Blame view
mm/page-writeback.c
39.1 KB
1da177e4c Linux-2.6.12-rc2 |
1 |
/* |
f30c22695 fix file specific... |
2 |
* mm/page-writeback.c |
1da177e4c Linux-2.6.12-rc2 |
3 4 |
* * Copyright (C) 2002, Linus Torvalds. |
04fbfdc14 mm: per device di... |
5 |
* Copyright (C) 2007 Red Hat, Inc., Peter Zijlstra <pzijlstr@redhat.com> |
1da177e4c Linux-2.6.12-rc2 |
6 7 8 9 |
* * Contains functions related to writing back dirty pages at the * address_space level. * |
e1f8e8744 Remove Andrew Mor... |
10 |
* 10Apr2002 Andrew Morton |
1da177e4c Linux-2.6.12-rc2 |
11 12 13 14 15 16 17 18 19 20 21 22 23 24 |
* Initial version */ #include <linux/kernel.h> #include <linux/module.h> #include <linux/spinlock.h> #include <linux/fs.h> #include <linux/mm.h> #include <linux/swap.h> #include <linux/slab.h> #include <linux/pagemap.h> #include <linux/writeback.h> #include <linux/init.h> #include <linux/backing-dev.h> |
55e829af0 [PATCH] io-accoun... |
25 |
#include <linux/task_io_accounting_ops.h> |
1da177e4c Linux-2.6.12-rc2 |
26 27 |
#include <linux/blkdev.h> #include <linux/mpage.h> |
d08b3851d [PATCH] mm: track... |
28 |
#include <linux/rmap.h> |
1da177e4c Linux-2.6.12-rc2 |
29 30 31 32 33 34 |
#include <linux/percpu.h> #include <linux/notifier.h> #include <linux/smp.h> #include <linux/sysctl.h> #include <linux/cpu.h> #include <linux/syscalls.h> |
cf9a2ae8d [PATCH] BLOCK: Mo... |
35 |
#include <linux/buffer_head.h> |
811d736f9 [PATCH] BLOCK: Di... |
36 |
#include <linux/pagevec.h> |
028c2dd18 writeback: Add tr... |
37 |
#include <trace/events/writeback.h> |
1da177e4c Linux-2.6.12-rc2 |
38 39 |
/* |
1da177e4c Linux-2.6.12-rc2 |
40 41 42 43 |
* After a CPU has dirtied this many pages, balance_dirty_pages_ratelimited * will look to see if it needs to force writeback or throttling. */ static long ratelimit_pages = 32; |
1da177e4c Linux-2.6.12-rc2 |
44 45 46 |
/* * When balance_dirty_pages decides that the caller needs to perform some * non-background writeback, this is how many pages it will attempt to write. |
3a2e9a5a2 writeback: balanc... |
47 |
* It should be somewhat larger than dirtied pages to ensure that reasonably |
1da177e4c Linux-2.6.12-rc2 |
48 49 |
* large amounts of I/O are submitted. */ |
3a2e9a5a2 writeback: balanc... |
50 |
static inline long sync_writeback_pages(unsigned long dirtied) |
1da177e4c Linux-2.6.12-rc2 |
51 |
{ |
3a2e9a5a2 writeback: balanc... |
52 53 54 55 |
if (dirtied < ratelimit_pages) dirtied = ratelimit_pages; return dirtied + dirtied / 2; |
1da177e4c Linux-2.6.12-rc2 |
56 57 58 59 60 |
} /* The following parameters are exported via /proc/sys/vm */ /* |
5b0830cb9 writeback: get ri... |
61 |
* Start background writeback (via writeback threads) at this percentage |
1da177e4c Linux-2.6.12-rc2 |
62 |
*/ |
1b5e62b42 writeback: double... |
63 |
int dirty_background_ratio = 10; |
1da177e4c Linux-2.6.12-rc2 |
64 65 |
/* |
2da02997e mm: add dirty_bac... |
66 67 68 69 70 71 |
* dirty_background_bytes starts at 0 (disabled) so that it is a function of * dirty_background_ratio * the amount of dirtyable memory */ unsigned long dirty_background_bytes; /* |
195cf453d mm/page-writeback... |
72 73 74 75 76 77 |
* free highmem will not be subtracted from the total free memory * for calculating free ratios if vm_highmem_is_dirtyable is true */ int vm_highmem_is_dirtyable; /* |
1da177e4c Linux-2.6.12-rc2 |
78 79 |
* The generator of dirty data starts writeback at this percentage */ |
1b5e62b42 writeback: double... |
80 |
int vm_dirty_ratio = 20; |
1da177e4c Linux-2.6.12-rc2 |
81 82 |
/* |
2da02997e mm: add dirty_bac... |
83 84 85 86 87 88 |
* vm_dirty_bytes starts at 0 (disabled) so that it is a function of * vm_dirty_ratio * the amount of dirtyable memory */ unsigned long vm_dirty_bytes; /* |
704503d83 mm: fix proc_doin... |
89 |
* The interval between `kupdate'-style writebacks |
1da177e4c Linux-2.6.12-rc2 |
90 |
*/ |
22ef37eed page-writeback: f... |
91 |
unsigned int dirty_writeback_interval = 5 * 100; /* centiseconds */ |
1da177e4c Linux-2.6.12-rc2 |
92 93 |
/* |
704503d83 mm: fix proc_doin... |
94 |
* The longest time for which data is allowed to remain dirty |
1da177e4c Linux-2.6.12-rc2 |
95 |
*/ |
22ef37eed page-writeback: f... |
96 |
unsigned int dirty_expire_interval = 30 * 100; /* centiseconds */ |
1da177e4c Linux-2.6.12-rc2 |
97 98 99 100 101 102 103 |
/* * Flag that makes the machine dump writes/reads and block dirtyings. */ int block_dump; /* |
ed5b43f15 [PATCH] Represent... |
104 105 |
* Flag that puts the machine in "laptop mode". Doubles as a timeout in jiffies: * a full sync is triggered after this time elapses without any disk activity. |
1da177e4c Linux-2.6.12-rc2 |
106 107 108 109 110 111 |
*/ int laptop_mode; EXPORT_SYMBOL(laptop_mode); /* End of sysctl-exported parameters */ |
1da177e4c Linux-2.6.12-rc2 |
112 |
/* |
04fbfdc14 mm: per device di... |
113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 |
* Scale the writeback cache size proportional to the relative writeout speeds. * * We do this by keeping a floating proportion between BDIs, based on page * writeback completions [end_page_writeback()]. Those devices that write out * pages fastest will get the larger share, while the slower will get a smaller * share. * * We use page writeout completions because we are interested in getting rid of * dirty pages. Having them written out is the primary goal. * * We introduce a concept of time, a period over which we measure these events, * because demand can/will vary over time. The length of this period itself is * measured in page writeback completions. * */ static struct prop_descriptor vm_completions; |
3e26c149c mm: dirty balanci... |
129 |
static struct prop_descriptor vm_dirties; |
04fbfdc14 mm: per device di... |
130 |
|
04fbfdc14 mm: per device di... |
131 132 133 134 135 136 137 138 |
/* * couple the period to the dirty_ratio: * * period/2 ~ roundup_pow_of_two(dirty limit) */ static int calc_period_shift(void) { unsigned long dirty_total; |
2da02997e mm: add dirty_bac... |
139 140 141 142 143 |
if (vm_dirty_bytes) dirty_total = vm_dirty_bytes / PAGE_SIZE; else dirty_total = (vm_dirty_ratio * determine_dirtyable_memory()) / 100; |
04fbfdc14 mm: per device di... |
144 145 146 147 |
return 2 + ilog2(dirty_total - 1); } /* |
2da02997e mm: add dirty_bac... |
148 |
* update the period when the dirty threshold changes. |
04fbfdc14 mm: per device di... |
149 |
*/ |
2da02997e mm: add dirty_bac... |
150 151 152 153 154 155 156 157 |
static void update_completion_period(void) { int shift = calc_period_shift(); prop_change_shift(&vm_completions, shift); prop_change_shift(&vm_dirties, shift); } int dirty_background_ratio_handler(struct ctl_table *table, int write, |
8d65af789 sysctl: remove "s... |
158 |
void __user *buffer, size_t *lenp, |
2da02997e mm: add dirty_bac... |
159 160 161 |
loff_t *ppos) { int ret; |
8d65af789 sysctl: remove "s... |
162 |
ret = proc_dointvec_minmax(table, write, buffer, lenp, ppos); |
2da02997e mm: add dirty_bac... |
163 164 165 166 167 168 |
if (ret == 0 && write) dirty_background_bytes = 0; return ret; } int dirty_background_bytes_handler(struct ctl_table *table, int write, |
8d65af789 sysctl: remove "s... |
169 |
void __user *buffer, size_t *lenp, |
2da02997e mm: add dirty_bac... |
170 171 172 |
loff_t *ppos) { int ret; |
8d65af789 sysctl: remove "s... |
173 |
ret = proc_doulongvec_minmax(table, write, buffer, lenp, ppos); |
2da02997e mm: add dirty_bac... |
174 175 176 177 |
if (ret == 0 && write) dirty_background_ratio = 0; return ret; } |
04fbfdc14 mm: per device di... |
178 |
int dirty_ratio_handler(struct ctl_table *table, int write, |
8d65af789 sysctl: remove "s... |
179 |
void __user *buffer, size_t *lenp, |
04fbfdc14 mm: per device di... |
180 181 182 |
loff_t *ppos) { int old_ratio = vm_dirty_ratio; |
2da02997e mm: add dirty_bac... |
183 |
int ret; |
8d65af789 sysctl: remove "s... |
184 |
ret = proc_dointvec_minmax(table, write, buffer, lenp, ppos); |
04fbfdc14 mm: per device di... |
185 |
if (ret == 0 && write && vm_dirty_ratio != old_ratio) { |
2da02997e mm: add dirty_bac... |
186 187 188 189 190 191 192 193 |
update_completion_period(); vm_dirty_bytes = 0; } return ret; } int dirty_bytes_handler(struct ctl_table *table, int write, |
8d65af789 sysctl: remove "s... |
194 |
void __user *buffer, size_t *lenp, |
2da02997e mm: add dirty_bac... |
195 196 |
loff_t *ppos) { |
fc3501d41 mm: fix dirty_byt... |
197 |
unsigned long old_bytes = vm_dirty_bytes; |
2da02997e mm: add dirty_bac... |
198 |
int ret; |
8d65af789 sysctl: remove "s... |
199 |
ret = proc_doulongvec_minmax(table, write, buffer, lenp, ppos); |
2da02997e mm: add dirty_bac... |
200 201 202 |
if (ret == 0 && write && vm_dirty_bytes != old_bytes) { update_completion_period(); vm_dirty_ratio = 0; |
04fbfdc14 mm: per device di... |
203 204 205 206 207 208 209 210 211 212 |
} return ret; } /* * Increment the BDI's writeout completion count and the global writeout * completion count. Called from test_clear_page_writeback(). */ static inline void __bdi_writeout_inc(struct backing_dev_info *bdi) { |
a42dde041 mm: bdi: allow se... |
213 214 |
__prop_inc_percpu_max(&vm_completions, &bdi->completions, bdi->max_prop_frac); |
04fbfdc14 mm: per device di... |
215 |
} |
dd5656e59 mm: bdi: export b... |
216 217 218 219 220 221 222 223 224 |
void bdi_writeout_inc(struct backing_dev_info *bdi) { unsigned long flags; local_irq_save(flags); __bdi_writeout_inc(bdi); local_irq_restore(flags); } EXPORT_SYMBOL_GPL(bdi_writeout_inc); |
1cf6e7d83 mm: task dirty ac... |
225 |
void task_dirty_inc(struct task_struct *tsk) |
3e26c149c mm: dirty balanci... |
226 227 228 |
{ prop_inc_single(&vm_dirties, &tsk->dirties); } |
04fbfdc14 mm: per device di... |
229 230 231 232 233 234 235 236 237 238 239 240 241 242 |
/* * Obtain an accurate fraction of the BDI's portion. */ static void bdi_writeout_fraction(struct backing_dev_info *bdi, long *numerator, long *denominator) { if (bdi_cap_writeback_dirty(bdi)) { prop_fraction_percpu(&vm_completions, &bdi->completions, numerator, denominator); } else { *numerator = 0; *denominator = 1; } } |
3e26c149c mm: dirty balanci... |
243 244 245 246 247 248 249 250 |
static inline void task_dirties_fraction(struct task_struct *tsk, long *numerator, long *denominator) { prop_fraction_single(&vm_dirties, &tsk->dirties, numerator, denominator); } /* |
1babe1838 writeback: add co... |
251 |
* task_dirty_limit - scale down dirty throttling threshold for one task |
3e26c149c mm: dirty balanci... |
252 253 254 255 |
* * task specific dirty limit: * * dirty -= (dirty/8) * p_{t} |
1babe1838 writeback: add co... |
256 257 258 259 260 261 262 |
* * To protect light/slow dirtying tasks from heavier/fast ones, we start * throttling individual tasks before reaching the bdi dirty limit. * Relatively low thresholds will be allocated to heavy dirtiers. So when * dirty pages grow large, heavy dirtiers will be throttled first, which will * effectively curb the growth of dirty pages. Light dirtiers with high enough * dirty threshold may never get throttled. |
3e26c149c mm: dirty balanci... |
263 |
*/ |
16c4042f0 writeback: avoid ... |
264 265 |
static unsigned long task_dirty_limit(struct task_struct *tsk, unsigned long bdi_dirty) |
3e26c149c mm: dirty balanci... |
266 267 |
{ long numerator, denominator; |
16c4042f0 writeback: avoid ... |
268 |
unsigned long dirty = bdi_dirty; |
3e26c149c mm: dirty balanci... |
269 270 271 272 273 274 275 |
u64 inv = dirty >> 3; task_dirties_fraction(tsk, &numerator, &denominator); inv *= numerator; do_div(inv, denominator); dirty -= inv; |
3e26c149c mm: dirty balanci... |
276 |
|
16c4042f0 writeback: avoid ... |
277 |
return max(dirty, bdi_dirty/2); |
3e26c149c mm: dirty balanci... |
278 |
} |
04fbfdc14 mm: per device di... |
279 |
/* |
189d3c4a9 mm: bdi: allow se... |
280 281 |
* */ |
189d3c4a9 mm: bdi: allow se... |
282 283 284 285 286 |
static unsigned int bdi_min_ratio; int bdi_set_min_ratio(struct backing_dev_info *bdi, unsigned int min_ratio) { int ret = 0; |
189d3c4a9 mm: bdi: allow se... |
287 |
|
cfc4ba536 writeback: use RC... |
288 |
spin_lock_bh(&bdi_lock); |
a42dde041 mm: bdi: allow se... |
289 |
if (min_ratio > bdi->max_ratio) { |
189d3c4a9 mm: bdi: allow se... |
290 |
ret = -EINVAL; |
a42dde041 mm: bdi: allow se... |
291 292 293 294 295 296 297 298 299 |
} else { min_ratio -= bdi->min_ratio; if (bdi_min_ratio + min_ratio < 100) { bdi_min_ratio += min_ratio; bdi->min_ratio += min_ratio; } else { ret = -EINVAL; } } |
cfc4ba536 writeback: use RC... |
300 |
spin_unlock_bh(&bdi_lock); |
a42dde041 mm: bdi: allow se... |
301 302 303 304 305 306 |
return ret; } int bdi_set_max_ratio(struct backing_dev_info *bdi, unsigned max_ratio) { |
a42dde041 mm: bdi: allow se... |
307 308 309 310 |
int ret = 0; if (max_ratio > 100) return -EINVAL; |
cfc4ba536 writeback: use RC... |
311 |
spin_lock_bh(&bdi_lock); |
a42dde041 mm: bdi: allow se... |
312 313 314 315 316 317 |
if (bdi->min_ratio > max_ratio) { ret = -EINVAL; } else { bdi->max_ratio = max_ratio; bdi->max_prop_frac = (PROP_FRAC_BASE * max_ratio) / 100; } |
cfc4ba536 writeback: use RC... |
318 |
spin_unlock_bh(&bdi_lock); |
189d3c4a9 mm: bdi: allow se... |
319 320 321 |
return ret; } |
a42dde041 mm: bdi: allow se... |
322 |
EXPORT_SYMBOL(bdi_set_max_ratio); |
189d3c4a9 mm: bdi: allow se... |
323 324 |
/* |
1da177e4c Linux-2.6.12-rc2 |
325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 |
* Work out the current dirty-memory clamping and background writeout * thresholds. * * The main aim here is to lower them aggressively if there is a lot of mapped * memory around. To avoid stressing page reclaim with lots of unreclaimable * pages. It is better to clamp down on writers than to start swapping, and * performing lots of scanning. * * We only allow 1/2 of the currently-unmapped memory to be dirtied. * * We don't permit the clamping level to fall below 5% - that is getting rather * excessive. * * We make sure that the background writeout level is below the adjusted * clamping level. */ |
1b4244647 Use ZVC counters ... |
341 342 343 344 345 346 |
static unsigned long highmem_dirtyable_memory(unsigned long total) { #ifdef CONFIG_HIGHMEM int node; unsigned long x = 0; |
37b07e416 memoryless nodes:... |
347 |
for_each_node_state(node, N_HIGH_MEMORY) { |
1b4244647 Use ZVC counters ... |
348 349 |
struct zone *z = &NODE_DATA(node)->node_zones[ZONE_HIGHMEM]; |
adea02a1b mm: count only re... |
350 351 |
x += zone_page_state(z, NR_FREE_PAGES) + zone_reclaimable_pages(z); |
1b4244647 Use ZVC counters ... |
352 353 354 355 356 357 358 359 360 361 362 363 |
} /* * Make sure that the number of highmem pages is never larger * than the number of the total dirtyable memory. This can only * occur in very strange VM situations but we want to make sure * that this does not occur. */ return min(x, total); #else return 0; #endif } |
3eefae994 ftrace: limit tra... |
364 365 366 367 368 369 370 |
/** * determine_dirtyable_memory - amount of memory that may be used * * Returns the numebr of pages that can currently be freed and used * by the kernel for direct mappings. */ unsigned long determine_dirtyable_memory(void) |
1b4244647 Use ZVC counters ... |
371 372 |
{ unsigned long x; |
adea02a1b mm: count only re... |
373 |
x = global_page_state(NR_FREE_PAGES) + global_reclaimable_pages(); |
195cf453d mm/page-writeback... |
374 375 376 |
if (!vm_highmem_is_dirtyable) x -= highmem_dirtyable_memory(x); |
1b4244647 Use ZVC counters ... |
377 378 |
return x + 1; /* Ensure that we never return 0 */ } |
03ab450f0 mm/page-writeback... |
379 |
/* |
1babe1838 writeback: add co... |
380 381 382 383 384 385 |
* global_dirty_limits - background-writeback and dirty-throttling thresholds * * Calculate the dirty thresholds based on sysctl parameters * - vm.dirty_background_ratio or vm.dirty_background_bytes * - vm.dirty_ratio or vm.dirty_bytes * The dirty limits will be lifted by 1/4 for PF_LESS_THROTTLE (ie. nfsd) and |
ebd1373d4 writeback: fix gl... |
386 |
* real-time tasks. |
1babe1838 writeback: add co... |
387 |
*/ |
16c4042f0 writeback: avoid ... |
388 |
void global_dirty_limits(unsigned long *pbackground, unsigned long *pdirty) |
1da177e4c Linux-2.6.12-rc2 |
389 |
{ |
364aeb284 mm: change dirty ... |
390 391 |
unsigned long background; unsigned long dirty; |
240c879f2 writeback: avoid ... |
392 |
unsigned long uninitialized_var(available_memory); |
1da177e4c Linux-2.6.12-rc2 |
393 |
struct task_struct *tsk; |
240c879f2 writeback: avoid ... |
394 395 |
if (!vm_dirty_bytes || !dirty_background_bytes) available_memory = determine_dirtyable_memory(); |
2da02997e mm: add dirty_bac... |
396 397 |
if (vm_dirty_bytes) dirty = DIV_ROUND_UP(vm_dirty_bytes, PAGE_SIZE); |
4cbec4c8b writeback: remove... |
398 399 |
else dirty = (vm_dirty_ratio * available_memory) / 100; |
1da177e4c Linux-2.6.12-rc2 |
400 |
|
2da02997e mm: add dirty_bac... |
401 402 403 404 |
if (dirty_background_bytes) background = DIV_ROUND_UP(dirty_background_bytes, PAGE_SIZE); else background = (dirty_background_ratio * available_memory) / 100; |
1da177e4c Linux-2.6.12-rc2 |
405 |
|
2da02997e mm: add dirty_bac... |
406 407 |
if (background >= dirty) background = dirty / 2; |
1da177e4c Linux-2.6.12-rc2 |
408 409 410 411 412 413 414 |
tsk = current; if (tsk->flags & PF_LESS_THROTTLE || rt_task(tsk)) { background += background / 4; dirty += dirty / 4; } *pbackground = background; *pdirty = dirty; |
16c4042f0 writeback: avoid ... |
415 |
} |
04fbfdc14 mm: per device di... |
416 |
|
03ab450f0 mm/page-writeback... |
417 |
/* |
1babe1838 writeback: add co... |
418 419 420 421 422 423 424 425 426 427 |
* bdi_dirty_limit - @bdi's share of dirty throttling threshold * * Allocate high/low dirty limits to fast/slow devices, in order to prevent * - starving fast devices * - piling up dirty pages (that will take long time to sync) on slow devices * * The bdi's share of dirty limit will be adapting to its throughput and * bounded by the bdi->min_ratio and/or bdi->max_ratio parameters, if set. */ unsigned long bdi_dirty_limit(struct backing_dev_info *bdi, unsigned long dirty) |
16c4042f0 writeback: avoid ... |
428 429 430 |
{ u64 bdi_dirty; long numerator, denominator; |
04fbfdc14 mm: per device di... |
431 |
|
16c4042f0 writeback: avoid ... |
432 433 434 435 |
/* * Calculate this BDI's share of the dirty ratio. */ bdi_writeout_fraction(bdi, &numerator, &denominator); |
04fbfdc14 mm: per device di... |
436 |
|
16c4042f0 writeback: avoid ... |
437 438 439 |
bdi_dirty = (dirty * (100 - bdi_min_ratio)) / 100; bdi_dirty *= numerator; do_div(bdi_dirty, denominator); |
04fbfdc14 mm: per device di... |
440 |
|
16c4042f0 writeback: avoid ... |
441 442 443 444 445 |
bdi_dirty += (dirty * bdi->min_ratio) / 100; if (bdi_dirty > (dirty * bdi->max_ratio) / 100) bdi_dirty = dirty * bdi->max_ratio / 100; return bdi_dirty; |
1da177e4c Linux-2.6.12-rc2 |
446 447 448 449 450 451 |
} /* * balance_dirty_pages() must be called by processes which are generating dirty * data. It looks at the number of dirty pages in the machine and will force * the caller to perform writeback if the system is over `vm_dirty_ratio'. |
5b0830cb9 writeback: get ri... |
452 453 |
* If we're over `background_thresh' then the writeback threads are woken to * perform some writeout. |
1da177e4c Linux-2.6.12-rc2 |
454 |
*/ |
3a2e9a5a2 writeback: balanc... |
455 456 |
static void balance_dirty_pages(struct address_space *mapping, unsigned long write_chunk) |
1da177e4c Linux-2.6.12-rc2 |
457 |
{ |
5fce25a9d mm: speed up writ... |
458 459 |
long nr_reclaimable, bdi_nr_reclaimable; long nr_writeback, bdi_nr_writeback; |
364aeb284 mm: change dirty ... |
460 461 462 |
unsigned long background_thresh; unsigned long dirty_thresh; unsigned long bdi_thresh; |
1da177e4c Linux-2.6.12-rc2 |
463 |
unsigned long pages_written = 0; |
87c6a9b25 writeback: make b... |
464 |
unsigned long pause = 1; |
e50e37201 writeback: balanc... |
465 |
bool dirty_exceeded = false; |
1da177e4c Linux-2.6.12-rc2 |
466 467 468 469 |
struct backing_dev_info *bdi = mapping->backing_dev_info; for (;;) { struct writeback_control wbc = { |
1da177e4c Linux-2.6.12-rc2 |
470 471 472 |
.sync_mode = WB_SYNC_NONE, .older_than_this = NULL, .nr_to_write = write_chunk, |
111ebb6e6 [PATCH] writeback... |
473 |
.range_cyclic = 1, |
1da177e4c Linux-2.6.12-rc2 |
474 |
}; |
5fce25a9d mm: speed up writ... |
475 476 477 |
nr_reclaimable = global_page_state(NR_FILE_DIRTY) + global_page_state(NR_UNSTABLE_NFS); nr_writeback = global_page_state(NR_WRITEBACK); |
16c4042f0 writeback: avoid ... |
478 479 480 481 482 483 484 |
global_dirty_limits(&background_thresh, &dirty_thresh); /* * Throttle it only when the background writeback cannot * catch-up. This avoids (excessively) small writeouts * when the bdi limits are ramping up. */ |
4cbec4c8b writeback: remove... |
485 |
if (nr_reclaimable + nr_writeback <= |
16c4042f0 writeback: avoid ... |
486 487 488 489 490 |
(background_thresh + dirty_thresh) / 2) break; bdi_thresh = bdi_dirty_limit(bdi, dirty_thresh); bdi_thresh = task_dirty_limit(current, bdi_thresh); |
e50e37201 writeback: balanc... |
491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 |
/* * In order to avoid the stacked BDI deadlock we need * to ensure we accurately count the 'dirty' pages when * the threshold is low. * * Otherwise it would be possible to get thresh+n pages * reported dirty, even though there are thresh-m pages * actually dirty; with m+n sitting in the percpu * deltas. */ if (bdi_thresh < 2*bdi_stat_error(bdi)) { bdi_nr_reclaimable = bdi_stat_sum(bdi, BDI_RECLAIMABLE); bdi_nr_writeback = bdi_stat_sum(bdi, BDI_WRITEBACK); } else { bdi_nr_reclaimable = bdi_stat(bdi, BDI_RECLAIMABLE); bdi_nr_writeback = bdi_stat(bdi, BDI_WRITEBACK); } |
5fce25a9d mm: speed up writ... |
508 |
|
e50e37201 writeback: balanc... |
509 510 511 512 513 514 515 |
/* * The bdi thresh is somehow "soft" limit derived from the * global "hard" limit. The former helps to prevent heavy IO * bdi or process from holding back light ones; The latter is * the last resort safeguard. */ dirty_exceeded = |
4cbec4c8b writeback: remove... |
516 517 |
(bdi_nr_reclaimable + bdi_nr_writeback > bdi_thresh) || (nr_reclaimable + nr_writeback > dirty_thresh); |
e50e37201 writeback: balanc... |
518 519 |
if (!dirty_exceeded) |
04fbfdc14 mm: per device di... |
520 |
break; |
1da177e4c Linux-2.6.12-rc2 |
521 |
|
04fbfdc14 mm: per device di... |
522 523 |
if (!bdi->dirty_exceeded) bdi->dirty_exceeded = 1; |
1da177e4c Linux-2.6.12-rc2 |
524 525 526 527 528 529 |
/* Note: nr_reclaimable denotes nr_dirty + nr_unstable. * Unstable writes are a feature of certain networked * filesystems (i.e. NFS) in which data may have been * written to the server's write cache, but has not yet * been flushed to permanent storage. |
d7831a0bd mm: prevent balan... |
530 531 532 |
* Only move pages to writeback if this bdi is over its * threshold otherwise wait until the disk writes catch * up. |
1da177e4c Linux-2.6.12-rc2 |
533 |
*/ |
028c2dd18 writeback: Add tr... |
534 |
trace_wbc_balance_dirty_start(&wbc, bdi); |
d7831a0bd mm: prevent balan... |
535 |
if (bdi_nr_reclaimable > bdi_thresh) { |
9c3a8ee8a writeback: remove... |
536 |
writeback_inodes_wb(&bdi->wb, &wbc); |
1da177e4c Linux-2.6.12-rc2 |
537 |
pages_written += write_chunk - wbc.nr_to_write; |
028c2dd18 writeback: Add tr... |
538 |
trace_wbc_balance_dirty_written(&wbc, bdi); |
e50e37201 writeback: balanc... |
539 540 |
if (pages_written >= write_chunk) break; /* We've done our duty */ |
04fbfdc14 mm: per device di... |
541 |
} |
028c2dd18 writeback: Add tr... |
542 |
trace_wbc_balance_dirty_wait(&wbc, bdi); |
d153ba644 writeback: do uni... |
543 |
__set_current_state(TASK_UNINTERRUPTIBLE); |
d25105e89 writeback: accoun... |
544 |
io_schedule_timeout(pause); |
87c6a9b25 writeback: make b... |
545 546 547 548 549 550 551 552 |
/* * Increase the delay for each loop, up to our previous * default of taking a 100ms nap. */ pause <<= 1; if (pause > HZ / 10) pause = HZ / 10; |
1da177e4c Linux-2.6.12-rc2 |
553 |
} |
e50e37201 writeback: balanc... |
554 |
if (!dirty_exceeded && bdi->dirty_exceeded) |
04fbfdc14 mm: per device di... |
555 |
bdi->dirty_exceeded = 0; |
1da177e4c Linux-2.6.12-rc2 |
556 557 |
if (writeback_in_progress(bdi)) |
5b0830cb9 writeback: get ri... |
558 |
return; |
1da177e4c Linux-2.6.12-rc2 |
559 560 561 562 563 564 565 566 567 568 |
/* * In laptop mode, we wait until hitting the higher threshold before * starting background writeout, and then write out all the way down * to the lower threshold. So slow writers cause minimal disk activity. * * In normal mode, we start background writeout at the lower * background_thresh, to keep the amount of dirty memory low. */ if ((laptop_mode && pages_written) || |
e50e37201 writeback: balanc... |
569 |
(!laptop_mode && (nr_reclaimable > background_thresh))) |
c5444198c writeback: simpli... |
570 |
bdi_start_background_writeback(bdi); |
1da177e4c Linux-2.6.12-rc2 |
571 |
} |
a200ee182 mm: set_page_dirt... |
572 |
void set_page_dirty_balance(struct page *page, int page_mkwrite) |
edc79b2a4 [PATCH] mm: balan... |
573 |
{ |
a200ee182 mm: set_page_dirt... |
574 |
if (set_page_dirty(page) || page_mkwrite) { |
edc79b2a4 [PATCH] mm: balan... |
575 576 577 578 579 580 |
struct address_space *mapping = page_mapping(page); if (mapping) balance_dirty_pages_ratelimited(mapping); } } |
245b2e70e percpu: clean up ... |
581 |
static DEFINE_PER_CPU(unsigned long, bdp_ratelimits) = 0; |
1da177e4c Linux-2.6.12-rc2 |
582 |
/** |
fa5a734e4 [PATCH] balance_d... |
583 |
* balance_dirty_pages_ratelimited_nr - balance dirty memory state |
67be2dd1b [PATCH] DocBook: ... |
584 |
* @mapping: address_space which was dirtied |
a580290c3 Documentation: fi... |
585 |
* @nr_pages_dirtied: number of pages which the caller has just dirtied |
1da177e4c Linux-2.6.12-rc2 |
586 587 588 589 590 591 592 593 594 595 |
* * Processes which are dirtying memory should call in here once for each page * which was newly dirtied. The function will periodically check the system's * dirty state and will initiate writeback if needed. * * On really big machines, get_writeback_state is expensive, so try to avoid * calling it too often (ratelimiting). But once we're over the dirty memory * limit we decrease the ratelimiting by a lot, to prevent individual processes * from overshooting the limit by (ratelimit_pages) each. */ |
fa5a734e4 [PATCH] balance_d... |
596 597 |
void balance_dirty_pages_ratelimited_nr(struct address_space *mapping, unsigned long nr_pages_dirtied) |
1da177e4c Linux-2.6.12-rc2 |
598 |
{ |
fa5a734e4 [PATCH] balance_d... |
599 600 |
unsigned long ratelimit; unsigned long *p; |
1da177e4c Linux-2.6.12-rc2 |
601 602 |
ratelimit = ratelimit_pages; |
04fbfdc14 mm: per device di... |
603 |
if (mapping->backing_dev_info->dirty_exceeded) |
1da177e4c Linux-2.6.12-rc2 |
604 605 606 607 608 609 |
ratelimit = 8; /* * Check the rate limiting. Also, we do not want to throttle real-time * tasks in balance_dirty_pages(). Period. */ |
fa5a734e4 [PATCH] balance_d... |
610 |
preempt_disable(); |
245b2e70e percpu: clean up ... |
611 |
p = &__get_cpu_var(bdp_ratelimits); |
fa5a734e4 [PATCH] balance_d... |
612 613 |
*p += nr_pages_dirtied; if (unlikely(*p >= ratelimit)) { |
3a2e9a5a2 writeback: balanc... |
614 |
ratelimit = sync_writeback_pages(*p); |
fa5a734e4 [PATCH] balance_d... |
615 616 |
*p = 0; preempt_enable(); |
3a2e9a5a2 writeback: balanc... |
617 |
balance_dirty_pages(mapping, ratelimit); |
1da177e4c Linux-2.6.12-rc2 |
618 619 |
return; } |
fa5a734e4 [PATCH] balance_d... |
620 |
preempt_enable(); |
1da177e4c Linux-2.6.12-rc2 |
621 |
} |
fa5a734e4 [PATCH] balance_d... |
622 |
EXPORT_SYMBOL(balance_dirty_pages_ratelimited_nr); |
1da177e4c Linux-2.6.12-rc2 |
623 |
|
232ea4d69 [PATCH] throttle_... |
624 |
void throttle_vm_writeout(gfp_t gfp_mask) |
1da177e4c Linux-2.6.12-rc2 |
625 |
{ |
364aeb284 mm: change dirty ... |
626 627 |
unsigned long background_thresh; unsigned long dirty_thresh; |
1da177e4c Linux-2.6.12-rc2 |
628 629 |
for ( ; ; ) { |
16c4042f0 writeback: avoid ... |
630 |
global_dirty_limits(&background_thresh, &dirty_thresh); |
1da177e4c Linux-2.6.12-rc2 |
631 632 633 634 635 636 |
/* * Boost the allowable dirty threshold a bit for page * allocators so they don't get DoS'ed by heavy writers */ dirty_thresh += dirty_thresh / 10; /* wheeee... */ |
c24f21bda [PATCH] zoned vm ... |
637 638 639 |
if (global_page_state(NR_UNSTABLE_NFS) + global_page_state(NR_WRITEBACK) <= dirty_thresh) break; |
8aa7e847d Fix congestion_wa... |
640 |
congestion_wait(BLK_RW_ASYNC, HZ/10); |
369f2389e writeback: remove... |
641 642 643 644 645 646 647 648 |
/* * The caller might hold locks which can prevent IO completion * or progress in the filesystem. So we cannot just sit here * waiting for IO to complete. */ if ((gfp_mask & (__GFP_FS|__GFP_IO)) != (__GFP_FS|__GFP_IO)) break; |
1da177e4c Linux-2.6.12-rc2 |
649 650 |
} } |
1da177e4c Linux-2.6.12-rc2 |
651 |
/* |
1da177e4c Linux-2.6.12-rc2 |
652 653 654 |
* sysctl handler for /proc/sys/vm/dirty_writeback_centisecs */ int dirty_writeback_centisecs_handler(ctl_table *table, int write, |
8d65af789 sysctl: remove "s... |
655 |
void __user *buffer, size_t *length, loff_t *ppos) |
1da177e4c Linux-2.6.12-rc2 |
656 |
{ |
8d65af789 sysctl: remove "s... |
657 |
proc_dointvec(table, write, buffer, length, ppos); |
6423104b6 writeback: fixups... |
658 |
bdi_arm_supers_timer(); |
1da177e4c Linux-2.6.12-rc2 |
659 660 |
return 0; } |
c2c4986ed writeback: fix pr... |
661 |
#ifdef CONFIG_BLOCK |
31373d09d laptop-mode: Make... |
662 |
void laptop_mode_timer_fn(unsigned long data) |
1da177e4c Linux-2.6.12-rc2 |
663 |
{ |
31373d09d laptop-mode: Make... |
664 665 666 |
struct request_queue *q = (struct request_queue *)data; int nr_pages = global_page_state(NR_FILE_DIRTY) + global_page_state(NR_UNSTABLE_NFS); |
1da177e4c Linux-2.6.12-rc2 |
667 |
|
31373d09d laptop-mode: Make... |
668 669 670 671 |
/* * We want to write everything out, not just down to the dirty * threshold */ |
31373d09d laptop-mode: Make... |
672 |
if (bdi_has_dirty_io(&q->backing_dev_info)) |
c5444198c writeback: simpli... |
673 |
bdi_start_writeback(&q->backing_dev_info, nr_pages); |
1da177e4c Linux-2.6.12-rc2 |
674 675 676 677 678 679 680 |
} /* * We've spun up the disk and we're in laptop mode: schedule writeback * of all dirty data a few seconds from now. If the flush is already scheduled * then push it back - the user is still using the disk. */ |
31373d09d laptop-mode: Make... |
681 |
void laptop_io_completion(struct backing_dev_info *info) |
1da177e4c Linux-2.6.12-rc2 |
682 |
{ |
31373d09d laptop-mode: Make... |
683 |
mod_timer(&info->laptop_mode_wb_timer, jiffies + laptop_mode); |
1da177e4c Linux-2.6.12-rc2 |
684 685 686 687 688 689 690 691 692 |
} /* * We're in laptop mode and we've just synced. The sync's writes will have * caused another writeback to be scheduled by laptop_io_completion. * Nothing needs to be written back anymore, so we unschedule the writeback. */ void laptop_sync_completion(void) { |
31373d09d laptop-mode: Make... |
693 694 695 696 697 698 699 700 |
struct backing_dev_info *bdi; rcu_read_lock(); list_for_each_entry_rcu(bdi, &bdi_list, bdi_list) del_timer(&bdi->laptop_mode_wb_timer); rcu_read_unlock(); |
1da177e4c Linux-2.6.12-rc2 |
701 |
} |
c2c4986ed writeback: fix pr... |
702 |
#endif |
1da177e4c Linux-2.6.12-rc2 |
703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 |
/* * If ratelimit_pages is too high then we can get into dirty-data overload * if a large number of processes all perform writes at the same time. * If it is too low then SMP machines will call the (expensive) * get_writeback_state too often. * * Here we set ratelimit_pages to a level which ensures that when all CPUs are * dirtying in parallel, we cannot go more than 3% (1/32) over the dirty memory * thresholds before writeback cuts in. * * But the limit should not be set too high. Because it also controls the * amount of memory which the balance_dirty_pages() caller has to write back. * If this is too large then the caller will block on the IO queue all the * time. So limit it to four megabytes - the balance_dirty_pages() caller * will write six megabyte chunks, max. */ |
2d1d43f6a [PATCH] call mm/p... |
720 |
void writeback_set_ratelimit(void) |
1da177e4c Linux-2.6.12-rc2 |
721 |
{ |
40c99aae2 [PATCH] remove st... |
722 |
ratelimit_pages = vm_total_pages / (num_online_cpus() * 32); |
1da177e4c Linux-2.6.12-rc2 |
723 724 725 726 727 |
if (ratelimit_pages < 16) ratelimit_pages = 16; if (ratelimit_pages * PAGE_CACHE_SIZE > 4096 * 1024) ratelimit_pages = (4096 * 1024) / PAGE_CACHE_SIZE; } |
26c2143b6 [PATCH] cpu hotpl... |
728 |
static int __cpuinit |
1da177e4c Linux-2.6.12-rc2 |
729 730 |
ratelimit_handler(struct notifier_block *self, unsigned long u, void *v) { |
2d1d43f6a [PATCH] call mm/p... |
731 |
writeback_set_ratelimit(); |
aa0f03037 [PATCH] Change co... |
732 |
return NOTIFY_DONE; |
1da177e4c Linux-2.6.12-rc2 |
733 |
} |
74b85f379 [PATCH] cpu hotpl... |
734 |
static struct notifier_block __cpuinitdata ratelimit_nb = { |
1da177e4c Linux-2.6.12-rc2 |
735 736 737 738 739 |
.notifier_call = ratelimit_handler, .next = NULL, }; /* |
dc6e29da9 Fix balance_dirty... |
740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 |
* Called early on to tune the page writeback dirty limits. * * We used to scale dirty pages according to how total memory * related to pages that could be allocated for buffers (by * comparing nr_free_buffer_pages() to vm_total_pages. * * However, that was when we used "dirty_ratio" to scale with * all memory, and we don't do that any more. "dirty_ratio" * is now applied to total non-HIGHPAGE memory (by subtracting * totalhigh_pages from vm_total_pages), and as such we can't * get into the old insane situation any more where we had * large amounts of dirty pages compared to a small amount of * non-HIGHMEM memory. * * But we might still want to scale the dirty_ratio by how * much memory the box has.. |
1da177e4c Linux-2.6.12-rc2 |
756 757 758 |
*/ void __init page_writeback_init(void) { |
04fbfdc14 mm: per device di... |
759 |
int shift; |
2d1d43f6a [PATCH] call mm/p... |
760 |
writeback_set_ratelimit(); |
1da177e4c Linux-2.6.12-rc2 |
761 |
register_cpu_notifier(&ratelimit_nb); |
04fbfdc14 mm: per device di... |
762 763 764 |
shift = calc_period_shift(); prop_descriptor_init(&vm_completions, shift); |
3e26c149c mm: dirty balanci... |
765 |
prop_descriptor_init(&vm_dirties, shift); |
1da177e4c Linux-2.6.12-rc2 |
766 |
} |
811d736f9 [PATCH] BLOCK: Di... |
767 |
/** |
f446daaea mm: implement wri... |
768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 |
* tag_pages_for_writeback - tag pages to be written by write_cache_pages * @mapping: address space structure to write * @start: starting page index * @end: ending page index (inclusive) * * This function scans the page range from @start to @end (inclusive) and tags * all pages that have DIRTY tag set with a special TOWRITE tag. The idea is * that write_cache_pages (or whoever calls this function) will then use * TOWRITE tag to identify pages eligible for writeback. This mechanism is * used to avoid livelocking of writeback by a process steadily creating new * dirty pages in the file (thus it is important for this function to be quick * so that it can tag pages faster than a dirtying process can create them). */ /* * We tag pages in batches of WRITEBACK_TAG_BATCH to reduce tree_lock latency. */ |
f446daaea mm: implement wri... |
784 785 786 |
void tag_pages_for_writeback(struct address_space *mapping, pgoff_t start, pgoff_t end) { |
3c111a071 mm: fix fatal ker... |
787 |
#define WRITEBACK_TAG_BATCH 4096 |
f446daaea mm: implement wri... |
788 789 790 791 792 793 794 795 796 797 |
unsigned long tagged; do { spin_lock_irq(&mapping->tree_lock); tagged = radix_tree_range_tag_if_tagged(&mapping->page_tree, &start, end, WRITEBACK_TAG_BATCH, PAGECACHE_TAG_DIRTY, PAGECACHE_TAG_TOWRITE); spin_unlock_irq(&mapping->tree_lock); WARN_ON_ONCE(tagged > WRITEBACK_TAG_BATCH); cond_resched(); |
d5ed3a4af lib/radix-tree.c:... |
798 799 |
/* We check 'start' to handle wrapping when end == ~0UL */ } while (tagged >= WRITEBACK_TAG_BATCH && start); |
f446daaea mm: implement wri... |
800 801 802 803 |
} EXPORT_SYMBOL(tag_pages_for_writeback); /** |
0ea971801 consolidate gener... |
804 |
* write_cache_pages - walk the list of dirty pages of the given address space and write all of them. |
811d736f9 [PATCH] BLOCK: Di... |
805 806 |
* @mapping: address space structure to write * @wbc: subtract the number of written pages from *@wbc->nr_to_write |
0ea971801 consolidate gener... |
807 808 |
* @writepage: function called for each page * @data: data passed to writepage function |
811d736f9 [PATCH] BLOCK: Di... |
809 |
* |
0ea971801 consolidate gener... |
810 |
* If a page is already under I/O, write_cache_pages() skips it, even |
811d736f9 [PATCH] BLOCK: Di... |
811 812 813 814 815 816 |
* if it's dirty. This is desirable behaviour for memory-cleaning writeback, * but it is INCORRECT for data-integrity system calls such as fsync(). fsync() * and msync() need to guarantee that all the data which was dirty at the time * the call was made get new I/O started against them. If wbc->sync_mode is * WB_SYNC_ALL then we were called for data integrity and we must wait for * existing IO to complete. |
f446daaea mm: implement wri... |
817 818 819 820 821 822 823 |
* * To avoid livelocks (when other process dirties new pages), we first tag * pages which should be written back with TOWRITE tag and only then start * writing them. For data-integrity sync we have to be careful so that we do * not miss some pages (e.g., because some other process has cleared TOWRITE * tag we set). The rule we follow is that TOWRITE tag can be cleared only * by the process clearing the DIRTY tag (and submitting the page for IO). |
811d736f9 [PATCH] BLOCK: Di... |
824 |
*/ |
0ea971801 consolidate gener... |
825 826 827 |
int write_cache_pages(struct address_space *mapping, struct writeback_control *wbc, writepage_t writepage, void *data) |
811d736f9 [PATCH] BLOCK: Di... |
828 |
{ |
811d736f9 [PATCH] BLOCK: Di... |
829 830 |
int ret = 0; int done = 0; |
811d736f9 [PATCH] BLOCK: Di... |
831 832 |
struct pagevec pvec; int nr_pages; |
31a12666d mm: write_cache_p... |
833 |
pgoff_t uninitialized_var(writeback_index); |
811d736f9 [PATCH] BLOCK: Di... |
834 835 |
pgoff_t index; pgoff_t end; /* Inclusive */ |
bd19e012f mm: write_cache_p... |
836 |
pgoff_t done_index; |
31a12666d mm: write_cache_p... |
837 |
int cycled; |
811d736f9 [PATCH] BLOCK: Di... |
838 |
int range_whole = 0; |
f446daaea mm: implement wri... |
839 |
int tag; |
811d736f9 [PATCH] BLOCK: Di... |
840 |
|
811d736f9 [PATCH] BLOCK: Di... |
841 842 |
pagevec_init(&pvec, 0); if (wbc->range_cyclic) { |
31a12666d mm: write_cache_p... |
843 844 845 846 847 848 |
writeback_index = mapping->writeback_index; /* prev offset */ index = writeback_index; if (index == 0) cycled = 1; else cycled = 0; |
811d736f9 [PATCH] BLOCK: Di... |
849 850 851 852 853 854 |
end = -1; } else { index = wbc->range_start >> PAGE_CACHE_SHIFT; end = wbc->range_end >> PAGE_CACHE_SHIFT; if (wbc->range_start == 0 && wbc->range_end == LLONG_MAX) range_whole = 1; |
31a12666d mm: write_cache_p... |
855 |
cycled = 1; /* ignore range_cyclic tests */ |
811d736f9 [PATCH] BLOCK: Di... |
856 |
} |
f446daaea mm: implement wri... |
857 858 859 860 |
if (wbc->sync_mode == WB_SYNC_ALL) tag = PAGECACHE_TAG_TOWRITE; else tag = PAGECACHE_TAG_DIRTY; |
811d736f9 [PATCH] BLOCK: Di... |
861 |
retry: |
f446daaea mm: implement wri... |
862 863 |
if (wbc->sync_mode == WB_SYNC_ALL) tag_pages_for_writeback(mapping, index, end); |
bd19e012f mm: write_cache_p... |
864 |
done_index = index; |
5a3d5c981 mm: write_cache_p... |
865 866 |
while (!done && (index <= end)) { int i; |
f446daaea mm: implement wri... |
867 |
nr_pages = pagevec_lookup_tag(&pvec, mapping, &index, tag, |
5a3d5c981 mm: write_cache_p... |
868 869 870 |
min(end - index, (pgoff_t)PAGEVEC_SIZE-1) + 1); if (nr_pages == 0) break; |
811d736f9 [PATCH] BLOCK: Di... |
871 |
|
811d736f9 [PATCH] BLOCK: Di... |
872 873 874 875 |
for (i = 0; i < nr_pages; i++) { struct page *page = pvec.pages[i]; /* |
d5482cdf8 mm: write_cache_p... |
876 877 878 879 880 |
* At this point, the page may be truncated or * invalidated (changing page->mapping to NULL), or * even swizzled back from swapper_space to tmpfs file * mapping. However, page->index will not change * because we have a reference on the page. |
811d736f9 [PATCH] BLOCK: Di... |
881 |
*/ |
d5482cdf8 mm: write_cache_p... |
882 883 884 885 886 887 888 889 890 891 |
if (page->index > end) { /* * can't be range_cyclic (1st pass) because * end == -1 in that case. */ done = 1; break; } done_index = page->index + 1; |
811d736f9 [PATCH] BLOCK: Di... |
892 |
lock_page(page); |
5a3d5c981 mm: write_cache_p... |
893 894 895 896 897 898 899 900 |
/* * Page truncated or invalidated. We can freely skip it * then, even for data integrity operations: the page * has disappeared concurrently, so there could be no * real expectation of this data interity operation * even if there is now a new, dirty page at the same * pagecache address. */ |
811d736f9 [PATCH] BLOCK: Di... |
901 |
if (unlikely(page->mapping != mapping)) { |
5a3d5c981 mm: write_cache_p... |
902 |
continue_unlock: |
811d736f9 [PATCH] BLOCK: Di... |
903 904 905 |
unlock_page(page); continue; } |
515f4a037 mm: write_cache_p... |
906 907 908 909 910 911 912 913 914 915 916 |
if (!PageDirty(page)) { /* someone wrote it for us */ goto continue_unlock; } if (PageWriteback(page)) { if (wbc->sync_mode != WB_SYNC_NONE) wait_on_page_writeback(page); else goto continue_unlock; } |
811d736f9 [PATCH] BLOCK: Di... |
917 |
|
515f4a037 mm: write_cache_p... |
918 919 |
BUG_ON(PageWriteback(page)); if (!clear_page_dirty_for_io(page)) |
5a3d5c981 mm: write_cache_p... |
920 |
goto continue_unlock; |
811d736f9 [PATCH] BLOCK: Di... |
921 |
|
9e094383b writeback: Add tr... |
922 |
trace_wbc_writepage(wbc, mapping->backing_dev_info); |
0ea971801 consolidate gener... |
923 |
ret = (*writepage)(page, wbc, data); |
00266770b mm: write_cache_p... |
924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 |
if (unlikely(ret)) { if (ret == AOP_WRITEPAGE_ACTIVATE) { unlock_page(page); ret = 0; } else { /* * done_index is set past this page, * so media errors will not choke * background writeout for the entire * file. This has consequences for * range_cyclic semantics (ie. it may * not be suitable for data integrity * writeout). */ done = 1; break; } |
0b5649278 writeback: pay at... |
941 |
} |
00266770b mm: write_cache_p... |
942 |
|
546a19242 writeback: write_... |
943 944 945 946 947 948 949 950 951 952 |
/* * We stop writing back only if we are not doing * integrity sync. In case of integrity sync we have to * keep going until we have written all the pages * we tagged for writeback prior to entering this loop. */ if (--wbc->nr_to_write <= 0 && wbc->sync_mode == WB_SYNC_NONE) { done = 1; break; |
05fe478dd mm: write_cache_p... |
953 |
} |
811d736f9 [PATCH] BLOCK: Di... |
954 955 956 957 |
} pagevec_release(&pvec); cond_resched(); } |
3a4c6800f Fix page writebac... |
958 |
if (!cycled && !done) { |
811d736f9 [PATCH] BLOCK: Di... |
959 |
/* |
31a12666d mm: write_cache_p... |
960 |
* range_cyclic: |
811d736f9 [PATCH] BLOCK: Di... |
961 962 963 |
* We hit the last page and there is more work to be done: wrap * back to the start of the file */ |
31a12666d mm: write_cache_p... |
964 |
cycled = 1; |
811d736f9 [PATCH] BLOCK: Di... |
965 |
index = 0; |
31a12666d mm: write_cache_p... |
966 |
end = writeback_index - 1; |
811d736f9 [PATCH] BLOCK: Di... |
967 968 |
goto retry; } |
0b5649278 writeback: pay at... |
969 970 |
if (wbc->range_cyclic || (range_whole && wbc->nr_to_write > 0)) mapping->writeback_index = done_index; |
06d6cf695 mm: Add range_con... |
971 |
|
811d736f9 [PATCH] BLOCK: Di... |
972 973 |
return ret; } |
0ea971801 consolidate gener... |
974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 |
EXPORT_SYMBOL(write_cache_pages); /* * Function used by generic_writepages to call the real writepage * function and set the mapping flags on error */ static int __writepage(struct page *page, struct writeback_control *wbc, void *data) { struct address_space *mapping = data; int ret = mapping->a_ops->writepage(page, wbc); mapping_set_error(mapping, ret); return ret; } /** * generic_writepages - walk the list of dirty pages of the given address space and writepage() all of them. * @mapping: address space structure to write * @wbc: subtract the number of written pages from *@wbc->nr_to_write * * This is a library function, which implements the writepages() * address_space_operation. */ int generic_writepages(struct address_space *mapping, struct writeback_control *wbc) { /* deal with chardevs and other special file */ if (!mapping->a_ops->writepage) return 0; return write_cache_pages(mapping, wbc, __writepage, mapping); } |
811d736f9 [PATCH] BLOCK: Di... |
1006 1007 |
EXPORT_SYMBOL(generic_writepages); |
1da177e4c Linux-2.6.12-rc2 |
1008 1009 |
int do_writepages(struct address_space *mapping, struct writeback_control *wbc) { |
22905f775 identify multipag... |
1010 |
int ret; |
1da177e4c Linux-2.6.12-rc2 |
1011 1012 1013 |
if (wbc->nr_to_write <= 0) return 0; if (mapping->a_ops->writepages) |
d08b3851d [PATCH] mm: track... |
1014 |
ret = mapping->a_ops->writepages(mapping, wbc); |
22905f775 identify multipag... |
1015 1016 |
else ret = generic_writepages(mapping, wbc); |
22905f775 identify multipag... |
1017 |
return ret; |
1da177e4c Linux-2.6.12-rc2 |
1018 1019 1020 1021 |
} /** * write_one_page - write out a single page and optionally wait on I/O |
67be2dd1b [PATCH] DocBook: ... |
1022 1023 |
* @page: the page to write * @wait: if true, wait on writeout |
1da177e4c Linux-2.6.12-rc2 |
1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 |
* * The page must be locked by the caller and will be unlocked upon return. * * write_one_page() returns a negative error code if I/O failed. */ int write_one_page(struct page *page, int wait) { struct address_space *mapping = page->mapping; int ret = 0; struct writeback_control wbc = { .sync_mode = WB_SYNC_ALL, .nr_to_write = 1, }; BUG_ON(!PageLocked(page)); if (wait) wait_on_page_writeback(page); if (clear_page_dirty_for_io(page)) { page_cache_get(page); ret = mapping->a_ops->writepage(page, &wbc); if (ret == 0 && wait) { wait_on_page_writeback(page); if (PageError(page)) ret = -EIO; } page_cache_release(page); } else { unlock_page(page); } return ret; } EXPORT_SYMBOL(write_one_page); /* |
767193253 [PATCH] simplify ... |
1060 1061 1062 1063 1064 |
* For address_spaces which do not use buffers nor write back. */ int __set_page_dirty_no_writeback(struct page *page) { if (!PageDirty(page)) |
c3f0da631 mm/page-writeback... |
1065 |
return !TestSetPageDirty(page); |
767193253 [PATCH] simplify ... |
1066 1067 1068 1069 |
return 0; } /* |
e3a7cca1e vfs: add/use acco... |
1070 1071 1072 1073 1074 1075 1076 |
* Helper function for set_page_dirty family. * NOTE: This relies on being atomic wrt interrupts. */ void account_page_dirtied(struct page *page, struct address_space *mapping) { if (mapping_cap_account_dirty(mapping)) { __inc_zone_page_state(page, NR_FILE_DIRTY); |
ea941f0e2 writeback: add nr... |
1077 |
__inc_zone_page_state(page, NR_DIRTIED); |
e3a7cca1e vfs: add/use acco... |
1078 1079 1080 1081 1082 |
__inc_bdi_stat(mapping->backing_dev_info, BDI_RECLAIMABLE); task_dirty_inc(current); task_io_account_write(PAGE_CACHE_SIZE); } } |
679ceace8 mm: exporting acc... |
1083 |
EXPORT_SYMBOL(account_page_dirtied); |
e3a7cca1e vfs: add/use acco... |
1084 1085 |
/* |
f629d1c9b mm: add account_p... |
1086 1087 1088 1089 1090 1091 1092 |
* Helper function for set_page_writeback family. * NOTE: Unlike account_page_dirtied this does not rely on being atomic * wrt interrupts. */ void account_page_writeback(struct page *page) { inc_zone_page_state(page, NR_WRITEBACK); |
ea941f0e2 writeback: add nr... |
1093 |
inc_zone_page_state(page, NR_WRITTEN); |
f629d1c9b mm: add account_p... |
1094 1095 1096 1097 |
} EXPORT_SYMBOL(account_page_writeback); /* |
1da177e4c Linux-2.6.12-rc2 |
1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 |
* For address_spaces which do not use buffers. Just tag the page as dirty in * its radix tree. * * This is also used when a single buffer is being dirtied: we want to set the * page dirty in that case, but not all the buffers. This is a "bottom-up" * dirtying, whereas __set_page_dirty_buffers() is a "top-down" dirtying. * * Most callers have locked the page, which pins the address_space in memory. * But zap_pte_range() does not lock the page, however in that case the * mapping is pinned by the vma's ->vm_file reference. * * We take care to handle the case where the page was truncated from the |
183ff22bb spelling fixes: mm/ |
1110 |
* mapping by re-checking page_mapping() inside tree_lock. |
1da177e4c Linux-2.6.12-rc2 |
1111 1112 1113 |
*/ int __set_page_dirty_nobuffers(struct page *page) { |
1da177e4c Linux-2.6.12-rc2 |
1114 1115 1116 |
if (!TestSetPageDirty(page)) { struct address_space *mapping = page_mapping(page); struct address_space *mapping2; |
8c08540f8 [PATCH] clean up ... |
1117 1118 |
if (!mapping) return 1; |
19fd62312 mm: spinlock tree... |
1119 |
spin_lock_irq(&mapping->tree_lock); |
8c08540f8 [PATCH] clean up ... |
1120 1121 1122 |
mapping2 = page_mapping(page); if (mapping2) { /* Race with truncate? */ BUG_ON(mapping2 != mapping); |
787d2214c fs: introduce som... |
1123 |
WARN_ON_ONCE(!PagePrivate(page) && !PageUptodate(page)); |
e3a7cca1e vfs: add/use acco... |
1124 |
account_page_dirtied(page, mapping); |
8c08540f8 [PATCH] clean up ... |
1125 1126 1127 |
radix_tree_tag_set(&mapping->page_tree, page_index(page), PAGECACHE_TAG_DIRTY); } |
19fd62312 mm: spinlock tree... |
1128 |
spin_unlock_irq(&mapping->tree_lock); |
8c08540f8 [PATCH] clean up ... |
1129 1130 1131 |
if (mapping->host) { /* !PageAnon && !swapper_space */ __mark_inode_dirty(mapping->host, I_DIRTY_PAGES); |
1da177e4c Linux-2.6.12-rc2 |
1132 |
} |
4741c9fd3 [PATCH] set_page_... |
1133 |
return 1; |
1da177e4c Linux-2.6.12-rc2 |
1134 |
} |
4741c9fd3 [PATCH] set_page_... |
1135 |
return 0; |
1da177e4c Linux-2.6.12-rc2 |
1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 |
} EXPORT_SYMBOL(__set_page_dirty_nobuffers); /* * When a writepage implementation decides that it doesn't want to write this * page for some reason, it should redirty the locked page via * redirty_page_for_writepage() and it should then unlock the page and return 0 */ int redirty_page_for_writepage(struct writeback_control *wbc, struct page *page) { wbc->pages_skipped++; return __set_page_dirty_nobuffers(page); } EXPORT_SYMBOL(redirty_page_for_writepage); /* |
6746aff74 HWPOISON: shmem: ... |
1152 1153 1154 1155 1156 1157 1158 |
* Dirty a page. * * For pages with a mapping this should be done under the page lock * for the benefit of asynchronous memory errors who prefer a consistent * dirty state. This rule can be broken in some special cases, * but should be better not to. * |
1da177e4c Linux-2.6.12-rc2 |
1159 1160 1161 |
* If the mapping doesn't provide a set_page_dirty a_op, then * just fall through and assume that it wants buffer_heads. */ |
1cf6e7d83 mm: task dirty ac... |
1162 |
int set_page_dirty(struct page *page) |
1da177e4c Linux-2.6.12-rc2 |
1163 1164 1165 1166 1167 |
{ struct address_space *mapping = page_mapping(page); if (likely(mapping)) { int (*spd)(struct page *) = mapping->a_ops->set_page_dirty; |
9361401eb [PATCH] BLOCK: Ma... |
1168 1169 1170 1171 1172 |
#ifdef CONFIG_BLOCK if (!spd) spd = __set_page_dirty_buffers; #endif return (*spd)(page); |
1da177e4c Linux-2.6.12-rc2 |
1173 |
} |
4741c9fd3 [PATCH] set_page_... |
1174 1175 1176 1177 |
if (!PageDirty(page)) { if (!TestSetPageDirty(page)) return 1; } |
1da177e4c Linux-2.6.12-rc2 |
1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 |
return 0; } EXPORT_SYMBOL(set_page_dirty); /* * set_page_dirty() is racy if the caller has no reference against * page->mapping->host, and if the page is unlocked. This is because another * CPU could truncate the page off the mapping and then free the mapping. * * Usually, the page _is_ locked, or the caller is a user-space process which * holds a reference on the inode by having an open file. * * In other cases, the page should be locked before running set_page_dirty(). */ int set_page_dirty_lock(struct page *page) { int ret; |
db37648cd [PATCH] mm: non s... |
1195 |
lock_page_nosync(page); |
1da177e4c Linux-2.6.12-rc2 |
1196 1197 1198 1199 1200 1201 1202 |
ret = set_page_dirty(page); unlock_page(page); return ret; } EXPORT_SYMBOL(set_page_dirty_lock); /* |
1da177e4c Linux-2.6.12-rc2 |
1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 |
* Clear a page's dirty flag, while caring for dirty memory accounting. * Returns true if the page was previously dirty. * * This is for preparing to put the page under writeout. We leave the page * tagged as dirty in the radix tree so that a concurrent write-for-sync * can discover it via a PAGECACHE_TAG_DIRTY walk. The ->writepage * implementation will run either set_page_writeback() or set_page_dirty(), * at which stage we bring the page's dirty flag and radix-tree dirty tag * back into sync. * * This incoherency between the page's dirty flag and radix-tree tag is * unfortunate, but it only exists while the page is locked. */ int clear_page_dirty_for_io(struct page *page) { struct address_space *mapping = page_mapping(page); |
79352894b mm: fix clear_pag... |
1219 |
BUG_ON(!PageLocked(page)); |
fe3cba17c mm: share PG_read... |
1220 |
ClearPageReclaim(page); |
7658cc289 VM: Fix nasty and... |
1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 |
if (mapping && mapping_cap_account_dirty(mapping)) { /* * Yes, Virginia, this is indeed insane. * * We use this sequence to make sure that * (a) we account for dirty stats properly * (b) we tell the low-level filesystem to * mark the whole page dirty if it was * dirty in a pagetable. Only to then * (c) clean the page again and return 1 to * cause the writeback. * * This way we avoid all nasty races with the * dirty bit in multiple places and clearing * them concurrently from different threads. * * Note! Normally the "set_page_dirty(page)" * has no effect on the actual dirty bit - since * that will already usually be set. But we * need the side effects, and it can help us * avoid races. * * We basically use the page "master dirty bit" * as a serialization point for all the different * threads doing their things. |
7658cc289 VM: Fix nasty and... |
1246 1247 1248 |
*/ if (page_mkclean(page)) set_page_dirty(page); |
79352894b mm: fix clear_pag... |
1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 |
/* * We carefully synchronise fault handlers against * installing a dirty pte and marking the page dirty * at this point. We do this by having them hold the * page lock at some point after installing their * pte, but before marking the page dirty. * Pages are always locked coming in here, so we get * the desired exclusion. See mm/memory.c:do_wp_page() * for more comments. */ |
7658cc289 VM: Fix nasty and... |
1259 |
if (TestClearPageDirty(page)) { |
8c08540f8 [PATCH] clean up ... |
1260 |
dec_zone_page_state(page, NR_FILE_DIRTY); |
c9e51e418 mm: count reclaim... |
1261 1262 |
dec_bdi_stat(mapping->backing_dev_info, BDI_RECLAIMABLE); |
7658cc289 VM: Fix nasty and... |
1263 |
return 1; |
1da177e4c Linux-2.6.12-rc2 |
1264 |
} |
7658cc289 VM: Fix nasty and... |
1265 |
return 0; |
1da177e4c Linux-2.6.12-rc2 |
1266 |
} |
7658cc289 VM: Fix nasty and... |
1267 |
return TestClearPageDirty(page); |
1da177e4c Linux-2.6.12-rc2 |
1268 |
} |
58bb01a9c [PATCH] re-export... |
1269 |
EXPORT_SYMBOL(clear_page_dirty_for_io); |
1da177e4c Linux-2.6.12-rc2 |
1270 1271 1272 1273 1274 1275 1276 |
int test_clear_page_writeback(struct page *page) { struct address_space *mapping = page_mapping(page); int ret; if (mapping) { |
69cb51d18 mm: count writeba... |
1277 |
struct backing_dev_info *bdi = mapping->backing_dev_info; |
1da177e4c Linux-2.6.12-rc2 |
1278 |
unsigned long flags; |
19fd62312 mm: spinlock tree... |
1279 |
spin_lock_irqsave(&mapping->tree_lock, flags); |
1da177e4c Linux-2.6.12-rc2 |
1280 |
ret = TestClearPageWriteback(page); |
69cb51d18 mm: count writeba... |
1281 |
if (ret) { |
1da177e4c Linux-2.6.12-rc2 |
1282 1283 1284 |
radix_tree_tag_clear(&mapping->page_tree, page_index(page), PAGECACHE_TAG_WRITEBACK); |
e4ad08fe6 mm: bdi: add sepa... |
1285 |
if (bdi_cap_account_writeback(bdi)) { |
69cb51d18 mm: count writeba... |
1286 |
__dec_bdi_stat(bdi, BDI_WRITEBACK); |
04fbfdc14 mm: per device di... |
1287 1288 |
__bdi_writeout_inc(bdi); } |
69cb51d18 mm: count writeba... |
1289 |
} |
19fd62312 mm: spinlock tree... |
1290 |
spin_unlock_irqrestore(&mapping->tree_lock, flags); |
1da177e4c Linux-2.6.12-rc2 |
1291 1292 1293 |
} else { ret = TestClearPageWriteback(page); } |
d688abf50 move page writeba... |
1294 1295 |
if (ret) dec_zone_page_state(page, NR_WRITEBACK); |
1da177e4c Linux-2.6.12-rc2 |
1296 1297 1298 1299 1300 1301 1302 1303 1304 |
return ret; } int test_set_page_writeback(struct page *page) { struct address_space *mapping = page_mapping(page); int ret; if (mapping) { |
69cb51d18 mm: count writeba... |
1305 |
struct backing_dev_info *bdi = mapping->backing_dev_info; |
1da177e4c Linux-2.6.12-rc2 |
1306 |
unsigned long flags; |
19fd62312 mm: spinlock tree... |
1307 |
spin_lock_irqsave(&mapping->tree_lock, flags); |
1da177e4c Linux-2.6.12-rc2 |
1308 |
ret = TestSetPageWriteback(page); |
69cb51d18 mm: count writeba... |
1309 |
if (!ret) { |
1da177e4c Linux-2.6.12-rc2 |
1310 1311 1312 |
radix_tree_tag_set(&mapping->page_tree, page_index(page), PAGECACHE_TAG_WRITEBACK); |
e4ad08fe6 mm: bdi: add sepa... |
1313 |
if (bdi_cap_account_writeback(bdi)) |
69cb51d18 mm: count writeba... |
1314 1315 |
__inc_bdi_stat(bdi, BDI_WRITEBACK); } |
1da177e4c Linux-2.6.12-rc2 |
1316 1317 1318 1319 |
if (!PageDirty(page)) radix_tree_tag_clear(&mapping->page_tree, page_index(page), PAGECACHE_TAG_DIRTY); |
f446daaea mm: implement wri... |
1320 1321 1322 |
radix_tree_tag_clear(&mapping->page_tree, page_index(page), PAGECACHE_TAG_TOWRITE); |
19fd62312 mm: spinlock tree... |
1323 |
spin_unlock_irqrestore(&mapping->tree_lock, flags); |
1da177e4c Linux-2.6.12-rc2 |
1324 1325 1326 |
} else { ret = TestSetPageWriteback(page); } |
d688abf50 move page writeba... |
1327 |
if (!ret) |
f629d1c9b mm: add account_p... |
1328 |
account_page_writeback(page); |
1da177e4c Linux-2.6.12-rc2 |
1329 1330 1331 1332 1333 1334 |
return ret; } EXPORT_SYMBOL(test_set_page_writeback); /* |
001281881 mm: use lockless ... |
1335 |
* Return true if any of the pages in the mapping are marked with the |
1da177e4c Linux-2.6.12-rc2 |
1336 1337 1338 1339 |
* passed tag. */ int mapping_tagged(struct address_space *mapping, int tag) { |
1da177e4c Linux-2.6.12-rc2 |
1340 |
int ret; |
001281881 mm: use lockless ... |
1341 |
rcu_read_lock(); |
1da177e4c Linux-2.6.12-rc2 |
1342 |
ret = radix_tree_tagged(&mapping->page_tree, tag); |
001281881 mm: use lockless ... |
1343 |
rcu_read_unlock(); |
1da177e4c Linux-2.6.12-rc2 |
1344 1345 1346 |
return ret; } EXPORT_SYMBOL(mapping_tagged); |