qemu_domain_address.c 102.8 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
/*
 * qemu_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"
28
#include "network/bridge_driver.h"
29 30 31 32 33 34 35 36 37 38 39 40 41 42
#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


43 44
/**
 * @def: Domain definition
45
 * @cont: Domain controller def
46 47
 * @qemuCaps: qemu capabilities
 *
48 49
 * If the controller model is already defined, return it immediately;
 * otherwise, based on the @qemuCaps return a default model value.
50
 *
51
 * Returns model on success, -1 on failure with error set.
52
 */
53
int
54 55 56
qemuDomainGetSCSIControllerModel(const virDomainDef *def,
                                 const virDomainControllerDef *cont,
                                 virQEMUCapsPtr qemuCaps)
57
{
58 59 60 61 62
    if (cont->model > 0)
        return cont->model;

    if (qemuDomainIsPSeries(def))
        return VIR_DOMAIN_CONTROLLER_MODEL_SCSI_IBMVSCSI;
63 64
    else if (ARCH_IS_S390(def->os.arch))
        return VIR_DOMAIN_CONTROLLER_MODEL_SCSI_VIRTIO_SCSI;
65 66 67 68 69 70 71 72 73
    else if (virQEMUCapsGet(qemuCaps, QEMU_CAPS_SCSI_LSI))
        return VIR_DOMAIN_CONTROLLER_MODEL_SCSI_LSILOGIC;
    else if (virQEMUCapsGet(qemuCaps, QEMU_CAPS_VIRTIO_SCSI))
        return VIR_DOMAIN_CONTROLLER_MODEL_SCSI_VIRTIO_SCSI;

    virReportError(VIR_ERR_INTERNAL_ERROR,
                   _("Unable to determine model for SCSI controller idx=%d"),
                   cont->idx);
    return -1;
74 75 76
}


77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101
/**
 * @def: Domain definition
 * @cont: Domain controller def
 * @qemuCaps: qemu capabilities
 *
 * Set the controller model based on the existing value and the
 * capabilities if possible.
 *
 * Returns 0 on success, -1 on failure with error set.
 */
int
qemuDomainSetSCSIControllerModel(const virDomainDef *def,
                                 virDomainControllerDefPtr cont,
                                 virQEMUCapsPtr qemuCaps)
{
    int model = qemuDomainGetSCSIControllerModel(def, cont, qemuCaps);

    if (model < 0)
        return -1;

    cont->model = model;
    return 0;
}


102 103 104 105 106 107 108 109 110 111 112
/**
 * @def: Domain definition
 * @info: Domain device info
 *
 * Using the device info, find the controller related to the
 * device by index and use that controller to return the model.
 *
 * Returns the model if found, -1 if not with an error message set
 */
int
qemuDomainFindSCSIControllerModel(const virDomainDef *def,
113
                                  virDomainDeviceInfoPtr info)
114 115 116 117 118 119 120 121 122 123
{
    virDomainControllerDefPtr cont;

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

124
    return cont->model;
125 126 127
}


128
static int
129
qemuDomainAssignVirtioSerialAddresses(virDomainDefPtr def)
130 131 132 133 134
{
    int ret = -1;
    size_t i;
    virDomainVirtioSerialAddrSetPtr addrs = NULL;

135
    if (!(addrs = virDomainVirtioSerialAddrSetCreateFromDomain(def)))
136 137 138 139 140 141 142 143 144
        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) &&
145 146
            virDomainVirtioSerialAddrAutoAssignFromCache(def, addrs,
                                                         &chr->info, true) < 0)
147 148 149 150 151 152 153 154
            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) &&
155 156
            virDomainVirtioSerialAddrAutoAssignFromCache(def, addrs,
                                                         &chr->info, false) < 0)
157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225
            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
226
qemuDomainAssignSpaprVIOAddresses(virDomainDefPtr def)
227 228 229 230 231 232 233
{
    size_t i;
    int ret = -1;

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

    for (i = 0; i < def->nnets; i++) {
234 235 236 237 238 239 240 241
        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)
242 243 244 245
            goto cleanup;
    }

    for (i = 0; i < def->ncontrollers; i++) {
246 247
        virDomainControllerDefPtr cont = def->controllers[i];

248
        if (cont->model == VIR_DOMAIN_CONTROLLER_MODEL_SCSI_IBMVSCSI &&
249 250 251 252 253
            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) {
254
            goto cleanup;
255
        }
256 257 258 259
    }

    for (i = 0; i < def->nserials; i++) {
        if (def->serials[i]->deviceType == VIR_DOMAIN_CHR_DEVICE_TYPE_SERIAL &&
260
            def->serials[i]->targetType == VIR_DOMAIN_CHR_SERIAL_TARGET_TYPE_SPAPR_VIO) {
261
            def->serials[i]->info.type = VIR_DOMAIN_DEVICE_ADDRESS_TYPE_SPAPRVIO;
262
        }
263 264 265 266 267 268
        if (qemuDomainAssignSpaprVIOAddress(def, &def->serials[i]->info,
                                            VIO_ADDR_SERIAL) < 0)
            goto cleanup;
    }

    if (def->nvram) {
269
        if (qemuDomainIsPSeries(def))
270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303
            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++) {
304 305 306 307 308
        virDomainNetDefPtr net = def->nets[i];

        if (STREQ(net->model, "virtio") &&
            net->info.type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_NONE) {
            net->info.type = type;
309 310 311 312 313 314 315 316 317 318
        }
    }

    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++) {
319 320 321 322 323 324 325
        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;
        }
326 327
    }

328 329 330 331 332 333 334 335
    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;
    }

336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354
    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;
        }
    }
}

355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376
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;
}
377 378 379 380 381 382 383 384 385

/*
 * 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
386
                              virQEMUCapsPtr qemuCaps)
387 388 389 390
{
    int ret = -1;
    virDomainCCWAddressSetPtr addrs = NULL;

391
    if (qemuDomainIsS390CCW(def) &&
392 393 394 395
        virQEMUCapsGet(qemuCaps, QEMU_CAPS_VIRTIO_CCW)) {
        qemuDomainPrimeVirtioDeviceAddresses(
            def, VIR_DOMAIN_DEVICE_ADDRESS_TYPE_CCW);

396
        if (!(addrs = qemuDomainCCWAddrSetCreateFromDomain(def)))
397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413
            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;
}


414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451
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;
}


452
static void
453 454 455
qemuDomainAssignARMVirtioMMIOAddresses(virDomainDefPtr def,
                                       virQEMUCapsPtr qemuCaps)
{
456 457 458 459 460
    if (def->os.arch != VIR_ARCH_ARMV7L &&
        def->os.arch != VIR_ARCH_AARCH64)
        return;

    if (!(STRPREFIX(def->os.machine, "vexpress-") ||
461
          qemuDomainIsVirt(def)))
462 463
        return;

464 465 466
    /* 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 */
467
    if (qemuDomainHasPCIeRoot(def) &&
468 469 470 471 472 473
        !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);
474 475 476 477
    }
}


478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496
/**
 * 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,
497
                                         virQEMUDriverPtr driver,
498
                                         virDomainPCIConnectFlags pcieFlags,
499
                                         virDomainPCIConnectFlags virtioFlags)
500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517
{
    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:
518
            case VIR_DOMAIN_CONTROLLER_MODEL_USB_QEMU_XHCI:
519
                return pcieFlags;
520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535

            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:
536
                return 0;
537 538 539
            }

        case VIR_DOMAIN_CONTROLLER_TYPE_IDE:
540 541
            return pciFlags;

542
        case VIR_DOMAIN_CONTROLLER_TYPE_SCSI:
543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559
            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;
            }

560
        case VIR_DOMAIN_CONTROLLER_TYPE_VIRTIO_SERIAL:
561
            return virtioFlags;
562 563 564 565 566 567 568 569

        case VIR_DOMAIN_CONTROLLER_TYPE_FDC:
        case VIR_DOMAIN_CONTROLLER_TYPE_CCID:
        case VIR_DOMAIN_CONTROLLER_TYPE_LAST:
            /* should be 0 */
            return pciFlags;
        }
    }
570
        break;
571 572 573

    case VIR_DOMAIN_DEVICE_FS:
        /* the only type of filesystem so far is virtio-9p-pci */
574
        return virtioFlags;
575 576 577 578 579 580 581 582 583 584

    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")) {
585
            return 0;
586
        }
587 588 589 590

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

591 592 593
        if (STREQ(net->model, "e1000e"))
            return pcieFlags;

594 595 596 597 598 599 600 601 602 603 604 605 606 607 608
        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:
609
            return 0;
610 611 612 613 614
        }

    case VIR_DOMAIN_DEVICE_DISK:
        switch ((virDomainDiskBus) dev->data.disk->bus) {
        case VIR_DOMAIN_DISK_BUS_VIRTIO:
615
            return virtioFlags; /* only virtio disks use PCI */
616 617 618 619 620 621 622 623 624 625

        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:
626
            return 0;
627 628
        }

629 630 631 632 633 634 635
    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 ||
636
            (hostdev->source.subsys.type != VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_PCI &&
637 638
             hostdev->source.subsys.type != VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_MDEV &&
             hostdev->source.subsys.type != VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_SCSI_HOST)) {
639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661
            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;
        }

662 663 664 665 666
        /* 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 */
667 668 669
        if (hostdev->source.subsys.type == VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_MDEV)
            return pcieFlags;

670 671 672 673 674 675 676
        /* according to pbonzini, from the guest PoV vhost-scsi devices
         * are the same as virtio-scsi, so they should use virtioFlags
         * (same as virtio-scsi) to determine Express vs. legacy placement
         */
        if (hostdev->source.subsys.type == VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_SCSI_HOST)
            return virtioFlags;

677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726
        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;

727
        return pciFlags;
728
    }
729 730 731 732

    case VIR_DOMAIN_DEVICE_MEMBALLOON:
        switch ((virDomainMemballoonModel) dev->data.memballoon->model) {
        case VIR_DOMAIN_MEMBALLOON_MODEL_VIRTIO:
733
            return virtioFlags;
734 735 736 737

        case VIR_DOMAIN_MEMBALLOON_MODEL_XEN:
        case VIR_DOMAIN_MEMBALLOON_MODEL_NONE:
        case VIR_DOMAIN_MEMBALLOON_MODEL_LAST:
738
            return 0;
739 740 741 742 743
        }

    case VIR_DOMAIN_DEVICE_RNG:
        switch ((virDomainRNGModel) dev->data.rng->model) {
        case VIR_DOMAIN_RNG_MODEL_VIRTIO:
744
            return virtioFlags;
745 746

        case VIR_DOMAIN_RNG_MODEL_LAST:
747
            return 0;
748 749 750 751 752 753 754 755 756 757 758
        }

    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:
759
            return 0;
760 761 762 763 764
        }

    case VIR_DOMAIN_DEVICE_VIDEO:
        switch ((virDomainVideoType) dev->data.video->type) {
        case VIR_DOMAIN_VIDEO_TYPE_VIRTIO:
765 766
            return virtioFlags;

767 768 769 770 771 772 773 774 775
        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;

776
        case VIR_DOMAIN_VIDEO_TYPE_DEFAULT:
F
Fabian Freyer 已提交
777
        case VIR_DOMAIN_VIDEO_TYPE_GOP:
778
        case VIR_DOMAIN_VIDEO_TYPE_LAST:
779
            return 0;
780 781 782 783 784 785 786 787
        }

    case VIR_DOMAIN_DEVICE_SHMEM:
        return pciFlags;

    case VIR_DOMAIN_DEVICE_INPUT:
        switch ((virDomainInputBus) dev->data.input->bus) {
        case VIR_DOMAIN_INPUT_BUS_VIRTIO:
788
            return virtioFlags;
789 790 791 792 793 794

        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:
795
            return 0;
796 797 798 799 800 801 802 803 804
        }

    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:
805
        case VIR_DOMAIN_CHR_SERIAL_TARGET_TYPE_SPAPR_VIO:
806
        case VIR_DOMAIN_CHR_SERIAL_TARGET_TYPE_SYSTEM:
807
        case VIR_DOMAIN_CHR_SERIAL_TARGET_TYPE_SCLP:
808
        case VIR_DOMAIN_CHR_SERIAL_TARGET_TYPE_NONE:
809
        case VIR_DOMAIN_CHR_SERIAL_TARGET_TYPE_LAST:
810
            return 0;
811 812
        }

813
        /* These devices don't ever connect with PCI */
814 815 816 817 818 819 820
    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:
821
        /* These devices don't even have a DeviceInfo */
822 823 824 825 826
    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:
827
        return 0;
828 829 830 831 832 833 834 835 836 837 838 839 840 841
    }

    /* 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;
842
    virQEMUDriverPtr driver;
843 844 845 846 847 848 849 850 851 852 853 854
} qemuDomainFillDevicePCIConnectFlagsIterData;


/**
 * qemuDomainFillDevicePCIConnectFlagsIterInit:
 *
 * Initialize the iterator data that is used when calling
 * qemuDomainCalculateDevicePCIConnectFlags().
 */
static void
qemuDomainFillDevicePCIConnectFlagsIterInit(virDomainDefPtr def,
                                            virQEMUCapsPtr qemuCaps,
855
                                            virQEMUDriverPtr driver,
856 857
                                            qemuDomainFillDevicePCIConnectFlagsIterData *data)
{
858 859 860

    data->driver = driver;

861
    if (qemuDomainHasPCIeRoot(def)) {
862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899
        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
900 901
        = qemuDomainDeviceCalculatePCIConnectFlags(dev, data->driver,
                                                   data->pcieFlags,
902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918
                                                   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())
 */
919
static int
920
qemuDomainFillAllPCIConnectFlags(virDomainDefPtr def,
921 922
                                 virQEMUCapsPtr qemuCaps,
                                 virQEMUDriverPtr driver)
923 924 925
{
    qemuDomainFillDevicePCIConnectFlagsIterData data;

926
    qemuDomainFillDevicePCIConnectFlagsIterInit(def, qemuCaps, driver, &data);
927 928 929 930 931 932 933

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


934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037
/**
 * qemuDomainFindUnusedIsolationGroupIter:
 * @def: domain definition
 * @dev: device definition
 * @info: device information
 * @opaque: user data
 *
 * Used to implement qemuDomainFindUnusedIsolationGroup(). You probably
 * don't want to call this directly.
 *
 * Return: 0 if the isolation group is not used by the device, <1 otherwise.
 */
static int
qemuDomainFindUnusedIsolationGroupIter(virDomainDefPtr def ATTRIBUTE_UNUSED,
                                       virDomainDeviceDefPtr dev ATTRIBUTE_UNUSED,
                                       virDomainDeviceInfoPtr info,
                                       void *opaque)
{
    unsigned int *isolationGroup = opaque;

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

    return 0;
}


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

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

    return isolationGroup;
}


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

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

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

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

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

        if (tmp < 0) {
1038 1039 1040 1041 1042
            VIR_WARN("Can't look up isolation group for host device "
                     "%04x:%02x:%02x.%x, device won't be isolated",
                     hostAddr->domain, hostAddr->bus,
                     hostAddr->slot, hostAddr->function);
            goto skip;
1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081
        }

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

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

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

        /* Network interfaces can ultimately result in the guest being
         * assigned a host device if the libvirt network they're connected
         * to is of type hostdev. All other kinds of network interfaces don't
         * require us to isolate the guest device, so we can skip them */
        if (iface->type != VIR_DOMAIN_NET_TYPE_NETWORK ||
            networkGetActualType(iface) != VIR_DOMAIN_NET_TYPE_HOSTDEV) {
            goto skip;
        }

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

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

        if (tmp == 0) {
1082 1083 1084 1085 1086
            VIR_WARN("Can't obtain usable isolation group for interface "
                     "configured to use hostdev-backed network '%s', "
                     "device won't be isolated",
                     iface->data.network.name);
            goto skip;
1087 1088 1089 1090 1091 1092 1093 1094 1095 1096
        }

        info->isolationGroup = tmp;

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

 skip:
1097
    return 0;
1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166
}


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


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

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

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

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

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

    ret = 0;

 cleanup:
    return ret;
}


1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177
/**
 * 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.
 */
1178
static void
1179 1180
qemuDomainFillDevicePCIConnectFlags(virDomainDefPtr def,
                                    virDomainDeviceDefPtr dev,
1181 1182
                                    virQEMUCapsPtr qemuCaps,
                                    virQEMUDriverPtr driver)
1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197
{
    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;

1198
        qemuDomainFillDevicePCIConnectFlagsIterInit(def, qemuCaps, driver, &data);
1199 1200

        info->pciConnectFlags
1201 1202
            = qemuDomainDeviceCalculatePCIConnectFlags(dev, data.driver,
                                                       data.pcieFlags,
1203 1204 1205 1206 1207
                                                       data.virtioFlags);
    }
}


1208
static int
1209
qemuDomainPCIAddressReserveNextAddr(virDomainPCIAddressSetPtr addrs,
1210
                                    virDomainDeviceInfoPtr dev)
1211
{
1212 1213
    return virDomainPCIAddressReserveNextAddr(addrs, dev,
                                              dev->pciConnectFlags, -1);
1214 1215 1216
}


1217 1218 1219 1220 1221 1222 1223 1224
static int
qemuDomainCollectPCIAddress(virDomainDefPtr def ATTRIBUTE_UNUSED,
                            virDomainDeviceDefPtr device,
                            virDomainDeviceInfoPtr info,
                            void *opaque)
{
    virDomainPCIAddressSetPtr addrs = opaque;
    int ret = -1;
1225
    virPCIDeviceAddressPtr addr = &info->addr.pci;
1226

1227 1228 1229
    if (!virDeviceInfoPCIAddressPresent(info) ||
        ((device->type == VIR_DOMAIN_DEVICE_HOSTDEV) &&
         (device->data.hostdev->parent.type != VIR_DOMAIN_DEVICE_NONE))) {
1230 1231 1232 1233 1234 1235 1236
        /* 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;
    }

1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258
    /* 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;
    }

1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286
    /* 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 &&
1287
                !(addrs->buses[0].flags & VIR_PCI_CONNECT_TYPE_PCI_DEVICE)) {
1288 1289 1290 1291 1292 1293 1294 1295 1296 1297
                virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                               _("Bus 0 must be PCI for integrated PIIX3 "
                                 "USB or IDE controllers"));
                return -1;
            } else {
                return 0;
            }
        }
    }

1298
    if (virDomainPCIAddressReserveAddr(addrs, addr,
1299 1300
                                       info->pciConnectFlags,
                                       info->isolationGroup) < 0) {
1301
        goto cleanup;
1302
    }
1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315

    ret = 0;
 cleanup:
    return ret;
}

static virDomainPCIAddressSetPtr
qemuDomainPCIAddressSetCreate(virDomainDefPtr def,
                              unsigned int nbuses,
                              bool dryRun)
{
    virDomainPCIAddressSetPtr addrs;
    size_t i;
1316 1317
    bool hasPCIeRoot = false;
    virDomainControllerModelPCI defaultModel;
1318 1319 1320 1321 1322 1323

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

    addrs->dryRun = dryRun;

1324 1325 1326 1327
    /* pSeries domains support multiple pci-root controllers */
    if (qemuDomainIsPSeries(def))
        addrs->multipleRootsSupported = true;

1328
    for (i = 0; i < def->ncontrollers; i++) {
1329 1330
        virDomainControllerDefPtr cont = def->controllers[i];
        size_t idx = cont->idx;
1331

1332
        if (cont->type != VIR_DOMAIN_CONTROLLER_TYPE_PCI)
1333 1334 1335 1336 1337
            continue;

        if (idx >= addrs->nbuses) {
            virReportError(VIR_ERR_INTERNAL_ERROR,
                           _("Inappropriate new pci controller index %zu "
1338
                             "exceeds addrs array length"), idx);
1339 1340 1341
            goto error;
        }

1342
        if (virDomainPCIAddressBusSetModel(&addrs->buses[idx], cont->model) < 0)
1343
            goto error;
1344

1345 1346 1347 1348
        /* Forward the information about isolation groups */
        addrs->buses[idx].isolationGroup = cont->info.isolationGroup;
        addrs->buses[idx].isolationGroupLocked = cont->info.isolationGroupLocked;

1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367
        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.
     */
1368 1369 1370 1371
    if (qemuDomainIsPSeries(def)) {
        /* pSeries guests should use PHBs (pci-root controllers) */
        defaultModel = VIR_DOMAIN_CONTROLLER_MODEL_PCI_ROOT;
    } else if (hasPCIeRoot) {
1372
        defaultModel = VIR_DOMAIN_CONTROLLER_MODEL_PCIE_ROOT_PORT;
1373
    } else {
1374
        defaultModel = VIR_DOMAIN_CONTROLLER_MODEL_PCI_BRIDGE;
1375
    }
1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387

    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);
    }
1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406

    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;
1407
    virPCIDeviceAddress tmp_addr;
1408 1409
    bool qemuDeviceVideoUsable = virQEMUCapsGet(qemuCaps, QEMU_CAPS_DEVICE_VIDEO_PRIMARY);
    char *addrStr = NULL;
1410 1411
    virDomainPCIConnectFlags flags = (VIR_PCI_CONNECT_HOTPLUGGABLE
                                      | VIR_PCI_CONNECT_TYPE_PCI_DEVICE);
1412 1413 1414

    /* Verify that first IDE and USB controllers (if any) is on the PIIX3, fn 1 */
    for (i = 0; i < def->ncontrollers; i++) {
1415 1416
        virDomainControllerDefPtr cont = def->controllers[i];

1417
        /* First IDE controller lives on the PIIX3 at slot=1, function=1 */
1418 1419 1420 1421 1422 1423 1424
        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) {
1425 1426 1427 1428 1429
                    virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                                   _("Primary IDE controller must have PCI address 0:0:1.1"));
                    goto cleanup;
                }
            } else {
1430 1431 1432 1433 1434
                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;
1435
            }
1436 1437 1438 1439 1440 1441 1442 1443 1444
        } 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) {
1445
                    virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
1446
                                   _("PIIX3 USB controller at index 0 must have PCI address 0:0:1.2"));
1447 1448 1449
                    goto cleanup;
                }
            } else {
1450 1451 1452 1453 1454
                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;
1455
            }
1456 1457 1458
        } else {
            /* this controller is not skipped in qemuDomainCollectPCIAddress */
            continue;
1459
        }
1460 1461 1462
        if (addrs->nbuses &&
            virDomainPCIAddressReserveAddr(addrs, &cont->info.addr.pci, flags, 0) < 0)
            goto cleanup;
1463 1464
    }

1465
    /* Implicit PIIX3 devices living on slot 1 not handled above */
1466 1467 1468
    if (addrs->nbuses) {
        memset(&tmp_addr, 0, sizeof(tmp_addr));
        tmp_addr.slot = 1;
1469 1470 1471 1472 1473
        /* ISA Bridge at 00:01.0 */
        if (virDomainPCIAddressReserveAddr(addrs, &tmp_addr, flags, 0) < 0)
            goto cleanup;
        /* Bridge at 00:01.3 */
        tmp_addr.function = 3;
1474
        if (virDomainPCIAddressReserveAddr(addrs, &tmp_addr, flags, 0) < 0)
1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485
            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];
1486
        if (virDeviceInfoPCIAddressWanted(&primaryVideo->info)) {
1487 1488 1489 1490 1491 1492
            memset(&tmp_addr, 0, sizeof(tmp_addr));
            tmp_addr.slot = 2;

            if (!(addrStr = virDomainPCIAddressAsString(&tmp_addr)))
                goto cleanup;
            if (!virDomainPCIAddressValidate(addrs, &tmp_addr,
1493
                                             addrStr, flags, true))
1494 1495 1496 1497
                goto cleanup;

            if (virDomainPCIAddressSlotInUse(addrs, &tmp_addr)) {
                if (qemuDeviceVideoUsable) {
1498
                    if (qemuDomainPCIAddressReserveNextAddr(addrs,
1499
                                                            &primaryVideo->info) < 0) {
1500
                        goto cleanup;
1501
                    }
1502 1503 1504 1505 1506 1507 1508
                } 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 {
1509
                if (virDomainPCIAddressReserveAddr(addrs, &tmp_addr, flags, 0) < 0)
1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533
                    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");
1534
        } else if (virDomainPCIAddressReserveAddr(addrs, &tmp_addr, flags, 0) < 0) {
1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551
            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;
1552
    virPCIDeviceAddress tmp_addr;
1553 1554
    bool qemuDeviceVideoUsable = virQEMUCapsGet(qemuCaps, QEMU_CAPS_DEVICE_VIDEO_PRIMARY);
    char *addrStr = NULL;
1555
    virDomainPCIConnectFlags flags = VIR_PCI_CONNECT_TYPE_PCIE_DEVICE;
1556 1557

    for (i = 0; i < def->ncontrollers; i++) {
1558 1559 1560
        virDomainControllerDefPtr cont = def->controllers[i];

        switch (cont->type) {
1561 1562 1563 1564 1565
        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.
             */
1566 1567 1568 1569 1570 1571
            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) {
1572 1573 1574 1575 1576
                        virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                                       _("Primary SATA controller must have PCI address 0:0:1f.2"));
                        goto cleanup;
                    }
                } else {
1577 1578 1579 1580 1581
                    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;
1582 1583 1584 1585 1586
                }
            }
            break;

        case VIR_DOMAIN_CONTROLLER_TYPE_USB:
1587 1588
            if ((cont->model == VIR_DOMAIN_CONTROLLER_MODEL_USB_ICH9_UHCI1) &&
                (cont->info.type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_NONE)) {
1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609
                /* 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) {
1610
                    if (virDomainPCIAddressReserveAddr(addrs, &tmp_addr, flags, 0) < 0)
1611
                        goto cleanup;
1612

1613 1614 1615 1616 1617 1618
                    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;
1619 1620 1621 1622 1623
                }
            }
            break;

        case VIR_DOMAIN_CONTROLLER_TYPE_PCI:
1624 1625
            if (cont->model == VIR_DOMAIN_CONTROLLER_MODEL_DMI_TO_PCI_BRIDGE &&
                cont->info.type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_NONE) {
1626 1627 1628 1629 1630 1631 1632
                /* 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)) {
1633
                    if (virDomainPCIAddressReserveAddr(addrs, &tmp_addr, flags, 0) < 0)
1634
                        goto cleanup;
1635

1636 1637 1638 1639 1640
                    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;
1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656
                }
            }
            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;
1657
        if (virDomainPCIAddressReserveAddr(addrs, &tmp_addr, flags, 0) < 0)
1658
           goto cleanup;
1659

1660 1661
        tmp_addr.function = 3;
        tmp_addr.multi = VIR_TRISTATE_SWITCH_ABSENT;
1662
        if (virDomainPCIAddressReserveAddr(addrs, &tmp_addr, flags, 0) < 0)
1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673
           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];
1674
        if (virDeviceInfoPCIAddressWanted(&primaryVideo->info)) {
1675 1676 1677 1678 1679 1680
            memset(&tmp_addr, 0, sizeof(tmp_addr));
            tmp_addr.slot = 1;

            if (!(addrStr = virDomainPCIAddressAsString(&tmp_addr)))
                goto cleanup;
            if (!virDomainPCIAddressValidate(addrs, &tmp_addr,
1681
                                             addrStr, flags, true))
1682 1683 1684 1685
                goto cleanup;

            if (virDomainPCIAddressSlotInUse(addrs, &tmp_addr)) {
                if (qemuDeviceVideoUsable) {
1686
                    if (qemuDomainPCIAddressReserveNextAddr(addrs,
1687
                                                            &primaryVideo->info) < 0)
1688 1689 1690 1691 1692 1693 1694 1695
                        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 {
1696
                if (virDomainPCIAddressReserveAddr(addrs, &tmp_addr, flags, 0) < 0)
1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721
                    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();
1722
        } else if (virDomainPCIAddressReserveAddr(addrs, &tmp_addr, flags, 0) < 0) {
1723 1724 1725
            goto cleanup;
        }
    }
1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742

    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;
            }
1743
            if (virDomainPCIAddressReserveAddr(addrs, &tmp_addr, flags, 0) < 0)
1744 1745 1746 1747 1748 1749 1750 1751
                goto cleanup;

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

1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763
    ret = 0;
 cleanup:
    VIR_FREE(addrStr);
    return ret;
}


static int
qemuDomainValidateDevicePCISlotsChipsets(virDomainDefPtr def,
                                         virQEMUCapsPtr qemuCaps,
                                         virDomainPCIAddressSetPtr addrs)
{
1764
    if (qemuDomainIsI440FX(def) &&
1765 1766 1767 1768
        qemuDomainValidateDevicePCISlotsPIIX3(def, qemuCaps, addrs) < 0) {
        return -1;
    }

1769
    if (qemuDomainIsQ35(def) &&
1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822
        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++) {
1823 1824 1825 1826
        virDomainControllerDefPtr cont = def->controllers[i];

        if (cont->type == VIR_DOMAIN_CONTROLLER_TYPE_PCI) {
            virDomainControllerModelPCI model = cont->model;
1827

1828 1829
            if (model == VIR_DOMAIN_CONTROLLER_MODEL_PCI_ROOT ||
                model == VIR_DOMAIN_CONTROLLER_MODEL_PCIE_ROOT ||
1830
                !virDeviceInfoPCIAddressWanted(&cont->info))
1831
                continue;
1832

1833
            if (qemuDomainPCIAddressReserveNextAddr(addrs, &cont->info) < 0)
1834 1835 1836 1837 1838
                goto error;
        }
    }

    for (i = 0; i < def->nfss; i++) {
1839
        if (!virDeviceInfoPCIAddressWanted(&def->fss[i]->info))
1840 1841 1842 1843
            continue;

        /* Only support VirtIO-9p-pci so far. If that changes,
         * we might need to skip devices here */
1844
        if (qemuDomainPCIAddressReserveNextAddr(addrs, &def->fss[i]->info) < 0)
1845 1846 1847 1848 1849
            goto error;
    }

    /* Network interfaces */
    for (i = 0; i < def->nnets; i++) {
1850 1851
        virDomainNetDefPtr net = def->nets[i];

1852 1853 1854 1855
        /* type='hostdev' network devices might be USB, and are also
         * in hostdevs list anyway, so handle them with other hostdevs
         * instead of here.
         */
1856 1857
        if ((net->type == VIR_DOMAIN_NET_TYPE_HOSTDEV) ||
            !virDeviceInfoPCIAddressWanted(&net->info)) {
1858 1859
            continue;
        }
1860

1861
        if (qemuDomainPCIAddressReserveNextAddr(addrs, &net->info) < 0)
1862 1863 1864 1865 1866
            goto error;
    }

    /* Sound cards */
    for (i = 0; i < def->nsounds; i++) {
1867 1868 1869
        virDomainSoundDefPtr sound = def->sounds[i];

        if (!virDeviceInfoPCIAddressWanted(&sound->info))
1870
            continue;
1871

1872
        /* Skip ISA sound card, PCSPK and usb-audio */
1873 1874 1875
        if (sound->model == VIR_DOMAIN_SOUND_MODEL_SB16 ||
            sound->model == VIR_DOMAIN_SOUND_MODEL_PCSPK ||
            sound->model == VIR_DOMAIN_SOUND_MODEL_USB) {
1876
            continue;
1877
        }
1878

1879
        if (qemuDomainPCIAddressReserveNextAddr(addrs, &sound->info) < 0)
1880 1881 1882 1883 1884
            goto error;
    }

    /* Device controllers (SCSI, USB, but not IDE, FDC or CCID) */
    for (i = 0; i < def->ncontrollers; i++) {
1885 1886
        virDomainControllerDefPtr cont = def->controllers[i];

1887
        /* PCI controllers have been dealt with earlier */
1888
        if (cont->type == VIR_DOMAIN_CONTROLLER_TYPE_PCI)
1889 1890 1891
            continue;

        /* USB controller model 'none' doesn't need a PCI address */
1892 1893
        if (cont->type == VIR_DOMAIN_CONTROLLER_TYPE_USB &&
            cont->model == VIR_DOMAIN_CONTROLLER_MODEL_USB_NONE)
1894 1895 1896
            continue;

        /* FDC lives behind the ISA bridge; CCID is a usb device */
1897 1898
        if (cont->type == VIR_DOMAIN_CONTROLLER_TYPE_FDC ||
            cont->type == VIR_DOMAIN_CONTROLLER_TYPE_CCID)
1899 1900 1901 1902
            continue;

        /* First IDE controller lives on the PIIX3 at slot=1, function=1,
           dealt with earlier on*/
1903 1904
        if (cont->type == VIR_DOMAIN_CONTROLLER_TYPE_IDE &&
            cont->idx == 0)
1905 1906
            continue;

1907
        if (!virDeviceInfoPCIAddressWanted(&cont->info))
1908 1909 1910
            continue;

        /* USB2 needs special handling to put all companions in the same slot */
1911
        if (IS_USB2_CONTROLLER(cont)) {
1912
            virPCIDeviceAddress addr = {0};
1913 1914 1915 1916
            bool foundAddr = false;

            for (j = 0; j < def->ncontrollers; j++) {
                if (IS_USB2_CONTROLLER(def->controllers[j]) &&
1917
                    def->controllers[j]->idx == cont->idx &&
1918
                    virDeviceInfoPCIAddressPresent(&def->controllers[j]->info)) {
1919 1920 1921 1922 1923 1924
                    addr = def->controllers[j]->info.addr.pci;
                    foundAddr = true;
                    break;
                }
            }

1925
            switch (cont->model) {
1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943
            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;
            }

1944 1945
            if (foundAddr) {
                /* Reserve this function on the slot we found */
1946
                if (virDomainPCIAddressReserveAddr(addrs, &addr,
1947 1948
                                                   cont->info.pciConnectFlags,
                                                   cont->info.isolationGroup) < 0) {
1949
                    goto error;
1950
                }
1951

1952 1953 1954 1955 1956
                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 */
1957 1958 1959
                if (virDomainPCIAddressReserveNextAddr(addrs, &cont->info,
                                                       cont->info.pciConnectFlags,
                                                       addr.function) < 0) {
1960 1961
                    goto error;
                }
1962

1963
                cont->info.addr.pci.multi = addr.multi;
1964 1965
            }
        } else {
1966
            if (qemuDomainPCIAddressReserveNextAddr(addrs, &cont->info) < 0)
1967
                 goto error;
1968 1969 1970 1971 1972 1973 1974 1975 1976 1977
        }
    }

    /* 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 */
1978
        if (virDeviceInfoPCIAddressPresent(&def->disks[i]->info) ||
1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990
            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;

1991
        if (!virDeviceInfoPCIAddressWanted(&def->disks[i]->info)) {
1992 1993 1994 1995 1996 1997
            virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
                           _("virtio disk cannot have an address of type '%s'"),
                           virDomainDeviceAddressTypeToString(def->disks[i]->info.type));
            goto error;
        }

1998
        if (qemuDomainPCIAddressReserveNextAddr(addrs, &def->disks[i]->info) < 0)
1999 2000 2001 2002 2003
            goto error;
    }

    /* Host PCI devices */
    for (i = 0; i < def->nhostdevs; i++) {
2004
        virDomainHostdevSubsysPtr subsys = &def->hostdevs[i]->source.subsys;
2005
        if (!virDeviceInfoPCIAddressWanted(def->hostdevs[i]->info))
2006
            continue;
2007 2008
        if (def->hostdevs[i]->mode != VIR_DOMAIN_HOSTDEV_MODE_SUBSYS)
            continue;
2009 2010 2011 2012
        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)) {
2013
            continue;
2014
        }
2015

2016
        if (qemuDomainPCIAddressReserveNextAddr(addrs,
2017
                                                def->hostdevs[i]->info) < 0)
2018 2019 2020 2021 2022 2023
            goto error;
    }

    /* VirtIO balloon */
    if (def->memballoon &&
        def->memballoon->model == VIR_DOMAIN_MEMBALLOON_MODEL_VIRTIO &&
2024
        virDeviceInfoPCIAddressWanted(&def->memballoon->info)) {
2025

2026
        if (qemuDomainPCIAddressReserveNextAddr(addrs,
2027
                                                &def->memballoon->info) < 0)
2028 2029 2030 2031 2032 2033
            goto error;
    }

    /* VirtIO RNG */
    for (i = 0; i < def->nrngs; i++) {
        if (def->rngs[i]->model != VIR_DOMAIN_RNG_MODEL_VIRTIO ||
2034
            !virDeviceInfoPCIAddressWanted(&def->rngs[i]->info))
2035 2036
            continue;

2037
        if (qemuDomainPCIAddressReserveNextAddr(addrs, &def->rngs[i]->info) < 0)
2038 2039 2040 2041 2042 2043
            goto error;
    }

    /* A watchdog - check if it is a PCI device */
    if (def->watchdog &&
        def->watchdog->model == VIR_DOMAIN_WATCHDOG_MODEL_I6300ESB &&
2044
        virDeviceInfoPCIAddressWanted(&def->watchdog->info)) {
2045
        if (qemuDomainPCIAddressReserveNextAddr(addrs, &def->watchdog->info) < 0)
2046 2047 2048 2049 2050 2051
            goto error;
    }

    /* Assign a PCI slot to the primary video card if there is not an
     * assigned address. */
    if (def->nvideos > 0 &&
2052
        virDeviceInfoPCIAddressWanted(&def->videos[0]->info)) {
2053
        if (qemuDomainPCIAddressReserveNextAddr(addrs, &def->videos[0]->info) < 0)
2054 2055 2056 2057
            goto error;
    }

    for (i = 1; i < def->nvideos; i++) {
2058
        if (!virDeviceInfoPCIAddressWanted(&def->videos[i]->info))
2059
            continue;
2060

2061
        if (qemuDomainPCIAddressReserveNextAddr(addrs, &def->videos[i]->info) < 0)
2062 2063 2064 2065 2066
            goto error;
    }

    /* Shared Memory */
    for (i = 0; i < def->nshmems; i++) {
2067
        if (!virDeviceInfoPCIAddressWanted(&def->shmems[i]->info))
2068 2069
            continue;

2070
        if (qemuDomainPCIAddressReserveNextAddr(addrs, &def->shmems[i]->info) < 0)
2071 2072 2073
            goto error;
    }
    for (i = 0; i < def->ninputs; i++) {
2074 2075
        if (def->inputs[i]->bus != VIR_DOMAIN_INPUT_BUS_VIRTIO ||
            !virDeviceInfoPCIAddressWanted(&def->inputs[i]->info))
2076 2077
            continue;

2078
        if (qemuDomainPCIAddressReserveNextAddr(addrs, &def->inputs[i]->info) < 0)
2079 2080 2081 2082 2083 2084 2085 2086
            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];

2087 2088
        if (chr->targetType != VIR_DOMAIN_CHR_SERIAL_TARGET_TYPE_PCI ||
            !virDeviceInfoPCIAddressWanted(&chr->info))
2089 2090
            continue;

2091
        if (qemuDomainPCIAddressReserveNextAddr(addrs, &chr->info) < 0)
2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117
            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;

2118
    if (qemuDomainIsVirt(def) &&
2119 2120 2121 2122 2123 2124 2125
        virQEMUCapsGet(qemuCaps, QEMU_CAPS_OBJECT_GPEX))
        return true;

    return false;
}


2126
static void
2127
qemuDomainPCIControllerSetDefaultModelName(virDomainControllerDefPtr cont,
2128
                                           virDomainDefPtr def,
2129
                                           virQEMUCapsPtr qemuCaps)
2130 2131 2132 2133 2134 2135
{
    int *modelName = &cont->opts.pciopts.modelName;

    /* make sure it's not already set */
    if (*modelName != VIR_DOMAIN_CONTROLLER_PCI_MODEL_NAME_NONE)
        return;
2136

2137 2138 2139 2140 2141 2142 2143 2144
    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:
2145 2146 2147 2148 2149 2150
        /* 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;
2151 2152 2153 2154 2155 2156 2157
        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;
2158
    case VIR_DOMAIN_CONTROLLER_MODEL_PCI_EXPANDER_BUS:
2159 2160
        *modelName = VIR_DOMAIN_CONTROLLER_PCI_MODEL_NAME_PXB;
        break;
2161
    case VIR_DOMAIN_CONTROLLER_MODEL_PCIE_EXPANDER_BUS:
2162 2163
        *modelName = VIR_DOMAIN_CONTROLLER_PCI_MODEL_NAME_PXB_PCIE;
        break;
2164
    case VIR_DOMAIN_CONTROLLER_MODEL_PCI_ROOT:
2165 2166 2167
        if (qemuDomainIsPSeries(def))
            *modelName = VIR_DOMAIN_CONTROLLER_PCI_MODEL_NAME_SPAPR_PCI_HOST_BRIDGE;
        break;
2168 2169 2170 2171 2172 2173 2174
    case VIR_DOMAIN_CONTROLLER_MODEL_PCIE_ROOT:
    case VIR_DOMAIN_CONTROLLER_MODEL_PCI_LAST:
        break;
    }
}


2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199
/**
 * qemuDomainAddressFindNewTargetIndex:
 * @def: domain definition
 *
 * Find a target index that can be used for a PCI controller.
 *
 * Returns: an unused target index, or -1 if all available target
 *          indexes are already taken.
 */
static int
qemuDomainAddressFindNewTargetIndex(virDomainDefPtr def)
{
    int targetIndex;
    int ret = -1;

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

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

            /* Skip everything but PHBs */
2200
            if (!virDomainControllerIsPSeriesPHB(cont))
2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221 2222
                continue;

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

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

    return ret;
}


2223 2224 2225 2226 2227 2228 2229 2230 2231 2232
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
2233
 *    implements the pci-expander-bus) includes a pci-bridge within
2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271
 *    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++) {
2272 2273 2274 2275
        virDomainControllerDefPtr cont = def->controllers[i];

        if (cont->type == VIR_DOMAIN_CONTROLLER_TYPE_PCI) {
            int thisBusNr = cont->opts.pciopts.busNr;
2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292

            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;
}


2293 2294 2295
static int
qemuDomainAssignPCIAddresses(virDomainDefPtr def,
                             virQEMUCapsPtr qemuCaps,
2296
                             virQEMUDriverPtr driver,
2297 2298 2299 2300 2301
                             virDomainObjPtr obj)
{
    int ret = -1;
    virDomainPCIAddressSetPtr addrs = NULL;
    qemuDomainObjPrivatePtr priv = NULL;
2302 2303 2304 2305
    int max_idx = -1;
    int nbuses = 0;
    size_t i;
    int rv;
2306

2307
    for (i = 0; i < def->ncontrollers; i++) {
2308 2309 2310 2311 2312
        virDomainControllerDefPtr cont = def->controllers[i];

        if (cont->type == VIR_DOMAIN_CONTROLLER_TYPE_PCI) {
            if ((int) cont->idx > max_idx)
                max_idx = cont->idx;
2313
        }
2314
    }
2315

2316
    nbuses = max_idx + 1;
2317

2318 2319 2320 2321
    /* 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.
     */
2322
    if (qemuDomainFillAllPCIConnectFlags(def, qemuCaps, driver) < 0)
2323 2324
        goto cleanup;

2325 2326 2327
    if (qemuDomainSetupIsolationGroups(def) < 0)
        goto cleanup;

2328
    if (nbuses > 0) {
2329 2330 2331
        /* 1st pass to figure out how many PCI bridges we need */
        if (!(addrs = qemuDomainPCIAddressSetCreate(def, nbuses, true)))
            goto cleanup;
2332

2333 2334 2335
        if (qemuDomainValidateDevicePCISlotsChipsets(def, qemuCaps,
                                                     addrs) < 0)
            goto cleanup;
2336

2337 2338 2339 2340 2341 2342 2343 2344
        /* 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)
2345
         *
2346 2347 2348 2349
         * 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.
2350
         */
2351

2352
        if (qemuDomainHasPCIRoot(def)) {
2353 2354 2355 2356 2357 2358 2359 2360 2361 2362 2363 2364
            /* 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++) {
2365
                if (!virDomainPCIAddressBusIsFullyReserved(&addrs->buses[i])) {
2366 2367 2368 2369 2370
                    buses_reserved = false;
                    break;
                }
            }
            if (!buses_reserved &&
2371
                qemuDomainPCIAddressReserveNextAddr(addrs, &info) < 0)
2372 2373
                goto cleanup;
        }
2374

2375 2376
        if (qemuDomainAssignDevicePCISlots(def, qemuCaps, addrs) < 0)
            goto cleanup;
2377

2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393
        /* 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 &&
2394
            qemuDomainHasPCIeRoot(def)) {
2395 2396 2397 2398 2399 2400 2401 2402
            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
             */
2403
            if (qemuDomainPCIAddressReserveNextAddr(addrs, &info) < 0)
2404 2405 2406 2407 2408 2409
               goto cleanup;
        }

        /* now reflect any controllers auto-added to addrs into the
         * domain controllers list
         */
2410
        for (i = 1; i < addrs->nbuses; i++) {
2411 2412
            virDomainDeviceDef dev;
            int contIndex;
2413
            virDomainPCIAddressBusPtr bus = &addrs->buses[i];
2414

2415 2416 2417 2418
            if ((rv = virDomainDefMaybeAddController(
                     def, VIR_DOMAIN_CONTROLLER_TYPE_PCI,
                     i, bus->model)) < 0)
                goto cleanup;
2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441

            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 */
2442
            qemuDomainFillDevicePCIConnectFlags(def, &dev, qemuCaps, driver);
2443 2444 2445 2446 2447 2448 2449

            /* Reserve an address for the controller. pci-root and pcie-root
             * controllers don't plug into any other PCI controller, hence
             * they should skip this step */
            if (bus->model != VIR_DOMAIN_CONTROLLER_MODEL_PCI_ROOT &&
                bus->model != VIR_DOMAIN_CONTROLLER_MODEL_PCIE_ROOT &&
                qemuDomainPCIAddressReserveNextAddr(addrs,
2450
                                                    &dev.data.controller->info) < 0) {
2451
                goto cleanup;
2452
            }
2453
        }
2454

2455 2456 2457 2458
        nbuses = addrs->nbuses;
        virDomainPCIAddressSetFree(addrs);
        addrs = NULL;
    }
2459

2460 2461
    if (!(addrs = qemuDomainPCIAddressSetCreate(def, nbuses, false)))
        goto cleanup;
2462

2463 2464 2465 2466
    if (qemuDomainSupportsPCI(def, qemuCaps)) {
        if (qemuDomainValidateDevicePCISlotsChipsets(def, qemuCaps,
                                                     addrs) < 0)
            goto cleanup;
2467

2468 2469
        if (qemuDomainAssignDevicePCISlots(def, qemuCaps, addrs) < 0)
            goto cleanup;
2470

2471 2472 2473 2474 2475
        /* set multi attribute for devices at function 0 of
         * any slot that has multiple functions in use
         */
        virDomainPCIAddressSetAllMulti(def);

2476 2477 2478 2479 2480
        for (i = 0; i < def->ncontrollers; i++) {
            virDomainControllerDefPtr cont = def->controllers[i];
            int idx = cont->idx;
            virPCIDeviceAddressPtr addr;
            virDomainPCIControllerOptsPtr options;
2481

2482 2483
            if (cont->type != VIR_DOMAIN_CONTROLLER_TYPE_PCI)
                continue;
2484

2485 2486
            addr = &cont->info.addr.pci;
            options = &cont->opts.pciopts;
2487

2488 2489 2490 2491
            /* set default model name (the actual name of the
             * device in qemu) for any controller that doesn't yet
             * have it set.
             */
2492
            qemuDomainPCIControllerSetDefaultModelName(cont, def, qemuCaps);
2493

2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519
            /* 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) {
2520
                    virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
2521 2522 2523 2524 2525
                                   _("No free busNr lower than current "
                                     "lowest busNr is available to "
                                     "auto-assign to bus %d. Must be "
                                     "manually assigned"),
                                   addr->bus);
2526 2527
                    goto cleanup;
                }
2528
                break;
2529 2530 2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 2552
            case VIR_DOMAIN_CONTROLLER_MODEL_PCI_ROOT:
                if (!qemuDomainIsPSeries(def))
                    break;
                if (options->targetIndex == -1) {
                    if (cont->idx == 0) {
                        /* The pci-root controller with controller index 0
                         * must always be assigned target index 0, because
                         * it represents the implicit PHB which is treated
                         * differently than all other PHBs */
                        options->targetIndex = 0;
                    } else {
                        /* For all other PHBs the target index doesn't need
                         * to match the controller index or have any
                         * particular value, really */
                        options->targetIndex = qemuDomainAddressFindNewTargetIndex(def);
                    }
                }
                if (options->targetIndex == -1) {
                    virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
                                   _("No usable target index found for %d"),
                                   addr->bus);
                    goto cleanup;
                }
                break;
2553 2554 2555 2556 2557 2558 2559
            case VIR_DOMAIN_CONTROLLER_MODEL_DMI_TO_PCI_BRIDGE:
            case VIR_DOMAIN_CONTROLLER_MODEL_PCIE_SWITCH_UPSTREAM_PORT:
            case VIR_DOMAIN_CONTROLLER_MODEL_PCIE_ROOT:
            case VIR_DOMAIN_CONTROLLER_MODEL_PCI_LAST:
                break;
            }

2560
            /* check if every PCI bridge controller's index is larger than
2561 2562 2563 2564 2565 2566
             * 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 "
2567 2568
                                 "bus='0x%02x', but index must be "
                                 "larger than bus"),
2569 2570
                               idx, idx, addr->bus);
                goto cleanup;
2571 2572 2573 2574 2575 2576
            }
        }
    }

    if (obj && obj->privateData) {
        priv = obj->privateData;
J
John Ferlan 已提交
2577 2578 2579
        /* if this is the live domain object, we persist the PCI addresses */
        priv->pciaddrs = addrs;
        addrs = NULL;
2580 2581 2582 2583 2584 2585 2586 2587 2588 2589 2590
    }

    ret = 0;

 cleanup:
    virDomainPCIAddressSetFree(addrs);

    return ret;
}


J
Ján Tomko 已提交
2591 2592 2593 2594 2595 2596 2597 2598 2599 2600 2601 2602
struct qemuAssignUSBIteratorInfo {
    virDomainUSBAddressSetPtr addrs;
    size_t count;
};


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

2603 2604 2605 2606 2607 2608
    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
2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 2624 2625 2626 2627 2628 2629 2630 2631 2632 2633 2634 2635 2636 2637 2638 2639 2640 2641 2642 2643 2644 2645 2646 2647 2648 2649 2650 2651 2652
        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);
}


2653 2654 2655 2656 2657 2658 2659 2660 2661 2662 2663 2664 2665 2666 2667 2668 2669
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;
2670
    size_t hubs_needed = 0;
2671
    int ret = -1;
2672
    size_t i;
2673 2674 2675 2676 2677 2678 2679

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

2680 2681 2682 2683 2684 2685 2686 2687
    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++) {
2688 2689 2690 2691 2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702
        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;
}


2703 2704 2705 2706 2707 2708 2709 2710 2711 2712 2713 2714 2715 2716 2717 2718 2719 2720 2721 2722 2723 2724 2725 2726 2727 2728 2729 2730 2731 2732 2733 2734
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
2735
                       _("failed to find an empty memory slot"));
2736 2737 2738 2739 2740 2741 2742 2743 2744 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766 2767 2768 2769 2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790
        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
2791 2792
static int
qemuDomainAssignUSBAddresses(virDomainDefPtr def,
2793 2794
                             virDomainObjPtr obj,
                             bool newDomain)
J
Ján Tomko 已提交
2795 2796 2797 2798 2799
{
    int ret = -1;
    virDomainUSBAddressSetPtr addrs = NULL;
    qemuDomainObjPrivatePtr priv = NULL;

2800 2801 2802 2803 2804 2805 2806 2807 2808 2809 2810
    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
2811 2812 2813
    if (!(addrs = virDomainUSBAddressSetCreate()))
        goto cleanup;

2814 2815 2816
    if (qemuDomainUSBAddressAddHubs(def) < 0)
        goto cleanup;

J
Ján Tomko 已提交
2817 2818 2819 2820 2821 2822 2823 2824 2825
    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
2826 2827 2828 2829 2830 2831 2832 2833
    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
2834 2835 2836 2837 2838 2839 2840 2841 2842 2843 2844 2845 2846
    if (obj && obj->privateData) {
        priv = obj->privateData;
        priv->usbaddrs = addrs;
        addrs = NULL;
    }
    ret = 0;

 cleanup:
    virDomainUSBAddressSetFree(addrs);
    return ret;
}


2847 2848 2849
int
qemuDomainAssignAddresses(virDomainDefPtr def,
                          virQEMUCapsPtr qemuCaps,
2850
                          virQEMUDriverPtr driver,
2851
                          virDomainObjPtr obj,
J
Ján Tomko 已提交
2852
                          bool newDomain)
2853
{
2854
    if (qemuDomainAssignVirtioSerialAddresses(def) < 0)
2855
        return -1;
2856

2857
    if (qemuDomainAssignSpaprVIOAddresses(def) < 0)
2858
        return -1;
2859

T
Tomasz Flendrich 已提交
2860
    if (qemuDomainAssignS390Addresses(def, qemuCaps) < 0)
2861
        return -1;
2862

2863
    qemuDomainAssignARMVirtioMMIOAddresses(def, qemuCaps);
2864

2865
    if (qemuDomainAssignPCIAddresses(def, qemuCaps, driver, obj) < 0)
2866 2867
        return -1;

2868
    if (qemuDomainAssignUSBAddresses(def, obj, newDomain) < 0)
J
Ján Tomko 已提交
2869 2870
        return -1;

2871 2872 2873
    if (qemuDomainAssignMemorySlots(def) < 0)
        return -1;

2874
    return 0;
2875 2876
}

2877 2878 2879 2880 2881 2882 2883 2884 2885 2886 2887 2888 2889 2890 2891 2892 2893
/**
 * 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,
2894 2895
                           virDomainDeviceDefPtr dev,
                           virQEMUDriverPtr driver)
2896 2897 2898 2899 2900 2901 2902
{
    qemuDomainObjPrivatePtr priv = obj->privateData;
    virDomainDeviceInfoPtr info = virDomainDeviceGetInfo(dev);

    if (!info)
        return 0;

2903
    qemuDomainFillDevicePCIConnectFlags(obj->def, dev, priv->qemuCaps, driver);
2904 2905 2906 2907

    return virDomainPCIAddressEnsureAddr(priv->pciaddrs, info,
                                         info->pciConnectFlags);
}
2908 2909 2910 2911 2912 2913 2914 2915 2916 2917 2918

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

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

2919 2920
    if (virDeviceInfoPCIAddressPresent(info))
        virDomainPCIAddressReleaseAddr(priv->pciaddrs, &info->addr.pci);
2921

2922 2923 2924 2925 2926
    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));
2927
}
2928 2929 2930 2931 2932 2933 2934 2935 2936 2937 2938 2939 2940 2941 2942 2943 2944 2945 2946 2947 2948 2949 2950 2951 2952 2953 2954 2955 2956 2957 2958 2959 2960 2961 2962 2963 2964 2965 2966 2967 2968 2969 2970 2971 2972


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

    if (!info->type) {
        if (qemuDomainIsS390CCW(vm->def) &&
            virQEMUCapsGet(priv->qemuCaps, QEMU_CAPS_VIRTIO_CCW))
            info->type = VIR_DOMAIN_DEVICE_ADDRESS_TYPE_CCW;
        else if (virQEMUCapsGet(priv->qemuCaps, QEMU_CAPS_VIRTIO_S390))
            info->type = VIR_DOMAIN_DEVICE_ADDRESS_TYPE_VIRTIO_S390;
    } else {
        if (!qemuDomainCheckCCWS390AddressSupport(vm->def, *info, priv->qemuCaps,
                                                  devicename))
            return -1;
    }

    if (info->type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_CCW) {
        if (!(ccwaddrs = qemuDomainCCWAddrSetCreateFromDomain(vm->def)))
            goto cleanup;
        if (virDomainCCWAddressAssign(info, ccwaddrs,
                                      !info->addr.ccw.assigned) < 0)
            goto cleanup;
    } else if (!info->type ||
               info->type == VIR_DOMAIN_DEVICE_ADDRESS_TYPE_PCI) {
        if (qemuDomainEnsurePCIAddress(vm, dev, driver) < 0)
            goto cleanup;
        *releaseAddr = true;
    }

    ret = 0;

 cleanup:
    virDomainCCWAddressSetFree(ccwaddrs);
    return ret;
}