src/lxc/namespace.c

   1 /*
   2  * lxc: linux Container library
   3  *
   4  * (C) Copyright IBM Corp. 2007, 2009
   5  *
   6  * Authors:
   7  * Daniel Lezcano <daniel.lezcano at free.fr>
   8  *
   9  * This library is free software; you can redistribute it and/or
  10  * modify it under the terms of the GNU Lesser General Public
  11  * License as published by the Free Software Foundation; either
  12  * version 2.1 of the License, or (at your option) any later version.
  13  *
  14  * This library is distributed in the hope that it will be useful,
  15  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  16  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
  17  * Lesser General Public License for more details.
  18  *
  19  * You should have received a copy of the GNU Lesser General Public
  20  * License along with this library; if not, write to the Free Software
  21  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
  22  */
  23
  24 #include <alloca.h>
  25 #include <errno.h>
  26 #include <fcntl.h>
  27 #include <signal.h>
  28 #include <unistd.h>
  29 #include <sys/param.h>
  30 #include <sys/stat.h>
  31 #include <sys/types.h>
  32
  33 #include "log.h"
  34 #include "namespace.h"
  35 #include "utils.h"
  36
  37 lxc_log_define(lxc_namespace, lxc);
  38
  39 struct clone_arg {
  40         int (*fn)(void *);
  41         void *arg;
  42 };
  43
  44 static int do_clone(void *arg)
  45 {
  46         struct clone_arg *clone_arg = arg;
  47         return clone_arg->fn(clone_arg->arg);
  48 }
  49
  50 pid_t lxc_clone(int (*fn)(void *), void *arg, int flags)
  51 {
  52         struct clone_arg clone_arg = {
  53                 .fn = fn,
  54                 .arg = arg,
  55         };
  56
  57         size_t stack_size = lxc_getpagesize();
  58         void *stack = alloca(stack_size);
  59         pid_t ret;
  60
  61 #ifdef __ia64__
  62         ret = __clone2(do_clone, stack,
  63                        stack_size, flags | SIGCHLD, &clone_arg);
  64 #else
  65         ret = clone(do_clone, stack  + stack_size, flags | SIGCHLD, &clone_arg);
  66 #endif
  67         if (ret < 0)
  68                 ERROR("Failed to clone (%#x): %s.", flags, strerror(errno));
  69
  70         return ret;
  71 }
  72
  73 /* Leave the user namespace at the first position in the array of structs so
  74  * that we always attach to it first when iterating over the struct and using
  75  * setns() to switch namespaces. This especially affects lxc_attach(): Suppose
  76  * you cloned a new user namespace and mount namespace as an unprivileged user
  77  * on the host and want to setns() to the mount namespace. This requires you to
  78  * attach to the user namespace first otherwise the kernel will fail this check:
  79  *
  80  *        if (!ns_capable(mnt_ns->user_ns, CAP_SYS_ADMIN) ||
  81  *            !ns_capable(current_user_ns(), CAP_SYS_CHROOT) ||
  82  *            !ns_capable(current_user_ns(), CAP_SYS_ADMIN))
  83  *            return -EPERM;
  84  *
  85  *    in
  86  *
  87  *        linux/fs/namespace.c:mntns_install().
  88  */
  89 const struct ns_info ns_info[LXC_NS_MAX] = {
  90         [LXC_NS_USER]    = { "user",   CLONE_NEWUSER,   "CLONE_NEWUSER",   "LXC_USER_NS"    },
  91         [LXC_NS_MNT]    =  { "mnt",    CLONE_NEWNS,     "CLONE_NEWNS",     "LXC_MNT_NS"     },
  92         [LXC_NS_PID]    =  { "pid",    CLONE_NEWPID,    "CLONE_NEWPID",    "LXC_PID_NS"     },
  93         [LXC_NS_UTS]    =  { "uts",    CLONE_NEWUTS,    "CLONE_NEWUTS",    "LXC_UTS_NS"     },
  94         [LXC_NS_IPC]    =  { "ipc",    CLONE_NEWIPC,    "CLONE_NEWIPC",    "LXC_IPC_NS"     },
  95         [LXC_NS_NET]    =  { "net",    CLONE_NEWNET,    "CLONE_NEWNET",    "LXC_NET_NS"     },
  96         [LXC_NS_CGROUP] =  { "cgroup", CLONE_NEWCGROUP, "CLONE_NEWCGROUP", "LXC_CGROUP_NS"  }
  97 };
  98
  99 int lxc_namespace_2_cloneflag(const char *namespace)
 100 {
 101         int i;
 102         for (i = 0; i < LXC_NS_MAX; i++)
 103                 if (!strcasecmp(ns_info[i].proc_name, namespace))
 104                         return ns_info[i].clone_flag;
 105
 106         ERROR("Invalid namespace name \"%s\"", namespace);
 107         return -EINVAL;
 108 }
 109
 110 int lxc_namespace_2_ns_idx(const char *namespace)
 111 {
 112         int i;
 113         for (i = 0; i < LXC_NS_MAX; i++)
 114                 if (!strcmp(ns_info[i].proc_name, namespace))
 115                         return i;
 116
 117         ERROR("Invalid namespace name \"%s\"", namespace);
 118         return -EINVAL;
 119 }
 120
 121 int lxc_fill_namespace_flags(char *flaglist, int *flags)
 122 {
 123         char *token, *saveptr = NULL;
 124         int aflag;
 125
 126         if (!flaglist) {
 127                 ERROR("At least one namespace is needed.");
 128                 return -1;
 129         }
 130
 131         token = strtok_r(flaglist, "|", &saveptr);
 132         while (token) {
 133
 134                 aflag = lxc_namespace_2_cloneflag(token);
 135                 if (aflag < 0)
 136                         return -1;
 137
 138                 *flags |= aflag;
 139
 140                 token = strtok_r(NULL, "|", &saveptr);
 141         }
 142         return 0;
 143 }