[deliverable/linux.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 void net_generic_release(struct rcu_head *rcu)
{
	struct net_generic *ng;

	ng = container_of(rcu, struct net_generic, rcu);
	kfree(ng);
}

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);
	call_rcu(&old_ng->rcu, net_generic_release);
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);
}

static struct net *net_create(void)
{
	struct net *net;
	int rv;

	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;
}

struct net *copy_net_ns(unsigned long flags, struct net *old_net)
{
	if (!(flags & CLONE_NEWNET))
		return get_net(old_net);
	return net_create();
}

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);

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

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;
}

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