qemu_domain_address.c 112.3 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41
/*
 * qemu_domain_address.c: QEMU domain address
 *
 * Copyright (C) 2006-2016 Red Hat, Inc.
 * Copyright (C) 2006 Daniel P. Berrange
 *
 * This library is free software; you can redistribute it and/or
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
 * version 2.1 of the License, or (at your option) any later version.
 *
 * This library is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * Lesser General Public License for more details.
 *
 * You should have received a copy of the GNU Lesser General Public
 * License along with this library.  If not, see
 * <http://www.gnu.org/licenses/>.
 *
 * Author: Daniel P. Berrange <berrange@redhat.com>
 */

#include <config.h>

#include "qemu_domain_address.h"
#include "qemu_domain.h"
#include "viralloc.h"
#include "virerror.h"
#include "virlog.h"

#define VIR_FROM_THIS VIR_FROM_QEMU

VIR_LOG_INIT("qemu.qemu_domain_address");

#define VIO_ADDR_NET 0x1000ul
#define VIO_ADDR_SCSI 0x2000ul
#define VIO_ADDR_SERIAL 0x30000000ul
#define VIO_ADDR_NVRAM 0x3000ul


42 43
/**
 * @def: Domain definition
44
 * @cont: Domain controller def
45 46
 * @qemuCaps: qemu capabilities
 *
47 48
 * If the controller model is already defined, return it immediately;
 * otherwise, based on the @qemuCaps return a default model value.
49
 *
50
 * Returns model on success, -1 on failure with error set.
51
 */
52
int
53 54 55
qemuDomainGetSCSIControllerModel(const virDomainDef *def,
                                 const virDomainControllerDef *cont,
                                 virQEMUCapsPtr qemuCaps)
56
{
57 58 59 60 61
    if (cont->model > 0)
        return cont->model;

    if (qemuDomainIsPSeries(def))
        return VIR_DOMAIN_CONTROLLER_MODEL_SCSI_IBMVSCSI;
62 63
    else if (ARCH_IS_S390(def->os.arch))
        return VIR_DOMAIN_CONTROLLER_MODEL_SCSI_VIRTIO_SCSI;
64 65 66 67 68 69 70 71 72
    else if (virQEMUCapsGet(qemuCaps, QEMU_CAPS_SCSI_LSI))
        return VIR_DOMAIN_CONTROLLER_MODEL_SCSI_LSILOGIC;
    else if (virQEMUCapsGet(qemuCaps, QEMU_CAPS_VIRTIO_SCSI))
        return VIR_DOMAIN_CONTROLLER_MODEL_SCSI_VIRTIO_SCSI;

    virReportError(VIR_ERR_INTERNAL_ERROR,
                   _("Unable to determine model for SCSI controller idx=%d"),
                   cont->idx);
    return -1;
73 74 75
}


76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100
/**
 * @def: Domain definition
 * @cont: Domain controller def
 * @qemuCaps: qemu capabilities
 *
 * Set the controller model based on the existing value and the
 * capabilities if possible.
 *
 * Returns 0 on success, -1 on failure with error set.
 */
int
qemuDomainSetSCSIControllerModel(const virDomainDef *def,
                                 virDomainControllerDefPtr cont,
                                 virQEMUCapsPtr qemuCaps)
{
    int model = qemuDomainGetSCSIControllerModel(def, cont, qemuCaps);

    if (model < 0)
        return -1;

    cont->model = model;
    return 0;
}


101 102 103 104 105 106 107 108 109 110 111
/**
 * @def: Domain definition
 * @info: Domain device info
 *
 * Using the device info, find the controller related to the
 * device by index and use that controller to return the model.
 *
 * Returns the model if found, -1 if not with an error message set
 */
int
qemuDomainFindSCSIControllerModel(const virDomainDef *def,
112
                                  virDomainDeviceInfoPtr info)
113 114 115 116 117 118 119 120 121 122
{
    virDomainControllerDefPtr cont;

    if (!(cont = virDomainDeviceFindSCSIController(def, info))) {
        virReportError(VIR_ERR_INTERNAL_ERROR,
                       _("unable to find a SCSI controller for idx=%d"),
                       info->addr.drive.controller);
        return -1;
    }

123
    return cont->model;
124 125 126
}


127
static int
128
qemuDomainAssignVirtioSerialAddresses(virDomainDefPtr def)
129 130 131 132 133
{
    int ret = -1;
    size_t i;
    virDomainVirtioSerialAddrSetPtr addrs = NULL;

134
    if (!(addrs = virDomainVirtioSerialAddrSetCreateFromDomain(def)))
135 136 137 138 139 140 141 142 143
        goto cleanup;

    VIR_DEBUG("Finished reserving existing ports");

    for (i = 0; i < def->nconsoles; i++) {
        virDomainChrDefPtr chr = def->consoles[i];
        if (chr->deviceType == VIR_DOMAIN_CHR_DEVICE_TYPE_CONSOLE &&
            chr->targetType == VIR_DOMAIN_CHR_CONSOLE_TARGET_TYPE_VIRTIO &&
            !virDomainVirtioSerialAddrIsComplete(&chr->info) &&
144 145
            virDomainVirtioSerialAddrAutoAssignFromCache(def, addrs,
                                                         &chr->info, true) < 0)
146 147 148 149 150 151 152 153
            goto cleanup;
    }

    for (i = 0; i < def->nchannels; i++) {
        virDomainChrDefPtr chr = def->channels[i];
        if (chr->deviceType == VIR_DOMAIN_CHR_DEVICE_TYPE_CHANNEL &&
            chr->targetType == VIR_DOMAIN_CHR_CHANNEL_TARGET_TYPE_VIRTIO &&
            !virDomainVirtioSerialAddrIsComplete(&chr->info) &&
154 155
            virDomainVirtioSerialAddrAutoAssignFromCache(def, addrs,
                                                         &chr->info, false) < 0)
156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224
            goto cleanup;
    }

    ret = 0;

 cleanup:
    virDomainVirtioSerialAddrSetFree(addrs);
    return ret;
}


static int
qemuDomainSpaprVIOFindByReg(virDomainDefPtr def ATTRIBUTE_UNUSED,
                            virDomainDeviceDefPtr device ATTRIBUTE_UNUSED,
                            virDomainDeviceInfoPtr info, void *opaque)
{
    virDomainDeviceInfoPtr target = opaque;

    if (info->type != VIR_DOMAIN_DEVICE_ADDRESS_TYPE_SPAPRVIO)
        return 0;

    /* Match a dev that has a reg, is not us, and has a matching reg */
    if (info->addr.spaprvio.has_reg && info != target &&
        info->addr.spaprvio.reg == target->addr.spaprvio.reg)
        /* Has to be < 0 so virDomainDeviceInfoIterate() will exit */
        return -1;

    return 0;
}


static int
qemuDomainAssignSpaprVIOAddress(virDomainDefPtr def,
                                virDomainDeviceInfoPtr info,
                                unsigned long long default_reg)
{
    bool user_reg;
    int ret;

    if (info->type != VIR_DOMAIN_DEVICE_ADDRESS_TYPE_SPAPRVIO)
        return 0;

    /* Check if the user has assigned the reg already, if so use it */
    user_reg = info->addr.spaprvio.has_reg;
    if (!user_reg) {
        info->addr.spaprvio.reg = default_reg;
        info->addr.spaprvio.has_reg = true;
    }

    ret = virDomainDeviceInfoIterate(def, qemuDomainSpaprVIOFindByReg, info);
    while (ret != 0) {
        if (user_reg) {
            virReportError(VIR_ERR_XML_ERROR,
                           _("spapr-vio address %#llx already in use"),
                           info->addr.spaprvio.reg);
            return -EEXIST;
        }

        /* We assigned the reg, so try a new value */
        info->addr.spaprvio.reg += 0x1000;
        ret = virDomainDeviceInfoIterate(def, qemuDomainSpaprVIOFindByReg,
                                         info);
    }

    return 0;
}


static int
225
qemuDomainAssignSpaprVIOAddresses(virDomainDefPtr def)
226 227 228 229 230 231 232
{
    size_t i;
    int ret = -1;

    /* Default values match QEMU. See spapr_(llan|vscsi|vty).c */

    for (i = 0; i < def->nnets; i++) {
233 234 235 236 237 238 239 240
        virDomainNetDefPtr net = def->nets[i];

        if (net->model &&
            STREQ(net->model, "spapr-vlan")) {
            net->info.type = VIR_DOMAIN_DEVICE_ADDRESS_TYPE_SPAPRVIO;
        }

        if (qemuDomainAssignSpaprVIOAddress(def, &net->info, VIO_ADDR_NET) < 0)
241 242 243 244
            goto cleanup;
    }

    for (i = 0; i < def->ncontrollers; i++) {
245 246
        virDomainControllerDefPtr cont = def->controllers[i];

247
        if (cont->model == VIR_DOMAIN_CONTROLLER_MODEL_SCSI_IBMVSCSI &&
248 249 250 251 252
            cont->type == VIR_DOMAIN_CONTROLLER_TYPE_SCSI) {
            cont->info.type = VIR_DOMAIN_DEVICE_ADDRESS_TYPE_SPAPRVIO;
        }
        if (qemuDomainAssignSpaprVIOAddress(def, &cont->info,
                                            VIO_ADDR_SCSI) < 0) {
253
            goto cleanup;
254
        }
255 256 257 258
    }

    for (i = 0; i < def->nserials; i++) {
        if (def->serials[i]->deviceType == VIR_DOMAIN_CHR_DEVICE_TYPE_SERIAL &&
259
            def->serials[i]->targetType == VIR_DOMAIN_CHR_SERIAL_TARGET_TYPE_SPAPR_VIO) {
260
            def->serials[i]->info.type = VIR_DOMAIN_DEVICE_ADDRESS_TYPE_SPAPRVIO;
261
        }
262 263 264 265 266 267
        if (qemuDomainAssignSpaprVIOAddress(def, &def->serials[i]->info,
                                            VIO_ADDR_SERIAL) < 0)
            goto cleanup;
    }

    if (def->nvram) {
268
        if (qemuDomainIsPSeries(def))
269 270 271 272 273 274 275 276 277 278 279 280 281 282 283
            def->nvram->info.type = VIR_DOMAIN_DEVICE_ADDRESS_TYPE_SPAPRVIO;
        if (qemuDomainAssignSpaprVIOAddress(def, &def->nvram->info,
                                            VIO_ADDR_NVRAM) < 0)
            goto cleanup;
    }

    /* No other devices are currently supported on spapr-vio */

    ret = 0;

 cleanup:
    return ret;
}


284 285 286 287 288 289 290 291 292 293 294 295 296
static void
qemuDomainPrimeVfioDeviceAddresses(virDomainDefPtr def,
                                   virDomainDeviceAddressType type)
{
    size_t i;

    for (i = 0; i < def->nhostdevs; i++) {
        virDomainHostdevSubsysPtr subsys = &def->hostdevs[i]->source.subsys;

        if (virHostdevIsMdevDevice(def->hostdevs[i]) &&
            subsys->u.mdev.model == VIR_MDEV_MODEL_TYPE_VFIO_CCW &&
            def->hostdevs[i]->info->type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_NONE)
            def->hostdevs[i]->info->type = type;
B
Boris Fiuczynski 已提交
297 298 299 300

        if (virHostdevIsMdevDevice(def->hostdevs[i]) &&
            subsys->u.mdev.model == VIR_MDEV_MODEL_TYPE_VFIO_AP)
            def->hostdevs[i]->info->type = VIR_DOMAIN_DEVICE_ADDRESS_TYPE_NONE;
301 302 303 304
    }
}


305 306 307 308 309 310
static void
qemuDomainPrimeVirtioDeviceAddresses(virDomainDefPtr def,
                                     virDomainDeviceAddressType type)
{
    /*
       declare address-less virtio devices to be of address type 'type'
311 312
       disks, networks, videos, consoles, controllers, memballoon and rng
       in this order
313 314
       if type is ccw filesystem and vsock devices are declared to be of
       address type ccw
315 316 317 318 319 320 321 322 323 324
    */
    size_t i;

    for (i = 0; i < def->ndisks; i++) {
        if (def->disks[i]->bus == VIR_DOMAIN_DISK_BUS_VIRTIO &&
            def->disks[i]->info.type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_NONE)
            def->disks[i]->info.type = type;
    }

    for (i = 0; i < def->nnets; i++) {
325 326
        virDomainNetDefPtr net = def->nets[i];

327 328
        if (net->model &&
            STREQ(net->model, "virtio") &&
329 330
            net->info.type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_NONE) {
            net->info.type = type;
331 332 333
        }
    }

334 335 336 337 338 339 340 341
    for (i = 0; i < def->nvideos; i++) {
        virDomainVideoDefPtr video = def->videos[i];

        if (video->info.type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_NONE &&
            video->type == VIR_DOMAIN_VIDEO_TYPE_VIRTIO)
            video->info.type = type;
    }

342
    for (i = 0; i < def->ninputs; i++) {
343
        if (def->inputs[i]->bus == VIR_DOMAIN_INPUT_BUS_VIRTIO &&
344 345 346 347 348
            def->inputs[i]->info.type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_NONE)
            def->inputs[i]->info.type = type;
    }

    for (i = 0; i < def->ncontrollers; i++) {
349 350 351 352 353 354 355
        virDomainControllerDefPtr cont = def->controllers[i];

        if ((cont->type == VIR_DOMAIN_CONTROLLER_TYPE_VIRTIO_SERIAL ||
             cont->type == VIR_DOMAIN_CONTROLLER_TYPE_SCSI) &&
            cont->info.type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_NONE) {
            cont->info.type = type;
        }
356 357
    }

358 359 360 361 362 363 364 365
    for (i = 0; i < def->nhostdevs; i++) {
        if (def->hostdevs[i]->mode == VIR_DOMAIN_HOSTDEV_MODE_SUBSYS &&
            def->hostdevs[i]->source.subsys.type ==
            VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_SCSI_HOST &&
            def->hostdevs[i]->info->type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_NONE)
            def->hostdevs[i]->info->type = type;
    }

366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381
    if (def->memballoon &&
        def->memballoon->model == VIR_DOMAIN_MEMBALLOON_MODEL_VIRTIO &&
        def->memballoon->info.type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_NONE)
        def->memballoon->info.type = type;

    for (i = 0; i < def->nrngs; i++) {
        if (def->rngs[i]->model == VIR_DOMAIN_RNG_MODEL_VIRTIO &&
            def->rngs[i]->info.type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_NONE)
            def->rngs[i]->info.type = type;
    }

    if (type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_CCW) {
        for (i = 0; i < def->nfss; i++) {
            if (def->fss[i]->info.type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_NONE)
                def->fss[i]->info.type = type;
        }
382 383 384 385
        if (def->vsock &&
            def->vsock->info.type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_NONE) {
            def->vsock->info.type = type;
        }
386 387 388 389 390 391 392 393 394 395 396 397
    }
}


/*
 * Three steps populating CCW devnos
 * 1. Allocate empty address set
 * 2. Gather addresses with explicit devno
 * 3. Assign defaults to the rest
 */
static int
qemuDomainAssignS390Addresses(virDomainDefPtr def,
T
Tomasz Flendrich 已提交
398
                              virQEMUCapsPtr qemuCaps)
399 400 401 402
{
    int ret = -1;
    virDomainCCWAddressSetPtr addrs = NULL;

403
    if (qemuDomainIsS390CCW(def) &&
404
        virQEMUCapsGet(qemuCaps, QEMU_CAPS_CCW)) {
405 406 407
        if (virQEMUCapsGet(qemuCaps, QEMU_CAPS_DEVICE_VFIO_CCW))
            qemuDomainPrimeVfioDeviceAddresses(
                def, VIR_DOMAIN_DEVICE_ADDRESS_TYPE_CCW);
408 409 410
        qemuDomainPrimeVirtioDeviceAddresses(
            def, VIR_DOMAIN_DEVICE_ADDRESS_TYPE_CCW);

411
        if (!(addrs = virDomainCCWAddressSetCreateFromDomain(def)))
412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428
            goto cleanup;

    } else if (virQEMUCapsGet(qemuCaps, QEMU_CAPS_VIRTIO_S390)) {
        /* deal with legacy virtio-s390 */
        qemuDomainPrimeVirtioDeviceAddresses(
            def, VIR_DOMAIN_DEVICE_ADDRESS_TYPE_VIRTIO_S390);
    }

    ret = 0;

 cleanup:
    virDomainCCWAddressSetFree(addrs);

    return ret;
}


429 430 431 432 433 434 435 436 437
static int
qemuDomainHasVirtioMMIODevicesCallback(virDomainDefPtr def ATTRIBUTE_UNUSED,
                                       virDomainDeviceDefPtr dev ATTRIBUTE_UNUSED,
                                       virDomainDeviceInfoPtr info,
                                       void *opaque)
{
    if (info->type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_VIRTIO_MMIO) {
        /* We can stop iterating as soon as we find the first
         * virtio-mmio device */
438
        *((bool *)opaque) = true;
439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466
        return -1;
    }

    return 0;
}


/**
 * qemuDomainHasVirtioMMIODevices:
 * @def: domain definition
 *
 * Scan @def looking for devices with a virtio-mmio address.
 *
 * Returns: true if there are any, false otherwise
 */
static bool
qemuDomainHasVirtioMMIODevices(virDomainDefPtr def)
{
    bool result = false;

    virDomainDeviceInfoIterate(def,
                               qemuDomainHasVirtioMMIODevicesCallback,
                               &result);

    return result;
}


467
static void
468 469 470
qemuDomainAssignARMVirtioMMIOAddresses(virDomainDefPtr def,
                                       virQEMUCapsPtr qemuCaps)
{
471 472 473 474 475
    if (def->os.arch != VIR_ARCH_ARMV7L &&
        def->os.arch != VIR_ARCH_AARCH64)
        return;

    if (!(STRPREFIX(def->os.machine, "vexpress-") ||
476
          qemuDomainIsARMVirt(def)))
477 478
        return;

479 480 481
    /* We use virtio-mmio by default on mach-virt guests only if they already
     * have at least one virtio-mmio device: in all other cases, we prefer
     * virtio-pci */
482
    if (qemuDomainHasPCIeRoot(def) &&
483 484 485 486 487 488
        !qemuDomainHasVirtioMMIODevices(def)) {
        qemuDomainPrimeVirtioDeviceAddresses(def,
                                             VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI);
    } else if (virQEMUCapsGet(qemuCaps, QEMU_CAPS_DEVICE_VIRTIO_MMIO)) {
        qemuDomainPrimeVirtioDeviceAddresses(def,
                                             VIR_DOMAIN_DEVICE_ADDRESS_TYPE_VIRTIO_MMIO);
489 490 491 492
    }
}


493 494 495 496 497 498 499 500 501 502 503 504 505 506
static void
qemuDomainAssignRISCVVirtioMMIOAddresses(virDomainDefPtr def,
                                         virQEMUCapsPtr qemuCaps)
{
    if (!qemuDomainIsRISCVVirt(def))
        return;

    if (virQEMUCapsGet(qemuCaps, QEMU_CAPS_DEVICE_VIRTIO_MMIO)) {
        qemuDomainPrimeVirtioDeviceAddresses(def,
                                             VIR_DOMAIN_DEVICE_ADDRESS_TYPE_VIRTIO_MMIO);
    }
}


507 508 509 510 511
static void
qemuDomainAssignVirtioMMIOAddresses(virDomainDefPtr def,
                                    virQEMUCapsPtr qemuCaps)
{
    qemuDomainAssignARMVirtioMMIOAddresses(def, qemuCaps);
512 513

    qemuDomainAssignRISCVVirtioMMIOAddresses(def, qemuCaps);
514 515 516
}


517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574
static bool
qemuDomainDeviceSupportZPCI(virDomainDeviceDefPtr device)
{
    switch ((virDomainDeviceType)device->type) {
    case VIR_DOMAIN_DEVICE_CHR:
        return false;

    case VIR_DOMAIN_DEVICE_CONTROLLER:
    case VIR_DOMAIN_DEVICE_DISK:
    case VIR_DOMAIN_DEVICE_LEASE:
    case VIR_DOMAIN_DEVICE_FS:
    case VIR_DOMAIN_DEVICE_NET:
    case VIR_DOMAIN_DEVICE_INPUT:
    case VIR_DOMAIN_DEVICE_SOUND:
    case VIR_DOMAIN_DEVICE_VIDEO:
    case VIR_DOMAIN_DEVICE_HOSTDEV:
    case VIR_DOMAIN_DEVICE_WATCHDOG:
    case VIR_DOMAIN_DEVICE_GRAPHICS:
    case VIR_DOMAIN_DEVICE_HUB:
    case VIR_DOMAIN_DEVICE_REDIRDEV:
    case VIR_DOMAIN_DEVICE_SMARTCARD:
    case VIR_DOMAIN_DEVICE_MEMBALLOON:
    case VIR_DOMAIN_DEVICE_NVRAM:
    case VIR_DOMAIN_DEVICE_RNG:
    case VIR_DOMAIN_DEVICE_SHMEM:
    case VIR_DOMAIN_DEVICE_TPM:
    case VIR_DOMAIN_DEVICE_PANIC:
    case VIR_DOMAIN_DEVICE_MEMORY:
    case VIR_DOMAIN_DEVICE_IOMMU:
    case VIR_DOMAIN_DEVICE_VSOCK:
        break;

    case VIR_DOMAIN_DEVICE_NONE:
    case VIR_DOMAIN_DEVICE_LAST:
    default:
        virReportEnumRangeError(virDomainDeviceType, device->type);
        return false;
    }

    return true;
}


static virPCIDeviceAddressExtensionFlags
qemuDomainDeviceCalculatePCIAddressExtensionFlags(virQEMUCapsPtr qemuCaps,
                                                  virDomainDeviceDefPtr dev)
{
    virPCIDeviceAddressExtensionFlags extFlags = 0;

    if (virQEMUCapsGet(qemuCaps, QEMU_CAPS_DEVICE_ZPCI) &&
        qemuDomainDeviceSupportZPCI(dev)) {
        extFlags |= VIR_PCI_ADDRESS_EXTENSION_ZPCI;
    }

    return extFlags;
}


575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593
/**
 * qemuDomainDeviceCalculatePCIConnectFlags:
 *
 * @dev: The device to be checked
 * @pcieFlags: flags to use for a known PCI Express device
 * @virtioFlags: flags to use for a virtio device (properly vetted
 *       for the current qemu binary and arch/machinetype)
 *
 * Lowest level function to determine PCI connectFlags for a
 * device. This function relies on the next higher-level function
 * determining the value for pcieFlags and virtioFlags in advance -
 * this is to make it more efficient to call multiple times.
 *
 * Returns appropriate virDomainPCIConnectFlags for this device in
 * this domain, or 0 if the device doesn't connect using PCI. There
 * is no failure.
 */
static virDomainPCIConnectFlags
qemuDomainDeviceCalculatePCIConnectFlags(virDomainDeviceDefPtr dev,
594
                                         virQEMUDriverPtr driver,
595
                                         virDomainPCIConnectFlags pcieFlags,
596
                                         virDomainPCIConnectFlags virtioFlags)
597 598 599 600
{
    virDomainPCIConnectFlags pciFlags = (VIR_PCI_CONNECT_TYPE_PCI_DEVICE |
                                         VIR_PCI_CONNECT_HOTPLUGGABLE);

601
    switch ((virDomainDeviceType)dev->type) {
602 603 604
    case VIR_DOMAIN_DEVICE_CONTROLLER: {
        virDomainControllerDefPtr cont = dev->data.controller;

605
        switch ((virDomainControllerType)cont->type) {
606 607 608 609 610 611 612 613
        case VIR_DOMAIN_CONTROLLER_TYPE_PCI:
            return virDomainPCIControllerModelToConnectType(cont->model);

        case VIR_DOMAIN_CONTROLLER_TYPE_SATA:
            return pciFlags;

        case VIR_DOMAIN_CONTROLLER_TYPE_USB:
            switch ((virDomainControllerModelUSB) cont->model) {
614 615 616 617 618 619 620 621 622
            case VIR_DOMAIN_CONTROLLER_MODEL_USB_DEFAULT:
                /* qemuDomainControllerDefPostParse should have
                 * changed 'model' to an explicit USB model in
                 * most cases. Since we're still on the default
                 * though, we must be going to use "-usb", which
                 * is assumed to be a PCI default
                 */
                return pciFlags;

623
            case VIR_DOMAIN_CONTROLLER_MODEL_USB_NEC_XHCI:
624
            case VIR_DOMAIN_CONTROLLER_MODEL_USB_QEMU_XHCI:
625
                return pcieFlags;
626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641

            case VIR_DOMAIN_CONTROLLER_MODEL_USB_EHCI:
            case VIR_DOMAIN_CONTROLLER_MODEL_USB_ICH9_EHCI1:
            case VIR_DOMAIN_CONTROLLER_MODEL_USB_ICH9_UHCI1:
            case VIR_DOMAIN_CONTROLLER_MODEL_USB_ICH9_UHCI2:
            case VIR_DOMAIN_CONTROLLER_MODEL_USB_ICH9_UHCI3:
            case VIR_DOMAIN_CONTROLLER_MODEL_USB_VT82C686B_UHCI:
            case VIR_DOMAIN_CONTROLLER_MODEL_USB_PIIX3_UHCI:
            case VIR_DOMAIN_CONTROLLER_MODEL_USB_PIIX4_UHCI:
            case VIR_DOMAIN_CONTROLLER_MODEL_USB_PCI_OHCI:
                return pciFlags;

            case VIR_DOMAIN_CONTROLLER_MODEL_USB_QUSB1: /* xen only */
            case VIR_DOMAIN_CONTROLLER_MODEL_USB_QUSB2: /* xen only */
            case VIR_DOMAIN_CONTROLLER_MODEL_USB_NONE:
            case VIR_DOMAIN_CONTROLLER_MODEL_USB_LAST:
642
                return 0;
643
            }
644
            break;
645 646

        case VIR_DOMAIN_CONTROLLER_TYPE_IDE:
647 648
            return pciFlags;

649
        case VIR_DOMAIN_CONTROLLER_TYPE_SCSI:
650
            switch ((virDomainControllerModelSCSI) cont->model) {
651 652 653
            case VIR_DOMAIN_CONTROLLER_MODEL_SCSI_DEFAULT:
                return 0;

654 655 656 657 658 659 660 661 662 663 664 665 666 667 668
            case VIR_DOMAIN_CONTROLLER_MODEL_SCSI_VIRTIO_SCSI:
                return virtioFlags;

            case VIR_DOMAIN_CONTROLLER_MODEL_SCSI_AUTO:
            case VIR_DOMAIN_CONTROLLER_MODEL_SCSI_BUSLOGIC:
            case VIR_DOMAIN_CONTROLLER_MODEL_SCSI_LSILOGIC:
            case VIR_DOMAIN_CONTROLLER_MODEL_SCSI_LSISAS1068:
            case VIR_DOMAIN_CONTROLLER_MODEL_SCSI_VMPVSCSI:
            case VIR_DOMAIN_CONTROLLER_MODEL_SCSI_IBMVSCSI:
            case VIR_DOMAIN_CONTROLLER_MODEL_SCSI_LSISAS1078:
                return pciFlags;

            case VIR_DOMAIN_CONTROLLER_MODEL_SCSI_LAST:
                return 0;
            }
669
            break;
670

671
        case VIR_DOMAIN_CONTROLLER_TYPE_VIRTIO_SERIAL:
672
            return virtioFlags;
673 674 675 676 677 678 679 680

        case VIR_DOMAIN_CONTROLLER_TYPE_FDC:
        case VIR_DOMAIN_CONTROLLER_TYPE_CCID:
        case VIR_DOMAIN_CONTROLLER_TYPE_LAST:
            /* should be 0 */
            return pciFlags;
        }
    }
681
        break;
682 683 684

    case VIR_DOMAIN_DEVICE_FS:
        /* the only type of filesystem so far is virtio-9p-pci */
685
        return virtioFlags;
686 687 688 689 690 691 692 693 694 695

    case VIR_DOMAIN_DEVICE_NET: {
        virDomainNetDefPtr net = dev->data.net;

        /* NB: a type='hostdev' will use PCI, but its
         * address is assigned when we're assigning the
         * addresses for other hostdev devices.
         */
        if (net->type == VIR_DOMAIN_NET_TYPE_HOSTDEV ||
            STREQ(net->model, "usb-net")) {
696
            return 0;
697
        }
698 699 700 701

        if (STREQ(net->model, "virtio"))
            return  virtioFlags;

702 703 704
        if (STREQ(net->model, "e1000e"))
            return pcieFlags;

705 706 707 708 709 710 711 712 713 714 715 716 717 718 719
        return pciFlags;
    }

    case VIR_DOMAIN_DEVICE_SOUND:
        switch ((virDomainSoundModel) dev->data.sound->model) {
        case VIR_DOMAIN_SOUND_MODEL_ES1370:
        case VIR_DOMAIN_SOUND_MODEL_AC97:
        case VIR_DOMAIN_SOUND_MODEL_ICH6:
        case VIR_DOMAIN_SOUND_MODEL_ICH9:
            return pciFlags;

        case VIR_DOMAIN_SOUND_MODEL_SB16:
        case VIR_DOMAIN_SOUND_MODEL_PCSPK:
        case VIR_DOMAIN_SOUND_MODEL_USB:
        case VIR_DOMAIN_SOUND_MODEL_LAST:
720
            return 0;
721
        }
722
        break;
723 724 725 726

    case VIR_DOMAIN_DEVICE_DISK:
        switch ((virDomainDiskBus) dev->data.disk->bus) {
        case VIR_DOMAIN_DISK_BUS_VIRTIO:
727
            return virtioFlags; /* only virtio disks use PCI */
728 729 730 731 732 733 734 735 736 737

        case VIR_DOMAIN_DISK_BUS_IDE:
        case VIR_DOMAIN_DISK_BUS_FDC:
        case VIR_DOMAIN_DISK_BUS_SCSI:
        case VIR_DOMAIN_DISK_BUS_XEN:
        case VIR_DOMAIN_DISK_BUS_USB:
        case VIR_DOMAIN_DISK_BUS_UML:
        case VIR_DOMAIN_DISK_BUS_SATA:
        case VIR_DOMAIN_DISK_BUS_SD:
        case VIR_DOMAIN_DISK_BUS_LAST:
738
            return 0;
739
        }
740
        break;
741

742 743 744 745 746 747
    case VIR_DOMAIN_DEVICE_HOSTDEV: {
        virDomainHostdevDefPtr hostdev = dev->data.hostdev;
        bool isExpress = false;
        virPCIDevicePtr pciDev;
        virPCIDeviceAddressPtr hostAddr = &hostdev->source.subsys.u.pci.addr;

748 749 750 751
        if (!virHostdevIsMdevDevice(hostdev) &&
            (hostdev->mode != VIR_DOMAIN_HOSTDEV_MODE_SUBSYS ||
             (hostdev->source.subsys.type != VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_PCI &&
              hostdev->source.subsys.type != VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_SCSI_HOST))) {
752 753 754 755 756 757 758 759 760 761 762
            return 0;
        }

        if (pciFlags == pcieFlags) {
            /* This arch/qemu only supports legacy PCI, so there
             * is no point in checking if the device is an Express
             * device.
             */
            return pciFlags;
        }

763
        if (virDeviceInfoPCIAddressIsPresent(hostdev->info)) {
764 765 766 767 768 769 770 771 772 773 774
            /* A guest-side address has already been assigned, so
             * we can avoid reading the PCI config, and just use
             * pcieFlags, since the pciConnectFlags checking is
             * more relaxed when an address is already assigned
             * than it is when we're looking for a new address (so
             * validation will pass regardless of whether we set
             * the flags to PCI or PCIe).
             */
            return pcieFlags;
        }

775 776 777 778 779
        /* mdevs don't have corresponding files in /sys that we can poke to
         * try and figure out whether they are legacy PCI or PCI Express, so
         * the logic below would never work; instead, we just go ahead and
         * assume they're PCI Express. This is a very reasonable assumption,
         * as all current mdev-capable devices are indeed PCI Express */
780 781 782
        if (hostdev->source.subsys.type == VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_MDEV)
            return pcieFlags;

783 784 785 786 787 788 789
        /* according to pbonzini, from the guest PoV vhost-scsi devices
         * are the same as virtio-scsi, so they should use virtioFlags
         * (same as virtio-scsi) to determine Express vs. legacy placement
         */
        if (hostdev->source.subsys.type == VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_SCSI_HOST)
            return virtioFlags;

790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839
        if (!(pciDev = virPCIDeviceNew(hostAddr->domain,
                                       hostAddr->bus,
                                       hostAddr->slot,
                                       hostAddr->function))) {
            /* libvirt should be able to perform all the
             * operations in virPCIDeviceNew() even if it's
             * running unprivileged, so if this fails, the device
             * apparently doesn't currently exist on the host.
             * Since the overwhelming majority of assignable host
             * devices are PCIe, assume this one is too.
             */
            return pcieFlags;
        }

        if (!driver->privileged) {
            /* unprivileged libvirtd is unable to read *all* of a
             * device's PCI config (it can only read the first 64
             * bytes, which isn't enough for the check that's done
             * in virPCIDeviceIsPCIExpress()), so instead of
             * trying and failing, we make an educated guess based
             * on the length of the device's config file - if it
             * is 256 bytes, then it is definitely a legacy PCI
             * device. If it's larger than that, then it is
             * *probably PCIe (although it could be PCI-x, but
             * those are extremely rare). If the config file can't
             * be found (in which case the "length" will be -1),
             * then we blindly assume the most likely outcome -
             * PCIe.
             */
            off_t configLen
               = virFileLength(virPCIDeviceGetConfigPath(pciDev), -1);

            virPCIDeviceFree(pciDev);

            if (configLen == 256)
                return pciFlags;

            return pcieFlags;
        }

        /* If we are running with privileges, we can examine the
         * PCI config contents with virPCIDeviceIsPCIExpress() for
         * a definitive answer.
         */
        isExpress = virPCIDeviceIsPCIExpress(pciDev);
        virPCIDeviceFree(pciDev);

        if (isExpress)
            return pcieFlags;

840
        return pciFlags;
841
    }
842 843 844 845

    case VIR_DOMAIN_DEVICE_MEMBALLOON:
        switch ((virDomainMemballoonModel) dev->data.memballoon->model) {
        case VIR_DOMAIN_MEMBALLOON_MODEL_VIRTIO:
846
            return virtioFlags;
847 848 849 850

        case VIR_DOMAIN_MEMBALLOON_MODEL_XEN:
        case VIR_DOMAIN_MEMBALLOON_MODEL_NONE:
        case VIR_DOMAIN_MEMBALLOON_MODEL_LAST:
851
            return 0;
852
        }
853
        break;
854 855 856 857

    case VIR_DOMAIN_DEVICE_RNG:
        switch ((virDomainRNGModel) dev->data.rng->model) {
        case VIR_DOMAIN_RNG_MODEL_VIRTIO:
858
            return virtioFlags;
859 860

        case VIR_DOMAIN_RNG_MODEL_LAST:
861
            return 0;
862
        }
863
        break;
864 865 866 867 868 869 870 871 872 873

    case VIR_DOMAIN_DEVICE_WATCHDOG:
        /* only one model connects using PCI */
        switch ((virDomainWatchdogModel) dev->data.watchdog->model) {
        case VIR_DOMAIN_WATCHDOG_MODEL_I6300ESB:
            return pciFlags;

        case VIR_DOMAIN_WATCHDOG_MODEL_IB700:
        case VIR_DOMAIN_WATCHDOG_MODEL_DIAG288:
        case VIR_DOMAIN_WATCHDOG_MODEL_LAST:
874
            return 0;
875
        }
876
        break;
877 878

    case VIR_DOMAIN_DEVICE_VIDEO:
879
        switch ((virDomainVideoType)dev->data.video->type) {
880
        case VIR_DOMAIN_VIDEO_TYPE_VIRTIO:
881 882
            return virtioFlags;

883 884 885 886 887 888 889 890 891
        case VIR_DOMAIN_VIDEO_TYPE_VGA:
        case VIR_DOMAIN_VIDEO_TYPE_CIRRUS:
        case VIR_DOMAIN_VIDEO_TYPE_VMVGA:
        case VIR_DOMAIN_VIDEO_TYPE_XEN:
        case VIR_DOMAIN_VIDEO_TYPE_VBOX:
        case VIR_DOMAIN_VIDEO_TYPE_QXL:
        case VIR_DOMAIN_VIDEO_TYPE_PARALLELS:
            return pciFlags;

892
        case VIR_DOMAIN_VIDEO_TYPE_DEFAULT:
F
Fabian Freyer 已提交
893
        case VIR_DOMAIN_VIDEO_TYPE_GOP:
894
        case VIR_DOMAIN_VIDEO_TYPE_NONE:
895
        case VIR_DOMAIN_VIDEO_TYPE_LAST:
896
            return 0;
897
        }
898
        break;
899 900 901 902 903 904 905

    case VIR_DOMAIN_DEVICE_SHMEM:
        return pciFlags;

    case VIR_DOMAIN_DEVICE_INPUT:
        switch ((virDomainInputBus) dev->data.input->bus) {
        case VIR_DOMAIN_INPUT_BUS_VIRTIO:
906
            return virtioFlags;
907 908 909 910 911 912

        case VIR_DOMAIN_INPUT_BUS_PS2:
        case VIR_DOMAIN_INPUT_BUS_USB:
        case VIR_DOMAIN_INPUT_BUS_XEN:
        case VIR_DOMAIN_INPUT_BUS_PARALLELS:
        case VIR_DOMAIN_INPUT_BUS_LAST:
913
            return 0;
914
        }
915
        break;
916 917

    case VIR_DOMAIN_DEVICE_CHR:
918
        switch ((virDomainChrSerialTargetType)dev->data.chr->targetType) {
919 920 921 922 923
        case VIR_DOMAIN_CHR_SERIAL_TARGET_TYPE_PCI:
            return pciFlags;

        case VIR_DOMAIN_CHR_SERIAL_TARGET_TYPE_ISA:
        case VIR_DOMAIN_CHR_SERIAL_TARGET_TYPE_USB:
924
        case VIR_DOMAIN_CHR_SERIAL_TARGET_TYPE_SPAPR_VIO:
925
        case VIR_DOMAIN_CHR_SERIAL_TARGET_TYPE_SYSTEM:
926
        case VIR_DOMAIN_CHR_SERIAL_TARGET_TYPE_SCLP:
927
        case VIR_DOMAIN_CHR_SERIAL_TARGET_TYPE_NONE:
928
        case VIR_DOMAIN_CHR_SERIAL_TARGET_TYPE_LAST:
929
            return 0;
930
        }
931
        break;
932

J
Ján Tomko 已提交
933 934 935
    case VIR_DOMAIN_DEVICE_VSOCK:
        return virtioFlags;

936
        /* These devices don't ever connect with PCI */
937 938 939 940 941 942 943
    case VIR_DOMAIN_DEVICE_NVRAM:
    case VIR_DOMAIN_DEVICE_TPM:
    case VIR_DOMAIN_DEVICE_PANIC:
    case VIR_DOMAIN_DEVICE_MEMORY:
    case VIR_DOMAIN_DEVICE_HUB:
    case VIR_DOMAIN_DEVICE_REDIRDEV:
    case VIR_DOMAIN_DEVICE_SMARTCARD:
944
        /* These devices don't even have a DeviceInfo */
945 946 947 948 949
    case VIR_DOMAIN_DEVICE_LEASE:
    case VIR_DOMAIN_DEVICE_GRAPHICS:
    case VIR_DOMAIN_DEVICE_IOMMU:
    case VIR_DOMAIN_DEVICE_LAST:
    case VIR_DOMAIN_DEVICE_NONE:
950
        return 0;
951 952 953 954 955 956 957 958 959 960 961 962 963 964
    }

    /* We can never get here, because all cases are covered in the
     * switch, and they all return, but the compiler will still
     * complain "control reaches end of non-void function" unless
     * we add the following return.
     */
    return 0;
}


typedef struct {
    virDomainPCIConnectFlags virtioFlags;
    virDomainPCIConnectFlags pcieFlags;
965
    virQEMUDriverPtr driver;
966 967 968 969 970 971 972 973 974 975 976 977
} qemuDomainFillDevicePCIConnectFlagsIterData;


/**
 * qemuDomainFillDevicePCIConnectFlagsIterInit:
 *
 * Initialize the iterator data that is used when calling
 * qemuDomainCalculateDevicePCIConnectFlags().
 */
static void
qemuDomainFillDevicePCIConnectFlagsIterInit(virDomainDefPtr def,
                                            virQEMUCapsPtr qemuCaps,
978
                                            virQEMUDriverPtr driver,
979 980
                                            qemuDomainFillDevicePCIConnectFlagsIterData *data)
{
981 982
    data->driver = driver;

983
    if (qemuDomainHasPCIeRoot(def)) {
984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021
        data->pcieFlags = (VIR_PCI_CONNECT_TYPE_PCIE_DEVICE |
                           VIR_PCI_CONNECT_HOTPLUGGABLE);
    } else {
        data->pcieFlags = (VIR_PCI_CONNECT_TYPE_PCI_DEVICE |
                           VIR_PCI_CONNECT_HOTPLUGGABLE);
    }

    if (virQEMUCapsGet(qemuCaps, QEMU_CAPS_VIRTIO_PCI_DISABLE_LEGACY)) {
        data->virtioFlags = data->pcieFlags;
    } else {
        data->virtioFlags = (VIR_PCI_CONNECT_TYPE_PCI_DEVICE |
                             VIR_PCI_CONNECT_HOTPLUGGABLE);
    }
}


/**
 * qemuDomainFillDevicePCIConnectFlagsIter:
 *
 * @def: the entire DomainDef
 * @dev: The device to be checked
 * @info: virDomainDeviceInfo within the device
 * @opaque: points to iterator data setup beforehand.
 *
 * Sets the pciConnectFlags for a single device's info. Has properly
 * formatted arguments to be called by virDomainDeviceInfoIterate().
 *
 * Always returns 0 - there is no failure.
 */
static int
qemuDomainFillDevicePCIConnectFlagsIter(virDomainDefPtr def ATTRIBUTE_UNUSED,
                                        virDomainDeviceDefPtr dev,
                                        virDomainDeviceInfoPtr info,
                                        void *opaque)
{
    qemuDomainFillDevicePCIConnectFlagsIterData *data = opaque;

    info->pciConnectFlags
1022 1023
        = qemuDomainDeviceCalculatePCIConnectFlags(dev, data->driver,
                                                   data->pcieFlags,
1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040
                                                   data->virtioFlags);
    return 0;
}


/**
 * qemuDomainFillAllPCIConnectFlags:
 *
 * @def: the entire DomainDef
 * @qemuCaps: as you'd expect
 *
 * Set the info->pciConnectFlags for all devices in the domain.
 *
 * Returns 0 on success or -1 on failure (the only possibility of
 * failure would be some internal problem with
 * virDomainDeviceInfoIterate())
 */
1041
static int
1042
qemuDomainFillAllPCIConnectFlags(virDomainDefPtr def,
1043 1044
                                 virQEMUCapsPtr qemuCaps,
                                 virQEMUDriverPtr driver)
1045 1046 1047
{
    qemuDomainFillDevicePCIConnectFlagsIterData data;

1048
    qemuDomainFillDevicePCIConnectFlagsIterInit(def, qemuCaps, driver, &data);
1049 1050 1051 1052 1053 1054 1055

    return virDomainDeviceInfoIterate(def,
                                      qemuDomainFillDevicePCIConnectFlagsIter,
                                      &data);
}


1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105
/**
 * qemuDomainFillDevicePCIExtensionFlagsIter:
 *
 * @def: the entire DomainDef
 * @dev: The device to be checked
 * @info: virDomainDeviceInfo within the device
 * @opaque: qemu capabilities
 *
 * Sets the pciAddressExtFlags for a single device's info. Has properly
 * formatted arguments to be called by virDomainDeviceInfoIterate().
 *
 * Always returns 0 - there is no failure.
 */
static int
qemuDomainFillDevicePCIExtensionFlagsIter(virDomainDefPtr def ATTRIBUTE_UNUSED,
                                          virDomainDeviceDefPtr dev,
                                          virDomainDeviceInfoPtr info,
                                          void *opaque)
{
    virQEMUCapsPtr qemuCaps = opaque;

    info->pciAddrExtFlags =
        qemuDomainDeviceCalculatePCIAddressExtensionFlags(qemuCaps, dev);

    return 0;
}


/**
 * qemuDomainFillAllPCIExtensionFlags:
 *
 * @def: the entire DomainDef
 * @qemuCaps: as you'd expect
 *
 * Set the info->pciAddressExtFlags for all devices in the domain.
 *
 * Returns 0 on success or -1 on failure (the only possibility of
 * failure would be some internal problem with
 * virDomainDeviceInfoIterate())
 */
static int
qemuDomainFillAllPCIExtensionFlags(virDomainDefPtr def,
                                   virQEMUCapsPtr qemuCaps)
{
    return virDomainDeviceInfoIterate(def,
                                      qemuDomainFillDevicePCIExtensionFlagsIter,
                                      qemuCaps);
}


1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209
/**
 * qemuDomainFindUnusedIsolationGroupIter:
 * @def: domain definition
 * @dev: device definition
 * @info: device information
 * @opaque: user data
 *
 * Used to implement qemuDomainFindUnusedIsolationGroup(). You probably
 * don't want to call this directly.
 *
 * Return: 0 if the isolation group is not used by the device, <1 otherwise.
 */
static int
qemuDomainFindUnusedIsolationGroupIter(virDomainDefPtr def ATTRIBUTE_UNUSED,
                                       virDomainDeviceDefPtr dev ATTRIBUTE_UNUSED,
                                       virDomainDeviceInfoPtr info,
                                       void *opaque)
{
    unsigned int *isolationGroup = opaque;

    if (info->isolationGroup == *isolationGroup)
        return -1;

    return 0;
}


/**
 * qemuDomainFindUnusedIsolationGroup:
 * @def: domain definition
 *
 * Find an isolation group that is not used by any device in @def yet.
 *
 * Normally, we'd look up the device's IOMMU group and base its isolation
 * group on that; however, when a network interface uses a network backed
 * by SR-IOV Virtual Functions, we can't know at PCI address assignment
 * time which host device will be used so we can't look up its IOMMU group.
 *
 * We still want such a device to be isolated: this function can be used
 * to obtain a synthetic isolation group usable for the purpose.
 *
 * Return: unused isolation group
 */
static unsigned int
qemuDomainFindUnusedIsolationGroup(virDomainDefPtr def)
{
    unsigned int isolationGroup = UINT_MAX;

    /* We start from the highest possible isolation group and work our
     * way backwards so that we're working in a completely different range
     * from IOMMU groups, thus avoiding clashes. We're realistically going
     * to call this function just a few times per guest anyway */
    while (isolationGroup > 0 &&
           virDomainDeviceInfoIterate(def,
                                      qemuDomainFindUnusedIsolationGroupIter,
                                      &isolationGroup) < 0) {
        isolationGroup--;
    }

    return isolationGroup;
}


/**
 * qemuDomainFillDeviceIsolationGroup:
 * @def: domain definition
 * @dev: device definition
 *
 * Fill isolation group information for a single device.
 *
 * Return: 0 on success, <0 on failure
 * */
int
qemuDomainFillDeviceIsolationGroup(virDomainDefPtr def,
                                   virDomainDeviceDefPtr dev)
{
    /* Only host devices need their isolation group to be different from
     * the default. Interfaces of type hostdev are just host devices in
     * disguise, but we don't need to handle them separately because for
     * each such interface a corresponding hostdev is also added to the
     * guest configuration */
    if (dev->type == VIR_DOMAIN_DEVICE_HOSTDEV) {
        virDomainHostdevDefPtr hostdev = dev->data.hostdev;
        virDomainDeviceInfoPtr info = hostdev->info;
        virPCIDeviceAddressPtr hostAddr;
        int tmp;

        /* Only PCI host devices are subject to isolation */
        if (hostdev->mode != VIR_DOMAIN_HOSTDEV_MODE_SUBSYS ||
            hostdev->source.subsys.type != VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_PCI) {
            goto skip;
        }

        hostAddr = &hostdev->source.subsys.u.pci.addr;

        /* If a non-default isolation has already been assigned to the
         * device, we can avoid looking up the information again */
        if (info->isolationGroup > 0)
            goto skip;

        /* The isolation group depends on the IOMMU group assigned by the host */
        tmp = virPCIDeviceAddressGetIOMMUGroupNum(hostAddr);

        if (tmp < 0) {
1210 1211 1212 1213 1214
            VIR_WARN("Can't look up isolation group for host device "
                     "%04x:%02x:%02x.%x, device won't be isolated",
                     hostAddr->domain, hostAddr->bus,
                     hostAddr->slot, hostAddr->function);
            goto skip;
1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238
        }

        /* The isolation group for a host device is its IOMMU group,
         * increased by one: this is because zero is a valid IOMMU group but
         * that's also the default isolation group, which we want to save
         * for emulated devices. Shifting isolation groups for host devices
         * by one ensures there is no overlap */
        info->isolationGroup = tmp + 1;

        VIR_DEBUG("Isolation group for host device %04x:%02x:%02x.%x is %u",
                  hostAddr->domain, hostAddr->bus,
                  hostAddr->slot, hostAddr->function,
                  info->isolationGroup);

    } else if (dev->type == VIR_DOMAIN_DEVICE_NET) {
        virDomainNetDefPtr iface = dev->data.net;
        virDomainDeviceInfoPtr info = &iface->info;
        unsigned int tmp;

        /* Network interfaces can ultimately result in the guest being
         * assigned a host device if the libvirt network they're connected
         * to is of type hostdev. All other kinds of network interfaces don't
         * require us to isolate the guest device, so we can skip them */
        if (iface->type != VIR_DOMAIN_NET_TYPE_NETWORK ||
1239
            virDomainNetResolveActualType(iface) != VIR_DOMAIN_NET_TYPE_HOSTDEV) {
1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253
            goto skip;
        }

        /* If a non-default isolation has already been assigned to the
         * device, we can avoid looking up the information again */
        if (info->isolationGroup > 0)
            goto skip;

        /* Obtain a synthetic isolation group for the device, since at this
         * point in time we don't have access to the IOMMU group of the host
         * device that will eventually be used by the guest */
        tmp = qemuDomainFindUnusedIsolationGroup(def);

        if (tmp == 0) {
1254 1255 1256 1257 1258
            VIR_WARN("Can't obtain usable isolation group for interface "
                     "configured to use hostdev-backed network '%s', "
                     "device won't be isolated",
                     iface->data.network.name);
            goto skip;
1259 1260 1261 1262 1263 1264 1265 1266 1267 1268
        }

        info->isolationGroup = tmp;

        VIR_DEBUG("Isolation group for interface configured to use "
                  "hostdev-backed network '%s' is %u",
                  iface->data.network.name, info->isolationGroup);
    }

 skip:
1269
    return 0;
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 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338
}


/**
 * qemuDomainFillDeviceIsolationGroupIter:
 * @def: domain definition
 * @dev: device definition
 * @info: device information
 * @opaque: user data
 *
 * A version of qemuDomainFillDeviceIsolationGroup() to be used
 * with virDomainDeviceInfoIterate()
 *
 * Return: 0 on success, <0 on failure
 */
static int
qemuDomainFillDeviceIsolationGroupIter(virDomainDefPtr def,
                                       virDomainDeviceDefPtr dev,
                                       virDomainDeviceInfoPtr info ATTRIBUTE_UNUSED,
                                       void *opaque ATTRIBUTE_UNUSED)
{
    return qemuDomainFillDeviceIsolationGroup(def, dev);
}


/**
 * qemuDomainSetupIsolationGroups:
 * @def: domain definition
 *
 * High-level function to set up isolation groups for all devices
 * and controllers in @def. Isolation groups will only be set up if
 * the guest architecture and machine type require it, so this
 * function can and should be called unconditionally before attempting
 * to assign any PCI address.
 *
 * Return: 0 on success, <0 on failure
 */
static int
qemuDomainSetupIsolationGroups(virDomainDefPtr def)
{
    int idx;
    int ret = -1;

    /* Only pSeries guests care about isolation groups at the moment */
    if (!qemuDomainIsPSeries(def))
        return 0;

    idx = virDomainControllerFind(def, VIR_DOMAIN_CONTROLLER_TYPE_PCI, 0);
    if (idx < 0)
        goto cleanup;

    /* We want to prevent hostdevs from being plugged into the default PHB:
     * we can make sure that doesn't happen by locking its isolation group */
    def->controllers[idx]->info.isolationGroupLocked = true;

    /* Fill in isolation groups for all other devices */
    if (virDomainDeviceInfoIterate(def,
                                   qemuDomainFillDeviceIsolationGroupIter,
                                   NULL) < 0) {
        goto cleanup;
    }

    ret = 0;

 cleanup:
    return ret;
}


1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349
/**
 * qemuDomainFillDevicePCIConnectFlags:
 *
 * @def: the entire DomainDef
 * @dev: The device to be checked
 * @qemuCaps: as you'd expect
 *
 * Set the info->pciConnectFlags for a single device.
 *
 * No return value.
 */
1350
static void
1351 1352
qemuDomainFillDevicePCIConnectFlags(virDomainDefPtr def,
                                    virDomainDeviceDefPtr dev,
1353 1354
                                    virQEMUCapsPtr qemuCaps,
                                    virQEMUDriverPtr driver)
1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369
{
    virDomainDeviceInfoPtr info = virDomainDeviceGetInfo(dev);

    if (info) {
        /* qemuDomainDeviceCalculatePCIConnectFlags() is called with
         * the data setup in the ...IterData by ...IterInit() rather
         * than setting the values directly here.  It may seem like
         * pointless posturing, but it's done this way to eliminate
         * duplicated setup code while allowing more efficient
         * operation when it's being done repeatedly with the device
         * iterator (since qemuDomainFillAllPCIConnectFlags() only
         * calls ...IterInit() once for all devices).
         */
        qemuDomainFillDevicePCIConnectFlagsIterData data;

1370
        qemuDomainFillDevicePCIConnectFlagsIterInit(def, qemuCaps, driver, &data);
1371 1372

        info->pciConnectFlags
1373 1374
            = qemuDomainDeviceCalculatePCIConnectFlags(dev, data.driver,
                                                       data.pcieFlags,
1375 1376 1377 1378 1379
                                                       data.virtioFlags);
    }
}


1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400
/**
 * qemuDomainFillDevicePCIExtensionFlags:
 *
 * @dev: The device to be checked
 * @info: virDomainDeviceInfo within the device
 * @qemuCaps: as you'd expect
 *
 * Set the info->pciAddressExtFlags for a single device.
 *
 * No return value.
 */
static void
qemuDomainFillDevicePCIExtensionFlags(virDomainDeviceDefPtr dev,
                                      virDomainDeviceInfoPtr info,
                                      virQEMUCapsPtr qemuCaps)
{
    info->pciAddrExtFlags =
        qemuDomainDeviceCalculatePCIAddressExtensionFlags(qemuCaps, dev);
}


1401
static int
1402
qemuDomainPCIAddressReserveNextAddr(virDomainPCIAddressSetPtr addrs,
1403
                                    virDomainDeviceInfoPtr dev)
1404
{
1405 1406
    return virDomainPCIAddressReserveNextAddr(addrs, dev,
                                              dev->pciConnectFlags, -1);
1407 1408 1409
}


1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427
static int
qemuDomainAssignPCIAddressExtension(virDomainDefPtr def ATTRIBUTE_UNUSED,
                                    virDomainDeviceDefPtr device ATTRIBUTE_UNUSED,
                                    virDomainDeviceInfoPtr info,
                                    void *opaque)
{
    virDomainPCIAddressSetPtr addrs = opaque;
    virPCIDeviceAddressPtr addr = &info->addr.pci;

    if (info->type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI)
        addr->extFlags = info->pciAddrExtFlags;

    if (virDeviceInfoPCIAddressExtensionIsWanted(info))
        return virDomainPCIAddressExtensionReserveNextAddr(addrs, addr);

    return 0;
}

1428 1429 1430 1431 1432 1433 1434 1435
static int
qemuDomainCollectPCIAddress(virDomainDefPtr def ATTRIBUTE_UNUSED,
                            virDomainDeviceDefPtr device,
                            virDomainDeviceInfoPtr info,
                            void *opaque)
{
    virDomainPCIAddressSetPtr addrs = opaque;
    int ret = -1;
1436
    virPCIDeviceAddressPtr addr = &info->addr.pci;
1437

1438
    if (!virDeviceInfoPCIAddressIsPresent(info) ||
1439 1440
        ((device->type == VIR_DOMAIN_DEVICE_HOSTDEV) &&
         (device->data.hostdev->parent.type != VIR_DOMAIN_DEVICE_NONE))) {
1441 1442 1443 1444 1445 1446 1447
        /* If a hostdev has a parent, its info will be a part of the
         * parent, and will have its address collected during the scan
         * of the parent's device type.
        */
        return 0;
    }

1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459
    /* If we get to here, the device has a PCI address assigned in the
     * config and we should mark it as in-use. But if the
     * pciConnectFlags are 0, then this device shouldn't have a PCI
     * address associated with it. *BUT* since there are cases in the
     * past where we've apparently allowed that, we need to pretend
     * for now that it's okay, otherwise an existing domain could
     * "disappear" from the list of domains due to a parse failure. We
     * can fix this by just forcing the pciConnectFlags to be
     * PCI_DEVICE (and then relying on validation functions to report
     * inappropriate address types.
     */
    if (!info->pciConnectFlags) {
1460
        char *addrStr = virPCIDeviceAddressAsString(&info->addr.pci);
1461 1462 1463 1464 1465 1466 1467 1468 1469

        VIR_WARN("qemuDomainDeviceCalculatePCIConnectFlags() thinks that the "
                 "device with PCI address %s should not have a PCI address",
                 addrStr ? addrStr : "(unknown)");
        VIR_FREE(addrStr);

        info->pciConnectFlags = VIR_PCI_CONNECT_TYPE_PCI_DEVICE;
    }

1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489
    /* Ignore implicit controllers on slot 0:0:1.0:
     * implicit IDE controller on 0:0:1.1 (no qemu command line)
     * implicit USB controller on 0:0:1.2 (-usb)
     *
     * If the machine does have a PCI bus, they will get reserved
     * in qemuDomainAssignDevicePCISlots().
     */

    /* These are the IDE and USB controllers in the PIIX3, hardcoded
     * to bus 0 slot 1.  They cannot be attached to a PCIe slot, only
     * PCI.
     */
    if (device->type == VIR_DOMAIN_DEVICE_CONTROLLER && addr->domain == 0 &&
        addr->bus == 0 && addr->slot == 1) {
        virDomainControllerDefPtr cont = device->data.controller;

        if ((cont->type == VIR_DOMAIN_CONTROLLER_TYPE_IDE && cont->idx == 0 &&
             addr->function == 1) ||
            (cont->type == VIR_DOMAIN_CONTROLLER_TYPE_USB && cont->idx == 0 &&
             (cont->model == VIR_DOMAIN_CONTROLLER_MODEL_USB_PIIX3_UHCI ||
1490 1491
              cont->model == VIR_DOMAIN_CONTROLLER_MODEL_USB_DEFAULT) &&
             addr->function == 2)) {
1492 1493 1494 1495 1496 1497 1498
            /* Note the check for nbuses > 0 - if there are no PCI
             * buses, we skip this check. This is a quirk required for
             * some machinetypes such as s390, which pretend to have a
             * PCI bus for long enough to generate the "-usb" on the
             * commandline, but that don't really care if a PCI bus
             * actually exists. */
            if (addrs->nbuses > 0 &&
1499
                !(addrs->buses[0].flags & VIR_PCI_CONNECT_TYPE_PCI_DEVICE)) {
1500 1501 1502 1503 1504 1505 1506 1507 1508 1509
                virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                               _("Bus 0 must be PCI for integrated PIIX3 "
                                 "USB or IDE controllers"));
                return -1;
            } else {
                return 0;
            }
        }
    }

1510
    if (virDomainPCIAddressReserveAddr(addrs, addr,
1511 1512
                                       info->pciConnectFlags,
                                       info->isolationGroup) < 0) {
1513
        goto cleanup;
1514
    }
1515 1516 1517 1518 1519 1520

    ret = 0;
 cleanup:
    return ret;
}

1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545
static int
qemuDomainCollectPCIAddressExtension(virDomainDefPtr def ATTRIBUTE_UNUSED,
                                     virDomainDeviceDefPtr device,
                                     virDomainDeviceInfoPtr info,
                                     void *opaque)
{
    virDomainPCIAddressSetPtr addrs = opaque;
    virPCIDeviceAddressPtr addr = &info->addr.pci;

    if (info->type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI)
        addr->extFlags = info->pciAddrExtFlags;

    if (!virDeviceInfoPCIAddressExtensionIsPresent(info) ||
        ((device->type == VIR_DOMAIN_DEVICE_HOSTDEV) &&
         (device->data.hostdev->parent.type != VIR_DOMAIN_DEVICE_NONE))) {
        /* If a hostdev has a parent, its info will be a part of the
         * parent, and will have its address collected during the scan
         * of the parent's device type.
        */
        return 0;
    }

    return virDomainPCIAddressExtensionReserveAddr(addrs, addr);
}

1546 1547
static virDomainPCIAddressSetPtr
qemuDomainPCIAddressSetCreate(virDomainDefPtr def,
1548
                              virQEMUCapsPtr qemuCaps,
1549 1550 1551 1552 1553
                              unsigned int nbuses,
                              bool dryRun)
{
    virDomainPCIAddressSetPtr addrs;
    size_t i;
1554 1555
    bool hasPCIeRoot = false;
    virDomainControllerModelPCI defaultModel;
1556
    virPCIDeviceAddressExtensionFlags extFlags = VIR_PCI_ADDRESS_EXTENSION_NONE;
1557

1558 1559 1560 1561
    if (virQEMUCapsGet(qemuCaps, QEMU_CAPS_DEVICE_ZPCI))
        extFlags |= VIR_PCI_ADDRESS_EXTENSION_ZPCI;

    if ((addrs = virDomainPCIAddressSetAlloc(nbuses, extFlags)) == NULL)
1562 1563 1564 1565
        return NULL;

    addrs->dryRun = dryRun;

1566 1567
    /* pSeries domains support multiple pci-root controllers */
    if (qemuDomainIsPSeries(def))
1568
        addrs->areMultipleRootsSupported = true;
1569

1570 1571 1572
    if (virQEMUCapsGet(qemuCaps, QEMU_CAPS_DEVICE_PCIE_PCI_BRIDGE))
        addrs->isPCIeToPCIBridgeSupported = true;

1573
    for (i = 0; i < def->ncontrollers; i++) {
1574 1575
        virDomainControllerDefPtr cont = def->controllers[i];
        size_t idx = cont->idx;
1576

1577
        if (cont->type != VIR_DOMAIN_CONTROLLER_TYPE_PCI)
1578 1579 1580 1581 1582
            continue;

        if (idx >= addrs->nbuses) {
            virReportError(VIR_ERR_INTERNAL_ERROR,
                           _("Inappropriate new pci controller index %zu "
1583
                             "exceeds addrs array length"), idx);
1584 1585 1586
            goto error;
        }

1587
        if (virDomainPCIAddressBusSetModel(&addrs->buses[idx], cont->model) < 0)
1588
            goto error;
1589

1590 1591 1592 1593
        /* Forward the information about isolation groups */
        addrs->buses[idx].isolationGroup = cont->info.isolationGroup;
        addrs->buses[idx].isolationGroupLocked = cont->info.isolationGroupLocked;

1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612
        if (cont->model == VIR_DOMAIN_CONTROLLER_MODEL_PCIE_ROOT)
            hasPCIeRoot = true;
    }

    if (nbuses > 0 && !addrs->buses[0].model) {
        /* This is just here to replicate a safety measure already in
         * an older version of this code. In practice, the root bus
         * should have already been added at index 0 prior to
         * assigning addresses to devices.
         */
        if (virDomainPCIAddressBusSetModel(&addrs->buses[0],
                                           VIR_DOMAIN_CONTROLLER_MODEL_PCI_ROOT) < 0)
            goto error;
    }

    /* Now fill in a reasonable model for all the buses in the set
     * that don't yet have a corresponding controller in the domain
     * config.
     */
1613 1614 1615 1616
    if (qemuDomainIsPSeries(def)) {
        /* pSeries guests should use PHBs (pci-root controllers) */
        defaultModel = VIR_DOMAIN_CONTROLLER_MODEL_PCI_ROOT;
    } else if (hasPCIeRoot) {
1617
        defaultModel = VIR_DOMAIN_CONTROLLER_MODEL_PCIE_ROOT_PORT;
1618
    } else {
1619
        defaultModel = VIR_DOMAIN_CONTROLLER_MODEL_PCI_BRIDGE;
1620
    }
1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632

    for (i = 1; i < addrs->nbuses; i++) {

        if (addrs->buses[i].model)
            continue;

        if (virDomainPCIAddressBusSetModel(&addrs->buses[i], defaultModel) < 0)
            goto error;

        VIR_DEBUG("Auto-adding <controller type='pci' model='%s' index='%zu'/>",
                  virDomainControllerModelPCITypeToString(defaultModel), i);
    }
1633 1634 1635 1636

    if (virDomainDeviceInfoIterate(def, qemuDomainCollectPCIAddress, addrs) < 0)
        goto error;

1637 1638 1639 1640 1641 1642
    if (virDomainDeviceInfoIterate(def,
                                   qemuDomainCollectPCIAddressExtension,
                                   addrs) < 0) {
        goto error;
    }

1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657
    return addrs;

 error:
    virDomainPCIAddressSetFree(addrs);
    return NULL;
}


static int
qemuDomainValidateDevicePCISlotsPIIX3(virDomainDefPtr def,
                                      virQEMUCapsPtr qemuCaps,
                                      virDomainPCIAddressSetPtr addrs)
{
    int ret = -1;
    size_t i;
1658
    virPCIDeviceAddress tmp_addr;
1659 1660
    bool qemuDeviceVideoUsable = virQEMUCapsGet(qemuCaps, QEMU_CAPS_DEVICE_VIDEO_PRIMARY);
    char *addrStr = NULL;
1661 1662
    virDomainPCIConnectFlags flags = (VIR_PCI_CONNECT_HOTPLUGGABLE
                                      | VIR_PCI_CONNECT_TYPE_PCI_DEVICE);
1663 1664 1665

    /* Verify that first IDE and USB controllers (if any) is on the PIIX3, fn 1 */
    for (i = 0; i < def->ncontrollers; i++) {
1666 1667
        virDomainControllerDefPtr cont = def->controllers[i];

1668
        /* First IDE controller lives on the PIIX3 at slot=1, function=1 */
1669 1670
        if (cont->type == VIR_DOMAIN_CONTROLLER_TYPE_IDE &&
            cont->idx == 0) {
1671
            if (virDeviceInfoPCIAddressIsPresent(&cont->info)) {
1672 1673 1674 1675
                if (cont->info.addr.pci.domain != 0 ||
                    cont->info.addr.pci.bus != 0 ||
                    cont->info.addr.pci.slot != 1 ||
                    cont->info.addr.pci.function != 1) {
1676 1677 1678 1679 1680
                    virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                                   _("Primary IDE controller must have PCI address 0:0:1.1"));
                    goto cleanup;
                }
            } else {
1681 1682 1683 1684 1685
                cont->info.type = VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI;
                cont->info.addr.pci.domain = 0;
                cont->info.addr.pci.bus = 0;
                cont->info.addr.pci.slot = 1;
                cont->info.addr.pci.function = 1;
1686
            }
1687 1688 1689
        } else if (cont->type == VIR_DOMAIN_CONTROLLER_TYPE_USB &&
                   cont->idx == 0 &&
                   (cont->model == VIR_DOMAIN_CONTROLLER_MODEL_USB_PIIX3_UHCI ||
1690
                    cont->model == VIR_DOMAIN_CONTROLLER_MODEL_USB_DEFAULT)) {
1691
            if (virDeviceInfoPCIAddressIsPresent(&cont->info)) {
1692 1693 1694 1695
                if (cont->info.addr.pci.domain != 0 ||
                    cont->info.addr.pci.bus != 0 ||
                    cont->info.addr.pci.slot != 1 ||
                    cont->info.addr.pci.function != 2) {
1696
                    virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
1697
                                   _("PIIX3 USB controller at index 0 must have PCI address 0:0:1.2"));
1698 1699 1700
                    goto cleanup;
                }
            } else {
1701 1702 1703 1704 1705
                cont->info.type = VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI;
                cont->info.addr.pci.domain = 0;
                cont->info.addr.pci.bus = 0;
                cont->info.addr.pci.slot = 1;
                cont->info.addr.pci.function = 2;
1706
            }
1707 1708 1709
        } else {
            /* this controller is not skipped in qemuDomainCollectPCIAddress */
            continue;
1710
        }
1711 1712 1713
        if (addrs->nbuses &&
            virDomainPCIAddressReserveAddr(addrs, &cont->info.addr.pci, flags, 0) < 0)
            goto cleanup;
1714 1715
    }

1716
    /* Implicit PIIX3 devices living on slot 1 not handled above */
1717 1718 1719
    if (addrs->nbuses) {
        memset(&tmp_addr, 0, sizeof(tmp_addr));
        tmp_addr.slot = 1;
1720 1721 1722 1723 1724
        /* ISA Bridge at 00:01.0 */
        if (virDomainPCIAddressReserveAddr(addrs, &tmp_addr, flags, 0) < 0)
            goto cleanup;
        /* Bridge at 00:01.3 */
        tmp_addr.function = 3;
1725
        if (virDomainPCIAddressReserveAddr(addrs, &tmp_addr, flags, 0) < 0)
1726 1727 1728
            goto cleanup;
    }

1729 1730
    if (def->nvideos > 0 &&
        def->videos[0]->type != VIR_DOMAIN_VIDEO_TYPE_NONE) {
1731 1732 1733 1734 1735 1736 1737
        /* Because the PIIX3 integrated IDE/USB controllers are
         * already at slot 1, when qemu looks for the first free slot
         * to place the VGA controller (which is always the first
         * device added after integrated devices), it *always* ends up
         * at slot 2.
         */
        virDomainVideoDefPtr primaryVideo = def->videos[0];
1738

1739
        if (virDeviceInfoPCIAddressIsWanted(&primaryVideo->info)) {
1740 1741 1742
            memset(&tmp_addr, 0, sizeof(tmp_addr));
            tmp_addr.slot = 2;

1743
            if (!(addrStr = virPCIDeviceAddressAsString(&tmp_addr)))
1744 1745
                goto cleanup;
            if (!virDomainPCIAddressValidate(addrs, &tmp_addr,
1746
                                             addrStr, flags, true))
1747 1748 1749 1750
                goto cleanup;

            if (virDomainPCIAddressSlotInUse(addrs, &tmp_addr)) {
                if (qemuDeviceVideoUsable) {
1751
                    if (qemuDomainPCIAddressReserveNextAddr(addrs,
1752
                                                            &primaryVideo->info) < 0) {
1753
                        goto cleanup;
1754
                    }
1755 1756 1757 1758 1759 1760 1761
                } else {
                    virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                                   _("PCI address 0:0:2.0 is in use, "
                                     "QEMU needs it for primary video"));
                    goto cleanup;
                }
            } else {
1762
                if (virDomainPCIAddressReserveAddr(addrs, &tmp_addr, flags, 0) < 0)
1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786
                    goto cleanup;
                primaryVideo->info.addr.pci = tmp_addr;
                primaryVideo->info.type = VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI;
            }
        } else if (!qemuDeviceVideoUsable) {
            if (primaryVideo->info.addr.pci.domain != 0 ||
                primaryVideo->info.addr.pci.bus != 0 ||
                primaryVideo->info.addr.pci.slot != 2 ||
                primaryVideo->info.addr.pci.function != 0) {
                virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                               _("Primary video card must have PCI address 0:0:2.0"));
                goto cleanup;
            }
            /* If TYPE == PCI, then qemuDomainCollectPCIAddress() function
             * has already reserved the address, so we must skip */
        }
    } else if (addrs->nbuses && !qemuDeviceVideoUsable) {
        memset(&tmp_addr, 0, sizeof(tmp_addr));
        tmp_addr.slot = 2;

        if (virDomainPCIAddressSlotInUse(addrs, &tmp_addr)) {
            VIR_DEBUG("PCI address 0:0:2.0 in use, future addition of a video"
                      " device will not be possible without manual"
                      " intervention");
1787
        } else if (virDomainPCIAddressReserveAddr(addrs, &tmp_addr, flags, 0) < 0) {
1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804
            goto cleanup;
        }
    }
    ret = 0;
 cleanup:
    VIR_FREE(addrStr);
    return ret;
}


static int
qemuDomainValidateDevicePCISlotsQ35(virDomainDefPtr def,
                                    virQEMUCapsPtr qemuCaps,
                                    virDomainPCIAddressSetPtr addrs)
{
    int ret = -1;
    size_t i;
1805
    virPCIDeviceAddress tmp_addr;
1806 1807
    bool qemuDeviceVideoUsable = virQEMUCapsGet(qemuCaps, QEMU_CAPS_DEVICE_VIDEO_PRIMARY);
    char *addrStr = NULL;
1808
    virDomainPCIConnectFlags flags = VIR_PCI_CONNECT_TYPE_PCIE_DEVICE;
1809 1810

    for (i = 0; i < def->ncontrollers; i++) {
1811 1812 1813
        virDomainControllerDefPtr cont = def->controllers[i];

        switch (cont->type) {
1814 1815 1816 1817 1818
        case VIR_DOMAIN_CONTROLLER_TYPE_SATA:
            /* Verify that the first SATA controller is at 00:1F.2 the
             * q35 machine type *always* has a SATA controller at this
             * address.
             */
1819
            if (cont->idx == 0) {
1820
                if (virDeviceInfoPCIAddressIsPresent(&cont->info)) {
1821 1822 1823 1824
                    if (cont->info.addr.pci.domain != 0 ||
                        cont->info.addr.pci.bus != 0 ||
                        cont->info.addr.pci.slot != 0x1F ||
                        cont->info.addr.pci.function != 2) {
1825 1826 1827 1828 1829
                        virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                                       _("Primary SATA controller must have PCI address 0:0:1f.2"));
                        goto cleanup;
                    }
                } else {
1830 1831 1832 1833 1834
                    cont->info.type = VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI;
                    cont->info.addr.pci.domain = 0;
                    cont->info.addr.pci.bus = 0;
                    cont->info.addr.pci.slot = 0x1F;
                    cont->info.addr.pci.function = 2;
1835 1836 1837 1838 1839
                }
            }
            break;

        case VIR_DOMAIN_CONTROLLER_TYPE_USB:
1840 1841
            if ((cont->model == VIR_DOMAIN_CONTROLLER_MODEL_USB_ICH9_UHCI1) &&
                (cont->info.type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_NONE)) {
1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862
                /* Try to assign the first found USB2 controller to
                 * 00:1D.0 and 2nd to 00:1A.0 (because that is their
                 * standard location on real Q35 hardware) unless they
                 * are already taken, but don't insist on it.
                 *
                 * (NB: all other controllers at the same index will
                 * get assigned to the same slot as the UHCI1 when
                 * addresses are later assigned to all devices.)
                 */
                bool assign = false;

                memset(&tmp_addr, 0, sizeof(tmp_addr));
                tmp_addr.slot = 0x1D;
                if (!virDomainPCIAddressSlotInUse(addrs, &tmp_addr)) {
                    assign = true;
                } else {
                    tmp_addr.slot = 0x1A;
                    if (!virDomainPCIAddressSlotInUse(addrs, &tmp_addr))
                        assign = true;
                }
                if (assign) {
1863
                    if (virDomainPCIAddressReserveAddr(addrs, &tmp_addr, flags, 0) < 0)
1864
                        goto cleanup;
1865

1866 1867 1868 1869 1870 1871
                    cont->info.type = VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI;
                    cont->info.addr.pci.domain = 0;
                    cont->info.addr.pci.bus = 0;
                    cont->info.addr.pci.slot = tmp_addr.slot;
                    cont->info.addr.pci.function = 0;
                    cont->info.addr.pci.multi = VIR_TRISTATE_SWITCH_ON;
1872 1873 1874 1875 1876
                }
            }
            break;

        case VIR_DOMAIN_CONTROLLER_TYPE_PCI:
1877 1878
            if (cont->model == VIR_DOMAIN_CONTROLLER_MODEL_DMI_TO_PCI_BRIDGE &&
                cont->info.type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_NONE) {
1879 1880 1881 1882 1883 1884 1885
                /* Try to assign this bridge to 00:1E.0 (because that
                * is its standard location on real hardware) unless
                * it's already taken, but don't insist on it.
                */
                memset(&tmp_addr, 0, sizeof(tmp_addr));
                tmp_addr.slot = 0x1E;
                if (!virDomainPCIAddressSlotInUse(addrs, &tmp_addr)) {
1886
                    if (virDomainPCIAddressReserveAddr(addrs, &tmp_addr, flags, 0) < 0)
1887
                        goto cleanup;
1888

1889 1890 1891 1892 1893
                    cont->info.type = VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI;
                    cont->info.addr.pci.domain = 0;
                    cont->info.addr.pci.bus = 0;
                    cont->info.addr.pci.slot = 0x1E;
                    cont->info.addr.pci.function = 0;
1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909
                }
            }
            break;
        }
    }

    /* Reserve slot 0x1F function 0 (ISA bridge, not in config model)
     * and function 3 (SMBus, also not (yet) in config model). As with
     * the SATA controller, these devices are always present in a q35
     * machine; there is no way to not have them.
     */
    if (addrs->nbuses) {
        memset(&tmp_addr, 0, sizeof(tmp_addr));
        tmp_addr.slot = 0x1F;
        tmp_addr.function = 0;
        tmp_addr.multi = VIR_TRISTATE_SWITCH_ON;
1910
        if (virDomainPCIAddressReserveAddr(addrs, &tmp_addr, flags, 0) < 0)
1911
           goto cleanup;
1912

1913 1914
        tmp_addr.function = 3;
        tmp_addr.multi = VIR_TRISTATE_SWITCH_ABSENT;
1915
        if (virDomainPCIAddressReserveAddr(addrs, &tmp_addr, flags, 0) < 0)
1916 1917 1918
           goto cleanup;
    }

1919 1920
    if (def->nvideos > 0 &&
        def->videos[0]->type != VIR_DOMAIN_VIDEO_TYPE_NONE) {
1921 1922
        /* NB: unlike the pc machinetypes, on q35 machinetypes the
         * integrated devices are at slot 0x1f, so when qemu looks for
1923
         * the first free slot for the first VGA, it will always be at
1924 1925 1926 1927
         * slot 1 (which was used up by the integrated PIIX3 devices
         * on pc machinetypes).
         */
        virDomainVideoDefPtr primaryVideo = def->videos[0];
1928
        if (virDeviceInfoPCIAddressIsWanted(&primaryVideo->info)) {
1929 1930 1931
            memset(&tmp_addr, 0, sizeof(tmp_addr));
            tmp_addr.slot = 1;

1932
            if (!(addrStr = virPCIDeviceAddressAsString(&tmp_addr)))
1933 1934
                goto cleanup;
            if (!virDomainPCIAddressValidate(addrs, &tmp_addr,
1935
                                             addrStr, flags, true))
1936 1937 1938 1939
                goto cleanup;

            if (virDomainPCIAddressSlotInUse(addrs, &tmp_addr)) {
                if (qemuDeviceVideoUsable) {
1940
                    if (qemuDomainPCIAddressReserveNextAddr(addrs,
1941
                                                            &primaryVideo->info) < 0)
1942 1943 1944 1945 1946 1947 1948 1949
                        goto cleanup;
                } else {
                    virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                                   _("PCI address 0:0:1.0 is in use, "
                                     "QEMU needs it for primary video"));
                    goto cleanup;
                }
            } else {
1950
                if (virDomainPCIAddressReserveAddr(addrs, &tmp_addr, flags, 0) < 0)
1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975
                    goto cleanup;
                primaryVideo->info.type = VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI;
                primaryVideo->info.addr.pci = tmp_addr;
            }
        } else if (!qemuDeviceVideoUsable) {
            if (primaryVideo->info.addr.pci.domain != 0 ||
                primaryVideo->info.addr.pci.bus != 0 ||
                primaryVideo->info.addr.pci.slot != 1 ||
                primaryVideo->info.addr.pci.function != 0) {
                virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                               _("Primary video card must have PCI address 0:0:1.0"));
                goto cleanup;
            }
            /* If TYPE == PCI, then qemuDomainCollectPCIAddress() function
             * has already reserved the address, so we must skip */
        }
    } else if (addrs->nbuses && !qemuDeviceVideoUsable) {
        memset(&tmp_addr, 0, sizeof(tmp_addr));
        tmp_addr.slot = 1;

        if (virDomainPCIAddressSlotInUse(addrs, &tmp_addr)) {
            VIR_DEBUG("PCI address 0:0:1.0 in use, future addition of a video"
                      " device will not be possible without manual"
                      " intervention");
            virResetLastError();
1976
        } else if (virDomainPCIAddressReserveAddr(addrs, &tmp_addr, flags, 0) < 0) {
1977 1978 1979
            goto cleanup;
        }
    }
1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993

    memset(&tmp_addr, 0, sizeof(tmp_addr));
    tmp_addr.slot = 0x1B;
    if (!virDomainPCIAddressSlotInUse(addrs, &tmp_addr)) {
        /* Since real Q35 hardware has an ICH9 chip that has an
         * integrated HD audio device at 0000:00:1B.0 put any
         * unaddressed ICH9 audio device at that address if it's not
         * already taken. If there's something already there, let the
         * normal device addressing assign something later.
         */
        for (i = 0; i < def->nsounds; i++) {
            virDomainSoundDefPtr sound = def->sounds[i];

            if (sound->model != VIR_DOMAIN_SOUND_MODEL_ICH9 ||
1994
                !virDeviceInfoPCIAddressIsWanted(&sound->info)) {
1995 1996
                continue;
            }
1997
            if (virDomainPCIAddressReserveAddr(addrs, &tmp_addr, flags, 0) < 0)
1998 1999 2000 2001 2002 2003 2004 2005
                goto cleanup;

            sound->info.type = VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI;
            sound->info.addr.pci = tmp_addr;
            break;
        }
    }

2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017
    ret = 0;
 cleanup:
    VIR_FREE(addrStr);
    return ret;
}


static int
qemuDomainValidateDevicePCISlotsChipsets(virDomainDefPtr def,
                                         virQEMUCapsPtr qemuCaps,
                                         virDomainPCIAddressSetPtr addrs)
{
2018
    if (qemuDomainIsI440FX(def) &&
2019 2020 2021 2022
        qemuDomainValidateDevicePCISlotsPIIX3(def, qemuCaps, addrs) < 0) {
        return -1;
    }

2023
    if (qemuDomainIsQ35(def) &&
2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049
        qemuDomainValidateDevicePCISlotsQ35(def, qemuCaps, addrs) < 0) {
        return -1;
    }

    return 0;
}


/*
 * This assigns static PCI slots to all configured devices.
 * The ordering here is chosen to match the ordering used
 * with old QEMU < 0.12, so that if a user updates a QEMU
 * host from old QEMU to QEMU >= 0.12, their guests should
 * get PCI addresses in the same order as before.
 *
 * NB, if they previously hotplugged devices then all bets
 * are off. Hotplug for old QEMU was unfixably broken wrt
 * to stable PCI addressing.
 *
 * Order is:
 *
 *  - Host bridge (slot 0)
 *  - PIIX3 ISA bridge, IDE controller, something else unknown, USB controller (slot 1)
 *  - Video (slot 2)
 *
 *  - These integrated devices were already added by
2050
 *    qemuDomainValidateDevicePCISlotsChipsets invoked right before this function
2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076
 *
 * Incrementally assign slots from 3 onwards:
 *
 *  - Net
 *  - Sound
 *  - SCSI controllers
 *  - VirtIO block
 *  - VirtIO balloon
 *  - Host device passthrough
 *  - Watchdog
 *  - pci serial devices
 *
 * Prior to this function being invoked, qemuDomainCollectPCIAddress() will have
 * added all existing PCI addresses from the 'def' to 'addrs'. Thus this
 * function must only try to reserve addresses if info.type == NONE and
 * skip over info.type == PCI
 */
static int
qemuDomainAssignDevicePCISlots(virDomainDefPtr def,
                               virQEMUCapsPtr qemuCaps,
                               virDomainPCIAddressSetPtr addrs)
{
    size_t i, j;

    /* PCI controllers */
    for (i = 0; i < def->ncontrollers; i++) {
2077 2078 2079 2080
        virDomainControllerDefPtr cont = def->controllers[i];

        if (cont->type == VIR_DOMAIN_CONTROLLER_TYPE_PCI) {
            virDomainControllerModelPCI model = cont->model;
2081

2082 2083
            if (model == VIR_DOMAIN_CONTROLLER_MODEL_PCI_ROOT ||
                model == VIR_DOMAIN_CONTROLLER_MODEL_PCIE_ROOT ||
2084
                !virDeviceInfoPCIAddressIsWanted(&cont->info))
2085
                continue;
2086

2087
            if (qemuDomainPCIAddressReserveNextAddr(addrs, &cont->info) < 0)
2088 2089 2090 2091 2092
                goto error;
        }
    }

    for (i = 0; i < def->nfss; i++) {
2093
        if (!virDeviceInfoPCIAddressIsWanted(&def->fss[i]->info))
2094 2095 2096 2097
            continue;

        /* Only support VirtIO-9p-pci so far. If that changes,
         * we might need to skip devices here */
2098
        if (qemuDomainPCIAddressReserveNextAddr(addrs, &def->fss[i]->info) < 0)
2099 2100 2101 2102 2103
            goto error;
    }

    /* Network interfaces */
    for (i = 0; i < def->nnets; i++) {
2104 2105
        virDomainNetDefPtr net = def->nets[i];

2106 2107 2108 2109
        /* type='hostdev' network devices might be USB, and are also
         * in hostdevs list anyway, so handle them with other hostdevs
         * instead of here.
         */
2110
        if ((net->type == VIR_DOMAIN_NET_TYPE_HOSTDEV) ||
2111
            !virDeviceInfoPCIAddressIsWanted(&net->info)) {
2112 2113
            continue;
        }
2114

2115
        if (qemuDomainPCIAddressReserveNextAddr(addrs, &net->info) < 0)
2116 2117 2118 2119 2120
            goto error;
    }

    /* Sound cards */
    for (i = 0; i < def->nsounds; i++) {
2121 2122
        virDomainSoundDefPtr sound = def->sounds[i];

2123
        if (!virDeviceInfoPCIAddressIsWanted(&sound->info))
2124
            continue;
2125

2126
        /* Skip ISA sound card, PCSPK and usb-audio */
2127 2128 2129
        if (sound->model == VIR_DOMAIN_SOUND_MODEL_SB16 ||
            sound->model == VIR_DOMAIN_SOUND_MODEL_PCSPK ||
            sound->model == VIR_DOMAIN_SOUND_MODEL_USB) {
2130
            continue;
2131
        }
2132

2133
        if (qemuDomainPCIAddressReserveNextAddr(addrs, &sound->info) < 0)
2134 2135 2136 2137 2138
            goto error;
    }

    /* Device controllers (SCSI, USB, but not IDE, FDC or CCID) */
    for (i = 0; i < def->ncontrollers; i++) {
2139 2140
        virDomainControllerDefPtr cont = def->controllers[i];

2141
        /* PCI controllers have been dealt with earlier */
2142
        if (cont->type == VIR_DOMAIN_CONTROLLER_TYPE_PCI)
2143 2144 2145
            continue;

        /* USB controller model 'none' doesn't need a PCI address */
2146 2147
        if (cont->type == VIR_DOMAIN_CONTROLLER_TYPE_USB &&
            cont->model == VIR_DOMAIN_CONTROLLER_MODEL_USB_NONE)
2148 2149 2150
            continue;

        /* FDC lives behind the ISA bridge; CCID is a usb device */
2151 2152
        if (cont->type == VIR_DOMAIN_CONTROLLER_TYPE_FDC ||
            cont->type == VIR_DOMAIN_CONTROLLER_TYPE_CCID)
2153 2154 2155 2156
            continue;

        /* First IDE controller lives on the PIIX3 at slot=1, function=1,
           dealt with earlier on*/
2157 2158
        if (cont->type == VIR_DOMAIN_CONTROLLER_TYPE_IDE &&
            cont->idx == 0)
2159 2160
            continue;

2161
        if (!virDeviceInfoPCIAddressIsWanted(&cont->info))
2162 2163 2164
            continue;

        /* USB2 needs special handling to put all companions in the same slot */
2165
        if (IS_USB2_CONTROLLER(cont)) {
2166
            virPCIDeviceAddress addr = {0};
2167 2168 2169 2170
            bool foundAddr = false;

            for (j = 0; j < def->ncontrollers; j++) {
                if (IS_USB2_CONTROLLER(def->controllers[j]) &&
2171
                    def->controllers[j]->idx == cont->idx &&
2172
                    virDeviceInfoPCIAddressIsPresent(&def->controllers[j]->info)) {
2173 2174 2175 2176 2177 2178
                    addr = def->controllers[j]->info.addr.pci;
                    foundAddr = true;
                    break;
                }
            }

2179
            switch (cont->model) {
2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197
            case VIR_DOMAIN_CONTROLLER_MODEL_USB_ICH9_EHCI1:
                addr.function = 7;
                addr.multi = VIR_TRISTATE_SWITCH_ABSENT;
                break;
            case VIR_DOMAIN_CONTROLLER_MODEL_USB_ICH9_UHCI1:
                addr.function = 0;
                addr.multi = VIR_TRISTATE_SWITCH_ON;
                break;
            case VIR_DOMAIN_CONTROLLER_MODEL_USB_ICH9_UHCI2:
                addr.function = 1;
                addr.multi = VIR_TRISTATE_SWITCH_ABSENT;
                break;
            case VIR_DOMAIN_CONTROLLER_MODEL_USB_ICH9_UHCI3:
                addr.function = 2;
                addr.multi = VIR_TRISTATE_SWITCH_ABSENT;
                break;
            }

2198 2199
            if (foundAddr) {
                /* Reserve this function on the slot we found */
2200
                if (virDomainPCIAddressReserveAddr(addrs, &addr,
2201 2202
                                                   cont->info.pciConnectFlags,
                                                   cont->info.isolationGroup) < 0) {
2203
                    goto error;
2204
                }
2205

2206 2207 2208 2209 2210
                cont->info.type = VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI;
                cont->info.addr.pci = addr;
            } else {
                /* This is the first part of the controller, so need
                 * to find a free slot & then reserve this function */
2211 2212 2213
                if (virDomainPCIAddressReserveNextAddr(addrs, &cont->info,
                                                       cont->info.pciConnectFlags,
                                                       addr.function) < 0) {
2214 2215
                    goto error;
                }
2216

2217
                cont->info.addr.pci.multi = addr.multi;
2218 2219
            }
        } else {
2220
            if (qemuDomainPCIAddressReserveNextAddr(addrs, &cont->info) < 0)
2221
                 goto error;
2222 2223 2224 2225 2226 2227 2228 2229 2230 2231
        }
    }

    /* Disks (VirtIO only for now) */
    for (i = 0; i < def->ndisks; i++) {
        /* Only VirtIO disks use PCI addrs */
        if (def->disks[i]->bus != VIR_DOMAIN_DISK_BUS_VIRTIO)
            continue;

        /* don't touch s390 devices */
2232
        if (virDeviceInfoPCIAddressIsPresent(&def->disks[i]->info) ||
2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244
            def->disks[i]->info.type ==
            VIR_DOMAIN_DEVICE_ADDRESS_TYPE_VIRTIO_S390 ||
            def->disks[i]->info.type ==
            VIR_DOMAIN_DEVICE_ADDRESS_TYPE_CCW)
            continue;

        /* Also ignore virtio-mmio disks if our machine allows them */
        if (def->disks[i]->info.type ==
            VIR_DOMAIN_DEVICE_ADDRESS_TYPE_VIRTIO_MMIO &&
            virQEMUCapsGet(qemuCaps, QEMU_CAPS_DEVICE_VIRTIO_MMIO))
            continue;

2245
        if (!virDeviceInfoPCIAddressIsWanted(&def->disks[i]->info)) {
2246 2247 2248 2249 2250 2251
            virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
                           _("virtio disk cannot have an address of type '%s'"),
                           virDomainDeviceAddressTypeToString(def->disks[i]->info.type));
            goto error;
        }

2252
        if (qemuDomainPCIAddressReserveNextAddr(addrs, &def->disks[i]->info) < 0)
2253 2254 2255 2256 2257
            goto error;
    }

    /* Host PCI devices */
    for (i = 0; i < def->nhostdevs; i++) {
2258
        virDomainHostdevSubsysPtr subsys = &def->hostdevs[i]->source.subsys;
2259
        if (!virDeviceInfoPCIAddressIsWanted(def->hostdevs[i]->info))
2260
            continue;
2261 2262
        if (def->hostdevs[i]->mode != VIR_DOMAIN_HOSTDEV_MODE_SUBSYS)
            continue;
2263 2264 2265 2266
        if (subsys->type != VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_PCI &&
            subsys->type != VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_SCSI_HOST &&
            !(subsys->type == VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_MDEV &&
              subsys->u.mdev.model == VIR_MDEV_MODEL_TYPE_VFIO_PCI)) {
2267
            continue;
2268
        }
2269

2270
        if (qemuDomainPCIAddressReserveNextAddr(addrs,
2271
                                                def->hostdevs[i]->info) < 0)
2272 2273 2274 2275 2276 2277
            goto error;
    }

    /* VirtIO balloon */
    if (def->memballoon &&
        def->memballoon->model == VIR_DOMAIN_MEMBALLOON_MODEL_VIRTIO &&
2278
        virDeviceInfoPCIAddressIsWanted(&def->memballoon->info)) {
2279

2280
        if (qemuDomainPCIAddressReserveNextAddr(addrs,
2281
                                                &def->memballoon->info) < 0)
2282 2283 2284 2285 2286 2287
            goto error;
    }

    /* VirtIO RNG */
    for (i = 0; i < def->nrngs; i++) {
        if (def->rngs[i]->model != VIR_DOMAIN_RNG_MODEL_VIRTIO ||
2288
            !virDeviceInfoPCIAddressIsWanted(&def->rngs[i]->info))
2289 2290
            continue;

2291
        if (qemuDomainPCIAddressReserveNextAddr(addrs, &def->rngs[i]->info) < 0)
2292 2293 2294 2295 2296 2297
            goto error;
    }

    /* A watchdog - check if it is a PCI device */
    if (def->watchdog &&
        def->watchdog->model == VIR_DOMAIN_WATCHDOG_MODEL_I6300ESB &&
2298
        virDeviceInfoPCIAddressIsWanted(&def->watchdog->info)) {
2299
        if (qemuDomainPCIAddressReserveNextAddr(addrs, &def->watchdog->info) < 0)
2300 2301 2302
            goto error;
    }

2303 2304
    /* Video devices */
    for (i = 0; i < def->nvideos; i++) {
2305 2306
        if (def->videos[i]->type == VIR_DOMAIN_VIDEO_TYPE_NONE)
            continue;
2307

2308
        if (!virDeviceInfoPCIAddressIsWanted(&def->videos[i]->info))
2309
            continue;
2310

2311
        if (qemuDomainPCIAddressReserveNextAddr(addrs, &def->videos[i]->info) < 0)
2312 2313 2314 2315 2316
            goto error;
    }

    /* Shared Memory */
    for (i = 0; i < def->nshmems; i++) {
2317
        if (!virDeviceInfoPCIAddressIsWanted(&def->shmems[i]->info))
2318 2319
            continue;

2320
        if (qemuDomainPCIAddressReserveNextAddr(addrs, &def->shmems[i]->info) < 0)
2321 2322 2323
            goto error;
    }
    for (i = 0; i < def->ninputs; i++) {
2324
        if (def->inputs[i]->bus != VIR_DOMAIN_INPUT_BUS_VIRTIO ||
2325
            !virDeviceInfoPCIAddressIsWanted(&def->inputs[i]->info))
2326 2327
            continue;

2328
        if (qemuDomainPCIAddressReserveNextAddr(addrs, &def->inputs[i]->info) < 0)
2329 2330 2331 2332 2333 2334 2335 2336
            goto error;
    }
    for (i = 0; i < def->nparallels; i++) {
        /* Nada - none are PCI based (yet) */
    }
    for (i = 0; i < def->nserials; i++) {
        virDomainChrDefPtr chr = def->serials[i];

2337
        if (chr->targetType != VIR_DOMAIN_CHR_SERIAL_TARGET_TYPE_PCI ||
2338
            !virDeviceInfoPCIAddressIsWanted(&chr->info))
2339 2340
            continue;

2341
        if (qemuDomainPCIAddressReserveNextAddr(addrs, &chr->info) < 0)
2342 2343 2344 2345 2346 2347 2348 2349 2350
            goto error;
    }
    for (i = 0; i < def->nchannels; i++) {
        /* Nada - none are PCI based (yet) */
    }
    for (i = 0; i < def->nhubs; i++) {
        /* Nada - none are PCI based (yet) */
    }

J
Ján Tomko 已提交
2351
    if (def->vsock &&
2352
        virDeviceInfoPCIAddressIsWanted(&def->vsock->info)) {
J
Ján Tomko 已提交
2353 2354 2355 2356 2357 2358

        if (qemuDomainPCIAddressReserveNextAddr(addrs,
                                                &def->vsock->info) < 0)
            goto error;
    }

2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369
    return 0;

 error:
    return -1;
}


static bool
qemuDomainSupportsPCI(virDomainDefPtr def,
                      virQEMUCapsPtr qemuCaps)
{
2370 2371 2372
    if ((def->os.arch != VIR_ARCH_ARMV7L) &&
        (def->os.arch != VIR_ARCH_AARCH64) &&
        !ARCH_IS_RISCV(def->os.arch))
2373 2374 2375 2376 2377
        return true;

    if (STREQ(def->os.machine, "versatilepb"))
        return true;

2378
    if (qemuDomainIsARMVirt(def) &&
2379 2380 2381 2382 2383 2384 2385
        virQEMUCapsGet(qemuCaps, QEMU_CAPS_OBJECT_GPEX))
        return true;

    return false;
}


2386
static void
2387
qemuDomainPCIControllerSetDefaultModelName(virDomainControllerDefPtr cont,
2388
                                           virDomainDefPtr def,
2389
                                           virQEMUCapsPtr qemuCaps)
2390 2391 2392 2393 2394 2395
{
    int *modelName = &cont->opts.pciopts.modelName;

    /* make sure it's not already set */
    if (*modelName != VIR_DOMAIN_CONTROLLER_PCI_MODEL_NAME_NONE)
        return;
2396

2397 2398 2399 2400 2401 2402 2403
    switch ((virDomainControllerModelPCI)cont->model) {
    case VIR_DOMAIN_CONTROLLER_MODEL_PCI_BRIDGE:
        *modelName = VIR_DOMAIN_CONTROLLER_PCI_MODEL_NAME_PCI_BRIDGE;
        break;
    case VIR_DOMAIN_CONTROLLER_MODEL_DMI_TO_PCI_BRIDGE:
        *modelName = VIR_DOMAIN_CONTROLLER_PCI_MODEL_NAME_I82801B11_BRIDGE;
        break;
2404 2405 2406
    case VIR_DOMAIN_CONTROLLER_MODEL_PCIE_TO_PCI_BRIDGE:
        *modelName = VIR_DOMAIN_CONTROLLER_PCI_MODEL_NAME_PCIE_PCI_BRIDGE;
        break;
2407
    case VIR_DOMAIN_CONTROLLER_MODEL_PCIE_ROOT_PORT:
2408 2409 2410 2411 2412 2413
        /* Use generic PCIe Root Ports if available, falling back to
         * ioh3420 otherwise */
        if (virQEMUCapsGet(qemuCaps, QEMU_CAPS_DEVICE_PCIE_ROOT_PORT))
            *modelName = VIR_DOMAIN_CONTROLLER_PCI_MODEL_NAME_PCIE_ROOT_PORT;
        else
            *modelName = VIR_DOMAIN_CONTROLLER_PCI_MODEL_NAME_IOH3420;
2414 2415 2416 2417 2418 2419 2420
        break;
    case VIR_DOMAIN_CONTROLLER_MODEL_PCIE_SWITCH_UPSTREAM_PORT:
        *modelName = VIR_DOMAIN_CONTROLLER_PCI_MODEL_NAME_X3130_UPSTREAM;
        break;
    case VIR_DOMAIN_CONTROLLER_MODEL_PCIE_SWITCH_DOWNSTREAM_PORT:
        *modelName = VIR_DOMAIN_CONTROLLER_PCI_MODEL_NAME_XIO3130_DOWNSTREAM;
        break;
2421
    case VIR_DOMAIN_CONTROLLER_MODEL_PCI_EXPANDER_BUS:
2422 2423
        *modelName = VIR_DOMAIN_CONTROLLER_PCI_MODEL_NAME_PXB;
        break;
2424
    case VIR_DOMAIN_CONTROLLER_MODEL_PCIE_EXPANDER_BUS:
2425 2426
        *modelName = VIR_DOMAIN_CONTROLLER_PCI_MODEL_NAME_PXB_PCIE;
        break;
2427
    case VIR_DOMAIN_CONTROLLER_MODEL_PCI_ROOT:
2428 2429 2430
        if (qemuDomainIsPSeries(def))
            *modelName = VIR_DOMAIN_CONTROLLER_PCI_MODEL_NAME_SPAPR_PCI_HOST_BRIDGE;
        break;
2431
    case VIR_DOMAIN_CONTROLLER_MODEL_PCIE_ROOT:
2432
    case VIR_DOMAIN_CONTROLLER_MODEL_PCI_DEFAULT:
2433 2434 2435 2436 2437 2438
    case VIR_DOMAIN_CONTROLLER_MODEL_PCI_LAST:
        break;
    }
}


2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463
/**
 * qemuDomainAddressFindNewTargetIndex:
 * @def: domain definition
 *
 * Find a target index that can be used for a PCI controller.
 *
 * Returns: an unused target index, or -1 if all available target
 *          indexes are already taken.
 */
static int
qemuDomainAddressFindNewTargetIndex(virDomainDefPtr def)
{
    int targetIndex;
    int ret = -1;

    /* Try all indexes between 1 and 31 - QEMU only supports 32
     * PHBs, and 0 is reserved for the default, implicit one */
    for (targetIndex = 1; targetIndex <= 31; targetIndex++) {
        bool found = false;
        size_t i;

        for (i = 0; i < def->ncontrollers; i++) {
            virDomainControllerDefPtr cont = def->controllers[i];

            /* Skip everything but PHBs */
2464
            if (!virDomainControllerIsPSeriesPHB(cont))
2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486
                continue;

            /* Stop looking as soon as we find a PHB that's
             * already using this specific target index */
            if (cont->opts.pciopts.targetIndex == targetIndex) {
                found = true;
                break;
            }
        }

        /* If no existing PCI controller uses this index, great,
         * it means it's free and we can return it to the caller */
        if (!found) {
            ret = targetIndex;
            break;
        }
    }

    return ret;
}


2487 2488 2489
static int
qemuDomainAddressFindNewBusNr(virDomainDefPtr def)
{
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 2531 2532
    /* Try to find a nice default for busNr for a new pci-expander-bus.
     * This is a bit tricky, since you need to satisfy the following:
     *
     * 1) There need to be enough unused bus numbers between busNr of this
     *    bus and busNr of the next highest bus for the guest to assign a
     *    unique bus number to each PCI bus that is a child of this
     *    bus. Each PCI controller. On top of this, the pxb device (which
     *    implements the pci-expander-bus) includes a pci-bridge within
     *    it, and that bridge also uses one bus number (so each pxb device
     *    requires at least 2 bus numbers).
     *
     * 2) There need to be enough bus numbers *below* this for all the
     *    child controllers of the pci-expander-bus with the next lower
     *    busNr (or the pci-root bus if there are no lower
     *    pci-expander-buses).
     *
     * 3) If at all possible, we want to avoid needing to change the busNr
     *    of a bus in the future, as that changes the guest's device ABI,
     *    which could potentially lead to issues with a guest OS that is
     *    picky about such things.
     *
     *  Due to the impossibility of predicting what might be added to the
     *  config in the future, we can't make a foolproof choice, but since
     *  a pci-expander-bus (pxb) has slots for 32 devices, and the only
     *  practical use for it is to assign real devices on a particular
     *  NUMA node in the host, it's reasonably safe to assume it should
     *  never need any additional child buses (probably only a few of the
     *  32 will ever be used). So for pci-expander-bus we find the lowest
     *  existing busNr, and set this one to the current lowest - 2 (one
     *  for the pxb, one for the intergrated pci-bridge), thus leaving the
     *  maximum possible bus numbers available for other buses plugged
     *  into pci-root (i.e. pci-bridges and other
     *  pci-expander-buses). Anyone who needs more than 32 devices
     *  descended from one pci-expander-bus should set the busNr manually
     *  in the config.
     *
     *  There is room for more error checking here - in particular we
     *  can/should determine the ultimate parent (root-bus) of each PCI
     *  controller and determine if there is enough space for all the
     *  buses within the current range allotted to the bus just prior to
     *  this one.
     */

2533 2534 2535 2536
    size_t i;
    int lowestBusNr = 256;

    for (i = 0; i < def->ncontrollers; i++) {
2537 2538 2539 2540
        virDomainControllerDefPtr cont = def->controllers[i];

        if (cont->type == VIR_DOMAIN_CONTROLLER_TYPE_PCI) {
            int thisBusNr = cont->opts.pciopts.busNr;
2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 2552 2553 2554 2555 2556 2557

            if (thisBusNr >= 0 && thisBusNr < lowestBusNr)
                lowestBusNr = thisBusNr;
        }
    }

    /* If we already have a busNR = 1, then we can't auto-assign (0 is
     * the pci[e]-root, and the others may have been assigned
     * purposefully).
     */
    if (lowestBusNr <= 2)
        return -1;

    return lowestBusNr - 2;
}


2558 2559 2560
static int
qemuDomainAssignPCIAddresses(virDomainDefPtr def,
                             virQEMUCapsPtr qemuCaps,
2561
                             virQEMUDriverPtr driver,
2562 2563 2564 2565 2566
                             virDomainObjPtr obj)
{
    int ret = -1;
    virDomainPCIAddressSetPtr addrs = NULL;
    qemuDomainObjPrivatePtr priv = NULL;
2567 2568 2569 2570
    int max_idx = -1;
    int nbuses = 0;
    size_t i;
    int rv;
2571

2572
    for (i = 0; i < def->ncontrollers; i++) {
2573 2574 2575
        virDomainControllerDefPtr cont = def->controllers[i];

        if (cont->type == VIR_DOMAIN_CONTROLLER_TYPE_PCI) {
2576
            if ((int)cont->idx > max_idx)
2577
                max_idx = cont->idx;
2578
        }
2579
    }
2580

2581
    nbuses = max_idx + 1;
2582

2583 2584 2585 2586
    /* set the connect type flags (pci vs. pcie) in the DeviceInfo
     * of all devices. This will be used to pick an appropriate
     * bus when assigning addresses.
     */
2587
    if (qemuDomainFillAllPCIConnectFlags(def, qemuCaps, driver) < 0)
2588 2589
        goto cleanup;

2590 2591 2592
    if (qemuDomainFillAllPCIExtensionFlags(def, qemuCaps) < 0)
        goto cleanup;

2593 2594 2595
    if (qemuDomainSetupIsolationGroups(def) < 0)
        goto cleanup;

2596
    if (nbuses > 0) {
2597
        /* 1st pass to figure out how many PCI bridges we need */
2598
        if (!(addrs = qemuDomainPCIAddressSetCreate(def, qemuCaps, nbuses, true)))
2599
            goto cleanup;
2600

2601 2602 2603
        if (qemuDomainValidateDevicePCISlotsChipsets(def, qemuCaps,
                                                     addrs) < 0)
            goto cleanup;
2604

2605 2606 2607 2608 2609 2610 2611 2612
        /* For domains that have pci-root, reserve 1 extra slot for a
         * (potential) bridge (for future expansion) only if buses are
         * not fully reserved yet (if all buses are fully reserved
         * with manually/previously assigned addresses, any attempt to
         * reserve an extra slot would fail anyway. But if all buses
         * are *not* fully reserved, this extra reservation might push
         * the config to add a new pci-bridge to plug into the final
         * available slot, thus preserving the ability to expand)
2613
         *
2614 2615 2616 2617
         * We only do this for those domains that have pci-root, since
         * those with pcie-root will usually want to expand using PCIe
         * controllers, which we will do after assigning addresses for
         * all *actual* devices.
2618
         */
2619

2620
        if (qemuDomainHasPCIRoot(def)) {
2621 2622 2623 2624 2625 2626 2627
            /* This is a dummy info used to reserve a slot for a
             * legacy PCI device that doesn't exist, but may in the
             * future, e.g.  if another device is hotplugged into the
             * domain.
             */
            virDomainDeviceInfo info = {
                .pciConnectFlags = (VIR_PCI_CONNECT_HOTPLUGGABLE |
2628 2629
                                    VIR_PCI_CONNECT_TYPE_PCI_DEVICE),
                .pciAddrExtFlags = VIR_PCI_ADDRESS_EXTENSION_NONE
2630 2631 2632 2633
            };
            bool buses_reserved = true;

            for (i = 0; i < addrs->nbuses; i++) {
2634
                if (!virDomainPCIAddressBusIsFullyReserved(&addrs->buses[i])) {
2635 2636 2637 2638 2639
                    buses_reserved = false;
                    break;
                }
            }
            if (!buses_reserved &&
2640
                qemuDomainPCIAddressReserveNextAddr(addrs, &info) < 0)
2641 2642
                goto cleanup;
        }
2643

2644 2645
        if (qemuDomainAssignDevicePCISlots(def, qemuCaps, addrs) < 0)
            goto cleanup;
2646

2647 2648 2649
        if (virDomainDeviceInfoIterate(def, qemuDomainAssignPCIAddressExtension, addrs) < 0)
            goto cleanup;

2650 2651 2652 2653 2654 2655 2656 2657 2658 2659 2660 2661 2662 2663 2664 2665
        /* Only for *new* domains with pcie-root (and no other
         * manually specified PCI controllers in the definition): If,
         * after assigning addresses/reserving slots for all devices,
         * we see that any extra buses have been auto-added, we
         * understand that the application has left management of PCI
         * addresses and controllers up to libvirt. In order to allow
         * such applications to easily support hotplug, we will do a
         * "one time" reservation of one extra PCIE|HOTPLUGGABLE
         * slots, which should cause us to auto-add 1 extra
         * pcie-root-port. The single slot in this root-port will be
         * available for hotplug, or may also be used when a device is
         * added to the config offline.
         */

        if (max_idx <= 0 &&
            addrs->nbuses > max_idx + 1 &&
2666
            qemuDomainHasPCIeRoot(def)) {
2667 2668
            virDomainDeviceInfo info = {
                .pciConnectFlags = (VIR_PCI_CONNECT_HOTPLUGGABLE |
2669 2670
                                    VIR_PCI_CONNECT_TYPE_PCIE_DEVICE),
                .pciAddrExtFlags = VIR_PCI_ADDRESS_EXTENSION_NONE
2671 2672 2673 2674 2675
            };

            /* if there isn't an empty pcie-root-port, this will
             * cause one to be added
             */
2676
            if (qemuDomainPCIAddressReserveNextAddr(addrs, &info) < 0)
2677 2678 2679 2680 2681 2682
               goto cleanup;
        }

        /* now reflect any controllers auto-added to addrs into the
         * domain controllers list
         */
2683
        for (i = 1; i < addrs->nbuses; i++) {
2684 2685
            virDomainDeviceDef dev;
            int contIndex;
2686
            virDomainPCIAddressBusPtr bus = &addrs->buses[i];
2687

2688 2689 2690 2691
            if ((rv = virDomainDefMaybeAddController(
                     def, VIR_DOMAIN_CONTROLLER_TYPE_PCI,
                     i, bus->model)) < 0)
                goto cleanup;
2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704 2705 2706 2707 2708 2709 2710 2711 2712 2713 2714

            if (rv == 0)
                continue; /* no new controller added */

            /* We did add a new controller, so we will need one more
             * address (and we need to set the new controller's
             * pciConnectFlags)
             */
            contIndex = virDomainControllerFind(def,
                                                VIR_DOMAIN_CONTROLLER_TYPE_PCI,
                                                i);
            if (contIndex < 0) {
                /* this should never happen - we just added it */
                virReportError(VIR_ERR_INTERNAL_ERROR,
                               _("Could not find auto-added %s controller "
                                 "with index %zu"),
                               virDomainControllerModelPCITypeToString(bus->model),
                               i);
                goto cleanup;
            }
            dev.type = VIR_DOMAIN_DEVICE_CONTROLLER;
            dev.data.controller = def->controllers[contIndex];
            /* set connect flags so it will be properly addressed */
2715
            qemuDomainFillDevicePCIConnectFlags(def, &dev, qemuCaps, driver);
2716 2717 2718 2719 2720 2721 2722

            /* Reserve an address for the controller. pci-root and pcie-root
             * controllers don't plug into any other PCI controller, hence
             * they should skip this step */
            if (bus->model != VIR_DOMAIN_CONTROLLER_MODEL_PCI_ROOT &&
                bus->model != VIR_DOMAIN_CONTROLLER_MODEL_PCIE_ROOT &&
                qemuDomainPCIAddressReserveNextAddr(addrs,
2723
                                                    &dev.data.controller->info) < 0) {
2724
                goto cleanup;
2725
            }
2726
        }
2727

2728 2729 2730 2731
        nbuses = addrs->nbuses;
        virDomainPCIAddressSetFree(addrs);
        addrs = NULL;
    }
2732

2733
    if (!(addrs = qemuDomainPCIAddressSetCreate(def, qemuCaps, nbuses, false)))
2734
        goto cleanup;
2735

2736 2737 2738 2739
    if (qemuDomainSupportsPCI(def, qemuCaps)) {
        if (qemuDomainValidateDevicePCISlotsChipsets(def, qemuCaps,
                                                     addrs) < 0)
            goto cleanup;
2740

2741 2742
        if (qemuDomainAssignDevicePCISlots(def, qemuCaps, addrs) < 0)
            goto cleanup;
2743

2744 2745 2746
        if (virDomainDeviceInfoIterate(def, qemuDomainAssignPCIAddressExtension, addrs) < 0)
            goto cleanup;

2747 2748 2749 2750 2751
        /* set multi attribute for devices at function 0 of
         * any slot that has multiple functions in use
         */
        virDomainPCIAddressSetAllMulti(def);

2752 2753 2754 2755 2756
        for (i = 0; i < def->ncontrollers; i++) {
            virDomainControllerDefPtr cont = def->controllers[i];
            int idx = cont->idx;
            virPCIDeviceAddressPtr addr;
            virDomainPCIControllerOptsPtr options;
2757

2758 2759
            if (cont->type != VIR_DOMAIN_CONTROLLER_TYPE_PCI)
                continue;
2760

2761 2762
            addr = &cont->info.addr.pci;
            options = &cont->opts.pciopts;
2763

2764 2765 2766 2767
            /* set default model name (the actual name of the
             * device in qemu) for any controller that doesn't yet
             * have it set.
             */
2768
            qemuDomainPCIControllerSetDefaultModelName(cont, def, qemuCaps);
2769

2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793 2794 2795
            /* set defaults for any other auto-generated config
             * options for this controller that haven't been
             * specified in config.
             */
            switch ((virDomainControllerModelPCI)cont->model) {
            case VIR_DOMAIN_CONTROLLER_MODEL_PCI_BRIDGE:
                if (options->chassisNr == -1)
                    options->chassisNr = cont->idx;
                break;
            case VIR_DOMAIN_CONTROLLER_MODEL_PCIE_ROOT_PORT:
                if (options->chassis == -1)
                   options->chassis = cont->idx;
                if (options->port == -1)
                   options->port = (addr->slot << 3) + addr->function;
                break;
            case VIR_DOMAIN_CONTROLLER_MODEL_PCIE_SWITCH_DOWNSTREAM_PORT:
                if (options->chassis == -1)
                   options->chassis = cont->idx;
                if (options->port == -1)
                   options->port = addr->slot;
                break;
            case VIR_DOMAIN_CONTROLLER_MODEL_PCI_EXPANDER_BUS:
            case VIR_DOMAIN_CONTROLLER_MODEL_PCIE_EXPANDER_BUS:
                if (options->busNr == -1)
                    options->busNr = qemuDomainAddressFindNewBusNr(def);
                if (options->busNr == -1) {
2796
                    virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
2797 2798 2799 2800 2801
                                   _("No free busNr lower than current "
                                     "lowest busNr is available to "
                                     "auto-assign to bus %d. Must be "
                                     "manually assigned"),
                                   addr->bus);
2802 2803
                    goto cleanup;
                }
2804
                break;
2805 2806 2807 2808 2809 2810 2811 2812 2813 2814 2815 2816 2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827 2828
            case VIR_DOMAIN_CONTROLLER_MODEL_PCI_ROOT:
                if (!qemuDomainIsPSeries(def))
                    break;
                if (options->targetIndex == -1) {
                    if (cont->idx == 0) {
                        /* The pci-root controller with controller index 0
                         * must always be assigned target index 0, because
                         * it represents the implicit PHB which is treated
                         * differently than all other PHBs */
                        options->targetIndex = 0;
                    } else {
                        /* For all other PHBs the target index doesn't need
                         * to match the controller index or have any
                         * particular value, really */
                        options->targetIndex = qemuDomainAddressFindNewTargetIndex(def);
                    }
                }
                if (options->targetIndex == -1) {
                    virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
                                   _("No usable target index found for %d"),
                                   addr->bus);
                    goto cleanup;
                }
                break;
2829
            case VIR_DOMAIN_CONTROLLER_MODEL_DMI_TO_PCI_BRIDGE:
2830
            case VIR_DOMAIN_CONTROLLER_MODEL_PCIE_TO_PCI_BRIDGE:
2831 2832
            case VIR_DOMAIN_CONTROLLER_MODEL_PCIE_SWITCH_UPSTREAM_PORT:
            case VIR_DOMAIN_CONTROLLER_MODEL_PCIE_ROOT:
2833
            case VIR_DOMAIN_CONTROLLER_MODEL_PCI_DEFAULT:
2834 2835 2836 2837
            case VIR_DOMAIN_CONTROLLER_MODEL_PCI_LAST:
                break;
            }

2838
            /* check if every PCI bridge controller's index is larger than
2839 2840 2841 2842 2843 2844
             * the bus it is placed onto
             */
            if (cont->model == VIR_DOMAIN_CONTROLLER_MODEL_PCI_BRIDGE &&
                idx <= addr->bus) {
                virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
                               _("PCI controller at index %d (0x%02x) has "
2845 2846
                                 "bus='0x%02x', but index must be "
                                 "larger than bus"),
2847 2848
                               idx, idx, addr->bus);
                goto cleanup;
2849 2850 2851 2852 2853 2854
            }
        }
    }

    if (obj && obj->privateData) {
        priv = obj->privateData;
J
John Ferlan 已提交
2855 2856 2857
        /* if this is the live domain object, we persist the PCI addresses */
        priv->pciaddrs = addrs;
        addrs = NULL;
2858 2859 2860 2861 2862 2863 2864 2865 2866 2867 2868
    }

    ret = 0;

 cleanup:
    virDomainPCIAddressSetFree(addrs);

    return ret;
}


J
Ján Tomko 已提交
2869 2870 2871 2872 2873 2874 2875 2876 2877 2878 2879 2880
struct qemuAssignUSBIteratorInfo {
    virDomainUSBAddressSetPtr addrs;
    size_t count;
};


static int
qemuDomainAssignUSBPortsIterator(virDomainDeviceInfoPtr info,
                                 void *opaque)
{
    struct qemuAssignUSBIteratorInfo *data = opaque;

2881 2882 2883 2884 2885 2886
    if (info->type != VIR_DOMAIN_DEVICE_ADDRESS_TYPE_NONE &&
        info->type != VIR_DOMAIN_DEVICE_ADDRESS_TYPE_USB)
        return 0;

    if (info->type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_USB &&
        virDomainUSBAddressPortIsValid(info->addr.usb.port))
J
Ján Tomko 已提交
2887 2888 2889 2890 2891 2892 2893 2894 2895 2896 2897 2898 2899 2900 2901 2902 2903 2904 2905 2906 2907 2908 2909 2910 2911 2912 2913 2914 2915 2916 2917 2918 2919 2920 2921 2922 2923 2924 2925 2926 2927 2928 2929 2930
        return 0;

    return virDomainUSBAddressAssign(data->addrs, info);
}


static int
qemuDomainAssignUSBHubs(virDomainUSBAddressSetPtr addrs,
                        virDomainDefPtr def)
{
    size_t i;

    for (i = 0; i < def->nhubs; i++) {
        virDomainHubDefPtr hub = def->hubs[i];
        if (hub->type != VIR_DOMAIN_HUB_TYPE_USB)
            continue;

        if (hub->info.type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_USB &&
            virDomainUSBAddressPortIsValid(hub->info.addr.usb.port))
            continue;
        if (virDomainUSBAddressAssign(addrs, &hub->info) < 0)
            return -1;

        if (virDomainUSBAddressSetAddHub(addrs, hub) < 0)
            return -1;
    }

    return 0;
}


static int
qemuDomainAssignUSBPorts(virDomainUSBAddressSetPtr addrs,
                         virDomainDefPtr def)
{
    struct qemuAssignUSBIteratorInfo data = { .addrs = addrs };

    return virDomainUSBDeviceDefForeach(def,
                                        qemuDomainAssignUSBPortsIterator,
                                        &data,
                                        true);
}


2931 2932 2933 2934 2935 2936 2937 2938 2939 2940 2941 2942 2943 2944 2945 2946 2947
static int
qemuDomainAssignUSBPortsCounter(virDomainDeviceInfoPtr info ATTRIBUTE_UNUSED,
                                void *opaque)
{
    struct qemuAssignUSBIteratorInfo *data = opaque;

    data->count++;
    return 0;
}


static int
qemuDomainUSBAddressAddHubs(virDomainDefPtr def)
{
    struct qemuAssignUSBIteratorInfo data = { .count = 0 };
    virDomainHubDefPtr hub = NULL;
    size_t available_ports;
2948
    size_t hubs_needed = 0;
2949
    int ret = -1;
2950
    size_t i;
2951 2952 2953 2954 2955 2956 2957

    available_ports = virDomainUSBAddressCountAllPorts(def);
    ignore_value(virDomainUSBDeviceDefForeach(def,
                                              qemuDomainAssignUSBPortsCounter,
                                              &data,
                                              false));

2958
    if (data.count > 0 && !virDomainDefHasUSB(def)) {
2959 2960 2961 2962 2963 2964
        virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                       _("USB is disabled for this domain, but USB devices "
                         "are present in the domain XML"));
        return -1;
    }

2965 2966 2967 2968 2969 2970 2971 2972
    if (data.count > available_ports)
        hubs_needed = VIR_DIV_UP(data.count - available_ports + 1,
                                 VIR_DOMAIN_USB_HUB_PORTS - 1);

    VIR_DEBUG("Found %zu USB devices and %zu provided USB ports; adding %zu hubs",
              data.count, available_ports, hubs_needed);

    for (i = 0; i < hubs_needed; i++) {
2973 2974 2975 2976 2977 2978 2979 2980 2981 2982 2983 2984 2985 2986 2987
        if (VIR_ALLOC(hub) < 0)
            return -1;
        hub->type = VIR_DOMAIN_HUB_TYPE_USB;

        if (VIR_APPEND_ELEMENT(def->hubs, def->nhubs, hub) < 0)
            goto cleanup;
    }

    ret = 0;
 cleanup:
    VIR_FREE(hub);
    return ret;
}


2988 2989 2990 2991 2992 2993 2994 2995 2996 2997 2998 2999 3000 3001 3002 3003 3004 3005 3006 3007 3008 3009 3010 3011 3012 3013 3014 3015 3016 3017 3018 3019
static virBitmapPtr
qemuDomainGetMemorySlotMap(const virDomainDef *def)
{
    virBitmapPtr ret;
    virDomainMemoryDefPtr mem;
    size_t i;

    if (!(ret = virBitmapNew(def->mem.memory_slots)))
        return NULL;

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

        if (mem->info.type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_DIMM)
            ignore_value(virBitmapSetBit(ret, mem->info.addr.dimm.slot));
    }

    return ret;
}


static int
qemuAssignMemoryDeviceSlot(virDomainMemoryDefPtr mem,
                           virBitmapPtr slotmap)
{
    ssize_t nextslot = -1;

    if (mem->info.type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_DIMM)
        return 0;

    if ((nextslot = virBitmapNextClearBit(slotmap, -1)) < 0) {
        virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
Y
Yuri Chornoivan 已提交
3020
                       _("failed to find an empty memory slot"));
3021 3022 3023 3024 3025 3026 3027 3028 3029 3030 3031 3032 3033 3034 3035 3036 3037 3038 3039 3040 3041 3042 3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054 3055 3056 3057 3058 3059 3060 3061 3062 3063 3064 3065 3066 3067 3068 3069 3070 3071 3072 3073 3074 3075
        return -1;
    }

    ignore_value(virBitmapSetBit(slotmap, nextslot));
    mem->info.type = VIR_DOMAIN_DEVICE_ADDRESS_TYPE_DIMM;
    mem->info.addr.dimm.slot = nextslot;

    return 0;
}


int
qemuDomainAssignMemoryDeviceSlot(virDomainDefPtr def,
                                 virDomainMemoryDefPtr mem)
{
    virBitmapPtr slotmap = NULL;
    int ret;

    if (!(slotmap = qemuDomainGetMemorySlotMap(def)))
        return -1;

    ret = qemuAssignMemoryDeviceSlot(mem, slotmap);

    virBitmapFree(slotmap);
    return ret;
}


static int
qemuDomainAssignMemorySlots(virDomainDefPtr def)
{
    virBitmapPtr slotmap = NULL;
    int ret = -1;
    size_t i;

    if (!virDomainDefHasMemoryHotplug(def))
        return 0;

    if (!(slotmap = qemuDomainGetMemorySlotMap(def)))
        return -1;

    for (i = 0; i < def->nmems; i++) {
        if (qemuAssignMemoryDeviceSlot(def->mems[i], slotmap) < 0)
            goto cleanup;
    }

    ret = 0;

 cleanup:
    virBitmapFree(slotmap);
    return ret;

}


J
Ján Tomko 已提交
3076 3077
static int
qemuDomainAssignUSBAddresses(virDomainDefPtr def,
3078 3079
                             virDomainObjPtr obj,
                             bool newDomain)
J
Ján Tomko 已提交
3080 3081 3082 3083 3084
{
    int ret = -1;
    virDomainUSBAddressSetPtr addrs = NULL;
    qemuDomainObjPrivatePtr priv = NULL;

3085 3086 3087 3088 3089 3090 3091 3092 3093 3094 3095
    if (!newDomain) {
        /* only create the address cache for:
         *  new domains
         *  domains that already have all the addresses specified
         * otherwise libvirt's attempt to recreate the USB topology via
         * QEMU command line might fail */
        if (virDomainUSBDeviceDefForeach(def, virDomainUSBAddressPresent, NULL,
                                         false) < 0)
            return 0;
    }

J
Ján Tomko 已提交
3096 3097 3098
    if (!(addrs = virDomainUSBAddressSetCreate()))
        goto cleanup;

3099 3100 3101
    if (qemuDomainUSBAddressAddHubs(def) < 0)
        goto cleanup;

J
Ján Tomko 已提交
3102 3103 3104 3105 3106 3107 3108 3109 3110
    if (virDomainUSBAddressSetAddControllers(addrs, def) < 0)
        goto cleanup;

    if (virDomainUSBDeviceDefForeach(def, virDomainUSBAddressReserve, addrs,
                                     true) < 0)
        goto cleanup;

    VIR_DEBUG("Existing USB addresses have been reserved");

J
Ján Tomko 已提交
3111 3112 3113 3114 3115 3116 3117 3118
    if (qemuDomainAssignUSBHubs(addrs, def) < 0)
        goto cleanup;

    if (qemuDomainAssignUSBPorts(addrs, def) < 0)
        goto cleanup;

    VIR_DEBUG("Finished assigning USB ports");

J
Ján Tomko 已提交
3119 3120 3121 3122 3123 3124 3125 3126 3127 3128 3129 3130 3131
    if (obj && obj->privateData) {
        priv = obj->privateData;
        priv->usbaddrs = addrs;
        addrs = NULL;
    }
    ret = 0;

 cleanup:
    virDomainUSBAddressSetFree(addrs);
    return ret;
}


3132 3133 3134
int
qemuDomainAssignAddresses(virDomainDefPtr def,
                          virQEMUCapsPtr qemuCaps,
3135
                          virQEMUDriverPtr driver,
3136
                          virDomainObjPtr obj,
J
Ján Tomko 已提交
3137
                          bool newDomain)
3138
{
3139
    if (qemuDomainAssignVirtioSerialAddresses(def) < 0)
3140
        return -1;
3141

3142
    if (qemuDomainAssignSpaprVIOAddresses(def) < 0)
3143
        return -1;
3144

T
Tomasz Flendrich 已提交
3145
    if (qemuDomainAssignS390Addresses(def, qemuCaps) < 0)
3146
        return -1;
3147

3148
    qemuDomainAssignVirtioMMIOAddresses(def, qemuCaps);
3149

3150
    if (qemuDomainAssignPCIAddresses(def, qemuCaps, driver, obj) < 0)
3151 3152
        return -1;

3153
    if (qemuDomainAssignUSBAddresses(def, obj, newDomain) < 0)
J
Ján Tomko 已提交
3154 3155
        return -1;

3156 3157 3158
    if (qemuDomainAssignMemorySlots(def) < 0)
        return -1;

3159
    return 0;
3160 3161
}

3162 3163 3164 3165 3166 3167 3168 3169 3170 3171 3172 3173 3174 3175 3176 3177 3178
/**
 * qemuDomainEnsurePCIAddress:
 *
 * @obj: the virDomainObjPtr for the domain. This will include
 *       qemuCaps and address cache (if there is one)
 *
 * @dev: the device that we need to ensure has a PCI address
 *
 * if @dev should have a PCI address but doesn't, assign an address on
 * a compatible PCI bus, and set it in @dev->...info. If there is an
 * address already, validate that it is on a compatible bus, based on
 * @dev->...info.pciConnectFlags.
 *
 * returns 0 on success -1 on failure.
 */
int
qemuDomainEnsurePCIAddress(virDomainObjPtr obj,
3179 3180
                           virDomainDeviceDefPtr dev,
                           virQEMUDriverPtr driver)
3181 3182 3183 3184 3185 3186 3187
{
    qemuDomainObjPrivatePtr priv = obj->privateData;
    virDomainDeviceInfoPtr info = virDomainDeviceGetInfo(dev);

    if (!info)
        return 0;

3188
    qemuDomainFillDevicePCIConnectFlags(obj->def, dev, priv->qemuCaps, driver);
3189

3190 3191
    qemuDomainFillDevicePCIExtensionFlags(dev, info, priv->qemuCaps);

3192 3193 3194
    return virDomainPCIAddressEnsureAddr(priv->pciaddrs, info,
                                         info->pciConnectFlags);
}
3195 3196 3197 3198 3199 3200 3201 3202 3203 3204 3205

void
qemuDomainReleaseDeviceAddress(virDomainObjPtr vm,
                               virDomainDeviceInfoPtr info,
                               const char *devstr)
{
    qemuDomainObjPrivatePtr priv = vm->privateData;

    if (!devstr)
        devstr = info->alias;

3206
    if (virDeviceInfoPCIAddressIsPresent(info)) {
3207
        virDomainPCIAddressReleaseAddr(priv->pciaddrs, &info->addr.pci);
3208 3209
        virDomainPCIAddressExtensionReleaseAddr(priv->pciaddrs, &info->addr.pci);
    }
3210

3211 3212
    if (virDomainUSBAddressRelease(priv->usbaddrs, info) < 0)
        VIR_WARN("Unable to release USB address on %s", NULLSTR(devstr));
3213
}
3214 3215 3216 3217 3218 3219 3220 3221 3222 3223 3224 3225 3226 3227 3228 3229


int
qemuDomainEnsureVirtioAddress(bool *releaseAddr,
                              virDomainObjPtr vm,
                              virDomainDeviceDefPtr dev,
                              const char *devicename)
{
    virDomainDeviceInfoPtr info = virDomainDeviceGetInfo(dev);
    qemuDomainObjPrivatePtr priv = vm->privateData;
    virDomainCCWAddressSetPtr ccwaddrs = NULL;
    virQEMUDriverPtr driver = priv->driver;
    int ret = -1;

    if (!info->type) {
        if (qemuDomainIsS390CCW(vm->def) &&
3230
            virQEMUCapsGet(priv->qemuCaps, QEMU_CAPS_CCW))
3231 3232 3233 3234 3235 3236 3237 3238 3239 3240
            info->type = VIR_DOMAIN_DEVICE_ADDRESS_TYPE_CCW;
        else if (virQEMUCapsGet(priv->qemuCaps, QEMU_CAPS_VIRTIO_S390))
            info->type = VIR_DOMAIN_DEVICE_ADDRESS_TYPE_VIRTIO_S390;
    } else {
        if (!qemuDomainCheckCCWS390AddressSupport(vm->def, *info, priv->qemuCaps,
                                                  devicename))
            return -1;
    }

    if (info->type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_CCW) {
3241
        if (!(ccwaddrs = virDomainCCWAddressSetCreateFromDomain(vm->def)))
3242 3243 3244 3245 3246 3247 3248 3249 3250 3251 3252 3253 3254 3255 3256 3257 3258
            goto cleanup;
        if (virDomainCCWAddressAssign(info, ccwaddrs,
                                      !info->addr.ccw.assigned) < 0)
            goto cleanup;
    } else if (!info->type ||
               info->type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI) {
        if (qemuDomainEnsurePCIAddress(vm, dev, driver) < 0)
            goto cleanup;
        *releaseAddr = true;
    }

    ret = 0;

 cleanup:
    virDomainCCWAddressSetFree(ccwaddrs);
    return ret;
}