Blame view
mm/page-writeback.c
45.2 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 |
/* |
ffd1f609a writeback: introd... |
40 41 42 43 44 |
* Sleep at most 200ms at a time in balance_dirty_pages(). */ #define MAX_PAUSE max(HZ/5, 1) /* |
e98be2d59 writeback: bdi wr... |
45 46 47 48 49 |
* Estimate write bandwidth at 200ms intervals. */ #define BANDWIDTH_INTERVAL max(HZ/5, 1) /* |
1da177e4c Linux-2.6.12-rc2 |
50 51 52 53 |
* 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 |
54 55 56 |
/* * 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... |
57 |
* It should be somewhat larger than dirtied pages to ensure that reasonably |
1da177e4c Linux-2.6.12-rc2 |
58 59 |
* large amounts of I/O are submitted. */ |
3a2e9a5a2 writeback: balanc... |
60 |
static inline long sync_writeback_pages(unsigned long dirtied) |
1da177e4c Linux-2.6.12-rc2 |
61 |
{ |
3a2e9a5a2 writeback: balanc... |
62 63 64 65 |
if (dirtied < ratelimit_pages) dirtied = ratelimit_pages; return dirtied + dirtied / 2; |
1da177e4c Linux-2.6.12-rc2 |
66 67 68 69 70 |
} /* The following parameters are exported via /proc/sys/vm */ /* |
5b0830cb9 writeback: get ri... |
71 |
* Start background writeback (via writeback threads) at this percentage |
1da177e4c Linux-2.6.12-rc2 |
72 |
*/ |
1b5e62b42 writeback: double... |
73 |
int dirty_background_ratio = 10; |
1da177e4c Linux-2.6.12-rc2 |
74 75 |
/* |
2da02997e mm: add dirty_bac... |
76 77 78 79 80 81 |
* 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... |
82 83 84 85 86 87 |
* 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 |
88 89 |
* The generator of dirty data starts writeback at this percentage */ |
1b5e62b42 writeback: double... |
90 |
int vm_dirty_ratio = 20; |
1da177e4c Linux-2.6.12-rc2 |
91 92 |
/* |
2da02997e mm: add dirty_bac... |
93 94 95 96 97 98 |
* 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... |
99 |
* The interval between `kupdate'-style writebacks |
1da177e4c Linux-2.6.12-rc2 |
100 |
*/ |
22ef37eed page-writeback: f... |
101 |
unsigned int dirty_writeback_interval = 5 * 100; /* centiseconds */ |
1da177e4c Linux-2.6.12-rc2 |
102 103 |
/* |
704503d83 mm: fix proc_doin... |
104 |
* The longest time for which data is allowed to remain dirty |
1da177e4c Linux-2.6.12-rc2 |
105 |
*/ |
22ef37eed page-writeback: f... |
106 |
unsigned int dirty_expire_interval = 30 * 100; /* centiseconds */ |
1da177e4c Linux-2.6.12-rc2 |
107 108 109 110 111 112 113 |
/* * Flag that makes the machine dump writes/reads and block dirtyings. */ int block_dump; /* |
ed5b43f15 [PATCH] Represent... |
114 115 |
* 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 |
116 117 118 119 120 121 |
*/ int laptop_mode; EXPORT_SYMBOL(laptop_mode); /* End of sysctl-exported parameters */ |
c42843f2f writeback: introd... |
122 |
unsigned long global_dirty_limit; |
1da177e4c Linux-2.6.12-rc2 |
123 |
|
1da177e4c Linux-2.6.12-rc2 |
124 |
/* |
04fbfdc14 mm: per device di... |
125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 |
* 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... |
141 |
static struct prop_descriptor vm_dirties; |
04fbfdc14 mm: per device di... |
142 |
|
04fbfdc14 mm: per device di... |
143 144 145 146 147 148 149 150 |
/* * 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... |
151 152 153 154 155 |
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... |
156 157 158 159 |
return 2 + ilog2(dirty_total - 1); } /* |
2da02997e mm: add dirty_bac... |
160 |
* update the period when the dirty threshold changes. |
04fbfdc14 mm: per device di... |
161 |
*/ |
2da02997e mm: add dirty_bac... |
162 163 164 165 166 167 168 169 |
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... |
170 |
void __user *buffer, size_t *lenp, |
2da02997e mm: add dirty_bac... |
171 172 173 |
loff_t *ppos) { int ret; |
8d65af789 sysctl: remove "s... |
174 |
ret = proc_dointvec_minmax(table, write, buffer, lenp, ppos); |
2da02997e mm: add dirty_bac... |
175 176 177 178 179 180 |
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... |
181 |
void __user *buffer, size_t *lenp, |
2da02997e mm: add dirty_bac... |
182 183 184 |
loff_t *ppos) { int ret; |
8d65af789 sysctl: remove "s... |
185 |
ret = proc_doulongvec_minmax(table, write, buffer, lenp, ppos); |
2da02997e mm: add dirty_bac... |
186 187 188 189 |
if (ret == 0 && write) dirty_background_ratio = 0; return ret; } |
04fbfdc14 mm: per device di... |
190 |
int dirty_ratio_handler(struct ctl_table *table, int write, |
8d65af789 sysctl: remove "s... |
191 |
void __user *buffer, size_t *lenp, |
04fbfdc14 mm: per device di... |
192 193 194 |
loff_t *ppos) { int old_ratio = vm_dirty_ratio; |
2da02997e mm: add dirty_bac... |
195 |
int ret; |
8d65af789 sysctl: remove "s... |
196 |
ret = proc_dointvec_minmax(table, write, buffer, lenp, ppos); |
04fbfdc14 mm: per device di... |
197 |
if (ret == 0 && write && vm_dirty_ratio != old_ratio) { |
2da02997e mm: add dirty_bac... |
198 199 200 201 202 203 204 205 |
update_completion_period(); vm_dirty_bytes = 0; } return ret; } int dirty_bytes_handler(struct ctl_table *table, int write, |
8d65af789 sysctl: remove "s... |
206 |
void __user *buffer, size_t *lenp, |
2da02997e mm: add dirty_bac... |
207 208 |
loff_t *ppos) { |
fc3501d41 mm: fix dirty_byt... |
209 |
unsigned long old_bytes = vm_dirty_bytes; |
2da02997e mm: add dirty_bac... |
210 |
int ret; |
8d65af789 sysctl: remove "s... |
211 |
ret = proc_doulongvec_minmax(table, write, buffer, lenp, ppos); |
2da02997e mm: add dirty_bac... |
212 213 214 |
if (ret == 0 && write && vm_dirty_bytes != old_bytes) { update_completion_period(); vm_dirty_ratio = 0; |
04fbfdc14 mm: per device di... |
215 216 217 218 219 220 221 222 223 224 |
} 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) { |
f7d2b1ecd writeback: accoun... |
225 |
__inc_bdi_stat(bdi, BDI_WRITTEN); |
a42dde041 mm: bdi: allow se... |
226 227 |
__prop_inc_percpu_max(&vm_completions, &bdi->completions, bdi->max_prop_frac); |
04fbfdc14 mm: per device di... |
228 |
} |
dd5656e59 mm: bdi: export b... |
229 230 231 232 233 234 235 236 237 |
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... |
238 |
void task_dirty_inc(struct task_struct *tsk) |
3e26c149c mm: dirty balanci... |
239 240 241 |
{ prop_inc_single(&vm_dirties, &tsk->dirties); } |
04fbfdc14 mm: per device di... |
242 243 244 245 246 247 |
/* * Obtain an accurate fraction of the BDI's portion. */ static void bdi_writeout_fraction(struct backing_dev_info *bdi, long *numerator, long *denominator) { |
3efaf0fab writeback: skip b... |
248 |
prop_fraction_percpu(&vm_completions, &bdi->completions, |
04fbfdc14 mm: per device di... |
249 |
numerator, denominator); |
04fbfdc14 mm: per device di... |
250 |
} |
3e26c149c mm: dirty balanci... |
251 252 253 254 255 256 257 258 |
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... |
259 |
* task_dirty_limit - scale down dirty throttling threshold for one task |
3e26c149c mm: dirty balanci... |
260 261 262 263 |
* * task specific dirty limit: * * dirty -= (dirty/8) * p_{t} |
1babe1838 writeback: add co... |
264 265 266 267 268 269 270 |
* * 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... |
271 |
*/ |
bcff25fc8 mm: properly refl... |
272 |
#define TASK_LIMIT_FRACTION 8 |
16c4042f0 writeback: avoid ... |
273 274 |
static unsigned long task_dirty_limit(struct task_struct *tsk, unsigned long bdi_dirty) |
3e26c149c mm: dirty balanci... |
275 276 |
{ long numerator, denominator; |
16c4042f0 writeback: avoid ... |
277 |
unsigned long dirty = bdi_dirty; |
bcff25fc8 mm: properly refl... |
278 |
u64 inv = dirty / TASK_LIMIT_FRACTION; |
3e26c149c mm: dirty balanci... |
279 280 281 282 283 284 |
task_dirties_fraction(tsk, &numerator, &denominator); inv *= numerator; do_div(inv, denominator); dirty -= inv; |
3e26c149c mm: dirty balanci... |
285 |
|
16c4042f0 writeback: avoid ... |
286 |
return max(dirty, bdi_dirty/2); |
3e26c149c mm: dirty balanci... |
287 |
} |
bcff25fc8 mm: properly refl... |
288 289 290 291 292 |
/* Minimum limit for any task */ static unsigned long task_min_dirty_limit(unsigned long bdi_dirty) { return bdi_dirty - bdi_dirty / TASK_LIMIT_FRACTION; } |
04fbfdc14 mm: per device di... |
293 |
/* |
189d3c4a9 mm: bdi: allow se... |
294 295 |
* */ |
189d3c4a9 mm: bdi: allow se... |
296 297 298 299 300 |
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... |
301 |
|
cfc4ba536 writeback: use RC... |
302 |
spin_lock_bh(&bdi_lock); |
a42dde041 mm: bdi: allow se... |
303 |
if (min_ratio > bdi->max_ratio) { |
189d3c4a9 mm: bdi: allow se... |
304 |
ret = -EINVAL; |
a42dde041 mm: bdi: allow se... |
305 306 307 308 309 310 311 312 313 |
} 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... |
314 |
spin_unlock_bh(&bdi_lock); |
a42dde041 mm: bdi: allow se... |
315 316 317 318 319 320 |
return ret; } int bdi_set_max_ratio(struct backing_dev_info *bdi, unsigned max_ratio) { |
a42dde041 mm: bdi: allow se... |
321 322 323 324 |
int ret = 0; if (max_ratio > 100) return -EINVAL; |
cfc4ba536 writeback: use RC... |
325 |
spin_lock_bh(&bdi_lock); |
a42dde041 mm: bdi: allow se... |
326 327 328 329 330 331 |
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... |
332 |
spin_unlock_bh(&bdi_lock); |
189d3c4a9 mm: bdi: allow se... |
333 334 335 |
return ret; } |
a42dde041 mm: bdi: allow se... |
336 |
EXPORT_SYMBOL(bdi_set_max_ratio); |
189d3c4a9 mm: bdi: allow se... |
337 338 |
/* |
1da177e4c Linux-2.6.12-rc2 |
339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 |
* 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 ... |
355 356 357 358 359 360 |
static unsigned long highmem_dirtyable_memory(unsigned long total) { #ifdef CONFIG_HIGHMEM int node; unsigned long x = 0; |
37b07e416 memoryless nodes:... |
361 |
for_each_node_state(node, N_HIGH_MEMORY) { |
1b4244647 Use ZVC counters ... |
362 363 |
struct zone *z = &NODE_DATA(node)->node_zones[ZONE_HIGHMEM]; |
adea02a1b mm: count only re... |
364 365 |
x += zone_page_state(z, NR_FREE_PAGES) + zone_reclaimable_pages(z); |
1b4244647 Use ZVC counters ... |
366 367 368 369 370 371 372 373 374 375 376 377 |
} /* * 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... |
378 379 380 381 382 383 384 |
/** * 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 ... |
385 386 |
{ unsigned long x; |
adea02a1b mm: count only re... |
387 |
x = global_page_state(NR_FREE_PAGES) + global_reclaimable_pages(); |
195cf453d mm/page-writeback... |
388 389 390 |
if (!vm_highmem_is_dirtyable) x -= highmem_dirtyable_memory(x); |
1b4244647 Use ZVC counters ... |
391 392 |
return x + 1; /* Ensure that we never return 0 */ } |
ffd1f609a writeback: introd... |
393 394 395 396 |
static unsigned long hard_dirty_limit(unsigned long thresh) { return max(thresh, global_dirty_limit); } |
03ab450f0 mm/page-writeback... |
397 |
/* |
1babe1838 writeback: add co... |
398 399 400 401 402 403 |
* 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... |
404 |
* real-time tasks. |
1babe1838 writeback: add co... |
405 |
*/ |
16c4042f0 writeback: avoid ... |
406 |
void global_dirty_limits(unsigned long *pbackground, unsigned long *pdirty) |
1da177e4c Linux-2.6.12-rc2 |
407 |
{ |
364aeb284 mm: change dirty ... |
408 409 |
unsigned long background; unsigned long dirty; |
240c879f2 writeback: avoid ... |
410 |
unsigned long uninitialized_var(available_memory); |
1da177e4c Linux-2.6.12-rc2 |
411 |
struct task_struct *tsk; |
240c879f2 writeback: avoid ... |
412 413 |
if (!vm_dirty_bytes || !dirty_background_bytes) available_memory = determine_dirtyable_memory(); |
2da02997e mm: add dirty_bac... |
414 415 |
if (vm_dirty_bytes) dirty = DIV_ROUND_UP(vm_dirty_bytes, PAGE_SIZE); |
4cbec4c8b writeback: remove... |
416 417 |
else dirty = (vm_dirty_ratio * available_memory) / 100; |
1da177e4c Linux-2.6.12-rc2 |
418 |
|
2da02997e mm: add dirty_bac... |
419 420 421 422 |
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 |
423 |
|
2da02997e mm: add dirty_bac... |
424 425 |
if (background >= dirty) background = dirty / 2; |
1da177e4c Linux-2.6.12-rc2 |
426 427 428 429 430 431 432 |
tsk = current; if (tsk->flags & PF_LESS_THROTTLE || rt_task(tsk)) { background += background / 4; dirty += dirty / 4; } *pbackground = background; *pdirty = dirty; |
e1cbe2360 writeback: trace ... |
433 |
trace_global_dirty_state(background, dirty); |
16c4042f0 writeback: avoid ... |
434 |
} |
04fbfdc14 mm: per device di... |
435 |
|
6f7186562 writeback: add bd... |
436 |
/** |
1babe1838 writeback: add co... |
437 |
* bdi_dirty_limit - @bdi's share of dirty throttling threshold |
6f7186562 writeback: add bd... |
438 439 |
* @bdi: the backing_dev_info to query * @dirty: global dirty limit in pages |
1babe1838 writeback: add co... |
440 |
* |
6f7186562 writeback: add bd... |
441 442 443 444 |
* Returns @bdi's dirty limit in pages. The term "dirty" in the context of * dirty balancing includes all PG_dirty, PG_writeback and NFS unstable pages. * And the "limit" in the name is not seriously taken as hard limit in * balance_dirty_pages(). |
1babe1838 writeback: add co... |
445 |
* |
6f7186562 writeback: add bd... |
446 |
* It allocates high/low dirty limits to fast/slow devices, in order to prevent |
1babe1838 writeback: add co... |
447 448 449 450 451 452 453 |
* - 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 ... |
454 455 456 |
{ u64 bdi_dirty; long numerator, denominator; |
04fbfdc14 mm: per device di... |
457 |
|
16c4042f0 writeback: avoid ... |
458 459 460 461 |
/* * Calculate this BDI's share of the dirty ratio. */ bdi_writeout_fraction(bdi, &numerator, &denominator); |
04fbfdc14 mm: per device di... |
462 |
|
16c4042f0 writeback: avoid ... |
463 464 465 |
bdi_dirty = (dirty * (100 - bdi_min_ratio)) / 100; bdi_dirty *= numerator; do_div(bdi_dirty, denominator); |
04fbfdc14 mm: per device di... |
466 |
|
16c4042f0 writeback: avoid ... |
467 468 469 470 471 |
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 |
472 |
} |
e98be2d59 writeback: bdi wr... |
473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 |
static void bdi_update_write_bandwidth(struct backing_dev_info *bdi, unsigned long elapsed, unsigned long written) { const unsigned long period = roundup_pow_of_two(3 * HZ); unsigned long avg = bdi->avg_write_bandwidth; unsigned long old = bdi->write_bandwidth; u64 bw; /* * bw = written * HZ / elapsed * * bw * elapsed + write_bandwidth * (period - elapsed) * write_bandwidth = --------------------------------------------------- * period */ bw = written - bdi->written_stamp; bw *= HZ; if (unlikely(elapsed > period)) { do_div(bw, elapsed); avg = bw; goto out; } bw += (u64)bdi->write_bandwidth * (period - elapsed); bw >>= ilog2(period); /* * one more level of smoothing, for filtering out sudden spikes */ if (avg > old && old >= (unsigned long)bw) avg -= (avg - old) >> 3; if (avg < old && old <= (unsigned long)bw) avg += (old - avg) >> 3; out: bdi->write_bandwidth = bw; bdi->avg_write_bandwidth = avg; } |
c42843f2f writeback: introd... |
512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 |
/* * The global dirtyable memory and dirty threshold could be suddenly knocked * down by a large amount (eg. on the startup of KVM in a swapless system). * This may throw the system into deep dirty exceeded state and throttle * heavy/light dirtiers alike. To retain good responsiveness, maintain * global_dirty_limit for tracking slowly down to the knocked down dirty * threshold. */ static void update_dirty_limit(unsigned long thresh, unsigned long dirty) { unsigned long limit = global_dirty_limit; /* * Follow up in one step. */ if (limit < thresh) { limit = thresh; goto update; } /* * Follow down slowly. Use the higher one as the target, because thresh * may drop below dirty. This is exactly the reason to introduce * global_dirty_limit which is guaranteed to lie above the dirty pages. */ thresh = max(thresh, dirty); if (limit > thresh) { limit -= (limit - thresh) >> 5; goto update; } return; update: global_dirty_limit = limit; } static void global_update_bandwidth(unsigned long thresh, unsigned long dirty, unsigned long now) { static DEFINE_SPINLOCK(dirty_lock); static unsigned long update_time; /* * check locklessly first to optimize away locking for the most time */ if (time_before(now, update_time + BANDWIDTH_INTERVAL)) return; spin_lock(&dirty_lock); if (time_after_eq(now, update_time + BANDWIDTH_INTERVAL)) { update_dirty_limit(thresh, dirty); update_time = now; } spin_unlock(&dirty_lock); } |
e98be2d59 writeback: bdi wr... |
567 |
void __bdi_update_bandwidth(struct backing_dev_info *bdi, |
c42843f2f writeback: introd... |
568 569 570 571 |
unsigned long thresh, unsigned long dirty, unsigned long bdi_thresh, unsigned long bdi_dirty, |
e98be2d59 writeback: bdi wr... |
572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 |
unsigned long start_time) { unsigned long now = jiffies; unsigned long elapsed = now - bdi->bw_time_stamp; unsigned long written; /* * rate-limit, only update once every 200ms. */ if (elapsed < BANDWIDTH_INTERVAL) return; written = percpu_counter_read(&bdi->bdi_stat[BDI_WRITTEN]); /* * Skip quiet periods when disk bandwidth is under-utilized. * (at least 1s idle time between two flusher runs) */ if (elapsed > HZ && time_before(bdi->bw_time_stamp, start_time)) goto snapshot; |
c42843f2f writeback: introd... |
592 593 |
if (thresh) global_update_bandwidth(thresh, dirty, now); |
e98be2d59 writeback: bdi wr... |
594 595 596 597 598 599 600 601 |
bdi_update_write_bandwidth(bdi, elapsed, written); snapshot: bdi->written_stamp = written; bdi->bw_time_stamp = now; } static void bdi_update_bandwidth(struct backing_dev_info *bdi, |
c42843f2f writeback: introd... |
602 603 604 605 |
unsigned long thresh, unsigned long dirty, unsigned long bdi_thresh, unsigned long bdi_dirty, |
e98be2d59 writeback: bdi wr... |
606 607 608 609 610 |
unsigned long start_time) { if (time_is_after_eq_jiffies(bdi->bw_time_stamp + BANDWIDTH_INTERVAL)) return; spin_lock(&bdi->wb.list_lock); |
c42843f2f writeback: introd... |
611 612 |
__bdi_update_bandwidth(bdi, thresh, dirty, bdi_thresh, bdi_dirty, start_time); |
e98be2d59 writeback: bdi wr... |
613 614 |
spin_unlock(&bdi->wb.list_lock); } |
1da177e4c Linux-2.6.12-rc2 |
615 616 617 618 |
/* * 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... |
619 620 |
* If we're over `background_thresh' then the writeback threads are woken to * perform some writeout. |
1da177e4c Linux-2.6.12-rc2 |
621 |
*/ |
3a2e9a5a2 writeback: balanc... |
622 623 |
static void balance_dirty_pages(struct address_space *mapping, unsigned long write_chunk) |
1da177e4c Linux-2.6.12-rc2 |
624 |
{ |
7762741e3 writeback: consol... |
625 626 627 |
unsigned long nr_reclaimable, bdi_nr_reclaimable; unsigned long nr_dirty; /* = file_dirty + writeback + unstable_nfs */ unsigned long bdi_dirty; |
364aeb284 mm: change dirty ... |
628 629 630 |
unsigned long background_thresh; unsigned long dirty_thresh; unsigned long bdi_thresh; |
bcff25fc8 mm: properly refl... |
631 632 |
unsigned long task_bdi_thresh; unsigned long min_task_bdi_thresh; |
1da177e4c Linux-2.6.12-rc2 |
633 |
unsigned long pages_written = 0; |
87c6a9b25 writeback: make b... |
634 |
unsigned long pause = 1; |
e50e37201 writeback: balanc... |
635 |
bool dirty_exceeded = false; |
bcff25fc8 mm: properly refl... |
636 |
bool clear_dirty_exceeded = true; |
1da177e4c Linux-2.6.12-rc2 |
637 |
struct backing_dev_info *bdi = mapping->backing_dev_info; |
e98be2d59 writeback: bdi wr... |
638 |
unsigned long start_time = jiffies; |
1da177e4c Linux-2.6.12-rc2 |
639 640 |
for (;;) { |
5fce25a9d mm: speed up writ... |
641 642 |
nr_reclaimable = global_page_state(NR_FILE_DIRTY) + global_page_state(NR_UNSTABLE_NFS); |
7762741e3 writeback: consol... |
643 |
nr_dirty = nr_reclaimable + global_page_state(NR_WRITEBACK); |
5fce25a9d mm: speed up writ... |
644 |
|
16c4042f0 writeback: avoid ... |
645 646 647 648 649 650 651 |
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. */ |
7762741e3 writeback: consol... |
652 |
if (nr_dirty <= (background_thresh + dirty_thresh) / 2) |
16c4042f0 writeback: avoid ... |
653 654 655 |
break; bdi_thresh = bdi_dirty_limit(bdi, dirty_thresh); |
bcff25fc8 mm: properly refl... |
656 657 |
min_task_bdi_thresh = task_min_dirty_limit(bdi_thresh); task_bdi_thresh = task_dirty_limit(current, bdi_thresh); |
16c4042f0 writeback: avoid ... |
658 |
|
e50e37201 writeback: balanc... |
659 660 661 662 663 664 665 666 667 668 |
/* * 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. */ |
bcff25fc8 mm: properly refl... |
669 |
if (task_bdi_thresh < 2 * bdi_stat_error(bdi)) { |
e50e37201 writeback: balanc... |
670 |
bdi_nr_reclaimable = bdi_stat_sum(bdi, BDI_RECLAIMABLE); |
7762741e3 writeback: consol... |
671 672 |
bdi_dirty = bdi_nr_reclaimable + bdi_stat_sum(bdi, BDI_WRITEBACK); |
e50e37201 writeback: balanc... |
673 674 |
} else { bdi_nr_reclaimable = bdi_stat(bdi, BDI_RECLAIMABLE); |
7762741e3 writeback: consol... |
675 676 |
bdi_dirty = bdi_nr_reclaimable + bdi_stat(bdi, BDI_WRITEBACK); |
e50e37201 writeback: balanc... |
677 |
} |
5fce25a9d mm: speed up writ... |
678 |
|
e50e37201 writeback: balanc... |
679 680 681 682 683 684 |
/* * 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. */ |
bcff25fc8 mm: properly refl... |
685 |
dirty_exceeded = (bdi_dirty > task_bdi_thresh) || |
7762741e3 writeback: consol... |
686 |
(nr_dirty > dirty_thresh); |
bcff25fc8 mm: properly refl... |
687 688 |
clear_dirty_exceeded = (bdi_dirty <= min_task_bdi_thresh) && (nr_dirty <= dirty_thresh); |
e50e37201 writeback: balanc... |
689 690 |
if (!dirty_exceeded) |
04fbfdc14 mm: per device di... |
691 |
break; |
1da177e4c Linux-2.6.12-rc2 |
692 |
|
04fbfdc14 mm: per device di... |
693 694 |
if (!bdi->dirty_exceeded) bdi->dirty_exceeded = 1; |
1da177e4c Linux-2.6.12-rc2 |
695 |
|
c42843f2f writeback: introd... |
696 697 |
bdi_update_bandwidth(bdi, dirty_thresh, nr_dirty, bdi_thresh, bdi_dirty, start_time); |
e98be2d59 writeback: bdi wr... |
698 |
|
1da177e4c Linux-2.6.12-rc2 |
699 700 701 702 703 |
/* 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... |
704 705 706 |
* 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 |
707 |
*/ |
d46db3d58 writeback: make w... |
708 |
trace_balance_dirty_start(bdi); |
bcff25fc8 mm: properly refl... |
709 |
if (bdi_nr_reclaimable > task_bdi_thresh) { |
d46db3d58 writeback: make w... |
710 711 712 |
pages_written += writeback_inodes_wb(&bdi->wb, write_chunk); trace_balance_dirty_written(bdi, pages_written); |
e50e37201 writeback: balanc... |
713 714 |
if (pages_written >= write_chunk) break; /* We've done our duty */ |
04fbfdc14 mm: per device di... |
715 |
} |
d153ba644 writeback: do uni... |
716 |
__set_current_state(TASK_UNINTERRUPTIBLE); |
d25105e89 writeback: accoun... |
717 |
io_schedule_timeout(pause); |
d46db3d58 writeback: make w... |
718 |
trace_balance_dirty_wait(bdi); |
87c6a9b25 writeback: make b... |
719 |
|
ffd1f609a writeback: introd... |
720 721 722 723 724 725 726 |
dirty_thresh = hard_dirty_limit(dirty_thresh); /* * max-pause area. If dirty exceeded but still within this * area, no need to sleep for more than 200ms: (a) 8 pages per * 200ms is typically more than enough to curb heavy dirtiers; * (b) the pause time limit makes the dirtiers more responsive. */ |
bb0822954 squeeze max-pause... |
727 728 |
if (nr_dirty < dirty_thresh && bdi_dirty < (task_bdi_thresh + bdi_thresh) / 2 && |
ffd1f609a writeback: introd... |
729 730 |
time_after(jiffies, start_time + MAX_PAUSE)) break; |
87c6a9b25 writeback: make b... |
731 732 733 734 735 736 737 738 |
/* * 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 |
739 |
} |
bcff25fc8 mm: properly refl... |
740 741 |
/* Clear dirty_exceeded flag only when no task can exceed the limit */ if (clear_dirty_exceeded && bdi->dirty_exceeded) |
04fbfdc14 mm: per device di... |
742 |
bdi->dirty_exceeded = 0; |
1da177e4c Linux-2.6.12-rc2 |
743 744 |
if (writeback_in_progress(bdi)) |
5b0830cb9 writeback: get ri... |
745 |
return; |
1da177e4c Linux-2.6.12-rc2 |
746 747 748 749 750 751 752 753 754 755 |
/* * 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... |
756 |
(!laptop_mode && (nr_reclaimable > background_thresh))) |
c5444198c writeback: simpli... |
757 |
bdi_start_background_writeback(bdi); |
1da177e4c Linux-2.6.12-rc2 |
758 |
} |
a200ee182 mm: set_page_dirt... |
759 |
void set_page_dirty_balance(struct page *page, int page_mkwrite) |
edc79b2a4 [PATCH] mm: balan... |
760 |
{ |
a200ee182 mm: set_page_dirt... |
761 |
if (set_page_dirty(page) || page_mkwrite) { |
edc79b2a4 [PATCH] mm: balan... |
762 763 764 765 766 767 |
struct address_space *mapping = page_mapping(page); if (mapping) balance_dirty_pages_ratelimited(mapping); } } |
245b2e70e percpu: clean up ... |
768 |
static DEFINE_PER_CPU(unsigned long, bdp_ratelimits) = 0; |
1da177e4c Linux-2.6.12-rc2 |
769 |
/** |
fa5a734e4 [PATCH] balance_d... |
770 |
* balance_dirty_pages_ratelimited_nr - balance dirty memory state |
67be2dd1b [PATCH] DocBook: ... |
771 |
* @mapping: address_space which was dirtied |
a580290c3 Documentation: fi... |
772 |
* @nr_pages_dirtied: number of pages which the caller has just dirtied |
1da177e4c Linux-2.6.12-rc2 |
773 774 775 776 777 778 779 780 781 782 |
* * 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... |
783 784 |
void balance_dirty_pages_ratelimited_nr(struct address_space *mapping, unsigned long nr_pages_dirtied) |
1da177e4c Linux-2.6.12-rc2 |
785 |
{ |
36715cef0 writeback: skip t... |
786 |
struct backing_dev_info *bdi = mapping->backing_dev_info; |
fa5a734e4 [PATCH] balance_d... |
787 788 |
unsigned long ratelimit; unsigned long *p; |
1da177e4c Linux-2.6.12-rc2 |
789 |
|
36715cef0 writeback: skip t... |
790 791 |
if (!bdi_cap_account_dirty(bdi)) return; |
1da177e4c Linux-2.6.12-rc2 |
792 |
ratelimit = ratelimit_pages; |
04fbfdc14 mm: per device di... |
793 |
if (mapping->backing_dev_info->dirty_exceeded) |
1da177e4c Linux-2.6.12-rc2 |
794 795 796 797 798 799 |
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... |
800 |
preempt_disable(); |
245b2e70e percpu: clean up ... |
801 |
p = &__get_cpu_var(bdp_ratelimits); |
fa5a734e4 [PATCH] balance_d... |
802 803 |
*p += nr_pages_dirtied; if (unlikely(*p >= ratelimit)) { |
3a2e9a5a2 writeback: balanc... |
804 |
ratelimit = sync_writeback_pages(*p); |
fa5a734e4 [PATCH] balance_d... |
805 806 |
*p = 0; preempt_enable(); |
3a2e9a5a2 writeback: balanc... |
807 |
balance_dirty_pages(mapping, ratelimit); |
1da177e4c Linux-2.6.12-rc2 |
808 809 |
return; } |
fa5a734e4 [PATCH] balance_d... |
810 |
preempt_enable(); |
1da177e4c Linux-2.6.12-rc2 |
811 |
} |
fa5a734e4 [PATCH] balance_d... |
812 |
EXPORT_SYMBOL(balance_dirty_pages_ratelimited_nr); |
1da177e4c Linux-2.6.12-rc2 |
813 |
|
232ea4d69 [PATCH] throttle_... |
814 |
void throttle_vm_writeout(gfp_t gfp_mask) |
1da177e4c Linux-2.6.12-rc2 |
815 |
{ |
364aeb284 mm: change dirty ... |
816 817 |
unsigned long background_thresh; unsigned long dirty_thresh; |
1da177e4c Linux-2.6.12-rc2 |
818 819 |
for ( ; ; ) { |
16c4042f0 writeback: avoid ... |
820 |
global_dirty_limits(&background_thresh, &dirty_thresh); |
1da177e4c Linux-2.6.12-rc2 |
821 822 823 824 825 826 |
/* * 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 ... |
827 828 829 |
if (global_page_state(NR_UNSTABLE_NFS) + global_page_state(NR_WRITEBACK) <= dirty_thresh) break; |
8aa7e847d Fix congestion_wa... |
830 |
congestion_wait(BLK_RW_ASYNC, HZ/10); |
369f2389e writeback: remove... |
831 832 833 834 835 836 837 838 |
/* * 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 |
839 840 |
} } |
1da177e4c Linux-2.6.12-rc2 |
841 |
/* |
1da177e4c Linux-2.6.12-rc2 |
842 843 844 |
* sysctl handler for /proc/sys/vm/dirty_writeback_centisecs */ int dirty_writeback_centisecs_handler(ctl_table *table, int write, |
8d65af789 sysctl: remove "s... |
845 |
void __user *buffer, size_t *length, loff_t *ppos) |
1da177e4c Linux-2.6.12-rc2 |
846 |
{ |
8d65af789 sysctl: remove "s... |
847 |
proc_dointvec(table, write, buffer, length, ppos); |
6423104b6 writeback: fixups... |
848 |
bdi_arm_supers_timer(); |
1da177e4c Linux-2.6.12-rc2 |
849 850 |
return 0; } |
c2c4986ed writeback: fix pr... |
851 |
#ifdef CONFIG_BLOCK |
31373d09d laptop-mode: Make... |
852 |
void laptop_mode_timer_fn(unsigned long data) |
1da177e4c Linux-2.6.12-rc2 |
853 |
{ |
31373d09d laptop-mode: Make... |
854 855 856 |
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 |
857 |
|
31373d09d laptop-mode: Make... |
858 859 860 861 |
/* * We want to write everything out, not just down to the dirty * threshold */ |
31373d09d laptop-mode: Make... |
862 |
if (bdi_has_dirty_io(&q->backing_dev_info)) |
c5444198c writeback: simpli... |
863 |
bdi_start_writeback(&q->backing_dev_info, nr_pages); |
1da177e4c Linux-2.6.12-rc2 |
864 865 866 867 868 869 870 |
} /* * 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... |
871 |
void laptop_io_completion(struct backing_dev_info *info) |
1da177e4c Linux-2.6.12-rc2 |
872 |
{ |
31373d09d laptop-mode: Make... |
873 |
mod_timer(&info->laptop_mode_wb_timer, jiffies + laptop_mode); |
1da177e4c Linux-2.6.12-rc2 |
874 875 876 877 878 879 880 881 882 |
} /* * 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... |
883 884 885 886 887 888 889 890 |
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 |
891 |
} |
c2c4986ed writeback: fix pr... |
892 |
#endif |
1da177e4c Linux-2.6.12-rc2 |
893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 |
/* * 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... |
910 |
void writeback_set_ratelimit(void) |
1da177e4c Linux-2.6.12-rc2 |
911 |
{ |
40c99aae2 [PATCH] remove st... |
912 |
ratelimit_pages = vm_total_pages / (num_online_cpus() * 32); |
1da177e4c Linux-2.6.12-rc2 |
913 914 915 916 917 |
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... |
918 |
static int __cpuinit |
1da177e4c Linux-2.6.12-rc2 |
919 920 |
ratelimit_handler(struct notifier_block *self, unsigned long u, void *v) { |
2d1d43f6a [PATCH] call mm/p... |
921 |
writeback_set_ratelimit(); |
aa0f03037 [PATCH] Change co... |
922 |
return NOTIFY_DONE; |
1da177e4c Linux-2.6.12-rc2 |
923 |
} |
74b85f379 [PATCH] cpu hotpl... |
924 |
static struct notifier_block __cpuinitdata ratelimit_nb = { |
1da177e4c Linux-2.6.12-rc2 |
925 926 927 928 929 |
.notifier_call = ratelimit_handler, .next = NULL, }; /* |
dc6e29da9 Fix balance_dirty... |
930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 |
* 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 |
946 947 948 |
*/ void __init page_writeback_init(void) { |
04fbfdc14 mm: per device di... |
949 |
int shift; |
2d1d43f6a [PATCH] call mm/p... |
950 |
writeback_set_ratelimit(); |
1da177e4c Linux-2.6.12-rc2 |
951 |
register_cpu_notifier(&ratelimit_nb); |
04fbfdc14 mm: per device di... |
952 953 954 |
shift = calc_period_shift(); prop_descriptor_init(&vm_completions, shift); |
3e26c149c mm: dirty balanci... |
955 |
prop_descriptor_init(&vm_dirties, shift); |
1da177e4c Linux-2.6.12-rc2 |
956 |
} |
811d736f9 [PATCH] BLOCK: Di... |
957 |
/** |
f446daaea mm: implement wri... |
958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 |
* 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... |
974 975 976 |
void tag_pages_for_writeback(struct address_space *mapping, pgoff_t start, pgoff_t end) { |
3c111a071 mm: fix fatal ker... |
977 |
#define WRITEBACK_TAG_BATCH 4096 |
f446daaea mm: implement wri... |
978 979 980 981 982 983 984 985 986 987 |
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:... |
988 989 |
/* We check 'start' to handle wrapping when end == ~0UL */ } while (tagged >= WRITEBACK_TAG_BATCH && start); |
f446daaea mm: implement wri... |
990 991 992 993 |
} EXPORT_SYMBOL(tag_pages_for_writeback); /** |
0ea971801 consolidate gener... |
994 |
* write_cache_pages - walk the list of dirty pages of the given address space and write all of them. |
811d736f9 [PATCH] BLOCK: Di... |
995 996 |
* @mapping: address space structure to write * @wbc: subtract the number of written pages from *@wbc->nr_to_write |
0ea971801 consolidate gener... |
997 998 |
* @writepage: function called for each page * @data: data passed to writepage function |
811d736f9 [PATCH] BLOCK: Di... |
999 |
* |
0ea971801 consolidate gener... |
1000 |
* If a page is already under I/O, write_cache_pages() skips it, even |
811d736f9 [PATCH] BLOCK: Di... |
1001 1002 1003 1004 1005 1006 |
* 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... |
1007 1008 1009 1010 1011 1012 1013 |
* * 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... |
1014 |
*/ |
0ea971801 consolidate gener... |
1015 1016 1017 |
int write_cache_pages(struct address_space *mapping, struct writeback_control *wbc, writepage_t writepage, void *data) |
811d736f9 [PATCH] BLOCK: Di... |
1018 |
{ |
811d736f9 [PATCH] BLOCK: Di... |
1019 1020 |
int ret = 0; int done = 0; |
811d736f9 [PATCH] BLOCK: Di... |
1021 1022 |
struct pagevec pvec; int nr_pages; |
31a12666d mm: write_cache_p... |
1023 |
pgoff_t uninitialized_var(writeback_index); |
811d736f9 [PATCH] BLOCK: Di... |
1024 1025 |
pgoff_t index; pgoff_t end; /* Inclusive */ |
bd19e012f mm: write_cache_p... |
1026 |
pgoff_t done_index; |
31a12666d mm: write_cache_p... |
1027 |
int cycled; |
811d736f9 [PATCH] BLOCK: Di... |
1028 |
int range_whole = 0; |
f446daaea mm: implement wri... |
1029 |
int tag; |
811d736f9 [PATCH] BLOCK: Di... |
1030 |
|
811d736f9 [PATCH] BLOCK: Di... |
1031 1032 |
pagevec_init(&pvec, 0); if (wbc->range_cyclic) { |
31a12666d mm: write_cache_p... |
1033 1034 1035 1036 1037 1038 |
writeback_index = mapping->writeback_index; /* prev offset */ index = writeback_index; if (index == 0) cycled = 1; else cycled = 0; |
811d736f9 [PATCH] BLOCK: Di... |
1039 1040 1041 1042 1043 1044 |
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... |
1045 |
cycled = 1; /* ignore range_cyclic tests */ |
811d736f9 [PATCH] BLOCK: Di... |
1046 |
} |
6e6938b6d writeback: introd... |
1047 |
if (wbc->sync_mode == WB_SYNC_ALL || wbc->tagged_writepages) |
f446daaea mm: implement wri... |
1048 1049 1050 |
tag = PAGECACHE_TAG_TOWRITE; else tag = PAGECACHE_TAG_DIRTY; |
811d736f9 [PATCH] BLOCK: Di... |
1051 |
retry: |
6e6938b6d writeback: introd... |
1052 |
if (wbc->sync_mode == WB_SYNC_ALL || wbc->tagged_writepages) |
f446daaea mm: implement wri... |
1053 |
tag_pages_for_writeback(mapping, index, end); |
bd19e012f mm: write_cache_p... |
1054 |
done_index = index; |
5a3d5c981 mm: write_cache_p... |
1055 1056 |
while (!done && (index <= end)) { int i; |
f446daaea mm: implement wri... |
1057 |
nr_pages = pagevec_lookup_tag(&pvec, mapping, &index, tag, |
5a3d5c981 mm: write_cache_p... |
1058 1059 1060 |
min(end - index, (pgoff_t)PAGEVEC_SIZE-1) + 1); if (nr_pages == 0) break; |
811d736f9 [PATCH] BLOCK: Di... |
1061 |
|
811d736f9 [PATCH] BLOCK: Di... |
1062 1063 1064 1065 |
for (i = 0; i < nr_pages; i++) { struct page *page = pvec.pages[i]; /* |
d5482cdf8 mm: write_cache_p... |
1066 1067 1068 1069 1070 |
* 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... |
1071 |
*/ |
d5482cdf8 mm: write_cache_p... |
1072 1073 1074 1075 1076 1077 1078 1079 |
if (page->index > end) { /* * can't be range_cyclic (1st pass) because * end == -1 in that case. */ done = 1; break; } |
cf15b07cf writeback: make m... |
1080 |
done_index = page->index; |
d5482cdf8 mm: write_cache_p... |
1081 |
|
811d736f9 [PATCH] BLOCK: Di... |
1082 |
lock_page(page); |
5a3d5c981 mm: write_cache_p... |
1083 1084 1085 1086 1087 1088 1089 1090 |
/* * 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... |
1091 |
if (unlikely(page->mapping != mapping)) { |
5a3d5c981 mm: write_cache_p... |
1092 |
continue_unlock: |
811d736f9 [PATCH] BLOCK: Di... |
1093 1094 1095 |
unlock_page(page); continue; } |
515f4a037 mm: write_cache_p... |
1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 |
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... |
1107 |
|
515f4a037 mm: write_cache_p... |
1108 1109 |
BUG_ON(PageWriteback(page)); if (!clear_page_dirty_for_io(page)) |
5a3d5c981 mm: write_cache_p... |
1110 |
goto continue_unlock; |
811d736f9 [PATCH] BLOCK: Di... |
1111 |
|
9e094383b writeback: Add tr... |
1112 |
trace_wbc_writepage(wbc, mapping->backing_dev_info); |
0ea971801 consolidate gener... |
1113 |
ret = (*writepage)(page, wbc, data); |
00266770b mm: write_cache_p... |
1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 |
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). */ |
cf15b07cf writeback: make m... |
1128 |
done_index = page->index + 1; |
00266770b mm: write_cache_p... |
1129 1130 1131 |
done = 1; break; } |
0b5649278 writeback: pay at... |
1132 |
} |
00266770b mm: write_cache_p... |
1133 |
|
546a19242 writeback: write_... |
1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 |
/* * 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... |
1144 |
} |
811d736f9 [PATCH] BLOCK: Di... |
1145 1146 1147 1148 |
} pagevec_release(&pvec); cond_resched(); } |
3a4c6800f Fix page writebac... |
1149 |
if (!cycled && !done) { |
811d736f9 [PATCH] BLOCK: Di... |
1150 |
/* |
31a12666d mm: write_cache_p... |
1151 |
* range_cyclic: |
811d736f9 [PATCH] BLOCK: Di... |
1152 1153 1154 |
* 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... |
1155 |
cycled = 1; |
811d736f9 [PATCH] BLOCK: Di... |
1156 |
index = 0; |
31a12666d mm: write_cache_p... |
1157 |
end = writeback_index - 1; |
811d736f9 [PATCH] BLOCK: Di... |
1158 1159 |
goto retry; } |
0b5649278 writeback: pay at... |
1160 1161 |
if (wbc->range_cyclic || (range_whole && wbc->nr_to_write > 0)) mapping->writeback_index = done_index; |
06d6cf695 mm: Add range_con... |
1162 |
|
811d736f9 [PATCH] BLOCK: Di... |
1163 1164 |
return ret; } |
0ea971801 consolidate gener... |
1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 |
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) { |
9b6096a65 mm: make generic_... |
1191 1192 |
struct blk_plug plug; int ret; |
0ea971801 consolidate gener... |
1193 1194 1195 |
/* deal with chardevs and other special file */ if (!mapping->a_ops->writepage) return 0; |
9b6096a65 mm: make generic_... |
1196 1197 1198 1199 |
blk_start_plug(&plug); ret = write_cache_pages(mapping, wbc, __writepage, mapping); blk_finish_plug(&plug); return ret; |
0ea971801 consolidate gener... |
1200 |
} |
811d736f9 [PATCH] BLOCK: Di... |
1201 1202 |
EXPORT_SYMBOL(generic_writepages); |
1da177e4c Linux-2.6.12-rc2 |
1203 1204 |
int do_writepages(struct address_space *mapping, struct writeback_control *wbc) { |
22905f775 identify multipag... |
1205 |
int ret; |
1da177e4c Linux-2.6.12-rc2 |
1206 1207 1208 |
if (wbc->nr_to_write <= 0) return 0; if (mapping->a_ops->writepages) |
d08b3851d [PATCH] mm: track... |
1209 |
ret = mapping->a_ops->writepages(mapping, wbc); |
22905f775 identify multipag... |
1210 1211 |
else ret = generic_writepages(mapping, wbc); |
22905f775 identify multipag... |
1212 |
return ret; |
1da177e4c Linux-2.6.12-rc2 |
1213 1214 1215 1216 |
} /** * write_one_page - write out a single page and optionally wait on I/O |
67be2dd1b [PATCH] DocBook: ... |
1217 1218 |
* @page: the page to write * @wait: if true, wait on writeout |
1da177e4c Linux-2.6.12-rc2 |
1219 1220 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 1246 1247 1248 1249 1250 1251 1252 1253 1254 |
* * 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 ... |
1255 1256 1257 1258 1259 |
* 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... |
1260 |
return !TestSetPageDirty(page); |
767193253 [PATCH] simplify ... |
1261 1262 1263 1264 |
return 0; } /* |
e3a7cca1e vfs: add/use acco... |
1265 1266 1267 1268 1269 1270 1271 |
* 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... |
1272 |
__inc_zone_page_state(page, NR_DIRTIED); |
e3a7cca1e vfs: add/use acco... |
1273 1274 1275 1276 1277 |
__inc_bdi_stat(mapping->backing_dev_info, BDI_RECLAIMABLE); task_dirty_inc(current); task_io_account_write(PAGE_CACHE_SIZE); } } |
679ceace8 mm: exporting acc... |
1278 |
EXPORT_SYMBOL(account_page_dirtied); |
e3a7cca1e vfs: add/use acco... |
1279 1280 |
/* |
f629d1c9b mm: add account_p... |
1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 |
* 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); } EXPORT_SYMBOL(account_page_writeback); /* |
1da177e4c Linux-2.6.12-rc2 |
1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 |
* 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/ |
1304 |
* mapping by re-checking page_mapping() inside tree_lock. |
1da177e4c Linux-2.6.12-rc2 |
1305 1306 1307 |
*/ int __set_page_dirty_nobuffers(struct page *page) { |
1da177e4c Linux-2.6.12-rc2 |
1308 1309 1310 |
if (!TestSetPageDirty(page)) { struct address_space *mapping = page_mapping(page); struct address_space *mapping2; |
8c08540f8 [PATCH] clean up ... |
1311 1312 |
if (!mapping) return 1; |
19fd62312 mm: spinlock tree... |
1313 |
spin_lock_irq(&mapping->tree_lock); |
8c08540f8 [PATCH] clean up ... |
1314 1315 1316 |
mapping2 = page_mapping(page); if (mapping2) { /* Race with truncate? */ BUG_ON(mapping2 != mapping); |
787d2214c fs: introduce som... |
1317 |
WARN_ON_ONCE(!PagePrivate(page) && !PageUptodate(page)); |
e3a7cca1e vfs: add/use acco... |
1318 |
account_page_dirtied(page, mapping); |
8c08540f8 [PATCH] clean up ... |
1319 1320 1321 |
radix_tree_tag_set(&mapping->page_tree, page_index(page), PAGECACHE_TAG_DIRTY); } |
19fd62312 mm: spinlock tree... |
1322 |
spin_unlock_irq(&mapping->tree_lock); |
8c08540f8 [PATCH] clean up ... |
1323 1324 1325 |
if (mapping->host) { /* !PageAnon && !swapper_space */ __mark_inode_dirty(mapping->host, I_DIRTY_PAGES); |
1da177e4c Linux-2.6.12-rc2 |
1326 |
} |
4741c9fd3 [PATCH] set_page_... |
1327 |
return 1; |
1da177e4c Linux-2.6.12-rc2 |
1328 |
} |
4741c9fd3 [PATCH] set_page_... |
1329 |
return 0; |
1da177e4c Linux-2.6.12-rc2 |
1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 |
} 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: ... |
1346 1347 1348 1349 1350 1351 1352 |
* 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 |
1353 1354 1355 |
* 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... |
1356 |
int set_page_dirty(struct page *page) |
1da177e4c Linux-2.6.12-rc2 |
1357 1358 1359 1360 1361 |
{ struct address_space *mapping = page_mapping(page); if (likely(mapping)) { int (*spd)(struct page *) = mapping->a_ops->set_page_dirty; |
278df9f45 mm: reclaim inval... |
1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 |
/* * readahead/lru_deactivate_page could remain * PG_readahead/PG_reclaim due to race with end_page_writeback * About readahead, if the page is written, the flags would be * reset. So no problem. * About lru_deactivate_page, if the page is redirty, the flag * will be reset. So no problem. but if the page is used by readahead * it will confuse readahead and make it restart the size rampup * process. But it's a trivial problem. */ ClearPageReclaim(page); |
9361401eb [PATCH] BLOCK: Ma... |
1373 1374 1375 1376 1377 |
#ifdef CONFIG_BLOCK if (!spd) spd = __set_page_dirty_buffers; #endif return (*spd)(page); |
1da177e4c Linux-2.6.12-rc2 |
1378 |
} |
4741c9fd3 [PATCH] set_page_... |
1379 1380 1381 1382 |
if (!PageDirty(page)) { if (!TestSetPageDirty(page)) return 1; } |
1da177e4c Linux-2.6.12-rc2 |
1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 |
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; |
7eaceacca block: remove per... |
1400 |
lock_page(page); |
1da177e4c Linux-2.6.12-rc2 |
1401 1402 1403 1404 1405 1406 1407 |
ret = set_page_dirty(page); unlock_page(page); return ret; } EXPORT_SYMBOL(set_page_dirty_lock); /* |
1da177e4c Linux-2.6.12-rc2 |
1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 |
* 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... |
1424 |
BUG_ON(!PageLocked(page)); |
7658cc289 VM: Fix nasty and... |
1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 |
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... |
1450 1451 1452 |
*/ if (page_mkclean(page)) set_page_dirty(page); |
79352894b mm: fix clear_pag... |
1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 |
/* * 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... |
1463 |
if (TestClearPageDirty(page)) { |
8c08540f8 [PATCH] clean up ... |
1464 |
dec_zone_page_state(page, NR_FILE_DIRTY); |
c9e51e418 mm: count reclaim... |
1465 1466 |
dec_bdi_stat(mapping->backing_dev_info, BDI_RECLAIMABLE); |
7658cc289 VM: Fix nasty and... |
1467 |
return 1; |
1da177e4c Linux-2.6.12-rc2 |
1468 |
} |
7658cc289 VM: Fix nasty and... |
1469 |
return 0; |
1da177e4c Linux-2.6.12-rc2 |
1470 |
} |
7658cc289 VM: Fix nasty and... |
1471 |
return TestClearPageDirty(page); |
1da177e4c Linux-2.6.12-rc2 |
1472 |
} |
58bb01a9c [PATCH] re-export... |
1473 |
EXPORT_SYMBOL(clear_page_dirty_for_io); |
1da177e4c Linux-2.6.12-rc2 |
1474 1475 1476 1477 1478 1479 1480 |
int test_clear_page_writeback(struct page *page) { struct address_space *mapping = page_mapping(page); int ret; if (mapping) { |
69cb51d18 mm: count writeba... |
1481 |
struct backing_dev_info *bdi = mapping->backing_dev_info; |
1da177e4c Linux-2.6.12-rc2 |
1482 |
unsigned long flags; |
19fd62312 mm: spinlock tree... |
1483 |
spin_lock_irqsave(&mapping->tree_lock, flags); |
1da177e4c Linux-2.6.12-rc2 |
1484 |
ret = TestClearPageWriteback(page); |
69cb51d18 mm: count writeba... |
1485 |
if (ret) { |
1da177e4c Linux-2.6.12-rc2 |
1486 1487 1488 |
radix_tree_tag_clear(&mapping->page_tree, page_index(page), PAGECACHE_TAG_WRITEBACK); |
e4ad08fe6 mm: bdi: add sepa... |
1489 |
if (bdi_cap_account_writeback(bdi)) { |
69cb51d18 mm: count writeba... |
1490 |
__dec_bdi_stat(bdi, BDI_WRITEBACK); |
04fbfdc14 mm: per device di... |
1491 1492 |
__bdi_writeout_inc(bdi); } |
69cb51d18 mm: count writeba... |
1493 |
} |
19fd62312 mm: spinlock tree... |
1494 |
spin_unlock_irqrestore(&mapping->tree_lock, flags); |
1da177e4c Linux-2.6.12-rc2 |
1495 1496 1497 |
} else { ret = TestClearPageWriteback(page); } |
99b12e3d8 writeback: accoun... |
1498 |
if (ret) { |
d688abf50 move page writeba... |
1499 |
dec_zone_page_state(page, NR_WRITEBACK); |
99b12e3d8 writeback: accoun... |
1500 1501 |
inc_zone_page_state(page, NR_WRITTEN); } |
1da177e4c Linux-2.6.12-rc2 |
1502 1503 1504 1505 1506 1507 1508 1509 1510 |
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... |
1511 |
struct backing_dev_info *bdi = mapping->backing_dev_info; |
1da177e4c Linux-2.6.12-rc2 |
1512 |
unsigned long flags; |
19fd62312 mm: spinlock tree... |
1513 |
spin_lock_irqsave(&mapping->tree_lock, flags); |
1da177e4c Linux-2.6.12-rc2 |
1514 |
ret = TestSetPageWriteback(page); |
69cb51d18 mm: count writeba... |
1515 |
if (!ret) { |
1da177e4c Linux-2.6.12-rc2 |
1516 1517 1518 |
radix_tree_tag_set(&mapping->page_tree, page_index(page), PAGECACHE_TAG_WRITEBACK); |
e4ad08fe6 mm: bdi: add sepa... |
1519 |
if (bdi_cap_account_writeback(bdi)) |
69cb51d18 mm: count writeba... |
1520 1521 |
__inc_bdi_stat(bdi, BDI_WRITEBACK); } |
1da177e4c Linux-2.6.12-rc2 |
1522 1523 1524 1525 |
if (!PageDirty(page)) radix_tree_tag_clear(&mapping->page_tree, page_index(page), PAGECACHE_TAG_DIRTY); |
f446daaea mm: implement wri... |
1526 1527 1528 |
radix_tree_tag_clear(&mapping->page_tree, page_index(page), PAGECACHE_TAG_TOWRITE); |
19fd62312 mm: spinlock tree... |
1529 |
spin_unlock_irqrestore(&mapping->tree_lock, flags); |
1da177e4c Linux-2.6.12-rc2 |
1530 1531 1532 |
} else { ret = TestSetPageWriteback(page); } |
d688abf50 move page writeba... |
1533 |
if (!ret) |
f629d1c9b mm: add account_p... |
1534 |
account_page_writeback(page); |
1da177e4c Linux-2.6.12-rc2 |
1535 1536 1537 1538 1539 1540 |
return ret; } EXPORT_SYMBOL(test_set_page_writeback); /* |
001281881 mm: use lockless ... |
1541 |
* Return true if any of the pages in the mapping are marked with the |
1da177e4c Linux-2.6.12-rc2 |
1542 1543 1544 1545 |
* passed tag. */ int mapping_tagged(struct address_space *mapping, int tag) { |
72c478321 mm: remove useles... |
1546 |
return radix_tree_tagged(&mapping->page_tree, tag); |
1da177e4c Linux-2.6.12-rc2 |
1547 1548 |
} EXPORT_SYMBOL(mapping_tagged); |