Blame view
mm/compaction.c
83.6 KB
b24413180
|
1 |
// SPDX-License-Identifier: GPL-2.0 |
748446bb6
|
2 3 4 5 6 7 8 9 10 |
/* * linux/mm/compaction.c * * Memory compaction for the reduction of external fragmentation. Note that * this heavily depends upon page migration to do all the real heavy * lifting * * Copyright IBM Corp. 2007-2010 Mel Gorman <mel@csn.ul.ie> */ |
698b1b306
|
11 |
#include <linux/cpu.h> |
748446bb6
|
12 13 14 15 |
#include <linux/swap.h> #include <linux/migrate.h> #include <linux/compaction.h> #include <linux/mm_inline.h> |
174cd4b1e
|
16 |
#include <linux/sched/signal.h> |
748446bb6
|
17 |
#include <linux/backing-dev.h> |
76ab0f530
|
18 |
#include <linux/sysctl.h> |
ed4a6d7f0
|
19 |
#include <linux/sysfs.h> |
194159fbc
|
20 |
#include <linux/page-isolation.h> |
b8c73fc24
|
21 |
#include <linux/kasan.h> |
698b1b306
|
22 23 |
#include <linux/kthread.h> #include <linux/freezer.h> |
83358ece2
|
24 |
#include <linux/page_owner.h> |
eb414681d
|
25 |
#include <linux/psi.h> |
748446bb6
|
26 |
#include "internal.h" |
010fc29a4
|
27 28 29 30 31 32 33 34 35 36 37 38 39 40 |
#ifdef CONFIG_COMPACTION static inline void count_compact_event(enum vm_event_item item) { count_vm_event(item); } static inline void count_compact_events(enum vm_event_item item, long delta) { count_vm_events(item, delta); } #else #define count_compact_event(item) do { } while (0) #define count_compact_events(item, delta) do { } while (0) #endif |
ff9543fd3
|
41 |
#if defined CONFIG_COMPACTION || defined CONFIG_CMA |
b7aba6984
|
42 43 |
#define CREATE_TRACE_POINTS #include <trace/events/compaction.h> |
06b6640a3
|
44 45 46 47 |
#define block_start_pfn(pfn, order) round_down(pfn, 1UL << (order)) #define block_end_pfn(pfn, order) ALIGN((pfn) + 1, 1UL << (order)) #define pageblock_start_pfn(pfn) block_start_pfn(pfn, pageblock_order) #define pageblock_end_pfn(pfn) block_end_pfn(pfn, pageblock_order) |
facdaa917
|
48 49 50 |
/* * Fragmentation score check interval for proactive compaction purposes. */ |
d34c0a759
|
51 |
static const unsigned int HPAGE_FRAG_CHECK_INTERVAL_MSEC = 500; |
facdaa917
|
52 53 54 55 56 57 58 59 |
/* * Page order with-respect-to which proactive compaction * calculates external fragmentation, which is used as * the "fragmentation score" of a node/zone. */ #if defined CONFIG_TRANSPARENT_HUGEPAGE #define COMPACTION_HPAGE_ORDER HPAGE_PMD_ORDER |
25788738e
|
60 |
#elif defined CONFIG_HUGETLBFS |
facdaa917
|
61 62 63 64 |
#define COMPACTION_HPAGE_ORDER HUGETLB_PAGE_ORDER #else #define COMPACTION_HPAGE_ORDER (PMD_SHIFT - PAGE_SHIFT) #endif |
748446bb6
|
65 66 67 |
static unsigned long release_freepages(struct list_head *freelist) { struct page *page, *next; |
6bace090a
|
68 |
unsigned long high_pfn = 0; |
748446bb6
|
69 70 |
list_for_each_entry_safe(page, next, freelist, lru) { |
6bace090a
|
71 |
unsigned long pfn = page_to_pfn(page); |
748446bb6
|
72 73 |
list_del(&page->lru); __free_page(page); |
6bace090a
|
74 75 |
if (pfn > high_pfn) high_pfn = pfn; |
748446bb6
|
76 |
} |
6bace090a
|
77 |
return high_pfn; |
748446bb6
|
78 |
} |
4469ab984
|
79 |
static void split_map_pages(struct list_head *list) |
ff9543fd3
|
80 |
{ |
66c64223a
|
81 82 83 84 85 86 87 88 89 |
unsigned int i, order, nr_pages; struct page *page, *next; LIST_HEAD(tmp_list); list_for_each_entry_safe(page, next, list, lru) { list_del(&page->lru); order = page_private(page); nr_pages = 1 << order; |
66c64223a
|
90 |
|
46f24fd85
|
91 |
post_alloc_hook(page, order, __GFP_MOVABLE); |
66c64223a
|
92 93 |
if (order) split_page(page, order); |
ff9543fd3
|
94 |
|
66c64223a
|
95 96 97 98 |
for (i = 0; i < nr_pages; i++) { list_add(&page->lru, &tmp_list); page++; } |
ff9543fd3
|
99 |
} |
66c64223a
|
100 101 |
list_splice(&tmp_list, list); |
ff9543fd3
|
102 |
} |
bb13ffeb9
|
103 |
#ifdef CONFIG_COMPACTION |
24e2716f6
|
104 |
|
bda807d44
|
105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 |
int PageMovable(struct page *page) { struct address_space *mapping; VM_BUG_ON_PAGE(!PageLocked(page), page); if (!__PageMovable(page)) return 0; mapping = page_mapping(page); if (mapping && mapping->a_ops && mapping->a_ops->isolate_page) return 1; return 0; } EXPORT_SYMBOL(PageMovable); void __SetPageMovable(struct page *page, struct address_space *mapping) { VM_BUG_ON_PAGE(!PageLocked(page), page); VM_BUG_ON_PAGE((unsigned long)mapping & PAGE_MAPPING_MOVABLE, page); page->mapping = (void *)((unsigned long)mapping | PAGE_MAPPING_MOVABLE); } EXPORT_SYMBOL(__SetPageMovable); void __ClearPageMovable(struct page *page) { |
bda807d44
|
131 132 133 134 135 136 137 138 139 140 |
VM_BUG_ON_PAGE(!PageMovable(page), page); /* * Clear registered address_space val with keeping PAGE_MAPPING_MOVABLE * flag so that VM can catch up released page by driver after isolation. * With it, VM migration doesn't try to put it back. */ page->mapping = (void *)((unsigned long)page->mapping & PAGE_MAPPING_MOVABLE); } EXPORT_SYMBOL(__ClearPageMovable); |
24e2716f6
|
141 142 143 144 145 |
/* Do not skip compaction more than 64 times */ #define COMPACT_MAX_DEFER_SHIFT 6 /* * Compaction is deferred when compaction fails to result in a page |
860b32729
|
146 |
* allocation success. 1 << compact_defer_shift, compactions are skipped up |
24e2716f6
|
147 148 |
* to a limit of 1 << COMPACT_MAX_DEFER_SHIFT */ |
2271b016b
|
149 |
static void defer_compaction(struct zone *zone, int order) |
24e2716f6
|
150 151 152 153 154 155 156 157 158 159 160 161 162 163 |
{ zone->compact_considered = 0; zone->compact_defer_shift++; if (order < zone->compact_order_failed) zone->compact_order_failed = order; if (zone->compact_defer_shift > COMPACT_MAX_DEFER_SHIFT) zone->compact_defer_shift = COMPACT_MAX_DEFER_SHIFT; trace_mm_compaction_defer_compaction(zone, order); } /* Returns true if compaction should be skipped this time */ |
2271b016b
|
164 |
static bool compaction_deferred(struct zone *zone, int order) |
24e2716f6
|
165 166 167 168 169 170 171 |
{ unsigned long defer_limit = 1UL << zone->compact_defer_shift; if (order < zone->compact_order_failed) return false; /* Avoid possible overflow */ |
62b35fe0e
|
172 |
if (++zone->compact_considered >= defer_limit) { |
24e2716f6
|
173 |
zone->compact_considered = defer_limit; |
24e2716f6
|
174 |
return false; |
62b35fe0e
|
175 |
} |
24e2716f6
|
176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 |
trace_mm_compaction_deferred(zone, order); return true; } /* * Update defer tracking counters after successful compaction of given order, * which means an allocation either succeeded (alloc_success == true) or is * expected to succeed. */ void compaction_defer_reset(struct zone *zone, int order, bool alloc_success) { if (alloc_success) { zone->compact_considered = 0; zone->compact_defer_shift = 0; } if (order >= zone->compact_order_failed) zone->compact_order_failed = order + 1; trace_mm_compaction_defer_reset(zone, order); } /* Returns true if restarting compaction after many failures */ |
2271b016b
|
201 |
static bool compaction_restarting(struct zone *zone, int order) |
24e2716f6
|
202 203 204 205 206 207 208 |
{ if (order < zone->compact_order_failed) return false; return zone->compact_defer_shift == COMPACT_MAX_DEFER_SHIFT && zone->compact_considered >= 1UL << zone->compact_defer_shift; } |
bb13ffeb9
|
209 210 211 212 213 214 215 216 217 |
/* Returns true if the pageblock should be scanned for pages to isolate. */ static inline bool isolation_suitable(struct compact_control *cc, struct page *page) { if (cc->ignore_skip_hint) return true; return !get_pageblock_skip(page); } |
02333641e
|
218 219 220 221 |
static void reset_cached_positions(struct zone *zone) { zone->compact_cached_migrate_pfn[0] = zone->zone_start_pfn; zone->compact_cached_migrate_pfn[1] = zone->zone_start_pfn; |
623446e4d
|
222 |
zone->compact_cached_free_pfn = |
06b6640a3
|
223 |
pageblock_start_pfn(zone_end_pfn(zone) - 1); |
02333641e
|
224 |
} |
bb13ffeb9
|
225 |
/* |
2271b016b
|
226 |
* Compound pages of >= pageblock_order should consistently be skipped until |
b527cfe5b
|
227 228 |
* released. It is always pointless to compact pages of such order (if they are * migratable), and the pageblocks they occupy cannot contain any free pages. |
21dc7e023
|
229 |
*/ |
b527cfe5b
|
230 |
static bool pageblock_skip_persistent(struct page *page) |
21dc7e023
|
231 |
{ |
b527cfe5b
|
232 |
if (!PageCompound(page)) |
21dc7e023
|
233 |
return false; |
b527cfe5b
|
234 235 236 237 238 239 240 |
page = compound_head(page); if (compound_order(page) >= pageblock_order) return true; return false; |
21dc7e023
|
241 |
} |
e332f741a
|
242 243 244 245 246 |
static bool __reset_isolation_pfn(struct zone *zone, unsigned long pfn, bool check_source, bool check_target) { struct page *page = pfn_to_online_page(pfn); |
6b0868c82
|
247 |
struct page *block_page; |
e332f741a
|
248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 |
struct page *end_page; unsigned long block_pfn; if (!page) return false; if (zone != page_zone(page)) return false; if (pageblock_skip_persistent(page)) return false; /* * If skip is already cleared do no further checking once the * restart points have been set. */ if (check_source && check_target && !get_pageblock_skip(page)) return true; /* * If clearing skip for the target scanner, do not select a * non-movable pageblock as the starting point. */ if (!check_source && check_target && get_pageblock_migratetype(page) != MIGRATE_MOVABLE) return false; |
6b0868c82
|
272 273 |
/* Ensure the start of the pageblock or zone is online and valid */ block_pfn = pageblock_start_pfn(pfn); |
a2e9a5afc
|
274 275 |
block_pfn = max(block_pfn, zone->zone_start_pfn); block_page = pfn_to_online_page(block_pfn); |
6b0868c82
|
276 277 278 279 280 281 |
if (block_page) { page = block_page; pfn = block_pfn; } /* Ensure the end of the pageblock or zone is online and valid */ |
a2e9a5afc
|
282 |
block_pfn = pageblock_end_pfn(pfn) - 1; |
6b0868c82
|
283 284 285 286 |
block_pfn = min(block_pfn, zone_end_pfn(zone) - 1); end_page = pfn_to_online_page(block_pfn); if (!end_page) return false; |
e332f741a
|
287 288 289 290 291 |
/* * Only clear the hint if a sample indicates there is either a * free page or an LRU page in the block. One or other condition * is necessary for the block to be a migration source/target. */ |
e332f741a
|
292 |
do { |
859a85ddf
|
293 294 295 296 |
if (check_source && PageLRU(page)) { clear_pageblock_skip(page); return true; } |
e332f741a
|
297 |
|
859a85ddf
|
298 299 300 |
if (check_target && PageBuddy(page)) { clear_pageblock_skip(page); return true; |
e332f741a
|
301 302 303 304 |
} page += (1 << PAGE_ALLOC_COSTLY_ORDER); pfn += (1 << PAGE_ALLOC_COSTLY_ORDER); |
a2e9a5afc
|
305 |
} while (page <= end_page); |
e332f741a
|
306 307 308 |
return false; } |
21dc7e023
|
309 |
/* |
bb13ffeb9
|
310 311 312 313 |
* This function is called to clear all cached information on pageblocks that * should be skipped for page isolation when the migrate and free page scanner * meet. */ |
62997027c
|
314 |
static void __reset_isolation_suitable(struct zone *zone) |
bb13ffeb9
|
315 |
{ |
e332f741a
|
316 |
unsigned long migrate_pfn = zone->zone_start_pfn; |
6b0868c82
|
317 |
unsigned long free_pfn = zone_end_pfn(zone) - 1; |
e332f741a
|
318 319 320 321 322 323 324 |
unsigned long reset_migrate = free_pfn; unsigned long reset_free = migrate_pfn; bool source_set = false; bool free_set = false; if (!zone->compact_blockskip_flush) return; |
bb13ffeb9
|
325 |
|
62997027c
|
326 |
zone->compact_blockskip_flush = false; |
bb13ffeb9
|
327 |
|
e332f741a
|
328 329 330 331 332 333 334 335 |
/* * Walk the zone and update pageblock skip information. Source looks * for PageLRU while target looks for PageBuddy. When the scanner * is found, both PageBuddy and PageLRU are checked as the pageblock * is suitable as both source and target. */ for (; migrate_pfn < free_pfn; migrate_pfn += pageblock_nr_pages, free_pfn -= pageblock_nr_pages) { |
bb13ffeb9
|
336 |
cond_resched(); |
e332f741a
|
337 338 339 340 341 342 343 344 345 |
/* Update the migrate PFN */ if (__reset_isolation_pfn(zone, migrate_pfn, true, source_set) && migrate_pfn < reset_migrate) { source_set = true; reset_migrate = migrate_pfn; zone->compact_init_migrate_pfn = reset_migrate; zone->compact_cached_migrate_pfn[0] = reset_migrate; zone->compact_cached_migrate_pfn[1] = reset_migrate; } |
bb13ffeb9
|
346 |
|
e332f741a
|
347 348 349 350 351 352 353 354 |
/* Update the free PFN */ if (__reset_isolation_pfn(zone, free_pfn, free_set, true) && free_pfn > reset_free) { free_set = true; reset_free = free_pfn; zone->compact_init_free_pfn = reset_free; zone->compact_cached_free_pfn = reset_free; } |
bb13ffeb9
|
355 |
} |
02333641e
|
356 |
|
e332f741a
|
357 358 359 360 361 362 |
/* Leave no distance if no suitable block was reset */ if (reset_migrate >= reset_free) { zone->compact_cached_migrate_pfn[0] = migrate_pfn; zone->compact_cached_migrate_pfn[1] = migrate_pfn; zone->compact_cached_free_pfn = free_pfn; } |
bb13ffeb9
|
363 |
} |
62997027c
|
364 365 366 367 368 369 370 371 372 373 374 375 376 377 |
void reset_isolation_suitable(pg_data_t *pgdat) { int zoneid; for (zoneid = 0; zoneid < MAX_NR_ZONES; zoneid++) { struct zone *zone = &pgdat->node_zones[zoneid]; if (!populated_zone(zone)) continue; /* Only flush if a full compaction finished recently */ if (zone->compact_blockskip_flush) __reset_isolation_suitable(zone); } } |
bb13ffeb9
|
378 |
/* |
e380bebe4
|
379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 |
* Sets the pageblock skip bit if it was clear. Note that this is a hint as * locks are not required for read/writers. Returns true if it was already set. */ static bool test_and_set_skip(struct compact_control *cc, struct page *page, unsigned long pfn) { bool skip; /* Do no update if skip hint is being ignored */ if (cc->ignore_skip_hint) return false; if (!IS_ALIGNED(pfn, pageblock_nr_pages)) return false; skip = get_pageblock_skip(page); if (!skip && !cc->no_set_skip_hint) set_pageblock_skip(page); return skip; } static void update_cached_migrate(struct compact_control *cc, unsigned long pfn) { struct zone *zone = cc->zone; pfn = pageblock_end_pfn(pfn); /* Set for isolation rather than compaction */ if (cc->no_set_skip_hint) return; if (pfn > zone->compact_cached_migrate_pfn[0]) zone->compact_cached_migrate_pfn[0] = pfn; if (cc->mode != MIGRATE_ASYNC && pfn > zone->compact_cached_migrate_pfn[1]) zone->compact_cached_migrate_pfn[1] = pfn; } /* |
bb13ffeb9
|
419 |
* If no pages were isolated then mark this pageblock to be skipped in the |
62997027c
|
420 |
* future. The information is later cleared by __reset_isolation_suitable(). |
bb13ffeb9
|
421 |
*/ |
c89511ab2
|
422 |
static void update_pageblock_skip(struct compact_control *cc, |
d097a6f63
|
423 |
struct page *page, unsigned long pfn) |
bb13ffeb9
|
424 |
{ |
c89511ab2
|
425 |
struct zone *zone = cc->zone; |
6815bf3f2
|
426 |
|
2583d6713
|
427 |
if (cc->no_set_skip_hint) |
6815bf3f2
|
428 |
return; |
bb13ffeb9
|
429 430 |
if (!page) return; |
edc2ca612
|
431 |
set_pageblock_skip(page); |
c89511ab2
|
432 |
|
35979ef33
|
433 |
/* Update where async and sync compaction should restart */ |
e380bebe4
|
434 435 |
if (pfn < zone->compact_cached_free_pfn) zone->compact_cached_free_pfn = pfn; |
bb13ffeb9
|
436 437 438 439 440 441 442 |
} #else static inline bool isolation_suitable(struct compact_control *cc, struct page *page) { return true; } |
b527cfe5b
|
443 |
static inline bool pageblock_skip_persistent(struct page *page) |
21dc7e023
|
444 445 446 447 448 |
{ return false; } static inline void update_pageblock_skip(struct compact_control *cc, |
d097a6f63
|
449 |
struct page *page, unsigned long pfn) |
bb13ffeb9
|
450 451 |
{ } |
e380bebe4
|
452 453 454 455 456 457 458 459 460 461 |
static void update_cached_migrate(struct compact_control *cc, unsigned long pfn) { } static bool test_and_set_skip(struct compact_control *cc, struct page *page, unsigned long pfn) { return false; } |
bb13ffeb9
|
462 |
#endif /* CONFIG_COMPACTION */ |
8b44d2791
|
463 464 |
/* * Compaction requires the taking of some coarse locks that are potentially |
cb2dcaf02
|
465 466 467 468 |
* very heavily contended. For async compaction, trylock and record if the * lock is contended. The lock will still be acquired but compaction will * abort when the current block is finished regardless of success rate. * Sync compaction acquires the lock. |
8b44d2791
|
469 |
* |
cb2dcaf02
|
470 |
* Always returns true which makes it easier to track lock state in callers. |
8b44d2791
|
471 |
*/ |
cb2dcaf02
|
472 |
static bool compact_lock_irqsave(spinlock_t *lock, unsigned long *flags, |
8b44d2791
|
473 |
struct compact_control *cc) |
77337edee
|
474 |
__acquires(lock) |
2a1402aa0
|
475 |
{ |
cb2dcaf02
|
476 477 478 479 480 481 |
/* Track if the lock is contended in async mode */ if (cc->mode == MIGRATE_ASYNC && !cc->contended) { if (spin_trylock_irqsave(lock, *flags)) return true; cc->contended = true; |
8b44d2791
|
482 |
} |
1f9efdef4
|
483 |
|
cb2dcaf02
|
484 |
spin_lock_irqsave(lock, *flags); |
8b44d2791
|
485 |
return true; |
2a1402aa0
|
486 |
} |
85aa125f0
|
487 |
/* |
c67fe3752
|
488 |
* Compaction requires the taking of some coarse locks that are potentially |
8b44d2791
|
489 490 491 492 493 494 495 |
* very heavily contended. The lock should be periodically unlocked to avoid * having disabled IRQs for a long time, even when there is nobody waiting on * the lock. It might also be that allowing the IRQs will result in * need_resched() becoming true. If scheduling is needed, async compaction * aborts. Sync compaction schedules. * Either compaction type will also abort if a fatal signal is pending. * In either case if the lock was locked, it is dropped and not regained. |
c67fe3752
|
496 |
* |
8b44d2791
|
497 498 499 500 |
* Returns true if compaction should abort due to fatal signal pending, or * async compaction due to need_resched() * Returns false when compaction can continue (sync compaction might have * scheduled) |
c67fe3752
|
501 |
*/ |
8b44d2791
|
502 503 |
static bool compact_unlock_should_abort(spinlock_t *lock, unsigned long flags, bool *locked, struct compact_control *cc) |
c67fe3752
|
504 |
{ |
8b44d2791
|
505 506 507 508 |
if (*locked) { spin_unlock_irqrestore(lock, flags); *locked = false; } |
1f9efdef4
|
509 |
|
8b44d2791
|
510 |
if (fatal_signal_pending(current)) { |
c3486f537
|
511 |
cc->contended = true; |
8b44d2791
|
512 513 |
return true; } |
c67fe3752
|
514 |
|
cf66f0700
|
515 |
cond_resched(); |
be9765722
|
516 517 518 |
return false; } |
c67fe3752
|
519 |
/* |
9e4be4708
|
520 521 522 |
* Isolate free pages onto a private freelist. If @strict is true, will abort * returning 0 on any invalid PFNs or non-free pages inside of the pageblock * (even though it may still end up isolating some pages). |
85aa125f0
|
523 |
*/ |
f40d1e42b
|
524 |
static unsigned long isolate_freepages_block(struct compact_control *cc, |
e14c720ef
|
525 |
unsigned long *start_pfn, |
85aa125f0
|
526 527 |
unsigned long end_pfn, struct list_head *freelist, |
4fca9730c
|
528 |
unsigned int stride, |
85aa125f0
|
529 |
bool strict) |
748446bb6
|
530 |
{ |
b7aba6984
|
531 |
int nr_scanned = 0, total_isolated = 0; |
d097a6f63
|
532 |
struct page *cursor; |
b8b2d8253
|
533 |
unsigned long flags = 0; |
f40d1e42b
|
534 |
bool locked = false; |
e14c720ef
|
535 |
unsigned long blockpfn = *start_pfn; |
66c64223a
|
536 |
unsigned int order; |
748446bb6
|
537 |
|
4fca9730c
|
538 539 540 |
/* Strict mode is for isolation, speed is secondary */ if (strict) stride = 1; |
748446bb6
|
541 |
cursor = pfn_to_page(blockpfn); |
f40d1e42b
|
542 |
/* Isolate free pages. */ |
4fca9730c
|
543 |
for (; blockpfn < end_pfn; blockpfn += stride, cursor += stride) { |
66c64223a
|
544 |
int isolated; |
748446bb6
|
545 |
struct page *page = cursor; |
8b44d2791
|
546 547 548 549 550 551 552 553 554 |
/* * Periodically drop the lock (if held) regardless of its * contention, to give chance to IRQs. Abort if fatal signal * pending or async compaction detects need_resched() */ if (!(blockpfn % SWAP_CLUSTER_MAX) && compact_unlock_should_abort(&cc->zone->lock, flags, &locked, cc)) break; |
b7aba6984
|
555 |
nr_scanned++; |
2af120bc0
|
556 |
|
9fcd6d2e0
|
557 558 559 560 561 562 563 |
/* * For compound pages such as THP and hugetlbfs, we can save * potentially a lot of iterations if we skip them at once. * The check is racy, but we can consider only valid values * and the only danger is skipping too much. */ if (PageCompound(page)) { |
21dc7e023
|
564 |
const unsigned int order = compound_order(page); |
d3c85bad8
|
565 |
if (likely(order < MAX_ORDER)) { |
21dc7e023
|
566 567 |
blockpfn += (1UL << order) - 1; cursor += (1UL << order) - 1; |
9fcd6d2e0
|
568 |
} |
9fcd6d2e0
|
569 570 |
goto isolate_fail; } |
f40d1e42b
|
571 |
if (!PageBuddy(page)) |
2af120bc0
|
572 |
goto isolate_fail; |
f40d1e42b
|
573 574 |
/* |
69b7189f1
|
575 576 577 578 579 |
* If we already hold the lock, we can skip some rechecking. * Note that if we hold the lock now, checked_pageblock was * already set in some previous iteration (or strict is true), * so it is correct to skip the suitable migration target * recheck as well. |
f40d1e42b
|
580 |
*/ |
69b7189f1
|
581 |
if (!locked) { |
cb2dcaf02
|
582 |
locked = compact_lock_irqsave(&cc->zone->lock, |
8b44d2791
|
583 |
&flags, cc); |
f40d1e42b
|
584 |
|
69b7189f1
|
585 586 587 588 |
/* Recheck this is a buddy page under lock */ if (!PageBuddy(page)) goto isolate_fail; } |
748446bb6
|
589 |
|
66c64223a
|
590 |
/* Found a free page, will break it into order-0 pages */ |
ab130f910
|
591 |
order = buddy_order(page); |
66c64223a
|
592 |
isolated = __isolate_free_page(page, order); |
a4f04f2c6
|
593 594 |
if (!isolated) break; |
66c64223a
|
595 |
set_page_private(page, order); |
a4f04f2c6
|
596 |
|
748446bb6
|
597 |
total_isolated += isolated; |
a4f04f2c6
|
598 |
cc->nr_freepages += isolated; |
66c64223a
|
599 |
list_add_tail(&page->lru, freelist); |
a4f04f2c6
|
600 601 602 |
if (!strict && cc->nr_migratepages <= cc->nr_freepages) { blockpfn += isolated; break; |
748446bb6
|
603 |
} |
a4f04f2c6
|
604 605 606 607 |
/* Advance to the end of split page */ blockpfn += isolated - 1; cursor += isolated - 1; continue; |
2af120bc0
|
608 609 610 611 612 613 |
isolate_fail: if (strict) break; else continue; |
748446bb6
|
614 |
} |
a4f04f2c6
|
615 616 |
if (locked) spin_unlock_irqrestore(&cc->zone->lock, flags); |
9fcd6d2e0
|
617 618 619 620 621 622 |
/* * There is a tiny chance that we have read bogus compound_order(), * so be careful to not go outside of the pageblock. */ if (unlikely(blockpfn > end_pfn)) blockpfn = end_pfn; |
e34d85f0e
|
623 624 |
trace_mm_compaction_isolate_freepages(*start_pfn, blockpfn, nr_scanned, total_isolated); |
e14c720ef
|
625 626 |
/* Record how far we have got within the block */ *start_pfn = blockpfn; |
f40d1e42b
|
627 628 629 630 631 |
/* * If strict isolation is requested by CMA then check that all the * pages requested were isolated. If there were any failures, 0 is * returned and CMA will fail. */ |
2af120bc0
|
632 |
if (strict && blockpfn < end_pfn) |
f40d1e42b
|
633 |
total_isolated = 0; |
7f354a548
|
634 |
cc->total_free_scanned += nr_scanned; |
397487db6
|
635 |
if (total_isolated) |
010fc29a4
|
636 |
count_compact_events(COMPACTISOLATED, total_isolated); |
748446bb6
|
637 638 |
return total_isolated; } |
85aa125f0
|
639 640 |
/** * isolate_freepages_range() - isolate free pages. |
e8b098fc5
|
641 |
* @cc: Compaction control structure. |
85aa125f0
|
642 643 644 645 646 647 648 649 650 651 652 |
* @start_pfn: The first PFN to start isolating. * @end_pfn: The one-past-last PFN. * * Non-free pages, invalid PFNs, or zone boundaries within the * [start_pfn, end_pfn) range are considered errors, cause function to * undo its actions and return zero. * * Otherwise, function returns one-past-the-last PFN of isolated page * (which may be greater then end_pfn if end fell in a middle of * a free page). */ |
ff9543fd3
|
653 |
unsigned long |
bb13ffeb9
|
654 655 |
isolate_freepages_range(struct compact_control *cc, unsigned long start_pfn, unsigned long end_pfn) |
85aa125f0
|
656 |
{ |
e1409c325
|
657 |
unsigned long isolated, pfn, block_start_pfn, block_end_pfn; |
85aa125f0
|
658 |
LIST_HEAD(freelist); |
7d49d8868
|
659 |
pfn = start_pfn; |
06b6640a3
|
660 |
block_start_pfn = pageblock_start_pfn(pfn); |
e1409c325
|
661 662 |
if (block_start_pfn < cc->zone->zone_start_pfn) block_start_pfn = cc->zone->zone_start_pfn; |
06b6640a3
|
663 |
block_end_pfn = pageblock_end_pfn(pfn); |
7d49d8868
|
664 665 |
for (; pfn < end_pfn; pfn += isolated, |
e1409c325
|
666 |
block_start_pfn = block_end_pfn, |
7d49d8868
|
667 |
block_end_pfn += pageblock_nr_pages) { |
e14c720ef
|
668 669 |
/* Protect pfn from changing by isolate_freepages_block */ unsigned long isolate_start_pfn = pfn; |
85aa125f0
|
670 |
|
85aa125f0
|
671 |
block_end_pfn = min(block_end_pfn, end_pfn); |
584200163
|
672 673 674 675 676 677 |
/* * pfn could pass the block_end_pfn if isolated freepage * is more than pageblock order. In this case, we adjust * scanning range to right one. */ if (pfn >= block_end_pfn) { |
06b6640a3
|
678 679 |
block_start_pfn = pageblock_start_pfn(pfn); block_end_pfn = pageblock_end_pfn(pfn); |
584200163
|
680 681 |
block_end_pfn = min(block_end_pfn, end_pfn); } |
e1409c325
|
682 683 |
if (!pageblock_pfn_to_page(block_start_pfn, block_end_pfn, cc->zone)) |
7d49d8868
|
684 |
break; |
e14c720ef
|
685 |
isolated = isolate_freepages_block(cc, &isolate_start_pfn, |
4fca9730c
|
686 |
block_end_pfn, &freelist, 0, true); |
85aa125f0
|
687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 |
/* * In strict mode, isolate_freepages_block() returns 0 if * there are any holes in the block (ie. invalid PFNs or * non-free pages). */ if (!isolated) break; /* * If we managed to isolate pages, it is always (1 << n) * * pageblock_nr_pages for some non-negative n. (Max order * page may span two pageblocks). */ } |
66c64223a
|
702 |
/* __isolate_free_page() does not map the pages */ |
4469ab984
|
703 |
split_map_pages(&freelist); |
85aa125f0
|
704 705 706 707 708 709 710 711 712 713 |
if (pfn < end_pfn) { /* Loop terminated early, cleanup. */ release_freepages(&freelist); return 0; } /* We don't use freelists for anything. */ return pfn; } |
748446bb6
|
714 |
/* Similar to reclaim, but different enough that they don't share logic */ |
5f438eee8
|
715 |
static bool too_many_isolated(pg_data_t *pgdat) |
748446bb6
|
716 |
{ |
bc6930457
|
717 |
unsigned long active, inactive, isolated; |
748446bb6
|
718 |
|
5f438eee8
|
719 720 721 722 723 724 |
inactive = node_page_state(pgdat, NR_INACTIVE_FILE) + node_page_state(pgdat, NR_INACTIVE_ANON); active = node_page_state(pgdat, NR_ACTIVE_FILE) + node_page_state(pgdat, NR_ACTIVE_ANON); isolated = node_page_state(pgdat, NR_ISOLATED_FILE) + node_page_state(pgdat, NR_ISOLATED_ANON); |
748446bb6
|
725 |
|
bc6930457
|
726 |
return isolated > (inactive + active) / 2; |
748446bb6
|
727 |
} |
2fe86e000
|
728 |
/** |
edc2ca612
|
729 730 |
* isolate_migratepages_block() - isolate all migrate-able pages within * a single pageblock |
2fe86e000
|
731 |
* @cc: Compaction control structure. |
edc2ca612
|
732 733 734 |
* @low_pfn: The first PFN to isolate * @end_pfn: The one-past-the-last PFN to isolate, within same pageblock * @isolate_mode: Isolation mode to be used. |
2fe86e000
|
735 736 |
* * Isolate all pages that can be migrated from the range specified by |
edc2ca612
|
737 |
* [low_pfn, end_pfn). The range is expected to be within same pageblock. |
c2ad7a1ff
|
738 |
* Returns errno, like -EAGAIN or -EINTR in case e.g signal pending or congestion, |
369fa227c
|
739 |
* -ENOMEM in case we could not allocate a page, or 0. |
c2ad7a1ff
|
740 |
* cc->migrate_pfn will contain the next pfn to scan. |
2fe86e000
|
741 |
* |
edc2ca612
|
742 |
* The pages are isolated on cc->migratepages list (not required to be empty), |
c2ad7a1ff
|
743 |
* and cc->nr_migratepages is updated accordingly. |
748446bb6
|
744 |
*/ |
c2ad7a1ff
|
745 |
static int |
edc2ca612
|
746 747 |
isolate_migratepages_block(struct compact_control *cc, unsigned long low_pfn, unsigned long end_pfn, isolate_mode_t isolate_mode) |
748446bb6
|
748 |
{ |
5f438eee8
|
749 |
pg_data_t *pgdat = cc->zone->zone_pgdat; |
b7aba6984
|
750 |
unsigned long nr_scanned = 0, nr_isolated = 0; |
fa9add641
|
751 |
struct lruvec *lruvec; |
b8b2d8253
|
752 |
unsigned long flags = 0; |
6168d0da2
|
753 |
struct lruvec *locked = NULL; |
bb13ffeb9
|
754 |
struct page *page = NULL, *valid_page = NULL; |
e34d85f0e
|
755 |
unsigned long start_pfn = low_pfn; |
fdd048e12
|
756 757 |
bool skip_on_failure = false; unsigned long next_skip_pfn = 0; |
e380bebe4
|
758 |
bool skip_updated = false; |
c2ad7a1ff
|
759 760 761 |
int ret = 0; cc->migrate_pfn = low_pfn; |
748446bb6
|
762 |
|
748446bb6
|
763 764 765 766 767 |
/* * Ensure that there are not too many pages isolated from the LRU * list by either parallel reclaimers or compaction. If there are, * delay for some time until fewer pages are isolated */ |
5f438eee8
|
768 |
while (unlikely(too_many_isolated(pgdat))) { |
d20bdd571
|
769 770 |
/* stop isolation if there are still pages not migrated */ if (cc->nr_migratepages) |
c2ad7a1ff
|
771 |
return -EAGAIN; |
d20bdd571
|
772 |
|
f9e35b3b4
|
773 |
/* async migration should just abort */ |
e0b9daeb4
|
774 |
if (cc->mode == MIGRATE_ASYNC) |
c2ad7a1ff
|
775 |
return -EAGAIN; |
f9e35b3b4
|
776 |
|
748446bb6
|
777 778 779 |
congestion_wait(BLK_RW_ASYNC, HZ/10); if (fatal_signal_pending(current)) |
c2ad7a1ff
|
780 |
return -EINTR; |
748446bb6
|
781 |
} |
cf66f0700
|
782 |
cond_resched(); |
aeef4b838
|
783 |
|
fdd048e12
|
784 785 786 787 |
if (cc->direct_compaction && (cc->mode == MIGRATE_ASYNC)) { skip_on_failure = true; next_skip_pfn = block_end_pfn(low_pfn, cc->order); } |
748446bb6
|
788 |
/* Time to isolate some pages for migration */ |
748446bb6
|
789 |
for (; low_pfn < end_pfn; low_pfn++) { |
29c0dde83
|
790 |
|
fdd048e12
|
791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 |
if (skip_on_failure && low_pfn >= next_skip_pfn) { /* * We have isolated all migration candidates in the * previous order-aligned block, and did not skip it due * to failure. We should migrate the pages now and * hopefully succeed compaction. */ if (nr_isolated) break; /* * We failed to isolate in the previous order-aligned * block. Set the new boundary to the end of the * current block. Note we can't simply increase * next_skip_pfn by 1 << order, as low_pfn might have * been incremented by a higher number due to skipping * a compound or a high-order buddy page in the * previous loop iteration. */ next_skip_pfn = block_end_pfn(low_pfn, cc->order); } |
8b44d2791
|
812 813 |
/* * Periodically drop the lock (if held) regardless of its |
670105a25
|
814 815 |
* contention, to give chance to IRQs. Abort completely if * a fatal signal is pending. |
8b44d2791
|
816 |
*/ |
6168d0da2
|
817 818 819 820 821 822 823 824 |
if (!(low_pfn % SWAP_CLUSTER_MAX)) { if (locked) { unlock_page_lruvec_irqrestore(locked, flags); locked = NULL; } if (fatal_signal_pending(current)) { cc->contended = true; |
c2ad7a1ff
|
825 |
ret = -EINTR; |
6168d0da2
|
826 |
|
6168d0da2
|
827 828 829 830 |
goto fatal_pending; } cond_resched(); |
670105a25
|
831 |
} |
c67fe3752
|
832 |
|
b7aba6984
|
833 |
nr_scanned++; |
748446bb6
|
834 |
|
748446bb6
|
835 |
page = pfn_to_page(low_pfn); |
dc9086004
|
836 |
|
e380bebe4
|
837 838 839 840 841 842 843 844 845 |
/* * Check if the pageblock has already been marked skipped. * Only the aligned PFN is checked as the caller isolates * COMPACT_CLUSTER_MAX at a time so the second call must * not falsely conclude that the block should be skipped. */ if (!valid_page && IS_ALIGNED(low_pfn, pageblock_nr_pages)) { if (!cc->ignore_skip_hint && get_pageblock_skip(page)) { low_pfn = end_pfn; |
9df413143
|
846 |
page = NULL; |
e380bebe4
|
847 848 |
goto isolate_abort; } |
bb13ffeb9
|
849 |
valid_page = page; |
e380bebe4
|
850 |
} |
bb13ffeb9
|
851 |
|
369fa227c
|
852 |
if (PageHuge(page) && cc->alloc_contig) { |
ae37c7ff7
|
853 |
ret = isolate_or_dissolve_huge_page(page, &cc->migratepages); |
369fa227c
|
854 855 856 857 858 859 860 861 862 863 864 865 |
/* * Fail isolation in case isolate_or_dissolve_huge_page() * reports an error. In case of -ENOMEM, abort right away. */ if (ret < 0) { /* Do not report -EBUSY down the chain */ if (ret == -EBUSY) ret = 0; low_pfn += (1UL << compound_order(page)) - 1; goto isolate_fail; } |
ae37c7ff7
|
866 867 868 869 870 871 872 873 |
if (PageHuge(page)) { /* * Hugepage was successfully isolated and placed * on the cc->migratepages list. */ low_pfn += compound_nr(page) - 1; goto isolate_success_no_list; } |
369fa227c
|
874 875 876 877 878 879 880 |
/* * Ok, the hugepage was dissolved. Now these pages are * Buddy and cannot be re-allocated because they are * isolated. Fall-through as the check below handles * Buddy pages. */ } |
6c14466cc
|
881 |
/* |
99c0fd5e5
|
882 883 884 885 |
* Skip if free. We read page order here without zone lock * which is generally unsafe, but the race window is small and * the worst thing that can happen is that we skip some * potential isolation targets. |
6c14466cc
|
886 |
*/ |
99c0fd5e5
|
887 |
if (PageBuddy(page)) { |
ab130f910
|
888 |
unsigned long freepage_order = buddy_order_unsafe(page); |
99c0fd5e5
|
889 890 891 892 893 894 895 896 |
/* * Without lock, we cannot be sure that what we got is * a valid page order. Consider only values in the * valid order range to prevent low_pfn overflow. */ if (freepage_order > 0 && freepage_order < MAX_ORDER) low_pfn += (1UL << freepage_order) - 1; |
748446bb6
|
897 |
continue; |
99c0fd5e5
|
898 |
} |
748446bb6
|
899 |
|
9927af740
|
900 |
/* |
29c0dde83
|
901 |
* Regardless of being on LRU, compound pages such as THP and |
1da2f328f
|
902 903 904 905 906 |
* hugetlbfs are not to be compacted unless we are attempting * an allocation much larger than the huge page size (eg CMA). * We can potentially save a lot of iterations if we skip them * at once. The check is racy, but we can consider only valid * values and the only danger is skipping too much. |
bc835011a
|
907 |
*/ |
1da2f328f
|
908 |
if (PageCompound(page) && !cc->alloc_contig) { |
21dc7e023
|
909 |
const unsigned int order = compound_order(page); |
edc2ca612
|
910 |
|
d3c85bad8
|
911 |
if (likely(order < MAX_ORDER)) |
21dc7e023
|
912 |
low_pfn += (1UL << order) - 1; |
fdd048e12
|
913 |
goto isolate_fail; |
2a1402aa0
|
914 |
} |
bda807d44
|
915 916 917 918 919 920 |
/* * Check may be lockless but that's ok as we recheck later. * It's possible to migrate LRU and non-lru movable pages. * Skip any other type of page */ if (!PageLRU(page)) { |
bda807d44
|
921 922 923 924 925 926 927 |
/* * __PageMovable can return false positive so we need * to verify it under page_lock. */ if (unlikely(__PageMovable(page)) && !PageIsolated(page)) { if (locked) { |
6168d0da2
|
928 929 |
unlock_page_lruvec_irqrestore(locked, flags); locked = NULL; |
bda807d44
|
930 |
} |
9e5bcd610
|
931 |
if (!isolate_movable_page(page, isolate_mode)) |
bda807d44
|
932 933 |
goto isolate_success; } |
fdd048e12
|
934 |
goto isolate_fail; |
bda807d44
|
935 |
} |
29c0dde83
|
936 |
|
119d6d59d
|
937 938 939 940 941 942 943 |
/* * Migration will fail if an anonymous page is pinned in memory, * so avoid taking lru_lock and isolating it unnecessarily in an * admittedly racy check. */ if (!page_mapping(page) && page_count(page) > page_mapcount(page)) |
fdd048e12
|
944 |
goto isolate_fail; |
119d6d59d
|
945 |
|
73e64c51a
|
946 947 948 949 950 951 |
/* * Only allow to migrate anonymous pages in GFP_NOFS context * because those do not depend on fs locks. */ if (!(cc->gfp_mask & __GFP_FS) && page_mapping(page)) goto isolate_fail; |
9df413143
|
952 953 954 955 956 957 958 |
/* * 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. */ if (unlikely(!get_page_unless_zero(page))) goto isolate_fail; |
c2135f7c5
|
959 |
if (!__isolate_lru_page_prepare(page, isolate_mode)) |
9df413143
|
960 961 962 963 964 |
goto isolate_fail_put; /* Try isolate the page */ if (!TestClearPageLRU(page)) goto isolate_fail_put; |
a984226f4
|
965 |
lruvec = mem_cgroup_page_lruvec(page); |
6168d0da2
|
966 |
|
69b7189f1
|
967 |
/* If we already hold the lock, we can skip some rechecking */ |
6168d0da2
|
968 969 970 971 972 973 |
if (lruvec != locked) { if (locked) unlock_page_lruvec_irqrestore(locked, flags); compact_lock_irqsave(&lruvec->lru_lock, &flags, cc); locked = lruvec; |
6168d0da2
|
974 975 |
lruvec_memcg_debug(lruvec, page); |
e380bebe4
|
976 |
|
e380bebe4
|
977 978 979 980 981 982 |
/* Try get exclusive access under lock */ if (!skip_updated) { skip_updated = true; if (test_and_set_skip(cc, page, low_pfn)) goto isolate_abort; } |
2a1402aa0
|
983 |
|
29c0dde83
|
984 985 986 987 988 |
/* * Page become compound since the non-locked check, * and it's on LRU. It can only be a THP so the order * is safe to read and it's 0 for tail pages. */ |
1da2f328f
|
989 |
if (unlikely(PageCompound(page) && !cc->alloc_contig)) { |
d8c6546b1
|
990 |
low_pfn += compound_nr(page) - 1; |
9df413143
|
991 992 |
SetPageLRU(page); goto isolate_fail_put; |
69b7189f1
|
993 |
} |
d99fd5feb
|
994 |
} |
fa9add641
|
995 |
|
1da2f328f
|
996 997 998 |
/* The whole page is taken off the LRU; skip the tail pages. */ if (PageCompound(page)) low_pfn += compound_nr(page) - 1; |
bc835011a
|
999 |
|
748446bb6
|
1000 |
/* Successfully isolated */ |
46ae6b2cc
|
1001 |
del_page_from_lru_list(page, lruvec); |
1da2f328f
|
1002 |
mod_node_page_state(page_pgdat(page), |
9de4f22a6
|
1003 |
NR_ISOLATED_ANON + page_is_file_lru(page), |
6c357848b
|
1004 |
thp_nr_pages(page)); |
b6c750163
|
1005 1006 |
isolate_success: |
fdd048e12
|
1007 |
list_add(&page->lru, &cc->migratepages); |
ae37c7ff7
|
1008 |
isolate_success_no_list: |
38935861d
|
1009 1010 |
cc->nr_migratepages += compound_nr(page); nr_isolated += compound_nr(page); |
748446bb6
|
1011 |
|
804d3121b
|
1012 1013 |
/* * Avoid isolating too much unless this block is being |
cb2dcaf02
|
1014 1015 1016 |
* rescanned (e.g. dirty/writeback pages, parallel allocation) * or a lock is contended. For contention, isolate quickly to * potentially remove one source of contention. |
804d3121b
|
1017 |
*/ |
38935861d
|
1018 |
if (cc->nr_migratepages >= COMPACT_CLUSTER_MAX && |
cb2dcaf02
|
1019 |
!cc->rescan && !cc->contended) { |
31b8384a5
|
1020 |
++low_pfn; |
748446bb6
|
1021 |
break; |
31b8384a5
|
1022 |
} |
fdd048e12
|
1023 1024 |
continue; |
9df413143
|
1025 1026 1027 1028 |
isolate_fail_put: /* Avoid potential deadlock in freeing page under lru_lock */ if (locked) { |
6168d0da2
|
1029 1030 |
unlock_page_lruvec_irqrestore(locked, flags); locked = NULL; |
9df413143
|
1031 1032 |
} put_page(page); |
fdd048e12
|
1033 |
isolate_fail: |
369fa227c
|
1034 |
if (!skip_on_failure && ret != -ENOMEM) |
fdd048e12
|
1035 1036 1037 1038 1039 1040 1041 1042 1043 |
continue; /* * We have isolated some pages, but then failed. Release them * instead of migrating, as we cannot form the cc->order buddy * page anyway. */ if (nr_isolated) { if (locked) { |
6168d0da2
|
1044 1045 |
unlock_page_lruvec_irqrestore(locked, flags); locked = NULL; |
fdd048e12
|
1046 |
} |
fdd048e12
|
1047 1048 |
putback_movable_pages(&cc->migratepages); cc->nr_migratepages = 0; |
fdd048e12
|
1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 |
nr_isolated = 0; } if (low_pfn < next_skip_pfn) { low_pfn = next_skip_pfn - 1; /* * The check near the loop beginning would have updated * next_skip_pfn too, but this is a bit simpler. */ next_skip_pfn += 1UL << cc->order; } |
369fa227c
|
1060 1061 1062 |
if (ret == -ENOMEM) break; |
748446bb6
|
1063 |
} |
99c0fd5e5
|
1064 1065 1066 1067 1068 1069 |
/* * The PageBuddy() check could have potentially brought us outside * the range to be scanned. */ if (unlikely(low_pfn > end_pfn)) low_pfn = end_pfn; |
9df413143
|
1070 |
page = NULL; |
e380bebe4
|
1071 |
isolate_abort: |
c67fe3752
|
1072 |
if (locked) |
6168d0da2
|
1073 |
unlock_page_lruvec_irqrestore(locked, flags); |
9df413143
|
1074 1075 1076 1077 |
if (page) { SetPageLRU(page); put_page(page); } |
748446bb6
|
1078 |
|
50b5b094e
|
1079 |
/* |
804d3121b
|
1080 1081 1082 1083 1084 1085 |
* Updated the cached scanner pfn once the pageblock has been scanned * Pages will either be migrated in which case there is no point * scanning in the near future or migration failed in which case the * failure reason may persist. The block is marked for skipping if * there were no pages isolated in the block or if the block is * rescanned twice in a row. |
50b5b094e
|
1086 |
*/ |
804d3121b
|
1087 |
if (low_pfn == end_pfn && (!nr_isolated || cc->rescan)) { |
e380bebe4
|
1088 1089 1090 1091 |
if (valid_page && !skip_updated) set_pageblock_skip(valid_page); update_cached_migrate(cc, low_pfn); } |
bb13ffeb9
|
1092 |
|
e34d85f0e
|
1093 1094 |
trace_mm_compaction_isolate_migratepages(start_pfn, low_pfn, nr_scanned, nr_isolated); |
b7aba6984
|
1095 |
|
670105a25
|
1096 |
fatal_pending: |
7f354a548
|
1097 |
cc->total_migrate_scanned += nr_scanned; |
397487db6
|
1098 |
if (nr_isolated) |
010fc29a4
|
1099 |
count_compact_events(COMPACTISOLATED, nr_isolated); |
397487db6
|
1100 |
|
c2ad7a1ff
|
1101 1102 1103 |
cc->migrate_pfn = low_pfn; return ret; |
2fe86e000
|
1104 |
} |
edc2ca612
|
1105 1106 1107 1108 1109 1110 |
/** * isolate_migratepages_range() - isolate migrate-able pages in a PFN range * @cc: Compaction control structure. * @start_pfn: The first PFN to start isolating. * @end_pfn: The one-past-last PFN. * |
369fa227c
|
1111 1112 |
* Returns -EAGAIN when contented, -EINTR in case of a signal pending, -ENOMEM * in case we could not allocate a page, or 0. |
edc2ca612
|
1113 |
*/ |
c2ad7a1ff
|
1114 |
int |
edc2ca612
|
1115 1116 1117 |
isolate_migratepages_range(struct compact_control *cc, unsigned long start_pfn, unsigned long end_pfn) { |
e1409c325
|
1118 |
unsigned long pfn, block_start_pfn, block_end_pfn; |
c2ad7a1ff
|
1119 |
int ret = 0; |
edc2ca612
|
1120 1121 1122 |
/* Scan block by block. First and last block may be incomplete */ pfn = start_pfn; |
06b6640a3
|
1123 |
block_start_pfn = pageblock_start_pfn(pfn); |
e1409c325
|
1124 1125 |
if (block_start_pfn < cc->zone->zone_start_pfn) block_start_pfn = cc->zone->zone_start_pfn; |
06b6640a3
|
1126 |
block_end_pfn = pageblock_end_pfn(pfn); |
edc2ca612
|
1127 1128 |
for (; pfn < end_pfn; pfn = block_end_pfn, |
e1409c325
|
1129 |
block_start_pfn = block_end_pfn, |
edc2ca612
|
1130 1131 1132 |
block_end_pfn += pageblock_nr_pages) { block_end_pfn = min(block_end_pfn, end_pfn); |
e1409c325
|
1133 1134 |
if (!pageblock_pfn_to_page(block_start_pfn, block_end_pfn, cc->zone)) |
edc2ca612
|
1135 |
continue; |
c2ad7a1ff
|
1136 1137 |
ret = isolate_migratepages_block(cc, pfn, block_end_pfn, ISOLATE_UNEVICTABLE); |
edc2ca612
|
1138 |
|
c2ad7a1ff
|
1139 |
if (ret) |
edc2ca612
|
1140 |
break; |
6ea41c0c0
|
1141 |
|
38935861d
|
1142 |
if (cc->nr_migratepages >= COMPACT_CLUSTER_MAX) |
6ea41c0c0
|
1143 |
break; |
edc2ca612
|
1144 |
} |
edc2ca612
|
1145 |
|
c2ad7a1ff
|
1146 |
return ret; |
edc2ca612
|
1147 |
} |
ff9543fd3
|
1148 1149 |
#endif /* CONFIG_COMPACTION || CONFIG_CMA */ #ifdef CONFIG_COMPACTION |
018e9a49a
|
1150 |
|
b682debd9
|
1151 1152 1153 |
static bool suitable_migration_source(struct compact_control *cc, struct page *page) { |
282722b0d
|
1154 |
int block_mt; |
9bebefd59
|
1155 1156 |
if (pageblock_skip_persistent(page)) return false; |
282722b0d
|
1157 |
if ((cc->mode != MIGRATE_ASYNC) || !cc->direct_compaction) |
b682debd9
|
1158 |
return true; |
282722b0d
|
1159 1160 1161 1162 1163 1164 |
block_mt = get_pageblock_migratetype(page); if (cc->migratetype == MIGRATE_MOVABLE) return is_migrate_movable(block_mt); else return block_mt == cc->migratetype; |
b682debd9
|
1165 |
} |
018e9a49a
|
1166 |
/* Returns true if the page is within a block suitable for migration to */ |
9f7e33879
|
1167 1168 |
static bool suitable_migration_target(struct compact_control *cc, struct page *page) |
018e9a49a
|
1169 1170 1171 1172 1173 1174 1175 1176 |
{ /* If the page is a large free page, then disallow migration */ if (PageBuddy(page)) { /* * We are checking page_order without zone->lock taken. But * the only small danger is that we skip a potentially suitable * pageblock, so it's not worth to check order for valid range. */ |
ab130f910
|
1177 |
if (buddy_order_unsafe(page) >= pageblock_order) |
018e9a49a
|
1178 1179 |
return false; } |
1ef36db2a
|
1180 1181 |
if (cc->ignore_block_suitable) return true; |
018e9a49a
|
1182 |
/* If the block is MIGRATE_MOVABLE or MIGRATE_CMA, allow migration */ |
b682debd9
|
1183 |
if (is_migrate_movable(get_pageblock_migratetype(page))) |
018e9a49a
|
1184 1185 1186 1187 1188 |
return true; /* Otherwise skip the block */ return false; } |
70b44595e
|
1189 1190 1191 |
static inline unsigned int freelist_scan_limit(struct compact_control *cc) { |
dd7ef7bd1
|
1192 1193 1194 |
unsigned short shift = BITS_PER_LONG - 1; return (COMPACT_CLUSTER_MAX >> min(shift, cc->fast_search_fail)) + 1; |
70b44595e
|
1195 |
} |
2fe86e000
|
1196 |
/* |
f2849aa09
|
1197 1198 1199 1200 1201 1202 1203 1204 |
* Test whether the free scanner has reached the same or lower pageblock than * the migration scanner, and compaction should thus terminate. */ static inline bool compact_scanners_met(struct compact_control *cc) { return (cc->free_pfn >> pageblock_order) <= (cc->migrate_pfn >> pageblock_order); } |
5a811889d
|
1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 |
/* * Used when scanning for a suitable migration target which scans freelists * in reverse. Reorders the list such as the unscanned pages are scanned * first on the next iteration of the free scanner */ static void move_freelist_head(struct list_head *freelist, struct page *freepage) { LIST_HEAD(sublist); if (!list_is_last(freelist, &freepage->lru)) { list_cut_before(&sublist, freelist, &freepage->lru); |
d2155fe54
|
1217 |
list_splice_tail(&sublist, freelist); |
5a811889d
|
1218 1219 1220 1221 1222 1223 1224 1225 1226 |
} } /* * Similar to move_freelist_head except used by the migration scanner * when scanning forward. It's possible for these list operations to * move against each other if they search the free list exactly in * lockstep. */ |
70b44595e
|
1227 1228 1229 1230 1231 1232 1233 |
static void move_freelist_tail(struct list_head *freelist, struct page *freepage) { LIST_HEAD(sublist); if (!list_is_first(freelist, &freepage->lru)) { list_cut_position(&sublist, freelist, &freepage->lru); |
d2155fe54
|
1234 |
list_splice_tail(&sublist, freelist); |
70b44595e
|
1235 1236 |
} } |
5a811889d
|
1237 1238 1239 1240 |
static void fast_isolate_around(struct compact_control *cc, unsigned long pfn, unsigned long nr_isolated) { unsigned long start_pfn, end_pfn; |
6e2b7044c
|
1241 |
struct page *page; |
5a811889d
|
1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 |
/* Do not search around if there are enough pages already */ if (cc->nr_freepages >= cc->nr_migratepages) return; /* Minimise scanning during async compaction */ if (cc->direct_compaction && cc->mode == MIGRATE_ASYNC) return; /* Pageblock boundaries */ |
6e2b7044c
|
1252 1253 1254 1255 1256 1257 |
start_pfn = max(pageblock_start_pfn(pfn), cc->zone->zone_start_pfn); end_pfn = min(pageblock_end_pfn(pfn), zone_end_pfn(cc->zone)); page = pageblock_pfn_to_page(start_pfn, end_pfn, cc->zone); if (!page) return; |
5a811889d
|
1258 1259 1260 |
/* Scan before */ if (start_pfn != pfn) { |
4fca9730c
|
1261 |
isolate_freepages_block(cc, &start_pfn, pfn, &cc->freepages, 1, false); |
5a811889d
|
1262 1263 1264 1265 1266 1267 |
if (cc->nr_freepages >= cc->nr_migratepages) return; } /* Scan after */ start_pfn = pfn + nr_isolated; |
60fce36af
|
1268 |
if (start_pfn < end_pfn) |
4fca9730c
|
1269 |
isolate_freepages_block(cc, &start_pfn, end_pfn, &cc->freepages, 1, false); |
5a811889d
|
1270 1271 1272 1273 1274 |
/* Skip this pageblock in the future as it's full or nearly full */ if (cc->nr_freepages < cc->nr_migratepages) set_pageblock_skip(page); } |
dbe2d4e4f
|
1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 |
/* Search orders in round-robin fashion */ static int next_search_order(struct compact_control *cc, int order) { order--; if (order < 0) order = cc->order - 1; /* Search wrapped around? */ if (order == cc->search_order) { cc->search_order--; if (cc->search_order < 0) cc->search_order = cc->order - 1; return -1; } return order; } |
5a811889d
|
1292 1293 1294 |
static unsigned long fast_isolate_freepages(struct compact_control *cc) { |
b55ca5264
|
1295 |
unsigned int limit = max(1U, freelist_scan_limit(cc) >> 1); |
5a811889d
|
1296 |
unsigned int nr_scanned = 0; |
74e21484e
|
1297 |
unsigned long low_pfn, min_pfn, highest = 0; |
5a811889d
|
1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 |
unsigned long nr_isolated = 0; unsigned long distance; struct page *page = NULL; bool scan_start = false; int order; /* Full compaction passes in a negative order */ if (cc->order <= 0) return cc->free_pfn; /* * If starting the scan, use a deeper search and use the highest * PFN found if a suitable one is not found. */ |
e332f741a
|
1312 |
if (cc->free_pfn >= cc->zone->compact_init_free_pfn) { |
5a811889d
|
1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 |
limit = pageblock_nr_pages >> 1; scan_start = true; } /* * Preferred point is in the top quarter of the scan space but take * a pfn from the top half if the search is problematic. */ distance = (cc->free_pfn - cc->migrate_pfn); low_pfn = pageblock_start_pfn(cc->free_pfn - (distance >> 2)); min_pfn = pageblock_start_pfn(cc->free_pfn - (distance >> 1)); if (WARN_ON_ONCE(min_pfn > low_pfn)) low_pfn = min_pfn; |
dbe2d4e4f
|
1327 1328 1329 1330 1331 1332 1333 1334 1335 |
/* * Search starts from the last successful isolation order or the next * order to search after a previous failure */ cc->search_order = min_t(unsigned int, cc->order - 1, cc->search_order); for (order = cc->search_order; !page && order >= 0; order = next_search_order(cc, order)) { |
5a811889d
|
1336 1337 1338 1339 1340 |
struct free_area *area = &cc->zone->free_area[order]; struct list_head *freelist; struct page *freepage; unsigned long flags; unsigned int order_scanned = 0; |
74e21484e
|
1341 |
unsigned long high_pfn = 0; |
5a811889d
|
1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 |
if (!area->nr_free) continue; spin_lock_irqsave(&cc->zone->lock, flags); freelist = &area->free_list[MIGRATE_MOVABLE]; list_for_each_entry_reverse(freepage, freelist, lru) { unsigned long pfn; order_scanned++; nr_scanned++; pfn = page_to_pfn(freepage); if (pfn >= highest) |
6e2b7044c
|
1356 1357 |
highest = max(pageblock_start_pfn(pfn), cc->zone->zone_start_pfn); |
5a811889d
|
1358 1359 1360 |
if (pfn >= low_pfn) { cc->fast_search_fail = 0; |
dbe2d4e4f
|
1361 |
cc->search_order = order; |
5a811889d
|
1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 |
page = freepage; break; } if (pfn >= min_pfn && pfn > high_pfn) { high_pfn = pfn; /* Shorten the scan if a candidate is found */ limit >>= 1; } if (order_scanned >= limit) break; } /* Use a minimum pfn if a preferred one was not found */ if (!page && high_pfn) { page = pfn_to_page(high_pfn); /* Update freepage for the list reorder below */ freepage = page; } /* Reorder to so a future search skips recent pages */ move_freelist_head(freelist, freepage); /* Isolate the page if available */ if (page) { if (__isolate_free_page(page, order)) { set_page_private(page, order); nr_isolated = 1 << order; cc->nr_freepages += nr_isolated; list_add_tail(&page->lru, &cc->freepages); count_compact_events(COMPACTISOLATED, nr_isolated); } else { /* If isolation fails, abort the search */ |
5b56d996d
|
1398 |
order = cc->search_order + 1; |
5a811889d
|
1399 1400 1401 1402 1403 1404 1405 |
page = NULL; } } spin_unlock_irqrestore(&cc->zone->lock, flags); /* |
b55ca5264
|
1406 |
* Smaller scan on next order so the total scan is related |
5a811889d
|
1407 1408 1409 |
* to freelist_scan_limit. */ if (order_scanned >= limit) |
b55ca5264
|
1410 |
limit = max(1U, limit >> 1); |
5a811889d
|
1411 1412 1413 1414 1415 1416 1417 |
} if (!page) { cc->fast_search_fail++; if (scan_start) { /* * Use the highest PFN found above min. If one was |
f38677551
|
1418 |
* not found, be pessimistic for direct compaction |
5a811889d
|
1419 1420 1421 1422 1423 1424 |
* and use the min mark. */ if (highest) { page = pfn_to_page(highest); cc->free_pfn = highest; } else { |
e577c8b64
|
1425 |
if (cc->direct_compaction && pfn_valid(min_pfn)) { |
73a6e474c
|
1426 |
page = pageblock_pfn_to_page(min_pfn, |
6e2b7044c
|
1427 1428 |
min(pageblock_end_pfn(min_pfn), zone_end_pfn(cc->zone)), |
73a6e474c
|
1429 |
cc->zone); |
5a811889d
|
1430 1431 1432 1433 1434 |
cc->free_pfn = min_pfn; } } } } |
d097a6f63
|
1435 1436 |
if (highest && highest >= cc->zone->compact_cached_free_pfn) { highest -= pageblock_nr_pages; |
5a811889d
|
1437 |
cc->zone->compact_cached_free_pfn = highest; |
d097a6f63
|
1438 |
} |
5a811889d
|
1439 1440 1441 1442 1443 1444 1445 1446 1447 |
cc->total_free_scanned += nr_scanned; if (!page) return cc->free_pfn; low_pfn = page_to_pfn(page); fast_isolate_around(cc, low_pfn, nr_isolated); return low_pfn; } |
f2849aa09
|
1448 |
/* |
ff9543fd3
|
1449 1450 |
* Based on information in the current compact_control, find blocks * suitable for isolating free pages from and then isolate them. |
2fe86e000
|
1451 |
*/ |
edc2ca612
|
1452 |
static void isolate_freepages(struct compact_control *cc) |
2fe86e000
|
1453 |
{ |
edc2ca612
|
1454 |
struct zone *zone = cc->zone; |
ff9543fd3
|
1455 |
struct page *page; |
c96b9e508
|
1456 |
unsigned long block_start_pfn; /* start of current pageblock */ |
e14c720ef
|
1457 |
unsigned long isolate_start_pfn; /* exact pfn we start at */ |
c96b9e508
|
1458 1459 |
unsigned long block_end_pfn; /* end of current pageblock */ unsigned long low_pfn; /* lowest pfn scanner is able to scan */ |
ff9543fd3
|
1460 |
struct list_head *freelist = &cc->freepages; |
4fca9730c
|
1461 |
unsigned int stride; |
2fe86e000
|
1462 |
|
5a811889d
|
1463 1464 1465 1466 |
/* Try a small search of the free lists for a candidate */ isolate_start_pfn = fast_isolate_freepages(cc); if (cc->nr_freepages) goto splitmap; |
ff9543fd3
|
1467 1468 |
/* * Initialise the free scanner. The starting point is where we last |
49e068f0b
|
1469 |
* successfully isolated from, zone-cached value, or the end of the |
e14c720ef
|
1470 1471 |
* zone when isolating for the first time. For looping we also need * this pfn aligned down to the pageblock boundary, because we do |
c96b9e508
|
1472 1473 |
* block_start_pfn -= pageblock_nr_pages in the for loop. * For ending point, take care when isolating in last pageblock of a |
a1c1dbeb2
|
1474 |
* zone which ends in the middle of a pageblock. |
49e068f0b
|
1475 1476 |
* The low boundary is the end of the pageblock the migration scanner * is using. |
ff9543fd3
|
1477 |
*/ |
e14c720ef
|
1478 |
isolate_start_pfn = cc->free_pfn; |
5a811889d
|
1479 |
block_start_pfn = pageblock_start_pfn(isolate_start_pfn); |
c96b9e508
|
1480 1481 |
block_end_pfn = min(block_start_pfn + pageblock_nr_pages, zone_end_pfn(zone)); |
06b6640a3
|
1482 |
low_pfn = pageblock_end_pfn(cc->migrate_pfn); |
4fca9730c
|
1483 |
stride = cc->mode == MIGRATE_ASYNC ? COMPACT_CLUSTER_MAX : 1; |
2fe86e000
|
1484 |
|
ff9543fd3
|
1485 |
/* |
ff9543fd3
|
1486 1487 1488 1489 |
* Isolate free pages until enough are available to migrate the * pages on cc->migratepages. We stop searching if the migrate * and free page scanners meet or enough free pages are isolated. */ |
f5f61a320
|
1490 |
for (; block_start_pfn >= low_pfn; |
c96b9e508
|
1491 |
block_end_pfn = block_start_pfn, |
e14c720ef
|
1492 1493 |
block_start_pfn -= pageblock_nr_pages, isolate_start_pfn = block_start_pfn) { |
4fca9730c
|
1494 |
unsigned long nr_isolated; |
f6ea3adb7
|
1495 1496 |
/* * This can iterate a massively long zone without finding any |
cb810ad29
|
1497 |
* suitable migration targets, so periodically check resched. |
f6ea3adb7
|
1498 |
*/ |
cb810ad29
|
1499 |
if (!(block_start_pfn % (SWAP_CLUSTER_MAX * pageblock_nr_pages))) |
cf66f0700
|
1500 |
cond_resched(); |
f6ea3adb7
|
1501 |
|
7d49d8868
|
1502 1503 1504 |
page = pageblock_pfn_to_page(block_start_pfn, block_end_pfn, zone); if (!page) |
ff9543fd3
|
1505 1506 1507 |
continue; /* Check the block is suitable for migration */ |
9f7e33879
|
1508 |
if (!suitable_migration_target(cc, page)) |
ff9543fd3
|
1509 |
continue; |
68e3e9262
|
1510 |
|
bb13ffeb9
|
1511 1512 1513 |
/* If isolation recently failed, do not retry */ if (!isolation_suitable(cc, page)) continue; |
e14c720ef
|
1514 |
/* Found a block suitable for isolating free pages from. */ |
4fca9730c
|
1515 1516 |
nr_isolated = isolate_freepages_block(cc, &isolate_start_pfn, block_end_pfn, freelist, stride, false); |
ff9543fd3
|
1517 |
|
d097a6f63
|
1518 1519 1520 |
/* Update the skip hint if the full pageblock was scanned */ if (isolate_start_pfn == block_end_pfn) update_pageblock_skip(cc, page, block_start_pfn); |
cb2dcaf02
|
1521 1522 |
/* Are enough freepages isolated? */ if (cc->nr_freepages >= cc->nr_migratepages) { |
a46cbf3bc
|
1523 1524 1525 1526 1527 |
if (isolate_start_pfn >= block_end_pfn) { /* * Restart at previous pageblock if more * freepages can be isolated next time. */ |
f5f61a320
|
1528 1529 |
isolate_start_pfn = block_start_pfn - pageblock_nr_pages; |
a46cbf3bc
|
1530 |
} |
be9765722
|
1531 |
break; |
a46cbf3bc
|
1532 |
} else if (isolate_start_pfn < block_end_pfn) { |
f5f61a320
|
1533 |
/* |
a46cbf3bc
|
1534 1535 |
* If isolation failed early, do not continue * needlessly. |
f5f61a320
|
1536 |
*/ |
a46cbf3bc
|
1537 |
break; |
f5f61a320
|
1538 |
} |
4fca9730c
|
1539 1540 1541 1542 1543 1544 1545 |
/* Adjust stride depending on isolation */ if (nr_isolated) { stride = 1; continue; } stride = min_t(unsigned int, COMPACT_CLUSTER_MAX, stride << 1); |
ff9543fd3
|
1546 |
} |
7ed695e06
|
1547 |
/* |
f5f61a320
|
1548 1549 1550 1551 |
* Record where the free scanner will restart next time. Either we * broke from the loop and set isolate_start_pfn based on the last * call to isolate_freepages_block(), or we met the migration scanner * and the loop terminated due to isolate_start_pfn < low_pfn |
7ed695e06
|
1552 |
*/ |
f5f61a320
|
1553 |
cc->free_pfn = isolate_start_pfn; |
5a811889d
|
1554 1555 1556 1557 |
splitmap: /* __isolate_free_page() does not map the pages */ split_map_pages(freelist); |
748446bb6
|
1558 1559 1560 1561 1562 1563 1564 |
} /* * This is a migrate-callback that "allocates" freepages by taking pages * from the isolated freelists in the block we are migrating to. */ static struct page *compaction_alloc(struct page *migratepage, |
666feb21a
|
1565 |
unsigned long data) |
748446bb6
|
1566 1567 1568 |
{ struct compact_control *cc = (struct compact_control *)data; struct page *freepage; |
748446bb6
|
1569 |
if (list_empty(&cc->freepages)) { |
cb2dcaf02
|
1570 |
isolate_freepages(cc); |
748446bb6
|
1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 |
if (list_empty(&cc->freepages)) return NULL; } freepage = list_entry(cc->freepages.next, struct page, lru); list_del(&freepage->lru); cc->nr_freepages--; return freepage; } /* |
d53aea3d4
|
1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 |
* This is a migrate-callback that "frees" freepages back to the isolated * freelist. All pages on the freelist are from the same zone, so there is no * special handling needed for NUMA. */ static void compaction_free(struct page *page, unsigned long data) { struct compact_control *cc = (struct compact_control *)data; list_add(&page->lru, &cc->freepages); cc->nr_freepages++; } |
ff9543fd3
|
1595 1596 1597 1598 1599 1600 1601 1602 |
/* possible outcome of isolate_migratepages */ typedef enum { ISOLATE_ABORT, /* Abort compaction now */ ISOLATE_NONE, /* No pages isolated, continue scanning */ ISOLATE_SUCCESS, /* Pages isolated, migrate */ } isolate_migrate_t; /* |
5bbe3547a
|
1603 1604 1605 |
* Allow userspace to control policy on scanning the unevictable LRU for * compactable pages. */ |
6923aa0d8
|
1606 1607 1608 |
#ifdef CONFIG_PREEMPT_RT int sysctl_compact_unevictable_allowed __read_mostly = 0; #else |
5bbe3547a
|
1609 |
int sysctl_compact_unevictable_allowed __read_mostly = 1; |
6923aa0d8
|
1610 |
#endif |
5bbe3547a
|
1611 |
|
70b44595e
|
1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 |
static inline void update_fast_start_pfn(struct compact_control *cc, unsigned long pfn) { if (cc->fast_start_pfn == ULONG_MAX) return; if (!cc->fast_start_pfn) cc->fast_start_pfn = pfn; cc->fast_start_pfn = min(cc->fast_start_pfn, pfn); } static inline unsigned long reinit_migrate_pfn(struct compact_control *cc) { if (!cc->fast_start_pfn || cc->fast_start_pfn == ULONG_MAX) return cc->migrate_pfn; cc->migrate_pfn = cc->fast_start_pfn; cc->fast_start_pfn = ULONG_MAX; return cc->migrate_pfn; } /* * Briefly search the free lists for a migration source that already has * some free pages to reduce the number of pages that need migration * before a pageblock is free. */ static unsigned long fast_find_migrateblock(struct compact_control *cc) { unsigned int limit = freelist_scan_limit(cc); unsigned int nr_scanned = 0; unsigned long distance; unsigned long pfn = cc->migrate_pfn; unsigned long high_pfn; int order; |
15d28d0d1
|
1649 |
bool found_block = false; |
70b44595e
|
1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 |
/* Skip hints are relied on to avoid repeats on the fast search */ if (cc->ignore_skip_hint) return pfn; /* * If the migrate_pfn is not at the start of a zone or the start * of a pageblock then assume this is a continuation of a previous * scan restarted due to COMPACT_CLUSTER_MAX. */ if (pfn != cc->zone->zone_start_pfn && pfn != pageblock_start_pfn(pfn)) return pfn; /* * For smaller orders, just linearly scan as the number of pages * to migrate should be relatively small and does not necessarily * justify freeing up a large block for a small allocation. */ if (cc->order <= PAGE_ALLOC_COSTLY_ORDER) return pfn; /* * Only allow kcompactd and direct requests for movable pages to * quickly clear out a MOVABLE pageblock for allocation. This * reduces the risk that a large movable pageblock is freed for * an unmovable/reclaimable small allocation. */ if (cc->direct_compaction && cc->migratetype != MIGRATE_MOVABLE) return pfn; /* * When starting the migration scanner, pick any pageblock within the * first half of the search space. Otherwise try and pick a pageblock * within the first eighth to reduce the chances that a migration * target later becomes a source. */ distance = (cc->free_pfn - cc->migrate_pfn) >> 1; if (cc->migrate_pfn != cc->zone->zone_start_pfn) distance >>= 2; high_pfn = pageblock_start_pfn(cc->migrate_pfn + distance); for (order = cc->order - 1; |
15d28d0d1
|
1692 |
order >= PAGE_ALLOC_COSTLY_ORDER && !found_block && nr_scanned < limit; |
70b44595e
|
1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 |
order--) { struct free_area *area = &cc->zone->free_area[order]; struct list_head *freelist; unsigned long flags; struct page *freepage; if (!area->nr_free) continue; spin_lock_irqsave(&cc->zone->lock, flags); freelist = &area->free_list[MIGRATE_MOVABLE]; list_for_each_entry(freepage, freelist, lru) { unsigned long free_pfn; |
15d28d0d1
|
1706 1707 1708 1709 |
if (nr_scanned++ >= limit) { move_freelist_tail(freelist, freepage); break; } |
70b44595e
|
1710 1711 |
free_pfn = page_to_pfn(freepage); if (free_pfn < high_pfn) { |
70b44595e
|
1712 1713 1714 1715 1716 1717 |
/* * Avoid if skipped recently. Ideally it would * move to the tail but even safe iteration of * the list assumes an entry is deleted, not * reordered. */ |
15d28d0d1
|
1718 |
if (get_pageblock_skip(freepage)) |
70b44595e
|
1719 |
continue; |
70b44595e
|
1720 1721 1722 |
/* Reorder to so a future search skips recent pages */ move_freelist_tail(freelist, freepage); |
e380bebe4
|
1723 |
update_fast_start_pfn(cc, free_pfn); |
70b44595e
|
1724 |
pfn = pageblock_start_pfn(free_pfn); |
20e6ec76a
|
1725 1726 |
if (pfn < cc->zone->zone_start_pfn) pfn = cc->zone->zone_start_pfn; |
70b44595e
|
1727 |
cc->fast_search_fail = 0; |
15d28d0d1
|
1728 |
found_block = true; |
70b44595e
|
1729 1730 1731 |
set_pageblock_skip(freepage); break; } |
70b44595e
|
1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 |
} spin_unlock_irqrestore(&cc->zone->lock, flags); } cc->total_migrate_scanned += nr_scanned; /* * If fast scanning failed then use a cached entry for a page block * that had free pages as the basis for starting a linear scan. */ |
15d28d0d1
|
1742 1743 |
if (!found_block) { cc->fast_search_fail++; |
70b44595e
|
1744 |
pfn = reinit_migrate_pfn(cc); |
15d28d0d1
|
1745 |
} |
70b44595e
|
1746 1747 |
return pfn; } |
5bbe3547a
|
1748 |
/* |
edc2ca612
|
1749 1750 1751 |
* Isolate all pages that can be migrated from the first suitable block, * starting at the block pointed to by the migrate scanner pfn within * compact_control. |
ff9543fd3
|
1752 |
*/ |
32aaf0553
|
1753 |
static isolate_migrate_t isolate_migratepages(struct compact_control *cc) |
ff9543fd3
|
1754 |
{ |
e1409c325
|
1755 1756 1757 |
unsigned long block_start_pfn; unsigned long block_end_pfn; unsigned long low_pfn; |
edc2ca612
|
1758 1759 |
struct page *page; const isolate_mode_t isolate_mode = |
5bbe3547a
|
1760 |
(sysctl_compact_unevictable_allowed ? ISOLATE_UNEVICTABLE : 0) | |
1d2047fef
|
1761 |
(cc->mode != MIGRATE_SYNC ? ISOLATE_ASYNC_MIGRATE : 0); |
70b44595e
|
1762 |
bool fast_find_block; |
ff9543fd3
|
1763 |
|
edc2ca612
|
1764 1765 |
/* * Start at where we last stopped, or beginning of the zone as |
70b44595e
|
1766 1767 |
* initialized by compact_zone(). The first failure will use * the lowest PFN as the starting point for linear scanning. |
edc2ca612
|
1768 |
*/ |
70b44595e
|
1769 |
low_pfn = fast_find_migrateblock(cc); |
06b6640a3
|
1770 |
block_start_pfn = pageblock_start_pfn(low_pfn); |
32aaf0553
|
1771 1772 |
if (block_start_pfn < cc->zone->zone_start_pfn) block_start_pfn = cc->zone->zone_start_pfn; |
ff9543fd3
|
1773 |
|
70b44595e
|
1774 1775 1776 1777 1778 1779 |
/* * fast_find_migrateblock marks a pageblock skipped so to avoid * the isolation_suitable check below, check whether the fast * search was successful. */ fast_find_block = low_pfn != cc->migrate_pfn && !cc->fast_search_fail; |
ff9543fd3
|
1780 |
/* Only scan within a pageblock boundary */ |
06b6640a3
|
1781 |
block_end_pfn = pageblock_end_pfn(low_pfn); |
ff9543fd3
|
1782 |
|
edc2ca612
|
1783 1784 1785 1786 |
/* * Iterate over whole pageblocks until we find the first suitable. * Do not cross the free scanner. */ |
e1409c325
|
1787 |
for (; block_end_pfn <= cc->free_pfn; |
70b44595e
|
1788 |
fast_find_block = false, |
c2ad7a1ff
|
1789 |
cc->migrate_pfn = low_pfn = block_end_pfn, |
e1409c325
|
1790 1791 |
block_start_pfn = block_end_pfn, block_end_pfn += pageblock_nr_pages) { |
ff9543fd3
|
1792 |
|
edc2ca612
|
1793 1794 1795 |
/* * This can potentially iterate a massively long zone with * many pageblocks unsuitable, so periodically check if we |
cb810ad29
|
1796 |
* need to schedule. |
edc2ca612
|
1797 |
*/ |
cb810ad29
|
1798 |
if (!(low_pfn % (SWAP_CLUSTER_MAX * pageblock_nr_pages))) |
cf66f0700
|
1799 |
cond_resched(); |
ff9543fd3
|
1800 |
|
32aaf0553
|
1801 1802 |
page = pageblock_pfn_to_page(block_start_pfn, block_end_pfn, cc->zone); |
7d49d8868
|
1803 |
if (!page) |
edc2ca612
|
1804 |
continue; |
e380bebe4
|
1805 1806 1807 1808 1809 1810 1811 1812 1813 |
/* * If isolation recently failed, do not retry. Only check the * pageblock once. COMPACT_CLUSTER_MAX causes a pageblock * to be visited multiple times. Assume skip was checked * before making it "skip" so other compaction instances do * not scan the same block. */ if (IS_ALIGNED(low_pfn, pageblock_nr_pages) && !fast_find_block && !isolation_suitable(cc, page)) |
edc2ca612
|
1814 1815 1816 |
continue; /* |
9bebefd59
|
1817 1818 1819 1820 1821 1822 |
* For async compaction, also only scan in MOVABLE blocks * without huge pages. Async compaction is optimistic to see * if the minimum amount of work satisfies the allocation. * The cached PFN is updated as it's possible that all * remaining blocks between source and target are unsuitable * and the compaction scanners fail to meet. |
edc2ca612
|
1823 |
*/ |
9bebefd59
|
1824 1825 |
if (!suitable_migration_source(cc, page)) { update_cached_migrate(cc, block_end_pfn); |
edc2ca612
|
1826 |
continue; |
9bebefd59
|
1827 |
} |
edc2ca612
|
1828 1829 |
/* Perform the isolation */ |
c2ad7a1ff
|
1830 1831 |
if (isolate_migratepages_block(cc, low_pfn, block_end_pfn, isolate_mode)) |
edc2ca612
|
1832 1833 1834 1835 1836 1837 1838 1839 1840 |
return ISOLATE_ABORT; /* * Either we isolated something and proceed with migration. Or * we failed and compact_zone should decide if we should * continue or not. */ break; } |
edc2ca612
|
1841 |
return cc->nr_migratepages ? ISOLATE_SUCCESS : ISOLATE_NONE; |
ff9543fd3
|
1842 |
} |
21c527a3c
|
1843 1844 1845 1846 1847 1848 1849 1850 |
/* * order == -1 is expected when compacting via * /proc/sys/vm/compact_memory */ static inline bool is_via_compact_memory(int order) { return order == -1; } |
facdaa917
|
1851 1852 |
static bool kswapd_is_running(pg_data_t *pgdat) { |
b03fbd4ff
|
1853 |
return pgdat->kswapd && task_is_running(pgdat->kswapd); |
facdaa917
|
1854 1855 1856 1857 |
} /* * A zone's fragmentation score is the external fragmentation wrt to the |
40d7e2032
|
1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 |
* COMPACTION_HPAGE_ORDER. It returns a value in the range [0, 100]. */ static unsigned int fragmentation_score_zone(struct zone *zone) { return extfrag_for_order(zone, COMPACTION_HPAGE_ORDER); } /* * A weighted zone's fragmentation score is the external fragmentation * wrt to the COMPACTION_HPAGE_ORDER scaled by the zone's size. It * returns a value in the range [0, 100]. |
facdaa917
|
1869 1870 1871 1872 1873 1874 |
* * The scaling factor ensures that proactive compaction focuses on larger * zones like ZONE_NORMAL, rather than smaller, specialized zones like * ZONE_DMA32. For smaller zones, the score value remains close to zero, * and thus never exceeds the high threshold for proactive compaction. */ |
40d7e2032
|
1875 |
static unsigned int fragmentation_score_zone_weighted(struct zone *zone) |
facdaa917
|
1876 1877 |
{ unsigned long score; |
40d7e2032
|
1878 |
score = zone->present_pages * fragmentation_score_zone(zone); |
facdaa917
|
1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 |
return div64_ul(score, zone->zone_pgdat->node_present_pages + 1); } /* * The per-node proactive (background) compaction process is started by its * corresponding kcompactd thread when the node's fragmentation score * exceeds the high threshold. The compaction process remains active till * the node's score falls below the low threshold, or one of the back-off * conditions is met. */ |
d34c0a759
|
1889 |
static unsigned int fragmentation_score_node(pg_data_t *pgdat) |
facdaa917
|
1890 |
{ |
d34c0a759
|
1891 |
unsigned int score = 0; |
facdaa917
|
1892 1893 1894 1895 1896 1897 |
int zoneid; for (zoneid = 0; zoneid < MAX_NR_ZONES; zoneid++) { struct zone *zone; zone = &pgdat->node_zones[zoneid]; |
40d7e2032
|
1898 |
score += fragmentation_score_zone_weighted(zone); |
facdaa917
|
1899 1900 1901 1902 |
} return score; } |
d34c0a759
|
1903 |
static unsigned int fragmentation_score_wmark(pg_data_t *pgdat, bool low) |
facdaa917
|
1904 |
{ |
d34c0a759
|
1905 |
unsigned int wmark_low; |
facdaa917
|
1906 1907 |
/* |
f0953a1bb
|
1908 1909 |
* Cap the low watermark to avoid excessive compaction * activity in case a user sets the proactiveness tunable |
facdaa917
|
1910 1911 |
* close to 100 (maximum). */ |
d34c0a759
|
1912 1913 |
wmark_low = max(100U - sysctl_compaction_proactiveness, 5U); return low ? wmark_low : min(wmark_low + 10, 100U); |
facdaa917
|
1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 |
} static bool should_proactive_compact_node(pg_data_t *pgdat) { int wmark_high; if (!sysctl_compaction_proactiveness || kswapd_is_running(pgdat)) return false; wmark_high = fragmentation_score_wmark(pgdat, false); return fragmentation_score_node(pgdat) > wmark_high; } |
40cacbcb3
|
1926 |
static enum compact_result __compact_finished(struct compact_control *cc) |
748446bb6
|
1927 |
{ |
8fb74b9fb
|
1928 |
unsigned int order; |
d39773a06
|
1929 |
const int migratetype = cc->migratetype; |
cb2dcaf02
|
1930 |
int ret; |
748446bb6
|
1931 |
|
753341a4b
|
1932 |
/* Compaction run completes if the migrate and free scanner meet */ |
f2849aa09
|
1933 |
if (compact_scanners_met(cc)) { |
55b7c4c99
|
1934 |
/* Let the next compaction start anew. */ |
40cacbcb3
|
1935 |
reset_cached_positions(cc->zone); |
55b7c4c99
|
1936 |
|
62997027c
|
1937 1938 |
/* * Mark that the PG_migrate_skip information should be cleared |
accf62422
|
1939 |
* by kswapd when it goes to sleep. kcompactd does not set the |
62997027c
|
1940 1941 1942 |
* flag itself as the decision to be clear should be directly * based on an allocation request. */ |
accf62422
|
1943 |
if (cc->direct_compaction) |
40cacbcb3
|
1944 |
cc->zone->compact_blockskip_flush = true; |
62997027c
|
1945 |
|
c8f7de0bf
|
1946 1947 1948 1949 |
if (cc->whole_zone) return COMPACT_COMPLETE; else return COMPACT_PARTIAL_SKIPPED; |
bb13ffeb9
|
1950 |
} |
748446bb6
|
1951 |
|
facdaa917
|
1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 |
if (cc->proactive_compaction) { int score, wmark_low; pg_data_t *pgdat; pgdat = cc->zone->zone_pgdat; if (kswapd_is_running(pgdat)) return COMPACT_PARTIAL_SKIPPED; score = fragmentation_score_zone(cc->zone); wmark_low = fragmentation_score_wmark(pgdat, true); if (score > wmark_low) ret = COMPACT_CONTINUE; else ret = COMPACT_SUCCESS; goto out; } |
21c527a3c
|
1970 |
if (is_via_compact_memory(cc->order)) |
56de7263f
|
1971 |
return COMPACT_CONTINUE; |
efe771c76
|
1972 1973 1974 1975 1976 1977 1978 1979 |
/* * Always finish scanning a pageblock to reduce the possibility of * fallbacks in the future. This is particularly important when * migration source is unmovable/reclaimable but it's not worth * special casing. */ if (!IS_ALIGNED(cc->migrate_pfn, pageblock_nr_pages)) return COMPACT_CONTINUE; |
baf6a9a1d
|
1980 |
|
56de7263f
|
1981 |
/* Direct compactor: Is a suitable page free? */ |
cb2dcaf02
|
1982 |
ret = COMPACT_NO_SUITABLE_PAGE; |
8fb74b9fb
|
1983 |
for (order = cc->order; order < MAX_ORDER; order++) { |
40cacbcb3
|
1984 |
struct free_area *area = &cc->zone->free_area[order]; |
2149cdaef
|
1985 |
bool can_steal; |
8fb74b9fb
|
1986 1987 |
/* Job done if page is free of the right migratetype */ |
b03641af6
|
1988 |
if (!free_area_empty(area, migratetype)) |
cf378319d
|
1989 |
return COMPACT_SUCCESS; |
8fb74b9fb
|
1990 |
|
2149cdaef
|
1991 1992 1993 |
#ifdef CONFIG_CMA /* MIGRATE_MOVABLE can fallback on MIGRATE_CMA */ if (migratetype == MIGRATE_MOVABLE && |
b03641af6
|
1994 |
!free_area_empty(area, MIGRATE_CMA)) |
cf378319d
|
1995 |
return COMPACT_SUCCESS; |
2149cdaef
|
1996 1997 1998 1999 2000 2001 |
#endif /* * Job done if allocation would steal freepages from * other migratetype buddy lists. */ if (find_suitable_fallback(area, order, migratetype, |
baf6a9a1d
|
2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 |
true, &can_steal) != -1) { /* movable pages are OK in any pageblock */ if (migratetype == MIGRATE_MOVABLE) return COMPACT_SUCCESS; /* * We are stealing for a non-movable allocation. Make * sure we finish compacting the current pageblock * first so it is as free as possible and we won't * have to steal another one soon. This only applies * to sync compaction, as async compaction operates * on pageblocks of the same migratetype. */ if (cc->mode == MIGRATE_ASYNC || IS_ALIGNED(cc->migrate_pfn, pageblock_nr_pages)) { return COMPACT_SUCCESS; } |
cb2dcaf02
|
2021 2022 |
ret = COMPACT_CONTINUE; break; |
baf6a9a1d
|
2023 |
} |
56de7263f
|
2024 |
} |
facdaa917
|
2025 |
out: |
cb2dcaf02
|
2026 2027 2028 2029 |
if (cc->contended || fatal_signal_pending(current)) ret = COMPACT_CONTENDED; return ret; |
837d026d5
|
2030 |
} |
40cacbcb3
|
2031 |
static enum compact_result compact_finished(struct compact_control *cc) |
837d026d5
|
2032 2033 |
{ int ret; |
40cacbcb3
|
2034 2035 |
ret = __compact_finished(cc); trace_mm_compaction_finished(cc->zone, cc->order, ret); |
837d026d5
|
2036 2037 2038 2039 |
if (ret == COMPACT_NO_SUITABLE_PAGE) ret = COMPACT_CONTINUE; return ret; |
748446bb6
|
2040 |
} |
ea7ab982b
|
2041 |
static enum compact_result __compaction_suitable(struct zone *zone, int order, |
c603844bd
|
2042 |
unsigned int alloc_flags, |
97a225e69
|
2043 |
int highest_zoneidx, |
86a294a81
|
2044 |
unsigned long wmark_target) |
3e7d34497
|
2045 |
{ |
3e7d34497
|
2046 |
unsigned long watermark; |
21c527a3c
|
2047 |
if (is_via_compact_memory(order)) |
3957c7768
|
2048 |
return COMPACT_CONTINUE; |
a92144438
|
2049 |
watermark = wmark_pages(zone, alloc_flags & ALLOC_WMARK_MASK); |
ebff39801
|
2050 2051 2052 2053 |
/* * If watermarks for high-order allocation are already met, there * should be no need for compaction at all. */ |
97a225e69
|
2054 |
if (zone_watermark_ok(zone, order, watermark, highest_zoneidx, |
ebff39801
|
2055 |
alloc_flags)) |
cf378319d
|
2056 |
return COMPACT_SUCCESS; |
ebff39801
|
2057 |
|
3957c7768
|
2058 |
/* |
9861a62c3
|
2059 |
* Watermarks for order-0 must be met for compaction to be able to |
984fdba6a
|
2060 2061 2062 2063 |
* isolate free pages for migration targets. This means that the * watermark and alloc_flags have to match, or be more pessimistic than * the check in __isolate_free_page(). We don't use the direct * compactor's alloc_flags, as they are not relevant for freepage |
97a225e69
|
2064 2065 2066 |
* isolation. We however do use the direct compactor's highest_zoneidx * to skip over zones where lowmem reserves would prevent allocation * even if compaction succeeds. |
8348faf91
|
2067 2068 |
* For costly orders, we require low watermark instead of min for * compaction to proceed to increase its chances. |
d883c6cf3
|
2069 2070 |
* ALLOC_CMA is used, as pages in CMA pageblocks are considered * suitable migration targets |
3e7d34497
|
2071 |
*/ |
8348faf91
|
2072 2073 2074 |
watermark = (order > PAGE_ALLOC_COSTLY_ORDER) ? low_wmark_pages(zone) : min_wmark_pages(zone); watermark += compact_gap(order); |
97a225e69
|
2075 |
if (!__zone_watermark_ok(zone, 0, watermark, highest_zoneidx, |
d883c6cf3
|
2076 |
ALLOC_CMA, wmark_target)) |
3e7d34497
|
2077 |
return COMPACT_SKIPPED; |
cc5c9f098
|
2078 2079 |
return COMPACT_CONTINUE; } |
2b1a20c3a
|
2080 2081 2082 2083 2084 2085 2086 |
/* * compaction_suitable: Is this suitable to run compaction on this zone now? * Returns * COMPACT_SKIPPED - If there are too few free pages for compaction * COMPACT_SUCCESS - If the allocation would succeed without compaction * COMPACT_CONTINUE - If compaction should run now */ |
cc5c9f098
|
2087 2088 |
enum compact_result compaction_suitable(struct zone *zone, int order, unsigned int alloc_flags, |
97a225e69
|
2089 |
int highest_zoneidx) |
cc5c9f098
|
2090 2091 2092 |
{ enum compact_result ret; int fragindex; |
97a225e69
|
2093 |
ret = __compaction_suitable(zone, order, alloc_flags, highest_zoneidx, |
cc5c9f098
|
2094 |
zone_page_state(zone, NR_FREE_PAGES)); |
3e7d34497
|
2095 2096 2097 2098 |
/* * fragmentation index determines if allocation failures are due to * low memory or external fragmentation * |
ebff39801
|
2099 2100 |
* index of -1000 would imply allocations might succeed depending on * watermarks, but we already failed the high-order watermark check |
3e7d34497
|
2101 2102 2103 |
* index towards 0 implies failure is due to lack of memory * index towards 1000 implies failure is due to fragmentation * |
203114202
|
2104 2105 2106 2107 2108 2109 |
* Only compact if a failure would be due to fragmentation. Also * ignore fragindex for non-costly orders where the alternative to * a successful reclaim/compaction is OOM. Fragindex and the * vm.extfrag_threshold sysctl is meant as a heuristic to prevent * excessive compaction for costly orders, but it should not be at the * expense of system stability. |
3e7d34497
|
2110 |
*/ |
203114202
|
2111 |
if (ret == COMPACT_CONTINUE && (order > PAGE_ALLOC_COSTLY_ORDER)) { |
cc5c9f098
|
2112 2113 2114 2115 |
fragindex = fragmentation_index(zone, order); if (fragindex >= 0 && fragindex <= sysctl_extfrag_threshold) ret = COMPACT_NOT_SUITABLE_ZONE; } |
837d026d5
|
2116 |
|
837d026d5
|
2117 2118 2119 2120 2121 2122 |
trace_mm_compaction_suitable(zone, order, ret); if (ret == COMPACT_NOT_SUITABLE_ZONE) ret = COMPACT_SKIPPED; return ret; } |
86a294a81
|
2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 |
bool compaction_zonelist_suitable(struct alloc_context *ac, int order, int alloc_flags) { struct zone *zone; struct zoneref *z; /* * Make sure at least one zone would pass __compaction_suitable if we continue * retrying the reclaim. */ |
97a225e69
|
2133 2134 |
for_each_zone_zonelist_nodemask(zone, z, ac->zonelist, ac->highest_zoneidx, ac->nodemask) { |
86a294a81
|
2135 2136 2137 2138 2139 2140 2141 2142 2143 |
unsigned long available; enum compact_result compact_result; /* * Do not consider all the reclaimable memory because we do not * want to trash just for a single high order allocation which * is even not guaranteed to appear even if __compaction_suitable * is happy about the watermark check. */ |
5a1c84b40
|
2144 |
available = zone_reclaimable_pages(zone) / order; |
86a294a81
|
2145 2146 |
available += zone_page_state_snapshot(zone, NR_FREE_PAGES); compact_result = __compaction_suitable(zone, order, alloc_flags, |
97a225e69
|
2147 |
ac->highest_zoneidx, available); |
cc5c9f098
|
2148 |
if (compact_result != COMPACT_SKIPPED) |
86a294a81
|
2149 2150 2151 2152 2153 |
return true; } return false; } |
5e1f0f098
|
2154 2155 |
static enum compact_result compact_zone(struct compact_control *cc, struct capture_control *capc) |
748446bb6
|
2156 |
{ |
ea7ab982b
|
2157 |
enum compact_result ret; |
40cacbcb3
|
2158 2159 |
unsigned long start_pfn = cc->zone->zone_start_pfn; unsigned long end_pfn = zone_end_pfn(cc->zone); |
566e54e11
|
2160 |
unsigned long last_migrated_pfn; |
e0b9daeb4
|
2161 |
const bool sync = cc->mode != MIGRATE_ASYNC; |
8854c55f5
|
2162 |
bool update_cached; |
748446bb6
|
2163 |
|
a94b52524
|
2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 |
/* * These counters track activities during zone compaction. Initialize * them before compacting a new zone. */ cc->total_migrate_scanned = 0; cc->total_free_scanned = 0; cc->nr_migratepages = 0; cc->nr_freepages = 0; INIT_LIST_HEAD(&cc->freepages); INIT_LIST_HEAD(&cc->migratepages); |
01c0bfe06
|
2174 |
cc->migratetype = gfp_migratetype(cc->gfp_mask); |
40cacbcb3
|
2175 |
ret = compaction_suitable(cc->zone, cc->order, cc->alloc_flags, |
97a225e69
|
2176 |
cc->highest_zoneidx); |
c46649dea
|
2177 |
/* Compaction is likely to fail */ |
cf378319d
|
2178 |
if (ret == COMPACT_SUCCESS || ret == COMPACT_SKIPPED) |
3e7d34497
|
2179 |
return ret; |
c46649dea
|
2180 2181 2182 |
/* huh, compaction_suitable is returning something unexpected */ VM_BUG_ON(ret != COMPACT_CONTINUE); |
3e7d34497
|
2183 |
|
c89511ab2
|
2184 |
/* |
d3132e4b8
|
2185 |
* Clear pageblock skip if there were failures recently and compaction |
accf62422
|
2186 |
* is about to be retried after being deferred. |
d3132e4b8
|
2187 |
*/ |
40cacbcb3
|
2188 2189 |
if (compaction_restarting(cc->zone, cc->order)) __reset_isolation_suitable(cc->zone); |
d3132e4b8
|
2190 2191 |
/* |
c89511ab2
|
2192 |
* Setup to move all movable pages to the end of the zone. Used cached |
06ed29989
|
2193 2194 2195 |
* information on where the scanners should start (unless we explicitly * want to compact the whole zone), but check that it is initialised * by ensuring the values are within zone boundaries. |
c89511ab2
|
2196 |
*/ |
70b44595e
|
2197 |
cc->fast_start_pfn = 0; |
06ed29989
|
2198 |
if (cc->whole_zone) { |
c89511ab2
|
2199 |
cc->migrate_pfn = start_pfn; |
06ed29989
|
2200 2201 |
cc->free_pfn = pageblock_start_pfn(end_pfn - 1); } else { |
40cacbcb3
|
2202 2203 |
cc->migrate_pfn = cc->zone->compact_cached_migrate_pfn[sync]; cc->free_pfn = cc->zone->compact_cached_free_pfn; |
06ed29989
|
2204 2205 |
if (cc->free_pfn < start_pfn || cc->free_pfn >= end_pfn) { cc->free_pfn = pageblock_start_pfn(end_pfn - 1); |
40cacbcb3
|
2206 |
cc->zone->compact_cached_free_pfn = cc->free_pfn; |
06ed29989
|
2207 2208 2209 |
} if (cc->migrate_pfn < start_pfn || cc->migrate_pfn >= end_pfn) { cc->migrate_pfn = start_pfn; |
40cacbcb3
|
2210 2211 |
cc->zone->compact_cached_migrate_pfn[0] = cc->migrate_pfn; cc->zone->compact_cached_migrate_pfn[1] = cc->migrate_pfn; |
06ed29989
|
2212 |
} |
c8f7de0bf
|
2213 |
|
e332f741a
|
2214 |
if (cc->migrate_pfn <= cc->zone->compact_init_migrate_pfn) |
06ed29989
|
2215 2216 |
cc->whole_zone = true; } |
c8f7de0bf
|
2217 |
|
566e54e11
|
2218 |
last_migrated_pfn = 0; |
748446bb6
|
2219 |
|
8854c55f5
|
2220 2221 2222 2223 2224 2225 2226 2227 2228 2229 |
/* * Migrate has separate cached PFNs for ASYNC and SYNC* migration on * the basis that some migrations will fail in ASYNC mode. However, * if the cached PFNs match and pageblocks are skipped due to having * no isolation candidates, then the sync state does not matter. * Until a pageblock with isolation candidates is found, keep the * cached PFNs in sync to avoid revisiting the same blocks. */ update_cached = !sync && cc->zone->compact_cached_migrate_pfn[0] == cc->zone->compact_cached_migrate_pfn[1]; |
16c4a097a
|
2230 2231 |
trace_mm_compaction_begin(start_pfn, cc->migrate_pfn, cc->free_pfn, end_pfn, sync); |
0eb927c0a
|
2232 |
|
361a2a229
|
2233 2234 |
/* lru_add_drain_all could be expensive with involving other CPUs */ lru_add_drain(); |
748446bb6
|
2235 |
|
40cacbcb3
|
2236 |
while ((ret = compact_finished(cc)) == COMPACT_CONTINUE) { |
9d502c1c8
|
2237 |
int err; |
19d3cf9de
|
2238 |
unsigned long iteration_start_pfn = cc->migrate_pfn; |
748446bb6
|
2239 |
|
804d3121b
|
2240 2241 2242 2243 2244 2245 2246 2247 2248 2249 |
/* * Avoid multiple rescans which can happen if a page cannot be * isolated (dirty/writeback in async mode) or if the migrated * pages are being allocated before the pageblock is cleared. * The first rescan will capture the entire pageblock for * migration. If it fails, it'll be marked skip and scanning * will proceed as normal. */ cc->rescan = false; if (pageblock_start_pfn(last_migrated_pfn) == |
19d3cf9de
|
2250 |
pageblock_start_pfn(iteration_start_pfn)) { |
804d3121b
|
2251 2252 |
cc->rescan = true; } |
32aaf0553
|
2253 |
switch (isolate_migratepages(cc)) { |
f9e35b3b4
|
2254 |
case ISOLATE_ABORT: |
2d1e10412
|
2255 |
ret = COMPACT_CONTENDED; |
5733c7d11
|
2256 |
putback_movable_pages(&cc->migratepages); |
e64c5237c
|
2257 |
cc->nr_migratepages = 0; |
f9e35b3b4
|
2258 2259 |
goto out; case ISOLATE_NONE: |
8854c55f5
|
2260 2261 2262 2263 |
if (update_cached) { cc->zone->compact_cached_migrate_pfn[1] = cc->zone->compact_cached_migrate_pfn[0]; } |
fdaf7f5c4
|
2264 2265 2266 2267 2268 2269 |
/* * We haven't isolated and migrated anything, but * there might still be unflushed migrations from * previous cc->order aligned block. */ goto check_drain; |
f9e35b3b4
|
2270 |
case ISOLATE_SUCCESS: |
8854c55f5
|
2271 |
update_cached = false; |
19d3cf9de
|
2272 |
last_migrated_pfn = iteration_start_pfn; |
f9e35b3b4
|
2273 |
} |
748446bb6
|
2274 |
|
d53aea3d4
|
2275 |
err = migrate_pages(&cc->migratepages, compaction_alloc, |
e0b9daeb4
|
2276 |
compaction_free, (unsigned long)cc, cc->mode, |
5ac95884a
|
2277 |
MR_COMPACTION, NULL); |
748446bb6
|
2278 |
|
f8c9301fa
|
2279 2280 |
trace_mm_compaction_migratepages(cc->nr_migratepages, err, &cc->migratepages); |
748446bb6
|
2281 |
|
f8c9301fa
|
2282 2283 |
/* All pages were either migrated or will be released */ cc->nr_migratepages = 0; |
9d502c1c8
|
2284 |
if (err) { |
5733c7d11
|
2285 |
putback_movable_pages(&cc->migratepages); |
7ed695e06
|
2286 2287 2288 2289 |
/* * migrate_pages() may return -ENOMEM when scanners meet * and we want compact_finished() to detect it */ |
f2849aa09
|
2290 |
if (err == -ENOMEM && !compact_scanners_met(cc)) { |
2d1e10412
|
2291 |
ret = COMPACT_CONTENDED; |
4bf2bba37
|
2292 2293 |
goto out; } |
fdd048e12
|
2294 2295 2296 2297 2298 2299 2300 2301 2302 |
/* * We failed to migrate at least one page in the current * order-aligned block, so skip the rest of it. */ if (cc->direct_compaction && (cc->mode == MIGRATE_ASYNC)) { cc->migrate_pfn = block_end_pfn( cc->migrate_pfn - 1, cc->order); /* Draining pcplists is useless in this case */ |
566e54e11
|
2303 |
last_migrated_pfn = 0; |
fdd048e12
|
2304 |
} |
748446bb6
|
2305 |
} |
fdaf7f5c4
|
2306 |
|
fdaf7f5c4
|
2307 2308 2309 2310 2311 2312 2313 2314 |
check_drain: /* * Has the migration scanner moved away from the previous * cc->order aligned block where we migrated from? If yes, * flush the pages that were freed, so that they can merge and * compact_finished() can detect immediately if allocation * would succeed. */ |
566e54e11
|
2315 |
if (cc->order > 0 && last_migrated_pfn) { |
fdaf7f5c4
|
2316 |
unsigned long current_block_start = |
06b6640a3
|
2317 |
block_start_pfn(cc->migrate_pfn, cc->order); |
fdaf7f5c4
|
2318 |
|
566e54e11
|
2319 |
if (last_migrated_pfn < current_block_start) { |
b01b21419
|
2320 |
lru_add_drain_cpu_zone(cc->zone); |
fdaf7f5c4
|
2321 |
/* No more flushing until we migrate again */ |
566e54e11
|
2322 |
last_migrated_pfn = 0; |
fdaf7f5c4
|
2323 2324 |
} } |
5e1f0f098
|
2325 2326 2327 2328 2329 |
/* Stop if a page has been captured */ if (capc && capc->page) { ret = COMPACT_SUCCESS; break; } |
748446bb6
|
2330 |
} |
f9e35b3b4
|
2331 |
out: |
6bace090a
|
2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 |
/* * Release free pages and update where the free scanner should restart, * so we don't leave any returned pages behind in the next attempt. */ if (cc->nr_freepages > 0) { unsigned long free_pfn = release_freepages(&cc->freepages); cc->nr_freepages = 0; VM_BUG_ON(free_pfn == 0); /* The cached pfn is always the first in a pageblock */ |
06b6640a3
|
2342 |
free_pfn = pageblock_start_pfn(free_pfn); |
6bace090a
|
2343 2344 2345 2346 |
/* * Only go back, not forward. The cached pfn might have been * already reset to zone end in compact_finished() */ |
40cacbcb3
|
2347 2348 |
if (free_pfn > cc->zone->compact_cached_free_pfn) cc->zone->compact_cached_free_pfn = free_pfn; |
6bace090a
|
2349 |
} |
748446bb6
|
2350 |
|
7f354a548
|
2351 2352 |
count_compact_events(COMPACTMIGRATE_SCANNED, cc->total_migrate_scanned); count_compact_events(COMPACTFREE_SCANNED, cc->total_free_scanned); |
16c4a097a
|
2353 2354 |
trace_mm_compaction_end(start_pfn, cc->migrate_pfn, cc->free_pfn, end_pfn, sync, ret); |
0eb927c0a
|
2355 |
|
748446bb6
|
2356 2357 |
return ret; } |
76ab0f530
|
2358 |
|
ea7ab982b
|
2359 |
static enum compact_result compact_zone_order(struct zone *zone, int order, |
c3486f537
|
2360 |
gfp_t gfp_mask, enum compact_priority prio, |
97a225e69
|
2361 |
unsigned int alloc_flags, int highest_zoneidx, |
5e1f0f098
|
2362 |
struct page **capture) |
56de7263f
|
2363 |
{ |
ea7ab982b
|
2364 |
enum compact_result ret; |
56de7263f
|
2365 |
struct compact_control cc = { |
56de7263f
|
2366 |
.order = order, |
dbe2d4e4f
|
2367 |
.search_order = order, |
6d7ce5594
|
2368 |
.gfp_mask = gfp_mask, |
56de7263f
|
2369 |
.zone = zone, |
a5508cd83
|
2370 2371 |
.mode = (prio == COMPACT_PRIO_ASYNC) ? MIGRATE_ASYNC : MIGRATE_SYNC_LIGHT, |
ebff39801
|
2372 |
.alloc_flags = alloc_flags, |
97a225e69
|
2373 |
.highest_zoneidx = highest_zoneidx, |
accf62422
|
2374 |
.direct_compaction = true, |
a8e025e55
|
2375 |
.whole_zone = (prio == MIN_COMPACT_PRIORITY), |
9f7e33879
|
2376 2377 |
.ignore_skip_hint = (prio == MIN_COMPACT_PRIORITY), .ignore_block_suitable = (prio == MIN_COMPACT_PRIORITY) |
56de7263f
|
2378 |
}; |
5e1f0f098
|
2379 2380 2381 2382 |
struct capture_control capc = { .cc = &cc, .page = NULL, }; |
b9e20f0da
|
2383 2384 2385 2386 2387 2388 2389 |
/* * Make sure the structs are really initialized before we expose the * capture control, in case we are interrupted and the interrupt handler * frees a page. */ barrier(); WRITE_ONCE(current->capture_control, &capc); |
56de7263f
|
2390 |
|
5e1f0f098
|
2391 |
ret = compact_zone(&cc, &capc); |
e64c5237c
|
2392 2393 2394 |
VM_BUG_ON(!list_empty(&cc.freepages)); VM_BUG_ON(!list_empty(&cc.migratepages)); |
b9e20f0da
|
2395 2396 2397 2398 2399 2400 2401 |
/* * Make sure we hide capture control first before we read the captured * page pointer, otherwise an interrupt could free and capture a page * and we would leak it. */ WRITE_ONCE(current->capture_control, NULL); *capture = READ_ONCE(capc.page); |
06dac2f46
|
2402 2403 2404 2405 2406 2407 2408 2409 |
/* * Technically, it is also possible that compaction is skipped but * the page is still captured out of luck(IRQ came and freed the page). * Returning COMPACT_SUCCESS in such cases helps in properly accounting * the COMPACT[STALL|FAIL] when compaction is skipped. */ if (*capture) ret = COMPACT_SUCCESS; |
5e1f0f098
|
2410 |
|
e64c5237c
|
2411 |
return ret; |
56de7263f
|
2412 |
} |
5e7719058
|
2413 |
int sysctl_extfrag_threshold = 500; |
56de7263f
|
2414 2415 |
/** * try_to_compact_pages - Direct compact to satisfy a high-order allocation |
56de7263f
|
2416 |
* @gfp_mask: The GFP mask of the current allocation |
1a6d53a10
|
2417 2418 2419 |
* @order: The order of the current allocation * @alloc_flags: The allocation flags of the current allocation * @ac: The context of current allocation |
112d2d29f
|
2420 |
* @prio: Determines how hard direct compaction should try to succeed |
6467552ca
|
2421 |
* @capture: Pointer to free page created by compaction will be stored here |
56de7263f
|
2422 2423 2424 |
* * This is the main entry point for direct page compaction. */ |
ea7ab982b
|
2425 |
enum compact_result try_to_compact_pages(gfp_t gfp_mask, unsigned int order, |
c603844bd
|
2426 |
unsigned int alloc_flags, const struct alloc_context *ac, |
5e1f0f098
|
2427 |
enum compact_priority prio, struct page **capture) |
56de7263f
|
2428 |
{ |
56de7263f
|
2429 |
int may_perform_io = gfp_mask & __GFP_IO; |
56de7263f
|
2430 2431 |
struct zoneref *z; struct zone *zone; |
1d4746d39
|
2432 |
enum compact_result rc = COMPACT_SKIPPED; |
56de7263f
|
2433 |
|
73e64c51a
|
2434 2435 2436 2437 2438 |
/* * Check if the GFP flags allow compaction - GFP_NOIO is really * tricky context because the migration might require IO */ if (!may_perform_io) |
53853e2d2
|
2439 |
return COMPACT_SKIPPED; |
56de7263f
|
2440 |
|
a5508cd83
|
2441 |
trace_mm_compaction_try_to_compact_pages(order, gfp_mask, prio); |
837d026d5
|
2442 |
|
56de7263f
|
2443 |
/* Compact each zone in the list */ |
97a225e69
|
2444 2445 |
for_each_zone_zonelist_nodemask(zone, z, ac->zonelist, ac->highest_zoneidx, ac->nodemask) { |
ea7ab982b
|
2446 |
enum compact_result status; |
56de7263f
|
2447 |
|
a8e025e55
|
2448 2449 |
if (prio > MIN_COMPACT_PRIORITY && compaction_deferred(zone, order)) { |
1d4746d39
|
2450 |
rc = max_t(enum compact_result, COMPACT_DEFERRED, rc); |
53853e2d2
|
2451 |
continue; |
1d4746d39
|
2452 |
} |
53853e2d2
|
2453 |
|
a5508cd83
|
2454 |
status = compact_zone_order(zone, order, gfp_mask, prio, |
97a225e69
|
2455 |
alloc_flags, ac->highest_zoneidx, capture); |
56de7263f
|
2456 |
rc = max(status, rc); |
7ceb009a2
|
2457 2458 |
/* The allocation should succeed, stop compacting */ if (status == COMPACT_SUCCESS) { |
53853e2d2
|
2459 2460 2461 2462 2463 2464 2465 |
/* * We think the allocation will succeed in this zone, * but it is not certain, hence the false. The caller * will repeat this with true if allocation indeed * succeeds in this zone. */ compaction_defer_reset(zone, order, false); |
1f9efdef4
|
2466 |
|
c3486f537
|
2467 |
break; |
1f9efdef4
|
2468 |
} |
a5508cd83
|
2469 |
if (prio != COMPACT_PRIO_ASYNC && (status == COMPACT_COMPLETE || |
c3486f537
|
2470 |
status == COMPACT_PARTIAL_SKIPPED)) |
53853e2d2
|
2471 2472 2473 2474 2475 2476 |
/* * We think that allocation won't succeed in this zone * so we defer compaction there. If it ends up * succeeding after all, it will be reset. */ defer_compaction(zone, order); |
1f9efdef4
|
2477 2478 2479 2480 |
/* * We might have stopped compacting due to need_resched() in * async compaction, or due to a fatal signal detected. In that |
c3486f537
|
2481 |
* case do not try further zones |
1f9efdef4
|
2482 |
*/ |
c3486f537
|
2483 2484 2485 |
if ((prio == COMPACT_PRIO_ASYNC && need_resched()) || fatal_signal_pending(current)) break; |
56de7263f
|
2486 2487 2488 2489 |
} return rc; } |
facdaa917
|
2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 2506 2507 2508 2509 2510 2511 2512 2513 2514 2515 2516 2517 2518 2519 2520 2521 2522 2523 2524 |
/* * Compact all zones within a node till each zone's fragmentation score * reaches within proactive compaction thresholds (as determined by the * proactiveness tunable). * * It is possible that the function returns before reaching score targets * due to various back-off conditions, such as, contention on per-node or * per-zone locks. */ static void proactive_compact_node(pg_data_t *pgdat) { int zoneid; struct zone *zone; struct compact_control cc = { .order = -1, .mode = MIGRATE_SYNC_LIGHT, .ignore_skip_hint = true, .whole_zone = true, .gfp_mask = GFP_KERNEL, .proactive_compaction = true, }; for (zoneid = 0; zoneid < MAX_NR_ZONES; zoneid++) { zone = &pgdat->node_zones[zoneid]; if (!populated_zone(zone)) continue; cc.zone = zone; compact_zone(&cc, NULL); VM_BUG_ON(!list_empty(&cc.freepages)); VM_BUG_ON(!list_empty(&cc.migratepages)); } } |
56de7263f
|
2525 |
|
76ab0f530
|
2526 |
/* Compact all zones within a node */ |
791cae962
|
2527 |
static void compact_node(int nid) |
76ab0f530
|
2528 |
{ |
791cae962
|
2529 |
pg_data_t *pgdat = NODE_DATA(nid); |
76ab0f530
|
2530 |
int zoneid; |
76ab0f530
|
2531 |
struct zone *zone; |
791cae962
|
2532 2533 2534 2535 2536 |
struct compact_control cc = { .order = -1, .mode = MIGRATE_SYNC, .ignore_skip_hint = true, .whole_zone = true, |
73e64c51a
|
2537 |
.gfp_mask = GFP_KERNEL, |
791cae962
|
2538 |
}; |
76ab0f530
|
2539 |
|
76ab0f530
|
2540 |
for (zoneid = 0; zoneid < MAX_NR_ZONES; zoneid++) { |
76ab0f530
|
2541 2542 2543 2544 |
zone = &pgdat->node_zones[zoneid]; if (!populated_zone(zone)) continue; |
791cae962
|
2545 |
cc.zone = zone; |
76ab0f530
|
2546 |
|
5e1f0f098
|
2547 |
compact_zone(&cc, NULL); |
754693457
|
2548 |
|
791cae962
|
2549 2550 |
VM_BUG_ON(!list_empty(&cc.freepages)); VM_BUG_ON(!list_empty(&cc.migratepages)); |
76ab0f530
|
2551 |
} |
76ab0f530
|
2552 2553 2554 |
} /* Compact all nodes in the system */ |
7964c06d6
|
2555 |
static void compact_nodes(void) |
76ab0f530
|
2556 2557 |
{ int nid; |
8575ec29f
|
2558 2559 |
/* Flush pending updates to the LRU lists */ lru_add_drain_all(); |
76ab0f530
|
2560 2561 |
for_each_online_node(nid) compact_node(nid); |
76ab0f530
|
2562 |
} |
fec4eb2c8
|
2563 |
/* |
facdaa917
|
2564 2565 2566 2567 |
* Tunable for proactive compaction. It determines how * aggressively the kernel should compact memory in the * background. It takes values in the range [0, 100]. */ |
d34c0a759
|
2568 |
unsigned int __read_mostly sysctl_compaction_proactiveness = 20; |
facdaa917
|
2569 |
|
65d759c8f
|
2570 2571 2572 2573 2574 2575 2576 2577 2578 2579 2580 2581 2582 2583 2584 2585 2586 2587 2588 2589 2590 2591 2592 |
int compaction_proactiveness_sysctl_handler(struct ctl_table *table, int write, void *buffer, size_t *length, loff_t *ppos) { int rc, nid; rc = proc_dointvec_minmax(table, write, buffer, length, ppos); if (rc) return rc; if (write && sysctl_compaction_proactiveness) { for_each_online_node(nid) { pg_data_t *pgdat = NODE_DATA(nid); if (pgdat->proactive_compact_trigger) continue; pgdat->proactive_compact_trigger = true; wake_up_interruptible(&pgdat->kcompactd_wait); } } return 0; } |
facdaa917
|
2593 |
/* |
fec4eb2c8
|
2594 2595 2596 |
* This is the entry point for compacting all nodes via * /proc/sys/vm/compact_memory */ |
76ab0f530
|
2597 |
int sysctl_compaction_handler(struct ctl_table *table, int write, |
32927393d
|
2598 |
void *buffer, size_t *length, loff_t *ppos) |
76ab0f530
|
2599 2600 |
{ if (write) |
7964c06d6
|
2601 |
compact_nodes(); |
76ab0f530
|
2602 2603 2604 |
return 0; } |
ed4a6d7f0
|
2605 2606 |
#if defined(CONFIG_SYSFS) && defined(CONFIG_NUMA) |
17adb230d
|
2607 2608 2609 |
static ssize_t compact_store(struct device *dev, struct device_attribute *attr, const char *buf, size_t count) |
ed4a6d7f0
|
2610 |
{ |
8575ec29f
|
2611 2612 2613 2614 2615 2616 2617 2618 |
int nid = dev->id; if (nid >= 0 && nid < nr_node_ids && node_online(nid)) { /* Flush pending updates to the LRU lists */ lru_add_drain_all(); compact_node(nid); } |
ed4a6d7f0
|
2619 2620 2621 |
return count; } |
17adb230d
|
2622 |
static DEVICE_ATTR_WO(compact); |
ed4a6d7f0
|
2623 2624 2625 |
int compaction_register_node(struct node *node) { |
10fbcf4c6
|
2626 |
return device_create_file(&node->dev, &dev_attr_compact); |
ed4a6d7f0
|
2627 2628 2629 2630 |
} void compaction_unregister_node(struct node *node) { |
10fbcf4c6
|
2631 |
return device_remove_file(&node->dev, &dev_attr_compact); |
ed4a6d7f0
|
2632 2633 |
} #endif /* CONFIG_SYSFS && CONFIG_NUMA */ |
ff9543fd3
|
2634 |
|
698b1b306
|
2635 2636 |
static inline bool kcompactd_work_requested(pg_data_t *pgdat) { |
65d759c8f
|
2637 2638 |
return pgdat->kcompactd_max_order > 0 || kthread_should_stop() || pgdat->proactive_compact_trigger; |
698b1b306
|
2639 2640 2641 2642 2643 2644 |
} static bool kcompactd_node_suitable(pg_data_t *pgdat) { int zoneid; struct zone *zone; |
97a225e69
|
2645 |
enum zone_type highest_zoneidx = pgdat->kcompactd_highest_zoneidx; |
698b1b306
|
2646 |
|
97a225e69
|
2647 |
for (zoneid = 0; zoneid <= highest_zoneidx; zoneid++) { |
698b1b306
|
2648 2649 2650 2651 2652 2653 |
zone = &pgdat->node_zones[zoneid]; if (!populated_zone(zone)) continue; if (compaction_suitable(zone, pgdat->kcompactd_max_order, 0, |
97a225e69
|
2654 |
highest_zoneidx) == COMPACT_CONTINUE) |
698b1b306
|
2655 2656 2657 2658 2659 2660 2661 2662 2663 2664 2665 2666 2667 2668 2669 2670 |
return true; } return false; } static void kcompactd_do_work(pg_data_t *pgdat) { /* * With no special task, compact all zones so that a page of requested * order is allocatable. */ int zoneid; struct zone *zone; struct compact_control cc = { .order = pgdat->kcompactd_max_order, |
dbe2d4e4f
|
2671 |
.search_order = pgdat->kcompactd_max_order, |
97a225e69
|
2672 |
.highest_zoneidx = pgdat->kcompactd_highest_zoneidx, |
698b1b306
|
2673 |
.mode = MIGRATE_SYNC_LIGHT, |
a0647dc92
|
2674 |
.ignore_skip_hint = false, |
73e64c51a
|
2675 |
.gfp_mask = GFP_KERNEL, |
698b1b306
|
2676 |
}; |
698b1b306
|
2677 |
trace_mm_compaction_kcompactd_wake(pgdat->node_id, cc.order, |
97a225e69
|
2678 |
cc.highest_zoneidx); |
7f354a548
|
2679 |
count_compact_event(KCOMPACTD_WAKE); |
698b1b306
|
2680 |
|
97a225e69
|
2681 |
for (zoneid = 0; zoneid <= cc.highest_zoneidx; zoneid++) { |
698b1b306
|
2682 2683 2684 2685 2686 2687 2688 2689 2690 2691 2692 2693 |
int status; zone = &pgdat->node_zones[zoneid]; if (!populated_zone(zone)) continue; if (compaction_deferred(zone, cc.order)) continue; if (compaction_suitable(zone, cc.order, 0, zoneid) != COMPACT_CONTINUE) continue; |
172400c69
|
2694 2695 |
if (kthread_should_stop()) return; |
a94b52524
|
2696 2697 |
cc.zone = zone; |
5e1f0f098
|
2698 |
status = compact_zone(&cc, NULL); |
698b1b306
|
2699 |
|
7ceb009a2
|
2700 |
if (status == COMPACT_SUCCESS) { |
698b1b306
|
2701 |
compaction_defer_reset(zone, cc.order, false); |
c8f7de0bf
|
2702 |
} else if (status == COMPACT_PARTIAL_SKIPPED || status == COMPACT_COMPLETE) { |
698b1b306
|
2703 |
/* |
bc3106b26
|
2704 2705 2706 2707 2708 2709 2710 2711 |
* Buddy pages may become stranded on pcps that could * otherwise coalesce on the zone's free area for * order >= cc.order. This is ratelimited by the * upcoming deferral. */ drain_all_pages(zone); /* |
698b1b306
|
2712 2713 2714 2715 2716 |
* We use sync migration mode here, so we defer like * sync direct compaction does. */ defer_compaction(zone, cc.order); } |
7f354a548
|
2717 2718 2719 2720 |
count_compact_events(KCOMPACTD_MIGRATE_SCANNED, cc.total_migrate_scanned); count_compact_events(KCOMPACTD_FREE_SCANNED, cc.total_free_scanned); |
698b1b306
|
2721 2722 2723 2724 2725 2726 |
VM_BUG_ON(!list_empty(&cc.freepages)); VM_BUG_ON(!list_empty(&cc.migratepages)); } /* * Regardless of success, we are done until woken up next. But remember |
97a225e69
|
2727 2728 |
* the requested order/highest_zoneidx in case it was higher/tighter * than our current ones |
698b1b306
|
2729 2730 2731 |
*/ if (pgdat->kcompactd_max_order <= cc.order) pgdat->kcompactd_max_order = 0; |
97a225e69
|
2732 2733 |
if (pgdat->kcompactd_highest_zoneidx >= cc.highest_zoneidx) pgdat->kcompactd_highest_zoneidx = pgdat->nr_zones - 1; |
698b1b306
|
2734 |
} |
97a225e69
|
2735 |
void wakeup_kcompactd(pg_data_t *pgdat, int order, int highest_zoneidx) |
698b1b306
|
2736 2737 2738 2739 2740 2741 |
{ if (!order) return; if (pgdat->kcompactd_max_order < order) pgdat->kcompactd_max_order = order; |
97a225e69
|
2742 2743 |
if (pgdat->kcompactd_highest_zoneidx > highest_zoneidx) pgdat->kcompactd_highest_zoneidx = highest_zoneidx; |
698b1b306
|
2744 |
|
6818600ff
|
2745 2746 2747 2748 2749 |
/* * Pairs with implicit barrier in wait_event_freezable() * such that wakeups are not missed. */ if (!wq_has_sleeper(&pgdat->kcompactd_wait)) |
698b1b306
|
2750 2751 2752 2753 2754 2755 |
return; if (!kcompactd_node_suitable(pgdat)) return; trace_mm_compaction_wakeup_kcompactd(pgdat->node_id, order, |
97a225e69
|
2756 |
highest_zoneidx); |
698b1b306
|
2757 2758 2759 2760 2761 2762 2763 2764 2765 |
wake_up_interruptible(&pgdat->kcompactd_wait); } /* * The background compaction daemon, started as a kernel thread * from the init process. */ static int kcompactd(void *p) { |
68d68ff6e
|
2766 |
pg_data_t *pgdat = (pg_data_t *)p; |
698b1b306
|
2767 |
struct task_struct *tsk = current; |
e1e92bfa3
|
2768 2769 |
long default_timeout = msecs_to_jiffies(HPAGE_FRAG_CHECK_INTERVAL_MSEC); long timeout = default_timeout; |
698b1b306
|
2770 2771 2772 2773 2774 2775 2776 2777 2778 |
const struct cpumask *cpumask = cpumask_of_node(pgdat->node_id); if (!cpumask_empty(cpumask)) set_cpus_allowed_ptr(tsk, cpumask); set_freezable(); pgdat->kcompactd_max_order = 0; |
97a225e69
|
2779 |
pgdat->kcompactd_highest_zoneidx = pgdat->nr_zones - 1; |
698b1b306
|
2780 2781 |
while (!kthread_should_stop()) { |
eb414681d
|
2782 |
unsigned long pflags; |
65d759c8f
|
2783 2784 2785 2786 2787 2788 |
/* * Avoid the unnecessary wakeup for proactive compaction * when it is disabled. */ if (!sysctl_compaction_proactiveness) timeout = MAX_SCHEDULE_TIMEOUT; |
698b1b306
|
2789 |
trace_mm_compaction_kcompactd_sleep(pgdat->node_id); |
facdaa917
|
2790 |
if (wait_event_freezable_timeout(pgdat->kcompactd_wait, |
65d759c8f
|
2791 2792 |
kcompactd_work_requested(pgdat), timeout) && !pgdat->proactive_compact_trigger) { |
facdaa917
|
2793 2794 2795 2796 |
psi_memstall_enter(&pflags); kcompactd_do_work(pgdat); psi_memstall_leave(&pflags); |
e1e92bfa3
|
2797 2798 2799 2800 2801 2802 2803 2804 |
/* * Reset the timeout value. The defer timeout from * proactive compaction is lost here but that is fine * as the condition of the zone changing substantionally * then carrying on with the previous defer interval is * not useful. */ timeout = default_timeout; |
facdaa917
|
2805 2806 |
continue; } |
698b1b306
|
2807 |
|
e1e92bfa3
|
2808 2809 2810 2811 2812 |
/* * Start the proactive work with default timeout. Based * on the fragmentation score, this timeout is updated. */ timeout = default_timeout; |
facdaa917
|
2813 2814 |
if (should_proactive_compact_node(pgdat)) { unsigned int prev_score, score; |
facdaa917
|
2815 2816 2817 2818 2819 2820 2821 |
prev_score = fragmentation_score_node(pgdat); proactive_compact_node(pgdat); score = fragmentation_score_node(pgdat); /* * Defer proactive compaction if the fragmentation * score did not go down i.e. no progress made. */ |
e1e92bfa3
|
2822 2823 2824 |
if (unlikely(score >= prev_score)) timeout = default_timeout << COMPACT_MAX_DEFER_SHIFT; |
facdaa917
|
2825 |
} |
65d759c8f
|
2826 2827 |
if (unlikely(pgdat->proactive_compact_trigger)) pgdat->proactive_compact_trigger = false; |
698b1b306
|
2828 2829 2830 2831 2832 2833 2834 2835 2836 2837 2838 2839 2840 2841 2842 2843 2844 2845 2846 2847 2848 2849 2850 2851 2852 2853 2854 2855 2856 2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 2867 2868 2869 2870 2871 2872 2873 2874 |
} return 0; } /* * This kcompactd start function will be called by init and node-hot-add. * On node-hot-add, kcompactd will moved to proper cpus if cpus are hot-added. */ int kcompactd_run(int nid) { pg_data_t *pgdat = NODE_DATA(nid); int ret = 0; if (pgdat->kcompactd) return 0; pgdat->kcompactd = kthread_run(kcompactd, pgdat, "kcompactd%d", nid); if (IS_ERR(pgdat->kcompactd)) { pr_err("Failed to start kcompactd on node %d ", nid); ret = PTR_ERR(pgdat->kcompactd); pgdat->kcompactd = NULL; } return ret; } /* * Called by memory hotplug when all memory in a node is offlined. Caller must * hold mem_hotplug_begin/end(). */ void kcompactd_stop(int nid) { struct task_struct *kcompactd = NODE_DATA(nid)->kcompactd; if (kcompactd) { kthread_stop(kcompactd); NODE_DATA(nid)->kcompactd = NULL; } } /* * It's optimal to keep kcompactd 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. */ |
e46b1db24
|
2875 |
static int kcompactd_cpu_online(unsigned int cpu) |
698b1b306
|
2876 2877 |
{ int nid; |
e46b1db24
|
2878 2879 2880 |
for_each_node_state(nid, N_MEMORY) { pg_data_t *pgdat = NODE_DATA(nid); const struct cpumask *mask; |
698b1b306
|
2881 |
|
e46b1db24
|
2882 |
mask = cpumask_of_node(pgdat->node_id); |
698b1b306
|
2883 |
|
e46b1db24
|
2884 2885 2886 |
if (cpumask_any_and(cpu_online_mask, mask) < nr_cpu_ids) /* One of our CPUs online: restore mask */ set_cpus_allowed_ptr(pgdat->kcompactd, mask); |
698b1b306
|
2887 |
} |
e46b1db24
|
2888 |
return 0; |
698b1b306
|
2889 2890 2891 2892 2893 |
} static int __init kcompactd_init(void) { int nid; |
e46b1db24
|
2894 2895 2896 2897 2898 2899 2900 2901 2902 2903 |
int ret; ret = cpuhp_setup_state_nocalls(CPUHP_AP_ONLINE_DYN, "mm/compaction:online", kcompactd_cpu_online, NULL); if (ret < 0) { pr_err("kcompactd: failed to register hotplug callbacks. "); return ret; } |
698b1b306
|
2904 2905 2906 |
for_each_node_state(nid, N_MEMORY) kcompactd_run(nid); |
698b1b306
|
2907 2908 2909 |
return 0; } subsys_initcall(kcompactd_init) |
ff9543fd3
|
2910 |
#endif /* CONFIG_COMPACTION */ |