memcontrol.h 21.4 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 29 30
#include <linux/page_counter.h>
#include <linux/vmpressure.h>
#include <linux/eventfd.h>
#include <linux/mmzone.h>
#include <linux/writeback.h>
31

32
struct mem_cgroup;
33 34
struct page;
struct mm_struct;
35
struct kmem_cache;
36

S
Sha Zhengju 已提交
37 38 39 40 41 42 43 44 45 46 47 48
/*
 * The corresponding mem_cgroup_stat_names is defined in mm/memcontrol.c,
 * These two lists should keep in accord with each other.
 */
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 anon rss */
	MEM_CGROUP_STAT_RSS_HUGE,	/* # of pages charged as anon huge */
	MEM_CGROUP_STAT_FILE_MAPPED,	/* # of pages charged as file rss */
49
	MEM_CGROUP_STAT_DIRTY,          /* # of dirty pages in page cache */
50
	MEM_CGROUP_STAT_WRITEBACK,	/* # of pages under writeback */
S
Sha Zhengju 已提交
51 52
	MEM_CGROUP_STAT_SWAP,		/* # of pages, swapped out */
	MEM_CGROUP_STAT_NSTATS,
53 54
};

55 56 57 58 59 60
struct mem_cgroup_reclaim_cookie {
	struct zone *zone;
	int priority;
	unsigned int generation;
};

61 62 63 64 65 66 67 68 69 70 71 72 73 74
enum mem_cgroup_events_index {
	MEM_CGROUP_EVENTS_PGPGIN,	/* # of pages paged in */
	MEM_CGROUP_EVENTS_PGPGOUT,	/* # of pages paged out */
	MEM_CGROUP_EVENTS_PGFAULT,	/* # of page-faults */
	MEM_CGROUP_EVENTS_PGMAJFAULT,	/* # of major page-faults */
	MEM_CGROUP_EVENTS_NSTATS,
	/* default hierarchy events */
	MEMCG_LOW = MEM_CGROUP_EVENTS_NSTATS,
	MEMCG_HIGH,
	MEMCG_MAX,
	MEMCG_OOM,
	MEMCG_NR_EVENTS,
};

M
Michal Hocko 已提交
75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91
/*
 * 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 cg_proto {
	struct page_counter	memory_allocated;	/* Current allocated memory. */
	struct percpu_counter	sockets_allocated;	/* Current number of sockets. */
	int			memory_pressure;
92
	bool			active;
M
Michal Hocko 已提交
93 94 95 96 97 98 99 100 101 102 103 104 105
	long			sysctl_mem[3];
	/*
	 * memcg field is used to find which memcg we belong directly
	 * Each memcg struct can hold more than one cg_proto, so container_of
	 * won't really cut.
	 *
	 * The elegant solution would be having an inverse function to
	 * proto_cgroup in struct proto, but that means polluting the structure
	 * for everybody, instead of just for memcg users.
	 */
	struct mem_cgroup	*memcg;
};

A
Andrew Morton 已提交
106
#ifdef CONFIG_MEMCG
M
Michal Hocko 已提交
107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205
struct mem_cgroup_stat_cpu {
	long count[MEM_CGROUP_STAT_NSTATS];
	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;
};

/*
 * per-zone information in memory controller.
 */
struct mem_cgroup_per_zone {
	struct lruvec		lruvec;
	unsigned long		lru_size[NR_LRU_LISTS];

	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_per_node {
	struct mem_cgroup_per_zone zoneinfo[MAX_NR_ZONES];
};

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;
};

/*
 * 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;

	/* Accounted resources */
	struct page_counter memory;
	struct page_counter memsw;
	struct page_counter kmem;

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

	unsigned long soft_limit;

	/* vmpressure notifications */
	struct vmpressure vmpressure;

	/* css_online() has been completed */
	int initialized;

	/*
	 * 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;

206 207 208
	/* handle for "memory.events" */
	struct cgroup_file events_file;

M
Michal Hocko 已提交
209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267
	/* 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;
	/*
	 * percpu counter.
	 */
	struct mem_cgroup_stat_cpu __percpu *stat;

#if defined(CONFIG_MEMCG_KMEM) && defined(CONFIG_INET)
	struct cg_proto tcp_mem;
#endif
#if defined(CONFIG_MEMCG_KMEM)
        /* Index in the kmem_cache->memcg_params.memcg_caches array */
	int kmemcg_id;
	bool kmem_acct_activated;
	bool kmem_acct_active;
#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 */
};
T
Tejun Heo 已提交
268 269
extern struct cgroup_subsys_state *mem_cgroup_root_css;

M
Michal Hocko 已提交
270 271 272 273 274 275 276
/**
 * mem_cgroup_events - count memory events against a cgroup
 * @memcg: the memory cgroup
 * @idx: the event index
 * @nr: the number of events to account for
 */
static inline void mem_cgroup_events(struct mem_cgroup *memcg,
277
		       enum mem_cgroup_events_index idx,
M
Michal Hocko 已提交
278 279 280
		       unsigned int nr)
{
	this_cpu_add(memcg->stat->events[idx], nr);
281
	cgroup_file_notify(&memcg->events_file);
M
Michal Hocko 已提交
282
}
283 284 285

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

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

294
void mem_cgroup_replace_page(struct page *oldpage, struct page *newpage);
295

296 297
struct lruvec *mem_cgroup_zone_lruvec(struct zone *, struct mem_cgroup *);
struct lruvec *mem_cgroup_page_lruvec(struct page *, struct zone *);
298

299
bool task_in_mem_cgroup(struct task_struct *task, struct mem_cgroup *memcg);
300 301
struct mem_cgroup *mem_cgroup_from_task(struct task_struct *p);
struct mem_cgroup *parent_mem_cgroup(struct mem_cgroup *memcg);
302

M
Michal Hocko 已提交
303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321
static inline
struct mem_cgroup *mem_cgroup_from_css(struct cgroup_subsys_state *css){
	return css ? container_of(css, struct mem_cgroup, css) : NULL;
}

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 *);

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
322

323 324
static inline bool mm_match_cgroup(struct mm_struct *mm,
				   struct mem_cgroup *memcg)
325
{
326
	struct mem_cgroup *task_memcg;
327
	bool match = false;
328

329
	rcu_read_lock();
330
	task_memcg = mem_cgroup_from_task(rcu_dereference(mm->owner));
331
	if (task_memcg)
332
		match = mem_cgroup_is_descendant(task_memcg, memcg);
333
	rcu_read_unlock();
334
	return match;
335
}
336

337
struct cgroup_subsys_state *mem_cgroup_css_from_page(struct page *page);
338
ino_t page_cgroup_ino(struct page *page);
339

M
Michal Hocko 已提交
340 341
static inline bool mem_cgroup_disabled(void)
{
342
	return !cgroup_subsys_enabled(memory_cgrp_subsys);
M
Michal Hocko 已提交
343
}
344

345 346 347
/*
 * For memory reclaim.
 */
348
int mem_cgroup_select_victim_node(struct mem_cgroup *memcg);
M
Michal Hocko 已提交
349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375

void mem_cgroup_update_lru_size(struct lruvec *lruvec, enum lru_list lru,
		int nr_pages);

static inline bool mem_cgroup_lruvec_online(struct lruvec *lruvec)
{
	struct mem_cgroup_per_zone *mz;
	struct mem_cgroup *memcg;

	if (mem_cgroup_disabled())
		return true;

	mz = container_of(lruvec, struct mem_cgroup_per_zone, lruvec);
	memcg = mz->memcg;

	return !!(memcg->css.flags & CSS_ONLINE);
}

static inline
unsigned long mem_cgroup_get_lru_size(struct lruvec *lruvec, enum lru_list lru)
{
	struct mem_cgroup_per_zone *mz;

	mz = container_of(lruvec, struct mem_cgroup_per_zone, lruvec);
	return mz->lru_size[lru];
}

376
static inline bool mem_cgroup_inactive_anon_is_low(struct lruvec *lruvec)
M
Michal Hocko 已提交
377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394
{
	unsigned long inactive_ratio;
	unsigned long inactive;
	unsigned long active;
	unsigned long gb;

	inactive = mem_cgroup_get_lru_size(lruvec, LRU_INACTIVE_ANON);
	active = mem_cgroup_get_lru_size(lruvec, LRU_ACTIVE_ANON);

	gb = (inactive + active) >> (30 - PAGE_SHIFT);
	if (gb)
		inactive_ratio = int_sqrt(10 * gb);
	else
		inactive_ratio = 1;

	return inactive * inactive_ratio < active;
}

395 396
void mem_cgroup_handle_over_high(void);

397 398
void mem_cgroup_print_oom_info(struct mem_cgroup *memcg,
				struct task_struct *p);
399

400
static inline void mem_cgroup_oom_enable(void)
401
{
T
Tejun Heo 已提交
402 403
	WARN_ON(current->memcg_may_oom);
	current->memcg_may_oom = 1;
404 405
}

406
static inline void mem_cgroup_oom_disable(void)
407
{
T
Tejun Heo 已提交
408 409
	WARN_ON(!current->memcg_may_oom);
	current->memcg_may_oom = 0;
410 411
}

412 413
static inline bool task_in_memcg_oom(struct task_struct *p)
{
T
Tejun Heo 已提交
414
	return p->memcg_in_oom;
415 416
}

417
bool mem_cgroup_oom_synchronize(bool wait);
418

A
Andrew Morton 已提交
419
#ifdef CONFIG_MEMCG_SWAP
420 421
extern int do_swap_account;
#endif
422

423 424
struct mem_cgroup *mem_cgroup_begin_page_stat(struct page *page);
void mem_cgroup_end_page_stat(struct mem_cgroup *memcg);
425

M
Michal Hocko 已提交
426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442
/**
 * mem_cgroup_update_page_stat - update page state statistics
 * @memcg: memcg to account against
 * @idx: page state item to account
 * @val: number of pages (positive or negative)
 *
 * See mem_cgroup_begin_page_stat() for locking requirements.
 */
static inline void mem_cgroup_update_page_stat(struct mem_cgroup *memcg,
				 enum mem_cgroup_stat_index idx, int val)
{
	VM_BUG_ON(!rcu_read_lock_held());

	if (memcg)
		this_cpu_add(memcg->stat->count[idx], val);
}

443
static inline void mem_cgroup_inc_page_stat(struct mem_cgroup *memcg,
S
Sha Zhengju 已提交
444
					    enum mem_cgroup_stat_index idx)
445
{
446
	mem_cgroup_update_page_stat(memcg, idx, 1);
447 448
}

449
static inline void mem_cgroup_dec_page_stat(struct mem_cgroup *memcg,
S
Sha Zhengju 已提交
450
					    enum mem_cgroup_stat_index idx)
451
{
452
	mem_cgroup_update_page_stat(memcg, idx, -1);
453 454
}

455 456 457
unsigned long mem_cgroup_soft_limit_reclaim(struct zone *zone, int order,
						gfp_t gfp_mask,
						unsigned long *total_scanned);
D
David Rientjes 已提交
458

459 460 461
static inline void mem_cgroup_count_vm_event(struct mm_struct *mm,
					     enum vm_event_item idx)
{
M
Michal Hocko 已提交
462 463
	struct mem_cgroup *memcg;

464 465
	if (mem_cgroup_disabled())
		return;
M
Michal Hocko 已提交
466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483

	rcu_read_lock();
	memcg = mem_cgroup_from_task(rcu_dereference(mm->owner));
	if (unlikely(!memcg))
		goto out;

	switch (idx) {
	case PGFAULT:
		this_cpu_inc(memcg->stat->events[MEM_CGROUP_EVENTS_PGFAULT]);
		break;
	case PGMAJFAULT:
		this_cpu_inc(memcg->stat->events[MEM_CGROUP_EVENTS_PGMAJFAULT]);
		break;
	default:
		BUG();
	}
out:
	rcu_read_unlock();
484
}
485
#ifdef CONFIG_TRANSPARENT_HUGEPAGE
486
void mem_cgroup_split_huge_fixup(struct page *head);
487 488
#endif

A
Andrew Morton 已提交
489
#else /* CONFIG_MEMCG */
490 491
struct mem_cgroup;

492 493 494 495 496 497 498 499 500 501 502 503
static inline void mem_cgroup_events(struct mem_cgroup *memcg,
				     enum mem_cgroup_events_index idx,
				     unsigned int nr)
{
}

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

504 505 506
static inline int mem_cgroup_try_charge(struct page *page, struct mm_struct *mm,
					gfp_t gfp_mask,
					struct mem_cgroup **memcgp)
507
{
508
	*memcgp = NULL;
509 510 511
	return 0;
}

512 513 514
static inline void mem_cgroup_commit_charge(struct page *page,
					    struct mem_cgroup *memcg,
					    bool lrucare)
515 516 517
{
}

518 519
static inline void mem_cgroup_cancel_charge(struct page *page,
					    struct mem_cgroup *memcg)
520 521 522
{
}

523
static inline void mem_cgroup_uncharge(struct page *page)
524 525 526
{
}

527
static inline void mem_cgroup_uncharge_list(struct list_head *page_list)
528 529 530
{
}

531
static inline void mem_cgroup_replace_page(struct page *old, struct page *new)
532 533 534
{
}

535 536
static inline struct lruvec *mem_cgroup_zone_lruvec(struct zone *zone,
						    struct mem_cgroup *memcg)
K
KAMEZAWA Hiroyuki 已提交
537
{
538
	return &zone->lruvec;
K
KAMEZAWA Hiroyuki 已提交
539 540
}

541 542
static inline struct lruvec *mem_cgroup_page_lruvec(struct page *page,
						    struct zone *zone)
543
{
544
	return &zone->lruvec;
545 546
}

547
static inline bool mm_match_cgroup(struct mm_struct *mm,
548
		struct mem_cgroup *memcg)
549
{
550
	return true;
551 552
}

553 554
static inline bool task_in_mem_cgroup(struct task_struct *task,
				      const struct mem_cgroup *memcg)
555
{
556
	return true;
557 558
}

559 560 561 562 563 564 565 566 567 568 569 570 571
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)
{
}

572 573 574 575
static inline bool mem_cgroup_disabled(void)
{
	return true;
}
576

577
static inline bool
578
mem_cgroup_inactive_anon_is_low(struct lruvec *lruvec)
579
{
580
	return true;
581 582
}

583 584 585 586 587
static inline bool mem_cgroup_lruvec_online(struct lruvec *lruvec)
{
	return true;
}

588
static inline unsigned long
589
mem_cgroup_get_lru_size(struct lruvec *lruvec, enum lru_list lru)
590 591 592 593
{
	return 0;
}

594 595 596
static inline void
mem_cgroup_update_lru_size(struct lruvec *lruvec, enum lru_list lru,
			      int increment)
K
KOSAKI Motohiro 已提交
597 598 599
{
}

600 601 602 603 604
static inline void
mem_cgroup_print_oom_info(struct mem_cgroup *memcg, struct task_struct *p)
{
}

605
static inline struct mem_cgroup *mem_cgroup_begin_page_stat(struct page *page)
606
{
607
	return NULL;
608 609
}

610
static inline void mem_cgroup_end_page_stat(struct mem_cgroup *memcg)
611 612 613
{
}

614 615 616 617
static inline void mem_cgroup_handle_over_high(void)
{
}

618
static inline void mem_cgroup_oom_enable(void)
619 620 621
{
}

622
static inline void mem_cgroup_oom_disable(void)
623 624 625
{
}

626 627 628 629 630
static inline bool task_in_memcg_oom(struct task_struct *p)
{
	return false;
}

631
static inline bool mem_cgroup_oom_synchronize(bool wait)
632 633 634 635
{
	return false;
}

636
static inline void mem_cgroup_inc_page_stat(struct mem_cgroup *memcg,
S
Sha Zhengju 已提交
637
					    enum mem_cgroup_stat_index idx)
638 639 640
{
}

641
static inline void mem_cgroup_dec_page_stat(struct mem_cgroup *memcg,
S
Sha Zhengju 已提交
642
					    enum mem_cgroup_stat_index idx)
643 644 645
{
}

646
static inline
647 648 649
unsigned long mem_cgroup_soft_limit_reclaim(struct zone *zone, int order,
					    gfp_t gfp_mask,
					    unsigned long *total_scanned)
650
{
651
	return 0;
652 653
}

654
static inline void mem_cgroup_split_huge_fixup(struct page *head)
655 656 657
{
}

658 659 660 661
static inline
void mem_cgroup_count_vm_event(struct mm_struct *mm, enum vm_event_item idx)
{
}
A
Andrew Morton 已提交
662
#endif /* CONFIG_MEMCG */
663

G
Glauber Costa 已提交
664 665 666 667 668 669
enum {
	UNDER_LIMIT,
	SOFT_LIMIT,
	OVER_LIMIT,
};

670
#ifdef CONFIG_CGROUP_WRITEBACK
T
Tejun Heo 已提交
671

672
struct list_head *mem_cgroup_cgwb_list(struct mem_cgroup *memcg);
T
Tejun Heo 已提交
673
struct wb_domain *mem_cgroup_wb_domain(struct bdi_writeback *wb);
674 675 676
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
677 678 679 680 681 682 683 684

#else	/* CONFIG_CGROUP_WRITEBACK */

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

685
static inline void mem_cgroup_wb_stats(struct bdi_writeback *wb,
686 687
				       unsigned long *pfilepages,
				       unsigned long *pheadroom,
688 689 690 691 692
				       unsigned long *pdirty,
				       unsigned long *pwriteback)
{
}

T
Tejun Heo 已提交
693
#endif	/* CONFIG_CGROUP_WRITEBACK */
694

G
Glauber Costa 已提交
695
struct sock;
696
#if defined(CONFIG_INET) && defined(CONFIG_MEMCG_KMEM)
G
Glauber Costa 已提交
697 698 699 700 701 702 703 704 705
void sock_update_memcg(struct sock *sk);
void sock_release_memcg(struct sock *sk);
#else
static inline void sock_update_memcg(struct sock *sk)
{
}
static inline void sock_release_memcg(struct sock *sk)
{
}
706
#endif /* CONFIG_INET && CONFIG_MEMCG_KMEM */
707 708

#ifdef CONFIG_MEMCG_KMEM
709
extern struct static_key memcg_kmem_enabled_key;
710

711
extern int memcg_nr_cache_ids;
712 713
void memcg_get_cache_ids(void);
void memcg_put_cache_ids(void);
714 715 716 717 718 719

/*
 * 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
 */
720
#define for_each_memcg_cache_index(_idx)	\
721
	for ((_idx) = 0; (_idx) < memcg_nr_cache_ids; (_idx)++)
722

723 724
static inline bool memcg_kmem_enabled(void)
{
725
	return static_key_false(&memcg_kmem_enabled_key);
726 727
}

M
Michal Hocko 已提交
728 729 730 731
static inline bool memcg_kmem_is_active(struct mem_cgroup *memcg)
{
	return memcg->kmem_acct_active;
}
732

733 734 735 736 737 738 739 740 741 742 743
/*
 * In general, we'll do everything in our power to not incur in any overhead
 * for non-memcg users for the kmem functions. Not even a function call, if we
 * can avoid it.
 *
 * Therefore, we'll inline all those functions so that in the best case, we'll
 * see that kmemcg is off for everybody and proceed quickly.  If it is on,
 * we'll still do most of the flag checking inline. We check a lot of
 * conditions, but because they are pretty simple, they are expected to be
 * fast.
 */
744 745
int __memcg_kmem_charge_memcg(struct page *page, gfp_t gfp, int order,
			      struct mem_cgroup *memcg);
746 747
int __memcg_kmem_charge(struct page *page, gfp_t gfp, int order);
void __memcg_kmem_uncharge(struct page *page, int order);
748

M
Michal Hocko 已提交
749 750 751 752 753 754 755 756 757
/*
 * helper for acessing a memcg's index. It will be used as an index in the
 * 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;
}
758

V
Vladimir Davydov 已提交
759
struct kmem_cache *__memcg_kmem_get_cache(struct kmem_cache *cachep, gfp_t gfp);
760
void __memcg_kmem_put_cache(struct kmem_cache *cachep);
761

V
Vladimir Davydov 已提交
762
static inline bool __memcg_kmem_bypass(void)
763 764 765 766 767
{
	if (!memcg_kmem_enabled())
		return true;
	if (in_interrupt() || (!current->mm) || (current->flags & PF_KTHREAD))
		return true;
768
	return false;
769 770 771
}

/**
772 773 774 775
 * memcg_kmem_charge: charge a kmem page
 * @page: page to charge
 * @gfp: reclaim mode
 * @order: allocation order
776
 *
777
 * Returns 0 on success, an error code on failure.
778
 */
779 780
static __always_inline int memcg_kmem_charge(struct page *page,
					     gfp_t gfp, int order)
781
{
V
Vladimir Davydov 已提交
782 783 784
	if (__memcg_kmem_bypass())
		return 0;
	if (!(gfp & __GFP_ACCOUNT))
785 786
		return 0;
	return __memcg_kmem_charge(page, gfp, order);
787 788 789
}

/**
790 791 792
 * memcg_kmem_uncharge: uncharge a kmem page
 * @page: page to uncharge
 * @order: allocation order
793
 */
794
static __always_inline void memcg_kmem_uncharge(struct page *page, int order)
795 796
{
	if (memcg_kmem_enabled())
797
		__memcg_kmem_uncharge(page, order);
798 799
}

800 801 802 803
/**
 * memcg_kmem_get_cache: selects the correct per-memcg cache for allocation
 * @cachep: the original global kmem cache
 *
804
 * All memory allocated from a per-memcg cache is charged to the owner memcg.
805 806 807 808
 */
static __always_inline struct kmem_cache *
memcg_kmem_get_cache(struct kmem_cache *cachep, gfp_t gfp)
{
V
Vladimir Davydov 已提交
809
	if (__memcg_kmem_bypass())
810
		return cachep;
V
Vladimir Davydov 已提交
811
	return __memcg_kmem_get_cache(cachep, gfp);
812
}
813 814 815 816 817 818

static __always_inline void memcg_kmem_put_cache(struct kmem_cache *cachep)
{
	if (memcg_kmem_enabled())
		__memcg_kmem_put_cache(cachep);
}
819
#else
820 821 822
#define for_each_memcg_cache_index(_idx)	\
	for (; NULL; )

823 824 825 826 827
static inline bool memcg_kmem_enabled(void)
{
	return false;
}

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

833
static inline int memcg_kmem_charge(struct page *page, gfp_t gfp, int order)
834
{
835
	return 0;
836 837
}

838
static inline void memcg_kmem_uncharge(struct page *page, int order)
839 840
{
}
841 842 843 844 845 846

static inline int memcg_cache_id(struct mem_cgroup *memcg)
{
	return -1;
}

847 848 849 850 851 852 853 854
static inline void memcg_get_cache_ids(void)
{
}

static inline void memcg_put_cache_ids(void)
{
}

855 856 857 858 859
static inline struct kmem_cache *
memcg_kmem_get_cache(struct kmem_cache *cachep, gfp_t gfp)
{
	return cachep;
}
860 861 862 863

static inline void memcg_kmem_put_cache(struct kmem_cache *cachep)
{
}
864
#endif /* CONFIG_MEMCG_KMEM */
B
Balbir Singh 已提交
865
#endif /* _LINUX_MEMCONTROL_H */