blk-throttle.c 43.1 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11
/*
 * Interface for controlling IO bandwidth on a request queue
 *
 * Copyright (C) 2010 Vivek Goyal <vgoyal@redhat.com>
 */

#include <linux/module.h>
#include <linux/slab.h>
#include <linux/blkdev.h>
#include <linux/bio.h>
#include <linux/blktrace_api.h>
12
#include <linux/blk-cgroup.h>
13
#include "blk.h"
14 15 16 17 18 19 20 21 22 23

/* Max dispatch from a group in 1 round */
static int throtl_grp_quantum = 8;

/* Total max dispatch from all groups in one round */
static int throtl_quantum = 32;

/* Throttling is performed over 100ms slice and after that slice is renewed */
static unsigned long throtl_slice = HZ/10;	/* 100 ms */

Tejun Heo's avatar
Tejun Heo committed
24
static struct blkcg_policy blkcg_policy_throtl;
25

26 27 28
/* A workqueue to queue throttle related work */
static struct workqueue_struct *kthrotld_workqueue;

29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57
/*
 * To implement hierarchical throttling, throtl_grps form a tree and bios
 * are dispatched upwards level by level until they reach the top and get
 * issued.  When dispatching bios from the children and local group at each
 * level, if the bios are dispatched into a single bio_list, there's a risk
 * of a local or child group which can queue many bios at once filling up
 * the list starving others.
 *
 * To avoid such starvation, dispatched bios are queued separately
 * according to where they came from.  When they are again dispatched to
 * the parent, they're popped in round-robin order so that no single source
 * hogs the dispatch window.
 *
 * throtl_qnode is used to keep the queued bios separated by their sources.
 * Bios are queued to throtl_qnode which in turn is queued to
 * throtl_service_queue and then dispatched in round-robin order.
 *
 * It's also used to track the reference counts on blkg's.  A qnode always
 * belongs to a throtl_grp and gets queued on itself or the parent, so
 * incrementing the reference of the associated throtl_grp when a qnode is
 * queued and decrementing when dequeued is enough to keep the whole blkg
 * tree pinned while bios are in flight.
 */
struct throtl_qnode {
	struct list_head	node;		/* service_queue->queued[] */
	struct bio_list		bios;		/* queued bios */
	struct throtl_grp	*tg;		/* tg this qnode belongs to */
};

58
struct throtl_service_queue {
59 60
	struct throtl_service_queue *parent_sq;	/* the parent service_queue */

61 62 63 64
	/*
	 * Bios queued directly to this service_queue or dispatched from
	 * children throtl_grp's.
	 */
65
	struct list_head	queued[2];	/* throtl_qnode [READ/WRITE] */
66 67 68 69 70 71
	unsigned int		nr_queued[2];	/* number of queued bios */

	/*
	 * RB tree of active children throtl_grp's, which are sorted by
	 * their ->disptime.
	 */
72 73 74 75
	struct rb_root		pending_tree;	/* RB tree of active tgs */
	struct rb_node		*first_pending;	/* first node in the tree */
	unsigned int		nr_pending;	/* # queued in the tree */
	unsigned long		first_pending_disptime;	/* disptime of the first tg */
76
	struct timer_list	pending_timer;	/* fires on first_pending_disptime */
77 78
};

79 80
enum tg_state_flags {
	THROTL_TG_PENDING	= 1 << 0,	/* on parent's pending tree */
81
	THROTL_TG_WAS_EMPTY	= 1 << 1,	/* bio_lists[] became non-empty */
82 83
};

84 85 86
#define rb_entry_tg(node)	rb_entry((node), struct throtl_grp, rb_node)

struct throtl_grp {
87 88 89
	/* must be the first member */
	struct blkg_policy_data pd;

90
	/* active throtl group service_queue member */
91 92
	struct rb_node rb_node;

93 94 95
	/* throtl_data this group belongs to */
	struct throtl_data *td;

96 97 98
	/* this group's service queue */
	struct throtl_service_queue service_queue;

99 100 101 102 103 104 105 106 107 108 109
	/*
	 * qnode_on_self is used when bios are directly queued to this
	 * throtl_grp so that local bios compete fairly with bios
	 * dispatched from children.  qnode_on_parent is used when bios are
	 * dispatched from this throtl_grp into its parent and will compete
	 * with the sibling qnode_on_parents and the parent's
	 * qnode_on_self.
	 */
	struct throtl_qnode qnode_on_self[2];
	struct throtl_qnode qnode_on_parent[2];

110 111 112 113 114 115 116 117 118
	/*
	 * Dispatch time in jiffies. This is the estimated time when group
	 * will unthrottle and is ready to dispatch more bio. It is used as
	 * key to sort active groups in service tree.
	 */
	unsigned long disptime;

	unsigned int flags;

119 120 121
	/* are there any throtl rules between this group and td? */
	bool has_rules[2];

122 123 124
	/* bytes per second rate limits */
	uint64_t bps[2];

125 126 127
	/* IOPS limits */
	unsigned int iops[2];

128 129
	/* Number of bytes disptached in current slice */
	uint64_t bytes_disp[2];
130 131
	/* Number of bio's dispatched in current slice */
	unsigned int io_disp[2];
132 133 134 135 136 137 138 139 140

	/* When did we start a new slice */
	unsigned long slice_start[2];
	unsigned long slice_end[2];
};

struct throtl_data
{
	/* service tree for active throtl groups */
141
	struct throtl_service_queue service_queue;
142 143 144 145 146 147 148

	struct request_queue *queue;

	/* Total Number of queued bios on READ and WRITE lists */
	unsigned int nr_queued[2];

	/*
149
	 * number of total undestroyed groups
150 151 152 153
	 */
	unsigned int nr_undestroyed_grps;

	/* Work for dispatching throttled bios */
154
	struct work_struct dispatch_work;
155 156
};

157 158
static void throtl_pending_timer_fn(unsigned long arg);

159 160 161 162 163
static inline struct throtl_grp *pd_to_tg(struct blkg_policy_data *pd)
{
	return pd ? container_of(pd, struct throtl_grp, pd) : NULL;
}

Tejun Heo's avatar
Tejun Heo committed
164
static inline struct throtl_grp *blkg_to_tg(struct blkcg_gq *blkg)
165
{
166
	return pd_to_tg(blkg_to_pd(blkg, &blkcg_policy_throtl));
167 168
}

Tejun Heo's avatar
Tejun Heo committed
169
static inline struct blkcg_gq *tg_to_blkg(struct throtl_grp *tg)
170
{
171
	return pd_to_blkg(&tg->pd);
172 173
}

174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224
/**
 * sq_to_tg - return the throl_grp the specified service queue belongs to
 * @sq: the throtl_service_queue of interest
 *
 * Return the throtl_grp @sq belongs to.  If @sq is the top-level one
 * embedded in throtl_data, %NULL is returned.
 */
static struct throtl_grp *sq_to_tg(struct throtl_service_queue *sq)
{
	if (sq && sq->parent_sq)
		return container_of(sq, struct throtl_grp, service_queue);
	else
		return NULL;
}

/**
 * sq_to_td - return throtl_data the specified service queue belongs to
 * @sq: the throtl_service_queue of interest
 *
 * A service_queue can be embeded in either a throtl_grp or throtl_data.
 * Determine the associated throtl_data accordingly and return it.
 */
static struct throtl_data *sq_to_td(struct throtl_service_queue *sq)
{
	struct throtl_grp *tg = sq_to_tg(sq);

	if (tg)
		return tg->td;
	else
		return container_of(sq, struct throtl_data, service_queue);
}

/**
 * throtl_log - log debug message via blktrace
 * @sq: the service_queue being reported
 * @fmt: printf format string
 * @args: printf args
 *
 * The messages are prefixed with "throtl BLKG_NAME" if @sq belongs to a
 * throtl_grp; otherwise, just "throtl".
 *
 * TODO: this should be made a function and name formatting should happen
 * after testing whether blktrace is enabled.
 */
#define throtl_log(sq, fmt, args...)	do {				\
	struct throtl_grp *__tg = sq_to_tg((sq));			\
	struct throtl_data *__td = sq_to_td((sq));			\
									\
	(void)__td;							\
	if ((__tg)) {							\
		char __pbuf[128];					\
225
									\
226 227 228 229 230
		blkg_path(tg_to_blkg(__tg), __pbuf, sizeof(__pbuf));	\
		blk_add_trace_msg(__td->queue, "throtl %s " fmt, __pbuf, ##args); \
	} else {							\
		blk_add_trace_msg(__td->queue, "throtl " fmt, ##args);	\
	}								\
231
} while (0)
232

233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315
static void throtl_qnode_init(struct throtl_qnode *qn, struct throtl_grp *tg)
{
	INIT_LIST_HEAD(&qn->node);
	bio_list_init(&qn->bios);
	qn->tg = tg;
}

/**
 * throtl_qnode_add_bio - add a bio to a throtl_qnode and activate it
 * @bio: bio being added
 * @qn: qnode to add bio to
 * @queued: the service_queue->queued[] list @qn belongs to
 *
 * Add @bio to @qn and put @qn on @queued if it's not already on.
 * @qn->tg's reference count is bumped when @qn is activated.  See the
 * comment on top of throtl_qnode definition for details.
 */
static void throtl_qnode_add_bio(struct bio *bio, struct throtl_qnode *qn,
				 struct list_head *queued)
{
	bio_list_add(&qn->bios, bio);
	if (list_empty(&qn->node)) {
		list_add_tail(&qn->node, queued);
		blkg_get(tg_to_blkg(qn->tg));
	}
}

/**
 * throtl_peek_queued - peek the first bio on a qnode list
 * @queued: the qnode list to peek
 */
static struct bio *throtl_peek_queued(struct list_head *queued)
{
	struct throtl_qnode *qn = list_first_entry(queued, struct throtl_qnode, node);
	struct bio *bio;

	if (list_empty(queued))
		return NULL;

	bio = bio_list_peek(&qn->bios);
	WARN_ON_ONCE(!bio);
	return bio;
}

/**
 * throtl_pop_queued - pop the first bio form a qnode list
 * @queued: the qnode list to pop a bio from
 * @tg_to_put: optional out argument for throtl_grp to put
 *
 * Pop the first bio from the qnode list @queued.  After popping, the first
 * qnode is removed from @queued if empty or moved to the end of @queued so
 * that the popping order is round-robin.
 *
 * When the first qnode is removed, its associated throtl_grp should be put
 * too.  If @tg_to_put is NULL, this function automatically puts it;
 * otherwise, *@tg_to_put is set to the throtl_grp to put and the caller is
 * responsible for putting it.
 */
static struct bio *throtl_pop_queued(struct list_head *queued,
				     struct throtl_grp **tg_to_put)
{
	struct throtl_qnode *qn = list_first_entry(queued, struct throtl_qnode, node);
	struct bio *bio;

	if (list_empty(queued))
		return NULL;

	bio = bio_list_pop(&qn->bios);
	WARN_ON_ONCE(!bio);

	if (bio_list_empty(&qn->bios)) {
		list_del_init(&qn->node);
		if (tg_to_put)
			*tg_to_put = qn->tg;
		else
			blkg_put(tg_to_blkg(qn->tg));
	} else {
		list_move_tail(&qn->node, queued);
	}

	return bio;
}

316
/* init a service_queue, assumes the caller zeroed it */
317
static void throtl_service_queue_init(struct throtl_service_queue *sq)
318
{
319 320
	INIT_LIST_HEAD(&sq->queued[0]);
	INIT_LIST_HEAD(&sq->queued[1]);
321
	sq->pending_tree = RB_ROOT;
322 323 324 325
	setup_timer(&sq->pending_timer, throtl_pending_timer_fn,
		    (unsigned long)sq);
}

326 327
static struct blkg_policy_data *throtl_pd_alloc(gfp_t gfp, int node)
{
328
	struct throtl_grp *tg;
329
	int rw;
330 331 332

	tg = kzalloc_node(sizeof(*tg), gfp, node);
	if (!tg)
333
		return NULL;
334

335 336 337 338 339 340 341 342 343 344 345 346 347
	throtl_service_queue_init(&tg->service_queue);

	for (rw = READ; rw <= WRITE; rw++) {
		throtl_qnode_init(&tg->qnode_on_self[rw], tg);
		throtl_qnode_init(&tg->qnode_on_parent[rw], tg);
	}

	RB_CLEAR_NODE(&tg->rb_node);
	tg->bps[READ] = -1;
	tg->bps[WRITE] = -1;
	tg->iops[READ] = -1;
	tg->iops[WRITE] = -1;

348
	return &tg->pd;
349 350
}

351
static void throtl_pd_init(struct blkg_policy_data *pd)
352
{
353 354
	struct throtl_grp *tg = pd_to_tg(pd);
	struct blkcg_gq *blkg = tg_to_blkg(tg);
355
	struct throtl_data *td = blkg->q->td;
356
	struct throtl_service_queue *sq = &tg->service_queue;
357

358
	/*
359
	 * If on the default hierarchy, we switch to properly hierarchical
360 361 362 363 364
	 * behavior where limits on a given throtl_grp are applied to the
	 * whole subtree rather than just the group itself.  e.g. If 16M
	 * read_bps limit is set on the root group, the whole system can't
	 * exceed 16M for the device.
	 *
365
	 * If not on the default hierarchy, the broken flat hierarchy
366 367 368 369 370
	 * behavior is retained where all throtl_grps are treated as if
	 * they're all separate root groups right below throtl_data.
	 * Limits of a group don't interact with limits of other groups
	 * regardless of the position of the group in the hierarchy.
	 */
371
	sq->parent_sq = &td->service_queue;
372
	if (cgroup_subsys_on_dfl(io_cgrp_subsys) && blkg->parent)
373
		sq->parent_sq = &blkg_to_tg(blkg->parent)->service_queue;
374
	tg->td = td;
375 376
}

377 378 379 380 381 382 383 384 385 386 387 388 389 390 391
/*
 * Set has_rules[] if @tg or any of its parents have limits configured.
 * This doesn't require walking up to the top of the hierarchy as the
 * parent's has_rules[] is guaranteed to be correct.
 */
static void tg_update_has_rules(struct throtl_grp *tg)
{
	struct throtl_grp *parent_tg = sq_to_tg(tg->service_queue.parent_sq);
	int rw;

	for (rw = READ; rw <= WRITE; rw++)
		tg->has_rules[rw] = (parent_tg && parent_tg->has_rules[rw]) ||
				    (tg->bps[rw] != -1 || tg->iops[rw] != -1);
}

392
static void throtl_pd_online(struct blkg_policy_data *pd)
393 394 395 396 397
{
	/*
	 * We don't want new groups to escape the limits of its ancestors.
	 * Update has_rules[] after a new group is brought online.
	 */
398
	tg_update_has_rules(pd_to_tg(pd));
399 400
}

401 402
static void throtl_pd_free(struct blkg_policy_data *pd)
{
403 404
	struct throtl_grp *tg = pd_to_tg(pd);

405
	del_timer_sync(&tg->service_queue.pending_timer);
406
	kfree(tg);
407 408
}

409 410
static struct throtl_grp *
throtl_rb_first(struct throtl_service_queue *parent_sq)
411 412
{
	/* Service tree is empty */
413
	if (!parent_sq->nr_pending)
414 415
		return NULL;

416 417
	if (!parent_sq->first_pending)
		parent_sq->first_pending = rb_first(&parent_sq->pending_tree);
418

419 420
	if (parent_sq->first_pending)
		return rb_entry_tg(parent_sq->first_pending);
421 422 423 424 425 426 427 428 429 430

	return NULL;
}

static void rb_erase_init(struct rb_node *n, struct rb_root *root)
{
	rb_erase(n, root);
	RB_CLEAR_NODE(n);
}

431 432
static void throtl_rb_erase(struct rb_node *n,
			    struct throtl_service_queue *parent_sq)
433
{
434 435 436 437
	if (parent_sq->first_pending == n)
		parent_sq->first_pending = NULL;
	rb_erase_init(n, &parent_sq->pending_tree);
	--parent_sq->nr_pending;
438 439
}

440
static void update_min_dispatch_time(struct throtl_service_queue *parent_sq)
441 442 443
{
	struct throtl_grp *tg;

444
	tg = throtl_rb_first(parent_sq);
445 446 447
	if (!tg)
		return;

448
	parent_sq->first_pending_disptime = tg->disptime;
449 450
}

451
static void tg_service_queue_add(struct throtl_grp *tg)
452
{
453
	struct throtl_service_queue *parent_sq = tg->service_queue.parent_sq;
454
	struct rb_node **node = &parent_sq->pending_tree.rb_node;
455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472
	struct rb_node *parent = NULL;
	struct throtl_grp *__tg;
	unsigned long key = tg->disptime;
	int left = 1;

	while (*node != NULL) {
		parent = *node;
		__tg = rb_entry_tg(parent);

		if (time_before(key, __tg->disptime))
			node = &parent->rb_left;
		else {
			node = &parent->rb_right;
			left = 0;
		}
	}

	if (left)
473
		parent_sq->first_pending = &tg->rb_node;
474 475

	rb_link_node(&tg->rb_node, parent, node);
476
	rb_insert_color(&tg->rb_node, &parent_sq->pending_tree);
477 478
}

479
static void __throtl_enqueue_tg(struct throtl_grp *tg)
480
{
481
	tg_service_queue_add(tg);
482
	tg->flags |= THROTL_TG_PENDING;
483
	tg->service_queue.parent_sq->nr_pending++;
484 485
}

486
static void throtl_enqueue_tg(struct throtl_grp *tg)
487
{
488
	if (!(tg->flags & THROTL_TG_PENDING))
489
		__throtl_enqueue_tg(tg);
490 491
}

492
static void __throtl_dequeue_tg(struct throtl_grp *tg)
493
{
494
	throtl_rb_erase(&tg->rb_node, tg->service_queue.parent_sq);
495
	tg->flags &= ~THROTL_TG_PENDING;
496 497
}

498
static void throtl_dequeue_tg(struct throtl_grp *tg)
499
{
500
	if (tg->flags & THROTL_TG_PENDING)
501
		__throtl_dequeue_tg(tg);
502 503
}

504
/* Call with queue lock held */
505 506
static void throtl_schedule_pending_timer(struct throtl_service_queue *sq,
					  unsigned long expires)
507
{
508 509 510 511 512 513 514 515 516 517 518
	unsigned long max_expire = jiffies + 8 * throtl_slice;

	/*
	 * Since we are adjusting the throttle limit dynamically, the sleep
	 * time calculated according to previous limit might be invalid. It's
	 * possible the cgroup sleep time is very long and no other cgroups
	 * have IO running so notify the limit changes. Make sure the cgroup
	 * doesn't sleep too long to avoid the missed notification.
	 */
	if (time_after(expires, max_expire))
		expires = max_expire;
519 520 521
	mod_timer(&sq->pending_timer, expires);
	throtl_log(sq, "schedule timer. delay=%lu jiffies=%lu",
		   expires - jiffies, jiffies);
522 523
}

524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543
/**
 * throtl_schedule_next_dispatch - schedule the next dispatch cycle
 * @sq: the service_queue to schedule dispatch for
 * @force: force scheduling
 *
 * Arm @sq->pending_timer so that the next dispatch cycle starts on the
 * dispatch time of the first pending child.  Returns %true if either timer
 * is armed or there's no pending child left.  %false if the current
 * dispatch window is still open and the caller should continue
 * dispatching.
 *
 * If @force is %true, the dispatch timer is always scheduled and this
 * function is guaranteed to return %true.  This is to be used when the
 * caller can't dispatch itself and needs to invoke pending_timer
 * unconditionally.  Note that forced scheduling is likely to induce short
 * delay before dispatch starts even if @sq->first_pending_disptime is not
 * in the future and thus shouldn't be used in hot paths.
 */
static bool throtl_schedule_next_dispatch(struct throtl_service_queue *sq,
					  bool force)
544
{
545
	/* any pending children left? */
546
	if (!sq->nr_pending)
547
		return true;
548

549
	update_min_dispatch_time(sq);
550

551
	/* is the next dispatch time in the future? */
552
	if (force || time_after(sq->first_pending_disptime, jiffies)) {
553
		throtl_schedule_pending_timer(sq, sq->first_pending_disptime);
554
		return true;
555 556
	}

557 558
	/* tell the caller to continue dispatching */
	return false;
559 560
}

561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582
static inline void throtl_start_new_slice_with_credit(struct throtl_grp *tg,
		bool rw, unsigned long start)
{
	tg->bytes_disp[rw] = 0;
	tg->io_disp[rw] = 0;

	/*
	 * Previous slice has expired. We must have trimmed it after last
	 * bio dispatch. That means since start of last slice, we never used
	 * that bandwidth. Do try to make use of that bandwidth while giving
	 * credit.
	 */
	if (time_after_eq(start, tg->slice_start[rw]))
		tg->slice_start[rw] = start;

	tg->slice_end[rw] = jiffies + throtl_slice;
	throtl_log(&tg->service_queue,
		   "[%c] new slice with credit start=%lu end=%lu jiffies=%lu",
		   rw == READ ? 'R' : 'W', tg->slice_start[rw],
		   tg->slice_end[rw], jiffies);
}

583
static inline void throtl_start_new_slice(struct throtl_grp *tg, bool rw)
584 585
{
	tg->bytes_disp[rw] = 0;
586
	tg->io_disp[rw] = 0;
587 588
	tg->slice_start[rw] = jiffies;
	tg->slice_end[rw] = jiffies + throtl_slice;
589 590 591 592
	throtl_log(&tg->service_queue,
		   "[%c] new slice start=%lu end=%lu jiffies=%lu",
		   rw == READ ? 'R' : 'W', tg->slice_start[rw],
		   tg->slice_end[rw], jiffies);
593 594
}

595 596
static inline void throtl_set_slice_end(struct throtl_grp *tg, bool rw,
					unsigned long jiffy_end)
597 598 599 600
{
	tg->slice_end[rw] = roundup(jiffy_end, throtl_slice);
}

601 602
static inline void throtl_extend_slice(struct throtl_grp *tg, bool rw,
				       unsigned long jiffy_end)
603 604
{
	tg->slice_end[rw] = roundup(jiffy_end, throtl_slice);
605 606 607 608
	throtl_log(&tg->service_queue,
		   "[%c] extend slice start=%lu end=%lu jiffies=%lu",
		   rw == READ ? 'R' : 'W', tg->slice_start[rw],
		   tg->slice_end[rw], jiffies);
609 610 611
}

/* Determine if previously allocated or extended slice is complete or not */
612
static bool throtl_slice_used(struct throtl_grp *tg, bool rw)
613 614
{
	if (time_in_range(jiffies, tg->slice_start[rw], tg->slice_end[rw]))
615
		return false;
616 617 618 619 620

	return 1;
}

/* Trim the used slices and adjust slice start accordingly */
621
static inline void throtl_trim_slice(struct throtl_grp *tg, bool rw)
622
{
623 624
	unsigned long nr_slices, time_elapsed, io_trim;
	u64 bytes_trim, tmp;
625 626 627 628 629 630 631 632

	BUG_ON(time_before(tg->slice_end[rw], tg->slice_start[rw]));

	/*
	 * If bps are unlimited (-1), then time slice don't get
	 * renewed. Don't try to trim the slice if slice is used. A new
	 * slice will start when appropriate.
	 */
633
	if (throtl_slice_used(tg, rw))
634 635
		return;

636 637 638 639 640 641 642 643
	/*
	 * A bio has been dispatched. Also adjust slice_end. It might happen
	 * that initially cgroup limit was very low resulting in high
	 * slice_end, but later limit was bumped up and bio was dispached
	 * sooner, then we need to reduce slice_end. A high bogus slice_end
	 * is bad because it does not allow new slice to start.
	 */

644
	throtl_set_slice_end(tg, rw, jiffies + throtl_slice);
645

646 647 648 649 650 651
	time_elapsed = jiffies - tg->slice_start[rw];

	nr_slices = time_elapsed / throtl_slice;

	if (!nr_slices)
		return;
652 653 654
	tmp = tg->bps[rw] * throtl_slice * nr_slices;
	do_div(tmp, HZ);
	bytes_trim = tmp;
655

656
	io_trim = (tg->iops[rw] * throtl_slice * nr_slices)/HZ;
657

658
	if (!bytes_trim && !io_trim)
659 660 661 662 663 664 665
		return;

	if (tg->bytes_disp[rw] >= bytes_trim)
		tg->bytes_disp[rw] -= bytes_trim;
	else
		tg->bytes_disp[rw] = 0;

666 667 668 669 670
	if (tg->io_disp[rw] >= io_trim)
		tg->io_disp[rw] -= io_trim;
	else
		tg->io_disp[rw] = 0;

671 672
	tg->slice_start[rw] += nr_slices * throtl_slice;

673 674 675 676
	throtl_log(&tg->service_queue,
		   "[%c] trim slice nr=%lu bytes=%llu io=%lu start=%lu end=%lu jiffies=%lu",
		   rw == READ ? 'R' : 'W', nr_slices, bytes_trim, io_trim,
		   tg->slice_start[rw], tg->slice_end[rw], jiffies);
677 678
}

679 680
static bool tg_with_in_iops_limit(struct throtl_grp *tg, struct bio *bio,
				  unsigned long *wait)
681 682
{
	bool rw = bio_data_dir(bio);
683
	unsigned int io_allowed;
684
	unsigned long jiffy_elapsed, jiffy_wait, jiffy_elapsed_rnd;
685
	u64 tmp;
686

687
	jiffy_elapsed = jiffy_elapsed_rnd = jiffies - tg->slice_start[rw];
688

689 690 691 692 693 694
	/* Slice has just started. Consider one slice interval */
	if (!jiffy_elapsed)
		jiffy_elapsed_rnd = throtl_slice;

	jiffy_elapsed_rnd = roundup(jiffy_elapsed_rnd, throtl_slice);

695 696 697 698 699 700 701 702 703 704 705 706 707 708
	/*
	 * jiffy_elapsed_rnd should not be a big value as minimum iops can be
	 * 1 then at max jiffy elapsed should be equivalent of 1 second as we
	 * will allow dispatch after 1 second and after that slice should
	 * have been trimmed.
	 */

	tmp = (u64)tg->iops[rw] * jiffy_elapsed_rnd;
	do_div(tmp, HZ);

	if (tmp > UINT_MAX)
		io_allowed = UINT_MAX;
	else
		io_allowed = tmp;
709 710

	if (tg->io_disp[rw] + 1 <= io_allowed) {
711 712
		if (wait)
			*wait = 0;
713
		return true;
714 715
	}

716 717 718 719 720 721 722 723 724 725 726 727 728
	/* Calc approx time to dispatch */
	jiffy_wait = ((tg->io_disp[rw] + 1) * HZ)/tg->iops[rw] + 1;

	if (jiffy_wait > jiffy_elapsed)
		jiffy_wait = jiffy_wait - jiffy_elapsed;
	else
		jiffy_wait = 1;

	if (wait)
		*wait = jiffy_wait;
	return 0;
}

729 730
static bool tg_with_in_bps_limit(struct throtl_grp *tg, struct bio *bio,
				 unsigned long *wait)
731 732
{
	bool rw = bio_data_dir(bio);
733
	u64 bytes_allowed, extra_bytes, tmp;
734
	unsigned long jiffy_elapsed, jiffy_wait, jiffy_elapsed_rnd;
735 736 737 738 739 740 741 742 743

	jiffy_elapsed = jiffy_elapsed_rnd = jiffies - tg->slice_start[rw];

	/* Slice has just started. Consider one slice interval */
	if (!jiffy_elapsed)
		jiffy_elapsed_rnd = throtl_slice;

	jiffy_elapsed_rnd = roundup(jiffy_elapsed_rnd, throtl_slice);

744 745
	tmp = tg->bps[rw] * jiffy_elapsed_rnd;
	do_div(tmp, HZ);
746
	bytes_allowed = tmp;
747

748
	if (tg->bytes_disp[rw] + bio->bi_iter.bi_size <= bytes_allowed) {
749 750
		if (wait)
			*wait = 0;
751
		return true;
752 753 754
	}

	/* Calc approx time to dispatch */
755
	extra_bytes = tg->bytes_disp[rw] + bio->bi_iter.bi_size - bytes_allowed;
756 757 758 759 760 761 762 763 764 765 766 767
	jiffy_wait = div64_u64(extra_bytes * HZ, tg->bps[rw]);

	if (!jiffy_wait)
		jiffy_wait = 1;

	/*
	 * This wait time is without taking into consideration the rounding
	 * up we did. Add that time also.
	 */
	jiffy_wait = jiffy_wait + (jiffy_elapsed_rnd - jiffy_elapsed);
	if (wait)
		*wait = jiffy_wait;
768 769 770 771 772 773 774
	return 0;
}

/*
 * Returns whether one can dispatch a bio or not. Also returns approx number
 * of jiffies to wait before this bio is with-in IO rate and can be dispatched
 */
775 776
static bool tg_may_dispatch(struct throtl_grp *tg, struct bio *bio,
			    unsigned long *wait)
777 778 779 780 781 782 783 784 785 786
{
	bool rw = bio_data_dir(bio);
	unsigned long bps_wait = 0, iops_wait = 0, max_wait = 0;

	/*
 	 * Currently whole state machine of group depends on first bio
	 * queued in the group bio list. So one should not be calling
	 * this function with a different bio if there are other bios
	 * queued.
	 */
787
	BUG_ON(tg->service_queue.nr_queued[rw] &&
788
	       bio != throtl_peek_queued(&tg->service_queue.queued[rw]));
789

790 791 792 793
	/* If tg->bps = -1, then BW is unlimited */
	if (tg->bps[rw] == -1 && tg->iops[rw] == -1) {
		if (wait)
			*wait = 0;
794
		return true;
795 796 797 798 799 800 801
	}

	/*
	 * If previous slice expired, start a new one otherwise renew/extend
	 * existing slice to make sure it is at least throtl_slice interval
	 * long since now.
	 */
802 803
	if (throtl_slice_used(tg, rw))
		throtl_start_new_slice(tg, rw);
804 805
	else {
		if (time_before(tg->slice_end[rw], jiffies + throtl_slice))
806
			throtl_extend_slice(tg, rw, jiffies + throtl_slice);
807 808
	}

809 810
	if (tg_with_in_bps_limit(tg, bio, &bps_wait) &&
	    tg_with_in_iops_limit(tg, bio, &iops_wait)) {
811 812 813 814 815 816 817 818 819 820 821
		if (wait)
			*wait = 0;
		return 1;
	}

	max_wait = max(bps_wait, iops_wait);

	if (wait)
		*wait = max_wait;

	if (time_before(tg->slice_end[rw], jiffies + max_wait))
822
		throtl_extend_slice(tg, rw, jiffies + max_wait);
823 824 825 826 827 828 829 830 831

	return 0;
}

static void throtl_charge_bio(struct throtl_grp *tg, struct bio *bio)
{
	bool rw = bio_data_dir(bio);

	/* Charge the bio to the group */
832
	tg->bytes_disp[rw] += bio->bi_iter.bi_size;
833
	tg->io_disp[rw]++;
834

835 836 837 838 839 840
	/*
	 * REQ_THROTTLED is used to prevent the same bio to be throttled
	 * more than once as a throttled bio will go through blk-throtl the
	 * second time when it eventually gets issued.  Set it when a bio
	 * is being charged to a tg.
	 */
841
	if (!(bio->bi_rw & REQ_THROTTLED))
842
		bio->bi_rw |= REQ_THROTTLED;
843 844
}

845 846 847 848 849 850 851 852 853 854 855
/**
 * throtl_add_bio_tg - add a bio to the specified throtl_grp
 * @bio: bio to add
 * @qn: qnode to use
 * @tg: the target throtl_grp
 *
 * Add @bio to @tg's service_queue using @qn.  If @qn is not specified,
 * tg->qnode_on_self[] is used.
 */
static void throtl_add_bio_tg(struct bio *bio, struct throtl_qnode *qn,
			      struct throtl_grp *tg)
856
{
857
	struct throtl_service_queue *sq = &tg->service_queue;
858 859
	bool rw = bio_data_dir(bio);

860 861 862
	if (!qn)
		qn = &tg->qnode_on_self[rw];

863 864 865 866 867 868 869 870 871
	/*
	 * If @tg doesn't currently have any bios queued in the same
	 * direction, queueing @bio can change when @tg should be
	 * dispatched.  Mark that @tg was empty.  This is automatically
	 * cleaered on the next tg_update_disptime().
	 */
	if (!sq->nr_queued[rw])
		tg->flags |= THROTL_TG_WAS_EMPTY;

872 873
	throtl_qnode_add_bio(bio, qn, &sq->queued[rw]);

874
	sq->nr_queued[rw]++;
875
	throtl_enqueue_tg(tg);
876 877
}

878
static void tg_update_disptime(struct throtl_grp *tg)
879
{
880
	struct throtl_service_queue *sq = &tg->service_queue;
881 882 883
	unsigned long read_wait = -1, write_wait = -1, min_wait = -1, disptime;
	struct bio *bio;

884
	if ((bio = throtl_peek_queued(&sq->queued[READ])))
885
		tg_may_dispatch(tg, bio, &read_wait);
886

887
	if ((bio = throtl_peek_queued(&sq->queued[WRITE])))
888
		tg_may_dispatch(tg, bio, &write_wait);
889 890 891 892 893

	min_wait = min(read_wait, write_wait);
	disptime = jiffies + min_wait;

	/* Update dispatch time */
894
	throtl_dequeue_tg(tg);
895
	tg->disptime = disptime;
896
	throtl_enqueue_tg(tg);
897 898 899

	/* see throtl_add_bio_tg() */
	tg->flags &= ~THROTL_TG_WAS_EMPTY;
900 901
}

902 903 904 905 906 907 908 909 910 911
static void start_parent_slice_with_credit(struct throtl_grp *child_tg,
					struct throtl_grp *parent_tg, bool rw)
{
	if (throtl_slice_used(parent_tg, rw)) {
		throtl_start_new_slice_with_credit(parent_tg, rw,
				child_tg->slice_start[rw]);
	}

}

912
static void tg_dispatch_one_bio(struct throtl_grp *tg, bool rw)
913
{
914
	struct throtl_service_queue *sq = &tg->service_queue;
915 916
	struct throtl_service_queue *parent_sq = sq->parent_sq;
	struct throtl_grp *parent_tg = sq_to_tg(parent_sq);
917
	struct throtl_grp *tg_to_put = NULL;
918 919
	struct bio *bio;

920 921 922 923 924 925 926
	/*
	 * @bio is being transferred from @tg to @parent_sq.  Popping a bio
	 * from @tg may put its reference and @parent_sq might end up
	 * getting released prematurely.  Remember the tg to put and put it
	 * after @bio is transferred to @parent_sq.
	 */
	bio = throtl_pop_queued(&sq->queued[rw], &tg_to_put);
927
	sq->nr_queued[rw]--;
928 929

	throtl_charge_bio(tg, bio);
930 931 932 933 934 935 936 937 938

	/*
	 * If our parent is another tg, we just need to transfer @bio to
	 * the parent using throtl_add_bio_tg().  If our parent is
	 * @td->service_queue, @bio is ready to be issued.  Put it on its
	 * bio_lists[] and decrease total number queued.  The caller is
	 * responsible for issuing these bios.
	 */
	if (parent_tg) {
939
		throtl_add_bio_tg(bio, &tg->qnode_on_parent[rw], parent_tg);
940
		start_parent_slice_with_credit(tg, parent_tg, rw);
941
	} else {
942 943
		throtl_qnode_add_bio(bio, &tg->qnode_on_parent[rw],
				     &parent_sq->queued[rw]);
944 945 946
		BUG_ON(tg->td->nr_queued[rw] <= 0);
		tg->td->nr_queued[rw]--;
	}
947

948
	throtl_trim_slice(tg, rw);
949

950 951
	if (tg_to_put)
		blkg_put(tg_to_blkg(tg_to_put));
952 953
}

954
static int throtl_dispatch_tg(struct throtl_grp *tg)
955
{
956
	struct throtl_service_queue *sq = &tg->service_queue;
957 958
	unsigned int nr_reads = 0, nr_writes = 0;
	unsigned int max_nr_reads = throtl_grp_quantum*3/4;
959
	unsigned int max_nr_writes = throtl_grp_quantum - max_nr_reads;
960 961 962 963
	struct bio *bio;

	/* Try to dispatch 75% READS and 25% WRITES */

964
	while ((bio = throtl_peek_queued(&sq->queued[READ])) &&
965
	       tg_may_dispatch(tg, bio, NULL)) {
966

967
		tg_dispatch_one_bio(tg, bio_data_dir(bio));
968 969 970 971 972 973
		nr_reads++;

		if (nr_reads >= max_nr_reads)
			break;
	}

974
	while ((bio = throtl_peek_queued(&sq->queued[WRITE])) &&
975
	       tg_may_dispatch(tg, bio, NULL)) {
976

977
		tg_dispatch_one_bio(tg, bio_data_dir(bio));
978 979 980 981 982 983 984 985 986
		nr_writes++;

		if (nr_writes >= max_nr_writes)
			break;
	}

	return nr_reads + nr_writes;
}

987
static int throtl_select_dispatch(struct throtl_service_queue *parent_sq)
988 989 990 991
{
	unsigned int nr_disp = 0;

	while (1) {
992 993
		struct throtl_grp *tg = throtl_rb_first(parent_sq);
		struct throtl_service_queue *sq = &tg->service_queue;
994 995 996 997 998 999 1000

		if (!tg)
			break;

		if (time_before(jiffies, tg->disptime))
			break;

1001
		throtl_dequeue_tg(tg);
1002

1003
		nr_disp += throtl_dispatch_tg(tg);
1004

1005
		if (sq->nr_queued[0] || sq->nr_queued[1])
1006
			tg_update_disptime(tg);
1007 1008 1009 1010 1011 1012 1013 1014

		if (nr_disp >= throtl_quantum)
			break;
	}

	return nr_disp;
}

1015 1016 1017 1018 1019 1020 1021
/**
 * throtl_pending_timer_fn - timer function for service_queue->pending_timer
 * @arg: the throtl_service_queue being serviced
 *
 * This timer is armed when a child throtl_grp with active bio's become
 * pending and queued on the service_queue's pending_tree and expires when
 * the first child throtl_grp should be dispatched.  This function
1022 1023 1024 1025 1026 1027 1028
 * dispatches bio's from the children throtl_grps to the parent
 * service_queue.
 *
 * If the parent's parent is another throtl_grp, dispatching is propagated
 * by either arming its pending_timer or repeating dispatch directly.  If
 * the top-level service_tree is reached, throtl_data->dispatch_work is
 * kicked so that the ready bio's are issued.
1029
 */
1030 1031 1032
static void throtl_pending_timer_fn(unsigned long arg)
{
	struct throtl_service_queue *sq = (void *)arg;
1033
	struct throtl_grp *tg = sq_to_tg(sq);
1034
	struct throtl_data *td = sq_to_td(sq);
1035
	struct request_queue *q = td->queue;
1036 1037
	struct throtl_service_queue *parent_sq;
	bool dispatched;
1038
	int ret;
1039 1040

	spin_lock_irq(q->queue_lock);
1041 1042 1043
again:
	parent_sq = sq->parent_sq;
	dispatched = false;
1044

1045 1046
	while (true) {
		throtl_log(sq, "dispatch nr_queued=%u read=%u write=%u",
1047 1048
			   sq->nr_queued[READ] + sq->nr_queued[WRITE],
			   sq->nr_queued[READ], sq->nr_queued[WRITE]);
1049 1050 1051 1052 1053 1054

		ret = throtl_select_dispatch(sq);
		if (ret) {
			throtl_log(sq, "bios disp=%u", ret);
			dispatched = true;
		}
1055

1056 1057
		if (throtl_schedule_next_dispatch(sq, false))
			break;
1058

1059 1060 1061 1062
		/* this dispatch windows is still open, relax and repeat */
		spin_unlock_irq(q->queue_lock);
		cpu_relax();
		spin_lock_irq(q->queue_lock);
1063
	}
1064

1065 1066
	if (!dispatched)
		goto out_unlock;
1067

1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083
	if (parent_sq) {
		/* @parent_sq is another throl_grp, propagate dispatch */
		if (tg->flags & THROTL_TG_WAS_EMPTY) {
			tg_update_disptime(tg);
			if (!throtl_schedule_next_dispatch(parent_sq, false)) {
				/* window is already open, repeat dispatching */
				sq = parent_sq;
				tg = sq_to_tg(sq);
				goto again;
			}
		}
	} else {
		/* reached the top-level, queue issueing */
		queue_work(kthrotld_workqueue, &td->dispatch_work);
	}
out_unlock:
1084
	spin_unlock_irq(q->queue_lock);
1085
}
1086

1087 1088 1089 1090 1091 1092 1093 1094
/**
 * blk_throtl_dispatch_work_fn - work function for throtl_data->dispatch_work
 * @work: work item being executed
 *
 * This function is queued for execution when bio's reach the bio_lists[]
 * of throtl_data->service_queue.  Those bio's are ready and issued by this
 * function.
 */
1095
static void blk_throtl_dispatch_work_fn(struct work_struct *work)
1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108
{
	struct throtl_data *td = container_of(work, struct throtl_data,
					      dispatch_work);
	struct throtl_service_queue *td_sq = &td->service_queue;
	struct request_queue *q = td->queue;
	struct bio_list bio_list_on_stack;
	struct bio *bio;
	struct blk_plug plug;
	int rw;

	bio_list_init(&bio_list_on_stack);

	spin_lock_irq(q->queue_lock);
1109 1110 1111
	for (rw = READ; rw <= WRITE; rw++)
		while ((bio = throtl_pop_queued(&td_sq->queued[rw], NULL)))
			bio_list_add(&bio_list_on_stack, bio);
1112 1113 1114
	spin_unlock_irq(q->queue_lock);

	if (!bio_list_empty(&bio_list_on_stack)) {
1115
		blk_start_plug(&plug);
1116 1117
		while((bio = bio_list_pop(&bio_list_on_stack)))
			generic_make_request(bio);
1118
		blk_finish_plug(&plug);
1119 1120 1121
	}
}

1122 1123
static u64 tg_prfill_conf_u64(struct seq_file *sf, struct blkg_policy_data *pd,
			      int off)
1124
{
1125 1126
	struct throtl_grp *tg = pd_to_tg(pd);
	u64 v = *(u64 *)((void *)tg + off);
1127

1128
	if (v == -1)
1129
		return 0;
1130
	return __blkg_prfill_u64(sf, pd, v);
1131 1132
}

1133 1134
static u64 tg_prfill_conf_uint(struct seq_file *sf, struct blkg_policy_data *pd,
			       int off)
1135
{
1136 1137
	struct throtl_grp *tg = pd_to_tg(pd);
	unsigned int v = *(unsigned int *)((void *)tg + off);
1138

1139 1140
	if (v == -1)
		return 0;
1141
	return __blkg_prfill_u64(sf, pd, v);
1142 1143
}

1144
static int tg_print_conf_u64(struct seq_file *sf, void *v)
1145
{
1146 1147
	blkcg_print_blkgs(sf, css_to_blkcg(seq_css(sf)), tg_prfill_conf_u64,
			  &blkcg_policy_throtl, seq_cft(sf)->private, false);
1148
	return 0;
1149 1150
}

1151
static int tg_print_conf_uint(struct seq_file *sf, void *v)
1152
{
1153 1154
	blkcg_print_blkgs(sf, css_to_blkcg(seq_css(sf)), tg_prfill_conf_uint,
			  &blkcg_policy_throtl, seq_cft(sf)->private, false);
1155
	return 0;
1156 1157
}

1158
static void tg_conf_updated(struct throtl_grp *tg)
1159
{
1160
	struct throtl_service_queue *sq = &tg->service_queue;
1161
	struct cgroup_subsys_state *pos_css;
1162
	struct blkcg_gq *blkg;
1163

1164 1165 1166 1167
	throtl_log(&tg->service_queue,
		   "limit change rbps=%llu wbps=%llu riops=%u wiops=%u",
		   tg->bps[READ], tg->bps[WRITE],
		   tg->iops[READ], tg->iops[WRITE]);
1168

1169 1170 1171 1172 1173 1174 1175
	/*
	 * Update has_rules[] flags for the updated tg's subtree.  A tg is
	 * considered to have rules if either the tg itself or any of its
	 * ancestors has rules.  This identifies groups without any
	 * restrictions in the whole hierarchy and allows them to bypass
	 * blk-throttle.
	 */
1176
	blkg_for_each_descendant_pre(blkg, pos_css, tg_to_blkg(tg))
1177 1178
		tg_update_has_rules(blkg_to_tg(blkg));

1179 1180 1181 1182 1183 1184 1185 1186
	/*
	 * We're already holding queue_lock and know @tg is valid.  Let's
	 * apply the new config directly.
	 *
	 * Restart the slices for both READ and WRITES. It might happen
	 * that a group's limit are dropped suddenly and we don't want to
	 * account recently dispatched IO with new low rate.
	 */
1187 1188
	throtl_start_new_slice(tg, 0);
	throtl_start_new_slice(tg, 1);
1189

1190
	if (tg->flags & THROTL_TG_PENDING) {
1191
		tg_update_disptime(tg);
1192
		throtl_schedule_next_dispatch(sq->parent_sq, true);
1193
	}
1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220
}

static ssize_t tg_set_conf(struct kernfs_open_file *of,
			   char *buf, size_t nbytes, loff_t off, bool is_u64)
{
	struct blkcg *blkcg = css_to_blkcg(of_css(of));
	struct blkg_conf_ctx ctx;
	struct throtl_grp *tg;
	int ret;
	u64 v;

	ret = blkg_conf_prep(blkcg, &blkcg_policy_throtl, buf, &ctx);
	if (ret)
		return ret;

	ret = -EINVAL;
	if (sscanf(ctx.body, "%llu", &v) != 1)
		goto out_finish;
	if (!v)
		v = -1;

	tg = blkg_to_tg(ctx.blkg);

	if (is_u64)
		*(u64 *)((void *)tg + of_cft(of)->private) = v;
	else
		*(unsigned int *)((void *)tg + of_cft(of)->private) = v;
1221

1222
	tg_conf_updated(tg);
1223 1224
	ret = 0;
out_finish:
1225
	blkg_conf_finish(&ctx);
1226
	return ret ?: nbytes;
1227 1228
}

1229 1230
static ssize_t tg_set_conf_u64(struct kernfs_open_file *of,
			       char *buf, size_t nbytes, loff_t off)
1231
{
1232
	return tg_set_conf(of, buf, nbytes, off, true);
1233 1234
}

1235 1236
static ssize_t tg_set_conf_uint(struct kernfs_open_file *of,
				char *buf, size_t nbytes, loff_t off)
1237
{
1238
	return tg_set_conf(of, buf, nbytes, off, false);
1239 1240
}

1241
static struct cftype throtl_legacy_files[] = {
1242 1243
	{
		.name = "throttle.read_bps_device",
1244
		.private = offsetof(struct throtl_grp, bps[READ]),
1245
		.seq_show = tg_print_conf_u64,
1246
		.write = tg_set_conf_u64,
1247 1248 1249
	},
	{
		.name = "throttle.write_bps_device",
1250
		.private = offsetof(struct throtl_grp, bps[WRITE]),
1251
		.seq_show = tg_print_conf_u64,
1252
		.write = tg_set_conf_u64,
1253 1254 1255
	},
	{
		.name = "throttle.read_iops_device",