Blame view
mm/page-writeback.c
36.4 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> |
1da177e4c Linux-2.6.12-rc2 |
37 38 |
/* |
1da177e4c Linux-2.6.12-rc2 |
39 40 41 42 |
* 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 |
43 44 45 |
/* * 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... |
46 |
* It should be somewhat larger than dirtied pages to ensure that reasonably |
1da177e4c Linux-2.6.12-rc2 |
47 48 |
* large amounts of I/O are submitted. */ |
3a2e9a5a2 writeback: balanc... |
49 |
static inline long sync_writeback_pages(unsigned long dirtied) |
1da177e4c Linux-2.6.12-rc2 |
50 |
{ |
3a2e9a5a2 writeback: balanc... |
51 52 53 54 |
if (dirtied < ratelimit_pages) dirtied = ratelimit_pages; return dirtied + dirtied / 2; |
1da177e4c Linux-2.6.12-rc2 |
55 56 57 58 59 |
} /* The following parameters are exported via /proc/sys/vm */ /* |
5b0830cb9 writeback: get ri... |
60 |
* Start background writeback (via writeback threads) at this percentage |
1da177e4c Linux-2.6.12-rc2 |
61 |
*/ |
1b5e62b42 writeback: double... |
62 |
int dirty_background_ratio = 10; |
1da177e4c Linux-2.6.12-rc2 |
63 64 |
/* |
2da02997e mm: add dirty_bac... |
65 66 67 68 69 70 |
* 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... |
71 72 73 74 75 76 |
* 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 |
77 78 |
* The generator of dirty data starts writeback at this percentage */ |
1b5e62b42 writeback: double... |
79 |
int vm_dirty_ratio = 20; |
1da177e4c Linux-2.6.12-rc2 |
80 81 |
/* |
2da02997e mm: add dirty_bac... |
82 83 84 85 86 87 |
* 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... |
88 |
* The interval between `kupdate'-style writebacks |
1da177e4c Linux-2.6.12-rc2 |
89 |
*/ |
22ef37eed page-writeback: f... |
90 |
unsigned int dirty_writeback_interval = 5 * 100; /* centiseconds */ |
1da177e4c Linux-2.6.12-rc2 |
91 92 |
/* |
704503d83 mm: fix proc_doin... |
93 |
* The longest time for which data is allowed to remain dirty |
1da177e4c Linux-2.6.12-rc2 |
94 |
*/ |
22ef37eed page-writeback: f... |
95 |
unsigned int dirty_expire_interval = 30 * 100; /* centiseconds */ |
1da177e4c Linux-2.6.12-rc2 |
96 97 98 99 100 101 102 |
/* * Flag that makes the machine dump writes/reads and block dirtyings. */ int block_dump; /* |
ed5b43f15 [PATCH] Represent... |
103 104 |
* 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 |
105 106 107 108 109 110 |
*/ int laptop_mode; EXPORT_SYMBOL(laptop_mode); /* End of sysctl-exported parameters */ |
1da177e4c Linux-2.6.12-rc2 |
111 |
/* |
04fbfdc14 mm: per device di... |
112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 |
* 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... |
128 |
static struct prop_descriptor vm_dirties; |
04fbfdc14 mm: per device di... |
129 |
|
04fbfdc14 mm: per device di... |
130 131 132 133 134 135 136 137 |
/* * 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... |
138 139 140 141 142 |
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... |
143 144 145 146 |
return 2 + ilog2(dirty_total - 1); } /* |
2da02997e mm: add dirty_bac... |
147 |
* update the period when the dirty threshold changes. |
04fbfdc14 mm: per device di... |
148 |
*/ |
2da02997e mm: add dirty_bac... |
149 150 151 152 153 154 155 156 |
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... |
157 |
void __user *buffer, size_t *lenp, |
2da02997e mm: add dirty_bac... |
158 159 160 |
loff_t *ppos) { int ret; |
8d65af789 sysctl: remove "s... |
161 |
ret = proc_dointvec_minmax(table, write, buffer, lenp, ppos); |
2da02997e mm: add dirty_bac... |
162 163 164 165 166 167 |
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... |
168 |
void __user *buffer, size_t *lenp, |
2da02997e mm: add dirty_bac... |
169 170 171 |
loff_t *ppos) { int ret; |
8d65af789 sysctl: remove "s... |
172 |
ret = proc_doulongvec_minmax(table, write, buffer, lenp, ppos); |
2da02997e mm: add dirty_bac... |
173 174 175 176 |
if (ret == 0 && write) dirty_background_ratio = 0; return ret; } |
04fbfdc14 mm: per device di... |
177 |
int dirty_ratio_handler(struct ctl_table *table, int write, |
8d65af789 sysctl: remove "s... |
178 |
void __user *buffer, size_t *lenp, |
04fbfdc14 mm: per device di... |
179 180 181 |
loff_t *ppos) { int old_ratio = vm_dirty_ratio; |
2da02997e mm: add dirty_bac... |
182 |
int ret; |
8d65af789 sysctl: remove "s... |
183 |
ret = proc_dointvec_minmax(table, write, buffer, lenp, ppos); |
04fbfdc14 mm: per device di... |
184 |
if (ret == 0 && write && vm_dirty_ratio != old_ratio) { |
2da02997e mm: add dirty_bac... |
185 186 187 188 189 190 191 192 |
update_completion_period(); vm_dirty_bytes = 0; } return ret; } int dirty_bytes_handler(struct ctl_table *table, int write, |
8d65af789 sysctl: remove "s... |
193 |
void __user *buffer, size_t *lenp, |
2da02997e mm: add dirty_bac... |
194 195 |
loff_t *ppos) { |
fc3501d41 mm: fix dirty_byt... |
196 |
unsigned long old_bytes = vm_dirty_bytes; |
2da02997e mm: add dirty_bac... |
197 |
int ret; |
8d65af789 sysctl: remove "s... |
198 |
ret = proc_doulongvec_minmax(table, write, buffer, lenp, ppos); |
2da02997e mm: add dirty_bac... |
199 200 201 |
if (ret == 0 && write && vm_dirty_bytes != old_bytes) { update_completion_period(); vm_dirty_ratio = 0; |
04fbfdc14 mm: per device di... |
202 203 204 205 206 207 208 209 210 211 |
} 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... |
212 213 |
__prop_inc_percpu_max(&vm_completions, &bdi->completions, bdi->max_prop_frac); |
04fbfdc14 mm: per device di... |
214 |
} |
dd5656e59 mm: bdi: export b... |
215 216 217 218 219 220 221 222 223 |
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... |
224 |
void task_dirty_inc(struct task_struct *tsk) |
3e26c149c mm: dirty balanci... |
225 226 227 |
{ prop_inc_single(&vm_dirties, &tsk->dirties); } |
04fbfdc14 mm: per device di... |
228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 |
/* * Obtain an accurate fraction of the BDI's portion. */ static void bdi_writeout_fraction(struct backing_dev_info *bdi, long *numerator, long *denominator) { if (bdi_cap_writeback_dirty(bdi)) { prop_fraction_percpu(&vm_completions, &bdi->completions, numerator, denominator); } else { *numerator = 0; *denominator = 1; } } /* * Clip the earned share of dirty pages to that which is actually available. * This avoids exceeding the total dirty_limit when the floating averages * fluctuate too quickly. */ |
dcf975d58 mm/page-writeback... |
248 249 |
static void clip_bdi_dirty_limit(struct backing_dev_info *bdi, unsigned long dirty, unsigned long *pbdi_dirty) |
04fbfdc14 mm: per device di... |
250 |
{ |
dcf975d58 mm/page-writeback... |
251 |
unsigned long avail_dirty; |
04fbfdc14 mm: per device di... |
252 |
|
dcf975d58 mm/page-writeback... |
253 |
avail_dirty = global_page_state(NR_FILE_DIRTY) + |
04fbfdc14 mm: per device di... |
254 |
global_page_state(NR_WRITEBACK) + |
fc3ba692a mm: Add NR_WRITEB... |
255 |
global_page_state(NR_UNSTABLE_NFS) + |
dcf975d58 mm/page-writeback... |
256 |
global_page_state(NR_WRITEBACK_TEMP); |
04fbfdc14 mm: per device di... |
257 |
|
dcf975d58 mm/page-writeback... |
258 259 260 |
if (avail_dirty < dirty) avail_dirty = dirty - avail_dirty; else |
04fbfdc14 mm: per device di... |
261 262 263 264 265 266 267 |
avail_dirty = 0; avail_dirty += bdi_stat(bdi, BDI_RECLAIMABLE) + bdi_stat(bdi, BDI_WRITEBACK); *pbdi_dirty = min(*pbdi_dirty, avail_dirty); } |
3e26c149c mm: dirty balanci... |
268 269 270 271 272 273 274 275 276 277 278 279 280 281 |
static inline void task_dirties_fraction(struct task_struct *tsk, long *numerator, long *denominator) { prop_fraction_single(&vm_dirties, &tsk->dirties, numerator, denominator); } /* * scale the dirty limit * * task specific dirty limit: * * dirty -= (dirty/8) * p_{t} */ |
dcf975d58 mm/page-writeback... |
282 |
static void task_dirty_limit(struct task_struct *tsk, unsigned long *pdirty) |
3e26c149c mm: dirty balanci... |
283 284 |
{ long numerator, denominator; |
dcf975d58 mm/page-writeback... |
285 |
unsigned long dirty = *pdirty; |
3e26c149c mm: dirty balanci... |
286 287 288 289 290 291 292 293 294 295 296 297 |
u64 inv = dirty >> 3; task_dirties_fraction(tsk, &numerator, &denominator); inv *= numerator; do_div(inv, denominator); dirty -= inv; if (dirty < *pdirty/2) dirty = *pdirty/2; *pdirty = dirty; } |
04fbfdc14 mm: per device di... |
298 |
/* |
189d3c4a9 mm: bdi: allow se... |
299 300 |
* */ |
189d3c4a9 mm: bdi: allow se... |
301 302 303 304 305 |
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... |
306 |
|
cfc4ba536 writeback: use RC... |
307 |
spin_lock_bh(&bdi_lock); |
a42dde041 mm: bdi: allow se... |
308 |
if (min_ratio > bdi->max_ratio) { |
189d3c4a9 mm: bdi: allow se... |
309 |
ret = -EINVAL; |
a42dde041 mm: bdi: allow se... |
310 311 312 313 314 315 316 317 318 |
} 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... |
319 |
spin_unlock_bh(&bdi_lock); |
a42dde041 mm: bdi: allow se... |
320 321 322 323 324 325 |
return ret; } int bdi_set_max_ratio(struct backing_dev_info *bdi, unsigned max_ratio) { |
a42dde041 mm: bdi: allow se... |
326 327 328 329 |
int ret = 0; if (max_ratio > 100) return -EINVAL; |
cfc4ba536 writeback: use RC... |
330 |
spin_lock_bh(&bdi_lock); |
a42dde041 mm: bdi: allow se... |
331 332 333 334 335 336 |
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... |
337 |
spin_unlock_bh(&bdi_lock); |
189d3c4a9 mm: bdi: allow se... |
338 339 340 |
return ret; } |
a42dde041 mm: bdi: allow se... |
341 |
EXPORT_SYMBOL(bdi_set_max_ratio); |
189d3c4a9 mm: bdi: allow se... |
342 343 |
/* |
1da177e4c Linux-2.6.12-rc2 |
344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 |
* 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 ... |
360 361 362 363 364 365 |
static unsigned long highmem_dirtyable_memory(unsigned long total) { #ifdef CONFIG_HIGHMEM int node; unsigned long x = 0; |
37b07e416 memoryless nodes:... |
366 |
for_each_node_state(node, N_HIGH_MEMORY) { |
1b4244647 Use ZVC counters ... |
367 368 |
struct zone *z = &NODE_DATA(node)->node_zones[ZONE_HIGHMEM]; |
adea02a1b mm: count only re... |
369 370 |
x += zone_page_state(z, NR_FREE_PAGES) + zone_reclaimable_pages(z); |
1b4244647 Use ZVC counters ... |
371 372 373 374 375 376 377 378 379 380 381 382 |
} /* * 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... |
383 384 385 386 387 388 389 |
/** * 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 ... |
390 391 |
{ unsigned long x; |
adea02a1b mm: count only re... |
392 |
x = global_page_state(NR_FREE_PAGES) + global_reclaimable_pages(); |
195cf453d mm/page-writeback... |
393 394 395 |
if (!vm_highmem_is_dirtyable) x -= highmem_dirtyable_memory(x); |
1b4244647 Use ZVC counters ... |
396 397 |
return x + 1; /* Ensure that we never return 0 */ } |
cf0ca9fe5 mm: bdi: export B... |
398 |
void |
364aeb284 mm: change dirty ... |
399 400 |
get_dirty_limits(unsigned long *pbackground, unsigned long *pdirty, unsigned long *pbdi_dirty, struct backing_dev_info *bdi) |
1da177e4c Linux-2.6.12-rc2 |
401 |
{ |
364aeb284 mm: change dirty ... |
402 403 |
unsigned long background; unsigned long dirty; |
1b4244647 Use ZVC counters ... |
404 |
unsigned long available_memory = determine_dirtyable_memory(); |
1da177e4c Linux-2.6.12-rc2 |
405 |
struct task_struct *tsk; |
2da02997e mm: add dirty_bac... |
406 407 408 409 410 411 412 413 414 415 |
if (vm_dirty_bytes) dirty = DIV_ROUND_UP(vm_dirty_bytes, PAGE_SIZE); else { int dirty_ratio; dirty_ratio = vm_dirty_ratio; if (dirty_ratio < 5) dirty_ratio = 5; dirty = (dirty_ratio * available_memory) / 100; } |
1da177e4c Linux-2.6.12-rc2 |
416 |
|
2da02997e mm: add dirty_bac... |
417 418 419 420 |
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 |
421 |
|
2da02997e mm: add dirty_bac... |
422 423 |
if (background >= dirty) background = dirty / 2; |
1da177e4c Linux-2.6.12-rc2 |
424 425 426 427 428 429 430 |
tsk = current; if (tsk->flags & PF_LESS_THROTTLE || rt_task(tsk)) { background += background / 4; dirty += dirty / 4; } *pbackground = background; *pdirty = dirty; |
04fbfdc14 mm: per device di... |
431 432 |
if (bdi) { |
189d3c4a9 mm: bdi: allow se... |
433 |
u64 bdi_dirty; |
04fbfdc14 mm: per device di... |
434 435 436 437 438 439 |
long numerator, denominator; /* * Calculate this BDI's share of the dirty ratio. */ bdi_writeout_fraction(bdi, &numerator, &denominator); |
189d3c4a9 mm: bdi: allow se... |
440 |
bdi_dirty = (dirty * (100 - bdi_min_ratio)) / 100; |
04fbfdc14 mm: per device di... |
441 442 |
bdi_dirty *= numerator; do_div(bdi_dirty, denominator); |
189d3c4a9 mm: bdi: allow se... |
443 |
bdi_dirty += (dirty * bdi->min_ratio) / 100; |
a42dde041 mm: bdi: allow se... |
444 445 |
if (bdi_dirty > (dirty * bdi->max_ratio) / 100) bdi_dirty = dirty * bdi->max_ratio / 100; |
04fbfdc14 mm: per device di... |
446 447 448 |
*pbdi_dirty = bdi_dirty; clip_bdi_dirty_limit(bdi, dirty, pbdi_dirty); |
3e26c149c mm: dirty balanci... |
449 |
task_dirty_limit(current, pbdi_dirty); |
04fbfdc14 mm: per device di... |
450 |
} |
1da177e4c Linux-2.6.12-rc2 |
451 452 453 454 455 456 |
} /* * 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... |
457 458 |
* If we're over `background_thresh' then the writeback threads are woken to * perform some writeout. |
1da177e4c Linux-2.6.12-rc2 |
459 |
*/ |
3a2e9a5a2 writeback: balanc... |
460 461 |
static void balance_dirty_pages(struct address_space *mapping, unsigned long write_chunk) |
1da177e4c Linux-2.6.12-rc2 |
462 |
{ |
5fce25a9d mm: speed up writ... |
463 464 |
long nr_reclaimable, bdi_nr_reclaimable; long nr_writeback, bdi_nr_writeback; |
364aeb284 mm: change dirty ... |
465 466 467 |
unsigned long background_thresh; unsigned long dirty_thresh; unsigned long bdi_thresh; |
1da177e4c Linux-2.6.12-rc2 |
468 |
unsigned long pages_written = 0; |
87c6a9b25 writeback: make b... |
469 |
unsigned long pause = 1; |
1da177e4c Linux-2.6.12-rc2 |
470 471 472 473 474 475 476 477 478 |
struct backing_dev_info *bdi = mapping->backing_dev_info; for (;;) { struct writeback_control wbc = { .bdi = bdi, .sync_mode = WB_SYNC_NONE, .older_than_this = NULL, .nr_to_write = write_chunk, |
111ebb6e6 [PATCH] writeback... |
479 |
.range_cyclic = 1, |
1da177e4c Linux-2.6.12-rc2 |
480 |
}; |
04fbfdc14 mm: per device di... |
481 482 |
get_dirty_limits(&background_thresh, &dirty_thresh, &bdi_thresh, bdi); |
5fce25a9d mm: speed up writ... |
483 484 485 486 |
nr_reclaimable = global_page_state(NR_FILE_DIRTY) + global_page_state(NR_UNSTABLE_NFS); nr_writeback = global_page_state(NR_WRITEBACK); |
04fbfdc14 mm: per device di... |
487 488 |
bdi_nr_reclaimable = bdi_stat(bdi, BDI_RECLAIMABLE); bdi_nr_writeback = bdi_stat(bdi, BDI_WRITEBACK); |
5fce25a9d mm: speed up writ... |
489 |
|
04fbfdc14 mm: per device di... |
490 491 |
if (bdi_nr_reclaimable + bdi_nr_writeback <= bdi_thresh) break; |
1da177e4c Linux-2.6.12-rc2 |
492 |
|
5fce25a9d mm: speed up writ... |
493 494 495 496 497 498 499 500 |
/* * Throttle it only when the background writeback cannot * catch-up. This avoids (excessively) small writeouts * when the bdi limits are ramping up. */ if (nr_reclaimable + nr_writeback < (background_thresh + dirty_thresh) / 2) break; |
04fbfdc14 mm: per device di... |
501 502 |
if (!bdi->dirty_exceeded) bdi->dirty_exceeded = 1; |
1da177e4c Linux-2.6.12-rc2 |
503 504 505 506 507 508 |
/* 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... |
509 510 511 |
* 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 |
512 |
*/ |
d7831a0bd mm: prevent balan... |
513 |
if (bdi_nr_reclaimable > bdi_thresh) { |
03ba3782e writeback: switch... |
514 |
writeback_inodes_wbc(&wbc); |
1da177e4c Linux-2.6.12-rc2 |
515 |
pages_written += write_chunk - wbc.nr_to_write; |
04fbfdc14 mm: per device di... |
516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 |
get_dirty_limits(&background_thresh, &dirty_thresh, &bdi_thresh, bdi); } /* * 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 if (bdi_nr_reclaimable) { bdi_nr_reclaimable = bdi_stat(bdi, BDI_RECLAIMABLE); bdi_nr_writeback = bdi_stat(bdi, BDI_WRITEBACK); |
1da177e4c Linux-2.6.12-rc2 |
536 |
} |
04fbfdc14 mm: per device di... |
537 538 539 540 541 |
if (bdi_nr_reclaimable + bdi_nr_writeback <= bdi_thresh) break; if (pages_written >= write_chunk) break; /* We've done our duty */ |
d25105e89 writeback: accoun... |
542 543 |
__set_current_state(TASK_INTERRUPTIBLE); io_schedule_timeout(pause); |
87c6a9b25 writeback: make b... |
544 545 546 547 548 549 550 551 |
/* * 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 |
552 |
} |
04fbfdc14 mm: per device di... |
553 554 555 |
if (bdi_nr_reclaimable + bdi_nr_writeback < bdi_thresh && bdi->dirty_exceeded) 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) || |
d3ddec763 writeback: stop b... |
569 570 |
(!laptop_mode && ((global_page_state(NR_FILE_DIRTY) + global_page_state(NR_UNSTABLE_NFS)) |
b6e51316d writeback: separa... |
571 |
> background_thresh))) |
a72bfd4de writeback: pass i... |
572 |
bdi_start_writeback(bdi, NULL, 0); |
1da177e4c Linux-2.6.12-rc2 |
573 |
} |
a200ee182 mm: set_page_dirt... |
574 |
void set_page_dirty_balance(struct page *page, int page_mkwrite) |
edc79b2a4 [PATCH] mm: balan... |
575 |
{ |
a200ee182 mm: set_page_dirt... |
576 |
if (set_page_dirty(page) || page_mkwrite) { |
edc79b2a4 [PATCH] mm: balan... |
577 578 579 580 581 582 |
struct address_space *mapping = page_mapping(page); if (mapping) balance_dirty_pages_ratelimited(mapping); } } |
245b2e70e percpu: clean up ... |
583 |
static DEFINE_PER_CPU(unsigned long, bdp_ratelimits) = 0; |
1da177e4c Linux-2.6.12-rc2 |
584 |
/** |
fa5a734e4 [PATCH] balance_d... |
585 |
* balance_dirty_pages_ratelimited_nr - balance dirty memory state |
67be2dd1b [PATCH] DocBook: ... |
586 |
* @mapping: address_space which was dirtied |
a580290c3 Documentation: fi... |
587 |
* @nr_pages_dirtied: number of pages which the caller has just dirtied |
1da177e4c Linux-2.6.12-rc2 |
588 589 590 591 592 593 594 595 596 597 |
* * 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... |
598 599 |
void balance_dirty_pages_ratelimited_nr(struct address_space *mapping, unsigned long nr_pages_dirtied) |
1da177e4c Linux-2.6.12-rc2 |
600 |
{ |
fa5a734e4 [PATCH] balance_d... |
601 602 |
unsigned long ratelimit; unsigned long *p; |
1da177e4c Linux-2.6.12-rc2 |
603 604 |
ratelimit = ratelimit_pages; |
04fbfdc14 mm: per device di... |
605 |
if (mapping->backing_dev_info->dirty_exceeded) |
1da177e4c Linux-2.6.12-rc2 |
606 607 608 609 610 611 |
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... |
612 |
preempt_disable(); |
245b2e70e percpu: clean up ... |
613 |
p = &__get_cpu_var(bdp_ratelimits); |
fa5a734e4 [PATCH] balance_d... |
614 615 |
*p += nr_pages_dirtied; if (unlikely(*p >= ratelimit)) { |
3a2e9a5a2 writeback: balanc... |
616 |
ratelimit = sync_writeback_pages(*p); |
fa5a734e4 [PATCH] balance_d... |
617 618 |
*p = 0; preempt_enable(); |
3a2e9a5a2 writeback: balanc... |
619 |
balance_dirty_pages(mapping, ratelimit); |
1da177e4c Linux-2.6.12-rc2 |
620 621 |
return; } |
fa5a734e4 [PATCH] balance_d... |
622 |
preempt_enable(); |
1da177e4c Linux-2.6.12-rc2 |
623 |
} |
fa5a734e4 [PATCH] balance_d... |
624 |
EXPORT_SYMBOL(balance_dirty_pages_ratelimited_nr); |
1da177e4c Linux-2.6.12-rc2 |
625 |
|
232ea4d69 [PATCH] throttle_... |
626 |
void throttle_vm_writeout(gfp_t gfp_mask) |
1da177e4c Linux-2.6.12-rc2 |
627 |
{ |
364aeb284 mm: change dirty ... |
628 629 |
unsigned long background_thresh; unsigned long dirty_thresh; |
1da177e4c Linux-2.6.12-rc2 |
630 631 |
for ( ; ; ) { |
04fbfdc14 mm: per device di... |
632 |
get_dirty_limits(&background_thresh, &dirty_thresh, NULL, NULL); |
1da177e4c Linux-2.6.12-rc2 |
633 634 635 636 637 638 |
/* * 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 ... |
639 640 641 |
if (global_page_state(NR_UNSTABLE_NFS) + global_page_state(NR_WRITEBACK) <= dirty_thresh) break; |
8aa7e847d Fix congestion_wa... |
642 |
congestion_wait(BLK_RW_ASYNC, HZ/10); |
369f2389e writeback: remove... |
643 644 645 646 647 648 649 650 |
/* * 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 |
651 652 |
} } |
1da177e4c Linux-2.6.12-rc2 |
653 |
static void laptop_timer_fn(unsigned long unused); |
8d06afab7 [PATCH] timer ini... |
654 |
static DEFINE_TIMER(laptop_mode_wb_timer, laptop_timer_fn, 0, 0); |
1da177e4c Linux-2.6.12-rc2 |
655 656 |
/* |
1da177e4c Linux-2.6.12-rc2 |
657 658 659 |
* sysctl handler for /proc/sys/vm/dirty_writeback_centisecs */ int dirty_writeback_centisecs_handler(ctl_table *table, int write, |
8d65af789 sysctl: remove "s... |
660 |
void __user *buffer, size_t *length, loff_t *ppos) |
1da177e4c Linux-2.6.12-rc2 |
661 |
{ |
8d65af789 sysctl: remove "s... |
662 |
proc_dointvec(table, write, buffer, length, ppos); |
1da177e4c Linux-2.6.12-rc2 |
663 664 |
return 0; } |
03ba3782e writeback: switch... |
665 |
static void do_laptop_sync(struct work_struct *work) |
1da177e4c Linux-2.6.12-rc2 |
666 |
{ |
03ba3782e writeback: switch... |
667 668 |
wakeup_flusher_threads(0); kfree(work); |
1da177e4c Linux-2.6.12-rc2 |
669 670 671 672 |
} static void laptop_timer_fn(unsigned long unused) { |
03ba3782e writeback: switch... |
673 674 675 676 677 678 679 |
struct work_struct *work; work = kmalloc(sizeof(*work), GFP_ATOMIC); if (work) { INIT_WORK(work, do_laptop_sync); schedule_work(work); } |
1da177e4c Linux-2.6.12-rc2 |
680 681 682 683 684 685 686 687 688 |
} /* * 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. */ void laptop_io_completion(void) { |
ed5b43f15 [PATCH] Represent... |
689 |
mod_timer(&laptop_mode_wb_timer, jiffies + laptop_mode); |
1da177e4c Linux-2.6.12-rc2 |
690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 |
} /* * 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) { del_timer(&laptop_mode_wb_timer); } /* * 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... |
718 |
void writeback_set_ratelimit(void) |
1da177e4c Linux-2.6.12-rc2 |
719 |
{ |
40c99aae2 [PATCH] remove st... |
720 |
ratelimit_pages = vm_total_pages / (num_online_cpus() * 32); |
1da177e4c Linux-2.6.12-rc2 |
721 722 723 724 725 |
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... |
726 |
static int __cpuinit |
1da177e4c Linux-2.6.12-rc2 |
727 728 |
ratelimit_handler(struct notifier_block *self, unsigned long u, void *v) { |
2d1d43f6a [PATCH] call mm/p... |
729 |
writeback_set_ratelimit(); |
aa0f03037 [PATCH] Change co... |
730 |
return NOTIFY_DONE; |
1da177e4c Linux-2.6.12-rc2 |
731 |
} |
74b85f379 [PATCH] cpu hotpl... |
732 |
static struct notifier_block __cpuinitdata ratelimit_nb = { |
1da177e4c Linux-2.6.12-rc2 |
733 734 735 736 737 |
.notifier_call = ratelimit_handler, .next = NULL, }; /* |
dc6e29da9 Fix balance_dirty... |
738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 |
* 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 |
754 755 756 |
*/ void __init page_writeback_init(void) { |
04fbfdc14 mm: per device di... |
757 |
int shift; |
2d1d43f6a [PATCH] call mm/p... |
758 |
writeback_set_ratelimit(); |
1da177e4c Linux-2.6.12-rc2 |
759 |
register_cpu_notifier(&ratelimit_nb); |
04fbfdc14 mm: per device di... |
760 761 762 |
shift = calc_period_shift(); prop_descriptor_init(&vm_completions, shift); |
3e26c149c mm: dirty balanci... |
763 |
prop_descriptor_init(&vm_dirties, shift); |
1da177e4c Linux-2.6.12-rc2 |
764 |
} |
811d736f9 [PATCH] BLOCK: Di... |
765 |
/** |
0ea971801 consolidate gener... |
766 |
* write_cache_pages - walk the list of dirty pages of the given address space and write all of them. |
811d736f9 [PATCH] BLOCK: Di... |
767 768 |
* @mapping: address space structure to write * @wbc: subtract the number of written pages from *@wbc->nr_to_write |
0ea971801 consolidate gener... |
769 770 |
* @writepage: function called for each page * @data: data passed to writepage function |
811d736f9 [PATCH] BLOCK: Di... |
771 |
* |
0ea971801 consolidate gener... |
772 |
* If a page is already under I/O, write_cache_pages() skips it, even |
811d736f9 [PATCH] BLOCK: Di... |
773 774 775 776 777 778 |
* 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. |
811d736f9 [PATCH] BLOCK: Di... |
779 |
*/ |
0ea971801 consolidate gener... |
780 781 782 |
int write_cache_pages(struct address_space *mapping, struct writeback_control *wbc, writepage_t writepage, void *data) |
811d736f9 [PATCH] BLOCK: Di... |
783 784 785 786 |
{ struct backing_dev_info *bdi = mapping->backing_dev_info; int ret = 0; int done = 0; |
811d736f9 [PATCH] BLOCK: Di... |
787 788 |
struct pagevec pvec; int nr_pages; |
31a12666d mm: write_cache_p... |
789 |
pgoff_t uninitialized_var(writeback_index); |
811d736f9 [PATCH] BLOCK: Di... |
790 791 |
pgoff_t index; pgoff_t end; /* Inclusive */ |
bd19e012f mm: write_cache_p... |
792 |
pgoff_t done_index; |
31a12666d mm: write_cache_p... |
793 |
int cycled; |
811d736f9 [PATCH] BLOCK: Di... |
794 |
int range_whole = 0; |
17bc6c30c vfs: Add no_nrwri... |
795 |
long nr_to_write = wbc->nr_to_write; |
811d736f9 [PATCH] BLOCK: Di... |
796 797 798 799 800 |
if (wbc->nonblocking && bdi_write_congested(bdi)) { wbc->encountered_congestion = 1; return 0; } |
811d736f9 [PATCH] BLOCK: Di... |
801 802 |
pagevec_init(&pvec, 0); if (wbc->range_cyclic) { |
31a12666d mm: write_cache_p... |
803 804 805 806 807 808 |
writeback_index = mapping->writeback_index; /* prev offset */ index = writeback_index; if (index == 0) cycled = 1; else cycled = 0; |
811d736f9 [PATCH] BLOCK: Di... |
809 810 811 812 813 814 |
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... |
815 |
cycled = 1; /* ignore range_cyclic tests */ |
811d736f9 [PATCH] BLOCK: Di... |
816 817 |
} retry: |
bd19e012f mm: write_cache_p... |
818 |
done_index = index; |
5a3d5c981 mm: write_cache_p... |
819 820 821 822 823 824 825 826 |
while (!done && (index <= end)) { int i; nr_pages = pagevec_lookup_tag(&pvec, mapping, &index, PAGECACHE_TAG_DIRTY, min(end - index, (pgoff_t)PAGEVEC_SIZE-1) + 1); if (nr_pages == 0) break; |
811d736f9 [PATCH] BLOCK: Di... |
827 |
|
811d736f9 [PATCH] BLOCK: Di... |
828 829 830 831 |
for (i = 0; i < nr_pages; i++) { struct page *page = pvec.pages[i]; /* |
d5482cdf8 mm: write_cache_p... |
832 833 834 835 836 |
* 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... |
837 |
*/ |
d5482cdf8 mm: write_cache_p... |
838 839 840 841 842 843 844 845 846 847 |
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... |
848 |
lock_page(page); |
5a3d5c981 mm: write_cache_p... |
849 850 851 852 853 854 855 856 |
/* * 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... |
857 |
if (unlikely(page->mapping != mapping)) { |
5a3d5c981 mm: write_cache_p... |
858 |
continue_unlock: |
811d736f9 [PATCH] BLOCK: Di... |
859 860 861 |
unlock_page(page); continue; } |
515f4a037 mm: write_cache_p... |
862 863 864 865 866 867 868 869 870 871 872 |
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... |
873 |
|
515f4a037 mm: write_cache_p... |
874 875 |
BUG_ON(PageWriteback(page)); if (!clear_page_dirty_for_io(page)) |
5a3d5c981 mm: write_cache_p... |
876 |
goto continue_unlock; |
811d736f9 [PATCH] BLOCK: Di... |
877 |
|
0ea971801 consolidate gener... |
878 |
ret = (*writepage)(page, wbc, data); |
00266770b mm: write_cache_p... |
879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 |
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; } } |
89e121900 writeback: fix br... |
897 |
if (nr_to_write > 0) { |
dcf6a79dd write-back: fix n... |
898 |
nr_to_write--; |
89e121900 writeback: fix br... |
899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 |
if (nr_to_write == 0 && wbc->sync_mode == WB_SYNC_NONE) { /* * We stop writing back only if we are * not doing integrity sync. In case of * integrity sync we have to keep going * because someone may be concurrently * dirtying pages, and we might have * synced a lot of newly appeared dirty * pages, but have not synced all of the * old dirty pages. */ done = 1; break; } |
05fe478dd mm: write_cache_p... |
914 |
} |
dcf6a79dd write-back: fix n... |
915 |
|
811d736f9 [PATCH] BLOCK: Di... |
916 917 918 |
if (wbc->nonblocking && bdi_write_congested(bdi)) { wbc->encountered_congestion = 1; done = 1; |
82fd1a9a8 mm: write_cache_p... |
919 |
break; |
811d736f9 [PATCH] BLOCK: Di... |
920 921 922 923 924 |
} } pagevec_release(&pvec); cond_resched(); } |
3a4c6800f Fix page writebac... |
925 |
if (!cycled && !done) { |
811d736f9 [PATCH] BLOCK: Di... |
926 |
/* |
31a12666d mm: write_cache_p... |
927 |
* range_cyclic: |
811d736f9 [PATCH] BLOCK: Di... |
928 929 930 |
* 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... |
931 |
cycled = 1; |
811d736f9 [PATCH] BLOCK: Di... |
932 |
index = 0; |
31a12666d mm: write_cache_p... |
933 |
end = writeback_index - 1; |
811d736f9 [PATCH] BLOCK: Di... |
934 935 |
goto retry; } |
17bc6c30c vfs: Add no_nrwri... |
936 937 |
if (!wbc->no_nrwrite_index_update) { if (wbc->range_cyclic || (range_whole && nr_to_write > 0)) |
bd19e012f mm: write_cache_p... |
938 |
mapping->writeback_index = done_index; |
17bc6c30c vfs: Add no_nrwri... |
939 940 |
wbc->nr_to_write = nr_to_write; } |
06d6cf695 mm: Add range_con... |
941 |
|
811d736f9 [PATCH] BLOCK: Di... |
942 943 |
return ret; } |
0ea971801 consolidate gener... |
944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 |
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... |
976 977 |
EXPORT_SYMBOL(generic_writepages); |
1da177e4c Linux-2.6.12-rc2 |
978 979 |
int do_writepages(struct address_space *mapping, struct writeback_control *wbc) { |
22905f775 identify multipag... |
980 |
int ret; |
1da177e4c Linux-2.6.12-rc2 |
981 982 983 |
if (wbc->nr_to_write <= 0) return 0; if (mapping->a_ops->writepages) |
d08b3851d [PATCH] mm: track... |
984 |
ret = mapping->a_ops->writepages(mapping, wbc); |
22905f775 identify multipag... |
985 986 |
else ret = generic_writepages(mapping, wbc); |
22905f775 identify multipag... |
987 |
return ret; |
1da177e4c Linux-2.6.12-rc2 |
988 989 990 991 |
} /** * write_one_page - write out a single page and optionally wait on I/O |
67be2dd1b [PATCH] DocBook: ... |
992 993 |
* @page: the page to write * @wait: if true, wait on writeout |
1da177e4c Linux-2.6.12-rc2 |
994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 |
* * 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 ... |
1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 |
* For address_spaces which do not use buffers nor write back. */ int __set_page_dirty_no_writeback(struct page *page) { if (!PageDirty(page)) SetPageDirty(page); return 0; } /* |
e3a7cca1e vfs: add/use acco... |
1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 |
* 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); __inc_bdi_stat(mapping->backing_dev_info, BDI_RECLAIMABLE); task_dirty_inc(current); task_io_account_write(PAGE_CACHE_SIZE); } } /* |
1da177e4c Linux-2.6.12-rc2 |
1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 |
* 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/ |
1066 |
* mapping by re-checking page_mapping() inside tree_lock. |
1da177e4c Linux-2.6.12-rc2 |
1067 1068 1069 |
*/ int __set_page_dirty_nobuffers(struct page *page) { |
1da177e4c Linux-2.6.12-rc2 |
1070 1071 1072 |
if (!TestSetPageDirty(page)) { struct address_space *mapping = page_mapping(page); struct address_space *mapping2; |
8c08540f8 [PATCH] clean up ... |
1073 1074 |
if (!mapping) return 1; |
19fd62312 mm: spinlock tree... |
1075 |
spin_lock_irq(&mapping->tree_lock); |
8c08540f8 [PATCH] clean up ... |
1076 1077 1078 |
mapping2 = page_mapping(page); if (mapping2) { /* Race with truncate? */ BUG_ON(mapping2 != mapping); |
787d2214c fs: introduce som... |
1079 |
WARN_ON_ONCE(!PagePrivate(page) && !PageUptodate(page)); |
e3a7cca1e vfs: add/use acco... |
1080 |
account_page_dirtied(page, mapping); |
8c08540f8 [PATCH] clean up ... |
1081 1082 1083 |
radix_tree_tag_set(&mapping->page_tree, page_index(page), PAGECACHE_TAG_DIRTY); } |
19fd62312 mm: spinlock tree... |
1084 |
spin_unlock_irq(&mapping->tree_lock); |
8c08540f8 [PATCH] clean up ... |
1085 1086 1087 |
if (mapping->host) { /* !PageAnon && !swapper_space */ __mark_inode_dirty(mapping->host, I_DIRTY_PAGES); |
1da177e4c Linux-2.6.12-rc2 |
1088 |
} |
4741c9fd3 [PATCH] set_page_... |
1089 |
return 1; |
1da177e4c Linux-2.6.12-rc2 |
1090 |
} |
4741c9fd3 [PATCH] set_page_... |
1091 |
return 0; |
1da177e4c Linux-2.6.12-rc2 |
1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 |
} 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: ... |
1108 1109 1110 1111 1112 1113 1114 |
* 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 |
1115 1116 1117 |
* 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... |
1118 |
int set_page_dirty(struct page *page) |
1da177e4c Linux-2.6.12-rc2 |
1119 1120 1121 1122 1123 |
{ struct address_space *mapping = page_mapping(page); if (likely(mapping)) { int (*spd)(struct page *) = mapping->a_ops->set_page_dirty; |
9361401eb [PATCH] BLOCK: Ma... |
1124 1125 1126 1127 1128 |
#ifdef CONFIG_BLOCK if (!spd) spd = __set_page_dirty_buffers; #endif return (*spd)(page); |
1da177e4c Linux-2.6.12-rc2 |
1129 |
} |
4741c9fd3 [PATCH] set_page_... |
1130 1131 1132 1133 |
if (!PageDirty(page)) { if (!TestSetPageDirty(page)) return 1; } |
1da177e4c Linux-2.6.12-rc2 |
1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 |
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... |
1151 |
lock_page_nosync(page); |
1da177e4c Linux-2.6.12-rc2 |
1152 1153 1154 1155 1156 1157 1158 |
ret = set_page_dirty(page); unlock_page(page); return ret; } EXPORT_SYMBOL(set_page_dirty_lock); /* |
1da177e4c Linux-2.6.12-rc2 |
1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 |
* 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... |
1175 |
BUG_ON(!PageLocked(page)); |
fe3cba17c mm: share PG_read... |
1176 |
ClearPageReclaim(page); |
7658cc289 VM: Fix nasty and... |
1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 |
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... |
1202 1203 1204 |
*/ if (page_mkclean(page)) set_page_dirty(page); |
79352894b mm: fix clear_pag... |
1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 |
/* * 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... |
1215 |
if (TestClearPageDirty(page)) { |
8c08540f8 [PATCH] clean up ... |
1216 |
dec_zone_page_state(page, NR_FILE_DIRTY); |
c9e51e418 mm: count reclaim... |
1217 1218 |
dec_bdi_stat(mapping->backing_dev_info, BDI_RECLAIMABLE); |
7658cc289 VM: Fix nasty and... |
1219 |
return 1; |
1da177e4c Linux-2.6.12-rc2 |
1220 |
} |
7658cc289 VM: Fix nasty and... |
1221 |
return 0; |
1da177e4c Linux-2.6.12-rc2 |
1222 |
} |
7658cc289 VM: Fix nasty and... |
1223 |
return TestClearPageDirty(page); |
1da177e4c Linux-2.6.12-rc2 |
1224 |
} |
58bb01a9c [PATCH] re-export... |
1225 |
EXPORT_SYMBOL(clear_page_dirty_for_io); |
1da177e4c Linux-2.6.12-rc2 |
1226 1227 1228 1229 1230 1231 1232 |
int test_clear_page_writeback(struct page *page) { struct address_space *mapping = page_mapping(page); int ret; if (mapping) { |
69cb51d18 mm: count writeba... |
1233 |
struct backing_dev_info *bdi = mapping->backing_dev_info; |
1da177e4c Linux-2.6.12-rc2 |
1234 |
unsigned long flags; |
19fd62312 mm: spinlock tree... |
1235 |
spin_lock_irqsave(&mapping->tree_lock, flags); |
1da177e4c Linux-2.6.12-rc2 |
1236 |
ret = TestClearPageWriteback(page); |
69cb51d18 mm: count writeba... |
1237 |
if (ret) { |
1da177e4c Linux-2.6.12-rc2 |
1238 1239 1240 |
radix_tree_tag_clear(&mapping->page_tree, page_index(page), PAGECACHE_TAG_WRITEBACK); |
e4ad08fe6 mm: bdi: add sepa... |
1241 |
if (bdi_cap_account_writeback(bdi)) { |
69cb51d18 mm: count writeba... |
1242 |
__dec_bdi_stat(bdi, BDI_WRITEBACK); |
04fbfdc14 mm: per device di... |
1243 1244 |
__bdi_writeout_inc(bdi); } |
69cb51d18 mm: count writeba... |
1245 |
} |
19fd62312 mm: spinlock tree... |
1246 |
spin_unlock_irqrestore(&mapping->tree_lock, flags); |
1da177e4c Linux-2.6.12-rc2 |
1247 1248 1249 |
} else { ret = TestClearPageWriteback(page); } |
d688abf50 move page writeba... |
1250 1251 |
if (ret) dec_zone_page_state(page, NR_WRITEBACK); |
1da177e4c Linux-2.6.12-rc2 |
1252 1253 1254 1255 1256 1257 1258 1259 1260 |
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... |
1261 |
struct backing_dev_info *bdi = mapping->backing_dev_info; |
1da177e4c Linux-2.6.12-rc2 |
1262 |
unsigned long flags; |
19fd62312 mm: spinlock tree... |
1263 |
spin_lock_irqsave(&mapping->tree_lock, flags); |
1da177e4c Linux-2.6.12-rc2 |
1264 |
ret = TestSetPageWriteback(page); |
69cb51d18 mm: count writeba... |
1265 |
if (!ret) { |
1da177e4c Linux-2.6.12-rc2 |
1266 1267 1268 |
radix_tree_tag_set(&mapping->page_tree, page_index(page), PAGECACHE_TAG_WRITEBACK); |
e4ad08fe6 mm: bdi: add sepa... |
1269 |
if (bdi_cap_account_writeback(bdi)) |
69cb51d18 mm: count writeba... |
1270 1271 |
__inc_bdi_stat(bdi, BDI_WRITEBACK); } |
1da177e4c Linux-2.6.12-rc2 |
1272 1273 1274 1275 |
if (!PageDirty(page)) radix_tree_tag_clear(&mapping->page_tree, page_index(page), PAGECACHE_TAG_DIRTY); |
19fd62312 mm: spinlock tree... |
1276 |
spin_unlock_irqrestore(&mapping->tree_lock, flags); |
1da177e4c Linux-2.6.12-rc2 |
1277 1278 1279 |
} else { ret = TestSetPageWriteback(page); } |
d688abf50 move page writeba... |
1280 1281 |
if (!ret) inc_zone_page_state(page, NR_WRITEBACK); |
1da177e4c Linux-2.6.12-rc2 |
1282 1283 1284 1285 1286 1287 |
return ret; } EXPORT_SYMBOL(test_set_page_writeback); /* |
001281881 mm: use lockless ... |
1288 |
* Return true if any of the pages in the mapping are marked with the |
1da177e4c Linux-2.6.12-rc2 |
1289 1290 1291 1292 |
* passed tag. */ int mapping_tagged(struct address_space *mapping, int tag) { |
1da177e4c Linux-2.6.12-rc2 |
1293 |
int ret; |
001281881 mm: use lockless ... |
1294 |
rcu_read_lock(); |
1da177e4c Linux-2.6.12-rc2 |
1295 |
ret = radix_tree_tagged(&mapping->page_tree, tag); |
001281881 mm: use lockless ... |
1296 |
rcu_read_unlock(); |
1da177e4c Linux-2.6.12-rc2 |
1297 1298 1299 |
return ret; } EXPORT_SYMBOL(mapping_tagged); |