[mirror_ubuntu-artful-kernel.git] / mm / page_cgroup.c

#include <linux/mm.h>
#include <linux/mmzone.h>
#include <linux/bootmem.h>
#include <linux/bit_spinlock.h>
#include <linux/page_cgroup.h>
#include <linux/hash.h>
#include <linux/slab.h>
#include <linux/memory.h>
#include <linux/vmalloc.h>
#include <linux/cgroup.h>
#include <linux/swapops.h>
#include <linux/kmemleak.h>

static void __meminit init_page_cgroup(struct page_cgroup *pc, unsigned long id)
{
	pc->flags = 0;
	set_page_cgroup_array_id(pc, id);
	pc->mem_cgroup = NULL;
	INIT_LIST_HEAD(&pc->lru);
}
static unsigned long total_usage;

#if !defined(CONFIG_SPARSEMEM)


void __meminit pgdat_page_cgroup_init(struct pglist_data *pgdat)
{
	pgdat->node_page_cgroup = NULL;
}

struct page_cgroup *lookup_page_cgroup(struct page *page)
{
	unsigned long pfn = page_to_pfn(page);
	unsigned long offset;
	struct page_cgroup *base;

	base = NODE_DATA(page_to_nid(page))->node_page_cgroup;
	if (unlikely(!base))
		return NULL;

	offset = pfn - NODE_DATA(page_to_nid(page))->node_start_pfn;
	return base + offset;
}

struct page *lookup_cgroup_page(struct page_cgroup *pc)
{
	unsigned long pfn;
	struct page *page;
	pg_data_t *pgdat;

	pgdat = NODE_DATA(page_cgroup_array_id(pc));
	pfn = pc - pgdat->node_page_cgroup + pgdat->node_start_pfn;
	page = pfn_to_page(pfn);
	VM_BUG_ON(pc != lookup_page_cgroup(page));
	return page;
}

static int __init alloc_node_page_cgroup(int nid)
{
	struct page_cgroup *base, *pc;
	unsigned long table_size;
	unsigned long start_pfn, nr_pages, index;

	start_pfn = NODE_DATA(nid)->node_start_pfn;
	nr_pages = NODE_DATA(nid)->node_spanned_pages;

	if (!nr_pages)
		return 0;

	table_size = sizeof(struct page_cgroup) * nr_pages;

	base = __alloc_bootmem_node_nopanic(NODE_DATA(nid),
			table_size, PAGE_SIZE, __pa(MAX_DMA_ADDRESS));
	if (!base)
		return -ENOMEM;
	for (index = 0; index < nr_pages; index++) {
		pc = base + index;
		init_page_cgroup(pc, nid);
	}
	NODE_DATA(nid)->node_page_cgroup = base;
	total_usage += table_size;
	return 0;
}

void __init page_cgroup_init_flatmem(void)
{

	int nid, fail;

	if (mem_cgroup_disabled())
		return;

	for_each_online_node(nid)  {
		fail = alloc_node_page_cgroup(nid);
		if (fail)
			goto fail;
	}
	printk(KERN_INFO "allocated %ld bytes of page_cgroup\n", total_usage);
	printk(KERN_INFO "please try 'cgroup_disable=memory' option if you"
	" don't want memory cgroups\n");
	return;
fail:
	printk(KERN_CRIT "allocation of page_cgroup failed.\n");
	printk(KERN_CRIT "please try 'cgroup_disable=memory' boot option\n");
	panic("Out of memory");
}

#else /* CONFIG_FLAT_NODE_MEM_MAP */

struct page_cgroup *lookup_page_cgroup(struct page *page)
{
	unsigned long pfn = page_to_pfn(page);
	struct mem_section *section = __pfn_to_section(pfn);

	if (!section->page_cgroup)
		return NULL;
	return section->page_cgroup + pfn;
}

struct page *lookup_cgroup_page(struct page_cgroup *pc)
{
	struct mem_section *section;
	struct page *page;
	unsigned long nr;

	nr = page_cgroup_array_id(pc);
	section = __nr_to_section(nr);
	page = pfn_to_page(pc - section->page_cgroup);
	VM_BUG_ON(pc != lookup_page_cgroup(page));
	return page;
}

static void *__meminit alloc_page_cgroup(size_t size, int nid)
{
	void *addr = NULL;
	gfp_t flags = GFP_KERNEL | __GFP_NOWARN;

	addr = alloc_pages_exact_nid(nid, size, flags);
	if (addr) {
		kmemleak_alloc(addr, size, 1, flags);
		return addr;
	}

	if (node_state(nid, N_HIGH_MEMORY))
		addr = vmalloc_node(size, nid);
	else
		addr = vmalloc(size);

	return addr;
}

#ifdef CONFIG_MEMORY_HOTPLUG
static void free_page_cgroup(void *addr)
{
	if (is_vmalloc_addr(addr)) {
		vfree(addr);
	} else {
		struct page *page = virt_to_page(addr);
		size_t table_size =
			sizeof(struct page_cgroup) * PAGES_PER_SECTION;

		BUG_ON(PageReserved(page));
		free_pages_exact(addr, table_size);
	}
}
#endif

static int __meminit init_section_page_cgroup(unsigned long pfn, int nid)
{
	struct page_cgroup *base, *pc;
	struct mem_section *section;
	unsigned long table_size;
	unsigned long nr;
	int index;

	nr = pfn_to_section_nr(pfn);
	section = __nr_to_section(nr);

	if (section->page_cgroup)
		return 0;

	table_size = sizeof(struct page_cgroup) * PAGES_PER_SECTION;
	base = alloc_page_cgroup(table_size, nid);

	/*
	 * The value stored in section->page_cgroup is (base - pfn)
	 * and it does not point to the memory block allocated above,
	 * causing kmemleak false positives.
	 */
	kmemleak_not_leak(base);

	if (!base) {
		printk(KERN_ERR "page cgroup allocation failure\n");
		return -ENOMEM;
	}

	for (index = 0; index < PAGES_PER_SECTION; index++) {
		pc = base + index;
		init_page_cgroup(pc, nr);
	}
	/*
	 * The passed "pfn" may not be aligned to SECTION.  For the calculation
	 * we need to apply a mask.
	 */
	pfn &= PAGE_SECTION_MASK;
	section->page_cgroup = base - pfn;
	total_usage += table_size;
	return 0;
}
#ifdef CONFIG_MEMORY_HOTPLUG
void __free_page_cgroup(unsigned long pfn)
{
	struct mem_section *ms;
	struct page_cgroup *base;

	ms = __pfn_to_section(pfn);
	if (!ms || !ms->page_cgroup)
		return;
	base = ms->page_cgroup + pfn;
	free_page_cgroup(base);
	ms->page_cgroup = NULL;
}

int __meminit online_page_cgroup(unsigned long start_pfn,
			unsigned long nr_pages,
			int nid)
{
	unsigned long start, end, pfn;
	int fail = 0;

	start = SECTION_ALIGN_DOWN(start_pfn);
	end = SECTION_ALIGN_UP(start_pfn + nr_pages);

	if (nid == -1) {
		/*
		 * In this case, "nid" already exists and contains valid memory.
		 * "start_pfn" passed to us is a pfn which is an arg for
		 * online__pages(), and start_pfn should exist.
		 */
		nid = pfn_to_nid(start_pfn);
		VM_BUG_ON(!node_state(nid, N_ONLINE));
	}

	for (pfn = start; !fail && pfn < end; pfn += PAGES_PER_SECTION) {
		if (!pfn_present(pfn))
			continue;
		fail = init_section_page_cgroup(pfn, nid);
	}
	if (!fail)
		return 0;

	/* rollback */
	for (pfn = start; pfn < end; pfn += PAGES_PER_SECTION)
		__free_page_cgroup(pfn);

	return -ENOMEM;
}

int __meminit offline_page_cgroup(unsigned long start_pfn,
		unsigned long nr_pages, int nid)
{
	unsigned long start, end, pfn;

	start = SECTION_ALIGN_DOWN(start_pfn);
	end = SECTION_ALIGN_UP(start_pfn + nr_pages);

	for (pfn = start; pfn < end; pfn += PAGES_PER_SECTION)
		__free_page_cgroup(pfn);
	return 0;

}

static int __meminit page_cgroup_callback(struct notifier_block *self,
			       unsigned long action, void *arg)
{
	struct memory_notify *mn = arg;
	int ret = 0;
	switch (action) {
	case MEM_GOING_ONLINE:
		ret = online_page_cgroup(mn->start_pfn,
				   mn->nr_pages, mn->status_change_nid);
		break;
	case MEM_OFFLINE:
		offline_page_cgroup(mn->start_pfn,
				mn->nr_pages, mn->status_change_nid);
		break;
	case MEM_CANCEL_ONLINE:
	case MEM_GOING_OFFLINE:
		break;
	case MEM_ONLINE:
	case MEM_CANCEL_OFFLINE:
		break;
	}

	return notifier_from_errno(ret);
}

#endif

void __init page_cgroup_init(void)
{
	unsigned long pfn;
	int nid;

	if (mem_cgroup_disabled())
		return;

	for_each_node_state(nid, N_HIGH_MEMORY) {
		unsigned long start_pfn, end_pfn;

		start_pfn = node_start_pfn(nid);
		end_pfn = node_end_pfn(nid);
		/*
		 * start_pfn and end_pfn may not be aligned to SECTION and the
		 * page->flags of out of node pages are not initialized.  So we
		 * scan [start_pfn, the biggest section's pfn < end_pfn) here.
		 */
		for (pfn = start_pfn;
		     pfn < end_pfn;
                     pfn = ALIGN(pfn + 1, PAGES_PER_SECTION)) {

			if (!pfn_valid(pfn))
				continue;
			/*
			 * Nodes's pfns can be overlapping.
			 * We know some arch can have a nodes layout such as
			 * -------------pfn-------------->
			 * N0 | N1 | N2 | N0 | N1 | N2|....
			 */
			if (pfn_to_nid(pfn) != nid)
				continue;
			if (init_section_page_cgroup(pfn, nid))
				goto oom;
		}
	}
	hotplug_memory_notifier(page_cgroup_callback, 0);
	printk(KERN_INFO "allocated %ld bytes of page_cgroup\n", total_usage);
	printk(KERN_INFO "please try 'cgroup_disable=memory' option if you "
			 "don't want memory cgroups\n");
	return;
oom:
	printk(KERN_CRIT "try 'cgroup_disable=memory' boot option\n");
	panic("Out of memory");
}

void __meminit pgdat_page_cgroup_init(struct pglist_data *pgdat)
{
	return;
}

#endif


#ifdef CONFIG_CGROUP_MEM_RES_CTLR_SWAP

static DEFINE_MUTEX(swap_cgroup_mutex);
struct swap_cgroup_ctrl {
	struct page **map;
	unsigned long length;
	spinlock_t	lock;
};

struct swap_cgroup_ctrl swap_cgroup_ctrl[MAX_SWAPFILES];

struct swap_cgroup {
	unsigned short		id;
};
#define SC_PER_PAGE	(PAGE_SIZE/sizeof(struct swap_cgroup))
#define SC_POS_MASK	(SC_PER_PAGE - 1)

/*
 * SwapCgroup implements "lookup" and "exchange" operations.
 * In typical usage, this swap_cgroup is accessed via memcg's charge/uncharge
 * against SwapCache. At swap_free(), this is accessed directly from swap.
 *
 * This means,
 *  - we have no race in "exchange" when we're accessed via SwapCache because
 *    SwapCache(and its swp_entry) is under lock.
 *  - When called via swap_free(), there is no user of this entry and no race.
 * Then, we don't need lock around "exchange".
 *
 * TODO: we can push these buffers out to HIGHMEM.
 */

/*
 * allocate buffer for swap_cgroup.
 */
static int swap_cgroup_prepare(int type)
{
	struct page *page;
	struct swap_cgroup_ctrl *ctrl;
	unsigned long idx, max;

	ctrl = &swap_cgroup_ctrl[type];

	for (idx = 0; idx < ctrl->length; idx++) {
		page = alloc_page(GFP_KERNEL | __GFP_ZERO);
		if (!page)
			goto not_enough_page;
		ctrl->map[idx] = page;
	}
	return 0;
not_enough_page:
	max = idx;
	for (idx = 0; idx < max; idx++)
		__free_page(ctrl->map[idx]);

	return -ENOMEM;
}

/**
 * swap_cgroup_cmpxchg - cmpxchg mem_cgroup's id for this swp_entry.
 * @end: swap entry to be cmpxchged
 * @old: old id
 * @new: new id
 *
 * Returns old id at success, 0 at failure.
 * (There is no mem_cgroup using 0 as its id)
 */
unsigned short swap_cgroup_cmpxchg(swp_entry_t ent,
					unsigned short old, unsigned short new)
{
	int type = swp_type(ent);
	unsigned long offset = swp_offset(ent);
	unsigned long idx = offset / SC_PER_PAGE;
	unsigned long pos = offset & SC_POS_MASK;
	struct swap_cgroup_ctrl *ctrl;
	struct page *mappage;
	struct swap_cgroup *sc;
	unsigned long flags;
	unsigned short retval;

	ctrl = &swap_cgroup_ctrl[type];

	mappage = ctrl->map[idx];
	sc = page_address(mappage);
	sc += pos;
	spin_lock_irqsave(&ctrl->lock, flags);
	retval = sc->id;
	if (retval == old)
		sc->id = new;
	else
		retval = 0;
	spin_unlock_irqrestore(&ctrl->lock, flags);
	return retval;
}

/**
 * swap_cgroup_record - record mem_cgroup for this swp_entry.
 * @ent: swap entry to be recorded into
 * @mem: mem_cgroup to be recorded
 *
 * Returns old value at success, 0 at failure.
 * (Of course, old value can be 0.)
 */
unsigned short swap_cgroup_record(swp_entry_t ent, unsigned short id)
{
	int type = swp_type(ent);
	unsigned long offset = swp_offset(ent);
	unsigned long idx = offset / SC_PER_PAGE;
	unsigned long pos = offset & SC_POS_MASK;
	struct swap_cgroup_ctrl *ctrl;
	struct page *mappage;
	struct swap_cgroup *sc;
	unsigned short old;
	unsigned long flags;

	ctrl = &swap_cgroup_ctrl[type];

	mappage = ctrl->map[idx];
	sc = page_address(mappage);
	sc += pos;
	spin_lock_irqsave(&ctrl->lock, flags);
	old = sc->id;
	sc->id = id;
	spin_unlock_irqrestore(&ctrl->lock, flags);

	return old;
}

/**
 * lookup_swap_cgroup - lookup mem_cgroup tied to swap entry
 * @ent: swap entry to be looked up.
 *
 * Returns CSS ID of mem_cgroup at success. 0 at failure. (0 is invalid ID)
 */
unsigned short lookup_swap_cgroup(swp_entry_t ent)
{
	int type = swp_type(ent);
	unsigned long offset = swp_offset(ent);
	unsigned long idx = offset / SC_PER_PAGE;
	unsigned long pos = offset & SC_POS_MASK;
	struct swap_cgroup_ctrl *ctrl;
	struct page *mappage;
	struct swap_cgroup *sc;
	unsigned short ret;

	ctrl = &swap_cgroup_ctrl[type];
	mappage = ctrl->map[idx];
	sc = page_address(mappage);
	sc += pos;
	ret = sc->id;
	return ret;
}

int swap_cgroup_swapon(int type, unsigned long max_pages)
{
	void *array;
	unsigned long array_size;
	unsigned long length;
	struct swap_cgroup_ctrl *ctrl;

	if (!do_swap_account)
		return 0;

	length = DIV_ROUND_UP(max_pages, SC_PER_PAGE);
	array_size = length * sizeof(void *);

	array = vzalloc(array_size);
	if (!array)
		goto nomem;

	ctrl = &swap_cgroup_ctrl[type];
	mutex_lock(&swap_cgroup_mutex);
	ctrl->length = length;
	ctrl->map = array;
	spin_lock_init(&ctrl->lock);
	if (swap_cgroup_prepare(type)) {
		/* memory shortage */
		ctrl->map = NULL;
		ctrl->length = 0;
		mutex_unlock(&swap_cgroup_mutex);
		vfree(array);
		goto nomem;
	}
	mutex_unlock(&swap_cgroup_mutex);

	return 0;
nomem:
	printk(KERN_INFO "couldn't allocate enough memory for swap_cgroup.\n");
	printk(KERN_INFO
		"swap_cgroup can be disabled by swapaccount=0 boot option\n");
	return -ENOMEM;
}

void swap_cgroup_swapoff(int type)
{
	struct page **map;
	unsigned long i, length;
	struct swap_cgroup_ctrl *ctrl;

	if (!do_swap_account)
		return;

	mutex_lock(&swap_cgroup_mutex);
	ctrl = &swap_cgroup_ctrl[type];
	map = ctrl->map;
	length = ctrl->length;
	ctrl->map = NULL;
	ctrl->length = 0;
	mutex_unlock(&swap_cgroup_mutex);

	if (map) {
		for (i = 0; i < length; i++) {
			struct page *page = map[i];
			if (page)
				__free_page(page);
		}
		vfree(map);
	}
}

#endif
Commit	Line	Data
52d4b9ac KH	1	#include <linux/mm.h>
	2	#include <linux/mmzone.h>
	3	#include <linux/bootmem.h>
	4	#include <linux/bit_spinlock.h>
	5	#include <linux/page_cgroup.h>
	6	#include <linux/hash.h>
94b6da5a	7	#include <linux/slab.h>
52d4b9ac	8	#include <linux/memory.h>
4c821042	9	#include <linux/vmalloc.h>
94b6da5a	10	#include <linux/cgroup.h>
27a7faa0	11	#include <linux/swapops.h>
7952f988	12	#include <linux/kmemleak.h>
52d4b9ac	13
6b3ae58e	14	static void __meminit init_page_cgroup(struct page_cgroup *pc, unsigned long id)
52d4b9ac KH	15	{
52d4b9ac KH	16	pc->flags = 0;
6b3ae58e	17	set_page_cgroup_array_id(pc, id);
52d4b9ac	18	pc->mem_cgroup = NULL;
08e552c6	19	INIT_LIST_HEAD(&pc->lru);
52d4b9ac KH	20	}
	21	static unsigned long total_usage;
	22
	23	#if !defined(CONFIG_SPARSEMEM)
	24
	25
31168481	26	void __meminit pgdat_page_cgroup_init(struct pglist_data *pgdat)
52d4b9ac KH	27	{
	28	pgdat->node_page_cgroup = NULL;
	29	}
	30
	31	struct page_cgroup lookup_page_cgroup(struct page page)
	32	{
	33	unsigned long pfn = page_to_pfn(page);
	34	unsigned long offset;
	35	struct page_cgroup *base;
	36
	37	base = NODE_DATA(page_to_nid(page))->node_page_cgroup;
	38	if (unlikely(!base))
	39	return NULL;
	40
	41	offset = pfn - NODE_DATA(page_to_nid(page))->node_start_pfn;
	42	return base + offset;
	43	}
	44
6b3ae58e JW	45	struct page lookup_cgroup_page(struct page_cgroup pc)
	46	{
	47	unsigned long pfn;
	48	struct page *page;
	49	pg_data_t *pgdat;
	50
	51	pgdat = NODE_DATA(page_cgroup_array_id(pc));
	52	pfn = pc - pgdat->node_page_cgroup + pgdat->node_start_pfn;
	53	page = pfn_to_page(pfn);
	54	VM_BUG_ON(pc != lookup_page_cgroup(page));
	55	return page;
	56	}
	57
52d4b9ac KH	58	static int __init alloc_node_page_cgroup(int nid)
	59	{
	60	struct page_cgroup base, pc;
	61	unsigned long table_size;
	62	unsigned long start_pfn, nr_pages, index;
	63
	64	start_pfn = NODE_DATA(nid)->node_start_pfn;
	65	nr_pages = NODE_DATA(nid)->node_spanned_pages;
	66
653d22c0 KH	67	if (!nr_pages)
	68	return 0;
	69
52d4b9ac	70	table_size = sizeof(struct page_cgroup) * nr_pages;
ca371c0d KH	71
	72	base = __alloc_bootmem_node_nopanic(NODE_DATA(nid),
	73	table_size, PAGE_SIZE, __pa(MAX_DMA_ADDRESS));
	74	if (!base)
52d4b9ac KH	75	return -ENOMEM;
	76	for (index = 0; index < nr_pages; index++) {
	77	pc = base + index;
6b3ae58e	78	init_page_cgroup(pc, nid);
52d4b9ac KH	79	}
	80	NODE_DATA(nid)->node_page_cgroup = base;
	81	total_usage += table_size;
	82	return 0;
	83	}
	84
ca371c0d	85	void __init page_cgroup_init_flatmem(void)
52d4b9ac KH	86	{
	87
	88	int nid, fail;
	89
f8d66542	90	if (mem_cgroup_disabled())
94b6da5a KH	91	return;
94b6da5a KH	92
52d4b9ac KH	93	for_each_online_node(nid) {
	94	fail = alloc_node_page_cgroup(nid);
	95	if (fail)
	96	goto fail;
	97	}
	98	printk(KERN_INFO "allocated %ld bytes of page_cgroup\n", total_usage);
8ca739e3 RD	99	printk(KERN_INFO "please try 'cgroup_disable=memory' option if you"
8ca739e3 RD	100	" don't want memory cgroups\n");
52d4b9ac KH	101	return;
52d4b9ac KH	102	fail:
8ca739e3 RD	103	printk(KERN_CRIT "allocation of page_cgroup failed.\n");
8ca739e3 RD	104	printk(KERN_CRIT "please try 'cgroup_disable=memory' boot option\n");
52d4b9ac KH	105	panic("Out of memory");
	106	}
	107
	108	#else /* CONFIG_FLAT_NODE_MEM_MAP */
	109
	110	struct page_cgroup lookup_page_cgroup(struct page page)
	111	{
	112	unsigned long pfn = page_to_pfn(page);
	113	struct mem_section *section = __pfn_to_section(pfn);
	114
d69b042f BS	115	if (!section->page_cgroup)
d69b042f BS	116	return NULL;
52d4b9ac KH	117	return section->page_cgroup + pfn;
	118	}
	119
6b3ae58e JW	120	struct page lookup_cgroup_page(struct page_cgroup pc)
	121	{
	122	struct mem_section *section;
	123	struct page *page;
	124	unsigned long nr;
	125
	126	nr = page_cgroup_array_id(pc);
	127	section = __nr_to_section(nr);
	128	page = pfn_to_page(pc - section->page_cgroup);
	129	VM_BUG_ON(pc != lookup_page_cgroup(page));
	130	return page;
	131	}
	132
268433b8	133	static void *__meminit alloc_page_cgroup(size_t size, int nid)
dde79e00 MH	134	{
dde79e00 MH	135	void *addr = NULL;
ff7ee93f	136	gfp_t flags = GFP_KERNEL \| __GFP_NOWARN;
dde79e00	137
ff7ee93f SR	138	addr = alloc_pages_exact_nid(nid, size, flags);
	139	if (addr) {
	140	kmemleak_alloc(addr, size, 1, flags);
dde79e00	141	return addr;
ff7ee93f	142	}
dde79e00 MH	143
	144	if (node_state(nid, N_HIGH_MEMORY))
	145	addr = vmalloc_node(size, nid);
	146	else
	147	addr = vmalloc(size);
	148
	149	return addr;
	150	}
	151
	152	#ifdef CONFIG_MEMORY_HOTPLUG
	153	static void free_page_cgroup(void *addr)
	154	{
	155	if (is_vmalloc_addr(addr)) {
	156	vfree(addr);
	157	} else {
	158	struct page *page = virt_to_page(addr);
6cfddb26 MH	159	size_t table_size =
	160	sizeof(struct page_cgroup) * PAGES_PER_SECTION;
	161
	162	BUG_ON(PageReserved(page));
	163	free_pages_exact(addr, table_size);
dde79e00 MH	164	}
	165	}
	166	#endif
	167
37573e8c	168	static int __meminit init_section_page_cgroup(unsigned long pfn, int nid)
52d4b9ac	169	{
52d4b9ac	170	struct page_cgroup base, pc;
6b3ae58e	171	struct mem_section *section;
52d4b9ac	172	unsigned long table_size;
6b3ae58e	173	unsigned long nr;
37573e8c	174	int index;
52d4b9ac	175
6b3ae58e JW	176	nr = pfn_to_section_nr(pfn);
	177	section = __nr_to_section(nr);
	178
	179	if (section->page_cgroup)
	180	return 0;
	181
6b3ae58e	182	table_size = sizeof(struct page_cgroup) * PAGES_PER_SECTION;
dde79e00 MH	183	base = alloc_page_cgroup(table_size, nid);
dde79e00 MH	184
6b3ae58e JW	185	/*
	186	* The value stored in section->page_cgroup is (base - pfn)
	187	* and it does not point to the memory block allocated above,
	188	* causing kmemleak false positives.
	189	*/
	190	kmemleak_not_leak(base);
52d4b9ac KH	191
	192	if (!base) {
	193	printk(KERN_ERR "page cgroup allocation failure\n");
	194	return -ENOMEM;
	195	}
	196
	197	for (index = 0; index < PAGES_PER_SECTION; index++) {
	198	pc = base + index;
6b3ae58e	199	init_page_cgroup(pc, nr);
52d4b9ac	200	}
37573e8c KH	201	/*
	202	* The passed "pfn" may not be aligned to SECTION. For the calculation
	203	* we need to apply a mask.
	204	*/
	205	pfn &= PAGE_SECTION_MASK;
52d4b9ac KH	206	section->page_cgroup = base - pfn;
	207	total_usage += table_size;
	208	return 0;
	209	}
	210	#ifdef CONFIG_MEMORY_HOTPLUG
	211	void __free_page_cgroup(unsigned long pfn)
	212	{
	213	struct mem_section *ms;
	214	struct page_cgroup *base;
	215
	216	ms = __pfn_to_section(pfn);
	217	if (!ms \|\| !ms->page_cgroup)
	218	return;
	219	base = ms->page_cgroup + pfn;
dde79e00 MH	220	free_page_cgroup(base);
dde79e00 MH	221	ms->page_cgroup = NULL;
52d4b9ac KH	222	}
52d4b9ac KH	223
31168481	224	int __meminit online_page_cgroup(unsigned long start_pfn,
52d4b9ac KH	225	unsigned long nr_pages,
	226	int nid)
	227	{
	228	unsigned long start, end, pfn;
	229	int fail = 0;
	230
1bb36fbd DK	231	start = SECTION_ALIGN_DOWN(start_pfn);
1bb36fbd DK	232	end = SECTION_ALIGN_UP(start_pfn + nr_pages);
52d4b9ac	233
37573e8c KH	234	if (nid == -1) {
	235	/*
	236	* In this case, "nid" already exists and contains valid memory.
	237	* "start_pfn" passed to us is a pfn which is an arg for
	238	* online__pages(), and start_pfn should exist.
	239	*/
	240	nid = pfn_to_nid(start_pfn);
	241	VM_BUG_ON(!node_state(nid, N_ONLINE));
	242	}
	243
52d4b9ac KH	244	for (pfn = start; !fail && pfn < end; pfn += PAGES_PER_SECTION) {
	245	if (!pfn_present(pfn))
	246	continue;
37573e8c	247	fail = init_section_page_cgroup(pfn, nid);
52d4b9ac KH	248	}
	249	if (!fail)
	250	return 0;
	251
	252	/* rollback */
	253	for (pfn = start; pfn < end; pfn += PAGES_PER_SECTION)
	254	__free_page_cgroup(pfn);
	255
	256	return -ENOMEM;
	257	}
	258
31168481	259	int __meminit offline_page_cgroup(unsigned long start_pfn,
52d4b9ac KH	260	unsigned long nr_pages, int nid)
	261	{
	262	unsigned long start, end, pfn;
	263
1bb36fbd DK	264	start = SECTION_ALIGN_DOWN(start_pfn);
1bb36fbd DK	265	end = SECTION_ALIGN_UP(start_pfn + nr_pages);
52d4b9ac KH	266
	267	for (pfn = start; pfn < end; pfn += PAGES_PER_SECTION)
	268	__free_page_cgroup(pfn);
	269	return 0;
	270
	271	}
	272
31168481	273	static int __meminit page_cgroup_callback(struct notifier_block *self,
52d4b9ac KH	274	unsigned long action, void *arg)
	275	{
	276	struct memory_notify *mn = arg;
	277	int ret = 0;
	278	switch (action) {
	279	case MEM_GOING_ONLINE:
	280	ret = online_page_cgroup(mn->start_pfn,
	281	mn->nr_pages, mn->status_change_nid);
	282	break;
52d4b9ac KH	283	case MEM_OFFLINE:
	284	offline_page_cgroup(mn->start_pfn,
	285	mn->nr_pages, mn->status_change_nid);
	286	break;
dc19f9db	287	case MEM_CANCEL_ONLINE:
52d4b9ac KH	288	case MEM_GOING_OFFLINE:
	289	break;
	290	case MEM_ONLINE:
	291	case MEM_CANCEL_OFFLINE:
	292	break;
	293	}
dc19f9db	294
5fda1bd5	295	return notifier_from_errno(ret);
52d4b9ac KH	296	}
	297
	298	#endif
	299
	300	void __init page_cgroup_init(void)
	301	{
	302	unsigned long pfn;
37573e8c	303	int nid;
52d4b9ac	304
f8d66542	305	if (mem_cgroup_disabled())
94b6da5a KH	306	return;
94b6da5a KH	307
37573e8c KH	308	for_each_node_state(nid, N_HIGH_MEMORY) {
	309	unsigned long start_pfn, end_pfn;
	310
	311	start_pfn = node_start_pfn(nid);
	312	end_pfn = node_end_pfn(nid);
	313	/*
	314	* start_pfn and end_pfn may not be aligned to SECTION and the
	315	* page->flags of out of node pages are not initialized. So we
	316	* scan [start_pfn, the biggest section's pfn < end_pfn) here.
	317	*/
	318	for (pfn = start_pfn;
	319	pfn < end_pfn;
	320	pfn = ALIGN(pfn + 1, PAGES_PER_SECTION)) {
	321
	322	if (!pfn_valid(pfn))
	323	continue;
	324	/*
	325	* Nodes's pfns can be overlapping.
	326	* We know some arch can have a nodes layout such as
	327	* -------------pfn-------------->
	328	* N0 \| N1 \| N2 \| N0 \| N1 \| N2\|....
	329	*/
	330	if (pfn_to_nid(pfn) != nid)
	331	continue;
	332	if (init_section_page_cgroup(pfn, nid))
	333	goto oom;
	334	}
52d4b9ac	335	}
37573e8c	336	hotplug_memory_notifier(page_cgroup_callback, 0);
52d4b9ac	337	printk(KERN_INFO "allocated %ld bytes of page_cgroup\n", total_usage);
37573e8c KH	338	printk(KERN_INFO "please try 'cgroup_disable=memory' option if you "
	339	"don't want memory cgroups\n");
	340	return;
	341	oom:
	342	printk(KERN_CRIT "try 'cgroup_disable=memory' boot option\n");
	343	panic("Out of memory");
52d4b9ac KH	344	}
52d4b9ac KH	345
31168481	346	void __meminit pgdat_page_cgroup_init(struct pglist_data *pgdat)
52d4b9ac KH	347	{
	348	return;
	349	}
	350
	351	#endif
27a7faa0 KH	352
	353
	354	#ifdef CONFIG_CGROUP_MEM_RES_CTLR_SWAP
	355
	356	static DEFINE_MUTEX(swap_cgroup_mutex);
	357	struct swap_cgroup_ctrl {
	358	struct page **map;
	359	unsigned long length;
e9e58a4e	360	spinlock_t lock;
27a7faa0 KH	361	};
	362
	363	struct swap_cgroup_ctrl swap_cgroup_ctrl[MAX_SWAPFILES];
	364
27a7faa0	365	struct swap_cgroup {
a3b2d692	366	unsigned short id;
27a7faa0 KH	367	};
	368	#define SC_PER_PAGE (PAGE_SIZE/sizeof(struct swap_cgroup))
	369	#define SC_POS_MASK (SC_PER_PAGE - 1)
	370
	371	/*
	372	* SwapCgroup implements "lookup" and "exchange" operations.
	373	* In typical usage, this swap_cgroup is accessed via memcg's charge/uncharge
	374	* against SwapCache. At swap_free(), this is accessed directly from swap.
	375	*
	376	* This means,
	377	* - we have no race in "exchange" when we're accessed via SwapCache because
	378	* SwapCache(and its swp_entry) is under lock.
	379	* - When called via swap_free(), there is no user of this entry and no race.
	380	* Then, we don't need lock around "exchange".
	381	*
	382	* TODO: we can push these buffers out to HIGHMEM.
	383	*/
	384
	385	/*
	386	* allocate buffer for swap_cgroup.
	387	*/
	388	static int swap_cgroup_prepare(int type)
	389	{
	390	struct page *page;
	391	struct swap_cgroup_ctrl *ctrl;
	392	unsigned long idx, max;
	393
27a7faa0 KH	394	ctrl = &swap_cgroup_ctrl[type];
	395
	396	for (idx = 0; idx < ctrl->length; idx++) {
	397	page = alloc_page(GFP_KERNEL \| __GFP_ZERO);
	398	if (!page)
	399	goto not_enough_page;
	400	ctrl->map[idx] = page;
	401	}
	402	return 0;
	403	not_enough_page:
	404	max = idx;
	405	for (idx = 0; idx < max; idx++)
	406	__free_page(ctrl->map[idx]);
	407
	408	return -ENOMEM;
	409	}
	410
02491447 DN	411	/**
	412	* swap_cgroup_cmpxchg - cmpxchg mem_cgroup's id for this swp_entry.
	413	* @end: swap entry to be cmpxchged
	414	* @old: old id
	415	* @new: new id
	416	*
	417	* Returns old id at success, 0 at failure.
25985edc	418	* (There is no mem_cgroup using 0 as its id)
02491447 DN	419	*/
	420	unsigned short swap_cgroup_cmpxchg(swp_entry_t ent,
	421	unsigned short old, unsigned short new)
	422	{
	423	int type = swp_type(ent);
	424	unsigned long offset = swp_offset(ent);
	425	unsigned long idx = offset / SC_PER_PAGE;
	426	unsigned long pos = offset & SC_POS_MASK;
	427	struct swap_cgroup_ctrl *ctrl;
	428	struct page *mappage;
	429	struct swap_cgroup *sc;
e9e58a4e KH	430	unsigned long flags;
e9e58a4e KH	431	unsigned short retval;
02491447 DN	432
	433	ctrl = &swap_cgroup_ctrl[type];
	434
	435	mappage = ctrl->map[idx];
	436	sc = page_address(mappage);
	437	sc += pos;
e9e58a4e KH	438	spin_lock_irqsave(&ctrl->lock, flags);
	439	retval = sc->id;
	440	if (retval == old)
	441	sc->id = new;
02491447	442	else
e9e58a4e KH	443	retval = 0;
	444	spin_unlock_irqrestore(&ctrl->lock, flags);
	445	return retval;
02491447 DN	446	}
02491447 DN	447
27a7faa0 KH	448	/**
	449	* swap_cgroup_record - record mem_cgroup for this swp_entry.
	450	* @ent: swap entry to be recorded into
	451	* @mem: mem_cgroup to be recorded
	452	*
a3b2d692 KH	453	* Returns old value at success, 0 at failure.
a3b2d692 KH	454	* (Of course, old value can be 0.)
27a7faa0	455	*/
a3b2d692	456	unsigned short swap_cgroup_record(swp_entry_t ent, unsigned short id)
27a7faa0 KH	457	{
	458	int type = swp_type(ent);
	459	unsigned long offset = swp_offset(ent);
	460	unsigned long idx = offset / SC_PER_PAGE;
	461	unsigned long pos = offset & SC_POS_MASK;
	462	struct swap_cgroup_ctrl *ctrl;
	463	struct page *mappage;
	464	struct swap_cgroup *sc;
a3b2d692	465	unsigned short old;
e9e58a4e	466	unsigned long flags;
27a7faa0	467
27a7faa0 KH	468	ctrl = &swap_cgroup_ctrl[type];
	469
	470	mappage = ctrl->map[idx];
	471	sc = page_address(mappage);
	472	sc += pos;
e9e58a4e KH	473	spin_lock_irqsave(&ctrl->lock, flags);
	474	old = sc->id;
	475	sc->id = id;
	476	spin_unlock_irqrestore(&ctrl->lock, flags);
27a7faa0 KH	477
	478	return old;
	479	}
	480
	481	/**
	482	* lookup_swap_cgroup - lookup mem_cgroup tied to swap entry
	483	* @ent: swap entry to be looked up.
	484	*
a3b2d692	485	* Returns CSS ID of mem_cgroup at success. 0 at failure. (0 is invalid ID)
27a7faa0	486	*/
a3b2d692	487	unsigned short lookup_swap_cgroup(swp_entry_t ent)
27a7faa0 KH	488	{
	489	int type = swp_type(ent);
	490	unsigned long offset = swp_offset(ent);
	491	unsigned long idx = offset / SC_PER_PAGE;
	492	unsigned long pos = offset & SC_POS_MASK;
	493	struct swap_cgroup_ctrl *ctrl;
	494	struct page *mappage;
	495	struct swap_cgroup *sc;
a3b2d692	496	unsigned short ret;
27a7faa0	497
27a7faa0 KH	498	ctrl = &swap_cgroup_ctrl[type];
	499	mappage = ctrl->map[idx];
	500	sc = page_address(mappage);
	501	sc += pos;
a3b2d692	502	ret = sc->id;
27a7faa0 KH	503	return ret;
	504	}
	505
	506	int swap_cgroup_swapon(int type, unsigned long max_pages)
	507	{
	508	void *array;
	509	unsigned long array_size;
	510	unsigned long length;
	511	struct swap_cgroup_ctrl *ctrl;
	512
	513	if (!do_swap_account)
	514	return 0;
	515
33278f7f	516	length = DIV_ROUND_UP(max_pages, SC_PER_PAGE);
27a7faa0 KH	517	array_size = length * sizeof(void *);
27a7faa0 KH	518
8c1fec1b	519	array = vzalloc(array_size);
27a7faa0 KH	520	if (!array)
	521	goto nomem;
	522
27a7faa0 KH	523	ctrl = &swap_cgroup_ctrl[type];
	524	mutex_lock(&swap_cgroup_mutex);
	525	ctrl->length = length;
	526	ctrl->map = array;
e9e58a4e	527	spin_lock_init(&ctrl->lock);
27a7faa0 KH	528	if (swap_cgroup_prepare(type)) {
	529	/* memory shortage */
	530	ctrl->map = NULL;
	531	ctrl->length = 0;
27a7faa0	532	mutex_unlock(&swap_cgroup_mutex);
6a5b18d2	533	vfree(array);
27a7faa0 KH	534	goto nomem;
	535	}
	536	mutex_unlock(&swap_cgroup_mutex);
	537
27a7faa0 KH	538	return 0;
	539	nomem:
	540	printk(KERN_INFO "couldn't allocate enough memory for swap_cgroup.\n");
	541	printk(KERN_INFO
00a66d29	542	"swap_cgroup can be disabled by swapaccount=0 boot option\n");
27a7faa0 KH	543	return -ENOMEM;
	544	}
	545
	546	void swap_cgroup_swapoff(int type)
	547	{
6a5b18d2 NK	548	struct page **map;
6a5b18d2 NK	549	unsigned long i, length;
27a7faa0 KH	550	struct swap_cgroup_ctrl *ctrl;
	551
	552	if (!do_swap_account)
	553	return;
	554
	555	mutex_lock(&swap_cgroup_mutex);
	556	ctrl = &swap_cgroup_ctrl[type];
6a5b18d2 NK	557	map = ctrl->map;
	558	length = ctrl->length;
	559	ctrl->map = NULL;
	560	ctrl->length = 0;
	561	mutex_unlock(&swap_cgroup_mutex);
	562
	563	if (map) {
	564	for (i = 0; i < length; i++) {
	565	struct page *page = map[i];
27a7faa0 KH	566	if (page)
	567	__free_page(page);
	568	}
6a5b18d2	569	vfree(map);
27a7faa0	570	}
27a7faa0 KH	571	}
	572
	573	#endif