Blame view
mm/memcontrol.c
142 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 * |
8cdea7c05 Memory controller... |
13 14 15 16 17 18 19 20 21 22 23 24 25 26 |
* This program is free software; you can redistribute it and/or modify * it under the terms of the GNU General Public License as published by * the Free Software Foundation; either version 2 of the License, or * (at your option) any later version. * * This program is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. */ #include <linux/res_counter.h> #include <linux/memcontrol.h> #include <linux/cgroup.h> |
78fb74669 Memory controller... |
27 |
#include <linux/mm.h> |
4ffef5fef memcg: move charg... |
28 |
#include <linux/hugetlb.h> |
d13d14430 memcg: handle swa... |
29 |
#include <linux/pagemap.h> |
d52aa412d memory cgroup enh... |
30 |
#include <linux/smp.h> |
8a9f3ccd2 Memory controller... |
31 |
#include <linux/page-flags.h> |
66e1707bc Memory controller... |
32 |
#include <linux/backing-dev.h> |
8a9f3ccd2 Memory controller... |
33 34 |
#include <linux/bit_spinlock.h> #include <linux/rcupdate.h> |
e222432bf memcg: show memcg... |
35 |
#include <linux/limits.h> |
b9e15bafd mm: Add export.h ... |
36 |
#include <linux/export.h> |
8c7c6e34a memcg: mem+swap c... |
37 |
#include <linux/mutex.h> |
f64c3f549 memory controller... |
38 |
#include <linux/rbtree.h> |
b6ac57d50 memcgroup: move m... |
39 |
#include <linux/slab.h> |
66e1707bc Memory controller... |
40 |
#include <linux/swap.h> |
024914477 memcg: move charg... |
41 |
#include <linux/swapops.h> |
66e1707bc Memory controller... |
42 |
#include <linux/spinlock.h> |
2e72b6347 memcg: implement ... |
43 44 |
#include <linux/eventfd.h> #include <linux/sort.h> |
66e1707bc Memory controller... |
45 |
#include <linux/fs.h> |
d2ceb9b7d memory cgroup enh... |
46 |
#include <linux/seq_file.h> |
333279487 memcgroup: use vm... |
47 |
#include <linux/vmalloc.h> |
b69408e88 vmscan: Use an in... |
48 |
#include <linux/mm_inline.h> |
52d4b9ac0 memcg: allocate a... |
49 |
#include <linux/page_cgroup.h> |
cdec2e426 memcg: coalesce c... |
50 |
#include <linux/cpu.h> |
158e0a2d1 memcg: use find_l... |
51 |
#include <linux/oom.h> |
08e552c69 memcg: synchroniz... |
52 |
#include "internal.h" |
d1a4c0b37 tcp memory pressu... |
53 54 |
#include <net/sock.h> #include <net/tcp_memcontrol.h> |
8cdea7c05 Memory controller... |
55 |
|
8697d3319 Memory controller... |
56 |
#include <asm/uaccess.h> |
cc8e970c3 memcg: add mm_vms... |
57 |
#include <trace/events/vmscan.h> |
a181b0e88 memcg: make globa... |
58 |
struct cgroup_subsys mem_cgroup_subsys __read_mostly; |
a181b0e88 memcg: make globa... |
59 |
#define MEM_CGROUP_RECLAIM_RETRIES 5 |
4b3bde4c9 memcg: remove the... |
60 |
struct mem_cgroup *root_mem_cgroup __read_mostly; |
8cdea7c05 Memory controller... |
61 |
|
c077719be memcg: mem+swap c... |
62 |
#ifdef CONFIG_CGROUP_MEM_RES_CTLR_SWAP |
338c84310 memcg: remove som... |
63 |
/* Turned on only when memory cgroup is enabled && really_do_swap_account = 1 */ |
c077719be memcg: mem+swap c... |
64 |
int do_swap_account __read_mostly; |
a42c390cf cgroups: make swa... |
65 66 67 68 69 70 71 |
/* for remember boot option*/ #ifdef CONFIG_CGROUP_MEM_RES_CTLR_SWAP_ENABLED static int really_do_swap_account __initdata = 1; #else static int really_do_swap_account __initdata = 0; #endif |
c077719be memcg: mem+swap c... |
72 73 74 |
#else #define do_swap_account (0) #endif |
8cdea7c05 Memory controller... |
75 |
/* |
d52aa412d memory cgroup enh... |
76 77 78 79 80 81 82 |
* Statistics for memory cgroup. */ enum mem_cgroup_stat_index { /* * For MEM_CONTAINER_TYPE_ALL, usage = pagecache + rss. */ MEM_CGROUP_STAT_CACHE, /* # of pages charged as cache */ |
d69b042f3 memcg: add file-b... |
83 |
MEM_CGROUP_STAT_RSS, /* # of pages charged as anon rss */ |
d8046582d memcg: make memcg... |
84 |
MEM_CGROUP_STAT_FILE_MAPPED, /* # of pages charged as file rss */ |
0c3e73e84 memcg: improve re... |
85 |
MEM_CGROUP_STAT_SWAPOUT, /* # of pages, swapped out */ |
711d3d2c9 memcg: cpu hotplu... |
86 |
MEM_CGROUP_STAT_DATA, /* end of data requires synchronization */ |
32047e2a8 memcg: avoid lock... |
87 |
MEM_CGROUP_ON_MOVE, /* someone is moving account between groups */ |
d52aa412d memory cgroup enh... |
88 89 |
MEM_CGROUP_STAT_NSTATS, }; |
e9f8974f2 memcg: break out ... |
90 91 92 93 |
enum mem_cgroup_events_index { MEM_CGROUP_EVENTS_PGPGIN, /* # of pages paged in */ MEM_CGROUP_EVENTS_PGPGOUT, /* # of pages paged out */ MEM_CGROUP_EVENTS_COUNT, /* # of pages paged in/out */ |
456f998ec memcg: add the pa... |
94 95 |
MEM_CGROUP_EVENTS_PGFAULT, /* # of page-faults */ MEM_CGROUP_EVENTS_PGMAJFAULT, /* # of major page-faults */ |
e9f8974f2 memcg: break out ... |
96 97 |
MEM_CGROUP_EVENTS_NSTATS, }; |
7a159cc9d memcg: use native... |
98 99 100 101 102 103 104 105 106 |
/* * Per memcg event counter is incremented at every pagein/pageout. With THP, * it will be incremated by the number of pages. This counter is used for * for trigger some periodic events. This is straightforward and better * than using jiffies etc. to handle periodic memcg event. */ enum mem_cgroup_events_target { MEM_CGROUP_TARGET_THRESH, MEM_CGROUP_TARGET_SOFTLIMIT, |
453a9bf34 memcg: fix numa s... |
107 |
MEM_CGROUP_TARGET_NUMAINFO, |
7a159cc9d memcg: use native... |
108 109 110 111 |
MEM_CGROUP_NTARGETS, }; #define THRESHOLDS_EVENTS_TARGET (128) #define SOFTLIMIT_EVENTS_TARGET (1024) |
453a9bf34 memcg: fix numa s... |
112 |
#define NUMAINFO_EVENTS_TARGET (1024) |
e9f8974f2 memcg: break out ... |
113 |
|
d52aa412d memory cgroup enh... |
114 |
struct mem_cgroup_stat_cpu { |
7a159cc9d memcg: use native... |
115 |
long count[MEM_CGROUP_STAT_NSTATS]; |
e9f8974f2 memcg: break out ... |
116 |
unsigned long events[MEM_CGROUP_EVENTS_NSTATS]; |
7a159cc9d memcg: use native... |
117 |
unsigned long targets[MEM_CGROUP_NTARGETS]; |
d52aa412d memory cgroup enh... |
118 |
}; |
527a5ec9a mm: memcg: per-pr... |
119 120 121 122 123 124 |
struct mem_cgroup_reclaim_iter { /* css_id of the last scanned hierarchy member */ int position; /* scan generation, increased every round-trip */ unsigned int generation; }; |
d52aa412d memory cgroup enh... |
125 |
/* |
6d12e2d8d per-zone and recl... |
126 127 |
* per-zone information in memory controller. */ |
6d12e2d8d per-zone and recl... |
128 |
struct mem_cgroup_per_zone { |
6290df545 mm: collect LRU l... |
129 |
struct lruvec lruvec; |
b69408e88 vmscan: Use an in... |
130 |
unsigned long count[NR_LRU_LISTS]; |
3e2f41f1f memcg: add zone_r... |
131 |
|
527a5ec9a mm: memcg: per-pr... |
132 |
struct mem_cgroup_reclaim_iter reclaim_iter[DEF_PRIORITY + 1]; |
3e2f41f1f memcg: add zone_r... |
133 |
struct zone_reclaim_stat reclaim_stat; |
f64c3f549 memory controller... |
134 135 136 137 |
struct rb_node tree_node; /* RB tree node */ unsigned long long usage_in_excess;/* Set to the value by which */ /* the soft limit is exceeded*/ bool on_tree; |
4e4169535 memory controller... |
138 139 |
struct mem_cgroup *mem; /* Back pointer, we cannot */ /* use container_of */ |
6d12e2d8d per-zone and recl... |
140 141 142 143 144 145 146 147 148 149 150 151 152 |
}; /* Macro for accessing counter */ #define MEM_CGROUP_ZSTAT(mz, idx) ((mz)->count[(idx)]) struct mem_cgroup_per_node { struct mem_cgroup_per_zone zoneinfo[MAX_NR_ZONES]; }; struct mem_cgroup_lru_info { struct mem_cgroup_per_node *nodeinfo[MAX_NUMNODES]; }; /* |
f64c3f549 memory controller... |
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 ... |
171 172 173 174 |
struct mem_cgroup_threshold { struct eventfd_ctx *eventfd; u64 threshold; }; |
9490ff275 memcg: oom notifier |
175 |
/* For threshold */ |
2e72b6347 memcg: implement ... |
176 177 |
struct mem_cgroup_threshold_ary { /* An array index points to threshold just below 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 |
|
c0ff4b854 memcg: rename mem... |
201 202 |
static void mem_cgroup_threshold(struct mem_cgroup *memcg); static void mem_cgroup_oom_notify(struct mem_cgroup *memcg); |
2e72b6347 memcg: implement ... |
203 |
|
f64c3f549 memory controller... |
204 |
/* |
8cdea7c05 Memory controller... |
205 206 207 208 209 210 |
* The memory controller data structure. The memory controller controls both * page cache and RSS per cgroup. We would eventually like to provide * statistics based on the statistics developed by Rik Van Riel for clock-pro, * to help the administrator determine what knobs to tune. * * TODO: Add a water mark for the memory controller. Reclaim will begin when |
8a9f3ccd2 Memory controller... |
211 212 213 |
* we hit the water mark. May be even add a low water mark, such that * no reclaim occurs from a cgroup at it's low water mark, this is * a feature that will be implemented much later in the future. |
8cdea7c05 Memory controller... |
214 215 216 217 218 219 220 |
*/ struct mem_cgroup { struct cgroup_subsys_state css; /* * the counter to account for memory usage */ struct res_counter res; |
78fb74669 Memory controller... |
221 |
/* |
8c7c6e34a memcg: mem+swap c... |
222 223 224 225 |
* the counter to account for mem+swap usage. */ struct res_counter memsw; /* |
78fb74669 Memory controller... |
226 227 |
* Per cgroup active and inactive list, similar to the * per zone LRU lists. |
78fb74669 Memory controller... |
228 |
*/ |
6d12e2d8d per-zone and recl... |
229 |
struct mem_cgroup_lru_info info; |
889976dbc memcg: reclaim me... |
230 231 232 |
int last_scanned_node; #if MAX_NUMNODES > 1 nodemask_t scan_nodes; |
453a9bf34 memcg: fix numa s... |
233 234 |
atomic_t numainfo_events; atomic_t numainfo_updating; |
889976dbc memcg: reclaim me... |
235 |
#endif |
18f59ea7d memcg: memory cgr... |
236 237 238 239 |
/* * Should the accounting and control be hierarchical, per subtree? */ bool use_hierarchy; |
79dfdaccd memcg: make oom_l... |
240 241 242 |
bool oom_lock; atomic_t under_oom; |
8c7c6e34a memcg: mem+swap c... |
243 |
atomic_t refcnt; |
14797e236 memcg: add inacti... |
244 |
|
1f4c025b5 memcg: export mem... |
245 |
int swappiness; |
3c11ecf44 memcg: oom kill d... |
246 247 |
/* OOM-Killer disable */ int oom_kill_disable; |
a7885eb8a memcg: swappiness |
248 |
|
22a668d7c memcg: fix behavi... |
249 250 |
/* set when res.limit == memsw.limit */ bool memsw_is_minimum; |
2e72b6347 memcg: implement ... |
251 252 253 254 |
/* protect arrays of thresholds */ struct mutex thresholds_lock; /* thresholds for memory usage. RCU-protected */ |
2c488db27 memcg: clean up m... |
255 |
struct mem_cgroup_thresholds thresholds; |
907860ed3 cgroups: make cft... |
256 |
|
2e72b6347 memcg: implement ... |
257 |
/* thresholds for mem+swap usage. RCU-protected */ |
2c488db27 memcg: clean up m... |
258 |
struct mem_cgroup_thresholds memsw_thresholds; |
907860ed3 cgroups: make cft... |
259 |
|
9490ff275 memcg: oom notifier |
260 261 |
/* For oom notifier event fd */ struct list_head oom_notify; |
185efc0f9 memcg: Revert "me... |
262 |
|
d52aa412d memory cgroup enh... |
263 |
/* |
7dc74be03 memcg: add interf... |
264 265 266 267 |
* Should we move charges of a task when a task is moved into this * mem_cgroup ? And what type of charges should we move ? */ unsigned long move_charge_at_immigrate; |
7dc74be03 memcg: add interf... |
268 |
/* |
c62b1a3b3 memcg: use generi... |
269 |
* percpu counter. |
d52aa412d memory cgroup enh... |
270 |
*/ |
c62b1a3b3 memcg: use generi... |
271 |
struct mem_cgroup_stat_cpu *stat; |
711d3d2c9 memcg: cpu hotplu... |
272 273 274 275 276 277 |
/* * 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... |
278 279 280 281 |
#ifdef CONFIG_INET struct tcp_memcontrol tcp_mem; #endif |
8cdea7c05 Memory controller... |
282 |
}; |
7dc74be03 memcg: add interf... |
283 284 285 286 287 288 |
/* Stuffs for move charges at task migration. */ /* * Types of charges to be moved. "move_charge_at_immitgrate" is treated as a * left-shifted bitmap of these types. */ enum move_type { |
4ffef5fef memcg: move charg... |
289 |
MOVE_CHARGE_TYPE_ANON, /* private anonymous page and swap of it */ |
87946a722 memcg: move charg... |
290 |
MOVE_CHARGE_TYPE_FILE, /* file page(including tmpfs) and swap of it */ |
7dc74be03 memcg: add interf... |
291 292 |
NR_MOVE_TYPE, }; |
4ffef5fef memcg: move charg... |
293 294 |
/* "mc" and its members are protected by cgroup_mutex */ static struct move_charge_struct { |
b1dd693e5 memcg: avoid dead... |
295 |
spinlock_t lock; /* for from, to */ |
4ffef5fef memcg: move charg... |
296 297 298 |
struct mem_cgroup *from; struct mem_cgroup *to; unsigned long precharge; |
854ffa8d1 memcg: improve pe... |
299 |
unsigned long moved_charge; |
483c30b51 memcg: improve pe... |
300 |
unsigned long moved_swap; |
8033b97c9 memcg: avoid oom ... |
301 302 303 |
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... |
304 |
.lock = __SPIN_LOCK_UNLOCKED(mc.lock), |
8033b97c9 memcg: avoid oom ... |
305 306 |
.waitq = __WAIT_QUEUE_HEAD_INITIALIZER(mc.waitq), }; |
4ffef5fef memcg: move charg... |
307 |
|
90254a658 memcg: clean up m... |
308 309 310 311 312 |
static bool move_anon(void) { return test_bit(MOVE_CHARGE_TYPE_ANON, &mc.to->move_charge_at_immigrate); } |
87946a722 memcg: move charg... |
313 314 315 316 317 |
static bool move_file(void) { return test_bit(MOVE_CHARGE_TYPE_FILE, &mc.to->move_charge_at_immigrate); } |
4e4169535 memory controller... |
318 319 320 321 322 323 |
/* * Maximum loops in mem_cgroup_hierarchical_reclaim(), used for soft * limit reclaim to prevent infinite loops, if they ever occur. */ #define MEM_CGROUP_MAX_RECLAIM_LOOPS (100) #define MEM_CGROUP_MAX_SOFT_LIMIT_RECLAIM_LOOPS (2) |
217bc3194 memory cgroup enh... |
324 325 326 |
enum charge_type { MEM_CGROUP_CHARGE_TYPE_CACHE = 0, MEM_CGROUP_CHARGE_TYPE_MAPPED, |
4f98a2fee vmscan: split LRU... |
327 |
MEM_CGROUP_CHARGE_TYPE_SHMEM, /* used by page migration of shmem */ |
c05555b57 memcg: atomic ops... |
328 |
MEM_CGROUP_CHARGE_TYPE_FORCE, /* used by force_empty */ |
d13d14430 memcg: handle swa... |
329 |
MEM_CGROUP_CHARGE_TYPE_SWAPOUT, /* for accounting swapcache */ |
8a9478ca7 memcg: fix swap a... |
330 |
MEM_CGROUP_CHARGE_TYPE_DROP, /* a page was unused swap cache */ |
c05555b57 memcg: atomic ops... |
331 332 |
NR_CHARGE_TYPE, }; |
8c7c6e34a memcg: mem+swap c... |
333 |
/* for encoding cft->private value on file */ |
65c64ce8e Partial revert "B... |
334 335 336 |
#define _MEM (0) #define _MEMSWAP (1) #define _OOM_TYPE (2) |
8c7c6e34a memcg: mem+swap c... |
337 338 339 |
#define MEMFILE_PRIVATE(x, val) (((x) << 16) | (val)) #define MEMFILE_TYPE(val) (((val) >> 16) & 0xffff) #define MEMFILE_ATTR(val) ((val) & 0xffff) |
9490ff275 memcg: oom notifier |
340 341 |
/* Used for OOM nofiier */ #define OOM_CONTROL (0) |
8c7c6e34a memcg: mem+swap c... |
342 |
|
75822b449 memory controller... |
343 344 345 346 347 348 349 |
/* * Reclaim flags for mem_cgroup_hierarchical_reclaim */ #define MEM_CGROUP_RECLAIM_NOSWAP_BIT 0x0 #define MEM_CGROUP_RECLAIM_NOSWAP (1 << MEM_CGROUP_RECLAIM_NOSWAP_BIT) #define MEM_CGROUP_RECLAIM_SHRINK_BIT 0x1 #define MEM_CGROUP_RECLAIM_SHRINK (1 << MEM_CGROUP_RECLAIM_SHRINK_BIT) |
c0ff4b854 memcg: rename mem... |
350 351 |
static void mem_cgroup_get(struct mem_cgroup *memcg); static void mem_cgroup_put(struct mem_cgroup *memcg); |
e1aab161e socket: initial c... |
352 353 354 355 356 |
/* Writing them here to avoid exposing memcg's inner layout */ #ifdef CONFIG_CGROUP_MEM_RES_CTLR_KMEM #ifdef CONFIG_INET #include <net/sock.h> |
d1a4c0b37 tcp memory pressu... |
357 |
#include <net/ip.h> |
e1aab161e socket: initial c... |
358 359 360 361 |
static bool mem_cgroup_is_root(struct mem_cgroup *memcg); void sock_update_memcg(struct sock *sk) { |
e1aab161e socket: initial c... |
362 363 364 365 |
if (static_branch(&memcg_socket_limit_enabled)) { struct mem_cgroup *memcg; BUG_ON(!sk->sk_prot->proto_cgroup); |
f3f511e1c net: fix sock_clo... |
366 367 368 369 370 371 372 373 374 375 376 377 378 |
/* 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)); mem_cgroup_get(sk->sk_cgrp->memcg); return; } |
e1aab161e socket: initial c... |
379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 |
rcu_read_lock(); memcg = mem_cgroup_from_task(current); if (!mem_cgroup_is_root(memcg)) { mem_cgroup_get(memcg); sk->sk_cgrp = sk->sk_prot->proto_cgroup(memcg); } rcu_read_unlock(); } } EXPORT_SYMBOL(sock_update_memcg); void sock_release_memcg(struct sock *sk) { if (static_branch(&memcg_socket_limit_enabled) && sk->sk_cgrp) { struct mem_cgroup *memcg; WARN_ON(!sk->sk_cgrp->memcg); memcg = sk->sk_cgrp->memcg; mem_cgroup_put(memcg); } } |
d1a4c0b37 tcp memory pressu... |
399 400 401 402 403 404 405 406 407 |
struct cg_proto *tcp_proto_cgroup(struct mem_cgroup *memcg) { if (!memcg || mem_cgroup_is_root(memcg)) return NULL; return &memcg->tcp_mem.cg_proto; } EXPORT_SYMBOL(tcp_proto_cgroup); |
e1aab161e socket: initial c... |
408 409 |
#endif /* CONFIG_INET */ #endif /* CONFIG_CGROUP_MEM_RES_CTLR_KMEM */ |
c0ff4b854 memcg: rename mem... |
410 |
static void drain_all_stock_async(struct mem_cgroup *memcg); |
8c7c6e34a memcg: mem+swap c... |
411 |
|
f64c3f549 memory controller... |
412 |
static struct mem_cgroup_per_zone * |
c0ff4b854 memcg: rename mem... |
413 |
mem_cgroup_zoneinfo(struct mem_cgroup *memcg, int nid, int zid) |
f64c3f549 memory controller... |
414 |
{ |
c0ff4b854 memcg: rename mem... |
415 |
return &memcg->info.nodeinfo[nid]->zoneinfo[zid]; |
f64c3f549 memory controller... |
416 |
} |
c0ff4b854 memcg: rename mem... |
417 |
struct cgroup_subsys_state *mem_cgroup_css(struct mem_cgroup *memcg) |
d324236b3 memcg: add access... |
418 |
{ |
c0ff4b854 memcg: rename mem... |
419 |
return &memcg->css; |
d324236b3 memcg: add access... |
420 |
} |
f64c3f549 memory controller... |
421 |
static struct mem_cgroup_per_zone * |
c0ff4b854 memcg: rename mem... |
422 |
page_cgroup_zoneinfo(struct mem_cgroup *memcg, struct page *page) |
f64c3f549 memory controller... |
423 |
{ |
97a6c37b3 memcg: change pag... |
424 425 |
int nid = page_to_nid(page); int zid = page_zonenum(page); |
f64c3f549 memory controller... |
426 |
|
c0ff4b854 memcg: rename mem... |
427 |
return mem_cgroup_zoneinfo(memcg, nid, zid); |
f64c3f549 memory controller... |
428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 |
} static struct mem_cgroup_tree_per_zone * soft_limit_tree_node_zone(int nid, int zid) { return &soft_limit_tree.rb_tree_per_node[nid]->rb_tree_per_zone[zid]; } static struct mem_cgroup_tree_per_zone * soft_limit_tree_from_page(struct page *page) { int nid = page_to_nid(page); int zid = page_zonenum(page); return &soft_limit_tree.rb_tree_per_node[nid]->rb_tree_per_zone[zid]; } static void |
c0ff4b854 memcg: rename mem... |
446 |
__mem_cgroup_insert_exceeded(struct mem_cgroup *memcg, |
f64c3f549 memory controller... |
447 |
struct mem_cgroup_per_zone *mz, |
ef8745c1e memcg: reduce che... |
448 449 |
struct mem_cgroup_tree_per_zone *mctz, unsigned long long new_usage_in_excess) |
f64c3f549 memory controller... |
450 451 452 453 454 455 456 |
{ 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; |
ef8745c1e memcg: reduce che... |
457 458 459 |
mz->usage_in_excess = new_usage_in_excess; if (!mz->usage_in_excess) return; |
f64c3f549 memory controller... |
460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 |
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; |
4e4169535 memory controller... |
476 477 478 |
} static void |
c0ff4b854 memcg: rename mem... |
479 |
__mem_cgroup_remove_exceeded(struct mem_cgroup *memcg, |
4e4169535 memory controller... |
480 481 482 483 484 485 486 487 488 489 |
struct mem_cgroup_per_zone *mz, struct mem_cgroup_tree_per_zone *mctz) { if (!mz->on_tree) return; rb_erase(&mz->tree_node, &mctz->rb_root); mz->on_tree = false; } static void |
c0ff4b854 memcg: rename mem... |
490 |
mem_cgroup_remove_exceeded(struct mem_cgroup *memcg, |
f64c3f549 memory controller... |
491 492 493 494 |
struct mem_cgroup_per_zone *mz, struct mem_cgroup_tree_per_zone *mctz) { spin_lock(&mctz->lock); |
c0ff4b854 memcg: rename mem... |
495 |
__mem_cgroup_remove_exceeded(memcg, mz, mctz); |
f64c3f549 memory controller... |
496 497 |
spin_unlock(&mctz->lock); } |
f64c3f549 memory controller... |
498 |
|
c0ff4b854 memcg: rename mem... |
499 |
static void mem_cgroup_update_tree(struct mem_cgroup *memcg, struct page *page) |
f64c3f549 memory controller... |
500 |
{ |
ef8745c1e memcg: reduce che... |
501 |
unsigned long long excess; |
f64c3f549 memory controller... |
502 503 |
struct mem_cgroup_per_zone *mz; struct mem_cgroup_tree_per_zone *mctz; |
4e649152c memcg: some modif... |
504 505 |
int nid = page_to_nid(page); int zid = page_zonenum(page); |
f64c3f549 memory controller... |
506 507 508 |
mctz = soft_limit_tree_from_page(page); /* |
4e649152c memcg: some modif... |
509 510 |
* Necessary to update all ancestors when hierarchy is used. * because their event counter is not touched. |
f64c3f549 memory controller... |
511 |
*/ |
c0ff4b854 memcg: rename mem... |
512 513 514 |
for (; memcg; memcg = parent_mem_cgroup(memcg)) { mz = mem_cgroup_zoneinfo(memcg, nid, zid); excess = res_counter_soft_limit_excess(&memcg->res); |
4e649152c memcg: some modif... |
515 516 517 518 |
/* * We have to update the tree if mz is on RB-tree or * mem is over its softlimit. */ |
ef8745c1e memcg: reduce che... |
519 |
if (excess || mz->on_tree) { |
4e649152c memcg: some modif... |
520 521 522 |
spin_lock(&mctz->lock); /* if on-tree, remove it */ if (mz->on_tree) |
c0ff4b854 memcg: rename mem... |
523 |
__mem_cgroup_remove_exceeded(memcg, mz, mctz); |
4e649152c memcg: some modif... |
524 |
/* |
ef8745c1e memcg: reduce che... |
525 526 |
* Insert again. mz->usage_in_excess will be updated. * If excess is 0, no tree ops. |
4e649152c memcg: some modif... |
527 |
*/ |
c0ff4b854 memcg: rename mem... |
528 |
__mem_cgroup_insert_exceeded(memcg, mz, mctz, excess); |
4e649152c memcg: some modif... |
529 530 |
spin_unlock(&mctz->lock); } |
f64c3f549 memory controller... |
531 532 |
} } |
c0ff4b854 memcg: rename mem... |
533 |
static void mem_cgroup_remove_from_trees(struct mem_cgroup *memcg) |
f64c3f549 memory controller... |
534 535 536 537 |
{ int node, zone; struct mem_cgroup_per_zone *mz; struct mem_cgroup_tree_per_zone *mctz; |
3ed28fa10 memcg: cleanup fo... |
538 |
for_each_node(node) { |
f64c3f549 memory controller... |
539 |
for (zone = 0; zone < MAX_NR_ZONES; zone++) { |
c0ff4b854 memcg: rename mem... |
540 |
mz = mem_cgroup_zoneinfo(memcg, node, zone); |
f64c3f549 memory controller... |
541 |
mctz = soft_limit_tree_node_zone(node, zone); |
c0ff4b854 memcg: rename mem... |
542 |
mem_cgroup_remove_exceeded(memcg, mz, mctz); |
f64c3f549 memory controller... |
543 544 545 |
} } } |
4e4169535 memory controller... |
546 547 548 549 |
static struct mem_cgroup_per_zone * __mem_cgroup_largest_soft_limit_node(struct mem_cgroup_tree_per_zone *mctz) { struct rb_node *rightmost = NULL; |
26251eaf9 memcg: fix refcnt... |
550 |
struct mem_cgroup_per_zone *mz; |
4e4169535 memory controller... |
551 552 |
retry: |
26251eaf9 memcg: fix refcnt... |
553 |
mz = NULL; |
4e4169535 memory controller... |
554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 |
rightmost = rb_last(&mctz->rb_root); if (!rightmost) goto done; /* Nothing to reclaim from */ mz = rb_entry(rightmost, struct mem_cgroup_per_zone, tree_node); /* * Remove the node now but someone else can add it back, * we will to add it back at the end of reclaim to its correct * position in the tree. */ __mem_cgroup_remove_exceeded(mz->mem, mz, mctz); if (!res_counter_soft_limit_excess(&mz->mem->res) || !css_tryget(&mz->mem->css)) goto retry; done: return mz; } static struct mem_cgroup_per_zone * mem_cgroup_largest_soft_limit_node(struct mem_cgroup_tree_per_zone *mctz) { struct mem_cgroup_per_zone *mz; spin_lock(&mctz->lock); mz = __mem_cgroup_largest_soft_limit_node(mctz); spin_unlock(&mctz->lock); return mz; } |
711d3d2c9 memcg: cpu hotplu... |
582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 |
/* * 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... |
601 |
static long mem_cgroup_read_stat(struct mem_cgroup *memcg, |
7a159cc9d memcg: use native... |
602 |
enum mem_cgroup_stat_index idx) |
c62b1a3b3 memcg: use generi... |
603 |
{ |
7a159cc9d memcg: use native... |
604 |
long val = 0; |
c62b1a3b3 memcg: use generi... |
605 |
int cpu; |
c62b1a3b3 memcg: use generi... |
606 |
|
711d3d2c9 memcg: cpu hotplu... |
607 608 |
get_online_cpus(); for_each_online_cpu(cpu) |
c0ff4b854 memcg: rename mem... |
609 |
val += per_cpu(memcg->stat->count[idx], cpu); |
711d3d2c9 memcg: cpu hotplu... |
610 |
#ifdef CONFIG_HOTPLUG_CPU |
c0ff4b854 memcg: rename mem... |
611 612 613 |
spin_lock(&memcg->pcp_counter_lock); val += memcg->nocpu_base.count[idx]; spin_unlock(&memcg->pcp_counter_lock); |
711d3d2c9 memcg: cpu hotplu... |
614 615 |
#endif put_online_cpus(); |
c62b1a3b3 memcg: use generi... |
616 617 |
return val; } |
c0ff4b854 memcg: rename mem... |
618 |
static void mem_cgroup_swap_statistics(struct mem_cgroup *memcg, |
0c3e73e84 memcg: improve re... |
619 620 621 |
bool charge) { int val = (charge) ? 1 : -1; |
c0ff4b854 memcg: rename mem... |
622 |
this_cpu_add(memcg->stat->count[MEM_CGROUP_STAT_SWAPOUT], val); |
0c3e73e84 memcg: improve re... |
623 |
} |
c0ff4b854 memcg: rename mem... |
624 |
static unsigned long mem_cgroup_read_events(struct mem_cgroup *memcg, |
e9f8974f2 memcg: break out ... |
625 626 627 628 629 630 |
enum mem_cgroup_events_index idx) { unsigned long val = 0; int cpu; for_each_online_cpu(cpu) |
c0ff4b854 memcg: rename mem... |
631 |
val += per_cpu(memcg->stat->events[idx], cpu); |
e9f8974f2 memcg: break out ... |
632 |
#ifdef CONFIG_HOTPLUG_CPU |
c0ff4b854 memcg: rename mem... |
633 634 635 |
spin_lock(&memcg->pcp_counter_lock); val += memcg->nocpu_base.events[idx]; spin_unlock(&memcg->pcp_counter_lock); |
e9f8974f2 memcg: break out ... |
636 637 638 |
#endif return val; } |
c0ff4b854 memcg: rename mem... |
639 |
static void mem_cgroup_charge_statistics(struct mem_cgroup *memcg, |
e401f1761 memcg: modify acc... |
640 |
bool file, int nr_pages) |
d52aa412d memory cgroup enh... |
641 |
{ |
c62b1a3b3 memcg: use generi... |
642 |
preempt_disable(); |
e401f1761 memcg: modify acc... |
643 |
if (file) |
c0ff4b854 memcg: rename mem... |
644 645 |
__this_cpu_add(memcg->stat->count[MEM_CGROUP_STAT_CACHE], nr_pages); |
d52aa412d memory cgroup enh... |
646 |
else |
c0ff4b854 memcg: rename mem... |
647 648 |
__this_cpu_add(memcg->stat->count[MEM_CGROUP_STAT_RSS], nr_pages); |
55e462b05 memcg: simple sta... |
649 |
|
e401f1761 memcg: modify acc... |
650 651 |
/* pagein of a big page is an event. So, ignore page size */ if (nr_pages > 0) |
c0ff4b854 memcg: rename mem... |
652 |
__this_cpu_inc(memcg->stat->events[MEM_CGROUP_EVENTS_PGPGIN]); |
3751d6043 memcg: fix event ... |
653 |
else { |
c0ff4b854 memcg: rename mem... |
654 |
__this_cpu_inc(memcg->stat->events[MEM_CGROUP_EVENTS_PGPGOUT]); |
3751d6043 memcg: fix event ... |
655 656 |
nr_pages = -nr_pages; /* for event */ } |
e401f1761 memcg: modify acc... |
657 |
|
c0ff4b854 memcg: rename mem... |
658 |
__this_cpu_add(memcg->stat->events[MEM_CGROUP_EVENTS_COUNT], nr_pages); |
2e72b6347 memcg: implement ... |
659 |
|
c62b1a3b3 memcg: use generi... |
660 |
preempt_enable(); |
6d12e2d8d per-zone and recl... |
661 |
} |
bb2a0de92 memcg: consolidat... |
662 |
unsigned long |
c0ff4b854 memcg: rename mem... |
663 |
mem_cgroup_zone_nr_lru_pages(struct mem_cgroup *memcg, int nid, int zid, |
bb2a0de92 memcg: consolidat... |
664 |
unsigned int lru_mask) |
889976dbc memcg: reclaim me... |
665 666 |
{ struct mem_cgroup_per_zone *mz; |
bb2a0de92 memcg: consolidat... |
667 668 |
enum lru_list l; unsigned long ret = 0; |
c0ff4b854 memcg: rename mem... |
669 |
mz = mem_cgroup_zoneinfo(memcg, nid, zid); |
bb2a0de92 memcg: consolidat... |
670 671 672 673 674 675 676 677 678 |
for_each_lru(l) { if (BIT(l) & lru_mask) ret += MEM_CGROUP_ZSTAT(mz, l); } return ret; } static unsigned long |
c0ff4b854 memcg: rename mem... |
679 |
mem_cgroup_node_nr_lru_pages(struct mem_cgroup *memcg, |
bb2a0de92 memcg: consolidat... |
680 681 |
int nid, unsigned int lru_mask) { |
889976dbc memcg: reclaim me... |
682 683 |
u64 total = 0; int zid; |
bb2a0de92 memcg: consolidat... |
684 |
for (zid = 0; zid < MAX_NR_ZONES; zid++) |
c0ff4b854 memcg: rename mem... |
685 686 |
total += mem_cgroup_zone_nr_lru_pages(memcg, nid, zid, lru_mask); |
bb2a0de92 memcg: consolidat... |
687 |
|
889976dbc memcg: reclaim me... |
688 689 |
return total; } |
bb2a0de92 memcg: consolidat... |
690 |
|
c0ff4b854 memcg: rename mem... |
691 |
static unsigned long mem_cgroup_nr_lru_pages(struct mem_cgroup *memcg, |
bb2a0de92 memcg: consolidat... |
692 |
unsigned int lru_mask) |
6d12e2d8d per-zone and recl... |
693 |
{ |
889976dbc memcg: reclaim me... |
694 |
int nid; |
6d12e2d8d per-zone and recl... |
695 |
u64 total = 0; |
bb2a0de92 memcg: consolidat... |
696 |
for_each_node_state(nid, N_HIGH_MEMORY) |
c0ff4b854 memcg: rename mem... |
697 |
total += mem_cgroup_node_nr_lru_pages(memcg, nid, lru_mask); |
6d12e2d8d per-zone and recl... |
698 |
return total; |
d52aa412d memory cgroup enh... |
699 |
} |
f53d7ce32 mm: memcg: shorte... |
700 701 |
static bool mem_cgroup_event_ratelimit(struct mem_cgroup *memcg, enum mem_cgroup_events_target target) |
7a159cc9d memcg: use native... |
702 703 |
{ unsigned long val, next; |
4799401fe memcg: Fix race c... |
704 705 |
val = __this_cpu_read(memcg->stat->events[MEM_CGROUP_EVENTS_COUNT]); next = __this_cpu_read(memcg->stat->targets[target]); |
7a159cc9d memcg: use native... |
706 |
/* from time_after() in jiffies.h */ |
f53d7ce32 mm: memcg: shorte... |
707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 |
if ((long)next - (long)val < 0) { switch (target) { case MEM_CGROUP_TARGET_THRESH: next = val + THRESHOLDS_EVENTS_TARGET; break; case MEM_CGROUP_TARGET_SOFTLIMIT: next = val + SOFTLIMIT_EVENTS_TARGET; break; 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... |
723 |
} |
f53d7ce32 mm: memcg: shorte... |
724 |
return false; |
d2265e6fa memcg : share eve... |
725 726 727 728 729 730 |
} /* * Check events in order. * */ |
c0ff4b854 memcg: rename mem... |
731 |
static void memcg_check_events(struct mem_cgroup *memcg, struct page *page) |
d2265e6fa memcg : share eve... |
732 |
{ |
4799401fe memcg: Fix race c... |
733 |
preempt_disable(); |
d2265e6fa memcg : share eve... |
734 |
/* threshold event is triggered in finer grain than soft limit */ |
f53d7ce32 mm: memcg: shorte... |
735 736 737 738 739 740 741 742 743 744 745 |
if (unlikely(mem_cgroup_event_ratelimit(memcg, MEM_CGROUP_TARGET_THRESH))) { bool do_softlimit, do_numainfo; do_softlimit = mem_cgroup_event_ratelimit(memcg, MEM_CGROUP_TARGET_SOFTLIMIT); #if MAX_NUMNODES > 1 do_numainfo = mem_cgroup_event_ratelimit(memcg, MEM_CGROUP_TARGET_NUMAINFO); #endif preempt_enable(); |
c0ff4b854 memcg: rename mem... |
746 |
mem_cgroup_threshold(memcg); |
f53d7ce32 mm: memcg: shorte... |
747 |
if (unlikely(do_softlimit)) |
c0ff4b854 memcg: rename mem... |
748 |
mem_cgroup_update_tree(memcg, page); |
453a9bf34 memcg: fix numa s... |
749 |
#if MAX_NUMNODES > 1 |
f53d7ce32 mm: memcg: shorte... |
750 |
if (unlikely(do_numainfo)) |
c0ff4b854 memcg: rename mem... |
751 |
atomic_inc(&memcg->numainfo_events); |
453a9bf34 memcg: fix numa s... |
752 |
#endif |
f53d7ce32 mm: memcg: shorte... |
753 754 |
} else preempt_enable(); |
d2265e6fa memcg : share eve... |
755 |
} |
d1a4c0b37 tcp memory pressu... |
756 |
struct mem_cgroup *mem_cgroup_from_cont(struct cgroup *cont) |
8cdea7c05 Memory controller... |
757 758 759 760 761 |
{ return container_of(cgroup_subsys_state(cont, mem_cgroup_subsys_id), struct mem_cgroup, css); } |
cf475ad28 cgroups: add an o... |
762 |
struct mem_cgroup *mem_cgroup_from_task(struct task_struct *p) |
78fb74669 Memory controller... |
763 |
{ |
31a78f23b mm owner: fix rac... |
764 765 766 767 768 769 770 |
/* * 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; |
78fb74669 Memory controller... |
771 772 773 |
return container_of(task_subsys_state(p, mem_cgroup_subsys_id), struct mem_cgroup, css); } |
a433658c3 vmscan,memcg: mem... |
774 |
struct mem_cgroup *try_get_mem_cgroup_from_mm(struct mm_struct *mm) |
54595fe26 memcg: use css_tr... |
775 |
{ |
c0ff4b854 memcg: rename mem... |
776 |
struct mem_cgroup *memcg = NULL; |
0b7f569e4 memcg: fix OOM ki... |
777 778 779 |
if (!mm) return NULL; |
54595fe26 memcg: use css_tr... |
780 781 782 783 784 785 786 |
/* * Because we have no locks, mm->owner's may be being moved to other * cgroup. We use css_tryget() here even if this looks * pessimistic (rather than adding locks here). */ rcu_read_lock(); do { |
c0ff4b854 memcg: rename mem... |
787 788 |
memcg = mem_cgroup_from_task(rcu_dereference(mm->owner)); if (unlikely(!memcg)) |
54595fe26 memcg: use css_tr... |
789 |
break; |
c0ff4b854 memcg: rename mem... |
790 |
} while (!css_tryget(&memcg->css)); |
54595fe26 memcg: use css_tr... |
791 |
rcu_read_unlock(); |
c0ff4b854 memcg: rename mem... |
792 |
return memcg; |
54595fe26 memcg: use css_tr... |
793 |
} |
5660048cc mm: move memcg hi... |
794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 |
/** * 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. */ struct mem_cgroup *mem_cgroup_iter(struct mem_cgroup *root, struct mem_cgroup *prev, struct mem_cgroup_reclaim_cookie *reclaim) |
14067bb3e memcg: hierarchic... |
814 |
{ |
9f3a0d093 mm: memcg: consol... |
815 816 |
struct mem_cgroup *memcg = NULL; int id = 0; |
711d3d2c9 memcg: cpu hotplu... |
817 |
|
5660048cc mm: move memcg hi... |
818 819 |
if (mem_cgroup_disabled()) return NULL; |
9f3a0d093 mm: memcg: consol... |
820 821 |
if (!root) root = root_mem_cgroup; |
7d74b06f2 memcg: use for_ea... |
822 |
|
9f3a0d093 mm: memcg: consol... |
823 824 |
if (prev && !reclaim) id = css_id(&prev->css); |
14067bb3e memcg: hierarchic... |
825 |
|
9f3a0d093 mm: memcg: consol... |
826 827 |
if (prev && prev != root) css_put(&prev->css); |
14067bb3e memcg: hierarchic... |
828 |
|
9f3a0d093 mm: memcg: consol... |
829 830 831 832 833 |
if (!root->use_hierarchy && root != root_mem_cgroup) { if (prev) return NULL; return root; } |
14067bb3e memcg: hierarchic... |
834 |
|
9f3a0d093 mm: memcg: consol... |
835 |
while (!memcg) { |
527a5ec9a mm: memcg: per-pr... |
836 |
struct mem_cgroup_reclaim_iter *uninitialized_var(iter); |
9f3a0d093 mm: memcg: consol... |
837 |
struct cgroup_subsys_state *css; |
711d3d2c9 memcg: cpu hotplu... |
838 |
|
527a5ec9a mm: memcg: per-pr... |
839 840 841 842 843 844 845 846 847 848 849 |
if (reclaim) { int nid = zone_to_nid(reclaim->zone); int zid = zone_idx(reclaim->zone); struct mem_cgroup_per_zone *mz; mz = mem_cgroup_zoneinfo(root, nid, zid); iter = &mz->reclaim_iter[reclaim->priority]; if (prev && reclaim->generation != iter->generation) return NULL; id = iter->position; } |
7d74b06f2 memcg: use for_ea... |
850 |
|
9f3a0d093 mm: memcg: consol... |
851 852 853 854 855 856 857 858 |
rcu_read_lock(); css = css_get_next(&mem_cgroup_subsys, id + 1, &root->css, &id); if (css) { if (css == &root->css || css_tryget(css)) memcg = container_of(css, struct mem_cgroup, css); } else id = 0; |
14067bb3e memcg: hierarchic... |
859 |
rcu_read_unlock(); |
14067bb3e memcg: hierarchic... |
860 |
|
527a5ec9a mm: memcg: per-pr... |
861 862 863 864 865 866 867 |
if (reclaim) { iter->position = id; if (!css) iter->generation++; else if (!prev && memcg) reclaim->generation = iter->generation; } |
9f3a0d093 mm: memcg: consol... |
868 869 870 871 872 |
if (prev && !css) return NULL; } return memcg; |
14067bb3e memcg: hierarchic... |
873 |
} |
7d74b06f2 memcg: use for_ea... |
874 |
|
5660048cc mm: move memcg hi... |
875 876 877 878 879 880 881 |
/** * 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... |
882 883 884 885 886 887 |
{ if (!root) root = root_mem_cgroup; if (prev && prev != root) css_put(&prev->css); } |
7d74b06f2 memcg: use for_ea... |
888 |
|
9f3a0d093 mm: memcg: consol... |
889 890 891 892 893 894 |
/* * 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... |
895 |
for (iter = mem_cgroup_iter(root, NULL, NULL); \ |
9f3a0d093 mm: memcg: consol... |
896 |
iter != NULL; \ |
527a5ec9a mm: memcg: per-pr... |
897 |
iter = mem_cgroup_iter(root, iter, NULL)) |
711d3d2c9 memcg: cpu hotplu... |
898 |
|
9f3a0d093 mm: memcg: consol... |
899 |
#define for_each_mem_cgroup(iter) \ |
527a5ec9a mm: memcg: per-pr... |
900 |
for (iter = mem_cgroup_iter(NULL, NULL, NULL); \ |
9f3a0d093 mm: memcg: consol... |
901 |
iter != NULL; \ |
527a5ec9a mm: memcg: per-pr... |
902 |
iter = mem_cgroup_iter(NULL, iter, NULL)) |
14067bb3e memcg: hierarchic... |
903 |
|
c0ff4b854 memcg: rename mem... |
904 |
static inline bool mem_cgroup_is_root(struct mem_cgroup *memcg) |
4b3bde4c9 memcg: remove the... |
905 |
{ |
c0ff4b854 memcg: rename mem... |
906 |
return (memcg == root_mem_cgroup); |
4b3bde4c9 memcg: remove the... |
907 |
} |
456f998ec memcg: add the pa... |
908 909 |
void mem_cgroup_count_vm_event(struct mm_struct *mm, enum vm_event_item idx) { |
c0ff4b854 memcg: rename mem... |
910 |
struct mem_cgroup *memcg; |
456f998ec memcg: add the pa... |
911 912 913 914 915 |
if (!mm) return; rcu_read_lock(); |
c0ff4b854 memcg: rename mem... |
916 917 |
memcg = mem_cgroup_from_task(rcu_dereference(mm->owner)); if (unlikely(!memcg)) |
456f998ec memcg: add the pa... |
918 919 920 |
goto out; switch (idx) { |
456f998ec memcg: add the pa... |
921 |
case PGFAULT: |
0e574a932 mm: memcg: clean ... |
922 923 924 925 |
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... |
926 927 928 929 930 931 932 933 |
break; default: BUG(); } out: rcu_read_unlock(); } EXPORT_SYMBOL(mem_cgroup_count_vm_event); |
925b7673c mm: make per-memc... |
934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 |
/** * mem_cgroup_zone_lruvec - get the lru list vector for a zone and memcg * @zone: zone of the wanted lruvec * @mem: memcg of the wanted lruvec * * 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; if (mem_cgroup_disabled()) return &zone->lruvec; mz = mem_cgroup_zoneinfo(memcg, zone_to_nid(zone), zone_idx(zone)); return &mz->lruvec; } |
08e552c69 memcg: synchroniz... |
954 955 956 957 958 959 960 961 962 963 964 965 966 |
/* * Following LRU functions are allowed to be used without PCG_LOCK. * Operations are called by routine of global LRU independently from memcg. * What we have to take care of here is validness of pc->mem_cgroup. * * Changes to pc->mem_cgroup happens when * 1. charge * 2. moving account * In typical case, "charge" is done before add-to-lru. Exception is SwapCache. * It is added to LRU before charge. * If PCG_USED bit is not set, page_cgroup is not added to this private LRU. * When moving account, the page is not on LRU. It's isolated. */ |
4f98a2fee vmscan: split LRU... |
967 |
|
925b7673c mm: make per-memc... |
968 969 970 971 972 973 974 975 976 977 978 979 980 981 |
/** * mem_cgroup_lru_add_list - account for adding an lru page and return lruvec * @zone: zone of the page * @page: the page * @lru: current lru * * This function accounts for @page being added to @lru, and returns * the lruvec for the given @zone and the memcg @page is charged to. * * The callsite is then responsible for physically linking the page to * the returned lruvec->lists[@lru]. */ struct lruvec *mem_cgroup_lru_add_list(struct zone *zone, struct page *page, enum lru_list lru) |
08e552c69 memcg: synchroniz... |
982 |
{ |
08e552c69 memcg: synchroniz... |
983 |
struct mem_cgroup_per_zone *mz; |
925b7673c mm: make per-memc... |
984 985 |
struct mem_cgroup *memcg; struct page_cgroup *pc; |
6d12e2d8d per-zone and recl... |
986 |
|
f8d665422 memcg: add mem_cg... |
987 |
if (mem_cgroup_disabled()) |
925b7673c mm: make per-memc... |
988 |
return &zone->lruvec; |
08e552c69 memcg: synchroniz... |
989 |
pc = lookup_page_cgroup(page); |
38c5d72f3 memcg: simplify L... |
990 |
memcg = pc->mem_cgroup; |
925b7673c mm: make per-memc... |
991 992 993 994 |
mz = page_cgroup_zoneinfo(memcg, page); /* compound_order() is stabilized through lru_lock */ MEM_CGROUP_ZSTAT(mz, lru) += 1 << compound_order(page); return &mz->lruvec; |
08e552c69 memcg: synchroniz... |
995 |
} |
b69408e88 vmscan: Use an in... |
996 |
|
925b7673c mm: make per-memc... |
997 998 999 1000 1001 1002 1003 1004 1005 |
/** * mem_cgroup_lru_del_list - account for removing an lru page * @page: the page * @lru: target lru * * This function accounts for @page being removed from @lru. * * The callsite is then responsible for physically unlinking * @page->lru. |
3f58a8294 memcg: move memcg... |
1006 |
*/ |
925b7673c mm: make per-memc... |
1007 |
void mem_cgroup_lru_del_list(struct page *page, enum lru_list lru) |
3f58a8294 memcg: move memcg... |
1008 1009 |
{ struct mem_cgroup_per_zone *mz; |
925b7673c mm: make per-memc... |
1010 |
struct mem_cgroup *memcg; |
3f58a8294 memcg: move memcg... |
1011 |
struct page_cgroup *pc; |
3f58a8294 memcg: move memcg... |
1012 1013 1014 1015 1016 |
if (mem_cgroup_disabled()) return; pc = lookup_page_cgroup(page); |
38c5d72f3 memcg: simplify L... |
1017 1018 |
memcg = pc->mem_cgroup; VM_BUG_ON(!memcg); |
925b7673c mm: make per-memc... |
1019 1020 |
mz = page_cgroup_zoneinfo(memcg, page); /* huge page split is done under lru_lock. so, we have no races. */ |
12d271078 memcg: fix split_... |
1021 |
VM_BUG_ON(MEM_CGROUP_ZSTAT(mz, lru) < (1 << compound_order(page))); |
925b7673c mm: make per-memc... |
1022 |
MEM_CGROUP_ZSTAT(mz, lru) -= 1 << compound_order(page); |
3f58a8294 memcg: move memcg... |
1023 |
} |
925b7673c mm: make per-memc... |
1024 |
void mem_cgroup_lru_del(struct page *page) |
08e552c69 memcg: synchroniz... |
1025 |
{ |
925b7673c mm: make per-memc... |
1026 |
mem_cgroup_lru_del_list(page, page_lru(page)); |
6d12e2d8d per-zone and recl... |
1027 |
} |
925b7673c mm: make per-memc... |
1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 |
/** * mem_cgroup_lru_move_lists - account for moving a page between lrus * @zone: zone of the page * @page: the page * @from: current lru * @to: target lru * * This function accounts for @page being moved between the lrus @from * and @to, and returns the lruvec for the given @zone and the memcg * @page is charged to. * * The callsite is then responsible for physically relinking * @page->lru to the returned lruvec->lists[@to]. */ struct lruvec *mem_cgroup_lru_move_lists(struct zone *zone, struct page *page, enum lru_list from, enum lru_list to) |
66e1707bc Memory controller... |
1046 |
{ |
925b7673c mm: make per-memc... |
1047 1048 1049 |
/* XXX: Optimize this, especially for @from == @to */ mem_cgroup_lru_del_list(page, from); return mem_cgroup_lru_add_list(zone, page, to); |
08e552c69 memcg: synchroniz... |
1050 |
} |
544122e5e memcg: fix LRU ac... |
1051 |
|
08e552c69 memcg: synchroniz... |
1052 |
/* |
c0ff4b854 memcg: rename mem... |
1053 |
* Checks whether given mem is same or in the root_mem_cgroup's |
3e92041d6 memcg: add mem_cg... |
1054 1055 |
* hierarchy subtree */ |
c0ff4b854 memcg: rename mem... |
1056 1057 |
static bool mem_cgroup_same_or_subtree(const struct mem_cgroup *root_memcg, struct mem_cgroup *memcg) |
3e92041d6 memcg: add mem_cg... |
1058 |
{ |
c0ff4b854 memcg: rename mem... |
1059 1060 1061 |
if (root_memcg != memcg) { return (root_memcg->use_hierarchy && css_is_ancestor(&memcg->css, &root_memcg->css)); |
3e92041d6 memcg: add mem_cg... |
1062 1063 1064 1065 |
} return true; } |
c0ff4b854 memcg: rename mem... |
1066 |
int task_in_mem_cgroup(struct task_struct *task, const struct mem_cgroup *memcg) |
4c4a22148 memcontrol: move ... |
1067 1068 |
{ int ret; |
0b7f569e4 memcg: fix OOM ki... |
1069 |
struct mem_cgroup *curr = NULL; |
158e0a2d1 memcg: use find_l... |
1070 |
struct task_struct *p; |
4c4a22148 memcontrol: move ... |
1071 |
|
158e0a2d1 memcg: use find_l... |
1072 |
p = find_lock_task_mm(task); |
de077d222 oom, memcg: fix e... |
1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 |
if (p) { curr = try_get_mem_cgroup_from_mm(p->mm); 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. */ task_lock(task); curr = mem_cgroup_from_task(task); if (curr) css_get(&curr->css); task_unlock(task); } |
0b7f569e4 memcg: fix OOM ki... |
1088 1089 |
if (!curr) return 0; |
d31f56dbf memcg: avoid oom-... |
1090 |
/* |
c0ff4b854 memcg: rename mem... |
1091 |
* We should check use_hierarchy of "memcg" not "curr". Because checking |
d31f56dbf memcg: avoid oom-... |
1092 |
* use_hierarchy of "curr" here make this function true if hierarchy is |
c0ff4b854 memcg: rename mem... |
1093 1094 |
* enabled in "curr" and "curr" is a child of "memcg" in *cgroup* * hierarchy(even if use_hierarchy is disabled in "memcg"). |
d31f56dbf memcg: avoid oom-... |
1095 |
*/ |
c0ff4b854 memcg: rename mem... |
1096 |
ret = mem_cgroup_same_or_subtree(memcg, curr); |
0b7f569e4 memcg: fix OOM ki... |
1097 |
css_put(&curr->css); |
4c4a22148 memcontrol: move ... |
1098 1099 |
return ret; } |
9b272977e memcg: skip scann... |
1100 |
int mem_cgroup_inactive_anon_is_low(struct mem_cgroup *memcg, struct zone *zone) |
14797e236 memcg: add inacti... |
1101 |
{ |
9b272977e memcg: skip scann... |
1102 1103 1104 |
unsigned long inactive_ratio; int nid = zone_to_nid(zone); int zid = zone_idx(zone); |
14797e236 memcg: add inacti... |
1105 |
unsigned long inactive; |
9b272977e memcg: skip scann... |
1106 |
unsigned long active; |
c772be939 memcg: fix calcul... |
1107 |
unsigned long gb; |
14797e236 memcg: add inacti... |
1108 |
|
9b272977e memcg: skip scann... |
1109 1110 1111 1112 |
inactive = mem_cgroup_zone_nr_lru_pages(memcg, nid, zid, BIT(LRU_INACTIVE_ANON)); active = mem_cgroup_zone_nr_lru_pages(memcg, nid, zid, BIT(LRU_ACTIVE_ANON)); |
14797e236 memcg: add inacti... |
1113 |
|
c772be939 memcg: fix calcul... |
1114 1115 1116 1117 1118 |
gb = (inactive + active) >> (30 - PAGE_SHIFT); if (gb) inactive_ratio = int_sqrt(10 * gb); else inactive_ratio = 1; |
9b272977e memcg: skip scann... |
1119 |
return inactive * inactive_ratio < active; |
14797e236 memcg: add inacti... |
1120 |
} |
9b272977e memcg: skip scann... |
1121 |
int mem_cgroup_inactive_file_is_low(struct mem_cgroup *memcg, struct zone *zone) |
56e49d218 vmscan: evict use... |
1122 1123 1124 |
{ unsigned long active; unsigned long inactive; |
9b272977e memcg: skip scann... |
1125 1126 |
int zid = zone_idx(zone); int nid = zone_to_nid(zone); |
56e49d218 vmscan: evict use... |
1127 |
|
9b272977e memcg: skip scann... |
1128 1129 1130 1131 |
inactive = mem_cgroup_zone_nr_lru_pages(memcg, nid, zid, BIT(LRU_INACTIVE_FILE)); active = mem_cgroup_zone_nr_lru_pages(memcg, nid, zid, BIT(LRU_ACTIVE_FILE)); |
56e49d218 vmscan: evict use... |
1132 1133 1134 |
return (active > inactive); } |
3e2f41f1f memcg: add zone_r... |
1135 1136 1137 |
struct zone_reclaim_stat *mem_cgroup_get_reclaim_stat(struct mem_cgroup *memcg, struct zone *zone) { |
13d7e3a2d memcg: convert to... |
1138 |
int nid = zone_to_nid(zone); |
3e2f41f1f memcg: add zone_r... |
1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 |
int zid = zone_idx(zone); struct mem_cgroup_per_zone *mz = mem_cgroup_zoneinfo(memcg, nid, zid); return &mz->reclaim_stat; } struct zone_reclaim_stat * mem_cgroup_get_reclaim_stat_from_page(struct page *page) { struct page_cgroup *pc; struct mem_cgroup_per_zone *mz; if (mem_cgroup_disabled()) return NULL; pc = lookup_page_cgroup(page); |
bd112db87 memcg: fix mem_cg... |
1155 1156 |
if (!PageCgroupUsed(pc)) return NULL; |
713735b42 memcg: correctly ... |
1157 1158 |
/* Ensure pc->mem_cgroup is visible after reading PCG_USED. */ smp_rmb(); |
97a6c37b3 memcg: change pag... |
1159 |
mz = page_cgroup_zoneinfo(pc->mem_cgroup, page); |
3e2f41f1f memcg: add zone_r... |
1160 1161 |
return &mz->reclaim_stat; } |
6d61ef409 memcg: memory cgr... |
1162 1163 |
#define mem_cgroup_from_res_counter(counter, member) \ container_of(counter, struct mem_cgroup, member) |
19942822d memcg: prevent en... |
1164 |
/** |
9d11ea9f1 memcg: simplify t... |
1165 1166 |
* mem_cgroup_margin - calculate chargeable space of a memory cgroup * @mem: the memory cgroup |
19942822d memcg: prevent en... |
1167 |
* |
9d11ea9f1 memcg: simplify t... |
1168 |
* Returns the maximum amount of memory @mem can be charged with, in |
7ec99d621 memcg: unify char... |
1169 |
* pages. |
19942822d memcg: prevent en... |
1170 |
*/ |
c0ff4b854 memcg: rename mem... |
1171 |
static unsigned long mem_cgroup_margin(struct mem_cgroup *memcg) |
19942822d memcg: prevent en... |
1172 |
{ |
9d11ea9f1 memcg: simplify t... |
1173 |
unsigned long long margin; |
c0ff4b854 memcg: rename mem... |
1174 |
margin = res_counter_margin(&memcg->res); |
9d11ea9f1 memcg: simplify t... |
1175 |
if (do_swap_account) |
c0ff4b854 memcg: rename mem... |
1176 |
margin = min(margin, res_counter_margin(&memcg->memsw)); |
7ec99d621 memcg: unify char... |
1177 |
return margin >> PAGE_SHIFT; |
19942822d memcg: prevent en... |
1178 |
} |
1f4c025b5 memcg: export mem... |
1179 |
int mem_cgroup_swappiness(struct mem_cgroup *memcg) |
a7885eb8a memcg: swappiness |
1180 1181 |
{ struct cgroup *cgrp = memcg->css.cgroup; |
a7885eb8a memcg: swappiness |
1182 1183 1184 1185 |
/* root ? */ if (cgrp->parent == NULL) return vm_swappiness; |
bf1ff2635 memcg: remove mem... |
1186 |
return memcg->swappiness; |
a7885eb8a memcg: swappiness |
1187 |
} |
c0ff4b854 memcg: rename mem... |
1188 |
static void mem_cgroup_start_move(struct mem_cgroup *memcg) |
32047e2a8 memcg: avoid lock... |
1189 1190 |
{ int cpu; |
1489ebad8 memcg: cpu hotplu... |
1191 1192 |
get_online_cpus(); |
c0ff4b854 memcg: rename mem... |
1193 |
spin_lock(&memcg->pcp_counter_lock); |
1489ebad8 memcg: cpu hotplu... |
1194 |
for_each_online_cpu(cpu) |
c0ff4b854 memcg: rename mem... |
1195 1196 1197 |
per_cpu(memcg->stat->count[MEM_CGROUP_ON_MOVE], cpu) += 1; memcg->nocpu_base.count[MEM_CGROUP_ON_MOVE] += 1; spin_unlock(&memcg->pcp_counter_lock); |
1489ebad8 memcg: cpu hotplu... |
1198 |
put_online_cpus(); |
32047e2a8 memcg: avoid lock... |
1199 1200 1201 |
synchronize_rcu(); } |
c0ff4b854 memcg: rename mem... |
1202 |
static void mem_cgroup_end_move(struct mem_cgroup *memcg) |
32047e2a8 memcg: avoid lock... |
1203 1204 |
{ int cpu; |
c0ff4b854 memcg: rename mem... |
1205 |
if (!memcg) |
32047e2a8 memcg: avoid lock... |
1206 |
return; |
1489ebad8 memcg: cpu hotplu... |
1207 |
get_online_cpus(); |
c0ff4b854 memcg: rename mem... |
1208 |
spin_lock(&memcg->pcp_counter_lock); |
1489ebad8 memcg: cpu hotplu... |
1209 |
for_each_online_cpu(cpu) |
c0ff4b854 memcg: rename mem... |
1210 1211 1212 |
per_cpu(memcg->stat->count[MEM_CGROUP_ON_MOVE], cpu) -= 1; memcg->nocpu_base.count[MEM_CGROUP_ON_MOVE] -= 1; spin_unlock(&memcg->pcp_counter_lock); |
1489ebad8 memcg: cpu hotplu... |
1213 |
put_online_cpus(); |
32047e2a8 memcg: avoid lock... |
1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 |
} /* * 2 routines for checking "mem" is under move_account() or not. * * mem_cgroup_stealed() - checking a cgroup is mc.from or not. This is used * for avoiding race in accounting. If true, * pc->mem_cgroup may be overwritten. * * mem_cgroup_under_move() - checking a cgroup is mc.from or mc.to or * under hierarchy of moving cgroups. This is for * waiting at hith-memory prressure caused by "move". */ |
c0ff4b854 memcg: rename mem... |
1226 |
static bool mem_cgroup_stealed(struct mem_cgroup *memcg) |
32047e2a8 memcg: avoid lock... |
1227 1228 |
{ VM_BUG_ON(!rcu_read_lock_held()); |
c0ff4b854 memcg: rename mem... |
1229 |
return this_cpu_read(memcg->stat->count[MEM_CGROUP_ON_MOVE]) > 0; |
32047e2a8 memcg: avoid lock... |
1230 |
} |
4b5343346 memcg: clean up t... |
1231 |
|
c0ff4b854 memcg: rename mem... |
1232 |
static bool mem_cgroup_under_move(struct mem_cgroup *memcg) |
4b5343346 memcg: clean up t... |
1233 |
{ |
2bd9bb206 memcg: clean up w... |
1234 1235 |
struct mem_cgroup *from; struct mem_cgroup *to; |
4b5343346 memcg: clean up t... |
1236 |
bool ret = false; |
2bd9bb206 memcg: clean up w... |
1237 1238 1239 1240 1241 1242 1243 1244 1245 |
/* * 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... |
1246 |
|
c0ff4b854 memcg: rename mem... |
1247 1248 |
ret = mem_cgroup_same_or_subtree(memcg, from) || mem_cgroup_same_or_subtree(memcg, to); |
2bd9bb206 memcg: clean up w... |
1249 1250 |
unlock: spin_unlock(&mc.lock); |
4b5343346 memcg: clean up t... |
1251 1252 |
return ret; } |
c0ff4b854 memcg: rename mem... |
1253 |
static bool mem_cgroup_wait_acct_move(struct mem_cgroup *memcg) |
4b5343346 memcg: clean up t... |
1254 1255 |
{ if (mc.moving_task && current != mc.moving_task) { |
c0ff4b854 memcg: rename mem... |
1256 |
if (mem_cgroup_under_move(memcg)) { |
4b5343346 memcg: clean up t... |
1257 1258 1259 1260 1261 1262 1263 1264 1265 1266 1267 |
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; } |
e222432bf memcg: show memcg... |
1268 |
/** |
6a6135b64 memcg: typo in co... |
1269 |
* mem_cgroup_print_oom_info: Called from OOM with tasklist_lock held in read mode. |
e222432bf memcg: show memcg... |
1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 |
* @memcg: The memory cgroup that went over limit * @p: Task that is going to be killed * * NOTE: @memcg and @p's mem_cgroup can be different when hierarchy is * enabled */ void mem_cgroup_print_oom_info(struct mem_cgroup *memcg, struct task_struct *p) { struct cgroup *task_cgrp; struct cgroup *mem_cgrp; /* * Need a buffer in BSS, can't rely on allocations. The code relies * on the assumption that OOM is serialized for memory controller. * If this assumption is broken, revisit this code. */ static char memcg_name[PATH_MAX]; int ret; |
d31f56dbf memcg: avoid oom-... |
1287 |
if (!memcg || !p) |
e222432bf memcg: show memcg... |
1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 |
return; rcu_read_lock(); mem_cgrp = memcg->css.cgroup; task_cgrp = task_cgroup(p, mem_cgroup_subsys_id); ret = cgroup_path(task_cgrp, memcg_name, PATH_MAX); if (ret < 0) { /* * Unfortunately, we are unable to convert to a useful name * But we'll still print out the usage information */ rcu_read_unlock(); goto done; } rcu_read_unlock(); printk(KERN_INFO "Task in %s killed", memcg_name); rcu_read_lock(); ret = cgroup_path(mem_cgrp, memcg_name, PATH_MAX); if (ret < 0) { rcu_read_unlock(); goto done; } rcu_read_unlock(); /* * Continues from above, so we don't need an KERN_ level */ printk(KERN_CONT " as a result of limit of %s ", memcg_name); done: printk(KERN_INFO "memory: usage %llukB, limit %llukB, failcnt %llu ", res_counter_read_u64(&memcg->res, RES_USAGE) >> 10, res_counter_read_u64(&memcg->res, RES_LIMIT) >> 10, res_counter_read_u64(&memcg->res, RES_FAILCNT)); printk(KERN_INFO "memory+swap: usage %llukB, limit %llukB, " "failcnt %llu ", res_counter_read_u64(&memcg->memsw, RES_USAGE) >> 10, res_counter_read_u64(&memcg->memsw, RES_LIMIT) >> 10, res_counter_read_u64(&memcg->memsw, RES_FAILCNT)); } |
81d39c20f memcg: fix shrink... |
1336 1337 1338 1339 |
/* * This function returns the number of memcg under hierarchy tree. Returns * 1(self count) if no children. */ |
c0ff4b854 memcg: rename mem... |
1340 |
static int mem_cgroup_count_children(struct mem_cgroup *memcg) |
81d39c20f memcg: fix shrink... |
1341 1342 |
{ int num = 0; |
7d74b06f2 memcg: use for_ea... |
1343 |
struct mem_cgroup *iter; |
c0ff4b854 memcg: rename mem... |
1344 |
for_each_mem_cgroup_tree(iter, memcg) |
7d74b06f2 memcg: use for_ea... |
1345 |
num++; |
81d39c20f memcg: fix shrink... |
1346 1347 |
return num; } |
6d61ef409 memcg: memory cgr... |
1348 |
/* |
a63d83f42 oom: badness heur... |
1349 1350 1351 1352 1353 1354 |
* Return the memory (and swap, if configured) limit for a memcg. */ u64 mem_cgroup_get_limit(struct mem_cgroup *memcg) { u64 limit; u64 memsw; |
f3e8eb70b memcg: fix unit m... |
1355 1356 |
limit = res_counter_read_u64(&memcg->res, RES_LIMIT); limit += total_swap_pages << PAGE_SHIFT; |
a63d83f42 oom: badness heur... |
1357 1358 1359 1360 1361 1362 1363 |
memsw = res_counter_read_u64(&memcg->memsw, RES_LIMIT); /* * If memsw is finite and limits the amount of swap space available * to this memcg, return that limit. */ return min(limit, memsw); } |
5660048cc mm: move memcg hi... |
1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 |
static unsigned long mem_cgroup_reclaim(struct mem_cgroup *memcg, gfp_t gfp_mask, unsigned long flags) { unsigned long total = 0; bool noswap = false; int loop; if (flags & MEM_CGROUP_RECLAIM_NOSWAP) noswap = true; if (!(flags & MEM_CGROUP_RECLAIM_SHRINK) && memcg->memsw_is_minimum) noswap = true; for (loop = 0; loop < MEM_CGROUP_MAX_RECLAIM_LOOPS; loop++) { if (loop) drain_all_stock_async(memcg); total += try_to_free_mem_cgroup_pages(memcg, gfp_mask, noswap); /* * Allow limit shrinkers, which are triggered directly * by userspace, to catch signals and stop reclaim * after minimal progress, regardless of the margin. */ if (total && (flags & MEM_CGROUP_RECLAIM_SHRINK)) break; if (mem_cgroup_margin(memcg)) break; /* * If nothing was reclaimed after two attempts, there * may be no reclaimable pages in this hierarchy. */ if (loop && !total) break; } return total; } |
4d0c066d2 memcg: fix reclai... |
1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 |
/** * test_mem_cgroup_node_reclaimable * @mem: the target memcg * @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... |
1409 |
static bool test_mem_cgroup_node_reclaimable(struct mem_cgroup *memcg, |
4d0c066d2 memcg: fix reclai... |
1410 1411 |
int nid, bool noswap) { |
c0ff4b854 memcg: rename mem... |
1412 |
if (mem_cgroup_node_nr_lru_pages(memcg, nid, LRU_ALL_FILE)) |
4d0c066d2 memcg: fix reclai... |
1413 1414 1415 |
return true; if (noswap || !total_swap_pages) return false; |
c0ff4b854 memcg: rename mem... |
1416 |
if (mem_cgroup_node_nr_lru_pages(memcg, nid, LRU_ALL_ANON)) |
4d0c066d2 memcg: fix reclai... |
1417 1418 1419 1420 |
return true; return false; } |
889976dbc memcg: reclaim me... |
1421 1422 1423 1424 1425 1426 1427 1428 |
#if MAX_NUMNODES > 1 /* * 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... |
1429 |
static void mem_cgroup_may_update_nodemask(struct mem_cgroup *memcg) |
889976dbc memcg: reclaim me... |
1430 1431 |
{ int nid; |
453a9bf34 memcg: fix numa s... |
1432 1433 1434 1435 |
/* * numainfo_events > 0 means there was at least NUMAINFO_EVENTS_TARGET * pagein/pageout changes since the last update. */ |
c0ff4b854 memcg: rename mem... |
1436 |
if (!atomic_read(&memcg->numainfo_events)) |
453a9bf34 memcg: fix numa s... |
1437 |
return; |
c0ff4b854 memcg: rename mem... |
1438 |
if (atomic_inc_return(&memcg->numainfo_updating) > 1) |
889976dbc memcg: reclaim me... |
1439 |
return; |
889976dbc memcg: reclaim me... |
1440 |
/* make a nodemask where this memcg uses memory from */ |
c0ff4b854 memcg: rename mem... |
1441 |
memcg->scan_nodes = node_states[N_HIGH_MEMORY]; |
889976dbc memcg: reclaim me... |
1442 1443 |
for_each_node_mask(nid, node_states[N_HIGH_MEMORY]) { |
c0ff4b854 memcg: rename mem... |
1444 1445 |
if (!test_mem_cgroup_node_reclaimable(memcg, nid, false)) node_clear(nid, memcg->scan_nodes); |
889976dbc memcg: reclaim me... |
1446 |
} |
453a9bf34 memcg: fix numa s... |
1447 |
|
c0ff4b854 memcg: rename mem... |
1448 1449 |
atomic_set(&memcg->numainfo_events, 0); atomic_set(&memcg->numainfo_updating, 0); |
889976dbc memcg: reclaim me... |
1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 |
} /* * 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... |
1464 |
int mem_cgroup_select_victim_node(struct mem_cgroup *memcg) |
889976dbc memcg: reclaim me... |
1465 1466 |
{ int node; |
c0ff4b854 memcg: rename mem... |
1467 1468 |
mem_cgroup_may_update_nodemask(memcg); node = memcg->last_scanned_node; |
889976dbc memcg: reclaim me... |
1469 |
|
c0ff4b854 memcg: rename mem... |
1470 |
node = next_node(node, memcg->scan_nodes); |
889976dbc memcg: reclaim me... |
1471 |
if (node == MAX_NUMNODES) |
c0ff4b854 memcg: rename mem... |
1472 |
node = first_node(memcg->scan_nodes); |
889976dbc memcg: reclaim me... |
1473 1474 1475 1476 1477 1478 1479 1480 |
/* * 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... |
1481 |
memcg->last_scanned_node = node; |
889976dbc memcg: reclaim me... |
1482 1483 |
return node; } |
4d0c066d2 memcg: fix reclai... |
1484 1485 1486 1487 1488 1489 |
/* * Check all nodes whether it contains reclaimable pages or not. * For quick scan, we make use of scan_nodes. This will allow us to skip * unused nodes. But scan_nodes is lazily updated and may not cotain * enough new information. We need to do double check. */ |
c0ff4b854 memcg: rename mem... |
1490 |
bool mem_cgroup_reclaimable(struct mem_cgroup *memcg, bool noswap) |
4d0c066d2 memcg: fix reclai... |
1491 1492 1493 1494 1495 1496 1497 |
{ int nid; /* * quick check...making use of scan_node. * We can skip unused nodes. */ |
c0ff4b854 memcg: rename mem... |
1498 1499 |
if (!nodes_empty(memcg->scan_nodes)) { for (nid = first_node(memcg->scan_nodes); |
4d0c066d2 memcg: fix reclai... |
1500 |
nid < MAX_NUMNODES; |
c0ff4b854 memcg: rename mem... |
1501 |
nid = next_node(nid, memcg->scan_nodes)) { |
4d0c066d2 memcg: fix reclai... |
1502 |
|
c0ff4b854 memcg: rename mem... |
1503 |
if (test_mem_cgroup_node_reclaimable(memcg, nid, noswap)) |
4d0c066d2 memcg: fix reclai... |
1504 1505 1506 1507 1508 1509 1510 |
return true; } } /* * Check rest of nodes. */ for_each_node_state(nid, N_HIGH_MEMORY) { |
c0ff4b854 memcg: rename mem... |
1511 |
if (node_isset(nid, memcg->scan_nodes)) |
4d0c066d2 memcg: fix reclai... |
1512 |
continue; |
c0ff4b854 memcg: rename mem... |
1513 |
if (test_mem_cgroup_node_reclaimable(memcg, nid, noswap)) |
4d0c066d2 memcg: fix reclai... |
1514 1515 1516 1517 |
return true; } return false; } |
889976dbc memcg: reclaim me... |
1518 |
#else |
c0ff4b854 memcg: rename mem... |
1519 |
int mem_cgroup_select_victim_node(struct mem_cgroup *memcg) |
889976dbc memcg: reclaim me... |
1520 1521 1522 |
{ return 0; } |
4d0c066d2 memcg: fix reclai... |
1523 |
|
c0ff4b854 memcg: rename mem... |
1524 |
bool mem_cgroup_reclaimable(struct mem_cgroup *memcg, bool noswap) |
4d0c066d2 memcg: fix reclai... |
1525 |
{ |
c0ff4b854 memcg: rename mem... |
1526 |
return test_mem_cgroup_node_reclaimable(memcg, 0, noswap); |
4d0c066d2 memcg: fix reclai... |
1527 |
} |
889976dbc memcg: reclaim me... |
1528 |
#endif |
5660048cc mm: move memcg hi... |
1529 1530 1531 1532 |
static int mem_cgroup_soft_reclaim(struct mem_cgroup *root_memcg, struct zone *zone, gfp_t gfp_mask, unsigned long *total_scanned) |
6d61ef409 memcg: memory cgr... |
1533 |
{ |
9f3a0d093 mm: memcg: consol... |
1534 |
struct mem_cgroup *victim = NULL; |
5660048cc mm: move memcg hi... |
1535 |
int total = 0; |
04046e1a0 memcg: use CSS ID |
1536 |
int loop = 0; |
9d11ea9f1 memcg: simplify t... |
1537 |
unsigned long excess; |
185efc0f9 memcg: Revert "me... |
1538 |
unsigned long nr_scanned; |
527a5ec9a mm: memcg: per-pr... |
1539 1540 1541 1542 |
struct mem_cgroup_reclaim_cookie reclaim = { .zone = zone, .priority = 0, }; |
9d11ea9f1 memcg: simplify t... |
1543 |
|
c0ff4b854 memcg: rename mem... |
1544 |
excess = res_counter_soft_limit_excess(&root_memcg->res) >> PAGE_SHIFT; |
04046e1a0 memcg: use CSS ID |
1545 |
|
4e4169535 memory controller... |
1546 |
while (1) { |
527a5ec9a mm: memcg: per-pr... |
1547 |
victim = mem_cgroup_iter(root_memcg, victim, &reclaim); |
9f3a0d093 mm: memcg: consol... |
1548 |
if (!victim) { |
04046e1a0 memcg: use CSS ID |
1549 |
loop++; |
4e4169535 memory controller... |
1550 1551 1552 1553 1554 1555 |
if (loop >= 2) { /* * If we have not been able to reclaim * anything, it might because there are * no reclaimable pages under this hierarchy */ |
5660048cc mm: move memcg hi... |
1556 |
if (!total) |
4e4169535 memory controller... |
1557 |
break; |
4e4169535 memory controller... |
1558 |
/* |
25985edce Fix common misspe... |
1559 |
* We want to do more targeted reclaim. |
4e4169535 memory controller... |
1560 1561 1562 1563 1564 |
* 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) || |
9f3a0d093 mm: memcg: consol... |
1565 |
(loop > MEM_CGROUP_MAX_RECLAIM_LOOPS)) |
4e4169535 memory controller... |
1566 |
break; |
4e4169535 memory controller... |
1567 |
} |
9f3a0d093 mm: memcg: consol... |
1568 |
continue; |
4e4169535 memory controller... |
1569 |
} |
5660048cc mm: move memcg hi... |
1570 |
if (!mem_cgroup_reclaimable(victim, false)) |
6d61ef409 memcg: memory cgr... |
1571 |
continue; |
5660048cc mm: move memcg hi... |
1572 1573 1574 1575 |
total += mem_cgroup_shrink_node_zone(victim, gfp_mask, false, zone, &nr_scanned); *total_scanned += nr_scanned; if (!res_counter_soft_limit_excess(&root_memcg->res)) |
9f3a0d093 mm: memcg: consol... |
1576 |
break; |
6d61ef409 memcg: memory cgr... |
1577 |
} |
9f3a0d093 mm: memcg: consol... |
1578 |
mem_cgroup_iter_break(root_memcg, victim); |
04046e1a0 memcg: use CSS ID |
1579 |
return total; |
6d61ef409 memcg: memory cgr... |
1580 |
} |
867578cbc memcg: fix oom ki... |
1581 1582 1583 |
/* * Check OOM-Killer is already running under our hierarchy. * If someone is running, return false. |
1af8efe96 memcg: change mem... |
1584 |
* Has to be called with memcg_oom_lock |
867578cbc memcg: fix oom ki... |
1585 |
*/ |
c0ff4b854 memcg: rename mem... |
1586 |
static bool mem_cgroup_oom_lock(struct mem_cgroup *memcg) |
867578cbc memcg: fix oom ki... |
1587 |
{ |
79dfdaccd memcg: make oom_l... |
1588 |
struct mem_cgroup *iter, *failed = NULL; |
a636b327f memcg: avoid unne... |
1589 |
|
9f3a0d093 mm: memcg: consol... |
1590 |
for_each_mem_cgroup_tree(iter, memcg) { |
23751be00 memcg: fix hierar... |
1591 |
if (iter->oom_lock) { |
79dfdaccd memcg: make oom_l... |
1592 1593 1594 1595 |
/* * this subtree of our hierarchy is already locked * so we cannot give a lock. */ |
79dfdaccd memcg: make oom_l... |
1596 |
failed = iter; |
9f3a0d093 mm: memcg: consol... |
1597 1598 |
mem_cgroup_iter_break(memcg, iter); break; |
23751be00 memcg: fix hierar... |
1599 1600 |
} else iter->oom_lock = true; |
7d74b06f2 memcg: use for_ea... |
1601 |
} |
867578cbc memcg: fix oom ki... |
1602 |
|
79dfdaccd memcg: make oom_l... |
1603 |
if (!failed) |
23751be00 memcg: fix hierar... |
1604 |
return true; |
79dfdaccd memcg: make oom_l... |
1605 1606 1607 1608 1609 |
/* * OK, we failed to lock the whole subtree so we have to clean up * what we set up to the failing subtree */ |
9f3a0d093 mm: memcg: consol... |
1610 |
for_each_mem_cgroup_tree(iter, memcg) { |
79dfdaccd memcg: make oom_l... |
1611 |
if (iter == failed) { |
9f3a0d093 mm: memcg: consol... |
1612 1613 |
mem_cgroup_iter_break(memcg, iter); break; |
79dfdaccd memcg: make oom_l... |
1614 1615 1616 |
} iter->oom_lock = false; } |
23751be00 memcg: fix hierar... |
1617 |
return false; |
a636b327f memcg: avoid unne... |
1618 |
} |
0b7f569e4 memcg: fix OOM ki... |
1619 |
|
79dfdaccd memcg: make oom_l... |
1620 |
/* |
1af8efe96 memcg: change mem... |
1621 |
* Has to be called with memcg_oom_lock |
79dfdaccd memcg: make oom_l... |
1622 |
*/ |
c0ff4b854 memcg: rename mem... |
1623 |
static int mem_cgroup_oom_unlock(struct mem_cgroup *memcg) |
0b7f569e4 memcg: fix OOM ki... |
1624 |
{ |
7d74b06f2 memcg: use for_ea... |
1625 |
struct mem_cgroup *iter; |
c0ff4b854 memcg: rename mem... |
1626 |
for_each_mem_cgroup_tree(iter, memcg) |
79dfdaccd memcg: make oom_l... |
1627 1628 1629 |
iter->oom_lock = false; return 0; } |
c0ff4b854 memcg: rename mem... |
1630 |
static void mem_cgroup_mark_under_oom(struct mem_cgroup *memcg) |
79dfdaccd memcg: make oom_l... |
1631 1632 |
{ struct mem_cgroup *iter; |
c0ff4b854 memcg: rename mem... |
1633 |
for_each_mem_cgroup_tree(iter, memcg) |
79dfdaccd memcg: make oom_l... |
1634 1635 |
atomic_inc(&iter->under_oom); } |
c0ff4b854 memcg: rename mem... |
1636 |
static void mem_cgroup_unmark_under_oom(struct mem_cgroup *memcg) |
79dfdaccd memcg: make oom_l... |
1637 1638 |
{ struct mem_cgroup *iter; |
867578cbc memcg: fix oom ki... |
1639 1640 1641 1642 1643 |
/* * 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... |
1644 |
for_each_mem_cgroup_tree(iter, memcg) |
79dfdaccd memcg: make oom_l... |
1645 |
atomic_add_unless(&iter->under_oom, -1, 0); |
0b7f569e4 memcg: fix OOM ki... |
1646 |
} |
1af8efe96 memcg: change mem... |
1647 |
static DEFINE_SPINLOCK(memcg_oom_lock); |
867578cbc memcg: fix oom ki... |
1648 |
static DECLARE_WAIT_QUEUE_HEAD(memcg_oom_waitq); |
dc98df5a1 memcg: oom wakeup... |
1649 1650 1651 1652 1653 1654 1655 1656 |
struct oom_wait_info { struct mem_cgroup *mem; wait_queue_t wait; }; static int memcg_oom_wake_function(wait_queue_t *wait, unsigned mode, int sync, void *arg) { |
c0ff4b854 memcg: rename mem... |
1657 1658 |
struct mem_cgroup *wake_memcg = (struct mem_cgroup *)arg, *oom_wait_memcg; |
dc98df5a1 memcg: oom wakeup... |
1659 1660 1661 |
struct oom_wait_info *oom_wait_info; oom_wait_info = container_of(wait, struct oom_wait_info, wait); |
c0ff4b854 memcg: rename mem... |
1662 |
oom_wait_memcg = oom_wait_info->mem; |
dc98df5a1 memcg: oom wakeup... |
1663 |
|
dc98df5a1 memcg: oom wakeup... |
1664 1665 1666 1667 |
/* * Both of oom_wait_info->mem and wake_mem are stable under us. * Then we can use css_is_ancestor without taking care of RCU. */ |
c0ff4b854 memcg: rename mem... |
1668 1669 |
if (!mem_cgroup_same_or_subtree(oom_wait_memcg, wake_memcg) && !mem_cgroup_same_or_subtree(wake_memcg, oom_wait_memcg)) |
dc98df5a1 memcg: oom wakeup... |
1670 |
return 0; |
dc98df5a1 memcg: oom wakeup... |
1671 1672 |
return autoremove_wake_function(wait, mode, sync, arg); } |
c0ff4b854 memcg: rename mem... |
1673 |
static void memcg_wakeup_oom(struct mem_cgroup *memcg) |
dc98df5a1 memcg: oom wakeup... |
1674 |
{ |
c0ff4b854 memcg: rename mem... |
1675 1676 |
/* for filtering, pass "memcg" as argument. */ __wake_up(&memcg_oom_waitq, TASK_NORMAL, 0, memcg); |
dc98df5a1 memcg: oom wakeup... |
1677 |
} |
c0ff4b854 memcg: rename mem... |
1678 |
static void memcg_oom_recover(struct mem_cgroup *memcg) |
3c11ecf44 memcg: oom kill d... |
1679 |
{ |
c0ff4b854 memcg: rename mem... |
1680 1681 |
if (memcg && atomic_read(&memcg->under_oom)) memcg_wakeup_oom(memcg); |
3c11ecf44 memcg: oom kill d... |
1682 |
} |
867578cbc memcg: fix oom ki... |
1683 1684 1685 |
/* * try to call OOM killer. returns false if we should exit memory-reclaim loop. */ |
c0ff4b854 memcg: rename mem... |
1686 |
bool mem_cgroup_handle_oom(struct mem_cgroup *memcg, gfp_t mask) |
0b7f569e4 memcg: fix OOM ki... |
1687 |
{ |
dc98df5a1 memcg: oom wakeup... |
1688 |
struct oom_wait_info owait; |
3c11ecf44 memcg: oom kill d... |
1689 |
bool locked, need_to_kill; |
867578cbc memcg: fix oom ki... |
1690 |
|
c0ff4b854 memcg: rename mem... |
1691 |
owait.mem = memcg; |
dc98df5a1 memcg: oom wakeup... |
1692 1693 1694 1695 |
owait.wait.flags = 0; owait.wait.func = memcg_oom_wake_function; owait.wait.private = current; INIT_LIST_HEAD(&owait.wait.task_list); |
3c11ecf44 memcg: oom kill d... |
1696 |
need_to_kill = true; |
c0ff4b854 memcg: rename mem... |
1697 |
mem_cgroup_mark_under_oom(memcg); |
79dfdaccd memcg: make oom_l... |
1698 |
|
c0ff4b854 memcg: rename mem... |
1699 |
/* At first, try to OOM lock hierarchy under memcg.*/ |
1af8efe96 memcg: change mem... |
1700 |
spin_lock(&memcg_oom_lock); |
c0ff4b854 memcg: rename mem... |
1701 |
locked = mem_cgroup_oom_lock(memcg); |
867578cbc memcg: fix oom ki... |
1702 1703 1704 1705 1706 |
/* * Even if signal_pending(), we can't quit charge() loop without * accounting. So, UNINTERRUPTIBLE is appropriate. But SIGKILL * under OOM is always welcomed, use TASK_KILLABLE here. */ |
3c11ecf44 memcg: oom kill d... |
1707 |
prepare_to_wait(&memcg_oom_waitq, &owait.wait, TASK_KILLABLE); |
c0ff4b854 memcg: rename mem... |
1708 |
if (!locked || memcg->oom_kill_disable) |
3c11ecf44 memcg: oom kill d... |
1709 1710 |
need_to_kill = false; if (locked) |
c0ff4b854 memcg: rename mem... |
1711 |
mem_cgroup_oom_notify(memcg); |
1af8efe96 memcg: change mem... |
1712 |
spin_unlock(&memcg_oom_lock); |
867578cbc memcg: fix oom ki... |
1713 |
|
3c11ecf44 memcg: oom kill d... |
1714 1715 |
if (need_to_kill) { finish_wait(&memcg_oom_waitq, &owait.wait); |
c0ff4b854 memcg: rename mem... |
1716 |
mem_cgroup_out_of_memory(memcg, mask); |
3c11ecf44 memcg: oom kill d... |
1717 |
} else { |
867578cbc memcg: fix oom ki... |
1718 |
schedule(); |
dc98df5a1 memcg: oom wakeup... |
1719 |
finish_wait(&memcg_oom_waitq, &owait.wait); |
867578cbc memcg: fix oom ki... |
1720 |
} |
1af8efe96 memcg: change mem... |
1721 |
spin_lock(&memcg_oom_lock); |
79dfdaccd memcg: make oom_l... |
1722 |
if (locked) |
c0ff4b854 memcg: rename mem... |
1723 1724 |
mem_cgroup_oom_unlock(memcg); memcg_wakeup_oom(memcg); |
1af8efe96 memcg: change mem... |
1725 |
spin_unlock(&memcg_oom_lock); |
867578cbc memcg: fix oom ki... |
1726 |
|
c0ff4b854 memcg: rename mem... |
1727 |
mem_cgroup_unmark_under_oom(memcg); |
79dfdaccd memcg: make oom_l... |
1728 |
|
867578cbc memcg: fix oom ki... |
1729 1730 1731 |
if (test_thread_flag(TIF_MEMDIE) || fatal_signal_pending(current)) return false; /* Give chance to dying process */ |
715a5ee82 memcg: fix oom sc... |
1732 |
schedule_timeout_uninterruptible(1); |
867578cbc memcg: fix oom ki... |
1733 |
return true; |
0b7f569e4 memcg: fix OOM ki... |
1734 |
} |
d69b042f3 memcg: add file-b... |
1735 1736 1737 |
/* * Currently used to update mapped file statistics, but the routine can be * generalized to update other statistics as well. |
32047e2a8 memcg: avoid lock... |
1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 |
* * Notes: Race condition * * We usually use page_cgroup_lock() for accessing page_cgroup member but * it tends to be costly. But considering some conditions, we doesn't need * to do so _always_. * * Considering "charge", lock_page_cgroup() is not required because all * file-stat operations happen after a page is attached to radix-tree. There * are no race with "charge". * * Considering "uncharge", we know that memcg doesn't clear pc->mem_cgroup * at "uncharge" intentionally. So, we always see valid pc->mem_cgroup even * if there are race with "uncharge". Statistics itself is properly handled * by flags. * * Considering "move", this is an only case we see a race. To make the race * small, we check MEM_CGROUP_ON_MOVE percpu value and detect there are * possibility of race condition. If there is, we take a lock. |
d69b042f3 memcg: add file-b... |
1757 |
*/ |
26174efd4 memcg: generic fi... |
1758 |
|
2a7106f2c memcg: create ext... |
1759 1760 |
void mem_cgroup_update_page_stat(struct page *page, enum mem_cgroup_page_stat_item idx, int val) |
d69b042f3 memcg: add file-b... |
1761 |
{ |
c0ff4b854 memcg: rename mem... |
1762 |
struct mem_cgroup *memcg; |
32047e2a8 memcg: avoid lock... |
1763 1764 |
struct page_cgroup *pc = lookup_page_cgroup(page); bool need_unlock = false; |
dbd4ea78f memcg: add lock t... |
1765 |
unsigned long uninitialized_var(flags); |
d69b042f3 memcg: add file-b... |
1766 |
|
cfa449461 mm: memcg: lookup... |
1767 |
if (mem_cgroup_disabled()) |
d69b042f3 memcg: add file-b... |
1768 |
return; |
32047e2a8 memcg: avoid lock... |
1769 |
rcu_read_lock(); |
c0ff4b854 memcg: rename mem... |
1770 1771 |
memcg = pc->mem_cgroup; if (unlikely(!memcg || !PageCgroupUsed(pc))) |
32047e2a8 memcg: avoid lock... |
1772 1773 |
goto out; /* pc->mem_cgroup is unstable ? */ |
c0ff4b854 memcg: rename mem... |
1774 |
if (unlikely(mem_cgroup_stealed(memcg)) || PageTransHuge(page)) { |
32047e2a8 memcg: avoid lock... |
1775 |
/* take a lock against to access pc->mem_cgroup */ |
dbd4ea78f memcg: add lock t... |
1776 |
move_lock_page_cgroup(pc, &flags); |
32047e2a8 memcg: avoid lock... |
1777 |
need_unlock = true; |
c0ff4b854 memcg: rename mem... |
1778 1779 |
memcg = pc->mem_cgroup; if (!memcg || !PageCgroupUsed(pc)) |
32047e2a8 memcg: avoid lock... |
1780 1781 |
goto out; } |
26174efd4 memcg: generic fi... |
1782 |
|
26174efd4 memcg: generic fi... |
1783 |
switch (idx) { |
2a7106f2c memcg: create ext... |
1784 |
case MEMCG_NR_FILE_MAPPED: |
26174efd4 memcg: generic fi... |
1785 1786 1787 |
if (val > 0) SetPageCgroupFileMapped(pc); else if (!page_mapped(page)) |
0c270f8f9 memcg: fix race i... |
1788 |
ClearPageCgroupFileMapped(pc); |
2a7106f2c memcg: create ext... |
1789 |
idx = MEM_CGROUP_STAT_FILE_MAPPED; |
26174efd4 memcg: generic fi... |
1790 1791 1792 |
break; default: BUG(); |
8725d5416 memcg: fix race i... |
1793 |
} |
d69b042f3 memcg: add file-b... |
1794 |
|
c0ff4b854 memcg: rename mem... |
1795 |
this_cpu_add(memcg->stat->count[idx], val); |
2a7106f2c memcg: create ext... |
1796 |
|
32047e2a8 memcg: avoid lock... |
1797 1798 |
out: if (unlikely(need_unlock)) |
dbd4ea78f memcg: add lock t... |
1799 |
move_unlock_page_cgroup(pc, &flags); |
32047e2a8 memcg: avoid lock... |
1800 1801 |
rcu_read_unlock(); return; |
d69b042f3 memcg: add file-b... |
1802 |
} |
2a7106f2c memcg: create ext... |
1803 |
EXPORT_SYMBOL(mem_cgroup_update_page_stat); |
26174efd4 memcg: generic fi... |
1804 |
|
f817ed485 memcg: move all a... |
1805 |
/* |
cdec2e426 memcg: coalesce c... |
1806 1807 1808 |
* 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... |
1809 |
#define CHARGE_BATCH 32U |
cdec2e426 memcg: coalesce c... |
1810 1811 |
struct memcg_stock_pcp { struct mem_cgroup *cached; /* this never be root cgroup */ |
11c9ea4e8 memcg: convert pe... |
1812 |
unsigned int nr_pages; |
cdec2e426 memcg: coalesce c... |
1813 |
struct work_struct work; |
26fe61684 memcg: fix percpu... |
1814 1815 |
unsigned long flags; #define FLUSHING_CACHED_CHARGE (0) |
cdec2e426 memcg: coalesce c... |
1816 1817 |
}; static DEFINE_PER_CPU(struct memcg_stock_pcp, memcg_stock); |
9f50fad65 Revert "memcg: ge... |
1818 |
static DEFINE_MUTEX(percpu_charge_mutex); |
cdec2e426 memcg: coalesce c... |
1819 1820 |
/* |
11c9ea4e8 memcg: convert pe... |
1821 |
* Try to consume stocked charge on this cpu. If success, one page is consumed |
cdec2e426 memcg: coalesce c... |
1822 1823 1824 1825 |
* from local stock and true is returned. If the stock is 0 or charges from a * cgroup which is not current target, returns false. This stock will be * refilled. */ |
c0ff4b854 memcg: rename mem... |
1826 |
static bool consume_stock(struct mem_cgroup *memcg) |
cdec2e426 memcg: coalesce c... |
1827 1828 1829 1830 1831 |
{ struct memcg_stock_pcp *stock; bool ret = true; stock = &get_cpu_var(memcg_stock); |
c0ff4b854 memcg: rename mem... |
1832 |
if (memcg == stock->cached && stock->nr_pages) |
11c9ea4e8 memcg: convert pe... |
1833 |
stock->nr_pages--; |
cdec2e426 memcg: coalesce c... |
1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 |
else /* need to call res_counter_charge */ ret = false; put_cpu_var(memcg_stock); return ret; } /* * Returns stocks cached in percpu to res_counter and reset cached information. */ static void drain_stock(struct memcg_stock_pcp *stock) { struct mem_cgroup *old = stock->cached; |
11c9ea4e8 memcg: convert pe... |
1846 1847 1848 1849 |
if (stock->nr_pages) { unsigned long bytes = stock->nr_pages * PAGE_SIZE; res_counter_uncharge(&old->res, bytes); |
cdec2e426 memcg: coalesce c... |
1850 |
if (do_swap_account) |
11c9ea4e8 memcg: convert pe... |
1851 1852 |
res_counter_uncharge(&old->memsw, bytes); stock->nr_pages = 0; |
cdec2e426 memcg: coalesce c... |
1853 1854 |
} stock->cached = NULL; |
cdec2e426 memcg: coalesce c... |
1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 |
} /* * 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) { struct memcg_stock_pcp *stock = &__get_cpu_var(memcg_stock); drain_stock(stock); |
26fe61684 memcg: fix percpu... |
1865 |
clear_bit(FLUSHING_CACHED_CHARGE, &stock->flags); |
cdec2e426 memcg: coalesce c... |
1866 1867 1868 1869 |
} /* * Cache charges(val) which is from res_counter, to local per_cpu area. |
320cc51d9 mm: fix typo in r... |
1870 |
* This will be consumed by consume_stock() function, later. |
cdec2e426 memcg: coalesce c... |
1871 |
*/ |
c0ff4b854 memcg: rename mem... |
1872 |
static void refill_stock(struct mem_cgroup *memcg, unsigned int nr_pages) |
cdec2e426 memcg: coalesce c... |
1873 1874 |
{ struct memcg_stock_pcp *stock = &get_cpu_var(memcg_stock); |
c0ff4b854 memcg: rename mem... |
1875 |
if (stock->cached != memcg) { /* reset if necessary */ |
cdec2e426 memcg: coalesce c... |
1876 |
drain_stock(stock); |
c0ff4b854 memcg: rename mem... |
1877 |
stock->cached = memcg; |
cdec2e426 memcg: coalesce c... |
1878 |
} |
11c9ea4e8 memcg: convert pe... |
1879 |
stock->nr_pages += nr_pages; |
cdec2e426 memcg: coalesce c... |
1880 1881 1882 1883 |
put_cpu_var(memcg_stock); } /* |
c0ff4b854 memcg: rename mem... |
1884 |
* Drains all per-CPU charge caches for given root_memcg resp. subtree |
d38144b7a memcg: unify sync... |
1885 1886 |
* of the hierarchy under it. sync flag says whether we should block * until the work is done. |
cdec2e426 memcg: coalesce c... |
1887 |
*/ |
c0ff4b854 memcg: rename mem... |
1888 |
static void drain_all_stock(struct mem_cgroup *root_memcg, bool sync) |
cdec2e426 memcg: coalesce c... |
1889 |
{ |
26fe61684 memcg: fix percpu... |
1890 |
int cpu, curcpu; |
d38144b7a memcg: unify sync... |
1891 |
|
cdec2e426 memcg: coalesce c... |
1892 |
/* Notify other cpus that system-wide "drain" is running */ |
cdec2e426 memcg: coalesce c... |
1893 |
get_online_cpus(); |
5af12d0ef memcg: pin execut... |
1894 |
curcpu = get_cpu(); |
cdec2e426 memcg: coalesce c... |
1895 1896 |
for_each_online_cpu(cpu) { struct memcg_stock_pcp *stock = &per_cpu(memcg_stock, cpu); |
c0ff4b854 memcg: rename mem... |
1897 |
struct mem_cgroup *memcg; |
26fe61684 memcg: fix percpu... |
1898 |
|
c0ff4b854 memcg: rename mem... |
1899 1900 |
memcg = stock->cached; if (!memcg || !stock->nr_pages) |
26fe61684 memcg: fix percpu... |
1901 |
continue; |
c0ff4b854 memcg: rename mem... |
1902 |
if (!mem_cgroup_same_or_subtree(root_memcg, memcg)) |
3e92041d6 memcg: add mem_cg... |
1903 |
continue; |
d1a05b697 memcg: do not try... |
1904 1905 1906 1907 1908 1909 |
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... |
1910 |
} |
5af12d0ef memcg: pin execut... |
1911 |
put_cpu(); |
d38144b7a memcg: unify sync... |
1912 1913 1914 1915 1916 1917 |
if (!sync) goto out; for_each_online_cpu(cpu) { struct memcg_stock_pcp *stock = &per_cpu(memcg_stock, cpu); |
9f50fad65 Revert "memcg: ge... |
1918 |
if (test_bit(FLUSHING_CACHED_CHARGE, &stock->flags)) |
d38144b7a memcg: unify sync... |
1919 1920 1921 |
flush_work(&stock->work); } out: |
cdec2e426 memcg: coalesce c... |
1922 |
put_online_cpus(); |
d38144b7a memcg: unify sync... |
1923 1924 1925 1926 1927 1928 1929 1930 |
} /* * Tries to drain stocked charges in other cpus. This function is asynchronous * and just put a work per cpu for draining localy on each cpu. Caller can * expects some charges will be back to res_counter later but cannot wait for * it. */ |
c0ff4b854 memcg: rename mem... |
1931 |
static void drain_all_stock_async(struct mem_cgroup *root_memcg) |
d38144b7a memcg: unify sync... |
1932 |
{ |
9f50fad65 Revert "memcg: ge... |
1933 1934 1935 1936 1937 |
/* * If someone calls draining, avoid adding more kworker runs. */ if (!mutex_trylock(&percpu_charge_mutex)) return; |
c0ff4b854 memcg: rename mem... |
1938 |
drain_all_stock(root_memcg, false); |
9f50fad65 Revert "memcg: ge... |
1939 |
mutex_unlock(&percpu_charge_mutex); |
cdec2e426 memcg: coalesce c... |
1940 1941 1942 |
} /* This is a synchronous drain interface. */ |
c0ff4b854 memcg: rename mem... |
1943 |
static void drain_all_stock_sync(struct mem_cgroup *root_memcg) |
cdec2e426 memcg: coalesce c... |
1944 1945 |
{ /* called when force_empty is called */ |
9f50fad65 Revert "memcg: ge... |
1946 |
mutex_lock(&percpu_charge_mutex); |
c0ff4b854 memcg: rename mem... |
1947 |
drain_all_stock(root_memcg, true); |
9f50fad65 Revert "memcg: ge... |
1948 |
mutex_unlock(&percpu_charge_mutex); |
cdec2e426 memcg: coalesce c... |
1949 |
} |
711d3d2c9 memcg: cpu hotplu... |
1950 1951 1952 1953 |
/* * 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... |
1954 |
static void mem_cgroup_drain_pcp_counter(struct mem_cgroup *memcg, int cpu) |
711d3d2c9 memcg: cpu hotplu... |
1955 1956 |
{ int i; |
c0ff4b854 memcg: rename mem... |
1957 |
spin_lock(&memcg->pcp_counter_lock); |
711d3d2c9 memcg: cpu hotplu... |
1958 |
for (i = 0; i < MEM_CGROUP_STAT_DATA; i++) { |
c0ff4b854 memcg: rename mem... |
1959 |
long x = per_cpu(memcg->stat->count[i], cpu); |
711d3d2c9 memcg: cpu hotplu... |
1960 |
|
c0ff4b854 memcg: rename mem... |
1961 1962 |
per_cpu(memcg->stat->count[i], cpu) = 0; memcg->nocpu_base.count[i] += x; |
711d3d2c9 memcg: cpu hotplu... |
1963 |
} |
e9f8974f2 memcg: break out ... |
1964 |
for (i = 0; i < MEM_CGROUP_EVENTS_NSTATS; i++) { |
c0ff4b854 memcg: rename mem... |
1965 |
unsigned long x = per_cpu(memcg->stat->events[i], cpu); |
e9f8974f2 memcg: break out ... |
1966 |
|
c0ff4b854 memcg: rename mem... |
1967 1968 |
per_cpu(memcg->stat->events[i], cpu) = 0; memcg->nocpu_base.events[i] += x; |
e9f8974f2 memcg: break out ... |
1969 |
} |
1489ebad8 memcg: cpu hotplu... |
1970 |
/* need to clear ON_MOVE value, works as a kind of lock. */ |
c0ff4b854 memcg: rename mem... |
1971 1972 |
per_cpu(memcg->stat->count[MEM_CGROUP_ON_MOVE], cpu) = 0; spin_unlock(&memcg->pcp_counter_lock); |
1489ebad8 memcg: cpu hotplu... |
1973 |
} |
c0ff4b854 memcg: rename mem... |
1974 |
static void synchronize_mem_cgroup_on_move(struct mem_cgroup *memcg, int cpu) |
1489ebad8 memcg: cpu hotplu... |
1975 1976 |
{ int idx = MEM_CGROUP_ON_MOVE; |
c0ff4b854 memcg: rename mem... |
1977 1978 1979 |
spin_lock(&memcg->pcp_counter_lock); per_cpu(memcg->stat->count[idx], cpu) = memcg->nocpu_base.count[idx]; spin_unlock(&memcg->pcp_counter_lock); |
711d3d2c9 memcg: cpu hotplu... |
1980 1981 1982 |
} static int __cpuinit memcg_cpu_hotplug_callback(struct notifier_block *nb, |
cdec2e426 memcg: coalesce c... |
1983 1984 1985 1986 1987 |
unsigned long action, void *hcpu) { int cpu = (unsigned long)hcpu; struct memcg_stock_pcp *stock; |
711d3d2c9 memcg: cpu hotplu... |
1988 |
struct mem_cgroup *iter; |
cdec2e426 memcg: coalesce c... |
1989 |
|
1489ebad8 memcg: cpu hotplu... |
1990 |
if ((action == CPU_ONLINE)) { |
9f3a0d093 mm: memcg: consol... |
1991 |
for_each_mem_cgroup(iter) |
1489ebad8 memcg: cpu hotplu... |
1992 1993 1994 |
synchronize_mem_cgroup_on_move(iter, cpu); return NOTIFY_OK; } |
711d3d2c9 memcg: cpu hotplu... |
1995 |
if ((action != CPU_DEAD) || action != CPU_DEAD_FROZEN) |
cdec2e426 memcg: coalesce c... |
1996 |
return NOTIFY_OK; |
711d3d2c9 memcg: cpu hotplu... |
1997 |
|
9f3a0d093 mm: memcg: consol... |
1998 |
for_each_mem_cgroup(iter) |
711d3d2c9 memcg: cpu hotplu... |
1999 |
mem_cgroup_drain_pcp_counter(iter, cpu); |
cdec2e426 memcg: coalesce c... |
2000 2001 2002 2003 |
stock = &per_cpu(memcg_stock, cpu); drain_stock(stock); return NOTIFY_OK; } |
4b5343346 memcg: clean up t... |
2004 2005 2006 2007 2008 2009 2010 2011 2012 |
/* See __mem_cgroup_try_charge() for details */ enum { CHARGE_OK, /* success */ CHARGE_RETRY, /* need to retry but retry is not bad */ CHARGE_NOMEM, /* we can't do more. return -ENOMEM */ CHARGE_WOULDBLOCK, /* GFP_WAIT wasn't set and no enough res. */ CHARGE_OOM_DIE, /* the current is killed because of OOM */ }; |
c0ff4b854 memcg: rename mem... |
2013 |
static int mem_cgroup_do_charge(struct mem_cgroup *memcg, gfp_t gfp_mask, |
7ec99d621 memcg: unify char... |
2014 |
unsigned int nr_pages, bool oom_check) |
4b5343346 memcg: clean up t... |
2015 |
{ |
7ec99d621 memcg: unify char... |
2016 |
unsigned long csize = nr_pages * PAGE_SIZE; |
4b5343346 memcg: clean up t... |
2017 2018 2019 2020 |
struct mem_cgroup *mem_over_limit; struct res_counter *fail_res; unsigned long flags = 0; int ret; |
c0ff4b854 memcg: rename mem... |
2021 |
ret = res_counter_charge(&memcg->res, csize, &fail_res); |
4b5343346 memcg: clean up t... |
2022 2023 2024 2025 |
if (likely(!ret)) { if (!do_swap_account) return CHARGE_OK; |
c0ff4b854 memcg: rename mem... |
2026 |
ret = res_counter_charge(&memcg->memsw, csize, &fail_res); |
4b5343346 memcg: clean up t... |
2027 2028 |
if (likely(!ret)) return CHARGE_OK; |
c0ff4b854 memcg: rename mem... |
2029 |
res_counter_uncharge(&memcg->res, csize); |
4b5343346 memcg: clean up t... |
2030 2031 2032 2033 |
mem_over_limit = mem_cgroup_from_res_counter(fail_res, memsw); flags |= MEM_CGROUP_RECLAIM_NOSWAP; } else mem_over_limit = mem_cgroup_from_res_counter(fail_res, res); |
9221edb71 memcg: prevent en... |
2034 |
/* |
7ec99d621 memcg: unify char... |
2035 2036 |
* nr_pages can be either a huge page (HPAGE_PMD_NR), a batch * of regular pages (CHARGE_BATCH), or a single regular page (1). |
9221edb71 memcg: prevent en... |
2037 2038 2039 2040 |
* * Never reclaim on behalf of optional batching, retry with a * single page instead. */ |
7ec99d621 memcg: unify char... |
2041 |
if (nr_pages == CHARGE_BATCH) |
4b5343346 memcg: clean up t... |
2042 2043 2044 2045 |
return CHARGE_RETRY; if (!(gfp_mask & __GFP_WAIT)) return CHARGE_WOULDBLOCK; |
5660048cc mm: move memcg hi... |
2046 |
ret = mem_cgroup_reclaim(mem_over_limit, gfp_mask, flags); |
7ec99d621 memcg: unify char... |
2047 |
if (mem_cgroup_margin(mem_over_limit) >= nr_pages) |
19942822d memcg: prevent en... |
2048 |
return CHARGE_RETRY; |
4b5343346 memcg: clean up t... |
2049 |
/* |
19942822d memcg: prevent en... |
2050 2051 2052 2053 2054 2055 2056 |
* 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. |
4b5343346 memcg: clean up t... |
2057 |
*/ |
7ec99d621 memcg: unify char... |
2058 |
if (nr_pages == 1 && ret) |
4b5343346 memcg: clean up t... |
2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 |
return CHARGE_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)) return CHARGE_RETRY; /* If we don't need to call oom-killer at el, return immediately */ if (!oom_check) return CHARGE_NOMEM; /* check OOM */ if (!mem_cgroup_handle_oom(mem_over_limit, gfp_mask)) return CHARGE_OOM_DIE; return CHARGE_RETRY; } |
cdec2e426 memcg: coalesce c... |
2077 |
/* |
38c5d72f3 memcg: simplify L... |
2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 |
* __mem_cgroup_try_charge() does * 1. detect memcg to be charged against from passed *mm and *ptr, * 2. update res_counter * 3. call memory reclaim if necessary. * * In some special case, if the task is fatal, fatal_signal_pending() or * has TIF_MEMDIE, this function returns -EINTR while writing root_mem_cgroup * to *ptr. There are two reasons for this. 1: fatal threads should quit as soon * as possible without any hazards. 2: all pages should have a valid * pc->mem_cgroup. If mm is NULL and the caller doesn't pass a valid memcg * pointer, that is treated as a charge to root_mem_cgroup. * * So __mem_cgroup_try_charge() will return * 0 ... on success, filling *ptr with a valid memcg pointer. * -ENOMEM ... charge failure because of resource limits. * -EINTR ... if thread is fatal. *ptr is filled with root_mem_cgroup. * * Unlike the exported interface, an "oom" parameter is added. if oom==true, * the oom-killer can be invoked. |
8a9f3ccd2 Memory controller... |
2097 |
*/ |
f817ed485 memcg: move all a... |
2098 |
static int __mem_cgroup_try_charge(struct mm_struct *mm, |
ec1685109 thp: memcg compound |
2099 |
gfp_t gfp_mask, |
7ec99d621 memcg: unify char... |
2100 |
unsigned int nr_pages, |
c0ff4b854 memcg: rename mem... |
2101 |
struct mem_cgroup **ptr, |
7ec99d621 memcg: unify char... |
2102 |
bool oom) |
8a9f3ccd2 Memory controller... |
2103 |
{ |
7ec99d621 memcg: unify char... |
2104 |
unsigned int batch = max(CHARGE_BATCH, nr_pages); |
4b5343346 memcg: clean up t... |
2105 |
int nr_oom_retries = MEM_CGROUP_RECLAIM_RETRIES; |
c0ff4b854 memcg: rename mem... |
2106 |
struct mem_cgroup *memcg = NULL; |
4b5343346 memcg: clean up t... |
2107 |
int ret; |
a636b327f memcg: avoid unne... |
2108 |
|
867578cbc memcg: fix oom ki... |
2109 2110 2111 2112 2113 2114 2115 2116 |
/* * Unlike gloval-vm's OOM-kill, we're not in memory shortage * in system level. So, allow to go ahead dying process in addition to * MEMDIE process. */ if (unlikely(test_thread_flag(TIF_MEMDIE) || fatal_signal_pending(current))) goto bypass; |
a636b327f memcg: avoid unne... |
2117 |
|
8a9f3ccd2 Memory controller... |
2118 |
/* |
3be91277e memcgroup: tidy u... |
2119 2120 |
* We always charge the cgroup the mm_struct belongs to. * The mm_struct's mem_cgroup changes on task migration if the |
8a9f3ccd2 Memory controller... |
2121 2122 2123 |
* thread group leader migrates. It's possible that mm is not * set, if so charge the init_mm (happens for pagecache usage). */ |
c0ff4b854 memcg: rename mem... |
2124 |
if (!*ptr && !mm) |
38c5d72f3 memcg: simplify L... |
2125 |
*ptr = root_mem_cgroup; |
f75ca9620 memcg: avoid css_... |
2126 |
again: |
c0ff4b854 memcg: rename mem... |
2127 2128 2129 2130 |
if (*ptr) { /* css should be a valid one */ memcg = *ptr; VM_BUG_ON(css_is_removed(&memcg->css)); if (mem_cgroup_is_root(memcg)) |
f75ca9620 memcg: avoid css_... |
2131 |
goto done; |
c0ff4b854 memcg: rename mem... |
2132 |
if (nr_pages == 1 && consume_stock(memcg)) |
f75ca9620 memcg: avoid css_... |
2133 |
goto done; |
c0ff4b854 memcg: rename mem... |
2134 |
css_get(&memcg->css); |
4b5343346 memcg: clean up t... |
2135 |
} else { |
f75ca9620 memcg: avoid css_... |
2136 |
struct task_struct *p; |
54595fe26 memcg: use css_tr... |
2137 |
|
f75ca9620 memcg: avoid css_... |
2138 2139 |
rcu_read_lock(); p = rcu_dereference(mm->owner); |
f75ca9620 memcg: avoid css_... |
2140 |
/* |
ebb76ce16 memcg: fix wrong ... |
2141 |
* Because we don't have task_lock(), "p" can exit. |
c0ff4b854 memcg: rename mem... |
2142 |
* In that case, "memcg" can point to root or p can be NULL with |
ebb76ce16 memcg: fix wrong ... |
2143 2144 2145 2146 2147 2148 |
* race with swapoff. Then, we have small risk of mis-accouning. * But such kind of mis-account by race always happens because * we don't have cgroup_mutex(). It's overkill and we allo that * small race, here. * (*) swapoff at el will charge against mm-struct not against * task-struct. So, mm->owner can be NULL. |
f75ca9620 memcg: avoid css_... |
2149 |
*/ |
c0ff4b854 memcg: rename mem... |
2150 |
memcg = mem_cgroup_from_task(p); |
38c5d72f3 memcg: simplify L... |
2151 2152 2153 |
if (!memcg) memcg = root_mem_cgroup; if (mem_cgroup_is_root(memcg)) { |
f75ca9620 memcg: avoid css_... |
2154 2155 2156 |
rcu_read_unlock(); goto done; } |
c0ff4b854 memcg: rename mem... |
2157 |
if (nr_pages == 1 && consume_stock(memcg)) { |
f75ca9620 memcg: avoid css_... |
2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 |
/* * It seems dagerous to access memcg without css_get(). * But considering how consume_stok works, it's not * necessary. If consume_stock success, some charges * from this memcg are cached on this cpu. So, we * don't need to call css_get()/css_tryget() before * calling consume_stock(). */ rcu_read_unlock(); goto done; } /* after here, we may be blocked. we need to get refcnt */ |
c0ff4b854 memcg: rename mem... |
2170 |
if (!css_tryget(&memcg->css)) { |
f75ca9620 memcg: avoid css_... |
2171 2172 2173 2174 2175 |
rcu_read_unlock(); goto again; } rcu_read_unlock(); } |
8a9f3ccd2 Memory controller... |
2176 |
|
4b5343346 memcg: clean up t... |
2177 2178 |
do { bool oom_check; |
7a81b88cb memcg: introduce ... |
2179 |
|
4b5343346 memcg: clean up t... |
2180 |
/* If killed, bypass charge */ |
f75ca9620 memcg: avoid css_... |
2181 |
if (fatal_signal_pending(current)) { |
c0ff4b854 memcg: rename mem... |
2182 |
css_put(&memcg->css); |
4b5343346 memcg: clean up t... |
2183 |
goto bypass; |
f75ca9620 memcg: avoid css_... |
2184 |
} |
6d61ef409 memcg: memory cgr... |
2185 |
|
4b5343346 memcg: clean up t... |
2186 2187 2188 2189 |
oom_check = false; if (oom && !nr_oom_retries) { oom_check = true; nr_oom_retries = MEM_CGROUP_RECLAIM_RETRIES; |
cdec2e426 memcg: coalesce c... |
2190 |
} |
66e1707bc Memory controller... |
2191 |
|
c0ff4b854 memcg: rename mem... |
2192 |
ret = mem_cgroup_do_charge(memcg, gfp_mask, batch, oom_check); |
4b5343346 memcg: clean up t... |
2193 2194 2195 2196 |
switch (ret) { case CHARGE_OK: break; case CHARGE_RETRY: /* not in OOM situation but retry */ |
7ec99d621 memcg: unify char... |
2197 |
batch = nr_pages; |
c0ff4b854 memcg: rename mem... |
2198 2199 |
css_put(&memcg->css); memcg = NULL; |
f75ca9620 memcg: avoid css_... |
2200 |
goto again; |
4b5343346 memcg: clean up t... |
2201 |
case CHARGE_WOULDBLOCK: /* !__GFP_WAIT */ |
c0ff4b854 memcg: rename mem... |
2202 |
css_put(&memcg->css); |
4b5343346 memcg: clean up t... |
2203 2204 |
goto nomem; case CHARGE_NOMEM: /* OOM routine works */ |
f75ca9620 memcg: avoid css_... |
2205 |
if (!oom) { |
c0ff4b854 memcg: rename mem... |
2206 |
css_put(&memcg->css); |
867578cbc memcg: fix oom ki... |
2207 |
goto nomem; |
f75ca9620 memcg: avoid css_... |
2208 |
} |
4b5343346 memcg: clean up t... |
2209 2210 2211 2212 |
/* If oom, we never return -ENOMEM */ nr_oom_retries--; break; case CHARGE_OOM_DIE: /* Killed by OOM Killer */ |
c0ff4b854 memcg: rename mem... |
2213 |
css_put(&memcg->css); |
867578cbc memcg: fix oom ki... |
2214 |
goto bypass; |
66e1707bc Memory controller... |
2215 |
} |
4b5343346 memcg: clean up t... |
2216 |
} while (ret != CHARGE_OK); |
7ec99d621 memcg: unify char... |
2217 |
if (batch > nr_pages) |
c0ff4b854 memcg: rename mem... |
2218 2219 |
refill_stock(memcg, batch - nr_pages); css_put(&memcg->css); |
0c3e73e84 memcg: improve re... |
2220 |
done: |
c0ff4b854 memcg: rename mem... |
2221 |
*ptr = memcg; |
7a81b88cb memcg: introduce ... |
2222 2223 |
return 0; nomem: |
c0ff4b854 memcg: rename mem... |
2224 |
*ptr = NULL; |
7a81b88cb memcg: introduce ... |
2225 |
return -ENOMEM; |
867578cbc memcg: fix oom ki... |
2226 |
bypass: |
38c5d72f3 memcg: simplify L... |
2227 2228 |
*ptr = root_mem_cgroup; return -EINTR; |
7a81b88cb memcg: introduce ... |
2229 |
} |
8a9f3ccd2 Memory controller... |
2230 |
|
a3b2d6926 cgroups: use css ... |
2231 |
/* |
a3032a2c1 memcg: add mem_cg... |
2232 2233 2234 2235 |
* Somemtimes we have to undo a charge we got by try_charge(). * This function is for that and do uncharge, put css's refcnt. * gotten by try_charge(). */ |
c0ff4b854 memcg: rename mem... |
2236 |
static void __mem_cgroup_cancel_charge(struct mem_cgroup *memcg, |
e7018b8d2 memcg: keep only ... |
2237 |
unsigned int nr_pages) |
a3032a2c1 memcg: add mem_cg... |
2238 |
{ |
c0ff4b854 memcg: rename mem... |
2239 |
if (!mem_cgroup_is_root(memcg)) { |
e7018b8d2 memcg: keep only ... |
2240 |
unsigned long bytes = nr_pages * PAGE_SIZE; |
c0ff4b854 memcg: rename mem... |
2241 |
res_counter_uncharge(&memcg->res, bytes); |
a3032a2c1 memcg: add mem_cg... |
2242 |
if (do_swap_account) |
c0ff4b854 memcg: rename mem... |
2243 |
res_counter_uncharge(&memcg->memsw, bytes); |
a3032a2c1 memcg: add mem_cg... |
2244 |
} |
854ffa8d1 memcg: improve pe... |
2245 |
} |
a3032a2c1 memcg: add mem_cg... |
2246 |
/* |
a3b2d6926 cgroups: use css ... |
2247 2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 |
* A helper function to get mem_cgroup from ID. must be called under * rcu_read_lock(). The caller must check css_is_removed() or some if * it's concern. (dropping refcnt from swap can be called against removed * memcg.) */ static struct mem_cgroup *mem_cgroup_lookup(unsigned short id) { struct cgroup_subsys_state *css; /* ID 0 is unused ID */ if (!id) return NULL; css = css_lookup(&mem_cgroup_subsys, id); if (!css) return NULL; return container_of(css, struct mem_cgroup, css); } |
e42d9d5d4 memcg: rename and... |
2264 |
struct mem_cgroup *try_get_mem_cgroup_from_page(struct page *page) |
b5a84319a memcg: fix shmem'... |
2265 |
{ |
c0ff4b854 memcg: rename mem... |
2266 |
struct mem_cgroup *memcg = NULL; |
3c776e646 memcg: charge swa... |
2267 |
struct page_cgroup *pc; |
a3b2d6926 cgroups: use css ... |
2268 |
unsigned short id; |
b5a84319a memcg: fix shmem'... |
2269 |
swp_entry_t ent; |
3c776e646 memcg: charge swa... |
2270 |
VM_BUG_ON(!PageLocked(page)); |
3c776e646 memcg: charge swa... |
2271 |
pc = lookup_page_cgroup(page); |
c0bd3f63c memcg: fix try_ge... |
2272 |
lock_page_cgroup(pc); |
a3b2d6926 cgroups: use css ... |
2273 |
if (PageCgroupUsed(pc)) { |
c0ff4b854 memcg: rename mem... |
2274 2275 2276 |
memcg = pc->mem_cgroup; if (memcg && !css_tryget(&memcg->css)) memcg = NULL; |
e42d9d5d4 memcg: rename and... |
2277 |
} else if (PageSwapCache(page)) { |
3c776e646 memcg: charge swa... |
2278 |
ent.val = page_private(page); |
9fb4b7cc0 page_cgroup: add ... |
2279 |
id = lookup_swap_cgroup_id(ent); |
a3b2d6926 cgroups: use css ... |
2280 |
rcu_read_lock(); |
c0ff4b854 memcg: rename mem... |
2281 2282 2283 |
memcg = mem_cgroup_lookup(id); if (memcg && !css_tryget(&memcg->css)) memcg = NULL; |
a3b2d6926 cgroups: use css ... |
2284 |
rcu_read_unlock(); |
3c776e646 memcg: charge swa... |
2285 |
} |
c0bd3f63c memcg: fix try_ge... |
2286 |
unlock_page_cgroup(pc); |
c0ff4b854 memcg: rename mem... |
2287 |
return memcg; |
b5a84319a memcg: fix shmem'... |
2288 |
} |
c0ff4b854 memcg: rename mem... |
2289 |
static void __mem_cgroup_commit_charge(struct mem_cgroup *memcg, |
5564e88ba memcg: condense p... |
2290 |
struct page *page, |
7ec99d621 memcg: unify char... |
2291 |
unsigned int nr_pages, |
ca3e02141 memcg: fix USED b... |
2292 |
struct page_cgroup *pc, |
7ec99d621 memcg: unify char... |
2293 |
enum charge_type ctype) |
7a81b88cb memcg: introduce ... |
2294 |
{ |
ca3e02141 memcg: fix USED b... |
2295 2296 2297 |
lock_page_cgroup(pc); if (unlikely(PageCgroupUsed(pc))) { unlock_page_cgroup(pc); |
c0ff4b854 memcg: rename mem... |
2298 |
__mem_cgroup_cancel_charge(memcg, nr_pages); |
ca3e02141 memcg: fix USED b... |
2299 2300 2301 2302 2303 2304 |
return; } /* * we don't need page_cgroup_lock about tail pages, becase they are not * accessed by any other context at this point. */ |
c0ff4b854 memcg: rename mem... |
2305 |
pc->mem_cgroup = memcg; |
261fb61a8 memcg: add commen... |
2306 2307 2308 2309 2310 2311 2312 |
/* * We access a page_cgroup asynchronously without lock_page_cgroup(). * Especially when a page_cgroup is taken from a page, pc->mem_cgroup * is accessed after testing USED bit. To make pc->mem_cgroup visible * before USED bit, we need memory barrier here. * See mem_cgroup_add_lru_list(), etc. */ |
08e552c69 memcg: synchroniz... |
2313 |
smp_wmb(); |
4b3bde4c9 memcg: remove the... |
2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 |
switch (ctype) { case MEM_CGROUP_CHARGE_TYPE_CACHE: case MEM_CGROUP_CHARGE_TYPE_SHMEM: SetPageCgroupCache(pc); SetPageCgroupUsed(pc); break; case MEM_CGROUP_CHARGE_TYPE_MAPPED: ClearPageCgroupCache(pc); SetPageCgroupUsed(pc); break; default: break; } |
3be91277e memcgroup: tidy u... |
2327 |
|
c0ff4b854 memcg: rename mem... |
2328 |
mem_cgroup_charge_statistics(memcg, PageCgroupCache(pc), nr_pages); |
52d4b9ac0 memcg: allocate a... |
2329 |
unlock_page_cgroup(pc); |
38c5d72f3 memcg: simplify L... |
2330 |
WARN_ON_ONCE(PageLRU(page)); |
430e48631 memcg: update thr... |
2331 2332 2333 2334 2335 |
/* * "charge_statistics" updated event counter. Then, check it. * Insert ancestor (and ancestor's ancestors), to softlimit RB-tree. * if they exceeds softlimit. */ |
c0ff4b854 memcg: rename mem... |
2336 |
memcg_check_events(memcg, page); |
7a81b88cb memcg: introduce ... |
2337 |
} |
66e1707bc Memory controller... |
2338 |
|
ca3e02141 memcg: fix USED b... |
2339 2340 2341 |
#ifdef CONFIG_TRANSPARENT_HUGEPAGE #define PCGF_NOCOPY_AT_SPLIT ((1 << PCG_LOCK) | (1 << PCG_MOVE_LOCK) |\ |
38c5d72f3 memcg: simplify L... |
2342 |
(1 << PCG_MIGRATION)) |
ca3e02141 memcg: fix USED b... |
2343 2344 |
/* * Because tail pages are not marked as "used", set it. We're under |
e94c8a9cb memcg: make mem_c... |
2345 2346 2347 |
* 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... |
2348 |
*/ |
e94c8a9cb memcg: make mem_c... |
2349 |
void mem_cgroup_split_huge_fixup(struct page *head) |
ca3e02141 memcg: fix USED b... |
2350 2351 |
{ struct page_cgroup *head_pc = lookup_page_cgroup(head); |
e94c8a9cb memcg: make mem_c... |
2352 2353 |
struct page_cgroup *pc; int i; |
ca3e02141 memcg: fix USED b... |
2354 |
|
3d37c4a91 memcg: bugfix che... |
2355 2356 |
if (mem_cgroup_disabled()) return; |
e94c8a9cb memcg: make mem_c... |
2357 2358 2359 2360 |
for (i = 1; i < HPAGE_PMD_NR; i++) { pc = head_pc + i; pc->mem_cgroup = head_pc->mem_cgroup; smp_wmb();/* see __commit_charge() */ |
e94c8a9cb memcg: make mem_c... |
2361 2362 |
pc->flags = head_pc->flags & ~PCGF_NOCOPY_AT_SPLIT; } |
ca3e02141 memcg: fix USED b... |
2363 |
} |
12d271078 memcg: fix split_... |
2364 |
#endif /* CONFIG_TRANSPARENT_HUGEPAGE */ |
ca3e02141 memcg: fix USED b... |
2365 |
|
f817ed485 memcg: move all a... |
2366 |
/** |
de3638d9c memcg: fold __mem... |
2367 |
* mem_cgroup_move_account - move account of the page |
5564e88ba memcg: condense p... |
2368 |
* @page: the page |
7ec99d621 memcg: unify char... |
2369 |
* @nr_pages: number of regular pages (>1 for huge pages) |
f817ed485 memcg: move all a... |
2370 2371 2372 |
* @pc: page_cgroup of the page. * @from: mem_cgroup which the page is moved from. * @to: mem_cgroup which the page is moved to. @from != @to. |
854ffa8d1 memcg: improve pe... |
2373 |
* @uncharge: whether we should call uncharge and css_put against @from. |
f817ed485 memcg: move all a... |
2374 2375 |
* * The caller must confirm following. |
08e552c69 memcg: synchroniz... |
2376 |
* - page is not on LRU (isolate_page() is useful.) |
7ec99d621 memcg: unify char... |
2377 |
* - compound_lock is held when nr_pages > 1 |
f817ed485 memcg: move all a... |
2378 |
* |
854ffa8d1 memcg: improve pe... |
2379 |
* This function doesn't do "charge" nor css_get to new cgroup. It should be |
25985edce Fix common misspe... |
2380 |
* done by a caller(__mem_cgroup_try_charge would be useful). If @uncharge is |
854ffa8d1 memcg: improve pe... |
2381 2382 |
* true, this function does "uncharge" from old cgroup, but it doesn't if * @uncharge is false, so a caller should do "uncharge". |
f817ed485 memcg: move all a... |
2383 |
*/ |
7ec99d621 memcg: unify char... |
2384 2385 2386 2387 2388 2389 |
static int mem_cgroup_move_account(struct page *page, unsigned int nr_pages, struct page_cgroup *pc, struct mem_cgroup *from, struct mem_cgroup *to, bool uncharge) |
f817ed485 memcg: move all a... |
2390 |
{ |
de3638d9c memcg: fold __mem... |
2391 2392 |
unsigned long flags; int ret; |
987eba66e memcg: fix rmdir,... |
2393 |
|
f817ed485 memcg: move all a... |
2394 |
VM_BUG_ON(from == to); |
5564e88ba memcg: condense p... |
2395 |
VM_BUG_ON(PageLRU(page)); |
de3638d9c memcg: fold __mem... |
2396 2397 2398 2399 2400 2401 2402 |
/* * 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; |
7ec99d621 memcg: unify char... |
2403 |
if (nr_pages > 1 && !PageTransHuge(page)) |
de3638d9c memcg: fold __mem... |
2404 2405 2406 2407 2408 2409 2410 2411 2412 |
goto out; lock_page_cgroup(pc); ret = -EINVAL; if (!PageCgroupUsed(pc) || pc->mem_cgroup != from) goto unlock; move_lock_page_cgroup(pc, &flags); |
f817ed485 memcg: move all a... |
2413 |
|
8725d5416 memcg: fix race i... |
2414 |
if (PageCgroupFileMapped(pc)) { |
c62b1a3b3 memcg: use generi... |
2415 2416 2417 2418 2419 |
/* Update mapped_file data for mem_cgroup */ preempt_disable(); __this_cpu_dec(from->stat->count[MEM_CGROUP_STAT_FILE_MAPPED]); __this_cpu_inc(to->stat->count[MEM_CGROUP_STAT_FILE_MAPPED]); preempt_enable(); |
d69b042f3 memcg: add file-b... |
2420 |
} |
987eba66e memcg: fix rmdir,... |
2421 |
mem_cgroup_charge_statistics(from, PageCgroupCache(pc), -nr_pages); |
854ffa8d1 memcg: improve pe... |
2422 2423 |
if (uncharge) /* This is not "cancel", but cancel_charge does all we need. */ |
e7018b8d2 memcg: keep only ... |
2424 |
__mem_cgroup_cancel_charge(from, nr_pages); |
d69b042f3 memcg: add file-b... |
2425 |
|
854ffa8d1 memcg: improve pe... |
2426 |
/* caller should have done css_get */ |
08e552c69 memcg: synchroniz... |
2427 |
pc->mem_cgroup = to; |
987eba66e memcg: fix rmdir,... |
2428 |
mem_cgroup_charge_statistics(to, PageCgroupCache(pc), nr_pages); |
887032670 cgroup avoid perm... |
2429 2430 2431 |
/* * We charges against "to" which may not have any tasks. Then, "to" * can be under rmdir(). But in current implementation, caller of |
4ffef5fef memcg: move charg... |
2432 |
* this function is just force_empty() and move charge, so it's |
25985edce Fix common misspe... |
2433 |
* guaranteed that "to" is never removed. So, we don't check rmdir |
4ffef5fef memcg: move charg... |
2434 |
* status here. |
887032670 cgroup avoid perm... |
2435 |
*/ |
de3638d9c memcg: fold __mem... |
2436 2437 2438 |
move_unlock_page_cgroup(pc, &flags); ret = 0; unlock: |
57f9fd7d2 memcg: cleanup me... |
2439 |
unlock_page_cgroup(pc); |
d2265e6fa memcg : share eve... |
2440 2441 2442 |
/* * check events */ |
5564e88ba memcg: condense p... |
2443 2444 |
memcg_check_events(to, page); memcg_check_events(from, page); |
de3638d9c memcg: fold __mem... |
2445 |
out: |
f817ed485 memcg: move all a... |
2446 2447 2448 2449 2450 2451 |
return ret; } /* * move charges to its parent. */ |
5564e88ba memcg: condense p... |
2452 2453 |
static int mem_cgroup_move_parent(struct page *page, struct page_cgroup *pc, |
f817ed485 memcg: move all a... |
2454 2455 2456 2457 2458 2459 |
struct mem_cgroup *child, gfp_t gfp_mask) { struct cgroup *cg = child->css.cgroup; struct cgroup *pcg = cg->parent; struct mem_cgroup *parent; |
7ec99d621 memcg: unify char... |
2460 |
unsigned int nr_pages; |
4be4489fe mm/memcontrol.c: ... |
2461 |
unsigned long uninitialized_var(flags); |
f817ed485 memcg: move all a... |
2462 2463 2464 2465 2466 |
int ret; /* Is ROOT ? */ if (!pcg) return -EINVAL; |
57f9fd7d2 memcg: cleanup me... |
2467 2468 2469 2470 2471 |
ret = -EBUSY; if (!get_page_unless_zero(page)) goto out; if (isolate_lru_page(page)) goto put; |
52dbb9050 memcg: fix race a... |
2472 |
|
7ec99d621 memcg: unify char... |
2473 |
nr_pages = hpage_nr_pages(page); |
08e552c69 memcg: synchroniz... |
2474 |
|
f817ed485 memcg: move all a... |
2475 |
parent = mem_cgroup_from_cont(pcg); |
7ec99d621 memcg: unify char... |
2476 |
ret = __mem_cgroup_try_charge(NULL, gfp_mask, nr_pages, &parent, false); |
38c5d72f3 memcg: simplify L... |
2477 |
if (ret) |
57f9fd7d2 memcg: cleanup me... |
2478 |
goto put_back; |
f817ed485 memcg: move all a... |
2479 |
|
7ec99d621 memcg: unify char... |
2480 |
if (nr_pages > 1) |
987eba66e memcg: fix rmdir,... |
2481 |
flags = compound_lock_irqsave(page); |
7ec99d621 memcg: unify char... |
2482 |
ret = mem_cgroup_move_account(page, nr_pages, pc, child, parent, true); |
854ffa8d1 memcg: improve pe... |
2483 |
if (ret) |
7ec99d621 memcg: unify char... |
2484 |
__mem_cgroup_cancel_charge(parent, nr_pages); |
8dba474f0 mm/memcontrol.c: ... |
2485 |
|
7ec99d621 memcg: unify char... |
2486 |
if (nr_pages > 1) |
987eba66e memcg: fix rmdir,... |
2487 |
compound_unlock_irqrestore(page, flags); |
8dba474f0 mm/memcontrol.c: ... |
2488 |
put_back: |
08e552c69 memcg: synchroniz... |
2489 |
putback_lru_page(page); |
57f9fd7d2 memcg: cleanup me... |
2490 |
put: |
40d58138f memcg: fix error ... |
2491 |
put_page(page); |
57f9fd7d2 memcg: cleanup me... |
2492 |
out: |
f817ed485 memcg: move all a... |
2493 2494 |
return ret; } |
7a81b88cb memcg: introduce ... |
2495 2496 2497 2498 2499 2500 2501 |
/* * Charge the memory controller for page usage. * Return * 0 if the charge was successful * < 0 if the cgroup is over its limit */ static int mem_cgroup_charge_common(struct page *page, struct mm_struct *mm, |
73045c47b memcg: remove mem... |
2502 |
gfp_t gfp_mask, enum charge_type ctype) |
7a81b88cb memcg: introduce ... |
2503 |
{ |
c0ff4b854 memcg: rename mem... |
2504 |
struct mem_cgroup *memcg = NULL; |
7ec99d621 memcg: unify char... |
2505 |
unsigned int nr_pages = 1; |
7a81b88cb memcg: introduce ... |
2506 |
struct page_cgroup *pc; |
8493ae439 memcg: never OOM ... |
2507 |
bool oom = true; |
7a81b88cb memcg: introduce ... |
2508 |
int ret; |
ec1685109 thp: memcg compound |
2509 |
|
37c2ac787 thp: compound_tra... |
2510 |
if (PageTransHuge(page)) { |
7ec99d621 memcg: unify char... |
2511 |
nr_pages <<= compound_order(page); |
37c2ac787 thp: compound_tra... |
2512 |
VM_BUG_ON(!PageTransHuge(page)); |
8493ae439 memcg: never OOM ... |
2513 2514 2515 2516 2517 |
/* * Never OOM-kill a process for a huge page. The * fault handler will fall back to regular pages. */ oom = false; |
37c2ac787 thp: compound_tra... |
2518 |
} |
7a81b88cb memcg: introduce ... |
2519 2520 |
pc = lookup_page_cgroup(page); |
c0ff4b854 memcg: rename mem... |
2521 |
ret = __mem_cgroup_try_charge(mm, gfp_mask, nr_pages, &memcg, oom); |
38c5d72f3 memcg: simplify L... |
2522 |
if (ret == -ENOMEM) |
7a81b88cb memcg: introduce ... |
2523 |
return ret; |
c0ff4b854 memcg: rename mem... |
2524 |
__mem_cgroup_commit_charge(memcg, page, nr_pages, pc, ctype); |
8a9f3ccd2 Memory controller... |
2525 |
return 0; |
8a9f3ccd2 Memory controller... |
2526 |
} |
7a81b88cb memcg: introduce ... |
2527 2528 |
int mem_cgroup_newpage_charge(struct page *page, struct mm_struct *mm, gfp_t gfp_mask) |
217bc3194 memory cgroup enh... |
2529 |
{ |
f8d665422 memcg: add mem_cg... |
2530 |
if (mem_cgroup_disabled()) |
cede86acd memcg: clean up c... |
2531 |
return 0; |
7a0524cfc mm: memcg: remove... |
2532 2533 2534 |
VM_BUG_ON(page_mapped(page)); VM_BUG_ON(page->mapping && !PageAnon(page)); VM_BUG_ON(!mm); |
217bc3194 memory cgroup enh... |
2535 |
return mem_cgroup_charge_common(page, mm, gfp_mask, |
7a0524cfc mm: memcg: remove... |
2536 |
MEM_CGROUP_CHARGE_TYPE_MAPPED); |
217bc3194 memory cgroup enh... |
2537 |
} |
83aae4c73 memcg: cleanup ca... |
2538 2539 2540 |
static void __mem_cgroup_commit_charge_swapin(struct page *page, struct mem_cgroup *ptr, enum charge_type ctype); |
5a6475a4e memcg: fix leak o... |
2541 |
static void |
c0ff4b854 memcg: rename mem... |
2542 |
__mem_cgroup_commit_charge_lrucare(struct page *page, struct mem_cgroup *memcg, |
5a6475a4e memcg: fix leak o... |
2543 2544 2545 |
enum charge_type ctype) { struct page_cgroup *pc = lookup_page_cgroup(page); |
36b62ad53 memcg: simplify c... |
2546 2547 2548 |
struct zone *zone = page_zone(page); unsigned long flags; bool removed = false; |
5a6475a4e memcg: fix leak o... |
2549 2550 2551 2552 2553 |
/* * In some case, SwapCache, FUSE(splice_buf->radixtree), the page * is already on LRU. It means the page may on some other page_cgroup's * LRU. Take care of it. */ |
36b62ad53 memcg: simplify c... |
2554 2555 2556 2557 2558 2559 |
spin_lock_irqsave(&zone->lru_lock, flags); if (PageLRU(page)) { del_page_from_lru_list(zone, page, page_lru(page)); ClearPageLRU(page); removed = true; } |
c0ff4b854 memcg: rename mem... |
2560 |
__mem_cgroup_commit_charge(memcg, page, 1, pc, ctype); |
36b62ad53 memcg: simplify c... |
2561 2562 2563 2564 2565 |
if (removed) { add_page_to_lru_list(zone, page, page_lru(page)); SetPageLRU(page); } spin_unlock_irqrestore(&zone->lru_lock, flags); |
5a6475a4e memcg: fix leak o... |
2566 2567 |
return; } |
e1a1cd590 Memory controller... |
2568 2569 |
int mem_cgroup_cache_charge(struct page *page, struct mm_struct *mm, gfp_t gfp_mask) |
8697d3319 Memory controller... |
2570 |
{ |
c0ff4b854 memcg: rename mem... |
2571 |
struct mem_cgroup *memcg = NULL; |
dc67d5046 memcg: simplify p... |
2572 |
enum charge_type type = MEM_CGROUP_CHARGE_TYPE_CACHE; |
b5a84319a memcg: fix shmem'... |
2573 |
int ret; |
f8d665422 memcg: add mem_cg... |
2574 |
if (mem_cgroup_disabled()) |
cede86acd memcg: clean up c... |
2575 |
return 0; |
52d4b9ac0 memcg: allocate a... |
2576 2577 |
if (PageCompound(page)) return 0; |
accf163e6 memcg: remove a r... |
2578 |
|
73045c47b memcg: remove mem... |
2579 |
if (unlikely(!mm)) |
8697d3319 Memory controller... |
2580 |
mm = &init_mm; |
dc67d5046 memcg: simplify p... |
2581 2582 |
if (!page_is_file_cache(page)) type = MEM_CGROUP_CHARGE_TYPE_SHMEM; |
accf163e6 memcg: remove a r... |
2583 |
|
38c5d72f3 memcg: simplify L... |
2584 |
if (!PageSwapCache(page)) |
dc67d5046 memcg: simplify p... |
2585 |
ret = mem_cgroup_charge_common(page, mm, gfp_mask, type); |
38c5d72f3 memcg: simplify L... |
2586 |
else { /* page is swapcache/shmem */ |
c0ff4b854 memcg: rename mem... |
2587 |
ret = mem_cgroup_try_charge_swapin(mm, page, gfp_mask, &memcg); |
83aae4c73 memcg: cleanup ca... |
2588 |
if (!ret) |
dc67d5046 memcg: simplify p... |
2589 2590 |
__mem_cgroup_commit_charge_swapin(page, memcg, type); } |
b5a84319a memcg: fix shmem'... |
2591 |
return ret; |
e8589cc18 memcg: better mig... |
2592 |
} |
54595fe26 memcg: use css_tr... |
2593 2594 2595 |
/* * While swap-in, try_charge -> commit or cancel, the page is locked. * And when try_charge() successfully returns, one refcnt to memcg without |
21ae2956c tree-wide: fix ty... |
2596 |
* struct page_cgroup is acquired. This refcnt will be consumed by |
54595fe26 memcg: use css_tr... |
2597 2598 |
* "commit()" or removed by "cancel()" */ |
8c7c6e34a memcg: mem+swap c... |
2599 2600 |
int mem_cgroup_try_charge_swapin(struct mm_struct *mm, struct page *page, |
72835c86c mm: unify remaini... |
2601 |
gfp_t mask, struct mem_cgroup **memcgp) |
8c7c6e34a memcg: mem+swap c... |
2602 |
{ |
c0ff4b854 memcg: rename mem... |
2603 |
struct mem_cgroup *memcg; |
54595fe26 memcg: use css_tr... |
2604 |
int ret; |
8c7c6e34a memcg: mem+swap c... |
2605 |
|
72835c86c mm: unify remaini... |
2606 |
*memcgp = NULL; |
56039efa1 memcg: fix ugly i... |
2607 |
|
f8d665422 memcg: add mem_cg... |
2608 |
if (mem_cgroup_disabled()) |
8c7c6e34a memcg: mem+swap c... |
2609 2610 2611 2612 |
return 0; if (!do_swap_account) goto charge_cur_mm; |
8c7c6e34a memcg: mem+swap c... |
2613 2614 |
/* * A racing thread's fault, or swapoff, may have already updated |
407f9c8b0 ksm: mem cgroup c... |
2615 2616 2617 |
* the pte, and even removed page from swap cache: in those cases * do_swap_page()'s pte_same() test will fail; but there's also a * KSM case which does need to charge the page. |
8c7c6e34a memcg: mem+swap c... |
2618 2619 |
*/ if (!PageSwapCache(page)) |
407f9c8b0 ksm: mem cgroup c... |
2620 |
goto charge_cur_mm; |
c0ff4b854 memcg: rename mem... |
2621 2622 |
memcg = try_get_mem_cgroup_from_page(page); if (!memcg) |
54595fe26 memcg: use css_tr... |
2623 |
goto charge_cur_mm; |
72835c86c mm: unify remaini... |
2624 2625 |
*memcgp = memcg; ret = __mem_cgroup_try_charge(NULL, mask, 1, memcgp, true); |
c0ff4b854 memcg: rename mem... |
2626 |
css_put(&memcg->css); |
38c5d72f3 memcg: simplify L... |
2627 2628 |
if (ret == -EINTR) ret = 0; |
54595fe26 memcg: use css_tr... |
2629 |
return ret; |
8c7c6e34a memcg: mem+swap c... |
2630 2631 2632 |
charge_cur_mm: if (unlikely(!mm)) mm = &init_mm; |
38c5d72f3 memcg: simplify L... |
2633 2634 2635 2636 |
ret = __mem_cgroup_try_charge(mm, mask, 1, memcgp, true); if (ret == -EINTR) ret = 0; return ret; |
8c7c6e34a memcg: mem+swap c... |
2637 |
} |
83aae4c73 memcg: cleanup ca... |
2638 |
static void |
72835c86c mm: unify remaini... |
2639 |
__mem_cgroup_commit_charge_swapin(struct page *page, struct mem_cgroup *memcg, |
83aae4c73 memcg: cleanup ca... |
2640 |
enum charge_type ctype) |
7a81b88cb memcg: introduce ... |
2641 |
{ |
f8d665422 memcg: add mem_cg... |
2642 |
if (mem_cgroup_disabled()) |
7a81b88cb memcg: introduce ... |
2643 |
return; |
72835c86c mm: unify remaini... |
2644 |
if (!memcg) |
7a81b88cb memcg: introduce ... |
2645 |
return; |
72835c86c mm: unify remaini... |
2646 |
cgroup_exclude_rmdir(&memcg->css); |
5a6475a4e memcg: fix leak o... |
2647 |
|
72835c86c mm: unify remaini... |
2648 |
__mem_cgroup_commit_charge_lrucare(page, memcg, ctype); |
8c7c6e34a memcg: mem+swap c... |
2649 2650 2651 |
/* * Now swap is on-memory. This means this page may be * counted both as mem and swap....double count. |
03f3c4336 memcg: fix swap a... |
2652 2653 2654 |
* Fix it by uncharging from memsw. Basically, this SwapCache is stable * under lock_page(). But in do_swap_page()::memory.c, reuse_swap_page() * may call delete_from_swap_cache() before reach here. |
8c7c6e34a memcg: mem+swap c... |
2655 |
*/ |
03f3c4336 memcg: fix swap a... |
2656 |
if (do_swap_account && PageSwapCache(page)) { |
8c7c6e34a memcg: mem+swap c... |
2657 |
swp_entry_t ent = {.val = page_private(page)}; |
72835c86c mm: unify remaini... |
2658 |
struct mem_cgroup *swap_memcg; |
a3b2d6926 cgroups: use css ... |
2659 |
unsigned short id; |
a3b2d6926 cgroups: use css ... |
2660 2661 2662 |
id = swap_cgroup_record(ent, 0); rcu_read_lock(); |
72835c86c mm: unify remaini... |
2663 2664 |
swap_memcg = mem_cgroup_lookup(id); if (swap_memcg) { |
a3b2d6926 cgroups: use css ... |
2665 2666 2667 2668 |
/* * This recorded memcg can be obsolete one. So, avoid * calling css_tryget */ |
72835c86c mm: unify remaini... |
2669 2670 2671 2672 2673 |
if (!mem_cgroup_is_root(swap_memcg)) res_counter_uncharge(&swap_memcg->memsw, PAGE_SIZE); mem_cgroup_swap_statistics(swap_memcg, false); mem_cgroup_put(swap_memcg); |
8c7c6e34a memcg: mem+swap c... |
2674 |
} |
a3b2d6926 cgroups: use css ... |
2675 |
rcu_read_unlock(); |
8c7c6e34a memcg: mem+swap c... |
2676 |
} |
887032670 cgroup avoid perm... |
2677 2678 2679 2680 2681 |
/* * At swapin, we may charge account against cgroup which has no tasks. * So, rmdir()->pre_destroy() can be called while we do this charge. * In that case, we need to call pre_destroy() again. check it here. */ |
72835c86c mm: unify remaini... |
2682 |
cgroup_release_and_wakeup_rmdir(&memcg->css); |
7a81b88cb memcg: introduce ... |
2683 |
} |
72835c86c mm: unify remaini... |
2684 2685 |
void mem_cgroup_commit_charge_swapin(struct page *page, struct mem_cgroup *memcg) |
83aae4c73 memcg: cleanup ca... |
2686 |
{ |
72835c86c mm: unify remaini... |
2687 2688 |
__mem_cgroup_commit_charge_swapin(page, memcg, MEM_CGROUP_CHARGE_TYPE_MAPPED); |
83aae4c73 memcg: cleanup ca... |
2689 |
} |
c0ff4b854 memcg: rename mem... |
2690 |
void mem_cgroup_cancel_charge_swapin(struct mem_cgroup *memcg) |
7a81b88cb memcg: introduce ... |
2691 |
{ |
f8d665422 memcg: add mem_cg... |
2692 |
if (mem_cgroup_disabled()) |
7a81b88cb memcg: introduce ... |
2693 |
return; |
c0ff4b854 memcg: rename mem... |
2694 |
if (!memcg) |
7a81b88cb memcg: introduce ... |
2695 |
return; |
c0ff4b854 memcg: rename mem... |
2696 |
__mem_cgroup_cancel_charge(memcg, 1); |
7a81b88cb memcg: introduce ... |
2697 |
} |
c0ff4b854 memcg: rename mem... |
2698 |
static void mem_cgroup_do_uncharge(struct mem_cgroup *memcg, |
7ec99d621 memcg: unify char... |
2699 2700 |
unsigned int nr_pages, const enum charge_type ctype) |
569b846df memcg: coalesce u... |
2701 2702 2703 |
{ struct memcg_batch_info *batch = NULL; bool uncharge_memsw = true; |
7ec99d621 memcg: unify char... |
2704 |
|
569b846df memcg: coalesce u... |
2705 2706 2707 |
/* If swapout, usage of swap doesn't decrease */ if (!do_swap_account || ctype == MEM_CGROUP_CHARGE_TYPE_SWAPOUT) uncharge_memsw = false; |
569b846df memcg: coalesce u... |
2708 2709 2710 2711 2712 2713 2714 2715 |
batch = ¤t->memcg_batch; /* * In usual, we do css_get() when we remember memcg pointer. * But in this case, we keep res->usage until end of a series of * uncharges. Then, it's ok to ignore memcg's refcnt. */ if (!batch->memcg) |
c0ff4b854 memcg: rename mem... |
2716 |
batch->memcg = memcg; |
569b846df memcg: coalesce u... |
2717 |
/* |
3c11ecf44 memcg: oom kill d... |
2718 |
* do_batch > 0 when unmapping pages or inode invalidate/truncate. |
25985edce Fix common misspe... |
2719 |
* In those cases, all pages freed continuously can be expected to be in |
3c11ecf44 memcg: oom kill d... |
2720 2721 2722 2723 2724 2725 2726 |
* the same cgroup and we have chance to coalesce uncharges. * But we do uncharge one by one if this is killed by OOM(TIF_MEMDIE) * because we want to do uncharge as soon as possible. */ if (!batch->do_batch || test_thread_flag(TIF_MEMDIE)) goto direct_uncharge; |
7ec99d621 memcg: unify char... |
2727 |
if (nr_pages > 1) |
ec1685109 thp: memcg compound |
2728 |
goto direct_uncharge; |
3c11ecf44 memcg: oom kill d... |
2729 |
/* |
569b846df memcg: coalesce u... |
2730 2731 2732 2733 |
* In typical case, batch->memcg == mem. This means we can * merge a series of uncharges to an uncharge of res_counter. * If not, we uncharge res_counter ony by one. */ |
c0ff4b854 memcg: rename mem... |
2734 |
if (batch->memcg != memcg) |
569b846df memcg: coalesce u... |
2735 2736 |
goto direct_uncharge; /* remember freed charge and uncharge it later */ |
7ffd4ca7a memcg: convert un... |
2737 |
batch->nr_pages++; |
569b846df memcg: coalesce u... |
2738 |
if (uncharge_memsw) |
7ffd4ca7a memcg: convert un... |
2739 |
batch->memsw_nr_pages++; |
569b846df memcg: coalesce u... |
2740 2741 |
return; direct_uncharge: |
c0ff4b854 memcg: rename mem... |
2742 |
res_counter_uncharge(&memcg->res, nr_pages * PAGE_SIZE); |
569b846df memcg: coalesce u... |
2743 |
if (uncharge_memsw) |
c0ff4b854 memcg: rename mem... |
2744 2745 2746 |
res_counter_uncharge(&memcg->memsw, nr_pages * PAGE_SIZE); if (unlikely(batch->memcg != memcg)) memcg_oom_recover(memcg); |
569b846df memcg: coalesce u... |
2747 2748 |
return; } |
7a81b88cb memcg: introduce ... |
2749 |
|
8697d3319 Memory controller... |
2750 |
/* |
69029cd55 memcg: remove ref... |
2751 |
* uncharge if !page_mapped(page) |
8a9f3ccd2 Memory controller... |
2752 |
*/ |
8c7c6e34a memcg: mem+swap c... |
2753 |
static struct mem_cgroup * |
69029cd55 memcg: remove ref... |
2754 |
__mem_cgroup_uncharge_common(struct page *page, enum charge_type ctype) |
8a9f3ccd2 Memory controller... |
2755 |
{ |
c0ff4b854 memcg: rename mem... |
2756 |
struct mem_cgroup *memcg = NULL; |
7ec99d621 memcg: unify char... |
2757 2758 |
unsigned int nr_pages = 1; struct page_cgroup *pc; |
8a9f3ccd2 Memory controller... |
2759 |
|
f8d665422 memcg: add mem_cg... |
2760 |
if (mem_cgroup_disabled()) |
8c7c6e34a memcg: mem+swap c... |
2761 |
return NULL; |
4077960e2 memory controller... |
2762 |
|
d13d14430 memcg: handle swa... |
2763 |
if (PageSwapCache(page)) |
8c7c6e34a memcg: mem+swap c... |
2764 |
return NULL; |
d13d14430 memcg: handle swa... |
2765 |
|
37c2ac787 thp: compound_tra... |
2766 |
if (PageTransHuge(page)) { |
7ec99d621 memcg: unify char... |
2767 |
nr_pages <<= compound_order(page); |
37c2ac787 thp: compound_tra... |
2768 2769 |
VM_BUG_ON(!PageTransHuge(page)); } |
8697d3319 Memory controller... |
2770 |
/* |
3c541e14b Memory controller... |
2771 |
* Check if our page_cgroup is valid |
8697d3319 Memory controller... |
2772 |
*/ |
52d4b9ac0 memcg: allocate a... |
2773 |
pc = lookup_page_cgroup(page); |
cfa449461 mm: memcg: lookup... |
2774 |
if (unlikely(!PageCgroupUsed(pc))) |
8c7c6e34a memcg: mem+swap c... |
2775 |
return NULL; |
b9c565d5a memcg: remove cle... |
2776 |
|
52d4b9ac0 memcg: allocate a... |
2777 |
lock_page_cgroup(pc); |
d13d14430 memcg: handle swa... |
2778 |
|
c0ff4b854 memcg: rename mem... |
2779 |
memcg = pc->mem_cgroup; |
8c7c6e34a memcg: mem+swap c... |
2780 |
|
d13d14430 memcg: handle swa... |
2781 2782 2783 2784 2785 |
if (!PageCgroupUsed(pc)) goto unlock_out; switch (ctype) { case MEM_CGROUP_CHARGE_TYPE_MAPPED: |
8a9478ca7 memcg: fix swap a... |
2786 |
case MEM_CGROUP_CHARGE_TYPE_DROP: |
ac39cf8cb memcg: fix mis-ac... |
2787 2788 |
/* See mem_cgroup_prepare_migration() */ if (page_mapped(page) || PageCgroupMigration(pc)) |
d13d14430 memcg: handle swa... |
2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799 |
goto unlock_out; break; case MEM_CGROUP_CHARGE_TYPE_SWAPOUT: if (!PageAnon(page)) { /* Shared memory */ if (page->mapping && !page_is_file_cache(page)) goto unlock_out; } else if (page_mapped(page)) /* Anon */ goto unlock_out; break; default: break; |
52d4b9ac0 memcg: allocate a... |
2800 |
} |
d13d14430 memcg: handle swa... |
2801 |
|
c0ff4b854 memcg: rename mem... |
2802 |
mem_cgroup_charge_statistics(memcg, PageCgroupCache(pc), -nr_pages); |
04046e1a0 memcg: use CSS ID |
2803 |
|
52d4b9ac0 memcg: allocate a... |
2804 |
ClearPageCgroupUsed(pc); |
544122e5e memcg: fix LRU ac... |
2805 2806 2807 2808 2809 2810 |
/* * pc->mem_cgroup is not cleared here. It will be accessed when it's * freed from LRU. This is safe because uncharged page is expected not * to be reused (freed soon). Exception is SwapCache, it's handled by * special functions. */ |
b9c565d5a memcg: remove cle... |
2811 |
|
52d4b9ac0 memcg: allocate a... |
2812 |
unlock_page_cgroup(pc); |
f75ca9620 memcg: avoid css_... |
2813 |
/* |
c0ff4b854 memcg: rename mem... |
2814 |
* even after unlock, we have memcg->res.usage here and this memcg |
f75ca9620 memcg: avoid css_... |
2815 2816 |
* will never be freed. */ |
c0ff4b854 memcg: rename mem... |
2817 |
memcg_check_events(memcg, page); |
f75ca9620 memcg: avoid css_... |
2818 |
if (do_swap_account && ctype == MEM_CGROUP_CHARGE_TYPE_SWAPOUT) { |
c0ff4b854 memcg: rename mem... |
2819 2820 |
mem_cgroup_swap_statistics(memcg, true); mem_cgroup_get(memcg); |
f75ca9620 memcg: avoid css_... |
2821 |
} |
c0ff4b854 memcg: rename mem... |
2822 2823 |
if (!mem_cgroup_is_root(memcg)) mem_cgroup_do_uncharge(memcg, nr_pages, ctype); |
6d12e2d8d per-zone and recl... |
2824 |
|
c0ff4b854 memcg: rename mem... |
2825 |
return memcg; |
d13d14430 memcg: handle swa... |
2826 2827 2828 |
unlock_out: unlock_page_cgroup(pc); |
8c7c6e34a memcg: mem+swap c... |
2829 |
return NULL; |
3c541e14b Memory controller... |
2830 |
} |
69029cd55 memcg: remove ref... |
2831 2832 |
void mem_cgroup_uncharge_page(struct page *page) { |
52d4b9ac0 memcg: allocate a... |
2833 2834 2835 |
/* early check. */ if (page_mapped(page)) return; |
40f23a21a mm: memcg: remove... |
2836 |
VM_BUG_ON(page->mapping && !PageAnon(page)); |
69029cd55 memcg: remove ref... |
2837 2838 2839 2840 2841 2842 |
__mem_cgroup_uncharge_common(page, MEM_CGROUP_CHARGE_TYPE_MAPPED); } void mem_cgroup_uncharge_cache_page(struct page *page) { VM_BUG_ON(page_mapped(page)); |
b7abea963 memcg: make page-... |
2843 |
VM_BUG_ON(page->mapping); |
69029cd55 memcg: remove ref... |
2844 2845 |
__mem_cgroup_uncharge_common(page, MEM_CGROUP_CHARGE_TYPE_CACHE); } |
569b846df memcg: coalesce u... |
2846 2847 2848 2849 2850 2851 2852 2853 2854 2855 2856 2857 2858 2859 |
/* * Batch_start/batch_end is called in unmap_page_range/invlidate/trucate. * In that cases, pages are freed continuously and we can expect pages * are in the same memcg. All these calls itself limits the number of * pages freed at once, then uncharge_start/end() is called properly. * This may be called prural(2) times in a context, */ void mem_cgroup_uncharge_start(void) { current->memcg_batch.do_batch++; /* We can do nest. */ if (current->memcg_batch.do_batch == 1) { current->memcg_batch.memcg = NULL; |
7ffd4ca7a memcg: convert un... |
2860 2861 |
current->memcg_batch.nr_pages = 0; current->memcg_batch.memsw_nr_pages = 0; |
569b846df memcg: coalesce u... |
2862 2863 2864 2865 2866 2867 2868 2869 2870 2871 2872 2873 2874 2875 2876 2877 2878 2879 2880 2881 |
} } void mem_cgroup_uncharge_end(void) { struct memcg_batch_info *batch = ¤t->memcg_batch; if (!batch->do_batch) return; batch->do_batch--; if (batch->do_batch) /* If stacked, do nothing. */ return; if (!batch->memcg) return; /* * This "batch->memcg" is valid without any css_get/put etc... * bacause we hide charges behind us. */ |
7ffd4ca7a memcg: convert un... |
2882 2883 2884 2885 2886 2887 |
if (batch->nr_pages) res_counter_uncharge(&batch->memcg->res, batch->nr_pages * PAGE_SIZE); if (batch->memsw_nr_pages) res_counter_uncharge(&batch->memcg->memsw, batch->memsw_nr_pages * PAGE_SIZE); |
3c11ecf44 memcg: oom kill d... |
2888 |
memcg_oom_recover(batch->memcg); |
569b846df memcg: coalesce u... |
2889 2890 2891 |
/* forget this pointer (for sanity check) */ batch->memcg = NULL; } |
4e5f01c2b memcg: clear pc->... |
2892 2893 2894 2895 2896 2897 2898 2899 2900 2901 2902 2903 2904 2905 2906 2907 |
/* * A function for resetting pc->mem_cgroup for newly allocated pages. * This function should be called if the newpage will be added to LRU * before start accounting. */ void mem_cgroup_reset_owner(struct page *newpage) { struct page_cgroup *pc; if (mem_cgroup_disabled()) return; pc = lookup_page_cgroup(newpage); VM_BUG_ON(PageCgroupUsed(pc)); pc->mem_cgroup = root_mem_cgroup; } |
e767e0561 memcg: fix deadlo... |
2908 |
#ifdef CONFIG_SWAP |
8c7c6e34a memcg: mem+swap c... |
2909 |
/* |
e767e0561 memcg: fix deadlo... |
2910 |
* called after __delete_from_swap_cache() and drop "page" account. |
8c7c6e34a memcg: mem+swap c... |
2911 2912 |
* memcg information is recorded to swap_cgroup of "ent" */ |
8a9478ca7 memcg: fix swap a... |
2913 2914 |
void mem_cgroup_uncharge_swapcache(struct page *page, swp_entry_t ent, bool swapout) |
8c7c6e34a memcg: mem+swap c... |
2915 2916 |
{ struct mem_cgroup *memcg; |
8a9478ca7 memcg: fix swap a... |
2917 2918 2919 2920 2921 2922 |
int ctype = MEM_CGROUP_CHARGE_TYPE_SWAPOUT; if (!swapout) /* this was a swap cache but the swap is unused ! */ ctype = MEM_CGROUP_CHARGE_TYPE_DROP; memcg = __mem_cgroup_uncharge_common(page, ctype); |
8c7c6e34a memcg: mem+swap c... |
2923 |
|
f75ca9620 memcg: avoid css_... |
2924 2925 2926 2927 2928 |
/* * record memcg information, if swapout && memcg != NULL, * mem_cgroup_get() was called in uncharge(). */ if (do_swap_account && swapout && memcg) |
a3b2d6926 cgroups: use css ... |
2929 |
swap_cgroup_record(ent, css_id(&memcg->css)); |
8c7c6e34a memcg: mem+swap c... |
2930 |
} |
e767e0561 memcg: fix deadlo... |
2931 |
#endif |
8c7c6e34a memcg: mem+swap c... |
2932 2933 2934 2935 2936 2937 2938 |
#ifdef CONFIG_CGROUP_MEM_RES_CTLR_SWAP /* * called from swap_entry_free(). remove record in swap_cgroup and * uncharge "memsw" account. */ void mem_cgroup_uncharge_swap(swp_entry_t ent) |
d13d14430 memcg: handle swa... |
2939 |
{ |
8c7c6e34a memcg: mem+swap c... |
2940 |
struct mem_cgroup *memcg; |
a3b2d6926 cgroups: use css ... |
2941 |
unsigned short id; |
8c7c6e34a memcg: mem+swap c... |
2942 2943 2944 |
if (!do_swap_account) return; |
a3b2d6926 cgroups: use css ... |
2945 2946 2947 |
id = swap_cgroup_record(ent, 0); rcu_read_lock(); memcg = mem_cgroup_lookup(id); |
8c7c6e34a memcg: mem+swap c... |
2948 |
if (memcg) { |
a3b2d6926 cgroups: use css ... |
2949 2950 2951 2952 |
/* * We uncharge this because swap is freed. * This memcg can be obsolete one. We avoid calling css_tryget */ |
0c3e73e84 memcg: improve re... |
2953 |
if (!mem_cgroup_is_root(memcg)) |
4e649152c memcg: some modif... |
2954 |
res_counter_uncharge(&memcg->memsw, PAGE_SIZE); |
0c3e73e84 memcg: improve re... |
2955 |
mem_cgroup_swap_statistics(memcg, false); |
8c7c6e34a memcg: mem+swap c... |
2956 2957 |
mem_cgroup_put(memcg); } |
a3b2d6926 cgroups: use css ... |
2958 |
rcu_read_unlock(); |
d13d14430 memcg: handle swa... |
2959 |
} |
024914477 memcg: move charg... |
2960 2961 2962 2963 2964 2965 |
/** * 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 |
483c30b51 memcg: improve pe... |
2966 |
* @need_fixup: whether we should fixup res_counters and refcounts. |
024914477 memcg: move charg... |
2967 2968 2969 2970 2971 2972 2973 2974 2975 2976 |
* * 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. * * The caller must have charged to @to, IOW, called res_counter_charge() about * both res and memsw, and called css_get(). */ static int mem_cgroup_move_swap_account(swp_entry_t entry, |
483c30b51 memcg: improve pe... |
2977 |
struct mem_cgroup *from, struct mem_cgroup *to, bool need_fixup) |
024914477 memcg: move charg... |
2978 2979 2980 2981 2982 2983 2984 |
{ unsigned short old_id, new_id; old_id = css_id(&from->css); new_id = css_id(&to->css); if (swap_cgroup_cmpxchg(entry, old_id, new_id) == old_id) { |
024914477 memcg: move charg... |
2985 |
mem_cgroup_swap_statistics(from, false); |
483c30b51 memcg: improve pe... |
2986 |
mem_cgroup_swap_statistics(to, true); |
024914477 memcg: move charg... |
2987 |
/* |
483c30b51 memcg: improve pe... |
2988 2989 2990 2991 2992 2993 |
* This function is only called from task migration context now. * It postpones res_counter and refcount handling till the end * of task migration(mem_cgroup_clear_mc()) for performance * improvement. But we cannot postpone mem_cgroup_get(to) * because if the process that has been moved to @to does * swap-in, the refcount of @to might be decreased to 0. |
024914477 memcg: move charg... |
2994 |
*/ |
024914477 memcg: move charg... |
2995 |
mem_cgroup_get(to); |
483c30b51 memcg: improve pe... |
2996 2997 2998 2999 3000 3001 3002 3003 3004 3005 |
if (need_fixup) { if (!mem_cgroup_is_root(from)) res_counter_uncharge(&from->memsw, PAGE_SIZE); mem_cgroup_put(from); /* * we charged both to->res and to->memsw, so we should * uncharge to->res. */ if (!mem_cgroup_is_root(to)) res_counter_uncharge(&to->res, PAGE_SIZE); |
483c30b51 memcg: improve pe... |
3006 |
} |
024914477 memcg: move charg... |
3007 3008 3009 3010 3011 3012 |
return 0; } return -EINVAL; } #else static inline int mem_cgroup_move_swap_account(swp_entry_t entry, |
483c30b51 memcg: improve pe... |
3013 |
struct mem_cgroup *from, struct mem_cgroup *to, bool need_fixup) |
024914477 memcg: move charg... |
3014 3015 3016 |
{ return -EINVAL; } |
8c7c6e34a memcg: mem+swap c... |
3017 |
#endif |
d13d14430 memcg: handle swa... |
3018 |
|
ae41be374 bugfix for memory... |
3019 |
/* |
01b1ae63c memcg: simple mig... |
3020 3021 |
* Before starting migration, account PAGE_SIZE to mem_cgroup that the old * page belongs to. |
ae41be374 bugfix for memory... |
3022 |
*/ |
ac39cf8cb memcg: fix mis-ac... |
3023 |
int mem_cgroup_prepare_migration(struct page *page, |
72835c86c mm: unify remaini... |
3024 |
struct page *newpage, struct mem_cgroup **memcgp, gfp_t gfp_mask) |
ae41be374 bugfix for memory... |
3025 |
{ |
c0ff4b854 memcg: rename mem... |
3026 |
struct mem_cgroup *memcg = NULL; |
7ec99d621 memcg: unify char... |
3027 |
struct page_cgroup *pc; |
ac39cf8cb memcg: fix mis-ac... |
3028 |
enum charge_type ctype; |
e8589cc18 memcg: better mig... |
3029 |
int ret = 0; |
8869b8f6e memcg: memcontrol... |
3030 |
|
72835c86c mm: unify remaini... |
3031 |
*memcgp = NULL; |
56039efa1 memcg: fix ugly i... |
3032 |
|
ec1685109 thp: memcg compound |
3033 |
VM_BUG_ON(PageTransHuge(page)); |
f8d665422 memcg: add mem_cg... |
3034 |
if (mem_cgroup_disabled()) |
4077960e2 memory controller... |
3035 |
return 0; |
52d4b9ac0 memcg: allocate a... |
3036 3037 3038 |
pc = lookup_page_cgroup(page); lock_page_cgroup(pc); if (PageCgroupUsed(pc)) { |
c0ff4b854 memcg: rename mem... |
3039 3040 |
memcg = pc->mem_cgroup; css_get(&memcg->css); |
ac39cf8cb memcg: fix mis-ac... |
3041 3042 3043 3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054 3055 3056 3057 3058 3059 3060 3061 3062 3063 3064 3065 3066 3067 3068 3069 3070 3071 |
/* * At migrating an anonymous page, its mapcount goes down * to 0 and uncharge() will be called. But, even if it's fully * unmapped, migration may fail and this page has to be * charged again. We set MIGRATION flag here and delay uncharge * until end_migration() is called * * Corner Case Thinking * A) * When the old page was mapped as Anon and it's unmap-and-freed * while migration was ongoing. * If unmap finds the old page, uncharge() of it will be delayed * until end_migration(). If unmap finds a new page, it's * uncharged when it make mapcount to be 1->0. If unmap code * finds swap_migration_entry, the new page will not be mapped * and end_migration() will find it(mapcount==0). * * B) * When the old page was mapped but migraion fails, the kernel * remaps it. A charge for it is kept by MIGRATION flag even * if mapcount goes down to 0. We can do remap successfully * without charging it again. * * C) * The "old" page is under lock_page() until the end of * migration, so, the old page itself will not be swapped-out. * If the new page is swapped out before end_migraton, our * hook to usual swap-out path will catch the event. */ if (PageAnon(page)) SetPageCgroupMigration(pc); |
e8589cc18 memcg: better mig... |
3072 |
} |
52d4b9ac0 memcg: allocate a... |
3073 |
unlock_page_cgroup(pc); |
ac39cf8cb memcg: fix mis-ac... |
3074 3075 3076 3077 |
/* * If the page is not charged at this point, * we return here. */ |
c0ff4b854 memcg: rename mem... |
3078 |
if (!memcg) |
ac39cf8cb memcg: fix mis-ac... |
3079 |
return 0; |
01b1ae63c memcg: simple mig... |
3080 |
|
72835c86c mm: unify remaini... |
3081 3082 |
*memcgp = memcg; ret = __mem_cgroup_try_charge(NULL, gfp_mask, 1, memcgp, false); |
c0ff4b854 memcg: rename mem... |
3083 |
css_put(&memcg->css);/* drop extra refcnt */ |
38c5d72f3 memcg: simplify L... |
3084 |
if (ret) { |
ac39cf8cb memcg: fix mis-ac... |
3085 3086 3087 3088 3089 3090 3091 3092 3093 |
if (PageAnon(page)) { lock_page_cgroup(pc); ClearPageCgroupMigration(pc); unlock_page_cgroup(pc); /* * The old page may be fully unmapped while we kept it. */ mem_cgroup_uncharge_page(page); } |
38c5d72f3 memcg: simplify L... |
3094 |
/* we'll need to revisit this error code (we have -EINTR) */ |
ac39cf8cb memcg: fix mis-ac... |
3095 |
return -ENOMEM; |
e8589cc18 memcg: better mig... |
3096 |
} |
ac39cf8cb memcg: fix mis-ac... |
3097 3098 3099 3100 3101 3102 3103 3104 3105 3106 3107 3108 3109 |
/* * We charge new page before it's used/mapped. So, even if unlock_page() * is called before end_migration, we can catch all events on this new * page. In the case new page is migrated but not remapped, new page's * mapcount will be finally 0 and we call uncharge in end_migration(). */ pc = lookup_page_cgroup(newpage); if (PageAnon(page)) ctype = MEM_CGROUP_CHARGE_TYPE_MAPPED; else if (page_is_file_cache(page)) ctype = MEM_CGROUP_CHARGE_TYPE_CACHE; else ctype = MEM_CGROUP_CHARGE_TYPE_SHMEM; |
c0ff4b854 memcg: rename mem... |
3110 |
__mem_cgroup_commit_charge(memcg, page, 1, pc, ctype); |
e8589cc18 memcg: better mig... |
3111 |
return ret; |
ae41be374 bugfix for memory... |
3112 |
} |
8869b8f6e memcg: memcontrol... |
3113 |
|
69029cd55 memcg: remove ref... |
3114 |
/* remove redundant charge if migration failed*/ |
c0ff4b854 memcg: rename mem... |
3115 |
void mem_cgroup_end_migration(struct mem_cgroup *memcg, |
50de1dd96 memcg: fix memory... |
3116 |
struct page *oldpage, struct page *newpage, bool migration_ok) |
ae41be374 bugfix for memory... |
3117 |
{ |
ac39cf8cb memcg: fix mis-ac... |
3118 |
struct page *used, *unused; |
01b1ae63c memcg: simple mig... |
3119 |
struct page_cgroup *pc; |
01b1ae63c memcg: simple mig... |
3120 |
|
c0ff4b854 memcg: rename mem... |
3121 |
if (!memcg) |
01b1ae63c memcg: simple mig... |
3122 |
return; |
ac39cf8cb memcg: fix mis-ac... |
3123 |
/* blocks rmdir() */ |
c0ff4b854 memcg: rename mem... |
3124 |
cgroup_exclude_rmdir(&memcg->css); |
50de1dd96 memcg: fix memory... |
3125 |
if (!migration_ok) { |
ac39cf8cb memcg: fix mis-ac... |
3126 3127 |
used = oldpage; unused = newpage; |
01b1ae63c memcg: simple mig... |
3128 |
} else { |
ac39cf8cb memcg: fix mis-ac... |
3129 |
used = newpage; |
01b1ae63c memcg: simple mig... |
3130 3131 |
unused = oldpage; } |
69029cd55 memcg: remove ref... |
3132 |
/* |
ac39cf8cb memcg: fix mis-ac... |
3133 3134 3135 |
* We disallowed uncharge of pages under migration because mapcount * of the page goes down to zero, temporarly. * Clear the flag and check the page should be charged. |
01b1ae63c memcg: simple mig... |
3136 |
*/ |
ac39cf8cb memcg: fix mis-ac... |
3137 3138 3139 3140 |
pc = lookup_page_cgroup(oldpage); lock_page_cgroup(pc); ClearPageCgroupMigration(pc); unlock_page_cgroup(pc); |
01b1ae63c memcg: simple mig... |
3141 |
|
ac39cf8cb memcg: fix mis-ac... |
3142 |
__mem_cgroup_uncharge_common(unused, MEM_CGROUP_CHARGE_TYPE_FORCE); |
01b1ae63c memcg: simple mig... |
3143 |
/* |
ac39cf8cb memcg: fix mis-ac... |
3144 3145 3146 3147 3148 3149 |
* If a page is a file cache, radix-tree replacement is very atomic * and we can skip this check. When it was an Anon page, its mapcount * goes down to 0. But because we added MIGRATION flage, it's not * uncharged yet. There are several case but page->mapcount check * and USED bit check in mem_cgroup_uncharge_page() will do enough * check. (see prepare_charge() also) |
69029cd55 memcg: remove ref... |
3150 |
*/ |
ac39cf8cb memcg: fix mis-ac... |
3151 3152 |
if (PageAnon(used)) mem_cgroup_uncharge_page(used); |
887032670 cgroup avoid perm... |
3153 |
/* |
ac39cf8cb memcg: fix mis-ac... |
3154 3155 |
* At migration, we may charge account against cgroup which has no * tasks. |
887032670 cgroup avoid perm... |
3156 3157 3158 |
* So, rmdir()->pre_destroy() can be called while we do this charge. * In that case, we need to call pre_destroy() again. check it here. */ |
c0ff4b854 memcg: rename mem... |
3159 |
cgroup_release_and_wakeup_rmdir(&memcg->css); |
ae41be374 bugfix for memory... |
3160 |
} |
78fb74669 Memory controller... |
3161 |
|
ab936cbcd memcg: add mem_cg... |
3162 3163 3164 3165 3166 3167 3168 3169 3170 3171 |
/* * At replace page cache, newpage is not under any memcg but it's on * LRU. So, this function doesn't touch res_counter but handles LRU * in correct way. Both pages are locked so we cannot race with uncharge. */ void mem_cgroup_replace_page_cache(struct page *oldpage, struct page *newpage) { struct mem_cgroup *memcg; struct page_cgroup *pc; |
ab936cbcd memcg: add mem_cg... |
3172 |
enum charge_type type = MEM_CGROUP_CHARGE_TYPE_CACHE; |
ab936cbcd memcg: add mem_cg... |
3173 3174 3175 3176 3177 3178 3179 3180 3181 3182 3183 3184 3185 3186 |
if (mem_cgroup_disabled()) return; pc = lookup_page_cgroup(oldpage); /* fix accounting on old pages */ lock_page_cgroup(pc); memcg = pc->mem_cgroup; mem_cgroup_charge_statistics(memcg, PageCgroupCache(pc), -1); ClearPageCgroupUsed(pc); unlock_page_cgroup(pc); if (PageSwapBacked(oldpage)) type = MEM_CGROUP_CHARGE_TYPE_SHMEM; |
ab936cbcd memcg: add mem_cg... |
3187 3188 3189 3190 3191 |
/* * Even if newpage->mapping was NULL before starting replacement, * the newpage may be on LRU(or pagevec for LRU) already. We lock * LRU while we overwrite pc->mem_cgroup. */ |
36b62ad53 memcg: simplify c... |
3192 |
__mem_cgroup_commit_charge_lrucare(newpage, memcg, type); |
ab936cbcd memcg: add mem_cg... |
3193 |
} |
f212ad7cf memcg: add memcg ... |
3194 3195 3196 3197 3198 3199 |
#ifdef CONFIG_DEBUG_VM static struct page_cgroup *lookup_page_cgroup_used(struct page *page) { struct page_cgroup *pc; pc = lookup_page_cgroup(page); |
cfa449461 mm: memcg: lookup... |
3200 3201 3202 3203 3204 |
/* * Can be NULL while feeding pages into the page allocator for * the first time, i.e. during boot or memory hotplug; * or when mem_cgroup_disabled(). */ |
f212ad7cf memcg: add memcg ... |
3205 3206 3207 3208 3209 3210 3211 3212 3213 3214 3215 3216 3217 3218 3219 3220 3221 3222 3223 |
if (likely(pc) && PageCgroupUsed(pc)) return pc; return NULL; } bool mem_cgroup_bad_page_check(struct page *page) { if (mem_cgroup_disabled()) return false; return lookup_page_cgroup_used(page) != NULL; } void mem_cgroup_print_bad_page(struct page *page) { struct page_cgroup *pc; pc = lookup_page_cgroup_used(page); if (pc) { |
90b3feaec memcg: fix mem_cg... |
3224 3225 |
printk(KERN_ALERT "pc:%p pc->flags:%lx pc->mem_cgroup:%p ", |
f212ad7cf memcg: add memcg ... |
3226 |
pc, pc->flags, pc->mem_cgroup); |
f212ad7cf memcg: add memcg ... |
3227 3228 3229 |
} } #endif |
8c7c6e34a memcg: mem+swap c... |
3230 |
static DEFINE_MUTEX(set_limit_mutex); |
d38d2a758 mm: make mem_cgro... |
3231 |
static int mem_cgroup_resize_limit(struct mem_cgroup *memcg, |
8c7c6e34a memcg: mem+swap c... |
3232 |
unsigned long long val) |
628f42355 memcg: limit chan... |
3233 |
{ |
81d39c20f memcg: fix shrink... |
3234 |
int retry_count; |
3c11ecf44 memcg: oom kill d... |
3235 |
u64 memswlimit, memlimit; |
628f42355 memcg: limit chan... |
3236 |
int ret = 0; |
81d39c20f memcg: fix shrink... |
3237 3238 |
int children = mem_cgroup_count_children(memcg); u64 curusage, oldusage; |
3c11ecf44 memcg: oom kill d... |
3239 |
int enlarge; |
81d39c20f memcg: fix shrink... |
3240 3241 3242 3243 3244 3245 3246 3247 3248 |
/* * For keeping hierarchical_reclaim simple, how long we should retry * is depends on callers. We set our retry-count to be function * of # of children which we should visit in this loop. */ retry_count = MEM_CGROUP_RECLAIM_RETRIES * children; oldusage = res_counter_read_u64(&memcg->res, RES_USAGE); |
628f42355 memcg: limit chan... |
3249 |
|
3c11ecf44 memcg: oom kill d... |
3250 |
enlarge = 0; |
8c7c6e34a memcg: mem+swap c... |
3251 |
while (retry_count) { |
628f42355 memcg: limit chan... |
3252 3253 3254 3255 |
if (signal_pending(current)) { ret = -EINTR; break; } |
8c7c6e34a memcg: mem+swap c... |
3256 3257 3258 |
/* * Rather than hide all in some function, I do this in * open coded manner. You see what this really does. |
c0ff4b854 memcg: rename mem... |
3259 |
* We have to guarantee memcg->res.limit < memcg->memsw.limit. |
8c7c6e34a memcg: mem+swap c... |
3260 3261 3262 3263 3264 3265 |
*/ mutex_lock(&set_limit_mutex); memswlimit = res_counter_read_u64(&memcg->memsw, RES_LIMIT); if (memswlimit < val) { ret = -EINVAL; mutex_unlock(&set_limit_mutex); |
628f42355 memcg: limit chan... |
3266 3267 |
break; } |
3c11ecf44 memcg: oom kill d... |
3268 3269 3270 3271 |
memlimit = res_counter_read_u64(&memcg->res, RES_LIMIT); if (memlimit < val) enlarge = 1; |
8c7c6e34a memcg: mem+swap c... |
3272 |
ret = res_counter_set_limit(&memcg->res, val); |
22a668d7c memcg: fix behavi... |
3273 3274 3275 3276 3277 3278 |
if (!ret) { if (memswlimit == val) memcg->memsw_is_minimum = true; else memcg->memsw_is_minimum = false; } |
8c7c6e34a memcg: mem+swap c... |
3279 3280 3281 3282 |
mutex_unlock(&set_limit_mutex); if (!ret) break; |
5660048cc mm: move memcg hi... |
3283 3284 |
mem_cgroup_reclaim(memcg, GFP_KERNEL, MEM_CGROUP_RECLAIM_SHRINK); |
81d39c20f memcg: fix shrink... |
3285 3286 3287 3288 3289 3290 |
curusage = res_counter_read_u64(&memcg->res, RES_USAGE); /* Usage is reduced ? */ if (curusage >= oldusage) retry_count--; else oldusage = curusage; |
8c7c6e34a memcg: mem+swap c... |
3291 |
} |
3c11ecf44 memcg: oom kill d... |
3292 3293 |
if (!ret && enlarge) memcg_oom_recover(memcg); |
14797e236 memcg: add inacti... |
3294 |
|
8c7c6e34a memcg: mem+swap c... |
3295 3296 |
return ret; } |
338c84310 memcg: remove som... |
3297 3298 |
static int mem_cgroup_resize_memsw_limit(struct mem_cgroup *memcg, unsigned long long val) |
8c7c6e34a memcg: mem+swap c... |
3299 |
{ |
81d39c20f memcg: fix shrink... |
3300 |
int retry_count; |
3c11ecf44 memcg: oom kill d... |
3301 |
u64 memlimit, memswlimit, oldusage, curusage; |
81d39c20f memcg: fix shrink... |
3302 3303 |
int children = mem_cgroup_count_children(memcg); int ret = -EBUSY; |
3c11ecf44 memcg: oom kill d... |
3304 |
int enlarge = 0; |
8c7c6e34a memcg: mem+swap c... |
3305 |
|
81d39c20f memcg: fix shrink... |
3306 3307 3308 |
/* see mem_cgroup_resize_res_limit */ retry_count = children * MEM_CGROUP_RECLAIM_RETRIES; oldusage = res_counter_read_u64(&memcg->memsw, RES_USAGE); |
8c7c6e34a memcg: mem+swap c... |
3309 3310 3311 3312 3313 3314 3315 3316 |
while (retry_count) { if (signal_pending(current)) { ret = -EINTR; break; } /* * Rather than hide all in some function, I do this in * open coded manner. You see what this really does. |
c0ff4b854 memcg: rename mem... |
3317 |
* We have to guarantee memcg->res.limit < memcg->memsw.limit. |
8c7c6e34a memcg: mem+swap c... |
3318 3319 3320 3321 3322 3323 3324 3325 |
*/ mutex_lock(&set_limit_mutex); memlimit = res_counter_read_u64(&memcg->res, RES_LIMIT); if (memlimit > val) { ret = -EINVAL; mutex_unlock(&set_limit_mutex); break; } |
3c11ecf44 memcg: oom kill d... |
3326 3327 3328 |
memswlimit = res_counter_read_u64(&memcg->memsw, RES_LIMIT); if (memswlimit < val) enlarge = 1; |
8c7c6e34a memcg: mem+swap c... |
3329 |
ret = res_counter_set_limit(&memcg->memsw, val); |
22a668d7c memcg: fix behavi... |
3330 3331 3332 3333 3334 3335 |
if (!ret) { if (memlimit == val) memcg->memsw_is_minimum = true; else memcg->memsw_is_minimum = false; } |
8c7c6e34a memcg: mem+swap c... |
3336 3337 3338 3339 |
mutex_unlock(&set_limit_mutex); if (!ret) break; |
5660048cc mm: move memcg hi... |
3340 3341 3342 |
mem_cgroup_reclaim(memcg, GFP_KERNEL, MEM_CGROUP_RECLAIM_NOSWAP | MEM_CGROUP_RECLAIM_SHRINK); |
8c7c6e34a memcg: mem+swap c... |
3343 |
curusage = res_counter_read_u64(&memcg->memsw, RES_USAGE); |
81d39c20f memcg: fix shrink... |
3344 |
/* Usage is reduced ? */ |
8c7c6e34a memcg: mem+swap c... |
3345 |
if (curusage >= oldusage) |
628f42355 memcg: limit chan... |
3346 |
retry_count--; |
81d39c20f memcg: fix shrink... |
3347 3348 |
else oldusage = curusage; |
628f42355 memcg: limit chan... |
3349 |
} |
3c11ecf44 memcg: oom kill d... |
3350 3351 |
if (!ret && enlarge) memcg_oom_recover(memcg); |
628f42355 memcg: limit chan... |
3352 3353 |
return ret; } |
4e4169535 memory controller... |
3354 |
unsigned long mem_cgroup_soft_limit_reclaim(struct zone *zone, int order, |
0ae5e89c6 memcg: count the ... |
3355 3356 |
gfp_t gfp_mask, unsigned long *total_scanned) |
4e4169535 memory controller... |
3357 3358 3359 3360 3361 3362 |
{ 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; |
ef8745c1e memcg: reduce che... |
3363 |
unsigned long long excess; |
0ae5e89c6 memcg: count the ... |
3364 |
unsigned long nr_scanned; |
4e4169535 memory controller... |
3365 3366 3367 |
if (order > 0) return 0; |
00918b6ab memcg: remove nid... |
3368 |
mctz = soft_limit_tree_node_zone(zone_to_nid(zone), zone_idx(zone)); |
4e4169535 memory controller... |
3369 3370 3371 3372 3373 3374 3375 3376 3377 3378 3379 3380 |
/* * 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; |
0ae5e89c6 memcg: count the ... |
3381 |
nr_scanned = 0; |
5660048cc mm: move memcg hi... |
3382 3383 |
reclaimed = mem_cgroup_soft_reclaim(mz->mem, zone, gfp_mask, &nr_scanned); |
4e4169535 memory controller... |
3384 |
nr_reclaimed += reclaimed; |
0ae5e89c6 memcg: count the ... |
3385 |
*total_scanned += nr_scanned; |
4e4169535 memory controller... |
3386 3387 3388 3389 3390 3391 3392 3393 3394 3395 3396 3397 3398 3399 3400 3401 3402 3403 3404 3405 3406 3407 |
spin_lock(&mctz->lock); /* * If we failed to reclaim anything from this memory cgroup * it is time to move on to the next cgroup */ next_mz = NULL; if (!reclaimed) { do { /* * Loop until we find yet another one. * * By the time we get the soft_limit lock * again, someone might have aded the * group back on the RB tree. Iterate to * make sure we get a different mem. * mem_cgroup_largest_soft_limit_node returns * NULL if no other cgroup is present on * the tree */ next_mz = __mem_cgroup_largest_soft_limit_node(mctz); |
39cc98f1f memcg: remove poi... |
3408 |
if (next_mz == mz) |
4e4169535 memory controller... |
3409 |
css_put(&next_mz->mem->css); |
39cc98f1f memcg: remove poi... |
3410 |
else /* next_mz == NULL or other memcg */ |
4e4169535 memory controller... |
3411 3412 3413 |
break; } while (1); } |
4e4169535 memory controller... |
3414 |
__mem_cgroup_remove_exceeded(mz->mem, mz, mctz); |
ef8745c1e memcg: reduce che... |
3415 |
excess = res_counter_soft_limit_excess(&mz->mem->res); |
4e4169535 memory controller... |
3416 3417 3418 3419 3420 3421 3422 3423 |
/* * 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. */ |
ef8745c1e memcg: reduce che... |
3424 3425 |
/* If excess == 0, no tree ops */ __mem_cgroup_insert_exceeded(mz->mem, mz, mctz, excess); |
4e4169535 memory controller... |
3426 3427 3428 3429 3430 3431 3432 3433 3434 3435 3436 3437 3438 3439 3440 3441 3442 |
spin_unlock(&mctz->lock); css_put(&mz->mem->css); loop++; /* * Could not reclaim anything and there are no more * mem cgroups to try or we seem to be looping without * reclaiming anything. */ if (!nr_reclaimed && (next_mz == NULL || loop > MEM_CGROUP_MAX_SOFT_LIMIT_RECLAIM_LOOPS)) break; } while (!nr_reclaimed); if (next_mz) css_put(&next_mz->mem->css); return nr_reclaimed; } |
c9b0ed514 memcg: helper fun... |
3443 |
/* |
cc8475822 memory cgroup enh... |
3444 |
* This routine traverse page_cgroup in given list and drop them all. |
cc8475822 memory cgroup enh... |
3445 3446 |
* *And* this routine doesn't reclaim page itself, just removes page_cgroup. */ |
c0ff4b854 memcg: rename mem... |
3447 |
static int mem_cgroup_force_empty_list(struct mem_cgroup *memcg, |
08e552c69 memcg: synchroniz... |
3448 |
int node, int zid, enum lru_list lru) |
cc8475822 memory cgroup enh... |
3449 |
{ |
08e552c69 memcg: synchroniz... |
3450 |
struct mem_cgroup_per_zone *mz; |
08e552c69 memcg: synchroniz... |
3451 |
unsigned long flags, loop; |
072c56c13 per-zone and recl... |
3452 |
struct list_head *list; |
925b7673c mm: make per-memc... |
3453 3454 |
struct page *busy; struct zone *zone; |
f817ed485 memcg: move all a... |
3455 |
int ret = 0; |
072c56c13 per-zone and recl... |
3456 |
|
08e552c69 memcg: synchroniz... |
3457 |
zone = &NODE_DATA(node)->node_zones[zid]; |
c0ff4b854 memcg: rename mem... |
3458 |
mz = mem_cgroup_zoneinfo(memcg, node, zid); |
6290df545 mm: collect LRU l... |
3459 |
list = &mz->lruvec.lists[lru]; |
cc8475822 memory cgroup enh... |
3460 |
|
f817ed485 memcg: move all a... |
3461 3462 3463 3464 3465 |
loop = MEM_CGROUP_ZSTAT(mz, lru); /* give some margin against EBUSY etc...*/ loop += 256; busy = NULL; while (loop--) { |
925b7673c mm: make per-memc... |
3466 |
struct page_cgroup *pc; |
5564e88ba memcg: condense p... |
3467 |
struct page *page; |
f817ed485 memcg: move all a... |
3468 |
ret = 0; |
08e552c69 memcg: synchroniz... |
3469 |
spin_lock_irqsave(&zone->lru_lock, flags); |
f817ed485 memcg: move all a... |
3470 |
if (list_empty(list)) { |
08e552c69 memcg: synchroniz... |
3471 |
spin_unlock_irqrestore(&zone->lru_lock, flags); |
52d4b9ac0 memcg: allocate a... |
3472 |
break; |
f817ed485 memcg: move all a... |
3473 |
} |
925b7673c mm: make per-memc... |
3474 3475 3476 |
page = list_entry(list->prev, struct page, lru); if (busy == page) { list_move(&page->lru, list); |
648bcc771 mm/memcontrol.c: ... |
3477 |
busy = NULL; |
08e552c69 memcg: synchroniz... |
3478 |
spin_unlock_irqrestore(&zone->lru_lock, flags); |
f817ed485 memcg: move all a... |
3479 3480 |
continue; } |
08e552c69 memcg: synchroniz... |
3481 |
spin_unlock_irqrestore(&zone->lru_lock, flags); |
f817ed485 memcg: move all a... |
3482 |
|
925b7673c mm: make per-memc... |
3483 |
pc = lookup_page_cgroup(page); |
5564e88ba memcg: condense p... |
3484 |
|
c0ff4b854 memcg: rename mem... |
3485 |
ret = mem_cgroup_move_parent(page, pc, memcg, GFP_KERNEL); |
38c5d72f3 memcg: simplify L... |
3486 |
if (ret == -ENOMEM || ret == -EINTR) |
52d4b9ac0 memcg: allocate a... |
3487 |
break; |
f817ed485 memcg: move all a... |
3488 3489 3490 |
if (ret == -EBUSY || ret == -EINVAL) { /* found lock contention or "pc" is obsolete. */ |
925b7673c mm: make per-memc... |
3491 |
busy = page; |
f817ed485 memcg: move all a... |
3492 3493 3494 |
cond_resched(); } else busy = NULL; |
cc8475822 memory cgroup enh... |
3495 |
} |
08e552c69 memcg: synchroniz... |
3496 |
|
f817ed485 memcg: move all a... |
3497 3498 3499 |
if (!ret && !list_empty(list)) return -EBUSY; return ret; |
cc8475822 memory cgroup enh... |
3500 3501 3502 3503 3504 3505 |
} /* * make mem_cgroup's charge to be 0 if there is no task. * This enables deleting this mem_cgroup. */ |
c0ff4b854 memcg: rename mem... |
3506 |
static int mem_cgroup_force_empty(struct mem_cgroup *memcg, bool free_all) |
cc8475822 memory cgroup enh... |
3507 |
{ |
f817ed485 memcg: move all a... |
3508 3509 3510 |
int ret; int node, zid, shrink; int nr_retries = MEM_CGROUP_RECLAIM_RETRIES; |
c0ff4b854 memcg: rename mem... |
3511 |
struct cgroup *cgrp = memcg->css.cgroup; |
8869b8f6e memcg: memcontrol... |
3512 |
|
c0ff4b854 memcg: rename mem... |
3513 |
css_get(&memcg->css); |
f817ed485 memcg: move all a... |
3514 3515 |
shrink = 0; |
c1e862c1f memcg: new force_... |
3516 3517 3518 |
/* should free all ? */ if (free_all) goto try_to_free; |
f817ed485 memcg: move all a... |
3519 |
move_account: |
fce664775 memcg: ensure lis... |
3520 |
do { |
f817ed485 memcg: move all a... |
3521 |
ret = -EBUSY; |
c1e862c1f memcg: new force_... |
3522 3523 3524 3525 |
if (cgroup_task_count(cgrp) || !list_empty(&cgrp->children)) goto out; ret = -EINTR; if (signal_pending(current)) |
cc8475822 memory cgroup enh... |
3526 |
goto out; |
52d4b9ac0 memcg: allocate a... |
3527 3528 |
/* This is for making all *used* pages to be on LRU. */ lru_add_drain_all(); |
c0ff4b854 memcg: rename mem... |
3529 |
drain_all_stock_sync(memcg); |
f817ed485 memcg: move all a... |
3530 |
ret = 0; |
c0ff4b854 memcg: rename mem... |
3531 |
mem_cgroup_start_move(memcg); |
299b4eaa3 memcg: NULL point... |
3532 |
for_each_node_state(node, N_HIGH_MEMORY) { |
f817ed485 memcg: move all a... |
3533 |
for (zid = 0; !ret && zid < MAX_NR_ZONES; zid++) { |
b69408e88 vmscan: Use an in... |
3534 |
enum lru_list l; |
f817ed485 memcg: move all a... |
3535 |
for_each_lru(l) { |
c0ff4b854 memcg: rename mem... |
3536 |
ret = mem_cgroup_force_empty_list(memcg, |
08e552c69 memcg: synchroniz... |
3537 |
node, zid, l); |
f817ed485 memcg: move all a... |
3538 3539 3540 |
if (ret) break; } |
1ecaab2bd per-zone and recl... |
3541 |
} |
f817ed485 memcg: move all a... |
3542 3543 3544 |
if (ret) break; } |
c0ff4b854 memcg: rename mem... |
3545 3546 |
mem_cgroup_end_move(memcg); memcg_oom_recover(memcg); |
f817ed485 memcg: move all a... |
3547 3548 3549 |
/* it seems parent cgroup doesn't have enough mem */ if (ret == -ENOMEM) goto try_to_free; |
52d4b9ac0 memcg: allocate a... |
3550 |
cond_resched(); |
fce664775 memcg: ensure lis... |
3551 |
/* "ret" should also be checked to ensure all lists are empty. */ |
c0ff4b854 memcg: rename mem... |
3552 |
} while (memcg->res.usage > 0 || ret); |
cc8475822 memory cgroup enh... |
3553 |
out: |
c0ff4b854 memcg: rename mem... |
3554 |
css_put(&memcg->css); |
cc8475822 memory cgroup enh... |
3555 |
return ret; |
f817ed485 memcg: move all a... |
3556 3557 |
try_to_free: |
c1e862c1f memcg: new force_... |
3558 3559 |
/* returns EBUSY if there is a task or if we come here twice. */ if (cgroup_task_count(cgrp) || !list_empty(&cgrp->children) || shrink) { |
f817ed485 memcg: move all a... |
3560 3561 3562 |
ret = -EBUSY; goto out; } |
c1e862c1f memcg: new force_... |
3563 3564 |
/* we call try-to-free pages for make this cgroup empty */ lru_add_drain_all(); |
f817ed485 memcg: move all a... |
3565 3566 |
/* try to free all pages in this cgroup */ shrink = 1; |
c0ff4b854 memcg: rename mem... |
3567 |
while (nr_retries && memcg->res.usage > 0) { |
f817ed485 memcg: move all a... |
3568 |
int progress; |
c1e862c1f memcg: new force_... |
3569 3570 3571 3572 3573 |
if (signal_pending(current)) { ret = -EINTR; goto out; } |
c0ff4b854 memcg: rename mem... |
3574 |
progress = try_to_free_mem_cgroup_pages(memcg, GFP_KERNEL, |
185efc0f9 memcg: Revert "me... |
3575 |
false); |
c1e862c1f memcg: new force_... |
3576 |
if (!progress) { |
f817ed485 memcg: move all a... |
3577 |
nr_retries--; |
c1e862c1f memcg: new force_... |
3578 |
/* maybe some writeback is necessary */ |
8aa7e847d Fix congestion_wa... |
3579 |
congestion_wait(BLK_RW_ASYNC, HZ/10); |
c1e862c1f memcg: new force_... |
3580 |
} |
f817ed485 memcg: move all a... |
3581 3582 |
} |
08e552c69 memcg: synchroniz... |
3583 |
lru_add_drain(); |
f817ed485 memcg: move all a... |
3584 |
/* try move_account...there may be some *locked* pages. */ |
fce664775 memcg: ensure lis... |
3585 |
goto move_account; |
cc8475822 memory cgroup enh... |
3586 |
} |
c1e862c1f memcg: new force_... |
3587 3588 3589 3590 |
int mem_cgroup_force_empty_write(struct cgroup *cont, unsigned int event) { return mem_cgroup_force_empty(mem_cgroup_from_cont(cont), true); } |
18f59ea7d memcg: memory cgr... |
3591 3592 3593 3594 3595 3596 3597 3598 3599 |
static u64 mem_cgroup_hierarchy_read(struct cgroup *cont, struct cftype *cft) { return mem_cgroup_from_cont(cont)->use_hierarchy; } static int mem_cgroup_hierarchy_write(struct cgroup *cont, struct cftype *cft, u64 val) { int retval = 0; |
c0ff4b854 memcg: rename mem... |
3600 |
struct mem_cgroup *memcg = mem_cgroup_from_cont(cont); |
18f59ea7d memcg: memory cgr... |
3601 |
struct cgroup *parent = cont->parent; |
c0ff4b854 memcg: rename mem... |
3602 |
struct mem_cgroup *parent_memcg = NULL; |
18f59ea7d memcg: memory cgr... |
3603 3604 |
if (parent) |
c0ff4b854 memcg: rename mem... |
3605 |
parent_memcg = mem_cgroup_from_cont(parent); |
18f59ea7d memcg: memory cgr... |
3606 3607 3608 |
cgroup_lock(); /* |
af901ca18 tree-wide: fix as... |
3609 |
* If parent's use_hierarchy is set, we can't make any modifications |
18f59ea7d memcg: memory cgr... |
3610 3611 3612 3613 3614 3615 |
* 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... |
3616 |
if ((!parent_memcg || !parent_memcg->use_hierarchy) && |
18f59ea7d memcg: memory cgr... |
3617 3618 |
(val == 1 || val == 0)) { if (list_empty(&cont->children)) |
c0ff4b854 memcg: rename mem... |
3619 |
memcg->use_hierarchy = val; |
18f59ea7d memcg: memory cgr... |
3620 3621 3622 3623 3624 3625 3626 3627 |
else retval = -EBUSY; } else retval = -EINVAL; cgroup_unlock(); return retval; } |
0c3e73e84 memcg: improve re... |
3628 |
|
c0ff4b854 memcg: rename mem... |
3629 |
static unsigned long mem_cgroup_recursive_stat(struct mem_cgroup *memcg, |
7a159cc9d memcg: use native... |
3630 |
enum mem_cgroup_stat_index idx) |
0c3e73e84 memcg: improve re... |
3631 |
{ |
7d74b06f2 memcg: use for_ea... |
3632 |
struct mem_cgroup *iter; |
7a159cc9d memcg: use native... |
3633 |
long val = 0; |
0c3e73e84 memcg: improve re... |
3634 |
|
7a159cc9d memcg: use native... |
3635 |
/* Per-cpu values can be negative, use a signed accumulator */ |
c0ff4b854 memcg: rename mem... |
3636 |
for_each_mem_cgroup_tree(iter, memcg) |
7d74b06f2 memcg: use for_ea... |
3637 3638 3639 3640 3641 |
val += mem_cgroup_read_stat(iter, idx); if (val < 0) /* race ? */ val = 0; return val; |
0c3e73e84 memcg: improve re... |
3642 |
} |
c0ff4b854 memcg: rename mem... |
3643 |
static inline u64 mem_cgroup_usage(struct mem_cgroup *memcg, bool swap) |
104f39284 memcg: extract me... |
3644 |
{ |
7d74b06f2 memcg: use for_ea... |
3645 |
u64 val; |
104f39284 memcg: extract me... |
3646 |
|
c0ff4b854 memcg: rename mem... |
3647 |
if (!mem_cgroup_is_root(memcg)) { |
104f39284 memcg: extract me... |
3648 |
if (!swap) |
65c64ce8e Partial revert "B... |
3649 |
return res_counter_read_u64(&memcg->res, RES_USAGE); |
104f39284 memcg: extract me... |
3650 |
else |
65c64ce8e Partial revert "B... |
3651 |
return res_counter_read_u64(&memcg->memsw, RES_USAGE); |
104f39284 memcg: extract me... |
3652 |
} |
c0ff4b854 memcg: rename mem... |
3653 3654 |
val = mem_cgroup_recursive_stat(memcg, MEM_CGROUP_STAT_CACHE); val += mem_cgroup_recursive_stat(memcg, MEM_CGROUP_STAT_RSS); |
104f39284 memcg: extract me... |
3655 |
|
7d74b06f2 memcg: use for_ea... |
3656 |
if (swap) |
c0ff4b854 memcg: rename mem... |
3657 |
val += mem_cgroup_recursive_stat(memcg, MEM_CGROUP_STAT_SWAPOUT); |
104f39284 memcg: extract me... |
3658 3659 3660 |
return val << PAGE_SHIFT; } |
2c3daa722 CGroup API files:... |
3661 |
static u64 mem_cgroup_read(struct cgroup *cont, struct cftype *cft) |
8cdea7c05 Memory controller... |
3662 |
{ |
c0ff4b854 memcg: rename mem... |
3663 |
struct mem_cgroup *memcg = mem_cgroup_from_cont(cont); |
104f39284 memcg: extract me... |
3664 |
u64 val; |
8c7c6e34a memcg: mem+swap c... |
3665 3666 3667 3668 3669 3670 |
int type, name; type = MEMFILE_TYPE(cft->private); name = MEMFILE_ATTR(cft->private); switch (type) { case _MEM: |
104f39284 memcg: extract me... |
3671 |
if (name == RES_USAGE) |
c0ff4b854 memcg: rename mem... |
3672 |
val = mem_cgroup_usage(memcg, false); |
104f39284 memcg: extract me... |
3673 |
else |
c0ff4b854 memcg: rename mem... |
3674 |
val = res_counter_read_u64(&memcg->res, name); |
8c7c6e34a memcg: mem+swap c... |
3675 3676 |
break; case _MEMSWAP: |
104f39284 memcg: extract me... |
3677 |
if (name == RES_USAGE) |
c0ff4b854 memcg: rename mem... |
3678 |
val = mem_cgroup_usage(memcg, true); |
104f39284 memcg: extract me... |
3679 |
else |
c0ff4b854 memcg: rename mem... |
3680 |
val = res_counter_read_u64(&memcg->memsw, name); |
8c7c6e34a memcg: mem+swap c... |
3681 3682 3683 3684 3685 3686 |
break; default: BUG(); break; } return val; |
8cdea7c05 Memory controller... |
3687 |
} |
628f42355 memcg: limit chan... |
3688 3689 3690 3691 |
/* * The user of this function is... * RES_LIMIT. */ |
856c13aa1 cgroup files: con... |
3692 3693 |
static int mem_cgroup_write(struct cgroup *cont, struct cftype *cft, const char *buffer) |
8cdea7c05 Memory controller... |
3694 |
{ |
628f42355 memcg: limit chan... |
3695 |
struct mem_cgroup *memcg = mem_cgroup_from_cont(cont); |
8c7c6e34a memcg: mem+swap c... |
3696 |
int type, name; |
628f42355 memcg: limit chan... |
3697 3698 |
unsigned long long val; int ret; |
8c7c6e34a memcg: mem+swap c... |
3699 3700 3701 |
type = MEMFILE_TYPE(cft->private); name = MEMFILE_ATTR(cft->private); switch (name) { |
628f42355 memcg: limit chan... |
3702 |
case RES_LIMIT: |
4b3bde4c9 memcg: remove the... |
3703 3704 3705 3706 |
if (mem_cgroup_is_root(memcg)) { /* Can't set limit on root */ ret = -EINVAL; break; } |
628f42355 memcg: limit chan... |
3707 3708 |
/* This function does all necessary parse...reuse it */ ret = res_counter_memparse_write_strategy(buffer, &val); |
8c7c6e34a memcg: mem+swap c... |
3709 3710 3711 |
if (ret) break; if (type == _MEM) |
628f42355 memcg: limit chan... |
3712 |
ret = mem_cgroup_resize_limit(memcg, val); |
8c7c6e34a memcg: mem+swap c... |
3713 3714 |
else ret = mem_cgroup_resize_memsw_limit(memcg, val); |
628f42355 memcg: limit chan... |
3715 |
break; |
296c81d89 memory controller... |
3716 3717 3718 3719 3720 3721 3722 3723 3724 3725 3726 3727 3728 3729 |
case RES_SOFT_LIMIT: ret = res_counter_memparse_write_strategy(buffer, &val); if (ret) break; /* * For memsw, soft limits are hard to implement in terms * of semantics, for now, we support soft limits for * control without swap */ if (type == _MEM) ret = res_counter_set_soft_limit(&memcg->res, val); else ret = -EINVAL; break; |
628f42355 memcg: limit chan... |
3730 3731 3732 3733 3734 |
default: ret = -EINVAL; /* should be BUG() ? */ break; } return ret; |
8cdea7c05 Memory controller... |
3735 |
} |
fee7b548e memcg: show real ... |
3736 3737 3738 3739 3740 3741 3742 3743 3744 3745 3746 3747 3748 3749 3750 3751 3752 3753 3754 3755 3756 3757 3758 3759 3760 3761 3762 |
static void memcg_get_hierarchical_limit(struct mem_cgroup *memcg, unsigned long long *mem_limit, unsigned long long *memsw_limit) { struct cgroup *cgroup; unsigned long long min_limit, min_memsw_limit, tmp; min_limit = res_counter_read_u64(&memcg->res, RES_LIMIT); min_memsw_limit = res_counter_read_u64(&memcg->memsw, RES_LIMIT); cgroup = memcg->css.cgroup; if (!memcg->use_hierarchy) goto out; while (cgroup->parent) { cgroup = cgroup->parent; memcg = mem_cgroup_from_cont(cgroup); if (!memcg->use_hierarchy) break; tmp = res_counter_read_u64(&memcg->res, RES_LIMIT); min_limit = min(min_limit, tmp); tmp = res_counter_read_u64(&memcg->memsw, RES_LIMIT); min_memsw_limit = min(min_memsw_limit, tmp); } out: *mem_limit = min_limit; *memsw_limit = min_memsw_limit; return; } |
29f2a4dac memcgroup: implem... |
3763 |
static int mem_cgroup_reset(struct cgroup *cont, unsigned int event) |
c84872e16 memcgroup: add th... |
3764 |
{ |
c0ff4b854 memcg: rename mem... |
3765 |
struct mem_cgroup *memcg; |
8c7c6e34a memcg: mem+swap c... |
3766 |
int type, name; |
c84872e16 memcgroup: add th... |
3767 |
|
c0ff4b854 memcg: rename mem... |
3768 |
memcg = mem_cgroup_from_cont(cont); |
8c7c6e34a memcg: mem+swap c... |
3769 3770 3771 |
type = MEMFILE_TYPE(event); name = MEMFILE_ATTR(event); switch (name) { |
29f2a4dac memcgroup: implem... |
3772 |
case RES_MAX_USAGE: |
8c7c6e34a memcg: mem+swap c... |
3773 |
if (type == _MEM) |
c0ff4b854 memcg: rename mem... |
3774 |
res_counter_reset_max(&memcg->res); |
8c7c6e34a memcg: mem+swap c... |
3775 |
else |
c0ff4b854 memcg: rename mem... |
3776 |
res_counter_reset_max(&memcg->memsw); |
29f2a4dac memcgroup: implem... |
3777 3778 |
break; case RES_FAILCNT: |
8c7c6e34a memcg: mem+swap c... |
3779 |
if (type == _MEM) |
c0ff4b854 memcg: rename mem... |
3780 |
res_counter_reset_failcnt(&memcg->res); |
8c7c6e34a memcg: mem+swap c... |
3781 |
else |
c0ff4b854 memcg: rename mem... |
3782 |
res_counter_reset_failcnt(&memcg->memsw); |
29f2a4dac memcgroup: implem... |
3783 3784 |
break; } |
f64c3f549 memory controller... |
3785 |
|
85cc59db1 memcgroup: use tr... |
3786 |
return 0; |
c84872e16 memcgroup: add th... |
3787 |
} |
7dc74be03 memcg: add interf... |
3788 3789 3790 3791 3792 |
static u64 mem_cgroup_move_charge_read(struct cgroup *cgrp, struct cftype *cft) { return mem_cgroup_from_cont(cgrp)->move_charge_at_immigrate; } |
024914477 memcg: move charg... |
3793 |
#ifdef CONFIG_MMU |
7dc74be03 memcg: add interf... |
3794 3795 3796 |
static int mem_cgroup_move_charge_write(struct cgroup *cgrp, struct cftype *cft, u64 val) { |
c0ff4b854 memcg: rename mem... |
3797 |
struct mem_cgroup *memcg = mem_cgroup_from_cont(cgrp); |
7dc74be03 memcg: add interf... |
3798 3799 3800 3801 3802 3803 3804 3805 3806 |
if (val >= (1 << NR_MOVE_TYPE)) return -EINVAL; /* * We check this value several times in both in can_attach() and * attach(), so we need cgroup lock to prevent this value from being * inconsistent. */ cgroup_lock(); |
c0ff4b854 memcg: rename mem... |
3807 |
memcg->move_charge_at_immigrate = val; |
7dc74be03 memcg: add interf... |
3808 3809 3810 3811 |
cgroup_unlock(); return 0; } |
024914477 memcg: move charg... |
3812 3813 3814 3815 3816 3817 3818 |
#else static int mem_cgroup_move_charge_write(struct cgroup *cgrp, struct cftype *cft, u64 val) { return -ENOSYS; } #endif |
7dc74be03 memcg: add interf... |
3819 |
|
14067bb3e memcg: hierarchic... |
3820 3821 3822 3823 3824 |
/* For read statistics */ enum { MCS_CACHE, MCS_RSS, |
d8046582d memcg: make memcg... |
3825 |
MCS_FILE_MAPPED, |
14067bb3e memcg: hierarchic... |
3826 3827 |
MCS_PGPGIN, MCS_PGPGOUT, |
1dd3a2732 memcg: show swap ... |
3828 |
MCS_SWAP, |
456f998ec memcg: add the pa... |
3829 3830 |
MCS_PGFAULT, MCS_PGMAJFAULT, |
14067bb3e memcg: hierarchic... |
3831 3832 3833 3834 3835 3836 3837 3838 3839 3840 |
MCS_INACTIVE_ANON, MCS_ACTIVE_ANON, MCS_INACTIVE_FILE, MCS_ACTIVE_FILE, MCS_UNEVICTABLE, NR_MCS_STAT, }; struct mcs_total_stat { s64 stat[NR_MCS_STAT]; |
d2ceb9b7d memory cgroup enh... |
3841 |
}; |
14067bb3e memcg: hierarchic... |
3842 3843 3844 3845 3846 3847 |
struct { char *local_name; char *total_name; } memcg_stat_strings[NR_MCS_STAT] = { {"cache", "total_cache"}, {"rss", "total_rss"}, |
d69b042f3 memcg: add file-b... |
3848 |
{"mapped_file", "total_mapped_file"}, |
14067bb3e memcg: hierarchic... |
3849 3850 |
{"pgpgin", "total_pgpgin"}, {"pgpgout", "total_pgpgout"}, |
1dd3a2732 memcg: show swap ... |
3851 |
{"swap", "total_swap"}, |
456f998ec memcg: add the pa... |
3852 3853 |
{"pgfault", "total_pgfault"}, {"pgmajfault", "total_pgmajfault"}, |
14067bb3e memcg: hierarchic... |
3854 3855 3856 3857 3858 3859 |
{"inactive_anon", "total_inactive_anon"}, {"active_anon", "total_active_anon"}, {"inactive_file", "total_inactive_file"}, {"active_file", "total_active_file"}, {"unevictable", "total_unevictable"} }; |
7d74b06f2 memcg: use for_ea... |
3860 |
static void |
c0ff4b854 memcg: rename mem... |
3861 |
mem_cgroup_get_local_stat(struct mem_cgroup *memcg, struct mcs_total_stat *s) |
14067bb3e memcg: hierarchic... |
3862 |
{ |
14067bb3e memcg: hierarchic... |
3863 3864 3865 |
s64 val; /* per cpu stat */ |
c0ff4b854 memcg: rename mem... |
3866 |
val = mem_cgroup_read_stat(memcg, MEM_CGROUP_STAT_CACHE); |
14067bb3e memcg: hierarchic... |
3867 |
s->stat[MCS_CACHE] += val * PAGE_SIZE; |
c0ff4b854 memcg: rename mem... |
3868 |
val = mem_cgroup_read_stat(memcg, MEM_CGROUP_STAT_RSS); |
14067bb3e memcg: hierarchic... |
3869 |
s->stat[MCS_RSS] += val * PAGE_SIZE; |
c0ff4b854 memcg: rename mem... |
3870 |
val = mem_cgroup_read_stat(memcg, MEM_CGROUP_STAT_FILE_MAPPED); |
d8046582d memcg: make memcg... |
3871 |
s->stat[MCS_FILE_MAPPED] += val * PAGE_SIZE; |
c0ff4b854 memcg: rename mem... |
3872 |
val = mem_cgroup_read_events(memcg, MEM_CGROUP_EVENTS_PGPGIN); |
14067bb3e memcg: hierarchic... |
3873 |
s->stat[MCS_PGPGIN] += val; |
c0ff4b854 memcg: rename mem... |
3874 |
val = mem_cgroup_read_events(memcg, MEM_CGROUP_EVENTS_PGPGOUT); |
14067bb3e memcg: hierarchic... |
3875 |
s->stat[MCS_PGPGOUT] += val; |
1dd3a2732 memcg: show swap ... |
3876 |
if (do_swap_account) { |
c0ff4b854 memcg: rename mem... |
3877 |
val = mem_cgroup_read_stat(memcg, MEM_CGROUP_STAT_SWAPOUT); |
1dd3a2732 memcg: show swap ... |
3878 3879 |
s->stat[MCS_SWAP] += val * PAGE_SIZE; } |
c0ff4b854 memcg: rename mem... |
3880 |
val = mem_cgroup_read_events(memcg, MEM_CGROUP_EVENTS_PGFAULT); |
456f998ec memcg: add the pa... |
3881 |
s->stat[MCS_PGFAULT] += val; |
c0ff4b854 memcg: rename mem... |
3882 |
val = mem_cgroup_read_events(memcg, MEM_CGROUP_EVENTS_PGMAJFAULT); |
456f998ec memcg: add the pa... |
3883 |
s->stat[MCS_PGMAJFAULT] += val; |
14067bb3e memcg: hierarchic... |
3884 3885 |
/* per zone stat */ |
c0ff4b854 memcg: rename mem... |
3886 |
val = mem_cgroup_nr_lru_pages(memcg, BIT(LRU_INACTIVE_ANON)); |
14067bb3e memcg: hierarchic... |
3887 |
s->stat[MCS_INACTIVE_ANON] += val * PAGE_SIZE; |
c0ff4b854 memcg: rename mem... |
3888 |
val = mem_cgroup_nr_lru_pages(memcg, BIT(LRU_ACTIVE_ANON)); |
14067bb3e memcg: hierarchic... |
3889 |
s->stat[MCS_ACTIVE_ANON] += val * PAGE_SIZE; |
c0ff4b854 memcg: rename mem... |
3890 |
val = mem_cgroup_nr_lru_pages(memcg, BIT(LRU_INACTIVE_FILE)); |
14067bb3e memcg: hierarchic... |
3891 |
s->stat[MCS_INACTIVE_FILE] += val * PAGE_SIZE; |
c0ff4b854 memcg: rename mem... |
3892 |
val = mem_cgroup_nr_lru_pages(memcg, BIT(LRU_ACTIVE_FILE)); |
14067bb3e memcg: hierarchic... |
3893 |
s->stat[MCS_ACTIVE_FILE] += val * PAGE_SIZE; |
c0ff4b854 memcg: rename mem... |
3894 |
val = mem_cgroup_nr_lru_pages(memcg, BIT(LRU_UNEVICTABLE)); |
14067bb3e memcg: hierarchic... |
3895 |
s->stat[MCS_UNEVICTABLE] += val * PAGE_SIZE; |
14067bb3e memcg: hierarchic... |
3896 3897 3898 |
} static void |
c0ff4b854 memcg: rename mem... |
3899 |
mem_cgroup_get_total_stat(struct mem_cgroup *memcg, struct mcs_total_stat *s) |
14067bb3e memcg: hierarchic... |
3900 |
{ |
7d74b06f2 memcg: use for_ea... |
3901 |
struct mem_cgroup *iter; |
c0ff4b854 memcg: rename mem... |
3902 |
for_each_mem_cgroup_tree(iter, memcg) |
7d74b06f2 memcg: use for_ea... |
3903 |
mem_cgroup_get_local_stat(iter, s); |
14067bb3e memcg: hierarchic... |
3904 |
} |
406eb0c9b memcg: add memory... |
3905 3906 3907 3908 3909 3910 3911 3912 |
#ifdef CONFIG_NUMA static int mem_control_numa_stat_show(struct seq_file *m, void *arg) { int nid; unsigned long total_nr, file_nr, anon_nr, unevictable_nr; unsigned long node_nr; struct cgroup *cont = m->private; struct mem_cgroup *mem_cont = mem_cgroup_from_cont(cont); |
bb2a0de92 memcg: consolidat... |
3913 |
total_nr = mem_cgroup_nr_lru_pages(mem_cont, LRU_ALL); |
406eb0c9b memcg: add memory... |
3914 3915 |
seq_printf(m, "total=%lu", total_nr); for_each_node_state(nid, N_HIGH_MEMORY) { |
bb2a0de92 memcg: consolidat... |
3916 |
node_nr = mem_cgroup_node_nr_lru_pages(mem_cont, nid, LRU_ALL); |
406eb0c9b memcg: add memory... |
3917 3918 3919 3920 |
seq_printf(m, " N%d=%lu", nid, node_nr); } seq_putc(m, ' '); |
bb2a0de92 memcg: consolidat... |
3921 |
file_nr = mem_cgroup_nr_lru_pages(mem_cont, LRU_ALL_FILE); |
406eb0c9b memcg: add memory... |
3922 3923 |
seq_printf(m, "file=%lu", file_nr); for_each_node_state(nid, N_HIGH_MEMORY) { |
bb2a0de92 memcg: consolidat... |
3924 3925 |
node_nr = mem_cgroup_node_nr_lru_pages(mem_cont, nid, LRU_ALL_FILE); |
406eb0c9b memcg: add memory... |
3926 3927 3928 3929 |
seq_printf(m, " N%d=%lu", nid, node_nr); } seq_putc(m, ' '); |
bb2a0de92 memcg: consolidat... |
3930 |
anon_nr = mem_cgroup_nr_lru_pages(mem_cont, LRU_ALL_ANON); |
406eb0c9b memcg: add memory... |
3931 3932 |
seq_printf(m, "anon=%lu", anon_nr); for_each_node_state(nid, N_HIGH_MEMORY) { |
bb2a0de92 memcg: consolidat... |
3933 3934 |
node_nr = mem_cgroup_node_nr_lru_pages(mem_cont, nid, LRU_ALL_ANON); |
406eb0c9b memcg: add memory... |
3935 3936 3937 3938 |
seq_printf(m, " N%d=%lu", nid, node_nr); } seq_putc(m, ' '); |
bb2a0de92 memcg: consolidat... |
3939 |
unevictable_nr = mem_cgroup_nr_lru_pages(mem_cont, BIT(LRU_UNEVICTABLE)); |
406eb0c9b memcg: add memory... |
3940 3941 |
seq_printf(m, "unevictable=%lu", unevictable_nr); for_each_node_state(nid, N_HIGH_MEMORY) { |
bb2a0de92 memcg: consolidat... |
3942 3943 |
node_nr = mem_cgroup_node_nr_lru_pages(mem_cont, nid, BIT(LRU_UNEVICTABLE)); |
406eb0c9b memcg: add memory... |
3944 3945 3946 3947 3948 3949 3950 |
seq_printf(m, " N%d=%lu", nid, node_nr); } seq_putc(m, ' '); return 0; } #endif /* CONFIG_NUMA */ |
c64745cf0 CGroup API files:... |
3951 3952 |
static int mem_control_stat_show(struct cgroup *cont, struct cftype *cft, struct cgroup_map_cb *cb) |
d2ceb9b7d memory cgroup enh... |
3953 |
{ |
d2ceb9b7d memory cgroup enh... |
3954 |
struct mem_cgroup *mem_cont = mem_cgroup_from_cont(cont); |
14067bb3e memcg: hierarchic... |
3955 |
struct mcs_total_stat mystat; |
d2ceb9b7d memory cgroup enh... |
3956 |
int i; |
14067bb3e memcg: hierarchic... |
3957 3958 |
memset(&mystat, 0, sizeof(mystat)); mem_cgroup_get_local_stat(mem_cont, &mystat); |
d2ceb9b7d memory cgroup enh... |
3959 |
|
406eb0c9b memcg: add memory... |
3960 |
|
1dd3a2732 memcg: show swap ... |
3961 3962 3963 |
for (i = 0; i < NR_MCS_STAT; i++) { if (i == MCS_SWAP && !do_swap_account) continue; |
14067bb3e memcg: hierarchic... |
3964 |
cb->fill(cb, memcg_stat_strings[i].local_name, mystat.stat[i]); |
1dd3a2732 memcg: show swap ... |
3965 |
} |
7b854121e Unevictable LRU P... |
3966 |
|
14067bb3e memcg: hierarchic... |
3967 |
/* Hierarchical information */ |
fee7b548e memcg: show real ... |
3968 3969 3970 3971 3972 3973 3974 |
{ unsigned long long limit, memsw_limit; memcg_get_hierarchical_limit(mem_cont, &limit, &memsw_limit); cb->fill(cb, "hierarchical_memory_limit", limit); if (do_swap_account) cb->fill(cb, "hierarchical_memsw_limit", memsw_limit); } |
7f016ee8b memcg: show recla... |
3975 |
|
14067bb3e memcg: hierarchic... |
3976 3977 |
memset(&mystat, 0, sizeof(mystat)); mem_cgroup_get_total_stat(mem_cont, &mystat); |
1dd3a2732 memcg: show swap ... |
3978 3979 3980 |
for (i = 0; i < NR_MCS_STAT; i++) { if (i == MCS_SWAP && !do_swap_account) continue; |
14067bb3e memcg: hierarchic... |
3981 |
cb->fill(cb, memcg_stat_strings[i].total_name, mystat.stat[i]); |
1dd3a2732 memcg: show swap ... |
3982 |
} |
14067bb3e memcg: hierarchic... |
3983 |
|
7f016ee8b memcg: show recla... |
3984 |
#ifdef CONFIG_DEBUG_VM |
7f016ee8b memcg: show recla... |
3985 3986 3987 3988 3989 3990 3991 3992 3993 3994 3995 3996 3997 3998 3999 4000 4001 4002 4003 4004 4005 4006 4007 4008 4009 |
{ int nid, zid; struct mem_cgroup_per_zone *mz; unsigned long recent_rotated[2] = {0, 0}; unsigned long recent_scanned[2] = {0, 0}; for_each_online_node(nid) for (zid = 0; zid < MAX_NR_ZONES; zid++) { mz = mem_cgroup_zoneinfo(mem_cont, nid, zid); recent_rotated[0] += mz->reclaim_stat.recent_rotated[0]; recent_rotated[1] += mz->reclaim_stat.recent_rotated[1]; recent_scanned[0] += mz->reclaim_stat.recent_scanned[0]; recent_scanned[1] += mz->reclaim_stat.recent_scanned[1]; } cb->fill(cb, "recent_rotated_anon", recent_rotated[0]); cb->fill(cb, "recent_rotated_file", recent_rotated[1]); cb->fill(cb, "recent_scanned_anon", recent_scanned[0]); cb->fill(cb, "recent_scanned_file", recent_scanned[1]); } #endif |
d2ceb9b7d memory cgroup enh... |
4010 4011 |
return 0; } |
a7885eb8a memcg: swappiness |
4012 4013 4014 |
static u64 mem_cgroup_swappiness_read(struct cgroup *cgrp, struct cftype *cft) { struct mem_cgroup *memcg = mem_cgroup_from_cont(cgrp); |
1f4c025b5 memcg: export mem... |
4015 |
return mem_cgroup_swappiness(memcg); |
a7885eb8a memcg: swappiness |
4016 4017 4018 4019 4020 4021 4022 |
} static int mem_cgroup_swappiness_write(struct cgroup *cgrp, struct cftype *cft, u64 val) { struct mem_cgroup *memcg = mem_cgroup_from_cont(cgrp); struct mem_cgroup *parent; |
068b38c1f memcg: fix a race... |
4023 |
|
a7885eb8a memcg: swappiness |
4024 4025 4026 4027 4028 4029 4030 |
if (val > 100) return -EINVAL; if (cgrp->parent == NULL) return -EINVAL; parent = mem_cgroup_from_cont(cgrp->parent); |
068b38c1f memcg: fix a race... |
4031 4032 |
cgroup_lock(); |
a7885eb8a memcg: swappiness |
4033 4034 |
/* If under hierarchy, only empty-root can set this value */ if ((parent->use_hierarchy) || |
068b38c1f memcg: fix a race... |
4035 4036 |
(memcg->use_hierarchy && !list_empty(&cgrp->children))) { cgroup_unlock(); |
a7885eb8a memcg: swappiness |
4037 |
return -EINVAL; |
068b38c1f memcg: fix a race... |
4038 |
} |
a7885eb8a memcg: swappiness |
4039 |
|
a7885eb8a memcg: swappiness |
4040 |
memcg->swappiness = val; |
a7885eb8a memcg: swappiness |
4041 |
|
068b38c1f memcg: fix a race... |
4042 |
cgroup_unlock(); |
a7885eb8a memcg: swappiness |
4043 4044 |
return 0; } |
2e72b6347 memcg: implement ... |
4045 4046 4047 4048 4049 4050 4051 4052 |
static void __mem_cgroup_threshold(struct mem_cgroup *memcg, bool swap) { struct mem_cgroup_threshold_ary *t; u64 usage; int i; rcu_read_lock(); if (!swap) |
2c488db27 memcg: clean up m... |
4053 |
t = rcu_dereference(memcg->thresholds.primary); |
2e72b6347 memcg: implement ... |
4054 |
else |
2c488db27 memcg: clean up m... |
4055 |
t = rcu_dereference(memcg->memsw_thresholds.primary); |
2e72b6347 memcg: implement ... |
4056 4057 4058 4059 4060 4061 4062 4063 4064 4065 4066 |
if (!t) goto unlock; usage = mem_cgroup_usage(memcg, swap); /* * current_threshold points to threshold just below usage. * If it's not true, a threshold was crossed after last * call of __mem_cgroup_threshold(). */ |
5407a5625 mm: remove unnece... |
4067 |
i = t->current_threshold; |
2e72b6347 memcg: implement ... |
4068 4069 4070 4071 4072 4073 4074 4075 4076 4077 4078 4079 4080 4081 4082 4083 4084 4085 4086 4087 4088 4089 4090 |
/* * 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... |
4091 |
t->current_threshold = i - 1; |
2e72b6347 memcg: implement ... |
4092 4093 4094 4095 4096 4097 |
unlock: rcu_read_unlock(); } static void mem_cgroup_threshold(struct mem_cgroup *memcg) { |
ad4ca5f4b memcg: fix thresh... |
4098 4099 4100 4101 4102 4103 4104 |
while (memcg) { __mem_cgroup_threshold(memcg, false); if (do_swap_account) __mem_cgroup_threshold(memcg, true); memcg = parent_mem_cgroup(memcg); } |
2e72b6347 memcg: implement ... |
4105 4106 4107 4108 4109 4110 4111 4112 4113 |
} static int compare_thresholds(const void *a, const void *b) { const struct mem_cgroup_threshold *_a = a; const struct mem_cgroup_threshold *_b = b; return _a->threshold - _b->threshold; } |
c0ff4b854 memcg: rename mem... |
4114 |
static int mem_cgroup_oom_notify_cb(struct mem_cgroup *memcg) |
9490ff275 memcg: oom notifier |
4115 4116 |
{ struct mem_cgroup_eventfd_list *ev; |
c0ff4b854 memcg: rename mem... |
4117 |
list_for_each_entry(ev, &memcg->oom_notify, list) |
9490ff275 memcg: oom notifier |
4118 4119 4120 |
eventfd_signal(ev->eventfd, 1); return 0; } |
c0ff4b854 memcg: rename mem... |
4121 |
static void mem_cgroup_oom_notify(struct mem_cgroup *memcg) |
9490ff275 memcg: oom notifier |
4122 |
{ |
7d74b06f2 memcg: use for_ea... |
4123 |
struct mem_cgroup *iter; |
c0ff4b854 memcg: rename mem... |
4124 |
for_each_mem_cgroup_tree(iter, memcg) |
7d74b06f2 memcg: use for_ea... |
4125 |
mem_cgroup_oom_notify_cb(iter); |
9490ff275 memcg: oom notifier |
4126 4127 4128 4129 |
} static int mem_cgroup_usage_register_event(struct cgroup *cgrp, struct cftype *cft, struct eventfd_ctx *eventfd, const char *args) |
2e72b6347 memcg: implement ... |
4130 4131 |
{ struct mem_cgroup *memcg = mem_cgroup_from_cont(cgrp); |
2c488db27 memcg: clean up m... |
4132 4133 |
struct mem_cgroup_thresholds *thresholds; struct mem_cgroup_threshold_ary *new; |
2e72b6347 memcg: implement ... |
4134 4135 |
int type = MEMFILE_TYPE(cft->private); u64 threshold, usage; |
2c488db27 memcg: clean up m... |
4136 |
int i, size, ret; |
2e72b6347 memcg: implement ... |
4137 4138 4139 4140 4141 4142 |
ret = res_counter_memparse_write_strategy(args, &threshold); if (ret) return ret; mutex_lock(&memcg->thresholds_lock); |
2c488db27 memcg: clean up m... |
4143 |
|
2e72b6347 memcg: implement ... |
4144 |
if (type == _MEM) |
2c488db27 memcg: clean up m... |
4145 |
thresholds = &memcg->thresholds; |
2e72b6347 memcg: implement ... |
4146 |
else if (type == _MEMSWAP) |
2c488db27 memcg: clean up m... |
4147 |
thresholds = &memcg->memsw_thresholds; |
2e72b6347 memcg: implement ... |
4148 4149 4150 4151 4152 4153 |
else BUG(); usage = mem_cgroup_usage(memcg, type == _MEMSWAP); /* Check if a threshold crossed before adding a new one */ |
2c488db27 memcg: clean up m... |
4154 |
if (thresholds->primary) |
2e72b6347 memcg: implement ... |
4155 |
__mem_cgroup_threshold(memcg, type == _MEMSWAP); |
2c488db27 memcg: clean up m... |
4156 |
size = thresholds->primary ? thresholds->primary->size + 1 : 1; |
2e72b6347 memcg: implement ... |
4157 4158 |
/* Allocate memory for new array of thresholds */ |
2c488db27 memcg: clean up m... |
4159 |
new = kmalloc(sizeof(*new) + size * sizeof(struct mem_cgroup_threshold), |
2e72b6347 memcg: implement ... |
4160 |
GFP_KERNEL); |
2c488db27 memcg: clean up m... |
4161 |
if (!new) { |
2e72b6347 memcg: implement ... |
4162 4163 4164 |
ret = -ENOMEM; goto unlock; } |
2c488db27 memcg: clean up m... |
4165 |
new->size = size; |
2e72b6347 memcg: implement ... |
4166 4167 |
/* Copy thresholds (if any) to new array */ |
2c488db27 memcg: clean up m... |
4168 4169 |
if (thresholds->primary) { memcpy(new->entries, thresholds->primary->entries, (size - 1) * |
2e72b6347 memcg: implement ... |
4170 |
sizeof(struct mem_cgroup_threshold)); |
2c488db27 memcg: clean up m... |
4171 |
} |
2e72b6347 memcg: implement ... |
4172 |
/* Add new threshold */ |
2c488db27 memcg: clean up m... |
4173 4174 |
new->entries[size - 1].eventfd = eventfd; new->entries[size - 1].threshold = threshold; |
2e72b6347 memcg: implement ... |
4175 4176 |
/* Sort thresholds. Registering of new threshold isn't time-critical */ |
2c488db27 memcg: clean up m... |
4177 |
sort(new->entries, size, sizeof(struct mem_cgroup_threshold), |
2e72b6347 memcg: implement ... |
4178 4179 4180 |
compare_thresholds, NULL); /* Find current threshold */ |
2c488db27 memcg: clean up m... |
4181 |
new->current_threshold = -1; |
2e72b6347 memcg: implement ... |
4182 |
for (i = 0; i < size; i++) { |
2c488db27 memcg: clean up m... |
4183 |
if (new->entries[i].threshold < usage) { |
2e72b6347 memcg: implement ... |
4184 |
/* |
2c488db27 memcg: clean up m... |
4185 4186 |
* new->current_threshold will not be used until * rcu_assign_pointer(), so it's safe to increment |
2e72b6347 memcg: implement ... |
4187 4188 |
* it here. */ |
2c488db27 memcg: clean up m... |
4189 |
++new->current_threshold; |
2e72b6347 memcg: implement ... |
4190 4191 |
} } |
2c488db27 memcg: clean up m... |
4192 4193 4194 4195 4196 |
/* 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 ... |
4197 |
|
907860ed3 cgroups: make cft... |
4198 |
/* To be sure that nobody uses thresholds */ |
2e72b6347 memcg: implement ... |
4199 |
synchronize_rcu(); |
2e72b6347 memcg: implement ... |
4200 4201 4202 4203 4204 |
unlock: mutex_unlock(&memcg->thresholds_lock); return ret; } |
907860ed3 cgroups: make cft... |
4205 |
static void mem_cgroup_usage_unregister_event(struct cgroup *cgrp, |
9490ff275 memcg: oom notifier |
4206 |
struct cftype *cft, struct eventfd_ctx *eventfd) |
2e72b6347 memcg: implement ... |
4207 4208 |
{ struct mem_cgroup *memcg = mem_cgroup_from_cont(cgrp); |
2c488db27 memcg: clean up m... |
4209 4210 |
struct mem_cgroup_thresholds *thresholds; struct mem_cgroup_threshold_ary *new; |
2e72b6347 memcg: implement ... |
4211 4212 |
int type = MEMFILE_TYPE(cft->private); u64 usage; |
2c488db27 memcg: clean up m... |
4213 |
int i, j, size; |
2e72b6347 memcg: implement ... |
4214 4215 4216 |
mutex_lock(&memcg->thresholds_lock); if (type == _MEM) |
2c488db27 memcg: clean up m... |
4217 |
thresholds = &memcg->thresholds; |
2e72b6347 memcg: implement ... |
4218 |
else if (type == _MEMSWAP) |
2c488db27 memcg: clean up m... |
4219 |
thresholds = &memcg->memsw_thresholds; |
2e72b6347 memcg: implement ... |
4220 4221 4222 4223 4224 4225 4226 4227 4228 4229 4230 4231 4232 4233 4234 |
else BUG(); /* * Something went wrong if we trying to unregister a threshold * if we don't have thresholds */ BUG_ON(!thresholds); usage = mem_cgroup_usage(memcg, type == _MEMSWAP); /* Check if a threshold crossed before removing */ __mem_cgroup_threshold(memcg, type == _MEMSWAP); /* Calculate new number of threshold */ |
2c488db27 memcg: clean up m... |
4235 4236 4237 |
size = 0; for (i = 0; i < thresholds->primary->size; i++) { if (thresholds->primary->entries[i].eventfd != eventfd) |
2e72b6347 memcg: implement ... |
4238 4239 |
size++; } |
2c488db27 memcg: clean up m... |
4240 |
new = thresholds->spare; |
907860ed3 cgroups: make cft... |
4241 |
|
2e72b6347 memcg: implement ... |
4242 4243 |
/* Set thresholds array to NULL if we don't have thresholds */ if (!size) { |
2c488db27 memcg: clean up m... |
4244 4245 |
kfree(new); new = NULL; |
907860ed3 cgroups: make cft... |
4246 |
goto swap_buffers; |
2e72b6347 memcg: implement ... |
4247 |
} |
2c488db27 memcg: clean up m... |
4248 |
new->size = size; |
2e72b6347 memcg: implement ... |
4249 4250 |
/* Copy thresholds and find current threshold */ |
2c488db27 memcg: clean up m... |
4251 4252 4253 |
new->current_threshold = -1; for (i = 0, j = 0; i < thresholds->primary->size; i++) { if (thresholds->primary->entries[i].eventfd == eventfd) |
2e72b6347 memcg: implement ... |
4254 |
continue; |
2c488db27 memcg: clean up m... |
4255 4256 |
new->entries[j] = thresholds->primary->entries[i]; if (new->entries[j].threshold < usage) { |
2e72b6347 memcg: implement ... |
4257 |
/* |
2c488db27 memcg: clean up m... |
4258 |
* new->current_threshold will not be used |
2e72b6347 memcg: implement ... |
4259 4260 4261 |
* until rcu_assign_pointer(), so it's safe to increment * it here. */ |
2c488db27 memcg: clean up m... |
4262 |
++new->current_threshold; |
2e72b6347 memcg: implement ... |
4263 4264 4265 |
} j++; } |
907860ed3 cgroups: make cft... |
4266 |
swap_buffers: |
2c488db27 memcg: clean up m... |
4267 4268 4269 |
/* Swap primary and spare array */ thresholds->spare = thresholds->primary; rcu_assign_pointer(thresholds->primary, new); |
2e72b6347 memcg: implement ... |
4270 |
|
907860ed3 cgroups: make cft... |
4271 |
/* To be sure that nobody uses thresholds */ |
2e72b6347 memcg: implement ... |
4272 |
synchronize_rcu(); |
2e72b6347 memcg: implement ... |
4273 |
mutex_unlock(&memcg->thresholds_lock); |
2e72b6347 memcg: implement ... |
4274 |
} |
c1e862c1f memcg: new force_... |
4275 |
|
9490ff275 memcg: oom notifier |
4276 4277 4278 4279 4280 4281 4282 4283 4284 4285 4286 |
static int mem_cgroup_oom_register_event(struct cgroup *cgrp, struct cftype *cft, struct eventfd_ctx *eventfd, const char *args) { struct mem_cgroup *memcg = mem_cgroup_from_cont(cgrp); struct mem_cgroup_eventfd_list *event; int type = MEMFILE_TYPE(cft->private); BUG_ON(type != _OOM_TYPE); event = kmalloc(sizeof(*event), GFP_KERNEL); if (!event) return -ENOMEM; |
1af8efe96 memcg: change mem... |
4287 |
spin_lock(&memcg_oom_lock); |
9490ff275 memcg: oom notifier |
4288 4289 4290 4291 4292 |
event->eventfd = eventfd; list_add(&event->list, &memcg->oom_notify); /* already in OOM ? */ |
79dfdaccd memcg: make oom_l... |
4293 |
if (atomic_read(&memcg->under_oom)) |
9490ff275 memcg: oom notifier |
4294 |
eventfd_signal(eventfd, 1); |
1af8efe96 memcg: change mem... |
4295 |
spin_unlock(&memcg_oom_lock); |
9490ff275 memcg: oom notifier |
4296 4297 4298 |
return 0; } |
907860ed3 cgroups: make cft... |
4299 |
static void mem_cgroup_oom_unregister_event(struct cgroup *cgrp, |
9490ff275 memcg: oom notifier |
4300 4301 |
struct cftype *cft, struct eventfd_ctx *eventfd) { |
c0ff4b854 memcg: rename mem... |
4302 |
struct mem_cgroup *memcg = mem_cgroup_from_cont(cgrp); |
9490ff275 memcg: oom notifier |
4303 4304 4305 4306 |
struct mem_cgroup_eventfd_list *ev, *tmp; int type = MEMFILE_TYPE(cft->private); BUG_ON(type != _OOM_TYPE); |
1af8efe96 memcg: change mem... |
4307 |
spin_lock(&memcg_oom_lock); |
9490ff275 memcg: oom notifier |
4308 |
|
c0ff4b854 memcg: rename mem... |
4309 |
list_for_each_entry_safe(ev, tmp, &memcg->oom_notify, list) { |
9490ff275 memcg: oom notifier |
4310 4311 4312 4313 4314 |
if (ev->eventfd == eventfd) { list_del(&ev->list); kfree(ev); } } |
1af8efe96 memcg: change mem... |
4315 |
spin_unlock(&memcg_oom_lock); |
9490ff275 memcg: oom notifier |
4316 |
} |
3c11ecf44 memcg: oom kill d... |
4317 4318 4319 |
static int mem_cgroup_oom_control_read(struct cgroup *cgrp, struct cftype *cft, struct cgroup_map_cb *cb) { |
c0ff4b854 memcg: rename mem... |
4320 |
struct mem_cgroup *memcg = mem_cgroup_from_cont(cgrp); |
3c11ecf44 memcg: oom kill d... |
4321 |
|
c0ff4b854 memcg: rename mem... |
4322 |
cb->fill(cb, "oom_kill_disable", memcg->oom_kill_disable); |
3c11ecf44 memcg: oom kill d... |
4323 |
|
c0ff4b854 memcg: rename mem... |
4324 |
if (atomic_read(&memcg->under_oom)) |
3c11ecf44 memcg: oom kill d... |
4325 4326 4327 4328 4329 |
cb->fill(cb, "under_oom", 1); else cb->fill(cb, "under_oom", 0); return 0; } |
3c11ecf44 memcg: oom kill d... |
4330 4331 4332 |
static int mem_cgroup_oom_control_write(struct cgroup *cgrp, struct cftype *cft, u64 val) { |
c0ff4b854 memcg: rename mem... |
4333 |
struct mem_cgroup *memcg = mem_cgroup_from_cont(cgrp); |
3c11ecf44 memcg: oom kill d... |
4334 4335 4336 4337 4338 4339 4340 4341 4342 4343 4344 |
struct mem_cgroup *parent; /* cannot set to root cgroup and only 0 and 1 are allowed */ if (!cgrp->parent || !((val == 0) || (val == 1))) return -EINVAL; parent = mem_cgroup_from_cont(cgrp->parent); cgroup_lock(); /* oom-kill-disable is a flag for subhierarchy. */ if ((parent->use_hierarchy) || |
c0ff4b854 memcg: rename mem... |
4345 |
(memcg->use_hierarchy && !list_empty(&cgrp->children))) { |
3c11ecf44 memcg: oom kill d... |
4346 4347 4348 |
cgroup_unlock(); return -EINVAL; } |
c0ff4b854 memcg: rename mem... |
4349 |
memcg->oom_kill_disable = val; |
4d845ebf4 memcg: fix wake u... |
4350 |
if (!val) |
c0ff4b854 memcg: rename mem... |
4351 |
memcg_oom_recover(memcg); |
3c11ecf44 memcg: oom kill d... |
4352 4353 4354 |
cgroup_unlock(); return 0; } |
406eb0c9b memcg: add memory... |
4355 4356 4357 4358 4359 4360 4361 4362 4363 4364 4365 4366 4367 4368 4369 |
#ifdef CONFIG_NUMA static const struct file_operations mem_control_numa_stat_file_operations = { .read = seq_read, .llseek = seq_lseek, .release = single_release, }; static int mem_control_numa_stat_open(struct inode *unused, struct file *file) { struct cgroup *cont = file->f_dentry->d_parent->d_fsdata; file->f_op = &mem_control_numa_stat_file_operations; return single_open(file, mem_control_numa_stat_show, cont); } #endif /* CONFIG_NUMA */ |
e5671dfae Basic kernel memo... |
4370 |
#ifdef CONFIG_CGROUP_MEM_RES_CTLR_KMEM |
e5671dfae Basic kernel memo... |
4371 4372 |
static int register_kmem_files(struct cgroup *cont, struct cgroup_subsys *ss) { |
d1a4c0b37 tcp memory pressu... |
4373 4374 4375 4376 4377 4378 4379 |
/* * Part of this would be better living in a separate allocation * function, leaving us with just the cgroup tree population work. * We, however, depend on state such as network's proto_list that * is only initialized after cgroup creation. I found the less * cumbersome way to deal with it to defer it all to populate time */ |
65c64ce8e Partial revert "B... |
4380 |
return mem_cgroup_sockets_init(cont, ss); |
e5671dfae Basic kernel memo... |
4381 |
}; |
d1a4c0b37 tcp memory pressu... |
4382 4383 4384 4385 4386 |
static void kmem_cgroup_destroy(struct cgroup_subsys *ss, struct cgroup *cont) { mem_cgroup_sockets_destroy(cont, ss); } |
e5671dfae Basic kernel memo... |
4387 4388 4389 4390 4391 |
#else static int register_kmem_files(struct cgroup *cont, struct cgroup_subsys *ss) { return 0; } |
d1a4c0b37 tcp memory pressu... |
4392 4393 4394 4395 4396 |
static void kmem_cgroup_destroy(struct cgroup_subsys *ss, struct cgroup *cont) { } |
e5671dfae Basic kernel memo... |
4397 |
#endif |
8cdea7c05 Memory controller... |
4398 4399 |
static struct cftype mem_cgroup_files[] = { { |
0eea10301 Memory controller... |
4400 |
.name = "usage_in_bytes", |
8c7c6e34a memcg: mem+swap c... |
4401 |
.private = MEMFILE_PRIVATE(_MEM, RES_USAGE), |
2c3daa722 CGroup API files:... |
4402 |
.read_u64 = mem_cgroup_read, |
9490ff275 memcg: oom notifier |
4403 4404 |
.register_event = mem_cgroup_usage_register_event, .unregister_event = mem_cgroup_usage_unregister_event, |
8cdea7c05 Memory controller... |
4405 4406 |
}, { |
c84872e16 memcgroup: add th... |
4407 |
.name = "max_usage_in_bytes", |
8c7c6e34a memcg: mem+swap c... |
4408 |
.private = MEMFILE_PRIVATE(_MEM, RES_MAX_USAGE), |
29f2a4dac memcgroup: implem... |
4409 |
.trigger = mem_cgroup_reset, |
c84872e16 memcgroup: add th... |
4410 4411 4412 |
.read_u64 = mem_cgroup_read, }, { |
0eea10301 Memory controller... |
4413 |
.name = "limit_in_bytes", |
8c7c6e34a memcg: mem+swap c... |
4414 |
.private = MEMFILE_PRIVATE(_MEM, RES_LIMIT), |
856c13aa1 cgroup files: con... |
4415 |
.write_string = mem_cgroup_write, |
2c3daa722 CGroup API files:... |
4416 |
.read_u64 = mem_cgroup_read, |
8cdea7c05 Memory controller... |
4417 4418 |
}, { |
296c81d89 memory controller... |
4419 4420 4421 4422 4423 4424 |
.name = "soft_limit_in_bytes", .private = MEMFILE_PRIVATE(_MEM, RES_SOFT_LIMIT), .write_string = mem_cgroup_write, .read_u64 = mem_cgroup_read, }, { |
8cdea7c05 Memory controller... |
4425 |
.name = "failcnt", |
8c7c6e34a memcg: mem+swap c... |
4426 |
.private = MEMFILE_PRIVATE(_MEM, RES_FAILCNT), |
29f2a4dac memcgroup: implem... |
4427 |
.trigger = mem_cgroup_reset, |
2c3daa722 CGroup API files:... |
4428 |
.read_u64 = mem_cgroup_read, |
8cdea7c05 Memory controller... |
4429 |
}, |
8697d3319 Memory controller... |
4430 |
{ |
d2ceb9b7d memory cgroup enh... |
4431 |
.name = "stat", |
c64745cf0 CGroup API files:... |
4432 |
.read_map = mem_control_stat_show, |
d2ceb9b7d memory cgroup enh... |
4433 |
}, |
c1e862c1f memcg: new force_... |
4434 4435 4436 4437 |
{ .name = "force_empty", .trigger = mem_cgroup_force_empty_write, }, |
18f59ea7d memcg: memory cgr... |
4438 4439 4440 4441 4442 |
{ .name = "use_hierarchy", .write_u64 = mem_cgroup_hierarchy_write, .read_u64 = mem_cgroup_hierarchy_read, }, |
a7885eb8a memcg: swappiness |
4443 4444 4445 4446 4447 |
{ .name = "swappiness", .read_u64 = mem_cgroup_swappiness_read, .write_u64 = mem_cgroup_swappiness_write, }, |
7dc74be03 memcg: add interf... |
4448 4449 4450 4451 4452 |
{ .name = "move_charge_at_immigrate", .read_u64 = mem_cgroup_move_charge_read, .write_u64 = mem_cgroup_move_charge_write, }, |
9490ff275 memcg: oom notifier |
4453 4454 |
{ .name = "oom_control", |
3c11ecf44 memcg: oom kill d... |
4455 4456 |
.read_map = mem_cgroup_oom_control_read, .write_u64 = mem_cgroup_oom_control_write, |
9490ff275 memcg: oom notifier |
4457 4458 4459 4460 |
.register_event = mem_cgroup_oom_register_event, .unregister_event = mem_cgroup_oom_unregister_event, .private = MEMFILE_PRIVATE(_OOM_TYPE, OOM_CONTROL), }, |
406eb0c9b memcg: add memory... |
4461 4462 4463 4464 |
#ifdef CONFIG_NUMA { .name = "numa_stat", .open = mem_control_numa_stat_open, |
895771271 mm: memory.numa_s... |
4465 |
.mode = S_IRUGO, |
406eb0c9b memcg: add memory... |
4466 4467 |
}, #endif |
8cdea7c05 Memory controller... |
4468 |
}; |
8c7c6e34a memcg: mem+swap c... |
4469 4470 4471 4472 4473 4474 |
#ifdef CONFIG_CGROUP_MEM_RES_CTLR_SWAP static struct cftype memsw_cgroup_files[] = { { .name = "memsw.usage_in_bytes", .private = MEMFILE_PRIVATE(_MEMSWAP, RES_USAGE), .read_u64 = mem_cgroup_read, |
9490ff275 memcg: oom notifier |
4475 4476 |
.register_event = mem_cgroup_usage_register_event, .unregister_event = mem_cgroup_usage_unregister_event, |
8c7c6e34a memcg: mem+swap c... |
4477 4478 4479 4480 4481 4482 4483 4484 4485 4486 4487 4488 4489 4490 4491 4492 4493 4494 4495 4496 4497 4498 4499 4500 4501 4502 4503 4504 4505 4506 4507 4508 4509 4510 |
}, { .name = "memsw.max_usage_in_bytes", .private = MEMFILE_PRIVATE(_MEMSWAP, RES_MAX_USAGE), .trigger = mem_cgroup_reset, .read_u64 = mem_cgroup_read, }, { .name = "memsw.limit_in_bytes", .private = MEMFILE_PRIVATE(_MEMSWAP, RES_LIMIT), .write_string = mem_cgroup_write, .read_u64 = mem_cgroup_read, }, { .name = "memsw.failcnt", .private = MEMFILE_PRIVATE(_MEMSWAP, RES_FAILCNT), .trigger = mem_cgroup_reset, .read_u64 = mem_cgroup_read, }, }; static int register_memsw_files(struct cgroup *cont, struct cgroup_subsys *ss) { if (!do_swap_account) return 0; return cgroup_add_files(cont, ss, memsw_cgroup_files, ARRAY_SIZE(memsw_cgroup_files)); }; #else static int register_memsw_files(struct cgroup *cont, struct cgroup_subsys *ss) { return 0; } #endif |
c0ff4b854 memcg: rename mem... |
4511 |
static int alloc_mem_cgroup_per_zone_info(struct mem_cgroup *memcg, int node) |
6d12e2d8d per-zone and recl... |
4512 4513 |
{ struct mem_cgroup_per_node *pn; |
1ecaab2bd per-zone and recl... |
4514 |
struct mem_cgroup_per_zone *mz; |
b69408e88 vmscan: Use an in... |
4515 |
enum lru_list l; |
41e3355de memcg: fix node_s... |
4516 |
int zone, tmp = node; |
1ecaab2bd per-zone and recl... |
4517 4518 4519 4520 4521 4522 4523 4524 |
/* * 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... |
4525 4526 |
if (!node_state(node, N_NORMAL_MEMORY)) tmp = -1; |
17295c88a memcg: use [kv]za... |
4527 |
pn = kzalloc_node(sizeof(*pn), GFP_KERNEL, tmp); |
6d12e2d8d per-zone and recl... |
4528 4529 |
if (!pn) return 1; |
1ecaab2bd per-zone and recl... |
4530 |
|
1ecaab2bd per-zone and recl... |
4531 4532 |
for (zone = 0; zone < MAX_NR_ZONES; zone++) { mz = &pn->zoneinfo[zone]; |
b69408e88 vmscan: Use an in... |
4533 |
for_each_lru(l) |
6290df545 mm: collect LRU l... |
4534 |
INIT_LIST_HEAD(&mz->lruvec.lists[l]); |
f64c3f549 memory controller... |
4535 |
mz->usage_in_excess = 0; |
4e4169535 memory controller... |
4536 |
mz->on_tree = false; |
c0ff4b854 memcg: rename mem... |
4537 |
mz->mem = memcg; |
1ecaab2bd per-zone and recl... |
4538 |
} |
0a619e587 memcg: do not exp... |
4539 |
memcg->info.nodeinfo[node] = pn; |
6d12e2d8d per-zone and recl... |
4540 4541 |
return 0; } |
c0ff4b854 memcg: rename mem... |
4542 |
static void free_mem_cgroup_per_zone_info(struct mem_cgroup *memcg, int node) |
1ecaab2bd per-zone and recl... |
4543 |
{ |
c0ff4b854 memcg: rename mem... |
4544 |
kfree(memcg->info.nodeinfo[node]); |
1ecaab2bd per-zone and recl... |
4545 |
} |
333279487 memcgroup: use vm... |
4546 4547 4548 |
static struct mem_cgroup *mem_cgroup_alloc(void) { struct mem_cgroup *mem; |
c62b1a3b3 memcg: use generi... |
4549 |
int size = sizeof(struct mem_cgroup); |
333279487 memcgroup: use vm... |
4550 |
|
c62b1a3b3 memcg: use generi... |
4551 |
/* Can be very big if MAX_NUMNODES is very big */ |
c8dad2bb6 memcg: reduce siz... |
4552 |
if (size < PAGE_SIZE) |
17295c88a memcg: use [kv]za... |
4553 |
mem = kzalloc(size, GFP_KERNEL); |
333279487 memcgroup: use vm... |
4554 |
else |
17295c88a memcg: use [kv]za... |
4555 |
mem = vzalloc(size); |
333279487 memcgroup: use vm... |
4556 |
|
e7bbcdf37 memcontrol: fix p... |
4557 4558 |
if (!mem) return NULL; |
c62b1a3b3 memcg: use generi... |
4559 |
mem->stat = alloc_percpu(struct mem_cgroup_stat_cpu); |
d2e61b8dc memcg: null deref... |
4560 4561 |
if (!mem->stat) goto out_free; |
711d3d2c9 memcg: cpu hotplu... |
4562 |
spin_lock_init(&mem->pcp_counter_lock); |
333279487 memcgroup: use vm... |
4563 |
return mem; |
d2e61b8dc memcg: null deref... |
4564 4565 4566 4567 4568 4569 4570 |
out_free: if (size < PAGE_SIZE) kfree(mem); else vfree(mem); return NULL; |
333279487 memcgroup: use vm... |
4571 |
} |
8c7c6e34a memcg: mem+swap c... |
4572 4573 4574 4575 4576 4577 4578 4579 |
/* * 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. * |
8c7c6e34a memcg: mem+swap c... |
4580 4581 |
* Removal of cgroup itself succeeds regardless of refs from swap. */ |
c0ff4b854 memcg: rename mem... |
4582 |
static void __mem_cgroup_free(struct mem_cgroup *memcg) |
333279487 memcgroup: use vm... |
4583 |
{ |
08e552c69 memcg: synchroniz... |
4584 |
int node; |
c0ff4b854 memcg: rename mem... |
4585 4586 |
mem_cgroup_remove_from_trees(memcg); free_css_id(&mem_cgroup_subsys, &memcg->css); |
04046e1a0 memcg: use CSS ID |
4587 |
|
3ed28fa10 memcg: cleanup fo... |
4588 |
for_each_node(node) |
c0ff4b854 memcg: rename mem... |
4589 |
free_mem_cgroup_per_zone_info(memcg, node); |
08e552c69 memcg: synchroniz... |
4590 |
|
c0ff4b854 memcg: rename mem... |
4591 |
free_percpu(memcg->stat); |
c62b1a3b3 memcg: use generi... |
4592 |
if (sizeof(struct mem_cgroup) < PAGE_SIZE) |
c0ff4b854 memcg: rename mem... |
4593 |
kfree(memcg); |
333279487 memcgroup: use vm... |
4594 |
else |
c0ff4b854 memcg: rename mem... |
4595 |
vfree(memcg); |
333279487 memcgroup: use vm... |
4596 |
} |
c0ff4b854 memcg: rename mem... |
4597 |
static void mem_cgroup_get(struct mem_cgroup *memcg) |
8c7c6e34a memcg: mem+swap c... |
4598 |
{ |
c0ff4b854 memcg: rename mem... |
4599 |
atomic_inc(&memcg->refcnt); |
8c7c6e34a memcg: mem+swap c... |
4600 |
} |
c0ff4b854 memcg: rename mem... |
4601 |
static void __mem_cgroup_put(struct mem_cgroup *memcg, int count) |
8c7c6e34a memcg: mem+swap c... |
4602 |
{ |
c0ff4b854 memcg: rename mem... |
4603 4604 4605 |
if (atomic_sub_and_test(count, &memcg->refcnt)) { struct mem_cgroup *parent = parent_mem_cgroup(memcg); __mem_cgroup_free(memcg); |
7bcc1bb12 memcg: get/put pa... |
4606 4607 4608 |
if (parent) mem_cgroup_put(parent); } |
8c7c6e34a memcg: mem+swap c... |
4609 |
} |
c0ff4b854 memcg: rename mem... |
4610 |
static void mem_cgroup_put(struct mem_cgroup *memcg) |
483c30b51 memcg: improve pe... |
4611 |
{ |
c0ff4b854 memcg: rename mem... |
4612 |
__mem_cgroup_put(memcg, 1); |
483c30b51 memcg: improve pe... |
4613 |
} |
7bcc1bb12 memcg: get/put pa... |
4614 4615 4616 |
/* * Returns the parent mem_cgroup in memcgroup hierarchy with hierarchy enabled. */ |
e1aab161e socket: initial c... |
4617 |
struct mem_cgroup *parent_mem_cgroup(struct mem_cgroup *memcg) |
7bcc1bb12 memcg: get/put pa... |
4618 |
{ |
c0ff4b854 memcg: rename mem... |
4619 |
if (!memcg->res.parent) |
7bcc1bb12 memcg: get/put pa... |
4620 |
return NULL; |
c0ff4b854 memcg: rename mem... |
4621 |
return mem_cgroup_from_res_counter(memcg->res.parent, res); |
7bcc1bb12 memcg: get/put pa... |
4622 |
} |
e1aab161e socket: initial c... |
4623 |
EXPORT_SYMBOL(parent_mem_cgroup); |
333279487 memcgroup: use vm... |
4624 |
|
c077719be memcg: mem+swap c... |
4625 4626 4627 |
#ifdef CONFIG_CGROUP_MEM_RES_CTLR_SWAP static void __init enable_swap_cgroup(void) { |
f8d665422 memcg: add mem_cg... |
4628 |
if (!mem_cgroup_disabled() && really_do_swap_account) |
c077719be memcg: mem+swap c... |
4629 4630 4631 4632 4633 4634 4635 |
do_swap_account = 1; } #else static void __init enable_swap_cgroup(void) { } #endif |
f64c3f549 memory controller... |
4636 4637 4638 4639 4640 |
static int mem_cgroup_soft_limit_tree_init(void) { struct mem_cgroup_tree_per_node *rtpn; struct mem_cgroup_tree_per_zone *rtpz; int tmp, node, zone; |
3ed28fa10 memcg: cleanup fo... |
4641 |
for_each_node(node) { |
f64c3f549 memory controller... |
4642 4643 4644 4645 4646 |
tmp = node; if (!node_state(node, N_NORMAL_MEMORY)) tmp = -1; rtpn = kzalloc_node(sizeof(*rtpn), GFP_KERNEL, tmp); if (!rtpn) |
c3cecc683 memcg: free entri... |
4647 |
goto err_cleanup; |
f64c3f549 memory controller... |
4648 4649 4650 4651 4652 4653 4654 4655 4656 4657 |
soft_limit_tree.rb_tree_per_node[node] = rtpn; for (zone = 0; zone < MAX_NR_ZONES; zone++) { rtpz = &rtpn->rb_tree_per_zone[zone]; rtpz->rb_root = RB_ROOT; spin_lock_init(&rtpz->lock); } } return 0; |
c3cecc683 memcg: free entri... |
4658 4659 |
err_cleanup: |
3ed28fa10 memcg: cleanup fo... |
4660 |
for_each_node(node) { |
c3cecc683 memcg: free entri... |
4661 4662 4663 4664 4665 4666 |
if (!soft_limit_tree.rb_tree_per_node[node]) break; kfree(soft_limit_tree.rb_tree_per_node[node]); soft_limit_tree.rb_tree_per_node[node] = NULL; } return 1; |
f64c3f549 memory controller... |
4667 |
} |
0eb253e22 memcg: fix sectio... |
4668 |
static struct cgroup_subsys_state * __ref |
8cdea7c05 Memory controller... |
4669 4670 |
mem_cgroup_create(struct cgroup_subsys *ss, struct cgroup *cont) { |
c0ff4b854 memcg: rename mem... |
4671 |
struct mem_cgroup *memcg, *parent; |
04046e1a0 memcg: use CSS ID |
4672 |
long error = -ENOMEM; |
6d12e2d8d per-zone and recl... |
4673 |
int node; |
8cdea7c05 Memory controller... |
4674 |
|
c0ff4b854 memcg: rename mem... |
4675 4676 |
memcg = mem_cgroup_alloc(); if (!memcg) |
04046e1a0 memcg: use CSS ID |
4677 |
return ERR_PTR(error); |
78fb74669 Memory controller... |
4678 |
|
3ed28fa10 memcg: cleanup fo... |
4679 |
for_each_node(node) |
c0ff4b854 memcg: rename mem... |
4680 |
if (alloc_mem_cgroup_per_zone_info(memcg, node)) |
6d12e2d8d per-zone and recl... |
4681 |
goto free_out; |
f64c3f549 memory controller... |
4682 |
|
c077719be memcg: mem+swap c... |
4683 |
/* root ? */ |
28dbc4b6a memcg: memory cgr... |
4684 |
if (cont->parent == NULL) { |
cdec2e426 memcg: coalesce c... |
4685 |
int cpu; |
c077719be memcg: mem+swap c... |
4686 |
enable_swap_cgroup(); |
28dbc4b6a memcg: memory cgr... |
4687 |
parent = NULL; |
f64c3f549 memory controller... |
4688 4689 |
if (mem_cgroup_soft_limit_tree_init()) goto free_out; |
a41c58a66 memcg: keep root ... |
4690 |
root_mem_cgroup = memcg; |
cdec2e426 memcg: coalesce c... |
4691 4692 4693 4694 4695 |
for_each_possible_cpu(cpu) { struct memcg_stock_pcp *stock = &per_cpu(memcg_stock, cpu); INIT_WORK(&stock->work, drain_local_stock); } |
711d3d2c9 memcg: cpu hotplu... |
4696 |
hotcpu_notifier(memcg_cpu_hotplug_callback, 0); |
18f59ea7d memcg: memory cgr... |
4697 |
} else { |
28dbc4b6a memcg: memory cgr... |
4698 |
parent = mem_cgroup_from_cont(cont->parent); |
c0ff4b854 memcg: rename mem... |
4699 4700 |
memcg->use_hierarchy = parent->use_hierarchy; memcg->oom_kill_disable = parent->oom_kill_disable; |
18f59ea7d memcg: memory cgr... |
4701 |
} |
28dbc4b6a memcg: memory cgr... |
4702 |
|
18f59ea7d memcg: memory cgr... |
4703 |
if (parent && parent->use_hierarchy) { |
c0ff4b854 memcg: rename mem... |
4704 4705 |
res_counter_init(&memcg->res, &parent->res); res_counter_init(&memcg->memsw, &parent->memsw); |
7bcc1bb12 memcg: get/put pa... |
4706 4707 4708 4709 4710 4711 4712 |
/* * We increment refcnt of the parent to ensure that we can * safely access it on res_counter_charge/uncharge. * This refcnt will be decremented when freeing this * mem_cgroup(see mem_cgroup_put). */ mem_cgroup_get(parent); |
18f59ea7d memcg: memory cgr... |
4713 |
} else { |
c0ff4b854 memcg: rename mem... |
4714 4715 |
res_counter_init(&memcg->res, NULL); res_counter_init(&memcg->memsw, NULL); |
18f59ea7d memcg: memory cgr... |
4716 |
} |
c0ff4b854 memcg: rename mem... |
4717 4718 |
memcg->last_scanned_node = MAX_NUMNODES; INIT_LIST_HEAD(&memcg->oom_notify); |
6d61ef409 memcg: memory cgr... |
4719 |
|
a7885eb8a memcg: swappiness |
4720 |
if (parent) |
c0ff4b854 memcg: rename mem... |
4721 4722 4723 4724 4725 |
memcg->swappiness = mem_cgroup_swappiness(parent); atomic_set(&memcg->refcnt, 1); memcg->move_charge_at_immigrate = 0; mutex_init(&memcg->thresholds_lock); return &memcg->css; |
6d12e2d8d per-zone and recl... |
4726 |
free_out: |
c0ff4b854 memcg: rename mem... |
4727 |
__mem_cgroup_free(memcg); |
04046e1a0 memcg: use CSS ID |
4728 |
return ERR_PTR(error); |
8cdea7c05 Memory controller... |
4729 |
} |
ec64f5154 cgroup: fix frequ... |
4730 |
static int mem_cgroup_pre_destroy(struct cgroup_subsys *ss, |
df878fb04 memory cgroup enh... |
4731 4732 |
struct cgroup *cont) { |
c0ff4b854 memcg: rename mem... |
4733 |
struct mem_cgroup *memcg = mem_cgroup_from_cont(cont); |
ec64f5154 cgroup: fix frequ... |
4734 |
|
c0ff4b854 memcg: rename mem... |
4735 |
return mem_cgroup_force_empty(memcg, false); |
df878fb04 memory cgroup enh... |
4736 |
} |
8cdea7c05 Memory controller... |
4737 4738 4739 |
static void mem_cgroup_destroy(struct cgroup_subsys *ss, struct cgroup *cont) { |
c0ff4b854 memcg: rename mem... |
4740 |
struct mem_cgroup *memcg = mem_cgroup_from_cont(cont); |
c268e9946 memcg: fix hierar... |
4741 |
|
d1a4c0b37 tcp memory pressu... |
4742 |
kmem_cgroup_destroy(ss, cont); |
c0ff4b854 memcg: rename mem... |
4743 |
mem_cgroup_put(memcg); |
8cdea7c05 Memory controller... |
4744 4745 4746 4747 4748 |
} static int mem_cgroup_populate(struct cgroup_subsys *ss, struct cgroup *cont) { |
8c7c6e34a memcg: mem+swap c... |
4749 4750 4751 4752 4753 4754 4755 |
int ret; ret = cgroup_add_files(cont, ss, mem_cgroup_files, ARRAY_SIZE(mem_cgroup_files)); if (!ret) ret = register_memsw_files(cont, ss); |
e5671dfae Basic kernel memo... |
4756 4757 4758 |
if (!ret) ret = register_kmem_files(cont, ss); |
8c7c6e34a memcg: mem+swap c... |
4759 |
return ret; |
8cdea7c05 Memory controller... |
4760 |
} |
024914477 memcg: move charg... |
4761 |
#ifdef CONFIG_MMU |
7dc74be03 memcg: add interf... |
4762 |
/* Handlers for move charge at task migration. */ |
854ffa8d1 memcg: improve pe... |
4763 4764 |
#define PRECHARGE_COUNT_AT_ONCE 256 static int mem_cgroup_do_precharge(unsigned long count) |
7dc74be03 memcg: add interf... |
4765 |
{ |
854ffa8d1 memcg: improve pe... |
4766 4767 |
int ret = 0; int batch_count = PRECHARGE_COUNT_AT_ONCE; |
c0ff4b854 memcg: rename mem... |
4768 |
struct mem_cgroup *memcg = mc.to; |
4ffef5fef memcg: move charg... |
4769 |
|
c0ff4b854 memcg: rename mem... |
4770 |
if (mem_cgroup_is_root(memcg)) { |
854ffa8d1 memcg: improve pe... |
4771 4772 4773 4774 4775 4776 4777 4778 |
mc.precharge += count; /* we don't need css_get for root */ return ret; } /* try to charge at once */ if (count > 1) { struct res_counter *dummy; /* |
c0ff4b854 memcg: rename mem... |
4779 |
* "memcg" cannot be under rmdir() because we've already checked |
854ffa8d1 memcg: improve pe... |
4780 4781 4782 4783 |
* by cgroup_lock_live_cgroup() that it is not removed and we * are still under the same cgroup_mutex. So we can postpone * css_get(). */ |
c0ff4b854 memcg: rename mem... |
4784 |
if (res_counter_charge(&memcg->res, PAGE_SIZE * count, &dummy)) |
854ffa8d1 memcg: improve pe... |
4785 |
goto one_by_one; |
c0ff4b854 memcg: rename mem... |
4786 |
if (do_swap_account && res_counter_charge(&memcg->memsw, |
854ffa8d1 memcg: improve pe... |
4787 |
PAGE_SIZE * count, &dummy)) { |
c0ff4b854 memcg: rename mem... |
4788 |
res_counter_uncharge(&memcg->res, PAGE_SIZE * count); |
854ffa8d1 memcg: improve pe... |
4789 4790 4791 |
goto one_by_one; } mc.precharge += count; |
854ffa8d1 memcg: improve pe... |
4792 4793 4794 4795 4796 4797 4798 4799 4800 4801 4802 4803 4804 |
return ret; } one_by_one: /* fall back to one by one charge */ while (count--) { if (signal_pending(current)) { ret = -EINTR; break; } if (!batch_count--) { batch_count = PRECHARGE_COUNT_AT_ONCE; cond_resched(); } |
c0ff4b854 memcg: rename mem... |
4805 4806 |
ret = __mem_cgroup_try_charge(NULL, GFP_KERNEL, 1, &memcg, false); |
38c5d72f3 memcg: simplify L... |
4807 |
if (ret) |
854ffa8d1 memcg: improve pe... |
4808 |
/* mem_cgroup_clear_mc() will do uncharge later */ |
38c5d72f3 memcg: simplify L... |
4809 |
return ret; |
854ffa8d1 memcg: improve pe... |
4810 4811 |
mc.precharge++; } |
4ffef5fef memcg: move charg... |
4812 4813 4814 4815 4816 4817 4818 4819 |
return ret; } /** * is_target_pte_for_mc - check a pte whether it is valid for move charge * @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... |
4820 |
* @target: the pointer the target page or swap ent will be stored(can be NULL) |
4ffef5fef memcg: move charg... |
4821 4822 4823 4824 4825 4826 |
* * 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... |
4827 4828 4829 |
* 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... |
4830 4831 4832 |
* * Called with pte lock held. */ |
4ffef5fef memcg: move charg... |
4833 4834 |
union mc_target { struct page *page; |
024914477 memcg: move charg... |
4835 |
swp_entry_t ent; |
4ffef5fef memcg: move charg... |
4836 |
}; |
4ffef5fef memcg: move charg... |
4837 4838 4839 |
enum mc_target_type { MC_TARGET_NONE, /* not used */ MC_TARGET_PAGE, |
024914477 memcg: move charg... |
4840 |
MC_TARGET_SWAP, |
4ffef5fef memcg: move charg... |
4841 |
}; |
90254a658 memcg: clean up m... |
4842 4843 |
static struct page *mc_handle_present_pte(struct vm_area_struct *vma, unsigned long addr, pte_t ptent) |
4ffef5fef memcg: move charg... |
4844 |
{ |
90254a658 memcg: clean up m... |
4845 |
struct page *page = vm_normal_page(vma, addr, ptent); |
4ffef5fef memcg: move charg... |
4846 |
|
90254a658 memcg: clean up m... |
4847 4848 4849 4850 4851 4852 |
if (!page || !page_mapped(page)) return NULL; if (PageAnon(page)) { /* we don't move shared anon */ if (!move_anon() || page_mapcount(page) > 2) return NULL; |
87946a722 memcg: move charg... |
4853 4854 |
} else if (!move_file()) /* we ignore mapcount for file pages */ |
90254a658 memcg: clean up m... |
4855 4856 4857 4858 4859 4860 4861 4862 4863 4864 4865 4866 4867 4868 4869 4870 4871 4872 |
return NULL; if (!get_page_unless_zero(page)) return NULL; return page; } static struct page *mc_handle_swap_pte(struct vm_area_struct *vma, unsigned long addr, pte_t ptent, swp_entry_t *entry) { int usage_count; struct page *page = NULL; swp_entry_t ent = pte_to_swp_entry(ptent); if (!move_anon() || non_swap_entry(ent)) return NULL; usage_count = mem_cgroup_count_swap_user(ent, &page); if (usage_count > 1) { /* we don't move shared anon */ |
024914477 memcg: move charg... |
4873 4874 |
if (page) put_page(page); |
90254a658 memcg: clean up m... |
4875 |
return NULL; |
024914477 memcg: move charg... |
4876 |
} |
90254a658 memcg: clean up m... |
4877 4878 4879 4880 4881 |
if (do_swap_account) entry->val = ent.val; return page; } |
87946a722 memcg: move charg... |
4882 4883 4884 4885 4886 4887 4888 4889 4890 4891 4892 4893 4894 4895 4896 4897 4898 4899 4900 4901 4902 |
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; struct inode *inode; struct address_space *mapping; pgoff_t pgoff; if (!vma->vm_file) /* anonymous vma */ return NULL; if (!move_file()) return NULL; inode = vma->vm_file->f_path.dentry->d_inode; mapping = vma->vm_file->f_mapping; if (pte_none(ptent)) pgoff = linear_page_index(vma, addr); else /* pte_file(ptent) is true */ pgoff = pte_to_pgoff(ptent); /* page is moved even if it's not RSS of this task(page-faulted). */ |
aa3b18955 tmpfs: convert me... |
4903 4904 4905 4906 4907 4908 |
page = find_get_page(mapping, pgoff); #ifdef CONFIG_SWAP /* shmem/tmpfs may report page out on swap: account for that too. */ if (radix_tree_exceptional_entry(page)) { swp_entry_t swap = radix_to_swp_entry(page); |
87946a722 memcg: move charg... |
4909 |
if (do_swap_account) |
aa3b18955 tmpfs: convert me... |
4910 4911 |
*entry = swap; page = find_get_page(&swapper_space, swap.val); |
87946a722 memcg: move charg... |
4912 |
} |
aa3b18955 tmpfs: convert me... |
4913 |
#endif |
87946a722 memcg: move charg... |
4914 4915 |
return page; } |
90254a658 memcg: clean up m... |
4916 4917 4918 4919 4920 4921 4922 4923 4924 4925 4926 4927 |
static int is_target_pte_for_mc(struct vm_area_struct *vma, unsigned long addr, pte_t ptent, union mc_target *target) { struct page *page = NULL; struct page_cgroup *pc; int ret = 0; 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); |
87946a722 memcg: move charg... |
4928 4929 |
else if (pte_none(ptent) || pte_file(ptent)) page = mc_handle_file_pte(vma, addr, ptent, &ent); |
90254a658 memcg: clean up m... |
4930 4931 4932 |
if (!page && !ent.val) return 0; |
024914477 memcg: move charg... |
4933 4934 4935 4936 4937 4938 4939 4940 4941 4942 4943 4944 4945 4946 4947 |
if (page) { pc = lookup_page_cgroup(page); /* * Do only loose check w/o page_cgroup lock. * mem_cgroup_move_account() checks the pc is valid or not under * the lock. */ if (PageCgroupUsed(pc) && pc->mem_cgroup == mc.from) { ret = MC_TARGET_PAGE; if (target) target->page = page; } if (!ret || !target) put_page(page); } |
90254a658 memcg: clean up m... |
4948 4949 |
/* There is a swap entry and a page doesn't exist or isn't charged */ if (ent.val && !ret && |
9fb4b7cc0 page_cgroup: add ... |
4950 |
css_id(&mc.from->css) == lookup_swap_cgroup_id(ent)) { |
7f0f15464 memcg: fix css_id... |
4951 4952 4953 |
ret = MC_TARGET_SWAP; if (target) target->ent = ent; |
4ffef5fef memcg: move charg... |
4954 |
} |
4ffef5fef memcg: move charg... |
4955 4956 4957 4958 4959 4960 4961 4962 4963 4964 |
return ret; } static int mem_cgroup_count_precharge_pte_range(pmd_t *pmd, unsigned long addr, unsigned long end, struct mm_walk *walk) { struct vm_area_struct *vma = walk->private; pte_t *pte; spinlock_t *ptl; |
033193275 pagewalk: only sp... |
4965 |
split_huge_page_pmd(walk->mm, pmd); |
4ffef5fef memcg: move charg... |
4966 4967 4968 4969 4970 4971 |
pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl); for (; addr != end; pte++, addr += PAGE_SIZE) if (is_target_pte_for_mc(vma, addr, *pte, NULL)) mc.precharge++; /* increment precharge temporarily */ pte_unmap_unlock(pte - 1, ptl); cond_resched(); |
7dc74be03 memcg: add interf... |
4972 4973 |
return 0; } |
4ffef5fef memcg: move charg... |
4974 4975 4976 4977 |
static unsigned long mem_cgroup_count_precharge(struct mm_struct *mm) { unsigned long precharge; struct vm_area_struct *vma; |
dfe076b09 memcg: fix deadlo... |
4978 |
down_read(&mm->mmap_sem); |
4ffef5fef memcg: move charg... |
4979 4980 4981 4982 4983 4984 4985 4986 |
for (vma = mm->mmap; vma; vma = vma->vm_next) { struct mm_walk mem_cgroup_count_precharge_walk = { .pmd_entry = mem_cgroup_count_precharge_pte_range, .mm = mm, .private = vma, }; if (is_vm_hugetlb_page(vma)) continue; |
4ffef5fef memcg: move charg... |
4987 4988 4989 |
walk_page_range(vma->vm_start, vma->vm_end, &mem_cgroup_count_precharge_walk); } |
dfe076b09 memcg: fix deadlo... |
4990 |
up_read(&mm->mmap_sem); |
4ffef5fef memcg: move charg... |
4991 4992 4993 4994 4995 4996 |
precharge = mc.precharge; mc.precharge = 0; return precharge; } |
4ffef5fef memcg: move charg... |
4997 4998 |
static int mem_cgroup_precharge_mc(struct mm_struct *mm) { |
dfe076b09 memcg: fix deadlo... |
4999 5000 5001 5002 5003 |
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... |
5004 |
} |
dfe076b09 memcg: fix deadlo... |
5005 5006 |
/* 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... |
5007 |
{ |
2bd9bb206 memcg: clean up w... |
5008 5009 |
struct mem_cgroup *from = mc.from; struct mem_cgroup *to = mc.to; |
4ffef5fef memcg: move charg... |
5010 |
/* we must uncharge all the leftover precharges from mc.to */ |
854ffa8d1 memcg: improve pe... |
5011 5012 5013 5014 5015 5016 5017 5018 5019 5020 5021 |
if (mc.precharge) { __mem_cgroup_cancel_charge(mc.to, mc.precharge); mc.precharge = 0; } /* * we didn't uncharge from mc.from at mem_cgroup_move_account(), so * we must uncharge here. */ if (mc.moved_charge) { __mem_cgroup_cancel_charge(mc.from, mc.moved_charge); mc.moved_charge = 0; |
4ffef5fef memcg: move charg... |
5022 |
} |
483c30b51 memcg: improve pe... |
5023 5024 |
/* we must fixup refcnts and charges */ if (mc.moved_swap) { |
483c30b51 memcg: improve pe... |
5025 5026 5027 5028 5029 5030 5031 5032 5033 5034 5035 5036 5037 |
/* uncharge swap account from the old cgroup */ if (!mem_cgroup_is_root(mc.from)) res_counter_uncharge(&mc.from->memsw, PAGE_SIZE * mc.moved_swap); __mem_cgroup_put(mc.from, mc.moved_swap); if (!mem_cgroup_is_root(mc.to)) { /* * we charged both to->res and to->memsw, so we should * uncharge to->res. */ res_counter_uncharge(&mc.to->res, PAGE_SIZE * mc.moved_swap); |
483c30b51 memcg: improve pe... |
5038 5039 |
} /* we've already done mem_cgroup_get(mc.to) */ |
483c30b51 memcg: improve pe... |
5040 5041 |
mc.moved_swap = 0; } |
dfe076b09 memcg: fix deadlo... |
5042 5043 5044 5045 5046 5047 5048 5049 5050 5051 5052 5053 5054 5055 5056 |
memcg_oom_recover(from); memcg_oom_recover(to); wake_up_all(&mc.waitq); } static void mem_cgroup_clear_mc(void) { struct mem_cgroup *from = mc.from; /* * 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... |
5057 |
spin_lock(&mc.lock); |
4ffef5fef memcg: move charg... |
5058 5059 |
mc.from = NULL; mc.to = NULL; |
2bd9bb206 memcg: clean up w... |
5060 |
spin_unlock(&mc.lock); |
32047e2a8 memcg: avoid lock... |
5061 |
mem_cgroup_end_move(from); |
4ffef5fef memcg: move charg... |
5062 |
} |
7dc74be03 memcg: add interf... |
5063 5064 |
static int mem_cgroup_can_attach(struct cgroup_subsys *ss, struct cgroup *cgroup, |
2f7ee5691 cgroup: introduce... |
5065 |
struct cgroup_taskset *tset) |
7dc74be03 memcg: add interf... |
5066 |
{ |
2f7ee5691 cgroup: introduce... |
5067 |
struct task_struct *p = cgroup_taskset_first(tset); |
7dc74be03 memcg: add interf... |
5068 |
int ret = 0; |
c0ff4b854 memcg: rename mem... |
5069 |
struct mem_cgroup *memcg = mem_cgroup_from_cont(cgroup); |
7dc74be03 memcg: add interf... |
5070 |
|
c0ff4b854 memcg: rename mem... |
5071 |
if (memcg->move_charge_at_immigrate) { |
7dc74be03 memcg: add interf... |
5072 5073 |
struct mm_struct *mm; struct mem_cgroup *from = mem_cgroup_from_task(p); |
c0ff4b854 memcg: rename mem... |
5074 |
VM_BUG_ON(from == memcg); |
7dc74be03 memcg: add interf... |
5075 5076 5077 5078 |
mm = get_task_mm(p); if (!mm) return 0; |
7dc74be03 memcg: add interf... |
5079 |
/* We move charges only when we move a owner of the mm */ |
4ffef5fef memcg: move charg... |
5080 5081 5082 5083 |
if (mm->owner == p) { VM_BUG_ON(mc.from); VM_BUG_ON(mc.to); VM_BUG_ON(mc.precharge); |
854ffa8d1 memcg: improve pe... |
5084 |
VM_BUG_ON(mc.moved_charge); |
483c30b51 memcg: improve pe... |
5085 |
VM_BUG_ON(mc.moved_swap); |
32047e2a8 memcg: avoid lock... |
5086 |
mem_cgroup_start_move(from); |
2bd9bb206 memcg: clean up w... |
5087 |
spin_lock(&mc.lock); |
4ffef5fef memcg: move charg... |
5088 |
mc.from = from; |
c0ff4b854 memcg: rename mem... |
5089 |
mc.to = memcg; |
2bd9bb206 memcg: clean up w... |
5090 |
spin_unlock(&mc.lock); |
dfe076b09 memcg: fix deadlo... |
5091 |
/* We set mc.moving_task later */ |
4ffef5fef memcg: move charg... |
5092 5093 5094 5095 |
ret = mem_cgroup_precharge_mc(mm); if (ret) mem_cgroup_clear_mc(); |
dfe076b09 memcg: fix deadlo... |
5096 5097 |
} mmput(mm); |
7dc74be03 memcg: add interf... |
5098 5099 5100 5101 5102 5103 |
} return ret; } static void mem_cgroup_cancel_attach(struct cgroup_subsys *ss, struct cgroup *cgroup, |
2f7ee5691 cgroup: introduce... |
5104 |
struct cgroup_taskset *tset) |
7dc74be03 memcg: add interf... |
5105 |
{ |
4ffef5fef memcg: move charg... |
5106 |
mem_cgroup_clear_mc(); |
7dc74be03 memcg: add interf... |
5107 |
} |
4ffef5fef memcg: move charg... |
5108 5109 5110 |
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... |
5111 |
{ |
4ffef5fef memcg: move charg... |
5112 5113 5114 5115 |
int ret = 0; struct vm_area_struct *vma = walk->private; pte_t *pte; spinlock_t *ptl; |
033193275 pagewalk: only sp... |
5116 |
split_huge_page_pmd(walk->mm, pmd); |
4ffef5fef memcg: move charg... |
5117 5118 5119 5120 5121 5122 5123 5124 |
retry: pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl); for (; addr != end; addr += PAGE_SIZE) { pte_t ptent = *(pte++); union mc_target target; int type; struct page *page; struct page_cgroup *pc; |
024914477 memcg: move charg... |
5125 |
swp_entry_t ent; |
4ffef5fef memcg: move charg... |
5126 5127 5128 5129 5130 5131 5132 5133 5134 5135 5136 |
if (!mc.precharge) break; type = is_target_pte_for_mc(vma, addr, ptent, &target); switch (type) { case MC_TARGET_PAGE: page = target.page; if (isolate_lru_page(page)) goto put; pc = lookup_page_cgroup(page); |
7ec99d621 memcg: unify char... |
5137 5138 |
if (!mem_cgroup_move_account(page, 1, pc, mc.from, mc.to, false)) { |
4ffef5fef memcg: move charg... |
5139 |
mc.precharge--; |
854ffa8d1 memcg: improve pe... |
5140 5141 |
/* we uncharge from mc.from later. */ mc.moved_charge++; |
4ffef5fef memcg: move charg... |
5142 5143 5144 5145 5146 |
} putback_lru_page(page); put: /* is_target_pte_for_mc() gets the page */ put_page(page); break; |
024914477 memcg: move charg... |
5147 5148 |
case MC_TARGET_SWAP: ent = target.ent; |
483c30b51 memcg: improve pe... |
5149 5150 |
if (!mem_cgroup_move_swap_account(ent, mc.from, mc.to, false)) { |
024914477 memcg: move charg... |
5151 |
mc.precharge--; |
483c30b51 memcg: improve pe... |
5152 5153 5154 |
/* we fixup refcnts and charges later. */ mc.moved_swap++; } |
024914477 memcg: move charg... |
5155 |
break; |
4ffef5fef memcg: move charg... |
5156 5157 5158 5159 5160 5161 5162 5163 5164 5165 5166 5167 5168 5169 |
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... |
5170 |
ret = mem_cgroup_do_precharge(1); |
4ffef5fef memcg: move charg... |
5171 5172 5173 5174 5175 5176 5177 5178 5179 5180 5181 5182 |
if (!ret) goto retry; } return ret; } static void mem_cgroup_move_charge(struct mm_struct *mm) { struct vm_area_struct *vma; lru_add_drain_all(); |
dfe076b09 memcg: fix deadlo... |
5183 5184 5185 5186 5187 5188 5189 5190 5191 5192 5193 5194 5195 |
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; } |
4ffef5fef memcg: move charg... |
5196 5197 5198 5199 5200 5201 5202 5203 5204 |
for (vma = mm->mmap; vma; vma = vma->vm_next) { int ret; struct mm_walk mem_cgroup_move_charge_walk = { .pmd_entry = mem_cgroup_move_charge_pte_range, .mm = mm, .private = vma, }; if (is_vm_hugetlb_page(vma)) continue; |
4ffef5fef memcg: move charg... |
5205 5206 5207 5208 5209 5210 5211 5212 5213 |
ret = walk_page_range(vma->vm_start, vma->vm_end, &mem_cgroup_move_charge_walk); if (ret) /* * means we have consumed all precharges and failed in * doing additional charge. Just abandon here. */ break; } |
dfe076b09 memcg: fix deadlo... |
5214 |
up_read(&mm->mmap_sem); |
7dc74be03 memcg: add interf... |
5215 |
} |
67e465a77 Memory controller... |
5216 5217 |
static void mem_cgroup_move_task(struct cgroup_subsys *ss, struct cgroup *cont, |
2f7ee5691 cgroup: introduce... |
5218 |
struct cgroup_taskset *tset) |
67e465a77 Memory controller... |
5219 |
{ |
2f7ee5691 cgroup: introduce... |
5220 |
struct task_struct *p = cgroup_taskset_first(tset); |
a433658c3 vmscan,memcg: mem... |
5221 |
struct mm_struct *mm = get_task_mm(p); |
dfe076b09 memcg: fix deadlo... |
5222 |
|
dfe076b09 memcg: fix deadlo... |
5223 |
if (mm) { |
a433658c3 vmscan,memcg: mem... |
5224 5225 5226 |
if (mc.to) mem_cgroup_move_charge(mm); put_swap_token(mm); |
dfe076b09 memcg: fix deadlo... |
5227 5228 |
mmput(mm); } |
a433658c3 vmscan,memcg: mem... |
5229 5230 |
if (mc.to) mem_cgroup_clear_mc(); |
67e465a77 Memory controller... |
5231 |
} |
5cfb80a73 memcg: disable mo... |
5232 5233 5234 |
#else /* !CONFIG_MMU */ static int mem_cgroup_can_attach(struct cgroup_subsys *ss, struct cgroup *cgroup, |
2f7ee5691 cgroup: introduce... |
5235 |
struct cgroup_taskset *tset) |
5cfb80a73 memcg: disable mo... |
5236 5237 5238 5239 5240 |
{ return 0; } static void mem_cgroup_cancel_attach(struct cgroup_subsys *ss, struct cgroup *cgroup, |
2f7ee5691 cgroup: introduce... |
5241 |
struct cgroup_taskset *tset) |
5cfb80a73 memcg: disable mo... |
5242 5243 5244 5245 |
{ } static void mem_cgroup_move_task(struct cgroup_subsys *ss, struct cgroup *cont, |
2f7ee5691 cgroup: introduce... |
5246 |
struct cgroup_taskset *tset) |
5cfb80a73 memcg: disable mo... |
5247 5248 5249 |
{ } #endif |
67e465a77 Memory controller... |
5250 |
|
8cdea7c05 Memory controller... |
5251 5252 5253 5254 |
struct cgroup_subsys mem_cgroup_subsys = { .name = "memory", .subsys_id = mem_cgroup_subsys_id, .create = mem_cgroup_create, |
df878fb04 memory cgroup enh... |
5255 |
.pre_destroy = mem_cgroup_pre_destroy, |
8cdea7c05 Memory controller... |
5256 5257 |
.destroy = mem_cgroup_destroy, .populate = mem_cgroup_populate, |
7dc74be03 memcg: add interf... |
5258 5259 |
.can_attach = mem_cgroup_can_attach, .cancel_attach = mem_cgroup_cancel_attach, |
67e465a77 Memory controller... |
5260 |
.attach = mem_cgroup_move_task, |
6d12e2d8d per-zone and recl... |
5261 |
.early_init = 0, |
04046e1a0 memcg: use CSS ID |
5262 |
.use_id = 1, |
8cdea7c05 Memory controller... |
5263 |
}; |
c077719be memcg: mem+swap c... |
5264 5265 |
#ifdef CONFIG_CGROUP_MEM_RES_CTLR_SWAP |
a42c390cf cgroups: make swa... |
5266 5267 5268 |
static int __init enable_swap_account(char *s) { /* consider enabled if no parameter or 1 is given */ |
a2c8990ae memsw: remove nos... |
5269 |
if (!strcmp(s, "1")) |
a42c390cf cgroups: make swa... |
5270 |
really_do_swap_account = 1; |
a2c8990ae memsw: remove nos... |
5271 |
else if (!strcmp(s, "0")) |
a42c390cf cgroups: make swa... |
5272 5273 5274 |
really_do_swap_account = 0; return 1; } |
a2c8990ae memsw: remove nos... |
5275 |
__setup("swapaccount=", enable_swap_account); |
c077719be memcg: mem+swap c... |
5276 |
|
c077719be memcg: mem+swap c... |
5277 |
#endif |