[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 <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);

#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);
}

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_free(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_free(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;

	net = ERR_PTR(-EINVAL);
	file = proc_ns_fget(fd);
	if (!file)
		goto out;

	ei = PROC_I(file->f_dentry->d_inode);
	if (ei->ns_ops != &netns_operations)
		goto out;

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