[linux-2.6-block.git] / drivers / md / bcache / io.c

/*
 * Some low level IO code, and hacks for various block layer limitations
 *
 * Copyright 2010, 2011 Kent Overstreet <kent.overstreet@gmail.com>
 * Copyright 2012 Google, Inc.
 */

#include "bcache.h"
#include "bset.h"
#include "debug.h"

#include <linux/blkdev.h>

static unsigned bch_bio_max_sectors(struct bio *bio)
{
	struct request_queue *q = bdev_get_queue(bio->bi_bdev);
	struct bio_vec bv;
	struct bvec_iter iter;
	unsigned ret = 0, seg = 0;

	if (bio->bi_rw & REQ_DISCARD)
		return min(bio_sectors(bio), q->limits.max_discard_sectors);

	bio_for_each_segment(bv, bio, iter) {
		struct bvec_merge_data bvm = {
			.bi_bdev	= bio->bi_bdev,
			.bi_sector	= bio->bi_iter.bi_sector,
			.bi_size	= ret << 9,
			.bi_rw		= bio->bi_rw,
		};

		if (seg == min_t(unsigned, BIO_MAX_PAGES,
				 queue_max_segments(q)))
			break;

		if (q->merge_bvec_fn &&
		    q->merge_bvec_fn(q, &bvm, &bv) < (int) bv.bv_len)
			break;

		seg++;
		ret += bv.bv_len >> 9;
	}

	ret = min(ret, queue_max_sectors(q));

	WARN_ON(!ret);
	ret = max_t(int, ret, bio_iovec(bio).bv_len >> 9);

	return ret;
}

static void bch_bio_submit_split_done(struct closure *cl)
{
	struct bio_split_hook *s = container_of(cl, struct bio_split_hook, cl);

	s->bio->bi_end_io = s->bi_end_io;
	s->bio->bi_private = s->bi_private;
	bio_endio(s->bio, 0);

	closure_debug_destroy(&s->cl);
	mempool_free(s, s->p->bio_split_hook);
}

static void bch_bio_submit_split_endio(struct bio *bio, int error)
{
	struct closure *cl = bio->bi_private;
	struct bio_split_hook *s = container_of(cl, struct bio_split_hook, cl);

	if (error)
		clear_bit(BIO_UPTODATE, &s->bio->bi_flags);

	bio_put(bio);
	closure_put(cl);
}

void bch_generic_make_request(struct bio *bio, struct bio_split_pool *p)
{
	struct bio_split_hook *s;
	struct bio *n;

	if (!bio_has_data(bio) && !(bio->bi_rw & REQ_DISCARD))
		goto submit;

	if (bio_sectors(bio) <= bch_bio_max_sectors(bio))
		goto submit;

	s = mempool_alloc(p->bio_split_hook, GFP_NOIO);
	closure_init(&s->cl, NULL);

	s->bio		= bio;
	s->p		= p;
	s->bi_end_io	= bio->bi_end_io;
	s->bi_private	= bio->bi_private;
	bio_get(bio);

	do {
		n = bio_next_split(bio, bch_bio_max_sectors(bio),
				   GFP_NOIO, s->p->bio_split);

		n->bi_end_io	= bch_bio_submit_split_endio;
		n->bi_private	= &s->cl;

		closure_get(&s->cl);
		generic_make_request(n);
	} while (n != bio);

	continue_at(&s->cl, bch_bio_submit_split_done, NULL);
	return;
submit:
	generic_make_request(bio);
}

/* Bios with headers */

void bch_bbio_free(struct bio *bio, struct cache_set *c)
{
	struct bbio *b = container_of(bio, struct bbio, bio);
	mempool_free(b, c->bio_meta);
}

struct bio *bch_bbio_alloc(struct cache_set *c)
{
	struct bbio *b = mempool_alloc(c->bio_meta, GFP_NOIO);
	struct bio *bio = &b->bio;

	bio_init(bio);
	bio->bi_flags		|= BIO_POOL_NONE << BIO_POOL_OFFSET;
	bio->bi_max_vecs	 = bucket_pages(c);
	bio->bi_io_vec		 = bio->bi_inline_vecs;

	return bio;
}

void __bch_submit_bbio(struct bio *bio, struct cache_set *c)
{
	struct bbio *b = container_of(bio, struct bbio, bio);

	bio->bi_iter.bi_sector	= PTR_OFFSET(&b->key, 0);
	bio->bi_bdev		= PTR_CACHE(c, &b->key, 0)->bdev;

	b->submit_time_us = local_clock_us();
	closure_bio_submit(bio, bio->bi_private, PTR_CACHE(c, &b->key, 0));
}

void bch_submit_bbio(struct bio *bio, struct cache_set *c,
		     struct bkey *k, unsigned ptr)
{
	struct bbio *b = container_of(bio, struct bbio, bio);
	bch_bkey_copy_single_ptr(&b->key, k, ptr);
	__bch_submit_bbio(bio, c);
}

/* IO errors */

void bch_count_io_errors(struct cache *ca, int error, const char *m)
{
	/*
	 * The halflife of an error is:
	 * log2(1/2)/log2(127/128) * refresh ~= 88 * refresh
	 */

	if (ca->set->error_decay) {
		unsigned count = atomic_inc_return(&ca->io_count);

		while (count > ca->set->error_decay) {
			unsigned errors;
			unsigned old = count;
			unsigned new = count - ca->set->error_decay;

			/*
			 * First we subtract refresh from count; each time we
			 * succesfully do so, we rescale the errors once:
			 */

			count = atomic_cmpxchg(&ca->io_count, old, new);

			if (count == old) {
				count = new;

				errors = atomic_read(&ca->io_errors);
				do {
					old = errors;
					new = ((uint64_t) errors * 127) / 128;
					errors = atomic_cmpxchg(&ca->io_errors,
								old, new);
				} while (old != errors);
			}
		}
	}

	if (error) {
		char buf[BDEVNAME_SIZE];
		unsigned errors = atomic_add_return(1 << IO_ERROR_SHIFT,
						    &ca->io_errors);
		errors >>= IO_ERROR_SHIFT;

		if (errors < ca->set->error_limit)
			pr_err("%s: IO error on %s, recovering",
			       bdevname(ca->bdev, buf), m);
		else
			bch_cache_set_error(ca->set,
					    "%s: too many IO errors %s",
					    bdevname(ca->bdev, buf), m);
	}
}

void bch_bbio_count_io_errors(struct cache_set *c, struct bio *bio,
			      int error, const char *m)
{
	struct bbio *b = container_of(bio, struct bbio, bio);
	struct cache *ca = PTR_CACHE(c, &b->key, 0);

	unsigned threshold = bio->bi_rw & REQ_WRITE
		? c->congested_write_threshold_us
		: c->congested_read_threshold_us;

	if (threshold) {
		unsigned t = local_clock_us();

		int us = t - b->submit_time_us;
		int congested = atomic_read(&c->congested);

		if (us > (int) threshold) {
			int ms = us / 1024;
			c->congested_last_us = t;

			ms = min(ms, CONGESTED_MAX + congested);
			atomic_sub(ms, &c->congested);
		} else if (congested < 0)
			atomic_inc(&c->congested);
	}

	bch_count_io_errors(ca, error, m);
}

void bch_bbio_endio(struct cache_set *c, struct bio *bio,
		    int error, const char *m)
{
	struct closure *cl = bio->bi_private;

	bch_bbio_count_io_errors(c, bio, error, m);
	bio_put(bio);
	closure_put(cl);
}
Commit	Line	Data
cafe5635 KO	1	/*
	2	* Some low level IO code, and hacks for various block layer limitations
	3	*
	4	* Copyright 2010, 2011 Kent Overstreet <kent.overstreet@gmail.com>
	5	* Copyright 2012 Google, Inc.
	6	*/
	7
	8	#include "bcache.h"
	9	#include "bset.h"
	10	#include "debug.h"
	11
c37511b8 KO	12	#include <linux/blkdev.h>
c37511b8 KO	13
cafe5635 KO	14	static unsigned bch_bio_max_sectors(struct bio *bio)
cafe5635 KO	15	{
cafe5635	16	struct request_queue *q = bdev_get_queue(bio->bi_bdev);
458b76ed KO	17	struct bio_vec bv;
	18	struct bvec_iter iter;
	19	unsigned ret = 0, seg = 0;
cafe5635	20
cafe5635	21	if (bio->bi_rw & REQ_DISCARD)
458b76ed	22	return min(bio_sectors(bio), q->limits.max_discard_sectors);
cafe5635	23
458b76ed KO	24	bio_for_each_segment(bv, bio, iter) {
	25	struct bvec_merge_data bvm = {
	26	.bi_bdev = bio->bi_bdev,
	27	.bi_sector = bio->bi_iter.bi_sector,
	28	.bi_size = ret << 9,
	29	.bi_rw = bio->bi_rw,
	30	};
	31
	32	if (seg == min_t(unsigned, BIO_MAX_PAGES,
	33	queue_max_segments(q)))
	34	break;
8e51e414	35
458b76ed KO	36	if (q->merge_bvec_fn &&
	37	q->merge_bvec_fn(q, &bvm, &bv) < (int) bv.bv_len)
	38	break;
cafe5635	39
458b76ed KO	40	seg++;
458b76ed KO	41	ret += bv.bv_len >> 9;
cafe5635 KO	42	}
	43
	44	ret = min(ret, queue_max_sectors(q));
	45
	46	WARN_ON(!ret);
a4ad39b1	47	ret = max_t(int, ret, bio_iovec(bio).bv_len >> 9);
cafe5635 KO	48
	49	return ret;
	50	}
	51
	52	static void bch_bio_submit_split_done(struct closure *cl)
	53	{
	54	struct bio_split_hook *s = container_of(cl, struct bio_split_hook, cl);
	55
	56	s->bio->bi_end_io = s->bi_end_io;
	57	s->bio->bi_private = s->bi_private;
326e1dbb	58	bio_endio(s->bio, 0);
cafe5635 KO	59
	60	closure_debug_destroy(&s->cl);
	61	mempool_free(s, s->p->bio_split_hook);
	62	}
	63
	64	static void bch_bio_submit_split_endio(struct bio *bio, int error)
	65	{
	66	struct closure *cl = bio->bi_private;
	67	struct bio_split_hook *s = container_of(cl, struct bio_split_hook, cl);
	68
	69	if (error)
	70	clear_bit(BIO_UPTODATE, &s->bio->bi_flags);
	71
	72	bio_put(bio);
	73	closure_put(cl);
	74	}
	75
cafe5635 KO	76	void bch_generic_make_request(struct bio bio, struct bio_split_pool p)
	77	{
	78	struct bio_split_hook *s;
8e51e414	79	struct bio *n;
cafe5635 KO	80
	81	if (!bio_has_data(bio) && !(bio->bi_rw & REQ_DISCARD))
	82	goto submit;
	83
	84	if (bio_sectors(bio) <= bch_bio_max_sectors(bio))
	85	goto submit;
	86
	87	s = mempool_alloc(p->bio_split_hook, GFP_NOIO);
8e51e414	88	closure_init(&s->cl, NULL);
cafe5635 KO	89
	90	s->bio = bio;
	91	s->p = p;
	92	s->bi_end_io = bio->bi_end_io;
	93	s->bi_private = bio->bi_private;
	94	bio_get(bio);
	95
8e51e414	96	do {
20d0189b KO	97	n = bio_next_split(bio, bch_bio_max_sectors(bio),
20d0189b KO	98	GFP_NOIO, s->p->bio_split);
8e51e414 KO	99
	100	n->bi_end_io = bch_bio_submit_split_endio;
	101	n->bi_private = &s->cl;
	102
	103	closure_get(&s->cl);
e90abc8e	104	generic_make_request(n);
8e51e414 KO	105	} while (n != bio);
	106
	107	continue_at(&s->cl, bch_bio_submit_split_done, NULL);
77b5a084	108	return;
cafe5635	109	submit:
e90abc8e	110	generic_make_request(bio);
cafe5635 KO	111	}
	112
	113	/* Bios with headers */
	114
	115	void bch_bbio_free(struct bio bio, struct cache_set c)
	116	{
	117	struct bbio *b = container_of(bio, struct bbio, bio);
	118	mempool_free(b, c->bio_meta);
	119	}
	120
	121	struct bio bch_bbio_alloc(struct cache_set c)
	122	{
	123	struct bbio *b = mempool_alloc(c->bio_meta, GFP_NOIO);
	124	struct bio *bio = &b->bio;
	125
	126	bio_init(bio);
	127	bio->bi_flags \|= BIO_POOL_NONE << BIO_POOL_OFFSET;
	128	bio->bi_max_vecs = bucket_pages(c);
	129	bio->bi_io_vec = bio->bi_inline_vecs;
	130
	131	return bio;
	132	}
	133
	134	void __bch_submit_bbio(struct bio bio, struct cache_set c)
	135	{
	136	struct bbio *b = container_of(bio, struct bbio, bio);
	137
4f024f37 KO	138	bio->bi_iter.bi_sector = PTR_OFFSET(&b->key, 0);
4f024f37 KO	139	bio->bi_bdev = PTR_CACHE(c, &b->key, 0)->bdev;
cafe5635 KO	140
	141	b->submit_time_us = local_clock_us();
	142	closure_bio_submit(bio, bio->bi_private, PTR_CACHE(c, &b->key, 0));
	143	}
	144
	145	void bch_submit_bbio(struct bio bio, struct cache_set c,
	146	struct bkey *k, unsigned ptr)
	147	{
	148	struct bbio *b = container_of(bio, struct bbio, bio);
	149	bch_bkey_copy_single_ptr(&b->key, k, ptr);
	150	__bch_submit_bbio(bio, c);
	151	}
	152
	153	/* IO errors */
	154
	155	void bch_count_io_errors(struct cache ca, int error, const char m)
	156	{
	157	/*
	158	* The halflife of an error is:
	159	* log2(1/2)/log2(127/128) * refresh ~= 88 * refresh
	160	*/
	161
	162	if (ca->set->error_decay) {
	163	unsigned count = atomic_inc_return(&ca->io_count);
	164
	165	while (count > ca->set->error_decay) {
	166	unsigned errors;
	167	unsigned old = count;
	168	unsigned new = count - ca->set->error_decay;
	169
	170	/*
	171	* First we subtract refresh from count; each time we
	172	* succesfully do so, we rescale the errors once:
	173	*/
	174
	175	count = atomic_cmpxchg(&ca->io_count, old, new);
	176
	177	if (count == old) {
	178	count = new;
	179
	180	errors = atomic_read(&ca->io_errors);
	181	do {
	182	old = errors;
	183	new = ((uint64_t) errors * 127) / 128;
	184	errors = atomic_cmpxchg(&ca->io_errors,
	185	old, new);
	186	} while (old != errors);
	187	}
	188	}
	189	}
	190
	191	if (error) {
	192	char buf[BDEVNAME_SIZE];
	193	unsigned errors = atomic_add_return(1 << IO_ERROR_SHIFT,
	194	&ca->io_errors);
	195	errors >>= IO_ERROR_SHIFT;
	196
	197	if (errors < ca->set->error_limit)
	198	pr_err("%s: IO error on %s, recovering",
	199	bdevname(ca->bdev, buf), m);
	200	else
	201	bch_cache_set_error(ca->set,
	202	"%s: too many IO errors %s",
	203	bdevname(ca->bdev, buf), m);
204	}
205	}
206
207	void bch_bbio_count_io_errors(struct cache_set c, struct bio bio,
208	int error, const char *m)
209	{
210	struct bbio *b = container_of(bio, struct bbio, bio);
211	struct cache *ca = PTR_CACHE(c, &b->key, 0);
212
213	unsigned threshold = bio->bi_rw & REQ_WRITE
214	? c->congested_write_threshold_us
215	: c->congested_read_threshold_us;
216
217	if (threshold) {
218	unsigned t = local_clock_us();
219
220	int us = t - b->submit_time_us;
221	int congested = atomic_read(&c->congested);
222
223	if (us > (int) threshold) {
224	int ms = us / 1024;
225	c->congested_last_us = t;
226
227	ms = min(ms, CONGESTED_MAX + congested);
228	atomic_sub(ms, &c->congested);
229	} else if (congested < 0)
230	atomic_inc(&c->congested);
231	}
232
233	bch_count_io_errors(ca, error, m);
234	}
235
236	void bch_bbio_endio(struct cache_set c, struct bio bio,
237	int error, const char *m)
238	{
239	struct closure *cl = bio->bi_private;
240
241	bch_bbio_count_io_errors(c, bio, error, m);
242	bio_put(bio);
243	closure_put(cl);
244	}