[linux-2.6-block.git] / mm / list_lru.c

/*
 * Copyright (c) 2013 Red Hat, Inc. and Parallels Inc. All rights reserved.
 * Authors: David Chinner and Glauber Costa
 *
 * Generic LRU infrastructure
 */
#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/mm.h>
#include <linux/list_lru.h>
#include <linux/slab.h>
#include <linux/mutex.h>
#include <linux/memcontrol.h>

#ifdef CONFIG_MEMCG_KMEM
static LIST_HEAD(list_lrus);
static DEFINE_MUTEX(list_lrus_mutex);

static void list_lru_register(struct list_lru *lru)
{
	mutex_lock(&list_lrus_mutex);
	list_add(&lru->list, &list_lrus);
	mutex_unlock(&list_lrus_mutex);
}

static void list_lru_unregister(struct list_lru *lru)
{
	mutex_lock(&list_lrus_mutex);
	list_del(&lru->list);
	mutex_unlock(&list_lrus_mutex);
}
#else
static void list_lru_register(struct list_lru *lru)
{
}

static void list_lru_unregister(struct list_lru *lru)
{
}
#endif /* CONFIG_MEMCG_KMEM */

#ifdef CONFIG_MEMCG_KMEM
static inline bool list_lru_memcg_aware(struct list_lru *lru)
{
	return !!lru->node[0].memcg_lrus;
}

static inline struct list_lru_one *
list_lru_from_memcg_idx(struct list_lru_node *nlru, int idx)
{
	/*
	 * The lock protects the array of per cgroup lists from relocation
	 * (see memcg_update_list_lru_node).
	 */
	lockdep_assert_held(&nlru->lock);
	if (nlru->memcg_lrus && idx >= 0)
		return nlru->memcg_lrus->lru[idx];

	return &nlru->lru;
}

static inline struct list_lru_one *
list_lru_from_kmem(struct list_lru_node *nlru, void *ptr)
{
	struct mem_cgroup *memcg;

	if (!nlru->memcg_lrus)
		return &nlru->lru;

	memcg = mem_cgroup_from_kmem(ptr);
	if (!memcg)
		return &nlru->lru;

	return list_lru_from_memcg_idx(nlru, memcg_cache_id(memcg));
}
#else
static inline bool list_lru_memcg_aware(struct list_lru *lru)
{
	return false;
}

static inline struct list_lru_one *
list_lru_from_memcg_idx(struct list_lru_node *nlru, int idx)
{
	return &nlru->lru;
}

static inline struct list_lru_one *
list_lru_from_kmem(struct list_lru_node *nlru, void *ptr)
{
	return &nlru->lru;
}
#endif /* CONFIG_MEMCG_KMEM */

bool list_lru_add(struct list_lru *lru, struct list_head *item)
{
	int nid = page_to_nid(virt_to_page(item));
	struct list_lru_node *nlru = &lru->node[nid];
	struct list_lru_one *l;

	spin_lock(&nlru->lock);
	l = list_lru_from_kmem(nlru, item);
	if (list_empty(item)) {
		list_add_tail(item, &l->list);
		l->nr_items++;
		spin_unlock(&nlru->lock);
		return true;
	}
	spin_unlock(&nlru->lock);
	return false;
}
EXPORT_SYMBOL_GPL(list_lru_add);

bool list_lru_del(struct list_lru *lru, struct list_head *item)
{
	int nid = page_to_nid(virt_to_page(item));
	struct list_lru_node *nlru = &lru->node[nid];
	struct list_lru_one *l;

	spin_lock(&nlru->lock);
	l = list_lru_from_kmem(nlru, item);
	if (!list_empty(item)) {
		list_del_init(item);
		l->nr_items--;
		spin_unlock(&nlru->lock);
		return true;
	}
	spin_unlock(&nlru->lock);
	return false;
}
EXPORT_SYMBOL_GPL(list_lru_del);

void list_lru_isolate(struct list_lru_one *list, struct list_head *item)
{
	list_del_init(item);
	list->nr_items--;
}
EXPORT_SYMBOL_GPL(list_lru_isolate);

void list_lru_isolate_move(struct list_lru_one *list, struct list_head *item,
			   struct list_head *head)
{
	list_move(item, head);
	list->nr_items--;
}
EXPORT_SYMBOL_GPL(list_lru_isolate_move);

static unsigned long __list_lru_count_one(struct list_lru *lru,
					  int nid, int memcg_idx)
{
	struct list_lru_node *nlru = &lru->node[nid];
	struct list_lru_one *l;
	unsigned long count;

	spin_lock(&nlru->lock);
	l = list_lru_from_memcg_idx(nlru, memcg_idx);
	count = l->nr_items;
	spin_unlock(&nlru->lock);

	return count;
}

unsigned long list_lru_count_one(struct list_lru *lru,
				 int nid, struct mem_cgroup *memcg)
{
	return __list_lru_count_one(lru, nid, memcg_cache_id(memcg));
}
EXPORT_SYMBOL_GPL(list_lru_count_one);

unsigned long list_lru_count_node(struct list_lru *lru, int nid)
{
	long count = 0;
	int memcg_idx;

	count += __list_lru_count_one(lru, nid, -1);
	if (list_lru_memcg_aware(lru)) {
		for_each_memcg_cache_index(memcg_idx)
			count += __list_lru_count_one(lru, nid, memcg_idx);
	}
	return count;
}
EXPORT_SYMBOL_GPL(list_lru_count_node);

static unsigned long
__list_lru_walk_one(struct list_lru *lru, int nid, int memcg_idx,
		    list_lru_walk_cb isolate, void *cb_arg,
		    unsigned long *nr_to_walk)
{

	struct list_lru_node *nlru = &lru->node[nid];
	struct list_lru_one *l;
	struct list_head *item, *n;
	unsigned long isolated = 0;

	spin_lock(&nlru->lock);
	l = list_lru_from_memcg_idx(nlru, memcg_idx);
restart:
	list_for_each_safe(item, n, &l->list) {
		enum lru_status ret;

		/*
		 * decrement nr_to_walk first so that we don't livelock if we
		 * get stuck on large numbesr of LRU_RETRY items
		 */
		if (!*nr_to_walk)
			break;
		--*nr_to_walk;

		ret = isolate(item, l, &nlru->lock, cb_arg);
		switch (ret) {
		case LRU_REMOVED_RETRY:
			assert_spin_locked(&nlru->lock);
		case LRU_REMOVED:
			isolated++;
			/*
			 * If the lru lock has been dropped, our list
			 * traversal is now invalid and so we have to
			 * restart from scratch.
			 */
			if (ret == LRU_REMOVED_RETRY)
				goto restart;
			break;
		case LRU_ROTATE:
			list_move_tail(item, &l->list);
			break;
		case LRU_SKIP:
			break;
		case LRU_RETRY:
			/*
			 * The lru lock has been dropped, our list traversal is
			 * now invalid and so we have to restart from scratch.
			 */
			assert_spin_locked(&nlru->lock);
			goto restart;
		default:
			BUG();
		}
	}

	spin_unlock(&nlru->lock);
	return isolated;
}

unsigned long
list_lru_walk_one(struct list_lru *lru, int nid, struct mem_cgroup *memcg,
		  list_lru_walk_cb isolate, void *cb_arg,
		  unsigned long *nr_to_walk)
{
	return __list_lru_walk_one(lru, nid, memcg_cache_id(memcg),
				   isolate, cb_arg, nr_to_walk);
}
EXPORT_SYMBOL_GPL(list_lru_walk_one);

unsigned long list_lru_walk_node(struct list_lru *lru, int nid,
				 list_lru_walk_cb isolate, void *cb_arg,
				 unsigned long *nr_to_walk)
{
	long isolated = 0;
	int memcg_idx;

	isolated += __list_lru_walk_one(lru, nid, -1, isolate, cb_arg,
					nr_to_walk);
	if (*nr_to_walk > 0 && list_lru_memcg_aware(lru)) {
		for_each_memcg_cache_index(memcg_idx) {
			isolated += __list_lru_walk_one(lru, nid, memcg_idx,
						isolate, cb_arg, nr_to_walk);
			if (*nr_to_walk <= 0)
				break;
		}
	}
	return isolated;
}
EXPORT_SYMBOL_GPL(list_lru_walk_node);

static void init_one_lru(struct list_lru_one *l)
{
	INIT_LIST_HEAD(&l->list);
	l->nr_items = 0;
}

#ifdef CONFIG_MEMCG_KMEM
static void __memcg_destroy_list_lru_node(struct list_lru_memcg *memcg_lrus,
					  int begin, int end)
{
	int i;

	for (i = begin; i < end; i++)
		kfree(memcg_lrus->lru[i]);
}

static int __memcg_init_list_lru_node(struct list_lru_memcg *memcg_lrus,
				      int begin, int end)
{
	int i;

	for (i = begin; i < end; i++) {
		struct list_lru_one *l;

		l = kmalloc(sizeof(struct list_lru_one), GFP_KERNEL);
		if (!l)
			goto fail;

		init_one_lru(l);
		memcg_lrus->lru[i] = l;
	}
	return 0;
fail:
	__memcg_destroy_list_lru_node(memcg_lrus, begin, i - 1);
	return -ENOMEM;
}

static int memcg_init_list_lru_node(struct list_lru_node *nlru)
{
	int size = memcg_nr_cache_ids;

	nlru->memcg_lrus = kmalloc(size * sizeof(void *), GFP_KERNEL);
	if (!nlru->memcg_lrus)
		return -ENOMEM;

	if (__memcg_init_list_lru_node(nlru->memcg_lrus, 0, size)) {
		kfree(nlru->memcg_lrus);
		return -ENOMEM;
	}

	return 0;
}

static void memcg_destroy_list_lru_node(struct list_lru_node *nlru)
{
	__memcg_destroy_list_lru_node(nlru->memcg_lrus, 0, memcg_nr_cache_ids);
	kfree(nlru->memcg_lrus);
}

static int memcg_update_list_lru_node(struct list_lru_node *nlru,
				      int old_size, int new_size)
{
	struct list_lru_memcg *old, *new;

	BUG_ON(old_size > new_size);

	old = nlru->memcg_lrus;
	new = kmalloc(new_size * sizeof(void *), GFP_KERNEL);
	if (!new)
		return -ENOMEM;

	if (__memcg_init_list_lru_node(new, old_size, new_size)) {
		kfree(new);
		return -ENOMEM;
	}

	memcpy(new, old, old_size * sizeof(void *));

	/*
	 * The lock guarantees that we won't race with a reader
	 * (see list_lru_from_memcg_idx).
	 *
	 * Since list_lru_{add,del} may be called under an IRQ-safe lock,
	 * we have to use IRQ-safe primitives here to avoid deadlock.
	 */
	spin_lock_irq(&nlru->lock);
	nlru->memcg_lrus = new;
	spin_unlock_irq(&nlru->lock);

	kfree(old);
	return 0;
}

static void memcg_cancel_update_list_lru_node(struct list_lru_node *nlru,
					      int old_size, int new_size)
{
	/* do not bother shrinking the array back to the old size, because we
	 * cannot handle allocation failures here */
	__memcg_destroy_list_lru_node(nlru->memcg_lrus, old_size, new_size);
}

static int memcg_init_list_lru(struct list_lru *lru, bool memcg_aware)
{
	int i;

	for (i = 0; i < nr_node_ids; i++) {
		if (!memcg_aware)
			lru->node[i].memcg_lrus = NULL;
		else if (memcg_init_list_lru_node(&lru->node[i]))
			goto fail;
	}
	return 0;
fail:
	for (i = i - 1; i >= 0; i--)
		memcg_destroy_list_lru_node(&lru->node[i]);
	return -ENOMEM;
}

static void memcg_destroy_list_lru(struct list_lru *lru)
{
	int i;

	if (!list_lru_memcg_aware(lru))
		return;

	for (i = 0; i < nr_node_ids; i++)
		memcg_destroy_list_lru_node(&lru->node[i]);
}

static int memcg_update_list_lru(struct list_lru *lru,
				 int old_size, int new_size)
{
	int i;

	if (!list_lru_memcg_aware(lru))
		return 0;

	for (i = 0; i < nr_node_ids; i++) {
		if (memcg_update_list_lru_node(&lru->node[i],
					       old_size, new_size))
			goto fail;
	}
	return 0;
fail:
	for (i = i - 1; i >= 0; i--)
		memcg_cancel_update_list_lru_node(&lru->node[i],
						  old_size, new_size);
	return -ENOMEM;
}

static void memcg_cancel_update_list_lru(struct list_lru *lru,
					 int old_size, int new_size)
{
	int i;

	if (!list_lru_memcg_aware(lru))
		return;

	for (i = 0; i < nr_node_ids; i++)
		memcg_cancel_update_list_lru_node(&lru->node[i],
						  old_size, new_size);
}

int memcg_update_all_list_lrus(int new_size)
{
	int ret = 0;
	struct list_lru *lru;
	int old_size = memcg_nr_cache_ids;

	mutex_lock(&list_lrus_mutex);
	list_for_each_entry(lru, &list_lrus, list) {
		ret = memcg_update_list_lru(lru, old_size, new_size);
		if (ret)
			goto fail;
	}
out:
	mutex_unlock(&list_lrus_mutex);
	return ret;
fail:
	list_for_each_entry_continue_reverse(lru, &list_lrus, list)
		memcg_cancel_update_list_lru(lru, old_size, new_size);
	goto out;
}

static void memcg_drain_list_lru_node(struct list_lru_node *nlru,
				      int src_idx, int dst_idx)
{
	struct list_lru_one *src, *dst;

	/*
	 * Since list_lru_{add,del} may be called under an IRQ-safe lock,
	 * we have to use IRQ-safe primitives here to avoid deadlock.
	 */
	spin_lock_irq(&nlru->lock);

	src = list_lru_from_memcg_idx(nlru, src_idx);
	dst = list_lru_from_memcg_idx(nlru, dst_idx);

	list_splice_init(&src->list, &dst->list);
	dst->nr_items += src->nr_items;
	src->nr_items = 0;

	spin_unlock_irq(&nlru->lock);
}

static void memcg_drain_list_lru(struct list_lru *lru,
				 int src_idx, int dst_idx)
{
	int i;

	if (!list_lru_memcg_aware(lru))
		return;

	for (i = 0; i < nr_node_ids; i++)
		memcg_drain_list_lru_node(&lru->node[i], src_idx, dst_idx);
}

void memcg_drain_all_list_lrus(int src_idx, int dst_idx)
{
	struct list_lru *lru;

	mutex_lock(&list_lrus_mutex);
	list_for_each_entry(lru, &list_lrus, list)
		memcg_drain_list_lru(lru, src_idx, dst_idx);
	mutex_unlock(&list_lrus_mutex);
}
#else
static int memcg_init_list_lru(struct list_lru *lru, bool memcg_aware)
{
	return 0;
}

static void memcg_destroy_list_lru(struct list_lru *lru)
{
}
#endif /* CONFIG_MEMCG_KMEM */

int __list_lru_init(struct list_lru *lru, bool memcg_aware,
		    struct lock_class_key *key)
{
	int i;
	size_t size = sizeof(*lru->node) * nr_node_ids;
	int err = -ENOMEM;

	memcg_get_cache_ids();

	lru->node = kzalloc(size, GFP_KERNEL);
	if (!lru->node)
		goto out;

	for (i = 0; i < nr_node_ids; i++) {
		spin_lock_init(&lru->node[i].lock);
		if (key)
			lockdep_set_class(&lru->node[i].lock, key);
		init_one_lru(&lru->node[i].lru);
	}

	err = memcg_init_list_lru(lru, memcg_aware);
	if (err) {
		kfree(lru->node);
		goto out;
	}

	list_lru_register(lru);
out:
	memcg_put_cache_ids();
	return err;
}
EXPORT_SYMBOL_GPL(__list_lru_init);

void list_lru_destroy(struct list_lru *lru)
{
	/* Already destroyed or not yet initialized? */
	if (!lru->node)
		return;

	memcg_get_cache_ids();

	list_lru_unregister(lru);

	memcg_destroy_list_lru(lru);
	kfree(lru->node);
	lru->node = NULL;

	memcg_put_cache_ids();
}
EXPORT_SYMBOL_GPL(list_lru_destroy);
Commit	Line	Data
a38e4082 DC	1	/*
	2	* Copyright (c) 2013 Red Hat, Inc. and Parallels Inc. All rights reserved.
	3	* Authors: David Chinner and Glauber Costa
	4	*
	5	* Generic LRU infrastructure
	6	*/
	7	#include <linux/kernel.h>
	8	#include <linux/module.h>
3b1d58a4	9	#include <linux/mm.h>
a38e4082	10	#include <linux/list_lru.h>
5ca302c8	11	#include <linux/slab.h>
c0a5b560	12	#include <linux/mutex.h>
60d3fd32	13	#include <linux/memcontrol.h>
c0a5b560 VD	14
	15	#ifdef CONFIG_MEMCG_KMEM
	16	static LIST_HEAD(list_lrus);
	17	static DEFINE_MUTEX(list_lrus_mutex);
	18
	19	static void list_lru_register(struct list_lru *lru)
	20	{
	21	mutex_lock(&list_lrus_mutex);
	22	list_add(&lru->list, &list_lrus);
	23	mutex_unlock(&list_lrus_mutex);
	24	}
	25
	26	static void list_lru_unregister(struct list_lru *lru)
	27	{
	28	mutex_lock(&list_lrus_mutex);
	29	list_del(&lru->list);
	30	mutex_unlock(&list_lrus_mutex);
	31	}
	32	#else
	33	static void list_lru_register(struct list_lru *lru)
	34	{
	35	}
	36
	37	static void list_lru_unregister(struct list_lru *lru)
	38	{
	39	}
	40	#endif /* CONFIG_MEMCG_KMEM */
a38e4082	41
60d3fd32 VD	42	#ifdef CONFIG_MEMCG_KMEM
	43	static inline bool list_lru_memcg_aware(struct list_lru *lru)
	44	{
	45	return !!lru->node[0].memcg_lrus;
	46	}
	47
	48	static inline struct list_lru_one *
	49	list_lru_from_memcg_idx(struct list_lru_node *nlru, int idx)
	50	{
	51	/*
	52	* The lock protects the array of per cgroup lists from relocation
	53	* (see memcg_update_list_lru_node).
	54	*/
	55	lockdep_assert_held(&nlru->lock);
	56	if (nlru->memcg_lrus && idx >= 0)
	57	return nlru->memcg_lrus->lru[idx];
	58
	59	return &nlru->lru;
	60	}
	61
	62	static inline struct list_lru_one *
	63	list_lru_from_kmem(struct list_lru_node nlru, void ptr)
	64	{
	65	struct mem_cgroup *memcg;
	66
	67	if (!nlru->memcg_lrus)
	68	return &nlru->lru;
	69
	70	memcg = mem_cgroup_from_kmem(ptr);
	71	if (!memcg)
	72	return &nlru->lru;
	73
	74	return list_lru_from_memcg_idx(nlru, memcg_cache_id(memcg));
	75	}
	76	#else
	77	static inline bool list_lru_memcg_aware(struct list_lru *lru)
	78	{
	79	return false;
	80	}
	81
	82	static inline struct list_lru_one *
	83	list_lru_from_memcg_idx(struct list_lru_node *nlru, int idx)
	84	{
	85	return &nlru->lru;
	86	}
	87
	88	static inline struct list_lru_one *
	89	list_lru_from_kmem(struct list_lru_node nlru, void ptr)
	90	{
	91	return &nlru->lru;
	92	}
	93	#endif /* CONFIG_MEMCG_KMEM */
	94
a38e4082 DC	95	bool list_lru_add(struct list_lru lru, struct list_head item)
a38e4082 DC	96	{
3b1d58a4 DC	97	int nid = page_to_nid(virt_to_page(item));
3b1d58a4 DC	98	struct list_lru_node *nlru = &lru->node[nid];
60d3fd32	99	struct list_lru_one *l;
3b1d58a4 DC	100
3b1d58a4 DC	101	spin_lock(&nlru->lock);
60d3fd32	102	l = list_lru_from_kmem(nlru, item);
a38e4082	103	if (list_empty(item)) {
60d3fd32 VD	104	list_add_tail(item, &l->list);
60d3fd32 VD	105	l->nr_items++;
3b1d58a4	106	spin_unlock(&nlru->lock);
a38e4082 DC	107	return true;
a38e4082 DC	108	}
3b1d58a4	109	spin_unlock(&nlru->lock);
a38e4082 DC	110	return false;
	111	}
	112	EXPORT_SYMBOL_GPL(list_lru_add);
	113
	114	bool list_lru_del(struct list_lru lru, struct list_head item)
	115	{
3b1d58a4 DC	116	int nid = page_to_nid(virt_to_page(item));
3b1d58a4 DC	117	struct list_lru_node *nlru = &lru->node[nid];
60d3fd32	118	struct list_lru_one *l;
3b1d58a4 DC	119
3b1d58a4 DC	120	spin_lock(&nlru->lock);
60d3fd32	121	l = list_lru_from_kmem(nlru, item);
a38e4082 DC	122	if (!list_empty(item)) {
a38e4082 DC	123	list_del_init(item);
60d3fd32	124	l->nr_items--;
3b1d58a4	125	spin_unlock(&nlru->lock);
a38e4082 DC	126	return true;
a38e4082 DC	127	}
3b1d58a4	128	spin_unlock(&nlru->lock);
a38e4082 DC	129	return false;
	130	}
	131	EXPORT_SYMBOL_GPL(list_lru_del);
	132
3f97b163 VD	133	void list_lru_isolate(struct list_lru_one list, struct list_head item)
	134	{
	135	list_del_init(item);
	136	list->nr_items--;
	137	}
	138	EXPORT_SYMBOL_GPL(list_lru_isolate);
	139
	140	void list_lru_isolate_move(struct list_lru_one list, struct list_head item,
	141	struct list_head *head)
	142	{
	143	list_move(item, head);
	144	list->nr_items--;
	145	}
	146	EXPORT_SYMBOL_GPL(list_lru_isolate_move);
	147
60d3fd32 VD	148	static unsigned long __list_lru_count_one(struct list_lru *lru,
60d3fd32 VD	149	int nid, int memcg_idx)
a38e4082	150	{
6a4f496f	151	struct list_lru_node *nlru = &lru->node[nid];
60d3fd32 VD	152	struct list_lru_one *l;
60d3fd32 VD	153	unsigned long count;
3b1d58a4	154
6a4f496f	155	spin_lock(&nlru->lock);
60d3fd32	156	l = list_lru_from_memcg_idx(nlru, memcg_idx);
60d3fd32	157	count = l->nr_items;
6a4f496f	158	spin_unlock(&nlru->lock);
3b1d58a4 DC	159
	160	return count;
	161	}
60d3fd32 VD	162
	163	unsigned long list_lru_count_one(struct list_lru *lru,
	164	int nid, struct mem_cgroup *memcg)
	165	{
	166	return __list_lru_count_one(lru, nid, memcg_cache_id(memcg));
	167	}
	168	EXPORT_SYMBOL_GPL(list_lru_count_one);
	169
	170	unsigned long list_lru_count_node(struct list_lru *lru, int nid)
	171	{
	172	long count = 0;
	173	int memcg_idx;
	174
	175	count += __list_lru_count_one(lru, nid, -1);
	176	if (list_lru_memcg_aware(lru)) {
	177	for_each_memcg_cache_index(memcg_idx)
	178	count += __list_lru_count_one(lru, nid, memcg_idx);
	179	}
	180	return count;
	181	}
6a4f496f	182	EXPORT_SYMBOL_GPL(list_lru_count_node);
3b1d58a4	183
60d3fd32 VD	184	static unsigned long
	185	__list_lru_walk_one(struct list_lru *lru, int nid, int memcg_idx,
	186	list_lru_walk_cb isolate, void *cb_arg,
	187	unsigned long *nr_to_walk)
3b1d58a4 DC	188	{
3b1d58a4 DC	189
60d3fd32 VD	190	struct list_lru_node *nlru = &lru->node[nid];
60d3fd32 VD	191	struct list_lru_one *l;
a38e4082	192	struct list_head item, n;
3b1d58a4	193	unsigned long isolated = 0;
a38e4082	194
3b1d58a4	195	spin_lock(&nlru->lock);
60d3fd32	196	l = list_lru_from_memcg_idx(nlru, memcg_idx);
a38e4082	197	restart:
60d3fd32	198	list_for_each_safe(item, n, &l->list) {
a38e4082	199	enum lru_status ret;
5cedf721 DC	200
	201	/*
	202	* decrement nr_to_walk first so that we don't livelock if we
	203	* get stuck on large numbesr of LRU_RETRY items
	204	*/
c56b097a	205	if (!*nr_to_walk)
5cedf721	206	break;
c56b097a	207	--*nr_to_walk;
5cedf721	208
3f97b163	209	ret = isolate(item, l, &nlru->lock, cb_arg);
a38e4082	210	switch (ret) {
449dd698 JW	211	case LRU_REMOVED_RETRY:
449dd698 JW	212	assert_spin_locked(&nlru->lock);
a38e4082	213	case LRU_REMOVED:
3b1d58a4	214	isolated++;
449dd698 JW	215	/*
	216	* If the lru lock has been dropped, our list
	217	* traversal is now invalid and so we have to
	218	* restart from scratch.
	219	*/
	220	if (ret == LRU_REMOVED_RETRY)
	221	goto restart;
a38e4082 DC	222	break;
a38e4082 DC	223	case LRU_ROTATE:
60d3fd32	224	list_move_tail(item, &l->list);
a38e4082 DC	225	break;
	226	case LRU_SKIP:
	227	break;
	228	case LRU_RETRY:
5cedf721 DC	229	/*
	230	* The lru lock has been dropped, our list traversal is
	231	* now invalid and so we have to restart from scratch.
	232	*/
449dd698	233	assert_spin_locked(&nlru->lock);
a38e4082 DC	234	goto restart;
	235	default:
	236	BUG();
	237	}
a38e4082	238	}
3b1d58a4 DC	239
	240	spin_unlock(&nlru->lock);
	241	return isolated;
	242	}
60d3fd32 VD	243
	244	unsigned long
	245	list_lru_walk_one(struct list_lru lru, int nid, struct mem_cgroup memcg,
	246	list_lru_walk_cb isolate, void *cb_arg,
	247	unsigned long *nr_to_walk)
	248	{
	249	return __list_lru_walk_one(lru, nid, memcg_cache_id(memcg),
	250	isolate, cb_arg, nr_to_walk);
	251	}
	252	EXPORT_SYMBOL_GPL(list_lru_walk_one);
	253
	254	unsigned long list_lru_walk_node(struct list_lru *lru, int nid,
	255	list_lru_walk_cb isolate, void *cb_arg,
	256	unsigned long *nr_to_walk)
	257	{
	258	long isolated = 0;
	259	int memcg_idx;
	260
	261	isolated += __list_lru_walk_one(lru, nid, -1, isolate, cb_arg,
	262	nr_to_walk);
	263	if (*nr_to_walk > 0 && list_lru_memcg_aware(lru)) {
	264	for_each_memcg_cache_index(memcg_idx) {
	265	isolated += __list_lru_walk_one(lru, nid, memcg_idx,
	266	isolate, cb_arg, nr_to_walk);
	267	if (*nr_to_walk <= 0)
	268	break;
	269	}
	270	}
	271	return isolated;
	272	}
3b1d58a4 DC	273	EXPORT_SYMBOL_GPL(list_lru_walk_node);
3b1d58a4 DC	274
60d3fd32 VD	275	static void init_one_lru(struct list_lru_one *l)
	276	{
	277	INIT_LIST_HEAD(&l->list);
	278	l->nr_items = 0;
	279	}
	280
	281	#ifdef CONFIG_MEMCG_KMEM
	282	static void __memcg_destroy_list_lru_node(struct list_lru_memcg *memcg_lrus,
	283	int begin, int end)
	284	{
	285	int i;
	286
	287	for (i = begin; i < end; i++)
	288	kfree(memcg_lrus->lru[i]);
	289	}
	290
	291	static int __memcg_init_list_lru_node(struct list_lru_memcg *memcg_lrus,
	292	int begin, int end)
	293	{
	294	int i;
	295
	296	for (i = begin; i < end; i++) {
	297	struct list_lru_one *l;
	298
	299	l = kmalloc(sizeof(struct list_lru_one), GFP_KERNEL);
	300	if (!l)
	301	goto fail;
	302
	303	init_one_lru(l);
	304	memcg_lrus->lru[i] = l;
	305	}
	306	return 0;
	307	fail:
	308	__memcg_destroy_list_lru_node(memcg_lrus, begin, i - 1);
	309	return -ENOMEM;
	310	}
	311
	312	static int memcg_init_list_lru_node(struct list_lru_node *nlru)
	313	{
	314	int size = memcg_nr_cache_ids;
	315
	316	nlru->memcg_lrus = kmalloc(size * sizeof(void *), GFP_KERNEL);
	317	if (!nlru->memcg_lrus)
	318	return -ENOMEM;
	319
	320	if (__memcg_init_list_lru_node(nlru->memcg_lrus, 0, size)) {
	321	kfree(nlru->memcg_lrus);
	322	return -ENOMEM;
	323	}
	324
	325	return 0;
	326	}
	327
	328	static void memcg_destroy_list_lru_node(struct list_lru_node *nlru)
	329	{
	330	__memcg_destroy_list_lru_node(nlru->memcg_lrus, 0, memcg_nr_cache_ids);
	331	kfree(nlru->memcg_lrus);
	332	}
	333
	334	static int memcg_update_list_lru_node(struct list_lru_node *nlru,
	335	int old_size, int new_size)
	336	{
	337	struct list_lru_memcg old, new;
	338
339	BUG_ON(old_size > new_size);
340
341	old = nlru->memcg_lrus;
342	new = kmalloc(new_size * sizeof(void *), GFP_KERNEL);
343	if (!new)
344	return -ENOMEM;
345
346	if (__memcg_init_list_lru_node(new, old_size, new_size)) {
347	kfree(new);
348	return -ENOMEM;
349	}
350
351	memcpy(new, old, old_size * sizeof(void *));
352
353	/*
354	* The lock guarantees that we won't race with a reader
355	* (see list_lru_from_memcg_idx).
356	*
357	* Since list_lru_{add,del} may be called under an IRQ-safe lock,
358	* we have to use IRQ-safe primitives here to avoid deadlock.
359	*/
360	spin_lock_irq(&nlru->lock);
361	nlru->memcg_lrus = new;
362	spin_unlock_irq(&nlru->lock);
363
364	kfree(old);
365	return 0;
366	}
367
368	static void memcg_cancel_update_list_lru_node(struct list_lru_node *nlru,
369	int old_size, int new_size)
370	{
371	/* do not bother shrinking the array back to the old size, because we
372	* cannot handle allocation failures here */
373	__memcg_destroy_list_lru_node(nlru->memcg_lrus, old_size, new_size);
374	}
375
376	static int memcg_init_list_lru(struct list_lru *lru, bool memcg_aware)
377	{
378	int i;
379
380	for (i = 0; i < nr_node_ids; i++) {
381	if (!memcg_aware)
382	lru->node[i].memcg_lrus = NULL;
383	else if (memcg_init_list_lru_node(&lru->node[i]))
384	goto fail;
385	}
386	return 0;
387	fail:
388	for (i = i - 1; i >= 0; i--)
389	memcg_destroy_list_lru_node(&lru->node[i]);
390	return -ENOMEM;
391	}
392
393	static void memcg_destroy_list_lru(struct list_lru *lru)
394	{
395	int i;
396
397	if (!list_lru_memcg_aware(lru))
398	return;
399
400	for (i = 0; i < nr_node_ids; i++)
401	memcg_destroy_list_lru_node(&lru->node[i]);
402	}
403
404	static int memcg_update_list_lru(struct list_lru *lru,
405	int old_size, int new_size)
406	{
407	int i;
408
409	if (!list_lru_memcg_aware(lru))
410	return 0;
411
412	for (i = 0; i < nr_node_ids; i++) {
413	if (memcg_update_list_lru_node(&lru->node[i],
414	old_size, new_size))
415	goto fail;
416	}
417	return 0;
418	fail:
419	for (i = i - 1; i >= 0; i--)
420	memcg_cancel_update_list_lru_node(&lru->node[i],
421	old_size, new_size);
422	return -ENOMEM;
423	}
424
425	static void memcg_cancel_update_list_lru(struct list_lru *lru,
426	int old_size, int new_size)
427	{
428	int i;
429
430	if (!list_lru_memcg_aware(lru))
431	return;
432
433	for (i = 0; i < nr_node_ids; i++)
434	memcg_cancel_update_list_lru_node(&lru->node[i],
435	old_size, new_size);
436	}
437
438	int memcg_update_all_list_lrus(int new_size)
439	{
440	int ret = 0;
441	struct list_lru *lru;
442	int old_size = memcg_nr_cache_ids;
443
444	mutex_lock(&list_lrus_mutex);
445	list_for_each_entry(lru, &list_lrus, list) {
446	ret = memcg_update_list_lru(lru, old_size, new_size);
447	if (ret)
448	goto fail;
449	}
450	out:
451	mutex_unlock(&list_lrus_mutex);
452	return ret;
453	fail:
454	list_for_each_entry_continue_reverse(lru, &list_lrus, list)
455	memcg_cancel_update_list_lru(lru, old_size, new_size);
456	goto out;
457	}
2788cf0c VD	458
	459	static void memcg_drain_list_lru_node(struct list_lru_node *nlru,
	460	int src_idx, int dst_idx)
	461	{
	462	struct list_lru_one src, dst;
	463
	464	/*
	465	* Since list_lru_{add,del} may be called under an IRQ-safe lock,
	466	* we have to use IRQ-safe primitives here to avoid deadlock.
	467	*/
	468	spin_lock_irq(&nlru->lock);
	469
	470	src = list_lru_from_memcg_idx(nlru, src_idx);
	471	dst = list_lru_from_memcg_idx(nlru, dst_idx);
	472
	473	list_splice_init(&src->list, &dst->list);
	474	dst->nr_items += src->nr_items;
	475	src->nr_items = 0;
	476
	477	spin_unlock_irq(&nlru->lock);
	478	}
	479
	480	static void memcg_drain_list_lru(struct list_lru *lru,
	481	int src_idx, int dst_idx)
	482	{
	483	int i;
	484
	485	if (!list_lru_memcg_aware(lru))
	486	return;
	487
	488	for (i = 0; i < nr_node_ids; i++)
	489	memcg_drain_list_lru_node(&lru->node[i], src_idx, dst_idx);
	490	}
	491
	492	void memcg_drain_all_list_lrus(int src_idx, int dst_idx)
	493	{
	494	struct list_lru *lru;
	495
	496	mutex_lock(&list_lrus_mutex);
	497	list_for_each_entry(lru, &list_lrus, list)
	498	memcg_drain_list_lru(lru, src_idx, dst_idx);
	499	mutex_unlock(&list_lrus_mutex);
	500	}
60d3fd32 VD	501	#else
	502	static int memcg_init_list_lru(struct list_lru *lru, bool memcg_aware)
	503	{
	504	return 0;
	505	}
	506
	507	static void memcg_destroy_list_lru(struct list_lru *lru)
	508	{
	509	}
	510	#endif /* CONFIG_MEMCG_KMEM */
	511
	512	int __list_lru_init(struct list_lru *lru, bool memcg_aware,
	513	struct lock_class_key *key)
a38e4082	514	{
3b1d58a4	515	int i;
5ca302c8	516	size_t size = sizeof(lru->node) nr_node_ids;
60d3fd32 VD	517	int err = -ENOMEM;
	518
	519	memcg_get_cache_ids();
5ca302c8 GC	520
	521	lru->node = kzalloc(size, GFP_KERNEL);
	522	if (!lru->node)
60d3fd32	523	goto out;
a38e4082	524
5ca302c8	525	for (i = 0; i < nr_node_ids; i++) {
3b1d58a4	526	spin_lock_init(&lru->node[i].lock);
449dd698 JW	527	if (key)
449dd698 JW	528	lockdep_set_class(&lru->node[i].lock, key);
60d3fd32 VD	529	init_one_lru(&lru->node[i].lru);
	530	}
	531
	532	err = memcg_init_list_lru(lru, memcg_aware);
	533	if (err) {
	534	kfree(lru->node);
	535	goto out;
3b1d58a4	536	}
60d3fd32	537
c0a5b560	538	list_lru_register(lru);
60d3fd32 VD	539	out:
	540	memcg_put_cache_ids();
	541	return err;
a38e4082	542	}
60d3fd32	543	EXPORT_SYMBOL_GPL(__list_lru_init);
5ca302c8 GC	544
	545	void list_lru_destroy(struct list_lru *lru)
	546	{
c0a5b560 VD	547	/* Already destroyed or not yet initialized? */
	548	if (!lru->node)
	549	return;
60d3fd32 VD	550
	551	memcg_get_cache_ids();
	552
c0a5b560	553	list_lru_unregister(lru);
60d3fd32 VD	554
60d3fd32 VD	555	memcg_destroy_list_lru(lru);
5ca302c8	556	kfree(lru->node);
c0a5b560	557	lru->node = NULL;
60d3fd32 VD	558
60d3fd32 VD	559	memcg_put_cache_ids();
5ca302c8 GC	560	}
5ca302c8 GC	561	EXPORT_SYMBOL_GPL(list_lru_destroy);