Blame view
mm/vmscan.c
102 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 |
|
f16015fbf mm: vmscan: disti... |
97 98 99 100 101 |
/* * The memory cgroup that hit its limit and as a result is the * primary target of this reclaim invocation. */ struct mem_cgroup *target_mem_cgroup; |
66e1707bc Memory controller... |
102 |
|
327c0e968 vmscan: fix it to... |
103 104 105 106 107 |
/* * Nodemask of nodes allowed by the caller. If NULL, all nodes * are scanned. */ nodemask_t *nodemask; |
1da177e4c Linux-2.6.12-rc2 |
108 |
}; |
f16015fbf mm: vmscan: disti... |
109 110 111 112 |
struct mem_cgroup_zone { struct mem_cgroup *mem_cgroup; struct zone *zone; }; |
1da177e4c Linux-2.6.12-rc2 |
113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 |
#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... |
147 |
long vm_total_pages; /* The total number of pages which the VM controls */ |
1da177e4c Linux-2.6.12-rc2 |
148 149 150 |
static LIST_HEAD(shrinker_list); static DECLARE_RWSEM(shrinker_rwsem); |
00f0b8259 Memory controller... |
151 |
#ifdef CONFIG_CGROUP_MEM_RES_CTLR |
89b5fae53 mm: vmscan: disti... |
152 153 |
static bool global_reclaim(struct scan_control *sc) { |
f16015fbf mm: vmscan: disti... |
154 |
return !sc->target_mem_cgroup; |
89b5fae53 mm: vmscan: disti... |
155 |
} |
f16015fbf mm: vmscan: disti... |
156 |
static bool scanning_global_lru(struct mem_cgroup_zone *mz) |
89b5fae53 mm: vmscan: disti... |
157 |
{ |
f16015fbf mm: vmscan: disti... |
158 |
return !mz->mem_cgroup; |
89b5fae53 mm: vmscan: disti... |
159 |
} |
91a45470f per-zone and recl... |
160 |
#else |
89b5fae53 mm: vmscan: disti... |
161 162 163 164 |
static bool global_reclaim(struct scan_control *sc) { return true; } |
f16015fbf mm: vmscan: disti... |
165 |
static bool scanning_global_lru(struct mem_cgroup_zone *mz) |
89b5fae53 mm: vmscan: disti... |
166 167 168 |
{ return true; } |
91a45470f per-zone and recl... |
169 |
#endif |
f16015fbf mm: vmscan: disti... |
170 |
static struct zone_reclaim_stat *get_reclaim_stat(struct mem_cgroup_zone *mz) |
6e9015716 mm: introduce zon... |
171 |
{ |
f16015fbf mm: vmscan: disti... |
172 173 |
if (!scanning_global_lru(mz)) return mem_cgroup_get_reclaim_stat(mz->mem_cgroup, mz->zone); |
3e2f41f1f memcg: add zone_r... |
174 |
|
f16015fbf mm: vmscan: disti... |
175 |
return &mz->zone->reclaim_stat; |
6e9015716 mm: introduce zon... |
176 |
} |
f16015fbf mm: vmscan: disti... |
177 178 |
static unsigned long zone_nr_lru_pages(struct mem_cgroup_zone *mz, enum lru_list lru) |
c9f299d98 mm: add zone nr_p... |
179 |
{ |
f16015fbf mm: vmscan: disti... |
180 181 182 183 184 |
if (!scanning_global_lru(mz)) return mem_cgroup_zone_nr_lru_pages(mz->mem_cgroup, zone_to_nid(mz->zone), zone_idx(mz->zone), BIT(lru)); |
a3d8e0549 memcg: add mem_cg... |
185 |
|
f16015fbf mm: vmscan: disti... |
186 |
return zone_page_state(mz->zone, NR_LRU_BASE + lru); |
c9f299d98 mm: add zone nr_p... |
187 |
} |
1da177e4c Linux-2.6.12-rc2 |
188 189 190 |
/* * Add a shrinker callback to be called from the vm */ |
8e1f936b7 mm: clean up and ... |
191 |
void register_shrinker(struct shrinker *shrinker) |
1da177e4c Linux-2.6.12-rc2 |
192 |
{ |
83aeeada7 vmscan: use atomi... |
193 |
atomic_long_set(&shrinker->nr_in_batch, 0); |
8e1f936b7 mm: clean up and ... |
194 195 196 |
down_write(&shrinker_rwsem); list_add_tail(&shrinker->list, &shrinker_list); up_write(&shrinker_rwsem); |
1da177e4c Linux-2.6.12-rc2 |
197 |
} |
8e1f936b7 mm: clean up and ... |
198 |
EXPORT_SYMBOL(register_shrinker); |
1da177e4c Linux-2.6.12-rc2 |
199 200 201 202 |
/* * Remove one */ |
8e1f936b7 mm: clean up and ... |
203 |
void unregister_shrinker(struct shrinker *shrinker) |
1da177e4c Linux-2.6.12-rc2 |
204 205 206 207 |
{ down_write(&shrinker_rwsem); list_del(&shrinker->list); up_write(&shrinker_rwsem); |
1da177e4c Linux-2.6.12-rc2 |
208 |
} |
8e1f936b7 mm: clean up and ... |
209 |
EXPORT_SYMBOL(unregister_shrinker); |
1da177e4c Linux-2.6.12-rc2 |
210 |
|
1495f230f vmscan: change sh... |
211 212 213 214 215 216 217 |
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 |
218 219 220 221 222 223 224 225 226 |
#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/ |
227 |
* If the vm encountered mapped pages on the LRU it increase the pressure on |
1da177e4c Linux-2.6.12-rc2 |
228 229 230 231 232 233 234 |
* 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... |
235 236 |
* * Returns the number of slab objects which we shrunk. |
1da177e4c Linux-2.6.12-rc2 |
237 |
*/ |
a09ed5e00 vmscan: change sh... |
238 |
unsigned long shrink_slab(struct shrink_control *shrink, |
1495f230f vmscan: change sh... |
239 |
unsigned long nr_pages_scanned, |
a09ed5e00 vmscan: change sh... |
240 |
unsigned long lru_pages) |
1da177e4c Linux-2.6.12-rc2 |
241 242 |
{ struct shrinker *shrinker; |
69e05944a [PATCH] vmscan: u... |
243 |
unsigned long ret = 0; |
1da177e4c Linux-2.6.12-rc2 |
244 |
|
1495f230f vmscan: change sh... |
245 246 |
if (nr_pages_scanned == 0) nr_pages_scanned = SWAP_CLUSTER_MAX; |
1da177e4c Linux-2.6.12-rc2 |
247 |
|
f06590bd7 mm: vmscan: corre... |
248 249 250 251 252 |
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 |
253 254 255 |
list_for_each_entry(shrinker, &shrinker_list, list) { unsigned long long delta; |
635697c66 vmscan: fix initi... |
256 257 |
long total_scan; long max_pass; |
095760730 vmscan: add shrin... |
258 |
int shrink_ret = 0; |
acf92b485 vmscan: shrinker-... |
259 260 |
long nr; long new_nr; |
e9299f505 vmscan: add custo... |
261 262 |
long batch_size = shrinker->batch ? shrinker->batch : SHRINK_BATCH; |
1da177e4c Linux-2.6.12-rc2 |
263 |
|
635697c66 vmscan: fix initi... |
264 265 266 |
max_pass = do_shrinker_shrink(shrinker, shrink, 0); if (max_pass <= 0) continue; |
acf92b485 vmscan: shrinker-... |
267 268 269 270 271 |
/* * 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... |
272 |
nr = atomic_long_xchg(&shrinker->nr_in_batch, 0); |
acf92b485 vmscan: shrinker-... |
273 274 |
total_scan = nr; |
1495f230f vmscan: change sh... |
275 |
delta = (4 * nr_pages_scanned) / shrinker->seeks; |
ea164d73a [PATCH] shrinker-... |
276 |
delta *= max_pass; |
1da177e4c Linux-2.6.12-rc2 |
277 |
do_div(delta, lru_pages + 1); |
acf92b485 vmscan: shrinker-... |
278 279 |
total_scan += delta; if (total_scan < 0) { |
88c3bd707 vmscan: print shr... |
280 281 282 |
printk(KERN_ERR "shrink_slab: %pF negative objects to " "delete nr=%ld ", |
acf92b485 vmscan: shrinker-... |
283 284 |
shrinker->shrink, total_scan); total_scan = max_pass; |
ea164d73a [PATCH] shrinker-... |
285 286 287 |
} /* |
3567b59aa vmscan: reduce wi... |
288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 |
* 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-... |
303 304 305 306 |
* 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-... |
307 308 |
if (total_scan > max_pass * 2) total_scan = max_pass * 2; |
1da177e4c Linux-2.6.12-rc2 |
309 |
|
acf92b485 vmscan: shrinker-... |
310 |
trace_mm_shrink_slab_start(shrinker, shrink, nr, |
095760730 vmscan: add shrin... |
311 312 |
nr_pages_scanned, lru_pages, max_pass, delta, total_scan); |
e9299f505 vmscan: add custo... |
313 |
while (total_scan >= batch_size) { |
b15e0905f [PATCH] vmscan: n... |
314 |
int nr_before; |
1da177e4c Linux-2.6.12-rc2 |
315 |
|
1495f230f vmscan: change sh... |
316 317 |
nr_before = do_shrinker_shrink(shrinker, shrink, 0); shrink_ret = do_shrinker_shrink(shrinker, shrink, |
e9299f505 vmscan: add custo... |
318 |
batch_size); |
1da177e4c Linux-2.6.12-rc2 |
319 320 |
if (shrink_ret == -1) break; |
b15e0905f [PATCH] vmscan: n... |
321 322 |
if (shrink_ret < nr_before) ret += nr_before - shrink_ret; |
e9299f505 vmscan: add custo... |
323 324 |
count_vm_events(SLABS_SCANNED, batch_size); total_scan -= batch_size; |
1da177e4c Linux-2.6.12-rc2 |
325 326 327 |
cond_resched(); } |
acf92b485 vmscan: shrinker-... |
328 329 330 331 332 |
/* * 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... |
333 334 335 336 337 |
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-... |
338 339 |
trace_mm_shrink_slab_end(shrinker, shrink_ret, nr, new_nr); |
1da177e4c Linux-2.6.12-rc2 |
340 341 |
} up_read(&shrinker_rwsem); |
f06590bd7 mm: vmscan: corre... |
342 343 |
out: cond_resched(); |
b15e0905f [PATCH] vmscan: n... |
344 |
return ret; |
1da177e4c Linux-2.6.12-rc2 |
345 |
} |
f3a310bc4 mm: vmscan: renam... |
346 |
static void set_reclaim_mode(int priority, struct scan_control *sc, |
7d3579e8e vmscan: narrow th... |
347 348 |
bool sync) { |
f3a310bc4 mm: vmscan: renam... |
349 |
reclaim_mode_t syncmode = sync ? RECLAIM_MODE_SYNC : RECLAIM_MODE_ASYNC; |
7d3579e8e vmscan: narrow th... |
350 351 |
/* |
3e7d34497 mm: vmscan: recla... |
352 353 354 |
* 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... |
355 |
*/ |
3e7d34497 mm: vmscan: recla... |
356 |
if (COMPACTION_BUILD) |
f3a310bc4 mm: vmscan: renam... |
357 |
sc->reclaim_mode = RECLAIM_MODE_COMPACTION; |
3e7d34497 mm: vmscan: recla... |
358 |
else |
f3a310bc4 mm: vmscan: renam... |
359 |
sc->reclaim_mode = RECLAIM_MODE_LUMPYRECLAIM; |
7d3579e8e vmscan: narrow th... |
360 361 |
/* |
3e7d34497 mm: vmscan: recla... |
362 363 364 |
* 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... |
365 366 |
*/ if (sc->order > PAGE_ALLOC_COSTLY_ORDER) |
f3a310bc4 mm: vmscan: renam... |
367 |
sc->reclaim_mode |= syncmode; |
7d3579e8e vmscan: narrow th... |
368 |
else if (sc->order && priority < DEF_PRIORITY - 2) |
f3a310bc4 mm: vmscan: renam... |
369 |
sc->reclaim_mode |= syncmode; |
7d3579e8e vmscan: narrow th... |
370 |
else |
f3a310bc4 mm: vmscan: renam... |
371 |
sc->reclaim_mode = RECLAIM_MODE_SINGLE | RECLAIM_MODE_ASYNC; |
7d3579e8e vmscan: narrow th... |
372 |
} |
f3a310bc4 mm: vmscan: renam... |
373 |
static void reset_reclaim_mode(struct scan_control *sc) |
7d3579e8e vmscan: narrow th... |
374 |
{ |
f3a310bc4 mm: vmscan: renam... |
375 |
sc->reclaim_mode = RECLAIM_MODE_SINGLE | RECLAIM_MODE_ASYNC; |
7d3579e8e vmscan: narrow th... |
376 |
} |
1da177e4c Linux-2.6.12-rc2 |
377 378 |
static inline int is_page_cache_freeable(struct page *page) { |
ceddc3a52 mm: document is_p... |
379 380 381 382 383 |
/* * 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... |
384 |
return page_count(page) - page_has_private(page) == 2; |
1da177e4c Linux-2.6.12-rc2 |
385 |
} |
7d3579e8e vmscan: narrow th... |
386 387 |
static int may_write_to_queue(struct backing_dev_info *bdi, struct scan_control *sc) |
1da177e4c Linux-2.6.12-rc2 |
388 |
{ |
930d91525 [PATCH] Swap Migr... |
389 |
if (current->flags & PF_SWAPWRITE) |
1da177e4c Linux-2.6.12-rc2 |
390 391 392 393 394 |
return 1; if (!bdi_write_congested(bdi)) return 1; if (bdi == current->backing_dev_info) return 1; |
7d3579e8e vmscan: narrow th... |
395 396 397 398 |
/* 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 |
399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 |
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... |
417 |
lock_page(page); |
3e9f45bd1 Factor outstandin... |
418 419 |
if (page_mapping(page) == mapping) mapping_set_error(mapping, error); |
1da177e4c Linux-2.6.12-rc2 |
420 421 |
unlock_page(page); } |
04e62a29b [PATCH] More page... |
422 423 424 425 426 427 428 429 430 431 432 |
/* 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 |
433 |
/* |
1742f19fa [PATCH] vmscan: r... |
434 435 |
* pageout is called by shrink_page_list() for each dirty page. * Calls ->writepage(). |
1da177e4c Linux-2.6.12-rc2 |
436 |
*/ |
c661b078f synchronous lumpy... |
437 |
static pageout_t pageout(struct page *page, struct address_space *mapping, |
7d3579e8e vmscan: narrow th... |
438 |
struct scan_control *sc) |
1da177e4c Linux-2.6.12-rc2 |
439 440 441 442 443 444 445 446 |
{ /* * 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... |
447 |
* If this process is currently in __generic_file_aio_write() against |
1da177e4c Linux-2.6.12-rc2 |
448 449 450 451 452 453 454 |
* 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 |
455 456 457 458 459 460 461 462 |
*/ 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... |
463 |
if (page_has_private(page)) { |
1da177e4c Linux-2.6.12-rc2 |
464 465 |
if (try_to_free_buffers(page)) { ClearPageDirty(page); |
d40cee245 mm: remove remain... |
466 467 |
printk("%s: orphaned page ", __func__); |
1da177e4c Linux-2.6.12-rc2 |
468 469 470 471 472 473 474 |
return PAGE_CLEAN; } } return PAGE_KEEP; } if (mapping->a_ops->writepage == NULL) return PAGE_ACTIVATE; |
0e093d997 writeback: do not... |
475 |
if (!may_write_to_queue(mapping->backing_dev_info, sc)) |
1da177e4c Linux-2.6.12-rc2 |
476 477 478 479 480 481 482 |
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... |
483 484 |
.range_start = 0, .range_end = LLONG_MAX, |
1da177e4c Linux-2.6.12-rc2 |
485 486 487 488 489 490 491 |
.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... |
492 |
if (res == AOP_WRITEPAGE_ACTIVATE) { |
1da177e4c Linux-2.6.12-rc2 |
493 494 495 |
ClearPageReclaim(page); return PAGE_ACTIVATE; } |
c661b078f synchronous lumpy... |
496 |
|
1da177e4c Linux-2.6.12-rc2 |
497 498 499 500 |
if (!PageWriteback(page)) { /* synchronous write or broken a_ops? */ ClearPageReclaim(page); } |
755f0225e vmscan: tracing: ... |
501 |
trace_mm_vmscan_writepage(page, |
f3a310bc4 mm: vmscan: renam... |
502 |
trace_reclaim_flags(page, sc->reclaim_mode)); |
e129b5c23 [PATCH] vm: add p... |
503 |
inc_zone_page_state(page, NR_VMSCAN_WRITE); |
1da177e4c Linux-2.6.12-rc2 |
504 505 506 507 508 |
return PAGE_SUCCESS; } return PAGE_CLEAN; } |
a649fd927 [PATCH] invalidat... |
509 |
/* |
e286781d5 mm: speculative p... |
510 511 |
* Same as remove_mapping, but if the page is removed from the mapping, it * gets returned with a refcount of 0. |
a649fd927 [PATCH] invalidat... |
512 |
*/ |
e286781d5 mm: speculative p... |
513 |
static int __remove_mapping(struct address_space *mapping, struct page *page) |
49d2e9cc4 [PATCH] Swap Migr... |
514 |
{ |
28e4d965e [PATCH] mm: remov... |
515 516 |
BUG_ON(!PageLocked(page)); BUG_ON(mapping != page_mapping(page)); |
49d2e9cc4 [PATCH] Swap Migr... |
517 |
|
19fd62312 mm: spinlock tree... |
518 |
spin_lock_irq(&mapping->tree_lock); |
49d2e9cc4 [PATCH] Swap Migr... |
519 |
/* |
0fd0e6b05 [PATCH] page inva... |
520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 |
* 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... |
543 |
*/ |
e286781d5 mm: speculative p... |
544 |
if (!page_freeze_refs(page, 2)) |
49d2e9cc4 [PATCH] Swap Migr... |
545 |
goto cannot_free; |
e286781d5 mm: speculative p... |
546 547 548 |
/* note: atomic_cmpxchg in page_freeze_refs provides the smp_rmb */ if (unlikely(PageDirty(page))) { page_unfreeze_refs(page, 2); |
49d2e9cc4 [PATCH] Swap Migr... |
549 |
goto cannot_free; |
e286781d5 mm: speculative p... |
550 |
} |
49d2e9cc4 [PATCH] Swap Migr... |
551 552 553 554 |
if (PageSwapCache(page)) { swp_entry_t swap = { .val = page_private(page) }; __delete_from_swap_cache(page); |
19fd62312 mm: spinlock tree... |
555 |
spin_unlock_irq(&mapping->tree_lock); |
cb4b86ba4 mm: add swap cach... |
556 |
swapcache_free(swap, page); |
e286781d5 mm: speculative p... |
557 |
} else { |
6072d13c4 Call the filesyst... |
558 559 560 |
void (*freepage)(struct page *); freepage = mapping->a_ops->freepage; |
e64a782fe mm: change __remo... |
561 |
__delete_from_page_cache(page); |
19fd62312 mm: spinlock tree... |
562 |
spin_unlock_irq(&mapping->tree_lock); |
e767e0561 memcg: fix deadlo... |
563 |
mem_cgroup_uncharge_cache_page(page); |
6072d13c4 Call the filesyst... |
564 565 566 |
if (freepage != NULL) freepage(page); |
49d2e9cc4 [PATCH] Swap Migr... |
567 |
} |
49d2e9cc4 [PATCH] Swap Migr... |
568 569 570 |
return 1; cannot_free: |
19fd62312 mm: spinlock tree... |
571 |
spin_unlock_irq(&mapping->tree_lock); |
49d2e9cc4 [PATCH] Swap Migr... |
572 573 |
return 0; } |
1da177e4c Linux-2.6.12-rc2 |
574 |
/* |
e286781d5 mm: speculative p... |
575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 |
* 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... |
593 594 595 596 597 598 599 600 601 |
/** * 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... |
602 603 604 605 |
void putback_lru_page(struct page *page) { int lru; int active = !!TestClearPageActive(page); |
bbfd28eee unevictable lru: ... |
606 |
int was_unevictable = PageUnevictable(page); |
894bc3104 Unevictable LRU I... |
607 608 609 610 611 612 613 614 615 616 617 618 619 |
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... |
620 |
lru = active + page_lru_base_type(page); |
894bc3104 Unevictable LRU I... |
621 622 623 624 625 626 627 628 |
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... |
629 |
/* |
21ee9f398 vmscan: add barri... |
630 631 632 633 634 |
* 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... |
635 636 |
* the page back to the evictable list. * |
21ee9f398 vmscan: add barri... |
637 |
* The other side is TestClearPageMlocked() or shmem_lock(). |
6a7b95481 vmscan: order evi... |
638 639 |
*/ smp_mb(); |
894bc3104 Unevictable LRU I... |
640 |
} |
894bc3104 Unevictable LRU I... |
641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 |
/* * 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: ... |
657 658 659 660 |
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... |
661 662 |
put_page(page); /* drop ref from isolate */ } |
dfc8d636c vmscan: factor ou... |
663 664 665 |
enum page_references { PAGEREF_RECLAIM, PAGEREF_RECLAIM_CLEAN, |
645747462 vmscan: detect ma... |
666 |
PAGEREF_KEEP, |
dfc8d636c vmscan: factor ou... |
667 668 669 670 |
PAGEREF_ACTIVATE, }; static enum page_references page_check_references(struct page *page, |
f16015fbf mm: vmscan: disti... |
671 |
struct mem_cgroup_zone *mz, |
dfc8d636c vmscan: factor ou... |
672 673 |
struct scan_control *sc) { |
645747462 vmscan: detect ma... |
674 |
int referenced_ptes, referenced_page; |
dfc8d636c vmscan: factor ou... |
675 |
unsigned long vm_flags; |
dfc8d636c vmscan: factor ou... |
676 |
|
f16015fbf mm: vmscan: disti... |
677 |
referenced_ptes = page_referenced(page, 1, mz->mem_cgroup, &vm_flags); |
645747462 vmscan: detect ma... |
678 |
referenced_page = TestClearPageReferenced(page); |
dfc8d636c vmscan: factor ou... |
679 680 |
/* Lumpy reclaim - ignore references */ |
f3a310bc4 mm: vmscan: renam... |
681 |
if (sc->reclaim_mode & RECLAIM_MODE_LUMPYRECLAIM) |
dfc8d636c vmscan: factor ou... |
682 683 684 685 686 687 688 689 |
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... |
690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 |
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... |
708 |
if (referenced_page || referenced_ptes > 1) |
645747462 vmscan: detect ma... |
709 |
return PAGEREF_ACTIVATE; |
c909e9936 vmscan: activate ... |
710 711 712 713 714 |
/* * Activate file-backed executable pages after first usage. */ if (vm_flags & VM_EXEC) return PAGEREF_ACTIVATE; |
645747462 vmscan: detect ma... |
715 716 |
return PAGEREF_KEEP; } |
dfc8d636c vmscan: factor ou... |
717 718 |
/* Reclaim if clean, defer dirty pages to writeback */ |
2e30244a7 vmscan,tmpfs: tre... |
719 |
if (referenced_page && !PageSwapBacked(page)) |
645747462 vmscan: detect ma... |
720 721 722 |
return PAGEREF_RECLAIM_CLEAN; return PAGEREF_RECLAIM; |
dfc8d636c vmscan: factor ou... |
723 |
} |
e286781d5 mm: speculative p... |
724 |
/* |
1742f19fa [PATCH] vmscan: r... |
725 |
* shrink_page_list() returns the number of reclaimed pages |
1da177e4c Linux-2.6.12-rc2 |
726 |
*/ |
1742f19fa [PATCH] vmscan: r... |
727 |
static unsigned long shrink_page_list(struct list_head *page_list, |
f16015fbf mm: vmscan: disti... |
728 |
struct mem_cgroup_zone *mz, |
f84f6e2b0 mm: vmscan: do no... |
729 |
struct scan_control *sc, |
92df3a723 mm: vmscan: throt... |
730 731 732 |
int priority, unsigned long *ret_nr_dirty, unsigned long *ret_nr_writeback) |
1da177e4c Linux-2.6.12-rc2 |
733 734 |
{ LIST_HEAD(ret_pages); |
abe4c3b50 vmscan: set up pa... |
735 |
LIST_HEAD(free_pages); |
1da177e4c Linux-2.6.12-rc2 |
736 |
int pgactivate = 0; |
0e093d997 writeback: do not... |
737 738 |
unsigned long nr_dirty = 0; unsigned long nr_congested = 0; |
05ff51376 [PATCH] vmscan re... |
739 |
unsigned long nr_reclaimed = 0; |
92df3a723 mm: vmscan: throt... |
740 |
unsigned long nr_writeback = 0; |
1da177e4c Linux-2.6.12-rc2 |
741 742 |
cond_resched(); |
1da177e4c Linux-2.6.12-rc2 |
743 |
while (!list_empty(page_list)) { |
dfc8d636c vmscan: factor ou... |
744 |
enum page_references references; |
1da177e4c Linux-2.6.12-rc2 |
745 746 747 |
struct address_space *mapping; struct page *page; int may_enter_fs; |
1da177e4c Linux-2.6.12-rc2 |
748 749 750 751 752 |
cond_resched(); page = lru_to_page(page_list); list_del(&page->lru); |
529ae9aaa mm: rename page t... |
753 |
if (!trylock_page(page)) |
1da177e4c Linux-2.6.12-rc2 |
754 |
goto keep; |
725d704ec [PATCH] mm: VM_BU... |
755 |
VM_BUG_ON(PageActive(page)); |
f16015fbf mm: vmscan: disti... |
756 |
VM_BUG_ON(page_zone(page) != mz->zone); |
1da177e4c Linux-2.6.12-rc2 |
757 758 |
sc->nr_scanned++; |
80e434260 [PATCH] zone recl... |
759 |
|
b291f0003 mlock: mlocked pa... |
760 761 |
if (unlikely(!page_evictable(page, NULL))) goto cull_mlocked; |
894bc3104 Unevictable LRU I... |
762 |
|
a6dc60f89 vmscan: rename sc... |
763 |
if (!sc->may_unmap && page_mapped(page)) |
80e434260 [PATCH] zone recl... |
764 |
goto keep_locked; |
1da177e4c Linux-2.6.12-rc2 |
765 766 767 |
/* Double the slab pressure for mapped and swapcache pages */ if (page_mapped(page) || PageSwapCache(page)) sc->nr_scanned++; |
c661b078f synchronous lumpy... |
768 769 770 771 |
may_enter_fs = (sc->gfp_mask & __GFP_FS) || (PageSwapCache(page) && (sc->gfp_mask & __GFP_IO)); if (PageWriteback(page)) { |
92df3a723 mm: vmscan: throt... |
772 |
nr_writeback++; |
c661b078f synchronous lumpy... |
773 |
/* |
a18bba061 mm: vmscan: remov... |
774 775 776 777 |
* 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... |
778 |
*/ |
f3a310bc4 mm: vmscan: renam... |
779 |
if ((sc->reclaim_mode & RECLAIM_MODE_SYNC) && |
7d3579e8e vmscan: narrow th... |
780 |
may_enter_fs) |
c661b078f synchronous lumpy... |
781 |
wait_on_page_writeback(page); |
7d3579e8e vmscan: narrow th... |
782 783 784 785 |
else { unlock_page(page); goto keep_lumpy; } |
c661b078f synchronous lumpy... |
786 |
} |
1da177e4c Linux-2.6.12-rc2 |
787 |
|
f16015fbf mm: vmscan: disti... |
788 |
references = page_check_references(page, mz, sc); |
dfc8d636c vmscan: factor ou... |
789 790 |
switch (references) { case PAGEREF_ACTIVATE: |
1da177e4c Linux-2.6.12-rc2 |
791 |
goto activate_locked; |
645747462 vmscan: detect ma... |
792 793 |
case PAGEREF_KEEP: goto keep_locked; |
dfc8d636c vmscan: factor ou... |
794 795 796 797 |
case PAGEREF_RECLAIM: case PAGEREF_RECLAIM_CLEAN: ; /* try to reclaim the page below */ } |
1da177e4c Linux-2.6.12-rc2 |
798 |
|
1da177e4c Linux-2.6.12-rc2 |
799 800 801 802 |
/* * Anonymous process memory has backing store? * Try to allocate it some swap space here. */ |
b291f0003 mlock: mlocked pa... |
803 |
if (PageAnon(page) && !PageSwapCache(page)) { |
63eb6b93c vmscan: let GFP_N... |
804 805 |
if (!(sc->gfp_mask & __GFP_IO)) goto keep_locked; |
ac47b003d mm: remove gfp_ma... |
806 |
if (!add_to_swap(page)) |
1da177e4c Linux-2.6.12-rc2 |
807 |
goto activate_locked; |
63eb6b93c vmscan: let GFP_N... |
808 |
may_enter_fs = 1; |
b291f0003 mlock: mlocked pa... |
809 |
} |
1da177e4c Linux-2.6.12-rc2 |
810 811 |
mapping = page_mapping(page); |
1da177e4c Linux-2.6.12-rc2 |
812 813 814 815 816 817 |
/* * 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... |
818 |
switch (try_to_unmap(page, TTU_UNMAP)) { |
1da177e4c Linux-2.6.12-rc2 |
819 820 821 822 |
case SWAP_FAIL: goto activate_locked; case SWAP_AGAIN: goto keep_locked; |
b291f0003 mlock: mlocked pa... |
823 824 |
case SWAP_MLOCK: goto cull_mlocked; |
1da177e4c Linux-2.6.12-rc2 |
825 826 827 828 829 830 |
case SWAP_SUCCESS: ; /* try to free the page below */ } } if (PageDirty(page)) { |
0e093d997 writeback: do not... |
831 |
nr_dirty++; |
ee72886d8 mm: vmscan: do no... |
832 833 |
/* * Only kswapd can writeback filesystem pages to |
f84f6e2b0 mm: vmscan: do no... |
834 835 |
* avoid risk of stack overflow but do not writeback * unless under significant pressure. |
ee72886d8 mm: vmscan: do no... |
836 |
*/ |
f84f6e2b0 mm: vmscan: do no... |
837 838 |
if (page_is_file_cache(page) && (!current_is_kswapd() || priority >= DEF_PRIORITY - 2)) { |
49ea7eb65 mm: vmscan: immed... |
839 840 841 842 843 844 845 846 |
/* * 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... |
847 848 |
goto keep_locked; } |
dfc8d636c vmscan: factor ou... |
849 |
if (references == PAGEREF_RECLAIM_CLEAN) |
1da177e4c Linux-2.6.12-rc2 |
850 |
goto keep_locked; |
4dd4b9202 revert "kswapd sh... |
851 |
if (!may_enter_fs) |
1da177e4c Linux-2.6.12-rc2 |
852 |
goto keep_locked; |
52a8363ea [PATCH] mm: impro... |
853 |
if (!sc->may_writepage) |
1da177e4c Linux-2.6.12-rc2 |
854 855 856 |
goto keep_locked; /* Page is dirty, try to write it out here */ |
7d3579e8e vmscan: narrow th... |
857 |
switch (pageout(page, mapping, sc)) { |
1da177e4c Linux-2.6.12-rc2 |
858 |
case PAGE_KEEP: |
0e093d997 writeback: do not... |
859 |
nr_congested++; |
1da177e4c Linux-2.6.12-rc2 |
860 861 862 863 |
goto keep_locked; case PAGE_ACTIVATE: goto activate_locked; case PAGE_SUCCESS: |
7d3579e8e vmscan: narrow th... |
864 865 866 |
if (PageWriteback(page)) goto keep_lumpy; if (PageDirty(page)) |
1da177e4c Linux-2.6.12-rc2 |
867 |
goto keep; |
7d3579e8e vmscan: narrow th... |
868 |
|
1da177e4c Linux-2.6.12-rc2 |
869 870 871 872 |
/* * A synchronous write - probably a ramdisk. Go * ahead and try to reclaim the page. */ |
529ae9aaa mm: rename page t... |
873 |
if (!trylock_page(page)) |
1da177e4c Linux-2.6.12-rc2 |
874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 |
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... |
893 |
* will do this, as well as the blockdev mapping. |
1da177e4c Linux-2.6.12-rc2 |
894 895 896 897 898 899 900 901 902 903 |
* 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... |
904 |
if (page_has_private(page)) { |
1da177e4c Linux-2.6.12-rc2 |
905 906 |
if (!try_to_release_page(page, sc->gfp_mask)) goto activate_locked; |
e286781d5 mm: speculative p... |
907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 |
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 |
923 |
} |
e286781d5 mm: speculative p... |
924 |
if (!mapping || !__remove_mapping(mapping, page)) |
49d2e9cc4 [PATCH] Swap Migr... |
925 |
goto keep_locked; |
1da177e4c Linux-2.6.12-rc2 |
926 |
|
a978d6f52 mm: unlockless re... |
927 928 929 930 931 932 933 934 |
/* * 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... |
935 |
free_it: |
05ff51376 [PATCH] vmscan re... |
936 |
nr_reclaimed++; |
abe4c3b50 vmscan: set up pa... |
937 938 939 940 941 942 |
/* * 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 |
943 |
continue; |
b291f0003 mlock: mlocked pa... |
944 |
cull_mlocked: |
63d6c5ad7 mm: remove try_to... |
945 946 |
if (PageSwapCache(page)) try_to_free_swap(page); |
b291f0003 mlock: mlocked pa... |
947 948 |
unlock_page(page); putback_lru_page(page); |
f3a310bc4 mm: vmscan: renam... |
949 |
reset_reclaim_mode(sc); |
b291f0003 mlock: mlocked pa... |
950 |
continue; |
1da177e4c Linux-2.6.12-rc2 |
951 |
activate_locked: |
68a22394c vmscan: free swap... |
952 953 |
/* Not a candidate for swapping, so reclaim swap space. */ if (PageSwapCache(page) && vm_swap_full()) |
a2c43eed8 mm: try_to_free_s... |
954 |
try_to_free_swap(page); |
894bc3104 Unevictable LRU I... |
955 |
VM_BUG_ON(PageActive(page)); |
1da177e4c Linux-2.6.12-rc2 |
956 957 958 959 960 |
SetPageActive(page); pgactivate++; keep_locked: unlock_page(page); keep: |
f3a310bc4 mm: vmscan: renam... |
961 |
reset_reclaim_mode(sc); |
7d3579e8e vmscan: narrow th... |
962 |
keep_lumpy: |
1da177e4c Linux-2.6.12-rc2 |
963 |
list_add(&page->lru, &ret_pages); |
b291f0003 mlock: mlocked pa... |
964 |
VM_BUG_ON(PageLRU(page) || PageUnevictable(page)); |
1da177e4c Linux-2.6.12-rc2 |
965 |
} |
abe4c3b50 vmscan: set up pa... |
966 |
|
0e093d997 writeback: do not... |
967 968 969 970 971 972 |
/* * 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 */ |
89b5fae53 mm: vmscan: disti... |
973 |
if (nr_dirty && nr_dirty == nr_congested && global_reclaim(sc)) |
f16015fbf mm: vmscan: disti... |
974 |
zone_set_flag(mz->zone, ZONE_CONGESTED); |
0e093d997 writeback: do not... |
975 |
|
cc59850ef mm: add free_hot_... |
976 |
free_hot_cold_page_list(&free_pages, 1); |
abe4c3b50 vmscan: set up pa... |
977 |
|
1da177e4c Linux-2.6.12-rc2 |
978 |
list_splice(&ret_pages, page_list); |
f8891e5e1 [PATCH] Light wei... |
979 |
count_vm_events(PGACTIVATE, pgactivate); |
92df3a723 mm: vmscan: throt... |
980 981 |
*ret_nr_dirty += nr_dirty; *ret_nr_writeback += nr_writeback; |
05ff51376 [PATCH] vmscan re... |
982 |
return nr_reclaimed; |
1da177e4c Linux-2.6.12-rc2 |
983 |
} |
5ad333eb6 Lumpy Reclaim V4 |
984 985 986 987 988 989 990 991 992 993 |
/* * 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... |
994 |
int __isolate_lru_page(struct page *page, isolate_mode_t mode, int file) |
5ad333eb6 Lumpy Reclaim V4 |
995 |
{ |
4356f21d0 mm: change isolat... |
996 |
bool all_lru_mode; |
5ad333eb6 Lumpy Reclaim V4 |
997 998 999 1000 1001 |
int ret = -EINVAL; /* Only take pages on the LRU. */ if (!PageLRU(page)) return ret; |
4356f21d0 mm: change isolat... |
1002 1003 |
all_lru_mode = (mode & (ISOLATE_ACTIVE|ISOLATE_INACTIVE)) == (ISOLATE_ACTIVE|ISOLATE_INACTIVE); |
5ad333eb6 Lumpy Reclaim V4 |
1004 1005 1006 1007 1008 |
/* * 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... |
1009 |
if (!all_lru_mode && !PageActive(page) != !(mode & ISOLATE_ACTIVE)) |
5ad333eb6 Lumpy Reclaim V4 |
1010 |
return ret; |
4356f21d0 mm: change isolat... |
1011 |
if (!all_lru_mode && !!page_is_file_cache(page) != file) |
4f98a2fee vmscan: split LRU... |
1012 |
return ret; |
894bc3104 Unevictable LRU I... |
1013 1014 1015 1016 1017 1018 1019 |
/* * 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 |
1020 |
ret = -EBUSY; |
08e552c69 memcg: synchroniz... |
1021 |
|
c82449352 mm: compaction: m... |
1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 |
/* * To minimise LRU disruption, the caller can indicate that it only * wants to isolate pages it will be able to operate on without * blocking - clean pages for the most part. * * ISOLATE_CLEAN means that only clean pages should be isolated. This * is used by reclaim when it is cannot write to backing storage * * ISOLATE_ASYNC_MIGRATE is used to indicate that it only wants to pages * that it is possible to migrate without blocking */ if (mode & (ISOLATE_CLEAN|ISOLATE_ASYNC_MIGRATE)) { /* All the caller can do on PageWriteback is block */ if (PageWriteback(page)) return ret; if (PageDirty(page)) { struct address_space *mapping; /* ISOLATE_CLEAN means only clean pages */ if (mode & ISOLATE_CLEAN) return ret; /* * Only pages without mappings or that have a * ->migratepage callback are possible to migrate * without blocking */ mapping = page_mapping(page); if (mapping && !mapping->a_ops->migratepage) return ret; } } |
39deaf858 mm: compaction: m... |
1055 |
|
f80c06736 mm: zone_reclaim:... |
1056 1057 |
if ((mode & ISOLATE_UNMAPPED) && page_mapped(page)) return ret; |
5ad333eb6 Lumpy Reclaim V4 |
1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 |
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... |
1070 |
/* |
1da177e4c Linux-2.6.12-rc2 |
1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 |
* 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. |
f626012db mm: remove isolat... |
1081 |
* @mz: The mem_cgroup_zone to pull pages from. |
1da177e4c Linux-2.6.12-rc2 |
1082 |
* @dst: The temp list to put pages on to. |
f626012db mm: remove isolat... |
1083 |
* @nr_scanned: The number of pages that were scanned. |
5ad333eb6 Lumpy Reclaim V4 |
1084 1085 |
* @order: The caller's attempted allocation order * @mode: One of the LRU isolation modes |
f626012db mm: remove isolat... |
1086 |
* @active: True [1] if isolating active pages |
4f98a2fee vmscan: split LRU... |
1087 |
* @file: True [1] if isolating file [!anon] pages |
1da177e4c Linux-2.6.12-rc2 |
1088 1089 1090 |
* * returns how many pages were moved onto *@dst. */ |
69e05944a [PATCH] vmscan: u... |
1091 |
static unsigned long isolate_lru_pages(unsigned long nr_to_scan, |
f626012db mm: remove isolat... |
1092 1093 1094 |
struct mem_cgroup_zone *mz, struct list_head *dst, unsigned long *nr_scanned, int order, isolate_mode_t mode, int active, int file) |
1da177e4c Linux-2.6.12-rc2 |
1095 |
{ |
f626012db mm: remove isolat... |
1096 1097 |
struct lruvec *lruvec; struct list_head *src; |
69e05944a [PATCH] vmscan: u... |
1098 |
unsigned long nr_taken = 0; |
a8a94d151 vmscan: tracing: ... |
1099 1100 1101 |
unsigned long nr_lumpy_taken = 0; unsigned long nr_lumpy_dirty = 0; unsigned long nr_lumpy_failed = 0; |
c9b02d970 [PATCH] mm: isola... |
1102 |
unsigned long scan; |
f626012db mm: remove isolat... |
1103 1104 1105 1106 1107 1108 1109 1110 |
int lru = LRU_BASE; lruvec = mem_cgroup_zone_lruvec(mz->zone, mz->mem_cgroup); if (active) lru += LRU_ACTIVE; if (file) lru += LRU_FILE; src = &lruvec->lists[lru]; |
1da177e4c Linux-2.6.12-rc2 |
1111 |
|
c9b02d970 [PATCH] mm: isola... |
1112 |
for (scan = 0; scan < nr_to_scan && !list_empty(src); scan++) { |
5ad333eb6 Lumpy Reclaim V4 |
1113 1114 1115 1116 1117 |
struct page *page; unsigned long pfn; unsigned long end_pfn; unsigned long page_pfn; int zone_id; |
1da177e4c Linux-2.6.12-rc2 |
1118 1119 |
page = lru_to_page(src); prefetchw_prev_lru_page(page, src, flags); |
725d704ec [PATCH] mm: VM_BU... |
1120 |
VM_BUG_ON(!PageLRU(page)); |
8d438f96d [PATCH] mm: PageL... |
1121 |
|
4f98a2fee vmscan: split LRU... |
1122 |
switch (__isolate_lru_page(page, mode, file)) { |
5ad333eb6 Lumpy Reclaim V4 |
1123 |
case 0: |
925b7673c mm: make per-memc... |
1124 |
mem_cgroup_lru_del(page); |
5ad333eb6 Lumpy Reclaim V4 |
1125 |
list_move(&page->lru, dst); |
2c888cfbc thp: fix anon mem... |
1126 |
nr_taken += hpage_nr_pages(page); |
5ad333eb6 Lumpy Reclaim V4 |
1127 1128 1129 1130 1131 1132 |
break; case -EBUSY: /* else it is being freed elsewhere */ list_move(&page->lru, src); continue; |
46453a6e1 [PATCH] mm: never... |
1133 |
|
5ad333eb6 Lumpy Reclaim V4 |
1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 |
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... |
1146 |
* as the mem_map is guaranteed valid out to MAX_ORDER, |
5ad333eb6 Lumpy Reclaim V4 |
1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 |
* 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... |
1166 |
|
5ad333eb6 Lumpy Reclaim V4 |
1167 1168 |
/* Check that we have not crossed a zone boundary. */ if (unlikely(page_zone_id(cursor_page) != zone_id)) |
08fc468f4 vmscan: isolate_l... |
1169 |
break; |
de2e7567c vmscan: don't att... |
1170 1171 1172 1173 1174 1175 |
/* * 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... |
1176 |
if (nr_swap_pages <= 0 && PageSwapBacked(cursor_page) && |
08fc468f4 vmscan: isolate_l... |
1177 1178 |
!PageSwapCache(cursor_page)) break; |
de2e7567c vmscan: don't att... |
1179 |
|
ee993b135 mm: fix lumpy rec... |
1180 |
if (__isolate_lru_page(cursor_page, mode, file) == 0) { |
501347315 mm: vmscan: check... |
1181 |
unsigned int isolated_pages; |
925b7673c mm: make per-memc... |
1182 |
mem_cgroup_lru_del(cursor_page); |
5ad333eb6 Lumpy Reclaim V4 |
1183 |
list_move(&cursor_page->lru, dst); |
501347315 mm: vmscan: check... |
1184 1185 1186 |
isolated_pages = hpage_nr_pages(cursor_page); nr_taken += isolated_pages; nr_lumpy_taken += isolated_pages; |
a8a94d151 vmscan: tracing: ... |
1187 |
if (PageDirty(cursor_page)) |
501347315 mm: vmscan: check... |
1188 |
nr_lumpy_dirty += isolated_pages; |
5ad333eb6 Lumpy Reclaim V4 |
1189 |
scan++; |
501347315 mm: vmscan: check... |
1190 |
pfn += isolated_pages - 1; |
a8a94d151 vmscan: tracing: ... |
1191 |
} else { |
d179e84ba mm: vmscan: do no... |
1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 |
/* * 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... |
1206 1207 |
continue; break; |
5ad333eb6 Lumpy Reclaim V4 |
1208 1209 |
} } |
08fc468f4 vmscan: isolate_l... |
1210 1211 1212 1213 |
/* If we break out of the loop above, lumpy reclaim failed */ if (pfn < end_pfn) nr_lumpy_failed++; |
1da177e4c Linux-2.6.12-rc2 |
1214 |
} |
f626012db mm: remove isolat... |
1215 |
*nr_scanned = scan; |
a8a94d151 vmscan: tracing: ... |
1216 1217 1218 1219 1220 |
trace_mm_vmscan_lru_isolate(order, nr_to_scan, scan, nr_taken, nr_lumpy_taken, nr_lumpy_dirty, nr_lumpy_failed, |
ea4d349ff vmscan/trace: Add... |
1221 |
mode, file); |
1da177e4c Linux-2.6.12-rc2 |
1222 1223 |
return nr_taken; } |
62695a84e vmscan: move isol... |
1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 |
/** * 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... |
1235 1236 1237 |
* 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... |
1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 |
* * 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... |
1252 |
VM_BUG_ON(!page_count(page)); |
62695a84e vmscan: move isol... |
1253 1254 1255 1256 |
if (PageLRU(page)) { struct zone *zone = page_zone(page); spin_lock_irq(&zone->lru_lock); |
0c917313a mm: strictly requ... |
1257 |
if (PageLRU(page)) { |
894bc3104 Unevictable LRU I... |
1258 |
int lru = page_lru(page); |
62695a84e vmscan: move isol... |
1259 |
ret = 0; |
0c917313a mm: strictly requ... |
1260 |
get_page(page); |
62695a84e vmscan: move isol... |
1261 |
ClearPageLRU(page); |
4f98a2fee vmscan: split LRU... |
1262 |
|
4f98a2fee vmscan: split LRU... |
1263 |
del_page_from_lru_list(zone, page, lru); |
62695a84e vmscan: move isol... |
1264 1265 1266 1267 1268 |
} spin_unlock_irq(&zone->lru_lock); } return ret; } |
5ad333eb6 Lumpy Reclaim V4 |
1269 |
/* |
35cd78156 vmscan: throttle ... |
1270 1271 1272 1273 1274 1275 1276 1277 1278 |
* 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; |
89b5fae53 mm: vmscan: disti... |
1279 |
if (!global_reclaim(sc)) |
35cd78156 vmscan: throttle ... |
1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 |
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... |
1292 |
static noinline_for_stack void |
3f79768f2 mm: rearrange put... |
1293 1294 |
putback_inactive_pages(struct mem_cgroup_zone *mz, struct list_head *page_list) |
666356297 vmscan: set up pa... |
1295 |
{ |
f16015fbf mm: vmscan: disti... |
1296 |
struct zone_reclaim_stat *reclaim_stat = get_reclaim_stat(mz); |
3f79768f2 mm: rearrange put... |
1297 1298 |
struct zone *zone = mz->zone; LIST_HEAD(pages_to_free); |
666356297 vmscan: set up pa... |
1299 |
|
666356297 vmscan: set up pa... |
1300 1301 1302 |
/* * Put back any unfreeable pages. */ |
666356297 vmscan: set up pa... |
1303 |
while (!list_empty(page_list)) { |
3f79768f2 mm: rearrange put... |
1304 |
struct page *page = lru_to_page(page_list); |
666356297 vmscan: set up pa... |
1305 |
int lru; |
3f79768f2 mm: rearrange put... |
1306 |
|
666356297 vmscan: set up pa... |
1307 1308 1309 1310 1311 1312 1313 1314 |
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... |
1315 |
SetPageLRU(page); |
666356297 vmscan: set up pa... |
1316 |
lru = page_lru(page); |
7a608572a Revert "mm: batch... |
1317 |
add_page_to_lru_list(zone, page, lru); |
666356297 vmscan: set up pa... |
1318 1319 |
if (is_active_lru(lru)) { int file = is_file_lru(lru); |
9992af102 thp: scale nr_rot... |
1320 1321 |
int numpages = hpage_nr_pages(page); reclaim_stat->recent_rotated[file] += numpages; |
666356297 vmscan: set up pa... |
1322 |
} |
2bcf88796 mm: take pagevecs... |
1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 |
if (put_page_testzero(page)) { __ClearPageLRU(page); __ClearPageActive(page); del_page_from_lru_list(zone, page, lru); if (unlikely(PageCompound(page))) { spin_unlock_irq(&zone->lru_lock); (*get_compound_page_dtor(page))(page); spin_lock_irq(&zone->lru_lock); } else list_add(&page->lru, &pages_to_free); |
666356297 vmscan: set up pa... |
1334 1335 |
} } |
666356297 vmscan: set up pa... |
1336 |
|
3f79768f2 mm: rearrange put... |
1337 1338 1339 1340 |
/* * To save our caller's stack, now use input list for pages to free. */ list_splice(&pages_to_free, page_list); |
666356297 vmscan: set up pa... |
1341 |
} |
f16015fbf mm: vmscan: disti... |
1342 1343 |
static noinline_for_stack void update_isolated_counts(struct mem_cgroup_zone *mz, |
3f79768f2 mm: rearrange put... |
1344 |
struct list_head *page_list, |
f16015fbf mm: vmscan: disti... |
1345 |
unsigned long *nr_anon, |
3f79768f2 mm: rearrange put... |
1346 |
unsigned long *nr_file) |
1489fa14c vmscan: update is... |
1347 |
{ |
3f79768f2 mm: rearrange put... |
1348 |
struct zone_reclaim_stat *reclaim_stat = get_reclaim_stat(mz); |
f16015fbf mm: vmscan: disti... |
1349 |
struct zone *zone = mz->zone; |
1489fa14c vmscan: update is... |
1350 |
unsigned int count[NR_LRU_LISTS] = { 0, }; |
3f79768f2 mm: rearrange put... |
1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 |
unsigned long nr_active = 0; struct page *page; int lru; /* * Count pages and clear active flags */ list_for_each_entry(page, page_list, lru) { int numpages = hpage_nr_pages(page); lru = page_lru_base_type(page); if (PageActive(page)) { lru += LRU_ACTIVE; ClearPageActive(page); nr_active += numpages; } count[lru] += numpages; } |
1489fa14c vmscan: update is... |
1368 |
|
1489fa14c vmscan: update is... |
1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 |
__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]; |
1489fa14c vmscan: update is... |
1382 1383 1384 1385 |
reclaim_stat->recent_scanned[0] += *nr_anon; reclaim_stat->recent_scanned[1] += *nr_file; } |
666356297 vmscan: set up pa... |
1386 |
/* |
a18bba061 mm: vmscan: remov... |
1387 |
* Returns true if a direct reclaim should wait on pages under writeback. |
e31f3698c vmscan: raise the... |
1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 |
* * 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... |
1406 |
if (sc->reclaim_mode & RECLAIM_MODE_SINGLE) |
e31f3698c vmscan: raise the... |
1407 |
return false; |
81d66c70b mm/vmscan.c: fix ... |
1408 |
/* If we have reclaimed everything on the isolated list, no stall */ |
e31f3698c vmscan: raise the... |
1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 |
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... |
1427 1428 |
* shrink_inactive_list() is a helper for shrink_zone(). It returns the number * of reclaimed pages |
1da177e4c Linux-2.6.12-rc2 |
1429 |
*/ |
666356297 vmscan: set up pa... |
1430 |
static noinline_for_stack unsigned long |
f16015fbf mm: vmscan: disti... |
1431 1432 |
shrink_inactive_list(unsigned long nr_to_scan, struct mem_cgroup_zone *mz, struct scan_control *sc, int priority, int file) |
1da177e4c Linux-2.6.12-rc2 |
1433 1434 |
{ LIST_HEAD(page_list); |
e247dbce5 vmscan: simplify ... |
1435 |
unsigned long nr_scanned; |
05ff51376 [PATCH] vmscan re... |
1436 |
unsigned long nr_reclaimed = 0; |
e247dbce5 vmscan: simplify ... |
1437 |
unsigned long nr_taken; |
e247dbce5 vmscan: simplify ... |
1438 1439 |
unsigned long nr_anon; unsigned long nr_file; |
92df3a723 mm: vmscan: throt... |
1440 1441 |
unsigned long nr_dirty = 0; unsigned long nr_writeback = 0; |
4356f21d0 mm: change isolat... |
1442 |
isolate_mode_t reclaim_mode = ISOLATE_INACTIVE; |
f16015fbf mm: vmscan: disti... |
1443 |
struct zone *zone = mz->zone; |
78dc583d3 vmscan: low order... |
1444 |
|
35cd78156 vmscan: throttle ... |
1445 |
while (unlikely(too_many_isolated(zone, file, sc))) { |
58355c787 congestion_wait()... |
1446 |
congestion_wait(BLK_RW_ASYNC, HZ/10); |
35cd78156 vmscan: throttle ... |
1447 1448 1449 1450 1451 |
/* We are about to die and free our memory. Return now. */ if (fatal_signal_pending(current)) return SWAP_CLUSTER_MAX; } |
f3a310bc4 mm: vmscan: renam... |
1452 |
set_reclaim_mode(priority, sc, false); |
4356f21d0 mm: change isolat... |
1453 1454 |
if (sc->reclaim_mode & RECLAIM_MODE_LUMPYRECLAIM) reclaim_mode |= ISOLATE_ACTIVE; |
1da177e4c Linux-2.6.12-rc2 |
1455 |
lru_add_drain(); |
f80c06736 mm: zone_reclaim:... |
1456 1457 1458 1459 1460 |
if (!sc->may_unmap) reclaim_mode |= ISOLATE_UNMAPPED; if (!sc->may_writepage) reclaim_mode |= ISOLATE_CLEAN; |
1da177e4c Linux-2.6.12-rc2 |
1461 |
spin_lock_irq(&zone->lru_lock); |
b35ea17b7 mm: shrink_inacti... |
1462 |
|
f626012db mm: remove isolat... |
1463 1464 1465 |
nr_taken = isolate_lru_pages(nr_to_scan, mz, &page_list, &nr_scanned, sc->order, reclaim_mode, 0, file); |
89b5fae53 mm: vmscan: disti... |
1466 |
if (global_reclaim(sc)) { |
e247dbce5 vmscan: simplify ... |
1467 1468 1469 1470 1471 1472 1473 |
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); |
e247dbce5 vmscan: simplify ... |
1474 |
} |
b35ea17b7 mm: shrink_inacti... |
1475 |
|
666356297 vmscan: set up pa... |
1476 1477 1478 1479 |
if (nr_taken == 0) { spin_unlock_irq(&zone->lru_lock); return 0; } |
5ad333eb6 Lumpy Reclaim V4 |
1480 |
|
3f79768f2 mm: rearrange put... |
1481 1482 1483 1484 |
update_isolated_counts(mz, &page_list, &nr_anon, &nr_file); __mod_zone_page_state(zone, NR_ISOLATED_ANON, nr_anon); __mod_zone_page_state(zone, NR_ISOLATED_FILE, nr_file); |
1da177e4c Linux-2.6.12-rc2 |
1485 |
|
e247dbce5 vmscan: simplify ... |
1486 |
spin_unlock_irq(&zone->lru_lock); |
c661b078f synchronous lumpy... |
1487 |
|
f16015fbf mm: vmscan: disti... |
1488 |
nr_reclaimed = shrink_page_list(&page_list, mz, sc, priority, |
92df3a723 mm: vmscan: throt... |
1489 |
&nr_dirty, &nr_writeback); |
c661b078f synchronous lumpy... |
1490 |
|
e31f3698c vmscan: raise the... |
1491 1492 |
/* Check if we should syncronously wait for writeback */ if (should_reclaim_stall(nr_taken, nr_reclaimed, priority, sc)) { |
f3a310bc4 mm: vmscan: renam... |
1493 |
set_reclaim_mode(priority, sc, true); |
f16015fbf mm: vmscan: disti... |
1494 |
nr_reclaimed += shrink_page_list(&page_list, mz, sc, |
92df3a723 mm: vmscan: throt... |
1495 |
priority, &nr_dirty, &nr_writeback); |
e247dbce5 vmscan: simplify ... |
1496 |
} |
b35ea17b7 mm: shrink_inacti... |
1497 |
|
3f79768f2 mm: rearrange put... |
1498 |
spin_lock_irq(&zone->lru_lock); |
e247dbce5 vmscan: simplify ... |
1499 1500 1501 |
if (current_is_kswapd()) __count_vm_events(KSWAPD_STEAL, nr_reclaimed); __count_zone_vm_events(PGSTEAL, zone, nr_reclaimed); |
a74609faf [PATCH] mm: page_... |
1502 |
|
3f79768f2 mm: rearrange put... |
1503 1504 1505 1506 1507 1508 1509 1510 |
putback_inactive_pages(mz, &page_list); __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); free_hot_cold_page_list(&page_list, 1); |
e11da5b4f tracing, vmscan: ... |
1511 |
|
92df3a723 mm: vmscan: throt... |
1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 |
/* * 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: ... |
1537 1538 1539 1540 |
trace_mm_vmscan_lru_shrink_inactive(zone->zone_pgdat->node_id, zone_idx(zone), nr_scanned, nr_reclaimed, priority, |
f3a310bc4 mm: vmscan: renam... |
1541 |
trace_shrink_flags(file, sc->reclaim_mode)); |
05ff51376 [PATCH] vmscan re... |
1542 |
return nr_reclaimed; |
1da177e4c Linux-2.6.12-rc2 |
1543 |
} |
3bb1a852a [PATCH] vmscan: F... |
1544 |
/* |
1da177e4c Linux-2.6.12-rc2 |
1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 |
* 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... |
1561 |
|
3eb4140f0 vmscan: merge dup... |
1562 1563 |
static void move_active_pages_to_lru(struct zone *zone, struct list_head *list, |
2bcf88796 mm: take pagevecs... |
1564 |
struct list_head *pages_to_free, |
3eb4140f0 vmscan: merge dup... |
1565 1566 1567 |
enum lru_list lru) { unsigned long pgmoved = 0; |
3eb4140f0 vmscan: merge dup... |
1568 |
struct page *page; |
2bcf88796 mm: take pagevecs... |
1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 |
if (buffer_heads_over_limit) { spin_unlock_irq(&zone->lru_lock); list_for_each_entry(page, list, lru) { if (page_has_private(page) && trylock_page(page)) { if (page_has_private(page)) try_to_release_page(page, 0); unlock_page(page); } } spin_lock_irq(&zone->lru_lock); } |
3eb4140f0 vmscan: merge dup... |
1580 1581 |
while (!list_empty(list)) { |
925b7673c mm: make per-memc... |
1582 |
struct lruvec *lruvec; |
3eb4140f0 vmscan: merge dup... |
1583 |
page = lru_to_page(list); |
3eb4140f0 vmscan: merge dup... |
1584 1585 1586 |
VM_BUG_ON(PageLRU(page)); SetPageLRU(page); |
925b7673c mm: make per-memc... |
1587 1588 |
lruvec = mem_cgroup_lru_add_list(zone, page, lru); list_move(&page->lru, &lruvec->lists[lru]); |
2c888cfbc thp: fix anon mem... |
1589 |
pgmoved += hpage_nr_pages(page); |
3eb4140f0 vmscan: merge dup... |
1590 |
|
2bcf88796 mm: take pagevecs... |
1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 |
if (put_page_testzero(page)) { __ClearPageLRU(page); __ClearPageActive(page); del_page_from_lru_list(zone, page, lru); if (unlikely(PageCompound(page))) { spin_unlock_irq(&zone->lru_lock); (*get_compound_page_dtor(page))(page); spin_lock_irq(&zone->lru_lock); } else list_add(&page->lru, pages_to_free); |
3eb4140f0 vmscan: merge dup... |
1602 1603 1604 1605 1606 1607 |
} } __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... |
1608 |
|
f626012db mm: remove isolat... |
1609 |
static void shrink_active_list(unsigned long nr_to_scan, |
f16015fbf mm: vmscan: disti... |
1610 1611 1612 |
struct mem_cgroup_zone *mz, struct scan_control *sc, int priority, int file) |
1da177e4c Linux-2.6.12-rc2 |
1613 |
{ |
44c241f16 mm: rename pgmove... |
1614 |
unsigned long nr_taken; |
f626012db mm: remove isolat... |
1615 |
unsigned long nr_scanned; |
6fe6b7e35 vmscan: report vm... |
1616 |
unsigned long vm_flags; |
1da177e4c Linux-2.6.12-rc2 |
1617 |
LIST_HEAD(l_hold); /* The pages which were snipped off */ |
8cab4754d vmscan: make mapp... |
1618 |
LIST_HEAD(l_active); |
b69408e88 vmscan: Use an in... |
1619 |
LIST_HEAD(l_inactive); |
1da177e4c Linux-2.6.12-rc2 |
1620 |
struct page *page; |
f16015fbf mm: vmscan: disti... |
1621 |
struct zone_reclaim_stat *reclaim_stat = get_reclaim_stat(mz); |
44c241f16 mm: rename pgmove... |
1622 |
unsigned long nr_rotated = 0; |
f80c06736 mm: zone_reclaim:... |
1623 |
isolate_mode_t reclaim_mode = ISOLATE_ACTIVE; |
f16015fbf mm: vmscan: disti... |
1624 |
struct zone *zone = mz->zone; |
1da177e4c Linux-2.6.12-rc2 |
1625 1626 |
lru_add_drain(); |
f80c06736 mm: zone_reclaim:... |
1627 1628 1629 1630 1631 |
if (!sc->may_unmap) reclaim_mode |= ISOLATE_UNMAPPED; if (!sc->may_writepage) reclaim_mode |= ISOLATE_CLEAN; |
1da177e4c Linux-2.6.12-rc2 |
1632 |
spin_lock_irq(&zone->lru_lock); |
925b7673c mm: make per-memc... |
1633 |
|
f626012db mm: remove isolat... |
1634 1635 1636 |
nr_taken = isolate_lru_pages(nr_to_scan, mz, &l_hold, &nr_scanned, sc->order, reclaim_mode, 1, file); |
89b5fae53 mm: vmscan: disti... |
1637 |
if (global_reclaim(sc)) |
f626012db mm: remove isolat... |
1638 |
zone->pages_scanned += nr_scanned; |
89b5fae53 mm: vmscan: disti... |
1639 |
|
b7c46d151 mm: drop unneeded... |
1640 |
reclaim_stat->recent_scanned[file] += nr_taken; |
1cfb419b3 per-zone and recl... |
1641 |
|
f626012db mm: remove isolat... |
1642 |
__count_zone_vm_events(PGREFILL, zone, nr_scanned); |
4f98a2fee vmscan: split LRU... |
1643 |
if (file) |
44c241f16 mm: rename pgmove... |
1644 |
__mod_zone_page_state(zone, NR_ACTIVE_FILE, -nr_taken); |
4f98a2fee vmscan: split LRU... |
1645 |
else |
44c241f16 mm: rename pgmove... |
1646 |
__mod_zone_page_state(zone, NR_ACTIVE_ANON, -nr_taken); |
a731286de mm: vmstat: add i... |
1647 |
__mod_zone_page_state(zone, NR_ISOLATED_ANON + file, nr_taken); |
1da177e4c Linux-2.6.12-rc2 |
1648 |
spin_unlock_irq(&zone->lru_lock); |
1da177e4c Linux-2.6.12-rc2 |
1649 1650 1651 1652 |
while (!list_empty(&l_hold)) { cond_resched(); page = lru_to_page(&l_hold); list_del(&page->lru); |
7e9cd4842 vmscan: fix pagec... |
1653 |
|
894bc3104 Unevictable LRU I... |
1654 1655 1656 1657 |
if (unlikely(!page_evictable(page, NULL))) { putback_lru_page(page); continue; } |
f16015fbf mm: vmscan: disti... |
1658 |
if (page_referenced(page, 0, mz->mem_cgroup, &vm_flags)) { |
9992af102 thp: scale nr_rot... |
1659 |
nr_rotated += hpage_nr_pages(page); |
8cab4754d vmscan: make mapp... |
1660 1661 1662 1663 1664 1665 1666 1667 1668 |
/* * 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_... |
1669 |
if ((vm_flags & VM_EXEC) && page_is_file_cache(page)) { |
8cab4754d vmscan: make mapp... |
1670 1671 1672 1673 |
list_add(&page->lru, &l_active); continue; } } |
7e9cd4842 vmscan: fix pagec... |
1674 |
|
5205e56ee vmscan: move Clea... |
1675 |
ClearPageActive(page); /* we are de-activating */ |
1da177e4c Linux-2.6.12-rc2 |
1676 1677 |
list_add(&page->lru, &l_inactive); } |
b555749aa vmscan: shrink_ac... |
1678 |
/* |
8cab4754d vmscan: make mapp... |
1679 |
* Move pages back to the lru list. |
b555749aa vmscan: shrink_ac... |
1680 |
*/ |
2a1dc5097 vmscan: protect z... |
1681 |
spin_lock_irq(&zone->lru_lock); |
4f98a2fee vmscan: split LRU... |
1682 |
/* |
8cab4754d vmscan: make mapp... |
1683 1684 1685 1686 |
* 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... |
1687 |
*/ |
b7c46d151 mm: drop unneeded... |
1688 |
reclaim_stat->recent_rotated[file] += nr_rotated; |
556adecba vmscan: second ch... |
1689 |
|
2bcf88796 mm: take pagevecs... |
1690 |
move_active_pages_to_lru(zone, &l_active, &l_hold, |
3eb4140f0 vmscan: merge dup... |
1691 |
LRU_ACTIVE + file * LRU_FILE); |
2bcf88796 mm: take pagevecs... |
1692 |
move_active_pages_to_lru(zone, &l_inactive, &l_hold, |
3eb4140f0 vmscan: merge dup... |
1693 |
LRU_BASE + file * LRU_FILE); |
a731286de mm: vmstat: add i... |
1694 |
__mod_zone_page_state(zone, NR_ISOLATED_ANON + file, -nr_taken); |
f8891e5e1 [PATCH] Light wei... |
1695 |
spin_unlock_irq(&zone->lru_lock); |
2bcf88796 mm: take pagevecs... |
1696 1697 |
free_hot_cold_page_list(&l_hold, 1); |
1da177e4c Linux-2.6.12-rc2 |
1698 |
} |
74e3f3c33 vmscan: prevent b... |
1699 |
#ifdef CONFIG_SWAP |
14797e236 memcg: add inacti... |
1700 |
static int inactive_anon_is_low_global(struct zone *zone) |
f89eb90e3 inactive_anon_is_... |
1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 |
{ 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... |
1712 1713 1714 1715 1716 1717 1718 1719 |
/** * 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. */ |
f16015fbf mm: vmscan: disti... |
1720 |
static int inactive_anon_is_low(struct mem_cgroup_zone *mz) |
14797e236 memcg: add inacti... |
1721 |
{ |
74e3f3c33 vmscan: prevent b... |
1722 1723 1724 1725 1726 1727 |
/* * If we don't have swap space, anonymous page deactivation * is pointless. */ if (!total_swap_pages) return 0; |
f16015fbf mm: vmscan: disti... |
1728 1729 1730 1731 1732 |
if (!scanning_global_lru(mz)) return mem_cgroup_inactive_anon_is_low(mz->mem_cgroup, mz->zone); return inactive_anon_is_low_global(mz->zone); |
14797e236 memcg: add inacti... |
1733 |
} |
74e3f3c33 vmscan: prevent b... |
1734 |
#else |
f16015fbf mm: vmscan: disti... |
1735 |
static inline int inactive_anon_is_low(struct mem_cgroup_zone *mz) |
74e3f3c33 vmscan: prevent b... |
1736 1737 1738 1739 |
{ return 0; } #endif |
14797e236 memcg: add inacti... |
1740 |
|
56e49d218 vmscan: evict use... |
1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 |
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 |
f16015fbf mm: vmscan: disti... |
1753 |
* @mz: memory cgroup and zone to check |
56e49d218 vmscan: evict use... |
1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 |
* * 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. */ |
f16015fbf mm: vmscan: disti... |
1765 |
static int inactive_file_is_low(struct mem_cgroup_zone *mz) |
56e49d218 vmscan: evict use... |
1766 |
{ |
f16015fbf mm: vmscan: disti... |
1767 1768 1769 |
if (!scanning_global_lru(mz)) return mem_cgroup_inactive_file_is_low(mz->mem_cgroup, mz->zone); |
56e49d218 vmscan: evict use... |
1770 |
|
f16015fbf mm: vmscan: disti... |
1771 |
return inactive_file_is_low_global(mz->zone); |
56e49d218 vmscan: evict use... |
1772 |
} |
f16015fbf mm: vmscan: disti... |
1773 |
static int inactive_list_is_low(struct mem_cgroup_zone *mz, int file) |
b39415b27 vmscan: do not ev... |
1774 1775 |
{ if (file) |
f16015fbf mm: vmscan: disti... |
1776 |
return inactive_file_is_low(mz); |
b39415b27 vmscan: do not ev... |
1777 |
else |
f16015fbf mm: vmscan: disti... |
1778 |
return inactive_anon_is_low(mz); |
b39415b27 vmscan: do not ev... |
1779 |
} |
4f98a2fee vmscan: split LRU... |
1780 |
static unsigned long shrink_list(enum lru_list lru, unsigned long nr_to_scan, |
f16015fbf mm: vmscan: disti... |
1781 1782 |
struct mem_cgroup_zone *mz, struct scan_control *sc, int priority) |
b69408e88 vmscan: Use an in... |
1783 |
{ |
4f98a2fee vmscan: split LRU... |
1784 |
int file = is_file_lru(lru); |
b39415b27 vmscan: do not ev... |
1785 |
if (is_active_lru(lru)) { |
f16015fbf mm: vmscan: disti... |
1786 1787 |
if (inactive_list_is_low(mz, file)) shrink_active_list(nr_to_scan, mz, sc, priority, file); |
556adecba vmscan: second ch... |
1788 1789 |
return 0; } |
f16015fbf mm: vmscan: disti... |
1790 |
return shrink_inactive_list(nr_to_scan, mz, sc, priority, file); |
4f98a2fee vmscan: split LRU... |
1791 |
} |
f16015fbf mm: vmscan: disti... |
1792 1793 |
static int vmscan_swappiness(struct mem_cgroup_zone *mz, struct scan_control *sc) |
1f4c025b5 memcg: export mem... |
1794 |
{ |
89b5fae53 mm: vmscan: disti... |
1795 |
if (global_reclaim(sc)) |
1f4c025b5 memcg: export mem... |
1796 |
return vm_swappiness; |
f16015fbf mm: vmscan: disti... |
1797 |
return mem_cgroup_swappiness(mz->mem_cgroup); |
1f4c025b5 memcg: export mem... |
1798 |
} |
4f98a2fee vmscan: split LRU... |
1799 1800 1801 1802 1803 1804 |
/* * 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... |
1805 |
* nr[0] = anon pages to scan; nr[1] = file pages to scan |
4f98a2fee vmscan: split LRU... |
1806 |
*/ |
f16015fbf mm: vmscan: disti... |
1807 1808 |
static void get_scan_count(struct mem_cgroup_zone *mz, struct scan_control *sc, unsigned long *nr, int priority) |
4f98a2fee vmscan: split LRU... |
1809 1810 1811 1812 |
{ unsigned long anon, file, free; unsigned long anon_prio, file_prio; unsigned long ap, fp; |
f16015fbf mm: vmscan: disti... |
1813 |
struct zone_reclaim_stat *reclaim_stat = get_reclaim_stat(mz); |
76a33fc38 vmscan: prevent g... |
1814 |
u64 fraction[2], denominator; |
4111304da mm: enum lru_list... |
1815 |
enum lru_list lru; |
76a33fc38 vmscan: prevent g... |
1816 |
int noswap = 0; |
a4d3e9e76 mm: vmscan: fix f... |
1817 |
bool force_scan = false; |
246e87a93 memcg: fix get_sc... |
1818 |
|
f11c0ca50 mm: vmscan: drop ... |
1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 |
/* * 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. */ |
b95a2f2d4 mm: vmscan: conve... |
1829 |
if (current_is_kswapd() && mz->zone->all_unreclaimable) |
a4d3e9e76 mm: vmscan: fix f... |
1830 |
force_scan = true; |
89b5fae53 mm: vmscan: disti... |
1831 |
if (!global_reclaim(sc)) |
a4d3e9e76 mm: vmscan: fix f... |
1832 |
force_scan = true; |
76a33fc38 vmscan: prevent g... |
1833 1834 1835 1836 1837 1838 1839 1840 1841 |
/* 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... |
1842 |
|
f16015fbf mm: vmscan: disti... |
1843 1844 1845 1846 |
anon = zone_nr_lru_pages(mz, LRU_ACTIVE_ANON) + zone_nr_lru_pages(mz, LRU_INACTIVE_ANON); file = zone_nr_lru_pages(mz, LRU_ACTIVE_FILE) + zone_nr_lru_pages(mz, LRU_INACTIVE_FILE); |
a4d3e9e76 mm: vmscan: fix f... |
1847 |
|
89b5fae53 mm: vmscan: disti... |
1848 |
if (global_reclaim(sc)) { |
f16015fbf mm: vmscan: disti... |
1849 |
free = zone_page_state(mz->zone, NR_FREE_PAGES); |
eeee9a8cd mm: make get_scan... |
1850 1851 |
/* If we have very few page cache pages, force-scan anon pages. */ |
f16015fbf mm: vmscan: disti... |
1852 |
if (unlikely(file + free <= high_wmark_pages(mz->zone))) { |
76a33fc38 vmscan: prevent g... |
1853 1854 1855 1856 |
fraction[0] = 1; fraction[1] = 0; denominator = 1; goto out; |
eeee9a8cd mm: make get_scan... |
1857 |
} |
4f98a2fee vmscan: split LRU... |
1858 1859 1860 |
} /* |
58c37f6e0 vmscan: protect r... |
1861 1862 1863 |
* With swappiness at 100, anonymous and file have the same priority. * This scanning priority is essentially the inverse of IO cost. */ |
f16015fbf mm: vmscan: disti... |
1864 1865 |
anon_prio = vmscan_swappiness(mz, sc); file_prio = 200 - vmscan_swappiness(mz, sc); |
58c37f6e0 vmscan: protect r... |
1866 1867 |
/* |
4f98a2fee vmscan: split LRU... |
1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 |
* 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] */ |
f16015fbf mm: vmscan: disti... |
1878 |
spin_lock_irq(&mz->zone->lru_lock); |
6e9015716 mm: introduce zon... |
1879 |
if (unlikely(reclaim_stat->recent_scanned[0] > anon / 4)) { |
6e9015716 mm: introduce zon... |
1880 1881 |
reclaim_stat->recent_scanned[0] /= 2; reclaim_stat->recent_rotated[0] /= 2; |
4f98a2fee vmscan: split LRU... |
1882 |
} |
6e9015716 mm: introduce zon... |
1883 |
if (unlikely(reclaim_stat->recent_scanned[1] > file / 4)) { |
6e9015716 mm: introduce zon... |
1884 1885 |
reclaim_stat->recent_scanned[1] /= 2; reclaim_stat->recent_rotated[1] /= 2; |
4f98a2fee vmscan: split LRU... |
1886 1887 1888 |
} /* |
00d8089c5 vmscan: fix get_s... |
1889 1890 1891 |
* 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... |
1892 |
*/ |
6e9015716 mm: introduce zon... |
1893 1894 |
ap = (anon_prio + 1) * (reclaim_stat->recent_scanned[0] + 1); ap /= reclaim_stat->recent_rotated[0] + 1; |
4f98a2fee vmscan: split LRU... |
1895 |
|
6e9015716 mm: introduce zon... |
1896 1897 |
fp = (file_prio + 1) * (reclaim_stat->recent_scanned[1] + 1); fp /= reclaim_stat->recent_rotated[1] + 1; |
f16015fbf mm: vmscan: disti... |
1898 |
spin_unlock_irq(&mz->zone->lru_lock); |
4f98a2fee vmscan: split LRU... |
1899 |
|
76a33fc38 vmscan: prevent g... |
1900 1901 1902 1903 |
fraction[0] = ap; fraction[1] = fp; denominator = ap + fp + 1; out: |
4111304da mm: enum lru_list... |
1904 1905 |
for_each_evictable_lru(lru) { int file = is_file_lru(lru); |
76a33fc38 vmscan: prevent g... |
1906 |
unsigned long scan; |
6e08a369e vmscan: cleanup t... |
1907 |
|
4111304da mm: enum lru_list... |
1908 |
scan = zone_nr_lru_pages(mz, lru); |
76a33fc38 vmscan: prevent g... |
1909 1910 |
if (priority || noswap) { scan >>= priority; |
f11c0ca50 mm: vmscan: drop ... |
1911 1912 |
if (!scan && force_scan) scan = SWAP_CLUSTER_MAX; |
76a33fc38 vmscan: prevent g... |
1913 1914 |
scan = div64_u64(scan * fraction[file], denominator); } |
4111304da mm: enum lru_list... |
1915 |
nr[lru] = scan; |
76a33fc38 vmscan: prevent g... |
1916 |
} |
6e08a369e vmscan: cleanup t... |
1917 |
} |
4f98a2fee vmscan: split LRU... |
1918 |
|
1da177e4c Linux-2.6.12-rc2 |
1919 |
/* |
3e7d34497 mm: vmscan: recla... |
1920 1921 1922 1923 1924 1925 |
* 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 */ |
f16015fbf mm: vmscan: disti... |
1926 |
static inline bool should_continue_reclaim(struct mem_cgroup_zone *mz, |
3e7d34497 mm: vmscan: recla... |
1927 1928 1929 1930 1931 1932 1933 1934 |
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... |
1935 |
if (!(sc->reclaim_mode & RECLAIM_MODE_COMPACTION)) |
3e7d34497 mm: vmscan: recla... |
1936 |
return false; |
2876592f2 mm: vmscan: stop ... |
1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 |
/* 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... |
1959 1960 1961 1962 1963 1964 |
/* * If we have not reclaimed enough pages for compaction and the * inactive lists are large enough, continue reclaiming */ pages_for_compaction = (2UL << sc->order); |
f16015fbf mm: vmscan: disti... |
1965 |
inactive_lru_pages = zone_nr_lru_pages(mz, LRU_INACTIVE_FILE); |
86cfd3a45 mm/vmscan.c: cons... |
1966 |
if (nr_swap_pages > 0) |
f16015fbf mm: vmscan: disti... |
1967 |
inactive_lru_pages += zone_nr_lru_pages(mz, LRU_INACTIVE_ANON); |
3e7d34497 mm: vmscan: recla... |
1968 1969 1970 1971 1972 |
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 */ |
f16015fbf mm: vmscan: disti... |
1973 |
switch (compaction_suitable(mz->zone, sc->order)) { |
3e7d34497 mm: vmscan: recla... |
1974 1975 1976 1977 1978 1979 1980 1981 1982 |
case COMPACT_PARTIAL: case COMPACT_CONTINUE: return false; default: return true; } } /* |
1da177e4c Linux-2.6.12-rc2 |
1983 1984 |
* This is a basic per-zone page freer. Used by both kswapd and direct reclaim. */ |
f16015fbf mm: vmscan: disti... |
1985 1986 |
static void shrink_mem_cgroup_zone(int priority, struct mem_cgroup_zone *mz, struct scan_control *sc) |
1da177e4c Linux-2.6.12-rc2 |
1987 |
{ |
b69408e88 vmscan: Use an in... |
1988 |
unsigned long nr[NR_LRU_LISTS]; |
8695949a1 [PATCH] Thin out ... |
1989 |
unsigned long nr_to_scan; |
4111304da mm: enum lru_list... |
1990 |
enum lru_list lru; |
f0fdc5e8e vmscan: fix zone ... |
1991 |
unsigned long nr_reclaimed, nr_scanned; |
22fba3354 vmscan: separate ... |
1992 |
unsigned long nr_to_reclaim = sc->nr_to_reclaim; |
3da367c3e vmscan: add block... |
1993 |
struct blk_plug plug; |
e0f79b8f1 vmscan: don't acc... |
1994 |
|
3e7d34497 mm: vmscan: recla... |
1995 1996 |
restart: nr_reclaimed = 0; |
f0fdc5e8e vmscan: fix zone ... |
1997 |
nr_scanned = sc->nr_scanned; |
f16015fbf mm: vmscan: disti... |
1998 |
get_scan_count(mz, sc, nr, priority); |
1da177e4c Linux-2.6.12-rc2 |
1999 |
|
3da367c3e vmscan: add block... |
2000 |
blk_start_plug(&plug); |
556adecba vmscan: second ch... |
2001 2002 |
while (nr[LRU_INACTIVE_ANON] || nr[LRU_ACTIVE_FILE] || nr[LRU_INACTIVE_FILE]) { |
4111304da mm: enum lru_list... |
2003 2004 |
for_each_evictable_lru(lru) { if (nr[lru]) { |
ece74b2e7 vmscan: kill sc.s... |
2005 |
nr_to_scan = min_t(unsigned long, |
4111304da mm: enum lru_list... |
2006 2007 |
nr[lru], SWAP_CLUSTER_MAX); nr[lru] -= nr_to_scan; |
1da177e4c Linux-2.6.12-rc2 |
2008 |
|
4111304da mm: enum lru_list... |
2009 |
nr_reclaimed += shrink_list(lru, nr_to_scan, |
f16015fbf mm: vmscan: disti... |
2010 |
mz, sc, priority); |
b69408e88 vmscan: Use an in... |
2011 |
} |
1da177e4c Linux-2.6.12-rc2 |
2012 |
} |
a79311c14 vmscan: bail out ... |
2013 2014 2015 2016 2017 2018 2019 2020 |
/* * 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... |
2021 |
if (nr_reclaimed >= nr_to_reclaim && priority < DEF_PRIORITY) |
a79311c14 vmscan: bail out ... |
2022 |
break; |
1da177e4c Linux-2.6.12-rc2 |
2023 |
} |
3da367c3e vmscan: add block... |
2024 |
blk_finish_plug(&plug); |
3e7d34497 mm: vmscan: recla... |
2025 |
sc->nr_reclaimed += nr_reclaimed; |
01dbe5c9b vmscan: improve r... |
2026 |
|
556adecba vmscan: second ch... |
2027 2028 2029 2030 |
/* * Even if we did not try to evict anon pages at all, we want to * rebalance the anon lru active/inactive ratio. */ |
f16015fbf mm: vmscan: disti... |
2031 2032 |
if (inactive_anon_is_low(mz)) shrink_active_list(SWAP_CLUSTER_MAX, mz, sc, priority, 0); |
556adecba vmscan: second ch... |
2033 |
|
3e7d34497 mm: vmscan: recla... |
2034 |
/* reclaim/compaction might need reclaim to continue */ |
f16015fbf mm: vmscan: disti... |
2035 |
if (should_continue_reclaim(mz, nr_reclaimed, |
3e7d34497 mm: vmscan: recla... |
2036 2037 |
sc->nr_scanned - nr_scanned, sc)) goto restart; |
232ea4d69 [PATCH] throttle_... |
2038 |
throttle_vm_writeout(sc->gfp_mask); |
1da177e4c Linux-2.6.12-rc2 |
2039 |
} |
f16015fbf mm: vmscan: disti... |
2040 2041 2042 |
static void shrink_zone(int priority, struct zone *zone, struct scan_control *sc) { |
5660048cc mm: move memcg hi... |
2043 2044 |
struct mem_cgroup *root = sc->target_mem_cgroup; struct mem_cgroup_reclaim_cookie reclaim = { |
f16015fbf mm: vmscan: disti... |
2045 |
.zone = zone, |
5660048cc mm: move memcg hi... |
2046 |
.priority = priority, |
f16015fbf mm: vmscan: disti... |
2047 |
}; |
5660048cc mm: move memcg hi... |
2048 |
struct mem_cgroup *memcg; |
5660048cc mm: move memcg hi... |
2049 2050 2051 2052 2053 2054 |
memcg = mem_cgroup_iter(root, NULL, &reclaim); do { struct mem_cgroup_zone mz = { .mem_cgroup = memcg, .zone = zone, }; |
f16015fbf mm: vmscan: disti... |
2055 |
|
5660048cc mm: move memcg hi... |
2056 2057 2058 2059 2060 2061 |
shrink_mem_cgroup_zone(priority, &mz, sc); /* * Limit reclaim has historically picked one memcg and * scanned it with decreasing priority levels until * nr_to_reclaim had been reclaimed. This priority * cycle is thus over after a single memcg. |
b95a2f2d4 mm: vmscan: conve... |
2062 2063 2064 2065 |
* * Direct reclaim and kswapd, on the other hand, have * to scan all memory cgroups to fulfill the overall * scan target for the zone. |
5660048cc mm: move memcg hi... |
2066 2067 2068 2069 2070 2071 2072 |
*/ if (!global_reclaim(sc)) { mem_cgroup_iter_break(root, memcg); break; } memcg = mem_cgroup_iter(root, memcg, &reclaim); } while (memcg); |
f16015fbf mm: vmscan: disti... |
2073 |
} |
fe4b1b244 mm: vmscan: when ... |
2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 |
/* Returns true if compaction should go ahead for a high-order request */ static inline bool compaction_ready(struct zone *zone, struct scan_control *sc) { unsigned long balance_gap, watermark; bool watermark_ok; /* Do not consider compaction for orders reclaim is meant to satisfy */ if (sc->order <= PAGE_ALLOC_COSTLY_ORDER) return false; /* * Compaction takes time to run and there are potentially other * callers using the pages just freed. Continue reclaiming until * there is a buffer of free pages available to give compaction * a reasonable chance of completing and allocating the page */ balance_gap = min(low_wmark_pages(zone), (zone->present_pages + KSWAPD_ZONE_BALANCE_GAP_RATIO-1) / KSWAPD_ZONE_BALANCE_GAP_RATIO); watermark = high_wmark_pages(zone) + balance_gap + (2UL << sc->order); watermark_ok = zone_watermark_ok_safe(zone, 0, watermark, 0, 0); /* * If compaction is deferred, reclaim up to a point where * compaction will have a chance of success when re-enabled */ if (compaction_deferred(zone)) return watermark_ok; /* If compaction is not ready to start, keep reclaiming */ if (!compaction_suitable(zone, sc->order)) return false; return watermark_ok; } |
1da177e4c Linux-2.6.12-rc2 |
2109 2110 2111 2112 2113 |
/* * 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... |
2114 2115 |
* We reclaim from a zone even if that zone is over high_wmark_pages(zone). * Because: |
1da177e4c Linux-2.6.12-rc2 |
2116 2117 |
* a) The caller may be trying to free *extra* pages to satisfy a higher-order * allocation or |
418589663 page allocator: u... |
2118 2119 2120 |
* 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 |
2121 |
* |
1da177e4c Linux-2.6.12-rc2 |
2122 2123 |
* 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... |
2124 2125 |
* * This function returns true if a zone is being reclaimed for a costly |
fe4b1b244 mm: vmscan: when ... |
2126 |
* high-order allocation and compaction is ready to begin. This indicates to |
0cee34fd7 mm: vmscan: check... |
2127 2128 |
* the caller that it should consider retrying the allocation instead of * further reclaim. |
1da177e4c Linux-2.6.12-rc2 |
2129 |
*/ |
e0c23279c vmscan: abort rec... |
2130 |
static bool shrink_zones(int priority, struct zonelist *zonelist, |
05ff51376 [PATCH] vmscan re... |
2131 |
struct scan_control *sc) |
1da177e4c Linux-2.6.12-rc2 |
2132 |
{ |
dd1a239f6 mm: have zonelist... |
2133 |
struct zoneref *z; |
54a6eb5c4 mm: use two zonel... |
2134 |
struct zone *zone; |
d149e3b25 memcg: add the so... |
2135 2136 |
unsigned long nr_soft_reclaimed; unsigned long nr_soft_scanned; |
0cee34fd7 mm: vmscan: check... |
2137 |
bool aborted_reclaim = false; |
1cfb419b3 per-zone and recl... |
2138 |
|
d4debc66d vmscan: remove un... |
2139 2140 |
for_each_zone_zonelist_nodemask(zone, z, zonelist, gfp_zone(sc->gfp_mask), sc->nodemask) { |
f3fe65122 [PATCH] mm: add p... |
2141 |
if (!populated_zone(zone)) |
1da177e4c Linux-2.6.12-rc2 |
2142 |
continue; |
1cfb419b3 per-zone and recl... |
2143 2144 2145 2146 |
/* * Take care memory controller reclaiming has small influence * to global LRU. */ |
89b5fae53 mm: vmscan: disti... |
2147 |
if (global_reclaim(sc)) { |
1cfb419b3 per-zone and recl... |
2148 2149 |
if (!cpuset_zone_allowed_hardwall(zone, GFP_KERNEL)) continue; |
93e4a89a8 mm: restore zone-... |
2150 |
if (zone->all_unreclaimable && priority != DEF_PRIORITY) |
1cfb419b3 per-zone and recl... |
2151 |
continue; /* Let kswapd poll it */ |
e0887c19b vmscan: limit dir... |
2152 2153 |
if (COMPACTION_BUILD) { /* |
e0c23279c vmscan: abort rec... |
2154 2155 2156 2157 2158 2159 2160 |
* 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... |
2161 |
*/ |
fe4b1b244 mm: vmscan: when ... |
2162 |
if (compaction_ready(zone, sc)) { |
0cee34fd7 mm: vmscan: check... |
2163 |
aborted_reclaim = true; |
e0887c19b vmscan: limit dir... |
2164 |
continue; |
e0c23279c vmscan: abort rec... |
2165 |
} |
e0887c19b vmscan: limit dir... |
2166 |
} |
ac34a1a3c memcg: fix direct... |
2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 |
/* * 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... |
2180 |
} |
408d85441 [PATCH] oom: use ... |
2181 |
|
a79311c14 vmscan: bail out ... |
2182 |
shrink_zone(priority, zone, sc); |
1da177e4c Linux-2.6.12-rc2 |
2183 |
} |
e0c23279c vmscan: abort rec... |
2184 |
|
0cee34fd7 mm: vmscan: check... |
2185 |
return aborted_reclaim; |
d1908362a vmscan: check all... |
2186 2187 2188 2189 2190 2191 |
} static bool zone_reclaimable(struct zone *zone) { return zone->pages_scanned < zone_reclaimable_pages(zone) * 6; } |
929bea7c7 vmscan: all_unrec... |
2192 |
/* All zones in zonelist are unreclaimable? */ |
d1908362a vmscan: check all... |
2193 2194 2195 2196 2197 |
static bool all_unreclaimable(struct zonelist *zonelist, struct scan_control *sc) { struct zoneref *z; struct zone *zone; |
d1908362a vmscan: check all... |
2198 2199 2200 2201 2202 2203 2204 |
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... |
2205 2206 |
if (!zone->all_unreclaimable) return false; |
d1908362a vmscan: check all... |
2207 |
} |
929bea7c7 vmscan: all_unrec... |
2208 |
return true; |
1da177e4c Linux-2.6.12-rc2 |
2209 |
} |
4f98a2fee vmscan: split LRU... |
2210 |
|
1da177e4c Linux-2.6.12-rc2 |
2211 2212 2213 2214 2215 2216 2217 2218 |
/* * 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... |
2219 2220 2221 2222 |
* 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... |
2223 2224 2225 |
* * returns: 0, if no pages reclaimed * else, the number of pages reclaimed |
1da177e4c Linux-2.6.12-rc2 |
2226 |
*/ |
dac1d27bc mm: use zonelists... |
2227 |
static unsigned long do_try_to_free_pages(struct zonelist *zonelist, |
a09ed5e00 vmscan: change sh... |
2228 2229 |
struct scan_control *sc, struct shrink_control *shrink) |
1da177e4c Linux-2.6.12-rc2 |
2230 2231 |
{ int priority; |
69e05944a [PATCH] vmscan: u... |
2232 |
unsigned long total_scanned = 0; |
1da177e4c Linux-2.6.12-rc2 |
2233 |
struct reclaim_state *reclaim_state = current->reclaim_state; |
dd1a239f6 mm: have zonelist... |
2234 |
struct zoneref *z; |
54a6eb5c4 mm: use two zonel... |
2235 |
struct zone *zone; |
22fba3354 vmscan: separate ... |
2236 |
unsigned long writeback_threshold; |
0cee34fd7 mm: vmscan: check... |
2237 |
bool aborted_reclaim; |
1da177e4c Linux-2.6.12-rc2 |
2238 |
|
c0ff7453b cpuset,mm: fix no... |
2239 |
get_mems_allowed(); |
873b47717 per-task-delay-ac... |
2240 |
delayacct_freepages_start(); |
89b5fae53 mm: vmscan: disti... |
2241 |
if (global_reclaim(sc)) |
1cfb419b3 per-zone and recl... |
2242 |
count_vm_event(ALLOCSTALL); |
1da177e4c Linux-2.6.12-rc2 |
2243 2244 |
for (priority = DEF_PRIORITY; priority >= 0; priority--) { |
66e1707bc Memory controller... |
2245 |
sc->nr_scanned = 0; |
f7b7fd8f3 [PATCH] temporari... |
2246 |
if (!priority) |
f16015fbf mm: vmscan: disti... |
2247 |
disable_swap_token(sc->target_mem_cgroup); |
0cee34fd7 mm: vmscan: check... |
2248 |
aborted_reclaim = shrink_zones(priority, zonelist, sc); |
e0c23279c vmscan: abort rec... |
2249 |
|
66e1707bc Memory controller... |
2250 2251 2252 2253 |
/* * Don't shrink slabs when reclaiming memory from * over limit cgroups */ |
89b5fae53 mm: vmscan: disti... |
2254 |
if (global_reclaim(sc)) { |
c6a8a8c58 vmscan: recalcula... |
2255 |
unsigned long lru_pages = 0; |
d4debc66d vmscan: remove un... |
2256 2257 |
for_each_zone_zonelist(zone, z, zonelist, gfp_zone(sc->gfp_mask)) { |
c6a8a8c58 vmscan: recalcula... |
2258 2259 2260 2261 2262 |
if (!cpuset_zone_allowed_hardwall(zone, GFP_KERNEL)) continue; lru_pages += zone_reclaimable_pages(zone); } |
1495f230f vmscan: change sh... |
2263 |
shrink_slab(shrink, sc->nr_scanned, lru_pages); |
91a45470f per-zone and recl... |
2264 |
if (reclaim_state) { |
a79311c14 vmscan: bail out ... |
2265 |
sc->nr_reclaimed += reclaim_state->reclaimed_slab; |
91a45470f per-zone and recl... |
2266 2267 |
reclaim_state->reclaimed_slab = 0; } |
1da177e4c Linux-2.6.12-rc2 |
2268 |
} |
66e1707bc Memory controller... |
2269 |
total_scanned += sc->nr_scanned; |
bb21c7ce1 vmscan: fix do_tr... |
2270 |
if (sc->nr_reclaimed >= sc->nr_to_reclaim) |
1da177e4c Linux-2.6.12-rc2 |
2271 |
goto out; |
1da177e4c Linux-2.6.12-rc2 |
2272 2273 2274 2275 2276 2277 2278 2279 |
/* * 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 ... |
2280 2281 |
writeback_threshold = sc->nr_to_reclaim + sc->nr_to_reclaim / 2; if (total_scanned > writeback_threshold) { |
0e175a183 writeback: Add a ... |
2282 2283 |
wakeup_flusher_threads(laptop_mode ? 0 : total_scanned, WB_REASON_TRY_TO_FREE_PAGES); |
66e1707bc Memory controller... |
2284 |
sc->may_writepage = 1; |
1da177e4c Linux-2.6.12-rc2 |
2285 2286 2287 |
} /* Take a nap, wait for some writeback to complete */ |
7b51755c3 vmscan: kill hibe... |
2288 |
if (!sc->hibernation_mode && sc->nr_scanned && |
0e093d997 writeback: do not... |
2289 2290 2291 2292 |
priority < DEF_PRIORITY - 2) { struct zone *preferred_zone; first_zones_zonelist(zonelist, gfp_zone(sc->gfp_mask), |
f33261d75 mm: fix deferred ... |
2293 2294 |
&cpuset_current_mems_allowed, &preferred_zone); |
0e093d997 writeback: do not... |
2295 2296 |
wait_iff_congested(preferred_zone, BLK_RW_ASYNC, HZ/10); } |
1da177e4c Linux-2.6.12-rc2 |
2297 |
} |
bb21c7ce1 vmscan: fix do_tr... |
2298 |
|
1da177e4c Linux-2.6.12-rc2 |
2299 |
out: |
873b47717 per-task-delay-ac... |
2300 |
delayacct_freepages_end(); |
c0ff7453b cpuset,mm: fix no... |
2301 |
put_mems_allowed(); |
873b47717 per-task-delay-ac... |
2302 |
|
bb21c7ce1 vmscan: fix do_tr... |
2303 2304 |
if (sc->nr_reclaimed) return sc->nr_reclaimed; |
929bea7c7 vmscan: all_unrec... |
2305 2306 2307 2308 2309 2310 2311 |
/* * As hibernation is going on, kswapd is freezed so that it can't mark * the zone into all_unreclaimable. Thus bypassing all_unreclaimable * check. */ if (oom_killer_disabled) return 0; |
0cee34fd7 mm: vmscan: check... |
2312 2313 |
/* Aborted reclaim to try compaction? don't OOM, then */ if (aborted_reclaim) |
7335084d4 mm: vmscan: do no... |
2314 |
return 1; |
bb21c7ce1 vmscan: fix do_tr... |
2315 |
/* top priority shrink_zones still had more to do? don't OOM, then */ |
89b5fae53 mm: vmscan: disti... |
2316 |
if (global_reclaim(sc) && !all_unreclaimable(zonelist, sc)) |
bb21c7ce1 vmscan: fix do_tr... |
2317 2318 2319 |
return 1; return 0; |
1da177e4c Linux-2.6.12-rc2 |
2320 |
} |
dac1d27bc mm: use zonelists... |
2321 |
unsigned long try_to_free_pages(struct zonelist *zonelist, int order, |
327c0e968 vmscan: fix it to... |
2322 |
gfp_t gfp_mask, nodemask_t *nodemask) |
66e1707bc Memory controller... |
2323 |
{ |
33906bc5c vmscan: tracing: ... |
2324 |
unsigned long nr_reclaimed; |
66e1707bc Memory controller... |
2325 2326 2327 |
struct scan_control sc = { .gfp_mask = gfp_mask, .may_writepage = !laptop_mode, |
22fba3354 vmscan: separate ... |
2328 |
.nr_to_reclaim = SWAP_CLUSTER_MAX, |
a6dc60f89 vmscan: rename sc... |
2329 |
.may_unmap = 1, |
2e2e42598 vmscan,memcg: rei... |
2330 |
.may_swap = 1, |
66e1707bc Memory controller... |
2331 |
.order = order, |
f16015fbf mm: vmscan: disti... |
2332 |
.target_mem_cgroup = NULL, |
327c0e968 vmscan: fix it to... |
2333 |
.nodemask = nodemask, |
66e1707bc Memory controller... |
2334 |
}; |
a09ed5e00 vmscan: change sh... |
2335 2336 2337 |
struct shrink_control shrink = { .gfp_mask = sc.gfp_mask, }; |
66e1707bc Memory controller... |
2338 |
|
33906bc5c vmscan: tracing: ... |
2339 2340 2341 |
trace_mm_vmscan_direct_reclaim_begin(order, sc.may_writepage, gfp_mask); |
a09ed5e00 vmscan: change sh... |
2342 |
nr_reclaimed = do_try_to_free_pages(zonelist, &sc, &shrink); |
33906bc5c vmscan: tracing: ... |
2343 2344 2345 2346 |
trace_mm_vmscan_direct_reclaim_end(nr_reclaimed); return nr_reclaimed; |
66e1707bc Memory controller... |
2347 |
} |
00f0b8259 Memory controller... |
2348 |
#ifdef CONFIG_CGROUP_MEM_RES_CTLR |
66e1707bc Memory controller... |
2349 |
|
72835c86c mm: unify remaini... |
2350 |
unsigned long mem_cgroup_shrink_node_zone(struct mem_cgroup *memcg, |
4e4169535 memory controller... |
2351 |
gfp_t gfp_mask, bool noswap, |
0ae5e89c6 memcg: count the ... |
2352 2353 |
struct zone *zone, unsigned long *nr_scanned) |
4e4169535 memory controller... |
2354 2355 |
{ struct scan_control sc = { |
0ae5e89c6 memcg: count the ... |
2356 |
.nr_scanned = 0, |
b8f5c5664 memcg: sc.nr_to_r... |
2357 |
.nr_to_reclaim = SWAP_CLUSTER_MAX, |
4e4169535 memory controller... |
2358 2359 2360 |
.may_writepage = !laptop_mode, .may_unmap = 1, .may_swap = !noswap, |
4e4169535 memory controller... |
2361 |
.order = 0, |
72835c86c mm: unify remaini... |
2362 |
.target_mem_cgroup = memcg, |
4e4169535 memory controller... |
2363 |
}; |
5660048cc mm: move memcg hi... |
2364 |
struct mem_cgroup_zone mz = { |
72835c86c mm: unify remaini... |
2365 |
.mem_cgroup = memcg, |
5660048cc mm: move memcg hi... |
2366 2367 |
.zone = zone, }; |
0ae5e89c6 memcg: count the ... |
2368 |
|
4e4169535 memory controller... |
2369 2370 |
sc.gfp_mask = (gfp_mask & GFP_RECLAIM_MASK) | (GFP_HIGHUSER_MOVABLE & ~GFP_RECLAIM_MASK); |
bdce6d9eb memcg, vmscan: ad... |
2371 2372 2373 2374 |
trace_mm_vmscan_memcg_softlimit_reclaim_begin(0, sc.may_writepage, sc.gfp_mask); |
4e4169535 memory controller... |
2375 2376 2377 2378 2379 2380 2381 |
/* * 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. */ |
5660048cc mm: move memcg hi... |
2382 |
shrink_mem_cgroup_zone(0, &mz, &sc); |
bdce6d9eb memcg, vmscan: ad... |
2383 2384 |
trace_mm_vmscan_memcg_softlimit_reclaim_end(sc.nr_reclaimed); |
0ae5e89c6 memcg: count the ... |
2385 |
*nr_scanned = sc.nr_scanned; |
4e4169535 memory controller... |
2386 2387 |
return sc.nr_reclaimed; } |
72835c86c mm: unify remaini... |
2388 |
unsigned long try_to_free_mem_cgroup_pages(struct mem_cgroup *memcg, |
a7885eb8a memcg: swappiness |
2389 |
gfp_t gfp_mask, |
185efc0f9 memcg: Revert "me... |
2390 |
bool noswap) |
66e1707bc Memory controller... |
2391 |
{ |
4e4169535 memory controller... |
2392 |
struct zonelist *zonelist; |
bdce6d9eb memcg, vmscan: ad... |
2393 |
unsigned long nr_reclaimed; |
889976dbc memcg: reclaim me... |
2394 |
int nid; |
66e1707bc Memory controller... |
2395 |
struct scan_control sc = { |
66e1707bc Memory controller... |
2396 |
.may_writepage = !laptop_mode, |
a6dc60f89 vmscan: rename sc... |
2397 |
.may_unmap = 1, |
2e2e42598 vmscan,memcg: rei... |
2398 |
.may_swap = !noswap, |
22fba3354 vmscan: separate ... |
2399 |
.nr_to_reclaim = SWAP_CLUSTER_MAX, |
66e1707bc Memory controller... |
2400 |
.order = 0, |
72835c86c mm: unify remaini... |
2401 |
.target_mem_cgroup = memcg, |
327c0e968 vmscan: fix it to... |
2402 |
.nodemask = NULL, /* we don't care the placement */ |
a09ed5e00 vmscan: change sh... |
2403 2404 2405 2406 2407 |
.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... |
2408 |
}; |
66e1707bc Memory controller... |
2409 |
|
889976dbc memcg: reclaim me... |
2410 2411 2412 2413 2414 |
/* * Unlike direct reclaim via alloc_pages(), memcg's reclaim doesn't * take care of from where we get pages. So the node where we start the * scan does not need to be the current node. */ |
72835c86c mm: unify remaini... |
2415 |
nid = mem_cgroup_select_victim_node(memcg); |
889976dbc memcg: reclaim me... |
2416 2417 |
zonelist = NODE_DATA(nid)->node_zonelists; |
bdce6d9eb memcg, vmscan: ad... |
2418 2419 2420 2421 |
trace_mm_vmscan_memcg_reclaim_begin(0, sc.may_writepage, sc.gfp_mask); |
a09ed5e00 vmscan: change sh... |
2422 |
nr_reclaimed = do_try_to_free_pages(zonelist, &sc, &shrink); |
bdce6d9eb memcg, vmscan: ad... |
2423 2424 2425 2426 |
trace_mm_vmscan_memcg_reclaim_end(nr_reclaimed); return nr_reclaimed; |
66e1707bc Memory controller... |
2427 2428 |
} #endif |
f16015fbf mm: vmscan: disti... |
2429 2430 2431 |
static void age_active_anon(struct zone *zone, struct scan_control *sc, int priority) { |
b95a2f2d4 mm: vmscan: conve... |
2432 |
struct mem_cgroup *memcg; |
f16015fbf mm: vmscan: disti... |
2433 |
|
b95a2f2d4 mm: vmscan: conve... |
2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 |
if (!total_swap_pages) return; memcg = mem_cgroup_iter(NULL, NULL, NULL); do { struct mem_cgroup_zone mz = { .mem_cgroup = memcg, .zone = zone, }; if (inactive_anon_is_low(&mz)) shrink_active_list(SWAP_CLUSTER_MAX, &mz, sc, priority, 0); memcg = mem_cgroup_iter(NULL, memcg, NULL); } while (memcg); |
f16015fbf mm: vmscan: disti... |
2450 |
} |
1741c8775 mm: kswapd: keep ... |
2451 2452 2453 2454 2455 2456 2457 2458 2459 2460 2461 |
/* * 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... |
2462 |
* percentage of the middle zones. For example, on 32-bit x86, highmem |
1741c8775 mm: kswapd: keep ... |
2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 |
* 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... |
2475 2476 |
/* A special case here: if zone has no page, we think it's balanced */ return balanced_pages >= (present_pages >> 2); |
1741c8775 mm: kswapd: keep ... |
2477 |
} |
f50de2d38 vmscan: have kswa... |
2478 |
/* is kswapd sleeping prematurely? */ |
dc83edd94 mm: kswapd: use t... |
2479 2480 |
static bool sleeping_prematurely(pg_data_t *pgdat, int order, long remaining, int classzone_idx) |
f50de2d38 vmscan: have kswa... |
2481 |
{ |
bb3ab5968 vmscan: stop kswa... |
2482 |
int i; |
1741c8775 mm: kswapd: keep ... |
2483 2484 |
unsigned long balanced = 0; bool all_zones_ok = true; |
f50de2d38 vmscan: have kswa... |
2485 2486 2487 |
/* If a direct reclaimer woke kswapd within HZ/10, it's premature */ if (remaining) |
dc83edd94 mm: kswapd: use t... |
2488 |
return true; |
f50de2d38 vmscan: have kswa... |
2489 |
|
0abdee2bd mm: kswapd: use t... |
2490 |
/* Check the watermark levels */ |
08951e545 mm: vmscan: corre... |
2491 |
for (i = 0; i <= classzone_idx; i++) { |
bb3ab5968 vmscan: stop kswa... |
2492 2493 2494 2495 |
struct zone *zone = pgdat->node_zones + i; if (!populated_zone(zone)) continue; |
355b09c47 mm: kswapd: treat... |
2496 2497 2498 2499 2500 2501 2502 2503 |
/* * 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... |
2504 |
continue; |
355b09c47 mm: kswapd: treat... |
2505 |
} |
de3fab393 vmscan: kswapd: d... |
2506 |
|
88f5acf88 mm: page allocato... |
2507 |
if (!zone_watermark_ok_safe(zone, order, high_wmark_pages(zone), |
da175d06b mm: vmscan: evalu... |
2508 |
i, 0)) |
1741c8775 mm: kswapd: keep ... |
2509 2510 2511 |
all_zones_ok = false; else balanced += zone->present_pages; |
bb3ab5968 vmscan: stop kswa... |
2512 |
} |
f50de2d38 vmscan: have kswa... |
2513 |
|
1741c8775 mm: kswapd: keep ... |
2514 2515 2516 2517 2518 2519 |
/* * 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... |
2520 |
return !pgdat_balanced(pgdat, balanced, classzone_idx); |
1741c8775 mm: kswapd: keep ... |
2521 2522 |
else return !all_zones_ok; |
f50de2d38 vmscan: have kswa... |
2523 |
} |
1da177e4c Linux-2.6.12-rc2 |
2524 2525 |
/* * For kswapd, balance_pgdat() will work across all this node's zones until |
418589663 page allocator: u... |
2526 |
* they are all at high_wmark_pages(zone). |
1da177e4c Linux-2.6.12-rc2 |
2527 |
* |
0abdee2bd mm: kswapd: use t... |
2528 |
* Returns the final order kswapd was reclaiming at |
1da177e4c Linux-2.6.12-rc2 |
2529 2530 2531 2532 2533 2534 2535 2536 2537 2538 |
* * 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... |
2539 2540 2541 2542 2543 |
* 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 |
2544 |
*/ |
995047488 mm: kswapd: stop ... |
2545 |
static unsigned long balance_pgdat(pg_data_t *pgdat, int order, |
dc83edd94 mm: kswapd: use t... |
2546 |
int *classzone_idx) |
1da177e4c Linux-2.6.12-rc2 |
2547 |
{ |
1da177e4c Linux-2.6.12-rc2 |
2548 |
int all_zones_ok; |
1741c8775 mm: kswapd: keep ... |
2549 |
unsigned long balanced; |
1da177e4c Linux-2.6.12-rc2 |
2550 2551 |
int priority; int i; |
995047488 mm: kswapd: stop ... |
2552 |
int end_zone = 0; /* Inclusive. 0 = ZONE_DMA */ |
69e05944a [PATCH] vmscan: u... |
2553 |
unsigned long total_scanned; |
1da177e4c Linux-2.6.12-rc2 |
2554 |
struct reclaim_state *reclaim_state = current->reclaim_state; |
0ae5e89c6 memcg: count the ... |
2555 2556 |
unsigned long nr_soft_reclaimed; unsigned long nr_soft_scanned; |
179e96395 [PATCH] vmscan: s... |
2557 2558 |
struct scan_control sc = { .gfp_mask = GFP_KERNEL, |
a6dc60f89 vmscan: rename sc... |
2559 |
.may_unmap = 1, |
2e2e42598 vmscan,memcg: rei... |
2560 |
.may_swap = 1, |
22fba3354 vmscan: separate ... |
2561 2562 2563 2564 2565 |
/* * 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 |
2566 |
.order = order, |
f16015fbf mm: vmscan: disti... |
2567 |
.target_mem_cgroup = NULL, |
179e96395 [PATCH] vmscan: s... |
2568 |
}; |
a09ed5e00 vmscan: change sh... |
2569 2570 2571 |
struct shrink_control shrink = { .gfp_mask = sc.gfp_mask, }; |
1da177e4c Linux-2.6.12-rc2 |
2572 2573 |
loop_again: total_scanned = 0; |
a79311c14 vmscan: bail out ... |
2574 |
sc.nr_reclaimed = 0; |
c0bbbc73d [PATCH] typo in v... |
2575 |
sc.may_writepage = !laptop_mode; |
f8891e5e1 [PATCH] Light wei... |
2576 |
count_vm_event(PAGEOUTRUN); |
1da177e4c Linux-2.6.12-rc2 |
2577 |
|
1da177e4c Linux-2.6.12-rc2 |
2578 |
for (priority = DEF_PRIORITY; priority >= 0; priority--) { |
1da177e4c Linux-2.6.12-rc2 |
2579 |
unsigned long lru_pages = 0; |
bb3ab5968 vmscan: stop kswa... |
2580 |
int has_under_min_watermark_zone = 0; |
1da177e4c Linux-2.6.12-rc2 |
2581 |
|
f7b7fd8f3 [PATCH] temporari... |
2582 2583 |
/* The swap token gets in the way of swapout... */ if (!priority) |
a433658c3 vmscan,memcg: mem... |
2584 |
disable_swap_token(NULL); |
f7b7fd8f3 [PATCH] temporari... |
2585 |
|
1da177e4c Linux-2.6.12-rc2 |
2586 |
all_zones_ok = 1; |
1741c8775 mm: kswapd: keep ... |
2587 |
balanced = 0; |
1da177e4c Linux-2.6.12-rc2 |
2588 |
|
d6277db4a [PATCH] swsusp: r... |
2589 2590 2591 2592 2593 2594 |
/* * 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 |
2595 |
|
d6277db4a [PATCH] swsusp: r... |
2596 2597 |
if (!populated_zone(zone)) continue; |
1da177e4c Linux-2.6.12-rc2 |
2598 |
|
93e4a89a8 mm: restore zone-... |
2599 |
if (zone->all_unreclaimable && priority != DEF_PRIORITY) |
d6277db4a [PATCH] swsusp: r... |
2600 |
continue; |
1da177e4c Linux-2.6.12-rc2 |
2601 |
|
556adecba vmscan: second ch... |
2602 2603 2604 2605 |
/* * Do some background aging of the anon list, to give * pages a chance to be referenced before reclaiming. */ |
f16015fbf mm: vmscan: disti... |
2606 |
age_active_anon(zone, &sc, priority); |
556adecba vmscan: second ch... |
2607 |
|
88f5acf88 mm: page allocato... |
2608 |
if (!zone_watermark_ok_safe(zone, order, |
418589663 page allocator: u... |
2609 |
high_wmark_pages(zone), 0, 0)) { |
d6277db4a [PATCH] swsusp: r... |
2610 |
end_zone = i; |
e1dbeda60 [PATCH] balance_p... |
2611 |
break; |
439423f68 vmscan: clear ZON... |
2612 2613 2614 |
} else { /* If balanced, clear the congested flag */ zone_clear_flag(zone, ZONE_CONGESTED); |
1da177e4c Linux-2.6.12-rc2 |
2615 |
} |
1da177e4c Linux-2.6.12-rc2 |
2616 |
} |
e1dbeda60 [PATCH] balance_p... |
2617 2618 |
if (i < 0) goto out; |
1da177e4c Linux-2.6.12-rc2 |
2619 2620 |
for (i = 0; i <= end_zone; i++) { struct zone *zone = pgdat->node_zones + i; |
adea02a1b mm: count only re... |
2621 |
lru_pages += zone_reclaimable_pages(zone); |
1da177e4c Linux-2.6.12-rc2 |
2622 2623 2624 2625 2626 2627 2628 2629 2630 2631 2632 2633 2634 |
} /* * 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... |
2635 |
int nr_slab; |
8afdcece4 mm: vmscan: kswap... |
2636 |
unsigned long balance_gap; |
1da177e4c Linux-2.6.12-rc2 |
2637 |
|
f3fe65122 [PATCH] mm: add p... |
2638 |
if (!populated_zone(zone)) |
1da177e4c Linux-2.6.12-rc2 |
2639 |
continue; |
93e4a89a8 mm: restore zone-... |
2640 |
if (zone->all_unreclaimable && priority != DEF_PRIORITY) |
1da177e4c Linux-2.6.12-rc2 |
2641 |
continue; |
1da177e4c Linux-2.6.12-rc2 |
2642 |
sc.nr_scanned = 0; |
4e4169535 memory controller... |
2643 |
|
0ae5e89c6 memcg: count the ... |
2644 |
nr_soft_scanned = 0; |
4e4169535 memory controller... |
2645 2646 |
/* * Call soft limit reclaim before calling shrink_zone. |
4e4169535 memory controller... |
2647 |
*/ |
0ae5e89c6 memcg: count the ... |
2648 2649 2650 2651 2652 |
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... |
2653 |
|
32a4330d4 mm: prevent kswap... |
2654 |
/* |
8afdcece4 mm: vmscan: kswap... |
2655 2656 2657 2658 2659 2660 |
* 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... |
2661 |
*/ |
8afdcece4 mm: vmscan: kswap... |
2662 2663 2664 2665 |
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... |
2666 |
if (!zone_watermark_ok_safe(zone, order, |
8afdcece4 mm: vmscan: kswap... |
2667 |
high_wmark_pages(zone) + balance_gap, |
d7868dae8 mm: vmscan: do no... |
2668 |
end_zone, 0)) { |
a79311c14 vmscan: bail out ... |
2669 |
shrink_zone(priority, zone, &sc); |
5a03b051e thp: use compacti... |
2670 |
|
d7868dae8 mm: vmscan: do no... |
2671 2672 2673 2674 2675 2676 2677 2678 |
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 |
2679 2680 2681 2682 2683 2684 |
/* * 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 ... |
2685 |
total_scanned > sc.nr_reclaimed + sc.nr_reclaimed / 2) |
1da177e4c Linux-2.6.12-rc2 |
2686 |
sc.may_writepage = 1; |
bb3ab5968 vmscan: stop kswa... |
2687 |
|
215ddd666 mm: vmscan: only ... |
2688 2689 2690 |
if (zone->all_unreclaimable) { if (end_zone && end_zone == i) end_zone--; |
d7868dae8 mm: vmscan: do no... |
2691 |
continue; |
215ddd666 mm: vmscan: only ... |
2692 |
} |
d7868dae8 mm: vmscan: do no... |
2693 |
|
88f5acf88 mm: page allocato... |
2694 |
if (!zone_watermark_ok_safe(zone, order, |
45973d74f vmscan: check hig... |
2695 2696 2697 2698 2699 2700 2701 |
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... |
2702 |
if (!zone_watermark_ok_safe(zone, order, |
45973d74f vmscan: check hig... |
2703 2704 |
min_wmark_pages(zone), end_zone, 0)) has_under_min_watermark_zone = 1; |
0e093d997 writeback: do not... |
2705 2706 2707 2708 2709 2710 2711 2712 2713 |
} 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... |
2714 |
if (i <= *classzone_idx) |
1741c8775 mm: kswapd: keep ... |
2715 |
balanced += zone->present_pages; |
45973d74f vmscan: check hig... |
2716 |
} |
bb3ab5968 vmscan: stop kswa... |
2717 |
|
1da177e4c Linux-2.6.12-rc2 |
2718 |
} |
dc83edd94 mm: kswapd: use t... |
2719 |
if (all_zones_ok || (order && pgdat_balanced(pgdat, balanced, *classzone_idx))) |
1da177e4c Linux-2.6.12-rc2 |
2720 2721 2722 2723 2724 |
break; /* kswapd: all done */ /* * OK, kswapd is getting into trouble. Take a nap, then take * another pass across the zones. */ |
bb3ab5968 vmscan: stop kswa... |
2725 2726 2727 2728 2729 2730 |
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 |
2731 2732 2733 2734 2735 2736 2737 |
/* * 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 ... |
2738 |
if (sc.nr_reclaimed >= SWAP_CLUSTER_MAX) |
1da177e4c Linux-2.6.12-rc2 |
2739 2740 2741 |
break; } out: |
995047488 mm: kswapd: stop ... |
2742 2743 2744 |
/* * order-0: All zones must meet high watermark for a balanced node |
1741c8775 mm: kswapd: keep ... |
2745 2746 |
* high-order: Balanced zones must make up at least 25% of the node * for the node to be balanced |
995047488 mm: kswapd: stop ... |
2747 |
*/ |
dc83edd94 mm: kswapd: use t... |
2748 |
if (!(all_zones_ok || (order && pgdat_balanced(pgdat, balanced, *classzone_idx)))) { |
1da177e4c Linux-2.6.12-rc2 |
2749 |
cond_resched(); |
8357376d3 [PATCH] swsusp: I... |
2750 2751 |
try_to_freeze(); |
73ce02e96 mm: stop kswapd's... |
2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766 2767 |
/* * 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 |
2768 2769 |
goto loop_again; } |
995047488 mm: kswapd: stop ... |
2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793 2794 2795 2796 |
/* * 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... |
2797 2798 |
if (i <= *classzone_idx) balanced += zone->present_pages; |
995047488 mm: kswapd: stop ... |
2799 2800 |
} } |
0abdee2bd mm: kswapd: use t... |
2801 2802 2803 2804 2805 2806 |
/* * 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... |
2807 |
*classzone_idx = end_zone; |
0abdee2bd mm: kswapd: use t... |
2808 |
return order; |
1da177e4c Linux-2.6.12-rc2 |
2809 |
} |
dc83edd94 mm: kswapd: use t... |
2810 |
static void kswapd_try_to_sleep(pg_data_t *pgdat, int order, int classzone_idx) |
f0bc0a60b vmscan: factor ou... |
2811 2812 2813 2814 2815 2816 2817 2818 2819 2820 |
{ 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... |
2821 |
if (!sleeping_prematurely(pgdat, order, remaining, classzone_idx)) { |
f0bc0a60b vmscan: factor ou... |
2822 2823 2824 2825 2826 2827 2828 2829 2830 |
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... |
2831 |
if (!sleeping_prematurely(pgdat, order, remaining, classzone_idx)) { |
f0bc0a60b vmscan: factor ou... |
2832 2833 2834 2835 2836 2837 2838 2839 2840 2841 2842 2843 2844 2845 2846 2847 2848 2849 2850 2851 2852 |
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 |
2853 2854 |
/* * The background pageout daemon, started as a kernel thread |
4f98a2fee vmscan: split LRU... |
2855 |
* from the init process. |
1da177e4c Linux-2.6.12-rc2 |
2856 2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 2867 |
* * 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 ... |
2868 |
unsigned long order, new_order; |
d2ebd0f6b kswapd: avoid unn... |
2869 |
unsigned balanced_order; |
215ddd666 mm: vmscan: only ... |
2870 |
int classzone_idx, new_classzone_idx; |
d2ebd0f6b kswapd: avoid unn... |
2871 |
int balanced_classzone_idx; |
1da177e4c Linux-2.6.12-rc2 |
2872 2873 |
pg_data_t *pgdat = (pg_data_t*)p; struct task_struct *tsk = current; |
f0bc0a60b vmscan: factor ou... |
2874 |
|
1da177e4c Linux-2.6.12-rc2 |
2875 2876 2877 |
struct reclaim_state reclaim_state = { .reclaimed_slab = 0, }; |
a70f73028 cpumask: replace ... |
2878 |
const struct cpumask *cpumask = cpumask_of_node(pgdat->node_id); |
1da177e4c Linux-2.6.12-rc2 |
2879 |
|
cf40bd16f lockdep: annotate... |
2880 |
lockdep_set_current_reclaim_state(GFP_KERNEL); |
174596a0b cpumask: convert mm/ |
2881 |
if (!cpumask_empty(cpumask)) |
c5f59f083 nodemask: use new... |
2882 |
set_cpus_allowed_ptr(tsk, cpumask); |
1da177e4c Linux-2.6.12-rc2 |
2883 2884 2885 2886 2887 2888 2889 2890 2891 2892 2893 2894 2895 2896 |
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... |
2897 |
tsk->flags |= PF_MEMALLOC | PF_SWAPWRITE | PF_KSWAPD; |
831441862 Freezer: make ker... |
2898 |
set_freezable(); |
1da177e4c Linux-2.6.12-rc2 |
2899 |
|
215ddd666 mm: vmscan: only ... |
2900 |
order = new_order = 0; |
d2ebd0f6b kswapd: avoid unn... |
2901 |
balanced_order = 0; |
215ddd666 mm: vmscan: only ... |
2902 |
classzone_idx = new_classzone_idx = pgdat->nr_zones - 1; |
d2ebd0f6b kswapd: avoid unn... |
2903 |
balanced_classzone_idx = classzone_idx; |
1da177e4c Linux-2.6.12-rc2 |
2904 |
for ( ; ; ) { |
8fe23e057 mm: clear node in... |
2905 |
int ret; |
3e1d1d28d [PATCH] Cleanup p... |
2906 |
|
215ddd666 mm: vmscan: only ... |
2907 2908 2909 2910 2911 |
/* * 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... |
2912 2913 |
if (balanced_classzone_idx >= new_classzone_idx && balanced_order == new_order) { |
215ddd666 mm: vmscan: only ... |
2914 2915 2916 2917 2918 |
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 ... |
2919 |
if (order < new_order || classzone_idx > new_classzone_idx) { |
1da177e4c Linux-2.6.12-rc2 |
2920 2921 |
/* * Don't sleep if someone wants a larger 'order' |
995047488 mm: kswapd: stop ... |
2922 |
* allocation or has tigher zone constraints |
1da177e4c Linux-2.6.12-rc2 |
2923 2924 |
*/ order = new_order; |
995047488 mm: kswapd: stop ... |
2925 |
classzone_idx = new_classzone_idx; |
1da177e4c Linux-2.6.12-rc2 |
2926 |
} else { |
d2ebd0f6b kswapd: avoid unn... |
2927 2928 |
kswapd_try_to_sleep(pgdat, balanced_order, balanced_classzone_idx); |
1da177e4c Linux-2.6.12-rc2 |
2929 |
order = pgdat->kswapd_max_order; |
995047488 mm: kswapd: stop ... |
2930 |
classzone_idx = pgdat->classzone_idx; |
f0dfcde09 kswapd: assign ne... |
2931 2932 |
new_order = order; new_classzone_idx = classzone_idx; |
4d40502ea mm: kswapd: reset... |
2933 |
pgdat->kswapd_max_order = 0; |
215ddd666 mm: vmscan: only ... |
2934 |
pgdat->classzone_idx = pgdat->nr_zones - 1; |
1da177e4c Linux-2.6.12-rc2 |
2935 |
} |
1da177e4c Linux-2.6.12-rc2 |
2936 |
|
8fe23e057 mm: clear node in... |
2937 2938 2939 2940 2941 2942 2943 2944 |
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: ... |
2945 2946 |
if (!ret) { trace_mm_vmscan_kswapd_wake(pgdat->node_id, order); |
d2ebd0f6b kswapd: avoid unn... |
2947 2948 2949 |
balanced_classzone_idx = classzone_idx; balanced_order = balance_pgdat(pgdat, order, &balanced_classzone_idx); |
33906bc5c vmscan: tracing: ... |
2950 |
} |
1da177e4c Linux-2.6.12-rc2 |
2951 2952 2953 2954 2955 2956 2957 |
} return 0; } /* * A zone is low on free memory, so wake its kswapd task to service it. */ |
995047488 mm: kswapd: stop ... |
2958 |
void wakeup_kswapd(struct zone *zone, int order, enum zone_type classzone_idx) |
1da177e4c Linux-2.6.12-rc2 |
2959 2960 |
{ pg_data_t *pgdat; |
f3fe65122 [PATCH] mm: add p... |
2961 |
if (!populated_zone(zone)) |
1da177e4c Linux-2.6.12-rc2 |
2962 |
return; |
88f5acf88 mm: page allocato... |
2963 |
if (!cpuset_zone_allowed_hardwall(zone, GFP_KERNEL)) |
1da177e4c Linux-2.6.12-rc2 |
2964 |
return; |
88f5acf88 mm: page allocato... |
2965 |
pgdat = zone->zone_pgdat; |
995047488 mm: kswapd: stop ... |
2966 |
if (pgdat->kswapd_max_order < order) { |
1da177e4c Linux-2.6.12-rc2 |
2967 |
pgdat->kswapd_max_order = order; |
995047488 mm: kswapd: stop ... |
2968 2969 |
pgdat->classzone_idx = min(pgdat->classzone_idx, classzone_idx); } |
8d0986e28 [PATCH] vm: kswap... |
2970 |
if (!waitqueue_active(&pgdat->kswapd_wait)) |
1da177e4c Linux-2.6.12-rc2 |
2971 |
return; |
88f5acf88 mm: page allocato... |
2972 2973 2974 2975 |
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... |
2976 |
wake_up_interruptible(&pgdat->kswapd_wait); |
1da177e4c Linux-2.6.12-rc2 |
2977 |
} |
adea02a1b mm: count only re... |
2978 2979 2980 2981 2982 2983 2984 2985 |
/* * 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... |
2986 |
{ |
adea02a1b mm: count only re... |
2987 2988 2989 2990 2991 2992 2993 2994 2995 2996 2997 2998 2999 3000 3001 3002 3003 3004 3005 3006 3007 3008 3009 3010 |
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... |
3011 |
} |
c6f37f121 PM/Suspend: Do no... |
3012 |
#ifdef CONFIG_HIBERNATION |
1da177e4c Linux-2.6.12-rc2 |
3013 |
/* |
7b51755c3 vmscan: kill hibe... |
3014 |
* Try to free `nr_to_reclaim' of memory, system-wide, and return the number of |
d6277db4a [PATCH] swsusp: r... |
3015 3016 3017 3018 3019 |
* 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 |
3020 |
*/ |
7b51755c3 vmscan: kill hibe... |
3021 |
unsigned long shrink_all_memory(unsigned long nr_to_reclaim) |
1da177e4c Linux-2.6.12-rc2 |
3022 |
{ |
d6277db4a [PATCH] swsusp: r... |
3023 |
struct reclaim_state reclaim_state; |
d6277db4a [PATCH] swsusp: r... |
3024 |
struct scan_control sc = { |
7b51755c3 vmscan: kill hibe... |
3025 3026 3027 |
.gfp_mask = GFP_HIGHUSER_MOVABLE, .may_swap = 1, .may_unmap = 1, |
d6277db4a [PATCH] swsusp: r... |
3028 |
.may_writepage = 1, |
7b51755c3 vmscan: kill hibe... |
3029 3030 |
.nr_to_reclaim = nr_to_reclaim, .hibernation_mode = 1, |
7b51755c3 vmscan: kill hibe... |
3031 |
.order = 0, |
1da177e4c Linux-2.6.12-rc2 |
3032 |
}; |
a09ed5e00 vmscan: change sh... |
3033 3034 3035 3036 |
struct shrink_control shrink = { .gfp_mask = sc.gfp_mask, }; struct zonelist *zonelist = node_zonelist(numa_node_id(), sc.gfp_mask); |
7b51755c3 vmscan: kill hibe... |
3037 3038 |
struct task_struct *p = current; unsigned long nr_reclaimed; |
1da177e4c Linux-2.6.12-rc2 |
3039 |
|
7b51755c3 vmscan: kill hibe... |
3040 3041 3042 3043 |
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... |
3044 |
|
a09ed5e00 vmscan: change sh... |
3045 |
nr_reclaimed = do_try_to_free_pages(zonelist, &sc, &shrink); |
d979677c4 mm: shrink_all_me... |
3046 |
|
7b51755c3 vmscan: kill hibe... |
3047 3048 3049 |
p->reclaim_state = NULL; lockdep_clear_current_reclaim_state(); p->flags &= ~PF_MEMALLOC; |
d6277db4a [PATCH] swsusp: r... |
3050 |
|
7b51755c3 vmscan: kill hibe... |
3051 |
return nr_reclaimed; |
1da177e4c Linux-2.6.12-rc2 |
3052 |
} |
c6f37f121 PM/Suspend: Do no... |
3053 |
#endif /* CONFIG_HIBERNATION */ |
1da177e4c Linux-2.6.12-rc2 |
3054 |
|
1da177e4c Linux-2.6.12-rc2 |
3055 3056 3057 3058 |
/* 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... |
3059 |
static int __devinit cpu_callback(struct notifier_block *nfb, |
69e05944a [PATCH] vmscan: u... |
3060 |
unsigned long action, void *hcpu) |
1da177e4c Linux-2.6.12-rc2 |
3061 |
{ |
58c0a4a78 Fix panic of cpu ... |
3062 |
int nid; |
1da177e4c Linux-2.6.12-rc2 |
3063 |
|
8bb784428 Add suspend-relat... |
3064 |
if (action == CPU_ONLINE || action == CPU_ONLINE_FROZEN) { |
58c0a4a78 Fix panic of cpu ... |
3065 |
for_each_node_state(nid, N_HIGH_MEMORY) { |
c5f59f083 nodemask: use new... |
3066 |
pg_data_t *pgdat = NODE_DATA(nid); |
a70f73028 cpumask: replace ... |
3067 3068 3069 |
const struct cpumask *mask; mask = cpumask_of_node(pgdat->node_id); |
c5f59f083 nodemask: use new... |
3070 |
|
3e5979453 cpumask: remove a... |
3071 |
if (cpumask_any_and(cpu_online_mask, mask) < nr_cpu_ids) |
1da177e4c Linux-2.6.12-rc2 |
3072 |
/* One of our CPUs online: restore mask */ |
c5f59f083 nodemask: use new... |
3073 |
set_cpus_allowed_ptr(pgdat->kswapd, mask); |
1da177e4c Linux-2.6.12-rc2 |
3074 3075 3076 3077 |
} } return NOTIFY_OK; } |
1da177e4c Linux-2.6.12-rc2 |
3078 |
|
3218ae14b [PATCH] pgdat all... |
3079 3080 3081 3082 3083 3084 3085 3086 3087 3088 3089 3090 3091 3092 3093 3094 3095 3096 3097 3098 3099 3100 |
/* * 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... |
3101 3102 3103 3104 3105 3106 3107 3108 3109 3110 |
/* * 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 |
3111 3112 |
static int __init kswapd_init(void) { |
3218ae14b [PATCH] pgdat all... |
3113 |
int nid; |
69e05944a [PATCH] vmscan: u... |
3114 |
|
1da177e4c Linux-2.6.12-rc2 |
3115 |
swap_setup(); |
9422ffba4 Memoryless nodes:... |
3116 |
for_each_node_state(nid, N_HIGH_MEMORY) |
3218ae14b [PATCH] pgdat all... |
3117 |
kswapd_run(nid); |
1da177e4c Linux-2.6.12-rc2 |
3118 3119 3120 3121 3122 |
hotcpu_notifier(cpu_callback, 0); return 0; } module_init(kswapd_init) |
9eeff2395 [PATCH] Zone recl... |
3123 3124 3125 3126 3127 3128 3129 |
#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... |
3130 3131 |
*/ int zone_reclaim_mode __read_mostly; |
1b2ffb789 [PATCH] Zone recl... |
3132 |
#define RECLAIM_OFF 0 |
7d03431cf swapfile/vmscan: ... |
3133 |
#define RECLAIM_ZONE (1<<0) /* Run shrink_inactive_list on the zone */ |
1b2ffb789 [PATCH] Zone recl... |
3134 3135 |
#define RECLAIM_WRITE (1<<1) /* Writeout pages during reclaim */ #define RECLAIM_SWAP (1<<2) /* Swap pages out during reclaim */ |
9eeff2395 [PATCH] Zone recl... |
3136 |
/* |
a92f71263 [PATCH] zone_recl... |
3137 3138 3139 3140 3141 |
* 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... |
3142 |
/* |
9614634fe [PATCH] ZVC/zone_... |
3143 3144 3145 3146 3147 3148 |
* 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... |
3149 3150 3151 3152 |
* 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 ... |
3153 3154 3155 3156 3157 3158 3159 3160 3161 3162 3163 3164 3165 3166 3167 3168 3169 3170 3171 3172 3173 3174 3175 3176 3177 3178 3179 3180 3181 3182 3183 3184 3185 3186 3187 3188 3189 3190 3191 3192 3193 |
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... |
3194 |
/* |
9eeff2395 [PATCH] Zone recl... |
3195 3196 |
* Try to free up some pages from this zone through reclaim. */ |
179e96395 [PATCH] vmscan: s... |
3197 |
static int __zone_reclaim(struct zone *zone, gfp_t gfp_mask, unsigned int order) |
9eeff2395 [PATCH] Zone recl... |
3198 |
{ |
7fb2d46d3 [PATCH] zone_recl... |
3199 |
/* Minimum pages needed in order to stay on node */ |
69e05944a [PATCH] vmscan: u... |
3200 |
const unsigned long nr_pages = 1 << order; |
9eeff2395 [PATCH] Zone recl... |
3201 3202 |
struct task_struct *p = current; struct reclaim_state reclaim_state; |
8695949a1 [PATCH] Thin out ... |
3203 |
int priority; |
179e96395 [PATCH] vmscan: s... |
3204 3205 |
struct scan_control sc = { .may_writepage = !!(zone_reclaim_mode & RECLAIM_WRITE), |
a6dc60f89 vmscan: rename sc... |
3206 |
.may_unmap = !!(zone_reclaim_mode & RECLAIM_SWAP), |
2e2e42598 vmscan,memcg: rei... |
3207 |
.may_swap = 1, |
22fba3354 vmscan: separate ... |
3208 3209 |
.nr_to_reclaim = max_t(unsigned long, nr_pages, SWAP_CLUSTER_MAX), |
179e96395 [PATCH] vmscan: s... |
3210 |
.gfp_mask = gfp_mask, |
bd2f6199c vmscan: respect h... |
3211 |
.order = order, |
179e96395 [PATCH] vmscan: s... |
3212 |
}; |
a09ed5e00 vmscan: change sh... |
3213 3214 3215 |
struct shrink_control shrink = { .gfp_mask = sc.gfp_mask, }; |
157480489 vmscan: avoid sub... |
3216 |
unsigned long nr_slab_pages0, nr_slab_pages1; |
9eeff2395 [PATCH] Zone recl... |
3217 |
|
9eeff2395 [PATCH] Zone recl... |
3218 |
cond_resched(); |
d4f7796e9 [PATCH] vmscan: f... |
3219 3220 3221 3222 3223 3224 |
/* * 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... |
3225 |
lockdep_set_current_reclaim_state(gfp_mask); |
9eeff2395 [PATCH] Zone recl... |
3226 3227 |
reclaim_state.reclaimed_slab = 0; p->reclaim_state = &reclaim_state; |
c84db23c6 [PATCH] zone_recl... |
3228 |
|
90afa5de6 vmscan: properly ... |
3229 |
if (zone_pagecache_reclaimable(zone) > zone->min_unmapped_pages) { |
0ff38490c [PATCH] zone_recl... |
3230 3231 3232 3233 3234 3235 |
/* * Free memory by calling shrink zone with increasing * priorities until we have enough memory freed. */ priority = ZONE_RECLAIM_PRIORITY; do { |
a79311c14 vmscan: bail out ... |
3236 |
shrink_zone(priority, zone, &sc); |
0ff38490c [PATCH] zone_recl... |
3237 |
priority--; |
a79311c14 vmscan: bail out ... |
3238 |
} while (priority >= 0 && sc.nr_reclaimed < nr_pages); |
0ff38490c [PATCH] zone_recl... |
3239 |
} |
c84db23c6 [PATCH] zone_recl... |
3240 |
|
157480489 vmscan: avoid sub... |
3241 3242 |
nr_slab_pages0 = zone_page_state(zone, NR_SLAB_RECLAIMABLE); if (nr_slab_pages0 > zone->min_slab_pages) { |
2a16e3f4b [PATCH] Reclaim s... |
3243 |
/* |
7fb2d46d3 [PATCH] zone_recl... |
3244 |
* shrink_slab() does not currently allow us to determine how |
0ff38490c [PATCH] zone_recl... |
3245 3246 3247 3248 |
* 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... |
3249 |
* |
0ff38490c [PATCH] zone_recl... |
3250 3251 |
* Note that shrink_slab will free memory on all zones and may * take a long time. |
2a16e3f4b [PATCH] Reclaim s... |
3252 |
*/ |
4dc4b3d97 vmscan: shrink_sl... |
3253 3254 3255 3256 |
for (;;) { unsigned long lru_pages = zone_reclaimable_pages(zone); /* No reclaimable slab or very low memory pressure */ |
1495f230f vmscan: change sh... |
3257 |
if (!shrink_slab(&shrink, sc.nr_scanned, lru_pages)) |
4dc4b3d97 vmscan: shrink_sl... |
3258 3259 3260 3261 3262 3263 3264 3265 |
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... |
3266 3267 3268 3269 3270 |
/* * Update nr_reclaimed by the number of slab pages we * reclaimed from this zone. */ |
157480489 vmscan: avoid sub... |
3271 3272 3273 |
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... |
3274 |
} |
9eeff2395 [PATCH] Zone recl... |
3275 |
p->reclaim_state = NULL; |
d4f7796e9 [PATCH] vmscan: f... |
3276 |
current->flags &= ~(PF_MEMALLOC | PF_SWAPWRITE); |
76ca542d8 mm, lockdep: anno... |
3277 |
lockdep_clear_current_reclaim_state(); |
a79311c14 vmscan: bail out ... |
3278 |
return sc.nr_reclaimed >= nr_pages; |
9eeff2395 [PATCH] Zone recl... |
3279 |
} |
179e96395 [PATCH] vmscan: s... |
3280 3281 3282 |
int zone_reclaim(struct zone *zone, gfp_t gfp_mask, unsigned int order) { |
179e96395 [PATCH] vmscan: s... |
3283 |
int node_id; |
d773ed6b8 mm: test and set ... |
3284 |
int ret; |
179e96395 [PATCH] vmscan: s... |
3285 3286 |
/* |
0ff38490c [PATCH] zone_recl... |
3287 3288 |
* Zone reclaim reclaims unmapped file backed pages and * slab pages if we are over the defined limits. |
34aa1330f [PATCH] zoned vm ... |
3289 |
* |
9614634fe [PATCH] ZVC/zone_... |
3290 3291 3292 3293 3294 |
* 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... |
3295 |
*/ |
90afa5de6 vmscan: properly ... |
3296 3297 |
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... |
3298 |
return ZONE_RECLAIM_FULL; |
179e96395 [PATCH] vmscan: s... |
3299 |
|
93e4a89a8 mm: restore zone-... |
3300 |
if (zone->all_unreclaimable) |
fa5e084e4 vmscan: do not un... |
3301 |
return ZONE_RECLAIM_FULL; |
d773ed6b8 mm: test and set ... |
3302 |
|
179e96395 [PATCH] vmscan: s... |
3303 |
/* |
d773ed6b8 mm: test and set ... |
3304 |
* Do not scan if the allocation should not be delayed. |
179e96395 [PATCH] vmscan: s... |
3305 |
*/ |
d773ed6b8 mm: test and set ... |
3306 |
if (!(gfp_mask & __GFP_WAIT) || (current->flags & PF_MEMALLOC)) |
fa5e084e4 vmscan: do not un... |
3307 |
return ZONE_RECLAIM_NOSCAN; |
179e96395 [PATCH] vmscan: s... |
3308 3309 3310 3311 3312 3313 3314 |
/* * 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... |
3315 |
node_id = zone_to_nid(zone); |
37c0708db Memoryless nodes:... |
3316 |
if (node_state(node_id, N_CPU) && node_id != numa_node_id()) |
fa5e084e4 vmscan: do not un... |
3317 |
return ZONE_RECLAIM_NOSCAN; |
d773ed6b8 mm: test and set ... |
3318 3319 |
if (zone_test_and_set_flag(zone, ZONE_RECLAIM_LOCKED)) |
fa5e084e4 vmscan: do not un... |
3320 |
return ZONE_RECLAIM_NOSCAN; |
d773ed6b8 mm: test and set ... |
3321 3322 |
ret = __zone_reclaim(zone, gfp_mask, order); zone_clear_flag(zone, ZONE_RECLAIM_LOCKED); |
24cf72518 vmscan: count the... |
3323 3324 |
if (!ret) count_vm_event(PGSCAN_ZONE_RECLAIM_FAILED); |
d773ed6b8 mm: test and set ... |
3325 |
return ret; |
179e96395 [PATCH] vmscan: s... |
3326 |
} |
9eeff2395 [PATCH] Zone recl... |
3327 |
#endif |
894bc3104 Unevictable LRU I... |
3328 |
|
894bc3104 Unevictable LRU I... |
3329 3330 3331 3332 3333 3334 |
/* * 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... |
3335 3336 |
* 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... |
3337 3338 |
* * Reasons page might not be evictable: |
ba9ddf493 Ramfs and Ram Dis... |
3339 |
* (1) page's mapping marked unevictable |
b291f0003 mlock: mlocked pa... |
3340 |
* (2) page is part of an mlocked VMA |
ba9ddf493 Ramfs and Ram Dis... |
3341 |
* |
894bc3104 Unevictable LRU I... |
3342 3343 3344 |
*/ int page_evictable(struct page *page, struct vm_area_struct *vma) { |
ba9ddf493 Ramfs and Ram Dis... |
3345 3346 |
if (mapping_unevictable(page_mapping(page))) return 0; |
b291f0003 mlock: mlocked pa... |
3347 3348 |
if (PageMlocked(page) || (vma && is_mlocked_vma(vma, page))) return 0; |
894bc3104 Unevictable LRU I... |
3349 3350 3351 |
return 1; } |
89e004ea5 SHM_LOCKED pages ... |
3352 3353 3354 3355 3356 3357 3358 3359 3360 3361 3362 3363 3364 3365 |
/** * 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) { |
925b7673c mm: make per-memc... |
3366 |
struct lruvec *lruvec; |
89e004ea5 SHM_LOCKED pages ... |
3367 |
|
925b7673c mm: make per-memc... |
3368 |
VM_BUG_ON(PageActive(page)); |
89e004ea5 SHM_LOCKED pages ... |
3369 3370 3371 |
retry: ClearPageUnevictable(page); if (page_evictable(page, NULL)) { |
401a8e1c1 mm: introduce pag... |
3372 |
enum lru_list l = page_lru_base_type(page); |
af936a160 vmscan: unevictab... |
3373 |
|
89e004ea5 SHM_LOCKED pages ... |
3374 |
__dec_zone_state(zone, NR_UNEVICTABLE); |
925b7673c mm: make per-memc... |
3375 3376 3377 |
lruvec = mem_cgroup_lru_move_lists(zone, page, LRU_UNEVICTABLE, l); list_move(&page->lru, &lruvec->lists[l]); |
89e004ea5 SHM_LOCKED pages ... |
3378 3379 3380 3381 3382 3383 3384 |
__inc_zone_state(zone, NR_INACTIVE_ANON + l); __count_vm_event(UNEVICTABLE_PGRESCUED); } else { /* * rotate unevictable list */ SetPageUnevictable(page); |
925b7673c mm: make per-memc... |
3385 3386 3387 |
lruvec = mem_cgroup_lru_move_lists(zone, page, LRU_UNEVICTABLE, LRU_UNEVICTABLE); list_move(&page->lru, &lruvec->lists[LRU_UNEVICTABLE]); |
89e004ea5 SHM_LOCKED pages ... |
3388 3389 3390 3391 3392 3393 3394 3395 3396 3397 3398 3399 3400 3401 3402 3403 3404 3405 3406 3407 3408 3409 3410 3411 3412 3413 3414 3415 3416 3417 3418 3419 3420 3421 3422 3423 3424 3425 3426 3427 3428 3429 3430 3431 3432 3433 3434 3435 3436 3437 3438 3439 3440 3441 3442 3443 3444 3445 3446 |
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... |
3447 |
|
264e56d82 mm: disable user ... |
3448 |
static void warn_scan_unevictable_pages(void) |
af936a160 vmscan: unevictab... |
3449 |
{ |
264e56d82 mm: disable user ... |
3450 |
printk_once(KERN_WARNING |
25bd91bd2 vmscan: add task ... |
3451 |
"%s: The scan_unevictable_pages sysctl/node-interface has been " |
264e56d82 mm: disable user ... |
3452 |
"disabled for lack of a legitimate use case. If you have " |
25bd91bd2 vmscan: add task ... |
3453 3454 3455 |
"one, please send an email to linux-mm@kvack.org. ", current->comm); |
af936a160 vmscan: unevictab... |
3456 3457 3458 3459 3460 3461 3462 3463 3464 |
} /* * 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... |
3465 |
void __user *buffer, |
af936a160 vmscan: unevictab... |
3466 3467 |
size_t *length, loff_t *ppos) { |
264e56d82 mm: disable user ... |
3468 |
warn_scan_unevictable_pages(); |
8d65af789 sysctl: remove "s... |
3469 |
proc_doulongvec_minmax(table, write, buffer, length, ppos); |
af936a160 vmscan: unevictab... |
3470 3471 3472 |
scan_unevictable_pages = 0; return 0; } |
e4455abb5 mm: only build pe... |
3473 |
#ifdef CONFIG_NUMA |
af936a160 vmscan: unevictab... |
3474 3475 3476 3477 |
/* * 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' ... |
3478 3479 |
static ssize_t read_scan_unevictable_node(struct device *dev, struct device_attribute *attr, |
af936a160 vmscan: unevictab... |
3480 3481 |
char *buf) { |
264e56d82 mm: disable user ... |
3482 |
warn_scan_unevictable_pages(); |
af936a160 vmscan: unevictab... |
3483 3484 3485 |
return sprintf(buf, "0 "); /* always zero; should fit... */ } |
10fbcf4c6 convert 'memory' ... |
3486 3487 |
static ssize_t write_scan_unevictable_node(struct device *dev, struct device_attribute *attr, |
af936a160 vmscan: unevictab... |
3488 3489 |
const char *buf, size_t count) { |
264e56d82 mm: disable user ... |
3490 |
warn_scan_unevictable_pages(); |
af936a160 vmscan: unevictab... |
3491 3492 |
return 1; } |
10fbcf4c6 convert 'memory' ... |
3493 |
static DEVICE_ATTR(scan_unevictable_pages, S_IRUGO | S_IWUSR, |
af936a160 vmscan: unevictab... |
3494 3495 3496 3497 3498 |
read_scan_unevictable_node, write_scan_unevictable_node); int scan_unevictable_register_node(struct node *node) { |
10fbcf4c6 convert 'memory' ... |
3499 |
return device_create_file(&node->dev, &dev_attr_scan_unevictable_pages); |
af936a160 vmscan: unevictab... |
3500 3501 3502 3503 |
} void scan_unevictable_unregister_node(struct node *node) { |
10fbcf4c6 convert 'memory' ... |
3504 |
device_remove_file(&node->dev, &dev_attr_scan_unevictable_pages); |
af936a160 vmscan: unevictab... |
3505 |
} |
e4455abb5 mm: only build pe... |
3506 |
#endif |