Blame view
mm/vmscan.c
116 KB
b24413180 License cleanup: ... |
1 |
// SPDX-License-Identifier: GPL-2.0 |
1da177e4c Linux-2.6.12-rc2 |
2 3 4 5 6 7 8 9 10 11 12 13 |
/* * linux/mm/vmscan.c * * Copyright (C) 1991, 1992, 1993, 1994 Linus Torvalds * * Swap reorganised 29.12.95, Stephen Tweedie. * kswapd added: 7.1.96 sct * Removed kswapd_ctl limits, and swap out as many pages as needed * to bring the system back to freepages.high: 2.4.97, Rik van Riel. * Zone aware kswapd started 02/00, Kanoj Sarcar (kanoj@sgi.com). * Multiqueue VM started 5.8.00, Rik van Riel. */ |
b1de0d139 mm: convert some ... |
14 |
#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt |
1da177e4c Linux-2.6.12-rc2 |
15 |
#include <linux/mm.h> |
5b3cc15af sched/headers: Pr... |
16 |
#include <linux/sched/mm.h> |
1da177e4c Linux-2.6.12-rc2 |
17 |
#include <linux/module.h> |
5a0e3ad6a include cleanup: ... |
18 |
#include <linux/gfp.h> |
1da177e4c Linux-2.6.12-rc2 |
19 20 21 22 23 |
#include <linux/kernel_stat.h> #include <linux/swap.h> #include <linux/pagemap.h> #include <linux/init.h> #include <linux/highmem.h> |
70ddf637e memcg: add memory... |
24 |
#include <linux/vmpressure.h> |
e129b5c23 [PATCH] vm: add p... |
25 |
#include <linux/vmstat.h> |
1da177e4c Linux-2.6.12-rc2 |
26 27 28 29 30 31 |
#include <linux/file.h> #include <linux/writeback.h> #include <linux/blkdev.h> #include <linux/buffer_head.h> /* for try_to_release_page(), buffer_heads_over_limit */ #include <linux/mm_inline.h> |
1da177e4c Linux-2.6.12-rc2 |
32 33 34 35 36 |
#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... |
37 |
#include <linux/compaction.h> |
1da177e4c Linux-2.6.12-rc2 |
38 39 |
#include <linux/notifier.h> #include <linux/rwsem.h> |
248a0301e [PATCH] mm: make ... |
40 |
#include <linux/delay.h> |
3218ae14b [PATCH] pgdat all... |
41 |
#include <linux/kthread.h> |
7dfb71030 [PATCH] Add inclu... |
42 |
#include <linux/freezer.h> |
66e1707bc Memory controller... |
43 |
#include <linux/memcontrol.h> |
873b47717 per-task-delay-ac... |
44 |
#include <linux/delayacct.h> |
af936a160 vmscan: unevictab... |
45 |
#include <linux/sysctl.h> |
929bea7c7 vmscan: all_unrec... |
46 |
#include <linux/oom.h> |
268bb0ce3 sanitize <linux/p... |
47 |
#include <linux/prefetch.h> |
b1de0d139 mm: convert some ... |
48 |
#include <linux/printk.h> |
f9fe48bec dax: support dirt... |
49 |
#include <linux/dax.h> |
1da177e4c Linux-2.6.12-rc2 |
50 51 52 53 54 |
#include <asm/tlbflush.h> #include <asm/div64.h> #include <linux/swapops.h> |
117aad1e9 mm: avoid reinser... |
55 |
#include <linux/balloon_compaction.h> |
1da177e4c Linux-2.6.12-rc2 |
56 |
|
0f8053a50 [PATCH] mm: make ... |
57 |
#include "internal.h" |
33906bc5c vmscan: tracing: ... |
58 59 |
#define CREATE_TRACE_POINTS #include <trace/events/vmscan.h> |
1da177e4c Linux-2.6.12-rc2 |
60 |
struct scan_control { |
22fba3354 vmscan: separate ... |
61 62 |
/* How many pages shrink_list() should reclaim */ unsigned long nr_to_reclaim; |
1da177e4c Linux-2.6.12-rc2 |
63 |
/* This context's GFP mask */ |
6daa0e286 [PATCH] gfp_t: mm... |
64 |
gfp_t gfp_mask; |
1da177e4c Linux-2.6.12-rc2 |
65 |
|
ee814fe23 mm: vmscan: clean... |
66 |
/* Allocation order */ |
5ad333eb6 Lumpy Reclaim V4 |
67 |
int order; |
66e1707bc Memory controller... |
68 |
|
ee814fe23 mm: vmscan: clean... |
69 70 71 72 73 |
/* * Nodemask of nodes allowed by the caller. If NULL, all nodes * are scanned. */ nodemask_t *nodemask; |
9e3b2f8cd mm/vmscan: store ... |
74 |
|
5f53e7629 vmscan: page_chec... |
75 |
/* |
f16015fbf mm: vmscan: disti... |
76 77 78 79 |
* 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... |
80 |
|
ee814fe23 mm: vmscan: clean... |
81 82 |
/* Scan (total_size >> priority) pages at once */ int priority; |
b2e18757f mm, vmscan: begin... |
83 84 |
/* The highest zone to isolate pages for reclaim from */ enum zone_type reclaim_idx; |
1276ad68e mm: vmscan: scan ... |
85 |
/* Writepage batching in laptop mode; RECLAIM_WRITE */ |
ee814fe23 mm: vmscan: clean... |
86 87 88 89 90 91 92 |
unsigned int may_writepage:1; /* Can mapped pages be reclaimed? */ unsigned int may_unmap:1; /* Can pages be swapped as part of reclaim? */ unsigned int may_swap:1; |
d6622f636 mm/vmscan: more r... |
93 94 95 96 97 98 99 |
/* * Cgroups are not reclaimed below their configured memory.low, * unless we threaten to OOM. If any cgroups are skipped due to * memory.low and nothing was reclaimed, go back for memory.low. */ unsigned int memcg_low_reclaim:1; unsigned int memcg_low_skipped:1; |
241994ed8 mm: memcontrol: d... |
100 |
|
ee814fe23 mm: vmscan: clean... |
101 102 103 104 105 106 107 108 109 110 |
unsigned int hibernation_mode:1; /* One of the zones is ready for compaction */ unsigned int compaction_ready:1; /* Incremented by the number of inactive pages that were scanned */ unsigned long nr_scanned; /* Number of pages freed so far during a call to shrink_zones() */ unsigned long nr_reclaimed; |
1da177e4c Linux-2.6.12-rc2 |
111 |
}; |
1da177e4c Linux-2.6.12-rc2 |
112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 |
#ifdef ARCH_HAS_PREFETCH #define prefetch_prev_lru_page(_page, _base, _field) \ do { \ if ((_page)->lru.prev != _base) { \ struct page *prev; \ \ prev = lru_to_page(&(_page->lru)); \ prefetch(&prev->_field); \ } \ } while (0) #else #define prefetch_prev_lru_page(_page, _base, _field) do { } while (0) #endif #ifdef ARCH_HAS_PREFETCHW #define prefetchw_prev_lru_page(_page, _base, _field) \ do { \ if ((_page)->lru.prev != _base) { \ struct page *prev; \ \ prev = lru_to_page(&(_page->lru)); \ prefetchw(&prev->_field); \ } \ } while (0) #else #define prefetchw_prev_lru_page(_page, _base, _field) do { } while (0) #endif /* * From 0 .. 100. Higher means more swappy. */ int vm_swappiness = 60; |
d0480be44 mm: update the de... |
144 145 146 147 148 |
/* * The total number of pages which are beyond the high watermark within all * zones. */ unsigned long vm_total_pages; |
1da177e4c Linux-2.6.12-rc2 |
149 150 151 |
static LIST_HEAD(shrinker_list); static DECLARE_RWSEM(shrinker_rwsem); |
c255a4580 memcg: rename con... |
152 |
#ifdef CONFIG_MEMCG |
89b5fae53 mm: vmscan: disti... |
153 154 |
static bool global_reclaim(struct scan_control *sc) { |
f16015fbf mm: vmscan: disti... |
155 |
return !sc->target_mem_cgroup; |
89b5fae53 mm: vmscan: disti... |
156 |
} |
97c9341f7 mm: vmscan: disab... |
157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 |
/** * sane_reclaim - is the usual dirty throttling mechanism operational? * @sc: scan_control in question * * The normal page dirty throttling mechanism in balance_dirty_pages() is * completely broken with the legacy memcg and direct stalling in * shrink_page_list() is used for throttling instead, which lacks all the * niceties such as fairness, adaptive pausing, bandwidth proportional * allocation and configurability. * * This function tests whether the vmscan currently in progress can assume * that the normal dirty throttling mechanism is operational. */ static bool sane_reclaim(struct scan_control *sc) { struct mem_cgroup *memcg = sc->target_mem_cgroup; if (!memcg) return true; #ifdef CONFIG_CGROUP_WRITEBACK |
69234acee Merge branch 'for... |
178 |
if (cgroup_subsys_on_dfl(memory_cgrp_subsys)) |
97c9341f7 mm: vmscan: disab... |
179 180 181 182 |
return true; #endif return false; } |
91a45470f per-zone and recl... |
183 |
#else |
89b5fae53 mm: vmscan: disti... |
184 185 186 187 |
static bool global_reclaim(struct scan_control *sc) { return true; } |
97c9341f7 mm: vmscan: disab... |
188 189 190 191 192 |
static bool sane_reclaim(struct scan_control *sc) { return true; } |
91a45470f per-zone and recl... |
193 |
#endif |
5a1c84b40 mm: remove reclai... |
194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 |
/* * This misses isolated pages which are not accounted for to save counters. * As the data only determines if reclaim or compaction continues, it is * not expected that isolated pages will be a dominating factor. */ unsigned long zone_reclaimable_pages(struct zone *zone) { unsigned long nr; nr = zone_page_state_snapshot(zone, NR_ZONE_INACTIVE_FILE) + zone_page_state_snapshot(zone, NR_ZONE_ACTIVE_FILE); if (get_nr_swap_pages() > 0) nr += zone_page_state_snapshot(zone, NR_ZONE_INACTIVE_ANON) + zone_page_state_snapshot(zone, NR_ZONE_ACTIVE_ANON); return nr; } |
599d0c954 mm, vmscan: move ... |
211 212 213 214 215 216 217 |
unsigned long pgdat_reclaimable_pages(struct pglist_data *pgdat) { unsigned long nr; nr = node_page_state_snapshot(pgdat, NR_ACTIVE_FILE) + node_page_state_snapshot(pgdat, NR_INACTIVE_FILE) + node_page_state_snapshot(pgdat, NR_ISOLATED_FILE); |
6e543d578 mm: vmscan: fix d... |
218 219 |
if (get_nr_swap_pages() > 0) |
599d0c954 mm, vmscan: move ... |
220 221 222 |
nr += node_page_state_snapshot(pgdat, NR_ACTIVE_ANON) + node_page_state_snapshot(pgdat, NR_INACTIVE_ANON) + node_page_state_snapshot(pgdat, NR_ISOLATED_ANON); |
6e543d578 mm: vmscan: fix d... |
223 224 225 |
return nr; } |
fd5388037 mm, vmscan: clean... |
226 227 228 229 230 231 232 |
/** * lruvec_lru_size - Returns the number of pages on the given LRU list. * @lruvec: lru vector * @lru: lru to use * @zone_idx: zones to consider (use MAX_NR_ZONES for the whole LRU list) */ unsigned long lruvec_lru_size(struct lruvec *lruvec, enum lru_list lru, int zone_idx) |
c9f299d98 mm: add zone nr_p... |
233 |
{ |
fd5388037 mm, vmscan: clean... |
234 235 |
unsigned long lru_size; int zid; |
c3c787e8c mm/memcg: scannin... |
236 |
if (!mem_cgroup_disabled()) |
fd5388037 mm, vmscan: clean... |
237 238 239 |
lru_size = mem_cgroup_get_lru_size(lruvec, lru); else lru_size = node_page_state(lruvec_pgdat(lruvec), NR_LRU_BASE + lru); |
a3d8e0549 memcg: add mem_cg... |
240 |
|
fd5388037 mm, vmscan: clean... |
241 242 243 |
for (zid = zone_idx + 1; zid < MAX_NR_ZONES; zid++) { struct zone *zone = &lruvec_pgdat(lruvec)->node_zones[zid]; unsigned long size; |
c9f299d98 mm: add zone nr_p... |
244 |
|
fd5388037 mm, vmscan: clean... |
245 246 247 248 249 250 251 252 253 254 255 256 |
if (!managed_zone(zone)) continue; if (!mem_cgroup_disabled()) size = mem_cgroup_get_zone_lru_size(lruvec, lru, zid); else size = zone_page_state(&lruvec_pgdat(lruvec)->node_zones[zid], NR_ZONE_LRU_BASE + lru); lru_size -= min(size, lru_size); } return lru_size; |
b4536f0c8 mm, memcg: fix th... |
257 |
|
b4536f0c8 mm, memcg: fix th... |
258 |
} |
1da177e4c Linux-2.6.12-rc2 |
259 |
/* |
1d3d4437e vmscan: per-node ... |
260 |
* Add a shrinker callback to be called from the vm. |
1da177e4c Linux-2.6.12-rc2 |
261 |
*/ |
1d3d4437e vmscan: per-node ... |
262 |
int register_shrinker(struct shrinker *shrinker) |
1da177e4c Linux-2.6.12-rc2 |
263 |
{ |
1d3d4437e vmscan: per-node ... |
264 |
size_t size = sizeof(*shrinker->nr_deferred); |
1d3d4437e vmscan: per-node ... |
265 266 267 268 269 270 |
if (shrinker->flags & SHRINKER_NUMA_AWARE) size *= nr_node_ids; shrinker->nr_deferred = kzalloc(size, GFP_KERNEL); if (!shrinker->nr_deferred) return -ENOMEM; |
8e1f936b7 mm: clean up and ... |
271 272 273 |
down_write(&shrinker_rwsem); list_add_tail(&shrinker->list, &shrinker_list); up_write(&shrinker_rwsem); |
1d3d4437e vmscan: per-node ... |
274 |
return 0; |
1da177e4c Linux-2.6.12-rc2 |
275 |
} |
8e1f936b7 mm: clean up and ... |
276 |
EXPORT_SYMBOL(register_shrinker); |
1da177e4c Linux-2.6.12-rc2 |
277 278 279 280 |
/* * Remove one */ |
8e1f936b7 mm: clean up and ... |
281 |
void unregister_shrinker(struct shrinker *shrinker) |
1da177e4c Linux-2.6.12-rc2 |
282 |
{ |
7880fc541 mm,vmscan: Make u... |
283 284 |
if (!shrinker->nr_deferred) return; |
1da177e4c Linux-2.6.12-rc2 |
285 286 287 |
down_write(&shrinker_rwsem); list_del(&shrinker->list); up_write(&shrinker_rwsem); |
ae3933216 mm/vmscan.c: don'... |
288 |
kfree(shrinker->nr_deferred); |
7880fc541 mm,vmscan: Make u... |
289 |
shrinker->nr_deferred = NULL; |
1da177e4c Linux-2.6.12-rc2 |
290 |
} |
8e1f936b7 mm: clean up and ... |
291 |
EXPORT_SYMBOL(unregister_shrinker); |
1da177e4c Linux-2.6.12-rc2 |
292 293 |
#define SHRINK_BATCH 128 |
1d3d4437e vmscan: per-node ... |
294 |
|
cb731d6c6 vmscan: per memor... |
295 296 297 298 |
static unsigned long do_shrink_slab(struct shrink_control *shrinkctl, struct shrinker *shrinker, unsigned long nr_scanned, unsigned long nr_eligible) |
1d3d4437e vmscan: per-node ... |
299 300 301 302 |
{ unsigned long freed = 0; unsigned long long delta; long total_scan; |
d5bc5fd3f mm: vmscan: shrin... |
303 |
long freeable; |
1d3d4437e vmscan: per-node ... |
304 305 306 307 308 |
long nr; long new_nr; int nid = shrinkctl->nid; long batch_size = shrinker->batch ? shrinker->batch : SHRINK_BATCH; |
5f33a0803 mm/vmscan.c: set ... |
309 |
long scanned = 0, next_deferred; |
1d3d4437e vmscan: per-node ... |
310 |
|
d5bc5fd3f mm: vmscan: shrin... |
311 312 |
freeable = shrinker->count_objects(shrinker, shrinkctl); if (freeable == 0) |
1d3d4437e vmscan: per-node ... |
313 314 315 316 317 318 319 320 321 322 |
return 0; /* * copy the current shrinker scan count into a local variable * and zero it so that other concurrent shrinker invocations * don't also do this scanning work. */ nr = atomic_long_xchg(&shrinker->nr_deferred[nid], 0); total_scan = nr; |
6b4f7799c mm: vmscan: invok... |
323 |
delta = (4 * nr_scanned) / shrinker->seeks; |
d5bc5fd3f mm: vmscan: shrin... |
324 |
delta *= freeable; |
6b4f7799c mm: vmscan: invok... |
325 |
do_div(delta, nr_eligible + 1); |
1d3d4437e vmscan: per-node ... |
326 327 |
total_scan += delta; if (total_scan < 0) { |
8612c6639 mm/vmscan.c: repl... |
328 329 |
pr_err("shrink_slab: %pF negative objects to delete nr=%ld ", |
a0b02131c shrinker: Kill ol... |
330 |
shrinker->scan_objects, total_scan); |
d5bc5fd3f mm: vmscan: shrin... |
331 |
total_scan = freeable; |
5f33a0803 mm/vmscan.c: set ... |
332 333 334 |
next_deferred = nr; } else next_deferred = total_scan; |
1d3d4437e vmscan: per-node ... |
335 336 337 338 339 340 341 |
/* * We need to avoid excessive windup on filesystem shrinkers * due to large numbers of GFP_NOFS allocations causing the * shrinkers to return -1 all the time. This results in a large * nr being built up so when a shrink that can do some work * comes along it empties the entire cache due to nr >>> |
d5bc5fd3f mm: vmscan: shrin... |
342 |
* freeable. This is bad for sustaining a working set in |
1d3d4437e vmscan: per-node ... |
343 344 345 346 347 |
* memory. * * Hence only allow the shrinker to scan the entire cache when * a large delta change is calculated directly. */ |
d5bc5fd3f mm: vmscan: shrin... |
348 349 |
if (delta < freeable / 4) total_scan = min(total_scan, freeable / 2); |
1d3d4437e vmscan: per-node ... |
350 351 352 353 354 355 |
/* * Avoid risking looping forever due to too large nr value: * never try to free more than twice the estimate number of * freeable entries. */ |
d5bc5fd3f mm: vmscan: shrin... |
356 357 |
if (total_scan > freeable * 2) total_scan = freeable * 2; |
1d3d4437e vmscan: per-node ... |
358 359 |
trace_mm_shrink_slab_start(shrinker, shrinkctl, nr, |
6b4f7799c mm: vmscan: invok... |
360 361 |
nr_scanned, nr_eligible, freeable, delta, total_scan); |
1d3d4437e vmscan: per-node ... |
362 |
|
0b1fb40a3 mm: vmscan: shrin... |
363 364 365 366 367 368 369 370 371 372 373 |
/* * Normally, we should not scan less than batch_size objects in one * pass to avoid too frequent shrinker calls, but if the slab has less * than batch_size objects in total and we are really tight on memory, * we will try to reclaim all available objects, otherwise we can end * up failing allocations although there are plenty of reclaimable * objects spread over several slabs with usage less than the * batch_size. * * We detect the "tight on memory" situations by looking at the total * number of objects we want to scan (total_scan). If it is greater |
d5bc5fd3f mm: vmscan: shrin... |
374 |
* than the total number of objects on slab (freeable), we must be |
0b1fb40a3 mm: vmscan: shrin... |
375 376 377 378 |
* scanning at high prio and therefore should try to reclaim as much as * possible. */ while (total_scan >= batch_size || |
d5bc5fd3f mm: vmscan: shrin... |
379 |
total_scan >= freeable) { |
a0b02131c shrinker: Kill ol... |
380 |
unsigned long ret; |
0b1fb40a3 mm: vmscan: shrin... |
381 |
unsigned long nr_to_scan = min(batch_size, total_scan); |
1d3d4437e vmscan: per-node ... |
382 |
|
0b1fb40a3 mm: vmscan: shrin... |
383 |
shrinkctl->nr_to_scan = nr_to_scan; |
d460acb5b mm: track actual ... |
384 |
shrinkctl->nr_scanned = nr_to_scan; |
a0b02131c shrinker: Kill ol... |
385 386 387 388 |
ret = shrinker->scan_objects(shrinker, shrinkctl); if (ret == SHRINK_STOP) break; freed += ret; |
1d3d4437e vmscan: per-node ... |
389 |
|
d460acb5b mm: track actual ... |
390 391 392 |
count_vm_events(SLABS_SCANNED, shrinkctl->nr_scanned); total_scan -= shrinkctl->nr_scanned; scanned += shrinkctl->nr_scanned; |
1d3d4437e vmscan: per-node ... |
393 394 395 |
cond_resched(); } |
5f33a0803 mm/vmscan.c: set ... |
396 397 398 399 |
if (next_deferred >= scanned) next_deferred -= scanned; else next_deferred = 0; |
1d3d4437e vmscan: per-node ... |
400 401 402 403 404 |
/* * 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. */ |
5f33a0803 mm/vmscan.c: set ... |
405 406 |
if (next_deferred > 0) new_nr = atomic_long_add_return(next_deferred, |
1d3d4437e vmscan: per-node ... |
407 408 409 |
&shrinker->nr_deferred[nid]); else new_nr = atomic_long_read(&shrinker->nr_deferred[nid]); |
df9024a8c mm: shrinker: add... |
410 |
trace_mm_shrink_slab_end(shrinker, nid, freed, nr, new_nr, total_scan); |
1d3d4437e vmscan: per-node ... |
411 |
return freed; |
1495f230f vmscan: change sh... |
412 |
} |
6b4f7799c mm: vmscan: invok... |
413 |
/** |
cb731d6c6 vmscan: per memor... |
414 |
* shrink_slab - shrink slab caches |
6b4f7799c mm: vmscan: invok... |
415 416 |
* @gfp_mask: allocation context * @nid: node whose slab caches to target |
cb731d6c6 vmscan: per memor... |
417 |
* @memcg: memory cgroup whose slab caches to target |
6b4f7799c mm: vmscan: invok... |
418 419 |
* @nr_scanned: pressure numerator * @nr_eligible: pressure denominator |
1da177e4c Linux-2.6.12-rc2 |
420 |
* |
6b4f7799c mm: vmscan: invok... |
421 |
* Call the shrink functions to age shrinkable caches. |
1da177e4c Linux-2.6.12-rc2 |
422 |
* |
6b4f7799c mm: vmscan: invok... |
423 424 |
* @nid is passed along to shrinkers with SHRINKER_NUMA_AWARE set, * unaware shrinkers will receive a node id of 0 instead. |
1da177e4c Linux-2.6.12-rc2 |
425 |
* |
cb731d6c6 vmscan: per memor... |
426 427 |
* @memcg specifies the memory cgroup to target. If it is not NULL, * only shrinkers with SHRINKER_MEMCG_AWARE set will be called to scan |
0fc9f58a9 mm: vmscan: pass ... |
428 429 |
* objects from the memory cgroup specified. Otherwise, only unaware * shrinkers are called. |
cb731d6c6 vmscan: per memor... |
430 |
* |
6b4f7799c mm: vmscan: invok... |
431 432 433 434 435 436 437 |
* @nr_scanned and @nr_eligible form a ratio that indicate how much of * the available objects should be scanned. Page reclaim for example * passes the number of pages scanned and the number of pages on the * LRU lists that it considered on @nid, plus a bias in @nr_scanned * when it encountered mapped pages. The ratio is further biased by * the ->seeks setting of the shrink function, which indicates the * cost to recreate an object relative to that of an LRU page. |
b15e0905f [PATCH] vmscan: n... |
438 |
* |
6b4f7799c mm: vmscan: invok... |
439 |
* Returns the number of reclaimed slab objects. |
1da177e4c Linux-2.6.12-rc2 |
440 |
*/ |
cb731d6c6 vmscan: per memor... |
441 442 443 444 |
static unsigned long shrink_slab(gfp_t gfp_mask, int nid, struct mem_cgroup *memcg, unsigned long nr_scanned, unsigned long nr_eligible) |
1da177e4c Linux-2.6.12-rc2 |
445 446 |
{ struct shrinker *shrinker; |
24f7c6b98 mm: new shrinker API |
447 |
unsigned long freed = 0; |
1da177e4c Linux-2.6.12-rc2 |
448 |
|
0fc9f58a9 mm: vmscan: pass ... |
449 |
if (memcg && (!memcg_kmem_enabled() || !mem_cgroup_online(memcg))) |
cb731d6c6 vmscan: per memor... |
450 |
return 0; |
6b4f7799c mm: vmscan: invok... |
451 452 |
if (nr_scanned == 0) nr_scanned = SWAP_CLUSTER_MAX; |
1da177e4c Linux-2.6.12-rc2 |
453 |
|
f06590bd7 mm: vmscan: corre... |
454 |
if (!down_read_trylock(&shrinker_rwsem)) { |
24f7c6b98 mm: new shrinker API |
455 456 457 458 459 460 461 |
/* * If we would return 0, our callers would understand that we * have nothing else to shrink and give up trying. By returning * 1 we keep it going and assume we'll be able to shrink next * time. */ freed = 1; |
f06590bd7 mm: vmscan: corre... |
462 463 |
goto out; } |
1da177e4c Linux-2.6.12-rc2 |
464 465 |
list_for_each_entry(shrinker, &shrinker_list, list) { |
6b4f7799c mm: vmscan: invok... |
466 467 468 |
struct shrink_control sc = { .gfp_mask = gfp_mask, .nid = nid, |
cb731d6c6 vmscan: per memor... |
469 |
.memcg = memcg, |
6b4f7799c mm: vmscan: invok... |
470 |
}; |
ec97097bc mm: vmscan: call ... |
471 |
|
0fc9f58a9 mm: vmscan: pass ... |
472 473 474 475 476 477 478 |
/* * If kernel memory accounting is disabled, we ignore * SHRINKER_MEMCG_AWARE flag and call all shrinkers * passing NULL for memcg. */ if (memcg_kmem_enabled() && !!memcg != !!(shrinker->flags & SHRINKER_MEMCG_AWARE)) |
cb731d6c6 vmscan: per memor... |
479 |
continue; |
6b4f7799c mm: vmscan: invok... |
480 481 |
if (!(shrinker->flags & SHRINKER_NUMA_AWARE)) sc.nid = 0; |
1da177e4c Linux-2.6.12-rc2 |
482 |
|
cb731d6c6 vmscan: per memor... |
483 |
freed += do_shrink_slab(&sc, shrinker, nr_scanned, nr_eligible); |
1da177e4c Linux-2.6.12-rc2 |
484 |
} |
6b4f7799c mm: vmscan: invok... |
485 |
|
1da177e4c Linux-2.6.12-rc2 |
486 |
up_read(&shrinker_rwsem); |
f06590bd7 mm: vmscan: corre... |
487 488 |
out: cond_resched(); |
24f7c6b98 mm: new shrinker API |
489 |
return freed; |
1da177e4c Linux-2.6.12-rc2 |
490 |
} |
cb731d6c6 vmscan: per memor... |
491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 |
void drop_slab_node(int nid) { unsigned long freed; do { struct mem_cgroup *memcg = NULL; freed = 0; do { freed += shrink_slab(GFP_KERNEL, nid, memcg, 1000, 1000); } while ((memcg = mem_cgroup_iter(NULL, memcg, NULL)) != NULL); } while (freed > 10); } void drop_slab(void) { int nid; for_each_online_node(nid) drop_slab_node(nid); } |
1da177e4c Linux-2.6.12-rc2 |
513 514 |
static inline int is_page_cache_freeable(struct page *page) { |
ceddc3a52 mm: document is_p... |
515 516 517 518 519 |
/* * 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. */ |
bd4c82c22 mm, THP, swap: de... |
520 521 522 |
int radix_pins = PageTransHuge(page) && PageSwapCache(page) ? HPAGE_PMD_NR : 1; return page_count(page) - page_has_private(page) == 1 + radix_pins; |
1da177e4c Linux-2.6.12-rc2 |
523 |
} |
703c27088 writeback: implem... |
524 |
static int may_write_to_inode(struct inode *inode, struct scan_control *sc) |
1da177e4c Linux-2.6.12-rc2 |
525 |
{ |
930d91525 [PATCH] Swap Migr... |
526 |
if (current->flags & PF_SWAPWRITE) |
1da177e4c Linux-2.6.12-rc2 |
527 |
return 1; |
703c27088 writeback: implem... |
528 |
if (!inode_write_congested(inode)) |
1da177e4c Linux-2.6.12-rc2 |
529 |
return 1; |
703c27088 writeback: implem... |
530 |
if (inode_to_bdi(inode) == current->backing_dev_info) |
1da177e4c Linux-2.6.12-rc2 |
531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 |
return 1; return 0; } /* * We detected a synchronous write error writing a page out. Probably * -ENOSPC. We need to propagate that into the address_space for a subsequent * fsync(), msync() or close(). * * The tricky part is that after writepage we cannot touch the mapping: nothing * prevents it from being freed up. But we have a ref on the page and once * that page is locked, the mapping is pinned. * * We're allowed to run sleeping lock_page() here because we know the caller has * __GFP_FS. */ static void handle_write_error(struct address_space *mapping, struct page *page, int error) { |
7eaceacca block: remove per... |
550 |
lock_page(page); |
3e9f45bd1 Factor outstandin... |
551 552 |
if (page_mapping(page) == mapping) mapping_set_error(mapping, error); |
1da177e4c Linux-2.6.12-rc2 |
553 554 |
unlock_page(page); } |
04e62a29b [PATCH] More page... |
555 556 557 558 559 560 561 562 563 564 565 |
/* 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 |
566 |
/* |
1742f19fa [PATCH] vmscan: r... |
567 568 |
* pageout is called by shrink_page_list() for each dirty page. * Calls ->writepage(). |
1da177e4c Linux-2.6.12-rc2 |
569 |
*/ |
c661b078f synchronous lumpy... |
570 |
static pageout_t pageout(struct page *page, struct address_space *mapping, |
7d3579e8e vmscan: narrow th... |
571 |
struct scan_control *sc) |
1da177e4c Linux-2.6.12-rc2 |
572 573 574 575 576 577 578 579 |
{ /* * If the page is dirty, only perform writeback if that write * will be non-blocking. To prevent this allocation from being * stalled by pagecache activity. But note that there may be * stalls if we need to run get_block(). We could test * PagePrivate for that. * |
8174202b3 write_iter varian... |
580 |
* If this process is currently in __generic_file_write_iter() against |
1da177e4c Linux-2.6.12-rc2 |
581 582 583 584 585 586 587 |
* 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 |
588 589 590 591 592 593 594 595 |
*/ 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... |
596 |
if (page_has_private(page)) { |
1da177e4c Linux-2.6.12-rc2 |
597 598 |
if (try_to_free_buffers(page)) { ClearPageDirty(page); |
b1de0d139 mm: convert some ... |
599 600 |
pr_info("%s: orphaned page ", __func__); |
1da177e4c Linux-2.6.12-rc2 |
601 602 603 604 605 606 607 |
return PAGE_CLEAN; } } return PAGE_KEEP; } if (mapping->a_ops->writepage == NULL) return PAGE_ACTIVATE; |
703c27088 writeback: implem... |
608 |
if (!may_write_to_inode(mapping->host, sc)) |
1da177e4c Linux-2.6.12-rc2 |
609 610 611 612 613 614 615 |
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... |
616 617 |
.range_start = 0, .range_end = LLONG_MAX, |
1da177e4c Linux-2.6.12-rc2 |
618 619 620 621 622 623 624 |
.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... |
625 |
if (res == AOP_WRITEPAGE_ACTIVATE) { |
1da177e4c Linux-2.6.12-rc2 |
626 627 628 |
ClearPageReclaim(page); return PAGE_ACTIVATE; } |
c661b078f synchronous lumpy... |
629 |
|
1da177e4c Linux-2.6.12-rc2 |
630 631 632 633 |
if (!PageWriteback(page)) { /* synchronous write or broken a_ops? */ ClearPageReclaim(page); } |
3aa238511 mm/vmscan.c: chan... |
634 |
trace_mm_vmscan_writepage(page); |
c4a25635b mm: move vmscan w... |
635 |
inc_node_page_state(page, NR_VMSCAN_WRITE); |
1da177e4c Linux-2.6.12-rc2 |
636 637 638 639 640 |
return PAGE_SUCCESS; } return PAGE_CLEAN; } |
a649fd927 [PATCH] invalidat... |
641 |
/* |
e286781d5 mm: speculative p... |
642 643 |
* Same as remove_mapping, but if the page is removed from the mapping, it * gets returned with a refcount of 0. |
a649fd927 [PATCH] invalidat... |
644 |
*/ |
a528910e1 mm: thrash detect... |
645 646 |
static int __remove_mapping(struct address_space *mapping, struct page *page, bool reclaimed) |
49d2e9cc4 [PATCH] Swap Migr... |
647 |
{ |
c4843a759 memcg: add per cg... |
648 |
unsigned long flags; |
bd4c82c22 mm, THP, swap: de... |
649 |
int refcount; |
c4843a759 memcg: add per cg... |
650 |
|
28e4d965e [PATCH] mm: remov... |
651 652 |
BUG_ON(!PageLocked(page)); BUG_ON(mapping != page_mapping(page)); |
49d2e9cc4 [PATCH] Swap Migr... |
653 |
|
c4843a759 memcg: add per cg... |
654 |
spin_lock_irqsave(&mapping->tree_lock, flags); |
49d2e9cc4 [PATCH] Swap Migr... |
655 |
/* |
0fd0e6b05 [PATCH] page inva... |
656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 |
* 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 |
0139aa7b7 mm: rename _count... |
675 |
* load is not satisfied before that of page->_refcount. |
0fd0e6b05 [PATCH] page inva... |
676 677 678 |
* * 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... |
679 |
*/ |
bd4c82c22 mm, THP, swap: de... |
680 681 682 683 684 |
if (unlikely(PageTransHuge(page)) && PageSwapCache(page)) refcount = 1 + HPAGE_PMD_NR; else refcount = 2; if (!page_ref_freeze(page, refcount)) |
49d2e9cc4 [PATCH] Swap Migr... |
685 |
goto cannot_free; |
e286781d5 mm: speculative p... |
686 687 |
/* note: atomic_cmpxchg in page_freeze_refs provides the smp_rmb */ if (unlikely(PageDirty(page))) { |
bd4c82c22 mm, THP, swap: de... |
688 |
page_ref_unfreeze(page, refcount); |
49d2e9cc4 [PATCH] Swap Migr... |
689 |
goto cannot_free; |
e286781d5 mm: speculative p... |
690 |
} |
49d2e9cc4 [PATCH] Swap Migr... |
691 692 693 |
if (PageSwapCache(page)) { swp_entry_t swap = { .val = page_private(page) }; |
0a31bc97c mm: memcontrol: r... |
694 |
mem_cgroup_swapout(page, swap); |
49d2e9cc4 [PATCH] Swap Migr... |
695 |
__delete_from_swap_cache(page); |
c4843a759 memcg: add per cg... |
696 |
spin_unlock_irqrestore(&mapping->tree_lock, flags); |
75f6d6d29 mm, THP, swap: un... |
697 |
put_swap_page(page, swap); |
e286781d5 mm: speculative p... |
698 |
} else { |
6072d13c4 Call the filesyst... |
699 |
void (*freepage)(struct page *); |
a528910e1 mm: thrash detect... |
700 |
void *shadow = NULL; |
6072d13c4 Call the filesyst... |
701 702 |
freepage = mapping->a_ops->freepage; |
a528910e1 mm: thrash detect... |
703 704 705 706 707 708 709 710 711 |
/* * Remember a shadow entry for reclaimed file cache in * order to detect refaults, thus thrashing, later on. * * But don't store shadows in an address space that is * already exiting. This is not just an optizimation, * inode reclaim needs to empty out the radix tree or * the nodes are lost. Don't plant shadows behind its * back. |
f9fe48bec dax: support dirt... |
712 713 714 715 716 717 |
* * We also don't store shadows for DAX mappings because the * only page cache pages found in these are zero pages * covering holes, and because we don't want to mix DAX * exceptional entries and shadow exceptional entries in the * same page_tree. |
a528910e1 mm: thrash detect... |
718 719 |
*/ if (reclaimed && page_is_file_cache(page) && |
f9fe48bec dax: support dirt... |
720 |
!mapping_exiting(mapping) && !dax_mapping(mapping)) |
a528910e1 mm: thrash detect... |
721 |
shadow = workingset_eviction(mapping, page); |
62cccb8c8 mm: simplify lock... |
722 |
__delete_from_page_cache(page, shadow); |
c4843a759 memcg: add per cg... |
723 |
spin_unlock_irqrestore(&mapping->tree_lock, flags); |
6072d13c4 Call the filesyst... |
724 725 726 |
if (freepage != NULL) freepage(page); |
49d2e9cc4 [PATCH] Swap Migr... |
727 |
} |
49d2e9cc4 [PATCH] Swap Migr... |
728 729 730 |
return 1; cannot_free: |
c4843a759 memcg: add per cg... |
731 |
spin_unlock_irqrestore(&mapping->tree_lock, flags); |
49d2e9cc4 [PATCH] Swap Migr... |
732 733 |
return 0; } |
1da177e4c Linux-2.6.12-rc2 |
734 |
/* |
e286781d5 mm: speculative p... |
735 736 737 738 739 740 741 |
* Attempt to detach a locked page from its ->mapping. If it is dirty or if * someone else has a ref on the page, abort and return 0. If it was * successfully detached, return 1. Assumes the caller has a single ref on * this page. */ int remove_mapping(struct address_space *mapping, struct page *page) { |
a528910e1 mm: thrash detect... |
742 |
if (__remove_mapping(mapping, page, false)) { |
e286781d5 mm: speculative p... |
743 744 745 746 747 |
/* * Unfreezing the refcount with 1 rather than 2 effectively * drops the pagecache ref for us without requiring another * atomic operation. */ |
fe896d187 mm: introduce pag... |
748 |
page_ref_unfreeze(page, 1); |
e286781d5 mm: speculative p... |
749 750 751 752 |
return 1; } return 0; } |
894bc3104 Unevictable LRU I... |
753 754 755 756 757 758 759 760 761 |
/** * 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... |
762 763 |
void putback_lru_page(struct page *page) { |
0ec3b74c7 mm: putback_lru_p... |
764 |
bool is_unevictable; |
bbfd28eee unevictable lru: ... |
765 |
int was_unevictable = PageUnevictable(page); |
894bc3104 Unevictable LRU I... |
766 |
|
309381fea mm: dump page whe... |
767 |
VM_BUG_ON_PAGE(PageLRU(page), page); |
894bc3104 Unevictable LRU I... |
768 769 770 |
redo: ClearPageUnevictable(page); |
39b5f29ac mm: remove vma ar... |
771 |
if (page_evictable(page)) { |
894bc3104 Unevictable LRU I... |
772 773 774 775 776 777 |
/* * For evictable pages, we can use the cache. * In event of a race, worst case is we end up with an * unevictable page on [in]active list. * We know how to handle that. */ |
0ec3b74c7 mm: putback_lru_p... |
778 |
is_unevictable = false; |
c53954a09 mm: remove lru pa... |
779 |
lru_cache_add(page); |
894bc3104 Unevictable LRU I... |
780 781 782 783 784 |
} else { /* * Put unevictable pages directly on zone's unevictable * list. */ |
0ec3b74c7 mm: putback_lru_p... |
785 |
is_unevictable = true; |
894bc3104 Unevictable LRU I... |
786 |
add_page_to_unevictable_list(page); |
6a7b95481 vmscan: order evi... |
787 |
/* |
21ee9f398 vmscan: add barri... |
788 789 790 |
* When racing with an mlock or AS_UNEVICTABLE clearing * (page is unlocked) make sure that if the other thread * does not observe our setting of PG_lru and fails |
245132643 SHM_UNLOCK: fix U... |
791 |
* isolation/check_move_unevictable_pages, |
21ee9f398 vmscan: add barri... |
792 |
* we see PG_mlocked/AS_UNEVICTABLE cleared below and move |
6a7b95481 vmscan: order evi... |
793 794 |
* the page back to the evictable list. * |
21ee9f398 vmscan: add barri... |
795 |
* The other side is TestClearPageMlocked() or shmem_lock(). |
6a7b95481 vmscan: order evi... |
796 797 |
*/ smp_mb(); |
894bc3104 Unevictable LRU I... |
798 |
} |
894bc3104 Unevictable LRU I... |
799 800 801 802 803 804 |
/* * page's status can change while we move it among lru. If an evictable * page is on unevictable list, it never be freed. To avoid that, * check after we added it to the list, again. */ |
0ec3b74c7 mm: putback_lru_p... |
805 |
if (is_unevictable && page_evictable(page)) { |
894bc3104 Unevictable LRU I... |
806 807 808 809 810 811 812 813 814 |
if (!isolate_lru_page(page)) { put_page(page); goto redo; } /* This means someone else dropped this page from LRU * So, it will be freed or putback to LRU again. There is * nothing to do here. */ } |
0ec3b74c7 mm: putback_lru_p... |
815 |
if (was_unevictable && !is_unevictable) |
bbfd28eee unevictable lru: ... |
816 |
count_vm_event(UNEVICTABLE_PGRESCUED); |
0ec3b74c7 mm: putback_lru_p... |
817 |
else if (!was_unevictable && is_unevictable) |
bbfd28eee unevictable lru: ... |
818 |
count_vm_event(UNEVICTABLE_PGCULLED); |
894bc3104 Unevictable LRU I... |
819 820 |
put_page(page); /* drop ref from isolate */ } |
dfc8d636c vmscan: factor ou... |
821 822 823 |
enum page_references { PAGEREF_RECLAIM, PAGEREF_RECLAIM_CLEAN, |
645747462 vmscan: detect ma... |
824 |
PAGEREF_KEEP, |
dfc8d636c vmscan: factor ou... |
825 826 827 828 829 830 |
PAGEREF_ACTIVATE, }; static enum page_references page_check_references(struct page *page, struct scan_control *sc) { |
645747462 vmscan: detect ma... |
831 |
int referenced_ptes, referenced_page; |
dfc8d636c vmscan: factor ou... |
832 |
unsigned long vm_flags; |
dfc8d636c vmscan: factor ou... |
833 |
|
c3ac9a8ad mm: memcg: count ... |
834 835 |
referenced_ptes = page_referenced(page, 1, sc->target_mem_cgroup, &vm_flags); |
645747462 vmscan: detect ma... |
836 |
referenced_page = TestClearPageReferenced(page); |
dfc8d636c vmscan: factor ou... |
837 |
|
dfc8d636c vmscan: factor ou... |
838 839 840 841 842 843 |
/* * 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... |
844 |
if (referenced_ptes) { |
e48982734 mm: consider all ... |
845 |
if (PageSwapBacked(page)) |
645747462 vmscan: detect ma... |
846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 |
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... |
862 |
if (referenced_page || referenced_ptes > 1) |
645747462 vmscan: detect ma... |
863 |
return PAGEREF_ACTIVATE; |
c909e9936 vmscan: activate ... |
864 865 866 867 868 |
/* * Activate file-backed executable pages after first usage. */ if (vm_flags & VM_EXEC) return PAGEREF_ACTIVATE; |
645747462 vmscan: detect ma... |
869 870 |
return PAGEREF_KEEP; } |
dfc8d636c vmscan: factor ou... |
871 872 |
/* Reclaim if clean, defer dirty pages to writeback */ |
2e30244a7 vmscan,tmpfs: tre... |
873 |
if (referenced_page && !PageSwapBacked(page)) |
645747462 vmscan: detect ma... |
874 875 876 |
return PAGEREF_RECLAIM_CLEAN; return PAGEREF_RECLAIM; |
dfc8d636c vmscan: factor ou... |
877 |
} |
e2be15f6c mm: vmscan: stall... |
878 879 880 881 |
/* Check if a page is dirty or under writeback */ static void page_check_dirty_writeback(struct page *page, bool *dirty, bool *writeback) { |
b45972265 mm: vmscan: take ... |
882 |
struct address_space *mapping; |
e2be15f6c mm: vmscan: stall... |
883 884 885 886 |
/* * Anonymous pages are not handled by flushers and must be written * from reclaim context. Do not stall reclaim based on them */ |
802a3a92a mm: reclaim MADV_... |
887 888 |
if (!page_is_file_cache(page) || (PageAnon(page) && !PageSwapBacked(page))) { |
e2be15f6c mm: vmscan: stall... |
889 890 891 892 893 894 895 896 |
*dirty = false; *writeback = false; return; } /* By default assume that the page flags are accurate */ *dirty = PageDirty(page); *writeback = PageWriteback(page); |
b45972265 mm: vmscan: take ... |
897 898 899 900 901 902 903 904 |
/* Verify dirty/writeback state if the filesystem supports it */ if (!page_has_private(page)) return; mapping = page_mapping(page); if (mapping && mapping->a_ops->is_dirty_writeback) mapping->a_ops->is_dirty_writeback(page, dirty, writeback); |
e2be15f6c mm: vmscan: stall... |
905 |
} |
3c710c1ad mm, vmscan: extra... |
906 907 908 909 910 911 |
struct reclaim_stat { unsigned nr_dirty; unsigned nr_unqueued_dirty; unsigned nr_congested; unsigned nr_writeback; unsigned nr_immediate; |
5bccd1665 mm, vmscan: enhan... |
912 913 914 |
unsigned nr_activate; unsigned nr_ref_keep; unsigned nr_unmap_fail; |
3c710c1ad mm, vmscan: extra... |
915 |
}; |
e286781d5 mm: speculative p... |
916 |
/* |
1742f19fa [PATCH] vmscan: r... |
917 |
* shrink_page_list() returns the number of reclaimed pages |
1da177e4c Linux-2.6.12-rc2 |
918 |
*/ |
1742f19fa [PATCH] vmscan: r... |
919 |
static unsigned long shrink_page_list(struct list_head *page_list, |
599d0c954 mm, vmscan: move ... |
920 |
struct pglist_data *pgdat, |
f84f6e2b0 mm: vmscan: do no... |
921 |
struct scan_control *sc, |
02c6de8d7 mm: cma: discard ... |
922 |
enum ttu_flags ttu_flags, |
3c710c1ad mm, vmscan: extra... |
923 |
struct reclaim_stat *stat, |
02c6de8d7 mm: cma: discard ... |
924 |
bool force_reclaim) |
1da177e4c Linux-2.6.12-rc2 |
925 926 |
{ LIST_HEAD(ret_pages); |
abe4c3b50 vmscan: set up pa... |
927 |
LIST_HEAD(free_pages); |
1da177e4c Linux-2.6.12-rc2 |
928 |
int pgactivate = 0; |
3c710c1ad mm, vmscan: extra... |
929 930 931 932 933 934 |
unsigned nr_unqueued_dirty = 0; unsigned nr_dirty = 0; unsigned nr_congested = 0; unsigned nr_reclaimed = 0; unsigned nr_writeback = 0; unsigned nr_immediate = 0; |
5bccd1665 mm, vmscan: enhan... |
935 936 |
unsigned nr_ref_keep = 0; unsigned nr_unmap_fail = 0; |
1da177e4c Linux-2.6.12-rc2 |
937 938 |
cond_resched(); |
1da177e4c Linux-2.6.12-rc2 |
939 940 941 942 |
while (!list_empty(page_list)) { struct address_space *mapping; struct page *page; int may_enter_fs; |
02c6de8d7 mm: cma: discard ... |
943 |
enum page_references references = PAGEREF_RECLAIM_CLEAN; |
e2be15f6c mm: vmscan: stall... |
944 |
bool dirty, writeback; |
1da177e4c Linux-2.6.12-rc2 |
945 946 947 948 949 |
cond_resched(); page = lru_to_page(page_list); list_del(&page->lru); |
529ae9aaa mm: rename page t... |
950 |
if (!trylock_page(page)) |
1da177e4c Linux-2.6.12-rc2 |
951 |
goto keep; |
309381fea mm: dump page whe... |
952 |
VM_BUG_ON_PAGE(PageActive(page), page); |
1da177e4c Linux-2.6.12-rc2 |
953 954 |
sc->nr_scanned++; |
80e434260 [PATCH] zone recl... |
955 |
|
39b5f29ac mm: remove vma ar... |
956 |
if (unlikely(!page_evictable(page))) |
ad6b67041 mm: remove SWAP_M... |
957 |
goto activate_locked; |
894bc3104 Unevictable LRU I... |
958 |
|
a6dc60f89 vmscan: rename sc... |
959 |
if (!sc->may_unmap && page_mapped(page)) |
80e434260 [PATCH] zone recl... |
960 |
goto keep_locked; |
1da177e4c Linux-2.6.12-rc2 |
961 |
/* Double the slab pressure for mapped and swapcache pages */ |
802a3a92a mm: reclaim MADV_... |
962 963 |
if ((page_mapped(page) || PageSwapCache(page)) && !(PageAnon(page) && !PageSwapBacked(page))) |
1da177e4c Linux-2.6.12-rc2 |
964 |
sc->nr_scanned++; |
c661b078f synchronous lumpy... |
965 966 |
may_enter_fs = (sc->gfp_mask & __GFP_FS) || (PageSwapCache(page) && (sc->gfp_mask & __GFP_IO)); |
283aba9f9 mm: vmscan: block... |
967 |
/* |
e2be15f6c mm: vmscan: stall... |
968 969 970 971 972 973 974 975 976 977 978 |
* The number of dirty pages determines if a zone is marked * reclaim_congested which affects wait_iff_congested. kswapd * will stall and start writing pages if the tail of the LRU * is all dirty unqueued pages. */ page_check_dirty_writeback(page, &dirty, &writeback); if (dirty || writeback) nr_dirty++; if (dirty && !writeback) nr_unqueued_dirty++; |
d04e8acd0 mm: vmscan: treat... |
979 980 981 982 983 984 |
/* * Treat this page as congested if the underlying BDI is or if * pages are cycling through the LRU so quickly that the * pages marked for immediate reclaim are making it to the * end of the LRU a second time. */ |
e2be15f6c mm: vmscan: stall... |
985 |
mapping = page_mapping(page); |
1da58ee2a mm: vmscan: count... |
986 |
if (((dirty || writeback) && mapping && |
703c27088 writeback: implem... |
987 |
inode_write_congested(mapping->host)) || |
d04e8acd0 mm: vmscan: treat... |
988 |
(writeback && PageReclaim(page))) |
e2be15f6c mm: vmscan: stall... |
989 990 991 |
nr_congested++; /* |
283aba9f9 mm: vmscan: block... |
992 993 994 995 996 997 998 999 1000 1001 |
* If a page at the tail of the LRU is under writeback, there * are three cases to consider. * * 1) If reclaim is encountering an excessive number of pages * under writeback and this page is both under writeback and * PageReclaim then it indicates that pages are being queued * for IO but are being recycled through the LRU before the * IO can complete. Waiting on the page itself risks an * indefinite stall if it is impossible to writeback the * page due to IO error or disconnected storage so instead |
b1a6f21e3 mm: vmscan: stall... |
1002 1003 |
* note that the LRU is being scanned too quickly and the * caller can stall after page list has been processed. |
283aba9f9 mm: vmscan: block... |
1004 |
* |
97c9341f7 mm: vmscan: disab... |
1005 |
* 2) Global or new memcg reclaim encounters a page that is |
ecf5fc6e9 mm, vmscan: Do no... |
1006 1007 1008 |
* not marked for immediate reclaim, or the caller does not * have __GFP_FS (or __GFP_IO if it's simply going to swap, * not to fs). In this case mark the page for immediate |
97c9341f7 mm: vmscan: disab... |
1009 |
* reclaim and continue scanning. |
283aba9f9 mm: vmscan: block... |
1010 |
* |
ecf5fc6e9 mm, vmscan: Do no... |
1011 1012 |
* Require may_enter_fs because we would wait on fs, which * may not have submitted IO yet. And the loop driver might |
283aba9f9 mm: vmscan: block... |
1013 1014 1015 1016 1017 |
* enter reclaim, and deadlock if it waits on a page for * which it is needed to do the write (loop masks off * __GFP_IO|__GFP_FS for this reason); but more thought * would probably show more reasons. * |
7fadc8202 mm, vmscan: unloc... |
1018 |
* 3) Legacy memcg encounters a page that is already marked |
283aba9f9 mm: vmscan: block... |
1019 1020 1021 1022 |
* PageReclaim. memcg does not have any dirty pages * throttling so we could easily OOM just because too many * pages are in writeback and there is nothing else to * reclaim. Wait for the writeback to complete. |
c55e8d035 mm: vmscan: move ... |
1023 1024 1025 1026 1027 1028 1029 1030 1031 |
* * In cases 1) and 2) we activate the pages to get them out of * the way while we continue scanning for clean pages on the * inactive list and refilling from the active list. The * observation here is that waiting for disk writes is more * expensive than potentially causing reloads down the line. * Since they're marked for immediate reclaim, they won't put * memory pressure on the cache working set any longer than it * takes to write them to disk. |
283aba9f9 mm: vmscan: block... |
1032 |
*/ |
c661b078f synchronous lumpy... |
1033 |
if (PageWriteback(page)) { |
283aba9f9 mm: vmscan: block... |
1034 1035 1036 |
/* Case 1 above */ if (current_is_kswapd() && PageReclaim(page) && |
599d0c954 mm, vmscan: move ... |
1037 |
test_bit(PGDAT_WRITEBACK, &pgdat->flags)) { |
b1a6f21e3 mm: vmscan: stall... |
1038 |
nr_immediate++; |
c55e8d035 mm: vmscan: move ... |
1039 |
goto activate_locked; |
283aba9f9 mm: vmscan: block... |
1040 1041 |
/* Case 2 above */ |
97c9341f7 mm: vmscan: disab... |
1042 |
} else if (sane_reclaim(sc) || |
ecf5fc6e9 mm, vmscan: Do no... |
1043 |
!PageReclaim(page) || !may_enter_fs) { |
c3b94f44f memcg: further pr... |
1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 |
/* * This is slightly racy - end_page_writeback() * might have just cleared PageReclaim, then * setting PageReclaim here end up interpreted * as PageReadahead - but that does not matter * enough to care. What we do want is for this * page to have PageReclaim set next time memcg * reclaim reaches the tests above, so it will * then wait_on_page_writeback() to avoid OOM; * and it's also appropriate in global reclaim. */ SetPageReclaim(page); |
e62e384e9 memcg: prevent OO... |
1056 |
nr_writeback++; |
c55e8d035 mm: vmscan: move ... |
1057 |
goto activate_locked; |
283aba9f9 mm: vmscan: block... |
1058 1059 1060 |
/* Case 3 above */ } else { |
7fadc8202 mm, vmscan: unloc... |
1061 |
unlock_page(page); |
283aba9f9 mm: vmscan: block... |
1062 |
wait_on_page_writeback(page); |
7fadc8202 mm, vmscan: unloc... |
1063 1064 1065 |
/* then go back and try same page again */ list_add_tail(&page->lru, page_list); continue; |
e62e384e9 memcg: prevent OO... |
1066 |
} |
c661b078f synchronous lumpy... |
1067 |
} |
1da177e4c Linux-2.6.12-rc2 |
1068 |
|
02c6de8d7 mm: cma: discard ... |
1069 1070 |
if (!force_reclaim) references = page_check_references(page, sc); |
dfc8d636c vmscan: factor ou... |
1071 1072 |
switch (references) { case PAGEREF_ACTIVATE: |
1da177e4c Linux-2.6.12-rc2 |
1073 |
goto activate_locked; |
645747462 vmscan: detect ma... |
1074 |
case PAGEREF_KEEP: |
5bccd1665 mm, vmscan: enhan... |
1075 |
nr_ref_keep++; |
645747462 vmscan: detect ma... |
1076 |
goto keep_locked; |
dfc8d636c vmscan: factor ou... |
1077 1078 1079 1080 |
case PAGEREF_RECLAIM: case PAGEREF_RECLAIM_CLEAN: ; /* try to reclaim the page below */ } |
1da177e4c Linux-2.6.12-rc2 |
1081 |
|
1da177e4c Linux-2.6.12-rc2 |
1082 1083 1084 |
/* * Anonymous process memory has backing store? * Try to allocate it some swap space here. |
802a3a92a mm: reclaim MADV_... |
1085 |
* Lazyfree page could be freed directly |
1da177e4c Linux-2.6.12-rc2 |
1086 |
*/ |
bd4c82c22 mm, THP, swap: de... |
1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 |
if (PageAnon(page) && PageSwapBacked(page)) { if (!PageSwapCache(page)) { if (!(sc->gfp_mask & __GFP_IO)) goto keep_locked; if (PageTransHuge(page)) { /* cannot split THP, skip it */ if (!can_split_huge_page(page, NULL)) goto activate_locked; /* * Split pages without a PMD map right * away. Chances are some or all of the * tail pages can be freed without IO. */ if (!compound_mapcount(page) && split_huge_page_to_list(page, page_list)) goto activate_locked; } if (!add_to_swap(page)) { if (!PageTransHuge(page)) goto activate_locked; /* Fallback to swap normal pages */ if (split_huge_page_to_list(page, page_list)) goto activate_locked; |
fe490cc0f mm, THP, swap: ad... |
1112 1113 1114 |
#ifdef CONFIG_TRANSPARENT_HUGEPAGE count_vm_event(THP_SWPOUT_FALLBACK); #endif |
bd4c82c22 mm, THP, swap: de... |
1115 1116 1117 |
if (!add_to_swap(page)) goto activate_locked; } |
0f0746589 mm, THP, swap: mo... |
1118 |
|
bd4c82c22 mm, THP, swap: de... |
1119 |
may_enter_fs = 1; |
1da177e4c Linux-2.6.12-rc2 |
1120 |
|
bd4c82c22 mm, THP, swap: de... |
1121 1122 1123 |
/* Adding to swap updated mapping */ mapping = page_mapping(page); } |
7751b2da6 vmscan: split fil... |
1124 1125 1126 1127 |
} else if (unlikely(PageTransHuge(page))) { /* Split file THP */ if (split_huge_page_to_list(page, page_list)) goto keep_locked; |
e2be15f6c mm: vmscan: stall... |
1128 |
} |
1da177e4c Linux-2.6.12-rc2 |
1129 1130 1131 1132 1133 |
/* * The page is mapped into the page tables of one or more * processes. Try to unmap it here. */ |
802a3a92a mm: reclaim MADV_... |
1134 |
if (page_mapped(page)) { |
bd4c82c22 mm, THP, swap: de... |
1135 1136 1137 1138 1139 |
enum ttu_flags flags = ttu_flags | TTU_BATCH_FLUSH; if (unlikely(PageTransHuge(page))) flags |= TTU_SPLIT_HUGE_PMD; if (!try_to_unmap(page, flags)) { |
5bccd1665 mm, vmscan: enhan... |
1140 |
nr_unmap_fail++; |
1da177e4c Linux-2.6.12-rc2 |
1141 |
goto activate_locked; |
1da177e4c Linux-2.6.12-rc2 |
1142 1143 1144 1145 |
} } if (PageDirty(page)) { |
ee72886d8 mm: vmscan: do no... |
1146 |
/* |
4eda48235 mm: vmscan: only ... |
1147 1148 1149 1150 1151 1152 1153 1154 |
* Only kswapd can writeback filesystem pages * to avoid risk of stack overflow. But avoid * injecting inefficient single-page IO into * flusher writeback as much as possible: only * write pages when we've encountered many * dirty pages, and when we've already scanned * the rest of the LRU for clean pages and see * the same dirty pages again (PageReclaim). |
ee72886d8 mm: vmscan: do no... |
1155 |
*/ |
f84f6e2b0 mm: vmscan: do no... |
1156 |
if (page_is_file_cache(page) && |
4eda48235 mm: vmscan: only ... |
1157 1158 |
(!current_is_kswapd() || !PageReclaim(page) || !test_bit(PGDAT_DIRTY, &pgdat->flags))) { |
49ea7eb65 mm: vmscan: immed... |
1159 1160 1161 1162 1163 1164 |
/* * Immediately reclaim when written back. * Similar in principal to deactivate_page() * except we already have the page isolated * and know it's dirty */ |
c4a25635b mm: move vmscan w... |
1165 |
inc_node_page_state(page, NR_VMSCAN_IMMEDIATE); |
49ea7eb65 mm: vmscan: immed... |
1166 |
SetPageReclaim(page); |
c55e8d035 mm: vmscan: move ... |
1167 |
goto activate_locked; |
ee72886d8 mm: vmscan: do no... |
1168 |
} |
dfc8d636c vmscan: factor ou... |
1169 |
if (references == PAGEREF_RECLAIM_CLEAN) |
1da177e4c Linux-2.6.12-rc2 |
1170 |
goto keep_locked; |
4dd4b9202 revert "kswapd sh... |
1171 |
if (!may_enter_fs) |
1da177e4c Linux-2.6.12-rc2 |
1172 |
goto keep_locked; |
52a8363ea [PATCH] mm: impro... |
1173 |
if (!sc->may_writepage) |
1da177e4c Linux-2.6.12-rc2 |
1174 |
goto keep_locked; |
d950c9477 mm: defer flush o... |
1175 1176 1177 1178 1179 1180 |
/* * Page is dirty. Flush the TLB if a writable entry * potentially exists to avoid CPU writes after IO * starts and then write it out here. */ try_to_unmap_flush_dirty(); |
7d3579e8e vmscan: narrow th... |
1181 |
switch (pageout(page, mapping, sc)) { |
1da177e4c Linux-2.6.12-rc2 |
1182 1183 1184 1185 1186 |
case PAGE_KEEP: goto keep_locked; case PAGE_ACTIVATE: goto activate_locked; case PAGE_SUCCESS: |
7d3579e8e vmscan: narrow th... |
1187 |
if (PageWriteback(page)) |
41ac1999c mm: vmscan: do no... |
1188 |
goto keep; |
7d3579e8e vmscan: narrow th... |
1189 |
if (PageDirty(page)) |
1da177e4c Linux-2.6.12-rc2 |
1190 |
goto keep; |
7d3579e8e vmscan: narrow th... |
1191 |
|
1da177e4c Linux-2.6.12-rc2 |
1192 1193 1194 1195 |
/* * A synchronous write - probably a ramdisk. Go * ahead and try to reclaim the page. */ |
529ae9aaa mm: rename page t... |
1196 |
if (!trylock_page(page)) |
1da177e4c Linux-2.6.12-rc2 |
1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 |
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... |
1216 |
* will do this, as well as the blockdev mapping. |
1da177e4c Linux-2.6.12-rc2 |
1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 |
* 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... |
1227 |
if (page_has_private(page)) { |
1da177e4c Linux-2.6.12-rc2 |
1228 1229 |
if (!try_to_release_page(page, sc->gfp_mask)) goto activate_locked; |
e286781d5 mm: speculative p... |
1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 |
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 |
1246 |
} |
802a3a92a mm: reclaim MADV_... |
1247 1248 1249 1250 1251 1252 1253 1254 |
if (PageAnon(page) && !PageSwapBacked(page)) { /* follow __remove_mapping for reference */ if (!page_ref_freeze(page, 1)) goto keep_locked; if (PageDirty(page)) { page_ref_unfreeze(page, 1); goto keep_locked; } |
1da177e4c Linux-2.6.12-rc2 |
1255 |
|
802a3a92a mm: reclaim MADV_... |
1256 |
count_vm_event(PGLAZYFREED); |
2262185c5 mm: per-cgroup me... |
1257 |
count_memcg_page_event(page, PGLAZYFREED); |
802a3a92a mm: reclaim MADV_... |
1258 1259 |
} else if (!mapping || !__remove_mapping(mapping, page, true)) goto keep_locked; |
a978d6f52 mm: unlockless re... |
1260 1261 1262 1263 1264 1265 1266 |
/* * 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. */ |
48c935ad8 page-flags: defin... |
1267 |
__ClearPageLocked(page); |
e286781d5 mm: speculative p... |
1268 |
free_it: |
05ff51376 [PATCH] vmscan re... |
1269 |
nr_reclaimed++; |
abe4c3b50 vmscan: set up pa... |
1270 1271 1272 1273 1274 |
/* * Is there need to periodically free_page_list? It would * appear not as the counts should be low */ |
bd4c82c22 mm, THP, swap: de... |
1275 1276 1277 1278 1279 |
if (unlikely(PageTransHuge(page))) { mem_cgroup_uncharge(page); (*get_compound_page_dtor(page))(page); } else list_add(&page->lru, &free_pages); |
1da177e4c Linux-2.6.12-rc2 |
1280 1281 1282 |
continue; activate_locked: |
68a22394c vmscan: free swap... |
1283 |
/* Not a candidate for swapping, so reclaim swap space. */ |
ad6b67041 mm: remove SWAP_M... |
1284 1285 |
if (PageSwapCache(page) && (mem_cgroup_swap_full(page) || PageMlocked(page))) |
a2c43eed8 mm: try_to_free_s... |
1286 |
try_to_free_swap(page); |
309381fea mm: dump page whe... |
1287 |
VM_BUG_ON_PAGE(PageActive(page), page); |
ad6b67041 mm: remove SWAP_M... |
1288 1289 1290 |
if (!PageMlocked(page)) { SetPageActive(page); pgactivate++; |
2262185c5 mm: per-cgroup me... |
1291 |
count_memcg_page_event(page, PGACTIVATE); |
ad6b67041 mm: remove SWAP_M... |
1292 |
} |
1da177e4c Linux-2.6.12-rc2 |
1293 1294 1295 1296 |
keep_locked: unlock_page(page); keep: list_add(&page->lru, &ret_pages); |
309381fea mm: dump page whe... |
1297 |
VM_BUG_ON_PAGE(PageLRU(page) || PageUnevictable(page), page); |
1da177e4c Linux-2.6.12-rc2 |
1298 |
} |
abe4c3b50 vmscan: set up pa... |
1299 |
|
747db954c mm: memcontrol: u... |
1300 |
mem_cgroup_uncharge_list(&free_pages); |
72b252aed mm: send one IPI ... |
1301 |
try_to_unmap_flush(); |
b745bc85f mm: page_alloc: c... |
1302 |
free_hot_cold_page_list(&free_pages, true); |
abe4c3b50 vmscan: set up pa... |
1303 |
|
1da177e4c Linux-2.6.12-rc2 |
1304 |
list_splice(&ret_pages, page_list); |
f8891e5e1 [PATCH] Light wei... |
1305 |
count_vm_events(PGACTIVATE, pgactivate); |
0a31bc97c mm: memcontrol: r... |
1306 |
|
3c710c1ad mm, vmscan: extra... |
1307 1308 1309 1310 1311 1312 |
if (stat) { stat->nr_dirty = nr_dirty; stat->nr_congested = nr_congested; stat->nr_unqueued_dirty = nr_unqueued_dirty; stat->nr_writeback = nr_writeback; stat->nr_immediate = nr_immediate; |
5bccd1665 mm, vmscan: enhan... |
1313 1314 1315 |
stat->nr_activate = pgactivate; stat->nr_ref_keep = nr_ref_keep; stat->nr_unmap_fail = nr_unmap_fail; |
3c710c1ad mm, vmscan: extra... |
1316 |
} |
05ff51376 [PATCH] vmscan re... |
1317 |
return nr_reclaimed; |
1da177e4c Linux-2.6.12-rc2 |
1318 |
} |
02c6de8d7 mm: cma: discard ... |
1319 1320 1321 1322 1323 1324 1325 1326 |
unsigned long reclaim_clean_pages_from_list(struct zone *zone, struct list_head *page_list) { struct scan_control sc = { .gfp_mask = GFP_KERNEL, .priority = DEF_PRIORITY, .may_unmap = 1, }; |
3c710c1ad mm, vmscan: extra... |
1327 |
unsigned long ret; |
02c6de8d7 mm: cma: discard ... |
1328 1329 1330 1331 |
struct page *page, *next; LIST_HEAD(clean_pages); list_for_each_entry_safe(page, next, page_list, lru) { |
117aad1e9 mm: avoid reinser... |
1332 |
if (page_is_file_cache(page) && !PageDirty(page) && |
b1123ea6d mm: balloon: use ... |
1333 |
!__PageMovable(page)) { |
02c6de8d7 mm: cma: discard ... |
1334 1335 1336 1337 |
ClearPageActive(page); list_move(&page->lru, &clean_pages); } } |
599d0c954 mm, vmscan: move ... |
1338 |
ret = shrink_page_list(&clean_pages, zone->zone_pgdat, &sc, |
a128ca71f mm: delete unnece... |
1339 |
TTU_IGNORE_ACCESS, NULL, true); |
02c6de8d7 mm: cma: discard ... |
1340 |
list_splice(&clean_pages, page_list); |
599d0c954 mm, vmscan: move ... |
1341 |
mod_node_page_state(zone->zone_pgdat, NR_ISOLATED_FILE, -ret); |
02c6de8d7 mm: cma: discard ... |
1342 1343 |
return ret; } |
5ad333eb6 Lumpy Reclaim V4 |
1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 |
/* * Attempt to remove the specified page from its LRU. Only take this page * if it is of the appropriate PageActive status. Pages which are being * freed elsewhere are also ignored. * * page: page to consider * mode: one of the LRU isolation modes defined above * * returns 0 on success, -ve errno on failure. */ |
f3fd4a619 mm: remove lru ty... |
1354 |
int __isolate_lru_page(struct page *page, isolate_mode_t mode) |
5ad333eb6 Lumpy Reclaim V4 |
1355 1356 1357 1358 1359 1360 |
{ int ret = -EINVAL; /* Only take pages on the LRU. */ if (!PageLRU(page)) return ret; |
e46a28790 CMA: migrate mloc... |
1361 1362 |
/* Compaction should not handle unevictable pages but CMA can do so */ if (PageUnevictable(page) && !(mode & ISOLATE_UNEVICTABLE)) |
894bc3104 Unevictable LRU I... |
1363 |
return ret; |
5ad333eb6 Lumpy Reclaim V4 |
1364 |
ret = -EBUSY; |
08e552c69 memcg: synchroniz... |
1365 |
|
c82449352 mm: compaction: m... |
1366 1367 1368 1369 1370 |
/* * 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. * |
c82449352 mm: compaction: m... |
1371 1372 1373 |
* ISOLATE_ASYNC_MIGRATE is used to indicate that it only wants to pages * that it is possible to migrate without blocking */ |
1276ad68e mm: vmscan: scan ... |
1374 |
if (mode & ISOLATE_ASYNC_MIGRATE) { |
c82449352 mm: compaction: m... |
1375 1376 1377 1378 1379 1380 |
/* All the caller can do on PageWriteback is block */ if (PageWriteback(page)) return ret; if (PageDirty(page)) { struct address_space *mapping; |
1f9c87e25 mm: pin address_s... |
1381 |
bool migrate_dirty; |
c82449352 mm: compaction: m... |
1382 |
|
c82449352 mm: compaction: m... |
1383 1384 1385 |
/* * Only pages without mappings or that have a * ->migratepage callback are possible to migrate |
1f9c87e25 mm: pin address_s... |
1386 1387 1388 1389 1390 |
* without blocking. However, we can be racing with * truncation so it's necessary to lock the page * to stabilise the mapping as truncation holds * the page lock until after the page is removed * from the page cache. |
c82449352 mm: compaction: m... |
1391 |
*/ |
1f9c87e25 mm: pin address_s... |
1392 1393 |
if (!trylock_page(page)) return ret; |
c82449352 mm: compaction: m... |
1394 |
mapping = page_mapping(page); |
b968dd765 mm: fix the NULL ... |
1395 |
migrate_dirty = !mapping || mapping->a_ops->migratepage; |
1f9c87e25 mm: pin address_s... |
1396 1397 |
unlock_page(page); if (!migrate_dirty) |
c82449352 mm: compaction: m... |
1398 1399 1400 |
return ret; } } |
39deaf858 mm: compaction: m... |
1401 |
|
f80c06736 mm: zone_reclaim:... |
1402 1403 |
if ((mode & ISOLATE_UNMAPPED) && page_mapped(page)) return ret; |
5ad333eb6 Lumpy Reclaim V4 |
1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 |
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; } |
7ee36a14f mm, vmscan: Updat... |
1416 1417 1418 1419 1420 1421 |
/* * Update LRU sizes after isolating pages. The LRU size updates must * be complete before mem_cgroup_update_lru_size due to a santity check. */ static __always_inline void update_lru_sizes(struct lruvec *lruvec, |
b4536f0c8 mm, memcg: fix th... |
1422 |
enum lru_list lru, unsigned long *nr_zone_taken) |
7ee36a14f mm, vmscan: Updat... |
1423 |
{ |
7ee36a14f mm, vmscan: Updat... |
1424 |
int zid; |
7ee36a14f mm, vmscan: Updat... |
1425 1426 1427 1428 1429 |
for (zid = 0; zid < MAX_NR_ZONES; zid++) { if (!nr_zone_taken[zid]) continue; __update_lru_size(lruvec, lru, zid, -nr_zone_taken[zid]); |
7ee36a14f mm, vmscan: Updat... |
1430 |
#ifdef CONFIG_MEMCG |
b4536f0c8 mm, memcg: fix th... |
1431 |
mem_cgroup_update_lru_size(lruvec, lru, zid, -nr_zone_taken[zid]); |
7ee36a14f mm, vmscan: Updat... |
1432 |
#endif |
b4536f0c8 mm, memcg: fix th... |
1433 |
} |
7ee36a14f mm, vmscan: Updat... |
1434 |
} |
49d2e9cc4 [PATCH] Swap Migr... |
1435 |
/* |
a52633d8e mm, vmscan: move ... |
1436 |
* zone_lru_lock is heavily contended. Some of the functions that |
1da177e4c Linux-2.6.12-rc2 |
1437 1438 1439 1440 1441 1442 1443 1444 |
* 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. * |
791b48b64 mm: vmscan: scan ... |
1445 |
* @nr_to_scan: The number of eligible pages to look through on the list. |
5dc35979e mm/vmscan: push l... |
1446 |
* @lruvec: The LRU vector to pull pages from. |
1da177e4c Linux-2.6.12-rc2 |
1447 |
* @dst: The temp list to put pages on to. |
f626012db mm: remove isolat... |
1448 |
* @nr_scanned: The number of pages that were scanned. |
fe2c2a106 vmscan: reclaim a... |
1449 |
* @sc: The scan_control struct for this reclaim session |
5ad333eb6 Lumpy Reclaim V4 |
1450 |
* @mode: One of the LRU isolation modes |
3cb994517 mm: push lru inde... |
1451 |
* @lru: LRU list id for isolating |
1da177e4c Linux-2.6.12-rc2 |
1452 1453 1454 |
* * returns how many pages were moved onto *@dst. */ |
69e05944a [PATCH] vmscan: u... |
1455 |
static unsigned long isolate_lru_pages(unsigned long nr_to_scan, |
5dc35979e mm/vmscan: push l... |
1456 |
struct lruvec *lruvec, struct list_head *dst, |
fe2c2a106 vmscan: reclaim a... |
1457 |
unsigned long *nr_scanned, struct scan_control *sc, |
3cb994517 mm: push lru inde... |
1458 |
isolate_mode_t mode, enum lru_list lru) |
1da177e4c Linux-2.6.12-rc2 |
1459 |
{ |
75b00af77 mm: trivial clean... |
1460 |
struct list_head *src = &lruvec->lists[lru]; |
69e05944a [PATCH] vmscan: u... |
1461 |
unsigned long nr_taken = 0; |
599d0c954 mm, vmscan: move ... |
1462 |
unsigned long nr_zone_taken[MAX_NR_ZONES] = { 0 }; |
7cc30fcfd mm: vmstat: accou... |
1463 |
unsigned long nr_skipped[MAX_NR_ZONES] = { 0, }; |
3db65812d Revert "mm, vmsca... |
1464 |
unsigned long skipped = 0; |
791b48b64 mm: vmscan: scan ... |
1465 |
unsigned long scan, total_scan, nr_pages; |
b2e18757f mm, vmscan: begin... |
1466 |
LIST_HEAD(pages_skipped); |
1da177e4c Linux-2.6.12-rc2 |
1467 |
|
791b48b64 mm: vmscan: scan ... |
1468 1469 1470 1471 |
scan = 0; for (total_scan = 0; scan < nr_to_scan && nr_taken < nr_to_scan && !list_empty(src); total_scan++) { |
5ad333eb6 Lumpy Reclaim V4 |
1472 |
struct page *page; |
5ad333eb6 Lumpy Reclaim V4 |
1473 |
|
1da177e4c Linux-2.6.12-rc2 |
1474 1475 |
page = lru_to_page(src); prefetchw_prev_lru_page(page, src, flags); |
309381fea mm: dump page whe... |
1476 |
VM_BUG_ON_PAGE(!PageLRU(page), page); |
8d438f96d [PATCH] mm: PageL... |
1477 |
|
b2e18757f mm, vmscan: begin... |
1478 1479 |
if (page_zonenum(page) > sc->reclaim_idx) { list_move(&page->lru, &pages_skipped); |
7cc30fcfd mm: vmstat: accou... |
1480 |
nr_skipped[page_zonenum(page)]++; |
b2e18757f mm, vmscan: begin... |
1481 1482 |
continue; } |
791b48b64 mm: vmscan: scan ... |
1483 1484 1485 1486 1487 1488 1489 |
/* * Do not count skipped pages because that makes the function * return with no isolated pages if the LRU mostly contains * ineligible pages. This causes the VM to not reclaim any * pages, triggering a premature OOM. */ scan++; |
f3fd4a619 mm: remove lru ty... |
1490 |
switch (__isolate_lru_page(page, mode)) { |
5ad333eb6 Lumpy Reclaim V4 |
1491 |
case 0: |
599d0c954 mm, vmscan: move ... |
1492 1493 1494 |
nr_pages = hpage_nr_pages(page); nr_taken += nr_pages; nr_zone_taken[page_zonenum(page)] += nr_pages; |
5ad333eb6 Lumpy Reclaim V4 |
1495 |
list_move(&page->lru, dst); |
5ad333eb6 Lumpy Reclaim V4 |
1496 1497 1498 1499 1500 1501 |
break; case -EBUSY: /* else it is being freed elsewhere */ list_move(&page->lru, src); continue; |
46453a6e1 [PATCH] mm: never... |
1502 |
|
5ad333eb6 Lumpy Reclaim V4 |
1503 1504 1505 |
default: BUG(); } |
1da177e4c Linux-2.6.12-rc2 |
1506 |
} |
b2e18757f mm, vmscan: begin... |
1507 1508 1509 1510 1511 1512 1513 |
/* * Splice any skipped pages to the start of the LRU list. Note that * this disrupts the LRU order when reclaiming for lower zones but * we cannot splice to the tail. If we did then the SWAP_CLUSTER_MAX * scanning would soon rescan the same pages to skip and put the * system at risk of premature OOM. */ |
7cc30fcfd mm: vmstat: accou... |
1514 1515 |
if (!list_empty(&pages_skipped)) { int zid; |
3db65812d Revert "mm, vmsca... |
1516 |
list_splice(&pages_skipped, src); |
7cc30fcfd mm: vmstat: accou... |
1517 1518 1519 1520 1521 |
for (zid = 0; zid < MAX_NR_ZONES; zid++) { if (!nr_skipped[zid]) continue; __count_zid_vm_events(PGSCAN_SKIP, zid, nr_skipped[zid]); |
1265e3a69 mm, vmscan: show ... |
1522 |
skipped += nr_skipped[zid]; |
7cc30fcfd mm: vmstat: accou... |
1523 1524 |
} } |
791b48b64 mm: vmscan: scan ... |
1525 |
*nr_scanned = total_scan; |
1265e3a69 mm, vmscan: show ... |
1526 |
trace_mm_vmscan_lru_isolate(sc->reclaim_idx, sc->order, nr_to_scan, |
791b48b64 mm: vmscan: scan ... |
1527 |
total_scan, skipped, nr_taken, mode, lru); |
b4536f0c8 mm, memcg: fix th... |
1528 |
update_lru_sizes(lruvec, lru, nr_zone_taken); |
1da177e4c Linux-2.6.12-rc2 |
1529 1530 |
return nr_taken; } |
62695a84e vmscan: move isol... |
1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 |
/** * 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... |
1542 1543 1544 |
* 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... |
1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 |
* * The vmstat statistic corresponding to the list on which the page was * found will be decremented. * * Restrictions: * (1) Must be called with an elevated refcount on the page. This is a * fundamentnal difference from isolate_lru_pages (which is called * without a stable reference). * (2) the lru_lock must not be held. * (3) interrupts must be enabled. */ int isolate_lru_page(struct page *page) { int ret = -EBUSY; |
309381fea mm: dump page whe... |
1559 |
VM_BUG_ON_PAGE(!page_count(page), page); |
cf2a82ee4 mm: downgrade VM_... |
1560 |
WARN_RATELIMIT(PageTail(page), "trying to isolate tail page"); |
0c917313a mm: strictly requ... |
1561 |
|
62695a84e vmscan: move isol... |
1562 1563 |
if (PageLRU(page)) { struct zone *zone = page_zone(page); |
fa9add641 mm/memcg: apply a... |
1564 |
struct lruvec *lruvec; |
62695a84e vmscan: move isol... |
1565 |
|
a52633d8e mm, vmscan: move ... |
1566 |
spin_lock_irq(zone_lru_lock(zone)); |
599d0c954 mm, vmscan: move ... |
1567 |
lruvec = mem_cgroup_page_lruvec(page, zone->zone_pgdat); |
0c917313a mm: strictly requ... |
1568 |
if (PageLRU(page)) { |
894bc3104 Unevictable LRU I... |
1569 |
int lru = page_lru(page); |
0c917313a mm: strictly requ... |
1570 |
get_page(page); |
62695a84e vmscan: move isol... |
1571 |
ClearPageLRU(page); |
fa9add641 mm/memcg: apply a... |
1572 1573 |
del_page_from_lru_list(page, lruvec, lru); ret = 0; |
62695a84e vmscan: move isol... |
1574 |
} |
a52633d8e mm, vmscan: move ... |
1575 |
spin_unlock_irq(zone_lru_lock(zone)); |
62695a84e vmscan: move isol... |
1576 1577 1578 |
} return ret; } |
5ad333eb6 Lumpy Reclaim V4 |
1579 |
/* |
d37dd5dcb vmscan: comment t... |
1580 1581 1582 1583 1584 |
* A direct reclaimer may isolate SWAP_CLUSTER_MAX pages from the LRU list and * then get resheduled. When there are massive number of tasks doing page * allocation, such sleeping direct reclaimers may keep piling up on each CPU, * the LRU list will go small and be scanned faster than necessary, leading to * unnecessary swapping, thrashing and OOM. |
35cd78156 vmscan: throttle ... |
1585 |
*/ |
599d0c954 mm, vmscan: move ... |
1586 |
static int too_many_isolated(struct pglist_data *pgdat, int file, |
35cd78156 vmscan: throttle ... |
1587 1588 1589 1590 1591 1592 |
struct scan_control *sc) { unsigned long inactive, isolated; if (current_is_kswapd()) return 0; |
97c9341f7 mm: vmscan: disab... |
1593 |
if (!sane_reclaim(sc)) |
35cd78156 vmscan: throttle ... |
1594 1595 1596 |
return 0; if (file) { |
599d0c954 mm, vmscan: move ... |
1597 1598 |
inactive = node_page_state(pgdat, NR_INACTIVE_FILE); isolated = node_page_state(pgdat, NR_ISOLATED_FILE); |
35cd78156 vmscan: throttle ... |
1599 |
} else { |
599d0c954 mm, vmscan: move ... |
1600 1601 |
inactive = node_page_state(pgdat, NR_INACTIVE_ANON); isolated = node_page_state(pgdat, NR_ISOLATED_ANON); |
35cd78156 vmscan: throttle ... |
1602 |
} |
3cf23841b mm/vmscan.c: avoi... |
1603 1604 1605 1606 1607 |
/* * GFP_NOIO/GFP_NOFS callers are allowed to isolate more pages, so they * won't get blocked by normal direct-reclaimers, forming a circular * deadlock. */ |
d0164adc8 mm, page_alloc: d... |
1608 |
if ((sc->gfp_mask & (__GFP_IO | __GFP_FS)) == (__GFP_IO | __GFP_FS)) |
3cf23841b mm/vmscan.c: avoi... |
1609 |
inactive >>= 3; |
35cd78156 vmscan: throttle ... |
1610 1611 |
return isolated > inactive; } |
666356297 vmscan: set up pa... |
1612 |
static noinline_for_stack void |
75b00af77 mm: trivial clean... |
1613 |
putback_inactive_pages(struct lruvec *lruvec, struct list_head *page_list) |
666356297 vmscan: set up pa... |
1614 |
{ |
27ac81d85 mm/vmscan: push l... |
1615 |
struct zone_reclaim_stat *reclaim_stat = &lruvec->reclaim_stat; |
599d0c954 mm, vmscan: move ... |
1616 |
struct pglist_data *pgdat = lruvec_pgdat(lruvec); |
3f79768f2 mm: rearrange put... |
1617 |
LIST_HEAD(pages_to_free); |
666356297 vmscan: set up pa... |
1618 |
|
666356297 vmscan: set up pa... |
1619 1620 1621 |
/* * Put back any unfreeable pages. */ |
666356297 vmscan: set up pa... |
1622 |
while (!list_empty(page_list)) { |
3f79768f2 mm: rearrange put... |
1623 |
struct page *page = lru_to_page(page_list); |
666356297 vmscan: set up pa... |
1624 |
int lru; |
3f79768f2 mm: rearrange put... |
1625 |
|
309381fea mm: dump page whe... |
1626 |
VM_BUG_ON_PAGE(PageLRU(page), page); |
666356297 vmscan: set up pa... |
1627 |
list_del(&page->lru); |
39b5f29ac mm: remove vma ar... |
1628 |
if (unlikely(!page_evictable(page))) { |
599d0c954 mm, vmscan: move ... |
1629 |
spin_unlock_irq(&pgdat->lru_lock); |
666356297 vmscan: set up pa... |
1630 |
putback_lru_page(page); |
599d0c954 mm, vmscan: move ... |
1631 |
spin_lock_irq(&pgdat->lru_lock); |
666356297 vmscan: set up pa... |
1632 1633 |
continue; } |
fa9add641 mm/memcg: apply a... |
1634 |
|
599d0c954 mm, vmscan: move ... |
1635 |
lruvec = mem_cgroup_page_lruvec(page, pgdat); |
fa9add641 mm/memcg: apply a... |
1636 |
|
7a608572a Revert "mm: batch... |
1637 |
SetPageLRU(page); |
666356297 vmscan: set up pa... |
1638 |
lru = page_lru(page); |
fa9add641 mm/memcg: apply a... |
1639 |
add_page_to_lru_list(page, lruvec, lru); |
666356297 vmscan: set up pa... |
1640 1641 |
if (is_active_lru(lru)) { int file = is_file_lru(lru); |
9992af102 thp: scale nr_rot... |
1642 1643 |
int numpages = hpage_nr_pages(page); reclaim_stat->recent_rotated[file] += numpages; |
666356297 vmscan: set up pa... |
1644 |
} |
2bcf88796 mm: take pagevecs... |
1645 1646 1647 |
if (put_page_testzero(page)) { __ClearPageLRU(page); __ClearPageActive(page); |
fa9add641 mm/memcg: apply a... |
1648 |
del_page_from_lru_list(page, lruvec, lru); |
2bcf88796 mm: take pagevecs... |
1649 1650 |
if (unlikely(PageCompound(page))) { |
599d0c954 mm, vmscan: move ... |
1651 |
spin_unlock_irq(&pgdat->lru_lock); |
747db954c mm: memcontrol: u... |
1652 |
mem_cgroup_uncharge(page); |
2bcf88796 mm: take pagevecs... |
1653 |
(*get_compound_page_dtor(page))(page); |
599d0c954 mm, vmscan: move ... |
1654 |
spin_lock_irq(&pgdat->lru_lock); |
2bcf88796 mm: take pagevecs... |
1655 1656 |
} else list_add(&page->lru, &pages_to_free); |
666356297 vmscan: set up pa... |
1657 1658 |
} } |
666356297 vmscan: set up pa... |
1659 |
|
3f79768f2 mm: rearrange put... |
1660 1661 1662 1663 |
/* * 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... |
1664 1665 1666 |
} /* |
399ba0b95 mm/vmscan.c: avoi... |
1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 |
* If a kernel thread (such as nfsd for loop-back mounts) services * a backing device by writing to the page cache it sets PF_LESS_THROTTLE. * In that case we should only throttle if the backing device it is * writing to is congested. In other cases it is safe to throttle. */ static int current_may_throttle(void) { return !(current->flags & PF_LESS_THROTTLE) || current->backing_dev_info == NULL || bdi_write_congested(current->backing_dev_info); } /* |
b2e18757f mm, vmscan: begin... |
1680 |
* shrink_inactive_list() is a helper for shrink_node(). It returns the number |
1742f19fa [PATCH] vmscan: r... |
1681 |
* of reclaimed pages |
1da177e4c Linux-2.6.12-rc2 |
1682 |
*/ |
666356297 vmscan: set up pa... |
1683 |
static noinline_for_stack unsigned long |
1a93be0e7 mm/vmscan: push l... |
1684 |
shrink_inactive_list(unsigned long nr_to_scan, struct lruvec *lruvec, |
9e3b2f8cd mm/vmscan: store ... |
1685 |
struct scan_control *sc, enum lru_list lru) |
1da177e4c Linux-2.6.12-rc2 |
1686 1687 |
{ LIST_HEAD(page_list); |
e247dbce5 vmscan: simplify ... |
1688 |
unsigned long nr_scanned; |
05ff51376 [PATCH] vmscan re... |
1689 |
unsigned long nr_reclaimed = 0; |
e247dbce5 vmscan: simplify ... |
1690 |
unsigned long nr_taken; |
3c710c1ad mm, vmscan: extra... |
1691 |
struct reclaim_stat stat = {}; |
f3fd4a619 mm: remove lru ty... |
1692 |
isolate_mode_t isolate_mode = 0; |
3cb994517 mm: push lru inde... |
1693 |
int file = is_file_lru(lru); |
599d0c954 mm, vmscan: move ... |
1694 |
struct pglist_data *pgdat = lruvec_pgdat(lruvec); |
1a93be0e7 mm/vmscan: push l... |
1695 |
struct zone_reclaim_stat *reclaim_stat = &lruvec->reclaim_stat; |
db73ee0d4 mm, vmscan: do no... |
1696 |
bool stalled = false; |
78dc583d3 vmscan: low order... |
1697 |
|
599d0c954 mm, vmscan: move ... |
1698 |
while (unlikely(too_many_isolated(pgdat, file, sc))) { |
db73ee0d4 mm, vmscan: do no... |
1699 1700 1701 1702 1703 1704 |
if (stalled) return 0; /* wait a bit for the reclaimer. */ msleep(100); stalled = true; |
35cd78156 vmscan: throttle ... |
1705 1706 1707 1708 1709 |
/* We are about to die and free our memory. Return now. */ if (fatal_signal_pending(current)) return SWAP_CLUSTER_MAX; } |
1da177e4c Linux-2.6.12-rc2 |
1710 |
lru_add_drain(); |
f80c06736 mm: zone_reclaim:... |
1711 1712 |
if (!sc->may_unmap) |
613172891 mm/vmscan.c: clea... |
1713 |
isolate_mode |= ISOLATE_UNMAPPED; |
f80c06736 mm: zone_reclaim:... |
1714 |
|
599d0c954 mm, vmscan: move ... |
1715 |
spin_lock_irq(&pgdat->lru_lock); |
b35ea17b7 mm: shrink_inacti... |
1716 |
|
5dc35979e mm/vmscan: push l... |
1717 1718 |
nr_taken = isolate_lru_pages(nr_to_scan, lruvec, &page_list, &nr_scanned, sc, isolate_mode, lru); |
95d918fc0 mm/vmscan: remove... |
1719 |
|
599d0c954 mm, vmscan: move ... |
1720 |
__mod_node_page_state(pgdat, NR_ISOLATED_ANON + file, nr_taken); |
9d5e6a9f2 mm: update_lru_si... |
1721 |
reclaim_stat->recent_scanned[file] += nr_taken; |
95d918fc0 mm/vmscan: remove... |
1722 |
|
2262185c5 mm: per-cgroup me... |
1723 1724 |
if (current_is_kswapd()) { if (global_reclaim(sc)) |
599d0c954 mm, vmscan: move ... |
1725 |
__count_vm_events(PGSCAN_KSWAPD, nr_scanned); |
2262185c5 mm: per-cgroup me... |
1726 1727 1728 1729 |
count_memcg_events(lruvec_memcg(lruvec), PGSCAN_KSWAPD, nr_scanned); } else { if (global_reclaim(sc)) |
599d0c954 mm, vmscan: move ... |
1730 |
__count_vm_events(PGSCAN_DIRECT, nr_scanned); |
2262185c5 mm: per-cgroup me... |
1731 1732 |
count_memcg_events(lruvec_memcg(lruvec), PGSCAN_DIRECT, nr_scanned); |
e247dbce5 vmscan: simplify ... |
1733 |
} |
599d0c954 mm, vmscan: move ... |
1734 |
spin_unlock_irq(&pgdat->lru_lock); |
b35ea17b7 mm: shrink_inacti... |
1735 |
|
d563c0501 vmscan: handle is... |
1736 |
if (nr_taken == 0) |
666356297 vmscan: set up pa... |
1737 |
return 0; |
5ad333eb6 Lumpy Reclaim V4 |
1738 |
|
a128ca71f mm: delete unnece... |
1739 |
nr_reclaimed = shrink_page_list(&page_list, pgdat, sc, 0, |
3c710c1ad mm, vmscan: extra... |
1740 |
&stat, false); |
c661b078f synchronous lumpy... |
1741 |
|
599d0c954 mm, vmscan: move ... |
1742 |
spin_lock_irq(&pgdat->lru_lock); |
3f79768f2 mm: rearrange put... |
1743 |
|
2262185c5 mm: per-cgroup me... |
1744 1745 |
if (current_is_kswapd()) { if (global_reclaim(sc)) |
599d0c954 mm, vmscan: move ... |
1746 |
__count_vm_events(PGSTEAL_KSWAPD, nr_reclaimed); |
2262185c5 mm: per-cgroup me... |
1747 1748 1749 1750 |
count_memcg_events(lruvec_memcg(lruvec), PGSTEAL_KSWAPD, nr_reclaimed); } else { if (global_reclaim(sc)) |
599d0c954 mm, vmscan: move ... |
1751 |
__count_vm_events(PGSTEAL_DIRECT, nr_reclaimed); |
2262185c5 mm: per-cgroup me... |
1752 1753 |
count_memcg_events(lruvec_memcg(lruvec), PGSTEAL_DIRECT, nr_reclaimed); |
904249aa6 mm: fix up the vm... |
1754 |
} |
a74609faf [PATCH] mm: page_... |
1755 |
|
27ac81d85 mm/vmscan: push l... |
1756 |
putback_inactive_pages(lruvec, &page_list); |
3f79768f2 mm: rearrange put... |
1757 |
|
599d0c954 mm, vmscan: move ... |
1758 |
__mod_node_page_state(pgdat, NR_ISOLATED_ANON + file, -nr_taken); |
3f79768f2 mm: rearrange put... |
1759 |
|
599d0c954 mm, vmscan: move ... |
1760 |
spin_unlock_irq(&pgdat->lru_lock); |
3f79768f2 mm: rearrange put... |
1761 |
|
747db954c mm: memcontrol: u... |
1762 |
mem_cgroup_uncharge_list(&page_list); |
b745bc85f mm: page_alloc: c... |
1763 |
free_hot_cold_page_list(&page_list, true); |
e11da5b4f tracing, vmscan: ... |
1764 |
|
92df3a723 mm: vmscan: throt... |
1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 |
/* * If reclaim is isolating dirty pages under writeback, it implies * that the long-lived page allocation rate is exceeding the page * laundering rate. Either the global limits are not being effective * at throttling processes due to the page distribution throughout * zones or there is heavy usage of a slow backing device. The * only option is to throttle from reclaim context which is not ideal * as there is no guarantee the dirtying process is throttled in the * same way balance_dirty_pages() manages. * |
8e9502828 mm: vmscan: move ... |
1775 1776 1777 |
* Once a zone is flagged ZONE_WRITEBACK, kswapd will count the number * of pages under pages flagged for immediate reclaim and stall if any * are encountered in the nr_immediate check below. |
92df3a723 mm: vmscan: throt... |
1778 |
*/ |
3c710c1ad mm, vmscan: extra... |
1779 |
if (stat.nr_writeback && stat.nr_writeback == nr_taken) |
599d0c954 mm, vmscan: move ... |
1780 |
set_bit(PGDAT_WRITEBACK, &pgdat->flags); |
92df3a723 mm: vmscan: throt... |
1781 |
|
d43006d50 mm: vmscan: have ... |
1782 |
/* |
b8b815180 mm/vmscan: wake u... |
1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 |
* If dirty pages are scanned that are not queued for IO, it * implies that flushers are not doing their job. This can * happen when memory pressure pushes dirty pages to the end of * the LRU before the dirty limits are breached and the dirty * data has expired. It can also happen when the proportion of * dirty pages grows not through writes but through memory * pressure reclaiming all the clean cache. And in some cases, * the flushers simply cannot keep up with the allocation * rate. Nudge the flusher threads in case they are asleep. */ if (stat.nr_unqueued_dirty == nr_taken) wakeup_flusher_threads(0, WB_REASON_VMSCAN); /* |
97c9341f7 mm: vmscan: disab... |
1797 1798 |
* Legacy memcg will stall in page writeback so avoid forcibly * stalling here. |
d43006d50 mm: vmscan: have ... |
1799 |
*/ |
97c9341f7 mm: vmscan: disab... |
1800 |
if (sane_reclaim(sc)) { |
b1a6f21e3 mm: vmscan: stall... |
1801 |
/* |
8e9502828 mm: vmscan: move ... |
1802 1803 1804 |
* Tag a zone as congested if all the dirty pages scanned were * backed by a congested BDI and wait_iff_congested will stall. */ |
3c710c1ad mm, vmscan: extra... |
1805 |
if (stat.nr_dirty && stat.nr_dirty == stat.nr_congested) |
599d0c954 mm, vmscan: move ... |
1806 |
set_bit(PGDAT_CONGESTED, &pgdat->flags); |
8e9502828 mm: vmscan: move ... |
1807 |
|
b8b815180 mm/vmscan: wake u... |
1808 1809 |
/* Allow kswapd to start writing pages during reclaim. */ if (stat.nr_unqueued_dirty == nr_taken) |
599d0c954 mm, vmscan: move ... |
1810 |
set_bit(PGDAT_DIRTY, &pgdat->flags); |
b1a6f21e3 mm: vmscan: stall... |
1811 1812 |
/* |
b738d7646 Don't trigger con... |
1813 1814 1815 |
* If kswapd scans pages marked marked for immediate * reclaim and under writeback (nr_immediate), it implies * that pages are cycling through the LRU faster than |
b1a6f21e3 mm: vmscan: stall... |
1816 1817 |
* they are written so also forcibly stall. */ |
3c710c1ad mm, vmscan: extra... |
1818 |
if (stat.nr_immediate && current_may_throttle()) |
b1a6f21e3 mm: vmscan: stall... |
1819 |
congestion_wait(BLK_RW_ASYNC, HZ/10); |
e2be15f6c mm: vmscan: stall... |
1820 |
} |
d43006d50 mm: vmscan: have ... |
1821 |
|
8e9502828 mm: vmscan: move ... |
1822 1823 1824 1825 1826 |
/* * Stall direct reclaim for IO completions if underlying BDIs or zone * is congested. Allow kswapd to continue until it starts encountering * unqueued dirty pages or cycling through the LRU too quickly. */ |
399ba0b95 mm/vmscan.c: avoi... |
1827 1828 |
if (!sc->hibernation_mode && !current_is_kswapd() && current_may_throttle()) |
599d0c954 mm, vmscan: move ... |
1829 |
wait_iff_congested(pgdat, BLK_RW_ASYNC, HZ/10); |
8e9502828 mm: vmscan: move ... |
1830 |
|
599d0c954 mm, vmscan: move ... |
1831 1832 |
trace_mm_vmscan_lru_shrink_inactive(pgdat->node_id, nr_scanned, nr_reclaimed, |
5bccd1665 mm, vmscan: enhan... |
1833 1834 1835 1836 |
stat.nr_dirty, stat.nr_writeback, stat.nr_congested, stat.nr_immediate, stat.nr_activate, stat.nr_ref_keep, stat.nr_unmap_fail, |
ba5e95794 mm: change mm_vms... |
1837 |
sc->priority, file); |
05ff51376 [PATCH] vmscan re... |
1838 |
return nr_reclaimed; |
1da177e4c Linux-2.6.12-rc2 |
1839 |
} |
3bb1a852a [PATCH] vmscan: F... |
1840 |
/* |
1da177e4c Linux-2.6.12-rc2 |
1841 1842 1843 1844 1845 1846 |
* 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 |
a52633d8e mm, vmscan: move ... |
1847 |
* appropriate to hold zone_lru_lock across the whole operation. But if |
1da177e4c Linux-2.6.12-rc2 |
1848 |
* the pages are mapped, the processing is slow (page_referenced()) so we |
a52633d8e mm, vmscan: move ... |
1849 |
* should drop zone_lru_lock around each page. It's impossible to balance |
1da177e4c Linux-2.6.12-rc2 |
1850 1851 1852 1853 |
* 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. * |
0139aa7b7 mm: rename _count... |
1854 |
* The downside is that we have to touch page->_refcount against each page. |
1da177e4c Linux-2.6.12-rc2 |
1855 |
* But we had to alter page->flags anyway. |
9d998b4f1 mm, vmscan: add a... |
1856 1857 |
* * Returns the number of pages moved to the given lru. |
1da177e4c Linux-2.6.12-rc2 |
1858 |
*/ |
1cfb419b3 per-zone and recl... |
1859 |
|
9d998b4f1 mm, vmscan: add a... |
1860 |
static unsigned move_active_pages_to_lru(struct lruvec *lruvec, |
3eb4140f0 vmscan: merge dup... |
1861 |
struct list_head *list, |
2bcf88796 mm: take pagevecs... |
1862 |
struct list_head *pages_to_free, |
3eb4140f0 vmscan: merge dup... |
1863 1864 |
enum lru_list lru) { |
599d0c954 mm, vmscan: move ... |
1865 |
struct pglist_data *pgdat = lruvec_pgdat(lruvec); |
3eb4140f0 vmscan: merge dup... |
1866 |
struct page *page; |
fa9add641 mm/memcg: apply a... |
1867 |
int nr_pages; |
9d998b4f1 mm, vmscan: add a... |
1868 |
int nr_moved = 0; |
3eb4140f0 vmscan: merge dup... |
1869 |
|
3eb4140f0 vmscan: merge dup... |
1870 1871 |
while (!list_empty(list)) { page = lru_to_page(list); |
599d0c954 mm, vmscan: move ... |
1872 |
lruvec = mem_cgroup_page_lruvec(page, pgdat); |
3eb4140f0 vmscan: merge dup... |
1873 |
|
309381fea mm: dump page whe... |
1874 |
VM_BUG_ON_PAGE(PageLRU(page), page); |
3eb4140f0 vmscan: merge dup... |
1875 |
SetPageLRU(page); |
fa9add641 mm/memcg: apply a... |
1876 |
nr_pages = hpage_nr_pages(page); |
599d0c954 mm, vmscan: move ... |
1877 |
update_lru_size(lruvec, lru, page_zonenum(page), nr_pages); |
925b7673c mm: make per-memc... |
1878 |
list_move(&page->lru, &lruvec->lists[lru]); |
3eb4140f0 vmscan: merge dup... |
1879 |
|
2bcf88796 mm: take pagevecs... |
1880 1881 1882 |
if (put_page_testzero(page)) { __ClearPageLRU(page); __ClearPageActive(page); |
fa9add641 mm/memcg: apply a... |
1883 |
del_page_from_lru_list(page, lruvec, lru); |
2bcf88796 mm: take pagevecs... |
1884 1885 |
if (unlikely(PageCompound(page))) { |
599d0c954 mm, vmscan: move ... |
1886 |
spin_unlock_irq(&pgdat->lru_lock); |
747db954c mm: memcontrol: u... |
1887 |
mem_cgroup_uncharge(page); |
2bcf88796 mm: take pagevecs... |
1888 |
(*get_compound_page_dtor(page))(page); |
599d0c954 mm, vmscan: move ... |
1889 |
spin_lock_irq(&pgdat->lru_lock); |
2bcf88796 mm: take pagevecs... |
1890 1891 |
} else list_add(&page->lru, pages_to_free); |
9d998b4f1 mm, vmscan: add a... |
1892 1893 |
} else { nr_moved += nr_pages; |
3eb4140f0 vmscan: merge dup... |
1894 1895 |
} } |
9d5e6a9f2 mm: update_lru_si... |
1896 |
|
2262185c5 mm: per-cgroup me... |
1897 |
if (!is_active_lru(lru)) { |
f0958906c mm, vmscan: do no... |
1898 |
__count_vm_events(PGDEACTIVATE, nr_moved); |
2262185c5 mm: per-cgroup me... |
1899 1900 1901 |
count_memcg_events(lruvec_memcg(lruvec), PGDEACTIVATE, nr_moved); } |
9d998b4f1 mm, vmscan: add a... |
1902 1903 |
return nr_moved; |
3eb4140f0 vmscan: merge dup... |
1904 |
} |
1cfb419b3 per-zone and recl... |
1905 |
|
f626012db mm: remove isolat... |
1906 |
static void shrink_active_list(unsigned long nr_to_scan, |
1a93be0e7 mm/vmscan: push l... |
1907 |
struct lruvec *lruvec, |
f16015fbf mm: vmscan: disti... |
1908 |
struct scan_control *sc, |
9e3b2f8cd mm/vmscan: store ... |
1909 |
enum lru_list lru) |
1da177e4c Linux-2.6.12-rc2 |
1910 |
{ |
44c241f16 mm: rename pgmove... |
1911 |
unsigned long nr_taken; |
f626012db mm: remove isolat... |
1912 |
unsigned long nr_scanned; |
6fe6b7e35 vmscan: report vm... |
1913 |
unsigned long vm_flags; |
1da177e4c Linux-2.6.12-rc2 |
1914 |
LIST_HEAD(l_hold); /* The pages which were snipped off */ |
8cab4754d vmscan: make mapp... |
1915 |
LIST_HEAD(l_active); |
b69408e88 vmscan: Use an in... |
1916 |
LIST_HEAD(l_inactive); |
1da177e4c Linux-2.6.12-rc2 |
1917 |
struct page *page; |
1a93be0e7 mm/vmscan: push l... |
1918 |
struct zone_reclaim_stat *reclaim_stat = &lruvec->reclaim_stat; |
9d998b4f1 mm, vmscan: add a... |
1919 1920 |
unsigned nr_deactivate, nr_activate; unsigned nr_rotated = 0; |
f3fd4a619 mm: remove lru ty... |
1921 |
isolate_mode_t isolate_mode = 0; |
3cb994517 mm: push lru inde... |
1922 |
int file = is_file_lru(lru); |
599d0c954 mm, vmscan: move ... |
1923 |
struct pglist_data *pgdat = lruvec_pgdat(lruvec); |
1da177e4c Linux-2.6.12-rc2 |
1924 1925 |
lru_add_drain(); |
f80c06736 mm: zone_reclaim:... |
1926 1927 |
if (!sc->may_unmap) |
613172891 mm/vmscan.c: clea... |
1928 |
isolate_mode |= ISOLATE_UNMAPPED; |
f80c06736 mm: zone_reclaim:... |
1929 |
|
599d0c954 mm, vmscan: move ... |
1930 |
spin_lock_irq(&pgdat->lru_lock); |
925b7673c mm: make per-memc... |
1931 |
|
5dc35979e mm/vmscan: push l... |
1932 1933 |
nr_taken = isolate_lru_pages(nr_to_scan, lruvec, &l_hold, &nr_scanned, sc, isolate_mode, lru); |
89b5fae53 mm: vmscan: disti... |
1934 |
|
599d0c954 mm, vmscan: move ... |
1935 |
__mod_node_page_state(pgdat, NR_ISOLATED_ANON + file, nr_taken); |
b7c46d151 mm: drop unneeded... |
1936 |
reclaim_stat->recent_scanned[file] += nr_taken; |
1cfb419b3 per-zone and recl... |
1937 |
|
599d0c954 mm, vmscan: move ... |
1938 |
__count_vm_events(PGREFILL, nr_scanned); |
2262185c5 mm: per-cgroup me... |
1939 |
count_memcg_events(lruvec_memcg(lruvec), PGREFILL, nr_scanned); |
9d5e6a9f2 mm: update_lru_si... |
1940 |
|
599d0c954 mm, vmscan: move ... |
1941 |
spin_unlock_irq(&pgdat->lru_lock); |
1da177e4c Linux-2.6.12-rc2 |
1942 |
|
1da177e4c Linux-2.6.12-rc2 |
1943 1944 1945 1946 |
while (!list_empty(&l_hold)) { cond_resched(); page = lru_to_page(&l_hold); list_del(&page->lru); |
7e9cd4842 vmscan: fix pagec... |
1947 |
|
39b5f29ac mm: remove vma ar... |
1948 |
if (unlikely(!page_evictable(page))) { |
894bc3104 Unevictable LRU I... |
1949 1950 1951 |
putback_lru_page(page); continue; } |
cc715d99e mm: vmscan: forci... |
1952 1953 1954 1955 1956 1957 1958 |
if (unlikely(buffer_heads_over_limit)) { if (page_has_private(page) && trylock_page(page)) { if (page_has_private(page)) try_to_release_page(page, 0); unlock_page(page); } } |
c3ac9a8ad mm: memcg: count ... |
1959 1960 |
if (page_referenced(page, 0, sc->target_mem_cgroup, &vm_flags)) { |
9992af102 thp: scale nr_rot... |
1961 |
nr_rotated += hpage_nr_pages(page); |
8cab4754d vmscan: make mapp... |
1962 1963 1964 1965 1966 1967 1968 1969 1970 |
/* * 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_... |
1971 |
if ((vm_flags & VM_EXEC) && page_is_file_cache(page)) { |
8cab4754d vmscan: make mapp... |
1972 1973 1974 1975 |
list_add(&page->lru, &l_active); continue; } } |
7e9cd4842 vmscan: fix pagec... |
1976 |
|
5205e56ee vmscan: move Clea... |
1977 |
ClearPageActive(page); /* we are de-activating */ |
1da177e4c Linux-2.6.12-rc2 |
1978 1979 |
list_add(&page->lru, &l_inactive); } |
b555749aa vmscan: shrink_ac... |
1980 |
/* |
8cab4754d vmscan: make mapp... |
1981 |
* Move pages back to the lru list. |
b555749aa vmscan: shrink_ac... |
1982 |
*/ |
599d0c954 mm, vmscan: move ... |
1983 |
spin_lock_irq(&pgdat->lru_lock); |
4f98a2fee vmscan: split LRU... |
1984 |
/* |
8cab4754d vmscan: make mapp... |
1985 1986 1987 |
* Count referenced pages from currently used mappings as rotated, * even though only some of them are actually re-activated. This * helps balance scan pressure between file and anonymous pages in |
7c0db9e91 mm, vmscan: fix a... |
1988 |
* get_scan_count. |
7e9cd4842 vmscan: fix pagec... |
1989 |
*/ |
b7c46d151 mm: drop unneeded... |
1990 |
reclaim_stat->recent_rotated[file] += nr_rotated; |
556adecba vmscan: second ch... |
1991 |
|
9d998b4f1 mm, vmscan: add a... |
1992 1993 |
nr_activate = move_active_pages_to_lru(lruvec, &l_active, &l_hold, lru); nr_deactivate = move_active_pages_to_lru(lruvec, &l_inactive, &l_hold, lru - LRU_ACTIVE); |
599d0c954 mm, vmscan: move ... |
1994 1995 |
__mod_node_page_state(pgdat, NR_ISOLATED_ANON + file, -nr_taken); spin_unlock_irq(&pgdat->lru_lock); |
2bcf88796 mm: take pagevecs... |
1996 |
|
747db954c mm: memcontrol: u... |
1997 |
mem_cgroup_uncharge_list(&l_hold); |
b745bc85f mm: page_alloc: c... |
1998 |
free_hot_cold_page_list(&l_hold, true); |
9d998b4f1 mm, vmscan: add a... |
1999 2000 |
trace_mm_vmscan_lru_shrink_active(pgdat->node_id, nr_taken, nr_activate, nr_deactivate, nr_rotated, sc->priority, file); |
1da177e4c Linux-2.6.12-rc2 |
2001 |
} |
59dc76b0d mm: vmscan: reduc... |
2002 2003 2004 |
/* * The inactive anon list should be small enough that the VM never has * to do too much work. |
14797e236 memcg: add inacti... |
2005 |
* |
59dc76b0d mm: vmscan: reduc... |
2006 2007 2008 |
* The inactive file list should be small enough to leave most memory * to the established workingset on the scan-resistant active list, * but large enough to avoid thrashing the aggregate readahead window. |
56e49d218 vmscan: evict use... |
2009 |
* |
59dc76b0d mm: vmscan: reduc... |
2010 2011 |
* Both inactive lists should also be large enough that each inactive * page has a chance to be referenced again before it is reclaimed. |
56e49d218 vmscan: evict use... |
2012 |
* |
2a2e48854 mm: vmscan: fix I... |
2013 2014 |
* If that fails and refaulting is observed, the inactive list grows. * |
59dc76b0d mm: vmscan: reduc... |
2015 2016 2017 |
* The inactive_ratio is the target ratio of ACTIVE to INACTIVE pages * on this LRU, maintained by the pageout code. A zone->inactive_ratio * of 3 means 3:1 or 25% of the pages are kept on the inactive list. |
56e49d218 vmscan: evict use... |
2018 |
* |
59dc76b0d mm: vmscan: reduc... |
2019 2020 2021 2022 2023 2024 2025 2026 2027 2028 |
* total target max * memory ratio inactive * ------------------------------------- * 10MB 1 5MB * 100MB 1 50MB * 1GB 3 250MB * 10GB 10 0.9GB * 100GB 31 3GB * 1TB 101 10GB * 10TB 320 32GB |
56e49d218 vmscan: evict use... |
2029 |
*/ |
f8d1a3116 mm: consider whet... |
2030 |
static bool inactive_list_is_low(struct lruvec *lruvec, bool file, |
2a2e48854 mm: vmscan: fix I... |
2031 2032 |
struct mem_cgroup *memcg, struct scan_control *sc, bool actual_reclaim) |
56e49d218 vmscan: evict use... |
2033 |
{ |
fd5388037 mm, vmscan: clean... |
2034 |
enum lru_list active_lru = file * LRU_FILE + LRU_ACTIVE; |
2a2e48854 mm: vmscan: fix I... |
2035 2036 2037 2038 2039 |
struct pglist_data *pgdat = lruvec_pgdat(lruvec); enum lru_list inactive_lru = file * LRU_FILE; unsigned long inactive, active; unsigned long inactive_ratio; unsigned long refaults; |
59dc76b0d mm: vmscan: reduc... |
2040 |
unsigned long gb; |
e3790144c mm: refactor inac... |
2041 |
|
59dc76b0d mm: vmscan: reduc... |
2042 2043 2044 2045 2046 2047 |
/* * If we don't have swap space, anonymous page deactivation * is pointless. */ if (!file && !total_swap_pages) return false; |
56e49d218 vmscan: evict use... |
2048 |
|
fd5388037 mm, vmscan: clean... |
2049 2050 |
inactive = lruvec_lru_size(lruvec, inactive_lru, sc->reclaim_idx); active = lruvec_lru_size(lruvec, active_lru, sc->reclaim_idx); |
f8d1a3116 mm: consider whet... |
2051 |
|
2a2e48854 mm: vmscan: fix I... |
2052 |
if (memcg) |
ccda7f436 mm: memcontrol: u... |
2053 |
refaults = memcg_page_state(memcg, WORKINGSET_ACTIVATE); |
b39415b27 vmscan: do not ev... |
2054 |
else |
2a2e48854 mm: vmscan: fix I... |
2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 |
refaults = node_page_state(pgdat, WORKINGSET_ACTIVATE); /* * When refaults are being observed, it means a new workingset * is being established. Disable active list protection to get * rid of the stale workingset quickly. */ if (file && actual_reclaim && lruvec->refaults != refaults) { inactive_ratio = 0; } else { gb = (inactive + active) >> (30 - PAGE_SHIFT); if (gb) inactive_ratio = int_sqrt(10 * gb); else inactive_ratio = 1; } |
59dc76b0d mm: vmscan: reduc... |
2071 |
|
2a2e48854 mm: vmscan: fix I... |
2072 2073 2074 2075 2076 |
if (actual_reclaim) trace_mm_vmscan_inactive_list_is_low(pgdat->node_id, sc->reclaim_idx, lruvec_lru_size(lruvec, inactive_lru, MAX_NR_ZONES), inactive, lruvec_lru_size(lruvec, active_lru, MAX_NR_ZONES), active, inactive_ratio, file); |
fd5388037 mm, vmscan: clean... |
2077 |
|
59dc76b0d mm: vmscan: reduc... |
2078 |
return inactive * inactive_ratio < active; |
b39415b27 vmscan: do not ev... |
2079 |
} |
4f98a2fee vmscan: split LRU... |
2080 |
static unsigned long shrink_list(enum lru_list lru, unsigned long nr_to_scan, |
2a2e48854 mm: vmscan: fix I... |
2081 2082 |
struct lruvec *lruvec, struct mem_cgroup *memcg, struct scan_control *sc) |
b69408e88 vmscan: Use an in... |
2083 |
{ |
b39415b27 vmscan: do not ev... |
2084 |
if (is_active_lru(lru)) { |
2a2e48854 mm: vmscan: fix I... |
2085 2086 |
if (inactive_list_is_low(lruvec, is_file_lru(lru), memcg, sc, true)) |
1a93be0e7 mm/vmscan: push l... |
2087 |
shrink_active_list(nr_to_scan, lruvec, sc, lru); |
556adecba vmscan: second ch... |
2088 2089 |
return 0; } |
1a93be0e7 mm/vmscan: push l... |
2090 |
return shrink_inactive_list(nr_to_scan, lruvec, sc, lru); |
4f98a2fee vmscan: split LRU... |
2091 |
} |
9a2651140 mm: vmscan: clean... |
2092 2093 2094 2095 2096 2097 |
enum scan_balance { SCAN_EQUAL, SCAN_FRACT, SCAN_ANON, SCAN_FILE, }; |
4f98a2fee vmscan: split LRU... |
2098 2099 2100 2101 2102 2103 |
/* * Determine how aggressively the anon and file LRU lists should be * scanned. The relative value of each set of LRU lists is determined * by looking at the fraction of the pages scanned we did rotate back * onto the active list instead of evict. * |
be7bd59db mm: fix page recl... |
2104 2105 |
* nr[0] = anon inactive pages to scan; nr[1] = anon active pages to scan * nr[2] = file inactive pages to scan; nr[3] = file active pages to scan |
4f98a2fee vmscan: split LRU... |
2106 |
*/ |
333776785 mm: vmscan: pass ... |
2107 |
static void get_scan_count(struct lruvec *lruvec, struct mem_cgroup *memcg, |
6b4f7799c mm: vmscan: invok... |
2108 2109 |
struct scan_control *sc, unsigned long *nr, unsigned long *lru_pages) |
4f98a2fee vmscan: split LRU... |
2110 |
{ |
333776785 mm: vmscan: pass ... |
2111 |
int swappiness = mem_cgroup_swappiness(memcg); |
9a2651140 mm: vmscan: clean... |
2112 2113 2114 |
struct zone_reclaim_stat *reclaim_stat = &lruvec->reclaim_stat; u64 fraction[2]; u64 denominator = 0; /* gcc */ |
599d0c954 mm, vmscan: move ... |
2115 |
struct pglist_data *pgdat = lruvec_pgdat(lruvec); |
4f98a2fee vmscan: split LRU... |
2116 |
unsigned long anon_prio, file_prio; |
9a2651140 mm: vmscan: clean... |
2117 |
enum scan_balance scan_balance; |
0bf1457f0 mm: vmscan: do no... |
2118 |
unsigned long anon, file; |
4f98a2fee vmscan: split LRU... |
2119 |
unsigned long ap, fp; |
4111304da mm: enum lru_list... |
2120 |
enum lru_list lru; |
76a33fc38 vmscan: prevent g... |
2121 2122 |
/* If we have no swap space, do not bother scanning anon pages. */ |
d8b38438a mm: vmscan: do no... |
2123 |
if (!sc->may_swap || mem_cgroup_get_nr_swap_pages(memcg) <= 0) { |
9a2651140 mm: vmscan: clean... |
2124 |
scan_balance = SCAN_FILE; |
76a33fc38 vmscan: prevent g... |
2125 2126 |
goto out; } |
4f98a2fee vmscan: split LRU... |
2127 |
|
10316b313 mm: vmscan: clari... |
2128 2129 2130 2131 2132 2133 2134 |
/* * Global reclaim will swap to prevent OOM even with no * swappiness, but memcg users want to use this knob to * disable swapping for individual groups completely when * using the memory controller's swap limit feature would be * too expensive. */ |
02695175c mm: vmscan: move ... |
2135 |
if (!global_reclaim(sc) && !swappiness) { |
9a2651140 mm: vmscan: clean... |
2136 |
scan_balance = SCAN_FILE; |
10316b313 mm: vmscan: clari... |
2137 2138 2139 2140 2141 2142 2143 2144 |
goto out; } /* * Do not apply any pressure balancing cleverness when the * system is close to OOM, scan both anon and file equally * (unless the swappiness setting disagrees with swapping). */ |
02695175c mm: vmscan: move ... |
2145 |
if (!sc->priority && swappiness) { |
9a2651140 mm: vmscan: clean... |
2146 |
scan_balance = SCAN_EQUAL; |
10316b313 mm: vmscan: clari... |
2147 2148 |
goto out; } |
11d16c25b mm: vmscan: impro... |
2149 |
/* |
623762517 revert "mm: vmsca... |
2150 2151 2152 2153 2154 2155 2156 2157 2158 |
* Prevent the reclaimer from falling into the cache trap: as * cache pages start out inactive, every cache fault will tip * the scan balance towards the file LRU. And as the file LRU * shrinks, so does the window for rotation from references. * This means we have a runaway feedback loop where a tiny * thrashing file LRU becomes infinitely more attractive than * anon pages. Try to detect this based on file LRU size. */ if (global_reclaim(sc)) { |
599d0c954 mm, vmscan: move ... |
2159 2160 2161 2162 |
unsigned long pgdatfile; unsigned long pgdatfree; int z; unsigned long total_high_wmark = 0; |
2ab051e11 memcg, vmscan: Fi... |
2163 |
|
599d0c954 mm, vmscan: move ... |
2164 2165 2166 2167 2168 2169 |
pgdatfree = sum_zone_node_page_state(pgdat->node_id, NR_FREE_PAGES); pgdatfile = node_page_state(pgdat, NR_ACTIVE_FILE) + node_page_state(pgdat, NR_INACTIVE_FILE); for (z = 0; z < MAX_NR_ZONES; z++) { struct zone *zone = &pgdat->node_zones[z]; |
6aa303def mm, vmscan: only ... |
2170 |
if (!managed_zone(zone)) |
599d0c954 mm, vmscan: move ... |
2171 2172 2173 2174 |
continue; total_high_wmark += high_wmark_pages(zone); } |
623762517 revert "mm: vmsca... |
2175 |
|
599d0c954 mm, vmscan: move ... |
2176 |
if (unlikely(pgdatfile + pgdatfree <= total_high_wmark)) { |
062262267 mm, vmscan: avoid... |
2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 |
/* * Force SCAN_ANON if there are enough inactive * anonymous pages on the LRU in eligible zones. * Otherwise, the small LRU gets thrashed. */ if (!inactive_list_is_low(lruvec, false, memcg, sc, false) && lruvec_lru_size(lruvec, LRU_INACTIVE_ANON, sc->reclaim_idx) >> sc->priority) { scan_balance = SCAN_ANON; goto out; } |
623762517 revert "mm: vmsca... |
2188 2189 2190 2191 |
} } /* |
316bda0e6 vmscan: do not fo... |
2192 2193 2194 2195 2196 2197 2198 |
* If there is enough inactive page cache, i.e. if the size of the * inactive list is greater than that of the active list *and* the * inactive list actually has some pages to scan on this priority, we * do not reclaim anything from the anonymous working set right now. * Without the second condition we could end up never scanning an * lruvec even if it has plenty of old anonymous pages unless the * system is under heavy pressure. |
7c5bd705d mm: memcg: only e... |
2199 |
*/ |
2a2e48854 mm: vmscan: fix I... |
2200 |
if (!inactive_list_is_low(lruvec, true, memcg, sc, false) && |
71ab6cfe8 mm, vmscan: consi... |
2201 |
lruvec_lru_size(lruvec, LRU_INACTIVE_FILE, sc->reclaim_idx) >> sc->priority) { |
9a2651140 mm: vmscan: clean... |
2202 |
scan_balance = SCAN_FILE; |
7c5bd705d mm: memcg: only e... |
2203 2204 |
goto out; } |
9a2651140 mm: vmscan: clean... |
2205 |
scan_balance = SCAN_FRACT; |
7c5bd705d mm: memcg: only e... |
2206 |
/* |
58c37f6e0 vmscan: protect r... |
2207 2208 2209 |
* With swappiness at 100, anonymous and file have the same priority. * This scanning priority is essentially the inverse of IO cost. */ |
02695175c mm: vmscan: move ... |
2210 |
anon_prio = swappiness; |
75b00af77 mm: trivial clean... |
2211 |
file_prio = 200 - anon_prio; |
58c37f6e0 vmscan: protect r... |
2212 2213 |
/* |
4f98a2fee vmscan: split LRU... |
2214 2215 2216 2217 2218 2219 2220 2221 2222 2223 |
* OK, so we have swap space and a fair amount of page cache * pages. We use the recently rotated / recently scanned * ratios to determine how valuable each cache is. * * Because workloads change over time (and to avoid overflow) * we keep these statistics as a floating average, which ends * up weighing recent references more than old ones. * * anon in [0], file in [1] */ |
2ab051e11 memcg, vmscan: Fi... |
2224 |
|
fd5388037 mm, vmscan: clean... |
2225 2226 2227 2228 |
anon = lruvec_lru_size(lruvec, LRU_ACTIVE_ANON, MAX_NR_ZONES) + lruvec_lru_size(lruvec, LRU_INACTIVE_ANON, MAX_NR_ZONES); file = lruvec_lru_size(lruvec, LRU_ACTIVE_FILE, MAX_NR_ZONES) + lruvec_lru_size(lruvec, LRU_INACTIVE_FILE, MAX_NR_ZONES); |
2ab051e11 memcg, vmscan: Fi... |
2229 |
|
599d0c954 mm, vmscan: move ... |
2230 |
spin_lock_irq(&pgdat->lru_lock); |
6e9015716 mm: introduce zon... |
2231 |
if (unlikely(reclaim_stat->recent_scanned[0] > anon / 4)) { |
6e9015716 mm: introduce zon... |
2232 2233 |
reclaim_stat->recent_scanned[0] /= 2; reclaim_stat->recent_rotated[0] /= 2; |
4f98a2fee vmscan: split LRU... |
2234 |
} |
6e9015716 mm: introduce zon... |
2235 |
if (unlikely(reclaim_stat->recent_scanned[1] > file / 4)) { |
6e9015716 mm: introduce zon... |
2236 2237 |
reclaim_stat->recent_scanned[1] /= 2; reclaim_stat->recent_rotated[1] /= 2; |
4f98a2fee vmscan: split LRU... |
2238 2239 2240 |
} /* |
00d8089c5 vmscan: fix get_s... |
2241 2242 2243 |
* 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... |
2244 |
*/ |
fe35004fb mm: avoid swappin... |
2245 |
ap = anon_prio * (reclaim_stat->recent_scanned[0] + 1); |
6e9015716 mm: introduce zon... |
2246 |
ap /= reclaim_stat->recent_rotated[0] + 1; |
4f98a2fee vmscan: split LRU... |
2247 |
|
fe35004fb mm: avoid swappin... |
2248 |
fp = file_prio * (reclaim_stat->recent_scanned[1] + 1); |
6e9015716 mm: introduce zon... |
2249 |
fp /= reclaim_stat->recent_rotated[1] + 1; |
599d0c954 mm, vmscan: move ... |
2250 |
spin_unlock_irq(&pgdat->lru_lock); |
4f98a2fee vmscan: split LRU... |
2251 |
|
76a33fc38 vmscan: prevent g... |
2252 2253 2254 2255 |
fraction[0] = ap; fraction[1] = fp; denominator = ap + fp + 1; out: |
688035f72 mm: don't avoid h... |
2256 2257 2258 2259 2260 |
*lru_pages = 0; for_each_evictable_lru(lru) { int file = is_file_lru(lru); unsigned long size; unsigned long scan; |
6b4f7799c mm: vmscan: invok... |
2261 |
|
688035f72 mm: don't avoid h... |
2262 2263 2264 2265 2266 2267 2268 2269 |
size = lruvec_lru_size(lruvec, lru, sc->reclaim_idx); scan = size >> sc->priority; /* * If the cgroup's already been deleted, make sure to * scrape out the remaining cache. */ if (!scan && !mem_cgroup_online(memcg)) scan = min(size, SWAP_CLUSTER_MAX); |
6b4f7799c mm: vmscan: invok... |
2270 |
|
688035f72 mm: don't avoid h... |
2271 2272 2273 2274 2275 |
switch (scan_balance) { case SCAN_EQUAL: /* Scan lists relative to size */ break; case SCAN_FRACT: |
9a2651140 mm: vmscan: clean... |
2276 |
/* |
688035f72 mm: don't avoid h... |
2277 2278 |
* Scan types proportional to swappiness and * their relative recent reclaim efficiency. |
9a2651140 mm: vmscan: clean... |
2279 |
*/ |
688035f72 mm: don't avoid h... |
2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 |
scan = div64_u64(scan * fraction[file], denominator); break; case SCAN_FILE: case SCAN_ANON: /* Scan one type exclusively */ if ((scan_balance == SCAN_FILE) != file) { size = 0; scan = 0; } break; default: /* Look ma, no brain */ BUG(); |
9a2651140 mm: vmscan: clean... |
2294 |
} |
688035f72 mm: don't avoid h... |
2295 2296 2297 |
*lru_pages += size; nr[lru] = scan; |
76a33fc38 vmscan: prevent g... |
2298 |
} |
6e08a369e vmscan: cleanup t... |
2299 |
} |
4f98a2fee vmscan: split LRU... |
2300 |
|
9b4f98cda mm: vmscan: compa... |
2301 |
/* |
a9dd0a831 mm, vmscan: make ... |
2302 |
* This is a basic per-node page freer. Used by both kswapd and direct reclaim. |
9b4f98cda mm: vmscan: compa... |
2303 |
*/ |
a9dd0a831 mm, vmscan: make ... |
2304 |
static void shrink_node_memcg(struct pglist_data *pgdat, struct mem_cgroup *memcg, |
333776785 mm: vmscan: pass ... |
2305 |
struct scan_control *sc, unsigned long *lru_pages) |
9b4f98cda mm: vmscan: compa... |
2306 |
{ |
ef8f23279 mm, memcg: move m... |
2307 |
struct lruvec *lruvec = mem_cgroup_lruvec(pgdat, memcg); |
9b4f98cda mm: vmscan: compa... |
2308 |
unsigned long nr[NR_LRU_LISTS]; |
e82e0561d mm: vmscan: obey ... |
2309 |
unsigned long targets[NR_LRU_LISTS]; |
9b4f98cda mm: vmscan: compa... |
2310 2311 2312 2313 2314 |
unsigned long nr_to_scan; enum lru_list lru; unsigned long nr_reclaimed = 0; unsigned long nr_to_reclaim = sc->nr_to_reclaim; struct blk_plug plug; |
1a501907b mm: vmscan: use p... |
2315 |
bool scan_adjusted; |
9b4f98cda mm: vmscan: compa... |
2316 |
|
333776785 mm: vmscan: pass ... |
2317 |
get_scan_count(lruvec, memcg, sc, nr, lru_pages); |
9b4f98cda mm: vmscan: compa... |
2318 |
|
e82e0561d mm: vmscan: obey ... |
2319 2320 |
/* Record the original scan target for proportional adjustments later */ memcpy(targets, nr, sizeof(nr)); |
1a501907b mm: vmscan: use p... |
2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 |
/* * Global reclaiming within direct reclaim at DEF_PRIORITY is a normal * event that can occur when there is little memory pressure e.g. * multiple streaming readers/writers. Hence, we do not abort scanning * when the requested number of pages are reclaimed when scanning at * DEF_PRIORITY on the assumption that the fact we are direct * reclaiming implies that kswapd is not keeping up and it is best to * do a batch of work at once. For memcg reclaim one check is made to * abort proportional reclaim if either the file or anon lru has already * dropped to zero at the first pass. */ scan_adjusted = (global_reclaim(sc) && !current_is_kswapd() && sc->priority == DEF_PRIORITY); |
9b4f98cda mm: vmscan: compa... |
2334 2335 2336 |
blk_start_plug(&plug); while (nr[LRU_INACTIVE_ANON] || nr[LRU_ACTIVE_FILE] || nr[LRU_INACTIVE_FILE]) { |
e82e0561d mm: vmscan: obey ... |
2337 2338 |
unsigned long nr_anon, nr_file, percentage; unsigned long nr_scanned; |
9b4f98cda mm: vmscan: compa... |
2339 2340 2341 2342 2343 2344 |
for_each_evictable_lru(lru) { if (nr[lru]) { nr_to_scan = min(nr[lru], SWAP_CLUSTER_MAX); nr[lru] -= nr_to_scan; nr_reclaimed += shrink_list(lru, nr_to_scan, |
2a2e48854 mm: vmscan: fix I... |
2345 |
lruvec, memcg, sc); |
9b4f98cda mm: vmscan: compa... |
2346 2347 |
} } |
e82e0561d mm: vmscan: obey ... |
2348 |
|
bd041733c mm, vmscan: add c... |
2349 |
cond_resched(); |
e82e0561d mm: vmscan: obey ... |
2350 2351 |
if (nr_reclaimed < nr_to_reclaim || scan_adjusted) continue; |
9b4f98cda mm: vmscan: compa... |
2352 |
/* |
e82e0561d mm: vmscan: obey ... |
2353 |
* For kswapd and memcg, reclaim at least the number of pages |
1a501907b mm: vmscan: use p... |
2354 |
* requested. Ensure that the anon and file LRUs are scanned |
e82e0561d mm: vmscan: obey ... |
2355 2356 2357 2358 2359 2360 |
* proportionally what was requested by get_scan_count(). We * stop reclaiming one LRU and reduce the amount scanning * proportional to the original scan target. */ nr_file = nr[LRU_INACTIVE_FILE] + nr[LRU_ACTIVE_FILE]; nr_anon = nr[LRU_INACTIVE_ANON] + nr[LRU_ACTIVE_ANON]; |
1a501907b mm: vmscan: use p... |
2361 2362 2363 2364 2365 2366 2367 2368 |
/* * It's just vindictive to attack the larger once the smaller * has gone to zero. And given the way we stop scanning the * smaller below, this makes sure that we only make one nudge * towards proportionality once we've got nr_to_reclaim. */ if (!nr_file || !nr_anon) break; |
e82e0561d mm: vmscan: obey ... |
2369 2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 2396 2397 2398 2399 |
if (nr_file > nr_anon) { unsigned long scan_target = targets[LRU_INACTIVE_ANON] + targets[LRU_ACTIVE_ANON] + 1; lru = LRU_BASE; percentage = nr_anon * 100 / scan_target; } else { unsigned long scan_target = targets[LRU_INACTIVE_FILE] + targets[LRU_ACTIVE_FILE] + 1; lru = LRU_FILE; percentage = nr_file * 100 / scan_target; } /* Stop scanning the smaller of the LRU */ nr[lru] = 0; nr[lru + LRU_ACTIVE] = 0; /* * Recalculate the other LRU scan count based on its original * scan target and the percentage scanning already complete */ lru = (lru == LRU_FILE) ? LRU_BASE : LRU_FILE; nr_scanned = targets[lru] - nr[lru]; nr[lru] = targets[lru] * (100 - percentage) / 100; nr[lru] -= min(nr[lru], nr_scanned); lru += LRU_ACTIVE; nr_scanned = targets[lru] - nr[lru]; nr[lru] = targets[lru] * (100 - percentage) / 100; nr[lru] -= min(nr[lru], nr_scanned); scan_adjusted = true; |
9b4f98cda mm: vmscan: compa... |
2400 2401 2402 2403 2404 2405 2406 2407 |
} blk_finish_plug(&plug); sc->nr_reclaimed += nr_reclaimed; /* * Even if we did not try to evict anon pages at all, we want to * rebalance the anon lru active/inactive ratio. */ |
2a2e48854 mm: vmscan: fix I... |
2408 |
if (inactive_list_is_low(lruvec, false, memcg, sc, true)) |
9b4f98cda mm: vmscan: compa... |
2409 2410 |
shrink_active_list(SWAP_CLUSTER_MAX, lruvec, sc, LRU_ACTIVE_ANON); |
9b4f98cda mm: vmscan: compa... |
2411 |
} |
23b9da55c mm: vmscan: remov... |
2412 |
/* Use reclaim/compaction for costly allocs or under memory pressure */ |
9e3b2f8cd mm/vmscan: store ... |
2413 |
static bool in_reclaim_compaction(struct scan_control *sc) |
23b9da55c mm: vmscan: remov... |
2414 |
{ |
d84da3f9e mm: use IS_ENABLE... |
2415 |
if (IS_ENABLED(CONFIG_COMPACTION) && sc->order && |
23b9da55c mm: vmscan: remov... |
2416 |
(sc->order > PAGE_ALLOC_COSTLY_ORDER || |
9e3b2f8cd mm/vmscan: store ... |
2417 |
sc->priority < DEF_PRIORITY - 2)) |
23b9da55c mm: vmscan: remov... |
2418 2419 2420 2421 |
return true; return false; } |
1da177e4c Linux-2.6.12-rc2 |
2422 |
/* |
23b9da55c mm: vmscan: remov... |
2423 2424 2425 2426 2427 |
* Reclaim/compaction is used for high-order allocation requests. It reclaims * order-0 pages before compacting the zone. should_continue_reclaim() returns * true if more pages should be reclaimed such that when the page allocator * calls try_to_compact_zone() that it will have enough free pages to succeed. * It will give up earlier than that if there is difficulty reclaiming pages. |
3e7d34497 mm: vmscan: recla... |
2428 |
*/ |
a9dd0a831 mm, vmscan: make ... |
2429 |
static inline bool should_continue_reclaim(struct pglist_data *pgdat, |
3e7d34497 mm: vmscan: recla... |
2430 2431 2432 2433 2434 2435 |
unsigned long nr_reclaimed, unsigned long nr_scanned, struct scan_control *sc) { unsigned long pages_for_compaction; unsigned long inactive_lru_pages; |
a9dd0a831 mm, vmscan: make ... |
2436 |
int z; |
3e7d34497 mm: vmscan: recla... |
2437 2438 |
/* If not in reclaim/compaction mode, stop */ |
9e3b2f8cd mm/vmscan: store ... |
2439 |
if (!in_reclaim_compaction(sc)) |
3e7d34497 mm: vmscan: recla... |
2440 |
return false; |
2876592f2 mm: vmscan: stop ... |
2441 |
/* Consider stopping depending on scan and reclaim activity */ |
dcda9b047 mm, tree wide: re... |
2442 |
if (sc->gfp_mask & __GFP_RETRY_MAYFAIL) { |
2876592f2 mm: vmscan: stop ... |
2443 |
/* |
dcda9b047 mm, tree wide: re... |
2444 |
* For __GFP_RETRY_MAYFAIL allocations, stop reclaiming if the |
2876592f2 mm: vmscan: stop ... |
2445 2446 |
* full LRU list has been scanned and we are still failing * to reclaim pages. This full LRU scan is potentially |
dcda9b047 mm, tree wide: re... |
2447 |
* expensive but a __GFP_RETRY_MAYFAIL caller really wants to succeed |
2876592f2 mm: vmscan: stop ... |
2448 2449 2450 2451 2452 |
*/ if (!nr_reclaimed && !nr_scanned) return false; } else { /* |
dcda9b047 mm, tree wide: re... |
2453 |
* For non-__GFP_RETRY_MAYFAIL allocations which can presumably |
2876592f2 mm: vmscan: stop ... |
2454 2455 2456 2457 2458 2459 2460 2461 2462 |
* 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... |
2463 2464 2465 2466 2467 |
/* * If we have not reclaimed enough pages for compaction and the * inactive lists are large enough, continue reclaiming */ |
9861a62c3 mm, compaction: c... |
2468 |
pages_for_compaction = compact_gap(sc->order); |
a9dd0a831 mm, vmscan: make ... |
2469 |
inactive_lru_pages = node_page_state(pgdat, NR_INACTIVE_FILE); |
ec8acf20a swap: add per-par... |
2470 |
if (get_nr_swap_pages() > 0) |
a9dd0a831 mm, vmscan: make ... |
2471 |
inactive_lru_pages += node_page_state(pgdat, NR_INACTIVE_ANON); |
3e7d34497 mm: vmscan: recla... |
2472 2473 2474 2475 2476 |
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 */ |
a9dd0a831 mm, vmscan: make ... |
2477 2478 |
for (z = 0; z <= sc->reclaim_idx; z++) { struct zone *zone = &pgdat->node_zones[z]; |
6aa303def mm, vmscan: only ... |
2479 |
if (!managed_zone(zone)) |
a9dd0a831 mm, vmscan: make ... |
2480 2481 2482 |
continue; switch (compaction_suitable(zone, sc->order, 0, sc->reclaim_idx)) { |
cf378319d mm, compaction: r... |
2483 |
case COMPACT_SUCCESS: |
a9dd0a831 mm, vmscan: make ... |
2484 2485 2486 2487 2488 2489 |
case COMPACT_CONTINUE: return false; default: /* check next zone */ ; } |
3e7d34497 mm: vmscan: recla... |
2490 |
} |
a9dd0a831 mm, vmscan: make ... |
2491 |
return true; |
3e7d34497 mm: vmscan: recla... |
2492 |
} |
970a39a36 mm, vmscan: avoid... |
2493 |
static bool shrink_node(pg_data_t *pgdat, struct scan_control *sc) |
1da177e4c Linux-2.6.12-rc2 |
2494 |
{ |
cb731d6c6 vmscan: per memor... |
2495 |
struct reclaim_state *reclaim_state = current->reclaim_state; |
f0fdc5e8e vmscan: fix zone ... |
2496 |
unsigned long nr_reclaimed, nr_scanned; |
2344d7e44 mm: vmscan: remov... |
2497 |
bool reclaimable = false; |
1da177e4c Linux-2.6.12-rc2 |
2498 |
|
9b4f98cda mm: vmscan: compa... |
2499 2500 2501 |
do { struct mem_cgroup *root = sc->target_mem_cgroup; struct mem_cgroup_reclaim_cookie reclaim = { |
ef8f23279 mm, memcg: move m... |
2502 |
.pgdat = pgdat, |
9b4f98cda mm: vmscan: compa... |
2503 2504 |
.priority = sc->priority, }; |
a9dd0a831 mm, vmscan: make ... |
2505 |
unsigned long node_lru_pages = 0; |
694fbc0fe revert "memcg: en... |
2506 |
struct mem_cgroup *memcg; |
3e7d34497 mm: vmscan: recla... |
2507 |
|
9b4f98cda mm: vmscan: compa... |
2508 2509 |
nr_reclaimed = sc->nr_reclaimed; nr_scanned = sc->nr_scanned; |
1da177e4c Linux-2.6.12-rc2 |
2510 |
|
694fbc0fe revert "memcg: en... |
2511 2512 |
memcg = mem_cgroup_iter(root, NULL, &reclaim); do { |
6b4f7799c mm: vmscan: invok... |
2513 |
unsigned long lru_pages; |
8e8ae6452 mm: memcontrol: h... |
2514 |
unsigned long reclaimed; |
cb731d6c6 vmscan: per memor... |
2515 |
unsigned long scanned; |
5660048cc mm: move memcg hi... |
2516 |
|
241994ed8 mm: memcontrol: d... |
2517 |
if (mem_cgroup_low(root, memcg)) { |
d6622f636 mm/vmscan: more r... |
2518 2519 |
if (!sc->memcg_low_reclaim) { sc->memcg_low_skipped = 1; |
241994ed8 mm: memcontrol: d... |
2520 |
continue; |
d6622f636 mm/vmscan: more r... |
2521 |
} |
31176c781 mm: memcontrol: c... |
2522 |
mem_cgroup_event(memcg, MEMCG_LOW); |
241994ed8 mm: memcontrol: d... |
2523 |
} |
8e8ae6452 mm: memcontrol: h... |
2524 |
reclaimed = sc->nr_reclaimed; |
cb731d6c6 vmscan: per memor... |
2525 |
scanned = sc->nr_scanned; |
f9be23d6d mm/vmscan: kill s... |
2526 |
|
a9dd0a831 mm, vmscan: make ... |
2527 2528 |
shrink_node_memcg(pgdat, memcg, sc, &lru_pages); node_lru_pages += lru_pages; |
f16015fbf mm: vmscan: disti... |
2529 |
|
b5afba297 mm: vmscan: fix m... |
2530 |
if (memcg) |
a9dd0a831 mm, vmscan: make ... |
2531 |
shrink_slab(sc->gfp_mask, pgdat->node_id, |
cb731d6c6 vmscan: per memor... |
2532 2533 |
memcg, sc->nr_scanned - scanned, lru_pages); |
8e8ae6452 mm: memcontrol: h... |
2534 2535 2536 2537 |
/* Record the group's reclaim efficiency */ vmpressure(sc->gfp_mask, memcg, false, sc->nr_scanned - scanned, sc->nr_reclaimed - reclaimed); |
9b4f98cda mm: vmscan: compa... |
2538 |
/* |
a394cb8ee memcg,vmscan: do ... |
2539 2540 |
* Direct reclaim and kswapd have to scan all memory * cgroups to fulfill the overall scan target for the |
a9dd0a831 mm, vmscan: make ... |
2541 |
* node. |
a394cb8ee memcg,vmscan: do ... |
2542 2543 2544 2545 2546 |
* * Limit reclaim, on the other hand, only cares about * nr_to_reclaim pages to be reclaimed and it will * retry with decreasing priority if one round over the * whole hierarchy is not sufficient. |
9b4f98cda mm: vmscan: compa... |
2547 |
*/ |
a394cb8ee memcg,vmscan: do ... |
2548 2549 |
if (!global_reclaim(sc) && sc->nr_reclaimed >= sc->nr_to_reclaim) { |
9b4f98cda mm: vmscan: compa... |
2550 2551 2552 |
mem_cgroup_iter_break(root, memcg); break; } |
241994ed8 mm: memcontrol: d... |
2553 |
} while ((memcg = mem_cgroup_iter(root, memcg, &reclaim))); |
70ddf637e memcg: add memory... |
2554 |
|
6b4f7799c mm: vmscan: invok... |
2555 2556 2557 2558 |
/* * Shrink the slab caches in the same proportion that * the eligible LRU pages were scanned. */ |
b2e18757f mm, vmscan: begin... |
2559 |
if (global_reclaim(sc)) |
a9dd0a831 mm, vmscan: make ... |
2560 |
shrink_slab(sc->gfp_mask, pgdat->node_id, NULL, |
cb731d6c6 vmscan: per memor... |
2561 |
sc->nr_scanned - nr_scanned, |
a9dd0a831 mm, vmscan: make ... |
2562 |
node_lru_pages); |
cb731d6c6 vmscan: per memor... |
2563 2564 2565 2566 |
if (reclaim_state) { sc->nr_reclaimed += reclaim_state->reclaimed_slab; reclaim_state->reclaimed_slab = 0; |
6b4f7799c mm: vmscan: invok... |
2567 |
} |
8e8ae6452 mm: memcontrol: h... |
2568 2569 |
/* Record the subtree's reclaim efficiency */ vmpressure(sc->gfp_mask, sc->target_mem_cgroup, true, |
70ddf637e memcg: add memory... |
2570 2571 |
sc->nr_scanned - nr_scanned, sc->nr_reclaimed - nr_reclaimed); |
2344d7e44 mm: vmscan: remov... |
2572 2573 |
if (sc->nr_reclaimed - nr_reclaimed) reclaimable = true; |
a9dd0a831 mm, vmscan: make ... |
2574 |
} while (should_continue_reclaim(pgdat, sc->nr_reclaimed - nr_reclaimed, |
9b4f98cda mm: vmscan: compa... |
2575 |
sc->nr_scanned - nr_scanned, sc)); |
2344d7e44 mm: vmscan: remov... |
2576 |
|
c73322d09 mm: fix 100% CPU ... |
2577 2578 2579 2580 2581 2582 2583 2584 |
/* * Kswapd gives up on balancing particular nodes after too * many failures to reclaim anything from them and goes to * sleep. On reclaim progress, reset the failure counter. A * successful direct reclaim run will revive a dormant kswapd. */ if (reclaimable) pgdat->kswapd_failures = 0; |
2344d7e44 mm: vmscan: remov... |
2585 |
return reclaimable; |
f16015fbf mm: vmscan: disti... |
2586 |
} |
53853e2d2 mm, compaction: d... |
2587 |
/* |
fdd4c6149 mm, vmscan: make ... |
2588 2589 2590 |
* Returns true if compaction should go ahead for a costly-order request, or * the allocation would already succeed without compaction. Return false if we * should reclaim first. |
53853e2d2 mm, compaction: d... |
2591 |
*/ |
4f588331b mm, vmscan: avoid... |
2592 |
static inline bool compaction_ready(struct zone *zone, struct scan_control *sc) |
fe4b1b244 mm: vmscan: when ... |
2593 |
{ |
31483b6ad mm, vmscan: remov... |
2594 |
unsigned long watermark; |
fdd4c6149 mm, vmscan: make ... |
2595 |
enum compact_result suitable; |
fe4b1b244 mm: vmscan: when ... |
2596 |
|
fdd4c6149 mm, vmscan: make ... |
2597 2598 2599 2600 2601 2602 2603 |
suitable = compaction_suitable(zone, sc->order, 0, sc->reclaim_idx); if (suitable == COMPACT_SUCCESS) /* Allocation should succeed already. Don't reclaim. */ return true; if (suitable == COMPACT_SKIPPED) /* Compaction cannot yet proceed. Do reclaim. */ return false; |
fe4b1b244 mm: vmscan: when ... |
2604 |
|
53853e2d2 mm, compaction: d... |
2605 |
/* |
fdd4c6149 mm, vmscan: make ... |
2606 2607 2608 2609 2610 2611 2612 |
* Compaction is already possible, but it takes time to run and there * are potentially other callers using the pages just freed. So proceed * with reclaim to make a buffer of free pages available to give * compaction a reasonable chance of completing and allocating the page. * Note that we won't actually reclaim the whole buffer in one attempt * as the target watermark in should_continue_reclaim() is lower. But if * we are already above the high+gap watermark, don't reclaim at all. |
53853e2d2 mm, compaction: d... |
2613 |
*/ |
fdd4c6149 mm, vmscan: make ... |
2614 |
watermark = high_wmark_pages(zone) + compact_gap(sc->order); |
fe4b1b244 mm: vmscan: when ... |
2615 |
|
fdd4c6149 mm, vmscan: make ... |
2616 |
return zone_watermark_ok_safe(zone, 0, watermark, sc->reclaim_idx); |
fe4b1b244 mm: vmscan: when ... |
2617 |
} |
1da177e4c Linux-2.6.12-rc2 |
2618 2619 2620 2621 2622 |
/* * 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. * |
1da177e4c Linux-2.6.12-rc2 |
2623 2624 2625 |
* If a zone is deemed to be full of pinned pages then just give it a light * scan then give up on it. */ |
0a0337e0d mm, oom: rework o... |
2626 |
static void shrink_zones(struct zonelist *zonelist, struct scan_control *sc) |
1da177e4c Linux-2.6.12-rc2 |
2627 |
{ |
dd1a239f6 mm: have zonelist... |
2628 |
struct zoneref *z; |
54a6eb5c4 mm: use two zonel... |
2629 |
struct zone *zone; |
0608f43da revert "memcg, vm... |
2630 2631 |
unsigned long nr_soft_reclaimed; unsigned long nr_soft_scanned; |
619d0d76c mm/vmscan: restor... |
2632 |
gfp_t orig_mask; |
79dafcdca mm, vmscan: by de... |
2633 |
pg_data_t *last_pgdat = NULL; |
1cfb419b3 per-zone and recl... |
2634 |
|
cc715d99e mm: vmscan: forci... |
2635 2636 2637 2638 2639 |
/* * If the number of buffer_heads in the machine exceeds the maximum * allowed level, force direct reclaim to scan the highmem zone as * highmem pages could be pinning lowmem pages storing buffer_heads */ |
619d0d76c mm/vmscan: restor... |
2640 |
orig_mask = sc->gfp_mask; |
b2e18757f mm, vmscan: begin... |
2641 |
if (buffer_heads_over_limit) { |
cc715d99e mm: vmscan: forci... |
2642 |
sc->gfp_mask |= __GFP_HIGHMEM; |
4f588331b mm, vmscan: avoid... |
2643 |
sc->reclaim_idx = gfp_zone(sc->gfp_mask); |
b2e18757f mm, vmscan: begin... |
2644 |
} |
cc715d99e mm: vmscan: forci... |
2645 |
|
d4debc66d vmscan: remove un... |
2646 |
for_each_zone_zonelist_nodemask(zone, z, zonelist, |
b2e18757f mm, vmscan: begin... |
2647 |
sc->reclaim_idx, sc->nodemask) { |
b2e18757f mm, vmscan: begin... |
2648 |
/* |
1cfb419b3 per-zone and recl... |
2649 2650 2651 |
* Take care memory controller reclaiming has small influence * to global LRU. */ |
89b5fae53 mm: vmscan: disti... |
2652 |
if (global_reclaim(sc)) { |
344736f29 cpuset: simplify ... |
2653 2654 |
if (!cpuset_zone_allowed(zone, GFP_KERNEL | __GFP_HARDWALL)) |
1cfb419b3 per-zone and recl... |
2655 |
continue; |
65ec02cb9 mm: vmscan: move ... |
2656 |
|
0b06496a3 mm: vmscan: rewor... |
2657 2658 2659 2660 2661 2662 2663 2664 2665 2666 2667 |
/* * If we already have plenty of memory free for * compaction in this zone, don't free any more. * Even though compaction is invoked for any * non-zero order, only frequent costly order * reclamation is disruptive enough to become a * noticeable problem, like transparent huge * page allocations. */ if (IS_ENABLED(CONFIG_COMPACTION) && sc->order > PAGE_ALLOC_COSTLY_ORDER && |
4f588331b mm, vmscan: avoid... |
2668 |
compaction_ready(zone, sc)) { |
0b06496a3 mm: vmscan: rewor... |
2669 2670 |
sc->compaction_ready = true; continue; |
e0887c19b vmscan: limit dir... |
2671 |
} |
0b06496a3 mm: vmscan: rewor... |
2672 |
|
0608f43da revert "memcg, vm... |
2673 |
/* |
79dafcdca mm, vmscan: by de... |
2674 2675 2676 2677 2678 2679 2680 2681 2682 |
* Shrink each node in the zonelist once. If the * zonelist is ordered by zone (not the default) then a * node may be shrunk multiple times but in that case * the user prefers lower zones being preserved. */ if (zone->zone_pgdat == last_pgdat) continue; /* |
0608f43da revert "memcg, vm... |
2683 2684 2685 2686 2687 2688 |
* 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; |
ef8f23279 mm, memcg: move m... |
2689 |
nr_soft_reclaimed = mem_cgroup_soft_limit_reclaim(zone->zone_pgdat, |
0608f43da revert "memcg, vm... |
2690 2691 2692 2693 |
sc->order, sc->gfp_mask, &nr_soft_scanned); sc->nr_reclaimed += nr_soft_reclaimed; sc->nr_scanned += nr_soft_scanned; |
ac34a1a3c memcg: fix direct... |
2694 |
/* need some check for avoid more shrink_zone() */ |
1cfb419b3 per-zone and recl... |
2695 |
} |
408d85441 [PATCH] oom: use ... |
2696 |
|
79dafcdca mm, vmscan: by de... |
2697 2698 2699 2700 |
/* See comment about same check for global reclaim above */ if (zone->zone_pgdat == last_pgdat) continue; last_pgdat = zone->zone_pgdat; |
970a39a36 mm, vmscan: avoid... |
2701 |
shrink_node(zone->zone_pgdat, sc); |
1da177e4c Linux-2.6.12-rc2 |
2702 |
} |
e0c23279c vmscan: abort rec... |
2703 |
|
65ec02cb9 mm: vmscan: move ... |
2704 |
/* |
619d0d76c mm/vmscan: restor... |
2705 2706 2707 2708 |
* Restore to original mask to avoid the impact on the caller if we * promoted it to __GFP_HIGHMEM. */ sc->gfp_mask = orig_mask; |
1da177e4c Linux-2.6.12-rc2 |
2709 |
} |
4f98a2fee vmscan: split LRU... |
2710 |
|
2a2e48854 mm: vmscan: fix I... |
2711 2712 2713 2714 2715 2716 2717 2718 2719 2720 |
static void snapshot_refaults(struct mem_cgroup *root_memcg, pg_data_t *pgdat) { struct mem_cgroup *memcg; memcg = mem_cgroup_iter(root_memcg, NULL, NULL); do { unsigned long refaults; struct lruvec *lruvec; if (memcg) |
ccda7f436 mm: memcontrol: u... |
2721 |
refaults = memcg_page_state(memcg, WORKINGSET_ACTIVATE); |
2a2e48854 mm: vmscan: fix I... |
2722 2723 2724 2725 2726 2727 2728 |
else refaults = node_page_state(pgdat, WORKINGSET_ACTIVATE); lruvec = mem_cgroup_lruvec(pgdat, memcg); lruvec->refaults = refaults; } while ((memcg = mem_cgroup_iter(root_memcg, memcg, NULL))); } |
1da177e4c Linux-2.6.12-rc2 |
2729 2730 2731 2732 2733 2734 2735 2736 |
/* * 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... |
2737 2738 2739 2740 |
* 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... |
2741 2742 2743 |
* * returns: 0, if no pages reclaimed * else, the number of pages reclaimed |
1da177e4c Linux-2.6.12-rc2 |
2744 |
*/ |
dac1d27bc mm: use zonelists... |
2745 |
static unsigned long do_try_to_free_pages(struct zonelist *zonelist, |
3115cd914 mm: vmscan: remov... |
2746 |
struct scan_control *sc) |
1da177e4c Linux-2.6.12-rc2 |
2747 |
{ |
241994ed8 mm: memcontrol: d... |
2748 |
int initial_priority = sc->priority; |
2a2e48854 mm: vmscan: fix I... |
2749 2750 2751 |
pg_data_t *last_pgdat; struct zoneref *z; struct zone *zone; |
241994ed8 mm: memcontrol: d... |
2752 |
retry: |
873b47717 per-task-delay-ac... |
2753 |
delayacct_freepages_start(); |
89b5fae53 mm: vmscan: disti... |
2754 |
if (global_reclaim(sc)) |
7cc30fcfd mm: vmstat: accou... |
2755 |
__count_zid_vm_events(ALLOCSTALL, sc->reclaim_idx, 1); |
1da177e4c Linux-2.6.12-rc2 |
2756 |
|
9e3b2f8cd mm/vmscan: store ... |
2757 |
do { |
70ddf637e memcg: add memory... |
2758 2759 |
vmpressure_prio(sc->gfp_mask, sc->target_mem_cgroup, sc->priority); |
66e1707bc Memory controller... |
2760 |
sc->nr_scanned = 0; |
0a0337e0d mm, oom: rework o... |
2761 |
shrink_zones(zonelist, sc); |
c6a8a8c58 vmscan: recalcula... |
2762 |
|
bb21c7ce1 vmscan: fix do_tr... |
2763 |
if (sc->nr_reclaimed >= sc->nr_to_reclaim) |
0b06496a3 mm: vmscan: rewor... |
2764 2765 2766 2767 |
break; if (sc->compaction_ready) break; |
1da177e4c Linux-2.6.12-rc2 |
2768 2769 |
/* |
0e50ce3b5 mm: use up free s... |
2770 2771 2772 2773 2774 |
* If we're getting trouble reclaiming, start doing * writepage even in laptop mode. */ if (sc->priority < DEF_PRIORITY - 2) sc->may_writepage = 1; |
0b06496a3 mm: vmscan: rewor... |
2775 |
} while (--sc->priority >= 0); |
bb21c7ce1 vmscan: fix do_tr... |
2776 |
|
2a2e48854 mm: vmscan: fix I... |
2777 2778 2779 2780 2781 2782 2783 2784 |
last_pgdat = NULL; for_each_zone_zonelist_nodemask(zone, z, zonelist, sc->reclaim_idx, sc->nodemask) { if (zone->zone_pgdat == last_pgdat) continue; last_pgdat = zone->zone_pgdat; snapshot_refaults(sc->target_mem_cgroup, zone->zone_pgdat); } |
873b47717 per-task-delay-ac... |
2785 |
delayacct_freepages_end(); |
bb21c7ce1 vmscan: fix do_tr... |
2786 2787 |
if (sc->nr_reclaimed) return sc->nr_reclaimed; |
0cee34fd7 mm: vmscan: check... |
2788 |
/* Aborted reclaim to try compaction? don't OOM, then */ |
0b06496a3 mm: vmscan: rewor... |
2789 |
if (sc->compaction_ready) |
7335084d4 mm: vmscan: do no... |
2790 |
return 1; |
241994ed8 mm: memcontrol: d... |
2791 |
/* Untapped cgroup reserves? Don't OOM, retry. */ |
d6622f636 mm/vmscan: more r... |
2792 |
if (sc->memcg_low_skipped) { |
241994ed8 mm: memcontrol: d... |
2793 |
sc->priority = initial_priority; |
d6622f636 mm/vmscan: more r... |
2794 2795 |
sc->memcg_low_reclaim = 1; sc->memcg_low_skipped = 0; |
241994ed8 mm: memcontrol: d... |
2796 2797 |
goto retry; } |
bb21c7ce1 vmscan: fix do_tr... |
2798 |
return 0; |
1da177e4c Linux-2.6.12-rc2 |
2799 |
} |
c73322d09 mm: fix 100% CPU ... |
2800 |
static bool allow_direct_reclaim(pg_data_t *pgdat) |
5515061d2 mm: throttle dire... |
2801 2802 2803 2804 2805 2806 |
{ struct zone *zone; unsigned long pfmemalloc_reserve = 0; unsigned long free_pages = 0; int i; bool wmark_ok; |
c73322d09 mm: fix 100% CPU ... |
2807 2808 |
if (pgdat->kswapd_failures >= MAX_RECLAIM_RETRIES) return true; |
5515061d2 mm: throttle dire... |
2809 2810 |
for (i = 0; i <= ZONE_NORMAL; i++) { zone = &pgdat->node_zones[i]; |
d450abd81 mm: fix check for... |
2811 2812 2813 2814 |
if (!managed_zone(zone)) continue; if (!zone_reclaimable_pages(zone)) |
675becce1 mm: vmscan: do no... |
2815 |
continue; |
5515061d2 mm: throttle dire... |
2816 2817 2818 |
pfmemalloc_reserve += min_wmark_pages(zone); free_pages += zone_page_state(zone, NR_FREE_PAGES); } |
675becce1 mm: vmscan: do no... |
2819 2820 2821 |
/* If there are no reserves (unexpected config) then do not throttle */ if (!pfmemalloc_reserve) return true; |
5515061d2 mm: throttle dire... |
2822 2823 2824 2825 |
wmark_ok = free_pages > pfmemalloc_reserve / 2; /* kswapd must be awake if processes are being throttled */ if (!wmark_ok && waitqueue_active(&pgdat->kswapd_wait)) { |
38087d9b0 mm, vmscan: simpl... |
2826 |
pgdat->kswapd_classzone_idx = min(pgdat->kswapd_classzone_idx, |
5515061d2 mm: throttle dire... |
2827 2828 2829 2830 2831 2832 2833 2834 2835 2836 2837 |
(enum zone_type)ZONE_NORMAL); wake_up_interruptible(&pgdat->kswapd_wait); } return wmark_ok; } /* * Throttle direct reclaimers if backing storage is backed by the network * and the PFMEMALLOC reserve for the preferred node is getting dangerously * depleted. kswapd will continue to make progress and wake the processes |
50694c28f mm: vmscan: check... |
2838 2839 2840 2841 |
* when the low watermark is reached. * * Returns true if a fatal signal was delivered during throttling. If this * happens, the page allocator should not consider triggering the OOM killer. |
5515061d2 mm: throttle dire... |
2842 |
*/ |
50694c28f mm: vmscan: check... |
2843 |
static bool throttle_direct_reclaim(gfp_t gfp_mask, struct zonelist *zonelist, |
5515061d2 mm: throttle dire... |
2844 2845 |
nodemask_t *nodemask) { |
675becce1 mm: vmscan: do no... |
2846 |
struct zoneref *z; |
5515061d2 mm: throttle dire... |
2847 |
struct zone *zone; |
675becce1 mm: vmscan: do no... |
2848 |
pg_data_t *pgdat = NULL; |
5515061d2 mm: throttle dire... |
2849 2850 2851 2852 2853 2854 2855 2856 2857 |
/* * Kernel threads should not be throttled as they may be indirectly * responsible for cleaning pages necessary for reclaim to make forward * progress. kjournald for example may enter direct reclaim while * committing a transaction where throttling it could forcing other * processes to block on log_wait_commit(). */ if (current->flags & PF_KTHREAD) |
50694c28f mm: vmscan: check... |
2858 2859 2860 2861 2862 2863 2864 2865 |
goto out; /* * If a fatal signal is pending, this process should not throttle. * It should return quickly so it can exit and free its memory */ if (fatal_signal_pending(current)) goto out; |
5515061d2 mm: throttle dire... |
2866 |
|
675becce1 mm: vmscan: do no... |
2867 2868 2869 2870 2871 2872 2873 2874 2875 2876 2877 2878 2879 2880 2881 |
/* * Check if the pfmemalloc reserves are ok by finding the first node * with a usable ZONE_NORMAL or lower zone. The expectation is that * GFP_KERNEL will be required for allocating network buffers when * swapping over the network so ZONE_HIGHMEM is unusable. * * Throttling is based on the first usable node and throttled processes * wait on a queue until kswapd makes progress and wakes them. There * is an affinity then between processes waking up and where reclaim * progress has been made assuming the process wakes on the same node. * More importantly, processes running on remote nodes will not compete * for remote pfmemalloc reserves and processes on different nodes * should make reasonable progress. */ for_each_zone_zonelist_nodemask(zone, z, zonelist, |
17636faad mm/vmscan: fix hi... |
2882 |
gfp_zone(gfp_mask), nodemask) { |
675becce1 mm: vmscan: do no... |
2883 2884 2885 2886 2887 |
if (zone_idx(zone) > ZONE_NORMAL) continue; /* Throttle based on the first usable node */ pgdat = zone->zone_pgdat; |
c73322d09 mm: fix 100% CPU ... |
2888 |
if (allow_direct_reclaim(pgdat)) |
675becce1 mm: vmscan: do no... |
2889 2890 2891 2892 2893 2894 |
goto out; break; } /* If no zone was usable by the allocation flags then do not throttle */ if (!pgdat) |
50694c28f mm: vmscan: check... |
2895 |
goto out; |
5515061d2 mm: throttle dire... |
2896 |
|
68243e76e mm: account for t... |
2897 2898 |
/* Account for the throttling */ count_vm_event(PGSCAN_DIRECT_THROTTLE); |
5515061d2 mm: throttle dire... |
2899 2900 2901 2902 2903 2904 2905 2906 2907 2908 |
/* * If the caller cannot enter the filesystem, it's possible that it * is due to the caller holding an FS lock or performing a journal * transaction in the case of a filesystem like ext[3|4]. In this case, * it is not safe to block on pfmemalloc_wait as kswapd could be * blocked waiting on the same lock. Instead, throttle for up to a * second before continuing. */ if (!(gfp_mask & __GFP_FS)) { wait_event_interruptible_timeout(pgdat->pfmemalloc_wait, |
c73322d09 mm: fix 100% CPU ... |
2909 |
allow_direct_reclaim(pgdat), HZ); |
50694c28f mm: vmscan: check... |
2910 2911 |
goto check_pending; |
5515061d2 mm: throttle dire... |
2912 2913 2914 2915 |
} /* Throttle until kswapd wakes the process */ wait_event_killable(zone->zone_pgdat->pfmemalloc_wait, |
c73322d09 mm: fix 100% CPU ... |
2916 |
allow_direct_reclaim(pgdat)); |
50694c28f mm: vmscan: check... |
2917 2918 2919 2920 2921 2922 2923 |
check_pending: if (fatal_signal_pending(current)) return true; out: return false; |
5515061d2 mm: throttle dire... |
2924 |
} |
dac1d27bc mm: use zonelists... |
2925 |
unsigned long try_to_free_pages(struct zonelist *zonelist, int order, |
327c0e968 vmscan: fix it to... |
2926 |
gfp_t gfp_mask, nodemask_t *nodemask) |
66e1707bc Memory controller... |
2927 |
{ |
33906bc5c vmscan: tracing: ... |
2928 |
unsigned long nr_reclaimed; |
66e1707bc Memory controller... |
2929 |
struct scan_control sc = { |
ee814fe23 mm: vmscan: clean... |
2930 |
.nr_to_reclaim = SWAP_CLUSTER_MAX, |
f2f43e566 mm/vmscan.c: fix ... |
2931 |
.gfp_mask = current_gfp_context(gfp_mask), |
b2e18757f mm, vmscan: begin... |
2932 |
.reclaim_idx = gfp_zone(gfp_mask), |
ee814fe23 mm: vmscan: clean... |
2933 2934 2935 |
.order = order, .nodemask = nodemask, .priority = DEF_PRIORITY, |
66e1707bc Memory controller... |
2936 |
.may_writepage = !laptop_mode, |
a6dc60f89 vmscan: rename sc... |
2937 |
.may_unmap = 1, |
2e2e42598 vmscan,memcg: rei... |
2938 |
.may_swap = 1, |
66e1707bc Memory controller... |
2939 |
}; |
5515061d2 mm: throttle dire... |
2940 |
/* |
50694c28f mm: vmscan: check... |
2941 2942 2943 |
* Do not enter reclaim if fatal signal was delivered while throttled. * 1 is returned so that the page allocator does not OOM kill at this * point. |
5515061d2 mm: throttle dire... |
2944 |
*/ |
f2f43e566 mm/vmscan.c: fix ... |
2945 |
if (throttle_direct_reclaim(sc.gfp_mask, zonelist, nodemask)) |
5515061d2 mm: throttle dire... |
2946 |
return 1; |
33906bc5c vmscan: tracing: ... |
2947 2948 |
trace_mm_vmscan_direct_reclaim_begin(order, sc.may_writepage, |
f2f43e566 mm/vmscan.c: fix ... |
2949 |
sc.gfp_mask, |
e5146b12e mm, vmscan: add c... |
2950 |
sc.reclaim_idx); |
33906bc5c vmscan: tracing: ... |
2951 |
|
3115cd914 mm: vmscan: remov... |
2952 |
nr_reclaimed = do_try_to_free_pages(zonelist, &sc); |
33906bc5c vmscan: tracing: ... |
2953 2954 2955 2956 |
trace_mm_vmscan_direct_reclaim_end(nr_reclaimed); return nr_reclaimed; |
66e1707bc Memory controller... |
2957 |
} |
c255a4580 memcg: rename con... |
2958 |
#ifdef CONFIG_MEMCG |
66e1707bc Memory controller... |
2959 |
|
a9dd0a831 mm, vmscan: make ... |
2960 |
unsigned long mem_cgroup_shrink_node(struct mem_cgroup *memcg, |
4e4169535 memory controller... |
2961 |
gfp_t gfp_mask, bool noswap, |
ef8f23279 mm, memcg: move m... |
2962 |
pg_data_t *pgdat, |
0ae5e89c6 memcg: count the ... |
2963 |
unsigned long *nr_scanned) |
4e4169535 memory controller... |
2964 2965 |
{ struct scan_control sc = { |
b8f5c5664 memcg: sc.nr_to_r... |
2966 |
.nr_to_reclaim = SWAP_CLUSTER_MAX, |
ee814fe23 mm: vmscan: clean... |
2967 |
.target_mem_cgroup = memcg, |
4e4169535 memory controller... |
2968 2969 |
.may_writepage = !laptop_mode, .may_unmap = 1, |
b2e18757f mm, vmscan: begin... |
2970 |
.reclaim_idx = MAX_NR_ZONES - 1, |
4e4169535 memory controller... |
2971 |
.may_swap = !noswap, |
4e4169535 memory controller... |
2972 |
}; |
6b4f7799c mm: vmscan: invok... |
2973 |
unsigned long lru_pages; |
0ae5e89c6 memcg: count the ... |
2974 |
|
4e4169535 memory controller... |
2975 2976 |
sc.gfp_mask = (gfp_mask & GFP_RECLAIM_MASK) | (GFP_HIGHUSER_MOVABLE & ~GFP_RECLAIM_MASK); |
bdce6d9eb memcg, vmscan: ad... |
2977 |
|
9e3b2f8cd mm/vmscan: store ... |
2978 |
trace_mm_vmscan_memcg_softlimit_reclaim_begin(sc.order, |
bdce6d9eb memcg, vmscan: ad... |
2979 |
sc.may_writepage, |
e5146b12e mm, vmscan: add c... |
2980 2981 |
sc.gfp_mask, sc.reclaim_idx); |
bdce6d9eb memcg, vmscan: ad... |
2982 |
|
4e4169535 memory controller... |
2983 2984 2985 |
/* * NOTE: Although we can get the priority field, using it * here is not a good idea, since it limits the pages we can scan. |
a9dd0a831 mm, vmscan: make ... |
2986 |
* if we don't reclaim here, the shrink_node from balance_pgdat |
4e4169535 memory controller... |
2987 2988 2989 |
* will pick up pages from other mem cgroup's as well. We hack * the priority and make it zero. */ |
ef8f23279 mm, memcg: move m... |
2990 |
shrink_node_memcg(pgdat, memcg, &sc, &lru_pages); |
bdce6d9eb memcg, vmscan: ad... |
2991 2992 |
trace_mm_vmscan_memcg_softlimit_reclaim_end(sc.nr_reclaimed); |
0ae5e89c6 memcg: count the ... |
2993 |
*nr_scanned = sc.nr_scanned; |
4e4169535 memory controller... |
2994 2995 |
return sc.nr_reclaimed; } |
72835c86c mm: unify remaini... |
2996 |
unsigned long try_to_free_mem_cgroup_pages(struct mem_cgroup *memcg, |
b70a2a21d mm: memcontrol: f... |
2997 |
unsigned long nr_pages, |
a7885eb8a memcg: swappiness |
2998 |
gfp_t gfp_mask, |
b70a2a21d mm: memcontrol: f... |
2999 |
bool may_swap) |
66e1707bc Memory controller... |
3000 |
{ |
4e4169535 memory controller... |
3001 |
struct zonelist *zonelist; |
bdce6d9eb memcg, vmscan: ad... |
3002 |
unsigned long nr_reclaimed; |
889976dbc memcg: reclaim me... |
3003 |
int nid; |
499118e96 mm: introduce mem... |
3004 |
unsigned int noreclaim_flag; |
66e1707bc Memory controller... |
3005 |
struct scan_control sc = { |
b70a2a21d mm: memcontrol: f... |
3006 |
.nr_to_reclaim = max(nr_pages, SWAP_CLUSTER_MAX), |
7dea19f9e mm: introduce mem... |
3007 |
.gfp_mask = (current_gfp_context(gfp_mask) & GFP_RECLAIM_MASK) | |
a09ed5e00 vmscan: change sh... |
3008 |
(GFP_HIGHUSER_MOVABLE & ~GFP_RECLAIM_MASK), |
b2e18757f mm, vmscan: begin... |
3009 |
.reclaim_idx = MAX_NR_ZONES - 1, |
ee814fe23 mm: vmscan: clean... |
3010 3011 3012 3013 |
.target_mem_cgroup = memcg, .priority = DEF_PRIORITY, .may_writepage = !laptop_mode, .may_unmap = 1, |
b70a2a21d mm: memcontrol: f... |
3014 |
.may_swap = may_swap, |
a09ed5e00 vmscan: change sh... |
3015 |
}; |
66e1707bc Memory controller... |
3016 |
|
889976dbc memcg: reclaim me... |
3017 3018 3019 3020 3021 |
/* * 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... |
3022 |
nid = mem_cgroup_select_victim_node(memcg); |
889976dbc memcg: reclaim me... |
3023 |
|
c9634cf01 mm: use zonelist ... |
3024 |
zonelist = &NODE_DATA(nid)->node_zonelists[ZONELIST_FALLBACK]; |
bdce6d9eb memcg, vmscan: ad... |
3025 3026 3027 |
trace_mm_vmscan_memcg_reclaim_begin(0, sc.may_writepage, |
e5146b12e mm, vmscan: add c... |
3028 3029 |
sc.gfp_mask, sc.reclaim_idx); |
bdce6d9eb memcg, vmscan: ad... |
3030 |
|
499118e96 mm: introduce mem... |
3031 |
noreclaim_flag = memalloc_noreclaim_save(); |
3115cd914 mm: vmscan: remov... |
3032 |
nr_reclaimed = do_try_to_free_pages(zonelist, &sc); |
499118e96 mm: introduce mem... |
3033 |
memalloc_noreclaim_restore(noreclaim_flag); |
bdce6d9eb memcg, vmscan: ad... |
3034 3035 3036 3037 |
trace_mm_vmscan_memcg_reclaim_end(nr_reclaimed); return nr_reclaimed; |
66e1707bc Memory controller... |
3038 3039 |
} #endif |
1d82de618 mm, vmscan: make ... |
3040 |
static void age_active_anon(struct pglist_data *pgdat, |
ef8f23279 mm, memcg: move m... |
3041 |
struct scan_control *sc) |
f16015fbf mm: vmscan: disti... |
3042 |
{ |
b95a2f2d4 mm: vmscan: conve... |
3043 |
struct mem_cgroup *memcg; |
f16015fbf mm: vmscan: disti... |
3044 |
|
b95a2f2d4 mm: vmscan: conve... |
3045 3046 3047 3048 3049 |
if (!total_swap_pages) return; memcg = mem_cgroup_iter(NULL, NULL, NULL); do { |
ef8f23279 mm, memcg: move m... |
3050 |
struct lruvec *lruvec = mem_cgroup_lruvec(pgdat, memcg); |
b95a2f2d4 mm: vmscan: conve... |
3051 |
|
2a2e48854 mm: vmscan: fix I... |
3052 |
if (inactive_list_is_low(lruvec, false, memcg, sc, true)) |
1a93be0e7 mm/vmscan: push l... |
3053 |
shrink_active_list(SWAP_CLUSTER_MAX, lruvec, |
9e3b2f8cd mm/vmscan: store ... |
3054 |
sc, LRU_ACTIVE_ANON); |
b95a2f2d4 mm: vmscan: conve... |
3055 3056 3057 |
memcg = mem_cgroup_iter(NULL, memcg, NULL); } while (memcg); |
f16015fbf mm: vmscan: disti... |
3058 |
} |
e716f2eb2 mm, vmscan: preve... |
3059 3060 3061 3062 3063 |
/* * Returns true if there is an eligible zone balanced for the request order * and classzone_idx */ static bool pgdat_balanced(pg_data_t *pgdat, int order, int classzone_idx) |
60cefed48 mm: vmscan: fix e... |
3064 |
{ |
e716f2eb2 mm, vmscan: preve... |
3065 3066 3067 |
int i; unsigned long mark = -1; struct zone *zone; |
60cefed48 mm: vmscan: fix e... |
3068 |
|
e716f2eb2 mm, vmscan: preve... |
3069 3070 |
for (i = 0; i <= classzone_idx; i++) { zone = pgdat->node_zones + i; |
6256c6b49 mm, vmscan: remov... |
3071 |
|
e716f2eb2 mm, vmscan: preve... |
3072 3073 3074 3075 3076 3077 3078 3079 3080 3081 3082 3083 3084 3085 3086 3087 3088 |
if (!managed_zone(zone)) continue; mark = high_wmark_pages(zone); if (zone_watermark_ok_safe(zone, order, mark, classzone_idx)) return true; } /* * If a node has no populated zone within classzone_idx, it does not * need balancing by definition. This can happen if a zone-restricted * allocation tries to wake a remote kswapd. */ if (mark == -1) return true; return false; |
60cefed48 mm: vmscan: fix e... |
3089 |
} |
631b6e083 mm, vmscan: only ... |
3090 3091 3092 3093 3094 3095 3096 |
/* Clear pgdat state for congested, dirty or under writeback. */ static void clear_pgdat_congested(pg_data_t *pgdat) { clear_bit(PGDAT_CONGESTED, &pgdat->flags); clear_bit(PGDAT_DIRTY, &pgdat->flags); clear_bit(PGDAT_WRITEBACK, &pgdat->flags); } |
1741c8775 mm: kswapd: keep ... |
3097 |
/* |
5515061d2 mm: throttle dire... |
3098 3099 3100 3101 3102 |
* Prepare kswapd for sleeping. This verifies that there are no processes * waiting in throttle_direct_reclaim() and that watermarks have been met. * * Returns true if kswapd is ready to sleep */ |
d9f21d426 mm, vmscan: avoid... |
3103 |
static bool prepare_kswapd_sleep(pg_data_t *pgdat, int order, int classzone_idx) |
f50de2d38 vmscan: have kswa... |
3104 |
{ |
5515061d2 mm: throttle dire... |
3105 |
/* |
9e5e36617 mm, vmscan: preve... |
3106 |
* The throttled processes are normally woken up in balance_pgdat() as |
c73322d09 mm: fix 100% CPU ... |
3107 |
* soon as allow_direct_reclaim() is true. But there is a potential |
9e5e36617 mm, vmscan: preve... |
3108 3109 3110 3111 3112 3113 3114 3115 3116 |
* race between when kswapd checks the watermarks and a process gets * throttled. There is also a potential race if processes get * throttled, kswapd wakes, a large process exits thereby balancing the * zones, which causes kswapd to exit balance_pgdat() before reaching * the wake up checks. If kswapd is going to sleep, no process should * be sleeping on pfmemalloc_wait, so wake them now if necessary. If * the wake up is premature, processes will wake kswapd and get * throttled again. The difference from wake ups in balance_pgdat() is * that here we are under prepare_to_wait(). |
5515061d2 mm: throttle dire... |
3117 |
*/ |
9e5e36617 mm, vmscan: preve... |
3118 3119 |
if (waitqueue_active(&pgdat->pfmemalloc_wait)) wake_up_all(&pgdat->pfmemalloc_wait); |
f50de2d38 vmscan: have kswa... |
3120 |
|
c73322d09 mm: fix 100% CPU ... |
3121 3122 3123 |
/* Hopeless node, leave it to direct reclaim */ if (pgdat->kswapd_failures >= MAX_RECLAIM_RETRIES) return true; |
e716f2eb2 mm, vmscan: preve... |
3124 3125 3126 |
if (pgdat_balanced(pgdat, order, classzone_idx)) { clear_pgdat_congested(pgdat); return true; |
1d82de618 mm, vmscan: make ... |
3127 |
} |
333b0a459 mm, vmscan: fix z... |
3128 |
return false; |
f50de2d38 vmscan: have kswa... |
3129 |
} |
1da177e4c Linux-2.6.12-rc2 |
3130 |
/* |
1d82de618 mm, vmscan: make ... |
3131 3132 |
* kswapd shrinks a node of pages that are at or below the highest usable * zone that is currently unbalanced. |
b8e83b942 mm: vmscan: flatt... |
3133 3134 |
* * Returns true if kswapd scanned at least the requested number of pages to |
283aba9f9 mm: vmscan: block... |
3135 3136 |
* reclaim or if the lack of progress was due to pages under writeback. * This is used to determine if the scanning priority needs to be raised. |
75485363c mm: vmscan: limit... |
3137 |
*/ |
1d82de618 mm, vmscan: make ... |
3138 |
static bool kswapd_shrink_node(pg_data_t *pgdat, |
accf62422 mm, kswapd: repla... |
3139 |
struct scan_control *sc) |
75485363c mm: vmscan: limit... |
3140 |
{ |
1d82de618 mm, vmscan: make ... |
3141 3142 |
struct zone *zone; int z; |
75485363c mm: vmscan: limit... |
3143 |
|
1d82de618 mm, vmscan: make ... |
3144 3145 |
/* Reclaim a number of pages proportional to the number of zones */ sc->nr_to_reclaim = 0; |
970a39a36 mm, vmscan: avoid... |
3146 |
for (z = 0; z <= sc->reclaim_idx; z++) { |
1d82de618 mm, vmscan: make ... |
3147 |
zone = pgdat->node_zones + z; |
6aa303def mm, vmscan: only ... |
3148 |
if (!managed_zone(zone)) |
1d82de618 mm, vmscan: make ... |
3149 |
continue; |
7c954f6de mm: vmscan: move ... |
3150 |
|
1d82de618 mm, vmscan: make ... |
3151 3152 |
sc->nr_to_reclaim += max(high_wmark_pages(zone), SWAP_CLUSTER_MAX); } |
7c954f6de mm: vmscan: move ... |
3153 3154 |
/* |
1d82de618 mm, vmscan: make ... |
3155 3156 |
* Historically care was taken to put equal pressure on all zones but * now pressure is applied based on node LRU order. |
7c954f6de mm: vmscan: move ... |
3157 |
*/ |
970a39a36 mm, vmscan: avoid... |
3158 |
shrink_node(pgdat, sc); |
283aba9f9 mm: vmscan: block... |
3159 |
|
7c954f6de mm: vmscan: move ... |
3160 |
/* |
1d82de618 mm, vmscan: make ... |
3161 3162 3163 3164 3165 |
* Fragmentation may mean that the system cannot be rebalanced for * high-order allocations. If twice the allocation size has been * reclaimed then recheck watermarks only at order-0 to prevent * excessive reclaim. Assume that a process requested a high-order * can direct reclaim/compact. |
7c954f6de mm: vmscan: move ... |
3166 |
*/ |
9861a62c3 mm, compaction: c... |
3167 |
if (sc->order && sc->nr_reclaimed >= compact_gap(sc->order)) |
1d82de618 mm, vmscan: make ... |
3168 |
sc->order = 0; |
7c954f6de mm: vmscan: move ... |
3169 |
|
b8e83b942 mm: vmscan: flatt... |
3170 |
return sc->nr_scanned >= sc->nr_to_reclaim; |
75485363c mm: vmscan: limit... |
3171 3172 3173 |
} /* |
1d82de618 mm, vmscan: make ... |
3174 3175 3176 |
* For kswapd, balance_pgdat() will reclaim pages across a node from zones * that are eligible for use by the caller until at least one zone is * balanced. |
1da177e4c Linux-2.6.12-rc2 |
3177 |
* |
1d82de618 mm, vmscan: make ... |
3178 |
* Returns the order kswapd finished reclaiming at. |
1da177e4c Linux-2.6.12-rc2 |
3179 3180 |
* * kswapd scans the zones in the highmem->normal->dma direction. It skips |
418589663 page allocator: u... |
3181 |
* zones which have free_pages > high_wmark_pages(zone), but once a zone is |
1d82de618 mm, vmscan: make ... |
3182 3183 3184 |
* found to have free_pages <= high_wmark_pages(zone), any page is that zone * or lower is eligible for reclaim until at least one usable zone is * balanced. |
1da177e4c Linux-2.6.12-rc2 |
3185 |
*/ |
accf62422 mm, kswapd: repla... |
3186 |
static int balance_pgdat(pg_data_t *pgdat, int order, int classzone_idx) |
1da177e4c Linux-2.6.12-rc2 |
3187 |
{ |
1da177e4c Linux-2.6.12-rc2 |
3188 |
int i; |
0608f43da revert "memcg, vm... |
3189 3190 |
unsigned long nr_soft_reclaimed; unsigned long nr_soft_scanned; |
1d82de618 mm, vmscan: make ... |
3191 |
struct zone *zone; |
179e96395 [PATCH] vmscan: s... |
3192 3193 |
struct scan_control sc = { .gfp_mask = GFP_KERNEL, |
ee814fe23 mm: vmscan: clean... |
3194 |
.order = order, |
b8e83b942 mm: vmscan: flatt... |
3195 |
.priority = DEF_PRIORITY, |
ee814fe23 mm: vmscan: clean... |
3196 |
.may_writepage = !laptop_mode, |
a6dc60f89 vmscan: rename sc... |
3197 |
.may_unmap = 1, |
2e2e42598 vmscan,memcg: rei... |
3198 |
.may_swap = 1, |
179e96395 [PATCH] vmscan: s... |
3199 |
}; |
f8891e5e1 [PATCH] Light wei... |
3200 |
count_vm_event(PAGEOUTRUN); |
1da177e4c Linux-2.6.12-rc2 |
3201 |
|
9e3b2f8cd mm/vmscan: store ... |
3202 |
do { |
c73322d09 mm: fix 100% CPU ... |
3203 |
unsigned long nr_reclaimed = sc.nr_reclaimed; |
b8e83b942 mm: vmscan: flatt... |
3204 |
bool raise_priority = true; |
84c7a7771 mm, vmscan: Have ... |
3205 |
sc.reclaim_idx = classzone_idx; |
1da177e4c Linux-2.6.12-rc2 |
3206 |
|
86c79f6b5 mm: vmscan: do no... |
3207 |
/* |
84c7a7771 mm, vmscan: Have ... |
3208 3209 3210 3211 3212 3213 3214 3215 |
* If the number of buffer_heads exceeds the maximum allowed * then consider reclaiming from all zones. This has a dual * purpose -- on 64-bit systems it is expected that * buffer_heads are stripped during active rotation. On 32-bit * systems, highmem pages can pin lowmem memory and shrinking * buffers can relieve lowmem pressure. Reclaim may still not * go ahead if all eligible zones for the original allocation * request are balanced to avoid excessive reclaim from kswapd. |
86c79f6b5 mm: vmscan: do no... |
3216 3217 3218 3219 |
*/ if (buffer_heads_over_limit) { for (i = MAX_NR_ZONES - 1; i >= 0; i--) { zone = pgdat->node_zones + i; |
6aa303def mm, vmscan: only ... |
3220 |
if (!managed_zone(zone)) |
86c79f6b5 mm: vmscan: do no... |
3221 |
continue; |
cc715d99e mm: vmscan: forci... |
3222 |
|
970a39a36 mm, vmscan: avoid... |
3223 |
sc.reclaim_idx = i; |
e1dbeda60 [PATCH] balance_p... |
3224 |
break; |
1da177e4c Linux-2.6.12-rc2 |
3225 |
} |
1da177e4c Linux-2.6.12-rc2 |
3226 |
} |
dafcb73e3 mm: avoid calling... |
3227 |
|
86c79f6b5 mm: vmscan: do no... |
3228 |
/* |
e716f2eb2 mm, vmscan: preve... |
3229 3230 3231 |
* Only reclaim if there are no eligible zones. Note that * sc.reclaim_idx is not used as buffer_heads_over_limit may * have adjusted it. |
86c79f6b5 mm: vmscan: do no... |
3232 |
*/ |
e716f2eb2 mm, vmscan: preve... |
3233 3234 |
if (pgdat_balanced(pgdat, sc.order, classzone_idx)) goto out; |
e1dbeda60 [PATCH] balance_p... |
3235 |
|
1da177e4c Linux-2.6.12-rc2 |
3236 |
/* |
1d82de618 mm, vmscan: make ... |
3237 3238 3239 3240 3241 |
* Do some background aging of the anon list, to give * pages a chance to be referenced before reclaiming. All * pages are rotated regardless of classzone as this is * about consistent aging. */ |
ef8f23279 mm, memcg: move m... |
3242 |
age_active_anon(pgdat, &sc); |
1d82de618 mm, vmscan: make ... |
3243 3244 |
/* |
b7ea3c417 mm: vmscan: check... |
3245 3246 3247 |
* If we're getting trouble reclaiming, start doing writepage * even in laptop mode. */ |
047d72c30 mm: remove seemin... |
3248 |
if (sc.priority < DEF_PRIORITY - 2) |
b7ea3c417 mm: vmscan: check... |
3249 |
sc.may_writepage = 1; |
1d82de618 mm, vmscan: make ... |
3250 3251 3252 |
/* Call soft limit reclaim before calling shrink_node. */ sc.nr_scanned = 0; nr_soft_scanned = 0; |
ef8f23279 mm, memcg: move m... |
3253 |
nr_soft_reclaimed = mem_cgroup_soft_limit_reclaim(pgdat, sc.order, |
1d82de618 mm, vmscan: make ... |
3254 3255 |
sc.gfp_mask, &nr_soft_scanned); sc.nr_reclaimed += nr_soft_reclaimed; |
b7ea3c417 mm: vmscan: check... |
3256 |
/* |
1d82de618 mm, vmscan: make ... |
3257 3258 3259 |
* There should be no need to raise the scanning priority if * enough pages are already being scanned that that high * watermark would be met at 100% efficiency. |
1da177e4c Linux-2.6.12-rc2 |
3260 |
*/ |
970a39a36 mm, vmscan: avoid... |
3261 |
if (kswapd_shrink_node(pgdat, &sc)) |
1d82de618 mm, vmscan: make ... |
3262 |
raise_priority = false; |
5515061d2 mm: throttle dire... |
3263 3264 3265 3266 3267 3268 3269 |
/* * If the low watermark is met there is no need for processes * to be throttled on pfmemalloc_wait as they should not be * able to safely make forward progress. Wake them */ if (waitqueue_active(&pgdat->pfmemalloc_wait) && |
c73322d09 mm: fix 100% CPU ... |
3270 |
allow_direct_reclaim(pgdat)) |
cfc511557 mm, vmscan: wake ... |
3271 |
wake_up_all(&pgdat->pfmemalloc_wait); |
5515061d2 mm: throttle dire... |
3272 |
|
b8e83b942 mm: vmscan: flatt... |
3273 3274 3275 |
/* Check if kswapd should be suspending */ if (try_to_freeze() || kthread_should_stop()) break; |
8357376d3 [PATCH] swsusp: I... |
3276 |
|
73ce02e96 mm: stop kswapd's... |
3277 |
/* |
b8e83b942 mm: vmscan: flatt... |
3278 3279 |
* Raise priority if scanning rate is too low or there was no * progress in reclaiming pages |
73ce02e96 mm: stop kswapd's... |
3280 |
*/ |
c73322d09 mm: fix 100% CPU ... |
3281 3282 |
nr_reclaimed = sc.nr_reclaimed - nr_reclaimed; if (raise_priority || !nr_reclaimed) |
b8e83b942 mm: vmscan: flatt... |
3283 |
sc.priority--; |
1d82de618 mm, vmscan: make ... |
3284 |
} while (sc.priority >= 1); |
1da177e4c Linux-2.6.12-rc2 |
3285 |
|
c73322d09 mm: fix 100% CPU ... |
3286 3287 |
if (!sc.nr_reclaimed) pgdat->kswapd_failures++; |
b8e83b942 mm: vmscan: flatt... |
3288 |
out: |
2a2e48854 mm: vmscan: fix I... |
3289 |
snapshot_refaults(NULL, pgdat); |
0abdee2bd mm: kswapd: use t... |
3290 |
/* |
1d82de618 mm, vmscan: make ... |
3291 3292 3293 3294 |
* Return the order kswapd stopped reclaiming at as * prepare_kswapd_sleep() takes it into account. If another caller * entered the allocator slow path while kswapd was awake, order will * remain at the higher level. |
0abdee2bd mm: kswapd: use t... |
3295 |
*/ |
1d82de618 mm, vmscan: make ... |
3296 |
return sc.order; |
1da177e4c Linux-2.6.12-rc2 |
3297 |
} |
e716f2eb2 mm, vmscan: preve... |
3298 3299 3300 3301 3302 3303 3304 3305 3306 3307 3308 3309 3310 3311 3312 |
/* * pgdat->kswapd_classzone_idx is the highest zone index that a recent * allocation request woke kswapd for. When kswapd has not woken recently, * the value is MAX_NR_ZONES which is not a valid index. This compares a * given classzone and returns it or the highest classzone index kswapd * was recently woke for. */ static enum zone_type kswapd_classzone_idx(pg_data_t *pgdat, enum zone_type classzone_idx) { if (pgdat->kswapd_classzone_idx == MAX_NR_ZONES) return classzone_idx; return max(pgdat->kswapd_classzone_idx, classzone_idx); } |
38087d9b0 mm, vmscan: simpl... |
3313 3314 |
static void kswapd_try_to_sleep(pg_data_t *pgdat, int alloc_order, int reclaim_order, unsigned int classzone_idx) |
f0bc0a60b vmscan: factor ou... |
3315 3316 3317 3318 3319 3320 3321 3322 |
{ long remaining = 0; DEFINE_WAIT(wait); if (freezing(current) || kthread_should_stop()) return; prepare_to_wait(&pgdat->kswapd_wait, &wait, TASK_INTERRUPTIBLE); |
333b0a459 mm, vmscan: fix z... |
3323 3324 3325 3326 3327 3328 3329 |
/* * Try to sleep for a short interval. Note that kcompactd will only be * woken if it is possible to sleep for a short interval. This is * deliberate on the assumption that if reclaim cannot keep an * eligible zone balanced that it's also unlikely that compaction will * succeed. */ |
d9f21d426 mm, vmscan: avoid... |
3330 |
if (prepare_kswapd_sleep(pgdat, reclaim_order, classzone_idx)) { |
fd901c953 mm: wake kcompact... |
3331 3332 3333 3334 3335 3336 3337 3338 3339 3340 3341 3342 |
/* * Compaction records what page blocks it recently failed to * isolate pages from and skips them in the future scanning. * When kswapd is going to sleep, it is reasonable to assume * that pages and compaction may succeed so reset the cache. */ reset_isolation_suitable(pgdat); /* * We have freed the memory, now we should compact it to make * allocation of the requested order possible. */ |
38087d9b0 mm, vmscan: simpl... |
3343 |
wakeup_kcompactd(pgdat, alloc_order, classzone_idx); |
fd901c953 mm: wake kcompact... |
3344 |
|
f0bc0a60b vmscan: factor ou... |
3345 |
remaining = schedule_timeout(HZ/10); |
38087d9b0 mm, vmscan: simpl... |
3346 3347 3348 3349 3350 3351 3352 |
/* * If woken prematurely then reset kswapd_classzone_idx and * order. The values will either be from a wakeup request or * the previous request that slept prematurely. */ if (remaining) { |
e716f2eb2 mm, vmscan: preve... |
3353 |
pgdat->kswapd_classzone_idx = kswapd_classzone_idx(pgdat, classzone_idx); |
38087d9b0 mm, vmscan: simpl... |
3354 3355 |
pgdat->kswapd_order = max(pgdat->kswapd_order, reclaim_order); } |
f0bc0a60b vmscan: factor ou... |
3356 3357 3358 3359 3360 3361 3362 3363 |
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. */ |
d9f21d426 mm, vmscan: avoid... |
3364 3365 |
if (!remaining && prepare_kswapd_sleep(pgdat, reclaim_order, classzone_idx)) { |
f0bc0a60b vmscan: factor ou... |
3366 3367 3368 3369 3370 3371 3372 3373 3374 3375 3376 |
trace_mm_vmscan_kswapd_sleep(pgdat->node_id); /* * vmstat counters are not perfectly accurate and the estimated * value for counters such as NR_FREE_PAGES can deviate from the * true value by nr_online_cpus * threshold. To avoid the zone * watermarks being breached while under pressure, we reduce the * per-cpu vmstat threshold while kswapd is awake and restore * them before going back to sleep. */ set_pgdat_percpu_threshold(pgdat, calculate_normal_threshold); |
1c7e7f6c0 mm: fix lost kswa... |
3377 3378 3379 |
if (!kthread_should_stop()) schedule(); |
f0bc0a60b vmscan: factor ou... |
3380 3381 3382 3383 3384 3385 3386 3387 3388 |
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 |
3389 3390 |
/* * The background pageout daemon, started as a kernel thread |
4f98a2fee vmscan: split LRU... |
3391 |
* from the init process. |
1da177e4c Linux-2.6.12-rc2 |
3392 3393 3394 3395 3396 3397 3398 3399 3400 3401 3402 3403 |
* * 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) { |
e716f2eb2 mm, vmscan: preve... |
3404 3405 |
unsigned int alloc_order, reclaim_order; unsigned int classzone_idx = MAX_NR_ZONES - 1; |
1da177e4c Linux-2.6.12-rc2 |
3406 3407 |
pg_data_t *pgdat = (pg_data_t*)p; struct task_struct *tsk = current; |
f0bc0a60b vmscan: factor ou... |
3408 |
|
1da177e4c Linux-2.6.12-rc2 |
3409 3410 3411 |
struct reclaim_state reclaim_state = { .reclaimed_slab = 0, }; |
a70f73028 cpumask: replace ... |
3412 |
const struct cpumask *cpumask = cpumask_of_node(pgdat->node_id); |
1da177e4c Linux-2.6.12-rc2 |
3413 |
|
174596a0b cpumask: convert mm/ |
3414 |
if (!cpumask_empty(cpumask)) |
c5f59f083 nodemask: use new... |
3415 |
set_cpus_allowed_ptr(tsk, cpumask); |
1da177e4c Linux-2.6.12-rc2 |
3416 3417 3418 3419 3420 3421 3422 3423 3424 3425 3426 3427 3428 3429 |
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... |
3430 |
tsk->flags |= PF_MEMALLOC | PF_SWAPWRITE | PF_KSWAPD; |
831441862 Freezer: make ker... |
3431 |
set_freezable(); |
1da177e4c Linux-2.6.12-rc2 |
3432 |
|
e716f2eb2 mm, vmscan: preve... |
3433 3434 |
pgdat->kswapd_order = 0; pgdat->kswapd_classzone_idx = MAX_NR_ZONES; |
1da177e4c Linux-2.6.12-rc2 |
3435 |
for ( ; ; ) { |
6f6313d48 mm/vmscan.c: try_... |
3436 |
bool ret; |
3e1d1d28d [PATCH] Cleanup p... |
3437 |
|
e716f2eb2 mm, vmscan: preve... |
3438 3439 |
alloc_order = reclaim_order = pgdat->kswapd_order; classzone_idx = kswapd_classzone_idx(pgdat, classzone_idx); |
38087d9b0 mm, vmscan: simpl... |
3440 3441 3442 |
kswapd_try_sleep: kswapd_try_to_sleep(pgdat, alloc_order, reclaim_order, classzone_idx); |
215ddd666 mm: vmscan: only ... |
3443 |
|
38087d9b0 mm, vmscan: simpl... |
3444 3445 |
/* Read the new order and classzone_idx */ alloc_order = reclaim_order = pgdat->kswapd_order; |
e716f2eb2 mm, vmscan: preve... |
3446 |
classzone_idx = kswapd_classzone_idx(pgdat, 0); |
38087d9b0 mm, vmscan: simpl... |
3447 |
pgdat->kswapd_order = 0; |
e716f2eb2 mm, vmscan: preve... |
3448 |
pgdat->kswapd_classzone_idx = MAX_NR_ZONES; |
1da177e4c Linux-2.6.12-rc2 |
3449 |
|
8fe23e057 mm: clear node in... |
3450 3451 3452 3453 3454 3455 3456 3457 |
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 */ |
38087d9b0 mm, vmscan: simpl... |
3458 3459 3460 3461 3462 3463 3464 3465 3466 3467 3468 |
if (ret) continue; /* * Reclaim begins at the requested order but if a high-order * reclaim fails then kswapd falls back to reclaiming for * order-0. If that happens, kswapd will consider sleeping * for the order it finished reclaiming at (reclaim_order) * but kcompactd is woken to compact for the original * request (alloc_order). */ |
e5146b12e mm, vmscan: add c... |
3469 3470 |
trace_mm_vmscan_kswapd_wake(pgdat->node_id, classzone_idx, alloc_order); |
d92a8cfcb locking/lockdep: ... |
3471 |
fs_reclaim_acquire(GFP_KERNEL); |
38087d9b0 mm, vmscan: simpl... |
3472 |
reclaim_order = balance_pgdat(pgdat, alloc_order, classzone_idx); |
d92a8cfcb locking/lockdep: ... |
3473 |
fs_reclaim_release(GFP_KERNEL); |
38087d9b0 mm, vmscan: simpl... |
3474 3475 |
if (reclaim_order < alloc_order) goto kswapd_try_sleep; |
1da177e4c Linux-2.6.12-rc2 |
3476 |
} |
b0a8cc58e mm: bugfix: set c... |
3477 |
|
71abdc15a mm: vmscan: clear... |
3478 |
tsk->flags &= ~(PF_MEMALLOC | PF_SWAPWRITE | PF_KSWAPD); |
b0a8cc58e mm: bugfix: set c... |
3479 |
current->reclaim_state = NULL; |
71abdc15a mm: vmscan: clear... |
3480 |
|
1da177e4c Linux-2.6.12-rc2 |
3481 3482 3483 3484 3485 3486 |
return 0; } /* * A zone is low on free memory, so wake its kswapd task to service it. */ |
995047488 mm: kswapd: stop ... |
3487 |
void wakeup_kswapd(struct zone *zone, int order, enum zone_type classzone_idx) |
1da177e4c Linux-2.6.12-rc2 |
3488 3489 |
{ pg_data_t *pgdat; |
6aa303def mm, vmscan: only ... |
3490 |
if (!managed_zone(zone)) |
1da177e4c Linux-2.6.12-rc2 |
3491 |
return; |
344736f29 cpuset: simplify ... |
3492 |
if (!cpuset_zone_allowed(zone, GFP_KERNEL | __GFP_HARDWALL)) |
1da177e4c Linux-2.6.12-rc2 |
3493 |
return; |
88f5acf88 mm: page allocato... |
3494 |
pgdat = zone->zone_pgdat; |
e716f2eb2 mm, vmscan: preve... |
3495 3496 |
pgdat->kswapd_classzone_idx = kswapd_classzone_idx(pgdat, classzone_idx); |
38087d9b0 mm, vmscan: simpl... |
3497 |
pgdat->kswapd_order = max(pgdat->kswapd_order, order); |
8d0986e28 [PATCH] vm: kswap... |
3498 |
if (!waitqueue_active(&pgdat->kswapd_wait)) |
1da177e4c Linux-2.6.12-rc2 |
3499 |
return; |
e1a556374 mm, vmscan: only ... |
3500 |
|
c73322d09 mm: fix 100% CPU ... |
3501 3502 3503 |
/* Hopeless node, leave it to direct reclaim */ if (pgdat->kswapd_failures >= MAX_RECLAIM_RETRIES) return; |
e716f2eb2 mm, vmscan: preve... |
3504 3505 |
if (pgdat_balanced(pgdat, order, classzone_idx)) return; |
88f5acf88 mm: page allocato... |
3506 |
|
e716f2eb2 mm, vmscan: preve... |
3507 |
trace_mm_vmscan_wakeup_kswapd(pgdat->node_id, classzone_idx, order); |
8d0986e28 [PATCH] vm: kswap... |
3508 |
wake_up_interruptible(&pgdat->kswapd_wait); |
1da177e4c Linux-2.6.12-rc2 |
3509 |
} |
c6f37f121 PM/Suspend: Do no... |
3510 |
#ifdef CONFIG_HIBERNATION |
1da177e4c Linux-2.6.12-rc2 |
3511 |
/* |
7b51755c3 vmscan: kill hibe... |
3512 |
* Try to free `nr_to_reclaim' of memory, system-wide, and return the number of |
d6277db4a [PATCH] swsusp: r... |
3513 3514 3515 3516 3517 |
* 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 |
3518 |
*/ |
7b51755c3 vmscan: kill hibe... |
3519 |
unsigned long shrink_all_memory(unsigned long nr_to_reclaim) |
1da177e4c Linux-2.6.12-rc2 |
3520 |
{ |
d6277db4a [PATCH] swsusp: r... |
3521 |
struct reclaim_state reclaim_state; |
d6277db4a [PATCH] swsusp: r... |
3522 |
struct scan_control sc = { |
ee814fe23 mm: vmscan: clean... |
3523 |
.nr_to_reclaim = nr_to_reclaim, |
7b51755c3 vmscan: kill hibe... |
3524 |
.gfp_mask = GFP_HIGHUSER_MOVABLE, |
b2e18757f mm, vmscan: begin... |
3525 |
.reclaim_idx = MAX_NR_ZONES - 1, |
ee814fe23 mm: vmscan: clean... |
3526 |
.priority = DEF_PRIORITY, |
d6277db4a [PATCH] swsusp: r... |
3527 |
.may_writepage = 1, |
ee814fe23 mm: vmscan: clean... |
3528 3529 |
.may_unmap = 1, .may_swap = 1, |
7b51755c3 vmscan: kill hibe... |
3530 |
.hibernation_mode = 1, |
1da177e4c Linux-2.6.12-rc2 |
3531 |
}; |
a09ed5e00 vmscan: change sh... |
3532 |
struct zonelist *zonelist = node_zonelist(numa_node_id(), sc.gfp_mask); |
7b51755c3 vmscan: kill hibe... |
3533 3534 |
struct task_struct *p = current; unsigned long nr_reclaimed; |
499118e96 mm: introduce mem... |
3535 |
unsigned int noreclaim_flag; |
1da177e4c Linux-2.6.12-rc2 |
3536 |
|
499118e96 mm: introduce mem... |
3537 |
noreclaim_flag = memalloc_noreclaim_save(); |
d92a8cfcb locking/lockdep: ... |
3538 |
fs_reclaim_acquire(sc.gfp_mask); |
7b51755c3 vmscan: kill hibe... |
3539 3540 |
reclaim_state.reclaimed_slab = 0; p->reclaim_state = &reclaim_state; |
d6277db4a [PATCH] swsusp: r... |
3541 |
|
3115cd914 mm: vmscan: remov... |
3542 |
nr_reclaimed = do_try_to_free_pages(zonelist, &sc); |
d979677c4 mm: shrink_all_me... |
3543 |
|
7b51755c3 vmscan: kill hibe... |
3544 |
p->reclaim_state = NULL; |
d92a8cfcb locking/lockdep: ... |
3545 |
fs_reclaim_release(sc.gfp_mask); |
499118e96 mm: introduce mem... |
3546 |
memalloc_noreclaim_restore(noreclaim_flag); |
d6277db4a [PATCH] swsusp: r... |
3547 |
|
7b51755c3 vmscan: kill hibe... |
3548 |
return nr_reclaimed; |
1da177e4c Linux-2.6.12-rc2 |
3549 |
} |
c6f37f121 PM/Suspend: Do no... |
3550 |
#endif /* CONFIG_HIBERNATION */ |
1da177e4c Linux-2.6.12-rc2 |
3551 |
|
1da177e4c Linux-2.6.12-rc2 |
3552 3553 3554 3555 |
/* 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. */ |
517bbed90 mm/vmscan: Conver... |
3556 |
static int kswapd_cpu_online(unsigned int cpu) |
1da177e4c Linux-2.6.12-rc2 |
3557 |
{ |
58c0a4a78 Fix panic of cpu ... |
3558 |
int nid; |
1da177e4c Linux-2.6.12-rc2 |
3559 |
|
517bbed90 mm/vmscan: Conver... |
3560 3561 3562 |
for_each_node_state(nid, N_MEMORY) { pg_data_t *pgdat = NODE_DATA(nid); const struct cpumask *mask; |
a70f73028 cpumask: replace ... |
3563 |
|
517bbed90 mm/vmscan: Conver... |
3564 |
mask = cpumask_of_node(pgdat->node_id); |
c5f59f083 nodemask: use new... |
3565 |
|
517bbed90 mm/vmscan: Conver... |
3566 3567 3568 |
if (cpumask_any_and(cpu_online_mask, mask) < nr_cpu_ids) /* One of our CPUs online: restore mask */ set_cpus_allowed_ptr(pgdat->kswapd, mask); |
1da177e4c Linux-2.6.12-rc2 |
3569 |
} |
517bbed90 mm/vmscan: Conver... |
3570 |
return 0; |
1da177e4c Linux-2.6.12-rc2 |
3571 |
} |
1da177e4c Linux-2.6.12-rc2 |
3572 |
|
3218ae14b [PATCH] pgdat all... |
3573 3574 3575 3576 3577 3578 3579 3580 3581 3582 3583 3584 3585 3586 3587 |
/* * 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 */ |
c6202adf3 mm/vmscan: Adjust... |
3588 |
BUG_ON(system_state < SYSTEM_RUNNING); |
d5dc0ad92 mm/vmscan: fix er... |
3589 3590 3591 |
pr_err("Failed to start kswapd on node %d ", nid); ret = PTR_ERR(pgdat->kswapd); |
d72515b85 mm/vmscan: fix er... |
3592 |
pgdat->kswapd = NULL; |
3218ae14b [PATCH] pgdat all... |
3593 3594 3595 |
} return ret; } |
8fe23e057 mm: clear node in... |
3596 |
/* |
d8adde17e memory hotplug: f... |
3597 |
* Called by memory hotplug when all memory in a node is offlined. Caller must |
bfc8c9013 mem-hotplug: impl... |
3598 |
* hold mem_hotplug_begin/end(). |
8fe23e057 mm: clear node in... |
3599 3600 3601 3602 |
*/ void kswapd_stop(int nid) { struct task_struct *kswapd = NODE_DATA(nid)->kswapd; |
d8adde17e memory hotplug: f... |
3603 |
if (kswapd) { |
8fe23e057 mm: clear node in... |
3604 |
kthread_stop(kswapd); |
d8adde17e memory hotplug: f... |
3605 3606 |
NODE_DATA(nid)->kswapd = NULL; } |
8fe23e057 mm: clear node in... |
3607 |
} |
1da177e4c Linux-2.6.12-rc2 |
3608 3609 |
static int __init kswapd_init(void) { |
517bbed90 mm/vmscan: Conver... |
3610 |
int nid, ret; |
69e05944a [PATCH] vmscan: u... |
3611 |
|
1da177e4c Linux-2.6.12-rc2 |
3612 |
swap_setup(); |
48fb2e240 vmscan: use N_MEM... |
3613 |
for_each_node_state(nid, N_MEMORY) |
3218ae14b [PATCH] pgdat all... |
3614 |
kswapd_run(nid); |
517bbed90 mm/vmscan: Conver... |
3615 3616 3617 3618 |
ret = cpuhp_setup_state_nocalls(CPUHP_AP_ONLINE_DYN, "mm/vmscan:online", kswapd_cpu_online, NULL); WARN_ON(ret < 0); |
1da177e4c Linux-2.6.12-rc2 |
3619 3620 3621 3622 |
return 0; } module_init(kswapd_init) |
9eeff2395 [PATCH] Zone recl... |
3623 3624 3625 |
#ifdef CONFIG_NUMA /* |
a5f5f91da mm: convert zone_... |
3626 |
* Node reclaim mode |
9eeff2395 [PATCH] Zone recl... |
3627 |
* |
a5f5f91da mm: convert zone_... |
3628 |
* If non-zero call node_reclaim when the number of free pages falls below |
9eeff2395 [PATCH] Zone recl... |
3629 |
* the watermarks. |
9eeff2395 [PATCH] Zone recl... |
3630 |
*/ |
a5f5f91da mm: convert zone_... |
3631 |
int node_reclaim_mode __read_mostly; |
9eeff2395 [PATCH] Zone recl... |
3632 |
|
1b2ffb789 [PATCH] Zone recl... |
3633 |
#define RECLAIM_OFF 0 |
7d03431cf swapfile/vmscan: ... |
3634 |
#define RECLAIM_ZONE (1<<0) /* Run shrink_inactive_list on the zone */ |
1b2ffb789 [PATCH] Zone recl... |
3635 |
#define RECLAIM_WRITE (1<<1) /* Writeout pages during reclaim */ |
95bbc0c72 mm: rename RECLAI... |
3636 |
#define RECLAIM_UNMAP (1<<2) /* Unmap pages during reclaim */ |
1b2ffb789 [PATCH] Zone recl... |
3637 |
|
9eeff2395 [PATCH] Zone recl... |
3638 |
/* |
a5f5f91da mm: convert zone_... |
3639 |
* Priority for NODE_RECLAIM. This determines the fraction of pages |
a92f71263 [PATCH] zone_recl... |
3640 3641 3642 |
* of a node considered for each zone_reclaim. 4 scans 1/16th of * a zone. */ |
a5f5f91da mm: convert zone_... |
3643 |
#define NODE_RECLAIM_PRIORITY 4 |
a92f71263 [PATCH] zone_recl... |
3644 |
|
9eeff2395 [PATCH] Zone recl... |
3645 |
/* |
a5f5f91da mm: convert zone_... |
3646 |
* Percentage of pages in a zone that must be unmapped for node_reclaim to |
9614634fe [PATCH] ZVC/zone_... |
3647 3648 3649 3650 3651 |
* occur. */ int sysctl_min_unmapped_ratio = 1; /* |
0ff38490c [PATCH] zone_recl... |
3652 3653 3654 3655 |
* 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; |
11fb99898 mm: move most fil... |
3656 |
static inline unsigned long node_unmapped_file_pages(struct pglist_data *pgdat) |
90afa5de6 vmscan: properly ... |
3657 |
{ |
11fb99898 mm: move most fil... |
3658 3659 3660 |
unsigned long file_mapped = node_page_state(pgdat, NR_FILE_MAPPED); unsigned long file_lru = node_page_state(pgdat, NR_INACTIVE_FILE) + node_page_state(pgdat, NR_ACTIVE_FILE); |
90afa5de6 vmscan: properly ... |
3661 3662 3663 3664 3665 3666 3667 3668 3669 3670 |
/* * 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 */ |
a5f5f91da mm: convert zone_... |
3671 |
static unsigned long node_pagecache_reclaimable(struct pglist_data *pgdat) |
90afa5de6 vmscan: properly ... |
3672 |
{ |
d031a1579 mm/vmscan.c: fix ... |
3673 3674 |
unsigned long nr_pagecache_reclaimable; unsigned long delta = 0; |
90afa5de6 vmscan: properly ... |
3675 3676 |
/* |
95bbc0c72 mm: rename RECLAI... |
3677 |
* If RECLAIM_UNMAP is set, then all file pages are considered |
90afa5de6 vmscan: properly ... |
3678 |
* potentially reclaimable. Otherwise, we have to worry about |
11fb99898 mm: move most fil... |
3679 |
* pages like swapcache and node_unmapped_file_pages() provides |
90afa5de6 vmscan: properly ... |
3680 3681 |
* a better estimate */ |
a5f5f91da mm: convert zone_... |
3682 3683 |
if (node_reclaim_mode & RECLAIM_UNMAP) nr_pagecache_reclaimable = node_page_state(pgdat, NR_FILE_PAGES); |
90afa5de6 vmscan: properly ... |
3684 |
else |
a5f5f91da mm: convert zone_... |
3685 |
nr_pagecache_reclaimable = node_unmapped_file_pages(pgdat); |
90afa5de6 vmscan: properly ... |
3686 3687 |
/* If we can't clean pages, remove dirty pages from consideration */ |
a5f5f91da mm: convert zone_... |
3688 3689 |
if (!(node_reclaim_mode & RECLAIM_WRITE)) delta += node_page_state(pgdat, NR_FILE_DIRTY); |
90afa5de6 vmscan: properly ... |
3690 3691 3692 3693 3694 3695 3696 |
/* 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... |
3697 |
/* |
a5f5f91da mm: convert zone_... |
3698 |
* Try to free up some pages from this node through reclaim. |
9eeff2395 [PATCH] Zone recl... |
3699 |
*/ |
a5f5f91da mm: convert zone_... |
3700 |
static int __node_reclaim(struct pglist_data *pgdat, gfp_t gfp_mask, unsigned int order) |
9eeff2395 [PATCH] Zone recl... |
3701 |
{ |
7fb2d46d3 [PATCH] zone_recl... |
3702 |
/* Minimum pages needed in order to stay on node */ |
69e05944a [PATCH] vmscan: u... |
3703 |
const unsigned long nr_pages = 1 << order; |
9eeff2395 [PATCH] Zone recl... |
3704 3705 |
struct task_struct *p = current; struct reclaim_state reclaim_state; |
499118e96 mm: introduce mem... |
3706 |
unsigned int noreclaim_flag; |
179e96395 [PATCH] vmscan: s... |
3707 |
struct scan_control sc = { |
62b726c1b mm/vmscan.c:__zon... |
3708 |
.nr_to_reclaim = max(nr_pages, SWAP_CLUSTER_MAX), |
f2f43e566 mm/vmscan.c: fix ... |
3709 |
.gfp_mask = current_gfp_context(gfp_mask), |
bd2f6199c vmscan: respect h... |
3710 |
.order = order, |
a5f5f91da mm: convert zone_... |
3711 3712 3713 |
.priority = NODE_RECLAIM_PRIORITY, .may_writepage = !!(node_reclaim_mode & RECLAIM_WRITE), .may_unmap = !!(node_reclaim_mode & RECLAIM_UNMAP), |
ee814fe23 mm: vmscan: clean... |
3714 |
.may_swap = 1, |
f2f43e566 mm/vmscan.c: fix ... |
3715 |
.reclaim_idx = gfp_zone(gfp_mask), |
179e96395 [PATCH] vmscan: s... |
3716 |
}; |
9eeff2395 [PATCH] Zone recl... |
3717 |
|
9eeff2395 [PATCH] Zone recl... |
3718 |
cond_resched(); |
d4f7796e9 [PATCH] vmscan: f... |
3719 |
/* |
95bbc0c72 mm: rename RECLAI... |
3720 |
* We need to be able to allocate from the reserves for RECLAIM_UNMAP |
d4f7796e9 [PATCH] vmscan: f... |
3721 |
* and we also need to be able to write out pages for RECLAIM_WRITE |
95bbc0c72 mm: rename RECLAI... |
3722 |
* and RECLAIM_UNMAP. |
d4f7796e9 [PATCH] vmscan: f... |
3723 |
*/ |
499118e96 mm: introduce mem... |
3724 3725 |
noreclaim_flag = memalloc_noreclaim_save(); p->flags |= PF_SWAPWRITE; |
d92a8cfcb locking/lockdep: ... |
3726 |
fs_reclaim_acquire(sc.gfp_mask); |
9eeff2395 [PATCH] Zone recl... |
3727 3728 |
reclaim_state.reclaimed_slab = 0; p->reclaim_state = &reclaim_state; |
c84db23c6 [PATCH] zone_recl... |
3729 |
|
a5f5f91da mm: convert zone_... |
3730 |
if (node_pagecache_reclaimable(pgdat) > pgdat->min_unmapped_pages) { |
0ff38490c [PATCH] zone_recl... |
3731 3732 3733 3734 |
/* * Free memory by calling shrink zone with increasing * priorities until we have enough memory freed. */ |
0ff38490c [PATCH] zone_recl... |
3735 |
do { |
970a39a36 mm, vmscan: avoid... |
3736 |
shrink_node(pgdat, &sc); |
9e3b2f8cd mm/vmscan: store ... |
3737 |
} while (sc.nr_reclaimed < nr_pages && --sc.priority >= 0); |
0ff38490c [PATCH] zone_recl... |
3738 |
} |
c84db23c6 [PATCH] zone_recl... |
3739 |
|
9eeff2395 [PATCH] Zone recl... |
3740 |
p->reclaim_state = NULL; |
d92a8cfcb locking/lockdep: ... |
3741 |
fs_reclaim_release(gfp_mask); |
499118e96 mm: introduce mem... |
3742 3743 |
current->flags &= ~PF_SWAPWRITE; memalloc_noreclaim_restore(noreclaim_flag); |
a79311c14 vmscan: bail out ... |
3744 |
return sc.nr_reclaimed >= nr_pages; |
9eeff2395 [PATCH] Zone recl... |
3745 |
} |
179e96395 [PATCH] vmscan: s... |
3746 |
|
a5f5f91da mm: convert zone_... |
3747 |
int node_reclaim(struct pglist_data *pgdat, gfp_t gfp_mask, unsigned int order) |
179e96395 [PATCH] vmscan: s... |
3748 |
{ |
d773ed6b8 mm: test and set ... |
3749 |
int ret; |
179e96395 [PATCH] vmscan: s... |
3750 3751 |
/* |
a5f5f91da mm: convert zone_... |
3752 |
* Node reclaim reclaims unmapped file backed pages and |
0ff38490c [PATCH] zone_recl... |
3753 |
* slab pages if we are over the defined limits. |
34aa1330f [PATCH] zoned vm ... |
3754 |
* |
9614634fe [PATCH] ZVC/zone_... |
3755 3756 |
* A small portion of unmapped file backed pages is needed for * file I/O otherwise pages read by file I/O will be immediately |
a5f5f91da mm: convert zone_... |
3757 3758 |
* thrown out if the node is overallocated. So we do not reclaim * if less than a specified percentage of the node is used by |
9614634fe [PATCH] ZVC/zone_... |
3759 |
* unmapped file backed pages. |
179e96395 [PATCH] vmscan: s... |
3760 |
*/ |
a5f5f91da mm: convert zone_... |
3761 |
if (node_pagecache_reclaimable(pgdat) <= pgdat->min_unmapped_pages && |
385386cff mm: vmstat: move ... |
3762 |
node_page_state(pgdat, NR_SLAB_RECLAIMABLE) <= pgdat->min_slab_pages) |
a5f5f91da mm: convert zone_... |
3763 |
return NODE_RECLAIM_FULL; |
179e96395 [PATCH] vmscan: s... |
3764 3765 |
/* |
d773ed6b8 mm: test and set ... |
3766 |
* Do not scan if the allocation should not be delayed. |
179e96395 [PATCH] vmscan: s... |
3767 |
*/ |
d0164adc8 mm, page_alloc: d... |
3768 |
if (!gfpflags_allow_blocking(gfp_mask) || (current->flags & PF_MEMALLOC)) |
a5f5f91da mm: convert zone_... |
3769 |
return NODE_RECLAIM_NOSCAN; |
179e96395 [PATCH] vmscan: s... |
3770 3771 |
/* |
a5f5f91da mm: convert zone_... |
3772 |
* Only run node reclaim on the local node or on nodes that do not |
179e96395 [PATCH] vmscan: s... |
3773 3774 3775 3776 |
* have associated processors. This will favor the local processor * over remote processors and spread off node memory allocations * as wide as possible. */ |
a5f5f91da mm: convert zone_... |
3777 3778 |
if (node_state(pgdat->node_id, N_CPU) && pgdat->node_id != numa_node_id()) return NODE_RECLAIM_NOSCAN; |
d773ed6b8 mm: test and set ... |
3779 |
|
a5f5f91da mm: convert zone_... |
3780 3781 |
if (test_and_set_bit(PGDAT_RECLAIM_LOCKED, &pgdat->flags)) return NODE_RECLAIM_NOSCAN; |
fa5e084e4 vmscan: do not un... |
3782 |
|
a5f5f91da mm: convert zone_... |
3783 3784 |
ret = __node_reclaim(pgdat, gfp_mask, order); clear_bit(PGDAT_RECLAIM_LOCKED, &pgdat->flags); |
d773ed6b8 mm: test and set ... |
3785 |
|
24cf72518 vmscan: count the... |
3786 3787 |
if (!ret) count_vm_event(PGSCAN_ZONE_RECLAIM_FAILED); |
d773ed6b8 mm: test and set ... |
3788 |
return ret; |
179e96395 [PATCH] vmscan: s... |
3789 |
} |
9eeff2395 [PATCH] Zone recl... |
3790 |
#endif |
894bc3104 Unevictable LRU I... |
3791 |
|
894bc3104 Unevictable LRU I... |
3792 3793 3794 |
/* * page_evictable - test whether a page is evictable * @page: the page to test |
894bc3104 Unevictable LRU I... |
3795 3796 |
* * Test whether page is evictable--i.e., should be placed on active/inactive |
39b5f29ac mm: remove vma ar... |
3797 |
* lists vs unevictable list. |
894bc3104 Unevictable LRU I... |
3798 3799 |
* * Reasons page might not be evictable: |
ba9ddf493 Ramfs and Ram Dis... |
3800 |
* (1) page's mapping marked unevictable |
b291f0003 mlock: mlocked pa... |
3801 |
* (2) page is part of an mlocked VMA |
ba9ddf493 Ramfs and Ram Dis... |
3802 |
* |
894bc3104 Unevictable LRU I... |
3803 |
*/ |
39b5f29ac mm: remove vma ar... |
3804 |
int page_evictable(struct page *page) |
894bc3104 Unevictable LRU I... |
3805 |
{ |
8d700626f mm: fix races bet... |
3806 3807 3808 3809 3810 3811 3812 |
int ret; /* Prevent address_space of inode and swap cache from being freed */ rcu_read_lock(); ret = !mapping_unevictable(page_mapping(page)) && !PageMlocked(page); rcu_read_unlock(); return ret; |
894bc3104 Unevictable LRU I... |
3813 |
} |
89e004ea5 SHM_LOCKED pages ... |
3814 |
|
85046579b SHM_UNLOCK: fix l... |
3815 |
#ifdef CONFIG_SHMEM |
89e004ea5 SHM_LOCKED pages ... |
3816 |
/** |
245132643 SHM_UNLOCK: fix U... |
3817 3818 3819 |
* check_move_unevictable_pages - check pages for evictability and move to appropriate zone lru list * @pages: array of pages to check * @nr_pages: number of pages to check |
89e004ea5 SHM_LOCKED pages ... |
3820 |
* |
245132643 SHM_UNLOCK: fix U... |
3821 |
* Checks pages for evictability and moves them to the appropriate lru list. |
85046579b SHM_UNLOCK: fix l... |
3822 3823 |
* * This function is only used for SysV IPC SHM_UNLOCK. |
89e004ea5 SHM_LOCKED pages ... |
3824 |
*/ |
245132643 SHM_UNLOCK: fix U... |
3825 |
void check_move_unevictable_pages(struct page **pages, int nr_pages) |
89e004ea5 SHM_LOCKED pages ... |
3826 |
{ |
925b7673c mm: make per-memc... |
3827 |
struct lruvec *lruvec; |
785b99feb mm, vmscan: relea... |
3828 |
struct pglist_data *pgdat = NULL; |
245132643 SHM_UNLOCK: fix U... |
3829 3830 3831 |
int pgscanned = 0; int pgrescued = 0; int i; |
89e004ea5 SHM_LOCKED pages ... |
3832 |
|
245132643 SHM_UNLOCK: fix U... |
3833 3834 |
for (i = 0; i < nr_pages; i++) { struct page *page = pages[i]; |
785b99feb mm, vmscan: relea... |
3835 |
struct pglist_data *pagepgdat = page_pgdat(page); |
89e004ea5 SHM_LOCKED pages ... |
3836 |
|
245132643 SHM_UNLOCK: fix U... |
3837 |
pgscanned++; |
785b99feb mm, vmscan: relea... |
3838 3839 3840 3841 3842 |
if (pagepgdat != pgdat) { if (pgdat) spin_unlock_irq(&pgdat->lru_lock); pgdat = pagepgdat; spin_lock_irq(&pgdat->lru_lock); |
245132643 SHM_UNLOCK: fix U... |
3843 |
} |
785b99feb mm, vmscan: relea... |
3844 |
lruvec = mem_cgroup_page_lruvec(page, pgdat); |
89e004ea5 SHM_LOCKED pages ... |
3845 |
|
245132643 SHM_UNLOCK: fix U... |
3846 3847 |
if (!PageLRU(page) || !PageUnevictable(page)) continue; |
89e004ea5 SHM_LOCKED pages ... |
3848 |
|
39b5f29ac mm: remove vma ar... |
3849 |
if (page_evictable(page)) { |
245132643 SHM_UNLOCK: fix U... |
3850 |
enum lru_list lru = page_lru_base_type(page); |
309381fea mm: dump page whe... |
3851 |
VM_BUG_ON_PAGE(PageActive(page), page); |
245132643 SHM_UNLOCK: fix U... |
3852 |
ClearPageUnevictable(page); |
fa9add641 mm/memcg: apply a... |
3853 3854 |
del_page_from_lru_list(page, lruvec, LRU_UNEVICTABLE); add_page_to_lru_list(page, lruvec, lru); |
245132643 SHM_UNLOCK: fix U... |
3855 |
pgrescued++; |
89e004ea5 SHM_LOCKED pages ... |
3856 |
} |
245132643 SHM_UNLOCK: fix U... |
3857 |
} |
89e004ea5 SHM_LOCKED pages ... |
3858 |
|
785b99feb mm, vmscan: relea... |
3859 |
if (pgdat) { |
245132643 SHM_UNLOCK: fix U... |
3860 3861 |
__count_vm_events(UNEVICTABLE_PGRESCUED, pgrescued); __count_vm_events(UNEVICTABLE_PGSCANNED, pgscanned); |
785b99feb mm, vmscan: relea... |
3862 |
spin_unlock_irq(&pgdat->lru_lock); |
89e004ea5 SHM_LOCKED pages ... |
3863 |
} |
89e004ea5 SHM_LOCKED pages ... |
3864 |
} |
85046579b SHM_UNLOCK: fix l... |
3865 |
#endif /* CONFIG_SHMEM */ |