qemu_domain_address.c 89.3 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 42
/*
 * 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


int
43
qemuDomainSetSCSIControllerModel(const virDomainDef *def,
44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67
                                 virQEMUCapsPtr qemuCaps,
                                 int *model)
{
    if (*model > 0) {
        switch (*model) {
        case VIR_DOMAIN_CONTROLLER_MODEL_SCSI_LSILOGIC:
            if (!virQEMUCapsGet(qemuCaps, QEMU_CAPS_SCSI_LSI)) {
                virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                               _("This QEMU doesn't support "
                                 "the LSI 53C895A SCSI controller"));
                return -1;
            }
            break;
        case VIR_DOMAIN_CONTROLLER_MODEL_SCSI_VIRTIO_SCSI:
            if (!virQEMUCapsGet(qemuCaps, QEMU_CAPS_VIRTIO_SCSI)) {
                virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                               _("This QEMU doesn't support "
                                 "virtio scsi controller"));
                return -1;
            }
            break;
        case VIR_DOMAIN_CONTROLLER_MODEL_SCSI_IBMVSCSI:
            /*TODO: need checking work here if necessary */
            break;
68 69 70 71 72 73 74 75
        case VIR_DOMAIN_CONTROLLER_MODEL_SCSI_LSISAS1068:
            if (!virQEMUCapsGet(qemuCaps, QEMU_CAPS_SCSI_MPTSAS1068)) {
                virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                               _("This QEMU doesn't support "
                                 "the LSI SAS1068 (MPT Fusion) controller"));
                return -1;
            }
            break;
76 77 78 79
        case VIR_DOMAIN_CONTROLLER_MODEL_SCSI_LSISAS1078:
            if (!virQEMUCapsGet(qemuCaps, QEMU_CAPS_SCSI_MEGASAS)) {
                virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                               _("This QEMU doesn't support "
80
                                 "the LSI SAS1078 (MegaRAID) controller"));
81 82 83 84 85 86 87 88 89 90
                return -1;
            }
            break;
        default:
            virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
                           _("Unsupported controller model: %s"),
                           virDomainControllerModelSCSITypeToString(*model));
            return -1;
        }
    } else {
91
        if (qemuDomainIsPSeries(def)) {
92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108
            *model = VIR_DOMAIN_CONTROLLER_MODEL_SCSI_IBMVSCSI;
        } else if (virQEMUCapsGet(qemuCaps, QEMU_CAPS_SCSI_LSI)) {
            *model = VIR_DOMAIN_CONTROLLER_MODEL_SCSI_LSILOGIC;
        } else if (virQEMUCapsGet(qemuCaps, QEMU_CAPS_VIRTIO_SCSI)) {
            *model = VIR_DOMAIN_CONTROLLER_MODEL_SCSI_VIRTIO_SCSI;
        } else {
            virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                           _("Unable to determine model for scsi controller"));
            return -1;
        }
    }

    return 0;
}


static int
109
qemuDomainAssignVirtioSerialAddresses(virDomainDefPtr def)
110 111 112 113 114
{
    int ret = -1;
    size_t i;
    virDomainVirtioSerialAddrSetPtr addrs = NULL;

115
    if (!(addrs = virDomainVirtioSerialAddrSetCreateFromDomain(def)))
116 117 118 119 120 121 122 123 124
        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) &&
125 126
            virDomainVirtioSerialAddrAutoAssignFromCache(def, addrs,
                                                         &chr->info, true) < 0)
127 128 129 130 131 132 133 134
            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) &&
135 136
            virDomainVirtioSerialAddrAutoAssignFromCache(def, addrs,
                                                         &chr->info, false) < 0)
137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215
            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
qemuDomainAssignSpaprVIOAddresses(virDomainDefPtr def,
                                  virQEMUCapsPtr qemuCaps)
{
    size_t i;
    int ret = -1;
    int model;

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

    for (i = 0; i < def->nnets; i++) {
216 217 218 219 220 221 222 223
        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)
224 225 226 227
            goto cleanup;
    }

    for (i = 0; i < def->ncontrollers; i++) {
228 229 230 231
        virDomainControllerDefPtr cont = def->controllers[i];

        model = cont->model;
        if (cont->type == VIR_DOMAIN_CONTROLLER_TYPE_SCSI) {
232 233 234 235 236
            if (qemuDomainSetSCSIControllerModel(def, qemuCaps, &model) < 0)
                goto cleanup;
        }

        if (model == VIR_DOMAIN_CONTROLLER_MODEL_SCSI_IBMVSCSI &&
237 238 239 240 241
            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) {
242
            goto cleanup;
243
        }
244 245 246 247
    }

    for (i = 0; i < def->nserials; i++) {
        if (def->serials[i]->deviceType == VIR_DOMAIN_CHR_DEVICE_TYPE_SERIAL &&
248
            qemuDomainIsPSeries(def))
249 250 251 252 253 254 255
            def->serials[i]->info.type = VIR_DOMAIN_DEVICE_ADDRESS_TYPE_SPAPRVIO;
        if (qemuDomainAssignSpaprVIOAddress(def, &def->serials[i]->info,
                                            VIO_ADDR_SERIAL) < 0)
            goto cleanup;
    }

    if (def->nvram) {
256
        if (qemuDomainIsPSeries(def))
257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290
            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;
}


static void
qemuDomainPrimeVirtioDeviceAddresses(virDomainDefPtr def,
                                     virDomainDeviceAddressType type)
{
    /*
       declare address-less virtio devices to be of address type 'type'
       disks, networks, consoles, controllers, memballoon and rng in this
       order
       if type is ccw filesystem devices are declared to be of address type ccw
    */
    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++) {
291 292 293 294 295
        virDomainNetDefPtr net = def->nets[i];

        if (STREQ(net->model, "virtio") &&
            net->info.type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_NONE) {
            net->info.type = type;
296 297 298 299 300 301 302 303 304 305
        }
    }

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

    for (i = 0; i < def->ncontrollers; i++) {
306 307 308 309 310 311 312
        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;
        }
313 314
    }

315 316 317 318 319 320 321 322
    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;
    }

323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341
    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;
        }
    }
}

342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363
virDomainCCWAddressSetPtr
qemuDomainCCWAddrSetCreateFromDomain(virDomainDefPtr def)
{
    virDomainCCWAddressSetPtr addrs = NULL;

    if (!(addrs = virDomainCCWAddressSetCreate()))
        goto error;

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

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

    return addrs;

 error:
    virDomainCCWAddressSetFree(addrs);
    return NULL;
}
364 365 366 367 368 369 370 371 372

/*
 * 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
373
                              virQEMUCapsPtr qemuCaps)
374 375 376 377
{
    int ret = -1;
    virDomainCCWAddressSetPtr addrs = NULL;

378
    if (qemuDomainIsS390CCW(def) &&
379 380 381 382
        virQEMUCapsGet(qemuCaps, QEMU_CAPS_VIRTIO_CCW)) {
        qemuDomainPrimeVirtioDeviceAddresses(
            def, VIR_DOMAIN_DEVICE_ADDRESS_TYPE_CCW);

383
        if (!(addrs = qemuDomainCCWAddrSetCreateFromDomain(def)))
384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400
            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;
}


401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438
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 */
        *((bool *) opaque) = true;
        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;
}


439
static void
440 441 442
qemuDomainAssignARMVirtioMMIOAddresses(virDomainDefPtr def,
                                       virQEMUCapsPtr qemuCaps)
{
443 444 445 446 447
    if (def->os.arch != VIR_ARCH_ARMV7L &&
        def->os.arch != VIR_ARCH_AARCH64)
        return;

    if (!(STRPREFIX(def->os.machine, "vexpress-") ||
448
          qemuDomainIsVirt(def)))
449 450
        return;

451 452 453
    /* 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 */
454
    if (qemuDomainHasPCIeRoot(def) &&
455 456 457 458 459 460
        !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);
461 462 463 464
    }
}


465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483
/**
 * 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,
484
                                         virQEMUDriverPtr driver,
485
                                         virDomainPCIConnectFlags pcieFlags,
486
                                         virDomainPCIConnectFlags virtioFlags)
487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504
{
    virDomainPCIConnectFlags pciFlags = (VIR_PCI_CONNECT_TYPE_PCI_DEVICE |
                                         VIR_PCI_CONNECT_HOTPLUGGABLE);

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

        switch ((virDomainControllerType) cont->type) {
        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) {
            case VIR_DOMAIN_CONTROLLER_MODEL_USB_NEC_XHCI:
505
            case VIR_DOMAIN_CONTROLLER_MODEL_USB_QEMU_XHCI:
506
                return pcieFlags;
507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522

            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:
523
                return 0;
524 525 526
            }

        case VIR_DOMAIN_CONTROLLER_TYPE_IDE:
527 528
            return pciFlags;

529
        case VIR_DOMAIN_CONTROLLER_TYPE_SCSI:
530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546
            switch ((virDomainControllerModelSCSI) cont->model) {
            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;
            }

547
        case VIR_DOMAIN_CONTROLLER_TYPE_VIRTIO_SERIAL:
548
            return virtioFlags;
549 550 551 552 553 554 555 556

        case VIR_DOMAIN_CONTROLLER_TYPE_FDC:
        case VIR_DOMAIN_CONTROLLER_TYPE_CCID:
        case VIR_DOMAIN_CONTROLLER_TYPE_LAST:
            /* should be 0 */
            return pciFlags;
        }
    }
557
        break;
558 559 560

    case VIR_DOMAIN_DEVICE_FS:
        /* the only type of filesystem so far is virtio-9p-pci */
561
        return virtioFlags;
562 563 564 565 566 567 568 569 570 571

    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")) {
572
            return 0;
573
        }
574 575 576 577

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

578 579 580
        if (STREQ(net->model, "e1000e"))
            return pcieFlags;

581 582 583 584 585 586 587 588 589 590 591 592 593 594 595
        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:
596
            return 0;
597 598 599 600 601
        }

    case VIR_DOMAIN_DEVICE_DISK:
        switch ((virDomainDiskBus) dev->data.disk->bus) {
        case VIR_DOMAIN_DISK_BUS_VIRTIO:
602
            return virtioFlags; /* only virtio disks use PCI */
603 604 605 606 607 608 609 610 611 612

        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:
613
            return 0;
614 615
        }

616 617 618 619 620 621 622
    case VIR_DOMAIN_DEVICE_HOSTDEV: {
        virDomainHostdevDefPtr hostdev = dev->data.hostdev;
        bool isExpress = false;
        virPCIDevicePtr pciDev;
        virPCIDeviceAddressPtr hostAddr = &hostdev->source.subsys.u.pci.addr;

        if (hostdev->mode != VIR_DOMAIN_HOSTDEV_MODE_SUBSYS ||
623 624
            (hostdev->source.subsys.type != VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_PCI &&
             hostdev->source.subsys.type != VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_MDEV)) {
625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647
            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;
        }

        if (virDeviceInfoPCIAddressPresent(hostdev->info)) {
            /* 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;
        }

648 649 650 651 652
        /* 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 */
653 654 655
        if (hostdev->source.subsys.type == VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_MDEV)
            return pcieFlags;

656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705
        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;

706
        return pciFlags;
707
    }
708 709 710 711

    case VIR_DOMAIN_DEVICE_MEMBALLOON:
        switch ((virDomainMemballoonModel) dev->data.memballoon->model) {
        case VIR_DOMAIN_MEMBALLOON_MODEL_VIRTIO:
712
            return virtioFlags;
713 714 715 716

        case VIR_DOMAIN_MEMBALLOON_MODEL_XEN:
        case VIR_DOMAIN_MEMBALLOON_MODEL_NONE:
        case VIR_DOMAIN_MEMBALLOON_MODEL_LAST:
717
            return 0;
718 719 720 721 722
        }

    case VIR_DOMAIN_DEVICE_RNG:
        switch ((virDomainRNGModel) dev->data.rng->model) {
        case VIR_DOMAIN_RNG_MODEL_VIRTIO:
723
            return virtioFlags;
724 725

        case VIR_DOMAIN_RNG_MODEL_LAST:
726
            return 0;
727 728 729 730 731 732 733 734 735 736 737
        }

    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:
738
            return 0;
739 740 741 742 743
        }

    case VIR_DOMAIN_DEVICE_VIDEO:
        switch ((virDomainVideoType) dev->data.video->type) {
        case VIR_DOMAIN_VIDEO_TYPE_VIRTIO:
744 745
            return virtioFlags;

746 747 748 749 750 751 752 753 754
        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;

F
Fabian Freyer 已提交
755
        case VIR_DOMAIN_VIDEO_TYPE_GOP:
756
        case VIR_DOMAIN_VIDEO_TYPE_LAST:
757
            return 0;
758 759 760 761 762 763 764 765
        }

    case VIR_DOMAIN_DEVICE_SHMEM:
        return pciFlags;

    case VIR_DOMAIN_DEVICE_INPUT:
        switch ((virDomainInputBus) dev->data.input->bus) {
        case VIR_DOMAIN_INPUT_BUS_VIRTIO:
766
            return virtioFlags;
767 768 769 770 771 772

        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:
773
            return 0;
774 775 776 777 778 779 780 781 782 783
        }

    case VIR_DOMAIN_DEVICE_CHR:
        switch ((virDomainChrSerialTargetType) dev->data.chr->targetType) {
        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:
        case VIR_DOMAIN_CHR_SERIAL_TARGET_TYPE_LAST:
784
            return 0;
785 786
        }

787
        /* These devices don't ever connect with PCI */
788 789 790 791 792 793 794
    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:
795
        /* These devices don't even have a DeviceInfo */
796 797 798 799 800
    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:
801
        return 0;
802 803 804 805 806 807 808 809 810 811 812 813 814 815
    }

    /* 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;
816
    virQEMUDriverPtr driver;
817 818 819 820 821 822 823 824 825 826 827 828
} qemuDomainFillDevicePCIConnectFlagsIterData;


/**
 * qemuDomainFillDevicePCIConnectFlagsIterInit:
 *
 * Initialize the iterator data that is used when calling
 * qemuDomainCalculateDevicePCIConnectFlags().
 */
static void
qemuDomainFillDevicePCIConnectFlagsIterInit(virDomainDefPtr def,
                                            virQEMUCapsPtr qemuCaps,
829
                                            virQEMUDriverPtr driver,
830 831
                                            qemuDomainFillDevicePCIConnectFlagsIterData *data)
{
832 833 834

    data->driver = driver;

835
    if (qemuDomainHasPCIeRoot(def)) {
836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873
        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
874 875
        = qemuDomainDeviceCalculatePCIConnectFlags(dev, data->driver,
                                                   data->pcieFlags,
876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892
                                                   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())
 */
893
static int
894
qemuDomainFillAllPCIConnectFlags(virDomainDefPtr def,
895 896
                                 virQEMUCapsPtr qemuCaps,
                                 virQEMUDriverPtr driver)
897 898 899
{
    qemuDomainFillDevicePCIConnectFlagsIterData data;

900
    qemuDomainFillDevicePCIConnectFlagsIterInit(def, qemuCaps, driver, &data);
901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918

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


/**
 * 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.
 */
919
static void
920 921
qemuDomainFillDevicePCIConnectFlags(virDomainDefPtr def,
                                    virDomainDeviceDefPtr dev,
922 923
                                    virQEMUCapsPtr qemuCaps,
                                    virQEMUDriverPtr driver)
924 925 926 927 928 929 930 931 932 933 934 935 936 937 938
{
    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;

939
        qemuDomainFillDevicePCIConnectFlagsIterInit(def, qemuCaps, driver, &data);
940 941

        info->pciConnectFlags
942 943
            = qemuDomainDeviceCalculatePCIConnectFlags(dev, data.driver,
                                                       data.pcieFlags,
944 945 946 947 948
                                                       data.virtioFlags);
    }
}


949
static int
950
qemuDomainPCIAddressReserveNextAddr(virDomainPCIAddressSetPtr addrs,
951
                                    virDomainDeviceInfoPtr dev)
952
{
953 954
    return virDomainPCIAddressReserveNextAddr(addrs, dev,
                                              dev->pciConnectFlags, -1);
955 956 957
}


958 959 960 961 962 963 964 965
static int
qemuDomainCollectPCIAddress(virDomainDefPtr def ATTRIBUTE_UNUSED,
                            virDomainDeviceDefPtr device,
                            virDomainDeviceInfoPtr info,
                            void *opaque)
{
    virDomainPCIAddressSetPtr addrs = opaque;
    int ret = -1;
966
    virPCIDeviceAddressPtr addr = &info->addr.pci;
967

968 969 970
    if (!virDeviceInfoPCIAddressPresent(info) ||
        ((device->type == VIR_DOMAIN_DEVICE_HOSTDEV) &&
         (device->data.hostdev->parent.type != VIR_DOMAIN_DEVICE_NONE))) {
971 972 973 974 975 976 977
        /* 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;
    }

978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999
    /* 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) {
        char *addrStr = virDomainPCIAddressAsString(&info->addr.pci);

        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;
    }

1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027
    /* 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 ||
              cont->model == -1) && addr->function == 2)) {
            /* 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 &&
1028
                !(addrs->buses[0].flags & VIR_PCI_CONNECT_TYPE_PCI_DEVICE)) {
1029 1030 1031 1032 1033 1034 1035 1036 1037 1038
                virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                               _("Bus 0 must be PCI for integrated PIIX3 "
                                 "USB or IDE controllers"));
                return -1;
            } else {
                return 0;
            }
        }
    }

1039
    if (virDomainPCIAddressReserveAddr(addrs, addr,
1040
                                       info->pciConnectFlags) < 0) {
1041
        goto cleanup;
1042
    }
1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055

    ret = 0;
 cleanup:
    return ret;
}

static virDomainPCIAddressSetPtr
qemuDomainPCIAddressSetCreate(virDomainDefPtr def,
                              unsigned int nbuses,
                              bool dryRun)
{
    virDomainPCIAddressSetPtr addrs;
    size_t i;
1056 1057
    bool hasPCIeRoot = false;
    virDomainControllerModelPCI defaultModel;
1058 1059 1060 1061 1062 1063 1064

    if ((addrs = virDomainPCIAddressSetAlloc(nbuses)) == NULL)
        return NULL;

    addrs->dryRun = dryRun;

    for (i = 0; i < def->ncontrollers; i++) {
1065 1066
        virDomainControllerDefPtr cont = def->controllers[i];
        size_t idx = cont->idx;
1067

1068
        if (cont->type != VIR_DOMAIN_CONTROLLER_TYPE_PCI)
1069 1070 1071 1072 1073
            continue;

        if (idx >= addrs->nbuses) {
            virReportError(VIR_ERR_INTERNAL_ERROR,
                           _("Inappropriate new pci controller index %zu "
1074
                             "exceeds addrs array length"), idx);
1075 1076 1077
            goto error;
        }

1078
        if (virDomainPCIAddressBusSetModel(&addrs->buses[idx], cont->model) < 0)
1079
            goto error;
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 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115

        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.
     */
    if (hasPCIeRoot)
        defaultModel = VIR_DOMAIN_CONTROLLER_MODEL_PCIE_ROOT_PORT;
    else
        defaultModel = VIR_DOMAIN_CONTROLLER_MODEL_PCI_BRIDGE;

    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);
    }
1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134

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

    return addrs;

 error:
    virDomainPCIAddressSetFree(addrs);
    return NULL;
}


static int
qemuDomainValidateDevicePCISlotsPIIX3(virDomainDefPtr def,
                                      virQEMUCapsPtr qemuCaps,
                                      virDomainPCIAddressSetPtr addrs)
{
    int ret = -1;
    size_t i;
1135
    virPCIDeviceAddress tmp_addr;
1136 1137
    bool qemuDeviceVideoUsable = virQEMUCapsGet(qemuCaps, QEMU_CAPS_DEVICE_VIDEO_PRIMARY);
    char *addrStr = NULL;
1138 1139
    virDomainPCIConnectFlags flags = (VIR_PCI_CONNECT_HOTPLUGGABLE
                                      | VIR_PCI_CONNECT_TYPE_PCI_DEVICE);
1140 1141 1142

    /* Verify that first IDE and USB controllers (if any) is on the PIIX3, fn 1 */
    for (i = 0; i < def->ncontrollers; i++) {
1143 1144
        virDomainControllerDefPtr cont = def->controllers[i];

1145
        /* First IDE controller lives on the PIIX3 at slot=1, function=1 */
1146 1147 1148 1149 1150 1151 1152
        if (cont->type == VIR_DOMAIN_CONTROLLER_TYPE_IDE &&
            cont->idx == 0) {
            if (virDeviceInfoPCIAddressPresent(&cont->info)) {
                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) {
1153 1154 1155 1156 1157
                    virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                                   _("Primary IDE controller must have PCI address 0:0:1.1"));
                    goto cleanup;
                }
            } else {
1158 1159 1160 1161 1162
                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;
1163
            }
1164 1165 1166 1167 1168 1169 1170 1171 1172
        } else if (cont->type == VIR_DOMAIN_CONTROLLER_TYPE_USB &&
                   cont->idx == 0 &&
                   (cont->model == VIR_DOMAIN_CONTROLLER_MODEL_USB_PIIX3_UHCI ||
                    cont->model == -1)) {
            if (virDeviceInfoPCIAddressPresent(&cont->info)) {
                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) {
1173 1174 1175 1176 1177
                    virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                                   _("PIIX3 USB controller must have PCI address 0:0:1.2"));
                    goto cleanup;
                }
            } else {
1178 1179 1180 1181 1182
                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;
1183 1184 1185 1186 1187 1188 1189 1190 1191 1192
            }
        }
    }

    /* PIIX3 (ISA bridge, IDE controller, something else unknown, USB controller)
     * hardcoded slot=1, multifunction device
     */
    if (addrs->nbuses) {
        memset(&tmp_addr, 0, sizeof(tmp_addr));
        tmp_addr.slot = 1;
1193
        if (virDomainPCIAddressReserveAddr(addrs, &tmp_addr, flags) < 0)
1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204
            goto cleanup;
    }

    if (def->nvideos > 0) {
        /* 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];
1205
        if (virDeviceInfoPCIAddressWanted(&primaryVideo->info)) {
1206 1207 1208 1209 1210 1211
            memset(&tmp_addr, 0, sizeof(tmp_addr));
            tmp_addr.slot = 2;

            if (!(addrStr = virDomainPCIAddressAsString(&tmp_addr)))
                goto cleanup;
            if (!virDomainPCIAddressValidate(addrs, &tmp_addr,
1212
                                             addrStr, flags, true))
1213 1214 1215 1216
                goto cleanup;

            if (virDomainPCIAddressSlotInUse(addrs, &tmp_addr)) {
                if (qemuDeviceVideoUsable) {
1217
                    if (qemuDomainPCIAddressReserveNextAddr(addrs,
1218
                                                            &primaryVideo->info) < 0) {
1219
                        goto cleanup;
1220
                    }
1221 1222 1223 1224 1225 1226 1227
                } 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 {
1228
                if (virDomainPCIAddressReserveAddr(addrs, &tmp_addr, flags) < 0)
1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252
                    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");
1253
        } else if (virDomainPCIAddressReserveAddr(addrs, &tmp_addr, flags) < 0) {
1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270
            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;
1271
    virPCIDeviceAddress tmp_addr;
1272 1273
    bool qemuDeviceVideoUsable = virQEMUCapsGet(qemuCaps, QEMU_CAPS_DEVICE_VIDEO_PRIMARY);
    char *addrStr = NULL;
1274
    virDomainPCIConnectFlags flags = VIR_PCI_CONNECT_TYPE_PCIE_DEVICE;
1275 1276

    for (i = 0; i < def->ncontrollers; i++) {
1277 1278 1279
        virDomainControllerDefPtr cont = def->controllers[i];

        switch (cont->type) {
1280 1281 1282 1283 1284
        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.
             */
1285 1286 1287 1288 1289 1290
            if (cont->idx == 0) {
                if (virDeviceInfoPCIAddressPresent(&cont->info)) {
                    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) {
1291 1292 1293 1294 1295
                        virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                                       _("Primary SATA controller must have PCI address 0:0:1f.2"));
                        goto cleanup;
                    }
                } else {
1296 1297 1298 1299 1300
                    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;
1301 1302 1303 1304 1305
                }
            }
            break;

        case VIR_DOMAIN_CONTROLLER_TYPE_USB:
1306 1307
            if ((cont->model == VIR_DOMAIN_CONTROLLER_MODEL_USB_ICH9_UHCI1) &&
                (cont->info.type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_NONE)) {
1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328
                /* 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) {
1329
                    if (virDomainPCIAddressReserveAddr(addrs,
1330
                                                       &tmp_addr, flags) < 0) {
1331
                        goto cleanup;
1332 1333
                    }

1334 1335 1336 1337 1338 1339
                    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;
1340 1341 1342 1343 1344
                }
            }
            break;

        case VIR_DOMAIN_CONTROLLER_TYPE_PCI:
1345 1346
            if (cont->model == VIR_DOMAIN_CONTROLLER_MODEL_DMI_TO_PCI_BRIDGE &&
                cont->info.type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_NONE) {
1347 1348 1349 1350 1351 1352 1353
                /* 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)) {
1354
                    if (virDomainPCIAddressReserveAddr(addrs,
1355
                                                       &tmp_addr, flags) < 0) {
1356
                        goto cleanup;
1357 1358
                    }

1359 1360 1361 1362 1363
                    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;
1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379
                }
            }
            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;
1380
        if (virDomainPCIAddressReserveAddr(addrs, &tmp_addr, flags) < 0)
1381
           goto cleanup;
1382

1383 1384
        tmp_addr.function = 3;
        tmp_addr.multi = VIR_TRISTATE_SWITCH_ABSENT;
1385
        if (virDomainPCIAddressReserveAddr(addrs, &tmp_addr, flags) < 0)
1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396
           goto cleanup;
    }

    if (def->nvideos > 0) {
        /* NB: unlike the pc machinetypes, on q35 machinetypes the
         * integrated devices are at slot 0x1f, so when qemu looks for
         * the first free lot for the first VGA, it will always be at
         * slot 1 (which was used up by the integrated PIIX3 devices
         * on pc machinetypes).
         */
        virDomainVideoDefPtr primaryVideo = def->videos[0];
1397
        if (virDeviceInfoPCIAddressWanted(&primaryVideo->info)) {
1398 1399 1400 1401 1402 1403
            memset(&tmp_addr, 0, sizeof(tmp_addr));
            tmp_addr.slot = 1;

            if (!(addrStr = virDomainPCIAddressAsString(&tmp_addr)))
                goto cleanup;
            if (!virDomainPCIAddressValidate(addrs, &tmp_addr,
1404
                                             addrStr, flags, true))
1405 1406 1407 1408
                goto cleanup;

            if (virDomainPCIAddressSlotInUse(addrs, &tmp_addr)) {
                if (qemuDeviceVideoUsable) {
1409
                    if (qemuDomainPCIAddressReserveNextAddr(addrs,
1410
                                                            &primaryVideo->info) < 0)
1411 1412 1413 1414 1415 1416 1417 1418
                        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 {
1419
                if (virDomainPCIAddressReserveAddr(addrs, &tmp_addr, flags) < 0)
1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444
                    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();
1445
        } else if (virDomainPCIAddressReserveAddr(addrs,
1446
                                                  &tmp_addr, flags) < 0) {
1447 1448 1449
            goto cleanup;
        }
    }
1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466

    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 ||
                !virDeviceInfoPCIAddressWanted(&sound->info)) {
                continue;
            }
1467
            if (virDomainPCIAddressReserveAddr(addrs, &tmp_addr, flags) < 0)
1468 1469 1470 1471 1472 1473 1474 1475
                goto cleanup;

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

1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487
    ret = 0;
 cleanup:
    VIR_FREE(addrStr);
    return ret;
}


static int
qemuDomainValidateDevicePCISlotsChipsets(virDomainDefPtr def,
                                         virQEMUCapsPtr qemuCaps,
                                         virDomainPCIAddressSetPtr addrs)
{
1488
    if (qemuDomainIsI440FX(def) &&
1489 1490 1491 1492
        qemuDomainValidateDevicePCISlotsPIIX3(def, qemuCaps, addrs) < 0) {
        return -1;
    }

1493
    if (qemuDomainIsQ35(def) &&
1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546
        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
 *    qemuValidateDevicePCISlotsChipsets invoked right before this function
 *
 * 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++) {
1547 1548 1549 1550
        virDomainControllerDefPtr cont = def->controllers[i];

        if (cont->type == VIR_DOMAIN_CONTROLLER_TYPE_PCI) {
            virDomainControllerModelPCI model = cont->model;
1551

1552 1553
            if (model == VIR_DOMAIN_CONTROLLER_MODEL_PCI_ROOT ||
                model == VIR_DOMAIN_CONTROLLER_MODEL_PCIE_ROOT ||
1554
                !virDeviceInfoPCIAddressWanted(&cont->info))
1555
                continue;
1556

1557
            if (qemuDomainPCIAddressReserveNextAddr(addrs, &cont->info) < 0)
1558 1559 1560 1561 1562
                goto error;
        }
    }

    for (i = 0; i < def->nfss; i++) {
1563
        if (!virDeviceInfoPCIAddressWanted(&def->fss[i]->info))
1564 1565 1566 1567
            continue;

        /* Only support VirtIO-9p-pci so far. If that changes,
         * we might need to skip devices here */
1568
        if (qemuDomainPCIAddressReserveNextAddr(addrs, &def->fss[i]->info) < 0)
1569 1570 1571 1572 1573
            goto error;
    }

    /* Network interfaces */
    for (i = 0; i < def->nnets; i++) {
1574 1575
        virDomainNetDefPtr net = def->nets[i];

1576 1577 1578 1579
        /* type='hostdev' network devices might be USB, and are also
         * in hostdevs list anyway, so handle them with other hostdevs
         * instead of here.
         */
1580 1581
        if ((net->type == VIR_DOMAIN_NET_TYPE_HOSTDEV) ||
            !virDeviceInfoPCIAddressWanted(&net->info)) {
1582 1583
            continue;
        }
1584

1585
        if (qemuDomainPCIAddressReserveNextAddr(addrs, &net->info) < 0)
1586 1587 1588 1589 1590
            goto error;
    }

    /* Sound cards */
    for (i = 0; i < def->nsounds; i++) {
1591 1592 1593
        virDomainSoundDefPtr sound = def->sounds[i];

        if (!virDeviceInfoPCIAddressWanted(&sound->info))
1594
            continue;
1595

1596
        /* Skip ISA sound card, PCSPK and usb-audio */
1597 1598 1599
        if (sound->model == VIR_DOMAIN_SOUND_MODEL_SB16 ||
            sound->model == VIR_DOMAIN_SOUND_MODEL_PCSPK ||
            sound->model == VIR_DOMAIN_SOUND_MODEL_USB) {
1600
            continue;
1601
        }
1602

1603
        if (qemuDomainPCIAddressReserveNextAddr(addrs, &sound->info) < 0)
1604 1605 1606 1607 1608
            goto error;
    }

    /* Device controllers (SCSI, USB, but not IDE, FDC or CCID) */
    for (i = 0; i < def->ncontrollers; i++) {
1609 1610
        virDomainControllerDefPtr cont = def->controllers[i];

1611
        /* PCI controllers have been dealt with earlier */
1612
        if (cont->type == VIR_DOMAIN_CONTROLLER_TYPE_PCI)
1613 1614 1615
            continue;

        /* USB controller model 'none' doesn't need a PCI address */
1616 1617
        if (cont->type == VIR_DOMAIN_CONTROLLER_TYPE_USB &&
            cont->model == VIR_DOMAIN_CONTROLLER_MODEL_USB_NONE)
1618 1619 1620
            continue;

        /* FDC lives behind the ISA bridge; CCID is a usb device */
1621 1622
        if (cont->type == VIR_DOMAIN_CONTROLLER_TYPE_FDC ||
            cont->type == VIR_DOMAIN_CONTROLLER_TYPE_CCID)
1623 1624 1625 1626
            continue;

        /* First IDE controller lives on the PIIX3 at slot=1, function=1,
           dealt with earlier on*/
1627 1628
        if (cont->type == VIR_DOMAIN_CONTROLLER_TYPE_IDE &&
            cont->idx == 0)
1629 1630
            continue;

1631
        if (!virDeviceInfoPCIAddressWanted(&cont->info))
1632 1633 1634
            continue;

        /* USB2 needs special handling to put all companions in the same slot */
1635
        if (IS_USB2_CONTROLLER(cont)) {
1636
            virPCIDeviceAddress addr = {0};
1637 1638 1639 1640
            bool foundAddr = false;

            for (j = 0; j < def->ncontrollers; j++) {
                if (IS_USB2_CONTROLLER(def->controllers[j]) &&
1641
                    def->controllers[j]->idx == cont->idx &&
1642
                    virDeviceInfoPCIAddressPresent(&def->controllers[j]->info)) {
1643 1644 1645 1646 1647 1648
                    addr = def->controllers[j]->info.addr.pci;
                    foundAddr = true;
                    break;
                }
            }

1649
            switch (cont->model) {
1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667
            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;
            }

1668 1669
            if (foundAddr) {
                /* Reserve this function on the slot we found */
1670
                if (virDomainPCIAddressReserveAddr(addrs, &addr,
1671
                                                   cont->info.pciConnectFlags) < 0) {
1672
                    goto error;
1673
                }
1674

1675 1676 1677 1678 1679
                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 */
1680 1681 1682
                if (virDomainPCIAddressReserveNextAddr(addrs, &cont->info,
                                                       cont->info.pciConnectFlags,
                                                       addr.function) < 0) {
1683 1684
                    goto error;
                }
1685

1686
                cont->info.addr.pci.multi = addr.multi;
1687 1688
            }
        } else {
1689
            if (qemuDomainPCIAddressReserveNextAddr(addrs, &cont->info) < 0)
1690
                 goto error;
1691 1692 1693 1694 1695 1696 1697 1698 1699 1700
        }
    }

    /* 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 */
1701
        if (virDeviceInfoPCIAddressPresent(&def->disks[i]->info) ||
1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713
            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;

1714
        if (!virDeviceInfoPCIAddressWanted(&def->disks[i]->info)) {
1715 1716 1717 1718 1719 1720
            virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
                           _("virtio disk cannot have an address of type '%s'"),
                           virDomainDeviceAddressTypeToString(def->disks[i]->info.type));
            goto error;
        }

1721
        if (qemuDomainPCIAddressReserveNextAddr(addrs, &def->disks[i]->info) < 0)
1722 1723 1724 1725 1726
            goto error;
    }

    /* Host PCI devices */
    for (i = 0; i < def->nhostdevs; i++) {
1727
        virDomainHostdevSubsysPtr subsys = &def->hostdevs[i]->source.subsys;
1728
        if (!virDeviceInfoPCIAddressWanted(def->hostdevs[i]->info))
1729
            continue;
1730 1731
        if (def->hostdevs[i]->mode != VIR_DOMAIN_HOSTDEV_MODE_SUBSYS)
            continue;
1732 1733 1734 1735
        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)) {
1736
            continue;
1737
        }
1738

1739
        if (qemuDomainPCIAddressReserveNextAddr(addrs,
1740
                                                def->hostdevs[i]->info) < 0)
1741 1742 1743 1744 1745 1746
            goto error;
    }

    /* VirtIO balloon */
    if (def->memballoon &&
        def->memballoon->model == VIR_DOMAIN_MEMBALLOON_MODEL_VIRTIO &&
1747
        virDeviceInfoPCIAddressWanted(&def->memballoon->info)) {
1748

1749
        if (qemuDomainPCIAddressReserveNextAddr(addrs,
1750
                                                &def->memballoon->info) < 0)
1751 1752 1753 1754 1755 1756
            goto error;
    }

    /* VirtIO RNG */
    for (i = 0; i < def->nrngs; i++) {
        if (def->rngs[i]->model != VIR_DOMAIN_RNG_MODEL_VIRTIO ||
1757
            !virDeviceInfoPCIAddressWanted(&def->rngs[i]->info))
1758 1759
            continue;

1760
        if (qemuDomainPCIAddressReserveNextAddr(addrs, &def->rngs[i]->info) < 0)
1761 1762 1763 1764 1765 1766
            goto error;
    }

    /* A watchdog - check if it is a PCI device */
    if (def->watchdog &&
        def->watchdog->model == VIR_DOMAIN_WATCHDOG_MODEL_I6300ESB &&
1767
        virDeviceInfoPCIAddressWanted(&def->watchdog->info)) {
1768
        if (qemuDomainPCIAddressReserveNextAddr(addrs, &def->watchdog->info) < 0)
1769 1770 1771 1772 1773 1774
            goto error;
    }

    /* Assign a PCI slot to the primary video card if there is not an
     * assigned address. */
    if (def->nvideos > 0 &&
1775
        virDeviceInfoPCIAddressWanted(&def->videos[0]->info)) {
1776
        if (qemuDomainPCIAddressReserveNextAddr(addrs, &def->videos[0]->info) < 0)
1777 1778 1779 1780
            goto error;
    }

    for (i = 1; i < def->nvideos; i++) {
1781
        if (!virDeviceInfoPCIAddressWanted(&def->videos[i]->info))
1782
            continue;
1783

1784
        if (qemuDomainPCIAddressReserveNextAddr(addrs, &def->videos[i]->info) < 0)
1785 1786 1787 1788 1789
            goto error;
    }

    /* Shared Memory */
    for (i = 0; i < def->nshmems; i++) {
1790
        if (!virDeviceInfoPCIAddressWanted(&def->shmems[i]->info))
1791 1792
            continue;

1793
        if (qemuDomainPCIAddressReserveNextAddr(addrs, &def->shmems[i]->info) < 0)
1794 1795 1796
            goto error;
    }
    for (i = 0; i < def->ninputs; i++) {
1797 1798
        if (def->inputs[i]->bus != VIR_DOMAIN_INPUT_BUS_VIRTIO ||
            !virDeviceInfoPCIAddressWanted(&def->inputs[i]->info))
1799 1800
            continue;

1801
        if (qemuDomainPCIAddressReserveNextAddr(addrs, &def->inputs[i]->info) < 0)
1802 1803 1804 1805 1806 1807 1808 1809
            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];

1810 1811
        if (chr->targetType != VIR_DOMAIN_CHR_SERIAL_TARGET_TYPE_PCI ||
            !virDeviceInfoPCIAddressWanted(&chr->info))
1812 1813
            continue;

1814
        if (qemuDomainPCIAddressReserveNextAddr(addrs, &chr->info) < 0)
1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840
            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) */
    }

    return 0;

 error:
    return -1;
}


static bool
qemuDomainSupportsPCI(virDomainDefPtr def,
                      virQEMUCapsPtr qemuCaps)
{
    if ((def->os.arch != VIR_ARCH_ARMV7L) && (def->os.arch != VIR_ARCH_AARCH64))
        return true;

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

1841
    if (qemuDomainIsVirt(def) &&
1842 1843 1844 1845 1846 1847 1848
        virQEMUCapsGet(qemuCaps, QEMU_CAPS_OBJECT_GPEX))
        return true;

    return false;
}


1849
static void
1850 1851
qemuDomainPCIControllerSetDefaultModelName(virDomainControllerDefPtr cont,
                                           virQEMUCapsPtr qemuCaps)
1852 1853 1854 1855 1856 1857
{
    int *modelName = &cont->opts.pciopts.modelName;

    /* make sure it's not already set */
    if (*modelName != VIR_DOMAIN_CONTROLLER_PCI_MODEL_NAME_NONE)
        return;
1858

1859 1860 1861 1862 1863 1864 1865 1866
    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;
    case VIR_DOMAIN_CONTROLLER_MODEL_PCIE_ROOT_PORT:
1867 1868 1869 1870 1871 1872
        /* 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;
1873 1874 1875 1876 1877 1878 1879
        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;
1880
    case VIR_DOMAIN_CONTROLLER_MODEL_PCI_EXPANDER_BUS:
1881 1882
        *modelName = VIR_DOMAIN_CONTROLLER_PCI_MODEL_NAME_PXB;
        break;
1883
    case VIR_DOMAIN_CONTROLLER_MODEL_PCIE_EXPANDER_BUS:
1884 1885
        *modelName = VIR_DOMAIN_CONTROLLER_PCI_MODEL_NAME_PXB_PCIE;
        break;
1886 1887 1888 1889 1890 1891 1892 1893
    case VIR_DOMAIN_CONTROLLER_MODEL_PCI_ROOT:
    case VIR_DOMAIN_CONTROLLER_MODEL_PCIE_ROOT:
    case VIR_DOMAIN_CONTROLLER_MODEL_PCI_LAST:
        break;
    }
}


1894 1895 1896 1897 1898 1899 1900 1901 1902 1903
static int
qemuDomainAddressFindNewBusNr(virDomainDefPtr def)
{
/* 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
1904
 *    implements the pci-expander-bus) includes a pci-bridge within
1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942
 *    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.
 */
    size_t i;
    int lowestBusNr = 256;

    for (i = 0; i < def->ncontrollers; i++) {
1943 1944 1945 1946
        virDomainControllerDefPtr cont = def->controllers[i];

        if (cont->type == VIR_DOMAIN_CONTROLLER_TYPE_PCI) {
            int thisBusNr = cont->opts.pciopts.busNr;
1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963

            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;
}


1964 1965 1966
static int
qemuDomainAssignPCIAddresses(virDomainDefPtr def,
                             virQEMUCapsPtr qemuCaps,
1967
                             virQEMUDriverPtr driver,
1968 1969 1970 1971 1972
                             virDomainObjPtr obj)
{
    int ret = -1;
    virDomainPCIAddressSetPtr addrs = NULL;
    qemuDomainObjPrivatePtr priv = NULL;
1973 1974 1975 1976
    int max_idx = -1;
    int nbuses = 0;
    size_t i;
    int rv;
1977

1978
    for (i = 0; i < def->ncontrollers; i++) {
1979 1980 1981 1982 1983
        virDomainControllerDefPtr cont = def->controllers[i];

        if (cont->type == VIR_DOMAIN_CONTROLLER_TYPE_PCI) {
            if ((int) cont->idx > max_idx)
                max_idx = cont->idx;
1984
        }
1985
    }
1986

1987
    nbuses = max_idx + 1;
1988

1989 1990 1991 1992
    /* 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.
     */
1993
    if (qemuDomainFillAllPCIConnectFlags(def, qemuCaps, driver) < 0)
1994 1995
        goto cleanup;

1996
    if (nbuses > 0) {
1997 1998 1999
        /* 1st pass to figure out how many PCI bridges we need */
        if (!(addrs = qemuDomainPCIAddressSetCreate(def, nbuses, true)))
            goto cleanup;
2000

2001 2002 2003
        if (qemuDomainValidateDevicePCISlotsChipsets(def, qemuCaps,
                                                     addrs) < 0)
            goto cleanup;
2004

2005 2006 2007 2008 2009 2010 2011 2012
        /* 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)
2013
         *
2014 2015 2016 2017
         * 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.
2018
         */
2019

2020
        if (qemuDomainHasPCIRoot(def)) {
2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032
            /* 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 |
                                    VIR_PCI_CONNECT_TYPE_PCI_DEVICE)
            };
            bool buses_reserved = true;

            for (i = 0; i < addrs->nbuses; i++) {
2033
                if (!virDomainPCIAddressBusIsFullyReserved(&addrs->buses[i])) {
2034 2035 2036 2037 2038
                    buses_reserved = false;
                    break;
                }
            }
            if (!buses_reserved &&
2039
                qemuDomainPCIAddressReserveNextAddr(addrs, &info) < 0)
2040 2041
                goto cleanup;
        }
2042

2043 2044
        if (qemuDomainAssignDevicePCISlots(def, qemuCaps, addrs) < 0)
            goto cleanup;
2045

2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061
        /* 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 &&
2062
            qemuDomainHasPCIeRoot(def)) {
2063 2064 2065 2066 2067 2068 2069 2070
            virDomainDeviceInfo info = {
                .pciConnectFlags = (VIR_PCI_CONNECT_HOTPLUGGABLE |
                                    VIR_PCI_CONNECT_TYPE_PCIE_DEVICE)
            };

            /* if there isn't an empty pcie-root-port, this will
             * cause one to be added
             */
2071
            if (qemuDomainPCIAddressReserveNextAddr(addrs, &info) < 0)
2072 2073 2074 2075 2076 2077
               goto cleanup;
        }

        /* now reflect any controllers auto-added to addrs into the
         * domain controllers list
         */
2078
        for (i = 1; i < addrs->nbuses; i++) {
2079 2080
            virDomainDeviceDef dev;
            int contIndex;
2081
            virDomainPCIAddressBusPtr bus = &addrs->buses[i];
2082

2083 2084 2085 2086
            if ((rv = virDomainDefMaybeAddController(
                     def, VIR_DOMAIN_CONTROLLER_TYPE_PCI,
                     i, bus->model)) < 0)
                goto cleanup;
2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109

            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 */
2110
            qemuDomainFillDevicePCIConnectFlags(def, &dev, qemuCaps, driver);
2111 2112
            if (qemuDomainPCIAddressReserveNextAddr(addrs,
                                                    &dev.data.controller->info) < 0) {
2113
                goto cleanup;
2114
            }
2115
        }
2116

2117 2118 2119 2120
        nbuses = addrs->nbuses;
        virDomainPCIAddressSetFree(addrs);
        addrs = NULL;
    }
2121

2122 2123
    if (!(addrs = qemuDomainPCIAddressSetCreate(def, nbuses, false)))
        goto cleanup;
2124

2125 2126 2127 2128
    if (qemuDomainSupportsPCI(def, qemuCaps)) {
        if (qemuDomainValidateDevicePCISlotsChipsets(def, qemuCaps,
                                                     addrs) < 0)
            goto cleanup;
2129

2130 2131
        if (qemuDomainAssignDevicePCISlots(def, qemuCaps, addrs) < 0)
            goto cleanup;
2132

2133 2134 2135 2136 2137
        /* set multi attribute for devices at function 0 of
         * any slot that has multiple functions in use
         */
        virDomainPCIAddressSetAllMulti(def);

2138 2139 2140 2141 2142
        for (i = 0; i < def->ncontrollers; i++) {
            virDomainControllerDefPtr cont = def->controllers[i];
            int idx = cont->idx;
            virPCIDeviceAddressPtr addr;
            virDomainPCIControllerOptsPtr options;
2143

2144 2145
            if (cont->type != VIR_DOMAIN_CONTROLLER_TYPE_PCI)
                continue;
2146

2147 2148
            addr = &cont->info.addr.pci;
            options = &cont->opts.pciopts;
2149

2150 2151 2152 2153
            /* set default model name (the actual name of the
             * device in qemu) for any controller that doesn't yet
             * have it set.
             */
2154
            qemuDomainPCIControllerSetDefaultModelName(cont, qemuCaps);
2155

2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181
            /* 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) {
2182
                    virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
2183 2184 2185 2186 2187
                                   _("No free busNr lower than current "
                                     "lowest busNr is available to "
                                     "auto-assign to bus %d. Must be "
                                     "manually assigned"),
                                   addr->bus);
2188 2189
                    goto cleanup;
                }
2190 2191 2192 2193 2194 2195 2196 2197 2198
                break;
            case VIR_DOMAIN_CONTROLLER_MODEL_DMI_TO_PCI_BRIDGE:
            case VIR_DOMAIN_CONTROLLER_MODEL_PCIE_SWITCH_UPSTREAM_PORT:
            case VIR_DOMAIN_CONTROLLER_MODEL_PCI_ROOT:
            case VIR_DOMAIN_CONTROLLER_MODEL_PCIE_ROOT:
            case VIR_DOMAIN_CONTROLLER_MODEL_PCI_LAST:
                break;
            }

2199
            /* check if every PCI bridge controller's index is larger than
2200 2201 2202 2203 2204 2205
             * 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 "
2206 2207
                                 "bus='0x%02x', but index must be "
                                 "larger than bus"),
2208 2209
                               idx, idx, addr->bus);
                goto cleanup;
2210 2211 2212 2213 2214 2215
            }
        }
    }

    if (obj && obj->privateData) {
        priv = obj->privateData;
J
John Ferlan 已提交
2216 2217 2218
        /* if this is the live domain object, we persist the PCI addresses */
        priv->pciaddrs = addrs;
        addrs = NULL;
2219 2220 2221 2222 2223 2224 2225 2226 2227 2228 2229
    }

    ret = 0;

 cleanup:
    virDomainPCIAddressSetFree(addrs);

    return ret;
}


J
Ján Tomko 已提交
2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241
struct qemuAssignUSBIteratorInfo {
    virDomainUSBAddressSetPtr addrs;
    size_t count;
};


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

2242 2243 2244 2245 2246 2247
    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
2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291
        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);
}


2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308
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;
2309
    size_t hubs_needed = 0;
2310
    int ret = -1;
2311
    size_t i;
2312 2313 2314 2315 2316 2317 2318

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

2319 2320 2321 2322 2323 2324 2325 2326
    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++) {
2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 2337 2338 2339 2340 2341
        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;
}


2342 2343 2344 2345 2346 2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364 2365 2366 2367 2368 2369 2370 2371 2372 2373
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
2374
                       _("failed to find an empty memory slot"));
2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429
        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
2430 2431
static int
qemuDomainAssignUSBAddresses(virDomainDefPtr def,
2432 2433
                             virDomainObjPtr obj,
                             bool newDomain)
J
Ján Tomko 已提交
2434 2435 2436 2437 2438
{
    int ret = -1;
    virDomainUSBAddressSetPtr addrs = NULL;
    qemuDomainObjPrivatePtr priv = NULL;

2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449
    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
2450 2451 2452
    if (!(addrs = virDomainUSBAddressSetCreate()))
        goto cleanup;

2453 2454 2455
    if (qemuDomainUSBAddressAddHubs(def) < 0)
        goto cleanup;

J
Ján Tomko 已提交
2456 2457 2458 2459 2460 2461 2462 2463 2464
    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
2465 2466 2467 2468 2469 2470 2471 2472
    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
2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485
    if (obj && obj->privateData) {
        priv = obj->privateData;
        priv->usbaddrs = addrs;
        addrs = NULL;
    }
    ret = 0;

 cleanup:
    virDomainUSBAddressSetFree(addrs);
    return ret;
}


2486 2487 2488
int
qemuDomainAssignAddresses(virDomainDefPtr def,
                          virQEMUCapsPtr qemuCaps,
2489
                          virQEMUDriverPtr driver,
2490
                          virDomainObjPtr obj,
J
Ján Tomko 已提交
2491
                          bool newDomain)
2492
{
2493
    if (qemuDomainAssignVirtioSerialAddresses(def) < 0)
2494
        return -1;
2495

2496 2497
    if (qemuDomainAssignSpaprVIOAddresses(def, qemuCaps) < 0)
        return -1;
2498

T
Tomasz Flendrich 已提交
2499
    if (qemuDomainAssignS390Addresses(def, qemuCaps) < 0)
2500
        return -1;
2501

2502
    qemuDomainAssignARMVirtioMMIOAddresses(def, qemuCaps);
2503

2504
    if (qemuDomainAssignPCIAddresses(def, qemuCaps, driver, obj) < 0)
2505 2506
        return -1;

2507
    if (qemuDomainAssignUSBAddresses(def, obj, newDomain) < 0)
J
Ján Tomko 已提交
2508 2509
        return -1;

2510 2511 2512
    if (qemuDomainAssignMemorySlots(def) < 0)
        return -1;

2513
    return 0;
2514 2515
}

2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 2527 2528 2529 2530 2531 2532
/**
 * 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,
2533 2534
                           virDomainDeviceDefPtr dev,
                           virQEMUDriverPtr driver)
2535 2536 2537 2538 2539 2540 2541
{
    qemuDomainObjPrivatePtr priv = obj->privateData;
    virDomainDeviceInfoPtr info = virDomainDeviceGetInfo(dev);

    if (!info)
        return 0;

2542
    qemuDomainFillDevicePCIConnectFlags(obj->def, dev, priv->qemuCaps, driver);
2543 2544 2545 2546

    return virDomainPCIAddressEnsureAddr(priv->pciaddrs, info,
                                         info->pciConnectFlags);
}
2547 2548 2549 2550 2551 2552 2553 2554 2555 2556 2557

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

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

T
Tomasz Flendrich 已提交
2558
    if (virDeviceInfoPCIAddressPresent(info) &&
2559
        virDomainPCIAddressReleaseAddr(priv->pciaddrs,
T
Tomasz Flendrich 已提交
2560
                                       &info->addr.pci) < 0)
2561 2562
        VIR_WARN("Unable to release PCI address on %s",
                 NULLSTR(devstr));
2563

2564 2565 2566 2567 2568
    if (info->type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_USB &&
        priv->usbaddrs &&
        virDomainUSBAddressRelease(priv->usbaddrs, info) < 0)
        VIR_WARN("Unable to release USB address on %s",
                 NULLSTR(devstr));
2569
}