[mirror_ubuntu-bionic-kernel.git] / net / core / dst.c

/*
 * net/core/dst.c	Protocol independent destination cache.
 *
 * Authors:		Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
 *
 */

#include <linux/bitops.h>
#include <linux/errno.h>
#include <linux/init.h>
#include <linux/kernel.h>
#include <linux/workqueue.h>
#include <linux/mm.h>
#include <linux/module.h>
#include <linux/slab.h>
#include <linux/netdevice.h>
#include <linux/skbuff.h>
#include <linux/string.h>
#include <linux/types.h>
#include <net/net_namespace.h>
#include <linux/sched.h>
#include <linux/prefetch.h>
#include <net/lwtunnel.h>

#include <net/dst.h>
#include <net/dst_metadata.h>

/*
 * Theory of operations:
 * 1) We use a list, protected by a spinlock, to add
 *    new entries from both BH and non-BH context.
 * 2) In order to keep spinlock held for a small delay,
 *    we use a second list where are stored long lived
 *    entries, that are handled by the garbage collect thread
 *    fired by a workqueue.
 * 3) This list is guarded by a mutex,
 *    so that the gc_task and dst_dev_event() can be synchronized.
 */

/*
 * We want to keep lock & list close together
 * to dirty as few cache lines as possible in __dst_free().
 * As this is not a very strong hint, we dont force an alignment on SMP.
 */
int dst_discard_out(struct net *net, struct sock *sk, struct sk_buff *skb)
{
	kfree_skb(skb);
	return 0;
}
EXPORT_SYMBOL(dst_discard_out);

const struct dst_metrics dst_default_metrics = {
	/* This initializer is needed to force linker to place this variable
	 * into const section. Otherwise it might end into bss section.
	 * We really want to avoid false sharing on this variable, and catch
	 * any writes on it.
	 */
	.refcnt = REFCOUNT_INIT(1),
};

void dst_init(struct dst_entry *dst, struct dst_ops *ops,
	      struct net_device *dev, int initial_ref, int initial_obsolete,
	      unsigned short flags)
{
	dst->child = NULL;
	dst->dev = dev;
	if (dev)
		dev_hold(dev);
	dst->ops = ops;
	dst_init_metrics(dst, dst_default_metrics.metrics, true);
	dst->expires = 0UL;
	dst->path = dst;
	dst->from = NULL;
#ifdef CONFIG_XFRM
	dst->xfrm = NULL;
#endif
	dst->input = dst_discard;
	dst->output = dst_discard_out;
	dst->error = 0;
	dst->obsolete = initial_obsolete;
	dst->header_len = 0;
	dst->trailer_len = 0;
#ifdef CONFIG_IP_ROUTE_CLASSID
	dst->tclassid = 0;
#endif
	dst->lwtstate = NULL;
	atomic_set(&dst->__refcnt, initial_ref);
	dst->__use = 0;
	dst->lastuse = jiffies;
	dst->flags = flags;
	dst->next = NULL;
	if (!(flags & DST_NOCOUNT))
		dst_entries_add(ops, 1);
}
EXPORT_SYMBOL(dst_init);

void *dst_alloc(struct dst_ops *ops, struct net_device *dev,
		int initial_ref, int initial_obsolete, unsigned short flags)
{
	struct dst_entry *dst;

	if (ops->gc && dst_entries_get_fast(ops) > ops->gc_thresh) {
		if (ops->gc(ops))
			return NULL;
	}

	dst = kmem_cache_alloc(ops->kmem_cachep, GFP_ATOMIC);
	if (!dst)
		return NULL;

	dst_init(dst, ops, dev, initial_ref, initial_obsolete, flags);

	return dst;
}
EXPORT_SYMBOL(dst_alloc);

struct dst_entry *dst_destroy(struct dst_entry * dst)
{
	struct dst_entry *child;

	smp_rmb();

	child = dst->child;

	if (!(dst->flags & DST_NOCOUNT))
		dst_entries_add(dst->ops, -1);

	if (dst->ops->destroy)
		dst->ops->destroy(dst);
	if (dst->dev)
		dev_put(dst->dev);

	lwtstate_put(dst->lwtstate);

	if (dst->flags & DST_METADATA)
		metadata_dst_free((struct metadata_dst *)dst);
	else
		kmem_cache_free(dst->ops->kmem_cachep, dst);

	dst = child;
	if (dst)
		dst_release_immediate(dst);
	return NULL;
}
EXPORT_SYMBOL(dst_destroy);

static void dst_destroy_rcu(struct rcu_head *head)
{
	struct dst_entry *dst = container_of(head, struct dst_entry, rcu_head);

	dst = dst_destroy(dst);
}

/* Operations to mark dst as DEAD and clean up the net device referenced
 * by dst:
 * 1. put the dst under loopback interface and discard all tx/rx packets
 *    on this route.
 * 2. release the net_device
 * This function should be called when removing routes from the fib tree
 * in preparation for a NETDEV_DOWN/NETDEV_UNREGISTER event and also to
 * make the next dst_ops->check() fail.
 */
void dst_dev_put(struct dst_entry *dst)
{
	struct net_device *dev = dst->dev;

	dst->obsolete = DST_OBSOLETE_DEAD;
	if (dst->ops->ifdown)
		dst->ops->ifdown(dst, dev, true);
	dst->input = dst_discard;
	dst->output = dst_discard_out;
	dst->dev = dev_net(dst->dev)->loopback_dev;
	dev_hold(dst->dev);
	dev_put(dev);
}
EXPORT_SYMBOL(dst_dev_put);

void dst_release(struct dst_entry *dst)
{
	if (dst) {
		int newrefcnt;

		newrefcnt = atomic_dec_return(&dst->__refcnt);
		if (unlikely(newrefcnt < 0))
			net_warn_ratelimited("%s: dst:%p refcnt:%d\n",
					     __func__, dst, newrefcnt);
		if (!newrefcnt)
			call_rcu(&dst->rcu_head, dst_destroy_rcu);
	}
}
EXPORT_SYMBOL(dst_release);

void dst_release_immediate(struct dst_entry *dst)
{
	if (dst) {
		int newrefcnt;

		newrefcnt = atomic_dec_return(&dst->__refcnt);
		if (unlikely(newrefcnt < 0))
			net_warn_ratelimited("%s: dst:%p refcnt:%d\n",
					     __func__, dst, newrefcnt);
		if (!newrefcnt)
			dst_destroy(dst);
	}
}
EXPORT_SYMBOL(dst_release_immediate);

u32 *dst_cow_metrics_generic(struct dst_entry *dst, unsigned long old)
{
	struct dst_metrics *p = kmalloc(sizeof(*p), GFP_ATOMIC);

	if (p) {
		struct dst_metrics *old_p = (struct dst_metrics *)__DST_METRICS_PTR(old);
		unsigned long prev, new;

		refcount_set(&p->refcnt, 1);
		memcpy(p->metrics, old_p->metrics, sizeof(p->metrics));

		new = (unsigned long) p;
		prev = cmpxchg(&dst->_metrics, old, new);

		if (prev != old) {
			kfree(p);
			p = (struct dst_metrics *)__DST_METRICS_PTR(prev);
			if (prev & DST_METRICS_READ_ONLY)
				p = NULL;
		} else if (prev & DST_METRICS_REFCOUNTED) {
			if (refcount_dec_and_test(&old_p->refcnt))
				kfree(old_p);
		}
	}
	BUILD_BUG_ON(offsetof(struct dst_metrics, metrics) != 0);
	return (u32 *)p;
}
EXPORT_SYMBOL(dst_cow_metrics_generic);

/* Caller asserts that dst_metrics_read_only(dst) is false.  */
void __dst_destroy_metrics_generic(struct dst_entry *dst, unsigned long old)
{
	unsigned long prev, new;

	new = ((unsigned long) &dst_default_metrics) | DST_METRICS_READ_ONLY;
	prev = cmpxchg(&dst->_metrics, old, new);
	if (prev == old)
		kfree(__DST_METRICS_PTR(old));
}
EXPORT_SYMBOL(__dst_destroy_metrics_generic);

static struct dst_ops md_dst_ops = {
	.family =		AF_UNSPEC,
};

static int dst_md_discard_out(struct net *net, struct sock *sk, struct sk_buff *skb)
{
	WARN_ONCE(1, "Attempting to call output on metadata dst\n");
	kfree_skb(skb);
	return 0;
}

static int dst_md_discard(struct sk_buff *skb)
{
	WARN_ONCE(1, "Attempting to call input on metadata dst\n");
	kfree_skb(skb);
	return 0;
}

static void __metadata_dst_init(struct metadata_dst *md_dst,
				enum metadata_type type, u8 optslen)

{
	struct dst_entry *dst;

	dst = &md_dst->dst;
	dst_init(dst, &md_dst_ops, NULL, 1, DST_OBSOLETE_NONE,
		 DST_METADATA | DST_NOCOUNT);

	dst->input = dst_md_discard;
	dst->output = dst_md_discard_out;

	memset(dst + 1, 0, sizeof(*md_dst) + optslen - sizeof(*dst));
	md_dst->type = type;
}

struct metadata_dst *metadata_dst_alloc(u8 optslen, enum metadata_type type,
					gfp_t flags)
{
	struct metadata_dst *md_dst;

	md_dst = kmalloc(sizeof(*md_dst) + optslen, flags);
	if (!md_dst)
		return NULL;

	__metadata_dst_init(md_dst, type, optslen);

	return md_dst;
}
EXPORT_SYMBOL_GPL(metadata_dst_alloc);

void metadata_dst_free(struct metadata_dst *md_dst)
{
#ifdef CONFIG_DST_CACHE
	if (md_dst->type == METADATA_IP_TUNNEL)
		dst_cache_destroy(&md_dst->u.tun_info.dst_cache);
#endif
	kfree(md_dst);
}

struct metadata_dst __percpu *
metadata_dst_alloc_percpu(u8 optslen, enum metadata_type type, gfp_t flags)
{
	int cpu;
	struct metadata_dst __percpu *md_dst;

	md_dst = __alloc_percpu_gfp(sizeof(struct metadata_dst) + optslen,
				    __alignof__(struct metadata_dst), flags);
	if (!md_dst)
		return NULL;

	for_each_possible_cpu(cpu)
		__metadata_dst_init(per_cpu_ptr(md_dst, cpu), type, optslen);

	return md_dst;
}
EXPORT_SYMBOL_GPL(metadata_dst_alloc_percpu);

void metadata_dst_free_percpu(struct metadata_dst __percpu *md_dst)
{
#ifdef CONFIG_DST_CACHE
	int cpu;

	for_each_possible_cpu(cpu) {
		struct metadata_dst *one_md_dst = per_cpu_ptr(md_dst, cpu);

		if (one_md_dst->type == METADATA_IP_TUNNEL)
			dst_cache_destroy(&one_md_dst->u.tun_info.dst_cache);
	}
#endif
	free_percpu(md_dst);
}
EXPORT_SYMBOL_GPL(metadata_dst_free_percpu);
Commit	Line	Data
1da177e4 LT	1	/*
	2	* net/core/dst.c Protocol independent destination cache.
	3	*
	4	* Authors: Alexey Kuznetsov, <kuznet@ms2.inr.ac.ru>
	5	*
	6	*/
	7
	8	#include <linux/bitops.h>
	9	#include <linux/errno.h>
	10	#include <linux/init.h>
	11	#include <linux/kernel.h>
86bba269	12	#include <linux/workqueue.h>
1da177e4 LT	13	#include <linux/mm.h>
1da177e4 LT	14	#include <linux/module.h>
5a0e3ad6	15	#include <linux/slab.h>
1da177e4	16	#include <linux/netdevice.h>
1da177e4 LT	17	#include <linux/skbuff.h>
	18	#include <linux/string.h>
	19	#include <linux/types.h>
e9dc8653	20	#include <net/net_namespace.h>
2fc1b5dd	21	#include <linux/sched.h>
268bb0ce	22	#include <linux/prefetch.h>
61adedf3	23	#include <net/lwtunnel.h>
1da177e4 LT	24
1da177e4 LT	25	#include <net/dst.h>
f38a9eb1	26	#include <net/dst_metadata.h>
1da177e4	27
86bba269 ED	28	/*
	29	* Theory of operations:
	30	* 1) We use a list, protected by a spinlock, to add
	31	* new entries from both BH and non-BH context.
	32	* 2) In order to keep spinlock held for a small delay,
	33	* we use a second list where are stored long lived
	34	* entries, that are handled by the garbage collect thread
	35	* fired by a workqueue.
	36	* 3) This list is guarded by a mutex,
	37	* so that the gc_task and dst_dev_event() can be synchronized.
1da177e4	38	*/
1da177e4	39
86bba269 ED	40	/*
	41	* We want to keep lock & list close together
	42	* to dirty as few cache lines as possible in __dst_free().
	43	* As this is not a very strong hint, we dont force an alignment on SMP.
	44	*/
ede2059d	45	int dst_discard_out(struct net net, struct sock sk, struct sk_buff *skb)
1da177e4 LT	46	{
	47	kfree_skb(skb);
	48	return 0;
	49	}
ede2059d	50	EXPORT_SYMBOL(dst_discard_out);
1da177e4	51
3fb07daf	52	const struct dst_metrics dst_default_metrics = {
a37e6e34 ED	53	/* This initializer is needed to force linker to place this variable
	54	* into const section. Otherwise it might end into bss section.
	55	* We really want to avoid false sharing on this variable, and catch
	56	* any writes on it.
	57	*/
9620fef2	58	.refcnt = REFCOUNT_INIT(1),
a37e6e34 ED	59	};
a37e6e34 ED	60
f38a9eb1 TG	61	void dst_init(struct dst_entry dst, struct dst_ops ops,
	62	struct net_device *dev, int initial_ref, int initial_obsolete,
	63	unsigned short flags)
1da177e4	64	{
cf911662	65	dst->child = NULL;
5c1e6aa3 DM	66	dst->dev = dev;
	67	if (dev)
	68	dev_hold(dev);
1da177e4	69	dst->ops = ops;
3fb07daf	70	dst_init_metrics(dst, dst_default_metrics.metrics, true);
cf911662	71	dst->expires = 0UL;
5c1e6aa3	72	dst->path = dst;
ecd98837	73	dst->from = NULL;
cf911662 DM	74	#ifdef CONFIG_XFRM
	75	dst->xfrm = NULL;
	76	#endif
5c1e6aa3	77	dst->input = dst_discard;
ede2059d	78	dst->output = dst_discard_out;
cf911662	79	dst->error = 0;
5c1e6aa3	80	dst->obsolete = initial_obsolete;
cf911662 DM	81	dst->header_len = 0;
	82	dst->trailer_len = 0;
	83	#ifdef CONFIG_IP_ROUTE_CLASSID
	84	dst->tclassid = 0;
1da177e4	85	#endif
61adedf3	86	dst->lwtstate = NULL;
5c1e6aa3	87	atomic_set(&dst->__refcnt, initial_ref);
cf911662	88	dst->__use = 0;
5c1e6aa3 DM	89	dst->lastuse = jiffies;
5c1e6aa3 DM	90	dst->flags = flags;
cf911662	91	dst->next = NULL;
957c665f DM	92	if (!(flags & DST_NOCOUNT))
957c665f DM	93	dst_entries_add(ops, 1);
f38a9eb1 TG	94	}
	95	EXPORT_SYMBOL(dst_init);
	96
	97	void dst_alloc(struct dst_ops ops, struct net_device *dev,
	98	int initial_ref, int initial_obsolete, unsigned short flags)
	99	{
	100	struct dst_entry *dst;
	101
	102	if (ops->gc && dst_entries_get_fast(ops) > ops->gc_thresh) {
	103	if (ops->gc(ops))
	104	return NULL;
	105	}
	106
	107	dst = kmem_cache_alloc(ops->kmem_cachep, GFP_ATOMIC);
	108	if (!dst)
	109	return NULL;
	110
	111	dst_init(dst, ops, dev, initial_ref, initial_obsolete, flags);
	112
1da177e4 LT	113	return dst;
1da177e4 LT	114	}
598ed936	115	EXPORT_SYMBOL(dst_alloc);
1da177e4	116
1da177e4 LT	117	struct dst_entry dst_destroy(struct dst_entry dst)
	118	{
	119	struct dst_entry *child;
1da177e4 LT	120
	121	smp_rmb();
	122
1da177e4 LT	123	child = dst->child;
1da177e4 LT	124
957c665f DM	125	if (!(dst->flags & DST_NOCOUNT))
957c665f DM	126	dst_entries_add(dst->ops, -1);
1da177e4 LT	127
	128	if (dst->ops->destroy)
	129	dst->ops->destroy(dst);
	130	if (dst->dev)
	131	dev_put(dst->dev);
f38a9eb1	132
e252b3d1 WC	133	lwtstate_put(dst->lwtstate);
e252b3d1 WC	134
f38a9eb1	135	if (dst->flags & DST_METADATA)
d71785ff	136	metadata_dst_free((struct metadata_dst *)dst);
f38a9eb1 TG	137	else
f38a9eb1 TG	138	kmem_cache_free(dst->ops->kmem_cachep, dst);
1da177e4 LT	139
1da177e4 LT	140	dst = child;
52df157f WW	141	if (dst)
52df157f WW	142	dst_release_immediate(dst);
1da177e4 LT	143	return NULL;
1da177e4 LT	144	}
598ed936	145	EXPORT_SYMBOL(dst_destroy);
1da177e4	146
f8864972 ED	147	static void dst_destroy_rcu(struct rcu_head *head)
	148	{
	149	struct dst_entry *dst = container_of(head, struct dst_entry, rcu_head);
	150
	151	dst = dst_destroy(dst);
f8864972 ED	152	}
f8864972 ED	153
4a6ce2b6 WW	154	/* Operations to mark dst as DEAD and clean up the net device referenced
	155	* by dst:
	156	* 1. put the dst under loopback interface and discard all tx/rx packets
	157	* on this route.
	158	* 2. release the net_device
	159	* This function should be called when removing routes from the fib tree
	160	* in preparation for a NETDEV_DOWN/NETDEV_UNREGISTER event and also to
	161	* make the next dst_ops->check() fail.
	162	*/
	163	void dst_dev_put(struct dst_entry *dst)
	164	{
	165	struct net_device *dev = dst->dev;
	166
	167	dst->obsolete = DST_OBSOLETE_DEAD;
	168	if (dst->ops->ifdown)
	169	dst->ops->ifdown(dst, dev, true);
	170	dst->input = dst_discard;
	171	dst->output = dst_discard_out;
	172	dst->dev = dev_net(dst->dev)->loopback_dev;
	173	dev_hold(dst->dev);
	174	dev_put(dev);
	175	}
	176	EXPORT_SYMBOL(dst_dev_put);
	177
8d330868 IJ	178	void dst_release(struct dst_entry *dst)
	179	{
	180	if (dst) {
598ed936	181	int newrefcnt;
ef711cf1	182
598ed936	183	newrefcnt = atomic_dec_return(&dst->__refcnt);
8bf4ada2 KK	184	if (unlikely(newrefcnt < 0))
	185	net_warn_ratelimited("%s: dst:%p refcnt:%d\n",
	186	__func__, dst, newrefcnt);
b2a9c0ed	187	if (!newrefcnt)
f8864972	188	call_rcu(&dst->rcu_head, dst_destroy_rcu);
8d330868 IJ	189	}
	190	}
	191	EXPORT_SYMBOL(dst_release);
	192
5f56f409 WW	193	void dst_release_immediate(struct dst_entry *dst)
	194	{
	195	if (dst) {
	196	int newrefcnt;
	197
	198	newrefcnt = atomic_dec_return(&dst->__refcnt);
	199	if (unlikely(newrefcnt < 0))
	200	net_warn_ratelimited("%s: dst:%p refcnt:%d\n",
	201	__func__, dst, newrefcnt);
	202	if (!newrefcnt)
	203	dst_destroy(dst);
	204	}
	205	}
	206	EXPORT_SYMBOL(dst_release_immediate);
	207
62fa8a84 DM	208	u32 dst_cow_metrics_generic(struct dst_entry dst, unsigned long old)
62fa8a84 DM	209	{
3fb07daf	210	struct dst_metrics p = kmalloc(sizeof(p), GFP_ATOMIC);
62fa8a84 DM	211
62fa8a84 DM	212	if (p) {
3fb07daf	213	struct dst_metrics old_p = (struct dst_metrics )__DST_METRICS_PTR(old);
62fa8a84 DM	214	unsigned long prev, new;
62fa8a84 DM	215
9620fef2	216	refcount_set(&p->refcnt, 1);
3fb07daf	217	memcpy(p->metrics, old_p->metrics, sizeof(p->metrics));
62fa8a84 DM	218
	219	new = (unsigned long) p;
	220	prev = cmpxchg(&dst->_metrics, old, new);
	221
	222	if (prev != old) {
	223	kfree(p);
3fb07daf	224	p = (struct dst_metrics *)__DST_METRICS_PTR(prev);
62fa8a84 DM	225	if (prev & DST_METRICS_READ_ONLY)
62fa8a84 DM	226	p = NULL;
3fb07daf	227	} else if (prev & DST_METRICS_REFCOUNTED) {
9620fef2	228	if (refcount_dec_and_test(&old_p->refcnt))
3fb07daf	229	kfree(old_p);
62fa8a84 DM	230	}
62fa8a84 DM	231	}
3fb07daf ED	232	BUILD_BUG_ON(offsetof(struct dst_metrics, metrics) != 0);
3fb07daf ED	233	return (u32 *)p;
62fa8a84 DM	234	}
	235	EXPORT_SYMBOL(dst_cow_metrics_generic);
	236
	237	/* Caller asserts that dst_metrics_read_only(dst) is false. */
	238	void __dst_destroy_metrics_generic(struct dst_entry *dst, unsigned long old)
	239	{
	240	unsigned long prev, new;
	241
3fb07daf	242	new = ((unsigned long) &dst_default_metrics) \| DST_METRICS_READ_ONLY;
62fa8a84 DM	243	prev = cmpxchg(&dst->_metrics, old, new);
	244	if (prev == old)
	245	kfree(__DST_METRICS_PTR(old));
	246	}
	247	EXPORT_SYMBOL(__dst_destroy_metrics_generic);
	248
f38a9eb1 TG	249	static struct dst_ops md_dst_ops = {
	250	.family = AF_UNSPEC,
	251	};
	252
ede2059d	253	static int dst_md_discard_out(struct net net, struct sock sk, struct sk_buff *skb)
f38a9eb1 TG	254	{
	255	WARN_ONCE(1, "Attempting to call output on metadata dst\n");
	256	kfree_skb(skb);
	257	return 0;
	258	}
	259
	260	static int dst_md_discard(struct sk_buff *skb)
	261	{
	262	WARN_ONCE(1, "Attempting to call input on metadata dst\n");
	263	kfree_skb(skb);
	264	return 0;
	265	}
	266
3fcece12 JK	267	static void __metadata_dst_init(struct metadata_dst *md_dst,
	268	enum metadata_type type, u8 optslen)
	269
f38a9eb1	270	{
f38a9eb1 TG	271	struct dst_entry *dst;
f38a9eb1 TG	272
f38a9eb1 TG	273	dst = &md_dst->dst;
f38a9eb1 TG	274	dst_init(dst, &md_dst_ops, NULL, 1, DST_OBSOLETE_NONE,
a4c2fd7f	275	DST_METADATA \| DST_NOCOUNT);
f38a9eb1 TG	276
f38a9eb1 TG	277	dst->input = dst_md_discard;
ede2059d	278	dst->output = dst_md_discard_out;
f38a9eb1 TG	279
f38a9eb1 TG	280	memset(dst + 1, 0, sizeof(md_dst) + optslen - sizeof(dst));
3fcece12	281	md_dst->type = type;
d3aa45ce AS	282	}
d3aa45ce AS	283
3fcece12 JK	284	struct metadata_dst *metadata_dst_alloc(u8 optslen, enum metadata_type type,
3fcece12 JK	285	gfp_t flags)
d3aa45ce AS	286	{
	287	struct metadata_dst *md_dst;
	288
	289	md_dst = kmalloc(sizeof(*md_dst) + optslen, flags);
	290	if (!md_dst)
	291	return NULL;
	292
3fcece12	293	__metadata_dst_init(md_dst, type, optslen);
f38a9eb1 TG	294
	295	return md_dst;
	296	}
	297	EXPORT_SYMBOL_GPL(metadata_dst_alloc);
	298
d71785ff PA	299	void metadata_dst_free(struct metadata_dst *md_dst)
	300	{
	301	#ifdef CONFIG_DST_CACHE
e65a4955 DL	302	if (md_dst->type == METADATA_IP_TUNNEL)
e65a4955 DL	303	dst_cache_destroy(&md_dst->u.tun_info.dst_cache);
d71785ff PA	304	#endif
	305	kfree(md_dst);
	306	}
	307
3fcece12 JK	308	struct metadata_dst __percpu *
3fcece12 JK	309	metadata_dst_alloc_percpu(u8 optslen, enum metadata_type type, gfp_t flags)
d3aa45ce AS	310	{
	311	int cpu;
	312	struct metadata_dst __percpu *md_dst;
	313
	314	md_dst = __alloc_percpu_gfp(sizeof(struct metadata_dst) + optslen,
	315	__alignof__(struct metadata_dst), flags);
	316	if (!md_dst)
	317	return NULL;
	318
	319	for_each_possible_cpu(cpu)
3fcece12	320	__metadata_dst_init(per_cpu_ptr(md_dst, cpu), type, optslen);
d3aa45ce AS	321
	322	return md_dst;
	323	}
	324	EXPORT_SYMBOL_GPL(metadata_dst_alloc_percpu);
d66f2b91 JK	325
	326	void metadata_dst_free_percpu(struct metadata_dst __percpu *md_dst)
	327	{
833e0e2f	328	#ifdef CONFIG_DST_CACHE
d66f2b91 JK	329	int cpu;
d66f2b91 JK	330
d66f2b91 JK	331	for_each_possible_cpu(cpu) {
	332	struct metadata_dst *one_md_dst = per_cpu_ptr(md_dst, cpu);
	333
	334	if (one_md_dst->type == METADATA_IP_TUNNEL)
	335	dst_cache_destroy(&one_md_dst->u.tun_info.dst_cache);
	336	}
	337	#endif
	338	free_percpu(md_dst);
	339	}
	340	EXPORT_SYMBOL_GPL(metadata_dst_free_percpu);