page-writeback.c 67.4 KB
Newer Older
L
Linus Torvalds 已提交
1
/*
2
 * mm/page-writeback.c
L
Linus Torvalds 已提交
3 4
 *
 * Copyright (C) 2002, Linus Torvalds.
P
Peter Zijlstra 已提交
5
 * Copyright (C) 2007 Red Hat, Inc., Peter Zijlstra <pzijlstr@redhat.com>
L
Linus Torvalds 已提交
6 7 8 9
 *
 * Contains functions related to writing back dirty pages at the
 * address_space level.
 *
10
 * 10Apr2002	Andrew Morton
L
Linus Torvalds 已提交
11 12 13 14
 *		Initial version
 */

#include <linux/kernel.h>
15
#include <linux/export.h>
L
Linus Torvalds 已提交
16 17 18 19 20 21 22 23 24
#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>
25
#include <linux/task_io_accounting_ops.h>
L
Linus Torvalds 已提交
26 27
#include <linux/blkdev.h>
#include <linux/mpage.h>
28
#include <linux/rmap.h>
L
Linus Torvalds 已提交
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>
A
Al Viro 已提交
35
#include <linux/buffer_head.h> /* __set_page_dirty_buffers */
36
#include <linux/pagevec.h>
37
#include <trace/events/writeback.h>
L
Linus Torvalds 已提交
38

39 40 41 42 43
/*
 * Sleep at most 200ms at a time in balance_dirty_pages().
 */
#define MAX_PAUSE		max(HZ/5, 1)

44 45 46 47 48 49
/*
 * Try to keep balance_dirty_pages() call intervals higher than this many pages
 * by raising pause time to max_pause when falls below it.
 */
#define DIRTY_POLL_THRESH	(128 >> (PAGE_SHIFT - 10))

50 51 52 53 54
/*
 * Estimate write bandwidth at 200ms intervals.
 */
#define BANDWIDTH_INTERVAL	max(HZ/5, 1)

W
Wu Fengguang 已提交
55 56
#define RATELIMIT_CALC_SHIFT	10

L
Linus Torvalds 已提交
57 58 59 60 61 62 63 64 65
/*
 * 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;

/* The following parameters are exported via /proc/sys/vm */

/*
66
 * Start background writeback (via writeback threads) at this percentage
L
Linus Torvalds 已提交
67
 */
68
int dirty_background_ratio = 10;
L
Linus Torvalds 已提交
69

70 71 72 73 74 75
/*
 * 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;

76 77 78 79 80 81
/*
 * 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;

L
Linus Torvalds 已提交
82 83 84
/*
 * The generator of dirty data starts writeback at this percentage
 */
85
int vm_dirty_ratio = 20;
L
Linus Torvalds 已提交
86

87 88 89 90 91 92
/*
 * 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;

L
Linus Torvalds 已提交
93
/*
94
 * The interval between `kupdate'-style writebacks
L
Linus Torvalds 已提交
95
 */
96
unsigned int dirty_writeback_interval = 5 * 100; /* centiseconds */
L
Linus Torvalds 已提交
97

98 99
EXPORT_SYMBOL_GPL(dirty_writeback_interval);

L
Linus Torvalds 已提交
100
/*
101
 * The longest time for which data is allowed to remain dirty
L
Linus Torvalds 已提交
102
 */
103
unsigned int dirty_expire_interval = 30 * 100; /* centiseconds */
L
Linus Torvalds 已提交
104 105 106 107 108 109 110

/*
 * Flag that makes the machine dump writes/reads and block dirtyings.
 */
int block_dump;

/*
111 112
 * 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.
L
Linus Torvalds 已提交
113 114 115 116 117 118 119
 */
int laptop_mode;

EXPORT_SYMBOL(laptop_mode);

/* End of sysctl-exported parameters */

120
unsigned long global_dirty_limit;
L
Linus Torvalds 已提交
121

P
Peter Zijlstra 已提交
122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139
/*
 * 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;

140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156
/*
 * 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.
 */
157

158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175
/*
 * In a memory zone, there is a certain amount of pages we consider
 * available for the page cache, which is essentially the number of
 * free and reclaimable pages, minus some zone reserves to protect
 * lowmem and the ability to uphold the zone's watermarks without
 * requiring writeback.
 *
 * This number of dirtyable pages is the base value of which the
 * user-configurable dirty ratio is the effictive number of pages that
 * are allowed to be actually dirtied.  Per individual zone, or
 * globally by using the sum of dirtyable pages over all zones.
 *
 * Because the user is allowed to specify the dirty limit globally as
 * absolute number of bytes, calculating the per-zone dirty limit can
 * require translating the configured limit into a percentage of
 * global dirtyable memory first.
 */

176 177 178 179 180 181 182 183 184 185 186
static unsigned long highmem_dirtyable_memory(unsigned long total)
{
#ifdef CONFIG_HIGHMEM
	int node;
	unsigned long x = 0;

	for_each_node_state(node, N_HIGH_MEMORY) {
		struct zone *z =
			&NODE_DATA(node)->node_zones[ZONE_HIGHMEM];

		x += zone_page_state(z, NR_FREE_PAGES) +
187
		     zone_reclaimable_pages(z) - z->dirty_balance_reserve;
188 189 190 191 192 193 194 195 196 197 198 199 200 201
	}
	/*
	 * 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
}

/**
202
 * global_dirtyable_memory - number of globally dirtyable pages
203
 *
204 205
 * Returns the global number of pages potentially available for dirty
 * page cache.  This is the base value for the global dirty limits.
206
 */
207
unsigned long global_dirtyable_memory(void)
208 209 210
{
	unsigned long x;

211 212
	x = global_page_state(NR_FREE_PAGES) + global_reclaimable_pages() -
	    dirty_balance_reserve;
213 214 215 216 217 218 219

	if (!vm_highmem_is_dirtyable)
		x -= highmem_dirtyable_memory(x);

	return x + 1;	/* Ensure that we never return 0 */
}

220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260
/*
 * 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
 * real-time tasks.
 */
void global_dirty_limits(unsigned long *pbackground, unsigned long *pdirty)
{
	unsigned long background;
	unsigned long dirty;
	unsigned long uninitialized_var(available_memory);
	struct task_struct *tsk;

	if (!vm_dirty_bytes || !dirty_background_bytes)
		available_memory = global_dirtyable_memory();

	if (vm_dirty_bytes)
		dirty = DIV_ROUND_UP(vm_dirty_bytes, PAGE_SIZE);
	else
		dirty = (vm_dirty_ratio * available_memory) / 100;

	if (dirty_background_bytes)
		background = DIV_ROUND_UP(dirty_background_bytes, PAGE_SIZE);
	else
		background = (dirty_background_ratio * available_memory) / 100;

	if (background >= dirty)
		background = dirty / 2;
	tsk = current;
	if (tsk->flags & PF_LESS_THROTTLE || rt_task(tsk)) {
		background += background / 4;
		dirty += dirty / 4;
	}
	*pbackground = background;
	*pdirty = dirty;
	trace_global_dirty_state(background, dirty);
}

261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324
/**
 * zone_dirtyable_memory - number of dirtyable pages in a zone
 * @zone: the zone
 *
 * Returns the zone's number of pages potentially available for dirty
 * page cache.  This is the base value for the per-zone dirty limits.
 */
static unsigned long zone_dirtyable_memory(struct zone *zone)
{
	/*
	 * The effective global number of dirtyable pages may exclude
	 * highmem as a big-picture measure to keep the ratio between
	 * dirty memory and lowmem reasonable.
	 *
	 * But this function is purely about the individual zone and a
	 * highmem zone can hold its share of dirty pages, so we don't
	 * care about vm_highmem_is_dirtyable here.
	 */
	return zone_page_state(zone, NR_FREE_PAGES) +
	       zone_reclaimable_pages(zone) -
	       zone->dirty_balance_reserve;
}

/**
 * zone_dirty_limit - maximum number of dirty pages allowed in a zone
 * @zone: the zone
 *
 * Returns the maximum number of dirty pages allowed in a zone, based
 * on the zone's dirtyable memory.
 */
static unsigned long zone_dirty_limit(struct zone *zone)
{
	unsigned long zone_memory = zone_dirtyable_memory(zone);
	struct task_struct *tsk = current;
	unsigned long dirty;

	if (vm_dirty_bytes)
		dirty = DIV_ROUND_UP(vm_dirty_bytes, PAGE_SIZE) *
			zone_memory / global_dirtyable_memory();
	else
		dirty = vm_dirty_ratio * zone_memory / 100;

	if (tsk->flags & PF_LESS_THROTTLE || rt_task(tsk))
		dirty += dirty / 4;

	return dirty;
}

/**
 * zone_dirty_ok - tells whether a zone is within its dirty limits
 * @zone: the zone to check
 *
 * Returns %true when the dirty pages in @zone are within the zone's
 * dirty limit, %false if the limit is exceeded.
 */
bool zone_dirty_ok(struct zone *zone)
{
	unsigned long limit = zone_dirty_limit(zone);

	return zone_page_state(zone, NR_FILE_DIRTY) +
	       zone_page_state(zone, NR_UNSTABLE_NFS) +
	       zone_page_state(zone, NR_WRITEBACK) <= limit;
}

P
Peter Zijlstra 已提交
325 326 327 328 329 330 331 332 333
/*
 * 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;

334 335 336
	if (vm_dirty_bytes)
		dirty_total = vm_dirty_bytes / PAGE_SIZE;
	else
337
		dirty_total = (vm_dirty_ratio * global_dirtyable_memory()) /
338
				100;
P
Peter Zijlstra 已提交
339 340 341 342
	return 2 + ilog2(dirty_total - 1);
}

/*
343
 * update the period when the dirty threshold changes.
P
Peter Zijlstra 已提交
344
 */
345 346 347 348
static void update_completion_period(void)
{
	int shift = calc_period_shift();
	prop_change_shift(&vm_completions, shift);
349 350

	writeback_set_ratelimit();
351 352 353
}

int dirty_background_ratio_handler(struct ctl_table *table, int write,
354
		void __user *buffer, size_t *lenp,
355 356 357 358
		loff_t *ppos)
{
	int ret;

359
	ret = proc_dointvec_minmax(table, write, buffer, lenp, ppos);
360 361 362 363 364 365
	if (ret == 0 && write)
		dirty_background_bytes = 0;
	return ret;
}

int dirty_background_bytes_handler(struct ctl_table *table, int write,
366
		void __user *buffer, size_t *lenp,
367 368 369 370
		loff_t *ppos)
{
	int ret;

371
	ret = proc_doulongvec_minmax(table, write, buffer, lenp, ppos);
372 373 374 375 376
	if (ret == 0 && write)
		dirty_background_ratio = 0;
	return ret;
}

P
Peter Zijlstra 已提交
377
int dirty_ratio_handler(struct ctl_table *table, int write,
378
		void __user *buffer, size_t *lenp,
P
Peter Zijlstra 已提交
379 380 381
		loff_t *ppos)
{
	int old_ratio = vm_dirty_ratio;
382 383
	int ret;

384
	ret = proc_dointvec_minmax(table, write, buffer, lenp, ppos);
P
Peter Zijlstra 已提交
385
	if (ret == 0 && write && vm_dirty_ratio != old_ratio) {
386 387 388 389 390 391 392
		update_completion_period();
		vm_dirty_bytes = 0;
	}
	return ret;
}

int dirty_bytes_handler(struct ctl_table *table, int write,
393
		void __user *buffer, size_t *lenp,
394 395
		loff_t *ppos)
{
396
	unsigned long old_bytes = vm_dirty_bytes;
397 398
	int ret;

399
	ret = proc_doulongvec_minmax(table, write, buffer, lenp, ppos);
400 401 402
	if (ret == 0 && write && vm_dirty_bytes != old_bytes) {
		update_completion_period();
		vm_dirty_ratio = 0;
P
Peter Zijlstra 已提交
403 404 405 406 407 408 409 410 411 412
	}
	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)
{
413
	__inc_bdi_stat(bdi, BDI_WRITTEN);
414 415
	__prop_inc_percpu_max(&vm_completions, &bdi->completions,
			      bdi->max_prop_frac);
P
Peter Zijlstra 已提交
416 417
}

418 419 420 421 422 423 424 425 426 427
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);

P
Peter Zijlstra 已提交
428 429 430 431 432 433
/*
 * Obtain an accurate fraction of the BDI's portion.
 */
static void bdi_writeout_fraction(struct backing_dev_info *bdi,
		long *numerator, long *denominator)
{
434
	prop_fraction_percpu(&vm_completions, &bdi->completions,
P
Peter Zijlstra 已提交
435 436 437
				numerator, denominator);
}

438
/*
439 440 441
 * bdi_min_ratio keeps the sum of the minimum dirty shares of all
 * registered backing devices, which, for obvious reasons, can not
 * exceed 100%.
442 443 444 445 446 447 448
 */
static unsigned int bdi_min_ratio;

int bdi_set_min_ratio(struct backing_dev_info *bdi, unsigned int min_ratio)
{
	int ret = 0;

449
	spin_lock_bh(&bdi_lock);
450
	if (min_ratio > bdi->max_ratio) {
451
		ret = -EINVAL;
452 453 454 455 456 457 458 459 460
	} 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;
		}
	}
461
	spin_unlock_bh(&bdi_lock);
462 463 464 465 466 467 468 469 470 471 472

	return ret;
}

int bdi_set_max_ratio(struct backing_dev_info *bdi, unsigned max_ratio)
{
	int ret = 0;

	if (max_ratio > 100)
		return -EINVAL;

473
	spin_lock_bh(&bdi_lock);
474 475 476 477 478 479
	if (bdi->min_ratio > max_ratio) {
		ret = -EINVAL;
	} else {
		bdi->max_ratio = max_ratio;
		bdi->max_prop_frac = (PROP_FRAC_BASE * max_ratio) / 100;
	}
480
	spin_unlock_bh(&bdi_lock);
481 482 483

	return ret;
}
484
EXPORT_SYMBOL(bdi_set_max_ratio);
485

W
Wu Fengguang 已提交
486 487 488 489 490 491
static unsigned long dirty_freerun_ceiling(unsigned long thresh,
					   unsigned long bg_thresh)
{
	return (thresh + bg_thresh) / 2;
}

492 493 494 495 496
static unsigned long hard_dirty_limit(unsigned long thresh)
{
	return max(thresh, global_dirty_limit);
}

497
/**
498
 * bdi_dirty_limit - @bdi's share of dirty throttling threshold
499 500
 * @bdi: the backing_dev_info to query
 * @dirty: global dirty limit in pages
501
 *
502 503
 * 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.
504 505 506 507 508 509 510
 *
 * Note that balance_dirty_pages() will only seriously take it as a hard limit
 * when sleeping max_pause per page is not enough to keep the dirty pages under
 * control. For example, when the device is completely stalled due to some error
 * conditions, or when there are 1000 dd tasks writing to a slow 10MB/s USB key.
 * In the other normal situations, it acts more gently by throttling the tasks
 * more (rather than completely block them) when the bdi dirty pages go high.
511
 *
512
 * It allocates high/low dirty limits to fast/slow devices, in order to prevent
513 514 515 516 517 518 519
 * - 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)
520 521 522
{
	u64 bdi_dirty;
	long numerator, denominator;
P
Peter Zijlstra 已提交
523

524 525 526 527
	/*
	 * Calculate this BDI's share of the dirty ratio.
	 */
	bdi_writeout_fraction(bdi, &numerator, &denominator);
P
Peter Zijlstra 已提交
528

529 530 531
	bdi_dirty = (dirty * (100 - bdi_min_ratio)) / 100;
	bdi_dirty *= numerator;
	do_div(bdi_dirty, denominator);
P
Peter Zijlstra 已提交
532

533 534 535 536 537
	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;
L
Linus Torvalds 已提交
538 539
}

W
Wu Fengguang 已提交
540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691
/*
 * Dirty position control.
 *
 * (o) global/bdi setpoints
 *
 * We want the dirty pages be balanced around the global/bdi setpoints.
 * When the number of dirty pages is higher/lower than the setpoint, the
 * dirty position control ratio (and hence task dirty ratelimit) will be
 * decreased/increased to bring the dirty pages back to the setpoint.
 *
 *     pos_ratio = 1 << RATELIMIT_CALC_SHIFT
 *
 *     if (dirty < setpoint) scale up   pos_ratio
 *     if (dirty > setpoint) scale down pos_ratio
 *
 *     if (bdi_dirty < bdi_setpoint) scale up   pos_ratio
 *     if (bdi_dirty > bdi_setpoint) scale down pos_ratio
 *
 *     task_ratelimit = dirty_ratelimit * pos_ratio >> RATELIMIT_CALC_SHIFT
 *
 * (o) global control line
 *
 *     ^ pos_ratio
 *     |
 *     |            |<===== global dirty control scope ======>|
 * 2.0 .............*
 *     |            .*
 *     |            . *
 *     |            .   *
 *     |            .     *
 *     |            .        *
 *     |            .            *
 * 1.0 ................................*
 *     |            .                  .     *
 *     |            .                  .          *
 *     |            .                  .              *
 *     |            .                  .                 *
 *     |            .                  .                    *
 *   0 +------------.------------------.----------------------*------------->
 *           freerun^          setpoint^                 limit^   dirty pages
 *
 * (o) bdi control line
 *
 *     ^ pos_ratio
 *     |
 *     |            *
 *     |              *
 *     |                *
 *     |                  *
 *     |                    * |<=========== span ============>|
 * 1.0 .......................*
 *     |                      . *
 *     |                      .   *
 *     |                      .     *
 *     |                      .       *
 *     |                      .         *
 *     |                      .           *
 *     |                      .             *
 *     |                      .               *
 *     |                      .                 *
 *     |                      .                   *
 *     |                      .                     *
 * 1/4 ...............................................* * * * * * * * * * * *
 *     |                      .                         .
 *     |                      .                           .
 *     |                      .                             .
 *   0 +----------------------.-------------------------------.------------->
 *                bdi_setpoint^                    x_intercept^
 *
 * The bdi control line won't drop below pos_ratio=1/4, so that bdi_dirty can
 * be smoothly throttled down to normal if it starts high in situations like
 * - start writing to a slow SD card and a fast disk at the same time. The SD
 *   card's bdi_dirty may rush to many times higher than bdi_setpoint.
 * - the bdi dirty thresh drops quickly due to change of JBOD workload
 */
static unsigned long bdi_position_ratio(struct backing_dev_info *bdi,
					unsigned long thresh,
					unsigned long bg_thresh,
					unsigned long dirty,
					unsigned long bdi_thresh,
					unsigned long bdi_dirty)
{
	unsigned long write_bw = bdi->avg_write_bandwidth;
	unsigned long freerun = dirty_freerun_ceiling(thresh, bg_thresh);
	unsigned long limit = hard_dirty_limit(thresh);
	unsigned long x_intercept;
	unsigned long setpoint;		/* dirty pages' target balance point */
	unsigned long bdi_setpoint;
	unsigned long span;
	long long pos_ratio;		/* for scaling up/down the rate limit */
	long x;

	if (unlikely(dirty >= limit))
		return 0;

	/*
	 * global setpoint
	 *
	 *                           setpoint - dirty 3
	 *        f(dirty) := 1.0 + (----------------)
	 *                           limit - setpoint
	 *
	 * it's a 3rd order polynomial that subjects to
	 *
	 * (1) f(freerun)  = 2.0 => rampup dirty_ratelimit reasonably fast
	 * (2) f(setpoint) = 1.0 => the balance point
	 * (3) f(limit)    = 0   => the hard limit
	 * (4) df/dx      <= 0	 => negative feedback control
	 * (5) the closer to setpoint, the smaller |df/dx| (and the reverse)
	 *     => fast response on large errors; small oscillation near setpoint
	 */
	setpoint = (freerun + limit) / 2;
	x = div_s64((setpoint - dirty) << RATELIMIT_CALC_SHIFT,
		    limit - setpoint + 1);
	pos_ratio = x;
	pos_ratio = pos_ratio * x >> RATELIMIT_CALC_SHIFT;
	pos_ratio = pos_ratio * x >> RATELIMIT_CALC_SHIFT;
	pos_ratio += 1 << RATELIMIT_CALC_SHIFT;

	/*
	 * We have computed basic pos_ratio above based on global situation. If
	 * the bdi is over/under its share of dirty pages, we want to scale
	 * pos_ratio further down/up. That is done by the following mechanism.
	 */

	/*
	 * bdi setpoint
	 *
	 *        f(bdi_dirty) := 1.0 + k * (bdi_dirty - bdi_setpoint)
	 *
	 *                        x_intercept - bdi_dirty
	 *                     := --------------------------
	 *                        x_intercept - bdi_setpoint
	 *
	 * The main bdi control line is a linear function that subjects to
	 *
	 * (1) f(bdi_setpoint) = 1.0
	 * (2) k = - 1 / (8 * write_bw)  (in single bdi case)
	 *     or equally: x_intercept = bdi_setpoint + 8 * write_bw
	 *
	 * For single bdi case, the dirty pages are observed to fluctuate
	 * regularly within range
	 *        [bdi_setpoint - write_bw/2, bdi_setpoint + write_bw/2]
	 * for various filesystems, where (2) can yield in a reasonable 12.5%
	 * fluctuation range for pos_ratio.
	 *
	 * For JBOD case, bdi_thresh (not bdi_dirty!) could fluctuate up to its
	 * own size, so move the slope over accordingly and choose a slope that
	 * yields 100% pos_ratio fluctuation on suddenly doubled bdi_thresh.
	 */
	if (unlikely(bdi_thresh > thresh))
		bdi_thresh = thresh;
692 693 694 695 696 697 698
	/*
	 * It's very possible that bdi_thresh is close to 0 not because the
	 * device is slow, but that it has remained inactive for long time.
	 * Honour such devices a reasonable good (hopefully IO efficient)
	 * threshold, so that the occasional writes won't be blocked and active
	 * writes can rampup the threshold quickly.
	 */
699
	bdi_thresh = max(bdi_thresh, (limit - dirty) / 8);
W
Wu Fengguang 已提交
700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717
	/*
	 * scale global setpoint to bdi's:
	 *	bdi_setpoint = setpoint * bdi_thresh / thresh
	 */
	x = div_u64((u64)bdi_thresh << 16, thresh + 1);
	bdi_setpoint = setpoint * (u64)x >> 16;
	/*
	 * Use span=(8*write_bw) in single bdi case as indicated by
	 * (thresh - bdi_thresh ~= 0) and transit to bdi_thresh in JBOD case.
	 *
	 *        bdi_thresh                    thresh - bdi_thresh
	 * span = ---------- * (8 * write_bw) + ------------------- * bdi_thresh
	 *          thresh                            thresh
	 */
	span = (thresh - bdi_thresh + 8 * write_bw) * (u64)x >> 16;
	x_intercept = bdi_setpoint + span;

	if (bdi_dirty < x_intercept - span / 4) {
718 719
		pos_ratio = div_u64(pos_ratio * (x_intercept - bdi_dirty),
				    x_intercept - bdi_setpoint + 1);
W
Wu Fengguang 已提交
720 721 722
	} else
		pos_ratio /= 4;

723 724 725 726 727 728 729
	/*
	 * bdi reserve area, safeguard against dirty pool underrun and disk idle
	 * It may push the desired control point of global dirty pages higher
	 * than setpoint.
	 */
	x_intercept = bdi_thresh / 2;
	if (bdi_dirty < x_intercept) {
730 731 732
		if (bdi_dirty > x_intercept / 8)
			pos_ratio = div_u64(pos_ratio * x_intercept, bdi_dirty);
		else
733 734 735
			pos_ratio *= 8;
	}

W
Wu Fengguang 已提交
736 737 738
	return pos_ratio;
}

739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778
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;
}

779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834
/*
 * 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);
}

W
Wu Fengguang 已提交
835 836 837 838 839 840 841 842 843 844 845 846 847 848 849
/*
 * Maintain bdi->dirty_ratelimit, the base dirty throttle rate.
 *
 * Normal bdi tasks will be curbed at or below it in long term.
 * Obviously it should be around (write_bw / N) when there are N dd tasks.
 */
static void bdi_update_dirty_ratelimit(struct backing_dev_info *bdi,
				       unsigned long thresh,
				       unsigned long bg_thresh,
				       unsigned long dirty,
				       unsigned long bdi_thresh,
				       unsigned long bdi_dirty,
				       unsigned long dirtied,
				       unsigned long elapsed)
{
850 851 852
	unsigned long freerun = dirty_freerun_ceiling(thresh, bg_thresh);
	unsigned long limit = hard_dirty_limit(thresh);
	unsigned long setpoint = (freerun + limit) / 2;
W
Wu Fengguang 已提交
853 854 855 856 857 858
	unsigned long write_bw = bdi->avg_write_bandwidth;
	unsigned long dirty_ratelimit = bdi->dirty_ratelimit;
	unsigned long dirty_rate;
	unsigned long task_ratelimit;
	unsigned long balanced_dirty_ratelimit;
	unsigned long pos_ratio;
859 860
	unsigned long step;
	unsigned long x;
W
Wu Fengguang 已提交
861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908

	/*
	 * The dirty rate will match the writeout rate in long term, except
	 * when dirty pages are truncated by userspace or re-dirtied by FS.
	 */
	dirty_rate = (dirtied - bdi->dirtied_stamp) * HZ / elapsed;

	pos_ratio = bdi_position_ratio(bdi, thresh, bg_thresh, dirty,
				       bdi_thresh, bdi_dirty);
	/*
	 * task_ratelimit reflects each dd's dirty rate for the past 200ms.
	 */
	task_ratelimit = (u64)dirty_ratelimit *
					pos_ratio >> RATELIMIT_CALC_SHIFT;
	task_ratelimit++; /* it helps rampup dirty_ratelimit from tiny values */

	/*
	 * A linear estimation of the "balanced" throttle rate. The theory is,
	 * if there are N dd tasks, each throttled at task_ratelimit, the bdi's
	 * dirty_rate will be measured to be (N * task_ratelimit). So the below
	 * formula will yield the balanced rate limit (write_bw / N).
	 *
	 * Note that the expanded form is not a pure rate feedback:
	 *	rate_(i+1) = rate_(i) * (write_bw / dirty_rate)		     (1)
	 * but also takes pos_ratio into account:
	 *	rate_(i+1) = rate_(i) * (write_bw / dirty_rate) * pos_ratio  (2)
	 *
	 * (1) is not realistic because pos_ratio also takes part in balancing
	 * the dirty rate.  Consider the state
	 *	pos_ratio = 0.5						     (3)
	 *	rate = 2 * (write_bw / N)				     (4)
	 * If (1) is used, it will stuck in that state! Because each dd will
	 * be throttled at
	 *	task_ratelimit = pos_ratio * rate = (write_bw / N)	     (5)
	 * yielding
	 *	dirty_rate = N * task_ratelimit = write_bw		     (6)
	 * put (6) into (1) we get
	 *	rate_(i+1) = rate_(i)					     (7)
	 *
	 * So we end up using (2) to always keep
	 *	rate_(i+1) ~= (write_bw / N)				     (8)
	 * regardless of the value of pos_ratio. As long as (8) is satisfied,
	 * pos_ratio is able to drive itself to 1.0, which is not only where
	 * the dirty count meet the setpoint, but also where the slope of
	 * pos_ratio is most flat and hence task_ratelimit is least fluctuated.
	 */
	balanced_dirty_ratelimit = div_u64((u64)task_ratelimit * write_bw,
					   dirty_rate | 1);
909 910 911 912 913
	/*
	 * balanced_dirty_ratelimit ~= (write_bw / N) <= write_bw
	 */
	if (unlikely(balanced_dirty_ratelimit > write_bw))
		balanced_dirty_ratelimit = write_bw;
W
Wu Fengguang 已提交
914

915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979
	/*
	 * We could safely do this and return immediately:
	 *
	 *	bdi->dirty_ratelimit = balanced_dirty_ratelimit;
	 *
	 * However to get a more stable dirty_ratelimit, the below elaborated
	 * code makes use of task_ratelimit to filter out sigular points and
	 * limit the step size.
	 *
	 * The below code essentially only uses the relative value of
	 *
	 *	task_ratelimit - dirty_ratelimit
	 *	= (pos_ratio - 1) * dirty_ratelimit
	 *
	 * which reflects the direction and size of dirty position error.
	 */

	/*
	 * dirty_ratelimit will follow balanced_dirty_ratelimit iff
	 * task_ratelimit is on the same side of dirty_ratelimit, too.
	 * For example, when
	 * - dirty_ratelimit > balanced_dirty_ratelimit
	 * - dirty_ratelimit > task_ratelimit (dirty pages are above setpoint)
	 * lowering dirty_ratelimit will help meet both the position and rate
	 * control targets. Otherwise, don't update dirty_ratelimit if it will
	 * only help meet the rate target. After all, what the users ultimately
	 * feel and care are stable dirty rate and small position error.
	 *
	 * |task_ratelimit - dirty_ratelimit| is used to limit the step size
	 * and filter out the sigular points of balanced_dirty_ratelimit. Which
	 * keeps jumping around randomly and can even leap far away at times
	 * due to the small 200ms estimation period of dirty_rate (we want to
	 * keep that period small to reduce time lags).
	 */
	step = 0;
	if (dirty < setpoint) {
		x = min(bdi->balanced_dirty_ratelimit,
			 min(balanced_dirty_ratelimit, task_ratelimit));
		if (dirty_ratelimit < x)
			step = x - dirty_ratelimit;
	} else {
		x = max(bdi->balanced_dirty_ratelimit,
			 max(balanced_dirty_ratelimit, task_ratelimit));
		if (dirty_ratelimit > x)
			step = dirty_ratelimit - x;
	}

	/*
	 * Don't pursue 100% rate matching. It's impossible since the balanced
	 * rate itself is constantly fluctuating. So decrease the track speed
	 * when it gets close to the target. Helps eliminate pointless tremors.
	 */
	step >>= dirty_ratelimit / (2 * step + 1);
	/*
	 * Limit the tracking speed to avoid overshooting.
	 */
	step = (step + 7) / 8;

	if (dirty_ratelimit < balanced_dirty_ratelimit)
		dirty_ratelimit += step;
	else
		dirty_ratelimit -= step;

	bdi->dirty_ratelimit = max(dirty_ratelimit, 1UL);
	bdi->balanced_dirty_ratelimit = balanced_dirty_ratelimit;
980 981

	trace_bdi_dirty_ratelimit(bdi, dirty_rate, task_ratelimit);
W
Wu Fengguang 已提交
982 983
}

984
void __bdi_update_bandwidth(struct backing_dev_info *bdi,
985
			    unsigned long thresh,
986
			    unsigned long bg_thresh,
987 988 989
			    unsigned long dirty,
			    unsigned long bdi_thresh,
			    unsigned long bdi_dirty,
990 991 992 993
			    unsigned long start_time)
{
	unsigned long now = jiffies;
	unsigned long elapsed = now - bdi->bw_time_stamp;
W
Wu Fengguang 已提交
994
	unsigned long dirtied;
995 996 997 998 999 1000 1001 1002
	unsigned long written;

	/*
	 * rate-limit, only update once every 200ms.
	 */
	if (elapsed < BANDWIDTH_INTERVAL)
		return;

W
Wu Fengguang 已提交
1003
	dirtied = percpu_counter_read(&bdi->bdi_stat[BDI_DIRTIED]);
1004 1005 1006 1007 1008 1009 1010 1011 1012
	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;

W
Wu Fengguang 已提交
1013
	if (thresh) {
1014
		global_update_bandwidth(thresh, dirty, now);
W
Wu Fengguang 已提交
1015 1016 1017 1018
		bdi_update_dirty_ratelimit(bdi, thresh, bg_thresh, dirty,
					   bdi_thresh, bdi_dirty,
					   dirtied, elapsed);
	}
1019 1020 1021
	bdi_update_write_bandwidth(bdi, elapsed, written);

snapshot:
W
Wu Fengguang 已提交
1022
	bdi->dirtied_stamp = dirtied;
1023 1024 1025 1026 1027
	bdi->written_stamp = written;
	bdi->bw_time_stamp = now;
}

static void bdi_update_bandwidth(struct backing_dev_info *bdi,
1028
				 unsigned long thresh,
1029
				 unsigned long bg_thresh,
1030 1031 1032
				 unsigned long dirty,
				 unsigned long bdi_thresh,
				 unsigned long bdi_dirty,
1033 1034 1035 1036 1037
				 unsigned long start_time)
{
	if (time_is_after_eq_jiffies(bdi->bw_time_stamp + BANDWIDTH_INTERVAL))
		return;
	spin_lock(&bdi->wb.list_lock);
1038 1039
	__bdi_update_bandwidth(bdi, thresh, bg_thresh, dirty,
			       bdi_thresh, bdi_dirty, start_time);
1040 1041 1042
	spin_unlock(&bdi->wb.list_lock);
}

1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059
/*
 * After a task dirtied this many pages, balance_dirty_pages_ratelimited_nr()
 * will look to see if it needs to start dirty throttling.
 *
 * If dirty_poll_interval is too low, big NUMA machines will call the expensive
 * global_page_state() too often. So scale it near-sqrt to the safety margin
 * (the number of pages we may dirty without exceeding the dirty limits).
 */
static unsigned long dirty_poll_interval(unsigned long dirty,
					 unsigned long thresh)
{
	if (thresh > dirty)
		return 1UL << (ilog2(thresh - dirty) >> 1);

	return 1;
}

1060 1061
static long bdi_max_pause(struct backing_dev_info *bdi,
			  unsigned long bdi_dirty)
1062
{
1063 1064
	long bw = bdi->avg_write_bandwidth;
	long t;
1065

1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083
	/*
	 * Limit pause time for small memory systems. If sleeping for too long
	 * time, a small pool of dirty/writeback pages may go empty and disk go
	 * idle.
	 *
	 * 8 serves as the safety ratio.
	 */
	t = bdi_dirty / (1 + bw / roundup_pow_of_two(1 + HZ / 8));
	t++;

	return min_t(long, t, MAX_PAUSE);
}

static long bdi_min_pause(struct backing_dev_info *bdi,
			  long max_pause,
			  unsigned long task_ratelimit,
			  unsigned long dirty_ratelimit,
			  int *nr_dirtied_pause)
1084
{
1085 1086 1087 1088 1089
	long hi = ilog2(bdi->avg_write_bandwidth);
	long lo = ilog2(bdi->dirty_ratelimit);
	long t;		/* target pause */
	long pause;	/* estimated next pause */
	int pages;	/* target nr_dirtied_pause */
1090

1091 1092
	/* target for 10ms pause on 1-dd case */
	t = max(1, HZ / 100);
1093 1094 1095 1096 1097

	/*
	 * Scale up pause time for concurrent dirtiers in order to reduce CPU
	 * overheads.
	 *
1098
	 * (N * 10ms) on 2^N concurrent tasks.
1099 1100
	 */
	if (hi > lo)
1101
		t += (hi - lo) * (10 * HZ) / 1024;
1102 1103

	/*
1104 1105 1106 1107 1108 1109 1110 1111
	 * This is a bit convoluted. We try to base the next nr_dirtied_pause
	 * on the much more stable dirty_ratelimit. However the next pause time
	 * will be computed based on task_ratelimit and the two rate limits may
	 * depart considerably at some time. Especially if task_ratelimit goes
	 * below dirty_ratelimit/2 and the target pause is max_pause, the next
	 * pause time will be max_pause*2 _trimmed down_ to max_pause.  As a
	 * result task_ratelimit won't be executed faithfully, which could
	 * eventually bring down dirty_ratelimit.
1112
	 *
1113 1114 1115 1116 1117 1118 1119
	 * We apply two rules to fix it up:
	 * 1) try to estimate the next pause time and if necessary, use a lower
	 *    nr_dirtied_pause so as not to exceed max_pause. When this happens,
	 *    nr_dirtied_pause will be "dancing" with task_ratelimit.
	 * 2) limit the target pause time to max_pause/2, so that the normal
	 *    small fluctuations of task_ratelimit won't trigger rule (1) and
	 *    nr_dirtied_pause will remain as stable as dirty_ratelimit.
1120
	 */
1121 1122
	t = min(t, 1 + max_pause / 2);
	pages = dirty_ratelimit * t / roundup_pow_of_two(HZ);
1123 1124

	/*
1125 1126 1127 1128 1129 1130
	 * Tiny nr_dirtied_pause is found to hurt I/O performance in the test
	 * case fio-mmap-randwrite-64k, which does 16*{sync read, async write}.
	 * When the 16 consecutive reads are often interrupted by some dirty
	 * throttling pause during the async writes, cfq will go into idles
	 * (deadline is fine). So push nr_dirtied_pause as high as possible
	 * until reaches DIRTY_POLL_THRESH=32 pages.
1131
	 */
1132 1133 1134 1135 1136 1137 1138 1139 1140
	if (pages < DIRTY_POLL_THRESH) {
		t = max_pause;
		pages = dirty_ratelimit * t / roundup_pow_of_two(HZ);
		if (pages > DIRTY_POLL_THRESH) {
			pages = DIRTY_POLL_THRESH;
			t = HZ * DIRTY_POLL_THRESH / dirty_ratelimit;
		}
	}

1141 1142 1143 1144 1145
	pause = HZ * pages / (task_ratelimit + 1);
	if (pause > max_pause) {
		t = max_pause;
		pages = task_ratelimit * t / roundup_pow_of_two(HZ);
	}
1146

1147
	*nr_dirtied_pause = pages;
1148
	/*
1149
	 * The minimal pause time will normally be half the target pause time.
1150
	 */
1151
	return pages >= DIRTY_POLL_THRESH ? 1 + t / 2 : t;
1152 1153
}

L
Linus Torvalds 已提交
1154 1155 1156
/*
 * 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
1157
 * the caller to wait once crossing the (background_thresh + dirty_thresh) / 2.
1158 1159
 * If we're over `background_thresh' then the writeback threads are woken to
 * perform some writeout.
L
Linus Torvalds 已提交
1160
 */
1161
static void balance_dirty_pages(struct address_space *mapping,
1162
				unsigned long pages_dirtied)
L
Linus Torvalds 已提交
1163
{
1164 1165
	unsigned long nr_reclaimable;	/* = file_dirty + unstable_nfs */
	unsigned long bdi_reclaimable;
1166 1167
	unsigned long nr_dirty;  /* = file_dirty + writeback + unstable_nfs */
	unsigned long bdi_dirty;
W
Wu Fengguang 已提交
1168
	unsigned long freerun;
1169 1170 1171
	unsigned long background_thresh;
	unsigned long dirty_thresh;
	unsigned long bdi_thresh;
1172
	long period;
1173 1174 1175 1176
	long pause;
	long max_pause;
	long min_pause;
	int nr_dirtied_pause;
1177
	bool dirty_exceeded = false;
1178
	unsigned long task_ratelimit;
1179
	unsigned long dirty_ratelimit;
1180
	unsigned long pos_ratio;
L
Linus Torvalds 已提交
1181
	struct backing_dev_info *bdi = mapping->backing_dev_info;
1182
	unsigned long start_time = jiffies;
L
Linus Torvalds 已提交
1183 1184

	for (;;) {
1185 1186
		unsigned long now = jiffies;

1187 1188 1189 1190 1191 1192
		/*
		 * 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.
		 */
1193 1194
		nr_reclaimable = global_page_state(NR_FILE_DIRTY) +
					global_page_state(NR_UNSTABLE_NFS);
1195
		nr_dirty = nr_reclaimable + global_page_state(NR_WRITEBACK);
1196

1197 1198 1199 1200 1201 1202 1203
		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.
		 */
W
Wu Fengguang 已提交
1204 1205
		freerun = dirty_freerun_ceiling(dirty_thresh,
						background_thresh);
1206 1207 1208
		if (nr_dirty <= freerun) {
			current->dirty_paused_when = now;
			current->nr_dirtied = 0;
1209 1210
			current->nr_dirtied_pause =
				dirty_poll_interval(nr_dirty, dirty_thresh);
1211
			break;
1212
		}
1213

1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229
		if (unlikely(!writeback_in_progress(bdi)))
			bdi_start_background_writeback(bdi);

		/*
		 * bdi_thresh is not treated as some limiting factor as
		 * dirty_thresh, due to reasons
		 * - in JBOD setup, bdi_thresh can fluctuate a lot
		 * - in a system with HDD and USB key, the USB key may somehow
		 *   go into state (bdi_dirty >> bdi_thresh) either because
		 *   bdi_dirty starts high, or because bdi_thresh drops low.
		 *   In this case we don't want to hard throttle the USB key
		 *   dirtiers for 100 seconds until bdi_dirty drops under
		 *   bdi_thresh. Instead the auxiliary bdi control line in
		 *   bdi_position_ratio() will let the dirtier task progress
		 *   at some rate <= (write_bw / 2) for bringing down bdi_dirty.
		 */
1230 1231
		bdi_thresh = bdi_dirty_limit(bdi, dirty_thresh);

1232 1233 1234 1235 1236 1237 1238 1239 1240 1241
		/*
		 * 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.
		 */
1242 1243 1244
		if (bdi_thresh < 2 * bdi_stat_error(bdi)) {
			bdi_reclaimable = bdi_stat_sum(bdi, BDI_RECLAIMABLE);
			bdi_dirty = bdi_reclaimable +
1245
				    bdi_stat_sum(bdi, BDI_WRITEBACK);
1246
		} else {
1247 1248
			bdi_reclaimable = bdi_stat(bdi, BDI_RECLAIMABLE);
			bdi_dirty = bdi_reclaimable +
1249
				    bdi_stat(bdi, BDI_WRITEBACK);
1250
		}
1251

1252
		dirty_exceeded = (bdi_dirty > bdi_thresh) &&
1253
				  (nr_dirty > dirty_thresh);
1254
		if (dirty_exceeded && !bdi->dirty_exceeded)
P
Peter Zijlstra 已提交
1255
			bdi->dirty_exceeded = 1;
L
Linus Torvalds 已提交
1256

1257 1258 1259
		bdi_update_bandwidth(bdi, dirty_thresh, background_thresh,
				     nr_dirty, bdi_thresh, bdi_dirty,
				     start_time);
1260

1261 1262 1263 1264
		dirty_ratelimit = bdi->dirty_ratelimit;
		pos_ratio = bdi_position_ratio(bdi, dirty_thresh,
					       background_thresh, nr_dirty,
					       bdi_thresh, bdi_dirty);
1265 1266
		task_ratelimit = ((u64)dirty_ratelimit * pos_ratio) >>
							RATELIMIT_CALC_SHIFT;
1267 1268 1269 1270 1271
		max_pause = bdi_max_pause(bdi, bdi_dirty);
		min_pause = bdi_min_pause(bdi, max_pause,
					  task_ratelimit, dirty_ratelimit,
					  &nr_dirtied_pause);

1272
		if (unlikely(task_ratelimit == 0)) {
1273
			period = max_pause;
1274
			pause = max_pause;
1275
			goto pause;
P
Peter Zijlstra 已提交
1276
		}
1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287
		period = HZ * pages_dirtied / task_ratelimit;
		pause = period;
		if (current->dirty_paused_when)
			pause -= now - current->dirty_paused_when;
		/*
		 * For less than 1s think time (ext3/4 may block the dirtier
		 * for up to 800ms from time to time on 1-HDD; so does xfs,
		 * however at much less frequency), try to compensate it in
		 * future periods by updating the virtual time; otherwise just
		 * do a reset, as it may be a light dirtier.
		 */
1288
		if (pause < min_pause) {
1289 1290 1291 1292 1293 1294 1295 1296 1297
			trace_balance_dirty_pages(bdi,
						  dirty_thresh,
						  background_thresh,
						  nr_dirty,
						  bdi_thresh,
						  bdi_dirty,
						  dirty_ratelimit,
						  task_ratelimit,
						  pages_dirtied,
1298
						  period,
1299
						  min(pause, 0L),
1300
						  start_time);
1301 1302 1303 1304 1305 1306
			if (pause < -HZ) {
				current->dirty_paused_when = now;
				current->nr_dirtied = 0;
			} else if (period) {
				current->dirty_paused_when += period;
				current->nr_dirtied = 0;
1307 1308
			} else if (current->nr_dirtied_pause <= pages_dirtied)
				current->nr_dirtied_pause += pages_dirtied;
W
Wu Fengguang 已提交
1309
			break;
P
Peter Zijlstra 已提交
1310
		}
1311 1312 1313 1314 1315
		if (unlikely(pause > max_pause)) {
			/* for occasional dropped task_ratelimit */
			now += min(pause - max_pause, max_pause);
			pause = max_pause;
		}
1316 1317

pause:
1318 1319 1320 1321 1322 1323 1324 1325 1326
		trace_balance_dirty_pages(bdi,
					  dirty_thresh,
					  background_thresh,
					  nr_dirty,
					  bdi_thresh,
					  bdi_dirty,
					  dirty_ratelimit,
					  task_ratelimit,
					  pages_dirtied,
1327
					  period,
1328 1329
					  pause,
					  start_time);
1330
		__set_current_state(TASK_KILLABLE);
1331
		io_schedule_timeout(pause);
1332

1333 1334
		current->dirty_paused_when = now + pause;
		current->nr_dirtied = 0;
1335
		current->nr_dirtied_pause = nr_dirtied_pause;
1336

1337
		/*
1338 1339
		 * This is typically equal to (nr_dirty < dirty_thresh) and can
		 * also keep "1000+ dd on a slow USB stick" under control.
1340
		 */
1341
		if (task_ratelimit)
1342
			break;
1343

1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356
		/*
		 * In the case of an unresponding NFS server and the NFS dirty
		 * pages exceeds dirty_thresh, give the other good bdi's a pipe
		 * to go through, so that tasks on them still remain responsive.
		 *
		 * In theory 1 page is enough to keep the comsumer-producer
		 * pipe going: the flusher cleans 1 page => the task dirties 1
		 * more page. However bdi_dirty has accounting errors.  So use
		 * the larger and more IO friendly bdi_stat_error.
		 */
		if (bdi_dirty <= bdi_stat_error(bdi))
			break;

1357 1358
		if (fatal_signal_pending(current))
			break;
L
Linus Torvalds 已提交
1359 1360
	}

1361
	if (!dirty_exceeded && bdi->dirty_exceeded)
P
Peter Zijlstra 已提交
1362
		bdi->dirty_exceeded = 0;
L
Linus Torvalds 已提交
1363 1364

	if (writeback_in_progress(bdi))
1365
		return;
L
Linus Torvalds 已提交
1366 1367 1368 1369 1370 1371 1372 1373 1374

	/*
	 * 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.
	 */
1375 1376 1377 1378
	if (laptop_mode)
		return;

	if (nr_reclaimable > background_thresh)
1379
		bdi_start_background_writeback(bdi);
L
Linus Torvalds 已提交
1380 1381
}

1382
void set_page_dirty_balance(struct page *page, int page_mkwrite)
P
Peter Zijlstra 已提交
1383
{
1384
	if (set_page_dirty(page) || page_mkwrite) {
P
Peter Zijlstra 已提交
1385 1386 1387 1388 1389 1390 1391
		struct address_space *mapping = page_mapping(page);

		if (mapping)
			balance_dirty_pages_ratelimited(mapping);
	}
}

1392
static DEFINE_PER_CPU(int, bdp_ratelimits);
1393

1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409
/*
 * Normal tasks are throttled by
 *	loop {
 *		dirty tsk->nr_dirtied_pause pages;
 *		take a snap in balance_dirty_pages();
 *	}
 * However there is a worst case. If every task exit immediately when dirtied
 * (tsk->nr_dirtied_pause - 1) pages, balance_dirty_pages() will never be
 * called to throttle the page dirties. The solution is to save the not yet
 * throttled page dirties in dirty_throttle_leaks on task exit and charge them
 * randomly into the running tasks. This works well for the above worst case,
 * as the new task will pick up and accumulate the old task's leaked dirty
 * count and eventually get throttled.
 */
DEFINE_PER_CPU(int, dirty_throttle_leaks) = 0;

L
Linus Torvalds 已提交
1410
/**
1411
 * balance_dirty_pages_ratelimited_nr - balance dirty memory state
1412
 * @mapping: address_space which was dirtied
1413
 * @nr_pages_dirtied: number of pages which the caller has just dirtied
L
Linus Torvalds 已提交
1414 1415 1416 1417 1418 1419 1420 1421 1422 1423
 *
 * 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.
 */
1424 1425
void balance_dirty_pages_ratelimited_nr(struct address_space *mapping,
					unsigned long nr_pages_dirtied)
L
Linus Torvalds 已提交
1426
{
1427
	struct backing_dev_info *bdi = mapping->backing_dev_info;
1428 1429
	int ratelimit;
	int *p;
L
Linus Torvalds 已提交
1430

1431 1432 1433
	if (!bdi_cap_account_dirty(bdi))
		return;

1434 1435 1436 1437 1438
	ratelimit = current->nr_dirtied_pause;
	if (bdi->dirty_exceeded)
		ratelimit = min(ratelimit, 32 >> (PAGE_SHIFT - 10));

	preempt_disable();
L
Linus Torvalds 已提交
1439
	/*
1440 1441 1442 1443
	 * This prevents one CPU to accumulate too many dirtied pages without
	 * calling into balance_dirty_pages(), which can happen when there are
	 * 1000+ tasks, all of them start dirtying pages at exactly the same
	 * time, hence all honoured too large initial task->nr_dirtied_pause.
L
Linus Torvalds 已提交
1444
	 */
1445
	p =  &__get_cpu_var(bdp_ratelimits);
1446
	if (unlikely(current->nr_dirtied >= ratelimit))
1447
		*p = 0;
1448 1449 1450
	else if (unlikely(*p >= ratelimit_pages)) {
		*p = 0;
		ratelimit = 0;
L
Linus Torvalds 已提交
1451
	}
1452 1453 1454 1455 1456 1457 1458 1459 1460 1461
	/*
	 * Pick up the dirtied pages by the exited tasks. This avoids lots of
	 * short-lived tasks (eg. gcc invocations in a kernel build) escaping
	 * the dirty throttling and livelock other long-run dirtiers.
	 */
	p = &__get_cpu_var(dirty_throttle_leaks);
	if (*p > 0 && current->nr_dirtied < ratelimit) {
		nr_pages_dirtied = min(*p, ratelimit - current->nr_dirtied);
		*p -= nr_pages_dirtied;
		current->nr_dirtied += nr_pages_dirtied;
L
Linus Torvalds 已提交
1462
	}
1463
	preempt_enable();
1464 1465 1466

	if (unlikely(current->nr_dirtied >= ratelimit))
		balance_dirty_pages(mapping, current->nr_dirtied);
L
Linus Torvalds 已提交
1467
}
1468
EXPORT_SYMBOL(balance_dirty_pages_ratelimited_nr);
L
Linus Torvalds 已提交
1469

1470
void throttle_vm_writeout(gfp_t gfp_mask)
L
Linus Torvalds 已提交
1471
{
1472 1473
	unsigned long background_thresh;
	unsigned long dirty_thresh;
L
Linus Torvalds 已提交
1474 1475

        for ( ; ; ) {
1476
		global_dirty_limits(&background_thresh, &dirty_thresh);
1477
		dirty_thresh = hard_dirty_limit(dirty_thresh);
L
Linus Torvalds 已提交
1478 1479 1480 1481 1482 1483 1484

                /*
                 * 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... */

1485 1486 1487
                if (global_page_state(NR_UNSTABLE_NFS) +
			global_page_state(NR_WRITEBACK) <= dirty_thresh)
                        	break;
1488
                congestion_wait(BLK_RW_ASYNC, HZ/10);
1489 1490 1491 1492 1493 1494 1495 1496

		/*
		 * 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;
L
Linus Torvalds 已提交
1497 1498 1499 1500 1501 1502 1503
        }
}

/*
 * sysctl handler for /proc/sys/vm/dirty_writeback_centisecs
 */
int dirty_writeback_centisecs_handler(ctl_table *table, int write,
1504
	void __user *buffer, size_t *length, loff_t *ppos)
L
Linus Torvalds 已提交
1505
{
1506
	proc_dointvec(table, write, buffer, length, ppos);
1507
	bdi_arm_supers_timer();
L
Linus Torvalds 已提交
1508 1509 1510
	return 0;
}

1511
#ifdef CONFIG_BLOCK
1512
void laptop_mode_timer_fn(unsigned long data)
L
Linus Torvalds 已提交
1513
{
1514 1515 1516
	struct request_queue *q = (struct request_queue *)data;
	int nr_pages = global_page_state(NR_FILE_DIRTY) +
		global_page_state(NR_UNSTABLE_NFS);
L
Linus Torvalds 已提交
1517

1518 1519 1520 1521 1522
	/*
	 * We want to write everything out, not just down to the dirty
	 * threshold
	 */
	if (bdi_has_dirty_io(&q->backing_dev_info))
1523 1524
		bdi_start_writeback(&q->backing_dev_info, nr_pages,
					WB_REASON_LAPTOP_TIMER);
L
Linus Torvalds 已提交
1525 1526 1527 1528 1529 1530 1531
}

/*
 * 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.
 */
1532
void laptop_io_completion(struct backing_dev_info *info)
L
Linus Torvalds 已提交
1533
{
1534
	mod_timer(&info->laptop_mode_wb_timer, jiffies + laptop_mode);
L
Linus Torvalds 已提交
1535 1536 1537 1538 1539 1540 1541 1542 1543
}

/*
 * 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)
{
1544 1545 1546 1547 1548 1549 1550 1551
	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();
L
Linus Torvalds 已提交
1552
}
1553
#endif
L
Linus Torvalds 已提交
1554 1555 1556 1557 1558 1559 1560 1561 1562

/*
 * 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
1563
 * thresholds.
L
Linus Torvalds 已提交
1564 1565
 */

1566
void writeback_set_ratelimit(void)
L
Linus Torvalds 已提交
1567
{
1568 1569 1570 1571
	unsigned long background_thresh;
	unsigned long dirty_thresh;
	global_dirty_limits(&background_thresh, &dirty_thresh);
	ratelimit_pages = dirty_thresh / (num_online_cpus() * 32);
L
Linus Torvalds 已提交
1572 1573 1574 1575
	if (ratelimit_pages < 16)
		ratelimit_pages = 16;
}

1576
static int __cpuinit
L
Linus Torvalds 已提交
1577 1578
ratelimit_handler(struct notifier_block *self, unsigned long u, void *v)
{
1579
	writeback_set_ratelimit();
1580
	return NOTIFY_DONE;
L
Linus Torvalds 已提交
1581 1582
}

1583
static struct notifier_block __cpuinitdata ratelimit_nb = {
L
Linus Torvalds 已提交
1584 1585 1586 1587 1588
	.notifier_call	= ratelimit_handler,
	.next		= NULL,
};

/*
1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604
 * 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..
L
Linus Torvalds 已提交
1605 1606 1607
 */
void __init page_writeback_init(void)
{
P
Peter Zijlstra 已提交
1608 1609
	int shift;

1610
	writeback_set_ratelimit();
L
Linus Torvalds 已提交
1611
	register_cpu_notifier(&ratelimit_nb);
P
Peter Zijlstra 已提交
1612 1613 1614

	shift = calc_period_shift();
	prop_descriptor_init(&vm_completions, shift);
L
Linus Torvalds 已提交
1615 1616
}

1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636
/**
 * 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.
 */
void tag_pages_for_writeback(struct address_space *mapping,
			     pgoff_t start, pgoff_t end)
{
R
Randy Dunlap 已提交
1637
#define WRITEBACK_TAG_BATCH 4096
1638 1639 1640 1641 1642 1643 1644 1645 1646 1647
	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();
1648 1649
		/* We check 'start' to handle wrapping when end == ~0UL */
	} while (tagged >= WRITEBACK_TAG_BATCH && start);
1650 1651 1652
}
EXPORT_SYMBOL(tag_pages_for_writeback);

1653
/**
1654
 * write_cache_pages - walk the list of dirty pages of the given address space and write all of them.
1655 1656
 * @mapping: address space structure to write
 * @wbc: subtract the number of written pages from *@wbc->nr_to_write
1657 1658
 * @writepage: function called for each page
 * @data: data passed to writepage function
1659
 *
1660
 * If a page is already under I/O, write_cache_pages() skips it, even
1661 1662 1663 1664 1665 1666
 * 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.
1667 1668 1669 1670 1671 1672 1673
 *
 * 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).
1674
 */
1675 1676 1677
int write_cache_pages(struct address_space *mapping,
		      struct writeback_control *wbc, writepage_t writepage,
		      void *data)
1678 1679 1680 1681 1682
{
	int ret = 0;
	int done = 0;
	struct pagevec pvec;
	int nr_pages;
N
Nick Piggin 已提交
1683
	pgoff_t uninitialized_var(writeback_index);
1684 1685
	pgoff_t index;
	pgoff_t end;		/* Inclusive */
1686
	pgoff_t done_index;
N
Nick Piggin 已提交
1687
	int cycled;
1688
	int range_whole = 0;
1689
	int tag;
1690 1691 1692

	pagevec_init(&pvec, 0);
	if (wbc->range_cyclic) {
N
Nick Piggin 已提交
1693 1694 1695 1696 1697 1698
		writeback_index = mapping->writeback_index; /* prev offset */
		index = writeback_index;
		if (index == 0)
			cycled = 1;
		else
			cycled = 0;
1699 1700 1701 1702 1703 1704
		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;
N
Nick Piggin 已提交
1705
		cycled = 1; /* ignore range_cyclic tests */
1706
	}
1707
	if (wbc->sync_mode == WB_SYNC_ALL || wbc->tagged_writepages)
1708 1709 1710
		tag = PAGECACHE_TAG_TOWRITE;
	else
		tag = PAGECACHE_TAG_DIRTY;
1711
retry:
1712
	if (wbc->sync_mode == WB_SYNC_ALL || wbc->tagged_writepages)
1713
		tag_pages_for_writeback(mapping, index, end);
1714
	done_index = index;
N
Nick Piggin 已提交
1715 1716 1717
	while (!done && (index <= end)) {
		int i;

1718
		nr_pages = pagevec_lookup_tag(&pvec, mapping, &index, tag,
N
Nick Piggin 已提交
1719 1720 1721
			      min(end - index, (pgoff_t)PAGEVEC_SIZE-1) + 1);
		if (nr_pages == 0)
			break;
1722 1723 1724 1725 1726

		for (i = 0; i < nr_pages; i++) {
			struct page *page = pvec.pages[i];

			/*
1727 1728 1729 1730 1731
			 * 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.
1732
			 */
1733 1734 1735 1736 1737 1738 1739 1740 1741
			if (page->index > end) {
				/*
				 * can't be range_cyclic (1st pass) because
				 * end == -1 in that case.
				 */
				done = 1;
				break;
			}

1742
			done_index = page->index;
1743

1744 1745
			lock_page(page);

N
Nick Piggin 已提交
1746 1747 1748 1749 1750 1751 1752 1753
			/*
			 * 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.
			 */
1754
			if (unlikely(page->mapping != mapping)) {
N
Nick Piggin 已提交
1755
continue_unlock:
1756 1757 1758 1759
				unlock_page(page);
				continue;
			}

1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770
			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;
			}
1771

1772 1773
			BUG_ON(PageWriteback(page));
			if (!clear_page_dirty_for_io(page))
N
Nick Piggin 已提交
1774
				goto continue_unlock;
1775

1776
			trace_wbc_writepage(wbc, mapping->backing_dev_info);
1777
			ret = (*writepage)(page, wbc, data);
1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791
			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).
					 */
1792
					done_index = page->index + 1;
1793 1794 1795
					done = 1;
					break;
				}
1796
			}
1797

1798 1799 1800 1801 1802 1803 1804 1805 1806 1807
			/*
			 * 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;
1808
			}
1809 1810 1811 1812
		}
		pagevec_release(&pvec);
		cond_resched();
	}
1813
	if (!cycled && !done) {
1814
		/*
N
Nick Piggin 已提交
1815
		 * range_cyclic:
1816 1817 1818
		 * We hit the last page and there is more work to be done: wrap
		 * back to the start of the file
		 */
N
Nick Piggin 已提交
1819
		cycled = 1;
1820
		index = 0;
N
Nick Piggin 已提交
1821
		end = writeback_index - 1;
1822 1823
		goto retry;
	}
1824 1825
	if (wbc->range_cyclic || (range_whole && wbc->nr_to_write > 0))
		mapping->writeback_index = done_index;
1826

1827 1828
	return ret;
}
1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854
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)
{
1855 1856 1857
	struct blk_plug plug;
	int ret;

1858 1859 1860 1861
	/* deal with chardevs and other special file */
	if (!mapping->a_ops->writepage)
		return 0;

1862 1863 1864 1865
	blk_start_plug(&plug);
	ret = write_cache_pages(mapping, wbc, __writepage, mapping);
	blk_finish_plug(&plug);
	return ret;
1866
}
1867 1868 1869

EXPORT_SYMBOL(generic_writepages);

L
Linus Torvalds 已提交
1870 1871
int do_writepages(struct address_space *mapping, struct writeback_control *wbc)
{
1872 1873
	int ret;

L
Linus Torvalds 已提交
1874 1875 1876
	if (wbc->nr_to_write <= 0)
		return 0;
	if (mapping->a_ops->writepages)
1877
		ret = mapping->a_ops->writepages(mapping, wbc);
1878 1879 1880
	else
		ret = generic_writepages(mapping, wbc);
	return ret;
L
Linus Torvalds 已提交
1881 1882 1883 1884
}

/**
 * write_one_page - write out a single page and optionally wait on I/O
1885 1886
 * @page: the page to write
 * @wait: if true, wait on writeout
L
Linus Torvalds 已提交
1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921
 *
 * 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);

1922 1923 1924 1925 1926 1927
/*
 * For address_spaces which do not use buffers nor write back.
 */
int __set_page_dirty_no_writeback(struct page *page)
{
	if (!PageDirty(page))
1928
		return !TestSetPageDirty(page);
1929 1930 1931
	return 0;
}

1932 1933 1934 1935 1936 1937 1938 1939
/*
 * 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);
1940
		__inc_zone_page_state(page, NR_DIRTIED);
1941
		__inc_bdi_stat(mapping->backing_dev_info, BDI_RECLAIMABLE);
1942
		__inc_bdi_stat(mapping->backing_dev_info, BDI_DIRTIED);
1943
		task_io_account_write(PAGE_CACHE_SIZE);
1944 1945
		current->nr_dirtied++;
		this_cpu_inc(bdp_ratelimits);
1946 1947
	}
}
M
Michael Rubin 已提交
1948
EXPORT_SYMBOL(account_page_dirtied);
1949

M
Michael Rubin 已提交
1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960
/*
 * 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);

L
Linus Torvalds 已提交
1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973
/*
 * 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
S
Simon Arlott 已提交
1974
 * mapping by re-checking page_mapping() inside tree_lock.
L
Linus Torvalds 已提交
1975 1976 1977 1978 1979 1980 1981
 */
int __set_page_dirty_nobuffers(struct page *page)
{
	if (!TestSetPageDirty(page)) {
		struct address_space *mapping = page_mapping(page);
		struct address_space *mapping2;

1982 1983 1984
		if (!mapping)
			return 1;

N
Nick Piggin 已提交
1985
		spin_lock_irq(&mapping->tree_lock);
1986 1987 1988
		mapping2 = page_mapping(page);
		if (mapping2) { /* Race with truncate? */
			BUG_ON(mapping2 != mapping);
1989
			WARN_ON_ONCE(!PagePrivate(page) && !PageUptodate(page));
1990
			account_page_dirtied(page, mapping);
1991 1992 1993
			radix_tree_tag_set(&mapping->page_tree,
				page_index(page), PAGECACHE_TAG_DIRTY);
		}
N
Nick Piggin 已提交
1994
		spin_unlock_irq(&mapping->tree_lock);
1995 1996 1997
		if (mapping->host) {
			/* !PageAnon && !swapper_space */
			__mark_inode_dirty(mapping->host, I_DIRTY_PAGES);
L
Linus Torvalds 已提交
1998
		}
1999
		return 1;
L
Linus Torvalds 已提交
2000
	}
2001
	return 0;
L
Linus Torvalds 已提交
2002 2003 2004
}
EXPORT_SYMBOL(__set_page_dirty_nobuffers);

2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022
/*
 * Call this whenever redirtying a page, to de-account the dirty counters
 * (NR_DIRTIED, BDI_DIRTIED, tsk->nr_dirtied), so that they match the written
 * counters (NR_WRITTEN, BDI_WRITTEN) in long term. The mismatches will lead to
 * systematic errors in balanced_dirty_ratelimit and the dirty pages position
 * control.
 */
void account_page_redirty(struct page *page)
{
	struct address_space *mapping = page->mapping;
	if (mapping && mapping_cap_account_dirty(mapping)) {
		current->nr_dirtied--;
		dec_zone_page_state(page, NR_DIRTIED);
		dec_bdi_stat(mapping->backing_dev_info, BDI_DIRTIED);
	}
}
EXPORT_SYMBOL(account_page_redirty);

L
Linus Torvalds 已提交
2023 2024 2025 2026 2027 2028 2029 2030
/*
 * 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++;
2031
	account_page_redirty(page);
L
Linus Torvalds 已提交
2032 2033 2034 2035 2036
	return __set_page_dirty_nobuffers(page);
}
EXPORT_SYMBOL(redirty_page_for_writepage);

/*
2037 2038 2039 2040 2041 2042 2043
 * 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.
 *
L
Linus Torvalds 已提交
2044 2045 2046
 * If the mapping doesn't provide a set_page_dirty a_op, then
 * just fall through and assume that it wants buffer_heads.
 */
N
Nick Piggin 已提交
2047
int set_page_dirty(struct page *page)
L
Linus Torvalds 已提交
2048 2049 2050 2051 2052
{
	struct address_space *mapping = page_mapping(page);

	if (likely(mapping)) {
		int (*spd)(struct page *) = mapping->a_ops->set_page_dirty;
M
Minchan Kim 已提交
2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063
		/*
		 * 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);
2064 2065 2066 2067 2068
#ifdef CONFIG_BLOCK
		if (!spd)
			spd = __set_page_dirty_buffers;
#endif
		return (*spd)(page);
L
Linus Torvalds 已提交
2069
	}
2070 2071 2072 2073
	if (!PageDirty(page)) {
		if (!TestSetPageDirty(page))
			return 1;
	}
L
Linus Torvalds 已提交
2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091
	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;

J
Jens Axboe 已提交
2092
	lock_page(page);
L
Linus Torvalds 已提交
2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116
	ret = set_page_dirty(page);
	unlock_page(page);
	return ret;
}
EXPORT_SYMBOL(set_page_dirty_lock);

/*
 * 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);

2117 2118
	BUG_ON(!PageLocked(page));

2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131 2132 2133 2134 2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146
	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.
		 */
		if (page_mkclean(page))
			set_page_dirty(page);
2147 2148 2149 2150 2151 2152 2153 2154 2155 2156
		/*
		 * 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.
		 */
2157
		if (TestClearPageDirty(page)) {
2158
			dec_zone_page_state(page, NR_FILE_DIRTY);
2159 2160
			dec_bdi_stat(mapping->backing_dev_info,
					BDI_RECLAIMABLE);
2161
			return 1;
L
Linus Torvalds 已提交
2162
		}
2163
		return 0;
L
Linus Torvalds 已提交
2164
	}
2165
	return TestClearPageDirty(page);
L
Linus Torvalds 已提交
2166
}
2167
EXPORT_SYMBOL(clear_page_dirty_for_io);
L
Linus Torvalds 已提交
2168 2169 2170 2171 2172 2173 2174

int test_clear_page_writeback(struct page *page)
{
	struct address_space *mapping = page_mapping(page);
	int ret;

	if (mapping) {
P
Peter Zijlstra 已提交
2175
		struct backing_dev_info *bdi = mapping->backing_dev_info;
L
Linus Torvalds 已提交
2176 2177
		unsigned long flags;

N
Nick Piggin 已提交
2178
		spin_lock_irqsave(&mapping->tree_lock, flags);
L
Linus Torvalds 已提交
2179
		ret = TestClearPageWriteback(page);
P
Peter Zijlstra 已提交
2180
		if (ret) {
L
Linus Torvalds 已提交
2181 2182 2183
			radix_tree_tag_clear(&mapping->page_tree,
						page_index(page),
						PAGECACHE_TAG_WRITEBACK);
2184
			if (bdi_cap_account_writeback(bdi)) {
P
Peter Zijlstra 已提交
2185
				__dec_bdi_stat(bdi, BDI_WRITEBACK);
P
Peter Zijlstra 已提交
2186 2187
				__bdi_writeout_inc(bdi);
			}
P
Peter Zijlstra 已提交
2188
		}
N
Nick Piggin 已提交
2189
		spin_unlock_irqrestore(&mapping->tree_lock, flags);
L
Linus Torvalds 已提交
2190 2191 2192
	} else {
		ret = TestClearPageWriteback(page);
	}
2193
	if (ret) {
2194
		dec_zone_page_state(page, NR_WRITEBACK);
2195 2196
		inc_zone_page_state(page, NR_WRITTEN);
	}
L
Linus Torvalds 已提交
2197 2198 2199 2200 2201 2202 2203 2204 2205
	return ret;
}

int test_set_page_writeback(struct page *page)
{
	struct address_space *mapping = page_mapping(page);
	int ret;

	if (mapping) {
P
Peter Zijlstra 已提交
2206
		struct backing_dev_info *bdi = mapping->backing_dev_info;
L
Linus Torvalds 已提交
2207 2208
		unsigned long flags;

N
Nick Piggin 已提交
2209
		spin_lock_irqsave(&mapping->tree_lock, flags);
L
Linus Torvalds 已提交
2210
		ret = TestSetPageWriteback(page);
P
Peter Zijlstra 已提交
2211
		if (!ret) {
L
Linus Torvalds 已提交
2212 2213 2214
			radix_tree_tag_set(&mapping->page_tree,
						page_index(page),
						PAGECACHE_TAG_WRITEBACK);
2215
			if (bdi_cap_account_writeback(bdi))
P
Peter Zijlstra 已提交
2216 2217
				__inc_bdi_stat(bdi, BDI_WRITEBACK);
		}
L
Linus Torvalds 已提交
2218 2219 2220 2221
		if (!PageDirty(page))
			radix_tree_tag_clear(&mapping->page_tree,
						page_index(page),
						PAGECACHE_TAG_DIRTY);
2222 2223 2224
		radix_tree_tag_clear(&mapping->page_tree,
				     page_index(page),
				     PAGECACHE_TAG_TOWRITE);
N
Nick Piggin 已提交
2225
		spin_unlock_irqrestore(&mapping->tree_lock, flags);
L
Linus Torvalds 已提交
2226 2227 2228
	} else {
		ret = TestSetPageWriteback(page);
	}
2229
	if (!ret)
M
Michael Rubin 已提交
2230
		account_page_writeback(page);
L
Linus Torvalds 已提交
2231 2232 2233 2234 2235 2236
	return ret;

}
EXPORT_SYMBOL(test_set_page_writeback);

/*
N
Nick Piggin 已提交
2237
 * Return true if any of the pages in the mapping are marked with the
L
Linus Torvalds 已提交
2238 2239 2240 2241
 * passed tag.
 */
int mapping_tagged(struct address_space *mapping, int tag)
{
2242
	return radix_tree_tagged(&mapping->page_tree, tag);
L
Linus Torvalds 已提交
2243 2244
}
EXPORT_SYMBOL(mapping_tagged);