memcontrol.c 79.4 KB
Newer Older
B
Balbir Singh 已提交
1 2 3 4 5
/* memcontrol.c - Memory Controller
 *
 * Copyright IBM Corporation, 2007
 * Author Balbir Singh <balbir@linux.vnet.ibm.com>
 *
6 7 8
 * Copyright 2007 OpenVZ SWsoft Inc
 * Author: Pavel Emelianov <xemul@openvz.org>
 *
B
Balbir Singh 已提交
9 10 11 12 13 14 15 16 17 18 19 20 21 22
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 */

#include <linux/res_counter.h>
#include <linux/memcontrol.h>
#include <linux/cgroup.h>
23
#include <linux/mm.h>
K
KAMEZAWA Hiroyuki 已提交
24
#include <linux/pagemap.h>
25
#include <linux/smp.h>
26
#include <linux/page-flags.h>
27
#include <linux/backing-dev.h>
28 29
#include <linux/bit_spinlock.h>
#include <linux/rcupdate.h>
30
#include <linux/limits.h>
31
#include <linux/mutex.h>
32
#include <linux/rbtree.h>
33
#include <linux/slab.h>
34 35 36
#include <linux/swap.h>
#include <linux/spinlock.h>
#include <linux/fs.h>
37
#include <linux/seq_file.h>
38
#include <linux/vmalloc.h>
39
#include <linux/mm_inline.h>
40
#include <linux/page_cgroup.h>
K
KAMEZAWA Hiroyuki 已提交
41
#include "internal.h"
B
Balbir Singh 已提交
42

43 44
#include <asm/uaccess.h>

45 46
struct cgroup_subsys mem_cgroup_subsys __read_mostly;
#define MEM_CGROUP_RECLAIM_RETRIES	5
47
struct mem_cgroup *root_mem_cgroup __read_mostly;
B
Balbir Singh 已提交
48

49
#ifdef CONFIG_CGROUP_MEM_RES_CTLR_SWAP
L
Li Zefan 已提交
50
/* Turned on only when memory cgroup is enabled && really_do_swap_account = 1 */
51 52 53 54 55 56
int do_swap_account __read_mostly;
static int really_do_swap_account __initdata = 1; /* for remember boot option*/
#else
#define do_swap_account		(0)
#endif

57
static DEFINE_MUTEX(memcg_tasklist);	/* can be hold under cgroup_mutex */
58
#define SOFTLIMIT_EVENTS_THRESH (1000)
59

60 61 62 63 64 65 66 67
/*
 * Statistics for memory cgroup.
 */
enum mem_cgroup_stat_index {
	/*
	 * For MEM_CONTAINER_TYPE_ALL, usage = pagecache + rss.
	 */
	MEM_CGROUP_STAT_CACHE, 	   /* # of pages charged as cache */
68 69
	MEM_CGROUP_STAT_RSS,	   /* # of pages charged as anon rss */
	MEM_CGROUP_STAT_MAPPED_FILE,  /* # of pages charged as file rss */
70 71
	MEM_CGROUP_STAT_PGPGIN_COUNT,	/* # of pages paged in */
	MEM_CGROUP_STAT_PGPGOUT_COUNT,	/* # of pages paged out */
72
	MEM_CGROUP_STAT_EVENTS,	/* sum of pagein + pageout for internal use */
73
	MEM_CGROUP_STAT_SWAPOUT, /* # of pages, swapped out */
74 75 76 77 78 79 80 81 82

	MEM_CGROUP_STAT_NSTATS,
};

struct mem_cgroup_stat_cpu {
	s64 count[MEM_CGROUP_STAT_NSTATS];
} ____cacheline_aligned_in_smp;

struct mem_cgroup_stat {
83
	struct mem_cgroup_stat_cpu cpustat[0];
84 85
};

86 87 88 89 90 91 92 93 94 95 96 97 98 99
static inline void
__mem_cgroup_stat_reset_safe(struct mem_cgroup_stat_cpu *stat,
				enum mem_cgroup_stat_index idx)
{
	stat->count[idx] = 0;
}

static inline s64
__mem_cgroup_stat_read_local(struct mem_cgroup_stat_cpu *stat,
				enum mem_cgroup_stat_index idx)
{
	return stat->count[idx];
}

100 101 102
/*
 * For accounting under irq disable, no need for increment preempt count.
 */
103
static inline void __mem_cgroup_stat_add_safe(struct mem_cgroup_stat_cpu *stat,
104 105
		enum mem_cgroup_stat_index idx, int val)
{
106
	stat->count[idx] += val;
107 108 109 110 111 112 113 114 115 116 117 118
}

static s64 mem_cgroup_read_stat(struct mem_cgroup_stat *stat,
		enum mem_cgroup_stat_index idx)
{
	int cpu;
	s64 ret = 0;
	for_each_possible_cpu(cpu)
		ret += stat->cpustat[cpu].count[idx];
	return ret;
}

K
KAMEZAWA Hiroyuki 已提交
119 120 121 122 123 124 125 126 127
static s64 mem_cgroup_local_usage(struct mem_cgroup_stat *stat)
{
	s64 ret;

	ret = mem_cgroup_read_stat(stat, MEM_CGROUP_STAT_CACHE);
	ret += mem_cgroup_read_stat(stat, MEM_CGROUP_STAT_RSS);
	return ret;
}

128 129 130 131
/*
 * per-zone information in memory controller.
 */
struct mem_cgroup_per_zone {
132 133 134
	/*
	 * spin_lock to protect the per cgroup LRU
	 */
135 136
	struct list_head	lists[NR_LRU_LISTS];
	unsigned long		count[NR_LRU_LISTS];
K
KOSAKI Motohiro 已提交
137 138

	struct zone_reclaim_stat reclaim_stat;
139 140 141 142
	struct rb_node		tree_node;	/* RB tree node */
	unsigned long long	usage_in_excess;/* Set to the value by which */
						/* the soft limit is exceeded*/
	bool			on_tree;
143 144
	struct mem_cgroup	*mem;		/* Back pointer, we cannot */
						/* use container_of	   */
145 146 147 148 149 150 151 152 153 154 155 156
};
/* Macro for accessing counter */
#define MEM_CGROUP_ZSTAT(mz, idx)	((mz)->count[(idx)])

struct mem_cgroup_per_node {
	struct mem_cgroup_per_zone zoneinfo[MAX_NR_ZONES];
};

struct mem_cgroup_lru_info {
	struct mem_cgroup_per_node *nodeinfo[MAX_NUMNODES];
};

157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176
/*
 * Cgroups above their limits are maintained in a RB-Tree, independent of
 * their hierarchy representation
 */

struct mem_cgroup_tree_per_zone {
	struct rb_root rb_root;
	spinlock_t lock;
};

struct mem_cgroup_tree_per_node {
	struct mem_cgroup_tree_per_zone rb_tree_per_zone[MAX_NR_ZONES];
};

struct mem_cgroup_tree {
	struct mem_cgroup_tree_per_node *rb_tree_per_node[MAX_NUMNODES];
};

static struct mem_cgroup_tree soft_limit_tree __read_mostly;

B
Balbir Singh 已提交
177 178 179 180 181 182 183
/*
 * The memory controller data structure. The memory controller controls both
 * page cache and RSS per cgroup. We would eventually like to provide
 * statistics based on the statistics developed by Rik Van Riel for clock-pro,
 * to help the administrator determine what knobs to tune.
 *
 * TODO: Add a water mark for the memory controller. Reclaim will begin when
184 185 186
 * we hit the water mark. May be even add a low water mark, such that
 * no reclaim occurs from a cgroup at it's low water mark, this is
 * a feature that will be implemented much later in the future.
B
Balbir Singh 已提交
187 188 189 190 191 192 193
 */
struct mem_cgroup {
	struct cgroup_subsys_state css;
	/*
	 * the counter to account for memory usage
	 */
	struct res_counter res;
194 195 196 197
	/*
	 * the counter to account for mem+swap usage.
	 */
	struct res_counter memsw;
198 199 200 201
	/*
	 * Per cgroup active and inactive list, similar to the
	 * per zone LRU lists.
	 */
202
	struct mem_cgroup_lru_info info;
203

K
KOSAKI Motohiro 已提交
204 205 206 207 208
	/*
	  protect against reclaim related member.
	*/
	spinlock_t reclaim_param_lock;

209
	int	prev_priority;	/* for recording reclaim priority */
210 211

	/*
212
	 * While reclaiming in a hierarchy, we cache the last child we
K
KAMEZAWA Hiroyuki 已提交
213
	 * reclaimed from.
214
	 */
K
KAMEZAWA Hiroyuki 已提交
215
	int last_scanned_child;
216 217 218 219
	/*
	 * Should the accounting and control be hierarchical, per subtree?
	 */
	bool use_hierarchy;
220
	unsigned long	last_oom_jiffies;
221
	atomic_t	refcnt;
222

K
KOSAKI Motohiro 已提交
223 224
	unsigned int	swappiness;

225 226 227
	/* set when res.limit == memsw.limit */
	bool		memsw_is_minimum;

228
	/*
229
	 * statistics. This must be placed at the end of memcg.
230 231
	 */
	struct mem_cgroup_stat stat;
B
Balbir Singh 已提交
232 233
};

234 235 236 237 238 239 240
/*
 * Maximum loops in mem_cgroup_hierarchical_reclaim(), used for soft
 * limit reclaim to prevent infinite loops, if they ever occur.
 */
#define	MEM_CGROUP_MAX_RECLAIM_LOOPS		(100)
#define	MEM_CGROUP_MAX_SOFT_LIMIT_RECLAIM_LOOPS	(2)

241 242 243
enum charge_type {
	MEM_CGROUP_CHARGE_TYPE_CACHE = 0,
	MEM_CGROUP_CHARGE_TYPE_MAPPED,
244
	MEM_CGROUP_CHARGE_TYPE_SHMEM,	/* used by page migration of shmem */
245
	MEM_CGROUP_CHARGE_TYPE_FORCE,	/* used by force_empty */
K
KAMEZAWA Hiroyuki 已提交
246
	MEM_CGROUP_CHARGE_TYPE_SWAPOUT,	/* for accounting swapcache */
K
KAMEZAWA Hiroyuki 已提交
247
	MEM_CGROUP_CHARGE_TYPE_DROP,	/* a page was unused swap cache */
248 249 250
	NR_CHARGE_TYPE,
};

251 252 253 254
/* only for here (for easy reading.) */
#define PCGF_CACHE	(1UL << PCG_CACHE)
#define PCGF_USED	(1UL << PCG_USED)
#define PCGF_LOCK	(1UL << PCG_LOCK)
255 256
/* Not used, but added here for completeness */
#define PCGF_ACCT	(1UL << PCG_ACCT)
257

258 259 260 261 262 263 264
/* for encoding cft->private value on file */
#define _MEM			(0)
#define _MEMSWAP		(1)
#define MEMFILE_PRIVATE(x, val)	(((x) << 16) | (val))
#define MEMFILE_TYPE(val)	(((val) >> 16) & 0xffff)
#define MEMFILE_ATTR(val)	((val) & 0xffff)

265 266 267 268 269 270 271
/*
 * Reclaim flags for mem_cgroup_hierarchical_reclaim
 */
#define MEM_CGROUP_RECLAIM_NOSWAP_BIT	0x0
#define MEM_CGROUP_RECLAIM_NOSWAP	(1 << MEM_CGROUP_RECLAIM_NOSWAP_BIT)
#define MEM_CGROUP_RECLAIM_SHRINK_BIT	0x1
#define MEM_CGROUP_RECLAIM_SHRINK	(1 << MEM_CGROUP_RECLAIM_SHRINK_BIT)
272 273
#define MEM_CGROUP_RECLAIM_SOFT_BIT	0x2
#define MEM_CGROUP_RECLAIM_SOFT		(1 << MEM_CGROUP_RECLAIM_SOFT_BIT)
274

275 276
static void mem_cgroup_get(struct mem_cgroup *mem);
static void mem_cgroup_put(struct mem_cgroup *mem);
277
static struct mem_cgroup *parent_mem_cgroup(struct mem_cgroup *mem);
278

279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313
static struct mem_cgroup_per_zone *
mem_cgroup_zoneinfo(struct mem_cgroup *mem, int nid, int zid)
{
	return &mem->info.nodeinfo[nid]->zoneinfo[zid];
}

static struct mem_cgroup_per_zone *
page_cgroup_zoneinfo(struct page_cgroup *pc)
{
	struct mem_cgroup *mem = pc->mem_cgroup;
	int nid = page_cgroup_nid(pc);
	int zid = page_cgroup_zid(pc);

	if (!mem)
		return NULL;

	return mem_cgroup_zoneinfo(mem, nid, zid);
}

static struct mem_cgroup_tree_per_zone *
soft_limit_tree_node_zone(int nid, int zid)
{
	return &soft_limit_tree.rb_tree_per_node[nid]->rb_tree_per_zone[zid];
}

static struct mem_cgroup_tree_per_zone *
soft_limit_tree_from_page(struct page *page)
{
	int nid = page_to_nid(page);
	int zid = page_zonenum(page);

	return &soft_limit_tree.rb_tree_per_node[nid]->rb_tree_per_zone[zid];
}

static void
314
__mem_cgroup_insert_exceeded(struct mem_cgroup *mem,
315
				struct mem_cgroup_per_zone *mz,
316 317
				struct mem_cgroup_tree_per_zone *mctz,
				unsigned long long new_usage_in_excess)
318 319 320 321 322 323 324 325
{
	struct rb_node **p = &mctz->rb_root.rb_node;
	struct rb_node *parent = NULL;
	struct mem_cgroup_per_zone *mz_node;

	if (mz->on_tree)
		return;

326 327 328
	mz->usage_in_excess = new_usage_in_excess;
	if (!mz->usage_in_excess)
		return;
329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344
	while (*p) {
		parent = *p;
		mz_node = rb_entry(parent, struct mem_cgroup_per_zone,
					tree_node);
		if (mz->usage_in_excess < mz_node->usage_in_excess)
			p = &(*p)->rb_left;
		/*
		 * We can't avoid mem cgroups that are over their soft
		 * limit by the same amount
		 */
		else if (mz->usage_in_excess >= mz_node->usage_in_excess)
			p = &(*p)->rb_right;
	}
	rb_link_node(&mz->tree_node, parent, p);
	rb_insert_color(&mz->tree_node, &mctz->rb_root);
	mz->on_tree = true;
345 346 347 348 349 350 351 352 353 354 355 356 357
}

static void
__mem_cgroup_remove_exceeded(struct mem_cgroup *mem,
				struct mem_cgroup_per_zone *mz,
				struct mem_cgroup_tree_per_zone *mctz)
{
	if (!mz->on_tree)
		return;
	rb_erase(&mz->tree_node, &mctz->rb_root);
	mz->on_tree = false;
}

358 359 360 361 362 363
static void
mem_cgroup_remove_exceeded(struct mem_cgroup *mem,
				struct mem_cgroup_per_zone *mz,
				struct mem_cgroup_tree_per_zone *mctz)
{
	spin_lock(&mctz->lock);
364
	__mem_cgroup_remove_exceeded(mem, mz, mctz);
365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387
	spin_unlock(&mctz->lock);
}

static bool mem_cgroup_soft_limit_check(struct mem_cgroup *mem)
{
	bool ret = false;
	int cpu;
	s64 val;
	struct mem_cgroup_stat_cpu *cpustat;

	cpu = get_cpu();
	cpustat = &mem->stat.cpustat[cpu];
	val = __mem_cgroup_stat_read_local(cpustat, MEM_CGROUP_STAT_EVENTS);
	if (unlikely(val > SOFTLIMIT_EVENTS_THRESH)) {
		__mem_cgroup_stat_reset_safe(cpustat, MEM_CGROUP_STAT_EVENTS);
		ret = true;
	}
	put_cpu();
	return ret;
}

static void mem_cgroup_update_tree(struct mem_cgroup *mem, struct page *page)
{
388
	unsigned long long excess;
389 390
	struct mem_cgroup_per_zone *mz;
	struct mem_cgroup_tree_per_zone *mctz;
391 392
	int nid = page_to_nid(page);
	int zid = page_zonenum(page);
393 394 395
	mctz = soft_limit_tree_from_page(page);

	/*
396 397
	 * Necessary to update all ancestors when hierarchy is used.
	 * because their event counter is not touched.
398
	 */
399 400
	for (; mem; mem = parent_mem_cgroup(mem)) {
		mz = mem_cgroup_zoneinfo(mem, nid, zid);
401
		excess = res_counter_soft_limit_excess(&mem->res);
402 403 404 405
		/*
		 * We have to update the tree if mz is on RB-tree or
		 * mem is over its softlimit.
		 */
406
		if (excess || mz->on_tree) {
407 408 409 410 411
			spin_lock(&mctz->lock);
			/* if on-tree, remove it */
			if (mz->on_tree)
				__mem_cgroup_remove_exceeded(mem, mz, mctz);
			/*
412 413
			 * Insert again. mz->usage_in_excess will be updated.
			 * If excess is 0, no tree ops.
414
			 */
415
			__mem_cgroup_insert_exceeded(mem, mz, mctz, excess);
416 417
			spin_unlock(&mctz->lock);
		}
418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435
	}
}

static void mem_cgroup_remove_from_trees(struct mem_cgroup *mem)
{
	int node, zone;
	struct mem_cgroup_per_zone *mz;
	struct mem_cgroup_tree_per_zone *mctz;

	for_each_node_state(node, N_POSSIBLE) {
		for (zone = 0; zone < MAX_NR_ZONES; zone++) {
			mz = mem_cgroup_zoneinfo(mem, node, zone);
			mctz = soft_limit_tree_node_zone(node, zone);
			mem_cgroup_remove_exceeded(mem, mz, mctz);
		}
	}
}

436 437 438 439 440 441 442 443 444
static inline unsigned long mem_cgroup_get_excess(struct mem_cgroup *mem)
{
	return res_counter_soft_limit_excess(&mem->res) >> PAGE_SHIFT;
}

static struct mem_cgroup_per_zone *
__mem_cgroup_largest_soft_limit_node(struct mem_cgroup_tree_per_zone *mctz)
{
	struct rb_node *rightmost = NULL;
445
	struct mem_cgroup_per_zone *mz;
446 447

retry:
448
	mz = NULL;
449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477
	rightmost = rb_last(&mctz->rb_root);
	if (!rightmost)
		goto done;		/* Nothing to reclaim from */

	mz = rb_entry(rightmost, struct mem_cgroup_per_zone, tree_node);
	/*
	 * Remove the node now but someone else can add it back,
	 * we will to add it back at the end of reclaim to its correct
	 * position in the tree.
	 */
	__mem_cgroup_remove_exceeded(mz->mem, mz, mctz);
	if (!res_counter_soft_limit_excess(&mz->mem->res) ||
		!css_tryget(&mz->mem->css))
		goto retry;
done:
	return mz;
}

static struct mem_cgroup_per_zone *
mem_cgroup_largest_soft_limit_node(struct mem_cgroup_tree_per_zone *mctz)
{
	struct mem_cgroup_per_zone *mz;

	spin_lock(&mctz->lock);
	mz = __mem_cgroup_largest_soft_limit_node(mctz);
	spin_unlock(&mctz->lock);
	return mz;
}

478 479 480 481 482 483 484 485 486 487 488 489 490
static void mem_cgroup_swap_statistics(struct mem_cgroup *mem,
					 bool charge)
{
	int val = (charge) ? 1 : -1;
	struct mem_cgroup_stat *stat = &mem->stat;
	struct mem_cgroup_stat_cpu *cpustat;
	int cpu = get_cpu();

	cpustat = &stat->cpustat[cpu];
	__mem_cgroup_stat_add_safe(cpustat, MEM_CGROUP_STAT_SWAPOUT, val);
	put_cpu();
}

491 492 493
static void mem_cgroup_charge_statistics(struct mem_cgroup *mem,
					 struct page_cgroup *pc,
					 bool charge)
494
{
495
	int val = (charge) ? 1 : -1;
496
	struct mem_cgroup_stat *stat = &mem->stat;
497
	struct mem_cgroup_stat_cpu *cpustat;
K
KAMEZAWA Hiroyuki 已提交
498
	int cpu = get_cpu();
499

K
KAMEZAWA Hiroyuki 已提交
500
	cpustat = &stat->cpustat[cpu];
501
	if (PageCgroupCache(pc))
502
		__mem_cgroup_stat_add_safe(cpustat, MEM_CGROUP_STAT_CACHE, val);
503
	else
504
		__mem_cgroup_stat_add_safe(cpustat, MEM_CGROUP_STAT_RSS, val);
505 506

	if (charge)
507
		__mem_cgroup_stat_add_safe(cpustat,
508 509
				MEM_CGROUP_STAT_PGPGIN_COUNT, 1);
	else
510
		__mem_cgroup_stat_add_safe(cpustat,
511
				MEM_CGROUP_STAT_PGPGOUT_COUNT, 1);
512
	__mem_cgroup_stat_add_safe(cpustat, MEM_CGROUP_STAT_EVENTS, 1);
K
KAMEZAWA Hiroyuki 已提交
513
	put_cpu();
514 515
}

K
KAMEZAWA Hiroyuki 已提交
516
static unsigned long mem_cgroup_get_local_zonestat(struct mem_cgroup *mem,
517
					enum lru_list idx)
518 519 520 521 522 523 524 525 526 527 528
{
	int nid, zid;
	struct mem_cgroup_per_zone *mz;
	u64 total = 0;

	for_each_online_node(nid)
		for (zid = 0; zid < MAX_NR_ZONES; zid++) {
			mz = mem_cgroup_zoneinfo(mem, nid, zid);
			total += MEM_CGROUP_ZSTAT(mz, idx);
		}
	return total;
529 530
}

531
static struct mem_cgroup *mem_cgroup_from_cont(struct cgroup *cont)
B
Balbir Singh 已提交
532 533 534 535 536 537
{
	return container_of(cgroup_subsys_state(cont,
				mem_cgroup_subsys_id), struct mem_cgroup,
				css);
}

538
struct mem_cgroup *mem_cgroup_from_task(struct task_struct *p)
539
{
540 541 542 543 544 545 546 547
	/*
	 * mm_update_next_owner() may clear mm->owner to NULL
	 * if it races with swapoff, page migration, etc.
	 * So this can be called with p == NULL.
	 */
	if (unlikely(!p))
		return NULL;

548 549 550 551
	return container_of(task_subsys_state(p, mem_cgroup_subsys_id),
				struct mem_cgroup, css);
}

552 553 554
static struct mem_cgroup *try_get_mem_cgroup_from_mm(struct mm_struct *mm)
{
	struct mem_cgroup *mem = NULL;
555 556 557

	if (!mm)
		return NULL;
558 559 560 561 562 563 564 565 566 567 568 569 570 571 572
	/*
	 * Because we have no locks, mm->owner's may be being moved to other
	 * cgroup. We use css_tryget() here even if this looks
	 * pessimistic (rather than adding locks here).
	 */
	rcu_read_lock();
	do {
		mem = mem_cgroup_from_task(rcu_dereference(mm->owner));
		if (unlikely(!mem))
			break;
	} while (!css_tryget(&mem->css));
	rcu_read_unlock();
	return mem;
}

K
KAMEZAWA Hiroyuki 已提交
573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607
/*
 * Call callback function against all cgroup under hierarchy tree.
 */
static int mem_cgroup_walk_tree(struct mem_cgroup *root, void *data,
			  int (*func)(struct mem_cgroup *, void *))
{
	int found, ret, nextid;
	struct cgroup_subsys_state *css;
	struct mem_cgroup *mem;

	if (!root->use_hierarchy)
		return (*func)(root, data);

	nextid = 1;
	do {
		ret = 0;
		mem = NULL;

		rcu_read_lock();
		css = css_get_next(&mem_cgroup_subsys, nextid, &root->css,
				   &found);
		if (css && css_tryget(css))
			mem = container_of(css, struct mem_cgroup, css);
		rcu_read_unlock();

		if (mem) {
			ret = (*func)(mem, data);
			css_put(&mem->css);
		}
		nextid = found + 1;
	} while (!ret && css);

	return ret;
}

608 609 610 611 612
static inline bool mem_cgroup_is_root(struct mem_cgroup *mem)
{
	return (mem == root_mem_cgroup);
}

K
KAMEZAWA Hiroyuki 已提交
613 614 615 616 617 618 619 620 621 622 623 624 625
/*
 * Following LRU functions are allowed to be used without PCG_LOCK.
 * Operations are called by routine of global LRU independently from memcg.
 * What we have to take care of here is validness of pc->mem_cgroup.
 *
 * Changes to pc->mem_cgroup happens when
 * 1. charge
 * 2. moving account
 * In typical case, "charge" is done before add-to-lru. Exception is SwapCache.
 * It is added to LRU before charge.
 * If PCG_USED bit is not set, page_cgroup is not added to this private LRU.
 * When moving account, the page is not on LRU. It's isolated.
 */
626

K
KAMEZAWA Hiroyuki 已提交
627 628 629 630
void mem_cgroup_del_lru_list(struct page *page, enum lru_list lru)
{
	struct page_cgroup *pc;
	struct mem_cgroup_per_zone *mz;
631

632
	if (mem_cgroup_disabled())
K
KAMEZAWA Hiroyuki 已提交
633 634 635
		return;
	pc = lookup_page_cgroup(page);
	/* can happen while we handle swapcache. */
636
	if (!TestClearPageCgroupAcctLRU(pc))
K
KAMEZAWA Hiroyuki 已提交
637
		return;
638
	VM_BUG_ON(!pc->mem_cgroup);
639 640 641 642
	/*
	 * We don't check PCG_USED bit. It's cleared when the "page" is finally
	 * removed from global LRU.
	 */
K
KAMEZAWA Hiroyuki 已提交
643
	mz = page_cgroup_zoneinfo(pc);
644
	MEM_CGROUP_ZSTAT(mz, lru) -= 1;
645 646 647
	if (mem_cgroup_is_root(pc->mem_cgroup))
		return;
	VM_BUG_ON(list_empty(&pc->lru));
K
KAMEZAWA Hiroyuki 已提交
648 649
	list_del_init(&pc->lru);
	return;
650 651
}

K
KAMEZAWA Hiroyuki 已提交
652
void mem_cgroup_del_lru(struct page *page)
653
{
K
KAMEZAWA Hiroyuki 已提交
654 655
	mem_cgroup_del_lru_list(page, page_lru(page));
}
656

K
KAMEZAWA Hiroyuki 已提交
657 658 659 660
void mem_cgroup_rotate_lru_list(struct page *page, enum lru_list lru)
{
	struct mem_cgroup_per_zone *mz;
	struct page_cgroup *pc;
661

662
	if (mem_cgroup_disabled())
K
KAMEZAWA Hiroyuki 已提交
663
		return;
664

K
KAMEZAWA Hiroyuki 已提交
665
	pc = lookup_page_cgroup(page);
666 667 668 669
	/*
	 * Used bit is set without atomic ops but after smp_wmb().
	 * For making pc->mem_cgroup visible, insert smp_rmb() here.
	 */
K
KAMEZAWA Hiroyuki 已提交
670
	smp_rmb();
671 672
	/* unused or root page is not rotated. */
	if (!PageCgroupUsed(pc) || mem_cgroup_is_root(pc->mem_cgroup))
K
KAMEZAWA Hiroyuki 已提交
673 674 675
		return;
	mz = page_cgroup_zoneinfo(pc);
	list_move(&pc->lru, &mz->lists[lru]);
676 677
}

K
KAMEZAWA Hiroyuki 已提交
678
void mem_cgroup_add_lru_list(struct page *page, enum lru_list lru)
679
{
K
KAMEZAWA Hiroyuki 已提交
680 681
	struct page_cgroup *pc;
	struct mem_cgroup_per_zone *mz;
682

683
	if (mem_cgroup_disabled())
K
KAMEZAWA Hiroyuki 已提交
684 685
		return;
	pc = lookup_page_cgroup(page);
686
	VM_BUG_ON(PageCgroupAcctLRU(pc));
687 688 689 690
	/*
	 * Used bit is set without atomic ops but after smp_wmb().
	 * For making pc->mem_cgroup visible, insert smp_rmb() here.
	 */
K
KAMEZAWA Hiroyuki 已提交
691 692
	smp_rmb();
	if (!PageCgroupUsed(pc))
L
Lee Schermerhorn 已提交
693
		return;
694

K
KAMEZAWA Hiroyuki 已提交
695
	mz = page_cgroup_zoneinfo(pc);
696
	MEM_CGROUP_ZSTAT(mz, lru) += 1;
697 698 699
	SetPageCgroupAcctLRU(pc);
	if (mem_cgroup_is_root(pc->mem_cgroup))
		return;
K
KAMEZAWA Hiroyuki 已提交
700 701
	list_add(&pc->lru, &mz->lists[lru]);
}
702

K
KAMEZAWA Hiroyuki 已提交
703
/*
704 705 706 707 708
 * At handling SwapCache, pc->mem_cgroup may be changed while it's linked to
 * lru because the page may.be reused after it's fully uncharged (because of
 * SwapCache behavior).To handle that, unlink page_cgroup from LRU when charge
 * it again. This function is only used to charge SwapCache. It's done under
 * lock_page and expected that zone->lru_lock is never held.
K
KAMEZAWA Hiroyuki 已提交
709
 */
710
static void mem_cgroup_lru_del_before_commit_swapcache(struct page *page)
K
KAMEZAWA Hiroyuki 已提交
711
{
712 713 714 715 716 717 718 719 720 721 722 723
	unsigned long flags;
	struct zone *zone = page_zone(page);
	struct page_cgroup *pc = lookup_page_cgroup(page);

	spin_lock_irqsave(&zone->lru_lock, flags);
	/*
	 * Forget old LRU when this page_cgroup is *not* used. This Used bit
	 * is guarded by lock_page() because the page is SwapCache.
	 */
	if (!PageCgroupUsed(pc))
		mem_cgroup_del_lru_list(page, page_lru(page));
	spin_unlock_irqrestore(&zone->lru_lock, flags);
K
KAMEZAWA Hiroyuki 已提交
724 725
}

726 727 728 729 730 731 732 733
static void mem_cgroup_lru_add_after_commit_swapcache(struct page *page)
{
	unsigned long flags;
	struct zone *zone = page_zone(page);
	struct page_cgroup *pc = lookup_page_cgroup(page);

	spin_lock_irqsave(&zone->lru_lock, flags);
	/* link when the page is linked to LRU but page_cgroup isn't */
734
	if (PageLRU(page) && !PageCgroupAcctLRU(pc))
735 736 737 738 739
		mem_cgroup_add_lru_list(page, page_lru(page));
	spin_unlock_irqrestore(&zone->lru_lock, flags);
}


K
KAMEZAWA Hiroyuki 已提交
740 741 742
void mem_cgroup_move_lists(struct page *page,
			   enum lru_list from, enum lru_list to)
{
743
	if (mem_cgroup_disabled())
K
KAMEZAWA Hiroyuki 已提交
744 745 746
		return;
	mem_cgroup_del_lru_list(page, from);
	mem_cgroup_add_lru_list(page, to);
747 748
}

749 750 751
int task_in_mem_cgroup(struct task_struct *task, const struct mem_cgroup *mem)
{
	int ret;
752
	struct mem_cgroup *curr = NULL;
753 754

	task_lock(task);
755 756 757
	rcu_read_lock();
	curr = try_get_mem_cgroup_from_mm(task->mm);
	rcu_read_unlock();
758
	task_unlock(task);
759 760 761 762 763 764 765
	if (!curr)
		return 0;
	if (curr->use_hierarchy)
		ret = css_is_ancestor(&curr->css, &mem->css);
	else
		ret = (curr == mem);
	css_put(&curr->css);
766 767 768
	return ret;
}

769 770 771 772 773
/*
 * prev_priority control...this will be used in memory reclaim path.
 */
int mem_cgroup_get_reclaim_priority(struct mem_cgroup *mem)
{
K
KOSAKI Motohiro 已提交
774 775 776 777 778 779 780
	int prev_priority;

	spin_lock(&mem->reclaim_param_lock);
	prev_priority = mem->prev_priority;
	spin_unlock(&mem->reclaim_param_lock);

	return prev_priority;
781 782 783 784
}

void mem_cgroup_note_reclaim_priority(struct mem_cgroup *mem, int priority)
{
K
KOSAKI Motohiro 已提交
785
	spin_lock(&mem->reclaim_param_lock);
786 787
	if (priority < mem->prev_priority)
		mem->prev_priority = priority;
K
KOSAKI Motohiro 已提交
788
	spin_unlock(&mem->reclaim_param_lock);
789 790 791 792
}

void mem_cgroup_record_reclaim_priority(struct mem_cgroup *mem, int priority)
{
K
KOSAKI Motohiro 已提交
793
	spin_lock(&mem->reclaim_param_lock);
794
	mem->prev_priority = priority;
K
KOSAKI Motohiro 已提交
795
	spin_unlock(&mem->reclaim_param_lock);
796 797
}

798
static int calc_inactive_ratio(struct mem_cgroup *memcg, unsigned long *present_pages)
799 800 801
{
	unsigned long active;
	unsigned long inactive;
802 803
	unsigned long gb;
	unsigned long inactive_ratio;
804

K
KAMEZAWA Hiroyuki 已提交
805 806
	inactive = mem_cgroup_get_local_zonestat(memcg, LRU_INACTIVE_ANON);
	active = mem_cgroup_get_local_zonestat(memcg, LRU_ACTIVE_ANON);
807

808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834
	gb = (inactive + active) >> (30 - PAGE_SHIFT);
	if (gb)
		inactive_ratio = int_sqrt(10 * gb);
	else
		inactive_ratio = 1;

	if (present_pages) {
		present_pages[0] = inactive;
		present_pages[1] = active;
	}

	return inactive_ratio;
}

int mem_cgroup_inactive_anon_is_low(struct mem_cgroup *memcg)
{
	unsigned long active;
	unsigned long inactive;
	unsigned long present_pages[2];
	unsigned long inactive_ratio;

	inactive_ratio = calc_inactive_ratio(memcg, present_pages);

	inactive = present_pages[0];
	active = present_pages[1];

	if (inactive * inactive_ratio < active)
835 836 837 838 839
		return 1;

	return 0;
}

840 841 842 843 844 845 846 847 848 849 850
int mem_cgroup_inactive_file_is_low(struct mem_cgroup *memcg)
{
	unsigned long active;
	unsigned long inactive;

	inactive = mem_cgroup_get_local_zonestat(memcg, LRU_INACTIVE_FILE);
	active = mem_cgroup_get_local_zonestat(memcg, LRU_ACTIVE_FILE);

	return (active > inactive);
}

851 852 853 854 855 856 857 858 859 860 861
unsigned long mem_cgroup_zone_nr_pages(struct mem_cgroup *memcg,
				       struct zone *zone,
				       enum lru_list lru)
{
	int nid = zone->zone_pgdat->node_id;
	int zid = zone_idx(zone);
	struct mem_cgroup_per_zone *mz = mem_cgroup_zoneinfo(memcg, nid, zid);

	return MEM_CGROUP_ZSTAT(mz, lru);
}

K
KOSAKI Motohiro 已提交
862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881
struct zone_reclaim_stat *mem_cgroup_get_reclaim_stat(struct mem_cgroup *memcg,
						      struct zone *zone)
{
	int nid = zone->zone_pgdat->node_id;
	int zid = zone_idx(zone);
	struct mem_cgroup_per_zone *mz = mem_cgroup_zoneinfo(memcg, nid, zid);

	return &mz->reclaim_stat;
}

struct zone_reclaim_stat *
mem_cgroup_get_reclaim_stat_from_page(struct page *page)
{
	struct page_cgroup *pc;
	struct mem_cgroup_per_zone *mz;

	if (mem_cgroup_disabled())
		return NULL;

	pc = lookup_page_cgroup(page);
882 883 884 885 886 887 888 889
	/*
	 * Used bit is set without atomic ops but after smp_wmb().
	 * For making pc->mem_cgroup visible, insert smp_rmb() here.
	 */
	smp_rmb();
	if (!PageCgroupUsed(pc))
		return NULL;

K
KOSAKI Motohiro 已提交
890 891 892 893 894 895 896
	mz = page_cgroup_zoneinfo(pc);
	if (!mz)
		return NULL;

	return &mz->reclaim_stat;
}

897 898 899 900 901
unsigned long mem_cgroup_isolate_pages(unsigned long nr_to_scan,
					struct list_head *dst,
					unsigned long *scanned, int order,
					int mode, struct zone *z,
					struct mem_cgroup *mem_cont,
902
					int active, int file)
903 904 905 906 907 908
{
	unsigned long nr_taken = 0;
	struct page *page;
	unsigned long scan;
	LIST_HEAD(pc_list);
	struct list_head *src;
909
	struct page_cgroup *pc, *tmp;
910 911 912
	int nid = z->zone_pgdat->node_id;
	int zid = zone_idx(z);
	struct mem_cgroup_per_zone *mz;
913
	int lru = LRU_FILE * file + active;
914
	int ret;
915

916
	BUG_ON(!mem_cont);
917
	mz = mem_cgroup_zoneinfo(mem_cont, nid, zid);
918
	src = &mz->lists[lru];
919

920 921
	scan = 0;
	list_for_each_entry_safe_reverse(pc, tmp, src, lru) {
H
Hugh Dickins 已提交
922
		if (scan >= nr_to_scan)
923
			break;
K
KAMEZAWA Hiroyuki 已提交
924 925

		page = pc->page;
926 927
		if (unlikely(!PageCgroupUsed(pc)))
			continue;
H
Hugh Dickins 已提交
928
		if (unlikely(!PageLRU(page)))
929 930
			continue;

H
Hugh Dickins 已提交
931
		scan++;
932 933 934
		ret = __isolate_lru_page(page, mode, file);
		switch (ret) {
		case 0:
935
			list_move(&page->lru, dst);
936
			mem_cgroup_del_lru(page);
937
			nr_taken++;
938 939 940 941 942 943 944
			break;
		case -EBUSY:
			/* we don't affect global LRU but rotate in our LRU */
			mem_cgroup_rotate_lru_list(page, page_lru(page));
			break;
		default:
			break;
945 946 947 948 949 950 951
		}
	}

	*scanned = scan;
	return nr_taken;
}

952 953 954
#define mem_cgroup_from_res_counter(counter, member)	\
	container_of(counter, struct mem_cgroup, member)

955 956 957 958 959 960 961 962 963 964 965 966
static bool mem_cgroup_check_under_limit(struct mem_cgroup *mem)
{
	if (do_swap_account) {
		if (res_counter_check_under_limit(&mem->res) &&
			res_counter_check_under_limit(&mem->memsw))
			return true;
	} else
		if (res_counter_check_under_limit(&mem->res))
			return true;
	return false;
}

K
KOSAKI Motohiro 已提交
967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982
static unsigned int get_swappiness(struct mem_cgroup *memcg)
{
	struct cgroup *cgrp = memcg->css.cgroup;
	unsigned int swappiness;

	/* root ? */
	if (cgrp->parent == NULL)
		return vm_swappiness;

	spin_lock(&memcg->reclaim_param_lock);
	swappiness = memcg->swappiness;
	spin_unlock(&memcg->reclaim_param_lock);

	return swappiness;
}

983 984 985 986 987 988
static int mem_cgroup_count_children_cb(struct mem_cgroup *mem, void *data)
{
	int *val = data;
	(*val)++;
	return 0;
}
989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056

/**
 * mem_cgroup_print_mem_info: Called from OOM with tasklist_lock held in read mode.
 * @memcg: The memory cgroup that went over limit
 * @p: Task that is going to be killed
 *
 * NOTE: @memcg and @p's mem_cgroup can be different when hierarchy is
 * enabled
 */
void mem_cgroup_print_oom_info(struct mem_cgroup *memcg, struct task_struct *p)
{
	struct cgroup *task_cgrp;
	struct cgroup *mem_cgrp;
	/*
	 * Need a buffer in BSS, can't rely on allocations. The code relies
	 * on the assumption that OOM is serialized for memory controller.
	 * If this assumption is broken, revisit this code.
	 */
	static char memcg_name[PATH_MAX];
	int ret;

	if (!memcg)
		return;


	rcu_read_lock();

	mem_cgrp = memcg->css.cgroup;
	task_cgrp = task_cgroup(p, mem_cgroup_subsys_id);

	ret = cgroup_path(task_cgrp, memcg_name, PATH_MAX);
	if (ret < 0) {
		/*
		 * Unfortunately, we are unable to convert to a useful name
		 * But we'll still print out the usage information
		 */
		rcu_read_unlock();
		goto done;
	}
	rcu_read_unlock();

	printk(KERN_INFO "Task in %s killed", memcg_name);

	rcu_read_lock();
	ret = cgroup_path(mem_cgrp, memcg_name, PATH_MAX);
	if (ret < 0) {
		rcu_read_unlock();
		goto done;
	}
	rcu_read_unlock();

	/*
	 * Continues from above, so we don't need an KERN_ level
	 */
	printk(KERN_CONT " as a result of limit of %s\n", memcg_name);
done:

	printk(KERN_INFO "memory: usage %llukB, limit %llukB, failcnt %llu\n",
		res_counter_read_u64(&memcg->res, RES_USAGE) >> 10,
		res_counter_read_u64(&memcg->res, RES_LIMIT) >> 10,
		res_counter_read_u64(&memcg->res, RES_FAILCNT));
	printk(KERN_INFO "memory+swap: usage %llukB, limit %llukB, "
		"failcnt %llu\n",
		res_counter_read_u64(&memcg->memsw, RES_USAGE) >> 10,
		res_counter_read_u64(&memcg->memsw, RES_LIMIT) >> 10,
		res_counter_read_u64(&memcg->memsw, RES_FAILCNT));
}

1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067
/*
 * This function returns the number of memcg under hierarchy tree. Returns
 * 1(self count) if no children.
 */
static int mem_cgroup_count_children(struct mem_cgroup *mem)
{
	int num = 0;
 	mem_cgroup_walk_tree(mem, &num, mem_cgroup_count_children_cb);
	return num;
}

1068
/*
K
KAMEZAWA Hiroyuki 已提交
1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110
 * Visit the first child (need not be the first child as per the ordering
 * of the cgroup list, since we track last_scanned_child) of @mem and use
 * that to reclaim free pages from.
 */
static struct mem_cgroup *
mem_cgroup_select_victim(struct mem_cgroup *root_mem)
{
	struct mem_cgroup *ret = NULL;
	struct cgroup_subsys_state *css;
	int nextid, found;

	if (!root_mem->use_hierarchy) {
		css_get(&root_mem->css);
		ret = root_mem;
	}

	while (!ret) {
		rcu_read_lock();
		nextid = root_mem->last_scanned_child + 1;
		css = css_get_next(&mem_cgroup_subsys, nextid, &root_mem->css,
				   &found);
		if (css && css_tryget(css))
			ret = container_of(css, struct mem_cgroup, css);

		rcu_read_unlock();
		/* Updates scanning parameter */
		spin_lock(&root_mem->reclaim_param_lock);
		if (!css) {
			/* this means start scan from ID:1 */
			root_mem->last_scanned_child = 0;
		} else
			root_mem->last_scanned_child = found;
		spin_unlock(&root_mem->reclaim_param_lock);
	}

	return ret;
}

/*
 * Scan the hierarchy if needed to reclaim memory. We remember the last child
 * we reclaimed from, so that we don't end up penalizing one child extensively
 * based on its position in the children list.
1111 1112
 *
 * root_mem is the original ancestor that we've been reclaim from.
K
KAMEZAWA Hiroyuki 已提交
1113 1114 1115
 *
 * We give up and return to the caller when we visit root_mem twice.
 * (other groups can be removed while we're walking....)
1116 1117
 *
 * If shrink==true, for avoiding to free too much, this returns immedieately.
1118 1119
 */
static int mem_cgroup_hierarchical_reclaim(struct mem_cgroup *root_mem,
1120
						struct zone *zone,
1121 1122
						gfp_t gfp_mask,
						unsigned long reclaim_options)
1123
{
K
KAMEZAWA Hiroyuki 已提交
1124 1125 1126
	struct mem_cgroup *victim;
	int ret, total = 0;
	int loop = 0;
1127 1128
	bool noswap = reclaim_options & MEM_CGROUP_RECLAIM_NOSWAP;
	bool shrink = reclaim_options & MEM_CGROUP_RECLAIM_SHRINK;
1129 1130
	bool check_soft = reclaim_options & MEM_CGROUP_RECLAIM_SOFT;
	unsigned long excess = mem_cgroup_get_excess(root_mem);
K
KAMEZAWA Hiroyuki 已提交
1131

1132 1133 1134 1135
	/* If memsw_is_minimum==1, swap-out is of-no-use. */
	if (root_mem->memsw_is_minimum)
		noswap = true;

1136
	while (1) {
K
KAMEZAWA Hiroyuki 已提交
1137
		victim = mem_cgroup_select_victim(root_mem);
1138
		if (victim == root_mem) {
K
KAMEZAWA Hiroyuki 已提交
1139
			loop++;
1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162
			if (loop >= 2) {
				/*
				 * If we have not been able to reclaim
				 * anything, it might because there are
				 * no reclaimable pages under this hierarchy
				 */
				if (!check_soft || !total) {
					css_put(&victim->css);
					break;
				}
				/*
				 * We want to do more targetted reclaim.
				 * excess >> 2 is not to excessive so as to
				 * reclaim too much, nor too less that we keep
				 * coming back to reclaim from this cgroup
				 */
				if (total >= (excess >> 2) ||
					(loop > MEM_CGROUP_MAX_RECLAIM_LOOPS)) {
					css_put(&victim->css);
					break;
				}
			}
		}
K
KAMEZAWA Hiroyuki 已提交
1163 1164 1165
		if (!mem_cgroup_local_usage(&victim->stat)) {
			/* this cgroup's local usage == 0 */
			css_put(&victim->css);
1166 1167
			continue;
		}
K
KAMEZAWA Hiroyuki 已提交
1168
		/* we use swappiness of local cgroup */
1169 1170 1171 1172 1173 1174 1175
		if (check_soft)
			ret = mem_cgroup_shrink_node_zone(victim, gfp_mask,
				noswap, get_swappiness(victim), zone,
				zone->zone_pgdat->node_id);
		else
			ret = try_to_free_mem_cgroup_pages(victim, gfp_mask,
						noswap, get_swappiness(victim));
K
KAMEZAWA Hiroyuki 已提交
1176
		css_put(&victim->css);
1177 1178 1179 1180 1181 1182 1183
		/*
		 * At shrinking usage, we can't check we should stop here or
		 * reclaim more. It's depends on callers. last_scanned_child
		 * will work enough for keeping fairness under tree.
		 */
		if (shrink)
			return ret;
K
KAMEZAWA Hiroyuki 已提交
1184
		total += ret;
1185 1186 1187 1188
		if (check_soft) {
			if (res_counter_check_under_soft_limit(&root_mem->res))
				return total;
		} else if (mem_cgroup_check_under_limit(root_mem))
K
KAMEZAWA Hiroyuki 已提交
1189
			return 1 + total;
1190
	}
K
KAMEZAWA Hiroyuki 已提交
1191
	return total;
1192 1193
}

1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209
bool mem_cgroup_oom_called(struct task_struct *task)
{
	bool ret = false;
	struct mem_cgroup *mem;
	struct mm_struct *mm;

	rcu_read_lock();
	mm = task->mm;
	if (!mm)
		mm = &init_mm;
	mem = mem_cgroup_from_task(rcu_dereference(mm->owner));
	if (mem && time_before(jiffies, mem->last_oom_jiffies + HZ/10))
		ret = true;
	rcu_read_unlock();
	return ret;
}
1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221

static int record_last_oom_cb(struct mem_cgroup *mem, void *data)
{
	mem->last_oom_jiffies = jiffies;
	return 0;
}

static void record_last_oom(struct mem_cgroup *mem)
{
	mem_cgroup_walk_tree(mem, NULL, record_last_oom_cb);
}

1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259
/*
 * Currently used to update mapped file statistics, but the routine can be
 * generalized to update other statistics as well.
 */
void mem_cgroup_update_mapped_file_stat(struct page *page, int val)
{
	struct mem_cgroup *mem;
	struct mem_cgroup_stat *stat;
	struct mem_cgroup_stat_cpu *cpustat;
	int cpu;
	struct page_cgroup *pc;

	if (!page_is_file_cache(page))
		return;

	pc = lookup_page_cgroup(page);
	if (unlikely(!pc))
		return;

	lock_page_cgroup(pc);
	mem = pc->mem_cgroup;
	if (!mem)
		goto done;

	if (!PageCgroupUsed(pc))
		goto done;

	/*
	 * Preemption is already disabled, we don't need get_cpu()
	 */
	cpu = smp_processor_id();
	stat = &mem->stat;
	cpustat = &stat->cpustat[cpu];

	__mem_cgroup_stat_add_safe(cpustat, MEM_CGROUP_STAT_MAPPED_FILE, val);
done:
	unlock_page_cgroup(pc);
}
1260

1261 1262 1263
/*
 * Unlike exported interface, "oom" parameter is added. if oom==true,
 * oom-killer can be invoked.
1264
 */
1265
static int __mem_cgroup_try_charge(struct mm_struct *mm,
1266
			gfp_t gfp_mask, struct mem_cgroup **memcg,
1267
			bool oom, struct page *page)
1268
{
1269
	struct mem_cgroup *mem, *mem_over_limit;
1270
	int nr_retries = MEM_CGROUP_RECLAIM_RETRIES;
1271
	struct res_counter *fail_res;
1272 1273 1274 1275 1276 1277 1278

	if (unlikely(test_thread_flag(TIF_MEMDIE))) {
		/* Don't account this! */
		*memcg = NULL;
		return 0;
	}

1279
	/*
1280 1281
	 * We always charge the cgroup the mm_struct belongs to.
	 * The mm_struct's mem_cgroup changes on task migration if the
1282 1283 1284
	 * thread group leader migrates. It's possible that mm is not
	 * set, if so charge the init_mm (happens for pagecache usage).
	 */
1285 1286 1287
	mem = *memcg;
	if (likely(!mem)) {
		mem = try_get_mem_cgroup_from_mm(mm);
1288
		*memcg = mem;
1289
	} else {
1290
		css_get(&mem->css);
1291
	}
1292 1293 1294
	if (unlikely(!mem))
		return 0;

1295
	VM_BUG_ON(css_is_removed(&mem->css));
1296

1297
	while (1) {
1298
		int ret = 0;
1299
		unsigned long flags = 0;
1300

1301 1302
		if (mem_cgroup_is_root(mem))
			goto done;
1303
		ret = res_counter_charge(&mem->res, PAGE_SIZE, &fail_res);
1304 1305 1306
		if (likely(!ret)) {
			if (!do_swap_account)
				break;
1307
			ret = res_counter_charge(&mem->memsw, PAGE_SIZE,
1308
							&fail_res);
1309 1310 1311
			if (likely(!ret))
				break;
			/* mem+swap counter fails */
1312
			res_counter_uncharge(&mem->res, PAGE_SIZE);
1313
			flags |= MEM_CGROUP_RECLAIM_NOSWAP;
1314 1315 1316 1317 1318 1319 1320
			mem_over_limit = mem_cgroup_from_res_counter(fail_res,
									memsw);
		} else
			/* mem counter fails */
			mem_over_limit = mem_cgroup_from_res_counter(fail_res,
									res);

1321
		if (!(gfp_mask & __GFP_WAIT))
1322
			goto nomem;
1323

1324 1325
		ret = mem_cgroup_hierarchical_reclaim(mem_over_limit, NULL,
						gfp_mask, flags);
1326 1327
		if (ret)
			continue;
1328 1329

		/*
1330 1331 1332 1333 1334
		 * try_to_free_mem_cgroup_pages() might not give us a full
		 * picture of reclaim. Some pages are reclaimed and might be
		 * moved to swap cache or just unmapped from the cgroup.
		 * Check the limit again to see if the reclaim reduced the
		 * current usage of the cgroup before giving up
1335
		 *
1336
		 */
1337 1338
		if (mem_cgroup_check_under_limit(mem_over_limit))
			continue;
1339 1340

		if (!nr_retries--) {
1341
			if (oom) {
1342
				mutex_lock(&memcg_tasklist);
1343
				mem_cgroup_out_of_memory(mem_over_limit, gfp_mask);
1344
				mutex_unlock(&memcg_tasklist);
1345
				record_last_oom(mem_over_limit);
1346
			}
1347
			goto nomem;
1348
		}
1349
	}
1350
	/*
1351 1352
	 * Insert ancestor (and ancestor's ancestors), to softlimit RB-tree.
	 * if they exceeds softlimit.
1353
	 */
1354 1355
	if (mem_cgroup_soft_limit_check(mem))
		mem_cgroup_update_tree(mem, page);
1356
done:
1357 1358 1359 1360 1361
	return 0;
nomem:
	css_put(&mem->css);
	return -ENOMEM;
}
1362

1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381
/*
 * A helper function to get mem_cgroup from ID. must be called under
 * rcu_read_lock(). The caller must check css_is_removed() or some if
 * it's concern. (dropping refcnt from swap can be called against removed
 * memcg.)
 */
static struct mem_cgroup *mem_cgroup_lookup(unsigned short id)
{
	struct cgroup_subsys_state *css;

	/* ID 0 is unused ID */
	if (!id)
		return NULL;
	css = css_lookup(&mem_cgroup_subsys, id);
	if (!css)
		return NULL;
	return container_of(css, struct mem_cgroup, css);
}

1382 1383 1384
static struct mem_cgroup *try_get_mem_cgroup_from_swapcache(struct page *page)
{
	struct mem_cgroup *mem;
1385
	struct page_cgroup *pc;
1386
	unsigned short id;
1387 1388
	swp_entry_t ent;

1389 1390
	VM_BUG_ON(!PageLocked(page));

1391 1392 1393
	if (!PageSwapCache(page))
		return NULL;

1394
	pc = lookup_page_cgroup(page);
1395
	lock_page_cgroup(pc);
1396
	if (PageCgroupUsed(pc)) {
1397
		mem = pc->mem_cgroup;
1398 1399 1400
		if (mem && !css_tryget(&mem->css))
			mem = NULL;
	} else {
1401
		ent.val = page_private(page);
1402 1403 1404 1405 1406 1407
		id = lookup_swap_cgroup(ent);
		rcu_read_lock();
		mem = mem_cgroup_lookup(id);
		if (mem && !css_tryget(&mem->css))
			mem = NULL;
		rcu_read_unlock();
1408
	}
1409
	unlock_page_cgroup(pc);
1410 1411 1412
	return mem;
}

1413
/*
1414
 * commit a charge got by __mem_cgroup_try_charge() and makes page_cgroup to be
1415 1416 1417 1418 1419 1420 1421 1422 1423 1424
 * USED state. If already USED, uncharge and return.
 */

static void __mem_cgroup_commit_charge(struct mem_cgroup *mem,
				     struct page_cgroup *pc,
				     enum charge_type ctype)
{
	/* try_charge() can return NULL to *memcg, taking care of it. */
	if (!mem)
		return;
1425 1426 1427 1428

	lock_page_cgroup(pc);
	if (unlikely(PageCgroupUsed(pc))) {
		unlock_page_cgroup(pc);
1429
		if (!mem_cgroup_is_root(mem)) {
1430
			res_counter_uncharge(&mem->res, PAGE_SIZE);
1431
			if (do_swap_account)
1432
				res_counter_uncharge(&mem->memsw, PAGE_SIZE);
1433
		}
1434
		css_put(&mem->css);
1435
		return;
1436
	}
1437

1438
	pc->mem_cgroup = mem;
1439 1440 1441 1442 1443 1444 1445
	/*
	 * We access a page_cgroup asynchronously without lock_page_cgroup().
	 * Especially when a page_cgroup is taken from a page, pc->mem_cgroup
	 * is accessed after testing USED bit. To make pc->mem_cgroup visible
	 * before USED bit, we need memory barrier here.
	 * See mem_cgroup_add_lru_list(), etc.
 	 */
K
KAMEZAWA Hiroyuki 已提交
1446
	smp_wmb();
1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459
	switch (ctype) {
	case MEM_CGROUP_CHARGE_TYPE_CACHE:
	case MEM_CGROUP_CHARGE_TYPE_SHMEM:
		SetPageCgroupCache(pc);
		SetPageCgroupUsed(pc);
		break;
	case MEM_CGROUP_CHARGE_TYPE_MAPPED:
		ClearPageCgroupCache(pc);
		SetPageCgroupUsed(pc);
		break;
	default:
		break;
	}
1460

K
KAMEZAWA Hiroyuki 已提交
1461
	mem_cgroup_charge_statistics(mem, pc, true);
1462 1463

	unlock_page_cgroup(pc);
1464
}
1465

1466 1467 1468 1469 1470 1471 1472
/**
 * mem_cgroup_move_account - move account of the page
 * @pc:	page_cgroup of the page.
 * @from: mem_cgroup which the page is moved from.
 * @to:	mem_cgroup which the page is moved to. @from != @to.
 *
 * The caller must confirm following.
K
KAMEZAWA Hiroyuki 已提交
1473
 * - page is not on LRU (isolate_page() is useful.)
1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487
 *
 * returns 0 at success,
 * returns -EBUSY when lock is busy or "pc" is unstable.
 *
 * This function does "uncharge" from old cgroup but doesn't do "charge" to
 * new cgroup. It should be done by a caller.
 */

static int mem_cgroup_move_account(struct page_cgroup *pc,
	struct mem_cgroup *from, struct mem_cgroup *to)
{
	struct mem_cgroup_per_zone *from_mz, *to_mz;
	int nid, zid;
	int ret = -EBUSY;
1488 1489 1490 1491
	struct page *page;
	int cpu;
	struct mem_cgroup_stat *stat;
	struct mem_cgroup_stat_cpu *cpustat;
1492 1493

	VM_BUG_ON(from == to);
K
KAMEZAWA Hiroyuki 已提交
1494
	VM_BUG_ON(PageLRU(pc->page));
1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509

	nid = page_cgroup_nid(pc);
	zid = page_cgroup_zid(pc);
	from_mz =  mem_cgroup_zoneinfo(from, nid, zid);
	to_mz =  mem_cgroup_zoneinfo(to, nid, zid);

	if (!trylock_page_cgroup(pc))
		return ret;

	if (!PageCgroupUsed(pc))
		goto out;

	if (pc->mem_cgroup != from)
		goto out;

1510
	if (!mem_cgroup_is_root(from))
1511
		res_counter_uncharge(&from->res, PAGE_SIZE);
K
KAMEZAWA Hiroyuki 已提交
1512
	mem_cgroup_charge_statistics(from, pc, false);
1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529

	page = pc->page;
	if (page_is_file_cache(page) && page_mapped(page)) {
		cpu = smp_processor_id();
		/* Update mapped_file data for mem_cgroup "from" */
		stat = &from->stat;
		cpustat = &stat->cpustat[cpu];
		__mem_cgroup_stat_add_safe(cpustat, MEM_CGROUP_STAT_MAPPED_FILE,
						-1);

		/* Update mapped_file data for mem_cgroup "to" */
		stat = &to->stat;
		cpustat = &stat->cpustat[cpu];
		__mem_cgroup_stat_add_safe(cpustat, MEM_CGROUP_STAT_MAPPED_FILE,
						1);
	}

1530
	if (do_swap_account && !mem_cgroup_is_root(from))
1531
		res_counter_uncharge(&from->memsw, PAGE_SIZE);
1532 1533 1534
	css_put(&from->css);

	css_get(&to->css);
K
KAMEZAWA Hiroyuki 已提交
1535 1536 1537
	pc->mem_cgroup = to;
	mem_cgroup_charge_statistics(to, pc, true);
	ret = 0;
1538 1539
out:
	unlock_page_cgroup(pc);
1540 1541 1542 1543 1544 1545
	/*
	 * We charges against "to" which may not have any tasks. Then, "to"
	 * can be under rmdir(). But in current implementation, caller of
	 * this function is just force_empty() and it's garanteed that
	 * "to" is never removed. So, we don't check rmdir status here.
	 */
1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556
	return ret;
}

/*
 * move charges to its parent.
 */

static int mem_cgroup_move_parent(struct page_cgroup *pc,
				  struct mem_cgroup *child,
				  gfp_t gfp_mask)
{
K
KAMEZAWA Hiroyuki 已提交
1557
	struct page *page = pc->page;
1558 1559 1560 1561 1562 1563 1564 1565 1566
	struct cgroup *cg = child->css.cgroup;
	struct cgroup *pcg = cg->parent;
	struct mem_cgroup *parent;
	int ret;

	/* Is ROOT ? */
	if (!pcg)
		return -EINVAL;

K
KAMEZAWA Hiroyuki 已提交
1567

1568 1569
	parent = mem_cgroup_from_cont(pcg);

K
KAMEZAWA Hiroyuki 已提交
1570

1571
	ret = __mem_cgroup_try_charge(NULL, gfp_mask, &parent, false, page);
1572
	if (ret || !parent)
1573 1574
		return ret;

1575 1576 1577 1578
	if (!get_page_unless_zero(page)) {
		ret = -EBUSY;
		goto uncharge;
	}
K
KAMEZAWA Hiroyuki 已提交
1579 1580 1581 1582 1583

	ret = isolate_lru_page(page);

	if (ret)
		goto cancel;
1584 1585 1586

	ret = mem_cgroup_move_account(pc, child, parent);

K
KAMEZAWA Hiroyuki 已提交
1587 1588 1589
	putback_lru_page(page);
	if (!ret) {
		put_page(page);
1590 1591
		/* drop extra refcnt by try_charge() */
		css_put(&parent->css);
K
KAMEZAWA Hiroyuki 已提交
1592
		return 0;
1593
	}
1594

K
KAMEZAWA Hiroyuki 已提交
1595
cancel:
1596 1597 1598 1599 1600
	put_page(page);
uncharge:
	/* drop extra refcnt by try_charge() */
	css_put(&parent->css);
	/* uncharge if move fails */
1601
	if (!mem_cgroup_is_root(parent)) {
1602
		res_counter_uncharge(&parent->res, PAGE_SIZE);
1603
		if (do_swap_account)
1604
			res_counter_uncharge(&parent->memsw, PAGE_SIZE);
1605
	}
1606 1607 1608
	return ret;
}

1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629
/*
 * Charge the memory controller for page usage.
 * Return
 * 0 if the charge was successful
 * < 0 if the cgroup is over its limit
 */
static int mem_cgroup_charge_common(struct page *page, struct mm_struct *mm,
				gfp_t gfp_mask, enum charge_type ctype,
				struct mem_cgroup *memcg)
{
	struct mem_cgroup *mem;
	struct page_cgroup *pc;
	int ret;

	pc = lookup_page_cgroup(page);
	/* can happen at boot */
	if (unlikely(!pc))
		return 0;
	prefetchw(pc);

	mem = memcg;
1630
	ret = __mem_cgroup_try_charge(mm, gfp_mask, &mem, true, page);
1631
	if (ret || !mem)
1632 1633 1634
		return ret;

	__mem_cgroup_commit_charge(mem, pc, ctype);
1635 1636 1637
	return 0;
}

1638 1639
int mem_cgroup_newpage_charge(struct page *page,
			      struct mm_struct *mm, gfp_t gfp_mask)
1640
{
1641
	if (mem_cgroup_disabled())
1642
		return 0;
1643 1644
	if (PageCompound(page))
		return 0;
1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655
	/*
	 * If already mapped, we don't have to account.
	 * If page cache, page->mapping has address_space.
	 * But page->mapping may have out-of-use anon_vma pointer,
	 * detecit it by PageAnon() check. newly-mapped-anon's page->mapping
	 * is NULL.
  	 */
	if (page_mapped(page) || (page->mapping && !PageAnon(page)))
		return 0;
	if (unlikely(!mm))
		mm = &init_mm;
1656
	return mem_cgroup_charge_common(page, mm, gfp_mask,
1657
				MEM_CGROUP_CHARGE_TYPE_MAPPED, NULL);
1658 1659
}

D
Daisuke Nishimura 已提交
1660 1661 1662 1663
static void
__mem_cgroup_commit_charge_swapin(struct page *page, struct mem_cgroup *ptr,
					enum charge_type ctype);

1664 1665
int mem_cgroup_cache_charge(struct page *page, struct mm_struct *mm,
				gfp_t gfp_mask)
1666
{
1667 1668 1669
	struct mem_cgroup *mem = NULL;
	int ret;

1670
	if (mem_cgroup_disabled())
1671
		return 0;
1672 1673
	if (PageCompound(page))
		return 0;
1674 1675 1676 1677 1678 1679 1680 1681
	/*
	 * Corner case handling. This is called from add_to_page_cache()
	 * in usual. But some FS (shmem) precharges this page before calling it
	 * and call add_to_page_cache() with GFP_NOWAIT.
	 *
	 * For GFP_NOWAIT case, the page may be pre-charged before calling
	 * add_to_page_cache(). (See shmem.c) check it here and avoid to call
	 * charge twice. (It works but has to pay a bit larger cost.)
1682 1683
	 * And when the page is SwapCache, it should take swap information
	 * into account. This is under lock_page() now.
1684 1685 1686 1687
	 */
	if (!(gfp_mask & __GFP_WAIT)) {
		struct page_cgroup *pc;

1688 1689 1690 1691 1692 1693 1694

		pc = lookup_page_cgroup(page);
		if (!pc)
			return 0;
		lock_page_cgroup(pc);
		if (PageCgroupUsed(pc)) {
			unlock_page_cgroup(pc);
1695 1696
			return 0;
		}
1697
		unlock_page_cgroup(pc);
1698 1699
	}

1700
	if (unlikely(!mm && !mem))
1701
		mm = &init_mm;
1702

1703 1704
	if (page_is_file_cache(page))
		return mem_cgroup_charge_common(page, mm, gfp_mask,
1705
				MEM_CGROUP_CHARGE_TYPE_CACHE, NULL);
1706

D
Daisuke Nishimura 已提交
1707 1708 1709 1710 1711 1712 1713 1714 1715
	/* shmem */
	if (PageSwapCache(page)) {
		ret = mem_cgroup_try_charge_swapin(mm, page, gfp_mask, &mem);
		if (!ret)
			__mem_cgroup_commit_charge_swapin(page, mem,
					MEM_CGROUP_CHARGE_TYPE_SHMEM);
	} else
		ret = mem_cgroup_charge_common(page, mm, gfp_mask,
					MEM_CGROUP_CHARGE_TYPE_SHMEM, mem);
1716 1717

	return ret;
1718 1719
}

1720 1721 1722
/*
 * While swap-in, try_charge -> commit or cancel, the page is locked.
 * And when try_charge() successfully returns, one refcnt to memcg without
1723
 * struct page_cgroup is acquired. This refcnt will be consumed by
1724 1725
 * "commit()" or removed by "cancel()"
 */
1726 1727 1728 1729 1730
int mem_cgroup_try_charge_swapin(struct mm_struct *mm,
				 struct page *page,
				 gfp_t mask, struct mem_cgroup **ptr)
{
	struct mem_cgroup *mem;
1731
	int ret;
1732

1733
	if (mem_cgroup_disabled())
1734 1735 1736 1737 1738 1739
		return 0;

	if (!do_swap_account)
		goto charge_cur_mm;
	/*
	 * A racing thread's fault, or swapoff, may have already updated
H
Hugh Dickins 已提交
1740 1741 1742
	 * the pte, and even removed page from swap cache: in those cases
	 * do_swap_page()'s pte_same() test will fail; but there's also a
	 * KSM case which does need to charge the page.
1743 1744
	 */
	if (!PageSwapCache(page))
H
Hugh Dickins 已提交
1745
		goto charge_cur_mm;
1746
	mem = try_get_mem_cgroup_from_swapcache(page);
1747 1748
	if (!mem)
		goto charge_cur_mm;
1749
	*ptr = mem;
1750
	ret = __mem_cgroup_try_charge(NULL, mask, ptr, true, page);
1751 1752 1753
	/* drop extra refcnt from tryget */
	css_put(&mem->css);
	return ret;
1754 1755 1756
charge_cur_mm:
	if (unlikely(!mm))
		mm = &init_mm;
1757
	return __mem_cgroup_try_charge(mm, mask, ptr, true, page);
1758 1759
}

D
Daisuke Nishimura 已提交
1760 1761 1762
static void
__mem_cgroup_commit_charge_swapin(struct page *page, struct mem_cgroup *ptr,
					enum charge_type ctype)
1763 1764 1765
{
	struct page_cgroup *pc;

1766
	if (mem_cgroup_disabled())
1767 1768 1769
		return;
	if (!ptr)
		return;
1770
	cgroup_exclude_rmdir(&ptr->css);
1771
	pc = lookup_page_cgroup(page);
1772
	mem_cgroup_lru_del_before_commit_swapcache(page);
D
Daisuke Nishimura 已提交
1773
	__mem_cgroup_commit_charge(ptr, pc, ctype);
1774
	mem_cgroup_lru_add_after_commit_swapcache(page);
1775 1776 1777
	/*
	 * Now swap is on-memory. This means this page may be
	 * counted both as mem and swap....double count.
1778 1779 1780
	 * Fix it by uncharging from memsw. Basically, this SwapCache is stable
	 * under lock_page(). But in do_swap_page()::memory.c, reuse_swap_page()
	 * may call delete_from_swap_cache() before reach here.
1781
	 */
1782
	if (do_swap_account && PageSwapCache(page)) {
1783
		swp_entry_t ent = {.val = page_private(page)};
1784
		unsigned short id;
1785
		struct mem_cgroup *memcg;
1786 1787 1788 1789

		id = swap_cgroup_record(ent, 0);
		rcu_read_lock();
		memcg = mem_cgroup_lookup(id);
1790
		if (memcg) {
1791 1792 1793 1794
			/*
			 * This recorded memcg can be obsolete one. So, avoid
			 * calling css_tryget
			 */
1795
			if (!mem_cgroup_is_root(memcg))
1796
				res_counter_uncharge(&memcg->memsw, PAGE_SIZE);
1797
			mem_cgroup_swap_statistics(memcg, false);
1798 1799
			mem_cgroup_put(memcg);
		}
1800
		rcu_read_unlock();
1801
	}
1802 1803 1804 1805 1806 1807
	/*
	 * At swapin, we may charge account against cgroup which has no tasks.
	 * So, rmdir()->pre_destroy() can be called while we do this charge.
	 * In that case, we need to call pre_destroy() again. check it here.
	 */
	cgroup_release_and_wakeup_rmdir(&ptr->css);
1808 1809
}

D
Daisuke Nishimura 已提交
1810 1811 1812 1813 1814 1815
void mem_cgroup_commit_charge_swapin(struct page *page, struct mem_cgroup *ptr)
{
	__mem_cgroup_commit_charge_swapin(page, ptr,
					MEM_CGROUP_CHARGE_TYPE_MAPPED);
}

1816 1817
void mem_cgroup_cancel_charge_swapin(struct mem_cgroup *mem)
{
1818
	if (mem_cgroup_disabled())
1819 1820 1821
		return;
	if (!mem)
		return;
1822
	if (!mem_cgroup_is_root(mem)) {
1823
		res_counter_uncharge(&mem->res, PAGE_SIZE);
1824
		if (do_swap_account)
1825
			res_counter_uncharge(&mem->memsw, PAGE_SIZE);
1826
	}
1827 1828 1829 1830
	css_put(&mem->css);
}


1831
/*
1832
 * uncharge if !page_mapped(page)
1833
 */
1834
static struct mem_cgroup *
1835
__mem_cgroup_uncharge_common(struct page *page, enum charge_type ctype)
1836
{
H
Hugh Dickins 已提交
1837
	struct page_cgroup *pc;
1838
	struct mem_cgroup *mem = NULL;
1839
	struct mem_cgroup_per_zone *mz;
1840

1841
	if (mem_cgroup_disabled())
1842
		return NULL;
1843

K
KAMEZAWA Hiroyuki 已提交
1844
	if (PageSwapCache(page))
1845
		return NULL;
K
KAMEZAWA Hiroyuki 已提交
1846

1847
	/*
1848
	 * Check if our page_cgroup is valid
1849
	 */
1850 1851
	pc = lookup_page_cgroup(page);
	if (unlikely(!pc || !PageCgroupUsed(pc)))
1852
		return NULL;
1853

1854
	lock_page_cgroup(pc);
K
KAMEZAWA Hiroyuki 已提交
1855

1856 1857
	mem = pc->mem_cgroup;

K
KAMEZAWA Hiroyuki 已提交
1858 1859 1860 1861 1862
	if (!PageCgroupUsed(pc))
		goto unlock_out;

	switch (ctype) {
	case MEM_CGROUP_CHARGE_TYPE_MAPPED:
K
KAMEZAWA Hiroyuki 已提交
1863
	case MEM_CGROUP_CHARGE_TYPE_DROP:
K
KAMEZAWA Hiroyuki 已提交
1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875
		if (page_mapped(page))
			goto unlock_out;
		break;
	case MEM_CGROUP_CHARGE_TYPE_SWAPOUT:
		if (!PageAnon(page)) {	/* Shared memory */
			if (page->mapping && !page_is_file_cache(page))
				goto unlock_out;
		} else if (page_mapped(page)) /* Anon */
				goto unlock_out;
		break;
	default:
		break;
1876
	}
K
KAMEZAWA Hiroyuki 已提交
1877

1878
	if (!mem_cgroup_is_root(mem)) {
1879
		res_counter_uncharge(&mem->res, PAGE_SIZE);
1880 1881
		if (do_swap_account &&
				(ctype != MEM_CGROUP_CHARGE_TYPE_SWAPOUT))
1882
			res_counter_uncharge(&mem->memsw, PAGE_SIZE);
1883 1884 1885
	}
	if (ctype == MEM_CGROUP_CHARGE_TYPE_SWAPOUT)
		mem_cgroup_swap_statistics(mem, true);
K
KAMEZAWA Hiroyuki 已提交
1886
	mem_cgroup_charge_statistics(mem, pc, false);
K
KAMEZAWA Hiroyuki 已提交
1887

1888
	ClearPageCgroupUsed(pc);
1889 1890 1891 1892 1893 1894
	/*
	 * pc->mem_cgroup is not cleared here. It will be accessed when it's
	 * freed from LRU. This is safe because uncharged page is expected not
	 * to be reused (freed soon). Exception is SwapCache, it's handled by
	 * special functions.
	 */
1895

1896
	mz = page_cgroup_zoneinfo(pc);
1897
	unlock_page_cgroup(pc);
H
Hugh Dickins 已提交
1898

1899
	if (mem_cgroup_soft_limit_check(mem))
1900
		mem_cgroup_update_tree(mem, page);
K
KAMEZAWA Hiroyuki 已提交
1901 1902 1903
	/* at swapout, this memcg will be accessed to record to swap */
	if (ctype != MEM_CGROUP_CHARGE_TYPE_SWAPOUT)
		css_put(&mem->css);
1904

1905
	return mem;
K
KAMEZAWA Hiroyuki 已提交
1906 1907 1908

unlock_out:
	unlock_page_cgroup(pc);
1909
	return NULL;
1910 1911
}

1912 1913
void mem_cgroup_uncharge_page(struct page *page)
{
1914 1915 1916 1917 1918
	/* early check. */
	if (page_mapped(page))
		return;
	if (page->mapping && !PageAnon(page))
		return;
1919 1920 1921 1922 1923 1924
	__mem_cgroup_uncharge_common(page, MEM_CGROUP_CHARGE_TYPE_MAPPED);
}

void mem_cgroup_uncharge_cache_page(struct page *page)
{
	VM_BUG_ON(page_mapped(page));
1925
	VM_BUG_ON(page->mapping);
1926 1927 1928
	__mem_cgroup_uncharge_common(page, MEM_CGROUP_CHARGE_TYPE_CACHE);
}

1929
#ifdef CONFIG_SWAP
1930
/*
1931
 * called after __delete_from_swap_cache() and drop "page" account.
1932 1933
 * memcg information is recorded to swap_cgroup of "ent"
 */
K
KAMEZAWA Hiroyuki 已提交
1934 1935
void
mem_cgroup_uncharge_swapcache(struct page *page, swp_entry_t ent, bool swapout)
1936 1937
{
	struct mem_cgroup *memcg;
K
KAMEZAWA Hiroyuki 已提交
1938 1939 1940 1941 1942 1943
	int ctype = MEM_CGROUP_CHARGE_TYPE_SWAPOUT;

	if (!swapout) /* this was a swap cache but the swap is unused ! */
		ctype = MEM_CGROUP_CHARGE_TYPE_DROP;

	memcg = __mem_cgroup_uncharge_common(page, ctype);
1944 1945

	/* record memcg information */
K
KAMEZAWA Hiroyuki 已提交
1946
	if (do_swap_account && swapout && memcg) {
1947
		swap_cgroup_record(ent, css_id(&memcg->css));
1948 1949
		mem_cgroup_get(memcg);
	}
K
KAMEZAWA Hiroyuki 已提交
1950
	if (swapout && memcg)
K
KAMEZAWA Hiroyuki 已提交
1951
		css_put(&memcg->css);
1952
}
1953
#endif
1954 1955 1956 1957 1958 1959 1960

#ifdef CONFIG_CGROUP_MEM_RES_CTLR_SWAP
/*
 * called from swap_entry_free(). remove record in swap_cgroup and
 * uncharge "memsw" account.
 */
void mem_cgroup_uncharge_swap(swp_entry_t ent)
K
KAMEZAWA Hiroyuki 已提交
1961
{
1962
	struct mem_cgroup *memcg;
1963
	unsigned short id;
1964 1965 1966 1967

	if (!do_swap_account)
		return;

1968 1969 1970
	id = swap_cgroup_record(ent, 0);
	rcu_read_lock();
	memcg = mem_cgroup_lookup(id);
1971
	if (memcg) {
1972 1973 1974 1975
		/*
		 * We uncharge this because swap is freed.
		 * This memcg can be obsolete one. We avoid calling css_tryget
		 */
1976
		if (!mem_cgroup_is_root(memcg))
1977
			res_counter_uncharge(&memcg->memsw, PAGE_SIZE);
1978
		mem_cgroup_swap_statistics(memcg, false);
1979 1980
		mem_cgroup_put(memcg);
	}
1981
	rcu_read_unlock();
K
KAMEZAWA Hiroyuki 已提交
1982
}
1983
#endif
K
KAMEZAWA Hiroyuki 已提交
1984

1985
/*
1986 1987
 * Before starting migration, account PAGE_SIZE to mem_cgroup that the old
 * page belongs to.
1988
 */
1989
int mem_cgroup_prepare_migration(struct page *page, struct mem_cgroup **ptr)
1990 1991
{
	struct page_cgroup *pc;
1992 1993
	struct mem_cgroup *mem = NULL;
	int ret = 0;
1994

1995
	if (mem_cgroup_disabled())
1996 1997
		return 0;

1998 1999 2000
	pc = lookup_page_cgroup(page);
	lock_page_cgroup(pc);
	if (PageCgroupUsed(pc)) {
2001 2002 2003
		mem = pc->mem_cgroup;
		css_get(&mem->css);
	}
2004
	unlock_page_cgroup(pc);
2005

2006
	if (mem) {
2007 2008
		ret = __mem_cgroup_try_charge(NULL, GFP_KERNEL, &mem, false,
						page);
2009 2010
		css_put(&mem->css);
	}
2011
	*ptr = mem;
2012
	return ret;
2013
}
2014

2015
/* remove redundant charge if migration failed*/
2016 2017
void mem_cgroup_end_migration(struct mem_cgroup *mem,
		struct page *oldpage, struct page *newpage)
2018
{
2019 2020 2021 2022 2023 2024
	struct page *target, *unused;
	struct page_cgroup *pc;
	enum charge_type ctype;

	if (!mem)
		return;
2025
	cgroup_exclude_rmdir(&mem->css);
2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042
	/* at migration success, oldpage->mapping is NULL. */
	if (oldpage->mapping) {
		target = oldpage;
		unused = NULL;
	} else {
		target = newpage;
		unused = oldpage;
	}

	if (PageAnon(target))
		ctype = MEM_CGROUP_CHARGE_TYPE_MAPPED;
	else if (page_is_file_cache(target))
		ctype = MEM_CGROUP_CHARGE_TYPE_CACHE;
	else
		ctype = MEM_CGROUP_CHARGE_TYPE_SHMEM;

	/* unused page is not on radix-tree now. */
K
KAMEZAWA Hiroyuki 已提交
2043
	if (unused)
2044 2045 2046
		__mem_cgroup_uncharge_common(unused, ctype);

	pc = lookup_page_cgroup(target);
2047
	/*
2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061
	 * __mem_cgroup_commit_charge() check PCG_USED bit of page_cgroup.
	 * So, double-counting is effectively avoided.
	 */
	__mem_cgroup_commit_charge(mem, pc, ctype);

	/*
	 * Both of oldpage and newpage are still under lock_page().
	 * Then, we don't have to care about race in radix-tree.
	 * But we have to be careful that this page is unmapped or not.
	 *
	 * There is a case for !page_mapped(). At the start of
	 * migration, oldpage was mapped. But now, it's zapped.
	 * But we know *target* page is not freed/reused under us.
	 * mem_cgroup_uncharge_page() does all necessary checks.
2062
	 */
2063 2064
	if (ctype == MEM_CGROUP_CHARGE_TYPE_MAPPED)
		mem_cgroup_uncharge_page(target);
2065 2066 2067 2068 2069 2070
	/*
	 * At migration, we may charge account against cgroup which has no tasks
	 * So, rmdir()->pre_destroy() can be called while we do this charge.
	 * In that case, we need to call pre_destroy() again. check it here.
	 */
	cgroup_release_and_wakeup_rmdir(&mem->css);
2071
}
2072

2073
/*
2074 2075 2076 2077 2078 2079
 * A call to try to shrink memory usage on charge failure at shmem's swapin.
 * Calling hierarchical_reclaim is not enough because we should update
 * last_oom_jiffies to prevent pagefault_out_of_memory from invoking global OOM.
 * Moreover considering hierarchy, we should reclaim from the mem_over_limit,
 * not from the memcg which this page would be charged to.
 * try_charge_swapin does all of these works properly.
2080
 */
2081
int mem_cgroup_shmem_charge_fallback(struct page *page,
2082 2083
			    struct mm_struct *mm,
			    gfp_t gfp_mask)
2084
{
2085
	struct mem_cgroup *mem = NULL;
2086
	int ret;
2087

2088
	if (mem_cgroup_disabled())
2089
		return 0;
2090

2091 2092 2093
	ret = mem_cgroup_try_charge_swapin(mm, page, gfp_mask, &mem);
	if (!ret)
		mem_cgroup_cancel_charge_swapin(mem); /* it does !mem check */
2094

2095
	return ret;
2096 2097
}

2098 2099
static DEFINE_MUTEX(set_limit_mutex);

2100
static int mem_cgroup_resize_limit(struct mem_cgroup *memcg,
2101
				unsigned long long val)
2102
{
2103
	int retry_count;
2104
	int progress;
2105
	u64 memswlimit;
2106
	int ret = 0;
2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117
	int children = mem_cgroup_count_children(memcg);
	u64 curusage, oldusage;

	/*
	 * For keeping hierarchical_reclaim simple, how long we should retry
	 * is depends on callers. We set our retry-count to be function
	 * of # of children which we should visit in this loop.
	 */
	retry_count = MEM_CGROUP_RECLAIM_RETRIES * children;

	oldusage = res_counter_read_u64(&memcg->res, RES_USAGE);
2118

2119
	while (retry_count) {
2120 2121 2122 2123
		if (signal_pending(current)) {
			ret = -EINTR;
			break;
		}
2124 2125 2126 2127 2128 2129 2130 2131 2132 2133
		/*
		 * Rather than hide all in some function, I do this in
		 * open coded manner. You see what this really does.
		 * We have to guarantee mem->res.limit < mem->memsw.limit.
		 */
		mutex_lock(&set_limit_mutex);
		memswlimit = res_counter_read_u64(&memcg->memsw, RES_LIMIT);
		if (memswlimit < val) {
			ret = -EINVAL;
			mutex_unlock(&set_limit_mutex);
2134 2135
			break;
		}
2136
		ret = res_counter_set_limit(&memcg->res, val);
2137 2138 2139 2140 2141 2142
		if (!ret) {
			if (memswlimit == val)
				memcg->memsw_is_minimum = true;
			else
				memcg->memsw_is_minimum = false;
		}
2143 2144 2145 2146 2147
		mutex_unlock(&set_limit_mutex);

		if (!ret)
			break;

2148 2149 2150
		progress = mem_cgroup_hierarchical_reclaim(memcg, NULL,
						GFP_KERNEL,
						MEM_CGROUP_RECLAIM_SHRINK);
2151 2152 2153 2154 2155 2156
		curusage = res_counter_read_u64(&memcg->res, RES_USAGE);
		/* Usage is reduced ? */
  		if (curusage >= oldusage)
			retry_count--;
		else
			oldusage = curusage;
2157
	}
2158

2159 2160 2161
	return ret;
}

L
Li Zefan 已提交
2162 2163
static int mem_cgroup_resize_memsw_limit(struct mem_cgroup *memcg,
					unsigned long long val)
2164
{
2165
	int retry_count;
2166
	u64 memlimit, oldusage, curusage;
2167 2168
	int children = mem_cgroup_count_children(memcg);
	int ret = -EBUSY;
2169

2170 2171 2172
	/* see mem_cgroup_resize_res_limit */
 	retry_count = children * MEM_CGROUP_RECLAIM_RETRIES;
	oldusage = res_counter_read_u64(&memcg->memsw, RES_USAGE);
2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190
	while (retry_count) {
		if (signal_pending(current)) {
			ret = -EINTR;
			break;
		}
		/*
		 * Rather than hide all in some function, I do this in
		 * open coded manner. You see what this really does.
		 * We have to guarantee mem->res.limit < mem->memsw.limit.
		 */
		mutex_lock(&set_limit_mutex);
		memlimit = res_counter_read_u64(&memcg->res, RES_LIMIT);
		if (memlimit > val) {
			ret = -EINVAL;
			mutex_unlock(&set_limit_mutex);
			break;
		}
		ret = res_counter_set_limit(&memcg->memsw, val);
2191 2192 2193 2194 2195 2196
		if (!ret) {
			if (memlimit == val)
				memcg->memsw_is_minimum = true;
			else
				memcg->memsw_is_minimum = false;
		}
2197 2198 2199 2200 2201
		mutex_unlock(&set_limit_mutex);

		if (!ret)
			break;

2202
		mem_cgroup_hierarchical_reclaim(memcg, NULL, GFP_KERNEL,
2203 2204
						MEM_CGROUP_RECLAIM_NOSWAP |
						MEM_CGROUP_RECLAIM_SHRINK);
2205
		curusage = res_counter_read_u64(&memcg->memsw, RES_USAGE);
2206
		/* Usage is reduced ? */
2207
		if (curusage >= oldusage)
2208
			retry_count--;
2209 2210
		else
			oldusage = curusage;
2211 2212 2213 2214
	}
	return ret;
}

2215 2216 2217 2218 2219 2220 2221 2222 2223
unsigned long mem_cgroup_soft_limit_reclaim(struct zone *zone, int order,
						gfp_t gfp_mask, int nid,
						int zid)
{
	unsigned long nr_reclaimed = 0;
	struct mem_cgroup_per_zone *mz, *next_mz = NULL;
	unsigned long reclaimed;
	int loop = 0;
	struct mem_cgroup_tree_per_zone *mctz;
2224
	unsigned long long excess;
2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276

	if (order > 0)
		return 0;

	mctz = soft_limit_tree_node_zone(nid, zid);
	/*
	 * This loop can run a while, specially if mem_cgroup's continuously
	 * keep exceeding their soft limit and putting the system under
	 * pressure
	 */
	do {
		if (next_mz)
			mz = next_mz;
		else
			mz = mem_cgroup_largest_soft_limit_node(mctz);
		if (!mz)
			break;

		reclaimed = mem_cgroup_hierarchical_reclaim(mz->mem, zone,
						gfp_mask,
						MEM_CGROUP_RECLAIM_SOFT);
		nr_reclaimed += reclaimed;
		spin_lock(&mctz->lock);

		/*
		 * If we failed to reclaim anything from this memory cgroup
		 * it is time to move on to the next cgroup
		 */
		next_mz = NULL;
		if (!reclaimed) {
			do {
				/*
				 * Loop until we find yet another one.
				 *
				 * By the time we get the soft_limit lock
				 * again, someone might have aded the
				 * group back on the RB tree. Iterate to
				 * make sure we get a different mem.
				 * mem_cgroup_largest_soft_limit_node returns
				 * NULL if no other cgroup is present on
				 * the tree
				 */
				next_mz =
				__mem_cgroup_largest_soft_limit_node(mctz);
				if (next_mz == mz) {
					css_put(&next_mz->mem->css);
					next_mz = NULL;
				} else /* next_mz == NULL or other memcg */
					break;
			} while (1);
		}
		__mem_cgroup_remove_exceeded(mz->mem, mz, mctz);
2277
		excess = res_counter_soft_limit_excess(&mz->mem->res);
2278 2279 2280 2281 2282 2283 2284 2285
		/*
		 * One school of thought says that we should not add
		 * back the node to the tree if reclaim returns 0.
		 * But our reclaim could return 0, simply because due
		 * to priority we are exposing a smaller subset of
		 * memory to reclaim from. Consider this as a longer
		 * term TODO.
		 */
2286 2287
		/* If excess == 0, no tree ops */
		__mem_cgroup_insert_exceeded(mz->mem, mz, mctz, excess);
2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305
		spin_unlock(&mctz->lock);
		css_put(&mz->mem->css);
		loop++;
		/*
		 * Could not reclaim anything and there are no more
		 * mem cgroups to try or we seem to be looping without
		 * reclaiming anything.
		 */
		if (!nr_reclaimed &&
			(next_mz == NULL ||
			loop > MEM_CGROUP_MAX_SOFT_LIMIT_RECLAIM_LOOPS))
			break;
	} while (!nr_reclaimed);
	if (next_mz)
		css_put(&next_mz->mem->css);
	return nr_reclaimed;
}

2306 2307 2308 2309
/*
 * This routine traverse page_cgroup in given list and drop them all.
 * *And* this routine doesn't reclaim page itself, just removes page_cgroup.
 */
2310
static int mem_cgroup_force_empty_list(struct mem_cgroup *mem,
K
KAMEZAWA Hiroyuki 已提交
2311
				int node, int zid, enum lru_list lru)
2312
{
K
KAMEZAWA Hiroyuki 已提交
2313 2314
	struct zone *zone;
	struct mem_cgroup_per_zone *mz;
2315
	struct page_cgroup *pc, *busy;
K
KAMEZAWA Hiroyuki 已提交
2316
	unsigned long flags, loop;
2317
	struct list_head *list;
2318
	int ret = 0;
2319

K
KAMEZAWA Hiroyuki 已提交
2320 2321
	zone = &NODE_DATA(node)->node_zones[zid];
	mz = mem_cgroup_zoneinfo(mem, node, zid);
2322
	list = &mz->lists[lru];
2323

2324 2325 2326 2327 2328 2329
	loop = MEM_CGROUP_ZSTAT(mz, lru);
	/* give some margin against EBUSY etc...*/
	loop += 256;
	busy = NULL;
	while (loop--) {
		ret = 0;
K
KAMEZAWA Hiroyuki 已提交
2330
		spin_lock_irqsave(&zone->lru_lock, flags);
2331
		if (list_empty(list)) {
K
KAMEZAWA Hiroyuki 已提交
2332
			spin_unlock_irqrestore(&zone->lru_lock, flags);
2333
			break;
2334 2335 2336 2337 2338
		}
		pc = list_entry(list->prev, struct page_cgroup, lru);
		if (busy == pc) {
			list_move(&pc->lru, list);
			busy = 0;
K
KAMEZAWA Hiroyuki 已提交
2339
			spin_unlock_irqrestore(&zone->lru_lock, flags);
2340 2341
			continue;
		}
K
KAMEZAWA Hiroyuki 已提交
2342
		spin_unlock_irqrestore(&zone->lru_lock, flags);
2343

K
KAMEZAWA Hiroyuki 已提交
2344
		ret = mem_cgroup_move_parent(pc, mem, GFP_KERNEL);
2345
		if (ret == -ENOMEM)
2346
			break;
2347 2348 2349 2350 2351 2352 2353

		if (ret == -EBUSY || ret == -EINVAL) {
			/* found lock contention or "pc" is obsolete. */
			busy = pc;
			cond_resched();
		} else
			busy = NULL;
2354
	}
K
KAMEZAWA Hiroyuki 已提交
2355

2356 2357 2358
	if (!ret && !list_empty(list))
		return -EBUSY;
	return ret;
2359 2360 2361 2362 2363 2364
}

/*
 * make mem_cgroup's charge to be 0 if there is no task.
 * This enables deleting this mem_cgroup.
 */
2365
static int mem_cgroup_force_empty(struct mem_cgroup *mem, bool free_all)
2366
{
2367 2368 2369
	int ret;
	int node, zid, shrink;
	int nr_retries = MEM_CGROUP_RECLAIM_RETRIES;
2370
	struct cgroup *cgrp = mem->css.cgroup;
2371

2372
	css_get(&mem->css);
2373 2374

	shrink = 0;
2375 2376 2377
	/* should free all ? */
	if (free_all)
		goto try_to_free;
2378
move_account:
2379
	while (mem->res.usage > 0) {
2380
		ret = -EBUSY;
2381 2382 2383 2384
		if (cgroup_task_count(cgrp) || !list_empty(&cgrp->children))
			goto out;
		ret = -EINTR;
		if (signal_pending(current))
2385
			goto out;
2386 2387
		/* This is for making all *used* pages to be on LRU. */
		lru_add_drain_all();
2388
		ret = 0;
2389
		for_each_node_state(node, N_HIGH_MEMORY) {
2390
			for (zid = 0; !ret && zid < MAX_NR_ZONES; zid++) {
2391
				enum lru_list l;
2392 2393
				for_each_lru(l) {
					ret = mem_cgroup_force_empty_list(mem,
K
KAMEZAWA Hiroyuki 已提交
2394
							node, zid, l);
2395 2396 2397
					if (ret)
						break;
				}
2398
			}
2399 2400 2401 2402 2403 2404
			if (ret)
				break;
		}
		/* it seems parent cgroup doesn't have enough mem */
		if (ret == -ENOMEM)
			goto try_to_free;
2405
		cond_resched();
2406 2407 2408 2409 2410
	}
	ret = 0;
out:
	css_put(&mem->css);
	return ret;
2411 2412

try_to_free:
2413 2414
	/* returns EBUSY if there is a task or if we come here twice. */
	if (cgroup_task_count(cgrp) || !list_empty(&cgrp->children) || shrink) {
2415 2416 2417
		ret = -EBUSY;
		goto out;
	}
2418 2419
	/* we call try-to-free pages for make this cgroup empty */
	lru_add_drain_all();
2420 2421 2422 2423
	/* try to free all pages in this cgroup */
	shrink = 1;
	while (nr_retries && mem->res.usage > 0) {
		int progress;
2424 2425 2426 2427 2428

		if (signal_pending(current)) {
			ret = -EINTR;
			goto out;
		}
K
KOSAKI Motohiro 已提交
2429 2430
		progress = try_to_free_mem_cgroup_pages(mem, GFP_KERNEL,
						false, get_swappiness(mem));
2431
		if (!progress) {
2432
			nr_retries--;
2433
			/* maybe some writeback is necessary */
2434
			congestion_wait(BLK_RW_ASYNC, HZ/10);
2435
		}
2436 2437

	}
K
KAMEZAWA Hiroyuki 已提交
2438
	lru_add_drain();
2439 2440 2441 2442 2443
	/* try move_account...there may be some *locked* pages. */
	if (mem->res.usage)
		goto move_account;
	ret = 0;
	goto out;
2444 2445
}

2446 2447 2448 2449 2450 2451
int mem_cgroup_force_empty_write(struct cgroup *cont, unsigned int event)
{
	return mem_cgroup_force_empty(mem_cgroup_from_cont(cont), true);
}


2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469
static u64 mem_cgroup_hierarchy_read(struct cgroup *cont, struct cftype *cft)
{
	return mem_cgroup_from_cont(cont)->use_hierarchy;
}

static int mem_cgroup_hierarchy_write(struct cgroup *cont, struct cftype *cft,
					u64 val)
{
	int retval = 0;
	struct mem_cgroup *mem = mem_cgroup_from_cont(cont);
	struct cgroup *parent = cont->parent;
	struct mem_cgroup *parent_mem = NULL;

	if (parent)
		parent_mem = mem_cgroup_from_cont(parent);

	cgroup_lock();
	/*
2470
	 * If parent's use_hierarchy is set, we can't make any modifications
2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487 2488 2489
	 * in the child subtrees. If it is unset, then the change can
	 * occur, provided the current cgroup has no children.
	 *
	 * For the root cgroup, parent_mem is NULL, we allow value to be
	 * set if there are no children.
	 */
	if ((!parent_mem || !parent_mem->use_hierarchy) &&
				(val == 1 || val == 0)) {
		if (list_empty(&cont->children))
			mem->use_hierarchy = val;
		else
			retval = -EBUSY;
	} else
		retval = -EINVAL;
	cgroup_unlock();

	return retval;
}

2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510 2511 2512 2513
struct mem_cgroup_idx_data {
	s64 val;
	enum mem_cgroup_stat_index idx;
};

static int
mem_cgroup_get_idx_stat(struct mem_cgroup *mem, void *data)
{
	struct mem_cgroup_idx_data *d = data;
	d->val += mem_cgroup_read_stat(&mem->stat, d->idx);
	return 0;
}

static void
mem_cgroup_get_recursive_idx_stat(struct mem_cgroup *mem,
				enum mem_cgroup_stat_index idx, s64 *val)
{
	struct mem_cgroup_idx_data d;
	d.idx = idx;
	d.val = 0;
	mem_cgroup_walk_tree(mem, &d, mem_cgroup_get_idx_stat);
	*val = d.val;
}

2514
static u64 mem_cgroup_read(struct cgroup *cont, struct cftype *cft)
B
Balbir Singh 已提交
2515
{
2516
	struct mem_cgroup *mem = mem_cgroup_from_cont(cont);
2517
	u64 idx_val, val;
2518 2519 2520 2521 2522 2523
	int type, name;

	type = MEMFILE_TYPE(cft->private);
	name = MEMFILE_ATTR(cft->private);
	switch (type) {
	case _MEM:
2524 2525 2526 2527 2528 2529 2530 2531 2532 2533
		if (name == RES_USAGE && mem_cgroup_is_root(mem)) {
			mem_cgroup_get_recursive_idx_stat(mem,
				MEM_CGROUP_STAT_CACHE, &idx_val);
			val = idx_val;
			mem_cgroup_get_recursive_idx_stat(mem,
				MEM_CGROUP_STAT_RSS, &idx_val);
			val += idx_val;
			val <<= PAGE_SHIFT;
		} else
			val = res_counter_read_u64(&mem->res, name);
2534 2535
		break;
	case _MEMSWAP:
2536 2537 2538 2539 2540 2541 2542 2543 2544
		if (name == RES_USAGE && mem_cgroup_is_root(mem)) {
			mem_cgroup_get_recursive_idx_stat(mem,
				MEM_CGROUP_STAT_CACHE, &idx_val);
			val = idx_val;
			mem_cgroup_get_recursive_idx_stat(mem,
				MEM_CGROUP_STAT_RSS, &idx_val);
			val += idx_val;
			mem_cgroup_get_recursive_idx_stat(mem,
				MEM_CGROUP_STAT_SWAPOUT, &idx_val);
2545
			val += idx_val;
2546 2547 2548
			val <<= PAGE_SHIFT;
		} else
			val = res_counter_read_u64(&mem->memsw, name);
2549 2550 2551 2552 2553 2554
		break;
	default:
		BUG();
		break;
	}
	return val;
B
Balbir Singh 已提交
2555
}
2556 2557 2558 2559
/*
 * The user of this function is...
 * RES_LIMIT.
 */
2560 2561
static int mem_cgroup_write(struct cgroup *cont, struct cftype *cft,
			    const char *buffer)
B
Balbir Singh 已提交
2562
{
2563
	struct mem_cgroup *memcg = mem_cgroup_from_cont(cont);
2564
	int type, name;
2565 2566 2567
	unsigned long long val;
	int ret;

2568 2569 2570
	type = MEMFILE_TYPE(cft->private);
	name = MEMFILE_ATTR(cft->private);
	switch (name) {
2571
	case RES_LIMIT:
2572 2573 2574 2575
		if (mem_cgroup_is_root(memcg)) { /* Can't set limit on root */
			ret = -EINVAL;
			break;
		}
2576 2577
		/* This function does all necessary parse...reuse it */
		ret = res_counter_memparse_write_strategy(buffer, &val);
2578 2579 2580
		if (ret)
			break;
		if (type == _MEM)
2581
			ret = mem_cgroup_resize_limit(memcg, val);
2582 2583
		else
			ret = mem_cgroup_resize_memsw_limit(memcg, val);
2584
		break;
2585 2586 2587 2588 2589 2590 2591 2592 2593 2594 2595 2596 2597 2598
	case RES_SOFT_LIMIT:
		ret = res_counter_memparse_write_strategy(buffer, &val);
		if (ret)
			break;
		/*
		 * For memsw, soft limits are hard to implement in terms
		 * of semantics, for now, we support soft limits for
		 * control without swap
		 */
		if (type == _MEM)
			ret = res_counter_set_soft_limit(&memcg->res, val);
		else
			ret = -EINVAL;
		break;
2599 2600 2601 2602 2603
	default:
		ret = -EINVAL; /* should be BUG() ? */
		break;
	}
	return ret;
B
Balbir Singh 已提交
2604 2605
}

2606 2607 2608 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 2624 2625 2626 2627 2628 2629 2630 2631 2632 2633
static void memcg_get_hierarchical_limit(struct mem_cgroup *memcg,
		unsigned long long *mem_limit, unsigned long long *memsw_limit)
{
	struct cgroup *cgroup;
	unsigned long long min_limit, min_memsw_limit, tmp;

	min_limit = res_counter_read_u64(&memcg->res, RES_LIMIT);
	min_memsw_limit = res_counter_read_u64(&memcg->memsw, RES_LIMIT);
	cgroup = memcg->css.cgroup;
	if (!memcg->use_hierarchy)
		goto out;

	while (cgroup->parent) {
		cgroup = cgroup->parent;
		memcg = mem_cgroup_from_cont(cgroup);
		if (!memcg->use_hierarchy)
			break;
		tmp = res_counter_read_u64(&memcg->res, RES_LIMIT);
		min_limit = min(min_limit, tmp);
		tmp = res_counter_read_u64(&memcg->memsw, RES_LIMIT);
		min_memsw_limit = min(min_memsw_limit, tmp);
	}
out:
	*mem_limit = min_limit;
	*memsw_limit = min_memsw_limit;
	return;
}

2634
static int mem_cgroup_reset(struct cgroup *cont, unsigned int event)
2635 2636
{
	struct mem_cgroup *mem;
2637
	int type, name;
2638 2639

	mem = mem_cgroup_from_cont(cont);
2640 2641 2642
	type = MEMFILE_TYPE(event);
	name = MEMFILE_ATTR(event);
	switch (name) {
2643
	case RES_MAX_USAGE:
2644 2645 2646 2647
		if (type == _MEM)
			res_counter_reset_max(&mem->res);
		else
			res_counter_reset_max(&mem->memsw);
2648 2649
		break;
	case RES_FAILCNT:
2650 2651 2652 2653
		if (type == _MEM)
			res_counter_reset_failcnt(&mem->res);
		else
			res_counter_reset_failcnt(&mem->memsw);
2654 2655
		break;
	}
2656

2657
	return 0;
2658 2659
}

K
KAMEZAWA Hiroyuki 已提交
2660 2661 2662 2663 2664

/* For read statistics */
enum {
	MCS_CACHE,
	MCS_RSS,
2665
	MCS_MAPPED_FILE,
K
KAMEZAWA Hiroyuki 已提交
2666 2667
	MCS_PGPGIN,
	MCS_PGPGOUT,
2668
	MCS_SWAP,
K
KAMEZAWA Hiroyuki 已提交
2669 2670 2671 2672 2673 2674 2675 2676 2677 2678
	MCS_INACTIVE_ANON,
	MCS_ACTIVE_ANON,
	MCS_INACTIVE_FILE,
	MCS_ACTIVE_FILE,
	MCS_UNEVICTABLE,
	NR_MCS_STAT,
};

struct mcs_total_stat {
	s64 stat[NR_MCS_STAT];
2679 2680
};

K
KAMEZAWA Hiroyuki 已提交
2681 2682 2683 2684 2685 2686
struct {
	char *local_name;
	char *total_name;
} memcg_stat_strings[NR_MCS_STAT] = {
	{"cache", "total_cache"},
	{"rss", "total_rss"},
2687
	{"mapped_file", "total_mapped_file"},
K
KAMEZAWA Hiroyuki 已提交
2688 2689
	{"pgpgin", "total_pgpgin"},
	{"pgpgout", "total_pgpgout"},
2690
	{"swap", "total_swap"},
K
KAMEZAWA Hiroyuki 已提交
2691 2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704 2705 2706 2707 2708
	{"inactive_anon", "total_inactive_anon"},
	{"active_anon", "total_active_anon"},
	{"inactive_file", "total_inactive_file"},
	{"active_file", "total_active_file"},
	{"unevictable", "total_unevictable"}
};


static int mem_cgroup_get_local_stat(struct mem_cgroup *mem, void *data)
{
	struct mcs_total_stat *s = data;
	s64 val;

	/* per cpu stat */
	val = mem_cgroup_read_stat(&mem->stat, MEM_CGROUP_STAT_CACHE);
	s->stat[MCS_CACHE] += val * PAGE_SIZE;
	val = mem_cgroup_read_stat(&mem->stat, MEM_CGROUP_STAT_RSS);
	s->stat[MCS_RSS] += val * PAGE_SIZE;
2709 2710
	val = mem_cgroup_read_stat(&mem->stat, MEM_CGROUP_STAT_MAPPED_FILE);
	s->stat[MCS_MAPPED_FILE] += val * PAGE_SIZE;
K
KAMEZAWA Hiroyuki 已提交
2711 2712 2713 2714
	val = mem_cgroup_read_stat(&mem->stat, MEM_CGROUP_STAT_PGPGIN_COUNT);
	s->stat[MCS_PGPGIN] += val;
	val = mem_cgroup_read_stat(&mem->stat, MEM_CGROUP_STAT_PGPGOUT_COUNT);
	s->stat[MCS_PGPGOUT] += val;
2715 2716 2717 2718
	if (do_swap_account) {
		val = mem_cgroup_read_stat(&mem->stat, MEM_CGROUP_STAT_SWAPOUT);
		s->stat[MCS_SWAP] += val * PAGE_SIZE;
	}
K
KAMEZAWA Hiroyuki 已提交
2719 2720 2721 2722 2723 2724 2725 2726 2727 2728 2729 2730 2731 2732 2733 2734 2735 2736 2737 2738 2739

	/* per zone stat */
	val = mem_cgroup_get_local_zonestat(mem, LRU_INACTIVE_ANON);
	s->stat[MCS_INACTIVE_ANON] += val * PAGE_SIZE;
	val = mem_cgroup_get_local_zonestat(mem, LRU_ACTIVE_ANON);
	s->stat[MCS_ACTIVE_ANON] += val * PAGE_SIZE;
	val = mem_cgroup_get_local_zonestat(mem, LRU_INACTIVE_FILE);
	s->stat[MCS_INACTIVE_FILE] += val * PAGE_SIZE;
	val = mem_cgroup_get_local_zonestat(mem, LRU_ACTIVE_FILE);
	s->stat[MCS_ACTIVE_FILE] += val * PAGE_SIZE;
	val = mem_cgroup_get_local_zonestat(mem, LRU_UNEVICTABLE);
	s->stat[MCS_UNEVICTABLE] += val * PAGE_SIZE;
	return 0;
}

static void
mem_cgroup_get_total_stat(struct mem_cgroup *mem, struct mcs_total_stat *s)
{
	mem_cgroup_walk_tree(mem, s, mem_cgroup_get_local_stat);
}

2740 2741
static int mem_control_stat_show(struct cgroup *cont, struct cftype *cft,
				 struct cgroup_map_cb *cb)
2742 2743
{
	struct mem_cgroup *mem_cont = mem_cgroup_from_cont(cont);
K
KAMEZAWA Hiroyuki 已提交
2744
	struct mcs_total_stat mystat;
2745 2746
	int i;

K
KAMEZAWA Hiroyuki 已提交
2747 2748
	memset(&mystat, 0, sizeof(mystat));
	mem_cgroup_get_local_stat(mem_cont, &mystat);
2749

2750 2751 2752
	for (i = 0; i < NR_MCS_STAT; i++) {
		if (i == MCS_SWAP && !do_swap_account)
			continue;
K
KAMEZAWA Hiroyuki 已提交
2753
		cb->fill(cb, memcg_stat_strings[i].local_name, mystat.stat[i]);
2754
	}
L
Lee Schermerhorn 已提交
2755

K
KAMEZAWA Hiroyuki 已提交
2756
	/* Hierarchical information */
2757 2758 2759 2760 2761 2762 2763
	{
		unsigned long long limit, memsw_limit;
		memcg_get_hierarchical_limit(mem_cont, &limit, &memsw_limit);
		cb->fill(cb, "hierarchical_memory_limit", limit);
		if (do_swap_account)
			cb->fill(cb, "hierarchical_memsw_limit", memsw_limit);
	}
K
KOSAKI Motohiro 已提交
2764

K
KAMEZAWA Hiroyuki 已提交
2765 2766
	memset(&mystat, 0, sizeof(mystat));
	mem_cgroup_get_total_stat(mem_cont, &mystat);
2767 2768 2769
	for (i = 0; i < NR_MCS_STAT; i++) {
		if (i == MCS_SWAP && !do_swap_account)
			continue;
K
KAMEZAWA Hiroyuki 已提交
2770
		cb->fill(cb, memcg_stat_strings[i].total_name, mystat.stat[i]);
2771
	}
K
KAMEZAWA Hiroyuki 已提交
2772

K
KOSAKI Motohiro 已提交
2773
#ifdef CONFIG_DEBUG_VM
2774
	cb->fill(cb, "inactive_ratio", calc_inactive_ratio(mem_cont, NULL));
K
KOSAKI Motohiro 已提交
2775 2776 2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799 2800 2801

	{
		int nid, zid;
		struct mem_cgroup_per_zone *mz;
		unsigned long recent_rotated[2] = {0, 0};
		unsigned long recent_scanned[2] = {0, 0};

		for_each_online_node(nid)
			for (zid = 0; zid < MAX_NR_ZONES; zid++) {
				mz = mem_cgroup_zoneinfo(mem_cont, nid, zid);

				recent_rotated[0] +=
					mz->reclaim_stat.recent_rotated[0];
				recent_rotated[1] +=
					mz->reclaim_stat.recent_rotated[1];
				recent_scanned[0] +=
					mz->reclaim_stat.recent_scanned[0];
				recent_scanned[1] +=
					mz->reclaim_stat.recent_scanned[1];
			}
		cb->fill(cb, "recent_rotated_anon", recent_rotated[0]);
		cb->fill(cb, "recent_rotated_file", recent_rotated[1]);
		cb->fill(cb, "recent_scanned_anon", recent_scanned[0]);
		cb->fill(cb, "recent_scanned_file", recent_scanned[1]);
	}
#endif

2802 2803 2804
	return 0;
}

K
KOSAKI Motohiro 已提交
2805 2806 2807 2808 2809 2810 2811 2812 2813 2814 2815 2816
static u64 mem_cgroup_swappiness_read(struct cgroup *cgrp, struct cftype *cft)
{
	struct mem_cgroup *memcg = mem_cgroup_from_cont(cgrp);

	return get_swappiness(memcg);
}

static int mem_cgroup_swappiness_write(struct cgroup *cgrp, struct cftype *cft,
				       u64 val)
{
	struct mem_cgroup *memcg = mem_cgroup_from_cont(cgrp);
	struct mem_cgroup *parent;
2817

K
KOSAKI Motohiro 已提交
2818 2819 2820 2821 2822 2823 2824
	if (val > 100)
		return -EINVAL;

	if (cgrp->parent == NULL)
		return -EINVAL;

	parent = mem_cgroup_from_cont(cgrp->parent);
2825 2826 2827

	cgroup_lock();

K
KOSAKI Motohiro 已提交
2828 2829
	/* If under hierarchy, only empty-root can set this value */
	if ((parent->use_hierarchy) ||
2830 2831
	    (memcg->use_hierarchy && !list_empty(&cgrp->children))) {
		cgroup_unlock();
K
KOSAKI Motohiro 已提交
2832
		return -EINVAL;
2833
	}
K
KOSAKI Motohiro 已提交
2834 2835 2836 2837 2838

	spin_lock(&memcg->reclaim_param_lock);
	memcg->swappiness = val;
	spin_unlock(&memcg->reclaim_param_lock);

2839 2840
	cgroup_unlock();

K
KOSAKI Motohiro 已提交
2841 2842 2843
	return 0;
}

2844

B
Balbir Singh 已提交
2845 2846
static struct cftype mem_cgroup_files[] = {
	{
2847
		.name = "usage_in_bytes",
2848
		.private = MEMFILE_PRIVATE(_MEM, RES_USAGE),
2849
		.read_u64 = mem_cgroup_read,
B
Balbir Singh 已提交
2850
	},
2851 2852
	{
		.name = "max_usage_in_bytes",
2853
		.private = MEMFILE_PRIVATE(_MEM, RES_MAX_USAGE),
2854
		.trigger = mem_cgroup_reset,
2855 2856
		.read_u64 = mem_cgroup_read,
	},
B
Balbir Singh 已提交
2857
	{
2858
		.name = "limit_in_bytes",
2859
		.private = MEMFILE_PRIVATE(_MEM, RES_LIMIT),
2860
		.write_string = mem_cgroup_write,
2861
		.read_u64 = mem_cgroup_read,
B
Balbir Singh 已提交
2862
	},
2863 2864 2865 2866 2867 2868
	{
		.name = "soft_limit_in_bytes",
		.private = MEMFILE_PRIVATE(_MEM, RES_SOFT_LIMIT),
		.write_string = mem_cgroup_write,
		.read_u64 = mem_cgroup_read,
	},
B
Balbir Singh 已提交
2869 2870
	{
		.name = "failcnt",
2871
		.private = MEMFILE_PRIVATE(_MEM, RES_FAILCNT),
2872
		.trigger = mem_cgroup_reset,
2873
		.read_u64 = mem_cgroup_read,
B
Balbir Singh 已提交
2874
	},
2875 2876
	{
		.name = "stat",
2877
		.read_map = mem_control_stat_show,
2878
	},
2879 2880 2881 2882
	{
		.name = "force_empty",
		.trigger = mem_cgroup_force_empty_write,
	},
2883 2884 2885 2886 2887
	{
		.name = "use_hierarchy",
		.write_u64 = mem_cgroup_hierarchy_write,
		.read_u64 = mem_cgroup_hierarchy_read,
	},
K
KOSAKI Motohiro 已提交
2888 2889 2890 2891 2892
	{
		.name = "swappiness",
		.read_u64 = mem_cgroup_swappiness_read,
		.write_u64 = mem_cgroup_swappiness_write,
	},
B
Balbir Singh 已提交
2893 2894
};

2895 2896 2897 2898 2899 2900 2901 2902 2903 2904 2905 2906 2907 2908 2909 2910 2911 2912 2913 2914 2915 2916 2917 2918 2919 2920 2921 2922 2923 2924 2925 2926 2927 2928 2929 2930 2931 2932 2933 2934 2935
#ifdef CONFIG_CGROUP_MEM_RES_CTLR_SWAP
static struct cftype memsw_cgroup_files[] = {
	{
		.name = "memsw.usage_in_bytes",
		.private = MEMFILE_PRIVATE(_MEMSWAP, RES_USAGE),
		.read_u64 = mem_cgroup_read,
	},
	{
		.name = "memsw.max_usage_in_bytes",
		.private = MEMFILE_PRIVATE(_MEMSWAP, RES_MAX_USAGE),
		.trigger = mem_cgroup_reset,
		.read_u64 = mem_cgroup_read,
	},
	{
		.name = "memsw.limit_in_bytes",
		.private = MEMFILE_PRIVATE(_MEMSWAP, RES_LIMIT),
		.write_string = mem_cgroup_write,
		.read_u64 = mem_cgroup_read,
	},
	{
		.name = "memsw.failcnt",
		.private = MEMFILE_PRIVATE(_MEMSWAP, RES_FAILCNT),
		.trigger = mem_cgroup_reset,
		.read_u64 = mem_cgroup_read,
	},
};

static int register_memsw_files(struct cgroup *cont, struct cgroup_subsys *ss)
{
	if (!do_swap_account)
		return 0;
	return cgroup_add_files(cont, ss, memsw_cgroup_files,
				ARRAY_SIZE(memsw_cgroup_files));
};
#else
static int register_memsw_files(struct cgroup *cont, struct cgroup_subsys *ss)
{
	return 0;
}
#endif

2936 2937 2938
static int alloc_mem_cgroup_per_zone_info(struct mem_cgroup *mem, int node)
{
	struct mem_cgroup_per_node *pn;
2939
	struct mem_cgroup_per_zone *mz;
2940
	enum lru_list l;
2941
	int zone, tmp = node;
2942 2943 2944 2945 2946 2947 2948 2949
	/*
	 * This routine is called against possible nodes.
	 * But it's BUG to call kmalloc() against offline node.
	 *
	 * TODO: this routine can waste much memory for nodes which will
	 *       never be onlined. It's better to use memory hotplug callback
	 *       function.
	 */
2950 2951 2952
	if (!node_state(node, N_NORMAL_MEMORY))
		tmp = -1;
	pn = kmalloc_node(sizeof(*pn), GFP_KERNEL, tmp);
2953 2954
	if (!pn)
		return 1;
2955

2956 2957
	mem->info.nodeinfo[node] = pn;
	memset(pn, 0, sizeof(*pn));
2958 2959 2960

	for (zone = 0; zone < MAX_NR_ZONES; zone++) {
		mz = &pn->zoneinfo[zone];
2961 2962
		for_each_lru(l)
			INIT_LIST_HEAD(&mz->lists[l]);
2963
		mz->usage_in_excess = 0;
2964 2965
		mz->on_tree = false;
		mz->mem = mem;
2966
	}
2967 2968 2969
	return 0;
}

2970 2971 2972 2973 2974
static void free_mem_cgroup_per_zone_info(struct mem_cgroup *mem, int node)
{
	kfree(mem->info.nodeinfo[node]);
}

2975 2976 2977 2978 2979 2980
static int mem_cgroup_size(void)
{
	int cpustat_size = nr_cpu_ids * sizeof(struct mem_cgroup_stat_cpu);
	return sizeof(struct mem_cgroup) + cpustat_size;
}

2981 2982 2983
static struct mem_cgroup *mem_cgroup_alloc(void)
{
	struct mem_cgroup *mem;
2984
	int size = mem_cgroup_size();
2985

2986 2987
	if (size < PAGE_SIZE)
		mem = kmalloc(size, GFP_KERNEL);
2988
	else
2989
		mem = vmalloc(size);
2990 2991

	if (mem)
2992
		memset(mem, 0, size);
2993 2994 2995
	return mem;
}

2996 2997 2998 2999 3000 3001 3002 3003 3004 3005 3006
/*
 * At destroying mem_cgroup, references from swap_cgroup can remain.
 * (scanning all at force_empty is too costly...)
 *
 * Instead of clearing all references at force_empty, we remember
 * the number of reference from swap_cgroup and free mem_cgroup when
 * it goes down to 0.
 *
 * Removal of cgroup itself succeeds regardless of refs from swap.
 */

3007
static void __mem_cgroup_free(struct mem_cgroup *mem)
3008
{
K
KAMEZAWA Hiroyuki 已提交
3009 3010
	int node;

3011
	mem_cgroup_remove_from_trees(mem);
K
KAMEZAWA Hiroyuki 已提交
3012 3013
	free_css_id(&mem_cgroup_subsys, &mem->css);

K
KAMEZAWA Hiroyuki 已提交
3014 3015 3016
	for_each_node_state(node, N_POSSIBLE)
		free_mem_cgroup_per_zone_info(mem, node);

3017
	if (mem_cgroup_size() < PAGE_SIZE)
3018 3019 3020 3021 3022
		kfree(mem);
	else
		vfree(mem);
}

3023 3024 3025 3026 3027 3028 3029
static void mem_cgroup_get(struct mem_cgroup *mem)
{
	atomic_inc(&mem->refcnt);
}

static void mem_cgroup_put(struct mem_cgroup *mem)
{
3030 3031
	if (atomic_dec_and_test(&mem->refcnt)) {
		struct mem_cgroup *parent = parent_mem_cgroup(mem);
3032
		__mem_cgroup_free(mem);
3033 3034 3035
		if (parent)
			mem_cgroup_put(parent);
	}
3036 3037
}

3038 3039 3040 3041 3042 3043 3044 3045 3046
/*
 * Returns the parent mem_cgroup in memcgroup hierarchy with hierarchy enabled.
 */
static struct mem_cgroup *parent_mem_cgroup(struct mem_cgroup *mem)
{
	if (!mem->res.parent)
		return NULL;
	return mem_cgroup_from_res_counter(mem->res.parent, res);
}
3047

3048 3049 3050
#ifdef CONFIG_CGROUP_MEM_RES_CTLR_SWAP
static void __init enable_swap_cgroup(void)
{
3051
	if (!mem_cgroup_disabled() && really_do_swap_account)
3052 3053 3054 3055 3056 3057 3058 3059
		do_swap_account = 1;
}
#else
static void __init enable_swap_cgroup(void)
{
}
#endif

3060 3061 3062 3063 3064 3065 3066 3067 3068 3069 3070 3071 3072 3073 3074 3075 3076 3077 3078 3079 3080 3081 3082 3083 3084
static int mem_cgroup_soft_limit_tree_init(void)
{
	struct mem_cgroup_tree_per_node *rtpn;
	struct mem_cgroup_tree_per_zone *rtpz;
	int tmp, node, zone;

	for_each_node_state(node, N_POSSIBLE) {
		tmp = node;
		if (!node_state(node, N_NORMAL_MEMORY))
			tmp = -1;
		rtpn = kzalloc_node(sizeof(*rtpn), GFP_KERNEL, tmp);
		if (!rtpn)
			return 1;

		soft_limit_tree.rb_tree_per_node[node] = rtpn;

		for (zone = 0; zone < MAX_NR_ZONES; zone++) {
			rtpz = &rtpn->rb_tree_per_zone[zone];
			rtpz->rb_root = RB_ROOT;
			spin_lock_init(&rtpz->lock);
		}
	}
	return 0;
}

L
Li Zefan 已提交
3085
static struct cgroup_subsys_state * __ref
B
Balbir Singh 已提交
3086 3087
mem_cgroup_create(struct cgroup_subsys *ss, struct cgroup *cont)
{
3088
	struct mem_cgroup *mem, *parent;
K
KAMEZAWA Hiroyuki 已提交
3089
	long error = -ENOMEM;
3090
	int node;
B
Balbir Singh 已提交
3091

3092 3093
	mem = mem_cgroup_alloc();
	if (!mem)
K
KAMEZAWA Hiroyuki 已提交
3094
		return ERR_PTR(error);
3095

3096 3097 3098
	for_each_node_state(node, N_POSSIBLE)
		if (alloc_mem_cgroup_per_zone_info(mem, node))
			goto free_out;
3099

3100
	/* root ? */
3101
	if (cont->parent == NULL) {
3102
		enable_swap_cgroup();
3103
		parent = NULL;
3104
		root_mem_cgroup = mem;
3105 3106 3107
		if (mem_cgroup_soft_limit_tree_init())
			goto free_out;

3108
	} else {
3109
		parent = mem_cgroup_from_cont(cont->parent);
3110 3111
		mem->use_hierarchy = parent->use_hierarchy;
	}
3112

3113 3114 3115
	if (parent && parent->use_hierarchy) {
		res_counter_init(&mem->res, &parent->res);
		res_counter_init(&mem->memsw, &parent->memsw);
3116 3117 3118 3119 3120 3121 3122
		/*
		 * We increment refcnt of the parent to ensure that we can
		 * safely access it on res_counter_charge/uncharge.
		 * This refcnt will be decremented when freeing this
		 * mem_cgroup(see mem_cgroup_put).
		 */
		mem_cgroup_get(parent);
3123 3124 3125 3126
	} else {
		res_counter_init(&mem->res, NULL);
		res_counter_init(&mem->memsw, NULL);
	}
K
KAMEZAWA Hiroyuki 已提交
3127
	mem->last_scanned_child = 0;
K
KOSAKI Motohiro 已提交
3128
	spin_lock_init(&mem->reclaim_param_lock);
3129

K
KOSAKI Motohiro 已提交
3130 3131
	if (parent)
		mem->swappiness = get_swappiness(parent);
3132
	atomic_set(&mem->refcnt, 1);
B
Balbir Singh 已提交
3133
	return &mem->css;
3134
free_out:
3135
	__mem_cgroup_free(mem);
3136
	root_mem_cgroup = NULL;
K
KAMEZAWA Hiroyuki 已提交
3137
	return ERR_PTR(error);
B
Balbir Singh 已提交
3138 3139
}

3140
static int mem_cgroup_pre_destroy(struct cgroup_subsys *ss,
3141 3142 3143
					struct cgroup *cont)
{
	struct mem_cgroup *mem = mem_cgroup_from_cont(cont);
3144 3145

	return mem_cgroup_force_empty(mem, false);
3146 3147
}

B
Balbir Singh 已提交
3148 3149 3150
static void mem_cgroup_destroy(struct cgroup_subsys *ss,
				struct cgroup *cont)
{
3151 3152 3153
	struct mem_cgroup *mem = mem_cgroup_from_cont(cont);

	mem_cgroup_put(mem);
B
Balbir Singh 已提交
3154 3155 3156 3157 3158
}

static int mem_cgroup_populate(struct cgroup_subsys *ss,
				struct cgroup *cont)
{
3159 3160 3161 3162 3163 3164 3165 3166
	int ret;

	ret = cgroup_add_files(cont, ss, mem_cgroup_files,
				ARRAY_SIZE(mem_cgroup_files));

	if (!ret)
		ret = register_memsw_files(cont, ss);
	return ret;
B
Balbir Singh 已提交
3167 3168
}

B
Balbir Singh 已提交
3169 3170 3171
static void mem_cgroup_move_task(struct cgroup_subsys *ss,
				struct cgroup *cont,
				struct cgroup *old_cont,
3172 3173
				struct task_struct *p,
				bool threadgroup)
B
Balbir Singh 已提交
3174
{
3175
	mutex_lock(&memcg_tasklist);
B
Balbir Singh 已提交
3176
	/*
3177 3178
	 * FIXME: It's better to move charges of this process from old
	 * memcg to new memcg. But it's just on TODO-List now.
B
Balbir Singh 已提交
3179
	 */
3180
	mutex_unlock(&memcg_tasklist);
B
Balbir Singh 已提交
3181 3182
}

B
Balbir Singh 已提交
3183 3184 3185 3186
struct cgroup_subsys mem_cgroup_subsys = {
	.name = "memory",
	.subsys_id = mem_cgroup_subsys_id,
	.create = mem_cgroup_create,
3187
	.pre_destroy = mem_cgroup_pre_destroy,
B
Balbir Singh 已提交
3188 3189
	.destroy = mem_cgroup_destroy,
	.populate = mem_cgroup_populate,
B
Balbir Singh 已提交
3190
	.attach = mem_cgroup_move_task,
3191
	.early_init = 0,
K
KAMEZAWA Hiroyuki 已提交
3192
	.use_id = 1,
B
Balbir Singh 已提交
3193
};
3194 3195 3196 3197 3198 3199 3200 3201 3202 3203

#ifdef CONFIG_CGROUP_MEM_RES_CTLR_SWAP

static int __init disable_swap_account(char *s)
{
	really_do_swap_account = 0;
	return 1;
}
__setup("noswapaccount", disable_swap_account);
#endif