nsproxy.c 5.03 KB
Newer Older
1 2 3 4 5 6 7 8 9
/*
 *  Copyright (C) 2006 IBM Corporation
 *
 *  Author: Serge Hallyn <serue@us.ibm.com>
 *
 *  This program is free software; you can redistribute it and/or
 *  modify it under the terms of the GNU General Public License as
 *  published by the Free Software Foundation, version 2 of the
 *  License.
10 11 12 13
 *
 *  Jun 2006 - namespaces support
 *             OpenVZ, SWsoft Inc.
 *             Pavel Emelianov <xemul@openvz.org>
14 15 16 17
 */

#include <linux/module.h>
#include <linux/nsproxy.h>
18
#include <linux/init_task.h>
19
#include <linux/mnt_namespace.h>
20
#include <linux/utsname.h>
Cedric Le Goater's avatar
Cedric Le Goater committed
21
#include <linux/pid_namespace.h>
22
#include <net/net_namespace.h>
23
#include <linux/ipc_namespace.h>
24

25 26
static struct kmem_cache *nsproxy_cachep;

27
struct nsproxy init_nsproxy = INIT_NSPROXY(init_nsproxy);
28

29
static inline struct nsproxy *create_nsproxy(void)
30
{
31
	struct nsproxy *nsproxy;
32

33 34 35 36
	nsproxy = kmem_cache_alloc(nsproxy_cachep, GFP_KERNEL);
	if (nsproxy)
		atomic_set(&nsproxy->count, 1);
	return nsproxy;
37 38 39
}

/*
40 41 42
 * Create new nsproxy and all of its the associated namespaces.
 * Return the newly created nsproxy.  Do not attach this to the task,
 * leave it to the caller to do proper locking and attach it to task.
43
 */
44 45
static struct nsproxy *create_new_namespaces(unsigned long flags,
			struct task_struct *tsk, struct fs_struct *new_fs)
46
{
47
	struct nsproxy *new_nsp;
48
	int err;
49

50
	new_nsp = create_nsproxy();
51 52
	if (!new_nsp)
		return ERR_PTR(-ENOMEM);
53

54
	new_nsp->mnt_ns = copy_mnt_ns(flags, tsk->nsproxy->mnt_ns, new_fs);
55 56
	if (IS_ERR(new_nsp->mnt_ns)) {
		err = PTR_ERR(new_nsp->mnt_ns);
57
		goto out_ns;
58
	}
59 60

	new_nsp->uts_ns = copy_utsname(flags, tsk->nsproxy->uts_ns);
61 62
	if (IS_ERR(new_nsp->uts_ns)) {
		err = PTR_ERR(new_nsp->uts_ns);
63
		goto out_uts;
64
	}
65 66

	new_nsp->ipc_ns = copy_ipcs(flags, tsk->nsproxy->ipc_ns);
67 68
	if (IS_ERR(new_nsp->ipc_ns)) {
		err = PTR_ERR(new_nsp->ipc_ns);
69
		goto out_ipc;
70
	}
71

72
	new_nsp->pid_ns = copy_pid_ns(flags, task_active_pid_ns(tsk));
73 74
	if (IS_ERR(new_nsp->pid_ns)) {
		err = PTR_ERR(new_nsp->pid_ns);
75
		goto out_pid;
76
	}
77

78 79 80 81 82 83
	new_nsp->net_ns = copy_net_ns(flags, tsk->nsproxy->net_ns);
	if (IS_ERR(new_nsp->net_ns)) {
		err = PTR_ERR(new_nsp->net_ns);
		goto out_net;
	}

84 85
	return new_nsp;

86
out_net:
87 88
	if (new_nsp->pid_ns)
		put_pid_ns(new_nsp->pid_ns);
89 90 91 92 93 94 95 96 97 98
out_pid:
	if (new_nsp->ipc_ns)
		put_ipc_ns(new_nsp->ipc_ns);
out_ipc:
	if (new_nsp->uts_ns)
		put_uts_ns(new_nsp->uts_ns);
out_uts:
	if (new_nsp->mnt_ns)
		put_mnt_ns(new_nsp->mnt_ns);
out_ns:
99
	kmem_cache_free(nsproxy_cachep, new_nsp);
100
	return ERR_PTR(err);
101 102 103 104 105 106
}

/*
 * called from clone.  This now handles copy for nsproxy and all
 * namespaces therein.
 */
107
int copy_namespaces(unsigned long flags, struct task_struct *tsk)
108 109
{
	struct nsproxy *old_ns = tsk->nsproxy;
110 111
	struct nsproxy *new_ns;
	int err = 0;
112 113 114 115 116 117

	if (!old_ns)
		return 0;

	get_nsproxy(old_ns);

118
	if (!(flags & (CLONE_NEWNS | CLONE_NEWUTS | CLONE_NEWIPC |
119
				CLONE_NEWPID | CLONE_NEWNET)))
120 121
		return 0;

122 123
	if (!capable(CAP_SYS_ADMIN)) {
		err = -EPERM;
124 125 126
		goto out;
	}

127 128 129 130 131 132 133 134 135 136 137 138
	/*
	 * CLONE_NEWIPC must detach from the undolist: after switching
	 * to a new ipc namespace, the semaphore arrays from the old
	 * namespace are unreachable.  In clone parlance, CLONE_SYSVSEM
	 * means share undolist with parent, so we must forbid using
	 * it along with CLONE_NEWIPC.
	 */
	if ((flags & CLONE_NEWIPC) && (flags & CLONE_SYSVSEM)) {
		err = -EINVAL;
		goto out;
	}

139 140 141 142 143
	new_ns = create_new_namespaces(flags, tsk, tsk->fs);
	if (IS_ERR(new_ns)) {
		err = PTR_ERR(new_ns);
		goto out;
	}
Cedric Le Goater's avatar
Cedric Le Goater committed
144

145
	tsk->nsproxy = new_ns;
146

147
out:
148
	put_nsproxy(old_ns);
149
	return err;
150 151 152 153
}

void free_nsproxy(struct nsproxy *ns)
{
Cedric Le Goater's avatar
Cedric Le Goater committed
154 155 156 157 158 159 160 161
	if (ns->mnt_ns)
		put_mnt_ns(ns->mnt_ns);
	if (ns->uts_ns)
		put_uts_ns(ns->uts_ns);
	if (ns->ipc_ns)
		put_ipc_ns(ns->ipc_ns);
	if (ns->pid_ns)
		put_pid_ns(ns->pid_ns);
162
	put_net(ns->net_ns);
163
	kmem_cache_free(nsproxy_cachep, ns);
164
}
165 166 167

/*
 * Called from unshare. Unshare all the namespaces part of nsproxy.
168
 * On success, returns the new nsproxy.
169 170 171 172 173 174
 */
int unshare_nsproxy_namespaces(unsigned long unshare_flags,
		struct nsproxy **new_nsp, struct fs_struct *new_fs)
{
	int err = 0;

175
	if (!(unshare_flags & (CLONE_NEWNS | CLONE_NEWUTS | CLONE_NEWIPC |
176
			       CLONE_NEWNET)))
177 178 179 180 181 182 183
		return 0;

	if (!capable(CAP_SYS_ADMIN))
		return -EPERM;

	*new_nsp = create_new_namespaces(unshare_flags, current,
				new_fs ? new_fs : current->fs);
184
	if (IS_ERR(*new_nsp)) {
185
		err = PTR_ERR(*new_nsp);
186 187 188
		goto out;
	}

189
	err = ns_cgroup_clone(current, task_pid(current));
190 191 192 193
	if (err)
		put_nsproxy(*new_nsp);

out:
194 195
	return err;
}
196

197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223
void switch_task_namespaces(struct task_struct *p, struct nsproxy *new)
{
	struct nsproxy *ns;

	might_sleep();

	ns = p->nsproxy;

	rcu_assign_pointer(p->nsproxy, new);

	if (ns && atomic_dec_and_test(&ns->count)) {
		/*
		 * wait for others to get what they want from this nsproxy.
		 *
		 * cannot release this nsproxy via the call_rcu() since
		 * put_mnt_ns() will want to sleep
		 */
		synchronize_rcu();
		free_nsproxy(ns);
	}
}

void exit_task_namespaces(struct task_struct *p)
{
	switch_task_namespaces(p, NULL);
}

224 225
static int __init nsproxy_cache_init(void)
{
226
	nsproxy_cachep = KMEM_CACHE(nsproxy, SLAB_PANIC);
227 228 229 230
	return 0;
}

module_init(nsproxy_cache_init);