Blame view
mm/memcontrol.c
150 KB
8cdea7c05 Memory controller... |
1 2 3 4 5 |
/* memcontrol.c - Memory Controller * * Copyright IBM Corporation, 2007 * Author Balbir Singh <balbir@linux.vnet.ibm.com> * |
78fb74669 Memory controller... |
6 7 8 |
* Copyright 2007 OpenVZ SWsoft Inc * Author: Pavel Emelianov <xemul@openvz.org> * |
2e72b6347 memcg: implement ... |
9 10 11 12 |
* Memory thresholds * Copyright (C) 2009 Nokia Corporation * Author: Kirill A. Shutemov * |
7ae1e1d0f memcg: kmem contr... |
13 14 15 16 |
* Kernel Memory Controller * Copyright (C) 2012 Parallels Inc. and Google Inc. * Authors: Glauber Costa and Suleiman Souhlal * |
1575e68b3 mm: memcontrol: u... |
17 18 19 20 21 22 |
* Native page reclaim * Charge lifetime sanitation * Lockless page tracking & accounting * Unified hierarchy configuration model * Copyright (C) 2015 Red Hat, Inc., Johannes Weiner * |
8cdea7c05 Memory controller... |
23 24 25 26 27 28 29 30 31 32 |
* 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. */ |
3e32cb2e0 mm: memcontrol: l... |
33 |
#include <linux/page_counter.h> |
8cdea7c05 Memory controller... |
34 35 |
#include <linux/memcontrol.h> #include <linux/cgroup.h> |
78fb74669 Memory controller... |
36 |
#include <linux/mm.h> |
4ffef5fef memcg: move charg... |
37 |
#include <linux/hugetlb.h> |
d13d14430 memcg: handle swa... |
38 |
#include <linux/pagemap.h> |
d52aa412d memory cgroup enh... |
39 |
#include <linux/smp.h> |
8a9f3ccd2 Memory controller... |
40 |
#include <linux/page-flags.h> |
66e1707bc Memory controller... |
41 |
#include <linux/backing-dev.h> |
8a9f3ccd2 Memory controller... |
42 43 |
#include <linux/bit_spinlock.h> #include <linux/rcupdate.h> |
e222432bf memcg: show memcg... |
44 |
#include <linux/limits.h> |
b9e15bafd mm: Add export.h ... |
45 |
#include <linux/export.h> |
8c7c6e34a memcg: mem+swap c... |
46 |
#include <linux/mutex.h> |
bb4cc1a8b revert "memcg: ge... |
47 |
#include <linux/rbtree.h> |
b6ac57d50 memcgroup: move m... |
48 |
#include <linux/slab.h> |
66e1707bc Memory controller... |
49 |
#include <linux/swap.h> |
024914477 memcg: move charg... |
50 |
#include <linux/swapops.h> |
66e1707bc Memory controller... |
51 |
#include <linux/spinlock.h> |
2e72b6347 memcg: implement ... |
52 |
#include <linux/eventfd.h> |
79bd9814e cgroup, memcg: mo... |
53 |
#include <linux/poll.h> |
2e72b6347 memcg: implement ... |
54 |
#include <linux/sort.h> |
66e1707bc Memory controller... |
55 |
#include <linux/fs.h> |
d2ceb9b7d memory cgroup enh... |
56 |
#include <linux/seq_file.h> |
70ddf637e memcg: add memory... |
57 |
#include <linux/vmpressure.h> |
b69408e88 vmscan: Use an in... |
58 |
#include <linux/mm_inline.h> |
5d1ea48bd mm: page_cgroup: ... |
59 |
#include <linux/swap_cgroup.h> |
cdec2e426 memcg: coalesce c... |
60 |
#include <linux/cpu.h> |
158e0a2d1 memcg: use find_l... |
61 |
#include <linux/oom.h> |
0056f4e66 mm: memcg: lockde... |
62 |
#include <linux/lockdep.h> |
79bd9814e cgroup, memcg: mo... |
63 |
#include <linux/file.h> |
08e552c69 memcg: synchroniz... |
64 |
#include "internal.h" |
d1a4c0b37 tcp memory pressu... |
65 |
#include <net/sock.h> |
4bd2c1ee4 memcg: cleanup km... |
66 |
#include <net/ip.h> |
d1a4c0b37 tcp memory pressu... |
67 |
#include <net/tcp_memcontrol.h> |
f35c3a8ee memcg, kmem: use ... |
68 |
#include "slab.h" |
8cdea7c05 Memory controller... |
69 |
|
8697d3319 Memory controller... |
70 |
#include <asm/uaccess.h> |
cc8e970c3 memcg: add mm_vms... |
71 |
#include <trace/events/vmscan.h> |
073219e99 cgroup: clean up ... |
72 73 |
struct cgroup_subsys memory_cgrp_subsys __read_mostly; EXPORT_SYMBOL(memory_cgrp_subsys); |
68ae564bb mm, memcg: avoid ... |
74 |
|
a181b0e88 memcg: make globa... |
75 |
#define MEM_CGROUP_RECLAIM_RETRIES 5 |
6bbda35ce memcg: mark more ... |
76 |
static struct mem_cgroup *root_mem_cgroup __read_mostly; |
8cdea7c05 Memory controller... |
77 |
|
21afa38ee mm: memcontrol: c... |
78 |
/* Whether the swap controller is active */ |
c255a4580 memcg: rename con... |
79 |
#ifdef CONFIG_MEMCG_SWAP |
c077719be memcg: mem+swap c... |
80 |
int do_swap_account __read_mostly; |
c077719be memcg: mem+swap c... |
81 |
#else |
a0db00fcf memcg: remove red... |
82 |
#define do_swap_account 0 |
c077719be memcg: mem+swap c... |
83 |
#endif |
af7c4b0ec mm: memcg: print ... |
84 85 86 |
static const char * const mem_cgroup_stat_names[] = { "cache", "rss", |
b070e65c0 mm, memcg: add rs... |
87 |
"rss_huge", |
af7c4b0ec mm: memcg: print ... |
88 |
"mapped_file", |
3ea67d06e memcg: add per cg... |
89 |
"writeback", |
af7c4b0ec mm: memcg: print ... |
90 91 |
"swap", }; |
af7c4b0ec mm: memcg: print ... |
92 93 94 95 96 97 |
static const char * const mem_cgroup_events_names[] = { "pgpgin", "pgpgout", "pgfault", "pgmajfault", }; |
58cf188ed memcg, oom: provi... |
98 99 100 101 102 103 104 |
static const char * const mem_cgroup_lru_names[] = { "inactive_anon", "active_anon", "inactive_file", "active_file", "unevictable", }; |
7a159cc9d memcg: use native... |
105 106 107 108 109 110 111 112 |
/* * 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, |
bb4cc1a8b revert "memcg: ge... |
113 |
MEM_CGROUP_TARGET_SOFTLIMIT, |
453a9bf34 memcg: fix numa s... |
114 |
MEM_CGROUP_TARGET_NUMAINFO, |
7a159cc9d memcg: use native... |
115 116 |
MEM_CGROUP_NTARGETS, }; |
a0db00fcf memcg: remove red... |
117 118 119 |
#define THRESHOLDS_EVENTS_TARGET 128 #define SOFTLIMIT_EVENTS_TARGET 1024 #define NUMAINFO_EVENTS_TARGET 1024 |
e9f8974f2 memcg: break out ... |
120 |
|
d52aa412d memory cgroup enh... |
121 |
struct mem_cgroup_stat_cpu { |
7a159cc9d memcg: use native... |
122 |
long count[MEM_CGROUP_STAT_NSTATS]; |
241994ed8 mm: memcontrol: d... |
123 |
unsigned long events[MEMCG_NR_EVENTS]; |
13114716c mm: memcg: keep r... |
124 |
unsigned long nr_page_events; |
7a159cc9d memcg: use native... |
125 |
unsigned long targets[MEM_CGROUP_NTARGETS]; |
d52aa412d memory cgroup enh... |
126 |
}; |
5ac8fb31a mm: memcontrol: c... |
127 128 |
struct reclaim_iter { struct mem_cgroup *position; |
527a5ec9a mm: memcg: per-pr... |
129 130 131 |
/* scan generation, increased every round-trip */ unsigned int generation; }; |
d52aa412d memory cgroup enh... |
132 |
/* |
6d12e2d8d per-zone and recl... |
133 134 |
* per-zone information in memory controller. */ |
6d12e2d8d per-zone and recl... |
135 |
struct mem_cgroup_per_zone { |
6290df545 mm: collect LRU l... |
136 |
struct lruvec lruvec; |
1eb492725 memcg: lru_size i... |
137 |
unsigned long lru_size[NR_LRU_LISTS]; |
3e2f41f1f memcg: add zone_r... |
138 |
|
5ac8fb31a mm: memcontrol: c... |
139 |
struct reclaim_iter iter[DEF_PRIORITY + 1]; |
527a5ec9a mm: memcg: per-pr... |
140 |
|
bb4cc1a8b revert "memcg: ge... |
141 |
struct rb_node tree_node; /* RB tree node */ |
3e32cb2e0 mm: memcontrol: l... |
142 |
unsigned long usage_in_excess;/* Set to the value by which */ |
bb4cc1a8b revert "memcg: ge... |
143 144 |
/* the soft limit is exceeded*/ bool on_tree; |
d79154bb5 memcg: replace me... |
145 |
struct mem_cgroup *memcg; /* Back pointer, we cannot */ |
4e4169535 memory controller... |
146 |
/* use container_of */ |
6d12e2d8d per-zone and recl... |
147 |
}; |
6d12e2d8d per-zone and recl... |
148 149 150 151 |
struct mem_cgroup_per_node { struct mem_cgroup_per_zone zoneinfo[MAX_NR_ZONES]; }; |
bb4cc1a8b revert "memcg: ge... |
152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 |
/* * Cgroups above their limits are maintained in a RB-Tree, independent of * their hierarchy representation */ struct mem_cgroup_tree_per_zone { struct rb_root rb_root; spinlock_t lock; }; struct mem_cgroup_tree_per_node { struct mem_cgroup_tree_per_zone rb_tree_per_zone[MAX_NR_ZONES]; }; struct mem_cgroup_tree { struct mem_cgroup_tree_per_node *rb_tree_per_node[MAX_NUMNODES]; }; static struct mem_cgroup_tree soft_limit_tree __read_mostly; |
2e72b6347 memcg: implement ... |
171 172 |
struct mem_cgroup_threshold { struct eventfd_ctx *eventfd; |
3e32cb2e0 mm: memcontrol: l... |
173 |
unsigned long threshold; |
2e72b6347 memcg: implement ... |
174 |
}; |
9490ff275 memcg: oom notifier |
175 |
/* For threshold */ |
2e72b6347 memcg: implement ... |
176 |
struct mem_cgroup_threshold_ary { |
748dad36d memcg: make thres... |
177 |
/* An array index points to threshold just below or equal to usage. */ |
5407a5625 mm: remove unnece... |
178 |
int current_threshold; |
2e72b6347 memcg: implement ... |
179 180 181 182 183 |
/* Size of entries[] */ unsigned int size; /* Array of thresholds */ struct mem_cgroup_threshold entries[0]; }; |
2c488db27 memcg: clean up m... |
184 185 186 187 188 189 190 191 192 193 194 |
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; }; |
9490ff275 memcg: oom notifier |
195 196 197 198 199 |
/* for OOM */ struct mem_cgroup_eventfd_list { struct list_head list; struct eventfd_ctx *eventfd; }; |
2e72b6347 memcg: implement ... |
200 |
|
79bd9814e cgroup, memcg: mo... |
201 202 203 |
/* * cgroup_event represents events which userspace want to receive. */ |
3bc942f37 memcg: rename cgr... |
204 |
struct mem_cgroup_event { |
79bd9814e cgroup, memcg: mo... |
205 |
/* |
59b6f8734 memcg: make cgrou... |
206 |
* memcg which the event belongs to. |
79bd9814e cgroup, memcg: mo... |
207 |
*/ |
59b6f8734 memcg: make cgrou... |
208 |
struct mem_cgroup *memcg; |
79bd9814e cgroup, memcg: mo... |
209 |
/* |
79bd9814e cgroup, memcg: mo... |
210 211 212 213 214 215 216 217 |
* eventfd to signal userspace about the event. */ struct eventfd_ctx *eventfd; /* * Each of these stored in a list by the cgroup. */ struct list_head list; /* |
fba948078 cgroup, memcg: mo... |
218 219 220 221 |
* register_event() callback will be used to add new userspace * waiter for changes related to this event. Use eventfd_signal() * on eventfd to send notification to userspace. */ |
59b6f8734 memcg: make cgrou... |
222 |
int (*register_event)(struct mem_cgroup *memcg, |
347c4a874 memcg: remove cgr... |
223 |
struct eventfd_ctx *eventfd, const char *args); |
fba948078 cgroup, memcg: mo... |
224 225 226 227 228 |
/* * unregister_event() callback will be called when userspace closes * the eventfd or on cgroup removing. This callback must be set, * if you want provide notification functionality. */ |
59b6f8734 memcg: make cgrou... |
229 |
void (*unregister_event)(struct mem_cgroup *memcg, |
fba948078 cgroup, memcg: mo... |
230 231 |
struct eventfd_ctx *eventfd); /* |
79bd9814e cgroup, memcg: mo... |
232 233 234 235 236 237 238 239 |
* All fields below needed to unregister event when * userspace closes eventfd. */ poll_table pt; wait_queue_head_t *wqh; wait_queue_t wait; struct work_struct remove; }; |
c0ff4b854 memcg: rename mem... |
240 241 |
static void mem_cgroup_threshold(struct mem_cgroup *memcg); static void mem_cgroup_oom_notify(struct mem_cgroup *memcg); |
2e72b6347 memcg: implement ... |
242 |
|
f64c3f549 memory controller... |
243 |
/* |
8cdea7c05 Memory controller... |
244 245 246 247 |
* 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. |
8cdea7c05 Memory controller... |
248 249 250 |
*/ struct mem_cgroup { struct cgroup_subsys_state css; |
3e32cb2e0 mm: memcontrol: l... |
251 252 253 254 255 |
/* Accounted resources */ struct page_counter memory; struct page_counter memsw; struct page_counter kmem; |
241994ed8 mm: memcontrol: d... |
256 257 258 |
/* Normal memory consumption range */ unsigned long low; unsigned long high; |
3e32cb2e0 mm: memcontrol: l... |
259 |
unsigned long soft_limit; |
59927fb98 memcg: free mem_c... |
260 |
|
70ddf637e memcg: add memory... |
261 262 |
/* vmpressure notifications */ struct vmpressure vmpressure; |
2f7dd7a41 mm: memcontrol: d... |
263 264 |
/* css_online() has been completed */ int initialized; |
465939a1f memcg: don't need... |
265 |
/* |
18f59ea7d memcg: memory cgr... |
266 267 268 |
* Should the accounting and control be hierarchical, per subtree? */ bool use_hierarchy; |
79dfdaccd memcg: make oom_l... |
269 270 271 |
bool oom_lock; atomic_t under_oom; |
3812c8c8f mm: memcg: do not... |
272 |
atomic_t oom_wakeups; |
79dfdaccd memcg: make oom_l... |
273 |
|
1f4c025b5 memcg: export mem... |
274 |
int swappiness; |
3c11ecf44 memcg: oom kill d... |
275 276 |
/* OOM-Killer disable */ int oom_kill_disable; |
a7885eb8a memcg: swappiness |
277 |
|
2e72b6347 memcg: implement ... |
278 279 280 281 |
/* protect arrays of thresholds */ struct mutex thresholds_lock; /* thresholds for memory usage. RCU-protected */ |
2c488db27 memcg: clean up m... |
282 |
struct mem_cgroup_thresholds thresholds; |
907860ed3 cgroups: make cft... |
283 |
|
2e72b6347 memcg: implement ... |
284 |
/* thresholds for mem+swap usage. RCU-protected */ |
2c488db27 memcg: clean up m... |
285 |
struct mem_cgroup_thresholds memsw_thresholds; |
907860ed3 cgroups: make cft... |
286 |
|
9490ff275 memcg: oom notifier |
287 288 |
/* For oom notifier event fd */ struct list_head oom_notify; |
185efc0f9 memcg: Revert "me... |
289 |
|
d52aa412d memory cgroup enh... |
290 |
/* |
7dc74be03 memcg: add interf... |
291 292 293 |
* Should we move charges of a task when a task is moved into this * mem_cgroup ? And what type of charges should we move ? */ |
f894ffa86 memcg: trivial cl... |
294 |
unsigned long move_charge_at_immigrate; |
7dc74be03 memcg: add interf... |
295 |
/* |
619d094b5 memcg: simplify m... |
296 297 |
* set > 0 if pages under this cgroup are moving to other cgroup. */ |
6de226191 mm: memcontrol: t... |
298 |
atomic_t moving_account; |
312734c04 memcg: remove PCG... |
299 |
/* taken only while moving_account > 0 */ |
6de226191 mm: memcontrol: t... |
300 301 302 |
spinlock_t move_lock; struct task_struct *move_lock_task; unsigned long move_lock_flags; |
619d094b5 memcg: simplify m... |
303 |
/* |
c62b1a3b3 memcg: use generi... |
304 |
* percpu counter. |
d52aa412d memory cgroup enh... |
305 |
*/ |
3a7951b4c memcg: mark stat ... |
306 |
struct mem_cgroup_stat_cpu __percpu *stat; |
711d3d2c9 memcg: cpu hotplu... |
307 308 309 310 311 312 |
/* * used when a cpu is offlined or other synchronizations * See mem_cgroup_read_stat(). */ struct mem_cgroup_stat_cpu nocpu_base; spinlock_t pcp_counter_lock; |
d1a4c0b37 tcp memory pressu... |
313 |
|
4bd2c1ee4 memcg: cleanup km... |
314 |
#if defined(CONFIG_MEMCG_KMEM) && defined(CONFIG_INET) |
2e685cad5 tcp_memcontrol: K... |
315 |
struct cg_proto tcp_mem; |
d1a4c0b37 tcp memory pressu... |
316 |
#endif |
2633d7a02 slab/slub: consid... |
317 |
#if defined(CONFIG_MEMCG_KMEM) |
f7ce3190c slab: embed memcg... |
318 |
/* Index in the kmem_cache->memcg_params.memcg_caches array */ |
2633d7a02 slab/slub: consid... |
319 |
int kmemcg_id; |
2788cf0c4 memcg: reparent l... |
320 |
bool kmem_acct_activated; |
2a4db7eb9 memcg: free memcg... |
321 |
bool kmem_acct_active; |
2633d7a02 slab/slub: consid... |
322 |
#endif |
45cf7ebd5 memcg: reduce the... |
323 324 325 326 327 328 329 |
int last_scanned_node; #if MAX_NUMNODES > 1 nodemask_t scan_nodes; atomic_t numainfo_events; atomic_t numainfo_updating; #endif |
70ddf637e memcg: add memory... |
330 |
|
fba948078 cgroup, memcg: mo... |
331 332 333 |
/* List of events which userspace want to receive */ struct list_head event_list; spinlock_t event_list_lock; |
54f72fe02 memcg: clean up m... |
334 335 |
struct mem_cgroup_per_node *nodeinfo[0]; /* WARNING: nodeinfo must be the last member here */ |
8cdea7c05 Memory controller... |
336 |
}; |
510fc4e11 memcg: kmem accou... |
337 |
#ifdef CONFIG_MEMCG_KMEM |
cb731d6c6 vmscan: per memor... |
338 |
bool memcg_kmem_is_active(struct mem_cgroup *memcg) |
7de37682b memcg: kmem accou... |
339 |
{ |
2a4db7eb9 memcg: free memcg... |
340 |
return memcg->kmem_acct_active; |
7de37682b memcg: kmem accou... |
341 |
} |
510fc4e11 memcg: kmem accou... |
342 |
#endif |
7dc74be03 memcg: add interf... |
343 344 |
/* Stuffs for move charges at task migration. */ /* |
1dfab5abc mm: memcontrol: f... |
345 |
* Types of charges to be moved. |
7dc74be03 memcg: add interf... |
346 |
*/ |
1dfab5abc mm: memcontrol: f... |
347 348 349 |
#define MOVE_ANON 0x1U #define MOVE_FILE 0x2U #define MOVE_MASK (MOVE_ANON | MOVE_FILE) |
7dc74be03 memcg: add interf... |
350 |
|
4ffef5fef memcg: move charg... |
351 352 |
/* "mc" and its members are protected by cgroup_mutex */ static struct move_charge_struct { |
b1dd693e5 memcg: avoid dead... |
353 |
spinlock_t lock; /* for from, to */ |
4ffef5fef memcg: move charg... |
354 355 |
struct mem_cgroup *from; struct mem_cgroup *to; |
1dfab5abc mm: memcontrol: f... |
356 |
unsigned long flags; |
4ffef5fef memcg: move charg... |
357 |
unsigned long precharge; |
854ffa8d1 memcg: improve pe... |
358 |
unsigned long moved_charge; |
483c30b51 memcg: improve pe... |
359 |
unsigned long moved_swap; |
8033b97c9 memcg: avoid oom ... |
360 361 362 |
struct task_struct *moving_task; /* a task moving charges */ wait_queue_head_t waitq; /* a waitq for other context */ } mc = { |
2bd9bb206 memcg: clean up w... |
363 |
.lock = __SPIN_LOCK_UNLOCKED(mc.lock), |
8033b97c9 memcg: avoid oom ... |
364 365 |
.waitq = __WAIT_QUEUE_HEAD_INITIALIZER(mc.waitq), }; |
4ffef5fef memcg: move charg... |
366 |
|
4e4169535 memory controller... |
367 368 369 370 |
/* * Maximum loops in mem_cgroup_hierarchical_reclaim(), used for soft * limit reclaim to prevent infinite loops, if they ever occur. */ |
a0db00fcf memcg: remove red... |
371 |
#define MEM_CGROUP_MAX_RECLAIM_LOOPS 100 |
bb4cc1a8b revert "memcg: ge... |
372 |
#define MEM_CGROUP_MAX_SOFT_LIMIT_RECLAIM_LOOPS 2 |
4e4169535 memory controller... |
373 |
|
217bc3194 memory cgroup enh... |
374 375 |
enum charge_type { MEM_CGROUP_CHARGE_TYPE_CACHE = 0, |
41326c17f memcg: rename MEM... |
376 |
MEM_CGROUP_CHARGE_TYPE_ANON, |
d13d14430 memcg: handle swa... |
377 |
MEM_CGROUP_CHARGE_TYPE_SWAPOUT, /* for accounting swapcache */ |
8a9478ca7 memcg: fix swap a... |
378 |
MEM_CGROUP_CHARGE_TYPE_DROP, /* a page was unused swap cache */ |
c05555b57 memcg: atomic ops... |
379 380 |
NR_CHARGE_TYPE, }; |
8c7c6e34a memcg: mem+swap c... |
381 |
/* for encoding cft->private value on file */ |
86ae53e1a memcg: change def... |
382 383 384 385 |
enum res_type { _MEM, _MEMSWAP, _OOM_TYPE, |
510fc4e11 memcg: kmem accou... |
386 |
_KMEM, |
86ae53e1a memcg: change def... |
387 |
}; |
a0db00fcf memcg: remove red... |
388 389 |
#define MEMFILE_PRIVATE(x, val) ((x) << 16 | (val)) #define MEMFILE_TYPE(val) ((val) >> 16 & 0xffff) |
8c7c6e34a memcg: mem+swap c... |
390 |
#define MEMFILE_ATTR(val) ((val) & 0xffff) |
9490ff275 memcg: oom notifier |
391 392 |
/* Used for OOM nofiier */ #define OOM_CONTROL (0) |
8c7c6e34a memcg: mem+swap c... |
393 |
|
75822b449 memory controller... |
394 |
/* |
0999821b1 memcg: replace cg... |
395 396 397 398 399 |
* The memcg_create_mutex will be held whenever a new cgroup is created. * As a consequence, any change that needs to protect against new child cgroups * appearing has to hold it as well. */ static DEFINE_MUTEX(memcg_create_mutex); |
b21451459 memcg: add mem_cg... |
400 401 |
struct mem_cgroup *mem_cgroup_from_css(struct cgroup_subsys_state *s) { |
a7c6d554a cgroup: add/updat... |
402 |
return s ? container_of(s, struct mem_cgroup, css) : NULL; |
b21451459 memcg: add mem_cg... |
403 |
} |
70ddf637e memcg: add memory... |
404 405 406 407 408 409 410 411 412 413 414 415 |
/* Some nice accessors for the vmpressure. */ struct vmpressure *memcg_to_vmpressure(struct mem_cgroup *memcg) { if (!memcg) memcg = root_mem_cgroup; return &memcg->vmpressure; } struct cgroup_subsys_state *vmpressure_to_css(struct vmpressure *vmpr) { return &container_of(vmpr, struct mem_cgroup, vmpressure)->css; } |
7ffc0edc4 memcg: move mem_c... |
416 417 418 419 |
static inline bool mem_cgroup_is_root(struct mem_cgroup *memcg) { return (memcg == root_mem_cgroup); } |
4219b2da2 memcg: fail to cr... |
420 421 422 423 424 |
/* * We restrict the id in the range of [1, 65535], so it can fit into * an unsigned short. */ #define MEM_CGROUP_ID_MAX USHRT_MAX |
34c00c319 memcg: convert to... |
425 426 |
static inline unsigned short mem_cgroup_id(struct mem_cgroup *memcg) { |
15a4c835e cgroup, memcg: im... |
427 |
return memcg->css.id; |
34c00c319 memcg: convert to... |
428 |
} |
adbe427b9 memcg: zap mem_cg... |
429 430 431 432 433 434 |
/* * A helper function to get mem_cgroup from ID. must be called under * rcu_read_lock(). The caller is responsible for calling * css_tryget_online() if the mem_cgroup is used for charging. (dropping * refcnt from swap can be called against removed memcg.) */ |
34c00c319 memcg: convert to... |
435 436 437 |
static inline struct mem_cgroup *mem_cgroup_from_id(unsigned short id) { struct cgroup_subsys_state *css; |
7d699ddb2 cgroup, memcg: al... |
438 |
css = css_from_id(id, &memory_cgrp_subsys); |
34c00c319 memcg: convert to... |
439 440 |
return mem_cgroup_from_css(css); } |
e1aab161e socket: initial c... |
441 |
/* Writing them here to avoid exposing memcg's inner layout */ |
4bd2c1ee4 memcg: cleanup km... |
442 |
#if defined(CONFIG_INET) && defined(CONFIG_MEMCG_KMEM) |
e1aab161e socket: initial c... |
443 |
|
e1aab161e socket: initial c... |
444 445 |
void sock_update_memcg(struct sock *sk) { |
376be5ff8 net: fix socket m... |
446 |
if (mem_cgroup_sockets_enabled) { |
e1aab161e socket: initial c... |
447 |
struct mem_cgroup *memcg; |
3f1346193 memcg: decrement ... |
448 |
struct cg_proto *cg_proto; |
e1aab161e socket: initial c... |
449 450 |
BUG_ON(!sk->sk_prot->proto_cgroup); |
f3f511e1c net: fix sock_clo... |
451 452 453 454 455 456 457 458 459 460 |
/* Socket cloning can throw us here with sk_cgrp already * filled. It won't however, necessarily happen from * process context. So the test for root memcg given * the current task's memcg won't help us in this case. * * Respecting the original socket's memcg is a better * decision in this case. */ if (sk->sk_cgrp) { BUG_ON(mem_cgroup_is_root(sk->sk_cgrp->memcg)); |
5347e5ae1 memcg: use css_ge... |
461 |
css_get(&sk->sk_cgrp->memcg->css); |
f3f511e1c net: fix sock_clo... |
462 463 |
return; } |
e1aab161e socket: initial c... |
464 465 |
rcu_read_lock(); memcg = mem_cgroup_from_task(current); |
3f1346193 memcg: decrement ... |
466 |
cg_proto = sk->sk_prot->proto_cgroup(memcg); |
5347e5ae1 memcg: use css_ge... |
467 |
if (!mem_cgroup_is_root(memcg) && |
ec903c0c8 cgroup: rename cs... |
468 469 |
memcg_proto_active(cg_proto) && css_tryget_online(&memcg->css)) { |
3f1346193 memcg: decrement ... |
470 |
sk->sk_cgrp = cg_proto; |
e1aab161e socket: initial c... |
471 472 473 474 475 476 477 478 |
} rcu_read_unlock(); } } EXPORT_SYMBOL(sock_update_memcg); void sock_release_memcg(struct sock *sk) { |
376be5ff8 net: fix socket m... |
479 |
if (mem_cgroup_sockets_enabled && sk->sk_cgrp) { |
e1aab161e socket: initial c... |
480 481 482 |
struct mem_cgroup *memcg; WARN_ON(!sk->sk_cgrp->memcg); memcg = sk->sk_cgrp->memcg; |
5347e5ae1 memcg: use css_ge... |
483 |
css_put(&sk->sk_cgrp->memcg->css); |
e1aab161e socket: initial c... |
484 485 |
} } |
d1a4c0b37 tcp memory pressu... |
486 487 488 489 490 |
struct cg_proto *tcp_proto_cgroup(struct mem_cgroup *memcg) { if (!memcg || mem_cgroup_is_root(memcg)) return NULL; |
2e685cad5 tcp_memcontrol: K... |
491 |
return &memcg->tcp_mem; |
d1a4c0b37 tcp memory pressu... |
492 493 |
} EXPORT_SYMBOL(tcp_proto_cgroup); |
e1aab161e socket: initial c... |
494 |
|
3f1346193 memcg: decrement ... |
495 |
#endif |
a8964b9b8 memcg: use static... |
496 |
#ifdef CONFIG_MEMCG_KMEM |
55007d849 memcg: allocate m... |
497 |
/* |
f7ce3190c slab: embed memcg... |
498 |
* This will be the memcg's index in each cache's ->memcg_params.memcg_caches. |
b86278359 memcg: stop using... |
499 500 501 502 503 |
* The main reason for not using cgroup id for this: * this works better in sparse environments, where we have a lot of memcgs, * but only a few kmem-limited. Or also, if we have, for instance, 200 * memcgs, and none but the 200th is kmem-limited, we'd have to have a * 200 entry array for that. |
55007d849 memcg: allocate m... |
504 |
* |
dbcf73e26 memcg: rename som... |
505 506 |
* The current size of the caches array is stored in memcg_nr_cache_ids. It * will double each time we have to increase it. |
55007d849 memcg: allocate m... |
507 |
*/ |
dbcf73e26 memcg: rename som... |
508 509 |
static DEFINE_IDA(memcg_cache_ida); int memcg_nr_cache_ids; |
749c54151 memcg: aggregate ... |
510 |
|
05257a1a3 memcg: add rwsem ... |
511 512 513 514 515 516 517 518 519 520 521 522 |
/* Protects memcg_nr_cache_ids */ static DECLARE_RWSEM(memcg_cache_ids_sem); void memcg_get_cache_ids(void) { down_read(&memcg_cache_ids_sem); } void memcg_put_cache_ids(void) { up_read(&memcg_cache_ids_sem); } |
55007d849 memcg: allocate m... |
523 524 525 526 527 528 |
/* * MIN_SIZE is different than 1, because we would like to avoid going through * the alloc/free process all the time. In a small machine, 4 kmem-limited * cgroups is a reasonable guess. In the future, it could be a parameter or * tunable, but that is strictly not necessary. * |
b86278359 memcg: stop using... |
529 |
* MAX_SIZE should be as large as the number of cgrp_ids. Ideally, we could get |
55007d849 memcg: allocate m... |
530 531 |
* this constant directly from cgroup, but it is understandable that this is * better kept as an internal representation in cgroup.c. In any case, the |
b86278359 memcg: stop using... |
532 |
* cgrp_id space is not getting any smaller, and we don't have to necessarily |
55007d849 memcg: allocate m... |
533 534 535 |
* increase ours as well if it increases. */ #define MEMCG_CACHES_MIN_SIZE 4 |
b86278359 memcg: stop using... |
536 |
#define MEMCG_CACHES_MAX_SIZE MEM_CGROUP_ID_MAX |
55007d849 memcg: allocate m... |
537 |
|
d7f25f8a2 memcg: infrastruc... |
538 539 540 541 542 543 |
/* * A lot of the calls to the cache allocation functions are expected to be * inlined by the compiler. Since the calls to memcg_kmem_get_cache are * conditional to this static branch, we'll have to allow modules that does * kmem_cache_alloc and the such to see this symbol as well */ |
a8964b9b8 memcg: use static... |
544 |
struct static_key memcg_kmem_enabled_key; |
d7f25f8a2 memcg: infrastruc... |
545 |
EXPORT_SYMBOL(memcg_kmem_enabled_key); |
a8964b9b8 memcg: use static... |
546 |
|
a8964b9b8 memcg: use static... |
547 |
#endif /* CONFIG_MEMCG_KMEM */ |
f64c3f549 memory controller... |
548 |
static struct mem_cgroup_per_zone * |
e231875ba mm: memcontrol: c... |
549 |
mem_cgroup_zone_zoneinfo(struct mem_cgroup *memcg, struct zone *zone) |
f64c3f549 memory controller... |
550 |
{ |
e231875ba mm: memcontrol: c... |
551 552 |
int nid = zone_to_nid(zone); int zid = zone_idx(zone); |
54f72fe02 memcg: clean up m... |
553 |
return &memcg->nodeinfo[nid]->zoneinfo[zid]; |
f64c3f549 memory controller... |
554 |
} |
c0ff4b854 memcg: rename mem... |
555 |
struct cgroup_subsys_state *mem_cgroup_css(struct mem_cgroup *memcg) |
d324236b3 memcg: add access... |
556 |
{ |
c0ff4b854 memcg: rename mem... |
557 |
return &memcg->css; |
d324236b3 memcg: add access... |
558 |
} |
f64c3f549 memory controller... |
559 |
static struct mem_cgroup_per_zone * |
e231875ba mm: memcontrol: c... |
560 |
mem_cgroup_page_zoneinfo(struct mem_cgroup *memcg, struct page *page) |
f64c3f549 memory controller... |
561 |
{ |
97a6c37b3 memcg: change pag... |
562 563 |
int nid = page_to_nid(page); int zid = page_zonenum(page); |
f64c3f549 memory controller... |
564 |
|
e231875ba mm: memcontrol: c... |
565 |
return &memcg->nodeinfo[nid]->zoneinfo[zid]; |
f64c3f549 memory controller... |
566 |
} |
bb4cc1a8b revert "memcg: ge... |
567 568 569 570 571 572 573 574 575 576 577 578 579 580 |
static struct mem_cgroup_tree_per_zone * soft_limit_tree_node_zone(int nid, int zid) { return &soft_limit_tree.rb_tree_per_node[nid]->rb_tree_per_zone[zid]; } static struct mem_cgroup_tree_per_zone * soft_limit_tree_from_page(struct page *page) { int nid = page_to_nid(page); int zid = page_zonenum(page); return &soft_limit_tree.rb_tree_per_node[nid]->rb_tree_per_zone[zid]; } |
cf2c81279 mm: memcontrol: r... |
581 582 |
static void __mem_cgroup_insert_exceeded(struct mem_cgroup_per_zone *mz, struct mem_cgroup_tree_per_zone *mctz, |
3e32cb2e0 mm: memcontrol: l... |
583 |
unsigned long new_usage_in_excess) |
bb4cc1a8b revert "memcg: ge... |
584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 |
{ struct rb_node **p = &mctz->rb_root.rb_node; struct rb_node *parent = NULL; struct mem_cgroup_per_zone *mz_node; if (mz->on_tree) return; mz->usage_in_excess = new_usage_in_excess; if (!mz->usage_in_excess) return; while (*p) { parent = *p; mz_node = rb_entry(parent, struct mem_cgroup_per_zone, tree_node); if (mz->usage_in_excess < mz_node->usage_in_excess) p = &(*p)->rb_left; /* * We can't avoid mem cgroups that are over their soft * limit by the same amount */ else if (mz->usage_in_excess >= mz_node->usage_in_excess) p = &(*p)->rb_right; } rb_link_node(&mz->tree_node, parent, p); rb_insert_color(&mz->tree_node, &mctz->rb_root); mz->on_tree = true; } |
cf2c81279 mm: memcontrol: r... |
612 613 |
static void __mem_cgroup_remove_exceeded(struct mem_cgroup_per_zone *mz, struct mem_cgroup_tree_per_zone *mctz) |
bb4cc1a8b revert "memcg: ge... |
614 615 616 617 618 619 |
{ if (!mz->on_tree) return; rb_erase(&mz->tree_node, &mctz->rb_root); mz->on_tree = false; } |
cf2c81279 mm: memcontrol: r... |
620 621 |
static void mem_cgroup_remove_exceeded(struct mem_cgroup_per_zone *mz, struct mem_cgroup_tree_per_zone *mctz) |
bb4cc1a8b revert "memcg: ge... |
622 |
{ |
0a31bc97c mm: memcontrol: r... |
623 624 625 |
unsigned long flags; spin_lock_irqsave(&mctz->lock, flags); |
cf2c81279 mm: memcontrol: r... |
626 |
__mem_cgroup_remove_exceeded(mz, mctz); |
0a31bc97c mm: memcontrol: r... |
627 |
spin_unlock_irqrestore(&mctz->lock, flags); |
bb4cc1a8b revert "memcg: ge... |
628 |
} |
3e32cb2e0 mm: memcontrol: l... |
629 630 631 |
static unsigned long soft_limit_excess(struct mem_cgroup *memcg) { unsigned long nr_pages = page_counter_read(&memcg->memory); |
4db0c3c29 mm: remove rest o... |
632 |
unsigned long soft_limit = READ_ONCE(memcg->soft_limit); |
3e32cb2e0 mm: memcontrol: l... |
633 634 635 636 637 638 639 |
unsigned long excess = 0; if (nr_pages > soft_limit) excess = nr_pages - soft_limit; return excess; } |
bb4cc1a8b revert "memcg: ge... |
640 641 642 |
static void mem_cgroup_update_tree(struct mem_cgroup *memcg, struct page *page) { |
3e32cb2e0 mm: memcontrol: l... |
643 |
unsigned long excess; |
bb4cc1a8b revert "memcg: ge... |
644 645 |
struct mem_cgroup_per_zone *mz; struct mem_cgroup_tree_per_zone *mctz; |
bb4cc1a8b revert "memcg: ge... |
646 |
|
e231875ba mm: memcontrol: c... |
647 |
mctz = soft_limit_tree_from_page(page); |
bb4cc1a8b revert "memcg: ge... |
648 649 650 651 652 |
/* * Necessary to update all ancestors when hierarchy is used. * because their event counter is not touched. */ for (; memcg; memcg = parent_mem_cgroup(memcg)) { |
e231875ba mm: memcontrol: c... |
653 |
mz = mem_cgroup_page_zoneinfo(memcg, page); |
3e32cb2e0 mm: memcontrol: l... |
654 |
excess = soft_limit_excess(memcg); |
bb4cc1a8b revert "memcg: ge... |
655 656 657 658 659 |
/* * We have to update the tree if mz is on RB-tree or * mem is over its softlimit. */ if (excess || mz->on_tree) { |
0a31bc97c mm: memcontrol: r... |
660 661 662 |
unsigned long flags; spin_lock_irqsave(&mctz->lock, flags); |
bb4cc1a8b revert "memcg: ge... |
663 664 |
/* if on-tree, remove it */ if (mz->on_tree) |
cf2c81279 mm: memcontrol: r... |
665 |
__mem_cgroup_remove_exceeded(mz, mctz); |
bb4cc1a8b revert "memcg: ge... |
666 667 668 669 |
/* * Insert again. mz->usage_in_excess will be updated. * If excess is 0, no tree ops. */ |
cf2c81279 mm: memcontrol: r... |
670 |
__mem_cgroup_insert_exceeded(mz, mctz, excess); |
0a31bc97c mm: memcontrol: r... |
671 |
spin_unlock_irqrestore(&mctz->lock, flags); |
bb4cc1a8b revert "memcg: ge... |
672 673 674 675 676 677 |
} } } static void mem_cgroup_remove_from_trees(struct mem_cgroup *memcg) { |
bb4cc1a8b revert "memcg: ge... |
678 |
struct mem_cgroup_tree_per_zone *mctz; |
e231875ba mm: memcontrol: c... |
679 680 |
struct mem_cgroup_per_zone *mz; int nid, zid; |
bb4cc1a8b revert "memcg: ge... |
681 |
|
e231875ba mm: memcontrol: c... |
682 683 684 685 |
for_each_node(nid) { for (zid = 0; zid < MAX_NR_ZONES; zid++) { mz = &memcg->nodeinfo[nid]->zoneinfo[zid]; mctz = soft_limit_tree_node_zone(nid, zid); |
cf2c81279 mm: memcontrol: r... |
686 |
mem_cgroup_remove_exceeded(mz, mctz); |
bb4cc1a8b revert "memcg: ge... |
687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 |
} } } static struct mem_cgroup_per_zone * __mem_cgroup_largest_soft_limit_node(struct mem_cgroup_tree_per_zone *mctz) { struct rb_node *rightmost = NULL; struct mem_cgroup_per_zone *mz; retry: mz = NULL; rightmost = rb_last(&mctz->rb_root); if (!rightmost) goto done; /* Nothing to reclaim from */ mz = rb_entry(rightmost, struct mem_cgroup_per_zone, tree_node); /* * Remove the node now but someone else can add it back, * we will to add it back at the end of reclaim to its correct * position in the tree. */ |
cf2c81279 mm: memcontrol: r... |
709 |
__mem_cgroup_remove_exceeded(mz, mctz); |
3e32cb2e0 mm: memcontrol: l... |
710 |
if (!soft_limit_excess(mz->memcg) || |
ec903c0c8 cgroup: rename cs... |
711 |
!css_tryget_online(&mz->memcg->css)) |
bb4cc1a8b revert "memcg: ge... |
712 713 714 715 716 717 718 719 720 |
goto retry; done: return mz; } static struct mem_cgroup_per_zone * mem_cgroup_largest_soft_limit_node(struct mem_cgroup_tree_per_zone *mctz) { struct mem_cgroup_per_zone *mz; |
0a31bc97c mm: memcontrol: r... |
721 |
spin_lock_irq(&mctz->lock); |
bb4cc1a8b revert "memcg: ge... |
722 |
mz = __mem_cgroup_largest_soft_limit_node(mctz); |
0a31bc97c mm: memcontrol: r... |
723 |
spin_unlock_irq(&mctz->lock); |
bb4cc1a8b revert "memcg: ge... |
724 725 |
return mz; } |
711d3d2c9 memcg: cpu hotplu... |
726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 |
/* * Implementation Note: reading percpu statistics for memcg. * * Both of vmstat[] and percpu_counter has threshold and do periodic * synchronization to implement "quick" read. There are trade-off between * reading cost and precision of value. Then, we may have a chance to implement * a periodic synchronizion of counter in memcg's counter. * * But this _read() function is used for user interface now. The user accounts * memory usage by memory cgroup and he _always_ requires exact value because * he accounts memory. Even if we provide quick-and-fuzzy read, we always * have to visit all online cpus and make sum. So, for now, unnecessary * synchronization is not implemented. (just implemented for cpu hotplug) * * If there are kernel internal actions which can make use of some not-exact * value, and reading all cpu value can be performance bottleneck in some * common workload, threashold and synchonization as vmstat[] should be * implemented. */ |
c0ff4b854 memcg: rename mem... |
745 |
static long mem_cgroup_read_stat(struct mem_cgroup *memcg, |
7a159cc9d memcg: use native... |
746 |
enum mem_cgroup_stat_index idx) |
c62b1a3b3 memcg: use generi... |
747 |
{ |
7a159cc9d memcg: use native... |
748 |
long val = 0; |
c62b1a3b3 memcg: use generi... |
749 |
int cpu; |
c62b1a3b3 memcg: use generi... |
750 |
|
711d3d2c9 memcg: cpu hotplu... |
751 752 |
get_online_cpus(); for_each_online_cpu(cpu) |
c0ff4b854 memcg: rename mem... |
753 |
val += per_cpu(memcg->stat->count[idx], cpu); |
711d3d2c9 memcg: cpu hotplu... |
754 |
#ifdef CONFIG_HOTPLUG_CPU |
c0ff4b854 memcg: rename mem... |
755 756 757 |
spin_lock(&memcg->pcp_counter_lock); val += memcg->nocpu_base.count[idx]; spin_unlock(&memcg->pcp_counter_lock); |
711d3d2c9 memcg: cpu hotplu... |
758 759 |
#endif put_online_cpus(); |
c62b1a3b3 memcg: use generi... |
760 761 |
return val; } |
c0ff4b854 memcg: rename mem... |
762 |
static unsigned long mem_cgroup_read_events(struct mem_cgroup *memcg, |
e9f8974f2 memcg: break out ... |
763 764 765 766 |
enum mem_cgroup_events_index idx) { unsigned long val = 0; int cpu; |
9c5675127 mm, memcg: protec... |
767 |
get_online_cpus(); |
e9f8974f2 memcg: break out ... |
768 |
for_each_online_cpu(cpu) |
c0ff4b854 memcg: rename mem... |
769 |
val += per_cpu(memcg->stat->events[idx], cpu); |
e9f8974f2 memcg: break out ... |
770 |
#ifdef CONFIG_HOTPLUG_CPU |
c0ff4b854 memcg: rename mem... |
771 772 773 |
spin_lock(&memcg->pcp_counter_lock); val += memcg->nocpu_base.events[idx]; spin_unlock(&memcg->pcp_counter_lock); |
e9f8974f2 memcg: break out ... |
774 |
#endif |
9c5675127 mm, memcg: protec... |
775 |
put_online_cpus(); |
e9f8974f2 memcg: break out ... |
776 777 |
return val; } |
c0ff4b854 memcg: rename mem... |
778 |
static void mem_cgroup_charge_statistics(struct mem_cgroup *memcg, |
b070e65c0 mm, memcg: add rs... |
779 |
struct page *page, |
0a31bc97c mm: memcontrol: r... |
780 |
int nr_pages) |
d52aa412d memory cgroup enh... |
781 |
{ |
b24028572 memcg: remove PCG... |
782 783 784 785 |
/* * Here, RSS means 'mapped anon' and anon's SwapCache. Shmem/tmpfs is * counted as CACHE even if it's on ANON LRU. */ |
0a31bc97c mm: memcontrol: r... |
786 |
if (PageAnon(page)) |
b24028572 memcg: remove PCG... |
787 |
__this_cpu_add(memcg->stat->count[MEM_CGROUP_STAT_RSS], |
c0ff4b854 memcg: rename mem... |
788 |
nr_pages); |
d52aa412d memory cgroup enh... |
789 |
else |
b24028572 memcg: remove PCG... |
790 |
__this_cpu_add(memcg->stat->count[MEM_CGROUP_STAT_CACHE], |
c0ff4b854 memcg: rename mem... |
791 |
nr_pages); |
55e462b05 memcg: simple sta... |
792 |
|
b070e65c0 mm, memcg: add rs... |
793 794 795 |
if (PageTransHuge(page)) __this_cpu_add(memcg->stat->count[MEM_CGROUP_STAT_RSS_HUGE], nr_pages); |
e401f1761 memcg: modify acc... |
796 797 |
/* pagein of a big page is an event. So, ignore page size */ if (nr_pages > 0) |
c0ff4b854 memcg: rename mem... |
798 |
__this_cpu_inc(memcg->stat->events[MEM_CGROUP_EVENTS_PGPGIN]); |
3751d6043 memcg: fix event ... |
799 |
else { |
c0ff4b854 memcg: rename mem... |
800 |
__this_cpu_inc(memcg->stat->events[MEM_CGROUP_EVENTS_PGPGOUT]); |
3751d6043 memcg: fix event ... |
801 802 |
nr_pages = -nr_pages; /* for event */ } |
e401f1761 memcg: modify acc... |
803 |
|
13114716c mm: memcg: keep r... |
804 |
__this_cpu_add(memcg->stat->nr_page_events, nr_pages); |
6d12e2d8d per-zone and recl... |
805 |
} |
e231875ba mm: memcontrol: c... |
806 |
unsigned long mem_cgroup_get_lru_size(struct lruvec *lruvec, enum lru_list lru) |
074291fea mm/vmscan: replac... |
807 808 809 810 811 812 |
{ struct mem_cgroup_per_zone *mz; mz = container_of(lruvec, struct mem_cgroup_per_zone, lruvec); return mz->lru_size[lru]; } |
e231875ba mm: memcontrol: c... |
813 814 815 |
static unsigned long mem_cgroup_node_nr_lru_pages(struct mem_cgroup *memcg, int nid, unsigned int lru_mask) |
bb2a0de92 memcg: consolidat... |
816 |
{ |
e231875ba mm: memcontrol: c... |
817 |
unsigned long nr = 0; |
889976dbc memcg: reclaim me... |
818 |
int zid; |
e231875ba mm: memcontrol: c... |
819 |
VM_BUG_ON((unsigned)nid >= nr_node_ids); |
bb2a0de92 memcg: consolidat... |
820 |
|
e231875ba mm: memcontrol: c... |
821 822 823 824 825 826 827 828 829 830 831 832 |
for (zid = 0; zid < MAX_NR_ZONES; zid++) { struct mem_cgroup_per_zone *mz; enum lru_list lru; for_each_lru(lru) { if (!(BIT(lru) & lru_mask)) continue; mz = &memcg->nodeinfo[nid]->zoneinfo[zid]; nr += mz->lru_size[lru]; } } return nr; |
889976dbc memcg: reclaim me... |
833 |
} |
bb2a0de92 memcg: consolidat... |
834 |
|
c0ff4b854 memcg: rename mem... |
835 |
static unsigned long mem_cgroup_nr_lru_pages(struct mem_cgroup *memcg, |
bb2a0de92 memcg: consolidat... |
836 |
unsigned int lru_mask) |
6d12e2d8d per-zone and recl... |
837 |
{ |
e231875ba mm: memcontrol: c... |
838 |
unsigned long nr = 0; |
889976dbc memcg: reclaim me... |
839 |
int nid; |
6d12e2d8d per-zone and recl... |
840 |
|
31aaea4aa memcontrol: use N... |
841 |
for_each_node_state(nid, N_MEMORY) |
e231875ba mm: memcontrol: c... |
842 843 |
nr += mem_cgroup_node_nr_lru_pages(memcg, nid, lru_mask); return nr; |
d52aa412d memory cgroup enh... |
844 |
} |
f53d7ce32 mm: memcg: shorte... |
845 846 |
static bool mem_cgroup_event_ratelimit(struct mem_cgroup *memcg, enum mem_cgroup_events_target target) |
7a159cc9d memcg: use native... |
847 848 |
{ unsigned long val, next; |
13114716c mm: memcg: keep r... |
849 |
val = __this_cpu_read(memcg->stat->nr_page_events); |
4799401fe memcg: Fix race c... |
850 |
next = __this_cpu_read(memcg->stat->targets[target]); |
7a159cc9d memcg: use native... |
851 |
/* from time_after() in jiffies.h */ |
f53d7ce32 mm: memcg: shorte... |
852 853 854 855 856 |
if ((long)next - (long)val < 0) { switch (target) { case MEM_CGROUP_TARGET_THRESH: next = val + THRESHOLDS_EVENTS_TARGET; break; |
bb4cc1a8b revert "memcg: ge... |
857 858 859 |
case MEM_CGROUP_TARGET_SOFTLIMIT: next = val + SOFTLIMIT_EVENTS_TARGET; break; |
f53d7ce32 mm: memcg: shorte... |
860 861 862 863 864 865 866 867 |
case MEM_CGROUP_TARGET_NUMAINFO: next = val + NUMAINFO_EVENTS_TARGET; break; default: break; } __this_cpu_write(memcg->stat->targets[target], next); return true; |
7a159cc9d memcg: use native... |
868 |
} |
f53d7ce32 mm: memcg: shorte... |
869 |
return false; |
d2265e6fa memcg : share eve... |
870 871 872 873 874 875 |
} /* * Check events in order. * */ |
c0ff4b854 memcg: rename mem... |
876 |
static void memcg_check_events(struct mem_cgroup *memcg, struct page *page) |
d2265e6fa memcg : share eve... |
877 878 |
{ /* threshold event is triggered in finer grain than soft limit */ |
f53d7ce32 mm: memcg: shorte... |
879 880 |
if (unlikely(mem_cgroup_event_ratelimit(memcg, MEM_CGROUP_TARGET_THRESH))) { |
bb4cc1a8b revert "memcg: ge... |
881 |
bool do_softlimit; |
82b3f2a71 mm/memcontrol.c: ... |
882 |
bool do_numainfo __maybe_unused; |
f53d7ce32 mm: memcg: shorte... |
883 |
|
bb4cc1a8b revert "memcg: ge... |
884 885 |
do_softlimit = mem_cgroup_event_ratelimit(memcg, MEM_CGROUP_TARGET_SOFTLIMIT); |
f53d7ce32 mm: memcg: shorte... |
886 887 888 889 |
#if MAX_NUMNODES > 1 do_numainfo = mem_cgroup_event_ratelimit(memcg, MEM_CGROUP_TARGET_NUMAINFO); #endif |
c0ff4b854 memcg: rename mem... |
890 |
mem_cgroup_threshold(memcg); |
bb4cc1a8b revert "memcg: ge... |
891 892 |
if (unlikely(do_softlimit)) mem_cgroup_update_tree(memcg, page); |
453a9bf34 memcg: fix numa s... |
893 |
#if MAX_NUMNODES > 1 |
f53d7ce32 mm: memcg: shorte... |
894 |
if (unlikely(do_numainfo)) |
c0ff4b854 memcg: rename mem... |
895 |
atomic_inc(&memcg->numainfo_events); |
453a9bf34 memcg: fix numa s... |
896 |
#endif |
0a31bc97c mm: memcontrol: r... |
897 |
} |
d2265e6fa memcg : share eve... |
898 |
} |
cf475ad28 cgroups: add an o... |
899 |
struct mem_cgroup *mem_cgroup_from_task(struct task_struct *p) |
78fb74669 Memory controller... |
900 |
{ |
31a78f23b mm owner: fix rac... |
901 902 903 904 905 906 907 |
/* * 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; |
073219e99 cgroup: clean up ... |
908 |
return mem_cgroup_from_css(task_css(p, memory_cgrp_id)); |
78fb74669 Memory controller... |
909 |
} |
df3819754 memcg: get_mem_cg... |
910 |
static struct mem_cgroup *get_mem_cgroup_from_mm(struct mm_struct *mm) |
54595fe26 memcg: use css_tr... |
911 |
{ |
c0ff4b854 memcg: rename mem... |
912 |
struct mem_cgroup *memcg = NULL; |
0b7f569e4 memcg: fix OOM ki... |
913 |
|
54595fe26 memcg: use css_tr... |
914 915 |
rcu_read_lock(); do { |
6f6acb005 memcg: fix swapca... |
916 917 918 919 920 921 |
/* * Page cache insertions can happen withou an * actual mm context, e.g. during disk probing * on boot, loopback IO, acct() writes etc. */ if (unlikely(!mm)) |
df3819754 memcg: get_mem_cg... |
922 |
memcg = root_mem_cgroup; |
6f6acb005 memcg: fix swapca... |
923 924 925 926 927 |
else { memcg = mem_cgroup_from_task(rcu_dereference(mm->owner)); if (unlikely(!memcg)) memcg = root_mem_cgroup; } |
ec903c0c8 cgroup: rename cs... |
928 |
} while (!css_tryget_online(&memcg->css)); |
54595fe26 memcg: use css_tr... |
929 |
rcu_read_unlock(); |
c0ff4b854 memcg: rename mem... |
930 |
return memcg; |
54595fe26 memcg: use css_tr... |
931 |
} |
5660048cc mm: move memcg hi... |
932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 |
/** * mem_cgroup_iter - iterate over memory cgroup hierarchy * @root: hierarchy root * @prev: previously returned memcg, NULL on first invocation * @reclaim: cookie for shared reclaim walks, NULL for full walks * * Returns references to children of the hierarchy below @root, or * @root itself, or %NULL after a full round-trip. * * Caller must pass the return value in @prev on subsequent * invocations for reference counting, or use mem_cgroup_iter_break() * to cancel a hierarchy walk before the round-trip is complete. * * Reclaimers can specify a zone and a priority level in @reclaim to * divide up the memcgs in the hierarchy among all concurrent * reclaimers operating on the same zone and priority. */ |
694fbc0fe revert "memcg: en... |
949 |
struct mem_cgroup *mem_cgroup_iter(struct mem_cgroup *root, |
5660048cc mm: move memcg hi... |
950 |
struct mem_cgroup *prev, |
694fbc0fe revert "memcg: en... |
951 |
struct mem_cgroup_reclaim_cookie *reclaim) |
14067bb3e memcg: hierarchic... |
952 |
{ |
5ac8fb31a mm: memcontrol: c... |
953 954 |
struct reclaim_iter *uninitialized_var(iter); struct cgroup_subsys_state *css = NULL; |
9f3a0d093 mm: memcg: consol... |
955 |
struct mem_cgroup *memcg = NULL; |
5ac8fb31a mm: memcontrol: c... |
956 |
struct mem_cgroup *pos = NULL; |
711d3d2c9 memcg: cpu hotplu... |
957 |
|
694fbc0fe revert "memcg: en... |
958 959 |
if (mem_cgroup_disabled()) return NULL; |
5660048cc mm: move memcg hi... |
960 |
|
9f3a0d093 mm: memcg: consol... |
961 962 |
if (!root) root = root_mem_cgroup; |
7d74b06f2 memcg: use for_ea... |
963 |
|
9f3a0d093 mm: memcg: consol... |
964 |
if (prev && !reclaim) |
5ac8fb31a mm: memcontrol: c... |
965 |
pos = prev; |
14067bb3e memcg: hierarchic... |
966 |
|
9f3a0d093 mm: memcg: consol... |
967 968 |
if (!root->use_hierarchy && root != root_mem_cgroup) { if (prev) |
5ac8fb31a mm: memcontrol: c... |
969 |
goto out; |
694fbc0fe revert "memcg: en... |
970 |
return root; |
9f3a0d093 mm: memcg: consol... |
971 |
} |
14067bb3e memcg: hierarchic... |
972 |
|
542f85f9a memcg: rework mem... |
973 |
rcu_read_lock(); |
5f5781619 memcg: relax memc... |
974 |
|
5ac8fb31a mm: memcontrol: c... |
975 976 977 978 979 980 981 982 983 984 |
if (reclaim) { struct mem_cgroup_per_zone *mz; mz = mem_cgroup_zone_zoneinfo(root, reclaim->zone); iter = &mz->iter[reclaim->priority]; if (prev && reclaim->generation != iter->generation) goto out_unlock; do { |
4db0c3c29 mm: remove rest o... |
985 |
pos = READ_ONCE(iter->position); |
5ac8fb31a mm: memcontrol: c... |
986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 |
/* * A racing update may change the position and * put the last reference, hence css_tryget(), * or retry to see the updated position. */ } while (pos && !css_tryget(&pos->css)); } if (pos) css = &pos->css; for (;;) { css = css_next_descendant_pre(css, &root->css); if (!css) { /* * Reclaimers share the hierarchy walk, and a * new one might jump in right at the end of * the hierarchy - make sure they see at least * one group and restart from the beginning. */ if (!prev) continue; break; |
527a5ec9a mm: memcg: per-pr... |
1009 |
} |
7d74b06f2 memcg: use for_ea... |
1010 |
|
5ac8fb31a mm: memcontrol: c... |
1011 1012 1013 1014 1015 1016 |
/* * Verify the css and acquire a reference. The root * is provided by the caller, so we know it's alive * and kicking, and don't take an extra reference. */ memcg = mem_cgroup_from_css(css); |
14067bb3e memcg: hierarchic... |
1017 |
|
5ac8fb31a mm: memcontrol: c... |
1018 1019 |
if (css == &root->css) break; |
14067bb3e memcg: hierarchic... |
1020 |
|
b2052564e mm: memcontrol: c... |
1021 |
if (css_tryget(css)) { |
5ac8fb31a mm: memcontrol: c... |
1022 1023 1024 1025 1026 1027 1028 |
/* * Make sure the memcg is initialized: * mem_cgroup_css_online() orders the the * initialization against setting the flag. */ if (smp_load_acquire(&memcg->initialized)) break; |
542f85f9a memcg: rework mem... |
1029 |
|
5ac8fb31a mm: memcontrol: c... |
1030 |
css_put(css); |
527a5ec9a mm: memcg: per-pr... |
1031 |
} |
9f3a0d093 mm: memcg: consol... |
1032 |
|
5ac8fb31a mm: memcontrol: c... |
1033 |
memcg = NULL; |
9f3a0d093 mm: memcg: consol... |
1034 |
} |
5ac8fb31a mm: memcontrol: c... |
1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 |
if (reclaim) { if (cmpxchg(&iter->position, pos, memcg) == pos) { if (memcg) css_get(&memcg->css); if (pos) css_put(&pos->css); } /* * pairs with css_tryget when dereferencing iter->position * above. */ if (pos) css_put(&pos->css); if (!memcg) iter->generation++; else if (!prev) reclaim->generation = iter->generation; |
9f3a0d093 mm: memcg: consol... |
1055 |
} |
5ac8fb31a mm: memcontrol: c... |
1056 |
|
542f85f9a memcg: rework mem... |
1057 1058 |
out_unlock: rcu_read_unlock(); |
5ac8fb31a mm: memcontrol: c... |
1059 |
out: |
c40046f3a memcg: keep prev'... |
1060 1061 |
if (prev && prev != root) css_put(&prev->css); |
9f3a0d093 mm: memcg: consol... |
1062 |
return memcg; |
14067bb3e memcg: hierarchic... |
1063 |
} |
7d74b06f2 memcg: use for_ea... |
1064 |
|
5660048cc mm: move memcg hi... |
1065 1066 1067 1068 1069 1070 1071 |
/** * mem_cgroup_iter_break - abort a hierarchy walk prematurely * @root: hierarchy root * @prev: last visited hierarchy member as returned by mem_cgroup_iter() */ void mem_cgroup_iter_break(struct mem_cgroup *root, struct mem_cgroup *prev) |
9f3a0d093 mm: memcg: consol... |
1072 1073 1074 1075 1076 1077 |
{ if (!root) root = root_mem_cgroup; if (prev && prev != root) css_put(&prev->css); } |
7d74b06f2 memcg: use for_ea... |
1078 |
|
9f3a0d093 mm: memcg: consol... |
1079 1080 1081 1082 1083 1084 |
/* * Iteration constructs for visiting all cgroups (under a tree). If * loops are exited prematurely (break), mem_cgroup_iter_break() must * be used for reference counting. */ #define for_each_mem_cgroup_tree(iter, root) \ |
527a5ec9a mm: memcg: per-pr... |
1085 |
for (iter = mem_cgroup_iter(root, NULL, NULL); \ |
9f3a0d093 mm: memcg: consol... |
1086 |
iter != NULL; \ |
527a5ec9a mm: memcg: per-pr... |
1087 |
iter = mem_cgroup_iter(root, iter, NULL)) |
711d3d2c9 memcg: cpu hotplu... |
1088 |
|
9f3a0d093 mm: memcg: consol... |
1089 |
#define for_each_mem_cgroup(iter) \ |
527a5ec9a mm: memcg: per-pr... |
1090 |
for (iter = mem_cgroup_iter(NULL, NULL, NULL); \ |
9f3a0d093 mm: memcg: consol... |
1091 |
iter != NULL; \ |
527a5ec9a mm: memcg: per-pr... |
1092 |
iter = mem_cgroup_iter(NULL, iter, NULL)) |
14067bb3e memcg: hierarchic... |
1093 |
|
68ae564bb mm, memcg: avoid ... |
1094 |
void __mem_cgroup_count_vm_event(struct mm_struct *mm, enum vm_event_item idx) |
456f998ec memcg: add the pa... |
1095 |
{ |
c0ff4b854 memcg: rename mem... |
1096 |
struct mem_cgroup *memcg; |
456f998ec memcg: add the pa... |
1097 |
|
456f998ec memcg: add the pa... |
1098 |
rcu_read_lock(); |
c0ff4b854 memcg: rename mem... |
1099 1100 |
memcg = mem_cgroup_from_task(rcu_dereference(mm->owner)); if (unlikely(!memcg)) |
456f998ec memcg: add the pa... |
1101 1102 1103 |
goto out; switch (idx) { |
456f998ec memcg: add the pa... |
1104 |
case PGFAULT: |
0e574a932 mm: memcg: clean ... |
1105 1106 1107 1108 |
this_cpu_inc(memcg->stat->events[MEM_CGROUP_EVENTS_PGFAULT]); break; case PGMAJFAULT: this_cpu_inc(memcg->stat->events[MEM_CGROUP_EVENTS_PGMAJFAULT]); |
456f998ec memcg: add the pa... |
1109 1110 1111 1112 1113 1114 1115 |
break; default: BUG(); } out: rcu_read_unlock(); } |
68ae564bb mm, memcg: avoid ... |
1116 |
EXPORT_SYMBOL(__mem_cgroup_count_vm_event); |
456f998ec memcg: add the pa... |
1117 |
|
925b7673c mm: make per-memc... |
1118 1119 1120 |
/** * mem_cgroup_zone_lruvec - get the lru list vector for a zone and memcg * @zone: zone of the wanted lruvec |
fa9add641 mm/memcg: apply a... |
1121 |
* @memcg: memcg of the wanted lruvec |
925b7673c mm: make per-memc... |
1122 1123 1124 1125 1126 1127 1128 1129 1130 |
* * Returns the lru list vector holding pages for the given @zone and * @mem. This can be the global zone lruvec, if the memory controller * is disabled. */ struct lruvec *mem_cgroup_zone_lruvec(struct zone *zone, struct mem_cgroup *memcg) { struct mem_cgroup_per_zone *mz; |
bea8c150a memcg: fix hotplu... |
1131 |
struct lruvec *lruvec; |
925b7673c mm: make per-memc... |
1132 |
|
bea8c150a memcg: fix hotplu... |
1133 1134 1135 1136 |
if (mem_cgroup_disabled()) { lruvec = &zone->lruvec; goto out; } |
925b7673c mm: make per-memc... |
1137 |
|
e231875ba mm: memcontrol: c... |
1138 |
mz = mem_cgroup_zone_zoneinfo(memcg, zone); |
bea8c150a memcg: fix hotplu... |
1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 |
lruvec = &mz->lruvec; out: /* * Since a node can be onlined after the mem_cgroup was created, * we have to be prepared to initialize lruvec->zone here; * and if offlined then reonlined, we need to reinitialize it. */ if (unlikely(lruvec->zone != zone)) lruvec->zone = zone; return lruvec; |
925b7673c mm: make per-memc... |
1149 |
} |
925b7673c mm: make per-memc... |
1150 |
/** |
dfe0e773d mm: memcontrol: u... |
1151 |
* mem_cgroup_page_lruvec - return lruvec for isolating/putting an LRU page |
925b7673c mm: make per-memc... |
1152 |
* @page: the page |
fa9add641 mm/memcg: apply a... |
1153 |
* @zone: zone of the page |
dfe0e773d mm: memcontrol: u... |
1154 1155 1156 1157 |
* * This function is only safe when following the LRU page isolation * and putback protocol: the LRU lock must be held, and the page must * either be PageLRU() or the caller must have isolated/allocated it. |
925b7673c mm: make per-memc... |
1158 |
*/ |
fa9add641 mm/memcg: apply a... |
1159 |
struct lruvec *mem_cgroup_page_lruvec(struct page *page, struct zone *zone) |
08e552c69 memcg: synchroniz... |
1160 |
{ |
08e552c69 memcg: synchroniz... |
1161 |
struct mem_cgroup_per_zone *mz; |
925b7673c mm: make per-memc... |
1162 |
struct mem_cgroup *memcg; |
bea8c150a memcg: fix hotplu... |
1163 |
struct lruvec *lruvec; |
6d12e2d8d per-zone and recl... |
1164 |
|
bea8c150a memcg: fix hotplu... |
1165 1166 1167 1168 |
if (mem_cgroup_disabled()) { lruvec = &zone->lruvec; goto out; } |
925b7673c mm: make per-memc... |
1169 |
|
1306a85ae mm: embed the mem... |
1170 |
memcg = page->mem_cgroup; |
7512102cf memcg: fix GPF wh... |
1171 |
/* |
dfe0e773d mm: memcontrol: u... |
1172 |
* Swapcache readahead pages are added to the LRU - and |
298333157 mm: memcontrol: r... |
1173 |
* possibly migrated - before they are charged. |
7512102cf memcg: fix GPF wh... |
1174 |
*/ |
298333157 mm: memcontrol: r... |
1175 1176 |
if (!memcg) memcg = root_mem_cgroup; |
7512102cf memcg: fix GPF wh... |
1177 |
|
e231875ba mm: memcontrol: c... |
1178 |
mz = mem_cgroup_page_zoneinfo(memcg, page); |
bea8c150a memcg: fix hotplu... |
1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 |
lruvec = &mz->lruvec; out: /* * Since a node can be onlined after the mem_cgroup was created, * we have to be prepared to initialize lruvec->zone here; * and if offlined then reonlined, we need to reinitialize it. */ if (unlikely(lruvec->zone != zone)) lruvec->zone = zone; return lruvec; |
08e552c69 memcg: synchroniz... |
1189 |
} |
b69408e88 vmscan: Use an in... |
1190 |
|
925b7673c mm: make per-memc... |
1191 |
/** |
fa9add641 mm/memcg: apply a... |
1192 1193 1194 1195 |
* mem_cgroup_update_lru_size - account for adding or removing an lru page * @lruvec: mem_cgroup per zone lru vector * @lru: index of lru list the page is sitting on * @nr_pages: positive when adding or negative when removing |
925b7673c mm: make per-memc... |
1196 |
* |
fa9add641 mm/memcg: apply a... |
1197 1198 |
* This function must be called when a page is added to or removed from an * lru list. |
3f58a8294 memcg: move memcg... |
1199 |
*/ |
fa9add641 mm/memcg: apply a... |
1200 1201 |
void mem_cgroup_update_lru_size(struct lruvec *lruvec, enum lru_list lru, int nr_pages) |
3f58a8294 memcg: move memcg... |
1202 1203 |
{ struct mem_cgroup_per_zone *mz; |
fa9add641 mm/memcg: apply a... |
1204 |
unsigned long *lru_size; |
3f58a8294 memcg: move memcg... |
1205 1206 1207 |
if (mem_cgroup_disabled()) return; |
fa9add641 mm/memcg: apply a... |
1208 1209 1210 1211 |
mz = container_of(lruvec, struct mem_cgroup_per_zone, lruvec); lru_size = mz->lru_size + lru; *lru_size += nr_pages; VM_BUG_ON((long)(*lru_size) < 0); |
08e552c69 memcg: synchroniz... |
1212 |
} |
544122e5e memcg: fix LRU ac... |
1213 |
|
2314b42db mm: memcontrol: d... |
1214 |
bool mem_cgroup_is_descendant(struct mem_cgroup *memcg, struct mem_cgroup *root) |
3e92041d6 memcg: add mem_cg... |
1215 |
{ |
2314b42db mm: memcontrol: d... |
1216 |
if (root == memcg) |
91c63734f kernel: cgroup: p... |
1217 |
return true; |
2314b42db mm: memcontrol: d... |
1218 |
if (!root->use_hierarchy) |
91c63734f kernel: cgroup: p... |
1219 |
return false; |
2314b42db mm: memcontrol: d... |
1220 |
return cgroup_is_descendant(memcg->css.cgroup, root->css.cgroup); |
c3ac9a8ad mm: memcg: count ... |
1221 |
} |
2314b42db mm: memcontrol: d... |
1222 |
bool task_in_mem_cgroup(struct task_struct *task, struct mem_cgroup *memcg) |
c3ac9a8ad mm: memcg: count ... |
1223 |
{ |
2314b42db mm: memcontrol: d... |
1224 |
struct mem_cgroup *task_memcg; |
158e0a2d1 memcg: use find_l... |
1225 |
struct task_struct *p; |
ffbdccf5e mm, memcg: don't ... |
1226 |
bool ret; |
4c4a22148 memcontrol: move ... |
1227 |
|
158e0a2d1 memcg: use find_l... |
1228 |
p = find_lock_task_mm(task); |
de077d222 oom, memcg: fix e... |
1229 |
if (p) { |
2314b42db mm: memcontrol: d... |
1230 |
task_memcg = get_mem_cgroup_from_mm(p->mm); |
de077d222 oom, memcg: fix e... |
1231 1232 1233 1234 1235 1236 1237 |
task_unlock(p); } else { /* * All threads may have already detached their mm's, but the oom * killer still needs to detect if they have already been oom * killed to prevent needlessly killing additional tasks. */ |
ffbdccf5e mm, memcg: don't ... |
1238 |
rcu_read_lock(); |
2314b42db mm: memcontrol: d... |
1239 1240 |
task_memcg = mem_cgroup_from_task(task); css_get(&task_memcg->css); |
ffbdccf5e mm, memcg: don't ... |
1241 |
rcu_read_unlock(); |
de077d222 oom, memcg: fix e... |
1242 |
} |
2314b42db mm: memcontrol: d... |
1243 1244 |
ret = mem_cgroup_is_descendant(task_memcg, memcg); css_put(&task_memcg->css); |
4c4a22148 memcontrol: move ... |
1245 1246 |
return ret; } |
c56d5c7df mm/vmscan: push l... |
1247 |
int mem_cgroup_inactive_anon_is_low(struct lruvec *lruvec) |
14797e236 memcg: add inacti... |
1248 |
{ |
9b272977e memcg: skip scann... |
1249 |
unsigned long inactive_ratio; |
14797e236 memcg: add inacti... |
1250 |
unsigned long inactive; |
9b272977e memcg: skip scann... |
1251 |
unsigned long active; |
c772be939 memcg: fix calcul... |
1252 |
unsigned long gb; |
14797e236 memcg: add inacti... |
1253 |
|
4d7dcca21 mm/memcg: get_lru... |
1254 1255 |
inactive = mem_cgroup_get_lru_size(lruvec, LRU_INACTIVE_ANON); active = mem_cgroup_get_lru_size(lruvec, LRU_ACTIVE_ANON); |
14797e236 memcg: add inacti... |
1256 |
|
c772be939 memcg: fix calcul... |
1257 1258 1259 1260 1261 |
gb = (inactive + active) >> (30 - PAGE_SHIFT); if (gb) inactive_ratio = int_sqrt(10 * gb); else inactive_ratio = 1; |
9b272977e memcg: skip scann... |
1262 |
return inactive * inactive_ratio < active; |
14797e236 memcg: add inacti... |
1263 |
} |
90cbc2508 vmscan: force sca... |
1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 |
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); } |
3e32cb2e0 mm: memcontrol: l... |
1277 |
#define mem_cgroup_from_counter(counter, member) \ |
6d61ef409 memcg: memory cgr... |
1278 |
container_of(counter, struct mem_cgroup, member) |
19942822d memcg: prevent en... |
1279 |
/** |
9d11ea9f1 memcg: simplify t... |
1280 |
* mem_cgroup_margin - calculate chargeable space of a memory cgroup |
dad7557eb mm: fix kernel-do... |
1281 |
* @memcg: the memory cgroup |
19942822d memcg: prevent en... |
1282 |
* |
9d11ea9f1 memcg: simplify t... |
1283 |
* Returns the maximum amount of memory @mem can be charged with, in |
7ec99d621 memcg: unify char... |
1284 |
* pages. |
19942822d memcg: prevent en... |
1285 |
*/ |
c0ff4b854 memcg: rename mem... |
1286 |
static unsigned long mem_cgroup_margin(struct mem_cgroup *memcg) |
19942822d memcg: prevent en... |
1287 |
{ |
3e32cb2e0 mm: memcontrol: l... |
1288 1289 1290 |
unsigned long margin = 0; unsigned long count; unsigned long limit; |
9d11ea9f1 memcg: simplify t... |
1291 |
|
3e32cb2e0 mm: memcontrol: l... |
1292 |
count = page_counter_read(&memcg->memory); |
4db0c3c29 mm: remove rest o... |
1293 |
limit = READ_ONCE(memcg->memory.limit); |
3e32cb2e0 mm: memcontrol: l... |
1294 1295 1296 1297 1298 |
if (count < limit) margin = limit - count; if (do_swap_account) { count = page_counter_read(&memcg->memsw); |
4db0c3c29 mm: remove rest o... |
1299 |
limit = READ_ONCE(memcg->memsw.limit); |
3e32cb2e0 mm: memcontrol: l... |
1300 1301 1302 1303 1304 |
if (count <= limit) margin = min(margin, limit - count); } return margin; |
19942822d memcg: prevent en... |
1305 |
} |
1f4c025b5 memcg: export mem... |
1306 |
int mem_cgroup_swappiness(struct mem_cgroup *memcg) |
a7885eb8a memcg: swappiness |
1307 |
{ |
a7885eb8a memcg: swappiness |
1308 |
/* root ? */ |
14208b0ec Merge branch 'for... |
1309 |
if (mem_cgroup_disabled() || !memcg->css.parent) |
a7885eb8a memcg: swappiness |
1310 |
return vm_swappiness; |
bf1ff2635 memcg: remove mem... |
1311 |
return memcg->swappiness; |
a7885eb8a memcg: swappiness |
1312 |
} |
619d094b5 memcg: simplify m... |
1313 |
/* |
bdcbb659f memcg: fold mem_c... |
1314 |
* A routine for checking "mem" is under move_account() or not. |
32047e2a8 memcg: avoid lock... |
1315 |
* |
bdcbb659f memcg: fold mem_c... |
1316 1317 1318 |
* Checking a cgroup is mc.from or mc.to or under hierarchy of * moving cgroups. This is for waiting at high-memory pressure * caused by "move". |
32047e2a8 memcg: avoid lock... |
1319 |
*/ |
c0ff4b854 memcg: rename mem... |
1320 |
static bool mem_cgroup_under_move(struct mem_cgroup *memcg) |
4b5343346 memcg: clean up t... |
1321 |
{ |
2bd9bb206 memcg: clean up w... |
1322 1323 |
struct mem_cgroup *from; struct mem_cgroup *to; |
4b5343346 memcg: clean up t... |
1324 |
bool ret = false; |
2bd9bb206 memcg: clean up w... |
1325 1326 1327 1328 1329 1330 1331 1332 1333 |
/* * Unlike task_move routines, we access mc.to, mc.from not under * mutual exclusion by cgroup_mutex. Here, we take spinlock instead. */ spin_lock(&mc.lock); from = mc.from; to = mc.to; if (!from) goto unlock; |
3e92041d6 memcg: add mem_cg... |
1334 |
|
2314b42db mm: memcontrol: d... |
1335 1336 |
ret = mem_cgroup_is_descendant(from, memcg) || mem_cgroup_is_descendant(to, memcg); |
2bd9bb206 memcg: clean up w... |
1337 1338 |
unlock: spin_unlock(&mc.lock); |
4b5343346 memcg: clean up t... |
1339 1340 |
return ret; } |
c0ff4b854 memcg: rename mem... |
1341 |
static bool mem_cgroup_wait_acct_move(struct mem_cgroup *memcg) |
4b5343346 memcg: clean up t... |
1342 1343 |
{ if (mc.moving_task && current != mc.moving_task) { |
c0ff4b854 memcg: rename mem... |
1344 |
if (mem_cgroup_under_move(memcg)) { |
4b5343346 memcg: clean up t... |
1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 |
DEFINE_WAIT(wait); prepare_to_wait(&mc.waitq, &wait, TASK_INTERRUPTIBLE); /* moving charge context might have finished. */ if (mc.moving_task) schedule(); finish_wait(&mc.waitq, &wait); return true; } } return false; } |
58cf188ed memcg, oom: provi... |
1356 |
#define K(x) ((x) << (PAGE_SHIFT-10)) |
e222432bf memcg: show memcg... |
1357 |
/** |
58cf188ed memcg, oom: provi... |
1358 |
* mem_cgroup_print_oom_info: Print OOM information relevant to memory controller. |
e222432bf memcg: show memcg... |
1359 1360 1361 1362 1363 1364 1365 1366 |
* @memcg: The memory cgroup that went over limit * @p: Task that is going to be killed * * NOTE: @memcg and @p's mem_cgroup can be different when hierarchy is * enabled */ void mem_cgroup_print_oom_info(struct mem_cgroup *memcg, struct task_struct *p) { |
e61734c55 cgroup: remove cg... |
1367 |
/* oom_info_lock ensures that parallel ooms do not interleave */ |
08088cb9a memcg: change oom... |
1368 |
static DEFINE_MUTEX(oom_info_lock); |
58cf188ed memcg, oom: provi... |
1369 1370 |
struct mem_cgroup *iter; unsigned int i; |
e222432bf memcg: show memcg... |
1371 |
|
08088cb9a memcg: change oom... |
1372 |
mutex_lock(&oom_info_lock); |
e222432bf memcg: show memcg... |
1373 |
rcu_read_lock(); |
2415b9f5c memcg: print cgro... |
1374 1375 1376 1377 1378 1379 1380 |
if (p) { pr_info("Task in "); pr_cont_cgroup_path(task_cgroup(p, memory_cgrp_id)); pr_cont(" killed as a result of limit of "); } else { pr_info("Memory limit reached of cgroup "); } |
e61734c55 cgroup: remove cg... |
1381 |
pr_cont_cgroup_path(memcg->css.cgroup); |
0346dadbf memcg: remove ext... |
1382 1383 |
pr_cont(" "); |
e222432bf memcg: show memcg... |
1384 |
|
e222432bf memcg: show memcg... |
1385 |
rcu_read_unlock(); |
3e32cb2e0 mm: memcontrol: l... |
1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 |
pr_info("memory: usage %llukB, limit %llukB, failcnt %lu ", K((u64)page_counter_read(&memcg->memory)), K((u64)memcg->memory.limit), memcg->memory.failcnt); pr_info("memory+swap: usage %llukB, limit %llukB, failcnt %lu ", K((u64)page_counter_read(&memcg->memsw)), K((u64)memcg->memsw.limit), memcg->memsw.failcnt); pr_info("kmem: usage %llukB, limit %llukB, failcnt %lu ", K((u64)page_counter_read(&memcg->kmem)), K((u64)memcg->kmem.limit), memcg->kmem.failcnt); |
58cf188ed memcg, oom: provi... |
1398 1399 |
for_each_mem_cgroup_tree(iter, memcg) { |
e61734c55 cgroup: remove cg... |
1400 1401 |
pr_info("Memory cgroup stats for "); pr_cont_cgroup_path(iter->css.cgroup); |
58cf188ed memcg, oom: provi... |
1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 |
pr_cont(":"); for (i = 0; i < MEM_CGROUP_STAT_NSTATS; i++) { if (i == MEM_CGROUP_STAT_SWAP && !do_swap_account) continue; pr_cont(" %s:%ldKB", mem_cgroup_stat_names[i], K(mem_cgroup_read_stat(iter, i))); } for (i = 0; i < NR_LRU_LISTS; i++) pr_cont(" %s:%luKB", mem_cgroup_lru_names[i], K(mem_cgroup_nr_lru_pages(iter, BIT(i)))); pr_cont(" "); } |
08088cb9a memcg: change oom... |
1418 |
mutex_unlock(&oom_info_lock); |
e222432bf memcg: show memcg... |
1419 |
} |
81d39c20f memcg: fix shrink... |
1420 1421 1422 1423 |
/* * This function returns the number of memcg under hierarchy tree. Returns * 1(self count) if no children. */ |
c0ff4b854 memcg: rename mem... |
1424 |
static int mem_cgroup_count_children(struct mem_cgroup *memcg) |
81d39c20f memcg: fix shrink... |
1425 1426 |
{ int num = 0; |
7d74b06f2 memcg: use for_ea... |
1427 |
struct mem_cgroup *iter; |
c0ff4b854 memcg: rename mem... |
1428 |
for_each_mem_cgroup_tree(iter, memcg) |
7d74b06f2 memcg: use for_ea... |
1429 |
num++; |
81d39c20f memcg: fix shrink... |
1430 1431 |
return num; } |
6d61ef409 memcg: memory cgr... |
1432 |
/* |
a63d83f42 oom: badness heur... |
1433 1434 |
* Return the memory (and swap, if configured) limit for a memcg. */ |
3e32cb2e0 mm: memcontrol: l... |
1435 |
static unsigned long mem_cgroup_get_limit(struct mem_cgroup *memcg) |
a63d83f42 oom: badness heur... |
1436 |
{ |
3e32cb2e0 mm: memcontrol: l... |
1437 |
unsigned long limit; |
f3e8eb70b memcg: fix unit m... |
1438 |
|
3e32cb2e0 mm: memcontrol: l... |
1439 |
limit = memcg->memory.limit; |
9a5a8f19b memcg: oom: fix t... |
1440 |
if (mem_cgroup_swappiness(memcg)) { |
3e32cb2e0 mm: memcontrol: l... |
1441 |
unsigned long memsw_limit; |
9a5a8f19b memcg: oom: fix t... |
1442 |
|
3e32cb2e0 mm: memcontrol: l... |
1443 1444 |
memsw_limit = memcg->memsw.limit; limit = min(limit + total_swap_pages, memsw_limit); |
9a5a8f19b memcg: oom: fix t... |
1445 |
} |
9a5a8f19b memcg: oom: fix t... |
1446 |
return limit; |
a63d83f42 oom: badness heur... |
1447 |
} |
19965460e mm, memcg: make m... |
1448 1449 |
static void mem_cgroup_out_of_memory(struct mem_cgroup *memcg, gfp_t gfp_mask, int order) |
9cbb78bb3 mm, memcg: introd... |
1450 1451 1452 1453 1454 1455 |
{ struct mem_cgroup *iter; unsigned long chosen_points = 0; unsigned long totalpages; unsigned int points = 0; struct task_struct *chosen = NULL; |
876aafbfd mm, memcg: move a... |
1456 |
/* |
465adcf1e mm, memcg: give e... |
1457 1458 1459 |
* If current has a pending SIGKILL or is exiting, then automatically * select it. The goal is to allow it to allocate so that it may * quickly exit and free its memory. |
876aafbfd mm, memcg: move a... |
1460 |
*/ |
d003f371b oom: don't assume... |
1461 |
if (fatal_signal_pending(current) || task_will_free_mem(current)) { |
49550b605 oom: add helpers ... |
1462 |
mark_tsk_oom_victim(current); |
876aafbfd mm, memcg: move a... |
1463 1464 |
return; } |
2415b9f5c memcg: print cgro... |
1465 |
check_panic_on_oom(CONSTRAINT_MEMCG, gfp_mask, order, NULL, memcg); |
3e32cb2e0 mm: memcontrol: l... |
1466 |
totalpages = mem_cgroup_get_limit(memcg) ? : 1; |
9cbb78bb3 mm, memcg: introd... |
1467 |
for_each_mem_cgroup_tree(iter, memcg) { |
72ec70299 cgroup: make task... |
1468 |
struct css_task_iter it; |
9cbb78bb3 mm, memcg: introd... |
1469 |
struct task_struct *task; |
72ec70299 cgroup: make task... |
1470 1471 |
css_task_iter_start(&iter->css, &it); while ((task = css_task_iter_next(&it))) { |
9cbb78bb3 mm, memcg: introd... |
1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 |
switch (oom_scan_process_thread(task, totalpages, NULL, false)) { case OOM_SCAN_SELECT: if (chosen) put_task_struct(chosen); chosen = task; chosen_points = ULONG_MAX; get_task_struct(chosen); /* fall through */ case OOM_SCAN_CONTINUE: continue; case OOM_SCAN_ABORT: |
72ec70299 cgroup: make task... |
1484 |
css_task_iter_end(&it); |
9cbb78bb3 mm, memcg: introd... |
1485 1486 1487 1488 1489 1490 1491 1492 |
mem_cgroup_iter_break(memcg, iter); if (chosen) put_task_struct(chosen); return; case OOM_SCAN_OK: break; }; points = oom_badness(task, memcg, NULL, totalpages); |
d49ad9355 mm, oom: prefer t... |
1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 |
if (!points || points < chosen_points) continue; /* Prefer thread group leaders for display purposes */ if (points == chosen_points && thread_group_leader(chosen)) continue; if (chosen) put_task_struct(chosen); chosen = task; chosen_points = points; get_task_struct(chosen); |
9cbb78bb3 mm, memcg: introd... |
1505 |
} |
72ec70299 cgroup: make task... |
1506 |
css_task_iter_end(&it); |
9cbb78bb3 mm, memcg: introd... |
1507 1508 1509 1510 1511 |
} if (!chosen) return; points = chosen_points * 1000 / totalpages; |
9cbb78bb3 mm, memcg: introd... |
1512 1513 |
oom_kill_process(chosen, gfp_mask, order, points, totalpages, memcg, NULL, "Memory cgroup out of memory"); |
9cbb78bb3 mm, memcg: introd... |
1514 |
} |
ae6e71d3d mm/memcontrol.c: ... |
1515 |
#if MAX_NUMNODES > 1 |
4d0c066d2 memcg: fix reclai... |
1516 1517 |
/** * test_mem_cgroup_node_reclaimable |
dad7557eb mm: fix kernel-do... |
1518 |
* @memcg: the target memcg |
4d0c066d2 memcg: fix reclai... |
1519 1520 1521 1522 1523 1524 1525 |
* @nid: the node ID to be checked. * @noswap : specify true here if the user wants flle only information. * * This function returns whether the specified memcg contains any * reclaimable pages on a node. Returns true if there are any reclaimable * pages in the node. */ |
c0ff4b854 memcg: rename mem... |
1526 |
static bool test_mem_cgroup_node_reclaimable(struct mem_cgroup *memcg, |
4d0c066d2 memcg: fix reclai... |
1527 1528 |
int nid, bool noswap) { |
c0ff4b854 memcg: rename mem... |
1529 |
if (mem_cgroup_node_nr_lru_pages(memcg, nid, LRU_ALL_FILE)) |
4d0c066d2 memcg: fix reclai... |
1530 1531 1532 |
return true; if (noswap || !total_swap_pages) return false; |
c0ff4b854 memcg: rename mem... |
1533 |
if (mem_cgroup_node_nr_lru_pages(memcg, nid, LRU_ALL_ANON)) |
4d0c066d2 memcg: fix reclai... |
1534 1535 1536 1537 |
return true; return false; } |
889976dbc memcg: reclaim me... |
1538 1539 1540 1541 1542 1543 1544 |
/* * Always updating the nodemask is not very good - even if we have an empty * list or the wrong list here, we can start from some node and traverse all * nodes based on the zonelist. So update the list loosely once per 10 secs. * */ |
c0ff4b854 memcg: rename mem... |
1545 |
static void mem_cgroup_may_update_nodemask(struct mem_cgroup *memcg) |
889976dbc memcg: reclaim me... |
1546 1547 |
{ int nid; |
453a9bf34 memcg: fix numa s... |
1548 1549 1550 1551 |
/* * numainfo_events > 0 means there was at least NUMAINFO_EVENTS_TARGET * pagein/pageout changes since the last update. */ |
c0ff4b854 memcg: rename mem... |
1552 |
if (!atomic_read(&memcg->numainfo_events)) |
453a9bf34 memcg: fix numa s... |
1553 |
return; |
c0ff4b854 memcg: rename mem... |
1554 |
if (atomic_inc_return(&memcg->numainfo_updating) > 1) |
889976dbc memcg: reclaim me... |
1555 |
return; |
889976dbc memcg: reclaim me... |
1556 |
/* make a nodemask where this memcg uses memory from */ |
31aaea4aa memcontrol: use N... |
1557 |
memcg->scan_nodes = node_states[N_MEMORY]; |
889976dbc memcg: reclaim me... |
1558 |
|
31aaea4aa memcontrol: use N... |
1559 |
for_each_node_mask(nid, node_states[N_MEMORY]) { |
889976dbc memcg: reclaim me... |
1560 |
|
c0ff4b854 memcg: rename mem... |
1561 1562 |
if (!test_mem_cgroup_node_reclaimable(memcg, nid, false)) node_clear(nid, memcg->scan_nodes); |
889976dbc memcg: reclaim me... |
1563 |
} |
453a9bf34 memcg: fix numa s... |
1564 |
|
c0ff4b854 memcg: rename mem... |
1565 1566 |
atomic_set(&memcg->numainfo_events, 0); atomic_set(&memcg->numainfo_updating, 0); |
889976dbc memcg: reclaim me... |
1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 |
} /* * Selecting a node where we start reclaim from. Because what we need is just * reducing usage counter, start from anywhere is O,K. Considering * memory reclaim from current node, there are pros. and cons. * * Freeing memory from current node means freeing memory from a node which * we'll use or we've used. So, it may make LRU bad. And if several threads * hit limits, it will see a contention on a node. But freeing from remote * node means more costs for memory reclaim because of memory latency. * * Now, we use round-robin. Better algorithm is welcomed. */ |
c0ff4b854 memcg: rename mem... |
1581 |
int mem_cgroup_select_victim_node(struct mem_cgroup *memcg) |
889976dbc memcg: reclaim me... |
1582 1583 |
{ int node; |
c0ff4b854 memcg: rename mem... |
1584 1585 |
mem_cgroup_may_update_nodemask(memcg); node = memcg->last_scanned_node; |
889976dbc memcg: reclaim me... |
1586 |
|
c0ff4b854 memcg: rename mem... |
1587 |
node = next_node(node, memcg->scan_nodes); |
889976dbc memcg: reclaim me... |
1588 |
if (node == MAX_NUMNODES) |
c0ff4b854 memcg: rename mem... |
1589 |
node = first_node(memcg->scan_nodes); |
889976dbc memcg: reclaim me... |
1590 1591 1592 1593 1594 1595 1596 1597 |
/* * We call this when we hit limit, not when pages are added to LRU. * No LRU may hold pages because all pages are UNEVICTABLE or * memcg is too small and all pages are not on LRU. In that case, * we use curret node. */ if (unlikely(node == MAX_NUMNODES)) node = numa_node_id(); |
c0ff4b854 memcg: rename mem... |
1598 |
memcg->last_scanned_node = node; |
889976dbc memcg: reclaim me... |
1599 1600 |
return node; } |
889976dbc memcg: reclaim me... |
1601 |
#else |
c0ff4b854 memcg: rename mem... |
1602 |
int mem_cgroup_select_victim_node(struct mem_cgroup *memcg) |
889976dbc memcg: reclaim me... |
1603 1604 1605 1606 |
{ return 0; } #endif |
0608f43da revert "memcg, vm... |
1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 |
static int mem_cgroup_soft_reclaim(struct mem_cgroup *root_memcg, struct zone *zone, gfp_t gfp_mask, unsigned long *total_scanned) { struct mem_cgroup *victim = NULL; int total = 0; int loop = 0; unsigned long excess; unsigned long nr_scanned; struct mem_cgroup_reclaim_cookie reclaim = { .zone = zone, .priority = 0, }; |
3e32cb2e0 mm: memcontrol: l... |
1621 |
excess = soft_limit_excess(root_memcg); |
0608f43da revert "memcg, vm... |
1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 |
while (1) { victim = mem_cgroup_iter(root_memcg, victim, &reclaim); if (!victim) { loop++; if (loop >= 2) { /* * If we have not been able to reclaim * anything, it might because there are * no reclaimable pages under this hierarchy */ if (!total) break; /* * We want to do more targeted reclaim. * excess >> 2 is not to excessive so as to * reclaim too much, nor too less that we keep * coming back to reclaim from this cgroup */ if (total >= (excess >> 2) || (loop > MEM_CGROUP_MAX_RECLAIM_LOOPS)) break; } continue; } |
0608f43da revert "memcg, vm... |
1647 1648 1649 |
total += mem_cgroup_shrink_node_zone(victim, gfp_mask, false, zone, &nr_scanned); *total_scanned += nr_scanned; |
3e32cb2e0 mm: memcontrol: l... |
1650 |
if (!soft_limit_excess(root_memcg)) |
0608f43da revert "memcg, vm... |
1651 |
break; |
6d61ef409 memcg: memory cgr... |
1652 |
} |
0608f43da revert "memcg, vm... |
1653 1654 |
mem_cgroup_iter_break(root_memcg, victim); return total; |
6d61ef409 memcg: memory cgr... |
1655 |
} |
0056f4e66 mm: memcg: lockde... |
1656 1657 1658 1659 1660 |
#ifdef CONFIG_LOCKDEP static struct lockdep_map memcg_oom_lock_dep_map = { .name = "memcg_oom_lock", }; #endif |
fb2a6fc56 mm: memcg: rework... |
1661 |
static DEFINE_SPINLOCK(memcg_oom_lock); |
867578cbc memcg: fix oom ki... |
1662 1663 1664 1665 |
/* * Check OOM-Killer is already running under our hierarchy. * If someone is running, return false. */ |
fb2a6fc56 mm: memcg: rework... |
1666 |
static bool mem_cgroup_oom_trylock(struct mem_cgroup *memcg) |
867578cbc memcg: fix oom ki... |
1667 |
{ |
79dfdaccd memcg: make oom_l... |
1668 |
struct mem_cgroup *iter, *failed = NULL; |
a636b327f memcg: avoid unne... |
1669 |
|
fb2a6fc56 mm: memcg: rework... |
1670 |
spin_lock(&memcg_oom_lock); |
9f3a0d093 mm: memcg: consol... |
1671 |
for_each_mem_cgroup_tree(iter, memcg) { |
23751be00 memcg: fix hierar... |
1672 |
if (iter->oom_lock) { |
79dfdaccd memcg: make oom_l... |
1673 1674 1675 1676 |
/* * this subtree of our hierarchy is already locked * so we cannot give a lock. */ |
79dfdaccd memcg: make oom_l... |
1677 |
failed = iter; |
9f3a0d093 mm: memcg: consol... |
1678 1679 |
mem_cgroup_iter_break(memcg, iter); break; |
23751be00 memcg: fix hierar... |
1680 1681 |
} else iter->oom_lock = true; |
7d74b06f2 memcg: use for_ea... |
1682 |
} |
867578cbc memcg: fix oom ki... |
1683 |
|
fb2a6fc56 mm: memcg: rework... |
1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 |
if (failed) { /* * OK, we failed to lock the whole subtree so we have * to clean up what we set up to the failing subtree */ for_each_mem_cgroup_tree(iter, memcg) { if (iter == failed) { mem_cgroup_iter_break(memcg, iter); break; } iter->oom_lock = false; |
79dfdaccd memcg: make oom_l... |
1695 |
} |
0056f4e66 mm: memcg: lockde... |
1696 1697 |
} else mutex_acquire(&memcg_oom_lock_dep_map, 0, 1, _RET_IP_); |
fb2a6fc56 mm: memcg: rework... |
1698 1699 1700 1701 |
spin_unlock(&memcg_oom_lock); return !failed; |
a636b327f memcg: avoid unne... |
1702 |
} |
0b7f569e4 memcg: fix OOM ki... |
1703 |
|
fb2a6fc56 mm: memcg: rework... |
1704 |
static void mem_cgroup_oom_unlock(struct mem_cgroup *memcg) |
0b7f569e4 memcg: fix OOM ki... |
1705 |
{ |
7d74b06f2 memcg: use for_ea... |
1706 |
struct mem_cgroup *iter; |
fb2a6fc56 mm: memcg: rework... |
1707 |
spin_lock(&memcg_oom_lock); |
0056f4e66 mm: memcg: lockde... |
1708 |
mutex_release(&memcg_oom_lock_dep_map, 1, _RET_IP_); |
c0ff4b854 memcg: rename mem... |
1709 |
for_each_mem_cgroup_tree(iter, memcg) |
79dfdaccd memcg: make oom_l... |
1710 |
iter->oom_lock = false; |
fb2a6fc56 mm: memcg: rework... |
1711 |
spin_unlock(&memcg_oom_lock); |
79dfdaccd memcg: make oom_l... |
1712 |
} |
c0ff4b854 memcg: rename mem... |
1713 |
static void mem_cgroup_mark_under_oom(struct mem_cgroup *memcg) |
79dfdaccd memcg: make oom_l... |
1714 1715 |
{ struct mem_cgroup *iter; |
c0ff4b854 memcg: rename mem... |
1716 |
for_each_mem_cgroup_tree(iter, memcg) |
79dfdaccd memcg: make oom_l... |
1717 1718 |
atomic_inc(&iter->under_oom); } |
c0ff4b854 memcg: rename mem... |
1719 |
static void mem_cgroup_unmark_under_oom(struct mem_cgroup *memcg) |
79dfdaccd memcg: make oom_l... |
1720 1721 |
{ struct mem_cgroup *iter; |
867578cbc memcg: fix oom ki... |
1722 1723 1724 1725 1726 |
/* * When a new child is created while the hierarchy is under oom, * mem_cgroup_oom_lock() may not be called. We have to use * atomic_add_unless() here. */ |
c0ff4b854 memcg: rename mem... |
1727 |
for_each_mem_cgroup_tree(iter, memcg) |
79dfdaccd memcg: make oom_l... |
1728 |
atomic_add_unless(&iter->under_oom, -1, 0); |
0b7f569e4 memcg: fix OOM ki... |
1729 |
} |
867578cbc memcg: fix oom ki... |
1730 |
static DECLARE_WAIT_QUEUE_HEAD(memcg_oom_waitq); |
dc98df5a1 memcg: oom wakeup... |
1731 |
struct oom_wait_info { |
d79154bb5 memcg: replace me... |
1732 |
struct mem_cgroup *memcg; |
dc98df5a1 memcg: oom wakeup... |
1733 1734 1735 1736 1737 1738 |
wait_queue_t wait; }; static int memcg_oom_wake_function(wait_queue_t *wait, unsigned mode, int sync, void *arg) { |
d79154bb5 memcg: replace me... |
1739 1740 |
struct mem_cgroup *wake_memcg = (struct mem_cgroup *)arg; struct mem_cgroup *oom_wait_memcg; |
dc98df5a1 memcg: oom wakeup... |
1741 1742 1743 |
struct oom_wait_info *oom_wait_info; oom_wait_info = container_of(wait, struct oom_wait_info, wait); |
d79154bb5 memcg: replace me... |
1744 |
oom_wait_memcg = oom_wait_info->memcg; |
dc98df5a1 memcg: oom wakeup... |
1745 |
|
2314b42db mm: memcontrol: d... |
1746 1747 |
if (!mem_cgroup_is_descendant(wake_memcg, oom_wait_memcg) && !mem_cgroup_is_descendant(oom_wait_memcg, wake_memcg)) |
dc98df5a1 memcg: oom wakeup... |
1748 |
return 0; |
dc98df5a1 memcg: oom wakeup... |
1749 1750 |
return autoremove_wake_function(wait, mode, sync, arg); } |
c0ff4b854 memcg: rename mem... |
1751 |
static void memcg_wakeup_oom(struct mem_cgroup *memcg) |
dc98df5a1 memcg: oom wakeup... |
1752 |
{ |
3812c8c8f mm: memcg: do not... |
1753 |
atomic_inc(&memcg->oom_wakeups); |
c0ff4b854 memcg: rename mem... |
1754 1755 |
/* for filtering, pass "memcg" as argument. */ __wake_up(&memcg_oom_waitq, TASK_NORMAL, 0, memcg); |
dc98df5a1 memcg: oom wakeup... |
1756 |
} |
c0ff4b854 memcg: rename mem... |
1757 |
static void memcg_oom_recover(struct mem_cgroup *memcg) |
3c11ecf44 memcg: oom kill d... |
1758 |
{ |
c0ff4b854 memcg: rename mem... |
1759 1760 |
if (memcg && atomic_read(&memcg->under_oom)) memcg_wakeup_oom(memcg); |
3c11ecf44 memcg: oom kill d... |
1761 |
} |
3812c8c8f mm: memcg: do not... |
1762 |
static void mem_cgroup_oom(struct mem_cgroup *memcg, gfp_t mask, int order) |
0b7f569e4 memcg: fix OOM ki... |
1763 |
{ |
3812c8c8f mm: memcg: do not... |
1764 1765 |
if (!current->memcg_oom.may_oom) return; |
867578cbc memcg: fix oom ki... |
1766 |
/* |
494264208 mm: memcg: handle... |
1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 |
* We are in the middle of the charge context here, so we * don't want to block when potentially sitting on a callstack * that holds all kinds of filesystem and mm locks. * * Also, the caller may handle a failed allocation gracefully * (like optional page cache readahead) and so an OOM killer * invocation might not even be necessary. * * That's why we don't do anything here except remember the * OOM context and then deal with it at the end of the page * fault when the stack is unwound, the locks are released, * and when we know whether the fault was overall successful. |
867578cbc memcg: fix oom ki... |
1779 |
*/ |
494264208 mm: memcg: handle... |
1780 1781 1782 1783 |
css_get(&memcg->css); current->memcg_oom.memcg = memcg; current->memcg_oom.gfp_mask = mask; current->memcg_oom.order = order; |
3812c8c8f mm: memcg: do not... |
1784 1785 1786 1787 |
} /** * mem_cgroup_oom_synchronize - complete memcg OOM handling |
494264208 mm: memcg: handle... |
1788 |
* @handle: actually kill/wait or just clean up the OOM state |
3812c8c8f mm: memcg: do not... |
1789 |
* |
494264208 mm: memcg: handle... |
1790 1791 |
* This has to be called at the end of a page fault if the memcg OOM * handler was enabled. |
3812c8c8f mm: memcg: do not... |
1792 |
* |
494264208 mm: memcg: handle... |
1793 |
* Memcg supports userspace OOM handling where failed allocations must |
3812c8c8f mm: memcg: do not... |
1794 1795 1796 1797 |
* sleep on a waitqueue until the userspace task resolves the * situation. Sleeping directly in the charge context with all kinds * of locks held is not a good idea, instead we remember an OOM state * in the task and mem_cgroup_oom_synchronize() has to be called at |
494264208 mm: memcg: handle... |
1798 |
* the end of the page fault to complete the OOM handling. |
3812c8c8f mm: memcg: do not... |
1799 1800 |
* * Returns %true if an ongoing memcg OOM situation was detected and |
494264208 mm: memcg: handle... |
1801 |
* completed, %false otherwise. |
3812c8c8f mm: memcg: do not... |
1802 |
*/ |
494264208 mm: memcg: handle... |
1803 |
bool mem_cgroup_oom_synchronize(bool handle) |
3812c8c8f mm: memcg: do not... |
1804 |
{ |
494264208 mm: memcg: handle... |
1805 |
struct mem_cgroup *memcg = current->memcg_oom.memcg; |
3812c8c8f mm: memcg: do not... |
1806 |
struct oom_wait_info owait; |
494264208 mm: memcg: handle... |
1807 |
bool locked; |
3812c8c8f mm: memcg: do not... |
1808 1809 |
/* OOM is global, do not handle */ |
3812c8c8f mm: memcg: do not... |
1810 |
if (!memcg) |
494264208 mm: memcg: handle... |
1811 |
return false; |
3812c8c8f mm: memcg: do not... |
1812 |
|
c32b3cbe0 oom, PM: make OOM... |
1813 |
if (!handle || oom_killer_disabled) |
494264208 mm: memcg: handle... |
1814 |
goto cleanup; |
3812c8c8f mm: memcg: do not... |
1815 1816 1817 1818 1819 1820 |
owait.memcg = memcg; owait.wait.flags = 0; owait.wait.func = memcg_oom_wake_function; owait.wait.private = current; INIT_LIST_HEAD(&owait.wait.task_list); |
867578cbc memcg: fix oom ki... |
1821 |
|
3812c8c8f mm: memcg: do not... |
1822 |
prepare_to_wait(&memcg_oom_waitq, &owait.wait, TASK_KILLABLE); |
494264208 mm: memcg: handle... |
1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 |
mem_cgroup_mark_under_oom(memcg); locked = mem_cgroup_oom_trylock(memcg); if (locked) mem_cgroup_oom_notify(memcg); if (locked && !memcg->oom_kill_disable) { mem_cgroup_unmark_under_oom(memcg); finish_wait(&memcg_oom_waitq, &owait.wait); mem_cgroup_out_of_memory(memcg, current->memcg_oom.gfp_mask, current->memcg_oom.order); } else { |
3812c8c8f mm: memcg: do not... |
1836 |
schedule(); |
494264208 mm: memcg: handle... |
1837 1838 1839 1840 1841 |
mem_cgroup_unmark_under_oom(memcg); finish_wait(&memcg_oom_waitq, &owait.wait); } if (locked) { |
fb2a6fc56 mm: memcg: rework... |
1842 1843 1844 1845 1846 1847 1848 1849 |
mem_cgroup_oom_unlock(memcg); /* * There is no guarantee that an OOM-lock contender * sees the wakeups triggered by the OOM kill * uncharges. Wake any sleepers explicitely. */ memcg_oom_recover(memcg); } |
494264208 mm: memcg: handle... |
1850 1851 |
cleanup: current->memcg_oom.memcg = NULL; |
3812c8c8f mm: memcg: do not... |
1852 |
css_put(&memcg->css); |
867578cbc memcg: fix oom ki... |
1853 |
return true; |
0b7f569e4 memcg: fix OOM ki... |
1854 |
} |
d7365e783 mm: memcontrol: f... |
1855 1856 1857 |
/** * mem_cgroup_begin_page_stat - begin a page state statistics transaction * @page: page that is going to change accounted state |
32047e2a8 memcg: avoid lock... |
1858 |
* |
d7365e783 mm: memcontrol: f... |
1859 1860 1861 |
* This function must mark the beginning of an accounted page state * change to prevent double accounting when the page is concurrently * being moved to another memcg: |
32047e2a8 memcg: avoid lock... |
1862 |
* |
6de226191 mm: memcontrol: t... |
1863 |
* memcg = mem_cgroup_begin_page_stat(page); |
d7365e783 mm: memcontrol: f... |
1864 1865 |
* if (TestClearPageState(page)) * mem_cgroup_update_page_stat(memcg, state, -1); |
6de226191 mm: memcontrol: t... |
1866 |
* mem_cgroup_end_page_stat(memcg); |
d69b042f3 memcg: add file-b... |
1867 |
*/ |
6de226191 mm: memcontrol: t... |
1868 |
struct mem_cgroup *mem_cgroup_begin_page_stat(struct page *page) |
89c06bd52 memcg: use new lo... |
1869 1870 |
{ struct mem_cgroup *memcg; |
6de226191 mm: memcontrol: t... |
1871 |
unsigned long flags; |
89c06bd52 memcg: use new lo... |
1872 |
|
6de226191 mm: memcontrol: t... |
1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 |
/* * The RCU lock is held throughout the transaction. The fast * path can get away without acquiring the memcg->move_lock * because page moving starts with an RCU grace period. * * The RCU lock also protects the memcg from being freed when * the page state that is going to change is the only thing * preventing the page from being uncharged. * E.g. end-writeback clearing PageWriteback(), which allows * migration to go ahead and uncharge the page before the * account transaction might be complete. */ |
d7365e783 mm: memcontrol: f... |
1885 1886 1887 1888 |
rcu_read_lock(); if (mem_cgroup_disabled()) return NULL; |
89c06bd52 memcg: use new lo... |
1889 |
again: |
1306a85ae mm: embed the mem... |
1890 |
memcg = page->mem_cgroup; |
298333157 mm: memcontrol: r... |
1891 |
if (unlikely(!memcg)) |
d7365e783 mm: memcontrol: f... |
1892 |
return NULL; |
bdcbb659f memcg: fold mem_c... |
1893 |
if (atomic_read(&memcg->moving_account) <= 0) |
d7365e783 mm: memcontrol: f... |
1894 |
return memcg; |
89c06bd52 memcg: use new lo... |
1895 |
|
6de226191 mm: memcontrol: t... |
1896 |
spin_lock_irqsave(&memcg->move_lock, flags); |
1306a85ae mm: embed the mem... |
1897 |
if (memcg != page->mem_cgroup) { |
6de226191 mm: memcontrol: t... |
1898 |
spin_unlock_irqrestore(&memcg->move_lock, flags); |
89c06bd52 memcg: use new lo... |
1899 1900 |
goto again; } |
6de226191 mm: memcontrol: t... |
1901 1902 1903 1904 1905 1906 1907 1908 |
/* * When charge migration first begins, we can have locked and * unlocked page stat updates happening concurrently. Track * the task who has the lock for mem_cgroup_end_page_stat(). */ memcg->move_lock_task = current; memcg->move_lock_flags = flags; |
d7365e783 mm: memcontrol: f... |
1909 1910 |
return memcg; |
89c06bd52 memcg: use new lo... |
1911 |
} |
d7365e783 mm: memcontrol: f... |
1912 1913 1914 |
/** * mem_cgroup_end_page_stat - finish a page state statistics transaction * @memcg: the memcg that was accounted against |
d7365e783 mm: memcontrol: f... |
1915 |
*/ |
6de226191 mm: memcontrol: t... |
1916 |
void mem_cgroup_end_page_stat(struct mem_cgroup *memcg) |
89c06bd52 memcg: use new lo... |
1917 |
{ |
6de226191 mm: memcontrol: t... |
1918 1919 1920 1921 1922 1923 1924 1925 |
if (memcg && memcg->move_lock_task == current) { unsigned long flags = memcg->move_lock_flags; memcg->move_lock_task = NULL; memcg->move_lock_flags = 0; spin_unlock_irqrestore(&memcg->move_lock, flags); } |
89c06bd52 memcg: use new lo... |
1926 |
|
d7365e783 mm: memcontrol: f... |
1927 |
rcu_read_unlock(); |
89c06bd52 memcg: use new lo... |
1928 |
} |
d7365e783 mm: memcontrol: f... |
1929 1930 1931 1932 1933 1934 1935 1936 1937 |
/** * 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. */ void mem_cgroup_update_page_stat(struct mem_cgroup *memcg, |
68b4876d9 memcg: remove MEM... |
1938 |
enum mem_cgroup_stat_index idx, int val) |
d69b042f3 memcg: add file-b... |
1939 |
{ |
658b72c5a memcg: check for ... |
1940 |
VM_BUG_ON(!rcu_read_lock_held()); |
26174efd4 memcg: generic fi... |
1941 |
|
d7365e783 mm: memcontrol: f... |
1942 1943 |
if (memcg) this_cpu_add(memcg->stat->count[idx], val); |
d69b042f3 memcg: add file-b... |
1944 |
} |
26174efd4 memcg: generic fi... |
1945 |
|
f817ed485 memcg: move all a... |
1946 |
/* |
cdec2e426 memcg: coalesce c... |
1947 1948 1949 |
* size of first charge trial. "32" comes from vmscan.c's magic value. * TODO: maybe necessary to use big numbers in big irons. */ |
7ec99d621 memcg: unify char... |
1950 |
#define CHARGE_BATCH 32U |
cdec2e426 memcg: coalesce c... |
1951 1952 |
struct memcg_stock_pcp { struct mem_cgroup *cached; /* this never be root cgroup */ |
11c9ea4e8 memcg: convert pe... |
1953 |
unsigned int nr_pages; |
cdec2e426 memcg: coalesce c... |
1954 |
struct work_struct work; |
26fe61684 memcg: fix percpu... |
1955 |
unsigned long flags; |
a0db00fcf memcg: remove red... |
1956 |
#define FLUSHING_CACHED_CHARGE 0 |
cdec2e426 memcg: coalesce c... |
1957 1958 |
}; static DEFINE_PER_CPU(struct memcg_stock_pcp, memcg_stock); |
9f50fad65 Revert "memcg: ge... |
1959 |
static DEFINE_MUTEX(percpu_charge_mutex); |
cdec2e426 memcg: coalesce c... |
1960 |
|
a0956d544 memcg: make it po... |
1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 |
/** * consume_stock: Try to consume stocked charge on this cpu. * @memcg: memcg to consume from. * @nr_pages: how many pages to charge. * * The charges will only happen if @memcg matches the current cpu's memcg * stock, and at least @nr_pages are available in that stock. Failure to * service an allocation will refill the stock. * * returns true if successful, false otherwise. |
cdec2e426 memcg: coalesce c... |
1971 |
*/ |
a0956d544 memcg: make it po... |
1972 |
static bool consume_stock(struct mem_cgroup *memcg, unsigned int nr_pages) |
cdec2e426 memcg: coalesce c... |
1973 1974 |
{ struct memcg_stock_pcp *stock; |
3e32cb2e0 mm: memcontrol: l... |
1975 |
bool ret = false; |
cdec2e426 memcg: coalesce c... |
1976 |
|
a0956d544 memcg: make it po... |
1977 |
if (nr_pages > CHARGE_BATCH) |
3e32cb2e0 mm: memcontrol: l... |
1978 |
return ret; |
a0956d544 memcg: make it po... |
1979 |
|
cdec2e426 memcg: coalesce c... |
1980 |
stock = &get_cpu_var(memcg_stock); |
3e32cb2e0 mm: memcontrol: l... |
1981 |
if (memcg == stock->cached && stock->nr_pages >= nr_pages) { |
a0956d544 memcg: make it po... |
1982 |
stock->nr_pages -= nr_pages; |
3e32cb2e0 mm: memcontrol: l... |
1983 1984 |
ret = true; } |
cdec2e426 memcg: coalesce c... |
1985 1986 1987 1988 1989 |
put_cpu_var(memcg_stock); return ret; } /* |
3e32cb2e0 mm: memcontrol: l... |
1990 |
* Returns stocks cached in percpu and reset cached information. |
cdec2e426 memcg: coalesce c... |
1991 1992 1993 1994 |
*/ static void drain_stock(struct memcg_stock_pcp *stock) { struct mem_cgroup *old = stock->cached; |
11c9ea4e8 memcg: convert pe... |
1995 |
if (stock->nr_pages) { |
3e32cb2e0 mm: memcontrol: l... |
1996 |
page_counter_uncharge(&old->memory, stock->nr_pages); |
cdec2e426 memcg: coalesce c... |
1997 |
if (do_swap_account) |
3e32cb2e0 mm: memcontrol: l... |
1998 |
page_counter_uncharge(&old->memsw, stock->nr_pages); |
e8ea14cc6 mm: memcontrol: t... |
1999 |
css_put_many(&old->css, stock->nr_pages); |
11c9ea4e8 memcg: convert pe... |
2000 |
stock->nr_pages = 0; |
cdec2e426 memcg: coalesce c... |
2001 2002 |
} stock->cached = NULL; |
cdec2e426 memcg: coalesce c... |
2003 2004 2005 2006 2007 2008 2009 2010 |
} /* * This must be called under preempt disabled or must be called by * a thread which is pinned to local cpu. */ static void drain_local_stock(struct work_struct *dummy) { |
7c8e0181e mm: replace __get... |
2011 |
struct memcg_stock_pcp *stock = this_cpu_ptr(&memcg_stock); |
cdec2e426 memcg: coalesce c... |
2012 |
drain_stock(stock); |
26fe61684 memcg: fix percpu... |
2013 |
clear_bit(FLUSHING_CACHED_CHARGE, &stock->flags); |
cdec2e426 memcg: coalesce c... |
2014 2015 2016 |
} /* |
3e32cb2e0 mm: memcontrol: l... |
2017 |
* Cache charges(val) to local per_cpu area. |
320cc51d9 mm: fix typo in r... |
2018 |
* This will be consumed by consume_stock() function, later. |
cdec2e426 memcg: coalesce c... |
2019 |
*/ |
c0ff4b854 memcg: rename mem... |
2020 |
static void refill_stock(struct mem_cgroup *memcg, unsigned int nr_pages) |
cdec2e426 memcg: coalesce c... |
2021 2022 |
{ struct memcg_stock_pcp *stock = &get_cpu_var(memcg_stock); |
c0ff4b854 memcg: rename mem... |
2023 |
if (stock->cached != memcg) { /* reset if necessary */ |
cdec2e426 memcg: coalesce c... |
2024 |
drain_stock(stock); |
c0ff4b854 memcg: rename mem... |
2025 |
stock->cached = memcg; |
cdec2e426 memcg: coalesce c... |
2026 |
} |
11c9ea4e8 memcg: convert pe... |
2027 |
stock->nr_pages += nr_pages; |
cdec2e426 memcg: coalesce c... |
2028 2029 2030 2031 |
put_cpu_var(memcg_stock); } /* |
c0ff4b854 memcg: rename mem... |
2032 |
* Drains all per-CPU charge caches for given root_memcg resp. subtree |
6d3d6aa22 mm: memcontrol: r... |
2033 |
* of the hierarchy under it. |
cdec2e426 memcg: coalesce c... |
2034 |
*/ |
6d3d6aa22 mm: memcontrol: r... |
2035 |
static void drain_all_stock(struct mem_cgroup *root_memcg) |
cdec2e426 memcg: coalesce c... |
2036 |
{ |
26fe61684 memcg: fix percpu... |
2037 |
int cpu, curcpu; |
d38144b7a memcg: unify sync... |
2038 |
|
6d3d6aa22 mm: memcontrol: r... |
2039 2040 2041 |
/* If someone's already draining, avoid adding running more workers. */ if (!mutex_trylock(&percpu_charge_mutex)) return; |
cdec2e426 memcg: coalesce c... |
2042 |
/* Notify other cpus that system-wide "drain" is running */ |
cdec2e426 memcg: coalesce c... |
2043 |
get_online_cpus(); |
5af12d0ef memcg: pin execut... |
2044 |
curcpu = get_cpu(); |
cdec2e426 memcg: coalesce c... |
2045 2046 |
for_each_online_cpu(cpu) { struct memcg_stock_pcp *stock = &per_cpu(memcg_stock, cpu); |
c0ff4b854 memcg: rename mem... |
2047 |
struct mem_cgroup *memcg; |
26fe61684 memcg: fix percpu... |
2048 |
|
c0ff4b854 memcg: rename mem... |
2049 2050 |
memcg = stock->cached; if (!memcg || !stock->nr_pages) |
26fe61684 memcg: fix percpu... |
2051 |
continue; |
2314b42db mm: memcontrol: d... |
2052 |
if (!mem_cgroup_is_descendant(memcg, root_memcg)) |
3e92041d6 memcg: add mem_cg... |
2053 |
continue; |
d1a05b697 memcg: do not try... |
2054 2055 2056 2057 2058 2059 |
if (!test_and_set_bit(FLUSHING_CACHED_CHARGE, &stock->flags)) { if (cpu == curcpu) drain_local_stock(&stock->work); else schedule_work_on(cpu, &stock->work); } |
cdec2e426 memcg: coalesce c... |
2060 |
} |
5af12d0ef memcg: pin execut... |
2061 |
put_cpu(); |
f894ffa86 memcg: trivial cl... |
2062 |
put_online_cpus(); |
9f50fad65 Revert "memcg: ge... |
2063 |
mutex_unlock(&percpu_charge_mutex); |
cdec2e426 memcg: coalesce c... |
2064 |
} |
711d3d2c9 memcg: cpu hotplu... |
2065 2066 2067 2068 |
/* * This function drains percpu counter value from DEAD cpu and * move it to local cpu. Note that this function can be preempted. */ |
c0ff4b854 memcg: rename mem... |
2069 |
static void mem_cgroup_drain_pcp_counter(struct mem_cgroup *memcg, int cpu) |
711d3d2c9 memcg: cpu hotplu... |
2070 2071 |
{ int i; |
c0ff4b854 memcg: rename mem... |
2072 |
spin_lock(&memcg->pcp_counter_lock); |
6104621de mm: memcg: remove... |
2073 |
for (i = 0; i < MEM_CGROUP_STAT_NSTATS; i++) { |
c0ff4b854 memcg: rename mem... |
2074 |
long x = per_cpu(memcg->stat->count[i], cpu); |
711d3d2c9 memcg: cpu hotplu... |
2075 |
|
c0ff4b854 memcg: rename mem... |
2076 2077 |
per_cpu(memcg->stat->count[i], cpu) = 0; memcg->nocpu_base.count[i] += x; |
711d3d2c9 memcg: cpu hotplu... |
2078 |
} |
e9f8974f2 memcg: break out ... |
2079 |
for (i = 0; i < MEM_CGROUP_EVENTS_NSTATS; i++) { |
c0ff4b854 memcg: rename mem... |
2080 |
unsigned long x = per_cpu(memcg->stat->events[i], cpu); |
e9f8974f2 memcg: break out ... |
2081 |
|
c0ff4b854 memcg: rename mem... |
2082 2083 |
per_cpu(memcg->stat->events[i], cpu) = 0; memcg->nocpu_base.events[i] += x; |
e9f8974f2 memcg: break out ... |
2084 |
} |
c0ff4b854 memcg: rename mem... |
2085 |
spin_unlock(&memcg->pcp_counter_lock); |
711d3d2c9 memcg: cpu hotplu... |
2086 |
} |
0db0628d9 kernel: delete __... |
2087 |
static int memcg_cpu_hotplug_callback(struct notifier_block *nb, |
cdec2e426 memcg: coalesce c... |
2088 2089 2090 2091 2092 |
unsigned long action, void *hcpu) { int cpu = (unsigned long)hcpu; struct memcg_stock_pcp *stock; |
711d3d2c9 memcg: cpu hotplu... |
2093 |
struct mem_cgroup *iter; |
cdec2e426 memcg: coalesce c... |
2094 |
|
619d094b5 memcg: simplify m... |
2095 |
if (action == CPU_ONLINE) |
1489ebad8 memcg: cpu hotplu... |
2096 |
return NOTIFY_OK; |
1489ebad8 memcg: cpu hotplu... |
2097 |
|
d833049bd memcg: fix broken... |
2098 |
if (action != CPU_DEAD && action != CPU_DEAD_FROZEN) |
cdec2e426 memcg: coalesce c... |
2099 |
return NOTIFY_OK; |
711d3d2c9 memcg: cpu hotplu... |
2100 |
|
9f3a0d093 mm: memcg: consol... |
2101 |
for_each_mem_cgroup(iter) |
711d3d2c9 memcg: cpu hotplu... |
2102 |
mem_cgroup_drain_pcp_counter(iter, cpu); |
cdec2e426 memcg: coalesce c... |
2103 2104 2105 2106 |
stock = &per_cpu(memcg_stock, cpu); drain_stock(stock); return NOTIFY_OK; } |
00501b531 mm: memcontrol: r... |
2107 2108 |
static int try_charge(struct mem_cgroup *memcg, gfp_t gfp_mask, unsigned int nr_pages) |
8a9f3ccd2 Memory controller... |
2109 |
{ |
7ec99d621 memcg: unify char... |
2110 |
unsigned int batch = max(CHARGE_BATCH, nr_pages); |
9b1306192 mm: memcontrol: r... |
2111 |
int nr_retries = MEM_CGROUP_RECLAIM_RETRIES; |
6539cc053 mm: memcontrol: f... |
2112 |
struct mem_cgroup *mem_over_limit; |
3e32cb2e0 mm: memcontrol: l... |
2113 |
struct page_counter *counter; |
6539cc053 mm: memcontrol: f... |
2114 |
unsigned long nr_reclaimed; |
b70a2a21d mm: memcontrol: f... |
2115 2116 |
bool may_swap = true; bool drained = false; |
05b843012 mm: memcontrol: u... |
2117 |
int ret = 0; |
a636b327f memcg: avoid unne... |
2118 |
|
ce00a9673 mm: memcontrol: r... |
2119 2120 |
if (mem_cgroup_is_root(memcg)) goto done; |
6539cc053 mm: memcontrol: f... |
2121 |
retry: |
b6b6cc72b memcg: do not rep... |
2122 2123 |
if (consume_stock(memcg, nr_pages)) goto done; |
8a9f3ccd2 Memory controller... |
2124 |
|
3fbe72442 mm: memcontrol: s... |
2125 |
if (!do_swap_account || |
3e32cb2e0 mm: memcontrol: l... |
2126 2127 |
!page_counter_try_charge(&memcg->memsw, batch, &counter)) { if (!page_counter_try_charge(&memcg->memory, batch, &counter)) |
6539cc053 mm: memcontrol: f... |
2128 |
goto done_restock; |
3fbe72442 mm: memcontrol: s... |
2129 |
if (do_swap_account) |
3e32cb2e0 mm: memcontrol: l... |
2130 2131 |
page_counter_uncharge(&memcg->memsw, batch); mem_over_limit = mem_cgroup_from_counter(counter, memory); |
3fbe72442 mm: memcontrol: s... |
2132 |
} else { |
3e32cb2e0 mm: memcontrol: l... |
2133 |
mem_over_limit = mem_cgroup_from_counter(counter, memsw); |
b70a2a21d mm: memcontrol: f... |
2134 |
may_swap = false; |
3fbe72442 mm: memcontrol: s... |
2135 |
} |
7a81b88cb memcg: introduce ... |
2136 |
|
6539cc053 mm: memcontrol: f... |
2137 2138 2139 2140 |
if (batch > nr_pages) { batch = nr_pages; goto retry; } |
6d61ef409 memcg: memory cgr... |
2141 |
|
06b078fc0 mm: memcontrol: r... |
2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 |
/* * Unlike in global OOM situations, memcg is not in a physical * memory shortage. Allow dying and OOM-killed tasks to * bypass the last charges so that they can exit quickly and * free their memory. */ if (unlikely(test_thread_flag(TIF_MEMDIE) || fatal_signal_pending(current) || current->flags & PF_EXITING)) goto bypass; if (unlikely(task_in_memcg_oom(current))) goto nomem; |
6539cc053 mm: memcontrol: f... |
2155 2156 |
if (!(gfp_mask & __GFP_WAIT)) goto nomem; |
4b5343346 memcg: clean up t... |
2157 |
|
241994ed8 mm: memcontrol: d... |
2158 |
mem_cgroup_events(mem_over_limit, MEMCG_MAX, 1); |
b70a2a21d mm: memcontrol: f... |
2159 2160 |
nr_reclaimed = try_to_free_mem_cgroup_pages(mem_over_limit, nr_pages, gfp_mask, may_swap); |
6539cc053 mm: memcontrol: f... |
2161 |
|
61e02c745 mm: memcontrol: c... |
2162 |
if (mem_cgroup_margin(mem_over_limit) >= nr_pages) |
6539cc053 mm: memcontrol: f... |
2163 |
goto retry; |
28c34c291 mm: memcontrol: r... |
2164 |
|
b70a2a21d mm: memcontrol: f... |
2165 |
if (!drained) { |
6d3d6aa22 mm: memcontrol: r... |
2166 |
drain_all_stock(mem_over_limit); |
b70a2a21d mm: memcontrol: f... |
2167 2168 2169 |
drained = true; goto retry; } |
28c34c291 mm: memcontrol: r... |
2170 2171 |
if (gfp_mask & __GFP_NORETRY) goto nomem; |
6539cc053 mm: memcontrol: f... |
2172 2173 2174 2175 2176 2177 2178 2179 2180 |
/* * Even though the limit is exceeded at this point, reclaim * may have been able to free some pages. Retry the charge * before killing the task. * * Only for regular pages, though: huge pages are rather * unlikely to succeed so close to the limit, and we fall back * to regular pages anyway in case of failure. */ |
61e02c745 mm: memcontrol: c... |
2181 |
if (nr_reclaimed && nr_pages <= (1 << PAGE_ALLOC_COSTLY_ORDER)) |
6539cc053 mm: memcontrol: f... |
2182 2183 2184 2185 2186 2187 2188 |
goto retry; /* * At task move, charge accounts can be doubly counted. So, it's * better to wait until the end of task_move if something is going on. */ if (mem_cgroup_wait_acct_move(mem_over_limit)) goto retry; |
9b1306192 mm: memcontrol: r... |
2189 2190 |
if (nr_retries--) goto retry; |
06b078fc0 mm: memcontrol: r... |
2191 2192 |
if (gfp_mask & __GFP_NOFAIL) goto bypass; |
6539cc053 mm: memcontrol: f... |
2193 2194 |
if (fatal_signal_pending(current)) goto bypass; |
241994ed8 mm: memcontrol: d... |
2195 |
mem_cgroup_events(mem_over_limit, MEMCG_OOM, 1); |
61e02c745 mm: memcontrol: c... |
2196 |
mem_cgroup_oom(mem_over_limit, gfp_mask, get_order(nr_pages)); |
7a81b88cb memcg: introduce ... |
2197 |
nomem: |
6d1fdc489 memcg: sanitize _... |
2198 |
if (!(gfp_mask & __GFP_NOFAIL)) |
3168ecbe1 mm: memcg: use pr... |
2199 |
return -ENOMEM; |
867578cbc memcg: fix oom ki... |
2200 |
bypass: |
ce00a9673 mm: memcontrol: r... |
2201 |
return -EINTR; |
6539cc053 mm: memcontrol: f... |
2202 2203 |
done_restock: |
e8ea14cc6 mm: memcontrol: t... |
2204 |
css_get_many(&memcg->css, batch); |
6539cc053 mm: memcontrol: f... |
2205 2206 |
if (batch > nr_pages) refill_stock(memcg, batch - nr_pages); |
7d638093d memcg: do not cal... |
2207 2208 |
if (!(gfp_mask & __GFP_WAIT)) goto done; |
241994ed8 mm: memcontrol: d... |
2209 2210 2211 2212 2213 2214 2215 2216 2217 2218 |
/* * If the hierarchy is above the normal consumption range, * make the charging task trim their excess contribution. */ do { if (page_counter_read(&memcg->memory) <= memcg->high) continue; mem_cgroup_events(memcg, MEMCG_HIGH, 1); try_to_free_mem_cgroup_pages(memcg, nr_pages, gfp_mask, true); } while ((memcg = parent_mem_cgroup(memcg))); |
6539cc053 mm: memcontrol: f... |
2219 |
done: |
05b843012 mm: memcontrol: u... |
2220 |
return ret; |
7a81b88cb memcg: introduce ... |
2221 |
} |
8a9f3ccd2 Memory controller... |
2222 |
|
00501b531 mm: memcontrol: r... |
2223 |
static void cancel_charge(struct mem_cgroup *memcg, unsigned int nr_pages) |
a3032a2c1 memcg: add mem_cg... |
2224 |
{ |
ce00a9673 mm: memcontrol: r... |
2225 2226 |
if (mem_cgroup_is_root(memcg)) return; |
3e32cb2e0 mm: memcontrol: l... |
2227 |
page_counter_uncharge(&memcg->memory, nr_pages); |
05b843012 mm: memcontrol: u... |
2228 |
if (do_swap_account) |
3e32cb2e0 mm: memcontrol: l... |
2229 |
page_counter_uncharge(&memcg->memsw, nr_pages); |
ce00a9673 mm: memcontrol: r... |
2230 |
|
e8ea14cc6 mm: memcontrol: t... |
2231 |
css_put_many(&memcg->css, nr_pages); |
d01dd17f1 memcg: use res_co... |
2232 2233 2234 |
} /* |
0a31bc97c mm: memcontrol: r... |
2235 2236 2237 2238 2239 2240 2241 2242 2243 |
* try_get_mem_cgroup_from_page - look up page's memcg association * @page: the page * * Look up, get a css reference, and return the memcg that owns @page. * * The page must be locked to prevent racing with swap-in and page * cache charges. If coming from an unlocked page table, the caller * must ensure the page is on the LRU or this can race with charging. */ |
e42d9d5d4 memcg: rename and... |
2244 |
struct mem_cgroup *try_get_mem_cgroup_from_page(struct page *page) |
b5a84319a memcg: fix shmem'... |
2245 |
{ |
298333157 mm: memcontrol: r... |
2246 |
struct mem_cgroup *memcg; |
a3b2d6926 cgroups: use css ... |
2247 |
unsigned short id; |
b5a84319a memcg: fix shmem'... |
2248 |
swp_entry_t ent; |
309381fea mm: dump page whe... |
2249 |
VM_BUG_ON_PAGE(!PageLocked(page), page); |
3c776e646 memcg: charge swa... |
2250 |
|
1306a85ae mm: embed the mem... |
2251 |
memcg = page->mem_cgroup; |
298333157 mm: memcontrol: r... |
2252 2253 |
if (memcg) { if (!css_tryget_online(&memcg->css)) |
c0ff4b854 memcg: rename mem... |
2254 |
memcg = NULL; |
e42d9d5d4 memcg: rename and... |
2255 |
} else if (PageSwapCache(page)) { |
3c776e646 memcg: charge swa... |
2256 |
ent.val = page_private(page); |
9fb4b7cc0 page_cgroup: add ... |
2257 |
id = lookup_swap_cgroup_id(ent); |
a3b2d6926 cgroups: use css ... |
2258 |
rcu_read_lock(); |
adbe427b9 memcg: zap mem_cg... |
2259 |
memcg = mem_cgroup_from_id(id); |
ec903c0c8 cgroup: rename cs... |
2260 |
if (memcg && !css_tryget_online(&memcg->css)) |
c0ff4b854 memcg: rename mem... |
2261 |
memcg = NULL; |
a3b2d6926 cgroups: use css ... |
2262 |
rcu_read_unlock(); |
3c776e646 memcg: charge swa... |
2263 |
} |
c0ff4b854 memcg: rename mem... |
2264 |
return memcg; |
b5a84319a memcg: fix shmem'... |
2265 |
} |
0a31bc97c mm: memcontrol: r... |
2266 2267 2268 2269 2270 2271 2272 2273 2274 2275 2276 2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 |
static void lock_page_lru(struct page *page, int *isolated) { struct zone *zone = page_zone(page); spin_lock_irq(&zone->lru_lock); if (PageLRU(page)) { struct lruvec *lruvec; lruvec = mem_cgroup_page_lruvec(page, zone); ClearPageLRU(page); del_page_from_lru_list(page, lruvec, page_lru(page)); *isolated = 1; } else *isolated = 0; } static void unlock_page_lru(struct page *page, int isolated) { struct zone *zone = page_zone(page); if (isolated) { struct lruvec *lruvec; lruvec = mem_cgroup_page_lruvec(page, zone); VM_BUG_ON_PAGE(PageLRU(page), page); SetPageLRU(page); add_page_to_lru_list(page, lruvec, page_lru(page)); } spin_unlock_irq(&zone->lru_lock); } |
00501b531 mm: memcontrol: r... |
2296 |
static void commit_charge(struct page *page, struct mem_cgroup *memcg, |
6abb5a867 mm: memcontrol: a... |
2297 |
bool lrucare) |
7a81b88cb memcg: introduce ... |
2298 |
{ |
0a31bc97c mm: memcontrol: r... |
2299 |
int isolated; |
9ce70c024 memcg: fix deadlo... |
2300 |
|
1306a85ae mm: embed the mem... |
2301 |
VM_BUG_ON_PAGE(page->mem_cgroup, page); |
9ce70c024 memcg: fix deadlo... |
2302 2303 2304 2305 2306 |
/* * In some cases, SwapCache and FUSE(splice_buf->radixtree), the page * may already be on some other mem_cgroup's LRU. Take care of it. */ |
0a31bc97c mm: memcontrol: r... |
2307 2308 |
if (lrucare) lock_page_lru(page, &isolated); |
9ce70c024 memcg: fix deadlo... |
2309 |
|
0a31bc97c mm: memcontrol: r... |
2310 2311 |
/* * Nobody should be changing or seriously looking at |
1306a85ae mm: embed the mem... |
2312 |
* page->mem_cgroup at this point: |
0a31bc97c mm: memcontrol: r... |
2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 |
* * - the page is uncharged * * - the page is off-LRU * * - an anonymous fault has exclusive page access, except for * a locked page table * * - a page cache insertion, a swapin fault, or a migration * have the page locked */ |
1306a85ae mm: embed the mem... |
2324 |
page->mem_cgroup = memcg; |
9ce70c024 memcg: fix deadlo... |
2325 |
|
0a31bc97c mm: memcontrol: r... |
2326 2327 |
if (lrucare) unlock_page_lru(page, isolated); |
7a81b88cb memcg: introduce ... |
2328 |
} |
66e1707bc Memory controller... |
2329 |
|
7ae1e1d0f memcg: kmem contr... |
2330 |
#ifdef CONFIG_MEMCG_KMEM |
dbf22eb6d memcg: zap __memc... |
2331 2332 |
int memcg_charge_kmem(struct mem_cgroup *memcg, gfp_t gfp, unsigned long nr_pages) |
7ae1e1d0f memcg: kmem contr... |
2333 |
{ |
3e32cb2e0 mm: memcontrol: l... |
2334 |
struct page_counter *counter; |
7ae1e1d0f memcg: kmem contr... |
2335 |
int ret = 0; |
7ae1e1d0f memcg: kmem contr... |
2336 |
|
3e32cb2e0 mm: memcontrol: l... |
2337 2338 |
ret = page_counter_try_charge(&memcg->kmem, nr_pages, &counter); if (ret < 0) |
7ae1e1d0f memcg: kmem contr... |
2339 |
return ret; |
3e32cb2e0 mm: memcontrol: l... |
2340 |
ret = try_charge(memcg, gfp, nr_pages); |
7ae1e1d0f memcg: kmem contr... |
2341 2342 |
if (ret == -EINTR) { /* |
00501b531 mm: memcontrol: r... |
2343 2344 2345 2346 2347 2348 |
* try_charge() chose to bypass to root due to OOM kill or * fatal signal. Since our only options are to either fail * the allocation or charge it to this cgroup, do it as a * temporary condition. But we can't fail. From a kmem/slab * perspective, the cache has already been selected, by * mem_cgroup_kmem_get_cache(), so it is too late to change |
7ae1e1d0f memcg: kmem contr... |
2349 2350 2351 |
* our minds. * * This condition will only trigger if the task entered |
00501b531 mm: memcontrol: r... |
2352 2353 2354 |
* memcg_charge_kmem in a sane state, but was OOM-killed * during try_charge() above. Tasks that were already dying * when the allocation triggers should have been already |
7ae1e1d0f memcg: kmem contr... |
2355 2356 |
* directed to the root cgroup in memcontrol.h */ |
3e32cb2e0 mm: memcontrol: l... |
2357 |
page_counter_charge(&memcg->memory, nr_pages); |
7ae1e1d0f memcg: kmem contr... |
2358 |
if (do_swap_account) |
3e32cb2e0 mm: memcontrol: l... |
2359 |
page_counter_charge(&memcg->memsw, nr_pages); |
e8ea14cc6 mm: memcontrol: t... |
2360 |
css_get_many(&memcg->css, nr_pages); |
7ae1e1d0f memcg: kmem contr... |
2361 2362 |
ret = 0; } else if (ret) |
3e32cb2e0 mm: memcontrol: l... |
2363 |
page_counter_uncharge(&memcg->kmem, nr_pages); |
7ae1e1d0f memcg: kmem contr... |
2364 2365 2366 |
return ret; } |
dbf22eb6d memcg: zap __memc... |
2367 |
void memcg_uncharge_kmem(struct mem_cgroup *memcg, unsigned long nr_pages) |
7ae1e1d0f memcg: kmem contr... |
2368 |
{ |
3e32cb2e0 mm: memcontrol: l... |
2369 |
page_counter_uncharge(&memcg->memory, nr_pages); |
7ae1e1d0f memcg: kmem contr... |
2370 |
if (do_swap_account) |
3e32cb2e0 mm: memcontrol: l... |
2371 |
page_counter_uncharge(&memcg->memsw, nr_pages); |
7de37682b memcg: kmem accou... |
2372 |
|
64f219938 mm: memcontrol: r... |
2373 |
page_counter_uncharge(&memcg->kmem, nr_pages); |
7de37682b memcg: kmem accou... |
2374 |
|
e8ea14cc6 mm: memcontrol: t... |
2375 |
css_put_many(&memcg->css, nr_pages); |
7ae1e1d0f memcg: kmem contr... |
2376 |
} |
2633d7a02 slab/slub: consid... |
2377 2378 2379 2380 2381 2382 2383 2384 2385 |
/* * 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. */ int memcg_cache_id(struct mem_cgroup *memcg) { return memcg ? memcg->kmemcg_id : -1; } |
f3bb3043a memcg: don't call... |
2386 |
static int memcg_alloc_cache_id(void) |
55007d849 memcg: allocate m... |
2387 |
{ |
f3bb3043a memcg: don't call... |
2388 2389 |
int id, size; int err; |
dbcf73e26 memcg: rename som... |
2390 |
id = ida_simple_get(&memcg_cache_ida, |
f3bb3043a memcg: don't call... |
2391 2392 2393 |
0, MEMCG_CACHES_MAX_SIZE, GFP_KERNEL); if (id < 0) return id; |
55007d849 memcg: allocate m... |
2394 |
|
dbcf73e26 memcg: rename som... |
2395 |
if (id < memcg_nr_cache_ids) |
f3bb3043a memcg: don't call... |
2396 2397 2398 2399 2400 2401 |
return id; /* * There's no space for the new id in memcg_caches arrays, * so we have to grow them. */ |
05257a1a3 memcg: add rwsem ... |
2402 |
down_write(&memcg_cache_ids_sem); |
f3bb3043a memcg: don't call... |
2403 2404 |
size = 2 * (id + 1); |
55007d849 memcg: allocate m... |
2405 2406 2407 2408 |
if (size < MEMCG_CACHES_MIN_SIZE) size = MEMCG_CACHES_MIN_SIZE; else if (size > MEMCG_CACHES_MAX_SIZE) size = MEMCG_CACHES_MAX_SIZE; |
f3bb3043a memcg: don't call... |
2409 |
err = memcg_update_all_caches(size); |
05257a1a3 memcg: add rwsem ... |
2410 |
if (!err) |
60d3fd32a list_lru: introdu... |
2411 2412 |
err = memcg_update_all_list_lrus(size); if (!err) |
05257a1a3 memcg: add rwsem ... |
2413 2414 2415 |
memcg_nr_cache_ids = size; up_write(&memcg_cache_ids_sem); |
f3bb3043a memcg: don't call... |
2416 |
if (err) { |
dbcf73e26 memcg: rename som... |
2417 |
ida_simple_remove(&memcg_cache_ida, id); |
f3bb3043a memcg: don't call... |
2418 2419 2420 2421 2422 2423 2424 |
return err; } return id; } static void memcg_free_cache_id(int id) { |
dbcf73e26 memcg: rename som... |
2425 |
ida_simple_remove(&memcg_cache_ida, id); |
55007d849 memcg: allocate m... |
2426 |
} |
d5b3cf713 memcg: zap memcg_... |
2427 |
struct memcg_kmem_cache_create_work { |
5722d094a memcg, slab: clea... |
2428 2429 2430 2431 |
struct mem_cgroup *memcg; struct kmem_cache *cachep; struct work_struct work; }; |
d5b3cf713 memcg: zap memcg_... |
2432 |
static void memcg_kmem_cache_create_func(struct work_struct *w) |
d7f25f8a2 memcg: infrastruc... |
2433 |
{ |
d5b3cf713 memcg: zap memcg_... |
2434 2435 |
struct memcg_kmem_cache_create_work *cw = container_of(w, struct memcg_kmem_cache_create_work, work); |
5722d094a memcg, slab: clea... |
2436 2437 |
struct mem_cgroup *memcg = cw->memcg; struct kmem_cache *cachep = cw->cachep; |
d7f25f8a2 memcg: infrastruc... |
2438 |
|
d5b3cf713 memcg: zap memcg_... |
2439 |
memcg_create_kmem_cache(memcg, cachep); |
bd6731458 memcg, slab: simp... |
2440 |
|
5722d094a memcg, slab: clea... |
2441 |
css_put(&memcg->css); |
d7f25f8a2 memcg: infrastruc... |
2442 2443 2444 2445 2446 |
kfree(cw); } /* * Enqueue the creation of a per-memcg kmem_cache. |
d7f25f8a2 memcg: infrastruc... |
2447 |
*/ |
d5b3cf713 memcg: zap memcg_... |
2448 2449 |
static void __memcg_schedule_kmem_cache_create(struct mem_cgroup *memcg, struct kmem_cache *cachep) |
d7f25f8a2 memcg: infrastruc... |
2450 |
{ |
d5b3cf713 memcg: zap memcg_... |
2451 |
struct memcg_kmem_cache_create_work *cw; |
d7f25f8a2 memcg: infrastruc... |
2452 |
|
776ed0f03 memcg: cleanup km... |
2453 |
cw = kmalloc(sizeof(*cw), GFP_NOWAIT); |
8135be5a8 memcg: fix possib... |
2454 |
if (!cw) |
d7f25f8a2 memcg: infrastruc... |
2455 |
return; |
8135be5a8 memcg: fix possib... |
2456 2457 |
css_get(&memcg->css); |
d7f25f8a2 memcg: infrastruc... |
2458 2459 2460 |
cw->memcg = memcg; cw->cachep = cachep; |
d5b3cf713 memcg: zap memcg_... |
2461 |
INIT_WORK(&cw->work, memcg_kmem_cache_create_func); |
d7f25f8a2 memcg: infrastruc... |
2462 |
|
d7f25f8a2 memcg: infrastruc... |
2463 2464 |
schedule_work(&cw->work); } |
d5b3cf713 memcg: zap memcg_... |
2465 2466 |
static void memcg_schedule_kmem_cache_create(struct mem_cgroup *memcg, struct kmem_cache *cachep) |
0e9d92f2d memcg: skip memcg... |
2467 2468 2469 2470 |
{ /* * We need to stop accounting when we kmalloc, because if the * corresponding kmalloc cache is not yet created, the first allocation |
d5b3cf713 memcg: zap memcg_... |
2471 |
* in __memcg_schedule_kmem_cache_create will recurse. |
0e9d92f2d memcg: skip memcg... |
2472 2473 2474 2475 2476 2477 2478 |
* * However, it is better to enclose the whole function. Depending on * the debugging options enabled, INIT_WORK(), for instance, can * trigger an allocation. This too, will make us recurse. Because at * this point we can't allow ourselves back into memcg_kmem_get_cache, * the safest choice is to do it like this, wrapping the whole function. */ |
6f185c290 memcg: turn memcg... |
2479 |
current->memcg_kmem_skip_account = 1; |
d5b3cf713 memcg: zap memcg_... |
2480 |
__memcg_schedule_kmem_cache_create(memcg, cachep); |
6f185c290 memcg: turn memcg... |
2481 |
current->memcg_kmem_skip_account = 0; |
0e9d92f2d memcg: skip memcg... |
2482 |
} |
c67a8a685 memcg, slab: merg... |
2483 |
|
d7f25f8a2 memcg: infrastruc... |
2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 2496 |
/* * Return the kmem_cache we're supposed to use for a slab allocation. * We try to use the current memcg's version of the cache. * * If the cache does not exist yet, if we are the first user of it, * we either create it immediately, if possible, or create it asynchronously * in a workqueue. * In the latter case, we will let the current allocation go through with * the original cache. * * Can't be called in interrupt context or from kernel threads. * This function needs to be called with rcu_read_lock() held. */ |
056b7ccef mm/memcontrol.c: ... |
2497 |
struct kmem_cache *__memcg_kmem_get_cache(struct kmem_cache *cachep) |
d7f25f8a2 memcg: infrastruc... |
2498 2499 |
{ struct mem_cgroup *memcg; |
959c8963f memcg, slab: fix ... |
2500 |
struct kmem_cache *memcg_cachep; |
2a4db7eb9 memcg: free memcg... |
2501 |
int kmemcg_id; |
d7f25f8a2 memcg: infrastruc... |
2502 |
|
f7ce3190c slab: embed memcg... |
2503 |
VM_BUG_ON(!is_root_cache(cachep)); |
d7f25f8a2 memcg: infrastruc... |
2504 |
|
9d100c5e4 memcg: don't chec... |
2505 |
if (current->memcg_kmem_skip_account) |
0e9d92f2d memcg: skip memcg... |
2506 |
return cachep; |
8135be5a8 memcg: fix possib... |
2507 |
memcg = get_mem_cgroup_from_mm(current->mm); |
4db0c3c29 mm: remove rest o... |
2508 |
kmemcg_id = READ_ONCE(memcg->kmemcg_id); |
2a4db7eb9 memcg: free memcg... |
2509 |
if (kmemcg_id < 0) |
ca0dde971 memcg: take refer... |
2510 |
goto out; |
d7f25f8a2 memcg: infrastruc... |
2511 |
|
2a4db7eb9 memcg: free memcg... |
2512 |
memcg_cachep = cache_from_memcg_idx(cachep, kmemcg_id); |
8135be5a8 memcg: fix possib... |
2513 2514 |
if (likely(memcg_cachep)) return memcg_cachep; |
ca0dde971 memcg: take refer... |
2515 2516 2517 2518 2519 2520 2521 2522 2523 |
/* * If we are in a safe context (can wait, and not in interrupt * context), we could be be predictable and return right away. * This would guarantee that the allocation being performed * already belongs in the new cache. * * However, there are some clashes that can arrive from locking. * For instance, because we acquire the slab_mutex while doing |
776ed0f03 memcg: cleanup km... |
2524 2525 2526 |
* memcg_create_kmem_cache, this means no further allocation * could happen with the slab_mutex held. So it's better to * defer everything. |
ca0dde971 memcg: take refer... |
2527 |
*/ |
d5b3cf713 memcg: zap memcg_... |
2528 |
memcg_schedule_kmem_cache_create(memcg, cachep); |
ca0dde971 memcg: take refer... |
2529 |
out: |
8135be5a8 memcg: fix possib... |
2530 |
css_put(&memcg->css); |
ca0dde971 memcg: take refer... |
2531 |
return cachep; |
d7f25f8a2 memcg: infrastruc... |
2532 |
} |
d7f25f8a2 memcg: infrastruc... |
2533 |
|
8135be5a8 memcg: fix possib... |
2534 2535 2536 |
void __memcg_kmem_put_cache(struct kmem_cache *cachep) { if (!is_root_cache(cachep)) |
f7ce3190c slab: embed memcg... |
2537 |
css_put(&cachep->memcg_params.memcg->css); |
8135be5a8 memcg: fix possib... |
2538 |
} |
7ae1e1d0f memcg: kmem contr... |
2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 2552 2553 2554 2555 2556 2557 2558 2559 |
/* * We need to verify if the allocation against current->mm->owner's memcg is * possible for the given order. But the page is not allocated yet, so we'll * need a further commit step to do the final arrangements. * * It is possible for the task to switch cgroups in this mean time, so at * commit time, we can't rely on task conversion any longer. We'll then use * the handle argument to return to the caller which cgroup we should commit * against. We could also return the memcg directly and avoid the pointer * passing, but a boolean return value gives better semantics considering * the compiled-out case as well. * * Returning true means the allocation is possible. */ bool __memcg_kmem_newpage_charge(gfp_t gfp, struct mem_cgroup **_memcg, int order) { struct mem_cgroup *memcg; int ret; *_memcg = NULL; |
6d42c232b memcg: also test ... |
2560 |
|
df3819754 memcg: get_mem_cg... |
2561 |
memcg = get_mem_cgroup_from_mm(current->mm); |
7ae1e1d0f memcg: kmem contr... |
2562 |
|
cf2b8fbf1 memcg: zap memcg_... |
2563 |
if (!memcg_kmem_is_active(memcg)) { |
7ae1e1d0f memcg: kmem contr... |
2564 2565 2566 |
css_put(&memcg->css); return true; } |
3e32cb2e0 mm: memcontrol: l... |
2567 |
ret = memcg_charge_kmem(memcg, gfp, 1 << order); |
7ae1e1d0f memcg: kmem contr... |
2568 2569 |
if (!ret) *_memcg = memcg; |
7ae1e1d0f memcg: kmem contr... |
2570 2571 2572 2573 2574 2575 2576 2577 |
css_put(&memcg->css); return (ret == 0); } void __memcg_kmem_commit_charge(struct page *page, struct mem_cgroup *memcg, int order) { |
7ae1e1d0f memcg: kmem contr... |
2578 2579 2580 2581 |
VM_BUG_ON(mem_cgroup_is_root(memcg)); /* The page allocation failed. Revert */ if (!page) { |
3e32cb2e0 mm: memcontrol: l... |
2582 |
memcg_uncharge_kmem(memcg, 1 << order); |
7ae1e1d0f memcg: kmem contr... |
2583 2584 |
return; } |
1306a85ae mm: embed the mem... |
2585 |
page->mem_cgroup = memcg; |
7ae1e1d0f memcg: kmem contr... |
2586 2587 2588 2589 |
} void __memcg_kmem_uncharge_pages(struct page *page, int order) { |
1306a85ae mm: embed the mem... |
2590 |
struct mem_cgroup *memcg = page->mem_cgroup; |
7ae1e1d0f memcg: kmem contr... |
2591 |
|
7ae1e1d0f memcg: kmem contr... |
2592 2593 |
if (!memcg) return; |
309381fea mm: dump page whe... |
2594 |
VM_BUG_ON_PAGE(mem_cgroup_is_root(memcg), page); |
298333157 mm: memcontrol: r... |
2595 |
|
3e32cb2e0 mm: memcontrol: l... |
2596 |
memcg_uncharge_kmem(memcg, 1 << order); |
1306a85ae mm: embed the mem... |
2597 |
page->mem_cgroup = NULL; |
7ae1e1d0f memcg: kmem contr... |
2598 |
} |
60d3fd32a list_lru: introdu... |
2599 2600 2601 2602 2603 2604 2605 2606 2607 2608 2609 |
struct mem_cgroup *__mem_cgroup_from_kmem(void *ptr) { struct mem_cgroup *memcg = NULL; struct kmem_cache *cachep; struct page *page; page = virt_to_head_page(ptr); if (PageSlab(page)) { cachep = page->slab_cache; if (!is_root_cache(cachep)) |
f7ce3190c slab: embed memcg... |
2610 |
memcg = cachep->memcg_params.memcg; |
60d3fd32a list_lru: introdu... |
2611 2612 2613 2614 2615 2616 |
} else /* page allocated by alloc_kmem_pages */ memcg = page->mem_cgroup; return memcg; } |
7ae1e1d0f memcg: kmem contr... |
2617 |
#endif /* CONFIG_MEMCG_KMEM */ |
ca3e02141 memcg: fix USED b... |
2618 |
#ifdef CONFIG_TRANSPARENT_HUGEPAGE |
ca3e02141 memcg: fix USED b... |
2619 2620 |
/* * Because tail pages are not marked as "used", set it. We're under |
e94c8a9cb memcg: make mem_c... |
2621 2622 2623 |
* zone->lru_lock, 'splitting on pmd' and compound_lock. * charge/uncharge will be never happen and move_account() is done under * compound_lock(), so we don't have to take care of races. |
ca3e02141 memcg: fix USED b... |
2624 |
*/ |
e94c8a9cb memcg: make mem_c... |
2625 |
void mem_cgroup_split_huge_fixup(struct page *head) |
ca3e02141 memcg: fix USED b... |
2626 |
{ |
e94c8a9cb memcg: make mem_c... |
2627 |
int i; |
ca3e02141 memcg: fix USED b... |
2628 |
|
3d37c4a91 memcg: bugfix che... |
2629 2630 |
if (mem_cgroup_disabled()) return; |
b070e65c0 mm, memcg: add rs... |
2631 |
|
298333157 mm: memcontrol: r... |
2632 |
for (i = 1; i < HPAGE_PMD_NR; i++) |
1306a85ae mm: embed the mem... |
2633 |
head[i].mem_cgroup = head->mem_cgroup; |
b9982f8d2 mm: memcontrol: m... |
2634 |
|
1306a85ae mm: embed the mem... |
2635 |
__this_cpu_sub(head->mem_cgroup->stat->count[MEM_CGROUP_STAT_RSS_HUGE], |
b070e65c0 mm, memcg: add rs... |
2636 |
HPAGE_PMD_NR); |
ca3e02141 memcg: fix USED b... |
2637 |
} |
12d271078 memcg: fix split_... |
2638 |
#endif /* CONFIG_TRANSPARENT_HUGEPAGE */ |
ca3e02141 memcg: fix USED b... |
2639 |
|
c255a4580 memcg: rename con... |
2640 |
#ifdef CONFIG_MEMCG_SWAP |
0a31bc97c mm: memcontrol: r... |
2641 2642 |
static void mem_cgroup_swap_statistics(struct mem_cgroup *memcg, bool charge) |
d13d14430 memcg: handle swa... |
2643 |
{ |
0a31bc97c mm: memcontrol: r... |
2644 2645 |
int val = (charge) ? 1 : -1; this_cpu_add(memcg->stat->count[MEM_CGROUP_STAT_SWAP], val); |
d13d14430 memcg: handle swa... |
2646 |
} |
024914477 memcg: move charg... |
2647 2648 2649 2650 2651 2652 2653 2654 2655 2656 2657 2658 |
/** * mem_cgroup_move_swap_account - move swap charge and swap_cgroup's record. * @entry: swap entry to be moved * @from: mem_cgroup which the entry is moved from * @to: mem_cgroup which the entry is moved to * * It succeeds only when the swap_cgroup's record for this entry is the same * as the mem_cgroup's id of @from. * * Returns 0 on success, -EINVAL on failure. * |
3e32cb2e0 mm: memcontrol: l... |
2659 |
* The caller must have charged to @to, IOW, called page_counter_charge() about |
024914477 memcg: move charg... |
2660 2661 2662 |
* both res and memsw, and called css_get(). */ static int mem_cgroup_move_swap_account(swp_entry_t entry, |
e91cbb425 memcg swap: mem_c... |
2663 |
struct mem_cgroup *from, struct mem_cgroup *to) |
024914477 memcg: move charg... |
2664 2665 |
{ unsigned short old_id, new_id; |
34c00c319 memcg: convert to... |
2666 2667 |
old_id = mem_cgroup_id(from); new_id = mem_cgroup_id(to); |
024914477 memcg: move charg... |
2668 2669 |
if (swap_cgroup_cmpxchg(entry, old_id, new_id) == old_id) { |
024914477 memcg: move charg... |
2670 |
mem_cgroup_swap_statistics(from, false); |
483c30b51 memcg: improve pe... |
2671 |
mem_cgroup_swap_statistics(to, true); |
024914477 memcg: move charg... |
2672 2673 2674 2675 2676 2677 |
return 0; } return -EINVAL; } #else static inline int mem_cgroup_move_swap_account(swp_entry_t entry, |
e91cbb425 memcg swap: mem_c... |
2678 |
struct mem_cgroup *from, struct mem_cgroup *to) |
024914477 memcg: move charg... |
2679 2680 2681 |
{ return -EINVAL; } |
8c7c6e34a memcg: mem+swap c... |
2682 |
#endif |
d13d14430 memcg: handle swa... |
2683 |
|
3e32cb2e0 mm: memcontrol: l... |
2684 |
static DEFINE_MUTEX(memcg_limit_mutex); |
f212ad7cf memcg: add memcg ... |
2685 |
|
d38d2a758 mm: make mem_cgro... |
2686 |
static int mem_cgroup_resize_limit(struct mem_cgroup *memcg, |
3e32cb2e0 mm: memcontrol: l... |
2687 |
unsigned long limit) |
628f42355 memcg: limit chan... |
2688 |
{ |
3e32cb2e0 mm: memcontrol: l... |
2689 2690 2691 |
unsigned long curusage; unsigned long oldusage; bool enlarge = false; |
81d39c20f memcg: fix shrink... |
2692 |
int retry_count; |
3e32cb2e0 mm: memcontrol: l... |
2693 |
int ret; |
81d39c20f memcg: fix shrink... |
2694 2695 2696 2697 2698 2699 |
/* * For keeping hierarchical_reclaim simple, how long we should retry * is depends on callers. We set our retry-count to be function * of # of children which we should visit in this loop. */ |
3e32cb2e0 mm: memcontrol: l... |
2700 2701 |
retry_count = MEM_CGROUP_RECLAIM_RETRIES * mem_cgroup_count_children(memcg); |
81d39c20f memcg: fix shrink... |
2702 |
|
3e32cb2e0 mm: memcontrol: l... |
2703 |
oldusage = page_counter_read(&memcg->memory); |
628f42355 memcg: limit chan... |
2704 |
|
3e32cb2e0 mm: memcontrol: l... |
2705 |
do { |
628f42355 memcg: limit chan... |
2706 2707 2708 2709 |
if (signal_pending(current)) { ret = -EINTR; break; } |
3e32cb2e0 mm: memcontrol: l... |
2710 2711 2712 2713 |
mutex_lock(&memcg_limit_mutex); if (limit > memcg->memsw.limit) { mutex_unlock(&memcg_limit_mutex); |
8c7c6e34a memcg: mem+swap c... |
2714 |
ret = -EINVAL; |
628f42355 memcg: limit chan... |
2715 2716 |
break; } |
3e32cb2e0 mm: memcontrol: l... |
2717 2718 2719 2720 |
if (limit > memcg->memory.limit) enlarge = true; ret = page_counter_limit(&memcg->memory, limit); mutex_unlock(&memcg_limit_mutex); |
8c7c6e34a memcg: mem+swap c... |
2721 2722 2723 |
if (!ret) break; |
b70a2a21d mm: memcontrol: f... |
2724 |
try_to_free_mem_cgroup_pages(memcg, 1, GFP_KERNEL, true); |
3e32cb2e0 mm: memcontrol: l... |
2725 |
curusage = page_counter_read(&memcg->memory); |
81d39c20f memcg: fix shrink... |
2726 |
/* Usage is reduced ? */ |
f894ffa86 memcg: trivial cl... |
2727 |
if (curusage >= oldusage) |
81d39c20f memcg: fix shrink... |
2728 2729 2730 |
retry_count--; else oldusage = curusage; |
3e32cb2e0 mm: memcontrol: l... |
2731 |
} while (retry_count); |
3c11ecf44 memcg: oom kill d... |
2732 2733 |
if (!ret && enlarge) memcg_oom_recover(memcg); |
14797e236 memcg: add inacti... |
2734 |
|
8c7c6e34a memcg: mem+swap c... |
2735 2736 |
return ret; } |
338c84310 memcg: remove som... |
2737 |
static int mem_cgroup_resize_memsw_limit(struct mem_cgroup *memcg, |
3e32cb2e0 mm: memcontrol: l... |
2738 |
unsigned long limit) |
8c7c6e34a memcg: mem+swap c... |
2739 |
{ |
3e32cb2e0 mm: memcontrol: l... |
2740 2741 2742 |
unsigned long curusage; unsigned long oldusage; bool enlarge = false; |
81d39c20f memcg: fix shrink... |
2743 |
int retry_count; |
3e32cb2e0 mm: memcontrol: l... |
2744 |
int ret; |
8c7c6e34a memcg: mem+swap c... |
2745 |
|
81d39c20f memcg: fix shrink... |
2746 |
/* see mem_cgroup_resize_res_limit */ |
3e32cb2e0 mm: memcontrol: l... |
2747 2748 2749 2750 2751 2752 |
retry_count = MEM_CGROUP_RECLAIM_RETRIES * mem_cgroup_count_children(memcg); oldusage = page_counter_read(&memcg->memsw); do { |
8c7c6e34a memcg: mem+swap c... |
2753 2754 2755 2756 |
if (signal_pending(current)) { ret = -EINTR; break; } |
3e32cb2e0 mm: memcontrol: l... |
2757 2758 2759 2760 |
mutex_lock(&memcg_limit_mutex); if (limit < memcg->memory.limit) { mutex_unlock(&memcg_limit_mutex); |
8c7c6e34a memcg: mem+swap c... |
2761 |
ret = -EINVAL; |
8c7c6e34a memcg: mem+swap c... |
2762 2763 |
break; } |
3e32cb2e0 mm: memcontrol: l... |
2764 2765 2766 2767 |
if (limit > memcg->memsw.limit) enlarge = true; ret = page_counter_limit(&memcg->memsw, limit); mutex_unlock(&memcg_limit_mutex); |
8c7c6e34a memcg: mem+swap c... |
2768 2769 2770 |
if (!ret) break; |
b70a2a21d mm: memcontrol: f... |
2771 |
try_to_free_mem_cgroup_pages(memcg, 1, GFP_KERNEL, false); |
3e32cb2e0 mm: memcontrol: l... |
2772 |
curusage = page_counter_read(&memcg->memsw); |
81d39c20f memcg: fix shrink... |
2773 |
/* Usage is reduced ? */ |
8c7c6e34a memcg: mem+swap c... |
2774 |
if (curusage >= oldusage) |
628f42355 memcg: limit chan... |
2775 |
retry_count--; |
81d39c20f memcg: fix shrink... |
2776 2777 |
else oldusage = curusage; |
3e32cb2e0 mm: memcontrol: l... |
2778 |
} while (retry_count); |
3c11ecf44 memcg: oom kill d... |
2779 2780 |
if (!ret && enlarge) memcg_oom_recover(memcg); |
3e32cb2e0 mm: memcontrol: l... |
2781 |
|
628f42355 memcg: limit chan... |
2782 2783 |
return ret; } |
0608f43da revert "memcg, vm... |
2784 2785 2786 2787 2788 2789 2790 2791 2792 |
unsigned long mem_cgroup_soft_limit_reclaim(struct zone *zone, int order, gfp_t gfp_mask, unsigned long *total_scanned) { unsigned long nr_reclaimed = 0; struct mem_cgroup_per_zone *mz, *next_mz = NULL; unsigned long reclaimed; int loop = 0; struct mem_cgroup_tree_per_zone *mctz; |
3e32cb2e0 mm: memcontrol: l... |
2793 |
unsigned long excess; |
0608f43da revert "memcg, vm... |
2794 2795 2796 2797 2798 2799 2800 2801 2802 2803 2804 2805 2806 2807 2808 2809 2810 2811 2812 2813 2814 2815 2816 2817 |
unsigned long nr_scanned; if (order > 0) return 0; mctz = soft_limit_tree_node_zone(zone_to_nid(zone), zone_idx(zone)); /* * This loop can run a while, specially if mem_cgroup's continuously * keep exceeding their soft limit and putting the system under * pressure */ do { if (next_mz) mz = next_mz; else mz = mem_cgroup_largest_soft_limit_node(mctz); if (!mz) break; nr_scanned = 0; reclaimed = mem_cgroup_soft_reclaim(mz->memcg, zone, gfp_mask, &nr_scanned); nr_reclaimed += reclaimed; *total_scanned += nr_scanned; |
0a31bc97c mm: memcontrol: r... |
2818 |
spin_lock_irq(&mctz->lock); |
bc2f2e7ff memcg: simplify u... |
2819 |
__mem_cgroup_remove_exceeded(mz, mctz); |
0608f43da revert "memcg, vm... |
2820 2821 2822 2823 2824 2825 |
/* * If we failed to reclaim anything from this memory cgroup * it is time to move on to the next cgroup */ next_mz = NULL; |
bc2f2e7ff memcg: simplify u... |
2826 2827 |
if (!reclaimed) next_mz = __mem_cgroup_largest_soft_limit_node(mctz); |
3e32cb2e0 mm: memcontrol: l... |
2828 |
excess = soft_limit_excess(mz->memcg); |
0608f43da revert "memcg, vm... |
2829 2830 2831 2832 2833 2834 2835 2836 2837 |
/* * One school of thought says that we should not add * back the node to the tree if reclaim returns 0. * But our reclaim could return 0, simply because due * to priority we are exposing a smaller subset of * memory to reclaim from. Consider this as a longer * term TODO. */ /* If excess == 0, no tree ops */ |
cf2c81279 mm: memcontrol: r... |
2838 |
__mem_cgroup_insert_exceeded(mz, mctz, excess); |
0a31bc97c mm: memcontrol: r... |
2839 |
spin_unlock_irq(&mctz->lock); |
0608f43da revert "memcg, vm... |
2840 2841 2842 2843 2844 2845 2846 2847 2848 2849 2850 2851 2852 2853 2854 2855 |
css_put(&mz->memcg->css); loop++; /* * Could not reclaim anything and there are no more * mem cgroups to try or we seem to be looping without * reclaiming anything. */ if (!nr_reclaimed && (next_mz == NULL || loop > MEM_CGROUP_MAX_SOFT_LIMIT_RECLAIM_LOOPS)) break; } while (!nr_reclaimed); if (next_mz) css_put(&next_mz->memcg->css); return nr_reclaimed; } |
ea280e7b4 memcg: update mem... |
2856 2857 2858 2859 2860 2861 |
/* * Test whether @memcg has children, dead or alive. Note that this * function doesn't care whether @memcg has use_hierarchy enabled and * returns %true if there are child csses according to the cgroup * hierarchy. Testing use_hierarchy is the caller's responsiblity. */ |
b5f99b537 memcg: fast hiera... |
2862 2863 |
static inline bool memcg_has_children(struct mem_cgroup *memcg) { |
ea280e7b4 memcg: update mem... |
2864 |
bool ret; |
696ac172f mm: memcg: fix te... |
2865 |
/* |
ea280e7b4 memcg: update mem... |
2866 2867 2868 2869 |
* The lock does not prevent addition or deletion of children, but * it prevents a new child from being initialized based on this * parent in css_online(), so it's enough to decide whether * hierarchically inherited attributes can still be changed or not. |
696ac172f mm: memcg: fix te... |
2870 |
*/ |
ea280e7b4 memcg: update mem... |
2871 2872 2873 2874 2875 2876 |
lockdep_assert_held(&memcg_create_mutex); rcu_read_lock(); ret = css_next_child(NULL, &memcg->css); rcu_read_unlock(); return ret; |
b5f99b537 memcg: fast hiera... |
2877 2878 2879 |
} /* |
c26251f9f memcg: split mem_... |
2880 2881 2882 2883 2884 2885 2886 2887 |
* Reclaims as many pages from the given memcg as possible and moves * the rest to the parent. * * Caller is responsible for holding css reference for memcg. */ static int mem_cgroup_force_empty(struct mem_cgroup *memcg) { int nr_retries = MEM_CGROUP_RECLAIM_RETRIES; |
c26251f9f memcg: split mem_... |
2888 |
|
c1e862c1f memcg: new force_... |
2889 2890 |
/* we call try-to-free pages for make this cgroup empty */ lru_add_drain_all(); |
f817ed485 memcg: move all a... |
2891 |
/* try to free all pages in this cgroup */ |
3e32cb2e0 mm: memcontrol: l... |
2892 |
while (nr_retries && page_counter_read(&memcg->memory)) { |
f817ed485 memcg: move all a... |
2893 |
int progress; |
c1e862c1f memcg: new force_... |
2894 |
|
c26251f9f memcg: split mem_... |
2895 2896 |
if (signal_pending(current)) return -EINTR; |
b70a2a21d mm: memcontrol: f... |
2897 2898 |
progress = try_to_free_mem_cgroup_pages(memcg, 1, GFP_KERNEL, true); |
c1e862c1f memcg: new force_... |
2899 |
if (!progress) { |
f817ed485 memcg: move all a... |
2900 |
nr_retries--; |
c1e862c1f memcg: new force_... |
2901 |
/* maybe some writeback is necessary */ |
8aa7e847d Fix congestion_wa... |
2902 |
congestion_wait(BLK_RW_ASYNC, HZ/10); |
c1e862c1f memcg: new force_... |
2903 |
} |
f817ed485 memcg: move all a... |
2904 2905 |
} |
ab5196c20 memcg: make mem_c... |
2906 2907 |
return 0; |
cc8475822 memory cgroup enh... |
2908 |
} |
6770c64e5 cgroup: replace c... |
2909 2910 2911 |
static ssize_t mem_cgroup_force_empty_write(struct kernfs_open_file *of, char *buf, size_t nbytes, loff_t off) |
c1e862c1f memcg: new force_... |
2912 |
{ |
6770c64e5 cgroup: replace c... |
2913 |
struct mem_cgroup *memcg = mem_cgroup_from_css(of_css(of)); |
c26251f9f memcg: split mem_... |
2914 |
|
d84230118 memcg: root_cgrou... |
2915 2916 |
if (mem_cgroup_is_root(memcg)) return -EINVAL; |
6770c64e5 cgroup: replace c... |
2917 |
return mem_cgroup_force_empty(memcg) ?: nbytes; |
c1e862c1f memcg: new force_... |
2918 |
} |
182446d08 cgroup: pass arou... |
2919 2920 |
static u64 mem_cgroup_hierarchy_read(struct cgroup_subsys_state *css, struct cftype *cft) |
18f59ea7d memcg: memory cgr... |
2921 |
{ |
182446d08 cgroup: pass arou... |
2922 |
return mem_cgroup_from_css(css)->use_hierarchy; |
18f59ea7d memcg: memory cgr... |
2923 |
} |
182446d08 cgroup: pass arou... |
2924 2925 |
static int mem_cgroup_hierarchy_write(struct cgroup_subsys_state *css, struct cftype *cft, u64 val) |
18f59ea7d memcg: memory cgr... |
2926 2927 |
{ int retval = 0; |
182446d08 cgroup: pass arou... |
2928 |
struct mem_cgroup *memcg = mem_cgroup_from_css(css); |
5c9d535b8 cgroup: remove cs... |
2929 |
struct mem_cgroup *parent_memcg = mem_cgroup_from_css(memcg->css.parent); |
18f59ea7d memcg: memory cgr... |
2930 |
|
0999821b1 memcg: replace cg... |
2931 |
mutex_lock(&memcg_create_mutex); |
567fb435b memcg: fix bad be... |
2932 2933 2934 |
if (memcg->use_hierarchy == val) goto out; |
18f59ea7d memcg: memory cgr... |
2935 |
/* |
af901ca18 tree-wide: fix as... |
2936 |
* If parent's use_hierarchy is set, we can't make any modifications |
18f59ea7d memcg: memory cgr... |
2937 2938 2939 2940 2941 2942 |
* 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. */ |
c0ff4b854 memcg: rename mem... |
2943 |
if ((!parent_memcg || !parent_memcg->use_hierarchy) && |
18f59ea7d memcg: memory cgr... |
2944 |
(val == 1 || val == 0)) { |
ea280e7b4 memcg: update mem... |
2945 |
if (!memcg_has_children(memcg)) |
c0ff4b854 memcg: rename mem... |
2946 |
memcg->use_hierarchy = val; |
18f59ea7d memcg: memory cgr... |
2947 2948 2949 2950 |
else retval = -EBUSY; } else retval = -EINVAL; |
567fb435b memcg: fix bad be... |
2951 2952 |
out: |
0999821b1 memcg: replace cg... |
2953 |
mutex_unlock(&memcg_create_mutex); |
18f59ea7d memcg: memory cgr... |
2954 2955 2956 |
return retval; } |
3e32cb2e0 mm: memcontrol: l... |
2957 2958 |
static unsigned long tree_stat(struct mem_cgroup *memcg, enum mem_cgroup_stat_index idx) |
ce00a9673 mm: memcontrol: r... |
2959 2960 2961 2962 2963 2964 2965 2966 2967 2968 2969 2970 2971 2972 2973 2974 |
{ struct mem_cgroup *iter; long val = 0; /* Per-cpu values can be negative, use a signed accumulator */ for_each_mem_cgroup_tree(iter, memcg) val += mem_cgroup_read_stat(iter, idx); if (val < 0) /* race ? */ val = 0; return val; } static inline u64 mem_cgroup_usage(struct mem_cgroup *memcg, bool swap) { u64 val; |
3e32cb2e0 mm: memcontrol: l... |
2975 2976 2977 2978 2979 2980 |
if (mem_cgroup_is_root(memcg)) { val = tree_stat(memcg, MEM_CGROUP_STAT_CACHE); val += tree_stat(memcg, MEM_CGROUP_STAT_RSS); if (swap) val += tree_stat(memcg, MEM_CGROUP_STAT_SWAP); } else { |
ce00a9673 mm: memcontrol: r... |
2981 |
if (!swap) |
3e32cb2e0 mm: memcontrol: l... |
2982 |
val = page_counter_read(&memcg->memory); |
ce00a9673 mm: memcontrol: r... |
2983 |
else |
3e32cb2e0 mm: memcontrol: l... |
2984 |
val = page_counter_read(&memcg->memsw); |
ce00a9673 mm: memcontrol: r... |
2985 |
} |
ce00a9673 mm: memcontrol: r... |
2986 2987 |
return val << PAGE_SHIFT; } |
3e32cb2e0 mm: memcontrol: l... |
2988 2989 2990 2991 2992 2993 2994 |
enum { RES_USAGE, RES_LIMIT, RES_MAX_USAGE, RES_FAILCNT, RES_SOFT_LIMIT, }; |
ce00a9673 mm: memcontrol: r... |
2995 |
|
791badbdb memcg: convert aw... |
2996 |
static u64 mem_cgroup_read_u64(struct cgroup_subsys_state *css, |
05b843012 mm: memcontrol: u... |
2997 |
struct cftype *cft) |
8cdea7c05 Memory controller... |
2998 |
{ |
182446d08 cgroup: pass arou... |
2999 |
struct mem_cgroup *memcg = mem_cgroup_from_css(css); |
3e32cb2e0 mm: memcontrol: l... |
3000 |
struct page_counter *counter; |
af36f906c memcg: always cre... |
3001 |
|
3e32cb2e0 mm: memcontrol: l... |
3002 |
switch (MEMFILE_TYPE(cft->private)) { |
8c7c6e34a memcg: mem+swap c... |
3003 |
case _MEM: |
3e32cb2e0 mm: memcontrol: l... |
3004 3005 |
counter = &memcg->memory; break; |
8c7c6e34a memcg: mem+swap c... |
3006 |
case _MEMSWAP: |
3e32cb2e0 mm: memcontrol: l... |
3007 3008 |
counter = &memcg->memsw; break; |
510fc4e11 memcg: kmem accou... |
3009 |
case _KMEM: |
3e32cb2e0 mm: memcontrol: l... |
3010 |
counter = &memcg->kmem; |
510fc4e11 memcg: kmem accou... |
3011 |
break; |
8c7c6e34a memcg: mem+swap c... |
3012 3013 |
default: BUG(); |
8c7c6e34a memcg: mem+swap c... |
3014 |
} |
3e32cb2e0 mm: memcontrol: l... |
3015 3016 3017 3018 3019 3020 3021 3022 3023 3024 3025 3026 3027 3028 3029 3030 3031 3032 3033 |
switch (MEMFILE_ATTR(cft->private)) { case RES_USAGE: if (counter == &memcg->memory) return mem_cgroup_usage(memcg, false); if (counter == &memcg->memsw) return mem_cgroup_usage(memcg, true); return (u64)page_counter_read(counter) * PAGE_SIZE; case RES_LIMIT: return (u64)counter->limit * PAGE_SIZE; case RES_MAX_USAGE: return (u64)counter->watermark * PAGE_SIZE; case RES_FAILCNT: return counter->failcnt; case RES_SOFT_LIMIT: return (u64)memcg->soft_limit * PAGE_SIZE; default: BUG(); } |
8cdea7c05 Memory controller... |
3034 |
} |
510fc4e11 memcg: kmem accou... |
3035 |
|
510fc4e11 memcg: kmem accou... |
3036 |
#ifdef CONFIG_MEMCG_KMEM |
8c0145b62 memcg: remove act... |
3037 3038 |
static int memcg_activate_kmem(struct mem_cgroup *memcg, unsigned long nr_pages) |
d64416377 memcg: rework mem... |
3039 3040 3041 |
{ int err = 0; int memcg_id; |
2a4db7eb9 memcg: free memcg... |
3042 |
BUG_ON(memcg->kmemcg_id >= 0); |
2788cf0c4 memcg: reparent l... |
3043 |
BUG_ON(memcg->kmem_acct_activated); |
2a4db7eb9 memcg: free memcg... |
3044 |
BUG_ON(memcg->kmem_acct_active); |
d64416377 memcg: rework mem... |
3045 3046 |
/* |
510fc4e11 memcg: kmem accou... |
3047 3048 3049 3050 3051 3052 3053 3054 3055 3056 |
* For simplicity, we won't allow this to be disabled. It also can't * be changed if the cgroup has children already, or if tasks had * already joined. * * If tasks join before we set the limit, a person looking at * kmem.usage_in_bytes will have no way to determine when it took * place, which makes the value quite meaningless. * * After it first became limited, changes in the value of the limit are * of course permitted. |
510fc4e11 memcg: kmem accou... |
3057 |
*/ |
0999821b1 memcg: replace cg... |
3058 |
mutex_lock(&memcg_create_mutex); |
ea280e7b4 memcg: update mem... |
3059 3060 |
if (cgroup_has_tasks(memcg->css.cgroup) || (memcg->use_hierarchy && memcg_has_children(memcg))) |
d64416377 memcg: rework mem... |
3061 3062 3063 3064 |
err = -EBUSY; mutex_unlock(&memcg_create_mutex); if (err) goto out; |
510fc4e11 memcg: kmem accou... |
3065 |
|
f3bb3043a memcg: don't call... |
3066 |
memcg_id = memcg_alloc_cache_id(); |
d64416377 memcg: rework mem... |
3067 3068 3069 3070 |
if (memcg_id < 0) { err = memcg_id; goto out; } |
d64416377 memcg: rework mem... |
3071 |
/* |
900a38f02 memcg: zap kmem_a... |
3072 3073 |
* We couldn't have accounted to this cgroup, because it hasn't got * activated yet, so this should succeed. |
d64416377 memcg: rework mem... |
3074 |
*/ |
3e32cb2e0 mm: memcontrol: l... |
3075 |
err = page_counter_limit(&memcg->kmem, nr_pages); |
d64416377 memcg: rework mem... |
3076 3077 3078 3079 |
VM_BUG_ON(err); static_key_slow_inc(&memcg_kmem_enabled_key); /* |
900a38f02 memcg: zap kmem_a... |
3080 3081 |
* A memory cgroup is considered kmem-active as soon as it gets * kmemcg_id. Setting the id after enabling static branching will |
d64416377 memcg: rework mem... |
3082 3083 3084 |
* guarantee no one starts accounting before all call sites are * patched. */ |
900a38f02 memcg: zap kmem_a... |
3085 |
memcg->kmemcg_id = memcg_id; |
2788cf0c4 memcg: reparent l... |
3086 |
memcg->kmem_acct_activated = true; |
2a4db7eb9 memcg: free memcg... |
3087 |
memcg->kmem_acct_active = true; |
510fc4e11 memcg: kmem accou... |
3088 |
out: |
d64416377 memcg: rework mem... |
3089 |
return err; |
d64416377 memcg: rework mem... |
3090 |
} |
d64416377 memcg: rework mem... |
3091 |
static int memcg_update_kmem_limit(struct mem_cgroup *memcg, |
3e32cb2e0 mm: memcontrol: l... |
3092 |
unsigned long limit) |
d64416377 memcg: rework mem... |
3093 3094 |
{ int ret; |
3e32cb2e0 mm: memcontrol: l... |
3095 |
mutex_lock(&memcg_limit_mutex); |
d64416377 memcg: rework mem... |
3096 |
if (!memcg_kmem_is_active(memcg)) |
3e32cb2e0 mm: memcontrol: l... |
3097 |
ret = memcg_activate_kmem(memcg, limit); |
d64416377 memcg: rework mem... |
3098 |
else |
3e32cb2e0 mm: memcontrol: l... |
3099 3100 |
ret = page_counter_limit(&memcg->kmem, limit); mutex_unlock(&memcg_limit_mutex); |
510fc4e11 memcg: kmem accou... |
3101 3102 |
return ret; } |
55007d849 memcg: allocate m... |
3103 |
static int memcg_propagate_kmem(struct mem_cgroup *memcg) |
510fc4e11 memcg: kmem accou... |
3104 |
{ |
55007d849 memcg: allocate m... |
3105 |
int ret = 0; |
510fc4e11 memcg: kmem accou... |
3106 |
struct mem_cgroup *parent = parent_mem_cgroup(memcg); |
55007d849 memcg: allocate m... |
3107 |
|
d64416377 memcg: rework mem... |
3108 3109 |
if (!parent) return 0; |
55007d849 memcg: allocate m... |
3110 |
|
8c0145b62 memcg: remove act... |
3111 |
mutex_lock(&memcg_limit_mutex); |
55007d849 memcg: allocate m... |
3112 |
/* |
d64416377 memcg: rework mem... |
3113 3114 |
* If the parent cgroup is not kmem-active now, it cannot be activated * after this point, because it has at least one child already. |
55007d849 memcg: allocate m... |
3115 |
*/ |
d64416377 memcg: rework mem... |
3116 |
if (memcg_kmem_is_active(parent)) |
8c0145b62 memcg: remove act... |
3117 3118 |
ret = memcg_activate_kmem(memcg, PAGE_COUNTER_MAX); mutex_unlock(&memcg_limit_mutex); |
55007d849 memcg: allocate m... |
3119 |
return ret; |
510fc4e11 memcg: kmem accou... |
3120 |
} |
d64416377 memcg: rework mem... |
3121 3122 |
#else static int memcg_update_kmem_limit(struct mem_cgroup *memcg, |
3e32cb2e0 mm: memcontrol: l... |
3123 |
unsigned long limit) |
d64416377 memcg: rework mem... |
3124 3125 3126 |
{ return -EINVAL; } |
6d0439904 memcg: stop warni... |
3127 |
#endif /* CONFIG_MEMCG_KMEM */ |
510fc4e11 memcg: kmem accou... |
3128 |
|
628f42355 memcg: limit chan... |
3129 3130 3131 3132 |
/* * The user of this function is... * RES_LIMIT. */ |
451af504d cgroup: replace c... |
3133 3134 |
static ssize_t mem_cgroup_write(struct kernfs_open_file *of, char *buf, size_t nbytes, loff_t off) |
8cdea7c05 Memory controller... |
3135 |
{ |
451af504d cgroup: replace c... |
3136 |
struct mem_cgroup *memcg = mem_cgroup_from_css(of_css(of)); |
3e32cb2e0 mm: memcontrol: l... |
3137 |
unsigned long nr_pages; |
628f42355 memcg: limit chan... |
3138 |
int ret; |
451af504d cgroup: replace c... |
3139 |
buf = strstrip(buf); |
650c5e565 mm: page_counter:... |
3140 |
ret = page_counter_memparse(buf, "-1", &nr_pages); |
3e32cb2e0 mm: memcontrol: l... |
3141 3142 |
if (ret) return ret; |
af36f906c memcg: always cre... |
3143 |
|
3e32cb2e0 mm: memcontrol: l... |
3144 |
switch (MEMFILE_ATTR(of_cft(of)->private)) { |
628f42355 memcg: limit chan... |
3145 |
case RES_LIMIT: |
4b3bde4c9 memcg: remove the... |
3146 3147 3148 3149 |
if (mem_cgroup_is_root(memcg)) { /* Can't set limit on root */ ret = -EINVAL; break; } |
3e32cb2e0 mm: memcontrol: l... |
3150 3151 3152 |
switch (MEMFILE_TYPE(of_cft(of)->private)) { case _MEM: ret = mem_cgroup_resize_limit(memcg, nr_pages); |
8c7c6e34a memcg: mem+swap c... |
3153 |
break; |
3e32cb2e0 mm: memcontrol: l... |
3154 3155 |
case _MEMSWAP: ret = mem_cgroup_resize_memsw_limit(memcg, nr_pages); |
296c81d89 memory controller... |
3156 |
break; |
3e32cb2e0 mm: memcontrol: l... |
3157 3158 3159 3160 |
case _KMEM: ret = memcg_update_kmem_limit(memcg, nr_pages); break; } |
296c81d89 memory controller... |
3161 |
break; |
3e32cb2e0 mm: memcontrol: l... |
3162 3163 3164 |
case RES_SOFT_LIMIT: memcg->soft_limit = nr_pages; ret = 0; |
628f42355 memcg: limit chan... |
3165 3166 |
break; } |
451af504d cgroup: replace c... |
3167 |
return ret ?: nbytes; |
8cdea7c05 Memory controller... |
3168 |
} |
6770c64e5 cgroup: replace c... |
3169 3170 |
static ssize_t mem_cgroup_reset(struct kernfs_open_file *of, char *buf, size_t nbytes, loff_t off) |
c84872e16 memcgroup: add th... |
3171 |
{ |
6770c64e5 cgroup: replace c... |
3172 |
struct mem_cgroup *memcg = mem_cgroup_from_css(of_css(of)); |
3e32cb2e0 mm: memcontrol: l... |
3173 |
struct page_counter *counter; |
c84872e16 memcgroup: add th... |
3174 |
|
3e32cb2e0 mm: memcontrol: l... |
3175 3176 3177 3178 3179 3180 3181 3182 3183 3184 3185 3186 3187 |
switch (MEMFILE_TYPE(of_cft(of)->private)) { case _MEM: counter = &memcg->memory; break; case _MEMSWAP: counter = &memcg->memsw; break; case _KMEM: counter = &memcg->kmem; break; default: BUG(); } |
af36f906c memcg: always cre... |
3188 |
|
3e32cb2e0 mm: memcontrol: l... |
3189 |
switch (MEMFILE_ATTR(of_cft(of)->private)) { |
29f2a4dac memcgroup: implem... |
3190 |
case RES_MAX_USAGE: |
3e32cb2e0 mm: memcontrol: l... |
3191 |
page_counter_reset_watermark(counter); |
29f2a4dac memcgroup: implem... |
3192 3193 |
break; case RES_FAILCNT: |
3e32cb2e0 mm: memcontrol: l... |
3194 |
counter->failcnt = 0; |
29f2a4dac memcgroup: implem... |
3195 |
break; |
3e32cb2e0 mm: memcontrol: l... |
3196 3197 |
default: BUG(); |
29f2a4dac memcgroup: implem... |
3198 |
} |
f64c3f549 memory controller... |
3199 |
|
6770c64e5 cgroup: replace c... |
3200 |
return nbytes; |
c84872e16 memcgroup: add th... |
3201 |
} |
182446d08 cgroup: pass arou... |
3202 |
static u64 mem_cgroup_move_charge_read(struct cgroup_subsys_state *css, |
7dc74be03 memcg: add interf... |
3203 3204 |
struct cftype *cft) { |
182446d08 cgroup: pass arou... |
3205 |
return mem_cgroup_from_css(css)->move_charge_at_immigrate; |
7dc74be03 memcg: add interf... |
3206 |
} |
024914477 memcg: move charg... |
3207 |
#ifdef CONFIG_MMU |
182446d08 cgroup: pass arou... |
3208 |
static int mem_cgroup_move_charge_write(struct cgroup_subsys_state *css, |
7dc74be03 memcg: add interf... |
3209 3210 |
struct cftype *cft, u64 val) { |
182446d08 cgroup: pass arou... |
3211 |
struct mem_cgroup *memcg = mem_cgroup_from_css(css); |
7dc74be03 memcg: add interf... |
3212 |
|
1dfab5abc mm: memcontrol: f... |
3213 |
if (val & ~MOVE_MASK) |
7dc74be03 memcg: add interf... |
3214 |
return -EINVAL; |
ee5e8472b memcg: prevent ch... |
3215 |
|
7dc74be03 memcg: add interf... |
3216 |
/* |
ee5e8472b memcg: prevent ch... |
3217 3218 3219 3220 |
* No kind of locking is needed in here, because ->can_attach() will * check this value once in the beginning of the process, and then carry * on with stale data. This means that changes to this value will only * affect task migrations starting after the change. |
7dc74be03 memcg: add interf... |
3221 |
*/ |
c0ff4b854 memcg: rename mem... |
3222 |
memcg->move_charge_at_immigrate = val; |
7dc74be03 memcg: add interf... |
3223 3224 |
return 0; } |
024914477 memcg: move charg... |
3225 |
#else |
182446d08 cgroup: pass arou... |
3226 |
static int mem_cgroup_move_charge_write(struct cgroup_subsys_state *css, |
024914477 memcg: move charg... |
3227 3228 3229 3230 3231 |
struct cftype *cft, u64 val) { return -ENOSYS; } #endif |
7dc74be03 memcg: add interf... |
3232 |
|
406eb0c9b memcg: add memory... |
3233 |
#ifdef CONFIG_NUMA |
2da8ca822 cgroup: replace c... |
3234 |
static int memcg_numa_stat_show(struct seq_file *m, void *v) |
406eb0c9b memcg: add memory... |
3235 |
{ |
25485de6e memcg: refactor m... |
3236 3237 3238 3239 3240 3241 3242 3243 3244 3245 3246 3247 |
struct numa_stat { const char *name; unsigned int lru_mask; }; static const struct numa_stat stats[] = { { "total", LRU_ALL }, { "file", LRU_ALL_FILE }, { "anon", LRU_ALL_ANON }, { "unevictable", BIT(LRU_UNEVICTABLE) }, }; const struct numa_stat *stat; |
406eb0c9b memcg: add memory... |
3248 |
int nid; |
25485de6e memcg: refactor m... |
3249 |
unsigned long nr; |
2da8ca822 cgroup: replace c... |
3250 |
struct mem_cgroup *memcg = mem_cgroup_from_css(seq_css(m)); |
406eb0c9b memcg: add memory... |
3251 |
|
25485de6e memcg: refactor m... |
3252 3253 3254 3255 3256 3257 3258 3259 3260 3261 |
for (stat = stats; stat < stats + ARRAY_SIZE(stats); stat++) { nr = mem_cgroup_nr_lru_pages(memcg, stat->lru_mask); seq_printf(m, "%s=%lu", stat->name, nr); for_each_node_state(nid, N_MEMORY) { nr = mem_cgroup_node_nr_lru_pages(memcg, nid, stat->lru_mask); seq_printf(m, " N%d=%lu", nid, nr); } seq_putc(m, ' '); |
406eb0c9b memcg: add memory... |
3262 |
} |
406eb0c9b memcg: add memory... |
3263 |
|
071aee138 memcg: support hi... |
3264 3265 3266 3267 3268 3269 3270 3271 3272 3273 3274 3275 3276 3277 3278 3279 |
for (stat = stats; stat < stats + ARRAY_SIZE(stats); stat++) { struct mem_cgroup *iter; nr = 0; for_each_mem_cgroup_tree(iter, memcg) nr += mem_cgroup_nr_lru_pages(iter, stat->lru_mask); seq_printf(m, "hierarchical_%s=%lu", stat->name, nr); for_each_node_state(nid, N_MEMORY) { nr = 0; for_each_mem_cgroup_tree(iter, memcg) nr += mem_cgroup_node_nr_lru_pages( iter, nid, stat->lru_mask); seq_printf(m, " N%d=%lu", nid, nr); } seq_putc(m, ' '); |
406eb0c9b memcg: add memory... |
3280 |
} |
406eb0c9b memcg: add memory... |
3281 |
|
406eb0c9b memcg: add memory... |
3282 3283 3284 |
return 0; } #endif /* CONFIG_NUMA */ |
2da8ca822 cgroup: replace c... |
3285 |
static int memcg_stat_show(struct seq_file *m, void *v) |
d2ceb9b7d memory cgroup enh... |
3286 |
{ |
2da8ca822 cgroup: replace c... |
3287 |
struct mem_cgroup *memcg = mem_cgroup_from_css(seq_css(m)); |
3e32cb2e0 mm: memcontrol: l... |
3288 |
unsigned long memory, memsw; |
af7c4b0ec mm: memcg: print ... |
3289 3290 |
struct mem_cgroup *mi; unsigned int i; |
406eb0c9b memcg: add memory... |
3291 |
|
0ca44b148 memcg: add BUILD_... |
3292 3293 3294 3295 |
BUILD_BUG_ON(ARRAY_SIZE(mem_cgroup_stat_names) != MEM_CGROUP_STAT_NSTATS); BUILD_BUG_ON(ARRAY_SIZE(mem_cgroup_events_names) != MEM_CGROUP_EVENTS_NSTATS); |
70bc068c4 mm/memcontrol.c: ... |
3296 |
BUILD_BUG_ON(ARRAY_SIZE(mem_cgroup_lru_names) != NR_LRU_LISTS); |
af7c4b0ec mm: memcg: print ... |
3297 |
for (i = 0; i < MEM_CGROUP_STAT_NSTATS; i++) { |
bff6bb83f memcg: rename MEM... |
3298 |
if (i == MEM_CGROUP_STAT_SWAP && !do_swap_account) |
1dd3a2732 memcg: show swap ... |
3299 |
continue; |
af7c4b0ec mm: memcg: print ... |
3300 3301 3302 |
seq_printf(m, "%s %ld ", mem_cgroup_stat_names[i], mem_cgroup_read_stat(memcg, i) * PAGE_SIZE); |
1dd3a2732 memcg: show swap ... |
3303 |
} |
7b854121e Unevictable LRU P... |
3304 |
|
af7c4b0ec mm: memcg: print ... |
3305 3306 3307 3308 3309 3310 3311 3312 3313 |
for (i = 0; i < MEM_CGROUP_EVENTS_NSTATS; i++) seq_printf(m, "%s %lu ", mem_cgroup_events_names[i], mem_cgroup_read_events(memcg, i)); for (i = 0; i < NR_LRU_LISTS; i++) seq_printf(m, "%s %lu ", mem_cgroup_lru_names[i], mem_cgroup_nr_lru_pages(memcg, BIT(i)) * PAGE_SIZE); |
14067bb3e memcg: hierarchic... |
3314 |
/* Hierarchical information */ |
3e32cb2e0 mm: memcontrol: l... |
3315 3316 3317 3318 |
memory = memsw = PAGE_COUNTER_MAX; for (mi = memcg; mi; mi = parent_mem_cgroup(mi)) { memory = min(memory, mi->memory.limit); memsw = min(memsw, mi->memsw.limit); |
fee7b548e memcg: show real ... |
3319 |
} |
3e32cb2e0 mm: memcontrol: l... |
3320 3321 3322 3323 3324 3325 3326 |
seq_printf(m, "hierarchical_memory_limit %llu ", (u64)memory * PAGE_SIZE); if (do_swap_account) seq_printf(m, "hierarchical_memsw_limit %llu ", (u64)memsw * PAGE_SIZE); |
7f016ee8b memcg: show recla... |
3327 |
|
af7c4b0ec mm: memcg: print ... |
3328 3329 |
for (i = 0; i < MEM_CGROUP_STAT_NSTATS; i++) { long long val = 0; |
bff6bb83f memcg: rename MEM... |
3330 |
if (i == MEM_CGROUP_STAT_SWAP && !do_swap_account) |
1dd3a2732 memcg: show swap ... |
3331 |
continue; |
af7c4b0ec mm: memcg: print ... |
3332 3333 3334 3335 3336 3337 3338 3339 3340 3341 3342 3343 3344 3345 3346 3347 3348 3349 3350 3351 3352 3353 3354 |
for_each_mem_cgroup_tree(mi, memcg) val += mem_cgroup_read_stat(mi, i) * PAGE_SIZE; seq_printf(m, "total_%s %lld ", mem_cgroup_stat_names[i], val); } for (i = 0; i < MEM_CGROUP_EVENTS_NSTATS; i++) { unsigned long long val = 0; for_each_mem_cgroup_tree(mi, memcg) val += mem_cgroup_read_events(mi, i); seq_printf(m, "total_%s %llu ", mem_cgroup_events_names[i], val); } for (i = 0; i < NR_LRU_LISTS; i++) { unsigned long long val = 0; for_each_mem_cgroup_tree(mi, memcg) val += mem_cgroup_nr_lru_pages(mi, BIT(i)) * PAGE_SIZE; seq_printf(m, "total_%s %llu ", mem_cgroup_lru_names[i], val); |
1dd3a2732 memcg: show swap ... |
3355 |
} |
14067bb3e memcg: hierarchic... |
3356 |
|
7f016ee8b memcg: show recla... |
3357 |
#ifdef CONFIG_DEBUG_VM |
7f016ee8b memcg: show recla... |
3358 3359 3360 |
{ int nid, zid; struct mem_cgroup_per_zone *mz; |
89abfab13 mm/memcg: move re... |
3361 |
struct zone_reclaim_stat *rstat; |
7f016ee8b memcg: show recla... |
3362 3363 3364 3365 3366 |
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++) { |
e231875ba mm: memcontrol: c... |
3367 |
mz = &memcg->nodeinfo[nid]->zoneinfo[zid]; |
89abfab13 mm/memcg: move re... |
3368 |
rstat = &mz->lruvec.reclaim_stat; |
7f016ee8b memcg: show recla... |
3369 |
|
89abfab13 mm/memcg: move re... |
3370 3371 3372 3373 |
recent_rotated[0] += rstat->recent_rotated[0]; recent_rotated[1] += rstat->recent_rotated[1]; recent_scanned[0] += rstat->recent_scanned[0]; recent_scanned[1] += rstat->recent_scanned[1]; |
7f016ee8b memcg: show recla... |
3374 |
} |
78ccf5b5a mm: memcg: print ... |
3375 3376 3377 3378 3379 3380 3381 3382 |
seq_printf(m, "recent_rotated_anon %lu ", recent_rotated[0]); seq_printf(m, "recent_rotated_file %lu ", recent_rotated[1]); seq_printf(m, "recent_scanned_anon %lu ", recent_scanned[0]); seq_printf(m, "recent_scanned_file %lu ", recent_scanned[1]); |
7f016ee8b memcg: show recla... |
3383 3384 |
} #endif |
d2ceb9b7d memory cgroup enh... |
3385 3386 |
return 0; } |
182446d08 cgroup: pass arou... |
3387 3388 |
static u64 mem_cgroup_swappiness_read(struct cgroup_subsys_state *css, struct cftype *cft) |
a7885eb8a memcg: swappiness |
3389 |
{ |
182446d08 cgroup: pass arou... |
3390 |
struct mem_cgroup *memcg = mem_cgroup_from_css(css); |
a7885eb8a memcg: swappiness |
3391 |
|
1f4c025b5 memcg: export mem... |
3392 |
return mem_cgroup_swappiness(memcg); |
a7885eb8a memcg: swappiness |
3393 |
} |
182446d08 cgroup: pass arou... |
3394 3395 |
static int mem_cgroup_swappiness_write(struct cgroup_subsys_state *css, struct cftype *cft, u64 val) |
a7885eb8a memcg: swappiness |
3396 |
{ |
182446d08 cgroup: pass arou... |
3397 |
struct mem_cgroup *memcg = mem_cgroup_from_css(css); |
a7885eb8a memcg: swappiness |
3398 |
|
3dae7fec5 mm: memcontrol: r... |
3399 |
if (val > 100) |
a7885eb8a memcg: swappiness |
3400 |
return -EINVAL; |
14208b0ec Merge branch 'for... |
3401 |
if (css->parent) |
3dae7fec5 mm: memcontrol: r... |
3402 3403 3404 |
memcg->swappiness = val; else vm_swappiness = val; |
068b38c1f memcg: fix a race... |
3405 |
|
a7885eb8a memcg: swappiness |
3406 3407 |
return 0; } |
2e72b6347 memcg: implement ... |
3408 3409 3410 |
static void __mem_cgroup_threshold(struct mem_cgroup *memcg, bool swap) { struct mem_cgroup_threshold_ary *t; |
3e32cb2e0 mm: memcontrol: l... |
3411 |
unsigned long usage; |
2e72b6347 memcg: implement ... |
3412 3413 3414 3415 |
int i; rcu_read_lock(); if (!swap) |
2c488db27 memcg: clean up m... |
3416 |
t = rcu_dereference(memcg->thresholds.primary); |
2e72b6347 memcg: implement ... |
3417 |
else |
2c488db27 memcg: clean up m... |
3418 |
t = rcu_dereference(memcg->memsw_thresholds.primary); |
2e72b6347 memcg: implement ... |
3419 3420 3421 |
if (!t) goto unlock; |
ce00a9673 mm: memcontrol: r... |
3422 |
usage = mem_cgroup_usage(memcg, swap); |
2e72b6347 memcg: implement ... |
3423 3424 |
/* |
748dad36d memcg: make thres... |
3425 |
* current_threshold points to threshold just below or equal to usage. |
2e72b6347 memcg: implement ... |
3426 3427 3428 |
* If it's not true, a threshold was crossed after last * call of __mem_cgroup_threshold(). */ |
5407a5625 mm: remove unnece... |
3429 |
i = t->current_threshold; |
2e72b6347 memcg: implement ... |
3430 3431 3432 3433 3434 3435 3436 3437 3438 3439 3440 3441 3442 3443 3444 3445 3446 3447 3448 3449 3450 3451 3452 |
/* * Iterate backward over array of thresholds starting from * current_threshold and check if a threshold is crossed. * If none of thresholds below usage is crossed, we read * only one element of the array here. */ for (; i >= 0 && unlikely(t->entries[i].threshold > usage); i--) eventfd_signal(t->entries[i].eventfd, 1); /* i = current_threshold + 1 */ i++; /* * Iterate forward over array of thresholds starting from * current_threshold+1 and check if a threshold is crossed. * If none of thresholds above usage is crossed, we read * only one element of the array here. */ for (; i < t->size && unlikely(t->entries[i].threshold <= usage); i++) eventfd_signal(t->entries[i].eventfd, 1); /* Update current_threshold */ |
5407a5625 mm: remove unnece... |
3453 |
t->current_threshold = i - 1; |
2e72b6347 memcg: implement ... |
3454 3455 3456 3457 3458 3459 |
unlock: rcu_read_unlock(); } static void mem_cgroup_threshold(struct mem_cgroup *memcg) { |
ad4ca5f4b memcg: fix thresh... |
3460 3461 3462 3463 3464 3465 3466 |
while (memcg) { __mem_cgroup_threshold(memcg, false); if (do_swap_account) __mem_cgroup_threshold(memcg, true); memcg = parent_mem_cgroup(memcg); } |
2e72b6347 memcg: implement ... |
3467 3468 3469 3470 3471 3472 |
} static int compare_thresholds(const void *a, const void *b) { const struct mem_cgroup_threshold *_a = a; const struct mem_cgroup_threshold *_b = b; |
2bff24a37 memcg: fix multip... |
3473 3474 3475 3476 3477 3478 3479 |
if (_a->threshold > _b->threshold) return 1; if (_a->threshold < _b->threshold) return -1; return 0; |
2e72b6347 memcg: implement ... |
3480 |
} |
c0ff4b854 memcg: rename mem... |
3481 |
static int mem_cgroup_oom_notify_cb(struct mem_cgroup *memcg) |
9490ff275 memcg: oom notifier |
3482 3483 |
{ struct mem_cgroup_eventfd_list *ev; |
2bcf2e92c memcg: oom_notify... |
3484 |
spin_lock(&memcg_oom_lock); |
c0ff4b854 memcg: rename mem... |
3485 |
list_for_each_entry(ev, &memcg->oom_notify, list) |
9490ff275 memcg: oom notifier |
3486 |
eventfd_signal(ev->eventfd, 1); |
2bcf2e92c memcg: oom_notify... |
3487 3488 |
spin_unlock(&memcg_oom_lock); |
9490ff275 memcg: oom notifier |
3489 3490 |
return 0; } |
c0ff4b854 memcg: rename mem... |
3491 |
static void mem_cgroup_oom_notify(struct mem_cgroup *memcg) |
9490ff275 memcg: oom notifier |
3492 |
{ |
7d74b06f2 memcg: use for_ea... |
3493 |
struct mem_cgroup *iter; |
c0ff4b854 memcg: rename mem... |
3494 |
for_each_mem_cgroup_tree(iter, memcg) |
7d74b06f2 memcg: use for_ea... |
3495 |
mem_cgroup_oom_notify_cb(iter); |
9490ff275 memcg: oom notifier |
3496 |
} |
59b6f8734 memcg: make cgrou... |
3497 |
static int __mem_cgroup_usage_register_event(struct mem_cgroup *memcg, |
347c4a874 memcg: remove cgr... |
3498 |
struct eventfd_ctx *eventfd, const char *args, enum res_type type) |
2e72b6347 memcg: implement ... |
3499 |
{ |
2c488db27 memcg: clean up m... |
3500 3501 |
struct mem_cgroup_thresholds *thresholds; struct mem_cgroup_threshold_ary *new; |
3e32cb2e0 mm: memcontrol: l... |
3502 3503 |
unsigned long threshold; unsigned long usage; |
2c488db27 memcg: clean up m... |
3504 |
int i, size, ret; |
2e72b6347 memcg: implement ... |
3505 |
|
650c5e565 mm: page_counter:... |
3506 |
ret = page_counter_memparse(args, "-1", &threshold); |
2e72b6347 memcg: implement ... |
3507 3508 |
if (ret) return ret; |
2105f9aec memcg: convert th... |
3509 |
threshold <<= PAGE_SHIFT; |
2e72b6347 memcg: implement ... |
3510 3511 |
mutex_lock(&memcg->thresholds_lock); |
2c488db27 memcg: clean up m... |
3512 |
|
05b843012 mm: memcontrol: u... |
3513 |
if (type == _MEM) { |
2c488db27 memcg: clean up m... |
3514 |
thresholds = &memcg->thresholds; |
ce00a9673 mm: memcontrol: r... |
3515 |
usage = mem_cgroup_usage(memcg, false); |
05b843012 mm: memcontrol: u... |
3516 |
} else if (type == _MEMSWAP) { |
2c488db27 memcg: clean up m... |
3517 |
thresholds = &memcg->memsw_thresholds; |
ce00a9673 mm: memcontrol: r... |
3518 |
usage = mem_cgroup_usage(memcg, true); |
05b843012 mm: memcontrol: u... |
3519 |
} else |
2e72b6347 memcg: implement ... |
3520 |
BUG(); |
2e72b6347 memcg: implement ... |
3521 |
/* Check if a threshold crossed before adding a new one */ |
2c488db27 memcg: clean up m... |
3522 |
if (thresholds->primary) |
2e72b6347 memcg: implement ... |
3523 |
__mem_cgroup_threshold(memcg, type == _MEMSWAP); |
2c488db27 memcg: clean up m... |
3524 |
size = thresholds->primary ? thresholds->primary->size + 1 : 1; |
2e72b6347 memcg: implement ... |
3525 3526 |
/* Allocate memory for new array of thresholds */ |
2c488db27 memcg: clean up m... |
3527 |
new = kmalloc(sizeof(*new) + size * sizeof(struct mem_cgroup_threshold), |
2e72b6347 memcg: implement ... |
3528 |
GFP_KERNEL); |
2c488db27 memcg: clean up m... |
3529 |
if (!new) { |
2e72b6347 memcg: implement ... |
3530 3531 3532 |
ret = -ENOMEM; goto unlock; } |
2c488db27 memcg: clean up m... |
3533 |
new->size = size; |
2e72b6347 memcg: implement ... |
3534 3535 |
/* Copy thresholds (if any) to new array */ |
2c488db27 memcg: clean up m... |
3536 3537 |
if (thresholds->primary) { memcpy(new->entries, thresholds->primary->entries, (size - 1) * |
2e72b6347 memcg: implement ... |
3538 |
sizeof(struct mem_cgroup_threshold)); |
2c488db27 memcg: clean up m... |
3539 |
} |
2e72b6347 memcg: implement ... |
3540 |
/* Add new threshold */ |
2c488db27 memcg: clean up m... |
3541 3542 |
new->entries[size - 1].eventfd = eventfd; new->entries[size - 1].threshold = threshold; |
2e72b6347 memcg: implement ... |
3543 3544 |
/* Sort thresholds. Registering of new threshold isn't time-critical */ |
2c488db27 memcg: clean up m... |
3545 |
sort(new->entries, size, sizeof(struct mem_cgroup_threshold), |
2e72b6347 memcg: implement ... |
3546 3547 3548 |
compare_thresholds, NULL); /* Find current threshold */ |
2c488db27 memcg: clean up m... |
3549 |
new->current_threshold = -1; |
2e72b6347 memcg: implement ... |
3550 |
for (i = 0; i < size; i++) { |
748dad36d memcg: make thres... |
3551 |
if (new->entries[i].threshold <= usage) { |
2e72b6347 memcg: implement ... |
3552 |
/* |
2c488db27 memcg: clean up m... |
3553 3554 |
* new->current_threshold will not be used until * rcu_assign_pointer(), so it's safe to increment |
2e72b6347 memcg: implement ... |
3555 3556 |
* it here. */ |
2c488db27 memcg: clean up m... |
3557 |
++new->current_threshold; |
748dad36d memcg: make thres... |
3558 3559 |
} else break; |
2e72b6347 memcg: implement ... |
3560 |
} |
2c488db27 memcg: clean up m... |
3561 3562 3563 3564 3565 |
/* Free old spare buffer and save old primary buffer as spare */ kfree(thresholds->spare); thresholds->spare = thresholds->primary; rcu_assign_pointer(thresholds->primary, new); |
2e72b6347 memcg: implement ... |
3566 |
|
907860ed3 cgroups: make cft... |
3567 |
/* To be sure that nobody uses thresholds */ |
2e72b6347 memcg: implement ... |
3568 |
synchronize_rcu(); |
2e72b6347 memcg: implement ... |
3569 3570 3571 3572 3573 |
unlock: mutex_unlock(&memcg->thresholds_lock); return ret; } |
59b6f8734 memcg: make cgrou... |
3574 |
static int mem_cgroup_usage_register_event(struct mem_cgroup *memcg, |
347c4a874 memcg: remove cgr... |
3575 3576 |
struct eventfd_ctx *eventfd, const char *args) { |
59b6f8734 memcg: make cgrou... |
3577 |
return __mem_cgroup_usage_register_event(memcg, eventfd, args, _MEM); |
347c4a874 memcg: remove cgr... |
3578 |
} |
59b6f8734 memcg: make cgrou... |
3579 |
static int memsw_cgroup_usage_register_event(struct mem_cgroup *memcg, |
347c4a874 memcg: remove cgr... |
3580 3581 |
struct eventfd_ctx *eventfd, const char *args) { |
59b6f8734 memcg: make cgrou... |
3582 |
return __mem_cgroup_usage_register_event(memcg, eventfd, args, _MEMSWAP); |
347c4a874 memcg: remove cgr... |
3583 |
} |
59b6f8734 memcg: make cgrou... |
3584 |
static void __mem_cgroup_usage_unregister_event(struct mem_cgroup *memcg, |
347c4a874 memcg: remove cgr... |
3585 |
struct eventfd_ctx *eventfd, enum res_type type) |
2e72b6347 memcg: implement ... |
3586 |
{ |
2c488db27 memcg: clean up m... |
3587 3588 |
struct mem_cgroup_thresholds *thresholds; struct mem_cgroup_threshold_ary *new; |
3e32cb2e0 mm: memcontrol: l... |
3589 |
unsigned long usage; |
2c488db27 memcg: clean up m... |
3590 |
int i, j, size; |
2e72b6347 memcg: implement ... |
3591 3592 |
mutex_lock(&memcg->thresholds_lock); |
05b843012 mm: memcontrol: u... |
3593 3594 |
if (type == _MEM) { |
2c488db27 memcg: clean up m... |
3595 |
thresholds = &memcg->thresholds; |
ce00a9673 mm: memcontrol: r... |
3596 |
usage = mem_cgroup_usage(memcg, false); |
05b843012 mm: memcontrol: u... |
3597 |
} else if (type == _MEMSWAP) { |
2c488db27 memcg: clean up m... |
3598 |
thresholds = &memcg->memsw_thresholds; |
ce00a9673 mm: memcontrol: r... |
3599 |
usage = mem_cgroup_usage(memcg, true); |
05b843012 mm: memcontrol: u... |
3600 |
} else |
2e72b6347 memcg: implement ... |
3601 |
BUG(); |
371528cae mm: memcg: Correc... |
3602 3603 |
if (!thresholds->primary) goto unlock; |
2e72b6347 memcg: implement ... |
3604 3605 3606 3607 |
/* Check if a threshold crossed before removing */ __mem_cgroup_threshold(memcg, type == _MEMSWAP); /* Calculate new number of threshold */ |
2c488db27 memcg: clean up m... |
3608 3609 3610 |
size = 0; for (i = 0; i < thresholds->primary->size; i++) { if (thresholds->primary->entries[i].eventfd != eventfd) |
2e72b6347 memcg: implement ... |
3611 3612 |
size++; } |
2c488db27 memcg: clean up m... |
3613 |
new = thresholds->spare; |
907860ed3 cgroups: make cft... |
3614 |
|
2e72b6347 memcg: implement ... |
3615 3616 |
/* Set thresholds array to NULL if we don't have thresholds */ if (!size) { |
2c488db27 memcg: clean up m... |
3617 3618 |
kfree(new); new = NULL; |
907860ed3 cgroups: make cft... |
3619 |
goto swap_buffers; |
2e72b6347 memcg: implement ... |
3620 |
} |
2c488db27 memcg: clean up m... |
3621 |
new->size = size; |
2e72b6347 memcg: implement ... |
3622 3623 |
/* Copy thresholds and find current threshold */ |
2c488db27 memcg: clean up m... |
3624 3625 3626 |
new->current_threshold = -1; for (i = 0, j = 0; i < thresholds->primary->size; i++) { if (thresholds->primary->entries[i].eventfd == eventfd) |
2e72b6347 memcg: implement ... |
3627 |
continue; |
2c488db27 memcg: clean up m... |
3628 |
new->entries[j] = thresholds->primary->entries[i]; |
748dad36d memcg: make thres... |
3629 |
if (new->entries[j].threshold <= usage) { |
2e72b6347 memcg: implement ... |
3630 |
/* |
2c488db27 memcg: clean up m... |
3631 |
* new->current_threshold will not be used |
2e72b6347 memcg: implement ... |
3632 3633 3634 |
* until rcu_assign_pointer(), so it's safe to increment * it here. */ |
2c488db27 memcg: clean up m... |
3635 |
++new->current_threshold; |
2e72b6347 memcg: implement ... |
3636 3637 3638 |
} j++; } |
907860ed3 cgroups: make cft... |
3639 |
swap_buffers: |
2c488db27 memcg: clean up m... |
3640 3641 |
/* Swap primary and spare array */ thresholds->spare = thresholds->primary; |
8c7577637 memcg: free spare... |
3642 3643 3644 3645 3646 |
/* If all events are unregistered, free the spare array */ if (!new) { kfree(thresholds->spare); thresholds->spare = NULL; } |
2c488db27 memcg: clean up m... |
3647 |
rcu_assign_pointer(thresholds->primary, new); |
2e72b6347 memcg: implement ... |
3648 |
|
907860ed3 cgroups: make cft... |
3649 |
/* To be sure that nobody uses thresholds */ |
2e72b6347 memcg: implement ... |
3650 |
synchronize_rcu(); |
371528cae mm: memcg: Correc... |
3651 |
unlock: |
2e72b6347 memcg: implement ... |
3652 |
mutex_unlock(&memcg->thresholds_lock); |
2e72b6347 memcg: implement ... |
3653 |
} |
c1e862c1f memcg: new force_... |
3654 |
|
59b6f8734 memcg: make cgrou... |
3655 |
static void mem_cgroup_usage_unregister_event(struct mem_cgroup *memcg, |
347c4a874 memcg: remove cgr... |
3656 3657 |
struct eventfd_ctx *eventfd) { |
59b6f8734 memcg: make cgrou... |
3658 |
return __mem_cgroup_usage_unregister_event(memcg, eventfd, _MEM); |
347c4a874 memcg: remove cgr... |
3659 |
} |
59b6f8734 memcg: make cgrou... |
3660 |
static void memsw_cgroup_usage_unregister_event(struct mem_cgroup *memcg, |
347c4a874 memcg: remove cgr... |
3661 3662 |
struct eventfd_ctx *eventfd) { |
59b6f8734 memcg: make cgrou... |
3663 |
return __mem_cgroup_usage_unregister_event(memcg, eventfd, _MEMSWAP); |
347c4a874 memcg: remove cgr... |
3664 |
} |
59b6f8734 memcg: make cgrou... |
3665 |
static int mem_cgroup_oom_register_event(struct mem_cgroup *memcg, |
347c4a874 memcg: remove cgr... |
3666 |
struct eventfd_ctx *eventfd, const char *args) |
9490ff275 memcg: oom notifier |
3667 |
{ |
9490ff275 memcg: oom notifier |
3668 |
struct mem_cgroup_eventfd_list *event; |
9490ff275 memcg: oom notifier |
3669 |
|
9490ff275 memcg: oom notifier |
3670 3671 3672 |
event = kmalloc(sizeof(*event), GFP_KERNEL); if (!event) return -ENOMEM; |
1af8efe96 memcg: change mem... |
3673 |
spin_lock(&memcg_oom_lock); |
9490ff275 memcg: oom notifier |
3674 3675 3676 3677 3678 |
event->eventfd = eventfd; list_add(&event->list, &memcg->oom_notify); /* already in OOM ? */ |
79dfdaccd memcg: make oom_l... |
3679 |
if (atomic_read(&memcg->under_oom)) |
9490ff275 memcg: oom notifier |
3680 |
eventfd_signal(eventfd, 1); |
1af8efe96 memcg: change mem... |
3681 |
spin_unlock(&memcg_oom_lock); |
9490ff275 memcg: oom notifier |
3682 3683 3684 |
return 0; } |
59b6f8734 memcg: make cgrou... |
3685 |
static void mem_cgroup_oom_unregister_event(struct mem_cgroup *memcg, |
347c4a874 memcg: remove cgr... |
3686 |
struct eventfd_ctx *eventfd) |
9490ff275 memcg: oom notifier |
3687 |
{ |
9490ff275 memcg: oom notifier |
3688 |
struct mem_cgroup_eventfd_list *ev, *tmp; |
9490ff275 memcg: oom notifier |
3689 |
|
1af8efe96 memcg: change mem... |
3690 |
spin_lock(&memcg_oom_lock); |
9490ff275 memcg: oom notifier |
3691 |
|
c0ff4b854 memcg: rename mem... |
3692 |
list_for_each_entry_safe(ev, tmp, &memcg->oom_notify, list) { |
9490ff275 memcg: oom notifier |
3693 3694 3695 3696 3697 |
if (ev->eventfd == eventfd) { list_del(&ev->list); kfree(ev); } } |
1af8efe96 memcg: change mem... |
3698 |
spin_unlock(&memcg_oom_lock); |
9490ff275 memcg: oom notifier |
3699 |
} |
2da8ca822 cgroup: replace c... |
3700 |
static int mem_cgroup_oom_control_read(struct seq_file *sf, void *v) |
3c11ecf44 memcg: oom kill d... |
3701 |
{ |
2da8ca822 cgroup: replace c... |
3702 |
struct mem_cgroup *memcg = mem_cgroup_from_css(seq_css(sf)); |
3c11ecf44 memcg: oom kill d... |
3703 |
|
791badbdb memcg: convert aw... |
3704 3705 3706 3707 |
seq_printf(sf, "oom_kill_disable %d ", memcg->oom_kill_disable); seq_printf(sf, "under_oom %d ", (bool)atomic_read(&memcg->under_oom)); |
3c11ecf44 memcg: oom kill d... |
3708 3709 |
return 0; } |
182446d08 cgroup: pass arou... |
3710 |
static int mem_cgroup_oom_control_write(struct cgroup_subsys_state *css, |
3c11ecf44 memcg: oom kill d... |
3711 3712 |
struct cftype *cft, u64 val) { |
182446d08 cgroup: pass arou... |
3713 |
struct mem_cgroup *memcg = mem_cgroup_from_css(css); |
3c11ecf44 memcg: oom kill d... |
3714 3715 |
/* cannot set to root cgroup and only 0 and 1 are allowed */ |
14208b0ec Merge branch 'for... |
3716 |
if (!css->parent || !((val == 0) || (val == 1))) |
3c11ecf44 memcg: oom kill d... |
3717 |
return -EINVAL; |
c0ff4b854 memcg: rename mem... |
3718 |
memcg->oom_kill_disable = val; |
4d845ebf4 memcg: fix wake u... |
3719 |
if (!val) |
c0ff4b854 memcg: rename mem... |
3720 |
memcg_oom_recover(memcg); |
3dae7fec5 mm: memcontrol: r... |
3721 |
|
3c11ecf44 memcg: oom kill d... |
3722 3723 |
return 0; } |
c255a4580 memcg: rename con... |
3724 |
#ifdef CONFIG_MEMCG_KMEM |
cbe128e34 cgroup: get rid o... |
3725 |
static int memcg_init_kmem(struct mem_cgroup *memcg, struct cgroup_subsys *ss) |
e5671dfae Basic kernel memo... |
3726 |
{ |
55007d849 memcg: allocate m... |
3727 |
int ret; |
55007d849 memcg: allocate m... |
3728 3729 3730 |
ret = memcg_propagate_kmem(memcg); if (ret) return ret; |
2633d7a02 slab/slub: consid... |
3731 |
|
1d62e4365 cgroup: pass stru... |
3732 |
return mem_cgroup_sockets_init(memcg, ss); |
573b400d0 mm/memcontrol.c: ... |
3733 |
} |
e5671dfae Basic kernel memo... |
3734 |
|
2a4db7eb9 memcg: free memcg... |
3735 3736 |
static void memcg_deactivate_kmem(struct mem_cgroup *memcg) { |
2788cf0c4 memcg: reparent l... |
3737 3738 3739 |
struct cgroup_subsys_state *css; struct mem_cgroup *parent, *child; int kmemcg_id; |
2a4db7eb9 memcg: free memcg... |
3740 3741 3742 3743 3744 3745 3746 3747 3748 3749 3750 3751 |
if (!memcg->kmem_acct_active) return; /* * Clear the 'active' flag before clearing memcg_caches arrays entries. * Since we take the slab_mutex in memcg_deactivate_kmem_caches(), it * guarantees no cache will be created for this cgroup after we are * done (see memcg_create_kmem_cache()). */ memcg->kmem_acct_active = false; memcg_deactivate_kmem_caches(memcg); |
2788cf0c4 memcg: reparent l... |
3752 3753 3754 3755 3756 3757 3758 3759 3760 3761 3762 3763 3764 3765 3766 3767 3768 3769 3770 3771 3772 3773 3774 3775 3776 3777 |
kmemcg_id = memcg->kmemcg_id; BUG_ON(kmemcg_id < 0); parent = parent_mem_cgroup(memcg); if (!parent) parent = root_mem_cgroup; /* * Change kmemcg_id of this cgroup and all its descendants to the * parent's id, and then move all entries from this cgroup's list_lrus * to ones of the parent. After we have finished, all list_lrus * corresponding to this cgroup are guaranteed to remain empty. The * ordering is imposed by list_lru_node->lock taken by * memcg_drain_all_list_lrus(). */ css_for_each_descendant_pre(css, &memcg->css) { child = mem_cgroup_from_css(css); BUG_ON(child->kmemcg_id != kmemcg_id); child->kmemcg_id = parent->kmemcg_id; if (!memcg->use_hierarchy) break; } memcg_drain_all_list_lrus(kmemcg_id, parent->kmemcg_id); memcg_free_cache_id(kmemcg_id); |
2a4db7eb9 memcg: free memcg... |
3778 |
} |
10d5ebf40 memcg: use css_ge... |
3779 |
static void memcg_destroy_kmem(struct mem_cgroup *memcg) |
d1a4c0b37 tcp memory pressu... |
3780 |
{ |
f48b80a5e memcg: cleanup st... |
3781 3782 3783 3784 3785 |
if (memcg->kmem_acct_activated) { memcg_destroy_kmem_caches(memcg); static_key_slow_dec(&memcg_kmem_enabled_key); WARN_ON(page_counter_read(&memcg->kmem)); } |
1d62e4365 cgroup: pass stru... |
3786 |
mem_cgroup_sockets_destroy(memcg); |
10d5ebf40 memcg: use css_ge... |
3787 |
} |
e5671dfae Basic kernel memo... |
3788 |
#else |
cbe128e34 cgroup: get rid o... |
3789 |
static int memcg_init_kmem(struct mem_cgroup *memcg, struct cgroup_subsys *ss) |
e5671dfae Basic kernel memo... |
3790 3791 3792 |
{ return 0; } |
d1a4c0b37 tcp memory pressu... |
3793 |
|
2a4db7eb9 memcg: free memcg... |
3794 3795 3796 |
static void memcg_deactivate_kmem(struct mem_cgroup *memcg) { } |
10d5ebf40 memcg: use css_ge... |
3797 3798 3799 |
static void memcg_destroy_kmem(struct mem_cgroup *memcg) { } |
e5671dfae Basic kernel memo... |
3800 |
#endif |
79bd9814e cgroup, memcg: mo... |
3801 |
/* |
3bc942f37 memcg: rename cgr... |
3802 3803 3804 3805 3806 3807 3808 3809 3810 3811 3812 3813 3814 |
* DO NOT USE IN NEW FILES. * * "cgroup.event_control" implementation. * * This is way over-engineered. It tries to support fully configurable * events for each user. Such level of flexibility is completely * unnecessary especially in the light of the planned unified hierarchy. * * Please deprecate this and replace with something simpler if at all * possible. */ /* |
79bd9814e cgroup, memcg: mo... |
3815 3816 3817 3818 |
* Unregister event and free resources. * * Gets called from workqueue. */ |
3bc942f37 memcg: rename cgr... |
3819 |
static void memcg_event_remove(struct work_struct *work) |
79bd9814e cgroup, memcg: mo... |
3820 |
{ |
3bc942f37 memcg: rename cgr... |
3821 3822 |
struct mem_cgroup_event *event = container_of(work, struct mem_cgroup_event, remove); |
59b6f8734 memcg: make cgrou... |
3823 |
struct mem_cgroup *memcg = event->memcg; |
79bd9814e cgroup, memcg: mo... |
3824 3825 |
remove_wait_queue(event->wqh, &event->wait); |
59b6f8734 memcg: make cgrou... |
3826 |
event->unregister_event(memcg, event->eventfd); |
79bd9814e cgroup, memcg: mo... |
3827 3828 3829 3830 3831 3832 |
/* Notify userspace the event is going away. */ eventfd_signal(event->eventfd, 1); eventfd_ctx_put(event->eventfd); kfree(event); |
59b6f8734 memcg: make cgrou... |
3833 |
css_put(&memcg->css); |
79bd9814e cgroup, memcg: mo... |
3834 3835 3836 3837 3838 3839 3840 |
} /* * Gets called on POLLHUP on eventfd when user closes it. * * Called with wqh->lock held and interrupts disabled. */ |
3bc942f37 memcg: rename cgr... |
3841 3842 |
static int memcg_event_wake(wait_queue_t *wait, unsigned mode, int sync, void *key) |
79bd9814e cgroup, memcg: mo... |
3843 |
{ |
3bc942f37 memcg: rename cgr... |
3844 3845 |
struct mem_cgroup_event *event = container_of(wait, struct mem_cgroup_event, wait); |
59b6f8734 memcg: make cgrou... |
3846 |
struct mem_cgroup *memcg = event->memcg; |
79bd9814e cgroup, memcg: mo... |
3847 3848 3849 3850 3851 3852 3853 3854 3855 3856 3857 3858 |
unsigned long flags = (unsigned long)key; if (flags & POLLHUP) { /* * If the event has been detached at cgroup removal, we * can simply return knowing the other side will cleanup * for us. * * We can't race against event freeing since the other * side will require wqh->lock via remove_wait_queue(), * which we hold. */ |
fba948078 cgroup, memcg: mo... |
3859 |
spin_lock(&memcg->event_list_lock); |
79bd9814e cgroup, memcg: mo... |
3860 3861 3862 3863 3864 3865 3866 3867 |
if (!list_empty(&event->list)) { list_del_init(&event->list); /* * We are in atomic context, but cgroup_event_remove() * may sleep, so we have to call it in workqueue. */ schedule_work(&event->remove); } |
fba948078 cgroup, memcg: mo... |
3868 |
spin_unlock(&memcg->event_list_lock); |
79bd9814e cgroup, memcg: mo... |
3869 3870 3871 3872 |
} return 0; } |
3bc942f37 memcg: rename cgr... |
3873 |
static void memcg_event_ptable_queue_proc(struct file *file, |
79bd9814e cgroup, memcg: mo... |
3874 3875 |
wait_queue_head_t *wqh, poll_table *pt) { |
3bc942f37 memcg: rename cgr... |
3876 3877 |
struct mem_cgroup_event *event = container_of(pt, struct mem_cgroup_event, pt); |
79bd9814e cgroup, memcg: mo... |
3878 3879 3880 3881 3882 3883 |
event->wqh = wqh; add_wait_queue(wqh, &event->wait); } /* |
3bc942f37 memcg: rename cgr... |
3884 3885 |
* DO NOT USE IN NEW FILES. * |
79bd9814e cgroup, memcg: mo... |
3886 3887 3888 3889 3890 |
* Parse input and register new cgroup event handler. * * Input must be in format '<event_fd> <control_fd> <args>'. * Interpretation of args is defined by control file implementation. */ |
451af504d cgroup: replace c... |
3891 3892 |
static ssize_t memcg_write_event_control(struct kernfs_open_file *of, char *buf, size_t nbytes, loff_t off) |
79bd9814e cgroup, memcg: mo... |
3893 |
{ |
451af504d cgroup: replace c... |
3894 |
struct cgroup_subsys_state *css = of_css(of); |
fba948078 cgroup, memcg: mo... |
3895 |
struct mem_cgroup *memcg = mem_cgroup_from_css(css); |
3bc942f37 memcg: rename cgr... |
3896 |
struct mem_cgroup_event *event; |
79bd9814e cgroup, memcg: mo... |
3897 3898 3899 3900 |
struct cgroup_subsys_state *cfile_css; unsigned int efd, cfd; struct fd efile; struct fd cfile; |
fba948078 cgroup, memcg: mo... |
3901 |
const char *name; |
79bd9814e cgroup, memcg: mo... |
3902 3903 |
char *endp; int ret; |
451af504d cgroup: replace c... |
3904 3905 3906 |
buf = strstrip(buf); efd = simple_strtoul(buf, &endp, 10); |
79bd9814e cgroup, memcg: mo... |
3907 3908 |
if (*endp != ' ') return -EINVAL; |
451af504d cgroup: replace c... |
3909 |
buf = endp + 1; |
79bd9814e cgroup, memcg: mo... |
3910 |
|
451af504d cgroup: replace c... |
3911 |
cfd = simple_strtoul(buf, &endp, 10); |
79bd9814e cgroup, memcg: mo... |
3912 3913 |
if ((*endp != ' ') && (*endp != '\0')) return -EINVAL; |
451af504d cgroup: replace c... |
3914 |
buf = endp + 1; |
79bd9814e cgroup, memcg: mo... |
3915 3916 3917 3918 |
event = kzalloc(sizeof(*event), GFP_KERNEL); if (!event) return -ENOMEM; |
59b6f8734 memcg: make cgrou... |
3919 |
event->memcg = memcg; |
79bd9814e cgroup, memcg: mo... |
3920 |
INIT_LIST_HEAD(&event->list); |
3bc942f37 memcg: rename cgr... |
3921 3922 3923 |
init_poll_funcptr(&event->pt, memcg_event_ptable_queue_proc); init_waitqueue_func_entry(&event->wait, memcg_event_wake); INIT_WORK(&event->remove, memcg_event_remove); |
79bd9814e cgroup, memcg: mo... |
3924 3925 3926 3927 3928 3929 3930 3931 3932 3933 3934 3935 3936 3937 3938 3939 3940 3941 3942 3943 3944 3945 3946 3947 |
efile = fdget(efd); if (!efile.file) { ret = -EBADF; goto out_kfree; } event->eventfd = eventfd_ctx_fileget(efile.file); if (IS_ERR(event->eventfd)) { ret = PTR_ERR(event->eventfd); goto out_put_efile; } cfile = fdget(cfd); if (!cfile.file) { ret = -EBADF; goto out_put_eventfd; } /* the process need read permission on control file */ /* AV: shouldn't we check that it's been opened for read instead? */ ret = inode_permission(file_inode(cfile.file), MAY_READ); if (ret < 0) goto out_put_cfile; |
79bd9814e cgroup, memcg: mo... |
3948 |
/* |
fba948078 cgroup, memcg: mo... |
3949 3950 3951 3952 |
* Determine the event callbacks and set them in @event. This used * to be done via struct cftype but cgroup core no longer knows * about these events. The following is crude but the whole thing * is for compatibility anyway. |
3bc942f37 memcg: rename cgr... |
3953 3954 |
* * DO NOT ADD NEW FILES. |
fba948078 cgroup, memcg: mo... |
3955 |
*/ |
b583043e9 kill f_dentry uses |
3956 |
name = cfile.file->f_path.dentry->d_name.name; |
fba948078 cgroup, memcg: mo... |
3957 3958 3959 3960 3961 3962 3963 3964 3965 3966 3967 |
if (!strcmp(name, "memory.usage_in_bytes")) { event->register_event = mem_cgroup_usage_register_event; event->unregister_event = mem_cgroup_usage_unregister_event; } else if (!strcmp(name, "memory.oom_control")) { event->register_event = mem_cgroup_oom_register_event; event->unregister_event = mem_cgroup_oom_unregister_event; } else if (!strcmp(name, "memory.pressure_level")) { event->register_event = vmpressure_register_event; event->unregister_event = vmpressure_unregister_event; } else if (!strcmp(name, "memory.memsw.usage_in_bytes")) { |
347c4a874 memcg: remove cgr... |
3968 3969 |
event->register_event = memsw_cgroup_usage_register_event; event->unregister_event = memsw_cgroup_usage_unregister_event; |
fba948078 cgroup, memcg: mo... |
3970 3971 3972 3973 3974 3975 |
} else { ret = -EINVAL; goto out_put_cfile; } /* |
b5557c4c3 memcg: cgroup_wri... |
3976 3977 3978 |
* Verify @cfile should belong to @css. Also, remaining events are * automatically removed on cgroup destruction but the removal is * asynchronous, so take an extra ref on @css. |
79bd9814e cgroup, memcg: mo... |
3979 |
*/ |
b583043e9 kill f_dentry uses |
3980 |
cfile_css = css_tryget_online_from_dir(cfile.file->f_path.dentry->d_parent, |
ec903c0c8 cgroup: rename cs... |
3981 |
&memory_cgrp_subsys); |
79bd9814e cgroup, memcg: mo... |
3982 |
ret = -EINVAL; |
5a17f543e cgroup: improve c... |
3983 |
if (IS_ERR(cfile_css)) |
79bd9814e cgroup, memcg: mo... |
3984 |
goto out_put_cfile; |
5a17f543e cgroup: improve c... |
3985 3986 |
if (cfile_css != css) { css_put(cfile_css); |
79bd9814e cgroup, memcg: mo... |
3987 |
goto out_put_cfile; |
5a17f543e cgroup: improve c... |
3988 |
} |
79bd9814e cgroup, memcg: mo... |
3989 |
|
451af504d cgroup: replace c... |
3990 |
ret = event->register_event(memcg, event->eventfd, buf); |
79bd9814e cgroup, memcg: mo... |
3991 3992 3993 3994 |
if (ret) goto out_put_css; efile.file->f_op->poll(efile.file, &event->pt); |
fba948078 cgroup, memcg: mo... |
3995 3996 3997 |
spin_lock(&memcg->event_list_lock); list_add(&event->list, &memcg->event_list); spin_unlock(&memcg->event_list_lock); |
79bd9814e cgroup, memcg: mo... |
3998 3999 4000 |
fdput(cfile); fdput(efile); |
451af504d cgroup: replace c... |
4001 |
return nbytes; |
79bd9814e cgroup, memcg: mo... |
4002 4003 |
out_put_css: |
b5557c4c3 memcg: cgroup_wri... |
4004 |
css_put(css); |
79bd9814e cgroup, memcg: mo... |
4005 4006 4007 4008 4009 4010 4011 4012 4013 4014 4015 |
out_put_cfile: fdput(cfile); out_put_eventfd: eventfd_ctx_put(event->eventfd); out_put_efile: fdput(efile); out_kfree: kfree(event); return ret; } |
241994ed8 mm: memcontrol: d... |
4016 |
static struct cftype mem_cgroup_legacy_files[] = { |
8cdea7c05 Memory controller... |
4017 |
{ |
0eea10301 Memory controller... |
4018 |
.name = "usage_in_bytes", |
8c7c6e34a memcg: mem+swap c... |
4019 |
.private = MEMFILE_PRIVATE(_MEM, RES_USAGE), |
791badbdb memcg: convert aw... |
4020 |
.read_u64 = mem_cgroup_read_u64, |
8cdea7c05 Memory controller... |
4021 4022 |
}, { |
c84872e16 memcgroup: add th... |
4023 |
.name = "max_usage_in_bytes", |
8c7c6e34a memcg: mem+swap c... |
4024 |
.private = MEMFILE_PRIVATE(_MEM, RES_MAX_USAGE), |
6770c64e5 cgroup: replace c... |
4025 |
.write = mem_cgroup_reset, |
791badbdb memcg: convert aw... |
4026 |
.read_u64 = mem_cgroup_read_u64, |
c84872e16 memcgroup: add th... |
4027 4028 |
}, { |
0eea10301 Memory controller... |
4029 |
.name = "limit_in_bytes", |
8c7c6e34a memcg: mem+swap c... |
4030 |
.private = MEMFILE_PRIVATE(_MEM, RES_LIMIT), |
451af504d cgroup: replace c... |
4031 |
.write = mem_cgroup_write, |
791badbdb memcg: convert aw... |
4032 |
.read_u64 = mem_cgroup_read_u64, |
8cdea7c05 Memory controller... |
4033 4034 |
}, { |
296c81d89 memory controller... |
4035 4036 |
.name = "soft_limit_in_bytes", .private = MEMFILE_PRIVATE(_MEM, RES_SOFT_LIMIT), |
451af504d cgroup: replace c... |
4037 |
.write = mem_cgroup_write, |
791badbdb memcg: convert aw... |
4038 |
.read_u64 = mem_cgroup_read_u64, |
296c81d89 memory controller... |
4039 4040 |
}, { |
8cdea7c05 Memory controller... |
4041 |
.name = "failcnt", |
8c7c6e34a memcg: mem+swap c... |
4042 |
.private = MEMFILE_PRIVATE(_MEM, RES_FAILCNT), |
6770c64e5 cgroup: replace c... |
4043 |
.write = mem_cgroup_reset, |
791badbdb memcg: convert aw... |
4044 |
.read_u64 = mem_cgroup_read_u64, |
8cdea7c05 Memory controller... |
4045 |
}, |
8697d3319 Memory controller... |
4046 |
{ |
d2ceb9b7d memory cgroup enh... |
4047 |
.name = "stat", |
2da8ca822 cgroup: replace c... |
4048 |
.seq_show = memcg_stat_show, |
d2ceb9b7d memory cgroup enh... |
4049 |
}, |
c1e862c1f memcg: new force_... |
4050 4051 |
{ .name = "force_empty", |
6770c64e5 cgroup: replace c... |
4052 |
.write = mem_cgroup_force_empty_write, |
c1e862c1f memcg: new force_... |
4053 |
}, |
18f59ea7d memcg: memory cgr... |
4054 4055 4056 4057 4058 |
{ .name = "use_hierarchy", .write_u64 = mem_cgroup_hierarchy_write, .read_u64 = mem_cgroup_hierarchy_read, }, |
a7885eb8a memcg: swappiness |
4059 |
{ |
3bc942f37 memcg: rename cgr... |
4060 |
.name = "cgroup.event_control", /* XXX: for compat */ |
451af504d cgroup: replace c... |
4061 |
.write = memcg_write_event_control, |
79bd9814e cgroup, memcg: mo... |
4062 4063 4064 4065 |
.flags = CFTYPE_NO_PREFIX, .mode = S_IWUGO, }, { |
a7885eb8a memcg: swappiness |
4066 4067 4068 4069 |
.name = "swappiness", .read_u64 = mem_cgroup_swappiness_read, .write_u64 = mem_cgroup_swappiness_write, }, |
7dc74be03 memcg: add interf... |
4070 4071 4072 4073 4074 |
{ .name = "move_charge_at_immigrate", .read_u64 = mem_cgroup_move_charge_read, .write_u64 = mem_cgroup_move_charge_write, }, |
9490ff275 memcg: oom notifier |
4075 4076 |
{ .name = "oom_control", |
2da8ca822 cgroup: replace c... |
4077 |
.seq_show = mem_cgroup_oom_control_read, |
3c11ecf44 memcg: oom kill d... |
4078 |
.write_u64 = mem_cgroup_oom_control_write, |
9490ff275 memcg: oom notifier |
4079 4080 |
.private = MEMFILE_PRIVATE(_OOM_TYPE, OOM_CONTROL), }, |
70ddf637e memcg: add memory... |
4081 4082 |
{ .name = "pressure_level", |
70ddf637e memcg: add memory... |
4083 |
}, |
406eb0c9b memcg: add memory... |
4084 4085 4086 |
#ifdef CONFIG_NUMA { .name = "numa_stat", |
2da8ca822 cgroup: replace c... |
4087 |
.seq_show = memcg_numa_stat_show, |
406eb0c9b memcg: add memory... |
4088 4089 |
}, #endif |
510fc4e11 memcg: kmem accou... |
4090 4091 4092 4093 |
#ifdef CONFIG_MEMCG_KMEM { .name = "kmem.limit_in_bytes", .private = MEMFILE_PRIVATE(_KMEM, RES_LIMIT), |
451af504d cgroup: replace c... |
4094 |
.write = mem_cgroup_write, |
791badbdb memcg: convert aw... |
4095 |
.read_u64 = mem_cgroup_read_u64, |
510fc4e11 memcg: kmem accou... |
4096 4097 4098 4099 |
}, { .name = "kmem.usage_in_bytes", .private = MEMFILE_PRIVATE(_KMEM, RES_USAGE), |
791badbdb memcg: convert aw... |
4100 |
.read_u64 = mem_cgroup_read_u64, |
510fc4e11 memcg: kmem accou... |
4101 4102 4103 4104 |
}, { .name = "kmem.failcnt", .private = MEMFILE_PRIVATE(_KMEM, RES_FAILCNT), |
6770c64e5 cgroup: replace c... |
4105 |
.write = mem_cgroup_reset, |
791badbdb memcg: convert aw... |
4106 |
.read_u64 = mem_cgroup_read_u64, |
510fc4e11 memcg: kmem accou... |
4107 4108 4109 4110 |
}, { .name = "kmem.max_usage_in_bytes", .private = MEMFILE_PRIVATE(_KMEM, RES_MAX_USAGE), |
6770c64e5 cgroup: replace c... |
4111 |
.write = mem_cgroup_reset, |
791badbdb memcg: convert aw... |
4112 |
.read_u64 = mem_cgroup_read_u64, |
510fc4e11 memcg: kmem accou... |
4113 |
}, |
749c54151 memcg: aggregate ... |
4114 4115 4116 |
#ifdef CONFIG_SLABINFO { .name = "kmem.slabinfo", |
b047501cd memcg: use generi... |
4117 4118 4119 4120 |
.seq_start = slab_start, .seq_next = slab_next, .seq_stop = slab_stop, .seq_show = memcg_slab_show, |
749c54151 memcg: aggregate ... |
4121 4122 |
}, #endif |
510fc4e11 memcg: kmem accou... |
4123 |
#endif |
6bc103498 cgroup: convert m... |
4124 |
{ }, /* terminate */ |
af36f906c memcg: always cre... |
4125 |
}; |
8c7c6e34a memcg: mem+swap c... |
4126 |
|
c0ff4b854 memcg: rename mem... |
4127 |
static int alloc_mem_cgroup_per_zone_info(struct mem_cgroup *memcg, int node) |
6d12e2d8d per-zone and recl... |
4128 4129 |
{ struct mem_cgroup_per_node *pn; |
1ecaab2bd per-zone and recl... |
4130 |
struct mem_cgroup_per_zone *mz; |
41e3355de memcg: fix node_s... |
4131 |
int zone, tmp = node; |
1ecaab2bd per-zone and recl... |
4132 4133 4134 4135 4136 4137 4138 4139 |
/* * 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. */ |
41e3355de memcg: fix node_s... |
4140 4141 |
if (!node_state(node, N_NORMAL_MEMORY)) tmp = -1; |
17295c88a memcg: use [kv]za... |
4142 |
pn = kzalloc_node(sizeof(*pn), GFP_KERNEL, tmp); |
6d12e2d8d per-zone and recl... |
4143 4144 |
if (!pn) return 1; |
1ecaab2bd per-zone and recl... |
4145 |
|
1ecaab2bd per-zone and recl... |
4146 4147 |
for (zone = 0; zone < MAX_NR_ZONES; zone++) { mz = &pn->zoneinfo[zone]; |
bea8c150a memcg: fix hotplu... |
4148 |
lruvec_init(&mz->lruvec); |
bb4cc1a8b revert "memcg: ge... |
4149 4150 |
mz->usage_in_excess = 0; mz->on_tree = false; |
d79154bb5 memcg: replace me... |
4151 |
mz->memcg = memcg; |
1ecaab2bd per-zone and recl... |
4152 |
} |
54f72fe02 memcg: clean up m... |
4153 |
memcg->nodeinfo[node] = pn; |
6d12e2d8d per-zone and recl... |
4154 4155 |
return 0; } |
c0ff4b854 memcg: rename mem... |
4156 |
static void free_mem_cgroup_per_zone_info(struct mem_cgroup *memcg, int node) |
1ecaab2bd per-zone and recl... |
4157 |
{ |
54f72fe02 memcg: clean up m... |
4158 |
kfree(memcg->nodeinfo[node]); |
1ecaab2bd per-zone and recl... |
4159 |
} |
333279487 memcgroup: use vm... |
4160 4161 |
static struct mem_cgroup *mem_cgroup_alloc(void) { |
d79154bb5 memcg: replace me... |
4162 |
struct mem_cgroup *memcg; |
8ff69e2c8 memcg: do not use... |
4163 |
size_t size; |
333279487 memcgroup: use vm... |
4164 |
|
8ff69e2c8 memcg: do not use... |
4165 4166 |
size = sizeof(struct mem_cgroup); size += nr_node_ids * sizeof(struct mem_cgroup_per_node *); |
333279487 memcgroup: use vm... |
4167 |
|
8ff69e2c8 memcg: do not use... |
4168 |
memcg = kzalloc(size, GFP_KERNEL); |
d79154bb5 memcg: replace me... |
4169 |
if (!memcg) |
e7bbcdf37 memcontrol: fix p... |
4170 |
return NULL; |
d79154bb5 memcg: replace me... |
4171 4172 |
memcg->stat = alloc_percpu(struct mem_cgroup_stat_cpu); if (!memcg->stat) |
d2e61b8dc memcg: null deref... |
4173 |
goto out_free; |
d79154bb5 memcg: replace me... |
4174 4175 |
spin_lock_init(&memcg->pcp_counter_lock); return memcg; |
d2e61b8dc memcg: null deref... |
4176 4177 |
out_free: |
8ff69e2c8 memcg: do not use... |
4178 |
kfree(memcg); |
d2e61b8dc memcg: null deref... |
4179 |
return NULL; |
333279487 memcgroup: use vm... |
4180 |
} |
8c7c6e34a memcg: mem+swap c... |
4181 |
/* |
c8b2a36fb memcg: execute th... |
4182 4183 4184 4185 4186 4187 4188 4189 |
* 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. |
59927fb98 memcg: free mem_c... |
4190 |
*/ |
c8b2a36fb memcg: execute th... |
4191 4192 |
static void __mem_cgroup_free(struct mem_cgroup *memcg) |
59927fb98 memcg: free mem_c... |
4193 |
{ |
c8b2a36fb memcg: execute th... |
4194 |
int node; |
59927fb98 memcg: free mem_c... |
4195 |
|
bb4cc1a8b revert "memcg: ge... |
4196 |
mem_cgroup_remove_from_trees(memcg); |
c8b2a36fb memcg: execute th... |
4197 4198 4199 4200 4201 |
for_each_node(node) free_mem_cgroup_per_zone_info(memcg, node); free_percpu(memcg->stat); |
8ff69e2c8 memcg: do not use... |
4202 |
kfree(memcg); |
59927fb98 memcg: free mem_c... |
4203 |
} |
3afe36b1f memcg: always fre... |
4204 |
|
7bcc1bb12 memcg: get/put pa... |
4205 4206 4207 |
/* * Returns the parent mem_cgroup in memcgroup hierarchy with hierarchy enabled. */ |
e1aab161e socket: initial c... |
4208 |
struct mem_cgroup *parent_mem_cgroup(struct mem_cgroup *memcg) |
7bcc1bb12 memcg: get/put pa... |
4209 |
{ |
3e32cb2e0 mm: memcontrol: l... |
4210 |
if (!memcg->memory.parent) |
7bcc1bb12 memcg: get/put pa... |
4211 |
return NULL; |
3e32cb2e0 mm: memcontrol: l... |
4212 |
return mem_cgroup_from_counter(memcg->memory.parent, memory); |
7bcc1bb12 memcg: get/put pa... |
4213 |
} |
e1aab161e socket: initial c... |
4214 |
EXPORT_SYMBOL(parent_mem_cgroup); |
333279487 memcgroup: use vm... |
4215 |
|
0eb253e22 memcg: fix sectio... |
4216 |
static struct cgroup_subsys_state * __ref |
eb95419b0 cgroup: pass arou... |
4217 |
mem_cgroup_css_alloc(struct cgroup_subsys_state *parent_css) |
8cdea7c05 Memory controller... |
4218 |
{ |
d142e3e66 memcg: split part... |
4219 |
struct mem_cgroup *memcg; |
04046e1a0 memcg: use CSS ID |
4220 |
long error = -ENOMEM; |
6d12e2d8d per-zone and recl... |
4221 |
int node; |
8cdea7c05 Memory controller... |
4222 |
|
c0ff4b854 memcg: rename mem... |
4223 4224 |
memcg = mem_cgroup_alloc(); if (!memcg) |
04046e1a0 memcg: use CSS ID |
4225 |
return ERR_PTR(error); |
78fb74669 Memory controller... |
4226 |
|
3ed28fa10 memcg: cleanup fo... |
4227 |
for_each_node(node) |
c0ff4b854 memcg: rename mem... |
4228 |
if (alloc_mem_cgroup_per_zone_info(memcg, node)) |
6d12e2d8d per-zone and recl... |
4229 |
goto free_out; |
f64c3f549 memory controller... |
4230 |
|
c077719be memcg: mem+swap c... |
4231 |
/* root ? */ |
eb95419b0 cgroup: pass arou... |
4232 |
if (parent_css == NULL) { |
a41c58a66 memcg: keep root ... |
4233 |
root_mem_cgroup = memcg; |
3e32cb2e0 mm: memcontrol: l... |
4234 |
page_counter_init(&memcg->memory, NULL); |
241994ed8 mm: memcontrol: d... |
4235 |
memcg->high = PAGE_COUNTER_MAX; |
24d404dc1 mm: memcontrol: s... |
4236 |
memcg->soft_limit = PAGE_COUNTER_MAX; |
3e32cb2e0 mm: memcontrol: l... |
4237 4238 |
page_counter_init(&memcg->memsw, NULL); page_counter_init(&memcg->kmem, NULL); |
18f59ea7d memcg: memory cgr... |
4239 |
} |
28dbc4b6a memcg: memory cgr... |
4240 |
|
d142e3e66 memcg: split part... |
4241 4242 |
memcg->last_scanned_node = MAX_NUMNODES; INIT_LIST_HEAD(&memcg->oom_notify); |
d142e3e66 memcg: split part... |
4243 4244 4245 |
memcg->move_charge_at_immigrate = 0; mutex_init(&memcg->thresholds_lock); spin_lock_init(&memcg->move_lock); |
70ddf637e memcg: add memory... |
4246 |
vmpressure_init(&memcg->vmpressure); |
fba948078 cgroup, memcg: mo... |
4247 4248 |
INIT_LIST_HEAD(&memcg->event_list); spin_lock_init(&memcg->event_list_lock); |
900a38f02 memcg: zap kmem_a... |
4249 4250 |
#ifdef CONFIG_MEMCG_KMEM memcg->kmemcg_id = -1; |
900a38f02 memcg: zap kmem_a... |
4251 |
#endif |
d142e3e66 memcg: split part... |
4252 4253 4254 4255 4256 4257 4258 4259 4260 |
return &memcg->css; free_out: __mem_cgroup_free(memcg); return ERR_PTR(error); } static int |
eb95419b0 cgroup: pass arou... |
4261 |
mem_cgroup_css_online(struct cgroup_subsys_state *css) |
d142e3e66 memcg: split part... |
4262 |
{ |
eb95419b0 cgroup: pass arou... |
4263 |
struct mem_cgroup *memcg = mem_cgroup_from_css(css); |
5c9d535b8 cgroup: remove cs... |
4264 |
struct mem_cgroup *parent = mem_cgroup_from_css(css->parent); |
2f7dd7a41 mm: memcontrol: d... |
4265 |
int ret; |
d142e3e66 memcg: split part... |
4266 |
|
15a4c835e cgroup, memcg: im... |
4267 |
if (css->id > MEM_CGROUP_ID_MAX) |
4219b2da2 memcg: fail to cr... |
4268 |
return -ENOSPC; |
638769869 cgroup: add css_p... |
4269 |
if (!parent) |
d142e3e66 memcg: split part... |
4270 |
return 0; |
0999821b1 memcg: replace cg... |
4271 |
mutex_lock(&memcg_create_mutex); |
d142e3e66 memcg: split part... |
4272 4273 4274 4275 4276 4277 |
memcg->use_hierarchy = parent->use_hierarchy; memcg->oom_kill_disable = parent->oom_kill_disable; memcg->swappiness = mem_cgroup_swappiness(parent); if (parent->use_hierarchy) { |
3e32cb2e0 mm: memcontrol: l... |
4278 |
page_counter_init(&memcg->memory, &parent->memory); |
241994ed8 mm: memcontrol: d... |
4279 |
memcg->high = PAGE_COUNTER_MAX; |
24d404dc1 mm: memcontrol: s... |
4280 |
memcg->soft_limit = PAGE_COUNTER_MAX; |
3e32cb2e0 mm: memcontrol: l... |
4281 4282 |
page_counter_init(&memcg->memsw, &parent->memsw); page_counter_init(&memcg->kmem, &parent->kmem); |
55007d849 memcg: allocate m... |
4283 |
|
7bcc1bb12 memcg: get/put pa... |
4284 |
/* |
8d76a9797 memcg: don't need... |
4285 4286 |
* No need to take a reference to the parent because cgroup * core guarantees its existence. |
7bcc1bb12 memcg: get/put pa... |
4287 |
*/ |
18f59ea7d memcg: memory cgr... |
4288 |
} else { |
3e32cb2e0 mm: memcontrol: l... |
4289 |
page_counter_init(&memcg->memory, NULL); |
241994ed8 mm: memcontrol: d... |
4290 |
memcg->high = PAGE_COUNTER_MAX; |
24d404dc1 mm: memcontrol: s... |
4291 |
memcg->soft_limit = PAGE_COUNTER_MAX; |
3e32cb2e0 mm: memcontrol: l... |
4292 4293 |
page_counter_init(&memcg->memsw, NULL); page_counter_init(&memcg->kmem, NULL); |
8c7f6edbd cgroup: mark subs... |
4294 4295 4296 4297 4298 |
/* * Deeper hierachy with use_hierarchy == false doesn't make * much sense so let cgroup subsystem know about this * unfortunate state in our controller. */ |
d142e3e66 memcg: split part... |
4299 |
if (parent != root_mem_cgroup) |
073219e99 cgroup: clean up ... |
4300 |
memory_cgrp_subsys.broken_hierarchy = true; |
18f59ea7d memcg: memory cgr... |
4301 |
} |
0999821b1 memcg: replace cg... |
4302 |
mutex_unlock(&memcg_create_mutex); |
d64416377 memcg: rework mem... |
4303 |
|
2f7dd7a41 mm: memcontrol: d... |
4304 4305 4306 4307 4308 4309 4310 4311 4312 4313 4314 4315 |
ret = memcg_init_kmem(memcg, &memory_cgrp_subsys); if (ret) return ret; /* * Make sure the memcg is initialized: mem_cgroup_iter() * orders reading memcg->initialized against its callers * reading the memcg members. */ smp_store_release(&memcg->initialized, 1); return 0; |
8cdea7c05 Memory controller... |
4316 |
} |
eb95419b0 cgroup: pass arou... |
4317 |
static void mem_cgroup_css_offline(struct cgroup_subsys_state *css) |
df878fb04 memory cgroup enh... |
4318 |
{ |
eb95419b0 cgroup: pass arou... |
4319 |
struct mem_cgroup *memcg = mem_cgroup_from_css(css); |
3bc942f37 memcg: rename cgr... |
4320 |
struct mem_cgroup_event *event, *tmp; |
79bd9814e cgroup, memcg: mo... |
4321 4322 4323 4324 4325 4326 |
/* * Unregister events and notify userspace. * Notify userspace about cgroup removing only after rmdir of cgroup * directory to avoid race between userspace and kernelspace. */ |
fba948078 cgroup, memcg: mo... |
4327 4328 |
spin_lock(&memcg->event_list_lock); list_for_each_entry_safe(event, tmp, &memcg->event_list, list) { |
79bd9814e cgroup, memcg: mo... |
4329 4330 4331 |
list_del_init(&event->list); schedule_work(&event->remove); } |
fba948078 cgroup, memcg: mo... |
4332 |
spin_unlock(&memcg->event_list_lock); |
ec64f5154 cgroup: fix frequ... |
4333 |
|
33cb876e9 vmpressure: make ... |
4334 |
vmpressure_cleanup(&memcg->vmpressure); |
2a4db7eb9 memcg: free memcg... |
4335 4336 |
memcg_deactivate_kmem(memcg); |
df878fb04 memory cgroup enh... |
4337 |
} |
eb95419b0 cgroup: pass arou... |
4338 |
static void mem_cgroup_css_free(struct cgroup_subsys_state *css) |
8cdea7c05 Memory controller... |
4339 |
{ |
eb95419b0 cgroup: pass arou... |
4340 |
struct mem_cgroup *memcg = mem_cgroup_from_css(css); |
c268e9946 memcg: fix hierar... |
4341 |
|
10d5ebf40 memcg: use css_ge... |
4342 |
memcg_destroy_kmem(memcg); |
465939a1f memcg: don't need... |
4343 |
__mem_cgroup_free(memcg); |
8cdea7c05 Memory controller... |
4344 |
} |
1ced953b1 blkcg, memcg: mak... |
4345 4346 4347 4348 4349 4350 4351 4352 4353 4354 4355 4356 4357 4358 4359 4360 |
/** * mem_cgroup_css_reset - reset the states of a mem_cgroup * @css: the target css * * Reset the states of the mem_cgroup associated with @css. This is * invoked when the userland requests disabling on the default hierarchy * but the memcg is pinned through dependency. The memcg should stop * applying policies and should revert to the vanilla state as it may be * made visible again. * * The current implementation only resets the essential configurations. * This needs to be expanded to cover all the visible parts. */ static void mem_cgroup_css_reset(struct cgroup_subsys_state *css) { struct mem_cgroup *memcg = mem_cgroup_from_css(css); |
3e32cb2e0 mm: memcontrol: l... |
4361 4362 4363 |
mem_cgroup_resize_limit(memcg, PAGE_COUNTER_MAX); mem_cgroup_resize_memsw_limit(memcg, PAGE_COUNTER_MAX); memcg_update_kmem_limit(memcg, PAGE_COUNTER_MAX); |
241994ed8 mm: memcontrol: d... |
4364 4365 |
memcg->low = 0; memcg->high = PAGE_COUNTER_MAX; |
24d404dc1 mm: memcontrol: s... |
4366 |
memcg->soft_limit = PAGE_COUNTER_MAX; |
1ced953b1 blkcg, memcg: mak... |
4367 |
} |
024914477 memcg: move charg... |
4368 |
#ifdef CONFIG_MMU |
7dc74be03 memcg: add interf... |
4369 |
/* Handlers for move charge at task migration. */ |
854ffa8d1 memcg: improve pe... |
4370 |
static int mem_cgroup_do_precharge(unsigned long count) |
7dc74be03 memcg: add interf... |
4371 |
{ |
05b843012 mm: memcontrol: u... |
4372 |
int ret; |
9476db974 mm: memcontrol: s... |
4373 4374 |
/* Try a single bulk charge without reclaim first */ |
00501b531 mm: memcontrol: r... |
4375 |
ret = try_charge(mc.to, GFP_KERNEL & ~__GFP_WAIT, count); |
9476db974 mm: memcontrol: s... |
4376 |
if (!ret) { |
854ffa8d1 memcg: improve pe... |
4377 |
mc.precharge += count; |
854ffa8d1 memcg: improve pe... |
4378 4379 |
return ret; } |
692e7c45d mm: memcontrol: c... |
4380 |
if (ret == -EINTR) { |
00501b531 mm: memcontrol: r... |
4381 |
cancel_charge(root_mem_cgroup, count); |
692e7c45d mm: memcontrol: c... |
4382 4383 |
return ret; } |
9476db974 mm: memcontrol: s... |
4384 4385 |
/* Try charges one by one with reclaim */ |
854ffa8d1 memcg: improve pe... |
4386 |
while (count--) { |
00501b531 mm: memcontrol: r... |
4387 |
ret = try_charge(mc.to, GFP_KERNEL & ~__GFP_NORETRY, 1); |
9476db974 mm: memcontrol: s... |
4388 4389 4390 |
/* * In case of failure, any residual charges against * mc.to will be dropped by mem_cgroup_clear_mc() |
692e7c45d mm: memcontrol: c... |
4391 4392 |
* later on. However, cancel any charges that are * bypassed to root right away or they'll be lost. |
9476db974 mm: memcontrol: s... |
4393 |
*/ |
692e7c45d mm: memcontrol: c... |
4394 |
if (ret == -EINTR) |
00501b531 mm: memcontrol: r... |
4395 |
cancel_charge(root_mem_cgroup, 1); |
38c5d72f3 memcg: simplify L... |
4396 |
if (ret) |
38c5d72f3 memcg: simplify L... |
4397 |
return ret; |
854ffa8d1 memcg: improve pe... |
4398 |
mc.precharge++; |
9476db974 mm: memcontrol: s... |
4399 |
cond_resched(); |
854ffa8d1 memcg: improve pe... |
4400 |
} |
9476db974 mm: memcontrol: s... |
4401 |
return 0; |
4ffef5fef memcg: move charg... |
4402 4403 4404 |
} /** |
8d32ff844 memcg: clean up e... |
4405 |
* get_mctgt_type - get target type of moving charge |
4ffef5fef memcg: move charg... |
4406 4407 4408 |
* @vma: the vma the pte to be checked belongs * @addr: the address corresponding to the pte to be checked * @ptent: the pte to be checked |
024914477 memcg: move charg... |
4409 |
* @target: the pointer the target page or swap ent will be stored(can be NULL) |
4ffef5fef memcg: move charg... |
4410 4411 4412 4413 4414 4415 |
* * Returns * 0(MC_TARGET_NONE): if the pte is not a target for move charge. * 1(MC_TARGET_PAGE): if the page corresponding to this pte is a target for * move charge. if @target is not NULL, the page is stored in target->page * with extra refcnt got(Callers should handle it). |
024914477 memcg: move charg... |
4416 4417 4418 |
* 2(MC_TARGET_SWAP): if the swap entry corresponding to this pte is a * target for charge migration. if @target is not NULL, the entry is stored * in target->ent. |
4ffef5fef memcg: move charg... |
4419 4420 4421 |
* * Called with pte lock held. */ |
4ffef5fef memcg: move charg... |
4422 4423 |
union mc_target { struct page *page; |
024914477 memcg: move charg... |
4424 |
swp_entry_t ent; |
4ffef5fef memcg: move charg... |
4425 |
}; |
4ffef5fef memcg: move charg... |
4426 |
enum mc_target_type { |
8d32ff844 memcg: clean up e... |
4427 |
MC_TARGET_NONE = 0, |
4ffef5fef memcg: move charg... |
4428 |
MC_TARGET_PAGE, |
024914477 memcg: move charg... |
4429 |
MC_TARGET_SWAP, |
4ffef5fef memcg: move charg... |
4430 |
}; |
90254a658 memcg: clean up m... |
4431 4432 |
static struct page *mc_handle_present_pte(struct vm_area_struct *vma, unsigned long addr, pte_t ptent) |
4ffef5fef memcg: move charg... |
4433 |
{ |
90254a658 memcg: clean up m... |
4434 |
struct page *page = vm_normal_page(vma, addr, ptent); |
4ffef5fef memcg: move charg... |
4435 |
|
90254a658 memcg: clean up m... |
4436 4437 4438 |
if (!page || !page_mapped(page)) return NULL; if (PageAnon(page)) { |
1dfab5abc mm: memcontrol: f... |
4439 |
if (!(mc.flags & MOVE_ANON)) |
90254a658 memcg: clean up m... |
4440 |
return NULL; |
1dfab5abc mm: memcontrol: f... |
4441 4442 4443 4444 |
} else { if (!(mc.flags & MOVE_FILE)) return NULL; } |
90254a658 memcg: clean up m... |
4445 4446 4447 4448 4449 |
if (!get_page_unless_zero(page)) return NULL; return page; } |
4b91355e9 memcg: fix/change... |
4450 |
#ifdef CONFIG_SWAP |
90254a658 memcg: clean up m... |
4451 4452 4453 |
static struct page *mc_handle_swap_pte(struct vm_area_struct *vma, unsigned long addr, pte_t ptent, swp_entry_t *entry) { |
90254a658 memcg: clean up m... |
4454 4455 |
struct page *page = NULL; swp_entry_t ent = pte_to_swp_entry(ptent); |
1dfab5abc mm: memcontrol: f... |
4456 |
if (!(mc.flags & MOVE_ANON) || non_swap_entry(ent)) |
90254a658 memcg: clean up m... |
4457 |
return NULL; |
4b91355e9 memcg: fix/change... |
4458 4459 4460 4461 |
/* * Because lookup_swap_cache() updates some statistics counter, * we call find_get_page() with swapper_space directly. */ |
33806f06d swap: make each s... |
4462 |
page = find_get_page(swap_address_space(ent), ent.val); |
90254a658 memcg: clean up m... |
4463 4464 4465 4466 4467 |
if (do_swap_account) entry->val = ent.val; return page; } |
4b91355e9 memcg: fix/change... |
4468 4469 4470 4471 4472 4473 4474 |
#else static struct page *mc_handle_swap_pte(struct vm_area_struct *vma, unsigned long addr, pte_t ptent, swp_entry_t *entry) { return NULL; } #endif |
90254a658 memcg: clean up m... |
4475 |
|
87946a722 memcg: move charg... |
4476 4477 4478 4479 |
static struct page *mc_handle_file_pte(struct vm_area_struct *vma, unsigned long addr, pte_t ptent, swp_entry_t *entry) { struct page *page = NULL; |
87946a722 memcg: move charg... |
4480 4481 4482 4483 4484 |
struct address_space *mapping; pgoff_t pgoff; if (!vma->vm_file) /* anonymous vma */ return NULL; |
1dfab5abc mm: memcontrol: f... |
4485 |
if (!(mc.flags & MOVE_FILE)) |
87946a722 memcg: move charg... |
4486 |
return NULL; |
87946a722 memcg: move charg... |
4487 |
mapping = vma->vm_file->f_mapping; |
0661a3361 mm: remove rest u... |
4488 |
pgoff = linear_page_index(vma, addr); |
87946a722 memcg: move charg... |
4489 4490 |
/* page is moved even if it's not RSS of this task(page-faulted). */ |
aa3b18955 tmpfs: convert me... |
4491 4492 |
#ifdef CONFIG_SWAP /* shmem/tmpfs may report page out on swap: account for that too. */ |
139b6a6fb mm: filemap: upda... |
4493 4494 4495 4496 4497 4498 4499 4500 4501 4502 4503 4504 |
if (shmem_mapping(mapping)) { page = find_get_entry(mapping, pgoff); if (radix_tree_exceptional_entry(page)) { swp_entry_t swp = radix_to_swp_entry(page); if (do_swap_account) *entry = swp; page = find_get_page(swap_address_space(swp), swp.val); } } else page = find_get_page(mapping, pgoff); #else page = find_get_page(mapping, pgoff); |
aa3b18955 tmpfs: convert me... |
4505 |
#endif |
87946a722 memcg: move charg... |
4506 4507 |
return page; } |
b1b0deabb mm: memcontrol: l... |
4508 4509 4510 4511 4512 4513 4514 4515 4516 4517 4518 4519 4520 4521 4522 4523 4524 4525 4526 4527 4528 4529 4530 4531 4532 4533 4534 4535 4536 4537 4538 4539 4540 4541 4542 4543 4544 4545 4546 4547 4548 4549 4550 4551 4552 4553 4554 4555 4556 4557 4558 4559 4560 4561 4562 4563 4564 4565 4566 4567 4568 4569 4570 4571 4572 4573 4574 4575 4576 4577 4578 4579 4580 4581 4582 4583 4584 4585 4586 4587 4588 4589 4590 4591 4592 |
/** * mem_cgroup_move_account - move account of the page * @page: the page * @nr_pages: number of regular pages (>1 for huge pages) * @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. * - page is not on LRU (isolate_page() is useful.) * - compound_lock is held when nr_pages > 1 * * This function doesn't do "charge" to new cgroup and doesn't do "uncharge" * from old cgroup. */ static int mem_cgroup_move_account(struct page *page, unsigned int nr_pages, struct mem_cgroup *from, struct mem_cgroup *to) { unsigned long flags; int ret; VM_BUG_ON(from == to); VM_BUG_ON_PAGE(PageLRU(page), page); /* * The page is isolated from LRU. So, collapse function * will not handle this page. But page splitting can happen. * Do this check under compound_page_lock(). The caller should * hold it. */ ret = -EBUSY; if (nr_pages > 1 && !PageTransHuge(page)) goto out; /* * Prevent mem_cgroup_migrate() from looking at page->mem_cgroup * of its source page while we change it: page migration takes * both pages off the LRU, but page cache replacement doesn't. */ if (!trylock_page(page)) goto out; ret = -EINVAL; if (page->mem_cgroup != from) goto out_unlock; spin_lock_irqsave(&from->move_lock, flags); if (!PageAnon(page) && page_mapped(page)) { __this_cpu_sub(from->stat->count[MEM_CGROUP_STAT_FILE_MAPPED], nr_pages); __this_cpu_add(to->stat->count[MEM_CGROUP_STAT_FILE_MAPPED], nr_pages); } if (PageWriteback(page)) { __this_cpu_sub(from->stat->count[MEM_CGROUP_STAT_WRITEBACK], nr_pages); __this_cpu_add(to->stat->count[MEM_CGROUP_STAT_WRITEBACK], nr_pages); } /* * It is safe to change page->mem_cgroup here because the page * is referenced, charged, and isolated - we can't race with * uncharging, charging, migration, or LRU putback. */ /* caller should have done css_get */ page->mem_cgroup = to; spin_unlock_irqrestore(&from->move_lock, flags); ret = 0; local_irq_disable(); mem_cgroup_charge_statistics(to, page, nr_pages); memcg_check_events(to, page); mem_cgroup_charge_statistics(from, page, -nr_pages); memcg_check_events(from, page); local_irq_enable(); out_unlock: unlock_page(page); out: return ret; } |
8d32ff844 memcg: clean up e... |
4593 |
static enum mc_target_type get_mctgt_type(struct vm_area_struct *vma, |
90254a658 memcg: clean up m... |
4594 4595 4596 |
unsigned long addr, pte_t ptent, union mc_target *target) { struct page *page = NULL; |
8d32ff844 memcg: clean up e... |
4597 |
enum mc_target_type ret = MC_TARGET_NONE; |
90254a658 memcg: clean up m... |
4598 4599 4600 4601 4602 4603 |
swp_entry_t ent = { .val = 0 }; if (pte_present(ptent)) page = mc_handle_present_pte(vma, addr, ptent); else if (is_swap_pte(ptent)) page = mc_handle_swap_pte(vma, addr, ptent, &ent); |
0661a3361 mm: remove rest u... |
4604 |
else if (pte_none(ptent)) |
87946a722 memcg: move charg... |
4605 |
page = mc_handle_file_pte(vma, addr, ptent, &ent); |
90254a658 memcg: clean up m... |
4606 4607 |
if (!page && !ent.val) |
8d32ff844 memcg: clean up e... |
4608 |
return ret; |
024914477 memcg: move charg... |
4609 |
if (page) { |
024914477 memcg: move charg... |
4610 |
/* |
0a31bc97c mm: memcontrol: r... |
4611 |
* Do only loose check w/o serialization. |
1306a85ae mm: embed the mem... |
4612 |
* mem_cgroup_move_account() checks the page is valid or |
0a31bc97c mm: memcontrol: r... |
4613 |
* not under LRU exclusion. |
024914477 memcg: move charg... |
4614 |
*/ |
1306a85ae mm: embed the mem... |
4615 |
if (page->mem_cgroup == mc.from) { |
024914477 memcg: move charg... |
4616 4617 4618 4619 4620 4621 4622 |
ret = MC_TARGET_PAGE; if (target) target->page = page; } if (!ret || !target) put_page(page); } |
90254a658 memcg: clean up m... |
4623 4624 |
/* There is a swap entry and a page doesn't exist or isn't charged */ if (ent.val && !ret && |
34c00c319 memcg: convert to... |
4625 |
mem_cgroup_id(mc.from) == lookup_swap_cgroup_id(ent)) { |
7f0f15464 memcg: fix css_id... |
4626 4627 4628 |
ret = MC_TARGET_SWAP; if (target) target->ent = ent; |
4ffef5fef memcg: move charg... |
4629 |
} |
4ffef5fef memcg: move charg... |
4630 4631 |
return ret; } |
12724850e memcg: avoid THP ... |
4632 4633 4634 4635 4636 4637 4638 4639 4640 4641 |
#ifdef CONFIG_TRANSPARENT_HUGEPAGE /* * We don't consider swapping or file mapped pages because THP does not * support them for now. * Caller should make sure that pmd_trans_huge(pmd) is true. */ static enum mc_target_type get_mctgt_type_thp(struct vm_area_struct *vma, unsigned long addr, pmd_t pmd, union mc_target *target) { struct page *page = NULL; |
12724850e memcg: avoid THP ... |
4642 4643 4644 |
enum mc_target_type ret = MC_TARGET_NONE; page = pmd_page(pmd); |
309381fea mm: dump page whe... |
4645 |
VM_BUG_ON_PAGE(!page || !PageHead(page), page); |
1dfab5abc mm: memcontrol: f... |
4646 |
if (!(mc.flags & MOVE_ANON)) |
12724850e memcg: avoid THP ... |
4647 |
return ret; |
1306a85ae mm: embed the mem... |
4648 |
if (page->mem_cgroup == mc.from) { |
12724850e memcg: avoid THP ... |
4649 4650 4651 4652 4653 4654 4655 4656 4657 4658 4659 4660 4661 4662 4663 |
ret = MC_TARGET_PAGE; if (target) { get_page(page); target->page = page; } } return ret; } #else static inline enum mc_target_type get_mctgt_type_thp(struct vm_area_struct *vma, unsigned long addr, pmd_t pmd, union mc_target *target) { return MC_TARGET_NONE; } #endif |
4ffef5fef memcg: move charg... |
4664 4665 4666 4667 |
static int mem_cgroup_count_precharge_pte_range(pmd_t *pmd, unsigned long addr, unsigned long end, struct mm_walk *walk) { |
26bcd64aa memcg: cleanup pr... |
4668 |
struct vm_area_struct *vma = walk->vma; |
4ffef5fef memcg: move charg... |
4669 4670 |
pte_t *pte; spinlock_t *ptl; |
bf929152e mm, thp: change p... |
4671 |
if (pmd_trans_huge_lock(pmd, vma, &ptl) == 1) { |
12724850e memcg: avoid THP ... |
4672 4673 |
if (get_mctgt_type_thp(vma, addr, *pmd, NULL) == MC_TARGET_PAGE) mc.precharge += HPAGE_PMD_NR; |
bf929152e mm, thp: change p... |
4674 |
spin_unlock(ptl); |
1a5a9906d mm: thp: fix pmd_... |
4675 |
return 0; |
12724850e memcg: avoid THP ... |
4676 |
} |
033193275 pagewalk: only sp... |
4677 |
|
45f83cefe mm: thp: fix up p... |
4678 4679 |
if (pmd_trans_unstable(pmd)) return 0; |
4ffef5fef memcg: move charg... |
4680 4681 |
pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl); for (; addr != end; pte++, addr += PAGE_SIZE) |
8d32ff844 memcg: clean up e... |
4682 |
if (get_mctgt_type(vma, addr, *pte, NULL)) |
4ffef5fef memcg: move charg... |
4683 4684 4685 |
mc.precharge++; /* increment precharge temporarily */ pte_unmap_unlock(pte - 1, ptl); cond_resched(); |
7dc74be03 memcg: add interf... |
4686 4687 |
return 0; } |
4ffef5fef memcg: move charg... |
4688 4689 4690 |
static unsigned long mem_cgroup_count_precharge(struct mm_struct *mm) { unsigned long precharge; |
4ffef5fef memcg: move charg... |
4691 |
|
26bcd64aa memcg: cleanup pr... |
4692 4693 4694 4695 |
struct mm_walk mem_cgroup_count_precharge_walk = { .pmd_entry = mem_cgroup_count_precharge_pte_range, .mm = mm, }; |
dfe076b09 memcg: fix deadlo... |
4696 |
down_read(&mm->mmap_sem); |
26bcd64aa memcg: cleanup pr... |
4697 |
walk_page_range(0, ~0UL, &mem_cgroup_count_precharge_walk); |
dfe076b09 memcg: fix deadlo... |
4698 |
up_read(&mm->mmap_sem); |
4ffef5fef memcg: move charg... |
4699 4700 4701 4702 4703 4704 |
precharge = mc.precharge; mc.precharge = 0; return precharge; } |
4ffef5fef memcg: move charg... |
4705 4706 |
static int mem_cgroup_precharge_mc(struct mm_struct *mm) { |
dfe076b09 memcg: fix deadlo... |
4707 4708 4709 4710 4711 |
unsigned long precharge = mem_cgroup_count_precharge(mm); VM_BUG_ON(mc.moving_task); mc.moving_task = current; return mem_cgroup_do_precharge(precharge); |
4ffef5fef memcg: move charg... |
4712 |
} |
dfe076b09 memcg: fix deadlo... |
4713 4714 |
/* cancels all extra charges on mc.from and mc.to, and wakes up all waiters. */ static void __mem_cgroup_clear_mc(void) |
4ffef5fef memcg: move charg... |
4715 |
{ |
2bd9bb206 memcg: clean up w... |
4716 4717 |
struct mem_cgroup *from = mc.from; struct mem_cgroup *to = mc.to; |
4ffef5fef memcg: move charg... |
4718 |
/* we must uncharge all the leftover precharges from mc.to */ |
854ffa8d1 memcg: improve pe... |
4719 |
if (mc.precharge) { |
00501b531 mm: memcontrol: r... |
4720 |
cancel_charge(mc.to, mc.precharge); |
854ffa8d1 memcg: improve pe... |
4721 4722 4723 4724 4725 4726 4727 |
mc.precharge = 0; } /* * we didn't uncharge from mc.from at mem_cgroup_move_account(), so * we must uncharge here. */ if (mc.moved_charge) { |
00501b531 mm: memcontrol: r... |
4728 |
cancel_charge(mc.from, mc.moved_charge); |
854ffa8d1 memcg: improve pe... |
4729 |
mc.moved_charge = 0; |
4ffef5fef memcg: move charg... |
4730 |
} |
483c30b51 memcg: improve pe... |
4731 4732 |
/* we must fixup refcnts and charges */ if (mc.moved_swap) { |
483c30b51 memcg: improve pe... |
4733 |
/* uncharge swap account from the old cgroup */ |
ce00a9673 mm: memcontrol: r... |
4734 |
if (!mem_cgroup_is_root(mc.from)) |
3e32cb2e0 mm: memcontrol: l... |
4735 |
page_counter_uncharge(&mc.from->memsw, mc.moved_swap); |
483c30b51 memcg: improve pe... |
4736 |
|
05b843012 mm: memcontrol: u... |
4737 |
/* |
3e32cb2e0 mm: memcontrol: l... |
4738 4739 |
* we charged both to->memory and to->memsw, so we * should uncharge to->memory. |
05b843012 mm: memcontrol: u... |
4740 |
*/ |
ce00a9673 mm: memcontrol: r... |
4741 |
if (!mem_cgroup_is_root(mc.to)) |
3e32cb2e0 mm: memcontrol: l... |
4742 |
page_counter_uncharge(&mc.to->memory, mc.moved_swap); |
e8ea14cc6 mm: memcontrol: t... |
4743 |
css_put_many(&mc.from->css, mc.moved_swap); |
3e32cb2e0 mm: memcontrol: l... |
4744 |
|
4050377b5 memcg: use css_ge... |
4745 |
/* we've already done css_get(mc.to) */ |
483c30b51 memcg: improve pe... |
4746 4747 |
mc.moved_swap = 0; } |
dfe076b09 memcg: fix deadlo... |
4748 4749 4750 4751 4752 4753 4754 |
memcg_oom_recover(from); memcg_oom_recover(to); wake_up_all(&mc.waitq); } static void mem_cgroup_clear_mc(void) { |
dfe076b09 memcg: fix deadlo... |
4755 4756 4757 4758 4759 4760 |
/* * we must clear moving_task before waking up waiters at the end of * task migration. */ mc.moving_task = NULL; __mem_cgroup_clear_mc(); |
2bd9bb206 memcg: clean up w... |
4761 |
spin_lock(&mc.lock); |
4ffef5fef memcg: move charg... |
4762 4763 |
mc.from = NULL; mc.to = NULL; |
2bd9bb206 memcg: clean up w... |
4764 |
spin_unlock(&mc.lock); |
4ffef5fef memcg: move charg... |
4765 |
} |
eb95419b0 cgroup: pass arou... |
4766 |
static int mem_cgroup_can_attach(struct cgroup_subsys_state *css, |
761b3ef50 cgroup: remove cg... |
4767 |
struct cgroup_taskset *tset) |
7dc74be03 memcg: add interf... |
4768 |
{ |
2f7ee5691 cgroup: introduce... |
4769 |
struct task_struct *p = cgroup_taskset_first(tset); |
7dc74be03 memcg: add interf... |
4770 |
int ret = 0; |
eb95419b0 cgroup: pass arou... |
4771 |
struct mem_cgroup *memcg = mem_cgroup_from_css(css); |
1dfab5abc mm: memcontrol: f... |
4772 |
unsigned long move_flags; |
7dc74be03 memcg: add interf... |
4773 |
|
ee5e8472b memcg: prevent ch... |
4774 4775 4776 4777 4778 |
/* * We are now commited to this value whatever it is. Changes in this * tunable will only affect upcoming migrations, not the current one. * So we need to save it, and keep it going. */ |
4db0c3c29 mm: remove rest o... |
4779 |
move_flags = READ_ONCE(memcg->move_charge_at_immigrate); |
1dfab5abc mm: memcontrol: f... |
4780 |
if (move_flags) { |
7dc74be03 memcg: add interf... |
4781 4782 |
struct mm_struct *mm; struct mem_cgroup *from = mem_cgroup_from_task(p); |
c0ff4b854 memcg: rename mem... |
4783 |
VM_BUG_ON(from == memcg); |
7dc74be03 memcg: add interf... |
4784 4785 4786 4787 |
mm = get_task_mm(p); if (!mm) return 0; |
7dc74be03 memcg: add interf... |
4788 |
/* We move charges only when we move a owner of the mm */ |
4ffef5fef memcg: move charg... |
4789 4790 4791 4792 |
if (mm->owner == p) { VM_BUG_ON(mc.from); VM_BUG_ON(mc.to); VM_BUG_ON(mc.precharge); |
854ffa8d1 memcg: improve pe... |
4793 |
VM_BUG_ON(mc.moved_charge); |
483c30b51 memcg: improve pe... |
4794 |
VM_BUG_ON(mc.moved_swap); |
247b1447b mm: memcontrol: f... |
4795 |
|
2bd9bb206 memcg: clean up w... |
4796 |
spin_lock(&mc.lock); |
4ffef5fef memcg: move charg... |
4797 |
mc.from = from; |
c0ff4b854 memcg: rename mem... |
4798 |
mc.to = memcg; |
1dfab5abc mm: memcontrol: f... |
4799 |
mc.flags = move_flags; |
2bd9bb206 memcg: clean up w... |
4800 |
spin_unlock(&mc.lock); |
dfe076b09 memcg: fix deadlo... |
4801 |
/* We set mc.moving_task later */ |
4ffef5fef memcg: move charg... |
4802 4803 4804 4805 |
ret = mem_cgroup_precharge_mc(mm); if (ret) mem_cgroup_clear_mc(); |
dfe076b09 memcg: fix deadlo... |
4806 4807 |
} mmput(mm); |
7dc74be03 memcg: add interf... |
4808 4809 4810 |
} return ret; } |
eb95419b0 cgroup: pass arou... |
4811 |
static void mem_cgroup_cancel_attach(struct cgroup_subsys_state *css, |
761b3ef50 cgroup: remove cg... |
4812 |
struct cgroup_taskset *tset) |
7dc74be03 memcg: add interf... |
4813 |
{ |
4e2f245d3 mm: memcontrol: d... |
4814 4815 |
if (mc.to) mem_cgroup_clear_mc(); |
7dc74be03 memcg: add interf... |
4816 |
} |
4ffef5fef memcg: move charg... |
4817 4818 4819 |
static int mem_cgroup_move_charge_pte_range(pmd_t *pmd, unsigned long addr, unsigned long end, struct mm_walk *walk) |
7dc74be03 memcg: add interf... |
4820 |
{ |
4ffef5fef memcg: move charg... |
4821 |
int ret = 0; |
26bcd64aa memcg: cleanup pr... |
4822 |
struct vm_area_struct *vma = walk->vma; |
4ffef5fef memcg: move charg... |
4823 4824 |
pte_t *pte; spinlock_t *ptl; |
12724850e memcg: avoid THP ... |
4825 4826 4827 |
enum mc_target_type target_type; union mc_target target; struct page *page; |
4ffef5fef memcg: move charg... |
4828 |
|
12724850e memcg: avoid THP ... |
4829 4830 4831 4832 4833 4834 4835 4836 4837 4838 |
/* * We don't take compound_lock() here but no race with splitting thp * happens because: * - if pmd_trans_huge_lock() returns 1, the relevant thp is not * under splitting, which means there's no concurrent thp split, * - if another thread runs into split_huge_page() just after we * entered this if-block, the thread must wait for page table lock * to be unlocked in __split_huge_page_splitting(), where the main * part of thp split is not executed yet. */ |
bf929152e mm, thp: change p... |
4839 |
if (pmd_trans_huge_lock(pmd, vma, &ptl) == 1) { |
62ade86ab memcg,thp: fix re... |
4840 |
if (mc.precharge < HPAGE_PMD_NR) { |
bf929152e mm, thp: change p... |
4841 |
spin_unlock(ptl); |
12724850e memcg: avoid THP ... |
4842 4843 4844 4845 4846 4847 |
return 0; } target_type = get_mctgt_type_thp(vma, addr, *pmd, &target); if (target_type == MC_TARGET_PAGE) { page = target.page; if (!isolate_lru_page(page)) { |
12724850e memcg: avoid THP ... |
4848 |
if (!mem_cgroup_move_account(page, HPAGE_PMD_NR, |
1306a85ae mm: embed the mem... |
4849 |
mc.from, mc.to)) { |
12724850e memcg: avoid THP ... |
4850 4851 4852 4853 4854 4855 4856 |
mc.precharge -= HPAGE_PMD_NR; mc.moved_charge += HPAGE_PMD_NR; } putback_lru_page(page); } put_page(page); } |
bf929152e mm, thp: change p... |
4857 |
spin_unlock(ptl); |
1a5a9906d mm: thp: fix pmd_... |
4858 |
return 0; |
12724850e memcg: avoid THP ... |
4859 |
} |
45f83cefe mm: thp: fix up p... |
4860 4861 |
if (pmd_trans_unstable(pmd)) return 0; |
4ffef5fef memcg: move charg... |
4862 4863 4864 4865 |
retry: pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl); for (; addr != end; addr += PAGE_SIZE) { pte_t ptent = *(pte++); |
024914477 memcg: move charg... |
4866 |
swp_entry_t ent; |
4ffef5fef memcg: move charg... |
4867 4868 4869 |
if (!mc.precharge) break; |
8d32ff844 memcg: clean up e... |
4870 |
switch (get_mctgt_type(vma, addr, ptent, &target)) { |
4ffef5fef memcg: move charg... |
4871 4872 4873 4874 |
case MC_TARGET_PAGE: page = target.page; if (isolate_lru_page(page)) goto put; |
1306a85ae mm: embed the mem... |
4875 |
if (!mem_cgroup_move_account(page, 1, mc.from, mc.to)) { |
4ffef5fef memcg: move charg... |
4876 |
mc.precharge--; |
854ffa8d1 memcg: improve pe... |
4877 4878 |
/* we uncharge from mc.from later. */ mc.moved_charge++; |
4ffef5fef memcg: move charg... |
4879 4880 |
} putback_lru_page(page); |
8d32ff844 memcg: clean up e... |
4881 |
put: /* get_mctgt_type() gets the page */ |
4ffef5fef memcg: move charg... |
4882 4883 |
put_page(page); break; |
024914477 memcg: move charg... |
4884 4885 |
case MC_TARGET_SWAP: ent = target.ent; |
e91cbb425 memcg swap: mem_c... |
4886 |
if (!mem_cgroup_move_swap_account(ent, mc.from, mc.to)) { |
024914477 memcg: move charg... |
4887 |
mc.precharge--; |
483c30b51 memcg: improve pe... |
4888 4889 4890 |
/* we fixup refcnts and charges later. */ mc.moved_swap++; } |
024914477 memcg: move charg... |
4891 |
break; |
4ffef5fef memcg: move charg... |
4892 4893 4894 4895 4896 4897 4898 4899 4900 4901 4902 4903 4904 4905 |
default: break; } } pte_unmap_unlock(pte - 1, ptl); cond_resched(); if (addr != end) { /* * We have consumed all precharges we got in can_attach(). * We try charge one by one, but don't do any additional * charges to mc.to if we have failed in charge once in attach() * phase. */ |
854ffa8d1 memcg: improve pe... |
4906 |
ret = mem_cgroup_do_precharge(1); |
4ffef5fef memcg: move charg... |
4907 4908 4909 4910 4911 4912 4913 4914 4915 |
if (!ret) goto retry; } return ret; } static void mem_cgroup_move_charge(struct mm_struct *mm) { |
26bcd64aa memcg: cleanup pr... |
4916 4917 4918 4919 |
struct mm_walk mem_cgroup_move_charge_walk = { .pmd_entry = mem_cgroup_move_charge_pte_range, .mm = mm, }; |
4ffef5fef memcg: move charg... |
4920 4921 |
lru_add_drain_all(); |
312722cbb mm: memcontrol: s... |
4922 4923 4924 4925 4926 4927 4928 |
/* * Signal mem_cgroup_begin_page_stat() to take the memcg's * move_lock while we're moving its pages to another memcg. * Then wait for already started RCU-only updates to finish. */ atomic_inc(&mc.from->moving_account); synchronize_rcu(); |
dfe076b09 memcg: fix deadlo... |
4929 4930 4931 4932 4933 4934 4935 4936 4937 4938 4939 4940 4941 |
retry: if (unlikely(!down_read_trylock(&mm->mmap_sem))) { /* * Someone who are holding the mmap_sem might be waiting in * waitq. So we cancel all extra charges, wake up all waiters, * and retry. Because we cancel precharges, we might not be able * to move enough charges, but moving charge is a best-effort * feature anyway, so it wouldn't be a big problem. */ __mem_cgroup_clear_mc(); cond_resched(); goto retry; } |
26bcd64aa memcg: cleanup pr... |
4942 4943 4944 4945 4946 |
/* * When we have consumed all precharges and failed in doing * additional charge, the page walk just aborts. */ walk_page_range(0, ~0UL, &mem_cgroup_move_charge_walk); |
dfe076b09 memcg: fix deadlo... |
4947 |
up_read(&mm->mmap_sem); |
312722cbb mm: memcontrol: s... |
4948 |
atomic_dec(&mc.from->moving_account); |
7dc74be03 memcg: add interf... |
4949 |
} |
eb95419b0 cgroup: pass arou... |
4950 |
static void mem_cgroup_move_task(struct cgroup_subsys_state *css, |
761b3ef50 cgroup: remove cg... |
4951 |
struct cgroup_taskset *tset) |
67e465a77 Memory controller... |
4952 |
{ |
2f7ee5691 cgroup: introduce... |
4953 |
struct task_struct *p = cgroup_taskset_first(tset); |
a433658c3 vmscan,memcg: mem... |
4954 |
struct mm_struct *mm = get_task_mm(p); |
dfe076b09 memcg: fix deadlo... |
4955 |
|
dfe076b09 memcg: fix deadlo... |
4956 |
if (mm) { |
a433658c3 vmscan,memcg: mem... |
4957 4958 |
if (mc.to) mem_cgroup_move_charge(mm); |
dfe076b09 memcg: fix deadlo... |
4959 4960 |
mmput(mm); } |
a433658c3 vmscan,memcg: mem... |
4961 4962 |
if (mc.to) mem_cgroup_clear_mc(); |
67e465a77 Memory controller... |
4963 |
} |
5cfb80a73 memcg: disable mo... |
4964 |
#else /* !CONFIG_MMU */ |
eb95419b0 cgroup: pass arou... |
4965 |
static int mem_cgroup_can_attach(struct cgroup_subsys_state *css, |
761b3ef50 cgroup: remove cg... |
4966 |
struct cgroup_taskset *tset) |
5cfb80a73 memcg: disable mo... |
4967 4968 4969 |
{ return 0; } |
eb95419b0 cgroup: pass arou... |
4970 |
static void mem_cgroup_cancel_attach(struct cgroup_subsys_state *css, |
761b3ef50 cgroup: remove cg... |
4971 |
struct cgroup_taskset *tset) |
5cfb80a73 memcg: disable mo... |
4972 4973 |
{ } |
eb95419b0 cgroup: pass arou... |
4974 |
static void mem_cgroup_move_task(struct cgroup_subsys_state *css, |
761b3ef50 cgroup: remove cg... |
4975 |
struct cgroup_taskset *tset) |
5cfb80a73 memcg: disable mo... |
4976 4977 4978 |
{ } #endif |
67e465a77 Memory controller... |
4979 |
|
f00baae7a memcg: force use_... |
4980 4981 |
/* * Cgroup retains root cgroups across [un]mount cycles making it necessary |
aa6ec29be cgroup: remove sa... |
4982 4983 |
* to verify whether we're attached to the default hierarchy on each mount * attempt. |
f00baae7a memcg: force use_... |
4984 |
*/ |
eb95419b0 cgroup: pass arou... |
4985 |
static void mem_cgroup_bind(struct cgroup_subsys_state *root_css) |
f00baae7a memcg: force use_... |
4986 4987 |
{ /* |
aa6ec29be cgroup: remove sa... |
4988 |
* use_hierarchy is forced on the default hierarchy. cgroup core |
f00baae7a memcg: force use_... |
4989 4990 4991 |
* guarantees that @root doesn't have any children, so turning it * on for the root memcg is enough. */ |
aa6ec29be cgroup: remove sa... |
4992 |
if (cgroup_on_dfl(root_css->cgroup)) |
7feee590b memcg: disable hi... |
4993 4994 4995 |
root_mem_cgroup->use_hierarchy = true; else root_mem_cgroup->use_hierarchy = false; |
f00baae7a memcg: force use_... |
4996 |
} |
241994ed8 mm: memcontrol: d... |
4997 4998 4999 5000 5001 5002 5003 5004 5005 |
static u64 memory_current_read(struct cgroup_subsys_state *css, struct cftype *cft) { return mem_cgroup_usage(mem_cgroup_from_css(css), false); } static int memory_low_show(struct seq_file *m, void *v) { struct mem_cgroup *memcg = mem_cgroup_from_css(seq_css(m)); |
4db0c3c29 mm: remove rest o... |
5006 |
unsigned long low = READ_ONCE(memcg->low); |
241994ed8 mm: memcontrol: d... |
5007 5008 |
if (low == PAGE_COUNTER_MAX) |
d2973697b mm: memcontrol: u... |
5009 5010 |
seq_puts(m, "max "); |
241994ed8 mm: memcontrol: d... |
5011 5012 5013 5014 5015 5016 5017 5018 5019 5020 5021 5022 5023 5024 5025 |
else seq_printf(m, "%llu ", (u64)low * PAGE_SIZE); return 0; } static ssize_t memory_low_write(struct kernfs_open_file *of, char *buf, size_t nbytes, loff_t off) { struct mem_cgroup *memcg = mem_cgroup_from_css(of_css(of)); unsigned long low; int err; buf = strstrip(buf); |
d2973697b mm: memcontrol: u... |
5026 |
err = page_counter_memparse(buf, "max", &low); |
241994ed8 mm: memcontrol: d... |
5027 5028 5029 5030 5031 5032 5033 5034 5035 5036 5037 |
if (err) return err; memcg->low = low; return nbytes; } static int memory_high_show(struct seq_file *m, void *v) { struct mem_cgroup *memcg = mem_cgroup_from_css(seq_css(m)); |
4db0c3c29 mm: remove rest o... |
5038 |
unsigned long high = READ_ONCE(memcg->high); |
241994ed8 mm: memcontrol: d... |
5039 5040 |
if (high == PAGE_COUNTER_MAX) |
d2973697b mm: memcontrol: u... |
5041 5042 |
seq_puts(m, "max "); |
241994ed8 mm: memcontrol: d... |
5043 5044 5045 5046 5047 5048 5049 5050 5051 5052 5053 5054 5055 5056 5057 |
else seq_printf(m, "%llu ", (u64)high * PAGE_SIZE); return 0; } static ssize_t memory_high_write(struct kernfs_open_file *of, char *buf, size_t nbytes, loff_t off) { struct mem_cgroup *memcg = mem_cgroup_from_css(of_css(of)); unsigned long high; int err; buf = strstrip(buf); |
d2973697b mm: memcontrol: u... |
5058 |
err = page_counter_memparse(buf, "max", &high); |
241994ed8 mm: memcontrol: d... |
5059 5060 5061 5062 5063 5064 5065 5066 5067 5068 5069 |
if (err) return err; memcg->high = high; return nbytes; } static int memory_max_show(struct seq_file *m, void *v) { struct mem_cgroup *memcg = mem_cgroup_from_css(seq_css(m)); |
4db0c3c29 mm: remove rest o... |
5070 |
unsigned long max = READ_ONCE(memcg->memory.limit); |
241994ed8 mm: memcontrol: d... |
5071 5072 |
if (max == PAGE_COUNTER_MAX) |
d2973697b mm: memcontrol: u... |
5073 5074 |
seq_puts(m, "max "); |
241994ed8 mm: memcontrol: d... |
5075 5076 5077 5078 5079 5080 5081 5082 5083 5084 5085 5086 5087 5088 5089 |
else seq_printf(m, "%llu ", (u64)max * PAGE_SIZE); return 0; } static ssize_t memory_max_write(struct kernfs_open_file *of, char *buf, size_t nbytes, loff_t off) { struct mem_cgroup *memcg = mem_cgroup_from_css(of_css(of)); unsigned long max; int err; buf = strstrip(buf); |
d2973697b mm: memcontrol: u... |
5090 |
err = page_counter_memparse(buf, "max", &max); |
241994ed8 mm: memcontrol: d... |
5091 5092 5093 5094 5095 5096 5097 5098 5099 5100 5101 5102 5103 5104 5105 5106 5107 5108 5109 5110 5111 5112 5113 5114 5115 5116 5117 5118 5119 5120 5121 5122 5123 5124 5125 5126 5127 5128 5129 5130 5131 5132 5133 5134 5135 5136 5137 5138 5139 5140 5141 5142 5143 5144 5145 5146 |
if (err) return err; err = mem_cgroup_resize_limit(memcg, max); if (err) return err; return nbytes; } static int memory_events_show(struct seq_file *m, void *v) { struct mem_cgroup *memcg = mem_cgroup_from_css(seq_css(m)); seq_printf(m, "low %lu ", mem_cgroup_read_events(memcg, MEMCG_LOW)); seq_printf(m, "high %lu ", mem_cgroup_read_events(memcg, MEMCG_HIGH)); seq_printf(m, "max %lu ", mem_cgroup_read_events(memcg, MEMCG_MAX)); seq_printf(m, "oom %lu ", mem_cgroup_read_events(memcg, MEMCG_OOM)); return 0; } static struct cftype memory_files[] = { { .name = "current", .read_u64 = memory_current_read, }, { .name = "low", .flags = CFTYPE_NOT_ON_ROOT, .seq_show = memory_low_show, .write = memory_low_write, }, { .name = "high", .flags = CFTYPE_NOT_ON_ROOT, .seq_show = memory_high_show, .write = memory_high_write, }, { .name = "max", .flags = CFTYPE_NOT_ON_ROOT, .seq_show = memory_max_show, .write = memory_max_write, }, { .name = "events", .flags = CFTYPE_NOT_ON_ROOT, .seq_show = memory_events_show, }, { } /* terminate */ }; |
073219e99 cgroup: clean up ... |
5147 |
struct cgroup_subsys memory_cgrp_subsys = { |
92fb97487 cgroup: rename ->... |
5148 |
.css_alloc = mem_cgroup_css_alloc, |
d142e3e66 memcg: split part... |
5149 |
.css_online = mem_cgroup_css_online, |
92fb97487 cgroup: rename ->... |
5150 5151 |
.css_offline = mem_cgroup_css_offline, .css_free = mem_cgroup_css_free, |
1ced953b1 blkcg, memcg: mak... |
5152 |
.css_reset = mem_cgroup_css_reset, |
7dc74be03 memcg: add interf... |
5153 5154 |
.can_attach = mem_cgroup_can_attach, .cancel_attach = mem_cgroup_cancel_attach, |
67e465a77 Memory controller... |
5155 |
.attach = mem_cgroup_move_task, |
f00baae7a memcg: force use_... |
5156 |
.bind = mem_cgroup_bind, |
241994ed8 mm: memcontrol: d... |
5157 5158 |
.dfl_cftypes = memory_files, .legacy_cftypes = mem_cgroup_legacy_files, |
6d12e2d8d per-zone and recl... |
5159 |
.early_init = 0, |
8cdea7c05 Memory controller... |
5160 |
}; |
c077719be memcg: mem+swap c... |
5161 |
|
241994ed8 mm: memcontrol: d... |
5162 5163 5164 5165 5166 5167 5168 5169 5170 5171 5172 5173 5174 5175 5176 5177 5178 5179 5180 5181 5182 5183 5184 5185 5186 5187 5188 5189 5190 5191 5192 5193 5194 5195 |
/** * 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 */ void mem_cgroup_events(struct mem_cgroup *memcg, enum mem_cgroup_events_index idx, unsigned int nr) { this_cpu_add(memcg->stat->events[idx], nr); } /** * mem_cgroup_low - check if memory consumption is below the normal range * @root: the highest ancestor to consider * @memcg: the memory cgroup to check * * Returns %true if memory consumption of @memcg, and that of all * configurable ancestors up to @root, is below the normal range. */ bool mem_cgroup_low(struct mem_cgroup *root, struct mem_cgroup *memcg) { if (mem_cgroup_disabled()) return false; /* * The toplevel group doesn't have a configurable range, so * it's never low when looked at directly, and it is not * considered an ancestor when assessing the hierarchy. */ if (memcg == root_mem_cgroup) return false; |
4e54dede3 memcg: fix low li... |
5196 |
if (page_counter_read(&memcg->memory) >= memcg->low) |
241994ed8 mm: memcontrol: d... |
5197 5198 5199 5200 5201 5202 5203 |
return false; while (memcg != root) { memcg = parent_mem_cgroup(memcg); if (memcg == root_mem_cgroup) break; |
4e54dede3 memcg: fix low li... |
5204 |
if (page_counter_read(&memcg->memory) >= memcg->low) |
241994ed8 mm: memcontrol: d... |
5205 5206 5207 5208 |
return false; } return true; } |
00501b531 mm: memcontrol: r... |
5209 5210 5211 5212 5213 5214 5215 5216 5217 5218 5219 5220 5221 5222 5223 5224 5225 5226 5227 5228 5229 5230 5231 5232 5233 5234 5235 5236 |
/** * mem_cgroup_try_charge - try charging a page * @page: page to charge * @mm: mm context of the victim * @gfp_mask: reclaim mode * @memcgp: charged memcg return * * Try to charge @page to the memcg that @mm belongs to, reclaiming * pages according to @gfp_mask if necessary. * * Returns 0 on success, with *@memcgp pointing to the charged memcg. * Otherwise, an error code is returned. * * After page->mapping has been set up, the caller must finalize the * charge with mem_cgroup_commit_charge(). Or abort the transaction * with mem_cgroup_cancel_charge() in case page instantiation fails. */ int mem_cgroup_try_charge(struct page *page, struct mm_struct *mm, gfp_t gfp_mask, struct mem_cgroup **memcgp) { struct mem_cgroup *memcg = NULL; unsigned int nr_pages = 1; int ret = 0; if (mem_cgroup_disabled()) goto out; if (PageSwapCache(page)) { |
00501b531 mm: memcontrol: r... |
5237 5238 5239 5240 5241 5242 5243 |
/* * Every swap fault against a single page tries to charge the * page, bail as early as possible. shmem_unuse() encounters * already charged pages, too. The USED bit is protected by * the page lock, which serializes swap cache removal, which * in turn serializes uncharging. */ |
1306a85ae mm: embed the mem... |
5244 |
if (page->mem_cgroup) |
00501b531 mm: memcontrol: r... |
5245 5246 5247 5248 5249 5250 5251 5252 5253 5254 5255 5256 5257 5258 5259 5260 5261 5262 5263 5264 5265 5266 5267 5268 5269 5270 5271 5272 5273 5274 5275 5276 5277 5278 5279 5280 5281 5282 5283 5284 5285 5286 5287 5288 5289 5290 5291 5292 5293 5294 5295 5296 5297 5298 5299 5300 5301 5302 5303 |
goto out; } if (PageTransHuge(page)) { nr_pages <<= compound_order(page); VM_BUG_ON_PAGE(!PageTransHuge(page), page); } if (do_swap_account && PageSwapCache(page)) memcg = try_get_mem_cgroup_from_page(page); if (!memcg) memcg = get_mem_cgroup_from_mm(mm); ret = try_charge(memcg, gfp_mask, nr_pages); css_put(&memcg->css); if (ret == -EINTR) { memcg = root_mem_cgroup; ret = 0; } out: *memcgp = memcg; return ret; } /** * mem_cgroup_commit_charge - commit a page charge * @page: page to charge * @memcg: memcg to charge the page to * @lrucare: page might be on LRU already * * Finalize a charge transaction started by mem_cgroup_try_charge(), * after page->mapping has been set up. This must happen atomically * as part of the page instantiation, i.e. under the page table lock * for anonymous pages, under the page lock for page and swap cache. * * In addition, the page must not be on the LRU during the commit, to * prevent racing with task migration. If it might be, use @lrucare. * * Use mem_cgroup_cancel_charge() to cancel the transaction instead. */ void mem_cgroup_commit_charge(struct page *page, struct mem_cgroup *memcg, bool lrucare) { unsigned int nr_pages = 1; VM_BUG_ON_PAGE(!page->mapping, page); VM_BUG_ON_PAGE(PageLRU(page) && !lrucare, page); if (mem_cgroup_disabled()) return; /* * Swap faults will attempt to charge the same page multiple * times. But reuse_swap_page() might have removed the page * from swapcache already, so we can't check PageSwapCache(). */ if (!memcg) return; |
6abb5a867 mm: memcontrol: a... |
5304 |
commit_charge(page, memcg, lrucare); |
00501b531 mm: memcontrol: r... |
5305 5306 5307 5308 |
if (PageTransHuge(page)) { nr_pages <<= compound_order(page); VM_BUG_ON_PAGE(!PageTransHuge(page), page); } |
6abb5a867 mm: memcontrol: a... |
5309 5310 5311 5312 |
local_irq_disable(); mem_cgroup_charge_statistics(memcg, page, nr_pages); memcg_check_events(memcg, page); local_irq_enable(); |
00501b531 mm: memcontrol: r... |
5313 5314 5315 5316 5317 5318 5319 5320 5321 5322 5323 5324 5325 5326 5327 5328 5329 5330 5331 5332 5333 5334 5335 5336 5337 5338 5339 5340 5341 5342 5343 5344 5345 5346 5347 5348 5349 5350 5351 5352 |
if (do_swap_account && PageSwapCache(page)) { swp_entry_t entry = { .val = page_private(page) }; /* * The swap entry might not get freed for a long time, * let's not wait for it. The page already received a * memory+swap charge, drop the swap entry duplicate. */ mem_cgroup_uncharge_swap(entry); } } /** * mem_cgroup_cancel_charge - cancel a page charge * @page: page to charge * @memcg: memcg to charge the page to * * Cancel a charge transaction started by mem_cgroup_try_charge(). */ void mem_cgroup_cancel_charge(struct page *page, struct mem_cgroup *memcg) { unsigned int nr_pages = 1; if (mem_cgroup_disabled()) return; /* * Swap faults will attempt to charge the same page multiple * times. But reuse_swap_page() might have removed the page * from swapcache already, so we can't check PageSwapCache(). */ if (!memcg) return; if (PageTransHuge(page)) { nr_pages <<= compound_order(page); VM_BUG_ON_PAGE(!PageTransHuge(page), page); } cancel_charge(memcg, nr_pages); } |
747db954c mm: memcontrol: u... |
5353 |
static void uncharge_batch(struct mem_cgroup *memcg, unsigned long pgpgout, |
747db954c mm: memcontrol: u... |
5354 5355 5356 |
unsigned long nr_anon, unsigned long nr_file, unsigned long nr_huge, struct page *dummy_page) { |
18eca2e63 mm: memcontrol: r... |
5357 |
unsigned long nr_pages = nr_anon + nr_file; |
747db954c mm: memcontrol: u... |
5358 |
unsigned long flags; |
ce00a9673 mm: memcontrol: r... |
5359 |
if (!mem_cgroup_is_root(memcg)) { |
18eca2e63 mm: memcontrol: r... |
5360 5361 5362 |
page_counter_uncharge(&memcg->memory, nr_pages); if (do_swap_account) page_counter_uncharge(&memcg->memsw, nr_pages); |
ce00a9673 mm: memcontrol: r... |
5363 5364 |
memcg_oom_recover(memcg); } |
747db954c mm: memcontrol: u... |
5365 5366 5367 5368 5369 5370 |
local_irq_save(flags); __this_cpu_sub(memcg->stat->count[MEM_CGROUP_STAT_RSS], nr_anon); __this_cpu_sub(memcg->stat->count[MEM_CGROUP_STAT_CACHE], nr_file); __this_cpu_sub(memcg->stat->count[MEM_CGROUP_STAT_RSS_HUGE], nr_huge); __this_cpu_add(memcg->stat->events[MEM_CGROUP_EVENTS_PGPGOUT], pgpgout); |
18eca2e63 mm: memcontrol: r... |
5371 |
__this_cpu_add(memcg->stat->nr_page_events, nr_pages); |
747db954c mm: memcontrol: u... |
5372 5373 |
memcg_check_events(memcg, dummy_page); local_irq_restore(flags); |
e8ea14cc6 mm: memcontrol: t... |
5374 5375 |
if (!mem_cgroup_is_root(memcg)) |
18eca2e63 mm: memcontrol: r... |
5376 |
css_put_many(&memcg->css, nr_pages); |
747db954c mm: memcontrol: u... |
5377 5378 5379 5380 5381 |
} static void uncharge_list(struct list_head *page_list) { struct mem_cgroup *memcg = NULL; |
747db954c mm: memcontrol: u... |
5382 5383 5384 5385 |
unsigned long nr_anon = 0; unsigned long nr_file = 0; unsigned long nr_huge = 0; unsigned long pgpgout = 0; |
747db954c mm: memcontrol: u... |
5386 5387 5388 5389 5390 5391 |
struct list_head *next; struct page *page; next = page_list->next; do { unsigned int nr_pages = 1; |
747db954c mm: memcontrol: u... |
5392 5393 5394 5395 5396 5397 |
page = list_entry(next, struct page, lru); next = page->lru.next; VM_BUG_ON_PAGE(PageLRU(page), page); VM_BUG_ON_PAGE(page_count(page), page); |
1306a85ae mm: embed the mem... |
5398 |
if (!page->mem_cgroup) |
747db954c mm: memcontrol: u... |
5399 5400 5401 5402 |
continue; /* * Nobody should be changing or seriously looking at |
1306a85ae mm: embed the mem... |
5403 |
* page->mem_cgroup at this point, we have fully |
298333157 mm: memcontrol: r... |
5404 |
* exclusive access to the page. |
747db954c mm: memcontrol: u... |
5405 |
*/ |
1306a85ae mm: embed the mem... |
5406 |
if (memcg != page->mem_cgroup) { |
747db954c mm: memcontrol: u... |
5407 |
if (memcg) { |
18eca2e63 mm: memcontrol: r... |
5408 5409 5410 |
uncharge_batch(memcg, pgpgout, nr_anon, nr_file, nr_huge, page); pgpgout = nr_anon = nr_file = nr_huge = 0; |
747db954c mm: memcontrol: u... |
5411 |
} |
1306a85ae mm: embed the mem... |
5412 |
memcg = page->mem_cgroup; |
747db954c mm: memcontrol: u... |
5413 5414 5415 5416 5417 5418 5419 5420 5421 5422 5423 5424 |
} if (PageTransHuge(page)) { nr_pages <<= compound_order(page); VM_BUG_ON_PAGE(!PageTransHuge(page), page); nr_huge += nr_pages; } if (PageAnon(page)) nr_anon += nr_pages; else nr_file += nr_pages; |
1306a85ae mm: embed the mem... |
5425 |
page->mem_cgroup = NULL; |
747db954c mm: memcontrol: u... |
5426 5427 5428 5429 5430 |
pgpgout++; } while (next != page_list); if (memcg) |
18eca2e63 mm: memcontrol: r... |
5431 5432 |
uncharge_batch(memcg, pgpgout, nr_anon, nr_file, nr_huge, page); |
747db954c mm: memcontrol: u... |
5433 |
} |
0a31bc97c mm: memcontrol: r... |
5434 5435 5436 5437 5438 5439 5440 5441 5442 |
/** * mem_cgroup_uncharge - uncharge a page * @page: page to uncharge * * Uncharge a page previously charged with mem_cgroup_try_charge() and * mem_cgroup_commit_charge(). */ void mem_cgroup_uncharge(struct page *page) { |
0a31bc97c mm: memcontrol: r... |
5443 5444 |
if (mem_cgroup_disabled()) return; |
747db954c mm: memcontrol: u... |
5445 |
/* Don't touch page->lru of any random page, pre-check: */ |
1306a85ae mm: embed the mem... |
5446 |
if (!page->mem_cgroup) |
0a31bc97c mm: memcontrol: r... |
5447 |
return; |
747db954c mm: memcontrol: u... |
5448 5449 5450 |
INIT_LIST_HEAD(&page->lru); uncharge_list(&page->lru); } |
0a31bc97c mm: memcontrol: r... |
5451 |
|
747db954c mm: memcontrol: u... |
5452 5453 5454 5455 5456 5457 5458 5459 5460 5461 5462 |
/** * mem_cgroup_uncharge_list - uncharge a list of page * @page_list: list of pages to uncharge * * Uncharge a list of pages previously charged with * mem_cgroup_try_charge() and mem_cgroup_commit_charge(). */ void mem_cgroup_uncharge_list(struct list_head *page_list) { if (mem_cgroup_disabled()) return; |
0a31bc97c mm: memcontrol: r... |
5463 |
|
747db954c mm: memcontrol: u... |
5464 5465 |
if (!list_empty(page_list)) uncharge_list(page_list); |
0a31bc97c mm: memcontrol: r... |
5466 5467 5468 5469 5470 5471 |
} /** * mem_cgroup_migrate - migrate a charge to another page * @oldpage: currently charged page * @newpage: page to transfer the charge to |
f5e03a498 memcg, shmem: fix... |
5472 |
* @lrucare: either or both pages might be on the LRU already |
0a31bc97c mm: memcontrol: r... |
5473 5474 5475 5476 5477 5478 5479 5480 |
* * Migrate the charge from @oldpage to @newpage. * * Both pages must be locked, @newpage->mapping must be set up. */ void mem_cgroup_migrate(struct page *oldpage, struct page *newpage, bool lrucare) { |
298333157 mm: memcontrol: r... |
5481 |
struct mem_cgroup *memcg; |
0a31bc97c mm: memcontrol: r... |
5482 5483 5484 5485 5486 5487 5488 |
int isolated; VM_BUG_ON_PAGE(!PageLocked(oldpage), oldpage); VM_BUG_ON_PAGE(!PageLocked(newpage), newpage); VM_BUG_ON_PAGE(!lrucare && PageLRU(oldpage), oldpage); VM_BUG_ON_PAGE(!lrucare && PageLRU(newpage), newpage); VM_BUG_ON_PAGE(PageAnon(oldpage) != PageAnon(newpage), newpage); |
6abb5a867 mm: memcontrol: a... |
5489 5490 |
VM_BUG_ON_PAGE(PageTransHuge(oldpage) != PageTransHuge(newpage), newpage); |
0a31bc97c mm: memcontrol: r... |
5491 5492 5493 5494 5495 |
if (mem_cgroup_disabled()) return; /* Page cache replacement: new page already charged? */ |
1306a85ae mm: embed the mem... |
5496 |
if (newpage->mem_cgroup) |
0a31bc97c mm: memcontrol: r... |
5497 |
return; |
7d5e32457 mm: memcontrol: c... |
5498 5499 5500 5501 5502 5503 |
/* * Swapcache readahead pages can get migrated before being * charged, and migration from compaction can happen to an * uncharged page when the PFN walker finds a page that * reclaim just put back on the LRU but has not released yet. */ |
1306a85ae mm: embed the mem... |
5504 |
memcg = oldpage->mem_cgroup; |
298333157 mm: memcontrol: r... |
5505 |
if (!memcg) |
0a31bc97c mm: memcontrol: r... |
5506 |
return; |
0a31bc97c mm: memcontrol: r... |
5507 5508 |
if (lrucare) lock_page_lru(oldpage, &isolated); |
1306a85ae mm: embed the mem... |
5509 |
oldpage->mem_cgroup = NULL; |
0a31bc97c mm: memcontrol: r... |
5510 5511 5512 |
if (lrucare) unlock_page_lru(oldpage, isolated); |
298333157 mm: memcontrol: r... |
5513 |
commit_charge(newpage, memcg, lrucare); |
0a31bc97c mm: memcontrol: r... |
5514 |
} |
2d11085e4 memcg: do not cre... |
5515 |
/* |
1081312f9 memcg: cleanup me... |
5516 5517 5518 5519 5520 5521 |
* subsys_initcall() for memory controller. * * Some parts like hotcpu_notifier() have to be initialized from this context * because of lock dependencies (cgroup_lock -> cpu hotplug) but basically * everything that doesn't depend on a specific mem_cgroup structure should * be initialized from here. |
2d11085e4 memcg: do not cre... |
5522 5523 5524 |
*/ static int __init mem_cgroup_init(void) { |
95a045f63 mm: memcontrol: c... |
5525 |
int cpu, node; |
2d11085e4 memcg: do not cre... |
5526 |
hotcpu_notifier(memcg_cpu_hotplug_callback, 0); |
95a045f63 mm: memcontrol: c... |
5527 5528 5529 5530 5531 5532 5533 5534 5535 5536 5537 5538 5539 5540 5541 5542 5543 5544 5545 5546 5547 |
for_each_possible_cpu(cpu) INIT_WORK(&per_cpu_ptr(&memcg_stock, cpu)->work, drain_local_stock); for_each_node(node) { struct mem_cgroup_tree_per_node *rtpn; int zone; rtpn = kzalloc_node(sizeof(*rtpn), GFP_KERNEL, node_online(node) ? node : NUMA_NO_NODE); for (zone = 0; zone < MAX_NR_ZONES; zone++) { struct mem_cgroup_tree_per_zone *rtpz; rtpz = &rtpn->rb_tree_per_zone[zone]; rtpz->rb_root = RB_ROOT; spin_lock_init(&rtpz->lock); } soft_limit_tree.rb_tree_per_node[node] = rtpn; } |
2d11085e4 memcg: do not cre... |
5548 5549 5550 |
return 0; } subsys_initcall(mem_cgroup_init); |
21afa38ee mm: memcontrol: c... |
5551 5552 5553 5554 5555 5556 5557 5558 5559 5560 5561 5562 5563 5564 5565 5566 5567 5568 5569 5570 5571 5572 5573 5574 5575 5576 5577 5578 5579 5580 5581 5582 5583 5584 |
#ifdef CONFIG_MEMCG_SWAP /** * mem_cgroup_swapout - transfer a memsw charge to swap * @page: page whose memsw charge to transfer * @entry: swap entry to move the charge to * * Transfer the memsw charge of @page to @entry. */ void mem_cgroup_swapout(struct page *page, swp_entry_t entry) { struct mem_cgroup *memcg; unsigned short oldid; VM_BUG_ON_PAGE(PageLRU(page), page); VM_BUG_ON_PAGE(page_count(page), page); if (!do_swap_account) return; memcg = page->mem_cgroup; /* Readahead page, never charged */ if (!memcg) return; oldid = swap_cgroup_record(entry, mem_cgroup_id(memcg)); VM_BUG_ON_PAGE(oldid, page); mem_cgroup_swap_statistics(memcg, true); page->mem_cgroup = NULL; if (!mem_cgroup_is_root(memcg)) page_counter_uncharge(&memcg->memory, 1); |
f371763a7 mm: memcontrol: f... |
5585 |
/* Caller disabled preemption with mapping->tree_lock */ |
21afa38ee mm: memcontrol: c... |
5586 5587 5588 5589 5590 5591 5592 5593 5594 5595 5596 5597 5598 5599 5600 5601 5602 5603 5604 5605 |
mem_cgroup_charge_statistics(memcg, page, -1); memcg_check_events(memcg, page); } /** * mem_cgroup_uncharge_swap - uncharge a swap entry * @entry: swap entry to uncharge * * Drop the memsw charge associated with @entry. */ void mem_cgroup_uncharge_swap(swp_entry_t entry) { struct mem_cgroup *memcg; unsigned short id; if (!do_swap_account) return; id = swap_cgroup_record(entry, 0); rcu_read_lock(); |
adbe427b9 memcg: zap mem_cg... |
5606 |
memcg = mem_cgroup_from_id(id); |
21afa38ee mm: memcontrol: c... |
5607 5608 5609 5610 5611 5612 5613 5614 5615 5616 5617 5618 5619 5620 5621 5622 5623 5624 5625 5626 5627 5628 5629 5630 5631 5632 5633 5634 5635 5636 5637 5638 5639 5640 5641 5642 5643 5644 5645 5646 5647 5648 5649 5650 5651 5652 5653 5654 5655 5656 5657 5658 5659 5660 5661 5662 5663 5664 5665 5666 5667 5668 5669 5670 5671 |
if (memcg) { if (!mem_cgroup_is_root(memcg)) page_counter_uncharge(&memcg->memsw, 1); mem_cgroup_swap_statistics(memcg, false); css_put(&memcg->css); } rcu_read_unlock(); } /* for remember boot option*/ #ifdef CONFIG_MEMCG_SWAP_ENABLED static int really_do_swap_account __initdata = 1; #else static int really_do_swap_account __initdata; #endif static int __init enable_swap_account(char *s) { if (!strcmp(s, "1")) really_do_swap_account = 1; else if (!strcmp(s, "0")) really_do_swap_account = 0; return 1; } __setup("swapaccount=", enable_swap_account); static struct cftype memsw_cgroup_files[] = { { .name = "memsw.usage_in_bytes", .private = MEMFILE_PRIVATE(_MEMSWAP, RES_USAGE), .read_u64 = mem_cgroup_read_u64, }, { .name = "memsw.max_usage_in_bytes", .private = MEMFILE_PRIVATE(_MEMSWAP, RES_MAX_USAGE), .write = mem_cgroup_reset, .read_u64 = mem_cgroup_read_u64, }, { .name = "memsw.limit_in_bytes", .private = MEMFILE_PRIVATE(_MEMSWAP, RES_LIMIT), .write = mem_cgroup_write, .read_u64 = mem_cgroup_read_u64, }, { .name = "memsw.failcnt", .private = MEMFILE_PRIVATE(_MEMSWAP, RES_FAILCNT), .write = mem_cgroup_reset, .read_u64 = mem_cgroup_read_u64, }, { }, /* terminate */ }; static int __init mem_cgroup_swap_init(void) { if (!mem_cgroup_disabled() && really_do_swap_account) { do_swap_account = 1; WARN_ON(cgroup_add_legacy_cftypes(&memory_cgrp_subsys, memsw_cgroup_files)); } return 0; } subsys_initcall(mem_cgroup_swap_init); #endif /* CONFIG_MEMCG_SWAP */ |