blk-cgroup.c 46.0 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11
/*
 * Common Block IO controller cgroup interface
 *
 * Based on ideas and code from CFQ, CFS and BFQ:
 * Copyright (C) 2003 Jens Axboe <axboe@kernel.dk>
 *
 * Copyright (C) 2008 Fabio Checconi <fabio@gandalf.sssup.it>
 *		      Paolo Valente <paolo.valente@unimore.it>
 *
 * Copyright (C) 2009 Vivek Goyal <vgoyal@redhat.com>
 * 	              Nauman Rafique <nauman@google.com>
12 13 14 15
 *
 * For policy-specific per-blkcg data:
 * Copyright (C) 2015 Paolo Valente <paolo.valente@unimore.it>
 *                    Arianna Avanzini <avanzini.arianna@gmail.com>
16 17
 */
#include <linux/ioprio.h>
18
#include <linux/kdev_t.h>
19
#include <linux/module.h>
20
#include <linux/sched/signal.h>
21
#include <linux/err.h>
22
#include <linux/blkdev.h>
23
#include <linux/backing-dev.h>
24
#include <linux/slab.h>
25
#include <linux/genhd.h>
26
#include <linux/delay.h>
T
Tejun Heo 已提交
27
#include <linux/atomic.h>
28
#include <linux/ctype.h>
29
#include <linux/blk-cgroup.h>
30
#include <linux/tracehook.h>
31
#include "blk.h"
32

33 34
#define MAX_KEY_LEN 100

35 36 37 38 39 40 41 42
/*
 * blkcg_pol_mutex protects blkcg_policy[] and policy [de]activation.
 * blkcg_pol_register_mutex nests outside of it and synchronizes entire
 * policy [un]register operations including cgroup file additions /
 * removals.  Putting cgroup file registration outside blkcg_pol_mutex
 * allows grabbing it from cgroup callbacks.
 */
static DEFINE_MUTEX(blkcg_pol_register_mutex);
43
static DEFINE_MUTEX(blkcg_pol_mutex);
44

45
struct blkcg blkcg_root;
T
Tejun Heo 已提交
46
EXPORT_SYMBOL_GPL(blkcg_root);
47

T
Tejun Heo 已提交
48 49
struct cgroup_subsys_state * const blkcg_root_css = &blkcg_root.css;

T
Tejun Heo 已提交
50
static struct blkcg_policy *blkcg_policy[BLKCG_MAX_POLS];
51

T
Tejun Heo 已提交
52 53
static LIST_HEAD(all_blkcgs);		/* protected by blkcg_pol_mutex */

54 55
static bool blkcg_debug_stats = false;

56
static bool blkcg_policy_enabled(struct request_queue *q,
T
Tejun Heo 已提交
57
				 const struct blkcg_policy *pol)
58 59 60 61
{
	return pol && test_bit(pol->plid, q->blkcg_pols);
}

62 63 64 65 66 67
/**
 * blkg_free - free a blkg
 * @blkg: blkg to free
 *
 * Free @blkg which may be partially allocated.
 */
T
Tejun Heo 已提交
68
static void blkg_free(struct blkcg_gq *blkg)
69
{
70
	int i;
71 72 73 74

	if (!blkg)
		return;

75
	for (i = 0; i < BLKCG_MAX_POLS; i++)
76 77
		if (blkg->pd[i])
			blkcg_policy[i]->pd_free_fn(blkg->pd[i]);
78

79
	if (blkg->blkcg != &blkcg_root)
80
		blk_exit_rl(blkg->q, &blkg->rl);
81 82 83

	blkg_rwstat_exit(&blkg->stat_ios);
	blkg_rwstat_exit(&blkg->stat_bytes);
84
	kfree(blkg);
85 86 87 88 89 90
}

/**
 * blkg_alloc - allocate a blkg
 * @blkcg: block cgroup the new blkg is associated with
 * @q: request_queue the new blkg is associated with
91
 * @gfp_mask: allocation mask to use
92
 *
93
 * Allocate a new blkg assocating @blkcg and @q.
94
 */
95 96
static struct blkcg_gq *blkg_alloc(struct blkcg *blkcg, struct request_queue *q,
				   gfp_t gfp_mask)
97
{
T
Tejun Heo 已提交
98
	struct blkcg_gq *blkg;
99
	int i;
100 101

	/* alloc and init base part */
102
	blkg = kzalloc_node(sizeof(*blkg), gfp_mask, q->node);
103 104 105
	if (!blkg)
		return NULL;

106 107 108 109
	if (blkg_rwstat_init(&blkg->stat_bytes, gfp_mask) ||
	    blkg_rwstat_init(&blkg->stat_ios, gfp_mask))
		goto err_free;

T
Tejun Heo 已提交
110
	blkg->q = q;
111
	INIT_LIST_HEAD(&blkg->q_node);
112
	blkg->blkcg = blkcg;
113
	atomic_set(&blkg->refcnt, 1);
114

115 116 117 118 119 120 121
	/* root blkg uses @q->root_rl, init rl only for !root blkgs */
	if (blkcg != &blkcg_root) {
		if (blk_init_rl(&blkg->rl, q, gfp_mask))
			goto err_free;
		blkg->rl.blkg = blkg;
	}

T
Tejun Heo 已提交
122
	for (i = 0; i < BLKCG_MAX_POLS; i++) {
T
Tejun Heo 已提交
123
		struct blkcg_policy *pol = blkcg_policy[i];
124
		struct blkg_policy_data *pd;
125

126
		if (!blkcg_policy_enabled(q, pol))
127 128 129
			continue;

		/* alloc per-policy data and attach it to blkg */
130
		pd = pol->pd_alloc_fn(gfp_mask, q->node);
131 132
		if (!pd)
			goto err_free;
133

134 135
		blkg->pd[i] = pd;
		pd->blkg = blkg;
T
Tejun Heo 已提交
136
		pd->plid = i;
137 138
	}

139
	return blkg;
140 141 142 143

err_free:
	blkg_free(blkg);
	return NULL;
144 145
}

T
Tejun Heo 已提交
146 147
struct blkcg_gq *blkg_lookup_slowpath(struct blkcg *blkcg,
				      struct request_queue *q, bool update_hint)
148
{
T
Tejun Heo 已提交
149
	struct blkcg_gq *blkg;
150

151
	/*
152 153 154 155
	 * Hint didn't match.  Look up from the radix tree.  Note that the
	 * hint can only be updated under queue_lock as otherwise @blkg
	 * could have already been removed from blkg_tree.  The caller is
	 * responsible for grabbing queue_lock if @update_hint.
156 157
	 */
	blkg = radix_tree_lookup(&blkcg->blkg_tree, q->id);
158 159 160 161 162
	if (blkg && blkg->q == q) {
		if (update_hint) {
			lockdep_assert_held(q->queue_lock);
			rcu_assign_pointer(blkcg->blkg_hint, blkg);
		}
163
		return blkg;
164
	}
165

166 167
	return NULL;
}
168
EXPORT_SYMBOL_GPL(blkg_lookup_slowpath);
169

170
/*
171 172
 * If @new_blkg is %NULL, this function tries to allocate a new one as
 * necessary using %GFP_NOWAIT.  @new_blkg is always consumed on return.
173
 */
174
static struct blkcg_gq *blkg_create(struct blkcg *blkcg,
175 176
				    struct request_queue *q,
				    struct blkcg_gq *new_blkg)
177
{
178
	struct blkcg_gq *blkg;
179
	struct bdi_writeback_congested *wb_congested;
180
	int i, ret;
181

182 183 184
	WARN_ON_ONCE(!rcu_read_lock_held());
	lockdep_assert_held(q->queue_lock);

185
	/* blkg holds a reference to blkcg */
186
	if (!css_tryget_online(&blkcg->css)) {
187
		ret = -ENODEV;
188
		goto err_free_blkg;
189
	}
190

191
	wb_congested = wb_congested_get_create(q->backing_dev_info,
192 193 194
					       blkcg->css.id,
					       GFP_NOWAIT | __GFP_NOWARN);
	if (!wb_congested) {
195
		ret = -ENOMEM;
196
		goto err_put_css;
197 198
	}

199 200 201 202 203 204
	/* allocate */
	if (!new_blkg) {
		new_blkg = blkg_alloc(blkcg, q, GFP_NOWAIT | __GFP_NOWARN);
		if (unlikely(!new_blkg)) {
			ret = -ENOMEM;
			goto err_put_congested;
205 206
		}
	}
207 208
	blkg = new_blkg;
	blkg->wb_congested = wb_congested;
209

210
	/* link parent */
T
Tejun Heo 已提交
211 212 213
	if (blkcg_parent(blkcg)) {
		blkg->parent = __blkg_lookup(blkcg_parent(blkcg), q, false);
		if (WARN_ON_ONCE(!blkg->parent)) {
214
			ret = -ENODEV;
215
			goto err_put_congested;
T
Tejun Heo 已提交
216 217 218 219
		}
		blkg_get(blkg->parent);
	}

220 221 222 223 224
	/* invoke per-policy init */
	for (i = 0; i < BLKCG_MAX_POLS; i++) {
		struct blkcg_policy *pol = blkcg_policy[i];

		if (blkg->pd[i] && pol->pd_init_fn)
225
			pol->pd_init_fn(blkg->pd[i]);
226 227 228
	}

	/* insert */
229
	spin_lock(&blkcg->lock);
230 231 232 233
	ret = radix_tree_insert(&blkcg->blkg_tree, q->id, blkg);
	if (likely(!ret)) {
		hlist_add_head_rcu(&blkg->blkcg_node, &blkcg->blkg_list);
		list_add(&blkg->q_node, &q->blkg_list);
234 235 236 237 238

		for (i = 0; i < BLKCG_MAX_POLS; i++) {
			struct blkcg_policy *pol = blkcg_policy[i];

			if (blkg->pd[i] && pol->pd_online_fn)
239
				pol->pd_online_fn(blkg->pd[i]);
240
		}
241
	}
242
	blkg->online = true;
243
	spin_unlock(&blkcg->lock);
244

245
	if (!ret)
246
		return blkg;
247

T
Tejun Heo 已提交
248 249 250 251
	/* @blkg failed fully initialized, use the usual release path */
	blkg_put(blkg);
	return ERR_PTR(ret);

252 253 254
err_put_congested:
	wb_congested_put(wb_congested);
err_put_css:
255
	css_put(&blkcg->css);
256
err_free_blkg:
257
	blkg_free(new_blkg);
258
	return ERR_PTR(ret);
259
}
260

261
/**
262
 * blkg_lookup_create - lookup blkg, try to create one if not there
263 264 265 266
 * @blkcg: blkcg of interest
 * @q: request_queue of interest
 *
 * Lookup blkg for the @blkcg - @q pair.  If it doesn't exist, try to
T
Tejun Heo 已提交
267 268 269
 * create one.  blkg creation is performed recursively from blkcg_root such
 * that all non-root blkg's have access to the parent blkg.  This function
 * should be called under RCU read lock and @q->queue_lock.
270 271 272 273 274
 *
 * Returns pointer to the looked up or created blkg on success, ERR_PTR()
 * value on error.  If @q is dead, returns ERR_PTR(-EINVAL).  If @q is not
 * dead and bypassing, returns ERR_PTR(-EBUSY).
 */
275 276
struct blkcg_gq *blkg_lookup_create(struct blkcg *blkcg,
				    struct request_queue *q)
277
{
278 279 280 281 282
	struct blkcg_gq *blkg;

	WARN_ON_ONCE(!rcu_read_lock_held());
	lockdep_assert_held(q->queue_lock);

283 284 285 286 287 288 289
	/*
	 * This could be the first entry point of blkcg implementation and
	 * we shouldn't allow anything to go through for a bypassing queue.
	 */
	if (unlikely(blk_queue_bypass(q)))
		return ERR_PTR(blk_queue_dying(q) ? -ENODEV : -EBUSY);

290 291 292 293
	blkg = __blkg_lookup(blkcg, q, true);
	if (blkg)
		return blkg;

T
Tejun Heo 已提交
294 295 296 297 298 299 300 301 302 303 304 305 306
	/*
	 * Create blkgs walking down from blkcg_root to @blkcg, so that all
	 * non-root blkgs have access to their parents.
	 */
	while (true) {
		struct blkcg *pos = blkcg;
		struct blkcg *parent = blkcg_parent(blkcg);

		while (parent && !__blkg_lookup(parent, q, false)) {
			pos = parent;
			parent = blkcg_parent(parent);
		}

307
		blkg = blkg_create(pos, q, NULL);
T
Tejun Heo 已提交
308 309 310
		if (pos == blkcg || IS_ERR(blkg))
			return blkg;
	}
311
}
312

T
Tejun Heo 已提交
313
static void blkg_destroy(struct blkcg_gq *blkg)
314
{
T
Tejun Heo 已提交
315
	struct blkcg *blkcg = blkg->blkcg;
316
	struct blkcg_gq *parent = blkg->parent;
317
	int i;
318

319
	lockdep_assert_held(blkg->q->queue_lock);
320
	lockdep_assert_held(&blkcg->lock);
321 322

	/* Something wrong if we are trying to remove same group twice */
323
	WARN_ON_ONCE(list_empty(&blkg->q_node));
324
	WARN_ON_ONCE(hlist_unhashed(&blkg->blkcg_node));
325

326 327 328 329 330 331 332
	for (i = 0; i < BLKCG_MAX_POLS; i++) {
		struct blkcg_policy *pol = blkcg_policy[i];

		if (blkg->pd[i] && pol->pd_offline_fn)
			pol->pd_offline_fn(blkg->pd[i]);
	}

333 334 335 336 337
	if (parent) {
		blkg_rwstat_add_aux(&parent->stat_bytes, &blkg->stat_bytes);
		blkg_rwstat_add_aux(&parent->stat_ios, &blkg->stat_ios);
	}

338 339
	blkg->online = false;

340
	radix_tree_delete(&blkcg->blkg_tree, blkg->q->id);
341
	list_del_init(&blkg->q_node);
342
	hlist_del_init_rcu(&blkg->blkcg_node);
343

344 345 346 347 348
	/*
	 * Both setting lookup hint to and clearing it from @blkg are done
	 * under queue_lock.  If it's not pointing to @blkg now, it never
	 * will.  Hint assignment itself can race safely.
	 */
349
	if (rcu_access_pointer(blkcg->blkg_hint) == blkg)
350 351
		rcu_assign_pointer(blkcg->blkg_hint, NULL);

352 353 354 355 356 357 358
	/*
	 * Put the reference taken at the time of creation so that when all
	 * queues are gone, group can be destroyed.
	 */
	blkg_put(blkg);
}

359 360 361 362
/**
 * blkg_destroy_all - destroy all blkgs associated with a request_queue
 * @q: request_queue of interest
 *
363
 * Destroy all blkgs associated with @q.
364
 */
365
static void blkg_destroy_all(struct request_queue *q)
366
{
T
Tejun Heo 已提交
367
	struct blkcg_gq *blkg, *n;
368

369
	lockdep_assert_held(q->queue_lock);
370

371
	list_for_each_entry_safe(blkg, n, &q->blkg_list, q_node) {
T
Tejun Heo 已提交
372
		struct blkcg *blkcg = blkg->blkcg;
373

374 375 376
		spin_lock(&blkcg->lock);
		blkg_destroy(blkg);
		spin_unlock(&blkcg->lock);
377
	}
378 379 380

	q->root_blkg = NULL;
	q->root_rl.blkg = NULL;
381 382
}

383 384 385 386 387 388 389 390 391
/*
 * A group is RCU protected, but having an rcu lock does not mean that one
 * can access all the fields of blkg and assume these are valid.  For
 * example, don't try to follow throtl_data and request queue links.
 *
 * Having a reference to blkg under an rcu allows accesses to only values
 * local to groups like group stats and group rate limits.
 */
void __blkg_release_rcu(struct rcu_head *rcu_head)
T
Tejun Heo 已提交
392
{
393
	struct blkcg_gq *blkg = container_of(rcu_head, struct blkcg_gq, rcu_head);
394

T
Tejun Heo 已提交
395
	/* release the blkcg and parent blkg refs this blkg has been holding */
T
Tejun Heo 已提交
396
	css_put(&blkg->blkcg->css);
397
	if (blkg->parent)
T
Tejun Heo 已提交
398
		blkg_put(blkg->parent);
T
Tejun Heo 已提交
399

400 401
	wb_congested_put(blkg->wb_congested);

402
	blkg_free(blkg);
T
Tejun Heo 已提交
403
}
404
EXPORT_SYMBOL_GPL(__blkg_release_rcu);
T
Tejun Heo 已提交
405

406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421
/*
 * The next function used by blk_queue_for_each_rl().  It's a bit tricky
 * because the root blkg uses @q->root_rl instead of its own rl.
 */
struct request_list *__blk_queue_next_rl(struct request_list *rl,
					 struct request_queue *q)
{
	struct list_head *ent;
	struct blkcg_gq *blkg;

	/*
	 * Determine the current blkg list_head.  The first entry is
	 * root_rl which is off @q->blkg_list and mapped to the head.
	 */
	if (rl == &q->root_rl) {
		ent = &q->blkg_list;
422 423 424
		/* There are no more block groups, hence no request lists */
		if (list_empty(ent))
			return NULL;
425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440
	} else {
		blkg = container_of(rl, struct blkcg_gq, rl);
		ent = &blkg->q_node;
	}

	/* walk to the next list_head, skip root blkcg */
	ent = ent->next;
	if (ent == &q->root_blkg->q_node)
		ent = ent->next;
	if (ent == &q->blkg_list)
		return NULL;

	blkg = container_of(ent, struct blkcg_gq, q_node);
	return &blkg->rl;
}

441 442
static int blkcg_reset_stats(struct cgroup_subsys_state *css,
			     struct cftype *cftype, u64 val)
443
{
444
	struct blkcg *blkcg = css_to_blkcg(css);
T
Tejun Heo 已提交
445
	struct blkcg_gq *blkg;
446
	int i;
447

448
	mutex_lock(&blkcg_pol_mutex);
449
	spin_lock_irq(&blkcg->lock);
T
Tejun Heo 已提交
450 451 452 453 454 455

	/*
	 * Note that stat reset is racy - it doesn't synchronize against
	 * stat updates.  This is a debug feature which shouldn't exist
	 * anyway.  If you get hit by a race, retry.
	 */
456
	hlist_for_each_entry(blkg, &blkcg->blkg_list, blkcg_node) {
457 458 459
		blkg_rwstat_reset(&blkg->stat_bytes);
		blkg_rwstat_reset(&blkg->stat_ios);

T
Tejun Heo 已提交
460
		for (i = 0; i < BLKCG_MAX_POLS; i++) {
T
Tejun Heo 已提交
461
			struct blkcg_policy *pol = blkcg_policy[i];
462

463 464
			if (blkg->pd[i] && pol->pd_reset_stats_fn)
				pol->pd_reset_stats_fn(blkg->pd[i]);
465
		}
466
	}
467

468
	spin_unlock_irq(&blkcg->lock);
469
	mutex_unlock(&blkcg_pol_mutex);
470 471 472
	return 0;
}

473
const char *blkg_dev_name(struct blkcg_gq *blkg)
474
{
475
	/* some drivers (floppy) instantiate a queue w/o disk registered */
476 477
	if (blkg->q->backing_dev_info->dev)
		return dev_name(blkg->q->backing_dev_info->dev);
478
	return NULL;
479
}
480
EXPORT_SYMBOL_GPL(blkg_dev_name);
481

482 483 484 485 486 487 488 489 490 491 492
/**
 * blkcg_print_blkgs - helper for printing per-blkg data
 * @sf: seq_file to print to
 * @blkcg: blkcg of interest
 * @prfill: fill function to print out a blkg
 * @pol: policy in question
 * @data: data to be passed to @prfill
 * @show_total: to print out sum of prfill return values or not
 *
 * This function invokes @prfill on each blkg of @blkcg if pd for the
 * policy specified by @pol exists.  @prfill is invoked with @sf, the
493 494 495
 * policy data and @data and the matching queue lock held.  If @show_total
 * is %true, the sum of the return values from @prfill is printed with
 * "Total" label at the end.
496 497 498 499
 *
 * This is to be used to construct print functions for
 * cftype->read_seq_string method.
 */
T
Tejun Heo 已提交
500
void blkcg_print_blkgs(struct seq_file *sf, struct blkcg *blkcg,
501 502
		       u64 (*prfill)(struct seq_file *,
				     struct blkg_policy_data *, int),
T
Tejun Heo 已提交
503
		       const struct blkcg_policy *pol, int data,
504
		       bool show_total)
505
{
T
Tejun Heo 已提交
506
	struct blkcg_gq *blkg;
507
	u64 total = 0;
508

509
	rcu_read_lock();
510
	hlist_for_each_entry_rcu(blkg, &blkcg->blkg_list, blkcg_node) {
511
		spin_lock_irq(blkg->q->queue_lock);
512
		if (blkcg_policy_enabled(blkg->q, pol))
513
			total += prfill(sf, blkg->pd[pol->plid], data);
514 515 516
		spin_unlock_irq(blkg->q->queue_lock);
	}
	rcu_read_unlock();
517 518 519 520

	if (show_total)
		seq_printf(sf, "Total %llu\n", (unsigned long long)total);
}
521
EXPORT_SYMBOL_GPL(blkcg_print_blkgs);
522 523 524 525

/**
 * __blkg_prfill_u64 - prfill helper for a single u64 value
 * @sf: seq_file to print to
526
 * @pd: policy private data of interest
527 528
 * @v: value to print
 *
529
 * Print @v to @sf for the device assocaited with @pd.
530
 */
531
u64 __blkg_prfill_u64(struct seq_file *sf, struct blkg_policy_data *pd, u64 v)
532
{
533
	const char *dname = blkg_dev_name(pd->blkg);
534 535 536 537 538 539 540

	if (!dname)
		return 0;

	seq_printf(sf, "%s %llu\n", dname, (unsigned long long)v);
	return v;
}
541
EXPORT_SYMBOL_GPL(__blkg_prfill_u64);
542 543 544 545

/**
 * __blkg_prfill_rwstat - prfill helper for a blkg_rwstat
 * @sf: seq_file to print to
546
 * @pd: policy private data of interest
547 548
 * @rwstat: rwstat to print
 *
549
 * Print @rwstat to @sf for the device assocaited with @pd.
550
 */
551
u64 __blkg_prfill_rwstat(struct seq_file *sf, struct blkg_policy_data *pd,
552
			 const struct blkg_rwstat *rwstat)
553 554 555 556 557 558
{
	static const char *rwstr[] = {
		[BLKG_RWSTAT_READ]	= "Read",
		[BLKG_RWSTAT_WRITE]	= "Write",
		[BLKG_RWSTAT_SYNC]	= "Sync",
		[BLKG_RWSTAT_ASYNC]	= "Async",
559
		[BLKG_RWSTAT_DISCARD]	= "Discard",
560
	};
561
	const char *dname = blkg_dev_name(pd->blkg);
562 563 564 565 566 567 568 569
	u64 v;
	int i;

	if (!dname)
		return 0;

	for (i = 0; i < BLKG_RWSTAT_NR; i++)
		seq_printf(sf, "%s %s %llu\n", dname, rwstr[i],
T
Tejun Heo 已提交
570
			   (unsigned long long)atomic64_read(&rwstat->aux_cnt[i]));
571

T
Tejun Heo 已提交
572
	v = atomic64_read(&rwstat->aux_cnt[BLKG_RWSTAT_READ]) +
573 574
		atomic64_read(&rwstat->aux_cnt[BLKG_RWSTAT_WRITE]) +
		atomic64_read(&rwstat->aux_cnt[BLKG_RWSTAT_DISCARD]);
575 576 577
	seq_printf(sf, "%s Total %llu\n", dname, (unsigned long long)v);
	return v;
}
T
Tejun Heo 已提交
578
EXPORT_SYMBOL_GPL(__blkg_prfill_rwstat);
579

580 581 582
/**
 * blkg_prfill_stat - prfill callback for blkg_stat
 * @sf: seq_file to print to
583 584
 * @pd: policy private data of interest
 * @off: offset to the blkg_stat in @pd
585 586 587
 *
 * prfill callback for printing a blkg_stat.
 */
588
u64 blkg_prfill_stat(struct seq_file *sf, struct blkg_policy_data *pd, int off)
589
{
590
	return __blkg_prfill_u64(sf, pd, blkg_stat_read((void *)pd + off));
591
}
592
EXPORT_SYMBOL_GPL(blkg_prfill_stat);
593

594 595 596
/**
 * blkg_prfill_rwstat - prfill callback for blkg_rwstat
 * @sf: seq_file to print to
597 598
 * @pd: policy private data of interest
 * @off: offset to the blkg_rwstat in @pd
599 600 601
 *
 * prfill callback for printing a blkg_rwstat.
 */
602 603
u64 blkg_prfill_rwstat(struct seq_file *sf, struct blkg_policy_data *pd,
		       int off)
604
{
605
	struct blkg_rwstat rwstat = blkg_rwstat_read((void *)pd + off);
606

607
	return __blkg_prfill_rwstat(sf, pd, &rwstat);
608
}
609
EXPORT_SYMBOL_GPL(blkg_prfill_rwstat);
610

611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691
static u64 blkg_prfill_rwstat_field(struct seq_file *sf,
				    struct blkg_policy_data *pd, int off)
{
	struct blkg_rwstat rwstat = blkg_rwstat_read((void *)pd->blkg + off);

	return __blkg_prfill_rwstat(sf, pd, &rwstat);
}

/**
 * blkg_print_stat_bytes - seq_show callback for blkg->stat_bytes
 * @sf: seq_file to print to
 * @v: unused
 *
 * To be used as cftype->seq_show to print blkg->stat_bytes.
 * cftype->private must be set to the blkcg_policy.
 */
int blkg_print_stat_bytes(struct seq_file *sf, void *v)
{
	blkcg_print_blkgs(sf, css_to_blkcg(seq_css(sf)),
			  blkg_prfill_rwstat_field, (void *)seq_cft(sf)->private,
			  offsetof(struct blkcg_gq, stat_bytes), true);
	return 0;
}
EXPORT_SYMBOL_GPL(blkg_print_stat_bytes);

/**
 * blkg_print_stat_bytes - seq_show callback for blkg->stat_ios
 * @sf: seq_file to print to
 * @v: unused
 *
 * To be used as cftype->seq_show to print blkg->stat_ios.  cftype->private
 * must be set to the blkcg_policy.
 */
int blkg_print_stat_ios(struct seq_file *sf, void *v)
{
	blkcg_print_blkgs(sf, css_to_blkcg(seq_css(sf)),
			  blkg_prfill_rwstat_field, (void *)seq_cft(sf)->private,
			  offsetof(struct blkcg_gq, stat_ios), true);
	return 0;
}
EXPORT_SYMBOL_GPL(blkg_print_stat_ios);

static u64 blkg_prfill_rwstat_field_recursive(struct seq_file *sf,
					      struct blkg_policy_data *pd,
					      int off)
{
	struct blkg_rwstat rwstat = blkg_rwstat_recursive_sum(pd->blkg,
							      NULL, off);
	return __blkg_prfill_rwstat(sf, pd, &rwstat);
}

/**
 * blkg_print_stat_bytes_recursive - recursive version of blkg_print_stat_bytes
 * @sf: seq_file to print to
 * @v: unused
 */
int blkg_print_stat_bytes_recursive(struct seq_file *sf, void *v)
{
	blkcg_print_blkgs(sf, css_to_blkcg(seq_css(sf)),
			  blkg_prfill_rwstat_field_recursive,
			  (void *)seq_cft(sf)->private,
			  offsetof(struct blkcg_gq, stat_bytes), true);
	return 0;
}
EXPORT_SYMBOL_GPL(blkg_print_stat_bytes_recursive);

/**
 * blkg_print_stat_ios_recursive - recursive version of blkg_print_stat_ios
 * @sf: seq_file to print to
 * @v: unused
 */
int blkg_print_stat_ios_recursive(struct seq_file *sf, void *v)
{
	blkcg_print_blkgs(sf, css_to_blkcg(seq_css(sf)),
			  blkg_prfill_rwstat_field_recursive,
			  (void *)seq_cft(sf)->private,
			  offsetof(struct blkcg_gq, stat_ios), true);
	return 0;
}
EXPORT_SYMBOL_GPL(blkg_print_stat_ios_recursive);

692 693
/**
 * blkg_stat_recursive_sum - collect hierarchical blkg_stat
694 695 696
 * @blkg: blkg of interest
 * @pol: blkcg_policy which contains the blkg_stat
 * @off: offset to the blkg_stat in blkg_policy_data or @blkg
697
 *
698 699 700 701 702 703
 * Collect the blkg_stat specified by @blkg, @pol and @off and all its
 * online descendants and their aux counts.  The caller must be holding the
 * queue lock for online tests.
 *
 * If @pol is NULL, blkg_stat is at @off bytes into @blkg; otherwise, it is
 * at @off bytes into @blkg's blkg_policy_data of the policy.
704
 */
705 706
u64 blkg_stat_recursive_sum(struct blkcg_gq *blkg,
			    struct blkcg_policy *pol, int off)
707 708
{
	struct blkcg_gq *pos_blkg;
709
	struct cgroup_subsys_state *pos_css;
710
	u64 sum = 0;
711

712
	lockdep_assert_held(blkg->q->queue_lock);
713 714

	rcu_read_lock();
715 716 717 718 719
	blkg_for_each_descendant_pre(pos_blkg, pos_css, blkg) {
		struct blkg_stat *stat;

		if (!pos_blkg->online)
			continue;
720

721 722 723 724 725 726
		if (pol)
			stat = (void *)blkg_to_pd(pos_blkg, pol) + off;
		else
			stat = (void *)blkg + off;

		sum += blkg_stat_read(stat) + atomic64_read(&stat->aux_cnt);
727 728 729 730 731 732 733 734 735
	}
	rcu_read_unlock();

	return sum;
}
EXPORT_SYMBOL_GPL(blkg_stat_recursive_sum);

/**
 * blkg_rwstat_recursive_sum - collect hierarchical blkg_rwstat
736 737 738
 * @blkg: blkg of interest
 * @pol: blkcg_policy which contains the blkg_rwstat
 * @off: offset to the blkg_rwstat in blkg_policy_data or @blkg
739
 *
740 741 742 743 744 745
 * Collect the blkg_rwstat specified by @blkg, @pol and @off and all its
 * online descendants and their aux counts.  The caller must be holding the
 * queue lock for online tests.
 *
 * If @pol is NULL, blkg_rwstat is at @off bytes into @blkg; otherwise, it
 * is at @off bytes into @blkg's blkg_policy_data of the policy.
746
 */
747 748
struct blkg_rwstat blkg_rwstat_recursive_sum(struct blkcg_gq *blkg,
					     struct blkcg_policy *pol, int off)
749 750
{
	struct blkcg_gq *pos_blkg;
751
	struct cgroup_subsys_state *pos_css;
752
	struct blkg_rwstat sum = { };
753 754
	int i;

755
	lockdep_assert_held(blkg->q->queue_lock);
756 757

	rcu_read_lock();
758
	blkg_for_each_descendant_pre(pos_blkg, pos_css, blkg) {
759
		struct blkg_rwstat *rwstat;
760 761 762 763

		if (!pos_blkg->online)
			continue;

764 765 766 767 768
		if (pol)
			rwstat = (void *)blkg_to_pd(pos_blkg, pol) + off;
		else
			rwstat = (void *)pos_blkg + off;

769
		for (i = 0; i < BLKG_RWSTAT_NR; i++)
770 771 772
			atomic64_add(atomic64_read(&rwstat->aux_cnt[i]) +
				percpu_counter_sum_positive(&rwstat->cpu_cnt[i]),
				&sum.aux_cnt[i]);
773 774 775 776 777 778 779
	}
	rcu_read_unlock();

	return sum;
}
EXPORT_SYMBOL_GPL(blkg_rwstat_recursive_sum);

780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800
/* Performs queue bypass and policy enabled checks then looks up blkg. */
static struct blkcg_gq *blkg_lookup_check(struct blkcg *blkcg,
					  const struct blkcg_policy *pol,
					  struct request_queue *q)
{
	WARN_ON_ONCE(!rcu_read_lock_held());
	lockdep_assert_held(q->queue_lock);

	if (!blkcg_policy_enabled(q, pol))
		return ERR_PTR(-EOPNOTSUPP);

	/*
	 * This could be the first entry point of blkcg implementation and
	 * we shouldn't allow anything to go through for a bypassing queue.
	 */
	if (unlikely(blk_queue_bypass(q)))
		return ERR_PTR(blk_queue_dying(q) ? -ENODEV : -EBUSY);

	return __blkg_lookup(blkcg, q, true /* update_hint */);
}

801 802 803
/**
 * blkg_conf_prep - parse and prepare for per-blkg config update
 * @blkcg: target block cgroup
804
 * @pol: target policy
805 806 807 808
 * @input: input string
 * @ctx: blkg_conf_ctx to be filled
 *
 * Parse per-blkg config update from @input and initialize @ctx with the
809 810 811
 * result.  @ctx->blkg points to the blkg to be updated and @ctx->body the
 * part of @input following MAJ:MIN.  This function returns with RCU read
 * lock and queue lock held and must be paired with blkg_conf_finish().
812
 */
T
Tejun Heo 已提交
813
int blkg_conf_prep(struct blkcg *blkcg, const struct blkcg_policy *pol,
814
		   char *input, struct blkg_conf_ctx *ctx)
815
	__acquires(rcu) __acquires(disk->queue->queue_lock)
816
{
817
	struct gendisk *disk;
818
	struct request_queue *q;
T
Tejun Heo 已提交
819
	struct blkcg_gq *blkg;
T
Tejun Heo 已提交
820
	unsigned int major, minor;
821 822
	int key_len, part, ret;
	char *body;
823

824
	if (sscanf(input, "%u:%u%n", &major, &minor, &key_len) != 2)
T
Tejun Heo 已提交
825
		return -EINVAL;
826

827 828 829 830 831
	body = input + key_len;
	if (!isspace(*body))
		return -EINVAL;
	body = skip_spaces(body);

T
Tejun Heo 已提交
832
	disk = get_gendisk(MKDEV(major, minor), &part);
833
	if (!disk)
834
		return -ENODEV;
835
	if (part) {
836 837
		ret = -ENODEV;
		goto fail;
838
	}
839

840
	q = disk->queue;
841

842 843
	rcu_read_lock();
	spin_lock_irq(q->queue_lock);
844

845
	blkg = blkg_lookup_check(blkcg, pol, q);
T
Tejun Heo 已提交
846 847
	if (IS_ERR(blkg)) {
		ret = PTR_ERR(blkg);
848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870
		goto fail_unlock;
	}

	if (blkg)
		goto success;

	/*
	 * Create blkgs walking down from blkcg_root to @blkcg, so that all
	 * non-root blkgs have access to their parents.
	 */
	while (true) {
		struct blkcg *pos = blkcg;
		struct blkcg *parent;
		struct blkcg_gq *new_blkg;

		parent = blkcg_parent(blkcg);
		while (parent && !__blkg_lookup(parent, q, false)) {
			pos = parent;
			parent = blkcg_parent(parent);
		}

		/* Drop locks to do new blkg allocation with GFP_KERNEL. */
		spin_unlock_irq(q->queue_lock);
871
		rcu_read_unlock();
872 873 874 875 876

		new_blkg = blkg_alloc(pos, q, GFP_KERNEL);
		if (unlikely(!new_blkg)) {
			ret = -ENOMEM;
			goto fail;
877
		}
878

879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901
		rcu_read_lock();
		spin_lock_irq(q->queue_lock);

		blkg = blkg_lookup_check(pos, pol, q);
		if (IS_ERR(blkg)) {
			ret = PTR_ERR(blkg);
			goto fail_unlock;
		}

		if (blkg) {
			blkg_free(new_blkg);
		} else {
			blkg = blkg_create(pos, q, new_blkg);
			if (unlikely(IS_ERR(blkg))) {
				ret = PTR_ERR(blkg);
				goto fail_unlock;
			}
		}

		if (pos == blkcg)
			goto success;
	}
success:
902 903
	ctx->disk = disk;
	ctx->blkg = blkg;
904
	ctx->body = body;
T
Tejun Heo 已提交
905
	return 0;
906 907 908 909 910

fail_unlock:
	spin_unlock_irq(q->queue_lock);
	rcu_read_unlock();
fail:
911
	put_disk_and_module(disk);
912 913 914 915 916 917 918 919 920 921 922
	/*
	 * If queue was bypassing, we should retry.  Do so after a
	 * short msleep().  It isn't strictly necessary but queue
	 * can be bypassing for some time and it's always nice to
	 * avoid busy looping.
	 */
	if (ret == -EBUSY) {
		msleep(10);
		ret = restart_syscall();
	}
	return ret;
923
}
924
EXPORT_SYMBOL_GPL(blkg_conf_prep);
925

926 927 928 929 930 931 932
/**
 * blkg_conf_finish - finish up per-blkg config update
 * @ctx: blkg_conf_ctx intiailized by blkg_conf_prep()
 *
 * Finish up after per-blkg config update.  This function must be paired
 * with blkg_conf_prep().
 */
933
void blkg_conf_finish(struct blkg_conf_ctx *ctx)
934
	__releases(ctx->disk->queue->queue_lock) __releases(rcu)
935
{
936
	spin_unlock_irq(ctx->disk->queue->queue_lock);
937
	rcu_read_unlock();
938
	put_disk_and_module(ctx->disk);
939
}
940
EXPORT_SYMBOL_GPL(blkg_conf_finish);
941

942 943 944 945 946 947 948 949 950
static int blkcg_print_stat(struct seq_file *sf, void *v)
{
	struct blkcg *blkcg = css_to_blkcg(seq_css(sf));
	struct blkcg_gq *blkg;

	rcu_read_lock();

	hlist_for_each_entry_rcu(blkg, &blkcg->blkg_list, blkcg_node) {
		const char *dname;
951
		char *buf;
952
		struct blkg_rwstat rwstat;
953
		u64 rbytes, wbytes, rios, wios, dbytes, dios;
954 955 956
		size_t size = seq_get_buf(sf, &buf), off = 0;
		int i;
		bool has_stats = false;
957 958 959 960 961

		dname = blkg_dev_name(blkg);
		if (!dname)
			continue;

962 963 964 965 966 967 968 969
		/*
		 * Hooray string manipulation, count is the size written NOT
		 * INCLUDING THE \0, so size is now count+1 less than what we
		 * had before, but we want to start writing the next bit from
		 * the \0 so we only add count to buf.
		 */
		off += scnprintf(buf+off, size-off, "%s ", dname);

970 971 972 973 974 975
		spin_lock_irq(blkg->q->queue_lock);

		rwstat = blkg_rwstat_recursive_sum(blkg, NULL,
					offsetof(struct blkcg_gq, stat_bytes));
		rbytes = atomic64_read(&rwstat.aux_cnt[BLKG_RWSTAT_READ]);
		wbytes = atomic64_read(&rwstat.aux_cnt[BLKG_RWSTAT_WRITE]);
976
		dbytes = atomic64_read(&rwstat.aux_cnt[BLKG_RWSTAT_DISCARD]);
977 978 979 980 981

		rwstat = blkg_rwstat_recursive_sum(blkg, NULL,
					offsetof(struct blkcg_gq, stat_ios));
		rios = atomic64_read(&rwstat.aux_cnt[BLKG_RWSTAT_READ]);
		wios = atomic64_read(&rwstat.aux_cnt[BLKG_RWSTAT_WRITE]);
982
		dios = atomic64_read(&rwstat.aux_cnt[BLKG_RWSTAT_DISCARD]);
983 984 985

		spin_unlock_irq(blkg->q->queue_lock);

986 987 988
		if (rbytes || wbytes || rios || wios) {
			has_stats = true;
			off += scnprintf(buf+off, size-off,
989 990 991
					 "rbytes=%llu wbytes=%llu rios=%llu wios=%llu dbytes=%llu dios=%llu",
					 rbytes, wbytes, rios, wios,
					 dbytes, dios);
992 993 994 995 996
		}

		if (!blkcg_debug_stats)
			goto next;

997 998 999 1000 1001 1002 1003 1004
		if (atomic_read(&blkg->use_delay)) {
			has_stats = true;
			off += scnprintf(buf+off, size-off,
					 " use_delay=%d delay_nsec=%llu",
					 atomic_read(&blkg->use_delay),
					(unsigned long long)atomic64_read(&blkg->delay_nsec));
		}

1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021
		for (i = 0; i < BLKCG_MAX_POLS; i++) {
			struct blkcg_policy *pol = blkcg_policy[i];
			size_t written;

			if (!blkg->pd[i] || !pol->pd_stat_fn)
				continue;

			written = pol->pd_stat_fn(blkg->pd[i], buf+off, size-off);
			if (written)
				has_stats = true;
			off += written;
		}
next:
		if (has_stats) {
			off += scnprintf(buf+off, size-off, "\n");
			seq_commit(sf, off);
		}
1022 1023 1024 1025 1026 1027
	}

	rcu_read_unlock();
	return 0;
}

1028
static struct cftype blkcg_files[] = {
1029 1030
	{
		.name = "stat",
1031
		.flags = CFTYPE_NOT_ON_ROOT,
1032 1033 1034 1035 1036
		.seq_show = blkcg_print_stat,
	},
	{ }	/* terminate */
};

1037
static struct cftype blkcg_legacy_files[] = {
1038 1039
	{
		.name = "reset_stats",
T
Tejun Heo 已提交
1040
		.write_u64 = blkcg_reset_stats,
1041
	},
1042
	{ }	/* terminate */
1043 1044
};

1045
/**
1046
 * blkcg_css_offline - cgroup css_offline callback
1047
 * @css: css of interest
1048
 *
1049
 * This function is called when @css is about to go away and responsible
1050 1051 1052
 * for shooting down all blkgs associated with @css.  blkgs should be
 * removed while holding both q and blkcg locks.  As blkcg lock is nested
 * inside q lock, this function performs reverse double lock dancing.
1053 1054 1055
 *
 * This is the blkcg counterpart of ioc_release_fn().
 */
1056
static void blkcg_css_offline(struct cgroup_subsys_state *css)
1057
{
1058
	struct blkcg *blkcg = css_to_blkcg(css);
1059

1060
	spin_lock_irq(&blkcg->lock);
1061

1062 1063
	while (!hlist_empty(&blkcg->blkg_list)) {
		struct blkcg_gq *blkg = hlist_entry(blkcg->blkg_list.first,
1064
						struct blkcg_gq, blkcg_node);
1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075
		struct request_queue *q = blkg->q;

		if (spin_trylock(q->queue_lock)) {
			blkg_destroy(blkg);
			spin_unlock(q->queue_lock);
		} else {
			spin_unlock_irq(&blkcg->lock);
			cpu_relax();
			spin_lock_irq(&blkcg->lock);
		}
	}
1076

1077
	spin_unlock_irq(&blkcg->lock);
1078 1079

	wb_blkcg_offline(blkcg);
1080 1081
}

1082
static void blkcg_css_free(struct cgroup_subsys_state *css)
1083
{
1084
	struct blkcg *blkcg = css_to_blkcg(css);
1085
	int i;
1086

T
Tejun Heo 已提交
1087
	mutex_lock(&blkcg_pol_mutex);
1088

T
Tejun Heo 已提交
1089 1090
	list_del(&blkcg->all_blkcgs_node);

1091
	for (i = 0; i < BLKCG_MAX_POLS; i++)
1092 1093 1094 1095 1096
		if (blkcg->cpd[i])
			blkcg_policy[i]->cpd_free_fn(blkcg->cpd[i]);

	mutex_unlock(&blkcg_pol_mutex);

1097
	kfree(blkcg);
1098 1099
}

1100 1101
static struct cgroup_subsys_state *
blkcg_css_alloc(struct cgroup_subsys_state *parent_css)
1102
{
T
Tejun Heo 已提交
1103
	struct blkcg *blkcg;
1104 1105
	struct cgroup_subsys_state *ret;
	int i;
1106

T
Tejun Heo 已提交
1107 1108
	mutex_lock(&blkcg_pol_mutex);

1109
	if (!parent_css) {
T
Tejun Heo 已提交
1110
		blkcg = &blkcg_root;
1111 1112 1113 1114
	} else {
		blkcg = kzalloc(sizeof(*blkcg), GFP_KERNEL);
		if (!blkcg) {
			ret = ERR_PTR(-ENOMEM);
1115
			goto unlock;
1116
		}
1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128
	}

	for (i = 0; i < BLKCG_MAX_POLS ; i++) {
		struct blkcg_policy *pol = blkcg_policy[i];
		struct blkcg_policy_data *cpd;

		/*
		 * If the policy hasn't been attached yet, wait for it
		 * to be attached before doing anything else. Otherwise,
		 * check if the policy requires any specific per-cgroup
		 * data: if it does, allocate and initialize it.
		 */
1129
		if (!pol || !pol->cpd_alloc_fn)
1130 1131
			continue;

1132
		cpd = pol->cpd_alloc_fn(GFP_KERNEL);
1133 1134 1135 1136
		if (!cpd) {
			ret = ERR_PTR(-ENOMEM);
			goto free_pd_blkcg;
		}
1137 1138
		blkcg->cpd[i] = cpd;
		cpd->blkcg = blkcg;
1139
		cpd->plid = i;
1140 1141
		if (pol->cpd_init_fn)
			pol->cpd_init_fn(cpd);
1142
	}
1143 1144

	spin_lock_init(&blkcg->lock);
1145
	INIT_RADIX_TREE(&blkcg->blkg_tree, GFP_NOWAIT | __GFP_NOWARN);
1146
	INIT_HLIST_HEAD(&blkcg->blkg_list);
1147 1148 1149
#ifdef CONFIG_CGROUP_WRITEBACK
	INIT_LIST_HEAD(&blkcg->cgwb_list);
#endif
T
Tejun Heo 已提交
1150 1151 1152
	list_add_tail(&blkcg->all_blkcgs_node, &all_blkcgs);

	mutex_unlock(&blkcg_pol_mutex);
1153
	return &blkcg->css;
1154 1155 1156

free_pd_blkcg:
	for (i--; i >= 0; i--)
1157 1158
		if (blkcg->cpd[i])
			blkcg_policy[i]->cpd_free_fn(blkcg->cpd[i]);
1159 1160 1161 1162

	if (blkcg != &blkcg_root)
		kfree(blkcg);
unlock:
T
Tejun Heo 已提交
1163
	mutex_unlock(&blkcg_pol_mutex);
1164
	return ret;
1165 1166
}

1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178
/**
 * blkcg_init_queue - initialize blkcg part of request queue
 * @q: request_queue to initialize
 *
 * Called from blk_alloc_queue_node(). Responsible for initializing blkcg
 * part of new request_queue @q.
 *
 * RETURNS:
 * 0 on success, -errno on failure.
 */
int blkcg_init_queue(struct request_queue *q)
{
1179 1180
	struct blkcg_gq *new_blkg, *blkg;
	bool preloaded;
1181 1182
	int ret;

1183 1184 1185 1186 1187 1188
	new_blkg = blkg_alloc(&blkcg_root, q, GFP_KERNEL);
	if (!new_blkg)
		return -ENOMEM;

	preloaded = !radix_tree_preload(GFP_KERNEL);

1189
	/* Make sure the root blkg exists. */
1190 1191
	rcu_read_lock();
	spin_lock_irq(q->queue_lock);
1192
	blkg = blkg_create(&blkcg_root, q, new_blkg);
1193 1194 1195 1196
	if (IS_ERR(blkg))
		goto err_unlock;
	q->root_blkg = blkg;
	q->root_rl.blkg = blkg;
1197 1198 1199
	spin_unlock_irq(q->queue_lock);
	rcu_read_unlock();

1200 1201 1202
	if (preloaded)
		radix_tree_preload_end();

1203 1204 1205 1206 1207 1208 1209 1210
	ret = blk_iolatency_init(q);
	if (ret) {
		spin_lock_irq(q->queue_lock);
		blkg_destroy_all(q);
		spin_unlock_irq(q->queue_lock);
		return ret;
	}

1211 1212 1213 1214 1215 1216 1217
	ret = blk_throtl_init(q);
	if (ret) {
		spin_lock_irq(q->queue_lock);
		blkg_destroy_all(q);
		spin_unlock_irq(q->queue_lock);
	}
	return ret;
1218 1219 1220 1221 1222 1223 1224

err_unlock:
	spin_unlock_irq(q->queue_lock);
	rcu_read_unlock();
	if (preloaded)
		radix_tree_preload_end();
	return PTR_ERR(blkg);
1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236
}

/**
 * blkcg_drain_queue - drain blkcg part of request_queue
 * @q: request_queue to drain
 *
 * Called from blk_drain_queue().  Responsible for draining blkcg part.
 */
void blkcg_drain_queue(struct request_queue *q)
{
	lockdep_assert_held(q->queue_lock);

1237 1238 1239 1240 1241 1242 1243
	/*
	 * @q could be exiting and already have destroyed all blkgs as
	 * indicated by NULL root_blkg.  If so, don't confuse policies.
	 */
	if (!q->root_blkg)
		return;

1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254
	blk_throtl_drain(q);
}

/**
 * blkcg_exit_queue - exit and release blkcg part of request_queue
 * @q: request_queue being released
 *
 * Called from blk_release_queue().  Responsible for exiting blkcg part.
 */
void blkcg_exit_queue(struct request_queue *q)
{
1255
	spin_lock_irq(q->queue_lock);
1256
	blkg_destroy_all(q);
1257 1258
	spin_unlock_irq(q->queue_lock);

1259 1260 1261
	blk_throtl_exit(q);
}

1262 1263 1264 1265 1266 1267
/*
 * We cannot support shared io contexts, as we have no mean to support
 * two tasks with the same ioc in two different groups without major rework
 * of the main cic data structures.  For now we allow a task to change
 * its cgroup only if it's the only owner of its ioc.
 */
1268
static int blkcg_can_attach(struct cgroup_taskset *tset)
1269
{
1270
	struct task_struct *task;
1271
	struct cgroup_subsys_state *dst_css;
1272 1273 1274 1275
	struct io_context *ioc;
	int ret = 0;

	/* task_lock() is needed to avoid races with exit_io_context() */
1276
	cgroup_taskset_for_each(task, dst_css, tset) {
1277 1278 1279 1280 1281 1282 1283 1284
		task_lock(task);
		ioc = task->io_context;
		if (ioc && atomic_read(&ioc->nr_tasks) > 1)
			ret = -EINVAL;
		task_unlock(task);
		if (ret)
			break;
	}
1285 1286 1287
	return ret;
}

1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307
static void blkcg_bind(struct cgroup_subsys_state *root_css)
{
	int i;

	mutex_lock(&blkcg_pol_mutex);

	for (i = 0; i < BLKCG_MAX_POLS; i++) {
		struct blkcg_policy *pol = blkcg_policy[i];
		struct blkcg *blkcg;

		if (!pol || !pol->cpd_bind_fn)
			continue;

		list_for_each_entry(blkcg, &all_blkcgs, all_blkcgs_node)
			if (blkcg->cpd[pol->plid])
				pol->cpd_bind_fn(blkcg->cpd[pol->plid]);
	}
	mutex_unlock(&blkcg_pol_mutex);
}

1308 1309 1310 1311 1312 1313 1314
static void blkcg_exit(struct task_struct *tsk)
{
	if (tsk->throttle_queue)
		blk_put_queue(tsk->throttle_queue);
	tsk->throttle_queue = NULL;
}

1315
struct cgroup_subsys io_cgrp_subsys = {
1316 1317 1318
	.css_alloc = blkcg_css_alloc,
	.css_offline = blkcg_css_offline,
	.css_free = blkcg_css_free,
T
Tejun Heo 已提交
1319
	.can_attach = blkcg_can_attach,
1320
	.bind = blkcg_bind,
1321
	.dfl_cftypes = blkcg_files,
1322
	.legacy_cftypes = blkcg_legacy_files,
1323
	.legacy_name = "blkio",
1324
	.exit = blkcg_exit,
1325 1326 1327 1328 1329 1330 1331 1332
#ifdef CONFIG_MEMCG
	/*
	 * This ensures that, if available, memcg is automatically enabled
	 * together on the default hierarchy so that the owner cgroup can
	 * be retrieved from writeback pages.
	 */
	.depends_on = 1 << memory_cgrp_id,
#endif
1333
};
1334
EXPORT_SYMBOL_GPL(io_cgrp_subsys);
1335

1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352
/**
 * blkcg_activate_policy - activate a blkcg policy on a request_queue
 * @q: request_queue of interest
 * @pol: blkcg policy to activate
 *
 * Activate @pol on @q.  Requires %GFP_KERNEL context.  @q goes through
 * bypass mode to populate its blkgs with policy_data for @pol.
 *
 * Activation happens with @q bypassed, so nobody would be accessing blkgs
 * from IO path.  Update of each blkg is protected by both queue and blkcg
 * locks so that holding either lock and testing blkcg_policy_enabled() is
 * always enough for dereferencing policy data.
 *
 * The caller is responsible for synchronizing [de]activations and policy
 * [un]registerations.  Returns 0 on success, -errno on failure.
 */
int blkcg_activate_policy(struct request_queue *q,
T
Tejun Heo 已提交
1353
			  const struct blkcg_policy *pol)
1354
{
1355
	struct blkg_policy_data *pd_prealloc = NULL;
1356
	struct blkcg_gq *blkg;
1357
	int ret;
1358 1359 1360 1361

	if (blkcg_policy_enabled(q, pol))
		return 0;

1362
	if (q->mq_ops)
1363
		blk_mq_freeze_queue(q);
1364
	else
1365
		blk_queue_bypass_start(q);
1366 1367
pd_prealloc:
	if (!pd_prealloc) {
1368
		pd_prealloc = pol->pd_alloc_fn(GFP_KERNEL, q->node);
1369
		if (!pd_prealloc) {
1370
			ret = -ENOMEM;
1371
			goto out_bypass_end;
1372 1373 1374 1375 1376 1377
		}
	}

	spin_lock_irq(q->queue_lock);

	list_for_each_entry(blkg, &q->blkg_list, q_node) {
1378 1379 1380 1381
		struct blkg_policy_data *pd;

		if (blkg->pd[pol->plid])
			continue;
1382

1383
		pd = pol->pd_alloc_fn(GFP_NOWAIT | __GFP_NOWARN, q->node);
1384 1385 1386 1387 1388 1389
		if (!pd)
			swap(pd, pd_prealloc);
		if (!pd) {
			spin_unlock_irq(q->queue_lock);
			goto pd_prealloc;
		}
1390 1391 1392

		blkg->pd[pol->plid] = pd;
		pd->blkg = blkg;
T
Tejun Heo 已提交
1393
		pd->plid = pol->plid;
1394
		if (pol->pd_init_fn)
1395
			pol->pd_init_fn(pd);
1396 1397 1398 1399
	}

	__set_bit(pol->plid, q->blkcg_pols);
	ret = 0;
1400

1401
	spin_unlock_irq(q->queue_lock);
1402
out_bypass_end:
1403
	if (q->mq_ops)
1404
		blk_mq_unfreeze_queue(q);
1405
	else
1406
		blk_queue_bypass_end(q);
1407 1408
	if (pd_prealloc)
		pol->pd_free_fn(pd_prealloc);
1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421
	return ret;
}
EXPORT_SYMBOL_GPL(blkcg_activate_policy);

/**
 * blkcg_deactivate_policy - deactivate a blkcg policy on a request_queue
 * @q: request_queue of interest
 * @pol: blkcg policy to deactivate
 *
 * Deactivate @pol on @q.  Follows the same synchronization rules as
 * blkcg_activate_policy().
 */
void blkcg_deactivate_policy(struct request_queue *q,
T
Tejun Heo 已提交
1422
			     const struct blkcg_policy *pol)
1423
{
T
Tejun Heo 已提交
1424
	struct blkcg_gq *blkg;
1425 1426 1427 1428

	if (!blkcg_policy_enabled(q, pol))
		return;

1429
	if (q->mq_ops)
1430
		blk_mq_freeze_queue(q);
1431
	else
1432 1433
		blk_queue_bypass_start(q);

1434 1435 1436 1437 1438
	spin_lock_irq(q->queue_lock);

	__clear_bit(pol->plid, q->blkcg_pols);

	list_for_each_entry(blkg, &q->blkg_list, q_node) {
1439
		if (blkg->pd[pol->plid]) {
1440
			if (pol->pd_offline_fn)
1441
				pol->pd_offline_fn(blkg->pd[pol->plid]);
1442 1443 1444
			pol->pd_free_fn(blkg->pd[pol->plid]);
			blkg->pd[pol->plid] = NULL;
		}
1445 1446 1447
	}

	spin_unlock_irq(q->queue_lock);
1448

1449
	if (q->mq_ops)
1450
		blk_mq_unfreeze_queue(q);
1451
	else
1452
		blk_queue_bypass_end(q);
1453 1454 1455
}
EXPORT_SYMBOL_GPL(blkcg_deactivate_policy);

T
Tejun Heo 已提交
1456
/**
T
Tejun Heo 已提交
1457 1458
 * blkcg_policy_register - register a blkcg policy
 * @pol: blkcg policy to register
T
Tejun Heo 已提交
1459
 *
T
Tejun Heo 已提交
1460 1461
 * Register @pol with blkcg core.  Might sleep and @pol may be modified on
 * successful registration.  Returns 0 on success and -errno on failure.
T
Tejun Heo 已提交
1462
 */
1463
int blkcg_policy_register(struct blkcg_policy *pol)
1464
{
1465
	struct blkcg *blkcg;
T
Tejun Heo 已提交
1466
	int i, ret;
1467

1468
	mutex_lock(&blkcg_pol_register_mutex);
1469 1470
	mutex_lock(&blkcg_pol_mutex);

T
Tejun Heo 已提交
1471 1472 1473
	/* find an empty slot */
	ret = -ENOSPC;
	for (i = 0; i < BLKCG_MAX_POLS; i++)
T
Tejun Heo 已提交
1474
		if (!blkcg_policy[i])
T
Tejun Heo 已提交
1475 1476
			break;
	if (i >= BLKCG_MAX_POLS)
1477
		goto err_unlock;
1478

1479 1480 1481 1482 1483
	/* Make sure cpd/pd_alloc_fn and cpd/pd_free_fn in pairs */
	if ((!pol->cpd_alloc_fn ^ !pol->cpd_free_fn) ||
		(!pol->pd_alloc_fn ^ !pol->pd_free_fn))
		goto err_unlock;

1484
	/* register @pol */
T
Tejun Heo 已提交
1485
	pol->plid = i;
1486 1487 1488
	blkcg_policy[pol->plid] = pol;

	/* allocate and install cpd's */
1489
	if (pol->cpd_alloc_fn) {
1490 1491 1492
		list_for_each_entry(blkcg, &all_blkcgs, all_blkcgs_node) {
			struct blkcg_policy_data *cpd;

1493
			cpd = pol->cpd_alloc_fn(GFP_KERNEL);
1494
			if (!cpd)
1495 1496
				goto err_free_cpds;

1497 1498
			blkcg->cpd[pol->plid] = cpd;
			cpd->blkcg = blkcg;
1499
			cpd->plid = pol->plid;
1500
			pol->cpd_init_fn(cpd);
1501 1502 1503
		}
	}

1504
	mutex_unlock(&blkcg_pol_mutex);
T
Tejun Heo 已提交
1505 1506

	/* everything is in place, add intf files for the new policy */
1507 1508 1509
	if (pol->dfl_cftypes)
		WARN_ON(cgroup_add_dfl_cftypes(&io_cgrp_subsys,
					       pol->dfl_cftypes));
1510
	if (pol->legacy_cftypes)
1511
		WARN_ON(cgroup_add_legacy_cftypes(&io_cgrp_subsys,
1512
						  pol->legacy_cftypes));
1513 1514 1515
	mutex_unlock(&blkcg_pol_register_mutex);
	return 0;

1516
err_free_cpds:
1517
	if (pol->cpd_free_fn) {
1518
		list_for_each_entry(blkcg, &all_blkcgs, all_blkcgs_node) {
1519 1520 1521 1522
			if (blkcg->cpd[pol->plid]) {
				pol->cpd_free_fn(blkcg->cpd[pol->plid]);
				blkcg->cpd[pol->plid] = NULL;
			}
1523 1524 1525
		}
	}
	blkcg_policy[pol->plid] = NULL;
1526
err_unlock:
1527
	mutex_unlock(&blkcg_pol_mutex);
1528
	mutex_unlock(&blkcg_pol_register_mutex);
T
Tejun Heo 已提交
1529
	return ret;
1530
}
T
Tejun Heo 已提交
1531
EXPORT_SYMBOL_GPL(blkcg_policy_register);
1532

T
Tejun Heo 已提交
1533
/**
T
Tejun Heo 已提交
1534 1535
 * blkcg_policy_unregister - unregister a blkcg policy
 * @pol: blkcg policy to unregister
T
Tejun Heo 已提交
1536
 *
T
Tejun Heo 已提交
1537
 * Undo blkcg_policy_register(@pol).  Might sleep.
T
Tejun Heo 已提交
1538
 */
T
Tejun Heo 已提交
1539
void blkcg_policy_unregister(struct blkcg_policy *pol)
1540
{
1541 1542
	struct blkcg *blkcg;

1543
	mutex_lock(&blkcg_pol_register_mutex);
1544

T
Tejun Heo 已提交
1545
	if (WARN_ON(blkcg_policy[pol->plid] != pol))
T
Tejun Heo 已提交
1546 1547 1548
		goto out_unlock;

	/* kill the intf files first */
1549 1550
	if (pol->dfl_cftypes)
		cgroup_rm_cftypes(pol->dfl_cftypes);
1551 1552
	if (pol->legacy_cftypes)
		cgroup_rm_cftypes(pol->legacy_cftypes);
1553

1554
	/* remove cpds and unregister */
1555
	mutex_lock(&blkcg_pol_mutex);
1556

1557
	if (pol->cpd_free_fn) {
1558
		list_for_each_entry(blkcg, &all_blkcgs, all_blkcgs_node) {
1559 1560 1561 1562
			if (blkcg->cpd[pol->plid]) {
				pol->cpd_free_fn(blkcg->cpd[pol->plid]);
				blkcg->cpd[pol->plid] = NULL;
			}
1563 1564
		}
	}
T
Tejun Heo 已提交
1565
	blkcg_policy[pol->plid] = NULL;
1566

1567
	mutex_unlock(&blkcg_pol_mutex);
1568 1569
out_unlock:
	mutex_unlock(&blkcg_pol_register_mutex);
1570
}
T
Tejun Heo 已提交
1571
EXPORT_SYMBOL_GPL(blkcg_policy_unregister);
1572

1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 1693 1694 1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 1713 1714 1715 1716 1717 1718
/*
 * Scale the accumulated delay based on how long it has been since we updated
 * the delay.  We only call this when we are adding delay, in case it's been a
 * while since we added delay, and when we are checking to see if we need to
 * delay a task, to account for any delays that may have occurred.
 */
static void blkcg_scale_delay(struct blkcg_gq *blkg, u64 now)
{
	u64 old = atomic64_read(&blkg->delay_start);

	/*
	 * We only want to scale down every second.  The idea here is that we
	 * want to delay people for min(delay_nsec, NSEC_PER_SEC) in a certain
	 * time window.  We only want to throttle tasks for recent delay that
	 * has occurred, in 1 second time windows since that's the maximum
	 * things can be throttled.  We save the current delay window in
	 * blkg->last_delay so we know what amount is still left to be charged
	 * to the blkg from this point onward.  blkg->last_use keeps track of
	 * the use_delay counter.  The idea is if we're unthrottling the blkg we
	 * are ok with whatever is happening now, and we can take away more of
	 * the accumulated delay as we've already throttled enough that
	 * everybody is happy with their IO latencies.
	 */
	if (time_before64(old + NSEC_PER_SEC, now) &&
	    atomic64_cmpxchg(&blkg->delay_start, old, now) == old) {
		u64 cur = atomic64_read(&blkg->delay_nsec);
		u64 sub = min_t(u64, blkg->last_delay, now - old);
		int cur_use = atomic_read(&blkg->use_delay);

		/*
		 * We've been unthrottled, subtract a larger chunk of our
		 * accumulated delay.
		 */
		if (cur_use < blkg->last_use)
			sub = max_t(u64, sub, blkg->last_delay >> 1);

		/*
		 * This shouldn't happen, but handle it anyway.  Our delay_nsec
		 * should only ever be growing except here where we subtract out
		 * min(last_delay, 1 second), but lord knows bugs happen and I'd
		 * rather not end up with negative numbers.
		 */
		if (unlikely(cur < sub)) {
			atomic64_set(&blkg->delay_nsec, 0);
			blkg->last_delay = 0;
		} else {
			atomic64_sub(sub, &blkg->delay_nsec);
			blkg->last_delay = cur - sub;
		}
		blkg->last_use = cur_use;
	}
}

/*
 * This is called when we want to actually walk up the hierarchy and check to
 * see if we need to throttle, and then actually throttle if there is some
 * accumulated delay.  This should only be called upon return to user space so
 * we're not holding some lock that would induce a priority inversion.
 */
static void blkcg_maybe_throttle_blkg(struct blkcg_gq *blkg, bool use_memdelay)
{
	u64 now = ktime_to_ns(ktime_get());
	u64 exp;
	u64 delay_nsec = 0;
	int tok;

	while (blkg->parent) {
		if (atomic_read(&blkg->use_delay)) {
			blkcg_scale_delay(blkg, now);
			delay_nsec = max_t(u64, delay_nsec,
					   atomic64_read(&blkg->delay_nsec));
		}
		blkg = blkg->parent;
	}

	if (!delay_nsec)
		return;

	/*
	 * Let's not sleep for all eternity if we've amassed a huge delay.
	 * Swapping or metadata IO can accumulate 10's of seconds worth of
	 * delay, and we want userspace to be able to do _something_ so cap the
	 * delays at 1 second.  If there's 10's of seconds worth of delay then
	 * the tasks will be delayed for 1 second for every syscall.
	 */
	delay_nsec = min_t(u64, delay_nsec, 250 * NSEC_PER_MSEC);

	/*
	 * TODO: the use_memdelay flag is going to be for the upcoming psi stuff
	 * that hasn't landed upstream yet.  Once that stuff is in place we need
	 * to do a psi_memstall_enter/leave if memdelay is set.
	 */

	exp = ktime_add_ns(now, delay_nsec);
	tok = io_schedule_prepare();
	do {
		__set_current_state(TASK_KILLABLE);
		if (!schedule_hrtimeout(&exp, HRTIMER_MODE_ABS))
			break;
	} while (!fatal_signal_pending(current));
	io_schedule_finish(tok);
}

/**
 * blkcg_maybe_throttle_current - throttle the current task if it has been marked
 *
 * This is only called if we've been marked with set_notify_resume().  Obviously
 * we can be set_notify_resume() for reasons other than blkcg throttling, so we
 * check to see if current->throttle_queue is set and if not this doesn't do
 * anything.  This should only ever be called by the resume code, it's not meant
 * to be called by people willy-nilly as it will actually do the work to
 * throttle the task if it is setup for throttling.
 */
void blkcg_maybe_throttle_current(void)
{
	struct request_queue *q = current->throttle_queue;
	struct cgroup_subsys_state *css;
	struct blkcg *blkcg;
	struct blkcg_gq *blkg;
	bool use_memdelay = current->use_memdelay;

	if (!q)
		return;

	current->throttle_queue = NULL;
	current->use_memdelay = false;

	rcu_read_lock();
	css = kthread_blkcg();
	if (css)
		blkcg = css_to_blkcg(css);
	else
		blkcg = css_to_blkcg(task_css(current, io_cgrp_id));

	if (!blkcg)
		goto out;
	blkg = blkg_lookup(blkcg, q);
	if (!blkg)
		goto out;
	blkg = blkg_try_get(blkg);
	if (!blkg)
		goto out;
	rcu_read_unlock();

	blkcg_maybe_throttle_blkg(blkg, use_memdelay);
	blkg_put(blkg);
1719
	blk_put_queue(q);
1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 1775
	return;
out:
	rcu_read_unlock();
	blk_put_queue(q);
}
EXPORT_SYMBOL_GPL(blkcg_maybe_throttle_current);

/**
 * blkcg_schedule_throttle - this task needs to check for throttling
 * @q - the request queue IO was submitted on
 * @use_memdelay - do we charge this to memory delay for PSI
 *
 * This is called by the IO controller when we know there's delay accumulated
 * for the blkg for this task.  We do not pass the blkg because there are places
 * we call this that may not have that information, the swapping code for
 * instance will only have a request_queue at that point.  This set's the
 * notify_resume for the task to check and see if it requires throttling before
 * returning to user space.
 *
 * We will only schedule once per syscall.  You can call this over and over
 * again and it will only do the check once upon return to user space, and only
 * throttle once.  If the task needs to be throttled again it'll need to be
 * re-set at the next time we see the task.
 */
void blkcg_schedule_throttle(struct request_queue *q, bool use_memdelay)
{
	if (unlikely(current->flags & PF_KTHREAD))
		return;

	if (!blk_get_queue(q))
		return;

	if (current->throttle_queue)
		blk_put_queue(current->throttle_queue);
	current->throttle_queue = q;
	if (use_memdelay)
		current->use_memdelay = use_memdelay;
	set_notify_resume(current);
}
EXPORT_SYMBOL_GPL(blkcg_schedule_throttle);

/**
 * blkcg_add_delay - add delay to this blkg
 * @now - the current time in nanoseconds
 * @delta - how many nanoseconds of delay to add
 *
 * Charge @delta to the blkg's current delay accumulation.  This is used to
 * throttle tasks if an IO controller thinks we need more throttling.
 */
void blkcg_add_delay(struct blkcg_gq *blkg, u64 now, u64 delta)
{
	blkcg_scale_delay(blkg, now);
	atomic64_add(delta, &blkg->delay_nsec);
}
EXPORT_SYMBOL_GPL(blkcg_add_delay);

1776 1777
module_param(blkcg_debug_stats, bool, 0644);
MODULE_PARM_DESC(blkcg_debug_stats, "True if you want debug stats, false if not");