lxc_controller.c 57.4 KB
Newer Older
1
/*
2 3
 * Copyright (C) 2010-2011 Red Hat, Inc.
 * Copyright IBM Corp. 2008
4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
 *
 * lxc_controller.c: linux container process controller
 *
 * Authors:
 *  David L. Leskovec <dlesko at linux.vnet.ibm.com>
 *
 * 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
 */

#include <config.h>

#include <sys/epoll.h>
28 29
#include <sys/wait.h>
#include <sys/socket.h>
30 31
#include <sys/types.h>
#include <sys/un.h>
32 33
#include <sys/utsname.h>
#include <sys/personality.h>
34
#include <unistd.h>
35
#include <paths.h>
36
#include <errno.h>
37 38
#include <fcntl.h>
#include <signal.h>
39
#include <getopt.h>
40
#include <sys/mount.h>
E
Eric Blake 已提交
41
#include <locale.h>
42 43
#include <linux/loop.h>
#include <dirent.h>
44 45
#include <grp.h>
#include <sys/stat.h>
46
#include <time.h>
47

D
Daniel P. Berrange 已提交
48
#if HAVE_CAPNG
49
# include <cap-ng.h>
D
Daniel P. Berrange 已提交
50 51
#endif

52 53 54 55 56
#if HAVE_NUMACTL
# define NUMA_VERSION1_COMPATIBILITY 1
# include <numa.h>
#endif

57
#include "virterror_internal.h"
58
#include "logging.h"
59 60 61
#include "util.h"

#include "lxc_conf.h"
62
#include "lxc_container.h"
63 64
#include "virnetdev.h"
#include "virnetdevveth.h"
65 66
#include "memory.h"
#include "util.h"
E
Eric Blake 已提交
67
#include "virfile.h"
68
#include "virpidfile.h"
69
#include "command.h"
70 71
#include "processinfo.h"
#include "nodeinfo.h"
72
#include "virrandom.h"
73

74 75
#define VIR_FROM_THIS VIR_FROM_LXC

D
Dan Smith 已提交
76 77 78 79 80 81
struct cgroup_device_policy {
    char type;
    int major;
    int minor;
};

82

83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106
typedef struct _virLXCControllerConsole virLXCControllerConsole;
typedef virLXCControllerConsole *virLXCControllerConsolePtr;
struct _virLXCControllerConsole {
    int hostWatch;
    int hostFd;  /* PTY FD in the host OS */
    bool hostClosed;
    int hostEpoll;
    bool hostBlocking;

    int contWatch;
    int contFd;  /* PTY FD in the container */
    bool contClosed;
    int contEpoll;
    bool contBlocking;

    int epollWatch;
    int epollFd; /* epoll FD for dealing with EOF */

    size_t fromHostLen;
    char fromHostBuf[1024];
    size_t fromContLen;
    char fromContBuf[1024];
};

107 108 109 110 111
typedef struct _virLXCController virLXCController;
typedef virLXCController *virLXCControllerPtr;
struct _virLXCController {
    char *name;
    virDomainDefPtr def;
112

113 114
    int handshakeFd;

115 116
    pid_t initpid;

117 118
    size_t nveths;
    char **veths;
119 120 121

    size_t nconsoles;
    virLXCControllerConsolePtr consoles;
122
    char *devptmx;
123 124 125

    size_t nloopDevs;
    int *loopDevFds;
126 127

    virSecurityManagerPtr securityManager;
128 129 130 131 132

    int monitorServerFd;
    int monitorServerWatch;
    int monitorClientFd;
    int monitorClientWatch;
133 134 135 136 137 138 139 140 141 142 143 144 145
};

static void virLXCControllerFree(virLXCControllerPtr ctrl);

static virLXCControllerPtr virLXCControllerNew(const char *name)
{
    virLXCControllerPtr ctrl = NULL;
    virCapsPtr caps = NULL;
    char *configFile = NULL;

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

146 147 148 149 150
    ctrl->monitorServerFd = -1;
    ctrl->monitorServerWatch = -1;
    ctrl->monitorClientFd = -1;
    ctrl->monitorClientWatch = -1;

151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179
    if (!(ctrl->name = strdup(name)))
        goto no_memory;

    if ((caps = lxcCapsInit(NULL)) == NULL)
        goto error;

    if ((configFile = virDomainConfigFile(LXC_STATE_DIR,
                                          ctrl->name)) == NULL)
        goto error;

    if ((ctrl->def = virDomainDefParseFile(caps,
                                           configFile,
                                           1 << VIR_DOMAIN_VIRT_LXC,
                                           0)) == NULL)
        goto error;

cleanup:
    VIR_FREE(configFile);
    virCapabilitiesFree(caps);
    return ctrl;

no_memory:
    virReportOOMError();
error:
    virLXCControllerFree(ctrl);
    ctrl = NULL;
    goto cleanup;
}

180

181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201
static int virLXCControllerCloseLoopDevices(virLXCControllerPtr ctrl,
                                            bool force)
{
    size_t i;

    for (i = 0 ; i < ctrl->nloopDevs ; i++) {
        if (force) {
            VIR_FORCE_CLOSE(ctrl->loopDevFds[i]);
        } else {
            if (VIR_CLOSE(ctrl->loopDevFds[i]) < 0) {
                virReportSystemError(errno, "%s",
                                     _("Unable to close loop device"));
                return -1;
            }
        }
    }

    return 0;
}


202 203 204 205 206
static void virLXCControllerStopInit(virLXCControllerPtr ctrl)
{
    if (ctrl->initpid == 0)
        return;

207
    virLXCControllerCloseLoopDevices(ctrl, true);
208 209 210 211 212
    virPidAbort(ctrl->initpid);
    ctrl->initpid = 0;
}


213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228
static void virLXCControllerConsoleClose(virLXCControllerConsolePtr console)
{
    if (console->hostWatch != -1)
        virEventRemoveHandle(console->hostWatch);
    VIR_FORCE_CLOSE(console->hostFd);

    if (console->contWatch != -1)
        virEventRemoveHandle(console->contWatch);
    VIR_FORCE_CLOSE(console->contFd);

    if (console->epollWatch != -1)
        virEventRemoveHandle(console->epollWatch);
    VIR_FORCE_CLOSE(console->epollFd);
}


229 230
static void virLXCControllerFree(virLXCControllerPtr ctrl)
{
231 232
    size_t i;

233 234 235
    if (!ctrl)
        return;

236 237
    virLXCControllerStopInit(ctrl);

238 239
    virSecurityManagerFree(ctrl->securityManager);

240 241 242 243
    for (i = 0 ; i < ctrl->nveths ; i++)
        VIR_FREE(ctrl->veths[i]);
    VIR_FREE(ctrl->veths);

244 245 246 247
    for (i = 0 ; i < ctrl->nconsoles ; i++)
        virLXCControllerConsoleClose(&(ctrl->consoles[i]));
    VIR_FREE(ctrl->consoles);

248 249
    VIR_FORCE_CLOSE(ctrl->handshakeFd);

250 251
    VIR_FREE(ctrl->devptmx);

252 253 254
    virDomainDefFree(ctrl->def);
    VIR_FREE(ctrl->name);

255 256 257 258 259 260 261
    if (ctrl->monitorServerWatch != -1)
        virEventRemoveHandle(ctrl->monitorServerWatch);
    VIR_FORCE_CLOSE(ctrl->monitorServerFd);
    if (ctrl->monitorClientWatch != -1)
        virEventRemoveHandle(ctrl->monitorClientWatch);
    VIR_FORCE_CLOSE(ctrl->monitorClientFd);

262 263 264
    VIR_FREE(ctrl);
}

265

266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297
static int virLXCControllerAddConsole(virLXCControllerPtr ctrl,
                                      int hostFd)
{
    if (VIR_EXPAND_N(ctrl->consoles, ctrl->nconsoles, 1) < 0) {
        virReportOOMError();
        return -1;
    }
    ctrl->consoles[ctrl->nconsoles-1].hostFd = hostFd;
    ctrl->consoles[ctrl->nconsoles-1].hostWatch = -1;

    ctrl->consoles[ctrl->nconsoles-1].contFd = -1;
    ctrl->consoles[ctrl->nconsoles-1].contWatch = -1;

    ctrl->consoles[ctrl->nconsoles-1].epollFd = -1;
    ctrl->consoles[ctrl->nconsoles-1].epollWatch = -1;
    return 0;
}


static int virLXCControllerConsoleSetNonblocking(virLXCControllerConsolePtr console)
{
    if (virSetBlocking(console->hostFd, false) < 0 ||
        virSetBlocking(console->contFd, false) < 0) {
        virReportSystemError(errno, "%s",
                             _("Unable to set console file descriptor non-blocking"));
        return -1;
    }

    return 0;
}


298 299 300 301 302 303 304 305 306 307 308 309
static int virLXCControllerDaemonHandshake(virLXCControllerPtr ctrl)
{
    if (lxcContainerSendContinue(ctrl->handshakeFd) < 0) {
        virReportSystemError(errno, "%s",
                             _("error sending continue signal to daemon"));
        return -1;
    }
    VIR_FORCE_CLOSE(ctrl->handshakeFd);
    return 0;
}


310 311 312 313 314 315 316 317 318 319 320 321 322
static int virLXCControllerValidateNICs(virLXCControllerPtr ctrl)
{
    if (ctrl->def->nnets != ctrl->nveths) {
        lxcError(VIR_ERR_INTERNAL_ERROR,
                 _("expecting %d veths, but got %zu"),
                 ctrl->def->nnets, ctrl->nveths);
        return -1;
    }

    return 0;
}


323 324 325 326 327 328 329 330 331 332 333 334 335
static int virLXCControllerValidateConsoles(virLXCControllerPtr ctrl)
{
    if (ctrl->def->nconsoles != ctrl->nconsoles) {
        lxcError(VIR_ERR_INTERNAL_ERROR,
                 _("expecting %d consoles, but got %zu tty file handlers"),
                 ctrl->def->nconsoles, ctrl->nconsoles);
        return -1;
    }

    return 0;
}


336
static int lxcGetLoopFD(char **dev_name)
337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390
{
    int fd = -1;
    DIR *dh = NULL;
    struct dirent *de;
    char *looppath;
    struct loop_info64 lo;

    VIR_DEBUG("Looking for loop devices in /dev");

    if (!(dh = opendir("/dev"))) {
        virReportSystemError(errno, "%s",
                             _("Unable to read /dev"));
        goto cleanup;
    }

    while ((de = readdir(dh)) != NULL) {
        if (!STRPREFIX(de->d_name, "loop"))
            continue;

        if (virAsprintf(&looppath, "/dev/%s", de->d_name) < 0) {
            virReportOOMError();
            goto cleanup;
        }

        VIR_DEBUG("Checking up on device %s", looppath);
        if ((fd = open(looppath, O_RDWR)) < 0) {
            virReportSystemError(errno,
                                 _("Unable to open %s"), looppath);
            goto cleanup;
        }

        if (ioctl(fd, LOOP_GET_STATUS64, &lo) < 0) {
            /* Got a free device, return the fd */
            if (errno == ENXIO)
                goto cleanup;

            VIR_FORCE_CLOSE(fd);
            virReportSystemError(errno,
                                 _("Unable to get loop status on %s"),
                                 looppath);
            goto cleanup;
        }

        /* Oh well, try the next device */
        VIR_FORCE_CLOSE(fd);
        VIR_FREE(looppath);
    }

    lxcError(VIR_ERR_INTERNAL_ERROR, "%s",
             _("Unable to find a free loop device in /dev"));

cleanup:
    if (fd != -1) {
        VIR_DEBUG("Got free loop device %s %d", looppath, fd);
391
        *dev_name = looppath;
392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457
    } else {
        VIR_DEBUG("No free loop devices available");
        VIR_FREE(looppath);
    }
    if (dh)
        closedir(dh);
    return fd;
}

static int lxcSetupLoopDevice(virDomainFSDefPtr fs)
{
    int lofd = -1;
    int fsfd = -1;
    struct loop_info64 lo;
    char *loname = NULL;
    int ret = -1;

    if ((lofd = lxcGetLoopFD(&loname)) < 0)
        return -1;

    memset(&lo, 0, sizeof(lo));
    lo.lo_flags = LO_FLAGS_AUTOCLEAR;

    if ((fsfd = open(fs->src, O_RDWR)) < 0) {
        virReportSystemError(errno,
                             _("Unable to open %s"), fs->src);
        goto cleanup;
    }

    if (ioctl(lofd, LOOP_SET_FD, fsfd) < 0) {
        virReportSystemError(errno,
                             _("Unable to attach %s to loop device"),
                             fs->src);
        goto cleanup;
    }

    if (ioctl(lofd, LOOP_SET_STATUS64, &lo) < 0) {
        virReportSystemError(errno, "%s",
                             _("Unable to mark loop device as autoclear"));

        if (ioctl(lofd, LOOP_CLR_FD, 0) < 0)
            VIR_WARN("Unable to detach %s from loop device", fs->src);
        goto cleanup;
    }

    VIR_DEBUG("Attached loop device  %s %d to %s", fs->src, lofd, loname);
    /*
     * We now change it into a block device type, so that
     * the rest of container setup 'just works'
     */
    fs->type = VIR_DOMAIN_FS_TYPE_BLOCK;
    VIR_FREE(fs->src);
    fs->src = loname;
    loname = NULL;

    ret = 0;

cleanup:
    VIR_FREE(loname);
    VIR_FORCE_CLOSE(fsfd);
    if (ret == -1)
        VIR_FORCE_CLOSE(lofd);
    return lofd;
}


458
static int virLXCControllerSetupLoopDevices(virLXCControllerPtr ctrl)
459 460 461 462
{
    size_t i;
    int ret = -1;

463
    for (i = 0 ; i < ctrl->def->nfss ; i++) {
464 465
        int fd;

466
        if (ctrl->def->fss[i]->type != VIR_DOMAIN_FS_TYPE_FILE)
467 468
            continue;

469
        fd = lxcSetupLoopDevice(ctrl->def->fss[i]);
470 471 472 473
        if (fd < 0)
            goto cleanup;

        VIR_DEBUG("Saving loop fd %d", fd);
474
        if (VIR_EXPAND_N(ctrl->loopDevFds, ctrl->nloopDevs, 1) < 0) {
475 476 477 478
            VIR_FORCE_CLOSE(fd);
            virReportOOMError();
            goto cleanup;
        }
479
        ctrl->loopDevFds[ctrl->nloopDevs - 1] = fd;
480 481 482 483 484 485 486 487 488
    }

    VIR_DEBUG("Setup all loop devices");
    ret = 0;

cleanup:
    return ret;
}

489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581
#if HAVE_NUMACTL
static int lxcSetContainerNUMAPolicy(virDomainDefPtr def)
{
    nodemask_t mask;
    int mode = -1;
    int node = -1;
    int ret = -1;
    int i = 0;
    int maxnode = 0;
    bool warned = false;

    if (!def->numatune.memory.nodemask)
        return 0;

    VIR_DEBUG("Setting NUMA memory policy");

    if (numa_available() < 0) {
        lxcError(VIR_ERR_CONFIG_UNSUPPORTED,
                 "%s", _("Host kernel is not aware of NUMA."));
        return -1;
    }

    maxnode = numa_max_node() + 1;

    /* Convert nodemask to NUMA bitmask. */
    nodemask_zero(&mask);
    for (i = 0; i < VIR_DOMAIN_CPUMASK_LEN; i++) {
        if (def->numatune.memory.nodemask[i]) {
            if (i > NUMA_NUM_NODES) {
                lxcError(VIR_ERR_CONFIG_UNSUPPORTED,
                         _("Host cannot support NUMA node %d"), i);
                return -1;
            }
            if (i > maxnode && !warned) {
                VIR_WARN("nodeset is out of range, there is only %d NUMA "
                         "nodes on host", maxnode);
                warned = true;
            }
            nodemask_set(&mask, i);
        }
    }

    mode = def->numatune.memory.mode;

    if (mode == VIR_DOMAIN_NUMATUNE_MEM_STRICT) {
        numa_set_bind_policy(1);
        numa_set_membind(&mask);
        numa_set_bind_policy(0);
    } else if (mode == VIR_DOMAIN_NUMATUNE_MEM_PREFERRED) {
        int nnodes = 0;
        for (i = 0; i < NUMA_NUM_NODES; i++) {
            if (nodemask_isset(&mask, i)) {
                node = i;
                nnodes++;
            }
        }

        if (nnodes != 1) {
            lxcError(VIR_ERR_CONFIG_UNSUPPORTED,
                     "%s", _("NUMA memory tuning in 'preferred' mode "
                             "only supports single node"));
            goto cleanup;
        }

        numa_set_bind_policy(0);
        numa_set_preferred(node);
    } else if (mode == VIR_DOMAIN_NUMATUNE_MEM_INTERLEAVE) {
        numa_set_interleave_mask(&mask);
    } else {
        lxcError(VIR_ERR_CONFIG_UNSUPPORTED,
                 _("Unable to set NUMA policy %s"),
                 virDomainNumatuneMemModeTypeToString(mode));
        goto cleanup;
    }

    ret = 0;

cleanup:
    return ret;
}
#else
static int lxcSetContainerNUMAPolicy(virDomainDefPtr def)
{
    if (def->numatune.memory.nodemask) {
        lxcError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                 _("NUMA policy is not available on this platform"));
        return -1;
    }

    return 0;
}
#endif

582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639

/*
 * To be run while still single threaded
 */
static int lxcSetContainerCpuAffinity(virDomainDefPtr def)
{
    int i, hostcpus, maxcpu = CPU_SETSIZE;
    virNodeInfo nodeinfo;
    unsigned char *cpumap;
    int cpumaplen;

    VIR_DEBUG("Setting CPU affinity");

    if (nodeGetInfo(NULL, &nodeinfo) < 0)
        return -1;

    /* setaffinity fails if you set bits for CPUs which
     * aren't present, so we have to limit ourselves */
    hostcpus = VIR_NODEINFO_MAXCPUS(nodeinfo);
    if (maxcpu > hostcpus)
        maxcpu = hostcpus;

    cpumaplen = VIR_CPU_MAPLEN(maxcpu);
    if (VIR_ALLOC_N(cpumap, cpumaplen) < 0) {
        virReportOOMError();
        return -1;
    }

    if (def->cpumask) {
        /* XXX why don't we keep 'cpumask' in the libvirt cpumap
         * format to start with ?!?! */
        for (i = 0 ; i < maxcpu && i < def->cpumasklen ; i++)
            if (def->cpumask[i])
                VIR_USE_CPU(cpumap, i);
    } else {
        /* You may think this is redundant, but we can't assume libvirtd
         * itself is running on all pCPUs, so we need to explicitly set
         * the spawned LXC instance to all pCPUs if no map is given in
         * its config file */
        for (i = 0 ; i < maxcpu ; i++)
            VIR_USE_CPU(cpumap, i);
    }

    /* We are pressuming we are running between fork/exec of LXC
     * so use '0' to indicate our own process ID. No threads are
     * running at this point
     */
    if (virProcessInfoSetAffinity(0, /* Self */
                                  cpumap, cpumaplen, maxcpu) < 0) {
        VIR_FREE(cpumap);
        return -1;
    }
    VIR_FREE(cpumap);

    return 0;
}


640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675
static int lxcSetContainerCpuTune(virCgroupPtr cgroup, virDomainDefPtr def)
{
    int ret = -1;
    if (def->cputune.shares != 0) {
        int rc = virCgroupSetCpuShares(cgroup, def->cputune.shares);
        if (rc != 0) {
            virReportSystemError(-rc,
                                 _("Unable to set io cpu shares for domain %s"),
                                 def->name);
            goto cleanup;
        }
    }
    if (def->cputune.quota != 0) {
        int rc = virCgroupSetCpuCfsQuota(cgroup, def->cputune.quota);
        if (rc != 0) {
            virReportSystemError(-rc,
                                 _("Unable to set io cpu quota for domain %s"),
                                 def->name);
            goto cleanup;
        }
    }
    if (def->cputune.period != 0) {
        int rc = virCgroupSetCpuCfsPeriod(cgroup, def->cputune.period);
        if (rc != 0) {
            virReportSystemError(-rc,
                                 _("Unable to set io cpu period for domain %s"),
                                 def->name);
            goto cleanup;
        }
    }
    ret = 0;
cleanup:
    return ret;
}


676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695
static int lxcSetContainerBlkioTune(virCgroupPtr cgroup, virDomainDefPtr def)
{
    int ret = -1;

    if (def->blkio.weight) {
        int rc = virCgroupSetBlkioWeight(cgroup, def->blkio.weight);
        if (rc != 0) {
            virReportSystemError(-rc,
                                 _("Unable to set Blkio weight for domain %s"),
                                 def->name);
            goto cleanup;
        }
    }

    ret = 0;
cleanup:
    return ret;
}


696
static int lxcSetContainerMemTune(virCgroupPtr cgroup, virDomainDefPtr def)
D
Dan Smith 已提交
697
{
698 699
    int ret = -1;
    int rc;
700

701
    rc = virCgroupSetMemory(cgroup, def->mem.max_balloon);
702
    if (rc != 0) {
703
        virReportSystemError(-rc,
704 705
                             _("Unable to set memory limit for domain %s"),
                             def->name);
706
        goto cleanup;
707
    }
D
Dan Smith 已提交
708

709
    if (def->mem.hard_limit) {
710 711 712 713 714 715 716 717 718
        rc = virCgroupSetMemoryHardLimit(cgroup, def->mem.hard_limit);
        if (rc != 0) {
            virReportSystemError(-rc,
                                 _("Unable to set memory hard limit for domain %s"),
                                 def->name);
            goto cleanup;
        }
    }

719
    if (def->mem.soft_limit) {
720 721 722 723 724 725 726 727 728
        rc = virCgroupSetMemorySoftLimit(cgroup, def->mem.soft_limit);
        if (rc != 0) {
            virReportSystemError(-rc,
                                 _("Unable to set memory soft limit for domain %s"),
                                 def->name);
            goto cleanup;
        }
    }

729
    if (def->mem.swap_hard_limit) {
730
        rc = virCgroupSetMemSwapHardLimit(cgroup, def->mem.swap_hard_limit);
731 732 733 734 735 736 737 738
        if (rc != 0) {
            virReportSystemError(-rc,
                                 _("Unable to set swap hard limit for domain %s"),
                                 def->name);
            goto cleanup;
        }
    }

739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759
    ret = 0;
cleanup:
    return ret;
}


static int lxcSetContainerDeviceACL(virCgroupPtr cgroup, virDomainDefPtr def)
{
    int ret = -1;
    int rc;
    size_t i;
    static const struct cgroup_device_policy devices[] = {
        {'c', LXC_DEV_MAJ_MEMORY, LXC_DEV_MIN_NULL},
        {'c', LXC_DEV_MAJ_MEMORY, LXC_DEV_MIN_ZERO},
        {'c', LXC_DEV_MAJ_MEMORY, LXC_DEV_MIN_FULL},
        {'c', LXC_DEV_MAJ_MEMORY, LXC_DEV_MIN_RANDOM},
        {'c', LXC_DEV_MAJ_MEMORY, LXC_DEV_MIN_URANDOM},
        {'c', LXC_DEV_MAJ_TTY, LXC_DEV_MIN_TTY},
        {'c', LXC_DEV_MAJ_TTY, LXC_DEV_MIN_PTMX},
        {0,   0, 0}};

D
Dan Smith 已提交
760
    rc = virCgroupDenyAllDevices(cgroup);
761
    if (rc != 0) {
762
        virReportSystemError(-rc,
763 764 765 766
                             _("Unable to deny devices for domain %s"),
                             def->name);
        goto cleanup;
    }
D
Dan Smith 已提交
767 768

    for (i = 0; devices[i].type != 0; i++) {
769
        const struct cgroup_device_policy *dev = &devices[i];
D
Dan Smith 已提交
770 771 772
        rc = virCgroupAllowDevice(cgroup,
                                  dev->type,
                                  dev->major,
773 774
                                  dev->minor,
                                  VIR_CGROUP_DEVICE_RWM);
775
        if (rc != 0) {
776
            virReportSystemError(-rc,
777 778 779 780
                                 _("Unable to allow device %c:%d:%d for domain %s"),
                                 dev->type, dev->major, dev->minor, def->name);
            goto cleanup;
        }
D
Dan Smith 已提交
781 782
    }

783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799
    for (i = 0 ; i < def->nfss ; i++) {
        if (def->fss[i]->type != VIR_DOMAIN_FS_TYPE_BLOCK)
            continue;

        rc = virCgroupAllowDevicePath(cgroup,
                                      def->fss[i]->src,
                                      def->fss[i]->readonly ?
                                      VIR_CGROUP_DEVICE_READ :
                                      VIR_CGROUP_DEVICE_RW);
        if (rc != 0) {
            virReportSystemError(-rc,
                                 _("Unable to allow device %s for domain %s"),
                                 def->fss[i]->src, def->name);
            goto cleanup;
        }
    }

800 801
    rc = virCgroupAllowDeviceMajor(cgroup, 'c', LXC_DEV_MAJ_PTY,
                                   VIR_CGROUP_DEVICE_RWM);
802
    if (rc != 0) {
803
        virReportSystemError(-rc,
804
                             _("Unable to allow PTY devices for domain %s"),
805 806 807
                             def->name);
        goto cleanup;
    }
808

809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866
    ret = 0;
cleanup:
    return ret;
}


/**
 * lxcSetContainerResources
 * @def: pointer to virtual machine structure
 *
 * Creates a cgroup for the container, moves the task inside,
 * and sets resource limits
 *
 * Returns 0 on success or -1 in case of error
 */
static int lxcSetContainerResources(virDomainDefPtr def)
{
    virCgroupPtr driver;
    virCgroupPtr cgroup;
    int rc = -1;

    if (lxcSetContainerCpuAffinity(def) < 0)
        return -1;

    if (lxcSetContainerNUMAPolicy(def) < 0)
        return -1;

    rc = virCgroupForDriver("lxc", &driver, 1, 0);
    if (rc != 0) {
        /* Skip all if no driver cgroup is configured */
        if (rc == -ENXIO || rc == -ENOENT)
            return 0;

        virReportSystemError(-rc, "%s",
                             _("Unable to get cgroup for driver"));
        return rc;
    }

    rc = virCgroupForDomain(driver, def->name, &cgroup, 1);
    if (rc != 0) {
        virReportSystemError(-rc,
                             _("Unable to create cgroup for domain %s"),
                             def->name);
        goto cleanup;
    }

    if (lxcSetContainerCpuTune(cgroup, def) < 0)
        goto cleanup;

    if (lxcSetContainerBlkioTune(cgroup, def) < 0)
        goto cleanup;

    if (lxcSetContainerMemTune(cgroup, def) < 0)
        goto cleanup;

    if (lxcSetContainerDeviceACL(cgroup, def) < 0)
        goto cleanup;

D
Dan Smith 已提交
867 868
    rc = virCgroupAddTask(cgroup, getpid());
    if (rc != 0) {
869
        virReportSystemError(-rc,
870 871
                             _("Unable to add task %d to cgroup for domain %s"),
                             getpid(), def->name);
D
Dan Smith 已提交
872 873
    }

874 875
cleanup:
    virCgroupFree(&driver);
D
Dan Smith 已提交
876 877 878 879 880
    virCgroupFree(&cgroup);

    return rc;
}

881
static char*lxcMonitorPath(virLXCControllerPtr ctrl)
882 883
{
    char *sockpath;
884 885

    if (virAsprintf(&sockpath, "%s/%s.sock",
886
                    LXC_STATE_DIR, ctrl->def->name) < 0)
887
        virReportOOMError();
888 889 890 891 892 893 894 895 896
    return sockpath;
}

static int lxcMonitorServer(const char *sockpath)
{
    int fd;
    struct sockaddr_un addr;

    if ((fd = socket(PF_UNIX, SOCK_STREAM, 0)) < 0) {
897
        virReportSystemError(errno,
898 899
                             _("failed to create server socket '%s'"),
                             sockpath);
900 901 902 903 904 905
        goto error;
    }

    unlink(sockpath);
    memset(&addr, 0, sizeof(addr));
    addr.sun_family = AF_UNIX;
C
Chris Lalancette 已提交
906
    if (virStrcpyStatic(addr.sun_path, sockpath) == NULL) {
907
        lxcError(VIR_ERR_INTERNAL_ERROR,
C
Chris Lalancette 已提交
908 909 910
                 _("Socket path %s too long for destination"), sockpath);
        goto error;
    }
911 912

    if (bind(fd, (struct sockaddr *) &addr, sizeof(addr)) < 0) {
913
        virReportSystemError(errno,
914 915
                             _("failed to bind server socket '%s'"),
                             sockpath);
916 917 918
        goto error;
    }
    if (listen(fd, 30 /* backlog */ ) < 0) {
919
        virReportSystemError(errno,
920 921
                             _("failed to listen server socket %s"),
                             sockpath);
922 923 924 925 926 927
        goto error;
    }

    return fd;

error:
928
    VIR_FORCE_CLOSE(fd);
929 930
    return -1;
}
931

D
Daniel P. Berrange 已提交
932 933 934 935 936 937 938 939 940

static int lxcControllerClearCapabilities(void)
{
#if HAVE_CAPNG
    int ret;

    capng_clear(CAPNG_SELECT_BOTH);

    if ((ret = capng_apply(CAPNG_SELECT_BOTH)) < 0) {
941
        lxcError(VIR_ERR_INTERNAL_ERROR,
D
Daniel P. Berrange 已提交
942 943 944 945
                 _("failed to apply capabilities: %d"), ret);
        return -1;
    }
#else
946
    VIR_WARN("libcap-ng support not compiled in, unable to clear capabilities");
D
Daniel P. Berrange 已提交
947 948 949 950
#endif
    return 0;
}

951 952 953
/* Return true if it is ok to ignore an accept-after-epoll syscall
   that fails with the specified errno value.  Else false.  */
static bool
954
ignorable_accept_errno(int errnum)
955 956 957 958 959 960 961
{
  return (errnum == EINVAL
          || errnum == ECONNABORTED
          || errnum == EAGAIN
          || errnum == EWOULDBLOCK);
}

962 963 964 965
static bool quit = false;
static virMutex lock;
static int sigpipe[2];

966
static void virLXCControllerSignalChildHandler(int signum ATTRIBUTE_UNUSED)
967 968 969 970
{
    ignore_value(write(sigpipe[1], "1", 1));
}

971 972 973 974
static void virLXCControllerSignalChildIO(int watch ATTRIBUTE_UNUSED,
                                          int fd ATTRIBUTE_UNUSED,
                                          int events ATTRIBUTE_UNUSED,
                                          void *opaque)
975
{
976 977
    char buf[1];
    int ret;
978
    virLXCControllerPtr ctrl = opaque;
979 980 981

    ignore_value(read(sigpipe[0], buf, 1));
    ret = waitpid(-1, NULL, WNOHANG);
982
    if (ret == ctrl->initpid) {
983 984 985 986 987 988 989
        virMutexLock(&lock);
        quit = true;
        virMutexUnlock(&lock);
    }
}


990
static void virLXCControllerClientIO(int watch ATTRIBUTE_UNUSED, int fd, int events, void *opaque)
991
{
992
    virLXCControllerPtr ctrl = opaque;
993 994 995 996 997
    char buf[1024];
    ssize_t ret;

    if (events & (VIR_EVENT_HANDLE_HANGUP |
                  VIR_EVENT_HANDLE_ERROR)) {
998 999
        virEventRemoveHandle(ctrl->monitorClientWatch);
        ctrl->monitorClientWatch = -1;
1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018
        return;
    }

reread:
    ret = read(fd, buf, sizeof(buf));
    if (ret == -1 && errno == EINTR)
        goto reread;
    if (ret == -1 && errno == EAGAIN)
        return;
    if (ret == -1) {
        lxcError(VIR_ERR_INTERNAL_ERROR, "%s",
                 _("Unable to read from monitor client"));
        virMutexLock(&lock);
        quit = true;
        virMutexUnlock(&lock);
        return;
    }
    if (ret == 0) {
        VIR_DEBUG("Client %d gone", fd);
1019 1020 1021
        VIR_FORCE_CLOSE(ctrl->monitorClientFd);
        virEventRemoveHandle(ctrl->monitorClientWatch);
        ctrl->monitorClientWatch = -1;
1022 1023 1024 1025
    }
}


1026
static void virLXCControllerServerAccept(int watch ATTRIBUTE_UNUSED, int fd, int events ATTRIBUTE_UNUSED, void *opaque)
1027
{
1028
    virLXCControllerPtr ctrl = opaque;
1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045
    int client;

    if ((client = accept(fd, NULL, NULL)) < 0) {
        /* First reflex may be simply to declare accept failure
           to be a fatal error.  However, accept may fail when
           a client quits between the above poll and here.
           That case is not fatal, but rather to be expected,
           if not common, so ignore it.  */
        if (ignorable_accept_errno(errno))
            return;
        virReportSystemError(errno, "%s",
                             _("Unable to accept monitor client"));
        virMutexLock(&lock);
        quit = true;
        virMutexUnlock(&lock);
        return;
    }
1046 1047 1048
    VIR_DEBUG("New client %d (old %d)\n", client, ctrl->monitorClientFd);
    VIR_FORCE_CLOSE(ctrl->monitorClientFd);
    virEventRemoveHandle(ctrl->monitorClientWatch);
1049

1050 1051 1052 1053 1054 1055
    ctrl->monitorClientFd = client;
    if ((ctrl->monitorClientWatch = virEventAddHandle(ctrl->monitorClientFd,
                                                      VIR_EVENT_HANDLE_READABLE,
                                                      virLXCControllerClientIO,
                                                      ctrl,
                                                      NULL)) < 0) {
1056 1057 1058 1059 1060 1061 1062 1063 1064
        lxcError(VIR_ERR_INTERNAL_ERROR, "%s",
                 _("Unable to watch client socket"));
        virMutexLock(&lock);
        quit = true;
        virMutexUnlock(&lock);
        return;
    }
}

1065
static void virLXCControllerConsoleUpdateWatch(virLXCControllerConsolePtr console)
1066 1067 1068 1069
{
    int hostEvents = 0;
    int contEvents = 0;

1070
    if (!console->hostClosed || (!console->hostBlocking && console->fromContLen)) {
1071 1072 1073 1074 1075
        if (console->fromHostLen < sizeof(console->fromHostBuf))
            hostEvents |= VIR_EVENT_HANDLE_READABLE;
        if (console->fromContLen)
            hostEvents |= VIR_EVENT_HANDLE_WRITABLE;
    }
1076
    if (!console->contClosed || (!console->contBlocking && console->fromHostLen)) {
1077 1078 1079 1080 1081 1082
        if (console->fromContLen < sizeof(console->fromContBuf))
            contEvents |= VIR_EVENT_HANDLE_READABLE;
        if (console->fromHostLen)
            contEvents |= VIR_EVENT_HANDLE_WRITABLE;
    }

1083 1084 1085
    VIR_DEBUG("Container watch %d=%d host watch %d=%d",
              console->contWatch, contEvents,
              console->hostWatch, hostEvents);
1086 1087
    virEventUpdateHandle(console->contWatch, contEvents);
    virEventUpdateHandle(console->hostWatch, hostEvents);
1088

1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124
    if (console->hostClosed) {
        int events = EPOLLIN | EPOLLET;
        if (console->hostBlocking)
            events |= EPOLLOUT;

        if (events != console->hostEpoll) {
            struct epoll_event event;
            int action = EPOLL_CTL_ADD;
            if (console->hostEpoll)
                action = EPOLL_CTL_MOD;

            VIR_DEBUG("newHostEvents=%x oldHostEvents=%x", events, console->hostEpoll);

            event.events = events;
            event.data.fd = console->hostFd;
            if (epoll_ctl(console->epollFd, action, console->hostFd, &event) < 0) {
                VIR_DEBUG(":fail");
                virReportSystemError(errno, "%s",
                                     _("Unable to add epoll fd"));
                quit = true;
                goto cleanup;
            }
            console->hostEpoll = events;
            VIR_DEBUG("newHostEvents=%x oldHostEvents=%x", events, console->hostEpoll);
        }
    } else if (console->hostEpoll) {
        VIR_DEBUG("Stop epoll oldContEvents=%x", console->hostEpoll);
        if (epoll_ctl(console->epollFd, EPOLL_CTL_DEL, console->hostFd, NULL) < 0) {
            virReportSystemError(errno, "%s",
                                 _("Unable to remove epoll fd"));
                VIR_DEBUG(":fail");
            quit = true;
            goto cleanup;
        }
        console->hostEpoll = 0;
    }
1125

1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164
    if (console->contClosed) {
        int events = EPOLLIN | EPOLLET;
        if (console->contBlocking)
            events |= EPOLLOUT;

        if (events != console->contEpoll) {
            struct epoll_event event;
            int action = EPOLL_CTL_ADD;
            if (console->contEpoll)
                action = EPOLL_CTL_MOD;

            VIR_DEBUG("newContEvents=%x oldContEvents=%x", events, console->contEpoll);

            event.events = events;
            event.data.fd = console->contFd;
            if (epoll_ctl(console->epollFd, action, console->contFd, &event) < 0) {
                virReportSystemError(errno, "%s",
                                     _("Unable to add epoll fd"));
                VIR_DEBUG(":fail");
                quit = true;
                goto cleanup;
            }
            console->contEpoll = events;
            VIR_DEBUG("newHostEvents=%x oldHostEvents=%x", events, console->contEpoll);
        }
    } else if (console->contEpoll) {
        VIR_DEBUG("Stop epoll oldContEvents=%x", console->contEpoll);
        if (epoll_ctl(console->epollFd, EPOLL_CTL_DEL, console->contFd, NULL) < 0) {
            virReportSystemError(errno, "%s",
                                 _("Unable to remove epoll fd"));
                VIR_DEBUG(":fail");
            quit = true;
            goto cleanup;
        }
        console->contEpoll = 0;
    }
cleanup:
    return;
}
1165 1166


1167
static void virLXCControllerConsoleEPoll(int watch, int fd, int events, void *opaque)
1168
{
1169
    virLXCControllerConsolePtr console = opaque;
1170

1171 1172 1173 1174 1175 1176 1177 1178 1179 1180
    virMutexLock(&lock);
    VIR_DEBUG("IO event watch=%d fd=%d events=%d fromHost=%zu fromcont=%zu",
              watch, fd, events,
              console->fromHostLen,
              console->fromContLen);

    while (1) {
        struct epoll_event event;
        int ret;
        ret = epoll_wait(console->epollFd, &event, 1, 0);
1181
        if (ret < 0) {
S
Stefan Berger 已提交
1182
            if (errno == EINTR)
1183 1184 1185 1186 1187 1188 1189
                continue;
            virReportSystemError(errno, "%s",
                                 _("Unable to wait on epoll"));
            quit = true;
            goto cleanup;
        }

1190 1191 1192 1193 1194 1195 1196
        if (ret == 0)
            break;

        VIR_DEBUG("fd=%d hostFd=%d contFd=%d hostEpoll=%x contEpoll=%x",
                  event.data.fd, console->hostFd, console->contFd,
                  console->hostEpoll, console->contEpoll);

1197 1198 1199
        /* If we get HUP+dead PID, we just re-enable the main loop
         * which will see the PID has died and exit */
        if ((event.events & EPOLLIN)) {
1200 1201
            if (event.data.fd == console->hostFd) {
                console->hostClosed = false;
1202
            } else {
1203
                console->contClosed = false;
1204
            }
1205
            virLXCControllerConsoleUpdateWatch(console);
1206 1207 1208 1209 1210
            break;
        }
    }

cleanup:
1211
    virMutexUnlock(&lock);
1212 1213
}

1214
static void virLXCControllerConsoleIO(int watch, int fd, int events, void *opaque)
1215
{
1216
    virLXCControllerConsolePtr console = opaque;
1217 1218

    virMutexLock(&lock);
1219 1220 1221 1222
    VIR_DEBUG("IO event watch=%d fd=%d events=%d fromHost=%zu fromcont=%zu",
              watch, fd, events,
              console->fromHostLen,
              console->fromContLen);
1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278
    if (events & VIR_EVENT_HANDLE_READABLE) {
        char *buf;
        size_t *len;
        size_t avail;
        ssize_t done;
        if (watch == console->hostWatch) {
            buf = console->fromHostBuf;
            len = &console->fromHostLen;
            avail = sizeof(console->fromHostBuf) - *len;
        } else {
            buf = console->fromContBuf;
            len = &console->fromContLen;
            avail = sizeof(console->fromContBuf) - *len;
        }
    reread:
        done = read(fd, buf + *len, avail);
        if (done == -1 && errno == EINTR)
            goto reread;
        if (done == -1 && errno != EAGAIN) {
            virReportSystemError(errno, "%s",
                                 _("Unable to read container pty"));
            goto error;
        }
        if (done > 0) {
            *len += done;
        } else {
            VIR_DEBUG("Read fd %d done %d errno %d", fd, (int)done, errno);
        }
    }

    if (events & VIR_EVENT_HANDLE_WRITABLE) {
        char *buf;
        size_t *len;
        ssize_t done;
        if (watch == console->hostWatch) {
            buf = console->fromContBuf;
            len = &console->fromContLen;
        } else {
            buf = console->fromHostBuf;
            len = &console->fromHostLen;
        }

    rewrite:
        done = write(fd, buf, *len);
        if (done == -1 && errno == EINTR)
            goto rewrite;
        if (done == -1 && errno != EAGAIN) {
            virReportSystemError(errno, "%s",
                                 _("Unable to write to container pty"));
            goto error;
        }
        if (done > 0) {
            memmove(buf, buf + done, (*len - done));
            *len -= done;
        } else {
            VIR_DEBUG("Write fd %d done %d errno %d", fd, (int)done, errno);
1279 1280 1281 1282
            if (watch == console->hostWatch)
                console->hostBlocking = true;
            else
                console->contBlocking = true;
1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294
        }
    }

    if (events & VIR_EVENT_HANDLE_HANGUP) {
        if (watch == console->hostWatch) {
            console->hostClosed = true;
        } else {
            console->contClosed = true;
        }
        VIR_DEBUG("Got EOF on %d %d", watch, fd);
    }

1295
    virLXCControllerConsoleUpdateWatch(console);
1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307
    virMutexUnlock(&lock);
    return;

error:
    virEventRemoveHandle(console->contWatch);
    virEventRemoveHandle(console->hostWatch);
    console->contWatch = console->hostWatch = -1;
    quit = true;
    virMutexUnlock(&lock);
}


1308
/**
1309
 * lxcControllerMain
1310 1311
 * @serverFd: server socket fd to accept client requests
 * @clientFd: initial client which is the libvirtd daemon
1312
 *
1313
 * Processes I/O on consoles and the monitor
1314 1315 1316
 *
 * Returns 0 on success or -1 in case of error
 */
1317
static int virLXCControllerMain(virLXCControllerPtr ctrl)
1318
{
1319
    virErrorPtr err;
1320
    int rc = -1;
1321
    size_t i;
1322 1323 1324 1325 1326

    if (virMutexInit(&lock) < 0)
        goto cleanup2;

    if (pipe2(sigpipe, O_CLOEXEC|O_NONBLOCK) < 0) {
1327
        virReportSystemError(errno, "%s",
1328
                             _("Cannot create signal pipe"));
1329 1330 1331
        goto cleanup;
    }

1332 1333
    if (virEventAddHandle(sigpipe[0],
                          VIR_EVENT_HANDLE_READABLE,
1334 1335
                          virLXCControllerSignalChildIO,
                          ctrl,
1336 1337 1338
                          NULL) < 0) {
        lxcError(VIR_ERR_INTERNAL_ERROR, "%s",
                 _("Unable to watch signal pipe"));
1339 1340
        goto cleanup;
    }
1341

1342
    if (signal(SIGCHLD, virLXCControllerSignalChildHandler) == SIG_ERR) {
1343
        virReportSystemError(errno, "%s",
1344
                             _("Cannot install signal handler"));
1345 1346 1347
        goto cleanup;
    }

1348
    VIR_DEBUG("serverFd=%d clientFd=%d",
1349
              ctrl->monitorServerFd, ctrl->monitorClientFd);
1350 1351
    virResetLastError();

1352 1353 1354 1355 1356
    if ((ctrl->monitorServerWatch = virEventAddHandle(ctrl->monitorServerFd,
                                                      VIR_EVENT_HANDLE_READABLE,
                                                      virLXCControllerServerAccept,
                                                      ctrl,
                                                      NULL)) < 0) {
1357 1358
        lxcError(VIR_ERR_INTERNAL_ERROR, "%s",
                 _("Unable to watch monitor socket"));
1359 1360 1361
        goto cleanup;
    }

1362 1363 1364 1365 1366 1367
    if (ctrl->monitorClientFd != -1 &&
        (ctrl->monitorClientWatch = virEventAddHandle(ctrl->monitorClientFd,
                                                      VIR_EVENT_HANDLE_READABLE,
                                                      virLXCControllerClientIO,
                                                      ctrl,
                                                      NULL)) < 0) {
1368 1369
        lxcError(VIR_ERR_INTERNAL_ERROR, "%s",
                 _("Unable to watch client socket"));
1370 1371 1372
        goto cleanup;
    }

1373 1374
    for (i = 0 ; i < ctrl->nconsoles ; i++) {
        if ((ctrl->consoles[i].epollFd = epoll_create1(EPOLL_CLOEXEC)) < 0) {
1375 1376 1377 1378 1379
            virReportSystemError(errno, "%s",
                                 _("Unable to create epoll fd"));
            goto cleanup;
        }

1380 1381 1382 1383 1384
        if ((ctrl->consoles[i].epollWatch = virEventAddHandle(ctrl->consoles[i].epollFd,
                                                              VIR_EVENT_HANDLE_READABLE,
                                                              virLXCControllerConsoleEPoll,
                                                              &(ctrl->consoles[i]),
                                                              NULL)) < 0) {
1385 1386 1387 1388 1389
            lxcError(VIR_ERR_INTERNAL_ERROR, "%s",
                     _("Unable to watch epoll FD"));
            goto cleanup;
        }

1390 1391 1392 1393 1394
        if ((ctrl->consoles[i].hostWatch = virEventAddHandle(ctrl->consoles[i].hostFd,
                                                             VIR_EVENT_HANDLE_READABLE,
                                                             virLXCControllerConsoleIO,
                                                             &(ctrl->consoles[i]),
                                                             NULL)) < 0) {
1395 1396 1397 1398 1399
            lxcError(VIR_ERR_INTERNAL_ERROR, "%s",
                     _("Unable to watch host console PTY"));
            goto cleanup;
        }

1400 1401 1402 1403 1404
        if ((ctrl->consoles[i].contWatch = virEventAddHandle(ctrl->consoles[i].contFd,
                                                             VIR_EVENT_HANDLE_READABLE,
                                                             virLXCControllerConsoleIO,
                                                             &(ctrl->consoles[i]),
                                                             NULL)) < 0) {
1405 1406 1407 1408
            lxcError(VIR_ERR_INTERNAL_ERROR, "%s",
                     _("Unable to watch host console PTY"));
            goto cleanup;
        }
1409
    }
1410

1411 1412 1413 1414
    virMutexLock(&lock);
    while (!quit) {
        virMutexUnlock(&lock);
        if (virEventRunDefaultImpl() < 0)
1415
            goto cleanup;
1416
        virMutexLock(&lock);
1417
    }
1418
    virMutexUnlock(&lock);
1419

1420 1421 1422
    err = virGetLastError();
    if (!err || err->code == VIR_ERR_OK)
        rc = 0;
1423 1424

cleanup:
1425 1426 1427
    virMutexDestroy(&lock);
    signal(SIGCHLD, SIG_DFL);
cleanup2:
1428

1429 1430
    for (i = 0 ; i < ctrl->nconsoles ; i++)
        virLXCControllerConsoleClose(&(ctrl->consoles[i]));
1431

1432 1433 1434
    return rc;
}

1435 1436 1437


/**
1438
 * virLXCControllerMoveInterfaces
1439 1440 1441 1442 1443 1444 1445 1446
 * @nveths: number of interfaces
 * @veths: interface names
 * @container: pid of container
 *
 * Moves network interfaces into a container's namespace
 *
 * Returns 0 on success or -1 in case of error
 */
1447
static int virLXCControllerMoveInterfaces(virLXCControllerPtr ctrl)
1448
{
1449 1450 1451
    size_t i;

    for (i = 0 ; i < ctrl->nveths ; i++) {
1452
        if (virNetDevSetNamespace(ctrl->veths[i], ctrl->initpid) < 0)
1453
            return -1;
1454
    }
1455 1456 1457 1458 1459 1460

    return 0;
}


/**
1461 1462
 * virLXCControllerDeleteInterfaces:
 * @ctrl: the LXC controller
1463 1464 1465 1466 1467
 *
 * Cleans up the container interfaces by deleting the veth device pairs.
 *
 * Returns 0 on success or -1 in case of error
 */
1468
static int virLXCControllerDeleteInterfaces(virLXCControllerPtr ctrl)
1469
{
1470 1471
    size_t i;
    int ret = 0;
1472

1473 1474 1475 1476 1477 1478
    for (i = 0 ; i < ctrl->nveths ; i++) {
        if (virNetDevVethDelete(ctrl->veths[i]) < 0)
            ret = -1;
    }

    return ret;
1479 1480
}

1481

1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500
static int lxcSetPersonality(virDomainDefPtr def)
{
    struct utsname utsname;
    const char *altArch;

    uname(&utsname);

    altArch = lxcContainerGetAlt32bitArch(utsname.machine);
    if (altArch &&
        STREQ(def->os.arch, altArch)) {
        if (personality(PER_LINUX32) < 0) {
            virReportSystemError(errno, _("Unable to request personality for %s on %s"),
                                 altArch, utsname.machine);
            return -1;
        }
    }
    return 0;
}

1501
#ifndef MS_REC
1502
# define MS_REC          16384
1503 1504 1505
#endif

#ifndef MS_SLAVE
1506
# define MS_SLAVE              (1<<19)
1507
#endif
1508

1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552
/* Create a private tty using the private devpts at PTMX, returning
 * the master in *TTYMASTER and the name of the slave, _from the
 * perspective of the guest after remounting file systems_, in
 * *TTYNAME.  Heavily borrowed from glibc, but doesn't require that
 * devpts == "/dev/pts" */
static int
lxcCreateTty(char *ptmx, int *ttymaster, char **ttyName)
{
    int ret = -1;
    int ptyno;
    int unlock = 0;

    if ((*ttymaster = open(ptmx, O_RDWR|O_NOCTTY|O_NONBLOCK)) < 0)
        goto cleanup;

    if (ioctl(*ttymaster, TIOCSPTLCK, &unlock) < 0)
        goto cleanup;

    if (ioctl(*ttymaster, TIOCGPTN, &ptyno) < 0)
        goto cleanup;

    /* If mount() succeeded at honoring newinstance, then the kernel
     * was new enough to also honor the mode=0620,gid=5 options, which
     * guarantee that the new pty already has correct permissions; so
     * while glibc has to fstat(), fchmod(), and fchown() for older
     * kernels, we can skip those steps.  ptyno shouldn't currently be
     * anything other than 0, but let's play it safe.  */
    if (virAsprintf(ttyName, "/dev/pts/%d", ptyno) < 0) {
        virReportOOMError();
        errno = ENOMEM;
        goto cleanup;
    }

    ret = 0;

cleanup:
    if (ret != 0) {
        VIR_FORCE_CLOSE(*ttymaster);
        VIR_FREE(*ttyName);
    }

    return ret;
}

1553

1554
static int
1555
virLXCControllerSetupDevPTS(virLXCControllerPtr ctrl)
1556
{
1557
    virDomainFSDefPtr root = virDomainGetRootFilesystem(ctrl->def);
1558
    char *mount_options = NULL;
1559 1560 1561
    char *opts;
    char *devpts = NULL;
    int ret = -1;
1562

1563 1564 1565 1566 1567 1568 1569 1570
    if (!root) {
        if (ctrl->nconsoles != 1) {
            lxcError(VIR_ERR_CONFIG_UNSUPPORTED,
                     _("Expected exactly one console, but got %zu"),
                     ctrl->nconsoles);
            return -1;
        }
        return 0;
1571 1572
    }

1573
    VIR_DEBUG("Setting up private /dev/pts");
1574

1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594
    /*
     * If doing a chroot style setup, we need to prepare
     * a private /dev/pts for the child now, which they
     * will later move into position.
     *
     * This is complex because 'virsh console' needs to
     * use /dev/pts from the host OS, and the guest OS
     * needs to use /dev/pts from the guest.
     *
     * This means that we (libvirt_lxc) need to see and
     * use both /dev/pts instances. We're running in the
     * host OS context though and don't want to expose
     * the guest OS /dev/pts there.
     *
     * Thus we call unshare(CLONE_NS) so that we can see
     * the guest's new /dev/pts, without it becoming
     * visible to the host OS. We also put the root FS
     * into slave mode, just in case it was currently
     * marked as shared
     */
1595 1596
    mount_options = virSecurityManagerGetMountOptions(ctrl->securityManager,
                                                      ctrl->def);
1597

1598 1599 1600 1601 1602 1603
    if (!virFileExists(root->src)) {
        virReportSystemError(errno,
                             _("root source %s does not exist"),
                             root->src);
        goto cleanup;
    }
1604

1605 1606 1607 1608 1609
    if (unshare(CLONE_NEWNS) < 0) {
        virReportSystemError(errno, "%s",
                             _("Cannot unshare mount namespace"));
        goto cleanup;
    }
1610

1611 1612 1613 1614 1615
    if (mount("", "/", NULL, MS_SLAVE|MS_REC, NULL) < 0) {
        virReportSystemError(errno, "%s",
                             _("Failed to switch root mount into slave mode"));
        goto cleanup;
    }
1616

1617 1618 1619 1620 1621
    if (virAsprintf(&devpts, "%s/dev/pts", root->src) < 0 ||
        virAsprintf(&ctrl->devptmx, "%s/dev/pts/ptmx", root->src) < 0) {
        virReportOOMError();
        goto cleanup;
    }
1622

1623 1624 1625 1626 1627 1628
    if (virFileMakePath(devpts) < 0) {
        virReportSystemError(errno,
                             _("Failed to make path %s"),
                             devpts);
        goto cleanup;
    }
1629

1630 1631 1632 1633 1634 1635 1636
    /* XXX should we support gid=X for X!=5 for distros which use
     * a different gid for tty?  */
    if (virAsprintf(&opts, "newinstance,ptmxmode=0666,mode=0620,gid=5%s",
                    (mount_options ? mount_options : "")) < 0) {
        virReportOOMError();
        goto cleanup;
    }
1637

1638 1639 1640 1641 1642 1643 1644 1645
    VIR_DEBUG("Mount devpts on %s type=tmpfs flags=%x, opts=%s",
              devpts, MS_NOSUID, opts);
    if (mount("devpts", devpts, "devpts", MS_NOSUID, opts) < 0) {
        virReportSystemError(errno,
                             _("Failed to mount devpts on %s"),
                             devpts);
        goto cleanup;
    }
1646

1647 1648 1649
    if (access(ctrl->devptmx, R_OK) < 0) {
        VIR_WARN("Kernel does not support private devpts, using shared devpts");
        VIR_FREE(ctrl->devptmx);
1650 1651
    }

1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666
    ret = 0;

cleanup:
    VIR_FREE(opts);
    VIR_FREE(devpts);
    return ret;
}


static int
virLXCControllerSetupConsoles(virLXCControllerPtr ctrl,
                              char **containerTTYPaths)
{
    size_t i;

1667
    for (i = 0 ; i < ctrl->nconsoles ; i++) {
1668 1669 1670
        if (ctrl->devptmx) {
            VIR_DEBUG("Opening tty on private %s", ctrl->devptmx);
            if (lxcCreateTty(ctrl->devptmx,
1671 1672
                             &ctrl->consoles[i].contFd,
                             &containerTTYPaths[i]) < 0) {
1673 1674
                virReportSystemError(errno, "%s",
                                     _("Failed to allocate tty"));
1675
                return -1;
1676 1677 1678
            }
        } else {
            VIR_DEBUG("Opening tty on shared /dev/ptmx");
1679 1680
            if (virFileOpenTty(&ctrl->consoles[i].contFd,
                               &containerTTYPaths[i],
1681 1682 1683
                               0) < 0) {
                virReportSystemError(errno, "%s",
                                     _("Failed to allocate tty"));
1684
                return -1;
1685 1686 1687
            }
        }
    }
1688 1689 1690 1691 1692
    return 0;
}


static int
1693
virLXCControllerRun(virLXCControllerPtr ctrl)
1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728
{
    int rc = -1;
    int control[2] = { -1, -1};
    int containerhandshake[2] = { -1, -1 };
    char **containerTTYPaths = NULL;
    size_t i;

    if (VIR_ALLOC_N(containerTTYPaths, ctrl->nconsoles) < 0) {
        virReportOOMError();
        goto cleanup;
    }

    if (socketpair(PF_UNIX, SOCK_STREAM, 0, control) < 0) {
        virReportSystemError(errno, "%s",
                             _("sockpair failed"));
        goto cleanup;
    }

    if (socketpair(PF_UNIX, SOCK_STREAM, 0, containerhandshake) < 0) {
        virReportSystemError(errno, "%s",
                             _("socketpair failed"));
        goto cleanup;
    }

    if (virLXCControllerSetupLoopDevices(ctrl) < 0)
        goto cleanup;

    if (lxcSetContainerResources(ctrl->def) < 0)
        goto cleanup;

    if (virLXCControllerSetupDevPTS(ctrl) < 0)
        goto cleanup;

    if (virLXCControllerSetupConsoles(ctrl, containerTTYPaths) < 0)
        goto cleanup;
1729

1730
    if (lxcSetPersonality(ctrl->def) < 0)
1731
        goto cleanup;
1732

1733
    if ((ctrl->initpid = lxcContainerStart(ctrl->def,
1734
                                           ctrl->securityManager,
1735 1736 1737 1738
                                           ctrl->nveths,
                                           ctrl->veths,
                                           control[1],
                                           containerhandshake[1],
1739 1740
                                           containerTTYPaths,
                                           ctrl->nconsoles)) < 0)
1741
        goto cleanup;
1742
    VIR_FORCE_CLOSE(control[1]);
1743
    VIR_FORCE_CLOSE(containerhandshake[1]);
1744

1745
    if (virLXCControllerMoveInterfaces(ctrl) < 0)
1746 1747
        goto cleanup;

1748 1749 1750
    if (lxcContainerSendContinue(control[0]) < 0) {
        virReportSystemError(errno, "%s",
                             _("Unable to send container continue message"));
1751
        goto cleanup;
1752
    }
1753

1754 1755 1756 1757 1758 1759
    if (lxcContainerWaitForContinue(containerhandshake[0]) < 0) {
        virReportSystemError(errno, "%s",
                             _("error receiving signal from container"));
        goto cleanup;
    }

1760 1761 1762
    /* Now the container is fully setup... */

    /* ...we can close the loop devices... */
1763 1764
    if (virLXCControllerCloseLoopDevices(ctrl, false) < 0)
        goto cleanup;
1765 1766

    /* ...and reduce our privileges */
D
Daniel P. Berrange 已提交
1767 1768 1769
    if (lxcControllerClearCapabilities() < 0)
        goto cleanup;

1770
    if (virLXCControllerDaemonHandshake(ctrl) < 0)
1771 1772
        goto cleanup;

1773 1774
    if (virSetBlocking(ctrl->monitorServerFd, false) < 0 ||
        virSetBlocking(ctrl->monitorClientFd, false) < 0) {
1775
        virReportSystemError(errno, "%s",
1776
                             _("Unable to set file descriptor non-blocking"));
1777 1778
        goto cleanup;
    }
1779 1780
    for (i = 0 ; i < ctrl->nconsoles ; i++)
        if (virLXCControllerConsoleSetNonblocking(&(ctrl->consoles[i])) < 0)
1781
            goto cleanup;
1782

1783
    rc = virLXCControllerMain(ctrl);
1784 1785

cleanup:
1786 1787
    VIR_FORCE_CLOSE(control[0]);
    VIR_FORCE_CLOSE(control[1]);
1788 1789
    VIR_FORCE_CLOSE(containerhandshake[0]);
    VIR_FORCE_CLOSE(containerhandshake[1]);
1790

1791 1792 1793
    for (i = 0 ; i < ctrl->nconsoles ; i++)
        VIR_FREE(containerTTYPaths[i]);
    VIR_FREE(containerTTYPaths);
1794

1795
    virLXCControllerStopInit(ctrl);
1796

1797 1798 1799 1800
    return rc;
}


1801
int main(int argc, char *argv[])
1802 1803
{
    pid_t pid;
1804 1805
    int rc = 1;
    char *name = NULL;
1806
    size_t nveths = 0;
1807
    char **veths = NULL;
1808
    int handshakeFd = -1;
1809 1810
    int bg = 0;
    char *sockpath = NULL;
1811
    const struct option options[] = {
1812 1813 1814 1815
        { "background", 0, NULL, 'b' },
        { "name",   1, NULL, 'n' },
        { "veth",   1, NULL, 'v' },
        { "console", 1, NULL, 'c' },
1816
        { "handshakefd", 1, NULL, 's' },
1817
        { "security", 1, NULL, 'S' },
1818 1819 1820
        { "help", 0, NULL, 'h' },
        { 0, 0, 0, 0 },
    };
1821 1822
    int *ttyFDs = NULL;
    size_t nttyFDs = 0;
1823
    virLXCControllerPtr ctrl = NULL;
1824
    size_t i;
1825
    const char *securityDriver = "none";
1826

E
Eric Blake 已提交
1827 1828
    if (setlocale(LC_ALL, "") == NULL ||
        bindtextdomain(PACKAGE, LOCALEDIR) == NULL ||
1829 1830
        textdomain(PACKAGE) == NULL ||
        virRandomInitialize(time(NULL) ^ getpid())) {
E
Eric Blake 已提交
1831 1832 1833 1834
        fprintf(stderr, _("%s: initialization failed\n"), argv[0]);
        exit(EXIT_FAILURE);
    }

1835 1836 1837
    /* Initialize logging */
    virLogSetFromEnv();

1838 1839
    while (1) {
        int c;
1840

1841
        c = getopt_long(argc, argv, "dn:v:m:c:s:h:S:",
1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853
                       options, NULL);

        if (c == -1)
            break;

        switch (c) {
        case 'b':
            bg = 1;
            break;

        case 'n':
            if ((name = strdup(optarg)) == NULL) {
1854
                virReportOOMError();
1855
                goto cleanup;
1856
            }
1857 1858 1859 1860
            break;

        case 'v':
            if (VIR_REALLOC_N(veths, nveths+1) < 0) {
1861
                virReportOOMError();
1862
                goto cleanup;
1863
            }
1864
            if ((veths[nveths++] = strdup(optarg)) == NULL) {
1865
                virReportOOMError();
1866
                goto cleanup;
1867
            }
1868 1869 1870
            break;

        case 'c':
1871 1872 1873 1874 1875
            if (VIR_REALLOC_N(ttyFDs, nttyFDs + 1) < 0) {
                virReportOOMError();
                goto cleanup;
            }
            if (virStrToLong_i(optarg, NULL, 10, &ttyFDs[nttyFDs++]) < 0) {
1876 1877 1878 1879 1880
                fprintf(stderr, "malformed --console argument '%s'", optarg);
                goto cleanup;
            }
            break;

1881
        case 's':
1882
            if (virStrToLong_i(optarg, NULL, 10, &handshakeFd) < 0) {
1883 1884 1885 1886 1887 1888
                fprintf(stderr, "malformed --handshakefd argument '%s'",
                        optarg);
                goto cleanup;
            }
            break;

1889
        case 'S':
1890
            securityDriver = optarg;
1891 1892
            break;

1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903
        case 'h':
        case '?':
            fprintf(stderr, "\n");
            fprintf(stderr, "syntax: %s [OPTIONS]\n", argv[0]);
            fprintf(stderr, "\n");
            fprintf(stderr, "Options\n");
            fprintf(stderr, "\n");
            fprintf(stderr, "  -b, --background\n");
            fprintf(stderr, "  -n NAME, --name NAME\n");
            fprintf(stderr, "  -c FD, --console FD\n");
            fprintf(stderr, "  -v VETH, --veth VETH\n");
1904
            fprintf(stderr, "  -s FD, --handshakefd FD\n");
1905
            fprintf(stderr, "  -S NAME, --security NAME\n");
1906 1907 1908
            fprintf(stderr, "  -h, --help\n");
            fprintf(stderr, "\n");
            goto cleanup;
1909 1910 1911
        }
    }

1912 1913 1914 1915 1916
    if (name == NULL) {
        fprintf(stderr, "%s: missing --name argument for configuration\n", argv[0]);
        goto cleanup;
    }

1917
    if (handshakeFd < 0) {
1918 1919 1920 1921 1922
        fprintf(stderr, "%s: missing --handshake argument for container PTY\n",
                argv[0]);
        goto cleanup;
    }

1923
    if (getuid() != 0) {
1924 1925 1926
        fprintf(stderr, "%s: must be run as the 'root' user\n", argv[0]);
        goto cleanup;
    }
1927

1928 1929
    virEventRegisterDefaultImpl();

1930
    if (!(ctrl = virLXCControllerNew(name)))
1931
        goto cleanup;
1932

1933 1934
    ctrl->handshakeFd = handshakeFd;

1935 1936 1937 1938 1939
    if (!(ctrl->securityManager = virSecurityManagerNew(securityDriver,
                                                        LXC_DRIVER_NAME,
                                                        false, false, false)))
        goto cleanup;

1940
    VIR_DEBUG("Security model %s type %s label %s imagelabel %s",
1941 1942 1943 1944
              NULLSTR(ctrl->def->seclabel.model),
              virDomainSeclabelTypeToString(ctrl->def->seclabel.type),
              NULLSTR(ctrl->def->seclabel.label),
              NULLSTR(ctrl->def->seclabel.imagelabel));
1945

1946 1947 1948
    ctrl->veths = veths;
    ctrl->nveths = nveths;

1949 1950 1951 1952 1953 1954
    for (i = 0 ; i < nttyFDs ; i++) {
        if (virLXCControllerAddConsole(ctrl, ttyFDs[i]) < 0)
            goto cleanup;
        ttyFDs[i] = -1;
    }

1955
    if (virLXCControllerValidateNICs(ctrl) < 0)
1956
        goto cleanup;
1957

1958 1959 1960
    if (virLXCControllerValidateConsoles(ctrl) < 0)
        goto cleanup;

1961
    if ((sockpath = lxcMonitorPath(ctrl)) == NULL)
1962
        goto cleanup;
1963

1964
    if ((ctrl->monitorServerFd = lxcMonitorServer(sockpath)) < 0)
1965
        goto cleanup;
1966

1967 1968 1969
    if (bg) {
        if ((pid = fork()) < 0)
            goto cleanup;
1970

1971
        if (pid > 0) {
1972
            if ((rc = virPidFileWrite(LXC_STATE_DIR, name, pid)) < 0) {
1973
                virReportSystemError(-rc,
1974 1975
                                     _("Unable to write pid file '%s/%s.pid'"),
                                     LXC_STATE_DIR, name);
1976 1977
                _exit(1);
            }
1978

1979 1980 1981 1982
            /* First child now exits, allowing original caller
             * (ie libvirtd's LXC driver to complete their
             * waitpid & continue */
            _exit(0);
1983 1984
        }

1985 1986
        /* Don't hold onto any cwd we inherit from libvirtd either */
        if (chdir("/") < 0) {
1987
            virReportSystemError(errno, "%s",
1988
                                 _("Unable to change to root dir"));
1989 1990 1991 1992
            goto cleanup;
        }

        if (setsid() < 0) {
1993
            virReportSystemError(errno, "%s",
1994
                                 _("Unable to become session leader"));
1995 1996 1997
            goto cleanup;
        }
    }
1998 1999

    /* Accept initial client which is the libvirtd daemon */
2000
    if ((ctrl->monitorClientFd = accept(ctrl->monitorServerFd, NULL, 0)) < 0) {
2001
        virReportSystemError(errno, "%s",
2002
                             _("Failed to accept a connection from driver"));
2003
        goto cleanup;
2004 2005
    }

2006
    rc = virLXCControllerRun(ctrl);
2007

2008
cleanup:
2009
    virPidFileDelete(LXC_STATE_DIR, name);
2010
    virLXCControllerDeleteInterfaces(ctrl);
J
Jim Meyering 已提交
2011 2012
    if (sockpath)
        unlink(sockpath);
2013
    VIR_FREE(sockpath);
2014 2015 2016 2017
    for (i = 0 ; i < nttyFDs ; i++)
        VIR_FORCE_CLOSE(ttyFDs[i]);
    VIR_FREE(ttyFDs);

2018
    virLXCControllerFree(ctrl);
2019

2020
    return rc ? EXIT_FAILURE : EXIT_SUCCESS;
2021
}