bridge_driver.c 100.7 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
 * 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 <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>
44
#include <net/if.h>
45

46
#include "virterror_internal.h"
47
#include "datatypes.h"
48
#include "bridge_driver.h"
49 50 51
#include "network_conf.h"
#include "driver.h"
#include "buf.h"
52
#include "virpidfile.h"
53
#include "util.h"
54
#include "command.h"
55 56 57
#include "memory.h"
#include "uuid.h"
#include "iptables.h"
58
#include "logging.h"
59
#include "dnsmasq.h"
60
#include "configmake.h"
61
#include "ignore-value.h"
62 63 64
#include "virnetdev.h"
#include "virnetdevbridge.h"
#include "virnetdevtap.h"
65

66 67
#define NETWORK_PID_DIR LOCALSTATEDIR "/run/libvirt/network"
#define NETWORK_STATE_DIR LOCALSTATEDIR "/lib/libvirt/network"
68

69
#define DNSMASQ_STATE_DIR LOCALSTATEDIR "/lib/libvirt/dnsmasq"
70
#define RADVD_STATE_DIR LOCALSTATEDIR "/lib/libvirt/radvd"
71

72 73
#define VIR_FROM_THIS VIR_FROM_NETWORK

74
#define networkReportError(code, ...)                                   \
75
    virReportErrorHelper(VIR_FROM_NETWORK, code, __FILE__,              \
76
                         __FUNCTION__, __LINE__, __VA_ARGS__)
77

78 79
/* Main driver state */
struct network_driver {
80
    virMutex lock;
81

82
    virNetworkObjList networks;
83 84 85 86 87 88 89

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

90 91 92

static void networkDriverLock(struct network_driver *driver)
{
93
    virMutexLock(&driver->lock);
94 95 96
}
static void networkDriverUnlock(struct network_driver *driver)
{
97
    virMutexUnlock(&driver->lock);
98 99
}

100 101
static int networkShutdown(void);

102 103 104 105 106 107 108
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,
109
                                     virNetworkObjPtr network);
110

111 112 113 114 115 116 117
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,
118
                                        virNetworkObjPtr network);
119

120 121
static void networkReloadIptablesRules(struct network_driver *driver);

122 123
static struct network_driver *driverState = NULL;

124
static char *
125
networkDnsmasqLeaseFileNameDefault(const char *netname)
126 127 128
{
    char *leasefile;

129 130
    ignore_value(virAsprintf(&leasefile, DNSMASQ_STATE_DIR "/%s.leases",
                             netname));
131 132 133
    return leasefile;
}

134 135 136
networkDnsmasqLeaseFileNameFunc networkDnsmasqLeaseFileName =
    networkDnsmasqLeaseFileNameDefault;

137 138 139 140 141 142
static char *
networkRadvdPidfileBasename(const char *netname)
{
    /* this is simple but we want to be sure it's consistently done */
    char *pidfilebase;

143
    ignore_value(virAsprintf(&pidfilebase, "%s-radvd", netname));
144 145 146 147 148 149 150 151
    return pidfilebase;
}

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

152 153
    ignore_value(virAsprintf(&configfile, RADVD_STATE_DIR "/%s-radvd.conf",
                             netname));
154 155
    return configfile;
}
156

157 158 159
static char *
networkBridgeDummyNicName(const char *brname)
{
160
    static const char dummyNicSuffix[] = "-nic";
161 162
    char *nicname;

163 164 165 166 167 168 169
    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.
         */
170 171 172 173 174
        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));
175
    } else {
176
        ignore_value(virAsprintf(&nicname, "%s%s", brname, dummyNicSuffix));
177
    }
178 179 180
    return nicname;
}

181 182 183 184 185 186 187 188 189 190 191
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);

192
        if ((config = virNetworkConfigFile(NETWORK_STATE_DIR,
193 194 195 196 197 198 199 200 201 202 203 204
                                           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 */
205
        tmp = virNetworkDefParseFile(config);
206 207 208 209 210 211 212 213
        VIR_FREE(config);
        if (tmp) {
            obj->newDef = obj->def;
            obj->def = tmp;
        }

        /* If bridge exists, then mark it active */
        if (obj->def->bridge &&
H
Hu Tao 已提交
214
            virNetDevExists(obj->def->bridge) == 1) {
215 216
            obj->active = 1;

217 218
            /* Try and read dnsmasq/radvd pids if any */
            if (obj->def->ips && (obj->def->nips > 0)) {
219 220 221 222
                char *radvdpidbase;

                ignore_value(virPidFileReadIfAlive(NETWORK_PID_DIR, obj->def->name,
                                                   &obj->dnsmasqPid, DNSMASQ));
223

224
                if (!(radvdpidbase = networkRadvdPidfileBasename(obj->def->name))) {
225
                    virReportOOMError();
226 227
                    goto cleanup;
                }
228 229
                ignore_value(virPidFileReadIfAlive(NETWORK_PID_DIR, radvdpidbase,
                                                   &obj->radvdPid, RADVD));
230
                VIR_FREE(radvdpidbase);
231 232 233
            }
        }

234
    cleanup:
235 236 237 238 239
        virNetworkObjUnlock(obj);
    }
}


240 241 242
static void
networkAutostartConfigs(struct network_driver *driver) {
    unsigned int i;
243

244
    for (i = 0 ; i < driver->networks.count ; i++) {
245
        virNetworkObjLock(driver->networks.objs[i]);
246
        if (driver->networks.objs[i]->autostart &&
247 248
            !virNetworkObjIsActive(driver->networks.objs[i])) {
            if (networkStartNetwork(driver, driver->networks.objs[i]) < 0) {
249
            /* failed to start but already logged */
250
            }
251
        }
252
        virNetworkObjUnlock(driver->networks.objs[i]);
253 254 255 256 257 258 259 260 261
    }
}

/**
 * networkStartup:
 *
 * Initialization function for the QEmu daemon
 */
static int
262
networkStartup(int privileged) {
263 264 265 266
    uid_t uid = geteuid();
    char *base = NULL;

    if (VIR_ALLOC(driverState) < 0)
267
        goto error;
268

269 270 271 272
    if (virMutexInit(&driverState->lock) < 0) {
        VIR_FREE(driverState);
        goto error;
    }
273 274
    networkDriverLock(driverState);

275
    if (privileged) {
276
        if (virAsprintf(&driverState->logDir,
277
                        "%s/log/libvirt/qemu", LOCALSTATEDIR) == -1)
278 279
            goto out_of_memory;

280
        if ((base = strdup (SYSCONFDIR "/libvirt")) == NULL)
281 282
            goto out_of_memory;
    } else {
283
        char *userdir = virGetUserDirectory(uid);
284 285 286

        if (!userdir)
            goto error;
287

288
        if (virAsprintf(&driverState->logDir,
289 290
                        "%s/.libvirt/qemu/log", userdir) == -1) {
            VIR_FREE(userdir);
291
            goto out_of_memory;
292
        }
293

294 295
        if (virAsprintf(&base, "%s/.libvirt", userdir) == -1) {
            VIR_FREE(userdir);
296 297
            goto out_of_memory;
        }
298
        VIR_FREE(userdir);
299 300 301 302 303
    }

    /* Configuration paths are either ~/.libvirt/qemu/... (session) or
     * /etc/libvirt/qemu/... (system).
     */
304
    if (virAsprintf(&driverState->networkConfigDir, "%s/qemu/networks", base) == -1)
305 306
        goto out_of_memory;

307 308
    if (virAsprintf(&driverState->networkAutostartDir, "%s/qemu/networks/autostart",
                    base) == -1)
309 310 311 312
        goto out_of_memory;

    VIR_FREE(base);

313
    if (!(driverState->iptables = iptablesContextNew())) {
314
        goto out_of_memory;
315 316 317
    }


318
    if (virNetworkLoadAllConfigs(&driverState->networks,
319
                                 driverState->networkConfigDir,
320 321 322
                                 driverState->networkAutostartDir) < 0)
        goto error;

323
    networkFindActiveConfigs(driverState);
324
    networkReloadIptablesRules(driverState);
325 326
    networkAutostartConfigs(driverState);

327 328
    networkDriverUnlock(driverState);

329 330
    return 0;

331
out_of_memory:
332
    virReportOOMError();
333 334

error:
335 336 337
    if (driverState)
        networkDriverUnlock(driverState);

338
    VIR_FREE(base);
339
    networkShutdown();
340 341 342 343 344 345 346 347 348 349 350
    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) {
351 352 353
    if (!driverState)
        return 0;

354
    networkDriverLock(driverState);
355
    virNetworkLoadAllConfigs(&driverState->networks,
356 357
                             driverState->networkConfigDir,
                             driverState->networkAutostartDir);
358
    networkReloadIptablesRules(driverState);
359
    networkAutostartConfigs(driverState);
360
    networkDriverUnlock(driverState);
361 362 363 364 365 366 367 368 369 370 371 372 373
    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) {
374
    unsigned int i;
375
    int active = 0;
376

377 378 379
    if (!driverState)
        return 0;

380
    networkDriverLock(driverState);
381 382
    for (i = 0 ; i < driverState->networks.count ; i++) {
        virNetworkObjPtr net = driverState->networks.objs[i];
383
        virNetworkObjLock(net);
D
Daniel P. Berrange 已提交
384
        if (virNetworkObjIsActive(net))
385
            active = 1;
386
        virNetworkObjUnlock(net);
387
    }
388
    networkDriverUnlock(driverState);
389
    return active;
390 391 392 393 394 395 396 397 398 399 400 401
}

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

402 403
    networkDriverLock(driverState);

404
    /* free inactive networks */
405
    virNetworkObjListFree(&driverState->networks);
406 407 408 409 410 411 412 413

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

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

414
    networkDriverUnlock(driverState);
415
    virMutexDestroy(&driverState->lock);
416

417 418 419 420 421 422
    VIR_FREE(driverState);

    return 0;
}


423 424 425 426
static int
networkBuildDnsmasqHostsfile(dnsmasqContext *dctx,
                             virNetworkIpDefPtr ipdef,
                             virNetworkDNSDefPtr dnsdef)
427
{
428
    unsigned int i, j;
429

430 431
    for (i = 0; i < ipdef->nhosts; i++) {
        virNetworkDHCPHostDefPtr host = &(ipdef->hosts[i]);
432
        if ((host->mac) && VIR_SOCKET_ADDR_VALID(&host->ip))
433 434
            if (dnsmasqAddDhcpHost(dctx, host->mac, &host->ip, host->name) < 0)
                return -1;
435
    }
436

437 438 439
    if (dnsdef) {
        for (i = 0; i < dnsdef->nhosts; i++) {
            virNetworkDNSHostsDefPtr host = &(dnsdef->hosts[i]);
440
            if (VIR_SOCKET_ADDR_VALID(&host->ip)) {
441
                for (j = 0; j < host->nnames; j++)
442 443
                    if (dnsmasqAddHost(dctx, &host->ip, host->names[j]) < 0)
                        return -1;
444 445
            }
        }
446 447
    }

448
    return 0;
449 450 451
}


452
static int
453
networkBuildDnsmasqArgv(virNetworkObjPtr network,
454
                        virNetworkIpDefPtr ipdef,
455
                        const char *pidfile,
456 457 458
                        virCommandPtr cmd,
                        dnsmasqContext *dctx)
{
459
    int r, ret = -1;
460
    int nbleases = 0;
461 462
    int ii;
    virNetworkIpDefPtr tmpipdef;
463 464

    /*
465
     * NB, be careful about syntax for dnsmasq options in long format.
466 467 468 469 470 471 472 473 474 475 476 477 478
     *
     * 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,
479 480
     * without reading the dnsmasq source :-( The manpage is not
     * very explicit on this.
481
     */
482 483 484 485 486

    /*
     * Needed to ensure dnsmasq uses same algorithm for processing
     * multiple namedriver entries in /etc/resolv.conf as GLibC.
     */
487
    virCommandAddArgList(cmd, "--strict-order", "--bind-interfaces", NULL);
488

489 490
    if (network->def->domain)
        virCommandAddArgList(cmd, "--domain", network->def->domain, NULL);
491

492 493
    if (pidfile)
        virCommandAddArgPair(cmd, "--pid-file", pidfile);
494

495
    /* *no* conf file */
496
    virCommandAddArg(cmd, "--conf-file=");
497

498 499 500
    virCommandAddArgList(cmd,
                         "--except-interface", "lo",
                         NULL);
501

502 503
    /* If this is an isolated network, set the default route option
     * (3) to be empty to avoid setting a default route that's
504 505 506 507
     * 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).
508
     */
509 510 511 512
    if (network->def->forwardType == VIR_NETWORK_FORWARD_NONE) {
        virCommandAddArgList(cmd, "--dhcp-option=3",
                             "--no-resolv", NULL);
    }
513

514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531
    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);
        }
    }

532 533 534 535 536 537 538 539 540 541 542
    /*
     * --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++) {
543
        char *ipaddr = virSocketAddrFormat(&tmpipdef->address);
544 545 546 547 548 549
        if (!ipaddr)
            goto cleanup;
        virCommandAddArgList(cmd, "--listen-address", ipaddr, NULL);
        VIR_FREE(ipaddr);
    }

550
    if (ipdef) {
551
        for (r = 0 ; r < ipdef->nranges ; r++) {
552
            char *saddr = virSocketAddrFormat(&ipdef->ranges[r].start);
553 554
            if (!saddr)
                goto cleanup;
555
            char *eaddr = virSocketAddrFormat(&ipdef->ranges[r].end);
556 557 558 559 560 561
            if (!eaddr) {
                VIR_FREE(saddr);
                goto cleanup;
            }
            virCommandAddArg(cmd, "--dhcp-range");
            virCommandAddArgFormat(cmd, "%s,%s", saddr, eaddr);
562
            VIR_FREE(saddr);
563
            VIR_FREE(eaddr);
564 565
            nbleases += virSocketAddrGetRange(&ipdef->ranges[r].start,
                                              &ipdef->ranges[r].end);
566
        }
567

568 569 570 571 572 573
        /*
         * 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) {
574
            char *bridgeaddr = virSocketAddrFormat(&ipdef->address);
575 576 577 578 579 580
            if (!bridgeaddr)
                goto cleanup;
            virCommandAddArg(cmd, "--dhcp-range");
            virCommandAddArgFormat(cmd, "%s,static", bridgeaddr);
            VIR_FREE(bridgeaddr);
        }
581

582
        if (ipdef->nranges > 0) {
583 584 585 586 587
            char *leasefile = networkDnsmasqLeaseFileName(network->def->name);
            if (!leasefile)
                goto cleanup;
            virCommandAddArgFormat(cmd, "--dhcp-leasefile=%s", leasefile);
            VIR_FREE(leasefile);
588 589
            virCommandAddArgFormat(cmd, "--dhcp-lease-max=%d", nbleases);
        }
590

591 592
        if (ipdef->nranges || ipdef->nhosts)
            virCommandAddArg(cmd, "--dhcp-no-override");
593

594 595 596 597
        /* add domain to any non-qualified hostnames in /etc/hosts or addn-hosts */
        if (network->def->domain)
           virCommandAddArg(cmd, "--expand-hosts");

598 599 600 601 602 603 604 605 606
        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);
607

608 609 610 611 612 613 614
        if (ipdef->tftproot) {
            virCommandAddArgList(cmd, "--enable-tftp",
                                 "--tftp-root", ipdef->tftproot,
                                 NULL);
        }
        if (ipdef->bootfile) {
            virCommandAddArg(cmd, "--dhcp-boot");
615 616
            if (VIR_SOCKET_ADDR_VALID(&ipdef->bootserver)) {
                char *bootserver = virSocketAddrFormat(&ipdef->bootserver);
617

618 619 620 621 622 623 624 625
                if (!bootserver)
                    goto cleanup;
                virCommandAddArgFormat(cmd, "%s%s%s",
                                       ipdef->bootfile, ",,", bootserver);
                VIR_FREE(bootserver);
            } else {
                virCommandAddArg(cmd, ipdef->bootfile);
            }
626
        }
627 628
    }

629 630 631
    ret = 0;
cleanup:
    return ret;
632 633
}

634 635
int
networkBuildDhcpDaemonCommandLine(virNetworkObjPtr network, virCommandPtr *cmdout,
636
                                  char *pidfile, dnsmasqContext *dctx)
637
{
638
    virCommandPtr cmd = NULL;
639
    int ret = -1, ii;
640
    virNetworkIpDefPtr ipdef;
641 642

    network->dnsmasqPid = -1;
643

644 645 646 647 648 649 650
    /* 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;
651
    }
652
    /* If no IPv4 addresses had dhcp info, pick the first (if there were any). */
653
    if (!ipdef)
654 655 656 657 658 659 660 661
        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))
662
        return 0;
663

664
    cmd = virCommandNew(DNSMASQ);
665
    if (networkBuildDnsmasqArgv(network, ipdef, pidfile, cmd, dctx) < 0) {
666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683
        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;
684
    dnsmasqContext *dctx = NULL;
685

686 687
    if (virFileMakePath(NETWORK_PID_DIR) < 0) {
        virReportSystemError(errno,
688 689
                             _("cannot create directory %s"),
                             NETWORK_PID_DIR);
690
        goto cleanup;
691
    }
692 693
    if (virFileMakePath(NETWORK_STATE_DIR) < 0) {
        virReportSystemError(errno,
694 695
                             _("cannot create directory %s"),
                             NETWORK_STATE_DIR);
696
        goto cleanup;
697 698
    }

699
    if (!(pidfile = virPidFileBuildPath(NETWORK_PID_DIR, network->def->name))) {
700
        virReportOOMError();
701
        goto cleanup;
702 703
    }

704 705
    if (virFileMakePath(DNSMASQ_STATE_DIR) < 0) {
        virReportSystemError(errno,
706 707 708 709 710
                             _("cannot create directory %s"),
                             DNSMASQ_STATE_DIR);
        goto cleanup;
    }

711 712 713 714 715 716 717 718 719 720
    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)
721
        goto cleanup;
722

G
Guido Günther 已提交
723 724
    ret = virCommandRun(cmd, NULL);
    if (ret < 0) {
725
        goto cleanup;
G
Guido Günther 已提交
726
    }
727 728

    /*
729 730 731 732 733
     * 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
734 735
     */

736
    ret = virPidFileRead(NETWORK_PID_DIR, network->def->name,
737 738
                         &network->dnsmasqPid);
    if (ret < 0)
739
        goto cleanup;
740

741 742 743
    ret = 0;
cleanup:
    VIR_FREE(pidfile);
744
    virCommandFree(cmd);
745
    dnsmasqContextFree(dctx);
746 747 748
    return ret;
}

749 750 751 752 753 754 755 756 757
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;
758
    int ret = -1, ii;
759 760 761 762
    virNetworkIpDefPtr ipdef;

    network->radvdPid = -1;

E
Eric Blake 已提交
763
    if (!virFileIsExecutable(RADVD)) {
764 765 766 767 768 769 770
        virReportSystemError(errno,
                             _("Cannot find %s - "
                               "Possibly the package isn't installed"),
                             RADVD);
        goto cleanup;
    }

771 772
    if (virFileMakePath(NETWORK_PID_DIR) < 0) {
        virReportSystemError(errno,
773 774 775 776
                             _("cannot create directory %s"),
                             NETWORK_PID_DIR);
        goto cleanup;
    }
777 778
    if (virFileMakePath(RADVD_STATE_DIR) < 0) {
        virReportSystemError(errno,
779 780 781 782 783 784 785 786 787 788
                             _("cannot create directory %s"),
                             RADVD_STATE_DIR);
        goto cleanup;
    }

    /* construct pidfile name */
    if (!(radvdpidbase = networkRadvdPidfileBasename(network->def->name))) {
        virReportOOMError();
        goto cleanup;
    }
789
    if (!(pidfile = virPidFileBuildPath(NETWORK_PID_DIR, radvdpidbase))) {
790 791 792 793 794
        virReportOOMError();
        goto cleanup;
    }

    /* create radvd config file appropriate for this network */
795
    virBufferAsprintf(&configbuf, "interface %s\n"
796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814
                      "{\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;
        }
815
        if (!(netaddr = virSocketAddrFormat(&ipdef->address)))
816
            goto cleanup;
817
        virBufferAsprintf(&configbuf,
818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855
                          "  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
856
     * virPidFileRead() below will fail if we use them).
857 858 859 860 861 862 863 864 865 866 867 868 869 870 871
     * 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;

872
    if (virPidFileRead(NETWORK_PID_DIR, radvdpidbase,
873 874 875 876 877 878 879 880 881 882 883 884 885 886
                       &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;
}

887
static int
888
networkAddMasqueradingIptablesRules(struct network_driver *driver,
889 890
                                    virNetworkObjPtr network,
                                    virNetworkIpDefPtr ipdef)
891 892
{
    int prefix = virNetworkIpDefPrefix(ipdef);
893
    const char *forwardIf = virNetworkDefForwardIf(network->def, 0);
894 895 896 897 898 899 900

    if (prefix < 0) {
        networkReportError(VIR_ERR_INTERNAL_ERROR,
                           _("Invalid prefix or netmask for '%s'"),
                           network->def->bridge);
        goto masqerr1;
    }
901

902
    /* allow forwarding packets from the bridge interface */
903
    if (iptablesAddForwardAllowOut(driver->iptables,
904
                                   &ipdef->address,
905
                                   prefix,
906
                                   network->def->bridge,
907
                                   forwardIf) < 0) {
908 909 910
        networkReportError(VIR_ERR_SYSTEM_ERROR,
                           _("failed to add iptables rule to allow forwarding from '%s'"),
                           network->def->bridge);
911 912 913
        goto masqerr1;
    }

914 915 916
    /* allow forwarding packets to the bridge interface if they are
     * part of an existing connection
     */
917
    if (iptablesAddForwardAllowRelatedIn(driver->iptables,
918
                                         &ipdef->address,
919
                                         prefix,
920
                                         network->def->bridge,
921
                                         forwardIf) < 0) {
922 923 924
        networkReportError(VIR_ERR_SYSTEM_ERROR,
                           _("failed to add iptables rule to allow forwarding to '%s'"),
                           network->def->bridge);
925 926 927
        goto masqerr2;
    }

928 929 930 931 932
    /*
     * Enable masquerading.
     *
     * We need to end up with 3 rules in the table in this order
     *
E
Eric Blake 已提交
933 934
     *  1. protocol=tcp with sport mapping restriction
     *  2. protocol=udp with sport mapping restriction
935 936 937
     *  3. generic any protocol
     *
     * The sport mappings are required, because default IPtables
E
Eric Blake 已提交
938
     * MASQUERADE maintain port numbers unchanged where possible.
939 940 941 942 943 944 945 946 947 948 949 950 951
     *
     * 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 */
952
    if (iptablesAddForwardMasquerade(driver->iptables,
953
                                     &ipdef->address,
954
                                     prefix,
955
                                     forwardIf,
956 957
                                     NULL) < 0) {
        networkReportError(VIR_ERR_SYSTEM_ERROR,
958 959 960 961
                           forwardIf ?
                           _("failed to add iptables rule to enable masquerading to %s") :
                           _("failed to add iptables rule to enable masquerading"),
                           forwardIf);
962 963 964
        goto masqerr3;
    }

965
    /* UDP with a source port restriction */
966
    if (iptablesAddForwardMasquerade(driver->iptables,
967
                                     &ipdef->address,
968
                                     prefix,
969
                                     forwardIf,
970 971
                                     "udp") < 0) {
        networkReportError(VIR_ERR_SYSTEM_ERROR,
972 973 974 975
                           forwardIf ?
                           _("failed to add iptables rule to enable UDP masquerading to %s") :
                           _("failed to add iptables rule to enable UDP masquerading"),
                           forwardIf);
976 977 978 979
        goto masqerr4;
    }

    /* TCP with a source port restriction */
980
    if (iptablesAddForwardMasquerade(driver->iptables,
981
                                     &ipdef->address,
982
                                     prefix,
983
                                     forwardIf,
984 985
                                     "tcp") < 0) {
        networkReportError(VIR_ERR_SYSTEM_ERROR,
986 987 988 989
                           forwardIf ?
                           _("failed to add iptables rule to enable TCP masquerading to %s") :
                           _("failed to add iptables rule to enable TCP masquerading"),
                           forwardIf);
990 991 992
        goto masqerr5;
    }

993
    return 0;
994

995 996
 masqerr5:
    iptablesRemoveForwardMasquerade(driver->iptables,
997
                                    &ipdef->address,
998
                                    prefix,
999
                                    forwardIf,
1000 1001 1002
                                    "udp");
 masqerr4:
    iptablesRemoveForwardMasquerade(driver->iptables,
1003
                                    &ipdef->address,
1004
                                    prefix,
1005
                                    forwardIf,
1006
                                    NULL);
1007 1008
 masqerr3:
    iptablesRemoveForwardAllowRelatedIn(driver->iptables,
1009
                                        &ipdef->address,
1010
                                        prefix,
1011
                                        network->def->bridge,
1012
                                        forwardIf);
1013 1014
 masqerr2:
    iptablesRemoveForwardAllowOut(driver->iptables,
1015
                                  &ipdef->address,
1016
                                  prefix,
1017
                                  network->def->bridge,
1018
                                  forwardIf);
1019
 masqerr1:
1020
    return -1;
1021 1022
}

1023 1024 1025 1026 1027 1028
static void
networkRemoveMasqueradingIptablesRules(struct network_driver *driver,
                                       virNetworkObjPtr network,
                                       virNetworkIpDefPtr ipdef)
{
    int prefix = virNetworkIpDefPrefix(ipdef);
1029
    const char *forwardIf = virNetworkDefForwardIf(network->def, 0);
1030 1031 1032 1033 1034

    if (prefix >= 0) {
        iptablesRemoveForwardMasquerade(driver->iptables,
                                        &ipdef->address,
                                        prefix,
1035
                                        forwardIf,
1036 1037 1038 1039
                                        "tcp");
        iptablesRemoveForwardMasquerade(driver->iptables,
                                        &ipdef->address,
                                        prefix,
1040
                                        forwardIf,
1041 1042 1043 1044
                                        "udp");
        iptablesRemoveForwardMasquerade(driver->iptables,
                                        &ipdef->address,
                                        prefix,
1045
                                        forwardIf,
1046 1047 1048 1049 1050 1051
                                        NULL);

        iptablesRemoveForwardAllowRelatedIn(driver->iptables,
                                            &ipdef->address,
                                            prefix,
                                            network->def->bridge,
1052
                                            forwardIf);
1053 1054 1055 1056
        iptablesRemoveForwardAllowOut(driver->iptables,
                                      &ipdef->address,
                                      prefix,
                                      network->def->bridge,
1057
                                      forwardIf);
1058 1059 1060
    }
}

1061
static int
1062
networkAddRoutingIptablesRules(struct network_driver *driver,
1063
                               virNetworkObjPtr network,
1064 1065
                               virNetworkIpDefPtr ipdef)
{
1066
    int prefix = virNetworkIpDefPrefix(ipdef);
1067
    const char *forwardIf = virNetworkDefForwardIf(network->def, 0);
1068 1069 1070 1071 1072 1073 1074

    if (prefix < 0) {
        networkReportError(VIR_ERR_INTERNAL_ERROR,
                           _("Invalid prefix or netmask for '%s'"),
                           network->def->bridge);
        goto routeerr1;
    }
1075

1076
    /* allow routing packets from the bridge interface */
1077
    if (iptablesAddForwardAllowOut(driver->iptables,
1078
                                   &ipdef->address,
1079
                                   prefix,
1080
                                   network->def->bridge,
1081
                                   forwardIf) < 0) {
1082 1083 1084
        networkReportError(VIR_ERR_SYSTEM_ERROR,
                           _("failed to add iptables rule to allow routing from '%s'"),
                           network->def->bridge);
1085 1086 1087 1088
        goto routeerr1;
    }

    /* allow routing packets to the bridge interface */
1089
    if (iptablesAddForwardAllowIn(driver->iptables,
1090
                                  &ipdef->address,
1091
                                  prefix,
1092
                                  network->def->bridge,
1093
                                  forwardIf) < 0) {
1094 1095 1096
        networkReportError(VIR_ERR_SYSTEM_ERROR,
                           _("failed to add iptables rule to allow routing to '%s'"),
                           network->def->bridge);
1097 1098 1099
        goto routeerr2;
    }

1100
    return 0;
1101

1102
routeerr2:
1103
    iptablesRemoveForwardAllowOut(driver->iptables,
1104
                                  &ipdef->address,
1105
                                  prefix,
1106
                                  network->def->bridge,
1107
                                  forwardIf);
1108
routeerr1:
1109
    return -1;
1110 1111
}

1112 1113 1114 1115 1116 1117
static void
networkRemoveRoutingIptablesRules(struct network_driver *driver,
                                  virNetworkObjPtr network,
                                  virNetworkIpDefPtr ipdef)
{
    int prefix = virNetworkIpDefPrefix(ipdef);
1118
    const char *forwardIf = virNetworkDefForwardIf(network->def, 0);
1119 1120 1121 1122 1123 1124

    if (prefix >= 0) {
        iptablesRemoveForwardAllowIn(driver->iptables,
                                     &ipdef->address,
                                     prefix,
                                     network->def->bridge,
1125
                                     forwardIf);
1126 1127 1128 1129 1130

        iptablesRemoveForwardAllowOut(driver->iptables,
                                      &ipdef->address,
                                      prefix,
                                      network->def->bridge,
1131
                                      forwardIf);
1132 1133 1134
    }
}

1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170
/* 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;
    }

1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187
    /* 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;
    }

1188 1189 1190
    return 0;

    /* unwind in reverse order from the point of failure */
1191 1192 1193 1194
err5:
    iptablesRemoveTcpInput(driver->iptables, AF_INET6, network->def->bridge, 53);
err4:
    iptablesRemoveForwardAllowCross(driver->iptables, AF_INET6, network->def->bridge);
1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214
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);
}

1215
static int
1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229
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;
    }
1230 1231

    /* allow DHCP requests through to dnsmasq */
1232

1233 1234
    if (iptablesAddTcpInput(driver->iptables, AF_INET,
                            network->def->bridge, 67) < 0) {
1235 1236 1237
        networkReportError(VIR_ERR_SYSTEM_ERROR,
                           _("failed to add iptables rule to allow DHCP requests from '%s'"),
                           network->def->bridge);
1238 1239 1240
        goto err1;
    }

1241 1242
    if (iptablesAddUdpInput(driver->iptables, AF_INET,
                            network->def->bridge, 67) < 0) {
1243 1244 1245
        networkReportError(VIR_ERR_SYSTEM_ERROR,
                           _("failed to add iptables rule to allow DHCP requests from '%s'"),
                           network->def->bridge);
1246 1247 1248
        goto err2;
    }

1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259
    /* 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);
1260
        VIR_WARN("May need to update iptables package & kernel to support CHECKSUM rule.");
1261 1262
    }

1263
    /* allow DNS requests through to dnsmasq */
1264 1265
    if (iptablesAddTcpInput(driver->iptables, AF_INET,
                            network->def->bridge, 53) < 0) {
1266 1267 1268
        networkReportError(VIR_ERR_SYSTEM_ERROR,
                           _("failed to add iptables rule to allow DNS requests from '%s'"),
                           network->def->bridge);
1269 1270 1271
        goto err3;
    }

1272 1273
    if (iptablesAddUdpInput(driver->iptables, AF_INET,
                            network->def->bridge, 53) < 0) {
1274 1275 1276
        networkReportError(VIR_ERR_SYSTEM_ERROR,
                           _("failed to add iptables rule to allow DNS requests from '%s'"),
                           network->def->bridge);
1277 1278 1279
        goto err4;
    }

1280 1281
    /* allow TFTP requests through to dnsmasq if necessary */
    if (ipv4def && ipv4def->tftproot &&
1282 1283
        iptablesAddUdpInput(driver->iptables, AF_INET,
                            network->def->bridge, 69) < 0) {
1284 1285 1286
        networkReportError(VIR_ERR_SYSTEM_ERROR,
                           _("failed to add iptables rule to allow TFTP requests from '%s'"),
                           network->def->bridge);
1287
        goto err5;
1288 1289
    }

1290 1291
    /* Catch all rules to block forwarding to/from bridges */

1292 1293
    if (iptablesAddForwardRejectOut(driver->iptables, AF_INET,
                                    network->def->bridge) < 0) {
1294 1295 1296
        networkReportError(VIR_ERR_SYSTEM_ERROR,
                           _("failed to add iptables rule to block outbound traffic from '%s'"),
                           network->def->bridge);
1297
        goto err6;
1298 1299
    }

1300 1301
    if (iptablesAddForwardRejectIn(driver->iptables, AF_INET,
                                   network->def->bridge) < 0) {
1302 1303 1304
        networkReportError(VIR_ERR_SYSTEM_ERROR,
                           _("failed to add iptables rule to block inbound traffic to '%s'"),
                           network->def->bridge);
1305
        goto err7;
1306 1307 1308
    }

    /* Allow traffic between guests on the same bridge */
1309 1310
    if (iptablesAddForwardAllowCross(driver->iptables, AF_INET,
                                     network->def->bridge) < 0) {
1311 1312 1313
        networkReportError(VIR_ERR_SYSTEM_ERROR,
                           _("failed to add iptables rule to allow cross bridge traffic on '%s'"),
                           network->def->bridge);
1314
        goto err8;
1315 1316
    }

1317 1318 1319 1320 1321
    /* add IPv6 general rules, if needed */
    if (networkAddGeneralIp6tablesRules(driver, network) < 0) {
        goto err9;
    }

1322
    return 0;
1323

1324
    /* unwind in reverse order from the point of failure */
1325 1326
err9:
    iptablesRemoveForwardAllowCross(driver->iptables, AF_INET, network->def->bridge);
1327
err8:
1328
    iptablesRemoveForwardRejectIn(driver->iptables, AF_INET, network->def->bridge);
1329
err7:
1330
    iptablesRemoveForwardRejectOut(driver->iptables, AF_INET, network->def->bridge);
1331 1332
err6:
    if (ipv4def && ipv4def->tftproot) {
1333
        iptablesRemoveUdpInput(driver->iptables, AF_INET, network->def->bridge, 69);
1334
    }
1335
err5:
1336
    iptablesRemoveUdpInput(driver->iptables, AF_INET, network->def->bridge, 53);
1337
err4:
1338
    iptablesRemoveTcpInput(driver->iptables, AF_INET, network->def->bridge, 53);
1339
err3:
1340
    iptablesRemoveUdpInput(driver->iptables, AF_INET, network->def->bridge, 67);
1341
err2:
1342
    iptablesRemoveTcpInput(driver->iptables, AF_INET, network->def->bridge, 67);
1343
err1:
1344
    return -1;
1345 1346 1347
}

static void
1348 1349 1350 1351 1352
networkRemoveGeneralIptablesRules(struct network_driver *driver,
                                  virNetworkObjPtr network)
{
    int ii;
    virNetworkIpDefPtr ipv4def;
1353

1354 1355
    networkRemoveGeneralIp6tablesRules(driver, network);

1356 1357 1358 1359 1360
    for (ii = 0;
         (ipv4def = virNetworkDefGetIpByIndex(network->def, AF_INET, ii));
         ii++) {
        if (ipv4def->nranges || ipv4def->nhosts || ipv4def->tftproot)
            break;
1361
    }
1362

1363 1364 1365
    iptablesRemoveForwardAllowCross(driver->iptables, AF_INET, network->def->bridge);
    iptablesRemoveForwardRejectIn(driver->iptables, AF_INET, network->def->bridge);
    iptablesRemoveForwardRejectOut(driver->iptables, AF_INET, network->def->bridge);
1366
    if (ipv4def && ipv4def->tftproot) {
1367
        iptablesRemoveUdpInput(driver->iptables, AF_INET, network->def->bridge, 69);
1368
    }
1369 1370
    iptablesRemoveUdpInput(driver->iptables, AF_INET, network->def->bridge, 53);
    iptablesRemoveTcpInput(driver->iptables, AF_INET, network->def->bridge, 53);
1371 1372 1373 1374
    if (ipv4def && (ipv4def->nranges || ipv4def->nhosts)) {
        iptablesRemoveOutputFixUdpChecksum(driver->iptables,
                                           network->def->bridge, 68);
    }
1375 1376
    iptablesRemoveUdpInput(driver->iptables, AF_INET, network->def->bridge, 67);
    iptablesRemoveTcpInput(driver->iptables, AF_INET, network->def->bridge, 67);
1377 1378
}

1379 1380 1381 1382 1383
static int
networkAddIpSpecificIptablesRules(struct network_driver *driver,
                                  virNetworkObjPtr network,
                                  virNetworkIpDefPtr ipdef)
{
1384 1385 1386
    /* NB: in the case of IPv6, routing rules are added when the
     * forward mode is NAT. This is because IPv6 has no NAT.
     */
1387

1388
    if (network->def->forwardType == VIR_NETWORK_FORWARD_NAT) {
1389
        if (VIR_SOCKET_ADDR_IS_FAMILY(&ipdef->address, AF_INET))
1390
            return networkAddMasqueradingIptablesRules(driver, network, ipdef);
1391
        else if (VIR_SOCKET_ADDR_IS_FAMILY(&ipdef->address, AF_INET6))
1392 1393 1394 1395
            return networkAddRoutingIptablesRules(driver, network, ipdef);
    } else if (network->def->forwardType == VIR_NETWORK_FORWARD_ROUTE) {
        return networkAddRoutingIptablesRules(driver, network, ipdef);
    }
1396 1397 1398 1399 1400 1401 1402 1403
    return 0;
}

static void
networkRemoveIpSpecificIptablesRules(struct network_driver *driver,
                                     virNetworkObjPtr network,
                                     virNetworkIpDefPtr ipdef)
{
1404
    if (network->def->forwardType == VIR_NETWORK_FORWARD_NAT) {
1405
        if (VIR_SOCKET_ADDR_IS_FAMILY(&ipdef->address, AF_INET))
1406
            networkRemoveMasqueradingIptablesRules(driver, network, ipdef);
1407
        else if (VIR_SOCKET_ADDR_IS_FAMILY(&ipdef->address, AF_INET6))
1408 1409
            networkRemoveRoutingIptablesRules(driver, network, ipdef);
    } else if (network->def->forwardType == VIR_NETWORK_FORWARD_ROUTE) {
1410
        networkRemoveRoutingIptablesRules(driver, network, ipdef);
1411
    }
1412 1413 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
}

/* 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);
}

1465 1466 1467 1468 1469
static void
networkReloadIptablesRules(struct network_driver *driver)
{
    unsigned int i;

1470
    VIR_INFO("Reloading iptables rules");
1471 1472 1473 1474

    for (i = 0 ; i < driver->networks.count ; i++) {
        virNetworkObjLock(driver->networks.objs[i]);
        if (virNetworkObjIsActive(driver->networks.objs[i])) {
1475 1476 1477 1478
            networkRemoveIptablesRules(driver, driver->networks.objs[i]);
            if (networkAddIptablesRules(driver, driver->networks.objs[i]) < 0) {
                /* failed to add but already logged */
            }
1479 1480 1481 1482 1483
        }
        virNetworkObjUnlock(driver->networks.objs[i]);
    }
}

1484
/* Enable IP Forwarding. Return 0 for success, -1 for failure. */
1485
static int
1486
networkEnableIpForwarding(bool enableIPv4, bool enableIPv6)
1487
{
1488 1489 1490 1491 1492 1493
    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;
1494 1495
}

1496 1497
#define SYSCTL_PATH "/proc/sys"

1498 1499
static int
networkSetIPv6Sysctls(virNetworkObjPtr network)
1500 1501 1502 1503
{
    char *field = NULL;
    int ret = -1;

1504 1505 1506 1507 1508 1509 1510 1511 1512
    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;
        }
1513

1514 1515 1516 1517 1518 1519
        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;
        }
1520

1521 1522 1523 1524 1525 1526 1527
        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);
1528 1529
    }

1530 1531 1532 1533 1534 1535 1536 1537 1538
    /* 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) {
1539
        virReportOOMError();
1540 1541 1542
        goto cleanup;
    }

1543
    if (virFileWriteStr(field, "0", 0) < 0) {
1544
        virReportSystemError(errno,
1545 1546 1547 1548 1549
                             _("cannot disable %s"), field);
        goto cleanup;
    }
    VIR_FREE(field);

1550 1551 1552 1553 1554
    /* 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) {
1555
        virReportOOMError();
1556 1557 1558
        goto cleanup;
    }

1559
    if (virFileWriteStr(field, "0", 0) < 0) {
1560
        virReportSystemError(errno,
1561
                             _("cannot disable %s"), field);
1562 1563 1564 1565 1566 1567 1568 1569 1570
        goto cleanup;
    }

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

1571 1572 1573 1574 1575 1576
#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
 */
1577 1578
static int
networkCheckRouteCollision(virNetworkObjPtr network)
1579
{
1580
    int ret = 0, len;
1581 1582 1583 1584 1585
    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)
1586
        goto out;
1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604

    /* 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;
1605 1606
        virNetworkIpDefPtr ipdef;
        int num, ii;
1607 1608 1609 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

        /* 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;

1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658
        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;
            }
1659 1660 1661 1662 1663 1664 1665 1666
        }
    }

out:
    VIR_FREE(buf);
    return ret;
}

1667
static int
D
Daniel P. Berrange 已提交
1668
networkAddAddrToBridge(virNetworkObjPtr network,
1669
                       virNetworkIpDefPtr ipdef)
1670
{
1671 1672 1673 1674 1675 1676 1677 1678 1679
    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;
    }

1680 1681
    if (virNetDevSetIPv4Address(network->def->bridge,
                                &ipdef->address, prefix) < 0)
1682 1683 1684 1685 1686 1687
        return -1;

    return 0;
}

static int
1688
networkStartNetworkVirtual(struct network_driver *driver,
1689 1690
                          virNetworkObjPtr network)
{
1691
    int ii;
1692
    bool v4present = false, v6present = false;
1693 1694
    virErrorPtr save_err = NULL;
    virNetworkIpDefPtr ipdef;
1695
    char *macTapIfName = NULL;
1696

1697 1698
    /* Check to see if any network IP collides with an existing route */
    if (networkCheckRouteCollision(network) < 0)
1699 1700
        return -1;

1701
    /* Create and configure the bridge device */
1702
    if (virNetDevBridgeCreate(network->def->bridge) < 0)
1703 1704
        return -1;

1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716
    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;
        }
1717 1718
        if (virNetDevTapCreateInBridgePort(network->def->bridge,
                                           &macTapIfName, network->def->mac, 0, false, NULL) < 0) {
1719 1720 1721 1722 1723
            VIR_FREE(macTapIfName);
            goto err0;
        }
    }

1724
    /* Set bridge options */
1725
    if (virNetDevBridgeSetSTPDelay(network->def->bridge,
1726
                          network->def->delay) < 0)
1727
        goto err1;
1728

1729
    if (virNetDevBridgeSetSTP(network->def->bridge,
1730
                              network->def->stp ? true : false) < 0)
1731
        goto err1;
1732

1733 1734 1735 1736
    /* Disable IPv6 on the bridge if there are no IPv6 addresses
     * defined, and set other IPv6 sysctl tunables appropriately.
     */
    if (networkSetIPv6Sysctls(network) < 0)
1737
        goto err1;
1738

1739 1740 1741 1742 1743 1744 1745
    /* Add "once per network" rules */
    if (networkAddIptablesRules(driver, network) < 0)
        goto err1;

    for (ii = 0;
         (ipdef = virNetworkDefGetIpByIndex(network->def, AF_UNSPEC, ii));
         ii++) {
1746
        if (VIR_SOCKET_ADDR_IS_FAMILY(&ipdef->address, AF_INET))
1747
            v4present = true;
1748
        if (VIR_SOCKET_ADDR_IS_FAMILY(&ipdef->address, AF_INET6))
1749
            v6present = true;
1750

1751
        /* Add the IP address/netmask to the bridge */
D
Daniel P. Berrange 已提交
1752
        if (networkAddAddrToBridge(network, ipdef) < 0) {
1753
            goto err2;
1754
        }
1755 1756
    }

1757
    /* Bring up the bridge interface */
1758
    if (virNetDevSetOnline(network->def->bridge, 1) < 0)
1759
        goto err2;
1760

1761
    /* If forwardType != NONE, turn on global IP forwarding */
1762
    if (network->def->forwardType != VIR_NETWORK_FORWARD_NONE &&
1763
        networkEnableIpForwarding(v4present, v6present) < 0) {
1764
        virReportSystemError(errno, "%s",
1765
                             _("failed to enable IP forwarding"));
1766
        goto err3;
1767 1768
    }

1769

1770 1771
    /* start dnsmasq if there are any IP addresses (v4 or v6) */
    if ((v4present || v6present) && networkStartDhcpDaemon(network) < 0)
1772
        goto err3;
1773

1774 1775 1776 1777
    /* start radvd if there are any ipv6 addresses */
    if (v6present && networkStartRadvd(network) < 0)
        goto err4;

1778
    if (virNetDevBandwidthSet(network->def->bridge, network->def->bandwidth) < 0) {
1779 1780 1781 1782 1783 1784
        networkReportError(VIR_ERR_INTERNAL_ERROR,
                           _("cannot set bandwidth limits on %s"),
                           network->def->bridge);
        goto err5;
    }

1785
    VIR_FREE(macTapIfName);
1786 1787 1788

    return 0;

1789
 err5:
1790
    ignore_value(virNetDevBandwidthClear(network->def->bridge));
1791

1792 1793 1794 1795
 err4:
    if (!save_err)
        save_err = virSaveLastError();

1796 1797 1798 1799 1800
    if (network->dnsmasqPid > 0) {
        kill(network->dnsmasqPid, SIGTERM);
        network->dnsmasqPid = -1;
    }

1801 1802 1803
 err3:
    if (!save_err)
        save_err = virSaveLastError();
1804
    ignore_value(virNetDevSetOnline(network->def->bridge, 0));
1805

1806 1807 1808 1809 1810 1811
 err2:
    if (!save_err)
        save_err = virSaveLastError();
    networkRemoveIptablesRules(driver, network);

 err1:
1812 1813 1814
    if (!save_err)
        save_err = virSaveLastError();

H
Hu Tao 已提交
1815 1816 1817 1818
    if (macTapIfName) {
        ignore_value(virNetDevTapDelete(macTapIfName));
        VIR_FREE(macTapIfName);
    }
1819 1820

 err0:
1821 1822
    if (!save_err)
        save_err = virSaveLastError();
1823
    ignore_value(virNetDevBridgeDelete(network->def->bridge));
1824

1825 1826 1827 1828
    if (save_err) {
        virSetError(save_err);
        virFreeError(save_err);
    }
1829 1830 1831
    return -1;
}

1832
static int networkShutdownNetworkVirtual(struct network_driver *driver,
1833 1834
                                        virNetworkObjPtr network)
{
1835
    ignore_value(virNetDevBandwidthClear(network->def->bridge));
1836

1837 1838 1839 1840 1841 1842 1843 1844
    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 {
1845
            virPidFileDelete(NETWORK_PID_DIR, radvdpidbase);
1846 1847 1848 1849
            VIR_FREE(radvdpidbase);
        }
    }

1850 1851 1852
    if (network->dnsmasqPid > 0)
        kill(network->dnsmasqPid, SIGTERM);

1853
    if (network->def->mac_specified) {
1854
        char *macTapIfName = networkBridgeDummyNicName(network->def->bridge);
1855 1856 1857
        if (!macTapIfName) {
            virReportOOMError();
        } else {
1858
            ignore_value(virNetDevTapDelete(macTapIfName));
1859 1860 1861 1862
            VIR_FREE(macTapIfName);
        }
    }

1863
    ignore_value(virNetDevSetOnline(network->def->bridge, 0));
1864

1865 1866
    networkRemoveIptablesRules(driver, network);

1867
    ignore_value(virNetDevBridgeDelete(network->def->bridge));
1868

1869
    /* See if its still alive and really really kill it */
1870
    if (network->dnsmasqPid > 0 &&
1871
        (kill(network->dnsmasqPid, 0) == 0))
1872 1873
        kill(network->dnsmasqPid, SIGKILL);
    network->dnsmasqPid = -1;
1874 1875 1876 1877 1878 1879

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

1880 1881 1882 1883 1884 1885 1886 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
    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;
    }

1993 1994 1995 1996 1997 1998 1999 2000
    network->active = 0;

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

2001
    return ret;
2002 2003 2004
}


2005 2006 2007 2008 2009
static virNetworkPtr networkLookupByUUID(virConnectPtr conn,
                                         const unsigned char *uuid) {
    struct network_driver *driver = conn->networkPrivateData;
    virNetworkObjPtr network;
    virNetworkPtr ret = NULL;
2010

2011
    networkDriverLock(driver);
2012
    network = virNetworkFindByUUID(&driver->networks, uuid);
2013
    networkDriverUnlock(driver);
2014
    if (!network) {
2015 2016
        networkReportError(VIR_ERR_NO_NETWORK,
                           "%s", _("no network with matching uuid"));
2017
        goto cleanup;
2018 2019
    }

2020 2021 2022
    ret = virGetNetwork(conn, network->def->name, network->def->uuid);

cleanup:
2023 2024
    if (network)
        virNetworkObjUnlock(network);
2025
    return ret;
2026 2027
}

2028 2029 2030 2031 2032 2033
static virNetworkPtr networkLookupByName(virConnectPtr conn,
                                         const char *name) {
    struct network_driver *driver = conn->networkPrivateData;
    virNetworkObjPtr network;
    virNetworkPtr ret = NULL;

2034
    networkDriverLock(driver);
2035
    network = virNetworkFindByName(&driver->networks, name);
2036
    networkDriverUnlock(driver);
2037
    if (!network) {
2038 2039
        networkReportError(VIR_ERR_NO_NETWORK,
                           _("no network with matching name '%s'"), name);
2040
        goto cleanup;
2041 2042
    }

2043 2044 2045
    ret = virGetNetwork(conn, network->def->name, network->def->uuid);

cleanup:
2046 2047
    if (network)
        virNetworkObjUnlock(network);
2048
    return ret;
2049 2050 2051 2052
}

static virDrvOpenStatus networkOpenNetwork(virConnectPtr conn,
                                           virConnectAuthPtr auth ATTRIBUTE_UNUSED,
2053 2054 2055 2056
                                           unsigned int flags)
{
    virCheckFlags(VIR_CONNECT_RO, VIR_DRV_OPEN_ERROR);

2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069
    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) {
2070
    int nactive = 0, i;
2071
    struct network_driver *driver = conn->networkPrivateData;
2072

2073 2074 2075
    networkDriverLock(driver);
    for (i = 0 ; i < driver->networks.count ; i++) {
        virNetworkObjLock(driver->networks.objs[i]);
D
Daniel P. Berrange 已提交
2076
        if (virNetworkObjIsActive(driver->networks.objs[i]))
2077
            nactive++;
2078 2079 2080
        virNetworkObjUnlock(driver->networks.objs[i]);
    }
    networkDriverUnlock(driver);
2081

2082 2083 2084 2085
    return nactive;
}

static int networkListNetworks(virConnectPtr conn, char **const names, int nnames) {
2086
    struct network_driver *driver = conn->networkPrivateData;
2087
    int got = 0, i;
2088

2089
    networkDriverLock(driver);
2090
    for (i = 0 ; i < driver->networks.count && got < nnames ; i++) {
2091
        virNetworkObjLock(driver->networks.objs[i]);
D
Daniel P. Berrange 已提交
2092
        if (virNetworkObjIsActive(driver->networks.objs[i])) {
2093
            if (!(names[got] = strdup(driver->networks.objs[i]->def->name))) {
2094
                virNetworkObjUnlock(driver->networks.objs[i]);
2095
                virReportOOMError();
2096 2097 2098 2099
                goto cleanup;
            }
            got++;
        }
2100
        virNetworkObjUnlock(driver->networks.objs[i]);
2101
    }
2102 2103
    networkDriverUnlock(driver);

2104 2105 2106
    return got;

 cleanup:
2107
    networkDriverUnlock(driver);
2108 2109 2110 2111 2112 2113
    for (i = 0 ; i < got ; i++)
        VIR_FREE(names[i]);
    return -1;
}

static int networkNumDefinedNetworks(virConnectPtr conn) {
2114
    int ninactive = 0, i;
2115
    struct network_driver *driver = conn->networkPrivateData;
2116

2117 2118 2119
    networkDriverLock(driver);
    for (i = 0 ; i < driver->networks.count ; i++) {
        virNetworkObjLock(driver->networks.objs[i]);
D
Daniel P. Berrange 已提交
2120
        if (!virNetworkObjIsActive(driver->networks.objs[i]))
2121
            ninactive++;
2122 2123 2124
        virNetworkObjUnlock(driver->networks.objs[i]);
    }
    networkDriverUnlock(driver);
2125

2126 2127 2128 2129
    return ninactive;
}

static int networkListDefinedNetworks(virConnectPtr conn, char **const names, int nnames) {
2130
    struct network_driver *driver = conn->networkPrivateData;
2131
    int got = 0, i;
2132

2133
    networkDriverLock(driver);
2134
    for (i = 0 ; i < driver->networks.count && got < nnames ; i++) {
2135
        virNetworkObjLock(driver->networks.objs[i]);
D
Daniel P. Berrange 已提交
2136
        if (!virNetworkObjIsActive(driver->networks.objs[i])) {
2137
            if (!(names[got] = strdup(driver->networks.objs[i]->def->name))) {
2138
                virNetworkObjUnlock(driver->networks.objs[i]);
2139
                virReportOOMError();
2140 2141 2142 2143
                goto cleanup;
            }
            got++;
        }
2144
        virNetworkObjUnlock(driver->networks.objs[i]);
2145
    }
2146
    networkDriverUnlock(driver);
2147 2148 2149
    return got;

 cleanup:
2150
    networkDriverUnlock(driver);
2151 2152 2153 2154 2155
    for (i = 0 ; i < got ; i++)
        VIR_FREE(names[i]);
    return -1;
}

2156 2157 2158

static int networkIsActive(virNetworkPtr net)
{
2159
    struct network_driver *driver = net->conn->networkPrivateData;
2160 2161 2162 2163 2164 2165 2166
    virNetworkObjPtr obj;
    int ret = -1;

    networkDriverLock(driver);
    obj = virNetworkFindByUUID(&driver->networks, net->uuid);
    networkDriverUnlock(driver);
    if (!obj) {
2167
        networkReportError(VIR_ERR_NO_NETWORK, NULL);
2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179
        goto cleanup;
    }
    ret = virNetworkObjIsActive(obj);

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

static int networkIsPersistent(virNetworkPtr net)
{
2180
    struct network_driver *driver = net->conn->networkPrivateData;
2181 2182 2183 2184 2185 2186 2187
    virNetworkObjPtr obj;
    int ret = -1;

    networkDriverLock(driver);
    obj = virNetworkFindByUUID(&driver->networks, net->uuid);
    networkDriverUnlock(driver);
    if (!obj) {
2188
        networkReportError(VIR_ERR_NO_NETWORK, NULL);
2189 2190 2191 2192 2193 2194 2195 2196 2197 2198 2199
        goto cleanup;
    }
    ret = obj->persistent;

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


2200
static virNetworkPtr networkCreate(virConnectPtr conn, const char *xml) {
2201
    struct network_driver *driver = conn->networkPrivateData;
2202
    virNetworkDefPtr def;
2203
    virNetworkObjPtr network = NULL;
2204
    virNetworkPtr ret = NULL;
2205

2206 2207
    networkDriverLock(driver);

2208
    if (!(def = virNetworkDefParseString(xml)))
2209
        goto cleanup;
2210

2211 2212 2213
    if (virNetworkObjIsDuplicate(&driver->networks, def, 1) < 0)
        goto cleanup;

2214 2215 2216 2217 2218 2219
    /* 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) {
2220

2221 2222 2223 2224 2225
        if (virNetworkSetBridgeName(&driver->networks, def, 1))
            goto cleanup;

        virNetworkSetBridgeMacAddr(def);
    }
2226

2227
    if (!(network = virNetworkAssignDef(&driver->networks,
2228 2229 2230
                                        def)))
        goto cleanup;
    def = NULL;
2231

2232
    if (networkStartNetwork(driver, network) < 0) {
2233 2234
        virNetworkRemoveInactive(&driver->networks,
                                 network);
2235
        network = NULL;
2236
        goto cleanup;
2237 2238
    }

2239
    VIR_INFO("Creating network '%s'", network->def->name);
2240 2241 2242 2243
    ret = virGetNetwork(conn, network->def->name, network->def->uuid);

cleanup:
    virNetworkDefFree(def);
2244 2245 2246
    if (network)
        virNetworkObjUnlock(network);
    networkDriverUnlock(driver);
2247
    return ret;
2248 2249 2250
}

static virNetworkPtr networkDefine(virConnectPtr conn, const char *xml) {
2251
    struct network_driver *driver = conn->networkPrivateData;
2252
    virNetworkIpDefPtr ipdef, ipv4def = NULL;
2253
    virNetworkDefPtr def;
2254
    bool freeDef = true;
2255
    virNetworkObjPtr network = NULL;
2256
    virNetworkPtr ret = NULL;
2257
    int ii;
2258
    dnsmasqContext* dctx = NULL;
2259

2260 2261
    networkDriverLock(driver);

2262
    if (!(def = virNetworkDefParseString(xml)))
2263
        goto cleanup;
2264

E
Eric Blake 已提交
2265
    if (virNetworkObjIsDuplicate(&driver->networks, def, 0) < 0)
2266 2267
        goto cleanup;

2268 2269 2270 2271 2272 2273
    /* 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) {
2274

2275 2276 2277 2278 2279
        if (virNetworkSetBridgeName(&driver->networks, def, 1))
            goto cleanup;

        virNetworkSetBridgeMacAddr(def);
    }
2280

2281
    if (!(network = virNetworkAssignDef(&driver->networks,
2282 2283
                                        def)))
        goto cleanup;
2284
    freeDef = false;
2285

2286 2287
    network->persistent = 1;

2288 2289
    if (virNetworkSaveConfig(driver->networkConfigDir, def) < 0) {
        virNetworkRemoveInactive(&driver->networks, network);
2290
        network = NULL;
2291
        goto cleanup;
2292 2293
    }

2294
    /* We only support dhcp on one IPv4 address per defined network */
2295
    for (ii = 0;
2296
         (ipdef = virNetworkDefGetIpByIndex(def, AF_UNSPEC, ii));
2297
         ii++) {
2298
        if (VIR_SOCKET_ADDR_IS_FAMILY(&ipdef->address, AF_INET)) {
2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310
            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) {
2311
        dctx = dnsmasqContextNew(def->name, DNSMASQ_STATE_DIR);
2312
        if (dctx == NULL ||
2313
            networkBuildDnsmasqHostsfile(dctx, ipv4def, def->dns) < 0 ||
2314
            dnsmasqSave(dctx) < 0)
2315 2316 2317
            goto cleanup;
    }

2318 2319
    VIR_INFO("Defining network '%s'", def->name);
    ret = virGetNetwork(conn, def->name, def->uuid);
2320 2321

cleanup:
2322 2323
    if (freeDef)
       virNetworkDefFree(def);
2324
    dnsmasqContextFree(dctx);
2325 2326 2327
    if (network)
        virNetworkObjUnlock(network);
    networkDriverUnlock(driver);
2328
    return ret;
2329 2330 2331
}

static int networkUndefine(virNetworkPtr net) {
2332
    struct network_driver *driver = net->conn->networkPrivateData;
2333
    virNetworkObjPtr network;
2334 2335
    virNetworkIpDefPtr ipdef;
    bool dhcp_present = false, v6present = false;
2336
    int ret = -1, ii;
2337

2338 2339
    networkDriverLock(driver);

2340
    network = virNetworkFindByUUID(&driver->networks, net->uuid);
2341
    if (!network) {
2342
        networkReportError(VIR_ERR_NO_NETWORK,
2343 2344
                           "%s", _("no network with matching uuid"));
        goto cleanup;
2345 2346
    }

D
Daniel P. Berrange 已提交
2347
    if (virNetworkObjIsActive(network)) {
2348
        networkReportError(VIR_ERR_OPERATION_INVALID,
2349 2350
                           "%s", _("network is still active"));
        goto cleanup;
2351 2352
    }

2353
    if (virNetworkDeleteConfig(driver->networkConfigDir,
2354 2355
                               driver->networkAutostartDir,
                               network) < 0)
2356
        goto cleanup;
2357

2358 2359
    /* we only support dhcp on one IPv4 address per defined network */
    for (ii = 0;
2360
         (ipdef = virNetworkDefGetIpByIndex(network->def, AF_UNSPEC, ii));
2361
         ii++) {
2362
        if (VIR_SOCKET_ADDR_IS_FAMILY(&ipdef->address, AF_INET)) {
2363 2364
            if (ipdef->nranges || ipdef->nhosts)
                dhcp_present = true;
2365
        } else if (VIR_SOCKET_ADDR_IS_FAMILY(&ipdef->address, AF_INET6)) {
2366 2367
            v6present = true;
        }
2368
    }
2369 2370

    if (dhcp_present) {
2371
        char *leasefile;
2372 2373 2374 2375 2376 2377
        dnsmasqContext *dctx = dnsmasqContextNew(network->def->name, DNSMASQ_STATE_DIR);
        if (dctx == NULL)
            goto cleanup;

        dnsmasqDelete(dctx);
        dnsmasqContextFree(dctx);
2378 2379 2380 2381 2382 2383

        leasefile = networkDnsmasqLeaseFileName(network->def->name);
        if (!leasefile)
            goto cleanup;
        unlink(leasefile);
        VIR_FREE(leasefile);
2384 2385
    }

2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401
    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;
        }
2402
        virPidFileDelete(NETWORK_PID_DIR, radvdpidbase);
2403 2404 2405 2406
        VIR_FREE(radvdpidbase);

    }

2407
    VIR_INFO("Undefining network '%s'", network->def->name);
2408 2409
    virNetworkRemoveInactive(&driver->networks,
                             network);
2410
    network = NULL;
2411
    ret = 0;
2412

2413
cleanup:
2414 2415 2416
    if (network)
        virNetworkObjUnlock(network);
    networkDriverUnlock(driver);
2417
    return ret;
2418 2419 2420
}

static int networkStart(virNetworkPtr net) {
2421 2422 2423
    struct network_driver *driver = net->conn->networkPrivateData;
    virNetworkObjPtr network;
    int ret = -1;
2424

2425
    networkDriverLock(driver);
2426
    network = virNetworkFindByUUID(&driver->networks, net->uuid);
2427

2428
    if (!network) {
2429
        networkReportError(VIR_ERR_NO_NETWORK,
2430 2431
                           "%s", _("no network with matching uuid"));
        goto cleanup;
2432 2433
    }

2434
    ret = networkStartNetwork(driver, network);
2435 2436

cleanup:
2437 2438
    if (network)
        virNetworkObjUnlock(network);
2439
    networkDriverUnlock(driver);
2440
    return ret;
2441 2442 2443
}

static int networkDestroy(virNetworkPtr net) {
2444 2445 2446
    struct network_driver *driver = net->conn->networkPrivateData;
    virNetworkObjPtr network;
    int ret = -1;
2447

2448
    networkDriverLock(driver);
2449
    network = virNetworkFindByUUID(&driver->networks, net->uuid);
2450

2451
    if (!network) {
2452
        networkReportError(VIR_ERR_NO_NETWORK,
2453 2454
                           "%s", _("no network with matching uuid"));
        goto cleanup;
2455 2456
    }

D
Daniel P. Berrange 已提交
2457
    if (!virNetworkObjIsActive(network)) {
2458
        networkReportError(VIR_ERR_OPERATION_INVALID,
2459 2460 2461 2462
                           "%s", _("network is not active"));
        goto cleanup;
    }

2463
    ret = networkShutdownNetwork(driver, network);
2464
    if (!network->persistent) {
2465 2466 2467 2468
        virNetworkRemoveInactive(&driver->networks,
                                 network);
        network = NULL;
    }
2469

2470
cleanup:
2471 2472
    if (network)
        virNetworkObjUnlock(network);
2473
    networkDriverUnlock(driver);
2474 2475 2476
    return ret;
}

2477
static char *networkGetXMLDesc(virNetworkPtr net,
2478
                               unsigned int flags)
2479
{
2480 2481 2482
    struct network_driver *driver = net->conn->networkPrivateData;
    virNetworkObjPtr network;
    char *ret = NULL;
2483

2484 2485
    virCheckFlags(0, NULL);

2486
    networkDriverLock(driver);
2487
    network = virNetworkFindByUUID(&driver->networks, net->uuid);
2488 2489
    networkDriverUnlock(driver);

2490
    if (!network) {
2491
        networkReportError(VIR_ERR_NO_NETWORK,
2492 2493
                           "%s", _("no network with matching uuid"));
        goto cleanup;
2494 2495
    }

2496
    ret = virNetworkDefFormat(network->def);
2497 2498

cleanup:
2499 2500
    if (network)
        virNetworkObjUnlock(network);
2501
    return ret;
2502 2503 2504
}

static char *networkGetBridgeName(virNetworkPtr net) {
2505 2506 2507 2508
    struct network_driver *driver = net->conn->networkPrivateData;
    virNetworkObjPtr network;
    char *bridge = NULL;

2509
    networkDriverLock(driver);
2510
    network = virNetworkFindByUUID(&driver->networks, net->uuid);
2511 2512
    networkDriverUnlock(driver);

2513
    if (!network) {
2514
        networkReportError(VIR_ERR_NO_NETWORK,
2515 2516
                           "%s", _("no network with matching id"));
        goto cleanup;
2517 2518
    }

2519
    if (!(network->def->bridge)) {
2520
        networkReportError(VIR_ERR_INTERNAL_ERROR,
2521 2522 2523 2524 2525
                           _("network '%s' does not have a bridge name."),
                           network->def->name);
        goto cleanup;
    }

2526
    bridge = strdup(network->def->bridge);
2527
    if (!bridge)
2528
        virReportOOMError();
2529 2530

cleanup:
2531 2532
    if (network)
        virNetworkObjUnlock(network);
2533 2534 2535 2536 2537
    return bridge;
}

static int networkGetAutostart(virNetworkPtr net,
                             int *autostart) {
2538 2539 2540
    struct network_driver *driver = net->conn->networkPrivateData;
    virNetworkObjPtr network;
    int ret = -1;
2541

2542
    networkDriverLock(driver);
2543
    network = virNetworkFindByUUID(&driver->networks, net->uuid);
2544
    networkDriverUnlock(driver);
2545
    if (!network) {
2546
        networkReportError(VIR_ERR_NO_NETWORK,
2547
                           "%s", _("no network with matching uuid"));
2548
        goto cleanup;
2549 2550 2551
    }

    *autostart = network->autostart;
2552
    ret = 0;
2553

2554
cleanup:
2555 2556
    if (network)
        virNetworkObjUnlock(network);
2557
    return ret;
2558 2559 2560
}

static int networkSetAutostart(virNetworkPtr net,
2561
                               int autostart) {
2562 2563
    struct network_driver *driver = net->conn->networkPrivateData;
    virNetworkObjPtr network;
2564
    char *configFile = NULL, *autostartLink = NULL;
2565
    int ret = -1;
2566

2567
    networkDriverLock(driver);
2568
    network = virNetworkFindByUUID(&driver->networks, net->uuid);
2569

2570
    if (!network) {
2571
        networkReportError(VIR_ERR_NO_NETWORK,
2572
                           "%s", _("no network with matching uuid"));
2573
        goto cleanup;
2574 2575
    }

2576
    if (!network->persistent) {
2577
        networkReportError(VIR_ERR_OPERATION_INVALID,
2578
                           "%s", _("cannot set autostart for transient network"));
2579 2580 2581
        goto cleanup;
    }

2582 2583
    autostart = (autostart != 0);

2584
    if (network->autostart != autostart) {
2585
        if ((configFile = virNetworkConfigFile(driver->networkConfigDir, network->def->name)) == NULL)
2586
            goto cleanup;
2587
        if ((autostartLink = virNetworkConfigFile(driver->networkAutostartDir, network->def->name)) == NULL)
2588 2589
            goto cleanup;

2590
        if (autostart) {
2591
            if (virFileMakePath(driver->networkAutostartDir) < 0) {
2592
                virReportSystemError(errno,
2593 2594
                                     _("cannot create autostart directory '%s'"),
                                     driver->networkAutostartDir);
2595 2596
                goto cleanup;
            }
2597

2598
            if (symlink(configFile, autostartLink) < 0) {
2599
                virReportSystemError(errno,
2600
                                     _("Failed to create symlink '%s' to '%s'"),
2601
                                     autostartLink, configFile);
2602 2603 2604
                goto cleanup;
            }
        } else {
2605
            if (unlink(autostartLink) < 0 && errno != ENOENT && errno != ENOTDIR) {
2606
                virReportSystemError(errno,
2607
                                     _("Failed to delete symlink '%s'"),
2608
                                     autostartLink);
2609 2610
                goto cleanup;
            }
2611 2612
        }

2613
        network->autostart = autostart;
2614
    }
2615
    ret = 0;
2616

2617
cleanup:
2618 2619
    VIR_FREE(configFile);
    VIR_FREE(autostartLink);
2620 2621
    if (network)
        virNetworkObjUnlock(network);
2622
    networkDriverUnlock(driver);
2623
    return ret;
2624 2625 2626 2627 2628
}


static virNetworkDriver networkDriver = {
    "Network",
2629 2630 2631 2632 2633 2634 2635 2636 2637 2638 2639 2640 2641 2642 2643 2644 2645 2646 2647
    .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 */
2648 2649 2650
};

static virStateDriver networkStateDriver = {
2651
    "Network",
2652 2653 2654 2655 2656 2657 2658 2659 2660 2661 2662
    networkStartup,
    networkShutdown,
    networkReload,
    networkActive,
};

int networkRegister(void) {
    virRegisterNetworkDriver(&networkDriver);
    virRegisterStateDriver(&networkStateDriver);
    return 0;
}
2663 2664 2665 2666 2667 2668 2669 2670 2671 2672 2673 2674 2675 2676 2677 2678 2679 2680 2681 2682 2683 2684 2685 2686 2687 2688 2689 2690

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

/* 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;
2691
    virPortGroupDefPtr portgroup;
2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704 2705 2706 2707 2708 2709 2710
    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;
2711 2712 2713 2714 2715 2716 2717 2718 2719 2720 2721 2722 2723 2724 2725 2726 2727 2728

    /* 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;
        }

2729 2730
        if (virNetDevBandwidthCopy(&iface->data.network.actual->bandwidth,
                                   portgroup->bandwidth) < 0)
2731 2732 2733 2734 2735 2736 2737 2738 2739 2740 2741 2742 2743 2744
            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) {
2745 2746 2747 2748 2749

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

2750 2751
        if (!iface->data.network.actual
            && (VIR_ALLOC(iface->data.network.actual) < 0)) {
2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766
            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)) {
2767
        virNetDevVPortProfilePtr virtport = NULL;
2768 2769 2770 2771 2772

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

2773 2774
        if (!iface->data.network.actual
            && (VIR_ALLOC(iface->data.network.actual) < 0)) {
2775 2776 2777 2778 2779 2780 2781 2782
            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:
2783
            iface->data.network.actual->data.direct.mode = VIR_NETDEV_MACVLAN_MODE_BRIDGE;
2784 2785
            break;
        case VIR_NETWORK_FORWARD_PRIVATE:
2786
            iface->data.network.actual->data.direct.mode = VIR_NETDEV_MACVLAN_MODE_PRIVATE;
2787 2788
            break;
        case VIR_NETWORK_FORWARD_VEPA:
2789
            iface->data.network.actual->data.direct.mode = VIR_NETDEV_MACVLAN_MODE_VEPA;
2790 2791
            break;
        case VIR_NETWORK_FORWARD_PASSTHROUGH:
2792
            iface->data.network.actual->data.direct.mode = VIR_NETDEV_MACVLAN_MODE_PASSTHRU;
2793 2794 2795 2796 2797 2798 2799 2800 2801 2802 2803 2804 2805 2806 2807 2808 2809 2810 2811 2812 2813 2814
            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;
        }
2815

2816 2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834 2835 2836 2837 2838
        /* 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
2839
                  == VIR_NETDEV_VPORT_PROFILE_8021QBH))) {
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
                /* 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
2966
               == VIR_NETDEV_VPORT_PROFILE_8021QBH)))) {
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
            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;
}
3069 3070 3071 3072 3073 3074 3075 3076 3077 3078 3079 3080 3081 3082 3083 3084 3085 3086 3087 3088 3089 3090 3091 3092 3093 3094 3095 3096 3097 3098

/*
 * 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;
3099
    char *dev_name = NULL;
3100 3101 3102 3103 3104 3105 3106 3107 3108 3109 3110 3111 3112 3113 3114 3115 3116 3117 3118 3119 3120 3121 3122 3123 3124 3125 3126 3127 3128

    *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:
3129
        if ((dev_name = netdef->bridge))
3130 3131 3132 3133 3134 3135 3136 3137 3138
            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)
3139
            dev_name = netdef->forwardIfs[0].dev;
3140

3141
        if (!dev_name) {
3142 3143 3144 3145 3146 3147 3148
            networkReportError(VIR_ERR_INTERNAL_ERROR,
                               _("network '%s' has no associated interface or bridge"),
                               netdef->name);
        }
        break;
    }

3149
    if (dev_name) {
3150 3151 3152 3153
        if (virNetDevGetIPv4Address(dev_name, &addr) < 0)
            goto cleanup;

        addrptr = &addr;
3154 3155 3156
    }

    if (addrptr &&
3157
        (*netaddr = virSocketAddrFormat(addrptr))) {
3158 3159 3160 3161 3162 3163 3164 3165
        ret = 0;
    }

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