Blame view
mm/memcontrol.c
145 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 |
}; |
d52aa412d memory cgroup enh... |
119 |
/* |
6d12e2d8d per-zone and recl... |
120 121 |
* per-zone information in memory controller. */ |
6d12e2d8d per-zone and recl... |
122 |
struct mem_cgroup_per_zone { |
072c56c13 per-zone and recl... |
123 124 125 |
/* * spin_lock to protect the per cgroup LRU */ |
b69408e88 vmscan: Use an in... |
126 127 |
struct list_head lists[NR_LRU_LISTS]; unsigned long count[NR_LRU_LISTS]; |
3e2f41f1f memcg: add zone_r... |
128 129 |
struct zone_reclaim_stat reclaim_stat; |
f64c3f549 memory controller... |
130 131 132 133 |
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... |
134 135 |
struct mem_cgroup *mem; /* Back pointer, we cannot */ /* use container_of */ |
6d12e2d8d per-zone and recl... |
136 137 138 139 140 141 142 143 144 145 146 147 148 |
}; /* 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... |
149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 |
* 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 ... |
167 168 169 170 |
struct mem_cgroup_threshold { struct eventfd_ctx *eventfd; u64 threshold; }; |
9490ff275 memcg: oom notifier |
171 |
/* For threshold */ |
2e72b6347 memcg: implement ... |
172 173 |
struct mem_cgroup_threshold_ary { /* An array index points to threshold just below usage. */ |
5407a5625 mm: remove unnece... |
174 |
int current_threshold; |
2e72b6347 memcg: implement ... |
175 176 177 178 179 |
/* Size of entries[] */ unsigned int size; /* Array of thresholds */ struct mem_cgroup_threshold entries[0]; }; |
2c488db27 memcg: clean up m... |
180 181 182 183 184 185 186 187 188 189 190 |
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 |
191 192 193 194 195 |
/* for OOM */ struct mem_cgroup_eventfd_list { struct list_head list; struct eventfd_ctx *eventfd; }; |
2e72b6347 memcg: implement ... |
196 |
|
c0ff4b854 memcg: rename mem... |
197 198 |
static void mem_cgroup_threshold(struct mem_cgroup *memcg); static void mem_cgroup_oom_notify(struct mem_cgroup *memcg); |
2e72b6347 memcg: implement ... |
199 |
|
f64c3f549 memory controller... |
200 |
/* |
8cdea7c05 Memory controller... |
201 202 203 204 205 206 |
* 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... |
207 208 209 |
* 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... |
210 211 212 213 214 215 216 |
*/ struct mem_cgroup { struct cgroup_subsys_state css; /* * the counter to account for memory usage */ struct res_counter res; |
78fb74669 Memory controller... |
217 |
/* |
8c7c6e34a memcg: mem+swap c... |
218 219 220 221 |
* the counter to account for mem+swap usage. */ struct res_counter memsw; /* |
78fb74669 Memory controller... |
222 223 |
* Per cgroup active and inactive list, similar to the * per zone LRU lists. |
78fb74669 Memory controller... |
224 |
*/ |
6d12e2d8d per-zone and recl... |
225 |
struct mem_cgroup_lru_info info; |
6d61ef409 memcg: memory cgr... |
226 |
/* |
af901ca18 tree-wide: fix as... |
227 |
* While reclaiming in a hierarchy, we cache the last child we |
04046e1a0 memcg: use CSS ID |
228 |
* reclaimed from. |
6d61ef409 memcg: memory cgr... |
229 |
*/ |
04046e1a0 memcg: use CSS ID |
230 |
int last_scanned_child; |
889976dbc memcg: reclaim me... |
231 232 233 |
int last_scanned_node; #if MAX_NUMNODES > 1 nodemask_t scan_nodes; |
453a9bf34 memcg: fix numa s... |
234 235 |
atomic_t numainfo_events; atomic_t numainfo_updating; |
889976dbc memcg: reclaim me... |
236 |
#endif |
18f59ea7d memcg: memory cgr... |
237 238 239 240 |
/* * Should the accounting and control be hierarchical, per subtree? */ bool use_hierarchy; |
79dfdaccd memcg: make oom_l... |
241 242 243 |
bool oom_lock; atomic_t under_oom; |
8c7c6e34a memcg: mem+swap c... |
244 |
atomic_t refcnt; |
14797e236 memcg: add inacti... |
245 |
|
1f4c025b5 memcg: export mem... |
246 |
int swappiness; |
3c11ecf44 memcg: oom kill d... |
247 248 |
/* OOM-Killer disable */ int oom_kill_disable; |
a7885eb8a memcg: swappiness |
249 |
|
22a668d7c memcg: fix behavi... |
250 251 |
/* set when res.limit == memsw.limit */ bool memsw_is_minimum; |
2e72b6347 memcg: implement ... |
252 253 254 255 |
/* protect arrays of thresholds */ struct mutex thresholds_lock; /* thresholds for memory usage. RCU-protected */ |
2c488db27 memcg: clean up m... |
256 |
struct mem_cgroup_thresholds thresholds; |
907860ed3 cgroups: make cft... |
257 |
|
2e72b6347 memcg: implement ... |
258 |
/* thresholds for mem+swap usage. RCU-protected */ |
2c488db27 memcg: clean up m... |
259 |
struct mem_cgroup_thresholds memsw_thresholds; |
907860ed3 cgroups: make cft... |
260 |
|
9490ff275 memcg: oom notifier |
261 262 |
/* For oom notifier event fd */ struct list_head oom_notify; |
185efc0f9 memcg: Revert "me... |
263 |
|
d52aa412d memory cgroup enh... |
264 |
/* |
7dc74be03 memcg: add interf... |
265 266 267 268 |
* 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... |
269 |
/* |
c62b1a3b3 memcg: use generi... |
270 |
* percpu counter. |
d52aa412d memory cgroup enh... |
271 |
*/ |
c62b1a3b3 memcg: use generi... |
272 |
struct mem_cgroup_stat_cpu *stat; |
711d3d2c9 memcg: cpu hotplu... |
273 274 275 276 277 278 |
/* * 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... |
279 280 281 282 |
#ifdef CONFIG_INET struct tcp_memcontrol tcp_mem; #endif |
8cdea7c05 Memory controller... |
283 |
}; |
7dc74be03 memcg: add interf... |
284 285 286 287 288 289 |
/* 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... |
290 |
MOVE_CHARGE_TYPE_ANON, /* private anonymous page and swap of it */ |
87946a722 memcg: move charg... |
291 |
MOVE_CHARGE_TYPE_FILE, /* file page(including tmpfs) and swap of it */ |
7dc74be03 memcg: add interf... |
292 293 |
NR_MOVE_TYPE, }; |
4ffef5fef memcg: move charg... |
294 295 |
/* "mc" and its members are protected by cgroup_mutex */ static struct move_charge_struct { |
b1dd693e5 memcg: avoid dead... |
296 |
spinlock_t lock; /* for from, to */ |
4ffef5fef memcg: move charg... |
297 298 299 |
struct mem_cgroup *from; struct mem_cgroup *to; unsigned long precharge; |
854ffa8d1 memcg: improve pe... |
300 |
unsigned long moved_charge; |
483c30b51 memcg: improve pe... |
301 |
unsigned long moved_swap; |
8033b97c9 memcg: avoid oom ... |
302 303 304 |
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... |
305 |
.lock = __SPIN_LOCK_UNLOCKED(mc.lock), |
8033b97c9 memcg: avoid oom ... |
306 307 |
.waitq = __WAIT_QUEUE_HEAD_INITIALIZER(mc.waitq), }; |
4ffef5fef memcg: move charg... |
308 |
|
90254a658 memcg: clean up m... |
309 310 311 312 313 |
static bool move_anon(void) { return test_bit(MOVE_CHARGE_TYPE_ANON, &mc.to->move_charge_at_immigrate); } |
87946a722 memcg: move charg... |
314 315 316 317 318 |
static bool move_file(void) { return test_bit(MOVE_CHARGE_TYPE_FILE, &mc.to->move_charge_at_immigrate); } |
4e4169535 memory controller... |
319 320 321 322 323 324 |
/* * 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... |
325 326 327 |
enum charge_type { MEM_CGROUP_CHARGE_TYPE_CACHE = 0, MEM_CGROUP_CHARGE_TYPE_MAPPED, |
4f98a2fee vmscan: split LRU... |
328 |
MEM_CGROUP_CHARGE_TYPE_SHMEM, /* used by page migration of shmem */ |
c05555b57 memcg: atomic ops... |
329 |
MEM_CGROUP_CHARGE_TYPE_FORCE, /* used by force_empty */ |
d13d14430 memcg: handle swa... |
330 |
MEM_CGROUP_CHARGE_TYPE_SWAPOUT, /* for accounting swapcache */ |
8a9478ca7 memcg: fix swap a... |
331 |
MEM_CGROUP_CHARGE_TYPE_DROP, /* a page was unused swap cache */ |
c05555b57 memcg: atomic ops... |
332 333 |
NR_CHARGE_TYPE, }; |
8c7c6e34a memcg: mem+swap c... |
334 |
/* for encoding cft->private value on file */ |
65c64ce8e Partial revert "B... |
335 336 337 |
#define _MEM (0) #define _MEMSWAP (1) #define _OOM_TYPE (2) |
8c7c6e34a memcg: mem+swap c... |
338 339 340 |
#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 |
341 342 |
/* Used for OOM nofiier */ #define OOM_CONTROL (0) |
8c7c6e34a memcg: mem+swap c... |
343 |
|
75822b449 memory controller... |
344 345 346 347 348 349 350 |
/* * 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) |
4e4169535 memory controller... |
351 352 |
#define MEM_CGROUP_RECLAIM_SOFT_BIT 0x2 #define MEM_CGROUP_RECLAIM_SOFT (1 << MEM_CGROUP_RECLAIM_SOFT_BIT) |
75822b449 memory controller... |
353 |
|
c0ff4b854 memcg: rename mem... |
354 355 |
static void mem_cgroup_get(struct mem_cgroup *memcg); static void mem_cgroup_put(struct mem_cgroup *memcg); |
e1aab161e socket: initial c... |
356 357 358 359 360 |
/* 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... |
361 |
#include <net/ip.h> |
e1aab161e socket: initial c... |
362 363 364 365 |
static bool mem_cgroup_is_root(struct mem_cgroup *memcg); void sock_update_memcg(struct sock *sk) { |
e1aab161e socket: initial c... |
366 367 368 369 |
if (static_branch(&memcg_socket_limit_enabled)) { struct mem_cgroup *memcg; BUG_ON(!sk->sk_prot->proto_cgroup); |
f3f511e1c net: fix sock_clo... |
370 371 372 373 374 375 376 377 378 379 380 381 382 |
/* 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... |
383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 |
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... |
403 404 405 406 407 408 409 410 411 |
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... |
412 413 |
#endif /* CONFIG_INET */ #endif /* CONFIG_CGROUP_MEM_RES_CTLR_KMEM */ |
c0ff4b854 memcg: rename mem... |
414 |
static void drain_all_stock_async(struct mem_cgroup *memcg); |
8c7c6e34a memcg: mem+swap c... |
415 |
|
f64c3f549 memory controller... |
416 |
static struct mem_cgroup_per_zone * |
c0ff4b854 memcg: rename mem... |
417 |
mem_cgroup_zoneinfo(struct mem_cgroup *memcg, int nid, int zid) |
f64c3f549 memory controller... |
418 |
{ |
c0ff4b854 memcg: rename mem... |
419 |
return &memcg->info.nodeinfo[nid]->zoneinfo[zid]; |
f64c3f549 memory controller... |
420 |
} |
c0ff4b854 memcg: rename mem... |
421 |
struct cgroup_subsys_state *mem_cgroup_css(struct mem_cgroup *memcg) |
d324236b3 memcg: add access... |
422 |
{ |
c0ff4b854 memcg: rename mem... |
423 |
return &memcg->css; |
d324236b3 memcg: add access... |
424 |
} |
f64c3f549 memory controller... |
425 |
static struct mem_cgroup_per_zone * |
c0ff4b854 memcg: rename mem... |
426 |
page_cgroup_zoneinfo(struct mem_cgroup *memcg, struct page *page) |
f64c3f549 memory controller... |
427 |
{ |
97a6c37b3 memcg: change pag... |
428 429 |
int nid = page_to_nid(page); int zid = page_zonenum(page); |
f64c3f549 memory controller... |
430 |
|
c0ff4b854 memcg: rename mem... |
431 |
return mem_cgroup_zoneinfo(memcg, nid, zid); |
f64c3f549 memory controller... |
432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 |
} 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... |
450 |
__mem_cgroup_insert_exceeded(struct mem_cgroup *memcg, |
f64c3f549 memory controller... |
451 |
struct mem_cgroup_per_zone *mz, |
ef8745c1e memcg: reduce che... |
452 453 |
struct mem_cgroup_tree_per_zone *mctz, unsigned long long new_usage_in_excess) |
f64c3f549 memory controller... |
454 455 456 457 458 459 460 |
{ 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... |
461 462 463 |
mz->usage_in_excess = new_usage_in_excess; if (!mz->usage_in_excess) return; |
f64c3f549 memory controller... |
464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 |
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... |
480 481 482 |
} static void |
c0ff4b854 memcg: rename mem... |
483 |
__mem_cgroup_remove_exceeded(struct mem_cgroup *memcg, |
4e4169535 memory controller... |
484 485 486 487 488 489 490 491 492 493 |
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... |
494 |
mem_cgroup_remove_exceeded(struct mem_cgroup *memcg, |
f64c3f549 memory controller... |
495 496 497 498 |
struct mem_cgroup_per_zone *mz, struct mem_cgroup_tree_per_zone *mctz) { spin_lock(&mctz->lock); |
c0ff4b854 memcg: rename mem... |
499 |
__mem_cgroup_remove_exceeded(memcg, mz, mctz); |
f64c3f549 memory controller... |
500 501 |
spin_unlock(&mctz->lock); } |
f64c3f549 memory controller... |
502 |
|
c0ff4b854 memcg: rename mem... |
503 |
static void mem_cgroup_update_tree(struct mem_cgroup *memcg, struct page *page) |
f64c3f549 memory controller... |
504 |
{ |
ef8745c1e memcg: reduce che... |
505 |
unsigned long long excess; |
f64c3f549 memory controller... |
506 507 |
struct mem_cgroup_per_zone *mz; struct mem_cgroup_tree_per_zone *mctz; |
4e649152c memcg: some modif... |
508 509 |
int nid = page_to_nid(page); int zid = page_zonenum(page); |
f64c3f549 memory controller... |
510 511 512 |
mctz = soft_limit_tree_from_page(page); /* |
4e649152c memcg: some modif... |
513 514 |
* Necessary to update all ancestors when hierarchy is used. * because their event counter is not touched. |
f64c3f549 memory controller... |
515 |
*/ |
c0ff4b854 memcg: rename mem... |
516 517 518 |
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... |
519 520 521 522 |
/* * We have to update the tree if mz is on RB-tree or * mem is over its softlimit. */ |
ef8745c1e memcg: reduce che... |
523 |
if (excess || mz->on_tree) { |
4e649152c memcg: some modif... |
524 525 526 |
spin_lock(&mctz->lock); /* if on-tree, remove it */ if (mz->on_tree) |
c0ff4b854 memcg: rename mem... |
527 |
__mem_cgroup_remove_exceeded(memcg, mz, mctz); |
4e649152c memcg: some modif... |
528 |
/* |
ef8745c1e memcg: reduce che... |
529 530 |
* Insert again. mz->usage_in_excess will be updated. * If excess is 0, no tree ops. |
4e649152c memcg: some modif... |
531 |
*/ |
c0ff4b854 memcg: rename mem... |
532 |
__mem_cgroup_insert_exceeded(memcg, mz, mctz, excess); |
4e649152c memcg: some modif... |
533 534 |
spin_unlock(&mctz->lock); } |
f64c3f549 memory controller... |
535 536 |
} } |
c0ff4b854 memcg: rename mem... |
537 |
static void mem_cgroup_remove_from_trees(struct mem_cgroup *memcg) |
f64c3f549 memory controller... |
538 539 540 541 542 543 544 |
{ int node, zone; struct mem_cgroup_per_zone *mz; struct mem_cgroup_tree_per_zone *mctz; for_each_node_state(node, N_POSSIBLE) { for (zone = 0; zone < MAX_NR_ZONES; zone++) { |
c0ff4b854 memcg: rename mem... |
545 |
mz = mem_cgroup_zoneinfo(memcg, node, zone); |
f64c3f549 memory controller... |
546 |
mctz = soft_limit_tree_node_zone(node, zone); |
c0ff4b854 memcg: rename mem... |
547 |
mem_cgroup_remove_exceeded(memcg, mz, mctz); |
f64c3f549 memory controller... |
548 549 550 |
} } } |
4e4169535 memory controller... |
551 552 553 554 |
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... |
555 |
struct mem_cgroup_per_zone *mz; |
4e4169535 memory controller... |
556 557 |
retry: |
26251eaf9 memcg: fix refcnt... |
558 |
mz = NULL; |
4e4169535 memory controller... |
559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 |
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... |
587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 |
/* * 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... |
606 |
static long mem_cgroup_read_stat(struct mem_cgroup *memcg, |
7a159cc9d memcg: use native... |
607 |
enum mem_cgroup_stat_index idx) |
c62b1a3b3 memcg: use generi... |
608 |
{ |
7a159cc9d memcg: use native... |
609 |
long val = 0; |
c62b1a3b3 memcg: use generi... |
610 |
int cpu; |
c62b1a3b3 memcg: use generi... |
611 |
|
711d3d2c9 memcg: cpu hotplu... |
612 613 |
get_online_cpus(); for_each_online_cpu(cpu) |
c0ff4b854 memcg: rename mem... |
614 |
val += per_cpu(memcg->stat->count[idx], cpu); |
711d3d2c9 memcg: cpu hotplu... |
615 |
#ifdef CONFIG_HOTPLUG_CPU |
c0ff4b854 memcg: rename mem... |
616 617 618 |
spin_lock(&memcg->pcp_counter_lock); val += memcg->nocpu_base.count[idx]; spin_unlock(&memcg->pcp_counter_lock); |
711d3d2c9 memcg: cpu hotplu... |
619 620 |
#endif put_online_cpus(); |
c62b1a3b3 memcg: use generi... |
621 622 |
return val; } |
c0ff4b854 memcg: rename mem... |
623 |
static void mem_cgroup_swap_statistics(struct mem_cgroup *memcg, |
0c3e73e84 memcg: improve re... |
624 625 626 |
bool charge) { int val = (charge) ? 1 : -1; |
c0ff4b854 memcg: rename mem... |
627 |
this_cpu_add(memcg->stat->count[MEM_CGROUP_STAT_SWAPOUT], val); |
0c3e73e84 memcg: improve re... |
628 |
} |
c0ff4b854 memcg: rename mem... |
629 |
void mem_cgroup_pgfault(struct mem_cgroup *memcg, int val) |
456f998ec memcg: add the pa... |
630 |
{ |
c0ff4b854 memcg: rename mem... |
631 |
this_cpu_add(memcg->stat->events[MEM_CGROUP_EVENTS_PGFAULT], val); |
456f998ec memcg: add the pa... |
632 |
} |
c0ff4b854 memcg: rename mem... |
633 |
void mem_cgroup_pgmajfault(struct mem_cgroup *memcg, int val) |
456f998ec memcg: add the pa... |
634 |
{ |
c0ff4b854 memcg: rename mem... |
635 |
this_cpu_add(memcg->stat->events[MEM_CGROUP_EVENTS_PGMAJFAULT], val); |
456f998ec memcg: add the pa... |
636 |
} |
c0ff4b854 memcg: rename mem... |
637 |
static unsigned long mem_cgroup_read_events(struct mem_cgroup *memcg, |
e9f8974f2 memcg: break out ... |
638 639 640 641 642 643 |
enum mem_cgroup_events_index idx) { unsigned long val = 0; int cpu; for_each_online_cpu(cpu) |
c0ff4b854 memcg: rename mem... |
644 |
val += per_cpu(memcg->stat->events[idx], cpu); |
e9f8974f2 memcg: break out ... |
645 |
#ifdef CONFIG_HOTPLUG_CPU |
c0ff4b854 memcg: rename mem... |
646 647 648 |
spin_lock(&memcg->pcp_counter_lock); val += memcg->nocpu_base.events[idx]; spin_unlock(&memcg->pcp_counter_lock); |
e9f8974f2 memcg: break out ... |
649 650 651 |
#endif return val; } |
c0ff4b854 memcg: rename mem... |
652 |
static void mem_cgroup_charge_statistics(struct mem_cgroup *memcg, |
e401f1761 memcg: modify acc... |
653 |
bool file, int nr_pages) |
d52aa412d memory cgroup enh... |
654 |
{ |
c62b1a3b3 memcg: use generi... |
655 |
preempt_disable(); |
e401f1761 memcg: modify acc... |
656 |
if (file) |
c0ff4b854 memcg: rename mem... |
657 658 |
__this_cpu_add(memcg->stat->count[MEM_CGROUP_STAT_CACHE], nr_pages); |
d52aa412d memory cgroup enh... |
659 |
else |
c0ff4b854 memcg: rename mem... |
660 661 |
__this_cpu_add(memcg->stat->count[MEM_CGROUP_STAT_RSS], nr_pages); |
55e462b05 memcg: simple sta... |
662 |
|
e401f1761 memcg: modify acc... |
663 664 |
/* pagein of a big page is an event. So, ignore page size */ if (nr_pages > 0) |
c0ff4b854 memcg: rename mem... |
665 |
__this_cpu_inc(memcg->stat->events[MEM_CGROUP_EVENTS_PGPGIN]); |
3751d6043 memcg: fix event ... |
666 |
else { |
c0ff4b854 memcg: rename mem... |
667 |
__this_cpu_inc(memcg->stat->events[MEM_CGROUP_EVENTS_PGPGOUT]); |
3751d6043 memcg: fix event ... |
668 669 |
nr_pages = -nr_pages; /* for event */ } |
e401f1761 memcg: modify acc... |
670 |
|
c0ff4b854 memcg: rename mem... |
671 |
__this_cpu_add(memcg->stat->events[MEM_CGROUP_EVENTS_COUNT], nr_pages); |
2e72b6347 memcg: implement ... |
672 |
|
c62b1a3b3 memcg: use generi... |
673 |
preempt_enable(); |
6d12e2d8d per-zone and recl... |
674 |
} |
bb2a0de92 memcg: consolidat... |
675 |
unsigned long |
c0ff4b854 memcg: rename mem... |
676 |
mem_cgroup_zone_nr_lru_pages(struct mem_cgroup *memcg, int nid, int zid, |
bb2a0de92 memcg: consolidat... |
677 |
unsigned int lru_mask) |
889976dbc memcg: reclaim me... |
678 679 |
{ struct mem_cgroup_per_zone *mz; |
bb2a0de92 memcg: consolidat... |
680 681 |
enum lru_list l; unsigned long ret = 0; |
c0ff4b854 memcg: rename mem... |
682 |
mz = mem_cgroup_zoneinfo(memcg, nid, zid); |
bb2a0de92 memcg: consolidat... |
683 684 685 686 687 688 689 690 691 |
for_each_lru(l) { if (BIT(l) & lru_mask) ret += MEM_CGROUP_ZSTAT(mz, l); } return ret; } static unsigned long |
c0ff4b854 memcg: rename mem... |
692 |
mem_cgroup_node_nr_lru_pages(struct mem_cgroup *memcg, |
bb2a0de92 memcg: consolidat... |
693 694 |
int nid, unsigned int lru_mask) { |
889976dbc memcg: reclaim me... |
695 696 |
u64 total = 0; int zid; |
bb2a0de92 memcg: consolidat... |
697 |
for (zid = 0; zid < MAX_NR_ZONES; zid++) |
c0ff4b854 memcg: rename mem... |
698 699 |
total += mem_cgroup_zone_nr_lru_pages(memcg, nid, zid, lru_mask); |
bb2a0de92 memcg: consolidat... |
700 |
|
889976dbc memcg: reclaim me... |
701 702 |
return total; } |
bb2a0de92 memcg: consolidat... |
703 |
|
c0ff4b854 memcg: rename mem... |
704 |
static unsigned long mem_cgroup_nr_lru_pages(struct mem_cgroup *memcg, |
bb2a0de92 memcg: consolidat... |
705 |
unsigned int lru_mask) |
6d12e2d8d per-zone and recl... |
706 |
{ |
889976dbc memcg: reclaim me... |
707 |
int nid; |
6d12e2d8d per-zone and recl... |
708 |
u64 total = 0; |
bb2a0de92 memcg: consolidat... |
709 |
for_each_node_state(nid, N_HIGH_MEMORY) |
c0ff4b854 memcg: rename mem... |
710 |
total += mem_cgroup_node_nr_lru_pages(memcg, nid, lru_mask); |
6d12e2d8d per-zone and recl... |
711 |
return total; |
d52aa412d memory cgroup enh... |
712 |
} |
c0ff4b854 memcg: rename mem... |
713 |
static bool __memcg_event_check(struct mem_cgroup *memcg, int target) |
7a159cc9d memcg: use native... |
714 715 |
{ unsigned long val, next; |
4799401fe memcg: Fix race c... |
716 717 |
val = __this_cpu_read(memcg->stat->events[MEM_CGROUP_EVENTS_COUNT]); next = __this_cpu_read(memcg->stat->targets[target]); |
7a159cc9d memcg: use native... |
718 719 720 |
/* from time_after() in jiffies.h */ return ((long)next - (long)val < 0); } |
c0ff4b854 memcg: rename mem... |
721 |
static void __mem_cgroup_target_update(struct mem_cgroup *memcg, int target) |
d2265e6fa memcg : share eve... |
722 |
{ |
7a159cc9d memcg: use native... |
723 |
unsigned long val, next; |
d2265e6fa memcg : share eve... |
724 |
|
4799401fe memcg: Fix race c... |
725 |
val = __this_cpu_read(memcg->stat->events[MEM_CGROUP_EVENTS_COUNT]); |
d2265e6fa memcg : share eve... |
726 |
|
7a159cc9d memcg: use native... |
727 728 729 730 731 732 733 |
switch (target) { case MEM_CGROUP_TARGET_THRESH: next = val + THRESHOLDS_EVENTS_TARGET; break; case MEM_CGROUP_TARGET_SOFTLIMIT: next = val + SOFTLIMIT_EVENTS_TARGET; break; |
453a9bf34 memcg: fix numa s... |
734 735 736 |
case MEM_CGROUP_TARGET_NUMAINFO: next = val + NUMAINFO_EVENTS_TARGET; break; |
7a159cc9d memcg: use native... |
737 738 739 |
default: return; } |
4799401fe memcg: Fix race c... |
740 |
__this_cpu_write(memcg->stat->targets[target], next); |
d2265e6fa memcg : share eve... |
741 742 743 744 745 746 |
} /* * Check events in order. * */ |
c0ff4b854 memcg: rename mem... |
747 |
static void memcg_check_events(struct mem_cgroup *memcg, struct page *page) |
d2265e6fa memcg : share eve... |
748 |
{ |
4799401fe memcg: Fix race c... |
749 |
preempt_disable(); |
d2265e6fa memcg : share eve... |
750 |
/* threshold event is triggered in finer grain than soft limit */ |
c0ff4b854 memcg: rename mem... |
751 752 753 754 |
if (unlikely(__memcg_event_check(memcg, MEM_CGROUP_TARGET_THRESH))) { mem_cgroup_threshold(memcg); __mem_cgroup_target_update(memcg, MEM_CGROUP_TARGET_THRESH); if (unlikely(__memcg_event_check(memcg, |
453a9bf34 memcg: fix numa s... |
755 |
MEM_CGROUP_TARGET_SOFTLIMIT))) { |
c0ff4b854 memcg: rename mem... |
756 757 |
mem_cgroup_update_tree(memcg, page); __mem_cgroup_target_update(memcg, |
453a9bf34 memcg: fix numa s... |
758 759 760 |
MEM_CGROUP_TARGET_SOFTLIMIT); } #if MAX_NUMNODES > 1 |
c0ff4b854 memcg: rename mem... |
761 |
if (unlikely(__memcg_event_check(memcg, |
453a9bf34 memcg: fix numa s... |
762 |
MEM_CGROUP_TARGET_NUMAINFO))) { |
c0ff4b854 memcg: rename mem... |
763 764 |
atomic_inc(&memcg->numainfo_events); __mem_cgroup_target_update(memcg, |
453a9bf34 memcg: fix numa s... |
765 |
MEM_CGROUP_TARGET_NUMAINFO); |
7a159cc9d memcg: use native... |
766 |
} |
453a9bf34 memcg: fix numa s... |
767 |
#endif |
d2265e6fa memcg : share eve... |
768 |
} |
4799401fe memcg: Fix race c... |
769 |
preempt_enable(); |
d2265e6fa memcg : share eve... |
770 |
} |
d1a4c0b37 tcp memory pressu... |
771 |
struct mem_cgroup *mem_cgroup_from_cont(struct cgroup *cont) |
8cdea7c05 Memory controller... |
772 773 774 775 776 |
{ return container_of(cgroup_subsys_state(cont, mem_cgroup_subsys_id), struct mem_cgroup, css); } |
cf475ad28 cgroups: add an o... |
777 |
struct mem_cgroup *mem_cgroup_from_task(struct task_struct *p) |
78fb74669 Memory controller... |
778 |
{ |
31a78f23b mm owner: fix rac... |
779 780 781 782 783 784 785 |
/* * 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... |
786 787 788 |
return container_of(task_subsys_state(p, mem_cgroup_subsys_id), struct mem_cgroup, css); } |
a433658c3 vmscan,memcg: mem... |
789 |
struct mem_cgroup *try_get_mem_cgroup_from_mm(struct mm_struct *mm) |
54595fe26 memcg: use css_tr... |
790 |
{ |
c0ff4b854 memcg: rename mem... |
791 |
struct mem_cgroup *memcg = NULL; |
0b7f569e4 memcg: fix OOM ki... |
792 793 794 |
if (!mm) return NULL; |
54595fe26 memcg: use css_tr... |
795 796 797 798 799 800 801 |
/* * 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... |
802 803 |
memcg = mem_cgroup_from_task(rcu_dereference(mm->owner)); if (unlikely(!memcg)) |
54595fe26 memcg: use css_tr... |
804 |
break; |
c0ff4b854 memcg: rename mem... |
805 |
} while (!css_tryget(&memcg->css)); |
54595fe26 memcg: use css_tr... |
806 |
rcu_read_unlock(); |
c0ff4b854 memcg: rename mem... |
807 |
return memcg; |
54595fe26 memcg: use css_tr... |
808 |
} |
7d74b06f2 memcg: use for_ea... |
809 |
/* The caller has to guarantee "mem" exists before calling this */ |
c0ff4b854 memcg: rename mem... |
810 |
static struct mem_cgroup *mem_cgroup_start_loop(struct mem_cgroup *memcg) |
14067bb3e memcg: hierarchic... |
811 |
{ |
711d3d2c9 memcg: cpu hotplu... |
812 813 |
struct cgroup_subsys_state *css; int found; |
c0ff4b854 memcg: rename mem... |
814 |
if (!memcg) /* ROOT cgroup has the smallest ID */ |
711d3d2c9 memcg: cpu hotplu... |
815 |
return root_mem_cgroup; /*css_put/get against root is ignored*/ |
c0ff4b854 memcg: rename mem... |
816 817 818 |
if (!memcg->use_hierarchy) { if (css_tryget(&memcg->css)) return memcg; |
711d3d2c9 memcg: cpu hotplu... |
819 820 821 822 823 824 825 |
return NULL; } rcu_read_lock(); /* * searching a memory cgroup which has the smallest ID under given * ROOT cgroup. (ID >= 1) */ |
c0ff4b854 memcg: rename mem... |
826 |
css = css_get_next(&mem_cgroup_subsys, 1, &memcg->css, &found); |
711d3d2c9 memcg: cpu hotplu... |
827 |
if (css && css_tryget(css)) |
c0ff4b854 memcg: rename mem... |
828 |
memcg = container_of(css, struct mem_cgroup, css); |
711d3d2c9 memcg: cpu hotplu... |
829 |
else |
c0ff4b854 memcg: rename mem... |
830 |
memcg = NULL; |
711d3d2c9 memcg: cpu hotplu... |
831 |
rcu_read_unlock(); |
c0ff4b854 memcg: rename mem... |
832 |
return memcg; |
7d74b06f2 memcg: use for_ea... |
833 834 835 836 837 838 839 840 841 |
} static struct mem_cgroup *mem_cgroup_get_next(struct mem_cgroup *iter, struct mem_cgroup *root, bool cond) { int nextid = css_id(&iter->css) + 1; int found; int hierarchy_used; |
14067bb3e memcg: hierarchic... |
842 |
struct cgroup_subsys_state *css; |
14067bb3e memcg: hierarchic... |
843 |
|
7d74b06f2 memcg: use for_ea... |
844 |
hierarchy_used = iter->use_hierarchy; |
14067bb3e memcg: hierarchic... |
845 |
|
7d74b06f2 memcg: use for_ea... |
846 |
css_put(&iter->css); |
711d3d2c9 memcg: cpu hotplu... |
847 848 |
/* If no ROOT, walk all, ignore hierarchy */ if (!cond || (root && !hierarchy_used)) |
7d74b06f2 memcg: use for_ea... |
849 |
return NULL; |
14067bb3e memcg: hierarchic... |
850 |
|
711d3d2c9 memcg: cpu hotplu... |
851 852 |
if (!root) root = root_mem_cgroup; |
7d74b06f2 memcg: use for_ea... |
853 854 |
do { iter = NULL; |
14067bb3e memcg: hierarchic... |
855 |
rcu_read_lock(); |
7d74b06f2 memcg: use for_ea... |
856 857 858 |
css = css_get_next(&mem_cgroup_subsys, nextid, &root->css, &found); |
14067bb3e memcg: hierarchic... |
859 |
if (css && css_tryget(css)) |
7d74b06f2 memcg: use for_ea... |
860 |
iter = container_of(css, struct mem_cgroup, css); |
14067bb3e memcg: hierarchic... |
861 |
rcu_read_unlock(); |
7d74b06f2 memcg: use for_ea... |
862 |
/* If css is NULL, no more cgroups will be found */ |
14067bb3e memcg: hierarchic... |
863 |
nextid = found + 1; |
7d74b06f2 memcg: use for_ea... |
864 |
} while (css && !iter); |
14067bb3e memcg: hierarchic... |
865 |
|
7d74b06f2 memcg: use for_ea... |
866 |
return iter; |
14067bb3e memcg: hierarchic... |
867 |
} |
7d74b06f2 memcg: use for_ea... |
868 869 870 871 872 873 874 875 876 877 878 879 |
/* * for_eacn_mem_cgroup_tree() for visiting all cgroup under tree. Please * be careful that "break" loop is not allowed. We have reference count. * Instead of that modify "cond" to be false and "continue" to exit the loop. */ #define for_each_mem_cgroup_tree_cond(iter, root, cond) \ for (iter = mem_cgroup_start_loop(root);\ iter != NULL;\ iter = mem_cgroup_get_next(iter, root, cond)) #define for_each_mem_cgroup_tree(iter, root) \ for_each_mem_cgroup_tree_cond(iter, root, true) |
711d3d2c9 memcg: cpu hotplu... |
880 881 |
#define for_each_mem_cgroup_all(iter) \ for_each_mem_cgroup_tree_cond(iter, NULL, true) |
14067bb3e memcg: hierarchic... |
882 |
|
c0ff4b854 memcg: rename mem... |
883 |
static inline bool mem_cgroup_is_root(struct mem_cgroup *memcg) |
4b3bde4c9 memcg: remove the... |
884 |
{ |
c0ff4b854 memcg: rename mem... |
885 |
return (memcg == root_mem_cgroup); |
4b3bde4c9 memcg: remove the... |
886 |
} |
456f998ec memcg: add the pa... |
887 888 |
void mem_cgroup_count_vm_event(struct mm_struct *mm, enum vm_event_item idx) { |
c0ff4b854 memcg: rename mem... |
889 |
struct mem_cgroup *memcg; |
456f998ec memcg: add the pa... |
890 891 892 893 894 |
if (!mm) return; rcu_read_lock(); |
c0ff4b854 memcg: rename mem... |
895 896 |
memcg = mem_cgroup_from_task(rcu_dereference(mm->owner)); if (unlikely(!memcg)) |
456f998ec memcg: add the pa... |
897 898 899 900 |
goto out; switch (idx) { case PGMAJFAULT: |
c0ff4b854 memcg: rename mem... |
901 |
mem_cgroup_pgmajfault(memcg, 1); |
456f998ec memcg: add the pa... |
902 903 |
break; case PGFAULT: |
c0ff4b854 memcg: rename mem... |
904 |
mem_cgroup_pgfault(memcg, 1); |
456f998ec memcg: add the pa... |
905 906 907 908 909 910 911 912 |
break; default: BUG(); } out: rcu_read_unlock(); } EXPORT_SYMBOL(mem_cgroup_count_vm_event); |
08e552c69 memcg: synchroniz... |
913 914 915 916 917 918 919 920 921 922 923 924 925 |
/* * 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... |
926 |
|
08e552c69 memcg: synchroniz... |
927 928 929 |
void mem_cgroup_del_lru_list(struct page *page, enum lru_list lru) { struct page_cgroup *pc; |
08e552c69 memcg: synchroniz... |
930 |
struct mem_cgroup_per_zone *mz; |
6d12e2d8d per-zone and recl... |
931 |
|
f8d665422 memcg: add mem_cg... |
932 |
if (mem_cgroup_disabled()) |
08e552c69 memcg: synchroniz... |
933 934 935 |
return; pc = lookup_page_cgroup(page); /* can happen while we handle swapcache. */ |
4b3bde4c9 memcg: remove the... |
936 |
if (!TestClearPageCgroupAcctLRU(pc)) |
08e552c69 memcg: synchroniz... |
937 |
return; |
4b3bde4c9 memcg: remove the... |
938 |
VM_BUG_ON(!pc->mem_cgroup); |
544122e5e memcg: fix LRU ac... |
939 940 941 942 |
/* * We don't check PCG_USED bit. It's cleared when the "page" is finally * removed from global LRU. */ |
97a6c37b3 memcg: change pag... |
943 |
mz = page_cgroup_zoneinfo(pc->mem_cgroup, page); |
ece35ca81 memcg: fix LRU ac... |
944 945 |
/* huge page split is done under lru_lock. so, we have no races. */ MEM_CGROUP_ZSTAT(mz, lru) -= 1 << compound_order(page); |
4b3bde4c9 memcg: remove the... |
946 947 948 |
if (mem_cgroup_is_root(pc->mem_cgroup)) return; VM_BUG_ON(list_empty(&pc->lru)); |
08e552c69 memcg: synchroniz... |
949 |
list_del_init(&pc->lru); |
6d12e2d8d per-zone and recl... |
950 |
} |
08e552c69 memcg: synchroniz... |
951 |
void mem_cgroup_del_lru(struct page *page) |
6d12e2d8d per-zone and recl... |
952 |
{ |
08e552c69 memcg: synchroniz... |
953 954 |
mem_cgroup_del_lru_list(page, page_lru(page)); } |
b69408e88 vmscan: Use an in... |
955 |
|
3f58a8294 memcg: move memcg... |
956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 |
/* * Writeback is about to end against a page which has been marked for immediate * reclaim. If it still appears to be reclaimable, move it to the tail of the * inactive list. */ void mem_cgroup_rotate_reclaimable_page(struct page *page) { struct mem_cgroup_per_zone *mz; struct page_cgroup *pc; enum lru_list lru = page_lru(page); if (mem_cgroup_disabled()) return; pc = lookup_page_cgroup(page); /* unused or root page is not rotated. */ if (!PageCgroupUsed(pc)) return; /* Ensure pc->mem_cgroup is visible after reading PCG_USED. */ smp_rmb(); if (mem_cgroup_is_root(pc->mem_cgroup)) return; |
97a6c37b3 memcg: change pag... |
978 |
mz = page_cgroup_zoneinfo(pc->mem_cgroup, page); |
3f58a8294 memcg: move memcg... |
979 980 |
list_move_tail(&pc->lru, &mz->lists[lru]); } |
08e552c69 memcg: synchroniz... |
981 982 983 984 |
void mem_cgroup_rotate_lru_list(struct page *page, enum lru_list lru) { struct mem_cgroup_per_zone *mz; struct page_cgroup *pc; |
b69408e88 vmscan: Use an in... |
985 |
|
f8d665422 memcg: add mem_cg... |
986 |
if (mem_cgroup_disabled()) |
08e552c69 memcg: synchroniz... |
987 |
return; |
6d12e2d8d per-zone and recl... |
988 |
|
08e552c69 memcg: synchroniz... |
989 |
pc = lookup_page_cgroup(page); |
4b3bde4c9 memcg: remove the... |
990 |
/* unused or root page is not rotated. */ |
713735b42 memcg: correctly ... |
991 992 993 994 995 |
if (!PageCgroupUsed(pc)) return; /* Ensure pc->mem_cgroup is visible after reading PCG_USED. */ smp_rmb(); if (mem_cgroup_is_root(pc->mem_cgroup)) |
08e552c69 memcg: synchroniz... |
996 |
return; |
97a6c37b3 memcg: change pag... |
997 |
mz = page_cgroup_zoneinfo(pc->mem_cgroup, page); |
08e552c69 memcg: synchroniz... |
998 |
list_move(&pc->lru, &mz->lists[lru]); |
6d12e2d8d per-zone and recl... |
999 |
} |
08e552c69 memcg: synchroniz... |
1000 |
void mem_cgroup_add_lru_list(struct page *page, enum lru_list lru) |
66e1707bc Memory controller... |
1001 |
{ |
08e552c69 memcg: synchroniz... |
1002 1003 |
struct page_cgroup *pc; struct mem_cgroup_per_zone *mz; |
6d12e2d8d per-zone and recl... |
1004 |
|
f8d665422 memcg: add mem_cg... |
1005 |
if (mem_cgroup_disabled()) |
08e552c69 memcg: synchroniz... |
1006 1007 |
return; pc = lookup_page_cgroup(page); |
4b3bde4c9 memcg: remove the... |
1008 |
VM_BUG_ON(PageCgroupAcctLRU(pc)); |
a61ed3cec memcg: close race... |
1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 |
/* * putback: charge: * SetPageLRU SetPageCgroupUsed * smp_mb smp_mb * PageCgroupUsed && add to memcg LRU PageLRU && add to memcg LRU * * Ensure that one of the two sides adds the page to the memcg * LRU during a race. */ smp_mb(); |
08e552c69 memcg: synchroniz... |
1019 |
if (!PageCgroupUsed(pc)) |
894bc3104 Unevictable LRU I... |
1020 |
return; |
713735b42 memcg: correctly ... |
1021 1022 |
/* Ensure pc->mem_cgroup is visible after reading PCG_USED. */ smp_rmb(); |
97a6c37b3 memcg: change pag... |
1023 |
mz = page_cgroup_zoneinfo(pc->mem_cgroup, page); |
ece35ca81 memcg: fix LRU ac... |
1024 1025 |
/* huge page split is done under lru_lock. so, we have no races. */ MEM_CGROUP_ZSTAT(mz, lru) += 1 << compound_order(page); |
4b3bde4c9 memcg: remove the... |
1026 1027 1028 |
SetPageCgroupAcctLRU(pc); if (mem_cgroup_is_root(pc->mem_cgroup)) return; |
08e552c69 memcg: synchroniz... |
1029 1030 |
list_add(&pc->lru, &mz->lists[lru]); } |
544122e5e memcg: fix LRU ac... |
1031 |
|
08e552c69 memcg: synchroniz... |
1032 |
/* |
5a6475a4e memcg: fix leak o... |
1033 1034 1035 1036 |
* At handling SwapCache and other FUSE stuff, pc->mem_cgroup may be changed * while it's linked to lru because the page may be reused after it's fully * uncharged. To handle that, unlink page_cgroup from LRU when charge it again. * It's done under lock_page and expected that zone->lru_lock isnever held. |
08e552c69 memcg: synchroniz... |
1037 |
*/ |
5a6475a4e memcg: fix leak o... |
1038 |
static void mem_cgroup_lru_del_before_commit(struct page *page) |
08e552c69 memcg: synchroniz... |
1039 |
{ |
544122e5e memcg: fix LRU ac... |
1040 1041 1042 |
unsigned long flags; struct zone *zone = page_zone(page); struct page_cgroup *pc = lookup_page_cgroup(page); |
5a6475a4e memcg: fix leak o... |
1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 |
/* * Doing this check without taking ->lru_lock seems wrong but this * is safe. Because if page_cgroup's USED bit is unset, the page * will not be added to any memcg's LRU. If page_cgroup's USED bit is * set, the commit after this will fail, anyway. * This all charge/uncharge is done under some mutual execustion. * So, we don't need to taking care of changes in USED bit. */ if (likely(!PageLRU(page))) return; |
544122e5e memcg: fix LRU ac... |
1053 1054 1055 1056 1057 1058 1059 1060 |
spin_lock_irqsave(&zone->lru_lock, flags); /* * Forget old LRU when this page_cgroup is *not* used. This Used bit * is guarded by lock_page() because the page is SwapCache. */ if (!PageCgroupUsed(pc)) mem_cgroup_del_lru_list(page, page_lru(page)); spin_unlock_irqrestore(&zone->lru_lock, flags); |
08e552c69 memcg: synchroniz... |
1061 |
} |
5a6475a4e memcg: fix leak o... |
1062 |
static void mem_cgroup_lru_add_after_commit(struct page *page) |
544122e5e memcg: fix LRU ac... |
1063 1064 1065 1066 |
{ unsigned long flags; struct zone *zone = page_zone(page); struct page_cgroup *pc = lookup_page_cgroup(page); |
a61ed3cec memcg: close race... |
1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 |
/* * putback: charge: * SetPageLRU SetPageCgroupUsed * smp_mb smp_mb * PageCgroupUsed && add to memcg LRU PageLRU && add to memcg LRU * * Ensure that one of the two sides adds the page to the memcg * LRU during a race. */ smp_mb(); |
5a6475a4e memcg: fix leak o... |
1077 1078 1079 |
/* taking care of that the page is added to LRU while we commit it */ if (likely(!PageLRU(page))) return; |
544122e5e memcg: fix LRU ac... |
1080 1081 |
spin_lock_irqsave(&zone->lru_lock, flags); /* link when the page is linked to LRU but page_cgroup isn't */ |
4b3bde4c9 memcg: remove the... |
1082 |
if (PageLRU(page) && !PageCgroupAcctLRU(pc)) |
544122e5e memcg: fix LRU ac... |
1083 1084 1085 |
mem_cgroup_add_lru_list(page, page_lru(page)); spin_unlock_irqrestore(&zone->lru_lock, flags); } |
08e552c69 memcg: synchroniz... |
1086 1087 1088 |
void mem_cgroup_move_lists(struct page *page, enum lru_list from, enum lru_list to) { |
f8d665422 memcg: add mem_cg... |
1089 |
if (mem_cgroup_disabled()) |
08e552c69 memcg: synchroniz... |
1090 1091 1092 |
return; mem_cgroup_del_lru_list(page, from); mem_cgroup_add_lru_list(page, to); |
66e1707bc Memory controller... |
1093 |
} |
3e92041d6 memcg: add mem_cg... |
1094 |
/* |
c0ff4b854 memcg: rename mem... |
1095 |
* Checks whether given mem is same or in the root_mem_cgroup's |
3e92041d6 memcg: add mem_cg... |
1096 1097 |
* hierarchy subtree */ |
c0ff4b854 memcg: rename mem... |
1098 1099 |
static bool mem_cgroup_same_or_subtree(const struct mem_cgroup *root_memcg, struct mem_cgroup *memcg) |
3e92041d6 memcg: add mem_cg... |
1100 |
{ |
c0ff4b854 memcg: rename mem... |
1101 1102 1103 |
if (root_memcg != memcg) { return (root_memcg->use_hierarchy && css_is_ancestor(&memcg->css, &root_memcg->css)); |
3e92041d6 memcg: add mem_cg... |
1104 1105 1106 1107 |
} return true; } |
c0ff4b854 memcg: rename mem... |
1108 |
int task_in_mem_cgroup(struct task_struct *task, const struct mem_cgroup *memcg) |
4c4a22148 memcontrol: move ... |
1109 1110 |
{ int ret; |
0b7f569e4 memcg: fix OOM ki... |
1111 |
struct mem_cgroup *curr = NULL; |
158e0a2d1 memcg: use find_l... |
1112 |
struct task_struct *p; |
4c4a22148 memcontrol: move ... |
1113 |
|
158e0a2d1 memcg: use find_l... |
1114 1115 1116 1117 1118 |
p = find_lock_task_mm(task); if (!p) return 0; curr = try_get_mem_cgroup_from_mm(p->mm); task_unlock(p); |
0b7f569e4 memcg: fix OOM ki... |
1119 1120 |
if (!curr) return 0; |
d31f56dbf memcg: avoid oom-... |
1121 |
/* |
c0ff4b854 memcg: rename mem... |
1122 |
* We should check use_hierarchy of "memcg" not "curr". Because checking |
d31f56dbf memcg: avoid oom-... |
1123 |
* use_hierarchy of "curr" here make this function true if hierarchy is |
c0ff4b854 memcg: rename mem... |
1124 1125 |
* 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-... |
1126 |
*/ |
c0ff4b854 memcg: rename mem... |
1127 |
ret = mem_cgroup_same_or_subtree(memcg, curr); |
0b7f569e4 memcg: fix OOM ki... |
1128 |
css_put(&curr->css); |
4c4a22148 memcontrol: move ... |
1129 1130 |
return ret; } |
9b272977e memcg: skip scann... |
1131 |
int mem_cgroup_inactive_anon_is_low(struct mem_cgroup *memcg, struct zone *zone) |
14797e236 memcg: add inacti... |
1132 |
{ |
9b272977e memcg: skip scann... |
1133 1134 1135 |
unsigned long inactive_ratio; int nid = zone_to_nid(zone); int zid = zone_idx(zone); |
14797e236 memcg: add inacti... |
1136 |
unsigned long inactive; |
9b272977e memcg: skip scann... |
1137 |
unsigned long active; |
c772be939 memcg: fix calcul... |
1138 |
unsigned long gb; |
14797e236 memcg: add inacti... |
1139 |
|
9b272977e memcg: skip scann... |
1140 1141 1142 1143 |
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... |
1144 |
|
c772be939 memcg: fix calcul... |
1145 1146 1147 1148 1149 |
gb = (inactive + active) >> (30 - PAGE_SHIFT); if (gb) inactive_ratio = int_sqrt(10 * gb); else inactive_ratio = 1; |
9b272977e memcg: skip scann... |
1150 |
return inactive * inactive_ratio < active; |
14797e236 memcg: add inacti... |
1151 |
} |
9b272977e memcg: skip scann... |
1152 |
int mem_cgroup_inactive_file_is_low(struct mem_cgroup *memcg, struct zone *zone) |
56e49d218 vmscan: evict use... |
1153 1154 1155 |
{ unsigned long active; unsigned long inactive; |
9b272977e memcg: skip scann... |
1156 1157 |
int zid = zone_idx(zone); int nid = zone_to_nid(zone); |
56e49d218 vmscan: evict use... |
1158 |
|
9b272977e memcg: skip scann... |
1159 1160 1161 1162 |
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... |
1163 1164 1165 |
return (active > inactive); } |
3e2f41f1f memcg: add zone_r... |
1166 1167 1168 |
struct zone_reclaim_stat *mem_cgroup_get_reclaim_stat(struct mem_cgroup *memcg, struct zone *zone) { |
13d7e3a2d memcg: convert to... |
1169 |
int nid = zone_to_nid(zone); |
3e2f41f1f memcg: add zone_r... |
1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 |
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... |
1186 1187 |
if (!PageCgroupUsed(pc)) return NULL; |
713735b42 memcg: correctly ... |
1188 1189 |
/* Ensure pc->mem_cgroup is visible after reading PCG_USED. */ smp_rmb(); |
97a6c37b3 memcg: change pag... |
1190 |
mz = page_cgroup_zoneinfo(pc->mem_cgroup, page); |
3e2f41f1f memcg: add zone_r... |
1191 1192 |
return &mz->reclaim_stat; } |
66e1707bc Memory controller... |
1193 1194 1195 |
unsigned long mem_cgroup_isolate_pages(unsigned long nr_to_scan, struct list_head *dst, unsigned long *scanned, int order, |
4356f21d0 mm: change isolat... |
1196 1197 |
isolate_mode_t mode, struct zone *z, |
66e1707bc Memory controller... |
1198 |
struct mem_cgroup *mem_cont, |
4f98a2fee vmscan: split LRU... |
1199 |
int active, int file) |
66e1707bc Memory controller... |
1200 1201 1202 1203 1204 1205 |
{ unsigned long nr_taken = 0; struct page *page; unsigned long scan; LIST_HEAD(pc_list); struct list_head *src; |
ff7283fa3 bugfix for memory... |
1206 |
struct page_cgroup *pc, *tmp; |
13d7e3a2d memcg: convert to... |
1207 |
int nid = zone_to_nid(z); |
1ecaab2bd per-zone and recl... |
1208 1209 |
int zid = zone_idx(z); struct mem_cgroup_per_zone *mz; |
b7c46d151 mm: drop unneeded... |
1210 |
int lru = LRU_FILE * file + active; |
2ffebca6a memcg: fix lru ro... |
1211 |
int ret; |
66e1707bc Memory controller... |
1212 |
|
cf475ad28 cgroups: add an o... |
1213 |
BUG_ON(!mem_cont); |
1ecaab2bd per-zone and recl... |
1214 |
mz = mem_cgroup_zoneinfo(mem_cont, nid, zid); |
b69408e88 vmscan: Use an in... |
1215 |
src = &mz->lists[lru]; |
66e1707bc Memory controller... |
1216 |
|
ff7283fa3 bugfix for memory... |
1217 1218 |
scan = 0; list_for_each_entry_safe_reverse(pc, tmp, src, lru) { |
436c6541b memcgroup: fix zo... |
1219 |
if (scan >= nr_to_scan) |
ff7283fa3 bugfix for memory... |
1220 |
break; |
08e552c69 memcg: synchroniz... |
1221 |
|
52d4b9ac0 memcg: allocate a... |
1222 1223 |
if (unlikely(!PageCgroupUsed(pc))) continue; |
5564e88ba memcg: condense p... |
1224 |
|
6b3ae58ef memcg: remove dir... |
1225 |
page = lookup_cgroup_page(pc); |
5564e88ba memcg: condense p... |
1226 |
|
436c6541b memcgroup: fix zo... |
1227 |
if (unlikely(!PageLRU(page))) |
ff7283fa3 bugfix for memory... |
1228 |
continue; |
ff7283fa3 bugfix for memory... |
1229 |
|
436c6541b memcgroup: fix zo... |
1230 |
scan++; |
2ffebca6a memcg: fix lru ro... |
1231 1232 1233 |
ret = __isolate_lru_page(page, mode, file); switch (ret) { case 0: |
66e1707bc Memory controller... |
1234 |
list_move(&page->lru, dst); |
2ffebca6a memcg: fix lru ro... |
1235 |
mem_cgroup_del_lru(page); |
2c888cfbc thp: fix anon mem... |
1236 |
nr_taken += hpage_nr_pages(page); |
2ffebca6a memcg: fix lru ro... |
1237 1238 1239 1240 1241 1242 1243 |
break; case -EBUSY: /* we don't affect global LRU but rotate in our LRU */ mem_cgroup_rotate_lru_list(page, page_lru(page)); break; default: break; |
66e1707bc Memory controller... |
1244 1245 |
} } |
66e1707bc Memory controller... |
1246 |
*scanned = scan; |
cc8e970c3 memcg: add mm_vms... |
1247 1248 1249 |
trace_mm_vmscan_memcg_isolate(0, nr_to_scan, scan, nr_taken, 0, 0, 0, mode); |
66e1707bc Memory controller... |
1250 1251 |
return nr_taken; } |
6d61ef409 memcg: memory cgr... |
1252 1253 |
#define mem_cgroup_from_res_counter(counter, member) \ container_of(counter, struct mem_cgroup, member) |
19942822d memcg: prevent en... |
1254 |
/** |
9d11ea9f1 memcg: simplify t... |
1255 1256 |
* mem_cgroup_margin - calculate chargeable space of a memory cgroup * @mem: the memory cgroup |
19942822d memcg: prevent en... |
1257 |
* |
9d11ea9f1 memcg: simplify t... |
1258 |
* Returns the maximum amount of memory @mem can be charged with, in |
7ec99d621 memcg: unify char... |
1259 |
* pages. |
19942822d memcg: prevent en... |
1260 |
*/ |
c0ff4b854 memcg: rename mem... |
1261 |
static unsigned long mem_cgroup_margin(struct mem_cgroup *memcg) |
19942822d memcg: prevent en... |
1262 |
{ |
9d11ea9f1 memcg: simplify t... |
1263 |
unsigned long long margin; |
c0ff4b854 memcg: rename mem... |
1264 |
margin = res_counter_margin(&memcg->res); |
9d11ea9f1 memcg: simplify t... |
1265 |
if (do_swap_account) |
c0ff4b854 memcg: rename mem... |
1266 |
margin = min(margin, res_counter_margin(&memcg->memsw)); |
7ec99d621 memcg: unify char... |
1267 |
return margin >> PAGE_SHIFT; |
19942822d memcg: prevent en... |
1268 |
} |
1f4c025b5 memcg: export mem... |
1269 |
int mem_cgroup_swappiness(struct mem_cgroup *memcg) |
a7885eb8a memcg: swappiness |
1270 1271 |
{ struct cgroup *cgrp = memcg->css.cgroup; |
a7885eb8a memcg: swappiness |
1272 1273 1274 1275 |
/* root ? */ if (cgrp->parent == NULL) return vm_swappiness; |
bf1ff2635 memcg: remove mem... |
1276 |
return memcg->swappiness; |
a7885eb8a memcg: swappiness |
1277 |
} |
c0ff4b854 memcg: rename mem... |
1278 |
static void mem_cgroup_start_move(struct mem_cgroup *memcg) |
32047e2a8 memcg: avoid lock... |
1279 1280 |
{ int cpu; |
1489ebad8 memcg: cpu hotplu... |
1281 1282 |
get_online_cpus(); |
c0ff4b854 memcg: rename mem... |
1283 |
spin_lock(&memcg->pcp_counter_lock); |
1489ebad8 memcg: cpu hotplu... |
1284 |
for_each_online_cpu(cpu) |
c0ff4b854 memcg: rename mem... |
1285 1286 1287 |
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... |
1288 |
put_online_cpus(); |
32047e2a8 memcg: avoid lock... |
1289 1290 1291 |
synchronize_rcu(); } |
c0ff4b854 memcg: rename mem... |
1292 |
static void mem_cgroup_end_move(struct mem_cgroup *memcg) |
32047e2a8 memcg: avoid lock... |
1293 1294 |
{ int cpu; |
c0ff4b854 memcg: rename mem... |
1295 |
if (!memcg) |
32047e2a8 memcg: avoid lock... |
1296 |
return; |
1489ebad8 memcg: cpu hotplu... |
1297 |
get_online_cpus(); |
c0ff4b854 memcg: rename mem... |
1298 |
spin_lock(&memcg->pcp_counter_lock); |
1489ebad8 memcg: cpu hotplu... |
1299 |
for_each_online_cpu(cpu) |
c0ff4b854 memcg: rename mem... |
1300 1301 1302 |
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... |
1303 |
put_online_cpus(); |
32047e2a8 memcg: avoid lock... |
1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 |
} /* * 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... |
1316 |
static bool mem_cgroup_stealed(struct mem_cgroup *memcg) |
32047e2a8 memcg: avoid lock... |
1317 1318 |
{ VM_BUG_ON(!rcu_read_lock_held()); |
c0ff4b854 memcg: rename mem... |
1319 |
return this_cpu_read(memcg->stat->count[MEM_CGROUP_ON_MOVE]) > 0; |
32047e2a8 memcg: avoid lock... |
1320 |
} |
4b5343346 memcg: clean up t... |
1321 |
|
c0ff4b854 memcg: rename mem... |
1322 |
static bool mem_cgroup_under_move(struct mem_cgroup *memcg) |
4b5343346 memcg: clean up t... |
1323 |
{ |
2bd9bb206 memcg: clean up w... |
1324 1325 |
struct mem_cgroup *from; struct mem_cgroup *to; |
4b5343346 memcg: clean up t... |
1326 |
bool ret = false; |
2bd9bb206 memcg: clean up w... |
1327 1328 1329 1330 1331 1332 1333 1334 1335 |
/* * 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... |
1336 |
|
c0ff4b854 memcg: rename mem... |
1337 1338 |
ret = mem_cgroup_same_or_subtree(memcg, from) || mem_cgroup_same_or_subtree(memcg, to); |
2bd9bb206 memcg: clean up w... |
1339 1340 |
unlock: spin_unlock(&mc.lock); |
4b5343346 memcg: clean up t... |
1341 1342 |
return ret; } |
c0ff4b854 memcg: rename mem... |
1343 |
static bool mem_cgroup_wait_acct_move(struct mem_cgroup *memcg) |
4b5343346 memcg: clean up t... |
1344 1345 |
{ if (mc.moving_task && current != mc.moving_task) { |
c0ff4b854 memcg: rename mem... |
1346 |
if (mem_cgroup_under_move(memcg)) { |
4b5343346 memcg: clean up t... |
1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 |
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... |
1358 |
/** |
6a6135b64 memcg: typo in co... |
1359 |
* mem_cgroup_print_oom_info: Called from OOM with tasklist_lock held in read mode. |
e222432bf memcg: show memcg... |
1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 |
* @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-... |
1377 |
if (!memcg || !p) |
e222432bf memcg: show memcg... |
1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 |
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... |
1426 1427 1428 1429 |
/* * This function returns the number of memcg under hierarchy tree. Returns * 1(self count) if no children. */ |
c0ff4b854 memcg: rename mem... |
1430 |
static int mem_cgroup_count_children(struct mem_cgroup *memcg) |
81d39c20f memcg: fix shrink... |
1431 1432 |
{ int num = 0; |
7d74b06f2 memcg: use for_ea... |
1433 |
struct mem_cgroup *iter; |
c0ff4b854 memcg: rename mem... |
1434 |
for_each_mem_cgroup_tree(iter, memcg) |
7d74b06f2 memcg: use for_ea... |
1435 |
num++; |
81d39c20f memcg: fix shrink... |
1436 1437 |
return num; } |
6d61ef409 memcg: memory cgr... |
1438 |
/* |
a63d83f42 oom: badness heur... |
1439 1440 1441 1442 1443 1444 |
* 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... |
1445 1446 |
limit = res_counter_read_u64(&memcg->res, RES_LIMIT); limit += total_swap_pages << PAGE_SHIFT; |
a63d83f42 oom: badness heur... |
1447 1448 1449 1450 1451 1452 1453 1454 1455 |
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); } /* |
04046e1a0 memcg: use CSS ID |
1456 1457 1458 1459 1460 |
* Visit the first child (need not be the first child as per the ordering * of the cgroup list, since we track last_scanned_child) of @mem and use * that to reclaim free pages from. */ static struct mem_cgroup * |
c0ff4b854 memcg: rename mem... |
1461 |
mem_cgroup_select_victim(struct mem_cgroup *root_memcg) |
04046e1a0 memcg: use CSS ID |
1462 1463 1464 1465 |
{ struct mem_cgroup *ret = NULL; struct cgroup_subsys_state *css; int nextid, found; |
c0ff4b854 memcg: rename mem... |
1466 1467 1468 |
if (!root_memcg->use_hierarchy) { css_get(&root_memcg->css); ret = root_memcg; |
04046e1a0 memcg: use CSS ID |
1469 1470 1471 1472 |
} while (!ret) { rcu_read_lock(); |
c0ff4b854 memcg: rename mem... |
1473 1474 |
nextid = root_memcg->last_scanned_child + 1; css = css_get_next(&mem_cgroup_subsys, nextid, &root_memcg->css, |
04046e1a0 memcg: use CSS ID |
1475 1476 1477 1478 1479 1480 |
&found); if (css && css_tryget(css)) ret = container_of(css, struct mem_cgroup, css); rcu_read_unlock(); /* Updates scanning parameter */ |
04046e1a0 memcg: use CSS ID |
1481 1482 |
if (!css) { /* this means start scan from ID:1 */ |
c0ff4b854 memcg: rename mem... |
1483 |
root_memcg->last_scanned_child = 0; |
04046e1a0 memcg: use CSS ID |
1484 |
} else |
c0ff4b854 memcg: rename mem... |
1485 |
root_memcg->last_scanned_child = found; |
04046e1a0 memcg: use CSS ID |
1486 1487 1488 1489 |
} return ret; } |
4d0c066d2 memcg: fix reclai... |
1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 |
/** * 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... |
1500 |
static bool test_mem_cgroup_node_reclaimable(struct mem_cgroup *memcg, |
4d0c066d2 memcg: fix reclai... |
1501 1502 |
int nid, bool noswap) { |
c0ff4b854 memcg: rename mem... |
1503 |
if (mem_cgroup_node_nr_lru_pages(memcg, nid, LRU_ALL_FILE)) |
4d0c066d2 memcg: fix reclai... |
1504 1505 1506 |
return true; if (noswap || !total_swap_pages) return false; |
c0ff4b854 memcg: rename mem... |
1507 |
if (mem_cgroup_node_nr_lru_pages(memcg, nid, LRU_ALL_ANON)) |
4d0c066d2 memcg: fix reclai... |
1508 1509 1510 1511 |
return true; return false; } |
889976dbc memcg: reclaim me... |
1512 1513 1514 1515 1516 1517 1518 1519 |
#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... |
1520 |
static void mem_cgroup_may_update_nodemask(struct mem_cgroup *memcg) |
889976dbc memcg: reclaim me... |
1521 1522 |
{ int nid; |
453a9bf34 memcg: fix numa s... |
1523 1524 1525 1526 |
/* * numainfo_events > 0 means there was at least NUMAINFO_EVENTS_TARGET * pagein/pageout changes since the last update. */ |
c0ff4b854 memcg: rename mem... |
1527 |
if (!atomic_read(&memcg->numainfo_events)) |
453a9bf34 memcg: fix numa s... |
1528 |
return; |
c0ff4b854 memcg: rename mem... |
1529 |
if (atomic_inc_return(&memcg->numainfo_updating) > 1) |
889976dbc memcg: reclaim me... |
1530 |
return; |
889976dbc memcg: reclaim me... |
1531 |
/* make a nodemask where this memcg uses memory from */ |
c0ff4b854 memcg: rename mem... |
1532 |
memcg->scan_nodes = node_states[N_HIGH_MEMORY]; |
889976dbc memcg: reclaim me... |
1533 1534 |
for_each_node_mask(nid, node_states[N_HIGH_MEMORY]) { |
c0ff4b854 memcg: rename mem... |
1535 1536 |
if (!test_mem_cgroup_node_reclaimable(memcg, nid, false)) node_clear(nid, memcg->scan_nodes); |
889976dbc memcg: reclaim me... |
1537 |
} |
453a9bf34 memcg: fix numa s... |
1538 |
|
c0ff4b854 memcg: rename mem... |
1539 1540 |
atomic_set(&memcg->numainfo_events, 0); atomic_set(&memcg->numainfo_updating, 0); |
889976dbc memcg: reclaim me... |
1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 |
} /* * 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... |
1555 |
int mem_cgroup_select_victim_node(struct mem_cgroup *memcg) |
889976dbc memcg: reclaim me... |
1556 1557 |
{ int node; |
c0ff4b854 memcg: rename mem... |
1558 1559 |
mem_cgroup_may_update_nodemask(memcg); node = memcg->last_scanned_node; |
889976dbc memcg: reclaim me... |
1560 |
|
c0ff4b854 memcg: rename mem... |
1561 |
node = next_node(node, memcg->scan_nodes); |
889976dbc memcg: reclaim me... |
1562 |
if (node == MAX_NUMNODES) |
c0ff4b854 memcg: rename mem... |
1563 |
node = first_node(memcg->scan_nodes); |
889976dbc memcg: reclaim me... |
1564 1565 1566 1567 1568 1569 1570 1571 |
/* * 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... |
1572 |
memcg->last_scanned_node = node; |
889976dbc memcg: reclaim me... |
1573 1574 |
return node; } |
4d0c066d2 memcg: fix reclai... |
1575 1576 1577 1578 1579 1580 |
/* * 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... |
1581 |
bool mem_cgroup_reclaimable(struct mem_cgroup *memcg, bool noswap) |
4d0c066d2 memcg: fix reclai... |
1582 1583 1584 1585 1586 1587 1588 |
{ int nid; /* * quick check...making use of scan_node. * We can skip unused nodes. */ |
c0ff4b854 memcg: rename mem... |
1589 1590 |
if (!nodes_empty(memcg->scan_nodes)) { for (nid = first_node(memcg->scan_nodes); |
4d0c066d2 memcg: fix reclai... |
1591 |
nid < MAX_NUMNODES; |
c0ff4b854 memcg: rename mem... |
1592 |
nid = next_node(nid, memcg->scan_nodes)) { |
4d0c066d2 memcg: fix reclai... |
1593 |
|
c0ff4b854 memcg: rename mem... |
1594 |
if (test_mem_cgroup_node_reclaimable(memcg, nid, noswap)) |
4d0c066d2 memcg: fix reclai... |
1595 1596 1597 1598 1599 1600 1601 |
return true; } } /* * Check rest of nodes. */ for_each_node_state(nid, N_HIGH_MEMORY) { |
c0ff4b854 memcg: rename mem... |
1602 |
if (node_isset(nid, memcg->scan_nodes)) |
4d0c066d2 memcg: fix reclai... |
1603 |
continue; |
c0ff4b854 memcg: rename mem... |
1604 |
if (test_mem_cgroup_node_reclaimable(memcg, nid, noswap)) |
4d0c066d2 memcg: fix reclai... |
1605 1606 1607 1608 |
return true; } return false; } |
889976dbc memcg: reclaim me... |
1609 |
#else |
c0ff4b854 memcg: rename mem... |
1610 |
int mem_cgroup_select_victim_node(struct mem_cgroup *memcg) |
889976dbc memcg: reclaim me... |
1611 1612 1613 |
{ return 0; } |
4d0c066d2 memcg: fix reclai... |
1614 |
|
c0ff4b854 memcg: rename mem... |
1615 |
bool mem_cgroup_reclaimable(struct mem_cgroup *memcg, bool noswap) |
4d0c066d2 memcg: fix reclai... |
1616 |
{ |
c0ff4b854 memcg: rename mem... |
1617 |
return test_mem_cgroup_node_reclaimable(memcg, 0, noswap); |
4d0c066d2 memcg: fix reclai... |
1618 |
} |
889976dbc memcg: reclaim me... |
1619 |
#endif |
04046e1a0 memcg: use CSS ID |
1620 1621 1622 1623 |
/* * Scan the hierarchy if needed to reclaim memory. We remember the last child * we reclaimed from, so that we don't end up penalizing one child extensively * based on its position in the children list. |
6d61ef409 memcg: memory cgr... |
1624 |
* |
c0ff4b854 memcg: rename mem... |
1625 |
* root_memcg is the original ancestor that we've been reclaim from. |
04046e1a0 memcg: use CSS ID |
1626 |
* |
c0ff4b854 memcg: rename mem... |
1627 |
* We give up and return to the caller when we visit root_memcg twice. |
04046e1a0 memcg: use CSS ID |
1628 |
* (other groups can be removed while we're walking....) |
81d39c20f memcg: fix shrink... |
1629 1630 |
* * If shrink==true, for avoiding to free too much, this returns immedieately. |
6d61ef409 memcg: memory cgr... |
1631 |
*/ |
c0ff4b854 memcg: rename mem... |
1632 |
static int mem_cgroup_hierarchical_reclaim(struct mem_cgroup *root_memcg, |
4e4169535 memory controller... |
1633 |
struct zone *zone, |
75822b449 memory controller... |
1634 |
gfp_t gfp_mask, |
0ae5e89c6 memcg: count the ... |
1635 1636 |
unsigned long reclaim_options, unsigned long *total_scanned) |
6d61ef409 memcg: memory cgr... |
1637 |
{ |
04046e1a0 memcg: use CSS ID |
1638 1639 1640 |
struct mem_cgroup *victim; int ret, total = 0; int loop = 0; |
75822b449 memory controller... |
1641 1642 |
bool noswap = reclaim_options & MEM_CGROUP_RECLAIM_NOSWAP; bool shrink = reclaim_options & MEM_CGROUP_RECLAIM_SHRINK; |
4e4169535 memory controller... |
1643 |
bool check_soft = reclaim_options & MEM_CGROUP_RECLAIM_SOFT; |
9d11ea9f1 memcg: simplify t... |
1644 |
unsigned long excess; |
185efc0f9 memcg: Revert "me... |
1645 |
unsigned long nr_scanned; |
9d11ea9f1 memcg: simplify t... |
1646 |
|
c0ff4b854 memcg: rename mem... |
1647 |
excess = res_counter_soft_limit_excess(&root_memcg->res) >> PAGE_SHIFT; |
04046e1a0 memcg: use CSS ID |
1648 |
|
22a668d7c memcg: fix behavi... |
1649 |
/* If memsw_is_minimum==1, swap-out is of-no-use. */ |
c0ff4b854 memcg: rename mem... |
1650 |
if (!check_soft && !shrink && root_memcg->memsw_is_minimum) |
22a668d7c memcg: fix behavi... |
1651 |
noswap = true; |
4e4169535 memory controller... |
1652 |
while (1) { |
c0ff4b854 memcg: rename mem... |
1653 1654 |
victim = mem_cgroup_select_victim(root_memcg); if (victim == root_memcg) { |
04046e1a0 memcg: use CSS ID |
1655 |
loop++; |
fbc29a25e memcg: avoid perc... |
1656 1657 1658 1659 1660 1661 1662 |
/* * We are not draining per cpu cached charges during * soft limit reclaim because global reclaim doesn't * care about charges. It tries to free some memory and * charges will not give any. */ if (!check_soft && loop >= 1) |
c0ff4b854 memcg: rename mem... |
1663 |
drain_all_stock_async(root_memcg); |
4e4169535 memory controller... |
1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 |
if (loop >= 2) { /* * If we have not been able to reclaim * anything, it might because there are * no reclaimable pages under this hierarchy */ if (!check_soft || !total) { css_put(&victim->css); break; } /* |
25985edce Fix common misspe... |
1675 |
* We want to do more targeted reclaim. |
4e4169535 memory controller... |
1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 |
* excess >> 2 is not to excessive so as to * reclaim too much, nor too less that we keep * coming back to reclaim from this cgroup */ if (total >= (excess >> 2) || (loop > MEM_CGROUP_MAX_RECLAIM_LOOPS)) { css_put(&victim->css); break; } } } |
4d0c066d2 memcg: fix reclai... |
1687 |
if (!mem_cgroup_reclaimable(victim, noswap)) { |
04046e1a0 memcg: use CSS ID |
1688 1689 |
/* this cgroup's local usage == 0 */ css_put(&victim->css); |
6d61ef409 memcg: memory cgr... |
1690 1691 |
continue; } |
04046e1a0 memcg: use CSS ID |
1692 |
/* we use swappiness of local cgroup */ |
0ae5e89c6 memcg: count the ... |
1693 |
if (check_soft) { |
4e4169535 memory controller... |
1694 |
ret = mem_cgroup_shrink_node_zone(victim, gfp_mask, |
185efc0f9 memcg: Revert "me... |
1695 1696 |
noswap, zone, &nr_scanned); *total_scanned += nr_scanned; |
0ae5e89c6 memcg: count the ... |
1697 |
} else |
4e4169535 memory controller... |
1698 |
ret = try_to_free_mem_cgroup_pages(victim, gfp_mask, |
185efc0f9 memcg: Revert "me... |
1699 |
noswap); |
04046e1a0 memcg: use CSS ID |
1700 |
css_put(&victim->css); |
81d39c20f memcg: fix shrink... |
1701 1702 1703 1704 1705 1706 1707 |
/* * At shrinking usage, we can't check we should stop here or * reclaim more. It's depends on callers. last_scanned_child * will work enough for keeping fairness under tree. */ if (shrink) return ret; |
04046e1a0 memcg: use CSS ID |
1708 |
total += ret; |
4e4169535 memory controller... |
1709 |
if (check_soft) { |
c0ff4b854 memcg: rename mem... |
1710 |
if (!res_counter_soft_limit_excess(&root_memcg->res)) |
4e4169535 memory controller... |
1711 |
return total; |
c0ff4b854 memcg: rename mem... |
1712 |
} else if (mem_cgroup_margin(root_memcg)) |
4fd14ebf6 memcg: remove unu... |
1713 |
return total; |
6d61ef409 memcg: memory cgr... |
1714 |
} |
04046e1a0 memcg: use CSS ID |
1715 |
return total; |
6d61ef409 memcg: memory cgr... |
1716 |
} |
867578cbc memcg: fix oom ki... |
1717 1718 1719 |
/* * Check OOM-Killer is already running under our hierarchy. * If someone is running, return false. |
1af8efe96 memcg: change mem... |
1720 |
* Has to be called with memcg_oom_lock |
867578cbc memcg: fix oom ki... |
1721 |
*/ |
c0ff4b854 memcg: rename mem... |
1722 |
static bool mem_cgroup_oom_lock(struct mem_cgroup *memcg) |
867578cbc memcg: fix oom ki... |
1723 |
{ |
79dfdaccd memcg: make oom_l... |
1724 1725 |
struct mem_cgroup *iter, *failed = NULL; bool cond = true; |
a636b327f memcg: avoid unne... |
1726 |
|
c0ff4b854 memcg: rename mem... |
1727 |
for_each_mem_cgroup_tree_cond(iter, memcg, cond) { |
23751be00 memcg: fix hierar... |
1728 |
if (iter->oom_lock) { |
79dfdaccd memcg: make oom_l... |
1729 1730 1731 1732 |
/* * this subtree of our hierarchy is already locked * so we cannot give a lock. */ |
79dfdaccd memcg: make oom_l... |
1733 1734 |
failed = iter; cond = false; |
23751be00 memcg: fix hierar... |
1735 1736 |
} else iter->oom_lock = true; |
7d74b06f2 memcg: use for_ea... |
1737 |
} |
867578cbc memcg: fix oom ki... |
1738 |
|
79dfdaccd memcg: make oom_l... |
1739 |
if (!failed) |
23751be00 memcg: fix hierar... |
1740 |
return true; |
79dfdaccd memcg: make oom_l... |
1741 1742 1743 1744 1745 1746 |
/* * OK, we failed to lock the whole subtree so we have to clean up * what we set up to the failing subtree */ cond = true; |
c0ff4b854 memcg: rename mem... |
1747 |
for_each_mem_cgroup_tree_cond(iter, memcg, cond) { |
79dfdaccd memcg: make oom_l... |
1748 1749 1750 1751 1752 1753 |
if (iter == failed) { cond = false; continue; } iter->oom_lock = false; } |
23751be00 memcg: fix hierar... |
1754 |
return false; |
a636b327f memcg: avoid unne... |
1755 |
} |
0b7f569e4 memcg: fix OOM ki... |
1756 |
|
79dfdaccd memcg: make oom_l... |
1757 |
/* |
1af8efe96 memcg: change mem... |
1758 |
* Has to be called with memcg_oom_lock |
79dfdaccd memcg: make oom_l... |
1759 |
*/ |
c0ff4b854 memcg: rename mem... |
1760 |
static int mem_cgroup_oom_unlock(struct mem_cgroup *memcg) |
0b7f569e4 memcg: fix OOM ki... |
1761 |
{ |
7d74b06f2 memcg: use for_ea... |
1762 |
struct mem_cgroup *iter; |
c0ff4b854 memcg: rename mem... |
1763 |
for_each_mem_cgroup_tree(iter, memcg) |
79dfdaccd memcg: make oom_l... |
1764 1765 1766 |
iter->oom_lock = false; return 0; } |
c0ff4b854 memcg: rename mem... |
1767 |
static void mem_cgroup_mark_under_oom(struct mem_cgroup *memcg) |
79dfdaccd memcg: make oom_l... |
1768 1769 |
{ struct mem_cgroup *iter; |
c0ff4b854 memcg: rename mem... |
1770 |
for_each_mem_cgroup_tree(iter, memcg) |
79dfdaccd memcg: make oom_l... |
1771 1772 |
atomic_inc(&iter->under_oom); } |
c0ff4b854 memcg: rename mem... |
1773 |
static void mem_cgroup_unmark_under_oom(struct mem_cgroup *memcg) |
79dfdaccd memcg: make oom_l... |
1774 1775 |
{ struct mem_cgroup *iter; |
867578cbc memcg: fix oom ki... |
1776 1777 1778 1779 1780 |
/* * 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... |
1781 |
for_each_mem_cgroup_tree(iter, memcg) |
79dfdaccd memcg: make oom_l... |
1782 |
atomic_add_unless(&iter->under_oom, -1, 0); |
0b7f569e4 memcg: fix OOM ki... |
1783 |
} |
1af8efe96 memcg: change mem... |
1784 |
static DEFINE_SPINLOCK(memcg_oom_lock); |
867578cbc memcg: fix oom ki... |
1785 |
static DECLARE_WAIT_QUEUE_HEAD(memcg_oom_waitq); |
dc98df5a1 memcg: oom wakeup... |
1786 1787 1788 1789 1790 1791 1792 1793 |
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... |
1794 1795 |
struct mem_cgroup *wake_memcg = (struct mem_cgroup *)arg, *oom_wait_memcg; |
dc98df5a1 memcg: oom wakeup... |
1796 1797 1798 |
struct oom_wait_info *oom_wait_info; oom_wait_info = container_of(wait, struct oom_wait_info, wait); |
c0ff4b854 memcg: rename mem... |
1799 |
oom_wait_memcg = oom_wait_info->mem; |
dc98df5a1 memcg: oom wakeup... |
1800 |
|
dc98df5a1 memcg: oom wakeup... |
1801 1802 1803 1804 |
/* * 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... |
1805 1806 |
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... |
1807 |
return 0; |
dc98df5a1 memcg: oom wakeup... |
1808 1809 |
return autoremove_wake_function(wait, mode, sync, arg); } |
c0ff4b854 memcg: rename mem... |
1810 |
static void memcg_wakeup_oom(struct mem_cgroup *memcg) |
dc98df5a1 memcg: oom wakeup... |
1811 |
{ |
c0ff4b854 memcg: rename mem... |
1812 1813 |
/* for filtering, pass "memcg" as argument. */ __wake_up(&memcg_oom_waitq, TASK_NORMAL, 0, memcg); |
dc98df5a1 memcg: oom wakeup... |
1814 |
} |
c0ff4b854 memcg: rename mem... |
1815 |
static void memcg_oom_recover(struct mem_cgroup *memcg) |
3c11ecf44 memcg: oom kill d... |
1816 |
{ |
c0ff4b854 memcg: rename mem... |
1817 1818 |
if (memcg && atomic_read(&memcg->under_oom)) memcg_wakeup_oom(memcg); |
3c11ecf44 memcg: oom kill d... |
1819 |
} |
867578cbc memcg: fix oom ki... |
1820 1821 1822 |
/* * try to call OOM killer. returns false if we should exit memory-reclaim loop. */ |
c0ff4b854 memcg: rename mem... |
1823 |
bool mem_cgroup_handle_oom(struct mem_cgroup *memcg, gfp_t mask) |
0b7f569e4 memcg: fix OOM ki... |
1824 |
{ |
dc98df5a1 memcg: oom wakeup... |
1825 |
struct oom_wait_info owait; |
3c11ecf44 memcg: oom kill d... |
1826 |
bool locked, need_to_kill; |
867578cbc memcg: fix oom ki... |
1827 |
|
c0ff4b854 memcg: rename mem... |
1828 |
owait.mem = memcg; |
dc98df5a1 memcg: oom wakeup... |
1829 1830 1831 1832 |
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... |
1833 |
need_to_kill = true; |
c0ff4b854 memcg: rename mem... |
1834 |
mem_cgroup_mark_under_oom(memcg); |
79dfdaccd memcg: make oom_l... |
1835 |
|
c0ff4b854 memcg: rename mem... |
1836 |
/* At first, try to OOM lock hierarchy under memcg.*/ |
1af8efe96 memcg: change mem... |
1837 |
spin_lock(&memcg_oom_lock); |
c0ff4b854 memcg: rename mem... |
1838 |
locked = mem_cgroup_oom_lock(memcg); |
867578cbc memcg: fix oom ki... |
1839 1840 1841 1842 1843 |
/* * 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... |
1844 |
prepare_to_wait(&memcg_oom_waitq, &owait.wait, TASK_KILLABLE); |
c0ff4b854 memcg: rename mem... |
1845 |
if (!locked || memcg->oom_kill_disable) |
3c11ecf44 memcg: oom kill d... |
1846 1847 |
need_to_kill = false; if (locked) |
c0ff4b854 memcg: rename mem... |
1848 |
mem_cgroup_oom_notify(memcg); |
1af8efe96 memcg: change mem... |
1849 |
spin_unlock(&memcg_oom_lock); |
867578cbc memcg: fix oom ki... |
1850 |
|
3c11ecf44 memcg: oom kill d... |
1851 1852 |
if (need_to_kill) { finish_wait(&memcg_oom_waitq, &owait.wait); |
c0ff4b854 memcg: rename mem... |
1853 |
mem_cgroup_out_of_memory(memcg, mask); |
3c11ecf44 memcg: oom kill d... |
1854 |
} else { |
867578cbc memcg: fix oom ki... |
1855 |
schedule(); |
dc98df5a1 memcg: oom wakeup... |
1856 |
finish_wait(&memcg_oom_waitq, &owait.wait); |
867578cbc memcg: fix oom ki... |
1857 |
} |
1af8efe96 memcg: change mem... |
1858 |
spin_lock(&memcg_oom_lock); |
79dfdaccd memcg: make oom_l... |
1859 |
if (locked) |
c0ff4b854 memcg: rename mem... |
1860 1861 |
mem_cgroup_oom_unlock(memcg); memcg_wakeup_oom(memcg); |
1af8efe96 memcg: change mem... |
1862 |
spin_unlock(&memcg_oom_lock); |
867578cbc memcg: fix oom ki... |
1863 |
|
c0ff4b854 memcg: rename mem... |
1864 |
mem_cgroup_unmark_under_oom(memcg); |
79dfdaccd memcg: make oom_l... |
1865 |
|
867578cbc memcg: fix oom ki... |
1866 1867 1868 |
if (test_thread_flag(TIF_MEMDIE) || fatal_signal_pending(current)) return false; /* Give chance to dying process */ |
715a5ee82 memcg: fix oom sc... |
1869 |
schedule_timeout_uninterruptible(1); |
867578cbc memcg: fix oom ki... |
1870 |
return true; |
0b7f569e4 memcg: fix OOM ki... |
1871 |
} |
d69b042f3 memcg: add file-b... |
1872 1873 1874 |
/* * Currently used to update mapped file statistics, but the routine can be * generalized to update other statistics as well. |
32047e2a8 memcg: avoid lock... |
1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 |
* * 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... |
1894 |
*/ |
26174efd4 memcg: generic fi... |
1895 |
|
2a7106f2c memcg: create ext... |
1896 1897 |
void mem_cgroup_update_page_stat(struct page *page, enum mem_cgroup_page_stat_item idx, int val) |
d69b042f3 memcg: add file-b... |
1898 |
{ |
c0ff4b854 memcg: rename mem... |
1899 |
struct mem_cgroup *memcg; |
32047e2a8 memcg: avoid lock... |
1900 1901 |
struct page_cgroup *pc = lookup_page_cgroup(page); bool need_unlock = false; |
dbd4ea78f memcg: add lock t... |
1902 |
unsigned long uninitialized_var(flags); |
d69b042f3 memcg: add file-b... |
1903 |
|
d69b042f3 memcg: add file-b... |
1904 1905 |
if (unlikely(!pc)) return; |
32047e2a8 memcg: avoid lock... |
1906 |
rcu_read_lock(); |
c0ff4b854 memcg: rename mem... |
1907 1908 |
memcg = pc->mem_cgroup; if (unlikely(!memcg || !PageCgroupUsed(pc))) |
32047e2a8 memcg: avoid lock... |
1909 1910 |
goto out; /* pc->mem_cgroup is unstable ? */ |
c0ff4b854 memcg: rename mem... |
1911 |
if (unlikely(mem_cgroup_stealed(memcg)) || PageTransHuge(page)) { |
32047e2a8 memcg: avoid lock... |
1912 |
/* take a lock against to access pc->mem_cgroup */ |
dbd4ea78f memcg: add lock t... |
1913 |
move_lock_page_cgroup(pc, &flags); |
32047e2a8 memcg: avoid lock... |
1914 |
need_unlock = true; |
c0ff4b854 memcg: rename mem... |
1915 1916 |
memcg = pc->mem_cgroup; if (!memcg || !PageCgroupUsed(pc)) |
32047e2a8 memcg: avoid lock... |
1917 1918 |
goto out; } |
26174efd4 memcg: generic fi... |
1919 |
|
26174efd4 memcg: generic fi... |
1920 |
switch (idx) { |
2a7106f2c memcg: create ext... |
1921 |
case MEMCG_NR_FILE_MAPPED: |
26174efd4 memcg: generic fi... |
1922 1923 1924 |
if (val > 0) SetPageCgroupFileMapped(pc); else if (!page_mapped(page)) |
0c270f8f9 memcg: fix race i... |
1925 |
ClearPageCgroupFileMapped(pc); |
2a7106f2c memcg: create ext... |
1926 |
idx = MEM_CGROUP_STAT_FILE_MAPPED; |
26174efd4 memcg: generic fi... |
1927 1928 1929 |
break; default: BUG(); |
8725d5416 memcg: fix race i... |
1930 |
} |
d69b042f3 memcg: add file-b... |
1931 |
|
c0ff4b854 memcg: rename mem... |
1932 |
this_cpu_add(memcg->stat->count[idx], val); |
2a7106f2c memcg: create ext... |
1933 |
|
32047e2a8 memcg: avoid lock... |
1934 1935 |
out: if (unlikely(need_unlock)) |
dbd4ea78f memcg: add lock t... |
1936 |
move_unlock_page_cgroup(pc, &flags); |
32047e2a8 memcg: avoid lock... |
1937 1938 |
rcu_read_unlock(); return; |
d69b042f3 memcg: add file-b... |
1939 |
} |
2a7106f2c memcg: create ext... |
1940 |
EXPORT_SYMBOL(mem_cgroup_update_page_stat); |
26174efd4 memcg: generic fi... |
1941 |
|
f817ed485 memcg: move all a... |
1942 |
/* |
cdec2e426 memcg: coalesce c... |
1943 1944 1945 |
* 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... |
1946 |
#define CHARGE_BATCH 32U |
cdec2e426 memcg: coalesce c... |
1947 1948 |
struct memcg_stock_pcp { struct mem_cgroup *cached; /* this never be root cgroup */ |
11c9ea4e8 memcg: convert pe... |
1949 |
unsigned int nr_pages; |
cdec2e426 memcg: coalesce c... |
1950 |
struct work_struct work; |
26fe61684 memcg: fix percpu... |
1951 1952 |
unsigned long flags; #define FLUSHING_CACHED_CHARGE (0) |
cdec2e426 memcg: coalesce c... |
1953 1954 |
}; static DEFINE_PER_CPU(struct memcg_stock_pcp, memcg_stock); |
9f50fad65 Revert "memcg: ge... |
1955 |
static DEFINE_MUTEX(percpu_charge_mutex); |
cdec2e426 memcg: coalesce c... |
1956 1957 |
/* |
11c9ea4e8 memcg: convert pe... |
1958 |
* Try to consume stocked charge on this cpu. If success, one page is consumed |
cdec2e426 memcg: coalesce c... |
1959 1960 1961 1962 |
* 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... |
1963 |
static bool consume_stock(struct mem_cgroup *memcg) |
cdec2e426 memcg: coalesce c... |
1964 1965 1966 1967 1968 |
{ struct memcg_stock_pcp *stock; bool ret = true; stock = &get_cpu_var(memcg_stock); |
c0ff4b854 memcg: rename mem... |
1969 |
if (memcg == stock->cached && stock->nr_pages) |
11c9ea4e8 memcg: convert pe... |
1970 |
stock->nr_pages--; |
cdec2e426 memcg: coalesce c... |
1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 |
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... |
1983 1984 1985 1986 |
if (stock->nr_pages) { unsigned long bytes = stock->nr_pages * PAGE_SIZE; res_counter_uncharge(&old->res, bytes); |
cdec2e426 memcg: coalesce c... |
1987 |
if (do_swap_account) |
11c9ea4e8 memcg: convert pe... |
1988 1989 |
res_counter_uncharge(&old->memsw, bytes); stock->nr_pages = 0; |
cdec2e426 memcg: coalesce c... |
1990 1991 |
} stock->cached = NULL; |
cdec2e426 memcg: coalesce c... |
1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 |
} /* * 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... |
2002 |
clear_bit(FLUSHING_CACHED_CHARGE, &stock->flags); |
cdec2e426 memcg: coalesce c... |
2003 2004 2005 2006 |
} /* * Cache charges(val) which is from res_counter, to local per_cpu area. |
320cc51d9 mm: fix typo in r... |
2007 |
* This will be consumed by consume_stock() function, later. |
cdec2e426 memcg: coalesce c... |
2008 |
*/ |
c0ff4b854 memcg: rename mem... |
2009 |
static void refill_stock(struct mem_cgroup *memcg, unsigned int nr_pages) |
cdec2e426 memcg: coalesce c... |
2010 2011 |
{ struct memcg_stock_pcp *stock = &get_cpu_var(memcg_stock); |
c0ff4b854 memcg: rename mem... |
2012 |
if (stock->cached != memcg) { /* reset if necessary */ |
cdec2e426 memcg: coalesce c... |
2013 |
drain_stock(stock); |
c0ff4b854 memcg: rename mem... |
2014 |
stock->cached = memcg; |
cdec2e426 memcg: coalesce c... |
2015 |
} |
11c9ea4e8 memcg: convert pe... |
2016 |
stock->nr_pages += nr_pages; |
cdec2e426 memcg: coalesce c... |
2017 2018 2019 2020 |
put_cpu_var(memcg_stock); } /* |
c0ff4b854 memcg: rename mem... |
2021 |
* Drains all per-CPU charge caches for given root_memcg resp. subtree |
d38144b7a memcg: unify sync... |
2022 2023 |
* of the hierarchy under it. sync flag says whether we should block * until the work is done. |
cdec2e426 memcg: coalesce c... |
2024 |
*/ |
c0ff4b854 memcg: rename mem... |
2025 |
static void drain_all_stock(struct mem_cgroup *root_memcg, bool sync) |
cdec2e426 memcg: coalesce c... |
2026 |
{ |
26fe61684 memcg: fix percpu... |
2027 |
int cpu, curcpu; |
d38144b7a memcg: unify sync... |
2028 |
|
cdec2e426 memcg: coalesce c... |
2029 |
/* Notify other cpus that system-wide "drain" is running */ |
cdec2e426 memcg: coalesce c... |
2030 |
get_online_cpus(); |
5af12d0ef memcg: pin execut... |
2031 |
curcpu = get_cpu(); |
cdec2e426 memcg: coalesce c... |
2032 2033 |
for_each_online_cpu(cpu) { struct memcg_stock_pcp *stock = &per_cpu(memcg_stock, cpu); |
c0ff4b854 memcg: rename mem... |
2034 |
struct mem_cgroup *memcg; |
26fe61684 memcg: fix percpu... |
2035 |
|
c0ff4b854 memcg: rename mem... |
2036 2037 |
memcg = stock->cached; if (!memcg || !stock->nr_pages) |
26fe61684 memcg: fix percpu... |
2038 |
continue; |
c0ff4b854 memcg: rename mem... |
2039 |
if (!mem_cgroup_same_or_subtree(root_memcg, memcg)) |
3e92041d6 memcg: add mem_cg... |
2040 |
continue; |
d1a05b697 memcg: do not try... |
2041 2042 2043 2044 2045 2046 |
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... |
2047 |
} |
5af12d0ef memcg: pin execut... |
2048 |
put_cpu(); |
d38144b7a memcg: unify sync... |
2049 2050 2051 2052 2053 2054 |
if (!sync) goto out; for_each_online_cpu(cpu) { struct memcg_stock_pcp *stock = &per_cpu(memcg_stock, cpu); |
9f50fad65 Revert "memcg: ge... |
2055 |
if (test_bit(FLUSHING_CACHED_CHARGE, &stock->flags)) |
d38144b7a memcg: unify sync... |
2056 2057 2058 |
flush_work(&stock->work); } out: |
cdec2e426 memcg: coalesce c... |
2059 |
put_online_cpus(); |
d38144b7a memcg: unify sync... |
2060 2061 2062 2063 2064 2065 2066 2067 |
} /* * 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... |
2068 |
static void drain_all_stock_async(struct mem_cgroup *root_memcg) |
d38144b7a memcg: unify sync... |
2069 |
{ |
9f50fad65 Revert "memcg: ge... |
2070 2071 2072 2073 2074 |
/* * If someone calls draining, avoid adding more kworker runs. */ if (!mutex_trylock(&percpu_charge_mutex)) return; |
c0ff4b854 memcg: rename mem... |
2075 |
drain_all_stock(root_memcg, false); |
9f50fad65 Revert "memcg: ge... |
2076 |
mutex_unlock(&percpu_charge_mutex); |
cdec2e426 memcg: coalesce c... |
2077 2078 2079 |
} /* This is a synchronous drain interface. */ |
c0ff4b854 memcg: rename mem... |
2080 |
static void drain_all_stock_sync(struct mem_cgroup *root_memcg) |
cdec2e426 memcg: coalesce c... |
2081 2082 |
{ /* called when force_empty is called */ |
9f50fad65 Revert "memcg: ge... |
2083 |
mutex_lock(&percpu_charge_mutex); |
c0ff4b854 memcg: rename mem... |
2084 |
drain_all_stock(root_memcg, true); |
9f50fad65 Revert "memcg: ge... |
2085 |
mutex_unlock(&percpu_charge_mutex); |
cdec2e426 memcg: coalesce c... |
2086 |
} |
711d3d2c9 memcg: cpu hotplu... |
2087 2088 2089 2090 |
/* * 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... |
2091 |
static void mem_cgroup_drain_pcp_counter(struct mem_cgroup *memcg, int cpu) |
711d3d2c9 memcg: cpu hotplu... |
2092 2093 |
{ int i; |
c0ff4b854 memcg: rename mem... |
2094 |
spin_lock(&memcg->pcp_counter_lock); |
711d3d2c9 memcg: cpu hotplu... |
2095 |
for (i = 0; i < MEM_CGROUP_STAT_DATA; i++) { |
c0ff4b854 memcg: rename mem... |
2096 |
long x = per_cpu(memcg->stat->count[i], cpu); |
711d3d2c9 memcg: cpu hotplu... |
2097 |
|
c0ff4b854 memcg: rename mem... |
2098 2099 |
per_cpu(memcg->stat->count[i], cpu) = 0; memcg->nocpu_base.count[i] += x; |
711d3d2c9 memcg: cpu hotplu... |
2100 |
} |
e9f8974f2 memcg: break out ... |
2101 |
for (i = 0; i < MEM_CGROUP_EVENTS_NSTATS; i++) { |
c0ff4b854 memcg: rename mem... |
2102 |
unsigned long x = per_cpu(memcg->stat->events[i], cpu); |
e9f8974f2 memcg: break out ... |
2103 |
|
c0ff4b854 memcg: rename mem... |
2104 2105 |
per_cpu(memcg->stat->events[i], cpu) = 0; memcg->nocpu_base.events[i] += x; |
e9f8974f2 memcg: break out ... |
2106 |
} |
1489ebad8 memcg: cpu hotplu... |
2107 |
/* need to clear ON_MOVE value, works as a kind of lock. */ |
c0ff4b854 memcg: rename mem... |
2108 2109 |
per_cpu(memcg->stat->count[MEM_CGROUP_ON_MOVE], cpu) = 0; spin_unlock(&memcg->pcp_counter_lock); |
1489ebad8 memcg: cpu hotplu... |
2110 |
} |
c0ff4b854 memcg: rename mem... |
2111 |
static void synchronize_mem_cgroup_on_move(struct mem_cgroup *memcg, int cpu) |
1489ebad8 memcg: cpu hotplu... |
2112 2113 |
{ int idx = MEM_CGROUP_ON_MOVE; |
c0ff4b854 memcg: rename mem... |
2114 2115 2116 |
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... |
2117 2118 2119 |
} static int __cpuinit memcg_cpu_hotplug_callback(struct notifier_block *nb, |
cdec2e426 memcg: coalesce c... |
2120 2121 2122 2123 2124 |
unsigned long action, void *hcpu) { int cpu = (unsigned long)hcpu; struct memcg_stock_pcp *stock; |
711d3d2c9 memcg: cpu hotplu... |
2125 |
struct mem_cgroup *iter; |
cdec2e426 memcg: coalesce c... |
2126 |
|
1489ebad8 memcg: cpu hotplu... |
2127 2128 2129 2130 2131 |
if ((action == CPU_ONLINE)) { for_each_mem_cgroup_all(iter) synchronize_mem_cgroup_on_move(iter, cpu); return NOTIFY_OK; } |
711d3d2c9 memcg: cpu hotplu... |
2132 |
if ((action != CPU_DEAD) || action != CPU_DEAD_FROZEN) |
cdec2e426 memcg: coalesce c... |
2133 |
return NOTIFY_OK; |
711d3d2c9 memcg: cpu hotplu... |
2134 2135 2136 |
for_each_mem_cgroup_all(iter) mem_cgroup_drain_pcp_counter(iter, cpu); |
cdec2e426 memcg: coalesce c... |
2137 2138 2139 2140 |
stock = &per_cpu(memcg_stock, cpu); drain_stock(stock); return NOTIFY_OK; } |
4b5343346 memcg: clean up t... |
2141 2142 2143 2144 2145 2146 2147 2148 2149 |
/* 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... |
2150 |
static int mem_cgroup_do_charge(struct mem_cgroup *memcg, gfp_t gfp_mask, |
7ec99d621 memcg: unify char... |
2151 |
unsigned int nr_pages, bool oom_check) |
4b5343346 memcg: clean up t... |
2152 |
{ |
7ec99d621 memcg: unify char... |
2153 |
unsigned long csize = nr_pages * PAGE_SIZE; |
4b5343346 memcg: clean up t... |
2154 2155 2156 2157 |
struct mem_cgroup *mem_over_limit; struct res_counter *fail_res; unsigned long flags = 0; int ret; |
c0ff4b854 memcg: rename mem... |
2158 |
ret = res_counter_charge(&memcg->res, csize, &fail_res); |
4b5343346 memcg: clean up t... |
2159 2160 2161 2162 |
if (likely(!ret)) { if (!do_swap_account) return CHARGE_OK; |
c0ff4b854 memcg: rename mem... |
2163 |
ret = res_counter_charge(&memcg->memsw, csize, &fail_res); |
4b5343346 memcg: clean up t... |
2164 2165 |
if (likely(!ret)) return CHARGE_OK; |
c0ff4b854 memcg: rename mem... |
2166 |
res_counter_uncharge(&memcg->res, csize); |
4b5343346 memcg: clean up t... |
2167 2168 2169 2170 |
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... |
2171 |
/* |
7ec99d621 memcg: unify char... |
2172 2173 |
* 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... |
2174 2175 2176 2177 |
* * Never reclaim on behalf of optional batching, retry with a * single page instead. */ |
7ec99d621 memcg: unify char... |
2178 |
if (nr_pages == CHARGE_BATCH) |
4b5343346 memcg: clean up t... |
2179 2180 2181 2182 2183 2184 |
return CHARGE_RETRY; if (!(gfp_mask & __GFP_WAIT)) return CHARGE_WOULDBLOCK; ret = mem_cgroup_hierarchical_reclaim(mem_over_limit, NULL, |
0ae5e89c6 memcg: count the ... |
2185 |
gfp_mask, flags, NULL); |
7ec99d621 memcg: unify char... |
2186 |
if (mem_cgroup_margin(mem_over_limit) >= nr_pages) |
19942822d memcg: prevent en... |
2187 |
return CHARGE_RETRY; |
4b5343346 memcg: clean up t... |
2188 |
/* |
19942822d memcg: prevent en... |
2189 2190 2191 2192 2193 2194 2195 |
* 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... |
2196 |
*/ |
7ec99d621 memcg: unify char... |
2197 |
if (nr_pages == 1 && ret) |
4b5343346 memcg: clean up t... |
2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 |
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... |
2216 |
/* |
f817ed485 memcg: move all a... |
2217 2218 |
* Unlike exported interface, "oom" parameter is added. if oom==true, * oom-killer can be invoked. |
8a9f3ccd2 Memory controller... |
2219 |
*/ |
f817ed485 memcg: move all a... |
2220 |
static int __mem_cgroup_try_charge(struct mm_struct *mm, |
ec1685109 thp: memcg compound |
2221 |
gfp_t gfp_mask, |
7ec99d621 memcg: unify char... |
2222 |
unsigned int nr_pages, |
c0ff4b854 memcg: rename mem... |
2223 |
struct mem_cgroup **ptr, |
7ec99d621 memcg: unify char... |
2224 |
bool oom) |
8a9f3ccd2 Memory controller... |
2225 |
{ |
7ec99d621 memcg: unify char... |
2226 |
unsigned int batch = max(CHARGE_BATCH, nr_pages); |
4b5343346 memcg: clean up t... |
2227 |
int nr_oom_retries = MEM_CGROUP_RECLAIM_RETRIES; |
c0ff4b854 memcg: rename mem... |
2228 |
struct mem_cgroup *memcg = NULL; |
4b5343346 memcg: clean up t... |
2229 |
int ret; |
a636b327f memcg: avoid unne... |
2230 |
|
867578cbc memcg: fix oom ki... |
2231 2232 2233 2234 2235 2236 2237 2238 |
/* * 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... |
2239 |
|
8a9f3ccd2 Memory controller... |
2240 |
/* |
3be91277e memcgroup: tidy u... |
2241 2242 |
* 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... |
2243 2244 2245 |
* 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... |
2246 |
if (!*ptr && !mm) |
f75ca9620 memcg: avoid css_... |
2247 2248 |
goto bypass; again: |
c0ff4b854 memcg: rename mem... |
2249 2250 2251 2252 |
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_... |
2253 |
goto done; |
c0ff4b854 memcg: rename mem... |
2254 |
if (nr_pages == 1 && consume_stock(memcg)) |
f75ca9620 memcg: avoid css_... |
2255 |
goto done; |
c0ff4b854 memcg: rename mem... |
2256 |
css_get(&memcg->css); |
4b5343346 memcg: clean up t... |
2257 |
} else { |
f75ca9620 memcg: avoid css_... |
2258 |
struct task_struct *p; |
54595fe26 memcg: use css_tr... |
2259 |
|
f75ca9620 memcg: avoid css_... |
2260 2261 |
rcu_read_lock(); p = rcu_dereference(mm->owner); |
f75ca9620 memcg: avoid css_... |
2262 |
/* |
ebb76ce16 memcg: fix wrong ... |
2263 |
* Because we don't have task_lock(), "p" can exit. |
c0ff4b854 memcg: rename mem... |
2264 |
* In that case, "memcg" can point to root or p can be NULL with |
ebb76ce16 memcg: fix wrong ... |
2265 2266 2267 2268 2269 2270 |
* 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_... |
2271 |
*/ |
c0ff4b854 memcg: rename mem... |
2272 2273 |
memcg = mem_cgroup_from_task(p); if (!memcg || mem_cgroup_is_root(memcg)) { |
f75ca9620 memcg: avoid css_... |
2274 2275 2276 |
rcu_read_unlock(); goto done; } |
c0ff4b854 memcg: rename mem... |
2277 |
if (nr_pages == 1 && consume_stock(memcg)) { |
f75ca9620 memcg: avoid css_... |
2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 |
/* * 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... |
2290 |
if (!css_tryget(&memcg->css)) { |
f75ca9620 memcg: avoid css_... |
2291 2292 2293 2294 2295 |
rcu_read_unlock(); goto again; } rcu_read_unlock(); } |
8a9f3ccd2 Memory controller... |
2296 |
|
4b5343346 memcg: clean up t... |
2297 2298 |
do { bool oom_check; |
7a81b88cb memcg: introduce ... |
2299 |
|
4b5343346 memcg: clean up t... |
2300 |
/* If killed, bypass charge */ |
f75ca9620 memcg: avoid css_... |
2301 |
if (fatal_signal_pending(current)) { |
c0ff4b854 memcg: rename mem... |
2302 |
css_put(&memcg->css); |
4b5343346 memcg: clean up t... |
2303 |
goto bypass; |
f75ca9620 memcg: avoid css_... |
2304 |
} |
6d61ef409 memcg: memory cgr... |
2305 |
|
4b5343346 memcg: clean up t... |
2306 2307 2308 2309 |
oom_check = false; if (oom && !nr_oom_retries) { oom_check = true; nr_oom_retries = MEM_CGROUP_RECLAIM_RETRIES; |
cdec2e426 memcg: coalesce c... |
2310 |
} |
66e1707bc Memory controller... |
2311 |
|
c0ff4b854 memcg: rename mem... |
2312 |
ret = mem_cgroup_do_charge(memcg, gfp_mask, batch, oom_check); |
4b5343346 memcg: clean up t... |
2313 2314 2315 2316 |
switch (ret) { case CHARGE_OK: break; case CHARGE_RETRY: /* not in OOM situation but retry */ |
7ec99d621 memcg: unify char... |
2317 |
batch = nr_pages; |
c0ff4b854 memcg: rename mem... |
2318 2319 |
css_put(&memcg->css); memcg = NULL; |
f75ca9620 memcg: avoid css_... |
2320 |
goto again; |
4b5343346 memcg: clean up t... |
2321 |
case CHARGE_WOULDBLOCK: /* !__GFP_WAIT */ |
c0ff4b854 memcg: rename mem... |
2322 |
css_put(&memcg->css); |
4b5343346 memcg: clean up t... |
2323 2324 |
goto nomem; case CHARGE_NOMEM: /* OOM routine works */ |
f75ca9620 memcg: avoid css_... |
2325 |
if (!oom) { |
c0ff4b854 memcg: rename mem... |
2326 |
css_put(&memcg->css); |
867578cbc memcg: fix oom ki... |
2327 |
goto nomem; |
f75ca9620 memcg: avoid css_... |
2328 |
} |
4b5343346 memcg: clean up t... |
2329 2330 2331 2332 |
/* If oom, we never return -ENOMEM */ nr_oom_retries--; break; case CHARGE_OOM_DIE: /* Killed by OOM Killer */ |
c0ff4b854 memcg: rename mem... |
2333 |
css_put(&memcg->css); |
867578cbc memcg: fix oom ki... |
2334 |
goto bypass; |
66e1707bc Memory controller... |
2335 |
} |
4b5343346 memcg: clean up t... |
2336 |
} while (ret != CHARGE_OK); |
7ec99d621 memcg: unify char... |
2337 |
if (batch > nr_pages) |
c0ff4b854 memcg: rename mem... |
2338 2339 |
refill_stock(memcg, batch - nr_pages); css_put(&memcg->css); |
0c3e73e84 memcg: improve re... |
2340 |
done: |
c0ff4b854 memcg: rename mem... |
2341 |
*ptr = memcg; |
7a81b88cb memcg: introduce ... |
2342 2343 |
return 0; nomem: |
c0ff4b854 memcg: rename mem... |
2344 |
*ptr = NULL; |
7a81b88cb memcg: introduce ... |
2345 |
return -ENOMEM; |
867578cbc memcg: fix oom ki... |
2346 |
bypass: |
c0ff4b854 memcg: rename mem... |
2347 |
*ptr = NULL; |
867578cbc memcg: fix oom ki... |
2348 |
return 0; |
7a81b88cb memcg: introduce ... |
2349 |
} |
8a9f3ccd2 Memory controller... |
2350 |
|
a3b2d6926 cgroups: use css ... |
2351 |
/* |
a3032a2c1 memcg: add mem_cg... |
2352 2353 2354 2355 |
* 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... |
2356 |
static void __mem_cgroup_cancel_charge(struct mem_cgroup *memcg, |
e7018b8d2 memcg: keep only ... |
2357 |
unsigned int nr_pages) |
a3032a2c1 memcg: add mem_cg... |
2358 |
{ |
c0ff4b854 memcg: rename mem... |
2359 |
if (!mem_cgroup_is_root(memcg)) { |
e7018b8d2 memcg: keep only ... |
2360 |
unsigned long bytes = nr_pages * PAGE_SIZE; |
c0ff4b854 memcg: rename mem... |
2361 |
res_counter_uncharge(&memcg->res, bytes); |
a3032a2c1 memcg: add mem_cg... |
2362 |
if (do_swap_account) |
c0ff4b854 memcg: rename mem... |
2363 |
res_counter_uncharge(&memcg->memsw, bytes); |
a3032a2c1 memcg: add mem_cg... |
2364 |
} |
854ffa8d1 memcg: improve pe... |
2365 |
} |
a3032a2c1 memcg: add mem_cg... |
2366 |
/* |
a3b2d6926 cgroups: use css ... |
2367 2368 2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 |
* 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... |
2384 |
struct mem_cgroup *try_get_mem_cgroup_from_page(struct page *page) |
b5a84319a memcg: fix shmem'... |
2385 |
{ |
c0ff4b854 memcg: rename mem... |
2386 |
struct mem_cgroup *memcg = NULL; |
3c776e646 memcg: charge swa... |
2387 |
struct page_cgroup *pc; |
a3b2d6926 cgroups: use css ... |
2388 |
unsigned short id; |
b5a84319a memcg: fix shmem'... |
2389 |
swp_entry_t ent; |
3c776e646 memcg: charge swa... |
2390 |
VM_BUG_ON(!PageLocked(page)); |
3c776e646 memcg: charge swa... |
2391 |
pc = lookup_page_cgroup(page); |
c0bd3f63c memcg: fix try_ge... |
2392 |
lock_page_cgroup(pc); |
a3b2d6926 cgroups: use css ... |
2393 |
if (PageCgroupUsed(pc)) { |
c0ff4b854 memcg: rename mem... |
2394 2395 2396 |
memcg = pc->mem_cgroup; if (memcg && !css_tryget(&memcg->css)) memcg = NULL; |
e42d9d5d4 memcg: rename and... |
2397 |
} else if (PageSwapCache(page)) { |
3c776e646 memcg: charge swa... |
2398 |
ent.val = page_private(page); |
a3b2d6926 cgroups: use css ... |
2399 2400 |
id = lookup_swap_cgroup(ent); rcu_read_lock(); |
c0ff4b854 memcg: rename mem... |
2401 2402 2403 |
memcg = mem_cgroup_lookup(id); if (memcg && !css_tryget(&memcg->css)) memcg = NULL; |
a3b2d6926 cgroups: use css ... |
2404 |
rcu_read_unlock(); |
3c776e646 memcg: charge swa... |
2405 |
} |
c0bd3f63c memcg: fix try_ge... |
2406 |
unlock_page_cgroup(pc); |
c0ff4b854 memcg: rename mem... |
2407 |
return memcg; |
b5a84319a memcg: fix shmem'... |
2408 |
} |
c0ff4b854 memcg: rename mem... |
2409 |
static void __mem_cgroup_commit_charge(struct mem_cgroup *memcg, |
5564e88ba memcg: condense p... |
2410 |
struct page *page, |
7ec99d621 memcg: unify char... |
2411 |
unsigned int nr_pages, |
ca3e02141 memcg: fix USED b... |
2412 |
struct page_cgroup *pc, |
7ec99d621 memcg: unify char... |
2413 |
enum charge_type ctype) |
7a81b88cb memcg: introduce ... |
2414 |
{ |
ca3e02141 memcg: fix USED b... |
2415 2416 2417 |
lock_page_cgroup(pc); if (unlikely(PageCgroupUsed(pc))) { unlock_page_cgroup(pc); |
c0ff4b854 memcg: rename mem... |
2418 |
__mem_cgroup_cancel_charge(memcg, nr_pages); |
ca3e02141 memcg: fix USED b... |
2419 2420 2421 2422 2423 2424 |
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... |
2425 |
pc->mem_cgroup = memcg; |
261fb61a8 memcg: add commen... |
2426 2427 2428 2429 2430 2431 2432 |
/* * 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... |
2433 |
smp_wmb(); |
4b3bde4c9 memcg: remove the... |
2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 |
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... |
2447 |
|
c0ff4b854 memcg: rename mem... |
2448 |
mem_cgroup_charge_statistics(memcg, PageCgroupCache(pc), nr_pages); |
52d4b9ac0 memcg: allocate a... |
2449 |
unlock_page_cgroup(pc); |
430e48631 memcg: update thr... |
2450 2451 2452 2453 2454 |
/* * "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... |
2455 |
memcg_check_events(memcg, page); |
7a81b88cb memcg: introduce ... |
2456 |
} |
66e1707bc Memory controller... |
2457 |
|
ca3e02141 memcg: fix USED b... |
2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 |
#ifdef CONFIG_TRANSPARENT_HUGEPAGE #define PCGF_NOCOPY_AT_SPLIT ((1 << PCG_LOCK) | (1 << PCG_MOVE_LOCK) |\ (1 << PCG_ACCT_LRU) | (1 << PCG_MIGRATION)) /* * Because tail pages are not marked as "used", set it. We're under * zone->lru_lock, 'splitting on pmd' and compund_lock. */ void mem_cgroup_split_huge_fixup(struct page *head, struct page *tail) { struct page_cgroup *head_pc = lookup_page_cgroup(head); struct page_cgroup *tail_pc = lookup_page_cgroup(tail); unsigned long flags; |
3d37c4a91 memcg: bugfix che... |
2471 2472 |
if (mem_cgroup_disabled()) return; |
ca3e02141 memcg: fix USED b... |
2473 |
/* |
ece35ca81 memcg: fix LRU ac... |
2474 |
* We have no races with charge/uncharge but will have races with |
ca3e02141 memcg: fix USED b... |
2475 2476 2477 2478 2479 2480 |
* page state accounting. */ move_lock_page_cgroup(head_pc, &flags); tail_pc->mem_cgroup = head_pc->mem_cgroup; smp_wmb(); /* see __commit_charge() */ |
ece35ca81 memcg: fix LRU ac... |
2481 2482 2483 2484 2485 2486 2487 2488 2489 2490 |
if (PageCgroupAcctLRU(head_pc)) { enum lru_list lru; struct mem_cgroup_per_zone *mz; /* * LRU flags cannot be copied because we need to add tail *.page to LRU by generic call and our hook will be called. * We hold lru_lock, then, reduce counter directly. */ lru = page_lru(head); |
97a6c37b3 memcg: change pag... |
2491 |
mz = page_cgroup_zoneinfo(head_pc->mem_cgroup, head); |
ece35ca81 memcg: fix LRU ac... |
2492 2493 |
MEM_CGROUP_ZSTAT(mz, lru) -= 1; } |
ca3e02141 memcg: fix USED b... |
2494 2495 2496 2497 |
tail_pc->flags = head_pc->flags & ~PCGF_NOCOPY_AT_SPLIT; move_unlock_page_cgroup(head_pc, &flags); } #endif |
f817ed485 memcg: move all a... |
2498 |
/** |
de3638d9c memcg: fold __mem... |
2499 |
* mem_cgroup_move_account - move account of the page |
5564e88ba memcg: condense p... |
2500 |
* @page: the page |
7ec99d621 memcg: unify char... |
2501 |
* @nr_pages: number of regular pages (>1 for huge pages) |
f817ed485 memcg: move all a... |
2502 2503 2504 |
* @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... |
2505 |
* @uncharge: whether we should call uncharge and css_put against @from. |
f817ed485 memcg: move all a... |
2506 2507 |
* * The caller must confirm following. |
08e552c69 memcg: synchroniz... |
2508 |
* - page is not on LRU (isolate_page() is useful.) |
7ec99d621 memcg: unify char... |
2509 |
* - compound_lock is held when nr_pages > 1 |
f817ed485 memcg: move all a... |
2510 |
* |
854ffa8d1 memcg: improve pe... |
2511 |
* This function doesn't do "charge" nor css_get to new cgroup. It should be |
25985edce Fix common misspe... |
2512 |
* done by a caller(__mem_cgroup_try_charge would be useful). If @uncharge is |
854ffa8d1 memcg: improve pe... |
2513 2514 |
* 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... |
2515 |
*/ |
7ec99d621 memcg: unify char... |
2516 2517 2518 2519 2520 2521 |
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... |
2522 |
{ |
de3638d9c memcg: fold __mem... |
2523 2524 |
unsigned long flags; int ret; |
987eba66e memcg: fix rmdir,... |
2525 |
|
f817ed485 memcg: move all a... |
2526 |
VM_BUG_ON(from == to); |
5564e88ba memcg: condense p... |
2527 |
VM_BUG_ON(PageLRU(page)); |
de3638d9c memcg: fold __mem... |
2528 2529 2530 2531 2532 2533 2534 |
/* * 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... |
2535 |
if (nr_pages > 1 && !PageTransHuge(page)) |
de3638d9c memcg: fold __mem... |
2536 2537 2538 2539 2540 2541 2542 2543 2544 |
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... |
2545 |
|
8725d5416 memcg: fix race i... |
2546 |
if (PageCgroupFileMapped(pc)) { |
c62b1a3b3 memcg: use generi... |
2547 2548 2549 2550 2551 |
/* 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... |
2552 |
} |
987eba66e memcg: fix rmdir,... |
2553 |
mem_cgroup_charge_statistics(from, PageCgroupCache(pc), -nr_pages); |
854ffa8d1 memcg: improve pe... |
2554 2555 |
if (uncharge) /* This is not "cancel", but cancel_charge does all we need. */ |
e7018b8d2 memcg: keep only ... |
2556 |
__mem_cgroup_cancel_charge(from, nr_pages); |
d69b042f3 memcg: add file-b... |
2557 |
|
854ffa8d1 memcg: improve pe... |
2558 |
/* caller should have done css_get */ |
08e552c69 memcg: synchroniz... |
2559 |
pc->mem_cgroup = to; |
987eba66e memcg: fix rmdir,... |
2560 |
mem_cgroup_charge_statistics(to, PageCgroupCache(pc), nr_pages); |
887032670 cgroup avoid perm... |
2561 2562 2563 |
/* * 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... |
2564 |
* this function is just force_empty() and move charge, so it's |
25985edce Fix common misspe... |
2565 |
* guaranteed that "to" is never removed. So, we don't check rmdir |
4ffef5fef memcg: move charg... |
2566 |
* status here. |
887032670 cgroup avoid perm... |
2567 |
*/ |
de3638d9c memcg: fold __mem... |
2568 2569 2570 |
move_unlock_page_cgroup(pc, &flags); ret = 0; unlock: |
57f9fd7d2 memcg: cleanup me... |
2571 |
unlock_page_cgroup(pc); |
d2265e6fa memcg : share eve... |
2572 2573 2574 |
/* * check events */ |
5564e88ba memcg: condense p... |
2575 2576 |
memcg_check_events(to, page); memcg_check_events(from, page); |
de3638d9c memcg: fold __mem... |
2577 |
out: |
f817ed485 memcg: move all a... |
2578 2579 2580 2581 2582 2583 |
return ret; } /* * move charges to its parent. */ |
5564e88ba memcg: condense p... |
2584 2585 |
static int mem_cgroup_move_parent(struct page *page, struct page_cgroup *pc, |
f817ed485 memcg: move all a... |
2586 2587 2588 2589 2590 2591 |
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... |
2592 |
unsigned int nr_pages; |
4be4489fe mm/memcontrol.c: ... |
2593 |
unsigned long uninitialized_var(flags); |
f817ed485 memcg: move all a... |
2594 2595 2596 2597 2598 |
int ret; /* Is ROOT ? */ if (!pcg) return -EINVAL; |
57f9fd7d2 memcg: cleanup me... |
2599 2600 2601 2602 2603 |
ret = -EBUSY; if (!get_page_unless_zero(page)) goto out; if (isolate_lru_page(page)) goto put; |
52dbb9050 memcg: fix race a... |
2604 |
|
7ec99d621 memcg: unify char... |
2605 |
nr_pages = hpage_nr_pages(page); |
08e552c69 memcg: synchroniz... |
2606 |
|
f817ed485 memcg: move all a... |
2607 |
parent = mem_cgroup_from_cont(pcg); |
7ec99d621 memcg: unify char... |
2608 |
ret = __mem_cgroup_try_charge(NULL, gfp_mask, nr_pages, &parent, false); |
a636b327f memcg: avoid unne... |
2609 |
if (ret || !parent) |
57f9fd7d2 memcg: cleanup me... |
2610 |
goto put_back; |
f817ed485 memcg: move all a... |
2611 |
|
7ec99d621 memcg: unify char... |
2612 |
if (nr_pages > 1) |
987eba66e memcg: fix rmdir,... |
2613 |
flags = compound_lock_irqsave(page); |
7ec99d621 memcg: unify char... |
2614 |
ret = mem_cgroup_move_account(page, nr_pages, pc, child, parent, true); |
854ffa8d1 memcg: improve pe... |
2615 |
if (ret) |
7ec99d621 memcg: unify char... |
2616 |
__mem_cgroup_cancel_charge(parent, nr_pages); |
8dba474f0 mm/memcontrol.c: ... |
2617 |
|
7ec99d621 memcg: unify char... |
2618 |
if (nr_pages > 1) |
987eba66e memcg: fix rmdir,... |
2619 |
compound_unlock_irqrestore(page, flags); |
8dba474f0 mm/memcontrol.c: ... |
2620 |
put_back: |
08e552c69 memcg: synchroniz... |
2621 |
putback_lru_page(page); |
57f9fd7d2 memcg: cleanup me... |
2622 |
put: |
40d58138f memcg: fix error ... |
2623 |
put_page(page); |
57f9fd7d2 memcg: cleanup me... |
2624 |
out: |
f817ed485 memcg: move all a... |
2625 2626 |
return ret; } |
7a81b88cb memcg: introduce ... |
2627 2628 2629 2630 2631 2632 2633 |
/* * 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... |
2634 |
gfp_t gfp_mask, enum charge_type ctype) |
7a81b88cb memcg: introduce ... |
2635 |
{ |
c0ff4b854 memcg: rename mem... |
2636 |
struct mem_cgroup *memcg = NULL; |
7ec99d621 memcg: unify char... |
2637 |
unsigned int nr_pages = 1; |
7a81b88cb memcg: introduce ... |
2638 |
struct page_cgroup *pc; |
8493ae439 memcg: never OOM ... |
2639 |
bool oom = true; |
7a81b88cb memcg: introduce ... |
2640 |
int ret; |
ec1685109 thp: memcg compound |
2641 |
|
37c2ac787 thp: compound_tra... |
2642 |
if (PageTransHuge(page)) { |
7ec99d621 memcg: unify char... |
2643 |
nr_pages <<= compound_order(page); |
37c2ac787 thp: compound_tra... |
2644 |
VM_BUG_ON(!PageTransHuge(page)); |
8493ae439 memcg: never OOM ... |
2645 2646 2647 2648 2649 |
/* * Never OOM-kill a process for a huge page. The * fault handler will fall back to regular pages. */ oom = false; |
37c2ac787 thp: compound_tra... |
2650 |
} |
7a81b88cb memcg: introduce ... |
2651 2652 |
pc = lookup_page_cgroup(page); |
af4a66214 memcg: remove NUL... |
2653 |
BUG_ON(!pc); /* XXX: remove this and move pc lookup into commit */ |
7a81b88cb memcg: introduce ... |
2654 |
|
c0ff4b854 memcg: rename mem... |
2655 2656 |
ret = __mem_cgroup_try_charge(mm, gfp_mask, nr_pages, &memcg, oom); if (ret || !memcg) |
7a81b88cb memcg: introduce ... |
2657 |
return ret; |
c0ff4b854 memcg: rename mem... |
2658 |
__mem_cgroup_commit_charge(memcg, page, nr_pages, pc, ctype); |
8a9f3ccd2 Memory controller... |
2659 |
return 0; |
8a9f3ccd2 Memory controller... |
2660 |
} |
7a81b88cb memcg: introduce ... |
2661 2662 |
int mem_cgroup_newpage_charge(struct page *page, struct mm_struct *mm, gfp_t gfp_mask) |
217bc3194 memory cgroup enh... |
2663 |
{ |
f8d665422 memcg: add mem_cg... |
2664 |
if (mem_cgroup_disabled()) |
cede86acd memcg: clean up c... |
2665 |
return 0; |
69029cd55 memcg: remove ref... |
2666 2667 2668 2669 2670 2671 2672 2673 2674 2675 2676 |
/* * If already mapped, we don't have to account. * If page cache, page->mapping has address_space. * But page->mapping may have out-of-use anon_vma pointer, * detecit it by PageAnon() check. newly-mapped-anon's page->mapping * is NULL. */ if (page_mapped(page) || (page->mapping && !PageAnon(page))) return 0; if (unlikely(!mm)) mm = &init_mm; |
217bc3194 memory cgroup enh... |
2677 |
return mem_cgroup_charge_common(page, mm, gfp_mask, |
73045c47b memcg: remove mem... |
2678 |
MEM_CGROUP_CHARGE_TYPE_MAPPED); |
217bc3194 memory cgroup enh... |
2679 |
} |
83aae4c73 memcg: cleanup ca... |
2680 2681 2682 |
static void __mem_cgroup_commit_charge_swapin(struct page *page, struct mem_cgroup *ptr, enum charge_type ctype); |
5a6475a4e memcg: fix leak o... |
2683 |
static void |
c0ff4b854 memcg: rename mem... |
2684 |
__mem_cgroup_commit_charge_lrucare(struct page *page, struct mem_cgroup *memcg, |
5a6475a4e memcg: fix leak o... |
2685 2686 2687 2688 2689 2690 2691 2692 2693 |
enum charge_type ctype) { struct page_cgroup *pc = lookup_page_cgroup(page); /* * 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. */ mem_cgroup_lru_del_before_commit(page); |
c0ff4b854 memcg: rename mem... |
2694 |
__mem_cgroup_commit_charge(memcg, page, 1, pc, ctype); |
5a6475a4e memcg: fix leak o... |
2695 2696 2697 |
mem_cgroup_lru_add_after_commit(page); return; } |
e1a1cd590 Memory controller... |
2698 2699 |
int mem_cgroup_cache_charge(struct page *page, struct mm_struct *mm, gfp_t gfp_mask) |
8697d3319 Memory controller... |
2700 |
{ |
c0ff4b854 memcg: rename mem... |
2701 |
struct mem_cgroup *memcg = NULL; |
b5a84319a memcg: fix shmem'... |
2702 |
int ret; |
f8d665422 memcg: add mem_cg... |
2703 |
if (mem_cgroup_disabled()) |
cede86acd memcg: clean up c... |
2704 |
return 0; |
52d4b9ac0 memcg: allocate a... |
2705 2706 |
if (PageCompound(page)) return 0; |
accf163e6 memcg: remove a r... |
2707 |
|
73045c47b memcg: remove mem... |
2708 |
if (unlikely(!mm)) |
8697d3319 Memory controller... |
2709 |
mm = &init_mm; |
accf163e6 memcg: remove a r... |
2710 |
|
5a6475a4e memcg: fix leak o... |
2711 |
if (page_is_file_cache(page)) { |
c0ff4b854 memcg: rename mem... |
2712 2713 |
ret = __mem_cgroup_try_charge(mm, gfp_mask, 1, &memcg, true); if (ret || !memcg) |
5a6475a4e memcg: fix leak o... |
2714 |
return ret; |
b5a84319a memcg: fix shmem'... |
2715 |
|
5a6475a4e memcg: fix leak o... |
2716 2717 2718 2719 2720 |
/* * FUSE reuses pages without going through the final * put that would remove them from the LRU list, make * sure that they get relinked properly. */ |
c0ff4b854 memcg: rename mem... |
2721 |
__mem_cgroup_commit_charge_lrucare(page, memcg, |
5a6475a4e memcg: fix leak o... |
2722 2723 2724 |
MEM_CGROUP_CHARGE_TYPE_CACHE); return ret; } |
83aae4c73 memcg: cleanup ca... |
2725 2726 |
/* shmem */ if (PageSwapCache(page)) { |
c0ff4b854 memcg: rename mem... |
2727 |
ret = mem_cgroup_try_charge_swapin(mm, page, gfp_mask, &memcg); |
83aae4c73 memcg: cleanup ca... |
2728 |
if (!ret) |
c0ff4b854 memcg: rename mem... |
2729 |
__mem_cgroup_commit_charge_swapin(page, memcg, |
83aae4c73 memcg: cleanup ca... |
2730 2731 2732 |
MEM_CGROUP_CHARGE_TYPE_SHMEM); } else ret = mem_cgroup_charge_common(page, mm, gfp_mask, |
73045c47b memcg: remove mem... |
2733 |
MEM_CGROUP_CHARGE_TYPE_SHMEM); |
b5a84319a memcg: fix shmem'... |
2734 |
|
b5a84319a memcg: fix shmem'... |
2735 |
return ret; |
e8589cc18 memcg: better mig... |
2736 |
} |
54595fe26 memcg: use css_tr... |
2737 2738 2739 |
/* * 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... |
2740 |
* struct page_cgroup is acquired. This refcnt will be consumed by |
54595fe26 memcg: use css_tr... |
2741 2742 |
* "commit()" or removed by "cancel()" */ |
8c7c6e34a memcg: mem+swap c... |
2743 2744 2745 2746 |
int mem_cgroup_try_charge_swapin(struct mm_struct *mm, struct page *page, gfp_t mask, struct mem_cgroup **ptr) { |
c0ff4b854 memcg: rename mem... |
2747 |
struct mem_cgroup *memcg; |
54595fe26 memcg: use css_tr... |
2748 |
int ret; |
8c7c6e34a memcg: mem+swap c... |
2749 |
|
56039efa1 memcg: fix ugly i... |
2750 |
*ptr = NULL; |
f8d665422 memcg: add mem_cg... |
2751 |
if (mem_cgroup_disabled()) |
8c7c6e34a memcg: mem+swap c... |
2752 2753 2754 2755 |
return 0; if (!do_swap_account) goto charge_cur_mm; |
8c7c6e34a memcg: mem+swap c... |
2756 2757 |
/* * A racing thread's fault, or swapoff, may have already updated |
407f9c8b0 ksm: mem cgroup c... |
2758 2759 2760 |
* 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... |
2761 2762 |
*/ if (!PageSwapCache(page)) |
407f9c8b0 ksm: mem cgroup c... |
2763 |
goto charge_cur_mm; |
c0ff4b854 memcg: rename mem... |
2764 2765 |
memcg = try_get_mem_cgroup_from_page(page); if (!memcg) |
54595fe26 memcg: use css_tr... |
2766 |
goto charge_cur_mm; |
c0ff4b854 memcg: rename mem... |
2767 |
*ptr = memcg; |
7ec99d621 memcg: unify char... |
2768 |
ret = __mem_cgroup_try_charge(NULL, mask, 1, ptr, true); |
c0ff4b854 memcg: rename mem... |
2769 |
css_put(&memcg->css); |
54595fe26 memcg: use css_tr... |
2770 |
return ret; |
8c7c6e34a memcg: mem+swap c... |
2771 2772 2773 |
charge_cur_mm: if (unlikely(!mm)) mm = &init_mm; |
7ec99d621 memcg: unify char... |
2774 |
return __mem_cgroup_try_charge(mm, mask, 1, ptr, true); |
8c7c6e34a memcg: mem+swap c... |
2775 |
} |
83aae4c73 memcg: cleanup ca... |
2776 2777 2778 |
static void __mem_cgroup_commit_charge_swapin(struct page *page, struct mem_cgroup *ptr, enum charge_type ctype) |
7a81b88cb memcg: introduce ... |
2779 |
{ |
f8d665422 memcg: add mem_cg... |
2780 |
if (mem_cgroup_disabled()) |
7a81b88cb memcg: introduce ... |
2781 2782 2783 |
return; if (!ptr) return; |
887032670 cgroup avoid perm... |
2784 |
cgroup_exclude_rmdir(&ptr->css); |
5a6475a4e memcg: fix leak o... |
2785 2786 |
__mem_cgroup_commit_charge_lrucare(page, ptr, ctype); |
8c7c6e34a memcg: mem+swap c... |
2787 2788 2789 |
/* * Now swap is on-memory. This means this page may be * counted both as mem and swap....double count. |
03f3c4336 memcg: fix swap a... |
2790 2791 2792 |
* 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... |
2793 |
*/ |
03f3c4336 memcg: fix swap a... |
2794 |
if (do_swap_account && PageSwapCache(page)) { |
8c7c6e34a memcg: mem+swap c... |
2795 |
swp_entry_t ent = {.val = page_private(page)}; |
a3b2d6926 cgroups: use css ... |
2796 |
unsigned short id; |
8c7c6e34a memcg: mem+swap c... |
2797 |
struct mem_cgroup *memcg; |
a3b2d6926 cgroups: use css ... |
2798 2799 2800 2801 |
id = swap_cgroup_record(ent, 0); rcu_read_lock(); memcg = mem_cgroup_lookup(id); |
8c7c6e34a memcg: mem+swap c... |
2802 |
if (memcg) { |
a3b2d6926 cgroups: use css ... |
2803 2804 2805 2806 |
/* * This recorded memcg can be obsolete one. So, avoid * calling css_tryget */ |
0c3e73e84 memcg: improve re... |
2807 |
if (!mem_cgroup_is_root(memcg)) |
4e649152c memcg: some modif... |
2808 |
res_counter_uncharge(&memcg->memsw, PAGE_SIZE); |
0c3e73e84 memcg: improve re... |
2809 |
mem_cgroup_swap_statistics(memcg, false); |
8c7c6e34a memcg: mem+swap c... |
2810 2811 |
mem_cgroup_put(memcg); } |
a3b2d6926 cgroups: use css ... |
2812 |
rcu_read_unlock(); |
8c7c6e34a memcg: mem+swap c... |
2813 |
} |
887032670 cgroup avoid perm... |
2814 2815 2816 2817 2818 2819 |
/* * 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. */ cgroup_release_and_wakeup_rmdir(&ptr->css); |
7a81b88cb memcg: introduce ... |
2820 |
} |
83aae4c73 memcg: cleanup ca... |
2821 2822 2823 2824 2825 |
void mem_cgroup_commit_charge_swapin(struct page *page, struct mem_cgroup *ptr) { __mem_cgroup_commit_charge_swapin(page, ptr, MEM_CGROUP_CHARGE_TYPE_MAPPED); } |
c0ff4b854 memcg: rename mem... |
2826 |
void mem_cgroup_cancel_charge_swapin(struct mem_cgroup *memcg) |
7a81b88cb memcg: introduce ... |
2827 |
{ |
f8d665422 memcg: add mem_cg... |
2828 |
if (mem_cgroup_disabled()) |
7a81b88cb memcg: introduce ... |
2829 |
return; |
c0ff4b854 memcg: rename mem... |
2830 |
if (!memcg) |
7a81b88cb memcg: introduce ... |
2831 |
return; |
c0ff4b854 memcg: rename mem... |
2832 |
__mem_cgroup_cancel_charge(memcg, 1); |
7a81b88cb memcg: introduce ... |
2833 |
} |
c0ff4b854 memcg: rename mem... |
2834 |
static void mem_cgroup_do_uncharge(struct mem_cgroup *memcg, |
7ec99d621 memcg: unify char... |
2835 2836 |
unsigned int nr_pages, const enum charge_type ctype) |
569b846df memcg: coalesce u... |
2837 2838 2839 |
{ struct memcg_batch_info *batch = NULL; bool uncharge_memsw = true; |
7ec99d621 memcg: unify char... |
2840 |
|
569b846df memcg: coalesce u... |
2841 2842 2843 |
/* 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... |
2844 2845 2846 2847 2848 2849 2850 2851 |
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... |
2852 |
batch->memcg = memcg; |
569b846df memcg: coalesce u... |
2853 |
/* |
3c11ecf44 memcg: oom kill d... |
2854 |
* do_batch > 0 when unmapping pages or inode invalidate/truncate. |
25985edce Fix common misspe... |
2855 |
* In those cases, all pages freed continuously can be expected to be in |
3c11ecf44 memcg: oom kill d... |
2856 2857 2858 2859 2860 2861 2862 |
* 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... |
2863 |
if (nr_pages > 1) |
ec1685109 thp: memcg compound |
2864 |
goto direct_uncharge; |
3c11ecf44 memcg: oom kill d... |
2865 |
/* |
569b846df memcg: coalesce u... |
2866 2867 2868 2869 |
* 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... |
2870 |
if (batch->memcg != memcg) |
569b846df memcg: coalesce u... |
2871 2872 |
goto direct_uncharge; /* remember freed charge and uncharge it later */ |
7ffd4ca7a memcg: convert un... |
2873 |
batch->nr_pages++; |
569b846df memcg: coalesce u... |
2874 |
if (uncharge_memsw) |
7ffd4ca7a memcg: convert un... |
2875 |
batch->memsw_nr_pages++; |
569b846df memcg: coalesce u... |
2876 2877 |
return; direct_uncharge: |
c0ff4b854 memcg: rename mem... |
2878 |
res_counter_uncharge(&memcg->res, nr_pages * PAGE_SIZE); |
569b846df memcg: coalesce u... |
2879 |
if (uncharge_memsw) |
c0ff4b854 memcg: rename mem... |
2880 2881 2882 |
res_counter_uncharge(&memcg->memsw, nr_pages * PAGE_SIZE); if (unlikely(batch->memcg != memcg)) memcg_oom_recover(memcg); |
569b846df memcg: coalesce u... |
2883 2884 |
return; } |
7a81b88cb memcg: introduce ... |
2885 |
|
8697d3319 Memory controller... |
2886 |
/* |
69029cd55 memcg: remove ref... |
2887 |
* uncharge if !page_mapped(page) |
8a9f3ccd2 Memory controller... |
2888 |
*/ |
8c7c6e34a memcg: mem+swap c... |
2889 |
static struct mem_cgroup * |
69029cd55 memcg: remove ref... |
2890 |
__mem_cgroup_uncharge_common(struct page *page, enum charge_type ctype) |
8a9f3ccd2 Memory controller... |
2891 |
{ |
c0ff4b854 memcg: rename mem... |
2892 |
struct mem_cgroup *memcg = NULL; |
7ec99d621 memcg: unify char... |
2893 2894 |
unsigned int nr_pages = 1; struct page_cgroup *pc; |
8a9f3ccd2 Memory controller... |
2895 |
|
f8d665422 memcg: add mem_cg... |
2896 |
if (mem_cgroup_disabled()) |
8c7c6e34a memcg: mem+swap c... |
2897 |
return NULL; |
4077960e2 memory controller... |
2898 |
|
d13d14430 memcg: handle swa... |
2899 |
if (PageSwapCache(page)) |
8c7c6e34a memcg: mem+swap c... |
2900 |
return NULL; |
d13d14430 memcg: handle swa... |
2901 |
|
37c2ac787 thp: compound_tra... |
2902 |
if (PageTransHuge(page)) { |
7ec99d621 memcg: unify char... |
2903 |
nr_pages <<= compound_order(page); |
37c2ac787 thp: compound_tra... |
2904 2905 |
VM_BUG_ON(!PageTransHuge(page)); } |
8697d3319 Memory controller... |
2906 |
/* |
3c541e14b Memory controller... |
2907 |
* Check if our page_cgroup is valid |
8697d3319 Memory controller... |
2908 |
*/ |
52d4b9ac0 memcg: allocate a... |
2909 2910 |
pc = lookup_page_cgroup(page); if (unlikely(!pc || !PageCgroupUsed(pc))) |
8c7c6e34a memcg: mem+swap c... |
2911 |
return NULL; |
b9c565d5a memcg: remove cle... |
2912 |
|
52d4b9ac0 memcg: allocate a... |
2913 |
lock_page_cgroup(pc); |
d13d14430 memcg: handle swa... |
2914 |
|
c0ff4b854 memcg: rename mem... |
2915 |
memcg = pc->mem_cgroup; |
8c7c6e34a memcg: mem+swap c... |
2916 |
|
d13d14430 memcg: handle swa... |
2917 2918 2919 2920 2921 |
if (!PageCgroupUsed(pc)) goto unlock_out; switch (ctype) { case MEM_CGROUP_CHARGE_TYPE_MAPPED: |
8a9478ca7 memcg: fix swap a... |
2922 |
case MEM_CGROUP_CHARGE_TYPE_DROP: |
ac39cf8cb memcg: fix mis-ac... |
2923 2924 |
/* See mem_cgroup_prepare_migration() */ if (page_mapped(page) || PageCgroupMigration(pc)) |
d13d14430 memcg: handle swa... |
2925 2926 2927 2928 2929 2930 2931 2932 2933 2934 2935 |
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... |
2936 |
} |
d13d14430 memcg: handle swa... |
2937 |
|
c0ff4b854 memcg: rename mem... |
2938 |
mem_cgroup_charge_statistics(memcg, PageCgroupCache(pc), -nr_pages); |
04046e1a0 memcg: use CSS ID |
2939 |
|
52d4b9ac0 memcg: allocate a... |
2940 |
ClearPageCgroupUsed(pc); |
544122e5e memcg: fix LRU ac... |
2941 2942 2943 2944 2945 2946 |
/* * 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... |
2947 |
|
52d4b9ac0 memcg: allocate a... |
2948 |
unlock_page_cgroup(pc); |
f75ca9620 memcg: avoid css_... |
2949 |
/* |
c0ff4b854 memcg: rename mem... |
2950 |
* even after unlock, we have memcg->res.usage here and this memcg |
f75ca9620 memcg: avoid css_... |
2951 2952 |
* will never be freed. */ |
c0ff4b854 memcg: rename mem... |
2953 |
memcg_check_events(memcg, page); |
f75ca9620 memcg: avoid css_... |
2954 |
if (do_swap_account && ctype == MEM_CGROUP_CHARGE_TYPE_SWAPOUT) { |
c0ff4b854 memcg: rename mem... |
2955 2956 |
mem_cgroup_swap_statistics(memcg, true); mem_cgroup_get(memcg); |
f75ca9620 memcg: avoid css_... |
2957 |
} |
c0ff4b854 memcg: rename mem... |
2958 2959 |
if (!mem_cgroup_is_root(memcg)) mem_cgroup_do_uncharge(memcg, nr_pages, ctype); |
6d12e2d8d per-zone and recl... |
2960 |
|
c0ff4b854 memcg: rename mem... |
2961 |
return memcg; |
d13d14430 memcg: handle swa... |
2962 2963 2964 |
unlock_out: unlock_page_cgroup(pc); |
8c7c6e34a memcg: mem+swap c... |
2965 |
return NULL; |
3c541e14b Memory controller... |
2966 |
} |
69029cd55 memcg: remove ref... |
2967 2968 |
void mem_cgroup_uncharge_page(struct page *page) { |
52d4b9ac0 memcg: allocate a... |
2969 2970 2971 2972 2973 |
/* early check. */ if (page_mapped(page)) return; if (page->mapping && !PageAnon(page)) return; |
69029cd55 memcg: remove ref... |
2974 2975 2976 2977 2978 2979 |
__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-... |
2980 |
VM_BUG_ON(page->mapping); |
69029cd55 memcg: remove ref... |
2981 2982 |
__mem_cgroup_uncharge_common(page, MEM_CGROUP_CHARGE_TYPE_CACHE); } |
569b846df memcg: coalesce u... |
2983 2984 2985 2986 2987 2988 2989 2990 2991 2992 2993 2994 2995 2996 |
/* * 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... |
2997 2998 |
current->memcg_batch.nr_pages = 0; current->memcg_batch.memsw_nr_pages = 0; |
569b846df memcg: coalesce u... |
2999 3000 3001 3002 3003 3004 3005 3006 3007 3008 3009 3010 3011 3012 3013 3014 3015 3016 3017 3018 |
} } 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... |
3019 3020 3021 3022 3023 3024 |
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... |
3025 |
memcg_oom_recover(batch->memcg); |
569b846df memcg: coalesce u... |
3026 3027 3028 |
/* forget this pointer (for sanity check) */ batch->memcg = NULL; } |
e767e0561 memcg: fix deadlo... |
3029 |
#ifdef CONFIG_SWAP |
8c7c6e34a memcg: mem+swap c... |
3030 |
/* |
e767e0561 memcg: fix deadlo... |
3031 |
* called after __delete_from_swap_cache() and drop "page" account. |
8c7c6e34a memcg: mem+swap c... |
3032 3033 |
* memcg information is recorded to swap_cgroup of "ent" */ |
8a9478ca7 memcg: fix swap a... |
3034 3035 |
void mem_cgroup_uncharge_swapcache(struct page *page, swp_entry_t ent, bool swapout) |
8c7c6e34a memcg: mem+swap c... |
3036 3037 |
{ struct mem_cgroup *memcg; |
8a9478ca7 memcg: fix swap a... |
3038 3039 3040 3041 3042 3043 |
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... |
3044 |
|
f75ca9620 memcg: avoid css_... |
3045 3046 3047 3048 3049 |
/* * record memcg information, if swapout && memcg != NULL, * mem_cgroup_get() was called in uncharge(). */ if (do_swap_account && swapout && memcg) |
a3b2d6926 cgroups: use css ... |
3050 |
swap_cgroup_record(ent, css_id(&memcg->css)); |
8c7c6e34a memcg: mem+swap c... |
3051 |
} |
e767e0561 memcg: fix deadlo... |
3052 |
#endif |
8c7c6e34a memcg: mem+swap c... |
3053 3054 3055 3056 3057 3058 3059 |
#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... |
3060 |
{ |
8c7c6e34a memcg: mem+swap c... |
3061 |
struct mem_cgroup *memcg; |
a3b2d6926 cgroups: use css ... |
3062 |
unsigned short id; |
8c7c6e34a memcg: mem+swap c... |
3063 3064 3065 |
if (!do_swap_account) return; |
a3b2d6926 cgroups: use css ... |
3066 3067 3068 |
id = swap_cgroup_record(ent, 0); rcu_read_lock(); memcg = mem_cgroup_lookup(id); |
8c7c6e34a memcg: mem+swap c... |
3069 |
if (memcg) { |
a3b2d6926 cgroups: use css ... |
3070 3071 3072 3073 |
/* * We uncharge this because swap is freed. * This memcg can be obsolete one. We avoid calling css_tryget */ |
0c3e73e84 memcg: improve re... |
3074 |
if (!mem_cgroup_is_root(memcg)) |
4e649152c memcg: some modif... |
3075 |
res_counter_uncharge(&memcg->memsw, PAGE_SIZE); |
0c3e73e84 memcg: improve re... |
3076 |
mem_cgroup_swap_statistics(memcg, false); |
8c7c6e34a memcg: mem+swap c... |
3077 3078 |
mem_cgroup_put(memcg); } |
a3b2d6926 cgroups: use css ... |
3079 |
rcu_read_unlock(); |
d13d14430 memcg: handle swa... |
3080 |
} |
024914477 memcg: move charg... |
3081 3082 3083 3084 3085 3086 |
/** * 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... |
3087 |
* @need_fixup: whether we should fixup res_counters and refcounts. |
024914477 memcg: move charg... |
3088 3089 3090 3091 3092 3093 3094 3095 3096 3097 |
* * 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... |
3098 |
struct mem_cgroup *from, struct mem_cgroup *to, bool need_fixup) |
024914477 memcg: move charg... |
3099 3100 3101 3102 3103 3104 3105 |
{ 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... |
3106 |
mem_cgroup_swap_statistics(from, false); |
483c30b51 memcg: improve pe... |
3107 |
mem_cgroup_swap_statistics(to, true); |
024914477 memcg: move charg... |
3108 |
/* |
483c30b51 memcg: improve pe... |
3109 3110 3111 3112 3113 3114 |
* 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... |
3115 |
*/ |
024914477 memcg: move charg... |
3116 |
mem_cgroup_get(to); |
483c30b51 memcg: improve pe... |
3117 3118 3119 3120 3121 3122 3123 3124 3125 3126 |
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... |
3127 |
} |
024914477 memcg: move charg... |
3128 3129 3130 3131 3132 3133 |
return 0; } return -EINVAL; } #else static inline int mem_cgroup_move_swap_account(swp_entry_t entry, |
483c30b51 memcg: improve pe... |
3134 |
struct mem_cgroup *from, struct mem_cgroup *to, bool need_fixup) |
024914477 memcg: move charg... |
3135 3136 3137 |
{ return -EINVAL; } |
8c7c6e34a memcg: mem+swap c... |
3138 |
#endif |
d13d14430 memcg: handle swa... |
3139 |
|
ae41be374 bugfix for memory... |
3140 |
/* |
01b1ae63c memcg: simple mig... |
3141 3142 |
* Before starting migration, account PAGE_SIZE to mem_cgroup that the old * page belongs to. |
ae41be374 bugfix for memory... |
3143 |
*/ |
ac39cf8cb memcg: fix mis-ac... |
3144 |
int mem_cgroup_prepare_migration(struct page *page, |
ef6a3c631 mm: add replace_p... |
3145 |
struct page *newpage, struct mem_cgroup **ptr, gfp_t gfp_mask) |
ae41be374 bugfix for memory... |
3146 |
{ |
c0ff4b854 memcg: rename mem... |
3147 |
struct mem_cgroup *memcg = NULL; |
7ec99d621 memcg: unify char... |
3148 |
struct page_cgroup *pc; |
ac39cf8cb memcg: fix mis-ac... |
3149 |
enum charge_type ctype; |
e8589cc18 memcg: better mig... |
3150 |
int ret = 0; |
8869b8f6e memcg: memcontrol... |
3151 |
|
56039efa1 memcg: fix ugly i... |
3152 |
*ptr = NULL; |
ec1685109 thp: memcg compound |
3153 |
VM_BUG_ON(PageTransHuge(page)); |
f8d665422 memcg: add mem_cg... |
3154 |
if (mem_cgroup_disabled()) |
4077960e2 memory controller... |
3155 |
return 0; |
52d4b9ac0 memcg: allocate a... |
3156 3157 3158 |
pc = lookup_page_cgroup(page); lock_page_cgroup(pc); if (PageCgroupUsed(pc)) { |
c0ff4b854 memcg: rename mem... |
3159 3160 |
memcg = pc->mem_cgroup; css_get(&memcg->css); |
ac39cf8cb memcg: fix mis-ac... |
3161 3162 3163 3164 3165 3166 3167 3168 3169 3170 3171 3172 3173 3174 3175 3176 3177 3178 3179 3180 3181 3182 3183 3184 3185 3186 3187 3188 3189 3190 3191 |
/* * 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... |
3192 |
} |
52d4b9ac0 memcg: allocate a... |
3193 |
unlock_page_cgroup(pc); |
ac39cf8cb memcg: fix mis-ac... |
3194 3195 3196 3197 |
/* * If the page is not charged at this point, * we return here. */ |
c0ff4b854 memcg: rename mem... |
3198 |
if (!memcg) |
ac39cf8cb memcg: fix mis-ac... |
3199 |
return 0; |
01b1ae63c memcg: simple mig... |
3200 |
|
c0ff4b854 memcg: rename mem... |
3201 |
*ptr = memcg; |
7ec99d621 memcg: unify char... |
3202 |
ret = __mem_cgroup_try_charge(NULL, gfp_mask, 1, ptr, false); |
c0ff4b854 memcg: rename mem... |
3203 |
css_put(&memcg->css);/* drop extra refcnt */ |
ac39cf8cb memcg: fix mis-ac... |
3204 3205 3206 3207 3208 3209 3210 3211 3212 3213 3214 |
if (ret || *ptr == NULL) { 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); } return -ENOMEM; |
e8589cc18 memcg: better mig... |
3215 |
} |
ac39cf8cb memcg: fix mis-ac... |
3216 3217 3218 3219 3220 3221 3222 3223 3224 3225 3226 3227 3228 |
/* * 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... |
3229 |
__mem_cgroup_commit_charge(memcg, page, 1, pc, ctype); |
e8589cc18 memcg: better mig... |
3230 |
return ret; |
ae41be374 bugfix for memory... |
3231 |
} |
8869b8f6e memcg: memcontrol... |
3232 |
|
69029cd55 memcg: remove ref... |
3233 |
/* remove redundant charge if migration failed*/ |
c0ff4b854 memcg: rename mem... |
3234 |
void mem_cgroup_end_migration(struct mem_cgroup *memcg, |
50de1dd96 memcg: fix memory... |
3235 |
struct page *oldpage, struct page *newpage, bool migration_ok) |
ae41be374 bugfix for memory... |
3236 |
{ |
ac39cf8cb memcg: fix mis-ac... |
3237 |
struct page *used, *unused; |
01b1ae63c memcg: simple mig... |
3238 |
struct page_cgroup *pc; |
01b1ae63c memcg: simple mig... |
3239 |
|
c0ff4b854 memcg: rename mem... |
3240 |
if (!memcg) |
01b1ae63c memcg: simple mig... |
3241 |
return; |
ac39cf8cb memcg: fix mis-ac... |
3242 |
/* blocks rmdir() */ |
c0ff4b854 memcg: rename mem... |
3243 |
cgroup_exclude_rmdir(&memcg->css); |
50de1dd96 memcg: fix memory... |
3244 |
if (!migration_ok) { |
ac39cf8cb memcg: fix mis-ac... |
3245 3246 |
used = oldpage; unused = newpage; |
01b1ae63c memcg: simple mig... |
3247 |
} else { |
ac39cf8cb memcg: fix mis-ac... |
3248 |
used = newpage; |
01b1ae63c memcg: simple mig... |
3249 3250 |
unused = oldpage; } |
69029cd55 memcg: remove ref... |
3251 |
/* |
ac39cf8cb memcg: fix mis-ac... |
3252 3253 3254 |
* 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... |
3255 |
*/ |
ac39cf8cb memcg: fix mis-ac... |
3256 3257 3258 3259 |
pc = lookup_page_cgroup(oldpage); lock_page_cgroup(pc); ClearPageCgroupMigration(pc); unlock_page_cgroup(pc); |
01b1ae63c memcg: simple mig... |
3260 |
|
ac39cf8cb memcg: fix mis-ac... |
3261 |
__mem_cgroup_uncharge_common(unused, MEM_CGROUP_CHARGE_TYPE_FORCE); |
01b1ae63c memcg: simple mig... |
3262 |
/* |
ac39cf8cb memcg: fix mis-ac... |
3263 3264 3265 3266 3267 3268 |
* 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... |
3269 |
*/ |
ac39cf8cb memcg: fix mis-ac... |
3270 3271 |
if (PageAnon(used)) mem_cgroup_uncharge_page(used); |
887032670 cgroup avoid perm... |
3272 |
/* |
ac39cf8cb memcg: fix mis-ac... |
3273 3274 |
* At migration, we may charge account against cgroup which has no * tasks. |
887032670 cgroup avoid perm... |
3275 3276 3277 |
* 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... |
3278 |
cgroup_release_and_wakeup_rmdir(&memcg->css); |
ae41be374 bugfix for memory... |
3279 |
} |
78fb74669 Memory controller... |
3280 |
|
f212ad7cf memcg: add memcg ... |
3281 3282 3283 3284 3285 3286 3287 3288 3289 3290 3291 3292 3293 3294 3295 3296 3297 3298 3299 3300 3301 3302 3303 3304 3305 3306 3307 3308 3309 3310 3311 3312 3313 3314 3315 3316 3317 3318 3319 3320 3321 3322 3323 3324 3325 3326 |
#ifdef CONFIG_DEBUG_VM static struct page_cgroup *lookup_page_cgroup_used(struct page *page) { struct page_cgroup *pc; pc = lookup_page_cgroup(page); 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) { int ret = -1; char *path; printk(KERN_ALERT "pc:%p pc->flags:%lx pc->mem_cgroup:%p", pc, pc->flags, pc->mem_cgroup); path = kmalloc(PATH_MAX, GFP_KERNEL); if (path) { rcu_read_lock(); ret = cgroup_path(pc->mem_cgroup->css.cgroup, path, PATH_MAX); rcu_read_unlock(); } printk(KERN_CONT "(%s) ", (ret < 0) ? "cannot get the path" : path); kfree(path); } } #endif |
8c7c6e34a memcg: mem+swap c... |
3327 |
static DEFINE_MUTEX(set_limit_mutex); |
d38d2a758 mm: make mem_cgro... |
3328 |
static int mem_cgroup_resize_limit(struct mem_cgroup *memcg, |
8c7c6e34a memcg: mem+swap c... |
3329 |
unsigned long long val) |
628f42355 memcg: limit chan... |
3330 |
{ |
81d39c20f memcg: fix shrink... |
3331 |
int retry_count; |
3c11ecf44 memcg: oom kill d... |
3332 |
u64 memswlimit, memlimit; |
628f42355 memcg: limit chan... |
3333 |
int ret = 0; |
81d39c20f memcg: fix shrink... |
3334 3335 |
int children = mem_cgroup_count_children(memcg); u64 curusage, oldusage; |
3c11ecf44 memcg: oom kill d... |
3336 |
int enlarge; |
81d39c20f memcg: fix shrink... |
3337 3338 3339 3340 3341 3342 3343 3344 3345 |
/* * 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... |
3346 |
|
3c11ecf44 memcg: oom kill d... |
3347 |
enlarge = 0; |
8c7c6e34a memcg: mem+swap c... |
3348 |
while (retry_count) { |
628f42355 memcg: limit chan... |
3349 3350 3351 3352 |
if (signal_pending(current)) { ret = -EINTR; break; } |
8c7c6e34a memcg: mem+swap c... |
3353 3354 3355 |
/* * Rather than hide all in some function, I do this in * open coded manner. You see what this really does. |
c0ff4b854 memcg: rename mem... |
3356 |
* We have to guarantee memcg->res.limit < memcg->memsw.limit. |
8c7c6e34a memcg: mem+swap c... |
3357 3358 3359 3360 3361 3362 |
*/ 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... |
3363 3364 |
break; } |
3c11ecf44 memcg: oom kill d... |
3365 3366 3367 3368 |
memlimit = res_counter_read_u64(&memcg->res, RES_LIMIT); if (memlimit < val) enlarge = 1; |
8c7c6e34a memcg: mem+swap c... |
3369 |
ret = res_counter_set_limit(&memcg->res, val); |
22a668d7c memcg: fix behavi... |
3370 3371 3372 3373 3374 3375 |
if (!ret) { if (memswlimit == val) memcg->memsw_is_minimum = true; else memcg->memsw_is_minimum = false; } |
8c7c6e34a memcg: mem+swap c... |
3376 3377 3378 3379 |
mutex_unlock(&set_limit_mutex); if (!ret) break; |
aa20d489c memcg: code clean... |
3380 |
mem_cgroup_hierarchical_reclaim(memcg, NULL, GFP_KERNEL, |
0ae5e89c6 memcg: count the ... |
3381 3382 |
MEM_CGROUP_RECLAIM_SHRINK, NULL); |
81d39c20f memcg: fix shrink... |
3383 3384 3385 3386 3387 3388 |
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... |
3389 |
} |
3c11ecf44 memcg: oom kill d... |
3390 3391 |
if (!ret && enlarge) memcg_oom_recover(memcg); |
14797e236 memcg: add inacti... |
3392 |
|
8c7c6e34a memcg: mem+swap c... |
3393 3394 |
return ret; } |
338c84310 memcg: remove som... |
3395 3396 |
static int mem_cgroup_resize_memsw_limit(struct mem_cgroup *memcg, unsigned long long val) |
8c7c6e34a memcg: mem+swap c... |
3397 |
{ |
81d39c20f memcg: fix shrink... |
3398 |
int retry_count; |
3c11ecf44 memcg: oom kill d... |
3399 |
u64 memlimit, memswlimit, oldusage, curusage; |
81d39c20f memcg: fix shrink... |
3400 3401 |
int children = mem_cgroup_count_children(memcg); int ret = -EBUSY; |
3c11ecf44 memcg: oom kill d... |
3402 |
int enlarge = 0; |
8c7c6e34a memcg: mem+swap c... |
3403 |
|
81d39c20f memcg: fix shrink... |
3404 3405 3406 |
/* 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... |
3407 3408 3409 3410 3411 3412 3413 3414 |
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... |
3415 |
* We have to guarantee memcg->res.limit < memcg->memsw.limit. |
8c7c6e34a memcg: mem+swap c... |
3416 3417 3418 3419 3420 3421 3422 3423 |
*/ 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... |
3424 3425 3426 |
memswlimit = res_counter_read_u64(&memcg->memsw, RES_LIMIT); if (memswlimit < val) enlarge = 1; |
8c7c6e34a memcg: mem+swap c... |
3427 |
ret = res_counter_set_limit(&memcg->memsw, val); |
22a668d7c memcg: fix behavi... |
3428 3429 3430 3431 3432 3433 |
if (!ret) { if (memlimit == val) memcg->memsw_is_minimum = true; else memcg->memsw_is_minimum = false; } |
8c7c6e34a memcg: mem+swap c... |
3434 3435 3436 3437 |
mutex_unlock(&set_limit_mutex); if (!ret) break; |
4e4169535 memory controller... |
3438 |
mem_cgroup_hierarchical_reclaim(memcg, NULL, GFP_KERNEL, |
75822b449 memory controller... |
3439 |
MEM_CGROUP_RECLAIM_NOSWAP | |
0ae5e89c6 memcg: count the ... |
3440 3441 |
MEM_CGROUP_RECLAIM_SHRINK, NULL); |
8c7c6e34a memcg: mem+swap c... |
3442 |
curusage = res_counter_read_u64(&memcg->memsw, RES_USAGE); |
81d39c20f memcg: fix shrink... |
3443 |
/* Usage is reduced ? */ |
8c7c6e34a memcg: mem+swap c... |
3444 |
if (curusage >= oldusage) |
628f42355 memcg: limit chan... |
3445 |
retry_count--; |
81d39c20f memcg: fix shrink... |
3446 3447 |
else oldusage = curusage; |
628f42355 memcg: limit chan... |
3448 |
} |
3c11ecf44 memcg: oom kill d... |
3449 3450 |
if (!ret && enlarge) memcg_oom_recover(memcg); |
628f42355 memcg: limit chan... |
3451 3452 |
return ret; } |
4e4169535 memory controller... |
3453 |
unsigned long mem_cgroup_soft_limit_reclaim(struct zone *zone, int order, |
0ae5e89c6 memcg: count the ... |
3454 3455 |
gfp_t gfp_mask, unsigned long *total_scanned) |
4e4169535 memory controller... |
3456 3457 3458 3459 3460 3461 |
{ 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... |
3462 |
unsigned long long excess; |
0ae5e89c6 memcg: count the ... |
3463 |
unsigned long nr_scanned; |
4e4169535 memory controller... |
3464 3465 3466 |
if (order > 0) return 0; |
00918b6ab memcg: remove nid... |
3467 |
mctz = soft_limit_tree_node_zone(zone_to_nid(zone), zone_idx(zone)); |
4e4169535 memory controller... |
3468 3469 3470 3471 3472 3473 3474 3475 3476 3477 3478 3479 |
/* * 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 ... |
3480 |
nr_scanned = 0; |
4e4169535 memory controller... |
3481 3482 |
reclaimed = mem_cgroup_hierarchical_reclaim(mz->mem, zone, gfp_mask, |
0ae5e89c6 memcg: count the ... |
3483 3484 |
MEM_CGROUP_RECLAIM_SOFT, &nr_scanned); |
4e4169535 memory controller... |
3485 |
nr_reclaimed += reclaimed; |
0ae5e89c6 memcg: count the ... |
3486 |
*total_scanned += nr_scanned; |
4e4169535 memory controller... |
3487 3488 3489 3490 3491 3492 3493 3494 3495 3496 3497 3498 3499 3500 3501 3502 3503 3504 3505 3506 3507 3508 |
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... |
3509 |
if (next_mz == mz) |
4e4169535 memory controller... |
3510 |
css_put(&next_mz->mem->css); |
39cc98f1f memcg: remove poi... |
3511 |
else /* next_mz == NULL or other memcg */ |
4e4169535 memory controller... |
3512 3513 3514 |
break; } while (1); } |
4e4169535 memory controller... |
3515 |
__mem_cgroup_remove_exceeded(mz->mem, mz, mctz); |
ef8745c1e memcg: reduce che... |
3516 |
excess = res_counter_soft_limit_excess(&mz->mem->res); |
4e4169535 memory controller... |
3517 3518 3519 3520 3521 3522 3523 3524 |
/* * 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... |
3525 3526 |
/* If excess == 0, no tree ops */ __mem_cgroup_insert_exceeded(mz->mem, mz, mctz, excess); |
4e4169535 memory controller... |
3527 3528 3529 3530 3531 3532 3533 3534 3535 3536 3537 3538 3539 3540 3541 3542 3543 |
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... |
3544 |
/* |
cc8475822 memory cgroup enh... |
3545 |
* This routine traverse page_cgroup in given list and drop them all. |
cc8475822 memory cgroup enh... |
3546 3547 |
* *And* this routine doesn't reclaim page itself, just removes page_cgroup. */ |
c0ff4b854 memcg: rename mem... |
3548 |
static int mem_cgroup_force_empty_list(struct mem_cgroup *memcg, |
08e552c69 memcg: synchroniz... |
3549 |
int node, int zid, enum lru_list lru) |
cc8475822 memory cgroup enh... |
3550 |
{ |
08e552c69 memcg: synchroniz... |
3551 3552 |
struct zone *zone; struct mem_cgroup_per_zone *mz; |
f817ed485 memcg: move all a... |
3553 |
struct page_cgroup *pc, *busy; |
08e552c69 memcg: synchroniz... |
3554 |
unsigned long flags, loop; |
072c56c13 per-zone and recl... |
3555 |
struct list_head *list; |
f817ed485 memcg: move all a... |
3556 |
int ret = 0; |
072c56c13 per-zone and recl... |
3557 |
|
08e552c69 memcg: synchroniz... |
3558 |
zone = &NODE_DATA(node)->node_zones[zid]; |
c0ff4b854 memcg: rename mem... |
3559 |
mz = mem_cgroup_zoneinfo(memcg, node, zid); |
b69408e88 vmscan: Use an in... |
3560 |
list = &mz->lists[lru]; |
cc8475822 memory cgroup enh... |
3561 |
|
f817ed485 memcg: move all a... |
3562 3563 3564 3565 3566 |
loop = MEM_CGROUP_ZSTAT(mz, lru); /* give some margin against EBUSY etc...*/ loop += 256; busy = NULL; while (loop--) { |
5564e88ba memcg: condense p... |
3567 |
struct page *page; |
f817ed485 memcg: move all a... |
3568 |
ret = 0; |
08e552c69 memcg: synchroniz... |
3569 |
spin_lock_irqsave(&zone->lru_lock, flags); |
f817ed485 memcg: move all a... |
3570 |
if (list_empty(list)) { |
08e552c69 memcg: synchroniz... |
3571 |
spin_unlock_irqrestore(&zone->lru_lock, flags); |
52d4b9ac0 memcg: allocate a... |
3572 |
break; |
f817ed485 memcg: move all a... |
3573 3574 3575 3576 |
} pc = list_entry(list->prev, struct page_cgroup, lru); if (busy == pc) { list_move(&pc->lru, list); |
648bcc771 mm/memcontrol.c: ... |
3577 |
busy = NULL; |
08e552c69 memcg: synchroniz... |
3578 |
spin_unlock_irqrestore(&zone->lru_lock, flags); |
f817ed485 memcg: move all a... |
3579 3580 |
continue; } |
08e552c69 memcg: synchroniz... |
3581 |
spin_unlock_irqrestore(&zone->lru_lock, flags); |
f817ed485 memcg: move all a... |
3582 |
|
6b3ae58ef memcg: remove dir... |
3583 |
page = lookup_cgroup_page(pc); |
5564e88ba memcg: condense p... |
3584 |
|
c0ff4b854 memcg: rename mem... |
3585 |
ret = mem_cgroup_move_parent(page, pc, memcg, GFP_KERNEL); |
f817ed485 memcg: move all a... |
3586 |
if (ret == -ENOMEM) |
52d4b9ac0 memcg: allocate a... |
3587 |
break; |
f817ed485 memcg: move all a... |
3588 3589 3590 3591 3592 3593 3594 |
if (ret == -EBUSY || ret == -EINVAL) { /* found lock contention or "pc" is obsolete. */ busy = pc; cond_resched(); } else busy = NULL; |
cc8475822 memory cgroup enh... |
3595 |
} |
08e552c69 memcg: synchroniz... |
3596 |
|
f817ed485 memcg: move all a... |
3597 3598 3599 |
if (!ret && !list_empty(list)) return -EBUSY; return ret; |
cc8475822 memory cgroup enh... |
3600 3601 3602 3603 3604 3605 |
} /* * make mem_cgroup's charge to be 0 if there is no task. * This enables deleting this mem_cgroup. */ |
c0ff4b854 memcg: rename mem... |
3606 |
static int mem_cgroup_force_empty(struct mem_cgroup *memcg, bool free_all) |
cc8475822 memory cgroup enh... |
3607 |
{ |
f817ed485 memcg: move all a... |
3608 3609 3610 |
int ret; int node, zid, shrink; int nr_retries = MEM_CGROUP_RECLAIM_RETRIES; |
c0ff4b854 memcg: rename mem... |
3611 |
struct cgroup *cgrp = memcg->css.cgroup; |
8869b8f6e memcg: memcontrol... |
3612 |
|
c0ff4b854 memcg: rename mem... |
3613 |
css_get(&memcg->css); |
f817ed485 memcg: move all a... |
3614 3615 |
shrink = 0; |
c1e862c1f memcg: new force_... |
3616 3617 3618 |
/* should free all ? */ if (free_all) goto try_to_free; |
f817ed485 memcg: move all a... |
3619 |
move_account: |
fce664775 memcg: ensure lis... |
3620 |
do { |
f817ed485 memcg: move all a... |
3621 |
ret = -EBUSY; |
c1e862c1f memcg: new force_... |
3622 3623 3624 3625 |
if (cgroup_task_count(cgrp) || !list_empty(&cgrp->children)) goto out; ret = -EINTR; if (signal_pending(current)) |
cc8475822 memory cgroup enh... |
3626 |
goto out; |
52d4b9ac0 memcg: allocate a... |
3627 3628 |
/* This is for making all *used* pages to be on LRU. */ lru_add_drain_all(); |
c0ff4b854 memcg: rename mem... |
3629 |
drain_all_stock_sync(memcg); |
f817ed485 memcg: move all a... |
3630 |
ret = 0; |
c0ff4b854 memcg: rename mem... |
3631 |
mem_cgroup_start_move(memcg); |
299b4eaa3 memcg: NULL point... |
3632 |
for_each_node_state(node, N_HIGH_MEMORY) { |
f817ed485 memcg: move all a... |
3633 |
for (zid = 0; !ret && zid < MAX_NR_ZONES; zid++) { |
b69408e88 vmscan: Use an in... |
3634 |
enum lru_list l; |
f817ed485 memcg: move all a... |
3635 |
for_each_lru(l) { |
c0ff4b854 memcg: rename mem... |
3636 |
ret = mem_cgroup_force_empty_list(memcg, |
08e552c69 memcg: synchroniz... |
3637 |
node, zid, l); |
f817ed485 memcg: move all a... |
3638 3639 3640 |
if (ret) break; } |
1ecaab2bd per-zone and recl... |
3641 |
} |
f817ed485 memcg: move all a... |
3642 3643 3644 |
if (ret) break; } |
c0ff4b854 memcg: rename mem... |
3645 3646 |
mem_cgroup_end_move(memcg); memcg_oom_recover(memcg); |
f817ed485 memcg: move all a... |
3647 3648 3649 |
/* it seems parent cgroup doesn't have enough mem */ if (ret == -ENOMEM) goto try_to_free; |
52d4b9ac0 memcg: allocate a... |
3650 |
cond_resched(); |
fce664775 memcg: ensure lis... |
3651 |
/* "ret" should also be checked to ensure all lists are empty. */ |
c0ff4b854 memcg: rename mem... |
3652 |
} while (memcg->res.usage > 0 || ret); |
cc8475822 memory cgroup enh... |
3653 |
out: |
c0ff4b854 memcg: rename mem... |
3654 |
css_put(&memcg->css); |
cc8475822 memory cgroup enh... |
3655 |
return ret; |
f817ed485 memcg: move all a... |
3656 3657 |
try_to_free: |
c1e862c1f memcg: new force_... |
3658 3659 |
/* 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... |
3660 3661 3662 |
ret = -EBUSY; goto out; } |
c1e862c1f memcg: new force_... |
3663 3664 |
/* we call try-to-free pages for make this cgroup empty */ lru_add_drain_all(); |
f817ed485 memcg: move all a... |
3665 3666 |
/* try to free all pages in this cgroup */ shrink = 1; |
c0ff4b854 memcg: rename mem... |
3667 |
while (nr_retries && memcg->res.usage > 0) { |
f817ed485 memcg: move all a... |
3668 |
int progress; |
c1e862c1f memcg: new force_... |
3669 3670 3671 3672 3673 |
if (signal_pending(current)) { ret = -EINTR; goto out; } |
c0ff4b854 memcg: rename mem... |
3674 |
progress = try_to_free_mem_cgroup_pages(memcg, GFP_KERNEL, |
185efc0f9 memcg: Revert "me... |
3675 |
false); |
c1e862c1f memcg: new force_... |
3676 |
if (!progress) { |
f817ed485 memcg: move all a... |
3677 |
nr_retries--; |
c1e862c1f memcg: new force_... |
3678 |
/* maybe some writeback is necessary */ |
8aa7e847d Fix congestion_wa... |
3679 |
congestion_wait(BLK_RW_ASYNC, HZ/10); |
c1e862c1f memcg: new force_... |
3680 |
} |
f817ed485 memcg: move all a... |
3681 3682 |
} |
08e552c69 memcg: synchroniz... |
3683 |
lru_add_drain(); |
f817ed485 memcg: move all a... |
3684 |
/* try move_account...there may be some *locked* pages. */ |
fce664775 memcg: ensure lis... |
3685 |
goto move_account; |
cc8475822 memory cgroup enh... |
3686 |
} |
c1e862c1f memcg: new force_... |
3687 3688 3689 3690 |
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... |
3691 3692 3693 3694 3695 3696 3697 3698 3699 |
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... |
3700 |
struct mem_cgroup *memcg = mem_cgroup_from_cont(cont); |
18f59ea7d memcg: memory cgr... |
3701 |
struct cgroup *parent = cont->parent; |
c0ff4b854 memcg: rename mem... |
3702 |
struct mem_cgroup *parent_memcg = NULL; |
18f59ea7d memcg: memory cgr... |
3703 3704 |
if (parent) |
c0ff4b854 memcg: rename mem... |
3705 |
parent_memcg = mem_cgroup_from_cont(parent); |
18f59ea7d memcg: memory cgr... |
3706 3707 3708 |
cgroup_lock(); /* |
af901ca18 tree-wide: fix as... |
3709 |
* If parent's use_hierarchy is set, we can't make any modifications |
18f59ea7d memcg: memory cgr... |
3710 3711 3712 3713 3714 3715 |
* 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... |
3716 |
if ((!parent_memcg || !parent_memcg->use_hierarchy) && |
18f59ea7d memcg: memory cgr... |
3717 3718 |
(val == 1 || val == 0)) { if (list_empty(&cont->children)) |
c0ff4b854 memcg: rename mem... |
3719 |
memcg->use_hierarchy = val; |
18f59ea7d memcg: memory cgr... |
3720 3721 3722 3723 3724 3725 3726 3727 |
else retval = -EBUSY; } else retval = -EINVAL; cgroup_unlock(); return retval; } |
0c3e73e84 memcg: improve re... |
3728 |
|
c0ff4b854 memcg: rename mem... |
3729 |
static unsigned long mem_cgroup_recursive_stat(struct mem_cgroup *memcg, |
7a159cc9d memcg: use native... |
3730 |
enum mem_cgroup_stat_index idx) |
0c3e73e84 memcg: improve re... |
3731 |
{ |
7d74b06f2 memcg: use for_ea... |
3732 |
struct mem_cgroup *iter; |
7a159cc9d memcg: use native... |
3733 |
long val = 0; |
0c3e73e84 memcg: improve re... |
3734 |
|
7a159cc9d memcg: use native... |
3735 |
/* Per-cpu values can be negative, use a signed accumulator */ |
c0ff4b854 memcg: rename mem... |
3736 |
for_each_mem_cgroup_tree(iter, memcg) |
7d74b06f2 memcg: use for_ea... |
3737 3738 3739 3740 3741 |
val += mem_cgroup_read_stat(iter, idx); if (val < 0) /* race ? */ val = 0; return val; |
0c3e73e84 memcg: improve re... |
3742 |
} |
c0ff4b854 memcg: rename mem... |
3743 |
static inline u64 mem_cgroup_usage(struct mem_cgroup *memcg, bool swap) |
104f39284 memcg: extract me... |
3744 |
{ |
7d74b06f2 memcg: use for_ea... |
3745 |
u64 val; |
104f39284 memcg: extract me... |
3746 |
|
c0ff4b854 memcg: rename mem... |
3747 |
if (!mem_cgroup_is_root(memcg)) { |
104f39284 memcg: extract me... |
3748 |
if (!swap) |
65c64ce8e Partial revert "B... |
3749 |
return res_counter_read_u64(&memcg->res, RES_USAGE); |
104f39284 memcg: extract me... |
3750 |
else |
65c64ce8e Partial revert "B... |
3751 |
return res_counter_read_u64(&memcg->memsw, RES_USAGE); |
104f39284 memcg: extract me... |
3752 |
} |
c0ff4b854 memcg: rename mem... |
3753 3754 |
val = mem_cgroup_recursive_stat(memcg, MEM_CGROUP_STAT_CACHE); val += mem_cgroup_recursive_stat(memcg, MEM_CGROUP_STAT_RSS); |
104f39284 memcg: extract me... |
3755 |
|
7d74b06f2 memcg: use for_ea... |
3756 |
if (swap) |
c0ff4b854 memcg: rename mem... |
3757 |
val += mem_cgroup_recursive_stat(memcg, MEM_CGROUP_STAT_SWAPOUT); |
104f39284 memcg: extract me... |
3758 3759 3760 |
return val << PAGE_SHIFT; } |
2c3daa722 CGroup API files:... |
3761 |
static u64 mem_cgroup_read(struct cgroup *cont, struct cftype *cft) |
8cdea7c05 Memory controller... |
3762 |
{ |
c0ff4b854 memcg: rename mem... |
3763 |
struct mem_cgroup *memcg = mem_cgroup_from_cont(cont); |
104f39284 memcg: extract me... |
3764 |
u64 val; |
8c7c6e34a memcg: mem+swap c... |
3765 3766 3767 3768 3769 3770 |
int type, name; type = MEMFILE_TYPE(cft->private); name = MEMFILE_ATTR(cft->private); switch (type) { case _MEM: |
104f39284 memcg: extract me... |
3771 |
if (name == RES_USAGE) |
c0ff4b854 memcg: rename mem... |
3772 |
val = mem_cgroup_usage(memcg, false); |
104f39284 memcg: extract me... |
3773 |
else |
c0ff4b854 memcg: rename mem... |
3774 |
val = res_counter_read_u64(&memcg->res, name); |
8c7c6e34a memcg: mem+swap c... |
3775 3776 |
break; case _MEMSWAP: |
104f39284 memcg: extract me... |
3777 |
if (name == RES_USAGE) |
c0ff4b854 memcg: rename mem... |
3778 |
val = mem_cgroup_usage(memcg, true); |
104f39284 memcg: extract me... |
3779 |
else |
c0ff4b854 memcg: rename mem... |
3780 |
val = res_counter_read_u64(&memcg->memsw, name); |
8c7c6e34a memcg: mem+swap c... |
3781 3782 3783 3784 3785 3786 |
break; default: BUG(); break; } return val; |
8cdea7c05 Memory controller... |
3787 |
} |
628f42355 memcg: limit chan... |
3788 3789 3790 3791 |
/* * The user of this function is... * RES_LIMIT. */ |
856c13aa1 cgroup files: con... |
3792 3793 |
static int mem_cgroup_write(struct cgroup *cont, struct cftype *cft, const char *buffer) |
8cdea7c05 Memory controller... |
3794 |
{ |
628f42355 memcg: limit chan... |
3795 |
struct mem_cgroup *memcg = mem_cgroup_from_cont(cont); |
8c7c6e34a memcg: mem+swap c... |
3796 |
int type, name; |
628f42355 memcg: limit chan... |
3797 3798 |
unsigned long long val; int ret; |
8c7c6e34a memcg: mem+swap c... |
3799 3800 3801 |
type = MEMFILE_TYPE(cft->private); name = MEMFILE_ATTR(cft->private); switch (name) { |
628f42355 memcg: limit chan... |
3802 |
case RES_LIMIT: |
4b3bde4c9 memcg: remove the... |
3803 3804 3805 3806 |
if (mem_cgroup_is_root(memcg)) { /* Can't set limit on root */ ret = -EINVAL; break; } |
628f42355 memcg: limit chan... |
3807 3808 |
/* This function does all necessary parse...reuse it */ ret = res_counter_memparse_write_strategy(buffer, &val); |
8c7c6e34a memcg: mem+swap c... |
3809 3810 3811 |
if (ret) break; if (type == _MEM) |
628f42355 memcg: limit chan... |
3812 |
ret = mem_cgroup_resize_limit(memcg, val); |
8c7c6e34a memcg: mem+swap c... |
3813 3814 |
else ret = mem_cgroup_resize_memsw_limit(memcg, val); |
628f42355 memcg: limit chan... |
3815 |
break; |
296c81d89 memory controller... |
3816 3817 3818 3819 3820 3821 3822 3823 3824 3825 3826 3827 3828 3829 |
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... |
3830 3831 3832 3833 3834 |
default: ret = -EINVAL; /* should be BUG() ? */ break; } return ret; |
8cdea7c05 Memory controller... |
3835 |
} |
fee7b548e memcg: show real ... |
3836 3837 3838 3839 3840 3841 3842 3843 3844 3845 3846 3847 3848 3849 3850 3851 3852 3853 3854 3855 3856 3857 3858 3859 3860 3861 3862 |
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... |
3863 |
static int mem_cgroup_reset(struct cgroup *cont, unsigned int event) |
c84872e16 memcgroup: add th... |
3864 |
{ |
c0ff4b854 memcg: rename mem... |
3865 |
struct mem_cgroup *memcg; |
8c7c6e34a memcg: mem+swap c... |
3866 |
int type, name; |
c84872e16 memcgroup: add th... |
3867 |
|
c0ff4b854 memcg: rename mem... |
3868 |
memcg = mem_cgroup_from_cont(cont); |
8c7c6e34a memcg: mem+swap c... |
3869 3870 3871 |
type = MEMFILE_TYPE(event); name = MEMFILE_ATTR(event); switch (name) { |
29f2a4dac memcgroup: implem... |
3872 |
case RES_MAX_USAGE: |
8c7c6e34a memcg: mem+swap c... |
3873 |
if (type == _MEM) |
c0ff4b854 memcg: rename mem... |
3874 |
res_counter_reset_max(&memcg->res); |
8c7c6e34a memcg: mem+swap c... |
3875 |
else |
c0ff4b854 memcg: rename mem... |
3876 |
res_counter_reset_max(&memcg->memsw); |
29f2a4dac memcgroup: implem... |
3877 3878 |
break; case RES_FAILCNT: |
8c7c6e34a memcg: mem+swap c... |
3879 |
if (type == _MEM) |
c0ff4b854 memcg: rename mem... |
3880 |
res_counter_reset_failcnt(&memcg->res); |
8c7c6e34a memcg: mem+swap c... |
3881 |
else |
c0ff4b854 memcg: rename mem... |
3882 |
res_counter_reset_failcnt(&memcg->memsw); |
29f2a4dac memcgroup: implem... |
3883 3884 |
break; } |
f64c3f549 memory controller... |
3885 |
|
85cc59db1 memcgroup: use tr... |
3886 |
return 0; |
c84872e16 memcgroup: add th... |
3887 |
} |
7dc74be03 memcg: add interf... |
3888 3889 3890 3891 3892 |
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... |
3893 |
#ifdef CONFIG_MMU |
7dc74be03 memcg: add interf... |
3894 3895 3896 |
static int mem_cgroup_move_charge_write(struct cgroup *cgrp, struct cftype *cft, u64 val) { |
c0ff4b854 memcg: rename mem... |
3897 |
struct mem_cgroup *memcg = mem_cgroup_from_cont(cgrp); |
7dc74be03 memcg: add interf... |
3898 3899 3900 3901 3902 3903 3904 3905 3906 |
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... |
3907 |
memcg->move_charge_at_immigrate = val; |
7dc74be03 memcg: add interf... |
3908 3909 3910 3911 |
cgroup_unlock(); return 0; } |
024914477 memcg: move charg... |
3912 3913 3914 3915 3916 3917 3918 |
#else static int mem_cgroup_move_charge_write(struct cgroup *cgrp, struct cftype *cft, u64 val) { return -ENOSYS; } #endif |
7dc74be03 memcg: add interf... |
3919 |
|
14067bb3e memcg: hierarchic... |
3920 3921 3922 3923 3924 |
/* For read statistics */ enum { MCS_CACHE, MCS_RSS, |
d8046582d memcg: make memcg... |
3925 |
MCS_FILE_MAPPED, |
14067bb3e memcg: hierarchic... |
3926 3927 |
MCS_PGPGIN, MCS_PGPGOUT, |
1dd3a2732 memcg: show swap ... |
3928 |
MCS_SWAP, |
456f998ec memcg: add the pa... |
3929 3930 |
MCS_PGFAULT, MCS_PGMAJFAULT, |
14067bb3e memcg: hierarchic... |
3931 3932 3933 3934 3935 3936 3937 3938 3939 3940 |
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... |
3941 |
}; |
14067bb3e memcg: hierarchic... |
3942 3943 3944 3945 3946 3947 |
struct { char *local_name; char *total_name; } memcg_stat_strings[NR_MCS_STAT] = { {"cache", "total_cache"}, {"rss", "total_rss"}, |
d69b042f3 memcg: add file-b... |
3948 |
{"mapped_file", "total_mapped_file"}, |
14067bb3e memcg: hierarchic... |
3949 3950 |
{"pgpgin", "total_pgpgin"}, {"pgpgout", "total_pgpgout"}, |
1dd3a2732 memcg: show swap ... |
3951 |
{"swap", "total_swap"}, |
456f998ec memcg: add the pa... |
3952 3953 |
{"pgfault", "total_pgfault"}, {"pgmajfault", "total_pgmajfault"}, |
14067bb3e memcg: hierarchic... |
3954 3955 3956 3957 3958 3959 |
{"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... |
3960 |
static void |
c0ff4b854 memcg: rename mem... |
3961 |
mem_cgroup_get_local_stat(struct mem_cgroup *memcg, struct mcs_total_stat *s) |
14067bb3e memcg: hierarchic... |
3962 |
{ |
14067bb3e memcg: hierarchic... |
3963 3964 3965 |
s64 val; /* per cpu stat */ |
c0ff4b854 memcg: rename mem... |
3966 |
val = mem_cgroup_read_stat(memcg, MEM_CGROUP_STAT_CACHE); |
14067bb3e memcg: hierarchic... |
3967 |
s->stat[MCS_CACHE] += val * PAGE_SIZE; |
c0ff4b854 memcg: rename mem... |
3968 |
val = mem_cgroup_read_stat(memcg, MEM_CGROUP_STAT_RSS); |
14067bb3e memcg: hierarchic... |
3969 |
s->stat[MCS_RSS] += val * PAGE_SIZE; |
c0ff4b854 memcg: rename mem... |
3970 |
val = mem_cgroup_read_stat(memcg, MEM_CGROUP_STAT_FILE_MAPPED); |
d8046582d memcg: make memcg... |
3971 |
s->stat[MCS_FILE_MAPPED] += val * PAGE_SIZE; |
c0ff4b854 memcg: rename mem... |
3972 |
val = mem_cgroup_read_events(memcg, MEM_CGROUP_EVENTS_PGPGIN); |
14067bb3e memcg: hierarchic... |
3973 |
s->stat[MCS_PGPGIN] += val; |
c0ff4b854 memcg: rename mem... |
3974 |
val = mem_cgroup_read_events(memcg, MEM_CGROUP_EVENTS_PGPGOUT); |
14067bb3e memcg: hierarchic... |
3975 |
s->stat[MCS_PGPGOUT] += val; |
1dd3a2732 memcg: show swap ... |
3976 |
if (do_swap_account) { |
c0ff4b854 memcg: rename mem... |
3977 |
val = mem_cgroup_read_stat(memcg, MEM_CGROUP_STAT_SWAPOUT); |
1dd3a2732 memcg: show swap ... |
3978 3979 |
s->stat[MCS_SWAP] += val * PAGE_SIZE; } |
c0ff4b854 memcg: rename mem... |
3980 |
val = mem_cgroup_read_events(memcg, MEM_CGROUP_EVENTS_PGFAULT); |
456f998ec memcg: add the pa... |
3981 |
s->stat[MCS_PGFAULT] += val; |
c0ff4b854 memcg: rename mem... |
3982 |
val = mem_cgroup_read_events(memcg, MEM_CGROUP_EVENTS_PGMAJFAULT); |
456f998ec memcg: add the pa... |
3983 |
s->stat[MCS_PGMAJFAULT] += val; |
14067bb3e memcg: hierarchic... |
3984 3985 |
/* per zone stat */ |
c0ff4b854 memcg: rename mem... |
3986 |
val = mem_cgroup_nr_lru_pages(memcg, BIT(LRU_INACTIVE_ANON)); |
14067bb3e memcg: hierarchic... |
3987 |
s->stat[MCS_INACTIVE_ANON] += val * PAGE_SIZE; |
c0ff4b854 memcg: rename mem... |
3988 |
val = mem_cgroup_nr_lru_pages(memcg, BIT(LRU_ACTIVE_ANON)); |
14067bb3e memcg: hierarchic... |
3989 |
s->stat[MCS_ACTIVE_ANON] += val * PAGE_SIZE; |
c0ff4b854 memcg: rename mem... |
3990 |
val = mem_cgroup_nr_lru_pages(memcg, BIT(LRU_INACTIVE_FILE)); |
14067bb3e memcg: hierarchic... |
3991 |
s->stat[MCS_INACTIVE_FILE] += val * PAGE_SIZE; |
c0ff4b854 memcg: rename mem... |
3992 |
val = mem_cgroup_nr_lru_pages(memcg, BIT(LRU_ACTIVE_FILE)); |
14067bb3e memcg: hierarchic... |
3993 |
s->stat[MCS_ACTIVE_FILE] += val * PAGE_SIZE; |
c0ff4b854 memcg: rename mem... |
3994 |
val = mem_cgroup_nr_lru_pages(memcg, BIT(LRU_UNEVICTABLE)); |
14067bb3e memcg: hierarchic... |
3995 |
s->stat[MCS_UNEVICTABLE] += val * PAGE_SIZE; |
14067bb3e memcg: hierarchic... |
3996 3997 3998 |
} static void |
c0ff4b854 memcg: rename mem... |
3999 |
mem_cgroup_get_total_stat(struct mem_cgroup *memcg, struct mcs_total_stat *s) |
14067bb3e memcg: hierarchic... |
4000 |
{ |
7d74b06f2 memcg: use for_ea... |
4001 |
struct mem_cgroup *iter; |
c0ff4b854 memcg: rename mem... |
4002 |
for_each_mem_cgroup_tree(iter, memcg) |
7d74b06f2 memcg: use for_ea... |
4003 |
mem_cgroup_get_local_stat(iter, s); |
14067bb3e memcg: hierarchic... |
4004 |
} |
406eb0c9b memcg: add memory... |
4005 4006 4007 4008 4009 4010 4011 4012 |
#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... |
4013 |
total_nr = mem_cgroup_nr_lru_pages(mem_cont, LRU_ALL); |
406eb0c9b memcg: add memory... |
4014 4015 |
seq_printf(m, "total=%lu", total_nr); for_each_node_state(nid, N_HIGH_MEMORY) { |
bb2a0de92 memcg: consolidat... |
4016 |
node_nr = mem_cgroup_node_nr_lru_pages(mem_cont, nid, LRU_ALL); |
406eb0c9b memcg: add memory... |
4017 4018 4019 4020 |
seq_printf(m, " N%d=%lu", nid, node_nr); } seq_putc(m, ' '); |
bb2a0de92 memcg: consolidat... |
4021 |
file_nr = mem_cgroup_nr_lru_pages(mem_cont, LRU_ALL_FILE); |
406eb0c9b memcg: add memory... |
4022 4023 |
seq_printf(m, "file=%lu", file_nr); for_each_node_state(nid, N_HIGH_MEMORY) { |
bb2a0de92 memcg: consolidat... |
4024 4025 |
node_nr = mem_cgroup_node_nr_lru_pages(mem_cont, nid, LRU_ALL_FILE); |
406eb0c9b memcg: add memory... |
4026 4027 4028 4029 |
seq_printf(m, " N%d=%lu", nid, node_nr); } seq_putc(m, ' '); |
bb2a0de92 memcg: consolidat... |
4030 |
anon_nr = mem_cgroup_nr_lru_pages(mem_cont, LRU_ALL_ANON); |
406eb0c9b memcg: add memory... |
4031 4032 |
seq_printf(m, "anon=%lu", anon_nr); for_each_node_state(nid, N_HIGH_MEMORY) { |
bb2a0de92 memcg: consolidat... |
4033 4034 |
node_nr = mem_cgroup_node_nr_lru_pages(mem_cont, nid, LRU_ALL_ANON); |
406eb0c9b memcg: add memory... |
4035 4036 4037 4038 |
seq_printf(m, " N%d=%lu", nid, node_nr); } seq_putc(m, ' '); |
bb2a0de92 memcg: consolidat... |
4039 |
unevictable_nr = mem_cgroup_nr_lru_pages(mem_cont, BIT(LRU_UNEVICTABLE)); |
406eb0c9b memcg: add memory... |
4040 4041 |
seq_printf(m, "unevictable=%lu", unevictable_nr); for_each_node_state(nid, N_HIGH_MEMORY) { |
bb2a0de92 memcg: consolidat... |
4042 4043 |
node_nr = mem_cgroup_node_nr_lru_pages(mem_cont, nid, BIT(LRU_UNEVICTABLE)); |
406eb0c9b memcg: add memory... |
4044 4045 4046 4047 4048 4049 4050 |
seq_printf(m, " N%d=%lu", nid, node_nr); } seq_putc(m, ' '); return 0; } #endif /* CONFIG_NUMA */ |
c64745cf0 CGroup API files:... |
4051 4052 |
static int mem_control_stat_show(struct cgroup *cont, struct cftype *cft, struct cgroup_map_cb *cb) |
d2ceb9b7d memory cgroup enh... |
4053 |
{ |
d2ceb9b7d memory cgroup enh... |
4054 |
struct mem_cgroup *mem_cont = mem_cgroup_from_cont(cont); |
14067bb3e memcg: hierarchic... |
4055 |
struct mcs_total_stat mystat; |
d2ceb9b7d memory cgroup enh... |
4056 |
int i; |
14067bb3e memcg: hierarchic... |
4057 4058 |
memset(&mystat, 0, sizeof(mystat)); mem_cgroup_get_local_stat(mem_cont, &mystat); |
d2ceb9b7d memory cgroup enh... |
4059 |
|
406eb0c9b memcg: add memory... |
4060 |
|
1dd3a2732 memcg: show swap ... |
4061 4062 4063 |
for (i = 0; i < NR_MCS_STAT; i++) { if (i == MCS_SWAP && !do_swap_account) continue; |
14067bb3e memcg: hierarchic... |
4064 |
cb->fill(cb, memcg_stat_strings[i].local_name, mystat.stat[i]); |
1dd3a2732 memcg: show swap ... |
4065 |
} |
7b854121e Unevictable LRU P... |
4066 |
|
14067bb3e memcg: hierarchic... |
4067 |
/* Hierarchical information */ |
fee7b548e memcg: show real ... |
4068 4069 4070 4071 4072 4073 4074 |
{ 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... |
4075 |
|
14067bb3e memcg: hierarchic... |
4076 4077 |
memset(&mystat, 0, sizeof(mystat)); mem_cgroup_get_total_stat(mem_cont, &mystat); |
1dd3a2732 memcg: show swap ... |
4078 4079 4080 |
for (i = 0; i < NR_MCS_STAT; i++) { if (i == MCS_SWAP && !do_swap_account) continue; |
14067bb3e memcg: hierarchic... |
4081 |
cb->fill(cb, memcg_stat_strings[i].total_name, mystat.stat[i]); |
1dd3a2732 memcg: show swap ... |
4082 |
} |
14067bb3e memcg: hierarchic... |
4083 |
|
7f016ee8b memcg: show recla... |
4084 |
#ifdef CONFIG_DEBUG_VM |
7f016ee8b memcg: show recla... |
4085 4086 4087 4088 4089 4090 4091 4092 4093 4094 4095 4096 4097 4098 4099 4100 4101 4102 4103 4104 4105 4106 4107 4108 4109 |
{ 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... |
4110 4111 |
return 0; } |
a7885eb8a memcg: swappiness |
4112 4113 4114 |
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... |
4115 |
return mem_cgroup_swappiness(memcg); |
a7885eb8a memcg: swappiness |
4116 4117 4118 4119 4120 4121 4122 |
} 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... |
4123 |
|
a7885eb8a memcg: swappiness |
4124 4125 4126 4127 4128 4129 4130 |
if (val > 100) return -EINVAL; if (cgrp->parent == NULL) return -EINVAL; parent = mem_cgroup_from_cont(cgrp->parent); |
068b38c1f memcg: fix a race... |
4131 4132 |
cgroup_lock(); |
a7885eb8a memcg: swappiness |
4133 4134 |
/* If under hierarchy, only empty-root can set this value */ if ((parent->use_hierarchy) || |
068b38c1f memcg: fix a race... |
4135 4136 |
(memcg->use_hierarchy && !list_empty(&cgrp->children))) { cgroup_unlock(); |
a7885eb8a memcg: swappiness |
4137 |
return -EINVAL; |
068b38c1f memcg: fix a race... |
4138 |
} |
a7885eb8a memcg: swappiness |
4139 |
|
a7885eb8a memcg: swappiness |
4140 |
memcg->swappiness = val; |
a7885eb8a memcg: swappiness |
4141 |
|
068b38c1f memcg: fix a race... |
4142 |
cgroup_unlock(); |
a7885eb8a memcg: swappiness |
4143 4144 |
return 0; } |
2e72b6347 memcg: implement ... |
4145 4146 4147 4148 4149 4150 4151 4152 |
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... |
4153 |
t = rcu_dereference(memcg->thresholds.primary); |
2e72b6347 memcg: implement ... |
4154 |
else |
2c488db27 memcg: clean up m... |
4155 |
t = rcu_dereference(memcg->memsw_thresholds.primary); |
2e72b6347 memcg: implement ... |
4156 4157 4158 4159 4160 4161 4162 4163 4164 4165 4166 |
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... |
4167 |
i = t->current_threshold; |
2e72b6347 memcg: implement ... |
4168 4169 4170 4171 4172 4173 4174 4175 4176 4177 4178 4179 4180 4181 4182 4183 4184 4185 4186 4187 4188 4189 4190 |
/* * 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... |
4191 |
t->current_threshold = i - 1; |
2e72b6347 memcg: implement ... |
4192 4193 4194 4195 4196 4197 |
unlock: rcu_read_unlock(); } static void mem_cgroup_threshold(struct mem_cgroup *memcg) { |
ad4ca5f4b memcg: fix thresh... |
4198 4199 4200 4201 4202 4203 4204 |
while (memcg) { __mem_cgroup_threshold(memcg, false); if (do_swap_account) __mem_cgroup_threshold(memcg, true); memcg = parent_mem_cgroup(memcg); } |
2e72b6347 memcg: implement ... |
4205 4206 4207 4208 4209 4210 4211 4212 4213 |
} 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... |
4214 |
static int mem_cgroup_oom_notify_cb(struct mem_cgroup *memcg) |
9490ff275 memcg: oom notifier |
4215 4216 |
{ struct mem_cgroup_eventfd_list *ev; |
c0ff4b854 memcg: rename mem... |
4217 |
list_for_each_entry(ev, &memcg->oom_notify, list) |
9490ff275 memcg: oom notifier |
4218 4219 4220 |
eventfd_signal(ev->eventfd, 1); return 0; } |
c0ff4b854 memcg: rename mem... |
4221 |
static void mem_cgroup_oom_notify(struct mem_cgroup *memcg) |
9490ff275 memcg: oom notifier |
4222 |
{ |
7d74b06f2 memcg: use for_ea... |
4223 |
struct mem_cgroup *iter; |
c0ff4b854 memcg: rename mem... |
4224 |
for_each_mem_cgroup_tree(iter, memcg) |
7d74b06f2 memcg: use for_ea... |
4225 |
mem_cgroup_oom_notify_cb(iter); |
9490ff275 memcg: oom notifier |
4226 4227 4228 4229 |
} static int mem_cgroup_usage_register_event(struct cgroup *cgrp, struct cftype *cft, struct eventfd_ctx *eventfd, const char *args) |
2e72b6347 memcg: implement ... |
4230 4231 |
{ struct mem_cgroup *memcg = mem_cgroup_from_cont(cgrp); |
2c488db27 memcg: clean up m... |
4232 4233 |
struct mem_cgroup_thresholds *thresholds; struct mem_cgroup_threshold_ary *new; |
2e72b6347 memcg: implement ... |
4234 4235 |
int type = MEMFILE_TYPE(cft->private); u64 threshold, usage; |
2c488db27 memcg: clean up m... |
4236 |
int i, size, ret; |
2e72b6347 memcg: implement ... |
4237 4238 4239 4240 4241 4242 |
ret = res_counter_memparse_write_strategy(args, &threshold); if (ret) return ret; mutex_lock(&memcg->thresholds_lock); |
2c488db27 memcg: clean up m... |
4243 |
|
2e72b6347 memcg: implement ... |
4244 |
if (type == _MEM) |
2c488db27 memcg: clean up m... |
4245 |
thresholds = &memcg->thresholds; |
2e72b6347 memcg: implement ... |
4246 |
else if (type == _MEMSWAP) |
2c488db27 memcg: clean up m... |
4247 |
thresholds = &memcg->memsw_thresholds; |
2e72b6347 memcg: implement ... |
4248 4249 4250 4251 4252 4253 |
else BUG(); usage = mem_cgroup_usage(memcg, type == _MEMSWAP); /* Check if a threshold crossed before adding a new one */ |
2c488db27 memcg: clean up m... |
4254 |
if (thresholds->primary) |
2e72b6347 memcg: implement ... |
4255 |
__mem_cgroup_threshold(memcg, type == _MEMSWAP); |
2c488db27 memcg: clean up m... |
4256 |
size = thresholds->primary ? thresholds->primary->size + 1 : 1; |
2e72b6347 memcg: implement ... |
4257 4258 |
/* Allocate memory for new array of thresholds */ |
2c488db27 memcg: clean up m... |
4259 |
new = kmalloc(sizeof(*new) + size * sizeof(struct mem_cgroup_threshold), |
2e72b6347 memcg: implement ... |
4260 |
GFP_KERNEL); |
2c488db27 memcg: clean up m... |
4261 |
if (!new) { |
2e72b6347 memcg: implement ... |
4262 4263 4264 |
ret = -ENOMEM; goto unlock; } |
2c488db27 memcg: clean up m... |
4265 |
new->size = size; |
2e72b6347 memcg: implement ... |
4266 4267 |
/* Copy thresholds (if any) to new array */ |
2c488db27 memcg: clean up m... |
4268 4269 |
if (thresholds->primary) { memcpy(new->entries, thresholds->primary->entries, (size - 1) * |
2e72b6347 memcg: implement ... |
4270 |
sizeof(struct mem_cgroup_threshold)); |
2c488db27 memcg: clean up m... |
4271 |
} |
2e72b6347 memcg: implement ... |
4272 |
/* Add new threshold */ |
2c488db27 memcg: clean up m... |
4273 4274 |
new->entries[size - 1].eventfd = eventfd; new->entries[size - 1].threshold = threshold; |
2e72b6347 memcg: implement ... |
4275 4276 |
/* Sort thresholds. Registering of new threshold isn't time-critical */ |
2c488db27 memcg: clean up m... |
4277 |
sort(new->entries, size, sizeof(struct mem_cgroup_threshold), |
2e72b6347 memcg: implement ... |
4278 4279 4280 |
compare_thresholds, NULL); /* Find current threshold */ |
2c488db27 memcg: clean up m... |
4281 |
new->current_threshold = -1; |
2e72b6347 memcg: implement ... |
4282 |
for (i = 0; i < size; i++) { |
2c488db27 memcg: clean up m... |
4283 |
if (new->entries[i].threshold < usage) { |
2e72b6347 memcg: implement ... |
4284 |
/* |
2c488db27 memcg: clean up m... |
4285 4286 |
* new->current_threshold will not be used until * rcu_assign_pointer(), so it's safe to increment |
2e72b6347 memcg: implement ... |
4287 4288 |
* it here. */ |
2c488db27 memcg: clean up m... |
4289 |
++new->current_threshold; |
2e72b6347 memcg: implement ... |
4290 4291 |
} } |
2c488db27 memcg: clean up m... |
4292 4293 4294 4295 4296 |
/* 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 ... |
4297 |
|
907860ed3 cgroups: make cft... |
4298 |
/* To be sure that nobody uses thresholds */ |
2e72b6347 memcg: implement ... |
4299 |
synchronize_rcu(); |
2e72b6347 memcg: implement ... |
4300 4301 4302 4303 4304 |
unlock: mutex_unlock(&memcg->thresholds_lock); return ret; } |
907860ed3 cgroups: make cft... |
4305 |
static void mem_cgroup_usage_unregister_event(struct cgroup *cgrp, |
9490ff275 memcg: oom notifier |
4306 |
struct cftype *cft, struct eventfd_ctx *eventfd) |
2e72b6347 memcg: implement ... |
4307 4308 |
{ struct mem_cgroup *memcg = mem_cgroup_from_cont(cgrp); |
2c488db27 memcg: clean up m... |
4309 4310 |
struct mem_cgroup_thresholds *thresholds; struct mem_cgroup_threshold_ary *new; |
2e72b6347 memcg: implement ... |
4311 4312 |
int type = MEMFILE_TYPE(cft->private); u64 usage; |
2c488db27 memcg: clean up m... |
4313 |
int i, j, size; |
2e72b6347 memcg: implement ... |
4314 4315 4316 |
mutex_lock(&memcg->thresholds_lock); if (type == _MEM) |
2c488db27 memcg: clean up m... |
4317 |
thresholds = &memcg->thresholds; |
2e72b6347 memcg: implement ... |
4318 |
else if (type == _MEMSWAP) |
2c488db27 memcg: clean up m... |
4319 |
thresholds = &memcg->memsw_thresholds; |
2e72b6347 memcg: implement ... |
4320 4321 4322 4323 4324 4325 4326 4327 4328 4329 4330 4331 4332 4333 4334 |
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... |
4335 4336 4337 |
size = 0; for (i = 0; i < thresholds->primary->size; i++) { if (thresholds->primary->entries[i].eventfd != eventfd) |
2e72b6347 memcg: implement ... |
4338 4339 |
size++; } |
2c488db27 memcg: clean up m... |
4340 |
new = thresholds->spare; |
907860ed3 cgroups: make cft... |
4341 |
|
2e72b6347 memcg: implement ... |
4342 4343 |
/* Set thresholds array to NULL if we don't have thresholds */ if (!size) { |
2c488db27 memcg: clean up m... |
4344 4345 |
kfree(new); new = NULL; |
907860ed3 cgroups: make cft... |
4346 |
goto swap_buffers; |
2e72b6347 memcg: implement ... |
4347 |
} |
2c488db27 memcg: clean up m... |
4348 |
new->size = size; |
2e72b6347 memcg: implement ... |
4349 4350 |
/* Copy thresholds and find current threshold */ |
2c488db27 memcg: clean up m... |
4351 4352 4353 |
new->current_threshold = -1; for (i = 0, j = 0; i < thresholds->primary->size; i++) { if (thresholds->primary->entries[i].eventfd == eventfd) |
2e72b6347 memcg: implement ... |
4354 |
continue; |
2c488db27 memcg: clean up m... |
4355 4356 |
new->entries[j] = thresholds->primary->entries[i]; if (new->entries[j].threshold < usage) { |
2e72b6347 memcg: implement ... |
4357 |
/* |
2c488db27 memcg: clean up m... |
4358 |
* new->current_threshold will not be used |
2e72b6347 memcg: implement ... |
4359 4360 4361 |
* until rcu_assign_pointer(), so it's safe to increment * it here. */ |
2c488db27 memcg: clean up m... |
4362 |
++new->current_threshold; |
2e72b6347 memcg: implement ... |
4363 4364 4365 |
} j++; } |
907860ed3 cgroups: make cft... |
4366 |
swap_buffers: |
2c488db27 memcg: clean up m... |
4367 4368 4369 |
/* Swap primary and spare array */ thresholds->spare = thresholds->primary; rcu_assign_pointer(thresholds->primary, new); |
2e72b6347 memcg: implement ... |
4370 |
|
907860ed3 cgroups: make cft... |
4371 |
/* To be sure that nobody uses thresholds */ |
2e72b6347 memcg: implement ... |
4372 |
synchronize_rcu(); |
2e72b6347 memcg: implement ... |
4373 |
mutex_unlock(&memcg->thresholds_lock); |
2e72b6347 memcg: implement ... |
4374 |
} |
c1e862c1f memcg: new force_... |
4375 |
|
9490ff275 memcg: oom notifier |
4376 4377 4378 4379 4380 4381 4382 4383 4384 4385 4386 |
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... |
4387 |
spin_lock(&memcg_oom_lock); |
9490ff275 memcg: oom notifier |
4388 4389 4390 4391 4392 |
event->eventfd = eventfd; list_add(&event->list, &memcg->oom_notify); /* already in OOM ? */ |
79dfdaccd memcg: make oom_l... |
4393 |
if (atomic_read(&memcg->under_oom)) |
9490ff275 memcg: oom notifier |
4394 |
eventfd_signal(eventfd, 1); |
1af8efe96 memcg: change mem... |
4395 |
spin_unlock(&memcg_oom_lock); |
9490ff275 memcg: oom notifier |
4396 4397 4398 |
return 0; } |
907860ed3 cgroups: make cft... |
4399 |
static void mem_cgroup_oom_unregister_event(struct cgroup *cgrp, |
9490ff275 memcg: oom notifier |
4400 4401 |
struct cftype *cft, struct eventfd_ctx *eventfd) { |
c0ff4b854 memcg: rename mem... |
4402 |
struct mem_cgroup *memcg = mem_cgroup_from_cont(cgrp); |
9490ff275 memcg: oom notifier |
4403 4404 4405 4406 |
struct mem_cgroup_eventfd_list *ev, *tmp; int type = MEMFILE_TYPE(cft->private); BUG_ON(type != _OOM_TYPE); |
1af8efe96 memcg: change mem... |
4407 |
spin_lock(&memcg_oom_lock); |
9490ff275 memcg: oom notifier |
4408 |
|
c0ff4b854 memcg: rename mem... |
4409 |
list_for_each_entry_safe(ev, tmp, &memcg->oom_notify, list) { |
9490ff275 memcg: oom notifier |
4410 4411 4412 4413 4414 |
if (ev->eventfd == eventfd) { list_del(&ev->list); kfree(ev); } } |
1af8efe96 memcg: change mem... |
4415 |
spin_unlock(&memcg_oom_lock); |
9490ff275 memcg: oom notifier |
4416 |
} |
3c11ecf44 memcg: oom kill d... |
4417 4418 4419 |
static int mem_cgroup_oom_control_read(struct cgroup *cgrp, struct cftype *cft, struct cgroup_map_cb *cb) { |
c0ff4b854 memcg: rename mem... |
4420 |
struct mem_cgroup *memcg = mem_cgroup_from_cont(cgrp); |
3c11ecf44 memcg: oom kill d... |
4421 |
|
c0ff4b854 memcg: rename mem... |
4422 |
cb->fill(cb, "oom_kill_disable", memcg->oom_kill_disable); |
3c11ecf44 memcg: oom kill d... |
4423 |
|
c0ff4b854 memcg: rename mem... |
4424 |
if (atomic_read(&memcg->under_oom)) |
3c11ecf44 memcg: oom kill d... |
4425 4426 4427 4428 4429 |
cb->fill(cb, "under_oom", 1); else cb->fill(cb, "under_oom", 0); return 0; } |
3c11ecf44 memcg: oom kill d... |
4430 4431 4432 |
static int mem_cgroup_oom_control_write(struct cgroup *cgrp, struct cftype *cft, u64 val) { |
c0ff4b854 memcg: rename mem... |
4433 |
struct mem_cgroup *memcg = mem_cgroup_from_cont(cgrp); |
3c11ecf44 memcg: oom kill d... |
4434 4435 4436 4437 4438 4439 4440 4441 4442 4443 4444 |
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... |
4445 |
(memcg->use_hierarchy && !list_empty(&cgrp->children))) { |
3c11ecf44 memcg: oom kill d... |
4446 4447 4448 |
cgroup_unlock(); return -EINVAL; } |
c0ff4b854 memcg: rename mem... |
4449 |
memcg->oom_kill_disable = val; |
4d845ebf4 memcg: fix wake u... |
4450 |
if (!val) |
c0ff4b854 memcg: rename mem... |
4451 |
memcg_oom_recover(memcg); |
3c11ecf44 memcg: oom kill d... |
4452 4453 4454 |
cgroup_unlock(); return 0; } |
406eb0c9b memcg: add memory... |
4455 4456 4457 4458 4459 4460 4461 4462 4463 4464 4465 4466 4467 4468 4469 |
#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... |
4470 |
#ifdef CONFIG_CGROUP_MEM_RES_CTLR_KMEM |
e5671dfae Basic kernel memo... |
4471 4472 |
static int register_kmem_files(struct cgroup *cont, struct cgroup_subsys *ss) { |
d1a4c0b37 tcp memory pressu... |
4473 4474 4475 4476 4477 4478 4479 |
/* * 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... |
4480 |
return mem_cgroup_sockets_init(cont, ss); |
e5671dfae Basic kernel memo... |
4481 |
}; |
d1a4c0b37 tcp memory pressu... |
4482 4483 4484 4485 4486 |
static void kmem_cgroup_destroy(struct cgroup_subsys *ss, struct cgroup *cont) { mem_cgroup_sockets_destroy(cont, ss); } |
e5671dfae Basic kernel memo... |
4487 4488 4489 4490 4491 |
#else static int register_kmem_files(struct cgroup *cont, struct cgroup_subsys *ss) { return 0; } |
d1a4c0b37 tcp memory pressu... |
4492 4493 4494 4495 4496 |
static void kmem_cgroup_destroy(struct cgroup_subsys *ss, struct cgroup *cont) { } |
e5671dfae Basic kernel memo... |
4497 |
#endif |
8cdea7c05 Memory controller... |
4498 4499 |
static struct cftype mem_cgroup_files[] = { { |
0eea10301 Memory controller... |
4500 |
.name = "usage_in_bytes", |
8c7c6e34a memcg: mem+swap c... |
4501 |
.private = MEMFILE_PRIVATE(_MEM, RES_USAGE), |
2c3daa722 CGroup API files:... |
4502 |
.read_u64 = mem_cgroup_read, |
9490ff275 memcg: oom notifier |
4503 4504 |
.register_event = mem_cgroup_usage_register_event, .unregister_event = mem_cgroup_usage_unregister_event, |
8cdea7c05 Memory controller... |
4505 4506 |
}, { |
c84872e16 memcgroup: add th... |
4507 |
.name = "max_usage_in_bytes", |
8c7c6e34a memcg: mem+swap c... |
4508 |
.private = MEMFILE_PRIVATE(_MEM, RES_MAX_USAGE), |
29f2a4dac memcgroup: implem... |
4509 |
.trigger = mem_cgroup_reset, |
c84872e16 memcgroup: add th... |
4510 4511 4512 |
.read_u64 = mem_cgroup_read, }, { |
0eea10301 Memory controller... |
4513 |
.name = "limit_in_bytes", |
8c7c6e34a memcg: mem+swap c... |
4514 |
.private = MEMFILE_PRIVATE(_MEM, RES_LIMIT), |
856c13aa1 cgroup files: con... |
4515 |
.write_string = mem_cgroup_write, |
2c3daa722 CGroup API files:... |
4516 |
.read_u64 = mem_cgroup_read, |
8cdea7c05 Memory controller... |
4517 4518 |
}, { |
296c81d89 memory controller... |
4519 4520 4521 4522 4523 4524 |
.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... |
4525 |
.name = "failcnt", |
8c7c6e34a memcg: mem+swap c... |
4526 |
.private = MEMFILE_PRIVATE(_MEM, RES_FAILCNT), |
29f2a4dac memcgroup: implem... |
4527 |
.trigger = mem_cgroup_reset, |
2c3daa722 CGroup API files:... |
4528 |
.read_u64 = mem_cgroup_read, |
8cdea7c05 Memory controller... |
4529 |
}, |
8697d3319 Memory controller... |
4530 |
{ |
d2ceb9b7d memory cgroup enh... |
4531 |
.name = "stat", |
c64745cf0 CGroup API files:... |
4532 |
.read_map = mem_control_stat_show, |
d2ceb9b7d memory cgroup enh... |
4533 |
}, |
c1e862c1f memcg: new force_... |
4534 4535 4536 4537 |
{ .name = "force_empty", .trigger = mem_cgroup_force_empty_write, }, |
18f59ea7d memcg: memory cgr... |
4538 4539 4540 4541 4542 |
{ .name = "use_hierarchy", .write_u64 = mem_cgroup_hierarchy_write, .read_u64 = mem_cgroup_hierarchy_read, }, |
a7885eb8a memcg: swappiness |
4543 4544 4545 4546 4547 |
{ .name = "swappiness", .read_u64 = mem_cgroup_swappiness_read, .write_u64 = mem_cgroup_swappiness_write, }, |
7dc74be03 memcg: add interf... |
4548 4549 4550 4551 4552 |
{ .name = "move_charge_at_immigrate", .read_u64 = mem_cgroup_move_charge_read, .write_u64 = mem_cgroup_move_charge_write, }, |
9490ff275 memcg: oom notifier |
4553 4554 |
{ .name = "oom_control", |
3c11ecf44 memcg: oom kill d... |
4555 4556 |
.read_map = mem_cgroup_oom_control_read, .write_u64 = mem_cgroup_oom_control_write, |
9490ff275 memcg: oom notifier |
4557 4558 4559 4560 |
.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... |
4561 4562 4563 4564 |
#ifdef CONFIG_NUMA { .name = "numa_stat", .open = mem_control_numa_stat_open, |
895771271 mm: memory.numa_s... |
4565 |
.mode = S_IRUGO, |
406eb0c9b memcg: add memory... |
4566 4567 |
}, #endif |
8cdea7c05 Memory controller... |
4568 |
}; |
8c7c6e34a memcg: mem+swap c... |
4569 4570 4571 4572 4573 4574 |
#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 |
4575 4576 |
.register_event = mem_cgroup_usage_register_event, .unregister_event = mem_cgroup_usage_unregister_event, |
8c7c6e34a memcg: mem+swap c... |
4577 4578 4579 4580 4581 4582 4583 4584 4585 4586 4587 4588 4589 4590 4591 4592 4593 4594 4595 4596 4597 4598 4599 4600 4601 4602 4603 4604 4605 4606 4607 4608 4609 4610 |
}, { .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... |
4611 |
static int alloc_mem_cgroup_per_zone_info(struct mem_cgroup *memcg, int node) |
6d12e2d8d per-zone and recl... |
4612 4613 |
{ struct mem_cgroup_per_node *pn; |
1ecaab2bd per-zone and recl... |
4614 |
struct mem_cgroup_per_zone *mz; |
b69408e88 vmscan: Use an in... |
4615 |
enum lru_list l; |
41e3355de memcg: fix node_s... |
4616 |
int zone, tmp = node; |
1ecaab2bd per-zone and recl... |
4617 4618 4619 4620 4621 4622 4623 4624 |
/* * 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... |
4625 4626 |
if (!node_state(node, N_NORMAL_MEMORY)) tmp = -1; |
17295c88a memcg: use [kv]za... |
4627 |
pn = kzalloc_node(sizeof(*pn), GFP_KERNEL, tmp); |
6d12e2d8d per-zone and recl... |
4628 4629 |
if (!pn) return 1; |
1ecaab2bd per-zone and recl... |
4630 |
|
1ecaab2bd per-zone and recl... |
4631 4632 |
for (zone = 0; zone < MAX_NR_ZONES; zone++) { mz = &pn->zoneinfo[zone]; |
b69408e88 vmscan: Use an in... |
4633 4634 |
for_each_lru(l) INIT_LIST_HEAD(&mz->lists[l]); |
f64c3f549 memory controller... |
4635 |
mz->usage_in_excess = 0; |
4e4169535 memory controller... |
4636 |
mz->on_tree = false; |
c0ff4b854 memcg: rename mem... |
4637 |
mz->mem = memcg; |
1ecaab2bd per-zone and recl... |
4638 |
} |
0a619e587 memcg: do not exp... |
4639 |
memcg->info.nodeinfo[node] = pn; |
6d12e2d8d per-zone and recl... |
4640 4641 |
return 0; } |
c0ff4b854 memcg: rename mem... |
4642 |
static void free_mem_cgroup_per_zone_info(struct mem_cgroup *memcg, int node) |
1ecaab2bd per-zone and recl... |
4643 |
{ |
c0ff4b854 memcg: rename mem... |
4644 |
kfree(memcg->info.nodeinfo[node]); |
1ecaab2bd per-zone and recl... |
4645 |
} |
333279487 memcgroup: use vm... |
4646 4647 4648 |
static struct mem_cgroup *mem_cgroup_alloc(void) { struct mem_cgroup *mem; |
c62b1a3b3 memcg: use generi... |
4649 |
int size = sizeof(struct mem_cgroup); |
333279487 memcgroup: use vm... |
4650 |
|
c62b1a3b3 memcg: use generi... |
4651 |
/* Can be very big if MAX_NUMNODES is very big */ |
c8dad2bb6 memcg: reduce siz... |
4652 |
if (size < PAGE_SIZE) |
17295c88a memcg: use [kv]za... |
4653 |
mem = kzalloc(size, GFP_KERNEL); |
333279487 memcgroup: use vm... |
4654 |
else |
17295c88a memcg: use [kv]za... |
4655 |
mem = vzalloc(size); |
333279487 memcgroup: use vm... |
4656 |
|
e7bbcdf37 memcontrol: fix p... |
4657 4658 |
if (!mem) return NULL; |
c62b1a3b3 memcg: use generi... |
4659 |
mem->stat = alloc_percpu(struct mem_cgroup_stat_cpu); |
d2e61b8dc memcg: null deref... |
4660 4661 |
if (!mem->stat) goto out_free; |
711d3d2c9 memcg: cpu hotplu... |
4662 |
spin_lock_init(&mem->pcp_counter_lock); |
333279487 memcgroup: use vm... |
4663 |
return mem; |
d2e61b8dc memcg: null deref... |
4664 4665 4666 4667 4668 4669 4670 |
out_free: if (size < PAGE_SIZE) kfree(mem); else vfree(mem); return NULL; |
333279487 memcgroup: use vm... |
4671 |
} |
8c7c6e34a memcg: mem+swap c... |
4672 4673 4674 4675 4676 4677 4678 4679 |
/* * 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... |
4680 4681 |
* Removal of cgroup itself succeeds regardless of refs from swap. */ |
c0ff4b854 memcg: rename mem... |
4682 |
static void __mem_cgroup_free(struct mem_cgroup *memcg) |
333279487 memcgroup: use vm... |
4683 |
{ |
08e552c69 memcg: synchroniz... |
4684 |
int node; |
c0ff4b854 memcg: rename mem... |
4685 4686 |
mem_cgroup_remove_from_trees(memcg); free_css_id(&mem_cgroup_subsys, &memcg->css); |
04046e1a0 memcg: use CSS ID |
4687 |
|
08e552c69 memcg: synchroniz... |
4688 |
for_each_node_state(node, N_POSSIBLE) |
c0ff4b854 memcg: rename mem... |
4689 |
free_mem_cgroup_per_zone_info(memcg, node); |
08e552c69 memcg: synchroniz... |
4690 |
|
c0ff4b854 memcg: rename mem... |
4691 |
free_percpu(memcg->stat); |
c62b1a3b3 memcg: use generi... |
4692 |
if (sizeof(struct mem_cgroup) < PAGE_SIZE) |
c0ff4b854 memcg: rename mem... |
4693 |
kfree(memcg); |
333279487 memcgroup: use vm... |
4694 |
else |
c0ff4b854 memcg: rename mem... |
4695 |
vfree(memcg); |
333279487 memcgroup: use vm... |
4696 |
} |
c0ff4b854 memcg: rename mem... |
4697 |
static void mem_cgroup_get(struct mem_cgroup *memcg) |
8c7c6e34a memcg: mem+swap c... |
4698 |
{ |
c0ff4b854 memcg: rename mem... |
4699 |
atomic_inc(&memcg->refcnt); |
8c7c6e34a memcg: mem+swap c... |
4700 |
} |
c0ff4b854 memcg: rename mem... |
4701 |
static void __mem_cgroup_put(struct mem_cgroup *memcg, int count) |
8c7c6e34a memcg: mem+swap c... |
4702 |
{ |
c0ff4b854 memcg: rename mem... |
4703 4704 4705 |
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... |
4706 4707 4708 |
if (parent) mem_cgroup_put(parent); } |
8c7c6e34a memcg: mem+swap c... |
4709 |
} |
c0ff4b854 memcg: rename mem... |
4710 |
static void mem_cgroup_put(struct mem_cgroup *memcg) |
483c30b51 memcg: improve pe... |
4711 |
{ |
c0ff4b854 memcg: rename mem... |
4712 |
__mem_cgroup_put(memcg, 1); |
483c30b51 memcg: improve pe... |
4713 |
} |
7bcc1bb12 memcg: get/put pa... |
4714 4715 4716 |
/* * Returns the parent mem_cgroup in memcgroup hierarchy with hierarchy enabled. */ |
e1aab161e socket: initial c... |
4717 |
struct mem_cgroup *parent_mem_cgroup(struct mem_cgroup *memcg) |
7bcc1bb12 memcg: get/put pa... |
4718 |
{ |
c0ff4b854 memcg: rename mem... |
4719 |
if (!memcg->res.parent) |
7bcc1bb12 memcg: get/put pa... |
4720 |
return NULL; |
c0ff4b854 memcg: rename mem... |
4721 |
return mem_cgroup_from_res_counter(memcg->res.parent, res); |
7bcc1bb12 memcg: get/put pa... |
4722 |
} |
e1aab161e socket: initial c... |
4723 |
EXPORT_SYMBOL(parent_mem_cgroup); |
333279487 memcgroup: use vm... |
4724 |
|
c077719be memcg: mem+swap c... |
4725 4726 4727 |
#ifdef CONFIG_CGROUP_MEM_RES_CTLR_SWAP static void __init enable_swap_cgroup(void) { |
f8d665422 memcg: add mem_cg... |
4728 |
if (!mem_cgroup_disabled() && really_do_swap_account) |
c077719be memcg: mem+swap c... |
4729 4730 4731 4732 4733 4734 4735 |
do_swap_account = 1; } #else static void __init enable_swap_cgroup(void) { } #endif |
f64c3f549 memory controller... |
4736 4737 4738 4739 4740 4741 4742 4743 4744 4745 4746 4747 4748 4749 4750 4751 4752 4753 4754 4755 4756 4757 4758 4759 |
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; for_each_node_state(node, N_POSSIBLE) { tmp = node; if (!node_state(node, N_NORMAL_MEMORY)) tmp = -1; rtpn = kzalloc_node(sizeof(*rtpn), GFP_KERNEL, tmp); if (!rtpn) return 1; 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; } |
0eb253e22 memcg: fix sectio... |
4760 |
static struct cgroup_subsys_state * __ref |
8cdea7c05 Memory controller... |
4761 4762 |
mem_cgroup_create(struct cgroup_subsys *ss, struct cgroup *cont) { |
c0ff4b854 memcg: rename mem... |
4763 |
struct mem_cgroup *memcg, *parent; |
04046e1a0 memcg: use CSS ID |
4764 |
long error = -ENOMEM; |
6d12e2d8d per-zone and recl... |
4765 |
int node; |
8cdea7c05 Memory controller... |
4766 |
|
c0ff4b854 memcg: rename mem... |
4767 4768 |
memcg = mem_cgroup_alloc(); if (!memcg) |
04046e1a0 memcg: use CSS ID |
4769 |
return ERR_PTR(error); |
78fb74669 Memory controller... |
4770 |
|
6d12e2d8d per-zone and recl... |
4771 |
for_each_node_state(node, N_POSSIBLE) |
c0ff4b854 memcg: rename mem... |
4772 |
if (alloc_mem_cgroup_per_zone_info(memcg, node)) |
6d12e2d8d per-zone and recl... |
4773 |
goto free_out; |
f64c3f549 memory controller... |
4774 |
|
c077719be memcg: mem+swap c... |
4775 |
/* root ? */ |
28dbc4b6a memcg: memory cgr... |
4776 |
if (cont->parent == NULL) { |
cdec2e426 memcg: coalesce c... |
4777 |
int cpu; |
c077719be memcg: mem+swap c... |
4778 |
enable_swap_cgroup(); |
28dbc4b6a memcg: memory cgr... |
4779 |
parent = NULL; |
f64c3f549 memory controller... |
4780 4781 |
if (mem_cgroup_soft_limit_tree_init()) goto free_out; |
a41c58a66 memcg: keep root ... |
4782 |
root_mem_cgroup = memcg; |
cdec2e426 memcg: coalesce c... |
4783 4784 4785 4786 4787 |
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... |
4788 |
hotcpu_notifier(memcg_cpu_hotplug_callback, 0); |
18f59ea7d memcg: memory cgr... |
4789 |
} else { |
28dbc4b6a memcg: memory cgr... |
4790 |
parent = mem_cgroup_from_cont(cont->parent); |
c0ff4b854 memcg: rename mem... |
4791 4792 |
memcg->use_hierarchy = parent->use_hierarchy; memcg->oom_kill_disable = parent->oom_kill_disable; |
18f59ea7d memcg: memory cgr... |
4793 |
} |
28dbc4b6a memcg: memory cgr... |
4794 |
|
18f59ea7d memcg: memory cgr... |
4795 |
if (parent && parent->use_hierarchy) { |
c0ff4b854 memcg: rename mem... |
4796 4797 |
res_counter_init(&memcg->res, &parent->res); res_counter_init(&memcg->memsw, &parent->memsw); |
7bcc1bb12 memcg: get/put pa... |
4798 4799 4800 4801 4802 4803 4804 |
/* * 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... |
4805 |
} else { |
c0ff4b854 memcg: rename mem... |
4806 4807 |
res_counter_init(&memcg->res, NULL); res_counter_init(&memcg->memsw, NULL); |
18f59ea7d memcg: memory cgr... |
4808 |
} |
c0ff4b854 memcg: rename mem... |
4809 4810 4811 |
memcg->last_scanned_child = 0; memcg->last_scanned_node = MAX_NUMNODES; INIT_LIST_HEAD(&memcg->oom_notify); |
6d61ef409 memcg: memory cgr... |
4812 |
|
a7885eb8a memcg: swappiness |
4813 |
if (parent) |
c0ff4b854 memcg: rename mem... |
4814 4815 4816 4817 4818 |
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... |
4819 |
free_out: |
c0ff4b854 memcg: rename mem... |
4820 |
__mem_cgroup_free(memcg); |
04046e1a0 memcg: use CSS ID |
4821 |
return ERR_PTR(error); |
8cdea7c05 Memory controller... |
4822 |
} |
ec64f5154 cgroup: fix frequ... |
4823 |
static int mem_cgroup_pre_destroy(struct cgroup_subsys *ss, |
df878fb04 memory cgroup enh... |
4824 4825 |
struct cgroup *cont) { |
c0ff4b854 memcg: rename mem... |
4826 |
struct mem_cgroup *memcg = mem_cgroup_from_cont(cont); |
ec64f5154 cgroup: fix frequ... |
4827 |
|
c0ff4b854 memcg: rename mem... |
4828 |
return mem_cgroup_force_empty(memcg, false); |
df878fb04 memory cgroup enh... |
4829 |
} |
8cdea7c05 Memory controller... |
4830 4831 4832 |
static void mem_cgroup_destroy(struct cgroup_subsys *ss, struct cgroup *cont) { |
c0ff4b854 memcg: rename mem... |
4833 |
struct mem_cgroup *memcg = mem_cgroup_from_cont(cont); |
c268e9946 memcg: fix hierar... |
4834 |
|
d1a4c0b37 tcp memory pressu... |
4835 |
kmem_cgroup_destroy(ss, cont); |
c0ff4b854 memcg: rename mem... |
4836 |
mem_cgroup_put(memcg); |
8cdea7c05 Memory controller... |
4837 4838 4839 4840 4841 |
} static int mem_cgroup_populate(struct cgroup_subsys *ss, struct cgroup *cont) { |
8c7c6e34a memcg: mem+swap c... |
4842 4843 4844 4845 4846 4847 4848 |
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... |
4849 4850 4851 |
if (!ret) ret = register_kmem_files(cont, ss); |
8c7c6e34a memcg: mem+swap c... |
4852 |
return ret; |
8cdea7c05 Memory controller... |
4853 |
} |
024914477 memcg: move charg... |
4854 |
#ifdef CONFIG_MMU |
7dc74be03 memcg: add interf... |
4855 |
/* Handlers for move charge at task migration. */ |
854ffa8d1 memcg: improve pe... |
4856 4857 |
#define PRECHARGE_COUNT_AT_ONCE 256 static int mem_cgroup_do_precharge(unsigned long count) |
7dc74be03 memcg: add interf... |
4858 |
{ |
854ffa8d1 memcg: improve pe... |
4859 4860 |
int ret = 0; int batch_count = PRECHARGE_COUNT_AT_ONCE; |
c0ff4b854 memcg: rename mem... |
4861 |
struct mem_cgroup *memcg = mc.to; |
4ffef5fef memcg: move charg... |
4862 |
|
c0ff4b854 memcg: rename mem... |
4863 |
if (mem_cgroup_is_root(memcg)) { |
854ffa8d1 memcg: improve pe... |
4864 4865 4866 4867 4868 4869 4870 4871 |
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... |
4872 |
* "memcg" cannot be under rmdir() because we've already checked |
854ffa8d1 memcg: improve pe... |
4873 4874 4875 4876 |
* 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... |
4877 |
if (res_counter_charge(&memcg->res, PAGE_SIZE * count, &dummy)) |
854ffa8d1 memcg: improve pe... |
4878 |
goto one_by_one; |
c0ff4b854 memcg: rename mem... |
4879 |
if (do_swap_account && res_counter_charge(&memcg->memsw, |
854ffa8d1 memcg: improve pe... |
4880 |
PAGE_SIZE * count, &dummy)) { |
c0ff4b854 memcg: rename mem... |
4881 |
res_counter_uncharge(&memcg->res, PAGE_SIZE * count); |
854ffa8d1 memcg: improve pe... |
4882 4883 4884 |
goto one_by_one; } mc.precharge += count; |
854ffa8d1 memcg: improve pe... |
4885 4886 4887 4888 4889 4890 4891 4892 4893 4894 4895 4896 4897 |
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... |
4898 4899 4900 |
ret = __mem_cgroup_try_charge(NULL, GFP_KERNEL, 1, &memcg, false); if (ret || !memcg) |
854ffa8d1 memcg: improve pe... |
4901 4902 4903 4904 |
/* mem_cgroup_clear_mc() will do uncharge later */ return -ENOMEM; mc.precharge++; } |
4ffef5fef memcg: move charg... |
4905 4906 4907 4908 4909 4910 4911 4912 |
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... |
4913 |
* @target: the pointer the target page or swap ent will be stored(can be NULL) |
4ffef5fef memcg: move charg... |
4914 4915 4916 4917 4918 4919 |
* * 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... |
4920 4921 4922 |
* 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... |
4923 4924 4925 |
* * Called with pte lock held. */ |
4ffef5fef memcg: move charg... |
4926 4927 |
union mc_target { struct page *page; |
024914477 memcg: move charg... |
4928 |
swp_entry_t ent; |
4ffef5fef memcg: move charg... |
4929 |
}; |
4ffef5fef memcg: move charg... |
4930 4931 4932 |
enum mc_target_type { MC_TARGET_NONE, /* not used */ MC_TARGET_PAGE, |
024914477 memcg: move charg... |
4933 |
MC_TARGET_SWAP, |
4ffef5fef memcg: move charg... |
4934 |
}; |
90254a658 memcg: clean up m... |
4935 4936 |
static struct page *mc_handle_present_pte(struct vm_area_struct *vma, unsigned long addr, pte_t ptent) |
4ffef5fef memcg: move charg... |
4937 |
{ |
90254a658 memcg: clean up m... |
4938 |
struct page *page = vm_normal_page(vma, addr, ptent); |
4ffef5fef memcg: move charg... |
4939 |
|
90254a658 memcg: clean up m... |
4940 4941 4942 4943 4944 4945 |
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... |
4946 4947 |
} else if (!move_file()) /* we ignore mapcount for file pages */ |
90254a658 memcg: clean up m... |
4948 4949 4950 4951 4952 4953 4954 4955 4956 4957 4958 4959 4960 4961 4962 4963 4964 4965 |
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... |
4966 4967 |
if (page) put_page(page); |
90254a658 memcg: clean up m... |
4968 |
return NULL; |
024914477 memcg: move charg... |
4969 |
} |
90254a658 memcg: clean up m... |
4970 4971 4972 4973 4974 |
if (do_swap_account) entry->val = ent.val; return page; } |
87946a722 memcg: move charg... |
4975 4976 4977 4978 4979 4980 4981 4982 4983 4984 4985 4986 4987 4988 4989 4990 4991 4992 4993 4994 4995 |
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... |
4996 4997 4998 4999 5000 5001 |
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... |
5002 |
if (do_swap_account) |
aa3b18955 tmpfs: convert me... |
5003 5004 |
*entry = swap; page = find_get_page(&swapper_space, swap.val); |
87946a722 memcg: move charg... |
5005 |
} |
aa3b18955 tmpfs: convert me... |
5006 |
#endif |
87946a722 memcg: move charg... |
5007 5008 |
return page; } |
90254a658 memcg: clean up m... |
5009 5010 5011 5012 5013 5014 5015 5016 5017 5018 5019 5020 |
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... |
5021 5022 |
else if (pte_none(ptent) || pte_file(ptent)) page = mc_handle_file_pte(vma, addr, ptent, &ent); |
90254a658 memcg: clean up m... |
5023 5024 5025 |
if (!page && !ent.val) return 0; |
024914477 memcg: move charg... |
5026 5027 5028 5029 5030 5031 5032 5033 5034 5035 5036 5037 5038 5039 5040 |
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... |
5041 5042 |
/* There is a swap entry and a page doesn't exist or isn't charged */ if (ent.val && !ret && |
7f0f15464 memcg: fix css_id... |
5043 5044 5045 5046 |
css_id(&mc.from->css) == lookup_swap_cgroup(ent)) { ret = MC_TARGET_SWAP; if (target) target->ent = ent; |
4ffef5fef memcg: move charg... |
5047 |
} |
4ffef5fef memcg: move charg... |
5048 5049 5050 5051 5052 5053 5054 5055 5056 5057 |
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... |
5058 |
split_huge_page_pmd(walk->mm, pmd); |
4ffef5fef memcg: move charg... |
5059 5060 5061 5062 5063 5064 |
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... |
5065 5066 |
return 0; } |
4ffef5fef memcg: move charg... |
5067 5068 5069 5070 |
static unsigned long mem_cgroup_count_precharge(struct mm_struct *mm) { unsigned long precharge; struct vm_area_struct *vma; |
dfe076b09 memcg: fix deadlo... |
5071 |
down_read(&mm->mmap_sem); |
4ffef5fef memcg: move charg... |
5072 5073 5074 5075 5076 5077 5078 5079 |
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... |
5080 5081 5082 |
walk_page_range(vma->vm_start, vma->vm_end, &mem_cgroup_count_precharge_walk); } |
dfe076b09 memcg: fix deadlo... |
5083 |
up_read(&mm->mmap_sem); |
4ffef5fef memcg: move charg... |
5084 5085 5086 5087 5088 5089 |
precharge = mc.precharge; mc.precharge = 0; return precharge; } |
4ffef5fef memcg: move charg... |
5090 5091 |
static int mem_cgroup_precharge_mc(struct mm_struct *mm) { |
dfe076b09 memcg: fix deadlo... |
5092 5093 5094 5095 5096 |
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... |
5097 |
} |
dfe076b09 memcg: fix deadlo... |
5098 5099 |
/* 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... |
5100 |
{ |
2bd9bb206 memcg: clean up w... |
5101 5102 |
struct mem_cgroup *from = mc.from; struct mem_cgroup *to = mc.to; |
4ffef5fef memcg: move charg... |
5103 |
/* we must uncharge all the leftover precharges from mc.to */ |
854ffa8d1 memcg: improve pe... |
5104 5105 5106 5107 5108 5109 5110 5111 5112 5113 5114 |
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... |
5115 |
} |
483c30b51 memcg: improve pe... |
5116 5117 |
/* we must fixup refcnts and charges */ if (mc.moved_swap) { |
483c30b51 memcg: improve pe... |
5118 5119 5120 5121 5122 5123 5124 5125 5126 5127 5128 5129 5130 |
/* 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... |
5131 5132 |
} /* we've already done mem_cgroup_get(mc.to) */ |
483c30b51 memcg: improve pe... |
5133 5134 |
mc.moved_swap = 0; } |
dfe076b09 memcg: fix deadlo... |
5135 5136 5137 5138 5139 5140 5141 5142 5143 5144 5145 5146 5147 5148 5149 |
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... |
5150 |
spin_lock(&mc.lock); |
4ffef5fef memcg: move charg... |
5151 5152 |
mc.from = NULL; mc.to = NULL; |
2bd9bb206 memcg: clean up w... |
5153 |
spin_unlock(&mc.lock); |
32047e2a8 memcg: avoid lock... |
5154 |
mem_cgroup_end_move(from); |
4ffef5fef memcg: move charg... |
5155 |
} |
7dc74be03 memcg: add interf... |
5156 5157 |
static int mem_cgroup_can_attach(struct cgroup_subsys *ss, struct cgroup *cgroup, |
2f7ee5691 cgroup: introduce... |
5158 |
struct cgroup_taskset *tset) |
7dc74be03 memcg: add interf... |
5159 |
{ |
2f7ee5691 cgroup: introduce... |
5160 |
struct task_struct *p = cgroup_taskset_first(tset); |
7dc74be03 memcg: add interf... |
5161 |
int ret = 0; |
c0ff4b854 memcg: rename mem... |
5162 |
struct mem_cgroup *memcg = mem_cgroup_from_cont(cgroup); |
7dc74be03 memcg: add interf... |
5163 |
|
c0ff4b854 memcg: rename mem... |
5164 |
if (memcg->move_charge_at_immigrate) { |
7dc74be03 memcg: add interf... |
5165 5166 |
struct mm_struct *mm; struct mem_cgroup *from = mem_cgroup_from_task(p); |
c0ff4b854 memcg: rename mem... |
5167 |
VM_BUG_ON(from == memcg); |
7dc74be03 memcg: add interf... |
5168 5169 5170 5171 |
mm = get_task_mm(p); if (!mm) return 0; |
7dc74be03 memcg: add interf... |
5172 |
/* We move charges only when we move a owner of the mm */ |
4ffef5fef memcg: move charg... |
5173 5174 5175 5176 |
if (mm->owner == p) { VM_BUG_ON(mc.from); VM_BUG_ON(mc.to); VM_BUG_ON(mc.precharge); |
854ffa8d1 memcg: improve pe... |
5177 |
VM_BUG_ON(mc.moved_charge); |
483c30b51 memcg: improve pe... |
5178 |
VM_BUG_ON(mc.moved_swap); |
32047e2a8 memcg: avoid lock... |
5179 |
mem_cgroup_start_move(from); |
2bd9bb206 memcg: clean up w... |
5180 |
spin_lock(&mc.lock); |
4ffef5fef memcg: move charg... |
5181 |
mc.from = from; |
c0ff4b854 memcg: rename mem... |
5182 |
mc.to = memcg; |
2bd9bb206 memcg: clean up w... |
5183 |
spin_unlock(&mc.lock); |
dfe076b09 memcg: fix deadlo... |
5184 |
/* We set mc.moving_task later */ |
4ffef5fef memcg: move charg... |
5185 5186 5187 5188 |
ret = mem_cgroup_precharge_mc(mm); if (ret) mem_cgroup_clear_mc(); |
dfe076b09 memcg: fix deadlo... |
5189 5190 |
} mmput(mm); |
7dc74be03 memcg: add interf... |
5191 5192 5193 5194 5195 5196 |
} return ret; } static void mem_cgroup_cancel_attach(struct cgroup_subsys *ss, struct cgroup *cgroup, |
2f7ee5691 cgroup: introduce... |
5197 |
struct cgroup_taskset *tset) |
7dc74be03 memcg: add interf... |
5198 |
{ |
4ffef5fef memcg: move charg... |
5199 |
mem_cgroup_clear_mc(); |
7dc74be03 memcg: add interf... |
5200 |
} |
4ffef5fef memcg: move charg... |
5201 5202 5203 |
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... |
5204 |
{ |
4ffef5fef memcg: move charg... |
5205 5206 5207 5208 |
int ret = 0; struct vm_area_struct *vma = walk->private; pte_t *pte; spinlock_t *ptl; |
033193275 pagewalk: only sp... |
5209 |
split_huge_page_pmd(walk->mm, pmd); |
4ffef5fef memcg: move charg... |
5210 5211 5212 5213 5214 5215 5216 5217 |
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... |
5218 |
swp_entry_t ent; |
4ffef5fef memcg: move charg... |
5219 5220 5221 5222 5223 5224 5225 5226 5227 5228 5229 |
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... |
5230 5231 |
if (!mem_cgroup_move_account(page, 1, pc, mc.from, mc.to, false)) { |
4ffef5fef memcg: move charg... |
5232 |
mc.precharge--; |
854ffa8d1 memcg: improve pe... |
5233 5234 |
/* we uncharge from mc.from later. */ mc.moved_charge++; |
4ffef5fef memcg: move charg... |
5235 5236 5237 5238 5239 |
} putback_lru_page(page); put: /* is_target_pte_for_mc() gets the page */ put_page(page); break; |
024914477 memcg: move charg... |
5240 5241 |
case MC_TARGET_SWAP: ent = target.ent; |
483c30b51 memcg: improve pe... |
5242 5243 |
if (!mem_cgroup_move_swap_account(ent, mc.from, mc.to, false)) { |
024914477 memcg: move charg... |
5244 |
mc.precharge--; |
483c30b51 memcg: improve pe... |
5245 5246 5247 |
/* we fixup refcnts and charges later. */ mc.moved_swap++; } |
024914477 memcg: move charg... |
5248 |
break; |
4ffef5fef memcg: move charg... |
5249 5250 5251 5252 5253 5254 5255 5256 5257 5258 5259 5260 5261 5262 |
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... |
5263 |
ret = mem_cgroup_do_precharge(1); |
4ffef5fef memcg: move charg... |
5264 5265 5266 5267 5268 5269 5270 5271 5272 5273 5274 5275 |
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... |
5276 5277 5278 5279 5280 5281 5282 5283 5284 5285 5286 5287 5288 |
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... |
5289 5290 5291 5292 5293 5294 5295 5296 5297 |
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... |
5298 5299 5300 5301 5302 5303 5304 5305 5306 |
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... |
5307 |
up_read(&mm->mmap_sem); |
7dc74be03 memcg: add interf... |
5308 |
} |
67e465a77 Memory controller... |
5309 5310 |
static void mem_cgroup_move_task(struct cgroup_subsys *ss, struct cgroup *cont, |
2f7ee5691 cgroup: introduce... |
5311 |
struct cgroup_taskset *tset) |
67e465a77 Memory controller... |
5312 |
{ |
2f7ee5691 cgroup: introduce... |
5313 |
struct task_struct *p = cgroup_taskset_first(tset); |
a433658c3 vmscan,memcg: mem... |
5314 |
struct mm_struct *mm = get_task_mm(p); |
dfe076b09 memcg: fix deadlo... |
5315 |
|
dfe076b09 memcg: fix deadlo... |
5316 |
if (mm) { |
a433658c3 vmscan,memcg: mem... |
5317 5318 5319 |
if (mc.to) mem_cgroup_move_charge(mm); put_swap_token(mm); |
dfe076b09 memcg: fix deadlo... |
5320 5321 |
mmput(mm); } |
a433658c3 vmscan,memcg: mem... |
5322 5323 |
if (mc.to) mem_cgroup_clear_mc(); |
67e465a77 Memory controller... |
5324 |
} |
5cfb80a73 memcg: disable mo... |
5325 5326 5327 |
#else /* !CONFIG_MMU */ static int mem_cgroup_can_attach(struct cgroup_subsys *ss, struct cgroup *cgroup, |
2f7ee5691 cgroup: introduce... |
5328 |
struct cgroup_taskset *tset) |
5cfb80a73 memcg: disable mo... |
5329 5330 5331 5332 5333 |
{ return 0; } static void mem_cgroup_cancel_attach(struct cgroup_subsys *ss, struct cgroup *cgroup, |
2f7ee5691 cgroup: introduce... |
5334 |
struct cgroup_taskset *tset) |
5cfb80a73 memcg: disable mo... |
5335 5336 5337 5338 |
{ } static void mem_cgroup_move_task(struct cgroup_subsys *ss, struct cgroup *cont, |
2f7ee5691 cgroup: introduce... |
5339 |
struct cgroup_taskset *tset) |
5cfb80a73 memcg: disable mo... |
5340 5341 5342 |
{ } #endif |
67e465a77 Memory controller... |
5343 |
|
8cdea7c05 Memory controller... |
5344 5345 5346 5347 |
struct cgroup_subsys mem_cgroup_subsys = { .name = "memory", .subsys_id = mem_cgroup_subsys_id, .create = mem_cgroup_create, |
df878fb04 memory cgroup enh... |
5348 |
.pre_destroy = mem_cgroup_pre_destroy, |
8cdea7c05 Memory controller... |
5349 5350 |
.destroy = mem_cgroup_destroy, .populate = mem_cgroup_populate, |
7dc74be03 memcg: add interf... |
5351 5352 |
.can_attach = mem_cgroup_can_attach, .cancel_attach = mem_cgroup_cancel_attach, |
67e465a77 Memory controller... |
5353 |
.attach = mem_cgroup_move_task, |
6d12e2d8d per-zone and recl... |
5354 |
.early_init = 0, |
04046e1a0 memcg: use CSS ID |
5355 |
.use_id = 1, |
8cdea7c05 Memory controller... |
5356 |
}; |
c077719be memcg: mem+swap c... |
5357 5358 |
#ifdef CONFIG_CGROUP_MEM_RES_CTLR_SWAP |
a42c390cf cgroups: make swa... |
5359 5360 5361 |
static int __init enable_swap_account(char *s) { /* consider enabled if no parameter or 1 is given */ |
a2c8990ae memsw: remove nos... |
5362 |
if (!strcmp(s, "1")) |
a42c390cf cgroups: make swa... |
5363 |
really_do_swap_account = 1; |
a2c8990ae memsw: remove nos... |
5364 |
else if (!strcmp(s, "0")) |
a42c390cf cgroups: make swa... |
5365 5366 5367 |
really_do_swap_account = 0; return 1; } |
a2c8990ae memsw: remove nos... |
5368 |
__setup("swapaccount=", enable_swap_account); |
c077719be memcg: mem+swap c... |
5369 |
|
c077719be memcg: mem+swap c... |
5370 |
#endif |