so it can
     * be unref'd for other drivers that depend on that, but we still
     * need to reset a job and we have a reference from the API that
     * called this function.  So we need to lock it back.  This is
     * just a workaround for the qemu driver.
     *
     * XXX: Ideally, the global handling of domain objects and object
     *      lists would be refactored so we don't need hacks like
     *      this, but since that requires refactor of all drivers,
     *      it's a work for another day.
     */
    virObjectLock(vm);
4642
    virObjectUnref(cfg);
4643 4644

    if (haveJob)
4645
        qemuDomainObjEndJob(driver, vm);
4646 4647

    virObjectUnref(vm);
4648
}
4649 4650

void
4651
qemuDomainSetFakeReboot(virQEMUDriverPtr driver,
4652 4653 4654 4655
                        virDomainObjPtr vm,
                        bool value)
{
    qemuDomainObjPrivatePtr priv = vm->privateData;
4656
    virQEMUDriverConfigPtr cfg = virQEMUDriverGetConfig(driver);
4657 4658

    if (priv->fakeReboot == value)
4659
        goto cleanup;
4660 4661 4662

    priv->fakeReboot = value;

4663
    if (virDomainSaveStatus(driver->xmlopt, cfg->stateDir, vm, driver->caps) < 0)
4664
        VIR_WARN("Failed to save status on vm %s", vm->def->name);
4665

4666
 cleanup:
4667
    virObjectUnref(cfg);
4668
}
M
Michal Privoznik 已提交
4669

4670
static void
4671 4672
qemuDomainCheckRemoveOptionalDisk(virQEMUDriverPtr driver,
                                  virDomainObjPtr vm,
4673
                                  size_t diskIndex)
4674 4675
{
    char uuid[VIR_UUID_STRING_BUFLEN];
4676
    virObjectEventPtr event = NULL;
4677
    virDomainDiskDefPtr disk = vm->def->disks[diskIndex];
4678
    const char *src = virDomainDiskGetSource(disk);
4679 4680 4681 4682 4683

    virUUIDFormat(vm->def->uuid, uuid);

    VIR_DEBUG("Dropping disk '%s' on domain '%s' (UUID '%s') "
              "due to inaccessible source '%s'",
4684
              disk->dst, vm->def->name, uuid, src);
4685 4686 4687 4688

    if (disk->device == VIR_DOMAIN_DISK_DEVICE_CDROM ||
        disk->device == VIR_DOMAIN_DISK_DEVICE_FLOPPY) {

4689
        event = virDomainEventDiskChangeNewFromObj(vm, src, NULL,
4690 4691
                                                   disk->info.alias,
                                                   VIR_DOMAIN_EVENT_DISK_CHANGE_MISSING_ON_START);
4692
        ignore_value(virDomainDiskSetSource(disk, NULL));
4693 4694
        /* keeping the old startup policy would be invalid for new images */
        disk->startupPolicy = VIR_DOMAIN_STARTUP_POLICY_DEFAULT;
4695
    } else {
4696
        event = virDomainEventDiskChangeNewFromObj(vm, src, NULL,
4697 4698
                                                   disk->info.alias,
                                                   VIR_DOMAIN_EVENT_DISK_DROP_MISSING_ON_START);
4699 4700
        virDomainDiskRemove(vm->def, diskIndex);
        virDomainDiskDefFree(disk);
4701 4702
    }

4703
    qemuDomainEventQueue(driver, event);
4704 4705
}

4706 4707 4708
static int
qemuDomainCheckDiskStartupPolicy(virQEMUDriverPtr driver,
                                 virDomainObjPtr vm,
4709
                                 size_t diskIndex,
4710 4711
                                 bool cold_boot)
{
4712
    int startupPolicy = vm->def->disks[diskIndex]->startupPolicy;
4713
    int device = vm->def->disks[diskIndex]->device;
4714

4715
    switch ((virDomainStartupPolicy) startupPolicy) {
4716
        case VIR_DOMAIN_STARTUP_POLICY_OPTIONAL:
4717 4718 4719 4720 4721 4722
            /* Once started with an optional disk, qemu saves its section
             * in the migration stream, so later, when restoring from it
             * we must make sure the sections match. */
            if (!cold_boot &&
                device != VIR_DOMAIN_DISK_DEVICE_FLOPPY &&
                device != VIR_DOMAIN_DISK_DEVICE_CDROM)
4723
                return -1;
4724 4725
            break;

4726
        case VIR_DOMAIN_STARTUP_POLICY_DEFAULT:
4727
        case VIR_DOMAIN_STARTUP_POLICY_MANDATORY:
4728
            return -1;
4729 4730

        case VIR_DOMAIN_STARTUP_POLICY_REQUISITE:
4731
            if (cold_boot)
4732
                return -1;
4733 4734 4735 4736 4737 4738 4739
            break;

        case VIR_DOMAIN_STARTUP_POLICY_LAST:
            /* this should never happen */
            break;
    }

4740
    qemuDomainCheckRemoveOptionalDisk(driver, vm, diskIndex);
4741
    virResetLastError();
4742 4743 4744
    return 0;
}

4745

M
Michal Privoznik 已提交
4746
int
4747 4748
qemuDomainCheckDiskPresence(virConnectPtr conn,
                            virQEMUDriverPtr driver,
M
Michal Privoznik 已提交
4749
                            virDomainObjPtr vm,
4750
                            unsigned int flags)
M
Michal Privoznik 已提交
4751
{
4752
    size_t i;
4753 4754
    bool pretend = flags & VIR_QEMU_PROCESS_START_PRETEND;
    bool cold_boot = flags & VIR_QEMU_PROCESS_START_COLD;
M
Michal Privoznik 已提交
4755

4756
    VIR_DEBUG("Checking for disk presence");
4757
    for (i = vm->def->ndisks; i > 0; i--) {
4758 4759
        size_t idx = i - 1;
        virDomainDiskDefPtr disk = vm->def->disks[idx];
4760
        virStorageFileFormat format = virDomainDiskGetFormat(disk);
M
Michal Privoznik 已提交
4761

4762 4763 4764 4765 4766 4767 4768 4769 4770 4771
        if (virStorageTranslateDiskSourcePool(conn, vm->def->disks[idx]) < 0) {
            if (pretend ||
                qemuDomainCheckDiskStartupPolicy(driver, vm, idx, cold_boot) < 0)
                return -1;
            continue;
        }

        if (pretend)
            continue;

4772
        if (virStorageSourceIsEmpty(disk->src))
4773 4774 4775 4776 4777
            continue;

        /* There is no need to check the backing chain for disks
         * without backing support, the fact that the file exists is
         * more than enough */
4778
        if (virStorageSourceIsLocalStorage(disk->src) &&
4779
            format > VIR_STORAGE_FILE_NONE &&
4780
            format < VIR_STORAGE_FILE_BACKING &&
4781
            virFileExists(virDomainDiskGetSource(disk)))
M
Michal Privoznik 已提交
4782 4783
            continue;

4784
        if (qemuDomainDetermineDiskChain(driver, vm, disk, true, true) >= 0)
M
Michal Privoznik 已提交
4785
            continue;
4786

4787
        if (qemuDomainCheckDiskStartupPolicy(driver, vm, idx, cold_boot) >= 0)
4788
            continue;
M
Michal Privoznik 已提交
4789

4790
        return -1;
M
Michal Privoznik 已提交
4791 4792
    }

4793
    return 0;
M
Michal Privoznik 已提交
4794
}
4795 4796 4797 4798 4799 4800 4801 4802 4803 4804

/*
 * The vm must be locked when any of the following cleanup functions is
 * called.
 */
int
qemuDomainCleanupAdd(virDomainObjPtr vm,
                     qemuDomainCleanupCallback cb)
{
    qemuDomainObjPrivatePtr priv = vm->privateData;
4805
    size_t i;
4806 4807 4808 4809 4810 4811 4812 4813 4814 4815

    VIR_DEBUG("vm=%s, cb=%p", vm->def->name, cb);

    for (i = 0; i < priv->ncleanupCallbacks; i++) {
        if (priv->cleanupCallbacks[i] == cb)
            return 0;
    }

    if (VIR_RESIZE_N(priv->cleanupCallbacks,
                     priv->ncleanupCallbacks_max,
4816
                     priv->ncleanupCallbacks, 1) < 0)
4817 4818 4819 4820 4821 4822 4823 4824 4825 4826 4827
        return -1;

    priv->cleanupCallbacks[priv->ncleanupCallbacks++] = cb;
    return 0;
}

void
qemuDomainCleanupRemove(virDomainObjPtr vm,
                        qemuDomainCleanupCallback cb)
{
    qemuDomainObjPrivatePtr priv = vm->privateData;
4828
    size_t i;
4829 4830 4831 4832

    VIR_DEBUG("vm=%s, cb=%p", vm->def->name, cb);

    for (i = 0; i < priv->ncleanupCallbacks; i++) {
4833 4834 4835
        if (priv->cleanupCallbacks[i] == cb)
            VIR_DELETE_ELEMENT_INPLACE(priv->cleanupCallbacks,
                                       i, priv->ncleanupCallbacks);
4836 4837 4838 4839 4840 4841 4842 4843
    }

    VIR_SHRINK_N(priv->cleanupCallbacks,
                 priv->ncleanupCallbacks_max,
                 priv->ncleanupCallbacks_max - priv->ncleanupCallbacks);
}

void
4844
qemuDomainCleanupRun(virQEMUDriverPtr driver,
4845 4846 4847
                     virDomainObjPtr vm)
{
    qemuDomainObjPrivatePtr priv = vm->privateData;
4848
    size_t i;
4849 4850 4851 4852

    VIR_DEBUG("driver=%p, vm=%s", driver, vm->def->name);

    /* run cleanup callbacks in reverse order */
4853 4854
    for (i = 0; i < priv->ncleanupCallbacks; i++) {
        if (priv->cleanupCallbacks[priv->ncleanupCallbacks - (i + 1)])
4855 4856 4857 4858 4859 4860 4861
            priv->cleanupCallbacks[i](driver, vm);
    }

    VIR_FREE(priv->cleanupCallbacks);
    priv->ncleanupCallbacks = 0;
    priv->ncleanupCallbacks_max = 0;
}
4862

4863 4864 4865
static void
qemuDomainGetImageIds(virQEMUDriverConfigPtr cfg,
                      virDomainObjPtr vm,
4866
                      virStorageSourcePtr src,
4867 4868 4869 4870 4871 4872 4873 4874 4875 4876 4877 4878 4879 4880 4881 4882 4883 4884
                      uid_t *uid, gid_t *gid)
{
    virSecurityLabelDefPtr vmlabel;
    virSecurityDeviceLabelDefPtr disklabel;

    if (uid)
        *uid = -1;
    if (gid)
        *gid = -1;

    if (cfg) {
        if (uid)
            *uid = cfg->user;

        if (gid)
            *gid = cfg->group;
    }

4885 4886
    if (vm && (vmlabel = virDomainDefGetSecurityLabelDef(vm->def, "dac")) &&
        vmlabel->label)
4887 4888
        virParseOwnershipIds(vmlabel->label, uid, gid);

4889
    if ((disklabel = virStorageSourceGetSecurityLabelDef(src, "dac")) &&
4890
        disklabel->label)
4891 4892 4893 4894
        virParseOwnershipIds(disklabel->label, uid, gid);
}


4895 4896 4897 4898 4899 4900 4901 4902 4903 4904 4905 4906 4907 4908 4909 4910 4911 4912 4913 4914 4915 4916 4917
int
qemuDomainStorageFileInit(virQEMUDriverPtr driver,
                          virDomainObjPtr vm,
                          virStorageSourcePtr src)
{
    virQEMUDriverConfigPtr cfg = virQEMUDriverGetConfig(driver);
    uid_t uid;
    gid_t gid;
    int ret = -1;

    qemuDomainGetImageIds(cfg, vm, src, &uid, &gid);

    if (virStorageFileInitAs(src, uid, gid) < 0)
        goto cleanup;

    ret = 0;

 cleanup:
    virObjectUnref(cfg);
    return ret;
}


4918 4919 4920 4921 4922
char *
qemuDomainStorageAlias(const char *device, int depth)
{
    char *alias;

4923
    device = qemuAliasDiskDriveSkipPrefix(device);
4924 4925 4926 4927 4928 4929 4930 4931 4932

    if (!depth)
        ignore_value(VIR_STRDUP(alias, device));
    else
        ignore_value(virAsprintf(&alias, "%s.%d", device, depth));
    return alias;
}


4933
int
4934
qemuDomainDetermineDiskChain(virQEMUDriverPtr driver,
4935
                             virDomainObjPtr vm,
4936
                             virDomainDiskDefPtr disk,
4937 4938
                             bool force_probe,
                             bool report_broken)
4939
{
4940 4941
    virQEMUDriverConfigPtr cfg = virQEMUDriverGetConfig(driver);
    int ret = 0;
4942 4943
    uid_t uid;
    gid_t gid;
4944

4945
    if (virStorageSourceIsEmpty(disk->src))
4946
        goto cleanup;
4947

4948
    if (disk->src->backingStore) {
4949
        if (force_probe)
4950
            virStorageSourceBackingStoreClear(disk->src);
4951
        else
4952
            goto cleanup;
4953
    }
4954

4955
    qemuDomainGetImageIds(cfg, vm, disk->src, &uid, &gid);
4956

4957
    if (virStorageFileGetMetadata(disk->src,
4958
                                  uid, gid,
4959
                                  cfg->allowDiskFormatProbing,
4960
                                  report_broken) < 0)
4961 4962
        ret = -1;

4963
 cleanup:
4964 4965
    virObjectUnref(cfg);
    return ret;
4966
}
4967

4968

4969 4970 4971 4972 4973 4974 4975 4976 4977 4978 4979 4980 4981 4982 4983 4984 4985 4986 4987 4988 4989 4990 4991 4992 4993 4994 4995 4996 4997 4998 4999 5000 5001 5002 5003 5004 5005 5006 5007 5008 5009 5010 5011 5012 5013 5014 5015 5016 5017 5018 5019 5020 5021 5022 5023 5024 5025 5026 5027 5028 5029 5030 5031
/**
 * qemuDomainDiskChainElementRevoke:
 *
 * Revoke access to a single backing chain element. This restores the labels,
 * removes cgroup ACLs for devices and removes locks.
 */
void
qemuDomainDiskChainElementRevoke(virQEMUDriverPtr driver,
                                 virDomainObjPtr vm,
                                 virStorageSourcePtr elem)
{
    if (virSecurityManagerRestoreImageLabel(driver->securityManager,
                                            vm->def, elem) < 0)
        VIR_WARN("Unable to restore security label on %s", NULLSTR(elem->path));

    if (qemuTeardownImageCgroup(vm, elem) < 0)
        VIR_WARN("Failed to teardown cgroup for disk path %s",
                 NULLSTR(elem->path));

    if (virDomainLockImageDetach(driver->lockManager, vm, elem) < 0)
        VIR_WARN("Unable to release lock on %s", NULLSTR(elem->path));
}


/**
 * qemuDomainDiskChainElementPrepare:
 *
 * Allow a VM access to a single element of a disk backing chain; this helper
 * ensures that the lock manager, cgroup device controller, and security manager
 * labelling are all aware of each new file before it is added to a chain */
int
qemuDomainDiskChainElementPrepare(virQEMUDriverPtr driver,
                                  virDomainObjPtr vm,
                                  virStorageSourcePtr elem,
                                  bool readonly)
{
    bool was_readonly = elem->readonly;
    virQEMUDriverConfigPtr cfg = NULL;
    int ret = -1;

    cfg = virQEMUDriverGetConfig(driver);

    elem->readonly = readonly;

    if (virDomainLockImageAttach(driver->lockManager, cfg->uri, vm, elem) < 0)
        goto cleanup;

    if (qemuSetupImageCgroup(vm, elem) < 0)
        goto cleanup;

    if (virSecurityManagerSetImageLabel(driver->securityManager, vm->def,
                                        elem) < 0)
        goto cleanup;

    ret = 0;

 cleanup:
    elem->readonly = was_readonly;
    virObjectUnref(cfg);
    return ret;
}


5032
bool
5033
qemuDomainDiskSourceDiffers(virDomainDiskDefPtr disk,
5034 5035 5036 5037 5038 5039 5040 5041 5042 5043 5044 5045 5046 5047 5048
                            virDomainDiskDefPtr origDisk)
{
    char *diskSrc = NULL, *origDiskSrc = NULL;
    bool diskEmpty, origDiskEmpty;
    bool ret = true;

    diskEmpty = virStorageSourceIsEmpty(disk->src);
    origDiskEmpty = virStorageSourceIsEmpty(origDisk->src);

    if (diskEmpty && origDiskEmpty)
        return false;

    if (diskEmpty ^ origDiskEmpty)
        return true;

5049 5050 5051 5052
    /* This won't be a network storage, so no need to get the diskPriv
     * in order to fetch the secret, thus NULL for param2 */
    if (qemuGetDriveSourceString(disk->src, NULL, &diskSrc) < 0 ||
        qemuGetDriveSourceString(origDisk->src, NULL, &origDiskSrc) < 0)
5053 5054 5055 5056 5057 5058 5059 5060 5061 5062 5063 5064 5065
        goto cleanup;

    /* So far in qemu disk sources are considered different
     * if either path to disk or its format changes. */
    ret = virDomainDiskGetFormat(disk) != virDomainDiskGetFormat(origDisk) ||
          STRNEQ_NULLABLE(diskSrc, origDiskSrc);
 cleanup:
    VIR_FREE(diskSrc);
    VIR_FREE(origDiskSrc);
    return ret;
}


5066 5067 5068 5069 5070 5071 5072 5073 5074 5075 5076 5077 5078 5079 5080 5081 5082 5083 5084 5085 5086 5087 5088
/*
 * Makes sure the @disk differs from @orig_disk only by the source
 * path and nothing else.  Fields that are being checked and the
 * information whether they are nullable (may not be specified) or is
 * taken from the virDomainDiskDefFormat() code.
 */
bool
qemuDomainDiskChangeSupported(virDomainDiskDefPtr disk,
                              virDomainDiskDefPtr orig_disk)
{
#define CHECK_EQ(field, field_name, nullable)                           \
    do {                                                                \
        if (nullable && !disk->field)                                   \
            break;                                                      \
        if (disk->field != orig_disk->field) {                          \
            virReportError(VIR_ERR_OPERATION_UNSUPPORTED,               \
                           _("cannot modify field '%s' of the disk"),   \
                           field_name);                                 \
            return false;                                               \
        }                                                               \
    } while (0)

    CHECK_EQ(device, "device", false);
5089 5090 5091 5092
    CHECK_EQ(bus, "bus", false);
    if (STRNEQ(disk->dst, orig_disk->dst)) {
        virReportError(VIR_ERR_OPERATION_UNSUPPORTED,
                       _("cannot modify field '%s' of the disk"),
5093
                       "target");
5094 5095 5096 5097
        return false;
    }
    CHECK_EQ(tray_status, "tray", true);
    CHECK_EQ(removable, "removable", true);
5098 5099 5100 5101 5102 5103 5104 5105 5106 5107 5108 5109 5110 5111 5112 5113 5114 5115 5116 5117 5118 5119 5120 5121 5122 5123 5124 5125 5126 5127 5128 5129 5130 5131 5132 5133 5134 5135 5136 5137 5138 5139 5140 5141 5142 5143 5144 5145 5146 5147 5148 5149 5150 5151 5152 5153 5154 5155 5156 5157 5158 5159 5160 5161 5162 5163 5164 5165 5166 5167 5168 5169 5170 5171 5172 5173 5174 5175 5176 5177 5178 5179 5180

    if (disk->geometry.cylinders &&
        disk->geometry.heads &&
        disk->geometry.sectors) {
        CHECK_EQ(geometry.cylinders, "geometry cylinders", false);
        CHECK_EQ(geometry.heads, "geometry heads", false);
        CHECK_EQ(geometry.sectors, "geometry sectors", false);
        CHECK_EQ(geometry.trans, "BIOS-translation-modus", true);
    }

    CHECK_EQ(blockio.logical_block_size,
             "blockio logical_block_size", false);
    CHECK_EQ(blockio.physical_block_size,
             "blockio physical_block_size", false);

    CHECK_EQ(blkdeviotune.total_bytes_sec,
             "blkdeviotune total_bytes_sec",
             true);
    CHECK_EQ(blkdeviotune.read_bytes_sec,
             "blkdeviotune read_bytes_sec",
             true);
    CHECK_EQ(blkdeviotune.write_bytes_sec,
             "blkdeviotune write_bytes_sec",
             true);
    CHECK_EQ(blkdeviotune.total_iops_sec,
             "blkdeviotune total_iops_sec",
             true);
    CHECK_EQ(blkdeviotune.read_iops_sec,
             "blkdeviotune read_iops_sec",
             true);
    CHECK_EQ(blkdeviotune.write_iops_sec,
             "blkdeviotune write_iops_sec",
             true);
    CHECK_EQ(blkdeviotune.total_bytes_sec_max,
             "blkdeviotune total_bytes_sec_max",
             true);
    CHECK_EQ(blkdeviotune.read_bytes_sec_max,
             "blkdeviotune read_bytes_sec_max",
             true);
    CHECK_EQ(blkdeviotune.write_bytes_sec_max,
             "blkdeviotune write_bytes_sec_max",
             true);
    CHECK_EQ(blkdeviotune.total_iops_sec_max,
             "blkdeviotune total_iops_sec_max",
             true);
    CHECK_EQ(blkdeviotune.read_iops_sec_max,
             "blkdeviotune read_iops_sec_max",
             true);
    CHECK_EQ(blkdeviotune.write_iops_sec_max,
             "blkdeviotune write_iops_sec_max",
             true);
    CHECK_EQ(blkdeviotune.size_iops_sec,
             "blkdeviotune size_iops_sec",
             true);

    if (disk->serial && STRNEQ_NULLABLE(disk->serial, orig_disk->serial)) {
        virReportError(VIR_ERR_OPERATION_UNSUPPORTED,
                       _("cannot modify field '%s' of the disk"),
                       "serial");
        return false;
    }

    if (disk->wwn && STRNEQ_NULLABLE(disk->wwn, orig_disk->wwn)) {
        virReportError(VIR_ERR_OPERATION_UNSUPPORTED,
                       _("cannot modify field '%s' of the disk"),
                       "wwn");
        return false;
    }

    if (disk->vendor && STRNEQ_NULLABLE(disk->vendor, orig_disk->vendor)) {
        virReportError(VIR_ERR_OPERATION_UNSUPPORTED,
                       _("cannot modify field '%s' of the disk"),
                       "vendor");
        return false;
    }

    if (disk->product && STRNEQ_NULLABLE(disk->product, orig_disk->product)) {
        virReportError(VIR_ERR_OPERATION_UNSUPPORTED,
                       _("cannot modify field '%s' of the disk"),
                       "product");
        return false;
    }

5181 5182 5183 5184 5185 5186 5187
    CHECK_EQ(cachemode, "cache", true);
    CHECK_EQ(error_policy, "error_policy", true);
    CHECK_EQ(rerror_policy, "rerror_policy", true);
    CHECK_EQ(iomode, "io", true);
    CHECK_EQ(ioeventfd, "ioeventfd", true);
    CHECK_EQ(event_idx, "event_idx", true);
    CHECK_EQ(copy_on_read, "copy_on_read", true);
5188
    /* "snapshot" is a libvirt internal field and thus can be changed */
5189
    /* startupPolicy is allowed to be updated. Therefore not checked here. */
5190
    CHECK_EQ(transient, "transient", true);
5191
    CHECK_EQ(info.bootIndex, "boot order", true);
5192 5193 5194 5195 5196 5197 5198 5199 5200 5201 5202 5203
    CHECK_EQ(rawio, "rawio", true);
    CHECK_EQ(sgio, "sgio", true);
    CHECK_EQ(discard, "discard", true);
    CHECK_EQ(iothread, "iothread", true);

    if (disk->domain_name &&
        STRNEQ_NULLABLE(disk->domain_name, orig_disk->domain_name)) {
        virReportError(VIR_ERR_OPERATION_UNSUPPORTED,
                       _("cannot modify field '%s' of the disk"),
                       "backenddomain");
        return false;
    }
5204

5205 5206 5207 5208 5209 5210
    /* checks for fields stored in disk->src */
    /* unfortunately 'readonly' and 'shared' can't be converted to tristate
     * values thus we need to ignore the check if the new value is 'false' */
    CHECK_EQ(src->readonly, "readonly", true);
    CHECK_EQ(src->shared, "shared", true);

5211 5212 5213 5214 5215
#undef CHECK_EQ

    return true;
}

5216 5217 5218
bool
qemuDomainDiskBlockJobIsActive(virDomainDiskDefPtr disk)
{
5219 5220
    qemuDomainDiskPrivatePtr diskPriv = QEMU_DOMAIN_DISK_PRIVATE(disk);

5221 5222 5223 5224 5225 5226 5227 5228
    if (disk->mirror) {
        virReportError(VIR_ERR_BLOCK_COPY_ACTIVE,
                       _("disk '%s' already in active block job"),
                       disk->dst);

        return true;
    }

5229
    if (diskPriv->blockjob) {
5230 5231 5232 5233 5234 5235 5236 5237 5238 5239
        virReportError(VIR_ERR_OPERATION_UNSUPPORTED,
                       _("disk '%s' already in active block job"),
                       disk->dst);
        return true;
    }

    return false;
}


5240 5241 5242 5243 5244 5245 5246 5247 5248 5249 5250 5251 5252 5253 5254
/**
 * qemuDomainHasBlockjob:
 * @vm: domain object
 * @copy_only: Reject only block copy job
 *
 * Return true if @vm has at least one disk involved in a current block
 * copy/commit/pull job. If @copy_only is true this returns true only if the
 * disk is involved in a block copy.
 * */
bool
qemuDomainHasBlockjob(virDomainObjPtr vm,
                      bool copy_only)
{
    size_t i;
    for (i = 0; i < vm->def->ndisks; i++) {
5255 5256 5257 5258
        virDomainDiskDefPtr disk = vm->def->disks[i];
        qemuDomainDiskPrivatePtr diskPriv = QEMU_DOMAIN_DISK_PRIVATE(disk);

        if (!copy_only && diskPriv->blockjob)
5259 5260
            return true;

5261
        if (disk->mirror && disk->mirrorJob == VIR_DOMAIN_BLOCK_JOB_TYPE_COPY)
5262 5263 5264 5265 5266 5267 5268
            return true;
    }

    return false;
}


5269 5270
int
qemuDomainUpdateDeviceList(virQEMUDriverPtr driver,
5271 5272
                           virDomainObjPtr vm,
                           int asyncJob)
5273 5274 5275
{
    qemuDomainObjPrivatePtr priv = vm->privateData;
    char **aliases;
5276
    int rc;
5277 5278 5279 5280

    if (!virQEMUCapsGet(priv->qemuCaps, QEMU_CAPS_DEVICE_DEL_EVENT))
        return 0;

5281 5282
    if (qemuDomainObjEnterMonitorAsync(driver, vm, asyncJob) < 0)
        return -1;
5283 5284 5285 5286
    rc = qemuMonitorGetDeviceAliases(priv->mon, &aliases);
    if (qemuDomainObjExitMonitor(driver, vm) < 0)
        return -1;
    if (rc < 0)
5287 5288 5289 5290 5291 5292
        return -1;

    virStringFreeList(priv->qemuDevices);
    priv->qemuDevices = aliases;
    return 0;
}
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

int
qemuDomainUpdateMemoryDeviceInfo(virQEMUDriverPtr driver,
                                 virDomainObjPtr vm,
                                 int asyncJob)
{
    qemuDomainObjPrivatePtr priv = vm->privateData;
    virHashTablePtr meminfo = NULL;
    int rc;
    size_t i;

    if (vm->def->nmems == 0)
        return 0;

    if (qemuDomainObjEnterMonitorAsync(driver, vm, asyncJob) < 0)
        return -1;

    rc = qemuMonitorGetMemoryDeviceInfo(priv->mon, &meminfo);

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

    /* if qemu doesn't support the info request, just carry on */
    if (rc == -2)
        return 0;

    if (rc < 0)
        return -1;

    for (i = 0; i < vm->def->nmems; i++) {
        virDomainMemoryDefPtr mem = vm->def->mems[i];
        qemuMonitorMemoryDeviceInfoPtr dimm;

        if (!mem->info.alias)
            continue;

        if (!(dimm = virHashLookup(meminfo, mem->info.alias)))
            continue;

        mem->info.type = VIR_DOMAIN_DEVICE_ADDRESS_TYPE_DIMM;
        mem->info.addr.dimm.slot = dimm->slot;
        mem->info.addr.dimm.base = dimm->address;
    }

    virHashFree(meminfo);
    return 0;
}


5343 5344 5345 5346 5347 5348 5349
bool
qemuDomainDefCheckABIStability(virQEMUDriverPtr driver,
                               virDomainDefPtr src,
                               virDomainDefPtr dst)
{
    virDomainDefPtr migratableDefSrc = NULL;
    virDomainDefPtr migratableDefDst = NULL;
5350 5351 5352 5353
    const unsigned int flags = VIR_DOMAIN_XML_SECURE |
                               VIR_DOMAIN_XML_UPDATE_CPU |
                               VIR_DOMAIN_XML_MIGRATABLE;
    const unsigned int check_flags = VIR_DOMAIN_DEF_ABI_CHECK_SKIP_VOLATILE;
5354 5355 5356 5357 5358 5359
    bool ret = false;

    if (!(migratableDefSrc = qemuDomainDefCopy(driver, src, flags)) ||
        !(migratableDefDst = qemuDomainDefCopy(driver, dst, flags)))
        goto cleanup;

5360 5361 5362 5363 5364 5365 5366 5367 5368
    if (!virDomainDefCheckABIStabilityFlags(migratableDefSrc,
                                            migratableDefDst,
                                            check_flags))
        goto cleanup;

    /* Force update any skipped values from the volatile flag */
    dst->mem.cur_balloon = src->mem.cur_balloon;

    ret = true;
5369

5370
 cleanup:
5371 5372 5373 5374
    virDomainDefFree(migratableDefSrc);
    virDomainDefFree(migratableDefDst);
    return ret;
}
5375 5376

bool
5377
qemuDomainAgentAvailable(virDomainObjPtr vm,
5378 5379
                         bool reportError)
{
5380 5381
    qemuDomainObjPrivatePtr priv = vm->privateData;

5382 5383 5384 5385 5386 5387 5388
    if (virDomainObjGetState(vm, NULL) != VIR_DOMAIN_RUNNING) {
        if (reportError) {
            virReportError(VIR_ERR_OPERATION_INVALID, "%s",
                           _("domain is not running"));
        }
        return false;
    }
5389 5390 5391 5392 5393 5394 5395 5396 5397
    if (priv->agentError) {
        if (reportError) {
            virReportError(VIR_ERR_AGENT_UNRESPONSIVE, "%s",
                           _("QEMU guest agent is not "
                             "available due to an error"));
        }
        return false;
    }
    if (!priv->agent) {
5398 5399 5400 5401 5402 5403 5404 5405 5406 5407 5408 5409
        if (qemuFindAgentConfig(vm->def)) {
            if (reportError) {
                virReportError(VIR_ERR_AGENT_UNRESPONSIVE, "%s",
                               _("QEMU guest agent is not connected"));
            }
            return false;
        } else {
            if (reportError) {
                virReportError(VIR_ERR_ARGUMENT_UNSUPPORTED, "%s",
                               _("QEMU guest agent is not configured"));
            }
            return false;
5410 5411 5412 5413
        }
    }
    return true;
}
5414

5415

5416
static unsigned long long
5417
qemuDomainGetMemorySizeAlignment(virDomainDefPtr def)
5418
{
5419 5420 5421 5422 5423
    /* PPC requires the memory sizes to be rounded to 256MiB increments, so
     * round them to the size always. */
    if (ARCH_IS_PPC64(def->os.arch))
        return 256 * 1024;

5424 5425 5426 5427 5428 5429 5430
    /* Align memory size. QEMU requires rounding to next 4KiB block.
     * We'll take the "traditional" path and round it to 1MiB*/

    return 1024;
}


5431 5432 5433 5434 5435 5436 5437 5438 5439 5440 5441 5442 5443 5444 5445
static unsigned long long
qemuDomainGetMemoryModuleSizeAlignment(const virDomainDef *def,
                                       const virDomainMemoryDef *mem ATTRIBUTE_UNUSED)
{
    /* PPC requires the memory sizes to be rounded to 256MiB increments, so
     * round them to the size always. */
    if (ARCH_IS_PPC64(def->os.arch))
        return 256 * 1024;

    /* dimm memory modules require 2MiB alignment rather than the 1MiB we are
     * using elsewhere. */
    return 2048;
}


5446 5447 5448
int
qemuDomainAlignMemorySizes(virDomainDefPtr def)
{
5449 5450
    unsigned long long maxmemkb = virMemoryMaxValue(false) >> 10;
    unsigned long long maxmemcapped = virMemoryMaxValue(true) >> 10;
5451
    unsigned long long initialmem = 0;
5452
    unsigned long long hotplugmem = 0;
5453
    unsigned long long mem;
5454
    unsigned long long align = qemuDomainGetMemorySizeAlignment(def);
5455 5456 5457 5458 5459
    size_t ncells = virDomainNumaGetNodeCount(def->numa);
    size_t i;

    /* align NUMA cell sizes if relevant */
    for (i = 0; i < ncells; i++) {
5460 5461
        mem = VIR_ROUND_UP(virDomainNumaGetNodeMemorySize(def->numa, i), align);
        initialmem += mem;
5462 5463 5464 5465 5466 5467 5468

        if (mem > maxmemkb) {
            virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
                           _("memory size of NUMA node '%zu' overflowed after "
                             "alignment"), i);
            return -1;
        }
5469
        virDomainNumaSetNodeMemorySize(def->numa, i, mem);
5470 5471
    }

5472 5473 5474 5475 5476
    /* align initial memory size, if NUMA is present calculate it as total of
     * individual aligned NUMA node sizes */
    if (initialmem == 0)
        initialmem = VIR_ROUND_UP(virDomainDefGetMemoryInitial(def), align);

5477 5478 5479 5480 5481 5482
    if (initialmem > maxmemcapped) {
        virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                       _("initial memory size overflowed after alignment"));
        return -1;
    }

5483
    def->mem.max_memory = VIR_ROUND_UP(def->mem.max_memory, align);
5484 5485 5486 5487 5488
    if (def->mem.max_memory > maxmemkb) {
        virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                       _("maximum memory size overflowed after alignment"));
        return -1;
    }
5489

5490
    /* Align memory module sizes */
5491 5492
    for (i = 0; i < def->nmems; i++) {
        align = qemuDomainGetMemoryModuleSizeAlignment(def, def->mems[i]);
5493
        def->mems[i]->size = VIR_ROUND_UP(def->mems[i]->size, align);
5494
        hotplugmem += def->mems[i]->size;
5495 5496 5497 5498 5499 5500 5501

        if (def->mems[i]->size > maxmemkb) {
            virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
                           _("size of memory module '%zu' overflowed after "
                             "alignment"), i);
            return -1;
        }
5502
    }
5503

5504 5505
    virDomainDefSetMemoryTotal(def, initialmem + hotplugmem);

5506 5507
    return 0;
}
5508 5509 5510 5511 5512 5513 5514 5515 5516 5517 5518


/**
 * qemuDomainMemoryDeviceAlignSize:
 * @mem: memory device definition object
 *
 * Aligns the size of the memory module as qemu enforces it. The size is updated
 * inplace. Default rounding is now to 1 MiB (qemu requires rouding to page,
 * size so this should be safe).
 */
void
5519 5520
qemuDomainMemoryDeviceAlignSize(virDomainDefPtr def,
                                virDomainMemoryDefPtr mem)
5521
{
5522
    mem->size = VIR_ROUND_UP(mem->size, qemuDomainGetMemorySizeAlignment(def));
5523
}
5524 5525 5526 5527 5528 5529 5530 5531 5532 5533 5534 5535 5536


/**
 * qemuDomainGetMonitor:
 * @vm: domain object
 *
 * Returns the monitor pointer corresponding to the domain object @vm.
 */
qemuMonitorPtr
qemuDomainGetMonitor(virDomainObjPtr vm)
{
    return ((qemuDomainObjPrivatePtr) vm->privateData)->mon;
}
5537 5538 5539 5540 5541 5542 5543 5544 5545 5546 5547 5548 5549 5550 5551 5552


/**
 * qemuDomainSupportsBlockJobs:
 * @vm: domain object
 * @modern: pointer to bool that returns whether modern block jobs are supported
 *
 * Returns -1 in case when qemu does not support block jobs at all. Otherwise
 * returns 0 and optionally fills @modern to denote that modern (async) block
 * jobs are supported.
 */
int
qemuDomainSupportsBlockJobs(virDomainObjPtr vm,
                            bool *modern)
{
    qemuDomainObjPrivatePtr priv = vm->privateData;
5553 5554
    bool asynchronous = virQEMUCapsGet(priv->qemuCaps, QEMU_CAPS_BLOCKJOB_ASYNC);
    bool synchronous = virQEMUCapsGet(priv->qemuCaps, QEMU_CAPS_BLOCKJOB_SYNC);
5555

5556
    if (!synchronous && !asynchronous) {
5557 5558 5559 5560 5561 5562
        virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                       _("block jobs not supported with this QEMU binary"));
        return -1;
    }

    if (modern)
5563
        *modern = asynchronous;
5564 5565 5566

    return 0;
}
5567 5568 5569 5570 5571 5572 5573 5574 5575


/**
 * qemuFindAgentConfig:
 * @def: domain definition
 *
 * Returns the pointer to the channel definition that is used to access the
 * guest agent if the agent is configured or NULL otherwise.
 */
5576
virDomainChrDefPtr
5577 5578 5579 5580 5581 5582 5583 5584 5585 5586
qemuFindAgentConfig(virDomainDefPtr def)
{
    size_t i;

    for (i = 0; i < def->nchannels; i++) {
        virDomainChrDefPtr channel = def->channels[i];

        if (channel->targetType != VIR_DOMAIN_CHR_CHANNEL_TARGET_TYPE_VIRTIO)
            continue;

5587 5588
        if (STREQ_NULLABLE(channel->target.name, "org.qemu.guest_agent.0"))
            return channel;
5589 5590
    }

5591
    return NULL;
5592
}
5593 5594 5595 5596 5597 5598 5599 5600 5601 5602 5603 5604 5605 5606 5607 5608 5609 5610 5611


bool
qemuDomainMachineIsQ35(const virDomainDef *def)
{
    return (STRPREFIX(def->os.machine, "pc-q35") ||
            STREQ(def->os.machine, "q35"));
}


bool
qemuDomainMachineIsI440FX(const virDomainDef *def)
{
    return (STREQ(def->os.machine, "pc") ||
            STRPREFIX(def->os.machine, "pc-0.") ||
            STRPREFIX(def->os.machine, "pc-1.") ||
            STRPREFIX(def->os.machine, "pc-i440") ||
            STRPREFIX(def->os.machine, "rhel"));
}
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 5642 5643
bool
qemuDomainMachineHasPCIRoot(const virDomainDef *def)
{
    int root = virDomainControllerFind(def, VIR_DOMAIN_CONTROLLER_TYPE_PCI, 0);

    if (root < 0)
        return false;

    if (def->controllers[root]->model != VIR_DOMAIN_CONTROLLER_MODEL_PCI_ROOT)
        return false;

    return true;
}


bool
qemuDomainMachineHasPCIeRoot(const virDomainDef *def)
{
    int root = virDomainControllerFind(def, VIR_DOMAIN_CONTROLLER_TYPE_PCI, 0);

    if (root < 0)
        return false;

    if (def->controllers[root]->model != VIR_DOMAIN_CONTROLLER_MODEL_PCIE_ROOT)
        return false;

    return true;
}


5644 5645 5646 5647 5648 5649 5650 5651 5652 5653 5654 5655 5656 5657 5658 5659 5660 5661
bool
qemuDomainMachineNeedsFDC(const virDomainDef *def)
{
    char *p = STRSKIP(def->os.machine, "pc-q35-");

    if (p) {
        if (STRPREFIX(p, "1.") ||
            STRPREFIX(p, "2.0") ||
            STRPREFIX(p, "2.1") ||
            STRPREFIX(p, "2.2") ||
            STRPREFIX(p, "2.3"))
            return false;
        return true;
    }
    return false;
}


5662 5663 5664 5665 5666 5667
bool
qemuDomainMachineIsS390CCW(const virDomainDef *def)
{
    return STRPREFIX(def->os.machine, "s390-ccw");
}

5668

J
Ján Tomko 已提交
5669 5670 5671
bool
qemuDomainMachineIsVirt(const virDomainDef *def)
{
5672 5673 5674 5675 5676 5677 5678 5679 5680
    if (def->os.arch != VIR_ARCH_ARMV7L &&
        def->os.arch != VIR_ARCH_AARCH64)
        return false;

    if (STRNEQ(def->os.machine, "virt") &&
        !STRPREFIX(def->os.machine, "virt-"))
        return false;

    return true;
J
Ján Tomko 已提交
5681 5682 5683
}


5684 5685 5686 5687 5688 5689 5690 5691 5692 5693 5694 5695 5696 5697
bool
qemuDomainMachineIsPSeries(const virDomainDef *def)
{
    if (!ARCH_IS_PPC64(def->os.arch))
        return false;

    if (STRNEQ(def->os.machine, "pseries") &&
        !STRPREFIX(def->os.machine, "pseries-"))
        return false;

    return true;
}


5698 5699 5700 5701 5702 5703 5704 5705 5706 5707 5708 5709 5710 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 5737 5738 5739 5740 5741 5742 5743 5744
static bool
qemuCheckMemoryDimmConflict(const virDomainDef *def,
                            const virDomainMemoryDef *mem)
{
    size_t i;

    for (i = 0; i < def->nmems; i++) {
         virDomainMemoryDefPtr tmp = def->mems[i];

         if (tmp == mem ||
             tmp->info.type != VIR_DOMAIN_DEVICE_ADDRESS_TYPE_DIMM)
             continue;

         if (mem->info.addr.dimm.slot == tmp->info.addr.dimm.slot) {
             virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
                            _("memory device slot '%u' is already being "
                              "used by another memory device"),
                            mem->info.addr.dimm.slot);
             return true;
         }

         if (mem->info.addr.dimm.base != 0 &&
             mem->info.addr.dimm.base == tmp->info.addr.dimm.base) {
             virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
                            _("memory device base '0x%llx' is already being "
                              "used by another memory device"),
                            mem->info.addr.dimm.base);
             return true;
         }
    }

    return false;
}
static int
qemuDomainDefValidateMemoryHotplugDevice(const virDomainMemoryDef *mem,
                                         const virDomainDef *def)
{
    switch ((virDomainMemoryModel) mem->model) {
    case VIR_DOMAIN_MEMORY_MODEL_DIMM:
        if (mem->info.type != VIR_DOMAIN_DEVICE_ADDRESS_TYPE_DIMM &&
            mem->info.type != VIR_DOMAIN_DEVICE_ADDRESS_TYPE_NONE) {
            virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                           _("only 'dimm' addresses are supported for the "
                             "pc-dimm device"));
            return -1;
        }

5745 5746 5747
        if (virDomainNumaGetNodeCount(def->numa) != 0) {
            if (mem->targetNode == -1) {
                virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
Y
Yuri Chornoivan 已提交
5748
                               _("target NUMA node needs to be specified for "
5749 5750 5751
                                 "memory device"));
                return -1;
            }
5752 5753
        }

5754 5755 5756 5757 5758 5759 5760 5761 5762 5763 5764 5765 5766 5767 5768 5769 5770 5771 5772 5773 5774 5775 5776 5777
        if (mem->info.type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_DIMM) {
            if (mem->info.addr.dimm.slot >= def->mem.memory_slots) {
                virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
                               _("memory device slot '%u' exceeds slots "
                                 "count '%u'"),
                               mem->info.addr.dimm.slot, def->mem.memory_slots);
                return -1;
            }


            if (qemuCheckMemoryDimmConflict(def, mem))
                return -1;
        }
        break;

    case VIR_DOMAIN_MEMORY_MODEL_NONE:
    case VIR_DOMAIN_MEMORY_MODEL_LAST:
        return -1;
    }

    return 0;
}


5778 5779 5780 5781 5782 5783 5784 5785 5786 5787 5788 5789 5790 5791 5792 5793 5794 5795 5796 5797 5798 5799 5800 5801 5802 5803 5804 5805
/**
 * qemuDomainDefValidateMemoryHotplug:
 * @def: domain definition
 * @qemuCaps: qemu capabilities object
 * @mem: definition of memory device that is to be added to @def with hotplug,
 *       NULL in case of regular VM startup
 *
 * Validates that the domain definition and memory modules have valid
 * configuration and are possibly able to accept @mem via hotplug if it's
 * non-NULL.
 *
 * Returns 0 on success; -1 and a libvirt error on error.
 */
int
qemuDomainDefValidateMemoryHotplug(const virDomainDef *def,
                                   virQEMUCapsPtr qemuCaps,
                                   const virDomainMemoryDef *mem)
{
    unsigned int nmems = def->nmems;
    unsigned long long hotplugSpace;
    unsigned long long hotplugMemory = 0;
    size_t i;

    hotplugSpace = def->mem.max_memory - virDomainDefGetMemoryInitial(def);

    if (mem) {
        nmems++;
        hotplugMemory = mem->size;
5806 5807 5808

        if (qemuDomainDefValidateMemoryHotplugDevice(mem, def) < 0)
            return -1;
5809 5810 5811 5812 5813 5814 5815 5816 5817 5818 5819 5820 5821 5822 5823 5824 5825 5826 5827
    }

    if (!virDomainDefHasMemoryHotplug(def)) {
        if (nmems) {
            virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                           _("cannot use/hotplug a memory device when domain "
                             "'maxMemory' is not defined"));
            return -1;
        }

        return 0;
    }

    if (!virQEMUCapsGet(qemuCaps, QEMU_CAPS_DEVICE_PC_DIMM)) {
        virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                       _("memory hotplug isn't supported by this QEMU binary"));
        return -1;
    }

5828 5829 5830 5831 5832 5833 5834 5835 5836 5837 5838
    if (!ARCH_IS_PPC64(def->os.arch)) {
        /* due to guest support, qemu would silently enable NUMA with one node
         * once the memory hotplug backend is enabled. To avoid possible
         * confusion we will enforce user originated numa configuration along
         * with memory hotplug. */
        if (virDomainNumaGetNodeCount(def->numa) == 0) {
            virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                           _("At least one numa node has to be configured when "
                             "enabling memory hotplug"));
            return -1;
        }
5839 5840 5841 5842 5843 5844 5845 5846 5847
    }

    if (nmems > def->mem.memory_slots) {
        virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
                       _("memory device count '%u' exceeds slots count '%u'"),
                       nmems, def->mem.memory_slots);
        return -1;
    }

5848
    for (i = 0; i < def->nmems; i++) {
5849 5850
        hotplugMemory += def->mems[i]->size;

5851 5852 5853 5854 5855 5856
        /* already existing devices don't need to be checked on hotplug */
        if (!mem &&
            qemuDomainDefValidateMemoryHotplugDevice(def->mems[i], def) < 0)
            return -1;
    }

5857 5858 5859 5860 5861 5862 5863 5864 5865 5866
    if (hotplugMemory > hotplugSpace) {
        virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                       _("memory device total size exceeds hotplug space"));
        return -1;
    }

    return 0;
}


5867 5868 5869 5870 5871 5872 5873 5874 5875 5876
bool
qemuDomainMachineHasBuiltinIDE(const virDomainDef *def)
{
    return qemuDomainMachineIsI440FX(def) ||
        STREQ(def->os.machine, "malta") ||
        STREQ(def->os.machine, "sun4u") ||
        STREQ(def->os.machine, "g3beige");
}


5877 5878 5879 5880 5881 5882 5883 5884 5885 5886 5887 5888 5889 5890 5891 5892 5893 5894 5895 5896 5897 5898 5899 5900
/**
 * qemuDomainUpdateCurrentMemorySize:
 *
 * Updates the current balloon size from the monitor if necessary. In case when
 * the balloon is not present for the domain, the function recalculates the
 * maximum size to reflect possible changes.
 *
 * Returns 0 on success and updates vm->def->mem.cur_balloon if necessary, -1 on
 * error and reports libvirt error.
 */
int
qemuDomainUpdateCurrentMemorySize(virQEMUDriverPtr driver,
                                  virDomainObjPtr vm)
{
    qemuDomainObjPrivatePtr priv = vm->privateData;
    unsigned long long balloon;
    int ret = -1;

    /* inactive domain doesn't need size update */
    if (!virDomainObjIsActive(vm))
        return 0;

    /* if no balloning is available, the current size equals to the current
     * full memory size */
5901
    if (!virDomainDefHasMemballoon(vm->def)) {
5902
        vm->def->mem.cur_balloon = virDomainDefGetMemoryTotal(vm->def);
5903 5904 5905 5906 5907 5908 5909 5910 5911 5912 5913 5914 5915 5916 5917 5918 5919 5920 5921 5922 5923 5924 5925 5926 5927 5928 5929 5930 5931 5932 5933 5934
        return 0;
    }

    /* current size is always automagically updated via the event */
    if (virQEMUCapsGet(priv->qemuCaps, QEMU_CAPS_BALLOON_EVENT))
        return 0;

    /* here we need to ask the monitor */

    /* Don't delay if someone's using the monitor, just use existing most
     * recent data instead */
    if (qemuDomainJobAllowed(priv, QEMU_JOB_QUERY)) {
        if (qemuDomainObjBeginJob(driver, vm, QEMU_JOB_QUERY) < 0)
            return -1;

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

        qemuDomainObjEnterMonitor(driver, vm);
        ret = qemuMonitorGetBalloonInfo(priv->mon, &balloon);
        if (qemuDomainObjExitMonitor(driver, vm) < 0)
            ret = -1;

 endjob:
        qemuDomainObjEndJob(driver, vm);

        if (ret < 0)
            return -1;

5935 5936
        vm->def->mem.cur_balloon = balloon;
    }
5937 5938 5939

    return 0;
}
5940 5941 5942


/**
5943
 * qemuDomainGetMemLockLimitBytes:
5944 5945 5946 5947
 *
 * @def: domain definition
 *
 * Returns the size of the memory in bytes that needs to be set as
5948
 * RLIMIT_MEMLOCK for the QEMU process.
5949 5950 5951 5952
 * If a mem.hard_limit is set, then that value is preferred; otherwise, the
 * value returned may depend upon the architecture or devices present.
 */
unsigned long long
5953
qemuDomainGetMemLockLimitBytes(virDomainDefPtr def)
5954 5955 5956
{
    unsigned long long memKB;

5957 5958 5959 5960 5961 5962
    /* prefer the hard limit */
    if (virMemoryLimitIsSet(def->mem.hard_limit)) {
        memKB = def->mem.hard_limit;
        goto done;
    }

5963 5964 5965 5966 5967 5968 5969 5970 5971 5972 5973 5974 5975 5976 5977 5978 5979 5980 5981 5982 5983 5984 5985 5986
    if (ARCH_IS_PPC64(def->os.arch)) {
        unsigned long long maxMemory;
        unsigned long long memory;
        unsigned long long baseLimit;
        unsigned long long passthroughLimit;
        size_t nPCIHostBridges;
        size_t i;
        bool usesVFIO = false;

        /* TODO: Detect at runtime once we start using more than just
         *       the default PCI Host Bridge */
        nPCIHostBridges = 1;

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

            if (dev->mode == VIR_DOMAIN_HOSTDEV_MODE_SUBSYS &&
                dev->source.subsys.type == VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_PCI &&
                dev->source.subsys.u.pci.backend == VIR_DOMAIN_HOSTDEV_PCI_BACKEND_VFIO) {
                usesVFIO = true;
                break;
            }
        }

5987
        memory = virDomainDefGetMemoryTotal(def);
5988 5989 5990 5991 5992 5993 5994 5995 5996 5997 5998 5999 6000 6001 6002 6003 6004 6005 6006 6007 6008 6009 6010 6011 6012 6013 6014 6015 6016 6017 6018 6019 6020 6021 6022 6023 6024 6025 6026 6027 6028 6029 6030 6031 6032 6033 6034 6035 6036 6037 6038 6039 6040

        if (def->mem.max_memory)
            maxMemory = def->mem.max_memory;
        else
            maxMemory = memory;

        /* baseLimit := maxMemory / 128                                  (a)
         *              + 4 MiB * #PHBs + 8 MiB                          (b)
         *
         * (a) is the hash table
         *
         * (b) is accounting for the 32-bit DMA window - it could be either the
         * KVM accelerated TCE tables for emulated devices, or the VFIO
         * userspace view. The 4 MiB per-PHB (including the default one) covers
         * a 2GiB DMA window: default is 1GiB, but it's possible it'll be
         * increased to help performance. The 8 MiB extra should be plenty for
         * the TCE table index for any reasonable number of PHBs and several
         * spapr-vlan or spapr-vscsi devices (512kB + a tiny bit each) */
        baseLimit = maxMemory / 128 +
                    4096 * nPCIHostBridges +
                    8192;

        /* passthroughLimit := max( 2 GiB * #PHBs,                       (c)
         *                          memory                               (d)
         *                          + memory * 1/512 * #PHBs + 8 MiB )   (e)
         *
         * (c) is the pre-DDW VFIO DMA window accounting. We're allowing 2 GiB
         * rather than 1 GiB
         *
         * (d) is the with-DDW (and memory pre-registration and related
         * features) DMA window accounting - assuming that we only account RAM
         * once, even if mapped to multiple PHBs
         *
         * (e) is the with-DDW userspace view and overhead for the 64-bit DMA
         * window. This is based a bit on expected guest behaviour, but there
         * really isn't a way to completely avoid that. We assume the guest
         * requests a 64-bit DMA window (per PHB) just big enough to map all
         * its RAM. 4 kiB page size gives the 1/512; it will be less with 64
         * kiB pages, less still if the guest is mapped with hugepages (unlike
         * the default 32-bit DMA window, DDW windows can use large IOMMU
         * pages). 8 MiB is for second and further level overheads, like (b) */
        passthroughLimit = MAX(2 * 1024 * 1024 * nPCIHostBridges,
                               memory +
                               memory / 512 * nPCIHostBridges + 8192);

        if (usesVFIO)
            memKB = baseLimit + passthroughLimit;
        else
            memKB = baseLimit;

        goto done;
    }

6041 6042 6043 6044 6045 6046 6047 6048 6049 6050 6051 6052 6053 6054 6055 6056 6057 6058
    /* For device passthrough using VFIO the guest memory and MMIO memory
     * regions need to be locked persistent in order to allow DMA.
     *
     * Currently the below limit is based on assumptions about the x86 platform.
     *
     * The chosen value of 1GiB below originates from x86 systems where it was
     * used as space reserved for the MMIO region for the whole system.
     *
     * On x86_64 systems the MMIO regions of the IOMMU mapped devices don't
     * count towards the locked memory limit since the memory is owned by the
     * device. Emulated devices though do count, but the regions are usually
     * small. Although it's not guaranteed that the limit will be enough for all
     * configurations it didn't pose a problem for now.
     *
     * http://www.redhat.com/archives/libvir-list/2015-November/msg00329.html
     *
     * Note that this may not be valid for all platforms.
     */
6059
    memKB = virDomainDefGetMemoryTotal(def) + 1024 * 1024;
6060

6061
 done:
6062 6063
    return memKB << 10;
}
6064 6065 6066 6067 6068


/**
 * @def: domain definition
 *
6069 6070 6071
 * Returns true if the locked memory limit needs to be set or updated because
 * of domain configuration, VFIO passthrough devices or architecture-specific
 * requirements.
6072 6073
 * */
bool
6074
qemuDomainRequiresMemLock(virDomainDefPtr def)
6075 6076 6077 6078 6079 6080 6081 6082 6083 6084 6085 6086 6087 6088 6089
{
    size_t i;

    if (def->mem.locked)
        return true;

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

        if (dev->mode == VIR_DOMAIN_HOSTDEV_MODE_SUBSYS &&
            dev->source.subsys.type == VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_PCI &&
            dev->source.subsys.u.pci.backend == VIR_DOMAIN_HOSTDEV_PCI_BACKEND_VFIO)
            return true;
    }

6090 6091 6092 6093
    /* ppc64 KVM domains need to lock some memory even when VFIO is not used */
    if (ARCH_IS_PPC64(def->os.arch) && def->virtType == VIR_DOMAIN_VIRT_KVM)
        return true;

6094 6095
    return false;
}
6096

6097 6098 6099 6100 6101 6102 6103
/**
 * qemuDomainAdjustMaxMemLock:
 * @vm: domain
 *
 * Adjust the memory locking limit for the QEMU process associated to @vm, in
 * order to comply with VFIO or architecture requirements.
 *
6104 6105 6106 6107
 * The limit will not be changed unless doing so is needed; the first time
 * the limit is changed, the original (default) limit is stored in @vm and
 * that value will be restored if qemuDomainAdjustMaxMemLock() is called once
 * memory locking is no longer required.
6108 6109 6110 6111 6112 6113 6114 6115 6116
 *
 * Returns: 0 on success, <0 on failure
 */
int
qemuDomainAdjustMaxMemLock(virDomainObjPtr vm)
{
    unsigned long long bytes = 0;
    int ret = -1;

6117
    if (qemuDomainRequiresMemLock(vm->def)) {
6118 6119 6120 6121 6122 6123 6124 6125
        /* If this is the first time adjusting the limit, save the current
         * value so that we can restore it once memory locking is no longer
         * required. Failing to obtain the current limit is not a critical
         * failure, it just means we'll be unable to lower it later */
        if (!vm->original_memlock) {
            if (virProcessGetMaxMemLock(vm->pid, &(vm->original_memlock)) < 0)
                vm->original_memlock = 0;
        }
6126
        bytes = qemuDomainGetMemLockLimitBytes(vm->def);
6127 6128 6129 6130 6131 6132
    } else {
        /* Once memory locking is no longer required, we can restore the
         * original, usually very low, limit */
        bytes = vm->original_memlock;
        vm->original_memlock = 0;
    }
6133 6134 6135 6136 6137 6138 6139 6140 6141 6142

    /* Trying to set the memory locking limit to zero is a no-op */
    if (virProcessSetMaxMemLock(vm->pid, bytes) < 0)
        goto out;

    ret = 0;

 out:
     return ret;
}
6143 6144 6145 6146 6147 6148 6149 6150 6151 6152

/**
 * qemuDomainHasVcpuPids:
 * @vm: Domain object
 *
 * Returns true if we were able to successfully detect vCPU pids for the VM.
 */
bool
qemuDomainHasVcpuPids(virDomainObjPtr vm)
{
6153 6154 6155 6156 6157 6158
    size_t i;
    size_t maxvcpus = virDomainDefGetVcpusMax(vm->def);
    virDomainVcpuDefPtr vcpu;

    for (i = 0; i < maxvcpus; i++) {
        vcpu = virDomainDefGetVcpu(vm->def, i);
6159

6160 6161 6162 6163 6164
        if (QEMU_DOMAIN_VCPU_PRIVATE(vcpu)->tid > 0)
            return true;
    }

    return false;
6165
}
6166 6167 6168 6169 6170 6171 6172 6173 6174 6175 6176


/**
 * qemuDomainGetVcpuPid:
 * @vm: domain object
 * @vcpu: cpu id
 *
 * Returns the vCPU pid. If @vcpu is offline or out of range 0 is returned.
 */
pid_t
qemuDomainGetVcpuPid(virDomainObjPtr vm,
6177
                     unsigned int vcpuid)
6178
{
6179 6180
    virDomainVcpuDefPtr vcpu = virDomainDefGetVcpu(vm->def, vcpuid);
    return QEMU_DOMAIN_VCPU_PRIVATE(vcpu)->tid;
6181
}
6182 6183


6184 6185 6186 6187 6188 6189 6190 6191 6192 6193 6194 6195 6196 6197 6198 6199 6200 6201 6202 6203 6204 6205 6206 6207 6208 6209 6210 6211 6212 6213 6214 6215 6216 6217 6218 6219 6220 6221 6222 6223 6224 6225
/**
 * qemuDomainValidateVcpuInfo:
 *
 * Validates vcpu thread information. If vcpu thread IDs are reported by qemu,
 * this function validates that online vcpus have thread info present and
 * offline vcpus don't.
 *
 * Returns 0 on success -1 on error.
 */
int
qemuDomainValidateVcpuInfo(virDomainObjPtr vm)
{
    size_t maxvcpus = virDomainDefGetVcpusMax(vm->def);
    virDomainVcpuDefPtr vcpu;
    qemuDomainVcpuPrivatePtr vcpupriv;
    size_t i;

    if (!qemuDomainHasVcpuPids(vm))
        return 0;

    for (i = 0; i < maxvcpus; i++) {
        vcpu = virDomainDefGetVcpu(vm->def, i);
        vcpupriv = QEMU_DOMAIN_VCPU_PRIVATE(vcpu);

        if (vcpu->online && vcpupriv->tid == 0) {
            virReportError(VIR_ERR_INTERNAL_ERROR,
                           _("qemu didn't report thread id for vcpu '%zu'"), i);
            return -1;
        }

        if (!vcpu->online && vcpupriv->tid != 0) {
            virReportError(VIR_ERR_INTERNAL_ERROR,
                           _("qemu reported thread id for inactive vcpu '%zu'"),
                           i);
            return -1;
        }
    }

    return 0;
}


6226 6227 6228 6229 6230 6231 6232 6233 6234
bool
qemuDomainSupportsNewVcpuHotplug(virDomainObjPtr vm)
{
    qemuDomainObjPrivatePtr priv = vm->privateData;

    return virQEMUCapsGet(priv->qemuCaps, QEMU_CAPS_QUERY_HOTPLUGGABLE_CPUS);
}


6235
/**
6236
 * qemuDomainRefreshVcpuInfo:
6237 6238 6239
 * @driver: qemu driver data
 * @vm: domain object
 * @asyncJob: current asynchronous job type
6240
 * @state: refresh vcpu state
6241
 *
6242 6243
 * Updates vCPU information private data of @vm. Due to historical reasons this
 * function returns success even if some data were not reported by qemu.
6244
 *
6245 6246
 * If @state is true, the vcpu state is refreshed as reported by the monitor.
 *
6247
 * Returns 0 on success and -1 on fatal error.
6248 6249
 */
int
6250 6251
qemuDomainRefreshVcpuInfo(virQEMUDriverPtr driver,
                          virDomainObjPtr vm,
6252 6253
                          int asyncJob,
                          bool state)
6254
{
6255
    virDomainVcpuDefPtr vcpu;
6256 6257
    qemuDomainVcpuPrivatePtr vcpupriv;
    qemuMonitorCPUInfoPtr info = NULL;
6258 6259
    size_t maxvcpus = virDomainDefGetVcpusMax(vm->def);
    size_t i;
6260
    bool hotplug;
6261
    int rc;
6262
    int ret = -1;
6263

6264
    hotplug = qemuDomainSupportsNewVcpuHotplug(vm);
6265 6266 6267

    if (qemuDomainObjEnterMonitorAsync(driver, vm, asyncJob) < 0)
        return -1;
6268

6269
    rc = qemuMonitorGetCPUInfo(qemuDomainGetMonitor(vm), &info, maxvcpus, hotplug);
6270

6271
    if (qemuDomainObjExitMonitor(driver, vm) < 0)
6272 6273
        goto cleanup;

6274
    if (rc < 0)
6275 6276 6277 6278
        goto cleanup;

    for (i = 0; i < maxvcpus; i++) {
        vcpu = virDomainDefGetVcpu(vm->def, i);
6279
        vcpupriv = QEMU_DOMAIN_VCPU_PRIVATE(vcpu);
6280

6281 6282 6283 6284 6285 6286 6287 6288 6289 6290 6291 6292 6293 6294 6295 6296 6297 6298 6299 6300 6301 6302 6303 6304 6305 6306 6307 6308 6309 6310 6311 6312 6313 6314 6315 6316 6317 6318 6319 6320
        /*
         * Current QEMU *can* report info about host threads mapped
         * to vCPUs, but it is not in a manner we can correctly
         * deal with. The TCG CPU emulation does have a separate vCPU
         * thread, but it runs every vCPU in that same thread. So it
         * is impossible to setup different affinity per thread.
         *
         * What's more the 'query-cpus' command returns bizarre
         * data for the threads. It gives the TCG thread for the
         * vCPU 0, but for vCPUs 1-> N, it actually replies with
         * the main process thread ID.
         *
         * The result is that when we try to set affinity for
         * vCPU 1, it will actually change the affinity of the
         * emulator thread :-( When you try to set affinity for
         * vCPUs 2, 3.... it will fail if the affinity was
         * different from vCPU 1.
         *
         * We *could* allow vcpu pinning with TCG, if we made the
         * restriction that all vCPUs had the same mask. This would
         * at least let us separate emulator from vCPUs threads, as
         * we do for KVM. It would need some changes to our cgroups
         * CPU layout though, and error reporting for the config
         * restrictions.
         *
         * Just disable CPU pinning with TCG until someone wants
         * to try to do this hard work.
         */
        if (vm->def->virtType != VIR_DOMAIN_VIRT_QEMU)
            vcpupriv->tid = info[i].tid;

        vcpupriv->socket_id = info[i].socket_id;
        vcpupriv->core_id = info[i].core_id;
        vcpupriv->thread_id = info[i].thread_id;
        vcpupriv->vcpus = info[i].vcpus;
        VIR_FREE(vcpupriv->type);
        VIR_STEAL_PTR(vcpupriv->type, info[i].type);
        VIR_FREE(vcpupriv->alias);
        VIR_STEAL_PTR(vcpupriv->alias, info[i].alias);
        vcpupriv->enable_id = info[i].id;
6321
        vcpupriv->qemu_id = info[i].qemu_id;
6322

6323
        if (hotplug && state) {
6324 6325 6326 6327 6328
            vcpu->online = info[i].online;
            if (info[i].hotpluggable)
                vcpu->hotpluggable = VIR_TRISTATE_BOOL_YES;
            else
                vcpu->hotpluggable = VIR_TRISTATE_BOOL_NO;
6329
        }
6330 6331
    }

6332
    ret = 0;
6333 6334

 cleanup:
6335
    qemuMonitorCPUInfoFree(info, maxvcpus);
6336
    return ret;
6337
}
6338

6339 6340 6341 6342 6343 6344 6345 6346 6347 6348 6349 6350 6351 6352 6353 6354 6355 6356 6357 6358 6359 6360 6361 6362 6363 6364 6365 6366 6367 6368 6369
/**
 * qemuDomainGetVcpuHalted:
 * @vm: domain object
 * @vcpu: cpu id
 *
 * Returns the vCPU halted state.
  */
bool
qemuDomainGetVcpuHalted(virDomainObjPtr vm,
                        unsigned int vcpuid)
{
    virDomainVcpuDefPtr vcpu = virDomainDefGetVcpu(vm->def, vcpuid);
    return QEMU_DOMAIN_VCPU_PRIVATE(vcpu)->halted;
}

/**
 * qemuDomainRefreshVcpuHalted:
 * @driver: qemu driver data
 * @vm: domain object
 * @asyncJob: current asynchronous job type
 *
 * Updates vCPU halted state in the private data of @vm.
 *
 * Returns 0 on success and -1 on error
 */
int
qemuDomainRefreshVcpuHalted(virQEMUDriverPtr driver,
                            virDomainObjPtr vm,
                            int asyncJob)
{
    virDomainVcpuDefPtr vcpu;
6370
    qemuDomainVcpuPrivatePtr vcpupriv;
6371
    size_t maxvcpus = virDomainDefGetVcpusMax(vm->def);
6372
    virBitmapPtr haltedmap = NULL;
6373 6374 6375 6376 6377 6378 6379 6380 6381 6382
    size_t i;
    int ret = -1;

    /* Not supported currently for TCG, see qemuDomainRefreshVcpuInfo */
    if (vm->def->virtType == VIR_DOMAIN_VIRT_QEMU)
        return 0;

    if (qemuDomainObjEnterMonitorAsync(driver, vm, asyncJob) < 0)
        return -1;

6383
    haltedmap = qemuMonitorGetCpuHalted(qemuDomainGetMonitor(vm), maxvcpus);
6384

6385
    if (qemuDomainObjExitMonitor(driver, vm) < 0 || !haltedmap)
6386 6387 6388 6389
        goto cleanup;

    for (i = 0; i < maxvcpus; i++) {
        vcpu = virDomainDefGetVcpu(vm->def, i);
6390 6391
        vcpupriv = QEMU_DOMAIN_VCPU_PRIVATE(vcpu);
        vcpupriv->halted = virBitmapIsBitSet(haltedmap, vcpupriv->qemu_id);
6392 6393 6394 6395 6396
    }

    ret = 0;

 cleanup:
6397
    virBitmapFree(haltedmap);
6398 6399
    return ret;
}
6400 6401 6402 6403 6404 6405 6406 6407 6408 6409 6410 6411 6412 6413 6414 6415 6416 6417 6418 6419

bool
qemuDomainSupportsNicdev(virDomainDefPtr def,
                         virDomainNetDefPtr net)
{
    /* non-virtio ARM nics require legacy -net nic */
    if (((def->os.arch == VIR_ARCH_ARMV7L) ||
        (def->os.arch == VIR_ARCH_AARCH64)) &&
        net->info.type != VIR_DOMAIN_DEVICE_ADDRESS_TYPE_VIRTIO_MMIO &&
        net->info.type != VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI)
        return false;

    return true;
}

bool
qemuDomainSupportsNetdev(virDomainDefPtr def,
                         virQEMUCapsPtr qemuCaps,
                         virDomainNetDefPtr net)
{
6420
    if (!qemuDomainSupportsNicdev(def, net))
6421 6422 6423
        return false;
    return virQEMUCapsGet(qemuCaps, QEMU_CAPS_NETDEV);
}
J
John Ferlan 已提交
6424 6425 6426 6427 6428 6429 6430


int
qemuDomainNetVLAN(virDomainNetDefPtr def)
{
    return qemuDomainDeviceAliasIndex(&def->info, "net");
}
P
Peter Krempa 已提交
6431 6432 6433 6434 6435 6436 6437 6438 6439 6440 6441 6442 6443 6444 6445 6446


virDomainDiskDefPtr
qemuDomainDiskByName(virDomainDefPtr def,
                     const char *name)
{
    virDomainDiskDefPtr ret;

    if (!(ret = virDomainDiskByName(def, name, true))) {
        virReportError(VIR_ERR_INVALID_ARG, "%s",
                       _("No device found for specified path"));
        return NULL;
    }

    return ret;
}
6447 6448 6449 6450 6451 6452 6453 6454 6455 6456 6457 6458 6459 6460 6461 6462 6463 6464 6465 6466 6467 6468 6469 6470 6471 6472 6473 6474 6475 6476 6477


/**
 * qemuDomainDefValidateDiskLunSource:
 * @src: disk source struct
 *
 * Validate whether the disk source is valid for disk device='lun'.
 *
 * Returns 0 if the configuration is valid -1 and a libvirt error if the soure
 * is invalid.
 */
int
qemuDomainDefValidateDiskLunSource(const virStorageSource *src)
{
    if (virStorageSourceGetActualType(src) == VIR_STORAGE_TYPE_NETWORK) {
        if (src->protocol != VIR_STORAGE_NET_PROTOCOL_ISCSI) {
            virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
                           _("disk device='lun' is not supported "
                             "for protocol='%s'"),
                           virStorageNetProtocolTypeToString(src->protocol));
            return -1;
        }
    } else if (!virStorageSourceIsBlockLocal(src)) {
        virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                       _("disk device='lun' is only valid for block "
                         "type disk source"));
        return -1;
    }

    return 0;
}
6478 6479 6480 6481 6482 6483 6484


int
qemuDomainPrepareChannel(virDomainChrDefPtr channel,
                         const char *domainChannelTargetDir)
{
    if (channel->targetType == VIR_DOMAIN_CHR_CHANNEL_TARGET_TYPE_VIRTIO &&
6485 6486 6487
        channel->source->type == VIR_DOMAIN_CHR_TYPE_UNIX &&
        !channel->source->data.nix.path) {
        if (virAsprintf(&channel->source->data.nix.path,
6488 6489 6490 6491 6492
                        "%s/%s", domainChannelTargetDir,
                        channel->target.name ? channel->target.name
                        : "unknown.sock") < 0)
            return -1;

6493
        channel->source->data.nix.listen = true;
6494 6495 6496 6497
    }

    return 0;
}
6498 6499


6500 6501 6502 6503 6504 6505 6506 6507 6508 6509 6510 6511 6512 6513 6514 6515 6516 6517
/* qemuProcessPrepareDomainChardevSourceTLS:
 * @source: pointer to host interface data for char devices
 * @cfg: driver configuration
 *
 * Updates host interface TLS encryption setting based on qemu.conf
 * for char devices.  This will be presented as "tls='yes|no'" in
 * live XML of a guest.
 */
void
qemuDomainPrepareChardevSourceTLS(virDomainChrSourceDefPtr source,
                                  virQEMUDriverConfigPtr cfg)
{
    if (source->type == VIR_DOMAIN_CHR_TYPE_TCP) {
        if (source->data.tcp.haveTLS == VIR_TRISTATE_BOOL_ABSENT) {
            if (cfg->chardevTLS)
                source->data.tcp.haveTLS = VIR_TRISTATE_BOOL_YES;
            else
                source->data.tcp.haveTLS = VIR_TRISTATE_BOOL_NO;
6518
            source->data.tcp.tlsFromConfig = true;
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 6555 6556 6557 6558 6559 6560 6561 6562 6563 6564 6565 6566
        }
    }
}


/* qemuProcessPrepareDomainChardevSource:
 * @def: live domain definition
 * @driver: qemu driver
 *
 * Iterate through all devices that use virDomainChrSourceDefPtr as host
 * interface part.
 */
void
qemuDomainPrepareChardevSource(virDomainDefPtr def,
                               virQEMUDriverPtr driver)
{
    size_t i;
    virQEMUDriverConfigPtr cfg = virQEMUDriverGetConfig(driver);

    for (i = 0; i < def->nserials; i++)
        qemuDomainPrepareChardevSourceTLS(def->serials[i]->source, cfg);

    for (i = 0; i < def->nparallels; i++)
        qemuDomainPrepareChardevSourceTLS(def->parallels[i]->source, cfg);

    for (i = 0; i < def->nchannels; i++)
        qemuDomainPrepareChardevSourceTLS(def->channels[i]->source, cfg);

    for (i = 0; i < def->nconsoles; i++)
        qemuDomainPrepareChardevSourceTLS(def->consoles[i]->source, cfg);

    for (i = 0; i < def->nrngs; i++)
        if (def->rngs[i]->backend == VIR_DOMAIN_RNG_BACKEND_EGD)
            qemuDomainPrepareChardevSourceTLS(def->rngs[i]->source.chardev, cfg);

    for (i = 0; i < def->nsmartcards; i++)
        if (def->smartcards[i]->type == VIR_DOMAIN_SMARTCARD_TYPE_PASSTHROUGH)
            qemuDomainPrepareChardevSourceTLS(def->smartcards[i]->data.passthru,
                                              cfg);

    for (i = 0; i < def->nredirdevs; i++)
        qemuDomainPrepareChardevSourceTLS(def->redirdevs[i]->source, cfg);

    virObjectUnref(cfg);
}



6567 6568 6569 6570 6571 6572 6573 6574 6575 6576 6577 6578 6579
int
qemuDomainPrepareShmemChardev(virDomainShmemDefPtr shmem)
{
    if (!shmem->server.enabled ||
        shmem->server.chr.data.nix.path)
        return 0;

    return virAsprintf(&shmem->server.chr.data.nix.path,
                       "/var/lib/libvirt/shmem-%s-sock",
                       shmem->name);
}


6580 6581 6582 6583 6584 6585 6586 6587 6588 6589 6590 6591 6592 6593 6594 6595 6596 6597 6598 6599 6600 6601 6602 6603 6604 6605 6606 6607 6608 6609 6610 6611 6612
/**
 * qemuDomainVcpuHotplugIsInOrder:
 * @def: domain definition
 *
 * Returns true if online vcpus were added in order (clustered behind vcpu0
 * with increasing order).
 */
bool
qemuDomainVcpuHotplugIsInOrder(virDomainDefPtr def)
{
    size_t maxvcpus = virDomainDefGetVcpusMax(def);
    virDomainVcpuDefPtr vcpu;
    unsigned int prevorder = 0;
    size_t seenonlinevcpus = 0;
    size_t i;

    for (i = 0; i < maxvcpus; i++) {
        vcpu = virDomainDefGetVcpu(def, i);

        if (!vcpu->online)
            break;

        if (vcpu->order < prevorder)
            break;

        if (vcpu->order > prevorder)
            prevorder = vcpu->order;

        seenonlinevcpus++;
    }

    return seenonlinevcpus == virDomainDefGetVcpus(def);
}
6613 6614 6615 6616 6617 6618 6619 6620 6621 6622 6623 6624 6625 6626 6627 6628 6629 6630 6631 6632 6633 6634 6635 6636 6637 6638 6639 6640 6641 6642 6643 6644 6645 6646


/**
 * qemuDomainVcpuPersistOrder:
 * @def: domain definition
 *
 * Saves the order of vcpus detected from qemu to the domain definition.
 * The private data note the order only for the entry describing the
 * hotpluggable entity. This function copies the order into the definition part
 * of all sub entities.
 */
void
qemuDomainVcpuPersistOrder(virDomainDefPtr def)
{
    size_t maxvcpus = virDomainDefGetVcpusMax(def);
    virDomainVcpuDefPtr vcpu;
    qemuDomainVcpuPrivatePtr vcpupriv;
    unsigned int prevorder = 0;
    size_t i;

    for (i = 0; i < maxvcpus; i++) {
        vcpu = virDomainDefGetVcpu(def, i);
        vcpupriv = QEMU_DOMAIN_VCPU_PRIVATE(vcpu);

        if (!vcpu->online) {
            vcpu->order = 0;
        } else {
            if (vcpupriv->enable_id != 0)
                prevorder = vcpupriv->enable_id;

            vcpu->order = prevorder;
        }
    }
}
6647 6648 6649 6650 6651 6652 6653 6654 6655 6656 6657 6658 6659 6660 6661 6662 6663 6664 6665 6666


int
qemuDomainCheckMonitor(virQEMUDriverPtr driver,
                       virDomainObjPtr vm,
                       qemuDomainAsyncJob asyncJob)
{
    qemuDomainObjPrivatePtr priv = vm->privateData;
    int ret;

    if (qemuDomainObjEnterMonitorAsync(driver, vm, asyncJob) < 0)
        return -1;

    ret = qemuMonitorCheck(priv->mon);

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

    return ret;
}
6667 6668 6669 6670 6671 6672 6673 6674 6675 6676 6677 6678


bool
qemuDomainSupportsVideoVga(virDomainVideoDefPtr video,
                           virQEMUCapsPtr qemuCaps)
{
    if (video->type == VIR_DOMAIN_VIDEO_TYPE_VIRTIO &&
        !virQEMUCapsGet(qemuCaps, QEMU_CAPS_DEVICE_VIRTIO_VGA))
        return false;

    return true;
}