qemu_capabilities.c 34.9 KB
Newer Older
1 2 3
/*
 * qemu_capabilities.c: QEMU capabilities generation
 *
4
 * Copyright (C) 2006-2011 Red Hat, Inc.
5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35
 * Copyright (C) 2006 Daniel P. Berrange
 *
 * This library is free software; you can redistribute it and/or
 * modify it under the terms of the GNU Lesser General Public
 * License as published by the Free Software Foundation; either
 * version 2.1 of the License, or (at your option) any later version.
 *
 * This library is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
 * Lesser General Public License for more details.
 *
 * You should have received a copy of the GNU Lesser General Public
 * License along with this library; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307  USA
 *
 * Author: Daniel P. Berrange <berrange@redhat.com>
 */

#include <config.h>

#include "qemu_capabilities.h"
#include "memory.h"
#include "logging.h"
#include "virterror_internal.h"
#include "util.h"
#include "files.h"
#include "nodeinfo.h"
#include "cpu/cpu.h"
#include "domain_conf.h"
#include "qemu_conf.h"
36
#include "command.h"
37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171

#include <sys/stat.h>
#include <unistd.h>
#include <sys/wait.h>
#include <sys/utsname.h>

#define VIR_FROM_THIS VIR_FROM_QEMU

struct qemu_feature_flags {
    const char *name;
    const int default_on;
    const int toggle;
};

struct qemu_arch_info {
    const char *arch;
    int wordsize;
    const char *machine;
    const char *binary;
    const char *altbinary;
    const struct qemu_feature_flags *flags;
    int nflags;
};

/* Feature flags for the architecture info */
static const struct qemu_feature_flags const arch_info_i686_flags [] = {
    { "pae",  1, 0 },
    { "nonpae",  1, 0 },
    { "acpi", 1, 1 },
    { "apic", 1, 0 },
};

static const struct qemu_feature_flags const arch_info_x86_64_flags [] = {
    { "acpi", 1, 1 },
    { "apic", 1, 0 },
};

/* The archicture tables for supported QEMU archs */
static const struct qemu_arch_info const arch_info_hvm[] = {
    {  "i686",   32, NULL, "qemu",
       "qemu-system-x86_64", arch_info_i686_flags, 4 },
    {  "x86_64", 64, NULL, "qemu-system-x86_64",
       NULL, arch_info_x86_64_flags, 2 },
    {  "arm",    32, NULL, "qemu-system-arm",    NULL, NULL, 0 },
    {  "mips",   32, NULL, "qemu-system-mips",   NULL, NULL, 0 },
    {  "mipsel", 32, NULL, "qemu-system-mipsel", NULL, NULL, 0 },
    {  "sparc",  32, NULL, "qemu-system-sparc",  NULL, NULL, 0 },
    {  "ppc",    32, NULL, "qemu-system-ppc",    NULL, NULL, 0 },
    {  "itanium", 64, NULL, "qemu-system-ia64",  NULL, NULL, 0 },
    {  "s390x",  64, NULL, "qemu-system-s390x",  NULL, NULL, 0 },
};

static const struct qemu_arch_info const arch_info_xen[] = {
    {  "i686",   32, "xenner", "xenner", NULL, arch_info_i686_flags, 4 },
    {  "x86_64", 64, "xenner", "xenner", NULL, arch_info_x86_64_flags, 2 },
};


/* Format is:
 * <machine> <desc> [(default)|(alias of <canonical>)]
 */
static int
qemuCapsParseMachineTypesStr(const char *output,
                             virCapsGuestMachinePtr **machines,
                             int *nmachines)
{
    const char *p = output;
    const char *next;
    virCapsGuestMachinePtr *list = NULL;
    int nitems = 0;

    do {
        const char *t;
        virCapsGuestMachinePtr machine;

        if ((next = strchr(p, '\n')))
            ++next;

        if (STRPREFIX(p, "Supported machines are:"))
            continue;

        if (!(t = strchr(p, ' ')) || (next && t >= next))
            continue;

        if (VIR_ALLOC(machine) < 0)
            goto no_memory;

        if (!(machine->name = strndup(p, t - p))) {
            VIR_FREE(machine);
            goto no_memory;
        }

        if (VIR_REALLOC_N(list, nitems + 1) < 0) {
            VIR_FREE(machine->name);
            VIR_FREE(machine);
            goto no_memory;
        }

        p = t;
        if (!(t = strstr(p, "(default)")) || (next && t >= next)) {
            list[nitems++] = machine;
        } else {
            /* put the default first in the list */
            memmove(list + 1, list, sizeof(*list) * nitems);
            list[0] = machine;
            nitems++;
        }

        if ((t = strstr(p, "(alias of ")) && (!next || t < next)) {
            p = t + strlen("(alias of ");
            if (!(t = strchr(p, ')')) || (next && t >= next))
                continue;

            if (!(machine->canonical = strndup(p, t - p)))
                goto no_memory;
        }
    } while ((p = next));

    *machines = list;
    *nmachines = nitems;

    return 0;

  no_memory:
    virReportOOMError();
    virCapabilitiesFreeMachines(list, nitems);
    return -1;
}

int
qemuCapsProbeMachineTypes(const char *binary,
                          virCapsGuestMachinePtr **machines,
                          int *nmachines)
{
    char *output;
172 173
    int ret = -1;
    virCommandPtr cmd;
174

175 176 177 178 179 180 181 182 183
    /* Make sure the binary we are about to try exec'ing exists.
     * Technically we could catch the exec() failure, but that's
     * in a sub-process so it's hard to feed back a useful error.
     */
    if (access(binary, X_OK) < 0) {
        virReportSystemError(errno, _("Cannot find QEMU binary %s"), binary);
        return -1;
    }

184 185 186 187
    cmd = virCommandNewArgList(binary, "-M", "?", NULL);
    virCommandAddEnvPassCommon(cmd);
    virCommandSetOutputBuffer(cmd, &output);
    virCommandClearCaps(cmd);
188

189
    if (virCommandRun(cmd, NULL) < 0)
190 191 192
        goto cleanup;

    if (qemuCapsParseMachineTypesStr(output, machines, nmachines) < 0)
193
        goto cleanup;
194 195 196 197

    ret = 0;

cleanup:
198 199
    VIR_FREE(output);
    virCommandFree(cmd);
200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385

    return ret;
}

static int
qemuCapsGetOldMachinesFromInfo(virCapsGuestDomainInfoPtr info,
                               const char *emulator,
                               time_t emulator_mtime,
                               virCapsGuestMachinePtr **machines,
                               int *nmachines)
{
    virCapsGuestMachinePtr *list;
    int i;

    if (!info->nmachines)
        return 0;

    if (!info->emulator || !STREQ(emulator, info->emulator))
        return 0;

    if (emulator_mtime != info->emulator_mtime) {
        VIR_DEBUG("mtime on %s has changed, refreshing machine types",
                  info->emulator);
        return 0;
    }

    if (VIR_ALLOC_N(list, info->nmachines) < 0) {
        virReportOOMError();
        return 0;
    }

    for (i = 0; i < info->nmachines; i++) {
        if (VIR_ALLOC(list[i]) < 0) {
            goto no_memory;
        }
        if (info->machines[i]->name &&
            !(list[i]->name = strdup(info->machines[i]->name))) {
            goto no_memory;
        }
        if (info->machines[i]->canonical &&
            !(list[i]->canonical = strdup(info->machines[i]->canonical))) {
            goto no_memory;
        }
    }

    *machines = list;
    *nmachines = info->nmachines;

    return 1;

  no_memory:
    virReportOOMError();
    virCapabilitiesFreeMachines(list, info->nmachines);
    return 0;
}

static int
qemuCapsGetOldMachines(const char *ostype,
                       const char *arch,
                       int wordsize,
                       const char *emulator,
                       time_t emulator_mtime,
                       virCapsPtr old_caps,
                       virCapsGuestMachinePtr **machines,
                       int *nmachines)
{
    int i;

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

        if (!STREQ(ostype, guest->ostype) ||
            !STREQ(arch, guest->arch.name) ||
            wordsize != guest->arch.wordsize)
            continue;

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

            if (qemuCapsGetOldMachinesFromInfo(&dom->info,
                                               emulator, emulator_mtime,
                                               machines, nmachines))
                return 1;
        }

        if (qemuCapsGetOldMachinesFromInfo(&guest->arch.defaultInfo,
                                           emulator, emulator_mtime,
                                           machines, nmachines))
            return 1;
    }

    return 0;
}


typedef int
(*qemuCapsParseCPUModels)(const char *output,
                       unsigned int *retcount,
                       const char ***retcpus);

/* Format:
 *      <arch> <model>
 * qemu-0.13 encloses some model names in []:
 *      <arch> [<model>]
 */
static int
qemuCapsParseX86Models(const char *output,
                       unsigned int *retcount,
                       const char ***retcpus)
{
    const char *p = output;
    const char *next;
    unsigned int count = 0;
    const char **cpus = NULL;
    int i;

    do {
        const char *t;

        if ((next = strchr(p, '\n')))
            next++;

        if (!(t = strchr(p, ' ')) || (next && t >= next))
            continue;

        if (!STRPREFIX(p, "x86"))
            continue;

        p = t;
        while (*p == ' ')
            p++;

        if (*p == '\0' || *p == '\n')
            continue;

        if (retcpus) {
            unsigned int len;

            if (VIR_REALLOC_N(cpus, count + 1) < 0)
                goto error;

            if (next)
                len = next - p - 1;
            else
                len = strlen(p);

            if (len > 2 && *p == '[' && p[len - 1] == ']') {
                p++;
                len -= 2;
            }

            if (!(cpus[count] = strndup(p, len)))
                goto error;
        }
        count++;
    } while ((p = next));

    if (retcount)
        *retcount = count;
    if (retcpus)
        *retcpus = cpus;

    return 0;

error:
    if (cpus) {
        for (i = 0; i < count; i++)
            VIR_FREE(cpus[i]);
    }
    VIR_FREE(cpus);

    return -1;
}


int
qemuCapsProbeCPUModels(const char *qemu,
                       unsigned long long qemuCmdFlags,
                       const char *arch,
                       unsigned int *count,
                       const char ***cpus)
{
    char *output = NULL;
    int ret = -1;
    qemuCapsParseCPUModels parse;
386
    virCommandPtr cmd;
387 388 389 390 391 392 393 394 395 396 397 398 399

    if (count)
        *count = 0;
    if (cpus)
        *cpus = NULL;

    if (STREQ(arch, "i686") || STREQ(arch, "x86_64"))
        parse = qemuCapsParseX86Models;
    else {
        VIR_DEBUG("don't know how to parse %s CPU models", arch);
        return 0;
    }

400 401 402 403 404 405
    cmd = virCommandNewArgList(qemu, "-cpu", "?", NULL);
    if (qemuCmdFlags & QEMUD_CMD_FLAG_NODEFCONFIG)
        virCommandAddArg(cmd, "-nodefconfig");
    virCommandAddEnvPassCommon(cmd);
    virCommandSetOutputBuffer(cmd, &output);
    virCommandClearCaps(cmd);
406

407
    if (virCommandRun(cmd, NULL) < 0)
408 409 410 411 412 413 414 415 416 417 418
        goto cleanup;

    if (parse(output, count, cpus) < 0) {
        virReportOOMError();
        goto cleanup;
    }

    ret = 0;

cleanup:
    VIR_FREE(output);
419
    virCommandFree(cmd);
420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442

    return ret;
}


static int
qemuCapsInitGuest(virCapsPtr caps,
                  virCapsPtr old_caps,
                  const char *hostmachine,
                  const struct qemu_arch_info *info,
                  int hvm)
{
    virCapsGuestPtr guest;
    int i;
    int haskvm = 0;
    int haskqemu = 0;
    char *kvmbin = NULL;
    char *binary = NULL;
    time_t binary_mtime;
    virCapsGuestMachinePtr *machines = NULL;
    int nmachines = 0;
    struct stat st;
    unsigned int ncpus;
443
    unsigned long long qemuCmdFlags;
444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558
    int ret = -1;

    /* Check for existance of base emulator, or alternate base
     * which can be used with magic cpu choice
     */
    binary = virFindFileInPath(info->binary);

    if (binary == NULL || access(binary, X_OK) != 0) {
        VIR_FREE(binary);
        binary = virFindFileInPath(info->altbinary);

        if (binary != NULL && access(binary, X_OK) != 0)
            VIR_FREE(binary);
    }

    /* Can use acceleration for KVM/KQEMU if
     *  - host & guest arches match
     * Or
     *  - hostarch is x86_64 and guest arch is i686
     * The latter simply needs "-cpu qemu32"
     */
    if (STREQ(info->arch, hostmachine) ||
        (STREQ(hostmachine, "x86_64") && STREQ(info->arch, "i686"))) {
        if (access("/dev/kvm", F_OK) == 0) {
            const char *const kvmbins[] = { "/usr/libexec/qemu-kvm", /* RHEL */
                                            "qemu-kvm", /* Fedora */
                                            "kvm" }; /* Upstream .spec */

            for (i = 0; i < ARRAY_CARDINALITY(kvmbins); ++i) {
                kvmbin = virFindFileInPath(kvmbins[i]);

                if (kvmbin == NULL || access(kvmbin, X_OK) != 0) {
                    VIR_FREE(kvmbin);
                    continue;
                }

                haskvm = 1;
                if (!binary)
                    binary = kvmbin;

                break;
            }
        }

        if (access("/dev/kqemu", F_OK) == 0)
            haskqemu = 1;
    }

    if (!binary)
        return 0;

    if (stat(binary, &st) == 0) {
        binary_mtime = st.st_mtime;
    } else {
        char ebuf[1024];
        VIR_WARN("Failed to stat %s, most peculiar : %s",
                 binary, virStrerror(errno, ebuf, sizeof(ebuf)));
        binary_mtime = 0;
    }

    if (info->machine) {
        virCapsGuestMachinePtr machine;

        if (VIR_ALLOC(machine) < 0) {
            goto no_memory;
        }

        if (!(machine->name = strdup(info->machine))) {
            VIR_FREE(machine);
            goto no_memory;
        }

        nmachines = 1;

        if (VIR_ALLOC_N(machines, nmachines) < 0) {
            VIR_FREE(machine->name);
            VIR_FREE(machine);
            goto no_memory;
        }

        machines[0] = machine;
    } else {
        int probe = 1;
        if (old_caps && binary_mtime)
            probe = !qemuCapsGetOldMachines(hvm ? "hvm" : "xen", info->arch,
                                            info->wordsize, binary, binary_mtime,
                                            old_caps, &machines, &nmachines);
        if (probe &&
            qemuCapsProbeMachineTypes(binary, &machines, &nmachines) < 0)
            goto error;
    }

    /* We register kvm as the base emulator too, since we can
     * just give -no-kvm to disable acceleration if required */
    if ((guest = virCapabilitiesAddGuest(caps,
                                         hvm ? "hvm" : "xen",
                                         info->arch,
                                         info->wordsize,
                                         binary,
                                         NULL,
                                         nmachines,
                                         machines)) == NULL)
        goto error;

    machines = NULL;
    nmachines = 0;

    guest->arch.defaultInfo.emulator_mtime = binary_mtime;

    if (caps->host.cpu &&
        qemuCapsProbeCPUModels(binary, 0, info->arch, &ncpus, NULL) == 0 &&
        ncpus > 0 &&
        !virCapabilitiesAddGuestFeature(guest, "cpuselection", 1, 0))
        goto error;

559 560 561 562 563
    if (qemuCapsExtractVersionInfo(binary, NULL, &qemuCmdFlags) < 0 ||
        ((qemuCmdFlags & QEMUD_CMD_FLAG_BOOTINDEX) &&
         !virCapabilitiesAddGuestFeature(guest, "deviceboot", 1, 0)))
        goto error;

564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820
    if (hvm) {
        if (virCapabilitiesAddGuestDomain(guest,
                                          "qemu",
                                          NULL,
                                          NULL,
                                          0,
                                          NULL) == NULL)
            goto error;

        if (haskqemu &&
            virCapabilitiesAddGuestDomain(guest,
                                          "kqemu",
                                          NULL,
                                          NULL,
                                          0,
                                          NULL) == NULL)
            goto error;

        if (haskvm) {
            virCapsGuestDomainPtr dom;

            if (stat(kvmbin, &st) == 0) {
                binary_mtime = st.st_mtime;
            } else {
                char ebuf[1024];
                VIR_WARN("Failed to stat %s, most peculiar : %s",
                         binary, virStrerror(errno, ebuf, sizeof(ebuf)));
                binary_mtime = 0;
            }

            if (!STREQ(binary, kvmbin)) {
                int probe = 1;
                if (old_caps && binary_mtime)
                    probe = !qemuCapsGetOldMachines("hvm", info->arch, info->wordsize,
                                                    kvmbin, binary_mtime,
                                                    old_caps, &machines, &nmachines);
                if (probe &&
                    qemuCapsProbeMachineTypes(kvmbin, &machines, &nmachines) < 0)
                    goto error;
            }

            if ((dom = virCapabilitiesAddGuestDomain(guest,
                                                     "kvm",
                                                     kvmbin,
                                                     NULL,
                                                     nmachines,
                                                     machines)) == NULL) {
                goto error;
            }

            machines = NULL;
            nmachines = 0;

            dom->info.emulator_mtime = binary_mtime;
        }
    } else {
        if (virCapabilitiesAddGuestDomain(guest,
                                          "kvm",
                                          NULL,
                                          NULL,
                                          0,
                                          NULL) == NULL)
            goto error;
    }

    if (info->nflags) {
        for (i = 0 ; i < info->nflags ; i++) {
            if (virCapabilitiesAddGuestFeature(guest,
                                               info->flags[i].name,
                                               info->flags[i].default_on,
                                               info->flags[i].toggle) == NULL)
                goto error;
        }
    }

    ret = 0;

cleanup:
    if (binary == kvmbin) {
        /* don't double free */
        VIR_FREE(binary);
    } else {
        VIR_FREE(binary);
        VIR_FREE(kvmbin);
    }

    return ret;

no_memory:
    virReportOOMError();

error:
    virCapabilitiesFreeMachines(machines, nmachines);

    goto cleanup;
}


static int
qemuCapsInitCPU(virCapsPtr caps,
                 const char *arch)
{
    virCPUDefPtr cpu = NULL;
    union cpuData *data = NULL;
    virNodeInfo nodeinfo;
    int ret = -1;

    if (VIR_ALLOC(cpu) < 0
        || !(cpu->arch = strdup(arch))) {
        virReportOOMError();
        goto error;
    }

    if (nodeGetInfo(NULL, &nodeinfo))
        goto error;

    cpu->type = VIR_CPU_TYPE_HOST;
    cpu->sockets = nodeinfo.sockets;
    cpu->cores = nodeinfo.cores;
    cpu->threads = nodeinfo.threads;

    if (!(data = cpuNodeData(arch))
        || cpuDecode(cpu, data, NULL, 0, NULL) < 0)
        goto error;

    caps->host.cpu = cpu;

    ret = 0;

cleanup:
    cpuDataFree(arch, data);

    return ret;

error:
    virCPUDefFree(cpu);
    goto cleanup;
}


virCapsPtr qemuCapsInit(virCapsPtr old_caps)
{
    struct utsname utsname;
    virCapsPtr caps;
    int i;
    char *xenner = NULL;

    /* Really, this never fails - look at the man-page. */
    uname (&utsname);

    if ((caps = virCapabilitiesNew(utsname.machine,
                                   1, 1)) == NULL)
        goto no_memory;

    /* Using KVM's mac prefix for QEMU too */
    virCapabilitiesSetMacPrefix(caps, (unsigned char[]){ 0x52, 0x54, 0x00 });

    /* Some machines have problematic NUMA toplogy causing
     * unexpected failures. We don't want to break the QEMU
     * driver in this scenario, so log errors & carry on
     */
    if (nodeCapsInitNUMA(caps) < 0) {
        virCapabilitiesFreeNUMAInfo(caps);
        VIR_WARN0("Failed to query host NUMA topology, disabling NUMA capabilities");
    }

    if (old_caps == NULL || old_caps->host.cpu == NULL) {
        if (qemuCapsInitCPU(caps, utsname.machine) < 0)
            VIR_WARN0("Failed to get host CPU");
    }
    else {
        caps->host.cpu = old_caps->host.cpu;
        old_caps->host.cpu = NULL;
    }

    virCapabilitiesAddHostMigrateTransport(caps,
                                           "tcp");

    /* First the pure HVM guests */
    for (i = 0 ; i < ARRAY_CARDINALITY(arch_info_hvm) ; i++)
        if (qemuCapsInitGuest(caps, old_caps,
                              utsname.machine,
                              &arch_info_hvm[i], 1) < 0)
            goto no_memory;

    /* Then possibly the Xen paravirt guests (ie Xenner */
    xenner = virFindFileInPath("xenner");

    if (xenner != NULL && access(xenner, X_OK) == 0 &&
        access("/dev/kvm", F_OK) == 0) {
        for (i = 0 ; i < ARRAY_CARDINALITY(arch_info_xen) ; i++)
            /* Allow Xen 32-on-32, 32-on-64 and 64-on-64 */
            if (STREQ(arch_info_xen[i].arch, utsname.machine) ||
                (STREQ(utsname.machine, "x86_64") &&
                 STREQ(arch_info_xen[i].arch, "i686"))) {
                if (qemuCapsInitGuest(caps, old_caps,
                                      utsname.machine,
                                      &arch_info_xen[i], 0) < 0)
                    goto no_memory;
            }
    }

    VIR_FREE(xenner);

    /* QEMU Requires an emulator in the XML */
    virCapabilitiesSetEmulatorRequired(caps);

    caps->defaultConsoleTargetType = VIR_DOMAIN_CHR_CONSOLE_TARGET_TYPE_SERIAL;

    return caps;

 no_memory:
    VIR_FREE(xenner);
    virCapabilitiesFree(caps);
    return NULL;
}


static unsigned long long
qemuCapsComputeCmdFlags(const char *help,
                        unsigned int version,
                        unsigned int is_kvm,
                        unsigned int kvm_version)
{
    unsigned long long flags = 0;
    const char *p;

    if (strstr(help, "-no-kqemu"))
        flags |= QEMUD_CMD_FLAG_KQEMU;
    if (strstr(help, "-enable-kqemu"))
        flags |= QEMUD_CMD_FLAG_ENABLE_KQEMU;
    if (strstr(help, "-no-kvm"))
        flags |= QEMUD_CMD_FLAG_KVM;
    if (strstr(help, "-enable-kvm"))
        flags |= QEMUD_CMD_FLAG_ENABLE_KVM;
    if (strstr(help, "-no-reboot"))
        flags |= QEMUD_CMD_FLAG_NO_REBOOT;
    if (strstr(help, "-name")) {
        flags |= QEMUD_CMD_FLAG_NAME;
        if (strstr(help, ",process="))
            flags |= QEMUD_CMD_FLAG_NAME_PROCESS;
    }
    if (strstr(help, "-uuid"))
        flags |= QEMUD_CMD_FLAG_UUID;
    if (strstr(help, "-xen-domid"))
        flags |= QEMUD_CMD_FLAG_XEN_DOMID;
    else if (strstr(help, "-domid"))
        flags |= QEMUD_CMD_FLAG_DOMID;
    if (strstr(help, "-drive")) {
        flags |= QEMUD_CMD_FLAG_DRIVE;
        if (strstr(help, "cache=") &&
            !strstr(help, "cache=on|off"))
            flags |= QEMUD_CMD_FLAG_DRIVE_CACHE_V2;
        if (strstr(help, "format="))
            flags |= QEMUD_CMD_FLAG_DRIVE_FORMAT;
        if (strstr(help, "readonly="))
            flags |= QEMUD_CMD_FLAG_DRIVE_READONLY;
821 822
        if (strstr(help, "aio=threads|native"))
            flags |= QEMUD_CMD_FLAG_DRIVE_AIO;
823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843
    }
    if ((p = strstr(help, "-vga")) && !strstr(help, "-std-vga")) {
        const char *nl = strstr(p, "\n");

        flags |= QEMUD_CMD_FLAG_VGA;

        if (strstr(p, "|qxl"))
            flags |= QEMUD_CMD_FLAG_VGA_QXL;
        if ((p = strstr(p, "|none")) && p < nl)
            flags |= QEMUD_CMD_FLAG_VGA_NONE;
    }
    if (strstr(help, "-spice"))
        flags |= QEMUD_CMD_FLAG_SPICE;
    if (strstr(help, "boot=on"))
        flags |= QEMUD_CMD_FLAG_DRIVE_BOOT;
    if (strstr(help, "serial=s"))
        flags |= QEMUD_CMD_FLAG_DRIVE_SERIAL;
    if (strstr(help, "-pcidevice"))
        flags |= QEMUD_CMD_FLAG_PCIDEVICE;
    if (strstr(help, "-mem-path"))
        flags |= QEMUD_CMD_FLAG_MEM_PATH;
844
    if (strstr(help, "-chardev")) {
845
        flags |= QEMUD_CMD_FLAG_CHARDEV;
846 847 848
        if (strstr(help, "-chardev spicevmc"))
            flags |= QEMUD_CMD_FLAG_CHARDEV_SPICEVMC;
    }
849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910
    if (strstr(help, "-balloon"))
        flags |= QEMUD_CMD_FLAG_BALLOON;
    if (strstr(help, "-device")) {
        flags |= QEMUD_CMD_FLAG_DEVICE;
        /*
         * When -device was introduced, qemu already supported drive's
         * readonly option but didn't advertise that.
         */
        flags |= QEMUD_CMD_FLAG_DRIVE_READONLY;
    }
    if (strstr(help, "-nodefconfig"))
        flags |= QEMUD_CMD_FLAG_NODEFCONFIG;
    /* The trailing ' ' is important to avoid a bogus match */
    if (strstr(help, "-rtc "))
        flags |= QEMUD_CMD_FLAG_RTC;
    /* to wit */
    if (strstr(help, "-rtc-td-hack"))
        flags |= QEMUD_CMD_FLAG_RTC_TD_HACK;
    if (strstr(help, "-no-hpet"))
        flags |= QEMUD_CMD_FLAG_NO_HPET;
    if (strstr(help, "-no-kvm-pit-reinjection"))
        flags |= QEMUD_CMD_FLAG_NO_KVM_PIT;
    if (strstr(help, "-tdf"))
        flags |= QEMUD_CMD_FLAG_TDF;
    if (strstr(help, "-enable-nesting"))
        flags |= QEMUD_CMD_FLAG_NESTING;
    if (strstr(help, ",menu=on"))
        flags |= QEMUD_CMD_FLAG_BOOT_MENU;
    if (strstr(help, "-fsdev"))
        flags |= QEMUD_CMD_FLAG_FSDEV;
    if (strstr(help, "-smbios type"))
        flags |= QEMUD_CMD_FLAG_SMBIOS_TYPE;

    if (strstr(help, "-netdev")) {
        /* Disable -netdev on 0.12 since although it exists,
         * the corresponding netdev_add/remove monitor commands
         * do not, and we need them to be able todo hotplug */
        if (version >= 13000)
            flags |= QEMUD_CMD_FLAG_NETDEV;
    }

    if (strstr(help, "-sdl"))
        flags |= QEMUD_CMD_FLAG_SDL;
    if (strstr(help, "cores=") &&
        strstr(help, "threads=") &&
        strstr(help, "sockets="))
        flags |= QEMUD_CMD_FLAG_SMP_TOPOLOGY;

    if (version >= 9000)
        flags |= QEMUD_CMD_FLAG_VNC_COLON;

    if (is_kvm && (version >= 10000 || kvm_version >= 74))
        flags |= QEMUD_CMD_FLAG_VNET_HDR;

    if (is_kvm && strstr(help, ",vhost=")) {
        flags |= QEMUD_CMD_FLAG_VNET_HOST;
    }

    /*
     * Handling of -incoming arg with varying features
     *  -incoming tcp    (kvm >= 79, qemu >= 0.10.0)
     *  -incoming exec   (kvm >= 80, qemu >= 0.10.0)
911 912
     *  -incoming unix   (qemu >= 0.12.0)
     *  -incoming fd     (qemu >= 0.12.0)
913 914 915 916 917 918 919 920 921
     *  -incoming stdio  (all earlier kvm)
     *
     * NB, there was a pre-kvm-79 'tcp' support, but it
     * was broken, because it blocked the monitor console
     * while waiting for data, so pretend it doesn't exist
     */
    if (version >= 10000) {
        flags |= QEMUD_CMD_FLAG_MIGRATE_QEMU_TCP;
        flags |= QEMUD_CMD_FLAG_MIGRATE_QEMU_EXEC;
922
        if (version >= 12000) {
923
            flags |= QEMUD_CMD_FLAG_MIGRATE_QEMU_UNIX;
924 925
            flags |= QEMUD_CMD_FLAG_MIGRATE_QEMU_FD;
        }
926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 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 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054
    } else if (kvm_version >= 79) {
        flags |= QEMUD_CMD_FLAG_MIGRATE_QEMU_TCP;
        if (kvm_version >= 80)
            flags |= QEMUD_CMD_FLAG_MIGRATE_QEMU_EXEC;
    } else if (kvm_version > 0) {
        flags |= QEMUD_CMD_FLAG_MIGRATE_KVM_STDIO;
    }

    if (version >= 10000)
        flags |= QEMUD_CMD_FLAG_0_10;

    /* While JSON mode was available in 0.12.0, it was too
     * incomplete to contemplate using. The 0.13.0 release
     * is good enough to use, even though it lacks one or
     * two features. The benefits of JSON mode now outweigh
     * the downside.
     */
     if (version >= 13000)
        flags |= QEMUD_CMD_FLAG_MONITOR_JSON;

    return flags;
}

/* We parse the output of 'qemu -help' to get the QEMU
 * version number. The first bit is easy, just parse
 * 'QEMU PC emulator version x.y.z'
 * or
 * 'QEMU emulator version x.y.z'.
 *
 * With qemu-kvm, however, that is followed by a string
 * in parenthesis as follows:
 *  - qemu-kvm-x.y.z in stable releases
 *  - kvm-XX for kvm versions up to kvm-85
 *  - qemu-kvm-devel-XX for kvm version kvm-86 and later
 *
 * For qemu-kvm versions before 0.10.z, we need to detect
 * the KVM version number for some features. With 0.10.z
 * and later, we just need the QEMU version number and
 * whether it is KVM QEMU or mainline QEMU.
 */
#define QEMU_VERSION_STR_1  "QEMU emulator version"
#define QEMU_VERSION_STR_2  "QEMU PC emulator version"
#define QEMU_KVM_VER_PREFIX "(qemu-kvm-"
#define KVM_VER_PREFIX      "(kvm-"

#define SKIP_BLANKS(p) do { while ((*(p) == ' ') || (*(p) == '\t')) (p)++; } while (0)

int qemuCapsParseHelpStr(const char *qemu,
                         const char *help,
                         unsigned long long *flags,
                         unsigned int *version,
                         unsigned int *is_kvm,
                         unsigned int *kvm_version)
{
    unsigned major, minor, micro;
    const char *p = help;

    *flags = *version = *is_kvm = *kvm_version = 0;

    if (STRPREFIX(p, QEMU_VERSION_STR_1))
        p += strlen(QEMU_VERSION_STR_1);
    else if (STRPREFIX(p, QEMU_VERSION_STR_2))
        p += strlen(QEMU_VERSION_STR_2);
    else
        goto fail;

    SKIP_BLANKS(p);

    major = virParseNumber(&p);
    if (major == -1 || *p != '.')
        goto fail;

    ++p;

    minor = virParseNumber(&p);
    if (minor == -1 || *p != '.')
        goto fail;

    ++p;

    micro = virParseNumber(&p);
    if (micro == -1)
        goto fail;

    SKIP_BLANKS(p);

    if (STRPREFIX(p, QEMU_KVM_VER_PREFIX)) {
        *is_kvm = 1;
        p += strlen(QEMU_KVM_VER_PREFIX);
    } else if (STRPREFIX(p, KVM_VER_PREFIX)) {
        int ret;

        *is_kvm = 1;
        p += strlen(KVM_VER_PREFIX);

        ret = virParseNumber(&p);
        if (ret == -1)
            goto fail;

        *kvm_version = ret;
    }

    *version = (major * 1000 * 1000) + (minor * 1000) + micro;

    *flags = qemuCapsComputeCmdFlags(help, *version, *is_kvm, *kvm_version);

    VIR_DEBUG("Version %u.%u.%u, cooked version %u, flags 0x%llx",
              major, minor, micro, *version, *flags);
    if (*kvm_version)
        VIR_DEBUG("KVM version %d detected", *kvm_version);
    else if (*is_kvm)
        VIR_DEBUG("qemu-kvm version %u.%u.%u detected", major, minor, micro);

    return 0;

fail:
    p = strchr(help, '\n');
    if (p)
        p = strndup(help, p - help);

    qemuReportError(VIR_ERR_INTERNAL_ERROR,
                    _("cannot parse %s version number in '%s'"),
                    qemu, p ? p : help);

    VIR_FREE(p);

    return -1;
}

E
Eric Blake 已提交
1055 1056 1057
static int
qemuCapsExtractDeviceStr(const char *qemu,
                         unsigned long long *flags)
1058
{
E
Eric Blake 已提交
1059
    char *output = NULL;
1060
    virCommandPtr cmd;
E
Eric Blake 已提交
1061
    int ret = -1;
1062

E
Eric Blake 已提交
1063 1064
    /* Cram together all device-related queries into one invocation;
     * the output format makes it possible to distinguish what we
1065 1066
     * need.  With qemu 0.13.0 and later, unrecognized '-device
     * bogus,?' cause an error in isolation, but are silently ignored
1067
     * in combination with '-device ?'.  Upstream qemu 0.12.x doesn't
1068 1069
     * understand '-device name,?', and always exits with status 1 for
     * the simpler '-device ?', so this function is really only useful
1070
     * if -help includes "device driver,?".  */
E
Eric Blake 已提交
1071
    cmd = virCommandNewArgList(qemu,
1072
                               "-device", "?",
E
Eric Blake 已提交
1073
                               "-device", "pci-assign,?",
1074
                               "-device", "virtio-blk-pci,?",
E
Eric Blake 已提交
1075
                               NULL);
1076 1077
    virCommandAddEnvPassCommon(cmd);
    /* qemu -help goes to stdout, but qemu -device ? goes to stderr.  */
E
Eric Blake 已提交
1078
    virCommandSetErrorBuffer(cmd, &output);
1079
    virCommandClearCaps(cmd);
1080

1081
    if (virCommandRun(cmd, NULL) < 0)
1082 1083
        goto cleanup;

E
Eric Blake 已提交
1084
    ret = qemuCapsParseDeviceStr(output, flags);
1085 1086

cleanup:
E
Eric Blake 已提交
1087
    VIR_FREE(output);
1088
    virCommandFree(cmd);
E
Eric Blake 已提交
1089 1090 1091 1092 1093 1094 1095
    return ret;
}


int
qemuCapsParseDeviceStr(const char *str, unsigned long long *flags)
{
1096 1097 1098
    /* Which devices exist. */
    if (strstr(str, "name \"hda-duplex\""))
        *flags |= QEMUD_CMD_FLAG_HDA_DUPLEX;
1099 1100 1101 1102
    if (strstr(str, "name \"ccid-card-emulated\""))
        *flags |= QEMUD_CMD_FLAG_CCID_EMULATED;
    if (strstr(str, "name \"ccid-card-passthru\""))
        *flags |= QEMUD_CMD_FLAG_CCID_PASSTHRU;
1103 1104 1105 1106
    /* Prefer -chardev spicevmc (detected earlier) over -device spicevmc */
    if (!(*flags & QEMUD_CMD_FLAG_CHARDEV_SPICEVMC) &&
        strstr(str, "name \"spicevmc\""))
        *flags |= QEMUD_CMD_FLAG_DEVICE_SPICEVMC;
1107 1108

    /* Features of given devices. */
E
Eric Blake 已提交
1109 1110
    if (strstr(str, "pci-assign.configfd"))
        *flags |= QEMUD_CMD_FLAG_PCI_CONFIGFD;
1111
    if (strstr(str, "virtio-blk-pci.bootindex")) {
1112
        *flags |= QEMUD_CMD_FLAG_BOOTINDEX;
1113 1114 1115
        if (strstr(str, "pci-assign.bootindex"))
            *flags |= QEMUD_CMD_FLAG_PCI_BOOTINDEX;
    }
E
Eric Blake 已提交
1116 1117

    return 0;
1118 1119 1120 1121 1122 1123
}

int qemuCapsExtractVersionInfo(const char *qemu,
                               unsigned int *retversion,
                               unsigned long long *retflags)
{
1124
    int ret = -1;
1125 1126
    unsigned int version, is_kvm, kvm_version;
    unsigned long long flags = 0;
1127 1128
    char *help = NULL;
    virCommandPtr cmd;
1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143

    if (retflags)
        *retflags = 0;
    if (retversion)
        *retversion = 0;

    /* Make sure the binary we are about to try exec'ing exists.
     * Technically we could catch the exec() failure, but that's
     * in a sub-process so it's hard to feed back a useful error.
     */
    if (access(qemu, X_OK) < 0) {
        virReportSystemError(errno, _("Cannot find QEMU binary %s"), qemu);
        return -1;
    }

1144 1145 1146 1147
    cmd = virCommandNewArgList(qemu, "-help", NULL);
    virCommandAddEnvPassCommon(cmd);
    virCommandSetOutputBuffer(cmd, &help);
    virCommandClearCaps(cmd);
1148

1149 1150
    if (virCommandRun(cmd, NULL) < 0)
        goto cleanup;
1151 1152 1153

    if (qemuCapsParseHelpStr(qemu, help, &flags,
                             &version, &is_kvm, &kvm_version) == -1)
1154
        goto cleanup;
1155

1156 1157
    /* qemuCapsExtractDeviceStr will only set additional flags if qemu
     * understands the 0.13.0+ notion of "-device driver,".  */
E
Eric Blake 已提交
1158
    if ((flags & QEMUD_CMD_FLAG_DEVICE) &&
1159
        strstr(help, "-device driver,?") &&
E
Eric Blake 已提交
1160 1161
        qemuCapsExtractDeviceStr(qemu, &flags) < 0)
        goto cleanup;
1162 1163 1164 1165 1166 1167 1168 1169

    if (retversion)
        *retversion = version;
    if (retflags)
        *retflags = flags;

    ret = 0;

1170
cleanup:
1171
    VIR_FREE(help);
1172
    virCommandFree(cmd);
1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222

    return ret;
}

static void
uname_normalize (struct utsname *ut)
{
    uname(ut);

    /* Map i386, i486, i586 to i686.  */
    if (ut->machine[0] == 'i' &&
        ut->machine[1] != '\0' &&
        ut->machine[2] == '8' &&
        ut->machine[3] == '6' &&
        ut->machine[4] == '\0')
        ut->machine[1] = '6';
}

int qemuCapsExtractVersion(virCapsPtr caps,
                           unsigned int *version)
{
    const char *binary;
    struct stat sb;
    struct utsname ut;

    if (*version > 0)
        return 0;

    uname_normalize(&ut);
    if ((binary = virCapabilitiesDefaultGuestEmulator(caps,
                                                      "hvm",
                                                      ut.machine,
                                                      "qemu")) == NULL) {
        qemuReportError(VIR_ERR_INTERNAL_ERROR,
                        _("Cannot find suitable emulator for %s"), ut.machine);
        return -1;
    }

    if (stat(binary, &sb) < 0) {
        virReportSystemError(errno,
                             _("Cannot find QEMU binary %s"), binary);
        return -1;
    }

    if (qemuCapsExtractVersionInfo(binary, version, NULL) < 0) {
        return -1;
    }

    return 0;
}