Blame view
mm/vmscan.c
109 KB
1da177e4c
|
1 2 3 4 5 6 7 8 9 10 11 12 |
/* * linux/mm/vmscan.c * * Copyright (C) 1991, 1992, 1993, 1994 Linus Torvalds * * Swap reorganised 29.12.95, Stephen Tweedie. * kswapd added: 7.1.96 sct * Removed kswapd_ctl limits, and swap out as many pages as needed * to bring the system back to freepages.high: 2.4.97, Rik van Riel. * Zone aware kswapd started 02/00, Kanoj Sarcar (kanoj@sgi.com). * Multiqueue VM started 5.8.00, Rik van Riel. */ |
b1de0d139
|
13 |
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt |
1da177e4c
|
14 15 |
#include <linux/mm.h> #include <linux/module.h> |
5a0e3ad6a
|
16 |
#include <linux/gfp.h> |
1da177e4c
|
17 18 19 20 21 |
#include <linux/kernel_stat.h> #include <linux/swap.h> #include <linux/pagemap.h> #include <linux/init.h> #include <linux/highmem.h> |
70ddf637e
|
22 |
#include <linux/vmpressure.h> |
e129b5c23
|
23 |
#include <linux/vmstat.h> |
1da177e4c
|
24 25 26 27 28 29 |
#include <linux/file.h> #include <linux/writeback.h> #include <linux/blkdev.h> #include <linux/buffer_head.h> /* for try_to_release_page(), buffer_heads_over_limit */ #include <linux/mm_inline.h> |
1da177e4c
|
30 31 32 33 34 |
#include <linux/backing-dev.h> #include <linux/rmap.h> #include <linux/topology.h> #include <linux/cpu.h> #include <linux/cpuset.h> |
3e7d34497
|
35 |
#include <linux/compaction.h> |
1da177e4c
|
36 37 |
#include <linux/notifier.h> #include <linux/rwsem.h> |
248a0301e
|
38 |
#include <linux/delay.h> |
3218ae14b
|
39 |
#include <linux/kthread.h> |
7dfb71030
|
40 |
#include <linux/freezer.h> |
66e1707bc
|
41 |
#include <linux/memcontrol.h> |
873b47717
|
42 |
#include <linux/delayacct.h> |
af936a160
|
43 |
#include <linux/sysctl.h> |
929bea7c7
|
44 |
#include <linux/oom.h> |
268bb0ce3
|
45 |
#include <linux/prefetch.h> |
b1de0d139
|
46 |
#include <linux/printk.h> |
1da177e4c
|
47 48 49 50 51 |
#include <asm/tlbflush.h> #include <asm/div64.h> #include <linux/swapops.h> |
117aad1e9
|
52 |
#include <linux/balloon_compaction.h> |
1da177e4c
|
53 |
|
0f8053a50
|
54 |
#include "internal.h" |
33906bc5c
|
55 56 |
#define CREATE_TRACE_POINTS #include <trace/events/vmscan.h> |
1da177e4c
|
57 |
struct scan_control { |
22fba3354
|
58 59 |
/* How many pages shrink_list() should reclaim */ unsigned long nr_to_reclaim; |
1da177e4c
|
60 |
/* This context's GFP mask */ |
6daa0e286
|
61 |
gfp_t gfp_mask; |
1da177e4c
|
62 |
|
ee814fe23
|
63 |
/* Allocation order */ |
5ad333eb6
|
64 |
int order; |
66e1707bc
|
65 |
|
ee814fe23
|
66 67 68 69 70 |
/* * Nodemask of nodes allowed by the caller. If NULL, all nodes * are scanned. */ nodemask_t *nodemask; |
9e3b2f8cd
|
71 |
|
5f53e7629
|
72 |
/* |
f16015fbf
|
73 74 75 76 |
* The memory cgroup that hit its limit and as a result is the * primary target of this reclaim invocation. */ struct mem_cgroup *target_mem_cgroup; |
66e1707bc
|
77 |
|
ee814fe23
|
78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 |
/* Scan (total_size >> priority) pages at once */ int priority; unsigned int may_writepage:1; /* Can mapped pages be reclaimed? */ unsigned int may_unmap:1; /* Can pages be swapped as part of reclaim? */ unsigned int may_swap:1; unsigned int hibernation_mode:1; /* One of the zones is ready for compaction */ unsigned int compaction_ready:1; /* Incremented by the number of inactive pages that were scanned */ unsigned long nr_scanned; /* Number of pages freed so far during a call to shrink_zones() */ unsigned long nr_reclaimed; |
1da177e4c
|
99 |
}; |
1da177e4c
|
100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 |
#define lru_to_page(_head) (list_entry((_head)->prev, struct page, lru)) #ifdef ARCH_HAS_PREFETCH #define prefetch_prev_lru_page(_page, _base, _field) \ do { \ if ((_page)->lru.prev != _base) { \ struct page *prev; \ \ prev = lru_to_page(&(_page->lru)); \ prefetch(&prev->_field); \ } \ } while (0) #else #define prefetch_prev_lru_page(_page, _base, _field) do { } while (0) #endif #ifdef ARCH_HAS_PREFETCHW #define prefetchw_prev_lru_page(_page, _base, _field) \ do { \ if ((_page)->lru.prev != _base) { \ struct page *prev; \ \ prev = lru_to_page(&(_page->lru)); \ prefetchw(&prev->_field); \ } \ } while (0) #else #define prefetchw_prev_lru_page(_page, _base, _field) do { } while (0) #endif /* * From 0 .. 100. Higher means more swappy. */ int vm_swappiness = 60; |
d0480be44
|
134 135 136 137 138 |
/* * The total number of pages which are beyond the high watermark within all * zones. */ unsigned long vm_total_pages; |
1da177e4c
|
139 140 141 |
static LIST_HEAD(shrinker_list); static DECLARE_RWSEM(shrinker_rwsem); |
c255a4580
|
142 |
#ifdef CONFIG_MEMCG |
89b5fae53
|
143 144 |
static bool global_reclaim(struct scan_control *sc) { |
f16015fbf
|
145 |
return !sc->target_mem_cgroup; |
89b5fae53
|
146 |
} |
91a45470f
|
147 |
#else |
89b5fae53
|
148 149 150 151 |
static bool global_reclaim(struct scan_control *sc) { return true; } |
91a45470f
|
152 |
#endif |
a1c3bfb2f
|
153 |
static unsigned long zone_reclaimable_pages(struct zone *zone) |
6e543d578
|
154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 |
{ int nr; nr = zone_page_state(zone, NR_ACTIVE_FILE) + zone_page_state(zone, NR_INACTIVE_FILE); if (get_nr_swap_pages() > 0) nr += zone_page_state(zone, NR_ACTIVE_ANON) + zone_page_state(zone, NR_INACTIVE_ANON); return nr; } bool zone_reclaimable(struct zone *zone) { |
0d5d823ab
|
169 170 |
return zone_page_state(zone, NR_PAGES_SCANNED) < zone_reclaimable_pages(zone) * 6; |
6e543d578
|
171 |
} |
4d7dcca21
|
172 |
static unsigned long get_lru_size(struct lruvec *lruvec, enum lru_list lru) |
c9f299d98
|
173 |
{ |
c3c787e8c
|
174 |
if (!mem_cgroup_disabled()) |
4d7dcca21
|
175 |
return mem_cgroup_get_lru_size(lruvec, lru); |
a3d8e0549
|
176 |
|
074291fea
|
177 |
return zone_page_state(lruvec_zone(lruvec), NR_LRU_BASE + lru); |
c9f299d98
|
178 |
} |
1da177e4c
|
179 |
/* |
1d3d4437e
|
180 |
* Add a shrinker callback to be called from the vm. |
1da177e4c
|
181 |
*/ |
1d3d4437e
|
182 |
int register_shrinker(struct shrinker *shrinker) |
1da177e4c
|
183 |
{ |
1d3d4437e
|
184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 |
size_t size = sizeof(*shrinker->nr_deferred); /* * If we only have one possible node in the system anyway, save * ourselves the trouble and disable NUMA aware behavior. This way we * will save memory and some small loop time later. */ if (nr_node_ids == 1) shrinker->flags &= ~SHRINKER_NUMA_AWARE; if (shrinker->flags & SHRINKER_NUMA_AWARE) size *= nr_node_ids; shrinker->nr_deferred = kzalloc(size, GFP_KERNEL); if (!shrinker->nr_deferred) return -ENOMEM; |
8e1f936b7
|
200 201 202 |
down_write(&shrinker_rwsem); list_add_tail(&shrinker->list, &shrinker_list); up_write(&shrinker_rwsem); |
1d3d4437e
|
203 |
return 0; |
1da177e4c
|
204 |
} |
8e1f936b7
|
205 |
EXPORT_SYMBOL(register_shrinker); |
1da177e4c
|
206 207 208 209 |
/* * Remove one */ |
8e1f936b7
|
210 |
void unregister_shrinker(struct shrinker *shrinker) |
1da177e4c
|
211 212 213 214 |
{ down_write(&shrinker_rwsem); list_del(&shrinker->list); up_write(&shrinker_rwsem); |
ae3933216
|
215 |
kfree(shrinker->nr_deferred); |
1da177e4c
|
216 |
} |
8e1f936b7
|
217 |
EXPORT_SYMBOL(unregister_shrinker); |
1da177e4c
|
218 219 |
#define SHRINK_BATCH 128 |
1d3d4437e
|
220 221 222 223 224 225 226 227 |
static unsigned long shrink_slab_node(struct shrink_control *shrinkctl, struct shrinker *shrinker, unsigned long nr_pages_scanned, unsigned long lru_pages) { unsigned long freed = 0; unsigned long long delta; long total_scan; |
d5bc5fd3f
|
228 |
long freeable; |
1d3d4437e
|
229 230 231 232 233 |
long nr; long new_nr; int nid = shrinkctl->nid; long batch_size = shrinker->batch ? shrinker->batch : SHRINK_BATCH; |
d5bc5fd3f
|
234 235 |
freeable = shrinker->count_objects(shrinker, shrinkctl); if (freeable == 0) |
1d3d4437e
|
236 237 238 239 240 241 242 243 244 245 246 |
return 0; /* * copy the current shrinker scan count into a local variable * and zero it so that other concurrent shrinker invocations * don't also do this scanning work. */ nr = atomic_long_xchg(&shrinker->nr_deferred[nid], 0); total_scan = nr; delta = (4 * nr_pages_scanned) / shrinker->seeks; |
d5bc5fd3f
|
247 |
delta *= freeable; |
1d3d4437e
|
248 249 250 251 252 253 |
do_div(delta, lru_pages + 1); total_scan += delta; if (total_scan < 0) { printk(KERN_ERR "shrink_slab: %pF negative objects to delete nr=%ld ", |
a0b02131c
|
254 |
shrinker->scan_objects, total_scan); |
d5bc5fd3f
|
255 |
total_scan = freeable; |
1d3d4437e
|
256 257 258 259 260 261 262 263 |
} /* * We need to avoid excessive windup on filesystem shrinkers * due to large numbers of GFP_NOFS allocations causing the * shrinkers to return -1 all the time. This results in a large * nr being built up so when a shrink that can do some work * comes along it empties the entire cache due to nr >>> |
d5bc5fd3f
|
264 |
* freeable. This is bad for sustaining a working set in |
1d3d4437e
|
265 266 267 268 269 |
* memory. * * Hence only allow the shrinker to scan the entire cache when * a large delta change is calculated directly. */ |
d5bc5fd3f
|
270 271 |
if (delta < freeable / 4) total_scan = min(total_scan, freeable / 2); |
1d3d4437e
|
272 273 274 275 276 277 |
/* * Avoid risking looping forever due to too large nr value: * never try to free more than twice the estimate number of * freeable entries. */ |
d5bc5fd3f
|
278 279 |
if (total_scan > freeable * 2) total_scan = freeable * 2; |
1d3d4437e
|
280 281 282 |
trace_mm_shrink_slab_start(shrinker, shrinkctl, nr, nr_pages_scanned, lru_pages, |
d5bc5fd3f
|
283 |
freeable, delta, total_scan); |
1d3d4437e
|
284 |
|
0b1fb40a3
|
285 286 287 288 289 290 291 292 293 294 295 |
/* * Normally, we should not scan less than batch_size objects in one * pass to avoid too frequent shrinker calls, but if the slab has less * than batch_size objects in total and we are really tight on memory, * we will try to reclaim all available objects, otherwise we can end * up failing allocations although there are plenty of reclaimable * objects spread over several slabs with usage less than the * batch_size. * * We detect the "tight on memory" situations by looking at the total * number of objects we want to scan (total_scan). If it is greater |
d5bc5fd3f
|
296 |
* than the total number of objects on slab (freeable), we must be |
0b1fb40a3
|
297 298 299 300 |
* scanning at high prio and therefore should try to reclaim as much as * possible. */ while (total_scan >= batch_size || |
d5bc5fd3f
|
301 |
total_scan >= freeable) { |
a0b02131c
|
302 |
unsigned long ret; |
0b1fb40a3
|
303 |
unsigned long nr_to_scan = min(batch_size, total_scan); |
1d3d4437e
|
304 |
|
0b1fb40a3
|
305 |
shrinkctl->nr_to_scan = nr_to_scan; |
a0b02131c
|
306 307 308 309 |
ret = shrinker->scan_objects(shrinker, shrinkctl); if (ret == SHRINK_STOP) break; freed += ret; |
1d3d4437e
|
310 |
|
0b1fb40a3
|
311 312 |
count_vm_events(SLABS_SCANNED, nr_to_scan); total_scan -= nr_to_scan; |
1d3d4437e
|
313 314 315 316 317 318 319 320 321 322 323 324 325 326 |
cond_resched(); } /* * move the unused scan count back into the shrinker in a * manner that handles concurrent updates. If we exhausted the * scan, there is no need to do an update. */ if (total_scan > 0) new_nr = atomic_long_add_return(total_scan, &shrinker->nr_deferred[nid]); else new_nr = atomic_long_read(&shrinker->nr_deferred[nid]); |
df9024a8c
|
327 |
trace_mm_shrink_slab_end(shrinker, nid, freed, nr, new_nr, total_scan); |
1d3d4437e
|
328 |
return freed; |
1495f230f
|
329 |
} |
1da177e4c
|
330 331 332 333 334 335 336 337 |
/* * Call the shrink functions to age shrinkable caches * * Here we assume it costs one seek to replace a lru page and that it also * takes a seek to recreate a cache object. With this in mind we age equal * percentages of the lru and ageable caches. This should balance the seeks * generated by these structures. * |
183ff22bb
|
338 |
* If the vm encountered mapped pages on the LRU it increase the pressure on |
1da177e4c
|
339 340 341 342 343 344 345 |
* slab to avoid swapping. * * We do weird things to avoid (scanned*seeks*entries) overflowing 32 bits. * * `lru_pages' represents the number of on-LRU pages in all the zones which * are eligible for the caller's allocation attempt. It is used for balancing * slab reclaim versus page reclaim. |
b15e0905f
|
346 347 |
* * Returns the number of slab objects which we shrunk. |
1da177e4c
|
348 |
*/ |
24f7c6b98
|
349 |
unsigned long shrink_slab(struct shrink_control *shrinkctl, |
1495f230f
|
350 |
unsigned long nr_pages_scanned, |
a09ed5e00
|
351 |
unsigned long lru_pages) |
1da177e4c
|
352 353 |
{ struct shrinker *shrinker; |
24f7c6b98
|
354 |
unsigned long freed = 0; |
1da177e4c
|
355 |
|
1495f230f
|
356 357 |
if (nr_pages_scanned == 0) nr_pages_scanned = SWAP_CLUSTER_MAX; |
1da177e4c
|
358 |
|
f06590bd7
|
359 |
if (!down_read_trylock(&shrinker_rwsem)) { |
24f7c6b98
|
360 361 362 363 364 365 366 |
/* * If we would return 0, our callers would understand that we * have nothing else to shrink and give up trying. By returning * 1 we keep it going and assume we'll be able to shrink next * time. */ freed = 1; |
f06590bd7
|
367 368 |
goto out; } |
1da177e4c
|
369 370 |
list_for_each_entry(shrinker, &shrinker_list, list) { |
ec97097bc
|
371 372 |
if (!(shrinker->flags & SHRINKER_NUMA_AWARE)) { shrinkctl->nid = 0; |
1d3d4437e
|
373 |
freed += shrink_slab_node(shrinkctl, shrinker, |
ec97097bc
|
374 375 376 377 378 379 380 381 |
nr_pages_scanned, lru_pages); continue; } for_each_node_mask(shrinkctl->nid, shrinkctl->nodes_to_scan) { if (node_online(shrinkctl->nid)) freed += shrink_slab_node(shrinkctl, shrinker, nr_pages_scanned, lru_pages); |
1da177e4c
|
382 |
|
1da177e4c
|
383 |
} |
1da177e4c
|
384 385 |
} up_read(&shrinker_rwsem); |
f06590bd7
|
386 387 |
out: cond_resched(); |
24f7c6b98
|
388 |
return freed; |
1da177e4c
|
389 |
} |
1da177e4c
|
390 391 |
static inline int is_page_cache_freeable(struct page *page) { |
ceddc3a52
|
392 393 394 395 396 |
/* * A freeable page cache page is referenced only by the caller * that isolated the page, the page cache radix tree and * optional buffer heads at page->private. */ |
edcf4748c
|
397 |
return page_count(page) - page_has_private(page) == 2; |
1da177e4c
|
398 |
} |
7d3579e8e
|
399 400 |
static int may_write_to_queue(struct backing_dev_info *bdi, struct scan_control *sc) |
1da177e4c
|
401 |
{ |
930d91525
|
402 |
if (current->flags & PF_SWAPWRITE) |
1da177e4c
|
403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 |
return 1; if (!bdi_write_congested(bdi)) return 1; if (bdi == current->backing_dev_info) return 1; return 0; } /* * We detected a synchronous write error writing a page out. Probably * -ENOSPC. We need to propagate that into the address_space for a subsequent * fsync(), msync() or close(). * * The tricky part is that after writepage we cannot touch the mapping: nothing * prevents it from being freed up. But we have a ref on the page and once * that page is locked, the mapping is pinned. * * We're allowed to run sleeping lock_page() here because we know the caller has * __GFP_FS. */ static void handle_write_error(struct address_space *mapping, struct page *page, int error) { |
7eaceacca
|
426 |
lock_page(page); |
3e9f45bd1
|
427 428 |
if (page_mapping(page) == mapping) mapping_set_error(mapping, error); |
1da177e4c
|
429 430 |
unlock_page(page); } |
04e62a29b
|
431 432 433 434 435 436 437 438 439 440 441 |
/* possible outcome of pageout() */ typedef enum { /* failed to write page out, page is locked */ PAGE_KEEP, /* move page to the active list, page is locked */ PAGE_ACTIVATE, /* page has been sent to the disk successfully, page is unlocked */ PAGE_SUCCESS, /* page is clean and locked */ PAGE_CLEAN, } pageout_t; |
1da177e4c
|
442 |
/* |
1742f19fa
|
443 444 |
* pageout is called by shrink_page_list() for each dirty page. * Calls ->writepage(). |
1da177e4c
|
445 |
*/ |
c661b078f
|
446 |
static pageout_t pageout(struct page *page, struct address_space *mapping, |
7d3579e8e
|
447 |
struct scan_control *sc) |
1da177e4c
|
448 449 450 451 452 453 454 455 |
{ /* * If the page is dirty, only perform writeback if that write * will be non-blocking. To prevent this allocation from being * stalled by pagecache activity. But note that there may be * stalls if we need to run get_block(). We could test * PagePrivate for that. * |
8174202b3
|
456 |
* If this process is currently in __generic_file_write_iter() against |
1da177e4c
|
457 458 459 460 461 462 463 |
* this page's queue, we can perform writeback even if that * will block. * * If the page is swapcache, write it back even if that would * block, for some throttling. This happens by accident, because * swap_backing_dev_info is bust: it doesn't reflect the * congestion state of the swapdevs. Easy to fix, if needed. |
1da177e4c
|
464 465 466 467 468 469 470 471 |
*/ if (!is_page_cache_freeable(page)) return PAGE_KEEP; if (!mapping) { /* * Some data journaling orphaned pages can have * page->mapping == NULL while being dirty with clean buffers. */ |
266cf658e
|
472 |
if (page_has_private(page)) { |
1da177e4c
|
473 474 |
if (try_to_free_buffers(page)) { ClearPageDirty(page); |
b1de0d139
|
475 476 |
pr_info("%s: orphaned page ", __func__); |
1da177e4c
|
477 478 479 480 481 482 483 |
return PAGE_CLEAN; } } return PAGE_KEEP; } if (mapping->a_ops->writepage == NULL) return PAGE_ACTIVATE; |
0e093d997
|
484 |
if (!may_write_to_queue(mapping->backing_dev_info, sc)) |
1da177e4c
|
485 486 487 488 489 490 491 |
return PAGE_KEEP; if (clear_page_dirty_for_io(page)) { int res; struct writeback_control wbc = { .sync_mode = WB_SYNC_NONE, .nr_to_write = SWAP_CLUSTER_MAX, |
111ebb6e6
|
492 493 |
.range_start = 0, .range_end = LLONG_MAX, |
1da177e4c
|
494 495 496 497 498 499 500 |
.for_reclaim = 1, }; SetPageReclaim(page); res = mapping->a_ops->writepage(page, &wbc); if (res < 0) handle_write_error(mapping, page, res); |
994fc28c7
|
501 |
if (res == AOP_WRITEPAGE_ACTIVATE) { |
1da177e4c
|
502 503 504 |
ClearPageReclaim(page); return PAGE_ACTIVATE; } |
c661b078f
|
505 |
|
1da177e4c
|
506 507 508 509 |
if (!PageWriteback(page)) { /* synchronous write or broken a_ops? */ ClearPageReclaim(page); } |
23b9da55c
|
510 |
trace_mm_vmscan_writepage(page, trace_reclaim_flags(page)); |
e129b5c23
|
511 |
inc_zone_page_state(page, NR_VMSCAN_WRITE); |
1da177e4c
|
512 513 514 515 516 |
return PAGE_SUCCESS; } return PAGE_CLEAN; } |
a649fd927
|
517 |
/* |
e286781d5
|
518 519 |
* Same as remove_mapping, but if the page is removed from the mapping, it * gets returned with a refcount of 0. |
a649fd927
|
520 |
*/ |
a528910e1
|
521 522 |
static int __remove_mapping(struct address_space *mapping, struct page *page, bool reclaimed) |
49d2e9cc4
|
523 |
{ |
28e4d965e
|
524 525 |
BUG_ON(!PageLocked(page)); BUG_ON(mapping != page_mapping(page)); |
49d2e9cc4
|
526 |
|
19fd62312
|
527 |
spin_lock_irq(&mapping->tree_lock); |
49d2e9cc4
|
528 |
/* |
0fd0e6b05
|
529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 |
* The non racy check for a busy page. * * Must be careful with the order of the tests. When someone has * a ref to the page, it may be possible that they dirty it then * drop the reference. So if PageDirty is tested before page_count * here, then the following race may occur: * * get_user_pages(&page); * [user mapping goes away] * write_to(page); * !PageDirty(page) [good] * SetPageDirty(page); * put_page(page); * !page_count(page) [good, discard it] * * [oops, our write_to data is lost] * * Reversing the order of the tests ensures such a situation cannot * escape unnoticed. The smp_rmb is needed to ensure the page->flags * load is not satisfied before that of page->_count. * * Note that if SetPageDirty is always performed via set_page_dirty, * and thus under tree_lock, then this ordering is not required. |
49d2e9cc4
|
552 |
*/ |
e286781d5
|
553 |
if (!page_freeze_refs(page, 2)) |
49d2e9cc4
|
554 |
goto cannot_free; |
e286781d5
|
555 556 557 |
/* note: atomic_cmpxchg in page_freeze_refs provides the smp_rmb */ if (unlikely(PageDirty(page))) { page_unfreeze_refs(page, 2); |
49d2e9cc4
|
558 |
goto cannot_free; |
e286781d5
|
559 |
} |
49d2e9cc4
|
560 561 562 |
if (PageSwapCache(page)) { swp_entry_t swap = { .val = page_private(page) }; |
0a31bc97c
|
563 |
mem_cgroup_swapout(page, swap); |
49d2e9cc4
|
564 |
__delete_from_swap_cache(page); |
19fd62312
|
565 |
spin_unlock_irq(&mapping->tree_lock); |
0a31bc97c
|
566 |
swapcache_free(swap); |
e286781d5
|
567 |
} else { |
6072d13c4
|
568 |
void (*freepage)(struct page *); |
a528910e1
|
569 |
void *shadow = NULL; |
6072d13c4
|
570 571 |
freepage = mapping->a_ops->freepage; |
a528910e1
|
572 573 574 575 576 577 578 579 580 581 582 583 584 585 |
/* * Remember a shadow entry for reclaimed file cache in * order to detect refaults, thus thrashing, later on. * * But don't store shadows in an address space that is * already exiting. This is not just an optizimation, * inode reclaim needs to empty out the radix tree or * the nodes are lost. Don't plant shadows behind its * back. */ if (reclaimed && page_is_file_cache(page) && !mapping_exiting(mapping)) shadow = workingset_eviction(mapping, page); __delete_from_page_cache(page, shadow); |
19fd62312
|
586 |
spin_unlock_irq(&mapping->tree_lock); |
6072d13c4
|
587 588 589 |
if (freepage != NULL) freepage(page); |
49d2e9cc4
|
590 |
} |
49d2e9cc4
|
591 592 593 |
return 1; cannot_free: |
19fd62312
|
594 |
spin_unlock_irq(&mapping->tree_lock); |
49d2e9cc4
|
595 596 |
return 0; } |
1da177e4c
|
597 |
/* |
e286781d5
|
598 599 600 601 602 603 604 |
* Attempt to detach a locked page from its ->mapping. If it is dirty or if * someone else has a ref on the page, abort and return 0. If it was * successfully detached, return 1. Assumes the caller has a single ref on * this page. */ int remove_mapping(struct address_space *mapping, struct page *page) { |
a528910e1
|
605 |
if (__remove_mapping(mapping, page, false)) { |
e286781d5
|
606 607 608 609 610 611 612 613 614 615 |
/* * Unfreezing the refcount with 1 rather than 2 effectively * drops the pagecache ref for us without requiring another * atomic operation. */ page_unfreeze_refs(page, 1); return 1; } return 0; } |
894bc3104
|
616 617 618 619 620 621 622 623 624 |
/** * putback_lru_page - put previously isolated page onto appropriate LRU list * @page: page to be put back to appropriate lru list * * Add previously isolated @page to appropriate LRU list. * Page may still be unevictable for other reasons. * * lru_lock must not be held, interrupts must be enabled. */ |
894bc3104
|
625 626 |
void putback_lru_page(struct page *page) { |
0ec3b74c7
|
627 |
bool is_unevictable; |
bbfd28eee
|
628 |
int was_unevictable = PageUnevictable(page); |
894bc3104
|
629 |
|
309381fea
|
630 |
VM_BUG_ON_PAGE(PageLRU(page), page); |
894bc3104
|
631 632 633 |
redo: ClearPageUnevictable(page); |
39b5f29ac
|
634 |
if (page_evictable(page)) { |
894bc3104
|
635 636 637 638 639 640 |
/* * For evictable pages, we can use the cache. * In event of a race, worst case is we end up with an * unevictable page on [in]active list. * We know how to handle that. */ |
0ec3b74c7
|
641 |
is_unevictable = false; |
c53954a09
|
642 |
lru_cache_add(page); |
894bc3104
|
643 644 645 646 647 |
} else { /* * Put unevictable pages directly on zone's unevictable * list. */ |
0ec3b74c7
|
648 |
is_unevictable = true; |
894bc3104
|
649 |
add_page_to_unevictable_list(page); |
6a7b95481
|
650 |
/* |
21ee9f398
|
651 652 653 |
* When racing with an mlock or AS_UNEVICTABLE clearing * (page is unlocked) make sure that if the other thread * does not observe our setting of PG_lru and fails |
245132643
|
654 |
* isolation/check_move_unevictable_pages, |
21ee9f398
|
655 |
* we see PG_mlocked/AS_UNEVICTABLE cleared below and move |
6a7b95481
|
656 657 |
* the page back to the evictable list. * |
21ee9f398
|
658 |
* The other side is TestClearPageMlocked() or shmem_lock(). |
6a7b95481
|
659 660 |
*/ smp_mb(); |
894bc3104
|
661 |
} |
894bc3104
|
662 663 664 665 666 667 |
/* * page's status can change while we move it among lru. If an evictable * page is on unevictable list, it never be freed. To avoid that, * check after we added it to the list, again. */ |
0ec3b74c7
|
668 |
if (is_unevictable && page_evictable(page)) { |
894bc3104
|
669 670 671 672 673 674 675 676 677 |
if (!isolate_lru_page(page)) { put_page(page); goto redo; } /* This means someone else dropped this page from LRU * So, it will be freed or putback to LRU again. There is * nothing to do here. */ } |
0ec3b74c7
|
678 |
if (was_unevictable && !is_unevictable) |
bbfd28eee
|
679 |
count_vm_event(UNEVICTABLE_PGRESCUED); |
0ec3b74c7
|
680 |
else if (!was_unevictable && is_unevictable) |
bbfd28eee
|
681 |
count_vm_event(UNEVICTABLE_PGCULLED); |
894bc3104
|
682 683 |
put_page(page); /* drop ref from isolate */ } |
dfc8d636c
|
684 685 686 |
enum page_references { PAGEREF_RECLAIM, PAGEREF_RECLAIM_CLEAN, |
645747462
|
687 |
PAGEREF_KEEP, |
dfc8d636c
|
688 689 690 691 692 693 |
PAGEREF_ACTIVATE, }; static enum page_references page_check_references(struct page *page, struct scan_control *sc) { |
645747462
|
694 |
int referenced_ptes, referenced_page; |
dfc8d636c
|
695 |
unsigned long vm_flags; |
dfc8d636c
|
696 |
|
c3ac9a8ad
|
697 698 |
referenced_ptes = page_referenced(page, 1, sc->target_mem_cgroup, &vm_flags); |
645747462
|
699 |
referenced_page = TestClearPageReferenced(page); |
dfc8d636c
|
700 |
|
dfc8d636c
|
701 702 703 704 705 706 |
/* * Mlock lost the isolation race with us. Let try_to_unmap() * move the page to the unevictable list. */ if (vm_flags & VM_LOCKED) return PAGEREF_RECLAIM; |
645747462
|
707 |
if (referenced_ptes) { |
e48982734
|
708 |
if (PageSwapBacked(page)) |
645747462
|
709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 |
return PAGEREF_ACTIVATE; /* * All mapped pages start out with page table * references from the instantiating fault, so we need * to look twice if a mapped file page is used more * than once. * * Mark it and spare it for another trip around the * inactive list. Another page table reference will * lead to its activation. * * Note: the mark is set for activated pages as well * so that recently deactivated but used pages are * quickly recovered. */ SetPageReferenced(page); |
34dbc67a6
|
725 |
if (referenced_page || referenced_ptes > 1) |
645747462
|
726 |
return PAGEREF_ACTIVATE; |
c909e9936
|
727 728 729 730 731 |
/* * Activate file-backed executable pages after first usage. */ if (vm_flags & VM_EXEC) return PAGEREF_ACTIVATE; |
645747462
|
732 733 |
return PAGEREF_KEEP; } |
dfc8d636c
|
734 735 |
/* Reclaim if clean, defer dirty pages to writeback */ |
2e30244a7
|
736 |
if (referenced_page && !PageSwapBacked(page)) |
645747462
|
737 738 739 |
return PAGEREF_RECLAIM_CLEAN; return PAGEREF_RECLAIM; |
dfc8d636c
|
740 |
} |
e2be15f6c
|
741 742 743 744 |
/* Check if a page is dirty or under writeback */ static void page_check_dirty_writeback(struct page *page, bool *dirty, bool *writeback) { |
b45972265
|
745 |
struct address_space *mapping; |
e2be15f6c
|
746 747 748 749 750 751 752 753 754 755 756 757 758 |
/* * Anonymous pages are not handled by flushers and must be written * from reclaim context. Do not stall reclaim based on them */ if (!page_is_file_cache(page)) { *dirty = false; *writeback = false; return; } /* By default assume that the page flags are accurate */ *dirty = PageDirty(page); *writeback = PageWriteback(page); |
b45972265
|
759 760 761 762 763 764 765 766 |
/* Verify dirty/writeback state if the filesystem supports it */ if (!page_has_private(page)) return; mapping = page_mapping(page); if (mapping && mapping->a_ops->is_dirty_writeback) mapping->a_ops->is_dirty_writeback(page, dirty, writeback); |
e2be15f6c
|
767 |
} |
e286781d5
|
768 |
/* |
1742f19fa
|
769 |
* shrink_page_list() returns the number of reclaimed pages |
1da177e4c
|
770 |
*/ |
1742f19fa
|
771 |
static unsigned long shrink_page_list(struct list_head *page_list, |
6a18adb35
|
772 |
struct zone *zone, |
f84f6e2b0
|
773 |
struct scan_control *sc, |
02c6de8d7
|
774 |
enum ttu_flags ttu_flags, |
8e9502828
|
775 |
unsigned long *ret_nr_dirty, |
d43006d50
|
776 |
unsigned long *ret_nr_unqueued_dirty, |
8e9502828
|
777 |
unsigned long *ret_nr_congested, |
02c6de8d7
|
778 |
unsigned long *ret_nr_writeback, |
b1a6f21e3
|
779 |
unsigned long *ret_nr_immediate, |
02c6de8d7
|
780 |
bool force_reclaim) |
1da177e4c
|
781 782 |
{ LIST_HEAD(ret_pages); |
abe4c3b50
|
783 |
LIST_HEAD(free_pages); |
1da177e4c
|
784 |
int pgactivate = 0; |
d43006d50
|
785 |
unsigned long nr_unqueued_dirty = 0; |
0e093d997
|
786 787 |
unsigned long nr_dirty = 0; unsigned long nr_congested = 0; |
05ff51376
|
788 |
unsigned long nr_reclaimed = 0; |
92df3a723
|
789 |
unsigned long nr_writeback = 0; |
b1a6f21e3
|
790 |
unsigned long nr_immediate = 0; |
1da177e4c
|
791 792 |
cond_resched(); |
1da177e4c
|
793 794 795 796 |
while (!list_empty(page_list)) { struct address_space *mapping; struct page *page; int may_enter_fs; |
02c6de8d7
|
797 |
enum page_references references = PAGEREF_RECLAIM_CLEAN; |
e2be15f6c
|
798 |
bool dirty, writeback; |
1da177e4c
|
799 800 801 802 803 |
cond_resched(); page = lru_to_page(page_list); list_del(&page->lru); |
529ae9aaa
|
804 |
if (!trylock_page(page)) |
1da177e4c
|
805 |
goto keep; |
309381fea
|
806 807 |
VM_BUG_ON_PAGE(PageActive(page), page); VM_BUG_ON_PAGE(page_zone(page) != zone, page); |
1da177e4c
|
808 809 |
sc->nr_scanned++; |
80e434260
|
810 |
|
39b5f29ac
|
811 |
if (unlikely(!page_evictable(page))) |
b291f0003
|
812 |
goto cull_mlocked; |
894bc3104
|
813 |
|
a6dc60f89
|
814 |
if (!sc->may_unmap && page_mapped(page)) |
80e434260
|
815 |
goto keep_locked; |
1da177e4c
|
816 817 818 |
/* Double the slab pressure for mapped and swapcache pages */ if (page_mapped(page) || PageSwapCache(page)) sc->nr_scanned++; |
c661b078f
|
819 820 |
may_enter_fs = (sc->gfp_mask & __GFP_FS) || (PageSwapCache(page) && (sc->gfp_mask & __GFP_IO)); |
283aba9f9
|
821 |
/* |
e2be15f6c
|
822 823 824 825 826 827 828 829 830 831 832 |
* The number of dirty pages determines if a zone is marked * reclaim_congested which affects wait_iff_congested. kswapd * will stall and start writing pages if the tail of the LRU * is all dirty unqueued pages. */ page_check_dirty_writeback(page, &dirty, &writeback); if (dirty || writeback) nr_dirty++; if (dirty && !writeback) nr_unqueued_dirty++; |
d04e8acd0
|
833 834 835 836 837 838 |
/* * Treat this page as congested if the underlying BDI is or if * pages are cycling through the LRU so quickly that the * pages marked for immediate reclaim are making it to the * end of the LRU a second time. */ |
e2be15f6c
|
839 |
mapping = page_mapping(page); |
d04e8acd0
|
840 841 |
if ((mapping && bdi_write_congested(mapping->backing_dev_info)) || (writeback && PageReclaim(page))) |
e2be15f6c
|
842 843 844 |
nr_congested++; /* |
283aba9f9
|
845 846 847 848 849 850 851 852 853 854 |
* If a page at the tail of the LRU is under writeback, there * are three cases to consider. * * 1) If reclaim is encountering an excessive number of pages * under writeback and this page is both under writeback and * PageReclaim then it indicates that pages are being queued * for IO but are being recycled through the LRU before the * IO can complete. Waiting on the page itself risks an * indefinite stall if it is impossible to writeback the * page due to IO error or disconnected storage so instead |
b1a6f21e3
|
855 856 |
* note that the LRU is being scanned too quickly and the * caller can stall after page list has been processed. |
283aba9f9
|
857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 |
* * 2) Global reclaim encounters a page, memcg encounters a * page that is not marked for immediate reclaim or * the caller does not have __GFP_IO. In this case mark * the page for immediate reclaim and continue scanning. * * __GFP_IO is checked because a loop driver thread might * enter reclaim, and deadlock if it waits on a page for * which it is needed to do the write (loop masks off * __GFP_IO|__GFP_FS for this reason); but more thought * would probably show more reasons. * * Don't require __GFP_FS, since we're not going into the * FS, just waiting on its writeback completion. Worryingly, * ext4 gfs2 and xfs allocate pages with * grab_cache_page_write_begin(,,AOP_FLAG_NOFS), so testing * may_enter_fs here is liable to OOM on them. * * 3) memcg encounters a page that is not already marked * PageReclaim. memcg does not have any dirty pages * throttling so we could easily OOM just because too many * pages are in writeback and there is nothing else to * reclaim. Wait for the writeback to complete. */ |
c661b078f
|
881 |
if (PageWriteback(page)) { |
283aba9f9
|
882 883 884 |
/* Case 1 above */ if (current_is_kswapd() && PageReclaim(page) && |
570546517
|
885 |
test_bit(ZONE_WRITEBACK, &zone->flags)) { |
b1a6f21e3
|
886 887 |
nr_immediate++; goto keep_locked; |
283aba9f9
|
888 889 890 |
/* Case 2 above */ } else if (global_reclaim(sc) || |
c3b94f44f
|
891 892 893 894 895 896 897 898 899 900 901 902 903 |
!PageReclaim(page) || !(sc->gfp_mask & __GFP_IO)) { /* * This is slightly racy - end_page_writeback() * might have just cleared PageReclaim, then * setting PageReclaim here end up interpreted * as PageReadahead - but that does not matter * enough to care. What we do want is for this * page to have PageReclaim set next time memcg * reclaim reaches the tests above, so it will * then wait_on_page_writeback() to avoid OOM; * and it's also appropriate in global reclaim. */ SetPageReclaim(page); |
e62e384e9
|
904 |
nr_writeback++; |
283aba9f9
|
905 |
|
c3b94f44f
|
906 |
goto keep_locked; |
283aba9f9
|
907 908 909 910 |
/* Case 3 above */ } else { wait_on_page_writeback(page); |
e62e384e9
|
911 |
} |
c661b078f
|
912 |
} |
1da177e4c
|
913 |
|
02c6de8d7
|
914 915 |
if (!force_reclaim) references = page_check_references(page, sc); |
dfc8d636c
|
916 917 |
switch (references) { case PAGEREF_ACTIVATE: |
1da177e4c
|
918 |
goto activate_locked; |
645747462
|
919 920 |
case PAGEREF_KEEP: goto keep_locked; |
dfc8d636c
|
921 922 923 924 |
case PAGEREF_RECLAIM: case PAGEREF_RECLAIM_CLEAN: ; /* try to reclaim the page below */ } |
1da177e4c
|
925 |
|
1da177e4c
|
926 927 928 929 |
/* * Anonymous process memory has backing store? * Try to allocate it some swap space here. */ |
b291f0003
|
930 |
if (PageAnon(page) && !PageSwapCache(page)) { |
63eb6b93c
|
931 932 |
if (!(sc->gfp_mask & __GFP_IO)) goto keep_locked; |
5bc7b8aca
|
933 |
if (!add_to_swap(page, page_list)) |
1da177e4c
|
934 |
goto activate_locked; |
63eb6b93c
|
935 |
may_enter_fs = 1; |
1da177e4c
|
936 |
|
e2be15f6c
|
937 938 939 |
/* Adding to swap updated mapping */ mapping = page_mapping(page); } |
1da177e4c
|
940 941 942 943 944 945 |
/* * The page is mapped into the page tables of one or more * processes. Try to unmap it here. */ if (page_mapped(page) && mapping) { |
02c6de8d7
|
946 |
switch (try_to_unmap(page, ttu_flags)) { |
1da177e4c
|
947 948 949 950 |
case SWAP_FAIL: goto activate_locked; case SWAP_AGAIN: goto keep_locked; |
b291f0003
|
951 952 |
case SWAP_MLOCK: goto cull_mlocked; |
1da177e4c
|
953 954 955 956 957 958 |
case SWAP_SUCCESS: ; /* try to free the page below */ } } if (PageDirty(page)) { |
ee72886d8
|
959 960 |
/* * Only kswapd can writeback filesystem pages to |
d43006d50
|
961 962 |
* avoid risk of stack overflow but only writeback * if many dirty pages have been encountered. |
ee72886d8
|
963 |
*/ |
f84f6e2b0
|
964 |
if (page_is_file_cache(page) && |
9e3b2f8cd
|
965 |
(!current_is_kswapd() || |
570546517
|
966 |
!test_bit(ZONE_DIRTY, &zone->flags))) { |
49ea7eb65
|
967 968 969 970 971 972 973 974 |
/* * Immediately reclaim when written back. * Similar in principal to deactivate_page() * except we already have the page isolated * and know it's dirty */ inc_zone_page_state(page, NR_VMSCAN_IMMEDIATE); SetPageReclaim(page); |
ee72886d8
|
975 976 |
goto keep_locked; } |
dfc8d636c
|
977 |
if (references == PAGEREF_RECLAIM_CLEAN) |
1da177e4c
|
978 |
goto keep_locked; |
4dd4b9202
|
979 |
if (!may_enter_fs) |
1da177e4c
|
980 |
goto keep_locked; |
52a8363ea
|
981 |
if (!sc->may_writepage) |
1da177e4c
|
982 983 984 |
goto keep_locked; /* Page is dirty, try to write it out here */ |
7d3579e8e
|
985 |
switch (pageout(page, mapping, sc)) { |
1da177e4c
|
986 987 988 989 990 |
case PAGE_KEEP: goto keep_locked; case PAGE_ACTIVATE: goto activate_locked; case PAGE_SUCCESS: |
7d3579e8e
|
991 |
if (PageWriteback(page)) |
41ac1999c
|
992 |
goto keep; |
7d3579e8e
|
993 |
if (PageDirty(page)) |
1da177e4c
|
994 |
goto keep; |
7d3579e8e
|
995 |
|
1da177e4c
|
996 997 998 999 |
/* * A synchronous write - probably a ramdisk. Go * ahead and try to reclaim the page. */ |
529ae9aaa
|
1000 |
if (!trylock_page(page)) |
1da177e4c
|
1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 |
goto keep; if (PageDirty(page) || PageWriteback(page)) goto keep_locked; mapping = page_mapping(page); case PAGE_CLEAN: ; /* try to free the page below */ } } /* * If the page has buffers, try to free the buffer mappings * associated with this page. If we succeed we try to free * the page as well. * * We do this even if the page is PageDirty(). * try_to_release_page() does not perform I/O, but it is * possible for a page to have PageDirty set, but it is actually * clean (all its buffers are clean). This happens if the * buffers were written out directly, with submit_bh(). ext3 |
894bc3104
|
1020 |
* will do this, as well as the blockdev mapping. |
1da177e4c
|
1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 |
* try_to_release_page() will discover that cleanness and will * drop the buffers and mark the page clean - it can be freed. * * Rarely, pages can have buffers and no ->mapping. These are * the pages which were not successfully invalidated in * truncate_complete_page(). We try to drop those buffers here * and if that worked, and the page is no longer mapped into * process address space (page_count == 1) it can be freed. * Otherwise, leave the page on the LRU so it is swappable. */ |
266cf658e
|
1031 |
if (page_has_private(page)) { |
1da177e4c
|
1032 1033 |
if (!try_to_release_page(page, sc->gfp_mask)) goto activate_locked; |
e286781d5
|
1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 |
if (!mapping && page_count(page) == 1) { unlock_page(page); if (put_page_testzero(page)) goto free_it; else { /* * rare race with speculative reference. * the speculative reference will free * this page shortly, so we may * increment nr_reclaimed here (and * leave it off the LRU). */ nr_reclaimed++; continue; } } |
1da177e4c
|
1050 |
} |
a528910e1
|
1051 |
if (!mapping || !__remove_mapping(mapping, page, true)) |
49d2e9cc4
|
1052 |
goto keep_locked; |
1da177e4c
|
1053 |
|
a978d6f52
|
1054 1055 1056 1057 1058 1059 1060 1061 |
/* * At this point, we have no other references and there is * no way to pick any more up (removed from LRU, removed * from pagecache). Can use non-atomic bitops now (and * we obviously don't have to worry about waking up a process * waiting on the page lock, because there are no references. */ __clear_page_locked(page); |
e286781d5
|
1062 |
free_it: |
05ff51376
|
1063 |
nr_reclaimed++; |
abe4c3b50
|
1064 1065 1066 1067 1068 1069 |
/* * Is there need to periodically free_page_list? It would * appear not as the counts should be low */ list_add(&page->lru, &free_pages); |
1da177e4c
|
1070 |
continue; |
b291f0003
|
1071 |
cull_mlocked: |
63d6c5ad7
|
1072 1073 |
if (PageSwapCache(page)) try_to_free_swap(page); |
b291f0003
|
1074 1075 1076 |
unlock_page(page); putback_lru_page(page); continue; |
1da177e4c
|
1077 |
activate_locked: |
68a22394c
|
1078 1079 |
/* Not a candidate for swapping, so reclaim swap space. */ if (PageSwapCache(page) && vm_swap_full()) |
a2c43eed8
|
1080 |
try_to_free_swap(page); |
309381fea
|
1081 |
VM_BUG_ON_PAGE(PageActive(page), page); |
1da177e4c
|
1082 1083 1084 1085 1086 1087 |
SetPageActive(page); pgactivate++; keep_locked: unlock_page(page); keep: list_add(&page->lru, &ret_pages); |
309381fea
|
1088 |
VM_BUG_ON_PAGE(PageLRU(page) || PageUnevictable(page), page); |
1da177e4c
|
1089 |
} |
abe4c3b50
|
1090 |
|
747db954c
|
1091 |
mem_cgroup_uncharge_list(&free_pages); |
b745bc85f
|
1092 |
free_hot_cold_page_list(&free_pages, true); |
abe4c3b50
|
1093 |
|
1da177e4c
|
1094 |
list_splice(&ret_pages, page_list); |
f8891e5e1
|
1095 |
count_vm_events(PGACTIVATE, pgactivate); |
0a31bc97c
|
1096 |
|
8e9502828
|
1097 1098 |
*ret_nr_dirty += nr_dirty; *ret_nr_congested += nr_congested; |
d43006d50
|
1099 |
*ret_nr_unqueued_dirty += nr_unqueued_dirty; |
92df3a723
|
1100 |
*ret_nr_writeback += nr_writeback; |
b1a6f21e3
|
1101 |
*ret_nr_immediate += nr_immediate; |
05ff51376
|
1102 |
return nr_reclaimed; |
1da177e4c
|
1103 |
} |
02c6de8d7
|
1104 1105 1106 1107 1108 1109 1110 1111 |
unsigned long reclaim_clean_pages_from_list(struct zone *zone, struct list_head *page_list) { struct scan_control sc = { .gfp_mask = GFP_KERNEL, .priority = DEF_PRIORITY, .may_unmap = 1, }; |
8e9502828
|
1112 |
unsigned long ret, dummy1, dummy2, dummy3, dummy4, dummy5; |
02c6de8d7
|
1113 1114 1115 1116 |
struct page *page, *next; LIST_HEAD(clean_pages); list_for_each_entry_safe(page, next, page_list, lru) { |
117aad1e9
|
1117 1118 |
if (page_is_file_cache(page) && !PageDirty(page) && !isolated_balloon_page(page)) { |
02c6de8d7
|
1119 1120 1121 1122 1123 1124 |
ClearPageActive(page); list_move(&page->lru, &clean_pages); } } ret = shrink_page_list(&clean_pages, zone, &sc, |
8e9502828
|
1125 1126 |
TTU_UNMAP|TTU_IGNORE_ACCESS, &dummy1, &dummy2, &dummy3, &dummy4, &dummy5, true); |
02c6de8d7
|
1127 |
list_splice(&clean_pages, page_list); |
83da75100
|
1128 |
mod_zone_page_state(zone, NR_ISOLATED_FILE, -ret); |
02c6de8d7
|
1129 1130 |
return ret; } |
5ad333eb6
|
1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 |
/* * Attempt to remove the specified page from its LRU. Only take this page * if it is of the appropriate PageActive status. Pages which are being * freed elsewhere are also ignored. * * page: page to consider * mode: one of the LRU isolation modes defined above * * returns 0 on success, -ve errno on failure. */ |
f3fd4a619
|
1141 |
int __isolate_lru_page(struct page *page, isolate_mode_t mode) |
5ad333eb6
|
1142 1143 1144 1145 1146 1147 |
{ int ret = -EINVAL; /* Only take pages on the LRU. */ if (!PageLRU(page)) return ret; |
e46a28790
|
1148 1149 |
/* Compaction should not handle unevictable pages but CMA can do so */ if (PageUnevictable(page) && !(mode & ISOLATE_UNEVICTABLE)) |
894bc3104
|
1150 |
return ret; |
5ad333eb6
|
1151 |
ret = -EBUSY; |
08e552c69
|
1152 |
|
c82449352
|
1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 |
/* * To minimise LRU disruption, the caller can indicate that it only * wants to isolate pages it will be able to operate on without * blocking - clean pages for the most part. * * ISOLATE_CLEAN means that only clean pages should be isolated. This * is used by reclaim when it is cannot write to backing storage * * ISOLATE_ASYNC_MIGRATE is used to indicate that it only wants to pages * that it is possible to migrate without blocking */ if (mode & (ISOLATE_CLEAN|ISOLATE_ASYNC_MIGRATE)) { /* All the caller can do on PageWriteback is block */ if (PageWriteback(page)) return ret; if (PageDirty(page)) { struct address_space *mapping; /* ISOLATE_CLEAN means only clean pages */ if (mode & ISOLATE_CLEAN) return ret; /* * Only pages without mappings or that have a * ->migratepage callback are possible to migrate * without blocking */ mapping = page_mapping(page); if (mapping && !mapping->a_ops->migratepage) return ret; } } |
39deaf858
|
1186 |
|
f80c06736
|
1187 1188 |
if ((mode & ISOLATE_UNMAPPED) && page_mapped(page)) return ret; |
5ad333eb6
|
1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 |
if (likely(get_page_unless_zero(page))) { /* * Be careful not to clear PageLRU until after we're * sure the page is not being freed elsewhere -- the * page release code relies on it. */ ClearPageLRU(page); ret = 0; } return ret; } |
49d2e9cc4
|
1201 |
/* |
1da177e4c
|
1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 |
* zone->lru_lock is heavily contended. Some of the functions that * shrink the lists perform better by taking out a batch of pages * and working on them outside the LRU lock. * * For pagecache intensive workloads, this function is the hottest * spot in the kernel (apart from copy_*_user functions). * * Appropriate locks must be held before calling this function. * * @nr_to_scan: The number of pages to look through on the list. |
5dc35979e
|
1212 |
* @lruvec: The LRU vector to pull pages from. |
1da177e4c
|
1213 |
* @dst: The temp list to put pages on to. |
f626012db
|
1214 |
* @nr_scanned: The number of pages that were scanned. |
fe2c2a106
|
1215 |
* @sc: The scan_control struct for this reclaim session |
5ad333eb6
|
1216 |
* @mode: One of the LRU isolation modes |
3cb994517
|
1217 |
* @lru: LRU list id for isolating |
1da177e4c
|
1218 1219 1220 |
* * returns how many pages were moved onto *@dst. */ |
69e05944a
|
1221 |
static unsigned long isolate_lru_pages(unsigned long nr_to_scan, |
5dc35979e
|
1222 |
struct lruvec *lruvec, struct list_head *dst, |
fe2c2a106
|
1223 |
unsigned long *nr_scanned, struct scan_control *sc, |
3cb994517
|
1224 |
isolate_mode_t mode, enum lru_list lru) |
1da177e4c
|
1225 |
{ |
75b00af77
|
1226 |
struct list_head *src = &lruvec->lists[lru]; |
69e05944a
|
1227 |
unsigned long nr_taken = 0; |
c9b02d970
|
1228 |
unsigned long scan; |
1da177e4c
|
1229 |
|
c9b02d970
|
1230 |
for (scan = 0; scan < nr_to_scan && !list_empty(src); scan++) { |
5ad333eb6
|
1231 |
struct page *page; |
fa9add641
|
1232 |
int nr_pages; |
5ad333eb6
|
1233 |
|
1da177e4c
|
1234 1235 |
page = lru_to_page(src); prefetchw_prev_lru_page(page, src, flags); |
309381fea
|
1236 |
VM_BUG_ON_PAGE(!PageLRU(page), page); |
8d438f96d
|
1237 |
|
f3fd4a619
|
1238 |
switch (__isolate_lru_page(page, mode)) { |
5ad333eb6
|
1239 |
case 0: |
fa9add641
|
1240 1241 |
nr_pages = hpage_nr_pages(page); mem_cgroup_update_lru_size(lruvec, lru, -nr_pages); |
5ad333eb6
|
1242 |
list_move(&page->lru, dst); |
fa9add641
|
1243 |
nr_taken += nr_pages; |
5ad333eb6
|
1244 1245 1246 1247 1248 1249 |
break; case -EBUSY: /* else it is being freed elsewhere */ list_move(&page->lru, src); continue; |
46453a6e1
|
1250 |
|
5ad333eb6
|
1251 1252 1253 |
default: BUG(); } |
1da177e4c
|
1254 |
} |
f626012db
|
1255 |
*nr_scanned = scan; |
75b00af77
|
1256 1257 |
trace_mm_vmscan_lru_isolate(sc->order, nr_to_scan, scan, nr_taken, mode, is_file_lru(lru)); |
1da177e4c
|
1258 1259 |
return nr_taken; } |
62695a84e
|
1260 1261 1262 1263 1264 1265 1266 1267 1268 1269 1270 |
/** * isolate_lru_page - tries to isolate a page from its LRU list * @page: page to isolate from its LRU list * * Isolates a @page from an LRU list, clears PageLRU and adjusts the * vmstat statistic corresponding to whatever LRU list the page was on. * * Returns 0 if the page was removed from an LRU list. * Returns -EBUSY if the page was not on an LRU list. * * The returned page will have PageLRU() cleared. If it was found on |
894bc3104
|
1271 1272 1273 |
* the active list, it will have PageActive set. If it was found on * the unevictable list, it will have the PageUnevictable bit set. That flag * may need to be cleared by the caller before letting the page go. |
62695a84e
|
1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 |
* * The vmstat statistic corresponding to the list on which the page was * found will be decremented. * * Restrictions: * (1) Must be called with an elevated refcount on the page. This is a * fundamentnal difference from isolate_lru_pages (which is called * without a stable reference). * (2) the lru_lock must not be held. * (3) interrupts must be enabled. */ int isolate_lru_page(struct page *page) { int ret = -EBUSY; |
309381fea
|
1288 |
VM_BUG_ON_PAGE(!page_count(page), page); |
0c917313a
|
1289 |
|
62695a84e
|
1290 1291 |
if (PageLRU(page)) { struct zone *zone = page_zone(page); |
fa9add641
|
1292 |
struct lruvec *lruvec; |
62695a84e
|
1293 1294 |
spin_lock_irq(&zone->lru_lock); |
fa9add641
|
1295 |
lruvec = mem_cgroup_page_lruvec(page, zone); |
0c917313a
|
1296 |
if (PageLRU(page)) { |
894bc3104
|
1297 |
int lru = page_lru(page); |
0c917313a
|
1298 |
get_page(page); |
62695a84e
|
1299 |
ClearPageLRU(page); |
fa9add641
|
1300 1301 |
del_page_from_lru_list(page, lruvec, lru); ret = 0; |
62695a84e
|
1302 1303 1304 1305 1306 |
} spin_unlock_irq(&zone->lru_lock); } return ret; } |
5ad333eb6
|
1307 |
/* |
d37dd5dcb
|
1308 1309 1310 1311 1312 |
* A direct reclaimer may isolate SWAP_CLUSTER_MAX pages from the LRU list and * then get resheduled. When there are massive number of tasks doing page * allocation, such sleeping direct reclaimers may keep piling up on each CPU, * the LRU list will go small and be scanned faster than necessary, leading to * unnecessary swapping, thrashing and OOM. |
35cd78156
|
1313 1314 1315 1316 1317 1318 1319 1320 |
*/ static int too_many_isolated(struct zone *zone, int file, struct scan_control *sc) { unsigned long inactive, isolated; if (current_is_kswapd()) return 0; |
89b5fae53
|
1321 |
if (!global_reclaim(sc)) |
35cd78156
|
1322 1323 1324 1325 1326 1327 1328 1329 1330 |
return 0; if (file) { inactive = zone_page_state(zone, NR_INACTIVE_FILE); isolated = zone_page_state(zone, NR_ISOLATED_FILE); } else { inactive = zone_page_state(zone, NR_INACTIVE_ANON); isolated = zone_page_state(zone, NR_ISOLATED_ANON); } |
3cf23841b
|
1331 1332 1333 1334 1335 1336 1337 |
/* * GFP_NOIO/GFP_NOFS callers are allowed to isolate more pages, so they * won't get blocked by normal direct-reclaimers, forming a circular * deadlock. */ if ((sc->gfp_mask & GFP_IOFS) == GFP_IOFS) inactive >>= 3; |
35cd78156
|
1338 1339 |
return isolated > inactive; } |
666356297
|
1340 |
static noinline_for_stack void |
75b00af77
|
1341 |
putback_inactive_pages(struct lruvec *lruvec, struct list_head *page_list) |
666356297
|
1342 |
{ |
27ac81d85
|
1343 1344 |
struct zone_reclaim_stat *reclaim_stat = &lruvec->reclaim_stat; struct zone *zone = lruvec_zone(lruvec); |
3f79768f2
|
1345 |
LIST_HEAD(pages_to_free); |
666356297
|
1346 |
|
666356297
|
1347 1348 1349 |
/* * Put back any unfreeable pages. */ |
666356297
|
1350 |
while (!list_empty(page_list)) { |
3f79768f2
|
1351 |
struct page *page = lru_to_page(page_list); |
666356297
|
1352 |
int lru; |
3f79768f2
|
1353 |
|
309381fea
|
1354 |
VM_BUG_ON_PAGE(PageLRU(page), page); |
666356297
|
1355 |
list_del(&page->lru); |
39b5f29ac
|
1356 |
if (unlikely(!page_evictable(page))) { |
666356297
|
1357 1358 1359 1360 1361 |
spin_unlock_irq(&zone->lru_lock); putback_lru_page(page); spin_lock_irq(&zone->lru_lock); continue; } |
fa9add641
|
1362 1363 |
lruvec = mem_cgroup_page_lruvec(page, zone); |
7a608572a
|
1364 |
SetPageLRU(page); |
666356297
|
1365 |
lru = page_lru(page); |
fa9add641
|
1366 |
add_page_to_lru_list(page, lruvec, lru); |
666356297
|
1367 1368 |
if (is_active_lru(lru)) { int file = is_file_lru(lru); |
9992af102
|
1369 1370 |
int numpages = hpage_nr_pages(page); reclaim_stat->recent_rotated[file] += numpages; |
666356297
|
1371 |
} |
2bcf88796
|
1372 1373 1374 |
if (put_page_testzero(page)) { __ClearPageLRU(page); __ClearPageActive(page); |
fa9add641
|
1375 |
del_page_from_lru_list(page, lruvec, lru); |
2bcf88796
|
1376 1377 1378 |
if (unlikely(PageCompound(page))) { spin_unlock_irq(&zone->lru_lock); |
747db954c
|
1379 |
mem_cgroup_uncharge(page); |
2bcf88796
|
1380 1381 1382 1383 |
(*get_compound_page_dtor(page))(page); spin_lock_irq(&zone->lru_lock); } else list_add(&page->lru, &pages_to_free); |
666356297
|
1384 1385 |
} } |
666356297
|
1386 |
|
3f79768f2
|
1387 1388 1389 1390 |
/* * To save our caller's stack, now use input list for pages to free. */ list_splice(&pages_to_free, page_list); |
666356297
|
1391 1392 1393 |
} /* |
399ba0b95
|
1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 |
* If a kernel thread (such as nfsd for loop-back mounts) services * a backing device by writing to the page cache it sets PF_LESS_THROTTLE. * In that case we should only throttle if the backing device it is * writing to is congested. In other cases it is safe to throttle. */ static int current_may_throttle(void) { return !(current->flags & PF_LESS_THROTTLE) || current->backing_dev_info == NULL || bdi_write_congested(current->backing_dev_info); } /* |
1742f19fa
|
1407 1408 |
* shrink_inactive_list() is a helper for shrink_zone(). It returns the number * of reclaimed pages |
1da177e4c
|
1409 |
*/ |
666356297
|
1410 |
static noinline_for_stack unsigned long |
1a93be0e7
|
1411 |
shrink_inactive_list(unsigned long nr_to_scan, struct lruvec *lruvec, |
9e3b2f8cd
|
1412 |
struct scan_control *sc, enum lru_list lru) |
1da177e4c
|
1413 1414 |
{ LIST_HEAD(page_list); |
e247dbce5
|
1415 |
unsigned long nr_scanned; |
05ff51376
|
1416 |
unsigned long nr_reclaimed = 0; |
e247dbce5
|
1417 |
unsigned long nr_taken; |
8e9502828
|
1418 1419 |
unsigned long nr_dirty = 0; unsigned long nr_congested = 0; |
e2be15f6c
|
1420 |
unsigned long nr_unqueued_dirty = 0; |
92df3a723
|
1421 |
unsigned long nr_writeback = 0; |
b1a6f21e3
|
1422 |
unsigned long nr_immediate = 0; |
f3fd4a619
|
1423 |
isolate_mode_t isolate_mode = 0; |
3cb994517
|
1424 |
int file = is_file_lru(lru); |
1a93be0e7
|
1425 1426 |
struct zone *zone = lruvec_zone(lruvec); struct zone_reclaim_stat *reclaim_stat = &lruvec->reclaim_stat; |
78dc583d3
|
1427 |
|
35cd78156
|
1428 |
while (unlikely(too_many_isolated(zone, file, sc))) { |
58355c787
|
1429 |
congestion_wait(BLK_RW_ASYNC, HZ/10); |
35cd78156
|
1430 1431 1432 1433 1434 |
/* We are about to die and free our memory. Return now. */ if (fatal_signal_pending(current)) return SWAP_CLUSTER_MAX; } |
1da177e4c
|
1435 |
lru_add_drain(); |
f80c06736
|
1436 1437 |
if (!sc->may_unmap) |
613172891
|
1438 |
isolate_mode |= ISOLATE_UNMAPPED; |
f80c06736
|
1439 |
if (!sc->may_writepage) |
613172891
|
1440 |
isolate_mode |= ISOLATE_CLEAN; |
f80c06736
|
1441 |
|
1da177e4c
|
1442 |
spin_lock_irq(&zone->lru_lock); |
b35ea17b7
|
1443 |
|
5dc35979e
|
1444 1445 |
nr_taken = isolate_lru_pages(nr_to_scan, lruvec, &page_list, &nr_scanned, sc, isolate_mode, lru); |
95d918fc0
|
1446 1447 1448 |
__mod_zone_page_state(zone, NR_LRU_BASE + lru, -nr_taken); __mod_zone_page_state(zone, NR_ISOLATED_ANON + file, nr_taken); |
89b5fae53
|
1449 |
if (global_reclaim(sc)) { |
0d5d823ab
|
1450 |
__mod_zone_page_state(zone, NR_PAGES_SCANNED, nr_scanned); |
e247dbce5
|
1451 |
if (current_is_kswapd()) |
75b00af77
|
1452 |
__count_zone_vm_events(PGSCAN_KSWAPD, zone, nr_scanned); |
e247dbce5
|
1453 |
else |
75b00af77
|
1454 |
__count_zone_vm_events(PGSCAN_DIRECT, zone, nr_scanned); |
e247dbce5
|
1455 |
} |
d563c0501
|
1456 |
spin_unlock_irq(&zone->lru_lock); |
b35ea17b7
|
1457 |
|
d563c0501
|
1458 |
if (nr_taken == 0) |
666356297
|
1459 |
return 0; |
5ad333eb6
|
1460 |
|
02c6de8d7
|
1461 |
nr_reclaimed = shrink_page_list(&page_list, zone, sc, TTU_UNMAP, |
8e9502828
|
1462 1463 1464 |
&nr_dirty, &nr_unqueued_dirty, &nr_congested, &nr_writeback, &nr_immediate, false); |
c661b078f
|
1465 |
|
3f79768f2
|
1466 |
spin_lock_irq(&zone->lru_lock); |
95d918fc0
|
1467 |
reclaim_stat->recent_scanned[file] += nr_taken; |
d563c0501
|
1468 |
|
904249aa6
|
1469 1470 1471 1472 1473 1474 1475 1476 |
if (global_reclaim(sc)) { if (current_is_kswapd()) __count_zone_vm_events(PGSTEAL_KSWAPD, zone, nr_reclaimed); else __count_zone_vm_events(PGSTEAL_DIRECT, zone, nr_reclaimed); } |
a74609faf
|
1477 |
|
27ac81d85
|
1478 |
putback_inactive_pages(lruvec, &page_list); |
3f79768f2
|
1479 |
|
95d918fc0
|
1480 |
__mod_zone_page_state(zone, NR_ISOLATED_ANON + file, -nr_taken); |
3f79768f2
|
1481 1482 |
spin_unlock_irq(&zone->lru_lock); |
747db954c
|
1483 |
mem_cgroup_uncharge_list(&page_list); |
b745bc85f
|
1484 |
free_hot_cold_page_list(&page_list, true); |
e11da5b4f
|
1485 |
|
92df3a723
|
1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 |
/* * If reclaim is isolating dirty pages under writeback, it implies * that the long-lived page allocation rate is exceeding the page * laundering rate. Either the global limits are not being effective * at throttling processes due to the page distribution throughout * zones or there is heavy usage of a slow backing device. The * only option is to throttle from reclaim context which is not ideal * as there is no guarantee the dirtying process is throttled in the * same way balance_dirty_pages() manages. * |
8e9502828
|
1496 1497 1498 |
* Once a zone is flagged ZONE_WRITEBACK, kswapd will count the number * of pages under pages flagged for immediate reclaim and stall if any * are encountered in the nr_immediate check below. |
92df3a723
|
1499 |
*/ |
918fc718c
|
1500 |
if (nr_writeback && nr_writeback == nr_taken) |
570546517
|
1501 |
set_bit(ZONE_WRITEBACK, &zone->flags); |
92df3a723
|
1502 |
|
d43006d50
|
1503 |
/* |
b1a6f21e3
|
1504 1505 |
* memcg will stall in page writeback so only consider forcibly * stalling for global reclaim |
d43006d50
|
1506 |
*/ |
b1a6f21e3
|
1507 1508 |
if (global_reclaim(sc)) { /* |
8e9502828
|
1509 1510 1511 1512 |
* Tag a zone as congested if all the dirty pages scanned were * backed by a congested BDI and wait_iff_congested will stall. */ if (nr_dirty && nr_dirty == nr_congested) |
570546517
|
1513 |
set_bit(ZONE_CONGESTED, &zone->flags); |
8e9502828
|
1514 1515 |
/* |
b1a6f21e3
|
1516 1517 |
* If dirty pages are scanned that are not queued for IO, it * implies that flushers are not keeping up. In this case, flag |
570546517
|
1518 1519 |
* the zone ZONE_DIRTY and kswapd will start writing pages from * reclaim context. |
b1a6f21e3
|
1520 1521 |
*/ if (nr_unqueued_dirty == nr_taken) |
570546517
|
1522 |
set_bit(ZONE_DIRTY, &zone->flags); |
b1a6f21e3
|
1523 1524 |
/* |
b738d7646
|
1525 1526 1527 |
* If kswapd scans pages marked marked for immediate * reclaim and under writeback (nr_immediate), it implies * that pages are cycling through the LRU faster than |
b1a6f21e3
|
1528 1529 |
* they are written so also forcibly stall. */ |
b738d7646
|
1530 |
if (nr_immediate && current_may_throttle()) |
b1a6f21e3
|
1531 |
congestion_wait(BLK_RW_ASYNC, HZ/10); |
e2be15f6c
|
1532 |
} |
d43006d50
|
1533 |
|
8e9502828
|
1534 1535 1536 1537 1538 |
/* * Stall direct reclaim for IO completions if underlying BDIs or zone * is congested. Allow kswapd to continue until it starts encountering * unqueued dirty pages or cycling through the LRU too quickly. */ |
399ba0b95
|
1539 1540 |
if (!sc->hibernation_mode && !current_is_kswapd() && current_may_throttle()) |
8e9502828
|
1541 |
wait_iff_congested(zone, BLK_RW_ASYNC, HZ/10); |
e11da5b4f
|
1542 1543 1544 |
trace_mm_vmscan_lru_shrink_inactive(zone->zone_pgdat->node_id, zone_idx(zone), nr_scanned, nr_reclaimed, |
9e3b2f8cd
|
1545 |
sc->priority, |
23b9da55c
|
1546 |
trace_shrink_flags(file)); |
05ff51376
|
1547 |
return nr_reclaimed; |
1da177e4c
|
1548 |
} |
3bb1a852a
|
1549 |
/* |
1da177e4c
|
1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 |
* This moves pages from the active list to the inactive list. * * We move them the other way if the page is referenced by one or more * processes, from rmap. * * If the pages are mostly unmapped, the processing is fast and it is * appropriate to hold zone->lru_lock across the whole operation. But if * the pages are mapped, the processing is slow (page_referenced()) so we * should drop zone->lru_lock around each page. It's impossible to balance * this, so instead we remove the pages from the LRU while processing them. * It is safe to rely on PG_active against the non-LRU pages in here because * nobody will play with that bit on a non-LRU page. * * The downside is that we have to touch page->_count against each page. * But we had to alter page->flags anyway. */ |
1cfb419b3
|
1566 |
|
fa9add641
|
1567 |
static void move_active_pages_to_lru(struct lruvec *lruvec, |
3eb4140f0
|
1568 |
struct list_head *list, |
2bcf88796
|
1569 |
struct list_head *pages_to_free, |
3eb4140f0
|
1570 1571 |
enum lru_list lru) { |
fa9add641
|
1572 |
struct zone *zone = lruvec_zone(lruvec); |
3eb4140f0
|
1573 |
unsigned long pgmoved = 0; |
3eb4140f0
|
1574 |
struct page *page; |
fa9add641
|
1575 |
int nr_pages; |
3eb4140f0
|
1576 |
|
3eb4140f0
|
1577 1578 |
while (!list_empty(list)) { page = lru_to_page(list); |
fa9add641
|
1579 |
lruvec = mem_cgroup_page_lruvec(page, zone); |
3eb4140f0
|
1580 |
|
309381fea
|
1581 |
VM_BUG_ON_PAGE(PageLRU(page), page); |
3eb4140f0
|
1582 |
SetPageLRU(page); |
fa9add641
|
1583 1584 |
nr_pages = hpage_nr_pages(page); mem_cgroup_update_lru_size(lruvec, lru, nr_pages); |
925b7673c
|
1585 |
list_move(&page->lru, &lruvec->lists[lru]); |
fa9add641
|
1586 |
pgmoved += nr_pages; |
3eb4140f0
|
1587 |
|
2bcf88796
|
1588 1589 1590 |
if (put_page_testzero(page)) { __ClearPageLRU(page); __ClearPageActive(page); |
fa9add641
|
1591 |
del_page_from_lru_list(page, lruvec, lru); |
2bcf88796
|
1592 1593 1594 |
if (unlikely(PageCompound(page))) { spin_unlock_irq(&zone->lru_lock); |
747db954c
|
1595 |
mem_cgroup_uncharge(page); |
2bcf88796
|
1596 1597 1598 1599 |
(*get_compound_page_dtor(page))(page); spin_lock_irq(&zone->lru_lock); } else list_add(&page->lru, pages_to_free); |
3eb4140f0
|
1600 1601 1602 1603 1604 1605 |
} } __mod_zone_page_state(zone, NR_LRU_BASE + lru, pgmoved); if (!is_active_lru(lru)) __count_vm_events(PGDEACTIVATE, pgmoved); } |
1cfb419b3
|
1606 |
|
f626012db
|
1607 |
static void shrink_active_list(unsigned long nr_to_scan, |
1a93be0e7
|
1608 |
struct lruvec *lruvec, |
f16015fbf
|
1609 |
struct scan_control *sc, |
9e3b2f8cd
|
1610 |
enum lru_list lru) |
1da177e4c
|
1611 |
{ |
44c241f16
|
1612 |
unsigned long nr_taken; |
f626012db
|
1613 |
unsigned long nr_scanned; |
6fe6b7e35
|
1614 |
unsigned long vm_flags; |
1da177e4c
|
1615 |
LIST_HEAD(l_hold); /* The pages which were snipped off */ |
8cab4754d
|
1616 |
LIST_HEAD(l_active); |
b69408e88
|
1617 |
LIST_HEAD(l_inactive); |
1da177e4c
|
1618 |
struct page *page; |
1a93be0e7
|
1619 |
struct zone_reclaim_stat *reclaim_stat = &lruvec->reclaim_stat; |
44c241f16
|
1620 |
unsigned long nr_rotated = 0; |
f3fd4a619
|
1621 |
isolate_mode_t isolate_mode = 0; |
3cb994517
|
1622 |
int file = is_file_lru(lru); |
1a93be0e7
|
1623 |
struct zone *zone = lruvec_zone(lruvec); |
1da177e4c
|
1624 1625 |
lru_add_drain(); |
f80c06736
|
1626 1627 |
if (!sc->may_unmap) |
613172891
|
1628 |
isolate_mode |= ISOLATE_UNMAPPED; |
f80c06736
|
1629 |
if (!sc->may_writepage) |
613172891
|
1630 |
isolate_mode |= ISOLATE_CLEAN; |
f80c06736
|
1631 |
|
1da177e4c
|
1632 |
spin_lock_irq(&zone->lru_lock); |
925b7673c
|
1633 |
|
5dc35979e
|
1634 1635 |
nr_taken = isolate_lru_pages(nr_to_scan, lruvec, &l_hold, &nr_scanned, sc, isolate_mode, lru); |
89b5fae53
|
1636 |
if (global_reclaim(sc)) |
0d5d823ab
|
1637 |
__mod_zone_page_state(zone, NR_PAGES_SCANNED, nr_scanned); |
89b5fae53
|
1638 |
|
b7c46d151
|
1639 |
reclaim_stat->recent_scanned[file] += nr_taken; |
1cfb419b3
|
1640 |
|
f626012db
|
1641 |
__count_zone_vm_events(PGREFILL, zone, nr_scanned); |
3cb994517
|
1642 |
__mod_zone_page_state(zone, NR_LRU_BASE + lru, -nr_taken); |
a731286de
|
1643 |
__mod_zone_page_state(zone, NR_ISOLATED_ANON + file, nr_taken); |
1da177e4c
|
1644 |
spin_unlock_irq(&zone->lru_lock); |
1da177e4c
|
1645 1646 1647 1648 |
while (!list_empty(&l_hold)) { cond_resched(); page = lru_to_page(&l_hold); list_del(&page->lru); |
7e9cd4842
|
1649 |
|
39b5f29ac
|
1650 |
if (unlikely(!page_evictable(page))) { |
894bc3104
|
1651 1652 1653 |
putback_lru_page(page); continue; } |
cc715d99e
|
1654 1655 1656 1657 1658 1659 1660 |
if (unlikely(buffer_heads_over_limit)) { if (page_has_private(page) && trylock_page(page)) { if (page_has_private(page)) try_to_release_page(page, 0); unlock_page(page); } } |
c3ac9a8ad
|
1661 1662 |
if (page_referenced(page, 0, sc->target_mem_cgroup, &vm_flags)) { |
9992af102
|
1663 |
nr_rotated += hpage_nr_pages(page); |
8cab4754d
|
1664 1665 1666 1667 1668 1669 1670 1671 1672 |
/* * Identify referenced, file-backed active pages and * give them one more trip around the active list. So * that executable code get better chances to stay in * memory under moderate memory pressure. Anon pages * are not likely to be evicted by use-once streaming * IO, plus JVM can create lots of anon VM_EXEC pages, * so we ignore them here. */ |
41e20983f
|
1673 |
if ((vm_flags & VM_EXEC) && page_is_file_cache(page)) { |
8cab4754d
|
1674 1675 1676 1677 |
list_add(&page->lru, &l_active); continue; } } |
7e9cd4842
|
1678 |
|
5205e56ee
|
1679 |
ClearPageActive(page); /* we are de-activating */ |
1da177e4c
|
1680 1681 |
list_add(&page->lru, &l_inactive); } |
b555749aa
|
1682 |
/* |
8cab4754d
|
1683 |
* Move pages back to the lru list. |
b555749aa
|
1684 |
*/ |
2a1dc5097
|
1685 |
spin_lock_irq(&zone->lru_lock); |
4f98a2fee
|
1686 |
/* |
8cab4754d
|
1687 1688 1689 |
* Count referenced pages from currently used mappings as rotated, * even though only some of them are actually re-activated. This * helps balance scan pressure between file and anonymous pages in |
7c0db9e91
|
1690 |
* get_scan_count. |
7e9cd4842
|
1691 |
*/ |
b7c46d151
|
1692 |
reclaim_stat->recent_rotated[file] += nr_rotated; |
556adecba
|
1693 |
|
fa9add641
|
1694 1695 |
move_active_pages_to_lru(lruvec, &l_active, &l_hold, lru); move_active_pages_to_lru(lruvec, &l_inactive, &l_hold, lru - LRU_ACTIVE); |
a731286de
|
1696 |
__mod_zone_page_state(zone, NR_ISOLATED_ANON + file, -nr_taken); |
f8891e5e1
|
1697 |
spin_unlock_irq(&zone->lru_lock); |
2bcf88796
|
1698 |
|
747db954c
|
1699 |
mem_cgroup_uncharge_list(&l_hold); |
b745bc85f
|
1700 |
free_hot_cold_page_list(&l_hold, true); |
1da177e4c
|
1701 |
} |
74e3f3c33
|
1702 |
#ifdef CONFIG_SWAP |
14797e236
|
1703 |
static int inactive_anon_is_low_global(struct zone *zone) |
f89eb90e3
|
1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 |
{ unsigned long active, inactive; active = zone_page_state(zone, NR_ACTIVE_ANON); inactive = zone_page_state(zone, NR_INACTIVE_ANON); if (inactive * zone->inactive_ratio < active) return 1; return 0; } |
14797e236
|
1715 1716 |
/** * inactive_anon_is_low - check if anonymous pages need to be deactivated |
c56d5c7df
|
1717 |
* @lruvec: LRU vector to check |
14797e236
|
1718 1719 1720 1721 |
* * Returns true if the zone does not have enough inactive anon pages, * meaning some active anon pages need to be deactivated. */ |
c56d5c7df
|
1722 |
static int inactive_anon_is_low(struct lruvec *lruvec) |
14797e236
|
1723 |
{ |
74e3f3c33
|
1724 1725 1726 1727 1728 1729 |
/* * If we don't have swap space, anonymous page deactivation * is pointless. */ if (!total_swap_pages) return 0; |
c3c787e8c
|
1730 |
if (!mem_cgroup_disabled()) |
c56d5c7df
|
1731 |
return mem_cgroup_inactive_anon_is_low(lruvec); |
f16015fbf
|
1732 |
|
c56d5c7df
|
1733 |
return inactive_anon_is_low_global(lruvec_zone(lruvec)); |
14797e236
|
1734 |
} |
74e3f3c33
|
1735 |
#else |
c56d5c7df
|
1736 |
static inline int inactive_anon_is_low(struct lruvec *lruvec) |
74e3f3c33
|
1737 1738 1739 1740 |
{ return 0; } #endif |
14797e236
|
1741 |
|
56e49d218
|
1742 1743 |
/** * inactive_file_is_low - check if file pages need to be deactivated |
c56d5c7df
|
1744 |
* @lruvec: LRU vector to check |
56e49d218
|
1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 |
* * When the system is doing streaming IO, memory pressure here * ensures that active file pages get deactivated, until more * than half of the file pages are on the inactive list. * * Once we get to that situation, protect the system's working * set from being evicted by disabling active file page aging. * * This uses a different ratio than the anonymous pages, because * the page cache uses a use-once replacement algorithm. */ |
c56d5c7df
|
1756 |
static int inactive_file_is_low(struct lruvec *lruvec) |
56e49d218
|
1757 |
{ |
e3790144c
|
1758 1759 1760 1761 1762 |
unsigned long inactive; unsigned long active; inactive = get_lru_size(lruvec, LRU_INACTIVE_FILE); active = get_lru_size(lruvec, LRU_ACTIVE_FILE); |
56e49d218
|
1763 |
|
e3790144c
|
1764 |
return active > inactive; |
56e49d218
|
1765 |
} |
75b00af77
|
1766 |
static int inactive_list_is_low(struct lruvec *lruvec, enum lru_list lru) |
b39415b27
|
1767 |
{ |
75b00af77
|
1768 |
if (is_file_lru(lru)) |
c56d5c7df
|
1769 |
return inactive_file_is_low(lruvec); |
b39415b27
|
1770 |
else |
c56d5c7df
|
1771 |
return inactive_anon_is_low(lruvec); |
b39415b27
|
1772 |
} |
4f98a2fee
|
1773 |
static unsigned long shrink_list(enum lru_list lru, unsigned long nr_to_scan, |
1a93be0e7
|
1774 |
struct lruvec *lruvec, struct scan_control *sc) |
b69408e88
|
1775 |
{ |
b39415b27
|
1776 |
if (is_active_lru(lru)) { |
75b00af77
|
1777 |
if (inactive_list_is_low(lruvec, lru)) |
1a93be0e7
|
1778 |
shrink_active_list(nr_to_scan, lruvec, sc, lru); |
556adecba
|
1779 1780 |
return 0; } |
1a93be0e7
|
1781 |
return shrink_inactive_list(nr_to_scan, lruvec, sc, lru); |
4f98a2fee
|
1782 |
} |
9a2651140
|
1783 1784 1785 1786 1787 1788 |
enum scan_balance { SCAN_EQUAL, SCAN_FRACT, SCAN_ANON, SCAN_FILE, }; |
4f98a2fee
|
1789 1790 1791 1792 1793 1794 |
/* * Determine how aggressively the anon and file LRU lists should be * scanned. The relative value of each set of LRU lists is determined * by looking at the fraction of the pages scanned we did rotate back * onto the active list instead of evict. * |
be7bd59db
|
1795 1796 |
* nr[0] = anon inactive pages to scan; nr[1] = anon active pages to scan * nr[2] = file inactive pages to scan; nr[3] = file active pages to scan |
4f98a2fee
|
1797 |
*/ |
02695175c
|
1798 1799 |
static void get_scan_count(struct lruvec *lruvec, int swappiness, struct scan_control *sc, unsigned long *nr) |
4f98a2fee
|
1800 |
{ |
9a2651140
|
1801 1802 1803 1804 |
struct zone_reclaim_stat *reclaim_stat = &lruvec->reclaim_stat; u64 fraction[2]; u64 denominator = 0; /* gcc */ struct zone *zone = lruvec_zone(lruvec); |
4f98a2fee
|
1805 |
unsigned long anon_prio, file_prio; |
9a2651140
|
1806 |
enum scan_balance scan_balance; |
0bf1457f0
|
1807 |
unsigned long anon, file; |
9a2651140
|
1808 |
bool force_scan = false; |
4f98a2fee
|
1809 |
unsigned long ap, fp; |
4111304da
|
1810 |
enum lru_list lru; |
6f04f48dc
|
1811 1812 |
bool some_scanned; int pass; |
246e87a93
|
1813 |
|
f11c0ca50
|
1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 |
/* * If the zone or memcg is small, nr[l] can be 0. This * results in no scanning on this priority and a potential * priority drop. Global direct reclaim can go to the next * zone and tends to have no problems. Global kswapd is for * zone balancing and it needs to scan a minimum amount. When * reclaiming for a memcg, a priority drop can cause high * latencies, so it's better to scan a minimum amount there as * well. */ |
6e543d578
|
1824 |
if (current_is_kswapd() && !zone_reclaimable(zone)) |
a4d3e9e76
|
1825 |
force_scan = true; |
89b5fae53
|
1826 |
if (!global_reclaim(sc)) |
a4d3e9e76
|
1827 |
force_scan = true; |
76a33fc38
|
1828 1829 |
/* If we have no swap space, do not bother scanning anon pages. */ |
ec8acf20a
|
1830 |
if (!sc->may_swap || (get_nr_swap_pages() <= 0)) { |
9a2651140
|
1831 |
scan_balance = SCAN_FILE; |
76a33fc38
|
1832 1833 |
goto out; } |
4f98a2fee
|
1834 |
|
10316b313
|
1835 1836 1837 1838 1839 1840 1841 |
/* * Global reclaim will swap to prevent OOM even with no * swappiness, but memcg users want to use this knob to * disable swapping for individual groups completely when * using the memory controller's swap limit feature would be * too expensive. */ |
02695175c
|
1842 |
if (!global_reclaim(sc) && !swappiness) { |
9a2651140
|
1843 |
scan_balance = SCAN_FILE; |
10316b313
|
1844 1845 1846 1847 1848 1849 1850 1851 |
goto out; } /* * Do not apply any pressure balancing cleverness when the * system is close to OOM, scan both anon and file equally * (unless the swappiness setting disagrees with swapping). */ |
02695175c
|
1852 |
if (!sc->priority && swappiness) { |
9a2651140
|
1853 |
scan_balance = SCAN_EQUAL; |
10316b313
|
1854 1855 |
goto out; } |
11d16c25b
|
1856 |
/* |
623762517
|
1857 1858 1859 1860 1861 1862 1863 1864 1865 |
* Prevent the reclaimer from falling into the cache trap: as * cache pages start out inactive, every cache fault will tip * the scan balance towards the file LRU. And as the file LRU * shrinks, so does the window for rotation from references. * This means we have a runaway feedback loop where a tiny * thrashing file LRU becomes infinitely more attractive than * anon pages. Try to detect this based on file LRU size. */ if (global_reclaim(sc)) { |
2ab051e11
|
1866 1867 1868 1869 1870 1871 |
unsigned long zonefile; unsigned long zonefree; zonefree = zone_page_state(zone, NR_FREE_PAGES); zonefile = zone_page_state(zone, NR_ACTIVE_FILE) + zone_page_state(zone, NR_INACTIVE_FILE); |
623762517
|
1872 |
|
2ab051e11
|
1873 |
if (unlikely(zonefile + zonefree <= high_wmark_pages(zone))) { |
623762517
|
1874 1875 1876 1877 1878 1879 |
scan_balance = SCAN_ANON; goto out; } } /* |
7c5bd705d
|
1880 1881 1882 1883 |
* There is enough inactive page cache, do not reclaim * anything from the anonymous working set right now. */ if (!inactive_file_is_low(lruvec)) { |
9a2651140
|
1884 |
scan_balance = SCAN_FILE; |
7c5bd705d
|
1885 1886 |
goto out; } |
9a2651140
|
1887 |
scan_balance = SCAN_FRACT; |
7c5bd705d
|
1888 |
/* |
58c37f6e0
|
1889 1890 1891 |
* With swappiness at 100, anonymous and file have the same priority. * This scanning priority is essentially the inverse of IO cost. */ |
02695175c
|
1892 |
anon_prio = swappiness; |
75b00af77
|
1893 |
file_prio = 200 - anon_prio; |
58c37f6e0
|
1894 1895 |
/* |
4f98a2fee
|
1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 |
* OK, so we have swap space and a fair amount of page cache * pages. We use the recently rotated / recently scanned * ratios to determine how valuable each cache is. * * Because workloads change over time (and to avoid overflow) * we keep these statistics as a floating average, which ends * up weighing recent references more than old ones. * * anon in [0], file in [1] */ |
2ab051e11
|
1906 1907 1908 1909 1910 |
anon = get_lru_size(lruvec, LRU_ACTIVE_ANON) + get_lru_size(lruvec, LRU_INACTIVE_ANON); file = get_lru_size(lruvec, LRU_ACTIVE_FILE) + get_lru_size(lruvec, LRU_INACTIVE_FILE); |
90126375d
|
1911 |
spin_lock_irq(&zone->lru_lock); |
6e9015716
|
1912 |
if (unlikely(reclaim_stat->recent_scanned[0] > anon / 4)) { |
6e9015716
|
1913 1914 |
reclaim_stat->recent_scanned[0] /= 2; reclaim_stat->recent_rotated[0] /= 2; |
4f98a2fee
|
1915 |
} |
6e9015716
|
1916 |
if (unlikely(reclaim_stat->recent_scanned[1] > file / 4)) { |
6e9015716
|
1917 1918 |
reclaim_stat->recent_scanned[1] /= 2; reclaim_stat->recent_rotated[1] /= 2; |
4f98a2fee
|
1919 1920 1921 |
} /* |
00d8089c5
|
1922 1923 1924 |
* The amount of pressure on anon vs file pages is inversely * proportional to the fraction of recently scanned pages on * each list that were recently referenced and in active use. |
4f98a2fee
|
1925 |
*/ |
fe35004fb
|
1926 |
ap = anon_prio * (reclaim_stat->recent_scanned[0] + 1); |
6e9015716
|
1927 |
ap /= reclaim_stat->recent_rotated[0] + 1; |
4f98a2fee
|
1928 |
|
fe35004fb
|
1929 |
fp = file_prio * (reclaim_stat->recent_scanned[1] + 1); |
6e9015716
|
1930 |
fp /= reclaim_stat->recent_rotated[1] + 1; |
90126375d
|
1931 |
spin_unlock_irq(&zone->lru_lock); |
4f98a2fee
|
1932 |
|
76a33fc38
|
1933 1934 1935 1936 |
fraction[0] = ap; fraction[1] = fp; denominator = ap + fp + 1; out: |
6f04f48dc
|
1937 1938 1939 1940 1941 1942 1943 |
some_scanned = false; /* Only use force_scan on second pass. */ for (pass = 0; !some_scanned && pass < 2; pass++) { for_each_evictable_lru(lru) { int file = is_file_lru(lru); unsigned long size; unsigned long scan; |
6e08a369e
|
1944 |
|
6f04f48dc
|
1945 1946 |
size = get_lru_size(lruvec, lru); scan = size >> sc->priority; |
9a2651140
|
1947 |
|
6f04f48dc
|
1948 1949 |
if (!scan && pass && force_scan) scan = min(size, SWAP_CLUSTER_MAX); |
9a2651140
|
1950 |
|
6f04f48dc
|
1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 |
switch (scan_balance) { case SCAN_EQUAL: /* Scan lists relative to size */ break; case SCAN_FRACT: /* * Scan types proportional to swappiness and * their relative recent reclaim efficiency. */ scan = div64_u64(scan * fraction[file], denominator); break; case SCAN_FILE: case SCAN_ANON: /* Scan one type exclusively */ if ((scan_balance == SCAN_FILE) != file) scan = 0; break; default: /* Look ma, no brain */ BUG(); } nr[lru] = scan; |
9a2651140
|
1974 |
/* |
6f04f48dc
|
1975 1976 |
* Skip the second pass and don't force_scan, * if we found something to scan. |
9a2651140
|
1977 |
*/ |
6f04f48dc
|
1978 |
some_scanned |= !!scan; |
9a2651140
|
1979 |
} |
76a33fc38
|
1980 |
} |
6e08a369e
|
1981 |
} |
4f98a2fee
|
1982 |
|
9b4f98cda
|
1983 1984 1985 |
/* * This is a basic per-zone page freer. Used by both kswapd and direct reclaim. */ |
02695175c
|
1986 1987 |
static void shrink_lruvec(struct lruvec *lruvec, int swappiness, struct scan_control *sc) |
9b4f98cda
|
1988 1989 |
{ unsigned long nr[NR_LRU_LISTS]; |
e82e0561d
|
1990 |
unsigned long targets[NR_LRU_LISTS]; |
9b4f98cda
|
1991 1992 1993 1994 1995 |
unsigned long nr_to_scan; enum lru_list lru; unsigned long nr_reclaimed = 0; unsigned long nr_to_reclaim = sc->nr_to_reclaim; struct blk_plug plug; |
1a501907b
|
1996 |
bool scan_adjusted; |
9b4f98cda
|
1997 |
|
02695175c
|
1998 |
get_scan_count(lruvec, swappiness, sc, nr); |
9b4f98cda
|
1999 |
|
e82e0561d
|
2000 2001 |
/* Record the original scan target for proportional adjustments later */ memcpy(targets, nr, sizeof(nr)); |
1a501907b
|
2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 |
/* * Global reclaiming within direct reclaim at DEF_PRIORITY is a normal * event that can occur when there is little memory pressure e.g. * multiple streaming readers/writers. Hence, we do not abort scanning * when the requested number of pages are reclaimed when scanning at * DEF_PRIORITY on the assumption that the fact we are direct * reclaiming implies that kswapd is not keeping up and it is best to * do a batch of work at once. For memcg reclaim one check is made to * abort proportional reclaim if either the file or anon lru has already * dropped to zero at the first pass. */ scan_adjusted = (global_reclaim(sc) && !current_is_kswapd() && sc->priority == DEF_PRIORITY); |
9b4f98cda
|
2015 2016 2017 |
blk_start_plug(&plug); while (nr[LRU_INACTIVE_ANON] || nr[LRU_ACTIVE_FILE] || nr[LRU_INACTIVE_FILE]) { |
e82e0561d
|
2018 2019 |
unsigned long nr_anon, nr_file, percentage; unsigned long nr_scanned; |
9b4f98cda
|
2020 2021 2022 2023 2024 2025 2026 2027 2028 |
for_each_evictable_lru(lru) { if (nr[lru]) { nr_to_scan = min(nr[lru], SWAP_CLUSTER_MAX); nr[lru] -= nr_to_scan; nr_reclaimed += shrink_list(lru, nr_to_scan, lruvec, sc); } } |
e82e0561d
|
2029 2030 2031 |
if (nr_reclaimed < nr_to_reclaim || scan_adjusted) continue; |
9b4f98cda
|
2032 |
/* |
e82e0561d
|
2033 |
* For kswapd and memcg, reclaim at least the number of pages |
1a501907b
|
2034 |
* requested. Ensure that the anon and file LRUs are scanned |
e82e0561d
|
2035 2036 2037 2038 2039 2040 |
* proportionally what was requested by get_scan_count(). We * stop reclaiming one LRU and reduce the amount scanning * proportional to the original scan target. */ nr_file = nr[LRU_INACTIVE_FILE] + nr[LRU_ACTIVE_FILE]; nr_anon = nr[LRU_INACTIVE_ANON] + nr[LRU_ACTIVE_ANON]; |
1a501907b
|
2041 2042 2043 2044 2045 2046 2047 2048 |
/* * It's just vindictive to attack the larger once the smaller * has gone to zero. And given the way we stop scanning the * smaller below, this makes sure that we only make one nudge * towards proportionality once we've got nr_to_reclaim. */ if (!nr_file || !nr_anon) break; |
e82e0561d
|
2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 |
if (nr_file > nr_anon) { unsigned long scan_target = targets[LRU_INACTIVE_ANON] + targets[LRU_ACTIVE_ANON] + 1; lru = LRU_BASE; percentage = nr_anon * 100 / scan_target; } else { unsigned long scan_target = targets[LRU_INACTIVE_FILE] + targets[LRU_ACTIVE_FILE] + 1; lru = LRU_FILE; percentage = nr_file * 100 / scan_target; } /* Stop scanning the smaller of the LRU */ nr[lru] = 0; nr[lru + LRU_ACTIVE] = 0; /* * Recalculate the other LRU scan count based on its original * scan target and the percentage scanning already complete */ lru = (lru == LRU_FILE) ? LRU_BASE : LRU_FILE; nr_scanned = targets[lru] - nr[lru]; nr[lru] = targets[lru] * (100 - percentage) / 100; nr[lru] -= min(nr[lru], nr_scanned); lru += LRU_ACTIVE; nr_scanned = targets[lru] - nr[lru]; nr[lru] = targets[lru] * (100 - percentage) / 100; nr[lru] -= min(nr[lru], nr_scanned); scan_adjusted = true; |
9b4f98cda
|
2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 |
} blk_finish_plug(&plug); sc->nr_reclaimed += nr_reclaimed; /* * Even if we did not try to evict anon pages at all, we want to * rebalance the anon lru active/inactive ratio. */ if (inactive_anon_is_low(lruvec)) shrink_active_list(SWAP_CLUSTER_MAX, lruvec, sc, LRU_ACTIVE_ANON); throttle_vm_writeout(sc->gfp_mask); } |
23b9da55c
|
2094 |
/* Use reclaim/compaction for costly allocs or under memory pressure */ |
9e3b2f8cd
|
2095 |
static bool in_reclaim_compaction(struct scan_control *sc) |
23b9da55c
|
2096 |
{ |
d84da3f9e
|
2097 |
if (IS_ENABLED(CONFIG_COMPACTION) && sc->order && |
23b9da55c
|
2098 |
(sc->order > PAGE_ALLOC_COSTLY_ORDER || |
9e3b2f8cd
|
2099 |
sc->priority < DEF_PRIORITY - 2)) |
23b9da55c
|
2100 2101 2102 2103 |
return true; return false; } |
1da177e4c
|
2104 |
/* |
23b9da55c
|
2105 2106 2107 2108 2109 |
* Reclaim/compaction is used for high-order allocation requests. It reclaims * order-0 pages before compacting the zone. should_continue_reclaim() returns * true if more pages should be reclaimed such that when the page allocator * calls try_to_compact_zone() that it will have enough free pages to succeed. * It will give up earlier than that if there is difficulty reclaiming pages. |
3e7d34497
|
2110 |
*/ |
9b4f98cda
|
2111 |
static inline bool should_continue_reclaim(struct zone *zone, |
3e7d34497
|
2112 2113 2114 2115 2116 2117 2118 2119 |
unsigned long nr_reclaimed, unsigned long nr_scanned, struct scan_control *sc) { unsigned long pages_for_compaction; unsigned long inactive_lru_pages; /* If not in reclaim/compaction mode, stop */ |
9e3b2f8cd
|
2120 |
if (!in_reclaim_compaction(sc)) |
3e7d34497
|
2121 |
return false; |
2876592f2
|
2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 |
/* Consider stopping depending on scan and reclaim activity */ if (sc->gfp_mask & __GFP_REPEAT) { /* * For __GFP_REPEAT allocations, stop reclaiming if the * full LRU list has been scanned and we are still failing * to reclaim pages. This full LRU scan is potentially * expensive but a __GFP_REPEAT caller really wants to succeed */ if (!nr_reclaimed && !nr_scanned) return false; } else { /* * For non-__GFP_REPEAT allocations which can presumably * fail without consequence, stop if we failed to reclaim * any pages from the last SWAP_CLUSTER_MAX number of * pages that were scanned. This will return to the * caller faster at the risk reclaim/compaction and * the resulting allocation attempt fails */ if (!nr_reclaimed) return false; } |
3e7d34497
|
2144 2145 2146 2147 2148 2149 |
/* * If we have not reclaimed enough pages for compaction and the * inactive lists are large enough, continue reclaiming */ pages_for_compaction = (2UL << sc->order); |
9b4f98cda
|
2150 |
inactive_lru_pages = zone_page_state(zone, NR_INACTIVE_FILE); |
ec8acf20a
|
2151 |
if (get_nr_swap_pages() > 0) |
9b4f98cda
|
2152 |
inactive_lru_pages += zone_page_state(zone, NR_INACTIVE_ANON); |
3e7d34497
|
2153 2154 2155 2156 2157 |
if (sc->nr_reclaimed < pages_for_compaction && inactive_lru_pages > pages_for_compaction) return true; /* If compaction would go ahead or the allocation would succeed, stop */ |
9b4f98cda
|
2158 |
switch (compaction_suitable(zone, sc->order)) { |
3e7d34497
|
2159 2160 2161 2162 2163 2164 2165 |
case COMPACT_PARTIAL: case COMPACT_CONTINUE: return false; default: return true; } } |
2344d7e44
|
2166 |
static bool shrink_zone(struct zone *zone, struct scan_control *sc) |
1da177e4c
|
2167 |
{ |
f0fdc5e8e
|
2168 |
unsigned long nr_reclaimed, nr_scanned; |
2344d7e44
|
2169 |
bool reclaimable = false; |
1da177e4c
|
2170 |
|
9b4f98cda
|
2171 2172 2173 2174 2175 2176 |
do { struct mem_cgroup *root = sc->target_mem_cgroup; struct mem_cgroup_reclaim_cookie reclaim = { .zone = zone, .priority = sc->priority, }; |
694fbc0fe
|
2177 |
struct mem_cgroup *memcg; |
3e7d34497
|
2178 |
|
9b4f98cda
|
2179 2180 |
nr_reclaimed = sc->nr_reclaimed; nr_scanned = sc->nr_scanned; |
1da177e4c
|
2181 |
|
694fbc0fe
|
2182 2183 |
memcg = mem_cgroup_iter(root, NULL, &reclaim); do { |
9b4f98cda
|
2184 |
struct lruvec *lruvec; |
02695175c
|
2185 |
int swappiness; |
5660048cc
|
2186 |
|
9b4f98cda
|
2187 |
lruvec = mem_cgroup_zone_lruvec(zone, memcg); |
02695175c
|
2188 |
swappiness = mem_cgroup_swappiness(memcg); |
f9be23d6d
|
2189 |
|
02695175c
|
2190 |
shrink_lruvec(lruvec, swappiness, sc); |
f16015fbf
|
2191 |
|
9b4f98cda
|
2192 |
/* |
a394cb8ee
|
2193 2194 |
* Direct reclaim and kswapd have to scan all memory * cgroups to fulfill the overall scan target for the |
9b4f98cda
|
2195 |
* zone. |
a394cb8ee
|
2196 2197 2198 2199 2200 |
* * Limit reclaim, on the other hand, only cares about * nr_to_reclaim pages to be reclaimed and it will * retry with decreasing priority if one round over the * whole hierarchy is not sufficient. |
9b4f98cda
|
2201 |
*/ |
a394cb8ee
|
2202 2203 |
if (!global_reclaim(sc) && sc->nr_reclaimed >= sc->nr_to_reclaim) { |
9b4f98cda
|
2204 2205 2206 |
mem_cgroup_iter_break(root, memcg); break; } |
694fbc0fe
|
2207 2208 |
memcg = mem_cgroup_iter(root, memcg, &reclaim); } while (memcg); |
70ddf637e
|
2209 2210 2211 2212 |
vmpressure(sc->gfp_mask, sc->target_mem_cgroup, sc->nr_scanned - nr_scanned, sc->nr_reclaimed - nr_reclaimed); |
2344d7e44
|
2213 2214 |
if (sc->nr_reclaimed - nr_reclaimed) reclaimable = true; |
9b4f98cda
|
2215 2216 |
} while (should_continue_reclaim(zone, sc->nr_reclaimed - nr_reclaimed, sc->nr_scanned - nr_scanned, sc)); |
2344d7e44
|
2217 2218 |
return reclaimable; |
f16015fbf
|
2219 |
} |
53853e2d2
|
2220 2221 2222 2223 |
/* * Returns true if compaction should go ahead for a high-order request, or * the high-order allocation would succeed without compaction. */ |
0b06496a3
|
2224 |
static inline bool compaction_ready(struct zone *zone, int order) |
fe4b1b244
|
2225 2226 2227 |
{ unsigned long balance_gap, watermark; bool watermark_ok; |
fe4b1b244
|
2228 2229 2230 2231 2232 2233 |
/* * Compaction takes time to run and there are potentially other * callers using the pages just freed. Continue reclaiming until * there is a buffer of free pages available to give compaction * a reasonable chance of completing and allocating the page */ |
4be89a346
|
2234 2235 |
balance_gap = min(low_wmark_pages(zone), DIV_ROUND_UP( zone->managed_pages, KSWAPD_ZONE_BALANCE_GAP_RATIO)); |
0b06496a3
|
2236 |
watermark = high_wmark_pages(zone) + balance_gap + (2UL << order); |
fe4b1b244
|
2237 2238 2239 2240 2241 2242 |
watermark_ok = zone_watermark_ok_safe(zone, 0, watermark, 0, 0); /* * If compaction is deferred, reclaim up to a point where * compaction will have a chance of success when re-enabled */ |
0b06496a3
|
2243 |
if (compaction_deferred(zone, order)) |
fe4b1b244
|
2244 |
return watermark_ok; |
53853e2d2
|
2245 2246 2247 2248 2249 |
/* * If compaction is not ready to start and allocation is not likely * to succeed without it, then keep reclaiming. */ if (compaction_suitable(zone, order) == COMPACT_SKIPPED) |
fe4b1b244
|
2250 2251 2252 2253 |
return false; return watermark_ok; } |
1da177e4c
|
2254 2255 2256 2257 2258 |
/* * This is the direct reclaim path, for page-allocating processes. We only * try to reclaim pages from zones which will satisfy the caller's allocation * request. * |
418589663
|
2259 2260 |
* We reclaim from a zone even if that zone is over high_wmark_pages(zone). * Because: |
1da177e4c
|
2261 2262 |
* a) The caller may be trying to free *extra* pages to satisfy a higher-order * allocation or |
418589663
|
2263 2264 2265 |
* b) The target zone may be at high_wmark_pages(zone) but the lower zones * must go *over* high_wmark_pages(zone) to satisfy the `incremental min' * zone defense algorithm. |
1da177e4c
|
2266 |
* |
1da177e4c
|
2267 2268 |
* If a zone is deemed to be full of pinned pages then just give it a light * scan then give up on it. |
2344d7e44
|
2269 2270 |
* * Returns true if a zone was reclaimable. |
1da177e4c
|
2271 |
*/ |
2344d7e44
|
2272 |
static bool shrink_zones(struct zonelist *zonelist, struct scan_control *sc) |
1da177e4c
|
2273 |
{ |
dd1a239f6
|
2274 |
struct zoneref *z; |
54a6eb5c4
|
2275 |
struct zone *zone; |
0608f43da
|
2276 2277 |
unsigned long nr_soft_reclaimed; unsigned long nr_soft_scanned; |
65ec02cb9
|
2278 |
unsigned long lru_pages = 0; |
65ec02cb9
|
2279 |
struct reclaim_state *reclaim_state = current->reclaim_state; |
619d0d76c
|
2280 |
gfp_t orig_mask; |
3115cd914
|
2281 2282 2283 |
struct shrink_control shrink = { .gfp_mask = sc->gfp_mask, }; |
9bbc04eeb
|
2284 |
enum zone_type requested_highidx = gfp_zone(sc->gfp_mask); |
2344d7e44
|
2285 |
bool reclaimable = false; |
1cfb419b3
|
2286 |
|
cc715d99e
|
2287 2288 2289 2290 2291 |
/* * If the number of buffer_heads in the machine exceeds the maximum * allowed level, force direct reclaim to scan the highmem zone as * highmem pages could be pinning lowmem pages storing buffer_heads */ |
619d0d76c
|
2292 |
orig_mask = sc->gfp_mask; |
cc715d99e
|
2293 2294 |
if (buffer_heads_over_limit) sc->gfp_mask |= __GFP_HIGHMEM; |
3115cd914
|
2295 |
nodes_clear(shrink.nodes_to_scan); |
65ec02cb9
|
2296 |
|
d4debc66d
|
2297 2298 |
for_each_zone_zonelist_nodemask(zone, z, zonelist, gfp_zone(sc->gfp_mask), sc->nodemask) { |
f3fe65122
|
2299 |
if (!populated_zone(zone)) |
1da177e4c
|
2300 |
continue; |
1cfb419b3
|
2301 2302 2303 2304 |
/* * Take care memory controller reclaiming has small influence * to global LRU. */ |
89b5fae53
|
2305 |
if (global_reclaim(sc)) { |
1cfb419b3
|
2306 2307 |
if (!cpuset_zone_allowed_hardwall(zone, GFP_KERNEL)) continue; |
65ec02cb9
|
2308 2309 |
lru_pages += zone_reclaimable_pages(zone); |
3115cd914
|
2310 |
node_set(zone_to_nid(zone), shrink.nodes_to_scan); |
65ec02cb9
|
2311 |
|
6e543d578
|
2312 2313 |
if (sc->priority != DEF_PRIORITY && !zone_reclaimable(zone)) |
1cfb419b3
|
2314 |
continue; /* Let kswapd poll it */ |
0b06496a3
|
2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 |
/* * If we already have plenty of memory free for * compaction in this zone, don't free any more. * Even though compaction is invoked for any * non-zero order, only frequent costly order * reclamation is disruptive enough to become a * noticeable problem, like transparent huge * page allocations. */ if (IS_ENABLED(CONFIG_COMPACTION) && sc->order > PAGE_ALLOC_COSTLY_ORDER && zonelist_zone_idx(z) <= requested_highidx && compaction_ready(zone, sc->order)) { sc->compaction_ready = true; continue; |
e0887c19b
|
2331 |
} |
0b06496a3
|
2332 |
|
0608f43da
|
2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 |
/* * This steals pages from memory cgroups over softlimit * and returns the number of reclaimed pages and * scanned pages. This works for global memory pressure * and balancing, not for a memcg's limit. */ nr_soft_scanned = 0; nr_soft_reclaimed = mem_cgroup_soft_limit_reclaim(zone, sc->order, sc->gfp_mask, &nr_soft_scanned); sc->nr_reclaimed += nr_soft_reclaimed; sc->nr_scanned += nr_soft_scanned; |
2344d7e44
|
2345 2346 |
if (nr_soft_reclaimed) reclaimable = true; |
ac34a1a3c
|
2347 |
/* need some check for avoid more shrink_zone() */ |
1cfb419b3
|
2348 |
} |
408d85441
|
2349 |
|
2344d7e44
|
2350 2351 2352 2353 2354 2355 |
if (shrink_zone(zone, sc)) reclaimable = true; if (global_reclaim(sc) && !reclaimable && zone_reclaimable(zone)) reclaimable = true; |
1da177e4c
|
2356 |
} |
e0c23279c
|
2357 |
|
65ec02cb9
|
2358 2359 2360 2361 2362 2363 2364 |
/* * Don't shrink slabs when reclaiming memory from over limit cgroups * but do shrink slab at least once when aborting reclaim for * compaction to avoid unevenly scanning file/anon LRU pages over slab * pages. */ if (global_reclaim(sc)) { |
3115cd914
|
2365 |
shrink_slab(&shrink, sc->nr_scanned, lru_pages); |
65ec02cb9
|
2366 2367 2368 2369 2370 |
if (reclaim_state) { sc->nr_reclaimed += reclaim_state->reclaimed_slab; reclaim_state->reclaimed_slab = 0; } } |
619d0d76c
|
2371 2372 2373 2374 2375 |
/* * Restore to original mask to avoid the impact on the caller if we * promoted it to __GFP_HIGHMEM. */ sc->gfp_mask = orig_mask; |
d1908362a
|
2376 |
|
2344d7e44
|
2377 |
return reclaimable; |
1da177e4c
|
2378 |
} |
4f98a2fee
|
2379 |
|
1da177e4c
|
2380 2381 2382 2383 2384 2385 2386 2387 |
/* * This is the main entry point to direct page reclaim. * * If a full scan of the inactive list fails to free enough memory then we * are "out of memory" and something needs to be killed. * * If the caller is !__GFP_FS then the probability of a failure is reasonably * high - the zone may be full of dirty or under-writeback pages, which this |
5b0830cb9
|
2388 2389 2390 2391 |
* caller can't do much about. We kick the writeback threads and take explicit * naps in the hope that some of these pages can be written. But if the * allocating task holds filesystem locks which prevent writeout this might not * work, and the allocation attempt will fail. |
a41f24ea9
|
2392 2393 2394 |
* * returns: 0, if no pages reclaimed * else, the number of pages reclaimed |
1da177e4c
|
2395 |
*/ |
dac1d27bc
|
2396 |
static unsigned long do_try_to_free_pages(struct zonelist *zonelist, |
3115cd914
|
2397 |
struct scan_control *sc) |
1da177e4c
|
2398 |
{ |
69e05944a
|
2399 |
unsigned long total_scanned = 0; |
22fba3354
|
2400 |
unsigned long writeback_threshold; |
2344d7e44
|
2401 |
bool zones_reclaimable; |
1da177e4c
|
2402 |
|
873b47717
|
2403 |
delayacct_freepages_start(); |
89b5fae53
|
2404 |
if (global_reclaim(sc)) |
1cfb419b3
|
2405 |
count_vm_event(ALLOCSTALL); |
1da177e4c
|
2406 |
|
9e3b2f8cd
|
2407 |
do { |
70ddf637e
|
2408 2409 |
vmpressure_prio(sc->gfp_mask, sc->target_mem_cgroup, sc->priority); |
66e1707bc
|
2410 |
sc->nr_scanned = 0; |
2344d7e44
|
2411 |
zones_reclaimable = shrink_zones(zonelist, sc); |
c6a8a8c58
|
2412 |
|
66e1707bc
|
2413 |
total_scanned += sc->nr_scanned; |
bb21c7ce1
|
2414 |
if (sc->nr_reclaimed >= sc->nr_to_reclaim) |
0b06496a3
|
2415 2416 2417 2418 |
break; if (sc->compaction_ready) break; |
1da177e4c
|
2419 2420 |
/* |
0e50ce3b5
|
2421 2422 2423 2424 2425 2426 2427 |
* If we're getting trouble reclaiming, start doing * writepage even in laptop mode. */ if (sc->priority < DEF_PRIORITY - 2) sc->may_writepage = 1; /* |
1da177e4c
|
2428 2429 2430 2431 2432 2433 |
* Try to write back as many pages as we just scanned. This * tends to cause slow streaming writers to write data to the * disk smoothly, at the dirtying rate, which is nice. But * that's undesirable in laptop mode, where we *want* lumpy * writeout. So in laptop mode, write out the whole world. */ |
22fba3354
|
2434 2435 |
writeback_threshold = sc->nr_to_reclaim + sc->nr_to_reclaim / 2; if (total_scanned > writeback_threshold) { |
0e175a183
|
2436 2437 |
wakeup_flusher_threads(laptop_mode ? 0 : total_scanned, WB_REASON_TRY_TO_FREE_PAGES); |
66e1707bc
|
2438 |
sc->may_writepage = 1; |
1da177e4c
|
2439 |
} |
0b06496a3
|
2440 |
} while (--sc->priority >= 0); |
bb21c7ce1
|
2441 |
|
873b47717
|
2442 |
delayacct_freepages_end(); |
bb21c7ce1
|
2443 2444 |
if (sc->nr_reclaimed) return sc->nr_reclaimed; |
0cee34fd7
|
2445 |
/* Aborted reclaim to try compaction? don't OOM, then */ |
0b06496a3
|
2446 |
if (sc->compaction_ready) |
7335084d4
|
2447 |
return 1; |
2344d7e44
|
2448 2449 |
/* Any of the zones still reclaimable? Don't OOM. */ if (zones_reclaimable) |
bb21c7ce1
|
2450 2451 2452 |
return 1; return 0; |
1da177e4c
|
2453 |
} |
5515061d2
|
2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 |
static bool pfmemalloc_watermark_ok(pg_data_t *pgdat) { struct zone *zone; unsigned long pfmemalloc_reserve = 0; unsigned long free_pages = 0; int i; bool wmark_ok; for (i = 0; i <= ZONE_NORMAL; i++) { zone = &pgdat->node_zones[i]; |
675becce1
|
2464 2465 |
if (!populated_zone(zone)) continue; |
5515061d2
|
2466 2467 2468 |
pfmemalloc_reserve += min_wmark_pages(zone); free_pages += zone_page_state(zone, NR_FREE_PAGES); } |
675becce1
|
2469 2470 2471 |
/* If there are no reserves (unexpected config) then do not throttle */ if (!pfmemalloc_reserve) return true; |
5515061d2
|
2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 2483 2484 2485 2486 2487 |
wmark_ok = free_pages > pfmemalloc_reserve / 2; /* kswapd must be awake if processes are being throttled */ if (!wmark_ok && waitqueue_active(&pgdat->kswapd_wait)) { pgdat->classzone_idx = min(pgdat->classzone_idx, (enum zone_type)ZONE_NORMAL); wake_up_interruptible(&pgdat->kswapd_wait); } return wmark_ok; } /* * Throttle direct reclaimers if backing storage is backed by the network * and the PFMEMALLOC reserve for the preferred node is getting dangerously * depleted. kswapd will continue to make progress and wake the processes |
50694c28f
|
2488 2489 2490 2491 |
* when the low watermark is reached. * * Returns true if a fatal signal was delivered during throttling. If this * happens, the page allocator should not consider triggering the OOM killer. |
5515061d2
|
2492 |
*/ |
50694c28f
|
2493 |
static bool throttle_direct_reclaim(gfp_t gfp_mask, struct zonelist *zonelist, |
5515061d2
|
2494 2495 |
nodemask_t *nodemask) { |
675becce1
|
2496 |
struct zoneref *z; |
5515061d2
|
2497 |
struct zone *zone; |
675becce1
|
2498 |
pg_data_t *pgdat = NULL; |
5515061d2
|
2499 2500 2501 2502 2503 2504 2505 2506 2507 |
/* * Kernel threads should not be throttled as they may be indirectly * responsible for cleaning pages necessary for reclaim to make forward * progress. kjournald for example may enter direct reclaim while * committing a transaction where throttling it could forcing other * processes to block on log_wait_commit(). */ if (current->flags & PF_KTHREAD) |
50694c28f
|
2508 2509 2510 2511 2512 2513 2514 2515 |
goto out; /* * If a fatal signal is pending, this process should not throttle. * It should return quickly so it can exit and free its memory */ if (fatal_signal_pending(current)) goto out; |
5515061d2
|
2516 |
|
675becce1
|
2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 2527 2528 2529 2530 2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 |
/* * Check if the pfmemalloc reserves are ok by finding the first node * with a usable ZONE_NORMAL or lower zone. The expectation is that * GFP_KERNEL will be required for allocating network buffers when * swapping over the network so ZONE_HIGHMEM is unusable. * * Throttling is based on the first usable node and throttled processes * wait on a queue until kswapd makes progress and wakes them. There * is an affinity then between processes waking up and where reclaim * progress has been made assuming the process wakes on the same node. * More importantly, processes running on remote nodes will not compete * for remote pfmemalloc reserves and processes on different nodes * should make reasonable progress. */ for_each_zone_zonelist_nodemask(zone, z, zonelist, gfp_mask, nodemask) { if (zone_idx(zone) > ZONE_NORMAL) continue; /* Throttle based on the first usable node */ pgdat = zone->zone_pgdat; if (pfmemalloc_watermark_ok(pgdat)) goto out; break; } /* If no zone was usable by the allocation flags then do not throttle */ if (!pgdat) |
50694c28f
|
2545 |
goto out; |
5515061d2
|
2546 |
|
68243e76e
|
2547 2548 |
/* Account for the throttling */ count_vm_event(PGSCAN_DIRECT_THROTTLE); |
5515061d2
|
2549 2550 2551 2552 2553 2554 2555 2556 2557 2558 2559 |
/* * If the caller cannot enter the filesystem, it's possible that it * is due to the caller holding an FS lock or performing a journal * transaction in the case of a filesystem like ext[3|4]. In this case, * it is not safe to block on pfmemalloc_wait as kswapd could be * blocked waiting on the same lock. Instead, throttle for up to a * second before continuing. */ if (!(gfp_mask & __GFP_FS)) { wait_event_interruptible_timeout(pgdat->pfmemalloc_wait, pfmemalloc_watermark_ok(pgdat), HZ); |
50694c28f
|
2560 2561 |
goto check_pending; |
5515061d2
|
2562 2563 2564 2565 2566 |
} /* Throttle until kswapd wakes the process */ wait_event_killable(zone->zone_pgdat->pfmemalloc_wait, pfmemalloc_watermark_ok(pgdat)); |
50694c28f
|
2567 2568 2569 2570 2571 2572 2573 |
check_pending: if (fatal_signal_pending(current)) return true; out: return false; |
5515061d2
|
2574 |
} |
dac1d27bc
|
2575 |
unsigned long try_to_free_pages(struct zonelist *zonelist, int order, |
327c0e968
|
2576 |
gfp_t gfp_mask, nodemask_t *nodemask) |
66e1707bc
|
2577 |
{ |
33906bc5c
|
2578 |
unsigned long nr_reclaimed; |
66e1707bc
|
2579 |
struct scan_control sc = { |
ee814fe23
|
2580 |
.nr_to_reclaim = SWAP_CLUSTER_MAX, |
21caf2fc1
|
2581 |
.gfp_mask = (gfp_mask = memalloc_noio_flags(gfp_mask)), |
ee814fe23
|
2582 2583 2584 |
.order = order, .nodemask = nodemask, .priority = DEF_PRIORITY, |
66e1707bc
|
2585 |
.may_writepage = !laptop_mode, |
a6dc60f89
|
2586 |
.may_unmap = 1, |
2e2e42598
|
2587 |
.may_swap = 1, |
66e1707bc
|
2588 |
}; |
5515061d2
|
2589 |
/* |
50694c28f
|
2590 2591 2592 |
* Do not enter reclaim if fatal signal was delivered while throttled. * 1 is returned so that the page allocator does not OOM kill at this * point. |
5515061d2
|
2593 |
*/ |
50694c28f
|
2594 |
if (throttle_direct_reclaim(gfp_mask, zonelist, nodemask)) |
5515061d2
|
2595 |
return 1; |
33906bc5c
|
2596 2597 2598 |
trace_mm_vmscan_direct_reclaim_begin(order, sc.may_writepage, gfp_mask); |
3115cd914
|
2599 |
nr_reclaimed = do_try_to_free_pages(zonelist, &sc); |
33906bc5c
|
2600 2601 2602 2603 |
trace_mm_vmscan_direct_reclaim_end(nr_reclaimed); return nr_reclaimed; |
66e1707bc
|
2604 |
} |
c255a4580
|
2605 |
#ifdef CONFIG_MEMCG |
66e1707bc
|
2606 |
|
72835c86c
|
2607 |
unsigned long mem_cgroup_shrink_node_zone(struct mem_cgroup *memcg, |
4e4169535
|
2608 |
gfp_t gfp_mask, bool noswap, |
0ae5e89c6
|
2609 2610 |
struct zone *zone, unsigned long *nr_scanned) |
4e4169535
|
2611 2612 |
{ struct scan_control sc = { |
b8f5c5664
|
2613 |
.nr_to_reclaim = SWAP_CLUSTER_MAX, |
ee814fe23
|
2614 |
.target_mem_cgroup = memcg, |
4e4169535
|
2615 2616 2617 |
.may_writepage = !laptop_mode, .may_unmap = 1, .may_swap = !noswap, |
4e4169535
|
2618 |
}; |
f9be23d6d
|
2619 |
struct lruvec *lruvec = mem_cgroup_zone_lruvec(zone, memcg); |
02695175c
|
2620 |
int swappiness = mem_cgroup_swappiness(memcg); |
0ae5e89c6
|
2621 |
|
4e4169535
|
2622 2623 |
sc.gfp_mask = (gfp_mask & GFP_RECLAIM_MASK) | (GFP_HIGHUSER_MOVABLE & ~GFP_RECLAIM_MASK); |
bdce6d9eb
|
2624 |
|
9e3b2f8cd
|
2625 |
trace_mm_vmscan_memcg_softlimit_reclaim_begin(sc.order, |
bdce6d9eb
|
2626 2627 |
sc.may_writepage, sc.gfp_mask); |
4e4169535
|
2628 2629 2630 2631 2632 2633 2634 |
/* * NOTE: Although we can get the priority field, using it * here is not a good idea, since it limits the pages we can scan. * if we don't reclaim here, the shrink_zone from balance_pgdat * will pick up pages from other mem cgroup's as well. We hack * the priority and make it zero. */ |
02695175c
|
2635 |
shrink_lruvec(lruvec, swappiness, &sc); |
bdce6d9eb
|
2636 2637 |
trace_mm_vmscan_memcg_softlimit_reclaim_end(sc.nr_reclaimed); |
0ae5e89c6
|
2638 |
*nr_scanned = sc.nr_scanned; |
4e4169535
|
2639 2640 |
return sc.nr_reclaimed; } |
72835c86c
|
2641 |
unsigned long try_to_free_mem_cgroup_pages(struct mem_cgroup *memcg, |
b70a2a21d
|
2642 |
unsigned long nr_pages, |
a7885eb8a
|
2643 |
gfp_t gfp_mask, |
b70a2a21d
|
2644 |
bool may_swap) |
66e1707bc
|
2645 |
{ |
4e4169535
|
2646 |
struct zonelist *zonelist; |
bdce6d9eb
|
2647 |
unsigned long nr_reclaimed; |
889976dbc
|
2648 |
int nid; |
66e1707bc
|
2649 |
struct scan_control sc = { |
b70a2a21d
|
2650 |
.nr_to_reclaim = max(nr_pages, SWAP_CLUSTER_MAX), |
a09ed5e00
|
2651 2652 |
.gfp_mask = (gfp_mask & GFP_RECLAIM_MASK) | (GFP_HIGHUSER_MOVABLE & ~GFP_RECLAIM_MASK), |
ee814fe23
|
2653 2654 2655 2656 |
.target_mem_cgroup = memcg, .priority = DEF_PRIORITY, .may_writepage = !laptop_mode, .may_unmap = 1, |
b70a2a21d
|
2657 |
.may_swap = may_swap, |
a09ed5e00
|
2658 |
}; |
66e1707bc
|
2659 |
|
889976dbc
|
2660 2661 2662 2663 2664 |
/* * Unlike direct reclaim via alloc_pages(), memcg's reclaim doesn't * take care of from where we get pages. So the node where we start the * scan does not need to be the current node. */ |
72835c86c
|
2665 |
nid = mem_cgroup_select_victim_node(memcg); |
889976dbc
|
2666 2667 |
zonelist = NODE_DATA(nid)->node_zonelists; |
bdce6d9eb
|
2668 2669 2670 2671 |
trace_mm_vmscan_memcg_reclaim_begin(0, sc.may_writepage, sc.gfp_mask); |
3115cd914
|
2672 |
nr_reclaimed = do_try_to_free_pages(zonelist, &sc); |
bdce6d9eb
|
2673 2674 2675 2676 |
trace_mm_vmscan_memcg_reclaim_end(nr_reclaimed); return nr_reclaimed; |
66e1707bc
|
2677 2678 |
} #endif |
9e3b2f8cd
|
2679 |
static void age_active_anon(struct zone *zone, struct scan_control *sc) |
f16015fbf
|
2680 |
{ |
b95a2f2d4
|
2681 |
struct mem_cgroup *memcg; |
f16015fbf
|
2682 |
|
b95a2f2d4
|
2683 2684 2685 2686 2687 |
if (!total_swap_pages) return; memcg = mem_cgroup_iter(NULL, NULL, NULL); do { |
c56d5c7df
|
2688 |
struct lruvec *lruvec = mem_cgroup_zone_lruvec(zone, memcg); |
b95a2f2d4
|
2689 |
|
c56d5c7df
|
2690 |
if (inactive_anon_is_low(lruvec)) |
1a93be0e7
|
2691 |
shrink_active_list(SWAP_CLUSTER_MAX, lruvec, |
9e3b2f8cd
|
2692 |
sc, LRU_ACTIVE_ANON); |
b95a2f2d4
|
2693 2694 2695 |
memcg = mem_cgroup_iter(NULL, memcg, NULL); } while (memcg); |
f16015fbf
|
2696 |
} |
60cefed48
|
2697 2698 2699 2700 2701 2702 |
static bool zone_balanced(struct zone *zone, int order, unsigned long balance_gap, int classzone_idx) { if (!zone_watermark_ok_safe(zone, order, high_wmark_pages(zone) + balance_gap, classzone_idx, 0)) return false; |
d84da3f9e
|
2703 |
if (IS_ENABLED(CONFIG_COMPACTION) && order && |
53853e2d2
|
2704 |
compaction_suitable(zone, order) == COMPACT_SKIPPED) |
60cefed48
|
2705 2706 2707 2708 |
return false; return true; } |
1741c8775
|
2709 |
/* |
4ae0a48b5
|
2710 2711 2712 2713 2714 2715 2716 2717 2718 2719 |
* pgdat_balanced() is used when checking if a node is balanced. * * For order-0, all zones must be balanced! * * For high-order allocations only zones that meet watermarks and are in a * zone allowed by the callers classzone_idx are added to balanced_pages. The * total of balanced pages must be at least 25% of the zones allowed by * classzone_idx for the node to be considered balanced. Forcing all zones to * be balanced for high orders can cause excessive reclaim when there are * imbalanced zones. |
1741c8775
|
2720 2721 2722 2723 |
* The choice of 25% is due to * o a 16M DMA zone that is balanced will not balance a zone on any * reasonable sized machine * o On all other machines, the top zone must be at least a reasonable |
25985edce
|
2724 |
* percentage of the middle zones. For example, on 32-bit x86, highmem |
1741c8775
|
2725 2726 2727 2728 |
* would need to be at least 256M for it to be balance a whole node. * Similarly, on x86-64 the Normal zone would need to be at least 1G * to balance a node on its own. These seemed like reasonable ratios. */ |
4ae0a48b5
|
2729 |
static bool pgdat_balanced(pg_data_t *pgdat, int order, int classzone_idx) |
1741c8775
|
2730 |
{ |
b40da0494
|
2731 |
unsigned long managed_pages = 0; |
4ae0a48b5
|
2732 |
unsigned long balanced_pages = 0; |
1741c8775
|
2733 |
int i; |
4ae0a48b5
|
2734 2735 2736 |
/* Check the watermark levels */ for (i = 0; i <= classzone_idx; i++) { struct zone *zone = pgdat->node_zones + i; |
1741c8775
|
2737 |
|
4ae0a48b5
|
2738 2739 |
if (!populated_zone(zone)) continue; |
b40da0494
|
2740 |
managed_pages += zone->managed_pages; |
4ae0a48b5
|
2741 2742 2743 2744 2745 2746 2747 2748 |
/* * A special case here: * * balance_pgdat() skips over all_unreclaimable after * DEF_PRIORITY. Effectively, it considers them balanced so * they must be considered balanced here as well! */ |
6e543d578
|
2749 |
if (!zone_reclaimable(zone)) { |
b40da0494
|
2750 |
balanced_pages += zone->managed_pages; |
4ae0a48b5
|
2751 2752 2753 2754 |
continue; } if (zone_balanced(zone, order, 0, i)) |
b40da0494
|
2755 |
balanced_pages += zone->managed_pages; |
4ae0a48b5
|
2756 2757 2758 2759 2760 |
else if (!order) return false; } if (order) |
b40da0494
|
2761 |
return balanced_pages >= (managed_pages >> 2); |
4ae0a48b5
|
2762 2763 |
else return true; |
1741c8775
|
2764 |
} |
5515061d2
|
2765 2766 2767 2768 2769 2770 2771 |
/* * Prepare kswapd for sleeping. This verifies that there are no processes * waiting in throttle_direct_reclaim() and that watermarks have been met. * * Returns true if kswapd is ready to sleep */ static bool prepare_kswapd_sleep(pg_data_t *pgdat, int order, long remaining, |
dc83edd94
|
2772 |
int classzone_idx) |
f50de2d38
|
2773 |
{ |
f50de2d38
|
2774 2775 |
/* If a direct reclaimer woke kswapd within HZ/10, it's premature */ if (remaining) |
5515061d2
|
2776 2777 2778 |
return false; /* |
53bcf5c32
|
2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 |
* The throttled processes are normally woken up in balance_pgdat() as * soon as pfmemalloc_watermark_ok() is true. But there is a potential * race between when kswapd checks the watermarks and a process gets * throttled. There is also a potential race if processes get * throttled, kswapd wakes, a large process exits thereby balancing the * zones, which causes kswapd to exit balance_pgdat() before reaching * the wake up checks. If kswapd is going to sleep, no process should * be sleeping on pfmemalloc_wait, so wake them now if necessary. If * the wake up is premature, processes will wake kswapd and get * throttled again. The difference from wake ups in balance_pgdat() is * that here we are under prepare_to_wait(). |
5515061d2
|
2790 |
*/ |
53bcf5c32
|
2791 2792 |
if (waitqueue_active(&pgdat->pfmemalloc_wait)) wake_up_all(&pgdat->pfmemalloc_wait); |
f50de2d38
|
2793 |
|
4ae0a48b5
|
2794 |
return pgdat_balanced(pgdat, order, classzone_idx); |
f50de2d38
|
2795 |
} |
1da177e4c
|
2796 |
/* |
75485363c
|
2797 2798 |
* kswapd shrinks the zone by the number of pages required to reach * the high watermark. |
b8e83b942
|
2799 2800 |
* * Returns true if kswapd scanned at least the requested number of pages to |
283aba9f9
|
2801 2802 |
* reclaim or if the lack of progress was due to pages under writeback. * This is used to determine if the scanning priority needs to be raised. |
75485363c
|
2803 |
*/ |
b8e83b942
|
2804 |
static bool kswapd_shrink_zone(struct zone *zone, |
7c954f6de
|
2805 |
int classzone_idx, |
75485363c
|
2806 |
struct scan_control *sc, |
2ab44f434
|
2807 2808 |
unsigned long lru_pages, unsigned long *nr_attempted) |
75485363c
|
2809 |
{ |
7c954f6de
|
2810 2811 |
int testorder = sc->order; unsigned long balance_gap; |
75485363c
|
2812 2813 2814 2815 |
struct reclaim_state *reclaim_state = current->reclaim_state; struct shrink_control shrink = { .gfp_mask = sc->gfp_mask, }; |
7c954f6de
|
2816 |
bool lowmem_pressure; |
75485363c
|
2817 2818 2819 |
/* Reclaim above the high watermark. */ sc->nr_to_reclaim = max(SWAP_CLUSTER_MAX, high_wmark_pages(zone)); |
7c954f6de
|
2820 2821 2822 2823 2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834 2835 2836 2837 |
/* * Kswapd reclaims only single pages with compaction enabled. Trying * too hard to reclaim until contiguous free pages have become * available can hurt performance by evicting too much useful data * from memory. Do not reclaim more than needed for compaction. */ if (IS_ENABLED(CONFIG_COMPACTION) && sc->order && compaction_suitable(zone, sc->order) != COMPACT_SKIPPED) testorder = 0; /* * We put equal pressure on every zone, unless one zone has way too * many pages free already. The "too many pages" is defined as the * high wmark plus a "gap" where the gap is either the low * watermark or 1% of the zone, whichever is smaller. */ |
4be89a346
|
2838 2839 |
balance_gap = min(low_wmark_pages(zone), DIV_ROUND_UP( zone->managed_pages, KSWAPD_ZONE_BALANCE_GAP_RATIO)); |
7c954f6de
|
2840 2841 2842 2843 2844 2845 2846 2847 2848 |
/* * If there is no low memory pressure or the zone is balanced then no * reclaim is necessary */ lowmem_pressure = (buffer_heads_over_limit && is_highmem(zone)); if (!lowmem_pressure && zone_balanced(zone, testorder, balance_gap, classzone_idx)) return true; |
75485363c
|
2849 |
shrink_zone(zone, sc); |
0ce3d7445
|
2850 2851 |
nodes_clear(shrink.nodes_to_scan); node_set(zone_to_nid(zone), shrink.nodes_to_scan); |
75485363c
|
2852 2853 |
reclaim_state->reclaimed_slab = 0; |
6e543d578
|
2854 |
shrink_slab(&shrink, sc->nr_scanned, lru_pages); |
75485363c
|
2855 |
sc->nr_reclaimed += reclaim_state->reclaimed_slab; |
2ab44f434
|
2856 2857 |
/* Account for the number of pages attempted to reclaim */ *nr_attempted += sc->nr_to_reclaim; |
570546517
|
2858 |
clear_bit(ZONE_WRITEBACK, &zone->flags); |
283aba9f9
|
2859 |
|
7c954f6de
|
2860 2861 2862 2863 2864 2865 |
/* * If a zone reaches its high watermark, consider it to be no longer * congested. It's possible there are dirty pages backed by congested * BDIs but as pressure is relieved, speculatively avoid congestion * waits. */ |
6e543d578
|
2866 |
if (zone_reclaimable(zone) && |
7c954f6de
|
2867 |
zone_balanced(zone, testorder, 0, classzone_idx)) { |
570546517
|
2868 2869 |
clear_bit(ZONE_CONGESTED, &zone->flags); clear_bit(ZONE_DIRTY, &zone->flags); |
7c954f6de
|
2870 |
} |
b8e83b942
|
2871 |
return sc->nr_scanned >= sc->nr_to_reclaim; |
75485363c
|
2872 2873 2874 |
} /* |
1da177e4c
|
2875 |
* For kswapd, balance_pgdat() will work across all this node's zones until |
418589663
|
2876 |
* they are all at high_wmark_pages(zone). |
1da177e4c
|
2877 |
* |
0abdee2bd
|
2878 |
* Returns the final order kswapd was reclaiming at |
1da177e4c
|
2879 2880 2881 2882 2883 2884 2885 2886 2887 2888 |
* * There is special handling here for zones which are full of pinned pages. * This can happen if the pages are all mlocked, or if they are all used by * device drivers (say, ZONE_DMA). Or if they are all in use by hugetlb. * What we do is to detect the case where all pages in the zone have been * scanned twice and there has been zero successful reclaim. Mark the zone as * dead and from now on, only perform a short scan. Basically we're polling * the zone for when the problem goes away. * * kswapd scans the zones in the highmem->normal->dma direction. It skips |
418589663
|
2889 2890 2891 2892 2893 |
* zones which have free_pages > high_wmark_pages(zone), but once a zone is * found to have free_pages <= high_wmark_pages(zone), we scan that zone and the * lower zones regardless of the number of free pages in the lower zones. This * interoperates with the page allocator fallback scheme to ensure that aging * of pages is balanced across the zones. |
1da177e4c
|
2894 |
*/ |
995047488
|
2895 |
static unsigned long balance_pgdat(pg_data_t *pgdat, int order, |
dc83edd94
|
2896 |
int *classzone_idx) |
1da177e4c
|
2897 |
{ |
1da177e4c
|
2898 |
int i; |
995047488
|
2899 |
int end_zone = 0; /* Inclusive. 0 = ZONE_DMA */ |
0608f43da
|
2900 2901 |
unsigned long nr_soft_reclaimed; unsigned long nr_soft_scanned; |
179e96395
|
2902 2903 |
struct scan_control sc = { .gfp_mask = GFP_KERNEL, |
ee814fe23
|
2904 |
.order = order, |
b8e83b942
|
2905 |
.priority = DEF_PRIORITY, |
ee814fe23
|
2906 |
.may_writepage = !laptop_mode, |
a6dc60f89
|
2907 |
.may_unmap = 1, |
2e2e42598
|
2908 |
.may_swap = 1, |
179e96395
|
2909 |
}; |
f8891e5e1
|
2910 |
count_vm_event(PAGEOUTRUN); |
1da177e4c
|
2911 |
|
9e3b2f8cd
|
2912 |
do { |
1da177e4c
|
2913 |
unsigned long lru_pages = 0; |
2ab44f434
|
2914 |
unsigned long nr_attempted = 0; |
b8e83b942
|
2915 |
bool raise_priority = true; |
2ab44f434
|
2916 |
bool pgdat_needs_compaction = (order > 0); |
b8e83b942
|
2917 2918 |
sc.nr_reclaimed = 0; |
1da177e4c
|
2919 |
|
d6277db4a
|
2920 2921 2922 2923 2924 2925 |
/* * Scan in the highmem->dma direction for the highest * zone which needs scanning */ for (i = pgdat->nr_zones - 1; i >= 0; i--) { struct zone *zone = pgdat->node_zones + i; |
1da177e4c
|
2926 |
|
d6277db4a
|
2927 2928 |
if (!populated_zone(zone)) continue; |
1da177e4c
|
2929 |
|
6e543d578
|
2930 2931 |
if (sc.priority != DEF_PRIORITY && !zone_reclaimable(zone)) |
d6277db4a
|
2932 |
continue; |
1da177e4c
|
2933 |
|
556adecba
|
2934 2935 2936 2937 |
/* * Do some background aging of the anon list, to give * pages a chance to be referenced before reclaiming. */ |
9e3b2f8cd
|
2938 |
age_active_anon(zone, &sc); |
556adecba
|
2939 |
|
cc715d99e
|
2940 2941 2942 2943 2944 2945 2946 2947 2948 2949 |
/* * If the number of buffer_heads in the machine * exceeds the maximum allowed level and this node * has a highmem zone, force kswapd to reclaim from * it to relieve lowmem pressure. */ if (buffer_heads_over_limit && is_highmem_idx(i)) { end_zone = i; break; } |
60cefed48
|
2950 |
if (!zone_balanced(zone, order, 0, 0)) { |
d6277db4a
|
2951 |
end_zone = i; |
e1dbeda60
|
2952 |
break; |
439423f68
|
2953 |
} else { |
d43006d50
|
2954 2955 2956 2957 |
/* * If balanced, clear the dirty and congested * flags */ |
570546517
|
2958 2959 |
clear_bit(ZONE_CONGESTED, &zone->flags); clear_bit(ZONE_DIRTY, &zone->flags); |
1da177e4c
|
2960 |
} |
1da177e4c
|
2961 |
} |
dafcb73e3
|
2962 |
|
b8e83b942
|
2963 |
if (i < 0) |
e1dbeda60
|
2964 |
goto out; |
1da177e4c
|
2965 2966 |
for (i = 0; i <= end_zone; i++) { struct zone *zone = pgdat->node_zones + i; |
2ab44f434
|
2967 2968 |
if (!populated_zone(zone)) continue; |
adea02a1b
|
2969 |
lru_pages += zone_reclaimable_pages(zone); |
2ab44f434
|
2970 2971 2972 2973 2974 2975 2976 2977 2978 2979 2980 |
/* * If any zone is currently balanced then kswapd will * not call compaction as it is expected that the * necessary pages are already available. */ if (pgdat_needs_compaction && zone_watermark_ok(zone, order, low_wmark_pages(zone), *classzone_idx, 0)) pgdat_needs_compaction = false; |
1da177e4c
|
2981 2982 2983 |
} /* |
b7ea3c417
|
2984 2985 2986 2987 2988 2989 2990 |
* If we're getting trouble reclaiming, start doing writepage * even in laptop mode. */ if (sc.priority < DEF_PRIORITY - 2) sc.may_writepage = 1; /* |
1da177e4c
|
2991 2992 2993 2994 2995 2996 2997 2998 2999 3000 |
* Now scan the zone in the dma->highmem direction, stopping * at the last zone which needs scanning. * * We do this because the page allocator works in the opposite * direction. This prevents the page allocator from allocating * pages behind kswapd's direction of progress, which would * cause too much scanning of the lower zones. */ for (i = 0; i <= end_zone; i++) { struct zone *zone = pgdat->node_zones + i; |
f3fe65122
|
3001 |
if (!populated_zone(zone)) |
1da177e4c
|
3002 |
continue; |
6e543d578
|
3003 3004 |
if (sc.priority != DEF_PRIORITY && !zone_reclaimable(zone)) |
1da177e4c
|
3005 |
continue; |
1da177e4c
|
3006 |
sc.nr_scanned = 0; |
4e4169535
|
3007 |
|
0608f43da
|
3008 3009 3010 3011 3012 3013 3014 3015 |
nr_soft_scanned = 0; /* * Call soft limit reclaim before calling shrink_zone. */ nr_soft_reclaimed = mem_cgroup_soft_limit_reclaim(zone, order, sc.gfp_mask, &nr_soft_scanned); sc.nr_reclaimed += nr_soft_reclaimed; |
32a4330d4
|
3016 |
/* |
7c954f6de
|
3017 3018 3019 3020 |
* There should be no need to raise the scanning * priority if enough pages are already being scanned * that that high watermark would be met at 100% * efficiency. |
fe2c2a106
|
3021 |
*/ |
7c954f6de
|
3022 3023 3024 |
if (kswapd_shrink_zone(zone, end_zone, &sc, lru_pages, &nr_attempted)) raise_priority = false; |
1da177e4c
|
3025 |
} |
5515061d2
|
3026 3027 3028 3029 3030 3031 3032 3033 3034 |
/* * If the low watermark is met there is no need for processes * to be throttled on pfmemalloc_wait as they should not be * able to safely make forward progress. Wake them */ if (waitqueue_active(&pgdat->pfmemalloc_wait) && pfmemalloc_watermark_ok(pgdat)) wake_up(&pgdat->pfmemalloc_wait); |
1da177e4c
|
3035 |
/* |
b8e83b942
|
3036 3037 3038 3039 3040 3041 |
* Fragmentation may mean that the system cannot be rebalanced * for high-order allocations in all zones. If twice the * allocation size has been reclaimed and the zones are still * not balanced then recheck the watermarks at order-0 to * prevent kswapd reclaiming excessively. Assume that a * process requested a high-order can direct reclaim/compact. |
1da177e4c
|
3042 |
*/ |
b8e83b942
|
3043 3044 |
if (order && sc.nr_reclaimed >= 2UL << order) order = sc.order = 0; |
8357376d3
|
3045 |
|
b8e83b942
|
3046 3047 3048 |
/* Check if kswapd should be suspending */ if (try_to_freeze() || kthread_should_stop()) break; |
8357376d3
|
3049 |
|
73ce02e96
|
3050 |
/* |
2ab44f434
|
3051 3052 3053 3054 3055 3056 3057 |
* Compact if necessary and kswapd is reclaiming at least the * high watermark number of pages as requsted */ if (pgdat_needs_compaction && sc.nr_reclaimed > nr_attempted) compact_pgdat(pgdat, order); /* |
b8e83b942
|
3058 3059 |
* Raise priority if scanning rate is too low or there was no * progress in reclaiming pages |
73ce02e96
|
3060 |
*/ |
b8e83b942
|
3061 3062 |
if (raise_priority || !sc.nr_reclaimed) sc.priority--; |
9aa41348a
|
3063 |
} while (sc.priority >= 1 && |
b8e83b942
|
3064 |
!pgdat_balanced(pgdat, order, *classzone_idx)); |
1da177e4c
|
3065 |
|
b8e83b942
|
3066 |
out: |
0abdee2bd
|
3067 |
/* |
5515061d2
|
3068 |
* Return the order we were reclaiming at so prepare_kswapd_sleep() |
0abdee2bd
|
3069 3070 3071 3072 |
* makes a decision on the order we were last reclaiming at. However, * if another caller entered the allocator slow path while kswapd * was awake, order will remain at the higher level */ |
dc83edd94
|
3073 |
*classzone_idx = end_zone; |
0abdee2bd
|
3074 |
return order; |
1da177e4c
|
3075 |
} |
dc83edd94
|
3076 |
static void kswapd_try_to_sleep(pg_data_t *pgdat, int order, int classzone_idx) |
f0bc0a60b
|
3077 3078 3079 3080 3081 3082 3083 3084 3085 3086 |
{ long remaining = 0; DEFINE_WAIT(wait); if (freezing(current) || kthread_should_stop()) return; prepare_to_wait(&pgdat->kswapd_wait, &wait, TASK_INTERRUPTIBLE); /* Try to sleep for a short interval */ |
5515061d2
|
3087 |
if (prepare_kswapd_sleep(pgdat, order, remaining, classzone_idx)) { |
f0bc0a60b
|
3088 3089 3090 3091 3092 3093 3094 3095 3096 |
remaining = schedule_timeout(HZ/10); finish_wait(&pgdat->kswapd_wait, &wait); prepare_to_wait(&pgdat->kswapd_wait, &wait, TASK_INTERRUPTIBLE); } /* * After a short sleep, check if it was a premature sleep. If not, then * go fully to sleep until explicitly woken up. */ |
5515061d2
|
3097 |
if (prepare_kswapd_sleep(pgdat, order, remaining, classzone_idx)) { |
f0bc0a60b
|
3098 3099 3100 3101 3102 3103 3104 3105 3106 3107 3108 |
trace_mm_vmscan_kswapd_sleep(pgdat->node_id); /* * vmstat counters are not perfectly accurate and the estimated * value for counters such as NR_FREE_PAGES can deviate from the * true value by nr_online_cpus * threshold. To avoid the zone * watermarks being breached while under pressure, we reduce the * per-cpu vmstat threshold while kswapd is awake and restore * them before going back to sleep. */ set_pgdat_percpu_threshold(pgdat, calculate_normal_threshold); |
1c7e7f6c0
|
3109 |
|
62997027c
|
3110 3111 3112 3113 3114 3115 3116 |
/* * Compaction records what page blocks it recently failed to * isolate pages from and skips them in the future scanning. * When kswapd is going to sleep, it is reasonable to assume * that pages and compaction may succeed so reset the cache. */ reset_isolation_suitable(pgdat); |
1c7e7f6c0
|
3117 3118 |
if (!kthread_should_stop()) schedule(); |
f0bc0a60b
|
3119 3120 3121 3122 3123 3124 3125 3126 3127 |
set_pgdat_percpu_threshold(pgdat, calculate_pressure_threshold); } else { if (remaining) count_vm_event(KSWAPD_LOW_WMARK_HIT_QUICKLY); else count_vm_event(KSWAPD_HIGH_WMARK_HIT_QUICKLY); } finish_wait(&pgdat->kswapd_wait, &wait); } |
1da177e4c
|
3128 3129 |
/* * The background pageout daemon, started as a kernel thread |
4f98a2fee
|
3130 |
* from the init process. |
1da177e4c
|
3131 3132 3133 3134 3135 3136 3137 3138 3139 3140 3141 3142 |
* * This basically trickles out pages so that we have _some_ * free memory available even if there is no other activity * that frees anything up. This is needed for things like routing * etc, where we otherwise might have all activity going on in * asynchronous contexts that cannot page things out. * * If there are applications that are active memory-allocators * (most normal use), this basically shouldn't matter. */ static int kswapd(void *p) { |
215ddd666
|
3143 |
unsigned long order, new_order; |
d2ebd0f6b
|
3144 |
unsigned balanced_order; |
215ddd666
|
3145 |
int classzone_idx, new_classzone_idx; |
d2ebd0f6b
|
3146 |
int balanced_classzone_idx; |
1da177e4c
|
3147 3148 |
pg_data_t *pgdat = (pg_data_t*)p; struct task_struct *tsk = current; |
f0bc0a60b
|
3149 |
|
1da177e4c
|
3150 3151 3152 |
struct reclaim_state reclaim_state = { .reclaimed_slab = 0, }; |
a70f73028
|
3153 |
const struct cpumask *cpumask = cpumask_of_node(pgdat->node_id); |
1da177e4c
|
3154 |
|
cf40bd16f
|
3155 |
lockdep_set_current_reclaim_state(GFP_KERNEL); |
174596a0b
|
3156 |
if (!cpumask_empty(cpumask)) |
c5f59f083
|
3157 |
set_cpus_allowed_ptr(tsk, cpumask); |
1da177e4c
|
3158 3159 3160 3161 3162 3163 3164 3165 3166 3167 3168 3169 3170 3171 |
current->reclaim_state = &reclaim_state; /* * Tell the memory management that we're a "memory allocator", * and that if we need more memory we should get access to it * regardless (see "__alloc_pages()"). "kswapd" should * never get caught in the normal page freeing logic. * * (Kswapd normally doesn't need memory anyway, but sometimes * you need a small amount of memory in order to be able to * page out something else, and this flag essentially protects * us from recursively trying to free more memory as we're * trying to free the first piece of memory in the first place). */ |
930d91525
|
3172 |
tsk->flags |= PF_MEMALLOC | PF_SWAPWRITE | PF_KSWAPD; |
831441862
|
3173 |
set_freezable(); |
1da177e4c
|
3174 |
|
215ddd666
|
3175 |
order = new_order = 0; |
d2ebd0f6b
|
3176 |
balanced_order = 0; |
215ddd666
|
3177 |
classzone_idx = new_classzone_idx = pgdat->nr_zones - 1; |
d2ebd0f6b
|
3178 |
balanced_classzone_idx = classzone_idx; |
1da177e4c
|
3179 |
for ( ; ; ) { |
6f6313d48
|
3180 |
bool ret; |
3e1d1d28d
|
3181 |
|
215ddd666
|
3182 3183 3184 3185 3186 |
/* * If the last balance_pgdat was unsuccessful it's unlikely a * new request of a similar or harder type will succeed soon * so consider going to sleep on the basis we reclaimed at */ |
d2ebd0f6b
|
3187 3188 |
if (balanced_classzone_idx >= new_classzone_idx && balanced_order == new_order) { |
215ddd666
|
3189 3190 3191 3192 3193 |
new_order = pgdat->kswapd_max_order; new_classzone_idx = pgdat->classzone_idx; pgdat->kswapd_max_order = 0; pgdat->classzone_idx = pgdat->nr_zones - 1; } |
995047488
|
3194 |
if (order < new_order || classzone_idx > new_classzone_idx) { |
1da177e4c
|
3195 3196 |
/* * Don't sleep if someone wants a larger 'order' |
995047488
|
3197 |
* allocation or has tigher zone constraints |
1da177e4c
|
3198 3199 |
*/ order = new_order; |
995047488
|
3200 |
classzone_idx = new_classzone_idx; |
1da177e4c
|
3201 |
} else { |
d2ebd0f6b
|
3202 3203 |
kswapd_try_to_sleep(pgdat, balanced_order, balanced_classzone_idx); |
1da177e4c
|
3204 |
order = pgdat->kswapd_max_order; |
995047488
|
3205 |
classzone_idx = pgdat->classzone_idx; |
f0dfcde09
|
3206 3207 |
new_order = order; new_classzone_idx = classzone_idx; |
4d40502ea
|
3208 |
pgdat->kswapd_max_order = 0; |
215ddd666
|
3209 |
pgdat->classzone_idx = pgdat->nr_zones - 1; |
1da177e4c
|
3210 |
} |
1da177e4c
|
3211 |
|
8fe23e057
|
3212 3213 3214 3215 3216 3217 3218 3219 |
ret = try_to_freeze(); if (kthread_should_stop()) break; /* * We can speed up thawing tasks if we don't call balance_pgdat * after returning from the refrigerator */ |
33906bc5c
|
3220 3221 |
if (!ret) { trace_mm_vmscan_kswapd_wake(pgdat->node_id, order); |
d2ebd0f6b
|
3222 3223 3224 |
balanced_classzone_idx = classzone_idx; balanced_order = balance_pgdat(pgdat, order, &balanced_classzone_idx); |
33906bc5c
|
3225 |
} |
1da177e4c
|
3226 |
} |
b0a8cc58e
|
3227 |
|
71abdc15a
|
3228 |
tsk->flags &= ~(PF_MEMALLOC | PF_SWAPWRITE | PF_KSWAPD); |
b0a8cc58e
|
3229 |
current->reclaim_state = NULL; |
71abdc15a
|
3230 |
lockdep_clear_current_reclaim_state(); |
1da177e4c
|
3231 3232 3233 3234 3235 3236 |
return 0; } /* * A zone is low on free memory, so wake its kswapd task to service it. */ |
995047488
|
3237 |
void wakeup_kswapd(struct zone *zone, int order, enum zone_type classzone_idx) |
1da177e4c
|
3238 3239 |
{ pg_data_t *pgdat; |
f3fe65122
|
3240 |
if (!populated_zone(zone)) |
1da177e4c
|
3241 |
return; |
88f5acf88
|
3242 |
if (!cpuset_zone_allowed_hardwall(zone, GFP_KERNEL)) |
1da177e4c
|
3243 |
return; |
88f5acf88
|
3244 |
pgdat = zone->zone_pgdat; |
995047488
|
3245 |
if (pgdat->kswapd_max_order < order) { |
1da177e4c
|
3246 |
pgdat->kswapd_max_order = order; |
995047488
|
3247 3248 |
pgdat->classzone_idx = min(pgdat->classzone_idx, classzone_idx); } |
8d0986e28
|
3249 |
if (!waitqueue_active(&pgdat->kswapd_wait)) |
1da177e4c
|
3250 |
return; |
892f795df
|
3251 |
if (zone_balanced(zone, order, 0, 0)) |
88f5acf88
|
3252 3253 3254 |
return; trace_mm_vmscan_wakeup_kswapd(pgdat->node_id, zone_idx(zone), order); |
8d0986e28
|
3255 |
wake_up_interruptible(&pgdat->kswapd_wait); |
1da177e4c
|
3256 |
} |
c6f37f121
|
3257 |
#ifdef CONFIG_HIBERNATION |
1da177e4c
|
3258 |
/* |
7b51755c3
|
3259 |
* Try to free `nr_to_reclaim' of memory, system-wide, and return the number of |
d6277db4a
|
3260 3261 3262 3263 3264 |
* freed pages. * * Rather than trying to age LRUs the aim is to preserve the overall * LRU order by reclaiming preferentially * inactive > active > active referenced > active mapped |
1da177e4c
|
3265 |
*/ |
7b51755c3
|
3266 |
unsigned long shrink_all_memory(unsigned long nr_to_reclaim) |
1da177e4c
|
3267 |
{ |
d6277db4a
|
3268 |
struct reclaim_state reclaim_state; |
d6277db4a
|
3269 |
struct scan_control sc = { |
ee814fe23
|
3270 |
.nr_to_reclaim = nr_to_reclaim, |
7b51755c3
|
3271 |
.gfp_mask = GFP_HIGHUSER_MOVABLE, |
ee814fe23
|
3272 |
.priority = DEF_PRIORITY, |
d6277db4a
|
3273 |
.may_writepage = 1, |
ee814fe23
|
3274 3275 |
.may_unmap = 1, .may_swap = 1, |
7b51755c3
|
3276 |
.hibernation_mode = 1, |
1da177e4c
|
3277 |
}; |
a09ed5e00
|
3278 |
struct zonelist *zonelist = node_zonelist(numa_node_id(), sc.gfp_mask); |
7b51755c3
|
3279 3280 |
struct task_struct *p = current; unsigned long nr_reclaimed; |
1da177e4c
|
3281 |
|
7b51755c3
|
3282 3283 3284 3285 |
p->flags |= PF_MEMALLOC; lockdep_set_current_reclaim_state(sc.gfp_mask); reclaim_state.reclaimed_slab = 0; p->reclaim_state = &reclaim_state; |
d6277db4a
|
3286 |
|
3115cd914
|
3287 |
nr_reclaimed = do_try_to_free_pages(zonelist, &sc); |
d979677c4
|
3288 |
|
7b51755c3
|
3289 3290 3291 |
p->reclaim_state = NULL; lockdep_clear_current_reclaim_state(); p->flags &= ~PF_MEMALLOC; |
d6277db4a
|
3292 |
|
7b51755c3
|
3293 |
return nr_reclaimed; |
1da177e4c
|
3294 |
} |
c6f37f121
|
3295 |
#endif /* CONFIG_HIBERNATION */ |
1da177e4c
|
3296 |
|
1da177e4c
|
3297 3298 3299 3300 |
/* It's optimal to keep kswapds on the same CPUs as their memory, but not required for correctness. So if the last cpu in a node goes away, we get changed to run anywhere: as the first one comes back, restore their cpu bindings. */ |
fcb35a9ba
|
3301 3302 |
static int cpu_callback(struct notifier_block *nfb, unsigned long action, void *hcpu) |
1da177e4c
|
3303 |
{ |
58c0a4a78
|
3304 |
int nid; |
1da177e4c
|
3305 |
|
8bb784428
|
3306 |
if (action == CPU_ONLINE || action == CPU_ONLINE_FROZEN) { |
48fb2e240
|
3307 |
for_each_node_state(nid, N_MEMORY) { |
c5f59f083
|
3308 |
pg_data_t *pgdat = NODE_DATA(nid); |
a70f73028
|
3309 3310 3311 |
const struct cpumask *mask; mask = cpumask_of_node(pgdat->node_id); |
c5f59f083
|
3312 |
|
3e5979453
|
3313 |
if (cpumask_any_and(cpu_online_mask, mask) < nr_cpu_ids) |
1da177e4c
|
3314 |
/* One of our CPUs online: restore mask */ |
c5f59f083
|
3315 |
set_cpus_allowed_ptr(pgdat->kswapd, mask); |
1da177e4c
|
3316 3317 3318 3319 |
} } return NOTIFY_OK; } |
1da177e4c
|
3320 |
|
3218ae14b
|
3321 3322 3323 3324 3325 3326 3327 3328 3329 3330 3331 3332 3333 3334 3335 3336 |
/* * This kswapd start function will be called by init and node-hot-add. * On node-hot-add, kswapd will moved to proper cpus if cpus are hot-added. */ int kswapd_run(int nid) { pg_data_t *pgdat = NODE_DATA(nid); int ret = 0; if (pgdat->kswapd) return 0; pgdat->kswapd = kthread_run(kswapd, pgdat, "kswapd%d", nid); if (IS_ERR(pgdat->kswapd)) { /* failure at boot is fatal */ BUG_ON(system_state == SYSTEM_BOOTING); |
d5dc0ad92
|
3337 3338 3339 |
pr_err("Failed to start kswapd on node %d ", nid); ret = PTR_ERR(pgdat->kswapd); |
d72515b85
|
3340 |
pgdat->kswapd = NULL; |
3218ae14b
|
3341 3342 3343 |
} return ret; } |
8fe23e057
|
3344 |
/* |
d8adde17e
|
3345 |
* Called by memory hotplug when all memory in a node is offlined. Caller must |
bfc8c9013
|
3346 |
* hold mem_hotplug_begin/end(). |
8fe23e057
|
3347 3348 3349 3350 |
*/ void kswapd_stop(int nid) { struct task_struct *kswapd = NODE_DATA(nid)->kswapd; |
d8adde17e
|
3351 |
if (kswapd) { |
8fe23e057
|
3352 |
kthread_stop(kswapd); |
d8adde17e
|
3353 3354 |
NODE_DATA(nid)->kswapd = NULL; } |
8fe23e057
|
3355 |
} |
1da177e4c
|
3356 3357 |
static int __init kswapd_init(void) { |
3218ae14b
|
3358 |
int nid; |
69e05944a
|
3359 |
|
1da177e4c
|
3360 |
swap_setup(); |
48fb2e240
|
3361 |
for_each_node_state(nid, N_MEMORY) |
3218ae14b
|
3362 |
kswapd_run(nid); |
1da177e4c
|
3363 3364 3365 3366 3367 |
hotcpu_notifier(cpu_callback, 0); return 0; } module_init(kswapd_init) |
9eeff2395
|
3368 3369 3370 3371 3372 3373 3374 |
#ifdef CONFIG_NUMA /* * Zone reclaim mode * * If non-zero call zone_reclaim when the number of free pages falls below * the watermarks. |
9eeff2395
|
3375 3376 |
*/ int zone_reclaim_mode __read_mostly; |
1b2ffb789
|
3377 |
#define RECLAIM_OFF 0 |
7d03431cf
|
3378 |
#define RECLAIM_ZONE (1<<0) /* Run shrink_inactive_list on the zone */ |
1b2ffb789
|
3379 3380 |
#define RECLAIM_WRITE (1<<1) /* Writeout pages during reclaim */ #define RECLAIM_SWAP (1<<2) /* Swap pages out during reclaim */ |
9eeff2395
|
3381 |
/* |
a92f71263
|
3382 3383 3384 3385 3386 |
* Priority for ZONE_RECLAIM. This determines the fraction of pages * of a node considered for each zone_reclaim. 4 scans 1/16th of * a zone. */ #define ZONE_RECLAIM_PRIORITY 4 |
9eeff2395
|
3387 |
/* |
9614634fe
|
3388 3389 3390 3391 3392 3393 |
* Percentage of pages in a zone that must be unmapped for zone_reclaim to * occur. */ int sysctl_min_unmapped_ratio = 1; /* |
0ff38490c
|
3394 3395 3396 3397 |
* If the number of slab pages in a zone grows beyond this percentage then * slab reclaim needs to occur. */ int sysctl_min_slab_ratio = 5; |
90afa5de6
|
3398 3399 3400 3401 3402 3403 3404 3405 3406 3407 3408 3409 3410 3411 3412 3413 3414 3415 3416 3417 3418 3419 3420 3421 3422 3423 3424 3425 3426 3427 3428 3429 3430 3431 3432 3433 3434 3435 3436 3437 3438 |
static inline unsigned long zone_unmapped_file_pages(struct zone *zone) { unsigned long file_mapped = zone_page_state(zone, NR_FILE_MAPPED); unsigned long file_lru = zone_page_state(zone, NR_INACTIVE_FILE) + zone_page_state(zone, NR_ACTIVE_FILE); /* * It's possible for there to be more file mapped pages than * accounted for by the pages on the file LRU lists because * tmpfs pages accounted for as ANON can also be FILE_MAPPED */ return (file_lru > file_mapped) ? (file_lru - file_mapped) : 0; } /* Work out how many page cache pages we can reclaim in this reclaim_mode */ static long zone_pagecache_reclaimable(struct zone *zone) { long nr_pagecache_reclaimable; long delta = 0; /* * If RECLAIM_SWAP is set, then all file pages are considered * potentially reclaimable. Otherwise, we have to worry about * pages like swapcache and zone_unmapped_file_pages() provides * a better estimate */ if (zone_reclaim_mode & RECLAIM_SWAP) nr_pagecache_reclaimable = zone_page_state(zone, NR_FILE_PAGES); else nr_pagecache_reclaimable = zone_unmapped_file_pages(zone); /* If we can't clean pages, remove dirty pages from consideration */ if (!(zone_reclaim_mode & RECLAIM_WRITE)) delta += zone_page_state(zone, NR_FILE_DIRTY); /* Watch for any possible underflows due to delta */ if (unlikely(delta > nr_pagecache_reclaimable)) delta = nr_pagecache_reclaimable; return nr_pagecache_reclaimable - delta; } |
0ff38490c
|
3439 |
/* |
9eeff2395
|
3440 3441 |
* Try to free up some pages from this zone through reclaim. */ |
179e96395
|
3442 |
static int __zone_reclaim(struct zone *zone, gfp_t gfp_mask, unsigned int order) |
9eeff2395
|
3443 |
{ |
7fb2d46d3
|
3444 |
/* Minimum pages needed in order to stay on node */ |
69e05944a
|
3445 |
const unsigned long nr_pages = 1 << order; |
9eeff2395
|
3446 3447 |
struct task_struct *p = current; struct reclaim_state reclaim_state; |
179e96395
|
3448 |
struct scan_control sc = { |
62b726c1b
|
3449 |
.nr_to_reclaim = max(nr_pages, SWAP_CLUSTER_MAX), |
21caf2fc1
|
3450 |
.gfp_mask = (gfp_mask = memalloc_noio_flags(gfp_mask)), |
bd2f6199c
|
3451 |
.order = order, |
9e3b2f8cd
|
3452 |
.priority = ZONE_RECLAIM_PRIORITY, |
ee814fe23
|
3453 3454 3455 |
.may_writepage = !!(zone_reclaim_mode & RECLAIM_WRITE), .may_unmap = !!(zone_reclaim_mode & RECLAIM_SWAP), .may_swap = 1, |
179e96395
|
3456 |
}; |
a09ed5e00
|
3457 3458 3459 |
struct shrink_control shrink = { .gfp_mask = sc.gfp_mask, }; |
157480489
|
3460 |
unsigned long nr_slab_pages0, nr_slab_pages1; |
9eeff2395
|
3461 |
|
9eeff2395
|
3462 |
cond_resched(); |
d4f7796e9
|
3463 3464 3465 3466 3467 3468 |
/* * We need to be able to allocate from the reserves for RECLAIM_SWAP * and we also need to be able to write out pages for RECLAIM_WRITE * and RECLAIM_SWAP. */ p->flags |= PF_MEMALLOC | PF_SWAPWRITE; |
76ca542d8
|
3469 |
lockdep_set_current_reclaim_state(gfp_mask); |
9eeff2395
|
3470 3471 |
reclaim_state.reclaimed_slab = 0; p->reclaim_state = &reclaim_state; |
c84db23c6
|
3472 |
|
90afa5de6
|
3473 |
if (zone_pagecache_reclaimable(zone) > zone->min_unmapped_pages) { |
0ff38490c
|
3474 3475 3476 3477 |
/* * Free memory by calling shrink zone with increasing * priorities until we have enough memory freed. */ |
0ff38490c
|
3478 |
do { |
9e3b2f8cd
|
3479 3480 |
shrink_zone(zone, &sc); } while (sc.nr_reclaimed < nr_pages && --sc.priority >= 0); |
0ff38490c
|
3481 |
} |
c84db23c6
|
3482 |
|
157480489
|
3483 3484 |
nr_slab_pages0 = zone_page_state(zone, NR_SLAB_RECLAIMABLE); if (nr_slab_pages0 > zone->min_slab_pages) { |
2a16e3f4b
|
3485 |
/* |
7fb2d46d3
|
3486 |
* shrink_slab() does not currently allow us to determine how |
0ff38490c
|
3487 3488 3489 3490 |
* many pages were freed in this zone. So we take the current * number of slab pages and shake the slab until it is reduced * by the same nr_pages that we used for reclaiming unmapped * pages. |
2a16e3f4b
|
3491 |
*/ |
0ce3d7445
|
3492 3493 |
nodes_clear(shrink.nodes_to_scan); node_set(zone_to_nid(zone), shrink.nodes_to_scan); |
4dc4b3d97
|
3494 3495 3496 3497 |
for (;;) { unsigned long lru_pages = zone_reclaimable_pages(zone); /* No reclaimable slab or very low memory pressure */ |
1495f230f
|
3498 |
if (!shrink_slab(&shrink, sc.nr_scanned, lru_pages)) |
4dc4b3d97
|
3499 3500 3501 3502 3503 3504 3505 3506 |
break; /* Freed enough memory */ nr_slab_pages1 = zone_page_state(zone, NR_SLAB_RECLAIMABLE); if (nr_slab_pages1 + nr_pages <= nr_slab_pages0) break; } |
83e33a471
|
3507 3508 3509 3510 3511 |
/* * Update nr_reclaimed by the number of slab pages we * reclaimed from this zone. */ |
157480489
|
3512 3513 3514 |
nr_slab_pages1 = zone_page_state(zone, NR_SLAB_RECLAIMABLE); if (nr_slab_pages1 < nr_slab_pages0) sc.nr_reclaimed += nr_slab_pages0 - nr_slab_pages1; |
2a16e3f4b
|
3515 |
} |
9eeff2395
|
3516 |
p->reclaim_state = NULL; |
d4f7796e9
|
3517 |
current->flags &= ~(PF_MEMALLOC | PF_SWAPWRITE); |
76ca542d8
|
3518 |
lockdep_clear_current_reclaim_state(); |
a79311c14
|
3519 |
return sc.nr_reclaimed >= nr_pages; |
9eeff2395
|
3520 |
} |
179e96395
|
3521 3522 3523 |
int zone_reclaim(struct zone *zone, gfp_t gfp_mask, unsigned int order) { |
179e96395
|
3524 |
int node_id; |
d773ed6b8
|
3525 |
int ret; |
179e96395
|
3526 3527 |
/* |
0ff38490c
|
3528 3529 |
* Zone reclaim reclaims unmapped file backed pages and * slab pages if we are over the defined limits. |
34aa1330f
|
3530 |
* |
9614634fe
|
3531 3532 3533 3534 3535 |
* A small portion of unmapped file backed pages is needed for * file I/O otherwise pages read by file I/O will be immediately * thrown out if the zone is overallocated. So we do not reclaim * if less than a specified percentage of the zone is used by * unmapped file backed pages. |
179e96395
|
3536 |
*/ |
90afa5de6
|
3537 3538 |
if (zone_pagecache_reclaimable(zone) <= zone->min_unmapped_pages && zone_page_state(zone, NR_SLAB_RECLAIMABLE) <= zone->min_slab_pages) |
fa5e084e4
|
3539 |
return ZONE_RECLAIM_FULL; |
179e96395
|
3540 |
|
6e543d578
|
3541 |
if (!zone_reclaimable(zone)) |
fa5e084e4
|
3542 |
return ZONE_RECLAIM_FULL; |
d773ed6b8
|
3543 |
|
179e96395
|
3544 |
/* |
d773ed6b8
|
3545 |
* Do not scan if the allocation should not be delayed. |
179e96395
|
3546 |
*/ |
d773ed6b8
|
3547 |
if (!(gfp_mask & __GFP_WAIT) || (current->flags & PF_MEMALLOC)) |
fa5e084e4
|
3548 |
return ZONE_RECLAIM_NOSCAN; |
179e96395
|
3549 3550 3551 3552 3553 3554 3555 |
/* * Only run zone reclaim on the local zone or on zones that do not * have associated processors. This will favor the local processor * over remote processors and spread off node memory allocations * as wide as possible. */ |
89fa30242
|
3556 |
node_id = zone_to_nid(zone); |
37c0708db
|
3557 |
if (node_state(node_id, N_CPU) && node_id != numa_node_id()) |
fa5e084e4
|
3558 |
return ZONE_RECLAIM_NOSCAN; |
d773ed6b8
|
3559 |
|
570546517
|
3560 |
if (test_and_set_bit(ZONE_RECLAIM_LOCKED, &zone->flags)) |
fa5e084e4
|
3561 |
return ZONE_RECLAIM_NOSCAN; |
d773ed6b8
|
3562 |
ret = __zone_reclaim(zone, gfp_mask, order); |
570546517
|
3563 |
clear_bit(ZONE_RECLAIM_LOCKED, &zone->flags); |
d773ed6b8
|
3564 |
|
24cf72518
|
3565 3566 |
if (!ret) count_vm_event(PGSCAN_ZONE_RECLAIM_FAILED); |
d773ed6b8
|
3567 |
return ret; |
179e96395
|
3568 |
} |
9eeff2395
|
3569 |
#endif |
894bc3104
|
3570 |
|
894bc3104
|
3571 3572 3573 |
/* * page_evictable - test whether a page is evictable * @page: the page to test |
894bc3104
|
3574 3575 |
* * Test whether page is evictable--i.e., should be placed on active/inactive |
39b5f29ac
|
3576 |
* lists vs unevictable list. |
894bc3104
|
3577 3578 |
* * Reasons page might not be evictable: |
ba9ddf493
|
3579 |
* (1) page's mapping marked unevictable |
b291f0003
|
3580 |
* (2) page is part of an mlocked VMA |
ba9ddf493
|
3581 |
* |
894bc3104
|
3582 |
*/ |
39b5f29ac
|
3583 |
int page_evictable(struct page *page) |
894bc3104
|
3584 |
{ |
39b5f29ac
|
3585 |
return !mapping_unevictable(page_mapping(page)) && !PageMlocked(page); |
894bc3104
|
3586 |
} |
89e004ea5
|
3587 |
|
85046579b
|
3588 |
#ifdef CONFIG_SHMEM |
89e004ea5
|
3589 |
/** |
245132643
|
3590 3591 3592 |
* check_move_unevictable_pages - check pages for evictability and move to appropriate zone lru list * @pages: array of pages to check * @nr_pages: number of pages to check |
89e004ea5
|
3593 |
* |
245132643
|
3594 |
* Checks pages for evictability and moves them to the appropriate lru list. |
85046579b
|
3595 3596 |
* * This function is only used for SysV IPC SHM_UNLOCK. |
89e004ea5
|
3597 |
*/ |
245132643
|
3598 |
void check_move_unevictable_pages(struct page **pages, int nr_pages) |
89e004ea5
|
3599 |
{ |
925b7673c
|
3600 |
struct lruvec *lruvec; |
245132643
|
3601 3602 3603 3604 |
struct zone *zone = NULL; int pgscanned = 0; int pgrescued = 0; int i; |
89e004ea5
|
3605 |
|
245132643
|
3606 3607 3608 |
for (i = 0; i < nr_pages; i++) { struct page *page = pages[i]; struct zone *pagezone; |
89e004ea5
|
3609 |
|
245132643
|
3610 3611 3612 3613 3614 3615 3616 3617 |
pgscanned++; pagezone = page_zone(page); if (pagezone != zone) { if (zone) spin_unlock_irq(&zone->lru_lock); zone = pagezone; spin_lock_irq(&zone->lru_lock); } |
fa9add641
|
3618 |
lruvec = mem_cgroup_page_lruvec(page, zone); |
89e004ea5
|
3619 |
|
245132643
|
3620 3621 |
if (!PageLRU(page) || !PageUnevictable(page)) continue; |
89e004ea5
|
3622 |
|
39b5f29ac
|
3623 |
if (page_evictable(page)) { |
245132643
|
3624 |
enum lru_list lru = page_lru_base_type(page); |
309381fea
|
3625 |
VM_BUG_ON_PAGE(PageActive(page), page); |
245132643
|
3626 |
ClearPageUnevictable(page); |
fa9add641
|
3627 3628 |
del_page_from_lru_list(page, lruvec, LRU_UNEVICTABLE); add_page_to_lru_list(page, lruvec, lru); |
245132643
|
3629 |
pgrescued++; |
89e004ea5
|
3630 |
} |
245132643
|
3631 |
} |
89e004ea5
|
3632 |
|
245132643
|
3633 3634 3635 3636 |
if (zone) { __count_vm_events(UNEVICTABLE_PGRESCUED, pgrescued); __count_vm_events(UNEVICTABLE_PGSCANNED, pgscanned); spin_unlock_irq(&zone->lru_lock); |
89e004ea5
|
3637 |
} |
89e004ea5
|
3638 |
} |
85046579b
|
3639 |
#endif /* CONFIG_SHMEM */ |