Blame view
mm/memcontrol.c
151 KB
8cdea7c05 Memory controller... |
1 2 3 4 5 |
/* memcontrol.c - Memory Controller * * Copyright IBM Corporation, 2007 * Author Balbir Singh <balbir@linux.vnet.ibm.com> * |
78fb74669 Memory controller... |
6 7 8 |
* Copyright 2007 OpenVZ SWsoft Inc * Author: Pavel Emelianov <xemul@openvz.org> * |
2e72b6347 memcg: implement ... |
9 10 11 12 |
* Memory thresholds * Copyright (C) 2009 Nokia Corporation * Author: Kirill A. Shutemov * |
7ae1e1d0f memcg: kmem contr... |
13 14 15 16 |
* Kernel Memory Controller * Copyright (C) 2012 Parallels Inc. and Google Inc. * Authors: Glauber Costa and Suleiman Souhlal * |
1575e68b3 mm: memcontrol: u... |
17 18 19 20 21 22 |
* Native page reclaim * Charge lifetime sanitation * Lockless page tracking & accounting * Unified hierarchy configuration model * Copyright (C) 2015 Red Hat, Inc., Johannes Weiner * |
8cdea7c05 Memory controller... |
23 24 25 26 27 28 29 30 31 32 |
* This program is free software; you can redistribute it and/or modify * it under the terms of the GNU General Public License as published by * the Free Software Foundation; either version 2 of the License, or * (at your option) any later version. * * This program is distributed in the hope that it will be useful, * but WITHOUT ANY WARRANTY; without even the implied warranty of * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the * GNU General Public License for more details. */ |
3e32cb2e0 mm: memcontrol: l... |
33 |
#include <linux/page_counter.h> |
8cdea7c05 Memory controller... |
34 35 |
#include <linux/memcontrol.h> #include <linux/cgroup.h> |
78fb74669 Memory controller... |
36 |
#include <linux/mm.h> |
4ffef5fef memcg: move charg... |
37 |
#include <linux/hugetlb.h> |
d13d14430 memcg: handle swa... |
38 |
#include <linux/pagemap.h> |
d52aa412d memory cgroup enh... |
39 |
#include <linux/smp.h> |
8a9f3ccd2 Memory controller... |
40 |
#include <linux/page-flags.h> |
66e1707bc Memory controller... |
41 |
#include <linux/backing-dev.h> |
8a9f3ccd2 Memory controller... |
42 43 |
#include <linux/bit_spinlock.h> #include <linux/rcupdate.h> |
e222432bf memcg: show memcg... |
44 |
#include <linux/limits.h> |
b9e15bafd mm: Add export.h ... |
45 |
#include <linux/export.h> |
8c7c6e34a memcg: mem+swap c... |
46 |
#include <linux/mutex.h> |
bb4cc1a8b revert "memcg: ge... |
47 |
#include <linux/rbtree.h> |
b6ac57d50 memcgroup: move m... |
48 |
#include <linux/slab.h> |
66e1707bc Memory controller... |
49 |
#include <linux/swap.h> |
024914477 memcg: move charg... |
50 |
#include <linux/swapops.h> |
66e1707bc Memory controller... |
51 |
#include <linux/spinlock.h> |
2e72b6347 memcg: implement ... |
52 |
#include <linux/eventfd.h> |
79bd9814e cgroup, memcg: mo... |
53 |
#include <linux/poll.h> |
2e72b6347 memcg: implement ... |
54 |
#include <linux/sort.h> |
66e1707bc Memory controller... |
55 |
#include <linux/fs.h> |
d2ceb9b7d memory cgroup enh... |
56 |
#include <linux/seq_file.h> |
70ddf637e memcg: add memory... |
57 |
#include <linux/vmpressure.h> |
b69408e88 vmscan: Use an in... |
58 |
#include <linux/mm_inline.h> |
5d1ea48bd mm: page_cgroup: ... |
59 |
#include <linux/swap_cgroup.h> |
cdec2e426 memcg: coalesce c... |
60 |
#include <linux/cpu.h> |
158e0a2d1 memcg: use find_l... |
61 |
#include <linux/oom.h> |
0056f4e66 mm: memcg: lockde... |
62 |
#include <linux/lockdep.h> |
79bd9814e cgroup, memcg: mo... |
63 |
#include <linux/file.h> |
b23afb93d memcg: punt high ... |
64 |
#include <linux/tracehook.h> |
08e552c69 memcg: synchroniz... |
65 |
#include "internal.h" |
d1a4c0b37 tcp memory pressu... |
66 |
#include <net/sock.h> |
4bd2c1ee4 memcg: cleanup km... |
67 |
#include <net/ip.h> |
f35c3a8ee memcg, kmem: use ... |
68 |
#include "slab.h" |
8cdea7c05 Memory controller... |
69 |
|
8697d3319 Memory controller... |
70 |
#include <asm/uaccess.h> |
cc8e970c3 memcg: add mm_vms... |
71 |
#include <trace/events/vmscan.h> |
073219e99 cgroup: clean up ... |
72 73 |
struct cgroup_subsys memory_cgrp_subsys __read_mostly; EXPORT_SYMBOL(memory_cgrp_subsys); |
68ae564bb mm, memcg: avoid ... |
74 |
|
7d828602e mm: memcontrol: e... |
75 |
struct mem_cgroup *root_mem_cgroup __read_mostly; |
a181b0e88 memcg: make globa... |
76 |
#define MEM_CGROUP_RECLAIM_RETRIES 5 |
8cdea7c05 Memory controller... |
77 |
|
f7e1cb6ec mm: memcontrol: a... |
78 79 |
/* Socket memory accounting disabled? */ static bool cgroup_memory_nosocket; |
04823c833 mm: memcontrol: a... |
80 81 |
/* Kernel memory accounting disabled? */ static bool cgroup_memory_nokmem; |
21afa38ee mm: memcontrol: c... |
82 |
/* Whether the swap controller is active */ |
c255a4580 memcg: rename con... |
83 |
#ifdef CONFIG_MEMCG_SWAP |
c077719be memcg: mem+swap c... |
84 |
int do_swap_account __read_mostly; |
c077719be memcg: mem+swap c... |
85 |
#else |
a0db00fcf memcg: remove red... |
86 |
#define do_swap_account 0 |
c077719be memcg: mem+swap c... |
87 |
#endif |
7941d2145 mm: memcontrol: d... |
88 89 90 91 92 |
/* Whether legacy memory+swap accounting is active */ static bool do_memsw_account(void) { return !cgroup_subsys_on_dfl(memory_cgrp_subsys) && do_swap_account; } |
af7c4b0ec mm: memcg: print ... |
93 94 95 |
static const char * const mem_cgroup_stat_names[] = { "cache", "rss", |
b070e65c0 mm, memcg: add rs... |
96 |
"rss_huge", |
af7c4b0ec mm: memcg: print ... |
97 |
"mapped_file", |
c4843a759 memcg: add per cg... |
98 |
"dirty", |
3ea67d06e memcg: add per cg... |
99 |
"writeback", |
af7c4b0ec mm: memcg: print ... |
100 101 |
"swap", }; |
af7c4b0ec mm: memcg: print ... |
102 103 104 105 106 107 |
static const char * const mem_cgroup_events_names[] = { "pgpgin", "pgpgout", "pgfault", "pgmajfault", }; |
58cf188ed memcg, oom: provi... |
108 109 110 111 112 113 114 |
static const char * const mem_cgroup_lru_names[] = { "inactive_anon", "active_anon", "inactive_file", "active_file", "unevictable", }; |
a0db00fcf memcg: remove red... |
115 116 117 |
#define THRESHOLDS_EVENTS_TARGET 128 #define SOFTLIMIT_EVENTS_TARGET 1024 #define NUMAINFO_EVENTS_TARGET 1024 |
e9f8974f2 memcg: break out ... |
118 |
|
bb4cc1a8b revert "memcg: ge... |
119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 |
/* * 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; |
9490ff275 memcg: oom notifier |
138 139 140 141 142 |
/* for OOM */ struct mem_cgroup_eventfd_list { struct list_head list; struct eventfd_ctx *eventfd; }; |
2e72b6347 memcg: implement ... |
143 |
|
79bd9814e cgroup, memcg: mo... |
144 145 146 |
/* * cgroup_event represents events which userspace want to receive. */ |
3bc942f37 memcg: rename cgr... |
147 |
struct mem_cgroup_event { |
79bd9814e cgroup, memcg: mo... |
148 |
/* |
59b6f8734 memcg: make cgrou... |
149 |
* memcg which the event belongs to. |
79bd9814e cgroup, memcg: mo... |
150 |
*/ |
59b6f8734 memcg: make cgrou... |
151 |
struct mem_cgroup *memcg; |
79bd9814e cgroup, memcg: mo... |
152 |
/* |
79bd9814e cgroup, memcg: mo... |
153 154 155 156 157 158 159 160 |
* eventfd to signal userspace about the event. */ struct eventfd_ctx *eventfd; /* * Each of these stored in a list by the cgroup. */ struct list_head list; /* |
fba948078 cgroup, memcg: mo... |
161 162 163 164 |
* register_event() callback will be used to add new userspace * waiter for changes related to this event. Use eventfd_signal() * on eventfd to send notification to userspace. */ |
59b6f8734 memcg: make cgrou... |
165 |
int (*register_event)(struct mem_cgroup *memcg, |
347c4a874 memcg: remove cgr... |
166 |
struct eventfd_ctx *eventfd, const char *args); |
fba948078 cgroup, memcg: mo... |
167 168 169 170 171 |
/* * unregister_event() callback will be called when userspace closes * the eventfd or on cgroup removing. This callback must be set, * if you want provide notification functionality. */ |
59b6f8734 memcg: make cgrou... |
172 |
void (*unregister_event)(struct mem_cgroup *memcg, |
fba948078 cgroup, memcg: mo... |
173 174 |
struct eventfd_ctx *eventfd); /* |
79bd9814e cgroup, memcg: mo... |
175 176 177 178 179 180 181 182 |
* All fields below needed to unregister event when * userspace closes eventfd. */ poll_table pt; wait_queue_head_t *wqh; wait_queue_t wait; struct work_struct remove; }; |
c0ff4b854 memcg: rename mem... |
183 184 |
static void mem_cgroup_threshold(struct mem_cgroup *memcg); static void mem_cgroup_oom_notify(struct mem_cgroup *memcg); |
2e72b6347 memcg: implement ... |
185 |
|
7dc74be03 memcg: add interf... |
186 187 |
/* Stuffs for move charges at task migration. */ /* |
1dfab5abc mm: memcontrol: f... |
188 |
* Types of charges to be moved. |
7dc74be03 memcg: add interf... |
189 |
*/ |
1dfab5abc mm: memcontrol: f... |
190 191 192 |
#define MOVE_ANON 0x1U #define MOVE_FILE 0x2U #define MOVE_MASK (MOVE_ANON | MOVE_FILE) |
7dc74be03 memcg: add interf... |
193 |
|
4ffef5fef memcg: move charg... |
194 195 |
/* "mc" and its members are protected by cgroup_mutex */ static struct move_charge_struct { |
b1dd693e5 memcg: avoid dead... |
196 |
spinlock_t lock; /* for from, to */ |
264a0ae16 memcg: relocate c... |
197 |
struct mm_struct *mm; |
4ffef5fef memcg: move charg... |
198 199 |
struct mem_cgroup *from; struct mem_cgroup *to; |
1dfab5abc mm: memcontrol: f... |
200 |
unsigned long flags; |
4ffef5fef memcg: move charg... |
201 |
unsigned long precharge; |
854ffa8d1 memcg: improve pe... |
202 |
unsigned long moved_charge; |
483c30b51 memcg: improve pe... |
203 |
unsigned long moved_swap; |
8033b97c9 memcg: avoid oom ... |
204 205 206 |
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... |
207 |
.lock = __SPIN_LOCK_UNLOCKED(mc.lock), |
8033b97c9 memcg: avoid oom ... |
208 209 |
.waitq = __WAIT_QUEUE_HEAD_INITIALIZER(mc.waitq), }; |
4ffef5fef memcg: move charg... |
210 |
|
4e4169535 memory controller... |
211 212 213 214 |
/* * Maximum loops in mem_cgroup_hierarchical_reclaim(), used for soft * limit reclaim to prevent infinite loops, if they ever occur. */ |
a0db00fcf memcg: remove red... |
215 |
#define MEM_CGROUP_MAX_RECLAIM_LOOPS 100 |
bb4cc1a8b revert "memcg: ge... |
216 |
#define MEM_CGROUP_MAX_SOFT_LIMIT_RECLAIM_LOOPS 2 |
4e4169535 memory controller... |
217 |
|
217bc3194 memory cgroup enh... |
218 219 |
enum charge_type { MEM_CGROUP_CHARGE_TYPE_CACHE = 0, |
41326c17f memcg: rename MEM... |
220 |
MEM_CGROUP_CHARGE_TYPE_ANON, |
d13d14430 memcg: handle swa... |
221 |
MEM_CGROUP_CHARGE_TYPE_SWAPOUT, /* for accounting swapcache */ |
8a9478ca7 memcg: fix swap a... |
222 |
MEM_CGROUP_CHARGE_TYPE_DROP, /* a page was unused swap cache */ |
c05555b57 memcg: atomic ops... |
223 224 |
NR_CHARGE_TYPE, }; |
8c7c6e34a memcg: mem+swap c... |
225 |
/* for encoding cft->private value on file */ |
86ae53e1a memcg: change def... |
226 227 228 229 |
enum res_type { _MEM, _MEMSWAP, _OOM_TYPE, |
510fc4e11 memcg: kmem accou... |
230 |
_KMEM, |
d55f90bfa net: drop tcp_mem... |
231 |
_TCP, |
86ae53e1a memcg: change def... |
232 |
}; |
a0db00fcf memcg: remove red... |
233 234 |
#define MEMFILE_PRIVATE(x, val) ((x) << 16 | (val)) #define MEMFILE_TYPE(val) ((val) >> 16 & 0xffff) |
8c7c6e34a memcg: mem+swap c... |
235 |
#define MEMFILE_ATTR(val) ((val) & 0xffff) |
9490ff275 memcg: oom notifier |
236 237 |
/* Used for OOM nofiier */ #define OOM_CONTROL (0) |
8c7c6e34a memcg: mem+swap c... |
238 |
|
70ddf637e memcg: add memory... |
239 240 241 242 243 244 245 246 247 248 249 250 |
/* Some nice accessors for the vmpressure. */ struct vmpressure *memcg_to_vmpressure(struct mem_cgroup *memcg) { if (!memcg) memcg = root_mem_cgroup; return &memcg->vmpressure; } struct cgroup_subsys_state *vmpressure_to_css(struct vmpressure *vmpr) { return &container_of(vmpr, struct mem_cgroup, vmpressure)->css; } |
7ffc0edc4 memcg: move mem_c... |
251 252 253 254 |
static inline bool mem_cgroup_is_root(struct mem_cgroup *memcg) { return (memcg == root_mem_cgroup); } |
127424c86 mm: memcontrol: m... |
255 |
#ifndef CONFIG_SLOB |
55007d849 memcg: allocate m... |
256 |
/* |
f7ce3190c slab: embed memcg... |
257 |
* This will be the memcg's index in each cache's ->memcg_params.memcg_caches. |
b86278359 memcg: stop using... |
258 259 260 261 262 |
* The main reason for not using cgroup id for this: * this works better in sparse environments, where we have a lot of memcgs, * but only a few kmem-limited. Or also, if we have, for instance, 200 * memcgs, and none but the 200th is kmem-limited, we'd have to have a * 200 entry array for that. |
55007d849 memcg: allocate m... |
263 |
* |
dbcf73e26 memcg: rename som... |
264 265 |
* The current size of the caches array is stored in memcg_nr_cache_ids. It * will double each time we have to increase it. |
55007d849 memcg: allocate m... |
266 |
*/ |
dbcf73e26 memcg: rename som... |
267 268 |
static DEFINE_IDA(memcg_cache_ida); int memcg_nr_cache_ids; |
749c54151 memcg: aggregate ... |
269 |
|
05257a1a3 memcg: add rwsem ... |
270 271 272 273 274 275 276 277 278 279 280 281 |
/* Protects memcg_nr_cache_ids */ static DECLARE_RWSEM(memcg_cache_ids_sem); void memcg_get_cache_ids(void) { down_read(&memcg_cache_ids_sem); } void memcg_put_cache_ids(void) { up_read(&memcg_cache_ids_sem); } |
55007d849 memcg: allocate m... |
282 283 284 285 286 287 |
/* * MIN_SIZE is different than 1, because we would like to avoid going through * the alloc/free process all the time. In a small machine, 4 kmem-limited * cgroups is a reasonable guess. In the future, it could be a parameter or * tunable, but that is strictly not necessary. * |
b86278359 memcg: stop using... |
288 |
* MAX_SIZE should be as large as the number of cgrp_ids. Ideally, we could get |
55007d849 memcg: allocate m... |
289 290 |
* this constant directly from cgroup, but it is understandable that this is * better kept as an internal representation in cgroup.c. In any case, the |
b86278359 memcg: stop using... |
291 |
* cgrp_id space is not getting any smaller, and we don't have to necessarily |
55007d849 memcg: allocate m... |
292 293 294 |
* increase ours as well if it increases. */ #define MEMCG_CACHES_MIN_SIZE 4 |
b86278359 memcg: stop using... |
295 |
#define MEMCG_CACHES_MAX_SIZE MEM_CGROUP_ID_MAX |
55007d849 memcg: allocate m... |
296 |
|
d7f25f8a2 memcg: infrastruc... |
297 298 299 300 301 302 |
/* * A lot of the calls to the cache allocation functions are expected to be * inlined by the compiler. Since the calls to memcg_kmem_get_cache are * conditional to this static branch, we'll have to allow modules that does * kmem_cache_alloc and the such to see this symbol as well */ |
ef12947c9 mm: memcontrol: s... |
303 |
DEFINE_STATIC_KEY_FALSE(memcg_kmem_enabled_key); |
d7f25f8a2 memcg: infrastruc... |
304 |
EXPORT_SYMBOL(memcg_kmem_enabled_key); |
a8964b9b8 memcg: use static... |
305 |
|
127424c86 mm: memcontrol: m... |
306 |
#endif /* !CONFIG_SLOB */ |
a8964b9b8 memcg: use static... |
307 |
|
f64c3f549 memory controller... |
308 |
static struct mem_cgroup_per_zone * |
e231875ba mm: memcontrol: c... |
309 |
mem_cgroup_zone_zoneinfo(struct mem_cgroup *memcg, struct zone *zone) |
f64c3f549 memory controller... |
310 |
{ |
e231875ba mm: memcontrol: c... |
311 312 |
int nid = zone_to_nid(zone); int zid = zone_idx(zone); |
54f72fe02 memcg: clean up m... |
313 |
return &memcg->nodeinfo[nid]->zoneinfo[zid]; |
f64c3f549 memory controller... |
314 |
} |
ad7fa852d memcg: implement ... |
315 316 317 318 319 320 321 322 323 324 |
/** * mem_cgroup_css_from_page - css of the memcg associated with a page * @page: page of interest * * If memcg is bound to the default hierarchy, css of the memcg associated * with @page is returned. The returned css remains associated with @page * until it is released. * * If memcg is bound to a traditional hierarchy, the css of root_mem_cgroup * is returned. |
ad7fa852d memcg: implement ... |
325 326 327 328 |
*/ struct cgroup_subsys_state *mem_cgroup_css_from_page(struct page *page) { struct mem_cgroup *memcg; |
ad7fa852d memcg: implement ... |
329 |
memcg = page->mem_cgroup; |
9e10a130d cgroup: replace c... |
330 |
if (!memcg || !cgroup_subsys_on_dfl(memory_cgrp_subsys)) |
ad7fa852d memcg: implement ... |
331 |
memcg = root_mem_cgroup; |
ad7fa852d memcg: implement ... |
332 333 |
return &memcg->css; } |
2fc045247 memcg: add page_c... |
334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 |
/** * page_cgroup_ino - return inode number of the memcg a page is charged to * @page: the page * * Look up the closest online ancestor of the memory cgroup @page is charged to * and return its inode number or 0 if @page is not charged to any cgroup. It * is safe to call this function without holding a reference to @page. * * Note, this function is inherently racy, because there is nothing to prevent * the cgroup inode from getting torn down and potentially reallocated a moment * after page_cgroup_ino() returns, so it only should be used by callers that * do not care (such as procfs interfaces). */ ino_t page_cgroup_ino(struct page *page) { struct mem_cgroup *memcg; unsigned long ino = 0; rcu_read_lock(); memcg = READ_ONCE(page->mem_cgroup); while (memcg && !(memcg->css.flags & CSS_ONLINE)) memcg = parent_mem_cgroup(memcg); if (memcg) ino = cgroup_ino(memcg->css.cgroup); rcu_read_unlock(); return ino; } |
f64c3f549 memory controller... |
361 |
static struct mem_cgroup_per_zone * |
e231875ba mm: memcontrol: c... |
362 |
mem_cgroup_page_zoneinfo(struct mem_cgroup *memcg, struct page *page) |
f64c3f549 memory controller... |
363 |
{ |
97a6c37b3 memcg: change pag... |
364 365 |
int nid = page_to_nid(page); int zid = page_zonenum(page); |
f64c3f549 memory controller... |
366 |
|
e231875ba mm: memcontrol: c... |
367 |
return &memcg->nodeinfo[nid]->zoneinfo[zid]; |
f64c3f549 memory controller... |
368 |
} |
bb4cc1a8b revert "memcg: ge... |
369 370 371 372 373 374 375 376 377 378 379 380 381 382 |
static struct mem_cgroup_tree_per_zone * soft_limit_tree_node_zone(int nid, int zid) { return &soft_limit_tree.rb_tree_per_node[nid]->rb_tree_per_zone[zid]; } static struct mem_cgroup_tree_per_zone * soft_limit_tree_from_page(struct page *page) { int nid = page_to_nid(page); int zid = page_zonenum(page); return &soft_limit_tree.rb_tree_per_node[nid]->rb_tree_per_zone[zid]; } |
cf2c81279 mm: memcontrol: r... |
383 384 |
static void __mem_cgroup_insert_exceeded(struct mem_cgroup_per_zone *mz, struct mem_cgroup_tree_per_zone *mctz, |
3e32cb2e0 mm: memcontrol: l... |
385 |
unsigned long new_usage_in_excess) |
bb4cc1a8b revert "memcg: ge... |
386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 |
{ struct rb_node **p = &mctz->rb_root.rb_node; struct rb_node *parent = NULL; struct mem_cgroup_per_zone *mz_node; if (mz->on_tree) return; mz->usage_in_excess = new_usage_in_excess; if (!mz->usage_in_excess) return; while (*p) { parent = *p; mz_node = rb_entry(parent, struct mem_cgroup_per_zone, tree_node); if (mz->usage_in_excess < mz_node->usage_in_excess) p = &(*p)->rb_left; /* * We can't avoid mem cgroups that are over their soft * limit by the same amount */ else if (mz->usage_in_excess >= mz_node->usage_in_excess) p = &(*p)->rb_right; } rb_link_node(&mz->tree_node, parent, p); rb_insert_color(&mz->tree_node, &mctz->rb_root); mz->on_tree = true; } |
cf2c81279 mm: memcontrol: r... |
414 415 |
static void __mem_cgroup_remove_exceeded(struct mem_cgroup_per_zone *mz, struct mem_cgroup_tree_per_zone *mctz) |
bb4cc1a8b revert "memcg: ge... |
416 417 418 419 420 421 |
{ if (!mz->on_tree) return; rb_erase(&mz->tree_node, &mctz->rb_root); mz->on_tree = false; } |
cf2c81279 mm: memcontrol: r... |
422 423 |
static void mem_cgroup_remove_exceeded(struct mem_cgroup_per_zone *mz, struct mem_cgroup_tree_per_zone *mctz) |
bb4cc1a8b revert "memcg: ge... |
424 |
{ |
0a31bc97c mm: memcontrol: r... |
425 426 427 |
unsigned long flags; spin_lock_irqsave(&mctz->lock, flags); |
cf2c81279 mm: memcontrol: r... |
428 |
__mem_cgroup_remove_exceeded(mz, mctz); |
0a31bc97c mm: memcontrol: r... |
429 |
spin_unlock_irqrestore(&mctz->lock, flags); |
bb4cc1a8b revert "memcg: ge... |
430 |
} |
3e32cb2e0 mm: memcontrol: l... |
431 432 433 |
static unsigned long soft_limit_excess(struct mem_cgroup *memcg) { unsigned long nr_pages = page_counter_read(&memcg->memory); |
4db0c3c29 mm: remove rest o... |
434 |
unsigned long soft_limit = READ_ONCE(memcg->soft_limit); |
3e32cb2e0 mm: memcontrol: l... |
435 436 437 438 439 440 441 |
unsigned long excess = 0; if (nr_pages > soft_limit) excess = nr_pages - soft_limit; return excess; } |
bb4cc1a8b revert "memcg: ge... |
442 443 444 |
static void mem_cgroup_update_tree(struct mem_cgroup *memcg, struct page *page) { |
3e32cb2e0 mm: memcontrol: l... |
445 |
unsigned long excess; |
bb4cc1a8b revert "memcg: ge... |
446 447 |
struct mem_cgroup_per_zone *mz; struct mem_cgroup_tree_per_zone *mctz; |
bb4cc1a8b revert "memcg: ge... |
448 |
|
e231875ba mm: memcontrol: c... |
449 |
mctz = soft_limit_tree_from_page(page); |
bb4cc1a8b revert "memcg: ge... |
450 451 452 453 454 |
/* * Necessary to update all ancestors when hierarchy is used. * because their event counter is not touched. */ for (; memcg; memcg = parent_mem_cgroup(memcg)) { |
e231875ba mm: memcontrol: c... |
455 |
mz = mem_cgroup_page_zoneinfo(memcg, page); |
3e32cb2e0 mm: memcontrol: l... |
456 |
excess = soft_limit_excess(memcg); |
bb4cc1a8b revert "memcg: ge... |
457 458 459 460 461 |
/* * We have to update the tree if mz is on RB-tree or * mem is over its softlimit. */ if (excess || mz->on_tree) { |
0a31bc97c mm: memcontrol: r... |
462 463 464 |
unsigned long flags; spin_lock_irqsave(&mctz->lock, flags); |
bb4cc1a8b revert "memcg: ge... |
465 466 |
/* if on-tree, remove it */ if (mz->on_tree) |
cf2c81279 mm: memcontrol: r... |
467 |
__mem_cgroup_remove_exceeded(mz, mctz); |
bb4cc1a8b revert "memcg: ge... |
468 469 470 471 |
/* * Insert again. mz->usage_in_excess will be updated. * If excess is 0, no tree ops. */ |
cf2c81279 mm: memcontrol: r... |
472 |
__mem_cgroup_insert_exceeded(mz, mctz, excess); |
0a31bc97c mm: memcontrol: r... |
473 |
spin_unlock_irqrestore(&mctz->lock, flags); |
bb4cc1a8b revert "memcg: ge... |
474 475 476 477 478 479 |
} } } static void mem_cgroup_remove_from_trees(struct mem_cgroup *memcg) { |
bb4cc1a8b revert "memcg: ge... |
480 |
struct mem_cgroup_tree_per_zone *mctz; |
e231875ba mm: memcontrol: c... |
481 482 |
struct mem_cgroup_per_zone *mz; int nid, zid; |
bb4cc1a8b revert "memcg: ge... |
483 |
|
e231875ba mm: memcontrol: c... |
484 485 486 487 |
for_each_node(nid) { for (zid = 0; zid < MAX_NR_ZONES; zid++) { mz = &memcg->nodeinfo[nid]->zoneinfo[zid]; mctz = soft_limit_tree_node_zone(nid, zid); |
cf2c81279 mm: memcontrol: r... |
488 |
mem_cgroup_remove_exceeded(mz, mctz); |
bb4cc1a8b revert "memcg: ge... |
489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 |
} } } static struct mem_cgroup_per_zone * __mem_cgroup_largest_soft_limit_node(struct mem_cgroup_tree_per_zone *mctz) { struct rb_node *rightmost = NULL; struct mem_cgroup_per_zone *mz; retry: mz = NULL; rightmost = rb_last(&mctz->rb_root); if (!rightmost) goto done; /* Nothing to reclaim from */ mz = rb_entry(rightmost, struct mem_cgroup_per_zone, tree_node); /* * Remove the node now but someone else can add it back, * we will to add it back at the end of reclaim to its correct * position in the tree. */ |
cf2c81279 mm: memcontrol: r... |
511 |
__mem_cgroup_remove_exceeded(mz, mctz); |
3e32cb2e0 mm: memcontrol: l... |
512 |
if (!soft_limit_excess(mz->memcg) || |
ec903c0c8 cgroup: rename cs... |
513 |
!css_tryget_online(&mz->memcg->css)) |
bb4cc1a8b revert "memcg: ge... |
514 515 516 517 518 519 520 521 522 |
goto retry; done: return mz; } static struct mem_cgroup_per_zone * mem_cgroup_largest_soft_limit_node(struct mem_cgroup_tree_per_zone *mctz) { struct mem_cgroup_per_zone *mz; |
0a31bc97c mm: memcontrol: r... |
523 |
spin_lock_irq(&mctz->lock); |
bb4cc1a8b revert "memcg: ge... |
524 |
mz = __mem_cgroup_largest_soft_limit_node(mctz); |
0a31bc97c mm: memcontrol: r... |
525 |
spin_unlock_irq(&mctz->lock); |
bb4cc1a8b revert "memcg: ge... |
526 527 |
return mz; } |
711d3d2c9 memcg: cpu hotplu... |
528 |
/* |
484ebb3b8 memcg: make mem_c... |
529 530 |
* Return page count for single (non recursive) @memcg. * |
711d3d2c9 memcg: cpu hotplu... |
531 532 533 534 535 |
* 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 |
484ebb3b8 memcg: make mem_c... |
536 |
* a periodic synchronization of counter in memcg's counter. |
711d3d2c9 memcg: cpu hotplu... |
537 538 539 540 541 542 543 544 545 |
* * 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 |
484ebb3b8 memcg: make mem_c... |
546 |
* common workload, threshold and synchronization as vmstat[] should be |
711d3d2c9 memcg: cpu hotplu... |
547 548 |
* implemented. */ |
484ebb3b8 memcg: make mem_c... |
549 550 |
static unsigned long mem_cgroup_read_stat(struct mem_cgroup *memcg, enum mem_cgroup_stat_index idx) |
c62b1a3b3 memcg: use generi... |
551 |
{ |
7a159cc9d memcg: use native... |
552 |
long val = 0; |
c62b1a3b3 memcg: use generi... |
553 |
int cpu; |
c62b1a3b3 memcg: use generi... |
554 |
|
484ebb3b8 memcg: make mem_c... |
555 |
/* Per-cpu values can be negative, use a signed accumulator */ |
733a572e6 memcg: make mem_c... |
556 |
for_each_possible_cpu(cpu) |
c0ff4b854 memcg: rename mem... |
557 |
val += per_cpu(memcg->stat->count[idx], cpu); |
484ebb3b8 memcg: make mem_c... |
558 559 560 561 562 563 |
/* * Summing races with updates, so val may be negative. Avoid exposing * transient negative values. */ if (val < 0) val = 0; |
c62b1a3b3 memcg: use generi... |
564 565 |
return val; } |
c0ff4b854 memcg: rename mem... |
566 |
static unsigned long mem_cgroup_read_events(struct mem_cgroup *memcg, |
e9f8974f2 memcg: break out ... |
567 568 569 570 |
enum mem_cgroup_events_index idx) { unsigned long val = 0; int cpu; |
733a572e6 memcg: make mem_c... |
571 |
for_each_possible_cpu(cpu) |
c0ff4b854 memcg: rename mem... |
572 |
val += per_cpu(memcg->stat->events[idx], cpu); |
e9f8974f2 memcg: break out ... |
573 574 |
return val; } |
c0ff4b854 memcg: rename mem... |
575 |
static void mem_cgroup_charge_statistics(struct mem_cgroup *memcg, |
b070e65c0 mm, memcg: add rs... |
576 |
struct page *page, |
f627c2f53 memcg: adjust to ... |
577 |
bool compound, int nr_pages) |
d52aa412d memory cgroup enh... |
578 |
{ |
b24028572 memcg: remove PCG... |
579 580 581 582 |
/* * Here, RSS means 'mapped anon' and anon's SwapCache. Shmem/tmpfs is * counted as CACHE even if it's on ANON LRU. */ |
0a31bc97c mm: memcontrol: r... |
583 |
if (PageAnon(page)) |
b24028572 memcg: remove PCG... |
584 |
__this_cpu_add(memcg->stat->count[MEM_CGROUP_STAT_RSS], |
c0ff4b854 memcg: rename mem... |
585 |
nr_pages); |
d52aa412d memory cgroup enh... |
586 |
else |
b24028572 memcg: remove PCG... |
587 |
__this_cpu_add(memcg->stat->count[MEM_CGROUP_STAT_CACHE], |
c0ff4b854 memcg: rename mem... |
588 |
nr_pages); |
55e462b05 memcg: simple sta... |
589 |
|
f627c2f53 memcg: adjust to ... |
590 591 |
if (compound) { VM_BUG_ON_PAGE(!PageTransHuge(page), page); |
b070e65c0 mm, memcg: add rs... |
592 593 |
__this_cpu_add(memcg->stat->count[MEM_CGROUP_STAT_RSS_HUGE], nr_pages); |
f627c2f53 memcg: adjust to ... |
594 |
} |
b070e65c0 mm, memcg: add rs... |
595 |
|
e401f1761 memcg: modify acc... |
596 597 |
/* pagein of a big page is an event. So, ignore page size */ if (nr_pages > 0) |
c0ff4b854 memcg: rename mem... |
598 |
__this_cpu_inc(memcg->stat->events[MEM_CGROUP_EVENTS_PGPGIN]); |
3751d6043 memcg: fix event ... |
599 |
else { |
c0ff4b854 memcg: rename mem... |
600 |
__this_cpu_inc(memcg->stat->events[MEM_CGROUP_EVENTS_PGPGOUT]); |
3751d6043 memcg: fix event ... |
601 602 |
nr_pages = -nr_pages; /* for event */ } |
e401f1761 memcg: modify acc... |
603 |
|
13114716c mm: memcg: keep r... |
604 |
__this_cpu_add(memcg->stat->nr_page_events, nr_pages); |
6d12e2d8d per-zone and recl... |
605 |
} |
0a6b76dd2 mm: workingset: m... |
606 607 |
unsigned long mem_cgroup_node_nr_lru_pages(struct mem_cgroup *memcg, int nid, unsigned int lru_mask) |
bb2a0de92 memcg: consolidat... |
608 |
{ |
e231875ba mm: memcontrol: c... |
609 |
unsigned long nr = 0; |
889976dbc memcg: reclaim me... |
610 |
int zid; |
e231875ba mm: memcontrol: c... |
611 |
VM_BUG_ON((unsigned)nid >= nr_node_ids); |
bb2a0de92 memcg: consolidat... |
612 |
|
e231875ba mm: memcontrol: c... |
613 614 615 616 617 618 619 620 621 622 623 624 |
for (zid = 0; zid < MAX_NR_ZONES; zid++) { struct mem_cgroup_per_zone *mz; enum lru_list lru; for_each_lru(lru) { if (!(BIT(lru) & lru_mask)) continue; mz = &memcg->nodeinfo[nid]->zoneinfo[zid]; nr += mz->lru_size[lru]; } } return nr; |
889976dbc memcg: reclaim me... |
625 |
} |
bb2a0de92 memcg: consolidat... |
626 |
|
c0ff4b854 memcg: rename mem... |
627 |
static unsigned long mem_cgroup_nr_lru_pages(struct mem_cgroup *memcg, |
bb2a0de92 memcg: consolidat... |
628 |
unsigned int lru_mask) |
6d12e2d8d per-zone and recl... |
629 |
{ |
e231875ba mm: memcontrol: c... |
630 |
unsigned long nr = 0; |
889976dbc memcg: reclaim me... |
631 |
int nid; |
6d12e2d8d per-zone and recl... |
632 |
|
31aaea4aa memcontrol: use N... |
633 |
for_each_node_state(nid, N_MEMORY) |
e231875ba mm: memcontrol: c... |
634 635 |
nr += mem_cgroup_node_nr_lru_pages(memcg, nid, lru_mask); return nr; |
d52aa412d memory cgroup enh... |
636 |
} |
f53d7ce32 mm: memcg: shorte... |
637 638 |
static bool mem_cgroup_event_ratelimit(struct mem_cgroup *memcg, enum mem_cgroup_events_target target) |
7a159cc9d memcg: use native... |
639 640 |
{ unsigned long val, next; |
13114716c mm: memcg: keep r... |
641 |
val = __this_cpu_read(memcg->stat->nr_page_events); |
4799401fe memcg: Fix race c... |
642 |
next = __this_cpu_read(memcg->stat->targets[target]); |
7a159cc9d memcg: use native... |
643 |
/* from time_after() in jiffies.h */ |
f53d7ce32 mm: memcg: shorte... |
644 645 646 647 648 |
if ((long)next - (long)val < 0) { switch (target) { case MEM_CGROUP_TARGET_THRESH: next = val + THRESHOLDS_EVENTS_TARGET; break; |
bb4cc1a8b revert "memcg: ge... |
649 650 651 |
case MEM_CGROUP_TARGET_SOFTLIMIT: next = val + SOFTLIMIT_EVENTS_TARGET; break; |
f53d7ce32 mm: memcg: shorte... |
652 653 654 655 656 657 658 659 |
case MEM_CGROUP_TARGET_NUMAINFO: next = val + NUMAINFO_EVENTS_TARGET; break; default: break; } __this_cpu_write(memcg->stat->targets[target], next); return true; |
7a159cc9d memcg: use native... |
660 |
} |
f53d7ce32 mm: memcg: shorte... |
661 |
return false; |
d2265e6fa memcg : share eve... |
662 663 664 665 666 667 |
} /* * Check events in order. * */ |
c0ff4b854 memcg: rename mem... |
668 |
static void memcg_check_events(struct mem_cgroup *memcg, struct page *page) |
d2265e6fa memcg : share eve... |
669 670 |
{ /* threshold event is triggered in finer grain than soft limit */ |
f53d7ce32 mm: memcg: shorte... |
671 672 |
if (unlikely(mem_cgroup_event_ratelimit(memcg, MEM_CGROUP_TARGET_THRESH))) { |
bb4cc1a8b revert "memcg: ge... |
673 |
bool do_softlimit; |
82b3f2a71 mm/memcontrol.c: ... |
674 |
bool do_numainfo __maybe_unused; |
f53d7ce32 mm: memcg: shorte... |
675 |
|
bb4cc1a8b revert "memcg: ge... |
676 677 |
do_softlimit = mem_cgroup_event_ratelimit(memcg, MEM_CGROUP_TARGET_SOFTLIMIT); |
f53d7ce32 mm: memcg: shorte... |
678 679 680 681 |
#if MAX_NUMNODES > 1 do_numainfo = mem_cgroup_event_ratelimit(memcg, MEM_CGROUP_TARGET_NUMAINFO); #endif |
c0ff4b854 memcg: rename mem... |
682 |
mem_cgroup_threshold(memcg); |
bb4cc1a8b revert "memcg: ge... |
683 684 |
if (unlikely(do_softlimit)) mem_cgroup_update_tree(memcg, page); |
453a9bf34 memcg: fix numa s... |
685 |
#if MAX_NUMNODES > 1 |
f53d7ce32 mm: memcg: shorte... |
686 |
if (unlikely(do_numainfo)) |
c0ff4b854 memcg: rename mem... |
687 |
atomic_inc(&memcg->numainfo_events); |
453a9bf34 memcg: fix numa s... |
688 |
#endif |
0a31bc97c mm: memcontrol: r... |
689 |
} |
d2265e6fa memcg : share eve... |
690 |
} |
cf475ad28 cgroups: add an o... |
691 |
struct mem_cgroup *mem_cgroup_from_task(struct task_struct *p) |
78fb74669 Memory controller... |
692 |
{ |
31a78f23b mm owner: fix rac... |
693 694 695 696 697 698 699 |
/* * mm_update_next_owner() may clear mm->owner to NULL * if it races with swapoff, page migration, etc. * So this can be called with p == NULL. */ if (unlikely(!p)) return NULL; |
073219e99 cgroup: clean up ... |
700 |
return mem_cgroup_from_css(task_css(p, memory_cgrp_id)); |
78fb74669 Memory controller... |
701 |
} |
33398cf2f memcg: export str... |
702 |
EXPORT_SYMBOL(mem_cgroup_from_task); |
78fb74669 Memory controller... |
703 |
|
df3819754 memcg: get_mem_cg... |
704 |
static struct mem_cgroup *get_mem_cgroup_from_mm(struct mm_struct *mm) |
54595fe26 memcg: use css_tr... |
705 |
{ |
c0ff4b854 memcg: rename mem... |
706 |
struct mem_cgroup *memcg = NULL; |
0b7f569e4 memcg: fix OOM ki... |
707 |
|
54595fe26 memcg: use css_tr... |
708 709 |
rcu_read_lock(); do { |
6f6acb005 memcg: fix swapca... |
710 711 712 713 714 715 |
/* * Page cache insertions can happen withou an * actual mm context, e.g. during disk probing * on boot, loopback IO, acct() writes etc. */ if (unlikely(!mm)) |
df3819754 memcg: get_mem_cg... |
716 |
memcg = root_mem_cgroup; |
6f6acb005 memcg: fix swapca... |
717 718 719 720 721 |
else { memcg = mem_cgroup_from_task(rcu_dereference(mm->owner)); if (unlikely(!memcg)) memcg = root_mem_cgroup; } |
ec903c0c8 cgroup: rename cs... |
722 |
} while (!css_tryget_online(&memcg->css)); |
54595fe26 memcg: use css_tr... |
723 |
rcu_read_unlock(); |
c0ff4b854 memcg: rename mem... |
724 |
return memcg; |
54595fe26 memcg: use css_tr... |
725 |
} |
5660048cc mm: move memcg hi... |
726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 |
/** * mem_cgroup_iter - iterate over memory cgroup hierarchy * @root: hierarchy root * @prev: previously returned memcg, NULL on first invocation * @reclaim: cookie for shared reclaim walks, NULL for full walks * * Returns references to children of the hierarchy below @root, or * @root itself, or %NULL after a full round-trip. * * Caller must pass the return value in @prev on subsequent * invocations for reference counting, or use mem_cgroup_iter_break() * to cancel a hierarchy walk before the round-trip is complete. * * Reclaimers can specify a zone and a priority level in @reclaim to * divide up the memcgs in the hierarchy among all concurrent * reclaimers operating on the same zone and priority. */ |
694fbc0fe revert "memcg: en... |
743 |
struct mem_cgroup *mem_cgroup_iter(struct mem_cgroup *root, |
5660048cc mm: move memcg hi... |
744 |
struct mem_cgroup *prev, |
694fbc0fe revert "memcg: en... |
745 |
struct mem_cgroup_reclaim_cookie *reclaim) |
14067bb3e memcg: hierarchic... |
746 |
{ |
33398cf2f memcg: export str... |
747 |
struct mem_cgroup_reclaim_iter *uninitialized_var(iter); |
5ac8fb31a mm: memcontrol: c... |
748 |
struct cgroup_subsys_state *css = NULL; |
9f3a0d093 mm: memcg: consol... |
749 |
struct mem_cgroup *memcg = NULL; |
5ac8fb31a mm: memcontrol: c... |
750 |
struct mem_cgroup *pos = NULL; |
711d3d2c9 memcg: cpu hotplu... |
751 |
|
694fbc0fe revert "memcg: en... |
752 753 |
if (mem_cgroup_disabled()) return NULL; |
5660048cc mm: move memcg hi... |
754 |
|
9f3a0d093 mm: memcg: consol... |
755 756 |
if (!root) root = root_mem_cgroup; |
7d74b06f2 memcg: use for_ea... |
757 |
|
9f3a0d093 mm: memcg: consol... |
758 |
if (prev && !reclaim) |
5ac8fb31a mm: memcontrol: c... |
759 |
pos = prev; |
14067bb3e memcg: hierarchic... |
760 |
|
9f3a0d093 mm: memcg: consol... |
761 762 |
if (!root->use_hierarchy && root != root_mem_cgroup) { if (prev) |
5ac8fb31a mm: memcontrol: c... |
763 |
goto out; |
694fbc0fe revert "memcg: en... |
764 |
return root; |
9f3a0d093 mm: memcg: consol... |
765 |
} |
14067bb3e memcg: hierarchic... |
766 |
|
542f85f9a memcg: rework mem... |
767 |
rcu_read_lock(); |
5f5781619 memcg: relax memc... |
768 |
|
5ac8fb31a mm: memcontrol: c... |
769 770 771 772 773 774 775 776 |
if (reclaim) { struct mem_cgroup_per_zone *mz; mz = mem_cgroup_zone_zoneinfo(root, reclaim->zone); iter = &mz->iter[reclaim->priority]; if (prev && reclaim->generation != iter->generation) goto out_unlock; |
6df38689e mm: memcontrol: f... |
777 |
while (1) { |
4db0c3c29 mm: remove rest o... |
778 |
pos = READ_ONCE(iter->position); |
6df38689e mm: memcontrol: f... |
779 780 |
if (!pos || css_tryget(&pos->css)) break; |
5ac8fb31a mm: memcontrol: c... |
781 |
/* |
6df38689e mm: memcontrol: f... |
782 783 784 785 786 787 |
* css reference reached zero, so iter->position will * be cleared by ->css_released. However, we should not * rely on this happening soon, because ->css_released * is called from a work queue, and by busy-waiting we * might block it. So we clear iter->position right * away. |
5ac8fb31a mm: memcontrol: c... |
788 |
*/ |
6df38689e mm: memcontrol: f... |
789 790 |
(void)cmpxchg(&iter->position, pos, NULL); } |
5ac8fb31a mm: memcontrol: c... |
791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 |
} if (pos) css = &pos->css; for (;;) { css = css_next_descendant_pre(css, &root->css); if (!css) { /* * Reclaimers share the hierarchy walk, and a * new one might jump in right at the end of * the hierarchy - make sure they see at least * one group and restart from the beginning. */ if (!prev) continue; break; |
527a5ec9a mm: memcg: per-pr... |
808 |
} |
7d74b06f2 memcg: use for_ea... |
809 |
|
5ac8fb31a mm: memcontrol: c... |
810 811 812 813 814 815 |
/* * Verify the css and acquire a reference. The root * is provided by the caller, so we know it's alive * and kicking, and don't take an extra reference. */ memcg = mem_cgroup_from_css(css); |
14067bb3e memcg: hierarchic... |
816 |
|
5ac8fb31a mm: memcontrol: c... |
817 818 |
if (css == &root->css) break; |
14067bb3e memcg: hierarchic... |
819 |
|
0b8f73e10 mm: memcontrol: c... |
820 821 |
if (css_tryget(css)) break; |
9f3a0d093 mm: memcg: consol... |
822 |
|
5ac8fb31a mm: memcontrol: c... |
823 |
memcg = NULL; |
9f3a0d093 mm: memcg: consol... |
824 |
} |
5ac8fb31a mm: memcontrol: c... |
825 826 |
if (reclaim) { |
5ac8fb31a mm: memcontrol: c... |
827 |
/* |
6df38689e mm: memcontrol: f... |
828 829 830 |
* The position could have already been updated by a competing * thread, so check that the value hasn't changed since we read * it to avoid reclaiming from the same cgroup twice. |
5ac8fb31a mm: memcontrol: c... |
831 |
*/ |
6df38689e mm: memcontrol: f... |
832 |
(void)cmpxchg(&iter->position, pos, memcg); |
5ac8fb31a mm: memcontrol: c... |
833 834 835 836 837 838 839 |
if (pos) css_put(&pos->css); if (!memcg) iter->generation++; else if (!prev) reclaim->generation = iter->generation; |
9f3a0d093 mm: memcg: consol... |
840 |
} |
5ac8fb31a mm: memcontrol: c... |
841 |
|
542f85f9a memcg: rework mem... |
842 843 |
out_unlock: rcu_read_unlock(); |
5ac8fb31a mm: memcontrol: c... |
844 |
out: |
c40046f3a memcg: keep prev'... |
845 846 |
if (prev && prev != root) css_put(&prev->css); |
9f3a0d093 mm: memcg: consol... |
847 |
return memcg; |
14067bb3e memcg: hierarchic... |
848 |
} |
7d74b06f2 memcg: use for_ea... |
849 |
|
5660048cc mm: move memcg hi... |
850 851 852 853 854 855 856 |
/** * mem_cgroup_iter_break - abort a hierarchy walk prematurely * @root: hierarchy root * @prev: last visited hierarchy member as returned by mem_cgroup_iter() */ void mem_cgroup_iter_break(struct mem_cgroup *root, struct mem_cgroup *prev) |
9f3a0d093 mm: memcg: consol... |
857 858 859 860 861 862 |
{ if (!root) root = root_mem_cgroup; if (prev && prev != root) css_put(&prev->css); } |
7d74b06f2 memcg: use for_ea... |
863 |
|
6df38689e mm: memcontrol: f... |
864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 |
static void invalidate_reclaim_iterators(struct mem_cgroup *dead_memcg) { struct mem_cgroup *memcg = dead_memcg; struct mem_cgroup_reclaim_iter *iter; struct mem_cgroup_per_zone *mz; int nid, zid; int i; while ((memcg = parent_mem_cgroup(memcg))) { for_each_node(nid) { for (zid = 0; zid < MAX_NR_ZONES; zid++) { mz = &memcg->nodeinfo[nid]->zoneinfo[zid]; for (i = 0; i <= DEF_PRIORITY; i++) { iter = &mz->iter[i]; cmpxchg(&iter->position, dead_memcg, NULL); } } } } } |
9f3a0d093 mm: memcg: consol... |
885 886 887 888 889 890 |
/* * Iteration constructs for visiting all cgroups (under a tree). If * loops are exited prematurely (break), mem_cgroup_iter_break() must * be used for reference counting. */ #define for_each_mem_cgroup_tree(iter, root) \ |
527a5ec9a mm: memcg: per-pr... |
891 |
for (iter = mem_cgroup_iter(root, NULL, NULL); \ |
9f3a0d093 mm: memcg: consol... |
892 |
iter != NULL; \ |
527a5ec9a mm: memcg: per-pr... |
893 |
iter = mem_cgroup_iter(root, iter, NULL)) |
711d3d2c9 memcg: cpu hotplu... |
894 |
|
9f3a0d093 mm: memcg: consol... |
895 |
#define for_each_mem_cgroup(iter) \ |
527a5ec9a mm: memcg: per-pr... |
896 |
for (iter = mem_cgroup_iter(NULL, NULL, NULL); \ |
9f3a0d093 mm: memcg: consol... |
897 |
iter != NULL; \ |
527a5ec9a mm: memcg: per-pr... |
898 |
iter = mem_cgroup_iter(NULL, iter, NULL)) |
14067bb3e memcg: hierarchic... |
899 |
|
925b7673c mm: make per-memc... |
900 901 902 |
/** * mem_cgroup_zone_lruvec - get the lru list vector for a zone and memcg * @zone: zone of the wanted lruvec |
fa9add641 mm/memcg: apply a... |
903 |
* @memcg: memcg of the wanted lruvec |
925b7673c mm: make per-memc... |
904 905 906 907 908 909 910 911 912 |
* * Returns the lru list vector holding pages for the given @zone and * @mem. This can be the global zone lruvec, if the memory controller * is disabled. */ struct lruvec *mem_cgroup_zone_lruvec(struct zone *zone, struct mem_cgroup *memcg) { struct mem_cgroup_per_zone *mz; |
bea8c150a memcg: fix hotplu... |
913 |
struct lruvec *lruvec; |
925b7673c mm: make per-memc... |
914 |
|
bea8c150a memcg: fix hotplu... |
915 916 917 918 |
if (mem_cgroup_disabled()) { lruvec = &zone->lruvec; goto out; } |
925b7673c mm: make per-memc... |
919 |
|
e231875ba mm: memcontrol: c... |
920 |
mz = mem_cgroup_zone_zoneinfo(memcg, zone); |
bea8c150a memcg: fix hotplu... |
921 922 923 924 925 926 927 928 929 930 |
lruvec = &mz->lruvec; out: /* * Since a node can be onlined after the mem_cgroup was created, * we have to be prepared to initialize lruvec->zone here; * and if offlined then reonlined, we need to reinitialize it. */ if (unlikely(lruvec->zone != zone)) lruvec->zone = zone; return lruvec; |
925b7673c mm: make per-memc... |
931 |
} |
925b7673c mm: make per-memc... |
932 |
/** |
dfe0e773d mm: memcontrol: u... |
933 |
* mem_cgroup_page_lruvec - return lruvec for isolating/putting an LRU page |
925b7673c mm: make per-memc... |
934 |
* @page: the page |
fa9add641 mm/memcg: apply a... |
935 |
* @zone: zone of the page |
dfe0e773d mm: memcontrol: u... |
936 937 938 939 |
* * This function is only safe when following the LRU page isolation * and putback protocol: the LRU lock must be held, and the page must * either be PageLRU() or the caller must have isolated/allocated it. |
925b7673c mm: make per-memc... |
940 |
*/ |
fa9add641 mm/memcg: apply a... |
941 |
struct lruvec *mem_cgroup_page_lruvec(struct page *page, struct zone *zone) |
08e552c69 memcg: synchroniz... |
942 |
{ |
08e552c69 memcg: synchroniz... |
943 |
struct mem_cgroup_per_zone *mz; |
925b7673c mm: make per-memc... |
944 |
struct mem_cgroup *memcg; |
bea8c150a memcg: fix hotplu... |
945 |
struct lruvec *lruvec; |
6d12e2d8d per-zone and recl... |
946 |
|
bea8c150a memcg: fix hotplu... |
947 948 949 950 |
if (mem_cgroup_disabled()) { lruvec = &zone->lruvec; goto out; } |
925b7673c mm: make per-memc... |
951 |
|
1306a85ae mm: embed the mem... |
952 |
memcg = page->mem_cgroup; |
7512102cf memcg: fix GPF wh... |
953 |
/* |
dfe0e773d mm: memcontrol: u... |
954 |
* Swapcache readahead pages are added to the LRU - and |
298333157 mm: memcontrol: r... |
955 |
* possibly migrated - before they are charged. |
7512102cf memcg: fix GPF wh... |
956 |
*/ |
298333157 mm: memcontrol: r... |
957 958 |
if (!memcg) memcg = root_mem_cgroup; |
7512102cf memcg: fix GPF wh... |
959 |
|
e231875ba mm: memcontrol: c... |
960 |
mz = mem_cgroup_page_zoneinfo(memcg, page); |
bea8c150a memcg: fix hotplu... |
961 962 963 964 965 966 967 968 969 970 |
lruvec = &mz->lruvec; out: /* * Since a node can be onlined after the mem_cgroup was created, * we have to be prepared to initialize lruvec->zone here; * and if offlined then reonlined, we need to reinitialize it. */ if (unlikely(lruvec->zone != zone)) lruvec->zone = zone; return lruvec; |
08e552c69 memcg: synchroniz... |
971 |
} |
b69408e88 vmscan: Use an in... |
972 |
|
925b7673c mm: make per-memc... |
973 |
/** |
fa9add641 mm/memcg: apply a... |
974 975 976 977 |
* mem_cgroup_update_lru_size - account for adding or removing an lru page * @lruvec: mem_cgroup per zone lru vector * @lru: index of lru list the page is sitting on * @nr_pages: positive when adding or negative when removing |
925b7673c mm: make per-memc... |
978 |
* |
ca707239e mm: update_lru_si... |
979 980 981 |
* This function must be called under lru_lock, just before a page is added * to or just after a page is removed from an lru list (that ordering being * so as to allow it to check that lru_size 0 is consistent with list_empty). |
3f58a8294 memcg: move memcg... |
982 |
*/ |
fa9add641 mm/memcg: apply a... |
983 984 |
void mem_cgroup_update_lru_size(struct lruvec *lruvec, enum lru_list lru, int nr_pages) |
3f58a8294 memcg: move memcg... |
985 986 |
{ struct mem_cgroup_per_zone *mz; |
fa9add641 mm/memcg: apply a... |
987 |
unsigned long *lru_size; |
ca707239e mm: update_lru_si... |
988 989 |
long size; bool empty; |
3f58a8294 memcg: move memcg... |
990 |
|
9d5e6a9f2 mm: update_lru_si... |
991 |
__update_lru_size(lruvec, lru, nr_pages); |
3f58a8294 memcg: move memcg... |
992 993 |
if (mem_cgroup_disabled()) return; |
fa9add641 mm/memcg: apply a... |
994 995 |
mz = container_of(lruvec, struct mem_cgroup_per_zone, lruvec); lru_size = mz->lru_size + lru; |
ca707239e mm: update_lru_si... |
996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 |
empty = list_empty(lruvec->lists + lru); if (nr_pages < 0) *lru_size += nr_pages; size = *lru_size; if (WARN_ONCE(size < 0 || empty != !size, "%s(%p, %d, %d): lru_size %ld but %sempty ", __func__, lruvec, lru, nr_pages, size, empty ? "" : "not ")) { VM_BUG_ON(1); *lru_size = 0; } if (nr_pages > 0) *lru_size += nr_pages; |
08e552c69 memcg: synchroniz... |
1012 |
} |
544122e5e memcg: fix LRU ac... |
1013 |
|
2314b42db mm: memcontrol: d... |
1014 |
bool task_in_mem_cgroup(struct task_struct *task, struct mem_cgroup *memcg) |
c3ac9a8ad mm: memcg: count ... |
1015 |
{ |
2314b42db mm: memcontrol: d... |
1016 |
struct mem_cgroup *task_memcg; |
158e0a2d1 memcg: use find_l... |
1017 |
struct task_struct *p; |
ffbdccf5e mm, memcg: don't ... |
1018 |
bool ret; |
4c4a22148 memcontrol: move ... |
1019 |
|
158e0a2d1 memcg: use find_l... |
1020 |
p = find_lock_task_mm(task); |
de077d222 oom, memcg: fix e... |
1021 |
if (p) { |
2314b42db mm: memcontrol: d... |
1022 |
task_memcg = get_mem_cgroup_from_mm(p->mm); |
de077d222 oom, memcg: fix e... |
1023 1024 1025 1026 1027 1028 1029 |
task_unlock(p); } else { /* * All threads may have already detached their mm's, but the oom * killer still needs to detect if they have already been oom * killed to prevent needlessly killing additional tasks. */ |
ffbdccf5e mm, memcg: don't ... |
1030 |
rcu_read_lock(); |
2314b42db mm: memcontrol: d... |
1031 1032 |
task_memcg = mem_cgroup_from_task(task); css_get(&task_memcg->css); |
ffbdccf5e mm, memcg: don't ... |
1033 |
rcu_read_unlock(); |
de077d222 oom, memcg: fix e... |
1034 |
} |
2314b42db mm: memcontrol: d... |
1035 1036 |
ret = mem_cgroup_is_descendant(task_memcg, memcg); css_put(&task_memcg->css); |
4c4a22148 memcontrol: move ... |
1037 1038 |
return ret; } |
19942822d memcg: prevent en... |
1039 |
/** |
9d11ea9f1 memcg: simplify t... |
1040 |
* mem_cgroup_margin - calculate chargeable space of a memory cgroup |
dad7557eb mm: fix kernel-do... |
1041 |
* @memcg: the memory cgroup |
19942822d memcg: prevent en... |
1042 |
* |
9d11ea9f1 memcg: simplify t... |
1043 |
* Returns the maximum amount of memory @mem can be charged with, in |
7ec99d621 memcg: unify char... |
1044 |
* pages. |
19942822d memcg: prevent en... |
1045 |
*/ |
c0ff4b854 memcg: rename mem... |
1046 |
static unsigned long mem_cgroup_margin(struct mem_cgroup *memcg) |
19942822d memcg: prevent en... |
1047 |
{ |
3e32cb2e0 mm: memcontrol: l... |
1048 1049 1050 |
unsigned long margin = 0; unsigned long count; unsigned long limit; |
9d11ea9f1 memcg: simplify t... |
1051 |
|
3e32cb2e0 mm: memcontrol: l... |
1052 |
count = page_counter_read(&memcg->memory); |
4db0c3c29 mm: remove rest o... |
1053 |
limit = READ_ONCE(memcg->memory.limit); |
3e32cb2e0 mm: memcontrol: l... |
1054 1055 |
if (count < limit) margin = limit - count; |
7941d2145 mm: memcontrol: d... |
1056 |
if (do_memsw_account()) { |
3e32cb2e0 mm: memcontrol: l... |
1057 |
count = page_counter_read(&memcg->memsw); |
4db0c3c29 mm: remove rest o... |
1058 |
limit = READ_ONCE(memcg->memsw.limit); |
3e32cb2e0 mm: memcontrol: l... |
1059 1060 1061 1062 1063 |
if (count <= limit) margin = min(margin, limit - count); } return margin; |
19942822d memcg: prevent en... |
1064 |
} |
619d094b5 memcg: simplify m... |
1065 |
/* |
bdcbb659f memcg: fold mem_c... |
1066 |
* A routine for checking "mem" is under move_account() or not. |
32047e2a8 memcg: avoid lock... |
1067 |
* |
bdcbb659f memcg: fold mem_c... |
1068 1069 1070 |
* Checking a cgroup is mc.from or mc.to or under hierarchy of * moving cgroups. This is for waiting at high-memory pressure * caused by "move". |
32047e2a8 memcg: avoid lock... |
1071 |
*/ |
c0ff4b854 memcg: rename mem... |
1072 |
static bool mem_cgroup_under_move(struct mem_cgroup *memcg) |
4b5343346 memcg: clean up t... |
1073 |
{ |
2bd9bb206 memcg: clean up w... |
1074 1075 |
struct mem_cgroup *from; struct mem_cgroup *to; |
4b5343346 memcg: clean up t... |
1076 |
bool ret = false; |
2bd9bb206 memcg: clean up w... |
1077 1078 1079 1080 1081 1082 1083 1084 1085 |
/* * 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... |
1086 |
|
2314b42db mm: memcontrol: d... |
1087 1088 |
ret = mem_cgroup_is_descendant(from, memcg) || mem_cgroup_is_descendant(to, memcg); |
2bd9bb206 memcg: clean up w... |
1089 1090 |
unlock: spin_unlock(&mc.lock); |
4b5343346 memcg: clean up t... |
1091 1092 |
return ret; } |
c0ff4b854 memcg: rename mem... |
1093 |
static bool mem_cgroup_wait_acct_move(struct mem_cgroup *memcg) |
4b5343346 memcg: clean up t... |
1094 1095 |
{ if (mc.moving_task && current != mc.moving_task) { |
c0ff4b854 memcg: rename mem... |
1096 |
if (mem_cgroup_under_move(memcg)) { |
4b5343346 memcg: clean up t... |
1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 |
DEFINE_WAIT(wait); prepare_to_wait(&mc.waitq, &wait, TASK_INTERRUPTIBLE); /* moving charge context might have finished. */ if (mc.moving_task) schedule(); finish_wait(&mc.waitq, &wait); return true; } } return false; } |
58cf188ed memcg, oom: provi... |
1108 |
#define K(x) ((x) << (PAGE_SHIFT-10)) |
e222432bf memcg: show memcg... |
1109 |
/** |
58cf188ed memcg, oom: provi... |
1110 |
* mem_cgroup_print_oom_info: Print OOM information relevant to memory controller. |
e222432bf memcg: show memcg... |
1111 1112 1113 1114 1115 1116 1117 1118 |
* @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) { |
58cf188ed memcg, oom: provi... |
1119 1120 |
struct mem_cgroup *iter; unsigned int i; |
e222432bf memcg: show memcg... |
1121 |
|
e222432bf memcg: show memcg... |
1122 |
rcu_read_lock(); |
2415b9f5c memcg: print cgro... |
1123 1124 1125 1126 1127 1128 1129 |
if (p) { pr_info("Task in "); pr_cont_cgroup_path(task_cgroup(p, memory_cgrp_id)); pr_cont(" killed as a result of limit of "); } else { pr_info("Memory limit reached of cgroup "); } |
e61734c55 cgroup: remove cg... |
1130 |
pr_cont_cgroup_path(memcg->css.cgroup); |
0346dadbf memcg: remove ext... |
1131 1132 |
pr_cont(" "); |
e222432bf memcg: show memcg... |
1133 |
|
e222432bf memcg: show memcg... |
1134 |
rcu_read_unlock(); |
3e32cb2e0 mm: memcontrol: l... |
1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 |
pr_info("memory: usage %llukB, limit %llukB, failcnt %lu ", K((u64)page_counter_read(&memcg->memory)), K((u64)memcg->memory.limit), memcg->memory.failcnt); pr_info("memory+swap: usage %llukB, limit %llukB, failcnt %lu ", K((u64)page_counter_read(&memcg->memsw)), K((u64)memcg->memsw.limit), memcg->memsw.failcnt); pr_info("kmem: usage %llukB, limit %llukB, failcnt %lu ", K((u64)page_counter_read(&memcg->kmem)), K((u64)memcg->kmem.limit), memcg->kmem.failcnt); |
58cf188ed memcg, oom: provi... |
1147 1148 |
for_each_mem_cgroup_tree(iter, memcg) { |
e61734c55 cgroup: remove cg... |
1149 1150 |
pr_info("Memory cgroup stats for "); pr_cont_cgroup_path(iter->css.cgroup); |
58cf188ed memcg, oom: provi... |
1151 1152 1153 |
pr_cont(":"); for (i = 0; i < MEM_CGROUP_STAT_NSTATS; i++) { |
37e843511 mm: memcontrol: c... |
1154 |
if (i == MEM_CGROUP_STAT_SWAP && !do_swap_account) |
58cf188ed memcg, oom: provi... |
1155 |
continue; |
484ebb3b8 memcg: make mem_c... |
1156 |
pr_cont(" %s:%luKB", mem_cgroup_stat_names[i], |
58cf188ed memcg, oom: provi... |
1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 |
K(mem_cgroup_read_stat(iter, i))); } for (i = 0; i < NR_LRU_LISTS; i++) pr_cont(" %s:%luKB", mem_cgroup_lru_names[i], K(mem_cgroup_nr_lru_pages(iter, BIT(i)))); pr_cont(" "); } |
e222432bf memcg: show memcg... |
1167 |
} |
81d39c20f memcg: fix shrink... |
1168 1169 1170 1171 |
/* * This function returns the number of memcg under hierarchy tree. Returns * 1(self count) if no children. */ |
c0ff4b854 memcg: rename mem... |
1172 |
static int mem_cgroup_count_children(struct mem_cgroup *memcg) |
81d39c20f memcg: fix shrink... |
1173 1174 |
{ int num = 0; |
7d74b06f2 memcg: use for_ea... |
1175 |
struct mem_cgroup *iter; |
c0ff4b854 memcg: rename mem... |
1176 |
for_each_mem_cgroup_tree(iter, memcg) |
7d74b06f2 memcg: use for_ea... |
1177 |
num++; |
81d39c20f memcg: fix shrink... |
1178 1179 |
return num; } |
6d61ef409 memcg: memory cgr... |
1180 |
/* |
a63d83f42 oom: badness heur... |
1181 1182 |
* Return the memory (and swap, if configured) limit for a memcg. */ |
3e32cb2e0 mm: memcontrol: l... |
1183 |
static unsigned long mem_cgroup_get_limit(struct mem_cgroup *memcg) |
a63d83f42 oom: badness heur... |
1184 |
{ |
3e32cb2e0 mm: memcontrol: l... |
1185 |
unsigned long limit; |
f3e8eb70b memcg: fix unit m... |
1186 |
|
3e32cb2e0 mm: memcontrol: l... |
1187 |
limit = memcg->memory.limit; |
9a5a8f19b memcg: oom: fix t... |
1188 |
if (mem_cgroup_swappiness(memcg)) { |
3e32cb2e0 mm: memcontrol: l... |
1189 |
unsigned long memsw_limit; |
37e843511 mm: memcontrol: c... |
1190 |
unsigned long swap_limit; |
9a5a8f19b memcg: oom: fix t... |
1191 |
|
3e32cb2e0 mm: memcontrol: l... |
1192 |
memsw_limit = memcg->memsw.limit; |
37e843511 mm: memcontrol: c... |
1193 1194 1195 |
swap_limit = memcg->swap.limit; swap_limit = min(swap_limit, (unsigned long)total_swap_pages); limit = min(limit + swap_limit, memsw_limit); |
9a5a8f19b memcg: oom: fix t... |
1196 |
} |
9a5a8f19b memcg: oom: fix t... |
1197 |
return limit; |
a63d83f42 oom: badness heur... |
1198 |
} |
b6e6edcfa mm: memcontrol: r... |
1199 |
static bool mem_cgroup_out_of_memory(struct mem_cgroup *memcg, gfp_t gfp_mask, |
19965460e mm, memcg: make m... |
1200 |
int order) |
9cbb78bb3 mm, memcg: introd... |
1201 |
{ |
6e0fc46dc mm, oom: organize... |
1202 1203 1204 1205 1206 |
struct oom_control oc = { .zonelist = NULL, .nodemask = NULL, .gfp_mask = gfp_mask, .order = order, |
6e0fc46dc mm, oom: organize... |
1207 |
}; |
9cbb78bb3 mm, memcg: introd... |
1208 1209 1210 1211 1212 |
struct mem_cgroup *iter; unsigned long chosen_points = 0; unsigned long totalpages; unsigned int points = 0; struct task_struct *chosen = NULL; |
dc56401fc mm: oom_kill: sim... |
1213 |
mutex_lock(&oom_lock); |
876aafbfd mm, memcg: move a... |
1214 |
/* |
465adcf1e mm, memcg: give e... |
1215 1216 1217 |
* If current has a pending SIGKILL or is exiting, then automatically * select it. The goal is to allow it to allocate so that it may * quickly exit and free its memory. |
876aafbfd mm, memcg: move a... |
1218 |
*/ |
d003f371b oom: don't assume... |
1219 |
if (fatal_signal_pending(current) || task_will_free_mem(current)) { |
16e951966 mm: oom_kill: cle... |
1220 |
mark_oom_victim(current); |
3ef22dfff oom, oom_reaper: ... |
1221 |
try_oom_reaper(current); |
dc56401fc mm: oom_kill: sim... |
1222 |
goto unlock; |
876aafbfd mm, memcg: move a... |
1223 |
} |
6e0fc46dc mm, oom: organize... |
1224 |
check_panic_on_oom(&oc, CONSTRAINT_MEMCG, memcg); |
3e32cb2e0 mm: memcontrol: l... |
1225 |
totalpages = mem_cgroup_get_limit(memcg) ? : 1; |
9cbb78bb3 mm, memcg: introd... |
1226 |
for_each_mem_cgroup_tree(iter, memcg) { |
72ec70299 cgroup: make task... |
1227 |
struct css_task_iter it; |
9cbb78bb3 mm, memcg: introd... |
1228 |
struct task_struct *task; |
72ec70299 cgroup: make task... |
1229 1230 |
css_task_iter_start(&iter->css, &it); while ((task = css_task_iter_next(&it))) { |
6e0fc46dc mm, oom: organize... |
1231 |
switch (oom_scan_process_thread(&oc, task, totalpages)) { |
9cbb78bb3 mm, memcg: introd... |
1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 |
case OOM_SCAN_SELECT: if (chosen) put_task_struct(chosen); chosen = task; chosen_points = ULONG_MAX; get_task_struct(chosen); /* fall through */ case OOM_SCAN_CONTINUE: continue; case OOM_SCAN_ABORT: |
72ec70299 cgroup: make task... |
1242 |
css_task_iter_end(&it); |
9cbb78bb3 mm, memcg: introd... |
1243 1244 1245 |
mem_cgroup_iter_break(memcg, iter); if (chosen) put_task_struct(chosen); |
dc56401fc mm: oom_kill: sim... |
1246 |
goto unlock; |
9cbb78bb3 mm, memcg: introd... |
1247 1248 1249 1250 |
case OOM_SCAN_OK: break; }; points = oom_badness(task, memcg, NULL, totalpages); |
d49ad9355 mm, oom: prefer t... |
1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 |
if (!points || points < chosen_points) continue; /* Prefer thread group leaders for display purposes */ if (points == chosen_points && thread_group_leader(chosen)) continue; if (chosen) put_task_struct(chosen); chosen = task; chosen_points = points; get_task_struct(chosen); |
9cbb78bb3 mm, memcg: introd... |
1263 |
} |
72ec70299 cgroup: make task... |
1264 |
css_task_iter_end(&it); |
9cbb78bb3 mm, memcg: introd... |
1265 |
} |
dc56401fc mm: oom_kill: sim... |
1266 1267 |
if (chosen) { points = chosen_points * 1000 / totalpages; |
6e0fc46dc mm, oom: organize... |
1268 1269 |
oom_kill_process(&oc, chosen, points, totalpages, memcg, "Memory cgroup out of memory"); |
dc56401fc mm: oom_kill: sim... |
1270 1271 1272 |
} unlock: mutex_unlock(&oom_lock); |
b6e6edcfa mm: memcontrol: r... |
1273 |
return chosen; |
9cbb78bb3 mm, memcg: introd... |
1274 |
} |
ae6e71d3d mm/memcontrol.c: ... |
1275 |
#if MAX_NUMNODES > 1 |
4d0c066d2 memcg: fix reclai... |
1276 1277 |
/** * test_mem_cgroup_node_reclaimable |
dad7557eb mm: fix kernel-do... |
1278 |
* @memcg: the target memcg |
4d0c066d2 memcg: fix reclai... |
1279 1280 1281 1282 1283 1284 1285 |
* @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... |
1286 |
static bool test_mem_cgroup_node_reclaimable(struct mem_cgroup *memcg, |
4d0c066d2 memcg: fix reclai... |
1287 1288 |
int nid, bool noswap) { |
c0ff4b854 memcg: rename mem... |
1289 |
if (mem_cgroup_node_nr_lru_pages(memcg, nid, LRU_ALL_FILE)) |
4d0c066d2 memcg: fix reclai... |
1290 1291 1292 |
return true; if (noswap || !total_swap_pages) return false; |
c0ff4b854 memcg: rename mem... |
1293 |
if (mem_cgroup_node_nr_lru_pages(memcg, nid, LRU_ALL_ANON)) |
4d0c066d2 memcg: fix reclai... |
1294 1295 1296 1297 |
return true; return false; } |
889976dbc memcg: reclaim me... |
1298 1299 1300 1301 1302 1303 1304 |
/* * 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... |
1305 |
static void mem_cgroup_may_update_nodemask(struct mem_cgroup *memcg) |
889976dbc memcg: reclaim me... |
1306 1307 |
{ int nid; |
453a9bf34 memcg: fix numa s... |
1308 1309 1310 1311 |
/* * numainfo_events > 0 means there was at least NUMAINFO_EVENTS_TARGET * pagein/pageout changes since the last update. */ |
c0ff4b854 memcg: rename mem... |
1312 |
if (!atomic_read(&memcg->numainfo_events)) |
453a9bf34 memcg: fix numa s... |
1313 |
return; |
c0ff4b854 memcg: rename mem... |
1314 |
if (atomic_inc_return(&memcg->numainfo_updating) > 1) |
889976dbc memcg: reclaim me... |
1315 |
return; |
889976dbc memcg: reclaim me... |
1316 |
/* make a nodemask where this memcg uses memory from */ |
31aaea4aa memcontrol: use N... |
1317 |
memcg->scan_nodes = node_states[N_MEMORY]; |
889976dbc memcg: reclaim me... |
1318 |
|
31aaea4aa memcontrol: use N... |
1319 |
for_each_node_mask(nid, node_states[N_MEMORY]) { |
889976dbc memcg: reclaim me... |
1320 |
|
c0ff4b854 memcg: rename mem... |
1321 1322 |
if (!test_mem_cgroup_node_reclaimable(memcg, nid, false)) node_clear(nid, memcg->scan_nodes); |
889976dbc memcg: reclaim me... |
1323 |
} |
453a9bf34 memcg: fix numa s... |
1324 |
|
c0ff4b854 memcg: rename mem... |
1325 1326 |
atomic_set(&memcg->numainfo_events, 0); atomic_set(&memcg->numainfo_updating, 0); |
889976dbc memcg: reclaim me... |
1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 |
} /* * 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... |
1341 |
int mem_cgroup_select_victim_node(struct mem_cgroup *memcg) |
889976dbc memcg: reclaim me... |
1342 1343 |
{ int node; |
c0ff4b854 memcg: rename mem... |
1344 1345 |
mem_cgroup_may_update_nodemask(memcg); node = memcg->last_scanned_node; |
889976dbc memcg: reclaim me... |
1346 |
|
0edaf86cf include/linux/nod... |
1347 |
node = next_node_in(node, memcg->scan_nodes); |
889976dbc memcg: reclaim me... |
1348 |
/* |
fda3d69be mm/memcontrol.c:m... |
1349 1350 1351 |
* mem_cgroup_may_update_nodemask might have seen no reclaimmable pages * last time it really checked all the LRUs due to rate limiting. * Fallback to the current node in that case for simplicity. |
889976dbc memcg: reclaim me... |
1352 1353 1354 |
*/ if (unlikely(node == MAX_NUMNODES)) node = numa_node_id(); |
c0ff4b854 memcg: rename mem... |
1355 |
memcg->last_scanned_node = node; |
889976dbc memcg: reclaim me... |
1356 1357 |
return node; } |
889976dbc memcg: reclaim me... |
1358 |
#else |
c0ff4b854 memcg: rename mem... |
1359 |
int mem_cgroup_select_victim_node(struct mem_cgroup *memcg) |
889976dbc memcg: reclaim me... |
1360 1361 1362 1363 |
{ return 0; } #endif |
0608f43da revert "memcg, vm... |
1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 |
static int mem_cgroup_soft_reclaim(struct mem_cgroup *root_memcg, struct zone *zone, gfp_t gfp_mask, unsigned long *total_scanned) { struct mem_cgroup *victim = NULL; int total = 0; int loop = 0; unsigned long excess; unsigned long nr_scanned; struct mem_cgroup_reclaim_cookie reclaim = { .zone = zone, .priority = 0, }; |
3e32cb2e0 mm: memcontrol: l... |
1378 |
excess = soft_limit_excess(root_memcg); |
0608f43da revert "memcg, vm... |
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 |
while (1) { victim = mem_cgroup_iter(root_memcg, victim, &reclaim); if (!victim) { loop++; if (loop >= 2) { /* * If we have not been able to reclaim * anything, it might because there are * no reclaimable pages under this hierarchy */ if (!total) break; /* * We want to do more targeted reclaim. * excess >> 2 is not to excessive so as to * reclaim too much, nor too less that we keep * coming back to reclaim from this cgroup */ if (total >= (excess >> 2) || (loop > MEM_CGROUP_MAX_RECLAIM_LOOPS)) break; } continue; } |
0608f43da revert "memcg, vm... |
1404 1405 1406 |
total += mem_cgroup_shrink_node_zone(victim, gfp_mask, false, zone, &nr_scanned); *total_scanned += nr_scanned; |
3e32cb2e0 mm: memcontrol: l... |
1407 |
if (!soft_limit_excess(root_memcg)) |
0608f43da revert "memcg, vm... |
1408 |
break; |
6d61ef409 memcg: memory cgr... |
1409 |
} |
0608f43da revert "memcg, vm... |
1410 1411 |
mem_cgroup_iter_break(root_memcg, victim); return total; |
6d61ef409 memcg: memory cgr... |
1412 |
} |
0056f4e66 mm: memcg: lockde... |
1413 1414 1415 1416 1417 |
#ifdef CONFIG_LOCKDEP static struct lockdep_map memcg_oom_lock_dep_map = { .name = "memcg_oom_lock", }; #endif |
fb2a6fc56 mm: memcg: rework... |
1418 |
static DEFINE_SPINLOCK(memcg_oom_lock); |
867578cbc memcg: fix oom ki... |
1419 1420 1421 1422 |
/* * Check OOM-Killer is already running under our hierarchy. * If someone is running, return false. */ |
fb2a6fc56 mm: memcg: rework... |
1423 |
static bool mem_cgroup_oom_trylock(struct mem_cgroup *memcg) |
867578cbc memcg: fix oom ki... |
1424 |
{ |
79dfdaccd memcg: make oom_l... |
1425 |
struct mem_cgroup *iter, *failed = NULL; |
a636b327f memcg: avoid unne... |
1426 |
|
fb2a6fc56 mm: memcg: rework... |
1427 |
spin_lock(&memcg_oom_lock); |
9f3a0d093 mm: memcg: consol... |
1428 |
for_each_mem_cgroup_tree(iter, memcg) { |
23751be00 memcg: fix hierar... |
1429 |
if (iter->oom_lock) { |
79dfdaccd memcg: make oom_l... |
1430 1431 1432 1433 |
/* * this subtree of our hierarchy is already locked * so we cannot give a lock. */ |
79dfdaccd memcg: make oom_l... |
1434 |
failed = iter; |
9f3a0d093 mm: memcg: consol... |
1435 1436 |
mem_cgroup_iter_break(memcg, iter); break; |
23751be00 memcg: fix hierar... |
1437 1438 |
} else iter->oom_lock = true; |
7d74b06f2 memcg: use for_ea... |
1439 |
} |
867578cbc memcg: fix oom ki... |
1440 |
|
fb2a6fc56 mm: memcg: rework... |
1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 |
if (failed) { /* * OK, we failed to lock the whole subtree so we have * to clean up what we set up to the failing subtree */ for_each_mem_cgroup_tree(iter, memcg) { if (iter == failed) { mem_cgroup_iter_break(memcg, iter); break; } iter->oom_lock = false; |
79dfdaccd memcg: make oom_l... |
1452 |
} |
0056f4e66 mm: memcg: lockde... |
1453 1454 |
} else mutex_acquire(&memcg_oom_lock_dep_map, 0, 1, _RET_IP_); |
fb2a6fc56 mm: memcg: rework... |
1455 1456 1457 1458 |
spin_unlock(&memcg_oom_lock); return !failed; |
a636b327f memcg: avoid unne... |
1459 |
} |
0b7f569e4 memcg: fix OOM ki... |
1460 |
|
fb2a6fc56 mm: memcg: rework... |
1461 |
static void mem_cgroup_oom_unlock(struct mem_cgroup *memcg) |
0b7f569e4 memcg: fix OOM ki... |
1462 |
{ |
7d74b06f2 memcg: use for_ea... |
1463 |
struct mem_cgroup *iter; |
fb2a6fc56 mm: memcg: rework... |
1464 |
spin_lock(&memcg_oom_lock); |
0056f4e66 mm: memcg: lockde... |
1465 |
mutex_release(&memcg_oom_lock_dep_map, 1, _RET_IP_); |
c0ff4b854 memcg: rename mem... |
1466 |
for_each_mem_cgroup_tree(iter, memcg) |
79dfdaccd memcg: make oom_l... |
1467 |
iter->oom_lock = false; |
fb2a6fc56 mm: memcg: rework... |
1468 |
spin_unlock(&memcg_oom_lock); |
79dfdaccd memcg: make oom_l... |
1469 |
} |
c0ff4b854 memcg: rename mem... |
1470 |
static void mem_cgroup_mark_under_oom(struct mem_cgroup *memcg) |
79dfdaccd memcg: make oom_l... |
1471 1472 |
{ struct mem_cgroup *iter; |
c2b42d3ca memcg: convert me... |
1473 |
spin_lock(&memcg_oom_lock); |
c0ff4b854 memcg: rename mem... |
1474 |
for_each_mem_cgroup_tree(iter, memcg) |
c2b42d3ca memcg: convert me... |
1475 1476 |
iter->under_oom++; spin_unlock(&memcg_oom_lock); |
79dfdaccd memcg: make oom_l... |
1477 |
} |
c0ff4b854 memcg: rename mem... |
1478 |
static void mem_cgroup_unmark_under_oom(struct mem_cgroup *memcg) |
79dfdaccd memcg: make oom_l... |
1479 1480 |
{ struct mem_cgroup *iter; |
867578cbc memcg: fix oom ki... |
1481 1482 |
/* * When a new child is created while the hierarchy is under oom, |
c2b42d3ca memcg: convert me... |
1483 |
* mem_cgroup_oom_lock() may not be called. Watch for underflow. |
867578cbc memcg: fix oom ki... |
1484 |
*/ |
c2b42d3ca memcg: convert me... |
1485 |
spin_lock(&memcg_oom_lock); |
c0ff4b854 memcg: rename mem... |
1486 |
for_each_mem_cgroup_tree(iter, memcg) |
c2b42d3ca memcg: convert me... |
1487 1488 1489 |
if (iter->under_oom > 0) iter->under_oom--; spin_unlock(&memcg_oom_lock); |
0b7f569e4 memcg: fix OOM ki... |
1490 |
} |
867578cbc memcg: fix oom ki... |
1491 |
static DECLARE_WAIT_QUEUE_HEAD(memcg_oom_waitq); |
dc98df5a1 memcg: oom wakeup... |
1492 |
struct oom_wait_info { |
d79154bb5 memcg: replace me... |
1493 |
struct mem_cgroup *memcg; |
dc98df5a1 memcg: oom wakeup... |
1494 1495 1496 1497 1498 1499 |
wait_queue_t wait; }; static int memcg_oom_wake_function(wait_queue_t *wait, unsigned mode, int sync, void *arg) { |
d79154bb5 memcg: replace me... |
1500 1501 |
struct mem_cgroup *wake_memcg = (struct mem_cgroup *)arg; struct mem_cgroup *oom_wait_memcg; |
dc98df5a1 memcg: oom wakeup... |
1502 1503 1504 |
struct oom_wait_info *oom_wait_info; oom_wait_info = container_of(wait, struct oom_wait_info, wait); |
d79154bb5 memcg: replace me... |
1505 |
oom_wait_memcg = oom_wait_info->memcg; |
dc98df5a1 memcg: oom wakeup... |
1506 |
|
2314b42db mm: memcontrol: d... |
1507 1508 |
if (!mem_cgroup_is_descendant(wake_memcg, oom_wait_memcg) && !mem_cgroup_is_descendant(oom_wait_memcg, wake_memcg)) |
dc98df5a1 memcg: oom wakeup... |
1509 |
return 0; |
dc98df5a1 memcg: oom wakeup... |
1510 1511 |
return autoremove_wake_function(wait, mode, sync, arg); } |
c0ff4b854 memcg: rename mem... |
1512 |
static void memcg_oom_recover(struct mem_cgroup *memcg) |
3c11ecf44 memcg: oom kill d... |
1513 |
{ |
c2b42d3ca memcg: convert me... |
1514 1515 1516 1517 1518 1519 1520 1521 1522 |
/* * For the following lockless ->under_oom test, the only required * guarantee is that it must see the state asserted by an OOM when * this function is called as a result of userland actions * triggered by the notification of the OOM. This is trivially * achieved by invoking mem_cgroup_mark_under_oom() before * triggering notification. */ if (memcg && memcg->under_oom) |
f4b90b70b memcg: remove unu... |
1523 |
__wake_up(&memcg_oom_waitq, TASK_NORMAL, 0, memcg); |
3c11ecf44 memcg: oom kill d... |
1524 |
} |
3812c8c8f mm: memcg: do not... |
1525 |
static void mem_cgroup_oom(struct mem_cgroup *memcg, gfp_t mask, int order) |
0b7f569e4 memcg: fix OOM ki... |
1526 |
{ |
1383399d7 mm: memcontrol: f... |
1527 |
if (!current->memcg_may_oom || current->memcg_in_oom) |
3812c8c8f mm: memcg: do not... |
1528 |
return; |
867578cbc memcg: fix oom ki... |
1529 |
/* |
494264208 mm: memcg: handle... |
1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 |
* We are in the middle of the charge context here, so we * don't want to block when potentially sitting on a callstack * that holds all kinds of filesystem and mm locks. * * Also, the caller may handle a failed allocation gracefully * (like optional page cache readahead) and so an OOM killer * invocation might not even be necessary. * * That's why we don't do anything here except remember the * OOM context and then deal with it at the end of the page * fault when the stack is unwound, the locks are released, * and when we know whether the fault was overall successful. |
867578cbc memcg: fix oom ki... |
1542 |
*/ |
494264208 mm: memcg: handle... |
1543 |
css_get(&memcg->css); |
626ebc410 memcg: flatten ta... |
1544 1545 1546 |
current->memcg_in_oom = memcg; current->memcg_oom_gfp_mask = mask; current->memcg_oom_order = order; |
3812c8c8f mm: memcg: do not... |
1547 1548 1549 1550 |
} /** * mem_cgroup_oom_synchronize - complete memcg OOM handling |
494264208 mm: memcg: handle... |
1551 |
* @handle: actually kill/wait or just clean up the OOM state |
3812c8c8f mm: memcg: do not... |
1552 |
* |
494264208 mm: memcg: handle... |
1553 1554 |
* This has to be called at the end of a page fault if the memcg OOM * handler was enabled. |
3812c8c8f mm: memcg: do not... |
1555 |
* |
494264208 mm: memcg: handle... |
1556 |
* Memcg supports userspace OOM handling where failed allocations must |
3812c8c8f mm: memcg: do not... |
1557 1558 1559 1560 |
* sleep on a waitqueue until the userspace task resolves the * situation. Sleeping directly in the charge context with all kinds * of locks held is not a good idea, instead we remember an OOM state * in the task and mem_cgroup_oom_synchronize() has to be called at |
494264208 mm: memcg: handle... |
1561 |
* the end of the page fault to complete the OOM handling. |
3812c8c8f mm: memcg: do not... |
1562 1563 |
* * Returns %true if an ongoing memcg OOM situation was detected and |
494264208 mm: memcg: handle... |
1564 |
* completed, %false otherwise. |
3812c8c8f mm: memcg: do not... |
1565 |
*/ |
494264208 mm: memcg: handle... |
1566 |
bool mem_cgroup_oom_synchronize(bool handle) |
3812c8c8f mm: memcg: do not... |
1567 |
{ |
626ebc410 memcg: flatten ta... |
1568 |
struct mem_cgroup *memcg = current->memcg_in_oom; |
3812c8c8f mm: memcg: do not... |
1569 |
struct oom_wait_info owait; |
494264208 mm: memcg: handle... |
1570 |
bool locked; |
3812c8c8f mm: memcg: do not... |
1571 1572 |
/* OOM is global, do not handle */ |
3812c8c8f mm: memcg: do not... |
1573 |
if (!memcg) |
494264208 mm: memcg: handle... |
1574 |
return false; |
3812c8c8f mm: memcg: do not... |
1575 |
|
c32b3cbe0 oom, PM: make OOM... |
1576 |
if (!handle || oom_killer_disabled) |
494264208 mm: memcg: handle... |
1577 |
goto cleanup; |
3812c8c8f mm: memcg: do not... |
1578 1579 1580 1581 1582 1583 |
owait.memcg = memcg; owait.wait.flags = 0; owait.wait.func = memcg_oom_wake_function; owait.wait.private = current; INIT_LIST_HEAD(&owait.wait.task_list); |
867578cbc memcg: fix oom ki... |
1584 |
|
3812c8c8f mm: memcg: do not... |
1585 |
prepare_to_wait(&memcg_oom_waitq, &owait.wait, TASK_KILLABLE); |
494264208 mm: memcg: handle... |
1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 |
mem_cgroup_mark_under_oom(memcg); locked = mem_cgroup_oom_trylock(memcg); if (locked) mem_cgroup_oom_notify(memcg); if (locked && !memcg->oom_kill_disable) { mem_cgroup_unmark_under_oom(memcg); finish_wait(&memcg_oom_waitq, &owait.wait); |
626ebc410 memcg: flatten ta... |
1596 1597 |
mem_cgroup_out_of_memory(memcg, current->memcg_oom_gfp_mask, current->memcg_oom_order); |
494264208 mm: memcg: handle... |
1598 |
} else { |
3812c8c8f mm: memcg: do not... |
1599 |
schedule(); |
494264208 mm: memcg: handle... |
1600 1601 1602 1603 1604 |
mem_cgroup_unmark_under_oom(memcg); finish_wait(&memcg_oom_waitq, &owait.wait); } if (locked) { |
fb2a6fc56 mm: memcg: rework... |
1605 1606 1607 1608 1609 1610 1611 1612 |
mem_cgroup_oom_unlock(memcg); /* * There is no guarantee that an OOM-lock contender * sees the wakeups triggered by the OOM kill * uncharges. Wake any sleepers explicitely. */ memcg_oom_recover(memcg); } |
494264208 mm: memcg: handle... |
1613 |
cleanup: |
626ebc410 memcg: flatten ta... |
1614 |
current->memcg_in_oom = NULL; |
3812c8c8f mm: memcg: do not... |
1615 |
css_put(&memcg->css); |
867578cbc memcg: fix oom ki... |
1616 |
return true; |
0b7f569e4 memcg: fix OOM ki... |
1617 |
} |
d7365e783 mm: memcontrol: f... |
1618 |
/** |
81f8c3a46 mm: memcontrol: g... |
1619 1620 |
* lock_page_memcg - lock a page->mem_cgroup binding * @page: the page |
32047e2a8 memcg: avoid lock... |
1621 |
* |
81f8c3a46 mm: memcontrol: g... |
1622 1623 |
* This function protects unlocked LRU pages from being moved to * another cgroup and stabilizes their page->mem_cgroup binding. |
d69b042f3 memcg: add file-b... |
1624 |
*/ |
62cccb8c8 mm: simplify lock... |
1625 |
void lock_page_memcg(struct page *page) |
89c06bd52 memcg: use new lo... |
1626 1627 |
{ struct mem_cgroup *memcg; |
6de226191 mm: memcontrol: t... |
1628 |
unsigned long flags; |
89c06bd52 memcg: use new lo... |
1629 |
|
6de226191 mm: memcontrol: t... |
1630 1631 1632 1633 |
/* * The RCU lock is held throughout the transaction. The fast * path can get away without acquiring the memcg->move_lock * because page moving starts with an RCU grace period. |
6de226191 mm: memcontrol: t... |
1634 |
*/ |
d7365e783 mm: memcontrol: f... |
1635 1636 1637 |
rcu_read_lock(); if (mem_cgroup_disabled()) |
62cccb8c8 mm: simplify lock... |
1638 |
return; |
89c06bd52 memcg: use new lo... |
1639 |
again: |
1306a85ae mm: embed the mem... |
1640 |
memcg = page->mem_cgroup; |
298333157 mm: memcontrol: r... |
1641 |
if (unlikely(!memcg)) |
62cccb8c8 mm: simplify lock... |
1642 |
return; |
d7365e783 mm: memcontrol: f... |
1643 |
|
bdcbb659f memcg: fold mem_c... |
1644 |
if (atomic_read(&memcg->moving_account) <= 0) |
62cccb8c8 mm: simplify lock... |
1645 |
return; |
89c06bd52 memcg: use new lo... |
1646 |
|
6de226191 mm: memcontrol: t... |
1647 |
spin_lock_irqsave(&memcg->move_lock, flags); |
1306a85ae mm: embed the mem... |
1648 |
if (memcg != page->mem_cgroup) { |
6de226191 mm: memcontrol: t... |
1649 |
spin_unlock_irqrestore(&memcg->move_lock, flags); |
89c06bd52 memcg: use new lo... |
1650 1651 |
goto again; } |
6de226191 mm: memcontrol: t... |
1652 1653 1654 1655 |
/* * When charge migration first begins, we can have locked and * unlocked page stat updates happening concurrently. Track |
81f8c3a46 mm: memcontrol: g... |
1656 |
* the task who has the lock for unlock_page_memcg(). |
6de226191 mm: memcontrol: t... |
1657 1658 1659 |
*/ memcg->move_lock_task = current; memcg->move_lock_flags = flags; |
d7365e783 mm: memcontrol: f... |
1660 |
|
62cccb8c8 mm: simplify lock... |
1661 |
return; |
89c06bd52 memcg: use new lo... |
1662 |
} |
81f8c3a46 mm: memcontrol: g... |
1663 |
EXPORT_SYMBOL(lock_page_memcg); |
89c06bd52 memcg: use new lo... |
1664 |
|
d7365e783 mm: memcontrol: f... |
1665 |
/** |
81f8c3a46 mm: memcontrol: g... |
1666 |
* unlock_page_memcg - unlock a page->mem_cgroup binding |
62cccb8c8 mm: simplify lock... |
1667 |
* @page: the page |
d7365e783 mm: memcontrol: f... |
1668 |
*/ |
62cccb8c8 mm: simplify lock... |
1669 |
void unlock_page_memcg(struct page *page) |
89c06bd52 memcg: use new lo... |
1670 |
{ |
62cccb8c8 mm: simplify lock... |
1671 |
struct mem_cgroup *memcg = page->mem_cgroup; |
6de226191 mm: memcontrol: t... |
1672 1673 1674 1675 1676 1677 1678 1679 |
if (memcg && memcg->move_lock_task == current) { unsigned long flags = memcg->move_lock_flags; memcg->move_lock_task = NULL; memcg->move_lock_flags = 0; spin_unlock_irqrestore(&memcg->move_lock, flags); } |
89c06bd52 memcg: use new lo... |
1680 |
|
d7365e783 mm: memcontrol: f... |
1681 |
rcu_read_unlock(); |
89c06bd52 memcg: use new lo... |
1682 |
} |
81f8c3a46 mm: memcontrol: g... |
1683 |
EXPORT_SYMBOL(unlock_page_memcg); |
89c06bd52 memcg: use new lo... |
1684 |
|
f817ed485 memcg: move all a... |
1685 |
/* |
cdec2e426 memcg: coalesce c... |
1686 1687 1688 |
* 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... |
1689 |
#define CHARGE_BATCH 32U |
cdec2e426 memcg: coalesce c... |
1690 1691 |
struct memcg_stock_pcp { struct mem_cgroup *cached; /* this never be root cgroup */ |
11c9ea4e8 memcg: convert pe... |
1692 |
unsigned int nr_pages; |
cdec2e426 memcg: coalesce c... |
1693 |
struct work_struct work; |
26fe61684 memcg: fix percpu... |
1694 |
unsigned long flags; |
a0db00fcf memcg: remove red... |
1695 |
#define FLUSHING_CACHED_CHARGE 0 |
cdec2e426 memcg: coalesce c... |
1696 1697 |
}; static DEFINE_PER_CPU(struct memcg_stock_pcp, memcg_stock); |
9f50fad65 Revert "memcg: ge... |
1698 |
static DEFINE_MUTEX(percpu_charge_mutex); |
cdec2e426 memcg: coalesce c... |
1699 |
|
a0956d544 memcg: make it po... |
1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 |
/** * consume_stock: Try to consume stocked charge on this cpu. * @memcg: memcg to consume from. * @nr_pages: how many pages to charge. * * The charges will only happen if @memcg matches the current cpu's memcg * stock, and at least @nr_pages are available in that stock. Failure to * service an allocation will refill the stock. * * returns true if successful, false otherwise. |
cdec2e426 memcg: coalesce c... |
1710 |
*/ |
a0956d544 memcg: make it po... |
1711 |
static bool consume_stock(struct mem_cgroup *memcg, unsigned int nr_pages) |
cdec2e426 memcg: coalesce c... |
1712 1713 |
{ struct memcg_stock_pcp *stock; |
3e32cb2e0 mm: memcontrol: l... |
1714 |
bool ret = false; |
cdec2e426 memcg: coalesce c... |
1715 |
|
a0956d544 memcg: make it po... |
1716 |
if (nr_pages > CHARGE_BATCH) |
3e32cb2e0 mm: memcontrol: l... |
1717 |
return ret; |
a0956d544 memcg: make it po... |
1718 |
|
cdec2e426 memcg: coalesce c... |
1719 |
stock = &get_cpu_var(memcg_stock); |
3e32cb2e0 mm: memcontrol: l... |
1720 |
if (memcg == stock->cached && stock->nr_pages >= nr_pages) { |
a0956d544 memcg: make it po... |
1721 |
stock->nr_pages -= nr_pages; |
3e32cb2e0 mm: memcontrol: l... |
1722 1723 |
ret = true; } |
cdec2e426 memcg: coalesce c... |
1724 1725 1726 1727 1728 |
put_cpu_var(memcg_stock); return ret; } /* |
3e32cb2e0 mm: memcontrol: l... |
1729 |
* Returns stocks cached in percpu and reset cached information. |
cdec2e426 memcg: coalesce c... |
1730 1731 1732 1733 |
*/ static void drain_stock(struct memcg_stock_pcp *stock) { struct mem_cgroup *old = stock->cached; |
11c9ea4e8 memcg: convert pe... |
1734 |
if (stock->nr_pages) { |
3e32cb2e0 mm: memcontrol: l... |
1735 |
page_counter_uncharge(&old->memory, stock->nr_pages); |
7941d2145 mm: memcontrol: d... |
1736 |
if (do_memsw_account()) |
3e32cb2e0 mm: memcontrol: l... |
1737 |
page_counter_uncharge(&old->memsw, stock->nr_pages); |
e8ea14cc6 mm: memcontrol: t... |
1738 |
css_put_many(&old->css, stock->nr_pages); |
11c9ea4e8 memcg: convert pe... |
1739 |
stock->nr_pages = 0; |
cdec2e426 memcg: coalesce c... |
1740 1741 |
} stock->cached = NULL; |
cdec2e426 memcg: coalesce c... |
1742 1743 1744 1745 1746 1747 1748 1749 |
} /* * This must be called under preempt disabled or must be called by * a thread which is pinned to local cpu. */ static void drain_local_stock(struct work_struct *dummy) { |
7c8e0181e mm: replace __get... |
1750 |
struct memcg_stock_pcp *stock = this_cpu_ptr(&memcg_stock); |
cdec2e426 memcg: coalesce c... |
1751 |
drain_stock(stock); |
26fe61684 memcg: fix percpu... |
1752 |
clear_bit(FLUSHING_CACHED_CHARGE, &stock->flags); |
cdec2e426 memcg: coalesce c... |
1753 1754 1755 |
} /* |
3e32cb2e0 mm: memcontrol: l... |
1756 |
* Cache charges(val) to local per_cpu area. |
320cc51d9 mm: fix typo in r... |
1757 |
* This will be consumed by consume_stock() function, later. |
cdec2e426 memcg: coalesce c... |
1758 |
*/ |
c0ff4b854 memcg: rename mem... |
1759 |
static void refill_stock(struct mem_cgroup *memcg, unsigned int nr_pages) |
cdec2e426 memcg: coalesce c... |
1760 1761 |
{ struct memcg_stock_pcp *stock = &get_cpu_var(memcg_stock); |
c0ff4b854 memcg: rename mem... |
1762 |
if (stock->cached != memcg) { /* reset if necessary */ |
cdec2e426 memcg: coalesce c... |
1763 |
drain_stock(stock); |
c0ff4b854 memcg: rename mem... |
1764 |
stock->cached = memcg; |
cdec2e426 memcg: coalesce c... |
1765 |
} |
11c9ea4e8 memcg: convert pe... |
1766 |
stock->nr_pages += nr_pages; |
cdec2e426 memcg: coalesce c... |
1767 1768 1769 1770 |
put_cpu_var(memcg_stock); } /* |
c0ff4b854 memcg: rename mem... |
1771 |
* Drains all per-CPU charge caches for given root_memcg resp. subtree |
6d3d6aa22 mm: memcontrol: r... |
1772 |
* of the hierarchy under it. |
cdec2e426 memcg: coalesce c... |
1773 |
*/ |
6d3d6aa22 mm: memcontrol: r... |
1774 |
static void drain_all_stock(struct mem_cgroup *root_memcg) |
cdec2e426 memcg: coalesce c... |
1775 |
{ |
26fe61684 memcg: fix percpu... |
1776 |
int cpu, curcpu; |
d38144b7a memcg: unify sync... |
1777 |
|
6d3d6aa22 mm: memcontrol: r... |
1778 1779 1780 |
/* If someone's already draining, avoid adding running more workers. */ if (!mutex_trylock(&percpu_charge_mutex)) return; |
cdec2e426 memcg: coalesce c... |
1781 |
/* Notify other cpus that system-wide "drain" is running */ |
cdec2e426 memcg: coalesce c... |
1782 |
get_online_cpus(); |
5af12d0ef memcg: pin execut... |
1783 |
curcpu = get_cpu(); |
cdec2e426 memcg: coalesce c... |
1784 1785 |
for_each_online_cpu(cpu) { struct memcg_stock_pcp *stock = &per_cpu(memcg_stock, cpu); |
c0ff4b854 memcg: rename mem... |
1786 |
struct mem_cgroup *memcg; |
26fe61684 memcg: fix percpu... |
1787 |
|
c0ff4b854 memcg: rename mem... |
1788 1789 |
memcg = stock->cached; if (!memcg || !stock->nr_pages) |
26fe61684 memcg: fix percpu... |
1790 |
continue; |
2314b42db mm: memcontrol: d... |
1791 |
if (!mem_cgroup_is_descendant(memcg, root_memcg)) |
3e92041d6 memcg: add mem_cg... |
1792 |
continue; |
d1a05b697 memcg: do not try... |
1793 1794 1795 1796 1797 1798 |
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... |
1799 |
} |
5af12d0ef memcg: pin execut... |
1800 |
put_cpu(); |
f894ffa86 memcg: trivial cl... |
1801 |
put_online_cpus(); |
9f50fad65 Revert "memcg: ge... |
1802 |
mutex_unlock(&percpu_charge_mutex); |
cdec2e426 memcg: coalesce c... |
1803 |
} |
0db0628d9 kernel: delete __... |
1804 |
static int memcg_cpu_hotplug_callback(struct notifier_block *nb, |
cdec2e426 memcg: coalesce c... |
1805 1806 1807 1808 1809 |
unsigned long action, void *hcpu) { int cpu = (unsigned long)hcpu; struct memcg_stock_pcp *stock; |
619d094b5 memcg: simplify m... |
1810 |
if (action == CPU_ONLINE) |
1489ebad8 memcg: cpu hotplu... |
1811 |
return NOTIFY_OK; |
1489ebad8 memcg: cpu hotplu... |
1812 |
|
d833049bd memcg: fix broken... |
1813 |
if (action != CPU_DEAD && action != CPU_DEAD_FROZEN) |
cdec2e426 memcg: coalesce c... |
1814 |
return NOTIFY_OK; |
711d3d2c9 memcg: cpu hotplu... |
1815 |
|
cdec2e426 memcg: coalesce c... |
1816 1817 1818 1819 |
stock = &per_cpu(memcg_stock, cpu); drain_stock(stock); return NOTIFY_OK; } |
f7e1cb6ec mm: memcontrol: a... |
1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 |
static void reclaim_high(struct mem_cgroup *memcg, unsigned int nr_pages, gfp_t gfp_mask) { do { if (page_counter_read(&memcg->memory) <= memcg->high) continue; mem_cgroup_events(memcg, MEMCG_HIGH, 1); try_to_free_mem_cgroup_pages(memcg, nr_pages, gfp_mask, true); } while ((memcg = parent_mem_cgroup(memcg))); } static void high_work_func(struct work_struct *work) { struct mem_cgroup *memcg; memcg = container_of(work, struct mem_cgroup, high_work); reclaim_high(memcg, CHARGE_BATCH, GFP_KERNEL); } |
b23afb93d memcg: punt high ... |
1839 1840 1841 1842 1843 1844 1845 |
/* * Scheduled by try_charge() to be executed from the userland return path * and reclaims memory over the high limit. */ void mem_cgroup_handle_over_high(void) { unsigned int nr_pages = current->memcg_nr_pages_over_high; |
f7e1cb6ec mm: memcontrol: a... |
1846 |
struct mem_cgroup *memcg; |
b23afb93d memcg: punt high ... |
1847 1848 1849 |
if (likely(!nr_pages)) return; |
f7e1cb6ec mm: memcontrol: a... |
1850 1851 |
memcg = get_mem_cgroup_from_mm(current->mm); reclaim_high(memcg, nr_pages, GFP_KERNEL); |
b23afb93d memcg: punt high ... |
1852 1853 1854 |
css_put(&memcg->css); current->memcg_nr_pages_over_high = 0; } |
00501b531 mm: memcontrol: r... |
1855 1856 |
static int try_charge(struct mem_cgroup *memcg, gfp_t gfp_mask, unsigned int nr_pages) |
8a9f3ccd2 Memory controller... |
1857 |
{ |
7ec99d621 memcg: unify char... |
1858 |
unsigned int batch = max(CHARGE_BATCH, nr_pages); |
9b1306192 mm: memcontrol: r... |
1859 |
int nr_retries = MEM_CGROUP_RECLAIM_RETRIES; |
6539cc053 mm: memcontrol: f... |
1860 |
struct mem_cgroup *mem_over_limit; |
3e32cb2e0 mm: memcontrol: l... |
1861 |
struct page_counter *counter; |
6539cc053 mm: memcontrol: f... |
1862 |
unsigned long nr_reclaimed; |
b70a2a21d mm: memcontrol: f... |
1863 1864 |
bool may_swap = true; bool drained = false; |
a636b327f memcg: avoid unne... |
1865 |
|
ce00a9673 mm: memcontrol: r... |
1866 |
if (mem_cgroup_is_root(memcg)) |
10d53c748 memcg: ratify and... |
1867 |
return 0; |
6539cc053 mm: memcontrol: f... |
1868 |
retry: |
b6b6cc72b memcg: do not rep... |
1869 |
if (consume_stock(memcg, nr_pages)) |
10d53c748 memcg: ratify and... |
1870 |
return 0; |
8a9f3ccd2 Memory controller... |
1871 |
|
7941d2145 mm: memcontrol: d... |
1872 |
if (!do_memsw_account() || |
6071ca520 mm: page_counter:... |
1873 1874 |
page_counter_try_charge(&memcg->memsw, batch, &counter)) { if (page_counter_try_charge(&memcg->memory, batch, &counter)) |
6539cc053 mm: memcontrol: f... |
1875 |
goto done_restock; |
7941d2145 mm: memcontrol: d... |
1876 |
if (do_memsw_account()) |
3e32cb2e0 mm: memcontrol: l... |
1877 1878 |
page_counter_uncharge(&memcg->memsw, batch); mem_over_limit = mem_cgroup_from_counter(counter, memory); |
3fbe72442 mm: memcontrol: s... |
1879 |
} else { |
3e32cb2e0 mm: memcontrol: l... |
1880 |
mem_over_limit = mem_cgroup_from_counter(counter, memsw); |
b70a2a21d mm: memcontrol: f... |
1881 |
may_swap = false; |
3fbe72442 mm: memcontrol: s... |
1882 |
} |
7a81b88cb memcg: introduce ... |
1883 |
|
6539cc053 mm: memcontrol: f... |
1884 1885 1886 1887 |
if (batch > nr_pages) { batch = nr_pages; goto retry; } |
6d61ef409 memcg: memory cgr... |
1888 |
|
06b078fc0 mm: memcontrol: r... |
1889 1890 1891 1892 1893 1894 1895 1896 1897 |
/* * Unlike in global OOM situations, memcg is not in a physical * memory shortage. Allow dying and OOM-killed tasks to * bypass the last charges so that they can exit quickly and * free their memory. */ if (unlikely(test_thread_flag(TIF_MEMDIE) || fatal_signal_pending(current) || current->flags & PF_EXITING)) |
10d53c748 memcg: ratify and... |
1898 |
goto force; |
06b078fc0 mm: memcontrol: r... |
1899 1900 1901 |
if (unlikely(task_in_memcg_oom(current))) goto nomem; |
d0164adc8 mm, page_alloc: d... |
1902 |
if (!gfpflags_allow_blocking(gfp_mask)) |
6539cc053 mm: memcontrol: f... |
1903 |
goto nomem; |
4b5343346 memcg: clean up t... |
1904 |
|
241994ed8 mm: memcontrol: d... |
1905 |
mem_cgroup_events(mem_over_limit, MEMCG_MAX, 1); |
b70a2a21d mm: memcontrol: f... |
1906 1907 |
nr_reclaimed = try_to_free_mem_cgroup_pages(mem_over_limit, nr_pages, gfp_mask, may_swap); |
6539cc053 mm: memcontrol: f... |
1908 |
|
61e02c745 mm: memcontrol: c... |
1909 |
if (mem_cgroup_margin(mem_over_limit) >= nr_pages) |
6539cc053 mm: memcontrol: f... |
1910 |
goto retry; |
28c34c291 mm: memcontrol: r... |
1911 |
|
b70a2a21d mm: memcontrol: f... |
1912 |
if (!drained) { |
6d3d6aa22 mm: memcontrol: r... |
1913 |
drain_all_stock(mem_over_limit); |
b70a2a21d mm: memcontrol: f... |
1914 1915 1916 |
drained = true; goto retry; } |
28c34c291 mm: memcontrol: r... |
1917 1918 |
if (gfp_mask & __GFP_NORETRY) goto nomem; |
6539cc053 mm: memcontrol: f... |
1919 1920 1921 1922 1923 1924 1925 1926 1927 |
/* * Even though the limit is exceeded at this point, reclaim * may have been able to free some pages. Retry the charge * before killing the task. * * Only for regular pages, though: huge pages are rather * unlikely to succeed so close to the limit, and we fall back * to regular pages anyway in case of failure. */ |
61e02c745 mm: memcontrol: c... |
1928 |
if (nr_reclaimed && nr_pages <= (1 << PAGE_ALLOC_COSTLY_ORDER)) |
6539cc053 mm: memcontrol: f... |
1929 1930 1931 1932 1933 1934 1935 |
goto retry; /* * At task move, charge accounts can be doubly counted. So, it's * better to wait until the end of task_move if something is going on. */ if (mem_cgroup_wait_acct_move(mem_over_limit)) goto retry; |
9b1306192 mm: memcontrol: r... |
1936 1937 |
if (nr_retries--) goto retry; |
06b078fc0 mm: memcontrol: r... |
1938 |
if (gfp_mask & __GFP_NOFAIL) |
10d53c748 memcg: ratify and... |
1939 |
goto force; |
06b078fc0 mm: memcontrol: r... |
1940 |
|
6539cc053 mm: memcontrol: f... |
1941 |
if (fatal_signal_pending(current)) |
10d53c748 memcg: ratify and... |
1942 |
goto force; |
6539cc053 mm: memcontrol: f... |
1943 |
|
241994ed8 mm: memcontrol: d... |
1944 |
mem_cgroup_events(mem_over_limit, MEMCG_OOM, 1); |
3608de078 mm/memcontrol.c: ... |
1945 1946 |
mem_cgroup_oom(mem_over_limit, gfp_mask, get_order(nr_pages * PAGE_SIZE)); |
7a81b88cb memcg: introduce ... |
1947 |
nomem: |
6d1fdc489 memcg: sanitize _... |
1948 |
if (!(gfp_mask & __GFP_NOFAIL)) |
3168ecbe1 mm: memcg: use pr... |
1949 |
return -ENOMEM; |
10d53c748 memcg: ratify and... |
1950 1951 1952 1953 1954 1955 1956 |
force: /* * The allocation either can't fail or will lead to more memory * being freed very soon. Allow memory usage go over the limit * temporarily by force charging it. */ page_counter_charge(&memcg->memory, nr_pages); |
7941d2145 mm: memcontrol: d... |
1957 |
if (do_memsw_account()) |
10d53c748 memcg: ratify and... |
1958 1959 1960 1961 |
page_counter_charge(&memcg->memsw, nr_pages); css_get_many(&memcg->css, nr_pages); return 0; |
6539cc053 mm: memcontrol: f... |
1962 1963 |
done_restock: |
e8ea14cc6 mm: memcontrol: t... |
1964 |
css_get_many(&memcg->css, batch); |
6539cc053 mm: memcontrol: f... |
1965 1966 |
if (batch > nr_pages) refill_stock(memcg, batch - nr_pages); |
b23afb93d memcg: punt high ... |
1967 |
|
241994ed8 mm: memcontrol: d... |
1968 |
/* |
b23afb93d memcg: punt high ... |
1969 1970 |
* If the hierarchy is above the normal consumption range, schedule * reclaim on returning to userland. We can perform reclaim here |
71baba4b9 mm, page_alloc: r... |
1971 |
* if __GFP_RECLAIM but let's always punt for simplicity and so that |
b23afb93d memcg: punt high ... |
1972 1973 1974 1975 |
* GFP_KERNEL can consistently be used during reclaim. @memcg is * not recorded as it most likely matches current's and won't * change in the meantime. As high limit is checked again before * reclaim, the cost of mismatch is negligible. |
241994ed8 mm: memcontrol: d... |
1976 1977 |
*/ do { |
b23afb93d memcg: punt high ... |
1978 |
if (page_counter_read(&memcg->memory) > memcg->high) { |
f7e1cb6ec mm: memcontrol: a... |
1979 1980 1981 1982 1983 |
/* Don't bother a random interrupted task */ if (in_interrupt()) { schedule_work(&memcg->high_work); break; } |
9516a18a9 memcg: fix memory... |
1984 |
current->memcg_nr_pages_over_high += batch; |
b23afb93d memcg: punt high ... |
1985 1986 1987 |
set_notify_resume(current); break; } |
241994ed8 mm: memcontrol: d... |
1988 |
} while ((memcg = parent_mem_cgroup(memcg))); |
10d53c748 memcg: ratify and... |
1989 1990 |
return 0; |
7a81b88cb memcg: introduce ... |
1991 |
} |
8a9f3ccd2 Memory controller... |
1992 |
|
00501b531 mm: memcontrol: r... |
1993 |
static void cancel_charge(struct mem_cgroup *memcg, unsigned int nr_pages) |
a3032a2c1 memcg: add mem_cg... |
1994 |
{ |
ce00a9673 mm: memcontrol: r... |
1995 1996 |
if (mem_cgroup_is_root(memcg)) return; |
3e32cb2e0 mm: memcontrol: l... |
1997 |
page_counter_uncharge(&memcg->memory, nr_pages); |
7941d2145 mm: memcontrol: d... |
1998 |
if (do_memsw_account()) |
3e32cb2e0 mm: memcontrol: l... |
1999 |
page_counter_uncharge(&memcg->memsw, nr_pages); |
ce00a9673 mm: memcontrol: r... |
2000 |
|
e8ea14cc6 mm: memcontrol: t... |
2001 |
css_put_many(&memcg->css, nr_pages); |
d01dd17f1 memcg: use res_co... |
2002 |
} |
0a31bc97c mm: memcontrol: r... |
2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 |
static void lock_page_lru(struct page *page, int *isolated) { struct zone *zone = page_zone(page); spin_lock_irq(&zone->lru_lock); if (PageLRU(page)) { struct lruvec *lruvec; lruvec = mem_cgroup_page_lruvec(page, zone); ClearPageLRU(page); del_page_from_lru_list(page, lruvec, page_lru(page)); *isolated = 1; } else *isolated = 0; } static void unlock_page_lru(struct page *page, int isolated) { struct zone *zone = page_zone(page); if (isolated) { struct lruvec *lruvec; lruvec = mem_cgroup_page_lruvec(page, zone); VM_BUG_ON_PAGE(PageLRU(page), page); SetPageLRU(page); add_page_to_lru_list(page, lruvec, page_lru(page)); } spin_unlock_irq(&zone->lru_lock); } |
00501b531 mm: memcontrol: r... |
2033 |
static void commit_charge(struct page *page, struct mem_cgroup *memcg, |
6abb5a867 mm: memcontrol: a... |
2034 |
bool lrucare) |
7a81b88cb memcg: introduce ... |
2035 |
{ |
0a31bc97c mm: memcontrol: r... |
2036 |
int isolated; |
9ce70c024 memcg: fix deadlo... |
2037 |
|
1306a85ae mm: embed the mem... |
2038 |
VM_BUG_ON_PAGE(page->mem_cgroup, page); |
9ce70c024 memcg: fix deadlo... |
2039 2040 2041 2042 2043 |
/* * In some cases, SwapCache and FUSE(splice_buf->radixtree), the page * may already be on some other mem_cgroup's LRU. Take care of it. */ |
0a31bc97c mm: memcontrol: r... |
2044 2045 |
if (lrucare) lock_page_lru(page, &isolated); |
9ce70c024 memcg: fix deadlo... |
2046 |
|
0a31bc97c mm: memcontrol: r... |
2047 2048 |
/* * Nobody should be changing or seriously looking at |
1306a85ae mm: embed the mem... |
2049 |
* page->mem_cgroup at this point: |
0a31bc97c mm: memcontrol: r... |
2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 |
* * - the page is uncharged * * - the page is off-LRU * * - an anonymous fault has exclusive page access, except for * a locked page table * * - a page cache insertion, a swapin fault, or a migration * have the page locked */ |
1306a85ae mm: embed the mem... |
2061 |
page->mem_cgroup = memcg; |
9ce70c024 memcg: fix deadlo... |
2062 |
|
0a31bc97c mm: memcontrol: r... |
2063 2064 |
if (lrucare) unlock_page_lru(page, isolated); |
7a81b88cb memcg: introduce ... |
2065 |
} |
66e1707bc Memory controller... |
2066 |
|
127424c86 mm: memcontrol: m... |
2067 |
#ifndef CONFIG_SLOB |
f3bb3043a memcg: don't call... |
2068 |
static int memcg_alloc_cache_id(void) |
55007d849 memcg: allocate m... |
2069 |
{ |
f3bb3043a memcg: don't call... |
2070 2071 |
int id, size; int err; |
dbcf73e26 memcg: rename som... |
2072 |
id = ida_simple_get(&memcg_cache_ida, |
f3bb3043a memcg: don't call... |
2073 2074 2075 |
0, MEMCG_CACHES_MAX_SIZE, GFP_KERNEL); if (id < 0) return id; |
55007d849 memcg: allocate m... |
2076 |
|
dbcf73e26 memcg: rename som... |
2077 |
if (id < memcg_nr_cache_ids) |
f3bb3043a memcg: don't call... |
2078 2079 2080 2081 2082 2083 |
return id; /* * There's no space for the new id in memcg_caches arrays, * so we have to grow them. */ |
05257a1a3 memcg: add rwsem ... |
2084 |
down_write(&memcg_cache_ids_sem); |
f3bb3043a memcg: don't call... |
2085 2086 |
size = 2 * (id + 1); |
55007d849 memcg: allocate m... |
2087 2088 2089 2090 |
if (size < MEMCG_CACHES_MIN_SIZE) size = MEMCG_CACHES_MIN_SIZE; else if (size > MEMCG_CACHES_MAX_SIZE) size = MEMCG_CACHES_MAX_SIZE; |
f3bb3043a memcg: don't call... |
2091 |
err = memcg_update_all_caches(size); |
05257a1a3 memcg: add rwsem ... |
2092 |
if (!err) |
60d3fd32a list_lru: introdu... |
2093 2094 |
err = memcg_update_all_list_lrus(size); if (!err) |
05257a1a3 memcg: add rwsem ... |
2095 2096 2097 |
memcg_nr_cache_ids = size; up_write(&memcg_cache_ids_sem); |
f3bb3043a memcg: don't call... |
2098 |
if (err) { |
dbcf73e26 memcg: rename som... |
2099 |
ida_simple_remove(&memcg_cache_ida, id); |
f3bb3043a memcg: don't call... |
2100 2101 2102 2103 2104 2105 2106 |
return err; } return id; } static void memcg_free_cache_id(int id) { |
dbcf73e26 memcg: rename som... |
2107 |
ida_simple_remove(&memcg_cache_ida, id); |
55007d849 memcg: allocate m... |
2108 |
} |
d5b3cf713 memcg: zap memcg_... |
2109 |
struct memcg_kmem_cache_create_work { |
5722d094a memcg, slab: clea... |
2110 2111 2112 2113 |
struct mem_cgroup *memcg; struct kmem_cache *cachep; struct work_struct work; }; |
d5b3cf713 memcg: zap memcg_... |
2114 |
static void memcg_kmem_cache_create_func(struct work_struct *w) |
d7f25f8a2 memcg: infrastruc... |
2115 |
{ |
d5b3cf713 memcg: zap memcg_... |
2116 2117 |
struct memcg_kmem_cache_create_work *cw = container_of(w, struct memcg_kmem_cache_create_work, work); |
5722d094a memcg, slab: clea... |
2118 2119 |
struct mem_cgroup *memcg = cw->memcg; struct kmem_cache *cachep = cw->cachep; |
d7f25f8a2 memcg: infrastruc... |
2120 |
|
d5b3cf713 memcg: zap memcg_... |
2121 |
memcg_create_kmem_cache(memcg, cachep); |
bd6731458 memcg, slab: simp... |
2122 |
|
5722d094a memcg, slab: clea... |
2123 |
css_put(&memcg->css); |
d7f25f8a2 memcg: infrastruc... |
2124 2125 2126 2127 2128 |
kfree(cw); } /* * Enqueue the creation of a per-memcg kmem_cache. |
d7f25f8a2 memcg: infrastruc... |
2129 |
*/ |
d5b3cf713 memcg: zap memcg_... |
2130 2131 |
static void __memcg_schedule_kmem_cache_create(struct mem_cgroup *memcg, struct kmem_cache *cachep) |
d7f25f8a2 memcg: infrastruc... |
2132 |
{ |
d5b3cf713 memcg: zap memcg_... |
2133 |
struct memcg_kmem_cache_create_work *cw; |
d7f25f8a2 memcg: infrastruc... |
2134 |
|
776ed0f03 memcg: cleanup km... |
2135 |
cw = kmalloc(sizeof(*cw), GFP_NOWAIT); |
8135be5a8 memcg: fix possib... |
2136 |
if (!cw) |
d7f25f8a2 memcg: infrastruc... |
2137 |
return; |
8135be5a8 memcg: fix possib... |
2138 2139 |
css_get(&memcg->css); |
d7f25f8a2 memcg: infrastruc... |
2140 2141 2142 |
cw->memcg = memcg; cw->cachep = cachep; |
d5b3cf713 memcg: zap memcg_... |
2143 |
INIT_WORK(&cw->work, memcg_kmem_cache_create_func); |
d7f25f8a2 memcg: infrastruc... |
2144 |
|
d7f25f8a2 memcg: infrastruc... |
2145 2146 |
schedule_work(&cw->work); } |
d5b3cf713 memcg: zap memcg_... |
2147 2148 |
static void memcg_schedule_kmem_cache_create(struct mem_cgroup *memcg, struct kmem_cache *cachep) |
0e9d92f2d memcg: skip memcg... |
2149 2150 2151 2152 |
{ /* * We need to stop accounting when we kmalloc, because if the * corresponding kmalloc cache is not yet created, the first allocation |
d5b3cf713 memcg: zap memcg_... |
2153 |
* in __memcg_schedule_kmem_cache_create will recurse. |
0e9d92f2d memcg: skip memcg... |
2154 2155 2156 2157 2158 2159 2160 |
* * However, it is better to enclose the whole function. Depending on * the debugging options enabled, INIT_WORK(), for instance, can * trigger an allocation. This too, will make us recurse. Because at * this point we can't allow ourselves back into memcg_kmem_get_cache, * the safest choice is to do it like this, wrapping the whole function. */ |
6f185c290 memcg: turn memcg... |
2161 |
current->memcg_kmem_skip_account = 1; |
d5b3cf713 memcg: zap memcg_... |
2162 |
__memcg_schedule_kmem_cache_create(memcg, cachep); |
6f185c290 memcg: turn memcg... |
2163 |
current->memcg_kmem_skip_account = 0; |
0e9d92f2d memcg: skip memcg... |
2164 |
} |
c67a8a685 memcg, slab: merg... |
2165 |
|
d7f25f8a2 memcg: infrastruc... |
2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 |
/* * Return the kmem_cache we're supposed to use for a slab allocation. * We try to use the current memcg's version of the cache. * * If the cache does not exist yet, if we are the first user of it, * we either create it immediately, if possible, or create it asynchronously * in a workqueue. * In the latter case, we will let the current allocation go through with * the original cache. * * Can't be called in interrupt context or from kernel threads. * This function needs to be called with rcu_read_lock() held. */ |
230e9fc28 slab: add SLAB_AC... |
2179 |
struct kmem_cache *__memcg_kmem_get_cache(struct kmem_cache *cachep, gfp_t gfp) |
d7f25f8a2 memcg: infrastruc... |
2180 2181 |
{ struct mem_cgroup *memcg; |
959c8963f memcg, slab: fix ... |
2182 |
struct kmem_cache *memcg_cachep; |
2a4db7eb9 memcg: free memcg... |
2183 |
int kmemcg_id; |
d7f25f8a2 memcg: infrastruc... |
2184 |
|
f7ce3190c slab: embed memcg... |
2185 |
VM_BUG_ON(!is_root_cache(cachep)); |
d7f25f8a2 memcg: infrastruc... |
2186 |
|
230e9fc28 slab: add SLAB_AC... |
2187 2188 2189 2190 2191 |
if (cachep->flags & SLAB_ACCOUNT) gfp |= __GFP_ACCOUNT; if (!(gfp & __GFP_ACCOUNT)) return cachep; |
9d100c5e4 memcg: don't chec... |
2192 |
if (current->memcg_kmem_skip_account) |
0e9d92f2d memcg: skip memcg... |
2193 |
return cachep; |
8135be5a8 memcg: fix possib... |
2194 |
memcg = get_mem_cgroup_from_mm(current->mm); |
4db0c3c29 mm: remove rest o... |
2195 |
kmemcg_id = READ_ONCE(memcg->kmemcg_id); |
2a4db7eb9 memcg: free memcg... |
2196 |
if (kmemcg_id < 0) |
ca0dde971 memcg: take refer... |
2197 |
goto out; |
d7f25f8a2 memcg: infrastruc... |
2198 |
|
2a4db7eb9 memcg: free memcg... |
2199 |
memcg_cachep = cache_from_memcg_idx(cachep, kmemcg_id); |
8135be5a8 memcg: fix possib... |
2200 2201 |
if (likely(memcg_cachep)) return memcg_cachep; |
ca0dde971 memcg: take refer... |
2202 2203 2204 2205 2206 2207 2208 2209 2210 |
/* * If we are in a safe context (can wait, and not in interrupt * context), we could be be predictable and return right away. * This would guarantee that the allocation being performed * already belongs in the new cache. * * However, there are some clashes that can arrive from locking. * For instance, because we acquire the slab_mutex while doing |
776ed0f03 memcg: cleanup km... |
2211 2212 2213 |
* memcg_create_kmem_cache, this means no further allocation * could happen with the slab_mutex held. So it's better to * defer everything. |
ca0dde971 memcg: take refer... |
2214 |
*/ |
d5b3cf713 memcg: zap memcg_... |
2215 |
memcg_schedule_kmem_cache_create(memcg, cachep); |
ca0dde971 memcg: take refer... |
2216 |
out: |
8135be5a8 memcg: fix possib... |
2217 |
css_put(&memcg->css); |
ca0dde971 memcg: take refer... |
2218 |
return cachep; |
d7f25f8a2 memcg: infrastruc... |
2219 |
} |
d7f25f8a2 memcg: infrastruc... |
2220 |
|
8135be5a8 memcg: fix possib... |
2221 2222 2223 |
void __memcg_kmem_put_cache(struct kmem_cache *cachep) { if (!is_root_cache(cachep)) |
f7ce3190c slab: embed memcg... |
2224 |
css_put(&cachep->memcg_params.memcg->css); |
8135be5a8 memcg: fix possib... |
2225 |
} |
f3ccb2c42 memcg: unify slab... |
2226 2227 |
int __memcg_kmem_charge_memcg(struct page *page, gfp_t gfp, int order, struct mem_cgroup *memcg) |
7ae1e1d0f memcg: kmem contr... |
2228 |
{ |
f3ccb2c42 memcg: unify slab... |
2229 2230 |
unsigned int nr_pages = 1 << order; struct page_counter *counter; |
7ae1e1d0f memcg: kmem contr... |
2231 |
int ret; |
f3ccb2c42 memcg: unify slab... |
2232 |
ret = try_charge(memcg, gfp, nr_pages); |
52c29b048 mm: memcontrol: a... |
2233 |
if (ret) |
f3ccb2c42 memcg: unify slab... |
2234 |
return ret; |
52c29b048 mm: memcontrol: a... |
2235 2236 2237 2238 2239 |
if (!cgroup_subsys_on_dfl(memory_cgrp_subsys) && !page_counter_try_charge(&memcg->kmem, nr_pages, &counter)) { cancel_charge(memcg, nr_pages); return -ENOMEM; |
7ae1e1d0f memcg: kmem contr... |
2240 |
} |
f3ccb2c42 memcg: unify slab... |
2241 |
page->mem_cgroup = memcg; |
7ae1e1d0f memcg: kmem contr... |
2242 |
|
f3ccb2c42 memcg: unify slab... |
2243 |
return 0; |
7ae1e1d0f memcg: kmem contr... |
2244 |
} |
f3ccb2c42 memcg: unify slab... |
2245 |
int __memcg_kmem_charge(struct page *page, gfp_t gfp, int order) |
7ae1e1d0f memcg: kmem contr... |
2246 |
{ |
f3ccb2c42 memcg: unify slab... |
2247 |
struct mem_cgroup *memcg; |
fcff7d7ee mm: memcontrol: d... |
2248 |
int ret = 0; |
7ae1e1d0f memcg: kmem contr... |
2249 |
|
f3ccb2c42 memcg: unify slab... |
2250 |
memcg = get_mem_cgroup_from_mm(current->mm); |
b6ecd2dea mm: memcontrol: z... |
2251 |
if (!mem_cgroup_is_root(memcg)) |
fcff7d7ee mm: memcontrol: d... |
2252 |
ret = __memcg_kmem_charge_memcg(page, gfp, order, memcg); |
7ae1e1d0f memcg: kmem contr... |
2253 |
css_put(&memcg->css); |
d05e83a6f memcg: simplify c... |
2254 |
return ret; |
7ae1e1d0f memcg: kmem contr... |
2255 |
} |
d05e83a6f memcg: simplify c... |
2256 |
void __memcg_kmem_uncharge(struct page *page, int order) |
7ae1e1d0f memcg: kmem contr... |
2257 |
{ |
1306a85ae mm: embed the mem... |
2258 |
struct mem_cgroup *memcg = page->mem_cgroup; |
f3ccb2c42 memcg: unify slab... |
2259 |
unsigned int nr_pages = 1 << order; |
7ae1e1d0f memcg: kmem contr... |
2260 |
|
7ae1e1d0f memcg: kmem contr... |
2261 2262 |
if (!memcg) return; |
309381fea mm: dump page whe... |
2263 |
VM_BUG_ON_PAGE(mem_cgroup_is_root(memcg), page); |
298333157 mm: memcontrol: r... |
2264 |
|
52c29b048 mm: memcontrol: a... |
2265 2266 |
if (!cgroup_subsys_on_dfl(memory_cgrp_subsys)) page_counter_uncharge(&memcg->kmem, nr_pages); |
f3ccb2c42 memcg: unify slab... |
2267 |
page_counter_uncharge(&memcg->memory, nr_pages); |
7941d2145 mm: memcontrol: d... |
2268 |
if (do_memsw_account()) |
f3ccb2c42 memcg: unify slab... |
2269 |
page_counter_uncharge(&memcg->memsw, nr_pages); |
60d3fd32a list_lru: introdu... |
2270 |
|
1306a85ae mm: embed the mem... |
2271 |
page->mem_cgroup = NULL; |
f3ccb2c42 memcg: unify slab... |
2272 |
css_put_many(&memcg->css, nr_pages); |
60d3fd32a list_lru: introdu... |
2273 |
} |
127424c86 mm: memcontrol: m... |
2274 |
#endif /* !CONFIG_SLOB */ |
7ae1e1d0f memcg: kmem contr... |
2275 |
|
ca3e02141 memcg: fix USED b... |
2276 |
#ifdef CONFIG_TRANSPARENT_HUGEPAGE |
ca3e02141 memcg: fix USED b... |
2277 2278 |
/* * Because tail pages are not marked as "used", set it. We're under |
3ac808fdd mm, thp: remove c... |
2279 |
* zone->lru_lock and migration entries setup in all page mappings. |
ca3e02141 memcg: fix USED b... |
2280 |
*/ |
e94c8a9cb memcg: make mem_c... |
2281 |
void mem_cgroup_split_huge_fixup(struct page *head) |
ca3e02141 memcg: fix USED b... |
2282 |
{ |
e94c8a9cb memcg: make mem_c... |
2283 |
int i; |
ca3e02141 memcg: fix USED b... |
2284 |
|
3d37c4a91 memcg: bugfix che... |
2285 2286 |
if (mem_cgroup_disabled()) return; |
b070e65c0 mm, memcg: add rs... |
2287 |
|
298333157 mm: memcontrol: r... |
2288 |
for (i = 1; i < HPAGE_PMD_NR; i++) |
1306a85ae mm: embed the mem... |
2289 |
head[i].mem_cgroup = head->mem_cgroup; |
b9982f8d2 mm: memcontrol: m... |
2290 |
|
1306a85ae mm: embed the mem... |
2291 |
__this_cpu_sub(head->mem_cgroup->stat->count[MEM_CGROUP_STAT_RSS_HUGE], |
b070e65c0 mm, memcg: add rs... |
2292 |
HPAGE_PMD_NR); |
ca3e02141 memcg: fix USED b... |
2293 |
} |
12d271078 memcg: fix split_... |
2294 |
#endif /* CONFIG_TRANSPARENT_HUGEPAGE */ |
ca3e02141 memcg: fix USED b... |
2295 |
|
c255a4580 memcg: rename con... |
2296 |
#ifdef CONFIG_MEMCG_SWAP |
0a31bc97c mm: memcontrol: r... |
2297 2298 |
static void mem_cgroup_swap_statistics(struct mem_cgroup *memcg, bool charge) |
d13d14430 memcg: handle swa... |
2299 |
{ |
0a31bc97c mm: memcontrol: r... |
2300 2301 |
int val = (charge) ? 1 : -1; this_cpu_add(memcg->stat->count[MEM_CGROUP_STAT_SWAP], val); |
d13d14430 memcg: handle swa... |
2302 |
} |
024914477 memcg: move charg... |
2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 |
/** * mem_cgroup_move_swap_account - move swap charge and swap_cgroup's record. * @entry: swap entry to be moved * @from: mem_cgroup which the entry is moved from * @to: mem_cgroup which the entry is moved to * * It succeeds only when the swap_cgroup's record for this entry is the same * as the mem_cgroup's id of @from. * * Returns 0 on success, -EINVAL on failure. * |
3e32cb2e0 mm: memcontrol: l... |
2315 |
* The caller must have charged to @to, IOW, called page_counter_charge() about |
024914477 memcg: move charg... |
2316 2317 2318 |
* both res and memsw, and called css_get(). */ static int mem_cgroup_move_swap_account(swp_entry_t entry, |
e91cbb425 memcg swap: mem_c... |
2319 |
struct mem_cgroup *from, struct mem_cgroup *to) |
024914477 memcg: move charg... |
2320 2321 |
{ unsigned short old_id, new_id; |
34c00c319 memcg: convert to... |
2322 2323 |
old_id = mem_cgroup_id(from); new_id = mem_cgroup_id(to); |
024914477 memcg: move charg... |
2324 2325 |
if (swap_cgroup_cmpxchg(entry, old_id, new_id) == old_id) { |
024914477 memcg: move charg... |
2326 |
mem_cgroup_swap_statistics(from, false); |
483c30b51 memcg: improve pe... |
2327 |
mem_cgroup_swap_statistics(to, true); |
024914477 memcg: move charg... |
2328 2329 2330 2331 2332 2333 |
return 0; } return -EINVAL; } #else static inline int mem_cgroup_move_swap_account(swp_entry_t entry, |
e91cbb425 memcg swap: mem_c... |
2334 |
struct mem_cgroup *from, struct mem_cgroup *to) |
024914477 memcg: move charg... |
2335 2336 2337 |
{ return -EINVAL; } |
8c7c6e34a memcg: mem+swap c... |
2338 |
#endif |
d13d14430 memcg: handle swa... |
2339 |
|
3e32cb2e0 mm: memcontrol: l... |
2340 |
static DEFINE_MUTEX(memcg_limit_mutex); |
f212ad7cf memcg: add memcg ... |
2341 |
|
d38d2a758 mm: make mem_cgro... |
2342 |
static int mem_cgroup_resize_limit(struct mem_cgroup *memcg, |
3e32cb2e0 mm: memcontrol: l... |
2343 |
unsigned long limit) |
628f42355 memcg: limit chan... |
2344 |
{ |
3e32cb2e0 mm: memcontrol: l... |
2345 2346 2347 |
unsigned long curusage; unsigned long oldusage; bool enlarge = false; |
81d39c20f memcg: fix shrink... |
2348 |
int retry_count; |
3e32cb2e0 mm: memcontrol: l... |
2349 |
int ret; |
81d39c20f memcg: fix shrink... |
2350 2351 2352 2353 2354 2355 |
/* * For keeping hierarchical_reclaim simple, how long we should retry * is depends on callers. We set our retry-count to be function * of # of children which we should visit in this loop. */ |
3e32cb2e0 mm: memcontrol: l... |
2356 2357 |
retry_count = MEM_CGROUP_RECLAIM_RETRIES * mem_cgroup_count_children(memcg); |
81d39c20f memcg: fix shrink... |
2358 |
|
3e32cb2e0 mm: memcontrol: l... |
2359 |
oldusage = page_counter_read(&memcg->memory); |
628f42355 memcg: limit chan... |
2360 |
|
3e32cb2e0 mm: memcontrol: l... |
2361 |
do { |
628f42355 memcg: limit chan... |
2362 2363 2364 2365 |
if (signal_pending(current)) { ret = -EINTR; break; } |
3e32cb2e0 mm: memcontrol: l... |
2366 2367 2368 2369 |
mutex_lock(&memcg_limit_mutex); if (limit > memcg->memsw.limit) { mutex_unlock(&memcg_limit_mutex); |
8c7c6e34a memcg: mem+swap c... |
2370 |
ret = -EINVAL; |
628f42355 memcg: limit chan... |
2371 2372 |
break; } |
3e32cb2e0 mm: memcontrol: l... |
2373 2374 2375 2376 |
if (limit > memcg->memory.limit) enlarge = true; ret = page_counter_limit(&memcg->memory, limit); mutex_unlock(&memcg_limit_mutex); |
8c7c6e34a memcg: mem+swap c... |
2377 2378 2379 |
if (!ret) break; |
b70a2a21d mm: memcontrol: f... |
2380 |
try_to_free_mem_cgroup_pages(memcg, 1, GFP_KERNEL, true); |
3e32cb2e0 mm: memcontrol: l... |
2381 |
curusage = page_counter_read(&memcg->memory); |
81d39c20f memcg: fix shrink... |
2382 |
/* Usage is reduced ? */ |
f894ffa86 memcg: trivial cl... |
2383 |
if (curusage >= oldusage) |
81d39c20f memcg: fix shrink... |
2384 2385 2386 |
retry_count--; else oldusage = curusage; |
3e32cb2e0 mm: memcontrol: l... |
2387 |
} while (retry_count); |
3c11ecf44 memcg: oom kill d... |
2388 2389 |
if (!ret && enlarge) memcg_oom_recover(memcg); |
14797e236 memcg: add inacti... |
2390 |
|
8c7c6e34a memcg: mem+swap c... |
2391 2392 |
return ret; } |
338c84310 memcg: remove som... |
2393 |
static int mem_cgroup_resize_memsw_limit(struct mem_cgroup *memcg, |
3e32cb2e0 mm: memcontrol: l... |
2394 |
unsigned long limit) |
8c7c6e34a memcg: mem+swap c... |
2395 |
{ |
3e32cb2e0 mm: memcontrol: l... |
2396 2397 2398 |
unsigned long curusage; unsigned long oldusage; bool enlarge = false; |
81d39c20f memcg: fix shrink... |
2399 |
int retry_count; |
3e32cb2e0 mm: memcontrol: l... |
2400 |
int ret; |
8c7c6e34a memcg: mem+swap c... |
2401 |
|
81d39c20f memcg: fix shrink... |
2402 |
/* see mem_cgroup_resize_res_limit */ |
3e32cb2e0 mm: memcontrol: l... |
2403 2404 2405 2406 2407 2408 |
retry_count = MEM_CGROUP_RECLAIM_RETRIES * mem_cgroup_count_children(memcg); oldusage = page_counter_read(&memcg->memsw); do { |
8c7c6e34a memcg: mem+swap c... |
2409 2410 2411 2412 |
if (signal_pending(current)) { ret = -EINTR; break; } |
3e32cb2e0 mm: memcontrol: l... |
2413 2414 2415 2416 |
mutex_lock(&memcg_limit_mutex); if (limit < memcg->memory.limit) { mutex_unlock(&memcg_limit_mutex); |
8c7c6e34a memcg: mem+swap c... |
2417 |
ret = -EINVAL; |
8c7c6e34a memcg: mem+swap c... |
2418 2419 |
break; } |
3e32cb2e0 mm: memcontrol: l... |
2420 2421 2422 2423 |
if (limit > memcg->memsw.limit) enlarge = true; ret = page_counter_limit(&memcg->memsw, limit); mutex_unlock(&memcg_limit_mutex); |
8c7c6e34a memcg: mem+swap c... |
2424 2425 2426 |
if (!ret) break; |
b70a2a21d mm: memcontrol: f... |
2427 |
try_to_free_mem_cgroup_pages(memcg, 1, GFP_KERNEL, false); |
3e32cb2e0 mm: memcontrol: l... |
2428 |
curusage = page_counter_read(&memcg->memsw); |
81d39c20f memcg: fix shrink... |
2429 |
/* Usage is reduced ? */ |
8c7c6e34a memcg: mem+swap c... |
2430 |
if (curusage >= oldusage) |
628f42355 memcg: limit chan... |
2431 |
retry_count--; |
81d39c20f memcg: fix shrink... |
2432 2433 |
else oldusage = curusage; |
3e32cb2e0 mm: memcontrol: l... |
2434 |
} while (retry_count); |
3c11ecf44 memcg: oom kill d... |
2435 2436 |
if (!ret && enlarge) memcg_oom_recover(memcg); |
3e32cb2e0 mm: memcontrol: l... |
2437 |
|
628f42355 memcg: limit chan... |
2438 2439 |
return ret; } |
0608f43da revert "memcg, vm... |
2440 2441 2442 2443 2444 2445 2446 2447 2448 |
unsigned long mem_cgroup_soft_limit_reclaim(struct zone *zone, int order, gfp_t gfp_mask, unsigned long *total_scanned) { unsigned long nr_reclaimed = 0; struct mem_cgroup_per_zone *mz, *next_mz = NULL; unsigned long reclaimed; int loop = 0; struct mem_cgroup_tree_per_zone *mctz; |
3e32cb2e0 mm: memcontrol: l... |
2449 |
unsigned long excess; |
0608f43da revert "memcg, vm... |
2450 2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 |
unsigned long nr_scanned; if (order > 0) return 0; mctz = soft_limit_tree_node_zone(zone_to_nid(zone), zone_idx(zone)); /* * This loop can run a while, specially if mem_cgroup's continuously * keep exceeding their soft limit and putting the system under * pressure */ do { if (next_mz) mz = next_mz; else mz = mem_cgroup_largest_soft_limit_node(mctz); if (!mz) break; nr_scanned = 0; reclaimed = mem_cgroup_soft_reclaim(mz->memcg, zone, gfp_mask, &nr_scanned); nr_reclaimed += reclaimed; *total_scanned += nr_scanned; |
0a31bc97c mm: memcontrol: r... |
2474 |
spin_lock_irq(&mctz->lock); |
bc2f2e7ff memcg: simplify u... |
2475 |
__mem_cgroup_remove_exceeded(mz, mctz); |
0608f43da revert "memcg, vm... |
2476 2477 2478 2479 2480 2481 |
/* * If we failed to reclaim anything from this memory cgroup * it is time to move on to the next cgroup */ next_mz = NULL; |
bc2f2e7ff memcg: simplify u... |
2482 2483 |
if (!reclaimed) next_mz = __mem_cgroup_largest_soft_limit_node(mctz); |
3e32cb2e0 mm: memcontrol: l... |
2484 |
excess = soft_limit_excess(mz->memcg); |
0608f43da revert "memcg, vm... |
2485 2486 2487 2488 2489 2490 2491 2492 2493 |
/* * One school of thought says that we should not add * back the node to the tree if reclaim returns 0. * But our reclaim could return 0, simply because due * to priority we are exposing a smaller subset of * memory to reclaim from. Consider this as a longer * term TODO. */ /* If excess == 0, no tree ops */ |
cf2c81279 mm: memcontrol: r... |
2494 |
__mem_cgroup_insert_exceeded(mz, mctz, excess); |
0a31bc97c mm: memcontrol: r... |
2495 |
spin_unlock_irq(&mctz->lock); |
0608f43da revert "memcg, vm... |
2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510 2511 |
css_put(&mz->memcg->css); loop++; /* * Could not reclaim anything and there are no more * mem cgroups to try or we seem to be looping without * reclaiming anything. */ if (!nr_reclaimed && (next_mz == NULL || loop > MEM_CGROUP_MAX_SOFT_LIMIT_RECLAIM_LOOPS)) break; } while (!nr_reclaimed); if (next_mz) css_put(&next_mz->memcg->css); return nr_reclaimed; } |
ea280e7b4 memcg: update mem... |
2512 2513 2514 2515 2516 2517 |
/* * Test whether @memcg has children, dead or alive. Note that this * function doesn't care whether @memcg has use_hierarchy enabled and * returns %true if there are child csses according to the cgroup * hierarchy. Testing use_hierarchy is the caller's responsiblity. */ |
b5f99b537 memcg: fast hiera... |
2518 2519 |
static inline bool memcg_has_children(struct mem_cgroup *memcg) { |
ea280e7b4 memcg: update mem... |
2520 |
bool ret; |
ea280e7b4 memcg: update mem... |
2521 2522 2523 2524 |
rcu_read_lock(); ret = css_next_child(NULL, &memcg->css); rcu_read_unlock(); return ret; |
b5f99b537 memcg: fast hiera... |
2525 2526 2527 |
} /* |
51038171b memcg: fix stale ... |
2528 |
* Reclaims as many pages from the given memcg as possible. |
c26251f9f memcg: split mem_... |
2529 2530 2531 2532 2533 2534 |
* * Caller is responsible for holding css reference for memcg. */ static int mem_cgroup_force_empty(struct mem_cgroup *memcg) { int nr_retries = MEM_CGROUP_RECLAIM_RETRIES; |
c26251f9f memcg: split mem_... |
2535 |
|
c1e862c1f memcg: new force_... |
2536 2537 |
/* we call try-to-free pages for make this cgroup empty */ lru_add_drain_all(); |
f817ed485 memcg: move all a... |
2538 |
/* try to free all pages in this cgroup */ |
3e32cb2e0 mm: memcontrol: l... |
2539 |
while (nr_retries && page_counter_read(&memcg->memory)) { |
f817ed485 memcg: move all a... |
2540 |
int progress; |
c1e862c1f memcg: new force_... |
2541 |
|
c26251f9f memcg: split mem_... |
2542 2543 |
if (signal_pending(current)) return -EINTR; |
b70a2a21d mm: memcontrol: f... |
2544 2545 |
progress = try_to_free_mem_cgroup_pages(memcg, 1, GFP_KERNEL, true); |
c1e862c1f memcg: new force_... |
2546 |
if (!progress) { |
f817ed485 memcg: move all a... |
2547 |
nr_retries--; |
c1e862c1f memcg: new force_... |
2548 |
/* maybe some writeback is necessary */ |
8aa7e847d Fix congestion_wa... |
2549 |
congestion_wait(BLK_RW_ASYNC, HZ/10); |
c1e862c1f memcg: new force_... |
2550 |
} |
f817ed485 memcg: move all a... |
2551 2552 |
} |
ab5196c20 memcg: make mem_c... |
2553 2554 |
return 0; |
cc8475822 memory cgroup enh... |
2555 |
} |
6770c64e5 cgroup: replace c... |
2556 2557 2558 |
static ssize_t mem_cgroup_force_empty_write(struct kernfs_open_file *of, char *buf, size_t nbytes, loff_t off) |
c1e862c1f memcg: new force_... |
2559 |
{ |
6770c64e5 cgroup: replace c... |
2560 |
struct mem_cgroup *memcg = mem_cgroup_from_css(of_css(of)); |
c26251f9f memcg: split mem_... |
2561 |
|
d84230118 memcg: root_cgrou... |
2562 2563 |
if (mem_cgroup_is_root(memcg)) return -EINVAL; |
6770c64e5 cgroup: replace c... |
2564 |
return mem_cgroup_force_empty(memcg) ?: nbytes; |
c1e862c1f memcg: new force_... |
2565 |
} |
182446d08 cgroup: pass arou... |
2566 2567 |
static u64 mem_cgroup_hierarchy_read(struct cgroup_subsys_state *css, struct cftype *cft) |
18f59ea7d memcg: memory cgr... |
2568 |
{ |
182446d08 cgroup: pass arou... |
2569 |
return mem_cgroup_from_css(css)->use_hierarchy; |
18f59ea7d memcg: memory cgr... |
2570 |
} |
182446d08 cgroup: pass arou... |
2571 2572 |
static int mem_cgroup_hierarchy_write(struct cgroup_subsys_state *css, struct cftype *cft, u64 val) |
18f59ea7d memcg: memory cgr... |
2573 2574 |
{ int retval = 0; |
182446d08 cgroup: pass arou... |
2575 |
struct mem_cgroup *memcg = mem_cgroup_from_css(css); |
5c9d535b8 cgroup: remove cs... |
2576 |
struct mem_cgroup *parent_memcg = mem_cgroup_from_css(memcg->css.parent); |
18f59ea7d memcg: memory cgr... |
2577 |
|
567fb435b memcg: fix bad be... |
2578 |
if (memcg->use_hierarchy == val) |
0b8f73e10 mm: memcontrol: c... |
2579 |
return 0; |
567fb435b memcg: fix bad be... |
2580 |
|
18f59ea7d memcg: memory cgr... |
2581 |
/* |
af901ca18 tree-wide: fix as... |
2582 |
* If parent's use_hierarchy is set, we can't make any modifications |
18f59ea7d memcg: memory cgr... |
2583 2584 2585 2586 2587 2588 |
* 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... |
2589 |
if ((!parent_memcg || !parent_memcg->use_hierarchy) && |
18f59ea7d memcg: memory cgr... |
2590 |
(val == 1 || val == 0)) { |
ea280e7b4 memcg: update mem... |
2591 |
if (!memcg_has_children(memcg)) |
c0ff4b854 memcg: rename mem... |
2592 |
memcg->use_hierarchy = val; |
18f59ea7d memcg: memory cgr... |
2593 2594 2595 2596 |
else retval = -EBUSY; } else retval = -EINVAL; |
567fb435b memcg: fix bad be... |
2597 |
|
18f59ea7d memcg: memory cgr... |
2598 2599 |
return retval; } |
72b54e731 mm: memcontrol: m... |
2600 |
static void tree_stat(struct mem_cgroup *memcg, unsigned long *stat) |
ce00a9673 mm: memcontrol: r... |
2601 2602 |
{ struct mem_cgroup *iter; |
72b54e731 mm: memcontrol: m... |
2603 |
int i; |
ce00a9673 mm: memcontrol: r... |
2604 |
|
72b54e731 mm: memcontrol: m... |
2605 |
memset(stat, 0, sizeof(*stat) * MEMCG_NR_STAT); |
ce00a9673 mm: memcontrol: r... |
2606 |
|
72b54e731 mm: memcontrol: m... |
2607 2608 2609 2610 |
for_each_mem_cgroup_tree(iter, memcg) { for (i = 0; i < MEMCG_NR_STAT; i++) stat[i] += mem_cgroup_read_stat(iter, i); } |
ce00a9673 mm: memcontrol: r... |
2611 |
} |
72b54e731 mm: memcontrol: m... |
2612 |
static void tree_events(struct mem_cgroup *memcg, unsigned long *events) |
587d9f726 mm: memcontrol: b... |
2613 2614 |
{ struct mem_cgroup *iter; |
72b54e731 mm: memcontrol: m... |
2615 |
int i; |
587d9f726 mm: memcontrol: b... |
2616 |
|
72b54e731 mm: memcontrol: m... |
2617 |
memset(events, 0, sizeof(*events) * MEMCG_NR_EVENTS); |
587d9f726 mm: memcontrol: b... |
2618 |
|
72b54e731 mm: memcontrol: m... |
2619 2620 2621 2622 |
for_each_mem_cgroup_tree(iter, memcg) { for (i = 0; i < MEMCG_NR_EVENTS; i++) events[i] += mem_cgroup_read_events(iter, i); } |
587d9f726 mm: memcontrol: b... |
2623 |
} |
6f6461562 mm/memcontrol.c: ... |
2624 |
static unsigned long mem_cgroup_usage(struct mem_cgroup *memcg, bool swap) |
ce00a9673 mm: memcontrol: r... |
2625 |
{ |
72b54e731 mm: memcontrol: m... |
2626 |
unsigned long val = 0; |
ce00a9673 mm: memcontrol: r... |
2627 |
|
3e32cb2e0 mm: memcontrol: l... |
2628 |
if (mem_cgroup_is_root(memcg)) { |
72b54e731 mm: memcontrol: m... |
2629 2630 2631 2632 2633 2634 2635 2636 2637 2638 2639 |
struct mem_cgroup *iter; for_each_mem_cgroup_tree(iter, memcg) { val += mem_cgroup_read_stat(iter, MEM_CGROUP_STAT_CACHE); val += mem_cgroup_read_stat(iter, MEM_CGROUP_STAT_RSS); if (swap) val += mem_cgroup_read_stat(iter, MEM_CGROUP_STAT_SWAP); } |
3e32cb2e0 mm: memcontrol: l... |
2640 |
} else { |
ce00a9673 mm: memcontrol: r... |
2641 |
if (!swap) |
3e32cb2e0 mm: memcontrol: l... |
2642 |
val = page_counter_read(&memcg->memory); |
ce00a9673 mm: memcontrol: r... |
2643 |
else |
3e32cb2e0 mm: memcontrol: l... |
2644 |
val = page_counter_read(&memcg->memsw); |
ce00a9673 mm: memcontrol: r... |
2645 |
} |
c12176d33 memcg: fix thresh... |
2646 |
return val; |
ce00a9673 mm: memcontrol: r... |
2647 |
} |
3e32cb2e0 mm: memcontrol: l... |
2648 2649 2650 2651 2652 2653 2654 |
enum { RES_USAGE, RES_LIMIT, RES_MAX_USAGE, RES_FAILCNT, RES_SOFT_LIMIT, }; |
ce00a9673 mm: memcontrol: r... |
2655 |
|
791badbdb memcg: convert aw... |
2656 |
static u64 mem_cgroup_read_u64(struct cgroup_subsys_state *css, |
05b843012 mm: memcontrol: u... |
2657 |
struct cftype *cft) |
8cdea7c05 Memory controller... |
2658 |
{ |
182446d08 cgroup: pass arou... |
2659 |
struct mem_cgroup *memcg = mem_cgroup_from_css(css); |
3e32cb2e0 mm: memcontrol: l... |
2660 |
struct page_counter *counter; |
af36f906c memcg: always cre... |
2661 |
|
3e32cb2e0 mm: memcontrol: l... |
2662 |
switch (MEMFILE_TYPE(cft->private)) { |
8c7c6e34a memcg: mem+swap c... |
2663 |
case _MEM: |
3e32cb2e0 mm: memcontrol: l... |
2664 2665 |
counter = &memcg->memory; break; |
8c7c6e34a memcg: mem+swap c... |
2666 |
case _MEMSWAP: |
3e32cb2e0 mm: memcontrol: l... |
2667 2668 |
counter = &memcg->memsw; break; |
510fc4e11 memcg: kmem accou... |
2669 |
case _KMEM: |
3e32cb2e0 mm: memcontrol: l... |
2670 |
counter = &memcg->kmem; |
510fc4e11 memcg: kmem accou... |
2671 |
break; |
d55f90bfa net: drop tcp_mem... |
2672 |
case _TCP: |
0db152981 mm: memcontrol: f... |
2673 |
counter = &memcg->tcpmem; |
d55f90bfa net: drop tcp_mem... |
2674 |
break; |
8c7c6e34a memcg: mem+swap c... |
2675 2676 |
default: BUG(); |
8c7c6e34a memcg: mem+swap c... |
2677 |
} |
3e32cb2e0 mm: memcontrol: l... |
2678 2679 2680 2681 |
switch (MEMFILE_ATTR(cft->private)) { case RES_USAGE: if (counter == &memcg->memory) |
c12176d33 memcg: fix thresh... |
2682 |
return (u64)mem_cgroup_usage(memcg, false) * PAGE_SIZE; |
3e32cb2e0 mm: memcontrol: l... |
2683 |
if (counter == &memcg->memsw) |
c12176d33 memcg: fix thresh... |
2684 |
return (u64)mem_cgroup_usage(memcg, true) * PAGE_SIZE; |
3e32cb2e0 mm: memcontrol: l... |
2685 2686 2687 2688 2689 2690 2691 2692 2693 2694 2695 2696 |
return (u64)page_counter_read(counter) * PAGE_SIZE; case RES_LIMIT: return (u64)counter->limit * PAGE_SIZE; case RES_MAX_USAGE: return (u64)counter->watermark * PAGE_SIZE; case RES_FAILCNT: return counter->failcnt; case RES_SOFT_LIMIT: return (u64)memcg->soft_limit * PAGE_SIZE; default: BUG(); } |
8cdea7c05 Memory controller... |
2697 |
} |
510fc4e11 memcg: kmem accou... |
2698 |
|
127424c86 mm: memcontrol: m... |
2699 |
#ifndef CONFIG_SLOB |
567e9ab2e mm: memcontrol: g... |
2700 |
static int memcg_online_kmem(struct mem_cgroup *memcg) |
d64416377 memcg: rework mem... |
2701 |
{ |
d64416377 memcg: rework mem... |
2702 |
int memcg_id; |
b313aeee2 mm: memcontrol: e... |
2703 2704 |
if (cgroup_memory_nokmem) return 0; |
2a4db7eb9 memcg: free memcg... |
2705 |
BUG_ON(memcg->kmemcg_id >= 0); |
567e9ab2e mm: memcontrol: g... |
2706 |
BUG_ON(memcg->kmem_state); |
d64416377 memcg: rework mem... |
2707 |
|
f3bb3043a memcg: don't call... |
2708 |
memcg_id = memcg_alloc_cache_id(); |
0b8f73e10 mm: memcontrol: c... |
2709 2710 |
if (memcg_id < 0) return memcg_id; |
d64416377 memcg: rework mem... |
2711 |
|
ef12947c9 mm: memcontrol: s... |
2712 |
static_branch_inc(&memcg_kmem_enabled_key); |
d64416377 memcg: rework mem... |
2713 |
/* |
567e9ab2e mm: memcontrol: g... |
2714 |
* A memory cgroup is considered kmem-online as soon as it gets |
900a38f02 memcg: zap kmem_a... |
2715 |
* kmemcg_id. Setting the id after enabling static branching will |
d64416377 memcg: rework mem... |
2716 2717 2718 |
* guarantee no one starts accounting before all call sites are * patched. */ |
900a38f02 memcg: zap kmem_a... |
2719 |
memcg->kmemcg_id = memcg_id; |
567e9ab2e mm: memcontrol: g... |
2720 |
memcg->kmem_state = KMEM_ONLINE; |
0b8f73e10 mm: memcontrol: c... |
2721 2722 |
return 0; |
d64416377 memcg: rework mem... |
2723 |
} |
8e0a89121 mm: memcontrol: g... |
2724 2725 2726 2727 2728 2729 2730 2731 2732 2733 2734 2735 2736 2737 2738 2739 2740 2741 2742 2743 2744 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766 2767 2768 2769 2770 |
static void memcg_offline_kmem(struct mem_cgroup *memcg) { struct cgroup_subsys_state *css; struct mem_cgroup *parent, *child; int kmemcg_id; if (memcg->kmem_state != KMEM_ONLINE) return; /* * Clear the online state before clearing memcg_caches array * entries. The slab_mutex in memcg_deactivate_kmem_caches() * guarantees that no cache will be created for this cgroup * after we are done (see memcg_create_kmem_cache()). */ memcg->kmem_state = KMEM_ALLOCATED; memcg_deactivate_kmem_caches(memcg); kmemcg_id = memcg->kmemcg_id; BUG_ON(kmemcg_id < 0); parent = parent_mem_cgroup(memcg); if (!parent) parent = root_mem_cgroup; /* * Change kmemcg_id of this cgroup and all its descendants to the * parent's id, and then move all entries from this cgroup's list_lrus * to ones of the parent. After we have finished, all list_lrus * corresponding to this cgroup are guaranteed to remain empty. The * ordering is imposed by list_lru_node->lock taken by * memcg_drain_all_list_lrus(). */ css_for_each_descendant_pre(css, &memcg->css) { child = mem_cgroup_from_css(css); BUG_ON(child->kmemcg_id != kmemcg_id); child->kmemcg_id = parent->kmemcg_id; if (!memcg->use_hierarchy) break; } memcg_drain_all_list_lrus(kmemcg_id, parent->kmemcg_id); memcg_free_cache_id(kmemcg_id); } static void memcg_free_kmem(struct mem_cgroup *memcg) { |
0b8f73e10 mm: memcontrol: c... |
2771 2772 2773 |
/* css_alloc() failed, offlining didn't happen */ if (unlikely(memcg->kmem_state == KMEM_ONLINE)) memcg_offline_kmem(memcg); |
8e0a89121 mm: memcontrol: g... |
2774 2775 2776 2777 2778 |
if (memcg->kmem_state == KMEM_ALLOCATED) { memcg_destroy_kmem_caches(memcg); static_branch_dec(&memcg_kmem_enabled_key); WARN_ON(page_counter_read(&memcg->kmem)); } |
8e0a89121 mm: memcontrol: g... |
2779 |
} |
d64416377 memcg: rework mem... |
2780 |
#else |
0b8f73e10 mm: memcontrol: c... |
2781 |
static int memcg_online_kmem(struct mem_cgroup *memcg) |
127424c86 mm: memcontrol: m... |
2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 |
{ return 0; } static void memcg_offline_kmem(struct mem_cgroup *memcg) { } static void memcg_free_kmem(struct mem_cgroup *memcg) { } #endif /* !CONFIG_SLOB */ |
d64416377 memcg: rework mem... |
2792 |
static int memcg_update_kmem_limit(struct mem_cgroup *memcg, |
3e32cb2e0 mm: memcontrol: l... |
2793 |
unsigned long limit) |
d64416377 memcg: rework mem... |
2794 |
{ |
b313aeee2 mm: memcontrol: e... |
2795 |
int ret; |
127424c86 mm: memcontrol: m... |
2796 2797 |
mutex_lock(&memcg_limit_mutex); |
127424c86 mm: memcontrol: m... |
2798 |
ret = page_counter_limit(&memcg->kmem, limit); |
127424c86 mm: memcontrol: m... |
2799 2800 |
mutex_unlock(&memcg_limit_mutex); return ret; |
d64416377 memcg: rework mem... |
2801 |
} |
510fc4e11 memcg: kmem accou... |
2802 |
|
d55f90bfa net: drop tcp_mem... |
2803 2804 2805 2806 2807 |
static int memcg_update_tcp_limit(struct mem_cgroup *memcg, unsigned long limit) { int ret; mutex_lock(&memcg_limit_mutex); |
0db152981 mm: memcontrol: f... |
2808 |
ret = page_counter_limit(&memcg->tcpmem, limit); |
d55f90bfa net: drop tcp_mem... |
2809 2810 |
if (ret) goto out; |
0db152981 mm: memcontrol: f... |
2811 |
if (!memcg->tcpmem_active) { |
d55f90bfa net: drop tcp_mem... |
2812 2813 2814 2815 2816 2817 2818 2819 2820 2821 2822 2823 2824 2825 2826 2827 2828 |
/* * The active flag needs to be written after the static_key * update. This is what guarantees that the socket activation * function is the last one to run. See sock_update_memcg() for * details, and note that we don't mark any socket as belonging * to this memcg until that flag is up. * * We need to do this, because static_keys will span multiple * sites, but we can't control their order. If we mark a socket * as accounted, but the accounting functions are not patched in * yet, we'll lose accounting. * * We never race with the readers in sock_update_memcg(), * because when this value change, the code to process it is not * patched in yet. */ static_branch_inc(&memcg_sockets_enabled_key); |
0db152981 mm: memcontrol: f... |
2829 |
memcg->tcpmem_active = true; |
d55f90bfa net: drop tcp_mem... |
2830 2831 2832 2833 2834 |
} out: mutex_unlock(&memcg_limit_mutex); return ret; } |
d55f90bfa net: drop tcp_mem... |
2835 |
|
628f42355 memcg: limit chan... |
2836 2837 2838 2839 |
/* * The user of this function is... * RES_LIMIT. */ |
451af504d cgroup: replace c... |
2840 2841 |
static ssize_t mem_cgroup_write(struct kernfs_open_file *of, char *buf, size_t nbytes, loff_t off) |
8cdea7c05 Memory controller... |
2842 |
{ |
451af504d cgroup: replace c... |
2843 |
struct mem_cgroup *memcg = mem_cgroup_from_css(of_css(of)); |
3e32cb2e0 mm: memcontrol: l... |
2844 |
unsigned long nr_pages; |
628f42355 memcg: limit chan... |
2845 |
int ret; |
451af504d cgroup: replace c... |
2846 |
buf = strstrip(buf); |
650c5e565 mm: page_counter:... |
2847 |
ret = page_counter_memparse(buf, "-1", &nr_pages); |
3e32cb2e0 mm: memcontrol: l... |
2848 2849 |
if (ret) return ret; |
af36f906c memcg: always cre... |
2850 |
|
3e32cb2e0 mm: memcontrol: l... |
2851 |
switch (MEMFILE_ATTR(of_cft(of)->private)) { |
628f42355 memcg: limit chan... |
2852 |
case RES_LIMIT: |
4b3bde4c9 memcg: remove the... |
2853 2854 2855 2856 |
if (mem_cgroup_is_root(memcg)) { /* Can't set limit on root */ ret = -EINVAL; break; } |
3e32cb2e0 mm: memcontrol: l... |
2857 2858 2859 |
switch (MEMFILE_TYPE(of_cft(of)->private)) { case _MEM: ret = mem_cgroup_resize_limit(memcg, nr_pages); |
8c7c6e34a memcg: mem+swap c... |
2860 |
break; |
3e32cb2e0 mm: memcontrol: l... |
2861 2862 |
case _MEMSWAP: ret = mem_cgroup_resize_memsw_limit(memcg, nr_pages); |
296c81d89 memory controller... |
2863 |
break; |
3e32cb2e0 mm: memcontrol: l... |
2864 2865 2866 |
case _KMEM: ret = memcg_update_kmem_limit(memcg, nr_pages); break; |
d55f90bfa net: drop tcp_mem... |
2867 2868 2869 |
case _TCP: ret = memcg_update_tcp_limit(memcg, nr_pages); break; |
3e32cb2e0 mm: memcontrol: l... |
2870 |
} |
296c81d89 memory controller... |
2871 |
break; |
3e32cb2e0 mm: memcontrol: l... |
2872 2873 2874 |
case RES_SOFT_LIMIT: memcg->soft_limit = nr_pages; ret = 0; |
628f42355 memcg: limit chan... |
2875 2876 |
break; } |
451af504d cgroup: replace c... |
2877 |
return ret ?: nbytes; |
8cdea7c05 Memory controller... |
2878 |
} |
6770c64e5 cgroup: replace c... |
2879 2880 |
static ssize_t mem_cgroup_reset(struct kernfs_open_file *of, char *buf, size_t nbytes, loff_t off) |
c84872e16 memcgroup: add th... |
2881 |
{ |
6770c64e5 cgroup: replace c... |
2882 |
struct mem_cgroup *memcg = mem_cgroup_from_css(of_css(of)); |
3e32cb2e0 mm: memcontrol: l... |
2883 |
struct page_counter *counter; |
c84872e16 memcgroup: add th... |
2884 |
|
3e32cb2e0 mm: memcontrol: l... |
2885 2886 2887 2888 2889 2890 2891 2892 2893 2894 |
switch (MEMFILE_TYPE(of_cft(of)->private)) { case _MEM: counter = &memcg->memory; break; case _MEMSWAP: counter = &memcg->memsw; break; case _KMEM: counter = &memcg->kmem; break; |
d55f90bfa net: drop tcp_mem... |
2895 |
case _TCP: |
0db152981 mm: memcontrol: f... |
2896 |
counter = &memcg->tcpmem; |
d55f90bfa net: drop tcp_mem... |
2897 |
break; |
3e32cb2e0 mm: memcontrol: l... |
2898 2899 2900 |
default: BUG(); } |
af36f906c memcg: always cre... |
2901 |
|
3e32cb2e0 mm: memcontrol: l... |
2902 |
switch (MEMFILE_ATTR(of_cft(of)->private)) { |
29f2a4dac memcgroup: implem... |
2903 |
case RES_MAX_USAGE: |
3e32cb2e0 mm: memcontrol: l... |
2904 |
page_counter_reset_watermark(counter); |
29f2a4dac memcgroup: implem... |
2905 2906 |
break; case RES_FAILCNT: |
3e32cb2e0 mm: memcontrol: l... |
2907 |
counter->failcnt = 0; |
29f2a4dac memcgroup: implem... |
2908 |
break; |
3e32cb2e0 mm: memcontrol: l... |
2909 2910 |
default: BUG(); |
29f2a4dac memcgroup: implem... |
2911 |
} |
f64c3f549 memory controller... |
2912 |
|
6770c64e5 cgroup: replace c... |
2913 |
return nbytes; |
c84872e16 memcgroup: add th... |
2914 |
} |
182446d08 cgroup: pass arou... |
2915 |
static u64 mem_cgroup_move_charge_read(struct cgroup_subsys_state *css, |
7dc74be03 memcg: add interf... |
2916 2917 |
struct cftype *cft) { |
182446d08 cgroup: pass arou... |
2918 |
return mem_cgroup_from_css(css)->move_charge_at_immigrate; |
7dc74be03 memcg: add interf... |
2919 |
} |
024914477 memcg: move charg... |
2920 |
#ifdef CONFIG_MMU |
182446d08 cgroup: pass arou... |
2921 |
static int mem_cgroup_move_charge_write(struct cgroup_subsys_state *css, |
7dc74be03 memcg: add interf... |
2922 2923 |
struct cftype *cft, u64 val) { |
182446d08 cgroup: pass arou... |
2924 |
struct mem_cgroup *memcg = mem_cgroup_from_css(css); |
7dc74be03 memcg: add interf... |
2925 |
|
1dfab5abc mm: memcontrol: f... |
2926 |
if (val & ~MOVE_MASK) |
7dc74be03 memcg: add interf... |
2927 |
return -EINVAL; |
ee5e8472b memcg: prevent ch... |
2928 |
|
7dc74be03 memcg: add interf... |
2929 |
/* |
ee5e8472b memcg: prevent ch... |
2930 2931 2932 2933 |
* No kind of locking is needed in here, because ->can_attach() will * check this value once in the beginning of the process, and then carry * on with stale data. This means that changes to this value will only * affect task migrations starting after the change. |
7dc74be03 memcg: add interf... |
2934 |
*/ |
c0ff4b854 memcg: rename mem... |
2935 |
memcg->move_charge_at_immigrate = val; |
7dc74be03 memcg: add interf... |
2936 2937 |
return 0; } |
024914477 memcg: move charg... |
2938 |
#else |
182446d08 cgroup: pass arou... |
2939 |
static int mem_cgroup_move_charge_write(struct cgroup_subsys_state *css, |
024914477 memcg: move charg... |
2940 2941 2942 2943 2944 |
struct cftype *cft, u64 val) { return -ENOSYS; } #endif |
7dc74be03 memcg: add interf... |
2945 |
|
406eb0c9b memcg: add memory... |
2946 |
#ifdef CONFIG_NUMA |
2da8ca822 cgroup: replace c... |
2947 |
static int memcg_numa_stat_show(struct seq_file *m, void *v) |
406eb0c9b memcg: add memory... |
2948 |
{ |
25485de6e memcg: refactor m... |
2949 2950 2951 2952 2953 2954 2955 2956 2957 2958 2959 2960 |
struct numa_stat { const char *name; unsigned int lru_mask; }; static const struct numa_stat stats[] = { { "total", LRU_ALL }, { "file", LRU_ALL_FILE }, { "anon", LRU_ALL_ANON }, { "unevictable", BIT(LRU_UNEVICTABLE) }, }; const struct numa_stat *stat; |
406eb0c9b memcg: add memory... |
2961 |
int nid; |
25485de6e memcg: refactor m... |
2962 |
unsigned long nr; |
2da8ca822 cgroup: replace c... |
2963 |
struct mem_cgroup *memcg = mem_cgroup_from_css(seq_css(m)); |
406eb0c9b memcg: add memory... |
2964 |
|
25485de6e memcg: refactor m... |
2965 2966 2967 2968 2969 2970 2971 2972 2973 2974 |
for (stat = stats; stat < stats + ARRAY_SIZE(stats); stat++) { nr = mem_cgroup_nr_lru_pages(memcg, stat->lru_mask); seq_printf(m, "%s=%lu", stat->name, nr); for_each_node_state(nid, N_MEMORY) { nr = mem_cgroup_node_nr_lru_pages(memcg, nid, stat->lru_mask); seq_printf(m, " N%d=%lu", nid, nr); } seq_putc(m, ' '); |
406eb0c9b memcg: add memory... |
2975 |
} |
406eb0c9b memcg: add memory... |
2976 |
|
071aee138 memcg: support hi... |
2977 2978 2979 2980 2981 2982 2983 2984 2985 2986 2987 2988 2989 2990 2991 2992 |
for (stat = stats; stat < stats + ARRAY_SIZE(stats); stat++) { struct mem_cgroup *iter; nr = 0; for_each_mem_cgroup_tree(iter, memcg) nr += mem_cgroup_nr_lru_pages(iter, stat->lru_mask); seq_printf(m, "hierarchical_%s=%lu", stat->name, nr); for_each_node_state(nid, N_MEMORY) { nr = 0; for_each_mem_cgroup_tree(iter, memcg) nr += mem_cgroup_node_nr_lru_pages( iter, nid, stat->lru_mask); seq_printf(m, " N%d=%lu", nid, nr); } seq_putc(m, ' '); |
406eb0c9b memcg: add memory... |
2993 |
} |
406eb0c9b memcg: add memory... |
2994 |
|
406eb0c9b memcg: add memory... |
2995 2996 2997 |
return 0; } #endif /* CONFIG_NUMA */ |
2da8ca822 cgroup: replace c... |
2998 |
static int memcg_stat_show(struct seq_file *m, void *v) |
d2ceb9b7d memory cgroup enh... |
2999 |
{ |
2da8ca822 cgroup: replace c... |
3000 |
struct mem_cgroup *memcg = mem_cgroup_from_css(seq_css(m)); |
3e32cb2e0 mm: memcontrol: l... |
3001 |
unsigned long memory, memsw; |
af7c4b0ec mm: memcg: print ... |
3002 3003 |
struct mem_cgroup *mi; unsigned int i; |
406eb0c9b memcg: add memory... |
3004 |
|
0ca44b148 memcg: add BUILD_... |
3005 3006 3007 3008 |
BUILD_BUG_ON(ARRAY_SIZE(mem_cgroup_stat_names) != MEM_CGROUP_STAT_NSTATS); BUILD_BUG_ON(ARRAY_SIZE(mem_cgroup_events_names) != MEM_CGROUP_EVENTS_NSTATS); |
70bc068c4 mm/memcontrol.c: ... |
3009 |
BUILD_BUG_ON(ARRAY_SIZE(mem_cgroup_lru_names) != NR_LRU_LISTS); |
af7c4b0ec mm: memcg: print ... |
3010 |
for (i = 0; i < MEM_CGROUP_STAT_NSTATS; i++) { |
7941d2145 mm: memcontrol: d... |
3011 |
if (i == MEM_CGROUP_STAT_SWAP && !do_memsw_account()) |
1dd3a2732 memcg: show swap ... |
3012 |
continue; |
484ebb3b8 memcg: make mem_c... |
3013 3014 |
seq_printf(m, "%s %lu ", mem_cgroup_stat_names[i], |
af7c4b0ec mm: memcg: print ... |
3015 |
mem_cgroup_read_stat(memcg, i) * PAGE_SIZE); |
1dd3a2732 memcg: show swap ... |
3016 |
} |
7b854121e Unevictable LRU P... |
3017 |
|
af7c4b0ec mm: memcg: print ... |
3018 3019 3020 3021 3022 3023 3024 3025 3026 |
for (i = 0; i < MEM_CGROUP_EVENTS_NSTATS; i++) seq_printf(m, "%s %lu ", mem_cgroup_events_names[i], mem_cgroup_read_events(memcg, i)); for (i = 0; i < NR_LRU_LISTS; i++) seq_printf(m, "%s %lu ", mem_cgroup_lru_names[i], mem_cgroup_nr_lru_pages(memcg, BIT(i)) * PAGE_SIZE); |
14067bb3e memcg: hierarchic... |
3027 |
/* Hierarchical information */ |
3e32cb2e0 mm: memcontrol: l... |
3028 3029 3030 3031 |
memory = memsw = PAGE_COUNTER_MAX; for (mi = memcg; mi; mi = parent_mem_cgroup(mi)) { memory = min(memory, mi->memory.limit); memsw = min(memsw, mi->memsw.limit); |
fee7b548e memcg: show real ... |
3032 |
} |
3e32cb2e0 mm: memcontrol: l... |
3033 3034 3035 |
seq_printf(m, "hierarchical_memory_limit %llu ", (u64)memory * PAGE_SIZE); |
7941d2145 mm: memcontrol: d... |
3036 |
if (do_memsw_account()) |
3e32cb2e0 mm: memcontrol: l... |
3037 3038 3039 |
seq_printf(m, "hierarchical_memsw_limit %llu ", (u64)memsw * PAGE_SIZE); |
7f016ee8b memcg: show recla... |
3040 |
|
af7c4b0ec mm: memcg: print ... |
3041 |
for (i = 0; i < MEM_CGROUP_STAT_NSTATS; i++) { |
484ebb3b8 memcg: make mem_c... |
3042 |
unsigned long long val = 0; |
af7c4b0ec mm: memcg: print ... |
3043 |
|
7941d2145 mm: memcontrol: d... |
3044 |
if (i == MEM_CGROUP_STAT_SWAP && !do_memsw_account()) |
1dd3a2732 memcg: show swap ... |
3045 |
continue; |
af7c4b0ec mm: memcg: print ... |
3046 3047 |
for_each_mem_cgroup_tree(mi, memcg) val += mem_cgroup_read_stat(mi, i) * PAGE_SIZE; |
484ebb3b8 memcg: make mem_c... |
3048 3049 |
seq_printf(m, "total_%s %llu ", mem_cgroup_stat_names[i], val); |
af7c4b0ec mm: memcg: print ... |
3050 3051 3052 3053 3054 3055 3056 3057 3058 3059 3060 3061 3062 3063 3064 3065 3066 3067 3068 |
} for (i = 0; i < MEM_CGROUP_EVENTS_NSTATS; i++) { unsigned long long val = 0; for_each_mem_cgroup_tree(mi, memcg) val += mem_cgroup_read_events(mi, i); seq_printf(m, "total_%s %llu ", mem_cgroup_events_names[i], val); } for (i = 0; i < NR_LRU_LISTS; i++) { unsigned long long val = 0; for_each_mem_cgroup_tree(mi, memcg) val += mem_cgroup_nr_lru_pages(mi, BIT(i)) * PAGE_SIZE; seq_printf(m, "total_%s %llu ", mem_cgroup_lru_names[i], val); |
1dd3a2732 memcg: show swap ... |
3069 |
} |
14067bb3e memcg: hierarchic... |
3070 |
|
7f016ee8b memcg: show recla... |
3071 |
#ifdef CONFIG_DEBUG_VM |
7f016ee8b memcg: show recla... |
3072 3073 3074 |
{ int nid, zid; struct mem_cgroup_per_zone *mz; |
89abfab13 mm/memcg: move re... |
3075 |
struct zone_reclaim_stat *rstat; |
7f016ee8b memcg: show recla... |
3076 3077 3078 3079 3080 |
unsigned long recent_rotated[2] = {0, 0}; unsigned long recent_scanned[2] = {0, 0}; for_each_online_node(nid) for (zid = 0; zid < MAX_NR_ZONES; zid++) { |
e231875ba mm: memcontrol: c... |
3081 |
mz = &memcg->nodeinfo[nid]->zoneinfo[zid]; |
89abfab13 mm/memcg: move re... |
3082 |
rstat = &mz->lruvec.reclaim_stat; |
7f016ee8b memcg: show recla... |
3083 |
|
89abfab13 mm/memcg: move re... |
3084 3085 3086 3087 |
recent_rotated[0] += rstat->recent_rotated[0]; recent_rotated[1] += rstat->recent_rotated[1]; recent_scanned[0] += rstat->recent_scanned[0]; recent_scanned[1] += rstat->recent_scanned[1]; |
7f016ee8b memcg: show recla... |
3088 |
} |
78ccf5b5a mm: memcg: print ... |
3089 3090 3091 3092 3093 3094 3095 3096 |
seq_printf(m, "recent_rotated_anon %lu ", recent_rotated[0]); seq_printf(m, "recent_rotated_file %lu ", recent_rotated[1]); seq_printf(m, "recent_scanned_anon %lu ", recent_scanned[0]); seq_printf(m, "recent_scanned_file %lu ", recent_scanned[1]); |
7f016ee8b memcg: show recla... |
3097 3098 |
} #endif |
d2ceb9b7d memory cgroup enh... |
3099 3100 |
return 0; } |
182446d08 cgroup: pass arou... |
3101 3102 |
static u64 mem_cgroup_swappiness_read(struct cgroup_subsys_state *css, struct cftype *cft) |
a7885eb8a memcg: swappiness |
3103 |
{ |
182446d08 cgroup: pass arou... |
3104 |
struct mem_cgroup *memcg = mem_cgroup_from_css(css); |
a7885eb8a memcg: swappiness |
3105 |
|
1f4c025b5 memcg: export mem... |
3106 |
return mem_cgroup_swappiness(memcg); |
a7885eb8a memcg: swappiness |
3107 |
} |
182446d08 cgroup: pass arou... |
3108 3109 |
static int mem_cgroup_swappiness_write(struct cgroup_subsys_state *css, struct cftype *cft, u64 val) |
a7885eb8a memcg: swappiness |
3110 |
{ |
182446d08 cgroup: pass arou... |
3111 |
struct mem_cgroup *memcg = mem_cgroup_from_css(css); |
a7885eb8a memcg: swappiness |
3112 |
|
3dae7fec5 mm: memcontrol: r... |
3113 |
if (val > 100) |
a7885eb8a memcg: swappiness |
3114 |
return -EINVAL; |
14208b0ec Merge branch 'for... |
3115 |
if (css->parent) |
3dae7fec5 mm: memcontrol: r... |
3116 3117 3118 |
memcg->swappiness = val; else vm_swappiness = val; |
068b38c1f memcg: fix a race... |
3119 |
|
a7885eb8a memcg: swappiness |
3120 3121 |
return 0; } |
2e72b6347 memcg: implement ... |
3122 3123 3124 |
static void __mem_cgroup_threshold(struct mem_cgroup *memcg, bool swap) { struct mem_cgroup_threshold_ary *t; |
3e32cb2e0 mm: memcontrol: l... |
3125 |
unsigned long usage; |
2e72b6347 memcg: implement ... |
3126 3127 3128 3129 |
int i; rcu_read_lock(); if (!swap) |
2c488db27 memcg: clean up m... |
3130 |
t = rcu_dereference(memcg->thresholds.primary); |
2e72b6347 memcg: implement ... |
3131 |
else |
2c488db27 memcg: clean up m... |
3132 |
t = rcu_dereference(memcg->memsw_thresholds.primary); |
2e72b6347 memcg: implement ... |
3133 3134 3135 |
if (!t) goto unlock; |
ce00a9673 mm: memcontrol: r... |
3136 |
usage = mem_cgroup_usage(memcg, swap); |
2e72b6347 memcg: implement ... |
3137 3138 |
/* |
748dad36d memcg: make thres... |
3139 |
* current_threshold points to threshold just below or equal to usage. |
2e72b6347 memcg: implement ... |
3140 3141 3142 |
* If it's not true, a threshold was crossed after last * call of __mem_cgroup_threshold(). */ |
5407a5625 mm: remove unnece... |
3143 |
i = t->current_threshold; |
2e72b6347 memcg: implement ... |
3144 3145 3146 3147 3148 3149 3150 3151 3152 3153 3154 3155 3156 3157 3158 3159 3160 3161 3162 3163 3164 3165 3166 |
/* * 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... |
3167 |
t->current_threshold = i - 1; |
2e72b6347 memcg: implement ... |
3168 3169 3170 3171 3172 3173 |
unlock: rcu_read_unlock(); } static void mem_cgroup_threshold(struct mem_cgroup *memcg) { |
ad4ca5f4b memcg: fix thresh... |
3174 3175 |
while (memcg) { __mem_cgroup_threshold(memcg, false); |
7941d2145 mm: memcontrol: d... |
3176 |
if (do_memsw_account()) |
ad4ca5f4b memcg: fix thresh... |
3177 3178 3179 3180 |
__mem_cgroup_threshold(memcg, true); memcg = parent_mem_cgroup(memcg); } |
2e72b6347 memcg: implement ... |
3181 3182 3183 3184 3185 3186 |
} static int compare_thresholds(const void *a, const void *b) { const struct mem_cgroup_threshold *_a = a; const struct mem_cgroup_threshold *_b = b; |
2bff24a37 memcg: fix multip... |
3187 3188 3189 3190 3191 3192 3193 |
if (_a->threshold > _b->threshold) return 1; if (_a->threshold < _b->threshold) return -1; return 0; |
2e72b6347 memcg: implement ... |
3194 |
} |
c0ff4b854 memcg: rename mem... |
3195 |
static int mem_cgroup_oom_notify_cb(struct mem_cgroup *memcg) |
9490ff275 memcg: oom notifier |
3196 3197 |
{ struct mem_cgroup_eventfd_list *ev; |
2bcf2e92c memcg: oom_notify... |
3198 |
spin_lock(&memcg_oom_lock); |
c0ff4b854 memcg: rename mem... |
3199 |
list_for_each_entry(ev, &memcg->oom_notify, list) |
9490ff275 memcg: oom notifier |
3200 |
eventfd_signal(ev->eventfd, 1); |
2bcf2e92c memcg: oom_notify... |
3201 3202 |
spin_unlock(&memcg_oom_lock); |
9490ff275 memcg: oom notifier |
3203 3204 |
return 0; } |
c0ff4b854 memcg: rename mem... |
3205 |
static void mem_cgroup_oom_notify(struct mem_cgroup *memcg) |
9490ff275 memcg: oom notifier |
3206 |
{ |
7d74b06f2 memcg: use for_ea... |
3207 |
struct mem_cgroup *iter; |
c0ff4b854 memcg: rename mem... |
3208 |
for_each_mem_cgroup_tree(iter, memcg) |
7d74b06f2 memcg: use for_ea... |
3209 |
mem_cgroup_oom_notify_cb(iter); |
9490ff275 memcg: oom notifier |
3210 |
} |
59b6f8734 memcg: make cgrou... |
3211 |
static int __mem_cgroup_usage_register_event(struct mem_cgroup *memcg, |
347c4a874 memcg: remove cgr... |
3212 |
struct eventfd_ctx *eventfd, const char *args, enum res_type type) |
2e72b6347 memcg: implement ... |
3213 |
{ |
2c488db27 memcg: clean up m... |
3214 3215 |
struct mem_cgroup_thresholds *thresholds; struct mem_cgroup_threshold_ary *new; |
3e32cb2e0 mm: memcontrol: l... |
3216 3217 |
unsigned long threshold; unsigned long usage; |
2c488db27 memcg: clean up m... |
3218 |
int i, size, ret; |
2e72b6347 memcg: implement ... |
3219 |
|
650c5e565 mm: page_counter:... |
3220 |
ret = page_counter_memparse(args, "-1", &threshold); |
2e72b6347 memcg: implement ... |
3221 3222 3223 3224 |
if (ret) return ret; mutex_lock(&memcg->thresholds_lock); |
2c488db27 memcg: clean up m... |
3225 |
|
05b843012 mm: memcontrol: u... |
3226 |
if (type == _MEM) { |
2c488db27 memcg: clean up m... |
3227 |
thresholds = &memcg->thresholds; |
ce00a9673 mm: memcontrol: r... |
3228 |
usage = mem_cgroup_usage(memcg, false); |
05b843012 mm: memcontrol: u... |
3229 |
} else if (type == _MEMSWAP) { |
2c488db27 memcg: clean up m... |
3230 |
thresholds = &memcg->memsw_thresholds; |
ce00a9673 mm: memcontrol: r... |
3231 |
usage = mem_cgroup_usage(memcg, true); |
05b843012 mm: memcontrol: u... |
3232 |
} else |
2e72b6347 memcg: implement ... |
3233 |
BUG(); |
2e72b6347 memcg: implement ... |
3234 |
/* Check if a threshold crossed before adding a new one */ |
2c488db27 memcg: clean up m... |
3235 |
if (thresholds->primary) |
2e72b6347 memcg: implement ... |
3236 |
__mem_cgroup_threshold(memcg, type == _MEMSWAP); |
2c488db27 memcg: clean up m... |
3237 |
size = thresholds->primary ? thresholds->primary->size + 1 : 1; |
2e72b6347 memcg: implement ... |
3238 3239 |
/* Allocate memory for new array of thresholds */ |
2c488db27 memcg: clean up m... |
3240 |
new = kmalloc(sizeof(*new) + size * sizeof(struct mem_cgroup_threshold), |
2e72b6347 memcg: implement ... |
3241 |
GFP_KERNEL); |
2c488db27 memcg: clean up m... |
3242 |
if (!new) { |
2e72b6347 memcg: implement ... |
3243 3244 3245 |
ret = -ENOMEM; goto unlock; } |
2c488db27 memcg: clean up m... |
3246 |
new->size = size; |
2e72b6347 memcg: implement ... |
3247 3248 |
/* Copy thresholds (if any) to new array */ |
2c488db27 memcg: clean up m... |
3249 3250 |
if (thresholds->primary) { memcpy(new->entries, thresholds->primary->entries, (size - 1) * |
2e72b6347 memcg: implement ... |
3251 |
sizeof(struct mem_cgroup_threshold)); |
2c488db27 memcg: clean up m... |
3252 |
} |
2e72b6347 memcg: implement ... |
3253 |
/* Add new threshold */ |
2c488db27 memcg: clean up m... |
3254 3255 |
new->entries[size - 1].eventfd = eventfd; new->entries[size - 1].threshold = threshold; |
2e72b6347 memcg: implement ... |
3256 3257 |
/* Sort thresholds. Registering of new threshold isn't time-critical */ |
2c488db27 memcg: clean up m... |
3258 |
sort(new->entries, size, sizeof(struct mem_cgroup_threshold), |
2e72b6347 memcg: implement ... |
3259 3260 3261 |
compare_thresholds, NULL); /* Find current threshold */ |
2c488db27 memcg: clean up m... |
3262 |
new->current_threshold = -1; |
2e72b6347 memcg: implement ... |
3263 |
for (i = 0; i < size; i++) { |
748dad36d memcg: make thres... |
3264 |
if (new->entries[i].threshold <= usage) { |
2e72b6347 memcg: implement ... |
3265 |
/* |
2c488db27 memcg: clean up m... |
3266 3267 |
* new->current_threshold will not be used until * rcu_assign_pointer(), so it's safe to increment |
2e72b6347 memcg: implement ... |
3268 3269 |
* it here. */ |
2c488db27 memcg: clean up m... |
3270 |
++new->current_threshold; |
748dad36d memcg: make thres... |
3271 3272 |
} else break; |
2e72b6347 memcg: implement ... |
3273 |
} |
2c488db27 memcg: clean up m... |
3274 3275 3276 3277 3278 |
/* 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 ... |
3279 |
|
907860ed3 cgroups: make cft... |
3280 |
/* To be sure that nobody uses thresholds */ |
2e72b6347 memcg: implement ... |
3281 |
synchronize_rcu(); |
2e72b6347 memcg: implement ... |
3282 3283 3284 3285 3286 |
unlock: mutex_unlock(&memcg->thresholds_lock); return ret; } |
59b6f8734 memcg: make cgrou... |
3287 |
static int mem_cgroup_usage_register_event(struct mem_cgroup *memcg, |
347c4a874 memcg: remove cgr... |
3288 3289 |
struct eventfd_ctx *eventfd, const char *args) { |
59b6f8734 memcg: make cgrou... |
3290 |
return __mem_cgroup_usage_register_event(memcg, eventfd, args, _MEM); |
347c4a874 memcg: remove cgr... |
3291 |
} |
59b6f8734 memcg: make cgrou... |
3292 |
static int memsw_cgroup_usage_register_event(struct mem_cgroup *memcg, |
347c4a874 memcg: remove cgr... |
3293 3294 |
struct eventfd_ctx *eventfd, const char *args) { |
59b6f8734 memcg: make cgrou... |
3295 |
return __mem_cgroup_usage_register_event(memcg, eventfd, args, _MEMSWAP); |
347c4a874 memcg: remove cgr... |
3296 |
} |
59b6f8734 memcg: make cgrou... |
3297 |
static void __mem_cgroup_usage_unregister_event(struct mem_cgroup *memcg, |
347c4a874 memcg: remove cgr... |
3298 |
struct eventfd_ctx *eventfd, enum res_type type) |
2e72b6347 memcg: implement ... |
3299 |
{ |
2c488db27 memcg: clean up m... |
3300 3301 |
struct mem_cgroup_thresholds *thresholds; struct mem_cgroup_threshold_ary *new; |
3e32cb2e0 mm: memcontrol: l... |
3302 |
unsigned long usage; |
2c488db27 memcg: clean up m... |
3303 |
int i, j, size; |
2e72b6347 memcg: implement ... |
3304 3305 |
mutex_lock(&memcg->thresholds_lock); |
05b843012 mm: memcontrol: u... |
3306 3307 |
if (type == _MEM) { |
2c488db27 memcg: clean up m... |
3308 |
thresholds = &memcg->thresholds; |
ce00a9673 mm: memcontrol: r... |
3309 |
usage = mem_cgroup_usage(memcg, false); |
05b843012 mm: memcontrol: u... |
3310 |
} else if (type == _MEMSWAP) { |
2c488db27 memcg: clean up m... |
3311 |
thresholds = &memcg->memsw_thresholds; |
ce00a9673 mm: memcontrol: r... |
3312 |
usage = mem_cgroup_usage(memcg, true); |
05b843012 mm: memcontrol: u... |
3313 |
} else |
2e72b6347 memcg: implement ... |
3314 |
BUG(); |
371528cae mm: memcg: Correc... |
3315 3316 |
if (!thresholds->primary) goto unlock; |
2e72b6347 memcg: implement ... |
3317 3318 3319 3320 |
/* Check if a threshold crossed before removing */ __mem_cgroup_threshold(memcg, type == _MEMSWAP); /* Calculate new number of threshold */ |
2c488db27 memcg: clean up m... |
3321 3322 3323 |
size = 0; for (i = 0; i < thresholds->primary->size; i++) { if (thresholds->primary->entries[i].eventfd != eventfd) |
2e72b6347 memcg: implement ... |
3324 3325 |
size++; } |
2c488db27 memcg: clean up m... |
3326 |
new = thresholds->spare; |
907860ed3 cgroups: make cft... |
3327 |
|
2e72b6347 memcg: implement ... |
3328 3329 |
/* Set thresholds array to NULL if we don't have thresholds */ if (!size) { |
2c488db27 memcg: clean up m... |
3330 3331 |
kfree(new); new = NULL; |
907860ed3 cgroups: make cft... |
3332 |
goto swap_buffers; |
2e72b6347 memcg: implement ... |
3333 |
} |
2c488db27 memcg: clean up m... |
3334 |
new->size = size; |
2e72b6347 memcg: implement ... |
3335 3336 |
/* Copy thresholds and find current threshold */ |
2c488db27 memcg: clean up m... |
3337 3338 3339 |
new->current_threshold = -1; for (i = 0, j = 0; i < thresholds->primary->size; i++) { if (thresholds->primary->entries[i].eventfd == eventfd) |
2e72b6347 memcg: implement ... |
3340 |
continue; |
2c488db27 memcg: clean up m... |
3341 |
new->entries[j] = thresholds->primary->entries[i]; |
748dad36d memcg: make thres... |
3342 |
if (new->entries[j].threshold <= usage) { |
2e72b6347 memcg: implement ... |
3343 |
/* |
2c488db27 memcg: clean up m... |
3344 |
* new->current_threshold will not be used |
2e72b6347 memcg: implement ... |
3345 3346 3347 |
* until rcu_assign_pointer(), so it's safe to increment * it here. */ |
2c488db27 memcg: clean up m... |
3348 |
++new->current_threshold; |
2e72b6347 memcg: implement ... |
3349 3350 3351 |
} j++; } |
907860ed3 cgroups: make cft... |
3352 |
swap_buffers: |
2c488db27 memcg: clean up m... |
3353 3354 |
/* Swap primary and spare array */ thresholds->spare = thresholds->primary; |
8c7577637 memcg: free spare... |
3355 |
|
2c488db27 memcg: clean up m... |
3356 |
rcu_assign_pointer(thresholds->primary, new); |
2e72b6347 memcg: implement ... |
3357 |
|
907860ed3 cgroups: make cft... |
3358 |
/* To be sure that nobody uses thresholds */ |
2e72b6347 memcg: implement ... |
3359 |
synchronize_rcu(); |
6611d8d76 memcg: only free ... |
3360 3361 3362 3363 3364 3365 |
/* If all events are unregistered, free the spare array */ if (!new) { kfree(thresholds->spare); thresholds->spare = NULL; } |
371528cae mm: memcg: Correc... |
3366 |
unlock: |
2e72b6347 memcg: implement ... |
3367 |
mutex_unlock(&memcg->thresholds_lock); |
2e72b6347 memcg: implement ... |
3368 |
} |
c1e862c1f memcg: new force_... |
3369 |
|
59b6f8734 memcg: make cgrou... |
3370 |
static void mem_cgroup_usage_unregister_event(struct mem_cgroup *memcg, |
347c4a874 memcg: remove cgr... |
3371 3372 |
struct eventfd_ctx *eventfd) { |
59b6f8734 memcg: make cgrou... |
3373 |
return __mem_cgroup_usage_unregister_event(memcg, eventfd, _MEM); |
347c4a874 memcg: remove cgr... |
3374 |
} |
59b6f8734 memcg: make cgrou... |
3375 |
static void memsw_cgroup_usage_unregister_event(struct mem_cgroup *memcg, |
347c4a874 memcg: remove cgr... |
3376 3377 |
struct eventfd_ctx *eventfd) { |
59b6f8734 memcg: make cgrou... |
3378 |
return __mem_cgroup_usage_unregister_event(memcg, eventfd, _MEMSWAP); |
347c4a874 memcg: remove cgr... |
3379 |
} |
59b6f8734 memcg: make cgrou... |
3380 |
static int mem_cgroup_oom_register_event(struct mem_cgroup *memcg, |
347c4a874 memcg: remove cgr... |
3381 |
struct eventfd_ctx *eventfd, const char *args) |
9490ff275 memcg: oom notifier |
3382 |
{ |
9490ff275 memcg: oom notifier |
3383 |
struct mem_cgroup_eventfd_list *event; |
9490ff275 memcg: oom notifier |
3384 |
|
9490ff275 memcg: oom notifier |
3385 3386 3387 |
event = kmalloc(sizeof(*event), GFP_KERNEL); if (!event) return -ENOMEM; |
1af8efe96 memcg: change mem... |
3388 |
spin_lock(&memcg_oom_lock); |
9490ff275 memcg: oom notifier |
3389 3390 3391 3392 3393 |
event->eventfd = eventfd; list_add(&event->list, &memcg->oom_notify); /* already in OOM ? */ |
c2b42d3ca memcg: convert me... |
3394 |
if (memcg->under_oom) |
9490ff275 memcg: oom notifier |
3395 |
eventfd_signal(eventfd, 1); |
1af8efe96 memcg: change mem... |
3396 |
spin_unlock(&memcg_oom_lock); |
9490ff275 memcg: oom notifier |
3397 3398 3399 |
return 0; } |
59b6f8734 memcg: make cgrou... |
3400 |
static void mem_cgroup_oom_unregister_event(struct mem_cgroup *memcg, |
347c4a874 memcg: remove cgr... |
3401 |
struct eventfd_ctx *eventfd) |
9490ff275 memcg: oom notifier |
3402 |
{ |
9490ff275 memcg: oom notifier |
3403 |
struct mem_cgroup_eventfd_list *ev, *tmp; |
9490ff275 memcg: oom notifier |
3404 |
|
1af8efe96 memcg: change mem... |
3405 |
spin_lock(&memcg_oom_lock); |
9490ff275 memcg: oom notifier |
3406 |
|
c0ff4b854 memcg: rename mem... |
3407 |
list_for_each_entry_safe(ev, tmp, &memcg->oom_notify, list) { |
9490ff275 memcg: oom notifier |
3408 3409 3410 3411 3412 |
if (ev->eventfd == eventfd) { list_del(&ev->list); kfree(ev); } } |
1af8efe96 memcg: change mem... |
3413 |
spin_unlock(&memcg_oom_lock); |
9490ff275 memcg: oom notifier |
3414 |
} |
2da8ca822 cgroup: replace c... |
3415 |
static int mem_cgroup_oom_control_read(struct seq_file *sf, void *v) |
3c11ecf44 memcg: oom kill d... |
3416 |
{ |
2da8ca822 cgroup: replace c... |
3417 |
struct mem_cgroup *memcg = mem_cgroup_from_css(seq_css(sf)); |
3c11ecf44 memcg: oom kill d... |
3418 |
|
791badbdb memcg: convert aw... |
3419 3420 |
seq_printf(sf, "oom_kill_disable %d ", memcg->oom_kill_disable); |
c2b42d3ca memcg: convert me... |
3421 3422 |
seq_printf(sf, "under_oom %d ", (bool)memcg->under_oom); |
3c11ecf44 memcg: oom kill d... |
3423 3424 |
return 0; } |
182446d08 cgroup: pass arou... |
3425 |
static int mem_cgroup_oom_control_write(struct cgroup_subsys_state *css, |
3c11ecf44 memcg: oom kill d... |
3426 3427 |
struct cftype *cft, u64 val) { |
182446d08 cgroup: pass arou... |
3428 |
struct mem_cgroup *memcg = mem_cgroup_from_css(css); |
3c11ecf44 memcg: oom kill d... |
3429 3430 |
/* cannot set to root cgroup and only 0 and 1 are allowed */ |
14208b0ec Merge branch 'for... |
3431 |
if (!css->parent || !((val == 0) || (val == 1))) |
3c11ecf44 memcg: oom kill d... |
3432 |
return -EINVAL; |
c0ff4b854 memcg: rename mem... |
3433 |
memcg->oom_kill_disable = val; |
4d845ebf4 memcg: fix wake u... |
3434 |
if (!val) |
c0ff4b854 memcg: rename mem... |
3435 |
memcg_oom_recover(memcg); |
3dae7fec5 mm: memcontrol: r... |
3436 |
|
3c11ecf44 memcg: oom kill d... |
3437 3438 |
return 0; } |
52ebea749 writeback: make b... |
3439 3440 3441 3442 3443 3444 |
#ifdef CONFIG_CGROUP_WRITEBACK struct list_head *mem_cgroup_cgwb_list(struct mem_cgroup *memcg) { return &memcg->cgwb_list; } |
841710aa6 writeback: implem... |
3445 3446 3447 3448 3449 3450 3451 3452 3453 |
static int memcg_wb_domain_init(struct mem_cgroup *memcg, gfp_t gfp) { return wb_domain_init(&memcg->cgwb_domain, gfp); } static void memcg_wb_domain_exit(struct mem_cgroup *memcg) { wb_domain_exit(&memcg->cgwb_domain); } |
2529bb3aa writeback: reset ... |
3454 3455 3456 3457 |
static void memcg_wb_domain_size_changed(struct mem_cgroup *memcg) { wb_domain_size_changed(&memcg->cgwb_domain); } |
841710aa6 writeback: implem... |
3458 3459 3460 3461 3462 3463 3464 3465 3466 |
struct wb_domain *mem_cgroup_wb_domain(struct bdi_writeback *wb) { struct mem_cgroup *memcg = mem_cgroup_from_css(wb->memcg_css); if (!memcg->css.parent) return NULL; return &memcg->cgwb_domain; } |
c2aa723a6 writeback: implem... |
3467 3468 3469 |
/** * mem_cgroup_wb_stats - retrieve writeback related stats from its memcg * @wb: bdi_writeback in question |
c5edf9cdc writeback: fix in... |
3470 3471 |
* @pfilepages: out parameter for number of file pages * @pheadroom: out parameter for number of allocatable pages according to memcg |
c2aa723a6 writeback: implem... |
3472 3473 3474 |
* @pdirty: out parameter for number of dirty pages * @pwriteback: out parameter for number of pages under writeback * |
c5edf9cdc writeback: fix in... |
3475 3476 3477 |
* Determine the numbers of file, headroom, dirty, and writeback pages in * @wb's memcg. File, dirty and writeback are self-explanatory. Headroom * is a bit more involved. |
c2aa723a6 writeback: implem... |
3478 |
* |
c5edf9cdc writeback: fix in... |
3479 3480 3481 3482 3483 |
* A memcg's headroom is "min(max, high) - used". In the hierarchy, the * headroom is calculated as the lowest headroom of itself and the * ancestors. Note that this doesn't consider the actual amount of * available memory in the system. The caller should further cap * *@pheadroom accordingly. |
c2aa723a6 writeback: implem... |
3484 |
*/ |
c5edf9cdc writeback: fix in... |
3485 3486 3487 |
void mem_cgroup_wb_stats(struct bdi_writeback *wb, unsigned long *pfilepages, unsigned long *pheadroom, unsigned long *pdirty, unsigned long *pwriteback) |
c2aa723a6 writeback: implem... |
3488 3489 3490 |
{ struct mem_cgroup *memcg = mem_cgroup_from_css(wb->memcg_css); struct mem_cgroup *parent; |
c2aa723a6 writeback: implem... |
3491 3492 3493 3494 3495 |
*pdirty = mem_cgroup_read_stat(memcg, MEM_CGROUP_STAT_DIRTY); /* this should eventually include NR_UNSTABLE_NFS */ *pwriteback = mem_cgroup_read_stat(memcg, MEM_CGROUP_STAT_WRITEBACK); |
c5edf9cdc writeback: fix in... |
3496 3497 3498 |
*pfilepages = mem_cgroup_nr_lru_pages(memcg, (1 << LRU_INACTIVE_FILE) | (1 << LRU_ACTIVE_FILE)); *pheadroom = PAGE_COUNTER_MAX; |
c2aa723a6 writeback: implem... |
3499 |
|
c2aa723a6 writeback: implem... |
3500 3501 3502 |
while ((parent = parent_mem_cgroup(memcg))) { unsigned long ceiling = min(memcg->memory.limit, memcg->high); unsigned long used = page_counter_read(&memcg->memory); |
c5edf9cdc writeback: fix in... |
3503 |
*pheadroom = min(*pheadroom, ceiling - min(ceiling, used)); |
c2aa723a6 writeback: implem... |
3504 3505 |
memcg = parent; } |
c2aa723a6 writeback: implem... |
3506 |
} |
841710aa6 writeback: implem... |
3507 3508 3509 3510 3511 3512 3513 3514 3515 3516 |
#else /* CONFIG_CGROUP_WRITEBACK */ static int memcg_wb_domain_init(struct mem_cgroup *memcg, gfp_t gfp) { return 0; } static void memcg_wb_domain_exit(struct mem_cgroup *memcg) { } |
2529bb3aa writeback: reset ... |
3517 3518 3519 |
static void memcg_wb_domain_size_changed(struct mem_cgroup *memcg) { } |
52ebea749 writeback: make b... |
3520 |
#endif /* CONFIG_CGROUP_WRITEBACK */ |
79bd9814e cgroup, memcg: mo... |
3521 |
/* |
3bc942f37 memcg: rename cgr... |
3522 3523 3524 3525 3526 3527 3528 3529 3530 3531 3532 3533 3534 |
* DO NOT USE IN NEW FILES. * * "cgroup.event_control" implementation. * * This is way over-engineered. It tries to support fully configurable * events for each user. Such level of flexibility is completely * unnecessary especially in the light of the planned unified hierarchy. * * Please deprecate this and replace with something simpler if at all * possible. */ /* |
79bd9814e cgroup, memcg: mo... |
3535 3536 3537 3538 |
* Unregister event and free resources. * * Gets called from workqueue. */ |
3bc942f37 memcg: rename cgr... |
3539 |
static void memcg_event_remove(struct work_struct *work) |
79bd9814e cgroup, memcg: mo... |
3540 |
{ |
3bc942f37 memcg: rename cgr... |
3541 3542 |
struct mem_cgroup_event *event = container_of(work, struct mem_cgroup_event, remove); |
59b6f8734 memcg: make cgrou... |
3543 |
struct mem_cgroup *memcg = event->memcg; |
79bd9814e cgroup, memcg: mo... |
3544 3545 |
remove_wait_queue(event->wqh, &event->wait); |
59b6f8734 memcg: make cgrou... |
3546 |
event->unregister_event(memcg, event->eventfd); |
79bd9814e cgroup, memcg: mo... |
3547 3548 3549 3550 3551 3552 |
/* Notify userspace the event is going away. */ eventfd_signal(event->eventfd, 1); eventfd_ctx_put(event->eventfd); kfree(event); |
59b6f8734 memcg: make cgrou... |
3553 |
css_put(&memcg->css); |
79bd9814e cgroup, memcg: mo... |
3554 3555 3556 3557 3558 3559 3560 |
} /* * Gets called on POLLHUP on eventfd when user closes it. * * Called with wqh->lock held and interrupts disabled. */ |
3bc942f37 memcg: rename cgr... |
3561 3562 |
static int memcg_event_wake(wait_queue_t *wait, unsigned mode, int sync, void *key) |
79bd9814e cgroup, memcg: mo... |
3563 |
{ |
3bc942f37 memcg: rename cgr... |
3564 3565 |
struct mem_cgroup_event *event = container_of(wait, struct mem_cgroup_event, wait); |
59b6f8734 memcg: make cgrou... |
3566 |
struct mem_cgroup *memcg = event->memcg; |
79bd9814e cgroup, memcg: mo... |
3567 3568 3569 3570 3571 3572 3573 3574 3575 3576 3577 3578 |
unsigned long flags = (unsigned long)key; if (flags & POLLHUP) { /* * If the event has been detached at cgroup removal, we * can simply return knowing the other side will cleanup * for us. * * We can't race against event freeing since the other * side will require wqh->lock via remove_wait_queue(), * which we hold. */ |
fba948078 cgroup, memcg: mo... |
3579 |
spin_lock(&memcg->event_list_lock); |
79bd9814e cgroup, memcg: mo... |
3580 3581 3582 3583 3584 3585 3586 3587 |
if (!list_empty(&event->list)) { list_del_init(&event->list); /* * We are in atomic context, but cgroup_event_remove() * may sleep, so we have to call it in workqueue. */ schedule_work(&event->remove); } |
fba948078 cgroup, memcg: mo... |
3588 |
spin_unlock(&memcg->event_list_lock); |
79bd9814e cgroup, memcg: mo... |
3589 3590 3591 3592 |
} return 0; } |
3bc942f37 memcg: rename cgr... |
3593 |
static void memcg_event_ptable_queue_proc(struct file *file, |
79bd9814e cgroup, memcg: mo... |
3594 3595 |
wait_queue_head_t *wqh, poll_table *pt) { |
3bc942f37 memcg: rename cgr... |
3596 3597 |
struct mem_cgroup_event *event = container_of(pt, struct mem_cgroup_event, pt); |
79bd9814e cgroup, memcg: mo... |
3598 3599 3600 3601 3602 3603 |
event->wqh = wqh; add_wait_queue(wqh, &event->wait); } /* |
3bc942f37 memcg: rename cgr... |
3604 3605 |
* DO NOT USE IN NEW FILES. * |
79bd9814e cgroup, memcg: mo... |
3606 3607 3608 3609 3610 |
* Parse input and register new cgroup event handler. * * Input must be in format '<event_fd> <control_fd> <args>'. * Interpretation of args is defined by control file implementation. */ |
451af504d cgroup: replace c... |
3611 3612 |
static ssize_t memcg_write_event_control(struct kernfs_open_file *of, char *buf, size_t nbytes, loff_t off) |
79bd9814e cgroup, memcg: mo... |
3613 |
{ |
451af504d cgroup: replace c... |
3614 |
struct cgroup_subsys_state *css = of_css(of); |
fba948078 cgroup, memcg: mo... |
3615 |
struct mem_cgroup *memcg = mem_cgroup_from_css(css); |
3bc942f37 memcg: rename cgr... |
3616 |
struct mem_cgroup_event *event; |
79bd9814e cgroup, memcg: mo... |
3617 3618 3619 3620 |
struct cgroup_subsys_state *cfile_css; unsigned int efd, cfd; struct fd efile; struct fd cfile; |
fba948078 cgroup, memcg: mo... |
3621 |
const char *name; |
79bd9814e cgroup, memcg: mo... |
3622 3623 |
char *endp; int ret; |
451af504d cgroup: replace c... |
3624 3625 3626 |
buf = strstrip(buf); efd = simple_strtoul(buf, &endp, 10); |
79bd9814e cgroup, memcg: mo... |
3627 3628 |
if (*endp != ' ') return -EINVAL; |
451af504d cgroup: replace c... |
3629 |
buf = endp + 1; |
79bd9814e cgroup, memcg: mo... |
3630 |
|
451af504d cgroup: replace c... |
3631 |
cfd = simple_strtoul(buf, &endp, 10); |
79bd9814e cgroup, memcg: mo... |
3632 3633 |
if ((*endp != ' ') && (*endp != '\0')) return -EINVAL; |
451af504d cgroup: replace c... |
3634 |
buf = endp + 1; |
79bd9814e cgroup, memcg: mo... |
3635 3636 3637 3638 |
event = kzalloc(sizeof(*event), GFP_KERNEL); if (!event) return -ENOMEM; |
59b6f8734 memcg: make cgrou... |
3639 |
event->memcg = memcg; |
79bd9814e cgroup, memcg: mo... |
3640 |
INIT_LIST_HEAD(&event->list); |
3bc942f37 memcg: rename cgr... |
3641 3642 3643 |
init_poll_funcptr(&event->pt, memcg_event_ptable_queue_proc); init_waitqueue_func_entry(&event->wait, memcg_event_wake); INIT_WORK(&event->remove, memcg_event_remove); |
79bd9814e cgroup, memcg: mo... |
3644 3645 3646 3647 3648 3649 3650 3651 3652 3653 3654 3655 3656 3657 3658 3659 3660 3661 3662 3663 3664 3665 3666 3667 |
efile = fdget(efd); if (!efile.file) { ret = -EBADF; goto out_kfree; } event->eventfd = eventfd_ctx_fileget(efile.file); if (IS_ERR(event->eventfd)) { ret = PTR_ERR(event->eventfd); goto out_put_efile; } cfile = fdget(cfd); if (!cfile.file) { ret = -EBADF; goto out_put_eventfd; } /* the process need read permission on control file */ /* AV: shouldn't we check that it's been opened for read instead? */ ret = inode_permission(file_inode(cfile.file), MAY_READ); if (ret < 0) goto out_put_cfile; |
79bd9814e cgroup, memcg: mo... |
3668 |
/* |
fba948078 cgroup, memcg: mo... |
3669 3670 3671 3672 |
* Determine the event callbacks and set them in @event. This used * to be done via struct cftype but cgroup core no longer knows * about these events. The following is crude but the whole thing * is for compatibility anyway. |
3bc942f37 memcg: rename cgr... |
3673 3674 |
* * DO NOT ADD NEW FILES. |
fba948078 cgroup, memcg: mo... |
3675 |
*/ |
b583043e9 kill f_dentry uses |
3676 |
name = cfile.file->f_path.dentry->d_name.name; |
fba948078 cgroup, memcg: mo... |
3677 3678 3679 3680 3681 3682 3683 3684 3685 3686 3687 |
if (!strcmp(name, "memory.usage_in_bytes")) { event->register_event = mem_cgroup_usage_register_event; event->unregister_event = mem_cgroup_usage_unregister_event; } else if (!strcmp(name, "memory.oom_control")) { event->register_event = mem_cgroup_oom_register_event; event->unregister_event = mem_cgroup_oom_unregister_event; } else if (!strcmp(name, "memory.pressure_level")) { event->register_event = vmpressure_register_event; event->unregister_event = vmpressure_unregister_event; } else if (!strcmp(name, "memory.memsw.usage_in_bytes")) { |
347c4a874 memcg: remove cgr... |
3688 3689 |
event->register_event = memsw_cgroup_usage_register_event; event->unregister_event = memsw_cgroup_usage_unregister_event; |
fba948078 cgroup, memcg: mo... |
3690 3691 3692 3693 3694 3695 |
} else { ret = -EINVAL; goto out_put_cfile; } /* |
b5557c4c3 memcg: cgroup_wri... |
3696 3697 3698 |
* Verify @cfile should belong to @css. Also, remaining events are * automatically removed on cgroup destruction but the removal is * asynchronous, so take an extra ref on @css. |
79bd9814e cgroup, memcg: mo... |
3699 |
*/ |
b583043e9 kill f_dentry uses |
3700 |
cfile_css = css_tryget_online_from_dir(cfile.file->f_path.dentry->d_parent, |
ec903c0c8 cgroup: rename cs... |
3701 |
&memory_cgrp_subsys); |
79bd9814e cgroup, memcg: mo... |
3702 |
ret = -EINVAL; |
5a17f543e cgroup: improve c... |
3703 |
if (IS_ERR(cfile_css)) |
79bd9814e cgroup, memcg: mo... |
3704 |
goto out_put_cfile; |
5a17f543e cgroup: improve c... |
3705 3706 |
if (cfile_css != css) { css_put(cfile_css); |
79bd9814e cgroup, memcg: mo... |
3707 |
goto out_put_cfile; |
5a17f543e cgroup: improve c... |
3708 |
} |
79bd9814e cgroup, memcg: mo... |
3709 |
|
451af504d cgroup: replace c... |
3710 |
ret = event->register_event(memcg, event->eventfd, buf); |
79bd9814e cgroup, memcg: mo... |
3711 3712 3713 3714 |
if (ret) goto out_put_css; efile.file->f_op->poll(efile.file, &event->pt); |
fba948078 cgroup, memcg: mo... |
3715 3716 3717 |
spin_lock(&memcg->event_list_lock); list_add(&event->list, &memcg->event_list); spin_unlock(&memcg->event_list_lock); |
79bd9814e cgroup, memcg: mo... |
3718 3719 3720 |
fdput(cfile); fdput(efile); |
451af504d cgroup: replace c... |
3721 |
return nbytes; |
79bd9814e cgroup, memcg: mo... |
3722 3723 |
out_put_css: |
b5557c4c3 memcg: cgroup_wri... |
3724 |
css_put(css); |
79bd9814e cgroup, memcg: mo... |
3725 3726 3727 3728 3729 3730 3731 3732 3733 3734 3735 |
out_put_cfile: fdput(cfile); out_put_eventfd: eventfd_ctx_put(event->eventfd); out_put_efile: fdput(efile); out_kfree: kfree(event); return ret; } |
241994ed8 mm: memcontrol: d... |
3736 |
static struct cftype mem_cgroup_legacy_files[] = { |
8cdea7c05 Memory controller... |
3737 |
{ |
0eea10301 Memory controller... |
3738 |
.name = "usage_in_bytes", |
8c7c6e34a memcg: mem+swap c... |
3739 |
.private = MEMFILE_PRIVATE(_MEM, RES_USAGE), |
791badbdb memcg: convert aw... |
3740 |
.read_u64 = mem_cgroup_read_u64, |
8cdea7c05 Memory controller... |
3741 3742 |
}, { |
c84872e16 memcgroup: add th... |
3743 |
.name = "max_usage_in_bytes", |
8c7c6e34a memcg: mem+swap c... |
3744 |
.private = MEMFILE_PRIVATE(_MEM, RES_MAX_USAGE), |
6770c64e5 cgroup: replace c... |
3745 |
.write = mem_cgroup_reset, |
791badbdb memcg: convert aw... |
3746 |
.read_u64 = mem_cgroup_read_u64, |
c84872e16 memcgroup: add th... |
3747 3748 |
}, { |
0eea10301 Memory controller... |
3749 |
.name = "limit_in_bytes", |
8c7c6e34a memcg: mem+swap c... |
3750 |
.private = MEMFILE_PRIVATE(_MEM, RES_LIMIT), |
451af504d cgroup: replace c... |
3751 |
.write = mem_cgroup_write, |
791badbdb memcg: convert aw... |
3752 |
.read_u64 = mem_cgroup_read_u64, |
8cdea7c05 Memory controller... |
3753 3754 |
}, { |
296c81d89 memory controller... |
3755 3756 |
.name = "soft_limit_in_bytes", .private = MEMFILE_PRIVATE(_MEM, RES_SOFT_LIMIT), |
451af504d cgroup: replace c... |
3757 |
.write = mem_cgroup_write, |
791badbdb memcg: convert aw... |
3758 |
.read_u64 = mem_cgroup_read_u64, |
296c81d89 memory controller... |
3759 3760 |
}, { |
8cdea7c05 Memory controller... |
3761 |
.name = "failcnt", |
8c7c6e34a memcg: mem+swap c... |
3762 |
.private = MEMFILE_PRIVATE(_MEM, RES_FAILCNT), |
6770c64e5 cgroup: replace c... |
3763 |
.write = mem_cgroup_reset, |
791badbdb memcg: convert aw... |
3764 |
.read_u64 = mem_cgroup_read_u64, |
8cdea7c05 Memory controller... |
3765 |
}, |
8697d3319 Memory controller... |
3766 |
{ |
d2ceb9b7d memory cgroup enh... |
3767 |
.name = "stat", |
2da8ca822 cgroup: replace c... |
3768 |
.seq_show = memcg_stat_show, |
d2ceb9b7d memory cgroup enh... |
3769 |
}, |
c1e862c1f memcg: new force_... |
3770 3771 |
{ .name = "force_empty", |
6770c64e5 cgroup: replace c... |
3772 |
.write = mem_cgroup_force_empty_write, |
c1e862c1f memcg: new force_... |
3773 |
}, |
18f59ea7d memcg: memory cgr... |
3774 3775 3776 3777 3778 |
{ .name = "use_hierarchy", .write_u64 = mem_cgroup_hierarchy_write, .read_u64 = mem_cgroup_hierarchy_read, }, |
a7885eb8a memcg: swappiness |
3779 |
{ |
3bc942f37 memcg: rename cgr... |
3780 |
.name = "cgroup.event_control", /* XXX: for compat */ |
451af504d cgroup: replace c... |
3781 |
.write = memcg_write_event_control, |
7dbdb199d cgroup: replace c... |
3782 |
.flags = CFTYPE_NO_PREFIX | CFTYPE_WORLD_WRITABLE, |
79bd9814e cgroup, memcg: mo... |
3783 3784 |
}, { |
a7885eb8a memcg: swappiness |
3785 3786 3787 3788 |
.name = "swappiness", .read_u64 = mem_cgroup_swappiness_read, .write_u64 = mem_cgroup_swappiness_write, }, |
7dc74be03 memcg: add interf... |
3789 3790 3791 3792 3793 |
{ .name = "move_charge_at_immigrate", .read_u64 = mem_cgroup_move_charge_read, .write_u64 = mem_cgroup_move_charge_write, }, |
9490ff275 memcg: oom notifier |
3794 3795 |
{ .name = "oom_control", |
2da8ca822 cgroup: replace c... |
3796 |
.seq_show = mem_cgroup_oom_control_read, |
3c11ecf44 memcg: oom kill d... |
3797 |
.write_u64 = mem_cgroup_oom_control_write, |
9490ff275 memcg: oom notifier |
3798 3799 |
.private = MEMFILE_PRIVATE(_OOM_TYPE, OOM_CONTROL), }, |
70ddf637e memcg: add memory... |
3800 3801 |
{ .name = "pressure_level", |
70ddf637e memcg: add memory... |
3802 |
}, |
406eb0c9b memcg: add memory... |
3803 3804 3805 |
#ifdef CONFIG_NUMA { .name = "numa_stat", |
2da8ca822 cgroup: replace c... |
3806 |
.seq_show = memcg_numa_stat_show, |
406eb0c9b memcg: add memory... |
3807 3808 |
}, #endif |
510fc4e11 memcg: kmem accou... |
3809 3810 3811 |
{ .name = "kmem.limit_in_bytes", .private = MEMFILE_PRIVATE(_KMEM, RES_LIMIT), |
451af504d cgroup: replace c... |
3812 |
.write = mem_cgroup_write, |
791badbdb memcg: convert aw... |
3813 |
.read_u64 = mem_cgroup_read_u64, |
510fc4e11 memcg: kmem accou... |
3814 3815 3816 3817 |
}, { .name = "kmem.usage_in_bytes", .private = MEMFILE_PRIVATE(_KMEM, RES_USAGE), |
791badbdb memcg: convert aw... |
3818 |
.read_u64 = mem_cgroup_read_u64, |
510fc4e11 memcg: kmem accou... |
3819 3820 3821 3822 |
}, { .name = "kmem.failcnt", .private = MEMFILE_PRIVATE(_KMEM, RES_FAILCNT), |
6770c64e5 cgroup: replace c... |
3823 |
.write = mem_cgroup_reset, |
791badbdb memcg: convert aw... |
3824 |
.read_u64 = mem_cgroup_read_u64, |
510fc4e11 memcg: kmem accou... |
3825 3826 3827 3828 |
}, { .name = "kmem.max_usage_in_bytes", .private = MEMFILE_PRIVATE(_KMEM, RES_MAX_USAGE), |
6770c64e5 cgroup: replace c... |
3829 |
.write = mem_cgroup_reset, |
791badbdb memcg: convert aw... |
3830 |
.read_u64 = mem_cgroup_read_u64, |
510fc4e11 memcg: kmem accou... |
3831 |
}, |
749c54151 memcg: aggregate ... |
3832 3833 3834 |
#ifdef CONFIG_SLABINFO { .name = "kmem.slabinfo", |
b047501cd memcg: use generi... |
3835 3836 3837 3838 |
.seq_start = slab_start, .seq_next = slab_next, .seq_stop = slab_stop, .seq_show = memcg_slab_show, |
749c54151 memcg: aggregate ... |
3839 3840 |
}, #endif |
d55f90bfa net: drop tcp_mem... |
3841 3842 3843 3844 3845 3846 3847 3848 3849 3850 3851 3852 3853 3854 3855 3856 3857 3858 3859 3860 3861 3862 3863 |
{ .name = "kmem.tcp.limit_in_bytes", .private = MEMFILE_PRIVATE(_TCP, RES_LIMIT), .write = mem_cgroup_write, .read_u64 = mem_cgroup_read_u64, }, { .name = "kmem.tcp.usage_in_bytes", .private = MEMFILE_PRIVATE(_TCP, RES_USAGE), .read_u64 = mem_cgroup_read_u64, }, { .name = "kmem.tcp.failcnt", .private = MEMFILE_PRIVATE(_TCP, RES_FAILCNT), .write = mem_cgroup_reset, .read_u64 = mem_cgroup_read_u64, }, { .name = "kmem.tcp.max_usage_in_bytes", .private = MEMFILE_PRIVATE(_TCP, RES_MAX_USAGE), .write = mem_cgroup_reset, .read_u64 = mem_cgroup_read_u64, }, |
6bc103498 cgroup: convert m... |
3864 |
{ }, /* terminate */ |
af36f906c memcg: always cre... |
3865 |
}; |
8c7c6e34a memcg: mem+swap c... |
3866 |
|
c0ff4b854 memcg: rename mem... |
3867 |
static int alloc_mem_cgroup_per_zone_info(struct mem_cgroup *memcg, int node) |
6d12e2d8d per-zone and recl... |
3868 3869 |
{ struct mem_cgroup_per_node *pn; |
1ecaab2bd per-zone and recl... |
3870 |
struct mem_cgroup_per_zone *mz; |
41e3355de memcg: fix node_s... |
3871 |
int zone, tmp = node; |
1ecaab2bd per-zone and recl... |
3872 3873 3874 3875 3876 3877 3878 3879 |
/* * 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... |
3880 3881 |
if (!node_state(node, N_NORMAL_MEMORY)) tmp = -1; |
17295c88a memcg: use [kv]za... |
3882 |
pn = kzalloc_node(sizeof(*pn), GFP_KERNEL, tmp); |
6d12e2d8d per-zone and recl... |
3883 3884 |
if (!pn) return 1; |
1ecaab2bd per-zone and recl... |
3885 |
|
1ecaab2bd per-zone and recl... |
3886 3887 |
for (zone = 0; zone < MAX_NR_ZONES; zone++) { mz = &pn->zoneinfo[zone]; |
bea8c150a memcg: fix hotplu... |
3888 |
lruvec_init(&mz->lruvec); |
bb4cc1a8b revert "memcg: ge... |
3889 3890 |
mz->usage_in_excess = 0; mz->on_tree = false; |
d79154bb5 memcg: replace me... |
3891 |
mz->memcg = memcg; |
1ecaab2bd per-zone and recl... |
3892 |
} |
54f72fe02 memcg: clean up m... |
3893 |
memcg->nodeinfo[node] = pn; |
6d12e2d8d per-zone and recl... |
3894 3895 |
return 0; } |
c0ff4b854 memcg: rename mem... |
3896 |
static void free_mem_cgroup_per_zone_info(struct mem_cgroup *memcg, int node) |
1ecaab2bd per-zone and recl... |
3897 |
{ |
54f72fe02 memcg: clean up m... |
3898 |
kfree(memcg->nodeinfo[node]); |
1ecaab2bd per-zone and recl... |
3899 |
} |
0b8f73e10 mm: memcontrol: c... |
3900 |
static void mem_cgroup_free(struct mem_cgroup *memcg) |
59927fb98 memcg: free mem_c... |
3901 |
{ |
c8b2a36fb memcg: execute th... |
3902 |
int node; |
59927fb98 memcg: free mem_c... |
3903 |
|
0b8f73e10 mm: memcontrol: c... |
3904 |
memcg_wb_domain_exit(memcg); |
c8b2a36fb memcg: execute th... |
3905 3906 |
for_each_node(node) free_mem_cgroup_per_zone_info(memcg, node); |
c8b2a36fb memcg: execute th... |
3907 |
free_percpu(memcg->stat); |
8ff69e2c8 memcg: do not use... |
3908 |
kfree(memcg); |
59927fb98 memcg: free mem_c... |
3909 |
} |
3afe36b1f memcg: always fre... |
3910 |
|
0b8f73e10 mm: memcontrol: c... |
3911 |
static struct mem_cgroup *mem_cgroup_alloc(void) |
8cdea7c05 Memory controller... |
3912 |
{ |
d142e3e66 memcg: split part... |
3913 |
struct mem_cgroup *memcg; |
0b8f73e10 mm: memcontrol: c... |
3914 |
size_t size; |
6d12e2d8d per-zone and recl... |
3915 |
int node; |
8cdea7c05 Memory controller... |
3916 |
|
0b8f73e10 mm: memcontrol: c... |
3917 3918 3919 3920 |
size = sizeof(struct mem_cgroup); size += nr_node_ids * sizeof(struct mem_cgroup_per_node *); memcg = kzalloc(size, GFP_KERNEL); |
c0ff4b854 memcg: rename mem... |
3921 |
if (!memcg) |
0b8f73e10 mm: memcontrol: c... |
3922 3923 3924 3925 3926 |
return NULL; memcg->stat = alloc_percpu(struct mem_cgroup_stat_cpu); if (!memcg->stat) goto fail; |
78fb74669 Memory controller... |
3927 |
|
3ed28fa10 memcg: cleanup fo... |
3928 |
for_each_node(node) |
c0ff4b854 memcg: rename mem... |
3929 |
if (alloc_mem_cgroup_per_zone_info(memcg, node)) |
0b8f73e10 mm: memcontrol: c... |
3930 |
goto fail; |
f64c3f549 memory controller... |
3931 |
|
0b8f73e10 mm: memcontrol: c... |
3932 3933 |
if (memcg_wb_domain_init(memcg, GFP_KERNEL)) goto fail; |
28dbc4b6a memcg: memory cgr... |
3934 |
|
f7e1cb6ec mm: memcontrol: a... |
3935 |
INIT_WORK(&memcg->high_work, high_work_func); |
d142e3e66 memcg: split part... |
3936 3937 |
memcg->last_scanned_node = MAX_NUMNODES; INIT_LIST_HEAD(&memcg->oom_notify); |
d142e3e66 memcg: split part... |
3938 3939 |
mutex_init(&memcg->thresholds_lock); spin_lock_init(&memcg->move_lock); |
70ddf637e memcg: add memory... |
3940 |
vmpressure_init(&memcg->vmpressure); |
fba948078 cgroup, memcg: mo... |
3941 3942 |
INIT_LIST_HEAD(&memcg->event_list); spin_lock_init(&memcg->event_list_lock); |
d886f4e48 mm: memcontrol: r... |
3943 |
memcg->socket_pressure = jiffies; |
127424c86 mm: memcontrol: m... |
3944 |
#ifndef CONFIG_SLOB |
900a38f02 memcg: zap kmem_a... |
3945 |
memcg->kmemcg_id = -1; |
900a38f02 memcg: zap kmem_a... |
3946 |
#endif |
52ebea749 writeback: make b... |
3947 3948 3949 |
#ifdef CONFIG_CGROUP_WRITEBACK INIT_LIST_HEAD(&memcg->cgwb_list); #endif |
0b8f73e10 mm: memcontrol: c... |
3950 3951 3952 3953 |
return memcg; fail: mem_cgroup_free(memcg); return NULL; |
d142e3e66 memcg: split part... |
3954 |
} |
0b8f73e10 mm: memcontrol: c... |
3955 3956 |
static struct cgroup_subsys_state * __ref mem_cgroup_css_alloc(struct cgroup_subsys_state *parent_css) |
d142e3e66 memcg: split part... |
3957 |
{ |
0b8f73e10 mm: memcontrol: c... |
3958 3959 3960 |
struct mem_cgroup *parent = mem_cgroup_from_css(parent_css); struct mem_cgroup *memcg; long error = -ENOMEM; |
d142e3e66 memcg: split part... |
3961 |
|
0b8f73e10 mm: memcontrol: c... |
3962 3963 3964 |
memcg = mem_cgroup_alloc(); if (!memcg) return ERR_PTR(error); |
d142e3e66 memcg: split part... |
3965 |
|
0b8f73e10 mm: memcontrol: c... |
3966 3967 3968 3969 3970 3971 3972 3973 |
memcg->high = PAGE_COUNTER_MAX; memcg->soft_limit = PAGE_COUNTER_MAX; if (parent) { memcg->swappiness = mem_cgroup_swappiness(parent); memcg->oom_kill_disable = parent->oom_kill_disable; } if (parent && parent->use_hierarchy) { memcg->use_hierarchy = true; |
3e32cb2e0 mm: memcontrol: l... |
3974 |
page_counter_init(&memcg->memory, &parent->memory); |
37e843511 mm: memcontrol: c... |
3975 |
page_counter_init(&memcg->swap, &parent->swap); |
3e32cb2e0 mm: memcontrol: l... |
3976 3977 |
page_counter_init(&memcg->memsw, &parent->memsw); page_counter_init(&memcg->kmem, &parent->kmem); |
0db152981 mm: memcontrol: f... |
3978 |
page_counter_init(&memcg->tcpmem, &parent->tcpmem); |
18f59ea7d memcg: memory cgr... |
3979 |
} else { |
3e32cb2e0 mm: memcontrol: l... |
3980 |
page_counter_init(&memcg->memory, NULL); |
37e843511 mm: memcontrol: c... |
3981 |
page_counter_init(&memcg->swap, NULL); |
3e32cb2e0 mm: memcontrol: l... |
3982 3983 |
page_counter_init(&memcg->memsw, NULL); page_counter_init(&memcg->kmem, NULL); |
0db152981 mm: memcontrol: f... |
3984 |
page_counter_init(&memcg->tcpmem, NULL); |
8c7f6edbd cgroup: mark subs... |
3985 3986 3987 3988 3989 |
/* * Deeper hierachy with use_hierarchy == false doesn't make * much sense so let cgroup subsystem know about this * unfortunate state in our controller. */ |
d142e3e66 memcg: split part... |
3990 |
if (parent != root_mem_cgroup) |
073219e99 cgroup: clean up ... |
3991 |
memory_cgrp_subsys.broken_hierarchy = true; |
18f59ea7d memcg: memory cgr... |
3992 |
} |
d64416377 memcg: rework mem... |
3993 |
|
0b8f73e10 mm: memcontrol: c... |
3994 3995 3996 3997 3998 |
/* The following stuff does not apply to the root */ if (!parent) { root_mem_cgroup = memcg; return &memcg->css; } |
b313aeee2 mm: memcontrol: e... |
3999 |
error = memcg_online_kmem(memcg); |
0b8f73e10 mm: memcontrol: c... |
4000 4001 |
if (error) goto fail; |
127424c86 mm: memcontrol: m... |
4002 |
|
f7e1cb6ec mm: memcontrol: a... |
4003 |
if (cgroup_subsys_on_dfl(memory_cgrp_subsys) && !cgroup_memory_nosocket) |
ef12947c9 mm: memcontrol: s... |
4004 |
static_branch_inc(&memcg_sockets_enabled_key); |
f7e1cb6ec mm: memcontrol: a... |
4005 |
|
0b8f73e10 mm: memcontrol: c... |
4006 4007 4008 4009 4010 4011 4012 4013 4014 4015 4016 |
return &memcg->css; fail: mem_cgroup_free(memcg); return NULL; } static int mem_cgroup_css_online(struct cgroup_subsys_state *css) { if (css->id > MEM_CGROUP_ID_MAX) return -ENOSPC; |
2f7dd7a41 mm: memcontrol: d... |
4017 4018 |
return 0; |
8cdea7c05 Memory controller... |
4019 |
} |
eb95419b0 cgroup: pass arou... |
4020 |
static void mem_cgroup_css_offline(struct cgroup_subsys_state *css) |
df878fb04 memory cgroup enh... |
4021 |
{ |
eb95419b0 cgroup: pass arou... |
4022 |
struct mem_cgroup *memcg = mem_cgroup_from_css(css); |
3bc942f37 memcg: rename cgr... |
4023 |
struct mem_cgroup_event *event, *tmp; |
79bd9814e cgroup, memcg: mo... |
4024 4025 4026 4027 4028 4029 |
/* * Unregister events and notify userspace. * Notify userspace about cgroup removing only after rmdir of cgroup * directory to avoid race between userspace and kernelspace. */ |
fba948078 cgroup, memcg: mo... |
4030 4031 |
spin_lock(&memcg->event_list_lock); list_for_each_entry_safe(event, tmp, &memcg->event_list, list) { |
79bd9814e cgroup, memcg: mo... |
4032 4033 4034 |
list_del_init(&event->list); schedule_work(&event->remove); } |
fba948078 cgroup, memcg: mo... |
4035 |
spin_unlock(&memcg->event_list_lock); |
ec64f5154 cgroup: fix frequ... |
4036 |
|
567e9ab2e mm: memcontrol: g... |
4037 |
memcg_offline_kmem(memcg); |
52ebea749 writeback: make b... |
4038 |
wb_memcg_offline(memcg); |
df878fb04 memory cgroup enh... |
4039 |
} |
6df38689e mm: memcontrol: f... |
4040 4041 4042 4043 4044 4045 |
static void mem_cgroup_css_released(struct cgroup_subsys_state *css) { struct mem_cgroup *memcg = mem_cgroup_from_css(css); invalidate_reclaim_iterators(memcg); } |
eb95419b0 cgroup: pass arou... |
4046 |
static void mem_cgroup_css_free(struct cgroup_subsys_state *css) |
8cdea7c05 Memory controller... |
4047 |
{ |
eb95419b0 cgroup: pass arou... |
4048 |
struct mem_cgroup *memcg = mem_cgroup_from_css(css); |
c268e9946 memcg: fix hierar... |
4049 |
|
f7e1cb6ec mm: memcontrol: a... |
4050 |
if (cgroup_subsys_on_dfl(memory_cgrp_subsys) && !cgroup_memory_nosocket) |
ef12947c9 mm: memcontrol: s... |
4051 |
static_branch_dec(&memcg_sockets_enabled_key); |
127424c86 mm: memcontrol: m... |
4052 |
|
0db152981 mm: memcontrol: f... |
4053 |
if (!cgroup_subsys_on_dfl(memory_cgrp_subsys) && memcg->tcpmem_active) |
d55f90bfa net: drop tcp_mem... |
4054 |
static_branch_dec(&memcg_sockets_enabled_key); |
3893e302f mm: memcontrol: s... |
4055 |
|
0b8f73e10 mm: memcontrol: c... |
4056 4057 4058 |
vmpressure_cleanup(&memcg->vmpressure); cancel_work_sync(&memcg->high_work); mem_cgroup_remove_from_trees(memcg); |
d886f4e48 mm: memcontrol: r... |
4059 |
memcg_free_kmem(memcg); |
0b8f73e10 mm: memcontrol: c... |
4060 |
mem_cgroup_free(memcg); |
8cdea7c05 Memory controller... |
4061 |
} |
1ced953b1 blkcg, memcg: mak... |
4062 4063 4064 4065 4066 4067 4068 4069 4070 4071 4072 4073 4074 4075 4076 4077 |
/** * mem_cgroup_css_reset - reset the states of a mem_cgroup * @css: the target css * * Reset the states of the mem_cgroup associated with @css. This is * invoked when the userland requests disabling on the default hierarchy * but the memcg is pinned through dependency. The memcg should stop * applying policies and should revert to the vanilla state as it may be * made visible again. * * The current implementation only resets the essential configurations. * This needs to be expanded to cover all the visible parts. */ static void mem_cgroup_css_reset(struct cgroup_subsys_state *css) { struct mem_cgroup *memcg = mem_cgroup_from_css(css); |
d334c9bcb mm: memcontrol: c... |
4078 4079 4080 4081 4082 |
page_counter_limit(&memcg->memory, PAGE_COUNTER_MAX); page_counter_limit(&memcg->swap, PAGE_COUNTER_MAX); page_counter_limit(&memcg->memsw, PAGE_COUNTER_MAX); page_counter_limit(&memcg->kmem, PAGE_COUNTER_MAX); page_counter_limit(&memcg->tcpmem, PAGE_COUNTER_MAX); |
241994ed8 mm: memcontrol: d... |
4083 4084 |
memcg->low = 0; memcg->high = PAGE_COUNTER_MAX; |
24d404dc1 mm: memcontrol: s... |
4085 |
memcg->soft_limit = PAGE_COUNTER_MAX; |
2529bb3aa writeback: reset ... |
4086 |
memcg_wb_domain_size_changed(memcg); |
1ced953b1 blkcg, memcg: mak... |
4087 |
} |
024914477 memcg: move charg... |
4088 |
#ifdef CONFIG_MMU |
7dc74be03 memcg: add interf... |
4089 |
/* Handlers for move charge at task migration. */ |
854ffa8d1 memcg: improve pe... |
4090 |
static int mem_cgroup_do_precharge(unsigned long count) |
7dc74be03 memcg: add interf... |
4091 |
{ |
05b843012 mm: memcontrol: u... |
4092 |
int ret; |
9476db974 mm: memcontrol: s... |
4093 |
|
d0164adc8 mm, page_alloc: d... |
4094 4095 |
/* Try a single bulk charge without reclaim first, kswapd may wake */ ret = try_charge(mc.to, GFP_KERNEL & ~__GFP_DIRECT_RECLAIM, count); |
9476db974 mm: memcontrol: s... |
4096 |
if (!ret) { |
854ffa8d1 memcg: improve pe... |
4097 |
mc.precharge += count; |
854ffa8d1 memcg: improve pe... |
4098 4099 |
return ret; } |
9476db974 mm: memcontrol: s... |
4100 4101 |
/* Try charges one by one with reclaim */ |
854ffa8d1 memcg: improve pe... |
4102 |
while (count--) { |
00501b531 mm: memcontrol: r... |
4103 |
ret = try_charge(mc.to, GFP_KERNEL & ~__GFP_NORETRY, 1); |
38c5d72f3 memcg: simplify L... |
4104 |
if (ret) |
38c5d72f3 memcg: simplify L... |
4105 |
return ret; |
854ffa8d1 memcg: improve pe... |
4106 |
mc.precharge++; |
9476db974 mm: memcontrol: s... |
4107 |
cond_resched(); |
854ffa8d1 memcg: improve pe... |
4108 |
} |
9476db974 mm: memcontrol: s... |
4109 |
return 0; |
4ffef5fef memcg: move charg... |
4110 4111 4112 |
} /** |
8d32ff844 memcg: clean up e... |
4113 |
* get_mctgt_type - get target type of moving charge |
4ffef5fef memcg: move charg... |
4114 4115 4116 |
* @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... |
4117 |
* @target: the pointer the target page or swap ent will be stored(can be NULL) |
4ffef5fef memcg: move charg... |
4118 4119 4120 4121 4122 4123 |
* * 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... |
4124 4125 4126 |
* 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... |
4127 4128 4129 |
* * Called with pte lock held. */ |
4ffef5fef memcg: move charg... |
4130 4131 |
union mc_target { struct page *page; |
024914477 memcg: move charg... |
4132 |
swp_entry_t ent; |
4ffef5fef memcg: move charg... |
4133 |
}; |
4ffef5fef memcg: move charg... |
4134 |
enum mc_target_type { |
8d32ff844 memcg: clean up e... |
4135 |
MC_TARGET_NONE = 0, |
4ffef5fef memcg: move charg... |
4136 |
MC_TARGET_PAGE, |
024914477 memcg: move charg... |
4137 |
MC_TARGET_SWAP, |
4ffef5fef memcg: move charg... |
4138 |
}; |
90254a658 memcg: clean up m... |
4139 4140 |
static struct page *mc_handle_present_pte(struct vm_area_struct *vma, unsigned long addr, pte_t ptent) |
4ffef5fef memcg: move charg... |
4141 |
{ |
90254a658 memcg: clean up m... |
4142 |
struct page *page = vm_normal_page(vma, addr, ptent); |
4ffef5fef memcg: move charg... |
4143 |
|
90254a658 memcg: clean up m... |
4144 4145 4146 |
if (!page || !page_mapped(page)) return NULL; if (PageAnon(page)) { |
1dfab5abc mm: memcontrol: f... |
4147 |
if (!(mc.flags & MOVE_ANON)) |
90254a658 memcg: clean up m... |
4148 |
return NULL; |
1dfab5abc mm: memcontrol: f... |
4149 4150 4151 4152 |
} else { if (!(mc.flags & MOVE_FILE)) return NULL; } |
90254a658 memcg: clean up m... |
4153 4154 4155 4156 4157 |
if (!get_page_unless_zero(page)) return NULL; return page; } |
4b91355e9 memcg: fix/change... |
4158 |
#ifdef CONFIG_SWAP |
90254a658 memcg: clean up m... |
4159 4160 4161 |
static struct page *mc_handle_swap_pte(struct vm_area_struct *vma, unsigned long addr, pte_t ptent, swp_entry_t *entry) { |
90254a658 memcg: clean up m... |
4162 4163 |
struct page *page = NULL; swp_entry_t ent = pte_to_swp_entry(ptent); |
1dfab5abc mm: memcontrol: f... |
4164 |
if (!(mc.flags & MOVE_ANON) || non_swap_entry(ent)) |
90254a658 memcg: clean up m... |
4165 |
return NULL; |
4b91355e9 memcg: fix/change... |
4166 4167 4168 4169 |
/* * Because lookup_swap_cache() updates some statistics counter, * we call find_get_page() with swapper_space directly. */ |
33806f06d swap: make each s... |
4170 |
page = find_get_page(swap_address_space(ent), ent.val); |
7941d2145 mm: memcontrol: d... |
4171 |
if (do_memsw_account()) |
90254a658 memcg: clean up m... |
4172 4173 4174 4175 |
entry->val = ent.val; return page; } |
4b91355e9 memcg: fix/change... |
4176 4177 4178 4179 4180 4181 4182 |
#else static struct page *mc_handle_swap_pte(struct vm_area_struct *vma, unsigned long addr, pte_t ptent, swp_entry_t *entry) { return NULL; } #endif |
90254a658 memcg: clean up m... |
4183 |
|
87946a722 memcg: move charg... |
4184 4185 4186 4187 |
static struct page *mc_handle_file_pte(struct vm_area_struct *vma, unsigned long addr, pte_t ptent, swp_entry_t *entry) { struct page *page = NULL; |
87946a722 memcg: move charg... |
4188 4189 4190 4191 4192 |
struct address_space *mapping; pgoff_t pgoff; if (!vma->vm_file) /* anonymous vma */ return NULL; |
1dfab5abc mm: memcontrol: f... |
4193 |
if (!(mc.flags & MOVE_FILE)) |
87946a722 memcg: move charg... |
4194 |
return NULL; |
87946a722 memcg: move charg... |
4195 |
mapping = vma->vm_file->f_mapping; |
0661a3361 mm: remove rest u... |
4196 |
pgoff = linear_page_index(vma, addr); |
87946a722 memcg: move charg... |
4197 4198 |
/* page is moved even if it's not RSS of this task(page-faulted). */ |
aa3b18955 tmpfs: convert me... |
4199 4200 |
#ifdef CONFIG_SWAP /* shmem/tmpfs may report page out on swap: account for that too. */ |
139b6a6fb mm: filemap: upda... |
4201 4202 4203 4204 |
if (shmem_mapping(mapping)) { page = find_get_entry(mapping, pgoff); if (radix_tree_exceptional_entry(page)) { swp_entry_t swp = radix_to_swp_entry(page); |
7941d2145 mm: memcontrol: d... |
4205 |
if (do_memsw_account()) |
139b6a6fb mm: filemap: upda... |
4206 4207 4208 4209 4210 4211 4212 |
*entry = swp; page = find_get_page(swap_address_space(swp), swp.val); } } else page = find_get_page(mapping, pgoff); #else page = find_get_page(mapping, pgoff); |
aa3b18955 tmpfs: convert me... |
4213 |
#endif |
87946a722 memcg: move charg... |
4214 4215 |
return page; } |
b1b0deabb mm: memcontrol: l... |
4216 4217 4218 4219 4220 4221 4222 |
/** * mem_cgroup_move_account - move account of the page * @page: the page * @nr_pages: number of regular pages (>1 for huge pages) * @from: mem_cgroup which the page is moved from. * @to: mem_cgroup which the page is moved to. @from != @to. * |
3ac808fdd mm, thp: remove c... |
4223 |
* The caller must make sure the page is not on LRU (isolate_page() is useful.) |
b1b0deabb mm: memcontrol: l... |
4224 4225 4226 4227 4228 |
* * This function doesn't do "charge" to new cgroup and doesn't do "uncharge" * from old cgroup. */ static int mem_cgroup_move_account(struct page *page, |
f627c2f53 memcg: adjust to ... |
4229 |
bool compound, |
b1b0deabb mm: memcontrol: l... |
4230 4231 4232 4233 |
struct mem_cgroup *from, struct mem_cgroup *to) { unsigned long flags; |
f627c2f53 memcg: adjust to ... |
4234 |
unsigned int nr_pages = compound ? hpage_nr_pages(page) : 1; |
b1b0deabb mm: memcontrol: l... |
4235 |
int ret; |
c4843a759 memcg: add per cg... |
4236 |
bool anon; |
b1b0deabb mm: memcontrol: l... |
4237 4238 4239 |
VM_BUG_ON(from == to); VM_BUG_ON_PAGE(PageLRU(page), page); |
f627c2f53 memcg: adjust to ... |
4240 |
VM_BUG_ON(compound && !PageTransHuge(page)); |
b1b0deabb mm: memcontrol: l... |
4241 4242 |
/* |
6a93ca8fd mm: migrate: do n... |
4243 |
* Prevent mem_cgroup_migrate() from looking at |
45637bab3 mm: rename mem_cg... |
4244 |
* page->mem_cgroup of its source page while we change it. |
b1b0deabb mm: memcontrol: l... |
4245 |
*/ |
f627c2f53 memcg: adjust to ... |
4246 |
ret = -EBUSY; |
b1b0deabb mm: memcontrol: l... |
4247 4248 4249 4250 4251 4252 |
if (!trylock_page(page)) goto out; ret = -EINVAL; if (page->mem_cgroup != from) goto out_unlock; |
c4843a759 memcg: add per cg... |
4253 |
anon = PageAnon(page); |
b1b0deabb mm: memcontrol: l... |
4254 |
spin_lock_irqsave(&from->move_lock, flags); |
c4843a759 memcg: add per cg... |
4255 |
if (!anon && page_mapped(page)) { |
b1b0deabb mm: memcontrol: l... |
4256 4257 4258 4259 4260 |
__this_cpu_sub(from->stat->count[MEM_CGROUP_STAT_FILE_MAPPED], nr_pages); __this_cpu_add(to->stat->count[MEM_CGROUP_STAT_FILE_MAPPED], nr_pages); } |
c4843a759 memcg: add per cg... |
4261 4262 4263 4264 4265 4266 4267 4268 4269 4270 4271 4272 4273 4274 4275 |
/* * move_lock grabbed above and caller set from->moving_account, so * mem_cgroup_update_page_stat() will serialize updates to PageDirty. * So mapping should be stable for dirty pages. */ if (!anon && PageDirty(page)) { struct address_space *mapping = page_mapping(page); if (mapping_cap_account_dirty(mapping)) { __this_cpu_sub(from->stat->count[MEM_CGROUP_STAT_DIRTY], nr_pages); __this_cpu_add(to->stat->count[MEM_CGROUP_STAT_DIRTY], nr_pages); } } |
b1b0deabb mm: memcontrol: l... |
4276 4277 4278 4279 4280 4281 4282 4283 4284 4285 4286 4287 4288 4289 4290 4291 4292 4293 4294 4295 |
if (PageWriteback(page)) { __this_cpu_sub(from->stat->count[MEM_CGROUP_STAT_WRITEBACK], nr_pages); __this_cpu_add(to->stat->count[MEM_CGROUP_STAT_WRITEBACK], nr_pages); } /* * It is safe to change page->mem_cgroup here because the page * is referenced, charged, and isolated - we can't race with * uncharging, charging, migration, or LRU putback. */ /* caller should have done css_get */ page->mem_cgroup = to; spin_unlock_irqrestore(&from->move_lock, flags); ret = 0; local_irq_disable(); |
f627c2f53 memcg: adjust to ... |
4296 |
mem_cgroup_charge_statistics(to, page, compound, nr_pages); |
b1b0deabb mm: memcontrol: l... |
4297 |
memcg_check_events(to, page); |
f627c2f53 memcg: adjust to ... |
4298 |
mem_cgroup_charge_statistics(from, page, compound, -nr_pages); |
b1b0deabb mm: memcontrol: l... |
4299 4300 4301 4302 4303 4304 4305 |
memcg_check_events(from, page); local_irq_enable(); out_unlock: unlock_page(page); out: return ret; } |
8d32ff844 memcg: clean up e... |
4306 |
static enum mc_target_type get_mctgt_type(struct vm_area_struct *vma, |
90254a658 memcg: clean up m... |
4307 4308 4309 |
unsigned long addr, pte_t ptent, union mc_target *target) { struct page *page = NULL; |
8d32ff844 memcg: clean up e... |
4310 |
enum mc_target_type ret = MC_TARGET_NONE; |
90254a658 memcg: clean up m... |
4311 4312 4313 4314 4315 4316 |
swp_entry_t ent = { .val = 0 }; if (pte_present(ptent)) page = mc_handle_present_pte(vma, addr, ptent); else if (is_swap_pte(ptent)) page = mc_handle_swap_pte(vma, addr, ptent, &ent); |
0661a3361 mm: remove rest u... |
4317 |
else if (pte_none(ptent)) |
87946a722 memcg: move charg... |
4318 |
page = mc_handle_file_pte(vma, addr, ptent, &ent); |
90254a658 memcg: clean up m... |
4319 4320 |
if (!page && !ent.val) |
8d32ff844 memcg: clean up e... |
4321 |
return ret; |
024914477 memcg: move charg... |
4322 |
if (page) { |
024914477 memcg: move charg... |
4323 |
/* |
0a31bc97c mm: memcontrol: r... |
4324 |
* Do only loose check w/o serialization. |
1306a85ae mm: embed the mem... |
4325 |
* mem_cgroup_move_account() checks the page is valid or |
0a31bc97c mm: memcontrol: r... |
4326 |
* not under LRU exclusion. |
024914477 memcg: move charg... |
4327 |
*/ |
1306a85ae mm: embed the mem... |
4328 |
if (page->mem_cgroup == mc.from) { |
024914477 memcg: move charg... |
4329 4330 4331 4332 4333 4334 4335 |
ret = MC_TARGET_PAGE; if (target) target->page = page; } if (!ret || !target) put_page(page); } |
90254a658 memcg: clean up m... |
4336 4337 |
/* There is a swap entry and a page doesn't exist or isn't charged */ if (ent.val && !ret && |
34c00c319 memcg: convert to... |
4338 |
mem_cgroup_id(mc.from) == lookup_swap_cgroup_id(ent)) { |
7f0f15464 memcg: fix css_id... |
4339 4340 4341 |
ret = MC_TARGET_SWAP; if (target) target->ent = ent; |
4ffef5fef memcg: move charg... |
4342 |
} |
4ffef5fef memcg: move charg... |
4343 4344 |
return ret; } |
12724850e memcg: avoid THP ... |
4345 4346 4347 4348 4349 4350 4351 4352 4353 4354 |
#ifdef CONFIG_TRANSPARENT_HUGEPAGE /* * We don't consider swapping or file mapped pages because THP does not * support them for now. * Caller should make sure that pmd_trans_huge(pmd) is true. */ static enum mc_target_type get_mctgt_type_thp(struct vm_area_struct *vma, unsigned long addr, pmd_t pmd, union mc_target *target) { struct page *page = NULL; |
12724850e memcg: avoid THP ... |
4355 4356 4357 |
enum mc_target_type ret = MC_TARGET_NONE; page = pmd_page(pmd); |
309381fea mm: dump page whe... |
4358 |
VM_BUG_ON_PAGE(!page || !PageHead(page), page); |
1dfab5abc mm: memcontrol: f... |
4359 |
if (!(mc.flags & MOVE_ANON)) |
12724850e memcg: avoid THP ... |
4360 |
return ret; |
1306a85ae mm: embed the mem... |
4361 |
if (page->mem_cgroup == mc.from) { |
12724850e memcg: avoid THP ... |
4362 4363 4364 4365 4366 4367 4368 4369 4370 4371 4372 4373 4374 4375 4376 |
ret = MC_TARGET_PAGE; if (target) { get_page(page); target->page = page; } } return ret; } #else static inline enum mc_target_type get_mctgt_type_thp(struct vm_area_struct *vma, unsigned long addr, pmd_t pmd, union mc_target *target) { return MC_TARGET_NONE; } #endif |
4ffef5fef memcg: move charg... |
4377 4378 4379 4380 |
static int mem_cgroup_count_precharge_pte_range(pmd_t *pmd, unsigned long addr, unsigned long end, struct mm_walk *walk) { |
26bcd64aa memcg: cleanup pr... |
4381 |
struct vm_area_struct *vma = walk->vma; |
4ffef5fef memcg: move charg... |
4382 4383 |
pte_t *pte; spinlock_t *ptl; |
b6ec57f4b thp: change pmd_t... |
4384 4385 |
ptl = pmd_trans_huge_lock(pmd, vma); if (ptl) { |
12724850e memcg: avoid THP ... |
4386 4387 |
if (get_mctgt_type_thp(vma, addr, *pmd, NULL) == MC_TARGET_PAGE) mc.precharge += HPAGE_PMD_NR; |
bf929152e mm, thp: change p... |
4388 |
spin_unlock(ptl); |
1a5a9906d mm: thp: fix pmd_... |
4389 |
return 0; |
12724850e memcg: avoid THP ... |
4390 |
} |
033193275 pagewalk: only sp... |
4391 |
|
45f83cefe mm: thp: fix up p... |
4392 4393 |
if (pmd_trans_unstable(pmd)) return 0; |
4ffef5fef memcg: move charg... |
4394 4395 |
pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl); for (; addr != end; pte++, addr += PAGE_SIZE) |
8d32ff844 memcg: clean up e... |
4396 |
if (get_mctgt_type(vma, addr, *pte, NULL)) |
4ffef5fef memcg: move charg... |
4397 4398 4399 |
mc.precharge++; /* increment precharge temporarily */ pte_unmap_unlock(pte - 1, ptl); cond_resched(); |
7dc74be03 memcg: add interf... |
4400 4401 |
return 0; } |
4ffef5fef memcg: move charg... |
4402 4403 4404 |
static unsigned long mem_cgroup_count_precharge(struct mm_struct *mm) { unsigned long precharge; |
4ffef5fef memcg: move charg... |
4405 |
|
26bcd64aa memcg: cleanup pr... |
4406 4407 4408 4409 |
struct mm_walk mem_cgroup_count_precharge_walk = { .pmd_entry = mem_cgroup_count_precharge_pte_range, .mm = mm, }; |
dfe076b09 memcg: fix deadlo... |
4410 |
down_read(&mm->mmap_sem); |
26bcd64aa memcg: cleanup pr... |
4411 |
walk_page_range(0, ~0UL, &mem_cgroup_count_precharge_walk); |
dfe076b09 memcg: fix deadlo... |
4412 |
up_read(&mm->mmap_sem); |
4ffef5fef memcg: move charg... |
4413 4414 4415 4416 4417 4418 |
precharge = mc.precharge; mc.precharge = 0; return precharge; } |
4ffef5fef memcg: move charg... |
4419 4420 |
static int mem_cgroup_precharge_mc(struct mm_struct *mm) { |
dfe076b09 memcg: fix deadlo... |
4421 4422 4423 4424 4425 |
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... |
4426 |
} |
dfe076b09 memcg: fix deadlo... |
4427 4428 |
/* 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... |
4429 |
{ |
2bd9bb206 memcg: clean up w... |
4430 4431 |
struct mem_cgroup *from = mc.from; struct mem_cgroup *to = mc.to; |
4ffef5fef memcg: move charg... |
4432 |
/* we must uncharge all the leftover precharges from mc.to */ |
854ffa8d1 memcg: improve pe... |
4433 |
if (mc.precharge) { |
00501b531 mm: memcontrol: r... |
4434 |
cancel_charge(mc.to, mc.precharge); |
854ffa8d1 memcg: improve pe... |
4435 4436 4437 4438 4439 4440 4441 |
mc.precharge = 0; } /* * we didn't uncharge from mc.from at mem_cgroup_move_account(), so * we must uncharge here. */ if (mc.moved_charge) { |
00501b531 mm: memcontrol: r... |
4442 |
cancel_charge(mc.from, mc.moved_charge); |
854ffa8d1 memcg: improve pe... |
4443 |
mc.moved_charge = 0; |
4ffef5fef memcg: move charg... |
4444 |
} |
483c30b51 memcg: improve pe... |
4445 4446 |
/* we must fixup refcnts and charges */ if (mc.moved_swap) { |
483c30b51 memcg: improve pe... |
4447 |
/* uncharge swap account from the old cgroup */ |
ce00a9673 mm: memcontrol: r... |
4448 |
if (!mem_cgroup_is_root(mc.from)) |
3e32cb2e0 mm: memcontrol: l... |
4449 |
page_counter_uncharge(&mc.from->memsw, mc.moved_swap); |
483c30b51 memcg: improve pe... |
4450 |
|
05b843012 mm: memcontrol: u... |
4451 |
/* |
3e32cb2e0 mm: memcontrol: l... |
4452 4453 |
* we charged both to->memory and to->memsw, so we * should uncharge to->memory. |
05b843012 mm: memcontrol: u... |
4454 |
*/ |
ce00a9673 mm: memcontrol: r... |
4455 |
if (!mem_cgroup_is_root(mc.to)) |
3e32cb2e0 mm: memcontrol: l... |
4456 |
page_counter_uncharge(&mc.to->memory, mc.moved_swap); |
e8ea14cc6 mm: memcontrol: t... |
4457 |
css_put_many(&mc.from->css, mc.moved_swap); |
3e32cb2e0 mm: memcontrol: l... |
4458 |
|
4050377b5 memcg: use css_ge... |
4459 |
/* we've already done css_get(mc.to) */ |
483c30b51 memcg: improve pe... |
4460 4461 |
mc.moved_swap = 0; } |
dfe076b09 memcg: fix deadlo... |
4462 4463 4464 4465 4466 4467 4468 |
memcg_oom_recover(from); memcg_oom_recover(to); wake_up_all(&mc.waitq); } static void mem_cgroup_clear_mc(void) { |
264a0ae16 memcg: relocate c... |
4469 |
struct mm_struct *mm = mc.mm; |
dfe076b09 memcg: fix deadlo... |
4470 4471 4472 4473 4474 4475 |
/* * 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... |
4476 |
spin_lock(&mc.lock); |
4ffef5fef memcg: move charg... |
4477 4478 |
mc.from = NULL; mc.to = NULL; |
264a0ae16 memcg: relocate c... |
4479 |
mc.mm = NULL; |
2bd9bb206 memcg: clean up w... |
4480 |
spin_unlock(&mc.lock); |
264a0ae16 memcg: relocate c... |
4481 4482 |
mmput(mm); |
4ffef5fef memcg: move charg... |
4483 |
} |
1f7dd3e5a cgroup: fix handl... |
4484 |
static int mem_cgroup_can_attach(struct cgroup_taskset *tset) |
7dc74be03 memcg: add interf... |
4485 |
{ |
1f7dd3e5a cgroup: fix handl... |
4486 |
struct cgroup_subsys_state *css; |
eed67d75b cgroup: Fix unini... |
4487 |
struct mem_cgroup *memcg = NULL; /* unneeded init to make gcc happy */ |
9f2115f93 memcg: restructur... |
4488 |
struct mem_cgroup *from; |
4530eddb5 cgroup, memcg, cp... |
4489 |
struct task_struct *leader, *p; |
9f2115f93 memcg: restructur... |
4490 |
struct mm_struct *mm; |
1dfab5abc mm: memcontrol: f... |
4491 |
unsigned long move_flags; |
9f2115f93 memcg: restructur... |
4492 |
int ret = 0; |
7dc74be03 memcg: add interf... |
4493 |
|
1f7dd3e5a cgroup: fix handl... |
4494 4495 |
/* charge immigration isn't supported on the default hierarchy */ if (cgroup_subsys_on_dfl(memory_cgrp_subsys)) |
9f2115f93 memcg: restructur... |
4496 |
return 0; |
4530eddb5 cgroup, memcg, cp... |
4497 4498 4499 4500 4501 4502 4503 |
/* * Multi-process migrations only happen on the default hierarchy * where charge immigration is not used. Perform charge * immigration if @tset contains a leader and whine if there are * multiple. */ p = NULL; |
1f7dd3e5a cgroup: fix handl... |
4504 |
cgroup_taskset_for_each_leader(leader, css, tset) { |
4530eddb5 cgroup, memcg, cp... |
4505 4506 |
WARN_ON_ONCE(p); p = leader; |
1f7dd3e5a cgroup: fix handl... |
4507 |
memcg = mem_cgroup_from_css(css); |
4530eddb5 cgroup, memcg, cp... |
4508 4509 4510 |
} if (!p) return 0; |
1f7dd3e5a cgroup: fix handl... |
4511 4512 4513 4514 4515 4516 4517 4518 |
/* * We are now commited to this value whatever it is. Changes in this * tunable will only affect upcoming migrations, not the current one. * So we need to save it, and keep it going. */ move_flags = READ_ONCE(memcg->move_charge_at_immigrate); if (!move_flags) return 0; |
9f2115f93 memcg: restructur... |
4519 4520 4521 4522 4523 4524 4525 4526 4527 4528 4529 4530 4531 4532 4533 4534 |
from = mem_cgroup_from_task(p); VM_BUG_ON(from == memcg); mm = get_task_mm(p); if (!mm) return 0; /* We move charges only when we move a owner of the mm */ if (mm->owner == p) { VM_BUG_ON(mc.from); VM_BUG_ON(mc.to); VM_BUG_ON(mc.precharge); VM_BUG_ON(mc.moved_charge); VM_BUG_ON(mc.moved_swap); spin_lock(&mc.lock); |
264a0ae16 memcg: relocate c... |
4535 |
mc.mm = mm; |
9f2115f93 memcg: restructur... |
4536 4537 4538 4539 4540 4541 4542 4543 4544 |
mc.from = from; mc.to = memcg; mc.flags = move_flags; spin_unlock(&mc.lock); /* We set mc.moving_task later */ ret = mem_cgroup_precharge_mc(mm); if (ret) mem_cgroup_clear_mc(); |
264a0ae16 memcg: relocate c... |
4545 4546 |
} else { mmput(mm); |
7dc74be03 memcg: add interf... |
4547 4548 4549 |
} return ret; } |
1f7dd3e5a cgroup: fix handl... |
4550 |
static void mem_cgroup_cancel_attach(struct cgroup_taskset *tset) |
7dc74be03 memcg: add interf... |
4551 |
{ |
4e2f245d3 mm: memcontrol: d... |
4552 4553 |
if (mc.to) mem_cgroup_clear_mc(); |
7dc74be03 memcg: add interf... |
4554 |
} |
4ffef5fef memcg: move charg... |
4555 4556 4557 |
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... |
4558 |
{ |
4ffef5fef memcg: move charg... |
4559 |
int ret = 0; |
26bcd64aa memcg: cleanup pr... |
4560 |
struct vm_area_struct *vma = walk->vma; |
4ffef5fef memcg: move charg... |
4561 4562 |
pte_t *pte; spinlock_t *ptl; |
12724850e memcg: avoid THP ... |
4563 4564 4565 |
enum mc_target_type target_type; union mc_target target; struct page *page; |
4ffef5fef memcg: move charg... |
4566 |
|
b6ec57f4b thp: change pmd_t... |
4567 4568 |
ptl = pmd_trans_huge_lock(pmd, vma); if (ptl) { |
62ade86ab memcg,thp: fix re... |
4569 |
if (mc.precharge < HPAGE_PMD_NR) { |
bf929152e mm, thp: change p... |
4570 |
spin_unlock(ptl); |
12724850e memcg: avoid THP ... |
4571 4572 4573 4574 4575 4576 |
return 0; } target_type = get_mctgt_type_thp(vma, addr, *pmd, &target); if (target_type == MC_TARGET_PAGE) { page = target.page; if (!isolate_lru_page(page)) { |
f627c2f53 memcg: adjust to ... |
4577 |
if (!mem_cgroup_move_account(page, true, |
1306a85ae mm: embed the mem... |
4578 |
mc.from, mc.to)) { |
12724850e memcg: avoid THP ... |
4579 4580 4581 4582 4583 4584 4585 |
mc.precharge -= HPAGE_PMD_NR; mc.moved_charge += HPAGE_PMD_NR; } putback_lru_page(page); } put_page(page); } |
bf929152e mm, thp: change p... |
4586 |
spin_unlock(ptl); |
1a5a9906d mm: thp: fix pmd_... |
4587 |
return 0; |
12724850e memcg: avoid THP ... |
4588 |
} |
45f83cefe mm: thp: fix up p... |
4589 4590 |
if (pmd_trans_unstable(pmd)) return 0; |
4ffef5fef memcg: move charg... |
4591 4592 4593 4594 |
retry: pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl); for (; addr != end; addr += PAGE_SIZE) { pte_t ptent = *(pte++); |
024914477 memcg: move charg... |
4595 |
swp_entry_t ent; |
4ffef5fef memcg: move charg... |
4596 4597 4598 |
if (!mc.precharge) break; |
8d32ff844 memcg: clean up e... |
4599 |
switch (get_mctgt_type(vma, addr, ptent, &target)) { |
4ffef5fef memcg: move charg... |
4600 4601 |
case MC_TARGET_PAGE: page = target.page; |
53f9263ba mm: rework mapcou... |
4602 4603 4604 4605 4606 4607 4608 4609 |
/* * We can have a part of the split pmd here. Moving it * can be done but it would be too convoluted so simply * ignore such a partial THP and keep it in original * memcg. There should be somebody mapping the head. */ if (PageTransCompound(page)) goto put; |
4ffef5fef memcg: move charg... |
4610 4611 |
if (isolate_lru_page(page)) goto put; |
f627c2f53 memcg: adjust to ... |
4612 4613 |
if (!mem_cgroup_move_account(page, false, mc.from, mc.to)) { |
4ffef5fef memcg: move charg... |
4614 |
mc.precharge--; |
854ffa8d1 memcg: improve pe... |
4615 4616 |
/* we uncharge from mc.from later. */ mc.moved_charge++; |
4ffef5fef memcg: move charg... |
4617 4618 |
} putback_lru_page(page); |
8d32ff844 memcg: clean up e... |
4619 |
put: /* get_mctgt_type() gets the page */ |
4ffef5fef memcg: move charg... |
4620 4621 |
put_page(page); break; |
024914477 memcg: move charg... |
4622 4623 |
case MC_TARGET_SWAP: ent = target.ent; |
e91cbb425 memcg swap: mem_c... |
4624 |
if (!mem_cgroup_move_swap_account(ent, mc.from, mc.to)) { |
024914477 memcg: move charg... |
4625 |
mc.precharge--; |
483c30b51 memcg: improve pe... |
4626 4627 4628 |
/* we fixup refcnts and charges later. */ mc.moved_swap++; } |
024914477 memcg: move charg... |
4629 |
break; |
4ffef5fef memcg: move charg... |
4630 4631 4632 4633 4634 4635 4636 4637 4638 4639 4640 4641 4642 4643 |
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... |
4644 |
ret = mem_cgroup_do_precharge(1); |
4ffef5fef memcg: move charg... |
4645 4646 4647 4648 4649 4650 |
if (!ret) goto retry; } return ret; } |
264a0ae16 memcg: relocate c... |
4651 |
static void mem_cgroup_move_charge(void) |
4ffef5fef memcg: move charg... |
4652 |
{ |
26bcd64aa memcg: cleanup pr... |
4653 4654 |
struct mm_walk mem_cgroup_move_charge_walk = { .pmd_entry = mem_cgroup_move_charge_pte_range, |
264a0ae16 memcg: relocate c... |
4655 |
.mm = mc.mm, |
26bcd64aa memcg: cleanup pr... |
4656 |
}; |
4ffef5fef memcg: move charg... |
4657 4658 |
lru_add_drain_all(); |
312722cbb mm: memcontrol: s... |
4659 |
/* |
81f8c3a46 mm: memcontrol: g... |
4660 4661 4662 |
* Signal lock_page_memcg() to take the memcg's move_lock * while we're moving its pages to another memcg. Then wait * for already started RCU-only updates to finish. |
312722cbb mm: memcontrol: s... |
4663 4664 4665 |
*/ atomic_inc(&mc.from->moving_account); synchronize_rcu(); |
dfe076b09 memcg: fix deadlo... |
4666 |
retry: |
264a0ae16 memcg: relocate c... |
4667 |
if (unlikely(!down_read_trylock(&mc.mm->mmap_sem))) { |
dfe076b09 memcg: fix deadlo... |
4668 4669 4670 4671 4672 4673 4674 4675 4676 4677 4678 |
/* * Someone who are holding the mmap_sem might be waiting in * waitq. So we cancel all extra charges, wake up all waiters, * and retry. Because we cancel precharges, we might not be able * to move enough charges, but moving charge is a best-effort * feature anyway, so it wouldn't be a big problem. */ __mem_cgroup_clear_mc(); cond_resched(); goto retry; } |
26bcd64aa memcg: cleanup pr... |
4679 4680 4681 4682 4683 |
/* * When we have consumed all precharges and failed in doing * additional charge, the page walk just aborts. */ walk_page_range(0, ~0UL, &mem_cgroup_move_charge_walk); |
264a0ae16 memcg: relocate c... |
4684 |
up_read(&mc.mm->mmap_sem); |
312722cbb mm: memcontrol: s... |
4685 |
atomic_dec(&mc.from->moving_account); |
7dc74be03 memcg: add interf... |
4686 |
} |
264a0ae16 memcg: relocate c... |
4687 |
static void mem_cgroup_move_task(void) |
67e465a77 Memory controller... |
4688 |
{ |
264a0ae16 memcg: relocate c... |
4689 4690 |
if (mc.to) { mem_cgroup_move_charge(); |
a433658c3 vmscan,memcg: mem... |
4691 |
mem_cgroup_clear_mc(); |
264a0ae16 memcg: relocate c... |
4692 |
} |
67e465a77 Memory controller... |
4693 |
} |
5cfb80a73 memcg: disable mo... |
4694 |
#else /* !CONFIG_MMU */ |
1f7dd3e5a cgroup: fix handl... |
4695 |
static int mem_cgroup_can_attach(struct cgroup_taskset *tset) |
5cfb80a73 memcg: disable mo... |
4696 4697 4698 |
{ return 0; } |
1f7dd3e5a cgroup: fix handl... |
4699 |
static void mem_cgroup_cancel_attach(struct cgroup_taskset *tset) |
5cfb80a73 memcg: disable mo... |
4700 4701 |
{ } |
264a0ae16 memcg: relocate c... |
4702 |
static void mem_cgroup_move_task(void) |
5cfb80a73 memcg: disable mo... |
4703 4704 4705 |
{ } #endif |
67e465a77 Memory controller... |
4706 |
|
f00baae7a memcg: force use_... |
4707 4708 |
/* * Cgroup retains root cgroups across [un]mount cycles making it necessary |
aa6ec29be cgroup: remove sa... |
4709 4710 |
* to verify whether we're attached to the default hierarchy on each mount * attempt. |
f00baae7a memcg: force use_... |
4711 |
*/ |
eb95419b0 cgroup: pass arou... |
4712 |
static void mem_cgroup_bind(struct cgroup_subsys_state *root_css) |
f00baae7a memcg: force use_... |
4713 4714 |
{ /* |
aa6ec29be cgroup: remove sa... |
4715 |
* use_hierarchy is forced on the default hierarchy. cgroup core |
f00baae7a memcg: force use_... |
4716 4717 4718 |
* guarantees that @root doesn't have any children, so turning it * on for the root memcg is enough. */ |
9e10a130d cgroup: replace c... |
4719 |
if (cgroup_subsys_on_dfl(memory_cgrp_subsys)) |
7feee590b memcg: disable hi... |
4720 4721 4722 |
root_mem_cgroup->use_hierarchy = true; else root_mem_cgroup->use_hierarchy = false; |
f00baae7a memcg: force use_... |
4723 |
} |
241994ed8 mm: memcontrol: d... |
4724 4725 4726 |
static u64 memory_current_read(struct cgroup_subsys_state *css, struct cftype *cft) { |
f5fc3c5d8 mm: memcontrol: e... |
4727 4728 4729 |
struct mem_cgroup *memcg = mem_cgroup_from_css(css); return (u64)page_counter_read(&memcg->memory) * PAGE_SIZE; |
241994ed8 mm: memcontrol: d... |
4730 4731 4732 4733 4734 |
} static int memory_low_show(struct seq_file *m, void *v) { struct mem_cgroup *memcg = mem_cgroup_from_css(seq_css(m)); |
4db0c3c29 mm: remove rest o... |
4735 |
unsigned long low = READ_ONCE(memcg->low); |
241994ed8 mm: memcontrol: d... |
4736 4737 |
if (low == PAGE_COUNTER_MAX) |
d2973697b mm: memcontrol: u... |
4738 4739 |
seq_puts(m, "max "); |
241994ed8 mm: memcontrol: d... |
4740 4741 4742 4743 4744 4745 4746 4747 4748 4749 4750 4751 4752 4753 4754 |
else seq_printf(m, "%llu ", (u64)low * PAGE_SIZE); return 0; } static ssize_t memory_low_write(struct kernfs_open_file *of, char *buf, size_t nbytes, loff_t off) { struct mem_cgroup *memcg = mem_cgroup_from_css(of_css(of)); unsigned long low; int err; buf = strstrip(buf); |
d2973697b mm: memcontrol: u... |
4755 |
err = page_counter_memparse(buf, "max", &low); |
241994ed8 mm: memcontrol: d... |
4756 4757 4758 4759 4760 4761 4762 4763 4764 4765 4766 |
if (err) return err; memcg->low = low; return nbytes; } static int memory_high_show(struct seq_file *m, void *v) { struct mem_cgroup *memcg = mem_cgroup_from_css(seq_css(m)); |
4db0c3c29 mm: remove rest o... |
4767 |
unsigned long high = READ_ONCE(memcg->high); |
241994ed8 mm: memcontrol: d... |
4768 4769 |
if (high == PAGE_COUNTER_MAX) |
d2973697b mm: memcontrol: u... |
4770 4771 |
seq_puts(m, "max "); |
241994ed8 mm: memcontrol: d... |
4772 4773 4774 4775 4776 4777 4778 4779 4780 4781 4782 |
else seq_printf(m, "%llu ", (u64)high * PAGE_SIZE); return 0; } static ssize_t memory_high_write(struct kernfs_open_file *of, char *buf, size_t nbytes, loff_t off) { struct mem_cgroup *memcg = mem_cgroup_from_css(of_css(of)); |
588083bb3 mm: memcontrol: r... |
4783 |
unsigned long nr_pages; |
241994ed8 mm: memcontrol: d... |
4784 4785 4786 4787 |
unsigned long high; int err; buf = strstrip(buf); |
d2973697b mm: memcontrol: u... |
4788 |
err = page_counter_memparse(buf, "max", &high); |
241994ed8 mm: memcontrol: d... |
4789 4790 4791 4792 |
if (err) return err; memcg->high = high; |
588083bb3 mm: memcontrol: r... |
4793 4794 4795 4796 |
nr_pages = page_counter_read(&memcg->memory); if (nr_pages > high) try_to_free_mem_cgroup_pages(memcg, nr_pages - high, GFP_KERNEL, true); |
2529bb3aa writeback: reset ... |
4797 |
memcg_wb_domain_size_changed(memcg); |
241994ed8 mm: memcontrol: d... |
4798 4799 4800 4801 4802 4803 |
return nbytes; } static int memory_max_show(struct seq_file *m, void *v) { struct mem_cgroup *memcg = mem_cgroup_from_css(seq_css(m)); |
4db0c3c29 mm: remove rest o... |
4804 |
unsigned long max = READ_ONCE(memcg->memory.limit); |
241994ed8 mm: memcontrol: d... |
4805 4806 |
if (max == PAGE_COUNTER_MAX) |
d2973697b mm: memcontrol: u... |
4807 4808 |
seq_puts(m, "max "); |
241994ed8 mm: memcontrol: d... |
4809 4810 4811 4812 4813 4814 4815 4816 4817 4818 4819 |
else seq_printf(m, "%llu ", (u64)max * PAGE_SIZE); return 0; } static ssize_t memory_max_write(struct kernfs_open_file *of, char *buf, size_t nbytes, loff_t off) { struct mem_cgroup *memcg = mem_cgroup_from_css(of_css(of)); |
b6e6edcfa mm: memcontrol: r... |
4820 4821 |
unsigned int nr_reclaims = MEM_CGROUP_RECLAIM_RETRIES; bool drained = false; |
241994ed8 mm: memcontrol: d... |
4822 4823 4824 4825 |
unsigned long max; int err; buf = strstrip(buf); |
d2973697b mm: memcontrol: u... |
4826 |
err = page_counter_memparse(buf, "max", &max); |
241994ed8 mm: memcontrol: d... |
4827 4828 |
if (err) return err; |
b6e6edcfa mm: memcontrol: r... |
4829 4830 4831 4832 4833 4834 4835 4836 4837 4838 4839 4840 4841 4842 4843 4844 4845 4846 4847 4848 4849 4850 4851 4852 4853 4854 4855 4856 4857 4858 |
xchg(&memcg->memory.limit, max); for (;;) { unsigned long nr_pages = page_counter_read(&memcg->memory); if (nr_pages <= max) break; if (signal_pending(current)) { err = -EINTR; break; } if (!drained) { drain_all_stock(memcg); drained = true; continue; } if (nr_reclaims) { if (!try_to_free_mem_cgroup_pages(memcg, nr_pages - max, GFP_KERNEL, true)) nr_reclaims--; continue; } mem_cgroup_events(memcg, MEMCG_OOM, 1); if (!mem_cgroup_out_of_memory(memcg, GFP_KERNEL, 0)) break; } |
241994ed8 mm: memcontrol: d... |
4859 |
|
2529bb3aa writeback: reset ... |
4860 |
memcg_wb_domain_size_changed(memcg); |
241994ed8 mm: memcontrol: d... |
4861 4862 4863 4864 4865 4866 4867 4868 4869 4870 4871 4872 4873 4874 4875 4876 4877 4878 |
return nbytes; } static int memory_events_show(struct seq_file *m, void *v) { struct mem_cgroup *memcg = mem_cgroup_from_css(seq_css(m)); seq_printf(m, "low %lu ", mem_cgroup_read_events(memcg, MEMCG_LOW)); seq_printf(m, "high %lu ", mem_cgroup_read_events(memcg, MEMCG_HIGH)); seq_printf(m, "max %lu ", mem_cgroup_read_events(memcg, MEMCG_MAX)); seq_printf(m, "oom %lu ", mem_cgroup_read_events(memcg, MEMCG_OOM)); return 0; } |
587d9f726 mm: memcontrol: b... |
4879 4880 4881 |
static int memory_stat_show(struct seq_file *m, void *v) { struct mem_cgroup *memcg = mem_cgroup_from_css(seq_css(m)); |
72b54e731 mm: memcontrol: m... |
4882 4883 |
unsigned long stat[MEMCG_NR_STAT]; unsigned long events[MEMCG_NR_EVENTS]; |
587d9f726 mm: memcontrol: b... |
4884 4885 4886 4887 4888 4889 4890 4891 4892 4893 4894 4895 |
int i; /* * Provide statistics on the state of the memory subsystem as * well as cumulative event counters that show past behavior. * * This list is ordered following a combination of these gradients: * 1) generic big picture -> specifics and details * 2) reflecting userspace activity -> reflecting kernel heuristics * * Current memory state: */ |
72b54e731 mm: memcontrol: m... |
4896 4897 |
tree_stat(memcg, stat); tree_events(memcg, events); |
587d9f726 mm: memcontrol: b... |
4898 4899 |
seq_printf(m, "anon %llu ", |
72b54e731 mm: memcontrol: m... |
4900 |
(u64)stat[MEM_CGROUP_STAT_RSS] * PAGE_SIZE); |
587d9f726 mm: memcontrol: b... |
4901 4902 |
seq_printf(m, "file %llu ", |
72b54e731 mm: memcontrol: m... |
4903 |
(u64)stat[MEM_CGROUP_STAT_CACHE] * PAGE_SIZE); |
12580e4b5 mm: memcontrol: r... |
4904 4905 4906 |
seq_printf(m, "kernel_stack %llu ", (u64)stat[MEMCG_KERNEL_STACK] * PAGE_SIZE); |
27ee57c93 mm: memcontrol: r... |
4907 4908 4909 4910 |
seq_printf(m, "slab %llu ", (u64)(stat[MEMCG_SLAB_RECLAIMABLE] + stat[MEMCG_SLAB_UNRECLAIMABLE]) * PAGE_SIZE); |
b2807f07f mm: memcontrol: a... |
4911 4912 |
seq_printf(m, "sock %llu ", |
72b54e731 mm: memcontrol: m... |
4913 |
(u64)stat[MEMCG_SOCK] * PAGE_SIZE); |
587d9f726 mm: memcontrol: b... |
4914 4915 4916 |
seq_printf(m, "file_mapped %llu ", |
72b54e731 mm: memcontrol: m... |
4917 |
(u64)stat[MEM_CGROUP_STAT_FILE_MAPPED] * PAGE_SIZE); |
587d9f726 mm: memcontrol: b... |
4918 4919 |
seq_printf(m, "file_dirty %llu ", |
72b54e731 mm: memcontrol: m... |
4920 |
(u64)stat[MEM_CGROUP_STAT_DIRTY] * PAGE_SIZE); |
587d9f726 mm: memcontrol: b... |
4921 4922 |
seq_printf(m, "file_writeback %llu ", |
72b54e731 mm: memcontrol: m... |
4923 |
(u64)stat[MEM_CGROUP_STAT_WRITEBACK] * PAGE_SIZE); |
587d9f726 mm: memcontrol: b... |
4924 4925 4926 4927 4928 4929 4930 4931 4932 4933 4934 |
for (i = 0; i < NR_LRU_LISTS; i++) { struct mem_cgroup *mi; unsigned long val = 0; for_each_mem_cgroup_tree(mi, memcg) val += mem_cgroup_nr_lru_pages(mi, BIT(i)); seq_printf(m, "%s %llu ", mem_cgroup_lru_names[i], (u64)val * PAGE_SIZE); } |
27ee57c93 mm: memcontrol: r... |
4935 4936 4937 4938 4939 4940 |
seq_printf(m, "slab_reclaimable %llu ", (u64)stat[MEMCG_SLAB_RECLAIMABLE] * PAGE_SIZE); seq_printf(m, "slab_unreclaimable %llu ", (u64)stat[MEMCG_SLAB_UNRECLAIMABLE] * PAGE_SIZE); |
587d9f726 mm: memcontrol: b... |
4941 4942 4943 4944 |
/* Accumulated memory events */ seq_printf(m, "pgfault %lu ", |
72b54e731 mm: memcontrol: m... |
4945 |
events[MEM_CGROUP_EVENTS_PGFAULT]); |
587d9f726 mm: memcontrol: b... |
4946 4947 |
seq_printf(m, "pgmajfault %lu ", |
72b54e731 mm: memcontrol: m... |
4948 |
events[MEM_CGROUP_EVENTS_PGMAJFAULT]); |
587d9f726 mm: memcontrol: b... |
4949 4950 4951 |
return 0; } |
241994ed8 mm: memcontrol: d... |
4952 4953 4954 |
static struct cftype memory_files[] = { { .name = "current", |
f5fc3c5d8 mm: memcontrol: e... |
4955 |
.flags = CFTYPE_NOT_ON_ROOT, |
241994ed8 mm: memcontrol: d... |
4956 4957 4958 4959 4960 4961 4962 4963 4964 4965 4966 4967 4968 4969 4970 4971 4972 4973 4974 4975 4976 4977 4978 |
.read_u64 = memory_current_read, }, { .name = "low", .flags = CFTYPE_NOT_ON_ROOT, .seq_show = memory_low_show, .write = memory_low_write, }, { .name = "high", .flags = CFTYPE_NOT_ON_ROOT, .seq_show = memory_high_show, .write = memory_high_write, }, { .name = "max", .flags = CFTYPE_NOT_ON_ROOT, .seq_show = memory_max_show, .write = memory_max_write, }, { .name = "events", .flags = CFTYPE_NOT_ON_ROOT, |
472912a2b memcg: generate f... |
4979 |
.file_offset = offsetof(struct mem_cgroup, events_file), |
241994ed8 mm: memcontrol: d... |
4980 4981 |
.seq_show = memory_events_show, }, |
587d9f726 mm: memcontrol: b... |
4982 4983 4984 4985 4986 |
{ .name = "stat", .flags = CFTYPE_NOT_ON_ROOT, .seq_show = memory_stat_show, }, |
241994ed8 mm: memcontrol: d... |
4987 4988 |
{ } /* terminate */ }; |
073219e99 cgroup: clean up ... |
4989 |
struct cgroup_subsys memory_cgrp_subsys = { |
92fb97487 cgroup: rename ->... |
4990 |
.css_alloc = mem_cgroup_css_alloc, |
d142e3e66 memcg: split part... |
4991 |
.css_online = mem_cgroup_css_online, |
92fb97487 cgroup: rename ->... |
4992 |
.css_offline = mem_cgroup_css_offline, |
6df38689e mm: memcontrol: f... |
4993 |
.css_released = mem_cgroup_css_released, |
92fb97487 cgroup: rename ->... |
4994 |
.css_free = mem_cgroup_css_free, |
1ced953b1 blkcg, memcg: mak... |
4995 |
.css_reset = mem_cgroup_css_reset, |
7dc74be03 memcg: add interf... |
4996 4997 |
.can_attach = mem_cgroup_can_attach, .cancel_attach = mem_cgroup_cancel_attach, |
264a0ae16 memcg: relocate c... |
4998 |
.post_attach = mem_cgroup_move_task, |
f00baae7a memcg: force use_... |
4999 |
.bind = mem_cgroup_bind, |
241994ed8 mm: memcontrol: d... |
5000 5001 |
.dfl_cftypes = memory_files, .legacy_cftypes = mem_cgroup_legacy_files, |
6d12e2d8d per-zone and recl... |
5002 |
.early_init = 0, |
8cdea7c05 Memory controller... |
5003 |
}; |
c077719be memcg: mem+swap c... |
5004 |
|
241994ed8 mm: memcontrol: d... |
5005 |
/** |
241994ed8 mm: memcontrol: d... |
5006 5007 5008 5009 5010 5011 5012 5013 5014 5015 5016 5017 5018 5019 5020 5021 5022 5023 5024 5025 |
* mem_cgroup_low - check if memory consumption is below the normal range * @root: the highest ancestor to consider * @memcg: the memory cgroup to check * * Returns %true if memory consumption of @memcg, and that of all * configurable ancestors up to @root, is below the normal range. */ bool mem_cgroup_low(struct mem_cgroup *root, struct mem_cgroup *memcg) { if (mem_cgroup_disabled()) return false; /* * The toplevel group doesn't have a configurable range, so * it's never low when looked at directly, and it is not * considered an ancestor when assessing the hierarchy. */ if (memcg == root_mem_cgroup) return false; |
4e54dede3 memcg: fix low li... |
5026 |
if (page_counter_read(&memcg->memory) >= memcg->low) |
241994ed8 mm: memcontrol: d... |
5027 5028 5029 5030 5031 5032 5033 |
return false; while (memcg != root) { memcg = parent_mem_cgroup(memcg); if (memcg == root_mem_cgroup) break; |
4e54dede3 memcg: fix low li... |
5034 |
if (page_counter_read(&memcg->memory) >= memcg->low) |
241994ed8 mm: memcontrol: d... |
5035 5036 5037 5038 |
return false; } return true; } |
00501b531 mm: memcontrol: r... |
5039 5040 5041 5042 5043 5044 5045 5046 5047 5048 5049 5050 5051 5052 5053 5054 5055 5056 |
/** * mem_cgroup_try_charge - try charging a page * @page: page to charge * @mm: mm context of the victim * @gfp_mask: reclaim mode * @memcgp: charged memcg return * * Try to charge @page to the memcg that @mm belongs to, reclaiming * pages according to @gfp_mask if necessary. * * Returns 0 on success, with *@memcgp pointing to the charged memcg. * Otherwise, an error code is returned. * * After page->mapping has been set up, the caller must finalize the * charge with mem_cgroup_commit_charge(). Or abort the transaction * with mem_cgroup_cancel_charge() in case page instantiation fails. */ int mem_cgroup_try_charge(struct page *page, struct mm_struct *mm, |
f627c2f53 memcg: adjust to ... |
5057 5058 |
gfp_t gfp_mask, struct mem_cgroup **memcgp, bool compound) |
00501b531 mm: memcontrol: r... |
5059 5060 |
{ struct mem_cgroup *memcg = NULL; |
f627c2f53 memcg: adjust to ... |
5061 |
unsigned int nr_pages = compound ? hpage_nr_pages(page) : 1; |
00501b531 mm: memcontrol: r... |
5062 5063 5064 5065 5066 5067 |
int ret = 0; if (mem_cgroup_disabled()) goto out; if (PageSwapCache(page)) { |
00501b531 mm: memcontrol: r... |
5068 5069 5070 5071 5072 5073 5074 |
/* * Every swap fault against a single page tries to charge the * page, bail as early as possible. shmem_unuse() encounters * already charged pages, too. The USED bit is protected by * the page lock, which serializes swap cache removal, which * in turn serializes uncharging. */ |
e993d905c memcg: zap try_ge... |
5075 |
VM_BUG_ON_PAGE(!PageLocked(page), page); |
1306a85ae mm: embed the mem... |
5076 |
if (page->mem_cgroup) |
00501b531 mm: memcontrol: r... |
5077 |
goto out; |
e993d905c memcg: zap try_ge... |
5078 |
|
37e843511 mm: memcontrol: c... |
5079 |
if (do_swap_account) { |
e993d905c memcg: zap try_ge... |
5080 5081 5082 5083 5084 5085 5086 5087 5088 |
swp_entry_t ent = { .val = page_private(page), }; unsigned short id = lookup_swap_cgroup_id(ent); rcu_read_lock(); memcg = mem_cgroup_from_id(id); if (memcg && !css_tryget_online(&memcg->css)) memcg = NULL; rcu_read_unlock(); } |
00501b531 mm: memcontrol: r... |
5089 |
} |
00501b531 mm: memcontrol: r... |
5090 5091 5092 5093 5094 5095 |
if (!memcg) memcg = get_mem_cgroup_from_mm(mm); ret = try_charge(memcg, gfp_mask, nr_pages); css_put(&memcg->css); |
00501b531 mm: memcontrol: r... |
5096 5097 5098 5099 5100 5101 5102 5103 5104 5105 5106 5107 5108 5109 5110 5111 5112 5113 5114 5115 5116 5117 |
out: *memcgp = memcg; return ret; } /** * mem_cgroup_commit_charge - commit a page charge * @page: page to charge * @memcg: memcg to charge the page to * @lrucare: page might be on LRU already * * Finalize a charge transaction started by mem_cgroup_try_charge(), * after page->mapping has been set up. This must happen atomically * as part of the page instantiation, i.e. under the page table lock * for anonymous pages, under the page lock for page and swap cache. * * In addition, the page must not be on the LRU during the commit, to * prevent racing with task migration. If it might be, use @lrucare. * * Use mem_cgroup_cancel_charge() to cancel the transaction instead. */ void mem_cgroup_commit_charge(struct page *page, struct mem_cgroup *memcg, |
f627c2f53 memcg: adjust to ... |
5118 |
bool lrucare, bool compound) |
00501b531 mm: memcontrol: r... |
5119 |
{ |
f627c2f53 memcg: adjust to ... |
5120 |
unsigned int nr_pages = compound ? hpage_nr_pages(page) : 1; |
00501b531 mm: memcontrol: r... |
5121 5122 5123 5124 5125 5126 5127 5128 5129 5130 5131 5132 5133 |
VM_BUG_ON_PAGE(!page->mapping, page); VM_BUG_ON_PAGE(PageLRU(page) && !lrucare, page); if (mem_cgroup_disabled()) return; /* * Swap faults will attempt to charge the same page multiple * times. But reuse_swap_page() might have removed the page * from swapcache already, so we can't check PageSwapCache(). */ if (!memcg) return; |
6abb5a867 mm: memcontrol: a... |
5134 |
commit_charge(page, memcg, lrucare); |
6abb5a867 mm: memcontrol: a... |
5135 |
local_irq_disable(); |
f627c2f53 memcg: adjust to ... |
5136 |
mem_cgroup_charge_statistics(memcg, page, compound, nr_pages); |
6abb5a867 mm: memcontrol: a... |
5137 5138 |
memcg_check_events(memcg, page); local_irq_enable(); |
00501b531 mm: memcontrol: r... |
5139 |
|
7941d2145 mm: memcontrol: d... |
5140 |
if (do_memsw_account() && PageSwapCache(page)) { |
00501b531 mm: memcontrol: r... |
5141 5142 5143 5144 5145 5146 5147 5148 5149 5150 5151 5152 5153 5154 5155 5156 5157 |
swp_entry_t entry = { .val = page_private(page) }; /* * The swap entry might not get freed for a long time, * let's not wait for it. The page already received a * memory+swap charge, drop the swap entry duplicate. */ mem_cgroup_uncharge_swap(entry); } } /** * mem_cgroup_cancel_charge - cancel a page charge * @page: page to charge * @memcg: memcg to charge the page to * * Cancel a charge transaction started by mem_cgroup_try_charge(). */ |
f627c2f53 memcg: adjust to ... |
5158 5159 |
void mem_cgroup_cancel_charge(struct page *page, struct mem_cgroup *memcg, bool compound) |
00501b531 mm: memcontrol: r... |
5160 |
{ |
f627c2f53 memcg: adjust to ... |
5161 |
unsigned int nr_pages = compound ? hpage_nr_pages(page) : 1; |
00501b531 mm: memcontrol: r... |
5162 5163 5164 5165 5166 5167 5168 5169 5170 5171 |
if (mem_cgroup_disabled()) return; /* * Swap faults will attempt to charge the same page multiple * times. But reuse_swap_page() might have removed the page * from swapcache already, so we can't check PageSwapCache(). */ if (!memcg) return; |
00501b531 mm: memcontrol: r... |
5172 5173 |
cancel_charge(memcg, nr_pages); } |
747db954c mm: memcontrol: u... |
5174 |
static void uncharge_batch(struct mem_cgroup *memcg, unsigned long pgpgout, |
747db954c mm: memcontrol: u... |
5175 5176 5177 |
unsigned long nr_anon, unsigned long nr_file, unsigned long nr_huge, struct page *dummy_page) { |
18eca2e63 mm: memcontrol: r... |
5178 |
unsigned long nr_pages = nr_anon + nr_file; |
747db954c mm: memcontrol: u... |
5179 |
unsigned long flags; |
ce00a9673 mm: memcontrol: r... |
5180 |
if (!mem_cgroup_is_root(memcg)) { |
18eca2e63 mm: memcontrol: r... |
5181 |
page_counter_uncharge(&memcg->memory, nr_pages); |
7941d2145 mm: memcontrol: d... |
5182 |
if (do_memsw_account()) |
18eca2e63 mm: memcontrol: r... |
5183 |
page_counter_uncharge(&memcg->memsw, nr_pages); |
ce00a9673 mm: memcontrol: r... |
5184 5185 |
memcg_oom_recover(memcg); } |
747db954c mm: memcontrol: u... |
5186 5187 5188 5189 5190 5191 |
local_irq_save(flags); __this_cpu_sub(memcg->stat->count[MEM_CGROUP_STAT_RSS], nr_anon); __this_cpu_sub(memcg->stat->count[MEM_CGROUP_STAT_CACHE], nr_file); __this_cpu_sub(memcg->stat->count[MEM_CGROUP_STAT_RSS_HUGE], nr_huge); __this_cpu_add(memcg->stat->events[MEM_CGROUP_EVENTS_PGPGOUT], pgpgout); |
18eca2e63 mm: memcontrol: r... |
5192 |
__this_cpu_add(memcg->stat->nr_page_events, nr_pages); |
747db954c mm: memcontrol: u... |
5193 5194 |
memcg_check_events(memcg, dummy_page); local_irq_restore(flags); |
e8ea14cc6 mm: memcontrol: t... |
5195 5196 |
if (!mem_cgroup_is_root(memcg)) |
18eca2e63 mm: memcontrol: r... |
5197 |
css_put_many(&memcg->css, nr_pages); |
747db954c mm: memcontrol: u... |
5198 5199 5200 5201 5202 |
} static void uncharge_list(struct list_head *page_list) { struct mem_cgroup *memcg = NULL; |
747db954c mm: memcontrol: u... |
5203 5204 5205 5206 |
unsigned long nr_anon = 0; unsigned long nr_file = 0; unsigned long nr_huge = 0; unsigned long pgpgout = 0; |
747db954c mm: memcontrol: u... |
5207 5208 |
struct list_head *next; struct page *page; |
8b5926560 mm: memcontrol: c... |
5209 5210 5211 5212 |
/* * Note that the list can be a single page->lru; hence the * do-while loop instead of a simple list_for_each_entry(). */ |
747db954c mm: memcontrol: u... |
5213 5214 5215 |
next = page_list->next; do { unsigned int nr_pages = 1; |
747db954c mm: memcontrol: u... |
5216 5217 5218 5219 5220 5221 |
page = list_entry(next, struct page, lru); next = page->lru.next; VM_BUG_ON_PAGE(PageLRU(page), page); VM_BUG_ON_PAGE(page_count(page), page); |
1306a85ae mm: embed the mem... |
5222 |
if (!page->mem_cgroup) |
747db954c mm: memcontrol: u... |
5223 5224 5225 5226 |
continue; /* * Nobody should be changing or seriously looking at |
1306a85ae mm: embed the mem... |
5227 |
* page->mem_cgroup at this point, we have fully |
298333157 mm: memcontrol: r... |
5228 |
* exclusive access to the page. |
747db954c mm: memcontrol: u... |
5229 |
*/ |
1306a85ae mm: embed the mem... |
5230 |
if (memcg != page->mem_cgroup) { |
747db954c mm: memcontrol: u... |
5231 |
if (memcg) { |
18eca2e63 mm: memcontrol: r... |
5232 5233 5234 |
uncharge_batch(memcg, pgpgout, nr_anon, nr_file, nr_huge, page); pgpgout = nr_anon = nr_file = nr_huge = 0; |
747db954c mm: memcontrol: u... |
5235 |
} |
1306a85ae mm: embed the mem... |
5236 |
memcg = page->mem_cgroup; |
747db954c mm: memcontrol: u... |
5237 5238 5239 5240 5241 5242 5243 5244 5245 5246 5247 5248 |
} if (PageTransHuge(page)) { nr_pages <<= compound_order(page); VM_BUG_ON_PAGE(!PageTransHuge(page), page); nr_huge += nr_pages; } if (PageAnon(page)) nr_anon += nr_pages; else nr_file += nr_pages; |
1306a85ae mm: embed the mem... |
5249 |
page->mem_cgroup = NULL; |
747db954c mm: memcontrol: u... |
5250 5251 5252 5253 5254 |
pgpgout++; } while (next != page_list); if (memcg) |
18eca2e63 mm: memcontrol: r... |
5255 5256 |
uncharge_batch(memcg, pgpgout, nr_anon, nr_file, nr_huge, page); |
747db954c mm: memcontrol: u... |
5257 |
} |
0a31bc97c mm: memcontrol: r... |
5258 5259 5260 5261 5262 5263 5264 5265 5266 |
/** * mem_cgroup_uncharge - uncharge a page * @page: page to uncharge * * Uncharge a page previously charged with mem_cgroup_try_charge() and * mem_cgroup_commit_charge(). */ void mem_cgroup_uncharge(struct page *page) { |
0a31bc97c mm: memcontrol: r... |
5267 5268 |
if (mem_cgroup_disabled()) return; |
747db954c mm: memcontrol: u... |
5269 |
/* Don't touch page->lru of any random page, pre-check: */ |
1306a85ae mm: embed the mem... |
5270 |
if (!page->mem_cgroup) |
0a31bc97c mm: memcontrol: r... |
5271 |
return; |
747db954c mm: memcontrol: u... |
5272 5273 5274 |
INIT_LIST_HEAD(&page->lru); uncharge_list(&page->lru); } |
0a31bc97c mm: memcontrol: r... |
5275 |
|
747db954c mm: memcontrol: u... |
5276 5277 5278 5279 5280 5281 5282 5283 5284 5285 5286 |
/** * mem_cgroup_uncharge_list - uncharge a list of page * @page_list: list of pages to uncharge * * Uncharge a list of pages previously charged with * mem_cgroup_try_charge() and mem_cgroup_commit_charge(). */ void mem_cgroup_uncharge_list(struct list_head *page_list) { if (mem_cgroup_disabled()) return; |
0a31bc97c mm: memcontrol: r... |
5287 |
|
747db954c mm: memcontrol: u... |
5288 5289 |
if (!list_empty(page_list)) uncharge_list(page_list); |
0a31bc97c mm: memcontrol: r... |
5290 5291 5292 |
} /** |
6a93ca8fd mm: migrate: do n... |
5293 5294 5295 |
* mem_cgroup_migrate - charge a page's replacement * @oldpage: currently circulating page * @newpage: replacement page |
0a31bc97c mm: memcontrol: r... |
5296 |
* |
6a93ca8fd mm: migrate: do n... |
5297 5298 |
* Charge @newpage as a replacement page for @oldpage. @oldpage will * be uncharged upon free. |
0a31bc97c mm: memcontrol: r... |
5299 5300 5301 |
* * Both pages must be locked, @newpage->mapping must be set up. */ |
6a93ca8fd mm: migrate: do n... |
5302 |
void mem_cgroup_migrate(struct page *oldpage, struct page *newpage) |
0a31bc97c mm: memcontrol: r... |
5303 |
{ |
298333157 mm: memcontrol: r... |
5304 |
struct mem_cgroup *memcg; |
44b7a8d33 mm: memcontrol: d... |
5305 5306 |
unsigned int nr_pages; bool compound; |
0a31bc97c mm: memcontrol: r... |
5307 5308 5309 |
VM_BUG_ON_PAGE(!PageLocked(oldpage), oldpage); VM_BUG_ON_PAGE(!PageLocked(newpage), newpage); |
0a31bc97c mm: memcontrol: r... |
5310 |
VM_BUG_ON_PAGE(PageAnon(oldpage) != PageAnon(newpage), newpage); |
6abb5a867 mm: memcontrol: a... |
5311 5312 |
VM_BUG_ON_PAGE(PageTransHuge(oldpage) != PageTransHuge(newpage), newpage); |
0a31bc97c mm: memcontrol: r... |
5313 5314 5315 5316 5317 |
if (mem_cgroup_disabled()) return; /* Page cache replacement: new page already charged? */ |
1306a85ae mm: embed the mem... |
5318 |
if (newpage->mem_cgroup) |
0a31bc97c mm: memcontrol: r... |
5319 |
return; |
45637bab3 mm: rename mem_cg... |
5320 |
/* Swapcache readahead pages can get replaced before being charged */ |
1306a85ae mm: embed the mem... |
5321 |
memcg = oldpage->mem_cgroup; |
298333157 mm: memcontrol: r... |
5322 |
if (!memcg) |
0a31bc97c mm: memcontrol: r... |
5323 |
return; |
44b7a8d33 mm: memcontrol: d... |
5324 5325 5326 5327 5328 5329 5330 5331 |
/* Force-charge the new page. The old one will be freed soon */ compound = PageTransHuge(newpage); nr_pages = compound ? hpage_nr_pages(newpage) : 1; page_counter_charge(&memcg->memory, nr_pages); if (do_memsw_account()) page_counter_charge(&memcg->memsw, nr_pages); css_get_many(&memcg->css, nr_pages); |
0a31bc97c mm: memcontrol: r... |
5332 |
|
9cf7666ac mm: memcontrol: d... |
5333 |
commit_charge(newpage, memcg, false); |
44b7a8d33 mm: memcontrol: d... |
5334 5335 5336 5337 5338 |
local_irq_disable(); mem_cgroup_charge_statistics(memcg, newpage, compound, nr_pages); memcg_check_events(memcg, newpage); local_irq_enable(); |
0a31bc97c mm: memcontrol: r... |
5339 |
} |
ef12947c9 mm: memcontrol: s... |
5340 |
DEFINE_STATIC_KEY_FALSE(memcg_sockets_enabled_key); |
110920876 mm: memcontrol: m... |
5341 5342 5343 5344 5345 5346 5347 5348 5349 5350 5351 5352 5353 5354 5355 5356 5357 5358 5359 5360 5361 5362 |
EXPORT_SYMBOL(memcg_sockets_enabled_key); void sock_update_memcg(struct sock *sk) { struct mem_cgroup *memcg; /* 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_memcg) { BUG_ON(mem_cgroup_is_root(sk->sk_memcg)); css_get(&sk->sk_memcg->css); return; } rcu_read_lock(); memcg = mem_cgroup_from_task(current); |
f7e1cb6ec mm: memcontrol: a... |
5363 5364 |
if (memcg == root_mem_cgroup) goto out; |
0db152981 mm: memcontrol: f... |
5365 |
if (!cgroup_subsys_on_dfl(memory_cgrp_subsys) && !memcg->tcpmem_active) |
f7e1cb6ec mm: memcontrol: a... |
5366 |
goto out; |
f7e1cb6ec mm: memcontrol: a... |
5367 |
if (css_tryget_online(&memcg->css)) |
110920876 mm: memcontrol: m... |
5368 |
sk->sk_memcg = memcg; |
f7e1cb6ec mm: memcontrol: a... |
5369 |
out: |
110920876 mm: memcontrol: m... |
5370 5371 5372 5373 5374 5375 5376 5377 5378 5379 5380 5381 5382 5383 5384 5385 5386 5387 5388 5389 |
rcu_read_unlock(); } EXPORT_SYMBOL(sock_update_memcg); void sock_release_memcg(struct sock *sk) { WARN_ON(!sk->sk_memcg); css_put(&sk->sk_memcg->css); } /** * mem_cgroup_charge_skmem - charge socket memory * @memcg: memcg to charge * @nr_pages: number of pages to charge * * Charges @nr_pages to @memcg. Returns %true if the charge fit within * @memcg's configured limit, %false if the charge had to be forced. */ bool mem_cgroup_charge_skmem(struct mem_cgroup *memcg, unsigned int nr_pages) { |
f7e1cb6ec mm: memcontrol: a... |
5390 |
gfp_t gfp_mask = GFP_KERNEL; |
110920876 mm: memcontrol: m... |
5391 |
|
f7e1cb6ec mm: memcontrol: a... |
5392 |
if (!cgroup_subsys_on_dfl(memory_cgrp_subsys)) { |
0db152981 mm: memcontrol: f... |
5393 |
struct page_counter *fail; |
f7e1cb6ec mm: memcontrol: a... |
5394 |
|
0db152981 mm: memcontrol: f... |
5395 5396 |
if (page_counter_try_charge(&memcg->tcpmem, nr_pages, &fail)) { memcg->tcpmem_pressure = 0; |
f7e1cb6ec mm: memcontrol: a... |
5397 5398 |
return true; } |
0db152981 mm: memcontrol: f... |
5399 5400 |
page_counter_charge(&memcg->tcpmem, nr_pages); memcg->tcpmem_pressure = 1; |
f7e1cb6ec mm: memcontrol: a... |
5401 |
return false; |
110920876 mm: memcontrol: m... |
5402 |
} |
d886f4e48 mm: memcontrol: r... |
5403 |
|
f7e1cb6ec mm: memcontrol: a... |
5404 5405 5406 |
/* Don't block in the packet receive path */ if (in_softirq()) gfp_mask = GFP_NOWAIT; |
b2807f07f mm: memcontrol: a... |
5407 |
this_cpu_add(memcg->stat->count[MEMCG_SOCK], nr_pages); |
f7e1cb6ec mm: memcontrol: a... |
5408 5409 5410 5411 |
if (try_charge(memcg, gfp_mask, nr_pages) == 0) return true; try_charge(memcg, gfp_mask|__GFP_NOFAIL, nr_pages); |
110920876 mm: memcontrol: m... |
5412 5413 5414 5415 5416 5417 5418 5419 5420 5421 |
return false; } /** * mem_cgroup_uncharge_skmem - uncharge socket memory * @memcg - memcg to uncharge * @nr_pages - number of pages to uncharge */ void mem_cgroup_uncharge_skmem(struct mem_cgroup *memcg, unsigned int nr_pages) { |
f7e1cb6ec mm: memcontrol: a... |
5422 |
if (!cgroup_subsys_on_dfl(memory_cgrp_subsys)) { |
0db152981 mm: memcontrol: f... |
5423 |
page_counter_uncharge(&memcg->tcpmem, nr_pages); |
f7e1cb6ec mm: memcontrol: a... |
5424 5425 |
return; } |
d886f4e48 mm: memcontrol: r... |
5426 |
|
b2807f07f mm: memcontrol: a... |
5427 |
this_cpu_sub(memcg->stat->count[MEMCG_SOCK], nr_pages); |
f7e1cb6ec mm: memcontrol: a... |
5428 5429 |
page_counter_uncharge(&memcg->memory, nr_pages); css_put_many(&memcg->css, nr_pages); |
110920876 mm: memcontrol: m... |
5430 |
} |
f7e1cb6ec mm: memcontrol: a... |
5431 5432 5433 5434 5435 5436 5437 5438 5439 |
static int __init cgroup_memory(char *s) { char *token; while ((token = strsep(&s, ",")) != NULL) { if (!*token) continue; if (!strcmp(token, "nosocket")) cgroup_memory_nosocket = true; |
04823c833 mm: memcontrol: a... |
5440 5441 |
if (!strcmp(token, "nokmem")) cgroup_memory_nokmem = true; |
f7e1cb6ec mm: memcontrol: a... |
5442 5443 5444 5445 |
} return 0; } __setup("cgroup.memory=", cgroup_memory); |
110920876 mm: memcontrol: m... |
5446 |
|
2d11085e4 memcg: do not cre... |
5447 |
/* |
1081312f9 memcg: cleanup me... |
5448 5449 5450 5451 5452 5453 |
* subsys_initcall() for memory controller. * * Some parts like hotcpu_notifier() have to be initialized from this context * because of lock dependencies (cgroup_lock -> cpu hotplug) but basically * everything that doesn't depend on a specific mem_cgroup structure should * be initialized from here. |
2d11085e4 memcg: do not cre... |
5454 5455 5456 |
*/ static int __init mem_cgroup_init(void) { |
95a045f63 mm: memcontrol: c... |
5457 |
int cpu, node; |
2d11085e4 memcg: do not cre... |
5458 |
hotcpu_notifier(memcg_cpu_hotplug_callback, 0); |
95a045f63 mm: memcontrol: c... |
5459 5460 5461 5462 5463 5464 5465 5466 5467 5468 5469 5470 5471 5472 5473 5474 5475 5476 5477 5478 5479 |
for_each_possible_cpu(cpu) INIT_WORK(&per_cpu_ptr(&memcg_stock, cpu)->work, drain_local_stock); for_each_node(node) { struct mem_cgroup_tree_per_node *rtpn; int zone; rtpn = kzalloc_node(sizeof(*rtpn), GFP_KERNEL, node_online(node) ? node : NUMA_NO_NODE); for (zone = 0; zone < MAX_NR_ZONES; zone++) { struct mem_cgroup_tree_per_zone *rtpz; rtpz = &rtpn->rb_tree_per_zone[zone]; rtpz->rb_root = RB_ROOT; spin_lock_init(&rtpz->lock); } soft_limit_tree.rb_tree_per_node[node] = rtpn; } |
2d11085e4 memcg: do not cre... |
5480 5481 5482 |
return 0; } subsys_initcall(mem_cgroup_init); |
21afa38ee mm: memcontrol: c... |
5483 5484 5485 5486 5487 5488 5489 5490 5491 5492 5493 5494 5495 5496 5497 5498 |
#ifdef CONFIG_MEMCG_SWAP /** * mem_cgroup_swapout - transfer a memsw charge to swap * @page: page whose memsw charge to transfer * @entry: swap entry to move the charge to * * Transfer the memsw charge of @page to @entry. */ void mem_cgroup_swapout(struct page *page, swp_entry_t entry) { struct mem_cgroup *memcg; unsigned short oldid; VM_BUG_ON_PAGE(PageLRU(page), page); VM_BUG_ON_PAGE(page_count(page), page); |
7941d2145 mm: memcontrol: d... |
5499 |
if (!do_memsw_account()) |
21afa38ee mm: memcontrol: c... |
5500 5501 5502 5503 5504 5505 5506 5507 5508 5509 5510 5511 5512 5513 5514 5515 |
return; memcg = page->mem_cgroup; /* Readahead page, never charged */ if (!memcg) return; oldid = swap_cgroup_record(entry, mem_cgroup_id(memcg)); VM_BUG_ON_PAGE(oldid, page); mem_cgroup_swap_statistics(memcg, true); page->mem_cgroup = NULL; if (!mem_cgroup_is_root(memcg)) page_counter_uncharge(&memcg->memory, 1); |
ce9ce6659 mm: memcontrol: b... |
5516 5517 5518 5519 5520 5521 5522 |
/* * Interrupts should be disabled here because the caller holds the * mapping->tree_lock lock which is taken with interrupts-off. It is * important here to have the interrupts disabled because it is the * only synchronisation we have for udpating the per-CPU variables. */ VM_BUG_ON(!irqs_disabled()); |
f627c2f53 memcg: adjust to ... |
5523 |
mem_cgroup_charge_statistics(memcg, page, false, -1); |
21afa38ee mm: memcontrol: c... |
5524 5525 |
memcg_check_events(memcg, page); } |
37e843511 mm: memcontrol: c... |
5526 5527 5528 5529 5530 5531 5532 5533 5534 5535 5536 5537 5538 5539 5540 5541 5542 5543 5544 5545 5546 5547 5548 5549 5550 5551 5552 5553 5554 5555 5556 5557 5558 5559 5560 |
/* * mem_cgroup_try_charge_swap - try charging a swap entry * @page: page being added to swap * @entry: swap entry to charge * * Try to charge @entry to the memcg that @page belongs to. * * Returns 0 on success, -ENOMEM on failure. */ int mem_cgroup_try_charge_swap(struct page *page, swp_entry_t entry) { struct mem_cgroup *memcg; struct page_counter *counter; unsigned short oldid; if (!cgroup_subsys_on_dfl(memory_cgrp_subsys) || !do_swap_account) return 0; memcg = page->mem_cgroup; /* Readahead page, never charged */ if (!memcg) return 0; if (!mem_cgroup_is_root(memcg) && !page_counter_try_charge(&memcg->swap, 1, &counter)) return -ENOMEM; oldid = swap_cgroup_record(entry, mem_cgroup_id(memcg)); VM_BUG_ON_PAGE(oldid, page); mem_cgroup_swap_statistics(memcg, true); css_get(&memcg->css); return 0; } |
21afa38ee mm: memcontrol: c... |
5561 5562 5563 5564 |
/** * mem_cgroup_uncharge_swap - uncharge a swap entry * @entry: swap entry to uncharge * |
37e843511 mm: memcontrol: c... |
5565 |
* Drop the swap charge associated with @entry. |
21afa38ee mm: memcontrol: c... |
5566 5567 5568 5569 5570 |
*/ void mem_cgroup_uncharge_swap(swp_entry_t entry) { struct mem_cgroup *memcg; unsigned short id; |
37e843511 mm: memcontrol: c... |
5571 |
if (!do_swap_account) |
21afa38ee mm: memcontrol: c... |
5572 5573 5574 5575 |
return; id = swap_cgroup_record(entry, 0); rcu_read_lock(); |
adbe427b9 memcg: zap mem_cg... |
5576 |
memcg = mem_cgroup_from_id(id); |
21afa38ee mm: memcontrol: c... |
5577 |
if (memcg) { |
37e843511 mm: memcontrol: c... |
5578 5579 5580 5581 5582 5583 |
if (!mem_cgroup_is_root(memcg)) { if (cgroup_subsys_on_dfl(memory_cgrp_subsys)) page_counter_uncharge(&memcg->swap, 1); else page_counter_uncharge(&memcg->memsw, 1); } |
21afa38ee mm: memcontrol: c... |
5584 5585 5586 5587 5588 |
mem_cgroup_swap_statistics(memcg, false); css_put(&memcg->css); } rcu_read_unlock(); } |
d8b38438a mm: vmscan: do no... |
5589 5590 5591 5592 5593 5594 5595 5596 5597 5598 5599 5600 |
long mem_cgroup_get_nr_swap_pages(struct mem_cgroup *memcg) { long nr_swap_pages = get_nr_swap_pages(); if (!do_swap_account || !cgroup_subsys_on_dfl(memory_cgrp_subsys)) return nr_swap_pages; for (; memcg != root_mem_cgroup; memcg = parent_mem_cgroup(memcg)) nr_swap_pages = min_t(long, nr_swap_pages, READ_ONCE(memcg->swap.limit) - page_counter_read(&memcg->swap)); return nr_swap_pages; } |
5ccc5abaa mm: free swap cac... |
5601 5602 5603 5604 5605 5606 5607 5608 5609 5610 5611 5612 5613 5614 5615 5616 5617 5618 5619 5620 5621 |
bool mem_cgroup_swap_full(struct page *page) { struct mem_cgroup *memcg; VM_BUG_ON_PAGE(!PageLocked(page), page); if (vm_swap_full()) return true; if (!do_swap_account || !cgroup_subsys_on_dfl(memory_cgrp_subsys)) return false; memcg = page->mem_cgroup; if (!memcg) return false; for (; memcg != root_mem_cgroup; memcg = parent_mem_cgroup(memcg)) if (page_counter_read(&memcg->swap) * 2 >= memcg->swap.limit) return true; return false; } |
21afa38ee mm: memcontrol: c... |
5622 5623 5624 5625 5626 5627 5628 5629 5630 5631 5632 5633 5634 5635 5636 5637 |
/* for remember boot option*/ #ifdef CONFIG_MEMCG_SWAP_ENABLED static int really_do_swap_account __initdata = 1; #else static int really_do_swap_account __initdata; #endif static int __init enable_swap_account(char *s) { if (!strcmp(s, "1")) really_do_swap_account = 1; else if (!strcmp(s, "0")) really_do_swap_account = 0; return 1; } __setup("swapaccount=", enable_swap_account); |
37e843511 mm: memcontrol: c... |
5638 5639 5640 5641 5642 5643 5644 5645 5646 5647 5648 5649 5650 5651 5652 5653 5654 5655 5656 5657 5658 5659 5660 5661 5662 5663 5664 5665 5666 5667 5668 5669 5670 5671 5672 5673 5674 5675 5676 5677 5678 5679 5680 5681 5682 5683 5684 5685 5686 5687 5688 5689 5690 5691 5692 5693 5694 5695 |
static u64 swap_current_read(struct cgroup_subsys_state *css, struct cftype *cft) { struct mem_cgroup *memcg = mem_cgroup_from_css(css); return (u64)page_counter_read(&memcg->swap) * PAGE_SIZE; } static int swap_max_show(struct seq_file *m, void *v) { struct mem_cgroup *memcg = mem_cgroup_from_css(seq_css(m)); unsigned long max = READ_ONCE(memcg->swap.limit); if (max == PAGE_COUNTER_MAX) seq_puts(m, "max "); else seq_printf(m, "%llu ", (u64)max * PAGE_SIZE); return 0; } static ssize_t swap_max_write(struct kernfs_open_file *of, char *buf, size_t nbytes, loff_t off) { struct mem_cgroup *memcg = mem_cgroup_from_css(of_css(of)); unsigned long max; int err; buf = strstrip(buf); err = page_counter_memparse(buf, "max", &max); if (err) return err; mutex_lock(&memcg_limit_mutex); err = page_counter_limit(&memcg->swap, max); mutex_unlock(&memcg_limit_mutex); if (err) return err; return nbytes; } static struct cftype swap_files[] = { { .name = "swap.current", .flags = CFTYPE_NOT_ON_ROOT, .read_u64 = swap_current_read, }, { .name = "swap.max", .flags = CFTYPE_NOT_ON_ROOT, .seq_show = swap_max_show, .write = swap_max_write, }, { } /* terminate */ }; |
21afa38ee mm: memcontrol: c... |
5696 5697 5698 5699 5700 5701 5702 5703 5704 5705 5706 5707 5708 5709 5710 5711 5712 5713 5714 5715 5716 5717 5718 5719 5720 5721 5722 5723 5724 5725 5726 |
static struct cftype memsw_cgroup_files[] = { { .name = "memsw.usage_in_bytes", .private = MEMFILE_PRIVATE(_MEMSWAP, RES_USAGE), .read_u64 = mem_cgroup_read_u64, }, { .name = "memsw.max_usage_in_bytes", .private = MEMFILE_PRIVATE(_MEMSWAP, RES_MAX_USAGE), .write = mem_cgroup_reset, .read_u64 = mem_cgroup_read_u64, }, { .name = "memsw.limit_in_bytes", .private = MEMFILE_PRIVATE(_MEMSWAP, RES_LIMIT), .write = mem_cgroup_write, .read_u64 = mem_cgroup_read_u64, }, { .name = "memsw.failcnt", .private = MEMFILE_PRIVATE(_MEMSWAP, RES_FAILCNT), .write = mem_cgroup_reset, .read_u64 = mem_cgroup_read_u64, }, { }, /* terminate */ }; static int __init mem_cgroup_swap_init(void) { if (!mem_cgroup_disabled() && really_do_swap_account) { do_swap_account = 1; |
37e843511 mm: memcontrol: c... |
5727 5728 |
WARN_ON(cgroup_add_dfl_cftypes(&memory_cgrp_subsys, swap_files)); |
21afa38ee mm: memcontrol: c... |
5729 5730 5731 5732 5733 5734 5735 5736 |
WARN_ON(cgroup_add_legacy_cftypes(&memory_cgrp_subsys, memsw_cgroup_files)); } return 0; } subsys_initcall(mem_cgroup_swap_init); #endif /* CONFIG_MEMCG_SWAP */ |