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

/*
 *
 * Copyright IBM Corporation, 2012
 * Author Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
 *
 * This program is free software; you can redistribute it and/or modify it
 * under the terms of version 2.1 of the GNU Lesser General Public License
 * as published by the Free Software Foundation.
 *
 * This program is distributed in the hope that it would be useful, but
 * WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
 *
 */

#include <linux/cgroup.h>
#include <linux/slab.h>
#include <linux/hugetlb.h>
#include <linux/hugetlb_cgroup.h>

struct hugetlb_cgroup {
	struct cgroup_subsys_state css;
	/*
	 * the counter to account for hugepages from hugetlb.
	 */
	struct res_counter hugepage[HUGE_MAX_HSTATE];
};

#define MEMFILE_PRIVATE(x, val)	(((x) << 16) | (val))
#define MEMFILE_IDX(val)	(((val) >> 16) & 0xffff)
#define MEMFILE_ATTR(val)	((val) & 0xffff)

struct cgroup_subsys hugetlb_subsys __read_mostly;
static struct hugetlb_cgroup *root_h_cgroup __read_mostly;

static inline
struct hugetlb_cgroup *hugetlb_cgroup_from_css(struct cgroup_subsys_state *s)
{
	return s ? container_of(s, struct hugetlb_cgroup, css) : NULL;
}

static inline
struct hugetlb_cgroup *hugetlb_cgroup_from_task(struct task_struct *task)
{
	return hugetlb_cgroup_from_css(task_css(task, hugetlb_subsys_id));
}

static inline bool hugetlb_cgroup_is_root(struct hugetlb_cgroup *h_cg)
{
	return (h_cg == root_h_cgroup);
}

static inline struct hugetlb_cgroup *
parent_hugetlb_cgroup(struct hugetlb_cgroup *h_cg)
{
	return hugetlb_cgroup_from_css(css_parent(&h_cg->css));
}

static inline bool hugetlb_cgroup_have_usage(struct hugetlb_cgroup *h_cg)
{
	int idx;

	for (idx = 0; idx < hugetlb_max_hstate; idx++) {
		if ((res_counter_read_u64(&h_cg->hugepage[idx], RES_USAGE)) > 0)
			return true;
	}
	return false;
}

static struct cgroup_subsys_state *
hugetlb_cgroup_css_alloc(struct cgroup_subsys_state *parent_css)
{
	struct hugetlb_cgroup *parent_h_cgroup = hugetlb_cgroup_from_css(parent_css);
	struct hugetlb_cgroup *h_cgroup;
	int idx;

	h_cgroup = kzalloc(sizeof(*h_cgroup), GFP_KERNEL);
	if (!h_cgroup)
		return ERR_PTR(-ENOMEM);

	if (parent_h_cgroup) {
		for (idx = 0; idx < HUGE_MAX_HSTATE; idx++)
			res_counter_init(&h_cgroup->hugepage[idx],
					 &parent_h_cgroup->hugepage[idx]);
	} else {
		root_h_cgroup = h_cgroup;
		for (idx = 0; idx < HUGE_MAX_HSTATE; idx++)
			res_counter_init(&h_cgroup->hugepage[idx], NULL);
	}
	return &h_cgroup->css;
}

static void hugetlb_cgroup_css_free(struct cgroup_subsys_state *css)
{
	struct hugetlb_cgroup *h_cgroup;

	h_cgroup = hugetlb_cgroup_from_css(css);
	kfree(h_cgroup);
}


/*
 * Should be called with hugetlb_lock held.
 * Since we are holding hugetlb_lock, pages cannot get moved from
 * active list or uncharged from the cgroup, So no need to get
 * page reference and test for page active here. This function
 * cannot fail.
 */
static void hugetlb_cgroup_move_parent(int idx, struct hugetlb_cgroup *h_cg,
				       struct page *page)
{
	int csize;
	struct res_counter *counter;
	struct res_counter *fail_res;
	struct hugetlb_cgroup *page_hcg;
	struct hugetlb_cgroup *parent = parent_hugetlb_cgroup(h_cg);

	page_hcg = hugetlb_cgroup_from_page(page);
	/*
	 * We can have pages in active list without any cgroup
	 * ie, hugepage with less than 3 pages. We can safely
	 * ignore those pages.
	 */
	if (!page_hcg || page_hcg != h_cg)
		goto out;

	csize = PAGE_SIZE << compound_order(page);
	if (!parent) {
		parent = root_h_cgroup;
		/* root has no limit */
		res_counter_charge_nofail(&parent->hugepage[idx],
					  csize, &fail_res);
	}
	counter = &h_cg->hugepage[idx];
	res_counter_uncharge_until(counter, counter->parent, csize);

	set_hugetlb_cgroup(page, parent);
out:
	return;
}

/*
 * Force the hugetlb cgroup to empty the hugetlb resources by moving them to
 * the parent cgroup.
 */
static void hugetlb_cgroup_css_offline(struct cgroup_subsys_state *css)
{
	struct hugetlb_cgroup *h_cg = hugetlb_cgroup_from_css(css);
	struct hstate *h;
	struct page *page;
	int idx = 0;

	do {
		for_each_hstate(h) {
			spin_lock(&hugetlb_lock);
			list_for_each_entry(page, &h->hugepage_activelist, lru)
				hugetlb_cgroup_move_parent(idx, h_cg, page);

			spin_unlock(&hugetlb_lock);
			idx++;
		}
		cond_resched();
	} while (hugetlb_cgroup_have_usage(h_cg));
}

int hugetlb_cgroup_charge_cgroup(int idx, unsigned long nr_pages,
				 struct hugetlb_cgroup **ptr)
{
	int ret = 0;
	struct res_counter *fail_res;
	struct hugetlb_cgroup *h_cg = NULL;
	unsigned long csize = nr_pages * PAGE_SIZE;

	if (hugetlb_cgroup_disabled())
		goto done;
	/*
	 * We don't charge any cgroup if the compound page have less
	 * than 3 pages.
	 */
	if (huge_page_order(&hstates[idx]) < HUGETLB_CGROUP_MIN_ORDER)
		goto done;
again:
	rcu_read_lock();
	h_cg = hugetlb_cgroup_from_task(current);
	if (!css_tryget(&h_cg->css)) {
		rcu_read_unlock();
		goto again;
	}
	rcu_read_unlock();

	ret = res_counter_charge(&h_cg->hugepage[idx], csize, &fail_res);
	css_put(&h_cg->css);
done:
	*ptr = h_cg;
	return ret;
}

/* Should be called with hugetlb_lock held */
void hugetlb_cgroup_commit_charge(int idx, unsigned long nr_pages,
				  struct hugetlb_cgroup *h_cg,
				  struct page *page)
{
	if (hugetlb_cgroup_disabled() || !h_cg)
		return;

	set_hugetlb_cgroup(page, h_cg);
	return;
}

/*
 * Should be called with hugetlb_lock held
 */
void hugetlb_cgroup_uncharge_page(int idx, unsigned long nr_pages,
				  struct page *page)
{
	struct hugetlb_cgroup *h_cg;
	unsigned long csize = nr_pages * PAGE_SIZE;

	if (hugetlb_cgroup_disabled())
		return;
	VM_BUG_ON(!spin_is_locked(&hugetlb_lock));
	h_cg = hugetlb_cgroup_from_page(page);
	if (unlikely(!h_cg))
		return;
	set_hugetlb_cgroup(page, NULL);
	res_counter_uncharge(&h_cg->hugepage[idx], csize);
	return;
}

void hugetlb_cgroup_uncharge_cgroup(int idx, unsigned long nr_pages,
				    struct hugetlb_cgroup *h_cg)
{
	unsigned long csize = nr_pages * PAGE_SIZE;

	if (hugetlb_cgroup_disabled() || !h_cg)
		return;

	if (huge_page_order(&hstates[idx]) < HUGETLB_CGROUP_MIN_ORDER)
		return;

	res_counter_uncharge(&h_cg->hugepage[idx], csize);
	return;
}

static u64 hugetlb_cgroup_read_u64(struct cgroup_subsys_state *css,
				   struct cftype *cft)
{
	int idx, name;
	struct hugetlb_cgroup *h_cg = hugetlb_cgroup_from_css(css);

	idx = MEMFILE_IDX(cft->private);
	name = MEMFILE_ATTR(cft->private);

	return res_counter_read_u64(&h_cg->hugepage[idx], name);
}

static int hugetlb_cgroup_write(struct cgroup_subsys_state *css,
				struct cftype *cft, const char *buffer)
{
	int idx, name, ret;
	unsigned long long val;
	struct hugetlb_cgroup *h_cg = hugetlb_cgroup_from_css(css);

	idx = MEMFILE_IDX(cft->private);
	name = MEMFILE_ATTR(cft->private);

	switch (name) {
	case RES_LIMIT:
		if (hugetlb_cgroup_is_root(h_cg)) {
			/* Can't set limit on root */
			ret = -EINVAL;
			break;
		}
		/* This function does all necessary parse...reuse it */
		ret = res_counter_memparse_write_strategy(buffer, &val);
		if (ret)
			break;
		ret = res_counter_set_limit(&h_cg->hugepage[idx], val);
		break;
	default:
		ret = -EINVAL;
		break;
	}
	return ret;
}

static int hugetlb_cgroup_reset(struct cgroup_subsys_state *css,
				unsigned int event)
{
	int idx, name, ret = 0;
	struct hugetlb_cgroup *h_cg = hugetlb_cgroup_from_css(css);

	idx = MEMFILE_IDX(event);
	name = MEMFILE_ATTR(event);

	switch (name) {
	case RES_MAX_USAGE:
		res_counter_reset_max(&h_cg->hugepage[idx]);
		break;
	case RES_FAILCNT:
		res_counter_reset_failcnt(&h_cg->hugepage[idx]);
		break;
	default:
		ret = -EINVAL;
		break;
	}
	return ret;
}

static char *mem_fmt(char *buf, int size, unsigned long hsize)
{
	if (hsize >= (1UL << 30))
		snprintf(buf, size, "%luGB", hsize >> 30);
	else if (hsize >= (1UL << 20))
		snprintf(buf, size, "%luMB", hsize >> 20);
	else
		snprintf(buf, size, "%luKB", hsize >> 10);
	return buf;
}

static void __init __hugetlb_cgroup_file_init(int idx)
{
	char buf[32];
	struct cftype *cft;
	struct hstate *h = &hstates[idx];

	/* format the size */
	mem_fmt(buf, 32, huge_page_size(h));

	/* Add the limit file */
	cft = &h->cgroup_files[0];
	snprintf(cft->name, MAX_CFTYPE_NAME, "%s.limit_in_bytes", buf);
	cft->private = MEMFILE_PRIVATE(idx, RES_LIMIT);
	cft->read_u64 = hugetlb_cgroup_read_u64;
	cft->write_string = hugetlb_cgroup_write;

	/* Add the usage file */
	cft = &h->cgroup_files[1];
	snprintf(cft->name, MAX_CFTYPE_NAME, "%s.usage_in_bytes", buf);
	cft->private = MEMFILE_PRIVATE(idx, RES_USAGE);
	cft->read_u64 = hugetlb_cgroup_read_u64;

	/* Add the MAX usage file */
	cft = &h->cgroup_files[2];
	snprintf(cft->name, MAX_CFTYPE_NAME, "%s.max_usage_in_bytes", buf);
	cft->private = MEMFILE_PRIVATE(idx, RES_MAX_USAGE);
	cft->trigger = hugetlb_cgroup_reset;
	cft->read_u64 = hugetlb_cgroup_read_u64;

	/* Add the failcntfile */
	cft = &h->cgroup_files[3];
	snprintf(cft->name, MAX_CFTYPE_NAME, "%s.failcnt", buf);
	cft->private  = MEMFILE_PRIVATE(idx, RES_FAILCNT);
	cft->trigger  = hugetlb_cgroup_reset;
	cft->read_u64 = hugetlb_cgroup_read_u64;

	/* NULL terminate the last cft */
	cft = &h->cgroup_files[4];
	memset(cft, 0, sizeof(*cft));

	WARN_ON(cgroup_add_cftypes(&hugetlb_subsys, h->cgroup_files));

	return;
}

void __init hugetlb_cgroup_file_init(void)
{
	struct hstate *h;

	for_each_hstate(h) {
		/*
		 * Add cgroup control files only if the huge page consists
		 * of more than two normal pages. This is because we use
		 * page[2].lru.next for storing cgroup details.
		 */
		if (huge_page_order(h) >= HUGETLB_CGROUP_MIN_ORDER)
			__hugetlb_cgroup_file_init(hstate_index(h));
	}
}

/*
 * hugetlb_lock will make sure a parallel cgroup rmdir won't happen
 * when we migrate hugepages
 */
void hugetlb_cgroup_migrate(struct page *oldhpage, struct page *newhpage)
{
	struct hugetlb_cgroup *h_cg;
	struct hstate *h = page_hstate(oldhpage);

	if (hugetlb_cgroup_disabled())
		return;

	VM_BUG_ON(!PageHuge(oldhpage));
	spin_lock(&hugetlb_lock);
	h_cg = hugetlb_cgroup_from_page(oldhpage);
	set_hugetlb_cgroup(oldhpage, NULL);

	/* move the h_cg details to new cgroup */
	set_hugetlb_cgroup(newhpage, h_cg);
	list_move(&newhpage->lru, &h->hugepage_activelist);
	spin_unlock(&hugetlb_lock);
	return;
}

struct cgroup_subsys hugetlb_subsys = {
	.name = "hugetlb",
	.css_alloc	= hugetlb_cgroup_css_alloc,
	.css_offline	= hugetlb_cgroup_css_offline,
	.css_free	= hugetlb_cgroup_css_free,
	.subsys_id	= hugetlb_subsys_id,
};
Commit	Line	Data
2bc64a20 AK	1	/*
	2	*
	3	* Copyright IBM Corporation, 2012
	4	* Author Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
	5	*
	6	* This program is free software; you can redistribute it and/or modify it
	7	* under the terms of version 2.1 of the GNU Lesser General Public License
	8	* as published by the Free Software Foundation.
	9	*
	10	* This program is distributed in the hope that it would be useful, but
	11	* WITHOUT ANY WARRANTY; without even the implied warranty of
	12	* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
	13	*
	14	*/
	15
	16	#include <linux/cgroup.h>
	17	#include <linux/slab.h>
	18	#include <linux/hugetlb.h>
	19	#include <linux/hugetlb_cgroup.h>
	20
	21	struct hugetlb_cgroup {
	22	struct cgroup_subsys_state css;
	23	/*
	24	* the counter to account for hugepages from hugetlb.
	25	*/
	26	struct res_counter hugepage[HUGE_MAX_HSTATE];
	27	};
	28
abb8206c AK	29	#define MEMFILE_PRIVATE(x, val) (((x) << 16) \| (val))
	30	#define MEMFILE_IDX(val) (((val) >> 16) & 0xffff)
	31	#define MEMFILE_ATTR(val) ((val) & 0xffff)
	32
2bc64a20 AK	33	struct cgroup_subsys hugetlb_subsys __read_mostly;
	34	static struct hugetlb_cgroup *root_h_cgroup __read_mostly;
	35
	36	static inline
	37	struct hugetlb_cgroup hugetlb_cgroup_from_css(struct cgroup_subsys_state s)
	38	{
a7c6d554	39	return s ? container_of(s, struct hugetlb_cgroup, css) : NULL;
2bc64a20 AK	40	}
2bc64a20 AK	41
2bc64a20 AK	42	static inline
	43	struct hugetlb_cgroup hugetlb_cgroup_from_task(struct task_struct task)
	44	{
8af01f56	45	return hugetlb_cgroup_from_css(task_css(task, hugetlb_subsys_id));
2bc64a20 AK	46	}
	47
	48	static inline bool hugetlb_cgroup_is_root(struct hugetlb_cgroup *h_cg)
	49	{
	50	return (h_cg == root_h_cgroup);
	51	}
	52
3f798518 TH	53	static inline struct hugetlb_cgroup *
3f798518 TH	54	parent_hugetlb_cgroup(struct hugetlb_cgroup *h_cg)
2bc64a20	55	{
63876986	56	return hugetlb_cgroup_from_css(css_parent(&h_cg->css));
2bc64a20 AK	57	}
2bc64a20 AK	58
3f798518	59	static inline bool hugetlb_cgroup_have_usage(struct hugetlb_cgroup *h_cg)
2bc64a20 AK	60	{
2bc64a20 AK	61	int idx;
2bc64a20 AK	62
	63	for (idx = 0; idx < hugetlb_max_hstate; idx++) {
	64	if ((res_counter_read_u64(&h_cg->hugepage[idx], RES_USAGE)) > 0)
	65	return true;
	66	}
	67	return false;
	68	}
	69
eb95419b TH	70	static struct cgroup_subsys_state *
eb95419b TH	71	hugetlb_cgroup_css_alloc(struct cgroup_subsys_state *parent_css)
2bc64a20	72	{
eb95419b TH	73	struct hugetlb_cgroup *parent_h_cgroup = hugetlb_cgroup_from_css(parent_css);
eb95419b TH	74	struct hugetlb_cgroup *h_cgroup;
2bc64a20	75	int idx;
2bc64a20 AK	76
	77	h_cgroup = kzalloc(sizeof(*h_cgroup), GFP_KERNEL);
	78	if (!h_cgroup)
	79	return ERR_PTR(-ENOMEM);
	80
eb95419b	81	if (parent_h_cgroup) {
2bc64a20 AK	82	for (idx = 0; idx < HUGE_MAX_HSTATE; idx++)
	83	res_counter_init(&h_cgroup->hugepage[idx],
	84	&parent_h_cgroup->hugepage[idx]);
	85	} else {
	86	root_h_cgroup = h_cgroup;
	87	for (idx = 0; idx < HUGE_MAX_HSTATE; idx++)
	88	res_counter_init(&h_cgroup->hugepage[idx], NULL);
	89	}
	90	return &h_cgroup->css;
	91	}
	92
eb95419b	93	static void hugetlb_cgroup_css_free(struct cgroup_subsys_state *css)
2bc64a20 AK	94	{
	95	struct hugetlb_cgroup *h_cgroup;
	96
eb95419b	97	h_cgroup = hugetlb_cgroup_from_css(css);
2bc64a20 AK	98	kfree(h_cgroup);
	99	}
	100
da1def55 AK	101
	102	/*
	103	* Should be called with hugetlb_lock held.
	104	* Since we are holding hugetlb_lock, pages cannot get moved from
	105	* active list or uncharged from the cgroup, So no need to get
	106	* page reference and test for page active here. This function
	107	* cannot fail.
	108	*/
3f798518	109	static void hugetlb_cgroup_move_parent(int idx, struct hugetlb_cgroup *h_cg,
da1def55 AK	110	struct page *page)
	111	{
	112	int csize;
	113	struct res_counter *counter;
	114	struct res_counter *fail_res;
	115	struct hugetlb_cgroup *page_hcg;
3f798518	116	struct hugetlb_cgroup *parent = parent_hugetlb_cgroup(h_cg);
da1def55 AK	117
	118	page_hcg = hugetlb_cgroup_from_page(page);
	119	/*
	120	* We can have pages in active list without any cgroup
	121	* ie, hugepage with less than 3 pages. We can safely
	122	* ignore those pages.
	123	*/
	124	if (!page_hcg \|\| page_hcg != h_cg)
	125	goto out;
	126
	127	csize = PAGE_SIZE << compound_order(page);
	128	if (!parent) {
	129	parent = root_h_cgroup;
	130	/* root has no limit */
	131	res_counter_charge_nofail(&parent->hugepage[idx],
	132	csize, &fail_res);
	133	}
	134	counter = &h_cg->hugepage[idx];
	135	res_counter_uncharge_until(counter, counter->parent, csize);
	136
	137	set_hugetlb_cgroup(page, parent);
	138	out:
	139	return;
	140	}
	141
	142	/*
	143	* Force the hugetlb cgroup to empty the hugetlb resources by moving them to
	144	* the parent cgroup.
	145	*/
eb95419b	146	static void hugetlb_cgroup_css_offline(struct cgroup_subsys_state *css)
2bc64a20	147	{
eb95419b	148	struct hugetlb_cgroup *h_cg = hugetlb_cgroup_from_css(css);
da1def55 AK	149	struct hstate *h;
da1def55 AK	150	struct page *page;
9d093cb1	151	int idx = 0;
da1def55 AK	152
da1def55 AK	153	do {
da1def55 AK	154	for_each_hstate(h) {
	155	spin_lock(&hugetlb_lock);
	156	list_for_each_entry(page, &h->hugepage_activelist, lru)
3f798518	157	hugetlb_cgroup_move_parent(idx, h_cg, page);
da1def55 AK	158
	159	spin_unlock(&hugetlb_lock);
	160	idx++;
	161	}
	162	cond_resched();
3f798518	163	} while (hugetlb_cgroup_have_usage(h_cg));
2bc64a20 AK	164	}
2bc64a20 AK	165
6d76dcf4 AK	166	int hugetlb_cgroup_charge_cgroup(int idx, unsigned long nr_pages,
	167	struct hugetlb_cgroup **ptr)
	168	{
	169	int ret = 0;
	170	struct res_counter *fail_res;
	171	struct hugetlb_cgroup *h_cg = NULL;
	172	unsigned long csize = nr_pages * PAGE_SIZE;
	173
	174	if (hugetlb_cgroup_disabled())
	175	goto done;
	176	/*
	177	* We don't charge any cgroup if the compound page have less
	178	* than 3 pages.
	179	*/
	180	if (huge_page_order(&hstates[idx]) < HUGETLB_CGROUP_MIN_ORDER)
	181	goto done;
	182	again:
	183	rcu_read_lock();
	184	h_cg = hugetlb_cgroup_from_task(current);
	185	if (!css_tryget(&h_cg->css)) {
	186	rcu_read_unlock();
	187	goto again;
	188	}
	189	rcu_read_unlock();
	190
	191	ret = res_counter_charge(&h_cg->hugepage[idx], csize, &fail_res);
	192	css_put(&h_cg->css);
	193	done:
	194	*ptr = h_cg;
	195	return ret;
	196	}
	197
94ae8ba7	198	/* Should be called with hugetlb_lock held */
6d76dcf4 AK	199	void hugetlb_cgroup_commit_charge(int idx, unsigned long nr_pages,
	200	struct hugetlb_cgroup *h_cg,
	201	struct page *page)
	202	{
	203	if (hugetlb_cgroup_disabled() \|\| !h_cg)
	204	return;
	205
6d76dcf4	206	set_hugetlb_cgroup(page, h_cg);
6d76dcf4 AK	207	return;
	208	}
	209
	210	/*
	211	* Should be called with hugetlb_lock held
	212	*/
	213	void hugetlb_cgroup_uncharge_page(int idx, unsigned long nr_pages,
	214	struct page *page)
	215	{
	216	struct hugetlb_cgroup *h_cg;
	217	unsigned long csize = nr_pages * PAGE_SIZE;
	218
	219	if (hugetlb_cgroup_disabled())
	220	return;
	221	VM_BUG_ON(!spin_is_locked(&hugetlb_lock));
	222	h_cg = hugetlb_cgroup_from_page(page);
	223	if (unlikely(!h_cg))
	224	return;
	225	set_hugetlb_cgroup(page, NULL);
	226	res_counter_uncharge(&h_cg->hugepage[idx], csize);
	227	return;
	228	}
	229
	230	void hugetlb_cgroup_uncharge_cgroup(int idx, unsigned long nr_pages,
	231	struct hugetlb_cgroup *h_cg)
	232	{
	233	unsigned long csize = nr_pages * PAGE_SIZE;
	234
	235	if (hugetlb_cgroup_disabled() \|\| !h_cg)
	236	return;
	237
	238	if (huge_page_order(&hstates[idx]) < HUGETLB_CGROUP_MIN_ORDER)
	239	return;
	240
	241	res_counter_uncharge(&h_cg->hugepage[idx], csize);
	242	return;
	243	}
	244
716f479d TH	245	static u64 hugetlb_cgroup_read_u64(struct cgroup_subsys_state *css,
716f479d TH	246	struct cftype *cft)
abb8206c	247	{
716f479d	248	int idx, name;
182446d0	249	struct hugetlb_cgroup *h_cg = hugetlb_cgroup_from_css(css);
abb8206c AK	250
	251	idx = MEMFILE_IDX(cft->private);
	252	name = MEMFILE_ATTR(cft->private);
	253
716f479d	254	return res_counter_read_u64(&h_cg->hugepage[idx], name);
abb8206c AK	255	}
abb8206c AK	256
182446d0 TH	257	static int hugetlb_cgroup_write(struct cgroup_subsys_state *css,
182446d0 TH	258	struct cftype cft, const char buffer)
abb8206c AK	259	{
	260	int idx, name, ret;
	261	unsigned long long val;
182446d0	262	struct hugetlb_cgroup *h_cg = hugetlb_cgroup_from_css(css);
abb8206c AK	263
	264	idx = MEMFILE_IDX(cft->private);
	265	name = MEMFILE_ATTR(cft->private);
	266
	267	switch (name) {
	268	case RES_LIMIT:
	269	if (hugetlb_cgroup_is_root(h_cg)) {
	270	/* Can't set limit on root */
	271	ret = -EINVAL;
	272	break;
	273	}
	274	/* This function does all necessary parse...reuse it */
	275	ret = res_counter_memparse_write_strategy(buffer, &val);
	276	if (ret)
	277	break;
	278	ret = res_counter_set_limit(&h_cg->hugepage[idx], val);
	279	break;
	280	default:
	281	ret = -EINVAL;
	282	break;
	283	}
	284	return ret;
	285	}
	286
182446d0 TH	287	static int hugetlb_cgroup_reset(struct cgroup_subsys_state *css,
182446d0 TH	288	unsigned int event)
abb8206c AK	289	{
abb8206c AK	290	int idx, name, ret = 0;
182446d0	291	struct hugetlb_cgroup *h_cg = hugetlb_cgroup_from_css(css);
abb8206c AK	292
	293	idx = MEMFILE_IDX(event);
	294	name = MEMFILE_ATTR(event);
	295
	296	switch (name) {
	297	case RES_MAX_USAGE:
	298	res_counter_reset_max(&h_cg->hugepage[idx]);
	299	break;
	300	case RES_FAILCNT:
	301	res_counter_reset_failcnt(&h_cg->hugepage[idx]);
	302	break;
	303	default:
	304	ret = -EINVAL;
	305	break;
	306	}
	307	return ret;
	308	}
	309
	310	static char mem_fmt(char buf, int size, unsigned long hsize)
	311	{
	312	if (hsize >= (1UL << 30))
	313	snprintf(buf, size, "%luGB", hsize >> 30);
	314	else if (hsize >= (1UL << 20))
	315	snprintf(buf, size, "%luMB", hsize >> 20);
	316	else
	317	snprintf(buf, size, "%luKB", hsize >> 10);
	318	return buf;
	319	}
	320
7179e7bf	321	static void __init __hugetlb_cgroup_file_init(int idx)
abb8206c AK	322	{
	323	char buf[32];
	324	struct cftype *cft;
	325	struct hstate *h = &hstates[idx];
	326
	327	/* format the size */
	328	mem_fmt(buf, 32, huge_page_size(h));
	329
	330	/* Add the limit file */
	331	cft = &h->cgroup_files[0];
	332	snprintf(cft->name, MAX_CFTYPE_NAME, "%s.limit_in_bytes", buf);
	333	cft->private = MEMFILE_PRIVATE(idx, RES_LIMIT);
716f479d	334	cft->read_u64 = hugetlb_cgroup_read_u64;
abb8206c AK	335	cft->write_string = hugetlb_cgroup_write;
	336
	337	/* Add the usage file */
	338	cft = &h->cgroup_files[1];
	339	snprintf(cft->name, MAX_CFTYPE_NAME, "%s.usage_in_bytes", buf);
	340	cft->private = MEMFILE_PRIVATE(idx, RES_USAGE);
716f479d	341	cft->read_u64 = hugetlb_cgroup_read_u64;
abb8206c AK	342
	343	/* Add the MAX usage file */
	344	cft = &h->cgroup_files[2];
	345	snprintf(cft->name, MAX_CFTYPE_NAME, "%s.max_usage_in_bytes", buf);
	346	cft->private = MEMFILE_PRIVATE(idx, RES_MAX_USAGE);
	347	cft->trigger = hugetlb_cgroup_reset;
716f479d	348	cft->read_u64 = hugetlb_cgroup_read_u64;
abb8206c AK	349
	350	/* Add the failcntfile */
	351	cft = &h->cgroup_files[3];
	352	snprintf(cft->name, MAX_CFTYPE_NAME, "%s.failcnt", buf);
	353	cft->private = MEMFILE_PRIVATE(idx, RES_FAILCNT);
	354	cft->trigger = hugetlb_cgroup_reset;
716f479d	355	cft->read_u64 = hugetlb_cgroup_read_u64;
abb8206c AK	356
	357	/* NULL terminate the last cft */
	358	cft = &h->cgroup_files[4];
	359	memset(cft, 0, sizeof(*cft));
	360
	361	WARN_ON(cgroup_add_cftypes(&hugetlb_subsys, h->cgroup_files));
	362
7179e7bf JW	363	return;
	364	}
	365
	366	void __init hugetlb_cgroup_file_init(void)
	367	{
	368	struct hstate *h;
	369
	370	for_each_hstate(h) {
	371	/*
	372	* Add cgroup control files only if the huge page consists
	373	* of more than two normal pages. This is because we use
	374	* page[2].lru.next for storing cgroup details.
	375	*/
	376	if (huge_page_order(h) >= HUGETLB_CGROUP_MIN_ORDER)
	377	__hugetlb_cgroup_file_init(hstate_index(h));
	378	}
abb8206c AK	379	}
abb8206c AK	380
75754681 AK	381	/*
	382	* hugetlb_lock will make sure a parallel cgroup rmdir won't happen
	383	* when we migrate hugepages
	384	*/
8e6ac7fa AK	385	void hugetlb_cgroup_migrate(struct page oldhpage, struct page newhpage)
	386	{
	387	struct hugetlb_cgroup *h_cg;
94ae8ba7	388	struct hstate *h = page_hstate(oldhpage);
8e6ac7fa AK	389
	390	if (hugetlb_cgroup_disabled())
	391	return;
	392
	393	VM_BUG_ON(!PageHuge(oldhpage));
	394	spin_lock(&hugetlb_lock);
	395	h_cg = hugetlb_cgroup_from_page(oldhpage);
	396	set_hugetlb_cgroup(oldhpage, NULL);
8e6ac7fa AK	397
	398	/* move the h_cg details to new cgroup */
	399	set_hugetlb_cgroup(newhpage, h_cg);
94ae8ba7	400	list_move(&newhpage->lru, &h->hugepage_activelist);
8e6ac7fa	401	spin_unlock(&hugetlb_lock);
8e6ac7fa AK	402	return;
	403	}
	404
2bc64a20 AK	405	struct cgroup_subsys hugetlb_subsys = {
2bc64a20 AK	406	.name = "hugetlb",
92fb9748 TH	407	.css_alloc = hugetlb_cgroup_css_alloc,
	408	.css_offline = hugetlb_cgroup_css_offline,
	409	.css_free = hugetlb_cgroup_css_free,
	410	.subsys_id = hugetlb_subsys_id,
2bc64a20	411	};