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 #ifndef _GNU_SOURCE
  25 #define _GNU_SOURCE 1
  26 #endif
  27 #include <errno.h>
  28 #include <fcntl.h>
  29 #include <sched.h>
  30 #include <signal.h>
  31 #include <sys/param.h>
  32 #include <sys/stat.h>
  33 #include <sys/syscall.h>
  34 #include <sys/types.h>
  35 #include <unistd.h>
  36
  37 #include "config.h"
  38 #include "log.h"
  39 #include "memory_utils.h"
  40 #include "namespace.h"
  41 #include "utils.h"
  42
  43 lxc_log_define(namespace, lxc);
  44
  45 #define __LXC_STACK_SIZE (8 * 1024 * 1024)
  46 pid_t lxc_clone(int (*fn)(void *), void *arg, int flags, int *pidfd)
  47 {
  48         pid_t ret;
  49         void *stack;
  50
  51         stack = malloc(__LXC_STACK_SIZE);
  52         if (!stack) {
  53                 SYSERROR("Failed to allocate clone stack");
  54                 return -ENOMEM;
  55         }
  56
  57 #ifdef __ia64__
  58         ret = __clone2(fn, stack, __LXC_STACK_SIZE, flags | SIGCHLD, arg, pidfd);
  59 #else
  60         ret = clone(fn, stack + __LXC_STACK_SIZE, flags | SIGCHLD, arg, pidfd);
  61 #endif
  62         if (ret < 0)
  63                 SYSERROR("Failed to clone (%#x)", flags);
  64
  65         return ret;
  66 }
  67
  68 /* Leave the user namespace at the first position in the array of structs so
  69  * that we always attach to it first when iterating over the struct and using
  70  * setns() to switch namespaces. This especially affects lxc_attach(): Suppose
  71  * you cloned a new user namespace and mount namespace as an unprivileged user
  72  * on the host and want to setns() to the mount namespace. This requires you to
  73  * attach to the user namespace first otherwise the kernel will fail this check:
  74  *
  75  *        if (!ns_capable(mnt_ns->user_ns, CAP_SYS_ADMIN) ||
  76  *            !ns_capable(current_user_ns(), CAP_SYS_CHROOT) ||
  77  *            !ns_capable(current_user_ns(), CAP_SYS_ADMIN))
  78  *            return -EPERM;
  79  *
  80  *    in
  81  *
  82  *        linux/fs/namespace.c:mntns_install().
  83  */
  84 const struct ns_info ns_info[LXC_NS_MAX] = {
  85         [LXC_NS_USER]    = { "user",   CLONE_NEWUSER,   "CLONE_NEWUSER",   "LXC_USER_NS"    },
  86         [LXC_NS_MNT]    =  { "mnt",    CLONE_NEWNS,     "CLONE_NEWNS",     "LXC_MNT_NS"     },
  87         [LXC_NS_PID]    =  { "pid",    CLONE_NEWPID,    "CLONE_NEWPID",    "LXC_PID_NS"     },
  88         [LXC_NS_UTS]    =  { "uts",    CLONE_NEWUTS,    "CLONE_NEWUTS",    "LXC_UTS_NS"     },
  89         [LXC_NS_IPC]    =  { "ipc",    CLONE_NEWIPC,    "CLONE_NEWIPC",    "LXC_IPC_NS"     },
  90         [LXC_NS_NET]    =  { "net",    CLONE_NEWNET,    "CLONE_NEWNET",    "LXC_NET_NS"     },
  91         [LXC_NS_CGROUP] =  { "cgroup", CLONE_NEWCGROUP, "CLONE_NEWCGROUP", "LXC_CGROUP_NS"  }
  92 };
  93
  94 int lxc_namespace_2_cloneflag(const char *namespace)
  95 {
  96         int i;
  97
  98         for (i = 0; i < LXC_NS_MAX; i++)
  99                 if (!strcasecmp(ns_info[i].proc_name, namespace))
 100                         return ns_info[i].clone_flag;
 101
 102         ERROR("Invalid namespace name \"%s\"", namespace);
 103         return -EINVAL;
 104 }
 105
 106 int lxc_namespace_2_ns_idx(const char *namespace)
 107 {
 108         int i;
 109
 110         for (i = 0; i < LXC_NS_MAX; i++)
 111                 if (!strcmp(ns_info[i].proc_name, namespace))
 112                         return i;
 113
 114         ERROR("Invalid namespace name \"%s\"", namespace);
 115         return -EINVAL;
 116 }
 117
 118 extern int lxc_namespace_2_std_identifiers(char *namespaces)
 119 {
 120         char **it;
 121         char *del;
 122
 123         /* The identifiers for namespaces used with lxc-attach and lxc-unshare
 124          * as given on the manpage do not align with the standard identifiers.
 125          * This affects network, mount, and uts namespaces. The standard identifiers
 126          * are: "mnt", "uts", and "net" whereas lxc-attach and lxc-unshare uses
 127          * "MOUNT", "UTSNAME", and "NETWORK". So let's use some cheap memmove()s
 128          * to replace them by their standard identifiers.
 129          * Let's illustrate this with an example:
 130          * Assume the string:
 131          *
 132          *      "IPC|MOUNT|PID"
 133          *
 134          * then we memmove()
 135          *
 136          *      dest: del + 1 == OUNT|PID
 137          *      src:  del + 3 == NT|PID
 138          */
 139         if (!namespaces)
 140                 return -1;
 141
 142         while ((del = strstr(namespaces, "MOUNT")))
 143                 memmove(del + 1, del + 3, strlen(del) - 2);
 144
 145         for (it = (char *[]){"NETWORK", "UTSNAME", NULL}; it && *it; it++)
 146                 while ((del = strstr(namespaces, *it)))
 147                         memmove(del + 3, del + 7, strlen(del) - 6);
 148
 149         return 0;
 150 }
 151
 152 int lxc_fill_namespace_flags(char *flaglist, int *flags)
 153 {
 154         char *token;
 155         int aflag;
 156
 157         if (!flaglist) {
 158                 ERROR("At least one namespace is needed.");
 159                 return -1;
 160         }
 161
 162         lxc_iterate_parts(token, flaglist, "|") {
 163                 aflag = lxc_namespace_2_cloneflag(token);
 164                 if (aflag < 0)
 165                         return -1;
 166
 167                 *flags |= aflag;
 168         }
 169
 170         return 0;
 171 }