lxc_container.c 58.3 KB
Newer Older
1
/*
2
 * Copyright (C) 2008-2013 Red Hat, Inc.
3
 * Copyright (C) 2008 IBM Corp.
4 5 6 7 8
 *
 * lxc_container.c: file description
 *
 * Authors:
 *  David L. Leskovec <dlesko at linux.vnet.ibm.com>
9
 *  Daniel P. Berrange <berrange@redhat.com>
10 11 12 13 14 15 16 17 18 19 20 21
 *
 * 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
22
 * License along with this library.  If not, see
O
Osier Yang 已提交
23
 * <http://www.gnu.org/licenses/>.
24 25 26 27 28 29 30
 */

#include <config.h>

#include <fcntl.h>
#include <limits.h>
#include <stdlib.h>
31
#include <stdio.h>
32 33
#include <sys/ioctl.h>
#include <sys/mount.h>
34
#include <sys/wait.h>
35
#include <sys/stat.h>
36
#include <unistd.h>
37
#include <mntent.h>
38 39
#include <sys/reboot.h>
#include <linux/reboot.h>
40 41 42 43 44 45

/* Yes, we want linux private one, for _syscall2() macro */
#include <linux/unistd.h>

/* For MS_MOVE */
#include <linux/fs.h>
46

47
#if WITH_CAPNG
48
# include <cap-ng.h>
D
Daniel P. Berrange 已提交
49
#endif
50

51
#if WITH_BLKID
52 53 54
# include <blkid/blkid.h>
#endif

55 56 57 58
#if WITH_SELINUX
# include <selinux/selinux.h>
#endif

59
#include "virerror.h"
60
#include "virlog.h"
61
#include "lxc_container.h"
62
#include "viralloc.h"
63
#include "virnetdevveth.h"
64
#include "viruuid.h"
E
Eric Blake 已提交
65
#include "virfile.h"
66
#include "virusb.h"
67
#include "vircommand.h"
68
#include "virnetdev.h"
69
#include "virprocess.h"
70
#include "virstring.h"
71

72 73
#define VIR_FROM_THIS VIR_FROM_LXC

74 75 76 77 78 79
/*
 * GLibc headers are behind the kernel, so we define these
 * constants if they're not present already.
 */

#ifndef CLONE_NEWPID
80
# define CLONE_NEWPID  0x20000000
81 82
#endif
#ifndef CLONE_NEWUTS
83
# define CLONE_NEWUTS  0x04000000
84 85
#endif
#ifndef CLONE_NEWUSER
86
# define CLONE_NEWUSER 0x10000000
87 88
#endif
#ifndef CLONE_NEWIPC
89
# define CLONE_NEWIPC  0x08000000
90 91
#endif
#ifndef CLONE_NEWNET
92
# define CLONE_NEWNET  0x40000000 /* New network namespace */
93 94 95 96 97 98 99 100
#endif

/* messages between parent and container */
typedef char lxc_message_t;
#define LXC_CONTINUE_MSG 'c'

typedef struct __lxc_child_argv lxc_child_argv_t;
struct __lxc_child_argv {
101
    virDomainDefPtr config;
102
    virSecurityManagerPtr securityDriver;
103
    size_t nveths;
104
    char **veths;
105
    int monitor;
106 107
    size_t npassFDs;
    int *passFDs;
108
    size_t nttyPaths;
109
    char **ttyPaths;
110
    int handshakefd;
111 112
};

113 114 115
static int lxcContainerMountFSBlock(virDomainFSDefPtr fs,
                                    const char *srcprefix);

116

117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162
/*
 * reboot(LINUX_REBOOT_CMD_CAD_ON) will return -EINVAL
 * in a child pid namespace if container reboot support exists.
 * Otherwise, it will either succeed or return -EPERM.
 */
ATTRIBUTE_NORETURN static int
lxcContainerRebootChild(void *argv)
{
    int *cmd = argv;
    int ret;

    ret = reboot(*cmd);
    if (ret == -1 && errno == EINVAL)
        _exit(1);
    _exit(0);
}


static
int lxcContainerHasReboot(void)
{
    int flags = CLONE_NEWPID|CLONE_NEWNS|CLONE_NEWUTS|
        CLONE_NEWIPC|SIGCHLD;
    int cpid;
    char *childStack;
    char *stack;
    char *buf;
    int cmd, v;
    int status;
    char *tmp;

    if (virFileReadAll("/proc/sys/kernel/ctrl-alt-del", 10, &buf) < 0)
        return -1;

    if ((tmp = strchr(buf, '\n')))
        *tmp = '\0';

    if (virStrToLong_i(buf, NULL, 10, &v) < 0) {
        virReportError(VIR_ERR_INTERNAL_ERROR,
                       _("Malformed ctrl-alt-del setting '%s'"), buf);
        VIR_FREE(buf);
        return -1;
    }
    VIR_FREE(buf);
    cmd = v ? LINUX_REBOOT_CMD_CAD_ON : LINUX_REBOOT_CMD_CAD_OFF;

163
    if (VIR_ALLOC_N(stack, getpagesize() * 4) < 0)
164 165 166 167 168 169 170 171 172 173
        return -1;

    childStack = stack + (getpagesize() * 4);

    cpid = clone(lxcContainerRebootChild, childStack, flags, &cmd);
    VIR_FREE(stack);
    if (cpid < 0) {
        virReportSystemError(errno, "%s",
                             _("Unable to clone to check reboot support"));
        return -1;
174
    } else if (virProcessWait(cpid, &status) < 0) {
175 176 177 178 179 180 181 182 183 184 185 186 187 188
        return -1;
    }

    if (WEXITSTATUS(status) != 1) {
        VIR_DEBUG("Containerized reboot support is missing "
                  "(kernel probably too old < 3.4)");
        return 0;
    }

    VIR_DEBUG("Containerized reboot support is available");
    return 1;
}


189
/**
190
 * lxcContainerBuildInitCmd:
191
 * @vmDef: pointer to vm definition structure
192
 *
193
 * Build a virCommandPtr for launching the container 'init' process
194
 *
195
 * Returns a virCommandPtr
196
 */
197
static virCommandPtr lxcContainerBuildInitCmd(virDomainDefPtr vmDef)
198
{
199
    char uuidstr[VIR_UUID_STRING_BUFLEN];
200
    virCommandPtr cmd;
201 202 203

    virUUIDFormat(vmDef->uuid, uuidstr);

204 205
    cmd = virCommandNew(vmDef->os.init);

206 207 208
    if (vmDef->os.initargv && vmDef->os.initargv[0])
        virCommandAddArgSet(cmd, (const char **)vmDef->os.initargv);

209 210
    virCommandAddEnvString(cmd, "PATH=/bin:/sbin");
    virCommandAddEnvString(cmd, "TERM=linux");
211
    virCommandAddEnvString(cmd, "container=lxc-libvirt");
212
    virCommandAddEnvPair(cmd, "container_uuid", uuidstr);
213 214
    virCommandAddEnvPair(cmd, "LIBVIRT_LXC_UUID", uuidstr);
    virCommandAddEnvPair(cmd, "LIBVIRT_LXC_NAME", vmDef->name);
215 216
    if (vmDef->os.cmdline)
        virCommandAddEnvPair(cmd, "LIBVIRT_LXC_CMDLINE", vmDef->os.cmdline);
217

218
    return cmd;
219 220 221
}

/**
222
 * lxcContainerSetupFDs:
223 224
 * @control: control FD from parent
 * @ttyfd: FD of tty to set as the container console
225 226
 * @npassFDs: number of extra FDs
 * @passFDs: list of extra FDs
227
 *
228 229 230 231
 * Setup file descriptors in the container. @ttyfd is set to be
 * the container's stdin, stdout & stderr. Any FDs included in
 * @passFDs, will be dup()'d such that they start from stderr+1
 * with no gaps.
232 233 234
 *
 * Returns 0 on success or -1 in case of error
 */
235 236
static int lxcContainerSetupFDs(int *ttyfd,
                                size_t npassFDs, int *passFDs)
237 238
{
    int rc = -1;
239 240
    int open_max;
    int fd;
241 242 243
    int last_fd;
    size_t i;
    size_t j;
244 245

    if (setsid() < 0) {
246
        virReportSystemError(errno, "%s",
247
                             _("setsid failed"));
248
        goto cleanup;
249 250
    }

251
    if (ioctl(*ttyfd, TIOCSCTTY, NULL) < 0) {
252
        virReportSystemError(errno, "%s",
253
                             _("ioctl(TIOCSTTY) failed"));
254 255 256
        goto cleanup;
    }

257
    if (dup2(*ttyfd, STDIN_FILENO) < 0) {
J
John Ferlan 已提交
258
        virReportSystemError(errno, "%s",
259
                             _("dup2(stdin) failed"));
J
John Ferlan 已提交
260 261
        goto cleanup;
    }
262

263
    if (dup2(*ttyfd, STDOUT_FILENO) < 0) {
264
        virReportSystemError(errno, "%s",
265
                             _("dup2(stdout) failed"));
266 267 268
        goto cleanup;
    }

269
    if (dup2(*ttyfd, STDERR_FILENO) < 0) {
270
        virReportSystemError(errno, "%s",
271
                             _("dup2(stderr) failed"));
272 273 274
        goto cleanup;
    }

275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333
    VIR_FORCE_CLOSE(*ttyfd);

    /* Any FDs in @passFDs need to be moved around so that
     * they are numbered, without gaps, starting from
     * STDERR_FILENO + 1
     */
    for (i = 0; i < npassFDs; i++) {
        int wantfd;

        wantfd = STDERR_FILENO + i + 1;
        VIR_DEBUG("Pass %d onto %d", passFDs[i], wantfd);

        /* If we already have desired FD number, life
         * is easy. Nothing needs renumbering */
        if (passFDs[i] == wantfd)
            continue;

        /*
         * Lets check to see if any later FDs are occupying
         * our desired FD number. If so, we must move them
         * out of the way
         */
        for (j = i + 1; j < npassFDs; j++) {
            if (passFDs[j] == wantfd) {
                VIR_DEBUG("Clash %zu", j);
                int newfd = dup(passFDs[j]);
                if (newfd < 0) {
                    virReportSystemError(errno,
                                         _("Cannot move fd %d out of the way"),
                                         passFDs[j]);
                    goto cleanup;
                }
                /* We're intentionally not closing the
                 * old value of passFDs[j], because we
                 * don't want later iterations of the
                 * loop to take it back. dup2() will
                 * cause it to be closed shortly anyway
                 */
                VIR_DEBUG("Moved clash onto %d", newfd);
                passFDs[j] = newfd;
            }
        }

        /* Finally we can move into our desired FD number */
        if (dup2(passFDs[i], wantfd) < 0) {
            virReportSystemError(errno,
                                 _("Cannot duplicate fd %d onto fd %d"),
                                 passFDs[i], wantfd);
            goto cleanup;
        }
        VIR_FORCE_CLOSE(passFDs[i]);
    }

    last_fd = STDERR_FILENO + npassFDs;

    /* Just in case someone forget to set FD_CLOEXEC, explicitly
     * close all remaining FDs before executing the container */
    open_max = sysconf(_SC_OPEN_MAX);
    if (open_max < 0) {
334
        virReportSystemError(errno, "%s",
335
                             _("sysconf(_SC_OPEN_MAX) failed"));
336 337 338
        goto cleanup;
    }

339 340 341 342 343
    for (fd = last_fd + 1; fd < open_max; fd++) {
        int tmpfd = fd;
        VIR_MASS_CLOSE(tmpfd);
    }

344 345 346
    rc = 0;

cleanup:
347
    VIR_DEBUG("rc=%d", rc);
348 349 350 351
    return rc;
}

/**
352
 * lxcContainerSendContinue:
353
 * @control: control FD to child
354
 *
355 356
 * Sends the continue message via the socket pair stored in the vm
 * structure.
357 358 359
 *
 * Returns 0 on success or -1 in case of error
 */
360
int lxcContainerSendContinue(int control)
361 362
{
    int rc = -1;
363 364
    lxc_message_t msg = LXC_CONTINUE_MSG;
    int writeCount = 0;
365

366
    VIR_DEBUG("Send continue on fd %d", control);
367 368 369
    writeCount = safewrite(control, &msg, sizeof(msg));
    if (writeCount != sizeof(msg)) {
        goto error_out;
370 371
    }

372 373 374
    rc = 0;
error_out:
    return rc;
375 376
}

377
/**
378
 * lxcContainerWaitForContinue:
379
 * @control: Control FD from parent
380 381 382 383 384 385 386
 *
 * This function will wait for the container continue message from the
 * parent process.  It will send this message on the socket pair stored in
 * the vm structure once it has completed the post clone container setup.
 *
 * Returns 0 on success or -1 in case of error
 */
387
int lxcContainerWaitForContinue(int control)
388 389 390 391
{
    lxc_message_t msg;
    int readLen;

392
    VIR_DEBUG("Wait continue on fd %d", control);
393
    readLen = saferead(control, &msg, sizeof(msg));
394
    VIR_DEBUG("Got continue on fd %d %d", control, readLen);
395 396 397 398 399 400 401
    if (readLen != sizeof(msg)) {
        if (readLen >= 0)
            errno = EIO;
        return -1;
    }
    if (msg != LXC_CONTINUE_MSG) {
        errno = EINVAL;
402
        return -1;
403 404
    }

405
    return 0;
406 407
}

408

409 410 411 412 413 414 415 416 417 418 419 420 421 422
/**
 * lxcContainerSetID:
 *
 * This function calls setuid and setgid to create proper
 * cred for tasks running in container.
 *
 * Returns 0 on success or -1 in case of error
 */
static int lxcContainerSetID(virDomainDefPtr def)
{
    /* Only call virSetUIDGID when user namespace is enabled
     * for this container. And user namespace is only enabled
     * when nuidmap&ngidmap is not zero */

423 424 425 426 427
    if (!def->idmap.nuidmap)
        return 0;

    VIR_DEBUG("Setting UID/GID to 0/0");
    if (virSetUIDGID(0, 0, NULL, 0) < 0) {
428 429 430 431 432 433 434 435 436
        virReportSystemError(errno, "%s",
                             _("setuid or setgid failed"));
        return -1;
    }

    return 0;
}


437
/**
438
 * lxcContainerRenameAndEnableInterfaces:
439 440
 * @nveths: number of interfaces
 * @veths: interface names
441
 *
442 443 444
 * This function will rename the interfaces to ethN
 * with id ascending order from zero and enable the
 * renamed interfaces for this container.
445 446 447
 *
 * Returns 0 on success or nonzero in case of error
 */
448
static int lxcContainerRenameAndEnableInterfaces(bool privNet,
449
                                                 size_t nveths,
450
                                                 char **veths)
451 452
{
    int rc = 0;
453
    size_t i;
454
    char *newname = NULL;
455

456
    for (i = 0; i < nveths; i++) {
457
        if (virAsprintf(&newname, "eth%zu", i) < 0) {
458
            rc = -1;
459
            goto error_out;
460
        }
461

462
        VIR_DEBUG("Renaming %s to %s", veths[i], newname);
463
        rc = virNetDevSetName(veths[i], newname);
464
        if (rc < 0)
465 466
            goto error_out;

467
        VIR_DEBUG("Enabling %s", newname);
468
        rc = virNetDevSetOnline(newname, true);
469
        if (rc < 0)
470
            goto error_out;
471

472
        VIR_FREE(newname);
473 474 475
    }

    /* enable lo device only if there were other net devices */
476
    if (veths || privNet)
477
        rc = virNetDevSetOnline("lo", true);
478 479

error_out:
480
    VIR_FREE(newname);
481 482 483
    return rc;
}

484

485
/*_syscall2(int, pivot_root, char *, newroot, const char *, oldroot)*/
486 487 488 489 490 491 492
extern int pivot_root(const char * new_root,const char * put_old);

static int lxcContainerChildMountSort(const void *a, const void *b)
{
  const char **sa = (const char**)a;
  const char **sb = (const char**)b;

E
Eric Blake 已提交
493
  /* Deliberately reversed args - we need to unmount deepest
494 495 496 497
     children first */
  return strcmp(*sb, *sa);
}

498
#ifndef MS_REC
499
# define MS_REC          16384
500 501 502
#endif

#ifndef MNT_DETACH
503
# define MNT_DETACH      0x00000002
504 505 506
#endif

#ifndef MS_PRIVATE
507
# define MS_PRIVATE              (1<<18)
508 509 510
#endif

#ifndef MS_SLAVE
511
# define MS_SLAVE                (1<<19)
512 513
#endif

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
static int lxcContainerGetSubtree(const char *prefix,
                                  char ***mountsret,
                                  size_t *nmountsret)
{
    FILE *procmnt;
    struct mntent mntent;
    char mntbuf[1024];
    int ret = -1;
    char **mounts = NULL;
    size_t nmounts = 0;

    VIR_DEBUG("prefix=%s", prefix);

    *mountsret = NULL;
    *nmountsret = 0;

    if (!(procmnt = setmntent("/proc/mounts", "r"))) {
        virReportSystemError(errno, "%s",
                             _("Failed to read /proc/mounts"));
        return -1;
    }

    while (getmntent_r(procmnt, &mntent, mntbuf, sizeof(mntbuf)) != NULL) {
        if (!STRPREFIX(mntent.mnt_dir, prefix))
            continue;

540
        if (VIR_REALLOC_N(mounts, nmounts+1) < 0)
541
            goto cleanup;
542
        if (VIR_STRDUP(mounts[nmounts], mntent.mnt_dir) < 0)
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
            goto cleanup;
        nmounts++;
    }

    if (mounts)
        qsort(mounts, nmounts, sizeof(mounts[0]),
              lxcContainerChildMountSort);

    ret = 0;
cleanup:
    *mountsret = mounts;
    *nmountsret = nmounts;
    endmntent(procmnt);
    return ret;
}

static int lxcContainerUnmountSubtree(const char *prefix,
                                      bool isOldRootFS)
{
    char **mounts = NULL;
    size_t nmounts = 0;
    size_t i;
    int saveErrno;
    const char *failedUmount = NULL;
    int ret = -1;

    VIR_DEBUG("Unmount subtreee from %s", prefix);

    if (lxcContainerGetSubtree(prefix, &mounts, &nmounts) < 0)
        goto cleanup;
573
    for (i = 0; i < nmounts; i++) {
574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606
        VIR_DEBUG("Umount %s", mounts[i]);
        if (umount(mounts[i]) < 0) {
            char ebuf[1024];
            failedUmount = mounts[i];
            saveErrno = errno;
            VIR_WARN("Failed to unmount '%s', trying to detach subtree '%s': %s",
                     failedUmount, mounts[nmounts-1],
                     virStrerror(errno, ebuf, sizeof(ebuf)));
            break;
        }
    }

    if (failedUmount) {
        /* This detaches the subtree */
        if (umount2(mounts[nmounts-1], MNT_DETACH) < 0) {
            virReportSystemError(saveErrno,
                                 _("Failed to unmount '%s' and could not detach subtree '%s'"),
                                 failedUmount, mounts[nmounts-1]);
            goto cleanup;
        }
        /* This unmounts the tmpfs on which the old root filesystem was hosted */
        if (isOldRootFS &&
            umount(mounts[nmounts-1]) < 0) {
            virReportSystemError(saveErrno,
                                 _("Failed to unmount '%s' and could not unmount old root '%s'"),
                                 failedUmount, mounts[nmounts-1]);
            goto cleanup;
        }
    }

    ret = 0;

cleanup:
607
    for (i = 0; i < nmounts; i++)
608 609 610 611 612 613 614
        VIR_FREE(mounts[i]);
    VIR_FREE(mounts);

    return ret;
}


615 616 617 618 619 620
static int lxcContainerPrepareRoot(virDomainDefPtr def,
                                   virDomainFSDefPtr root)
{
    char *dst;
    char *tmp;

621 622
    VIR_DEBUG("Prepare root %d", root->type);

623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639
    if (root->type == VIR_DOMAIN_FS_TYPE_MOUNT)
        return 0;

    if (root->type == VIR_DOMAIN_FS_TYPE_FILE) {
        virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
                       _("Unexpected root filesystem without loop device"));
        return -1;
    }

    if (root->type != VIR_DOMAIN_FS_TYPE_BLOCK) {
        virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
                       _("Unsupported root filesystem type %s"),
                       virDomainFSTypeToString(root->type));
        return -1;
    }

    if (virAsprintf(&dst, "%s/%s.root",
640
                    LXC_STATE_DIR, def->name) < 0)
641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659
        return -1;

    tmp = root->dst;
    root->dst = dst;

    if (lxcContainerMountFSBlock(root, "") < 0) {
        root->dst = tmp;
        VIR_FREE(dst);
        return -1;
    }

    root->dst = tmp;
    root->type = VIR_DOMAIN_FS_TYPE_MOUNT;
    VIR_FREE(root->src);
    root->src = dst;

    return 0;
}

660 661
static int lxcContainerPivotRoot(virDomainFSDefPtr root)
{
662
    int ret;
663
    char *oldroot = NULL, *newroot = NULL;
664

M
Mark McLoughlin 已提交
665 666
    ret = -1;

667 668
    VIR_DEBUG("Pivot via %s", root->src);

669 670
    /* root->parent must be private, so make / private. */
    if (mount("", "/", NULL, MS_PRIVATE|MS_REC, NULL) < 0) {
671
        virReportSystemError(errno, "%s",
672
                             _("Failed to make root private"));
673
        goto err;
674 675
    }

676
    if (virAsprintf(&oldroot, "%s/.oldroot", root->src) < 0)
677
        goto err;
678

679 680
    if (virFileMakePath(oldroot) < 0) {
        virReportSystemError(errno,
681
                             _("Failed to create %s"),
682
                             oldroot);
683 684 685 686 687
        goto err;
    }

    /* Create a tmpfs root since old and new roots must be
     * on separate filesystems */
688
    if (mount("tmprootfs", oldroot, "tmpfs", 0, NULL) < 0) {
689
        virReportSystemError(errno,
690
                             _("Failed to mount empty tmpfs at %s"),
691 692 693
                             oldroot);
        goto err;
    }
M
Mark McLoughlin 已提交
694

695
    /* Create a directory called 'new' in tmpfs */
696
    if (virAsprintf(&newroot, "%s/new", oldroot) < 0)
697 698
        goto err;

699 700
    if (virFileMakePath(newroot) < 0) {
        virReportSystemError(errno,
701
                             _("Failed to create %s"),
702 703 704 705 706 707
                             newroot);
        goto err;
    }

    /* ... and mount our root onto it */
    if (mount(root->src, newroot, NULL, MS_BIND|MS_REC, NULL) < 0) {
708
        virReportSystemError(errno,
709
                             _("Failed to bind new root %s into tmpfs"),
710 711 712 713
                             root->src);
        goto err;
    }

714 715 716 717 718 719 720 721 722
    if (root->readonly) {
        if (mount(root->src, newroot, NULL, MS_BIND|MS_REC|MS_RDONLY|MS_REMOUNT, NULL) < 0) {
            virReportSystemError(errno,
                                 _("Failed to make new root %s readonly"),
                                 root->src);
            goto err;
        }
    }

723
    /* Now we chdir into the tmpfs, then pivot into the
724
     * root->src bind-mounted onto '/new' */
725
    if (chdir(newroot) < 0) {
726
        virReportSystemError(errno,
727
                             _("Failed to chdir into %s"), newroot);
728
        goto err;
729 730 731 732
    }

    /* The old root directory will live at /.oldroot after
     * this and will soon be unmounted completely */
733
    if (pivot_root(".", ".oldroot") < 0) {
734
        virReportSystemError(errno, "%s",
735
                             _("Failed to pivot root"));
736
        goto err;
737 738 739
    }

    /* CWD is undefined after pivot_root, so go to / */
740 741 742
    if (chdir("/") < 0)
        goto err;

M
Mark McLoughlin 已提交
743 744 745
    ret = 0;

err:
746 747 748
    VIR_FREE(oldroot);
    VIR_FREE(newroot);

M
Mark McLoughlin 已提交
749
    return ret;
750 751
}

752

753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774
typedef struct {
    const char *src;
    const char *dst;
    const char *type;
    const char *opts;
    int mflags;
} virLXCBasicMountInfo;

static const virLXCBasicMountInfo lxcBasicMounts[] = {
    /* When we want to make a bind mount readonly, for unknown reasons,
     * it is currently necessary to bind it once, and then remount the
     * bind with the readonly flag. If this is not done, then the original
     * mount point in the main OS becomes readonly too which is not what
     * we want. Hence some things have two entries here.
     */
    { "proc", "/proc", "proc", NULL, MS_NOSUID|MS_NOEXEC|MS_NODEV },
    { "/proc/sys", "/proc/sys", NULL, NULL, MS_BIND },
    { "/proc/sys", "/proc/sys", NULL, NULL, MS_BIND|MS_REMOUNT|MS_RDONLY },
    { "sysfs", "/sys", "sysfs", NULL, MS_NOSUID|MS_NOEXEC|MS_NODEV },
    { "sysfs", "/sys", "sysfs", NULL, MS_BIND|MS_REMOUNT|MS_RDONLY },
    { "securityfs", "/sys/kernel/security", "securityfs", NULL, MS_NOSUID|MS_NOEXEC|MS_NODEV },
    { "securityfs", "/sys/kernel/security", "securityfs", NULL, MS_BIND|MS_REMOUNT|MS_RDONLY },
775
#if WITH_SELINUX
776 777
    { SELINUX_MOUNT, SELINUX_MOUNT, "selinuxfs", NULL, MS_NOSUID|MS_NOEXEC|MS_NODEV },
    { SELINUX_MOUNT, SELINUX_MOUNT, NULL, NULL, MS_BIND|MS_REMOUNT|MS_RDONLY },
778
#endif
779 780 781
};


782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849
static bool lxcIsBasicMountLocation(const char *path)
{
    size_t i;

    for (i = 0; i < ARRAY_CARDINALITY(lxcBasicMounts); i++) {
        if (STREQ(path, lxcBasicMounts[i].dst))
            return true;
    }

    return false;
}


static int lxcContainerSetReadOnly(void)
{
    FILE *procmnt;
    struct mntent mntent;
    char mntbuf[1024];
    int ret = -1;
    char **mounts = NULL;
    size_t nmounts = 0;
    size_t i;

    if (!(procmnt = setmntent("/proc/mounts", "r"))) {
        virReportSystemError(errno, "%s",
                             _("Failed to read /proc/mounts"));
        return -1;
    }

    while (getmntent_r(procmnt, &mntent, mntbuf, sizeof(mntbuf)) != NULL) {
        if (STREQ(mntent.mnt_dir, "/") ||
            STREQ(mntent.mnt_dir, "/.oldroot") ||
            STRPREFIX(mntent.mnt_dir, "/.oldroot/") ||
            lxcIsBasicMountLocation(mntent.mnt_dir))
            continue;

        if (VIR_REALLOC_N(mounts, nmounts + 1) < 0)
            goto cleanup;
        if (VIR_STRDUP(mounts[nmounts], mntent.mnt_dir) < 0)
            goto cleanup;
        nmounts++;
    }

    if (mounts)
        qsort(mounts, nmounts, sizeof(mounts[0]),
              lxcContainerChildMountSort);

    for (i = 0; i < nmounts; i++) {
        VIR_DEBUG("Bind readonly %s", mounts[i]);
        if (mount(mounts[i], mounts[i], NULL, MS_BIND|MS_REC|MS_RDONLY|MS_REMOUNT, NULL) < 0) {
            virReportSystemError(errno,
                                 _("Failed to make mount %s readonly"),
                                 mounts[i]);
            goto cleanup;
        }
    }

    ret = 0;
cleanup:
    for (i = 0; i < nmounts; i++)
        VIR_FREE(mounts[i]);
    VIR_FREE(mounts);
    endmntent(procmnt);
    return ret;

}


850 851
static int lxcContainerMountBasicFS(bool userns_enabled)
{
852 853
    size_t i;
    int rc = -1;
854

855
    VIR_DEBUG("Mounting basic filesystems");
856

857 858
    for (i = 0; i < ARRAY_CARDINALITY(lxcBasicMounts); i++) {
        virLXCBasicMountInfo const *mnt = &lxcBasicMounts[i];
859
        const char *srcpath = NULL;
860

861
        VIR_DEBUG("Processing %s -> %s",
862
                  mnt->src, mnt->dst);
863

864
        srcpath = mnt->src;
865 866 867

        /* Skip if mount doesn't exist in source */
        if ((srcpath[0] == '/') &&
868
            (access(srcpath, R_OK) < 0))
869 870
            continue;

871
#if WITH_SELINUX
872
        if (STREQ(mnt->src, SELINUX_MOUNT) &&
873
            (!is_selinux_enabled() || userns_enabled))
874 875 876
            continue;
#endif

877
        if (STREQ(mnt->src, "securityfs") && userns_enabled)
878 879
            continue;

880
        if (virFileMakePath(mnt->dst) < 0) {
881 882
            virReportSystemError(errno,
                                 _("Failed to mkdir %s"),
883
                                 mnt->src);
884 885 886
            goto cleanup;
        }

887
        VIR_DEBUG("Mount %s on %s type=%s flags=%x, opts=%s",
888 889
                  srcpath, mnt->dst, mnt->type, mnt->mflags, mnt->opts);
        if (mount(srcpath, mnt->dst, mnt->type, mnt->mflags, mnt->opts) < 0) {
890
            virReportSystemError(errno,
891
                                 _("Failed to mount %s on %s type %s flags=%x opts=%s"),
892 893
                                 srcpath, mnt->dst, NULLSTR(mnt->type),
                                 mnt->mflags, NULLSTR(mnt->opts));
894
            goto cleanup;
895
        }
896 897 898 899 900 901 902 903 904
    }

    rc = 0;

cleanup:
    VIR_DEBUG("rc=%d", rc);
    return rc;
}

905
#if WITH_FUSE
906 907
static int lxcContainerMountProcFuse(virDomainDefPtr def,
                                     const char *stateDir)
908 909 910 911
{
    int ret;
    char *meminfo_path = NULL;

912 913
    VIR_DEBUG("Mount /proc/meminfo stateDir=%s", stateDir);

914
    if ((ret = virAsprintf(&meminfo_path,
915
                           "/.oldroot/%s/%s.fuse/meminfo",
916
                           stateDir,
917 918 919 920 921 922 923 924 925 926 927 928 929 930
                           def->name)) < 0)
        return ret;

    if ((ret = mount(meminfo_path, "/proc/meminfo",
                     NULL, MS_BIND, NULL)) < 0) {
        virReportSystemError(errno,
                             _("Failed to mount %s on /proc/meminfo"),
                             meminfo_path);
    }

    VIR_FREE(meminfo_path);
    return ret;
}
#else
931 932
static int lxcContainerMountProcFuse(virDomainDefPtr def ATTRIBUTE_UNUSED,
                                     const char *stateDir ATTRIBUTE_UNUSED)
933 934 935 936
{
    return 0;
}
#endif
937

938 939 940
static int lxcContainerMountFSDev(virDomainDefPtr def,
                                  const char *stateDir)
{
941
    int ret = -1;
942 943 944 945 946 947 948 949
    char *path = NULL;

    VIR_DEBUG("Mount /dev/ stateDir=%s", stateDir);

    if ((ret = virAsprintf(&path, "/.oldroot/%s/%s.dev",
                           stateDir, def->name)) < 0)
        return ret;

950 951 952 953 954 955 956
    if (virFileMakePath("/dev") < 0) {
        virReportSystemError(errno, "%s",
                             _("Cannot create /dev"));
        goto cleanup;
    }

    VIR_DEBUG("Trying to move %s to /dev", path);
957

958
    if (mount(path, "/dev", NULL, MS_MOVE, NULL) < 0) {
959 960 961
        virReportSystemError(errno,
                             _("Failed to mount %s on /dev"),
                             path);
962
        goto cleanup;
963 964
    }

965 966 967
    ret = 0;

cleanup:
968 969 970 971
    VIR_FREE(path);
    return ret;
}

972 973
static int lxcContainerMountFSDevPTS(virDomainDefPtr def,
                                     const char *stateDir)
974
{
975 976
    int ret;
    char *path = NULL;
977

978 979
    VIR_DEBUG("Mount /dev/pts stateDir=%s", stateDir);

980
    if ((ret = virAsprintf(&path,
981
                           "/.oldroot/%s/%s.devpts",
982
                           stateDir,
983 984
                           def->name)) < 0)
        return ret;
985

986 987
    if (virFileMakePath("/dev/pts") < 0) {
        virReportSystemError(errno, "%s",
988
                             _("Cannot create /dev/pts"));
989
        goto cleanup;
990
    }
991

992 993 994 995 996 997 998
    VIR_DEBUG("Trying to move %s to /dev/pts", path);

    if ((ret = mount(path, "/dev/pts",
                     NULL, MS_MOVE, NULL)) < 0) {
        virReportSystemError(errno,
                             _("Failed to mount %s on /dev/pts"),
                             path);
999
        goto cleanup;
1000
    }
1001

1002 1003
cleanup:
    VIR_FREE(path);
1004

1005
    return ret;
1006 1007
}

1008
static int lxcContainerSetupDevices(char **ttyPaths, size_t nttyPaths)
1009
{
1010
    size_t i;
1011 1012 1013 1014 1015 1016 1017 1018 1019
    const struct {
        const char *src;
        const char *dst;
    } links[] = {
        { "/proc/self/fd/0", "/dev/stdin" },
        { "/proc/self/fd/1", "/dev/stdout" },
        { "/proc/self/fd/2", "/dev/stderr" },
        { "/proc/self/fd", "/dev/fd" },
    };
1020

1021
    for (i = 0; i < ARRAY_CARDINALITY(links); i++) {
1022 1023 1024 1025 1026 1027 1028 1029
        if (symlink(links[i].src, links[i].dst) < 0) {
            virReportSystemError(errno,
                                 _("Failed to symlink device %s to %s"),
                                 links[i].dst, links[i].src);
            return -1;
        }
    }

1030 1031 1032
    /* We have private devpts capability, so bind that */
    if (virFileTouch("/dev/ptmx", 0666) < 0)
        return -1;
1033

1034 1035 1036 1037
    if (mount("/dev/pts/ptmx", "/dev/ptmx", "ptmx", MS_BIND, NULL) < 0) {
        virReportSystemError(errno, "%s",
                             _("Failed to bind /dev/pts/ptmx on to /dev/ptmx"));
        return -1;
1038 1039
    }

1040
    for (i = 0; i < nttyPaths; i++) {
1041
        char *tty;
1042
        if (virAsprintf(&tty, "/dev/tty%zu", i+1) < 0)
1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058
            return -1;
        if (symlink(ttyPaths[i], tty) < 0) {
            VIR_FREE(tty);
            virReportSystemError(errno,
                                 _("Failed to symlink %s to %s"),
                                 ttyPaths[i], tty);
            return -1;
        }
        VIR_FREE(tty);
        if (i == 0 &&
            symlink(ttyPaths[i], "/dev/console") < 0) {
            virReportSystemError(errno,
                                 _("Failed to symlink %s to /dev/console"),
                                 ttyPaths[i]);
            return -1;
        }
1059
    }
1060 1061 1062 1063
    return 0;
}


1064 1065
static int lxcContainerMountFSBind(virDomainFSDefPtr fs,
                                   const char *srcprefix)
1066
{
1067 1068
    char *src = NULL;
    int ret = -1;
1069
    struct stat st;
1070

1071 1072
    VIR_DEBUG("src=%s dst=%s", fs->src, fs->dst);

1073
    if (virAsprintf(&src, "%s%s", srcprefix, fs->src) < 0)
1074 1075
        goto cleanup;

1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 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
    if (stat(fs->dst, &st) < 0) {
        if (errno != ENOENT) {
            virReportSystemError(errno, _("Unable to stat bind target %s"),
                                 fs->dst);
            goto cleanup;
        }
        /* ENOENT => create the target dir or file */
        if (stat(src, &st) < 0) {
            virReportSystemError(errno, _("Unable to stat bind source %s"),
                                 src);
            goto cleanup;
        }
        if (S_ISDIR(st.st_mode)) {
            if (virFileMakePath(fs->dst) < 0) {
                virReportSystemError(errno,
                                     _("Failed to create %s"),
                                     fs->dst);
                goto cleanup;
            }
        } else {
            /* Create Empty file for target mount point */
            int fd = open(fs->dst, O_WRONLY|O_CREAT|O_NOCTTY|O_NONBLOCK, 0666);
            if (fd < 0) {
                if (errno != EEXIST) {
                    virReportSystemError(errno,
                                         _("Failed to create bind target %s"),
                                         fs->dst);
                    goto cleanup;
                }
            }
            if (VIR_CLOSE(fd) < 0) {
                virReportSystemError(errno,
                                     _("Failed to close bind target %s"),
                                     fs->dst);
                goto cleanup;
            }
        }
1113 1114 1115 1116 1117 1118 1119 1120 1121
    }

    if (mount(src, fs->dst, NULL, MS_BIND, NULL) < 0) {
        virReportSystemError(errno,
                             _("Failed to bind mount directory %s to %s"),
                             src, fs->dst);
        goto cleanup;
    }

1122 1123
    if (fs->readonly) {
        VIR_DEBUG("Binding %s readonly", fs->dst);
1124
        if (mount(src, fs->dst, NULL, MS_BIND|MS_REMOUNT|MS_RDONLY, NULL) < 0) {
1125 1126 1127
            virReportSystemError(errno,
                                 _("Failed to make directory %s readonly"),
                                 fs->dst);
1128
        }
1129 1130 1131 1132 1133 1134 1135
    } else {
        VIR_DEBUG("Binding %s readwrite", fs->dst);
        if (mount(src, fs->dst, NULL, MS_BIND|MS_REMOUNT, NULL) < 0) {
            virReportSystemError(errno,
                                 _("Failed to make directory %s readwrite"),
                                 fs->dst);
        }
1136 1137 1138 1139 1140 1141 1142 1143 1144 1145
    }

    ret = 0;

cleanup:
    VIR_FREE(src);
    return ret;
}


1146
#ifdef WITH_BLKID
1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203
static int
lxcContainerMountDetectFilesystem(const char *src, char **type)
{
    int fd;
    int ret = -1;
    int rc;
    const char *data = NULL;
    blkid_probe blkid = NULL;

    *type = NULL;

    if ((fd = open(src, O_RDONLY)) < 0) {
        virReportSystemError(errno,
                             _("Unable to open filesystem %s"), src);
        return -1;
    }

    if (!(blkid = blkid_new_probe())) {
        virReportSystemError(errno, "%s",
                             _("Unable to create blkid library handle"));
        goto cleanup;
    }
    if (blkid_probe_set_device(blkid, fd, 0, 0) < 0) {
        virReportSystemError(EINVAL,
                             _("Unable to associate device %s with blkid library"),
                             src);
        goto cleanup;
    }

    blkid_probe_enable_superblocks(blkid, 1);

    blkid_probe_set_superblocks_flags(blkid, BLKID_SUBLKS_TYPE);

    rc = blkid_do_safeprobe(blkid);
    if (rc != 0) {
        if (rc == 1) /* Nothing found, return success with *type == NULL */
            goto done;

        if (rc == -2) {
            virReportSystemError(EINVAL,
                                 _("Too many filesystems detected for %s"),
                                 src);
        } else {
            virReportSystemError(errno,
                                 _("Unable to detect filesystem for %s"),
                                 src);
        }
        goto cleanup;
    }

    if (blkid_probe_lookup_value(blkid, "TYPE", &data, NULL) < 0) {
        virReportSystemError(ENOENT,
                             _("Unable to find filesystem type for %s"),
                             src);
        goto cleanup;
    }

1204
    if (VIR_STRDUP(*type, data) < 0)
1205 1206 1207 1208 1209 1210 1211 1212 1213 1214
        goto cleanup;

done:
    ret = 0;
cleanup:
    VIR_FORCE_CLOSE(fd);
    if (blkid)
        blkid_free_probe(blkid);
    return ret;
}
1215
#else /* ! WITH_BLKID */
1216 1217 1218 1219 1220 1221 1222 1223
static int
lxcContainerMountDetectFilesystem(const char *src ATTRIBUTE_UNUSED,
                                  char **type)
{
    /* No libblkid, so just return success with no detected type */
    *type = NULL;
    return 0;
}
1224
#endif /* ! WITH_BLKID */
1225 1226

/*
C
Chen Hanxiao 已提交
1227
 * This function attempts to do automatic detection of filesystem
1228 1229 1230 1231 1232 1233 1234 1235
 * type following the same rules as the util-linux 'mount' binary.
 *
 * The main difference is that we don't (currently) try to use
 * libblkid to detect the format first. We go straight to using
 * /etc/filesystems, and then /proc/filesystems
 */
static int lxcContainerMountFSBlockAuto(virDomainFSDefPtr fs,
                                        int fsflags,
1236 1237
                                        const char *src,
                                        const char *srcprefix)
1238 1239 1240 1241 1242 1243 1244 1245 1246
{
    FILE *fp = NULL;
    int ret = -1;
    bool tryProc = false;
    bool gotStar = false;
    char *fslist = NULL;
    char *line = NULL;
    const char *type;

1247
    VIR_DEBUG("src=%s dst=%s srcprefix=%s", src, fs->dst, srcprefix);
1248 1249 1250

    /* First time around we use /etc/filesystems */
retry:
1251
    if (virAsprintf(&fslist, "%s%s", srcprefix,
1252
                    tryProc ? "/proc/filesystems" : "/etc/filesystems") < 0)
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 1279
        goto cleanup;

    VIR_DEBUG("Open fslist %s", fslist);
    if (!(fp = fopen(fslist, "r"))) {
        /* If /etc/filesystems does not exist, then we need to retry
         * with /proc/filesystems next
         */
        if (errno == ENOENT &&
            !tryProc) {
            tryProc = true;
            VIR_FREE(fslist);
            goto retry;
        }

        virReportSystemError(errno,
                             _("Unable to read %s"),
                             fslist);
        goto cleanup;
    }

    while (!feof(fp)) {
        size_t n;
        VIR_FREE(line);
        if (getline(&line, &n, fp) <= 0) {
            if (feof(fp))
                break;

1280 1281 1282
            goto cleanup;
        }

1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295
        if (strstr(line, "nodev"))
            continue;

        type = strchr(line, '\n');
        if (type)
            line[type-line] = '\0';

        type = line;
        virSkipSpaces(&type);

        /*
         * /etc/filesystems is only allowed to contain '*' on the last line
         */
1296
        if (gotStar && !tryProc) {
1297 1298 1299
            virReportError(VIR_ERR_INTERNAL_ERROR,
                           _("%s has unexpected '*' before last line"),
                           fslist);
1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320
            goto cleanup;
        }

        /* An '*' on the last line in /etc/filesystems
         * means try /proc/filesystems next. We don't
         * jump immediately though, since we need to see
         * if any more lines follow
         */
        if (!tryProc &&
            STREQ(type, "*"))
            gotStar = true;

        VIR_DEBUG("Trying mount %s with %s", src, type);
        if (mount(src, fs->dst, type, fsflags, NULL) < 0) {
            /* These errnos indicate a bogus filesystem type for
             * the image we have, so skip to the next type
             */
            if (errno == EINVAL || errno == ENODEV)
                continue;

            virReportSystemError(errno,
1321
                                 _("Failed to mount device %s to %s"),
1322 1323 1324 1325 1326 1327
                                 src, fs->dst);
            goto cleanup;
        }

        ret = 0;
        break;
1328 1329
    }

1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341
    /* We've got to the end of /etc/filesystems and saw
     * a '*', so we must try /proc/filesystems next
     */
    if (ret != 0 &&
        !tryProc &&
        gotStar) {
        tryProc = true;
        VIR_FREE(fslist);
        VIR_FORCE_FCLOSE(fp);
        goto retry;
    }

1342 1343 1344 1345 1346 1347
    if (ret != 0) {
        virReportSystemError(ENODEV,
                             _("Failed to mount device %s to %s, unable to detect filesystem"),
                             src, fs->dst);
    }

1348 1349 1350 1351
    VIR_DEBUG("Done mounting filesystem ret=%d tryProc=%d", ret, tryProc);

cleanup:
    VIR_FREE(line);
1352
    VIR_FREE(fslist);
1353 1354 1355 1356 1357 1358 1359 1360 1361 1362
    VIR_FORCE_FCLOSE(fp);
    return ret;
}


/*
 * Mount a block device 'src' on fs->dst, automatically
 * probing for filesystem type
 */
static int lxcContainerMountFSBlockHelper(virDomainFSDefPtr fs,
1363 1364
                                          const char *src,
                                          const char *srcprefix)
1365 1366 1367
{
    int fsflags = 0;
    int ret = -1;
1368 1369
    char *format = NULL;

1370 1371 1372 1373 1374 1375 1376 1377 1378 1379
    if (fs->readonly)
        fsflags |= MS_RDONLY;

    if (virFileMakePath(fs->dst) < 0) {
        virReportSystemError(errno,
                             _("Failed to create %s"),
                             fs->dst);
        goto cleanup;
    }

1380 1381 1382 1383
    if (lxcContainerMountDetectFilesystem(src, &format) < 0)
        goto cleanup;

    if (format) {
1384 1385
        VIR_DEBUG("Mount '%s' on '%s' with detected format '%s'",
                  src, fs->dst, format);
1386 1387 1388 1389 1390 1391 1392 1393
        if (mount(src, fs->dst, format, fsflags, NULL) < 0) {
            virReportSystemError(errno,
                                 _("Failed to mount device %s to %s as %s"),
                                 src, fs->dst, format);
            goto cleanup;
        }
        ret = 0;
    } else {
1394
        ret = lxcContainerMountFSBlockAuto(fs, fsflags, src, srcprefix);
1395
    }
1396 1397

cleanup:
1398
    VIR_FREE(format);
1399 1400 1401 1402 1403 1404 1405 1406 1407 1408
    return ret;
}


static int lxcContainerMountFSBlock(virDomainFSDefPtr fs,
                                    const char *srcprefix)
{
    char *src = NULL;
    int ret = -1;

1409 1410
    VIR_DEBUG("src=%s dst=%s", fs->src, fs->dst);

1411
    if (virAsprintf(&src, "%s%s", srcprefix, fs->src) < 0)
1412 1413
        goto cleanup;

1414
    ret = lxcContainerMountFSBlockHelper(fs, src, srcprefix);
1415 1416 1417 1418 1419 1420 1421 1422 1423

    VIR_DEBUG("Done mounting filesystem ret=%d", ret);

cleanup:
    VIR_FREE(src);
    return ret;
}


1424 1425
static int lxcContainerMountFSTmpfs(virDomainFSDefPtr fs,
                                    char *sec_mount_options)
1426 1427 1428 1429
{
    int ret = -1;
    char *data = NULL;

1430 1431
    VIR_DEBUG("usage=%lld sec=%s", fs->usage, sec_mount_options);

1432
    if (virAsprintf(&data,
1433
                    "size=%lld%s", fs->usage, sec_mount_options) < 0)
1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455
        goto cleanup;

    if (virFileMakePath(fs->dst) < 0) {
        virReportSystemError(errno,
                             _("Failed to create %s"),
                             fs->dst);
        goto cleanup;
    }

    if (mount("tmpfs", fs->dst, "tmpfs", MS_NOSUID|MS_NODEV, data) < 0) {
        virReportSystemError(errno,
                             _("Failed to mount directory %s as tmpfs"),
                             fs->dst);
        goto cleanup;
    }

    if (fs->readonly) {
        VIR_DEBUG("Binding %s readonly", fs->dst);
        if (mount(fs->dst, fs->dst, NULL, MS_BIND|MS_REMOUNT|MS_RDONLY, NULL) < 0) {
            virReportSystemError(errno,
                                 _("Failed to make directory %s readonly"),
                                 fs->dst);
1456
            goto cleanup;
1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467
        }
    }

    ret = 0;

cleanup:
    VIR_FREE(data);
    return ret;
}


1468
static int lxcContainerMountFS(virDomainFSDefPtr fs,
1469
                               char *sec_mount_options)
1470 1471 1472
{
    switch (fs->type) {
    case VIR_DOMAIN_FS_TYPE_MOUNT:
1473
        if (lxcContainerMountFSBind(fs, "/.oldroot") < 0)
1474 1475
            return -1;
        break;
1476
    case VIR_DOMAIN_FS_TYPE_BLOCK:
1477
        if (lxcContainerMountFSBlock(fs, "/.oldroot") < 0)
1478 1479
            return -1;
        break;
1480
    case VIR_DOMAIN_FS_TYPE_RAM:
1481
        if (lxcContainerMountFSTmpfs(fs, sec_mount_options) < 0)
1482 1483
            return -1;
        break;
1484 1485 1486 1487
    case VIR_DOMAIN_FS_TYPE_BIND:
        if (lxcContainerMountFSBind(fs, "") < 0)
            return -1;
        break;
1488
    case VIR_DOMAIN_FS_TYPE_FILE:
1489 1490 1491
        /* We do actually support this, but the lxc controller
         * should have associated the file with a loopback
         * device and changed this to TYPE_BLOCK for us */
1492 1493 1494
        virReportError(VIR_ERR_INTERNAL_ERROR,
                       _("Unexpected filesystem type %s"),
                       virDomainFSTypeToString(fs->type));
1495
        return -1;
1496
    default:
1497 1498 1499
        virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
                       _("Cannot mount filesystem type %s"),
                       virDomainFSTypeToString(fs->type));
1500
        return -1;
1501 1502 1503 1504 1505 1506
    }
    return 0;
}


static int lxcContainerMountAllFS(virDomainDefPtr vmDef,
1507
                                  char *sec_mount_options)
1508 1509
{
    size_t i;
1510
    VIR_DEBUG("Mounting all non-root filesystems");
1511 1512

    /* Pull in rest of container's mounts */
1513
    for (i = 0; i < vmDef->nfss; i++) {
1514
        if (STREQ(vmDef->fss[i]->dst, "/"))
1515 1516
            continue;

1517 1518 1519 1520
        if (lxcContainerUnmountSubtree(vmDef->fss[i]->dst,
                                       false) < 0)
            return -1;

1521
        if (lxcContainerMountFS(vmDef->fss[i], sec_mount_options) < 0)
1522 1523 1524
            return -1;
    }

1525
    VIR_DEBUG("Mounted all non-root filesystems");
1526 1527 1528 1529
    return 0;
}


1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555
int lxcContainerSetupHostdevCapsMakePath(const char *dev)
{
    int ret = -1;
    char *dir, *tmp;

    if (VIR_STRDUP(dir, dev) < 0)
        return -1;

    if ((tmp = strrchr(dir, '/'))) {
        *tmp = '\0';
        if (virFileMakePath(dir) < 0) {
            virReportSystemError(errno,
                                 _("Failed to create directory for '%s' dev '%s'"),
                                 dir, dev);
            goto cleanup;
        }
    }

    ret = 0;

cleanup:
    VIR_FREE(dir);
    return ret;
}


1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609
static int lxcContainerUnmountForSharedRoot(const char *stateDir,
                                            const char *domain)
{
    int ret = -1;
    char *tmp = NULL;

#if WITH_SELINUX
    /* Some versions of Linux kernel don't let you overmount
     * the selinux filesystem, so make sure we kill it first
     */
    /* Filed coverity bug for false positive 'USE_AFTER_FREE' due to swap
     * of root->src with root->dst and the VIR_FREE(root->src) prior to the
     * reset of root->src in lxcContainerPrepareRoot()
     */
    /* coverity[deref_arg] */
    if (lxcContainerUnmountSubtree(SELINUX_MOUNT, false) < 0)
        goto cleanup;
#endif

    /* These filesystems are created by libvirt temporarily, they
     * shouldn't appear in container. */
    if (virAsprintf(&tmp, "%s/%s.dev", stateDir, domain) < 0 ||
        lxcContainerUnmountSubtree(tmp, false) < 0)
        goto cleanup;

    VIR_FREE(tmp);
    if (virAsprintf(&tmp, "%s/%s.devpts", stateDir, domain) < 0 ||
        lxcContainerUnmountSubtree(tmp, false) < 0)
        goto cleanup;

#if WITH_FUSE
    VIR_FREE(tmp);
    if (virAsprintf(&tmp, "%s/%s.fuse", stateDir, domain) < 0 ||
        lxcContainerUnmountSubtree(tmp, false) < 0)
        goto cleanup;
#endif

    /* If we have the root source being '/', then we need to
     * get rid of any existing stuff under /proc, /sys & /tmp.
     * We need new namespace aware versions of those. We must
     * do /proc last otherwise we won't find /proc/mounts :-) */
    if (lxcContainerUnmountSubtree("/sys", false) < 0 ||
        lxcContainerUnmountSubtree("/dev", false) < 0 ||
        lxcContainerUnmountSubtree("/proc", false) < 0)
        goto cleanup;

    ret = 0;

cleanup:
    VIR_FREE(tmp);
    return ret;
}


1610 1611 1612 1613 1614
/* Got a FS mapped to /, we're going the pivot_root
 * approach to do a better-chroot-than-chroot
 * this is based on this thread http://lkml.org/lkml/2008/3/5/29
 */
static int lxcContainerSetupPivotRoot(virDomainDefPtr vmDef,
1615 1616
                                      virDomainFSDefPtr root,
                                      char **ttyPaths,
1617
                                      size_t nttyPaths,
1618
                                      virSecurityManagerPtr securityDriver)
1619
{
1620
    virCgroupPtr cgroup = NULL;
1621
    int ret = -1;
1622
    char *sec_mount_options;
1623
    char *stateDir = NULL;
1624

1625 1626
    VIR_DEBUG("Setup pivot root");

1627 1628
    if (!(sec_mount_options = virSecurityManagerGetMountOptions(securityDriver, vmDef)))
        return -1;
1629 1630 1631

    /* Before pivoting we need to identify any
     * cgroups controllers that are mounted */
1632
    if (virCgroupNewSelf(&cgroup) < 0)
1633
        goto cleanup;
1634

1635 1636 1637
    if (virFileResolveAllLinks(LXC_STATE_DIR, &stateDir) < 0)
        goto cleanup;

1638 1639 1640 1641
    /* Ensure the root filesystem is mounted */
    if (lxcContainerPrepareRoot(vmDef, root) < 0)
        goto cleanup;

1642
    /* Gives us a private root, leaving all parent OS mounts on /.oldroot */
1643
    if (lxcContainerPivotRoot(root) < 0)
1644
        goto cleanup;
1645

1646
    if (STREQ(root->src, "/") &&
1647
        lxcContainerUnmountForSharedRoot(stateDir, vmDef->name) < 0)
1648 1649
        goto cleanup;

1650
    /* Mounts the core /proc, /sys, etc filesystems */
1651
    if (lxcContainerMountBasicFS(vmDef->idmap.nuidmap) < 0)
1652 1653
        goto cleanup;

1654 1655 1656 1657 1658
    /* Ensure entire root filesystem (except /.oldroot) is readonly */
    if (root->readonly &&
        lxcContainerSetReadOnly() < 0)
        goto cleanup;

1659
    /* Mounts /proc/meminfo etc sysinfo */
1660
    if (lxcContainerMountProcFuse(vmDef, stateDir) < 0)
1661 1662
        goto cleanup;

1663 1664
    /* Now we can re-mount the cgroups controllers in the
     * same configuration as before */
1665
    if (virCgroupIsolateMount(cgroup, "/.oldroot/", sec_mount_options) < 0)
1666
        goto cleanup;
1667

1668 1669 1670 1671
    /* Mounts /dev */
    if (lxcContainerMountFSDev(vmDef, stateDir) < 0)
        goto cleanup;

1672
    /* Mounts /dev/pts */
1673
    if (lxcContainerMountFSDevPTS(vmDef, stateDir) < 0)
1674
        goto cleanup;
1675

1676 1677
    /* Setup device nodes in /dev/ */
    if (lxcContainerSetupDevices(ttyPaths, nttyPaths) < 0)
1678
        goto cleanup;
1679

1680
    /* Sets up any non-root mounts from guest config */
1681
    if (lxcContainerMountAllFS(vmDef, sec_mount_options) < 0)
1682
        goto cleanup;
1683

1684
   /* Gets rid of all remaining mounts from host OS, including /.oldroot itself */
1685
    if (lxcContainerUnmountSubtree("/.oldroot", true) < 0)
1686
        goto cleanup;
1687

1688 1689 1690
    ret = 0;

cleanup:
1691
    VIR_FREE(stateDir);
1692
    virCgroupFree(&cgroup);
1693
    VIR_FREE(sec_mount_options);
1694
    return ret;
1695 1696
}

1697

1698 1699 1700 1701 1702
static int lxcContainerResolveSymlinks(virDomainDefPtr vmDef)
{
    char *newroot;
    size_t i;

1703 1704
    VIR_DEBUG("Resolving symlinks");

1705
    for (i = 0; i < vmDef->nfss; i++) {
1706
        virDomainFSDefPtr fs = vmDef->fss[i];
1707 1708
        if (!fs->src)
            continue;
1709 1710

        if (access(fs->src, F_OK)) {
1711 1712
            virReportSystemError(errno,
                                 _("Failed to access '%s'"), fs->src);
1713 1714 1715
            return -1;
        }

1716
        VIR_DEBUG("Resolving '%s'", fs->src);
1717
        if (virFileResolveAllLinks(fs->src, &newroot) < 0) {
1718 1719 1720
            virReportSystemError(errno,
                                 _("Failed to resolve symlink at %s"),
                                 fs->src);
1721
            return -1;
1722
        }
1723 1724 1725 1726 1727 1728

        VIR_DEBUG("Resolved '%s' to %s", fs->src, newroot);

        VIR_FREE(fs->src);
        fs->src = newroot;
    }
1729
    VIR_DEBUG("Resolved all filesystem symlinks");
1730 1731 1732 1733

    return 0;
}

D
Daniel P. Berrange 已提交
1734
/*
C
Chen Hanxiao 已提交
1735
 * This is running as the 'init' process inside the container.
D
Daniel P. Berrange 已提交
1736 1737 1738
 * It removes some capabilities that could be dangerous to
 * host system, since they are not currently "containerized"
 */
1739
#if WITH_CAPNG
1740 1741
static int lxcContainerDropCapabilities(bool keepReboot)
{
D
Daniel P. Berrange 已提交
1742 1743 1744 1745 1746 1747 1748 1749 1750
    int ret;

    capng_get_caps_process();

    if ((ret = capng_updatev(CAPNG_DROP,
                             CAPNG_EFFECTIVE | CAPNG_PERMITTED |
                             CAPNG_INHERITABLE | CAPNG_BOUNDING_SET,
                             CAP_SYS_MODULE, /* No kernel module loading */
                             CAP_SYS_TIME, /* No changing the clock */
1751
                             CAP_MKNOD, /* No creating device nodes */
D
Daniel P. Berrange 已提交
1752 1753
                             CAP_AUDIT_CONTROL, /* No messing with auditing status */
                             CAP_MAC_ADMIN, /* No messing with LSM config */
1754
                             keepReboot ? -1 : CAP_SYS_BOOT, /* No use of reboot */
1755
                             -1)) < 0) {
1756 1757
        virReportError(VIR_ERR_INTERNAL_ERROR,
                       _("Failed to remove capabilities: %d"), ret);
D
Daniel P. Berrange 已提交
1758 1759
        return -1;
    }
1760

D
Daniel P. Berrange 已提交
1761
    if ((ret = capng_apply(CAPNG_SELECT_BOTH)) < 0) {
1762 1763
        virReportError(VIR_ERR_INTERNAL_ERROR,
                       _("Failed to apply capabilities: %d"), ret);
D
Daniel P. Berrange 已提交
1764
        return -1;
1765
    }
D
Daniel P. Berrange 已提交
1766

1767 1768 1769 1770 1771
    /* We do not need to call capng_lock() in this case. The bounding
     * set restriction will prevent them reacquiring sys_boot/module/time,
     * etc which is all that matters for the container. Once inside the
     * container it is fine for SECURE_NOROOT / SECURE_NO_SETUID_FIXUP to
     * be unmasked  - they can never escape the bounding set. */
D
Daniel P. Berrange 已提交
1772

1773 1774
    return 0;
}
D
Daniel P. Berrange 已提交
1775
#else
1776 1777
static int lxcContainerDropCapabilities(bool keepReboot ATTRIBUTE_UNUSED)
{
1778
    VIR_WARN("libcap-ng support not compiled in, unable to clear capabilities");
1779 1780
    return 0;
}
1781
#endif
1782 1783


1784
/**
1785 1786
 * lxcContainerChild:
 * @data: pointer to container arguments
1787 1788 1789 1790 1791 1792 1793 1794 1795
 *
 * This function is run in the process clone()'d in lxcStartContainer.
 * Perform a number of container setup tasks:
 *     Setup container file system
 *     mount container /proca
 * Then exec's the container init
 *
 * Returns 0 on success or -1 in case of error
 */
1796
static int lxcContainerChild(void *data)
1797
{
1798
    lxc_child_argv_t *argv = data;
1799
    virDomainDefPtr vmDef = argv->config;
1800
    int ttyfd = -1;
1801
    int ret = -1;
1802
    char *ttyPath = NULL;
1803
    virDomainFSDefPtr root;
1804
    virCommandPtr cmd = NULL;
1805
    int hasReboot;
1806 1807

    if (NULL == vmDef) {
1808 1809
        virReportError(VIR_ERR_INTERNAL_ERROR,
                       "%s", _("lxcChild() passed invalid vm definition"));
1810
        goto cleanup;
1811 1812
    }

1813 1814 1815 1816 1817 1818 1819 1820 1821 1822
    /* Wait for controller to finish setup tasks, including
     * things like move of network interfaces, uid/gid mapping
     */
    if (lxcContainerWaitForContinue(argv->monitor) < 0) {
        virReportSystemError(errno, "%s",
                             _("Failed to read the container continue message"));
        goto cleanup;
    }
    VIR_DEBUG("Received container continue message");

1823 1824 1825
    if ((hasReboot = lxcContainerHasReboot()) < 0)
        goto cleanup;

1826 1827 1828
    cmd = lxcContainerBuildInitCmd(vmDef);
    virCommandWriteArgLog(cmd, 1);

1829 1830 1831
    if (lxcContainerSetID(vmDef) < 0)
        goto cleanup;

1832
    root = virDomainGetRootFilesystem(vmDef);
1833

1834
    if (argv->nttyPaths) {
1835 1836 1837 1838
        const char *tty = argv->ttyPaths[0];
        if (STRPREFIX(tty, "/dev/pts/"))
            tty += strlen("/dev/pts/");
        if (virAsprintf(&ttyPath, "%s/%s.devpts/%s",
1839
                        LXC_STATE_DIR, vmDef->name, tty) < 0)
1840
            goto cleanup;
1841
    } else if (VIR_STRDUP(ttyPath, "/dev/null") < 0) {
1842
            goto cleanup;
1843
    }
1844

1845
    VIR_DEBUG("Container TTY path: %s", ttyPath);
1846 1847

    ttyfd = open(ttyPath, O_RDWR|O_NOCTTY);
1848
    if (ttyfd < 0) {
1849
        virReportSystemError(errno,
1850
                             _("Failed to open tty %s"),
1851
                             ttyPath);
1852
        goto cleanup;
1853
    }
1854

1855 1856 1857
    if (lxcContainerResolveSymlinks(vmDef) < 0)
        goto cleanup;

1858
    VIR_DEBUG("Setting up pivot");
1859 1860 1861
    if (lxcContainerSetupPivotRoot(vmDef, root,
                                   argv->ttyPaths, argv->nttyPaths,
                                   argv->securityDriver) < 0)
1862
        goto cleanup;
1863

1864 1865 1866 1867 1868 1869 1870
    if (!virFileExists(vmDef->os.init)) {
        virReportSystemError(errno,
                    _("cannot find init path '%s' relative to container root"),
                    vmDef->os.init);
        goto cleanup;
    }

1871
    /* rename and enable interfaces */
1872 1873 1874
    if (lxcContainerRenameAndEnableInterfaces(!!(vmDef->features &
                                                 (1 << VIR_DOMAIN_FEATURE_PRIVNET)),
                                              argv->nveths,
1875
                                              argv->veths) < 0) {
1876
        goto cleanup;
1877
    }
1878

1879
    /* drop a set of root capabilities */
1880
    if (lxcContainerDropCapabilities(!!hasReboot) < 0)
1881
        goto cleanup;
1882

1883 1884
    if (lxcContainerSendContinue(argv->handshakefd) < 0) {
        virReportSystemError(errno, "%s",
C
Chen Hanxiao 已提交
1885
                            _("Failed to send continue signal to controller"));
1886 1887 1888
        goto cleanup;
    }

1889 1890 1891 1892
    VIR_DEBUG("Setting up security labeling");
    if (virSecurityManagerSetProcessLabel(argv->securityDriver, vmDef) < 0)
        goto cleanup;

1893 1894 1895 1896
    VIR_FORCE_CLOSE(argv->handshakefd);
    VIR_FORCE_CLOSE(argv->monitor);
    if (lxcContainerSetupFDs(&ttyfd,
                             argv->npassFDs, argv->passFDs) < 0)
1897 1898
        goto cleanup;

1899
    ret = 0;
1900
cleanup:
1901 1902
    VIR_FREE(ttyPath);
    VIR_FORCE_CLOSE(ttyfd);
1903
    VIR_FORCE_CLOSE(argv->monitor);
1904
    VIR_FORCE_CLOSE(argv->handshakefd);
1905 1906

    if (ret == 0) {
E
Eric Blake 已提交
1907
        /* this function will only return if an error occurred */
1908 1909
        ret = virCommandExec(cmd);
    }
1910 1911 1912 1913 1914 1915 1916 1917 1918

    if (ret != 0) {
        virErrorPtr err = virGetLastError();
        if (err && err->message)
            fprintf(stderr, "%s\n", err->message);
        else
            fprintf(stderr, "%s\n",
                    _("Unknown failure in libvirt_lxc startup"));
    }
1919

1920 1921
    virCommandFree(cmd);
    return ret;
1922
}
1923

1924 1925 1926
static int userns_supported(void)
{
    return lxcContainerAvailable(LXC_CONTAINER_FEATURE_USER) == 0;
1927 1928 1929 1930 1931
}

static int userns_required(virDomainDefPtr def)
{
    return def->idmap.uidmap && def->idmap.gidmap;
1932 1933
}

1934
virArch lxcContainerGetAlt32bitArch(virArch arch)
1935 1936 1937
{
    /* Any Linux 64bit arch which has a 32bit
     * personality available should be listed here */
1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953
    if (arch == VIR_ARCH_X86_64)
        return VIR_ARCH_I686;
    if (arch == VIR_ARCH_S390X)
        return VIR_ARCH_S390;
    if (arch == VIR_ARCH_PPC64)
        return VIR_ARCH_PPC;
    if (arch == VIR_ARCH_PARISC64)
        return VIR_ARCH_PARISC;
    if (arch == VIR_ARCH_SPARC64)
        return VIR_ARCH_SPARC;
    if (arch == VIR_ARCH_MIPS64)
        return VIR_ARCH_MIPS;
    if (arch == VIR_ARCH_MIPS64EL)
        return VIR_ARCH_MIPSEL;

    return VIR_ARCH_NONE;
1954 1955 1956
}


1957 1958 1959 1960 1961 1962 1963 1964
static bool
lxcNeedNetworkNamespace(virDomainDefPtr def)
{
    size_t i;
    if (def->nets != NULL)
        return true;
    if (def->features & (1 << VIR_DOMAIN_FEATURE_PRIVNET))
        return true;
1965
    for (i = 0; i < def->nhostdevs; i++) {
1966 1967 1968 1969 1970 1971 1972
        if (def->hostdevs[i]->mode == VIR_DOMAIN_HOSTDEV_MODE_CAPABILITIES &&
            def->hostdevs[i]->source.caps.type == VIR_DOMAIN_HOSTDEV_CAPS_TYPE_NET)
            return true;
    }
    return false;
}

1973 1974
/**
 * lxcContainerStart:
1975 1976 1977 1978 1979
 * @def: pointer to virtual machine structure
 * @nveths: number of interfaces
 * @veths: interface names
 * @control: control FD to the container
 * @ttyPath: path of tty to set as the container console
1980 1981 1982 1983 1984
 *
 * Starts a container process by calling clone() with the namespace flags
 *
 * Returns PID of container on success or -1 in case of error
 */
1985
int lxcContainerStart(virDomainDefPtr def,
1986
                      virSecurityManagerPtr securityDriver,
1987
                      size_t nveths,
1988
                      char **veths,
1989 1990
                      size_t npassFDs,
                      int *passFDs,
1991
                      int control,
1992
                      int handshakefd,
1993 1994
                      size_t nttyPaths,
                      char **ttyPaths)
1995 1996
{
    pid_t pid;
E
Eric Blake 已提交
1997
    int cflags;
1998 1999
    int stacksize = getpagesize() * 4;
    char *stack, *stacktop;
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011
    lxc_child_argv_t args = {
        .config = def,
        .securityDriver = securityDriver,
        .nveths = nveths,
        .veths = veths,
        .npassFDs = npassFDs,
        .passFDs = passFDs,
        .monitor = control,
        .nttyPaths = nttyPaths,
        .ttyPaths = ttyPaths,
        .handshakefd = handshakefd
    };
2012 2013

    /* allocate a stack for the container */
2014
    if (VIR_ALLOC_N(stack, stacksize) < 0)
2015 2016 2017
        return -1;
    stacktop = stack + stacksize;

E
Eric Blake 已提交
2018
    cflags = CLONE_NEWPID|CLONE_NEWNS|CLONE_NEWUTS|CLONE_NEWIPC|SIGCHLD;
2019

2020 2021 2022 2023 2024 2025 2026
    if (userns_required(def)) {
        if (userns_supported()) {
            VIR_DEBUG("Enable user namespace");
            cflags |= CLONE_NEWUSER;
        } else {
            virReportSystemError(VIR_ERR_CONFIG_UNSUPPORTED, "%s",
                                 _("Kernel doesn't support user namespace"));
2027
            VIR_FREE(stack);
2028 2029
            return -1;
        }
2030
    }
2031

2032
    if (lxcNeedNetworkNamespace(def)) {
2033
        VIR_DEBUG("Enable network namespaces");
E
Eric Blake 已提交
2034
        cflags |= CLONE_NEWNET;
2035
    }
2036

E
Eric Blake 已提交
2037
    pid = clone(lxcContainerChild, stacktop, cflags, &args);
2038
    VIR_FREE(stack);
2039
    VIR_DEBUG("clone() completed, new container PID is %d", pid);
2040 2041

    if (pid < 0) {
2042
        virReportSystemError(errno, "%s",
2043
                             _("Failed to run clone container"));
2044 2045 2046 2047 2048 2049
        return -1;
    }

    return pid;
}

2050 2051
ATTRIBUTE_NORETURN static int
lxcContainerDummyChild(void *argv ATTRIBUTE_UNUSED)
2052 2053 2054 2055 2056 2057
{
    _exit(0);
}

int lxcContainerAvailable(int features)
{
2058
    int flags = CLONE_NEWPID|CLONE_NEWNS|CLONE_NEWUTS|
2059 2060 2061 2062 2063
        CLONE_NEWIPC|SIGCHLD;
    int cpid;
    char *childStack;
    char *stack;

2064 2065 2066
    if (features & LXC_CONTAINER_FEATURE_USER)
        flags |= CLONE_NEWUSER;

2067 2068 2069 2070
    if (features & LXC_CONTAINER_FEATURE_NET)
        flags |= CLONE_NEWNET;

    if (VIR_ALLOC_N(stack, getpagesize() * 4) < 0) {
2071
        VIR_DEBUG("Unable to allocate stack");
2072 2073 2074 2075 2076 2077 2078 2079
        return -1;
    }

    childStack = stack + (getpagesize() * 4);

    cpid = clone(lxcContainerDummyChild, childStack, flags, NULL);
    VIR_FREE(stack);
    if (cpid < 0) {
2080
        char ebuf[1024] ATTRIBUTE_UNUSED;
2081
        VIR_DEBUG("clone call returned %s, container support is not enabled",
2082
                  virStrerror(errno, ebuf, sizeof(ebuf)));
2083
        return -1;
2084
    } else if (virProcessWait(cpid, NULL) < 0) {
E
Eric Blake 已提交
2085
        return -1;
2086 2087
    }

G
Gao feng 已提交
2088
    VIR_DEBUG("container support is enabled");
2089
    return 0;
2090
}
2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111

int lxcContainerChown(virDomainDefPtr def, const char *path)
{
    uid_t uid;
    gid_t gid;

    if (!def->idmap.uidmap)
        return 0;

    uid = def->idmap.uidmap[0].target;
    gid = def->idmap.gidmap[0].target;

    if (chown(path, uid, gid) < 0) {
        virReportSystemError(errno,
                             _("Failed to change owner of %s to %u:%u"),
                             path, uid, gid);
        return -1;
    }

    return 0;
}