memcontrol.c 56.1 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/mutex.h>
31
#include <linux/slab.h>
32 33 34
#include <linux/swap.h>
#include <linux/spinlock.h>
#include <linux/fs.h>
35
#include <linux/seq_file.h>
36
#include <linux/vmalloc.h>
37
#include <linux/mm_inline.h>
38
#include <linux/page_cgroup.h>
K
KAMEZAWA Hiroyuki 已提交
39
#include "internal.h"
B
Balbir Singh 已提交
40

41 42
#include <asm/uaccess.h>

43 44
struct cgroup_subsys mem_cgroup_subsys __read_mostly;
#define MEM_CGROUP_RECLAIM_RETRIES	5
B
Balbir Singh 已提交
45

46 47 48 49 50 51 52 53
#ifdef CONFIG_CGROUP_MEM_RES_CTLR_SWAP
/* Turned on only when memory cgroup is enabled && really_do_swap_account = 0 */
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

54
static DEFINE_MUTEX(memcg_tasklist);	/* can be hold under cgroup_mutex */
55

56 57 58 59 60 61 62 63 64
/*
 * 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 */
	MEM_CGROUP_STAT_RSS,	   /* # of pages charged as rss */
65 66
	MEM_CGROUP_STAT_PGPGIN_COUNT,	/* # of pages paged in */
	MEM_CGROUP_STAT_PGPGOUT_COUNT,	/* # of pages paged out */
67 68 69 70 71 72 73 74 75

	MEM_CGROUP_STAT_NSTATS,
};

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

struct mem_cgroup_stat {
76
	struct mem_cgroup_stat_cpu cpustat[0];
77 78 79 80 81
};

/*
 * For accounting under irq disable, no need for increment preempt count.
 */
82
static inline void __mem_cgroup_stat_add_safe(struct mem_cgroup_stat_cpu *stat,
83 84
		enum mem_cgroup_stat_index idx, int val)
{
85
	stat->count[idx] += val;
86 87 88 89 90 91 92 93 94 95 96 97
}

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;
}

98 99 100 101
/*
 * per-zone information in memory controller.
 */
struct mem_cgroup_per_zone {
102 103 104
	/*
	 * spin_lock to protect the per cgroup LRU
	 */
105 106
	struct list_head	lists[NR_LRU_LISTS];
	unsigned long		count[NR_LRU_LISTS];
K
KOSAKI Motohiro 已提交
107 108

	struct zone_reclaim_stat reclaim_stat;
109 110 111 112 113 114 115 116 117 118 119 120
};
/* 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];
};

B
Balbir Singh 已提交
121 122 123 124 125 126 127
/*
 * 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
128 129 130
 * 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
131 132 133 134 135 136 137
 */
struct mem_cgroup {
	struct cgroup_subsys_state css;
	/*
	 * the counter to account for memory usage
	 */
	struct res_counter res;
138 139 140 141
	/*
	 * the counter to account for mem+swap usage.
	 */
	struct res_counter memsw;
142 143 144 145
	/*
	 * Per cgroup active and inactive list, similar to the
	 * per zone LRU lists.
	 */
146
	struct mem_cgroup_lru_info info;
147

K
KOSAKI Motohiro 已提交
148 149 150 151 152
	/*
	  protect against reclaim related member.
	*/
	spinlock_t reclaim_param_lock;

153
	int	prev_priority;	/* for recording reclaim priority */
154 155 156

	/*
	 * While reclaiming in a hiearchy, we cache the last child we
157
	 * reclaimed from. Protected by hierarchy_mutex
158 159
	 */
	struct mem_cgroup *last_scanned_child;
160 161 162 163
	/*
	 * Should the accounting and control be hierarchical, per subtree?
	 */
	bool use_hierarchy;
164
	unsigned long	last_oom_jiffies;
165
	atomic_t	refcnt;
166

K
KOSAKI Motohiro 已提交
167 168
	unsigned int	swappiness;

169
	/*
170
	 * statistics. This must be placed at the end of memcg.
171 172
	 */
	struct mem_cgroup_stat stat;
B
Balbir Singh 已提交
173 174
};

175 176 177
enum charge_type {
	MEM_CGROUP_CHARGE_TYPE_CACHE = 0,
	MEM_CGROUP_CHARGE_TYPE_MAPPED,
178
	MEM_CGROUP_CHARGE_TYPE_SHMEM,	/* used by page migration of shmem */
179
	MEM_CGROUP_CHARGE_TYPE_FORCE,	/* used by force_empty */
K
KAMEZAWA Hiroyuki 已提交
180
	MEM_CGROUP_CHARGE_TYPE_SWAPOUT,	/* for accounting swapcache */
181 182 183
	NR_CHARGE_TYPE,
};

184 185 186 187
/* only for here (for easy reading.) */
#define PCGF_CACHE	(1UL << PCG_CACHE)
#define PCGF_USED	(1UL << PCG_USED)
#define PCGF_LOCK	(1UL << PCG_LOCK)
188 189
static const unsigned long
pcg_default_flags[NR_CHARGE_TYPE] = {
K
KAMEZAWA Hiroyuki 已提交
190 191 192
	PCGF_CACHE | PCGF_USED | PCGF_LOCK, /* File Cache */
	PCGF_USED | PCGF_LOCK, /* Anon */
	PCGF_CACHE | PCGF_USED | PCGF_LOCK, /* Shmem */
193
	0, /* FORCE */
194 195
};

196 197 198 199 200 201 202 203 204 205
/* 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)

static void mem_cgroup_get(struct mem_cgroup *mem);
static void mem_cgroup_put(struct mem_cgroup *mem);

206 207 208
static void mem_cgroup_charge_statistics(struct mem_cgroup *mem,
					 struct page_cgroup *pc,
					 bool charge)
209 210 211
{
	int val = (charge)? 1 : -1;
	struct mem_cgroup_stat *stat = &mem->stat;
212
	struct mem_cgroup_stat_cpu *cpustat;
K
KAMEZAWA Hiroyuki 已提交
213
	int cpu = get_cpu();
214

K
KAMEZAWA Hiroyuki 已提交
215
	cpustat = &stat->cpustat[cpu];
216
	if (PageCgroupCache(pc))
217
		__mem_cgroup_stat_add_safe(cpustat, MEM_CGROUP_STAT_CACHE, val);
218
	else
219
		__mem_cgroup_stat_add_safe(cpustat, MEM_CGROUP_STAT_RSS, val);
220 221

	if (charge)
222
		__mem_cgroup_stat_add_safe(cpustat,
223 224
				MEM_CGROUP_STAT_PGPGIN_COUNT, 1);
	else
225
		__mem_cgroup_stat_add_safe(cpustat,
226
				MEM_CGROUP_STAT_PGPGOUT_COUNT, 1);
K
KAMEZAWA Hiroyuki 已提交
227
	put_cpu();
228 229
}

230
static struct mem_cgroup_per_zone *
231 232 233 234 235
mem_cgroup_zoneinfo(struct mem_cgroup *mem, int nid, int zid)
{
	return &mem->info.nodeinfo[nid]->zoneinfo[zid];
}

236
static struct mem_cgroup_per_zone *
237 238 239 240 241
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);
242

243 244 245
	if (!mem)
		return NULL;

246 247 248 249
	return mem_cgroup_zoneinfo(mem, nid, zid);
}

static unsigned long mem_cgroup_get_all_zonestat(struct mem_cgroup *mem,
250
					enum lru_list idx)
251 252 253 254 255 256 257 258 259 260 261
{
	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;
262 263
}

264
static struct mem_cgroup *mem_cgroup_from_cont(struct cgroup *cont)
B
Balbir Singh 已提交
265 266 267 268 269 270
{
	return container_of(cgroup_subsys_state(cont,
				mem_cgroup_subsys_id), struct mem_cgroup,
				css);
}

271
struct mem_cgroup *mem_cgroup_from_task(struct task_struct *p)
272
{
273 274 275 276 277 278 279 280
	/*
	 * 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;

281 282 283 284
	return container_of(task_subsys_state(p, mem_cgroup_subsys_id),
				struct mem_cgroup, css);
}

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
static struct mem_cgroup *try_get_mem_cgroup_from_mm(struct mm_struct *mm)
{
	struct mem_cgroup *mem = NULL;
	/*
	 * 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;
}

static bool mem_cgroup_is_obsolete(struct mem_cgroup *mem)
{
	if (!mem)
		return true;
	return css_is_removed(&mem->css);
}

K
KAMEZAWA Hiroyuki 已提交
310 311 312 313 314 315 316 317 318 319 320 321 322
/*
 * 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.
 */
323

K
KAMEZAWA Hiroyuki 已提交
324 325 326 327 328
void mem_cgroup_del_lru_list(struct page *page, enum lru_list lru)
{
	struct page_cgroup *pc;
	struct mem_cgroup *mem;
	struct mem_cgroup_per_zone *mz;
329

330
	if (mem_cgroup_disabled())
K
KAMEZAWA Hiroyuki 已提交
331 332 333
		return;
	pc = lookup_page_cgroup(page);
	/* can happen while we handle swapcache. */
334
	if (list_empty(&pc->lru) || !pc->mem_cgroup)
K
KAMEZAWA Hiroyuki 已提交
335
		return;
336 337 338 339
	/*
	 * 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
340 341
	mz = page_cgroup_zoneinfo(pc);
	mem = pc->mem_cgroup;
342
	MEM_CGROUP_ZSTAT(mz, lru) -= 1;
K
KAMEZAWA Hiroyuki 已提交
343 344
	list_del_init(&pc->lru);
	return;
345 346
}

K
KAMEZAWA Hiroyuki 已提交
347
void mem_cgroup_del_lru(struct page *page)
348
{
K
KAMEZAWA Hiroyuki 已提交
349 350
	mem_cgroup_del_lru_list(page, page_lru(page));
}
351

K
KAMEZAWA Hiroyuki 已提交
352 353 354 355
void mem_cgroup_rotate_lru_list(struct page *page, enum lru_list lru)
{
	struct mem_cgroup_per_zone *mz;
	struct page_cgroup *pc;
356

357
	if (mem_cgroup_disabled())
K
KAMEZAWA Hiroyuki 已提交
358
		return;
359

K
KAMEZAWA Hiroyuki 已提交
360
	pc = lookup_page_cgroup(page);
361 362 363 364
	/*
	 * 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
365 366 367 368 369 370
	smp_rmb();
	/* unused page is not rotated. */
	if (!PageCgroupUsed(pc))
		return;
	mz = page_cgroup_zoneinfo(pc);
	list_move(&pc->lru, &mz->lists[lru]);
371 372
}

K
KAMEZAWA Hiroyuki 已提交
373
void mem_cgroup_add_lru_list(struct page *page, enum lru_list lru)
374
{
K
KAMEZAWA Hiroyuki 已提交
375 376
	struct page_cgroup *pc;
	struct mem_cgroup_per_zone *mz;
377

378
	if (mem_cgroup_disabled())
K
KAMEZAWA Hiroyuki 已提交
379 380
		return;
	pc = lookup_page_cgroup(page);
381 382 383 384
	/*
	 * 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
385 386
	smp_rmb();
	if (!PageCgroupUsed(pc))
L
Lee Schermerhorn 已提交
387
		return;
388

K
KAMEZAWA Hiroyuki 已提交
389
	mz = page_cgroup_zoneinfo(pc);
390
	MEM_CGROUP_ZSTAT(mz, lru) += 1;
K
KAMEZAWA Hiroyuki 已提交
391 392
	list_add(&pc->lru, &mz->lists[lru]);
}
393

K
KAMEZAWA Hiroyuki 已提交
394
/*
395 396 397 398 399
 * 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
400
 */
401
static void mem_cgroup_lru_del_before_commit_swapcache(struct page *page)
K
KAMEZAWA Hiroyuki 已提交
402
{
403 404 405 406 407 408 409 410 411 412 413 414
	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
415 416
}

417 418 419 420 421 422 423 424 425 426 427 428 429 430
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 */
	if (PageLRU(page) && list_empty(&pc->lru))
		mem_cgroup_add_lru_list(page, page_lru(page));
	spin_unlock_irqrestore(&zone->lru_lock, flags);
}


K
KAMEZAWA Hiroyuki 已提交
431 432 433
void mem_cgroup_move_lists(struct page *page,
			   enum lru_list from, enum lru_list to)
{
434
	if (mem_cgroup_disabled())
K
KAMEZAWA Hiroyuki 已提交
435 436 437
		return;
	mem_cgroup_del_lru_list(page, from);
	mem_cgroup_add_lru_list(page, to);
438 439
}

440 441 442 443 444
int task_in_mem_cgroup(struct task_struct *task, const struct mem_cgroup *mem)
{
	int ret;

	task_lock(task);
445
	ret = task->mm && mm_match_cgroup(task->mm, mem);
446 447 448 449
	task_unlock(task);
	return ret;
}

450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465
/*
 * Calculate mapped_ratio under memory controller. This will be used in
 * vmscan.c for deteremining we have to reclaim mapped pages.
 */
int mem_cgroup_calc_mapped_ratio(struct mem_cgroup *mem)
{
	long total, rss;

	/*
	 * usage is recorded in bytes. But, here, we assume the number of
	 * physical pages can be represented by "long" on any arch.
	 */
	total = (long) (mem->res.usage >> PAGE_SHIFT) + 1L;
	rss = (long)mem_cgroup_read_stat(&mem->stat, MEM_CGROUP_STAT_RSS);
	return (int)((rss * 100L) / total);
}
466

467 468 469 470 471
/*
 * 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
472 473 474 475 476 477 478
	int prev_priority;

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

	return prev_priority;
479 480 481 482
}

void mem_cgroup_note_reclaim_priority(struct mem_cgroup *mem, int priority)
{
K
KOSAKI Motohiro 已提交
483
	spin_lock(&mem->reclaim_param_lock);
484 485
	if (priority < mem->prev_priority)
		mem->prev_priority = priority;
K
KOSAKI Motohiro 已提交
486
	spin_unlock(&mem->reclaim_param_lock);
487 488 489 490
}

void mem_cgroup_record_reclaim_priority(struct mem_cgroup *mem, int priority)
{
K
KOSAKI Motohiro 已提交
491
	spin_lock(&mem->reclaim_param_lock);
492
	mem->prev_priority = priority;
K
KOSAKI Motohiro 已提交
493
	spin_unlock(&mem->reclaim_param_lock);
494 495
}

496
static int calc_inactive_ratio(struct mem_cgroup *memcg, unsigned long *present_pages)
497 498 499
{
	unsigned long active;
	unsigned long inactive;
500 501
	unsigned long gb;
	unsigned long inactive_ratio;
502 503 504 505

	inactive = mem_cgroup_get_all_zonestat(memcg, LRU_INACTIVE_ANON);
	active = mem_cgroup_get_all_zonestat(memcg, LRU_ACTIVE_ANON);

506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532
	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)
533 534 535 536 537
		return 1;

	return 0;
}

538 539 540 541 542 543 544 545 546 547 548
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
549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568
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);
569 570 571 572 573 574 575 576
	/*
	 * 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
577 578 579 580 581 582 583
	mz = page_cgroup_zoneinfo(pc);
	if (!mz)
		return NULL;

	return &mz->reclaim_stat;
}

584 585 586 587 588
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,
589
					int active, int file)
590 591 592 593 594 595
{
	unsigned long nr_taken = 0;
	struct page *page;
	unsigned long scan;
	LIST_HEAD(pc_list);
	struct list_head *src;
596
	struct page_cgroup *pc, *tmp;
597 598 599
	int nid = z->zone_pgdat->node_id;
	int zid = zone_idx(z);
	struct mem_cgroup_per_zone *mz;
600
	int lru = LRU_FILE * !!file + !!active;
601

602
	BUG_ON(!mem_cont);
603
	mz = mem_cgroup_zoneinfo(mem_cont, nid, zid);
604
	src = &mz->lists[lru];
605

606 607
	scan = 0;
	list_for_each_entry_safe_reverse(pc, tmp, src, lru) {
H
Hugh Dickins 已提交
608
		if (scan >= nr_to_scan)
609
			break;
K
KAMEZAWA Hiroyuki 已提交
610 611

		page = pc->page;
612 613
		if (unlikely(!PageCgroupUsed(pc)))
			continue;
H
Hugh Dickins 已提交
614
		if (unlikely(!PageLRU(page)))
615 616
			continue;

H
Hugh Dickins 已提交
617
		scan++;
618
		if (__isolate_lru_page(page, mode, file) == 0) {
619 620 621 622 623 624 625 626 627
			list_move(&page->lru, dst);
			nr_taken++;
		}
	}

	*scanned = scan;
	return nr_taken;
}

628 629 630 631 632
#define mem_cgroup_from_res_counter(counter, member)	\
	container_of(counter, struct mem_cgroup, member)

/*
 * This routine finds the DFS walk successor. This routine should be
633
 * called with hierarchy_mutex held
634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695
 */
static struct mem_cgroup *
mem_cgroup_get_next_node(struct mem_cgroup *curr, struct mem_cgroup *root_mem)
{
	struct cgroup *cgroup, *curr_cgroup, *root_cgroup;

	curr_cgroup = curr->css.cgroup;
	root_cgroup = root_mem->css.cgroup;

	if (!list_empty(&curr_cgroup->children)) {
		/*
		 * Walk down to children
		 */
		mem_cgroup_put(curr);
		cgroup = list_entry(curr_cgroup->children.next,
						struct cgroup, sibling);
		curr = mem_cgroup_from_cont(cgroup);
		mem_cgroup_get(curr);
		goto done;
	}

visit_parent:
	if (curr_cgroup == root_cgroup) {
		mem_cgroup_put(curr);
		curr = root_mem;
		mem_cgroup_get(curr);
		goto done;
	}

	/*
	 * Goto next sibling
	 */
	if (curr_cgroup->sibling.next != &curr_cgroup->parent->children) {
		mem_cgroup_put(curr);
		cgroup = list_entry(curr_cgroup->sibling.next, struct cgroup,
						sibling);
		curr = mem_cgroup_from_cont(cgroup);
		mem_cgroup_get(curr);
		goto done;
	}

	/*
	 * Go up to next parent and next parent's sibling if need be
	 */
	curr_cgroup = curr_cgroup->parent;
	goto visit_parent;

done:
	root_mem->last_scanned_child = curr;
	return curr;
}

/*
 * 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_get_first_node(struct mem_cgroup *root_mem)
{
	struct cgroup *cgroup;
	struct mem_cgroup *ret;
696 697 698
	bool obsolete;

	obsolete = mem_cgroup_is_obsolete(root_mem->last_scanned_child);
699 700 701 702

	/*
	 * Scan all children under the mem_cgroup mem
	 */
703
	mutex_lock(&mem_cgroup_subsys.hierarchy_mutex);
704 705 706 707 708 709 710
	if (list_empty(&root_mem->css.cgroup->children)) {
		ret = root_mem;
		goto done;
	}

	if (!root_mem->last_scanned_child || obsolete) {

711
		if (obsolete && root_mem->last_scanned_child)
712 713 714 715 716 717 718 719 720 721 722 723
			mem_cgroup_put(root_mem->last_scanned_child);

		cgroup = list_first_entry(&root_mem->css.cgroup->children,
				struct cgroup, sibling);
		ret = mem_cgroup_from_cont(cgroup);
		mem_cgroup_get(ret);
	} else
		ret = mem_cgroup_get_next_node(root_mem->last_scanned_child,
						root_mem);

done:
	root_mem->last_scanned_child = ret;
724
	mutex_unlock(&mem_cgroup_subsys.hierarchy_mutex);
725 726 727
	return ret;
}

728 729 730 731 732 733 734 735 736 737 738 739
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
740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755
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;
}

756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775
/*
 * Dance down 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.
 *
 * root_mem is the original ancestor that we've been reclaim from.
 */
static int mem_cgroup_hierarchical_reclaim(struct mem_cgroup *root_mem,
						gfp_t gfp_mask, bool noswap)
{
	struct mem_cgroup *next_mem;
	int ret = 0;

	/*
	 * Reclaim unconditionally and don't check for return value.
	 * We need to reclaim in the current group and down the tree.
	 * One might think about checking for children before reclaiming,
	 * but there might be left over accounting, even after children
	 * have left.
	 */
K
KOSAKI Motohiro 已提交
776 777
	ret = try_to_free_mem_cgroup_pages(root_mem, gfp_mask, noswap,
					   get_swappiness(root_mem));
778
	if (mem_cgroup_check_under_limit(root_mem))
779
		return 0;
780 781
	if (!root_mem->use_hierarchy)
		return ret;
782 783 784 785

	next_mem = mem_cgroup_get_first_node(root_mem);

	while (next_mem != root_mem) {
786
		if (mem_cgroup_is_obsolete(next_mem)) {
787 788 789 790
			mem_cgroup_put(next_mem);
			next_mem = mem_cgroup_get_first_node(root_mem);
			continue;
		}
K
KOSAKI Motohiro 已提交
791 792
		ret = try_to_free_mem_cgroup_pages(next_mem, gfp_mask, noswap,
						   get_swappiness(next_mem));
793
		if (mem_cgroup_check_under_limit(root_mem))
794
			return 0;
795
		mutex_lock(&mem_cgroup_subsys.hierarchy_mutex);
796
		next_mem = mem_cgroup_get_next_node(next_mem, root_mem);
797
		mutex_unlock(&mem_cgroup_subsys.hierarchy_mutex);
798 799 800 801
	}
	return ret;
}

802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817
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;
}
818 819 820
/*
 * Unlike exported interface, "oom" parameter is added. if oom==true,
 * oom-killer can be invoked.
821
 */
822
static int __mem_cgroup_try_charge(struct mm_struct *mm,
823 824
			gfp_t gfp_mask, struct mem_cgroup **memcg,
			bool oom)
825
{
826
	struct mem_cgroup *mem, *mem_over_limit;
827
	int nr_retries = MEM_CGROUP_RECLAIM_RETRIES;
828
	struct res_counter *fail_res;
829 830 831 832 833 834 835

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

836
	/*
837 838
	 * We always charge the cgroup the mm_struct belongs to.
	 * The mm_struct's mem_cgroup changes on task migration if the
839 840 841
	 * thread group leader migrates. It's possible that mm is not
	 * set, if so charge the init_mm (happens for pagecache usage).
	 */
842 843 844
	mem = *memcg;
	if (likely(!mem)) {
		mem = try_get_mem_cgroup_from_mm(mm);
845
		*memcg = mem;
846
	} else {
847
		css_get(&mem->css);
848
	}
849 850 851 852
	if (unlikely(!mem))
		return 0;

	VM_BUG_ON(mem_cgroup_is_obsolete(mem));
853

854 855 856
	while (1) {
		int ret;
		bool noswap = false;
857

858
		ret = res_counter_charge(&mem->res, PAGE_SIZE, &fail_res);
859 860 861
		if (likely(!ret)) {
			if (!do_swap_account)
				break;
862 863
			ret = res_counter_charge(&mem->memsw, PAGE_SIZE,
							&fail_res);
864 865 866 867 868
			if (likely(!ret))
				break;
			/* mem+swap counter fails */
			res_counter_uncharge(&mem->res, PAGE_SIZE);
			noswap = true;
869 870 871 872 873 874 875
			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);

876
		if (!(gfp_mask & __GFP_WAIT))
877
			goto nomem;
878

879 880
		ret = mem_cgroup_hierarchical_reclaim(mem_over_limit, gfp_mask,
							noswap);
881 882

		/*
883 884 885 886 887
		 * 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
888
		 *
889
		 */
890 891
		if (mem_cgroup_check_under_limit(mem_over_limit))
			continue;
892 893

		if (!nr_retries--) {
894
			if (oom) {
895
				mutex_lock(&memcg_tasklist);
896
				mem_cgroup_out_of_memory(mem_over_limit, gfp_mask);
897
				mutex_unlock(&memcg_tasklist);
898
				mem_over_limit->last_oom_jiffies = jiffies;
899
			}
900
			goto nomem;
901
		}
902
	}
903 904 905 906 907
	return 0;
nomem:
	css_put(&mem->css);
	return -ENOMEM;
}
908

909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925
static struct mem_cgroup *try_get_mem_cgroup_from_swapcache(struct page *page)
{
	struct mem_cgroup *mem;
	swp_entry_t ent;

	if (!PageSwapCache(page))
		return NULL;

	ent.val = page_private(page);
	mem = lookup_swap_cgroup(ent);
	if (!mem)
		return NULL;
	if (!css_tryget(&mem->css))
		return NULL;
	return mem;
}

926
/*
927
 * commit a charge got by __mem_cgroup_try_charge() and makes page_cgroup to be
928 929 930 931 932 933 934 935 936 937
 * 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;
938 939 940 941 942

	lock_page_cgroup(pc);
	if (unlikely(PageCgroupUsed(pc))) {
		unlock_page_cgroup(pc);
		res_counter_uncharge(&mem->res, PAGE_SIZE);
943 944
		if (do_swap_account)
			res_counter_uncharge(&mem->memsw, PAGE_SIZE);
945
		css_put(&mem->css);
946
		return;
947
	}
948
	pc->mem_cgroup = mem;
K
KAMEZAWA Hiroyuki 已提交
949
	smp_wmb();
950
	pc->flags = pcg_default_flags[ctype];
951

K
KAMEZAWA Hiroyuki 已提交
952
	mem_cgroup_charge_statistics(mem, pc, true);
953 954

	unlock_page_cgroup(pc);
955
}
956

957 958 959 960 961 962 963
/**
 * 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
964
 * - page is not on LRU (isolate_page() is useful.)
965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980
 *
 * 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;

	VM_BUG_ON(from == to);
K
KAMEZAWA Hiroyuki 已提交
981
	VM_BUG_ON(PageLRU(pc->page));
982 983 984 985 986 987 988 989 990 991 992 993 994 995 996

	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;

K
KAMEZAWA Hiroyuki 已提交
997 998 999 1000
	res_counter_uncharge(&from->res, PAGE_SIZE);
	mem_cgroup_charge_statistics(from, pc, false);
	if (do_swap_account)
		res_counter_uncharge(&from->memsw, PAGE_SIZE);
1001 1002 1003
	css_put(&from->css);

	css_get(&to->css);
K
KAMEZAWA Hiroyuki 已提交
1004 1005 1006
	pc->mem_cgroup = to;
	mem_cgroup_charge_statistics(to, pc, true);
	ret = 0;
1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019
out:
	unlock_page_cgroup(pc);
	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
1020
	struct page *page = pc->page;
1021 1022 1023 1024 1025 1026 1027 1028 1029
	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
1030

1031 1032
	parent = mem_cgroup_from_cont(pcg);

K
KAMEZAWA Hiroyuki 已提交
1033

1034
	ret = __mem_cgroup_try_charge(NULL, gfp_mask, &parent, false);
1035
	if (ret || !parent)
1036 1037
		return ret;

1038 1039 1040 1041
	if (!get_page_unless_zero(page)) {
		ret = -EBUSY;
		goto uncharge;
	}
K
KAMEZAWA Hiroyuki 已提交
1042 1043 1044 1045 1046

	ret = isolate_lru_page(page);

	if (ret)
		goto cancel;
1047 1048 1049

	ret = mem_cgroup_move_account(pc, child, parent);

K
KAMEZAWA Hiroyuki 已提交
1050 1051 1052
	putback_lru_page(page);
	if (!ret) {
		put_page(page);
1053 1054
		/* drop extra refcnt by try_charge() */
		css_put(&parent->css);
K
KAMEZAWA Hiroyuki 已提交
1055
		return 0;
1056
	}
1057

K
KAMEZAWA Hiroyuki 已提交
1058
cancel:
1059 1060 1061 1062 1063
	put_page(page);
uncharge:
	/* drop extra refcnt by try_charge() */
	css_put(&parent->css);
	/* uncharge if move fails */
K
KAMEZAWA Hiroyuki 已提交
1064 1065 1066
	res_counter_uncharge(&parent->res, PAGE_SIZE);
	if (do_swap_account)
		res_counter_uncharge(&parent->memsw, PAGE_SIZE);
1067 1068 1069
	return ret;
}

1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090
/*
 * 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;
1091
	ret = __mem_cgroup_try_charge(mm, gfp_mask, &mem, true);
1092
	if (ret || !mem)
1093 1094 1095
		return ret;

	__mem_cgroup_commit_charge(mem, pc, ctype);
1096 1097 1098
	return 0;
}

1099 1100
int mem_cgroup_newpage_charge(struct page *page,
			      struct mm_struct *mm, gfp_t gfp_mask)
1101
{
1102
	if (mem_cgroup_disabled())
1103
		return 0;
1104 1105
	if (PageCompound(page))
		return 0;
1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116
	/*
	 * 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;
1117
	return mem_cgroup_charge_common(page, mm, gfp_mask,
1118
				MEM_CGROUP_CHARGE_TYPE_MAPPED, NULL);
1119 1120
}

1121 1122
int mem_cgroup_cache_charge(struct page *page, struct mm_struct *mm,
				gfp_t gfp_mask)
1123
{
1124 1125 1126
	struct mem_cgroup *mem = NULL;
	int ret;

1127
	if (mem_cgroup_disabled())
1128
		return 0;
1129 1130
	if (PageCompound(page))
		return 0;
1131 1132 1133 1134 1135 1136 1137 1138
	/*
	 * 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.)
1139 1140
	 * And when the page is SwapCache, it should take swap information
	 * into account. This is under lock_page() now.
1141 1142 1143 1144
	 */
	if (!(gfp_mask & __GFP_WAIT)) {
		struct page_cgroup *pc;

1145 1146 1147 1148 1149 1150 1151

		pc = lookup_page_cgroup(page);
		if (!pc)
			return 0;
		lock_page_cgroup(pc);
		if (PageCgroupUsed(pc)) {
			unlock_page_cgroup(pc);
1152 1153
			return 0;
		}
1154
		unlock_page_cgroup(pc);
1155 1156
	}

1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167
	if (do_swap_account && PageSwapCache(page)) {
		mem = try_get_mem_cgroup_from_swapcache(page);
		if (mem)
			mm = NULL;
		  else
			mem = NULL;
		/* SwapCache may be still linked to LRU now. */
		mem_cgroup_lru_del_before_commit_swapcache(page);
	}

	if (unlikely(!mm && !mem))
1168
		mm = &init_mm;
1169

1170 1171
	if (page_is_file_cache(page))
		return mem_cgroup_charge_common(page, mm, gfp_mask,
1172
				MEM_CGROUP_CHARGE_TYPE_CACHE, NULL);
1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190

	ret = mem_cgroup_charge_common(page, mm, gfp_mask,
				MEM_CGROUP_CHARGE_TYPE_SHMEM, mem);
	if (mem)
		css_put(&mem->css);
	if (PageSwapCache(page))
		mem_cgroup_lru_add_after_commit_swapcache(page);

	if (do_swap_account && !ret && PageSwapCache(page)) {
		swp_entry_t ent = {.val = page_private(page)};
		/* avoid double counting */
		mem = swap_cgroup_record(ent, NULL);
		if (mem) {
			res_counter_uncharge(&mem->memsw, PAGE_SIZE);
			mem_cgroup_put(mem);
		}
	}
	return ret;
1191 1192
}

1193 1194 1195 1196 1197 1198
/*
 * While swap-in, try_charge -> commit or cancel, the page is locked.
 * And when try_charge() successfully returns, one refcnt to memcg without
 * struct page_cgroup is aquired. This refcnt will be cumsumed by
 * "commit()" or removed by "cancel()"
 */
1199 1200 1201 1202 1203
int mem_cgroup_try_charge_swapin(struct mm_struct *mm,
				 struct page *page,
				 gfp_t mask, struct mem_cgroup **ptr)
{
	struct mem_cgroup *mem;
1204
	int ret;
1205

1206
	if (mem_cgroup_disabled())
1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217
		return 0;

	if (!do_swap_account)
		goto charge_cur_mm;
	/*
	 * A racing thread's fault, or swapoff, may have already updated
	 * the pte, and even removed page from swap cache: return success
	 * to go on to do_swap_page()'s pte_same() test, which should fail.
	 */
	if (!PageSwapCache(page))
		return 0;
1218
	mem = try_get_mem_cgroup_from_swapcache(page);
1219 1220
	if (!mem)
		goto charge_cur_mm;
1221
	*ptr = mem;
1222 1223 1224 1225
	ret = __mem_cgroup_try_charge(NULL, mask, ptr, true);
	/* drop extra refcnt from tryget */
	css_put(&mem->css);
	return ret;
1226 1227 1228 1229 1230 1231
charge_cur_mm:
	if (unlikely(!mm))
		mm = &init_mm;
	return __mem_cgroup_try_charge(mm, mask, ptr, true);
}

1232 1233 1234 1235
void mem_cgroup_commit_charge_swapin(struct page *page, struct mem_cgroup *ptr)
{
	struct page_cgroup *pc;

1236
	if (mem_cgroup_disabled())
1237 1238 1239 1240
		return;
	if (!ptr)
		return;
	pc = lookup_page_cgroup(page);
1241
	mem_cgroup_lru_del_before_commit_swapcache(page);
1242
	__mem_cgroup_commit_charge(ptr, pc, MEM_CGROUP_CHARGE_TYPE_MAPPED);
1243
	mem_cgroup_lru_add_after_commit_swapcache(page);
1244 1245 1246
	/*
	 * Now swap is on-memory. This means this page may be
	 * counted both as mem and swap....double count.
1247 1248 1249
	 * 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.
1250
	 */
1251
	if (do_swap_account && PageSwapCache(page)) {
1252 1253 1254 1255 1256 1257 1258 1259 1260
		swp_entry_t ent = {.val = page_private(page)};
		struct mem_cgroup *memcg;
		memcg = swap_cgroup_record(ent, NULL);
		if (memcg) {
			res_counter_uncharge(&memcg->memsw, PAGE_SIZE);
			mem_cgroup_put(memcg);
		}

	}
K
KAMEZAWA Hiroyuki 已提交
1261
	/* add this page(page_cgroup) to the LRU we want. */
1262

1263 1264 1265 1266
}

void mem_cgroup_cancel_charge_swapin(struct mem_cgroup *mem)
{
1267
	if (mem_cgroup_disabled())
1268 1269 1270 1271
		return;
	if (!mem)
		return;
	res_counter_uncharge(&mem->res, PAGE_SIZE);
1272 1273
	if (do_swap_account)
		res_counter_uncharge(&mem->memsw, PAGE_SIZE);
1274 1275 1276 1277
	css_put(&mem->css);
}


1278
/*
1279
 * uncharge if !page_mapped(page)
1280
 */
1281
static struct mem_cgroup *
1282
__mem_cgroup_uncharge_common(struct page *page, enum charge_type ctype)
1283
{
H
Hugh Dickins 已提交
1284
	struct page_cgroup *pc;
1285
	struct mem_cgroup *mem = NULL;
1286
	struct mem_cgroup_per_zone *mz;
1287

1288
	if (mem_cgroup_disabled())
1289
		return NULL;
1290

K
KAMEZAWA Hiroyuki 已提交
1291
	if (PageSwapCache(page))
1292
		return NULL;
K
KAMEZAWA Hiroyuki 已提交
1293

1294
	/*
1295
	 * Check if our page_cgroup is valid
1296
	 */
1297 1298
	pc = lookup_page_cgroup(page);
	if (unlikely(!pc || !PageCgroupUsed(pc)))
1299
		return NULL;
1300

1301
	lock_page_cgroup(pc);
K
KAMEZAWA Hiroyuki 已提交
1302

1303 1304
	mem = pc->mem_cgroup;

K
KAMEZAWA Hiroyuki 已提交
1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321
	if (!PageCgroupUsed(pc))
		goto unlock_out;

	switch (ctype) {
	case MEM_CGROUP_CHARGE_TYPE_MAPPED:
		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;
1322
	}
K
KAMEZAWA Hiroyuki 已提交
1323

1324 1325 1326 1327
	res_counter_uncharge(&mem->res, PAGE_SIZE);
	if (do_swap_account && (ctype != MEM_CGROUP_CHARGE_TYPE_SWAPOUT))
		res_counter_uncharge(&mem->memsw, PAGE_SIZE);

K
KAMEZAWA Hiroyuki 已提交
1328
	mem_cgroup_charge_statistics(mem, pc, false);
1329
	ClearPageCgroupUsed(pc);
1330 1331 1332 1333 1334 1335
	/*
	 * 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.
	 */
1336

1337
	mz = page_cgroup_zoneinfo(pc);
1338
	unlock_page_cgroup(pc);
H
Hugh Dickins 已提交
1339

K
KAMEZAWA Hiroyuki 已提交
1340 1341 1342
	/* at swapout, this memcg will be accessed to record to swap */
	if (ctype != MEM_CGROUP_CHARGE_TYPE_SWAPOUT)
		css_put(&mem->css);
1343

1344
	return mem;
K
KAMEZAWA Hiroyuki 已提交
1345 1346 1347

unlock_out:
	unlock_page_cgroup(pc);
1348
	return NULL;
1349 1350
}

1351 1352
void mem_cgroup_uncharge_page(struct page *page)
{
1353 1354 1355 1356 1357
	/* early check. */
	if (page_mapped(page))
		return;
	if (page->mapping && !PageAnon(page))
		return;
1358 1359 1360 1361 1362 1363
	__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));
1364
	VM_BUG_ON(page->mapping);
1365 1366 1367
	__mem_cgroup_uncharge_common(page, MEM_CGROUP_CHARGE_TYPE_CACHE);
}

1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382
/*
 * called from __delete_from_swap_cache() and drop "page" account.
 * memcg information is recorded to swap_cgroup of "ent"
 */
void mem_cgroup_uncharge_swapcache(struct page *page, swp_entry_t ent)
{
	struct mem_cgroup *memcg;

	memcg = __mem_cgroup_uncharge_common(page,
					MEM_CGROUP_CHARGE_TYPE_SWAPOUT);
	/* record memcg information */
	if (do_swap_account && memcg) {
		swap_cgroup_record(ent, memcg);
		mem_cgroup_get(memcg);
	}
K
KAMEZAWA Hiroyuki 已提交
1383 1384
	if (memcg)
		css_put(&memcg->css);
1385 1386 1387 1388 1389 1390 1391 1392
}

#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
1393
{
1394 1395 1396 1397 1398 1399 1400 1401 1402 1403
	struct mem_cgroup *memcg;

	if (!do_swap_account)
		return;

	memcg = swap_cgroup_record(ent, NULL);
	if (memcg) {
		res_counter_uncharge(&memcg->memsw, PAGE_SIZE);
		mem_cgroup_put(memcg);
	}
K
KAMEZAWA Hiroyuki 已提交
1404
}
1405
#endif
K
KAMEZAWA Hiroyuki 已提交
1406

1407
/*
1408 1409
 * Before starting migration, account PAGE_SIZE to mem_cgroup that the old
 * page belongs to.
1410
 */
1411
int mem_cgroup_prepare_migration(struct page *page, struct mem_cgroup **ptr)
1412 1413
{
	struct page_cgroup *pc;
1414 1415
	struct mem_cgroup *mem = NULL;
	int ret = 0;
1416

1417
	if (mem_cgroup_disabled())
1418 1419
		return 0;

1420 1421 1422
	pc = lookup_page_cgroup(page);
	lock_page_cgroup(pc);
	if (PageCgroupUsed(pc)) {
1423 1424 1425
		mem = pc->mem_cgroup;
		css_get(&mem->css);
	}
1426
	unlock_page_cgroup(pc);
1427

1428
	if (mem) {
1429
		ret = __mem_cgroup_try_charge(NULL, GFP_KERNEL, &mem, false);
1430 1431
		css_put(&mem->css);
	}
1432
	*ptr = mem;
1433
	return ret;
1434
}
1435

1436
/* remove redundant charge if migration failed*/
1437 1438
void mem_cgroup_end_migration(struct mem_cgroup *mem,
		struct page *oldpage, struct page *newpage)
1439
{
1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463
	struct page *target, *unused;
	struct page_cgroup *pc;
	enum charge_type ctype;

	if (!mem)
		return;

	/* 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
1464
	if (unused)
1465 1466 1467
		__mem_cgroup_uncharge_common(unused, ctype);

	pc = lookup_page_cgroup(target);
1468
	/*
1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482
	 * __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.
1483
	 */
1484 1485
	if (ctype == MEM_CGROUP_CHARGE_TYPE_MAPPED)
		mem_cgroup_uncharge_page(target);
1486
}
1487

1488 1489 1490 1491 1492
/*
 * A call to try to shrink memory usage under specified resource controller.
 * This is typically used for page reclaiming for shmem for reducing side
 * effect of page allocation from shmem, which is used by some mem_cgroup.
 */
1493 1494 1495
int mem_cgroup_shrink_usage(struct page *page,
			    struct mm_struct *mm,
			    gfp_t gfp_mask)
1496
{
1497
	struct mem_cgroup *mem = NULL;
1498 1499 1500
	int progress = 0;
	int retry = MEM_CGROUP_RECLAIM_RETRIES;

1501
	if (mem_cgroup_disabled())
1502
		return 0;
1503 1504 1505 1506
	if (page)
		mem = try_get_mem_cgroup_from_swapcache(page);
	if (!mem && mm)
		mem = try_get_mem_cgroup_from_mm(mm);
1507
	if (unlikely(!mem))
1508
		return 0;
1509 1510

	do {
1511
		progress = mem_cgroup_hierarchical_reclaim(mem, gfp_mask, true);
1512
		progress += mem_cgroup_check_under_limit(mem);
1513 1514 1515 1516 1517 1518 1519 1520
	} while (!progress && --retry);

	css_put(&mem->css);
	if (!retry)
		return -ENOMEM;
	return 0;
}

1521 1522
static DEFINE_MUTEX(set_limit_mutex);

1523
static int mem_cgroup_resize_limit(struct mem_cgroup *memcg,
1524
				unsigned long long val)
1525 1526 1527 1528
{

	int retry_count = MEM_CGROUP_RECLAIM_RETRIES;
	int progress;
1529
	u64 memswlimit;
1530 1531
	int ret = 0;

1532
	while (retry_count) {
1533 1534 1535 1536
		if (signal_pending(current)) {
			ret = -EINTR;
			break;
		}
1537 1538 1539 1540 1541 1542 1543 1544 1545 1546
		/*
		 * 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);
1547 1548
			break;
		}
1549 1550 1551 1552 1553 1554
		ret = res_counter_set_limit(&memcg->res, val);
		mutex_unlock(&set_limit_mutex);

		if (!ret)
			break;

1555 1556
		progress = mem_cgroup_hierarchical_reclaim(memcg, GFP_KERNEL,
							   false);
1557 1558
  		if (!progress)			retry_count--;
	}
1559

1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596
	return ret;
}

int mem_cgroup_resize_memsw_limit(struct mem_cgroup *memcg,
				unsigned long long val)
{
	int retry_count = MEM_CGROUP_RECLAIM_RETRIES;
	u64 memlimit, oldusage, curusage;
	int ret;

	if (!do_swap_account)
		return -EINVAL;

	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);
		mutex_unlock(&set_limit_mutex);

		if (!ret)
			break;

		oldusage = res_counter_read_u64(&memcg->memsw, RES_USAGE);
1597
		mem_cgroup_hierarchical_reclaim(memcg, GFP_KERNEL, true);
1598 1599
		curusage = res_counter_read_u64(&memcg->memsw, RES_USAGE);
		if (curusage >= oldusage)
1600 1601 1602 1603 1604
			retry_count--;
	}
	return ret;
}

1605 1606 1607 1608
/*
 * This routine traverse page_cgroup in given list and drop them all.
 * *And* this routine doesn't reclaim page itself, just removes page_cgroup.
 */
1609
static int mem_cgroup_force_empty_list(struct mem_cgroup *mem,
K
KAMEZAWA Hiroyuki 已提交
1610
				int node, int zid, enum lru_list lru)
1611
{
K
KAMEZAWA Hiroyuki 已提交
1612 1613
	struct zone *zone;
	struct mem_cgroup_per_zone *mz;
1614
	struct page_cgroup *pc, *busy;
K
KAMEZAWA Hiroyuki 已提交
1615
	unsigned long flags, loop;
1616
	struct list_head *list;
1617
	int ret = 0;
1618

K
KAMEZAWA Hiroyuki 已提交
1619 1620
	zone = &NODE_DATA(node)->node_zones[zid];
	mz = mem_cgroup_zoneinfo(mem, node, zid);
1621
	list = &mz->lists[lru];
1622

1623 1624 1625 1626 1627 1628
	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
1629
		spin_lock_irqsave(&zone->lru_lock, flags);
1630
		if (list_empty(list)) {
K
KAMEZAWA Hiroyuki 已提交
1631
			spin_unlock_irqrestore(&zone->lru_lock, flags);
1632
			break;
1633 1634 1635 1636 1637
		}
		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
1638
			spin_unlock_irqrestore(&zone->lru_lock, flags);
1639 1640
			continue;
		}
K
KAMEZAWA Hiroyuki 已提交
1641
		spin_unlock_irqrestore(&zone->lru_lock, flags);
1642

K
KAMEZAWA Hiroyuki 已提交
1643
		ret = mem_cgroup_move_parent(pc, mem, GFP_KERNEL);
1644
		if (ret == -ENOMEM)
1645
			break;
1646 1647 1648 1649 1650 1651 1652

		if (ret == -EBUSY || ret == -EINVAL) {
			/* found lock contention or "pc" is obsolete. */
			busy = pc;
			cond_resched();
		} else
			busy = NULL;
1653
	}
K
KAMEZAWA Hiroyuki 已提交
1654

1655 1656 1657
	if (!ret && !list_empty(list))
		return -EBUSY;
	return ret;
1658 1659 1660 1661 1662 1663
}

/*
 * make mem_cgroup's charge to be 0 if there is no task.
 * This enables deleting this mem_cgroup.
 */
1664
static int mem_cgroup_force_empty(struct mem_cgroup *mem, bool free_all)
1665
{
1666 1667 1668
	int ret;
	int node, zid, shrink;
	int nr_retries = MEM_CGROUP_RECLAIM_RETRIES;
1669
	struct cgroup *cgrp = mem->css.cgroup;
1670

1671
	css_get(&mem->css);
1672 1673

	shrink = 0;
1674 1675 1676
	/* should free all ? */
	if (free_all)
		goto try_to_free;
1677
move_account:
1678
	while (mem->res.usage > 0) {
1679
		ret = -EBUSY;
1680 1681 1682 1683
		if (cgroup_task_count(cgrp) || !list_empty(&cgrp->children))
			goto out;
		ret = -EINTR;
		if (signal_pending(current))
1684
			goto out;
1685 1686
		/* This is for making all *used* pages to be on LRU. */
		lru_add_drain_all();
1687 1688 1689
		ret = 0;
		for_each_node_state(node, N_POSSIBLE) {
			for (zid = 0; !ret && zid < MAX_NR_ZONES; zid++) {
1690
				enum lru_list l;
1691 1692
				for_each_lru(l) {
					ret = mem_cgroup_force_empty_list(mem,
K
KAMEZAWA Hiroyuki 已提交
1693
							node, zid, l);
1694 1695 1696
					if (ret)
						break;
				}
1697
			}
1698 1699 1700 1701 1702 1703
			if (ret)
				break;
		}
		/* it seems parent cgroup doesn't have enough mem */
		if (ret == -ENOMEM)
			goto try_to_free;
1704
		cond_resched();
1705 1706 1707 1708 1709
	}
	ret = 0;
out:
	css_put(&mem->css);
	return ret;
1710 1711

try_to_free:
1712 1713
	/* returns EBUSY if there is a task or if we come here twice. */
	if (cgroup_task_count(cgrp) || !list_empty(&cgrp->children) || shrink) {
1714 1715 1716
		ret = -EBUSY;
		goto out;
	}
1717 1718
	/* we call try-to-free pages for make this cgroup empty */
	lru_add_drain_all();
1719 1720 1721 1722
	/* try to free all pages in this cgroup */
	shrink = 1;
	while (nr_retries && mem->res.usage > 0) {
		int progress;
1723 1724 1725 1726 1727

		if (signal_pending(current)) {
			ret = -EINTR;
			goto out;
		}
K
KOSAKI Motohiro 已提交
1728 1729
		progress = try_to_free_mem_cgroup_pages(mem, GFP_KERNEL,
						false, get_swappiness(mem));
1730
		if (!progress) {
1731
			nr_retries--;
1732 1733 1734
			/* maybe some writeback is necessary */
			congestion_wait(WRITE, HZ/10);
		}
1735 1736

	}
K
KAMEZAWA Hiroyuki 已提交
1737
	lru_add_drain();
1738 1739 1740 1741 1742
	/* try move_account...there may be some *locked* pages. */
	if (mem->res.usage)
		goto move_account;
	ret = 0;
	goto out;
1743 1744
}

1745 1746 1747 1748 1749 1750
int mem_cgroup_force_empty_write(struct cgroup *cont, unsigned int event)
{
	return mem_cgroup_force_empty(mem_cgroup_from_cont(cont), true);
}


1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788
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();
	/*
	 * If parent's use_hiearchy is set, we can't make any modifications
	 * 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;
}

1789
static u64 mem_cgroup_read(struct cgroup *cont, struct cftype *cft)
B
Balbir Singh 已提交
1790
{
1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809
	struct mem_cgroup *mem = mem_cgroup_from_cont(cont);
	u64 val = 0;
	int type, name;

	type = MEMFILE_TYPE(cft->private);
	name = MEMFILE_ATTR(cft->private);
	switch (type) {
	case _MEM:
		val = res_counter_read_u64(&mem->res, name);
		break;
	case _MEMSWAP:
		if (do_swap_account)
			val = res_counter_read_u64(&mem->memsw, name);
		break;
	default:
		BUG();
		break;
	}
	return val;
B
Balbir Singh 已提交
1810
}
1811 1812 1813 1814
/*
 * The user of this function is...
 * RES_LIMIT.
 */
1815 1816
static int mem_cgroup_write(struct cgroup *cont, struct cftype *cft,
			    const char *buffer)
B
Balbir Singh 已提交
1817
{
1818
	struct mem_cgroup *memcg = mem_cgroup_from_cont(cont);
1819
	int type, name;
1820 1821 1822
	unsigned long long val;
	int ret;

1823 1824 1825
	type = MEMFILE_TYPE(cft->private);
	name = MEMFILE_ATTR(cft->private);
	switch (name) {
1826 1827 1828
	case RES_LIMIT:
		/* This function does all necessary parse...reuse it */
		ret = res_counter_memparse_write_strategy(buffer, &val);
1829 1830 1831
		if (ret)
			break;
		if (type == _MEM)
1832
			ret = mem_cgroup_resize_limit(memcg, val);
1833 1834
		else
			ret = mem_cgroup_resize_memsw_limit(memcg, val);
1835 1836 1837 1838 1839 1840
		break;
	default:
		ret = -EINVAL; /* should be BUG() ? */
		break;
	}
	return ret;
B
Balbir Singh 已提交
1841 1842
}

1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870
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;
}

1871
static int mem_cgroup_reset(struct cgroup *cont, unsigned int event)
1872 1873
{
	struct mem_cgroup *mem;
1874
	int type, name;
1875 1876

	mem = mem_cgroup_from_cont(cont);
1877 1878 1879
	type = MEMFILE_TYPE(event);
	name = MEMFILE_ATTR(event);
	switch (name) {
1880
	case RES_MAX_USAGE:
1881 1882 1883 1884
		if (type == _MEM)
			res_counter_reset_max(&mem->res);
		else
			res_counter_reset_max(&mem->memsw);
1885 1886
		break;
	case RES_FAILCNT:
1887 1888 1889 1890
		if (type == _MEM)
			res_counter_reset_failcnt(&mem->res);
		else
			res_counter_reset_failcnt(&mem->memsw);
1891 1892
		break;
	}
1893
	return 0;
1894 1895
}

1896 1897 1898 1899 1900 1901
static const struct mem_cgroup_stat_desc {
	const char *msg;
	u64 unit;
} mem_cgroup_stat_desc[] = {
	[MEM_CGROUP_STAT_CACHE] = { "cache", PAGE_SIZE, },
	[MEM_CGROUP_STAT_RSS] = { "rss", PAGE_SIZE, },
1902 1903
	[MEM_CGROUP_STAT_PGPGIN_COUNT] = {"pgpgin", 1, },
	[MEM_CGROUP_STAT_PGPGOUT_COUNT] = {"pgpgout", 1, },
1904 1905
};

1906 1907
static int mem_control_stat_show(struct cgroup *cont, struct cftype *cft,
				 struct cgroup_map_cb *cb)
1908 1909 1910 1911 1912 1913 1914 1915 1916 1917
{
	struct mem_cgroup *mem_cont = mem_cgroup_from_cont(cont);
	struct mem_cgroup_stat *stat = &mem_cont->stat;
	int i;

	for (i = 0; i < ARRAY_SIZE(stat->cpustat[0].count); i++) {
		s64 val;

		val = mem_cgroup_read_stat(stat, i);
		val *= mem_cgroup_stat_desc[i].unit;
1918
		cb->fill(cb, mem_cgroup_stat_desc[i].msg, val);
1919
	}
1920 1921
	/* showing # of active pages */
	{
1922 1923
		unsigned long active_anon, inactive_anon;
		unsigned long active_file, inactive_file;
L
Lee Schermerhorn 已提交
1924
		unsigned long unevictable;
1925 1926 1927 1928 1929 1930 1931 1932 1933

		inactive_anon = mem_cgroup_get_all_zonestat(mem_cont,
						LRU_INACTIVE_ANON);
		active_anon = mem_cgroup_get_all_zonestat(mem_cont,
						LRU_ACTIVE_ANON);
		inactive_file = mem_cgroup_get_all_zonestat(mem_cont,
						LRU_INACTIVE_FILE);
		active_file = mem_cgroup_get_all_zonestat(mem_cont,
						LRU_ACTIVE_FILE);
L
Lee Schermerhorn 已提交
1934 1935 1936
		unevictable = mem_cgroup_get_all_zonestat(mem_cont,
							LRU_UNEVICTABLE);

1937 1938 1939 1940
		cb->fill(cb, "active_anon", (active_anon) * PAGE_SIZE);
		cb->fill(cb, "inactive_anon", (inactive_anon) * PAGE_SIZE);
		cb->fill(cb, "active_file", (active_file) * PAGE_SIZE);
		cb->fill(cb, "inactive_file", (inactive_file) * PAGE_SIZE);
L
Lee Schermerhorn 已提交
1941 1942
		cb->fill(cb, "unevictable", unevictable * PAGE_SIZE);

1943
	}
1944 1945 1946 1947 1948 1949 1950
	{
		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
1951 1952

#ifdef CONFIG_DEBUG_VM
1953
	cb->fill(cb, "inactive_ratio", calc_inactive_ratio(mem_cont, NULL));
K
KOSAKI Motohiro 已提交
1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980

	{
		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

1981 1982 1983
	return 0;
}

K
KOSAKI Motohiro 已提交
1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
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;
	if (val > 100)
		return -EINVAL;

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

	parent = mem_cgroup_from_cont(cgrp->parent);
	/* If under hierarchy, only empty-root can set this value */
	if ((parent->use_hierarchy) ||
	    (memcg->use_hierarchy && !list_empty(&cgrp->children)))
		return -EINVAL;

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

	return 0;
}

2015

B
Balbir Singh 已提交
2016 2017
static struct cftype mem_cgroup_files[] = {
	{
2018
		.name = "usage_in_bytes",
2019
		.private = MEMFILE_PRIVATE(_MEM, RES_USAGE),
2020
		.read_u64 = mem_cgroup_read,
B
Balbir Singh 已提交
2021
	},
2022 2023
	{
		.name = "max_usage_in_bytes",
2024
		.private = MEMFILE_PRIVATE(_MEM, RES_MAX_USAGE),
2025
		.trigger = mem_cgroup_reset,
2026 2027
		.read_u64 = mem_cgroup_read,
	},
B
Balbir Singh 已提交
2028
	{
2029
		.name = "limit_in_bytes",
2030
		.private = MEMFILE_PRIVATE(_MEM, RES_LIMIT),
2031
		.write_string = mem_cgroup_write,
2032
		.read_u64 = mem_cgroup_read,
B
Balbir Singh 已提交
2033 2034 2035
	},
	{
		.name = "failcnt",
2036
		.private = MEMFILE_PRIVATE(_MEM, RES_FAILCNT),
2037
		.trigger = mem_cgroup_reset,
2038
		.read_u64 = mem_cgroup_read,
B
Balbir Singh 已提交
2039
	},
2040 2041
	{
		.name = "stat",
2042
		.read_map = mem_control_stat_show,
2043
	},
2044 2045 2046 2047
	{
		.name = "force_empty",
		.trigger = mem_cgroup_force_empty_write,
	},
2048 2049 2050 2051 2052
	{
		.name = "use_hierarchy",
		.write_u64 = mem_cgroup_hierarchy_write,
		.read_u64 = mem_cgroup_hierarchy_read,
	},
K
KOSAKI Motohiro 已提交
2053 2054 2055 2056 2057
	{
		.name = "swappiness",
		.read_u64 = mem_cgroup_swappiness_read,
		.write_u64 = mem_cgroup_swappiness_write,
	},
B
Balbir Singh 已提交
2058 2059
};

2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100
#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

2101 2102 2103
static int alloc_mem_cgroup_per_zone_info(struct mem_cgroup *mem, int node)
{
	struct mem_cgroup_per_node *pn;
2104
	struct mem_cgroup_per_zone *mz;
2105
	enum lru_list l;
2106
	int zone, tmp = node;
2107 2108 2109 2110 2111 2112 2113 2114
	/*
	 * 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.
	 */
2115 2116 2117
	if (!node_state(node, N_NORMAL_MEMORY))
		tmp = -1;
	pn = kmalloc_node(sizeof(*pn), GFP_KERNEL, tmp);
2118 2119
	if (!pn)
		return 1;
2120

2121 2122
	mem->info.nodeinfo[node] = pn;
	memset(pn, 0, sizeof(*pn));
2123 2124 2125

	for (zone = 0; zone < MAX_NR_ZONES; zone++) {
		mz = &pn->zoneinfo[zone];
2126 2127
		for_each_lru(l)
			INIT_LIST_HEAD(&mz->lists[l]);
2128
	}
2129 2130 2131
	return 0;
}

2132 2133 2134 2135 2136
static void free_mem_cgroup_per_zone_info(struct mem_cgroup *mem, int node)
{
	kfree(mem->info.nodeinfo[node]);
}

2137 2138 2139 2140 2141 2142
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;
}

2143 2144 2145
static struct mem_cgroup *mem_cgroup_alloc(void)
{
	struct mem_cgroup *mem;
2146
	int size = mem_cgroup_size();
2147

2148 2149
	if (size < PAGE_SIZE)
		mem = kmalloc(size, GFP_KERNEL);
2150
	else
2151
		mem = vmalloc(size);
2152 2153

	if (mem)
2154
		memset(mem, 0, size);
2155 2156 2157
	return mem;
}

2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168
/*
 * 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.
 */

2169
static void __mem_cgroup_free(struct mem_cgroup *mem)
2170
{
K
KAMEZAWA Hiroyuki 已提交
2171 2172 2173 2174 2175
	int node;

	for_each_node_state(node, N_POSSIBLE)
		free_mem_cgroup_per_zone_info(mem, node);

2176
	if (mem_cgroup_size() < PAGE_SIZE)
2177 2178 2179 2180 2181
		kfree(mem);
	else
		vfree(mem);
}

2182 2183 2184 2185 2186 2187 2188
static void mem_cgroup_get(struct mem_cgroup *mem)
{
	atomic_inc(&mem->refcnt);
}

static void mem_cgroup_put(struct mem_cgroup *mem)
{
2189 2190
	if (atomic_dec_and_test(&mem->refcnt))
		__mem_cgroup_free(mem);
2191 2192
}

2193

2194 2195 2196
#ifdef CONFIG_CGROUP_MEM_RES_CTLR_SWAP
static void __init enable_swap_cgroup(void)
{
2197
	if (!mem_cgroup_disabled() && really_do_swap_account)
2198 2199 2200 2201 2202 2203 2204 2205
		do_swap_account = 1;
}
#else
static void __init enable_swap_cgroup(void)
{
}
#endif

B
Balbir Singh 已提交
2206 2207 2208
static struct cgroup_subsys_state *
mem_cgroup_create(struct cgroup_subsys *ss, struct cgroup *cont)
{
2209
	struct mem_cgroup *mem, *parent;
2210
	int node;
B
Balbir Singh 已提交
2211

2212 2213 2214
	mem = mem_cgroup_alloc();
	if (!mem)
		return ERR_PTR(-ENOMEM);
2215

2216 2217 2218
	for_each_node_state(node, N_POSSIBLE)
		if (alloc_mem_cgroup_per_zone_info(mem, node))
			goto free_out;
2219
	/* root ? */
2220
	if (cont->parent == NULL) {
2221
		enable_swap_cgroup();
2222
		parent = NULL;
2223
	} else {
2224
		parent = mem_cgroup_from_cont(cont->parent);
2225 2226
		mem->use_hierarchy = parent->use_hierarchy;
	}
2227

2228 2229 2230 2231 2232 2233 2234
	if (parent && parent->use_hierarchy) {
		res_counter_init(&mem->res, &parent->res);
		res_counter_init(&mem->memsw, &parent->memsw);
	} else {
		res_counter_init(&mem->res, NULL);
		res_counter_init(&mem->memsw, NULL);
	}
2235
	mem->last_scanned_child = NULL;
K
KOSAKI Motohiro 已提交
2236
	spin_lock_init(&mem->reclaim_param_lock);
2237

K
KOSAKI Motohiro 已提交
2238 2239
	if (parent)
		mem->swappiness = get_swappiness(parent);
2240
	atomic_set(&mem->refcnt, 1);
B
Balbir Singh 已提交
2241
	return &mem->css;
2242
free_out:
2243
	__mem_cgroup_free(mem);
2244
	return ERR_PTR(-ENOMEM);
B
Balbir Singh 已提交
2245 2246
}

2247 2248 2249 2250
static void mem_cgroup_pre_destroy(struct cgroup_subsys *ss,
					struct cgroup *cont)
{
	struct mem_cgroup *mem = mem_cgroup_from_cont(cont);
2251
	mem_cgroup_force_empty(mem, false);
2252 2253
}

B
Balbir Singh 已提交
2254 2255 2256
static void mem_cgroup_destroy(struct cgroup_subsys *ss,
				struct cgroup *cont)
{
2257
	mem_cgroup_put(mem_cgroup_from_cont(cont));
B
Balbir Singh 已提交
2258 2259 2260 2261 2262
}

static int mem_cgroup_populate(struct cgroup_subsys *ss,
				struct cgroup *cont)
{
2263 2264 2265 2266 2267 2268 2269 2270
	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
2271 2272
}

B
Balbir Singh 已提交
2273 2274 2275 2276 2277
static void mem_cgroup_move_task(struct cgroup_subsys *ss,
				struct cgroup *cont,
				struct cgroup *old_cont,
				struct task_struct *p)
{
2278
	mutex_lock(&memcg_tasklist);
B
Balbir Singh 已提交
2279
	/*
2280 2281
	 * 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
2282
	 */
2283
	mutex_unlock(&memcg_tasklist);
B
Balbir Singh 已提交
2284 2285
}

B
Balbir Singh 已提交
2286 2287 2288 2289
struct cgroup_subsys mem_cgroup_subsys = {
	.name = "memory",
	.subsys_id = mem_cgroup_subsys_id,
	.create = mem_cgroup_create,
2290
	.pre_destroy = mem_cgroup_pre_destroy,
B
Balbir Singh 已提交
2291 2292
	.destroy = mem_cgroup_destroy,
	.populate = mem_cgroup_populate,
B
Balbir Singh 已提交
2293
	.attach = mem_cgroup_move_task,
2294
	.early_init = 0,
B
Balbir Singh 已提交
2295
};
2296 2297 2298 2299 2300 2301 2302 2303 2304 2305

#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