page-writeback.c 74.2 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 <linux/timer.h>
38
#include <linux/sched/rt.h>
39
#include <linux/mm_inline.h>
40
#include <trace/events/writeback.h>
L
Linus Torvalds 已提交
41

42 43
#include "internal.h"

44 45 46 47 48
/*
 * Sleep at most 200ms at a time in balance_dirty_pages().
 */
#define MAX_PAUSE		max(HZ/5, 1)

49 50 51 52 53 54
/*
 * 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))

55 56 57 58 59
/*
 * Estimate write bandwidth at 200ms intervals.
 */
#define BANDWIDTH_INTERVAL	max(HZ/5, 1)

W
Wu Fengguang 已提交
60 61
#define RATELIMIT_CALC_SHIFT	10

L
Linus Torvalds 已提交
62 63 64 65 66 67 68 69 70
/*
 * 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 */

/*
71
 * Start background writeback (via writeback threads) at this percentage
L
Linus Torvalds 已提交
72
 */
73
int dirty_background_ratio = 10;
L
Linus Torvalds 已提交
74

75 76 77 78 79 80
/*
 * 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;

81 82 83 84 85 86
/*
 * 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
87 88 89
/*
 * The generator of dirty data starts writeback at this percentage
 */
90
int vm_dirty_ratio = 20;
L
Linus Torvalds 已提交
91

92 93 94 95 96 97
/*
 * 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
98
/*
99
 * The interval between `kupdate'-style writebacks
L
Linus Torvalds 已提交
100
 */
101
unsigned int dirty_writeback_interval = 5 * 100; /* centiseconds */
L
Linus Torvalds 已提交
102

103 104
EXPORT_SYMBOL_GPL(dirty_writeback_interval);

L
Linus Torvalds 已提交
105
/*
106
 * The longest time for which data is allowed to remain dirty
L
Linus Torvalds 已提交
107
 */
108
unsigned int dirty_expire_interval = 30 * 100; /* centiseconds */
L
Linus Torvalds 已提交
109 110 111 112 113 114 115

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

/*
116 117
 * 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
118 119 120 121 122 123 124
 */
int laptop_mode;

EXPORT_SYMBOL(laptop_mode);

/* End of sysctl-exported parameters */

125
unsigned long global_dirty_limit;
L
Linus Torvalds 已提交
126

P
Peter Zijlstra 已提交
127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142
/*
 * 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.
 *
 */
143 144 145 146 147 148 149 150 151 152 153 154 155 156
static struct fprop_global writeout_completions;

static void writeout_period(unsigned long t);
/* Timer for aging of writeout_completions */
static struct timer_list writeout_period_timer =
		TIMER_DEFERRED_INITIALIZER(writeout_period, 0, 0);
static unsigned long writeout_period_time = 0;

/*
 * Length of period for aging writeout fractions of bdis. This is an
 * arbitrarily chosen number. The longer the period, the slower fractions will
 * reflect changes in current writeout rate.
 */
#define VM_COMPLETIONS_PERIOD_LEN (3*HZ)
P
Peter Zijlstra 已提交
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.
 */

176 177 178 179 180 181 182 183 184 185 186 187 188 189
/**
 * 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)
{
	unsigned long nr_pages;

	nr_pages = zone_page_state(zone, NR_FREE_PAGES);
	nr_pages -= min(nr_pages, zone->dirty_balance_reserve);

190 191
	nr_pages += zone_page_state(zone, NR_INACTIVE_FILE);
	nr_pages += zone_page_state(zone, NR_ACTIVE_FILE);
192 193 194 195

	return nr_pages;
}

196 197 198 199 200 201 202
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) {
203
		struct zone *z = &NODE_DATA(node)->node_zones[ZONE_HIGHMEM];
204

205
		x += zone_dirtyable_memory(z);
206
	}
207 208 209 210 211 212 213 214 215 216 217 218
	/*
	 * Unreclaimable memory (kernel memory or anonymous memory
	 * without swap) can bring down the dirtyable pages below
	 * the zone's dirty balance reserve and the above calculation
	 * will underflow.  However we still want to add in nodes
	 * which are below threshold (negative values) to get a more
	 * accurate calculation but make sure that the total never
	 * underflows.
	 */
	if ((long)x < 0)
		x = 0;

219 220 221 222 223 224 225 226 227 228 229 230 231
	/*
	 * 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
}

/**
232
 * global_dirtyable_memory - number of globally dirtyable pages
233
 *
234 235
 * Returns the global number of pages potentially available for dirty
 * page cache.  This is the base value for the global dirty limits.
236
 */
237
static unsigned long global_dirtyable_memory(void)
238 239 240
{
	unsigned long x;

241
	x = global_page_state(NR_FREE_PAGES);
242
	x -= min(x, dirty_balance_reserve);
243

244 245
	x += global_page_state(NR_INACTIVE_FILE);
	x += global_page_state(NR_ACTIVE_FILE);
246

247 248 249 250 251 252
	if (!vm_highmem_is_dirtyable)
		x -= highmem_dirtyable_memory(x);

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

253 254 255 256 257 258 259 260 261 262 263
/*
 * 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)
{
264
	const unsigned long available_memory = global_dirtyable_memory();
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
	unsigned long background;
	unsigned long dirty;
	struct task_struct *tsk;

	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);
}

291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331
/**
 * 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;
}

332
int dirty_background_ratio_handler(struct ctl_table *table, int write,
333
		void __user *buffer, size_t *lenp,
334 335 336 337
		loff_t *ppos)
{
	int ret;

338
	ret = proc_dointvec_minmax(table, write, buffer, lenp, ppos);
339 340 341 342 343 344
	if (ret == 0 && write)
		dirty_background_bytes = 0;
	return ret;
}

int dirty_background_bytes_handler(struct ctl_table *table, int write,
345
		void __user *buffer, size_t *lenp,
346 347 348 349
		loff_t *ppos)
{
	int ret;

350
	ret = proc_doulongvec_minmax(table, write, buffer, lenp, ppos);
351 352 353 354 355
	if (ret == 0 && write)
		dirty_background_ratio = 0;
	return ret;
}

P
Peter Zijlstra 已提交
356
int dirty_ratio_handler(struct ctl_table *table, int write,
357
		void __user *buffer, size_t *lenp,
P
Peter Zijlstra 已提交
358 359 360
		loff_t *ppos)
{
	int old_ratio = vm_dirty_ratio;
361 362
	int ret;

363
	ret = proc_dointvec_minmax(table, write, buffer, lenp, ppos);
P
Peter Zijlstra 已提交
364
	if (ret == 0 && write && vm_dirty_ratio != old_ratio) {
365
		writeback_set_ratelimit();
366 367 368 369 370 371
		vm_dirty_bytes = 0;
	}
	return ret;
}

int dirty_bytes_handler(struct ctl_table *table, int write,
372
		void __user *buffer, size_t *lenp,
373 374
		loff_t *ppos)
{
375
	unsigned long old_bytes = vm_dirty_bytes;
376 377
	int ret;

378
	ret = proc_doulongvec_minmax(table, write, buffer, lenp, ppos);
379
	if (ret == 0 && write && vm_dirty_bytes != old_bytes) {
380
		writeback_set_ratelimit();
381
		vm_dirty_ratio = 0;
P
Peter Zijlstra 已提交
382 383 384 385
	}
	return ret;
}

386 387 388 389 390 391 392 393 394
static unsigned long wp_next_time(unsigned long cur_time)
{
	cur_time += VM_COMPLETIONS_PERIOD_LEN;
	/* 0 has a special meaning... */
	if (!cur_time)
		return 1;
	return cur_time;
}

P
Peter Zijlstra 已提交
395 396 397 398 399 400
/*
 * 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)
{
401
	__inc_bdi_stat(bdi, BDI_WRITTEN);
402 403 404 405 406 407 408 409 410 411 412 413 414
	__fprop_inc_percpu_max(&writeout_completions, &bdi->completions,
			       bdi->max_prop_frac);
	/* First event after period switching was turned off? */
	if (!unlikely(writeout_period_time)) {
		/*
		 * We can race with other __bdi_writeout_inc calls here but
		 * it does not cause any harm since the resulting time when
		 * timer will fire and what is in writeout_period_time will be
		 * roughly the same.
		 */
		writeout_period_time = wp_next_time(jiffies);
		mod_timer(&writeout_period_timer, writeout_period_time);
	}
P
Peter Zijlstra 已提交
415 416
}

417 418 419 420 421 422 423 424 425 426
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
427 428 429 430 431 432
/*
 * Obtain an accurate fraction of the BDI's portion.
 */
static void bdi_writeout_fraction(struct backing_dev_info *bdi,
		long *numerator, long *denominator)
{
433
	fprop_fraction_percpu(&writeout_completions, &bdi->completions,
P
Peter Zijlstra 已提交
434 435 436
				numerator, denominator);
}

437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458
/*
 * On idle system, we can be called long after we scheduled because we use
 * deferred timers so count with missed periods.
 */
static void writeout_period(unsigned long t)
{
	int miss_periods = (jiffies - writeout_period_time) /
						 VM_COMPLETIONS_PERIOD_LEN;

	if (fprop_new_period(&writeout_completions, miss_periods + 1)) {
		writeout_period_time = wp_next_time(writeout_period_time +
				miss_periods * VM_COMPLETIONS_PERIOD_LEN);
		mod_timer(&writeout_period_timer, writeout_period_time);
	} else {
		/*
		 * Aging has zeroed all fractions. Stop wasting CPU on period
		 * updates.
		 */
		writeout_period_time = 0;
	}
}

459
/*
460 461 462
 * bdi_min_ratio keeps the sum of the minimum dirty shares of all
 * registered backing devices, which, for obvious reasons, can not
 * exceed 100%.
463 464 465 466 467 468 469
 */
static unsigned int bdi_min_ratio;

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

470
	spin_lock_bh(&bdi_lock);
471
	if (min_ratio > bdi->max_ratio) {
472
		ret = -EINVAL;
473 474 475 476 477 478 479 480 481
	} 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;
		}
	}
482
	spin_unlock_bh(&bdi_lock);
483 484 485 486 487 488 489 490 491 492 493

	return ret;
}

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

	if (max_ratio > 100)
		return -EINVAL;

494
	spin_lock_bh(&bdi_lock);
495 496 497 498
	if (bdi->min_ratio > max_ratio) {
		ret = -EINVAL;
	} else {
		bdi->max_ratio = max_ratio;
499
		bdi->max_prop_frac = (FPROP_FRAC_BASE * max_ratio) / 100;
500
	}
501
	spin_unlock_bh(&bdi_lock);
502 503 504

	return ret;
}
505
EXPORT_SYMBOL(bdi_set_max_ratio);
506

W
Wu Fengguang 已提交
507 508 509 510 511 512
static unsigned long dirty_freerun_ceiling(unsigned long thresh,
					   unsigned long bg_thresh)
{
	return (thresh + bg_thresh) / 2;
}

513 514 515 516 517
static unsigned long hard_dirty_limit(unsigned long thresh)
{
	return max(thresh, global_dirty_limit);
}

518
/**
519
 * bdi_dirty_limit - @bdi's share of dirty throttling threshold
520 521
 * @bdi: the backing_dev_info to query
 * @dirty: global dirty limit in pages
522
 *
523 524
 * 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.
525 526 527 528 529 530 531
 *
 * 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.
532
 *
533
 * It allocates high/low dirty limits to fast/slow devices, in order to prevent
534 535 536 537 538 539 540
 * - 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)
541 542 543
{
	u64 bdi_dirty;
	long numerator, denominator;
P
Peter Zijlstra 已提交
544

545 546 547 548
	/*
	 * Calculate this BDI's share of the dirty ratio.
	 */
	bdi_writeout_fraction(bdi, &numerator, &denominator);
P
Peter Zijlstra 已提交
549

550 551 552
	bdi_dirty = (dirty * (100 - bdi_min_ratio)) / 100;
	bdi_dirty *= numerator;
	do_div(bdi_dirty, denominator);
P
Peter Zijlstra 已提交
553

554 555 556 557 558
	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
559 560
}

561 562 563 564 565 566 567 568 569 570 571 572 573 574
/*
 *                           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
 */
575
static long long pos_ratio_polynom(unsigned long setpoint,
576 577 578 579 580 581
					  unsigned long dirty,
					  unsigned long limit)
{
	long long pos_ratio;
	long x;

582
	x = div64_s64(((s64)setpoint - (s64)dirty) << RATELIMIT_CALC_SHIFT,
583 584 585 586 587 588 589 590 591
		    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;

	return clamp(pos_ratio, 0LL, 2LL << RATELIMIT_CALC_SHIFT);
}

W
Wu Fengguang 已提交
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
/*
 * 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
	 *
690 691 692 693 694 695 696 697 698 699 700 701 702 703
	 * See comment for pos_ratio_polynom().
	 */
	setpoint = (freerun + limit) / 2;
	pos_ratio = pos_ratio_polynom(setpoint, dirty, limit);

	/*
	 * The strictlimit feature is a tool preventing mistrusted filesystems
	 * from growing a large number of dirty pages before throttling. For
	 * such filesystems balance_dirty_pages always checks bdi counters
	 * against bdi limits. Even if global "nr_dirty" is under "freerun".
	 * This is especially important for fuse which sets bdi->max_ratio to
	 * 1% by default. Without strictlimit feature, fuse writeback may
	 * consume arbitrary amount of RAM because it is accounted in
	 * NR_WRITEBACK_TEMP which is not involved in calculating "nr_dirty".
W
Wu Fengguang 已提交
704
	 *
705 706 707 708 709 710 711 712 713
	 * Here, in bdi_position_ratio(), we calculate pos_ratio based on
	 * two values: bdi_dirty and bdi_thresh. Let's consider an example:
	 * total amount of RAM is 16GB, bdi->max_ratio is equal to 1%, global
	 * limits are set by default to 10% and 20% (background and throttle).
	 * Then bdi_thresh is 1% of 20% of 16GB. This amounts to ~8K pages.
	 * bdi_dirty_limit(bdi, bg_thresh) is about ~4K pages. bdi_setpoint is
	 * about ~6K pages (as the average of background and throttle bdi
	 * limits). The 3rd order polynomial will provide positive feedback if
	 * bdi_dirty is under bdi_setpoint and vice versa.
W
Wu Fengguang 已提交
714
	 *
715 716 717 718
	 * Note, that we cannot use global counters in these calculations
	 * because we want to throttle process writing to a strictlimit BDI
	 * much earlier than global "freerun" is reached (~23MB vs. ~2.3GB
	 * in the example above).
W
Wu Fengguang 已提交
719
	 */
720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763
	if (unlikely(bdi->capabilities & BDI_CAP_STRICTLIMIT)) {
		long long bdi_pos_ratio;
		unsigned long bdi_bg_thresh;

		if (bdi_dirty < 8)
			return min_t(long long, pos_ratio * 2,
				     2 << RATELIMIT_CALC_SHIFT);

		if (bdi_dirty >= bdi_thresh)
			return 0;

		bdi_bg_thresh = div_u64((u64)bdi_thresh * bg_thresh, thresh);
		bdi_setpoint = dirty_freerun_ceiling(bdi_thresh,
						     bdi_bg_thresh);

		if (bdi_setpoint == 0 || bdi_setpoint == bdi_thresh)
			return 0;

		bdi_pos_ratio = pos_ratio_polynom(bdi_setpoint, bdi_dirty,
						  bdi_thresh);

		/*
		 * Typically, for strictlimit case, bdi_setpoint << setpoint
		 * and pos_ratio >> bdi_pos_ratio. In the other words global
		 * state ("dirty") is not limiting factor and we have to
		 * make decision based on bdi counters. But there is an
		 * important case when global pos_ratio should get precedence:
		 * global limits are exceeded (e.g. due to activities on other
		 * BDIs) while given strictlimit BDI is below limit.
		 *
		 * "pos_ratio * bdi_pos_ratio" would work for the case above,
		 * but it would look too non-natural for the case of all
		 * activity in the system coming from a single strictlimit BDI
		 * with bdi->max_ratio == 100%.
		 *
		 * Note that min() below somewhat changes the dynamics of the
		 * control system. Normally, pos_ratio value can be well over 3
		 * (when globally we are at freerun and bdi is well below bdi
		 * setpoint). Now the maximum pos_ratio in the same situation
		 * is 2. We might want to tweak this if we observe the control
		 * system is too slow to adapt.
		 */
		return min(pos_ratio, bdi_pos_ratio);
	}
W
Wu Fengguang 已提交
764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797

	/*
	 * 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;
798 799 800 801 802 803 804
	/*
	 * 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.
	 */
805
	bdi_thresh = max(bdi_thresh, (limit - dirty) / 8);
W
Wu Fengguang 已提交
806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823
	/*
	 * 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) {
824
		pos_ratio = div64_u64(pos_ratio * (x_intercept - bdi_dirty),
825
				    x_intercept - bdi_setpoint + 1);
W
Wu Fengguang 已提交
826 827 828
	} else
		pos_ratio /= 4;

829 830 831 832 833 834 835
	/*
	 * 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) {
836 837 838
		if (bdi_dirty > x_intercept / 8)
			pos_ratio = div_u64(pos_ratio * x_intercept, bdi_dirty);
		else
839 840 841
			pos_ratio *= 8;
	}

W
Wu Fengguang 已提交
842 843 844
	return pos_ratio;
}

845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884
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;
}

885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940
/*
 * 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
941 942 943 944 945 946 947 948 949 950 951 952 953 954 955
/*
 * 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)
{
956 957 958
	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
959 960 961 962 963 964
	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;
965 966
	unsigned long step;
	unsigned long x;
W
Wu Fengguang 已提交
967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014

	/*
	 * 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);
1015 1016 1017 1018 1019
	/*
	 * 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
1020

1021 1022 1023 1024 1025 1026
	/*
	 * 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
W
Wanpeng Li 已提交
1027
	 * code makes use of task_ratelimit to filter out singular points and
1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049
	 * 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
W
Wanpeng Li 已提交
1050
	 * and filter out the singular points of balanced_dirty_ratelimit. Which
1051 1052 1053 1054 1055
	 * 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;
1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076

	/*
	 * For strictlimit case, calculations above were based on bdi counters
	 * and limits (starting from pos_ratio = bdi_position_ratio() and up to
	 * balanced_dirty_ratelimit = task_ratelimit * write_bw / dirty_rate).
	 * Hence, to calculate "step" properly, we have to use bdi_dirty as
	 * "dirty" and bdi_setpoint as "setpoint".
	 *
	 * We rampup dirty_ratelimit forcibly if bdi_dirty is low because
	 * it's possible that bdi_thresh is close to zero due to inactivity
	 * of backing device (see the implementation of bdi_dirty_limit()).
	 */
	if (unlikely(bdi->capabilities & BDI_CAP_STRICTLIMIT)) {
		dirty = bdi_dirty;
		if (bdi_dirty < 8)
			setpoint = bdi_dirty + 1;
		else
			setpoint = (bdi_thresh +
				    bdi_dirty_limit(bdi, bg_thresh)) / 2;
	}

1077
	if (dirty < setpoint) {
1078 1079
		x = min3(bdi->balanced_dirty_ratelimit,
			 balanced_dirty_ratelimit, task_ratelimit);
1080 1081 1082
		if (dirty_ratelimit < x)
			step = x - dirty_ratelimit;
	} else {
1083 1084
		x = max3(bdi->balanced_dirty_ratelimit,
			 balanced_dirty_ratelimit, task_ratelimit);
1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106
		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;
1107 1108

	trace_bdi_dirty_ratelimit(bdi, dirty_rate, task_ratelimit);
W
Wu Fengguang 已提交
1109 1110
}

1111
void __bdi_update_bandwidth(struct backing_dev_info *bdi,
1112
			    unsigned long thresh,
1113
			    unsigned long bg_thresh,
1114 1115 1116
			    unsigned long dirty,
			    unsigned long bdi_thresh,
			    unsigned long bdi_dirty,
1117 1118 1119 1120
			    unsigned long start_time)
{
	unsigned long now = jiffies;
	unsigned long elapsed = now - bdi->bw_time_stamp;
W
Wu Fengguang 已提交
1121
	unsigned long dirtied;
1122 1123 1124 1125 1126 1127 1128 1129
	unsigned long written;

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

W
Wu Fengguang 已提交
1130
	dirtied = percpu_counter_read(&bdi->bdi_stat[BDI_DIRTIED]);
1131 1132 1133 1134 1135 1136 1137 1138 1139
	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
1140
	if (thresh) {
1141
		global_update_bandwidth(thresh, dirty, now);
W
Wu Fengguang 已提交
1142 1143 1144 1145
		bdi_update_dirty_ratelimit(bdi, thresh, bg_thresh, dirty,
					   bdi_thresh, bdi_dirty,
					   dirtied, elapsed);
	}
1146 1147 1148
	bdi_update_write_bandwidth(bdi, elapsed, written);

snapshot:
W
Wu Fengguang 已提交
1149
	bdi->dirtied_stamp = dirtied;
1150 1151 1152 1153 1154
	bdi->written_stamp = written;
	bdi->bw_time_stamp = now;
}

static void bdi_update_bandwidth(struct backing_dev_info *bdi,
1155
				 unsigned long thresh,
1156
				 unsigned long bg_thresh,
1157 1158 1159
				 unsigned long dirty,
				 unsigned long bdi_thresh,
				 unsigned long bdi_dirty,
1160 1161 1162 1163 1164
				 unsigned long start_time)
{
	if (time_is_after_eq_jiffies(bdi->bw_time_stamp + BANDWIDTH_INTERVAL))
		return;
	spin_lock(&bdi->wb.list_lock);
1165 1166
	__bdi_update_bandwidth(bdi, thresh, bg_thresh, dirty,
			       bdi_thresh, bdi_dirty, start_time);
1167 1168 1169
	spin_unlock(&bdi->wb.list_lock);
}

1170
/*
1171
 * After a task dirtied this many pages, balance_dirty_pages_ratelimited()
1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186
 * 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;
}

1187 1188
static unsigned long bdi_max_pause(struct backing_dev_info *bdi,
				   unsigned long bdi_dirty)
1189
{
1190 1191
	unsigned long bw = bdi->avg_write_bandwidth;
	unsigned long t;
1192

1193 1194 1195 1196 1197 1198 1199 1200 1201 1202
	/*
	 * 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++;

1203
	return min_t(unsigned long, t, MAX_PAUSE);
1204 1205 1206 1207 1208 1209 1210
}

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)
1211
{
1212 1213 1214 1215 1216
	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 */
1217

1218 1219
	/* target for 10ms pause on 1-dd case */
	t = max(1, HZ / 100);
1220 1221 1222 1223 1224

	/*
	 * Scale up pause time for concurrent dirtiers in order to reduce CPU
	 * overheads.
	 *
1225
	 * (N * 10ms) on 2^N concurrent tasks.
1226 1227
	 */
	if (hi > lo)
1228
		t += (hi - lo) * (10 * HZ) / 1024;
1229 1230

	/*
1231 1232 1233 1234 1235 1236 1237 1238
	 * 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.
1239
	 *
1240 1241 1242 1243 1244 1245 1246
	 * 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.
1247
	 */
1248 1249
	t = min(t, 1 + max_pause / 2);
	pages = dirty_ratelimit * t / roundup_pow_of_two(HZ);
1250 1251

	/*
1252 1253 1254 1255 1256 1257
	 * 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.
1258
	 */
1259 1260 1261 1262 1263 1264 1265 1266 1267
	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;
		}
	}

1268 1269 1270 1271 1272
	pause = HZ * pages / (task_ratelimit + 1);
	if (pause > max_pause) {
		t = max_pause;
		pages = task_ratelimit * t / roundup_pow_of_two(HZ);
	}
1273

1274
	*nr_dirtied_pause = pages;
1275
	/*
1276
	 * The minimal pause time will normally be half the target pause time.
1277
	 */
1278
	return pages >= DIRTY_POLL_THRESH ? 1 + t / 2 : t;
1279 1280
}

1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305
static inline void bdi_dirty_limits(struct backing_dev_info *bdi,
				    unsigned long dirty_thresh,
				    unsigned long background_thresh,
				    unsigned long *bdi_dirty,
				    unsigned long *bdi_thresh,
				    unsigned long *bdi_bg_thresh)
{
	unsigned long bdi_reclaimable;

	/*
	 * 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.
	 */
	*bdi_thresh = bdi_dirty_limit(bdi, dirty_thresh);

	if (bdi_bg_thresh)
1306 1307 1308
		*bdi_bg_thresh = dirty_thresh ? div_u64((u64)*bdi_thresh *
							background_thresh,
							dirty_thresh) : 0;
1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330

	/*
	 * 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_reclaimable = bdi_stat_sum(bdi, BDI_RECLAIMABLE);
		*bdi_dirty = bdi_reclaimable +
			bdi_stat_sum(bdi, BDI_WRITEBACK);
	} else {
		bdi_reclaimable = bdi_stat(bdi, BDI_RECLAIMABLE);
		*bdi_dirty = bdi_reclaimable +
			bdi_stat(bdi, BDI_WRITEBACK);
	}
}

L
Linus Torvalds 已提交
1331 1332 1333
/*
 * 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
1334
 * the caller to wait once crossing the (background_thresh + dirty_thresh) / 2.
1335 1336
 * If we're over `background_thresh' then the writeback threads are woken to
 * perform some writeout.
L
Linus Torvalds 已提交
1337
 */
1338
static void balance_dirty_pages(struct address_space *mapping,
1339
				unsigned long pages_dirtied)
L
Linus Torvalds 已提交
1340
{
1341
	unsigned long nr_reclaimable;	/* = file_dirty + unstable_nfs */
1342
	unsigned long nr_dirty;  /* = file_dirty + writeback + unstable_nfs */
1343 1344
	unsigned long background_thresh;
	unsigned long dirty_thresh;
1345
	long period;
1346 1347 1348 1349
	long pause;
	long max_pause;
	long min_pause;
	int nr_dirtied_pause;
1350
	bool dirty_exceeded = false;
1351
	unsigned long task_ratelimit;
1352
	unsigned long dirty_ratelimit;
1353
	unsigned long pos_ratio;
L
Linus Torvalds 已提交
1354
	struct backing_dev_info *bdi = mapping->backing_dev_info;
1355
	bool strictlimit = bdi->capabilities & BDI_CAP_STRICTLIMIT;
1356
	unsigned long start_time = jiffies;
L
Linus Torvalds 已提交
1357 1358

	for (;;) {
1359
		unsigned long now = jiffies;
1360 1361 1362 1363 1364
		unsigned long uninitialized_var(bdi_thresh);
		unsigned long thresh;
		unsigned long uninitialized_var(bdi_dirty);
		unsigned long dirty;
		unsigned long bg_thresh;
1365

1366 1367 1368 1369 1370 1371
		/*
		 * 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.
		 */
1372 1373
		nr_reclaimable = global_page_state(NR_FILE_DIRTY) +
					global_page_state(NR_UNSTABLE_NFS);
1374
		nr_dirty = nr_reclaimable + global_page_state(NR_WRITEBACK);
1375

1376 1377
		global_dirty_limits(&background_thresh, &dirty_thresh);

1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389
		if (unlikely(strictlimit)) {
			bdi_dirty_limits(bdi, dirty_thresh, background_thresh,
					 &bdi_dirty, &bdi_thresh, &bg_thresh);

			dirty = bdi_dirty;
			thresh = bdi_thresh;
		} else {
			dirty = nr_dirty;
			thresh = dirty_thresh;
			bg_thresh = background_thresh;
		}

1390 1391 1392
		/*
		 * Throttle it only when the background writeback cannot
		 * catch-up. This avoids (excessively) small writeouts
1393 1394 1395 1396 1397
		 * when the bdi limits are ramping up in case of !strictlimit.
		 *
		 * In strictlimit case make decision based on the bdi counters
		 * and limits. Small writeouts when the bdi limits are ramping
		 * up are the price we consciously pay for strictlimit-ing.
1398
		 */
1399
		if (dirty <= dirty_freerun_ceiling(thresh, bg_thresh)) {
1400 1401
			current->dirty_paused_when = now;
			current->nr_dirtied = 0;
1402
			current->nr_dirtied_pause =
1403
				dirty_poll_interval(dirty, thresh);
1404
			break;
1405
		}
1406

1407 1408 1409
		if (unlikely(!writeback_in_progress(bdi)))
			bdi_start_background_writeback(bdi);

1410 1411 1412
		if (!strictlimit)
			bdi_dirty_limits(bdi, dirty_thresh, background_thresh,
					 &bdi_dirty, &bdi_thresh, NULL);
1413

1414
		dirty_exceeded = (bdi_dirty > bdi_thresh) &&
1415
				 ((nr_dirty > dirty_thresh) || strictlimit);
1416
		if (dirty_exceeded && !bdi->dirty_exceeded)
P
Peter Zijlstra 已提交
1417
			bdi->dirty_exceeded = 1;
L
Linus Torvalds 已提交
1418

1419 1420 1421
		bdi_update_bandwidth(bdi, dirty_thresh, background_thresh,
				     nr_dirty, bdi_thresh, bdi_dirty,
				     start_time);
1422

1423 1424 1425 1426
		dirty_ratelimit = bdi->dirty_ratelimit;
		pos_ratio = bdi_position_ratio(bdi, dirty_thresh,
					       background_thresh, nr_dirty,
					       bdi_thresh, bdi_dirty);
1427 1428
		task_ratelimit = ((u64)dirty_ratelimit * pos_ratio) >>
							RATELIMIT_CALC_SHIFT;
1429 1430 1431 1432 1433
		max_pause = bdi_max_pause(bdi, bdi_dirty);
		min_pause = bdi_min_pause(bdi, max_pause,
					  task_ratelimit, dirty_ratelimit,
					  &nr_dirtied_pause);

1434
		if (unlikely(task_ratelimit == 0)) {
1435
			period = max_pause;
1436
			pause = max_pause;
1437
			goto pause;
P
Peter Zijlstra 已提交
1438
		}
1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449
		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.
		 */
1450
		if (pause < min_pause) {
1451 1452 1453 1454 1455 1456 1457 1458 1459
			trace_balance_dirty_pages(bdi,
						  dirty_thresh,
						  background_thresh,
						  nr_dirty,
						  bdi_thresh,
						  bdi_dirty,
						  dirty_ratelimit,
						  task_ratelimit,
						  pages_dirtied,
1460
						  period,
1461
						  min(pause, 0L),
1462
						  start_time);
1463 1464 1465 1466 1467 1468
			if (pause < -HZ) {
				current->dirty_paused_when = now;
				current->nr_dirtied = 0;
			} else if (period) {
				current->dirty_paused_when += period;
				current->nr_dirtied = 0;
1469 1470
			} else if (current->nr_dirtied_pause <= pages_dirtied)
				current->nr_dirtied_pause += pages_dirtied;
W
Wu Fengguang 已提交
1471
			break;
P
Peter Zijlstra 已提交
1472
		}
1473 1474 1475 1476 1477
		if (unlikely(pause > max_pause)) {
			/* for occasional dropped task_ratelimit */
			now += min(pause - max_pause, max_pause);
			pause = max_pause;
		}
1478 1479

pause:
1480 1481 1482 1483 1484 1485 1486 1487 1488
		trace_balance_dirty_pages(bdi,
					  dirty_thresh,
					  background_thresh,
					  nr_dirty,
					  bdi_thresh,
					  bdi_dirty,
					  dirty_ratelimit,
					  task_ratelimit,
					  pages_dirtied,
1489
					  period,
1490 1491
					  pause,
					  start_time);
1492
		__set_current_state(TASK_KILLABLE);
1493
		io_schedule_timeout(pause);
1494

1495 1496
		current->dirty_paused_when = now + pause;
		current->nr_dirtied = 0;
1497
		current->nr_dirtied_pause = nr_dirtied_pause;
1498

1499
		/*
1500 1501
		 * This is typically equal to (nr_dirty < dirty_thresh) and can
		 * also keep "1000+ dd on a slow USB stick" under control.
1502
		 */
1503
		if (task_ratelimit)
1504
			break;
1505

1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518
		/*
		 * 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;

1519 1520
		if (fatal_signal_pending(current))
			break;
L
Linus Torvalds 已提交
1521 1522
	}

1523
	if (!dirty_exceeded && bdi->dirty_exceeded)
P
Peter Zijlstra 已提交
1524
		bdi->dirty_exceeded = 0;
L
Linus Torvalds 已提交
1525 1526

	if (writeback_in_progress(bdi))
1527
		return;
L
Linus Torvalds 已提交
1528 1529 1530 1531 1532 1533 1534 1535 1536

	/*
	 * 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.
	 */
1537 1538 1539 1540
	if (laptop_mode)
		return;

	if (nr_reclaimable > background_thresh)
1541
		bdi_start_background_writeback(bdi);
L
Linus Torvalds 已提交
1542 1543
}

1544
void set_page_dirty_balance(struct page *page)
P
Peter Zijlstra 已提交
1545
{
1546
	if (set_page_dirty(page)) {
P
Peter Zijlstra 已提交
1547 1548 1549 1550 1551 1552 1553
		struct address_space *mapping = page_mapping(page);

		if (mapping)
			balance_dirty_pages_ratelimited(mapping);
	}
}

1554
static DEFINE_PER_CPU(int, bdp_ratelimits);
1555

1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571
/*
 * 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
1572
/**
1573
 * balance_dirty_pages_ratelimited - balance dirty memory state
1574
 * @mapping: address_space which was dirtied
L
Linus Torvalds 已提交
1575 1576 1577 1578 1579 1580 1581 1582 1583 1584
 *
 * 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.
 */
1585
void balance_dirty_pages_ratelimited(struct address_space *mapping)
L
Linus Torvalds 已提交
1586
{
1587
	struct backing_dev_info *bdi = mapping->backing_dev_info;
1588 1589
	int ratelimit;
	int *p;
L
Linus Torvalds 已提交
1590

1591 1592 1593
	if (!bdi_cap_account_dirty(bdi))
		return;

1594 1595 1596 1597 1598
	ratelimit = current->nr_dirtied_pause;
	if (bdi->dirty_exceeded)
		ratelimit = min(ratelimit, 32 >> (PAGE_SHIFT - 10));

	preempt_disable();
L
Linus Torvalds 已提交
1599
	/*
1600 1601 1602 1603
	 * 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
1604
	 */
1605
	p =  this_cpu_ptr(&bdp_ratelimits);
1606
	if (unlikely(current->nr_dirtied >= ratelimit))
1607
		*p = 0;
1608 1609 1610
	else if (unlikely(*p >= ratelimit_pages)) {
		*p = 0;
		ratelimit = 0;
L
Linus Torvalds 已提交
1611
	}
1612 1613 1614 1615 1616
	/*
	 * 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.
	 */
1617
	p = this_cpu_ptr(&dirty_throttle_leaks);
1618
	if (*p > 0 && current->nr_dirtied < ratelimit) {
1619
		unsigned long nr_pages_dirtied;
1620 1621 1622
		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
1623
	}
1624
	preempt_enable();
1625 1626 1627

	if (unlikely(current->nr_dirtied >= ratelimit))
		balance_dirty_pages(mapping, current->nr_dirtied);
L
Linus Torvalds 已提交
1628
}
1629
EXPORT_SYMBOL(balance_dirty_pages_ratelimited);
L
Linus Torvalds 已提交
1630

1631
void throttle_vm_writeout(gfp_t gfp_mask)
L
Linus Torvalds 已提交
1632
{
1633 1634
	unsigned long background_thresh;
	unsigned long dirty_thresh;
L
Linus Torvalds 已提交
1635 1636

        for ( ; ; ) {
1637
		global_dirty_limits(&background_thresh, &dirty_thresh);
1638
		dirty_thresh = hard_dirty_limit(dirty_thresh);
L
Linus Torvalds 已提交
1639 1640 1641 1642 1643 1644 1645

                /*
                 * 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... */

1646 1647 1648
                if (global_page_state(NR_UNSTABLE_NFS) +
			global_page_state(NR_WRITEBACK) <= dirty_thresh)
                        	break;
1649
                congestion_wait(BLK_RW_ASYNC, HZ/10);
1650 1651 1652 1653 1654 1655 1656 1657

		/*
		 * 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
1658 1659 1660 1661 1662 1663
        }
}

/*
 * sysctl handler for /proc/sys/vm/dirty_writeback_centisecs
 */
1664
int dirty_writeback_centisecs_handler(struct ctl_table *table, int write,
1665
	void __user *buffer, size_t *length, loff_t *ppos)
L
Linus Torvalds 已提交
1666
{
1667
	proc_dointvec(table, write, buffer, length, ppos);
L
Linus Torvalds 已提交
1668 1669 1670
	return 0;
}

1671
#ifdef CONFIG_BLOCK
1672
void laptop_mode_timer_fn(unsigned long data)
L
Linus Torvalds 已提交
1673
{
1674 1675 1676
	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
1677

1678 1679 1680 1681 1682
	/*
	 * We want to write everything out, not just down to the dirty
	 * threshold
	 */
	if (bdi_has_dirty_io(&q->backing_dev_info))
1683 1684
		bdi_start_writeback(&q->backing_dev_info, nr_pages,
					WB_REASON_LAPTOP_TIMER);
L
Linus Torvalds 已提交
1685 1686 1687 1688 1689 1690 1691
}

/*
 * 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.
 */
1692
void laptop_io_completion(struct backing_dev_info *info)
L
Linus Torvalds 已提交
1693
{
1694
	mod_timer(&info->laptop_mode_wb_timer, jiffies + laptop_mode);
L
Linus Torvalds 已提交
1695 1696 1697 1698 1699 1700 1701 1702 1703
}

/*
 * 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)
{
1704 1705 1706 1707 1708 1709 1710 1711
	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
1712
}
1713
#endif
L
Linus Torvalds 已提交
1714 1715 1716 1717 1718 1719 1720 1721 1722

/*
 * 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
1723
 * thresholds.
L
Linus Torvalds 已提交
1724 1725
 */

1726
void writeback_set_ratelimit(void)
L
Linus Torvalds 已提交
1727
{
1728 1729 1730
	unsigned long background_thresh;
	unsigned long dirty_thresh;
	global_dirty_limits(&background_thresh, &dirty_thresh);
1731
	global_dirty_limit = dirty_thresh;
1732
	ratelimit_pages = dirty_thresh / (num_online_cpus() * 32);
L
Linus Torvalds 已提交
1733 1734 1735 1736
	if (ratelimit_pages < 16)
		ratelimit_pages = 16;
}

1737
static int
1738 1739
ratelimit_handler(struct notifier_block *self, unsigned long action,
		  void *hcpu)
L
Linus Torvalds 已提交
1740
{
1741 1742 1743 1744 1745 1746 1747 1748 1749

	switch (action & ~CPU_TASKS_FROZEN) {
	case CPU_ONLINE:
	case CPU_DEAD:
		writeback_set_ratelimit();
		return NOTIFY_OK;
	default:
		return NOTIFY_DONE;
	}
L
Linus Torvalds 已提交
1750 1751
}

1752
static struct notifier_block ratelimit_nb = {
L
Linus Torvalds 已提交
1753 1754 1755 1756 1757
	.notifier_call	= ratelimit_handler,
	.next		= NULL,
};

/*
1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773
 * 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
1774 1775 1776
 */
void __init page_writeback_init(void)
{
1777
	writeback_set_ratelimit();
L
Linus Torvalds 已提交
1778
	register_cpu_notifier(&ratelimit_nb);
P
Peter Zijlstra 已提交
1779

1780
	fprop_global_init(&writeout_completions, GFP_KERNEL);
L
Linus Torvalds 已提交
1781 1782
}

1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802
/**
 * 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
1803
#define WRITEBACK_TAG_BATCH 4096
1804 1805 1806 1807 1808 1809 1810 1811 1812 1813
	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();
1814 1815
		/* We check 'start' to handle wrapping when end == ~0UL */
	} while (tagged >= WRITEBACK_TAG_BATCH && start);
1816 1817 1818
}
EXPORT_SYMBOL(tag_pages_for_writeback);

1819
/**
1820
 * write_cache_pages - walk the list of dirty pages of the given address space and write all of them.
1821 1822
 * @mapping: address space structure to write
 * @wbc: subtract the number of written pages from *@wbc->nr_to_write
1823 1824
 * @writepage: function called for each page
 * @data: data passed to writepage function
1825
 *
1826
 * If a page is already under I/O, write_cache_pages() skips it, even
1827 1828 1829 1830 1831 1832
 * 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.
1833 1834 1835 1836 1837 1838 1839
 *
 * 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).
1840
 */
1841 1842 1843
int write_cache_pages(struct address_space *mapping,
		      struct writeback_control *wbc, writepage_t writepage,
		      void *data)
1844 1845 1846 1847 1848
{
	int ret = 0;
	int done = 0;
	struct pagevec pvec;
	int nr_pages;
N
Nick Piggin 已提交
1849
	pgoff_t uninitialized_var(writeback_index);
1850 1851
	pgoff_t index;
	pgoff_t end;		/* Inclusive */
1852
	pgoff_t done_index;
N
Nick Piggin 已提交
1853
	int cycled;
1854
	int range_whole = 0;
1855
	int tag;
1856 1857 1858

	pagevec_init(&pvec, 0);
	if (wbc->range_cyclic) {
N
Nick Piggin 已提交
1859 1860 1861 1862 1863 1864
		writeback_index = mapping->writeback_index; /* prev offset */
		index = writeback_index;
		if (index == 0)
			cycled = 1;
		else
			cycled = 0;
1865 1866 1867 1868 1869 1870
		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
1871
		cycled = 1; /* ignore range_cyclic tests */
1872
	}
1873
	if (wbc->sync_mode == WB_SYNC_ALL || wbc->tagged_writepages)
1874 1875 1876
		tag = PAGECACHE_TAG_TOWRITE;
	else
		tag = PAGECACHE_TAG_DIRTY;
1877
retry:
1878
	if (wbc->sync_mode == WB_SYNC_ALL || wbc->tagged_writepages)
1879
		tag_pages_for_writeback(mapping, index, end);
1880
	done_index = index;
N
Nick Piggin 已提交
1881 1882 1883
	while (!done && (index <= end)) {
		int i;

1884
		nr_pages = pagevec_lookup_tag(&pvec, mapping, &index, tag,
N
Nick Piggin 已提交
1885 1886 1887
			      min(end - index, (pgoff_t)PAGEVEC_SIZE-1) + 1);
		if (nr_pages == 0)
			break;
1888 1889 1890 1891 1892

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

			/*
1893 1894 1895 1896 1897
			 * 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.
1898
			 */
1899 1900 1901 1902 1903 1904 1905 1906 1907
			if (page->index > end) {
				/*
				 * can't be range_cyclic (1st pass) because
				 * end == -1 in that case.
				 */
				done = 1;
				break;
			}

1908
			done_index = page->index;
1909

1910 1911
			lock_page(page);

N
Nick Piggin 已提交
1912 1913 1914 1915 1916 1917 1918 1919
			/*
			 * 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.
			 */
1920
			if (unlikely(page->mapping != mapping)) {
N
Nick Piggin 已提交
1921
continue_unlock:
1922 1923 1924 1925
				unlock_page(page);
				continue;
			}

1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936
			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;
			}
1937

1938 1939
			BUG_ON(PageWriteback(page));
			if (!clear_page_dirty_for_io(page))
N
Nick Piggin 已提交
1940
				goto continue_unlock;
1941

1942
			trace_wbc_writepage(wbc, mapping->backing_dev_info);
1943
			ret = (*writepage)(page, wbc, data);
1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957
			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).
					 */
1958
					done_index = page->index + 1;
1959 1960 1961
					done = 1;
					break;
				}
1962
			}
1963

1964 1965 1966 1967 1968 1969 1970 1971 1972 1973
			/*
			 * 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;
1974
			}
1975 1976 1977 1978
		}
		pagevec_release(&pvec);
		cond_resched();
	}
1979
	if (!cycled && !done) {
1980
		/*
N
Nick Piggin 已提交
1981
		 * range_cyclic:
1982 1983 1984
		 * 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
1985
		cycled = 1;
1986
		index = 0;
N
Nick Piggin 已提交
1987
		end = writeback_index - 1;
1988 1989
		goto retry;
	}
1990 1991
	if (wbc->range_cyclic || (range_whole && wbc->nr_to_write > 0))
		mapping->writeback_index = done_index;
1992

1993 1994
	return ret;
}
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020
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)
{
2021 2022 2023
	struct blk_plug plug;
	int ret;

2024 2025 2026 2027
	/* deal with chardevs and other special file */
	if (!mapping->a_ops->writepage)
		return 0;

2028 2029 2030 2031
	blk_start_plug(&plug);
	ret = write_cache_pages(mapping, wbc, __writepage, mapping);
	blk_finish_plug(&plug);
	return ret;
2032
}
2033 2034 2035

EXPORT_SYMBOL(generic_writepages);

L
Linus Torvalds 已提交
2036 2037
int do_writepages(struct address_space *mapping, struct writeback_control *wbc)
{
2038 2039
	int ret;

L
Linus Torvalds 已提交
2040 2041 2042
	if (wbc->nr_to_write <= 0)
		return 0;
	if (mapping->a_ops->writepages)
2043
		ret = mapping->a_ops->writepages(mapping, wbc);
2044 2045 2046
	else
		ret = generic_writepages(mapping, wbc);
	return ret;
L
Linus Torvalds 已提交
2047 2048 2049 2050
}

/**
 * write_one_page - write out a single page and optionally wait on I/O
2051 2052
 * @page: the page to write
 * @wait: if true, wait on writeout
L
Linus Torvalds 已提交
2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087
 *
 * 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);

2088 2089 2090 2091 2092 2093
/*
 * For address_spaces which do not use buffers nor write back.
 */
int __set_page_dirty_no_writeback(struct page *page)
{
	if (!PageDirty(page))
2094
		return !TestSetPageDirty(page);
2095 2096 2097
	return 0;
}

2098 2099 2100 2101 2102 2103
/*
 * 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)
{
T
Tejun Heo 已提交
2104 2105
	trace_writeback_dirty_page(page, mapping);

2106 2107
	if (mapping_cap_account_dirty(mapping)) {
		__inc_zone_page_state(page, NR_FILE_DIRTY);
2108
		__inc_zone_page_state(page, NR_DIRTIED);
2109
		__inc_bdi_stat(mapping->backing_dev_info, BDI_RECLAIMABLE);
2110
		__inc_bdi_stat(mapping->backing_dev_info, BDI_DIRTIED);
2111
		task_io_account_write(PAGE_CACHE_SIZE);
2112 2113
		current->nr_dirtied++;
		this_cpu_inc(bdp_ratelimits);
2114 2115
	}
}
M
Michael Rubin 已提交
2116
EXPORT_SYMBOL(account_page_dirtied);
2117

L
Linus Torvalds 已提交
2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130
/*
 * 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
2131
 * mapping by re-checking page_mapping() inside tree_lock.
L
Linus Torvalds 已提交
2132 2133 2134 2135 2136 2137
 */
int __set_page_dirty_nobuffers(struct page *page)
{
	if (!TestSetPageDirty(page)) {
		struct address_space *mapping = page_mapping(page);
		struct address_space *mapping2;
2138
		unsigned long flags;
L
Linus Torvalds 已提交
2139

2140 2141 2142
		if (!mapping)
			return 1;

2143
		spin_lock_irqsave(&mapping->tree_lock, flags);
2144 2145 2146
		mapping2 = page_mapping(page);
		if (mapping2) { /* Race with truncate? */
			BUG_ON(mapping2 != mapping);
2147
			WARN_ON_ONCE(!PagePrivate(page) && !PageUptodate(page));
2148
			account_page_dirtied(page, mapping);
2149 2150 2151
			radix_tree_tag_set(&mapping->page_tree,
				page_index(page), PAGECACHE_TAG_DIRTY);
		}
2152
		spin_unlock_irqrestore(&mapping->tree_lock, flags);
2153 2154 2155
		if (mapping->host) {
			/* !PageAnon && !swapper_space */
			__mark_inode_dirty(mapping->host, I_DIRTY_PAGES);
L
Linus Torvalds 已提交
2156
		}
2157
		return 1;
L
Linus Torvalds 已提交
2158
	}
2159
	return 0;
L
Linus Torvalds 已提交
2160 2161 2162
}
EXPORT_SYMBOL(__set_page_dirty_nobuffers);

2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180
/*
 * 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
2181 2182 2183 2184 2185 2186 2187 2188
/*
 * 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++;
2189
	account_page_redirty(page);
L
Linus Torvalds 已提交
2190 2191 2192 2193 2194
	return __set_page_dirty_nobuffers(page);
}
EXPORT_SYMBOL(redirty_page_for_writepage);

/*
2195 2196 2197 2198 2199 2200 2201
 * 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
2202 2203 2204
 * 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
2205
int set_page_dirty(struct page *page)
L
Linus Torvalds 已提交
2206 2207 2208 2209 2210
{
	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
2211 2212 2213 2214 2215 2216 2217 2218 2219 2220 2221
		/*
		 * 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);
2222 2223 2224 2225 2226
#ifdef CONFIG_BLOCK
		if (!spd)
			spd = __set_page_dirty_buffers;
#endif
		return (*spd)(page);
L
Linus Torvalds 已提交
2227
	}
2228 2229 2230 2231
	if (!PageDirty(page)) {
		if (!TestSetPageDirty(page))
			return 1;
	}
L
Linus Torvalds 已提交
2232 2233 2234 2235 2236 2237 2238 2239 2240 2241 2242 2243 2244 2245 2246 2247 2248 2249
	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
2250
	lock_page(page);
L
Linus Torvalds 已提交
2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272 2273 2274
	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);

2275 2276
	BUG_ON(!PageLocked(page));

2277 2278 2279 2280 2281 2282 2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304
	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);
2305 2306 2307 2308 2309 2310 2311 2312 2313 2314
		/*
		 * 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.
		 */
2315
		if (TestClearPageDirty(page)) {
2316
			dec_zone_page_state(page, NR_FILE_DIRTY);
2317 2318
			dec_bdi_stat(mapping->backing_dev_info,
					BDI_RECLAIMABLE);
2319
			return 1;
L
Linus Torvalds 已提交
2320
		}
2321
		return 0;
L
Linus Torvalds 已提交
2322
	}
2323
	return TestClearPageDirty(page);
L
Linus Torvalds 已提交
2324
}
2325
EXPORT_SYMBOL(clear_page_dirty_for_io);
L
Linus Torvalds 已提交
2326 2327 2328 2329

int test_clear_page_writeback(struct page *page)
{
	struct address_space *mapping = page_mapping(page);
2330
	unsigned long memcg_flags;
2331 2332 2333
	struct mem_cgroup *memcg;
	bool locked;
	int ret;
L
Linus Torvalds 已提交
2334

2335
	memcg = mem_cgroup_begin_page_stat(page, &locked, &memcg_flags);
L
Linus Torvalds 已提交
2336
	if (mapping) {
P
Peter Zijlstra 已提交
2337
		struct backing_dev_info *bdi = mapping->backing_dev_info;
L
Linus Torvalds 已提交
2338 2339
		unsigned long flags;

N
Nick Piggin 已提交
2340
		spin_lock_irqsave(&mapping->tree_lock, flags);
L
Linus Torvalds 已提交
2341
		ret = TestClearPageWriteback(page);
P
Peter Zijlstra 已提交
2342
		if (ret) {
L
Linus Torvalds 已提交
2343 2344 2345
			radix_tree_tag_clear(&mapping->page_tree,
						page_index(page),
						PAGECACHE_TAG_WRITEBACK);
2346
			if (bdi_cap_account_writeback(bdi)) {
P
Peter Zijlstra 已提交
2347
				__dec_bdi_stat(bdi, BDI_WRITEBACK);
P
Peter Zijlstra 已提交
2348 2349
				__bdi_writeout_inc(bdi);
			}
P
Peter Zijlstra 已提交
2350
		}
N
Nick Piggin 已提交
2351
		spin_unlock_irqrestore(&mapping->tree_lock, flags);
L
Linus Torvalds 已提交
2352 2353 2354
	} else {
		ret = TestClearPageWriteback(page);
	}
2355
	if (ret) {
2356
		mem_cgroup_dec_page_stat(memcg, MEM_CGROUP_STAT_WRITEBACK);
2357
		dec_zone_page_state(page, NR_WRITEBACK);
2358 2359
		inc_zone_page_state(page, NR_WRITTEN);
	}
2360
	mem_cgroup_end_page_stat(memcg, &locked, &memcg_flags);
L
Linus Torvalds 已提交
2361 2362 2363
	return ret;
}

2364
int __test_set_page_writeback(struct page *page, bool keep_write)
L
Linus Torvalds 已提交
2365 2366
{
	struct address_space *mapping = page_mapping(page);
2367
	unsigned long memcg_flags;
2368 2369 2370
	struct mem_cgroup *memcg;
	bool locked;
	int ret;
L
Linus Torvalds 已提交
2371

2372
	memcg = mem_cgroup_begin_page_stat(page, &locked, &memcg_flags);
L
Linus Torvalds 已提交
2373
	if (mapping) {
P
Peter Zijlstra 已提交
2374
		struct backing_dev_info *bdi = mapping->backing_dev_info;
L
Linus Torvalds 已提交
2375 2376
		unsigned long flags;

N
Nick Piggin 已提交
2377
		spin_lock_irqsave(&mapping->tree_lock, flags);
L
Linus Torvalds 已提交
2378
		ret = TestSetPageWriteback(page);
P
Peter Zijlstra 已提交
2379
		if (!ret) {
L
Linus Torvalds 已提交
2380 2381 2382
			radix_tree_tag_set(&mapping->page_tree,
						page_index(page),
						PAGECACHE_TAG_WRITEBACK);
2383
			if (bdi_cap_account_writeback(bdi))
P
Peter Zijlstra 已提交
2384 2385
				__inc_bdi_stat(bdi, BDI_WRITEBACK);
		}
L
Linus Torvalds 已提交
2386 2387 2388 2389
		if (!PageDirty(page))
			radix_tree_tag_clear(&mapping->page_tree,
						page_index(page),
						PAGECACHE_TAG_DIRTY);
2390 2391 2392 2393
		if (!keep_write)
			radix_tree_tag_clear(&mapping->page_tree,
						page_index(page),
						PAGECACHE_TAG_TOWRITE);
N
Nick Piggin 已提交
2394
		spin_unlock_irqrestore(&mapping->tree_lock, flags);
L
Linus Torvalds 已提交
2395 2396 2397
	} else {
		ret = TestSetPageWriteback(page);
	}
2398
	if (!ret) {
2399
		mem_cgroup_inc_page_stat(memcg, MEM_CGROUP_STAT_WRITEBACK);
2400 2401
		inc_zone_page_state(page, NR_WRITEBACK);
	}
2402
	mem_cgroup_end_page_stat(memcg, &locked, &memcg_flags);
L
Linus Torvalds 已提交
2403 2404 2405
	return ret;

}
2406
EXPORT_SYMBOL(__test_set_page_writeback);
L
Linus Torvalds 已提交
2407 2408

/*
N
Nick Piggin 已提交
2409
 * Return true if any of the pages in the mapping are marked with the
L
Linus Torvalds 已提交
2410 2411 2412 2413
 * passed tag.
 */
int mapping_tagged(struct address_space *mapping, int tag)
{
2414
	return radix_tree_tagged(&mapping->page_tree, tag);
L
Linus Torvalds 已提交
2415 2416
}
EXPORT_SYMBOL(mapping_tagged);
2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436

/**
 * wait_for_stable_page() - wait for writeback to finish, if necessary.
 * @page:	The page to wait on.
 *
 * This function determines if the given page is related to a backing device
 * that requires page contents to be held stable during writeback.  If so, then
 * it will wait for any pending writeback to complete.
 */
void wait_for_stable_page(struct page *page)
{
	struct address_space *mapping = page_mapping(page);
	struct backing_dev_info *bdi = mapping->backing_dev_info;

	if (!bdi_cap_stable_pages_required(bdi))
		return;

	wait_on_page_writeback(page);
}
EXPORT_SYMBOL_GPL(wait_for_stable_page);