Blame view
mm/vmscan.c
98.2 KB
1da177e4c Linux-2.6.12-rc2 |
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 include cleanup: ... |
16 |
#include <linux/gfp.h> |
1da177e4c Linux-2.6.12-rc2 |
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> |
e129b5c23 [PATCH] vm: add p... |
22 |
#include <linux/vmstat.h> |
1da177e4c Linux-2.6.12-rc2 |
23 24 25 26 27 28 29 30 31 32 33 34 |
#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> #include <linux/pagevec.h> #include <linux/backing-dev.h> #include <linux/rmap.h> #include <linux/topology.h> #include <linux/cpu.h> #include <linux/cpuset.h> |
3e7d34497 mm: vmscan: recla... |
35 |
#include <linux/compaction.h> |
1da177e4c Linux-2.6.12-rc2 |
36 37 |
#include <linux/notifier.h> #include <linux/rwsem.h> |
248a0301e [PATCH] mm: make ... |
38 |
#include <linux/delay.h> |
3218ae14b [PATCH] pgdat all... |
39 |
#include <linux/kthread.h> |
7dfb71030 [PATCH] Add inclu... |
40 |
#include <linux/freezer.h> |
66e1707bc Memory controller... |
41 |
#include <linux/memcontrol.h> |
873b47717 per-task-delay-ac... |
42 |
#include <linux/delayacct.h> |
af936a160 vmscan: unevictab... |
43 |
#include <linux/sysctl.h> |
929bea7c7 vmscan: all_unrec... |
44 |
#include <linux/oom.h> |
268bb0ce3 sanitize <linux/p... |
45 |
#include <linux/prefetch.h> |
1da177e4c Linux-2.6.12-rc2 |
46 47 48 49 50 |
#include <asm/tlbflush.h> #include <asm/div64.h> #include <linux/swapops.h> |
0f8053a50 [PATCH] mm: make ... |
51 |
#include "internal.h" |
33906bc5c vmscan: tracing: ... |
52 53 |
#define CREATE_TRACE_POINTS #include <trace/events/vmscan.h> |
ee64fc935 mm: vmscan: conve... |
54 |
/* |
f3a310bc4 mm: vmscan: renam... |
55 56 57 58 59 |
* reclaim_mode determines how the inactive list is shrunk * RECLAIM_MODE_SINGLE: Reclaim only order-0 pages * RECLAIM_MODE_ASYNC: Do not block * RECLAIM_MODE_SYNC: Allow blocking e.g. call wait_on_page_writeback * RECLAIM_MODE_LUMPYRECLAIM: For high-order allocations, take a reference |
ee64fc935 mm: vmscan: conve... |
60 61 |
* page from the LRU and reclaim all pages within a * naturally aligned range |
f3a310bc4 mm: vmscan: renam... |
62 |
* RECLAIM_MODE_COMPACTION: For high-order allocations, reclaim a number of |
3e7d34497 mm: vmscan: recla... |
63 |
* order-0 pages and then compact the zone |
ee64fc935 mm: vmscan: conve... |
64 |
*/ |
f3a310bc4 mm: vmscan: renam... |
65 66 67 68 69 70 |
typedef unsigned __bitwise__ reclaim_mode_t; #define RECLAIM_MODE_SINGLE ((__force reclaim_mode_t)0x01u) #define RECLAIM_MODE_ASYNC ((__force reclaim_mode_t)0x02u) #define RECLAIM_MODE_SYNC ((__force reclaim_mode_t)0x04u) #define RECLAIM_MODE_LUMPYRECLAIM ((__force reclaim_mode_t)0x08u) #define RECLAIM_MODE_COMPACTION ((__force reclaim_mode_t)0x10u) |
7d3579e8e vmscan: narrow th... |
71 |
|
1da177e4c Linux-2.6.12-rc2 |
72 |
struct scan_control { |
1da177e4c Linux-2.6.12-rc2 |
73 74 |
/* Incremented by the number of inactive pages that were scanned */ unsigned long nr_scanned; |
a79311c14 vmscan: bail out ... |
75 76 |
/* Number of pages freed so far during a call to shrink_zones() */ unsigned long nr_reclaimed; |
22fba3354 vmscan: separate ... |
77 78 |
/* How many pages shrink_list() should reclaim */ unsigned long nr_to_reclaim; |
7b51755c3 vmscan: kill hibe... |
79 |
unsigned long hibernation_mode; |
1da177e4c Linux-2.6.12-rc2 |
80 |
/* This context's GFP mask */ |
6daa0e286 [PATCH] gfp_t: mm... |
81 |
gfp_t gfp_mask; |
1da177e4c Linux-2.6.12-rc2 |
82 83 |
int may_writepage; |
a6dc60f89 vmscan: rename sc... |
84 85 |
/* Can mapped pages be reclaimed? */ int may_unmap; |
f1fd1067e [PATCH] Zone recl... |
86 |
|
2e2e42598 vmscan,memcg: rei... |
87 88 |
/* Can pages be swapped as part of reclaim? */ int may_swap; |
5ad333eb6 Lumpy Reclaim V4 |
89 |
int order; |
66e1707bc Memory controller... |
90 |
|
5f53e7629 vmscan: page_chec... |
91 |
/* |
415b54e37 Fix typo s/conten... |
92 93 |
* Intend to reclaim enough continuous memory rather than reclaim * enough amount of memory. i.e, mode for high order allocation. |
5f53e7629 vmscan: page_chec... |
94 |
*/ |
f3a310bc4 mm: vmscan: renam... |
95 |
reclaim_mode_t reclaim_mode; |
5f53e7629 vmscan: page_chec... |
96 |
|
66e1707bc Memory controller... |
97 98 |
/* Which cgroup do we reclaim from */ struct mem_cgroup *mem_cgroup; |
327c0e968 vmscan: fix it to... |
99 100 101 102 103 |
/* * Nodemask of nodes allowed by the caller. If NULL, all nodes * are scanned. */ nodemask_t *nodemask; |
1da177e4c Linux-2.6.12-rc2 |
104 |
}; |
1da177e4c Linux-2.6.12-rc2 |
105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 |
#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; |
bd1e22b8e [PATCH] initialis... |
139 |
long vm_total_pages; /* The total number of pages which the VM controls */ |
1da177e4c Linux-2.6.12-rc2 |
140 141 142 |
static LIST_HEAD(shrinker_list); static DECLARE_RWSEM(shrinker_rwsem); |
00f0b8259 Memory controller... |
143 |
#ifdef CONFIG_CGROUP_MEM_RES_CTLR |
e72e2bd67 memcg: rename sca... |
144 |
#define scanning_global_lru(sc) (!(sc)->mem_cgroup) |
91a45470f per-zone and recl... |
145 |
#else |
e72e2bd67 memcg: rename sca... |
146 |
#define scanning_global_lru(sc) (1) |
91a45470f per-zone and recl... |
147 |
#endif |
6e9015716 mm: introduce zon... |
148 149 150 |
static struct zone_reclaim_stat *get_reclaim_stat(struct zone *zone, struct scan_control *sc) { |
e72e2bd67 memcg: rename sca... |
151 |
if (!scanning_global_lru(sc)) |
3e2f41f1f memcg: add zone_r... |
152 |
return mem_cgroup_get_reclaim_stat(sc->mem_cgroup, zone); |
6e9015716 mm: introduce zon... |
153 154 |
return &zone->reclaim_stat; } |
0b2176763 mm/vmscan: rename... |
155 156 |
static unsigned long zone_nr_lru_pages(struct zone *zone, struct scan_control *sc, enum lru_list lru) |
c9f299d98 mm: add zone nr_p... |
157 |
{ |
e72e2bd67 memcg: rename sca... |
158 |
if (!scanning_global_lru(sc)) |
bb2a0de92 memcg: consolidat... |
159 160 |
return mem_cgroup_zone_nr_lru_pages(sc->mem_cgroup, zone_to_nid(zone), zone_idx(zone), BIT(lru)); |
a3d8e0549 memcg: add mem_cg... |
161 |
|
c9f299d98 mm: add zone nr_p... |
162 163 |
return zone_page_state(zone, NR_LRU_BASE + lru); } |
1da177e4c Linux-2.6.12-rc2 |
164 165 166 |
/* * Add a shrinker callback to be called from the vm */ |
8e1f936b7 mm: clean up and ... |
167 |
void register_shrinker(struct shrinker *shrinker) |
1da177e4c Linux-2.6.12-rc2 |
168 |
{ |
83aeeada7 vmscan: use atomi... |
169 |
atomic_long_set(&shrinker->nr_in_batch, 0); |
8e1f936b7 mm: clean up and ... |
170 171 172 |
down_write(&shrinker_rwsem); list_add_tail(&shrinker->list, &shrinker_list); up_write(&shrinker_rwsem); |
1da177e4c Linux-2.6.12-rc2 |
173 |
} |
8e1f936b7 mm: clean up and ... |
174 |
EXPORT_SYMBOL(register_shrinker); |
1da177e4c Linux-2.6.12-rc2 |
175 176 177 178 |
/* * Remove one */ |
8e1f936b7 mm: clean up and ... |
179 |
void unregister_shrinker(struct shrinker *shrinker) |
1da177e4c Linux-2.6.12-rc2 |
180 181 182 183 |
{ down_write(&shrinker_rwsem); list_del(&shrinker->list); up_write(&shrinker_rwsem); |
1da177e4c Linux-2.6.12-rc2 |
184 |
} |
8e1f936b7 mm: clean up and ... |
185 |
EXPORT_SYMBOL(unregister_shrinker); |
1da177e4c Linux-2.6.12-rc2 |
186 |
|
1495f230f vmscan: change sh... |
187 188 189 190 191 192 193 |
static inline int do_shrinker_shrink(struct shrinker *shrinker, struct shrink_control *sc, unsigned long nr_to_scan) { sc->nr_to_scan = nr_to_scan; return (*shrinker->shrink)(shrinker, sc); } |
1da177e4c Linux-2.6.12-rc2 |
194 195 196 197 198 199 200 201 202 |
#define SHRINK_BATCH 128 /* * 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 spelling fixes: mm/ |
203 |
* If the vm encountered mapped pages on the LRU it increase the pressure on |
1da177e4c Linux-2.6.12-rc2 |
204 205 206 207 208 209 210 |
* 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 [PATCH] vmscan: n... |
211 212 |
* * Returns the number of slab objects which we shrunk. |
1da177e4c Linux-2.6.12-rc2 |
213 |
*/ |
a09ed5e00 vmscan: change sh... |
214 |
unsigned long shrink_slab(struct shrink_control *shrink, |
1495f230f vmscan: change sh... |
215 |
unsigned long nr_pages_scanned, |
a09ed5e00 vmscan: change sh... |
216 |
unsigned long lru_pages) |
1da177e4c Linux-2.6.12-rc2 |
217 218 |
{ struct shrinker *shrinker; |
69e05944a [PATCH] vmscan: u... |
219 |
unsigned long ret = 0; |
1da177e4c Linux-2.6.12-rc2 |
220 |
|
1495f230f vmscan: change sh... |
221 222 |
if (nr_pages_scanned == 0) nr_pages_scanned = SWAP_CLUSTER_MAX; |
1da177e4c Linux-2.6.12-rc2 |
223 |
|
f06590bd7 mm: vmscan: corre... |
224 225 226 227 228 |
if (!down_read_trylock(&shrinker_rwsem)) { /* Assume we'll be able to shrink next time */ ret = 1; goto out; } |
1da177e4c Linux-2.6.12-rc2 |
229 230 231 |
list_for_each_entry(shrinker, &shrinker_list, list) { unsigned long long delta; |
635697c66 vmscan: fix initi... |
232 233 |
long total_scan; long max_pass; |
095760730 vmscan: add shrin... |
234 |
int shrink_ret = 0; |
acf92b485 vmscan: shrinker-... |
235 236 |
long nr; long new_nr; |
e9299f505 vmscan: add custo... |
237 238 |
long batch_size = shrinker->batch ? shrinker->batch : SHRINK_BATCH; |
1da177e4c Linux-2.6.12-rc2 |
239 |
|
635697c66 vmscan: fix initi... |
240 241 242 |
max_pass = do_shrinker_shrink(shrinker, shrink, 0); if (max_pass <= 0) continue; |
acf92b485 vmscan: shrinker-... |
243 244 245 246 247 |
/* * 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. */ |
83aeeada7 vmscan: use atomi... |
248 |
nr = atomic_long_xchg(&shrinker->nr_in_batch, 0); |
acf92b485 vmscan: shrinker-... |
249 250 |
total_scan = nr; |
1495f230f vmscan: change sh... |
251 |
delta = (4 * nr_pages_scanned) / shrinker->seeks; |
ea164d73a [PATCH] shrinker-... |
252 |
delta *= max_pass; |
1da177e4c Linux-2.6.12-rc2 |
253 |
do_div(delta, lru_pages + 1); |
acf92b485 vmscan: shrinker-... |
254 255 |
total_scan += delta; if (total_scan < 0) { |
88c3bd707 vmscan: print shr... |
256 257 258 |
printk(KERN_ERR "shrink_slab: %pF negative objects to " "delete nr=%ld ", |
acf92b485 vmscan: shrinker-... |
259 260 |
shrinker->shrink, total_scan); total_scan = max_pass; |
ea164d73a [PATCH] shrinker-... |
261 262 263 |
} /* |
3567b59aa vmscan: reduce wi... |
264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 |
* 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); /* |
ea164d73a [PATCH] shrinker-... |
279 280 281 282 |
* Avoid risking looping forever due to too large nr value: * never try to free more than twice the estimate number of * freeable entries. */ |
acf92b485 vmscan: shrinker-... |
283 284 |
if (total_scan > max_pass * 2) total_scan = max_pass * 2; |
1da177e4c Linux-2.6.12-rc2 |
285 |
|
acf92b485 vmscan: shrinker-... |
286 |
trace_mm_shrink_slab_start(shrinker, shrink, nr, |
095760730 vmscan: add shrin... |
287 288 |
nr_pages_scanned, lru_pages, max_pass, delta, total_scan); |
e9299f505 vmscan: add custo... |
289 |
while (total_scan >= batch_size) { |
b15e0905f [PATCH] vmscan: n... |
290 |
int nr_before; |
1da177e4c Linux-2.6.12-rc2 |
291 |
|
1495f230f vmscan: change sh... |
292 293 |
nr_before = do_shrinker_shrink(shrinker, shrink, 0); shrink_ret = do_shrinker_shrink(shrinker, shrink, |
e9299f505 vmscan: add custo... |
294 |
batch_size); |
1da177e4c Linux-2.6.12-rc2 |
295 296 |
if (shrink_ret == -1) break; |
b15e0905f [PATCH] vmscan: n... |
297 298 |
if (shrink_ret < nr_before) ret += nr_before - shrink_ret; |
e9299f505 vmscan: add custo... |
299 300 |
count_vm_events(SLABS_SCANNED, batch_size); total_scan -= batch_size; |
1da177e4c Linux-2.6.12-rc2 |
301 302 303 |
cond_resched(); } |
acf92b485 vmscan: shrinker-... |
304 305 306 307 308 |
/* * 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. */ |
83aeeada7 vmscan: use atomi... |
309 310 311 312 313 |
if (total_scan > 0) new_nr = atomic_long_add_return(total_scan, &shrinker->nr_in_batch); else new_nr = atomic_long_read(&shrinker->nr_in_batch); |
acf92b485 vmscan: shrinker-... |
314 315 |
trace_mm_shrink_slab_end(shrinker, shrink_ret, nr, new_nr); |
1da177e4c Linux-2.6.12-rc2 |
316 317 |
} up_read(&shrinker_rwsem); |
f06590bd7 mm: vmscan: corre... |
318 319 |
out: cond_resched(); |
b15e0905f [PATCH] vmscan: n... |
320 |
return ret; |
1da177e4c Linux-2.6.12-rc2 |
321 |
} |
f3a310bc4 mm: vmscan: renam... |
322 |
static void set_reclaim_mode(int priority, struct scan_control *sc, |
7d3579e8e vmscan: narrow th... |
323 324 |
bool sync) { |
f3a310bc4 mm: vmscan: renam... |
325 |
reclaim_mode_t syncmode = sync ? RECLAIM_MODE_SYNC : RECLAIM_MODE_ASYNC; |
7d3579e8e vmscan: narrow th... |
326 327 |
/* |
3e7d34497 mm: vmscan: recla... |
328 329 330 |
* Initially assume we are entering either lumpy reclaim or * reclaim/compaction.Depending on the order, we will either set the * sync mode or just reclaim order-0 pages later. |
7d3579e8e vmscan: narrow th... |
331 |
*/ |
3e7d34497 mm: vmscan: recla... |
332 |
if (COMPACTION_BUILD) |
f3a310bc4 mm: vmscan: renam... |
333 |
sc->reclaim_mode = RECLAIM_MODE_COMPACTION; |
3e7d34497 mm: vmscan: recla... |
334 |
else |
f3a310bc4 mm: vmscan: renam... |
335 |
sc->reclaim_mode = RECLAIM_MODE_LUMPYRECLAIM; |
7d3579e8e vmscan: narrow th... |
336 337 |
/* |
3e7d34497 mm: vmscan: recla... |
338 339 340 |
* Avoid using lumpy reclaim or reclaim/compaction if possible by * restricting when its set to either costly allocations or when * under memory pressure |
7d3579e8e vmscan: narrow th... |
341 342 |
*/ if (sc->order > PAGE_ALLOC_COSTLY_ORDER) |
f3a310bc4 mm: vmscan: renam... |
343 |
sc->reclaim_mode |= syncmode; |
7d3579e8e vmscan: narrow th... |
344 |
else if (sc->order && priority < DEF_PRIORITY - 2) |
f3a310bc4 mm: vmscan: renam... |
345 |
sc->reclaim_mode |= syncmode; |
7d3579e8e vmscan: narrow th... |
346 |
else |
f3a310bc4 mm: vmscan: renam... |
347 |
sc->reclaim_mode = RECLAIM_MODE_SINGLE | RECLAIM_MODE_ASYNC; |
7d3579e8e vmscan: narrow th... |
348 |
} |
f3a310bc4 mm: vmscan: renam... |
349 |
static void reset_reclaim_mode(struct scan_control *sc) |
7d3579e8e vmscan: narrow th... |
350 |
{ |
f3a310bc4 mm: vmscan: renam... |
351 |
sc->reclaim_mode = RECLAIM_MODE_SINGLE | RECLAIM_MODE_ASYNC; |
7d3579e8e vmscan: narrow th... |
352 |
} |
1da177e4c Linux-2.6.12-rc2 |
353 354 |
static inline int is_page_cache_freeable(struct page *page) { |
ceddc3a52 mm: document is_p... |
355 356 357 358 359 |
/* * 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 mm: return boolea... |
360 |
return page_count(page) - page_has_private(page) == 2; |
1da177e4c Linux-2.6.12-rc2 |
361 |
} |
7d3579e8e vmscan: narrow th... |
362 363 |
static int may_write_to_queue(struct backing_dev_info *bdi, struct scan_control *sc) |
1da177e4c Linux-2.6.12-rc2 |
364 |
{ |
930d91525 [PATCH] Swap Migr... |
365 |
if (current->flags & PF_SWAPWRITE) |
1da177e4c Linux-2.6.12-rc2 |
366 367 368 369 370 |
return 1; if (!bdi_write_congested(bdi)) return 1; if (bdi == current->backing_dev_info) return 1; |
7d3579e8e vmscan: narrow th... |
371 372 373 374 |
/* lumpy reclaim for hugepage often need a lot of write */ if (sc->order > PAGE_ALLOC_COSTLY_ORDER) return 1; |
1da177e4c Linux-2.6.12-rc2 |
375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 |
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 block: remove per... |
393 |
lock_page(page); |
3e9f45bd1 Factor outstandin... |
394 395 |
if (page_mapping(page) == mapping) mapping_set_error(mapping, error); |
1da177e4c Linux-2.6.12-rc2 |
396 397 |
unlock_page(page); } |
04e62a29b [PATCH] More page... |
398 399 400 401 402 403 404 405 406 407 408 |
/* 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 Linux-2.6.12-rc2 |
409 |
/* |
1742f19fa [PATCH] vmscan: r... |
410 411 |
* pageout is called by shrink_page_list() for each dirty page. * Calls ->writepage(). |
1da177e4c Linux-2.6.12-rc2 |
412 |
*/ |
c661b078f synchronous lumpy... |
413 |
static pageout_t pageout(struct page *page, struct address_space *mapping, |
7d3579e8e vmscan: narrow th... |
414 |
struct scan_control *sc) |
1da177e4c Linux-2.6.12-rc2 |
415 416 417 418 419 420 421 422 |
{ /* * 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 mm/vmscan: change... |
423 |
* If this process is currently in __generic_file_aio_write() against |
1da177e4c Linux-2.6.12-rc2 |
424 425 426 427 428 429 430 |
* 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 Linux-2.6.12-rc2 |
431 432 433 434 435 436 437 438 |
*/ 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 FS-Cache: Recruit... |
439 |
if (page_has_private(page)) { |
1da177e4c Linux-2.6.12-rc2 |
440 441 |
if (try_to_free_buffers(page)) { ClearPageDirty(page); |
d40cee245 mm: remove remain... |
442 443 |
printk("%s: orphaned page ", __func__); |
1da177e4c Linux-2.6.12-rc2 |
444 445 446 447 448 449 450 |
return PAGE_CLEAN; } } return PAGE_KEEP; } if (mapping->a_ops->writepage == NULL) return PAGE_ACTIVATE; |
0e093d997 writeback: do not... |
451 |
if (!may_write_to_queue(mapping->backing_dev_info, sc)) |
1da177e4c Linux-2.6.12-rc2 |
452 453 454 455 456 457 458 |
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 [PATCH] writeback... |
459 460 |
.range_start = 0, .range_end = LLONG_MAX, |
1da177e4c Linux-2.6.12-rc2 |
461 462 463 464 465 466 467 |
.for_reclaim = 1, }; SetPageReclaim(page); res = mapping->a_ops->writepage(page, &wbc); if (res < 0) handle_write_error(mapping, page, res); |
994fc28c7 [PATCH] add AOP_T... |
468 |
if (res == AOP_WRITEPAGE_ACTIVATE) { |
1da177e4c Linux-2.6.12-rc2 |
469 470 471 |
ClearPageReclaim(page); return PAGE_ACTIVATE; } |
c661b078f synchronous lumpy... |
472 |
|
1da177e4c Linux-2.6.12-rc2 |
473 474 475 476 |
if (!PageWriteback(page)) { /* synchronous write or broken a_ops? */ ClearPageReclaim(page); } |
755f0225e vmscan: tracing: ... |
477 |
trace_mm_vmscan_writepage(page, |
f3a310bc4 mm: vmscan: renam... |
478 |
trace_reclaim_flags(page, sc->reclaim_mode)); |
e129b5c23 [PATCH] vm: add p... |
479 |
inc_zone_page_state(page, NR_VMSCAN_WRITE); |
1da177e4c Linux-2.6.12-rc2 |
480 481 482 483 484 |
return PAGE_SUCCESS; } return PAGE_CLEAN; } |
a649fd927 [PATCH] invalidat... |
485 |
/* |
e286781d5 mm: speculative p... |
486 487 |
* Same as remove_mapping, but if the page is removed from the mapping, it * gets returned with a refcount of 0. |
a649fd927 [PATCH] invalidat... |
488 |
*/ |
e286781d5 mm: speculative p... |
489 |
static int __remove_mapping(struct address_space *mapping, struct page *page) |
49d2e9cc4 [PATCH] Swap Migr... |
490 |
{ |
28e4d965e [PATCH] mm: remov... |
491 492 |
BUG_ON(!PageLocked(page)); BUG_ON(mapping != page_mapping(page)); |
49d2e9cc4 [PATCH] Swap Migr... |
493 |
|
19fd62312 mm: spinlock tree... |
494 |
spin_lock_irq(&mapping->tree_lock); |
49d2e9cc4 [PATCH] Swap Migr... |
495 |
/* |
0fd0e6b05 [PATCH] page inva... |
496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 |
* 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 [PATCH] Swap Migr... |
519 |
*/ |
e286781d5 mm: speculative p... |
520 |
if (!page_freeze_refs(page, 2)) |
49d2e9cc4 [PATCH] Swap Migr... |
521 |
goto cannot_free; |
e286781d5 mm: speculative p... |
522 523 524 |
/* note: atomic_cmpxchg in page_freeze_refs provides the smp_rmb */ if (unlikely(PageDirty(page))) { page_unfreeze_refs(page, 2); |
49d2e9cc4 [PATCH] Swap Migr... |
525 |
goto cannot_free; |
e286781d5 mm: speculative p... |
526 |
} |
49d2e9cc4 [PATCH] Swap Migr... |
527 528 529 530 |
if (PageSwapCache(page)) { swp_entry_t swap = { .val = page_private(page) }; __delete_from_swap_cache(page); |
19fd62312 mm: spinlock tree... |
531 |
spin_unlock_irq(&mapping->tree_lock); |
cb4b86ba4 mm: add swap cach... |
532 |
swapcache_free(swap, page); |
e286781d5 mm: speculative p... |
533 |
} else { |
6072d13c4 Call the filesyst... |
534 535 536 |
void (*freepage)(struct page *); freepage = mapping->a_ops->freepage; |
e64a782fe mm: change __remo... |
537 |
__delete_from_page_cache(page); |
19fd62312 mm: spinlock tree... |
538 |
spin_unlock_irq(&mapping->tree_lock); |
e767e0561 memcg: fix deadlo... |
539 |
mem_cgroup_uncharge_cache_page(page); |
6072d13c4 Call the filesyst... |
540 541 542 |
if (freepage != NULL) freepage(page); |
49d2e9cc4 [PATCH] Swap Migr... |
543 |
} |
49d2e9cc4 [PATCH] Swap Migr... |
544 545 546 |
return 1; cannot_free: |
19fd62312 mm: spinlock tree... |
547 |
spin_unlock_irq(&mapping->tree_lock); |
49d2e9cc4 [PATCH] Swap Migr... |
548 549 |
return 0; } |
1da177e4c Linux-2.6.12-rc2 |
550 |
/* |
e286781d5 mm: speculative p... |
551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 |
* 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 Unevictable LRU I... |
569 570 571 572 573 574 575 576 577 |
/** * 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 Unevictable LRU I... |
578 579 580 581 |
void putback_lru_page(struct page *page) { int lru; int active = !!TestClearPageActive(page); |
bbfd28eee unevictable lru: ... |
582 |
int was_unevictable = PageUnevictable(page); |
894bc3104 Unevictable LRU I... |
583 584 585 586 587 588 589 590 591 592 593 594 595 |
VM_BUG_ON(PageLRU(page)); redo: ClearPageUnevictable(page); if (page_evictable(page, NULL)) { /* * 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. */ |
401a8e1c1 mm: introduce pag... |
596 |
lru = active + page_lru_base_type(page); |
894bc3104 Unevictable LRU I... |
597 598 599 600 601 602 603 604 |
lru_cache_add_lru(page, lru); } else { /* * Put unevictable pages directly on zone's unevictable * list. */ lru = LRU_UNEVICTABLE; add_page_to_unevictable_list(page); |
6a7b95481 vmscan: order evi... |
605 |
/* |
21ee9f398 vmscan: add barri... |
606 607 608 609 610 |
* 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 * isolation/check_move_unevictable_page, * we see PG_mlocked/AS_UNEVICTABLE cleared below and move |
6a7b95481 vmscan: order evi... |
611 612 |
* the page back to the evictable list. * |
21ee9f398 vmscan: add barri... |
613 |
* The other side is TestClearPageMlocked() or shmem_lock(). |
6a7b95481 vmscan: order evi... |
614 615 |
*/ smp_mb(); |
894bc3104 Unevictable LRU I... |
616 |
} |
894bc3104 Unevictable LRU I... |
617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 |
/* * 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. */ if (lru == LRU_UNEVICTABLE && page_evictable(page, NULL)) { 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. */ } |
bbfd28eee unevictable lru: ... |
633 634 635 636 |
if (was_unevictable && lru != LRU_UNEVICTABLE) count_vm_event(UNEVICTABLE_PGRESCUED); else if (!was_unevictable && lru == LRU_UNEVICTABLE) count_vm_event(UNEVICTABLE_PGCULLED); |
894bc3104 Unevictable LRU I... |
637 638 |
put_page(page); /* drop ref from isolate */ } |
dfc8d636c vmscan: factor ou... |
639 640 641 |
enum page_references { PAGEREF_RECLAIM, PAGEREF_RECLAIM_CLEAN, |
645747462 vmscan: detect ma... |
642 |
PAGEREF_KEEP, |
dfc8d636c vmscan: factor ou... |
643 644 645 646 647 648 |
PAGEREF_ACTIVATE, }; static enum page_references page_check_references(struct page *page, struct scan_control *sc) { |
645747462 vmscan: detect ma... |
649 |
int referenced_ptes, referenced_page; |
dfc8d636c vmscan: factor ou... |
650 |
unsigned long vm_flags; |
dfc8d636c vmscan: factor ou... |
651 |
|
645747462 vmscan: detect ma... |
652 653 |
referenced_ptes = page_referenced(page, 1, sc->mem_cgroup, &vm_flags); referenced_page = TestClearPageReferenced(page); |
dfc8d636c vmscan: factor ou... |
654 655 |
/* Lumpy reclaim - ignore references */ |
f3a310bc4 mm: vmscan: renam... |
656 |
if (sc->reclaim_mode & RECLAIM_MODE_LUMPYRECLAIM) |
dfc8d636c vmscan: factor ou... |
657 658 659 660 661 662 663 664 |
return PAGEREF_RECLAIM; /* * 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 vmscan: detect ma... |
665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 |
if (referenced_ptes) { if (PageAnon(page)) 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 vmscan: promote s... |
683 |
if (referenced_page || referenced_ptes > 1) |
645747462 vmscan: detect ma... |
684 |
return PAGEREF_ACTIVATE; |
c909e9936 vmscan: activate ... |
685 686 687 688 689 |
/* * Activate file-backed executable pages after first usage. */ if (vm_flags & VM_EXEC) return PAGEREF_ACTIVATE; |
645747462 vmscan: detect ma... |
690 691 |
return PAGEREF_KEEP; } |
dfc8d636c vmscan: factor ou... |
692 693 |
/* Reclaim if clean, defer dirty pages to writeback */ |
2e30244a7 vmscan,tmpfs: tre... |
694 |
if (referenced_page && !PageSwapBacked(page)) |
645747462 vmscan: detect ma... |
695 696 697 |
return PAGEREF_RECLAIM_CLEAN; return PAGEREF_RECLAIM; |
dfc8d636c vmscan: factor ou... |
698 |
} |
e286781d5 mm: speculative p... |
699 |
/* |
1742f19fa [PATCH] vmscan: r... |
700 |
* shrink_page_list() returns the number of reclaimed pages |
1da177e4c Linux-2.6.12-rc2 |
701 |
*/ |
1742f19fa [PATCH] vmscan: r... |
702 |
static unsigned long shrink_page_list(struct list_head *page_list, |
0e093d997 writeback: do not... |
703 |
struct zone *zone, |
f84f6e2b0 mm: vmscan: do no... |
704 |
struct scan_control *sc, |
92df3a723 mm: vmscan: throt... |
705 706 707 |
int priority, unsigned long *ret_nr_dirty, unsigned long *ret_nr_writeback) |
1da177e4c Linux-2.6.12-rc2 |
708 709 |
{ LIST_HEAD(ret_pages); |
abe4c3b50 vmscan: set up pa... |
710 |
LIST_HEAD(free_pages); |
1da177e4c Linux-2.6.12-rc2 |
711 |
int pgactivate = 0; |
0e093d997 writeback: do not... |
712 713 |
unsigned long nr_dirty = 0; unsigned long nr_congested = 0; |
05ff51376 [PATCH] vmscan re... |
714 |
unsigned long nr_reclaimed = 0; |
92df3a723 mm: vmscan: throt... |
715 |
unsigned long nr_writeback = 0; |
1da177e4c Linux-2.6.12-rc2 |
716 717 |
cond_resched(); |
1da177e4c Linux-2.6.12-rc2 |
718 |
while (!list_empty(page_list)) { |
dfc8d636c vmscan: factor ou... |
719 |
enum page_references references; |
1da177e4c Linux-2.6.12-rc2 |
720 721 722 |
struct address_space *mapping; struct page *page; int may_enter_fs; |
1da177e4c Linux-2.6.12-rc2 |
723 724 725 726 727 |
cond_resched(); page = lru_to_page(page_list); list_del(&page->lru); |
529ae9aaa mm: rename page t... |
728 |
if (!trylock_page(page)) |
1da177e4c Linux-2.6.12-rc2 |
729 |
goto keep; |
725d704ec [PATCH] mm: VM_BU... |
730 |
VM_BUG_ON(PageActive(page)); |
0e093d997 writeback: do not... |
731 |
VM_BUG_ON(page_zone(page) != zone); |
1da177e4c Linux-2.6.12-rc2 |
732 733 |
sc->nr_scanned++; |
80e434260 [PATCH] zone recl... |
734 |
|
b291f0003 mlock: mlocked pa... |
735 736 |
if (unlikely(!page_evictable(page, NULL))) goto cull_mlocked; |
894bc3104 Unevictable LRU I... |
737 |
|
a6dc60f89 vmscan: rename sc... |
738 |
if (!sc->may_unmap && page_mapped(page)) |
80e434260 [PATCH] zone recl... |
739 |
goto keep_locked; |
1da177e4c Linux-2.6.12-rc2 |
740 741 742 |
/* Double the slab pressure for mapped and swapcache pages */ if (page_mapped(page) || PageSwapCache(page)) sc->nr_scanned++; |
c661b078f synchronous lumpy... |
743 744 745 746 |
may_enter_fs = (sc->gfp_mask & __GFP_FS) || (PageSwapCache(page) && (sc->gfp_mask & __GFP_IO)); if (PageWriteback(page)) { |
92df3a723 mm: vmscan: throt... |
747 |
nr_writeback++; |
c661b078f synchronous lumpy... |
748 |
/* |
a18bba061 mm: vmscan: remov... |
749 750 751 752 |
* Synchronous reclaim cannot queue pages for * writeback due to the possibility of stack overflow * but if it encounters a page under writeback, wait * for the IO to complete. |
c661b078f synchronous lumpy... |
753 |
*/ |
f3a310bc4 mm: vmscan: renam... |
754 |
if ((sc->reclaim_mode & RECLAIM_MODE_SYNC) && |
7d3579e8e vmscan: narrow th... |
755 |
may_enter_fs) |
c661b078f synchronous lumpy... |
756 |
wait_on_page_writeback(page); |
7d3579e8e vmscan: narrow th... |
757 758 759 760 |
else { unlock_page(page); goto keep_lumpy; } |
c661b078f synchronous lumpy... |
761 |
} |
1da177e4c Linux-2.6.12-rc2 |
762 |
|
dfc8d636c vmscan: factor ou... |
763 764 765 |
references = page_check_references(page, sc); switch (references) { case PAGEREF_ACTIVATE: |
1da177e4c Linux-2.6.12-rc2 |
766 |
goto activate_locked; |
645747462 vmscan: detect ma... |
767 768 |
case PAGEREF_KEEP: goto keep_locked; |
dfc8d636c vmscan: factor ou... |
769 770 771 772 |
case PAGEREF_RECLAIM: case PAGEREF_RECLAIM_CLEAN: ; /* try to reclaim the page below */ } |
1da177e4c Linux-2.6.12-rc2 |
773 |
|
1da177e4c Linux-2.6.12-rc2 |
774 775 776 777 |
/* * Anonymous process memory has backing store? * Try to allocate it some swap space here. */ |
b291f0003 mlock: mlocked pa... |
778 |
if (PageAnon(page) && !PageSwapCache(page)) { |
63eb6b93c vmscan: let GFP_N... |
779 780 |
if (!(sc->gfp_mask & __GFP_IO)) goto keep_locked; |
ac47b003d mm: remove gfp_ma... |
781 |
if (!add_to_swap(page)) |
1da177e4c Linux-2.6.12-rc2 |
782 |
goto activate_locked; |
63eb6b93c vmscan: let GFP_N... |
783 |
may_enter_fs = 1; |
b291f0003 mlock: mlocked pa... |
784 |
} |
1da177e4c Linux-2.6.12-rc2 |
785 786 |
mapping = page_mapping(page); |
1da177e4c Linux-2.6.12-rc2 |
787 788 789 790 791 792 |
/* * The page is mapped into the page tables of one or more * processes. Try to unmap it here. */ if (page_mapped(page) && mapping) { |
14fa31b89 HWPOISON: Use bit... |
793 |
switch (try_to_unmap(page, TTU_UNMAP)) { |
1da177e4c Linux-2.6.12-rc2 |
794 795 796 797 |
case SWAP_FAIL: goto activate_locked; case SWAP_AGAIN: goto keep_locked; |
b291f0003 mlock: mlocked pa... |
798 799 |
case SWAP_MLOCK: goto cull_mlocked; |
1da177e4c Linux-2.6.12-rc2 |
800 801 802 803 804 805 |
case SWAP_SUCCESS: ; /* try to free the page below */ } } if (PageDirty(page)) { |
0e093d997 writeback: do not... |
806 |
nr_dirty++; |
ee72886d8 mm: vmscan: do no... |
807 808 |
/* * Only kswapd can writeback filesystem pages to |
f84f6e2b0 mm: vmscan: do no... |
809 810 |
* avoid risk of stack overflow but do not writeback * unless under significant pressure. |
ee72886d8 mm: vmscan: do no... |
811 |
*/ |
f84f6e2b0 mm: vmscan: do no... |
812 813 |
if (page_is_file_cache(page) && (!current_is_kswapd() || priority >= DEF_PRIORITY - 2)) { |
49ea7eb65 mm: vmscan: immed... |
814 815 816 817 818 819 820 821 |
/* * 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 mm: vmscan: do no... |
822 823 |
goto keep_locked; } |
dfc8d636c vmscan: factor ou... |
824 |
if (references == PAGEREF_RECLAIM_CLEAN) |
1da177e4c Linux-2.6.12-rc2 |
825 |
goto keep_locked; |
4dd4b9202 revert "kswapd sh... |
826 |
if (!may_enter_fs) |
1da177e4c Linux-2.6.12-rc2 |
827 |
goto keep_locked; |
52a8363ea [PATCH] mm: impro... |
828 |
if (!sc->may_writepage) |
1da177e4c Linux-2.6.12-rc2 |
829 830 831 |
goto keep_locked; /* Page is dirty, try to write it out here */ |
7d3579e8e vmscan: narrow th... |
832 |
switch (pageout(page, mapping, sc)) { |
1da177e4c Linux-2.6.12-rc2 |
833 |
case PAGE_KEEP: |
0e093d997 writeback: do not... |
834 |
nr_congested++; |
1da177e4c Linux-2.6.12-rc2 |
835 836 837 838 |
goto keep_locked; case PAGE_ACTIVATE: goto activate_locked; case PAGE_SUCCESS: |
7d3579e8e vmscan: narrow th... |
839 840 841 |
if (PageWriteback(page)) goto keep_lumpy; if (PageDirty(page)) |
1da177e4c Linux-2.6.12-rc2 |
842 |
goto keep; |
7d3579e8e vmscan: narrow th... |
843 |
|
1da177e4c Linux-2.6.12-rc2 |
844 845 846 847 |
/* * A synchronous write - probably a ramdisk. Go * ahead and try to reclaim the page. */ |
529ae9aaa mm: rename page t... |
848 |
if (!trylock_page(page)) |
1da177e4c Linux-2.6.12-rc2 |
849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 |
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 Unevictable LRU I... |
868 |
* will do this, as well as the blockdev mapping. |
1da177e4c Linux-2.6.12-rc2 |
869 870 871 872 873 874 875 876 877 878 |
* 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 FS-Cache: Recruit... |
879 |
if (page_has_private(page)) { |
1da177e4c Linux-2.6.12-rc2 |
880 881 |
if (!try_to_release_page(page, sc->gfp_mask)) goto activate_locked; |
e286781d5 mm: speculative p... |
882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 |
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 Linux-2.6.12-rc2 |
898 |
} |
e286781d5 mm: speculative p... |
899 |
if (!mapping || !__remove_mapping(mapping, page)) |
49d2e9cc4 [PATCH] Swap Migr... |
900 |
goto keep_locked; |
1da177e4c Linux-2.6.12-rc2 |
901 |
|
a978d6f52 mm: unlockless re... |
902 903 904 905 906 907 908 909 |
/* * 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 mm: speculative p... |
910 |
free_it: |
05ff51376 [PATCH] vmscan re... |
911 |
nr_reclaimed++; |
abe4c3b50 vmscan: set up pa... |
912 913 914 915 916 917 |
/* * 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 Linux-2.6.12-rc2 |
918 |
continue; |
b291f0003 mlock: mlocked pa... |
919 |
cull_mlocked: |
63d6c5ad7 mm: remove try_to... |
920 921 |
if (PageSwapCache(page)) try_to_free_swap(page); |
b291f0003 mlock: mlocked pa... |
922 923 |
unlock_page(page); putback_lru_page(page); |
f3a310bc4 mm: vmscan: renam... |
924 |
reset_reclaim_mode(sc); |
b291f0003 mlock: mlocked pa... |
925 |
continue; |
1da177e4c Linux-2.6.12-rc2 |
926 |
activate_locked: |
68a22394c vmscan: free swap... |
927 928 |
/* Not a candidate for swapping, so reclaim swap space. */ if (PageSwapCache(page) && vm_swap_full()) |
a2c43eed8 mm: try_to_free_s... |
929 |
try_to_free_swap(page); |
894bc3104 Unevictable LRU I... |
930 |
VM_BUG_ON(PageActive(page)); |
1da177e4c Linux-2.6.12-rc2 |
931 932 933 934 935 |
SetPageActive(page); pgactivate++; keep_locked: unlock_page(page); keep: |
f3a310bc4 mm: vmscan: renam... |
936 |
reset_reclaim_mode(sc); |
7d3579e8e vmscan: narrow th... |
937 |
keep_lumpy: |
1da177e4c Linux-2.6.12-rc2 |
938 |
list_add(&page->lru, &ret_pages); |
b291f0003 mlock: mlocked pa... |
939 |
VM_BUG_ON(PageLRU(page) || PageUnevictable(page)); |
1da177e4c Linux-2.6.12-rc2 |
940 |
} |
abe4c3b50 vmscan: set up pa... |
941 |
|
0e093d997 writeback: do not... |
942 943 944 945 946 947 |
/* * Tag a zone as congested if all the dirty pages encountered were * backed by a congested BDI. In this case, reclaimers should just * back off and wait for congestion to clear because further reclaim * will encounter the same problem */ |
d6c438b6c memcg: fix zone c... |
948 |
if (nr_dirty && nr_dirty == nr_congested && scanning_global_lru(sc)) |
0e093d997 writeback: do not... |
949 |
zone_set_flag(zone, ZONE_CONGESTED); |
cc59850ef mm: add free_hot_... |
950 |
free_hot_cold_page_list(&free_pages, 1); |
abe4c3b50 vmscan: set up pa... |
951 |
|
1da177e4c Linux-2.6.12-rc2 |
952 |
list_splice(&ret_pages, page_list); |
f8891e5e1 [PATCH] Light wei... |
953 |
count_vm_events(PGACTIVATE, pgactivate); |
92df3a723 mm: vmscan: throt... |
954 955 |
*ret_nr_dirty += nr_dirty; *ret_nr_writeback += nr_writeback; |
05ff51376 [PATCH] vmscan re... |
956 |
return nr_reclaimed; |
1da177e4c Linux-2.6.12-rc2 |
957 |
} |
5ad333eb6 Lumpy Reclaim V4 |
958 959 960 961 962 963 964 965 966 967 |
/* * 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. */ |
4356f21d0 mm: change isolat... |
968 |
int __isolate_lru_page(struct page *page, isolate_mode_t mode, int file) |
5ad333eb6 Lumpy Reclaim V4 |
969 |
{ |
4356f21d0 mm: change isolat... |
970 |
bool all_lru_mode; |
5ad333eb6 Lumpy Reclaim V4 |
971 972 973 974 975 |
int ret = -EINVAL; /* Only take pages on the LRU. */ if (!PageLRU(page)) return ret; |
4356f21d0 mm: change isolat... |
976 977 |
all_lru_mode = (mode & (ISOLATE_ACTIVE|ISOLATE_INACTIVE)) == (ISOLATE_ACTIVE|ISOLATE_INACTIVE); |
5ad333eb6 Lumpy Reclaim V4 |
978 979 980 981 982 |
/* * When checking the active state, we need to be sure we are * dealing with comparible boolean values. Take the logical not * of each. */ |
4356f21d0 mm: change isolat... |
983 |
if (!all_lru_mode && !PageActive(page) != !(mode & ISOLATE_ACTIVE)) |
5ad333eb6 Lumpy Reclaim V4 |
984 |
return ret; |
4356f21d0 mm: change isolat... |
985 |
if (!all_lru_mode && !!page_is_file_cache(page) != file) |
4f98a2fee vmscan: split LRU... |
986 |
return ret; |
894bc3104 Unevictable LRU I... |
987 988 989 990 991 992 993 |
/* * When this function is being called for lumpy reclaim, we * initially look into all LRU pages, active, inactive and * unevictable; only give shrink_page_list evictable pages. */ if (PageUnevictable(page)) return ret; |
5ad333eb6 Lumpy Reclaim V4 |
994 |
ret = -EBUSY; |
08e552c69 memcg: synchroniz... |
995 |
|
39deaf858 mm: compaction: m... |
996 997 |
if ((mode & ISOLATE_CLEAN) && (PageDirty(page) || PageWriteback(page))) return ret; |
f80c06736 mm: zone_reclaim:... |
998 999 |
if ((mode & ISOLATE_UNMAPPED) && page_mapped(page)) return ret; |
5ad333eb6 Lumpy Reclaim V4 |
1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 |
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 [PATCH] Swap Migr... |
1012 |
/* |
1da177e4c Linux-2.6.12-rc2 |
1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 |
* 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. * @src: The LRU list to pull pages off. * @dst: The temp list to put pages on to. * @scanned: The number of pages that were scanned. |
5ad333eb6 Lumpy Reclaim V4 |
1026 1027 |
* @order: The caller's attempted allocation order * @mode: One of the LRU isolation modes |
4f98a2fee vmscan: split LRU... |
1028 |
* @file: True [1] if isolating file [!anon] pages |
1da177e4c Linux-2.6.12-rc2 |
1029 1030 1031 |
* * returns how many pages were moved onto *@dst. */ |
69e05944a [PATCH] vmscan: u... |
1032 1033 |
static unsigned long isolate_lru_pages(unsigned long nr_to_scan, struct list_head *src, struct list_head *dst, |
4356f21d0 mm: change isolat... |
1034 1035 |
unsigned long *scanned, int order, isolate_mode_t mode, int file) |
1da177e4c Linux-2.6.12-rc2 |
1036 |
{ |
69e05944a [PATCH] vmscan: u... |
1037 |
unsigned long nr_taken = 0; |
a8a94d151 vmscan: tracing: ... |
1038 1039 1040 |
unsigned long nr_lumpy_taken = 0; unsigned long nr_lumpy_dirty = 0; unsigned long nr_lumpy_failed = 0; |
c9b02d970 [PATCH] mm: isola... |
1041 |
unsigned long scan; |
1da177e4c Linux-2.6.12-rc2 |
1042 |
|
c9b02d970 [PATCH] mm: isola... |
1043 |
for (scan = 0; scan < nr_to_scan && !list_empty(src); scan++) { |
5ad333eb6 Lumpy Reclaim V4 |
1044 1045 1046 1047 1048 |
struct page *page; unsigned long pfn; unsigned long end_pfn; unsigned long page_pfn; int zone_id; |
1da177e4c Linux-2.6.12-rc2 |
1049 1050 |
page = lru_to_page(src); prefetchw_prev_lru_page(page, src, flags); |
725d704ec [PATCH] mm: VM_BU... |
1051 |
VM_BUG_ON(!PageLRU(page)); |
8d438f96d [PATCH] mm: PageL... |
1052 |
|
4f98a2fee vmscan: split LRU... |
1053 |
switch (__isolate_lru_page(page, mode, file)) { |
5ad333eb6 Lumpy Reclaim V4 |
1054 1055 |
case 0: list_move(&page->lru, dst); |
2ffebca6a memcg: fix lru ro... |
1056 |
mem_cgroup_del_lru(page); |
2c888cfbc thp: fix anon mem... |
1057 |
nr_taken += hpage_nr_pages(page); |
5ad333eb6 Lumpy Reclaim V4 |
1058 1059 1060 1061 1062 |
break; case -EBUSY: /* else it is being freed elsewhere */ list_move(&page->lru, src); |
2ffebca6a memcg: fix lru ro... |
1063 |
mem_cgroup_rotate_lru_list(page, page_lru(page)); |
5ad333eb6 Lumpy Reclaim V4 |
1064 |
continue; |
46453a6e1 [PATCH] mm: never... |
1065 |
|
5ad333eb6 Lumpy Reclaim V4 |
1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 |
default: BUG(); } if (!order) continue; /* * Attempt to take all pages in the order aligned region * surrounding the tag page. Only take those pages of * the same active state as that tag page. We may safely * round the target page pfn down to the requested order |
25985edce Fix common misspe... |
1078 |
* as the mem_map is guaranteed valid out to MAX_ORDER, |
5ad333eb6 Lumpy Reclaim V4 |
1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 |
* where that page is in a different zone we will detect * it from its zone id and abort this block scan. */ zone_id = page_zone_id(page); page_pfn = page_to_pfn(page); pfn = page_pfn & ~((1 << order) - 1); end_pfn = pfn + (1 << order); for (; pfn < end_pfn; pfn++) { struct page *cursor_page; /* The target page is in the block, ignore it. */ if (unlikely(pfn == page_pfn)) continue; /* Avoid holes within the zone. */ if (unlikely(!pfn_valid_within(pfn))) break; cursor_page = pfn_to_page(pfn); |
4f98a2fee vmscan: split LRU... |
1098 |
|
5ad333eb6 Lumpy Reclaim V4 |
1099 1100 |
/* Check that we have not crossed a zone boundary. */ if (unlikely(page_zone_id(cursor_page) != zone_id)) |
08fc468f4 vmscan: isolate_l... |
1101 |
break; |
de2e7567c vmscan: don't att... |
1102 1103 1104 1105 1106 1107 |
/* * If we don't have enough swap space, reclaiming of * anon page which don't already have a swap slot is * pointless. */ |
043bcbe5e mm: test PageSwap... |
1108 |
if (nr_swap_pages <= 0 && PageSwapBacked(cursor_page) && |
08fc468f4 vmscan: isolate_l... |
1109 1110 |
!PageSwapCache(cursor_page)) break; |
de2e7567c vmscan: don't att... |
1111 |
|
ee993b135 mm: fix lumpy rec... |
1112 |
if (__isolate_lru_page(cursor_page, mode, file) == 0) { |
5ad333eb6 Lumpy Reclaim V4 |
1113 |
list_move(&cursor_page->lru, dst); |
cb4cbcf6b mm: fix incorrect... |
1114 |
mem_cgroup_del_lru(cursor_page); |
3770490ec mm: vmscan: fix t... |
1115 |
nr_taken += hpage_nr_pages(cursor_page); |
a8a94d151 vmscan: tracing: ... |
1116 1117 1118 |
nr_lumpy_taken++; if (PageDirty(cursor_page)) nr_lumpy_dirty++; |
5ad333eb6 Lumpy Reclaim V4 |
1119 |
scan++; |
a8a94d151 vmscan: tracing: ... |
1120 |
} else { |
d179e84ba mm: vmscan: do no... |
1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 |
/* * Check if the page is freed already. * * We can't use page_count() as that * requires compound_head and we don't * have a pin on the page here. If a * page is tail, we may or may not * have isolated the head, so assume * it's not free, it'd be tricky to * track the head status without a * page pin. */ if (!PageTail(cursor_page) && !atomic_read(&cursor_page->_count)) |
08fc468f4 vmscan: isolate_l... |
1135 1136 |
continue; break; |
5ad333eb6 Lumpy Reclaim V4 |
1137 1138 |
} } |
08fc468f4 vmscan: isolate_l... |
1139 1140 1141 1142 |
/* If we break out of the loop above, lumpy reclaim failed */ if (pfn < end_pfn) nr_lumpy_failed++; |
1da177e4c Linux-2.6.12-rc2 |
1143 1144 1145 |
} *scanned = scan; |
a8a94d151 vmscan: tracing: ... |
1146 1147 1148 1149 1150 1151 |
trace_mm_vmscan_lru_isolate(order, nr_to_scan, scan, nr_taken, nr_lumpy_taken, nr_lumpy_dirty, nr_lumpy_failed, mode); |
1da177e4c Linux-2.6.12-rc2 |
1152 1153 |
return nr_taken; } |
66e1707bc Memory controller... |
1154 1155 1156 |
static unsigned long isolate_pages_global(unsigned long nr, struct list_head *dst, unsigned long *scanned, int order, |
4356f21d0 mm: change isolat... |
1157 1158 |
isolate_mode_t mode, struct zone *z, int active, int file) |
66e1707bc Memory controller... |
1159 |
{ |
4f98a2fee vmscan: split LRU... |
1160 |
int lru = LRU_BASE; |
66e1707bc Memory controller... |
1161 |
if (active) |
4f98a2fee vmscan: split LRU... |
1162 1163 1164 1165 |
lru += LRU_ACTIVE; if (file) lru += LRU_FILE; return isolate_lru_pages(nr, &z->lru[lru].list, dst, scanned, order, |
b7c46d151 mm: drop unneeded... |
1166 |
mode, file); |
66e1707bc Memory controller... |
1167 |
} |
1da177e4c Linux-2.6.12-rc2 |
1168 |
/* |
5ad333eb6 Lumpy Reclaim V4 |
1169 1170 1171 |
* clear_active_flags() is a helper for shrink_active_list(), clearing * any active bits from the pages in the list. */ |
4f98a2fee vmscan: split LRU... |
1172 1173 |
static unsigned long clear_active_flags(struct list_head *page_list, unsigned int *count) |
5ad333eb6 Lumpy Reclaim V4 |
1174 1175 |
{ int nr_active = 0; |
4f98a2fee vmscan: split LRU... |
1176 |
int lru; |
5ad333eb6 Lumpy Reclaim V4 |
1177 |
struct page *page; |
4f98a2fee vmscan: split LRU... |
1178 |
list_for_each_entry(page, page_list, lru) { |
2c888cfbc thp: fix anon mem... |
1179 |
int numpages = hpage_nr_pages(page); |
401a8e1c1 mm: introduce pag... |
1180 |
lru = page_lru_base_type(page); |
5ad333eb6 Lumpy Reclaim V4 |
1181 |
if (PageActive(page)) { |
4f98a2fee vmscan: split LRU... |
1182 |
lru += LRU_ACTIVE; |
5ad333eb6 Lumpy Reclaim V4 |
1183 |
ClearPageActive(page); |
2c888cfbc thp: fix anon mem... |
1184 |
nr_active += numpages; |
5ad333eb6 Lumpy Reclaim V4 |
1185 |
} |
1489fa14c vmscan: update is... |
1186 |
if (count) |
2c888cfbc thp: fix anon mem... |
1187 |
count[lru] += numpages; |
4f98a2fee vmscan: split LRU... |
1188 |
} |
5ad333eb6 Lumpy Reclaim V4 |
1189 1190 1191 |
return nr_active; } |
62695a84e vmscan: move isol... |
1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 |
/** * 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 Unevictable LRU I... |
1203 1204 1205 |
* 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 vmscan: move isol... |
1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 |
* * 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; |
0c917313a mm: strictly requ... |
1220 |
VM_BUG_ON(!page_count(page)); |
62695a84e vmscan: move isol... |
1221 1222 1223 1224 |
if (PageLRU(page)) { struct zone *zone = page_zone(page); spin_lock_irq(&zone->lru_lock); |
0c917313a mm: strictly requ... |
1225 |
if (PageLRU(page)) { |
894bc3104 Unevictable LRU I... |
1226 |
int lru = page_lru(page); |
62695a84e vmscan: move isol... |
1227 |
ret = 0; |
0c917313a mm: strictly requ... |
1228 |
get_page(page); |
62695a84e vmscan: move isol... |
1229 |
ClearPageLRU(page); |
4f98a2fee vmscan: split LRU... |
1230 |
|
4f98a2fee vmscan: split LRU... |
1231 |
del_page_from_lru_list(zone, page, lru); |
62695a84e vmscan: move isol... |
1232 1233 1234 1235 1236 |
} spin_unlock_irq(&zone->lru_lock); } return ret; } |
5ad333eb6 Lumpy Reclaim V4 |
1237 |
/* |
35cd78156 vmscan: throttle ... |
1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 |
* Are there way too many processes in the direct reclaim path already? */ static int too_many_isolated(struct zone *zone, int file, struct scan_control *sc) { unsigned long inactive, isolated; if (current_is_kswapd()) return 0; if (!scanning_global_lru(sc)) 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); } return isolated > inactive; } /* |
666356297 vmscan: set up pa... |
1263 1264 1265 |
* TODO: Try merging with migrations version of putback_lru_pages */ static noinline_for_stack void |
1489fa14c vmscan: update is... |
1266 |
putback_lru_pages(struct zone *zone, struct scan_control *sc, |
666356297 vmscan: set up pa... |
1267 1268 1269 1270 1271 |
unsigned long nr_anon, unsigned long nr_file, struct list_head *page_list) { struct page *page; struct pagevec pvec; |
1489fa14c vmscan: update is... |
1272 |
struct zone_reclaim_stat *reclaim_stat = get_reclaim_stat(zone, sc); |
666356297 vmscan: set up pa... |
1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 |
pagevec_init(&pvec, 1); /* * Put back any unfreeable pages. */ spin_lock(&zone->lru_lock); while (!list_empty(page_list)) { int lru; page = lru_to_page(page_list); VM_BUG_ON(PageLRU(page)); list_del(&page->lru); if (unlikely(!page_evictable(page, NULL))) { spin_unlock_irq(&zone->lru_lock); putback_lru_page(page); spin_lock_irq(&zone->lru_lock); continue; } |
7a608572a Revert "mm: batch... |
1291 |
SetPageLRU(page); |
666356297 vmscan: set up pa... |
1292 |
lru = page_lru(page); |
7a608572a Revert "mm: batch... |
1293 |
add_page_to_lru_list(zone, page, lru); |
666356297 vmscan: set up pa... |
1294 1295 |
if (is_active_lru(lru)) { int file = is_file_lru(lru); |
9992af102 thp: scale nr_rot... |
1296 1297 |
int numpages = hpage_nr_pages(page); reclaim_stat->recent_rotated[file] += numpages; |
666356297 vmscan: set up pa... |
1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 |
} if (!pagevec_add(&pvec, page)) { spin_unlock_irq(&zone->lru_lock); __pagevec_release(&pvec); spin_lock_irq(&zone->lru_lock); } } __mod_zone_page_state(zone, NR_ISOLATED_ANON, -nr_anon); __mod_zone_page_state(zone, NR_ISOLATED_FILE, -nr_file); spin_unlock_irq(&zone->lru_lock); pagevec_release(&pvec); } |
1489fa14c vmscan: update is... |
1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 |
static noinline_for_stack void update_isolated_counts(struct zone *zone, struct scan_control *sc, unsigned long *nr_anon, unsigned long *nr_file, struct list_head *isolated_list) { unsigned long nr_active; unsigned int count[NR_LRU_LISTS] = { 0, }; struct zone_reclaim_stat *reclaim_stat = get_reclaim_stat(zone, sc); nr_active = clear_active_flags(isolated_list, count); __count_vm_events(PGDEACTIVATE, nr_active); __mod_zone_page_state(zone, NR_ACTIVE_FILE, -count[LRU_ACTIVE_FILE]); __mod_zone_page_state(zone, NR_INACTIVE_FILE, -count[LRU_INACTIVE_FILE]); __mod_zone_page_state(zone, NR_ACTIVE_ANON, -count[LRU_ACTIVE_ANON]); __mod_zone_page_state(zone, NR_INACTIVE_ANON, -count[LRU_INACTIVE_ANON]); *nr_anon = count[LRU_ACTIVE_ANON] + count[LRU_INACTIVE_ANON]; *nr_file = count[LRU_ACTIVE_FILE] + count[LRU_INACTIVE_FILE]; __mod_zone_page_state(zone, NR_ISOLATED_ANON, *nr_anon); __mod_zone_page_state(zone, NR_ISOLATED_FILE, *nr_file); reclaim_stat->recent_scanned[0] += *nr_anon; reclaim_stat->recent_scanned[1] += *nr_file; } |
666356297 vmscan: set up pa... |
1341 |
/* |
a18bba061 mm: vmscan: remov... |
1342 |
* Returns true if a direct reclaim should wait on pages under writeback. |
e31f3698c vmscan: raise the... |
1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 |
* * If we are direct reclaiming for contiguous pages and we do not reclaim * everything in the list, try again and wait for writeback IO to complete. * This will stall high-order allocations noticeably. Only do that when really * need to free the pages under high memory pressure. */ static inline bool should_reclaim_stall(unsigned long nr_taken, unsigned long nr_freed, int priority, struct scan_control *sc) { int lumpy_stall_priority; /* kswapd should not stall on sync IO */ if (current_is_kswapd()) return false; /* Only stall on lumpy reclaim */ |
f3a310bc4 mm: vmscan: renam... |
1361 |
if (sc->reclaim_mode & RECLAIM_MODE_SINGLE) |
e31f3698c vmscan: raise the... |
1362 |
return false; |
81d66c70b mm/vmscan.c: fix ... |
1363 |
/* If we have reclaimed everything on the isolated list, no stall */ |
e31f3698c vmscan: raise the... |
1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 |
if (nr_freed == nr_taken) return false; /* * For high-order allocations, there are two stall thresholds. * High-cost allocations stall immediately where as lower * order allocations such as stacks require the scanning * priority to be much higher before stalling. */ if (sc->order > PAGE_ALLOC_COSTLY_ORDER) lumpy_stall_priority = DEF_PRIORITY; else lumpy_stall_priority = DEF_PRIORITY / 3; return priority <= lumpy_stall_priority; } /* |
1742f19fa [PATCH] vmscan: r... |
1382 1383 |
* shrink_inactive_list() is a helper for shrink_zone(). It returns the number * of reclaimed pages |
1da177e4c Linux-2.6.12-rc2 |
1384 |
*/ |
666356297 vmscan: set up pa... |
1385 1386 1387 |
static noinline_for_stack unsigned long shrink_inactive_list(unsigned long nr_to_scan, struct zone *zone, struct scan_control *sc, int priority, int file) |
1da177e4c Linux-2.6.12-rc2 |
1388 1389 |
{ LIST_HEAD(page_list); |
e247dbce5 vmscan: simplify ... |
1390 |
unsigned long nr_scanned; |
05ff51376 [PATCH] vmscan re... |
1391 |
unsigned long nr_reclaimed = 0; |
e247dbce5 vmscan: simplify ... |
1392 |
unsigned long nr_taken; |
e247dbce5 vmscan: simplify ... |
1393 1394 |
unsigned long nr_anon; unsigned long nr_file; |
92df3a723 mm: vmscan: throt... |
1395 1396 |
unsigned long nr_dirty = 0; unsigned long nr_writeback = 0; |
4356f21d0 mm: change isolat... |
1397 |
isolate_mode_t reclaim_mode = ISOLATE_INACTIVE; |
78dc583d3 vmscan: low order... |
1398 |
|
35cd78156 vmscan: throttle ... |
1399 |
while (unlikely(too_many_isolated(zone, file, sc))) { |
58355c787 congestion_wait()... |
1400 |
congestion_wait(BLK_RW_ASYNC, HZ/10); |
35cd78156 vmscan: throttle ... |
1401 1402 1403 1404 1405 |
/* We are about to die and free our memory. Return now. */ if (fatal_signal_pending(current)) return SWAP_CLUSTER_MAX; } |
f3a310bc4 mm: vmscan: renam... |
1406 |
set_reclaim_mode(priority, sc, false); |
4356f21d0 mm: change isolat... |
1407 1408 |
if (sc->reclaim_mode & RECLAIM_MODE_LUMPYRECLAIM) reclaim_mode |= ISOLATE_ACTIVE; |
1da177e4c Linux-2.6.12-rc2 |
1409 |
lru_add_drain(); |
f80c06736 mm: zone_reclaim:... |
1410 1411 1412 1413 1414 |
if (!sc->may_unmap) reclaim_mode |= ISOLATE_UNMAPPED; if (!sc->may_writepage) reclaim_mode |= ISOLATE_CLEAN; |
1da177e4c Linux-2.6.12-rc2 |
1415 |
spin_lock_irq(&zone->lru_lock); |
b35ea17b7 mm: shrink_inacti... |
1416 |
|
e247dbce5 vmscan: simplify ... |
1417 |
if (scanning_global_lru(sc)) { |
4356f21d0 mm: change isolat... |
1418 1419 |
nr_taken = isolate_pages_global(nr_to_scan, &page_list, &nr_scanned, sc->order, reclaim_mode, zone, 0, file); |
e247dbce5 vmscan: simplify ... |
1420 1421 1422 1423 1424 1425 1426 1427 |
zone->pages_scanned += nr_scanned; if (current_is_kswapd()) __count_zone_vm_events(PGSCAN_KSWAPD, zone, nr_scanned); else __count_zone_vm_events(PGSCAN_DIRECT, zone, nr_scanned); } else { |
4356f21d0 mm: change isolat... |
1428 1429 1430 |
nr_taken = mem_cgroup_isolate_pages(nr_to_scan, &page_list, &nr_scanned, sc->order, reclaim_mode, zone, sc->mem_cgroup, 0, file); |
e247dbce5 vmscan: simplify ... |
1431 1432 1433 1434 1435 |
/* * mem_cgroup_isolate_pages() keeps track of * scanned pages on its own. */ } |
b35ea17b7 mm: shrink_inacti... |
1436 |
|
666356297 vmscan: set up pa... |
1437 1438 1439 1440 |
if (nr_taken == 0) { spin_unlock_irq(&zone->lru_lock); return 0; } |
5ad333eb6 Lumpy Reclaim V4 |
1441 |
|
1489fa14c vmscan: update is... |
1442 |
update_isolated_counts(zone, sc, &nr_anon, &nr_file, &page_list); |
1da177e4c Linux-2.6.12-rc2 |
1443 |
|
e247dbce5 vmscan: simplify ... |
1444 |
spin_unlock_irq(&zone->lru_lock); |
c661b078f synchronous lumpy... |
1445 |
|
92df3a723 mm: vmscan: throt... |
1446 1447 |
nr_reclaimed = shrink_page_list(&page_list, zone, sc, priority, &nr_dirty, &nr_writeback); |
c661b078f synchronous lumpy... |
1448 |
|
e31f3698c vmscan: raise the... |
1449 1450 |
/* Check if we should syncronously wait for writeback */ if (should_reclaim_stall(nr_taken, nr_reclaimed, priority, sc)) { |
f3a310bc4 mm: vmscan: renam... |
1451 |
set_reclaim_mode(priority, sc, true); |
92df3a723 mm: vmscan: throt... |
1452 1453 |
nr_reclaimed += shrink_page_list(&page_list, zone, sc, priority, &nr_dirty, &nr_writeback); |
e247dbce5 vmscan: simplify ... |
1454 |
} |
b35ea17b7 mm: shrink_inacti... |
1455 |
|
e247dbce5 vmscan: simplify ... |
1456 1457 1458 1459 |
local_irq_disable(); if (current_is_kswapd()) __count_vm_events(KSWAPD_STEAL, nr_reclaimed); __count_zone_vm_events(PGSTEAL, zone, nr_reclaimed); |
a74609faf [PATCH] mm: page_... |
1460 |
|
1489fa14c vmscan: update is... |
1461 |
putback_lru_pages(zone, sc, nr_anon, nr_file, &page_list); |
e11da5b4f tracing, vmscan: ... |
1462 |
|
92df3a723 mm: vmscan: throt... |
1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 |
/* * 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. * * This scales the number of dirty pages that must be under writeback * before throttling depending on priority. It is a simple backoff * function that has the most effect in the range DEF_PRIORITY to * DEF_PRIORITY-2 which is the priority reclaim is considered to be * in trouble and reclaim is considered to be in trouble. * * DEF_PRIORITY 100% isolated pages must be PageWriteback to throttle * DEF_PRIORITY-1 50% must be PageWriteback * DEF_PRIORITY-2 25% must be PageWriteback, kswapd in trouble * ... * DEF_PRIORITY-6 For SWAP_CLUSTER_MAX isolated pages, throttle if any * isolated page is PageWriteback */ if (nr_writeback && nr_writeback >= (nr_taken >> (DEF_PRIORITY-priority))) wait_iff_congested(zone, BLK_RW_ASYNC, HZ/10); |
e11da5b4f tracing, vmscan: ... |
1488 1489 1490 1491 |
trace_mm_vmscan_lru_shrink_inactive(zone->zone_pgdat->node_id, zone_idx(zone), nr_scanned, nr_reclaimed, priority, |
f3a310bc4 mm: vmscan: renam... |
1492 |
trace_shrink_flags(file, sc->reclaim_mode)); |
05ff51376 [PATCH] vmscan re... |
1493 |
return nr_reclaimed; |
1da177e4c Linux-2.6.12-rc2 |
1494 |
} |
3bb1a852a [PATCH] vmscan: F... |
1495 |
/* |
1da177e4c Linux-2.6.12-rc2 |
1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 |
* 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 per-zone and recl... |
1512 |
|
3eb4140f0 vmscan: merge dup... |
1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 |
static void move_active_pages_to_lru(struct zone *zone, struct list_head *list, enum lru_list lru) { unsigned long pgmoved = 0; struct pagevec pvec; struct page *page; pagevec_init(&pvec, 1); while (!list_empty(list)) { page = lru_to_page(list); |
3eb4140f0 vmscan: merge dup... |
1525 1526 1527 |
VM_BUG_ON(PageLRU(page)); SetPageLRU(page); |
3eb4140f0 vmscan: merge dup... |
1528 1529 |
list_move(&page->lru, &zone->lru[lru].list); mem_cgroup_add_lru_list(page, lru); |
2c888cfbc thp: fix anon mem... |
1530 |
pgmoved += hpage_nr_pages(page); |
3eb4140f0 vmscan: merge dup... |
1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 |
if (!pagevec_add(&pvec, page) || list_empty(list)) { spin_unlock_irq(&zone->lru_lock); if (buffer_heads_over_limit) pagevec_strip(&pvec); __pagevec_release(&pvec); spin_lock_irq(&zone->lru_lock); } } __mod_zone_page_state(zone, NR_LRU_BASE + lru, pgmoved); if (!is_active_lru(lru)) __count_vm_events(PGDEACTIVATE, pgmoved); } |
1cfb419b3 per-zone and recl... |
1544 |
|
1742f19fa [PATCH] vmscan: r... |
1545 |
static void shrink_active_list(unsigned long nr_pages, struct zone *zone, |
4f98a2fee vmscan: split LRU... |
1546 |
struct scan_control *sc, int priority, int file) |
1da177e4c Linux-2.6.12-rc2 |
1547 |
{ |
44c241f16 mm: rename pgmove... |
1548 |
unsigned long nr_taken; |
69e05944a [PATCH] vmscan: u... |
1549 |
unsigned long pgscanned; |
6fe6b7e35 vmscan: report vm... |
1550 |
unsigned long vm_flags; |
1da177e4c Linux-2.6.12-rc2 |
1551 |
LIST_HEAD(l_hold); /* The pages which were snipped off */ |
8cab4754d vmscan: make mapp... |
1552 |
LIST_HEAD(l_active); |
b69408e88 vmscan: Use an in... |
1553 |
LIST_HEAD(l_inactive); |
1da177e4c Linux-2.6.12-rc2 |
1554 |
struct page *page; |
6e9015716 mm: introduce zon... |
1555 |
struct zone_reclaim_stat *reclaim_stat = get_reclaim_stat(zone, sc); |
44c241f16 mm: rename pgmove... |
1556 |
unsigned long nr_rotated = 0; |
f80c06736 mm: zone_reclaim:... |
1557 |
isolate_mode_t reclaim_mode = ISOLATE_ACTIVE; |
1da177e4c Linux-2.6.12-rc2 |
1558 1559 |
lru_add_drain(); |
f80c06736 mm: zone_reclaim:... |
1560 1561 1562 1563 1564 |
if (!sc->may_unmap) reclaim_mode |= ISOLATE_UNMAPPED; if (!sc->may_writepage) reclaim_mode |= ISOLATE_CLEAN; |
1da177e4c Linux-2.6.12-rc2 |
1565 |
spin_lock_irq(&zone->lru_lock); |
e72e2bd67 memcg: rename sca... |
1566 |
if (scanning_global_lru(sc)) { |
8b25c6d22 vmscan: remove is... |
1567 1568 |
nr_taken = isolate_pages_global(nr_pages, &l_hold, &pgscanned, sc->order, |
f80c06736 mm: zone_reclaim:... |
1569 |
reclaim_mode, zone, |
8b25c6d22 vmscan: remove is... |
1570 |
1, file); |
1cfb419b3 per-zone and recl... |
1571 |
zone->pages_scanned += pgscanned; |
8b25c6d22 vmscan: remove is... |
1572 1573 1574 |
} else { nr_taken = mem_cgroup_isolate_pages(nr_pages, &l_hold, &pgscanned, sc->order, |
f80c06736 mm: zone_reclaim:... |
1575 |
reclaim_mode, zone, |
8b25c6d22 vmscan: remove is... |
1576 1577 1578 1579 1580 |
sc->mem_cgroup, 1, file); /* * mem_cgroup_isolate_pages() keeps track of * scanned pages on its own. */ |
4f98a2fee vmscan: split LRU... |
1581 |
} |
8b25c6d22 vmscan: remove is... |
1582 |
|
b7c46d151 mm: drop unneeded... |
1583 |
reclaim_stat->recent_scanned[file] += nr_taken; |
1cfb419b3 per-zone and recl... |
1584 |
|
3eb4140f0 vmscan: merge dup... |
1585 |
__count_zone_vm_events(PGREFILL, zone, pgscanned); |
4f98a2fee vmscan: split LRU... |
1586 |
if (file) |
44c241f16 mm: rename pgmove... |
1587 |
__mod_zone_page_state(zone, NR_ACTIVE_FILE, -nr_taken); |
4f98a2fee vmscan: split LRU... |
1588 |
else |
44c241f16 mm: rename pgmove... |
1589 |
__mod_zone_page_state(zone, NR_ACTIVE_ANON, -nr_taken); |
a731286de mm: vmstat: add i... |
1590 |
__mod_zone_page_state(zone, NR_ISOLATED_ANON + file, nr_taken); |
1da177e4c Linux-2.6.12-rc2 |
1591 |
spin_unlock_irq(&zone->lru_lock); |
1da177e4c Linux-2.6.12-rc2 |
1592 1593 1594 1595 |
while (!list_empty(&l_hold)) { cond_resched(); page = lru_to_page(&l_hold); list_del(&page->lru); |
7e9cd4842 vmscan: fix pagec... |
1596 |
|
894bc3104 Unevictable LRU I... |
1597 1598 1599 1600 |
if (unlikely(!page_evictable(page, NULL))) { putback_lru_page(page); continue; } |
645747462 vmscan: detect ma... |
1601 |
if (page_referenced(page, 0, sc->mem_cgroup, &vm_flags)) { |
9992af102 thp: scale nr_rot... |
1602 |
nr_rotated += hpage_nr_pages(page); |
8cab4754d vmscan: make mapp... |
1603 1604 1605 1606 1607 1608 1609 1610 1611 |
/* * 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 vmscan: limit VM_... |
1612 |
if ((vm_flags & VM_EXEC) && page_is_file_cache(page)) { |
8cab4754d vmscan: make mapp... |
1613 1614 1615 1616 |
list_add(&page->lru, &l_active); continue; } } |
7e9cd4842 vmscan: fix pagec... |
1617 |
|
5205e56ee vmscan: move Clea... |
1618 |
ClearPageActive(page); /* we are de-activating */ |
1da177e4c Linux-2.6.12-rc2 |
1619 1620 |
list_add(&page->lru, &l_inactive); } |
b555749aa vmscan: shrink_ac... |
1621 |
/* |
8cab4754d vmscan: make mapp... |
1622 |
* Move pages back to the lru list. |
b555749aa vmscan: shrink_ac... |
1623 |
*/ |
2a1dc5097 vmscan: protect z... |
1624 |
spin_lock_irq(&zone->lru_lock); |
4f98a2fee vmscan: split LRU... |
1625 |
/* |
8cab4754d vmscan: make mapp... |
1626 1627 1628 1629 |
* 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 vmscan: fix pagec... |
1630 |
*/ |
b7c46d151 mm: drop unneeded... |
1631 |
reclaim_stat->recent_rotated[file] += nr_rotated; |
556adecba vmscan: second ch... |
1632 |
|
3eb4140f0 vmscan: merge dup... |
1633 1634 1635 1636 |
move_active_pages_to_lru(zone, &l_active, LRU_ACTIVE + file * LRU_FILE); move_active_pages_to_lru(zone, &l_inactive, LRU_BASE + file * LRU_FILE); |
a731286de mm: vmstat: add i... |
1637 |
__mod_zone_page_state(zone, NR_ISOLATED_ANON + file, -nr_taken); |
f8891e5e1 [PATCH] Light wei... |
1638 |
spin_unlock_irq(&zone->lru_lock); |
1da177e4c Linux-2.6.12-rc2 |
1639 |
} |
74e3f3c33 vmscan: prevent b... |
1640 |
#ifdef CONFIG_SWAP |
14797e236 memcg: add inacti... |
1641 |
static int inactive_anon_is_low_global(struct zone *zone) |
f89eb90e3 inactive_anon_is_... |
1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 |
{ 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 memcg: add inacti... |
1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 |
/** * inactive_anon_is_low - check if anonymous pages need to be deactivated * @zone: zone to check * @sc: scan control of this context * * Returns true if the zone does not have enough inactive anon pages, * meaning some active anon pages need to be deactivated. */ static int inactive_anon_is_low(struct zone *zone, struct scan_control *sc) { int low; |
74e3f3c33 vmscan: prevent b... |
1664 1665 1666 1667 1668 1669 |
/* * If we don't have swap space, anonymous page deactivation * is pointless. */ if (!total_swap_pages) return 0; |
e72e2bd67 memcg: rename sca... |
1670 |
if (scanning_global_lru(sc)) |
14797e236 memcg: add inacti... |
1671 1672 |
low = inactive_anon_is_low_global(zone); else |
9b272977e memcg: skip scann... |
1673 |
low = mem_cgroup_inactive_anon_is_low(sc->mem_cgroup, zone); |
14797e236 memcg: add inacti... |
1674 1675 |
return low; } |
74e3f3c33 vmscan: prevent b... |
1676 1677 1678 1679 1680 1681 1682 |
#else static inline int inactive_anon_is_low(struct zone *zone, struct scan_control *sc) { return 0; } #endif |
14797e236 memcg: add inacti... |
1683 |
|
56e49d218 vmscan: evict use... |
1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 |
static int inactive_file_is_low_global(struct zone *zone) { unsigned long active, inactive; active = zone_page_state(zone, NR_ACTIVE_FILE); inactive = zone_page_state(zone, NR_INACTIVE_FILE); return (active > inactive); } /** * inactive_file_is_low - check if file pages need to be deactivated * @zone: zone to check * @sc: scan control of this context * * 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. */ static int inactive_file_is_low(struct zone *zone, struct scan_control *sc) { int low; if (scanning_global_lru(sc)) low = inactive_file_is_low_global(zone); else |
9b272977e memcg: skip scann... |
1716 |
low = mem_cgroup_inactive_file_is_low(sc->mem_cgroup, zone); |
56e49d218 vmscan: evict use... |
1717 1718 |
return low; } |
b39415b27 vmscan: do not ev... |
1719 1720 1721 1722 1723 1724 1725 1726 |
static int inactive_list_is_low(struct zone *zone, struct scan_control *sc, int file) { if (file) return inactive_file_is_low(zone, sc); else return inactive_anon_is_low(zone, sc); } |
4f98a2fee vmscan: split LRU... |
1727 |
static unsigned long shrink_list(enum lru_list lru, unsigned long nr_to_scan, |
b69408e88 vmscan: Use an in... |
1728 1729 |
struct zone *zone, struct scan_control *sc, int priority) { |
4f98a2fee vmscan: split LRU... |
1730 |
int file = is_file_lru(lru); |
b39415b27 vmscan: do not ev... |
1731 1732 1733 |
if (is_active_lru(lru)) { if (inactive_list_is_low(zone, sc, file)) shrink_active_list(nr_to_scan, zone, sc, priority, file); |
556adecba vmscan: second ch... |
1734 1735 |
return 0; } |
33c120ed2 more aggressively... |
1736 |
return shrink_inactive_list(nr_to_scan, zone, sc, priority, file); |
4f98a2fee vmscan: split LRU... |
1737 |
} |
1f4c025b5 memcg: export mem... |
1738 1739 1740 1741 1742 1743 |
static int vmscan_swappiness(struct scan_control *sc) { if (scanning_global_lru(sc)) return vm_swappiness; return mem_cgroup_swappiness(sc->mem_cgroup); } |
4f98a2fee vmscan: split LRU... |
1744 1745 1746 1747 1748 1749 |
/* * 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. * |
76a33fc38 vmscan: prevent g... |
1750 |
* nr[0] = anon pages to scan; nr[1] = file pages to scan |
4f98a2fee vmscan: split LRU... |
1751 |
*/ |
76a33fc38 vmscan: prevent g... |
1752 1753 |
static void get_scan_count(struct zone *zone, struct scan_control *sc, unsigned long *nr, int priority) |
4f98a2fee vmscan: split LRU... |
1754 1755 1756 1757 |
{ unsigned long anon, file, free; unsigned long anon_prio, file_prio; unsigned long ap, fp; |
6e9015716 mm: introduce zon... |
1758 |
struct zone_reclaim_stat *reclaim_stat = get_reclaim_stat(zone, sc); |
76a33fc38 vmscan: prevent g... |
1759 1760 1761 |
u64 fraction[2], denominator; enum lru_list l; int noswap = 0; |
a4d3e9e76 mm: vmscan: fix f... |
1762 |
bool force_scan = false; |
246e87a93 memcg: fix get_sc... |
1763 |
|
f11c0ca50 mm: vmscan: drop ... |
1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 |
/* * 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. */ |
a4d3e9e76 mm: vmscan: fix f... |
1774 1775 |
if (scanning_global_lru(sc) && current_is_kswapd()) force_scan = true; |
a4d3e9e76 mm: vmscan: fix f... |
1776 1777 |
if (!scanning_global_lru(sc)) force_scan = true; |
76a33fc38 vmscan: prevent g... |
1778 1779 1780 1781 1782 1783 1784 1785 1786 |
/* If we have no swap space, do not bother scanning anon pages. */ if (!sc->may_swap || (nr_swap_pages <= 0)) { noswap = 1; fraction[0] = 0; fraction[1] = 1; denominator = 1; goto out; } |
4f98a2fee vmscan: split LRU... |
1787 |
|
a4d3e9e76 mm: vmscan: fix f... |
1788 1789 1790 1791 |
anon = zone_nr_lru_pages(zone, sc, LRU_ACTIVE_ANON) + zone_nr_lru_pages(zone, sc, LRU_INACTIVE_ANON); file = zone_nr_lru_pages(zone, sc, LRU_ACTIVE_FILE) + zone_nr_lru_pages(zone, sc, LRU_INACTIVE_FILE); |
e72e2bd67 memcg: rename sca... |
1792 |
if (scanning_global_lru(sc)) { |
eeee9a8cd mm: make get_scan... |
1793 1794 1795 |
free = zone_page_state(zone, NR_FREE_PAGES); /* If we have very few page cache pages, force-scan anon pages. */ |
418589663 page allocator: u... |
1796 |
if (unlikely(file + free <= high_wmark_pages(zone))) { |
76a33fc38 vmscan: prevent g... |
1797 1798 1799 1800 |
fraction[0] = 1; fraction[1] = 0; denominator = 1; goto out; |
eeee9a8cd mm: make get_scan... |
1801 |
} |
4f98a2fee vmscan: split LRU... |
1802 1803 1804 |
} /* |
58c37f6e0 vmscan: protect r... |
1805 1806 1807 |
* With swappiness at 100, anonymous and file have the same priority. * This scanning priority is essentially the inverse of IO cost. */ |
1f4c025b5 memcg: export mem... |
1808 1809 |
anon_prio = vmscan_swappiness(sc); file_prio = 200 - vmscan_swappiness(sc); |
58c37f6e0 vmscan: protect r... |
1810 1811 |
/* |
4f98a2fee vmscan: split LRU... |
1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 |
* 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] */ |
58c37f6e0 vmscan: protect r... |
1822 |
spin_lock_irq(&zone->lru_lock); |
6e9015716 mm: introduce zon... |
1823 |
if (unlikely(reclaim_stat->recent_scanned[0] > anon / 4)) { |
6e9015716 mm: introduce zon... |
1824 1825 |
reclaim_stat->recent_scanned[0] /= 2; reclaim_stat->recent_rotated[0] /= 2; |
4f98a2fee vmscan: split LRU... |
1826 |
} |
6e9015716 mm: introduce zon... |
1827 |
if (unlikely(reclaim_stat->recent_scanned[1] > file / 4)) { |
6e9015716 mm: introduce zon... |
1828 1829 |
reclaim_stat->recent_scanned[1] /= 2; reclaim_stat->recent_rotated[1] /= 2; |
4f98a2fee vmscan: split LRU... |
1830 1831 1832 |
} /* |
00d8089c5 vmscan: fix get_s... |
1833 1834 1835 |
* 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 vmscan: split LRU... |
1836 |
*/ |
6e9015716 mm: introduce zon... |
1837 1838 |
ap = (anon_prio + 1) * (reclaim_stat->recent_scanned[0] + 1); ap /= reclaim_stat->recent_rotated[0] + 1; |
4f98a2fee vmscan: split LRU... |
1839 |
|
6e9015716 mm: introduce zon... |
1840 1841 |
fp = (file_prio + 1) * (reclaim_stat->recent_scanned[1] + 1); fp /= reclaim_stat->recent_rotated[1] + 1; |
58c37f6e0 vmscan: protect r... |
1842 |
spin_unlock_irq(&zone->lru_lock); |
4f98a2fee vmscan: split LRU... |
1843 |
|
76a33fc38 vmscan: prevent g... |
1844 1845 1846 1847 1848 1849 1850 |
fraction[0] = ap; fraction[1] = fp; denominator = ap + fp + 1; out: for_each_evictable_lru(l) { int file = is_file_lru(l); unsigned long scan; |
6e08a369e vmscan: cleanup t... |
1851 |
|
76a33fc38 vmscan: prevent g... |
1852 1853 1854 |
scan = zone_nr_lru_pages(zone, sc, l); if (priority || noswap) { scan >>= priority; |
f11c0ca50 mm: vmscan: drop ... |
1855 1856 |
if (!scan && force_scan) scan = SWAP_CLUSTER_MAX; |
76a33fc38 vmscan: prevent g... |
1857 1858 |
scan = div64_u64(scan * fraction[file], denominator); } |
246e87a93 memcg: fix get_sc... |
1859 |
nr[l] = scan; |
76a33fc38 vmscan: prevent g... |
1860 |
} |
6e08a369e vmscan: cleanup t... |
1861 |
} |
4f98a2fee vmscan: split LRU... |
1862 |
|
1da177e4c Linux-2.6.12-rc2 |
1863 |
/* |
3e7d34497 mm: vmscan: recla... |
1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 |
* Reclaim/compaction depends on a number of pages being freed. To avoid * disruption to the system, a small number of order-0 pages continue to be * rotated and reclaimed in the normal fashion. However, by the time we get * back to the allocator and call try_to_compact_zone(), we ensure that * there are enough free pages for it to be likely successful */ static inline bool should_continue_reclaim(struct zone *zone, 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 */ |
f3a310bc4 mm: vmscan: renam... |
1879 |
if (!(sc->reclaim_mode & RECLAIM_MODE_COMPACTION)) |
3e7d34497 mm: vmscan: recla... |
1880 |
return false; |
2876592f2 mm: vmscan: stop ... |
1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 |
/* 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 mm: vmscan: recla... |
1903 1904 1905 1906 1907 1908 |
/* * If we have not reclaimed enough pages for compaction and the * inactive lists are large enough, continue reclaiming */ pages_for_compaction = (2UL << sc->order); |
86cfd3a45 mm/vmscan.c: cons... |
1909 1910 1911 |
inactive_lru_pages = zone_nr_lru_pages(zone, sc, LRU_INACTIVE_FILE); if (nr_swap_pages > 0) inactive_lru_pages += zone_nr_lru_pages(zone, sc, LRU_INACTIVE_ANON); |
3e7d34497 mm: vmscan: recla... |
1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 |
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 */ switch (compaction_suitable(zone, sc->order)) { case COMPACT_PARTIAL: case COMPACT_CONTINUE: return false; default: return true; } } /* |
1da177e4c Linux-2.6.12-rc2 |
1927 1928 |
* This is a basic per-zone page freer. Used by both kswapd and direct reclaim. */ |
a79311c14 vmscan: bail out ... |
1929 |
static void shrink_zone(int priority, struct zone *zone, |
05ff51376 [PATCH] vmscan re... |
1930 |
struct scan_control *sc) |
1da177e4c Linux-2.6.12-rc2 |
1931 |
{ |
b69408e88 vmscan: Use an in... |
1932 |
unsigned long nr[NR_LRU_LISTS]; |
8695949a1 [PATCH] Thin out ... |
1933 |
unsigned long nr_to_scan; |
b69408e88 vmscan: Use an in... |
1934 |
enum lru_list l; |
f0fdc5e8e vmscan: fix zone ... |
1935 |
unsigned long nr_reclaimed, nr_scanned; |
22fba3354 vmscan: separate ... |
1936 |
unsigned long nr_to_reclaim = sc->nr_to_reclaim; |
3da367c3e vmscan: add block... |
1937 |
struct blk_plug plug; |
e0f79b8f1 vmscan: don't acc... |
1938 |
|
3e7d34497 mm: vmscan: recla... |
1939 1940 |
restart: nr_reclaimed = 0; |
f0fdc5e8e vmscan: fix zone ... |
1941 |
nr_scanned = sc->nr_scanned; |
76a33fc38 vmscan: prevent g... |
1942 |
get_scan_count(zone, sc, nr, priority); |
1da177e4c Linux-2.6.12-rc2 |
1943 |
|
3da367c3e vmscan: add block... |
1944 |
blk_start_plug(&plug); |
556adecba vmscan: second ch... |
1945 1946 |
while (nr[LRU_INACTIVE_ANON] || nr[LRU_ACTIVE_FILE] || nr[LRU_INACTIVE_FILE]) { |
894bc3104 Unevictable LRU I... |
1947 |
for_each_evictable_lru(l) { |
b69408e88 vmscan: Use an in... |
1948 |
if (nr[l]) { |
ece74b2e7 vmscan: kill sc.s... |
1949 1950 |
nr_to_scan = min_t(unsigned long, nr[l], SWAP_CLUSTER_MAX); |
b69408e88 vmscan: Use an in... |
1951 |
nr[l] -= nr_to_scan; |
1da177e4c Linux-2.6.12-rc2 |
1952 |
|
01dbe5c9b vmscan: improve r... |
1953 1954 |
nr_reclaimed += shrink_list(l, nr_to_scan, zone, sc, priority); |
b69408e88 vmscan: Use an in... |
1955 |
} |
1da177e4c Linux-2.6.12-rc2 |
1956 |
} |
a79311c14 vmscan: bail out ... |
1957 1958 1959 1960 1961 1962 1963 1964 |
/* * On large memory systems, scan >> priority can become * really large. This is fine for the starting priority; * we want to put equal scanning pressure on each zone. * However, if the VM has a harder time of freeing pages, * with multiple processes reclaiming pages, the total * freeing target can get unreasonably large. */ |
338fde909 vmscan: make cons... |
1965 |
if (nr_reclaimed >= nr_to_reclaim && priority < DEF_PRIORITY) |
a79311c14 vmscan: bail out ... |
1966 |
break; |
1da177e4c Linux-2.6.12-rc2 |
1967 |
} |
3da367c3e vmscan: add block... |
1968 |
blk_finish_plug(&plug); |
3e7d34497 mm: vmscan: recla... |
1969 |
sc->nr_reclaimed += nr_reclaimed; |
01dbe5c9b vmscan: improve r... |
1970 |
|
556adecba vmscan: second ch... |
1971 1972 1973 1974 |
/* * Even if we did not try to evict anon pages at all, we want to * rebalance the anon lru active/inactive ratio. */ |
74e3f3c33 vmscan: prevent b... |
1975 |
if (inactive_anon_is_low(zone, sc)) |
556adecba vmscan: second ch... |
1976 |
shrink_active_list(SWAP_CLUSTER_MAX, zone, sc, priority, 0); |
3e7d34497 mm: vmscan: recla... |
1977 1978 1979 1980 |
/* reclaim/compaction might need reclaim to continue */ if (should_continue_reclaim(zone, nr_reclaimed, sc->nr_scanned - nr_scanned, sc)) goto restart; |
232ea4d69 [PATCH] throttle_... |
1981 |
throttle_vm_writeout(sc->gfp_mask); |
1da177e4c Linux-2.6.12-rc2 |
1982 1983 1984 1985 1986 1987 1988 |
} /* * 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 page allocator: u... |
1989 1990 |
* We reclaim from a zone even if that zone is over high_wmark_pages(zone). * Because: |
1da177e4c Linux-2.6.12-rc2 |
1991 1992 |
* a) The caller may be trying to free *extra* pages to satisfy a higher-order * allocation or |
418589663 page allocator: u... |
1993 1994 1995 |
* 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 Linux-2.6.12-rc2 |
1996 |
* |
1da177e4c Linux-2.6.12-rc2 |
1997 1998 |
* If a zone is deemed to be full of pinned pages then just give it a light * scan then give up on it. |
e0c23279c vmscan: abort rec... |
1999 2000 2001 2002 |
* * This function returns true if a zone is being reclaimed for a costly * high-order allocation and compaction is either ready to begin or deferred. * This indicates to the caller that it should retry the allocation or fail. |
1da177e4c Linux-2.6.12-rc2 |
2003 |
*/ |
e0c23279c vmscan: abort rec... |
2004 |
static bool shrink_zones(int priority, struct zonelist *zonelist, |
05ff51376 [PATCH] vmscan re... |
2005 |
struct scan_control *sc) |
1da177e4c Linux-2.6.12-rc2 |
2006 |
{ |
dd1a239f6 mm: have zonelist... |
2007 |
struct zoneref *z; |
54a6eb5c4 mm: use two zonel... |
2008 |
struct zone *zone; |
d149e3b25 memcg: add the so... |
2009 2010 |
unsigned long nr_soft_reclaimed; unsigned long nr_soft_scanned; |
e0c23279c vmscan: abort rec... |
2011 |
bool should_abort_reclaim = false; |
1cfb419b3 per-zone and recl... |
2012 |
|
d4debc66d vmscan: remove un... |
2013 2014 |
for_each_zone_zonelist_nodemask(zone, z, zonelist, gfp_zone(sc->gfp_mask), sc->nodemask) { |
f3fe65122 [PATCH] mm: add p... |
2015 |
if (!populated_zone(zone)) |
1da177e4c Linux-2.6.12-rc2 |
2016 |
continue; |
1cfb419b3 per-zone and recl... |
2017 2018 2019 2020 |
/* * Take care memory controller reclaiming has small influence * to global LRU. */ |
e72e2bd67 memcg: rename sca... |
2021 |
if (scanning_global_lru(sc)) { |
1cfb419b3 per-zone and recl... |
2022 2023 |
if (!cpuset_zone_allowed_hardwall(zone, GFP_KERNEL)) continue; |
93e4a89a8 mm: restore zone-... |
2024 |
if (zone->all_unreclaimable && priority != DEF_PRIORITY) |
1cfb419b3 per-zone and recl... |
2025 |
continue; /* Let kswapd poll it */ |
e0887c19b vmscan: limit dir... |
2026 2027 |
if (COMPACTION_BUILD) { /* |
e0c23279c vmscan: abort rec... |
2028 2029 2030 2031 2032 2033 2034 |
* 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 * noticable problem, like transparent huge page * allocations. |
e0887c19b vmscan: limit dir... |
2035 2036 2037 |
*/ if (sc->order > PAGE_ALLOC_COSTLY_ORDER && (compaction_suitable(zone, sc->order) || |
e0c23279c vmscan: abort rec... |
2038 2039 |
compaction_deferred(zone))) { should_abort_reclaim = true; |
e0887c19b vmscan: limit dir... |
2040 |
continue; |
e0c23279c vmscan: abort rec... |
2041 |
} |
e0887c19b vmscan: limit dir... |
2042 |
} |
ac34a1a3c memcg: fix direct... |
2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 |
/* * 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; /* need some check for avoid more shrink_zone() */ |
1cfb419b3 per-zone and recl... |
2056 |
} |
408d85441 [PATCH] oom: use ... |
2057 |
|
a79311c14 vmscan: bail out ... |
2058 |
shrink_zone(priority, zone, sc); |
1da177e4c Linux-2.6.12-rc2 |
2059 |
} |
e0c23279c vmscan: abort rec... |
2060 2061 |
return should_abort_reclaim; |
d1908362a vmscan: check all... |
2062 2063 2064 2065 2066 2067 |
} static bool zone_reclaimable(struct zone *zone) { return zone->pages_scanned < zone_reclaimable_pages(zone) * 6; } |
929bea7c7 vmscan: all_unrec... |
2068 |
/* All zones in zonelist are unreclaimable? */ |
d1908362a vmscan: check all... |
2069 2070 2071 2072 2073 |
static bool all_unreclaimable(struct zonelist *zonelist, struct scan_control *sc) { struct zoneref *z; struct zone *zone; |
d1908362a vmscan: check all... |
2074 2075 2076 2077 2078 2079 2080 |
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; |
929bea7c7 vmscan: all_unrec... |
2081 2082 |
if (!zone->all_unreclaimable) return false; |
d1908362a vmscan: check all... |
2083 |
} |
929bea7c7 vmscan: all_unrec... |
2084 |
return true; |
1da177e4c Linux-2.6.12-rc2 |
2085 |
} |
4f98a2fee vmscan: split LRU... |
2086 |
|
1da177e4c Linux-2.6.12-rc2 |
2087 2088 2089 2090 2091 2092 2093 2094 |
/* * 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 writeback: get ri... |
2095 2096 2097 2098 |
* 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 page allocator: s... |
2099 2100 2101 |
* * returns: 0, if no pages reclaimed * else, the number of pages reclaimed |
1da177e4c Linux-2.6.12-rc2 |
2102 |
*/ |
dac1d27bc mm: use zonelists... |
2103 |
static unsigned long do_try_to_free_pages(struct zonelist *zonelist, |
a09ed5e00 vmscan: change sh... |
2104 2105 |
struct scan_control *sc, struct shrink_control *shrink) |
1da177e4c Linux-2.6.12-rc2 |
2106 2107 |
{ int priority; |
69e05944a [PATCH] vmscan: u... |
2108 |
unsigned long total_scanned = 0; |
1da177e4c Linux-2.6.12-rc2 |
2109 |
struct reclaim_state *reclaim_state = current->reclaim_state; |
dd1a239f6 mm: have zonelist... |
2110 |
struct zoneref *z; |
54a6eb5c4 mm: use two zonel... |
2111 |
struct zone *zone; |
22fba3354 vmscan: separate ... |
2112 |
unsigned long writeback_threshold; |
1da177e4c Linux-2.6.12-rc2 |
2113 |
|
c0ff7453b cpuset,mm: fix no... |
2114 |
get_mems_allowed(); |
873b47717 per-task-delay-ac... |
2115 |
delayacct_freepages_start(); |
e72e2bd67 memcg: rename sca... |
2116 |
if (scanning_global_lru(sc)) |
1cfb419b3 per-zone and recl... |
2117 |
count_vm_event(ALLOCSTALL); |
1da177e4c Linux-2.6.12-rc2 |
2118 2119 |
for (priority = DEF_PRIORITY; priority >= 0; priority--) { |
66e1707bc Memory controller... |
2120 |
sc->nr_scanned = 0; |
f7b7fd8f3 [PATCH] temporari... |
2121 |
if (!priority) |
a433658c3 vmscan,memcg: mem... |
2122 |
disable_swap_token(sc->mem_cgroup); |
e0c23279c vmscan: abort rec... |
2123 2124 |
if (shrink_zones(priority, zonelist, sc)) break; |
66e1707bc Memory controller... |
2125 2126 2127 2128 |
/* * Don't shrink slabs when reclaiming memory from * over limit cgroups */ |
e72e2bd67 memcg: rename sca... |
2129 |
if (scanning_global_lru(sc)) { |
c6a8a8c58 vmscan: recalcula... |
2130 |
unsigned long lru_pages = 0; |
d4debc66d vmscan: remove un... |
2131 2132 |
for_each_zone_zonelist(zone, z, zonelist, gfp_zone(sc->gfp_mask)) { |
c6a8a8c58 vmscan: recalcula... |
2133 2134 2135 2136 2137 |
if (!cpuset_zone_allowed_hardwall(zone, GFP_KERNEL)) continue; lru_pages += zone_reclaimable_pages(zone); } |
1495f230f vmscan: change sh... |
2138 |
shrink_slab(shrink, sc->nr_scanned, lru_pages); |
91a45470f per-zone and recl... |
2139 |
if (reclaim_state) { |
a79311c14 vmscan: bail out ... |
2140 |
sc->nr_reclaimed += reclaim_state->reclaimed_slab; |
91a45470f per-zone and recl... |
2141 2142 |
reclaim_state->reclaimed_slab = 0; } |
1da177e4c Linux-2.6.12-rc2 |
2143 |
} |
66e1707bc Memory controller... |
2144 |
total_scanned += sc->nr_scanned; |
bb21c7ce1 vmscan: fix do_tr... |
2145 |
if (sc->nr_reclaimed >= sc->nr_to_reclaim) |
1da177e4c Linux-2.6.12-rc2 |
2146 |
goto out; |
1da177e4c Linux-2.6.12-rc2 |
2147 2148 2149 2150 2151 2152 2153 2154 |
/* * 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 vmscan: separate ... |
2155 2156 |
writeback_threshold = sc->nr_to_reclaim + sc->nr_to_reclaim / 2; if (total_scanned > writeback_threshold) { |
0e175a183 writeback: Add a ... |
2157 2158 |
wakeup_flusher_threads(laptop_mode ? 0 : total_scanned, WB_REASON_TRY_TO_FREE_PAGES); |
66e1707bc Memory controller... |
2159 |
sc->may_writepage = 1; |
1da177e4c Linux-2.6.12-rc2 |
2160 2161 2162 |
} /* Take a nap, wait for some writeback to complete */ |
7b51755c3 vmscan: kill hibe... |
2163 |
if (!sc->hibernation_mode && sc->nr_scanned && |
0e093d997 writeback: do not... |
2164 2165 2166 2167 |
priority < DEF_PRIORITY - 2) { struct zone *preferred_zone; first_zones_zonelist(zonelist, gfp_zone(sc->gfp_mask), |
f33261d75 mm: fix deferred ... |
2168 2169 |
&cpuset_current_mems_allowed, &preferred_zone); |
0e093d997 writeback: do not... |
2170 2171 |
wait_iff_congested(preferred_zone, BLK_RW_ASYNC, HZ/10); } |
1da177e4c Linux-2.6.12-rc2 |
2172 |
} |
bb21c7ce1 vmscan: fix do_tr... |
2173 |
|
1da177e4c Linux-2.6.12-rc2 |
2174 |
out: |
873b47717 per-task-delay-ac... |
2175 |
delayacct_freepages_end(); |
c0ff7453b cpuset,mm: fix no... |
2176 |
put_mems_allowed(); |
873b47717 per-task-delay-ac... |
2177 |
|
bb21c7ce1 vmscan: fix do_tr... |
2178 2179 |
if (sc->nr_reclaimed) return sc->nr_reclaimed; |
929bea7c7 vmscan: all_unrec... |
2180 2181 2182 2183 2184 2185 2186 |
/* * 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; |
bb21c7ce1 vmscan: fix do_tr... |
2187 |
/* top priority shrink_zones still had more to do? don't OOM, then */ |
d1908362a vmscan: check all... |
2188 |
if (scanning_global_lru(sc) && !all_unreclaimable(zonelist, sc)) |
bb21c7ce1 vmscan: fix do_tr... |
2189 2190 2191 |
return 1; return 0; |
1da177e4c Linux-2.6.12-rc2 |
2192 |
} |
dac1d27bc mm: use zonelists... |
2193 |
unsigned long try_to_free_pages(struct zonelist *zonelist, int order, |
327c0e968 vmscan: fix it to... |
2194 |
gfp_t gfp_mask, nodemask_t *nodemask) |
66e1707bc Memory controller... |
2195 |
{ |
33906bc5c vmscan: tracing: ... |
2196 |
unsigned long nr_reclaimed; |
66e1707bc Memory controller... |
2197 2198 2199 |
struct scan_control sc = { .gfp_mask = gfp_mask, .may_writepage = !laptop_mode, |
22fba3354 vmscan: separate ... |
2200 |
.nr_to_reclaim = SWAP_CLUSTER_MAX, |
a6dc60f89 vmscan: rename sc... |
2201 |
.may_unmap = 1, |
2e2e42598 vmscan,memcg: rei... |
2202 |
.may_swap = 1, |
66e1707bc Memory controller... |
2203 2204 |
.order = order, .mem_cgroup = NULL, |
327c0e968 vmscan: fix it to... |
2205 |
.nodemask = nodemask, |
66e1707bc Memory controller... |
2206 |
}; |
a09ed5e00 vmscan: change sh... |
2207 2208 2209 |
struct shrink_control shrink = { .gfp_mask = sc.gfp_mask, }; |
66e1707bc Memory controller... |
2210 |
|
33906bc5c vmscan: tracing: ... |
2211 2212 2213 |
trace_mm_vmscan_direct_reclaim_begin(order, sc.may_writepage, gfp_mask); |
a09ed5e00 vmscan: change sh... |
2214 |
nr_reclaimed = do_try_to_free_pages(zonelist, &sc, &shrink); |
33906bc5c vmscan: tracing: ... |
2215 2216 2217 2218 |
trace_mm_vmscan_direct_reclaim_end(nr_reclaimed); return nr_reclaimed; |
66e1707bc Memory controller... |
2219 |
} |
00f0b8259 Memory controller... |
2220 |
#ifdef CONFIG_CGROUP_MEM_RES_CTLR |
66e1707bc Memory controller... |
2221 |
|
4e4169535 memory controller... |
2222 2223 |
unsigned long mem_cgroup_shrink_node_zone(struct mem_cgroup *mem, gfp_t gfp_mask, bool noswap, |
0ae5e89c6 memcg: count the ... |
2224 2225 |
struct zone *zone, unsigned long *nr_scanned) |
4e4169535 memory controller... |
2226 2227 |
{ struct scan_control sc = { |
0ae5e89c6 memcg: count the ... |
2228 |
.nr_scanned = 0, |
b8f5c5664 memcg: sc.nr_to_r... |
2229 |
.nr_to_reclaim = SWAP_CLUSTER_MAX, |
4e4169535 memory controller... |
2230 2231 2232 |
.may_writepage = !laptop_mode, .may_unmap = 1, .may_swap = !noswap, |
4e4169535 memory controller... |
2233 2234 |
.order = 0, .mem_cgroup = mem, |
4e4169535 memory controller... |
2235 |
}; |
0ae5e89c6 memcg: count the ... |
2236 |
|
4e4169535 memory controller... |
2237 2238 |
sc.gfp_mask = (gfp_mask & GFP_RECLAIM_MASK) | (GFP_HIGHUSER_MOVABLE & ~GFP_RECLAIM_MASK); |
bdce6d9eb memcg, vmscan: ad... |
2239 2240 2241 2242 |
trace_mm_vmscan_memcg_softlimit_reclaim_begin(0, sc.may_writepage, sc.gfp_mask); |
4e4169535 memory controller... |
2243 2244 2245 2246 2247 2248 2249 2250 |
/* * 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. */ shrink_zone(0, zone, &sc); |
bdce6d9eb memcg, vmscan: ad... |
2251 2252 |
trace_mm_vmscan_memcg_softlimit_reclaim_end(sc.nr_reclaimed); |
0ae5e89c6 memcg: count the ... |
2253 |
*nr_scanned = sc.nr_scanned; |
4e4169535 memory controller... |
2254 2255 |
return sc.nr_reclaimed; } |
e1a1cd590 Memory controller... |
2256 |
unsigned long try_to_free_mem_cgroup_pages(struct mem_cgroup *mem_cont, |
a7885eb8a memcg: swappiness |
2257 |
gfp_t gfp_mask, |
185efc0f9 memcg: Revert "me... |
2258 |
bool noswap) |
66e1707bc Memory controller... |
2259 |
{ |
4e4169535 memory controller... |
2260 |
struct zonelist *zonelist; |
bdce6d9eb memcg, vmscan: ad... |
2261 |
unsigned long nr_reclaimed; |
889976dbc memcg: reclaim me... |
2262 |
int nid; |
66e1707bc Memory controller... |
2263 |
struct scan_control sc = { |
66e1707bc Memory controller... |
2264 |
.may_writepage = !laptop_mode, |
a6dc60f89 vmscan: rename sc... |
2265 |
.may_unmap = 1, |
2e2e42598 vmscan,memcg: rei... |
2266 |
.may_swap = !noswap, |
22fba3354 vmscan: separate ... |
2267 |
.nr_to_reclaim = SWAP_CLUSTER_MAX, |
66e1707bc Memory controller... |
2268 2269 |
.order = 0, .mem_cgroup = mem_cont, |
327c0e968 vmscan: fix it to... |
2270 |
.nodemask = NULL, /* we don't care the placement */ |
a09ed5e00 vmscan: change sh... |
2271 2272 2273 2274 2275 |
.gfp_mask = (gfp_mask & GFP_RECLAIM_MASK) | (GFP_HIGHUSER_MOVABLE & ~GFP_RECLAIM_MASK), }; struct shrink_control shrink = { .gfp_mask = sc.gfp_mask, |
66e1707bc Memory controller... |
2276 |
}; |
66e1707bc Memory controller... |
2277 |
|
889976dbc memcg: reclaim me... |
2278 2279 2280 2281 2282 2283 2284 2285 |
/* * 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. */ nid = mem_cgroup_select_victim_node(mem_cont); zonelist = NODE_DATA(nid)->node_zonelists; |
bdce6d9eb memcg, vmscan: ad... |
2286 2287 2288 2289 |
trace_mm_vmscan_memcg_reclaim_begin(0, sc.may_writepage, sc.gfp_mask); |
a09ed5e00 vmscan: change sh... |
2290 |
nr_reclaimed = do_try_to_free_pages(zonelist, &sc, &shrink); |
bdce6d9eb memcg, vmscan: ad... |
2291 2292 2293 2294 |
trace_mm_vmscan_memcg_reclaim_end(nr_reclaimed); return nr_reclaimed; |
66e1707bc Memory controller... |
2295 2296 |
} #endif |
1741c8775 mm: kswapd: keep ... |
2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 |
/* * pgdat_balanced is used when checking if a node is 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. * 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 Fix common misspe... |
2308 |
* percentage of the middle zones. For example, on 32-bit x86, highmem |
1741c8775 mm: kswapd: keep ... |
2309 2310 2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 |
* 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. */ static bool pgdat_balanced(pg_data_t *pgdat, unsigned long balanced_pages, int classzone_idx) { unsigned long present_pages = 0; int i; for (i = 0; i <= classzone_idx; i++) present_pages += pgdat->node_zones[i].present_pages; |
4746efded vmscan: fix a liv... |
2321 2322 |
/* A special case here: if zone has no page, we think it's balanced */ return balanced_pages >= (present_pages >> 2); |
1741c8775 mm: kswapd: keep ... |
2323 |
} |
f50de2d38 vmscan: have kswa... |
2324 |
/* is kswapd sleeping prematurely? */ |
dc83edd94 mm: kswapd: use t... |
2325 2326 |
static bool sleeping_prematurely(pg_data_t *pgdat, int order, long remaining, int classzone_idx) |
f50de2d38 vmscan: have kswa... |
2327 |
{ |
bb3ab5968 vmscan: stop kswa... |
2328 |
int i; |
1741c8775 mm: kswapd: keep ... |
2329 2330 |
unsigned long balanced = 0; bool all_zones_ok = true; |
f50de2d38 vmscan: have kswa... |
2331 2332 2333 |
/* If a direct reclaimer woke kswapd within HZ/10, it's premature */ if (remaining) |
dc83edd94 mm: kswapd: use t... |
2334 |
return true; |
f50de2d38 vmscan: have kswa... |
2335 |
|
0abdee2bd mm: kswapd: use t... |
2336 |
/* Check the watermark levels */ |
08951e545 mm: vmscan: corre... |
2337 |
for (i = 0; i <= classzone_idx; i++) { |
bb3ab5968 vmscan: stop kswa... |
2338 2339 2340 2341 |
struct zone *zone = pgdat->node_zones + i; if (!populated_zone(zone)) continue; |
355b09c47 mm: kswapd: treat... |
2342 2343 2344 2345 2346 2347 2348 2349 |
/* * balance_pgdat() skips over all_unreclaimable after * DEF_PRIORITY. Effectively, it considers them balanced so * they must be considered balanced here as well if kswapd * is to sleep */ if (zone->all_unreclaimable) { balanced += zone->present_pages; |
de3fab393 vmscan: kswapd: d... |
2350 |
continue; |
355b09c47 mm: kswapd: treat... |
2351 |
} |
de3fab393 vmscan: kswapd: d... |
2352 |
|
88f5acf88 mm: page allocato... |
2353 |
if (!zone_watermark_ok_safe(zone, order, high_wmark_pages(zone), |
da175d06b mm: vmscan: evalu... |
2354 |
i, 0)) |
1741c8775 mm: kswapd: keep ... |
2355 2356 2357 |
all_zones_ok = false; else balanced += zone->present_pages; |
bb3ab5968 vmscan: stop kswa... |
2358 |
} |
f50de2d38 vmscan: have kswa... |
2359 |
|
1741c8775 mm: kswapd: keep ... |
2360 2361 2362 2363 2364 2365 |
/* * For high-order requests, the balanced zones must contain at least * 25% of the nodes pages for kswapd to sleep. For order-0, all zones * must be balanced */ if (order) |
afc7e326a mm: vmscan: corre... |
2366 |
return !pgdat_balanced(pgdat, balanced, classzone_idx); |
1741c8775 mm: kswapd: keep ... |
2367 2368 |
else return !all_zones_ok; |
f50de2d38 vmscan: have kswa... |
2369 |
} |
1da177e4c Linux-2.6.12-rc2 |
2370 2371 |
/* * For kswapd, balance_pgdat() will work across all this node's zones until |
418589663 page allocator: u... |
2372 |
* they are all at high_wmark_pages(zone). |
1da177e4c Linux-2.6.12-rc2 |
2373 |
* |
0abdee2bd mm: kswapd: use t... |
2374 |
* Returns the final order kswapd was reclaiming at |
1da177e4c Linux-2.6.12-rc2 |
2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 |
* * 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 page allocator: u... |
2385 2386 2387 2388 2389 |
* 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 Linux-2.6.12-rc2 |
2390 |
*/ |
995047488 mm: kswapd: stop ... |
2391 |
static unsigned long balance_pgdat(pg_data_t *pgdat, int order, |
dc83edd94 mm: kswapd: use t... |
2392 |
int *classzone_idx) |
1da177e4c Linux-2.6.12-rc2 |
2393 |
{ |
1da177e4c Linux-2.6.12-rc2 |
2394 |
int all_zones_ok; |
1741c8775 mm: kswapd: keep ... |
2395 |
unsigned long balanced; |
1da177e4c Linux-2.6.12-rc2 |
2396 2397 |
int priority; int i; |
995047488 mm: kswapd: stop ... |
2398 |
int end_zone = 0; /* Inclusive. 0 = ZONE_DMA */ |
69e05944a [PATCH] vmscan: u... |
2399 |
unsigned long total_scanned; |
1da177e4c Linux-2.6.12-rc2 |
2400 |
struct reclaim_state *reclaim_state = current->reclaim_state; |
0ae5e89c6 memcg: count the ... |
2401 2402 |
unsigned long nr_soft_reclaimed; unsigned long nr_soft_scanned; |
179e96395 [PATCH] vmscan: s... |
2403 2404 |
struct scan_control sc = { .gfp_mask = GFP_KERNEL, |
a6dc60f89 vmscan: rename sc... |
2405 |
.may_unmap = 1, |
2e2e42598 vmscan,memcg: rei... |
2406 |
.may_swap = 1, |
22fba3354 vmscan: separate ... |
2407 2408 2409 2410 2411 |
/* * kswapd doesn't want to be bailed out while reclaim. because * we want to put equal scanning pressure on each zone. */ .nr_to_reclaim = ULONG_MAX, |
5ad333eb6 Lumpy Reclaim V4 |
2412 |
.order = order, |
66e1707bc Memory controller... |
2413 |
.mem_cgroup = NULL, |
179e96395 [PATCH] vmscan: s... |
2414 |
}; |
a09ed5e00 vmscan: change sh... |
2415 2416 2417 |
struct shrink_control shrink = { .gfp_mask = sc.gfp_mask, }; |
1da177e4c Linux-2.6.12-rc2 |
2418 2419 |
loop_again: total_scanned = 0; |
a79311c14 vmscan: bail out ... |
2420 |
sc.nr_reclaimed = 0; |
c0bbbc73d [PATCH] typo in v... |
2421 |
sc.may_writepage = !laptop_mode; |
f8891e5e1 [PATCH] Light wei... |
2422 |
count_vm_event(PAGEOUTRUN); |
1da177e4c Linux-2.6.12-rc2 |
2423 |
|
1da177e4c Linux-2.6.12-rc2 |
2424 |
for (priority = DEF_PRIORITY; priority >= 0; priority--) { |
1da177e4c Linux-2.6.12-rc2 |
2425 |
unsigned long lru_pages = 0; |
bb3ab5968 vmscan: stop kswa... |
2426 |
int has_under_min_watermark_zone = 0; |
1da177e4c Linux-2.6.12-rc2 |
2427 |
|
f7b7fd8f3 [PATCH] temporari... |
2428 2429 |
/* The swap token gets in the way of swapout... */ if (!priority) |
a433658c3 vmscan,memcg: mem... |
2430 |
disable_swap_token(NULL); |
f7b7fd8f3 [PATCH] temporari... |
2431 |
|
1da177e4c Linux-2.6.12-rc2 |
2432 |
all_zones_ok = 1; |
1741c8775 mm: kswapd: keep ... |
2433 |
balanced = 0; |
1da177e4c Linux-2.6.12-rc2 |
2434 |
|
d6277db4a [PATCH] swsusp: r... |
2435 2436 2437 2438 2439 2440 |
/* * 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 Linux-2.6.12-rc2 |
2441 |
|
d6277db4a [PATCH] swsusp: r... |
2442 2443 |
if (!populated_zone(zone)) continue; |
1da177e4c Linux-2.6.12-rc2 |
2444 |
|
93e4a89a8 mm: restore zone-... |
2445 |
if (zone->all_unreclaimable && priority != DEF_PRIORITY) |
d6277db4a [PATCH] swsusp: r... |
2446 |
continue; |
1da177e4c Linux-2.6.12-rc2 |
2447 |
|
556adecba vmscan: second ch... |
2448 2449 2450 2451 |
/* * Do some background aging of the anon list, to give * pages a chance to be referenced before reclaiming. */ |
14797e236 memcg: add inacti... |
2452 |
if (inactive_anon_is_low(zone, &sc)) |
556adecba vmscan: second ch... |
2453 2454 |
shrink_active_list(SWAP_CLUSTER_MAX, zone, &sc, priority, 0); |
88f5acf88 mm: page allocato... |
2455 |
if (!zone_watermark_ok_safe(zone, order, |
418589663 page allocator: u... |
2456 |
high_wmark_pages(zone), 0, 0)) { |
d6277db4a [PATCH] swsusp: r... |
2457 |
end_zone = i; |
e1dbeda60 [PATCH] balance_p... |
2458 |
break; |
439423f68 vmscan: clear ZON... |
2459 2460 2461 |
} else { /* If balanced, clear the congested flag */ zone_clear_flag(zone, ZONE_CONGESTED); |
1da177e4c Linux-2.6.12-rc2 |
2462 |
} |
1da177e4c Linux-2.6.12-rc2 |
2463 |
} |
e1dbeda60 [PATCH] balance_p... |
2464 2465 |
if (i < 0) goto out; |
1da177e4c Linux-2.6.12-rc2 |
2466 2467 |
for (i = 0; i <= end_zone; i++) { struct zone *zone = pgdat->node_zones + i; |
adea02a1b mm: count only re... |
2468 |
lru_pages += zone_reclaimable_pages(zone); |
1da177e4c Linux-2.6.12-rc2 |
2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 |
} /* * 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; |
b15e0905f [PATCH] vmscan: n... |
2482 |
int nr_slab; |
8afdcece4 mm: vmscan: kswap... |
2483 |
unsigned long balance_gap; |
1da177e4c Linux-2.6.12-rc2 |
2484 |
|
f3fe65122 [PATCH] mm: add p... |
2485 |
if (!populated_zone(zone)) |
1da177e4c Linux-2.6.12-rc2 |
2486 |
continue; |
93e4a89a8 mm: restore zone-... |
2487 |
if (zone->all_unreclaimable && priority != DEF_PRIORITY) |
1da177e4c Linux-2.6.12-rc2 |
2488 |
continue; |
1da177e4c Linux-2.6.12-rc2 |
2489 |
sc.nr_scanned = 0; |
4e4169535 memory controller... |
2490 |
|
0ae5e89c6 memcg: count the ... |
2491 |
nr_soft_scanned = 0; |
4e4169535 memory controller... |
2492 2493 |
/* * Call soft limit reclaim before calling shrink_zone. |
4e4169535 memory controller... |
2494 |
*/ |
0ae5e89c6 memcg: count the ... |
2495 2496 2497 2498 2499 |
nr_soft_reclaimed = mem_cgroup_soft_limit_reclaim(zone, order, sc.gfp_mask, &nr_soft_scanned); sc.nr_reclaimed += nr_soft_reclaimed; total_scanned += nr_soft_scanned; |
00918b6ab memcg: remove nid... |
2500 |
|
32a4330d4 mm: prevent kswap... |
2501 |
/* |
8afdcece4 mm: vmscan: kswap... |
2502 2503 2504 2505 2506 2507 |
* 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. |
32a4330d4 mm: prevent kswap... |
2508 |
*/ |
8afdcece4 mm: vmscan: kswap... |
2509 2510 2511 2512 |
balance_gap = min(low_wmark_pages(zone), (zone->present_pages + KSWAPD_ZONE_BALANCE_GAP_RATIO-1) / KSWAPD_ZONE_BALANCE_GAP_RATIO); |
88f5acf88 mm: page allocato... |
2513 |
if (!zone_watermark_ok_safe(zone, order, |
8afdcece4 mm: vmscan: kswap... |
2514 |
high_wmark_pages(zone) + balance_gap, |
d7868dae8 mm: vmscan: do no... |
2515 |
end_zone, 0)) { |
a79311c14 vmscan: bail out ... |
2516 |
shrink_zone(priority, zone, &sc); |
5a03b051e thp: use compacti... |
2517 |
|
d7868dae8 mm: vmscan: do no... |
2518 2519 2520 2521 2522 2523 2524 2525 |
reclaim_state->reclaimed_slab = 0; nr_slab = shrink_slab(&shrink, sc.nr_scanned, lru_pages); sc.nr_reclaimed += reclaim_state->reclaimed_slab; total_scanned += sc.nr_scanned; if (nr_slab == 0 && !zone_reclaimable(zone)) zone->all_unreclaimable = 1; } |
1da177e4c Linux-2.6.12-rc2 |
2526 2527 2528 2529 2530 2531 |
/* * If we've done a decent amount of scanning and * the reclaim ratio is low, start doing writepage * even in laptop mode */ if (total_scanned > SWAP_CLUSTER_MAX * 2 && |
a79311c14 vmscan: bail out ... |
2532 |
total_scanned > sc.nr_reclaimed + sc.nr_reclaimed / 2) |
1da177e4c Linux-2.6.12-rc2 |
2533 |
sc.may_writepage = 1; |
bb3ab5968 vmscan: stop kswa... |
2534 |
|
215ddd666 mm: vmscan: only ... |
2535 2536 2537 |
if (zone->all_unreclaimable) { if (end_zone && end_zone == i) end_zone--; |
d7868dae8 mm: vmscan: do no... |
2538 |
continue; |
215ddd666 mm: vmscan: only ... |
2539 |
} |
d7868dae8 mm: vmscan: do no... |
2540 |
|
88f5acf88 mm: page allocato... |
2541 |
if (!zone_watermark_ok_safe(zone, order, |
45973d74f vmscan: check hig... |
2542 2543 2544 2545 2546 2547 2548 |
high_wmark_pages(zone), end_zone, 0)) { all_zones_ok = 0; /* * We are still under min water mark. This * means that we have a GFP_ATOMIC allocation * failure risk. Hurry up! */ |
88f5acf88 mm: page allocato... |
2549 |
if (!zone_watermark_ok_safe(zone, order, |
45973d74f vmscan: check hig... |
2550 2551 |
min_wmark_pages(zone), end_zone, 0)) has_under_min_watermark_zone = 1; |
0e093d997 writeback: do not... |
2552 2553 2554 2555 2556 2557 2558 2559 2560 |
} else { /* * 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, * spectulatively avoid congestion waits */ zone_clear_flag(zone, ZONE_CONGESTED); |
dc83edd94 mm: kswapd: use t... |
2561 |
if (i <= *classzone_idx) |
1741c8775 mm: kswapd: keep ... |
2562 |
balanced += zone->present_pages; |
45973d74f vmscan: check hig... |
2563 |
} |
bb3ab5968 vmscan: stop kswa... |
2564 |
|
1da177e4c Linux-2.6.12-rc2 |
2565 |
} |
dc83edd94 mm: kswapd: use t... |
2566 |
if (all_zones_ok || (order && pgdat_balanced(pgdat, balanced, *classzone_idx))) |
1da177e4c Linux-2.6.12-rc2 |
2567 2568 2569 2570 2571 |
break; /* kswapd: all done */ /* * OK, kswapd is getting into trouble. Take a nap, then take * another pass across the zones. */ |
bb3ab5968 vmscan: stop kswa... |
2572 2573 2574 2575 2576 2577 |
if (total_scanned && (priority < DEF_PRIORITY - 2)) { if (has_under_min_watermark_zone) count_vm_event(KSWAPD_SKIP_CONGESTION_WAIT); else congestion_wait(BLK_RW_ASYNC, HZ/10); } |
1da177e4c Linux-2.6.12-rc2 |
2578 2579 2580 2581 2582 2583 2584 |
/* * We do this so kswapd doesn't build up large priorities for * example when it is freeing in parallel with allocators. It * matches the direct reclaim path behaviour in terms of impact * on zone->*_priority. */ |
a79311c14 vmscan: bail out ... |
2585 |
if (sc.nr_reclaimed >= SWAP_CLUSTER_MAX) |
1da177e4c Linux-2.6.12-rc2 |
2586 2587 2588 |
break; } out: |
995047488 mm: kswapd: stop ... |
2589 2590 2591 |
/* * order-0: All zones must meet high watermark for a balanced node |
1741c8775 mm: kswapd: keep ... |
2592 2593 |
* high-order: Balanced zones must make up at least 25% of the node * for the node to be balanced |
995047488 mm: kswapd: stop ... |
2594 |
*/ |
dc83edd94 mm: kswapd: use t... |
2595 |
if (!(all_zones_ok || (order && pgdat_balanced(pgdat, balanced, *classzone_idx)))) { |
1da177e4c Linux-2.6.12-rc2 |
2596 |
cond_resched(); |
8357376d3 [PATCH] swsusp: I... |
2597 2598 |
try_to_freeze(); |
73ce02e96 mm: stop kswapd's... |
2599 2600 2601 2602 2603 2604 2605 2606 2607 2608 2609 2610 2611 2612 2613 2614 |
/* * Fragmentation may mean that the system cannot be * rebalanced for high-order allocations in all zones. * At this point, if nr_reclaimed < SWAP_CLUSTER_MAX, * it means the zones have been fully scanned and are still * not balanced. For high-order allocations, there is * little point trying all over again as kswapd may * infinite loop. * * Instead, recheck all watermarks at order-0 as they * are the most important. If watermarks are ok, kswapd will go * back to sleep. High-order users can still perform direct * reclaim if they wish. */ if (sc.nr_reclaimed < SWAP_CLUSTER_MAX) order = sc.order = 0; |
1da177e4c Linux-2.6.12-rc2 |
2615 2616 |
goto loop_again; } |
995047488 mm: kswapd: stop ... |
2617 2618 2619 2620 2621 2622 2623 2624 2625 2626 2627 2628 2629 2630 2631 2632 2633 2634 2635 2636 2637 2638 2639 2640 2641 2642 2643 |
/* * If kswapd was reclaiming at a higher order, it has the option of * sleeping without all zones being balanced. Before it does, it must * ensure that the watermarks for order-0 on *all* zones are met and * that the congestion flags are cleared. The congestion flag must * be cleared as kswapd is the only mechanism that clears the flag * and it is potentially going to sleep here. */ if (order) { for (i = 0; i <= end_zone; i++) { struct zone *zone = pgdat->node_zones + i; if (!populated_zone(zone)) continue; if (zone->all_unreclaimable && priority != DEF_PRIORITY) continue; /* Confirm the zone is balanced for order-0 */ if (!zone_watermark_ok(zone, 0, high_wmark_pages(zone), 0, 0)) { order = sc.order = 0; goto loop_again; } /* If balanced, clear the congested flag */ zone_clear_flag(zone, ZONE_CONGESTED); |
16fb95123 vmscan: count pag... |
2644 2645 |
if (i <= *classzone_idx) balanced += zone->present_pages; |
995047488 mm: kswapd: stop ... |
2646 2647 |
} } |
0abdee2bd mm: kswapd: use t... |
2648 2649 2650 2651 2652 2653 |
/* * Return the order we were reclaiming at so sleeping_prematurely() * 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 mm: kswapd: use t... |
2654 |
*classzone_idx = end_zone; |
0abdee2bd mm: kswapd: use t... |
2655 |
return order; |
1da177e4c Linux-2.6.12-rc2 |
2656 |
} |
dc83edd94 mm: kswapd: use t... |
2657 |
static void kswapd_try_to_sleep(pg_data_t *pgdat, int order, int classzone_idx) |
f0bc0a60b vmscan: factor ou... |
2658 2659 2660 2661 2662 2663 2664 2665 2666 2667 |
{ 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 */ |
dc83edd94 mm: kswapd: use t... |
2668 |
if (!sleeping_prematurely(pgdat, order, remaining, classzone_idx)) { |
f0bc0a60b vmscan: factor ou... |
2669 2670 2671 2672 2673 2674 2675 2676 2677 |
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. */ |
dc83edd94 mm: kswapd: use t... |
2678 |
if (!sleeping_prematurely(pgdat, order, remaining, classzone_idx)) { |
f0bc0a60b vmscan: factor ou... |
2679 2680 2681 2682 2683 2684 2685 2686 2687 2688 2689 2690 2691 2692 2693 2694 2695 2696 2697 2698 2699 |
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); schedule(); 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 Linux-2.6.12-rc2 |
2700 2701 |
/* * The background pageout daemon, started as a kernel thread |
4f98a2fee vmscan: split LRU... |
2702 |
* from the init process. |
1da177e4c Linux-2.6.12-rc2 |
2703 2704 2705 2706 2707 2708 2709 2710 2711 2712 2713 2714 |
* * 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 mm: vmscan: only ... |
2715 |
unsigned long order, new_order; |
d2ebd0f6b kswapd: avoid unn... |
2716 |
unsigned balanced_order; |
215ddd666 mm: vmscan: only ... |
2717 |
int classzone_idx, new_classzone_idx; |
d2ebd0f6b kswapd: avoid unn... |
2718 |
int balanced_classzone_idx; |
1da177e4c Linux-2.6.12-rc2 |
2719 2720 |
pg_data_t *pgdat = (pg_data_t*)p; struct task_struct *tsk = current; |
f0bc0a60b vmscan: factor ou... |
2721 |
|
1da177e4c Linux-2.6.12-rc2 |
2722 2723 2724 |
struct reclaim_state reclaim_state = { .reclaimed_slab = 0, }; |
a70f73028 cpumask: replace ... |
2725 |
const struct cpumask *cpumask = cpumask_of_node(pgdat->node_id); |
1da177e4c Linux-2.6.12-rc2 |
2726 |
|
cf40bd16f lockdep: annotate... |
2727 |
lockdep_set_current_reclaim_state(GFP_KERNEL); |
174596a0b cpumask: convert mm/ |
2728 |
if (!cpumask_empty(cpumask)) |
c5f59f083 nodemask: use new... |
2729 |
set_cpus_allowed_ptr(tsk, cpumask); |
1da177e4c Linux-2.6.12-rc2 |
2730 2731 2732 2733 2734 2735 2736 2737 2738 2739 2740 2741 2742 2743 |
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 [PATCH] Swap Migr... |
2744 |
tsk->flags |= PF_MEMALLOC | PF_SWAPWRITE | PF_KSWAPD; |
831441862 Freezer: make ker... |
2745 |
set_freezable(); |
1da177e4c Linux-2.6.12-rc2 |
2746 |
|
215ddd666 mm: vmscan: only ... |
2747 |
order = new_order = 0; |
d2ebd0f6b kswapd: avoid unn... |
2748 |
balanced_order = 0; |
215ddd666 mm: vmscan: only ... |
2749 |
classzone_idx = new_classzone_idx = pgdat->nr_zones - 1; |
d2ebd0f6b kswapd: avoid unn... |
2750 |
balanced_classzone_idx = classzone_idx; |
1da177e4c Linux-2.6.12-rc2 |
2751 |
for ( ; ; ) { |
8fe23e057 mm: clear node in... |
2752 |
int ret; |
3e1d1d28d [PATCH] Cleanup p... |
2753 |
|
215ddd666 mm: vmscan: only ... |
2754 2755 2756 2757 2758 |
/* * 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 kswapd: avoid unn... |
2759 2760 |
if (balanced_classzone_idx >= new_classzone_idx && balanced_order == new_order) { |
215ddd666 mm: vmscan: only ... |
2761 2762 2763 2764 2765 |
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 mm: kswapd: stop ... |
2766 |
if (order < new_order || classzone_idx > new_classzone_idx) { |
1da177e4c Linux-2.6.12-rc2 |
2767 2768 |
/* * Don't sleep if someone wants a larger 'order' |
995047488 mm: kswapd: stop ... |
2769 |
* allocation or has tigher zone constraints |
1da177e4c Linux-2.6.12-rc2 |
2770 2771 |
*/ order = new_order; |
995047488 mm: kswapd: stop ... |
2772 |
classzone_idx = new_classzone_idx; |
1da177e4c Linux-2.6.12-rc2 |
2773 |
} else { |
d2ebd0f6b kswapd: avoid unn... |
2774 2775 |
kswapd_try_to_sleep(pgdat, balanced_order, balanced_classzone_idx); |
1da177e4c Linux-2.6.12-rc2 |
2776 |
order = pgdat->kswapd_max_order; |
995047488 mm: kswapd: stop ... |
2777 |
classzone_idx = pgdat->classzone_idx; |
f0dfcde09 kswapd: assign ne... |
2778 2779 |
new_order = order; new_classzone_idx = classzone_idx; |
4d40502ea mm: kswapd: reset... |
2780 |
pgdat->kswapd_max_order = 0; |
215ddd666 mm: vmscan: only ... |
2781 |
pgdat->classzone_idx = pgdat->nr_zones - 1; |
1da177e4c Linux-2.6.12-rc2 |
2782 |
} |
1da177e4c Linux-2.6.12-rc2 |
2783 |
|
8fe23e057 mm: clear node in... |
2784 2785 2786 2787 2788 2789 2790 2791 |
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 vmscan: tracing: ... |
2792 2793 |
if (!ret) { trace_mm_vmscan_kswapd_wake(pgdat->node_id, order); |
d2ebd0f6b kswapd: avoid unn... |
2794 2795 2796 |
balanced_classzone_idx = classzone_idx; balanced_order = balance_pgdat(pgdat, order, &balanced_classzone_idx); |
33906bc5c vmscan: tracing: ... |
2797 |
} |
1da177e4c Linux-2.6.12-rc2 |
2798 2799 2800 2801 2802 2803 2804 |
} return 0; } /* * A zone is low on free memory, so wake its kswapd task to service it. */ |
995047488 mm: kswapd: stop ... |
2805 |
void wakeup_kswapd(struct zone *zone, int order, enum zone_type classzone_idx) |
1da177e4c Linux-2.6.12-rc2 |
2806 2807 |
{ pg_data_t *pgdat; |
f3fe65122 [PATCH] mm: add p... |
2808 |
if (!populated_zone(zone)) |
1da177e4c Linux-2.6.12-rc2 |
2809 |
return; |
88f5acf88 mm: page allocato... |
2810 |
if (!cpuset_zone_allowed_hardwall(zone, GFP_KERNEL)) |
1da177e4c Linux-2.6.12-rc2 |
2811 |
return; |
88f5acf88 mm: page allocato... |
2812 |
pgdat = zone->zone_pgdat; |
995047488 mm: kswapd: stop ... |
2813 |
if (pgdat->kswapd_max_order < order) { |
1da177e4c Linux-2.6.12-rc2 |
2814 |
pgdat->kswapd_max_order = order; |
995047488 mm: kswapd: stop ... |
2815 2816 |
pgdat->classzone_idx = min(pgdat->classzone_idx, classzone_idx); } |
8d0986e28 [PATCH] vm: kswap... |
2817 |
if (!waitqueue_active(&pgdat->kswapd_wait)) |
1da177e4c Linux-2.6.12-rc2 |
2818 |
return; |
88f5acf88 mm: page allocato... |
2819 2820 2821 2822 |
if (zone_watermark_ok_safe(zone, order, low_wmark_pages(zone), 0, 0)) return; trace_mm_vmscan_wakeup_kswapd(pgdat->node_id, zone_idx(zone), order); |
8d0986e28 [PATCH] vm: kswap... |
2823 |
wake_up_interruptible(&pgdat->kswapd_wait); |
1da177e4c Linux-2.6.12-rc2 |
2824 |
} |
adea02a1b mm: count only re... |
2825 2826 2827 2828 2829 2830 2831 2832 |
/* * The reclaimable count would be mostly accurate. * The less reclaimable pages may be * - mlocked pages, which will be moved to unevictable list when encountered * - mapped pages, which may require several travels to be reclaimed * - dirty pages, which is not "instantly" reclaimable */ unsigned long global_reclaimable_pages(void) |
4f98a2fee vmscan: split LRU... |
2833 |
{ |
adea02a1b mm: count only re... |
2834 2835 2836 2837 2838 2839 2840 2841 2842 2843 2844 2845 2846 2847 2848 2849 2850 2851 2852 2853 2854 2855 2856 2857 |
int nr; nr = global_page_state(NR_ACTIVE_FILE) + global_page_state(NR_INACTIVE_FILE); if (nr_swap_pages > 0) nr += global_page_state(NR_ACTIVE_ANON) + global_page_state(NR_INACTIVE_ANON); return nr; } unsigned long zone_reclaimable_pages(struct zone *zone) { int nr; nr = zone_page_state(zone, NR_ACTIVE_FILE) + zone_page_state(zone, NR_INACTIVE_FILE); if (nr_swap_pages > 0) nr += zone_page_state(zone, NR_ACTIVE_ANON) + zone_page_state(zone, NR_INACTIVE_ANON); return nr; |
4f98a2fee vmscan: split LRU... |
2858 |
} |
c6f37f121 PM/Suspend: Do no... |
2859 |
#ifdef CONFIG_HIBERNATION |
1da177e4c Linux-2.6.12-rc2 |
2860 |
/* |
7b51755c3 vmscan: kill hibe... |
2861 |
* Try to free `nr_to_reclaim' of memory, system-wide, and return the number of |
d6277db4a [PATCH] swsusp: r... |
2862 2863 2864 2865 2866 |
* 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 Linux-2.6.12-rc2 |
2867 |
*/ |
7b51755c3 vmscan: kill hibe... |
2868 |
unsigned long shrink_all_memory(unsigned long nr_to_reclaim) |
1da177e4c Linux-2.6.12-rc2 |
2869 |
{ |
d6277db4a [PATCH] swsusp: r... |
2870 |
struct reclaim_state reclaim_state; |
d6277db4a [PATCH] swsusp: r... |
2871 |
struct scan_control sc = { |
7b51755c3 vmscan: kill hibe... |
2872 2873 2874 |
.gfp_mask = GFP_HIGHUSER_MOVABLE, .may_swap = 1, .may_unmap = 1, |
d6277db4a [PATCH] swsusp: r... |
2875 |
.may_writepage = 1, |
7b51755c3 vmscan: kill hibe... |
2876 2877 |
.nr_to_reclaim = nr_to_reclaim, .hibernation_mode = 1, |
7b51755c3 vmscan: kill hibe... |
2878 |
.order = 0, |
1da177e4c Linux-2.6.12-rc2 |
2879 |
}; |
a09ed5e00 vmscan: change sh... |
2880 2881 2882 2883 |
struct shrink_control shrink = { .gfp_mask = sc.gfp_mask, }; struct zonelist *zonelist = node_zonelist(numa_node_id(), sc.gfp_mask); |
7b51755c3 vmscan: kill hibe... |
2884 2885 |
struct task_struct *p = current; unsigned long nr_reclaimed; |
1da177e4c Linux-2.6.12-rc2 |
2886 |
|
7b51755c3 vmscan: kill hibe... |
2887 2888 2889 2890 |
p->flags |= PF_MEMALLOC; lockdep_set_current_reclaim_state(sc.gfp_mask); reclaim_state.reclaimed_slab = 0; p->reclaim_state = &reclaim_state; |
d6277db4a [PATCH] swsusp: r... |
2891 |
|
a09ed5e00 vmscan: change sh... |
2892 |
nr_reclaimed = do_try_to_free_pages(zonelist, &sc, &shrink); |
d979677c4 mm: shrink_all_me... |
2893 |
|
7b51755c3 vmscan: kill hibe... |
2894 2895 2896 |
p->reclaim_state = NULL; lockdep_clear_current_reclaim_state(); p->flags &= ~PF_MEMALLOC; |
d6277db4a [PATCH] swsusp: r... |
2897 |
|
7b51755c3 vmscan: kill hibe... |
2898 |
return nr_reclaimed; |
1da177e4c Linux-2.6.12-rc2 |
2899 |
} |
c6f37f121 PM/Suspend: Do no... |
2900 |
#endif /* CONFIG_HIBERNATION */ |
1da177e4c Linux-2.6.12-rc2 |
2901 |
|
1da177e4c Linux-2.6.12-rc2 |
2902 2903 2904 2905 |
/* 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. */ |
9c7b216d2 [PATCH] cpu hotpl... |
2906 |
static int __devinit cpu_callback(struct notifier_block *nfb, |
69e05944a [PATCH] vmscan: u... |
2907 |
unsigned long action, void *hcpu) |
1da177e4c Linux-2.6.12-rc2 |
2908 |
{ |
58c0a4a78 Fix panic of cpu ... |
2909 |
int nid; |
1da177e4c Linux-2.6.12-rc2 |
2910 |
|
8bb784428 Add suspend-relat... |
2911 |
if (action == CPU_ONLINE || action == CPU_ONLINE_FROZEN) { |
58c0a4a78 Fix panic of cpu ... |
2912 |
for_each_node_state(nid, N_HIGH_MEMORY) { |
c5f59f083 nodemask: use new... |
2913 |
pg_data_t *pgdat = NODE_DATA(nid); |
a70f73028 cpumask: replace ... |
2914 2915 2916 |
const struct cpumask *mask; mask = cpumask_of_node(pgdat->node_id); |
c5f59f083 nodemask: use new... |
2917 |
|
3e5979453 cpumask: remove a... |
2918 |
if (cpumask_any_and(cpu_online_mask, mask) < nr_cpu_ids) |
1da177e4c Linux-2.6.12-rc2 |
2919 |
/* One of our CPUs online: restore mask */ |
c5f59f083 nodemask: use new... |
2920 |
set_cpus_allowed_ptr(pgdat->kswapd, mask); |
1da177e4c Linux-2.6.12-rc2 |
2921 2922 2923 2924 |
} } return NOTIFY_OK; } |
1da177e4c Linux-2.6.12-rc2 |
2925 |
|
3218ae14b [PATCH] pgdat all... |
2926 2927 2928 2929 2930 2931 2932 2933 2934 2935 2936 2937 2938 2939 2940 2941 2942 2943 2944 2945 2946 2947 |
/* * 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); printk("Failed to start kswapd on node %d ",nid); ret = -1; } return ret; } |
8fe23e057 mm: clear node in... |
2948 2949 2950 2951 2952 2953 2954 2955 2956 2957 |
/* * Called by memory hotplug when all memory in a node is offlined. */ void kswapd_stop(int nid) { struct task_struct *kswapd = NODE_DATA(nid)->kswapd; if (kswapd) kthread_stop(kswapd); } |
1da177e4c Linux-2.6.12-rc2 |
2958 2959 |
static int __init kswapd_init(void) { |
3218ae14b [PATCH] pgdat all... |
2960 |
int nid; |
69e05944a [PATCH] vmscan: u... |
2961 |
|
1da177e4c Linux-2.6.12-rc2 |
2962 |
swap_setup(); |
9422ffba4 Memoryless nodes:... |
2963 |
for_each_node_state(nid, N_HIGH_MEMORY) |
3218ae14b [PATCH] pgdat all... |
2964 |
kswapd_run(nid); |
1da177e4c Linux-2.6.12-rc2 |
2965 2966 2967 2968 2969 |
hotcpu_notifier(cpu_callback, 0); return 0; } module_init(kswapd_init) |
9eeff2395 [PATCH] Zone recl... |
2970 2971 2972 2973 2974 2975 2976 |
#ifdef CONFIG_NUMA /* * Zone reclaim mode * * If non-zero call zone_reclaim when the number of free pages falls below * the watermarks. |
9eeff2395 [PATCH] Zone recl... |
2977 2978 |
*/ int zone_reclaim_mode __read_mostly; |
1b2ffb789 [PATCH] Zone recl... |
2979 |
#define RECLAIM_OFF 0 |
7d03431cf swapfile/vmscan: ... |
2980 |
#define RECLAIM_ZONE (1<<0) /* Run shrink_inactive_list on the zone */ |
1b2ffb789 [PATCH] Zone recl... |
2981 2982 |
#define RECLAIM_WRITE (1<<1) /* Writeout pages during reclaim */ #define RECLAIM_SWAP (1<<2) /* Swap pages out during reclaim */ |
9eeff2395 [PATCH] Zone recl... |
2983 |
/* |
a92f71263 [PATCH] zone_recl... |
2984 2985 2986 2987 2988 |
* 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 [PATCH] Zone recl... |
2989 |
/* |
9614634fe [PATCH] ZVC/zone_... |
2990 2991 2992 2993 2994 2995 |
* Percentage of pages in a zone that must be unmapped for zone_reclaim to * occur. */ int sysctl_min_unmapped_ratio = 1; /* |
0ff38490c [PATCH] zone_recl... |
2996 2997 2998 2999 |
* 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 vmscan: properly ... |
3000 3001 3002 3003 3004 3005 3006 3007 3008 3009 3010 3011 3012 3013 3014 3015 3016 3017 3018 3019 3020 3021 3022 3023 3024 3025 3026 3027 3028 3029 3030 3031 3032 3033 3034 3035 3036 3037 3038 3039 3040 |
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 [PATCH] zone_recl... |
3041 |
/* |
9eeff2395 [PATCH] Zone recl... |
3042 3043 |
* Try to free up some pages from this zone through reclaim. */ |
179e96395 [PATCH] vmscan: s... |
3044 |
static int __zone_reclaim(struct zone *zone, gfp_t gfp_mask, unsigned int order) |
9eeff2395 [PATCH] Zone recl... |
3045 |
{ |
7fb2d46d3 [PATCH] zone_recl... |
3046 |
/* Minimum pages needed in order to stay on node */ |
69e05944a [PATCH] vmscan: u... |
3047 |
const unsigned long nr_pages = 1 << order; |
9eeff2395 [PATCH] Zone recl... |
3048 3049 |
struct task_struct *p = current; struct reclaim_state reclaim_state; |
8695949a1 [PATCH] Thin out ... |
3050 |
int priority; |
179e96395 [PATCH] vmscan: s... |
3051 3052 |
struct scan_control sc = { .may_writepage = !!(zone_reclaim_mode & RECLAIM_WRITE), |
a6dc60f89 vmscan: rename sc... |
3053 |
.may_unmap = !!(zone_reclaim_mode & RECLAIM_SWAP), |
2e2e42598 vmscan,memcg: rei... |
3054 |
.may_swap = 1, |
22fba3354 vmscan: separate ... |
3055 3056 |
.nr_to_reclaim = max_t(unsigned long, nr_pages, SWAP_CLUSTER_MAX), |
179e96395 [PATCH] vmscan: s... |
3057 |
.gfp_mask = gfp_mask, |
bd2f6199c vmscan: respect h... |
3058 |
.order = order, |
179e96395 [PATCH] vmscan: s... |
3059 |
}; |
a09ed5e00 vmscan: change sh... |
3060 3061 3062 |
struct shrink_control shrink = { .gfp_mask = sc.gfp_mask, }; |
157480489 vmscan: avoid sub... |
3063 |
unsigned long nr_slab_pages0, nr_slab_pages1; |
9eeff2395 [PATCH] Zone recl... |
3064 |
|
9eeff2395 [PATCH] Zone recl... |
3065 |
cond_resched(); |
d4f7796e9 [PATCH] vmscan: f... |
3066 3067 3068 3069 3070 3071 |
/* * 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 mm, lockdep: anno... |
3072 |
lockdep_set_current_reclaim_state(gfp_mask); |
9eeff2395 [PATCH] Zone recl... |
3073 3074 |
reclaim_state.reclaimed_slab = 0; p->reclaim_state = &reclaim_state; |
c84db23c6 [PATCH] zone_recl... |
3075 |
|
90afa5de6 vmscan: properly ... |
3076 |
if (zone_pagecache_reclaimable(zone) > zone->min_unmapped_pages) { |
0ff38490c [PATCH] zone_recl... |
3077 3078 3079 3080 3081 3082 |
/* * Free memory by calling shrink zone with increasing * priorities until we have enough memory freed. */ priority = ZONE_RECLAIM_PRIORITY; do { |
a79311c14 vmscan: bail out ... |
3083 |
shrink_zone(priority, zone, &sc); |
0ff38490c [PATCH] zone_recl... |
3084 |
priority--; |
a79311c14 vmscan: bail out ... |
3085 |
} while (priority >= 0 && sc.nr_reclaimed < nr_pages); |
0ff38490c [PATCH] zone_recl... |
3086 |
} |
c84db23c6 [PATCH] zone_recl... |
3087 |
|
157480489 vmscan: avoid sub... |
3088 3089 |
nr_slab_pages0 = zone_page_state(zone, NR_SLAB_RECLAIMABLE); if (nr_slab_pages0 > zone->min_slab_pages) { |
2a16e3f4b [PATCH] Reclaim s... |
3090 |
/* |
7fb2d46d3 [PATCH] zone_recl... |
3091 |
* shrink_slab() does not currently allow us to determine how |
0ff38490c [PATCH] zone_recl... |
3092 3093 3094 3095 |
* 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 [PATCH] Reclaim s... |
3096 |
* |
0ff38490c [PATCH] zone_recl... |
3097 3098 |
* Note that shrink_slab will free memory on all zones and may * take a long time. |
2a16e3f4b [PATCH] Reclaim s... |
3099 |
*/ |
4dc4b3d97 vmscan: shrink_sl... |
3100 3101 3102 3103 |
for (;;) { unsigned long lru_pages = zone_reclaimable_pages(zone); /* No reclaimable slab or very low memory pressure */ |
1495f230f vmscan: change sh... |
3104 |
if (!shrink_slab(&shrink, sc.nr_scanned, lru_pages)) |
4dc4b3d97 vmscan: shrink_sl... |
3105 3106 3107 3108 3109 3110 3111 3112 |
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 [PATCH] zone recl... |
3113 3114 3115 3116 3117 |
/* * Update nr_reclaimed by the number of slab pages we * reclaimed from this zone. */ |
157480489 vmscan: avoid sub... |
3118 3119 3120 |
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 [PATCH] Reclaim s... |
3121 |
} |
9eeff2395 [PATCH] Zone recl... |
3122 |
p->reclaim_state = NULL; |
d4f7796e9 [PATCH] vmscan: f... |
3123 |
current->flags &= ~(PF_MEMALLOC | PF_SWAPWRITE); |
76ca542d8 mm, lockdep: anno... |
3124 |
lockdep_clear_current_reclaim_state(); |
a79311c14 vmscan: bail out ... |
3125 |
return sc.nr_reclaimed >= nr_pages; |
9eeff2395 [PATCH] Zone recl... |
3126 |
} |
179e96395 [PATCH] vmscan: s... |
3127 3128 3129 |
int zone_reclaim(struct zone *zone, gfp_t gfp_mask, unsigned int order) { |
179e96395 [PATCH] vmscan: s... |
3130 |
int node_id; |
d773ed6b8 mm: test and set ... |
3131 |
int ret; |
179e96395 [PATCH] vmscan: s... |
3132 3133 |
/* |
0ff38490c [PATCH] zone_recl... |
3134 3135 |
* Zone reclaim reclaims unmapped file backed pages and * slab pages if we are over the defined limits. |
34aa1330f [PATCH] zoned vm ... |
3136 |
* |
9614634fe [PATCH] ZVC/zone_... |
3137 3138 3139 3140 3141 |
* 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 [PATCH] vmscan: s... |
3142 |
*/ |
90afa5de6 vmscan: properly ... |
3143 3144 |
if (zone_pagecache_reclaimable(zone) <= zone->min_unmapped_pages && zone_page_state(zone, NR_SLAB_RECLAIMABLE) <= zone->min_slab_pages) |
fa5e084e4 vmscan: do not un... |
3145 |
return ZONE_RECLAIM_FULL; |
179e96395 [PATCH] vmscan: s... |
3146 |
|
93e4a89a8 mm: restore zone-... |
3147 |
if (zone->all_unreclaimable) |
fa5e084e4 vmscan: do not un... |
3148 |
return ZONE_RECLAIM_FULL; |
d773ed6b8 mm: test and set ... |
3149 |
|
179e96395 [PATCH] vmscan: s... |
3150 |
/* |
d773ed6b8 mm: test and set ... |
3151 |
* Do not scan if the allocation should not be delayed. |
179e96395 [PATCH] vmscan: s... |
3152 |
*/ |
d773ed6b8 mm: test and set ... |
3153 |
if (!(gfp_mask & __GFP_WAIT) || (current->flags & PF_MEMALLOC)) |
fa5e084e4 vmscan: do not un... |
3154 |
return ZONE_RECLAIM_NOSCAN; |
179e96395 [PATCH] vmscan: s... |
3155 3156 3157 3158 3159 3160 3161 |
/* * 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 [PATCH] NUMA: Add... |
3162 |
node_id = zone_to_nid(zone); |
37c0708db Memoryless nodes:... |
3163 |
if (node_state(node_id, N_CPU) && node_id != numa_node_id()) |
fa5e084e4 vmscan: do not un... |
3164 |
return ZONE_RECLAIM_NOSCAN; |
d773ed6b8 mm: test and set ... |
3165 3166 |
if (zone_test_and_set_flag(zone, ZONE_RECLAIM_LOCKED)) |
fa5e084e4 vmscan: do not un... |
3167 |
return ZONE_RECLAIM_NOSCAN; |
d773ed6b8 mm: test and set ... |
3168 3169 |
ret = __zone_reclaim(zone, gfp_mask, order); zone_clear_flag(zone, ZONE_RECLAIM_LOCKED); |
24cf72518 vmscan: count the... |
3170 3171 |
if (!ret) count_vm_event(PGSCAN_ZONE_RECLAIM_FAILED); |
d773ed6b8 mm: test and set ... |
3172 |
return ret; |
179e96395 [PATCH] vmscan: s... |
3173 |
} |
9eeff2395 [PATCH] Zone recl... |
3174 |
#endif |
894bc3104 Unevictable LRU I... |
3175 |
|
894bc3104 Unevictable LRU I... |
3176 3177 3178 3179 3180 3181 |
/* * page_evictable - test whether a page is evictable * @page: the page to test * @vma: the VMA in which the page is or will be mapped, may be NULL * * Test whether page is evictable--i.e., should be placed on active/inactive |
b291f0003 mlock: mlocked pa... |
3182 3183 |
* lists vs unevictable list. The vma argument is !NULL when called from the * fault path to determine how to instantate a new page. |
894bc3104 Unevictable LRU I... |
3184 3185 |
* * Reasons page might not be evictable: |
ba9ddf493 Ramfs and Ram Dis... |
3186 |
* (1) page's mapping marked unevictable |
b291f0003 mlock: mlocked pa... |
3187 |
* (2) page is part of an mlocked VMA |
ba9ddf493 Ramfs and Ram Dis... |
3188 |
* |
894bc3104 Unevictable LRU I... |
3189 3190 3191 |
*/ int page_evictable(struct page *page, struct vm_area_struct *vma) { |
ba9ddf493 Ramfs and Ram Dis... |
3192 3193 |
if (mapping_unevictable(page_mapping(page))) return 0; |
b291f0003 mlock: mlocked pa... |
3194 3195 |
if (PageMlocked(page) || (vma && is_mlocked_vma(vma, page))) return 0; |
894bc3104 Unevictable LRU I... |
3196 3197 3198 |
return 1; } |
89e004ea5 SHM_LOCKED pages ... |
3199 3200 3201 3202 3203 3204 3205 3206 3207 3208 3209 3210 3211 3212 3213 3214 3215 3216 3217 |
/** * check_move_unevictable_page - check page for evictability and move to appropriate zone lru list * @page: page to check evictability and move to appropriate lru list * @zone: zone page is in * * Checks a page for evictability and moves the page to the appropriate * zone lru list. * * Restrictions: zone->lru_lock must be held, page must be on LRU and must * have PageUnevictable set. */ static void check_move_unevictable_page(struct page *page, struct zone *zone) { VM_BUG_ON(PageActive(page)); retry: ClearPageUnevictable(page); if (page_evictable(page, NULL)) { |
401a8e1c1 mm: introduce pag... |
3218 |
enum lru_list l = page_lru_base_type(page); |
af936a160 vmscan: unevictab... |
3219 |
|
89e004ea5 SHM_LOCKED pages ... |
3220 3221 |
__dec_zone_state(zone, NR_UNEVICTABLE); list_move(&page->lru, &zone->lru[l].list); |
08e552c69 memcg: synchroniz... |
3222 |
mem_cgroup_move_lists(page, LRU_UNEVICTABLE, l); |
89e004ea5 SHM_LOCKED pages ... |
3223 3224 3225 3226 3227 3228 3229 3230 |
__inc_zone_state(zone, NR_INACTIVE_ANON + l); __count_vm_event(UNEVICTABLE_PGRESCUED); } else { /* * rotate unevictable list */ SetPageUnevictable(page); list_move(&page->lru, &zone->lru[LRU_UNEVICTABLE].list); |
08e552c69 memcg: synchroniz... |
3231 |
mem_cgroup_rotate_lru_list(page, LRU_UNEVICTABLE); |
89e004ea5 SHM_LOCKED pages ... |
3232 3233 3234 3235 3236 3237 3238 3239 3240 3241 3242 3243 3244 3245 3246 3247 3248 3249 3250 3251 3252 3253 3254 3255 3256 3257 3258 3259 3260 3261 3262 3263 3264 3265 3266 3267 3268 3269 3270 3271 3272 3273 3274 3275 3276 3277 3278 3279 3280 3281 3282 3283 3284 3285 3286 3287 3288 3289 3290 |
if (page_evictable(page, NULL)) goto retry; } } /** * scan_mapping_unevictable_pages - scan an address space for evictable pages * @mapping: struct address_space to scan for evictable pages * * Scan all pages in mapping. Check unevictable pages for * evictability and move them to the appropriate zone lru list. */ void scan_mapping_unevictable_pages(struct address_space *mapping) { pgoff_t next = 0; pgoff_t end = (i_size_read(mapping->host) + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT; struct zone *zone; struct pagevec pvec; if (mapping->nrpages == 0) return; pagevec_init(&pvec, 0); while (next < end && pagevec_lookup(&pvec, mapping, next, PAGEVEC_SIZE)) { int i; int pg_scanned = 0; zone = NULL; for (i = 0; i < pagevec_count(&pvec); i++) { struct page *page = pvec.pages[i]; pgoff_t page_index = page->index; struct zone *pagezone = page_zone(page); pg_scanned++; if (page_index > next) next = page_index; next++; if (pagezone != zone) { if (zone) spin_unlock_irq(&zone->lru_lock); zone = pagezone; spin_lock_irq(&zone->lru_lock); } if (PageLRU(page) && PageUnevictable(page)) check_move_unevictable_page(page, zone); } if (zone) spin_unlock_irq(&zone->lru_lock); pagevec_release(&pvec); count_vm_events(UNEVICTABLE_PGSCANNED, pg_scanned); } } |
af936a160 vmscan: unevictab... |
3291 |
|
264e56d82 mm: disable user ... |
3292 |
static void warn_scan_unevictable_pages(void) |
af936a160 vmscan: unevictab... |
3293 |
{ |
264e56d82 mm: disable user ... |
3294 |
printk_once(KERN_WARNING |
25bd91bd2 vmscan: add task ... |
3295 |
"%s: The scan_unevictable_pages sysctl/node-interface has been " |
264e56d82 mm: disable user ... |
3296 |
"disabled for lack of a legitimate use case. If you have " |
25bd91bd2 vmscan: add task ... |
3297 3298 3299 |
"one, please send an email to linux-mm@kvack.org. ", current->comm); |
af936a160 vmscan: unevictab... |
3300 3301 3302 3303 3304 3305 3306 3307 3308 |
} /* * 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 sysctl: remove "s... |
3309 |
void __user *buffer, |
af936a160 vmscan: unevictab... |
3310 3311 |
size_t *length, loff_t *ppos) { |
264e56d82 mm: disable user ... |
3312 |
warn_scan_unevictable_pages(); |
8d65af789 sysctl: remove "s... |
3313 |
proc_doulongvec_minmax(table, write, buffer, length, ppos); |
af936a160 vmscan: unevictab... |
3314 3315 3316 |
scan_unevictable_pages = 0; return 0; } |
e4455abb5 mm: only build pe... |
3317 |
#ifdef CONFIG_NUMA |
af936a160 vmscan: unevictab... |
3318 3319 3320 3321 |
/* * per node 'scan_unevictable_pages' attribute. On demand re-scan of * a specified node's per zone unevictable lists for evictable pages. */ |
10fbcf4c6 convert 'memory' ... |
3322 3323 |
static ssize_t read_scan_unevictable_node(struct device *dev, struct device_attribute *attr, |
af936a160 vmscan: unevictab... |
3324 3325 |
char *buf) { |
264e56d82 mm: disable user ... |
3326 |
warn_scan_unevictable_pages(); |
af936a160 vmscan: unevictab... |
3327 3328 3329 |
return sprintf(buf, "0 "); /* always zero; should fit... */ } |
10fbcf4c6 convert 'memory' ... |
3330 3331 |
static ssize_t write_scan_unevictable_node(struct device *dev, struct device_attribute *attr, |
af936a160 vmscan: unevictab... |
3332 3333 |
const char *buf, size_t count) { |
264e56d82 mm: disable user ... |
3334 |
warn_scan_unevictable_pages(); |
af936a160 vmscan: unevictab... |
3335 3336 |
return 1; } |
10fbcf4c6 convert 'memory' ... |
3337 |
static DEVICE_ATTR(scan_unevictable_pages, S_IRUGO | S_IWUSR, |
af936a160 vmscan: unevictab... |
3338 3339 3340 3341 3342 |
read_scan_unevictable_node, write_scan_unevictable_node); int scan_unevictable_register_node(struct node *node) { |
10fbcf4c6 convert 'memory' ... |
3343 |
return device_create_file(&node->dev, &dev_attr_scan_unevictable_pages); |
af936a160 vmscan: unevictab... |
3344 3345 3346 3347 |
} void scan_unevictable_unregister_node(struct node *node) { |
10fbcf4c6 convert 'memory' ... |
3348 |
device_remove_file(&node->dev, &dev_attr_scan_unevictable_pages); |
af936a160 vmscan: unevictab... |
3349 |
} |
e4455abb5 mm: only build pe... |
3350 |
#endif |