Blame view
mm/oom_kill.c
22.3 KB
1da177e4c
|
1 2 3 4 5 6 |
/* * linux/mm/oom_kill.c * * Copyright (C) 1998,2000 Rik van Riel * Thanks go out to Claus Fischer for some serious inspiration and * for goading me into coding this file... |
a63d83f42
|
7 8 |
* Copyright (C) 2010 Google, Inc. * Rewritten by David Rientjes |
1da177e4c
|
9 10 |
* * The routines in this file are used to kill a process when |
a49335cce
|
11 12 |
* we're seriously out of memory. This gets called from __alloc_pages() * in mm/page_alloc.c when we really run out of memory. |
1da177e4c
|
13 14 15 16 17 18 |
* * Since we won't call these routines often (on a well-configured * machine) this file will double as a 'coding guide' and a signpost * for newbie kernel hackers. It features several pointers to major * kernel subsystems and hints as to where to find out what things do. */ |
8ac773b4f
|
19 |
#include <linux/oom.h> |
1da177e4c
|
20 |
#include <linux/mm.h> |
4e950f6f0
|
21 |
#include <linux/err.h> |
5a0e3ad6a
|
22 |
#include <linux/gfp.h> |
1da177e4c
|
23 24 25 26 |
#include <linux/sched.h> #include <linux/swap.h> #include <linux/timex.h> #include <linux/jiffies.h> |
ef08e3b49
|
27 |
#include <linux/cpuset.h> |
b95f1b31b
|
28 |
#include <linux/export.h> |
8bc719d3c
|
29 |
#include <linux/notifier.h> |
c7ba5c9e8
|
30 |
#include <linux/memcontrol.h> |
6f48d0ebd
|
31 |
#include <linux/mempolicy.h> |
5cd9c58fb
|
32 |
#include <linux/security.h> |
edd45544c
|
33 |
#include <linux/ptrace.h> |
f660daac4
|
34 |
#include <linux/freezer.h> |
43d2b1132
|
35 |
#include <linux/ftrace.h> |
dc3f21ead
|
36 |
#include <linux/ratelimit.h> |
43d2b1132
|
37 38 39 |
#define CREATE_TRACE_POINTS #include <trace/events/oom.h> |
1da177e4c
|
40 |
|
fadd8fbd1
|
41 |
int sysctl_panic_on_oom; |
fe071d7e8
|
42 |
int sysctl_oom_kill_allocating_task; |
ad915c432
|
43 |
int sysctl_oom_dump_tasks = 1; |
c7d4caeb1
|
44 |
static DEFINE_SPINLOCK(zone_scan_lock); |
1da177e4c
|
45 |
|
43362a497
|
46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 |
/* * compare_swap_oom_score_adj() - compare and swap current's oom_score_adj * @old_val: old oom_score_adj for compare * @new_val: new oom_score_adj for swap * * Sets the oom_score_adj value for current to @new_val iff its present value is * @old_val. Usually used to reinstate a previous value to prevent racing with * userspacing tuning the value in the interim. */ void compare_swap_oom_score_adj(int old_val, int new_val) { struct sighand_struct *sighand = current->sighand; spin_lock_irq(&sighand->siglock); if (current->signal->oom_score_adj == old_val) current->signal->oom_score_adj = new_val; |
43d2b1132
|
62 |
trace_oom_score_adj_update(current); |
43362a497
|
63 64 |
spin_unlock_irq(&sighand->siglock); } |
72788c385
|
65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 |
/** * test_set_oom_score_adj() - set current's oom_score_adj and return old value * @new_val: new oom_score_adj value * * Sets the oom_score_adj value for current to @new_val with proper * synchronization and returns the old value. Usually used to temporarily * set a value, save the old value in the caller, and then reinstate it later. */ int test_set_oom_score_adj(int new_val) { struct sighand_struct *sighand = current->sighand; int old_val; spin_lock_irq(&sighand->siglock); old_val = current->signal->oom_score_adj; |
c9f01245b
|
80 |
current->signal->oom_score_adj = new_val; |
43d2b1132
|
81 |
trace_oom_score_adj_update(current); |
72788c385
|
82 83 84 85 |
spin_unlock_irq(&sighand->siglock); return old_val; } |
6f48d0ebd
|
86 87 88 89 90 91 92 93 94 |
#ifdef CONFIG_NUMA /** * has_intersects_mems_allowed() - check task eligiblity for kill * @tsk: task struct of which task to consider * @mask: nodemask passed to page allocator for mempolicy ooms * * Task eligibility is determined by whether or not a candidate task, @tsk, * shares the same mempolicy nodes as current if it is bound by such a policy * and whether or not it has the same set of allowed cpuset nodes. |
495789a51
|
95 |
*/ |
6f48d0ebd
|
96 97 |
static bool has_intersects_mems_allowed(struct task_struct *tsk, const nodemask_t *mask) |
495789a51
|
98 |
{ |
6f48d0ebd
|
99 |
struct task_struct *start = tsk; |
495789a51
|
100 |
|
495789a51
|
101 |
do { |
6f48d0ebd
|
102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 |
if (mask) { /* * If this is a mempolicy constrained oom, tsk's * cpuset is irrelevant. Only return true if its * mempolicy intersects current, otherwise it may be * needlessly killed. */ if (mempolicy_nodemask_intersects(tsk, mask)) return true; } else { /* * This is not a mempolicy constrained oom, so only * check the mems of tsk's cpuset. */ if (cpuset_mems_allowed_intersects(current, tsk)) return true; } |
df1090a8d
|
119 |
} while_each_thread(start, tsk); |
6f48d0ebd
|
120 121 122 123 124 125 126 |
return false; } #else static bool has_intersects_mems_allowed(struct task_struct *tsk, const nodemask_t *mask) { return true; |
495789a51
|
127 |
} |
6f48d0ebd
|
128 |
#endif /* CONFIG_NUMA */ |
495789a51
|
129 |
|
6f48d0ebd
|
130 131 132 133 134 135 |
/* * The process p may have detached its own ->mm while exiting or through * use_mm(), but one or more of its subthreads may still have a valid * pointer. Return p, or any of its subthreads with a valid ->mm, with * task_lock() held. */ |
158e0a2d1
|
136 |
struct task_struct *find_lock_task_mm(struct task_struct *p) |
dd8e8f405
|
137 138 139 140 141 142 143 144 145 146 147 148 |
{ struct task_struct *t = p; do { task_lock(t); if (likely(t->mm)) return t; task_unlock(t); } while_each_thread(p, t); return NULL; } |
ab290adba
|
149 |
/* return true if the task is not adequate as candidate victim task. */ |
e85bfd3aa
|
150 |
static bool oom_unkillable_task(struct task_struct *p, |
72835c86c
|
151 |
const struct mem_cgroup *memcg, const nodemask_t *nodemask) |
ab290adba
|
152 153 154 155 156 157 158 |
{ if (is_global_init(p)) return true; if (p->flags & PF_KTHREAD) return true; /* When mem_cgroup_out_of_memory() and p is not member of the group */ |
72835c86c
|
159 |
if (memcg && !task_in_mem_cgroup(p, memcg)) |
ab290adba
|
160 161 162 163 164 165 166 167 |
return true; /* p may not have freeable memory in nodemask */ if (!has_intersects_mems_allowed(p, nodemask)) return true; return false; } |
1da177e4c
|
168 |
/** |
a63d83f42
|
169 |
* oom_badness - heuristic function to determine which candidate task to kill |
1da177e4c
|
170 |
* @p: task struct of which task we should calculate |
a63d83f42
|
171 |
* @totalpages: total present RAM allowed for page allocation |
1da177e4c
|
172 |
* |
a63d83f42
|
173 174 175 |
* The heuristic for determining which task to kill is made to be as simple and * predictable as possible. The goal is to return the highest value for the * task consuming the most memory to avoid subsequent oom failures. |
1da177e4c
|
176 |
*/ |
72835c86c
|
177 |
unsigned int oom_badness(struct task_struct *p, struct mem_cgroup *memcg, |
a63d83f42
|
178 |
const nodemask_t *nodemask, unsigned long totalpages) |
1da177e4c
|
179 |
{ |
ff05b6f7a
|
180 |
long points; |
28b83c519
|
181 |
|
72835c86c
|
182 |
if (oom_unkillable_task(p, memcg, nodemask)) |
26ebc9849
|
183 |
return 0; |
1da177e4c
|
184 |
|
dd8e8f405
|
185 186 |
p = find_lock_task_mm(p); if (!p) |
1da177e4c
|
187 |
return 0; |
5aecc85ab
|
188 189 190 191 |
if (p->signal->oom_score_adj == OOM_SCORE_ADJ_MIN) { task_unlock(p); return 0; } |
1da177e4c
|
192 |
/* |
a63d83f42
|
193 194 |
* The memory controller may have a limit of 0 bytes, so avoid a divide * by zero, if necessary. |
1da177e4c
|
195 |
*/ |
a63d83f42
|
196 197 |
if (!totalpages) totalpages = 1; |
1da177e4c
|
198 199 |
/* |
a63d83f42
|
200 |
* The baseline for the badness score is the proportion of RAM that each |
f755a042d
|
201 |
* task's rss, pagetable and swap space use. |
1da177e4c
|
202 |
*/ |
f755a042d
|
203 204 205 206 207 |
points = get_mm_rss(p->mm) + p->mm->nr_ptes; points += get_mm_counter(p->mm, MM_SWAPENTS); points *= 1000; points /= totalpages; |
a63d83f42
|
208 |
task_unlock(p); |
1da177e4c
|
209 210 |
/* |
a63d83f42
|
211 212 |
* Root processes get 3% bonus, just like the __vm_enough_memory() * implementation used by LSMs. |
1da177e4c
|
213 |
*/ |
a63d83f42
|
214 215 |
if (has_capability_noaudit(p, CAP_SYS_ADMIN)) points -= 30; |
1da177e4c
|
216 217 |
/* |
a63d83f42
|
218 219 220 |
* /proc/pid/oom_score_adj ranges from -1000 to +1000 such that it may * either completely disable oom killing or always prefer a certain * task. |
1da177e4c
|
221 |
*/ |
a63d83f42
|
222 |
points += p->signal->oom_score_adj; |
1da177e4c
|
223 |
|
f19e8aa11
|
224 225 226 227 228 229 230 |
/* * Never return 0 for an eligible task that may be killed since it's * possible that no single user task uses more than 0.1% of memory and * no single admin tasks uses more than 3.0%. */ if (points <= 0) return 1; |
a63d83f42
|
231 |
return (points < 1000) ? points : 1000; |
1da177e4c
|
232 233 234 |
} /* |
9b0f8b040
|
235 236 |
* Determine the type of allocation constraint. */ |
9b0f8b040
|
237 |
#ifdef CONFIG_NUMA |
4365a5676
|
238 |
static enum oom_constraint constrained_alloc(struct zonelist *zonelist, |
a63d83f42
|
239 240 |
gfp_t gfp_mask, nodemask_t *nodemask, unsigned long *totalpages) |
4365a5676
|
241 |
{ |
54a6eb5c4
|
242 |
struct zone *zone; |
dd1a239f6
|
243 |
struct zoneref *z; |
54a6eb5c4
|
244 |
enum zone_type high_zoneidx = gfp_zone(gfp_mask); |
a63d83f42
|
245 246 |
bool cpuset_limited = false; int nid; |
9b0f8b040
|
247 |
|
a63d83f42
|
248 249 250 251 252 |
/* Default to all available memory */ *totalpages = totalram_pages + total_swap_pages; if (!zonelist) return CONSTRAINT_NONE; |
4365a5676
|
253 254 255 256 257 258 259 |
/* * Reach here only when __GFP_NOFAIL is used. So, we should avoid * to kill current.We have to random task kill in this case. * Hopefully, CONSTRAINT_THISNODE...but no way to handle it, now. */ if (gfp_mask & __GFP_THISNODE) return CONSTRAINT_NONE; |
9b0f8b040
|
260 |
|
4365a5676
|
261 |
/* |
a63d83f42
|
262 263 264 |
* This is not a __GFP_THISNODE allocation, so a truncated nodemask in * the page allocator means a mempolicy is in effect. Cpuset policy * is enforced in get_page_from_freelist(). |
4365a5676
|
265 |
*/ |
a63d83f42
|
266 267 268 269 |
if (nodemask && !nodes_subset(node_states[N_HIGH_MEMORY], *nodemask)) { *totalpages = total_swap_pages; for_each_node_mask(nid, *nodemask) *totalpages += node_spanned_pages(nid); |
9b0f8b040
|
270 |
return CONSTRAINT_MEMORY_POLICY; |
a63d83f42
|
271 |
} |
4365a5676
|
272 273 274 275 276 |
/* Check this allocation failure is caused by cpuset's wall function */ for_each_zone_zonelist_nodemask(zone, z, zonelist, high_zoneidx, nodemask) if (!cpuset_zone_allowed_softwall(zone, gfp_mask)) |
a63d83f42
|
277 |
cpuset_limited = true; |
9b0f8b040
|
278 |
|
a63d83f42
|
279 280 281 282 283 284 |
if (cpuset_limited) { *totalpages = total_swap_pages; for_each_node_mask(nid, cpuset_current_mems_allowed) *totalpages += node_spanned_pages(nid); return CONSTRAINT_CPUSET; } |
9b0f8b040
|
285 286 |
return CONSTRAINT_NONE; } |
4365a5676
|
287 288 |
#else static enum oom_constraint constrained_alloc(struct zonelist *zonelist, |
a63d83f42
|
289 290 |
gfp_t gfp_mask, nodemask_t *nodemask, unsigned long *totalpages) |
4365a5676
|
291 |
{ |
a63d83f42
|
292 |
*totalpages = totalram_pages + total_swap_pages; |
4365a5676
|
293 294 295 |
return CONSTRAINT_NONE; } #endif |
9b0f8b040
|
296 297 |
/* |
1da177e4c
|
298 299 300 301 302 |
* Simple selection loop. We chose the process with the highest * number of 'points'. We expect the caller will lock the tasklist. * * (not docbooked, we don't want this one cluttering up the manual) */ |
a63d83f42
|
303 |
static struct task_struct *select_bad_process(unsigned int *ppoints, |
72835c86c
|
304 |
unsigned long totalpages, struct mem_cgroup *memcg, |
08ab9b10d
|
305 |
const nodemask_t *nodemask, bool force_kill) |
1da177e4c
|
306 |
{ |
3a5dda7a1
|
307 |
struct task_struct *g, *p; |
1da177e4c
|
308 |
struct task_struct *chosen = NULL; |
9827b781f
|
309 |
*ppoints = 0; |
1da177e4c
|
310 |
|
3a5dda7a1
|
311 |
do_each_thread(g, p) { |
a63d83f42
|
312 |
unsigned int points; |
a49335cce
|
313 |
|
c027a474a
|
314 |
if (p->exit_state) |
30e2b41f2
|
315 |
continue; |
72835c86c
|
316 |
if (oom_unkillable_task(p, memcg, nodemask)) |
6cf86ac6f
|
317 |
continue; |
ef08e3b49
|
318 |
|
a49335cce
|
319 |
/* |
b78483a4b
|
320 321 322 323 324 325 326 327 |
* This task already has access to memory reserves and is * being killed. Don't allow any other task access to the * memory reserve. * * Note: this may have a chance of deadlock if it gets * blocked waiting for another task which itself is waiting * for memory. Is there a better alternative? */ |
f660daac4
|
328 329 |
if (test_tsk_thread_flag(p, TIF_MEMDIE)) { if (unlikely(frozen(p))) |
a5be2d0d1
|
330 |
__thaw_task(p); |
08ab9b10d
|
331 332 |
if (!force_kill) return ERR_PTR(-1UL); |
f660daac4
|
333 |
} |
c027a474a
|
334 335 |
if (!p->mm) continue; |
b78483a4b
|
336 |
|
30e2b41f2
|
337 |
if (p->flags & PF_EXITING) { |
edd45544c
|
338 339 340 341 342 343 344 345 346 347 348 349 |
/* * If p is the current task and is in the process of * releasing memory, we allow the "kill" to set * TIF_MEMDIE, which will allow it to gain access to * memory reserves. Otherwise, it may stall forever. * * The loop isn't broken here, however, in case other * threads are found to have already been oom killed. */ if (p == current) { chosen = p; *ppoints = 1000; |
08ab9b10d
|
350 |
} else if (!force_kill) { |
edd45544c
|
351 352 353 354 355 |
/* * If this task is not being ptraced on exit, * then wait for it to finish before killing * some other task unnecessarily. */ |
d21142ece
|
356 |
if (!(p->group_leader->ptrace & PT_TRACE_EXIT)) |
edd45544c
|
357 358 |
return ERR_PTR(-1UL); } |
50ec3bbff
|
359 |
} |
972c4ea59
|
360 |
|
72835c86c
|
361 |
points = oom_badness(p, memcg, nodemask, totalpages); |
a63d83f42
|
362 |
if (points > *ppoints) { |
a49335cce
|
363 |
chosen = p; |
9827b781f
|
364 |
*ppoints = points; |
1da177e4c
|
365 |
} |
3a5dda7a1
|
366 |
} while_each_thread(g, p); |
972c4ea59
|
367 |
|
1da177e4c
|
368 369 370 371 |
return chosen; } /** |
1b578df02
|
372 |
* dump_tasks - dump current memory state of all system tasks |
74ab7f1d3
|
373 |
* @mem: current's memory controller, if constrained |
e85bfd3aa
|
374 |
* @nodemask: nodemask passed to page allocator for mempolicy ooms |
1b578df02
|
375 |
* |
e85bfd3aa
|
376 377 378 |
* Dumps the current memory state of all eligible tasks. Tasks not in the same * memcg, not in the same cpuset, or bound to a disjoint set of mempolicy nodes * are not shown. |
fef1bdd68
|
379 |
* State information includes task's pid, uid, tgid, vm size, rss, cpu, oom_adj |
a63d83f42
|
380 |
* value, oom_score_adj value, and name. |
fef1bdd68
|
381 |
* |
fef1bdd68
|
382 383 |
* Call with tasklist_lock read-locked. */ |
72835c86c
|
384 |
static void dump_tasks(const struct mem_cgroup *memcg, const nodemask_t *nodemask) |
fef1bdd68
|
385 |
{ |
c55db9578
|
386 387 |
struct task_struct *p; struct task_struct *task; |
fef1bdd68
|
388 |
|
a63d83f42
|
389 390 |
pr_info("[ pid ] uid tgid total_vm rss cpu oom_adj oom_score_adj name "); |
c55db9578
|
391 |
for_each_process(p) { |
72835c86c
|
392 |
if (oom_unkillable_task(p, memcg, nodemask)) |
b4416d2be
|
393 |
continue; |
fef1bdd68
|
394 |
|
c55db9578
|
395 396 |
task = find_lock_task_mm(p); if (!task) { |
6d2661ede
|
397 |
/* |
74ab7f1d3
|
398 399 |
* This is a kthread or all of p's threads have already * detached their mm's. There's no need to report |
c55db9578
|
400 |
* them; they can't be oom killed anyway. |
6d2661ede
|
401 |
*/ |
6d2661ede
|
402 403 |
continue; } |
c55db9578
|
404 |
|
a63d83f42
|
405 406 |
pr_info("[%5d] %5d %5d %8lu %8lu %3u %3d %5d %s ", |
078de5f70
|
407 408 |
task->pid, from_kuid(&init_user_ns, task_uid(task)), task->tgid, task->mm->total_vm, get_mm_rss(task->mm), |
a63d83f42
|
409 410 |
task_cpu(task), task->signal->oom_adj, task->signal->oom_score_adj, task->comm); |
c55db9578
|
411 412 |
task_unlock(task); } |
fef1bdd68
|
413 |
} |
d31f56dbf
|
414 |
static void dump_header(struct task_struct *p, gfp_t gfp_mask, int order, |
72835c86c
|
415 |
struct mem_cgroup *memcg, const nodemask_t *nodemask) |
1b604d75b
|
416 |
{ |
5e9d834a0
|
417 |
task_lock(current); |
1b604d75b
|
418 |
pr_warning("%s invoked oom-killer: gfp_mask=0x%x, order=%d, " |
a63d83f42
|
419 420 421 422 |
"oom_adj=%d, oom_score_adj=%d ", current->comm, gfp_mask, order, current->signal->oom_adj, current->signal->oom_score_adj); |
1b604d75b
|
423 424 425 |
cpuset_print_task_mems_allowed(current); task_unlock(current); dump_stack(); |
72835c86c
|
426 |
mem_cgroup_print_oom_info(memcg, p); |
b2b755b5f
|
427 |
show_mem(SHOW_MEM_FILTER_NODES); |
1b604d75b
|
428 |
if (sysctl_oom_dump_tasks) |
72835c86c
|
429 |
dump_tasks(memcg, nodemask); |
1b604d75b
|
430 |
} |
3b4798cbc
|
431 |
#define K(x) ((x) << (PAGE_SHIFT-10)) |
2a1c9b1fc
|
432 433 434 435 |
static void oom_kill_process(struct task_struct *p, gfp_t gfp_mask, int order, unsigned int points, unsigned long totalpages, struct mem_cgroup *memcg, nodemask_t *nodemask, const char *message) |
1da177e4c
|
436 |
{ |
52d3c0367
|
437 |
struct task_struct *victim = p; |
5e9d834a0
|
438 |
struct task_struct *child; |
52d3c0367
|
439 |
struct task_struct *t = p; |
647f2bdf4
|
440 |
struct mm_struct *mm; |
52d3c0367
|
441 |
unsigned int victim_points = 0; |
dc3f21ead
|
442 443 |
static DEFINE_RATELIMIT_STATE(oom_rs, DEFAULT_RATELIMIT_INTERVAL, DEFAULT_RATELIMIT_BURST); |
1da177e4c
|
444 |
|
50ec3bbff
|
445 446 447 448 |
/* * If the task is already exiting, don't alarm the sysadmin or kill * its children or threads, just set TIF_MEMDIE so it can die quickly */ |
0753ba01e
|
449 |
if (p->flags & PF_EXITING) { |
4358997ae
|
450 |
set_tsk_thread_flag(p, TIF_MEMDIE); |
2a1c9b1fc
|
451 |
return; |
50ec3bbff
|
452 |
} |
dc3f21ead
|
453 |
if (__ratelimit(&oom_rs)) |
8447d950e
|
454 |
dump_header(p, gfp_mask, order, memcg, nodemask); |
5e9d834a0
|
455 |
task_lock(p); |
a63d83f42
|
456 457 |
pr_err("%s: Kill process %d (%s) score %d or sacrifice child ", |
5e9d834a0
|
458 459 |
message, task_pid_nr(p), p->comm, points); task_unlock(p); |
f3af38d30
|
460 |
|
5e9d834a0
|
461 462 |
/* * If any of p's children has a different mm and is eligible for kill, |
11239836c
|
463 |
* the one with the highest oom_badness() score is sacrificed for its |
5e9d834a0
|
464 465 466 |
* parent. This attempts to lose the minimal amount of work done while * still freeing memory. */ |
dd8e8f405
|
467 |
do { |
5e9d834a0
|
468 |
list_for_each_entry(child, &t->children, sibling) { |
a63d83f42
|
469 |
unsigned int child_points; |
5e9d834a0
|
470 |
|
edd45544c
|
471 472 |
if (child->mm == p->mm) continue; |
a63d83f42
|
473 474 475 |
/* * oom_badness() returns 0 if the thread is unkillable */ |
72835c86c
|
476 |
child_points = oom_badness(child, memcg, nodemask, |
a63d83f42
|
477 |
totalpages); |
5e9d834a0
|
478 479 480 481 |
if (child_points > victim_points) { victim = child; victim_points = child_points; } |
dd8e8f405
|
482 483 |
} } while_each_thread(p, t); |
647f2bdf4
|
484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 |
victim = find_lock_task_mm(victim); if (!victim) return; /* mm cannot safely be dereferenced after task_unlock(victim) */ mm = victim->mm; pr_err("Killed process %d (%s) total-vm:%lukB, anon-rss:%lukB, file-rss:%lukB ", task_pid_nr(victim), victim->comm, K(victim->mm->total_vm), K(get_mm_counter(victim->mm, MM_ANONPAGES)), K(get_mm_counter(victim->mm, MM_FILEPAGES))); task_unlock(victim); /* * Kill all user processes sharing victim->mm in other thread groups, if * any. They don't get access to memory reserves, though, to avoid * depletion of all memory. This prevents mm->mmap_sem livelock when an * oom killed thread cannot exit because it requires the semaphore and * its contended by another thread trying to allocate memory itself. * That thread will now get access to memory reserves since it has a * pending fatal signal. */ for_each_process(p) if (p->mm == mm && !same_thread_group(p, victim) && !(p->flags & PF_KTHREAD)) { if (p->signal->oom_score_adj == OOM_SCORE_ADJ_MIN) continue; task_lock(p); /* Protect ->comm from prctl() */ pr_err("Kill process %d (%s) sharing same memory ", task_pid_nr(p), p->comm); task_unlock(p); |
d2d393099
|
517 |
do_send_sig_info(SIGKILL, SEND_SIG_FORCED, p, true); |
647f2bdf4
|
518 519 520 |
} set_tsk_thread_flag(victim, TIF_MEMDIE); |
d2d393099
|
521 |
do_send_sig_info(SIGKILL, SEND_SIG_FORCED, victim, true); |
1da177e4c
|
522 |
} |
647f2bdf4
|
523 |
#undef K |
1da177e4c
|
524 |
|
309ed8825
|
525 526 527 528 |
/* * Determines whether the kernel must panic because of the panic_on_oom sysctl. */ static void check_panic_on_oom(enum oom_constraint constraint, gfp_t gfp_mask, |
e85bfd3aa
|
529 |
int order, const nodemask_t *nodemask) |
309ed8825
|
530 531 532 533 534 535 536 537 538 539 540 541 542 |
{ if (likely(!sysctl_panic_on_oom)) return; if (sysctl_panic_on_oom != 2) { /* * panic_on_oom == 1 only affects CONSTRAINT_NONE, the kernel * does not panic for cpuset, mempolicy, or memcg allocation * failures. */ if (constraint != CONSTRAINT_NONE) return; } read_lock(&tasklist_lock); |
e85bfd3aa
|
543 |
dump_header(NULL, gfp_mask, order, NULL, nodemask); |
309ed8825
|
544 545 546 547 548 |
read_unlock(&tasklist_lock); panic("Out of memory: %s panic_on_oom is enabled ", sysctl_panic_on_oom == 2 ? "compulsory" : "system-wide"); } |
00f0b8259
|
549 |
#ifdef CONFIG_CGROUP_MEM_RES_CTLR |
e845e1993
|
550 551 |
void mem_cgroup_out_of_memory(struct mem_cgroup *memcg, gfp_t gfp_mask, int order) |
c7ba5c9e8
|
552 |
{ |
a63d83f42
|
553 554 |
unsigned long limit; unsigned int points = 0; |
c7ba5c9e8
|
555 |
struct task_struct *p; |
f9434ad15
|
556 557 558 559 560 561 562 |
/* * If current has a pending SIGKILL, then automatically select it. The * goal is to allow it to allocate so that it may quickly exit and free * its memory. */ if (fatal_signal_pending(current)) { set_thread_flag(TIF_MEMDIE); |
f9434ad15
|
563 564 |
return; } |
e845e1993
|
565 |
check_panic_on_oom(CONSTRAINT_MEMCG, gfp_mask, order, NULL); |
72835c86c
|
566 |
limit = mem_cgroup_get_limit(memcg) >> PAGE_SHIFT; |
e115f2d89
|
567 |
read_lock(&tasklist_lock); |
08ab9b10d
|
568 |
p = select_bad_process(&points, limit, memcg, NULL, false); |
2a1c9b1fc
|
569 |
if (p && PTR_ERR(p) != -1UL) |
e845e1993
|
570 |
oom_kill_process(p, gfp_mask, order, points, limit, memcg, NULL, |
2a1c9b1fc
|
571 |
"Memory cgroup out of memory"); |
e115f2d89
|
572 |
read_unlock(&tasklist_lock); |
c7ba5c9e8
|
573 574 |
} #endif |
8bc719d3c
|
575 576 577 578 579 580 581 582 583 584 585 586 587 |
static BLOCKING_NOTIFIER_HEAD(oom_notify_list); int register_oom_notifier(struct notifier_block *nb) { return blocking_notifier_chain_register(&oom_notify_list, nb); } EXPORT_SYMBOL_GPL(register_oom_notifier); int unregister_oom_notifier(struct notifier_block *nb) { return blocking_notifier_chain_unregister(&oom_notify_list, nb); } EXPORT_SYMBOL_GPL(unregister_oom_notifier); |
098d7f128
|
588 589 590 591 592 |
/* * Try to acquire the OOM killer lock for the zones in zonelist. Returns zero * if a parallel OOM killing is already taking place that includes a zone in * the zonelist. Otherwise, locks all zones in the zonelist and returns 1. */ |
ff321feac
|
593 |
int try_set_zonelist_oom(struct zonelist *zonelist, gfp_t gfp_mask) |
098d7f128
|
594 |
{ |
dd1a239f6
|
595 596 |
struct zoneref *z; struct zone *zone; |
098d7f128
|
597 |
int ret = 1; |
c7d4caeb1
|
598 |
spin_lock(&zone_scan_lock); |
dd1a239f6
|
599 600 |
for_each_zone_zonelist(zone, z, zonelist, gfp_zone(gfp_mask)) { if (zone_is_oom_locked(zone)) { |
098d7f128
|
601 602 603 |
ret = 0; goto out; } |
dd1a239f6
|
604 605 606 607 |
} for_each_zone_zonelist(zone, z, zonelist, gfp_zone(gfp_mask)) { /* |
c7d4caeb1
|
608 |
* Lock each zone in the zonelist under zone_scan_lock so a |
ff321feac
|
609 |
* parallel invocation of try_set_zonelist_oom() doesn't succeed |
dd1a239f6
|
610 611 612 613 |
* when it shouldn't. */ zone_set_flag(zone, ZONE_OOM_LOCKED); } |
098d7f128
|
614 |
|
098d7f128
|
615 |
out: |
c7d4caeb1
|
616 |
spin_unlock(&zone_scan_lock); |
098d7f128
|
617 618 619 620 621 622 623 624 |
return ret; } /* * Clears the ZONE_OOM_LOCKED flag for all zones in the zonelist so that failed * allocation attempts with zonelists containing them may now recall the OOM * killer, if necessary. */ |
dd1a239f6
|
625 |
void clear_zonelist_oom(struct zonelist *zonelist, gfp_t gfp_mask) |
098d7f128
|
626 |
{ |
dd1a239f6
|
627 628 |
struct zoneref *z; struct zone *zone; |
098d7f128
|
629 |
|
c7d4caeb1
|
630 |
spin_lock(&zone_scan_lock); |
dd1a239f6
|
631 632 633 |
for_each_zone_zonelist(zone, z, zonelist, gfp_zone(gfp_mask)) { zone_clear_flag(zone, ZONE_OOM_LOCKED); } |
c7d4caeb1
|
634 |
spin_unlock(&zone_scan_lock); |
098d7f128
|
635 |
} |
1c0fe6e3b
|
636 |
/* |
e36589323
|
637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 |
* Try to acquire the oom killer lock for all system zones. Returns zero if a * parallel oom killing is taking place, otherwise locks all zones and returns * non-zero. */ static int try_set_system_oom(void) { struct zone *zone; int ret = 1; spin_lock(&zone_scan_lock); for_each_populated_zone(zone) if (zone_is_oom_locked(zone)) { ret = 0; goto out; } for_each_populated_zone(zone) zone_set_flag(zone, ZONE_OOM_LOCKED); out: spin_unlock(&zone_scan_lock); return ret; } /* * Clears ZONE_OOM_LOCKED for all system zones so that failed allocation * attempts or page faults may now recall the oom killer, if necessary. */ static void clear_system_oom(void) { struct zone *zone; spin_lock(&zone_scan_lock); for_each_populated_zone(zone) zone_clear_flag(zone, ZONE_OOM_LOCKED); spin_unlock(&zone_scan_lock); } |
1da177e4c
|
672 |
/** |
6937a25cf
|
673 |
* out_of_memory - kill the "best" process when we run out of memory |
1b578df02
|
674 675 676 |
* @zonelist: zonelist pointer * @gfp_mask: memory allocation flags * @order: amount of memory being requested as a power of 2 |
6f48d0ebd
|
677 |
* @nodemask: nodemask passed to page allocator |
08ab9b10d
|
678 |
* @force_kill: true if a task must be killed, even if others are exiting |
1da177e4c
|
679 680 681 682 683 684 |
* * If we run out of memory, we have the choice between either * killing a random task (bad), letting the system crash (worse) * OR try to be smart about which process to kill. Note that we * don't have to be perfect here, we just have to be good. */ |
4365a5676
|
685 |
void out_of_memory(struct zonelist *zonelist, gfp_t gfp_mask, |
08ab9b10d
|
686 |
int order, nodemask_t *nodemask, bool force_kill) |
1da177e4c
|
687 |
{ |
e85bfd3aa
|
688 |
const nodemask_t *mpol_mask; |
0aad4b312
|
689 |
struct task_struct *p; |
a63d83f42
|
690 |
unsigned long totalpages; |
8bc719d3c
|
691 |
unsigned long freed = 0; |
a63d83f42
|
692 |
unsigned int points; |
e36589323
|
693 |
enum oom_constraint constraint = CONSTRAINT_NONE; |
b52723c56
|
694 |
int killed = 0; |
8bc719d3c
|
695 696 697 698 699 |
blocking_notifier_call_chain(&oom_notify_list, 0, &freed); if (freed > 0) /* Got some memory back in the last second. */ return; |
1da177e4c
|
700 |
|
7b98c2e40
|
701 702 703 704 705 706 707 708 709 |
/* * If current has a pending SIGKILL, then automatically select it. The * goal is to allow it to allocate so that it may quickly exit and free * its memory. */ if (fatal_signal_pending(current)) { set_thread_flag(TIF_MEMDIE); return; } |
9b0f8b040
|
710 711 712 713 |
/* * Check if there were limitations on the allocation (only relevant for * NUMA) that may require different handling. */ |
a63d83f42
|
714 715 |
constraint = constrained_alloc(zonelist, gfp_mask, nodemask, &totalpages); |
e85bfd3aa
|
716 717 |
mpol_mask = (constraint == CONSTRAINT_MEMORY_POLICY) ? nodemask : NULL; check_panic_on_oom(constraint, gfp_mask, order, mpol_mask); |
0aad4b312
|
718 |
|
2b45ab339
|
719 |
read_lock(&tasklist_lock); |
f88ccad58
|
720 |
if (sysctl_oom_kill_allocating_task && |
a96cfd6e9
|
721 |
!oom_unkillable_task(current, NULL, nodemask) && |
c9f01245b
|
722 |
current->mm) { |
2a1c9b1fc
|
723 724 725 726 |
oom_kill_process(current, gfp_mask, order, 0, totalpages, NULL, nodemask, "Out of memory (oom_kill_allocating_task)"); goto out; |
0aad4b312
|
727 |
} |
08ab9b10d
|
728 729 |
p = select_bad_process(&points, totalpages, NULL, mpol_mask, force_kill); |
0aad4b312
|
730 731 |
/* Found nothing?!?! Either we hang forever, or we panic. */ if (!p) { |
e85bfd3aa
|
732 |
dump_header(NULL, gfp_mask, order, NULL, mpol_mask); |
0aad4b312
|
733 734 735 736 |
read_unlock(&tasklist_lock); panic("Out of memory and no killable processes... "); } |
2a1c9b1fc
|
737 738 739 740 741 |
if (PTR_ERR(p) != -1UL) { oom_kill_process(p, gfp_mask, order, points, totalpages, NULL, nodemask, "Out of memory"); killed = 1; } |
b52723c56
|
742 |
out: |
140ffcec4
|
743 |
read_unlock(&tasklist_lock); |
1da177e4c
|
744 745 746 |
/* * Give "p" a good chance of killing itself before we |
2f659f462
|
747 |
* retry to allocate memory unless "p" is current |
1da177e4c
|
748 |
*/ |
b52723c56
|
749 |
if (killed && !test_thread_flag(TIF_MEMDIE)) |
140ffcec4
|
750 |
schedule_timeout_uninterruptible(1); |
1da177e4c
|
751 |
} |
e36589323
|
752 753 754 755 756 757 758 759 760 761 |
/* * The pagefault handler calls here because it is out of memory, so kill a * memory-hogging task. If a populated zone has ZONE_OOM_LOCKED set, a parallel * oom killing is already in progress so do nothing. If a task is found with * TIF_MEMDIE set, it has been killed so do nothing and allow it to exit. */ void pagefault_out_of_memory(void) { if (try_set_system_oom()) { |
08ab9b10d
|
762 |
out_of_memory(NULL, 0, 0, NULL, false); |
e36589323
|
763 764 765 766 767 |
clear_system_oom(); } if (!test_thread_flag(TIF_MEMDIE)) schedule_timeout_uninterruptible(1); } |