Blame view
mm/vmscan.c
107 KB
1da177e4c
|
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 |
/* * 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. */ #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> |
1da177e4c
|
46 47 48 49 50 |
#include <asm/tlbflush.h> #include <asm/div64.h> #include <linux/swapops.h> |
117aad1e9
|
51 |
#include <linux/balloon_compaction.h> |
1da177e4c
|
52 |
|
0f8053a50
|
53 |
#include "internal.h" |
33906bc5c
|
54 55 |
#define CREATE_TRACE_POINTS #include <trace/events/vmscan.h> |
1da177e4c
|
56 |
struct scan_control { |
1da177e4c
|
57 58 |
/* Incremented by the number of inactive pages that were scanned */ unsigned long nr_scanned; |
a79311c14
|
59 60 |
/* Number of pages freed so far during a call to shrink_zones() */ unsigned long nr_reclaimed; |
22fba3354
|
61 62 |
/* How many pages shrink_list() should reclaim */ unsigned long nr_to_reclaim; |
7b51755c3
|
63 |
unsigned long hibernation_mode; |
1da177e4c
|
64 |
/* This context's GFP mask */ |
6daa0e286
|
65 |
gfp_t gfp_mask; |
1da177e4c
|
66 67 |
int may_writepage; |
a6dc60f89
|
68 69 |
/* Can mapped pages be reclaimed? */ int may_unmap; |
f1fd1067e
|
70 |
|
2e2e42598
|
71 72 |
/* Can pages be swapped as part of reclaim? */ int may_swap; |
5ad333eb6
|
73 |
int order; |
66e1707bc
|
74 |
|
9e3b2f8cd
|
75 76 |
/* Scan (total_size >> priority) pages at once */ int priority; |
5f53e7629
|
77 |
/* |
f16015fbf
|
78 79 80 81 |
* 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
|
82 |
|
327c0e968
|
83 84 85 86 87 |
/* * Nodemask of nodes allowed by the caller. If NULL, all nodes * are scanned. */ nodemask_t *nodemask; |
1da177e4c
|
88 |
}; |
1da177e4c
|
89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 |
#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; |
b21e0b90c
|
123 |
unsigned long vm_total_pages; /* The total number of pages which the VM controls */ |
1da177e4c
|
124 125 126 |
static LIST_HEAD(shrinker_list); static DECLARE_RWSEM(shrinker_rwsem); |
c255a4580
|
127 |
#ifdef CONFIG_MEMCG |
89b5fae53
|
128 129 |
static bool global_reclaim(struct scan_control *sc) { |
f16015fbf
|
130 |
return !sc->target_mem_cgroup; |
89b5fae53
|
131 |
} |
91a45470f
|
132 |
#else |
89b5fae53
|
133 134 135 136 |
static bool global_reclaim(struct scan_control *sc) { return true; } |
91a45470f
|
137 |
#endif |
a1c3bfb2f
|
138 |
static unsigned long zone_reclaimable_pages(struct zone *zone) |
6e543d578
|
139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 |
{ 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) { return zone->pages_scanned < zone_reclaimable_pages(zone) * 6; } |
4d7dcca21
|
156 |
static unsigned long get_lru_size(struct lruvec *lruvec, enum lru_list lru) |
c9f299d98
|
157 |
{ |
c3c787e8c
|
158 |
if (!mem_cgroup_disabled()) |
4d7dcca21
|
159 |
return mem_cgroup_get_lru_size(lruvec, lru); |
a3d8e0549
|
160 |
|
074291fea
|
161 |
return zone_page_state(lruvec_zone(lruvec), NR_LRU_BASE + lru); |
c9f299d98
|
162 |
} |
1da177e4c
|
163 |
/* |
1d3d4437e
|
164 |
* Add a shrinker callback to be called from the vm. |
1da177e4c
|
165 |
*/ |
1d3d4437e
|
166 |
int register_shrinker(struct shrinker *shrinker) |
1da177e4c
|
167 |
{ |
1d3d4437e
|
168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 |
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
|
184 185 186 |
down_write(&shrinker_rwsem); list_add_tail(&shrinker->list, &shrinker_list); up_write(&shrinker_rwsem); |
1d3d4437e
|
187 |
return 0; |
1da177e4c
|
188 |
} |
8e1f936b7
|
189 |
EXPORT_SYMBOL(register_shrinker); |
1da177e4c
|
190 191 192 193 |
/* * Remove one */ |
8e1f936b7
|
194 |
void unregister_shrinker(struct shrinker *shrinker) |
1da177e4c
|
195 196 197 198 |
{ down_write(&shrinker_rwsem); list_del(&shrinker->list); up_write(&shrinker_rwsem); |
ae3933216
|
199 |
kfree(shrinker->nr_deferred); |
1da177e4c
|
200 |
} |
8e1f936b7
|
201 |
EXPORT_SYMBOL(unregister_shrinker); |
1da177e4c
|
202 203 |
#define SHRINK_BATCH 128 |
1d3d4437e
|
204 205 206 207 208 209 210 211 212 213 214 215 216 217 |
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; long max_pass; long nr; long new_nr; int nid = shrinkctl->nid; long batch_size = shrinker->batch ? shrinker->batch : SHRINK_BATCH; |
a0b02131c
|
218 |
max_pass = shrinker->count_objects(shrinker, shrinkctl); |
1d3d4437e
|
219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 |
if (max_pass == 0) 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; delta *= max_pass; 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
|
238 |
shrinker->scan_objects, total_scan); |
1d3d4437e
|
239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 |
total_scan = max_pass; } /* * 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 >>> * max_pass. This is bad for sustaining a working set in * memory. * * Hence only allow the shrinker to scan the entire cache when * a large delta change is calculated directly. */ if (delta < max_pass / 4) total_scan = min(total_scan, max_pass / 2); /* * Avoid risking looping forever due to too large nr value: * never try to free more than twice the estimate number of * freeable entries. */ if (total_scan > max_pass * 2) total_scan = max_pass * 2; trace_mm_shrink_slab_start(shrinker, shrinkctl, nr, nr_pages_scanned, lru_pages, max_pass, delta, total_scan); |
0b1fb40a3
|
268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 |
/* * 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 * than the total number of objects on slab (max_pass), we must be * scanning at high prio and therefore should try to reclaim as much as * possible. */ while (total_scan >= batch_size || total_scan >= max_pass) { |
a0b02131c
|
285 |
unsigned long ret; |
0b1fb40a3
|
286 |
unsigned long nr_to_scan = min(batch_size, total_scan); |
1d3d4437e
|
287 |
|
0b1fb40a3
|
288 |
shrinkctl->nr_to_scan = nr_to_scan; |
a0b02131c
|
289 290 291 292 |
ret = shrinker->scan_objects(shrinker, shrinkctl); if (ret == SHRINK_STOP) break; freed += ret; |
1d3d4437e
|
293 |
|
0b1fb40a3
|
294 295 |
count_vm_events(SLABS_SCANNED, nr_to_scan); total_scan -= nr_to_scan; |
1d3d4437e
|
296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 |
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]); trace_mm_shrink_slab_end(shrinker, freed, nr, new_nr); return freed; |
1495f230f
|
313 |
} |
1da177e4c
|
314 315 316 317 318 319 320 321 |
/* * 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
|
322 |
* If the vm encountered mapped pages on the LRU it increase the pressure on |
1da177e4c
|
323 324 325 326 327 328 329 |
* 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
|
330 331 |
* * Returns the number of slab objects which we shrunk. |
1da177e4c
|
332 |
*/ |
24f7c6b98
|
333 |
unsigned long shrink_slab(struct shrink_control *shrinkctl, |
1495f230f
|
334 |
unsigned long nr_pages_scanned, |
a09ed5e00
|
335 |
unsigned long lru_pages) |
1da177e4c
|
336 337 |
{ struct shrinker *shrinker; |
24f7c6b98
|
338 |
unsigned long freed = 0; |
1da177e4c
|
339 |
|
1495f230f
|
340 341 |
if (nr_pages_scanned == 0) nr_pages_scanned = SWAP_CLUSTER_MAX; |
1da177e4c
|
342 |
|
f06590bd7
|
343 |
if (!down_read_trylock(&shrinker_rwsem)) { |
24f7c6b98
|
344 345 346 347 348 349 350 |
/* * 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
|
351 352 |
goto out; } |
1da177e4c
|
353 354 |
list_for_each_entry(shrinker, &shrinker_list, list) { |
ec97097bc
|
355 356 |
if (!(shrinker->flags & SHRINKER_NUMA_AWARE)) { shrinkctl->nid = 0; |
1d3d4437e
|
357 |
freed += shrink_slab_node(shrinkctl, shrinker, |
ec97097bc
|
358 359 360 361 362 363 364 365 |
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
|
366 |
|
1da177e4c
|
367 |
} |
1da177e4c
|
368 369 |
} up_read(&shrinker_rwsem); |
f06590bd7
|
370 371 |
out: cond_resched(); |
24f7c6b98
|
372 |
return freed; |
1da177e4c
|
373 |
} |
1da177e4c
|
374 375 |
static inline int is_page_cache_freeable(struct page *page) { |
ceddc3a52
|
376 377 378 379 380 |
/* * 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
|
381 |
return page_count(page) - page_has_private(page) == 2; |
1da177e4c
|
382 |
} |
7d3579e8e
|
383 384 |
static int may_write_to_queue(struct backing_dev_info *bdi, struct scan_control *sc) |
1da177e4c
|
385 |
{ |
930d91525
|
386 |
if (current->flags & PF_SWAPWRITE) |
1da177e4c
|
387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 |
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
|
410 |
lock_page(page); |
3e9f45bd1
|
411 412 |
if (page_mapping(page) == mapping) mapping_set_error(mapping, error); |
1da177e4c
|
413 414 |
unlock_page(page); } |
04e62a29b
|
415 416 417 418 419 420 421 422 423 424 425 |
/* 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
|
426 |
/* |
1742f19fa
|
427 428 |
* pageout is called by shrink_page_list() for each dirty page. * Calls ->writepage(). |
1da177e4c
|
429 |
*/ |
c661b078f
|
430 |
static pageout_t pageout(struct page *page, struct address_space *mapping, |
7d3579e8e
|
431 |
struct scan_control *sc) |
1da177e4c
|
432 433 434 435 436 437 438 439 |
{ /* * 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. * |
6aceb53be
|
440 |
* If this process is currently in __generic_file_aio_write() against |
1da177e4c
|
441 442 443 444 445 446 447 |
* 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
|
448 449 450 451 452 453 454 455 |
*/ 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
|
456 |
if (page_has_private(page)) { |
1da177e4c
|
457 458 |
if (try_to_free_buffers(page)) { ClearPageDirty(page); |
d40cee245
|
459 460 |
printk("%s: orphaned page ", __func__); |
1da177e4c
|
461 462 463 464 465 466 467 |
return PAGE_CLEAN; } } return PAGE_KEEP; } if (mapping->a_ops->writepage == NULL) return PAGE_ACTIVATE; |
0e093d997
|
468 |
if (!may_write_to_queue(mapping->backing_dev_info, sc)) |
1da177e4c
|
469 470 471 472 473 474 475 |
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
|
476 477 |
.range_start = 0, .range_end = LLONG_MAX, |
1da177e4c
|
478 479 480 481 482 483 484 |
.for_reclaim = 1, }; SetPageReclaim(page); res = mapping->a_ops->writepage(page, &wbc); if (res < 0) handle_write_error(mapping, page, res); |
994fc28c7
|
485 |
if (res == AOP_WRITEPAGE_ACTIVATE) { |
1da177e4c
|
486 487 488 |
ClearPageReclaim(page); return PAGE_ACTIVATE; } |
c661b078f
|
489 |
|
1da177e4c
|
490 491 492 493 |
if (!PageWriteback(page)) { /* synchronous write or broken a_ops? */ ClearPageReclaim(page); } |
23b9da55c
|
494 |
trace_mm_vmscan_writepage(page, trace_reclaim_flags(page)); |
e129b5c23
|
495 |
inc_zone_page_state(page, NR_VMSCAN_WRITE); |
1da177e4c
|
496 497 498 499 500 |
return PAGE_SUCCESS; } return PAGE_CLEAN; } |
a649fd927
|
501 |
/* |
e286781d5
|
502 503 |
* Same as remove_mapping, but if the page is removed from the mapping, it * gets returned with a refcount of 0. |
a649fd927
|
504 |
*/ |
e286781d5
|
505 |
static int __remove_mapping(struct address_space *mapping, struct page *page) |
49d2e9cc4
|
506 |
{ |
28e4d965e
|
507 508 |
BUG_ON(!PageLocked(page)); BUG_ON(mapping != page_mapping(page)); |
49d2e9cc4
|
509 |
|
19fd62312
|
510 |
spin_lock_irq(&mapping->tree_lock); |
49d2e9cc4
|
511 |
/* |
0fd0e6b05
|
512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 |
* 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
|
535 |
*/ |
e286781d5
|
536 |
if (!page_freeze_refs(page, 2)) |
49d2e9cc4
|
537 |
goto cannot_free; |
e286781d5
|
538 539 540 |
/* note: atomic_cmpxchg in page_freeze_refs provides the smp_rmb */ if (unlikely(PageDirty(page))) { page_unfreeze_refs(page, 2); |
49d2e9cc4
|
541 |
goto cannot_free; |
e286781d5
|
542 |
} |
49d2e9cc4
|
543 544 545 546 |
if (PageSwapCache(page)) { swp_entry_t swap = { .val = page_private(page) }; __delete_from_swap_cache(page); |
19fd62312
|
547 |
spin_unlock_irq(&mapping->tree_lock); |
cb4b86ba4
|
548 |
swapcache_free(swap, page); |
e286781d5
|
549 |
} else { |
6072d13c4
|
550 551 552 |
void (*freepage)(struct page *); freepage = mapping->a_ops->freepage; |
e64a782fe
|
553 |
__delete_from_page_cache(page); |
19fd62312
|
554 |
spin_unlock_irq(&mapping->tree_lock); |
e767e0561
|
555 |
mem_cgroup_uncharge_cache_page(page); |
6072d13c4
|
556 557 558 |
if (freepage != NULL) freepage(page); |
49d2e9cc4
|
559 |
} |
49d2e9cc4
|
560 561 562 |
return 1; cannot_free: |
19fd62312
|
563 |
spin_unlock_irq(&mapping->tree_lock); |
49d2e9cc4
|
564 565 |
return 0; } |
1da177e4c
|
566 |
/* |
e286781d5
|
567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 |
* 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) { if (__remove_mapping(mapping, page)) { /* * 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
|
585 586 587 588 589 590 591 592 593 |
/** * 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
|
594 595 |
void putback_lru_page(struct page *page) { |
0ec3b74c7
|
596 |
bool is_unevictable; |
bbfd28eee
|
597 |
int was_unevictable = PageUnevictable(page); |
894bc3104
|
598 |
|
309381fea
|
599 |
VM_BUG_ON_PAGE(PageLRU(page), page); |
894bc3104
|
600 601 602 |
redo: ClearPageUnevictable(page); |
39b5f29ac
|
603 |
if (page_evictable(page)) { |
894bc3104
|
604 605 606 607 608 609 |
/* * 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
|
610 |
is_unevictable = false; |
c53954a09
|
611 |
lru_cache_add(page); |
894bc3104
|
612 613 614 615 616 |
} else { /* * Put unevictable pages directly on zone's unevictable * list. */ |
0ec3b74c7
|
617 |
is_unevictable = true; |
894bc3104
|
618 |
add_page_to_unevictable_list(page); |
6a7b95481
|
619 |
/* |
21ee9f398
|
620 621 622 |
* 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
|
623 |
* isolation/check_move_unevictable_pages, |
21ee9f398
|
624 |
* we see PG_mlocked/AS_UNEVICTABLE cleared below and move |
6a7b95481
|
625 626 |
* the page back to the evictable list. * |
21ee9f398
|
627 |
* The other side is TestClearPageMlocked() or shmem_lock(). |
6a7b95481
|
628 629 |
*/ smp_mb(); |
894bc3104
|
630 |
} |
894bc3104
|
631 632 633 634 635 636 |
/* * 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
|
637 |
if (is_unevictable && page_evictable(page)) { |
894bc3104
|
638 639 640 641 642 643 644 645 646 |
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
|
647 |
if (was_unevictable && !is_unevictable) |
bbfd28eee
|
648 |
count_vm_event(UNEVICTABLE_PGRESCUED); |
0ec3b74c7
|
649 |
else if (!was_unevictable && is_unevictable) |
bbfd28eee
|
650 |
count_vm_event(UNEVICTABLE_PGCULLED); |
894bc3104
|
651 652 |
put_page(page); /* drop ref from isolate */ } |
dfc8d636c
|
653 654 655 |
enum page_references { PAGEREF_RECLAIM, PAGEREF_RECLAIM_CLEAN, |
645747462
|
656 |
PAGEREF_KEEP, |
dfc8d636c
|
657 658 659 660 661 662 |
PAGEREF_ACTIVATE, }; static enum page_references page_check_references(struct page *page, struct scan_control *sc) { |
645747462
|
663 |
int referenced_ptes, referenced_page; |
dfc8d636c
|
664 |
unsigned long vm_flags; |
dfc8d636c
|
665 |
|
c3ac9a8ad
|
666 667 |
referenced_ptes = page_referenced(page, 1, sc->target_mem_cgroup, &vm_flags); |
645747462
|
668 |
referenced_page = TestClearPageReferenced(page); |
dfc8d636c
|
669 |
|
dfc8d636c
|
670 671 672 673 674 675 |
/* * 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
|
676 |
if (referenced_ptes) { |
e48982734
|
677 |
if (PageSwapBacked(page)) |
645747462
|
678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 |
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
|
694 |
if (referenced_page || referenced_ptes > 1) |
645747462
|
695 |
return PAGEREF_ACTIVATE; |
c909e9936
|
696 697 698 699 700 |
/* * Activate file-backed executable pages after first usage. */ if (vm_flags & VM_EXEC) return PAGEREF_ACTIVATE; |
645747462
|
701 702 |
return PAGEREF_KEEP; } |
dfc8d636c
|
703 704 |
/* Reclaim if clean, defer dirty pages to writeback */ |
2e30244a7
|
705 |
if (referenced_page && !PageSwapBacked(page)) |
645747462
|
706 707 708 |
return PAGEREF_RECLAIM_CLEAN; return PAGEREF_RECLAIM; |
dfc8d636c
|
709 |
} |
e2be15f6c
|
710 711 712 713 |
/* Check if a page is dirty or under writeback */ static void page_check_dirty_writeback(struct page *page, bool *dirty, bool *writeback) { |
b45972265
|
714 |
struct address_space *mapping; |
e2be15f6c
|
715 716 717 718 719 720 721 722 723 724 725 726 727 |
/* * 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
|
728 729 730 731 732 733 734 735 |
/* 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
|
736 |
} |
e286781d5
|
737 |
/* |
1742f19fa
|
738 |
* shrink_page_list() returns the number of reclaimed pages |
1da177e4c
|
739 |
*/ |
1742f19fa
|
740 |
static unsigned long shrink_page_list(struct list_head *page_list, |
6a18adb35
|
741 |
struct zone *zone, |
f84f6e2b0
|
742 |
struct scan_control *sc, |
02c6de8d7
|
743 |
enum ttu_flags ttu_flags, |
8e9502828
|
744 |
unsigned long *ret_nr_dirty, |
d43006d50
|
745 |
unsigned long *ret_nr_unqueued_dirty, |
8e9502828
|
746 |
unsigned long *ret_nr_congested, |
02c6de8d7
|
747 |
unsigned long *ret_nr_writeback, |
b1a6f21e3
|
748 |
unsigned long *ret_nr_immediate, |
02c6de8d7
|
749 |
bool force_reclaim) |
1da177e4c
|
750 751 |
{ LIST_HEAD(ret_pages); |
abe4c3b50
|
752 |
LIST_HEAD(free_pages); |
1da177e4c
|
753 |
int pgactivate = 0; |
d43006d50
|
754 |
unsigned long nr_unqueued_dirty = 0; |
0e093d997
|
755 756 |
unsigned long nr_dirty = 0; unsigned long nr_congested = 0; |
05ff51376
|
757 |
unsigned long nr_reclaimed = 0; |
92df3a723
|
758 |
unsigned long nr_writeback = 0; |
b1a6f21e3
|
759 |
unsigned long nr_immediate = 0; |
1da177e4c
|
760 761 |
cond_resched(); |
69980e317
|
762 |
mem_cgroup_uncharge_start(); |
1da177e4c
|
763 764 765 766 |
while (!list_empty(page_list)) { struct address_space *mapping; struct page *page; int may_enter_fs; |
02c6de8d7
|
767 |
enum page_references references = PAGEREF_RECLAIM_CLEAN; |
e2be15f6c
|
768 |
bool dirty, writeback; |
1da177e4c
|
769 770 771 772 773 |
cond_resched(); page = lru_to_page(page_list); list_del(&page->lru); |
529ae9aaa
|
774 |
if (!trylock_page(page)) |
1da177e4c
|
775 |
goto keep; |
309381fea
|
776 777 |
VM_BUG_ON_PAGE(PageActive(page), page); VM_BUG_ON_PAGE(page_zone(page) != zone, page); |
1da177e4c
|
778 779 |
sc->nr_scanned++; |
80e434260
|
780 |
|
39b5f29ac
|
781 |
if (unlikely(!page_evictable(page))) |
b291f0003
|
782 |
goto cull_mlocked; |
894bc3104
|
783 |
|
a6dc60f89
|
784 |
if (!sc->may_unmap && page_mapped(page)) |
80e434260
|
785 |
goto keep_locked; |
1da177e4c
|
786 787 788 |
/* Double the slab pressure for mapped and swapcache pages */ if (page_mapped(page) || PageSwapCache(page)) sc->nr_scanned++; |
c661b078f
|
789 790 |
may_enter_fs = (sc->gfp_mask & __GFP_FS) || (PageSwapCache(page) && (sc->gfp_mask & __GFP_IO)); |
283aba9f9
|
791 |
/* |
e2be15f6c
|
792 793 794 795 796 797 798 799 800 801 802 |
* 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
|
803 804 805 806 807 808 |
/* * 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
|
809 |
mapping = page_mapping(page); |
d04e8acd0
|
810 811 |
if ((mapping && bdi_write_congested(mapping->backing_dev_info)) || (writeback && PageReclaim(page))) |
e2be15f6c
|
812 813 814 |
nr_congested++; /* |
283aba9f9
|
815 816 817 818 819 820 821 822 823 824 |
* 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
|
825 826 |
* note that the LRU is being scanned too quickly and the * caller can stall after page list has been processed. |
283aba9f9
|
827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 |
* * 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
|
851 |
if (PageWriteback(page)) { |
283aba9f9
|
852 853 854 855 |
/* Case 1 above */ if (current_is_kswapd() && PageReclaim(page) && zone_is_reclaim_writeback(zone)) { |
b1a6f21e3
|
856 857 |
nr_immediate++; goto keep_locked; |
283aba9f9
|
858 859 860 |
/* Case 2 above */ } else if (global_reclaim(sc) || |
c3b94f44f
|
861 862 863 864 865 866 867 868 869 870 871 872 873 |
!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
|
874 |
nr_writeback++; |
283aba9f9
|
875 |
|
c3b94f44f
|
876 |
goto keep_locked; |
283aba9f9
|
877 878 879 880 |
/* Case 3 above */ } else { wait_on_page_writeback(page); |
e62e384e9
|
881 |
} |
c661b078f
|
882 |
} |
1da177e4c
|
883 |
|
02c6de8d7
|
884 885 |
if (!force_reclaim) references = page_check_references(page, sc); |
dfc8d636c
|
886 887 |
switch (references) { case PAGEREF_ACTIVATE: |
1da177e4c
|
888 |
goto activate_locked; |
645747462
|
889 890 |
case PAGEREF_KEEP: goto keep_locked; |
dfc8d636c
|
891 892 893 894 |
case PAGEREF_RECLAIM: case PAGEREF_RECLAIM_CLEAN: ; /* try to reclaim the page below */ } |
1da177e4c
|
895 |
|
1da177e4c
|
896 897 898 899 |
/* * Anonymous process memory has backing store? * Try to allocate it some swap space here. */ |
b291f0003
|
900 |
if (PageAnon(page) && !PageSwapCache(page)) { |
63eb6b93c
|
901 902 |
if (!(sc->gfp_mask & __GFP_IO)) goto keep_locked; |
5bc7b8aca
|
903 |
if (!add_to_swap(page, page_list)) |
1da177e4c
|
904 |
goto activate_locked; |
63eb6b93c
|
905 |
may_enter_fs = 1; |
1da177e4c
|
906 |
|
e2be15f6c
|
907 908 909 |
/* Adding to swap updated mapping */ mapping = page_mapping(page); } |
1da177e4c
|
910 911 912 913 914 915 |
/* * The page is mapped into the page tables of one or more * processes. Try to unmap it here. */ if (page_mapped(page) && mapping) { |
02c6de8d7
|
916 |
switch (try_to_unmap(page, ttu_flags)) { |
1da177e4c
|
917 918 919 920 |
case SWAP_FAIL: goto activate_locked; case SWAP_AGAIN: goto keep_locked; |
b291f0003
|
921 922 |
case SWAP_MLOCK: goto cull_mlocked; |
1da177e4c
|
923 924 925 926 927 928 |
case SWAP_SUCCESS: ; /* try to free the page below */ } } if (PageDirty(page)) { |
ee72886d8
|
929 930 |
/* * Only kswapd can writeback filesystem pages to |
d43006d50
|
931 932 |
* avoid risk of stack overflow but only writeback * if many dirty pages have been encountered. |
ee72886d8
|
933 |
*/ |
f84f6e2b0
|
934 |
if (page_is_file_cache(page) && |
9e3b2f8cd
|
935 |
(!current_is_kswapd() || |
d43006d50
|
936 |
!zone_is_reclaim_dirty(zone))) { |
49ea7eb65
|
937 938 939 940 941 942 943 944 |
/* * 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
|
945 946 |
goto keep_locked; } |
dfc8d636c
|
947 |
if (references == PAGEREF_RECLAIM_CLEAN) |
1da177e4c
|
948 |
goto keep_locked; |
4dd4b9202
|
949 |
if (!may_enter_fs) |
1da177e4c
|
950 |
goto keep_locked; |
52a8363ea
|
951 |
if (!sc->may_writepage) |
1da177e4c
|
952 953 954 |
goto keep_locked; /* Page is dirty, try to write it out here */ |
7d3579e8e
|
955 |
switch (pageout(page, mapping, sc)) { |
1da177e4c
|
956 957 958 959 960 |
case PAGE_KEEP: goto keep_locked; case PAGE_ACTIVATE: goto activate_locked; case PAGE_SUCCESS: |
7d3579e8e
|
961 |
if (PageWriteback(page)) |
41ac1999c
|
962 |
goto keep; |
7d3579e8e
|
963 |
if (PageDirty(page)) |
1da177e4c
|
964 |
goto keep; |
7d3579e8e
|
965 |
|
1da177e4c
|
966 967 968 969 |
/* * A synchronous write - probably a ramdisk. Go * ahead and try to reclaim the page. */ |
529ae9aaa
|
970 |
if (!trylock_page(page)) |
1da177e4c
|
971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 |
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
|
990 |
* will do this, as well as the blockdev mapping. |
1da177e4c
|
991 992 993 994 995 996 997 998 999 1000 |
* 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
|
1001 |
if (page_has_private(page)) { |
1da177e4c
|
1002 1003 |
if (!try_to_release_page(page, sc->gfp_mask)) goto activate_locked; |
e286781d5
|
1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 |
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
|
1020 |
} |
e286781d5
|
1021 |
if (!mapping || !__remove_mapping(mapping, page)) |
49d2e9cc4
|
1022 |
goto keep_locked; |
1da177e4c
|
1023 |
|
a978d6f52
|
1024 1025 1026 1027 1028 1029 1030 1031 |
/* * 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
|
1032 |
free_it: |
05ff51376
|
1033 |
nr_reclaimed++; |
abe4c3b50
|
1034 1035 1036 1037 1038 1039 |
/* * 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
|
1040 |
continue; |
b291f0003
|
1041 |
cull_mlocked: |
63d6c5ad7
|
1042 1043 |
if (PageSwapCache(page)) try_to_free_swap(page); |
b291f0003
|
1044 1045 1046 |
unlock_page(page); putback_lru_page(page); continue; |
1da177e4c
|
1047 |
activate_locked: |
68a22394c
|
1048 1049 |
/* Not a candidate for swapping, so reclaim swap space. */ if (PageSwapCache(page) && vm_swap_full()) |
a2c43eed8
|
1050 |
try_to_free_swap(page); |
309381fea
|
1051 |
VM_BUG_ON_PAGE(PageActive(page), page); |
1da177e4c
|
1052 1053 1054 1055 1056 1057 |
SetPageActive(page); pgactivate++; keep_locked: unlock_page(page); keep: list_add(&page->lru, &ret_pages); |
309381fea
|
1058 |
VM_BUG_ON_PAGE(PageLRU(page) || PageUnevictable(page), page); |
1da177e4c
|
1059 |
} |
abe4c3b50
|
1060 |
|
cc59850ef
|
1061 |
free_hot_cold_page_list(&free_pages, 1); |
abe4c3b50
|
1062 |
|
1da177e4c
|
1063 |
list_splice(&ret_pages, page_list); |
f8891e5e1
|
1064 |
count_vm_events(PGACTIVATE, pgactivate); |
69980e317
|
1065 |
mem_cgroup_uncharge_end(); |
8e9502828
|
1066 1067 |
*ret_nr_dirty += nr_dirty; *ret_nr_congested += nr_congested; |
d43006d50
|
1068 |
*ret_nr_unqueued_dirty += nr_unqueued_dirty; |
92df3a723
|
1069 |
*ret_nr_writeback += nr_writeback; |
b1a6f21e3
|
1070 |
*ret_nr_immediate += nr_immediate; |
05ff51376
|
1071 |
return nr_reclaimed; |
1da177e4c
|
1072 |
} |
02c6de8d7
|
1073 1074 1075 1076 1077 1078 1079 1080 |
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
|
1081 |
unsigned long ret, dummy1, dummy2, dummy3, dummy4, dummy5; |
02c6de8d7
|
1082 1083 1084 1085 |
struct page *page, *next; LIST_HEAD(clean_pages); list_for_each_entry_safe(page, next, page_list, lru) { |
117aad1e9
|
1086 1087 |
if (page_is_file_cache(page) && !PageDirty(page) && !isolated_balloon_page(page)) { |
02c6de8d7
|
1088 1089 1090 1091 1092 1093 |
ClearPageActive(page); list_move(&page->lru, &clean_pages); } } ret = shrink_page_list(&clean_pages, zone, &sc, |
8e9502828
|
1094 1095 |
TTU_UNMAP|TTU_IGNORE_ACCESS, &dummy1, &dummy2, &dummy3, &dummy4, &dummy5, true); |
02c6de8d7
|
1096 1097 1098 1099 |
list_splice(&clean_pages, page_list); __mod_zone_page_state(zone, NR_ISOLATED_FILE, -ret); return ret; } |
5ad333eb6
|
1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 |
/* * 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
|
1110 |
int __isolate_lru_page(struct page *page, isolate_mode_t mode) |
5ad333eb6
|
1111 1112 1113 1114 1115 1116 |
{ int ret = -EINVAL; /* Only take pages on the LRU. */ if (!PageLRU(page)) return ret; |
e46a28790
|
1117 1118 |
/* Compaction should not handle unevictable pages but CMA can do so */ if (PageUnevictable(page) && !(mode & ISOLATE_UNEVICTABLE)) |
894bc3104
|
1119 |
return ret; |
5ad333eb6
|
1120 |
ret = -EBUSY; |
08e552c69
|
1121 |
|
c82449352
|
1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 |
/* * 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
|
1155 |
|
f80c06736
|
1156 1157 |
if ((mode & ISOLATE_UNMAPPED) && page_mapped(page)) return ret; |
5ad333eb6
|
1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 |
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
|
1170 |
/* |
1da177e4c
|
1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 |
* 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
|
1181 |
* @lruvec: The LRU vector to pull pages from. |
1da177e4c
|
1182 |
* @dst: The temp list to put pages on to. |
f626012db
|
1183 |
* @nr_scanned: The number of pages that were scanned. |
fe2c2a106
|
1184 |
* @sc: The scan_control struct for this reclaim session |
5ad333eb6
|
1185 |
* @mode: One of the LRU isolation modes |
3cb994517
|
1186 |
* @lru: LRU list id for isolating |
1da177e4c
|
1187 1188 1189 |
* * returns how many pages were moved onto *@dst. */ |
69e05944a
|
1190 |
static unsigned long isolate_lru_pages(unsigned long nr_to_scan, |
5dc35979e
|
1191 |
struct lruvec *lruvec, struct list_head *dst, |
fe2c2a106
|
1192 |
unsigned long *nr_scanned, struct scan_control *sc, |
3cb994517
|
1193 |
isolate_mode_t mode, enum lru_list lru) |
1da177e4c
|
1194 |
{ |
75b00af77
|
1195 |
struct list_head *src = &lruvec->lists[lru]; |
69e05944a
|
1196 |
unsigned long nr_taken = 0; |
c9b02d970
|
1197 |
unsigned long scan; |
1da177e4c
|
1198 |
|
c9b02d970
|
1199 |
for (scan = 0; scan < nr_to_scan && !list_empty(src); scan++) { |
5ad333eb6
|
1200 |
struct page *page; |
fa9add641
|
1201 |
int nr_pages; |
5ad333eb6
|
1202 |
|
1da177e4c
|
1203 1204 |
page = lru_to_page(src); prefetchw_prev_lru_page(page, src, flags); |
309381fea
|
1205 |
VM_BUG_ON_PAGE(!PageLRU(page), page); |
8d438f96d
|
1206 |
|
f3fd4a619
|
1207 |
switch (__isolate_lru_page(page, mode)) { |
5ad333eb6
|
1208 |
case 0: |
fa9add641
|
1209 1210 |
nr_pages = hpage_nr_pages(page); mem_cgroup_update_lru_size(lruvec, lru, -nr_pages); |
5ad333eb6
|
1211 |
list_move(&page->lru, dst); |
fa9add641
|
1212 |
nr_taken += nr_pages; |
5ad333eb6
|
1213 1214 1215 1216 1217 1218 |
break; case -EBUSY: /* else it is being freed elsewhere */ list_move(&page->lru, src); continue; |
46453a6e1
|
1219 |
|
5ad333eb6
|
1220 1221 1222 |
default: BUG(); } |
1da177e4c
|
1223 |
} |
f626012db
|
1224 |
*nr_scanned = scan; |
75b00af77
|
1225 1226 |
trace_mm_vmscan_lru_isolate(sc->order, nr_to_scan, scan, nr_taken, mode, is_file_lru(lru)); |
1da177e4c
|
1227 1228 |
return nr_taken; } |
62695a84e
|
1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 |
/** * 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
|
1240 1241 1242 |
* 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
|
1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 |
* * 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
|
1257 |
VM_BUG_ON_PAGE(!page_count(page), page); |
0c917313a
|
1258 |
|
62695a84e
|
1259 1260 |
if (PageLRU(page)) { struct zone *zone = page_zone(page); |
fa9add641
|
1261 |
struct lruvec *lruvec; |
62695a84e
|
1262 1263 |
spin_lock_irq(&zone->lru_lock); |
fa9add641
|
1264 |
lruvec = mem_cgroup_page_lruvec(page, zone); |
0c917313a
|
1265 |
if (PageLRU(page)) { |
894bc3104
|
1266 |
int lru = page_lru(page); |
0c917313a
|
1267 |
get_page(page); |
62695a84e
|
1268 |
ClearPageLRU(page); |
fa9add641
|
1269 1270 |
del_page_from_lru_list(page, lruvec, lru); ret = 0; |
62695a84e
|
1271 1272 1273 1274 1275 |
} spin_unlock_irq(&zone->lru_lock); } return ret; } |
5ad333eb6
|
1276 |
/* |
d37dd5dcb
|
1277 1278 1279 1280 1281 |
* 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
|
1282 1283 1284 1285 1286 1287 1288 1289 |
*/ 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
|
1290 |
if (!global_reclaim(sc)) |
35cd78156
|
1291 1292 1293 1294 1295 1296 1297 1298 1299 |
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
|
1300 1301 1302 1303 1304 1305 1306 |
/* * 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
|
1307 1308 |
return isolated > inactive; } |
666356297
|
1309 |
static noinline_for_stack void |
75b00af77
|
1310 |
putback_inactive_pages(struct lruvec *lruvec, struct list_head *page_list) |
666356297
|
1311 |
{ |
27ac81d85
|
1312 1313 |
struct zone_reclaim_stat *reclaim_stat = &lruvec->reclaim_stat; struct zone *zone = lruvec_zone(lruvec); |
3f79768f2
|
1314 |
LIST_HEAD(pages_to_free); |
666356297
|
1315 |
|
666356297
|
1316 1317 1318 |
/* * Put back any unfreeable pages. */ |
666356297
|
1319 |
while (!list_empty(page_list)) { |
3f79768f2
|
1320 |
struct page *page = lru_to_page(page_list); |
666356297
|
1321 |
int lru; |
3f79768f2
|
1322 |
|
309381fea
|
1323 |
VM_BUG_ON_PAGE(PageLRU(page), page); |
666356297
|
1324 |
list_del(&page->lru); |
39b5f29ac
|
1325 |
if (unlikely(!page_evictable(page))) { |
666356297
|
1326 1327 1328 1329 1330 |
spin_unlock_irq(&zone->lru_lock); putback_lru_page(page); spin_lock_irq(&zone->lru_lock); continue; } |
fa9add641
|
1331 1332 |
lruvec = mem_cgroup_page_lruvec(page, zone); |
7a608572a
|
1333 |
SetPageLRU(page); |
666356297
|
1334 |
lru = page_lru(page); |
fa9add641
|
1335 |
add_page_to_lru_list(page, lruvec, lru); |
666356297
|
1336 1337 |
if (is_active_lru(lru)) { int file = is_file_lru(lru); |
9992af102
|
1338 1339 |
int numpages = hpage_nr_pages(page); reclaim_stat->recent_rotated[file] += numpages; |
666356297
|
1340 |
} |
2bcf88796
|
1341 1342 1343 |
if (put_page_testzero(page)) { __ClearPageLRU(page); __ClearPageActive(page); |
fa9add641
|
1344 |
del_page_from_lru_list(page, lruvec, lru); |
2bcf88796
|
1345 1346 1347 1348 1349 1350 1351 |
if (unlikely(PageCompound(page))) { spin_unlock_irq(&zone->lru_lock); (*get_compound_page_dtor(page))(page); spin_lock_irq(&zone->lru_lock); } else list_add(&page->lru, &pages_to_free); |
666356297
|
1352 1353 |
} } |
666356297
|
1354 |
|
3f79768f2
|
1355 1356 1357 1358 |
/* * To save our caller's stack, now use input list for pages to free. */ list_splice(&pages_to_free, page_list); |
666356297
|
1359 1360 1361 |
} /* |
1742f19fa
|
1362 1363 |
* shrink_inactive_list() is a helper for shrink_zone(). It returns the number * of reclaimed pages |
1da177e4c
|
1364 |
*/ |
666356297
|
1365 |
static noinline_for_stack unsigned long |
1a93be0e7
|
1366 |
shrink_inactive_list(unsigned long nr_to_scan, struct lruvec *lruvec, |
9e3b2f8cd
|
1367 |
struct scan_control *sc, enum lru_list lru) |
1da177e4c
|
1368 1369 |
{ LIST_HEAD(page_list); |
e247dbce5
|
1370 |
unsigned long nr_scanned; |
05ff51376
|
1371 |
unsigned long nr_reclaimed = 0; |
e247dbce5
|
1372 |
unsigned long nr_taken; |
8e9502828
|
1373 1374 |
unsigned long nr_dirty = 0; unsigned long nr_congested = 0; |
e2be15f6c
|
1375 |
unsigned long nr_unqueued_dirty = 0; |
92df3a723
|
1376 |
unsigned long nr_writeback = 0; |
b1a6f21e3
|
1377 |
unsigned long nr_immediate = 0; |
f3fd4a619
|
1378 |
isolate_mode_t isolate_mode = 0; |
3cb994517
|
1379 |
int file = is_file_lru(lru); |
1a93be0e7
|
1380 1381 |
struct zone *zone = lruvec_zone(lruvec); struct zone_reclaim_stat *reclaim_stat = &lruvec->reclaim_stat; |
78dc583d3
|
1382 |
|
35cd78156
|
1383 |
while (unlikely(too_many_isolated(zone, file, sc))) { |
58355c787
|
1384 |
congestion_wait(BLK_RW_ASYNC, HZ/10); |
35cd78156
|
1385 1386 1387 1388 1389 |
/* We are about to die and free our memory. Return now. */ if (fatal_signal_pending(current)) return SWAP_CLUSTER_MAX; } |
1da177e4c
|
1390 |
lru_add_drain(); |
f80c06736
|
1391 1392 |
if (!sc->may_unmap) |
613172891
|
1393 |
isolate_mode |= ISOLATE_UNMAPPED; |
f80c06736
|
1394 |
if (!sc->may_writepage) |
613172891
|
1395 |
isolate_mode |= ISOLATE_CLEAN; |
f80c06736
|
1396 |
|
1da177e4c
|
1397 |
spin_lock_irq(&zone->lru_lock); |
b35ea17b7
|
1398 |
|
5dc35979e
|
1399 1400 |
nr_taken = isolate_lru_pages(nr_to_scan, lruvec, &page_list, &nr_scanned, sc, isolate_mode, lru); |
95d918fc0
|
1401 1402 1403 |
__mod_zone_page_state(zone, NR_LRU_BASE + lru, -nr_taken); __mod_zone_page_state(zone, NR_ISOLATED_ANON + file, nr_taken); |
89b5fae53
|
1404 |
if (global_reclaim(sc)) { |
e247dbce5
|
1405 1406 |
zone->pages_scanned += nr_scanned; if (current_is_kswapd()) |
75b00af77
|
1407 |
__count_zone_vm_events(PGSCAN_KSWAPD, zone, nr_scanned); |
e247dbce5
|
1408 |
else |
75b00af77
|
1409 |
__count_zone_vm_events(PGSCAN_DIRECT, zone, nr_scanned); |
e247dbce5
|
1410 |
} |
d563c0501
|
1411 |
spin_unlock_irq(&zone->lru_lock); |
b35ea17b7
|
1412 |
|
d563c0501
|
1413 |
if (nr_taken == 0) |
666356297
|
1414 |
return 0; |
5ad333eb6
|
1415 |
|
02c6de8d7
|
1416 |
nr_reclaimed = shrink_page_list(&page_list, zone, sc, TTU_UNMAP, |
8e9502828
|
1417 1418 1419 |
&nr_dirty, &nr_unqueued_dirty, &nr_congested, &nr_writeback, &nr_immediate, false); |
c661b078f
|
1420 |
|
3f79768f2
|
1421 |
spin_lock_irq(&zone->lru_lock); |
95d918fc0
|
1422 |
reclaim_stat->recent_scanned[file] += nr_taken; |
d563c0501
|
1423 |
|
904249aa6
|
1424 1425 1426 1427 1428 1429 1430 1431 |
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
|
1432 |
|
27ac81d85
|
1433 |
putback_inactive_pages(lruvec, &page_list); |
3f79768f2
|
1434 |
|
95d918fc0
|
1435 |
__mod_zone_page_state(zone, NR_ISOLATED_ANON + file, -nr_taken); |
3f79768f2
|
1436 1437 1438 1439 |
spin_unlock_irq(&zone->lru_lock); free_hot_cold_page_list(&page_list, 1); |
e11da5b4f
|
1440 |
|
92df3a723
|
1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 |
/* * 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
|
1451 1452 1453 |
* 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
|
1454 |
*/ |
918fc718c
|
1455 |
if (nr_writeback && nr_writeback == nr_taken) |
283aba9f9
|
1456 |
zone_set_flag(zone, ZONE_WRITEBACK); |
92df3a723
|
1457 |
|
d43006d50
|
1458 |
/* |
b1a6f21e3
|
1459 1460 |
* memcg will stall in page writeback so only consider forcibly * stalling for global reclaim |
d43006d50
|
1461 |
*/ |
b1a6f21e3
|
1462 1463 |
if (global_reclaim(sc)) { /* |
8e9502828
|
1464 1465 1466 1467 1468 1469 1470 |
* 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) zone_set_flag(zone, ZONE_CONGESTED); /* |
b1a6f21e3
|
1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 |
* If dirty pages are scanned that are not queued for IO, it * implies that flushers are not keeping up. In this case, flag * the zone ZONE_TAIL_LRU_DIRTY and kswapd will start writing * pages from reclaim context. It will forcibly stall in the * next check. */ if (nr_unqueued_dirty == nr_taken) zone_set_flag(zone, ZONE_TAIL_LRU_DIRTY); /* * In addition, 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 * they are written so also forcibly stall. */ if (nr_unqueued_dirty == nr_taken || nr_immediate) congestion_wait(BLK_RW_ASYNC, HZ/10); |
e2be15f6c
|
1488 |
} |
d43006d50
|
1489 |
|
8e9502828
|
1490 1491 1492 1493 1494 1495 1496 |
/* * 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. */ if (!sc->hibernation_mode && !current_is_kswapd()) wait_iff_congested(zone, BLK_RW_ASYNC, HZ/10); |
e11da5b4f
|
1497 1498 1499 |
trace_mm_vmscan_lru_shrink_inactive(zone->zone_pgdat->node_id, zone_idx(zone), nr_scanned, nr_reclaimed, |
9e3b2f8cd
|
1500 |
sc->priority, |
23b9da55c
|
1501 |
trace_shrink_flags(file)); |
05ff51376
|
1502 |
return nr_reclaimed; |
1da177e4c
|
1503 |
} |
3bb1a852a
|
1504 |
/* |
1da177e4c
|
1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 |
* 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
|
1521 |
|
fa9add641
|
1522 |
static void move_active_pages_to_lru(struct lruvec *lruvec, |
3eb4140f0
|
1523 |
struct list_head *list, |
2bcf88796
|
1524 |
struct list_head *pages_to_free, |
3eb4140f0
|
1525 1526 |
enum lru_list lru) { |
fa9add641
|
1527 |
struct zone *zone = lruvec_zone(lruvec); |
3eb4140f0
|
1528 |
unsigned long pgmoved = 0; |
3eb4140f0
|
1529 |
struct page *page; |
fa9add641
|
1530 |
int nr_pages; |
3eb4140f0
|
1531 |
|
3eb4140f0
|
1532 1533 |
while (!list_empty(list)) { page = lru_to_page(list); |
fa9add641
|
1534 |
lruvec = mem_cgroup_page_lruvec(page, zone); |
3eb4140f0
|
1535 |
|
309381fea
|
1536 |
VM_BUG_ON_PAGE(PageLRU(page), page); |
3eb4140f0
|
1537 |
SetPageLRU(page); |
fa9add641
|
1538 1539 |
nr_pages = hpage_nr_pages(page); mem_cgroup_update_lru_size(lruvec, lru, nr_pages); |
925b7673c
|
1540 |
list_move(&page->lru, &lruvec->lists[lru]); |
fa9add641
|
1541 |
pgmoved += nr_pages; |
3eb4140f0
|
1542 |
|
2bcf88796
|
1543 1544 1545 |
if (put_page_testzero(page)) { __ClearPageLRU(page); __ClearPageActive(page); |
fa9add641
|
1546 |
del_page_from_lru_list(page, lruvec, lru); |
2bcf88796
|
1547 1548 1549 1550 1551 1552 1553 |
if (unlikely(PageCompound(page))) { spin_unlock_irq(&zone->lru_lock); (*get_compound_page_dtor(page))(page); spin_lock_irq(&zone->lru_lock); } else list_add(&page->lru, pages_to_free); |
3eb4140f0
|
1554 1555 1556 1557 1558 1559 |
} } __mod_zone_page_state(zone, NR_LRU_BASE + lru, pgmoved); if (!is_active_lru(lru)) __count_vm_events(PGDEACTIVATE, pgmoved); } |
1cfb419b3
|
1560 |
|
f626012db
|
1561 |
static void shrink_active_list(unsigned long nr_to_scan, |
1a93be0e7
|
1562 |
struct lruvec *lruvec, |
f16015fbf
|
1563 |
struct scan_control *sc, |
9e3b2f8cd
|
1564 |
enum lru_list lru) |
1da177e4c
|
1565 |
{ |
44c241f16
|
1566 |
unsigned long nr_taken; |
f626012db
|
1567 |
unsigned long nr_scanned; |
6fe6b7e35
|
1568 |
unsigned long vm_flags; |
1da177e4c
|
1569 |
LIST_HEAD(l_hold); /* The pages which were snipped off */ |
8cab4754d
|
1570 |
LIST_HEAD(l_active); |
b69408e88
|
1571 |
LIST_HEAD(l_inactive); |
1da177e4c
|
1572 |
struct page *page; |
1a93be0e7
|
1573 |
struct zone_reclaim_stat *reclaim_stat = &lruvec->reclaim_stat; |
44c241f16
|
1574 |
unsigned long nr_rotated = 0; |
f3fd4a619
|
1575 |
isolate_mode_t isolate_mode = 0; |
3cb994517
|
1576 |
int file = is_file_lru(lru); |
1a93be0e7
|
1577 |
struct zone *zone = lruvec_zone(lruvec); |
1da177e4c
|
1578 1579 |
lru_add_drain(); |
f80c06736
|
1580 1581 |
if (!sc->may_unmap) |
613172891
|
1582 |
isolate_mode |= ISOLATE_UNMAPPED; |
f80c06736
|
1583 |
if (!sc->may_writepage) |
613172891
|
1584 |
isolate_mode |= ISOLATE_CLEAN; |
f80c06736
|
1585 |
|
1da177e4c
|
1586 |
spin_lock_irq(&zone->lru_lock); |
925b7673c
|
1587 |
|
5dc35979e
|
1588 1589 |
nr_taken = isolate_lru_pages(nr_to_scan, lruvec, &l_hold, &nr_scanned, sc, isolate_mode, lru); |
89b5fae53
|
1590 |
if (global_reclaim(sc)) |
f626012db
|
1591 |
zone->pages_scanned += nr_scanned; |
89b5fae53
|
1592 |
|
b7c46d151
|
1593 |
reclaim_stat->recent_scanned[file] += nr_taken; |
1cfb419b3
|
1594 |
|
f626012db
|
1595 |
__count_zone_vm_events(PGREFILL, zone, nr_scanned); |
3cb994517
|
1596 |
__mod_zone_page_state(zone, NR_LRU_BASE + lru, -nr_taken); |
a731286de
|
1597 |
__mod_zone_page_state(zone, NR_ISOLATED_ANON + file, nr_taken); |
1da177e4c
|
1598 |
spin_unlock_irq(&zone->lru_lock); |
1da177e4c
|
1599 1600 1601 1602 |
while (!list_empty(&l_hold)) { cond_resched(); page = lru_to_page(&l_hold); list_del(&page->lru); |
7e9cd4842
|
1603 |
|
39b5f29ac
|
1604 |
if (unlikely(!page_evictable(page))) { |
894bc3104
|
1605 1606 1607 |
putback_lru_page(page); continue; } |
cc715d99e
|
1608 1609 1610 1611 1612 1613 1614 |
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
|
1615 1616 |
if (page_referenced(page, 0, sc->target_mem_cgroup, &vm_flags)) { |
9992af102
|
1617 |
nr_rotated += hpage_nr_pages(page); |
8cab4754d
|
1618 1619 1620 1621 1622 1623 1624 1625 1626 |
/* * 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
|
1627 |
if ((vm_flags & VM_EXEC) && page_is_file_cache(page)) { |
8cab4754d
|
1628 1629 1630 1631 |
list_add(&page->lru, &l_active); continue; } } |
7e9cd4842
|
1632 |
|
5205e56ee
|
1633 |
ClearPageActive(page); /* we are de-activating */ |
1da177e4c
|
1634 1635 |
list_add(&page->lru, &l_inactive); } |
b555749aa
|
1636 |
/* |
8cab4754d
|
1637 |
* Move pages back to the lru list. |
b555749aa
|
1638 |
*/ |
2a1dc5097
|
1639 |
spin_lock_irq(&zone->lru_lock); |
4f98a2fee
|
1640 |
/* |
8cab4754d
|
1641 1642 1643 1644 |
* 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 * get_scan_ratio. |
7e9cd4842
|
1645 |
*/ |
b7c46d151
|
1646 |
reclaim_stat->recent_rotated[file] += nr_rotated; |
556adecba
|
1647 |
|
fa9add641
|
1648 1649 |
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
|
1650 |
__mod_zone_page_state(zone, NR_ISOLATED_ANON + file, -nr_taken); |
f8891e5e1
|
1651 |
spin_unlock_irq(&zone->lru_lock); |
2bcf88796
|
1652 1653 |
free_hot_cold_page_list(&l_hold, 1); |
1da177e4c
|
1654 |
} |
74e3f3c33
|
1655 |
#ifdef CONFIG_SWAP |
14797e236
|
1656 |
static int inactive_anon_is_low_global(struct zone *zone) |
f89eb90e3
|
1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 |
{ 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
|
1668 1669 |
/** * inactive_anon_is_low - check if anonymous pages need to be deactivated |
c56d5c7df
|
1670 |
* @lruvec: LRU vector to check |
14797e236
|
1671 1672 1673 1674 |
* * Returns true if the zone does not have enough inactive anon pages, * meaning some active anon pages need to be deactivated. */ |
c56d5c7df
|
1675 |
static int inactive_anon_is_low(struct lruvec *lruvec) |
14797e236
|
1676 |
{ |
74e3f3c33
|
1677 1678 1679 1680 1681 1682 |
/* * If we don't have swap space, anonymous page deactivation * is pointless. */ if (!total_swap_pages) return 0; |
c3c787e8c
|
1683 |
if (!mem_cgroup_disabled()) |
c56d5c7df
|
1684 |
return mem_cgroup_inactive_anon_is_low(lruvec); |
f16015fbf
|
1685 |
|
c56d5c7df
|
1686 |
return inactive_anon_is_low_global(lruvec_zone(lruvec)); |
14797e236
|
1687 |
} |
74e3f3c33
|
1688 |
#else |
c56d5c7df
|
1689 |
static inline int inactive_anon_is_low(struct lruvec *lruvec) |
74e3f3c33
|
1690 1691 1692 1693 |
{ return 0; } #endif |
14797e236
|
1694 |
|
56e49d218
|
1695 1696 |
/** * inactive_file_is_low - check if file pages need to be deactivated |
c56d5c7df
|
1697 |
* @lruvec: LRU vector to check |
56e49d218
|
1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 |
* * 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
|
1709 |
static int inactive_file_is_low(struct lruvec *lruvec) |
56e49d218
|
1710 |
{ |
e3790144c
|
1711 1712 1713 1714 1715 |
unsigned long inactive; unsigned long active; inactive = get_lru_size(lruvec, LRU_INACTIVE_FILE); active = get_lru_size(lruvec, LRU_ACTIVE_FILE); |
56e49d218
|
1716 |
|
e3790144c
|
1717 |
return active > inactive; |
56e49d218
|
1718 |
} |
75b00af77
|
1719 |
static int inactive_list_is_low(struct lruvec *lruvec, enum lru_list lru) |
b39415b27
|
1720 |
{ |
75b00af77
|
1721 |
if (is_file_lru(lru)) |
c56d5c7df
|
1722 |
return inactive_file_is_low(lruvec); |
b39415b27
|
1723 |
else |
c56d5c7df
|
1724 |
return inactive_anon_is_low(lruvec); |
b39415b27
|
1725 |
} |
4f98a2fee
|
1726 |
static unsigned long shrink_list(enum lru_list lru, unsigned long nr_to_scan, |
1a93be0e7
|
1727 |
struct lruvec *lruvec, struct scan_control *sc) |
b69408e88
|
1728 |
{ |
b39415b27
|
1729 |
if (is_active_lru(lru)) { |
75b00af77
|
1730 |
if (inactive_list_is_low(lruvec, lru)) |
1a93be0e7
|
1731 |
shrink_active_list(nr_to_scan, lruvec, sc, lru); |
556adecba
|
1732 1733 |
return 0; } |
1a93be0e7
|
1734 |
return shrink_inactive_list(nr_to_scan, lruvec, sc, lru); |
4f98a2fee
|
1735 |
} |
3d58ab5c9
|
1736 |
static int vmscan_swappiness(struct scan_control *sc) |
1f4c025b5
|
1737 |
{ |
89b5fae53
|
1738 |
if (global_reclaim(sc)) |
1f4c025b5
|
1739 |
return vm_swappiness; |
3d58ab5c9
|
1740 |
return mem_cgroup_swappiness(sc->target_mem_cgroup); |
1f4c025b5
|
1741 |
} |
9a2651140
|
1742 1743 1744 1745 1746 1747 |
enum scan_balance { SCAN_EQUAL, SCAN_FRACT, SCAN_ANON, SCAN_FILE, }; |
4f98a2fee
|
1748 1749 1750 1751 1752 1753 |
/* * 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
|
1754 1755 |
* 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
|
1756 |
*/ |
90126375d
|
1757 |
static void get_scan_count(struct lruvec *lruvec, struct scan_control *sc, |
9e3b2f8cd
|
1758 |
unsigned long *nr) |
4f98a2fee
|
1759 |
{ |
9a2651140
|
1760 1761 1762 1763 |
struct zone_reclaim_stat *reclaim_stat = &lruvec->reclaim_stat; u64 fraction[2]; u64 denominator = 0; /* gcc */ struct zone *zone = lruvec_zone(lruvec); |
4f98a2fee
|
1764 |
unsigned long anon_prio, file_prio; |
9a2651140
|
1765 1766 1767 |
enum scan_balance scan_balance; unsigned long anon, file, free; bool force_scan = false; |
4f98a2fee
|
1768 |
unsigned long ap, fp; |
4111304da
|
1769 |
enum lru_list lru; |
246e87a93
|
1770 |
|
f11c0ca50
|
1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 |
/* * 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
|
1781 |
if (current_is_kswapd() && !zone_reclaimable(zone)) |
a4d3e9e76
|
1782 |
force_scan = true; |
89b5fae53
|
1783 |
if (!global_reclaim(sc)) |
a4d3e9e76
|
1784 |
force_scan = true; |
76a33fc38
|
1785 1786 |
/* If we have no swap space, do not bother scanning anon pages. */ |
ec8acf20a
|
1787 |
if (!sc->may_swap || (get_nr_swap_pages() <= 0)) { |
9a2651140
|
1788 |
scan_balance = SCAN_FILE; |
76a33fc38
|
1789 1790 |
goto out; } |
4f98a2fee
|
1791 |
|
10316b313
|
1792 1793 1794 1795 1796 1797 1798 1799 |
/* * 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. */ if (!global_reclaim(sc) && !vmscan_swappiness(sc)) { |
9a2651140
|
1800 |
scan_balance = SCAN_FILE; |
10316b313
|
1801 1802 1803 1804 1805 1806 1807 1808 1809 |
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). */ if (!sc->priority && vmscan_swappiness(sc)) { |
9a2651140
|
1810 |
scan_balance = SCAN_EQUAL; |
10316b313
|
1811 1812 |
goto out; } |
4d7dcca21
|
1813 1814 1815 1816 |
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); |
a4d3e9e76
|
1817 |
|
11d16c25b
|
1818 1819 1820 1821 1822 1823 |
/* * If it's foreseeable that reclaiming the file cache won't be * enough to get the zone back into a desirable shape, we have * to swap. Better start now and leave the - probably heavily * thrashing - remaining file pages alone. */ |
89b5fae53
|
1824 |
if (global_reclaim(sc)) { |
11d16c25b
|
1825 |
free = zone_page_state(zone, NR_FREE_PAGES); |
90126375d
|
1826 |
if (unlikely(file + free <= high_wmark_pages(zone))) { |
9a2651140
|
1827 |
scan_balance = SCAN_ANON; |
76a33fc38
|
1828 |
goto out; |
eeee9a8cd
|
1829 |
} |
4f98a2fee
|
1830 1831 1832 |
} /* |
7c5bd705d
|
1833 1834 1835 1836 |
* There is enough inactive page cache, do not reclaim * anything from the anonymous working set right now. */ if (!inactive_file_is_low(lruvec)) { |
9a2651140
|
1837 |
scan_balance = SCAN_FILE; |
7c5bd705d
|
1838 1839 |
goto out; } |
9a2651140
|
1840 |
scan_balance = SCAN_FRACT; |
7c5bd705d
|
1841 |
/* |
58c37f6e0
|
1842 1843 1844 |
* With swappiness at 100, anonymous and file have the same priority. * This scanning priority is essentially the inverse of IO cost. */ |
3d58ab5c9
|
1845 |
anon_prio = vmscan_swappiness(sc); |
75b00af77
|
1846 |
file_prio = 200 - anon_prio; |
58c37f6e0
|
1847 1848 |
/* |
4f98a2fee
|
1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 |
* 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] */ |
90126375d
|
1859 |
spin_lock_irq(&zone->lru_lock); |
6e9015716
|
1860 |
if (unlikely(reclaim_stat->recent_scanned[0] > anon / 4)) { |
6e9015716
|
1861 1862 |
reclaim_stat->recent_scanned[0] /= 2; reclaim_stat->recent_rotated[0] /= 2; |
4f98a2fee
|
1863 |
} |
6e9015716
|
1864 |
if (unlikely(reclaim_stat->recent_scanned[1] > file / 4)) { |
6e9015716
|
1865 1866 |
reclaim_stat->recent_scanned[1] /= 2; reclaim_stat->recent_rotated[1] /= 2; |
4f98a2fee
|
1867 1868 1869 |
} /* |
00d8089c5
|
1870 1871 1872 |
* 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
|
1873 |
*/ |
fe35004fb
|
1874 |
ap = anon_prio * (reclaim_stat->recent_scanned[0] + 1); |
6e9015716
|
1875 |
ap /= reclaim_stat->recent_rotated[0] + 1; |
4f98a2fee
|
1876 |
|
fe35004fb
|
1877 |
fp = file_prio * (reclaim_stat->recent_scanned[1] + 1); |
6e9015716
|
1878 |
fp /= reclaim_stat->recent_rotated[1] + 1; |
90126375d
|
1879 |
spin_unlock_irq(&zone->lru_lock); |
4f98a2fee
|
1880 |
|
76a33fc38
|
1881 1882 1883 1884 |
fraction[0] = ap; fraction[1] = fp; denominator = ap + fp + 1; out: |
4111304da
|
1885 1886 |
for_each_evictable_lru(lru) { int file = is_file_lru(lru); |
d778df51c
|
1887 |
unsigned long size; |
76a33fc38
|
1888 |
unsigned long scan; |
6e08a369e
|
1889 |
|
d778df51c
|
1890 |
size = get_lru_size(lruvec, lru); |
10316b313
|
1891 |
scan = size >> sc->priority; |
9a2651140
|
1892 |
|
10316b313
|
1893 1894 |
if (!scan && force_scan) scan = min(size, SWAP_CLUSTER_MAX); |
9a2651140
|
1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 |
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(); } |
4111304da
|
1917 |
nr[lru] = scan; |
76a33fc38
|
1918 |
} |
6e08a369e
|
1919 |
} |
4f98a2fee
|
1920 |
|
9b4f98cda
|
1921 1922 1923 1924 1925 1926 |
/* * This is a basic per-zone page freer. Used by both kswapd and direct reclaim. */ static void shrink_lruvec(struct lruvec *lruvec, struct scan_control *sc) { unsigned long nr[NR_LRU_LISTS]; |
e82e0561d
|
1927 |
unsigned long targets[NR_LRU_LISTS]; |
9b4f98cda
|
1928 1929 1930 1931 1932 |
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; |
e82e0561d
|
1933 |
bool scan_adjusted = false; |
9b4f98cda
|
1934 1935 |
get_scan_count(lruvec, sc, nr); |
e82e0561d
|
1936 1937 |
/* Record the original scan target for proportional adjustments later */ memcpy(targets, nr, sizeof(nr)); |
9b4f98cda
|
1938 1939 1940 |
blk_start_plug(&plug); while (nr[LRU_INACTIVE_ANON] || nr[LRU_ACTIVE_FILE] || nr[LRU_INACTIVE_FILE]) { |
e82e0561d
|
1941 1942 |
unsigned long nr_anon, nr_file, percentage; unsigned long nr_scanned; |
9b4f98cda
|
1943 1944 1945 1946 1947 1948 1949 1950 1951 |
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
|
1952 1953 1954 |
if (nr_reclaimed < nr_to_reclaim || scan_adjusted) continue; |
9b4f98cda
|
1955 |
/* |
e82e0561d
|
1956 1957 1958 1959 |
* For global direct reclaim, reclaim only the number of pages * requested. Less care is taken to scan proportionally as it * is more important to minimise direct reclaim stall latency * than it is to properly age the LRU lists. |
9b4f98cda
|
1960 |
*/ |
e82e0561d
|
1961 |
if (global_reclaim(sc) && !current_is_kswapd()) |
9b4f98cda
|
1962 |
break; |
e82e0561d
|
1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 |
/* * For kswapd and memcg, reclaim at least the number of pages * requested. Ensure that the anon and file LRUs shrink * 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]; 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
|
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 |
} 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
|
2019 |
/* Use reclaim/compaction for costly allocs or under memory pressure */ |
9e3b2f8cd
|
2020 |
static bool in_reclaim_compaction(struct scan_control *sc) |
23b9da55c
|
2021 |
{ |
d84da3f9e
|
2022 |
if (IS_ENABLED(CONFIG_COMPACTION) && sc->order && |
23b9da55c
|
2023 |
(sc->order > PAGE_ALLOC_COSTLY_ORDER || |
9e3b2f8cd
|
2024 |
sc->priority < DEF_PRIORITY - 2)) |
23b9da55c
|
2025 2026 2027 2028 |
return true; return false; } |
1da177e4c
|
2029 |
/* |
23b9da55c
|
2030 2031 2032 2033 2034 |
* 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
|
2035 |
*/ |
9b4f98cda
|
2036 |
static inline bool should_continue_reclaim(struct zone *zone, |
3e7d34497
|
2037 2038 2039 2040 2041 2042 2043 2044 |
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
|
2045 |
if (!in_reclaim_compaction(sc)) |
3e7d34497
|
2046 |
return false; |
2876592f2
|
2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 |
/* 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
|
2069 2070 2071 2072 2073 2074 |
/* * 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
|
2075 |
inactive_lru_pages = zone_page_state(zone, NR_INACTIVE_FILE); |
ec8acf20a
|
2076 |
if (get_nr_swap_pages() > 0) |
9b4f98cda
|
2077 |
inactive_lru_pages += zone_page_state(zone, NR_INACTIVE_ANON); |
3e7d34497
|
2078 2079 2080 2081 2082 |
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
|
2083 |
switch (compaction_suitable(zone, sc->order)) { |
3e7d34497
|
2084 2085 2086 2087 2088 2089 2090 |
case COMPACT_PARTIAL: case COMPACT_CONTINUE: return false; default: return true; } } |
0608f43da
|
2091 |
static void shrink_zone(struct zone *zone, struct scan_control *sc) |
1da177e4c
|
2092 |
{ |
f0fdc5e8e
|
2093 |
unsigned long nr_reclaimed, nr_scanned; |
1da177e4c
|
2094 |
|
9b4f98cda
|
2095 2096 2097 2098 2099 2100 |
do { struct mem_cgroup *root = sc->target_mem_cgroup; struct mem_cgroup_reclaim_cookie reclaim = { .zone = zone, .priority = sc->priority, }; |
694fbc0fe
|
2101 |
struct mem_cgroup *memcg; |
3e7d34497
|
2102 |
|
9b4f98cda
|
2103 2104 |
nr_reclaimed = sc->nr_reclaimed; nr_scanned = sc->nr_scanned; |
1da177e4c
|
2105 |
|
694fbc0fe
|
2106 2107 |
memcg = mem_cgroup_iter(root, NULL, &reclaim); do { |
9b4f98cda
|
2108 |
struct lruvec *lruvec; |
5660048cc
|
2109 |
|
9b4f98cda
|
2110 |
lruvec = mem_cgroup_zone_lruvec(zone, memcg); |
f9be23d6d
|
2111 |
|
9b4f98cda
|
2112 |
shrink_lruvec(lruvec, sc); |
f16015fbf
|
2113 |
|
9b4f98cda
|
2114 |
/* |
a394cb8ee
|
2115 2116 |
* Direct reclaim and kswapd have to scan all memory * cgroups to fulfill the overall scan target for the |
9b4f98cda
|
2117 |
* zone. |
a394cb8ee
|
2118 2119 2120 2121 2122 |
* * 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
|
2123 |
*/ |
a394cb8ee
|
2124 2125 |
if (!global_reclaim(sc) && sc->nr_reclaimed >= sc->nr_to_reclaim) { |
9b4f98cda
|
2126 2127 2128 |
mem_cgroup_iter_break(root, memcg); break; } |
694fbc0fe
|
2129 2130 |
memcg = mem_cgroup_iter(root, memcg, &reclaim); } while (memcg); |
70ddf637e
|
2131 2132 2133 2134 |
vmpressure(sc->gfp_mask, sc->target_mem_cgroup, sc->nr_scanned - nr_scanned, sc->nr_reclaimed - nr_reclaimed); |
9b4f98cda
|
2135 2136 |
} while (should_continue_reclaim(zone, sc->nr_reclaimed - nr_reclaimed, sc->nr_scanned - nr_scanned, sc)); |
f16015fbf
|
2137 |
} |
fe4b1b244
|
2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 |
/* Returns true if compaction should go ahead for a high-order request */ static inline bool compaction_ready(struct zone *zone, struct scan_control *sc) { unsigned long balance_gap, watermark; bool watermark_ok; /* Do not consider compaction for orders reclaim is meant to satisfy */ if (sc->order <= PAGE_ALLOC_COSTLY_ORDER) return false; /* * 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 */ balance_gap = min(low_wmark_pages(zone), |
b40da0494
|
2155 |
(zone->managed_pages + KSWAPD_ZONE_BALANCE_GAP_RATIO-1) / |
fe4b1b244
|
2156 2157 2158 2159 2160 2161 2162 2163 |
KSWAPD_ZONE_BALANCE_GAP_RATIO); watermark = high_wmark_pages(zone) + balance_gap + (2UL << sc->order); 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 */ |
aff622495
|
2164 |
if (compaction_deferred(zone, sc->order)) |
fe4b1b244
|
2165 2166 2167 2168 2169 2170 2171 2172 |
return watermark_ok; /* If compaction is not ready to start, keep reclaiming */ if (!compaction_suitable(zone, sc->order)) return false; return watermark_ok; } |
1da177e4c
|
2173 2174 2175 2176 2177 |
/* * 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
|
2178 2179 |
* We reclaim from a zone even if that zone is over high_wmark_pages(zone). * Because: |
1da177e4c
|
2180 2181 |
* a) The caller may be trying to free *extra* pages to satisfy a higher-order * allocation or |
418589663
|
2182 2183 2184 |
* 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
|
2185 |
* |
1da177e4c
|
2186 2187 |
* If a zone is deemed to be full of pinned pages then just give it a light * scan then give up on it. |
e0c23279c
|
2188 2189 |
* * This function returns true if a zone is being reclaimed for a costly |
fe4b1b244
|
2190 |
* high-order allocation and compaction is ready to begin. This indicates to |
0cee34fd7
|
2191 2192 |
* the caller that it should consider retrying the allocation instead of * further reclaim. |
1da177e4c
|
2193 |
*/ |
9e3b2f8cd
|
2194 |
static bool shrink_zones(struct zonelist *zonelist, struct scan_control *sc) |
1da177e4c
|
2195 |
{ |
dd1a239f6
|
2196 |
struct zoneref *z; |
54a6eb5c4
|
2197 |
struct zone *zone; |
0608f43da
|
2198 2199 |
unsigned long nr_soft_reclaimed; unsigned long nr_soft_scanned; |
0cee34fd7
|
2200 |
bool aborted_reclaim = false; |
1cfb419b3
|
2201 |
|
cc715d99e
|
2202 2203 2204 2205 2206 2207 2208 |
/* * 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 */ if (buffer_heads_over_limit) sc->gfp_mask |= __GFP_HIGHMEM; |
d4debc66d
|
2209 2210 |
for_each_zone_zonelist_nodemask(zone, z, zonelist, gfp_zone(sc->gfp_mask), sc->nodemask) { |
f3fe65122
|
2211 |
if (!populated_zone(zone)) |
1da177e4c
|
2212 |
continue; |
1cfb419b3
|
2213 2214 2215 2216 |
/* * Take care memory controller reclaiming has small influence * to global LRU. */ |
89b5fae53
|
2217 |
if (global_reclaim(sc)) { |
1cfb419b3
|
2218 2219 |
if (!cpuset_zone_allowed_hardwall(zone, GFP_KERNEL)) continue; |
6e543d578
|
2220 2221 |
if (sc->priority != DEF_PRIORITY && !zone_reclaimable(zone)) |
1cfb419b3
|
2222 |
continue; /* Let kswapd poll it */ |
d84da3f9e
|
2223 |
if (IS_ENABLED(CONFIG_COMPACTION)) { |
e0887c19b
|
2224 |
/* |
e0c23279c
|
2225 2226 2227 2228 2229 |
* 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 |
c7cfa37b7
|
2230 2231 |
* noticeable problem, like transparent huge * page allocations. |
e0887c19b
|
2232 |
*/ |
fe4b1b244
|
2233 |
if (compaction_ready(zone, sc)) { |
0cee34fd7
|
2234 |
aborted_reclaim = true; |
e0887c19b
|
2235 |
continue; |
e0c23279c
|
2236 |
} |
e0887c19b
|
2237 |
} |
0608f43da
|
2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 |
/* * 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; |
ac34a1a3c
|
2250 |
/* need some check for avoid more shrink_zone() */ |
1cfb419b3
|
2251 |
} |
408d85441
|
2252 |
|
9e3b2f8cd
|
2253 |
shrink_zone(zone, sc); |
1da177e4c
|
2254 |
} |
e0c23279c
|
2255 |
|
0cee34fd7
|
2256 |
return aborted_reclaim; |
d1908362a
|
2257 |
} |
929bea7c7
|
2258 |
/* All zones in zonelist are unreclaimable? */ |
d1908362a
|
2259 2260 2261 2262 2263 |
static bool all_unreclaimable(struct zonelist *zonelist, struct scan_control *sc) { struct zoneref *z; struct zone *zone; |
d1908362a
|
2264 2265 2266 2267 2268 2269 2270 |
for_each_zone_zonelist_nodemask(zone, z, zonelist, gfp_zone(sc->gfp_mask), sc->nodemask) { if (!populated_zone(zone)) continue; if (!cpuset_zone_allowed_hardwall(zone, GFP_KERNEL)) continue; |
6e543d578
|
2271 |
if (zone_reclaimable(zone)) |
929bea7c7
|
2272 |
return false; |
d1908362a
|
2273 |
} |
929bea7c7
|
2274 |
return true; |
1da177e4c
|
2275 |
} |
4f98a2fee
|
2276 |
|
1da177e4c
|
2277 2278 2279 2280 2281 2282 2283 2284 |
/* * 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
|
2285 2286 2287 2288 |
* 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
|
2289 2290 2291 |
* * returns: 0, if no pages reclaimed * else, the number of pages reclaimed |
1da177e4c
|
2292 |
*/ |
dac1d27bc
|
2293 |
static unsigned long do_try_to_free_pages(struct zonelist *zonelist, |
a09ed5e00
|
2294 2295 |
struct scan_control *sc, struct shrink_control *shrink) |
1da177e4c
|
2296 |
{ |
69e05944a
|
2297 |
unsigned long total_scanned = 0; |
1da177e4c
|
2298 |
struct reclaim_state *reclaim_state = current->reclaim_state; |
dd1a239f6
|
2299 |
struct zoneref *z; |
54a6eb5c4
|
2300 |
struct zone *zone; |
22fba3354
|
2301 |
unsigned long writeback_threshold; |
0cee34fd7
|
2302 |
bool aborted_reclaim; |
1da177e4c
|
2303 |
|
873b47717
|
2304 |
delayacct_freepages_start(); |
89b5fae53
|
2305 |
if (global_reclaim(sc)) |
1cfb419b3
|
2306 |
count_vm_event(ALLOCSTALL); |
1da177e4c
|
2307 |
|
9e3b2f8cd
|
2308 |
do { |
70ddf637e
|
2309 2310 |
vmpressure_prio(sc->gfp_mask, sc->target_mem_cgroup, sc->priority); |
66e1707bc
|
2311 |
sc->nr_scanned = 0; |
9e3b2f8cd
|
2312 |
aborted_reclaim = shrink_zones(zonelist, sc); |
e0c23279c
|
2313 |
|
66e1707bc
|
2314 |
/* |
5a1c9cbc1
|
2315 2316 2317 2318 |
* 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. |
66e1707bc
|
2319 |
*/ |
89b5fae53
|
2320 |
if (global_reclaim(sc)) { |
c6a8a8c58
|
2321 |
unsigned long lru_pages = 0; |
0ce3d7445
|
2322 2323 |
nodes_clear(shrink->nodes_to_scan); |
d4debc66d
|
2324 2325 |
for_each_zone_zonelist(zone, z, zonelist, gfp_zone(sc->gfp_mask)) { |
c6a8a8c58
|
2326 2327 2328 2329 |
if (!cpuset_zone_allowed_hardwall(zone, GFP_KERNEL)) continue; lru_pages += zone_reclaimable_pages(zone); |
0ce3d7445
|
2330 2331 |
node_set(zone_to_nid(zone), shrink->nodes_to_scan); |
c6a8a8c58
|
2332 |
} |
1495f230f
|
2333 |
shrink_slab(shrink, sc->nr_scanned, lru_pages); |
91a45470f
|
2334 |
if (reclaim_state) { |
a79311c14
|
2335 |
sc->nr_reclaimed += reclaim_state->reclaimed_slab; |
91a45470f
|
2336 2337 |
reclaim_state->reclaimed_slab = 0; } |
1da177e4c
|
2338 |
} |
66e1707bc
|
2339 |
total_scanned += sc->nr_scanned; |
bb21c7ce1
|
2340 |
if (sc->nr_reclaimed >= sc->nr_to_reclaim) |
1da177e4c
|
2341 |
goto out; |
1da177e4c
|
2342 2343 |
/* |
0e50ce3b5
|
2344 2345 2346 2347 2348 2349 2350 |
* If we're getting trouble reclaiming, start doing * writepage even in laptop mode. */ if (sc->priority < DEF_PRIORITY - 2) sc->may_writepage = 1; /* |
1da177e4c
|
2351 2352 2353 2354 2355 2356 |
* 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
|
2357 2358 |
writeback_threshold = sc->nr_to_reclaim + sc->nr_to_reclaim / 2; if (total_scanned > writeback_threshold) { |
0e175a183
|
2359 2360 |
wakeup_flusher_threads(laptop_mode ? 0 : total_scanned, WB_REASON_TRY_TO_FREE_PAGES); |
66e1707bc
|
2361 |
sc->may_writepage = 1; |
1da177e4c
|
2362 |
} |
5a1c9cbc1
|
2363 |
} while (--sc->priority >= 0 && !aborted_reclaim); |
bb21c7ce1
|
2364 |
|
1da177e4c
|
2365 |
out: |
873b47717
|
2366 |
delayacct_freepages_end(); |
bb21c7ce1
|
2367 2368 |
if (sc->nr_reclaimed) return sc->nr_reclaimed; |
929bea7c7
|
2369 2370 2371 2372 2373 2374 2375 |
/* * As hibernation is going on, kswapd is freezed so that it can't mark * the zone into all_unreclaimable. Thus bypassing all_unreclaimable * check. */ if (oom_killer_disabled) return 0; |
0cee34fd7
|
2376 2377 |
/* Aborted reclaim to try compaction? don't OOM, then */ if (aborted_reclaim) |
7335084d4
|
2378 |
return 1; |
bb21c7ce1
|
2379 |
/* top priority shrink_zones still had more to do? don't OOM, then */ |
89b5fae53
|
2380 |
if (global_reclaim(sc) && !all_unreclaimable(zonelist, sc)) |
bb21c7ce1
|
2381 2382 2383 |
return 1; return 0; |
1da177e4c
|
2384 |
} |
5515061d2
|
2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 2400 2401 2402 2403 2404 2405 2406 2407 2408 2409 2410 2411 2412 2413 2414 |
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]; pfmemalloc_reserve += min_wmark_pages(zone); free_pages += zone_page_state(zone, NR_FREE_PAGES); } 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
|
2415 2416 2417 2418 |
* 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
|
2419 |
*/ |
50694c28f
|
2420 |
static bool throttle_direct_reclaim(gfp_t gfp_mask, struct zonelist *zonelist, |
5515061d2
|
2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 |
nodemask_t *nodemask) { struct zone *zone; int high_zoneidx = gfp_zone(gfp_mask); pg_data_t *pgdat; /* * 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
|
2435 2436 2437 2438 2439 2440 2441 2442 |
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
|
2443 2444 2445 2446 2447 |
/* Check if the pfmemalloc reserves are ok */ first_zones_zonelist(zonelist, high_zoneidx, NULL, &zone); pgdat = zone->zone_pgdat; if (pfmemalloc_watermark_ok(pgdat)) |
50694c28f
|
2448 |
goto out; |
5515061d2
|
2449 |
|
68243e76e
|
2450 2451 |
/* Account for the throttling */ count_vm_event(PGSCAN_DIRECT_THROTTLE); |
5515061d2
|
2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 |
/* * 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
|
2463 2464 |
goto check_pending; |
5515061d2
|
2465 2466 2467 2468 2469 |
} /* Throttle until kswapd wakes the process */ wait_event_killable(zone->zone_pgdat->pfmemalloc_wait, pfmemalloc_watermark_ok(pgdat)); |
50694c28f
|
2470 2471 2472 2473 2474 2475 2476 |
check_pending: if (fatal_signal_pending(current)) return true; out: return false; |
5515061d2
|
2477 |
} |
dac1d27bc
|
2478 |
unsigned long try_to_free_pages(struct zonelist *zonelist, int order, |
327c0e968
|
2479 |
gfp_t gfp_mask, nodemask_t *nodemask) |
66e1707bc
|
2480 |
{ |
33906bc5c
|
2481 |
unsigned long nr_reclaimed; |
66e1707bc
|
2482 |
struct scan_control sc = { |
21caf2fc1
|
2483 |
.gfp_mask = (gfp_mask = memalloc_noio_flags(gfp_mask)), |
66e1707bc
|
2484 |
.may_writepage = !laptop_mode, |
22fba3354
|
2485 |
.nr_to_reclaim = SWAP_CLUSTER_MAX, |
a6dc60f89
|
2486 |
.may_unmap = 1, |
2e2e42598
|
2487 |
.may_swap = 1, |
66e1707bc
|
2488 |
.order = order, |
9e3b2f8cd
|
2489 |
.priority = DEF_PRIORITY, |
f16015fbf
|
2490 |
.target_mem_cgroup = NULL, |
327c0e968
|
2491 |
.nodemask = nodemask, |
66e1707bc
|
2492 |
}; |
a09ed5e00
|
2493 2494 2495 |
struct shrink_control shrink = { .gfp_mask = sc.gfp_mask, }; |
66e1707bc
|
2496 |
|
5515061d2
|
2497 |
/* |
50694c28f
|
2498 2499 2500 |
* 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
|
2501 |
*/ |
50694c28f
|
2502 |
if (throttle_direct_reclaim(gfp_mask, zonelist, nodemask)) |
5515061d2
|
2503 |
return 1; |
33906bc5c
|
2504 2505 2506 |
trace_mm_vmscan_direct_reclaim_begin(order, sc.may_writepage, gfp_mask); |
a09ed5e00
|
2507 |
nr_reclaimed = do_try_to_free_pages(zonelist, &sc, &shrink); |
33906bc5c
|
2508 2509 2510 2511 |
trace_mm_vmscan_direct_reclaim_end(nr_reclaimed); return nr_reclaimed; |
66e1707bc
|
2512 |
} |
c255a4580
|
2513 |
#ifdef CONFIG_MEMCG |
66e1707bc
|
2514 |
|
72835c86c
|
2515 |
unsigned long mem_cgroup_shrink_node_zone(struct mem_cgroup *memcg, |
4e4169535
|
2516 |
gfp_t gfp_mask, bool noswap, |
0ae5e89c6
|
2517 2518 |
struct zone *zone, unsigned long *nr_scanned) |
4e4169535
|
2519 2520 |
{ struct scan_control sc = { |
0ae5e89c6
|
2521 |
.nr_scanned = 0, |
b8f5c5664
|
2522 |
.nr_to_reclaim = SWAP_CLUSTER_MAX, |
4e4169535
|
2523 2524 2525 |
.may_writepage = !laptop_mode, .may_unmap = 1, .may_swap = !noswap, |
4e4169535
|
2526 |
.order = 0, |
9e3b2f8cd
|
2527 |
.priority = 0, |
72835c86c
|
2528 |
.target_mem_cgroup = memcg, |
4e4169535
|
2529 |
}; |
f9be23d6d
|
2530 |
struct lruvec *lruvec = mem_cgroup_zone_lruvec(zone, memcg); |
0ae5e89c6
|
2531 |
|
4e4169535
|
2532 2533 |
sc.gfp_mask = (gfp_mask & GFP_RECLAIM_MASK) | (GFP_HIGHUSER_MOVABLE & ~GFP_RECLAIM_MASK); |
bdce6d9eb
|
2534 |
|
9e3b2f8cd
|
2535 |
trace_mm_vmscan_memcg_softlimit_reclaim_begin(sc.order, |
bdce6d9eb
|
2536 2537 |
sc.may_writepage, sc.gfp_mask); |
4e4169535
|
2538 2539 2540 2541 2542 2543 2544 |
/* * 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. */ |
f9be23d6d
|
2545 |
shrink_lruvec(lruvec, &sc); |
bdce6d9eb
|
2546 2547 |
trace_mm_vmscan_memcg_softlimit_reclaim_end(sc.nr_reclaimed); |
0ae5e89c6
|
2548 |
*nr_scanned = sc.nr_scanned; |
4e4169535
|
2549 2550 |
return sc.nr_reclaimed; } |
72835c86c
|
2551 |
unsigned long try_to_free_mem_cgroup_pages(struct mem_cgroup *memcg, |
a7885eb8a
|
2552 |
gfp_t gfp_mask, |
185efc0f9
|
2553 |
bool noswap) |
66e1707bc
|
2554 |
{ |
4e4169535
|
2555 |
struct zonelist *zonelist; |
bdce6d9eb
|
2556 |
unsigned long nr_reclaimed; |
889976dbc
|
2557 |
int nid; |
66e1707bc
|
2558 |
struct scan_control sc = { |
66e1707bc
|
2559 |
.may_writepage = !laptop_mode, |
a6dc60f89
|
2560 |
.may_unmap = 1, |
2e2e42598
|
2561 |
.may_swap = !noswap, |
22fba3354
|
2562 |
.nr_to_reclaim = SWAP_CLUSTER_MAX, |
66e1707bc
|
2563 |
.order = 0, |
9e3b2f8cd
|
2564 |
.priority = DEF_PRIORITY, |
72835c86c
|
2565 |
.target_mem_cgroup = memcg, |
327c0e968
|
2566 |
.nodemask = NULL, /* we don't care the placement */ |
a09ed5e00
|
2567 2568 2569 2570 2571 |
.gfp_mask = (gfp_mask & GFP_RECLAIM_MASK) | (GFP_HIGHUSER_MOVABLE & ~GFP_RECLAIM_MASK), }; struct shrink_control shrink = { .gfp_mask = sc.gfp_mask, |
66e1707bc
|
2572 |
}; |
66e1707bc
|
2573 |
|
889976dbc
|
2574 2575 2576 2577 2578 |
/* * 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
|
2579 |
nid = mem_cgroup_select_victim_node(memcg); |
889976dbc
|
2580 2581 |
zonelist = NODE_DATA(nid)->node_zonelists; |
bdce6d9eb
|
2582 2583 2584 2585 |
trace_mm_vmscan_memcg_reclaim_begin(0, sc.may_writepage, sc.gfp_mask); |
a09ed5e00
|
2586 |
nr_reclaimed = do_try_to_free_pages(zonelist, &sc, &shrink); |
bdce6d9eb
|
2587 2588 2589 2590 |
trace_mm_vmscan_memcg_reclaim_end(nr_reclaimed); return nr_reclaimed; |
66e1707bc
|
2591 2592 |
} #endif |
9e3b2f8cd
|
2593 |
static void age_active_anon(struct zone *zone, struct scan_control *sc) |
f16015fbf
|
2594 |
{ |
b95a2f2d4
|
2595 |
struct mem_cgroup *memcg; |
f16015fbf
|
2596 |
|
b95a2f2d4
|
2597 2598 2599 2600 2601 |
if (!total_swap_pages) return; memcg = mem_cgroup_iter(NULL, NULL, NULL); do { |
c56d5c7df
|
2602 |
struct lruvec *lruvec = mem_cgroup_zone_lruvec(zone, memcg); |
b95a2f2d4
|
2603 |
|
c56d5c7df
|
2604 |
if (inactive_anon_is_low(lruvec)) |
1a93be0e7
|
2605 |
shrink_active_list(SWAP_CLUSTER_MAX, lruvec, |
9e3b2f8cd
|
2606 |
sc, LRU_ACTIVE_ANON); |
b95a2f2d4
|
2607 2608 2609 |
memcg = mem_cgroup_iter(NULL, memcg, NULL); } while (memcg); |
f16015fbf
|
2610 |
} |
60cefed48
|
2611 2612 2613 2614 2615 2616 |
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
|
2617 2618 |
if (IS_ENABLED(CONFIG_COMPACTION) && order && !compaction_suitable(zone, order)) |
60cefed48
|
2619 2620 2621 2622 |
return false; return true; } |
1741c8775
|
2623 |
/* |
4ae0a48b5
|
2624 2625 2626 2627 2628 2629 2630 2631 2632 2633 |
* 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
|
2634 2635 2636 2637 |
* 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
|
2638 |
* percentage of the middle zones. For example, on 32-bit x86, highmem |
1741c8775
|
2639 2640 2641 2642 |
* 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
|
2643 |
static bool pgdat_balanced(pg_data_t *pgdat, int order, int classzone_idx) |
1741c8775
|
2644 |
{ |
b40da0494
|
2645 |
unsigned long managed_pages = 0; |
4ae0a48b5
|
2646 |
unsigned long balanced_pages = 0; |
1741c8775
|
2647 |
int i; |
4ae0a48b5
|
2648 2649 2650 |
/* Check the watermark levels */ for (i = 0; i <= classzone_idx; i++) { struct zone *zone = pgdat->node_zones + i; |
1741c8775
|
2651 |
|
4ae0a48b5
|
2652 2653 |
if (!populated_zone(zone)) continue; |
b40da0494
|
2654 |
managed_pages += zone->managed_pages; |
4ae0a48b5
|
2655 2656 2657 2658 2659 2660 2661 2662 |
/* * 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
|
2663 |
if (!zone_reclaimable(zone)) { |
b40da0494
|
2664 |
balanced_pages += zone->managed_pages; |
4ae0a48b5
|
2665 2666 2667 2668 |
continue; } if (zone_balanced(zone, order, 0, i)) |
b40da0494
|
2669 |
balanced_pages += zone->managed_pages; |
4ae0a48b5
|
2670 2671 2672 2673 2674 |
else if (!order) return false; } if (order) |
b40da0494
|
2675 |
return balanced_pages >= (managed_pages >> 2); |
4ae0a48b5
|
2676 2677 |
else return true; |
1741c8775
|
2678 |
} |
5515061d2
|
2679 2680 2681 2682 2683 2684 2685 |
/* * 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
|
2686 |
int classzone_idx) |
f50de2d38
|
2687 |
{ |
f50de2d38
|
2688 2689 |
/* If a direct reclaimer woke kswapd within HZ/10, it's premature */ if (remaining) |
5515061d2
|
2690 2691 2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704 |
return false; /* * There is a potential race between when kswapd checks its watermarks * and a process gets throttled. There is also a potential race if * processes get throttled, kswapd wakes, a large process exits therby * balancing the zones that causes kswapd to miss a wakeup. If kswapd * is going to sleep, no process should be sleeping on pfmemalloc_wait * so wake them now if necessary. If necessary, processes will wake * kswapd and get throttled again */ if (waitqueue_active(&pgdat->pfmemalloc_wait)) { wake_up(&pgdat->pfmemalloc_wait); return false; } |
f50de2d38
|
2705 |
|
4ae0a48b5
|
2706 |
return pgdat_balanced(pgdat, order, classzone_idx); |
f50de2d38
|
2707 |
} |
1da177e4c
|
2708 |
/* |
75485363c
|
2709 2710 |
* kswapd shrinks the zone by the number of pages required to reach * the high watermark. |
b8e83b942
|
2711 2712 |
* * Returns true if kswapd scanned at least the requested number of pages to |
283aba9f9
|
2713 2714 |
* 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
|
2715 |
*/ |
b8e83b942
|
2716 |
static bool kswapd_shrink_zone(struct zone *zone, |
7c954f6de
|
2717 |
int classzone_idx, |
75485363c
|
2718 |
struct scan_control *sc, |
2ab44f434
|
2719 2720 |
unsigned long lru_pages, unsigned long *nr_attempted) |
75485363c
|
2721 |
{ |
7c954f6de
|
2722 2723 |
int testorder = sc->order; unsigned long balance_gap; |
75485363c
|
2724 2725 2726 2727 |
struct reclaim_state *reclaim_state = current->reclaim_state; struct shrink_control shrink = { .gfp_mask = sc->gfp_mask, }; |
7c954f6de
|
2728 |
bool lowmem_pressure; |
75485363c
|
2729 2730 2731 |
/* Reclaim above the high watermark. */ sc->nr_to_reclaim = max(SWAP_CLUSTER_MAX, high_wmark_pages(zone)); |
7c954f6de
|
2732 2733 2734 2735 2736 2737 2738 2739 2740 2741 2742 2743 2744 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 |
/* * 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. */ balance_gap = min(low_wmark_pages(zone), (zone->managed_pages + KSWAPD_ZONE_BALANCE_GAP_RATIO-1) / KSWAPD_ZONE_BALANCE_GAP_RATIO); /* * 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
|
2762 |
shrink_zone(zone, sc); |
0ce3d7445
|
2763 2764 |
nodes_clear(shrink.nodes_to_scan); node_set(zone_to_nid(zone), shrink.nodes_to_scan); |
75485363c
|
2765 2766 |
reclaim_state->reclaimed_slab = 0; |
6e543d578
|
2767 |
shrink_slab(&shrink, sc->nr_scanned, lru_pages); |
75485363c
|
2768 |
sc->nr_reclaimed += reclaim_state->reclaimed_slab; |
2ab44f434
|
2769 2770 |
/* Account for the number of pages attempted to reclaim */ *nr_attempted += sc->nr_to_reclaim; |
283aba9f9
|
2771 |
zone_clear_flag(zone, ZONE_WRITEBACK); |
7c954f6de
|
2772 2773 2774 2775 2776 2777 |
/* * 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
|
2778 |
if (zone_reclaimable(zone) && |
7c954f6de
|
2779 2780 2781 2782 |
zone_balanced(zone, testorder, 0, classzone_idx)) { zone_clear_flag(zone, ZONE_CONGESTED); zone_clear_flag(zone, ZONE_TAIL_LRU_DIRTY); } |
b8e83b942
|
2783 |
return sc->nr_scanned >= sc->nr_to_reclaim; |
75485363c
|
2784 2785 2786 |
} /* |
1da177e4c
|
2787 |
* For kswapd, balance_pgdat() will work across all this node's zones until |
418589663
|
2788 |
* they are all at high_wmark_pages(zone). |
1da177e4c
|
2789 |
* |
0abdee2bd
|
2790 |
* Returns the final order kswapd was reclaiming at |
1da177e4c
|
2791 2792 2793 2794 2795 2796 2797 2798 2799 2800 |
* * 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
|
2801 2802 2803 2804 2805 |
* 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
|
2806 |
*/ |
995047488
|
2807 |
static unsigned long balance_pgdat(pg_data_t *pgdat, int order, |
dc83edd94
|
2808 |
int *classzone_idx) |
1da177e4c
|
2809 |
{ |
1da177e4c
|
2810 |
int i; |
995047488
|
2811 |
int end_zone = 0; /* Inclusive. 0 = ZONE_DMA */ |
0608f43da
|
2812 2813 |
unsigned long nr_soft_reclaimed; unsigned long nr_soft_scanned; |
179e96395
|
2814 2815 |
struct scan_control sc = { .gfp_mask = GFP_KERNEL, |
b8e83b942
|
2816 |
.priority = DEF_PRIORITY, |
a6dc60f89
|
2817 |
.may_unmap = 1, |
2e2e42598
|
2818 |
.may_swap = 1, |
b8e83b942
|
2819 |
.may_writepage = !laptop_mode, |
5ad333eb6
|
2820 |
.order = order, |
f16015fbf
|
2821 |
.target_mem_cgroup = NULL, |
179e96395
|
2822 |
}; |
f8891e5e1
|
2823 |
count_vm_event(PAGEOUTRUN); |
1da177e4c
|
2824 |
|
9e3b2f8cd
|
2825 |
do { |
1da177e4c
|
2826 |
unsigned long lru_pages = 0; |
2ab44f434
|
2827 |
unsigned long nr_attempted = 0; |
b8e83b942
|
2828 |
bool raise_priority = true; |
2ab44f434
|
2829 |
bool pgdat_needs_compaction = (order > 0); |
b8e83b942
|
2830 2831 |
sc.nr_reclaimed = 0; |
1da177e4c
|
2832 |
|
d6277db4a
|
2833 2834 2835 2836 2837 2838 |
/* * 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
|
2839 |
|
d6277db4a
|
2840 2841 |
if (!populated_zone(zone)) continue; |
1da177e4c
|
2842 |
|
6e543d578
|
2843 2844 |
if (sc.priority != DEF_PRIORITY && !zone_reclaimable(zone)) |
d6277db4a
|
2845 |
continue; |
1da177e4c
|
2846 |
|
556adecba
|
2847 2848 2849 2850 |
/* * Do some background aging of the anon list, to give * pages a chance to be referenced before reclaiming. */ |
9e3b2f8cd
|
2851 |
age_active_anon(zone, &sc); |
556adecba
|
2852 |
|
cc715d99e
|
2853 2854 2855 2856 2857 2858 2859 2860 2861 2862 |
/* * 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
|
2863 |
if (!zone_balanced(zone, order, 0, 0)) { |
d6277db4a
|
2864 |
end_zone = i; |
e1dbeda60
|
2865 |
break; |
439423f68
|
2866 |
} else { |
d43006d50
|
2867 2868 2869 2870 |
/* * If balanced, clear the dirty and congested * flags */ |
439423f68
|
2871 |
zone_clear_flag(zone, ZONE_CONGESTED); |
d43006d50
|
2872 |
zone_clear_flag(zone, ZONE_TAIL_LRU_DIRTY); |
1da177e4c
|
2873 |
} |
1da177e4c
|
2874 |
} |
dafcb73e3
|
2875 |
|
b8e83b942
|
2876 |
if (i < 0) |
e1dbeda60
|
2877 |
goto out; |
1da177e4c
|
2878 2879 |
for (i = 0; i <= end_zone; i++) { struct zone *zone = pgdat->node_zones + i; |
2ab44f434
|
2880 2881 |
if (!populated_zone(zone)) continue; |
adea02a1b
|
2882 |
lru_pages += zone_reclaimable_pages(zone); |
2ab44f434
|
2883 2884 2885 2886 2887 2888 2889 2890 2891 2892 2893 |
/* * 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
|
2894 2895 2896 |
} /* |
b7ea3c417
|
2897 2898 2899 2900 2901 2902 2903 |
* If we're getting trouble reclaiming, start doing writepage * even in laptop mode. */ if (sc.priority < DEF_PRIORITY - 2) sc.may_writepage = 1; /* |
1da177e4c
|
2904 2905 2906 2907 2908 2909 2910 2911 2912 2913 |
* 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
|
2914 |
if (!populated_zone(zone)) |
1da177e4c
|
2915 |
continue; |
6e543d578
|
2916 2917 |
if (sc.priority != DEF_PRIORITY && !zone_reclaimable(zone)) |
1da177e4c
|
2918 |
continue; |
1da177e4c
|
2919 |
sc.nr_scanned = 0; |
4e4169535
|
2920 |
|
0608f43da
|
2921 2922 2923 2924 2925 2926 2927 2928 |
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
|
2929 |
/* |
7c954f6de
|
2930 2931 2932 2933 |
* 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
|
2934 |
*/ |
7c954f6de
|
2935 2936 2937 |
if (kswapd_shrink_zone(zone, end_zone, &sc, lru_pages, &nr_attempted)) raise_priority = false; |
1da177e4c
|
2938 |
} |
5515061d2
|
2939 2940 2941 2942 2943 2944 2945 2946 2947 |
/* * 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
|
2948 |
/* |
b8e83b942
|
2949 2950 2951 2952 2953 2954 |
* 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
|
2955 |
*/ |
b8e83b942
|
2956 2957 |
if (order && sc.nr_reclaimed >= 2UL << order) order = sc.order = 0; |
8357376d3
|
2958 |
|
b8e83b942
|
2959 2960 2961 |
/* Check if kswapd should be suspending */ if (try_to_freeze() || kthread_should_stop()) break; |
8357376d3
|
2962 |
|
73ce02e96
|
2963 |
/* |
2ab44f434
|
2964 2965 2966 2967 2968 2969 2970 |
* 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
|
2971 2972 |
* Raise priority if scanning rate is too low or there was no * progress in reclaiming pages |
73ce02e96
|
2973 |
*/ |
b8e83b942
|
2974 2975 |
if (raise_priority || !sc.nr_reclaimed) sc.priority--; |
9aa41348a
|
2976 |
} while (sc.priority >= 1 && |
b8e83b942
|
2977 |
!pgdat_balanced(pgdat, order, *classzone_idx)); |
1da177e4c
|
2978 |
|
b8e83b942
|
2979 |
out: |
0abdee2bd
|
2980 |
/* |
5515061d2
|
2981 |
* Return the order we were reclaiming at so prepare_kswapd_sleep() |
0abdee2bd
|
2982 2983 2984 2985 |
* 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
|
2986 |
*classzone_idx = end_zone; |
0abdee2bd
|
2987 |
return order; |
1da177e4c
|
2988 |
} |
dc83edd94
|
2989 |
static void kswapd_try_to_sleep(pg_data_t *pgdat, int order, int classzone_idx) |
f0bc0a60b
|
2990 2991 2992 2993 2994 2995 2996 2997 2998 2999 |
{ 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
|
3000 |
if (prepare_kswapd_sleep(pgdat, order, remaining, classzone_idx)) { |
f0bc0a60b
|
3001 3002 3003 3004 3005 3006 3007 3008 3009 |
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
|
3010 |
if (prepare_kswapd_sleep(pgdat, order, remaining, classzone_idx)) { |
f0bc0a60b
|
3011 3012 3013 3014 3015 3016 3017 3018 3019 3020 3021 |
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
|
3022 |
|
62997027c
|
3023 3024 3025 3026 3027 3028 3029 |
/* * 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
|
3030 3031 |
if (!kthread_should_stop()) schedule(); |
f0bc0a60b
|
3032 3033 3034 3035 3036 3037 3038 3039 3040 |
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
|
3041 3042 |
/* * The background pageout daemon, started as a kernel thread |
4f98a2fee
|
3043 |
* from the init process. |
1da177e4c
|
3044 3045 3046 3047 3048 3049 3050 3051 3052 3053 3054 3055 |
* * 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
|
3056 |
unsigned long order, new_order; |
d2ebd0f6b
|
3057 |
unsigned balanced_order; |
215ddd666
|
3058 |
int classzone_idx, new_classzone_idx; |
d2ebd0f6b
|
3059 |
int balanced_classzone_idx; |
1da177e4c
|
3060 3061 |
pg_data_t *pgdat = (pg_data_t*)p; struct task_struct *tsk = current; |
f0bc0a60b
|
3062 |
|
1da177e4c
|
3063 3064 3065 |
struct reclaim_state reclaim_state = { .reclaimed_slab = 0, }; |
a70f73028
|
3066 |
const struct cpumask *cpumask = cpumask_of_node(pgdat->node_id); |
1da177e4c
|
3067 |
|
cf40bd16f
|
3068 |
lockdep_set_current_reclaim_state(GFP_KERNEL); |
174596a0b
|
3069 |
if (!cpumask_empty(cpumask)) |
c5f59f083
|
3070 |
set_cpus_allowed_ptr(tsk, cpumask); |
1da177e4c
|
3071 3072 3073 3074 3075 3076 3077 3078 3079 3080 3081 3082 3083 3084 |
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
|
3085 |
tsk->flags |= PF_MEMALLOC | PF_SWAPWRITE | PF_KSWAPD; |
831441862
|
3086 |
set_freezable(); |
1da177e4c
|
3087 |
|
215ddd666
|
3088 |
order = new_order = 0; |
d2ebd0f6b
|
3089 |
balanced_order = 0; |
215ddd666
|
3090 |
classzone_idx = new_classzone_idx = pgdat->nr_zones - 1; |
d2ebd0f6b
|
3091 |
balanced_classzone_idx = classzone_idx; |
1da177e4c
|
3092 |
for ( ; ; ) { |
6f6313d48
|
3093 |
bool ret; |
3e1d1d28d
|
3094 |
|
215ddd666
|
3095 3096 3097 3098 3099 |
/* * 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
|
3100 3101 |
if (balanced_classzone_idx >= new_classzone_idx && balanced_order == new_order) { |
215ddd666
|
3102 3103 3104 3105 3106 |
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
|
3107 |
if (order < new_order || classzone_idx > new_classzone_idx) { |
1da177e4c
|
3108 3109 |
/* * Don't sleep if someone wants a larger 'order' |
995047488
|
3110 |
* allocation or has tigher zone constraints |
1da177e4c
|
3111 3112 |
*/ order = new_order; |
995047488
|
3113 |
classzone_idx = new_classzone_idx; |
1da177e4c
|
3114 |
} else { |
d2ebd0f6b
|
3115 3116 |
kswapd_try_to_sleep(pgdat, balanced_order, balanced_classzone_idx); |
1da177e4c
|
3117 |
order = pgdat->kswapd_max_order; |
995047488
|
3118 |
classzone_idx = pgdat->classzone_idx; |
f0dfcde09
|
3119 3120 |
new_order = order; new_classzone_idx = classzone_idx; |
4d40502ea
|
3121 |
pgdat->kswapd_max_order = 0; |
215ddd666
|
3122 |
pgdat->classzone_idx = pgdat->nr_zones - 1; |
1da177e4c
|
3123 |
} |
1da177e4c
|
3124 |
|
8fe23e057
|
3125 3126 3127 3128 3129 3130 3131 3132 |
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
|
3133 3134 |
if (!ret) { trace_mm_vmscan_kswapd_wake(pgdat->node_id, order); |
d2ebd0f6b
|
3135 3136 3137 |
balanced_classzone_idx = classzone_idx; balanced_order = balance_pgdat(pgdat, order, &balanced_classzone_idx); |
33906bc5c
|
3138 |
} |
1da177e4c
|
3139 |
} |
b0a8cc58e
|
3140 3141 |
current->reclaim_state = NULL; |
1da177e4c
|
3142 3143 3144 3145 3146 3147 |
return 0; } /* * A zone is low on free memory, so wake its kswapd task to service it. */ |
995047488
|
3148 |
void wakeup_kswapd(struct zone *zone, int order, enum zone_type classzone_idx) |
1da177e4c
|
3149 3150 |
{ pg_data_t *pgdat; |
f3fe65122
|
3151 |
if (!populated_zone(zone)) |
1da177e4c
|
3152 |
return; |
88f5acf88
|
3153 |
if (!cpuset_zone_allowed_hardwall(zone, GFP_KERNEL)) |
1da177e4c
|
3154 |
return; |
88f5acf88
|
3155 |
pgdat = zone->zone_pgdat; |
995047488
|
3156 |
if (pgdat->kswapd_max_order < order) { |
1da177e4c
|
3157 |
pgdat->kswapd_max_order = order; |
995047488
|
3158 3159 |
pgdat->classzone_idx = min(pgdat->classzone_idx, classzone_idx); } |
8d0986e28
|
3160 |
if (!waitqueue_active(&pgdat->kswapd_wait)) |
1da177e4c
|
3161 |
return; |
892f795df
|
3162 |
if (zone_balanced(zone, order, 0, 0)) |
88f5acf88
|
3163 3164 3165 |
return; trace_mm_vmscan_wakeup_kswapd(pgdat->node_id, zone_idx(zone), order); |
8d0986e28
|
3166 |
wake_up_interruptible(&pgdat->kswapd_wait); |
1da177e4c
|
3167 |
} |
c6f37f121
|
3168 |
#ifdef CONFIG_HIBERNATION |
1da177e4c
|
3169 |
/* |
7b51755c3
|
3170 |
* Try to free `nr_to_reclaim' of memory, system-wide, and return the number of |
d6277db4a
|
3171 3172 3173 3174 3175 |
* 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
|
3176 |
*/ |
7b51755c3
|
3177 |
unsigned long shrink_all_memory(unsigned long nr_to_reclaim) |
1da177e4c
|
3178 |
{ |
d6277db4a
|
3179 |
struct reclaim_state reclaim_state; |
d6277db4a
|
3180 |
struct scan_control sc = { |
7b51755c3
|
3181 3182 3183 |
.gfp_mask = GFP_HIGHUSER_MOVABLE, .may_swap = 1, .may_unmap = 1, |
d6277db4a
|
3184 |
.may_writepage = 1, |
7b51755c3
|
3185 3186 |
.nr_to_reclaim = nr_to_reclaim, .hibernation_mode = 1, |
7b51755c3
|
3187 |
.order = 0, |
9e3b2f8cd
|
3188 |
.priority = DEF_PRIORITY, |
1da177e4c
|
3189 |
}; |
a09ed5e00
|
3190 3191 3192 3193 |
struct shrink_control shrink = { .gfp_mask = sc.gfp_mask, }; struct zonelist *zonelist = node_zonelist(numa_node_id(), sc.gfp_mask); |
7b51755c3
|
3194 3195 |
struct task_struct *p = current; unsigned long nr_reclaimed; |
1da177e4c
|
3196 |
|
7b51755c3
|
3197 3198 3199 3200 |
p->flags |= PF_MEMALLOC; lockdep_set_current_reclaim_state(sc.gfp_mask); reclaim_state.reclaimed_slab = 0; p->reclaim_state = &reclaim_state; |
d6277db4a
|
3201 |
|
a09ed5e00
|
3202 |
nr_reclaimed = do_try_to_free_pages(zonelist, &sc, &shrink); |
d979677c4
|
3203 |
|
7b51755c3
|
3204 3205 3206 |
p->reclaim_state = NULL; lockdep_clear_current_reclaim_state(); p->flags &= ~PF_MEMALLOC; |
d6277db4a
|
3207 |
|
7b51755c3
|
3208 |
return nr_reclaimed; |
1da177e4c
|
3209 |
} |
c6f37f121
|
3210 |
#endif /* CONFIG_HIBERNATION */ |
1da177e4c
|
3211 |
|
1da177e4c
|
3212 3213 3214 3215 |
/* 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
|
3216 3217 |
static int cpu_callback(struct notifier_block *nfb, unsigned long action, void *hcpu) |
1da177e4c
|
3218 |
{ |
58c0a4a78
|
3219 |
int nid; |
1da177e4c
|
3220 |
|
8bb784428
|
3221 |
if (action == CPU_ONLINE || action == CPU_ONLINE_FROZEN) { |
48fb2e240
|
3222 |
for_each_node_state(nid, N_MEMORY) { |
c5f59f083
|
3223 |
pg_data_t *pgdat = NODE_DATA(nid); |
a70f73028
|
3224 3225 3226 |
const struct cpumask *mask; mask = cpumask_of_node(pgdat->node_id); |
c5f59f083
|
3227 |
|
3e5979453
|
3228 |
if (cpumask_any_and(cpu_online_mask, mask) < nr_cpu_ids) |
1da177e4c
|
3229 |
/* One of our CPUs online: restore mask */ |
c5f59f083
|
3230 |
set_cpus_allowed_ptr(pgdat->kswapd, mask); |
1da177e4c
|
3231 3232 3233 3234 |
} } return NOTIFY_OK; } |
1da177e4c
|
3235 |
|
3218ae14b
|
3236 3237 3238 3239 3240 3241 3242 3243 3244 3245 3246 3247 3248 3249 3250 3251 |
/* * 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
|
3252 3253 3254 |
pr_err("Failed to start kswapd on node %d ", nid); ret = PTR_ERR(pgdat->kswapd); |
d72515b85
|
3255 |
pgdat->kswapd = NULL; |
3218ae14b
|
3256 3257 3258 |
} return ret; } |
8fe23e057
|
3259 |
/* |
d8adde17e
|
3260 3261 |
* Called by memory hotplug when all memory in a node is offlined. Caller must * hold lock_memory_hotplug(). |
8fe23e057
|
3262 3263 3264 3265 |
*/ void kswapd_stop(int nid) { struct task_struct *kswapd = NODE_DATA(nid)->kswapd; |
d8adde17e
|
3266 |
if (kswapd) { |
8fe23e057
|
3267 |
kthread_stop(kswapd); |
d8adde17e
|
3268 3269 |
NODE_DATA(nid)->kswapd = NULL; } |
8fe23e057
|
3270 |
} |
1da177e4c
|
3271 3272 |
static int __init kswapd_init(void) { |
3218ae14b
|
3273 |
int nid; |
69e05944a
|
3274 |
|
1da177e4c
|
3275 |
swap_setup(); |
48fb2e240
|
3276 |
for_each_node_state(nid, N_MEMORY) |
3218ae14b
|
3277 |
kswapd_run(nid); |
1da177e4c
|
3278 3279 3280 3281 3282 |
hotcpu_notifier(cpu_callback, 0); return 0; } module_init(kswapd_init) |
9eeff2395
|
3283 3284 3285 3286 3287 3288 3289 |
#ifdef CONFIG_NUMA /* * Zone reclaim mode * * If non-zero call zone_reclaim when the number of free pages falls below * the watermarks. |
9eeff2395
|
3290 3291 |
*/ int zone_reclaim_mode __read_mostly; |
1b2ffb789
|
3292 |
#define RECLAIM_OFF 0 |
7d03431cf
|
3293 |
#define RECLAIM_ZONE (1<<0) /* Run shrink_inactive_list on the zone */ |
1b2ffb789
|
3294 3295 |
#define RECLAIM_WRITE (1<<1) /* Writeout pages during reclaim */ #define RECLAIM_SWAP (1<<2) /* Swap pages out during reclaim */ |
9eeff2395
|
3296 |
/* |
a92f71263
|
3297 3298 3299 3300 3301 |
* 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
|
3302 |
/* |
9614634fe
|
3303 3304 3305 3306 3307 3308 |
* Percentage of pages in a zone that must be unmapped for zone_reclaim to * occur. */ int sysctl_min_unmapped_ratio = 1; /* |
0ff38490c
|
3309 3310 3311 3312 |
* 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
|
3313 3314 3315 3316 3317 3318 3319 3320 3321 3322 3323 3324 3325 3326 3327 3328 3329 3330 3331 3332 3333 3334 3335 3336 3337 3338 3339 3340 3341 3342 3343 3344 3345 3346 3347 3348 3349 3350 3351 3352 3353 |
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
|
3354 |
/* |
9eeff2395
|
3355 3356 |
* Try to free up some pages from this zone through reclaim. */ |
179e96395
|
3357 |
static int __zone_reclaim(struct zone *zone, gfp_t gfp_mask, unsigned int order) |
9eeff2395
|
3358 |
{ |
7fb2d46d3
|
3359 |
/* Minimum pages needed in order to stay on node */ |
69e05944a
|
3360 |
const unsigned long nr_pages = 1 << order; |
9eeff2395
|
3361 3362 |
struct task_struct *p = current; struct reclaim_state reclaim_state; |
179e96395
|
3363 3364 |
struct scan_control sc = { .may_writepage = !!(zone_reclaim_mode & RECLAIM_WRITE), |
a6dc60f89
|
3365 |
.may_unmap = !!(zone_reclaim_mode & RECLAIM_SWAP), |
2e2e42598
|
3366 |
.may_swap = 1, |
62b726c1b
|
3367 |
.nr_to_reclaim = max(nr_pages, SWAP_CLUSTER_MAX), |
21caf2fc1
|
3368 |
.gfp_mask = (gfp_mask = memalloc_noio_flags(gfp_mask)), |
bd2f6199c
|
3369 |
.order = order, |
9e3b2f8cd
|
3370 |
.priority = ZONE_RECLAIM_PRIORITY, |
179e96395
|
3371 |
}; |
a09ed5e00
|
3372 3373 3374 |
struct shrink_control shrink = { .gfp_mask = sc.gfp_mask, }; |
157480489
|
3375 |
unsigned long nr_slab_pages0, nr_slab_pages1; |
9eeff2395
|
3376 |
|
9eeff2395
|
3377 |
cond_resched(); |
d4f7796e9
|
3378 3379 3380 3381 3382 3383 |
/* * 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
|
3384 |
lockdep_set_current_reclaim_state(gfp_mask); |
9eeff2395
|
3385 3386 |
reclaim_state.reclaimed_slab = 0; p->reclaim_state = &reclaim_state; |
c84db23c6
|
3387 |
|
90afa5de6
|
3388 |
if (zone_pagecache_reclaimable(zone) > zone->min_unmapped_pages) { |
0ff38490c
|
3389 3390 3391 3392 |
/* * Free memory by calling shrink zone with increasing * priorities until we have enough memory freed. */ |
0ff38490c
|
3393 |
do { |
9e3b2f8cd
|
3394 3395 |
shrink_zone(zone, &sc); } while (sc.nr_reclaimed < nr_pages && --sc.priority >= 0); |
0ff38490c
|
3396 |
} |
c84db23c6
|
3397 |
|
157480489
|
3398 3399 |
nr_slab_pages0 = zone_page_state(zone, NR_SLAB_RECLAIMABLE); if (nr_slab_pages0 > zone->min_slab_pages) { |
2a16e3f4b
|
3400 |
/* |
7fb2d46d3
|
3401 |
* shrink_slab() does not currently allow us to determine how |
0ff38490c
|
3402 3403 3404 3405 |
* 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
|
3406 |
*/ |
0ce3d7445
|
3407 3408 |
nodes_clear(shrink.nodes_to_scan); node_set(zone_to_nid(zone), shrink.nodes_to_scan); |
4dc4b3d97
|
3409 3410 3411 3412 |
for (;;) { unsigned long lru_pages = zone_reclaimable_pages(zone); /* No reclaimable slab or very low memory pressure */ |
1495f230f
|
3413 |
if (!shrink_slab(&shrink, sc.nr_scanned, lru_pages)) |
4dc4b3d97
|
3414 3415 3416 3417 3418 3419 3420 3421 |
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
|
3422 3423 3424 3425 3426 |
/* * Update nr_reclaimed by the number of slab pages we * reclaimed from this zone. */ |
157480489
|
3427 3428 3429 |
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
|
3430 |
} |
9eeff2395
|
3431 |
p->reclaim_state = NULL; |
d4f7796e9
|
3432 |
current->flags &= ~(PF_MEMALLOC | PF_SWAPWRITE); |
76ca542d8
|
3433 |
lockdep_clear_current_reclaim_state(); |
a79311c14
|
3434 |
return sc.nr_reclaimed >= nr_pages; |
9eeff2395
|
3435 |
} |
179e96395
|
3436 3437 3438 |
int zone_reclaim(struct zone *zone, gfp_t gfp_mask, unsigned int order) { |
179e96395
|
3439 |
int node_id; |
d773ed6b8
|
3440 |
int ret; |
179e96395
|
3441 3442 |
/* |
0ff38490c
|
3443 3444 |
* Zone reclaim reclaims unmapped file backed pages and * slab pages if we are over the defined limits. |
34aa1330f
|
3445 |
* |
9614634fe
|
3446 3447 3448 3449 3450 |
* 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
|
3451 |
*/ |
90afa5de6
|
3452 3453 |
if (zone_pagecache_reclaimable(zone) <= zone->min_unmapped_pages && zone_page_state(zone, NR_SLAB_RECLAIMABLE) <= zone->min_slab_pages) |
fa5e084e4
|
3454 |
return ZONE_RECLAIM_FULL; |
179e96395
|
3455 |
|
6e543d578
|
3456 |
if (!zone_reclaimable(zone)) |
fa5e084e4
|
3457 |
return ZONE_RECLAIM_FULL; |
d773ed6b8
|
3458 |
|
179e96395
|
3459 |
/* |
d773ed6b8
|
3460 |
* Do not scan if the allocation should not be delayed. |
179e96395
|
3461 |
*/ |
d773ed6b8
|
3462 |
if (!(gfp_mask & __GFP_WAIT) || (current->flags & PF_MEMALLOC)) |
fa5e084e4
|
3463 |
return ZONE_RECLAIM_NOSCAN; |
179e96395
|
3464 3465 3466 3467 3468 3469 3470 |
/* * 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
|
3471 |
node_id = zone_to_nid(zone); |
37c0708db
|
3472 |
if (node_state(node_id, N_CPU) && node_id != numa_node_id()) |
fa5e084e4
|
3473 |
return ZONE_RECLAIM_NOSCAN; |
d773ed6b8
|
3474 3475 |
if (zone_test_and_set_flag(zone, ZONE_RECLAIM_LOCKED)) |
fa5e084e4
|
3476 |
return ZONE_RECLAIM_NOSCAN; |
d773ed6b8
|
3477 3478 |
ret = __zone_reclaim(zone, gfp_mask, order); zone_clear_flag(zone, ZONE_RECLAIM_LOCKED); |
24cf72518
|
3479 3480 |
if (!ret) count_vm_event(PGSCAN_ZONE_RECLAIM_FAILED); |
d773ed6b8
|
3481 |
return ret; |
179e96395
|
3482 |
} |
9eeff2395
|
3483 |
#endif |
894bc3104
|
3484 |
|
894bc3104
|
3485 3486 3487 |
/* * page_evictable - test whether a page is evictable * @page: the page to test |
894bc3104
|
3488 3489 |
* * Test whether page is evictable--i.e., should be placed on active/inactive |
39b5f29ac
|
3490 |
* lists vs unevictable list. |
894bc3104
|
3491 3492 |
* * Reasons page might not be evictable: |
ba9ddf493
|
3493 |
* (1) page's mapping marked unevictable |
b291f0003
|
3494 |
* (2) page is part of an mlocked VMA |
ba9ddf493
|
3495 |
* |
894bc3104
|
3496 |
*/ |
39b5f29ac
|
3497 |
int page_evictable(struct page *page) |
894bc3104
|
3498 |
{ |
39b5f29ac
|
3499 |
return !mapping_unevictable(page_mapping(page)) && !PageMlocked(page); |
894bc3104
|
3500 |
} |
89e004ea5
|
3501 |
|
85046579b
|
3502 |
#ifdef CONFIG_SHMEM |
89e004ea5
|
3503 |
/** |
245132643
|
3504 3505 3506 |
* 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
|
3507 |
* |
245132643
|
3508 |
* Checks pages for evictability and moves them to the appropriate lru list. |
85046579b
|
3509 3510 |
* * This function is only used for SysV IPC SHM_UNLOCK. |
89e004ea5
|
3511 |
*/ |
245132643
|
3512 |
void check_move_unevictable_pages(struct page **pages, int nr_pages) |
89e004ea5
|
3513 |
{ |
925b7673c
|
3514 |
struct lruvec *lruvec; |
245132643
|
3515 3516 3517 3518 |
struct zone *zone = NULL; int pgscanned = 0; int pgrescued = 0; int i; |
89e004ea5
|
3519 |
|
245132643
|
3520 3521 3522 |
for (i = 0; i < nr_pages; i++) { struct page *page = pages[i]; struct zone *pagezone; |
89e004ea5
|
3523 |
|
245132643
|
3524 3525 3526 3527 3528 3529 3530 3531 |
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
|
3532 |
lruvec = mem_cgroup_page_lruvec(page, zone); |
89e004ea5
|
3533 |
|
245132643
|
3534 3535 |
if (!PageLRU(page) || !PageUnevictable(page)) continue; |
89e004ea5
|
3536 |
|
39b5f29ac
|
3537 |
if (page_evictable(page)) { |
245132643
|
3538 |
enum lru_list lru = page_lru_base_type(page); |
309381fea
|
3539 |
VM_BUG_ON_PAGE(PageActive(page), page); |
245132643
|
3540 |
ClearPageUnevictable(page); |
fa9add641
|
3541 3542 |
del_page_from_lru_list(page, lruvec, LRU_UNEVICTABLE); add_page_to_lru_list(page, lruvec, lru); |
245132643
|
3543 |
pgrescued++; |
89e004ea5
|
3544 |
} |
245132643
|
3545 |
} |
89e004ea5
|
3546 |
|
245132643
|
3547 3548 3549 3550 |
if (zone) { __count_vm_events(UNEVICTABLE_PGRESCUED, pgrescued); __count_vm_events(UNEVICTABLE_PGSCANNED, pgscanned); spin_unlock_irq(&zone->lru_lock); |
89e004ea5
|
3551 |
} |
89e004ea5
|
3552 |
} |
85046579b
|
3553 |
#endif /* CONFIG_SHMEM */ |
af936a160
|
3554 |
|
264e56d82
|
3555 |
static void warn_scan_unevictable_pages(void) |
af936a160
|
3556 |
{ |
264e56d82
|
3557 |
printk_once(KERN_WARNING |
25bd91bd2
|
3558 |
"%s: The scan_unevictable_pages sysctl/node-interface has been " |
264e56d82
|
3559 |
"disabled for lack of a legitimate use case. If you have " |
25bd91bd2
|
3560 3561 3562 |
"one, please send an email to linux-mm@kvack.org. ", current->comm); |
af936a160
|
3563 3564 3565 3566 3567 3568 3569 3570 3571 |
} /* * scan_unevictable_pages [vm] sysctl handler. On demand re-scan of * all nodes' unevictable lists for evictable pages */ unsigned long scan_unevictable_pages; int scan_unevictable_handler(struct ctl_table *table, int write, |
8d65af789
|
3572 |
void __user *buffer, |
af936a160
|
3573 3574 |
size_t *length, loff_t *ppos) { |
264e56d82
|
3575 |
warn_scan_unevictable_pages(); |
8d65af789
|
3576 |
proc_doulongvec_minmax(table, write, buffer, length, ppos); |
af936a160
|
3577 3578 3579 |
scan_unevictable_pages = 0; return 0; } |
e4455abb5
|
3580 |
#ifdef CONFIG_NUMA |
af936a160
|
3581 3582 3583 3584 |
/* * per node 'scan_unevictable_pages' attribute. On demand re-scan of * a specified node's per zone unevictable lists for evictable pages. */ |
10fbcf4c6
|
3585 3586 |
static ssize_t read_scan_unevictable_node(struct device *dev, struct device_attribute *attr, |
af936a160
|
3587 3588 |
char *buf) { |
264e56d82
|
3589 |
warn_scan_unevictable_pages(); |
af936a160
|
3590 3591 3592 |
return sprintf(buf, "0 "); /* always zero; should fit... */ } |
10fbcf4c6
|
3593 3594 |
static ssize_t write_scan_unevictable_node(struct device *dev, struct device_attribute *attr, |
af936a160
|
3595 3596 |
const char *buf, size_t count) { |
264e56d82
|
3597 |
warn_scan_unevictable_pages(); |
af936a160
|
3598 3599 |
return 1; } |
10fbcf4c6
|
3600 |
static DEVICE_ATTR(scan_unevictable_pages, S_IRUGO | S_IWUSR, |
af936a160
|
3601 3602 3603 3604 3605 |
read_scan_unevictable_node, write_scan_unevictable_node); int scan_unevictable_register_node(struct node *node) { |
10fbcf4c6
|
3606 |
return device_create_file(&node->dev, &dev_attr_scan_unevictable_pages); |
af936a160
|
3607 3608 3609 3610 |
} void scan_unevictable_unregister_node(struct node *node) { |
10fbcf4c6
|
3611 |
device_remove_file(&node->dev, &dev_attr_scan_unevictable_pages); |
af936a160
|
3612 |
} |
e4455abb5
|
3613 |
#endif |