bridge_driver.c 101.1 KB
Newer Older
1
/*
2
 * bridge_driver.c: core driver methods for managing network
3
 *
E
Eric Blake 已提交
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 36 37 38 39 40 41 42 43 44
 * 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 <sys/types.h>
#include <sys/poll.h>
#include <dirent.h>
#include <limits.h>
#include <string.h>
#include <stdio.h>
#include <stdarg.h>
#include <stdlib.h>
#include <unistd.h>
#include <errno.h>
#include <sys/utsname.h>
#include <sys/stat.h>
#include <fcntl.h>
#include <signal.h>
#include <paths.h>
#include <pwd.h>
#include <stdio.h>
#include <sys/wait.h>
#include <sys/ioctl.h>
45
#include <net/if.h>
46

47
#include "virterror_internal.h"
48
#include "datatypes.h"
49
#include "bridge_driver.h"
50 51 52
#include "network_conf.h"
#include "driver.h"
#include "buf.h"
53
#include "virpidfile.h"
54
#include "util.h"
55
#include "command.h"
56 57 58
#include "memory.h"
#include "uuid.h"
#include "iptables.h"
59
#include "interface.h"
60
#include "logging.h"
61
#include "dnsmasq.h"
62
#include "util/network.h"
63
#include "configmake.h"
64
#include "ignore-value.h"
65 66 67
#include "virnetdev.h"
#include "virnetdevbridge.h"
#include "virnetdevtap.h"
68

69 70
#define NETWORK_PID_DIR LOCALSTATEDIR "/run/libvirt/network"
#define NETWORK_STATE_DIR LOCALSTATEDIR "/lib/libvirt/network"
71

72
#define DNSMASQ_STATE_DIR LOCALSTATEDIR "/lib/libvirt/dnsmasq"
73
#define RADVD_STATE_DIR LOCALSTATEDIR "/lib/libvirt/radvd"
74

75 76
#define VIR_FROM_THIS VIR_FROM_NETWORK

77
#define networkReportError(code, ...)                                   \
78
    virReportErrorHelper(VIR_FROM_NETWORK, code, __FILE__,              \
79
                         __FUNCTION__, __LINE__, __VA_ARGS__)
80

81 82
/* Main driver state */
struct network_driver {
83
    virMutex lock;
84

85
    virNetworkObjList networks;
86 87 88 89 90 91 92

    iptablesContext *iptables;
    char *networkConfigDir;
    char *networkAutostartDir;
    char *logDir;
};

93 94 95

static void networkDriverLock(struct network_driver *driver)
{
96
    virMutexLock(&driver->lock);
97 98 99
}
static void networkDriverUnlock(struct network_driver *driver)
{
100
    virMutexUnlock(&driver->lock);
101 102
}

103 104
static int networkShutdown(void);

105 106 107 108 109 110 111
static int networkStartNetwork(struct network_driver *driver,
                               virNetworkObjPtr network);

static int networkShutdownNetwork(struct network_driver *driver,
                                  virNetworkObjPtr network);

static int networkStartNetworkVirtual(struct network_driver *driver,
112
                                     virNetworkObjPtr network);
113

114 115 116 117 118 119 120
static int networkShutdownNetworkVirtual(struct network_driver *driver,
                                        virNetworkObjPtr network);

static int networkStartNetworkExternal(struct network_driver *driver,
                                     virNetworkObjPtr network);

static int networkShutdownNetworkExternal(struct network_driver *driver,
121
                                        virNetworkObjPtr network);
122

123 124
static void networkReloadIptablesRules(struct network_driver *driver);

125 126
static struct network_driver *driverState = NULL;

127
static char *
128
networkDnsmasqLeaseFileNameDefault(const char *netname)
129 130 131
{
    char *leasefile;

132 133
    ignore_value(virAsprintf(&leasefile, DNSMASQ_STATE_DIR "/%s.leases",
                             netname));
134 135 136
    return leasefile;
}

137 138 139
networkDnsmasqLeaseFileNameFunc networkDnsmasqLeaseFileName =
    networkDnsmasqLeaseFileNameDefault;

140 141 142 143 144 145
static char *
networkRadvdPidfileBasename(const char *netname)
{
    /* this is simple but we want to be sure it's consistently done */
    char *pidfilebase;

146
    ignore_value(virAsprintf(&pidfilebase, "%s-radvd", netname));
147 148 149 150 151 152 153 154
    return pidfilebase;
}

static char *
networkRadvdConfigFileName(const char *netname)
{
    char *configfile;

155 156
    ignore_value(virAsprintf(&configfile, RADVD_STATE_DIR "/%s-radvd.conf",
                             netname));
157 158
    return configfile;
}
159

160 161 162
static char *
networkBridgeDummyNicName(const char *brname)
{
163
    static const char dummyNicSuffix[] = "-nic";
164 165
    char *nicname;

166 167 168 169 170 171 172
    if (strlen(brname) + sizeof(dummyNicSuffix) > IFNAMSIZ) {
        /* because the length of an ifname is limited to IFNAMSIZ-1
         * (usually 15), and we're adding 4 more characters, we must
         * truncate the original name to 11 to fit. In order to catch
         * a possible numeric ending (eg virbr0, virbr1, etc), we grab
         * the first 8 and last 3 characters of the string.
         */
173 174 175 176 177
        ignore_value(virAsprintf(&nicname, "%.*s%s%s",
                                 /* space for last 3 chars + "-nic" + NULL */
                                 (int)(IFNAMSIZ - (3 + sizeof(dummyNicSuffix))),
                                 brname, brname + strlen(brname) - 3,
                                 dummyNicSuffix));
178
    } else {
179
        ignore_value(virAsprintf(&nicname, "%s%s", brname, dummyNicSuffix));
180
    }
181 182 183
    return nicname;
}

184 185 186 187 188 189 190 191 192 193 194
static void
networkFindActiveConfigs(struct network_driver *driver) {
    unsigned int i;

    for (i = 0 ; i < driver->networks.count ; i++) {
        virNetworkObjPtr obj = driver->networks.objs[i];
        virNetworkDefPtr tmp;
        char *config;

        virNetworkObjLock(obj);

195
        if ((config = virNetworkConfigFile(NETWORK_STATE_DIR,
196 197 198 199 200 201 202 203 204 205 206 207
                                           obj->def->name)) == NULL) {
            virNetworkObjUnlock(obj);
            continue;
        }

        if (access(config, R_OK) < 0) {
            VIR_FREE(config);
            virNetworkObjUnlock(obj);
            continue;
        }

        /* Try and load the live config */
208
        tmp = virNetworkDefParseFile(config);
209 210 211 212 213 214 215 216
        VIR_FREE(config);
        if (tmp) {
            obj->newDef = obj->def;
            obj->def = tmp;
        }

        /* If bridge exists, then mark it active */
        if (obj->def->bridge &&
217
            virNetDevExists(obj->def->bridge) == 0) {
218 219
            obj->active = 1;

220 221
            /* Try and read dnsmasq/radvd pids if any */
            if (obj->def->ips && (obj->def->nips > 0)) {
222 223 224 225
                char *radvdpidbase;

                ignore_value(virPidFileReadIfAlive(NETWORK_PID_DIR, obj->def->name,
                                                   &obj->dnsmasqPid, DNSMASQ));
226

227
                if (!(radvdpidbase = networkRadvdPidfileBasename(obj->def->name))) {
228
                    virReportOOMError();
229 230
                    goto cleanup;
                }
231 232
                ignore_value(virPidFileReadIfAlive(NETWORK_PID_DIR, radvdpidbase,
                                                   &obj->radvdPid, RADVD));
233
                VIR_FREE(radvdpidbase);
234 235 236
            }
        }

237
    cleanup:
238 239 240 241 242
        virNetworkObjUnlock(obj);
    }
}


243 244 245
static void
networkAutostartConfigs(struct network_driver *driver) {
    unsigned int i;
246

247
    for (i = 0 ; i < driver->networks.count ; i++) {
248
        virNetworkObjLock(driver->networks.objs[i]);
249
        if (driver->networks.objs[i]->autostart &&
250 251
            !virNetworkObjIsActive(driver->networks.objs[i])) {
            if (networkStartNetwork(driver, driver->networks.objs[i]) < 0) {
252
            /* failed to start but already logged */
253
            }
254
        }
255
        virNetworkObjUnlock(driver->networks.objs[i]);
256 257 258 259 260 261 262 263 264
    }
}

/**
 * networkStartup:
 *
 * Initialization function for the QEmu daemon
 */
static int
265
networkStartup(int privileged) {
266 267 268 269
    uid_t uid = geteuid();
    char *base = NULL;

    if (VIR_ALLOC(driverState) < 0)
270
        goto error;
271

272 273 274 275
    if (virMutexInit(&driverState->lock) < 0) {
        VIR_FREE(driverState);
        goto error;
    }
276 277
    networkDriverLock(driverState);

278
    if (privileged) {
279
        if (virAsprintf(&driverState->logDir,
280
                        "%s/log/libvirt/qemu", LOCALSTATEDIR) == -1)
281 282
            goto out_of_memory;

283
        if ((base = strdup (SYSCONFDIR "/libvirt")) == NULL)
284 285
            goto out_of_memory;
    } else {
286
        char *userdir = virGetUserDirectory(uid);
287 288 289

        if (!userdir)
            goto error;
290

291
        if (virAsprintf(&driverState->logDir,
292 293
                        "%s/.libvirt/qemu/log", userdir) == -1) {
            VIR_FREE(userdir);
294
            goto out_of_memory;
295
        }
296

297 298
        if (virAsprintf(&base, "%s/.libvirt", userdir) == -1) {
            VIR_FREE(userdir);
299 300
            goto out_of_memory;
        }
301
        VIR_FREE(userdir);
302 303 304 305 306
    }

    /* Configuration paths are either ~/.libvirt/qemu/... (session) or
     * /etc/libvirt/qemu/... (system).
     */
307
    if (virAsprintf(&driverState->networkConfigDir, "%s/qemu/networks", base) == -1)
308 309
        goto out_of_memory;

310 311
    if (virAsprintf(&driverState->networkAutostartDir, "%s/qemu/networks/autostart",
                    base) == -1)
312 313 314 315
        goto out_of_memory;

    VIR_FREE(base);

316
    if (!(driverState->iptables = iptablesContextNew())) {
317
        goto out_of_memory;
318 319 320
    }


321
    if (virNetworkLoadAllConfigs(&driverState->networks,
322
                                 driverState->networkConfigDir,
323 324 325
                                 driverState->networkAutostartDir) < 0)
        goto error;

326
    networkFindActiveConfigs(driverState);
327
    networkReloadIptablesRules(driverState);
328 329
    networkAutostartConfigs(driverState);

330 331
    networkDriverUnlock(driverState);

332 333
    return 0;

334
out_of_memory:
335
    virReportOOMError();
336 337

error:
338 339 340
    if (driverState)
        networkDriverUnlock(driverState);

341
    VIR_FREE(base);
342
    networkShutdown();
343 344 345 346 347 348 349 350 351 352 353
    return -1;
}

/**
 * networkReload:
 *
 * Function to restart the QEmu daemon, it will recheck the configuration
 * files and update its state and the networking
 */
static int
networkReload(void) {
354 355 356
    if (!driverState)
        return 0;

357
    networkDriverLock(driverState);
358
    virNetworkLoadAllConfigs(&driverState->networks,
359 360
                             driverState->networkConfigDir,
                             driverState->networkAutostartDir);
361
    networkReloadIptablesRules(driverState);
362
    networkAutostartConfigs(driverState);
363
    networkDriverUnlock(driverState);
364 365 366 367 368 369 370 371 372 373 374 375 376
    return 0;
}

/**
 * networkActive:
 *
 * Checks if the QEmu daemon is active, i.e. has an active domain or
 * an active network
 *
 * Returns 1 if active, 0 otherwise
 */
static int
networkActive(void) {
377
    unsigned int i;
378
    int active = 0;
379

380 381 382
    if (!driverState)
        return 0;

383
    networkDriverLock(driverState);
384 385
    for (i = 0 ; i < driverState->networks.count ; i++) {
        virNetworkObjPtr net = driverState->networks.objs[i];
386
        virNetworkObjLock(net);
D
Daniel P. Berrange 已提交
387
        if (virNetworkObjIsActive(net))
388
            active = 1;
389
        virNetworkObjUnlock(net);
390
    }
391
    networkDriverUnlock(driverState);
392
    return active;
393 394 395 396 397 398 399 400 401 402 403 404
}

/**
 * networkShutdown:
 *
 * Shutdown the QEmu daemon, it will stop all active domains and networks
 */
static int
networkShutdown(void) {
    if (!driverState)
        return -1;

405 406
    networkDriverLock(driverState);

407
    /* free inactive networks */
408
    virNetworkObjListFree(&driverState->networks);
409 410 411 412 413 414 415 416

    VIR_FREE(driverState->logDir);
    VIR_FREE(driverState->networkConfigDir);
    VIR_FREE(driverState->networkAutostartDir);

    if (driverState->iptables)
        iptablesContextFree(driverState->iptables);

417
    networkDriverUnlock(driverState);
418
    virMutexDestroy(&driverState->lock);
419

420 421 422 423 424 425
    VIR_FREE(driverState);

    return 0;
}


426 427 428 429
static int
networkBuildDnsmasqHostsfile(dnsmasqContext *dctx,
                             virNetworkIpDefPtr ipdef,
                             virNetworkDNSDefPtr dnsdef)
430
{
431
    unsigned int i, j;
432

433 434
    for (i = 0; i < ipdef->nhosts; i++) {
        virNetworkDHCPHostDefPtr host = &(ipdef->hosts[i]);
435
        if ((host->mac) && VIR_SOCKET_ADDR_VALID(&host->ip))
436 437
            if (dnsmasqAddDhcpHost(dctx, host->mac, &host->ip, host->name) < 0)
                return -1;
438
    }
439

440 441 442
    if (dnsdef) {
        for (i = 0; i < dnsdef->nhosts; i++) {
            virNetworkDNSHostsDefPtr host = &(dnsdef->hosts[i]);
443
            if (VIR_SOCKET_ADDR_VALID(&host->ip)) {
444
                for (j = 0; j < host->nnames; j++)
445 446
                    if (dnsmasqAddHost(dctx, &host->ip, host->names[j]) < 0)
                        return -1;
447 448
            }
        }
449 450
    }

451
    return 0;
452 453 454
}


455
static int
456
networkBuildDnsmasqArgv(virNetworkObjPtr network,
457
                        virNetworkIpDefPtr ipdef,
458
                        const char *pidfile,
459 460 461
                        virCommandPtr cmd,
                        dnsmasqContext *dctx)
{
462
    int r, ret = -1;
463
    int nbleases = 0;
464 465
    int ii;
    virNetworkIpDefPtr tmpipdef;
466 467

    /*
468
     * NB, be careful about syntax for dnsmasq options in long format.
469 470 471 472 473 474 475 476 477 478 479 480 481
     *
     * If the flag has a mandatory argument, it can be given using
     * either syntax:
     *
     *     --foo bar
     *     --foo=bar
     *
     * If the flag has a optional argument, it *must* be given using
     * the syntax:
     *
     *     --foo=bar
     *
     * It is hard to determine whether a flag is optional or not,
482 483
     * without reading the dnsmasq source :-( The manpage is not
     * very explicit on this.
484
     */
485 486 487 488 489

    /*
     * Needed to ensure dnsmasq uses same algorithm for processing
     * multiple namedriver entries in /etc/resolv.conf as GLibC.
     */
490
    virCommandAddArgList(cmd, "--strict-order", "--bind-interfaces", NULL);
491

492 493
    if (network->def->domain)
        virCommandAddArgList(cmd, "--domain", network->def->domain, NULL);
494

495 496
    if (pidfile)
        virCommandAddArgPair(cmd, "--pid-file", pidfile);
497

498
    /* *no* conf file */
499
    virCommandAddArg(cmd, "--conf-file=");
500

501 502 503
    virCommandAddArgList(cmd,
                         "--except-interface", "lo",
                         NULL);
504

505 506
    /* If this is an isolated network, set the default route option
     * (3) to be empty to avoid setting a default route that's
507 508 509 510
     * guaranteed to not work, and set --no-resolv so that no dns
     * requests are forwarded on to the dns server listed in the
     * host's /etc/resolv.conf (since this could be used as a channel
     * to build a connection to the outside).
511
     */
512 513 514 515
    if (network->def->forwardType == VIR_NETWORK_FORWARD_NONE) {
        virCommandAddArgList(cmd, "--dhcp-option=3",
                             "--no-resolv", NULL);
    }
516

517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534
    if (network->def->dns != NULL) {
        virNetworkDNSDefPtr dns = network->def->dns;
        int i;

        for (i = 0; i < dns->ntxtrecords; i++) {
            char *record = NULL;
            if (virAsprintf(&record, "%s,%s",
                            dns->txtrecords[i].name,
                            dns->txtrecords[i].value) < 0) {
                virReportOOMError();
                goto cleanup;
            }

            virCommandAddArgPair(cmd, "--txt-record", record);
            VIR_FREE(record);
        }
    }

535 536 537 538 539 540 541 542 543 544 545
    /*
     * --interface does not actually work with dnsmasq < 2.47,
     * due to DAD for ipv6 addresses on the interface.
     *
     * virCommandAddArgList(cmd, "--interface", ipdef->bridge, NULL);
     *
     * So listen on all defined IPv[46] addresses
     */
    for (ii = 0;
         (tmpipdef = virNetworkDefGetIpByIndex(network->def, AF_UNSPEC, ii));
         ii++) {
546
        char *ipaddr = virSocketAddrFormat(&tmpipdef->address);
547 548 549 550 551 552
        if (!ipaddr)
            goto cleanup;
        virCommandAddArgList(cmd, "--listen-address", ipaddr, NULL);
        VIR_FREE(ipaddr);
    }

553
    if (ipdef) {
554
        for (r = 0 ; r < ipdef->nranges ; r++) {
555
            char *saddr = virSocketAddrFormat(&ipdef->ranges[r].start);
556 557
            if (!saddr)
                goto cleanup;
558
            char *eaddr = virSocketAddrFormat(&ipdef->ranges[r].end);
559 560 561 562 563 564
            if (!eaddr) {
                VIR_FREE(saddr);
                goto cleanup;
            }
            virCommandAddArg(cmd, "--dhcp-range");
            virCommandAddArgFormat(cmd, "%s,%s", saddr, eaddr);
565
            VIR_FREE(saddr);
566
            VIR_FREE(eaddr);
567 568
            nbleases += virSocketAddrGetRange(&ipdef->ranges[r].start,
                                              &ipdef->ranges[r].end);
569
        }
570

571 572 573 574 575 576
        /*
         * For static-only DHCP, i.e. with no range but at least one host element,
         * we have to add a special --dhcp-range option to enable the service in
         * dnsmasq.
         */
        if (!ipdef->nranges && ipdef->nhosts) {
577
            char *bridgeaddr = virSocketAddrFormat(&ipdef->address);
578 579 580 581 582 583
            if (!bridgeaddr)
                goto cleanup;
            virCommandAddArg(cmd, "--dhcp-range");
            virCommandAddArgFormat(cmd, "%s,static", bridgeaddr);
            VIR_FREE(bridgeaddr);
        }
584

585
        if (ipdef->nranges > 0) {
586 587 588 589 590
            char *leasefile = networkDnsmasqLeaseFileName(network->def->name);
            if (!leasefile)
                goto cleanup;
            virCommandAddArgFormat(cmd, "--dhcp-leasefile=%s", leasefile);
            VIR_FREE(leasefile);
591 592
            virCommandAddArgFormat(cmd, "--dhcp-lease-max=%d", nbleases);
        }
593

594 595
        if (ipdef->nranges || ipdef->nhosts)
            virCommandAddArg(cmd, "--dhcp-no-override");
596

597 598 599 600
        /* add domain to any non-qualified hostnames in /etc/hosts or addn-hosts */
        if (network->def->domain)
           virCommandAddArg(cmd, "--expand-hosts");

601 602 603 604 605 606 607 608 609
        if (networkBuildDnsmasqHostsfile(dctx, ipdef, network->def->dns) < 0)
            goto cleanup;

        if (dctx->hostsfile->nhosts)
            virCommandAddArgPair(cmd, "--dhcp-hostsfile",
                                 dctx->hostsfile->path);
        if (dctx->addnhostsfile->nhosts)
            virCommandAddArgPair(cmd, "--addn-hosts",
                                 dctx->addnhostsfile->path);
610

611 612 613 614 615 616 617
        if (ipdef->tftproot) {
            virCommandAddArgList(cmd, "--enable-tftp",
                                 "--tftp-root", ipdef->tftproot,
                                 NULL);
        }
        if (ipdef->bootfile) {
            virCommandAddArg(cmd, "--dhcp-boot");
618 619
            if (VIR_SOCKET_ADDR_VALID(&ipdef->bootserver)) {
                char *bootserver = virSocketAddrFormat(&ipdef->bootserver);
620

621 622 623 624 625 626 627 628
                if (!bootserver)
                    goto cleanup;
                virCommandAddArgFormat(cmd, "%s%s%s",
                                       ipdef->bootfile, ",,", bootserver);
                VIR_FREE(bootserver);
            } else {
                virCommandAddArg(cmd, ipdef->bootfile);
            }
629
        }
630 631
    }

632 633 634
    ret = 0;
cleanup:
    return ret;
635 636
}

637 638
int
networkBuildDhcpDaemonCommandLine(virNetworkObjPtr network, virCommandPtr *cmdout,
639
                                  char *pidfile, dnsmasqContext *dctx)
640
{
641
    virCommandPtr cmd = NULL;
642
    int ret = -1, ii;
643
    virNetworkIpDefPtr ipdef;
644 645

    network->dnsmasqPid = -1;
646

647 648 649 650 651 652 653
    /* Look for first IPv4 address that has dhcp defined. */
    /* We support dhcp config on 1 IPv4 interface only. */
    for (ii = 0;
         (ipdef = virNetworkDefGetIpByIndex(network->def, AF_INET, ii));
         ii++) {
        if (ipdef->nranges || ipdef->nhosts)
            break;
654
    }
655
    /* If no IPv4 addresses had dhcp info, pick the first (if there were any). */
656
    if (!ipdef)
657 658 659 660 661 662 663 664
        ipdef = virNetworkDefGetIpByIndex(network->def, AF_INET, 0);

    /* If there are no IP addresses at all (v4 or v6), return now, since
     * there won't be any address for dnsmasq to listen on anyway.
     * If there are any addresses, even if no dhcp ranges or static entries,
     * we should continue and run dnsmasq, just for the DNS capabilities.
     */
    if (!virNetworkDefGetIpByIndex(network->def, AF_UNSPEC, 0))
665
        return 0;
666

667
    cmd = virCommandNew(DNSMASQ);
668
    if (networkBuildDnsmasqArgv(network, ipdef, pidfile, cmd, dctx) < 0) {
669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686
        goto cleanup;
    }

    if (cmdout)
        *cmdout = cmd;
    ret = 0;
cleanup:
    if (ret < 0)
        virCommandFree(cmd);
    return ret;
}

static int
networkStartDhcpDaemon(virNetworkObjPtr network)
{
    virCommandPtr cmd = NULL;
    char *pidfile = NULL;
    int ret = -1;
687
    dnsmasqContext *dctx = NULL;
688

689 690
    if (virFileMakePath(NETWORK_PID_DIR) < 0) {
        virReportSystemError(errno,
691 692
                             _("cannot create directory %s"),
                             NETWORK_PID_DIR);
693
        goto cleanup;
694
    }
695 696
    if (virFileMakePath(NETWORK_STATE_DIR) < 0) {
        virReportSystemError(errno,
697 698
                             _("cannot create directory %s"),
                             NETWORK_STATE_DIR);
699
        goto cleanup;
700 701
    }

702
    if (!(pidfile = virPidFileBuildPath(NETWORK_PID_DIR, network->def->name))) {
703
        virReportOOMError();
704
        goto cleanup;
705 706
    }

707 708
    if (virFileMakePath(DNSMASQ_STATE_DIR) < 0) {
        virReportSystemError(errno,
709 710 711 712 713
                             _("cannot create directory %s"),
                             DNSMASQ_STATE_DIR);
        goto cleanup;
    }

714 715 716 717 718 719 720 721 722 723
    dctx = dnsmasqContextNew(network->def->name, DNSMASQ_STATE_DIR);
    if (dctx == NULL)
        goto cleanup;

    ret = networkBuildDhcpDaemonCommandLine(network, &cmd, pidfile, dctx);
    if (ret < 0)
        goto cleanup;

    ret = dnsmasqSave(dctx);
    if (ret < 0)
724
        goto cleanup;
725

G
Guido Günther 已提交
726 727
    ret = virCommandRun(cmd, NULL);
    if (ret < 0) {
728
        goto cleanup;
G
Guido Günther 已提交
729
    }
730 731

    /*
732 733 734 735 736
     * There really is no race here - when dnsmasq daemonizes, its
     * leader process stays around until its child has actually
     * written its pidfile. So by time virCommandRun exits it has
     * waitpid'd and guaranteed the proess has started and written a
     * pid
737 738
     */

739
    ret = virPidFileRead(NETWORK_PID_DIR, network->def->name,
740 741
                         &network->dnsmasqPid);
    if (ret < 0)
742
        goto cleanup;
743

744 745 746
    ret = 0;
cleanup:
    VIR_FREE(pidfile);
747
    virCommandFree(cmd);
748
    dnsmasqContextFree(dctx);
749 750 751
    return ret;
}

752 753 754 755 756 757 758 759 760
static int
networkStartRadvd(virNetworkObjPtr network)
{
    char *pidfile = NULL;
    char *radvdpidbase = NULL;
    virBuffer configbuf = VIR_BUFFER_INITIALIZER;;
    char *configstr = NULL;
    char *configfile = NULL;
    virCommandPtr cmd = NULL;
761
    int ret = -1, ii;
762 763 764 765
    virNetworkIpDefPtr ipdef;

    network->radvdPid = -1;

E
Eric Blake 已提交
766
    if (!virFileIsExecutable(RADVD)) {
767 768 769 770 771 772 773
        virReportSystemError(errno,
                             _("Cannot find %s - "
                               "Possibly the package isn't installed"),
                             RADVD);
        goto cleanup;
    }

774 775
    if (virFileMakePath(NETWORK_PID_DIR) < 0) {
        virReportSystemError(errno,
776 777 778 779
                             _("cannot create directory %s"),
                             NETWORK_PID_DIR);
        goto cleanup;
    }
780 781
    if (virFileMakePath(RADVD_STATE_DIR) < 0) {
        virReportSystemError(errno,
782 783 784 785 786 787 788 789 790 791
                             _("cannot create directory %s"),
                             RADVD_STATE_DIR);
        goto cleanup;
    }

    /* construct pidfile name */
    if (!(radvdpidbase = networkRadvdPidfileBasename(network->def->name))) {
        virReportOOMError();
        goto cleanup;
    }
792
    if (!(pidfile = virPidFileBuildPath(NETWORK_PID_DIR, radvdpidbase))) {
793 794 795 796 797
        virReportOOMError();
        goto cleanup;
    }

    /* create radvd config file appropriate for this network */
798
    virBufferAsprintf(&configbuf, "interface %s\n"
799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817
                      "{\n"
                      "  AdvSendAdvert on;\n"
                      "  AdvManagedFlag off;\n"
                      "  AdvOtherConfigFlag off;\n"
                      "\n",
                      network->def->bridge);
    for (ii = 0;
         (ipdef = virNetworkDefGetIpByIndex(network->def, AF_INET6, ii));
         ii++) {
        int prefix;
        char *netaddr;

        prefix = virNetworkIpDefPrefix(ipdef);
        if (prefix < 0) {
            networkReportError(VIR_ERR_INTERNAL_ERROR,
                               _("bridge  '%s' has an invalid prefix"),
                               network->def->bridge);
            goto cleanup;
        }
818
        if (!(netaddr = virSocketAddrFormat(&ipdef->address)))
819
            goto cleanup;
820
        virBufferAsprintf(&configbuf,
821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858
                          "  prefix %s/%d\n"
                          "  {\n"
                          "    AdvOnLink on;\n"
                          "    AdvAutonomous on;\n"
                          "    AdvRouterAddr off;\n"
                          "  };\n",
                          netaddr, prefix);
        VIR_FREE(netaddr);
    }

    virBufferAddLit(&configbuf, "};\n");

    if (virBufferError(&configbuf)) {
        virReportOOMError();
        goto cleanup;
    }
    if (!(configstr = virBufferContentAndReset(&configbuf))) {
        virReportOOMError();
        goto cleanup;
    }

    /* construct the filename */
    if (!(configfile = networkRadvdConfigFileName(network->def->name))) {
        virReportOOMError();
        goto cleanup;
    }
    /* write the file */
    if (virFileWriteStr(configfile, configstr, 0600) < 0) {
        virReportSystemError(errno,
                             _("couldn't write radvd config file '%s'"),
                             configfile);
        goto cleanup;
    }

    /* prevent radvd from daemonizing itself with "--debug 1", and use
     * a dummy pidfile name - virCommand will create the pidfile we
     * want to use (this is necessary because radvd's internal
     * daemonization and pidfile creation causes a race, and the
859
     * virPidFileRead() below will fail if we use them).
860 861 862 863 864 865 866 867 868 869 870 871 872 873 874
     * Unfortunately, it isn't possible to tell radvd to not create
     * its own pidfile, so we just let it do so, with a slightly
     * different name. Unused, but harmless.
     */
    cmd = virCommandNewArgList(RADVD, "--debug", "1",
                               "--config", configfile,
                               "--pidfile", NULL);
    virCommandAddArgFormat(cmd, "%s-bin", pidfile);

    virCommandSetPidFile(cmd, pidfile);
    virCommandDaemonize(cmd);

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

875
    if (virPidFileRead(NETWORK_PID_DIR, radvdpidbase,
876 877 878 879 880 881 882 883 884 885 886 887 888 889
                       &network->radvdPid) < 0)
        goto cleanup;

    ret = 0;
cleanup:
    virCommandFree(cmd);
    VIR_FREE(configfile);
    VIR_FREE(configstr);
    virBufferFreeAndReset(&configbuf);
    VIR_FREE(radvdpidbase);
    VIR_FREE(pidfile);
    return ret;
}

890
static int
891
networkAddMasqueradingIptablesRules(struct network_driver *driver,
892 893
                                    virNetworkObjPtr network,
                                    virNetworkIpDefPtr ipdef)
894 895
{
    int prefix = virNetworkIpDefPrefix(ipdef);
896
    const char *forwardIf = virNetworkDefForwardIf(network->def, 0);
897 898 899 900 901 902 903

    if (prefix < 0) {
        networkReportError(VIR_ERR_INTERNAL_ERROR,
                           _("Invalid prefix or netmask for '%s'"),
                           network->def->bridge);
        goto masqerr1;
    }
904

905
    /* allow forwarding packets from the bridge interface */
906
    if (iptablesAddForwardAllowOut(driver->iptables,
907
                                   &ipdef->address,
908
                                   prefix,
909
                                   network->def->bridge,
910
                                   forwardIf) < 0) {
911 912 913
        networkReportError(VIR_ERR_SYSTEM_ERROR,
                           _("failed to add iptables rule to allow forwarding from '%s'"),
                           network->def->bridge);
914 915 916
        goto masqerr1;
    }

917 918 919
    /* allow forwarding packets to the bridge interface if they are
     * part of an existing connection
     */
920
    if (iptablesAddForwardAllowRelatedIn(driver->iptables,
921
                                         &ipdef->address,
922
                                         prefix,
923
                                         network->def->bridge,
924
                                         forwardIf) < 0) {
925 926 927
        networkReportError(VIR_ERR_SYSTEM_ERROR,
                           _("failed to add iptables rule to allow forwarding to '%s'"),
                           network->def->bridge);
928 929 930
        goto masqerr2;
    }

931 932 933 934 935
    /*
     * Enable masquerading.
     *
     * We need to end up with 3 rules in the table in this order
     *
E
Eric Blake 已提交
936 937
     *  1. protocol=tcp with sport mapping restriction
     *  2. protocol=udp with sport mapping restriction
938 939 940
     *  3. generic any protocol
     *
     * The sport mappings are required, because default IPtables
E
Eric Blake 已提交
941
     * MASQUERADE maintain port numbers unchanged where possible.
942 943 944 945 946 947 948 949 950 951 952 953 954
     *
     * NFS can be configured to only "trust" port numbers < 1023.
     *
     * Guests using NAT thus need to be prevented from having port
     * numbers < 1023, otherwise they can bypass the NFS "security"
     * check on the source port number.
     *
     * Since we use '--insert' to add rules to the header of the
     * chain, we actually need to add them in the reverse of the
     * order just mentioned !
     */

    /* First the generic masquerade rule for other protocols */
955
    if (iptablesAddForwardMasquerade(driver->iptables,
956
                                     &ipdef->address,
957
                                     prefix,
958
                                     forwardIf,
959 960
                                     NULL) < 0) {
        networkReportError(VIR_ERR_SYSTEM_ERROR,
961 962 963 964
                           forwardIf ?
                           _("failed to add iptables rule to enable masquerading to %s") :
                           _("failed to add iptables rule to enable masquerading"),
                           forwardIf);
965 966 967
        goto masqerr3;
    }

968
    /* UDP with a source port restriction */
969
    if (iptablesAddForwardMasquerade(driver->iptables,
970
                                     &ipdef->address,
971
                                     prefix,
972
                                     forwardIf,
973 974
                                     "udp") < 0) {
        networkReportError(VIR_ERR_SYSTEM_ERROR,
975 976 977 978
                           forwardIf ?
                           _("failed to add iptables rule to enable UDP masquerading to %s") :
                           _("failed to add iptables rule to enable UDP masquerading"),
                           forwardIf);
979 980 981 982
        goto masqerr4;
    }

    /* TCP with a source port restriction */
983
    if (iptablesAddForwardMasquerade(driver->iptables,
984
                                     &ipdef->address,
985
                                     prefix,
986
                                     forwardIf,
987 988
                                     "tcp") < 0) {
        networkReportError(VIR_ERR_SYSTEM_ERROR,
989 990 991 992
                           forwardIf ?
                           _("failed to add iptables rule to enable TCP masquerading to %s") :
                           _("failed to add iptables rule to enable TCP masquerading"),
                           forwardIf);
993 994 995
        goto masqerr5;
    }

996
    return 0;
997

998 999
 masqerr5:
    iptablesRemoveForwardMasquerade(driver->iptables,
1000
                                    &ipdef->address,
1001
                                    prefix,
1002
                                    forwardIf,
1003 1004 1005
                                    "udp");
 masqerr4:
    iptablesRemoveForwardMasquerade(driver->iptables,
1006
                                    &ipdef->address,
1007
                                    prefix,
1008
                                    forwardIf,
1009
                                    NULL);
1010 1011
 masqerr3:
    iptablesRemoveForwardAllowRelatedIn(driver->iptables,
1012
                                        &ipdef->address,
1013
                                        prefix,
1014
                                        network->def->bridge,
1015
                                        forwardIf);
1016 1017
 masqerr2:
    iptablesRemoveForwardAllowOut(driver->iptables,
1018
                                  &ipdef->address,
1019
                                  prefix,
1020
                                  network->def->bridge,
1021
                                  forwardIf);
1022
 masqerr1:
1023
    return -1;
1024 1025
}

1026 1027 1028 1029 1030 1031
static void
networkRemoveMasqueradingIptablesRules(struct network_driver *driver,
                                       virNetworkObjPtr network,
                                       virNetworkIpDefPtr ipdef)
{
    int prefix = virNetworkIpDefPrefix(ipdef);
1032
    const char *forwardIf = virNetworkDefForwardIf(network->def, 0);
1033 1034 1035 1036 1037

    if (prefix >= 0) {
        iptablesRemoveForwardMasquerade(driver->iptables,
                                        &ipdef->address,
                                        prefix,
1038
                                        forwardIf,
1039 1040 1041 1042
                                        "tcp");
        iptablesRemoveForwardMasquerade(driver->iptables,
                                        &ipdef->address,
                                        prefix,
1043
                                        forwardIf,
1044 1045 1046 1047
                                        "udp");
        iptablesRemoveForwardMasquerade(driver->iptables,
                                        &ipdef->address,
                                        prefix,
1048
                                        forwardIf,
1049 1050 1051 1052 1053 1054
                                        NULL);

        iptablesRemoveForwardAllowRelatedIn(driver->iptables,
                                            &ipdef->address,
                                            prefix,
                                            network->def->bridge,
1055
                                            forwardIf);
1056 1057 1058 1059
        iptablesRemoveForwardAllowOut(driver->iptables,
                                      &ipdef->address,
                                      prefix,
                                      network->def->bridge,
1060
                                      forwardIf);
1061 1062 1063
    }
}

1064
static int
1065
networkAddRoutingIptablesRules(struct network_driver *driver,
1066
                               virNetworkObjPtr network,
1067 1068
                               virNetworkIpDefPtr ipdef)
{
1069
    int prefix = virNetworkIpDefPrefix(ipdef);
1070
    const char *forwardIf = virNetworkDefForwardIf(network->def, 0);
1071 1072 1073 1074 1075 1076 1077

    if (prefix < 0) {
        networkReportError(VIR_ERR_INTERNAL_ERROR,
                           _("Invalid prefix or netmask for '%s'"),
                           network->def->bridge);
        goto routeerr1;
    }
1078

1079
    /* allow routing packets from the bridge interface */
1080
    if (iptablesAddForwardAllowOut(driver->iptables,
1081
                                   &ipdef->address,
1082
                                   prefix,
1083
                                   network->def->bridge,
1084
                                   forwardIf) < 0) {
1085 1086 1087
        networkReportError(VIR_ERR_SYSTEM_ERROR,
                           _("failed to add iptables rule to allow routing from '%s'"),
                           network->def->bridge);
1088 1089 1090 1091
        goto routeerr1;
    }

    /* allow routing packets to the bridge interface */
1092
    if (iptablesAddForwardAllowIn(driver->iptables,
1093
                                  &ipdef->address,
1094
                                  prefix,
1095
                                  network->def->bridge,
1096
                                  forwardIf) < 0) {
1097 1098 1099
        networkReportError(VIR_ERR_SYSTEM_ERROR,
                           _("failed to add iptables rule to allow routing to '%s'"),
                           network->def->bridge);
1100 1101 1102
        goto routeerr2;
    }

1103
    return 0;
1104

1105
routeerr2:
1106
    iptablesRemoveForwardAllowOut(driver->iptables,
1107
                                  &ipdef->address,
1108
                                  prefix,
1109
                                  network->def->bridge,
1110
                                  forwardIf);
1111
routeerr1:
1112
    return -1;
1113 1114
}

1115 1116 1117 1118 1119 1120
static void
networkRemoveRoutingIptablesRules(struct network_driver *driver,
                                  virNetworkObjPtr network,
                                  virNetworkIpDefPtr ipdef)
{
    int prefix = virNetworkIpDefPrefix(ipdef);
1121
    const char *forwardIf = virNetworkDefForwardIf(network->def, 0);
1122 1123 1124 1125 1126 1127

    if (prefix >= 0) {
        iptablesRemoveForwardAllowIn(driver->iptables,
                                     &ipdef->address,
                                     prefix,
                                     network->def->bridge,
1128
                                     forwardIf);
1129 1130 1131 1132 1133

        iptablesRemoveForwardAllowOut(driver->iptables,
                                      &ipdef->address,
                                      prefix,
                                      network->def->bridge,
1134
                                      forwardIf);
1135 1136 1137
    }
}

1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173
/* Add all once/network rules required for IPv6 (if any IPv6 addresses are defined) */
static int
networkAddGeneralIp6tablesRules(struct network_driver *driver,
                               virNetworkObjPtr network)
{

    if (!virNetworkDefGetIpByIndex(network->def, AF_INET6, 0))
        return 0;

    /* Catch all rules to block forwarding to/from bridges */

    if (iptablesAddForwardRejectOut(driver->iptables, AF_INET6,
                                    network->def->bridge) < 0) {
        networkReportError(VIR_ERR_SYSTEM_ERROR,
                           _("failed to add ip6tables rule to block outbound traffic from '%s'"),
                           network->def->bridge);
        goto err1;
    }

    if (iptablesAddForwardRejectIn(driver->iptables, AF_INET6,
                                   network->def->bridge) < 0) {
        networkReportError(VIR_ERR_SYSTEM_ERROR,
                           _("failed to add ip6tables rule to block inbound traffic to '%s'"),
                           network->def->bridge);
        goto err2;
    }

    /* Allow traffic between guests on the same bridge */
    if (iptablesAddForwardAllowCross(driver->iptables, AF_INET6,
                                     network->def->bridge) < 0) {
        networkReportError(VIR_ERR_SYSTEM_ERROR,
                           _("failed to add ip6tables rule to allow cross bridge traffic on '%s'"),
                           network->def->bridge);
        goto err3;
    }

1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190
    /* allow DNS over IPv6 */
    if (iptablesAddTcpInput(driver->iptables, AF_INET6,
                            network->def->bridge, 53) < 0) {
        networkReportError(VIR_ERR_SYSTEM_ERROR,
                           _("failed to add ip6tables rule to allow DNS requests from '%s'"),
                           network->def->bridge);
        goto err4;
    }

    if (iptablesAddUdpInput(driver->iptables, AF_INET6,
                            network->def->bridge, 53) < 0) {
        networkReportError(VIR_ERR_SYSTEM_ERROR,
                           _("failed to add ip6tables rule to allow DNS requests from '%s'"),
                           network->def->bridge);
        goto err5;
    }

1191 1192 1193
    return 0;

    /* unwind in reverse order from the point of failure */
1194 1195 1196 1197
err5:
    iptablesRemoveTcpInput(driver->iptables, AF_INET6, network->def->bridge, 53);
err4:
    iptablesRemoveForwardAllowCross(driver->iptables, AF_INET6, network->def->bridge);
1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217
err3:
    iptablesRemoveForwardRejectIn(driver->iptables, AF_INET6, network->def->bridge);
err2:
    iptablesRemoveForwardRejectOut(driver->iptables, AF_INET6, network->def->bridge);
err1:
    return -1;
}

static void
networkRemoveGeneralIp6tablesRules(struct network_driver *driver,
                                  virNetworkObjPtr network)
{
    if (!virNetworkDefGetIpByIndex(network->def, AF_INET6, 0))
        return;

    iptablesRemoveForwardAllowCross(driver->iptables, AF_INET6, network->def->bridge);
    iptablesRemoveForwardRejectIn(driver->iptables, AF_INET6, network->def->bridge);
    iptablesRemoveForwardRejectOut(driver->iptables, AF_INET6, network->def->bridge);
}

1218
static int
1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232
networkAddGeneralIptablesRules(struct network_driver *driver,
                               virNetworkObjPtr network)
{
    int ii;
    virNetworkIpDefPtr ipv4def;

    /* First look for first IPv4 address that has dhcp or tftpboot defined. */
    /* We support dhcp config on 1 IPv4 interface only. */
    for (ii = 0;
         (ipv4def = virNetworkDefGetIpByIndex(network->def, AF_INET, ii));
         ii++) {
        if (ipv4def->nranges || ipv4def->nhosts || ipv4def->tftproot)
            break;
    }
1233 1234

    /* allow DHCP requests through to dnsmasq */
1235

1236 1237
    if (iptablesAddTcpInput(driver->iptables, AF_INET,
                            network->def->bridge, 67) < 0) {
1238 1239 1240
        networkReportError(VIR_ERR_SYSTEM_ERROR,
                           _("failed to add iptables rule to allow DHCP requests from '%s'"),
                           network->def->bridge);
1241 1242 1243
        goto err1;
    }

1244 1245
    if (iptablesAddUdpInput(driver->iptables, AF_INET,
                            network->def->bridge, 67) < 0) {
1246 1247 1248
        networkReportError(VIR_ERR_SYSTEM_ERROR,
                           _("failed to add iptables rule to allow DHCP requests from '%s'"),
                           network->def->bridge);
1249 1250 1251
        goto err2;
    }

1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262
    /* If we are doing local DHCP service on this network, attempt to
     * add a rule that will fixup the checksum of DHCP response
     * packets back to the guests (but report failure without
     * aborting, since not all iptables implementations support it).
     */

    if (ipv4def && (ipv4def->nranges || ipv4def->nhosts) &&
        (iptablesAddOutputFixUdpChecksum(driver->iptables,
                                         network->def->bridge, 68) < 0)) {
        VIR_WARN("Could not add rule to fixup DHCP response checksums "
                 "on network '%s'.", network->def->name);
1263
        VIR_WARN("May need to update iptables package & kernel to support CHECKSUM rule.");
1264 1265
    }

1266
    /* allow DNS requests through to dnsmasq */
1267 1268
    if (iptablesAddTcpInput(driver->iptables, AF_INET,
                            network->def->bridge, 53) < 0) {
1269 1270 1271
        networkReportError(VIR_ERR_SYSTEM_ERROR,
                           _("failed to add iptables rule to allow DNS requests from '%s'"),
                           network->def->bridge);
1272 1273 1274
        goto err3;
    }

1275 1276
    if (iptablesAddUdpInput(driver->iptables, AF_INET,
                            network->def->bridge, 53) < 0) {
1277 1278 1279
        networkReportError(VIR_ERR_SYSTEM_ERROR,
                           _("failed to add iptables rule to allow DNS requests from '%s'"),
                           network->def->bridge);
1280 1281 1282
        goto err4;
    }

1283 1284
    /* allow TFTP requests through to dnsmasq if necessary */
    if (ipv4def && ipv4def->tftproot &&
1285 1286
        iptablesAddUdpInput(driver->iptables, AF_INET,
                            network->def->bridge, 69) < 0) {
1287 1288 1289
        networkReportError(VIR_ERR_SYSTEM_ERROR,
                           _("failed to add iptables rule to allow TFTP requests from '%s'"),
                           network->def->bridge);
1290
        goto err5;
1291 1292
    }

1293 1294
    /* Catch all rules to block forwarding to/from bridges */

1295 1296
    if (iptablesAddForwardRejectOut(driver->iptables, AF_INET,
                                    network->def->bridge) < 0) {
1297 1298 1299
        networkReportError(VIR_ERR_SYSTEM_ERROR,
                           _("failed to add iptables rule to block outbound traffic from '%s'"),
                           network->def->bridge);
1300
        goto err6;
1301 1302
    }

1303 1304
    if (iptablesAddForwardRejectIn(driver->iptables, AF_INET,
                                   network->def->bridge) < 0) {
1305 1306 1307
        networkReportError(VIR_ERR_SYSTEM_ERROR,
                           _("failed to add iptables rule to block inbound traffic to '%s'"),
                           network->def->bridge);
1308
        goto err7;
1309 1310 1311
    }

    /* Allow traffic between guests on the same bridge */
1312 1313
    if (iptablesAddForwardAllowCross(driver->iptables, AF_INET,
                                     network->def->bridge) < 0) {
1314 1315 1316
        networkReportError(VIR_ERR_SYSTEM_ERROR,
                           _("failed to add iptables rule to allow cross bridge traffic on '%s'"),
                           network->def->bridge);
1317
        goto err8;
1318 1319
    }

1320 1321 1322 1323 1324
    /* add IPv6 general rules, if needed */
    if (networkAddGeneralIp6tablesRules(driver, network) < 0) {
        goto err9;
    }

1325
    return 0;
1326

1327
    /* unwind in reverse order from the point of failure */
1328 1329
err9:
    iptablesRemoveForwardAllowCross(driver->iptables, AF_INET, network->def->bridge);
1330
err8:
1331
    iptablesRemoveForwardRejectIn(driver->iptables, AF_INET, network->def->bridge);
1332
err7:
1333
    iptablesRemoveForwardRejectOut(driver->iptables, AF_INET, network->def->bridge);
1334 1335
err6:
    if (ipv4def && ipv4def->tftproot) {
1336
        iptablesRemoveUdpInput(driver->iptables, AF_INET, network->def->bridge, 69);
1337
    }
1338
err5:
1339
    iptablesRemoveUdpInput(driver->iptables, AF_INET, network->def->bridge, 53);
1340
err4:
1341
    iptablesRemoveTcpInput(driver->iptables, AF_INET, network->def->bridge, 53);
1342
err3:
1343
    iptablesRemoveUdpInput(driver->iptables, AF_INET, network->def->bridge, 67);
1344
err2:
1345
    iptablesRemoveTcpInput(driver->iptables, AF_INET, network->def->bridge, 67);
1346
err1:
1347
    return -1;
1348 1349 1350
}

static void
1351 1352 1353 1354 1355
networkRemoveGeneralIptablesRules(struct network_driver *driver,
                                  virNetworkObjPtr network)
{
    int ii;
    virNetworkIpDefPtr ipv4def;
1356

1357 1358
    networkRemoveGeneralIp6tablesRules(driver, network);

1359 1360 1361 1362 1363
    for (ii = 0;
         (ipv4def = virNetworkDefGetIpByIndex(network->def, AF_INET, ii));
         ii++) {
        if (ipv4def->nranges || ipv4def->nhosts || ipv4def->tftproot)
            break;
1364
    }
1365

1366 1367 1368
    iptablesRemoveForwardAllowCross(driver->iptables, AF_INET, network->def->bridge);
    iptablesRemoveForwardRejectIn(driver->iptables, AF_INET, network->def->bridge);
    iptablesRemoveForwardRejectOut(driver->iptables, AF_INET, network->def->bridge);
1369
    if (ipv4def && ipv4def->tftproot) {
1370
        iptablesRemoveUdpInput(driver->iptables, AF_INET, network->def->bridge, 69);
1371
    }
1372 1373
    iptablesRemoveUdpInput(driver->iptables, AF_INET, network->def->bridge, 53);
    iptablesRemoveTcpInput(driver->iptables, AF_INET, network->def->bridge, 53);
1374 1375 1376 1377
    if (ipv4def && (ipv4def->nranges || ipv4def->nhosts)) {
        iptablesRemoveOutputFixUdpChecksum(driver->iptables,
                                           network->def->bridge, 68);
    }
1378 1379
    iptablesRemoveUdpInput(driver->iptables, AF_INET, network->def->bridge, 67);
    iptablesRemoveTcpInput(driver->iptables, AF_INET, network->def->bridge, 67);
1380 1381
}

1382 1383 1384 1385 1386
static int
networkAddIpSpecificIptablesRules(struct network_driver *driver,
                                  virNetworkObjPtr network,
                                  virNetworkIpDefPtr ipdef)
{
1387 1388 1389
    /* NB: in the case of IPv6, routing rules are added when the
     * forward mode is NAT. This is because IPv6 has no NAT.
     */
1390

1391
    if (network->def->forwardType == VIR_NETWORK_FORWARD_NAT) {
1392
        if (VIR_SOCKET_ADDR_IS_FAMILY(&ipdef->address, AF_INET))
1393
            return networkAddMasqueradingIptablesRules(driver, network, ipdef);
1394
        else if (VIR_SOCKET_ADDR_IS_FAMILY(&ipdef->address, AF_INET6))
1395 1396 1397 1398
            return networkAddRoutingIptablesRules(driver, network, ipdef);
    } else if (network->def->forwardType == VIR_NETWORK_FORWARD_ROUTE) {
        return networkAddRoutingIptablesRules(driver, network, ipdef);
    }
1399 1400 1401 1402 1403 1404 1405 1406
    return 0;
}

static void
networkRemoveIpSpecificIptablesRules(struct network_driver *driver,
                                     virNetworkObjPtr network,
                                     virNetworkIpDefPtr ipdef)
{
1407
    if (network->def->forwardType == VIR_NETWORK_FORWARD_NAT) {
1408
        if (VIR_SOCKET_ADDR_IS_FAMILY(&ipdef->address, AF_INET))
1409
            networkRemoveMasqueradingIptablesRules(driver, network, ipdef);
1410
        else if (VIR_SOCKET_ADDR_IS_FAMILY(&ipdef->address, AF_INET6))
1411 1412
            networkRemoveRoutingIptablesRules(driver, network, ipdef);
    } else if (network->def->forwardType == VIR_NETWORK_FORWARD_ROUTE) {
1413
        networkRemoveRoutingIptablesRules(driver, network, ipdef);
1414
    }
1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467
}

/* Add all rules for all ip addresses (and general rules) on a network */
static int
networkAddIptablesRules(struct network_driver *driver,
                        virNetworkObjPtr network)
{
    int ii;
    virNetworkIpDefPtr ipdef;

    /* Add "once per network" rules */
    if (networkAddGeneralIptablesRules(driver, network) < 0)
        return -1;

    for (ii = 0;
         (ipdef = virNetworkDefGetIpByIndex(network->def, AF_UNSPEC, ii));
         ii++) {
        /* Add address-specific iptables rules */
        if (networkAddIpSpecificIptablesRules(driver, network, ipdef) < 0) {
            goto err;
        }
    }
    return 0;

err:
    /* The final failed call to networkAddIpSpecificIptablesRules will
     * have removed any rules it created, but we need to remove those
     * added for previous IP addresses.
     */
    while ((--ii >= 0) &&
           (ipdef = virNetworkDefGetIpByIndex(network->def, AF_UNSPEC, ii))) {
        networkRemoveIpSpecificIptablesRules(driver, network, ipdef);
    }
    networkRemoveGeneralIptablesRules(driver, network);
    return -1;
}

/* Remove all rules for all ip addresses (and general rules) on a network */
static void
networkRemoveIptablesRules(struct network_driver *driver,
                           virNetworkObjPtr network)
{
    int ii;
    virNetworkIpDefPtr ipdef;

    for (ii = 0;
         (ipdef = virNetworkDefGetIpByIndex(network->def, AF_UNSPEC, ii));
         ii++) {
        networkRemoveIpSpecificIptablesRules(driver, network, ipdef);
    }
    networkRemoveGeneralIptablesRules(driver, network);
}

1468 1469 1470 1471 1472
static void
networkReloadIptablesRules(struct network_driver *driver)
{
    unsigned int i;

1473
    VIR_INFO("Reloading iptables rules");
1474 1475 1476 1477

    for (i = 0 ; i < driver->networks.count ; i++) {
        virNetworkObjLock(driver->networks.objs[i]);
        if (virNetworkObjIsActive(driver->networks.objs[i])) {
1478 1479 1480 1481
            networkRemoveIptablesRules(driver, driver->networks.objs[i]);
            if (networkAddIptablesRules(driver, driver->networks.objs[i]) < 0) {
                /* failed to add but already logged */
            }
1482 1483 1484 1485 1486
        }
        virNetworkObjUnlock(driver->networks.objs[i]);
    }
}

1487
/* Enable IP Forwarding. Return 0 for success, -1 for failure. */
1488
static int
1489
networkEnableIpForwarding(bool enableIPv4, bool enableIPv6)
1490
{
1491 1492 1493 1494 1495 1496
    int ret = 0;
    if (enableIPv4)
        ret = virFileWriteStr("/proc/sys/net/ipv4/ip_forward", "1\n", 0);
    if (enableIPv6 && ret == 0)
        ret = virFileWriteStr("/proc/sys/net/ipv6/conf/all/forwarding", "1\n", 0);
    return ret;
1497 1498
}

1499 1500
#define SYSCTL_PATH "/proc/sys"

1501 1502
static int
networkSetIPv6Sysctls(virNetworkObjPtr network)
1503 1504 1505 1506
{
    char *field = NULL;
    int ret = -1;

1507 1508 1509 1510 1511 1512 1513 1514 1515
    if (!virNetworkDefGetIpByIndex(network->def, AF_INET6, 0)) {
        /* Only set disable_ipv6 if there are no ipv6 addresses defined for
         * the network.
         */
        if (virAsprintf(&field, SYSCTL_PATH "/net/ipv6/conf/%s/disable_ipv6",
                        network->def->bridge) < 0) {
            virReportOOMError();
            goto cleanup;
        }
1516

1517 1518 1519 1520 1521 1522
        if (access(field, W_OK) < 0 && errno == ENOENT) {
            VIR_DEBUG("ipv6 appears to already be disabled on %s",
                      network->def->bridge);
            ret = 0;
            goto cleanup;
        }
1523

1524 1525 1526 1527 1528 1529 1530
        if (virFileWriteStr(field, "1", 0) < 0) {
            virReportSystemError(errno,
                                 _("cannot write to %s to disable IPv6 on bridge %s"),
                                 field, network->def->bridge);
            goto cleanup;
        }
        VIR_FREE(field);
1531 1532
    }

1533 1534 1535 1536 1537 1538 1539 1540 1541
    /* The rest of the ipv6 sysctl tunables should always be set,
     * whether or not we're using ipv6 on this bridge.
     */

    /* Prevent guests from hijacking the host network by sending out
     * their own router advertisements.
     */
    if (virAsprintf(&field, SYSCTL_PATH "/net/ipv6/conf/%s/accept_ra",
                    network->def->bridge) < 0) {
1542
        virReportOOMError();
1543 1544 1545
        goto cleanup;
    }

1546
    if (virFileWriteStr(field, "0", 0) < 0) {
1547
        virReportSystemError(errno,
1548 1549 1550 1551 1552
                             _("cannot disable %s"), field);
        goto cleanup;
    }
    VIR_FREE(field);

1553 1554 1555 1556 1557
    /* All interfaces used as a gateway (which is what this is, by
     * definition), must always have autoconf=0.
     */
    if (virAsprintf(&field, SYSCTL_PATH "/net/ipv6/conf/%s/autoconf",
                    network->def->bridge) < 0) {
1558
        virReportOOMError();
1559 1560 1561
        goto cleanup;
    }

1562
    if (virFileWriteStr(field, "0", 0) < 0) {
1563
        virReportSystemError(errno,
1564
                             _("cannot disable %s"), field);
1565 1566 1567 1568 1569 1570 1571 1572 1573
        goto cleanup;
    }

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

1574 1575 1576 1577 1578 1579
#define PROC_NET_ROUTE "/proc/net/route"

/* XXX: This function can be a lot more exhaustive, there are certainly
 *      other scenarios where we can ruin host network connectivity.
 * XXX: Using a proper library is preferred over parsing /proc
 */
1580 1581
static int
networkCheckRouteCollision(virNetworkObjPtr network)
1582
{
1583
    int ret = 0, len;
1584 1585 1586 1587 1588
    char *cur, *buf = NULL;
    enum {MAX_ROUTE_SIZE = 1024*64};

    /* Read whole routing table into memory */
    if ((len = virFileReadAll(PROC_NET_ROUTE, MAX_ROUTE_SIZE, &buf)) < 0)
1589
        goto out;
1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607

    /* Dropping the last character shouldn't hurt */
    if (len > 0)
        buf[len-1] = '\0';

    VIR_DEBUG("%s output:\n%s", PROC_NET_ROUTE, buf);

    if (!STRPREFIX (buf, "Iface"))
        goto out;

    /* First line is just headings, skip it */
    cur = strchr(buf, '\n');
    if (cur)
        cur++;

    while (cur) {
        char iface[17], dest[128], mask[128];
        unsigned int addr_val, mask_val;
1608 1609
        virNetworkIpDefPtr ipdef;
        int num, ii;
1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637

        /* NUL-terminate the line, so sscanf doesn't go beyond a newline.  */
        char *nl = strchr(cur, '\n');
        if (nl) {
            *nl++ = '\0';
        }

        num = sscanf(cur, "%16s %127s %*s %*s %*s %*s %*s %127s",
                     iface, dest, mask);
        cur = nl;

        if (num != 3) {
            VIR_DEBUG("Failed to parse %s", PROC_NET_ROUTE);
            continue;
        }

        if (virStrToLong_ui(dest, NULL, 16, &addr_val) < 0) {
            VIR_DEBUG("Failed to convert network address %s to uint", dest);
            continue;
        }

        if (virStrToLong_ui(mask, NULL, 16, &mask_val) < 0) {
            VIR_DEBUG("Failed to convert network mask %s to uint", mask);
            continue;
        }

        addr_val &= mask_val;

1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661
        for (ii = 0;
             (ipdef = virNetworkDefGetIpByIndex(network->def, AF_INET, ii));
             ii++) {

            unsigned int net_dest;
            virSocketAddr netmask;

            if (virNetworkIpDefNetmask(ipdef, &netmask) < 0) {
                VIR_WARN("Failed to get netmask of '%s'",
                         network->def->bridge);
                continue;
            }

            net_dest = (ipdef->address.data.inet4.sin_addr.s_addr &
                        netmask.data.inet4.sin_addr.s_addr);

            if ((net_dest == addr_val) &&
                (netmask.data.inet4.sin_addr.s_addr == mask_val)) {
                networkReportError(VIR_ERR_INTERNAL_ERROR,
                                   _("Network is already in use by interface %s"),
                                   iface);
                ret = -1;
                goto out;
            }
1662 1663 1664 1665 1666 1667 1668 1669
        }
    }

out:
    VIR_FREE(buf);
    return ret;
}

1670
static int
D
Daniel P. Berrange 已提交
1671
networkAddAddrToBridge(virNetworkObjPtr network,
1672
                       virNetworkIpDefPtr ipdef)
1673
{
1674 1675 1676 1677 1678 1679 1680 1681 1682
    int prefix = virNetworkIpDefPrefix(ipdef);

    if (prefix < 0) {
        networkReportError(VIR_ERR_INTERNAL_ERROR,
                           _("bridge '%s' has an invalid netmask or IP address"),
                           network->def->bridge);
        return -1;
    }

1683 1684
    if (virNetDevSetIPv4Address(network->def->bridge,
                                &ipdef->address, prefix) < 0)
1685 1686 1687 1688 1689 1690
        return -1;

    return 0;
}

static int
1691
networkStartNetworkVirtual(struct network_driver *driver,
1692 1693
                          virNetworkObjPtr network)
{
1694
    int ii;
1695
    bool v4present = false, v6present = false;
1696 1697
    virErrorPtr save_err = NULL;
    virNetworkIpDefPtr ipdef;
1698
    char *macTapIfName = NULL;
1699

1700 1701
    /* Check to see if any network IP collides with an existing route */
    if (networkCheckRouteCollision(network) < 0)
1702 1703
        return -1;

1704
    /* Create and configure the bridge device */
1705
    if (virNetDevBridgeCreate(network->def->bridge) < 0)
1706 1707
        return -1;

1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719
    if (network->def->mac_specified) {
        /* To set a mac for the bridge, we need to define a dummy tap
         * device, set its mac, then attach it to the bridge. As long
         * as its mac address is lower than any other interface that
         * gets attached, the bridge will always maintain this mac
         * address.
         */
        macTapIfName = networkBridgeDummyNicName(network->def->bridge);
        if (!macTapIfName) {
            virReportOOMError();
            goto err0;
        }
1720 1721
        if (virNetDevTapCreateInBridgePort(network->def->bridge,
                                           &macTapIfName, network->def->mac, 0, false, NULL) < 0) {
1722 1723 1724 1725 1726
            VIR_FREE(macTapIfName);
            goto err0;
        }
    }

1727
    /* Set bridge options */
1728
    if (virNetDevBridgeSetSTPDelay(network->def->bridge,
1729
                          network->def->delay) < 0)
1730
        goto err1;
1731

1732
    if (virNetDevBridgeSetSTP(network->def->bridge,
1733
                              network->def->stp ? true : false) < 0)
1734
        goto err1;
1735

1736 1737 1738 1739
    /* Disable IPv6 on the bridge if there are no IPv6 addresses
     * defined, and set other IPv6 sysctl tunables appropriately.
     */
    if (networkSetIPv6Sysctls(network) < 0)
1740
        goto err1;
1741

1742 1743 1744 1745 1746 1747 1748
    /* Add "once per network" rules */
    if (networkAddIptablesRules(driver, network) < 0)
        goto err1;

    for (ii = 0;
         (ipdef = virNetworkDefGetIpByIndex(network->def, AF_UNSPEC, ii));
         ii++) {
1749
        if (VIR_SOCKET_ADDR_IS_FAMILY(&ipdef->address, AF_INET))
1750
            v4present = true;
1751
        if (VIR_SOCKET_ADDR_IS_FAMILY(&ipdef->address, AF_INET6))
1752
            v6present = true;
1753

1754
        /* Add the IP address/netmask to the bridge */
D
Daniel P. Berrange 已提交
1755
        if (networkAddAddrToBridge(network, ipdef) < 0) {
1756
            goto err2;
1757
        }
1758 1759
    }

1760
    /* Bring up the bridge interface */
1761
    if (virNetDevSetOnline(network->def->bridge, 1) < 0)
1762
        goto err2;
1763

1764
    /* If forwardType != NONE, turn on global IP forwarding */
1765
    if (network->def->forwardType != VIR_NETWORK_FORWARD_NONE &&
1766
        networkEnableIpForwarding(v4present, v6present) < 0) {
1767
        virReportSystemError(errno, "%s",
1768
                             _("failed to enable IP forwarding"));
1769
        goto err3;
1770 1771
    }

1772

1773 1774
    /* start dnsmasq if there are any IP addresses (v4 or v6) */
    if ((v4present || v6present) && networkStartDhcpDaemon(network) < 0)
1775
        goto err3;
1776

1777 1778 1779 1780
    /* start radvd if there are any ipv6 addresses */
    if (v6present && networkStartRadvd(network) < 0)
        goto err4;

1781 1782 1783 1784 1785 1786 1787
    if (virBandwidthEnable(network->def->bandwidth, network->def->bridge) < 0) {
        networkReportError(VIR_ERR_INTERNAL_ERROR,
                           _("cannot set bandwidth limits on %s"),
                           network->def->bridge);
        goto err5;
    }

1788
    VIR_FREE(macTapIfName);
1789 1790 1791

    return 0;

1792 1793 1794 1795 1796 1797
 err5:
    if (virBandwidthDisable(network->def->bridge, true) < 0) {
        VIR_WARN("Failed to disable QoS on %s",
                 network->def->bridge);
    }

1798 1799 1800 1801
 err4:
    if (!save_err)
        save_err = virSaveLastError();

1802 1803 1804 1805 1806
    if (network->dnsmasqPid > 0) {
        kill(network->dnsmasqPid, SIGTERM);
        network->dnsmasqPid = -1;
    }

1807 1808 1809
 err3:
    if (!save_err)
        save_err = virSaveLastError();
1810
    ignore_value(virNetDevSetOnline(network->def->bridge, 0));
1811

1812 1813 1814 1815 1816 1817
 err2:
    if (!save_err)
        save_err = virSaveLastError();
    networkRemoveIptablesRules(driver, network);

 err1:
1818 1819 1820
    if (!save_err)
        save_err = virSaveLastError();

1821
    ignore_value(virNetDevTapDelete(macTapIfName));
1822
    VIR_FREE(macTapIfName);
1823 1824

 err0:
1825 1826
    if (!save_err)
        save_err = virSaveLastError();
1827
    ignore_value(virNetDevBridgeDelete(network->def->bridge));
1828

1829 1830 1831 1832
    if (save_err) {
        virSetError(save_err);
        virFreeError(save_err);
    }
1833 1834 1835
    return -1;
}

1836
static int networkShutdownNetworkVirtual(struct network_driver *driver,
1837 1838
                                        virNetworkObjPtr network)
{
1839 1840 1841 1842 1843
    if (virBandwidthDisable(network->def->bridge, true) < 0) {
        VIR_WARN("Failed to disable QoS on %s",
                 network->def->name);
    }

1844 1845 1846 1847 1848 1849 1850 1851
    if (network->radvdPid > 0) {
        char *radvdpidbase;

        kill(network->radvdPid, SIGTERM);
        /* attempt to delete the pidfile we created */
        if (!(radvdpidbase = networkRadvdPidfileBasename(network->def->name))) {
            virReportOOMError();
        } else {
1852
            virPidFileDelete(NETWORK_PID_DIR, radvdpidbase);
1853 1854 1855 1856
            VIR_FREE(radvdpidbase);
        }
    }

1857 1858 1859
    if (network->dnsmasqPid > 0)
        kill(network->dnsmasqPid, SIGTERM);

1860
    if (network->def->mac_specified) {
1861
        char *macTapIfName = networkBridgeDummyNicName(network->def->bridge);
1862 1863 1864
        if (!macTapIfName) {
            virReportOOMError();
        } else {
1865
            ignore_value(virNetDevTapDelete(macTapIfName));
1866 1867 1868 1869
            VIR_FREE(macTapIfName);
        }
    }

1870
    ignore_value(virNetDevSetOnline(network->def->bridge, 0));
1871

1872 1873
    networkRemoveIptablesRules(driver, network);

1874
    ignore_value(virNetDevBridgeDelete(network->def->bridge));
1875

1876
    /* See if its still alive and really really kill it */
1877
    if (network->dnsmasqPid > 0 &&
1878
        (kill(network->dnsmasqPid, 0) == 0))
1879 1880
        kill(network->dnsmasqPid, SIGKILL);
    network->dnsmasqPid = -1;
1881 1882 1883 1884 1885 1886

    if (network->radvdPid > 0 &&
        (kill(network->radvdPid, 0) == 0))
        kill(network->radvdPid, SIGKILL);
    network->radvdPid = -1;

1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999
    return 0;
}

static int
networkStartNetworkExternal(struct network_driver *driver ATTRIBUTE_UNUSED,
                            virNetworkObjPtr network ATTRIBUTE_UNUSED)
{
    /* put anything here that needs to be done each time a network of
     * type BRIDGE, PRIVATE, VEPA, or PASSTHROUGH is started. On
     * failure, undo anything you've done, and return -1. On success
     * return 0.
     */
    return 0;
}

static int networkShutdownNetworkExternal(struct network_driver *driver ATTRIBUTE_UNUSED,
                                        virNetworkObjPtr network ATTRIBUTE_UNUSED)
{
    /* put anything here that needs to be done each time a network of
     * type BRIDGE, PRIVATE, VEPA, or PASSTHROUGH is shutdown. On
     * failure, undo anything you've done, and return -1. On success
     * return 0.
     */
    return 0;
}

static int
networkStartNetwork(struct network_driver *driver,
                    virNetworkObjPtr network)
{
    int ret = 0;

    if (virNetworkObjIsActive(network)) {
        networkReportError(VIR_ERR_OPERATION_INVALID,
                           "%s", _("network is already active"));
        return -1;
    }

    switch (network->def->forwardType) {

    case VIR_NETWORK_FORWARD_NONE:
    case VIR_NETWORK_FORWARD_NAT:
    case VIR_NETWORK_FORWARD_ROUTE:
        ret = networkStartNetworkVirtual(driver, network);
        break;

    case VIR_NETWORK_FORWARD_BRIDGE:
    case VIR_NETWORK_FORWARD_PRIVATE:
    case VIR_NETWORK_FORWARD_VEPA:
    case VIR_NETWORK_FORWARD_PASSTHROUGH:
        ret = networkStartNetworkExternal(driver, network);
        break;
    }

    if (ret < 0)
        return ret;

    /* Persist the live configuration now that anything autogenerated
     * is setup.
     */
    if ((ret = virNetworkSaveConfig(NETWORK_STATE_DIR, network->def)) < 0) {
        goto error;
    }

    VIR_INFO("Starting up network '%s'", network->def->name);
    network->active = 1;

error:
    if (ret < 0) {
        virErrorPtr save_err = virSaveLastError();
        int save_errno = errno;
        networkShutdownNetwork(driver, network);
        virSetError(save_err);
        virFreeError(save_err);
        errno = save_errno;
    }
    return ret;
}

static int networkShutdownNetwork(struct network_driver *driver,
                                        virNetworkObjPtr network)
{
    int ret = 0;
    char *stateFile;

    VIR_INFO("Shutting down network '%s'", network->def->name);

    if (!virNetworkObjIsActive(network))
        return 0;

    stateFile = virNetworkConfigFile(NETWORK_STATE_DIR, network->def->name);
    if (!stateFile)
        return -1;

    unlink(stateFile);
    VIR_FREE(stateFile);

    switch (network->def->forwardType) {

    case VIR_NETWORK_FORWARD_NONE:
    case VIR_NETWORK_FORWARD_NAT:
    case VIR_NETWORK_FORWARD_ROUTE:
        ret = networkShutdownNetworkVirtual(driver, network);
        break;

    case VIR_NETWORK_FORWARD_BRIDGE:
    case VIR_NETWORK_FORWARD_PRIVATE:
    case VIR_NETWORK_FORWARD_VEPA:
    case VIR_NETWORK_FORWARD_PASSTHROUGH:
        ret = networkShutdownNetworkExternal(driver, network);
        break;
    }

2000 2001 2002 2003 2004 2005 2006 2007
    network->active = 0;

    if (network->newDef) {
        virNetworkDefFree(network->def);
        network->def = network->newDef;
        network->newDef = NULL;
    }

2008
    return ret;
2009 2010 2011
}


2012 2013 2014 2015 2016
static virNetworkPtr networkLookupByUUID(virConnectPtr conn,
                                         const unsigned char *uuid) {
    struct network_driver *driver = conn->networkPrivateData;
    virNetworkObjPtr network;
    virNetworkPtr ret = NULL;
2017

2018
    networkDriverLock(driver);
2019
    network = virNetworkFindByUUID(&driver->networks, uuid);
2020
    networkDriverUnlock(driver);
2021
    if (!network) {
2022 2023
        networkReportError(VIR_ERR_NO_NETWORK,
                           "%s", _("no network with matching uuid"));
2024
        goto cleanup;
2025 2026
    }

2027 2028 2029
    ret = virGetNetwork(conn, network->def->name, network->def->uuid);

cleanup:
2030 2031
    if (network)
        virNetworkObjUnlock(network);
2032
    return ret;
2033 2034
}

2035 2036 2037 2038 2039 2040
static virNetworkPtr networkLookupByName(virConnectPtr conn,
                                         const char *name) {
    struct network_driver *driver = conn->networkPrivateData;
    virNetworkObjPtr network;
    virNetworkPtr ret = NULL;

2041
    networkDriverLock(driver);
2042
    network = virNetworkFindByName(&driver->networks, name);
2043
    networkDriverUnlock(driver);
2044
    if (!network) {
2045 2046
        networkReportError(VIR_ERR_NO_NETWORK,
                           _("no network with matching name '%s'"), name);
2047
        goto cleanup;
2048 2049
    }

2050 2051 2052
    ret = virGetNetwork(conn, network->def->name, network->def->uuid);

cleanup:
2053 2054
    if (network)
        virNetworkObjUnlock(network);
2055
    return ret;
2056 2057 2058 2059
}

static virDrvOpenStatus networkOpenNetwork(virConnectPtr conn,
                                           virConnectAuthPtr auth ATTRIBUTE_UNUSED,
2060 2061 2062 2063
                                           unsigned int flags)
{
    virCheckFlags(VIR_CONNECT_RO, VIR_DRV_OPEN_ERROR);

2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076
    if (!driverState)
        return VIR_DRV_OPEN_DECLINED;

    conn->networkPrivateData = driverState;
    return VIR_DRV_OPEN_SUCCESS;
}

static int networkCloseNetwork(virConnectPtr conn) {
    conn->networkPrivateData = NULL;
    return 0;
}

static int networkNumNetworks(virConnectPtr conn) {
2077
    int nactive = 0, i;
2078
    struct network_driver *driver = conn->networkPrivateData;
2079

2080 2081 2082
    networkDriverLock(driver);
    for (i = 0 ; i < driver->networks.count ; i++) {
        virNetworkObjLock(driver->networks.objs[i]);
D
Daniel P. Berrange 已提交
2083
        if (virNetworkObjIsActive(driver->networks.objs[i]))
2084
            nactive++;
2085 2086 2087
        virNetworkObjUnlock(driver->networks.objs[i]);
    }
    networkDriverUnlock(driver);
2088

2089 2090 2091 2092
    return nactive;
}

static int networkListNetworks(virConnectPtr conn, char **const names, int nnames) {
2093
    struct network_driver *driver = conn->networkPrivateData;
2094
    int got = 0, i;
2095

2096
    networkDriverLock(driver);
2097
    for (i = 0 ; i < driver->networks.count && got < nnames ; i++) {
2098
        virNetworkObjLock(driver->networks.objs[i]);
D
Daniel P. Berrange 已提交
2099
        if (virNetworkObjIsActive(driver->networks.objs[i])) {
2100
            if (!(names[got] = strdup(driver->networks.objs[i]->def->name))) {
2101
                virNetworkObjUnlock(driver->networks.objs[i]);
2102
                virReportOOMError();
2103 2104 2105 2106
                goto cleanup;
            }
            got++;
        }
2107
        virNetworkObjUnlock(driver->networks.objs[i]);
2108
    }
2109 2110
    networkDriverUnlock(driver);

2111 2112 2113
    return got;

 cleanup:
2114
    networkDriverUnlock(driver);
2115 2116 2117 2118 2119 2120
    for (i = 0 ; i < got ; i++)
        VIR_FREE(names[i]);
    return -1;
}

static int networkNumDefinedNetworks(virConnectPtr conn) {
2121
    int ninactive = 0, i;
2122
    struct network_driver *driver = conn->networkPrivateData;
2123

2124 2125 2126
    networkDriverLock(driver);
    for (i = 0 ; i < driver->networks.count ; i++) {
        virNetworkObjLock(driver->networks.objs[i]);
D
Daniel P. Berrange 已提交
2127
        if (!virNetworkObjIsActive(driver->networks.objs[i]))
2128
            ninactive++;
2129 2130 2131
        virNetworkObjUnlock(driver->networks.objs[i]);
    }
    networkDriverUnlock(driver);
2132

2133 2134 2135 2136
    return ninactive;
}

static int networkListDefinedNetworks(virConnectPtr conn, char **const names, int nnames) {
2137
    struct network_driver *driver = conn->networkPrivateData;
2138
    int got = 0, i;
2139

2140
    networkDriverLock(driver);
2141
    for (i = 0 ; i < driver->networks.count && got < nnames ; i++) {
2142
        virNetworkObjLock(driver->networks.objs[i]);
D
Daniel P. Berrange 已提交
2143
        if (!virNetworkObjIsActive(driver->networks.objs[i])) {
2144
            if (!(names[got] = strdup(driver->networks.objs[i]->def->name))) {
2145
                virNetworkObjUnlock(driver->networks.objs[i]);
2146
                virReportOOMError();
2147 2148 2149 2150
                goto cleanup;
            }
            got++;
        }
2151
        virNetworkObjUnlock(driver->networks.objs[i]);
2152
    }
2153
    networkDriverUnlock(driver);
2154 2155 2156
    return got;

 cleanup:
2157
    networkDriverUnlock(driver);
2158 2159 2160 2161 2162
    for (i = 0 ; i < got ; i++)
        VIR_FREE(names[i]);
    return -1;
}

2163 2164 2165

static int networkIsActive(virNetworkPtr net)
{
2166
    struct network_driver *driver = net->conn->networkPrivateData;
2167 2168 2169 2170 2171 2172 2173
    virNetworkObjPtr obj;
    int ret = -1;

    networkDriverLock(driver);
    obj = virNetworkFindByUUID(&driver->networks, net->uuid);
    networkDriverUnlock(driver);
    if (!obj) {
2174
        networkReportError(VIR_ERR_NO_NETWORK, NULL);
2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186
        goto cleanup;
    }
    ret = virNetworkObjIsActive(obj);

cleanup:
    if (obj)
        virNetworkObjUnlock(obj);
    return ret;
}

static int networkIsPersistent(virNetworkPtr net)
{
2187
    struct network_driver *driver = net->conn->networkPrivateData;
2188 2189 2190 2191 2192 2193 2194
    virNetworkObjPtr obj;
    int ret = -1;

    networkDriverLock(driver);
    obj = virNetworkFindByUUID(&driver->networks, net->uuid);
    networkDriverUnlock(driver);
    if (!obj) {
2195
        networkReportError(VIR_ERR_NO_NETWORK, NULL);
2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206
        goto cleanup;
    }
    ret = obj->persistent;

cleanup:
    if (obj)
        virNetworkObjUnlock(obj);
    return ret;
}


2207
static virNetworkPtr networkCreate(virConnectPtr conn, const char *xml) {
2208
    struct network_driver *driver = conn->networkPrivateData;
2209
    virNetworkDefPtr def;
2210
    virNetworkObjPtr network = NULL;
2211
    virNetworkPtr ret = NULL;
2212

2213 2214
    networkDriverLock(driver);

2215
    if (!(def = virNetworkDefParseString(xml)))
2216
        goto cleanup;
2217

2218 2219 2220
    if (virNetworkObjIsDuplicate(&driver->networks, def, 1) < 0)
        goto cleanup;

2221 2222 2223 2224 2225 2226
    /* Only the three L3 network types that are configured by libvirt
     * need to have a bridge device name / mac address provided
     */
    if (def->forwardType == VIR_NETWORK_FORWARD_NONE ||
        def->forwardType == VIR_NETWORK_FORWARD_NAT ||
        def->forwardType == VIR_NETWORK_FORWARD_ROUTE) {
2227

2228 2229 2230 2231 2232
        if (virNetworkSetBridgeName(&driver->networks, def, 1))
            goto cleanup;

        virNetworkSetBridgeMacAddr(def);
    }
2233

2234
    if (!(network = virNetworkAssignDef(&driver->networks,
2235 2236 2237
                                        def)))
        goto cleanup;
    def = NULL;
2238

2239
    if (networkStartNetwork(driver, network) < 0) {
2240 2241
        virNetworkRemoveInactive(&driver->networks,
                                 network);
2242
        network = NULL;
2243
        goto cleanup;
2244 2245
    }

2246
    VIR_INFO("Creating network '%s'", network->def->name);
2247 2248 2249 2250
    ret = virGetNetwork(conn, network->def->name, network->def->uuid);

cleanup:
    virNetworkDefFree(def);
2251 2252 2253
    if (network)
        virNetworkObjUnlock(network);
    networkDriverUnlock(driver);
2254
    return ret;
2255 2256 2257
}

static virNetworkPtr networkDefine(virConnectPtr conn, const char *xml) {
2258
    struct network_driver *driver = conn->networkPrivateData;
2259
    virNetworkIpDefPtr ipdef, ipv4def = NULL;
2260
    virNetworkDefPtr def;
2261
    bool freeDef = true;
2262
    virNetworkObjPtr network = NULL;
2263
    virNetworkPtr ret = NULL;
2264
    int ii;
2265
    dnsmasqContext* dctx = NULL;
2266

2267 2268
    networkDriverLock(driver);

2269
    if (!(def = virNetworkDefParseString(xml)))
2270
        goto cleanup;
2271

E
Eric Blake 已提交
2272
    if (virNetworkObjIsDuplicate(&driver->networks, def, 0) < 0)
2273 2274
        goto cleanup;

2275 2276 2277 2278 2279 2280
    /* Only the three L3 network types that are configured by libvirt
     * need to have a bridge device name / mac address provided
     */
    if (def->forwardType == VIR_NETWORK_FORWARD_NONE ||
        def->forwardType == VIR_NETWORK_FORWARD_NAT ||
        def->forwardType == VIR_NETWORK_FORWARD_ROUTE) {
2281

2282 2283 2284 2285 2286
        if (virNetworkSetBridgeName(&driver->networks, def, 1))
            goto cleanup;

        virNetworkSetBridgeMacAddr(def);
    }
2287

2288
    if (!(network = virNetworkAssignDef(&driver->networks,
2289 2290
                                        def)))
        goto cleanup;
2291
    freeDef = false;
2292

2293 2294
    network->persistent = 1;

2295 2296
    if (virNetworkSaveConfig(driver->networkConfigDir, def) < 0) {
        virNetworkRemoveInactive(&driver->networks, network);
2297
        network = NULL;
2298
        goto cleanup;
2299 2300
    }

2301
    /* We only support dhcp on one IPv4 address per defined network */
2302
    for (ii = 0;
2303
         (ipdef = virNetworkDefGetIpByIndex(def, AF_UNSPEC, ii));
2304
         ii++) {
2305
        if (VIR_SOCKET_ADDR_IS_FAMILY(&ipdef->address, AF_INET)) {
2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317
            if (ipdef->nranges || ipdef->nhosts) {
                if (ipv4def) {
                    networkReportError(VIR_ERR_CONFIG_UNSUPPORTED,
                                       "%s", _("Multiple dhcp sections found. dhcp is supported only for a single IPv4 address on each network"));
                    goto cleanup;
                } else {
                    ipv4def = ipdef;
                }
            }
        }
    }
    if (ipv4def) {
2318
        dctx = dnsmasqContextNew(def->name, DNSMASQ_STATE_DIR);
2319
        if (dctx == NULL ||
2320
            networkBuildDnsmasqHostsfile(dctx, ipv4def, def->dns) < 0 ||
2321
            dnsmasqSave(dctx) < 0)
2322 2323 2324
            goto cleanup;
    }

2325 2326
    VIR_INFO("Defining network '%s'", def->name);
    ret = virGetNetwork(conn, def->name, def->uuid);
2327 2328

cleanup:
2329 2330
    if (freeDef)
       virNetworkDefFree(def);
2331
    dnsmasqContextFree(dctx);
2332 2333 2334
    if (network)
        virNetworkObjUnlock(network);
    networkDriverUnlock(driver);
2335
    return ret;
2336 2337 2338
}

static int networkUndefine(virNetworkPtr net) {
2339
    struct network_driver *driver = net->conn->networkPrivateData;
2340
    virNetworkObjPtr network;
2341 2342
    virNetworkIpDefPtr ipdef;
    bool dhcp_present = false, v6present = false;
2343
    int ret = -1, ii;
2344

2345 2346
    networkDriverLock(driver);

2347
    network = virNetworkFindByUUID(&driver->networks, net->uuid);
2348
    if (!network) {
2349
        networkReportError(VIR_ERR_NO_NETWORK,
2350 2351
                           "%s", _("no network with matching uuid"));
        goto cleanup;
2352 2353
    }

D
Daniel P. Berrange 已提交
2354
    if (virNetworkObjIsActive(network)) {
2355
        networkReportError(VIR_ERR_OPERATION_INVALID,
2356 2357
                           "%s", _("network is still active"));
        goto cleanup;
2358 2359
    }

2360
    if (virNetworkDeleteConfig(driver->networkConfigDir,
2361 2362
                               driver->networkAutostartDir,
                               network) < 0)
2363
        goto cleanup;
2364

2365 2366
    /* we only support dhcp on one IPv4 address per defined network */
    for (ii = 0;
2367
         (ipdef = virNetworkDefGetIpByIndex(network->def, AF_UNSPEC, ii));
2368
         ii++) {
2369
        if (VIR_SOCKET_ADDR_IS_FAMILY(&ipdef->address, AF_INET)) {
2370 2371
            if (ipdef->nranges || ipdef->nhosts)
                dhcp_present = true;
2372
        } else if (VIR_SOCKET_ADDR_IS_FAMILY(&ipdef->address, AF_INET6)) {
2373 2374
            v6present = true;
        }
2375
    }
2376 2377

    if (dhcp_present) {
2378
        char *leasefile;
2379 2380 2381 2382 2383 2384
        dnsmasqContext *dctx = dnsmasqContextNew(network->def->name, DNSMASQ_STATE_DIR);
        if (dctx == NULL)
            goto cleanup;

        dnsmasqDelete(dctx);
        dnsmasqContextFree(dctx);
2385 2386 2387 2388 2389 2390

        leasefile = networkDnsmasqLeaseFileName(network->def->name);
        if (!leasefile)
            goto cleanup;
        unlink(leasefile);
        VIR_FREE(leasefile);
2391 2392
    }

2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408
    if (v6present) {
        char *configfile = networkRadvdConfigFileName(network->def->name);

        if (!configfile) {
            virReportOOMError();
            goto cleanup;
        }
        unlink(configfile);
        VIR_FREE(configfile);

        char *radvdpidbase = networkRadvdPidfileBasename(network->def->name);

        if (!(radvdpidbase)) {
            virReportOOMError();
            goto cleanup;
        }
2409
        virPidFileDelete(NETWORK_PID_DIR, radvdpidbase);
2410 2411 2412 2413
        VIR_FREE(radvdpidbase);

    }

2414
    VIR_INFO("Undefining network '%s'", network->def->name);
2415 2416
    virNetworkRemoveInactive(&driver->networks,
                             network);
2417
    network = NULL;
2418
    ret = 0;
2419

2420
cleanup:
2421 2422 2423
    if (network)
        virNetworkObjUnlock(network);
    networkDriverUnlock(driver);
2424
    return ret;
2425 2426 2427
}

static int networkStart(virNetworkPtr net) {
2428 2429 2430
    struct network_driver *driver = net->conn->networkPrivateData;
    virNetworkObjPtr network;
    int ret = -1;
2431

2432
    networkDriverLock(driver);
2433
    network = virNetworkFindByUUID(&driver->networks, net->uuid);
2434

2435
    if (!network) {
2436
        networkReportError(VIR_ERR_NO_NETWORK,
2437 2438
                           "%s", _("no network with matching uuid"));
        goto cleanup;
2439 2440
    }

2441
    ret = networkStartNetwork(driver, network);
2442 2443

cleanup:
2444 2445
    if (network)
        virNetworkObjUnlock(network);
2446
    networkDriverUnlock(driver);
2447
    return ret;
2448 2449 2450
}

static int networkDestroy(virNetworkPtr net) {
2451 2452 2453
    struct network_driver *driver = net->conn->networkPrivateData;
    virNetworkObjPtr network;
    int ret = -1;
2454

2455
    networkDriverLock(driver);
2456
    network = virNetworkFindByUUID(&driver->networks, net->uuid);
2457

2458
    if (!network) {
2459
        networkReportError(VIR_ERR_NO_NETWORK,
2460 2461
                           "%s", _("no network with matching uuid"));
        goto cleanup;
2462 2463
    }

D
Daniel P. Berrange 已提交
2464
    if (!virNetworkObjIsActive(network)) {
2465
        networkReportError(VIR_ERR_OPERATION_INVALID,
2466 2467 2468 2469
                           "%s", _("network is not active"));
        goto cleanup;
    }

2470
    ret = networkShutdownNetwork(driver, network);
2471
    if (!network->persistent) {
2472 2473 2474 2475
        virNetworkRemoveInactive(&driver->networks,
                                 network);
        network = NULL;
    }
2476

2477
cleanup:
2478 2479
    if (network)
        virNetworkObjUnlock(network);
2480
    networkDriverUnlock(driver);
2481 2482 2483
    return ret;
}

2484
static char *networkGetXMLDesc(virNetworkPtr net,
2485
                               unsigned int flags)
2486
{
2487 2488 2489
    struct network_driver *driver = net->conn->networkPrivateData;
    virNetworkObjPtr network;
    char *ret = NULL;
2490

2491 2492
    virCheckFlags(0, NULL);

2493
    networkDriverLock(driver);
2494
    network = virNetworkFindByUUID(&driver->networks, net->uuid);
2495 2496
    networkDriverUnlock(driver);

2497
    if (!network) {
2498
        networkReportError(VIR_ERR_NO_NETWORK,
2499 2500
                           "%s", _("no network with matching uuid"));
        goto cleanup;
2501 2502
    }

2503
    ret = virNetworkDefFormat(network->def);
2504 2505

cleanup:
2506 2507
    if (network)
        virNetworkObjUnlock(network);
2508
    return ret;
2509 2510 2511
}

static char *networkGetBridgeName(virNetworkPtr net) {
2512 2513 2514 2515
    struct network_driver *driver = net->conn->networkPrivateData;
    virNetworkObjPtr network;
    char *bridge = NULL;

2516
    networkDriverLock(driver);
2517
    network = virNetworkFindByUUID(&driver->networks, net->uuid);
2518 2519
    networkDriverUnlock(driver);

2520
    if (!network) {
2521
        networkReportError(VIR_ERR_NO_NETWORK,
2522 2523
                           "%s", _("no network with matching id"));
        goto cleanup;
2524 2525
    }

2526
    if (!(network->def->bridge)) {
2527
        networkReportError(VIR_ERR_INTERNAL_ERROR,
2528 2529 2530 2531 2532
                           _("network '%s' does not have a bridge name."),
                           network->def->name);
        goto cleanup;
    }

2533
    bridge = strdup(network->def->bridge);
2534
    if (!bridge)
2535
        virReportOOMError();
2536 2537

cleanup:
2538 2539
    if (network)
        virNetworkObjUnlock(network);
2540 2541 2542 2543 2544
    return bridge;
}

static int networkGetAutostart(virNetworkPtr net,
                             int *autostart) {
2545 2546 2547
    struct network_driver *driver = net->conn->networkPrivateData;
    virNetworkObjPtr network;
    int ret = -1;
2548

2549
    networkDriverLock(driver);
2550
    network = virNetworkFindByUUID(&driver->networks, net->uuid);
2551
    networkDriverUnlock(driver);
2552
    if (!network) {
2553
        networkReportError(VIR_ERR_NO_NETWORK,
2554
                           "%s", _("no network with matching uuid"));
2555
        goto cleanup;
2556 2557 2558
    }

    *autostart = network->autostart;
2559
    ret = 0;
2560

2561
cleanup:
2562 2563
    if (network)
        virNetworkObjUnlock(network);
2564
    return ret;
2565 2566 2567
}

static int networkSetAutostart(virNetworkPtr net,
2568
                               int autostart) {
2569 2570
    struct network_driver *driver = net->conn->networkPrivateData;
    virNetworkObjPtr network;
2571
    char *configFile = NULL, *autostartLink = NULL;
2572
    int ret = -1;
2573

2574
    networkDriverLock(driver);
2575
    network = virNetworkFindByUUID(&driver->networks, net->uuid);
2576

2577
    if (!network) {
2578
        networkReportError(VIR_ERR_NO_NETWORK,
2579
                           "%s", _("no network with matching uuid"));
2580
        goto cleanup;
2581 2582
    }

2583
    if (!network->persistent) {
2584
        networkReportError(VIR_ERR_OPERATION_INVALID,
2585
                           "%s", _("cannot set autostart for transient network"));
2586 2587 2588
        goto cleanup;
    }

2589 2590
    autostart = (autostart != 0);

2591
    if (network->autostart != autostart) {
2592
        if ((configFile = virNetworkConfigFile(driver->networkConfigDir, network->def->name)) == NULL)
2593
            goto cleanup;
2594
        if ((autostartLink = virNetworkConfigFile(driver->networkAutostartDir, network->def->name)) == NULL)
2595 2596
            goto cleanup;

2597
        if (autostart) {
2598
            if (virFileMakePath(driver->networkAutostartDir) < 0) {
2599
                virReportSystemError(errno,
2600 2601
                                     _("cannot create autostart directory '%s'"),
                                     driver->networkAutostartDir);
2602 2603
                goto cleanup;
            }
2604

2605
            if (symlink(configFile, autostartLink) < 0) {
2606
                virReportSystemError(errno,
2607
                                     _("Failed to create symlink '%s' to '%s'"),
2608
                                     autostartLink, configFile);
2609 2610 2611
                goto cleanup;
            }
        } else {
2612
            if (unlink(autostartLink) < 0 && errno != ENOENT && errno != ENOTDIR) {
2613
                virReportSystemError(errno,
2614
                                     _("Failed to delete symlink '%s'"),
2615
                                     autostartLink);
2616 2617
                goto cleanup;
            }
2618 2619
        }

2620
        network->autostart = autostart;
2621
    }
2622
    ret = 0;
2623

2624
cleanup:
2625 2626
    VIR_FREE(configFile);
    VIR_FREE(autostartLink);
2627 2628
    if (network)
        virNetworkObjUnlock(network);
2629
    networkDriverUnlock(driver);
2630
    return ret;
2631 2632 2633 2634 2635
}


static virNetworkDriver networkDriver = {
    "Network",
2636 2637 2638 2639 2640 2641 2642 2643 2644 2645 2646 2647 2648 2649 2650 2651 2652 2653 2654
    .open = networkOpenNetwork, /* 0.2.0 */
    .close = networkCloseNetwork, /* 0.2.0 */
    .numOfNetworks = networkNumNetworks, /* 0.2.0 */
    .listNetworks = networkListNetworks, /* 0.2.0 */
    .numOfDefinedNetworks = networkNumDefinedNetworks, /* 0.2.0 */
    .listDefinedNetworks = networkListDefinedNetworks, /* 0.2.0 */
    .networkLookupByUUID = networkLookupByUUID, /* 0.2.0 */
    .networkLookupByName = networkLookupByName, /* 0.2.0 */
    .networkCreateXML = networkCreate, /* 0.2.0 */
    .networkDefineXML = networkDefine, /* 0.2.0 */
    .networkUndefine = networkUndefine, /* 0.2.0 */
    .networkCreate = networkStart, /* 0.2.0 */
    .networkDestroy = networkDestroy, /* 0.2.0 */
    .networkGetXMLDesc = networkGetXMLDesc, /* 0.2.0 */
    .networkGetBridgeName = networkGetBridgeName, /* 0.2.0 */
    .networkGetAutostart = networkGetAutostart, /* 0.2.1 */
    .networkSetAutostart = networkSetAutostart, /* 0.2.1 */
    .networkIsActive = networkIsActive, /* 0.7.3 */
    .networkIsPersistent = networkIsPersistent, /* 0.7.3 */
2655 2656 2657
};

static virStateDriver networkStateDriver = {
2658
    "Network",
2659 2660 2661 2662 2663 2664 2665 2666 2667 2668 2669
    networkStartup,
    networkShutdown,
    networkReload,
    networkActive,
};

int networkRegister(void) {
    virRegisterNetworkDriver(&networkDriver);
    virRegisterStateDriver(&networkStateDriver);
    return 0;
}
2670 2671 2672 2673 2674 2675 2676 2677 2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688 2689 2690 2691 2692 2693 2694 2695 2696 2697

/********************************************************/

/* Private API to deal with logical switch capabilities.
 * These functions are exported so that other parts of libvirt can
 * call them, but are not part of the public API and not in the
 * driver's function table. If we ever have more than one network
 * driver, we will need to present these functions via a second
 * "backend" function table.
 */

/* networkAllocateActualDevice:
 * @iface: the original NetDef from the domain
 *
 * Looks up the network reference by iface, allocates a physical
 * device from that network (if appropriate), and returns with the
 * virDomainActualNetDef filled in accordingly. If there are no
 * changes to be made in the netdef, then just leave the actualdef
 * empty.
 *
 * Returns 0 on success, -1 on failure.
 */
int
networkAllocateActualDevice(virDomainNetDefPtr iface)
{
    struct network_driver *driver = driverState;
    virNetworkObjPtr network;
    virNetworkDefPtr netdef;
2698
    virPortGroupDefPtr portgroup;
2699 2700 2701 2702 2703 2704 2705 2706 2707 2708 2709 2710 2711 2712 2713 2714 2715 2716 2717
    int ret = -1;

    if (iface->type != VIR_DOMAIN_NET_TYPE_NETWORK)
        return 0;

    virDomainActualNetDefFree(iface->data.network.actual);
    iface->data.network.actual = NULL;

    networkDriverLock(driver);
    network = virNetworkFindByName(&driver->networks, iface->data.network.name);
    networkDriverUnlock(driver);
    if (!network) {
        networkReportError(VIR_ERR_NO_NETWORK,
                           _("no network with matching name '%s'"),
                           iface->data.network.name);
        goto cleanup;
    }

    netdef = network->def;
2718 2719 2720 2721 2722 2723 2724 2725 2726 2727 2728 2729 2730 2731 2732 2733 2734 2735 2736 2737 2738 2739 2740 2741 2742 2743 2744 2745 2746 2747 2748 2749 2750 2751 2752

    /* portgroup can be present for any type of network, in particular
     * for bandwidth information, so we need to check for that and
     * fill it in appropriately for all forward types.
    */
    portgroup = virPortGroupFindByName(netdef, iface->data.network.portgroup);

    /* If there is already interface-specific bandwidth, just use that
     * (already in NetDef). Otherwise, if there is bandwidth info in
     * the portgroup, fill that into the ActualDef.
     */
    if (portgroup && !iface->bandwidth) {
        if (!iface->data.network.actual
            && (VIR_ALLOC(iface->data.network.actual) < 0)) {
            virReportOOMError();
            goto cleanup;
        }

        if (virBandwidthCopy(&iface->data.network.actual->bandwidth,
                             portgroup->bandwidth) < 0) {
            goto cleanup;
        }
    }

    if ((netdef->forwardType == VIR_NETWORK_FORWARD_NONE) ||
        (netdef->forwardType == VIR_NETWORK_FORWARD_NAT) ||
        (netdef->forwardType == VIR_NETWORK_FORWARD_ROUTE)) {
        /* for these forward types, the actual net type really *is*
         *NETWORK; we just keep the info from the portgroup in
         * iface->data.network.actual
        */
        if (iface->data.network.actual)
            iface->data.network.actual->type = VIR_DOMAIN_NET_TYPE_NETWORK;
    } else if ((netdef->forwardType == VIR_NETWORK_FORWARD_BRIDGE) &&
               netdef->bridge) {
2753 2754 2755 2756 2757

        /* <forward type='bridge'/> <bridge name='xxx'/>
         * is VIR_DOMAIN_NET_TYPE_BRIDGE
         */

2758 2759
        if (!iface->data.network.actual
            && (VIR_ALLOC(iface->data.network.actual) < 0)) {
2760 2761 2762 2763 2764 2765 2766 2767 2768 2769 2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780
            virReportOOMError();
            goto cleanup;
        }

        iface->data.network.actual->type = VIR_DOMAIN_NET_TYPE_BRIDGE;
        iface->data.network.actual->data.bridge.brname = strdup(netdef->bridge);
        if (!iface->data.network.actual->data.bridge.brname) {
            virReportOOMError();
            goto cleanup;
        }

    } else if ((netdef->forwardType == VIR_NETWORK_FORWARD_BRIDGE) ||
               (netdef->forwardType == VIR_NETWORK_FORWARD_PRIVATE) ||
               (netdef->forwardType == VIR_NETWORK_FORWARD_VEPA) ||
               (netdef->forwardType == VIR_NETWORK_FORWARD_PASSTHROUGH)) {
        virVirtualPortProfileParamsPtr virtport = NULL;

        /* <forward type='bridge|private|vepa|passthrough'> are all
         * VIR_DOMAIN_NET_TYPE_DIRECT.
         */

2781 2782
        if (!iface->data.network.actual
            && (VIR_ALLOC(iface->data.network.actual) < 0)) {
2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799 2800 2801 2802 2803 2804 2805 2806 2807 2808 2809 2810 2811 2812 2813 2814 2815 2816 2817 2818 2819 2820 2821 2822
            virReportOOMError();
            goto cleanup;
        }

        /* Set type=direct and appropriate <source mode='xxx'/> */
        iface->data.network.actual->type = VIR_DOMAIN_NET_TYPE_DIRECT;
        switch (netdef->forwardType) {
        case VIR_NETWORK_FORWARD_BRIDGE:
            iface->data.network.actual->data.direct.mode = VIR_MACVTAP_MODE_BRIDGE;
            break;
        case VIR_NETWORK_FORWARD_PRIVATE:
            iface->data.network.actual->data.direct.mode = VIR_MACVTAP_MODE_PRIVATE;
            break;
        case VIR_NETWORK_FORWARD_VEPA:
            iface->data.network.actual->data.direct.mode = VIR_MACVTAP_MODE_VEPA;
            break;
        case VIR_NETWORK_FORWARD_PASSTHROUGH:
            iface->data.network.actual->data.direct.mode = VIR_MACVTAP_MODE_PASSTHRU;
            break;
        }

        /* Find the most specific virtportprofile and copy it */
        if (iface->data.network.virtPortProfile) {
            virtport = iface->data.network.virtPortProfile;
        } else {
            if (portgroup)
                virtport = portgroup->virtPortProfile;
            else
                virtport = netdef->virtPortProfile;
        }
        if (virtport) {
            if (VIR_ALLOC(iface->data.network.actual->data.direct.virtPortProfile) < 0) {
                virReportOOMError();
                goto cleanup;
            }
            /* There are no pointers in a virtualPortProfile, so a shallow copy
             * is sufficient
             */
            *iface->data.network.actual->data.direct.virtPortProfile = *virtport;
        }
2823

2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834 2835 2836 2837 2838 2839 2840 2841 2842 2843 2844 2845 2846 2847 2848 2849 2850 2851 2852 2853 2854 2855 2856 2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 2867 2868 2869 2870 2871 2872 2873 2874 2875 2876 2877 2878 2879 2880 2881 2882 2883 2884 2885 2886 2887 2888 2889 2890 2891 2892 2893 2894 2895 2896 2897 2898 2899 2900 2901 2902 2903 2904 2905 2906 2907 2908 2909 2910 2911 2912 2913 2914 2915 2916 2917 2918 2919 2920 2921 2922 2923 2924 2925 2926 2927 2928 2929 2930 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 2973 2974 2975 2976 2977 2978 2979 2980 2981 2982 2983 2984 2985 2986 2987 2988 2989 2990 2991 2992 2993 2994 2995 2996 2997 2998 2999 3000 3001 3002 3003 3004 3005 3006 3007 3008 3009 3010 3011 3012 3013 3014 3015 3016 3017 3018 3019 3020 3021 3022 3023 3024 3025 3026 3027 3028 3029 3030 3031 3032 3033 3034 3035 3036 3037 3038 3039 3040 3041 3042 3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054 3055 3056 3057 3058 3059 3060 3061 3062 3063 3064 3065 3066 3067 3068 3069 3070 3071 3072 3073 3074 3075 3076
        /* If there is only a single device, just return it (caller will detect
         * any error if exclusive use is required but could not be acquired).
         */
        if (netdef->nForwardIfs == 0) {
            networkReportError(VIR_ERR_INTERNAL_ERROR,
                               _("network '%s' uses a direct mode, but has no forward dev and no interface pool"),
                               netdef->name);
            goto cleanup;
        } else {
            int ii;
            virNetworkForwardIfDefPtr dev = NULL;

            /* pick an interface from the pool */

            /* PASSTHROUGH mode, and PRIVATE Mode + 802.1Qbh both require
             * exclusive access to a device, so current usageCount must be
             * 0.  Other modes can share, so just search for the one with
             * the lowest usageCount.
             */
            if ((netdef->forwardType == VIR_NETWORK_FORWARD_PASSTHROUGH) ||
                ((netdef->forwardType == VIR_NETWORK_FORWARD_PRIVATE) &&
                 iface->data.network.actual->data.direct.virtPortProfile &&
                 (iface->data.network.actual->data.direct.virtPortProfile->virtPortType
                  == VIR_VIRTUALPORT_8021QBH))) {
                /* pick first dev with 0 usageCount */

                for (ii = 0; ii < netdef->nForwardIfs; ii++) {
                    if (netdef->forwardIfs[ii].usageCount == 0) {
                        dev = &netdef->forwardIfs[ii];
                        break;
                    }
                }
            } else {
                /* pick least used dev */
                dev = &netdef->forwardIfs[0];
                for (ii = 1; ii < netdef->nForwardIfs; ii++) {
                    if (netdef->forwardIfs[ii].usageCount < dev->usageCount)
                        dev = &netdef->forwardIfs[ii];
                }
            }
            /* dev points at the physical device we want to use */
            if (!dev) {
                networkReportError(VIR_ERR_INTERNAL_ERROR,
                                   _("network '%s' requires exclusive access to interfaces, but none are available"),
                               netdef->name);
                goto cleanup;
            }
            iface->data.network.actual->data.direct.linkdev = strdup(dev->dev);
            if (!iface->data.network.actual->data.direct.linkdev) {
                virReportOOMError();
                goto cleanup;
            }
            /* we are now assured of success, so mark the allocation */
            dev->usageCount++;
            VIR_DEBUG("Using physical device %s, usageCount %d",
                      dev->dev, dev->usageCount);
        }
    }

    ret = 0;
cleanup:
    if (network)
        virNetworkObjUnlock(network);
    if (ret < 0) {
        virDomainActualNetDefFree(iface->data.network.actual);
        iface->data.network.actual = NULL;
    }
    return ret;
}

/* networkNotifyActualDevice:
 * @iface:  the domain's NetDef with an "actual" device already filled in.
 *
 * Called to notify the network driver when libvirtd is restarted and
 * finds an already running domain. If appropriate it will force an
 * allocation of the actual->direct.linkdev to get everything back in
 * order.
 *
 * Returns 0 on success, -1 on failure.
 */
int
networkNotifyActualDevice(virDomainNetDefPtr iface)
{
    struct network_driver *driver = driverState;
    virNetworkObjPtr network;
    virNetworkDefPtr netdef;
    char *actualDev;
    int ret = -1;

    if (iface->type != VIR_DOMAIN_NET_TYPE_NETWORK)
       return 0;

    if (!iface->data.network.actual ||
        (virDomainNetGetActualType(iface) != VIR_DOMAIN_NET_TYPE_DIRECT)) {
        VIR_DEBUG("Nothing to claim from network %s", iface->data.network.name);
        return 0;
    }

    networkDriverLock(driver);
    network = virNetworkFindByName(&driver->networks, iface->data.network.name);
    networkDriverUnlock(driver);
    if (!network) {
        networkReportError(VIR_ERR_NO_NETWORK,
                           _("no network with matching name '%s'"),
                           iface->data.network.name);
        goto cleanup;
    }

    actualDev = virDomainNetGetActualDirectDev(iface);
    if (!actualDev) {
        networkReportError(VIR_ERR_INTERNAL_ERROR,
                           "%s", _("the interface uses a direct mode, but has no source dev"));
            goto cleanup;
        }

    netdef = network->def;
    if (netdef->nForwardIfs == 0) {
        networkReportError(VIR_ERR_INTERNAL_ERROR,
                           _("network '%s' uses a direct mode, but has no forward dev and no interface pool"),
                           netdef->name);
        goto cleanup;
    } else {
        int ii;
        virNetworkForwardIfDefPtr dev = NULL;

        /* find the matching interface in the pool and increment its usageCount */

        for (ii = 0; ii < netdef->nForwardIfs; ii++) {
            if (STREQ(actualDev, netdef->forwardIfs[ii].dev)) {
                dev = &netdef->forwardIfs[ii];
                break;
            }
        }
        /* dev points at the physical device we want to use */
        if (!dev) {
            networkReportError(VIR_ERR_INTERNAL_ERROR,
                               _("network '%s' doesn't have dev='%s' in use by domain"),
                               netdef->name, actualDev);
            goto cleanup;
        }

        /* PASSTHROUGH mode, and PRIVATE Mode + 802.1Qbh both require
         * exclusive access to a device, so current usageCount must be
         * 0 in those cases.
         */
        if ((dev->usageCount > 0) &&
            ((netdef->forwardType == VIR_NETWORK_FORWARD_PASSTHROUGH) ||
             ((netdef->forwardType == VIR_NETWORK_FORWARD_PRIVATE) &&
              iface->data.network.actual->data.direct.virtPortProfile &&
              (iface->data.network.actual->data.direct.virtPortProfile->virtPortType
               == VIR_VIRTUALPORT_8021QBH)))) {
            networkReportError(VIR_ERR_INTERNAL_ERROR,
                               _("network '%s' claims dev='%s' is already in use by a different domain"),
                               netdef->name, actualDev);
            goto cleanup;
        }
        /* we are now assured of success, so mark the allocation */
        dev->usageCount++;
        VIR_DEBUG("Using physical device %s, usageCount %d",
                  dev->dev, dev->usageCount);
    }

    ret = 0;
cleanup:
    if (network)
        virNetworkObjUnlock(network);
    return ret;
}


/* networkReleaseActualDevice:
 * @iface:  a domain's NetDef (interface definition)
 *
 * Given a domain <interface> element that previously had its <actual>
 * element filled in (and possibly a physical device allocated to it),
 * free up the physical device for use by someone else, and free the
 * virDomainActualNetDef.
 *
 * Returns 0 on success, -1 on failure.
 */
int
networkReleaseActualDevice(virDomainNetDefPtr iface)
{
    struct network_driver *driver = driverState;
    virNetworkObjPtr network = NULL;
    virNetworkDefPtr netdef;
    char *actualDev;
    int ret = -1;

    if (iface->type != VIR_DOMAIN_NET_TYPE_NETWORK)
       return 0;

    if (!iface->data.network.actual ||
        (virDomainNetGetActualType(iface) != VIR_DOMAIN_NET_TYPE_DIRECT)) {
        VIR_DEBUG("Nothing to release to network %s", iface->data.network.name);
        ret = 0;
        goto cleanup;
    }

    networkDriverLock(driver);
    network = virNetworkFindByName(&driver->networks, iface->data.network.name);
    networkDriverUnlock(driver);
    if (!network) {
        networkReportError(VIR_ERR_NO_NETWORK,
                           _("no network with matching name '%s'"),
                           iface->data.network.name);
        goto cleanup;
    }

    actualDev = virDomainNetGetActualDirectDev(iface);
    if (!actualDev) {
        networkReportError(VIR_ERR_INTERNAL_ERROR,
                           "%s", _("the interface uses a direct mode, but has no source dev"));
            goto cleanup;
        }

    netdef = network->def;
    if (netdef->nForwardIfs == 0) {
        networkReportError(VIR_ERR_INTERNAL_ERROR,
                           _("network '%s' uses a direct mode, but has no forward dev and no interface pool"),
                           netdef->name);
        goto cleanup;
    } else {
        int ii;
        virNetworkForwardIfDefPtr dev = NULL;

        for (ii = 0; ii < netdef->nForwardIfs; ii++) {
            if (STREQ(actualDev, netdef->forwardIfs[ii].dev)) {
                dev = &netdef->forwardIfs[ii];
                break;
            }
        }
        /* dev points at the physical device we've been using */
        if (!dev) {
            networkReportError(VIR_ERR_INTERNAL_ERROR,
                               _("network '%s' doesn't have dev='%s' in use by domain"),
                               netdef->name, actualDev);
            goto cleanup;
        }

        dev->usageCount--;
        VIR_DEBUG("Releasing physical device %s, usageCount %d",
                  dev->dev, dev->usageCount);
    }

    ret = 0;
cleanup:
    if (network)
        virNetworkObjUnlock(network);
    virDomainActualNetDefFree(iface->data.network.actual);
    iface->data.network.actual = NULL;
    return ret;
}
3077 3078 3079 3080 3081 3082 3083 3084 3085 3086 3087 3088 3089 3090 3091 3092 3093 3094 3095 3096 3097 3098 3099 3100 3101 3102 3103 3104 3105 3106

/*
 * networkGetNetworkAddress:
 * @netname: the name of a network
 * @netaddr: string representation of IP address for that network.
 *
 * Attempt to return an IP (v4) address associated with the named
 * network. If a libvirt virtual network, that will be provided in the
 * configuration. For host bridge and direct (macvtap) networks, we
 * must do an ioctl to learn the address.
 *
 * Note: This function returns the 1st IPv4 address it finds. It might
 * be useful if it was more flexible, but the current use (getting a
 * listen address for qemu's vnc/spice graphics server) can only use a
 * single address anyway.
 *
 * Returns 0 on success, and puts a string (which must be free'd by
 * the caller) into *netaddr. Returns -1 on failure or -2 if
 * completely unsupported.
 */
int
networkGetNetworkAddress(const char *netname, char **netaddr)
{
    int ret = -1;
    struct network_driver *driver = driverState;
    virNetworkObjPtr network = NULL;
    virNetworkDefPtr netdef;
    virNetworkIpDefPtr ipdef;
    virSocketAddr addr;
    virSocketAddrPtr addrptr = NULL;
3107
    char *dev_name = NULL;
3108 3109 3110 3111 3112 3113 3114 3115 3116 3117 3118 3119 3120 3121 3122 3123 3124 3125 3126 3127 3128 3129 3130 3131 3132 3133 3134 3135 3136

    *netaddr = NULL;
    networkDriverLock(driver);
    network = virNetworkFindByName(&driver->networks, netname);
    networkDriverUnlock(driver);
    if (!network) {
        networkReportError(VIR_ERR_NO_NETWORK,
                           _("no network with matching name '%s'"),
                           netname);
        goto cleanup;
    }
    netdef = network->def;

    switch (netdef->forwardType) {
    case VIR_NETWORK_FORWARD_NONE:
    case VIR_NETWORK_FORWARD_NAT:
    case VIR_NETWORK_FORWARD_ROUTE:
        /* if there's an ipv4def, get it's address */
        ipdef = virNetworkDefGetIpByIndex(netdef, AF_INET, 0);
        if (!ipdef) {
            networkReportError(VIR_ERR_INTERNAL_ERROR,
                               _("network '%s' doesn't have an IPv4 address"),
                               netdef->name);
            break;
        }
        addrptr = &ipdef->address;
        break;

    case VIR_NETWORK_FORWARD_BRIDGE:
3137
        if ((dev_name = netdef->bridge))
3138 3139 3140 3141 3142 3143 3144 3145 3146
            break;
        /*
         * fall through if netdef->bridge wasn't set, since this is
         * also a direct-mode interface.
         */
    case VIR_NETWORK_FORWARD_PRIVATE:
    case VIR_NETWORK_FORWARD_VEPA:
    case VIR_NETWORK_FORWARD_PASSTHROUGH:
        if ((netdef->nForwardIfs > 0) && netdef->forwardIfs)
3147
            dev_name = netdef->forwardIfs[0].dev;
3148

3149
        if (!dev_name) {
3150 3151 3152 3153 3154 3155 3156
            networkReportError(VIR_ERR_INTERNAL_ERROR,
                               _("network '%s' has no associated interface or bridge"),
                               netdef->name);
        }
        break;
    }

3157 3158
    if (dev_name) {
        if (ifaceGetIPAddress(dev_name, &addr)) {
3159 3160
            virReportSystemError(errno,
                                 _("Failed to get IP address for '%s' (network '%s')"),
3161
                                 dev_name, netdef->name);
3162 3163 3164 3165 3166 3167
        } else {
            addrptr = &addr;
        }
    }

    if (addrptr &&
3168
        (*netaddr = virSocketAddrFormat(addrptr))) {
3169 3170 3171 3172 3173 3174 3175 3176
        ret = 0;
    }

cleanup:
    if (network)
        virNetworkObjUnlock(network);
    return ret;
}