[mirror_ubuntu-artful-kernel.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_cgroup(struct cgroup *cgroup)
{
	return hugetlb_cgroup_from_css(cgroup_css(cgroup, hugetlb_subsys_id));
}

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 *cgroup)
{
	int idx;
	struct cgroup *parent_cgroup;
	struct hugetlb_cgroup *h_cgroup, *parent_h_cgroup;

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

	parent_cgroup = cgroup->parent;
	if (parent_cgroup) {
		parent_h_cgroup = hugetlb_cgroup_from_cgroup(parent_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 *cgroup)
{
	struct hugetlb_cgroup *h_cgroup;

	h_cgroup = hugetlb_cgroup_from_cgroup(cgroup);
	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 *cgroup)
{
	struct hugetlb_cgroup *h_cg = hugetlb_cgroup_from_cgroup(cgroup);
	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 ssize_t hugetlb_cgroup_read(struct cgroup *cgroup, struct cftype *cft,
				   struct file *file, char __user *buf,
				   size_t nbytes, loff_t *ppos)
{
	u64 val;
	char str[64];
	int idx, name, len;
	struct hugetlb_cgroup *h_cg = hugetlb_cgroup_from_cgroup(cgroup);

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

	val = res_counter_read_u64(&h_cg->hugepage[idx], name);
	len = scnprintf(str, sizeof(str), "%llu\n", (unsigned long long)val);
	return simple_read_from_buffer(buf, nbytes, ppos, str, len);
}

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

	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 *cgroup, unsigned int event)
{
	int idx, name, ret = 0;
	struct hugetlb_cgroup *h_cg = hugetlb_cgroup_from_cgroup(cgroup);

	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 = hugetlb_cgroup_read;
	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 = hugetlb_cgroup_read;

	/* 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 = hugetlb_cgroup_read;

	/* 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 = hugetlb_cgroup_read;

	/* 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	}
	41
	42	static inline
	43	struct hugetlb_cgroup hugetlb_cgroup_from_cgroup(struct cgroup cgroup)
	44	{
8af01f56	45	return hugetlb_cgroup_from_css(cgroup_css(cgroup, hugetlb_subsys_id));
2bc64a20 AK	46	}
	47
	48	static inline
	49	struct hugetlb_cgroup hugetlb_cgroup_from_task(struct task_struct task)
	50	{
8af01f56	51	return hugetlb_cgroup_from_css(task_css(task, hugetlb_subsys_id));
2bc64a20 AK	52	}
	53
	54	static inline bool hugetlb_cgroup_is_root(struct hugetlb_cgroup *h_cg)
	55	{
	56	return (h_cg == root_h_cgroup);
	57	}
	58
3f798518 TH	59	static inline struct hugetlb_cgroup *
3f798518 TH	60	parent_hugetlb_cgroup(struct hugetlb_cgroup *h_cg)
2bc64a20	61	{
63876986	62	return hugetlb_cgroup_from_css(css_parent(&h_cg->css));
2bc64a20 AK	63	}
2bc64a20 AK	64
3f798518	65	static inline bool hugetlb_cgroup_have_usage(struct hugetlb_cgroup *h_cg)
2bc64a20 AK	66	{
2bc64a20 AK	67	int idx;
2bc64a20 AK	68
	69	for (idx = 0; idx < hugetlb_max_hstate; idx++) {
	70	if ((res_counter_read_u64(&h_cg->hugepage[idx], RES_USAGE)) > 0)
	71	return true;
	72	}
	73	return false;
	74	}
	75
92fb9748	76	static struct cgroup_subsys_state hugetlb_cgroup_css_alloc(struct cgroup cgroup)
2bc64a20 AK	77	{
	78	int idx;
	79	struct cgroup *parent_cgroup;
	80	struct hugetlb_cgroup h_cgroup, parent_h_cgroup;
	81
	82	h_cgroup = kzalloc(sizeof(*h_cgroup), GFP_KERNEL);
	83	if (!h_cgroup)
	84	return ERR_PTR(-ENOMEM);
	85
	86	parent_cgroup = cgroup->parent;
	87	if (parent_cgroup) {
	88	parent_h_cgroup = hugetlb_cgroup_from_cgroup(parent_cgroup);
	89	for (idx = 0; idx < HUGE_MAX_HSTATE; idx++)
	90	res_counter_init(&h_cgroup->hugepage[idx],
	91	&parent_h_cgroup->hugepage[idx]);
	92	} else {
	93	root_h_cgroup = h_cgroup;
	94	for (idx = 0; idx < HUGE_MAX_HSTATE; idx++)
	95	res_counter_init(&h_cgroup->hugepage[idx], NULL);
	96	}
	97	return &h_cgroup->css;
	98	}
	99
92fb9748	100	static void hugetlb_cgroup_css_free(struct cgroup *cgroup)
2bc64a20 AK	101	{
	102	struct hugetlb_cgroup *h_cgroup;
	103
	104	h_cgroup = hugetlb_cgroup_from_cgroup(cgroup);
	105	kfree(h_cgroup);
	106	}
	107
da1def55 AK	108
	109	/*
	110	* Should be called with hugetlb_lock held.
	111	* Since we are holding hugetlb_lock, pages cannot get moved from
	112	* active list or uncharged from the cgroup, So no need to get
	113	* page reference and test for page active here. This function
	114	* cannot fail.
	115	*/
3f798518	116	static void hugetlb_cgroup_move_parent(int idx, struct hugetlb_cgroup *h_cg,
da1def55 AK	117	struct page *page)
	118	{
	119	int csize;
	120	struct res_counter *counter;
	121	struct res_counter *fail_res;
	122	struct hugetlb_cgroup *page_hcg;
3f798518	123	struct hugetlb_cgroup *parent = parent_hugetlb_cgroup(h_cg);
da1def55 AK	124
	125	page_hcg = hugetlb_cgroup_from_page(page);
	126	/*
	127	* We can have pages in active list without any cgroup
	128	* ie, hugepage with less than 3 pages. We can safely
	129	* ignore those pages.
	130	*/
	131	if (!page_hcg \|\| page_hcg != h_cg)
	132	goto out;
	133
	134	csize = PAGE_SIZE << compound_order(page);
	135	if (!parent) {
	136	parent = root_h_cgroup;
	137	/* root has no limit */
	138	res_counter_charge_nofail(&parent->hugepage[idx],
	139	csize, &fail_res);
	140	}
	141	counter = &h_cg->hugepage[idx];
	142	res_counter_uncharge_until(counter, counter->parent, csize);
	143
	144	set_hugetlb_cgroup(page, parent);
	145	out:
	146	return;
	147	}
	148
	149	/*
	150	* Force the hugetlb cgroup to empty the hugetlb resources by moving them to
	151	* the parent cgroup.
	152	*/
92fb9748	153	static void hugetlb_cgroup_css_offline(struct cgroup *cgroup)
2bc64a20	154	{
3f798518	155	struct hugetlb_cgroup *h_cg = hugetlb_cgroup_from_cgroup(cgroup);
da1def55 AK	156	struct hstate *h;
da1def55 AK	157	struct page *page;
9d093cb1	158	int idx = 0;
da1def55 AK	159
da1def55 AK	160	do {
da1def55 AK	161	for_each_hstate(h) {
	162	spin_lock(&hugetlb_lock);
	163	list_for_each_entry(page, &h->hugepage_activelist, lru)
3f798518	164	hugetlb_cgroup_move_parent(idx, h_cg, page);
da1def55 AK	165
	166	spin_unlock(&hugetlb_lock);
	167	idx++;
	168	}
	169	cond_resched();
3f798518	170	} while (hugetlb_cgroup_have_usage(h_cg));
2bc64a20 AK	171	}
2bc64a20 AK	172
6d76dcf4 AK	173	int hugetlb_cgroup_charge_cgroup(int idx, unsigned long nr_pages,
	174	struct hugetlb_cgroup **ptr)
	175	{
	176	int ret = 0;
	177	struct res_counter *fail_res;
	178	struct hugetlb_cgroup *h_cg = NULL;
	179	unsigned long csize = nr_pages * PAGE_SIZE;
	180
	181	if (hugetlb_cgroup_disabled())
	182	goto done;
	183	/*
	184	* We don't charge any cgroup if the compound page have less
	185	* than 3 pages.
	186	*/
	187	if (huge_page_order(&hstates[idx]) < HUGETLB_CGROUP_MIN_ORDER)
	188	goto done;
	189	again:
	190	rcu_read_lock();
	191	h_cg = hugetlb_cgroup_from_task(current);
	192	if (!css_tryget(&h_cg->css)) {
	193	rcu_read_unlock();
	194	goto again;
	195	}
	196	rcu_read_unlock();
	197
	198	ret = res_counter_charge(&h_cg->hugepage[idx], csize, &fail_res);
	199	css_put(&h_cg->css);
	200	done:
	201	*ptr = h_cg;
	202	return ret;
	203	}
	204
94ae8ba7	205	/* Should be called with hugetlb_lock held */
6d76dcf4 AK	206	void hugetlb_cgroup_commit_charge(int idx, unsigned long nr_pages,
	207	struct hugetlb_cgroup *h_cg,
	208	struct page *page)
	209	{
	210	if (hugetlb_cgroup_disabled() \|\| !h_cg)
	211	return;
	212
6d76dcf4	213	set_hugetlb_cgroup(page, h_cg);
6d76dcf4 AK	214	return;
	215	}
	216
	217	/*
	218	* Should be called with hugetlb_lock held
	219	*/
	220	void hugetlb_cgroup_uncharge_page(int idx, unsigned long nr_pages,
	221	struct page *page)
	222	{
	223	struct hugetlb_cgroup *h_cg;
	224	unsigned long csize = nr_pages * PAGE_SIZE;
	225
	226	if (hugetlb_cgroup_disabled())
	227	return;
	228	VM_BUG_ON(!spin_is_locked(&hugetlb_lock));
	229	h_cg = hugetlb_cgroup_from_page(page);
	230	if (unlikely(!h_cg))
	231	return;
	232	set_hugetlb_cgroup(page, NULL);
	233	res_counter_uncharge(&h_cg->hugepage[idx], csize);
	234	return;
	235	}
	236
	237	void hugetlb_cgroup_uncharge_cgroup(int idx, unsigned long nr_pages,
	238	struct hugetlb_cgroup *h_cg)
	239	{
	240	unsigned long csize = nr_pages * PAGE_SIZE;
	241
	242	if (hugetlb_cgroup_disabled() \|\| !h_cg)
	243	return;
	244
	245	if (huge_page_order(&hstates[idx]) < HUGETLB_CGROUP_MIN_ORDER)
	246	return;
	247
	248	res_counter_uncharge(&h_cg->hugepage[idx], csize);
	249	return;
	250	}
	251
abb8206c AK	252	static ssize_t hugetlb_cgroup_read(struct cgroup cgroup, struct cftype cft,
	253	struct file file, char __user buf,
	254	size_t nbytes, loff_t *ppos)
	255	{
	256	u64 val;
	257	char str[64];
	258	int idx, name, len;
	259	struct hugetlb_cgroup *h_cg = hugetlb_cgroup_from_cgroup(cgroup);
	260
	261	idx = MEMFILE_IDX(cft->private);
	262	name = MEMFILE_ATTR(cft->private);
	263
	264	val = res_counter_read_u64(&h_cg->hugepage[idx], name);
	265	len = scnprintf(str, sizeof(str), "%llu\n", (unsigned long long)val);
	266	return simple_read_from_buffer(buf, nbytes, ppos, str, len);
	267	}
	268
	269	static int hugetlb_cgroup_write(struct cgroup cgroup, struct cftype cft,
	270	const char *buffer)
	271	{
	272	int idx, name, ret;
	273	unsigned long long val;
	274	struct hugetlb_cgroup *h_cg = hugetlb_cgroup_from_cgroup(cgroup);
	275
	276	idx = MEMFILE_IDX(cft->private);
	277	name = MEMFILE_ATTR(cft->private);
	278
	279	switch (name) {
	280	case RES_LIMIT:
	281	if (hugetlb_cgroup_is_root(h_cg)) {
	282	/* Can't set limit on root */
	283	ret = -EINVAL;
	284	break;
	285	}
	286	/* This function does all necessary parse...reuse it */
	287	ret = res_counter_memparse_write_strategy(buffer, &val);
	288	if (ret)
	289	break;
	290	ret = res_counter_set_limit(&h_cg->hugepage[idx], val);
	291	break;
	292	default:
	293	ret = -EINVAL;
	294	break;
	295	}
	296	return ret;
	297	}
	298
	299	static int hugetlb_cgroup_reset(struct cgroup *cgroup, unsigned int event)
	300	{
	301	int idx, name, ret = 0;
	302	struct hugetlb_cgroup *h_cg = hugetlb_cgroup_from_cgroup(cgroup);
	303
	304	idx = MEMFILE_IDX(event);
	305	name = MEMFILE_ATTR(event);
	306
	307	switch (name) {
	308	case RES_MAX_USAGE:
	309	res_counter_reset_max(&h_cg->hugepage[idx]);
	310	break;
	311	case RES_FAILCNT:
	312	res_counter_reset_failcnt(&h_cg->hugepage[idx]);
	313	break;
	314	default:
	315	ret = -EINVAL;
316	break;
317	}
318	return ret;
319	}
320
321	static char mem_fmt(char buf, int size, unsigned long hsize)
322	{
323	if (hsize >= (1UL << 30))
324	snprintf(buf, size, "%luGB", hsize >> 30);
325	else if (hsize >= (1UL << 20))
326	snprintf(buf, size, "%luMB", hsize >> 20);
327	else
328	snprintf(buf, size, "%luKB", hsize >> 10);
329	return buf;
330	}
331
7179e7bf	332	static void __init __hugetlb_cgroup_file_init(int idx)
abb8206c AK	333	{
	334	char buf[32];
	335	struct cftype *cft;
	336	struct hstate *h = &hstates[idx];
	337
	338	/* format the size */
	339	mem_fmt(buf, 32, huge_page_size(h));
	340
	341	/* Add the limit file */
	342	cft = &h->cgroup_files[0];
	343	snprintf(cft->name, MAX_CFTYPE_NAME, "%s.limit_in_bytes", buf);
	344	cft->private = MEMFILE_PRIVATE(idx, RES_LIMIT);
	345	cft->read = hugetlb_cgroup_read;
	346	cft->write_string = hugetlb_cgroup_write;
	347
	348	/* Add the usage file */
	349	cft = &h->cgroup_files[1];
	350	snprintf(cft->name, MAX_CFTYPE_NAME, "%s.usage_in_bytes", buf);
	351	cft->private = MEMFILE_PRIVATE(idx, RES_USAGE);
	352	cft->read = hugetlb_cgroup_read;
	353
	354	/* Add the MAX usage file */
	355	cft = &h->cgroup_files[2];
	356	snprintf(cft->name, MAX_CFTYPE_NAME, "%s.max_usage_in_bytes", buf);
	357	cft->private = MEMFILE_PRIVATE(idx, RES_MAX_USAGE);
	358	cft->trigger = hugetlb_cgroup_reset;
	359	cft->read = hugetlb_cgroup_read;
	360
	361	/* Add the failcntfile */
	362	cft = &h->cgroup_files[3];
	363	snprintf(cft->name, MAX_CFTYPE_NAME, "%s.failcnt", buf);
	364	cft->private = MEMFILE_PRIVATE(idx, RES_FAILCNT);
	365	cft->trigger = hugetlb_cgroup_reset;
	366	cft->read = hugetlb_cgroup_read;
	367
	368	/* NULL terminate the last cft */
	369	cft = &h->cgroup_files[4];
	370	memset(cft, 0, sizeof(*cft));
	371
	372	WARN_ON(cgroup_add_cftypes(&hugetlb_subsys, h->cgroup_files));
	373
7179e7bf JW	374	return;
	375	}
	376
	377	void __init hugetlb_cgroup_file_init(void)
	378	{
	379	struct hstate *h;
	380
	381	for_each_hstate(h) {
	382	/*
	383	* Add cgroup control files only if the huge page consists
	384	* of more than two normal pages. This is because we use
	385	* page[2].lru.next for storing cgroup details.
	386	*/
	387	if (huge_page_order(h) >= HUGETLB_CGROUP_MIN_ORDER)
	388	__hugetlb_cgroup_file_init(hstate_index(h));
	389	}
abb8206c AK	390	}
abb8206c AK	391
75754681 AK	392	/*
	393	* hugetlb_lock will make sure a parallel cgroup rmdir won't happen
	394	* when we migrate hugepages
	395	*/
8e6ac7fa AK	396	void hugetlb_cgroup_migrate(struct page oldhpage, struct page newhpage)
	397	{
	398	struct hugetlb_cgroup *h_cg;
94ae8ba7	399	struct hstate *h = page_hstate(oldhpage);
8e6ac7fa AK	400
	401	if (hugetlb_cgroup_disabled())
	402	return;
	403
	404	VM_BUG_ON(!PageHuge(oldhpage));
	405	spin_lock(&hugetlb_lock);
	406	h_cg = hugetlb_cgroup_from_page(oldhpage);
	407	set_hugetlb_cgroup(oldhpage, NULL);
8e6ac7fa AK	408
	409	/* move the h_cg details to new cgroup */
	410	set_hugetlb_cgroup(newhpage, h_cg);
94ae8ba7	411	list_move(&newhpage->lru, &h->hugepage_activelist);
8e6ac7fa	412	spin_unlock(&hugetlb_lock);
8e6ac7fa AK	413	return;
	414	}
	415
2bc64a20 AK	416	struct cgroup_subsys hugetlb_subsys = {
2bc64a20 AK	417	.name = "hugetlb",
92fb9748 TH	418	.css_alloc = hugetlb_cgroup_css_alloc,
	419	.css_offline = hugetlb_cgroup_css_offline,
	420	.css_free = hugetlb_cgroup_css_free,
	421	.subsys_id = hugetlb_subsys_id,
2bc64a20	422	};