Blame view
mm/compaction.c
80.4 KB
b24413180 License cleanup: ... |
1 |
// SPDX-License-Identifier: GPL-2.0 |
748446bb6 mm: compaction: m... |
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 mm, compaction: i... |
11 |
#include <linux/cpu.h> |
748446bb6 mm: compaction: m... |
12 13 14 15 |
#include <linux/swap.h> #include <linux/migrate.h> #include <linux/compaction.h> #include <linux/mm_inline.h> |
174cd4b1e sched/headers: Pr... |
16 |
#include <linux/sched/signal.h> |
748446bb6 mm: compaction: m... |
17 |
#include <linux/backing-dev.h> |
76ab0f530 mm: compaction: a... |
18 |
#include <linux/sysctl.h> |
ed4a6d7f0 mm: compaction: a... |
19 |
#include <linux/sysfs.h> |
194159fbc mm: remove MIGRAT... |
20 |
#include <linux/page-isolation.h> |
b8c73fc24 mm: page_alloc: a... |
21 |
#include <linux/kasan.h> |
698b1b306 mm, compaction: i... |
22 23 |
#include <linux/kthread.h> #include <linux/freezer.h> |
83358ece2 mm/page_owner: in... |
24 |
#include <linux/page_owner.h> |
eb414681d psi: pressure sta... |
25 |
#include <linux/psi.h> |
748446bb6 mm: compaction: m... |
26 |
#include "internal.h" |
010fc29a4 compaction: fix b... |
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 mm: compaction: e... |
41 |
#if defined CONFIG_COMPACTION || defined CONFIG_CMA |
b7aba6984 mm: compaction: a... |
42 43 |
#define CREATE_TRACE_POINTS #include <trace/events/compaction.h> |
06b6640a3 mm, compaction: w... |
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 mm: proactive com... |
48 49 50 |
/* * Fragmentation score check interval for proactive compaction purposes. */ |
d34c0a759 mm: use unsigned ... |
51 |
static const unsigned int HPAGE_FRAG_CHECK_INTERVAL_MSEC = 500; |
facdaa917 mm: proactive com... |
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 mm: fix compile e... |
60 |
#elif defined CONFIG_HUGETLBFS |
facdaa917 mm: proactive com... |
61 62 63 64 |
#define COMPACTION_HPAGE_ORDER HUGETLB_PAGE_ORDER #else #define COMPACTION_HPAGE_ORDER (PMD_SHIFT - PAGE_SHIFT) #endif |
748446bb6 mm: compaction: m... |
65 66 67 |
static unsigned long release_freepages(struct list_head *freelist) { struct page *page, *next; |
6bace090a mm, compaction: a... |
68 |
unsigned long high_pfn = 0; |
748446bb6 mm: compaction: m... |
69 70 |
list_for_each_entry_safe(page, next, freelist, lru) { |
6bace090a mm, compaction: a... |
71 |
unsigned long pfn = page_to_pfn(page); |
748446bb6 mm: compaction: m... |
72 73 |
list_del(&page->lru); __free_page(page); |
6bace090a mm, compaction: a... |
74 75 |
if (pfn > high_pfn) high_pfn = pfn; |
748446bb6 mm: compaction: m... |
76 |
} |
6bace090a mm, compaction: a... |
77 |
return high_pfn; |
748446bb6 mm: compaction: m... |
78 |
} |
4469ab984 mm, compaction: r... |
79 |
static void split_map_pages(struct list_head *list) |
ff9543fd3 mm: compaction: e... |
80 |
{ |
66c64223a mm/compaction: sp... |
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 mm/compaction: sp... |
90 |
|
46f24fd85 mm/page_alloc: in... |
91 |
post_alloc_hook(page, order, __GFP_MOVABLE); |
66c64223a mm/compaction: sp... |
92 93 |
if (order) split_page(page, order); |
ff9543fd3 mm: compaction: e... |
94 |
|
66c64223a mm/compaction: sp... |
95 96 97 98 |
for (i = 0; i < nr_pages; i++) { list_add(&page->lru, &tmp_list); page++; } |
ff9543fd3 mm: compaction: e... |
99 |
} |
66c64223a mm/compaction: sp... |
100 101 |
list_splice(&tmp_list, list); |
ff9543fd3 mm: compaction: e... |
102 |
} |
bb13ffeb9 mm: compaction: c... |
103 |
#ifdef CONFIG_COMPACTION |
24e2716f6 mm/compaction: ad... |
104 |
|
bda807d44 mm: migrate: supp... |
105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 |
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) { VM_BUG_ON_PAGE(!PageLocked(page), page); 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 mm/compaction: ad... |
142 143 144 145 146 |
/* 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 mm/compaction: co... |
147 |
* allocation success. 1 << compact_defer_shift, compactions are skipped up |
24e2716f6 mm/compaction: ad... |
148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 |
* to a limit of 1 << COMPACT_MAX_DEFER_SHIFT */ void defer_compaction(struct zone *zone, int order) { 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 */ bool compaction_deferred(struct zone *zone, int order) { unsigned long defer_limit = 1UL << zone->compact_defer_shift; if (order < zone->compact_order_failed) return false; /* Avoid possible overflow */ |
62b35fe0e mm/compaction.c: ... |
173 |
if (++zone->compact_considered >= defer_limit) { |
24e2716f6 mm/compaction: ad... |
174 |
zone->compact_considered = defer_limit; |
24e2716f6 mm/compaction: ad... |
175 |
return false; |
62b35fe0e mm/compaction.c: ... |
176 |
} |
24e2716f6 mm/compaction: ad... |
177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 |
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 */ bool compaction_restarting(struct zone *zone, int order) { 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 mm: compaction: c... |
210 211 212 213 214 215 216 217 218 |
/* 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 mm, compaction: e... |
219 220 221 222 |
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 mm/compaction: fi... |
223 |
zone->compact_cached_free_pfn = |
06b6640a3 mm, compaction: w... |
224 |
pageblock_start_pfn(zone_end_pfn(zone) - 1); |
02333641e mm, compaction: e... |
225 |
} |
bb13ffeb9 mm: compaction: c... |
226 |
/* |
b527cfe5b mm, compaction: e... |
227 228 229 |
* Compound pages of >= pageblock_order should consistenly be skipped until * 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 mm, compaction: p... |
230 |
*/ |
b527cfe5b mm, compaction: e... |
231 |
static bool pageblock_skip_persistent(struct page *page) |
21dc7e023 mm, compaction: p... |
232 |
{ |
b527cfe5b mm, compaction: e... |
233 |
if (!PageCompound(page)) |
21dc7e023 mm, compaction: p... |
234 |
return false; |
b527cfe5b mm, compaction: e... |
235 236 237 238 239 240 241 |
page = compound_head(page); if (compound_order(page) >= pageblock_order) return true; return false; |
21dc7e023 mm, compaction: p... |
242 |
} |
e332f741a mm, compaction: b... |
243 244 245 246 247 |
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 mm/compaction.c: ... |
248 |
struct page *block_page; |
e332f741a mm, compaction: b... |
249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 |
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 mm/compaction.c: ... |
273 274 |
/* Ensure the start of the pageblock or zone is online and valid */ block_pfn = pageblock_start_pfn(pfn); |
a2e9a5afc mm, compaction: f... |
275 276 |
block_pfn = max(block_pfn, zone->zone_start_pfn); block_page = pfn_to_online_page(block_pfn); |
6b0868c82 mm/compaction.c: ... |
277 278 279 280 281 282 |
if (block_page) { page = block_page; pfn = block_pfn; } /* Ensure the end of the pageblock or zone is online and valid */ |
a2e9a5afc mm, compaction: f... |
283 |
block_pfn = pageblock_end_pfn(pfn) - 1; |
6b0868c82 mm/compaction.c: ... |
284 285 286 287 |
block_pfn = min(block_pfn, zone_end_pfn(zone) - 1); end_page = pfn_to_online_page(block_pfn); if (!end_page) return false; |
e332f741a mm, compaction: b... |
288 289 290 291 292 |
/* * 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 mm, compaction: b... |
293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 |
do { if (pfn_valid_within(pfn)) { if (check_source && PageLRU(page)) { clear_pageblock_skip(page); return true; } if (check_target && PageBuddy(page)) { clear_pageblock_skip(page); return true; } } page += (1 << PAGE_ALLOC_COSTLY_ORDER); pfn += (1 << PAGE_ALLOC_COSTLY_ORDER); |
a2e9a5afc mm, compaction: f... |
308 |
} while (page <= end_page); |
e332f741a mm, compaction: b... |
309 310 311 |
return false; } |
21dc7e023 mm, compaction: p... |
312 |
/* |
bb13ffeb9 mm: compaction: c... |
313 314 315 316 |
* 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 mm: compaction: c... |
317 |
static void __reset_isolation_suitable(struct zone *zone) |
bb13ffeb9 mm: compaction: c... |
318 |
{ |
e332f741a mm, compaction: b... |
319 |
unsigned long migrate_pfn = zone->zone_start_pfn; |
6b0868c82 mm/compaction.c: ... |
320 |
unsigned long free_pfn = zone_end_pfn(zone) - 1; |
e332f741a mm, compaction: b... |
321 322 323 324 325 326 327 |
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 mm: compaction: c... |
328 |
|
62997027c mm: compaction: c... |
329 |
zone->compact_blockskip_flush = false; |
bb13ffeb9 mm: compaction: c... |
330 |
|
e332f741a mm, compaction: b... |
331 332 333 334 335 336 337 338 |
/* * 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 mm: compaction: c... |
339 |
cond_resched(); |
e332f741a mm, compaction: b... |
340 341 342 343 344 345 346 347 348 |
/* 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 mm: compaction: c... |
349 |
|
e332f741a mm, compaction: b... |
350 351 352 353 354 355 356 357 |
/* 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 mm: compaction: c... |
358 |
} |
02333641e mm, compaction: e... |
359 |
|
e332f741a mm, compaction: b... |
360 361 362 363 364 365 |
/* 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 mm: compaction: c... |
366 |
} |
62997027c mm: compaction: c... |
367 368 369 370 371 372 373 374 375 376 377 378 379 380 |
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 mm: compaction: c... |
381 |
/* |
e380bebe4 mm, compaction: k... |
382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 |
* 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 mm: compaction: c... |
422 |
* If no pages were isolated then mark this pageblock to be skipped in the |
62997027c mm: compaction: c... |
423 |
* future. The information is later cleared by __reset_isolation_suitable(). |
bb13ffeb9 mm: compaction: c... |
424 |
*/ |
c89511ab2 mm: compaction: R... |
425 |
static void update_pageblock_skip(struct compact_control *cc, |
d097a6f63 mm, compaction: r... |
426 |
struct page *page, unsigned long pfn) |
bb13ffeb9 mm: compaction: c... |
427 |
{ |
c89511ab2 mm: compaction: R... |
428 |
struct zone *zone = cc->zone; |
6815bf3f2 mm/compaction: re... |
429 |
|
2583d6713 mm, compaction: s... |
430 |
if (cc->no_set_skip_hint) |
6815bf3f2 mm/compaction: re... |
431 |
return; |
bb13ffeb9 mm: compaction: c... |
432 433 |
if (!page) return; |
edc2ca612 mm, compaction: m... |
434 |
set_pageblock_skip(page); |
c89511ab2 mm: compaction: R... |
435 |
|
35979ef33 mm, compaction: a... |
436 |
/* Update where async and sync compaction should restart */ |
e380bebe4 mm, compaction: k... |
437 438 |
if (pfn < zone->compact_cached_free_pfn) zone->compact_cached_free_pfn = pfn; |
bb13ffeb9 mm: compaction: c... |
439 440 441 442 443 444 445 |
} #else static inline bool isolation_suitable(struct compact_control *cc, struct page *page) { return true; } |
b527cfe5b mm, compaction: e... |
446 |
static inline bool pageblock_skip_persistent(struct page *page) |
21dc7e023 mm, compaction: p... |
447 448 449 450 451 |
{ return false; } static inline void update_pageblock_skip(struct compact_control *cc, |
d097a6f63 mm, compaction: r... |
452 |
struct page *page, unsigned long pfn) |
bb13ffeb9 mm: compaction: c... |
453 454 |
{ } |
e380bebe4 mm, compaction: k... |
455 456 457 458 459 460 461 462 463 464 |
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 mm: compaction: c... |
465 |
#endif /* CONFIG_COMPACTION */ |
8b44d2791 mm, compaction: p... |
466 467 |
/* * Compaction requires the taking of some coarse locks that are potentially |
cb2dcaf02 mm, compaction: f... |
468 469 470 471 |
* 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 mm, compaction: p... |
472 |
* |
cb2dcaf02 mm, compaction: f... |
473 |
* Always returns true which makes it easier to track lock state in callers. |
8b44d2791 mm, compaction: p... |
474 |
*/ |
cb2dcaf02 mm, compaction: f... |
475 |
static bool compact_lock_irqsave(spinlock_t *lock, unsigned long *flags, |
8b44d2791 mm, compaction: p... |
476 |
struct compact_control *cc) |
77337edee mm/compaction: ad... |
477 |
__acquires(lock) |
2a1402aa0 mm: compaction: a... |
478 |
{ |
cb2dcaf02 mm, compaction: f... |
479 480 481 482 483 484 |
/* 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 mm, compaction: p... |
485 |
} |
1f9efdef4 mm, compaction: k... |
486 |
|
cb2dcaf02 mm, compaction: f... |
487 |
spin_lock_irqsave(lock, *flags); |
8b44d2791 mm, compaction: p... |
488 |
return true; |
2a1402aa0 mm: compaction: a... |
489 |
} |
85aa125f0 mm: compaction: i... |
490 |
/* |
c67fe3752 mm: compaction: A... |
491 |
* Compaction requires the taking of some coarse locks that are potentially |
8b44d2791 mm, compaction: p... |
492 493 494 495 496 497 498 |
* 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 mm: compaction: A... |
499 |
* |
8b44d2791 mm, compaction: p... |
500 501 502 503 |
* 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 mm: compaction: A... |
504 |
*/ |
8b44d2791 mm, compaction: p... |
505 506 |
static bool compact_unlock_should_abort(spinlock_t *lock, unsigned long flags, bool *locked, struct compact_control *cc) |
c67fe3752 mm: compaction: A... |
507 |
{ |
8b44d2791 mm, compaction: p... |
508 509 510 511 |
if (*locked) { spin_unlock_irqrestore(lock, flags); *locked = false; } |
1f9efdef4 mm, compaction: k... |
512 |
|
8b44d2791 mm, compaction: p... |
513 |
if (fatal_signal_pending(current)) { |
c3486f537 mm, compaction: s... |
514 |
cc->contended = true; |
8b44d2791 mm, compaction: p... |
515 516 |
return true; } |
c67fe3752 mm: compaction: A... |
517 |
|
cf66f0700 mm, compaction: d... |
518 |
cond_resched(); |
be9765722 mm, compaction: p... |
519 520 521 |
return false; } |
c67fe3752 mm: compaction: A... |
522 |
/* |
9e4be4708 mm/compaction.c: ... |
523 524 525 |
* 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 mm: compaction: i... |
526 |
*/ |
f40d1e42b mm: compaction: a... |
527 |
static unsigned long isolate_freepages_block(struct compact_control *cc, |
e14c720ef mm, compaction: r... |
528 |
unsigned long *start_pfn, |
85aa125f0 mm: compaction: i... |
529 530 |
unsigned long end_pfn, struct list_head *freelist, |
4fca9730c mm, compaction: s... |
531 |
unsigned int stride, |
85aa125f0 mm: compaction: i... |
532 |
bool strict) |
748446bb6 mm: compaction: m... |
533 |
{ |
b7aba6984 mm: compaction: a... |
534 |
int nr_scanned = 0, total_isolated = 0; |
d097a6f63 mm, compaction: r... |
535 |
struct page *cursor; |
b8b2d8253 mm/compaction.c: ... |
536 |
unsigned long flags = 0; |
f40d1e42b mm: compaction: a... |
537 |
bool locked = false; |
e14c720ef mm, compaction: r... |
538 |
unsigned long blockpfn = *start_pfn; |
66c64223a mm/compaction: sp... |
539 |
unsigned int order; |
748446bb6 mm: compaction: m... |
540 |
|
4fca9730c mm, compaction: s... |
541 542 543 |
/* Strict mode is for isolation, speed is secondary */ if (strict) stride = 1; |
748446bb6 mm: compaction: m... |
544 |
cursor = pfn_to_page(blockpfn); |
f40d1e42b mm: compaction: a... |
545 |
/* Isolate free pages. */ |
4fca9730c mm, compaction: s... |
546 |
for (; blockpfn < end_pfn; blockpfn += stride, cursor += stride) { |
66c64223a mm/compaction: sp... |
547 |
int isolated; |
748446bb6 mm: compaction: m... |
548 |
struct page *page = cursor; |
8b44d2791 mm, compaction: p... |
549 550 551 552 553 554 555 556 557 |
/* * 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 mm: compaction: a... |
558 |
nr_scanned++; |
f40d1e42b mm: compaction: a... |
559 |
if (!pfn_valid_within(blockpfn)) |
2af120bc0 mm/compaction: br... |
560 |
goto isolate_fail; |
9fcd6d2e0 mm, compaction: s... |
561 562 563 564 565 566 567 |
/* * 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 mm, compaction: p... |
568 |
const unsigned int order = compound_order(page); |
d3c85bad8 mm, compaction: r... |
569 |
if (likely(order < MAX_ORDER)) { |
21dc7e023 mm, compaction: p... |
570 571 |
blockpfn += (1UL << order) - 1; cursor += (1UL << order) - 1; |
9fcd6d2e0 mm, compaction: s... |
572 |
} |
9fcd6d2e0 mm, compaction: s... |
573 574 |
goto isolate_fail; } |
f40d1e42b mm: compaction: a... |
575 |
if (!PageBuddy(page)) |
2af120bc0 mm/compaction: br... |
576 |
goto isolate_fail; |
f40d1e42b mm: compaction: a... |
577 578 |
/* |
69b7189f1 mm, compaction: s... |
579 580 581 582 583 |
* 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 mm: compaction: a... |
584 |
*/ |
69b7189f1 mm, compaction: s... |
585 |
if (!locked) { |
cb2dcaf02 mm, compaction: f... |
586 |
locked = compact_lock_irqsave(&cc->zone->lock, |
8b44d2791 mm, compaction: p... |
587 |
&flags, cc); |
f40d1e42b mm: compaction: a... |
588 |
|
69b7189f1 mm, compaction: s... |
589 590 591 592 |
/* Recheck this is a buddy page under lock */ if (!PageBuddy(page)) goto isolate_fail; } |
748446bb6 mm: compaction: m... |
593 |
|
66c64223a mm/compaction: sp... |
594 |
/* Found a free page, will break it into order-0 pages */ |
ab130f910 mm: rename page_o... |
595 |
order = buddy_order(page); |
66c64223a mm/compaction: sp... |
596 |
isolated = __isolate_free_page(page, order); |
a4f04f2c6 mm, compaction: a... |
597 598 |
if (!isolated) break; |
66c64223a mm/compaction: sp... |
599 |
set_page_private(page, order); |
a4f04f2c6 mm, compaction: a... |
600 |
|
748446bb6 mm: compaction: m... |
601 |
total_isolated += isolated; |
a4f04f2c6 mm, compaction: a... |
602 |
cc->nr_freepages += isolated; |
66c64223a mm/compaction: sp... |
603 |
list_add_tail(&page->lru, freelist); |
a4f04f2c6 mm, compaction: a... |
604 605 606 |
if (!strict && cc->nr_migratepages <= cc->nr_freepages) { blockpfn += isolated; break; |
748446bb6 mm: compaction: m... |
607 |
} |
a4f04f2c6 mm, compaction: a... |
608 609 610 611 |
/* Advance to the end of split page */ blockpfn += isolated - 1; cursor += isolated - 1; continue; |
2af120bc0 mm/compaction: br... |
612 613 614 615 616 617 |
isolate_fail: if (strict) break; else continue; |
748446bb6 mm: compaction: m... |
618 |
} |
a4f04f2c6 mm, compaction: a... |
619 620 |
if (locked) spin_unlock_irqrestore(&cc->zone->lock, flags); |
9fcd6d2e0 mm, compaction: s... |
621 622 623 624 625 626 |
/* * 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 mm/compaction: pr... |
627 628 |
trace_mm_compaction_isolate_freepages(*start_pfn, blockpfn, nr_scanned, total_isolated); |
e14c720ef mm, compaction: r... |
629 630 |
/* Record how far we have got within the block */ *start_pfn = blockpfn; |
f40d1e42b mm: compaction: a... |
631 632 633 634 635 |
/* * 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 mm/compaction: br... |
636 |
if (strict && blockpfn < end_pfn) |
f40d1e42b mm: compaction: a... |
637 |
total_isolated = 0; |
7f354a548 mm, compaction: a... |
638 |
cc->total_free_scanned += nr_scanned; |
397487db6 mm: compaction: A... |
639 |
if (total_isolated) |
010fc29a4 compaction: fix b... |
640 |
count_compact_events(COMPACTISOLATED, total_isolated); |
748446bb6 mm: compaction: m... |
641 642 |
return total_isolated; } |
85aa125f0 mm: compaction: i... |
643 644 |
/** * isolate_freepages_range() - isolate free pages. |
e8b098fc5 mm: kernel-doc: a... |
645 |
* @cc: Compaction control structure. |
85aa125f0 mm: compaction: i... |
646 647 648 649 650 651 652 653 654 655 656 |
* @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 mm: compaction: e... |
657 |
unsigned long |
bb13ffeb9 mm: compaction: c... |
658 659 |
isolate_freepages_range(struct compact_control *cc, unsigned long start_pfn, unsigned long end_pfn) |
85aa125f0 mm: compaction: i... |
660 |
{ |
e1409c325 mm/compaction: pa... |
661 |
unsigned long isolated, pfn, block_start_pfn, block_end_pfn; |
85aa125f0 mm: compaction: i... |
662 |
LIST_HEAD(freelist); |
7d49d8868 mm, compaction: r... |
663 |
pfn = start_pfn; |
06b6640a3 mm, compaction: w... |
664 |
block_start_pfn = pageblock_start_pfn(pfn); |
e1409c325 mm/compaction: pa... |
665 666 |
if (block_start_pfn < cc->zone->zone_start_pfn) block_start_pfn = cc->zone->zone_start_pfn; |
06b6640a3 mm, compaction: w... |
667 |
block_end_pfn = pageblock_end_pfn(pfn); |
7d49d8868 mm, compaction: r... |
668 669 |
for (; pfn < end_pfn; pfn += isolated, |
e1409c325 mm/compaction: pa... |
670 |
block_start_pfn = block_end_pfn, |
7d49d8868 mm, compaction: r... |
671 |
block_end_pfn += pageblock_nr_pages) { |
e14c720ef mm, compaction: r... |
672 673 |
/* Protect pfn from changing by isolate_freepages_block */ unsigned long isolate_start_pfn = pfn; |
85aa125f0 mm: compaction: i... |
674 |
|
85aa125f0 mm: compaction: i... |
675 |
block_end_pfn = min(block_end_pfn, end_pfn); |
584200163 mm/compaction: sk... |
676 677 678 679 680 681 |
/* * 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 mm, compaction: w... |
682 683 |
block_start_pfn = pageblock_start_pfn(pfn); block_end_pfn = pageblock_end_pfn(pfn); |
584200163 mm/compaction: sk... |
684 685 |
block_end_pfn = min(block_end_pfn, end_pfn); } |
e1409c325 mm/compaction: pa... |
686 687 |
if (!pageblock_pfn_to_page(block_start_pfn, block_end_pfn, cc->zone)) |
7d49d8868 mm, compaction: r... |
688 |
break; |
e14c720ef mm, compaction: r... |
689 |
isolated = isolate_freepages_block(cc, &isolate_start_pfn, |
4fca9730c mm, compaction: s... |
690 |
block_end_pfn, &freelist, 0, true); |
85aa125f0 mm: compaction: i... |
691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 |
/* * 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 mm/compaction: sp... |
706 |
/* __isolate_free_page() does not map the pages */ |
4469ab984 mm, compaction: r... |
707 |
split_map_pages(&freelist); |
85aa125f0 mm: compaction: i... |
708 709 710 711 712 713 714 715 716 717 |
if (pfn < end_pfn) { /* Loop terminated early, cleanup. */ release_freepages(&freelist); return 0; } /* We don't use freelists for anything. */ return pfn; } |
748446bb6 mm: compaction: m... |
718 |
/* Similar to reclaim, but different enough that they don't share logic */ |
5f438eee8 mm/compaction: pa... |
719 |
static bool too_many_isolated(pg_data_t *pgdat) |
748446bb6 mm: compaction: m... |
720 |
{ |
bc6930457 mm: compaction: h... |
721 |
unsigned long active, inactive, isolated; |
748446bb6 mm: compaction: m... |
722 |
|
5f438eee8 mm/compaction: pa... |
723 724 725 726 727 728 |
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 mm: compaction: m... |
729 |
|
bc6930457 mm: compaction: h... |
730 |
return isolated > (inactive + active) / 2; |
748446bb6 mm: compaction: m... |
731 |
} |
2fe86e000 mm: compaction: i... |
732 |
/** |
edc2ca612 mm, compaction: m... |
733 734 |
* isolate_migratepages_block() - isolate all migrate-able pages within * a single pageblock |
2fe86e000 mm: compaction: i... |
735 |
* @cc: Compaction control structure. |
edc2ca612 mm, compaction: m... |
736 737 738 |
* @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 mm: compaction: i... |
739 740 |
* * Isolate all pages that can be migrated from the range specified by |
edc2ca612 mm, compaction: m... |
741 742 743 744 |
* [low_pfn, end_pfn). The range is expected to be within same pageblock. * Returns zero if there is a fatal signal pending, otherwise PFN of the * first page that was not scanned (which may be both less, equal to or more * than end_pfn). |
2fe86e000 mm: compaction: i... |
745 |
* |
edc2ca612 mm, compaction: m... |
746 747 748 |
* The pages are isolated on cc->migratepages list (not required to be empty), * and cc->nr_migratepages is updated accordingly. The cc->migrate_pfn field * is neither read nor updated. |
748446bb6 mm: compaction: m... |
749 |
*/ |
edc2ca612 mm, compaction: m... |
750 751 752 |
static unsigned long isolate_migratepages_block(struct compact_control *cc, unsigned long low_pfn, unsigned long end_pfn, isolate_mode_t isolate_mode) |
748446bb6 mm: compaction: m... |
753 |
{ |
5f438eee8 mm/compaction: pa... |
754 |
pg_data_t *pgdat = cc->zone->zone_pgdat; |
b7aba6984 mm: compaction: a... |
755 |
unsigned long nr_scanned = 0, nr_isolated = 0; |
fa9add641 mm/memcg: apply a... |
756 |
struct lruvec *lruvec; |
b8b2d8253 mm/compaction.c: ... |
757 |
unsigned long flags = 0; |
2a1402aa0 mm: compaction: a... |
758 |
bool locked = false; |
bb13ffeb9 mm: compaction: c... |
759 |
struct page *page = NULL, *valid_page = NULL; |
e34d85f0e mm/compaction: pr... |
760 |
unsigned long start_pfn = low_pfn; |
fdd048e12 mm, compaction: s... |
761 762 |
bool skip_on_failure = false; unsigned long next_skip_pfn = 0; |
e380bebe4 mm, compaction: k... |
763 |
bool skip_updated = false; |
748446bb6 mm: compaction: m... |
764 |
|
748446bb6 mm: compaction: m... |
765 766 767 768 769 |
/* * 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 mm/compaction: pa... |
770 |
while (unlikely(too_many_isolated(pgdat))) { |
d20bdd571 mm/compaction: st... |
771 772 773 |
/* stop isolation if there are still pages not migrated */ if (cc->nr_migratepages) return 0; |
f9e35b3b4 mm: compaction: a... |
774 |
/* async migration should just abort */ |
e0b9daeb4 mm, compaction: e... |
775 |
if (cc->mode == MIGRATE_ASYNC) |
2fe86e000 mm: compaction: i... |
776 |
return 0; |
f9e35b3b4 mm: compaction: a... |
777 |
|
748446bb6 mm: compaction: m... |
778 779 780 |
congestion_wait(BLK_RW_ASYNC, HZ/10); if (fatal_signal_pending(current)) |
2fe86e000 mm: compaction: i... |
781 |
return 0; |
748446bb6 mm: compaction: m... |
782 |
} |
cf66f0700 mm, compaction: d... |
783 |
cond_resched(); |
aeef4b838 mm, compaction: t... |
784 |
|
fdd048e12 mm, compaction: s... |
785 786 787 788 |
if (cc->direct_compaction && (cc->mode == MIGRATE_ASYNC)) { skip_on_failure = true; next_skip_pfn = block_end_pfn(low_pfn, cc->order); } |
748446bb6 mm: compaction: m... |
789 |
/* Time to isolate some pages for migration */ |
748446bb6 mm: compaction: m... |
790 |
for (; low_pfn < end_pfn; low_pfn++) { |
29c0dde83 mm, compaction: a... |
791 |
|
fdd048e12 mm, compaction: s... |
792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 |
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 mm, compaction: p... |
813 814 |
/* * Periodically drop the lock (if held) regardless of its |
670105a25 mm: compaction: a... |
815 816 |
* contention, to give chance to IRQs. Abort completely if * a fatal signal is pending. |
8b44d2791 mm, compaction: p... |
817 818 |
*/ if (!(low_pfn % SWAP_CLUSTER_MAX) |
f4b7e272b mm: remove zone_l... |
819 |
&& compact_unlock_should_abort(&pgdat->lru_lock, |
670105a25 mm: compaction: a... |
820 821 822 823 |
flags, &locked, cc)) { low_pfn = 0; goto fatal_pending; } |
c67fe3752 mm: compaction: A... |
824 |
|
748446bb6 mm: compaction: m... |
825 |
if (!pfn_valid_within(low_pfn)) |
fdd048e12 mm, compaction: s... |
826 |
goto isolate_fail; |
b7aba6984 mm: compaction: a... |
827 |
nr_scanned++; |
748446bb6 mm: compaction: m... |
828 |
|
748446bb6 mm: compaction: m... |
829 |
page = pfn_to_page(low_pfn); |
dc9086004 mm: compaction: c... |
830 |
|
e380bebe4 mm, compaction: k... |
831 832 833 834 835 836 837 838 839 840 841 |
/* * 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; goto isolate_abort; } |
bb13ffeb9 mm: compaction: c... |
842 |
valid_page = page; |
e380bebe4 mm, compaction: k... |
843 |
} |
bb13ffeb9 mm: compaction: c... |
844 |
|
6c14466cc mm: improve docum... |
845 |
/* |
99c0fd5e5 mm, compaction: s... |
846 847 848 849 |
* 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 mm: improve docum... |
850 |
*/ |
99c0fd5e5 mm, compaction: s... |
851 |
if (PageBuddy(page)) { |
ab130f910 mm: rename page_o... |
852 |
unsigned long freepage_order = buddy_order_unsafe(page); |
99c0fd5e5 mm, compaction: s... |
853 854 855 856 857 858 859 860 |
/* * 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 mm: compaction: m... |
861 |
continue; |
99c0fd5e5 mm, compaction: s... |
862 |
} |
748446bb6 mm: compaction: m... |
863 |
|
9927af740 mm: compaction: p... |
864 |
/* |
29c0dde83 mm, compaction: a... |
865 |
* Regardless of being on LRU, compound pages such as THP and |
1da2f328f mm,thp,compaction... |
866 867 868 869 870 |
* 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 thp: transhuge is... |
871 |
*/ |
1da2f328f mm,thp,compaction... |
872 |
if (PageCompound(page) && !cc->alloc_contig) { |
21dc7e023 mm, compaction: p... |
873 |
const unsigned int order = compound_order(page); |
edc2ca612 mm, compaction: m... |
874 |
|
d3c85bad8 mm, compaction: r... |
875 |
if (likely(order < MAX_ORDER)) |
21dc7e023 mm, compaction: p... |
876 |
low_pfn += (1UL << order) - 1; |
fdd048e12 mm, compaction: s... |
877 |
goto isolate_fail; |
2a1402aa0 mm: compaction: a... |
878 |
} |
bda807d44 mm: migrate: supp... |
879 880 881 882 883 884 |
/* * 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 mm: migrate: supp... |
885 886 887 888 889 890 891 |
/* * __PageMovable can return false positive so we need * to verify it under page_lock. */ if (unlikely(__PageMovable(page)) && !PageIsolated(page)) { if (locked) { |
f4b7e272b mm: remove zone_l... |
892 |
spin_unlock_irqrestore(&pgdat->lru_lock, |
bda807d44 mm: migrate: supp... |
893 894 895 |
flags); locked = false; } |
9e5bcd610 mm/migration: mak... |
896 |
if (!isolate_movable_page(page, isolate_mode)) |
bda807d44 mm: migrate: supp... |
897 898 |
goto isolate_success; } |
fdd048e12 mm, compaction: s... |
899 |
goto isolate_fail; |
bda807d44 mm: migrate: supp... |
900 |
} |
29c0dde83 mm, compaction: a... |
901 |
|
119d6d59d mm, compaction: a... |
902 903 904 905 906 907 908 |
/* * 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 mm, compaction: s... |
909 |
goto isolate_fail; |
119d6d59d mm, compaction: a... |
910 |
|
73e64c51a mm, compaction: a... |
911 912 913 914 915 916 |
/* * 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; |
69b7189f1 mm, compaction: s... |
917 918 |
/* If we already hold the lock, we can skip some rechecking */ if (!locked) { |
f4b7e272b mm: remove zone_l... |
919 |
locked = compact_lock_irqsave(&pgdat->lru_lock, |
8b44d2791 mm, compaction: p... |
920 |
&flags, cc); |
e380bebe4 mm, compaction: k... |
921 |
|
e380bebe4 mm, compaction: k... |
922 923 924 925 926 927 |
/* 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 mm: compaction: a... |
928 |
|
29c0dde83 mm, compaction: a... |
929 |
/* Recheck PageLRU and PageCompound under lock */ |
69b7189f1 mm, compaction: s... |
930 |
if (!PageLRU(page)) |
fdd048e12 mm, compaction: s... |
931 |
goto isolate_fail; |
29c0dde83 mm, compaction: a... |
932 933 934 935 936 937 |
/* * 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 mm,thp,compaction... |
938 |
if (unlikely(PageCompound(page) && !cc->alloc_contig)) { |
d8c6546b1 mm: introduce com... |
939 |
low_pfn += compound_nr(page) - 1; |
fdd048e12 mm, compaction: s... |
940 |
goto isolate_fail; |
69b7189f1 mm, compaction: s... |
941 |
} |
bc835011a thp: transhuge is... |
942 |
} |
f4b7e272b mm: remove zone_l... |
943 |
lruvec = mem_cgroup_page_lruvec(page, pgdat); |
fa9add641 mm/memcg: apply a... |
944 |
|
748446bb6 mm: compaction: m... |
945 |
/* Try isolate the page */ |
edc2ca612 mm, compaction: m... |
946 |
if (__isolate_lru_page(page, isolate_mode) != 0) |
fdd048e12 mm, compaction: s... |
947 |
goto isolate_fail; |
748446bb6 mm: compaction: m... |
948 |
|
1da2f328f mm,thp,compaction... |
949 950 951 |
/* The whole page is taken off the LRU; skip the tail pages. */ if (PageCompound(page)) low_pfn += compound_nr(page) - 1; |
bc835011a thp: transhuge is... |
952 |
|
748446bb6 mm: compaction: m... |
953 |
/* Successfully isolated */ |
fa9add641 mm/memcg: apply a... |
954 |
del_page_from_lru_list(page, lruvec, page_lru(page)); |
1da2f328f mm,thp,compaction... |
955 |
mod_node_page_state(page_pgdat(page), |
9de4f22a6 mm: code cleanup ... |
956 |
NR_ISOLATED_ANON + page_is_file_lru(page), |
6c357848b mm: replace hpage... |
957 |
thp_nr_pages(page)); |
b6c750163 mm/compaction: cl... |
958 959 |
isolate_success: |
fdd048e12 mm, compaction: s... |
960 |
list_add(&page->lru, &cc->migratepages); |
38935861d mm/compaction: co... |
961 962 |
cc->nr_migratepages += compound_nr(page); nr_isolated += compound_nr(page); |
748446bb6 mm: compaction: m... |
963 |
|
804d3121b mm, compaction: a... |
964 965 |
/* * Avoid isolating too much unless this block is being |
cb2dcaf02 mm, compaction: f... |
966 967 968 |
* 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 mm, compaction: a... |
969 |
*/ |
38935861d mm/compaction: co... |
970 |
if (cc->nr_migratepages >= COMPACT_CLUSTER_MAX && |
cb2dcaf02 mm, compaction: f... |
971 |
!cc->rescan && !cc->contended) { |
31b8384a5 mm: compaction: p... |
972 |
++low_pfn; |
748446bb6 mm: compaction: m... |
973 |
break; |
31b8384a5 mm: compaction: p... |
974 |
} |
fdd048e12 mm, compaction: s... |
975 976 977 978 979 980 981 982 983 984 985 986 987 |
continue; isolate_fail: if (!skip_on_failure) 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) { |
f4b7e272b mm: remove zone_l... |
988 |
spin_unlock_irqrestore(&pgdat->lru_lock, flags); |
fdd048e12 mm, compaction: s... |
989 990 |
locked = false; } |
fdd048e12 mm, compaction: s... |
991 992 |
putback_movable_pages(&cc->migratepages); cc->nr_migratepages = 0; |
fdd048e12 mm, compaction: s... |
993 994 995 996 997 998 999 1000 1001 1002 1003 |
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; } |
748446bb6 mm: compaction: m... |
1004 |
} |
99c0fd5e5 mm, compaction: s... |
1005 1006 1007 1008 1009 1010 |
/* * The PageBuddy() check could have potentially brought us outside * the range to be scanned. */ if (unlikely(low_pfn > end_pfn)) low_pfn = end_pfn; |
e380bebe4 mm, compaction: k... |
1011 |
isolate_abort: |
c67fe3752 mm: compaction: A... |
1012 |
if (locked) |
f4b7e272b mm: remove zone_l... |
1013 |
spin_unlock_irqrestore(&pgdat->lru_lock, flags); |
748446bb6 mm: compaction: m... |
1014 |
|
50b5b094e mm: compaction: d... |
1015 |
/* |
804d3121b mm, compaction: a... |
1016 1017 1018 1019 1020 1021 |
* 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 mm: compaction: d... |
1022 |
*/ |
804d3121b mm, compaction: a... |
1023 |
if (low_pfn == end_pfn && (!nr_isolated || cc->rescan)) { |
e380bebe4 mm, compaction: k... |
1024 1025 1026 1027 |
if (valid_page && !skip_updated) set_pageblock_skip(valid_page); update_cached_migrate(cc, low_pfn); } |
bb13ffeb9 mm: compaction: c... |
1028 |
|
e34d85f0e mm/compaction: pr... |
1029 1030 |
trace_mm_compaction_isolate_migratepages(start_pfn, low_pfn, nr_scanned, nr_isolated); |
b7aba6984 mm: compaction: a... |
1031 |
|
670105a25 mm: compaction: a... |
1032 |
fatal_pending: |
7f354a548 mm, compaction: a... |
1033 |
cc->total_migrate_scanned += nr_scanned; |
397487db6 mm: compaction: A... |
1034 |
if (nr_isolated) |
010fc29a4 compaction: fix b... |
1035 |
count_compact_events(COMPACTISOLATED, nr_isolated); |
397487db6 mm: compaction: A... |
1036 |
|
2fe86e000 mm: compaction: i... |
1037 1038 |
return low_pfn; } |
edc2ca612 mm, compaction: m... |
1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 |
/** * 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. * * Returns zero if isolation fails fatally due to e.g. pending signal. * Otherwise, function returns one-past-the-last PFN of isolated page * (which may be greater than end_pfn if end fell in a middle of a THP page). */ unsigned long isolate_migratepages_range(struct compact_control *cc, unsigned long start_pfn, unsigned long end_pfn) { |
e1409c325 mm/compaction: pa... |
1053 |
unsigned long pfn, block_start_pfn, block_end_pfn; |
edc2ca612 mm, compaction: m... |
1054 1055 1056 |
/* Scan block by block. First and last block may be incomplete */ pfn = start_pfn; |
06b6640a3 mm, compaction: w... |
1057 |
block_start_pfn = pageblock_start_pfn(pfn); |
e1409c325 mm/compaction: pa... |
1058 1059 |
if (block_start_pfn < cc->zone->zone_start_pfn) block_start_pfn = cc->zone->zone_start_pfn; |
06b6640a3 mm, compaction: w... |
1060 |
block_end_pfn = pageblock_end_pfn(pfn); |
edc2ca612 mm, compaction: m... |
1061 1062 |
for (; pfn < end_pfn; pfn = block_end_pfn, |
e1409c325 mm/compaction: pa... |
1063 |
block_start_pfn = block_end_pfn, |
edc2ca612 mm, compaction: m... |
1064 1065 1066 |
block_end_pfn += pageblock_nr_pages) { block_end_pfn = min(block_end_pfn, end_pfn); |
e1409c325 mm/compaction: pa... |
1067 1068 |
if (!pageblock_pfn_to_page(block_start_pfn, block_end_pfn, cc->zone)) |
edc2ca612 mm, compaction: m... |
1069 1070 1071 1072 |
continue; pfn = isolate_migratepages_block(cc, pfn, block_end_pfn, ISOLATE_UNEVICTABLE); |
14af4a5e9 mm, cma: prevent ... |
1073 |
if (!pfn) |
edc2ca612 mm, compaction: m... |
1074 |
break; |
6ea41c0c0 mm/compaction.c: ... |
1075 |
|
38935861d mm/compaction: co... |
1076 |
if (cc->nr_migratepages >= COMPACT_CLUSTER_MAX) |
6ea41c0c0 mm/compaction.c: ... |
1077 |
break; |
edc2ca612 mm, compaction: m... |
1078 |
} |
edc2ca612 mm, compaction: m... |
1079 1080 1081 |
return pfn; } |
ff9543fd3 mm: compaction: e... |
1082 1083 |
#endif /* CONFIG_COMPACTION || CONFIG_CMA */ #ifdef CONFIG_COMPACTION |
018e9a49a mm/compaction.c: ... |
1084 |
|
b682debd9 mm, compaction: c... |
1085 1086 1087 |
static bool suitable_migration_source(struct compact_control *cc, struct page *page) { |
282722b0d mm, compaction: r... |
1088 |
int block_mt; |
9bebefd59 mm, compaction: c... |
1089 1090 |
if (pageblock_skip_persistent(page)) return false; |
282722b0d mm, compaction: r... |
1091 |
if ((cc->mode != MIGRATE_ASYNC) || !cc->direct_compaction) |
b682debd9 mm, compaction: c... |
1092 |
return true; |
282722b0d mm, compaction: r... |
1093 1094 1095 1096 1097 1098 |
block_mt = get_pageblock_migratetype(page); if (cc->migratetype == MIGRATE_MOVABLE) return is_migrate_movable(block_mt); else return block_mt == cc->migratetype; |
b682debd9 mm, compaction: c... |
1099 |
} |
018e9a49a mm/compaction.c: ... |
1100 |
/* Returns true if the page is within a block suitable for migration to */ |
9f7e33879 mm, compaction: m... |
1101 1102 |
static bool suitable_migration_target(struct compact_control *cc, struct page *page) |
018e9a49a mm/compaction.c: ... |
1103 1104 1105 1106 1107 1108 1109 1110 |
{ /* 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 mm: rename page_o... |
1111 |
if (buddy_order_unsafe(page) >= pageblock_order) |
018e9a49a mm/compaction.c: ... |
1112 1113 |
return false; } |
1ef36db2a mm/compaction: ig... |
1114 1115 |
if (cc->ignore_block_suitable) return true; |
018e9a49a mm/compaction.c: ... |
1116 |
/* If the block is MIGRATE_MOVABLE or MIGRATE_CMA, allow migration */ |
b682debd9 mm, compaction: c... |
1117 |
if (is_migrate_movable(get_pageblock_migratetype(page))) |
018e9a49a mm/compaction.c: ... |
1118 1119 1120 1121 1122 |
return true; /* Otherwise skip the block */ return false; } |
70b44595e mm, compaction: u... |
1123 1124 1125 |
static inline unsigned int freelist_scan_limit(struct compact_control *cc) { |
dd7ef7bd1 mm/compaction.c: ... |
1126 1127 1128 |
unsigned short shift = BITS_PER_LONG - 1; return (COMPACT_CLUSTER_MAX >> min(shift, cc->fast_search_fail)) + 1; |
70b44595e mm, compaction: u... |
1129 |
} |
2fe86e000 mm: compaction: i... |
1130 |
/* |
f2849aa09 mm, compaction: m... |
1131 1132 1133 1134 1135 1136 1137 1138 |
* 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 mm, compaction: u... |
1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 |
/* * 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); if (!list_empty(&sublist)) list_splice_tail(&sublist, freelist); } } /* * 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 mm, compaction: u... |
1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 |
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); if (!list_empty(&sublist)) list_splice_tail(&sublist, freelist); } } |
5a811889d mm, compaction: u... |
1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 |
static void fast_isolate_around(struct compact_control *cc, unsigned long pfn, unsigned long nr_isolated) { unsigned long start_pfn, end_pfn; struct page *page = pfn_to_page(pfn); /* 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 */ start_pfn = pageblock_start_pfn(pfn); |
60fce36af mm/compaction.c: ... |
1189 |
end_pfn = min(pageblock_end_pfn(pfn), zone_end_pfn(cc->zone)) - 1; |
5a811889d mm, compaction: u... |
1190 1191 1192 |
/* Scan before */ if (start_pfn != pfn) { |
4fca9730c mm, compaction: s... |
1193 |
isolate_freepages_block(cc, &start_pfn, pfn, &cc->freepages, 1, false); |
5a811889d mm, compaction: u... |
1194 1195 1196 1197 1198 1199 |
if (cc->nr_freepages >= cc->nr_migratepages) return; } /* Scan after */ start_pfn = pfn + nr_isolated; |
60fce36af mm/compaction.c: ... |
1200 |
if (start_pfn < end_pfn) |
4fca9730c mm, compaction: s... |
1201 |
isolate_freepages_block(cc, &start_pfn, end_pfn, &cc->freepages, 1, false); |
5a811889d mm, compaction: u... |
1202 1203 1204 1205 1206 |
/* 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 mm, compaction: r... |
1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 |
/* 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 mm, compaction: u... |
1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 |
static unsigned long fast_isolate_freepages(struct compact_control *cc) { unsigned int limit = min(1U, freelist_scan_limit(cc) >> 1); unsigned int nr_scanned = 0; unsigned long low_pfn, min_pfn, high_pfn = 0, highest = 0; 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 mm, compaction: b... |
1244 |
if (cc->free_pfn >= cc->zone->compact_init_free_pfn) { |
5a811889d mm, compaction: u... |
1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 |
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 mm, compaction: r... |
1259 1260 1261 1262 1263 1264 1265 1266 1267 |
/* * 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 mm, compaction: u... |
1268 1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 |
struct free_area *area = &cc->zone->free_area[order]; struct list_head *freelist; struct page *freepage; unsigned long flags; unsigned int order_scanned = 0; 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) highest = pageblock_start_pfn(pfn); if (pfn >= low_pfn) { cc->fast_search_fail = 0; |
dbe2d4e4f mm, compaction: r... |
1291 |
cc->search_order = order; |
5a811889d mm, compaction: u... |
1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 |
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 mm/compaction.c: ... |
1328 |
order = cc->search_order + 1; |
5a811889d mm, compaction: u... |
1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 |
page = NULL; } } spin_unlock_irqrestore(&cc->zone->lock, flags); /* * Smaller scan on next order so the total scan ig related * to freelist_scan_limit. */ if (order_scanned >= limit) limit = min(1U, limit >> 1); } if (!page) { cc->fast_search_fail++; if (scan_start) { /* * Use the highest PFN found above min. If one was |
f38677551 mm/compaction: fi... |
1348 |
* not found, be pessimistic for direct compaction |
5a811889d mm, compaction: u... |
1349 1350 1351 1352 1353 1354 |
* and use the min mark. */ if (highest) { page = pfn_to_page(highest); cc->free_pfn = highest; } else { |
e577c8b64 mm, compaction: m... |
1355 |
if (cc->direct_compaction && pfn_valid(min_pfn)) { |
73a6e474c mm: memmap_init: ... |
1356 1357 1358 |
page = pageblock_pfn_to_page(min_pfn, pageblock_end_pfn(min_pfn), cc->zone); |
5a811889d mm, compaction: u... |
1359 1360 1361 1362 1363 |
cc->free_pfn = min_pfn; } } } } |
d097a6f63 mm, compaction: r... |
1364 1365 |
if (highest && highest >= cc->zone->compact_cached_free_pfn) { highest -= pageblock_nr_pages; |
5a811889d mm, compaction: u... |
1366 |
cc->zone->compact_cached_free_pfn = highest; |
d097a6f63 mm, compaction: r... |
1367 |
} |
5a811889d mm, compaction: u... |
1368 1369 1370 1371 1372 1373 1374 1375 1376 |
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 mm, compaction: m... |
1377 |
/* |
ff9543fd3 mm: compaction: e... |
1378 1379 |
* Based on information in the current compact_control, find blocks * suitable for isolating free pages from and then isolate them. |
2fe86e000 mm: compaction: i... |
1380 |
*/ |
edc2ca612 mm, compaction: m... |
1381 |
static void isolate_freepages(struct compact_control *cc) |
2fe86e000 mm: compaction: i... |
1382 |
{ |
edc2ca612 mm, compaction: m... |
1383 |
struct zone *zone = cc->zone; |
ff9543fd3 mm: compaction: e... |
1384 |
struct page *page; |
c96b9e508 mm/compaction: cl... |
1385 |
unsigned long block_start_pfn; /* start of current pageblock */ |
e14c720ef mm, compaction: r... |
1386 |
unsigned long isolate_start_pfn; /* exact pfn we start at */ |
c96b9e508 mm/compaction: cl... |
1387 1388 |
unsigned long block_end_pfn; /* end of current pageblock */ unsigned long low_pfn; /* lowest pfn scanner is able to scan */ |
ff9543fd3 mm: compaction: e... |
1389 |
struct list_head *freelist = &cc->freepages; |
4fca9730c mm, compaction: s... |
1390 |
unsigned int stride; |
2fe86e000 mm: compaction: i... |
1391 |
|
5a811889d mm, compaction: u... |
1392 1393 1394 1395 |
/* 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 mm: compaction: e... |
1396 1397 |
/* * Initialise the free scanner. The starting point is where we last |
49e068f0b mm/compaction: ma... |
1398 |
* successfully isolated from, zone-cached value, or the end of the |
e14c720ef mm, compaction: r... |
1399 1400 |
* zone when isolating for the first time. For looping we also need * this pfn aligned down to the pageblock boundary, because we do |
c96b9e508 mm/compaction: cl... |
1401 1402 |
* block_start_pfn -= pageblock_nr_pages in the for loop. * For ending point, take care when isolating in last pageblock of a |
a1c1dbeb2 mm/compaction.c: ... |
1403 |
* zone which ends in the middle of a pageblock. |
49e068f0b mm/compaction: ma... |
1404 1405 |
* The low boundary is the end of the pageblock the migration scanner * is using. |
ff9543fd3 mm: compaction: e... |
1406 |
*/ |
e14c720ef mm, compaction: r... |
1407 |
isolate_start_pfn = cc->free_pfn; |
5a811889d mm, compaction: u... |
1408 |
block_start_pfn = pageblock_start_pfn(isolate_start_pfn); |
c96b9e508 mm/compaction: cl... |
1409 1410 |
block_end_pfn = min(block_start_pfn + pageblock_nr_pages, zone_end_pfn(zone)); |
06b6640a3 mm, compaction: w... |
1411 |
low_pfn = pageblock_end_pfn(cc->migrate_pfn); |
4fca9730c mm, compaction: s... |
1412 |
stride = cc->mode == MIGRATE_ASYNC ? COMPACT_CLUSTER_MAX : 1; |
2fe86e000 mm: compaction: i... |
1413 |
|
ff9543fd3 mm: compaction: e... |
1414 |
/* |
ff9543fd3 mm: compaction: e... |
1415 1416 1417 1418 |
* 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 mm, compaction: s... |
1419 |
for (; block_start_pfn >= low_pfn; |
c96b9e508 mm/compaction: cl... |
1420 |
block_end_pfn = block_start_pfn, |
e14c720ef mm, compaction: r... |
1421 1422 |
block_start_pfn -= pageblock_nr_pages, isolate_start_pfn = block_start_pfn) { |
4fca9730c mm, compaction: s... |
1423 |
unsigned long nr_isolated; |
f6ea3adb7 mm/compaction.c: ... |
1424 1425 |
/* * This can iterate a massively long zone without finding any |
cb810ad29 mm, compaction: r... |
1426 |
* suitable migration targets, so periodically check resched. |
f6ea3adb7 mm/compaction.c: ... |
1427 |
*/ |
cb810ad29 mm, compaction: r... |
1428 |
if (!(block_start_pfn % (SWAP_CLUSTER_MAX * pageblock_nr_pages))) |
cf66f0700 mm, compaction: d... |
1429 |
cond_resched(); |
f6ea3adb7 mm/compaction.c: ... |
1430 |
|
7d49d8868 mm, compaction: r... |
1431 1432 1433 |
page = pageblock_pfn_to_page(block_start_pfn, block_end_pfn, zone); if (!page) |
ff9543fd3 mm: compaction: e... |
1434 1435 1436 |
continue; /* Check the block is suitable for migration */ |
9f7e33879 mm, compaction: m... |
1437 |
if (!suitable_migration_target(cc, page)) |
ff9543fd3 mm: compaction: e... |
1438 |
continue; |
68e3e9262 Revert "mm: compa... |
1439 |
|
bb13ffeb9 mm: compaction: c... |
1440 1441 1442 |
/* If isolation recently failed, do not retry */ if (!isolation_suitable(cc, page)) continue; |
e14c720ef mm, compaction: r... |
1443 |
/* Found a block suitable for isolating free pages from. */ |
4fca9730c mm, compaction: s... |
1444 1445 |
nr_isolated = isolate_freepages_block(cc, &isolate_start_pfn, block_end_pfn, freelist, stride, false); |
ff9543fd3 mm: compaction: e... |
1446 |
|
d097a6f63 mm, compaction: r... |
1447 1448 1449 |
/* 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 mm, compaction: f... |
1450 1451 |
/* Are enough freepages isolated? */ if (cc->nr_freepages >= cc->nr_migratepages) { |
a46cbf3bc mm, compaction: p... |
1452 1453 1454 1455 1456 |
if (isolate_start_pfn >= block_end_pfn) { /* * Restart at previous pageblock if more * freepages can be isolated next time. */ |
f5f61a320 mm, compaction: s... |
1457 1458 |
isolate_start_pfn = block_start_pfn - pageblock_nr_pages; |
a46cbf3bc mm, compaction: p... |
1459 |
} |
be9765722 mm, compaction: p... |
1460 |
break; |
a46cbf3bc mm, compaction: p... |
1461 |
} else if (isolate_start_pfn < block_end_pfn) { |
f5f61a320 mm, compaction: s... |
1462 |
/* |
a46cbf3bc mm, compaction: p... |
1463 1464 |
* If isolation failed early, do not continue * needlessly. |
f5f61a320 mm, compaction: s... |
1465 |
*/ |
a46cbf3bc mm, compaction: p... |
1466 |
break; |
f5f61a320 mm, compaction: s... |
1467 |
} |
4fca9730c mm, compaction: s... |
1468 1469 1470 1471 1472 1473 1474 |
/* Adjust stride depending on isolation */ if (nr_isolated) { stride = 1; continue; } stride = min_t(unsigned int, COMPACT_CLUSTER_MAX, stride << 1); |
ff9543fd3 mm: compaction: e... |
1475 |
} |
7ed695e06 mm: compaction: d... |
1476 |
/* |
f5f61a320 mm, compaction: s... |
1477 1478 1479 1480 |
* 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 mm: compaction: d... |
1481 |
*/ |
f5f61a320 mm, compaction: s... |
1482 |
cc->free_pfn = isolate_start_pfn; |
5a811889d mm, compaction: u... |
1483 1484 1485 1486 |
splitmap: /* __isolate_free_page() does not map the pages */ split_map_pages(freelist); |
748446bb6 mm: compaction: m... |
1487 1488 1489 1490 1491 1492 1493 |
} /* * 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 mm, migrate: remo... |
1494 |
unsigned long data) |
748446bb6 mm: compaction: m... |
1495 1496 1497 |
{ struct compact_control *cc = (struct compact_control *)data; struct page *freepage; |
748446bb6 mm: compaction: m... |
1498 |
if (list_empty(&cc->freepages)) { |
cb2dcaf02 mm, compaction: f... |
1499 |
isolate_freepages(cc); |
748446bb6 mm: compaction: m... |
1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 |
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 mm, compaction: r... |
1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 |
* 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 mm: compaction: e... |
1524 1525 1526 1527 1528 1529 1530 1531 |
/* 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 mm: allow compact... |
1532 1533 1534 |
* Allow userspace to control policy on scanning the unevictable LRU for * compactable pages. */ |
6923aa0d8 mm/compaction: Di... |
1535 1536 1537 |
#ifdef CONFIG_PREEMPT_RT int sysctl_compact_unevictable_allowed __read_mostly = 0; #else |
5bbe3547a mm: allow compact... |
1538 |
int sysctl_compact_unevictable_allowed __read_mostly = 1; |
6923aa0d8 mm/compaction: Di... |
1539 |
#endif |
5bbe3547a mm: allow compact... |
1540 |
|
70b44595e mm, compaction: u... |
1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 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 |
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; /* 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; order >= PAGE_ALLOC_COSTLY_ORDER && pfn == cc->migrate_pfn && nr_scanned < limit; 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; nr_scanned++; free_pfn = page_to_pfn(freepage); if (free_pfn < high_pfn) { |
70b44595e mm, compaction: u... |
1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 |
/* * 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. */ if (get_pageblock_skip(freepage)) { if (list_is_last(freelist, &freepage->lru)) break; continue; } /* Reorder to so a future search skips recent pages */ move_freelist_tail(freelist, freepage); |
e380bebe4 mm, compaction: k... |
1653 |
update_fast_start_pfn(cc, free_pfn); |
70b44595e mm, compaction: u... |
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 |
pfn = pageblock_start_pfn(free_pfn); cc->fast_search_fail = 0; set_pageblock_skip(freepage); break; } if (nr_scanned >= limit) { cc->fast_search_fail++; move_freelist_tail(freelist, freepage); break; } } 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. */ if (pfn == cc->migrate_pfn) pfn = reinit_migrate_pfn(cc); return pfn; } |
5bbe3547a mm: allow compact... |
1680 |
/* |
edc2ca612 mm, compaction: m... |
1681 1682 1683 |
* 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 mm: compaction: e... |
1684 |
*/ |
32aaf0553 mm/compaction.c: ... |
1685 |
static isolate_migrate_t isolate_migratepages(struct compact_control *cc) |
ff9543fd3 mm: compaction: e... |
1686 |
{ |
e1409c325 mm/compaction: pa... |
1687 1688 1689 |
unsigned long block_start_pfn; unsigned long block_end_pfn; unsigned long low_pfn; |
edc2ca612 mm, compaction: m... |
1690 1691 |
struct page *page; const isolate_mode_t isolate_mode = |
5bbe3547a mm: allow compact... |
1692 |
(sysctl_compact_unevictable_allowed ? ISOLATE_UNEVICTABLE : 0) | |
1d2047fef mm, compaction: d... |
1693 |
(cc->mode != MIGRATE_SYNC ? ISOLATE_ASYNC_MIGRATE : 0); |
70b44595e mm, compaction: u... |
1694 |
bool fast_find_block; |
ff9543fd3 mm: compaction: e... |
1695 |
|
edc2ca612 mm, compaction: m... |
1696 1697 |
/* * Start at where we last stopped, or beginning of the zone as |
70b44595e mm, compaction: u... |
1698 1699 |
* initialized by compact_zone(). The first failure will use * the lowest PFN as the starting point for linear scanning. |
edc2ca612 mm, compaction: m... |
1700 |
*/ |
70b44595e mm, compaction: u... |
1701 |
low_pfn = fast_find_migrateblock(cc); |
06b6640a3 mm, compaction: w... |
1702 |
block_start_pfn = pageblock_start_pfn(low_pfn); |
32aaf0553 mm/compaction.c: ... |
1703 1704 |
if (block_start_pfn < cc->zone->zone_start_pfn) block_start_pfn = cc->zone->zone_start_pfn; |
ff9543fd3 mm: compaction: e... |
1705 |
|
70b44595e mm, compaction: u... |
1706 1707 1708 1709 1710 1711 |
/* * 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 mm: compaction: e... |
1712 |
/* Only scan within a pageblock boundary */ |
06b6640a3 mm, compaction: w... |
1713 |
block_end_pfn = pageblock_end_pfn(low_pfn); |
ff9543fd3 mm: compaction: e... |
1714 |
|
edc2ca612 mm, compaction: m... |
1715 1716 1717 1718 |
/* * Iterate over whole pageblocks until we find the first suitable. * Do not cross the free scanner. */ |
e1409c325 mm/compaction: pa... |
1719 |
for (; block_end_pfn <= cc->free_pfn; |
70b44595e mm, compaction: u... |
1720 |
fast_find_block = false, |
e1409c325 mm/compaction: pa... |
1721 1722 1723 |
low_pfn = block_end_pfn, block_start_pfn = block_end_pfn, block_end_pfn += pageblock_nr_pages) { |
ff9543fd3 mm: compaction: e... |
1724 |
|
edc2ca612 mm, compaction: m... |
1725 1726 1727 |
/* * This can potentially iterate a massively long zone with * many pageblocks unsuitable, so periodically check if we |
cb810ad29 mm, compaction: r... |
1728 |
* need to schedule. |
edc2ca612 mm, compaction: m... |
1729 |
*/ |
cb810ad29 mm, compaction: r... |
1730 |
if (!(low_pfn % (SWAP_CLUSTER_MAX * pageblock_nr_pages))) |
cf66f0700 mm, compaction: d... |
1731 |
cond_resched(); |
ff9543fd3 mm: compaction: e... |
1732 |
|
32aaf0553 mm/compaction.c: ... |
1733 1734 |
page = pageblock_pfn_to_page(block_start_pfn, block_end_pfn, cc->zone); |
7d49d8868 mm, compaction: r... |
1735 |
if (!page) |
edc2ca612 mm, compaction: m... |
1736 |
continue; |
e380bebe4 mm, compaction: k... |
1737 1738 1739 1740 1741 1742 1743 1744 1745 |
/* * 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 mm, compaction: m... |
1746 1747 1748 |
continue; /* |
9bebefd59 mm, compaction: c... |
1749 1750 1751 1752 1753 1754 |
* 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 mm, compaction: m... |
1755 |
*/ |
9bebefd59 mm, compaction: c... |
1756 1757 |
if (!suitable_migration_source(cc, page)) { update_cached_migrate(cc, block_end_pfn); |
edc2ca612 mm, compaction: m... |
1758 |
continue; |
9bebefd59 mm, compaction: c... |
1759 |
} |
edc2ca612 mm, compaction: m... |
1760 1761 |
/* Perform the isolation */ |
e1409c325 mm/compaction: pa... |
1762 1763 |
low_pfn = isolate_migratepages_block(cc, low_pfn, block_end_pfn, isolate_mode); |
edc2ca612 mm, compaction: m... |
1764 |
|
cb2dcaf02 mm, compaction: f... |
1765 |
if (!low_pfn) |
edc2ca612 mm, compaction: m... |
1766 1767 1768 1769 1770 1771 1772 1773 1774 |
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; } |
f2849aa09 mm, compaction: m... |
1775 1776 |
/* Record where migration scanner will be restarted. */ cc->migrate_pfn = low_pfn; |
ff9543fd3 mm: compaction: e... |
1777 |
|
edc2ca612 mm, compaction: m... |
1778 |
return cc->nr_migratepages ? ISOLATE_SUCCESS : ISOLATE_NONE; |
ff9543fd3 mm: compaction: e... |
1779 |
} |
21c527a3c mm/compaction.c: ... |
1780 1781 1782 1783 1784 1785 1786 1787 |
/* * 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 mm: proactive com... |
1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 |
static bool kswapd_is_running(pg_data_t *pgdat) { return pgdat->kswapd && (pgdat->kswapd->state == TASK_RUNNING); } /* * A 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]. * * 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. */ |
d34c0a759 mm: use unsigned ... |
1803 |
static unsigned int fragmentation_score_zone(struct zone *zone) |
facdaa917 mm: proactive com... |
1804 1805 1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 |
{ unsigned long score; score = zone->present_pages * extfrag_for_order(zone, COMPACTION_HPAGE_ORDER); 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 mm: use unsigned ... |
1819 |
static unsigned int fragmentation_score_node(pg_data_t *pgdat) |
facdaa917 mm: proactive com... |
1820 |
{ |
d34c0a759 mm: use unsigned ... |
1821 |
unsigned int score = 0; |
facdaa917 mm: proactive com... |
1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 |
int zoneid; for (zoneid = 0; zoneid < MAX_NR_ZONES; zoneid++) { struct zone *zone; zone = &pgdat->node_zones[zoneid]; score += fragmentation_score_zone(zone); } return score; } |
d34c0a759 mm: use unsigned ... |
1833 |
static unsigned int fragmentation_score_wmark(pg_data_t *pgdat, bool low) |
facdaa917 mm: proactive com... |
1834 |
{ |
d34c0a759 mm: use unsigned ... |
1835 |
unsigned int wmark_low; |
facdaa917 mm: proactive com... |
1836 1837 1838 1839 1840 1841 |
/* * Cap the low watermak to avoid excessive compaction * activity in case a user sets the proactivess tunable * close to 100 (maximum). */ |
d34c0a759 mm: use unsigned ... |
1842 1843 |
wmark_low = max(100U - sysctl_compaction_proactiveness, 5U); return low ? wmark_low : min(wmark_low + 10, 100U); |
facdaa917 mm: proactive com... |
1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 |
} 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 mm, compaction: r... |
1856 |
static enum compact_result __compact_finished(struct compact_control *cc) |
748446bb6 mm: compaction: m... |
1857 |
{ |
8fb74b9fb mm: compaction: p... |
1858 |
unsigned int order; |
d39773a06 mm, compaction: a... |
1859 |
const int migratetype = cc->migratetype; |
cb2dcaf02 mm, compaction: f... |
1860 |
int ret; |
748446bb6 mm: compaction: m... |
1861 |
|
753341a4b revert "mm: have ... |
1862 |
/* Compaction run completes if the migrate and free scanner meet */ |
f2849aa09 mm, compaction: m... |
1863 |
if (compact_scanners_met(cc)) { |
55b7c4c99 mm: compaction: r... |
1864 |
/* Let the next compaction start anew. */ |
40cacbcb3 mm, compaction: r... |
1865 |
reset_cached_positions(cc->zone); |
55b7c4c99 mm: compaction: r... |
1866 |
|
62997027c mm: compaction: c... |
1867 1868 |
/* * Mark that the PG_migrate_skip information should be cleared |
accf62422 mm, kswapd: repla... |
1869 |
* by kswapd when it goes to sleep. kcompactd does not set the |
62997027c mm: compaction: c... |
1870 1871 1872 |
* flag itself as the decision to be clear should be directly * based on an allocation request. */ |
accf62422 mm, kswapd: repla... |
1873 |
if (cc->direct_compaction) |
40cacbcb3 mm, compaction: r... |
1874 |
cc->zone->compact_blockskip_flush = true; |
62997027c mm: compaction: c... |
1875 |
|
c8f7de0bf mm, compaction: d... |
1876 1877 1878 1879 |
if (cc->whole_zone) return COMPACT_COMPLETE; else return COMPACT_PARTIAL_SKIPPED; |
bb13ffeb9 mm: compaction: c... |
1880 |
} |
748446bb6 mm: compaction: m... |
1881 |
|
facdaa917 mm: proactive com... |
1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 |
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 mm/compaction.c: ... |
1900 |
if (is_via_compact_memory(cc->order)) |
56de7263f mm: compaction: d... |
1901 |
return COMPACT_CONTINUE; |
efe771c76 mm, compaction: a... |
1902 1903 1904 1905 1906 1907 1908 1909 |
/* * 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 mm, compaction: f... |
1910 |
|
56de7263f mm: compaction: d... |
1911 |
/* Direct compactor: Is a suitable page free? */ |
cb2dcaf02 mm, compaction: f... |
1912 |
ret = COMPACT_NO_SUITABLE_PAGE; |
8fb74b9fb mm: compaction: p... |
1913 |
for (order = cc->order; order < MAX_ORDER; order++) { |
40cacbcb3 mm, compaction: r... |
1914 |
struct free_area *area = &cc->zone->free_area[order]; |
2149cdaef mm/compaction: en... |
1915 |
bool can_steal; |
8fb74b9fb mm: compaction: p... |
1916 1917 |
/* Job done if page is free of the right migratetype */ |
b03641af6 mm: move buddy li... |
1918 |
if (!free_area_empty(area, migratetype)) |
cf378319d mm, compaction: r... |
1919 |
return COMPACT_SUCCESS; |
8fb74b9fb mm: compaction: p... |
1920 |
|
2149cdaef mm/compaction: en... |
1921 1922 1923 |
#ifdef CONFIG_CMA /* MIGRATE_MOVABLE can fallback on MIGRATE_CMA */ if (migratetype == MIGRATE_MOVABLE && |
b03641af6 mm: move buddy li... |
1924 |
!free_area_empty(area, MIGRATE_CMA)) |
cf378319d mm, compaction: r... |
1925 |
return COMPACT_SUCCESS; |
2149cdaef mm/compaction: en... |
1926 1927 1928 1929 1930 1931 |
#endif /* * Job done if allocation would steal freepages from * other migratetype buddy lists. */ if (find_suitable_fallback(area, order, migratetype, |
baf6a9a1d mm, compaction: f... |
1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 |
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 mm, compaction: f... |
1951 1952 |
ret = COMPACT_CONTINUE; break; |
baf6a9a1d mm, compaction: f... |
1953 |
} |
56de7263f mm: compaction: d... |
1954 |
} |
facdaa917 mm: proactive com... |
1955 |
out: |
cb2dcaf02 mm, compaction: f... |
1956 1957 1958 1959 |
if (cc->contended || fatal_signal_pending(current)) ret = COMPACT_CONTENDED; return ret; |
837d026d5 mm/compaction: mo... |
1960 |
} |
40cacbcb3 mm, compaction: r... |
1961 |
static enum compact_result compact_finished(struct compact_control *cc) |
837d026d5 mm/compaction: mo... |
1962 1963 |
{ int ret; |
40cacbcb3 mm, compaction: r... |
1964 1965 |
ret = __compact_finished(cc); trace_mm_compaction_finished(cc->zone, cc->order, ret); |
837d026d5 mm/compaction: mo... |
1966 1967 1968 1969 |
if (ret == COMPACT_NO_SUITABLE_PAGE) ret = COMPACT_CONTINUE; return ret; |
748446bb6 mm: compaction: m... |
1970 |
} |
3e7d34497 mm: vmscan: recla... |
1971 1972 1973 1974 |
/* * compaction_suitable: Is this suitable to run compaction on this zone now? * Returns * COMPACT_SKIPPED - If there are too few free pages for compaction |
cf378319d mm, compaction: r... |
1975 |
* COMPACT_SUCCESS - If the allocation would succeed without compaction |
3e7d34497 mm: vmscan: recla... |
1976 1977 |
* COMPACT_CONTINUE - If compaction should run now */ |
ea7ab982b mm, compaction: c... |
1978 |
static enum compact_result __compaction_suitable(struct zone *zone, int order, |
c603844bd mm, page_alloc: c... |
1979 |
unsigned int alloc_flags, |
97a225e69 mm/page_alloc: in... |
1980 |
int highest_zoneidx, |
86a294a81 mm, oom, compacti... |
1981 |
unsigned long wmark_target) |
3e7d34497 mm: vmscan: recla... |
1982 |
{ |
3e7d34497 mm: vmscan: recla... |
1983 |
unsigned long watermark; |
21c527a3c mm/compaction.c: ... |
1984 |
if (is_via_compact_memory(order)) |
3957c7768 mm: compaction: f... |
1985 |
return COMPACT_CONTINUE; |
a92144438 mm: move zone wat... |
1986 |
watermark = wmark_pages(zone, alloc_flags & ALLOC_WMARK_MASK); |
ebff39801 mm, compaction: p... |
1987 1988 1989 1990 |
/* * If watermarks for high-order allocation are already met, there * should be no need for compaction at all. */ |
97a225e69 mm/page_alloc: in... |
1991 |
if (zone_watermark_ok(zone, order, watermark, highest_zoneidx, |
ebff39801 mm, compaction: p... |
1992 |
alloc_flags)) |
cf378319d mm, compaction: r... |
1993 |
return COMPACT_SUCCESS; |
ebff39801 mm, compaction: p... |
1994 |
|
3957c7768 mm: compaction: f... |
1995 |
/* |
9861a62c3 mm, compaction: c... |
1996 |
* Watermarks for order-0 must be met for compaction to be able to |
984fdba6a mm, compaction: u... |
1997 1998 1999 2000 |
* 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 mm/page_alloc: in... |
2001 2002 2003 |
* 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 mm, compaction: r... |
2004 2005 |
* For costly orders, we require low watermark instead of min for * compaction to proceed to increase its chances. |
d883c6cf3 Revert "mm/cma: m... |
2006 2007 |
* ALLOC_CMA is used, as pages in CMA pageblocks are considered * suitable migration targets |
3e7d34497 mm: vmscan: recla... |
2008 |
*/ |
8348faf91 mm, compaction: r... |
2009 2010 2011 |
watermark = (order > PAGE_ALLOC_COSTLY_ORDER) ? low_wmark_pages(zone) : min_wmark_pages(zone); watermark += compact_gap(order); |
97a225e69 mm/page_alloc: in... |
2012 |
if (!__zone_watermark_ok(zone, 0, watermark, highest_zoneidx, |
d883c6cf3 Revert "mm/cma: m... |
2013 |
ALLOC_CMA, wmark_target)) |
3e7d34497 mm: vmscan: recla... |
2014 |
return COMPACT_SKIPPED; |
cc5c9f098 mm, compaction: i... |
2015 2016 2017 2018 2019 |
return COMPACT_CONTINUE; } enum compact_result compaction_suitable(struct zone *zone, int order, unsigned int alloc_flags, |
97a225e69 mm/page_alloc: in... |
2020 |
int highest_zoneidx) |
cc5c9f098 mm, compaction: i... |
2021 2022 2023 |
{ enum compact_result ret; int fragindex; |
97a225e69 mm/page_alloc: in... |
2024 |
ret = __compaction_suitable(zone, order, alloc_flags, highest_zoneidx, |
cc5c9f098 mm, compaction: i... |
2025 |
zone_page_state(zone, NR_FREE_PAGES)); |
3e7d34497 mm: vmscan: recla... |
2026 2027 2028 2029 |
/* * fragmentation index determines if allocation failures are due to * low memory or external fragmentation * |
ebff39801 mm, compaction: p... |
2030 2031 |
* index of -1000 would imply allocations might succeed depending on * watermarks, but we already failed the high-order watermark check |
3e7d34497 mm: vmscan: recla... |
2032 2033 2034 |
* index towards 0 implies failure is due to lack of memory * index towards 1000 implies failure is due to fragmentation * |
203114202 mm, compaction: r... |
2035 2036 2037 2038 2039 2040 |
* 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 mm: vmscan: recla... |
2041 |
*/ |
203114202 mm, compaction: r... |
2042 |
if (ret == COMPACT_CONTINUE && (order > PAGE_ALLOC_COSTLY_ORDER)) { |
cc5c9f098 mm, compaction: i... |
2043 2044 2045 2046 |
fragindex = fragmentation_index(zone, order); if (fragindex >= 0 && fragindex <= sysctl_extfrag_threshold) ret = COMPACT_NOT_SUITABLE_ZONE; } |
837d026d5 mm/compaction: mo... |
2047 |
|
837d026d5 mm/compaction: mo... |
2048 2049 2050 2051 2052 2053 |
trace_mm_compaction_suitable(zone, order, ret); if (ret == COMPACT_NOT_SUITABLE_ZONE) ret = COMPACT_SKIPPED; return ret; } |
86a294a81 mm, oom, compacti... |
2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 |
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 mm/page_alloc: in... |
2064 2065 |
for_each_zone_zonelist_nodemask(zone, z, ac->zonelist, ac->highest_zoneidx, ac->nodemask) { |
86a294a81 mm, oom, compacti... |
2066 2067 2068 2069 2070 2071 2072 2073 2074 |
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 mm: remove reclai... |
2075 |
available = zone_reclaimable_pages(zone) / order; |
86a294a81 mm, oom, compacti... |
2076 2077 |
available += zone_page_state_snapshot(zone, NR_FREE_PAGES); compact_result = __compaction_suitable(zone, order, alloc_flags, |
97a225e69 mm/page_alloc: in... |
2078 |
ac->highest_zoneidx, available); |
cc5c9f098 mm, compaction: i... |
2079 |
if (compact_result != COMPACT_SKIPPED) |
86a294a81 mm, oom, compacti... |
2080 2081 2082 2083 2084 |
return true; } return false; } |
5e1f0f098 mm, compaction: c... |
2085 2086 |
static enum compact_result compact_zone(struct compact_control *cc, struct capture_control *capc) |
748446bb6 mm: compaction: m... |
2087 |
{ |
ea7ab982b mm, compaction: c... |
2088 |
enum compact_result ret; |
40cacbcb3 mm, compaction: r... |
2089 2090 |
unsigned long start_pfn = cc->zone->zone_start_pfn; unsigned long end_pfn = zone_end_pfn(cc->zone); |
566e54e11 mm, compaction: r... |
2091 |
unsigned long last_migrated_pfn; |
e0b9daeb4 mm, compaction: e... |
2092 |
const bool sync = cc->mode != MIGRATE_ASYNC; |
8854c55f5 mm, compaction: k... |
2093 |
bool update_cached; |
748446bb6 mm: compaction: m... |
2094 |
|
a94b52524 mm/compaction.c: ... |
2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 |
/* * 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 mm: rename gfpfla... |
2105 |
cc->migratetype = gfp_migratetype(cc->gfp_mask); |
40cacbcb3 mm, compaction: r... |
2106 |
ret = compaction_suitable(cc->zone, cc->order, cc->alloc_flags, |
97a225e69 mm/page_alloc: in... |
2107 |
cc->highest_zoneidx); |
c46649dea mm, compaction: c... |
2108 |
/* Compaction is likely to fail */ |
cf378319d mm, compaction: r... |
2109 |
if (ret == COMPACT_SUCCESS || ret == COMPACT_SKIPPED) |
3e7d34497 mm: vmscan: recla... |
2110 |
return ret; |
c46649dea mm, compaction: c... |
2111 2112 2113 |
/* huh, compaction_suitable is returning something unexpected */ VM_BUG_ON(ret != COMPACT_CONTINUE); |
3e7d34497 mm: vmscan: recla... |
2114 |
|
c89511ab2 mm: compaction: R... |
2115 |
/* |
d3132e4b8 mm: compaction: r... |
2116 |
* Clear pageblock skip if there were failures recently and compaction |
accf62422 mm, kswapd: repla... |
2117 |
* is about to be retried after being deferred. |
d3132e4b8 mm: compaction: r... |
2118 |
*/ |
40cacbcb3 mm, compaction: r... |
2119 2120 |
if (compaction_restarting(cc->zone, cc->order)) __reset_isolation_suitable(cc->zone); |
d3132e4b8 mm: compaction: r... |
2121 2122 |
/* |
c89511ab2 mm: compaction: R... |
2123 |
* Setup to move all movable pages to the end of the zone. Used cached |
06ed29989 mm, compaction: m... |
2124 2125 2126 |
* 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 mm: compaction: R... |
2127 |
*/ |
70b44595e mm, compaction: u... |
2128 |
cc->fast_start_pfn = 0; |
06ed29989 mm, compaction: m... |
2129 |
if (cc->whole_zone) { |
c89511ab2 mm: compaction: R... |
2130 |
cc->migrate_pfn = start_pfn; |
06ed29989 mm, compaction: m... |
2131 2132 |
cc->free_pfn = pageblock_start_pfn(end_pfn - 1); } else { |
40cacbcb3 mm, compaction: r... |
2133 2134 |
cc->migrate_pfn = cc->zone->compact_cached_migrate_pfn[sync]; cc->free_pfn = cc->zone->compact_cached_free_pfn; |
06ed29989 mm, compaction: m... |
2135 2136 |
if (cc->free_pfn < start_pfn || cc->free_pfn >= end_pfn) { cc->free_pfn = pageblock_start_pfn(end_pfn - 1); |
40cacbcb3 mm, compaction: r... |
2137 |
cc->zone->compact_cached_free_pfn = cc->free_pfn; |
06ed29989 mm, compaction: m... |
2138 2139 2140 |
} if (cc->migrate_pfn < start_pfn || cc->migrate_pfn >= end_pfn) { cc->migrate_pfn = start_pfn; |
40cacbcb3 mm, compaction: r... |
2141 2142 |
cc->zone->compact_cached_migrate_pfn[0] = cc->migrate_pfn; cc->zone->compact_cached_migrate_pfn[1] = cc->migrate_pfn; |
06ed29989 mm, compaction: m... |
2143 |
} |
c8f7de0bf mm, compaction: d... |
2144 |
|
e332f741a mm, compaction: b... |
2145 |
if (cc->migrate_pfn <= cc->zone->compact_init_migrate_pfn) |
06ed29989 mm, compaction: m... |
2146 2147 |
cc->whole_zone = true; } |
c8f7de0bf mm, compaction: d... |
2148 |
|
566e54e11 mm, compaction: r... |
2149 |
last_migrated_pfn = 0; |
748446bb6 mm: compaction: m... |
2150 |
|
8854c55f5 mm, compaction: k... |
2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 |
/* * 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 mm/compaction: en... |
2161 2162 |
trace_mm_compaction_begin(start_pfn, cc->migrate_pfn, cc->free_pfn, end_pfn, sync); |
0eb927c0a mm: compaction: t... |
2163 |
|
748446bb6 mm: compaction: m... |
2164 |
migrate_prep_local(); |
40cacbcb3 mm, compaction: r... |
2165 |
while ((ret = compact_finished(cc)) == COMPACT_CONTINUE) { |
9d502c1c8 mm/compaction: ch... |
2166 |
int err; |
566e54e11 mm, compaction: r... |
2167 |
unsigned long start_pfn = cc->migrate_pfn; |
748446bb6 mm: compaction: m... |
2168 |
|
804d3121b mm, compaction: a... |
2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 |
/* * 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) == pageblock_start_pfn(start_pfn)) { cc->rescan = true; } |
32aaf0553 mm/compaction.c: ... |
2182 |
switch (isolate_migratepages(cc)) { |
f9e35b3b4 mm: compaction: a... |
2183 |
case ISOLATE_ABORT: |
2d1e10412 mm, compaction: d... |
2184 |
ret = COMPACT_CONTENDED; |
5733c7d11 mm: introduce put... |
2185 |
putback_movable_pages(&cc->migratepages); |
e64c5237c mm: compaction: a... |
2186 |
cc->nr_migratepages = 0; |
f9e35b3b4 mm: compaction: a... |
2187 2188 |
goto out; case ISOLATE_NONE: |
8854c55f5 mm, compaction: k... |
2189 2190 2191 2192 |
if (update_cached) { cc->zone->compact_cached_migrate_pfn[1] = cc->zone->compact_cached_migrate_pfn[0]; } |
fdaf7f5c4 mm, compaction: m... |
2193 2194 2195 2196 2197 2198 |
/* * We haven't isolated and migrated anything, but * there might still be unflushed migrations from * previous cc->order aligned block. */ goto check_drain; |
f9e35b3b4 mm: compaction: a... |
2199 |
case ISOLATE_SUCCESS: |
8854c55f5 mm, compaction: k... |
2200 |
update_cached = false; |
566e54e11 mm, compaction: r... |
2201 |
last_migrated_pfn = start_pfn; |
f9e35b3b4 mm: compaction: a... |
2202 2203 |
; } |
748446bb6 mm: compaction: m... |
2204 |
|
d53aea3d4 mm, compaction: r... |
2205 |
err = migrate_pages(&cc->migratepages, compaction_alloc, |
e0b9daeb4 mm, compaction: e... |
2206 |
compaction_free, (unsigned long)cc, cc->mode, |
7b2a2d4a1 mm: migrate: Add ... |
2207 |
MR_COMPACTION); |
748446bb6 mm: compaction: m... |
2208 |
|
f8c9301fa mm/compaction: do... |
2209 2210 |
trace_mm_compaction_migratepages(cc->nr_migratepages, err, &cc->migratepages); |
748446bb6 mm: compaction: m... |
2211 |
|
f8c9301fa mm/compaction: do... |
2212 2213 |
/* All pages were either migrated or will be released */ cc->nr_migratepages = 0; |
9d502c1c8 mm/compaction: ch... |
2214 |
if (err) { |
5733c7d11 mm: introduce put... |
2215 |
putback_movable_pages(&cc->migratepages); |
7ed695e06 mm: compaction: d... |
2216 2217 2218 2219 |
/* * migrate_pages() may return -ENOMEM when scanners meet * and we want compact_finished() to detect it */ |
f2849aa09 mm, compaction: m... |
2220 |
if (err == -ENOMEM && !compact_scanners_met(cc)) { |
2d1e10412 mm, compaction: d... |
2221 |
ret = COMPACT_CONTENDED; |
4bf2bba37 mm, thp: abort co... |
2222 2223 |
goto out; } |
fdd048e12 mm, compaction: s... |
2224 2225 2226 2227 2228 2229 2230 2231 2232 |
/* * 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 mm, compaction: r... |
2233 |
last_migrated_pfn = 0; |
fdd048e12 mm, compaction: s... |
2234 |
} |
748446bb6 mm: compaction: m... |
2235 |
} |
fdaf7f5c4 mm, compaction: m... |
2236 |
|
fdaf7f5c4 mm, compaction: m... |
2237 2238 2239 2240 2241 2242 2243 2244 |
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 mm, compaction: r... |
2245 |
if (cc->order > 0 && last_migrated_pfn) { |
fdaf7f5c4 mm, compaction: m... |
2246 |
unsigned long current_block_start = |
06b6640a3 mm, compaction: w... |
2247 |
block_start_pfn(cc->migrate_pfn, cc->order); |
fdaf7f5c4 mm, compaction: m... |
2248 |
|
566e54e11 mm, compaction: r... |
2249 |
if (last_migrated_pfn < current_block_start) { |
b01b21419 mm/swap: Use loca... |
2250 |
lru_add_drain_cpu_zone(cc->zone); |
fdaf7f5c4 mm, compaction: m... |
2251 |
/* No more flushing until we migrate again */ |
566e54e11 mm, compaction: r... |
2252 |
last_migrated_pfn = 0; |
fdaf7f5c4 mm, compaction: m... |
2253 2254 |
} } |
5e1f0f098 mm, compaction: c... |
2255 2256 2257 2258 2259 |
/* Stop if a page has been captured */ if (capc && capc->page) { ret = COMPACT_SUCCESS; break; } |
748446bb6 mm: compaction: m... |
2260 |
} |
f9e35b3b4 mm: compaction: a... |
2261 |
out: |
6bace090a mm, compaction: a... |
2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 |
/* * 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 mm, compaction: w... |
2272 |
free_pfn = pageblock_start_pfn(free_pfn); |
6bace090a mm, compaction: a... |
2273 2274 2275 2276 |
/* * Only go back, not forward. The cached pfn might have been * already reset to zone end in compact_finished() */ |
40cacbcb3 mm, compaction: r... |
2277 2278 |
if (free_pfn > cc->zone->compact_cached_free_pfn) cc->zone->compact_cached_free_pfn = free_pfn; |
6bace090a mm, compaction: a... |
2279 |
} |
748446bb6 mm: compaction: m... |
2280 |
|
7f354a548 mm, compaction: a... |
2281 2282 |
count_compact_events(COMPACTMIGRATE_SCANNED, cc->total_migrate_scanned); count_compact_events(COMPACTFREE_SCANNED, cc->total_free_scanned); |
16c4a097a mm/compaction: en... |
2283 2284 |
trace_mm_compaction_end(start_pfn, cc->migrate_pfn, cc->free_pfn, end_pfn, sync, ret); |
0eb927c0a mm: compaction: t... |
2285 |
|
748446bb6 mm: compaction: m... |
2286 2287 |
return ret; } |
76ab0f530 mm: compaction: a... |
2288 |
|
ea7ab982b mm, compaction: c... |
2289 |
static enum compact_result compact_zone_order(struct zone *zone, int order, |
c3486f537 mm, compaction: s... |
2290 |
gfp_t gfp_mask, enum compact_priority prio, |
97a225e69 mm/page_alloc: in... |
2291 |
unsigned int alloc_flags, int highest_zoneidx, |
5e1f0f098 mm, compaction: c... |
2292 |
struct page **capture) |
56de7263f mm: compaction: d... |
2293 |
{ |
ea7ab982b mm, compaction: c... |
2294 |
enum compact_result ret; |
56de7263f mm: compaction: d... |
2295 |
struct compact_control cc = { |
56de7263f mm: compaction: d... |
2296 |
.order = order, |
dbe2d4e4f mm, compaction: r... |
2297 |
.search_order = order, |
6d7ce5594 mm, compaction: p... |
2298 |
.gfp_mask = gfp_mask, |
56de7263f mm: compaction: d... |
2299 |
.zone = zone, |
a5508cd83 mm, compaction: i... |
2300 2301 |
.mode = (prio == COMPACT_PRIO_ASYNC) ? MIGRATE_ASYNC : MIGRATE_SYNC_LIGHT, |
ebff39801 mm, compaction: p... |
2302 |
.alloc_flags = alloc_flags, |
97a225e69 mm/page_alloc: in... |
2303 |
.highest_zoneidx = highest_zoneidx, |
accf62422 mm, kswapd: repla... |
2304 |
.direct_compaction = true, |
a8e025e55 mm, compaction: a... |
2305 |
.whole_zone = (prio == MIN_COMPACT_PRIORITY), |
9f7e33879 mm, compaction: m... |
2306 2307 |
.ignore_skip_hint = (prio == MIN_COMPACT_PRIORITY), .ignore_block_suitable = (prio == MIN_COMPACT_PRIORITY) |
56de7263f mm: compaction: d... |
2308 |
}; |
5e1f0f098 mm, compaction: c... |
2309 2310 2311 2312 |
struct capture_control capc = { .cc = &cc, .page = NULL, }; |
b9e20f0da mm, compaction: m... |
2313 2314 2315 2316 2317 2318 2319 |
/* * 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 mm: compaction: d... |
2320 |
|
5e1f0f098 mm, compaction: c... |
2321 |
ret = compact_zone(&cc, &capc); |
e64c5237c mm: compaction: a... |
2322 2323 2324 |
VM_BUG_ON(!list_empty(&cc.freepages)); VM_BUG_ON(!list_empty(&cc.migratepages)); |
b9e20f0da mm, compaction: m... |
2325 2326 2327 2328 2329 2330 2331 |
/* * 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); |
5e1f0f098 mm, compaction: c... |
2332 |
|
e64c5237c mm: compaction: a... |
2333 |
return ret; |
56de7263f mm: compaction: d... |
2334 |
} |
5e7719058 mm: compaction: a... |
2335 |
int sysctl_extfrag_threshold = 500; |
56de7263f mm: compaction: d... |
2336 2337 |
/** * try_to_compact_pages - Direct compact to satisfy a high-order allocation |
56de7263f mm: compaction: d... |
2338 |
* @gfp_mask: The GFP mask of the current allocation |
1a6d53a10 mm: reduce try_to... |
2339 2340 2341 |
* @order: The order of the current allocation * @alloc_flags: The allocation flags of the current allocation * @ac: The context of current allocation |
112d2d29f mm/compaction.c: ... |
2342 |
* @prio: Determines how hard direct compaction should try to succeed |
6467552ca mm, compaction: f... |
2343 |
* @capture: Pointer to free page created by compaction will be stored here |
56de7263f mm: compaction: d... |
2344 2345 2346 |
* * This is the main entry point for direct page compaction. */ |
ea7ab982b mm, compaction: c... |
2347 |
enum compact_result try_to_compact_pages(gfp_t gfp_mask, unsigned int order, |
c603844bd mm, page_alloc: c... |
2348 |
unsigned int alloc_flags, const struct alloc_context *ac, |
5e1f0f098 mm, compaction: c... |
2349 |
enum compact_priority prio, struct page **capture) |
56de7263f mm: compaction: d... |
2350 |
{ |
56de7263f mm: compaction: d... |
2351 |
int may_perform_io = gfp_mask & __GFP_IO; |
56de7263f mm: compaction: d... |
2352 2353 |
struct zoneref *z; struct zone *zone; |
1d4746d39 mm, compaction: d... |
2354 |
enum compact_result rc = COMPACT_SKIPPED; |
56de7263f mm: compaction: d... |
2355 |
|
73e64c51a mm, compaction: a... |
2356 2357 2358 2359 2360 |
/* * Check if the GFP flags allow compaction - GFP_NOIO is really * tricky context because the migration might require IO */ if (!may_perform_io) |
53853e2d2 mm, compaction: d... |
2361 |
return COMPACT_SKIPPED; |
56de7263f mm: compaction: d... |
2362 |
|
a5508cd83 mm, compaction: i... |
2363 |
trace_mm_compaction_try_to_compact_pages(order, gfp_mask, prio); |
837d026d5 mm/compaction: mo... |
2364 |
|
56de7263f mm: compaction: d... |
2365 |
/* Compact each zone in the list */ |
97a225e69 mm/page_alloc: in... |
2366 2367 |
for_each_zone_zonelist_nodemask(zone, z, ac->zonelist, ac->highest_zoneidx, ac->nodemask) { |
ea7ab982b mm, compaction: c... |
2368 |
enum compact_result status; |
56de7263f mm: compaction: d... |
2369 |
|
a8e025e55 mm, compaction: a... |
2370 2371 |
if (prio > MIN_COMPACT_PRIORITY && compaction_deferred(zone, order)) { |
1d4746d39 mm, compaction: d... |
2372 |
rc = max_t(enum compact_result, COMPACT_DEFERRED, rc); |
53853e2d2 mm, compaction: d... |
2373 |
continue; |
1d4746d39 mm, compaction: d... |
2374 |
} |
53853e2d2 mm, compaction: d... |
2375 |
|
a5508cd83 mm, compaction: i... |
2376 |
status = compact_zone_order(zone, order, gfp_mask, prio, |
97a225e69 mm/page_alloc: in... |
2377 |
alloc_flags, ac->highest_zoneidx, capture); |
56de7263f mm: compaction: d... |
2378 |
rc = max(status, rc); |
7ceb009a2 mm, compaction: d... |
2379 2380 |
/* The allocation should succeed, stop compacting */ if (status == COMPACT_SUCCESS) { |
53853e2d2 mm, compaction: d... |
2381 2382 2383 2384 2385 2386 2387 |
/* * 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 mm, compaction: k... |
2388 |
|
c3486f537 mm, compaction: s... |
2389 |
break; |
1f9efdef4 mm, compaction: k... |
2390 |
} |
a5508cd83 mm, compaction: i... |
2391 |
if (prio != COMPACT_PRIO_ASYNC && (status == COMPACT_COMPLETE || |
c3486f537 mm, compaction: s... |
2392 |
status == COMPACT_PARTIAL_SKIPPED)) |
53853e2d2 mm, compaction: d... |
2393 2394 2395 2396 2397 2398 |
/* * 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 mm, compaction: k... |
2399 2400 2401 2402 |
/* * We might have stopped compacting due to need_resched() in * async compaction, or due to a fatal signal detected. In that |
c3486f537 mm, compaction: s... |
2403 |
* case do not try further zones |
1f9efdef4 mm, compaction: k... |
2404 |
*/ |
c3486f537 mm, compaction: s... |
2405 2406 2407 |
if ((prio == COMPACT_PRIO_ASYNC && need_resched()) || fatal_signal_pending(current)) break; |
56de7263f mm: compaction: d... |
2408 2409 2410 2411 |
} return rc; } |
facdaa917 mm: proactive com... |
2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 2442 2443 2444 2445 2446 |
/* * 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 mm: compaction: d... |
2447 |
|
76ab0f530 mm: compaction: a... |
2448 |
/* Compact all zones within a node */ |
791cae962 mm, compaction: c... |
2449 |
static void compact_node(int nid) |
76ab0f530 mm: compaction: a... |
2450 |
{ |
791cae962 mm, compaction: c... |
2451 |
pg_data_t *pgdat = NODE_DATA(nid); |
76ab0f530 mm: compaction: a... |
2452 |
int zoneid; |
76ab0f530 mm: compaction: a... |
2453 |
struct zone *zone; |
791cae962 mm, compaction: c... |
2454 2455 2456 2457 2458 |
struct compact_control cc = { .order = -1, .mode = MIGRATE_SYNC, .ignore_skip_hint = true, .whole_zone = true, |
73e64c51a mm, compaction: a... |
2459 |
.gfp_mask = GFP_KERNEL, |
791cae962 mm, compaction: c... |
2460 |
}; |
76ab0f530 mm: compaction: a... |
2461 |
|
76ab0f530 mm: compaction: a... |
2462 |
for (zoneid = 0; zoneid < MAX_NR_ZONES; zoneid++) { |
76ab0f530 mm: compaction: a... |
2463 2464 2465 2466 |
zone = &pgdat->node_zones[zoneid]; if (!populated_zone(zone)) continue; |
791cae962 mm, compaction: c... |
2467 |
cc.zone = zone; |
76ab0f530 mm: compaction: a... |
2468 |
|
5e1f0f098 mm, compaction: c... |
2469 |
compact_zone(&cc, NULL); |
754693457 mm/compaction.c: ... |
2470 |
|
791cae962 mm, compaction: c... |
2471 2472 |
VM_BUG_ON(!list_empty(&cc.freepages)); VM_BUG_ON(!list_empty(&cc.migratepages)); |
76ab0f530 mm: compaction: a... |
2473 |
} |
76ab0f530 mm: compaction: a... |
2474 2475 2476 |
} /* Compact all nodes in the system */ |
7964c06d6 mm: compaction: f... |
2477 |
static void compact_nodes(void) |
76ab0f530 mm: compaction: a... |
2478 2479 |
{ int nid; |
8575ec29f compact_pgdat: wo... |
2480 2481 |
/* Flush pending updates to the LRU lists */ lru_add_drain_all(); |
76ab0f530 mm: compaction: a... |
2482 2483 |
for_each_online_node(nid) compact_node(nid); |
76ab0f530 mm: compaction: a... |
2484 2485 2486 2487 |
} /* The written value is actually unused, all memory is compacted */ int sysctl_compact_memory; |
fec4eb2c8 mm/compaction: im... |
2488 |
/* |
facdaa917 mm: proactive com... |
2489 2490 2491 2492 |
* 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 mm: use unsigned ... |
2493 |
unsigned int __read_mostly sysctl_compaction_proactiveness = 20; |
facdaa917 mm: proactive com... |
2494 2495 |
/* |
fec4eb2c8 mm/compaction: im... |
2496 2497 2498 |
* This is the entry point for compacting all nodes via * /proc/sys/vm/compact_memory */ |
76ab0f530 mm: compaction: a... |
2499 |
int sysctl_compaction_handler(struct ctl_table *table, int write, |
32927393d sysctl: pass kern... |
2500 |
void *buffer, size_t *length, loff_t *ppos) |
76ab0f530 mm: compaction: a... |
2501 2502 |
{ if (write) |
7964c06d6 mm: compaction: f... |
2503 |
compact_nodes(); |
76ab0f530 mm: compaction: a... |
2504 2505 2506 |
return 0; } |
ed4a6d7f0 mm: compaction: a... |
2507 2508 |
#if defined(CONFIG_SYSFS) && defined(CONFIG_NUMA) |
74e77fb9a mm/compaction.c: ... |
2509 |
static ssize_t sysfs_compact_node(struct device *dev, |
10fbcf4c6 convert 'memory' ... |
2510 |
struct device_attribute *attr, |
ed4a6d7f0 mm: compaction: a... |
2511 2512 |
const char *buf, size_t count) { |
8575ec29f compact_pgdat: wo... |
2513 2514 2515 2516 2517 2518 2519 2520 |
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 mm: compaction: a... |
2521 2522 2523 |
return count; } |
0825a6f98 mm: use octal not... |
2524 |
static DEVICE_ATTR(compact, 0200, NULL, sysfs_compact_node); |
ed4a6d7f0 mm: compaction: a... |
2525 2526 2527 |
int compaction_register_node(struct node *node) { |
10fbcf4c6 convert 'memory' ... |
2528 |
return device_create_file(&node->dev, &dev_attr_compact); |
ed4a6d7f0 mm: compaction: a... |
2529 2530 2531 2532 |
} void compaction_unregister_node(struct node *node) { |
10fbcf4c6 convert 'memory' ... |
2533 |
return device_remove_file(&node->dev, &dev_attr_compact); |
ed4a6d7f0 mm: compaction: a... |
2534 2535 |
} #endif /* CONFIG_SYSFS && CONFIG_NUMA */ |
ff9543fd3 mm: compaction: e... |
2536 |
|
698b1b306 mm, compaction: i... |
2537 2538 |
static inline bool kcompactd_work_requested(pg_data_t *pgdat) { |
172400c69 mm: fix kcompactd... |
2539 |
return pgdat->kcompactd_max_order > 0 || kthread_should_stop(); |
698b1b306 mm, compaction: i... |
2540 2541 2542 2543 2544 2545 |
} static bool kcompactd_node_suitable(pg_data_t *pgdat) { int zoneid; struct zone *zone; |
97a225e69 mm/page_alloc: in... |
2546 |
enum zone_type highest_zoneidx = pgdat->kcompactd_highest_zoneidx; |
698b1b306 mm, compaction: i... |
2547 |
|
97a225e69 mm/page_alloc: in... |
2548 |
for (zoneid = 0; zoneid <= highest_zoneidx; zoneid++) { |
698b1b306 mm, compaction: i... |
2549 2550 2551 2552 2553 2554 |
zone = &pgdat->node_zones[zoneid]; if (!populated_zone(zone)) continue; if (compaction_suitable(zone, pgdat->kcompactd_max_order, 0, |
97a225e69 mm/page_alloc: in... |
2555 |
highest_zoneidx) == COMPACT_CONTINUE) |
698b1b306 mm, compaction: i... |
2556 2557 2558 2559 2560 2561 2562 2563 2564 2565 2566 2567 2568 2569 2570 2571 |
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 mm, compaction: r... |
2572 |
.search_order = pgdat->kcompactd_max_order, |
97a225e69 mm/page_alloc: in... |
2573 |
.highest_zoneidx = pgdat->kcompactd_highest_zoneidx, |
698b1b306 mm, compaction: i... |
2574 |
.mode = MIGRATE_SYNC_LIGHT, |
a0647dc92 mm, compaction: k... |
2575 |
.ignore_skip_hint = false, |
73e64c51a mm, compaction: a... |
2576 |
.gfp_mask = GFP_KERNEL, |
698b1b306 mm, compaction: i... |
2577 |
}; |
698b1b306 mm, compaction: i... |
2578 |
trace_mm_compaction_kcompactd_wake(pgdat->node_id, cc.order, |
97a225e69 mm/page_alloc: in... |
2579 |
cc.highest_zoneidx); |
7f354a548 mm, compaction: a... |
2580 |
count_compact_event(KCOMPACTD_WAKE); |
698b1b306 mm, compaction: i... |
2581 |
|
97a225e69 mm/page_alloc: in... |
2582 |
for (zoneid = 0; zoneid <= cc.highest_zoneidx; zoneid++) { |
698b1b306 mm, compaction: i... |
2583 2584 2585 2586 2587 2588 2589 2590 2591 2592 2593 2594 |
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 mm: fix kcompactd... |
2595 2596 |
if (kthread_should_stop()) return; |
a94b52524 mm/compaction.c: ... |
2597 2598 |
cc.zone = zone; |
5e1f0f098 mm, compaction: c... |
2599 |
status = compact_zone(&cc, NULL); |
698b1b306 mm, compaction: i... |
2600 |
|
7ceb009a2 mm, compaction: d... |
2601 |
if (status == COMPACT_SUCCESS) { |
698b1b306 mm, compaction: i... |
2602 |
compaction_defer_reset(zone, cc.order, false); |
c8f7de0bf mm, compaction: d... |
2603 |
} else if (status == COMPACT_PARTIAL_SKIPPED || status == COMPACT_COMPLETE) { |
698b1b306 mm, compaction: i... |
2604 |
/* |
bc3106b26 mm, compaction: d... |
2605 2606 2607 2608 2609 2610 2611 2612 |
* 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 mm, compaction: i... |
2613 2614 2615 2616 2617 |
* We use sync migration mode here, so we defer like * sync direct compaction does. */ defer_compaction(zone, cc.order); } |
7f354a548 mm, compaction: a... |
2618 2619 2620 2621 |
count_compact_events(KCOMPACTD_MIGRATE_SCANNED, cc.total_migrate_scanned); count_compact_events(KCOMPACTD_FREE_SCANNED, cc.total_free_scanned); |
698b1b306 mm, compaction: i... |
2622 2623 2624 2625 2626 2627 |
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 mm/page_alloc: in... |
2628 2629 |
* the requested order/highest_zoneidx in case it was higher/tighter * than our current ones |
698b1b306 mm, compaction: i... |
2630 2631 2632 |
*/ if (pgdat->kcompactd_max_order <= cc.order) pgdat->kcompactd_max_order = 0; |
97a225e69 mm/page_alloc: in... |
2633 2634 |
if (pgdat->kcompactd_highest_zoneidx >= cc.highest_zoneidx) pgdat->kcompactd_highest_zoneidx = pgdat->nr_zones - 1; |
698b1b306 mm, compaction: i... |
2635 |
} |
97a225e69 mm/page_alloc: in... |
2636 |
void wakeup_kcompactd(pg_data_t *pgdat, int order, int highest_zoneidx) |
698b1b306 mm, compaction: i... |
2637 2638 2639 2640 2641 2642 |
{ if (!order) return; if (pgdat->kcompactd_max_order < order) pgdat->kcompactd_max_order = order; |
97a225e69 mm/page_alloc: in... |
2643 2644 |
if (pgdat->kcompactd_highest_zoneidx > highest_zoneidx) pgdat->kcompactd_highest_zoneidx = highest_zoneidx; |
698b1b306 mm, compaction: i... |
2645 |
|
6818600ff mm,compaction: se... |
2646 2647 2648 2649 2650 |
/* * Pairs with implicit barrier in wait_event_freezable() * such that wakeups are not missed. */ if (!wq_has_sleeper(&pgdat->kcompactd_wait)) |
698b1b306 mm, compaction: i... |
2651 2652 2653 2654 2655 2656 |
return; if (!kcompactd_node_suitable(pgdat)) return; trace_mm_compaction_wakeup_kcompactd(pgdat->node_id, order, |
97a225e69 mm/page_alloc: in... |
2657 |
highest_zoneidx); |
698b1b306 mm, compaction: i... |
2658 2659 2660 2661 2662 2663 2664 2665 2666 2667 2668 |
wake_up_interruptible(&pgdat->kcompactd_wait); } /* * The background compaction daemon, started as a kernel thread * from the init process. */ static int kcompactd(void *p) { pg_data_t *pgdat = (pg_data_t*)p; struct task_struct *tsk = current; |
facdaa917 mm: proactive com... |
2669 |
unsigned int proactive_defer = 0; |
698b1b306 mm, compaction: i... |
2670 2671 2672 2673 2674 2675 2676 2677 2678 |
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 mm/page_alloc: in... |
2679 |
pgdat->kcompactd_highest_zoneidx = pgdat->nr_zones - 1; |
698b1b306 mm, compaction: i... |
2680 2681 |
while (!kthread_should_stop()) { |
eb414681d psi: pressure sta... |
2682 |
unsigned long pflags; |
698b1b306 mm, compaction: i... |
2683 |
trace_mm_compaction_kcompactd_sleep(pgdat->node_id); |
facdaa917 mm: proactive com... |
2684 2685 2686 2687 2688 2689 2690 2691 2692 |
if (wait_event_freezable_timeout(pgdat->kcompactd_wait, kcompactd_work_requested(pgdat), msecs_to_jiffies(HPAGE_FRAG_CHECK_INTERVAL_MSEC))) { psi_memstall_enter(&pflags); kcompactd_do_work(pgdat); psi_memstall_leave(&pflags); continue; } |
698b1b306 mm, compaction: i... |
2693 |
|
facdaa917 mm: proactive com... |
2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704 2705 2706 2707 2708 2709 2710 2711 |
/* kcompactd wait timeout */ if (should_proactive_compact_node(pgdat)) { unsigned int prev_score, score; if (proactive_defer) { proactive_defer--; continue; } 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. */ proactive_defer = score < prev_score ? 0 : 1 << COMPACT_MAX_DEFER_SHIFT; } |
698b1b306 mm, compaction: i... |
2712 2713 2714 2715 2716 2717 2718 2719 2720 2721 2722 2723 2724 2725 2726 2727 2728 2729 2730 2731 2732 2733 2734 2735 2736 2737 2738 2739 2740 2741 2742 2743 2744 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 |
} 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 mm/compaction: Co... |
2759 |
static int kcompactd_cpu_online(unsigned int cpu) |
698b1b306 mm, compaction: i... |
2760 2761 |
{ int nid; |
e46b1db24 mm/compaction: Co... |
2762 2763 2764 |
for_each_node_state(nid, N_MEMORY) { pg_data_t *pgdat = NODE_DATA(nid); const struct cpumask *mask; |
698b1b306 mm, compaction: i... |
2765 |
|
e46b1db24 mm/compaction: Co... |
2766 |
mask = cpumask_of_node(pgdat->node_id); |
698b1b306 mm, compaction: i... |
2767 |
|
e46b1db24 mm/compaction: Co... |
2768 2769 2770 |
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 mm, compaction: i... |
2771 |
} |
e46b1db24 mm/compaction: Co... |
2772 |
return 0; |
698b1b306 mm, compaction: i... |
2773 2774 2775 2776 2777 |
} static int __init kcompactd_init(void) { int nid; |
e46b1db24 mm/compaction: Co... |
2778 2779 2780 2781 2782 2783 2784 2785 2786 2787 |
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 mm, compaction: i... |
2788 2789 2790 |
for_each_node_state(nid, N_MEMORY) kcompactd_run(nid); |
698b1b306 mm, compaction: i... |
2791 2792 2793 |
return 0; } subsys_initcall(kcompactd_init) |
ff9543fd3 mm: compaction: e... |
2794 |
#endif /* CONFIG_COMPACTION */ |