[mirror_ubuntu-artful-kernel.git] / net / core / net_namespace.c

#include <linux/workqueue.h>
#include <linux/rtnetlink.h>
#include <linux/cache.h>
#include <linux/slab.h>
#include <linux/list.h>
#include <linux/delay.h>
#include <linux/sched.h>
#include <linux/idr.h>
#include <linux/rculist.h>
#include <linux/nsproxy.h>
#include <linux/proc_fs.h>
#include <linux/file.h>
#include <linux/export.h>
#include <net/net_namespace.h>
#include <net/netns/generic.h>

/*
 *	Our network namespace constructor/destructor lists
 */

static LIST_HEAD(pernet_list);
static struct list_head *first_device = &pernet_list;
static DEFINE_MUTEX(net_mutex);

LIST_HEAD(net_namespace_list);
EXPORT_SYMBOL_GPL(net_namespace_list);

struct net init_net;
EXPORT_SYMBOL(init_net);

#define INITIAL_NET_GEN_PTRS	13 /* +1 for len +2 for rcu_head */

static int net_assign_generic(struct net *net, int id, void *data)
{
	struct net_generic *ng, *old_ng;

	BUG_ON(!mutex_is_locked(&net_mutex));
	BUG_ON(id == 0);

	old_ng = rcu_dereference_protected(net->gen,
					   lockdep_is_held(&net_mutex));
	ng = old_ng;
	if (old_ng->len >= id)
		goto assign;

	ng = kzalloc(sizeof(struct net_generic) +
			id * sizeof(void *), GFP_KERNEL);
	if (ng == NULL)
		return -ENOMEM;

	/*
	 * Some synchronisation notes:
	 *
	 * The net_generic explores the net->gen array inside rcu
	 * read section. Besides once set the net->gen->ptr[x]
	 * pointer never changes (see rules in netns/generic.h).
	 *
	 * That said, we simply duplicate this array and schedule
	 * the old copy for kfree after a grace period.
	 */

	ng->len = id;
	memcpy(&ng->ptr, &old_ng->ptr, old_ng->len * sizeof(void*));

	rcu_assign_pointer(net->gen, ng);
	kfree_rcu(old_ng, rcu);
assign:
	ng->ptr[id - 1] = data;
	return 0;
}

static int ops_init(const struct pernet_operations *ops, struct net *net)
{
	int err;
	if (ops->id && ops->size) {
		void *data = kzalloc(ops->size, GFP_KERNEL);
		if (!data)
			return -ENOMEM;

		err = net_assign_generic(net, *ops->id, data);
		if (err) {
			kfree(data);
			return err;
		}
	}
	if (ops->init)
		return ops->init(net);
	return 0;
}

static void ops_free(const struct pernet_operations *ops, struct net *net)
{
	if (ops->id && ops->size) {
		int id = *ops->id;
		kfree(net_generic(net, id));
	}
}

static void ops_exit_list(const struct pernet_operations *ops,
			  struct list_head *net_exit_list)
{
	struct net *net;
	if (ops->exit) {
		list_for_each_entry(net, net_exit_list, exit_list)
			ops->exit(net);
	}
	if (ops->exit_batch)
		ops->exit_batch(net_exit_list);
}

static void ops_free_list(const struct pernet_operations *ops,
			  struct list_head *net_exit_list)
{
	struct net *net;
	if (ops->size && ops->id) {
		list_for_each_entry(net, net_exit_list, exit_list)
			ops_free(ops, net);
	}
}

/*
 * setup_net runs the initializers for the network namespace object.
 */
static __net_init int setup_net(struct net *net)
{
	/* Must be called with net_mutex held */
	const struct pernet_operations *ops, *saved_ops;
	int error = 0;
	LIST_HEAD(net_exit_list);

	atomic_set(&net->count, 1);
	atomic_set(&net->passive, 1);
	net->dev_base_seq = 1;

#ifdef NETNS_REFCNT_DEBUG
	atomic_set(&net->use_count, 0);
#endif

	list_for_each_entry(ops, &pernet_list, list) {
		error = ops_init(ops, net);
		if (error < 0)
			goto out_undo;
	}
out:
	return error;

out_undo:
	/* Walk through the list backwards calling the exit functions
	 * for the pernet modules whose init functions did not fail.
	 */
	list_add(&net->exit_list, &net_exit_list);
	saved_ops = ops;
	list_for_each_entry_continue_reverse(ops, &pernet_list, list)
		ops_exit_list(ops, &net_exit_list);

	ops = saved_ops;
	list_for_each_entry_continue_reverse(ops, &pernet_list, list)
		ops_free_list(ops, &net_exit_list);

	rcu_barrier();
	goto out;
}

static struct net_generic *net_alloc_generic(void)
{
	struct net_generic *ng;
	size_t generic_size = sizeof(struct net_generic) +
		INITIAL_NET_GEN_PTRS * sizeof(void *);

	ng = kzalloc(generic_size, GFP_KERNEL);
	if (ng)
		ng->len = INITIAL_NET_GEN_PTRS;

	return ng;
}

#ifdef CONFIG_NET_NS
static struct kmem_cache *net_cachep;
static struct workqueue_struct *netns_wq;

static struct net *net_alloc(void)
{
	struct net *net = NULL;
	struct net_generic *ng;

	ng = net_alloc_generic();
	if (!ng)
		goto out;

	net = kmem_cache_zalloc(net_cachep, GFP_KERNEL);
	if (!net)
		goto out_free;

	rcu_assign_pointer(net->gen, ng);
out:
	return net;

out_free:
	kfree(ng);
	goto out;
}

static void net_free(struct net *net)
{
#ifdef NETNS_REFCNT_DEBUG
	if (unlikely(atomic_read(&net->use_count) != 0)) {
		printk(KERN_EMERG "network namespace not free! Usage: %d\n",
			atomic_read(&net->use_count));
		return;
	}
#endif
	kfree(net->gen);
	kmem_cache_free(net_cachep, net);
}

void net_drop_ns(void *p)
{
	struct net *ns = p;
	if (ns && atomic_dec_and_test(&ns->passive))
		net_free(ns);
}

struct net *copy_net_ns(unsigned long flags, struct net *old_net)
{
	struct net *net;
	int rv;

	if (!(flags & CLONE_NEWNET))
		return get_net(old_net);

	net = net_alloc();
	if (!net)
		return ERR_PTR(-ENOMEM);
	mutex_lock(&net_mutex);
	rv = setup_net(net);
	if (rv == 0) {
		rtnl_lock();
		list_add_tail_rcu(&net->list, &net_namespace_list);
		rtnl_unlock();
	}
	mutex_unlock(&net_mutex);
	if (rv < 0) {
		net_drop_ns(net);
		return ERR_PTR(rv);
	}
	return net;
}

static DEFINE_SPINLOCK(cleanup_list_lock);
static LIST_HEAD(cleanup_list);  /* Must hold cleanup_list_lock to touch */

static void cleanup_net(struct work_struct *work)
{
	const struct pernet_operations *ops;
	struct net *net, *tmp;
	LIST_HEAD(net_kill_list);
	LIST_HEAD(net_exit_list);

	/* Atomically snapshot the list of namespaces to cleanup */
	spin_lock_irq(&cleanup_list_lock);
	list_replace_init(&cleanup_list, &net_kill_list);
	spin_unlock_irq(&cleanup_list_lock);

	mutex_lock(&net_mutex);

	/* Don't let anyone else find us. */
	rtnl_lock();
	list_for_each_entry(net, &net_kill_list, cleanup_list) {
		list_del_rcu(&net->list);
		list_add_tail(&net->exit_list, &net_exit_list);
	}
	rtnl_unlock();

	/*
	 * Another CPU might be rcu-iterating the list, wait for it.
	 * This needs to be before calling the exit() notifiers, so
	 * the rcu_barrier() below isn't sufficient alone.
	 */
	synchronize_rcu();

	/* Run all of the network namespace exit methods */
	list_for_each_entry_reverse(ops, &pernet_list, list)
		ops_exit_list(ops, &net_exit_list);

	/* Free the net generic variables */
	list_for_each_entry_reverse(ops, &pernet_list, list)
		ops_free_list(ops, &net_exit_list);

	mutex_unlock(&net_mutex);

	/* Ensure there are no outstanding rcu callbacks using this
	 * network namespace.
	 */
	rcu_barrier();

	/* Finally it is safe to free my network namespace structure */
	list_for_each_entry_safe(net, tmp, &net_exit_list, exit_list) {
		list_del_init(&net->exit_list);
		net_drop_ns(net);
	}
}
static DECLARE_WORK(net_cleanup_work, cleanup_net);

void __put_net(struct net *net)
{
	/* Cleanup the network namespace in process context */
	unsigned long flags;

	spin_lock_irqsave(&cleanup_list_lock, flags);
	list_add(&net->cleanup_list, &cleanup_list);
	spin_unlock_irqrestore(&cleanup_list_lock, flags);

	queue_work(netns_wq, &net_cleanup_work);
}
EXPORT_SYMBOL_GPL(__put_net);

struct net *get_net_ns_by_fd(int fd)
{
	struct proc_inode *ei;
	struct file *file;
	struct net *net;

	file = proc_ns_fget(fd);
	if (IS_ERR(file))
		return ERR_CAST(file);

	ei = PROC_I(file->f_dentry->d_inode);
	if (ei->ns_ops == &netns_operations)
		net = get_net(ei->ns);
	else
		net = ERR_PTR(-EINVAL);

	fput(file);
	return net;
}

#else
struct net *copy_net_ns(unsigned long flags, struct net *old_net)
{
	if (flags & CLONE_NEWNET)
		return ERR_PTR(-EINVAL);
	return old_net;
}

struct net *get_net_ns_by_fd(int fd)
{
	return ERR_PTR(-EINVAL);
}
#endif

struct net *get_net_ns_by_pid(pid_t pid)
{
	struct task_struct *tsk;
	struct net *net;

	/* Lookup the network namespace */
	net = ERR_PTR(-ESRCH);
	rcu_read_lock();
	tsk = find_task_by_vpid(pid);
	if (tsk) {
		struct nsproxy *nsproxy;
		nsproxy = task_nsproxy(tsk);
		if (nsproxy)
			net = get_net(nsproxy->net_ns);
	}
	rcu_read_unlock();
	return net;
}
EXPORT_SYMBOL_GPL(get_net_ns_by_pid);

static int __init net_ns_init(void)
{
	struct net_generic *ng;

#ifdef CONFIG_NET_NS
	net_cachep = kmem_cache_create("net_namespace", sizeof(struct net),
					SMP_CACHE_BYTES,
					SLAB_PANIC, NULL);

	/* Create workqueue for cleanup */
	netns_wq = create_singlethread_workqueue("netns");
	if (!netns_wq)
		panic("Could not create netns workq");
#endif

	ng = net_alloc_generic();
	if (!ng)
		panic("Could not allocate generic netns");

	rcu_assign_pointer(init_net.gen, ng);

	mutex_lock(&net_mutex);
	if (setup_net(&init_net))
		panic("Could not setup the initial network namespace");

	rtnl_lock();
	list_add_tail_rcu(&init_net.list, &net_namespace_list);
	rtnl_unlock();

	mutex_unlock(&net_mutex);

	return 0;
}

pure_initcall(net_ns_init);

#ifdef CONFIG_NET_NS
static int __register_pernet_operations(struct list_head *list,
					struct pernet_operations *ops)
{
	struct net *net;
	int error;
	LIST_HEAD(net_exit_list);

	list_add_tail(&ops->list, list);
	if (ops->init || (ops->id && ops->size)) {
		for_each_net(net) {
			error = ops_init(ops, net);
			if (error)
				goto out_undo;
			list_add_tail(&net->exit_list, &net_exit_list);
		}
	}
	return 0;

out_undo:
	/* If I have an error cleanup all namespaces I initialized */
	list_del(&ops->list);
	ops_exit_list(ops, &net_exit_list);
	ops_free_list(ops, &net_exit_list);
	return error;
}

static void __unregister_pernet_operations(struct pernet_operations *ops)
{
	struct net *net;
	LIST_HEAD(net_exit_list);

	list_del(&ops->list);
	for_each_net(net)
		list_add_tail(&net->exit_list, &net_exit_list);
	ops_exit_list(ops, &net_exit_list);
	ops_free_list(ops, &net_exit_list);
}

#else

static int __register_pernet_operations(struct list_head *list,
					struct pernet_operations *ops)
{
	int err = 0;
	err = ops_init(ops, &init_net);
	if (err)
		ops_free(ops, &init_net);
	return err;
	
}

static void __unregister_pernet_operations(struct pernet_operations *ops)
{
	LIST_HEAD(net_exit_list);
	list_add(&init_net.exit_list, &net_exit_list);
	ops_exit_list(ops, &net_exit_list);
	ops_free_list(ops, &net_exit_list);
}

#endif /* CONFIG_NET_NS */

static DEFINE_IDA(net_generic_ids);

static int register_pernet_operations(struct list_head *list,
				      struct pernet_operations *ops)
{
	int error;

	if (ops->id) {
again:
		error = ida_get_new_above(&net_generic_ids, 1, ops->id);
		if (error < 0) {
			if (error == -EAGAIN) {
				ida_pre_get(&net_generic_ids, GFP_KERNEL);
				goto again;
			}
			return error;
		}
	}
	error = __register_pernet_operations(list, ops);
	if (error) {
		rcu_barrier();
		if (ops->id)
			ida_remove(&net_generic_ids, *ops->id);
	}

	return error;
}

static void unregister_pernet_operations(struct pernet_operations *ops)
{
	
	__unregister_pernet_operations(ops);
	rcu_barrier();
	if (ops->id)
		ida_remove(&net_generic_ids, *ops->id);
}

/**
 *      register_pernet_subsys - register a network namespace subsystem
 *	@ops:  pernet operations structure for the subsystem
 *
 *	Register a subsystem which has init and exit functions
 *	that are called when network namespaces are created and
 *	destroyed respectively.
 *
 *	When registered all network namespace init functions are
 *	called for every existing network namespace.  Allowing kernel
 *	modules to have a race free view of the set of network namespaces.
 *
 *	When a new network namespace is created all of the init
 *	methods are called in the order in which they were registered.
 *
 *	When a network namespace is destroyed all of the exit methods
 *	are called in the reverse of the order with which they were
 *	registered.
 */
int register_pernet_subsys(struct pernet_operations *ops)
{
	int error;
	mutex_lock(&net_mutex);
	error =  register_pernet_operations(first_device, ops);
	mutex_unlock(&net_mutex);
	return error;
}
EXPORT_SYMBOL_GPL(register_pernet_subsys);

/**
 *      unregister_pernet_subsys - unregister a network namespace subsystem
 *	@ops: pernet operations structure to manipulate
 *
 *	Remove the pernet operations structure from the list to be
 *	used when network namespaces are created or destroyed.  In
 *	addition run the exit method for all existing network
 *	namespaces.
 */
void unregister_pernet_subsys(struct pernet_operations *ops)
{
	mutex_lock(&net_mutex);
	unregister_pernet_operations(ops);
	mutex_unlock(&net_mutex);
}
EXPORT_SYMBOL_GPL(unregister_pernet_subsys);

/**
 *      register_pernet_device - register a network namespace device
 *	@ops:  pernet operations structure for the subsystem
 *
 *	Register a device which has init and exit functions
 *	that are called when network namespaces are created and
 *	destroyed respectively.
 *
 *	When registered all network namespace init functions are
 *	called for every existing network namespace.  Allowing kernel
 *	modules to have a race free view of the set of network namespaces.
 *
 *	When a new network namespace is created all of the init
 *	methods are called in the order in which they were registered.
 *
 *	When a network namespace is destroyed all of the exit methods
 *	are called in the reverse of the order with which they were
 *	registered.
 */
int register_pernet_device(struct pernet_operations *ops)
{
	int error;
	mutex_lock(&net_mutex);
	error = register_pernet_operations(&pernet_list, ops);
	if (!error && (first_device == &pernet_list))
		first_device = &ops->list;
	mutex_unlock(&net_mutex);
	return error;
}
EXPORT_SYMBOL_GPL(register_pernet_device);

/**
 *      unregister_pernet_device - unregister a network namespace netdevice
 *	@ops: pernet operations structure to manipulate
 *
 *	Remove the pernet operations structure from the list to be
 *	used when network namespaces are created or destroyed.  In
 *	addition run the exit method for all existing network
 *	namespaces.
 */
void unregister_pernet_device(struct pernet_operations *ops)
{
	mutex_lock(&net_mutex);
	if (&ops->list == first_device)
		first_device = first_device->next;
	unregister_pernet_operations(ops);
	mutex_unlock(&net_mutex);
}
EXPORT_SYMBOL_GPL(unregister_pernet_device);

#ifdef CONFIG_NET_NS
static void *netns_get(struct task_struct *task)
{
	struct net *net = NULL;
	struct nsproxy *nsproxy;

	rcu_read_lock();
	nsproxy = task_nsproxy(task);
	if (nsproxy)
		net = get_net(nsproxy->net_ns);
	rcu_read_unlock();

	return net;
}

static void netns_put(void *ns)
{
	put_net(ns);
}

static int netns_install(struct nsproxy *nsproxy, void *ns)
{
	put_net(nsproxy->net_ns);
	nsproxy->net_ns = get_net(ns);
	return 0;
}

const struct proc_ns_operations netns_operations = {
	.name		= "net",
	.type		= CLONE_NEWNET,
	.get		= netns_get,
	.put		= netns_put,
	.install	= netns_install,
};
#endif
Commit	Line	Data
5f256bec EB	1	#include <linux/workqueue.h>
	2	#include <linux/rtnetlink.h>
	3	#include <linux/cache.h>
	4	#include <linux/slab.h>
	5	#include <linux/list.h>
	6	#include <linux/delay.h>
9dd776b6	7	#include <linux/sched.h>
c93cf61f	8	#include <linux/idr.h>
11a28d37	9	#include <linux/rculist.h>
30ffee84	10	#include <linux/nsproxy.h>
f0630529 EB	11	#include <linux/proc_fs.h>
f0630529 EB	12	#include <linux/file.h>
bc3b2d7f	13	#include <linux/export.h>
5f256bec	14	#include <net/net_namespace.h>
dec827d1	15	#include <net/netns/generic.h>
5f256bec EB	16
	17	/*
	18	* Our network namespace constructor/destructor lists
	19	*/
	20
	21	static LIST_HEAD(pernet_list);
	22	static struct list_head *first_device = &pernet_list;
	23	static DEFINE_MUTEX(net_mutex);
	24
5f256bec	25	LIST_HEAD(net_namespace_list);
b76a461f	26	EXPORT_SYMBOL_GPL(net_namespace_list);
5f256bec	27
5f256bec	28	struct net init_net;
ff4b9502	29	EXPORT_SYMBOL(init_net);
5f256bec	30
dec827d1 PE	31	#define INITIAL_NET_GEN_PTRS 13 /* +1 for len +2 for rcu_head */
dec827d1 PE	32
05fceb4a JP	33	static int net_assign_generic(struct net net, int id, void data)
	34	{
	35	struct net_generic ng, old_ng;
	36
	37	BUG_ON(!mutex_is_locked(&net_mutex));
	38	BUG_ON(id == 0);
	39
1c87733d ED	40	old_ng = rcu_dereference_protected(net->gen,
	41	lockdep_is_held(&net_mutex));
	42	ng = old_ng;
05fceb4a JP	43	if (old_ng->len >= id)
	44	goto assign;
	45
	46	ng = kzalloc(sizeof(struct net_generic) +
	47	id * sizeof(void *), GFP_KERNEL);
	48	if (ng == NULL)
	49	return -ENOMEM;
	50
	51	/*
	52	* Some synchronisation notes:
	53	*
	54	* The net_generic explores the net->gen array inside rcu
	55	* read section. Besides once set the net->gen->ptr[x]
	56	* pointer never changes (see rules in netns/generic.h).
	57	*
	58	* That said, we simply duplicate this array and schedule
	59	* the old copy for kfree after a grace period.
	60	*/
	61
	62	ng->len = id;
	63	memcpy(&ng->ptr, &old_ng->ptr, old_ng->len * sizeof(void*));
	64
	65	rcu_assign_pointer(net->gen, ng);
04d4dfed	66	kfree_rcu(old_ng, rcu);
05fceb4a JP	67	assign:
	68	ng->ptr[id - 1] = data;
	69	return 0;
	70	}
	71
f875bae0 EB	72	static int ops_init(const struct pernet_operations ops, struct net net)
	73	{
	74	int err;
	75	if (ops->id && ops->size) {
	76	void *data = kzalloc(ops->size, GFP_KERNEL);
	77	if (!data)
	78	return -ENOMEM;
	79
	80	err = net_assign_generic(net, *ops->id, data);
	81	if (err) {
	82	kfree(data);
	83	return err;
	84	}
	85	}
	86	if (ops->init)
	87	return ops->init(net);
	88	return 0;
	89	}
	90
	91	static void ops_free(const struct pernet_operations ops, struct net net)
	92	{
	93	if (ops->id && ops->size) {
	94	int id = *ops->id;
	95	kfree(net_generic(net, id));
	96	}
	97	}
	98
72ad937a EB	99	static void ops_exit_list(const struct pernet_operations *ops,
	100	struct list_head *net_exit_list)
	101	{
	102	struct net *net;
	103	if (ops->exit) {
	104	list_for_each_entry(net, net_exit_list, exit_list)
	105	ops->exit(net);
	106	}
72ad937a EB	107	if (ops->exit_batch)
	108	ops->exit_batch(net_exit_list);
	109	}
	110
	111	static void ops_free_list(const struct pernet_operations *ops,
	112	struct list_head *net_exit_list)
	113	{
	114	struct net *net;
	115	if (ops->size && ops->id) {
	116	list_for_each_entry(net, net_exit_list, exit_list)
	117	ops_free(ops, net);
	118	}
	119	}
	120
5f256bec EB	121	/*
	122	* setup_net runs the initializers for the network namespace object.
	123	*/
1a2ee93d	124	static __net_init int setup_net(struct net *net)
5f256bec EB	125	{
5f256bec EB	126	/* Must be called with net_mutex held */
f875bae0	127	const struct pernet_operations ops, saved_ops;
486a87f1	128	int error = 0;
72ad937a	129	LIST_HEAD(net_exit_list);
5f256bec	130
5f256bec	131	atomic_set(&net->count, 1);
a685e089	132	atomic_set(&net->passive, 1);
4e985ada	133	net->dev_base_seq = 1;
486a87f1	134
5d1e4468	135	#ifdef NETNS_REFCNT_DEBUG
5f256bec	136	atomic_set(&net->use_count, 0);
5d1e4468	137	#endif
5f256bec	138
768f3591	139	list_for_each_entry(ops, &pernet_list, list) {
f875bae0 EB	140	error = ops_init(ops, net);
	141	if (error < 0)
	142	goto out_undo;
5f256bec EB	143	}
	144	out:
	145	return error;
768f3591	146
5f256bec EB	147	out_undo:
	148	/* Walk through the list backwards calling the exit functions
	149	* for the pernet modules whose init functions did not fail.
	150	*/
72ad937a	151	list_add(&net->exit_list, &net_exit_list);
f875bae0	152	saved_ops = ops;
72ad937a EB	153	list_for_each_entry_continue_reverse(ops, &pernet_list, list)
	154	ops_exit_list(ops, &net_exit_list);
	155
f875bae0 EB	156	ops = saved_ops;
f875bae0 EB	157	list_for_each_entry_continue_reverse(ops, &pernet_list, list)
72ad937a	158	ops_free_list(ops, &net_exit_list);
310928d9 DL	159
310928d9 DL	160	rcu_barrier();
5f256bec EB	161	goto out;
	162	}
	163
486a87f1	164	static struct net_generic *net_alloc_generic(void)
6a1a3b9f	165	{
486a87f1 DL	166	struct net_generic *ng;
	167	size_t generic_size = sizeof(struct net_generic) +
	168	INITIAL_NET_GEN_PTRS * sizeof(void *);
	169
	170	ng = kzalloc(generic_size, GFP_KERNEL);
	171	if (ng)
	172	ng->len = INITIAL_NET_GEN_PTRS;
	173
	174	return ng;
6a1a3b9f PE	175	}
6a1a3b9f PE	176
ebe47d47 CN	177	#ifdef CONFIG_NET_NS
	178	static struct kmem_cache *net_cachep;
	179	static struct workqueue_struct *netns_wq;
	180
486a87f1	181	static struct net *net_alloc(void)
45a19b0a	182	{
486a87f1 DL	183	struct net *net = NULL;
	184	struct net_generic *ng;
	185
	186	ng = net_alloc_generic();
	187	if (!ng)
	188	goto out;
	189
	190	net = kmem_cache_zalloc(net_cachep, GFP_KERNEL);
45a19b0a	191	if (!net)
486a87f1	192	goto out_free;
45a19b0a	193
486a87f1 DL	194	rcu_assign_pointer(net->gen, ng);
	195	out:
	196	return net;
	197
	198	out_free:
	199	kfree(ng);
	200	goto out;
	201	}
	202
	203	static void net_free(struct net *net)
	204	{
5d1e4468	205	#ifdef NETNS_REFCNT_DEBUG
45a19b0a JFS	206	if (unlikely(atomic_read(&net->use_count) != 0)) {
	207	printk(KERN_EMERG "network namespace not free! Usage: %d\n",
	208	atomic_read(&net->use_count));
	209	return;
	210	}
5d1e4468	211	#endif
4ef079cc	212	kfree(net->gen);
45a19b0a JFS	213	kmem_cache_free(net_cachep, net);
	214	}
	215
a685e089 AV	216	void net_drop_ns(void *p)
	217	{
	218	struct net *ns = p;
	219	if (ns && atomic_dec_and_test(&ns->passive))
	220	net_free(ns);
	221	}
	222
911cb193	223	struct net copy_net_ns(unsigned long flags, struct net old_net)
9dd776b6	224	{
088eb2d9 AD	225	struct net *net;
088eb2d9 AD	226	int rv;
9dd776b6	227
911cb193 RL	228	if (!(flags & CLONE_NEWNET))
	229	return get_net(old_net);
	230
088eb2d9 AD	231	net = net_alloc();
	232	if (!net)
	233	return ERR_PTR(-ENOMEM);
9dd776b6	234	mutex_lock(&net_mutex);
088eb2d9 AD	235	rv = setup_net(net);
088eb2d9 AD	236	if (rv == 0) {
486a87f1	237	rtnl_lock();
11a28d37	238	list_add_tail_rcu(&net->list, &net_namespace_list);
486a87f1 DL	239	rtnl_unlock();
486a87f1 DL	240	}
9dd776b6	241	mutex_unlock(&net_mutex);
088eb2d9	242	if (rv < 0) {
a685e089	243	net_drop_ns(net);
088eb2d9 AD	244	return ERR_PTR(rv);
	245	}
	246	return net;
	247	}
486a87f1	248
2b035b39 EB	249	static DEFINE_SPINLOCK(cleanup_list_lock);
	250	static LIST_HEAD(cleanup_list); /* Must hold cleanup_list_lock to touch */
	251
6a1a3b9f PE	252	static void cleanup_net(struct work_struct *work)
6a1a3b9f PE	253	{
f875bae0	254	const struct pernet_operations *ops;
2b035b39 EB	255	struct net net, tmp;
2b035b39 EB	256	LIST_HEAD(net_kill_list);
72ad937a	257	LIST_HEAD(net_exit_list);
6a1a3b9f	258
2b035b39 EB	259	/* Atomically snapshot the list of namespaces to cleanup */
	260	spin_lock_irq(&cleanup_list_lock);
	261	list_replace_init(&cleanup_list, &net_kill_list);
	262	spin_unlock_irq(&cleanup_list_lock);
6a1a3b9f PE	263
	264	mutex_lock(&net_mutex);
	265
	266	/* Don't let anyone else find us. */
	267	rtnl_lock();
72ad937a	268	list_for_each_entry(net, &net_kill_list, cleanup_list) {
2b035b39	269	list_del_rcu(&net->list);
72ad937a EB	270	list_add_tail(&net->exit_list, &net_exit_list);
72ad937a EB	271	}
6a1a3b9f PE	272	rtnl_unlock();
6a1a3b9f PE	273
11a28d37 JB	274	/*
	275	* Another CPU might be rcu-iterating the list, wait for it.
	276	* This needs to be before calling the exit() notifiers, so
	277	* the rcu_barrier() below isn't sufficient alone.
	278	*/
	279	synchronize_rcu();
	280
6a1a3b9f	281	/* Run all of the network namespace exit methods */
72ad937a EB	282	list_for_each_entry_reverse(ops, &pernet_list, list)
	283	ops_exit_list(ops, &net_exit_list);
	284
f875bae0	285	/* Free the net generic variables */
72ad937a EB	286	list_for_each_entry_reverse(ops, &pernet_list, list)
72ad937a EB	287	ops_free_list(ops, &net_exit_list);
6a1a3b9f PE	288
	289	mutex_unlock(&net_mutex);
	290
	291	/* Ensure there are no outstanding rcu callbacks using this
	292	* network namespace.
	293	*/
	294	rcu_barrier();
	295
	296	/* Finally it is safe to free my network namespace structure */
72ad937a EB	297	list_for_each_entry_safe(net, tmp, &net_exit_list, exit_list) {
72ad937a EB	298	list_del_init(&net->exit_list);
a685e089	299	net_drop_ns(net);
2b035b39	300	}
6a1a3b9f	301	}
2b035b39	302	static DECLARE_WORK(net_cleanup_work, cleanup_net);
6a1a3b9f PE	303
	304	void __put_net(struct net *net)
	305	{
	306	/* Cleanup the network namespace in process context */
2b035b39 EB	307	unsigned long flags;
	308
	309	spin_lock_irqsave(&cleanup_list_lock, flags);
	310	list_add(&net->cleanup_list, &cleanup_list);
	311	spin_unlock_irqrestore(&cleanup_list_lock, flags);
	312
	313	queue_work(netns_wq, &net_cleanup_work);
6a1a3b9f PE	314	}
	315	EXPORT_SYMBOL_GPL(__put_net);
	316
956c9207 SR	317	struct net *get_net_ns_by_fd(int fd)
	318	{
	319	struct proc_inode *ei;
	320	struct file *file;
	321	struct net *net;
	322
956c9207	323	file = proc_ns_fget(fd);
c316e6a3 AV	324	if (IS_ERR(file))
c316e6a3 AV	325	return ERR_CAST(file);
956c9207 SR	326
956c9207 SR	327	ei = PROC_I(file->f_dentry->d_inode);
c316e6a3 AV	328	if (ei->ns_ops == &netns_operations)
	329	net = get_net(ei->ns);
	330	else
	331	net = ERR_PTR(-EINVAL);
956c9207	332
c316e6a3	333	fput(file);
956c9207 SR	334	return net;
	335	}
	336
6a1a3b9f PE	337	#else
	338	struct net copy_net_ns(unsigned long flags, struct net old_net)
	339	{
	340	if (flags & CLONE_NEWNET)
	341	return ERR_PTR(-EINVAL);
	342	return old_net;
	343	}
956c9207 SR	344
	345	struct net *get_net_ns_by_fd(int fd)
	346	{
	347	return ERR_PTR(-EINVAL);
	348	}
6a1a3b9f PE	349	#endif
6a1a3b9f PE	350
30ffee84 JB	351	struct net *get_net_ns_by_pid(pid_t pid)
	352	{
	353	struct task_struct *tsk;
	354	struct net *net;
	355
	356	/* Lookup the network namespace */
	357	net = ERR_PTR(-ESRCH);
	358	rcu_read_lock();
	359	tsk = find_task_by_vpid(pid);
	360	if (tsk) {
	361	struct nsproxy *nsproxy;
	362	nsproxy = task_nsproxy(tsk);
	363	if (nsproxy)
	364	net = get_net(nsproxy->net_ns);
	365	}
	366	rcu_read_unlock();
	367	return net;
	368	}
	369	EXPORT_SYMBOL_GPL(get_net_ns_by_pid);
	370
5f256bec EB	371	static int __init net_ns_init(void)
5f256bec EB	372	{
486a87f1	373	struct net_generic *ng;
5f256bec	374
d57a9212	375	#ifdef CONFIG_NET_NS
5f256bec EB	376	net_cachep = kmem_cache_create("net_namespace", sizeof(struct net),
	377	SMP_CACHE_BYTES,
	378	SLAB_PANIC, NULL);
3ef1355d BT	379
	380	/* Create workqueue for cleanup */
	381	netns_wq = create_singlethread_workqueue("netns");
	382	if (!netns_wq)
	383	panic("Could not create netns workq");
d57a9212	384	#endif
3ef1355d	385
486a87f1 DL	386	ng = net_alloc_generic();
	387	if (!ng)
	388	panic("Could not allocate generic netns");
	389
	390	rcu_assign_pointer(init_net.gen, ng);
	391
5f256bec	392	mutex_lock(&net_mutex);
ca0f3112 SH	393	if (setup_net(&init_net))
ca0f3112 SH	394	panic("Could not setup the initial network namespace");
5f256bec	395
f4618d39	396	rtnl_lock();
11a28d37	397	list_add_tail_rcu(&init_net.list, &net_namespace_list);
f4618d39	398	rtnl_unlock();
5f256bec EB	399
5f256bec EB	400	mutex_unlock(&net_mutex);
5f256bec EB	401
	402	return 0;
	403	}
	404
	405	pure_initcall(net_ns_init);
	406
ed160e83	407	#ifdef CONFIG_NET_NS
f875bae0 EB	408	static int __register_pernet_operations(struct list_head *list,
f875bae0 EB	409	struct pernet_operations *ops)
5f256bec	410	{
72ad937a	411	struct net *net;
5f256bec	412	int error;
72ad937a	413	LIST_HEAD(net_exit_list);
5f256bec	414
5f256bec	415	list_add_tail(&ops->list, list);
f875bae0	416	if (ops->init \|\| (ops->id && ops->size)) {
1dba323b	417	for_each_net(net) {
f875bae0	418	error = ops_init(ops, net);
5f256bec EB	419	if (error)
5f256bec EB	420	goto out_undo;
72ad937a	421	list_add_tail(&net->exit_list, &net_exit_list);
5f256bec EB	422	}
5f256bec EB	423	}
1dba323b	424	return 0;
5f256bec EB	425
	426	out_undo:
	427	/* If I have an error cleanup all namespaces I initialized */
	428	list_del(&ops->list);
72ad937a EB	429	ops_exit_list(ops, &net_exit_list);
72ad937a EB	430	ops_free_list(ops, &net_exit_list);
1dba323b	431	return error;
5f256bec EB	432	}
5f256bec EB	433
f875bae0	434	static void __unregister_pernet_operations(struct pernet_operations *ops)
5f256bec EB	435	{
5f256bec EB	436	struct net *net;
72ad937a	437	LIST_HEAD(net_exit_list);
5f256bec EB	438
5f256bec EB	439	list_del(&ops->list);
72ad937a EB	440	for_each_net(net)
	441	list_add_tail(&net->exit_list, &net_exit_list);
	442	ops_exit_list(ops, &net_exit_list);
	443	ops_free_list(ops, &net_exit_list);
5f256bec EB	444	}
5f256bec EB	445
ed160e83 DL	446	#else
ed160e83 DL	447
f875bae0 EB	448	static int __register_pernet_operations(struct list_head *list,
f875bae0 EB	449	struct pernet_operations *ops)
ed160e83	450	{
f875bae0 EB	451	int err = 0;
	452	err = ops_init(ops, &init_net);
	453	if (err)
	454	ops_free(ops, &init_net);
	455	return err;
	456
ed160e83 DL	457	}
ed160e83 DL	458
f875bae0	459	static void __unregister_pernet_operations(struct pernet_operations *ops)
ed160e83	460	{
72ad937a EB	461	LIST_HEAD(net_exit_list);
	462	list_add(&init_net.exit_list, &net_exit_list);
	463	ops_exit_list(ops, &net_exit_list);
	464	ops_free_list(ops, &net_exit_list);
ed160e83	465	}
f875bae0 EB	466
f875bae0 EB	467	#endif /* CONFIG_NET_NS */
ed160e83	468
c93cf61f PE	469	static DEFINE_IDA(net_generic_ids);
c93cf61f PE	470
f875bae0 EB	471	static int register_pernet_operations(struct list_head *list,
	472	struct pernet_operations *ops)
	473	{
	474	int error;
	475
	476	if (ops->id) {
	477	again:
	478	error = ida_get_new_above(&net_generic_ids, 1, ops->id);
	479	if (error < 0) {
	480	if (error == -EAGAIN) {
	481	ida_pre_get(&net_generic_ids, GFP_KERNEL);
	482	goto again;
	483	}
	484	return error;
	485	}
	486	}
	487	error = __register_pernet_operations(list, ops);
3a765eda EB	488	if (error) {
	489	rcu_barrier();
	490	if (ops->id)
	491	ida_remove(&net_generic_ids, *ops->id);
	492	}
f875bae0 EB	493
	494	return error;
	495	}
	496
	497	static void unregister_pernet_operations(struct pernet_operations *ops)
	498	{
	499
	500	__unregister_pernet_operations(ops);
3a765eda	501	rcu_barrier();
f875bae0 EB	502	if (ops->id)
	503	ida_remove(&net_generic_ids, *ops->id);
	504	}
	505
5f256bec EB	506	/**
	507	* register_pernet_subsys - register a network namespace subsystem
	508	* @ops: pernet operations structure for the subsystem
	509	*
	510	* Register a subsystem which has init and exit functions
	511	* that are called when network namespaces are created and
	512	* destroyed respectively.
	513	*
	514	* When registered all network namespace init functions are
	515	* called for every existing network namespace. Allowing kernel
	516	* modules to have a race free view of the set of network namespaces.
	517	*
	518	* When a new network namespace is created all of the init
	519	* methods are called in the order in which they were registered.
	520	*
	521	* When a network namespace is destroyed all of the exit methods
	522	* are called in the reverse of the order with which they were
	523	* registered.
	524	*/
	525	int register_pernet_subsys(struct pernet_operations *ops)
	526	{
	527	int error;
	528	mutex_lock(&net_mutex);
	529	error = register_pernet_operations(first_device, ops);
	530	mutex_unlock(&net_mutex);
	531	return error;
	532	}
	533	EXPORT_SYMBOL_GPL(register_pernet_subsys);
	534
	535	/**
	536	* unregister_pernet_subsys - unregister a network namespace subsystem
	537	* @ops: pernet operations structure to manipulate
	538	*
	539	* Remove the pernet operations structure from the list to be
53379e57	540	* used when network namespaces are created or destroyed. In
5f256bec EB	541	* addition run the exit method for all existing network
	542	* namespaces.
	543	*/
b3c981d2	544	void unregister_pernet_subsys(struct pernet_operations *ops)
5f256bec EB	545	{
5f256bec EB	546	mutex_lock(&net_mutex);
b3c981d2	547	unregister_pernet_operations(ops);
5f256bec EB	548	mutex_unlock(&net_mutex);
	549	}
	550	EXPORT_SYMBOL_GPL(unregister_pernet_subsys);
	551
	552	/**
	553	* register_pernet_device - register a network namespace device
	554	* @ops: pernet operations structure for the subsystem
	555	*
	556	* Register a device which has init and exit functions
	557	* that are called when network namespaces are created and
	558	* destroyed respectively.
	559	*
	560	* When registered all network namespace init functions are
	561	* called for every existing network namespace. Allowing kernel
	562	* modules to have a race free view of the set of network namespaces.
	563	*
	564	* When a new network namespace is created all of the init
	565	* methods are called in the order in which they were registered.
	566	*
	567	* When a network namespace is destroyed all of the exit methods
	568	* are called in the reverse of the order with which they were
	569	* registered.
	570	*/
	571	int register_pernet_device(struct pernet_operations *ops)
	572	{
	573	int error;
	574	mutex_lock(&net_mutex);
	575	error = register_pernet_operations(&pernet_list, ops);
	576	if (!error && (first_device == &pernet_list))
	577	first_device = &ops->list;
	578	mutex_unlock(&net_mutex);
	579	return error;
	580	}
	581	EXPORT_SYMBOL_GPL(register_pernet_device);
	582
	583	/**
	584	* unregister_pernet_device - unregister a network namespace netdevice
	585	* @ops: pernet operations structure to manipulate
	586	*
	587	* Remove the pernet operations structure from the list to be
53379e57	588	* used when network namespaces are created or destroyed. In
5f256bec EB	589	* addition run the exit method for all existing network
	590	* namespaces.
	591	*/
	592	void unregister_pernet_device(struct pernet_operations *ops)
	593	{
	594	mutex_lock(&net_mutex);
	595	if (&ops->list == first_device)
	596	first_device = first_device->next;
	597	unregister_pernet_operations(ops);
	598	mutex_unlock(&net_mutex);
	599	}
	600	EXPORT_SYMBOL_GPL(unregister_pernet_device);
13b6f576 EB	601
	602	#ifdef CONFIG_NET_NS
	603	static void netns_get(struct task_struct task)
	604	{
f0630529 EB	605	struct net *net = NULL;
	606	struct nsproxy *nsproxy;
	607
13b6f576	608	rcu_read_lock();
f0630529 EB	609	nsproxy = task_nsproxy(task);
	610	if (nsproxy)
	611	net = get_net(nsproxy->net_ns);
13b6f576	612	rcu_read_unlock();
f0630529	613
13b6f576 EB	614	return net;
	615	}
	616
	617	static void netns_put(void *ns)
	618	{
	619	put_net(ns);
	620	}
	621
	622	static int netns_install(struct nsproxy nsproxy, void ns)
	623	{
	624	put_net(nsproxy->net_ns);
	625	nsproxy->net_ns = get_net(ns);
	626	return 0;
	627	}
	628
	629	const struct proc_ns_operations netns_operations = {
	630	.name = "net",
	631	.type = CLONE_NEWNET,
	632	.get = netns_get,
	633	.put = netns_put,
	634	.install = netns_install,
	635	};
	636	#endif