memcontrol.h 29.0 KB
Newer Older
B
Balbir Singh 已提交
1 2 3 4 5
/* memcontrol.h - 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
 * 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.
 */

#ifndef _LINUX_MEMCONTROL_H
#define _LINUX_MEMCONTROL_H
22
#include <linux/cgroup.h>
23
#include <linux/vm_event_item.h>
24
#include <linux/hardirq.h>
25
#include <linux/jump_label.h>
M
Michal Hocko 已提交
26 27 28
#include <linux/page_counter.h>
#include <linux/vmpressure.h>
#include <linux/eventfd.h>
29 30
#include <linux/mm.h>
#include <linux/vmstat.h>
M
Michal Hocko 已提交
31
#include <linux/writeback.h>
32
#include <linux/page-flags.h>
33

34
struct mem_cgroup;
35 36
struct page;
struct mm_struct;
37
struct kmem_cache;
38

39 40 41 42 43 44 45 46 47
/* Cgroup-specific page state, on top of universal node page state */
enum memcg_stat_item {
	MEMCG_CACHE = NR_VM_NODE_STAT_ITEMS,
	MEMCG_RSS,
	MEMCG_RSS_HUGE,
	MEMCG_SWAP,
	MEMCG_SOCK,
	/* XXX: why are these zone and not node counters? */
	MEMCG_KERNEL_STACK_KB,
48
	MEMCG_NR_STAT,
49 50
};

51 52 53 54 55 56 57 58 59
/* Cgroup-specific events, on top of universal VM events */
enum memcg_event_item {
	MEMCG_LOW = NR_VM_EVENT_ITEMS,
	MEMCG_HIGH,
	MEMCG_MAX,
	MEMCG_OOM,
	MEMCG_NR_EVENTS,
};

60
struct mem_cgroup_reclaim_cookie {
61
	pg_data_t *pgdat;
62 63 64 65
	int priority;
	unsigned int generation;
};

66 67 68 69 70 71 72 73 74 75
#ifdef CONFIG_MEMCG

#define MEM_CGROUP_ID_SHIFT	16
#define MEM_CGROUP_ID_MAX	USHRT_MAX

struct mem_cgroup_id {
	int id;
	atomic_t ref;
};

M
Michal Hocko 已提交
76 77 78 79 80 81 82 83 84 85 86 87 88 89
/*
 * Per memcg event counter is incremented at every pagein/pageout. With THP,
 * it will be incremated by the number of pages. This counter is used for
 * for trigger some periodic events. This is straightforward and better
 * than using jiffies etc. to handle periodic memcg event.
 */
enum mem_cgroup_events_target {
	MEM_CGROUP_TARGET_THRESH,
	MEM_CGROUP_TARGET_SOFTLIMIT,
	MEM_CGROUP_TARGET_NUMAINFO,
	MEM_CGROUP_NTARGETS,
};

struct mem_cgroup_stat_cpu {
90
	long count[MEMCG_NR_STAT];
M
Michal Hocko 已提交
91 92 93 94 95 96 97 98 99 100 101
	unsigned long events[MEMCG_NR_EVENTS];
	unsigned long nr_page_events;
	unsigned long targets[MEM_CGROUP_NTARGETS];
};

struct mem_cgroup_reclaim_iter {
	struct mem_cgroup *position;
	/* scan generation, increased every round-trip */
	unsigned int generation;
};

102 103 104 105
struct lruvec_stat {
	long count[NR_VM_NODE_STAT_ITEMS];
};

M
Michal Hocko 已提交
106 107 108
/*
 * per-zone information in memory controller.
 */
109
struct mem_cgroup_per_node {
M
Michal Hocko 已提交
110
	struct lruvec		lruvec;
111 112 113 114

	struct lruvec_stat __percpu *lruvec_stat_cpu;
	atomic_long_t		lruvec_stat[NR_VM_NODE_STAT_ITEMS];

115
	unsigned long		lru_zone_size[MAX_NR_ZONES][NR_LRU_LISTS];
M
Michal Hocko 已提交
116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152

	struct mem_cgroup_reclaim_iter	iter[DEF_PRIORITY + 1];

	struct rb_node		tree_node;	/* RB tree node */
	unsigned long		usage_in_excess;/* Set to the value by which */
						/* the soft limit is exceeded*/
	bool			on_tree;
	struct mem_cgroup	*memcg;		/* Back pointer, we cannot */
						/* use container_of	   */
};

struct mem_cgroup_threshold {
	struct eventfd_ctx *eventfd;
	unsigned long threshold;
};

/* For threshold */
struct mem_cgroup_threshold_ary {
	/* An array index points to threshold just below or equal to usage. */
	int current_threshold;
	/* Size of entries[] */
	unsigned int size;
	/* Array of thresholds */
	struct mem_cgroup_threshold entries[0];
};

struct mem_cgroup_thresholds {
	/* Primary thresholds array */
	struct mem_cgroup_threshold_ary *primary;
	/*
	 * Spare threshold array.
	 * This is needed to make mem_cgroup_unregister_event() "never fail".
	 * It must be able to store at least primary->size - 1 entries.
	 */
	struct mem_cgroup_threshold_ary *spare;
};

153 154 155 156 157 158
enum memcg_kmem_state {
	KMEM_NONE,
	KMEM_ALLOCATED,
	KMEM_ONLINE,
};

M
Michal Hocko 已提交
159 160 161 162 163 164 165 166 167
/*
 * 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.
 */
struct mem_cgroup {
	struct cgroup_subsys_state css;

168 169 170
	/* Private memcg ID. Used to ID objects that outlive the cgroup */
	struct mem_cgroup_id id;

M
Michal Hocko 已提交
171 172
	/* Accounted resources */
	struct page_counter memory;
173
	struct page_counter swap;
174 175

	/* Legacy consumer-oriented counters */
M
Michal Hocko 已提交
176 177
	struct page_counter memsw;
	struct page_counter kmem;
178
	struct page_counter tcpmem;
M
Michal Hocko 已提交
179 180 181 182 183

	/* Normal memory consumption range */
	unsigned long low;
	unsigned long high;

184 185 186
	/* Range enforcement for interrupt charges */
	struct work_struct high_work;

M
Michal Hocko 已提交
187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204
	unsigned long soft_limit;

	/* vmpressure notifications */
	struct vmpressure vmpressure;

	/*
	 * Should the accounting and control be hierarchical, per subtree?
	 */
	bool use_hierarchy;

	/* protected by memcg_oom_lock */
	bool		oom_lock;
	int		under_oom;

	int	swappiness;
	/* OOM-Killer disable */
	int		oom_kill_disable;

205 206 207
	/* handle for "memory.events" */
	struct cgroup_file events_file;

M
Michal Hocko 已提交
208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232
	/* protect arrays of thresholds */
	struct mutex thresholds_lock;

	/* thresholds for memory usage. RCU-protected */
	struct mem_cgroup_thresholds thresholds;

	/* thresholds for mem+swap usage. RCU-protected */
	struct mem_cgroup_thresholds memsw_thresholds;

	/* For oom notifier event fd */
	struct list_head oom_notify;

	/*
	 * Should we move charges of a task when a task is moved into this
	 * mem_cgroup ? And what type of charges should we move ?
	 */
	unsigned long move_charge_at_immigrate;
	/*
	 * set > 0 if pages under this cgroup are moving to other cgroup.
	 */
	atomic_t		moving_account;
	/* taken only while moving_account > 0 */
	spinlock_t		move_lock;
	struct task_struct	*move_lock_task;
	unsigned long		move_lock_flags;
233 234 235 236

	struct mem_cgroup_stat_cpu __percpu *stat_cpu;
	atomic_long_t		stat[MEMCG_NR_STAT];
	atomic_long_t		events[MEMCG_NR_EVENTS];
M
Michal Hocko 已提交
237

238 239 240
	unsigned long		socket_pressure;

	/* Legacy tcp memory accounting */
241 242
	bool			tcpmem_active;
	int			tcpmem_pressure;
243

244
#ifndef CONFIG_SLOB
M
Michal Hocko 已提交
245 246
        /* Index in the kmem_cache->memcg_params.memcg_caches array */
	int kmemcg_id;
247
	enum memcg_kmem_state kmem_state;
248
	struct list_head kmem_caches;
M
Michal Hocko 已提交
249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269
#endif

	int last_scanned_node;
#if MAX_NUMNODES > 1
	nodemask_t	scan_nodes;
	atomic_t	numainfo_events;
	atomic_t	numainfo_updating;
#endif

#ifdef CONFIG_CGROUP_WRITEBACK
	struct list_head cgwb_list;
	struct wb_domain cgwb_domain;
#endif

	/* List of events which userspace want to receive */
	struct list_head event_list;
	spinlock_t event_list_lock;

	struct mem_cgroup_per_node *nodeinfo[0];
	/* WARNING: nodeinfo must be the last member here */
};
270

271 272 273 274 275 276
/*
 * size of first charge trial. "32" comes from vmscan.c's magic value.
 * TODO: maybe necessary to use big numbers in big irons.
 */
#define MEMCG_CHARGE_BATCH 32U

277
extern struct mem_cgroup *root_mem_cgroup;
T
Tejun Heo 已提交
278

279 280 281 282 283
static inline bool mem_cgroup_disabled(void)
{
	return !cgroup_subsys_enabled(memory_cgrp_subsys);
}

284 285
bool mem_cgroup_low(struct mem_cgroup *root, struct mem_cgroup *memcg);

286
int mem_cgroup_try_charge(struct page *page, struct mm_struct *mm,
287 288
			  gfp_t gfp_mask, struct mem_cgroup **memcgp,
			  bool compound);
289
void mem_cgroup_commit_charge(struct page *page, struct mem_cgroup *memcg,
290 291 292
			      bool lrucare, bool compound);
void mem_cgroup_cancel_charge(struct page *page, struct mem_cgroup *memcg,
		bool compound);
293
void mem_cgroup_uncharge(struct page *page);
294
void mem_cgroup_uncharge_list(struct list_head *page_list);
295

296
void mem_cgroup_migrate(struct page *oldpage, struct page *newpage);
297

298 299
static struct mem_cgroup_per_node *
mem_cgroup_nodeinfo(struct mem_cgroup *memcg, int nid)
300
{
301
	return memcg->nodeinfo[nid];
302 303 304
}

/**
305 306
 * mem_cgroup_lruvec - get the lru list vector for a node or a memcg zone
 * @node: node of the wanted lruvec
307 308
 * @memcg: memcg of the wanted lruvec
 *
309 310
 * Returns the lru list vector holding pages for a given @node or a given
 * @memcg and @zone. This can be the node lruvec, if the memory controller
311 312
 * is disabled.
 */
313
static inline struct lruvec *mem_cgroup_lruvec(struct pglist_data *pgdat,
314
				struct mem_cgroup *memcg)
315
{
316
	struct mem_cgroup_per_node *mz;
317 318 319
	struct lruvec *lruvec;

	if (mem_cgroup_disabled()) {
320
		lruvec = node_lruvec(pgdat);
321 322 323
		goto out;
	}

324
	mz = mem_cgroup_nodeinfo(memcg, pgdat->node_id);
325 326 327 328
	lruvec = &mz->lruvec;
out:
	/*
	 * Since a node can be onlined after the mem_cgroup was created,
M
Mel Gorman 已提交
329
	 * we have to be prepared to initialize lruvec->pgdat here;
330 331
	 * and if offlined then reonlined, we need to reinitialize it.
	 */
332 333
	if (unlikely(lruvec->pgdat != pgdat))
		lruvec->pgdat = pgdat;
334 335 336
	return lruvec;
}

M
Mel Gorman 已提交
337
struct lruvec *mem_cgroup_page_lruvec(struct page *, struct pglist_data *);
338

339
bool task_in_mem_cgroup(struct task_struct *task, struct mem_cgroup *memcg);
340
struct mem_cgroup *mem_cgroup_from_task(struct task_struct *p);
341

M
Michal Hocko 已提交
342 343 344 345 346
static inline
struct mem_cgroup *mem_cgroup_from_css(struct cgroup_subsys_state *css){
	return css ? container_of(css, struct mem_cgroup, css) : NULL;
}

347 348 349
#define mem_cgroup_from_counter(counter, member)	\
	container_of(counter, struct mem_cgroup, member)

M
Michal Hocko 已提交
350 351 352 353
struct mem_cgroup *mem_cgroup_iter(struct mem_cgroup *,
				   struct mem_cgroup *,
				   struct mem_cgroup_reclaim_cookie *);
void mem_cgroup_iter_break(struct mem_cgroup *, struct mem_cgroup *);
354 355
int mem_cgroup_scan_tasks(struct mem_cgroup *,
			  int (*)(struct task_struct *, void *), void *);
M
Michal Hocko 已提交
356

357 358 359 360 361
static inline unsigned short mem_cgroup_id(struct mem_cgroup *memcg)
{
	if (mem_cgroup_disabled())
		return 0;

362
	return memcg->id.id;
363
}
364
struct mem_cgroup *mem_cgroup_from_id(unsigned short id);
365

366 367 368 369 370 371 372 373 374 375 376
static inline struct mem_cgroup *lruvec_memcg(struct lruvec *lruvec)
{
	struct mem_cgroup_per_node *mz;

	if (mem_cgroup_disabled())
		return NULL;

	mz = container_of(lruvec, struct mem_cgroup_per_node, lruvec);
	return mz->memcg;
}

377 378 379 380 381 382 383 384 385 386 387 388 389 390
/**
 * parent_mem_cgroup - find the accounting parent of a memcg
 * @memcg: memcg whose parent to find
 *
 * Returns the parent memcg, or NULL if this is the root or the memory
 * controller is in legacy no-hierarchy mode.
 */
static inline struct mem_cgroup *parent_mem_cgroup(struct mem_cgroup *memcg)
{
	if (!memcg->memory.parent)
		return NULL;
	return mem_cgroup_from_counter(memcg->memory.parent, memory);
}

M
Michal Hocko 已提交
391 392 393 394 395 396 397 398 399
static inline bool mem_cgroup_is_descendant(struct mem_cgroup *memcg,
			      struct mem_cgroup *root)
{
	if (root == memcg)
		return true;
	if (!root->use_hierarchy)
		return false;
	return cgroup_is_descendant(memcg->css.cgroup, root->css.cgroup);
}
G
Glauber Costa 已提交
400

401 402
static inline bool mm_match_cgroup(struct mm_struct *mm,
				   struct mem_cgroup *memcg)
403
{
404
	struct mem_cgroup *task_memcg;
405
	bool match = false;
406

407
	rcu_read_lock();
408
	task_memcg = mem_cgroup_from_task(rcu_dereference(mm->owner));
409
	if (task_memcg)
410
		match = mem_cgroup_is_descendant(task_memcg, memcg);
411
	rcu_read_unlock();
412
	return match;
413
}
414

415
struct cgroup_subsys_state *mem_cgroup_css_from_page(struct page *page);
416
ino_t page_cgroup_ino(struct page *page);
417

418 419 420 421 422 423 424
static inline bool mem_cgroup_online(struct mem_cgroup *memcg)
{
	if (mem_cgroup_disabled())
		return true;
	return !!(memcg->css.flags & CSS_ONLINE);
}

425 426 427
/*
 * For memory reclaim.
 */
428
int mem_cgroup_select_victim_node(struct mem_cgroup *memcg);
M
Michal Hocko 已提交
429 430

void mem_cgroup_update_lru_size(struct lruvec *lruvec, enum lru_list lru,
431
		int zid, int nr_pages);
M
Michal Hocko 已提交
432

433 434 435
unsigned long mem_cgroup_node_nr_lru_pages(struct mem_cgroup *memcg,
					   int nid, unsigned int lru_mask);

M
Michal Hocko 已提交
436 437 438
static inline
unsigned long mem_cgroup_get_lru_size(struct lruvec *lruvec, enum lru_list lru)
{
439
	struct mem_cgroup_per_node *mz;
440 441
	unsigned long nr_pages = 0;
	int zid;
M
Michal Hocko 已提交
442

443
	mz = container_of(lruvec, struct mem_cgroup_per_node, lruvec);
444 445 446 447 448 449 450 451 452 453 454 455 456
	for (zid = 0; zid < MAX_NR_ZONES; zid++)
		nr_pages += mz->lru_zone_size[zid][lru];
	return nr_pages;
}

static inline
unsigned long mem_cgroup_get_zone_lru_size(struct lruvec *lruvec,
		enum lru_list lru, int zone_idx)
{
	struct mem_cgroup_per_node *mz;

	mz = container_of(lruvec, struct mem_cgroup_per_node, lruvec);
	return mz->lru_zone_size[zone_idx][lru];
M
Michal Hocko 已提交
457 458
}

459 460
void mem_cgroup_handle_over_high(void);

461 462
unsigned long mem_cgroup_get_limit(struct mem_cgroup *memcg);

463 464
void mem_cgroup_print_oom_info(struct mem_cgroup *memcg,
				struct task_struct *p);
465

466
static inline void mem_cgroup_oom_enable(void)
467
{
T
Tejun Heo 已提交
468 469
	WARN_ON(current->memcg_may_oom);
	current->memcg_may_oom = 1;
470 471
}

472
static inline void mem_cgroup_oom_disable(void)
473
{
T
Tejun Heo 已提交
474 475
	WARN_ON(!current->memcg_may_oom);
	current->memcg_may_oom = 0;
476 477
}

478 479
static inline bool task_in_memcg_oom(struct task_struct *p)
{
T
Tejun Heo 已提交
480
	return p->memcg_in_oom;
481 482
}

483
bool mem_cgroup_oom_synchronize(bool wait);
484

A
Andrew Morton 已提交
485
#ifdef CONFIG_MEMCG_SWAP
486 487
extern int do_swap_account;
#endif
488

489 490
struct mem_cgroup *lock_page_memcg(struct page *page);
void __unlock_page_memcg(struct mem_cgroup *memcg);
J
Johannes Weiner 已提交
491
void unlock_page_memcg(struct page *page);
492

493
/* idx can be of type enum memcg_stat_item or node_stat_item */
494
static inline unsigned long memcg_page_state(struct mem_cgroup *memcg,
495
					     int idx)
496
{
497 498 499 500 501 502
	long x = atomic_long_read(&memcg->stat[idx]);
#ifdef CONFIG_SMP
	if (x < 0)
		x = 0;
#endif
	return x;
503 504
}

505
/* idx can be of type enum memcg_stat_item or node_stat_item */
506
static inline void __mod_memcg_state(struct mem_cgroup *memcg,
507
				     int idx, int val)
508
{
509 510 511 512 513 514 515 516 517 518 519
	long x;

	if (mem_cgroup_disabled())
		return;

	x = val + __this_cpu_read(memcg->stat_cpu->count[idx]);
	if (unlikely(abs(x) > MEMCG_CHARGE_BATCH)) {
		atomic_long_add(x, &memcg->stat[idx]);
		x = 0;
	}
	__this_cpu_write(memcg->stat_cpu->count[idx], x);
520 521
}

522
/* idx can be of type enum memcg_stat_item or node_stat_item */
523
static inline void mod_memcg_state(struct mem_cgroup *memcg,
524
				   int idx, int val)
525
{
526 527 528
	unsigned long flags;

	local_irq_save(flags);
529
	__mod_memcg_state(memcg, idx, val);
530
	local_irq_restore(flags);
531 532
}

M
Michal Hocko 已提交
533
/**
534
 * mod_memcg_page_state - update page state statistics
J
Johannes Weiner 已提交
535
 * @page: the page
M
Michal Hocko 已提交
536 537 538
 * @idx: page state item to account
 * @val: number of pages (positive or negative)
 *
539 540 541
 * The @page must be locked or the caller must use lock_page_memcg()
 * to prevent double accounting when the page is concurrently being
 * moved to another memcg:
542
 *
543
 *   lock_page(page) or lock_page_memcg(page)
544
 *   if (TestClearPageState(page))
545
 *     mod_memcg_page_state(page, state, -1);
546
 *   unlock_page(page) or unlock_page_memcg(page)
547 548
 *
 * Kernel pages are an exception to this, since they'll never move.
M
Michal Hocko 已提交
549
 */
550
static inline void __mod_memcg_page_state(struct page *page,
551
					  int idx, int val)
552 553 554 555 556
{
	if (page->mem_cgroup)
		__mod_memcg_state(page->mem_cgroup, idx, val);
}

557
static inline void mod_memcg_page_state(struct page *page,
558
					int idx, int val)
M
Michal Hocko 已提交
559
{
J
Johannes Weiner 已提交
560
	if (page->mem_cgroup)
561
		mod_memcg_state(page->mem_cgroup, idx, val);
M
Michal Hocko 已提交
562 563
}

564 565
static inline unsigned long lruvec_page_state(struct lruvec *lruvec,
					      enum node_stat_item idx)
566
{
567
	struct mem_cgroup_per_node *pn;
568
	long x;
569 570 571 572 573

	if (mem_cgroup_disabled())
		return node_page_state(lruvec_pgdat(lruvec), idx);

	pn = container_of(lruvec, struct mem_cgroup_per_node, lruvec);
574 575 576 577 578 579
	x = atomic_long_read(&pn->lruvec_stat[idx]);
#ifdef CONFIG_SMP
	if (x < 0)
		x = 0;
#endif
	return x;
580 581
}

582 583
static inline void __mod_lruvec_state(struct lruvec *lruvec,
				      enum node_stat_item idx, int val)
584
{
585
	struct mem_cgroup_per_node *pn;
586
	long x;
587

588
	/* Update node */
589
	__mod_node_page_state(lruvec_pgdat(lruvec), idx, val);
590

591 592
	if (mem_cgroup_disabled())
		return;
593

594
	pn = container_of(lruvec, struct mem_cgroup_per_node, lruvec);
595 596

	/* Update memcg */
597
	__mod_memcg_state(pn->memcg, idx, val);
598 599

	/* Update lruvec */
600 601 602 603 604 605
	x = val + __this_cpu_read(pn->lruvec_stat_cpu->count[idx]);
	if (unlikely(abs(x) > MEMCG_CHARGE_BATCH)) {
		atomic_long_add(x, &pn->lruvec_stat[idx]);
		x = 0;
	}
	__this_cpu_write(pn->lruvec_stat_cpu->count[idx], x);
606 607 608 609 610
}

static inline void mod_lruvec_state(struct lruvec *lruvec,
				    enum node_stat_item idx, int val)
{
611 612 613
	unsigned long flags;

	local_irq_save(flags);
614
	__mod_lruvec_state(lruvec, idx, val);
615
	local_irq_restore(flags);
616 617 618 619 620
}

static inline void __mod_lruvec_page_state(struct page *page,
					   enum node_stat_item idx, int val)
{
621 622
	pg_data_t *pgdat = page_pgdat(page);
	struct lruvec *lruvec;
623

624 625 626
	/* Untracked pages have no memcg, no lruvec. Update only the node */
	if (!page->mem_cgroup) {
		__mod_node_page_state(pgdat, idx, val);
627
		return;
628 629 630 631
	}

	lruvec = mem_cgroup_lruvec(pgdat, page->mem_cgroup);
	__mod_lruvec_state(lruvec, idx, val);
632 633 634 635 636
}

static inline void mod_lruvec_page_state(struct page *page,
					 enum node_stat_item idx, int val)
{
637 638 639
	unsigned long flags;

	local_irq_save(flags);
640
	__mod_lruvec_page_state(page, idx, val);
641
	local_irq_restore(flags);
642 643
}

644
unsigned long mem_cgroup_soft_limit_reclaim(pg_data_t *pgdat, int order,
645 646
						gfp_t gfp_mask,
						unsigned long *total_scanned);
D
David Rientjes 已提交
647

648 649 650 651
/* idx can be of type enum memcg_event_item or vm_event_item */
static inline void __count_memcg_events(struct mem_cgroup *memcg,
					int idx, unsigned long count)
{
652 653 654 655 656 657 658 659 660 661 662
	unsigned long x;

	if (mem_cgroup_disabled())
		return;

	x = count + __this_cpu_read(memcg->stat_cpu->events[idx]);
	if (unlikely(x > MEMCG_CHARGE_BATCH)) {
		atomic_long_add(x, &memcg->events[idx]);
		x = 0;
	}
	__this_cpu_write(memcg->stat_cpu->events[idx], x);
663 664
}

665
static inline void count_memcg_events(struct mem_cgroup *memcg,
666
				      int idx, unsigned long count)
667
{
668 669 670
	unsigned long flags;

	local_irq_save(flags);
671
	__count_memcg_events(memcg, idx, count);
672
	local_irq_restore(flags);
673 674
}

675
/* idx can be of type enum memcg_event_item or vm_event_item */
676
static inline void count_memcg_page_event(struct page *page,
677
					  int idx)
678 679 680 681 682 683 684
{
	if (page->mem_cgroup)
		count_memcg_events(page->mem_cgroup, idx, 1);
}

static inline void count_memcg_event_mm(struct mm_struct *mm,
					enum vm_event_item idx)
685
{
M
Michal Hocko 已提交
686 687
	struct mem_cgroup *memcg;

688 689
	if (mem_cgroup_disabled())
		return;
M
Michal Hocko 已提交
690 691 692

	rcu_read_lock();
	memcg = mem_cgroup_from_task(rcu_dereference(mm->owner));
693
	if (likely(memcg)) {
694
		count_memcg_events(memcg, idx, 1);
695 696 697
		if (idx == OOM_KILL)
			cgroup_file_notify(&memcg->events_file);
	}
M
Michal Hocko 已提交
698
	rcu_read_unlock();
699
}
700 701 702 703 704 705 706 707

static inline void mem_cgroup_event(struct mem_cgroup *memcg,
				    enum memcg_event_item event)
{
	count_memcg_events(memcg, event, 1);
	cgroup_file_notify(&memcg->events_file);
}

708
#ifdef CONFIG_TRANSPARENT_HUGEPAGE
709
void mem_cgroup_split_huge_fixup(struct page *head);
710 711
#endif

A
Andrew Morton 已提交
712
#else /* CONFIG_MEMCG */
713 714 715 716

#define MEM_CGROUP_ID_SHIFT	0
#define MEM_CGROUP_ID_MAX	0

717 718
struct mem_cgroup;

719 720 721 722 723
static inline bool mem_cgroup_disabled(void)
{
	return true;
}

724
static inline void mem_cgroup_event(struct mem_cgroup *memcg,
725
				    enum memcg_event_item event)
726 727 728 729 730 731 732 733 734
{
}

static inline bool mem_cgroup_low(struct mem_cgroup *root,
				  struct mem_cgroup *memcg)
{
	return false;
}

735 736
static inline int mem_cgroup_try_charge(struct page *page, struct mm_struct *mm,
					gfp_t gfp_mask,
737 738
					struct mem_cgroup **memcgp,
					bool compound)
739
{
740
	*memcgp = NULL;
741 742 743
	return 0;
}

744 745
static inline void mem_cgroup_commit_charge(struct page *page,
					    struct mem_cgroup *memcg,
746
					    bool lrucare, bool compound)
747 748 749
{
}

750
static inline void mem_cgroup_cancel_charge(struct page *page,
751 752
					    struct mem_cgroup *memcg,
					    bool compound)
753 754 755
{
}

756
static inline void mem_cgroup_uncharge(struct page *page)
757 758 759
{
}

760
static inline void mem_cgroup_uncharge_list(struct list_head *page_list)
761 762 763
{
}

764
static inline void mem_cgroup_migrate(struct page *old, struct page *new)
765 766 767
{
}

768
static inline struct lruvec *mem_cgroup_lruvec(struct pglist_data *pgdat,
769
				struct mem_cgroup *memcg)
K
KAMEZAWA Hiroyuki 已提交
770
{
771
	return node_lruvec(pgdat);
K
KAMEZAWA Hiroyuki 已提交
772 773
}

774
static inline struct lruvec *mem_cgroup_page_lruvec(struct page *page,
M
Mel Gorman 已提交
775
						    struct pglist_data *pgdat)
776
{
M
Mel Gorman 已提交
777
	return &pgdat->lruvec;
778 779
}

780
static inline bool mm_match_cgroup(struct mm_struct *mm,
781
		struct mem_cgroup *memcg)
782
{
783
	return true;
784 785
}

786 787
static inline bool task_in_mem_cgroup(struct task_struct *task,
				      const struct mem_cgroup *memcg)
788
{
789
	return true;
790 791
}

792 793 794 795 796 797 798 799 800 801 802 803 804
static inline struct mem_cgroup *
mem_cgroup_iter(struct mem_cgroup *root,
		struct mem_cgroup *prev,
		struct mem_cgroup_reclaim_cookie *reclaim)
{
	return NULL;
}

static inline void mem_cgroup_iter_break(struct mem_cgroup *root,
					 struct mem_cgroup *prev)
{
}

805 806 807 808 809 810
static inline int mem_cgroup_scan_tasks(struct mem_cgroup *memcg,
		int (*fn)(struct task_struct *, void *), void *arg)
{
	return 0;
}

811
static inline unsigned short mem_cgroup_id(struct mem_cgroup *memcg)
812
{
813 814 815 816 817 818 819 820
	return 0;
}

static inline struct mem_cgroup *mem_cgroup_from_id(unsigned short id)
{
	WARN_ON_ONCE(id);
	/* XXX: This should always return root_mem_cgroup */
	return NULL;
821
}
822

823 824 825 826 827
static inline struct mem_cgroup *lruvec_memcg(struct lruvec *lruvec)
{
	return NULL;
}

828
static inline bool mem_cgroup_online(struct mem_cgroup *memcg)
829
{
830
	return true;
831 832
}

833
static inline unsigned long
834
mem_cgroup_get_lru_size(struct lruvec *lruvec, enum lru_list lru)
835 836 837
{
	return 0;
}
838 839 840 841 842 843
static inline
unsigned long mem_cgroup_get_zone_lru_size(struct lruvec *lruvec,
		enum lru_list lru, int zone_idx)
{
	return 0;
}
844

845 846 847 848 849 850 851
static inline unsigned long
mem_cgroup_node_nr_lru_pages(struct mem_cgroup *memcg,
			     int nid, unsigned int lru_mask)
{
	return 0;
}

852 853 854 855 856
static inline unsigned long mem_cgroup_get_limit(struct mem_cgroup *memcg)
{
	return 0;
}

857 858 859 860 861
static inline void
mem_cgroup_print_oom_info(struct mem_cgroup *memcg, struct task_struct *p)
{
}

862 863 864 865 866 867
static inline struct mem_cgroup *lock_page_memcg(struct page *page)
{
	return NULL;
}

static inline void __unlock_page_memcg(struct mem_cgroup *memcg)
868 869 870
{
}

J
Johannes Weiner 已提交
871
static inline void unlock_page_memcg(struct page *page)
872 873 874
{
}

875 876 877 878
static inline void mem_cgroup_handle_over_high(void)
{
}

879
static inline void mem_cgroup_oom_enable(void)
880 881 882
{
}

883
static inline void mem_cgroup_oom_disable(void)
884 885 886
{
}

887 888 889 890 891
static inline bool task_in_memcg_oom(struct task_struct *p)
{
	return false;
}

892
static inline bool mem_cgroup_oom_synchronize(bool wait)
893 894 895 896
{
	return false;
}

897
static inline unsigned long memcg_page_state(struct mem_cgroup *memcg,
898
					     int idx)
899 900 901 902
{
	return 0;
}

903
static inline void __mod_memcg_state(struct mem_cgroup *memcg,
904
				     int idx,
905
				     int nr)
906 907 908
{
}

909
static inline void mod_memcg_state(struct mem_cgroup *memcg,
910
				   int idx,
911
				   int nr)
912 913 914
{
}

915
static inline void __mod_memcg_page_state(struct page *page,
916
					  int idx,
917
					  int nr)
918 919 920
{
}

921
static inline void mod_memcg_page_state(struct page *page,
922
					int idx,
923
					int nr)
924 925 926
{
}

927 928
static inline unsigned long lruvec_page_state(struct lruvec *lruvec,
					      enum node_stat_item idx)
929
{
930
	return node_page_state(lruvec_pgdat(lruvec), idx);
931 932
}

933 934
static inline void __mod_lruvec_state(struct lruvec *lruvec,
				      enum node_stat_item idx, int val)
935
{
936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954
	__mod_node_page_state(lruvec_pgdat(lruvec), idx, val);
}

static inline void mod_lruvec_state(struct lruvec *lruvec,
				    enum node_stat_item idx, int val)
{
	mod_node_page_state(lruvec_pgdat(lruvec), idx, val);
}

static inline void __mod_lruvec_page_state(struct page *page,
					   enum node_stat_item idx, int val)
{
	__mod_node_page_state(page_pgdat(page), idx, val);
}

static inline void mod_lruvec_page_state(struct page *page,
					 enum node_stat_item idx, int val)
{
	mod_node_page_state(page_pgdat(page), idx, val);
955 956
}

957
static inline
958
unsigned long mem_cgroup_soft_limit_reclaim(pg_data_t *pgdat, int order,
959 960
					    gfp_t gfp_mask,
					    unsigned long *total_scanned)
961
{
962
	return 0;
963 964
}

965
static inline void mem_cgroup_split_huge_fixup(struct page *head)
966 967 968
{
}

969 970 971 972 973 974 975
static inline void count_memcg_events(struct mem_cgroup *memcg,
				      enum vm_event_item idx,
				      unsigned long count)
{
}

static inline void count_memcg_page_event(struct page *page,
976
					  int idx)
977 978 979
{
}

980
static inline
981
void count_memcg_event_mm(struct mm_struct *mm, enum vm_event_item idx)
982 983
{
}
A
Andrew Morton 已提交
984
#endif /* CONFIG_MEMCG */
985

986
/* idx can be of type enum memcg_stat_item or node_stat_item */
987
static inline void __inc_memcg_state(struct mem_cgroup *memcg,
988
				     int idx)
989 990 991 992
{
	__mod_memcg_state(memcg, idx, 1);
}

993
/* idx can be of type enum memcg_stat_item or node_stat_item */
994
static inline void __dec_memcg_state(struct mem_cgroup *memcg,
995
				     int idx)
996 997 998 999
{
	__mod_memcg_state(memcg, idx, -1);
}

1000
/* idx can be of type enum memcg_stat_item or node_stat_item */
1001
static inline void __inc_memcg_page_state(struct page *page,
1002
					  int idx)
1003 1004 1005 1006
{
	__mod_memcg_page_state(page, idx, 1);
}

1007
/* idx can be of type enum memcg_stat_item or node_stat_item */
1008
static inline void __dec_memcg_page_state(struct page *page,
1009
					  int idx)
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
{
	__mod_memcg_page_state(page, idx, -1);
}

static inline void __inc_lruvec_state(struct lruvec *lruvec,
				      enum node_stat_item idx)
{
	__mod_lruvec_state(lruvec, idx, 1);
}

static inline void __dec_lruvec_state(struct lruvec *lruvec,
				      enum node_stat_item idx)
{
	__mod_lruvec_state(lruvec, idx, -1);
}

static inline void __inc_lruvec_page_state(struct page *page,
					   enum node_stat_item idx)
{
	__mod_lruvec_page_state(page, idx, 1);
}

static inline void __dec_lruvec_page_state(struct page *page,
					   enum node_stat_item idx)
{
	__mod_lruvec_page_state(page, idx, -1);
}

1038
/* idx can be of type enum memcg_stat_item or node_stat_item */
1039
static inline void inc_memcg_state(struct mem_cgroup *memcg,
1040
				   int idx)
1041 1042 1043 1044
{
	mod_memcg_state(memcg, idx, 1);
}

1045
/* idx can be of type enum memcg_stat_item or node_stat_item */
1046
static inline void dec_memcg_state(struct mem_cgroup *memcg,
1047
				   int idx)
1048 1049 1050 1051
{
	mod_memcg_state(memcg, idx, -1);
}

1052
/* idx can be of type enum memcg_stat_item or node_stat_item */
1053
static inline void inc_memcg_page_state(struct page *page,
1054
					int idx)
1055 1056 1057 1058
{
	mod_memcg_page_state(page, idx, 1);
}

1059
/* idx can be of type enum memcg_stat_item or node_stat_item */
1060
static inline void dec_memcg_page_state(struct page *page,
1061
					int idx)
1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089
{
	mod_memcg_page_state(page, idx, -1);
}

static inline void inc_lruvec_state(struct lruvec *lruvec,
				    enum node_stat_item idx)
{
	mod_lruvec_state(lruvec, idx, 1);
}

static inline void dec_lruvec_state(struct lruvec *lruvec,
				    enum node_stat_item idx)
{
	mod_lruvec_state(lruvec, idx, -1);
}

static inline void inc_lruvec_page_state(struct page *page,
					 enum node_stat_item idx)
{
	mod_lruvec_page_state(page, idx, 1);
}

static inline void dec_lruvec_page_state(struct page *page,
					 enum node_stat_item idx)
{
	mod_lruvec_page_state(page, idx, -1);
}

1090
#ifdef CONFIG_CGROUP_WRITEBACK
T
Tejun Heo 已提交
1091

1092
struct list_head *mem_cgroup_cgwb_list(struct mem_cgroup *memcg);
T
Tejun Heo 已提交
1093
struct wb_domain *mem_cgroup_wb_domain(struct bdi_writeback *wb);
1094 1095 1096
void mem_cgroup_wb_stats(struct bdi_writeback *wb, unsigned long *pfilepages,
			 unsigned long *pheadroom, unsigned long *pdirty,
			 unsigned long *pwriteback);
T
Tejun Heo 已提交
1097 1098 1099 1100 1101 1102 1103 1104

#else	/* CONFIG_CGROUP_WRITEBACK */

static inline struct wb_domain *mem_cgroup_wb_domain(struct bdi_writeback *wb)
{
	return NULL;
}

1105
static inline void mem_cgroup_wb_stats(struct bdi_writeback *wb,
1106 1107
				       unsigned long *pfilepages,
				       unsigned long *pheadroom,
1108 1109 1110 1111 1112
				       unsigned long *pdirty,
				       unsigned long *pwriteback)
{
}

T
Tejun Heo 已提交
1113
#endif	/* CONFIG_CGROUP_WRITEBACK */
1114

G
Glauber Costa 已提交
1115
struct sock;
1116 1117
bool mem_cgroup_charge_skmem(struct mem_cgroup *memcg, unsigned int nr_pages);
void mem_cgroup_uncharge_skmem(struct mem_cgroup *memcg, unsigned int nr_pages);
1118
#ifdef CONFIG_MEMCG
1119 1120
extern struct static_key_false memcg_sockets_enabled_key;
#define mem_cgroup_sockets_enabled static_branch_unlikely(&memcg_sockets_enabled_key)
1121 1122
void mem_cgroup_sk_alloc(struct sock *sk);
void mem_cgroup_sk_free(struct sock *sk);
1123
static inline bool mem_cgroup_under_socket_pressure(struct mem_cgroup *memcg)
1124
{
1125
	if (!cgroup_subsys_on_dfl(memory_cgrp_subsys) && memcg->tcpmem_pressure)
1126 1127 1128 1129 1130 1131
		return true;
	do {
		if (time_before(jiffies, memcg->socket_pressure))
			return true;
	} while ((memcg = parent_mem_cgroup(memcg)));
	return false;
1132 1133
}
#else
1134
#define mem_cgroup_sockets_enabled 0
1135 1136
static inline void mem_cgroup_sk_alloc(struct sock *sk) { };
static inline void mem_cgroup_sk_free(struct sock *sk) { };
1137
static inline bool mem_cgroup_under_socket_pressure(struct mem_cgroup *memcg)
1138 1139 1140 1141
{
	return false;
}
#endif
1142

1143 1144 1145 1146 1147 1148 1149
struct kmem_cache *memcg_kmem_get_cache(struct kmem_cache *cachep);
void memcg_kmem_put_cache(struct kmem_cache *cachep);
int memcg_kmem_charge_memcg(struct page *page, gfp_t gfp, int order,
			    struct mem_cgroup *memcg);
int memcg_kmem_charge(struct page *page, gfp_t gfp, int order);
void memcg_kmem_uncharge(struct page *page, int order);

1150
#if defined(CONFIG_MEMCG) && !defined(CONFIG_SLOB)
1151
extern struct static_key_false memcg_kmem_enabled_key;
1152
extern struct workqueue_struct *memcg_kmem_cache_wq;
1153

1154
extern int memcg_nr_cache_ids;
1155 1156
void memcg_get_cache_ids(void);
void memcg_put_cache_ids(void);
1157 1158 1159 1160 1161 1162

/*
 * Helper macro to loop through all memcg-specific caches. Callers must still
 * check if the cache is valid (it is either valid or NULL).
 * the slab_mutex must be held when looping through those caches
 */
1163
#define for_each_memcg_cache_index(_idx)	\
1164
	for ((_idx) = 0; (_idx) < memcg_nr_cache_ids; (_idx)++)
1165

1166 1167
static inline bool memcg_kmem_enabled(void)
{
1168
	return static_branch_unlikely(&memcg_kmem_enabled_key);
1169 1170
}

M
Michal Hocko 已提交
1171
/*
J
Jesper Dangaard Brouer 已提交
1172
 * helper for accessing a memcg's index. It will be used as an index in the
M
Michal Hocko 已提交
1173 1174 1175 1176 1177 1178 1179
 * child cache array in kmem_cache, and also to derive its name. This function
 * will return -1 when this is not a kmem-limited memcg.
 */
static inline int memcg_cache_id(struct mem_cgroup *memcg)
{
	return memcg ? memcg->kmemcg_id : -1;
}
1180

1181
#else
1182 1183 1184
#define for_each_memcg_cache_index(_idx)	\
	for (; NULL; )

1185 1186 1187 1188 1189
static inline bool memcg_kmem_enabled(void)
{
	return false;
}

1190 1191 1192 1193 1194
static inline int memcg_cache_id(struct mem_cgroup *memcg)
{
	return -1;
}

1195 1196 1197 1198 1199 1200 1201 1202
static inline void memcg_get_cache_ids(void)
{
}

static inline void memcg_put_cache_ids(void)
{
}

1203 1204
#endif /* CONFIG_MEMCG && !CONFIG_SLOB */

B
Balbir Singh 已提交
1205
#endif /* _LINUX_MEMCONTROL_H */