Blame view
mm/readahead.c
16.5 KB
457c89965 treewide: Add SPD... |
1 |
// SPDX-License-Identifier: GPL-2.0-only |
1da177e4c Linux-2.6.12-rc2 |
2 3 4 5 6 |
/* * mm/readahead.c - address_space-level file readahead. * * Copyright (C) 2002, Linus Torvalds * |
e1f8e8744 Remove Andrew Mor... |
7 |
* 09Apr2002 Andrew Morton |
1da177e4c Linux-2.6.12-rc2 |
8 9 10 11 |
* Initial version. */ #include <linux/kernel.h> |
11bd969fd mm: silently skip... |
12 |
#include <linux/dax.h> |
5a0e3ad6a include cleanup: ... |
13 |
#include <linux/gfp.h> |
b95f1b31b mm: Map most file... |
14 |
#include <linux/export.h> |
1da177e4c Linux-2.6.12-rc2 |
15 16 |
#include <linux/blkdev.h> #include <linux/backing-dev.h> |
8bde37f08 [PATCH] io-accoun... |
17 |
#include <linux/task_io_accounting_ops.h> |
1da177e4c Linux-2.6.12-rc2 |
18 |
#include <linux/pagevec.h> |
f5ff8422b Fix warnings with... |
19 |
#include <linux/pagemap.h> |
782182e53 mm: move readahea... |
20 21 |
#include <linux/syscalls.h> #include <linux/file.h> |
d72ee9111 mm: move lru_to_p... |
22 |
#include <linux/mm_inline.h> |
ca47e8c72 mm: skip readahea... |
23 |
#include <linux/blk-cgroup.h> |
3d8f76153 vfs: implement re... |
24 |
#include <linux/fadvise.h> |
1da177e4c Linux-2.6.12-rc2 |
25 |
|
29f175d12 mm/readahead.c: i... |
26 |
#include "internal.h" |
1da177e4c Linux-2.6.12-rc2 |
27 28 29 30 31 32 33 |
/* * Initialise a struct file's readahead state. Assumes that the caller has * memset *ra to zero. */ void file_ra_state_init(struct file_ra_state *ra, struct address_space *mapping) { |
de1414a65 fs: export inode_... |
34 |
ra->ra_pages = inode_to_bdi(mapping->host)->ra_pages; |
f4e6b498d readahead: combin... |
35 |
ra->prev_pos = -1; |
1da177e4c Linux-2.6.12-rc2 |
36 |
} |
d41cc702c [GFS2] Export fil... |
37 |
EXPORT_SYMBOL_GPL(file_ra_state_init); |
1da177e4c Linux-2.6.12-rc2 |
38 |
|
03fb3d2af FS-Cache: Release... |
39 40 |
/* * see if a page needs releasing upon read_cache_pages() failure |
266cf658e FS-Cache: Recruit... |
41 42 43 44 |
* - the caller of read_cache_pages() may have set PG_private or PG_fscache * before calling, such as the NFS fs marking pages that are cached locally * on disk, thus we need to give the fs a chance to clean up in the event of * an error |
03fb3d2af FS-Cache: Release... |
45 46 47 48 |
*/ static void read_cache_pages_invalidate_page(struct address_space *mapping, struct page *page) { |
266cf658e FS-Cache: Recruit... |
49 |
if (page_has_private(page)) { |
03fb3d2af FS-Cache: Release... |
50 51 52 |
if (!trylock_page(page)) BUG(); page->mapping = mapping; |
09cbfeaf1 mm, fs: get rid o... |
53 |
do_invalidatepage(page, 0, PAGE_SIZE); |
03fb3d2af FS-Cache: Release... |
54 55 56 |
page->mapping = NULL; unlock_page(page); } |
09cbfeaf1 mm, fs: get rid o... |
57 |
put_page(page); |
03fb3d2af FS-Cache: Release... |
58 59 60 61 62 63 64 65 66 67 68 |
} /* * release a list of pages, invalidating them first if need be */ static void read_cache_pages_invalidate_pages(struct address_space *mapping, struct list_head *pages) { struct page *victim; while (!list_empty(pages)) { |
c8ad6302c mm/readahead.c, m... |
69 |
victim = lru_to_page(pages); |
03fb3d2af FS-Cache: Release... |
70 71 72 73 |
list_del(&victim->lru); read_cache_pages_invalidate_page(mapping, victim); } } |
1da177e4c Linux-2.6.12-rc2 |
74 |
/** |
bd40cddae [PATCH] kernel-do... |
75 |
* read_cache_pages - populate an address space with some pages & start reads against them |
1da177e4c Linux-2.6.12-rc2 |
76 77 78 79 80 81 82 |
* @mapping: the address_space * @pages: The address of a list_head which contains the target pages. These * pages have their ->index populated and are otherwise uninitialised. * @filler: callback routine for filling a single page. * @data: private data for the callback routine. * * Hides the details of the LRU cache etc from the filesystems. |
a862f68a8 docs/core-api/mm:... |
83 84 |
* * Returns: %0 on success, error return by @filler otherwise |
1da177e4c Linux-2.6.12-rc2 |
85 86 87 88 89 |
*/ int read_cache_pages(struct address_space *mapping, struct list_head *pages, int (*filler)(void *, struct page *), void *data) { struct page *page; |
1da177e4c Linux-2.6.12-rc2 |
90 |
int ret = 0; |
1da177e4c Linux-2.6.12-rc2 |
91 |
while (!list_empty(pages)) { |
c8ad6302c mm/readahead.c, m... |
92 |
page = lru_to_page(pages); |
1da177e4c Linux-2.6.12-rc2 |
93 |
list_del(&page->lru); |
063d99b4f mm, fs: obey gfp_... |
94 |
if (add_to_page_cache_lru(page, mapping, page->index, |
8a5c743e3 mm, memcg: use co... |
95 |
readahead_gfp_mask(mapping))) { |
03fb3d2af FS-Cache: Release... |
96 |
read_cache_pages_invalidate_page(mapping, page); |
1da177e4c Linux-2.6.12-rc2 |
97 98 |
continue; } |
09cbfeaf1 mm, fs: get rid o... |
99 |
put_page(page); |
eb2be1893 mm: buffered writ... |
100 |
|
1da177e4c Linux-2.6.12-rc2 |
101 |
ret = filler(data, page); |
eb2be1893 mm: buffered writ... |
102 |
if (unlikely(ret)) { |
03fb3d2af FS-Cache: Release... |
103 |
read_cache_pages_invalidate_pages(mapping, pages); |
1da177e4c Linux-2.6.12-rc2 |
104 105 |
break; } |
09cbfeaf1 mm, fs: get rid o... |
106 |
task_io_account_read(PAGE_SIZE); |
1da177e4c Linux-2.6.12-rc2 |
107 |
} |
1da177e4c Linux-2.6.12-rc2 |
108 109 110 111 112 113 |
return ret; } EXPORT_SYMBOL(read_cache_pages); static int read_pages(struct address_space *mapping, struct file *filp, |
8a5c743e3 mm, memcg: use co... |
114 |
struct list_head *pages, unsigned int nr_pages, gfp_t gfp) |
1da177e4c Linux-2.6.12-rc2 |
115 |
{ |
5b417b187 read-ahead: use p... |
116 |
struct blk_plug plug; |
1da177e4c Linux-2.6.12-rc2 |
117 |
unsigned page_idx; |
994fc28c7 [PATCH] add AOP_T... |
118 |
int ret; |
1da177e4c Linux-2.6.12-rc2 |
119 |
|
5b417b187 read-ahead: use p... |
120 |
blk_start_plug(&plug); |
1da177e4c Linux-2.6.12-rc2 |
121 122 |
if (mapping->a_ops->readpages) { ret = mapping->a_ops->readpages(filp, mapping, pages, nr_pages); |
029e332ea [PATCH] Cleanup r... |
123 124 |
/* Clean up the remaining pages */ put_pages_list(pages); |
1da177e4c Linux-2.6.12-rc2 |
125 126 |
goto out; } |
1da177e4c Linux-2.6.12-rc2 |
127 |
for (page_idx = 0; page_idx < nr_pages; page_idx++) { |
c8ad6302c mm/readahead.c, m... |
128 |
struct page *page = lru_to_page(pages); |
1da177e4c Linux-2.6.12-rc2 |
129 |
list_del(&page->lru); |
8a5c743e3 mm, memcg: use co... |
130 |
if (!add_to_page_cache_lru(page, mapping, page->index, gfp)) |
9f1a3cfcf [PATCH] AOP_TRUNC... |
131 |
mapping->a_ops->readpage(filp, page); |
09cbfeaf1 mm, fs: get rid o... |
132 |
put_page(page); |
1da177e4c Linux-2.6.12-rc2 |
133 |
} |
994fc28c7 [PATCH] add AOP_T... |
134 |
ret = 0; |
5b417b187 read-ahead: use p... |
135 |
|
1da177e4c Linux-2.6.12-rc2 |
136 |
out: |
5b417b187 read-ahead: use p... |
137 |
blk_finish_plug(&plug); |
1da177e4c Linux-2.6.12-rc2 |
138 139 140 141 |
return ret; } /* |
b3751e6ab mm: split ->readp... |
142 143 |
* __do_page_cache_readahead() actually reads a chunk of disk. It allocates * the pages first, then submits them for I/O. This avoids the very bad |
1da177e4c Linux-2.6.12-rc2 |
144 145 146 147 |
* behaviour which would occur if page allocations are causing VM writeback. * We really don't want to intermingle reads and writes like that. * * Returns the number of pages requested, or the maximum amount of I/O allowed. |
1da177e4c Linux-2.6.12-rc2 |
148 |
*/ |
c534aa3fd mm: return an uns... |
149 150 151 |
unsigned int __do_page_cache_readahead(struct address_space *mapping, struct file *filp, pgoff_t offset, unsigned long nr_to_read, unsigned long lookahead_size) |
1da177e4c Linux-2.6.12-rc2 |
152 153 154 155 156 157 |
{ struct inode *inode = mapping->host; struct page *page; unsigned long end_index; /* The last page we want to read */ LIST_HEAD(page_pool); int page_idx; |
c534aa3fd mm: return an uns... |
158 |
unsigned int nr_pages = 0; |
1da177e4c Linux-2.6.12-rc2 |
159 |
loff_t isize = i_size_read(inode); |
8a5c743e3 mm, memcg: use co... |
160 |
gfp_t gfp_mask = readahead_gfp_mask(mapping); |
1da177e4c Linux-2.6.12-rc2 |
161 162 163 |
if (isize == 0) goto out; |
09cbfeaf1 mm, fs: get rid o... |
164 |
end_index = ((isize - 1) >> PAGE_SHIFT); |
1da177e4c Linux-2.6.12-rc2 |
165 166 167 168 |
/* * Preallocate as many pages as we will need. */ |
1da177e4c Linux-2.6.12-rc2 |
169 |
for (page_idx = 0; page_idx < nr_to_read; page_idx++) { |
7361f4d8c [PATCH] readahead... |
170 |
pgoff_t page_offset = offset + page_idx; |
c743d96b6 readahead: remove... |
171 |
|
1da177e4c Linux-2.6.12-rc2 |
172 173 |
if (page_offset > end_index) break; |
560d454ba mm: Convert __do_... |
174 |
page = xa_load(&mapping->i_pages, page_offset); |
3159f943a xarray: Replace e... |
175 |
if (page && !xa_is_value(page)) { |
b3751e6ab mm: split ->readp... |
176 177 178 179 180 181 182 183 184 |
/* * Page already present? Kick off the current batch of * contiguous pages before continuing with the next * batch. */ if (nr_pages) read_pages(mapping, filp, &page_pool, nr_pages, gfp_mask); nr_pages = 0; |
1da177e4c Linux-2.6.12-rc2 |
185 |
continue; |
b3751e6ab mm: split ->readp... |
186 |
} |
1da177e4c Linux-2.6.12-rc2 |
187 |
|
8a5c743e3 mm, memcg: use co... |
188 |
page = __page_cache_alloc(gfp_mask); |
1da177e4c Linux-2.6.12-rc2 |
189 190 191 192 |
if (!page) break; page->index = page_offset; list_add(&page->lru, &page_pool); |
46fc3e7b4 readahead: add lo... |
193 194 |
if (page_idx == nr_to_read - lookahead_size) SetPageReadahead(page); |
836978b35 mm: give the 'ret... |
195 |
nr_pages++; |
1da177e4c Linux-2.6.12-rc2 |
196 |
} |
1da177e4c Linux-2.6.12-rc2 |
197 198 199 200 201 202 |
/* * Now start the IO. We ignore I/O errors - if the page is not * uptodate then the caller will launch readpage again, and * will then handle the error. */ |
836978b35 mm: give the 'ret... |
203 204 |
if (nr_pages) read_pages(mapping, filp, &page_pool, nr_pages, gfp_mask); |
1da177e4c Linux-2.6.12-rc2 |
205 206 |
BUG_ON(!list_empty(&page_pool)); out: |
836978b35 mm: give the 'ret... |
207 |
return nr_pages; |
1da177e4c Linux-2.6.12-rc2 |
208 209 210 211 212 213 214 |
} /* * Chunk the readahead into 2 megabyte units, so that we don't pin too much * memory at once. */ int force_page_cache_readahead(struct address_space *mapping, struct file *filp, |
9491ae4aa mm: don't cap req... |
215 |
pgoff_t offset, unsigned long nr_to_read) |
1da177e4c Linux-2.6.12-rc2 |
216 |
{ |
9491ae4aa mm: don't cap req... |
217 218 219 |
struct backing_dev_info *bdi = inode_to_bdi(mapping->host); struct file_ra_state *ra = &filp->f_ra; unsigned long max_pages; |
1da177e4c Linux-2.6.12-rc2 |
220 221 |
if (unlikely(!mapping->a_ops->readpage && !mapping->a_ops->readpages)) return -EINVAL; |
9491ae4aa mm: don't cap req... |
222 223 224 225 226 227 |
/* * If the request exceeds the readahead window, allow the read to * be up to the optimal hardware IO size */ max_pages = max_t(unsigned long, bdi->io_pages, ra->ra_pages); nr_to_read = min(nr_to_read, max_pages); |
1da177e4c Linux-2.6.12-rc2 |
228 |
while (nr_to_read) { |
09cbfeaf1 mm, fs: get rid o... |
229 |
unsigned long this_chunk = (2 * 1024 * 1024) / PAGE_SIZE; |
1da177e4c Linux-2.6.12-rc2 |
230 231 232 |
if (this_chunk > nr_to_read) this_chunk = nr_to_read; |
c534aa3fd mm: return an uns... |
233 |
__do_page_cache_readahead(mapping, filp, offset, this_chunk, 0); |
58d5640eb mm/readahead.c: f... |
234 |
|
1da177e4c Linux-2.6.12-rc2 |
235 236 237 |
offset += this_chunk; nr_to_read -= this_chunk; } |
58d5640eb mm/readahead.c: f... |
238 |
return 0; |
1da177e4c Linux-2.6.12-rc2 |
239 |
} |
5ce1110b9 readahead: data s... |
240 |
/* |
c743d96b6 readahead: remove... |
241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 |
* Set the initial window size, round to next power of 2 and square * for small size, x 4 for medium, and x 2 for large * for 128k (32 page) max ra * 1-8 page = 32k initial, > 8 page = 128k initial */ static unsigned long get_init_ra_size(unsigned long size, unsigned long max) { unsigned long newsize = roundup_pow_of_two(size); if (newsize <= max / 32) newsize = newsize * 4; else if (newsize <= max / 4) newsize = newsize * 2; else newsize = max; return newsize; } /* |
122a21d11 readahead: on-dem... |
261 262 263 |
* Get the previous window size, ramp it up, and * return it as the new window size. */ |
c743d96b6 readahead: remove... |
264 |
static unsigned long get_next_ra_size(struct file_ra_state *ra, |
20ff1c950 mm/readahead.c: s... |
265 |
unsigned long max) |
122a21d11 readahead: on-dem... |
266 |
{ |
f9acc8c7b readahead: sanify... |
267 |
unsigned long cur = ra->size; |
122a21d11 readahead: on-dem... |
268 269 |
if (cur < max / 16) |
20ff1c950 mm/readahead.c: s... |
270 271 272 273 |
return 4 * cur; if (cur <= max / 2) return 2 * cur; return max; |
122a21d11 readahead: on-dem... |
274 275 276 277 278 279 280 281 |
} /* * On-demand readahead design. * * The fields in struct file_ra_state represent the most-recently-executed * readahead attempt: * |
f9acc8c7b readahead: sanify... |
282 283 284 285 |
* |<----- async_size ---------| * |------------------- size -------------------->| * |==================#===========================| * ^start ^page marked with PG_readahead |
122a21d11 readahead: on-dem... |
286 287 288 289 |
* * To overlap application thinking time and disk I/O time, we do * `readahead pipelining': Do not wait until the application consumed all * readahead pages and stalled on the missing page at readahead_index; |
f9acc8c7b readahead: sanify... |
290 291 292 |
* Instead, submit an asynchronous readahead I/O as soon as there are * only async_size pages left in the readahead window. Normally async_size * will be equal to size, for maximum pipelining. |
122a21d11 readahead: on-dem... |
293 294 295 |
* * In interleaved sequential reads, concurrent streams on the same fd can * be invalidating each other's readahead state. So we flag the new readahead |
f9acc8c7b readahead: sanify... |
296 |
* page at (start+size-async_size) with PG_readahead, and use it as readahead |
122a21d11 readahead: on-dem... |
297 298 299 |
* indicator. The flag won't be set on already cached pages, to avoid the * readahead-for-nothing fuss, saving pointless page cache lookups. * |
f4e6b498d readahead: combin... |
300 |
* prev_pos tracks the last visited byte in the _previous_ read request. |
122a21d11 readahead: on-dem... |
301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 |
* It should be maintained by the caller, and will be used for detecting * small random reads. Note that the readahead algorithm checks loosely * for sequential patterns. Hence interleaved reads might be served as * sequential ones. * * There is a special-case: if the first page which the application tries to * read happens to be the first page of the file, it is assumed that a linear * read is about to happen and the window is immediately set to the initial size * based on I/O request size and the max_readahead. * * The code ramps up the readahead size aggressively at first, but slow down as * it approaches max_readhead. */ /* |
10be0b372 readahead: introd... |
316 317 318 319 320 321 |
* Count contiguously cached pages from @offset-1 to @offset-@max, * this count is a conservative estimation of * - length of the sequential read sequence, or * - thrashing threshold in memory tight systems */ static pgoff_t count_history_pages(struct address_space *mapping, |
10be0b372 readahead: introd... |
322 323 324 325 326 |
pgoff_t offset, unsigned long max) { pgoff_t head; rcu_read_lock(); |
0d3f92966 page cache: Conve... |
327 |
head = page_cache_prev_miss(mapping, offset - 1, max); |
10be0b372 readahead: introd... |
328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 |
rcu_read_unlock(); return offset - 1 - head; } /* * page cache context based read-ahead */ static int try_context_readahead(struct address_space *mapping, struct file_ra_state *ra, pgoff_t offset, unsigned long req_size, unsigned long max) { pgoff_t size; |
3e2faa085 mm/readahead.c: r... |
343 |
size = count_history_pages(mapping, offset, max); |
10be0b372 readahead: introd... |
344 345 |
/* |
2cad40180 readahead: make c... |
346 |
* not enough history pages: |
10be0b372 readahead: introd... |
347 348 |
* it could be a random read */ |
2cad40180 readahead: make c... |
349 |
if (size <= req_size) |
10be0b372 readahead: introd... |
350 351 352 353 354 355 356 357 358 359 |
return 0; /* * starts from beginning of file: * it is a strong indication of long-run stream (or whole-file-read) */ if (size >= offset) size *= 2; ra->start = offset; |
2cad40180 readahead: make c... |
360 361 |
ra->size = min(size + req_size, max); ra->async_size = 1; |
10be0b372 readahead: introd... |
362 363 364 365 366 |
return 1; } /* |
122a21d11 readahead: on-dem... |
367 368 369 370 371 |
* A minimal readahead algorithm for trivial sequential/random reads. */ static unsigned long ondemand_readahead(struct address_space *mapping, struct file_ra_state *ra, struct file *filp, |
cf914a7d6 readahead: split ... |
372 |
bool hit_readahead_marker, pgoff_t offset, |
122a21d11 readahead: on-dem... |
373 374 |
unsigned long req_size) { |
9491ae4aa mm: don't cap req... |
375 376 |
struct backing_dev_info *bdi = inode_to_bdi(mapping->host); unsigned long max_pages = ra->ra_pages; |
dc30b96ab readahead: strict... |
377 |
unsigned long add_pages; |
af248a0c6 readahead: fix se... |
378 |
pgoff_t prev_offset; |
045a2529a readahead: move t... |
379 380 |
/* |
9491ae4aa mm: don't cap req... |
381 382 383 384 385 386 387 |
* If the request exceeds the readahead window, allow the read to * be up to the optimal hardware IO size */ if (req_size > max_pages && bdi->io_pages > max_pages) max_pages = min(req_size, bdi->io_pages); /* |
045a2529a readahead: move t... |
388 389 390 391 |
* start of file */ if (!offset) goto initial_readahead; |
122a21d11 readahead: on-dem... |
392 393 |
/* |
f9acc8c7b readahead: sanify... |
394 |
* It's the expected callback offset, assume sequential access. |
122a21d11 readahead: on-dem... |
395 396 |
* Ramp up sizes, and push forward the readahead window. */ |
045a2529a readahead: move t... |
397 398 |
if ((offset == (ra->start + ra->size - ra->async_size) || offset == (ra->start + ra->size))) { |
f9acc8c7b readahead: sanify... |
399 |
ra->start += ra->size; |
9491ae4aa mm: don't cap req... |
400 |
ra->size = get_next_ra_size(ra, max_pages); |
f9acc8c7b readahead: sanify... |
401 402 |
ra->async_size = ra->size; goto readit; |
122a21d11 readahead: on-dem... |
403 |
} |
122a21d11 readahead: on-dem... |
404 |
/* |
6b10c6c9f readahead: basic ... |
405 406 407 408 409 410 411 |
* Hit a marked page without valid readahead state. * E.g. interleaved reads. * Query the pagecache for async_size, which normally equals to * readahead size. Ramp it up and use it as the new readahead size. */ if (hit_readahead_marker) { pgoff_t start; |
30002ed2e mm: readahead sca... |
412 |
rcu_read_lock(); |
0d3f92966 page cache: Conve... |
413 |
start = page_cache_next_miss(mapping, offset + 1, max_pages); |
30002ed2e mm: readahead sca... |
414 |
rcu_read_unlock(); |
6b10c6c9f readahead: basic ... |
415 |
|
9491ae4aa mm: don't cap req... |
416 |
if (!start || start - offset > max_pages) |
6b10c6c9f readahead: basic ... |
417 418 419 420 |
return 0; ra->start = start; ra->size = start - offset; /* old async_size */ |
160334a0c readahead: increa... |
421 |
ra->size += req_size; |
9491ae4aa mm: don't cap req... |
422 |
ra->size = get_next_ra_size(ra, max_pages); |
6b10c6c9f readahead: basic ... |
423 424 425 426 427 |
ra->async_size = ra->size; goto readit; } /* |
045a2529a readahead: move t... |
428 |
* oversize read |
122a21d11 readahead: on-dem... |
429 |
*/ |
9491ae4aa mm: don't cap req... |
430 |
if (req_size > max_pages) |
045a2529a readahead: move t... |
431 432 433 434 |
goto initial_readahead; /* * sequential cache miss |
af248a0c6 readahead: fix se... |
435 436 |
* trivial case: (offset - prev_offset) == 1 * unaligned reads: (offset - prev_offset) == 0 |
045a2529a readahead: move t... |
437 |
*/ |
09cbfeaf1 mm, fs: get rid o... |
438 |
prev_offset = (unsigned long long)ra->prev_pos >> PAGE_SHIFT; |
af248a0c6 readahead: fix se... |
439 |
if (offset - prev_offset <= 1UL) |
045a2529a readahead: move t... |
440 441 442 |
goto initial_readahead; /* |
10be0b372 readahead: introd... |
443 444 445 |
* Query the page cache and look for the traces(cached history pages) * that a sequential stream would leave behind. */ |
9491ae4aa mm: don't cap req... |
446 |
if (try_context_readahead(mapping, ra, offset, req_size, max_pages)) |
10be0b372 readahead: introd... |
447 448 449 |
goto readit; /* |
045a2529a readahead: move t... |
450 451 452 453 454 455 |
* standalone, small random read * Read as is, and do not pollute the readahead state. */ return __do_page_cache_readahead(mapping, filp, offset, req_size, 0); initial_readahead: |
f9acc8c7b readahead: sanify... |
456 |
ra->start = offset; |
9491ae4aa mm: don't cap req... |
457 |
ra->size = get_init_ra_size(req_size, max_pages); |
f9acc8c7b readahead: sanify... |
458 |
ra->async_size = ra->size > req_size ? ra->size - req_size : ra->size; |
122a21d11 readahead: on-dem... |
459 |
|
f9acc8c7b readahead: sanify... |
460 |
readit: |
51daa88eb readahead: remove... |
461 462 463 464 |
/* * Will this read hit the readahead marker made by itself? * If so, trigger the readahead marker hit now, and merge * the resulted next readahead window into the current one. |
dc30b96ab readahead: strict... |
465 |
* Take care of maximum IO pages as above. |
51daa88eb readahead: remove... |
466 467 |
*/ if (offset == ra->start && ra->size == ra->async_size) { |
dc30b96ab readahead: strict... |
468 469 470 471 472 473 474 475 |
add_pages = get_next_ra_size(ra, max_pages); if (ra->size + add_pages <= max_pages) { ra->async_size = add_pages; ra->size += add_pages; } else { ra->size = max_pages; ra->async_size = max_pages >> 1; } |
51daa88eb readahead: remove... |
476 |
} |
122a21d11 readahead: on-dem... |
477 478 479 480 |
return ra_submit(ra, mapping, filp); } /** |
cf914a7d6 readahead: split ... |
481 |
* page_cache_sync_readahead - generic file readahead |
122a21d11 readahead: on-dem... |
482 483 484 |
* @mapping: address_space which holds the pagecache and I/O vectors * @ra: file_ra_state which holds the readahead state * @filp: passed on to ->readpage() and ->readpages() |
cf914a7d6 readahead: split ... |
485 |
* @offset: start offset into @mapping, in pagecache page-sized units |
122a21d11 readahead: on-dem... |
486 |
* @req_size: hint: total size of the read which the caller is performing in |
cf914a7d6 readahead: split ... |
487 |
* pagecache pages |
122a21d11 readahead: on-dem... |
488 |
* |
cf914a7d6 readahead: split ... |
489 490 491 492 |
* page_cache_sync_readahead() should be called when a cache miss happened: * it will submit the read. The readahead logic may decide to piggyback more * pages onto the read request if access patterns suggest it will improve * performance. |
122a21d11 readahead: on-dem... |
493 |
*/ |
cf914a7d6 readahead: split ... |
494 495 496 |
void page_cache_sync_readahead(struct address_space *mapping, struct file_ra_state *ra, struct file *filp, pgoff_t offset, unsigned long req_size) |
122a21d11 readahead: on-dem... |
497 498 499 |
{ /* no read-ahead */ if (!ra->ra_pages) |
cf914a7d6 readahead: split ... |
500 |
return; |
ca47e8c72 mm: skip readahea... |
501 502 |
if (blk_cgroup_congested()) return; |
0141450f6 readahead: introd... |
503 |
/* be dumb */ |
70655c06b readahead: fix NU... |
504 |
if (filp && (filp->f_mode & FMODE_RANDOM)) { |
0141450f6 readahead: introd... |
505 506 507 |
force_page_cache_readahead(mapping, filp, offset, req_size); return; } |
cf914a7d6 readahead: split ... |
508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 |
/* do read-ahead */ ondemand_readahead(mapping, ra, filp, false, offset, req_size); } EXPORT_SYMBOL_GPL(page_cache_sync_readahead); /** * page_cache_async_readahead - file readahead for marked pages * @mapping: address_space which holds the pagecache and I/O vectors * @ra: file_ra_state which holds the readahead state * @filp: passed on to ->readpage() and ->readpages() * @page: the page at @offset which has the PG_readahead flag set * @offset: start offset into @mapping, in pagecache page-sized units * @req_size: hint: total size of the read which the caller is performing in * pagecache pages * |
bf8abe8b9 readahead.c: fix ... |
523 |
* page_cache_async_readahead() should be called when a page is used which |
f7850d932 mm/readahead: fix... |
524 |
* has the PG_readahead flag; this is a marker to suggest that the application |
cf914a7d6 readahead: split ... |
525 |
* has used up enough of the readahead window that we should start pulling in |
f7850d932 mm/readahead: fix... |
526 527 |
* more pages. */ |
cf914a7d6 readahead: split ... |
528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 |
void page_cache_async_readahead(struct address_space *mapping, struct file_ra_state *ra, struct file *filp, struct page *page, pgoff_t offset, unsigned long req_size) { /* no read-ahead */ if (!ra->ra_pages) return; /* * Same bit is used for PG_readahead and PG_reclaim. */ if (PageWriteback(page)) return; ClearPageReadahead(page); /* * Defer asynchronous read-ahead on IO congestion. */ |
703c27088 writeback: implem... |
549 |
if (inode_read_congested(mapping->host)) |
cf914a7d6 readahead: split ... |
550 |
return; |
122a21d11 readahead: on-dem... |
551 |
|
ca47e8c72 mm: skip readahea... |
552 553 |
if (blk_cgroup_congested()) return; |
122a21d11 readahead: on-dem... |
554 |
/* do read-ahead */ |
cf914a7d6 readahead: split ... |
555 |
ondemand_readahead(mapping, ra, filp, true, offset, req_size); |
122a21d11 readahead: on-dem... |
556 |
} |
cf914a7d6 readahead: split ... |
557 |
EXPORT_SYMBOL_GPL(page_cache_async_readahead); |
782182e53 mm: move readahea... |
558 |
|
c7b95d515 mm: add ksys_read... |
559 |
ssize_t ksys_readahead(int fd, loff_t offset, size_t count) |
782182e53 mm: move readahea... |
560 561 |
{ ssize_t ret; |
2903ff019 switch simple cas... |
562 |
struct fd f; |
782182e53 mm: move readahea... |
563 564 |
ret = -EBADF; |
2903ff019 switch simple cas... |
565 |
f = fdget(fd); |
3d8f76153 vfs: implement re... |
566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 |
if (!f.file || !(f.file->f_mode & FMODE_READ)) goto out; /* * The readahead() syscall is intended to run only on files * that can execute readahead. If readahead is not possible * on this file, then we must return -EINVAL. */ ret = -EINVAL; if (!f.file->f_mapping || !f.file->f_mapping->a_ops || !S_ISREG(file_inode(f.file)->i_mode)) goto out; ret = vfs_fadvise(f.file, offset, count, POSIX_FADV_WILLNEED); out: fdput(f); |
782182e53 mm: move readahea... |
582 583 |
return ret; } |
c7b95d515 mm: add ksys_read... |
584 585 586 587 588 |
SYSCALL_DEFINE3(readahead, int, fd, loff_t, offset, size_t, count) { return ksys_readahead(fd, offset, count); } |