[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/netdevice.h>
#include <net/net_namespace.h>
#include <net/netns/generic.h>
#include <net/rtnetlink.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 unregister_netdevices(struct net *net, struct list_head *list)
{
	struct net_device *dev;
	/* At exit all network devices most be removed from a network
	 * namespace.  Do this in the reverse order of registeration.
	 */
	for_each_netdev_reverse(net, dev) {
		if (dev->rtnl_link_ops)
			dev->rtnl_link_ops->dellink(dev, list);
		else
			unregister_netdevice_queue(dev, list);
	}
}

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->list == first_device) {
		LIST_HEAD(dev_kill_list);
		rtnl_lock();
		list_for_each_entry(net, net_exit_list, exit_list)
			unregister_netdevices(net, &dev_kill_list);
		unregister_netdevice_many(&dev_kill_list);
		rtnl_unlock();
	}
	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);

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 && ops->id)
		ida_remove(&net_generic_ids, *ops->id);

	return error;
}

static void unregister_pernet_operations(struct pernet_operations *ops)
{
	
	__unregister_pernet_operations(ops);
	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 *module)
{
	mutex_lock(&net_mutex);
	unregister_pernet_operations(module);
	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);

static void net_generic_release(struct rcu_head *rcu)
{
	struct net_generic *ng;

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

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);

	ng = old_ng = net->gen;
	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;
}
EXPORT_SYMBOL_GPL(net_assign_generic);
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>
2b035b39	11	#include <linux/netdevice.h>
5f256bec	12	#include <net/net_namespace.h>
dec827d1	13	#include <net/netns/generic.h>
2b035b39	14	#include <net/rtnetlink.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
2b035b39 EB	32	static void unregister_netdevices(struct net net, struct list_head list)
	33	{
	34	struct net_device *dev;
	35	/* At exit all network devices most be removed from a network
	36	* namespace. Do this in the reverse order of registeration.
	37	*/
	38	for_each_netdev_reverse(net, dev) {
	39	if (dev->rtnl_link_ops)
	40	dev->rtnl_link_ops->dellink(dev, list);
	41	else
	42	unregister_netdevice_queue(dev, list);
	43	}
	44	}
	45
f875bae0 EB	46	static int ops_init(const struct pernet_operations ops, struct net net)
	47	{
	48	int err;
	49	if (ops->id && ops->size) {
	50	void *data = kzalloc(ops->size, GFP_KERNEL);
	51	if (!data)
	52	return -ENOMEM;
	53
	54	err = net_assign_generic(net, *ops->id, data);
	55	if (err) {
	56	kfree(data);
	57	return err;
	58	}
	59	}
	60	if (ops->init)
	61	return ops->init(net);
	62	return 0;
	63	}
	64
	65	static void ops_free(const struct pernet_operations ops, struct net net)
	66	{
	67	if (ops->id && ops->size) {
	68	int id = *ops->id;
	69	kfree(net_generic(net, id));
	70	}
	71	}
	72
72ad937a EB	73	static void ops_exit_list(const struct pernet_operations *ops,
	74	struct list_head *net_exit_list)
	75	{
	76	struct net *net;
	77	if (ops->exit) {
	78	list_for_each_entry(net, net_exit_list, exit_list)
	79	ops->exit(net);
	80	}
	81	if (&ops->list == first_device) {
	82	LIST_HEAD(dev_kill_list);
	83	rtnl_lock();
	84	list_for_each_entry(net, net_exit_list, exit_list)
	85	unregister_netdevices(net, &dev_kill_list);
	86	unregister_netdevice_many(&dev_kill_list);
	87	rtnl_unlock();
	88	}
	89	if (ops->exit_batch)
	90	ops->exit_batch(net_exit_list);
	91	}
	92
	93	static void ops_free_list(const struct pernet_operations *ops,
	94	struct list_head *net_exit_list)
	95	{
	96	struct net *net;
	97	if (ops->size && ops->id) {
	98	list_for_each_entry(net, net_exit_list, exit_list)
	99	ops_free(ops, net);
	100	}
	101	}
	102
5f256bec EB	103	/*
	104	* setup_net runs the initializers for the network namespace object.
	105	*/
1a2ee93d	106	static __net_init int setup_net(struct net *net)
5f256bec EB	107	{
5f256bec EB	108	/* Must be called with net_mutex held */
f875bae0	109	const struct pernet_operations ops, saved_ops;
486a87f1	110	int error = 0;
72ad937a	111	LIST_HEAD(net_exit_list);
5f256bec	112
5f256bec	113	atomic_set(&net->count, 1);
486a87f1	114
5d1e4468	115	#ifdef NETNS_REFCNT_DEBUG
5f256bec	116	atomic_set(&net->use_count, 0);
5d1e4468	117	#endif
5f256bec	118
768f3591	119	list_for_each_entry(ops, &pernet_list, list) {
f875bae0 EB	120	error = ops_init(ops, net);
	121	if (error < 0)
	122	goto out_undo;
5f256bec EB	123	}
	124	out:
	125	return error;
768f3591	126
5f256bec EB	127	out_undo:
	128	/* Walk through the list backwards calling the exit functions
	129	* for the pernet modules whose init functions did not fail.
	130	*/
72ad937a	131	list_add(&net->exit_list, &net_exit_list);
f875bae0	132	saved_ops = ops;
72ad937a EB	133	list_for_each_entry_continue_reverse(ops, &pernet_list, list)
	134	ops_exit_list(ops, &net_exit_list);
	135
f875bae0 EB	136	ops = saved_ops;
f875bae0 EB	137	list_for_each_entry_continue_reverse(ops, &pernet_list, list)
72ad937a	138	ops_free_list(ops, &net_exit_list);
310928d9 DL	139
310928d9 DL	140	rcu_barrier();
5f256bec EB	141	goto out;
	142	}
	143
486a87f1	144	static struct net_generic *net_alloc_generic(void)
6a1a3b9f	145	{
486a87f1 DL	146	struct net_generic *ng;
	147	size_t generic_size = sizeof(struct net_generic) +
	148	INITIAL_NET_GEN_PTRS * sizeof(void *);
	149
	150	ng = kzalloc(generic_size, GFP_KERNEL);
	151	if (ng)
	152	ng->len = INITIAL_NET_GEN_PTRS;
	153
	154	return ng;
6a1a3b9f PE	155	}
6a1a3b9f PE	156
ebe47d47 CN	157	#ifdef CONFIG_NET_NS
	158	static struct kmem_cache *net_cachep;
	159	static struct workqueue_struct *netns_wq;
	160
486a87f1	161	static struct net *net_alloc(void)
45a19b0a	162	{
486a87f1 DL	163	struct net *net = NULL;
	164	struct net_generic *ng;
	165
	166	ng = net_alloc_generic();
	167	if (!ng)
	168	goto out;
	169
	170	net = kmem_cache_zalloc(net_cachep, GFP_KERNEL);
45a19b0a	171	if (!net)
486a87f1	172	goto out_free;
45a19b0a	173
486a87f1 DL	174	rcu_assign_pointer(net->gen, ng);
	175	out:
	176	return net;
	177
	178	out_free:
	179	kfree(ng);
	180	goto out;
	181	}
	182
	183	static void net_free(struct net *net)
	184	{
5d1e4468	185	#ifdef NETNS_REFCNT_DEBUG
45a19b0a JFS	186	if (unlikely(atomic_read(&net->use_count) != 0)) {
	187	printk(KERN_EMERG "network namespace not free! Usage: %d\n",
	188	atomic_read(&net->use_count));
	189	return;
	190	}
5d1e4468	191	#endif
4ef079cc	192	kfree(net->gen);
45a19b0a JFS	193	kmem_cache_free(net_cachep, net);
	194	}
	195
088eb2d9	196	static struct net *net_create(void)
9dd776b6	197	{
088eb2d9 AD	198	struct net *net;
088eb2d9 AD	199	int rv;
9dd776b6	200
088eb2d9 AD	201	net = net_alloc();
	202	if (!net)
	203	return ERR_PTR(-ENOMEM);
9dd776b6	204	mutex_lock(&net_mutex);
088eb2d9 AD	205	rv = setup_net(net);
088eb2d9 AD	206	if (rv == 0) {
486a87f1	207	rtnl_lock();
11a28d37	208	list_add_tail_rcu(&net->list, &net_namespace_list);
486a87f1 DL	209	rtnl_unlock();
486a87f1 DL	210	}
9dd776b6	211	mutex_unlock(&net_mutex);
088eb2d9 AD	212	if (rv < 0) {
	213	net_free(net);
	214	return ERR_PTR(rv);
	215	}
	216	return net;
	217	}
486a87f1	218
088eb2d9 AD	219	struct net copy_net_ns(unsigned long flags, struct net old_net)
	220	{
	221	if (!(flags & CLONE_NEWNET))
	222	return get_net(old_net);
	223	return net_create();
9dd776b6 EB	224	}
9dd776b6 EB	225
2b035b39 EB	226	static DEFINE_SPINLOCK(cleanup_list_lock);
	227	static LIST_HEAD(cleanup_list); /* Must hold cleanup_list_lock to touch */
	228
6a1a3b9f PE	229	static void cleanup_net(struct work_struct *work)
6a1a3b9f PE	230	{
f875bae0	231	const struct pernet_operations *ops;
2b035b39 EB	232	struct net net, tmp;
2b035b39 EB	233	LIST_HEAD(net_kill_list);
72ad937a	234	LIST_HEAD(net_exit_list);
6a1a3b9f	235
2b035b39 EB	236	/* Atomically snapshot the list of namespaces to cleanup */
	237	spin_lock_irq(&cleanup_list_lock);
	238	list_replace_init(&cleanup_list, &net_kill_list);
	239	spin_unlock_irq(&cleanup_list_lock);
6a1a3b9f PE	240
	241	mutex_lock(&net_mutex);
	242
	243	/* Don't let anyone else find us. */
	244	rtnl_lock();
72ad937a	245	list_for_each_entry(net, &net_kill_list, cleanup_list) {
2b035b39	246	list_del_rcu(&net->list);
72ad937a EB	247	list_add_tail(&net->exit_list, &net_exit_list);
72ad937a EB	248	}
6a1a3b9f PE	249	rtnl_unlock();
6a1a3b9f PE	250
11a28d37 JB	251	/*
	252	* Another CPU might be rcu-iterating the list, wait for it.
	253	* This needs to be before calling the exit() notifiers, so
	254	* the rcu_barrier() below isn't sufficient alone.
	255	*/
	256	synchronize_rcu();
	257
6a1a3b9f	258	/* Run all of the network namespace exit methods */
72ad937a EB	259	list_for_each_entry_reverse(ops, &pernet_list, list)
	260	ops_exit_list(ops, &net_exit_list);
	261
f875bae0	262	/* Free the net generic variables */
72ad937a EB	263	list_for_each_entry_reverse(ops, &pernet_list, list)
72ad937a EB	264	ops_free_list(ops, &net_exit_list);
6a1a3b9f PE	265
	266	mutex_unlock(&net_mutex);
	267
	268	/* Ensure there are no outstanding rcu callbacks using this
	269	* network namespace.
	270	*/
	271	rcu_barrier();
	272
	273	/* Finally it is safe to free my network namespace structure */
72ad937a EB	274	list_for_each_entry_safe(net, tmp, &net_exit_list, exit_list) {
72ad937a EB	275	list_del_init(&net->exit_list);
2b035b39 EB	276	net_free(net);
2b035b39 EB	277	}
6a1a3b9f	278	}
2b035b39	279	static DECLARE_WORK(net_cleanup_work, cleanup_net);
6a1a3b9f PE	280
	281	void __put_net(struct net *net)
	282	{
	283	/* Cleanup the network namespace in process context */
2b035b39 EB	284	unsigned long flags;
	285
	286	spin_lock_irqsave(&cleanup_list_lock, flags);
	287	list_add(&net->cleanup_list, &cleanup_list);
	288	spin_unlock_irqrestore(&cleanup_list_lock, flags);
	289
	290	queue_work(netns_wq, &net_cleanup_work);
6a1a3b9f PE	291	}
	292	EXPORT_SYMBOL_GPL(__put_net);
	293
	294	#else
	295	struct net copy_net_ns(unsigned long flags, struct net old_net)
	296	{
	297	if (flags & CLONE_NEWNET)
	298	return ERR_PTR(-EINVAL);
	299	return old_net;
	300	}
	301	#endif
	302
30ffee84 JB	303	struct net *get_net_ns_by_pid(pid_t pid)
	304	{
	305	struct task_struct *tsk;
	306	struct net *net;
	307
	308	/* Lookup the network namespace */
	309	net = ERR_PTR(-ESRCH);
	310	rcu_read_lock();
	311	tsk = find_task_by_vpid(pid);
	312	if (tsk) {
	313	struct nsproxy *nsproxy;
	314	nsproxy = task_nsproxy(tsk);
	315	if (nsproxy)
	316	net = get_net(nsproxy->net_ns);
	317	}
	318	rcu_read_unlock();
	319	return net;
	320	}
	321	EXPORT_SYMBOL_GPL(get_net_ns_by_pid);
	322
5f256bec EB	323	static int __init net_ns_init(void)
5f256bec EB	324	{
486a87f1	325	struct net_generic *ng;
5f256bec	326
d57a9212	327	#ifdef CONFIG_NET_NS
5f256bec EB	328	net_cachep = kmem_cache_create("net_namespace", sizeof(struct net),
	329	SMP_CACHE_BYTES,
	330	SLAB_PANIC, NULL);
3ef1355d BT	331
	332	/* Create workqueue for cleanup */
	333	netns_wq = create_singlethread_workqueue("netns");
	334	if (!netns_wq)
	335	panic("Could not create netns workq");
d57a9212	336	#endif
3ef1355d	337
486a87f1 DL	338	ng = net_alloc_generic();
	339	if (!ng)
	340	panic("Could not allocate generic netns");
	341
	342	rcu_assign_pointer(init_net.gen, ng);
	343
5f256bec	344	mutex_lock(&net_mutex);
ca0f3112 SH	345	if (setup_net(&init_net))
ca0f3112 SH	346	panic("Could not setup the initial network namespace");
5f256bec	347
f4618d39	348	rtnl_lock();
11a28d37	349	list_add_tail_rcu(&init_net.list, &net_namespace_list);
f4618d39	350	rtnl_unlock();
5f256bec EB	351
5f256bec EB	352	mutex_unlock(&net_mutex);
5f256bec EB	353
	354	return 0;
	355	}
	356
	357	pure_initcall(net_ns_init);
	358
ed160e83	359	#ifdef CONFIG_NET_NS
f875bae0 EB	360	static int __register_pernet_operations(struct list_head *list,
f875bae0 EB	361	struct pernet_operations *ops)
5f256bec	362	{
72ad937a	363	struct net *net;
5f256bec	364	int error;
72ad937a	365	LIST_HEAD(net_exit_list);
5f256bec	366
5f256bec	367	list_add_tail(&ops->list, list);
f875bae0	368	if (ops->init \|\| (ops->id && ops->size)) {
1dba323b	369	for_each_net(net) {
f875bae0	370	error = ops_init(ops, net);
5f256bec EB	371	if (error)
5f256bec EB	372	goto out_undo;
72ad937a	373	list_add_tail(&net->exit_list, &net_exit_list);
5f256bec EB	374	}
5f256bec EB	375	}
1dba323b	376	return 0;
5f256bec EB	377
	378	out_undo:
	379	/* If I have an error cleanup all namespaces I initialized */
	380	list_del(&ops->list);
72ad937a EB	381	ops_exit_list(ops, &net_exit_list);
72ad937a EB	382	ops_free_list(ops, &net_exit_list);
1dba323b	383	return error;
5f256bec EB	384	}
5f256bec EB	385
f875bae0	386	static void __unregister_pernet_operations(struct pernet_operations *ops)
5f256bec EB	387	{
5f256bec EB	388	struct net *net;
72ad937a	389	LIST_HEAD(net_exit_list);
5f256bec EB	390
5f256bec EB	391	list_del(&ops->list);
72ad937a EB	392	for_each_net(net)
	393	list_add_tail(&net->exit_list, &net_exit_list);
	394	ops_exit_list(ops, &net_exit_list);
	395	ops_free_list(ops, &net_exit_list);
5f256bec EB	396	}
5f256bec EB	397
ed160e83 DL	398	#else
ed160e83 DL	399
f875bae0 EB	400	static int __register_pernet_operations(struct list_head *list,
f875bae0 EB	401	struct pernet_operations *ops)
ed160e83	402	{
f875bae0 EB	403	int err = 0;
	404	err = ops_init(ops, &init_net);
	405	if (err)
	406	ops_free(ops, &init_net);
	407	return err;
	408
ed160e83 DL	409	}
ed160e83 DL	410
f875bae0	411	static void __unregister_pernet_operations(struct pernet_operations *ops)
ed160e83	412	{
72ad937a EB	413	LIST_HEAD(net_exit_list);
	414	list_add(&init_net.exit_list, &net_exit_list);
	415	ops_exit_list(ops, &net_exit_list);
	416	ops_free_list(ops, &net_exit_list);
ed160e83	417	}
f875bae0 EB	418
f875bae0 EB	419	#endif /* CONFIG_NET_NS */
ed160e83	420
c93cf61f PE	421	static DEFINE_IDA(net_generic_ids);
c93cf61f PE	422
f875bae0 EB	423	static int register_pernet_operations(struct list_head *list,
	424	struct pernet_operations *ops)
	425	{
	426	int error;
	427
	428	if (ops->id) {
	429	again:
	430	error = ida_get_new_above(&net_generic_ids, 1, ops->id);
	431	if (error < 0) {
	432	if (error == -EAGAIN) {
	433	ida_pre_get(&net_generic_ids, GFP_KERNEL);
	434	goto again;
	435	}
	436	return error;
	437	}
	438	}
	439	error = __register_pernet_operations(list, ops);
	440	if (error && ops->id)
	441	ida_remove(&net_generic_ids, *ops->id);
	442
	443	return error;
	444	}
	445
	446	static void unregister_pernet_operations(struct pernet_operations *ops)
	447	{
	448
	449	__unregister_pernet_operations(ops);
	450	if (ops->id)
	451	ida_remove(&net_generic_ids, *ops->id);
	452	}
	453
5f256bec EB	454	/**
	455	* register_pernet_subsys - register a network namespace subsystem
	456	* @ops: pernet operations structure for the subsystem
	457	*
	458	* Register a subsystem which has init and exit functions
	459	* that are called when network namespaces are created and
	460	* destroyed respectively.
	461	*
	462	* When registered all network namespace init functions are
	463	* called for every existing network namespace. Allowing kernel
	464	* modules to have a race free view of the set of network namespaces.
	465	*
	466	* When a new network namespace is created all of the init
	467	* methods are called in the order in which they were registered.
	468	*
	469	* When a network namespace is destroyed all of the exit methods
	470	* are called in the reverse of the order with which they were
	471	* registered.
	472	*/
	473	int register_pernet_subsys(struct pernet_operations *ops)
	474	{
	475	int error;
	476	mutex_lock(&net_mutex);
	477	error = register_pernet_operations(first_device, ops);
	478	mutex_unlock(&net_mutex);
	479	return error;
	480	}
	481	EXPORT_SYMBOL_GPL(register_pernet_subsys);
	482
	483	/**
	484	* unregister_pernet_subsys - unregister a network namespace subsystem
	485	* @ops: pernet operations structure to manipulate
	486	*
	487	* Remove the pernet operations structure from the list to be
53379e57	488	* used when network namespaces are created or destroyed. In
5f256bec EB	489	* addition run the exit method for all existing network
	490	* namespaces.
	491	*/
	492	void unregister_pernet_subsys(struct pernet_operations *module)
	493	{
	494	mutex_lock(&net_mutex);
	495	unregister_pernet_operations(module);
	496	mutex_unlock(&net_mutex);
	497	}
	498	EXPORT_SYMBOL_GPL(unregister_pernet_subsys);
	499
	500	/**
	501	* register_pernet_device - register a network namespace device
	502	* @ops: pernet operations structure for the subsystem
	503	*
	504	* Register a device which has init and exit functions
	505	* that are called when network namespaces are created and
	506	* destroyed respectively.
	507	*
	508	* When registered all network namespace init functions are
	509	* called for every existing network namespace. Allowing kernel
	510	* modules to have a race free view of the set of network namespaces.
	511	*
	512	* When a new network namespace is created all of the init
	513	* methods are called in the order in which they were registered.
	514	*
	515	* When a network namespace is destroyed all of the exit methods
	516	* are called in the reverse of the order with which they were
	517	* registered.
	518	*/
	519	int register_pernet_device(struct pernet_operations *ops)
	520	{
	521	int error;
	522	mutex_lock(&net_mutex);
	523	error = register_pernet_operations(&pernet_list, ops);
	524	if (!error && (first_device == &pernet_list))
	525	first_device = &ops->list;
	526	mutex_unlock(&net_mutex);
	527	return error;
	528	}
	529	EXPORT_SYMBOL_GPL(register_pernet_device);
	530
	531	/**
	532	* unregister_pernet_device - unregister a network namespace netdevice
	533	* @ops: pernet operations structure to manipulate
	534	*
	535	* Remove the pernet operations structure from the list to be
53379e57	536	* used when network namespaces are created or destroyed. In
5f256bec EB	537	* addition run the exit method for all existing network
	538	* namespaces.
	539	*/
	540	void unregister_pernet_device(struct pernet_operations *ops)
	541	{
	542	mutex_lock(&net_mutex);
	543	if (&ops->list == first_device)
	544	first_device = first_device->next;
	545	unregister_pernet_operations(ops);
	546	mutex_unlock(&net_mutex);
	547	}
	548	EXPORT_SYMBOL_GPL(unregister_pernet_device);
c93cf61f	549
dec827d1 PE	550	static void net_generic_release(struct rcu_head *rcu)
	551	{
	552	struct net_generic *ng;
	553
	554	ng = container_of(rcu, struct net_generic, rcu);
	555	kfree(ng);
	556	}
	557
	558	int net_assign_generic(struct net net, int id, void data)
	559	{
	560	struct net_generic ng, old_ng;
	561
	562	BUG_ON(!mutex_is_locked(&net_mutex));
	563	BUG_ON(id == 0);
	564
	565	ng = old_ng = net->gen;
	566	if (old_ng->len >= id)
	567	goto assign;
	568
	569	ng = kzalloc(sizeof(struct net_generic) +
	570	id * sizeof(void *), GFP_KERNEL);
	571	if (ng == NULL)
	572	return -ENOMEM;
	573
	574	/*
	575	* Some synchronisation notes:
	576	*
	577	* The net_generic explores the net->gen array inside rcu
	578	* read section. Besides once set the net->gen->ptr[x]
	579	* pointer never changes (see rules in netns/generic.h).
	580	*
	581	* That said, we simply duplicate this array and schedule
	582	* the old copy for kfree after a grace period.
	583	*/
	584
	585	ng->len = id;
14458630	586	memcpy(&ng->ptr, &old_ng->ptr, old_ng->len * sizeof(void*));
dec827d1 PE	587
	588	rcu_assign_pointer(net->gen, ng);
	589	call_rcu(&old_ng->rcu, net_generic_release);
	590	assign:
	591	ng->ptr[id - 1] = data;
	592	return 0;
	593	}
	594	EXPORT_SYMBOL_GPL(net_assign_generic);