Blame view
mm/swap_state.c
22.4 KB
b24413180
|
1 |
// SPDX-License-Identifier: GPL-2.0 |
1da177e4c
|
2 3 4 5 6 7 8 9 |
/* * linux/mm/swap_state.c * * Copyright (C) 1991, 1992, 1993, 1994 Linus Torvalds * Swap reorganised 29.12.95, Stephen Tweedie * * Rewritten to use page cache, (C) 1998 Stephen Tweedie */ |
1da177e4c
|
10 |
#include <linux/mm.h> |
5a0e3ad6a
|
11 |
#include <linux/gfp.h> |
1da177e4c
|
12 13 |
#include <linux/kernel_stat.h> #include <linux/swap.h> |
46017e954
|
14 |
#include <linux/swapops.h> |
1da177e4c
|
15 16 |
#include <linux/init.h> #include <linux/pagemap.h> |
1da177e4c
|
17 |
#include <linux/backing-dev.h> |
3fb5c298b
|
18 |
#include <linux/blkdev.h> |
c484d4104
|
19 |
#include <linux/pagevec.h> |
b20a35035
|
20 |
#include <linux/migrate.h> |
4b3ef9daa
|
21 |
#include <linux/vmalloc.h> |
67afa38e0
|
22 |
#include <linux/swap_slots.h> |
38d8b4e6b
|
23 |
#include <linux/huge_mm.h> |
1da177e4c
|
24 25 26 27 28 |
#include <asm/pgtable.h> /* * swapper_space is a fiction, retained to simplify the path through |
7eaceacca
|
29 |
* vmscan's shrink_page_list. |
1da177e4c
|
30 |
*/ |
f5e54d6e5
|
31 |
static const struct address_space_operations swap_aops = { |
1da177e4c
|
32 |
.writepage = swap_writepage, |
62c230bc1
|
33 |
.set_page_dirty = swap_set_page_dirty, |
1c93923cc
|
34 |
#ifdef CONFIG_MIGRATION |
e965f9630
|
35 |
.migratepage = migrate_page, |
1c93923cc
|
36 |
#endif |
1da177e4c
|
37 |
}; |
783cb68ee
|
38 39 |
struct address_space *swapper_spaces[MAX_SWAPFILES] __read_mostly; static unsigned int nr_swapper_spaces[MAX_SWAPFILES] __read_mostly; |
f5c754d63
|
40 |
static bool enable_vma_readahead __read_mostly = true; |
ec560175c
|
41 |
|
ec560175c
|
42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 |
#define SWAP_RA_WIN_SHIFT (PAGE_SHIFT / 2) #define SWAP_RA_HITS_MASK ((1UL << SWAP_RA_WIN_SHIFT) - 1) #define SWAP_RA_HITS_MAX SWAP_RA_HITS_MASK #define SWAP_RA_WIN_MASK (~PAGE_MASK & ~SWAP_RA_HITS_MASK) #define SWAP_RA_HITS(v) ((v) & SWAP_RA_HITS_MASK) #define SWAP_RA_WIN(v) (((v) & SWAP_RA_WIN_MASK) >> SWAP_RA_WIN_SHIFT) #define SWAP_RA_ADDR(v) ((v) & PAGE_MASK) #define SWAP_RA_VAL(addr, win, hits) \ (((addr) & PAGE_MASK) | \ (((win) << SWAP_RA_WIN_SHIFT) & SWAP_RA_WIN_MASK) | \ ((hits) & SWAP_RA_HITS_MASK)) /* Initial readahead hits is 4 to start up with a small window */ #define GET_SWAP_RA_VAL(vma) \ (atomic_long_read(&(vma)->swap_readahead_info) ? : 4) |
1da177e4c
|
59 60 |
#define INC_CACHE_INFO(x) do { swap_cache_info.x++; } while (0) |
38d8b4e6b
|
61 |
#define ADD_CACHE_INFO(x, nr) do { swap_cache_info.x += (nr); } while (0) |
1da177e4c
|
62 63 64 65 66 67 |
static struct { unsigned long add_total; unsigned long del_total; unsigned long find_success; unsigned long find_total; |
1da177e4c
|
68 |
} swap_cache_info; |
33806f06d
|
69 70 |
unsigned long total_swapcache_pages(void) { |
4b3ef9daa
|
71 |
unsigned int i, j, nr; |
33806f06d
|
72 |
unsigned long ret = 0; |
4b3ef9daa
|
73 |
struct address_space *spaces; |
054f1d1fa
|
74 |
struct swap_info_struct *si; |
33806f06d
|
75 |
|
4b3ef9daa
|
76 |
for (i = 0; i < MAX_SWAPFILES; i++) { |
054f1d1fa
|
77 78 79 80 81 82 83 84 |
swp_entry_t entry = swp_entry(i, 1); /* Avoid get_swap_device() to warn for bad swap entry */ if (!swp_swap_info(entry)) continue; /* Prevent swapoff to free swapper_spaces */ si = get_swap_device(entry); if (!si) |
4b3ef9daa
|
85 |
continue; |
054f1d1fa
|
86 87 |
nr = nr_swapper_spaces[i]; spaces = swapper_spaces[i]; |
4b3ef9daa
|
88 89 |
for (j = 0; j < nr; j++) ret += spaces[j].nrpages; |
054f1d1fa
|
90 |
put_swap_device(si); |
4b3ef9daa
|
91 |
} |
33806f06d
|
92 93 |
return ret; } |
579f82901
|
94 |
static atomic_t swapin_readahead_hits = ATOMIC_INIT(4); |
1da177e4c
|
95 96 |
void show_swap_cache_info(void) { |
33806f06d
|
97 98 |
printk("%lu pages in swap cache ", total_swapcache_pages()); |
2c97b7fc0
|
99 100 |
printk("Swap cache stats: add %lu, delete %lu, find %lu/%lu ", |
1da177e4c
|
101 |
swap_cache_info.add_total, swap_cache_info.del_total, |
bb63be0a0
|
102 |
swap_cache_info.find_success, swap_cache_info.find_total); |
ec8acf20a
|
103 104 105 |
printk("Free swap = %ldkB ", get_nr_swap_pages() << (PAGE_SHIFT - 10)); |
1da177e4c
|
106 107 108 109 110 |
printk("Total swap = %lukB ", total_swap_pages << (PAGE_SHIFT - 10)); } /* |
8d93b41c0
|
111 |
* add_to_swap_cache resembles add_to_page_cache_locked on swapper_space, |
1da177e4c
|
112 113 |
* but sets SwapCache flag and private instead of mapping and index. */ |
8d93b41c0
|
114 |
int add_to_swap_cache(struct page *page, swp_entry_t entry, gfp_t gfp) |
1da177e4c
|
115 |
{ |
8d93b41c0
|
116 |
struct address_space *address_space = swap_address_space(entry); |
38d8b4e6b
|
117 |
pgoff_t idx = swp_offset(entry); |
8d93b41c0
|
118 119 |
XA_STATE_ORDER(xas, &address_space->i_pages, idx, compound_order(page)); unsigned long i, nr = 1UL << compound_order(page); |
1da177e4c
|
120 |
|
309381fea
|
121 122 123 |
VM_BUG_ON_PAGE(!PageLocked(page), page); VM_BUG_ON_PAGE(PageSwapCache(page), page); VM_BUG_ON_PAGE(!PageSwapBacked(page), page); |
51726b122
|
124 |
|
38d8b4e6b
|
125 |
page_ref_add(page, nr); |
31a563962
|
126 |
SetPageSwapCache(page); |
31a563962
|
127 |
|
8d93b41c0
|
128 129 130 131 132 133 134 135 |
do { xas_lock_irq(&xas); xas_create_range(&xas); if (xas_error(&xas)) goto unlock; for (i = 0; i < nr; i++) { VM_BUG_ON_PAGE(xas.xa_index != idx + i, page); set_page_private(page + i, entry.val + i); |
69bf4b6b5
|
136 |
xas_store(&xas, page + i); |
8d93b41c0
|
137 138 |
xas_next(&xas); } |
38d8b4e6b
|
139 140 141 |
address_space->nrpages += nr; __mod_node_page_state(page_pgdat(page), NR_FILE_PAGES, nr); ADD_CACHE_INFO(add_total, nr); |
8d93b41c0
|
142 143 144 |
unlock: xas_unlock_irq(&xas); } while (xas_nomem(&xas, gfp)); |
31a563962
|
145 |
|
8d93b41c0
|
146 147 |
if (!xas_error(&xas)) return 0; |
31a563962
|
148 |
|
8d93b41c0
|
149 150 151 |
ClearPageSwapCache(page); page_ref_sub(page, nr); return xas_error(&xas); |
1da177e4c
|
152 |
} |
1da177e4c
|
153 154 155 156 |
/* * This must be called only on pages that have * been verified to be in the swap cache. */ |
4e17ec250
|
157 |
void __delete_from_swap_cache(struct page *page, swp_entry_t entry) |
1da177e4c
|
158 |
{ |
4e17ec250
|
159 |
struct address_space *address_space = swap_address_space(entry); |
38d8b4e6b
|
160 |
int i, nr = hpage_nr_pages(page); |
4e17ec250
|
161 162 |
pgoff_t idx = swp_offset(entry); XA_STATE(xas, &address_space->i_pages, idx); |
33806f06d
|
163 |
|
309381fea
|
164 165 166 |
VM_BUG_ON_PAGE(!PageLocked(page), page); VM_BUG_ON_PAGE(!PageSwapCache(page), page); VM_BUG_ON_PAGE(PageWriteback(page), page); |
1da177e4c
|
167 |
|
38d8b4e6b
|
168 |
for (i = 0; i < nr; i++) { |
4e17ec250
|
169 |
void *entry = xas_store(&xas, NULL); |
69bf4b6b5
|
170 |
VM_BUG_ON_PAGE(entry != page + i, entry); |
38d8b4e6b
|
171 |
set_page_private(page + i, 0); |
4e17ec250
|
172 |
xas_next(&xas); |
38d8b4e6b
|
173 |
} |
1da177e4c
|
174 |
ClearPageSwapCache(page); |
38d8b4e6b
|
175 176 177 |
address_space->nrpages -= nr; __mod_node_page_state(page_pgdat(page), NR_FILE_PAGES, -nr); ADD_CACHE_INFO(del_total, nr); |
1da177e4c
|
178 179 180 181 182 183 184 185 186 |
} /** * add_to_swap - allocate swap space for a page * @page: page we want to move to swap * * Allocate swap space for the page and add the page to the * swap cache. Caller needs to hold the page lock. */ |
0f0746589
|
187 |
int add_to_swap(struct page *page) |
1da177e4c
|
188 189 |
{ swp_entry_t entry; |
1da177e4c
|
190 |
int err; |
309381fea
|
191 192 |
VM_BUG_ON_PAGE(!PageLocked(page), page); VM_BUG_ON_PAGE(!PageUptodate(page), page); |
1da177e4c
|
193 |
|
38d8b4e6b
|
194 |
entry = get_swap_page(page); |
2ca4532a4
|
195 |
if (!entry.val) |
0f0746589
|
196 |
return 0; |
2ca4532a4
|
197 |
/* |
8d93b41c0
|
198 |
* XArray node allocations from PF_MEMALLOC contexts could |
2ca4532a4
|
199 200 201 202 203 204 205 |
* completely exhaust the page allocator. __GFP_NOMEMALLOC * stops emergency reserves from being allocated. * * TODO: this could cause a theoretical memory reclaim * deadlock in the swap out path. */ /* |
854e9ed09
|
206 |
* Add it to the swap cache. |
2ca4532a4
|
207 208 209 |
*/ err = add_to_swap_cache(page, entry, __GFP_HIGH|__GFP_NOMEMALLOC|__GFP_NOWARN); |
38d8b4e6b
|
210 |
if (err) |
bd53b714d
|
211 |
/* |
2ca4532a4
|
212 213 |
* add_to_swap_cache() doesn't return -EEXIST, so we can safely * clear SWAP_HAS_CACHE flag. |
1da177e4c
|
214 |
*/ |
0f0746589
|
215 |
goto fail; |
9625456cc
|
216 217 218 219 220 221 222 223 224 225 226 |
/* * Normally the page will be dirtied in unmap because its pte should be * dirty. A special case is MADV_FREE page. The page'e pte could have * dirty bit cleared but the page's SwapBacked bit is still set because * clearing the dirty bit and SwapBacked bit has no lock protected. For * such page, unmap will not set dirty bit for it, so page reclaim will * not write the page out. This can cause data corruption when the page * is swap in later. Always setting the dirty bit for the page solves * the problem. */ set_page_dirty(page); |
38d8b4e6b
|
227 228 |
return 1; |
38d8b4e6b
|
229 |
fail: |
0f0746589
|
230 |
put_swap_page(page, entry); |
38d8b4e6b
|
231 |
return 0; |
1da177e4c
|
232 233 234 235 236 237 238 239 240 241 |
} /* * This must be called only on pages that have * been verified to be in the swap cache and locked. * It will never put the page into the free list, * the caller has a reference on the page. */ void delete_from_swap_cache(struct page *page) { |
4e17ec250
|
242 243 |
swp_entry_t entry = { .val = page_private(page) }; struct address_space *address_space = swap_address_space(entry); |
1da177e4c
|
244 |
|
b93b01631
|
245 |
xa_lock_irq(&address_space->i_pages); |
4e17ec250
|
246 |
__delete_from_swap_cache(page, entry); |
b93b01631
|
247 |
xa_unlock_irq(&address_space->i_pages); |
1da177e4c
|
248 |
|
75f6d6d29
|
249 |
put_swap_page(page, entry); |
38d8b4e6b
|
250 |
page_ref_sub(page, hpage_nr_pages(page)); |
1da177e4c
|
251 |
} |
1da177e4c
|
252 253 254 255 |
/* * If we are the only user, then try to free up the swap cache. * * Its ok to check for PageSwapCache without the page lock |
a2c43eed8
|
256 257 |
* here because we are going to recheck again inside * try_to_free_swap() _with_ the lock. |
1da177e4c
|
258 259 260 261 |
* - Marcelo */ static inline void free_swap_cache(struct page *page) { |
a2c43eed8
|
262 263 |
if (PageSwapCache(page) && !page_mapped(page) && trylock_page(page)) { try_to_free_swap(page); |
1da177e4c
|
264 265 266 267 268 269 |
unlock_page(page); } } /* * Perform a free_page(), also freeing any swap cache associated with |
b8072f099
|
270 |
* this page if it is the last user of the page. |
1da177e4c
|
271 272 273 274 |
*/ void free_page_and_swap_cache(struct page *page) { free_swap_cache(page); |
6fcb52a56
|
275 |
if (!is_huge_zero_page(page)) |
770a53702
|
276 |
put_page(page); |
1da177e4c
|
277 278 279 280 281 282 283 284 |
} /* * Passed an array of pages, drop them all from swapcache and then release * them. They are removed from the LRU and freed if this is their last use. */ void free_pages_and_swap_cache(struct page **pages, int nr) { |
1da177e4c
|
285 |
struct page **pagep = pages; |
aabfb5729
|
286 |
int i; |
1da177e4c
|
287 288 |
lru_add_drain(); |
aabfb5729
|
289 290 |
for (i = 0; i < nr; i++) free_swap_cache(pagep[i]); |
c6f92f9fb
|
291 |
release_pages(pagep, nr); |
1da177e4c
|
292 |
} |
e9e9b7ece
|
293 294 295 296 |
static inline bool swap_use_vma_readahead(void) { return READ_ONCE(enable_vma_readahead) && !atomic_read(&nr_rotate_swap); } |
1da177e4c
|
297 298 299 300 301 302 |
/* * Lookup a swap entry in the swap cache. A found page will be returned * unlocked and with its refcount incremented - we rely on the kernel * lock getting page table operations atomic even if we drop the page * lock before returning. */ |
ec560175c
|
303 304 |
struct page *lookup_swap_cache(swp_entry_t entry, struct vm_area_struct *vma, unsigned long addr) |
1da177e4c
|
305 306 |
{ struct page *page; |
eb085574a
|
307 |
struct swap_info_struct *si; |
1da177e4c
|
308 |
|
eb085574a
|
309 310 311 |
si = get_swap_device(entry); if (!si) return NULL; |
f6ab1f7f6
|
312 |
page = find_get_page(swap_address_space(entry), swp_offset(entry)); |
eb085574a
|
313 |
put_swap_device(si); |
1da177e4c
|
314 |
|
ec560175c
|
315 316 |
INC_CACHE_INFO(find_total); if (page) { |
eaf649ebc
|
317 318 |
bool vma_ra = swap_use_vma_readahead(); bool readahead; |
1da177e4c
|
319 |
INC_CACHE_INFO(find_success); |
eaf649ebc
|
320 321 322 323 |
/* * At the moment, we don't support PG_readahead for anon THP * so let's bail out rather than confusing the readahead stat. */ |
ec560175c
|
324 325 |
if (unlikely(PageTransCompound(page))) return page; |
eaf649ebc
|
326 |
|
ec560175c
|
327 |
readahead = TestClearPageReadahead(page); |
eaf649ebc
|
328 329 330 331 332 333 334 |
if (vma && vma_ra) { unsigned long ra_val; int win, hits; ra_val = GET_SWAP_RA_VAL(vma); win = SWAP_RA_WIN(ra_val); hits = SWAP_RA_HITS(ra_val); |
ec560175c
|
335 336 337 338 339 |
if (readahead) hits = min_t(int, hits + 1, SWAP_RA_HITS_MAX); atomic_long_set(&vma->swap_readahead_info, SWAP_RA_VAL(addr, win, hits)); } |
eaf649ebc
|
340 |
|
ec560175c
|
341 |
if (readahead) { |
cbc65df24
|
342 |
count_vm_event(SWAP_RA_HIT); |
eaf649ebc
|
343 |
if (!vma || !vma_ra) |
ec560175c
|
344 |
atomic_inc(&swapin_readahead_hits); |
cbc65df24
|
345 |
} |
579f82901
|
346 |
} |
eaf649ebc
|
347 |
|
1da177e4c
|
348 349 |
return page; } |
5b999aadb
|
350 351 352 |
struct page *__read_swap_cache_async(swp_entry_t entry, gfp_t gfp_mask, struct vm_area_struct *vma, unsigned long addr, bool *new_page_allocated) |
1da177e4c
|
353 |
{ |
eb085574a
|
354 355 |
struct page *found_page = NULL, *new_page = NULL; struct swap_info_struct *si; |
1da177e4c
|
356 |
int err; |
5b999aadb
|
357 |
*new_page_allocated = false; |
1da177e4c
|
358 359 360 361 362 363 364 |
do { /* * First check the swap cache. Since this is normally * called after lookup_swap_cache() failed, re-calling * that would confuse statistics. */ |
eb085574a
|
365 366 367 368 369 370 |
si = get_swap_device(entry); if (!si) break; found_page = find_get_page(swap_address_space(entry), swp_offset(entry)); put_swap_device(si); |
1da177e4c
|
371 372 |
if (found_page) break; |
ba81f8384
|
373 374 375 376 377 378 379 380 381 382 |
/* * Just skip read ahead for unused swap slot. * During swap_off when swap_slot_cache is disabled, * we have to handle the race between putting * swap entry in swap cache and marking swap slot * as SWAP_HAS_CACHE. That's done in later part of code or * else swap_off will be aborted if we return NULL. */ if (!__swp_swapcount(entry) && swap_slot_cache_enabled) break; |
e8c26ab60
|
383 |
|
1da177e4c
|
384 385 386 387 |
/* * Get a new page to read into from swap. */ if (!new_page) { |
02098feaa
|
388 |
new_page = alloc_page_vma(gfp_mask, vma, addr); |
1da177e4c
|
389 390 391 392 393 |
if (!new_page) break; /* Out of memory */ } /* |
f000944d0
|
394 395 |
* Swap entry may have been freed since our caller observed it. */ |
355cfa73d
|
396 |
err = swapcache_prepare(entry); |
cbab0e4ee
|
397 |
if (err == -EEXIST) { |
cbab0e4ee
|
398 399 400 |
/* * We might race against get_swap_page() and stumble * across a SWAP_HAS_CACHE swap_map entry whose page |
9c1cc2e4f
|
401 |
* has not been brought into the swapcache yet. |
cbab0e4ee
|
402 403 |
*/ cond_resched(); |
355cfa73d
|
404 |
continue; |
8d93b41c0
|
405 |
} else if (err) /* swp entry is obsolete ? */ |
f000944d0
|
406 |
break; |
8d93b41c0
|
407 |
/* May fail (-ENOMEM) if XArray node allocation failed. */ |
48c935ad8
|
408 |
__SetPageLocked(new_page); |
fa9949da5
|
409 |
__SetPageSwapBacked(new_page); |
8d93b41c0
|
410 |
err = add_to_swap_cache(new_page, entry, gfp_mask & GFP_KERNEL); |
529ae9aaa
|
411 |
if (likely(!err)) { |
8d93b41c0
|
412 |
/* Initiate read into locked page */ |
1899ad18c
|
413 |
SetPageWorkingset(new_page); |
c5fdae469
|
414 |
lru_cache_add_anon(new_page); |
5b999aadb
|
415 |
*new_page_allocated = true; |
1da177e4c
|
416 417 |
return new_page; } |
48c935ad8
|
418 |
__ClearPageLocked(new_page); |
2ca4532a4
|
419 420 421 422 |
/* * add_to_swap_cache() doesn't return -EEXIST, so we can safely * clear SWAP_HAS_CACHE flag. */ |
75f6d6d29
|
423 |
put_swap_page(new_page, entry); |
f000944d0
|
424 |
} while (err != -ENOMEM); |
1da177e4c
|
425 426 |
if (new_page) |
09cbfeaf1
|
427 |
put_page(new_page); |
1da177e4c
|
428 429 |
return found_page; } |
46017e954
|
430 |
|
5b999aadb
|
431 432 433 434 435 436 437 |
/* * Locate a page of swap in physical memory, reserving swap cache space * and reading the disk if it is not already cached. * A failure return means that either the page allocation failed or that * the swap entry is no longer in use. */ struct page *read_swap_cache_async(swp_entry_t entry, gfp_t gfp_mask, |
23955622f
|
438 |
struct vm_area_struct *vma, unsigned long addr, bool do_poll) |
5b999aadb
|
439 440 441 442 443 444 |
{ bool page_was_allocated; struct page *retpage = __read_swap_cache_async(entry, gfp_mask, vma, addr, &page_was_allocated); if (page_was_allocated) |
23955622f
|
445 |
swap_readpage(retpage, do_poll); |
5b999aadb
|
446 447 448 |
return retpage; } |
ec560175c
|
449 450 451 452 453 |
static unsigned int __swapin_nr_pages(unsigned long prev_offset, unsigned long offset, int hits, int max_pages, int prev_win) |
579f82901
|
454 |
{ |
ec560175c
|
455 |
unsigned int pages, last_ra; |
579f82901
|
456 457 458 459 460 461 |
/* * This heuristic has been found to work well on both sequential and * random loads, swapping to hard disk or to SSD: please don't ask * what the "+ 2" means, it just happens to work well, that's all. */ |
ec560175c
|
462 |
pages = hits + 2; |
579f82901
|
463 464 465 466 467 468 469 470 |
if (pages == 2) { /* * We can have no readahead hits to judge by: but must not get * stuck here forever, so check for an adjacent offset instead * (and don't even bother to check whether swap type is same). */ if (offset != prev_offset + 1 && offset != prev_offset - 1) pages = 1; |
579f82901
|
471 472 473 474 475 476 477 478 479 480 481 |
} else { unsigned int roundup = 4; while (roundup < pages) roundup <<= 1; pages = roundup; } if (pages > max_pages) pages = max_pages; /* Don't shrink readahead too fast */ |
ec560175c
|
482 |
last_ra = prev_win / 2; |
579f82901
|
483 484 |
if (pages < last_ra) pages = last_ra; |
ec560175c
|
485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 |
return pages; } static unsigned long swapin_nr_pages(unsigned long offset) { static unsigned long prev_offset; unsigned int hits, pages, max_pages; static atomic_t last_readahead_pages; max_pages = 1 << READ_ONCE(page_cluster); if (max_pages <= 1) return 1; hits = atomic_xchg(&swapin_readahead_hits, 0); pages = __swapin_nr_pages(prev_offset, offset, hits, max_pages, atomic_read(&last_readahead_pages)); if (!hits) prev_offset = offset; |
579f82901
|
504 505 506 507 |
atomic_set(&last_readahead_pages, pages); return pages; } |
46017e954
|
508 |
/** |
e9e9b7ece
|
509 |
* swap_cluster_readahead - swap in pages in hope we need them soon |
46017e954
|
510 |
* @entry: swap entry of this memory |
7682486b3
|
511 |
* @gfp_mask: memory allocation flags |
e9e9b7ece
|
512 |
* @vmf: fault information |
46017e954
|
513 514 515 516 517 518 519 520 521 522 523 |
* * Returns the struct page for entry and addr, after queueing swapin. * * Primitive swap readahead code. We simply read an aligned block of * (1 << page_cluster) entries in the swap area. This method is chosen * because it doesn't cost us any seek time. We also make sure to queue * the 'original' request together with the readahead ones... * * This has been extended to use the NUMA policies from the mm triggering * the readahead. * |
e9f598730
|
524 |
* Caller must hold read mmap_sem if vmf->vma is not NULL. |
46017e954
|
525 |
*/ |
e9e9b7ece
|
526 527 |
struct page *swap_cluster_readahead(swp_entry_t entry, gfp_t gfp_mask, struct vm_fault *vmf) |
46017e954
|
528 |
{ |
46017e954
|
529 |
struct page *page; |
579f82901
|
530 531 |
unsigned long entry_offset = swp_offset(entry); unsigned long offset = entry_offset; |
67f96aa25
|
532 |
unsigned long start_offset, end_offset; |
579f82901
|
533 |
unsigned long mask; |
e9a6effa5
|
534 |
struct swap_info_struct *si = swp_swap_info(entry); |
3fb5c298b
|
535 |
struct blk_plug plug; |
c4fa63092
|
536 |
bool do_poll = true, page_allocated; |
e9e9b7ece
|
537 538 |
struct vm_area_struct *vma = vmf->vma; unsigned long addr = vmf->address; |
46017e954
|
539 |
|
579f82901
|
540 541 542 |
mask = swapin_nr_pages(offset) - 1; if (!mask) goto skip; |
8fd2e0b50
|
543 544 545 546 547 548 |
/* Test swap type to make sure the dereference is safe */ if (likely(si->flags & (SWP_BLKDEV | SWP_FS))) { struct inode *inode = si->swap_file->f_mapping->host; if (inode_read_congested(inode)) goto skip; } |
23955622f
|
549 |
do_poll = false; |
67f96aa25
|
550 551 552 553 554 |
/* Read a page_cluster sized and aligned cluster around offset. */ start_offset = offset & ~mask; end_offset = offset | mask; if (!start_offset) /* First page is swap header. */ start_offset++; |
e9a6effa5
|
555 556 |
if (end_offset >= si->max) end_offset = si->max - 1; |
67f96aa25
|
557 |
|
3fb5c298b
|
558 |
blk_start_plug(&plug); |
67f96aa25
|
559 |
for (offset = start_offset; offset <= end_offset ; offset++) { |
46017e954
|
560 |
/* Ok, do the async read-ahead now */ |
c4fa63092
|
561 562 563 |
page = __read_swap_cache_async( swp_entry(swp_type(entry), offset), gfp_mask, vma, addr, &page_allocated); |
46017e954
|
564 |
if (!page) |
67f96aa25
|
565 |
continue; |
c4fa63092
|
566 567 |
if (page_allocated) { swap_readpage(page, false); |
eaf649ebc
|
568 |
if (offset != entry_offset) { |
c4fa63092
|
569 570 571 |
SetPageReadahead(page); count_vm_event(SWAP_RA); } |
cbc65df24
|
572 |
} |
09cbfeaf1
|
573 |
put_page(page); |
46017e954
|
574 |
} |
3fb5c298b
|
575 |
blk_finish_plug(&plug); |
46017e954
|
576 |
lru_add_drain(); /* Push any new pages onto the LRU now */ |
579f82901
|
577 |
skip: |
23955622f
|
578 |
return read_swap_cache_async(entry, gfp_mask, vma, addr, do_poll); |
46017e954
|
579 |
} |
4b3ef9daa
|
580 581 582 583 584 585 586 |
int init_swap_address_space(unsigned int type, unsigned long nr_pages) { struct address_space *spaces, *space; unsigned int i, nr; nr = DIV_ROUND_UP(nr_pages, SWAP_ADDRESS_SPACE_PAGES); |
778e1cdd8
|
587 |
spaces = kvcalloc(nr, sizeof(struct address_space), GFP_KERNEL); |
4b3ef9daa
|
588 589 590 591 |
if (!spaces) return -ENOMEM; for (i = 0; i < nr; i++) { space = spaces + i; |
a28334862
|
592 |
xa_init_flags(&space->i_pages, XA_FLAGS_LOCK_IRQ); |
4b3ef9daa
|
593 594 595 596 |
atomic_set(&space->i_mmap_writable, 0); space->a_ops = &swap_aops; /* swap cache doesn't use writeback related tags */ mapping_set_no_writeback_tags(space); |
4b3ef9daa
|
597 598 |
} nr_swapper_spaces[type] = nr; |
054f1d1fa
|
599 |
swapper_spaces[type] = spaces; |
4b3ef9daa
|
600 601 602 603 604 605 |
return 0; } void exit_swap_address_space(unsigned int type) { |
054f1d1fa
|
606 |
kvfree(swapper_spaces[type]); |
4b3ef9daa
|
607 |
nr_swapper_spaces[type] = 0; |
054f1d1fa
|
608 |
swapper_spaces[type] = NULL; |
4b3ef9daa
|
609 |
} |
ec560175c
|
610 611 612 613 614 615 616 617 618 619 620 621 622 |
static inline void swap_ra_clamp_pfn(struct vm_area_struct *vma, unsigned long faddr, unsigned long lpfn, unsigned long rpfn, unsigned long *start, unsigned long *end) { *start = max3(lpfn, PFN_DOWN(vma->vm_start), PFN_DOWN(faddr & PMD_MASK)); *end = min3(rpfn, PFN_DOWN(vma->vm_end), PFN_DOWN((faddr & PMD_MASK) + PMD_SIZE)); } |
eaf649ebc
|
623 624 |
static void swap_ra_info(struct vm_fault *vmf, struct vma_swap_readahead *ra_info) |
ec560175c
|
625 626 |
{ struct vm_area_struct *vma = vmf->vma; |
eaf649ebc
|
627 |
unsigned long ra_val; |
ec560175c
|
628 629 630 |
swp_entry_t entry; unsigned long faddr, pfn, fpfn; unsigned long start, end; |
eaf649ebc
|
631 |
pte_t *pte, *orig_pte; |
ec560175c
|
632 633 634 635 |
unsigned int max_win, hits, prev_win, win, left; #ifndef CONFIG_64BIT pte_t *tpte; #endif |
61b639723
|
636 637 638 |
max_win = 1 << min_t(unsigned int, READ_ONCE(page_cluster), SWAP_RA_ORDER_CEILING); if (max_win == 1) { |
eaf649ebc
|
639 640 |
ra_info->win = 1; return; |
61b639723
|
641 |
} |
ec560175c
|
642 |
faddr = vmf->address; |
eaf649ebc
|
643 644 645 646 647 648 |
orig_pte = pte = pte_offset_map(vmf->pmd, faddr); entry = pte_to_swp_entry(*pte); if ((unlikely(non_swap_entry(entry)))) { pte_unmap(orig_pte); return; } |
ec560175c
|
649 |
|
ec560175c
|
650 |
fpfn = PFN_DOWN(faddr); |
eaf649ebc
|
651 652 653 654 655 |
ra_val = GET_SWAP_RA_VAL(vma); pfn = PFN_DOWN(SWAP_RA_ADDR(ra_val)); prev_win = SWAP_RA_WIN(ra_val); hits = SWAP_RA_HITS(ra_val); ra_info->win = win = __swapin_nr_pages(pfn, fpfn, hits, |
ec560175c
|
656 657 658 |
max_win, prev_win); atomic_long_set(&vma->swap_readahead_info, SWAP_RA_VAL(faddr, win, 0)); |
eaf649ebc
|
659 660 661 662 |
if (win == 1) { pte_unmap(orig_pte); return; } |
ec560175c
|
663 664 665 666 667 668 669 670 671 672 673 674 |
/* Copy the PTEs because the page table may be unmapped */ if (fpfn == pfn + 1) swap_ra_clamp_pfn(vma, faddr, fpfn, fpfn + win, &start, &end); else if (pfn == fpfn + 1) swap_ra_clamp_pfn(vma, faddr, fpfn - win + 1, fpfn + 1, &start, &end); else { left = (win - 1) / 2; swap_ra_clamp_pfn(vma, faddr, fpfn - left, fpfn + win - left, &start, &end); } |
eaf649ebc
|
675 676 677 |
ra_info->nr_pte = end - start; ra_info->offset = fpfn - start; pte -= ra_info->offset; |
ec560175c
|
678 |
#ifdef CONFIG_64BIT |
eaf649ebc
|
679 |
ra_info->ptes = pte; |
ec560175c
|
680 |
#else |
eaf649ebc
|
681 |
tpte = ra_info->ptes; |
ec560175c
|
682 683 684 |
for (pfn = start; pfn != end; pfn++) *tpte++ = *pte++; #endif |
eaf649ebc
|
685 |
pte_unmap(orig_pte); |
ec560175c
|
686 |
} |
e9f598730
|
687 688 689 690 691 692 693 694 695 696 697 698 699 700 |
/** * swap_vma_readahead - swap in pages in hope we need them soon * @entry: swap entry of this memory * @gfp_mask: memory allocation flags * @vmf: fault information * * Returns the struct page for entry and addr, after queueing swapin. * * Primitive swap readahead code. We simply read in a few pages whoes * virtual addresses are around the fault address in the same vma. * * Caller must hold read mmap_sem if vmf->vma is not NULL. * */ |
f5c754d63
|
701 702 |
static struct page *swap_vma_readahead(swp_entry_t fentry, gfp_t gfp_mask, struct vm_fault *vmf) |
ec560175c
|
703 704 705 706 707 708 709 710 |
{ struct blk_plug plug; struct vm_area_struct *vma = vmf->vma; struct page *page; pte_t *pte, pentry; swp_entry_t entry; unsigned int i; bool page_allocated; |
eaf649ebc
|
711 |
struct vma_swap_readahead ra_info = {0,}; |
ec560175c
|
712 |
|
eaf649ebc
|
713 714 |
swap_ra_info(vmf, &ra_info); if (ra_info.win == 1) |
ec560175c
|
715 716 717 |
goto skip; blk_start_plug(&plug); |
eaf649ebc
|
718 |
for (i = 0, pte = ra_info.ptes; i < ra_info.nr_pte; |
ec560175c
|
719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 |
i++, pte++) { pentry = *pte; if (pte_none(pentry)) continue; if (pte_present(pentry)) continue; entry = pte_to_swp_entry(pentry); if (unlikely(non_swap_entry(entry))) continue; page = __read_swap_cache_async(entry, gfp_mask, vma, vmf->address, &page_allocated); if (!page) continue; if (page_allocated) { swap_readpage(page, false); |
eaf649ebc
|
734 |
if (i != ra_info.offset) { |
ec560175c
|
735 736 737 738 739 740 741 742 743 744 |
SetPageReadahead(page); count_vm_event(SWAP_RA); } } put_page(page); } blk_finish_plug(&plug); lru_add_drain(); skip: return read_swap_cache_async(fentry, gfp_mask, vma, vmf->address, |
eaf649ebc
|
745 |
ra_info.win == 1); |
ec560175c
|
746 |
} |
d9bfcfdc4
|
747 |
|
e9e9b7ece
|
748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 |
/** * swapin_readahead - swap in pages in hope we need them soon * @entry: swap entry of this memory * @gfp_mask: memory allocation flags * @vmf: fault information * * Returns the struct page for entry and addr, after queueing swapin. * * It's a main entry function for swap readahead. By the configuration, * it will read ahead blocks by cluster-based(ie, physical disk based) * or vma-based(ie, virtual address based on faulty address) readahead. */ struct page *swapin_readahead(swp_entry_t entry, gfp_t gfp_mask, struct vm_fault *vmf) { return swap_use_vma_readahead() ? swap_vma_readahead(entry, gfp_mask, vmf) : swap_cluster_readahead(entry, gfp_mask, vmf); } |
d9bfcfdc4
|
767 768 769 770 |
#ifdef CONFIG_SYSFS static ssize_t vma_ra_enabled_show(struct kobject *kobj, struct kobj_attribute *attr, char *buf) { |
e9e9b7ece
|
771 772 |
return sprintf(buf, "%s ", enable_vma_readahead ? "true" : "false"); |
d9bfcfdc4
|
773 774 775 776 777 778 |
} static ssize_t vma_ra_enabled_store(struct kobject *kobj, struct kobj_attribute *attr, const char *buf, size_t count) { if (!strncmp(buf, "true", 4) || !strncmp(buf, "1", 1)) |
e9e9b7ece
|
779 |
enable_vma_readahead = true; |
d9bfcfdc4
|
780 |
else if (!strncmp(buf, "false", 5) || !strncmp(buf, "0", 1)) |
e9e9b7ece
|
781 |
enable_vma_readahead = false; |
d9bfcfdc4
|
782 783 784 785 786 787 788 789 |
else return -EINVAL; return count; } static struct kobj_attribute vma_ra_enabled_attr = __ATTR(vma_ra_enabled, 0644, vma_ra_enabled_show, vma_ra_enabled_store); |
d9bfcfdc4
|
790 791 |
static struct attribute *swap_attrs[] = { &vma_ra_enabled_attr.attr, |
d9bfcfdc4
|
792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 |
NULL, }; static struct attribute_group swap_attr_group = { .attrs = swap_attrs, }; static int __init swap_init_sysfs(void) { int err; struct kobject *swap_kobj; swap_kobj = kobject_create_and_add("swap", mm_kobj); if (!swap_kobj) { pr_err("failed to create swap kobject "); return -ENOMEM; } err = sysfs_create_group(swap_kobj, &swap_attr_group); if (err) { pr_err("failed to register swap group "); goto delete_obj; } return 0; delete_obj: kobject_put(swap_kobj); return err; } subsys_initcall(swap_init_sysfs); #endif |