backing-dev.c 22.0 KB
Newer Older
1 2 3

#include <linux/wait.h>
#include <linux/backing-dev.h>
4 5
#include <linux/kthread.h>
#include <linux/freezer.h>
6
#include <linux/fs.h>
7
#include <linux/pagemap.h>
8
#include <linux/mm.h>
9 10
#include <linux/sched.h>
#include <linux/module.h>
11 12
#include <linux/writeback.h>
#include <linux/device.h>
13
#include <trace/events/writeback.h>
14

15 16
static atomic_long_t bdi_seq = ATOMIC_LONG_INIT(0);

17
struct backing_dev_info default_backing_dev_info = {
18
	.name		= "default",
19 20 21 22 23
	.ra_pages	= VM_MAX_READAHEAD * 1024 / PAGE_CACHE_SIZE,
	.state		= 0,
	.capabilities	= BDI_CAP_MAP_COPY,
};
EXPORT_SYMBOL_GPL(default_backing_dev_info);
24

J
Jörn Engel 已提交
25 26
struct backing_dev_info noop_backing_dev_info = {
	.name		= "noop",
27
	.capabilities	= BDI_CAP_NO_ACCT_AND_WRITEBACK,
J
Jörn Engel 已提交
28 29 30
};
EXPORT_SYMBOL_GPL(noop_backing_dev_info);

31
static struct class *bdi_class;
32 33 34 35 36 37

/*
 * bdi_lock protects updates to bdi_list and bdi_pending_list, as well as
 * reader side protection for bdi_pending_list. bdi_list has RCU reader side
 * locking.
 */
38
DEFINE_SPINLOCK(bdi_lock);
39
LIST_HEAD(bdi_list);
40 41 42 43 44 45 46 47
LIST_HEAD(bdi_pending_list);

static struct task_struct *sync_supers_tsk;
static struct timer_list sync_supers_timer;

static int bdi_sync_supers(void *);
static void sync_supers_timer_fn(unsigned long);

48 49 50 51 52 53 54 55 56 57 58
void bdi_lock_two(struct bdi_writeback *wb1, struct bdi_writeback *wb2)
{
	if (wb1 < wb2) {
		spin_lock(&wb1->list_lock);
		spin_lock_nested(&wb2->list_lock, 1);
	} else {
		spin_lock(&wb2->list_lock);
		spin_lock_nested(&wb1->list_lock, 1);
	}
}

59 60 61 62 63 64 65 66 67 68 69 70 71 72
#ifdef CONFIG_DEBUG_FS
#include <linux/debugfs.h>
#include <linux/seq_file.h>

static struct dentry *bdi_debug_root;

static void bdi_debug_init(void)
{
	bdi_debug_root = debugfs_create_dir("bdi", NULL);
}

static int bdi_debug_stats_show(struct seq_file *m, void *v)
{
	struct backing_dev_info *bdi = m->private;
C
Christoph Hellwig 已提交
73
	struct bdi_writeback *wb = &bdi->wb;
74 75 76
	unsigned long background_thresh;
	unsigned long dirty_thresh;
	unsigned long bdi_thresh;
77
	unsigned long nr_dirty, nr_io, nr_more_io;
78 79
	struct inode *inode;

80
	nr_dirty = nr_io = nr_more_io = 0;
81
	spin_lock(&wb->list_lock);
N
Nick Piggin 已提交
82
	list_for_each_entry(inode, &wb->b_dirty, i_wb_list)
C
Christoph Hellwig 已提交
83
		nr_dirty++;
N
Nick Piggin 已提交
84
	list_for_each_entry(inode, &wb->b_io, i_wb_list)
C
Christoph Hellwig 已提交
85
		nr_io++;
N
Nick Piggin 已提交
86
	list_for_each_entry(inode, &wb->b_more_io, i_wb_list)
C
Christoph Hellwig 已提交
87
		nr_more_io++;
88
	spin_unlock(&wb->list_lock);
89

90 91
	global_dirty_limits(&background_thresh, &dirty_thresh);
	bdi_thresh = bdi_dirty_limit(bdi, dirty_thresh);
92 93 94

#define K(x) ((x) << (PAGE_SHIFT - 10))
	seq_printf(m,
95 96 97 98 99
		   "BdiWriteback:       %10lu kB\n"
		   "BdiReclaimable:     %10lu kB\n"
		   "BdiDirtyThresh:     %10lu kB\n"
		   "DirtyThresh:        %10lu kB\n"
		   "BackgroundThresh:   %10lu kB\n"
100
		   "BdiDirtied:         %10lu kB\n"
101 102 103 104 105 106 107
		   "BdiWritten:         %10lu kB\n"
		   "BdiWriteBandwidth:  %10lu kBps\n"
		   "b_dirty:            %10lu\n"
		   "b_io:               %10lu\n"
		   "b_more_io:          %10lu\n"
		   "bdi_list:           %10u\n"
		   "state:              %10lx\n",
108 109
		   (unsigned long) K(bdi_stat(bdi, BDI_WRITEBACK)),
		   (unsigned long) K(bdi_stat(bdi, BDI_RECLAIMABLE)),
110 111 112
		   K(bdi_thresh),
		   K(dirty_thresh),
		   K(background_thresh),
113
		   (unsigned long) K(bdi_stat(bdi, BDI_DIRTIED)),
114
		   (unsigned long) K(bdi_stat(bdi, BDI_WRITTEN)),
115
		   (unsigned long) K(bdi->write_bandwidth),
116 117 118
		   nr_dirty,
		   nr_io,
		   nr_more_io,
C
Christoph Hellwig 已提交
119
		   !list_empty(&bdi->bdi_list), bdi->state);
120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161
#undef K

	return 0;
}

static int bdi_debug_stats_open(struct inode *inode, struct file *file)
{
	return single_open(file, bdi_debug_stats_show, inode->i_private);
}

static const struct file_operations bdi_debug_stats_fops = {
	.open		= bdi_debug_stats_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
	.release	= single_release,
};

static void bdi_debug_register(struct backing_dev_info *bdi, const char *name)
{
	bdi->debug_dir = debugfs_create_dir(name, bdi_debug_root);
	bdi->debug_stats = debugfs_create_file("stats", 0444, bdi->debug_dir,
					       bdi, &bdi_debug_stats_fops);
}

static void bdi_debug_unregister(struct backing_dev_info *bdi)
{
	debugfs_remove(bdi->debug_stats);
	debugfs_remove(bdi->debug_dir);
}
#else
static inline void bdi_debug_init(void)
{
}
static inline void bdi_debug_register(struct backing_dev_info *bdi,
				      const char *name)
{
}
static inline void bdi_debug_unregister(struct backing_dev_info *bdi)
{
}
#endif

162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191
static ssize_t read_ahead_kb_store(struct device *dev,
				  struct device_attribute *attr,
				  const char *buf, size_t count)
{
	struct backing_dev_info *bdi = dev_get_drvdata(dev);
	char *end;
	unsigned long read_ahead_kb;
	ssize_t ret = -EINVAL;

	read_ahead_kb = simple_strtoul(buf, &end, 10);
	if (*buf && (end[0] == '\0' || (end[0] == '\n' && end[1] == '\0'))) {
		bdi->ra_pages = read_ahead_kb >> (PAGE_SHIFT - 10);
		ret = count;
	}
	return ret;
}

#define K(pages) ((pages) << (PAGE_SHIFT - 10))

#define BDI_SHOW(name, expr)						\
static ssize_t name##_show(struct device *dev,				\
			   struct device_attribute *attr, char *page)	\
{									\
	struct backing_dev_info *bdi = dev_get_drvdata(dev);		\
									\
	return snprintf(page, PAGE_SIZE-1, "%lld\n", (long long)expr);	\
}

BDI_SHOW(read_ahead_kb, K(bdi->ra_pages))

192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209
static ssize_t min_ratio_store(struct device *dev,
		struct device_attribute *attr, const char *buf, size_t count)
{
	struct backing_dev_info *bdi = dev_get_drvdata(dev);
	char *end;
	unsigned int ratio;
	ssize_t ret = -EINVAL;

	ratio = simple_strtoul(buf, &end, 10);
	if (*buf && (end[0] == '\0' || (end[0] == '\n' && end[1] == '\0'))) {
		ret = bdi_set_min_ratio(bdi, ratio);
		if (!ret)
			ret = count;
	}
	return ret;
}
BDI_SHOW(min_ratio, bdi->min_ratio)

210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227
static ssize_t max_ratio_store(struct device *dev,
		struct device_attribute *attr, const char *buf, size_t count)
{
	struct backing_dev_info *bdi = dev_get_drvdata(dev);
	char *end;
	unsigned int ratio;
	ssize_t ret = -EINVAL;

	ratio = simple_strtoul(buf, &end, 10);
	if (*buf && (end[0] == '\0' || (end[0] == '\n' && end[1] == '\0'))) {
		ret = bdi_set_max_ratio(bdi, ratio);
		if (!ret)
			ret = count;
	}
	return ret;
}
BDI_SHOW(max_ratio, bdi->max_ratio)

228 229 230 231
#define __ATTR_RW(attr) __ATTR(attr, 0644, attr##_show, attr##_store)

static struct device_attribute bdi_dev_attrs[] = {
	__ATTR_RW(read_ahead_kb),
232
	__ATTR_RW(min_ratio),
233
	__ATTR_RW(max_ratio),
234 235 236 237 238 239
	__ATTR_NULL,
};

static __init int bdi_class_init(void)
{
	bdi_class = class_create(THIS_MODULE, "bdi");
240 241 242
	if (IS_ERR(bdi_class))
		return PTR_ERR(bdi_class);

243
	bdi_class->dev_attrs = bdi_dev_attrs;
244
	bdi_debug_init();
245 246
	return 0;
}
247
postcore_initcall(bdi_class_init);
248

249 250 251 252
static int __init default_bdi_init(void)
{
	int err;

253 254 255 256
	sync_supers_tsk = kthread_run(bdi_sync_supers, NULL, "sync_supers");
	BUG_ON(IS_ERR(sync_supers_tsk));

	setup_timer(&sync_supers_timer, sync_supers_timer_fn, 0);
257
	bdi_arm_supers_timer();
258

259 260 261
	err = bdi_init(&default_backing_dev_info);
	if (!err)
		bdi_register(&default_backing_dev_info, NULL, "default");
262
	err = bdi_init(&noop_backing_dev_info);
263 264 265 266 267

	return err;
}
subsys_initcall(default_bdi_init);

268 269 270 271 272 273
int bdi_has_dirty_io(struct backing_dev_info *bdi)
{
	return wb_has_dirty_io(&bdi->wb);
}

/*
274
 * kupdated() used to do this. We cannot do it from the bdi_forker_thread()
275 276
 * or we risk deadlocking on ->s_umount. The longer term solution would be
 * to implement sync_supers_bdi() or similar and simply do it from the
277
 * bdi writeback thread individually.
278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295
 */
static int bdi_sync_supers(void *unused)
{
	set_user_nice(current, 0);

	while (!kthread_should_stop()) {
		set_current_state(TASK_INTERRUPTIBLE);
		schedule();

		/*
		 * Do this periodically, like kupdated() did before.
		 */
		sync_supers();
	}

	return 0;
}

296
void bdi_arm_supers_timer(void)
297 298 299
{
	unsigned long next;

300 301 302
	if (!dirty_writeback_interval)
		return;

303 304 305 306 307 308 309
	next = msecs_to_jiffies(dirty_writeback_interval * 10) + jiffies;
	mod_timer(&sync_supers_timer, round_jiffies_up(next));
}

static void sync_supers_timer_fn(unsigned long unused)
{
	wake_up_process(sync_supers_tsk);
310
	bdi_arm_supers_timer();
311 312
}

313 314 315 316 317 318
static void wakeup_timer_fn(unsigned long data)
{
	struct backing_dev_info *bdi = (struct backing_dev_info *)data;

	spin_lock_bh(&bdi->wb_lock);
	if (bdi->wb.task) {
319
		trace_writeback_wake_thread(bdi);
320 321 322 323 324 325 326
		wake_up_process(bdi->wb.task);
	} else {
		/*
		 * When bdi tasks are inactive for long time, they are killed.
		 * In this case we have to wake-up the forker thread which
		 * should create and run the bdi thread.
		 */
327
		trace_writeback_wake_forker_thread(bdi);
328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351
		wake_up_process(default_backing_dev_info.wb.task);
	}
	spin_unlock_bh(&bdi->wb_lock);
}

/*
 * This function is used when the first inode for this bdi is marked dirty. It
 * wakes-up the corresponding bdi thread which should then take care of the
 * periodic background write-out of dirty inodes. Since the write-out would
 * starts only 'dirty_writeback_interval' centisecs from now anyway, we just
 * set up a timer which wakes the bdi thread up later.
 *
 * Note, we wouldn't bother setting up the timer, but this function is on the
 * fast-path (used by '__mark_inode_dirty()'), so we save few context switches
 * by delaying the wake-up.
 */
void bdi_wakeup_thread_delayed(struct backing_dev_info *bdi)
{
	unsigned long timeout;

	timeout = msecs_to_jiffies(dirty_writeback_interval * 10);
	mod_timer(&bdi->wb.wakeup_timer, jiffies + timeout);
}

352 353 354 355 356 357 358 359 360 361 362 363
/*
 * Calculate the longest interval (jiffies) bdi threads are allowed to be
 * inactive.
 */
static unsigned long bdi_longest_inactive(void)
{
	unsigned long interval;

	interval = msecs_to_jiffies(dirty_writeback_interval * 10);
	return max(5UL * 60 * HZ, interval);
}

364 365 366 367 368 369 370 371 372 373 374
/*
 * Clear pending bit and wakeup anybody waiting for flusher thread creation or
 * shutdown
 */
static void bdi_clear_pending(struct backing_dev_info *bdi)
{
	clear_bit(BDI_pending, &bdi->state);
	smp_mb__after_clear_bit();
	wake_up_bit(&bdi->state, BDI_pending);
}

375
static int bdi_forker_thread(void *ptr)
376 377 378
{
	struct bdi_writeback *me = ptr;

P
Peter Zijlstra 已提交
379
	current->flags |= PF_SWAPWRITE;
C
Christoph Hellwig 已提交
380 381 382 383 384 385
	set_freezable();

	/*
	 * Our parent may run at a different priority, just set us to normal
	 */
	set_user_nice(current, 0);
386 387

	for (;;) {
388
		struct task_struct *task = NULL;
389
		struct backing_dev_info *bdi;
390 391 392
		enum {
			NO_ACTION,   /* Nothing to do */
			FORK_THREAD, /* Fork bdi thread */
393
			KILL_THREAD, /* Kill inactive bdi thread */
394
		} action = NO_ACTION;
395 396 397 398 399

		/*
		 * Temporary measure, we want to make sure we don't see
		 * dirty data on the default backing_dev_info
		 */
400 401
		if (wb_has_dirty_io(me) || !list_empty(&me->bdi->work_list)) {
			del_timer(&me->wakeup_timer);
402
			wb_do_writeback(me, 0);
403
		}
404

405
		spin_lock_bh(&bdi_lock);
406 407 408 409 410 411 412
		/*
		 * In the following loop we are going to check whether we have
		 * some work to do without any synchronization with tasks
		 * waking us up to do work for them. So we have to set task
		 * state already here so that we don't miss wakeups coming
		 * after we verify some condition.
		 */
413
		set_current_state(TASK_INTERRUPTIBLE);
414

415
		list_for_each_entry(bdi, &bdi_list, bdi_list) {
416 417 418 419
			bool have_dirty_io;

			if (!bdi_cap_writeback_dirty(bdi) ||
			     bdi_cap_flush_forker(bdi))
420 421
				continue;

422 423 424
			WARN(!test_bit(BDI_registered, &bdi->state),
			     "bdi %p/%s is not registered!\n", bdi, bdi->name);

425 426
			have_dirty_io = !list_empty(&bdi->work_list) ||
					wb_has_dirty_io(&bdi->wb);
427 428

			/*
429 430
			 * If the bdi has work to do, but the thread does not
			 * exist - create it.
431
			 */
432 433 434 435 436 437 438 439 440
			if (!bdi->wb.task && have_dirty_io) {
				/*
				 * Set the pending bit - if someone will try to
				 * unregister this bdi - it'll wait on this bit.
				 */
				set_bit(BDI_pending, &bdi->state);
				action = FORK_THREAD;
				break;
			}
441

442 443
			spin_lock(&bdi->wb_lock);

444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459
			/*
			 * If there is no work to do and the bdi thread was
			 * inactive long enough - kill it. The wb_lock is taken
			 * to make sure no-one adds more work to this bdi and
			 * wakes the bdi thread up.
			 */
			if (bdi->wb.task && !have_dirty_io &&
			    time_after(jiffies, bdi->wb.last_active +
						bdi_longest_inactive())) {
				task = bdi->wb.task;
				bdi->wb.task = NULL;
				spin_unlock(&bdi->wb_lock);
				set_bit(BDI_pending, &bdi->state);
				action = KILL_THREAD;
				break;
			}
460
			spin_unlock(&bdi->wb_lock);
461
		}
462
		spin_unlock_bh(&bdi_lock);
463

464 465 466 467
		/* Keep working if default bdi still has things to do */
		if (!list_empty(&me->bdi->work_list))
			__set_current_state(TASK_RUNNING);

468 469 470
		switch (action) {
		case FORK_THREAD:
			__set_current_state(TASK_RUNNING);
471 472
			task = kthread_create(bdi_writeback_thread, &bdi->wb,
					      "flush-%s", dev_name(bdi->dev));
473 474 475
			if (IS_ERR(task)) {
				/*
				 * If thread creation fails, force writeout of
476 477
				 * the bdi from the thread. Hopefully 1024 is
				 * large enough for efficient IO.
478
				 */
479
				writeback_inodes_wb(&bdi->wb, 1024);
480 481 482 483
			} else {
				/*
				 * The spinlock makes sure we do not lose
				 * wake-ups when racing with 'bdi_queue_work()'.
484 485
				 * And as soon as the bdi thread is visible, we
				 * can start it.
486
				 */
487
				spin_lock_bh(&bdi->wb_lock);
488
				bdi->wb.task = task;
489
				spin_unlock_bh(&bdi->wb_lock);
490
				wake_up_process(task);
491
			}
492
			bdi_clear_pending(bdi);
493 494 495 496 497
			break;

		case KILL_THREAD:
			__set_current_state(TASK_RUNNING);
			kthread_stop(task);
498
			bdi_clear_pending(bdi);
499
			break;
500

501
		case NO_ACTION:
502 503 504 505 506 507 508 509 510
			if (!wb_has_dirty_io(me) || !dirty_writeback_interval)
				/*
				 * There are no dirty data. The only thing we
				 * should now care about is checking for
				 * inactive bdi threads and killing them. Thus,
				 * let's sleep for longer time, save energy and
				 * be friendly for battery-driven devices.
				 */
				schedule_timeout(bdi_longest_inactive());
511
			else
512
				schedule_timeout(msecs_to_jiffies(dirty_writeback_interval * 10));
513
			try_to_freeze();
514
			break;
515 516 517 518 519 520
		}
	}

	return 0;
}

521 522 523 524 525 526 527 528 529
/*
 * Remove bdi from bdi_list, and ensure that it is no longer visible
 */
static void bdi_remove_from_list(struct backing_dev_info *bdi)
{
	spin_lock_bh(&bdi_lock);
	list_del_rcu(&bdi->bdi_list);
	spin_unlock_bh(&bdi_lock);

530
	synchronize_rcu_expedited();
531 532
}

533 534 535 536 537 538
int bdi_register(struct backing_dev_info *bdi, struct device *parent,
		const char *fmt, ...)
{
	va_list args;
	struct device *dev;

539
	if (bdi->dev)	/* The driver needs to use separate queues per device */
540
		return 0;
541

542
	va_start(args, fmt);
543
	dev = device_create_vargs(bdi_class, parent, MKDEV(0, 0), bdi, fmt, args);
544
	va_end(args);
545 546
	if (IS_ERR(dev))
		return PTR_ERR(dev);
547

548 549
	bdi->dev = dev;

550 551 552 553 554 555 556 557
	/*
	 * Just start the forker thread for our default backing_dev_info,
	 * and add other bdi's to the list. They will get a thread created
	 * on-demand when they need it.
	 */
	if (bdi_cap_flush_forker(bdi)) {
		struct bdi_writeback *wb = &bdi->wb;

558
		wb->task = kthread_run(bdi_forker_thread, wb, "bdi-%s",
559
						dev_name(dev));
560 561
		if (IS_ERR(wb->task))
			return PTR_ERR(wb->task);
562 563 564
	}

	bdi_debug_register(bdi, dev_name(dev));
565
	set_bit(BDI_registered, &bdi->state);
566 567 568 569 570

	spin_lock_bh(&bdi_lock);
	list_add_tail_rcu(&bdi->bdi_list, &bdi_list);
	spin_unlock_bh(&bdi_lock);

571
	trace_writeback_bdi_register(bdi);
572
	return 0;
573 574 575 576 577 578 579 580 581
}
EXPORT_SYMBOL(bdi_register);

int bdi_register_dev(struct backing_dev_info *bdi, dev_t dev)
{
	return bdi_register(bdi, NULL, "%u:%u", MAJOR(dev), MINOR(dev));
}
EXPORT_SYMBOL(bdi_register_dev);

582 583 584 585
/*
 * Remove bdi from the global list and shutdown any threads we have running
 */
static void bdi_wb_shutdown(struct backing_dev_info *bdi)
586
{
587 588 589 590
	if (!bdi_cap_writeback_dirty(bdi))
		return;

	/*
591
	 * Make sure nobody finds us on the bdi_list anymore
592
	 */
593
	bdi_remove_from_list(bdi);
594 595

	/*
596
	 * If setup is pending, wait for that to complete first
597
	 */
598 599
	wait_on_bit(&bdi->state, BDI_pending, bdi_sched_wait,
			TASK_UNINTERRUPTIBLE);
600 601

	/*
C
Christoph Hellwig 已提交
602
	 * Finally, kill the kernel thread. We don't need to be RCU
603 604 605
	 * safe anymore, since the bdi is gone from visibility. Force
	 * unfreeze of the thread before calling kthread_stop(), otherwise
	 * it would never exet if it is currently stuck in the refrigerator.
606
	 */
C
Christoph Hellwig 已提交
607 608 609
	if (bdi->wb.task) {
		thaw_process(bdi->wb.task);
		kthread_stop(bdi->wb.task);
610
	}
611 612
}

613 614 615 616 617 618 619 620 621 622
/*
 * This bdi is going away now, make sure that no super_blocks point to it
 */
static void bdi_prune_sb(struct backing_dev_info *bdi)
{
	struct super_block *sb;

	spin_lock(&sb_lock);
	list_for_each_entry(sb, &super_blocks, s_list) {
		if (sb->s_bdi == bdi)
623
			sb->s_bdi = &default_backing_dev_info;
624 625 626 627
	}
	spin_unlock(&sb_lock);
}

628 629 630
void bdi_unregister(struct backing_dev_info *bdi)
{
	if (bdi->dev) {
631
		bdi_set_min_ratio(bdi, 0);
632
		trace_writeback_bdi_unregister(bdi);
633
		bdi_prune_sb(bdi);
634
		del_timer_sync(&bdi->wb.wakeup_timer);
635

636 637
		if (!bdi_cap_flush_forker(bdi))
			bdi_wb_shutdown(bdi);
638
		bdi_debug_unregister(bdi);
639 640 641 642 643
		device_unregister(bdi->dev);
		bdi->dev = NULL;
	}
}
EXPORT_SYMBOL(bdi_unregister);
644

645 646 647 648 649 650 651 652 653
static void bdi_wb_init(struct bdi_writeback *wb, struct backing_dev_info *bdi)
{
	memset(wb, 0, sizeof(*wb));

	wb->bdi = bdi;
	wb->last_old_flush = jiffies;
	INIT_LIST_HEAD(&wb->b_dirty);
	INIT_LIST_HEAD(&wb->b_io);
	INIT_LIST_HEAD(&wb->b_more_io);
654
	spin_lock_init(&wb->list_lock);
655 656 657
	setup_timer(&wb->wakeup_timer, wakeup_timer_fn, (unsigned long)bdi);
}

658 659 660 661 662
/*
 * Initial write bandwidth: 100 MB/s
 */
#define INIT_BW		(100 << (20 - PAGE_SHIFT))

663 664
int bdi_init(struct backing_dev_info *bdi)
{
665
	int i, err;
666

667 668
	bdi->dev = NULL;

669
	bdi->min_ratio = 0;
670 671
	bdi->max_ratio = 100;
	bdi->max_prop_frac = PROP_FRAC_BASE;
672
	spin_lock_init(&bdi->wb_lock);
673
	INIT_LIST_HEAD(&bdi->bdi_list);
674 675 676 677
	INIT_LIST_HEAD(&bdi->work_list);

	bdi_wb_init(&bdi->wb, bdi);

678
	for (i = 0; i < NR_BDI_STAT_ITEMS; i++) {
679
		err = percpu_counter_init(&bdi->bdi_stat[i], 0);
P
Peter Zijlstra 已提交
680 681 682 683 684
		if (err)
			goto err;
	}

	bdi->dirty_exceeded = 0;
685 686 687 688

	bdi->bw_time_stamp = jiffies;
	bdi->written_stamp = 0;

689
	bdi->balanced_dirty_ratelimit = INIT_BW;
W
Wu Fengguang 已提交
690
	bdi->dirty_ratelimit = INIT_BW;
691 692 693
	bdi->write_bandwidth = INIT_BW;
	bdi->avg_write_bandwidth = INIT_BW;

P
Peter Zijlstra 已提交
694 695 696 697
	err = prop_local_init_percpu(&bdi->completions);

	if (err) {
err:
698
		while (i--)
P
Peter Zijlstra 已提交
699
			percpu_counter_destroy(&bdi->bdi_stat[i]);
700 701 702 703 704 705 706 707 708 709
	}

	return err;
}
EXPORT_SYMBOL(bdi_init);

void bdi_destroy(struct backing_dev_info *bdi)
{
	int i;

710 711 712 713 714 715 716
	/*
	 * Splice our entries to the default_backing_dev_info, if this
	 * bdi disappears
	 */
	if (bdi_has_dirty_io(bdi)) {
		struct bdi_writeback *dst = &default_backing_dev_info.wb;

717
		bdi_lock_two(&bdi->wb, dst);
718 719 720
		list_splice(&bdi->wb.b_dirty, &dst->b_dirty);
		list_splice(&bdi->wb.b_io, &dst->b_io);
		list_splice(&bdi->wb.b_more_io, &dst->b_more_io);
721 722
		spin_unlock(&bdi->wb.list_lock);
		spin_unlock(&dst->list_lock);
723
	}
724

725 726
	bdi_unregister(bdi);

727 728
	for (i = 0; i < NR_BDI_STAT_ITEMS; i++)
		percpu_counter_destroy(&bdi->bdi_stat[i]);
P
Peter Zijlstra 已提交
729 730

	prop_local_destroy_percpu(&bdi->completions);
731 732 733
}
EXPORT_SYMBOL(bdi_destroy);

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
/*
 * For use from filesystems to quickly init and register a bdi associated
 * with dirty writeback
 */
int bdi_setup_and_register(struct backing_dev_info *bdi, char *name,
			   unsigned int cap)
{
	char tmp[32];
	int err;

	bdi->name = name;
	bdi->capabilities = cap;
	err = bdi_init(bdi);
	if (err)
		return err;

	sprintf(tmp, "%.28s%s", name, "-%d");
	err = bdi_register(bdi, NULL, tmp, atomic_long_inc_return(&bdi_seq));
	if (err) {
		bdi_destroy(bdi);
		return err;
	}

	return 0;
}
EXPORT_SYMBOL(bdi_setup_and_register);

761 762 763 764
static wait_queue_head_t congestion_wqh[2] = {
		__WAIT_QUEUE_HEAD_INITIALIZER(congestion_wqh[0]),
		__WAIT_QUEUE_HEAD_INITIALIZER(congestion_wqh[1])
	};
765
static atomic_t nr_bdi_congested[2];
766

767
void clear_bdi_congested(struct backing_dev_info *bdi, int sync)
768 769
{
	enum bdi_state bit;
770
	wait_queue_head_t *wqh = &congestion_wqh[sync];
771

772
	bit = sync ? BDI_sync_congested : BDI_async_congested;
773 774
	if (test_and_clear_bit(bit, &bdi->state))
		atomic_dec(&nr_bdi_congested[sync]);
775 776 777 778 779 780
	smp_mb__after_clear_bit();
	if (waitqueue_active(wqh))
		wake_up(wqh);
}
EXPORT_SYMBOL(clear_bdi_congested);

781
void set_bdi_congested(struct backing_dev_info *bdi, int sync)
782 783 784
{
	enum bdi_state bit;

785
	bit = sync ? BDI_sync_congested : BDI_async_congested;
786 787
	if (!test_and_set_bit(bit, &bdi->state))
		atomic_inc(&nr_bdi_congested[sync]);
788 789 790 791 792
}
EXPORT_SYMBOL(set_bdi_congested);

/**
 * congestion_wait - wait for a backing_dev to become uncongested
793
 * @sync: SYNC or ASYNC IO
794 795 796 797 798 799
 * @timeout: timeout in jiffies
 *
 * Waits for up to @timeout jiffies for a backing_dev (any backing_dev) to exit
 * write congestion.  If no backing_devs are congested then just wait for the
 * next write to be completed.
 */
800
long congestion_wait(int sync, long timeout)
801 802
{
	long ret;
803
	unsigned long start = jiffies;
804
	DEFINE_WAIT(wait);
805
	wait_queue_head_t *wqh = &congestion_wqh[sync];
806 807 808 809

	prepare_to_wait(wqh, &wait, TASK_UNINTERRUPTIBLE);
	ret = io_schedule_timeout(timeout);
	finish_wait(wqh, &wait);
810 811 812 813

	trace_writeback_congestion_wait(jiffies_to_usecs(timeout),
					jiffies_to_usecs(jiffies - start));

814 815 816
	return ret;
}
EXPORT_SYMBOL(congestion_wait);
P
Peter Zijlstra 已提交
817

818 819 820 821 822 823 824 825 826 827 828
/**
 * wait_iff_congested - Conditionally wait for a backing_dev to become uncongested or a zone to complete writes
 * @zone: A zone to check if it is heavily congested
 * @sync: SYNC or ASYNC IO
 * @timeout: timeout in jiffies
 *
 * In the event of a congested backing_dev (any backing_dev) and the given
 * @zone has experienced recent congestion, this waits for up to @timeout
 * jiffies for either a BDI to exit congestion of the given @sync queue
 * or a write to complete.
 *
L
Lucas De Marchi 已提交
829
 * In the absence of zone congestion, cond_resched() is called to yield
830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871
 * the processor if necessary but otherwise does not sleep.
 *
 * The return value is 0 if the sleep is for the full timeout. Otherwise,
 * it is the number of jiffies that were still remaining when the function
 * returned. return_value == timeout implies the function did not sleep.
 */
long wait_iff_congested(struct zone *zone, int sync, long timeout)
{
	long ret;
	unsigned long start = jiffies;
	DEFINE_WAIT(wait);
	wait_queue_head_t *wqh = &congestion_wqh[sync];

	/*
	 * If there is no congestion, or heavy congestion is not being
	 * encountered in the current zone, yield if necessary instead
	 * of sleeping on the congestion queue
	 */
	if (atomic_read(&nr_bdi_congested[sync]) == 0 ||
			!zone_is_reclaim_congested(zone)) {
		cond_resched();

		/* In case we scheduled, work out time remaining */
		ret = timeout - (jiffies - start);
		if (ret < 0)
			ret = 0;

		goto out;
	}

	/* Sleep until uncongested or a write happens */
	prepare_to_wait(wqh, &wait, TASK_UNINTERRUPTIBLE);
	ret = io_schedule_timeout(timeout);
	finish_wait(wqh, &wait);

out:
	trace_writeback_wait_iff_congested(jiffies_to_usecs(timeout),
					jiffies_to_usecs(jiffies - start));

	return ret;
}
EXPORT_SYMBOL(wait_iff_congested);