[deliverable/linux.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>

/**
 * bch_bio_split - split a bio
 * @bio:	bio to split
 * @sectors:	number of sectors to split from the front of @bio
 * @gfp:	gfp mask
 * @bs:		bio set to allocate from
 *
 * Allocates and returns a new bio which represents @sectors from the start of
 * @bio, and updates @bio to represent the remaining sectors.
 *
 * If bio_sectors(@bio) was less than or equal to @sectors, returns @bio
 * unchanged.
 *
 * The newly allocated bio will point to @bio's bi_io_vec, if the split was on a
 * bvec boundry; it is the caller's responsibility to ensure that @bio is not
 * freed before the split.
 */
struct bio *bch_bio_split(struct bio *bio, int sectors,
			  gfp_t gfp, struct bio_set *bs)
{
	unsigned vcnt = 0, nbytes = sectors << 9;
	struct bio_vec bv;
	struct bvec_iter iter;
	struct bio *ret = NULL;

	BUG_ON(sectors <= 0);

	if (sectors >= bio_sectors(bio))
		return bio;

	if (bio->bi_rw & REQ_DISCARD) {
		ret = bio_alloc_bioset(gfp, 1, bs);
		if (!ret)
			return NULL;
		goto out;
	}

	bio_for_each_segment(bv, bio, iter) {
		vcnt++;

		if (nbytes <= bv.bv_len)
			break;

		nbytes -= bv.bv_len;
	}

	ret = bio_alloc_bioset(gfp, vcnt, bs);
	if (!ret)
		return NULL;

	bio_for_each_segment(bv, bio, iter) {
		ret->bi_io_vec[ret->bi_vcnt++] = bv;

		if (ret->bi_vcnt == vcnt)
			break;
	}

	ret->bi_io_vec[ret->bi_vcnt - 1].bv_len = nbytes;
out:
	ret->bi_bdev	= bio->bi_bdev;
	ret->bi_iter.bi_sector	= bio->bi_iter.bi_sector;
	ret->bi_iter.bi_size	= sectors << 9;
	ret->bi_rw	= bio->bi_rw;

	if (bio_integrity(bio)) {
		if (bio_integrity_clone(ret, bio, gfp)) {
			bio_put(ret);
			return NULL;
		}

		bio_integrity_trim(ret, 0, bio_sectors(ret));
	}

	bio_advance(bio, ret->bi_iter.bi_size);

	return ret;
}

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_nodec(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 = bch_bio_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);
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	/**
	15	* bch_bio_split - split a bio
	16	* @bio: bio to split
	17	* @sectors: number of sectors to split from the front of @bio
	18	* @gfp: gfp mask
	19	* @bs: bio set to allocate from
	20	*
	21	* Allocates and returns a new bio which represents @sectors from the start of
	22	* @bio, and updates @bio to represent the remaining sectors.
	23	*
	24	* If bio_sectors(@bio) was less than or equal to @sectors, returns @bio
	25	* unchanged.
	26	*
	27	* The newly allocated bio will point to @bio's bi_io_vec, if the split was on a
	28	* bvec boundry; it is the caller's responsibility to ensure that @bio is not
	29	* freed before the split.
cafe5635 KO	30	*/
	31	struct bio bch_bio_split(struct bio bio, int sectors,
	32	gfp_t gfp, struct bio_set *bs)
	33	{
7988613b KO	34	unsigned vcnt = 0, nbytes = sectors << 9;
	35	struct bio_vec bv;
	36	struct bvec_iter iter;
cafe5635 KO	37	struct bio *ret = NULL;
	38
	39	BUG_ON(sectors <= 0);
	40
cafe5635 KO	41	if (sectors >= bio_sectors(bio))
	42	return bio;
	43
	44	if (bio->bi_rw & REQ_DISCARD) {
	45	ret = bio_alloc_bioset(gfp, 1, bs);
5c694129 KAM	46	if (!ret)
5c694129 KAM	47	return NULL;
cafe5635 KO	48	goto out;
	49	}
	50
7988613b KO	51	bio_for_each_segment(bv, bio, iter) {
7988613b KO	52	vcnt++;
cafe5635	53
7988613b KO	54	if (nbytes <= bv.bv_len)
7988613b KO	55	break;
cafe5635	56
7988613b KO	57	nbytes -= bv.bv_len;
7988613b KO	58	}
cafe5635	59
7988613b KO	60	ret = bio_alloc_bioset(gfp, vcnt, bs);
	61	if (!ret)
	62	return NULL;
cafe5635	63
7988613b KO	64	bio_for_each_segment(bv, bio, iter) {
7988613b KO	65	ret->bi_io_vec[ret->bi_vcnt++] = bv;
cafe5635	66
7988613b	67	if (ret->bi_vcnt == vcnt)
cafe5635	68	break;
cafe5635	69	}
7988613b KO	70
7988613b KO	71	ret->bi_io_vec[ret->bi_vcnt - 1].bv_len = nbytes;
cafe5635 KO	72	out:
cafe5635 KO	73	ret->bi_bdev = bio->bi_bdev;
4f024f37 KO	74	ret->bi_iter.bi_sector = bio->bi_iter.bi_sector;
4f024f37 KO	75	ret->bi_iter.bi_size = sectors << 9;
cafe5635	76	ret->bi_rw = bio->bi_rw;
cafe5635 KO	77
	78	if (bio_integrity(bio)) {
	79	if (bio_integrity_clone(ret, bio, gfp)) {
	80	bio_put(ret);
	81	return NULL;
	82	}
	83
	84	bio_integrity_trim(ret, 0, bio_sectors(ret));
cafe5635 KO	85	}
cafe5635 KO	86
7988613b KO	87	bio_advance(bio, ret->bi_iter.bi_size);
7988613b KO	88
cafe5635 KO	89	return ret;
	90	}
	91
	92	static unsigned bch_bio_max_sectors(struct bio *bio)
	93	{
cafe5635	94	struct request_queue *q = bdev_get_queue(bio->bi_bdev);
458b76ed KO	95	struct bio_vec bv;
	96	struct bvec_iter iter;
	97	unsigned ret = 0, seg = 0;
cafe5635	98
cafe5635	99	if (bio->bi_rw & REQ_DISCARD)
458b76ed	100	return min(bio_sectors(bio), q->limits.max_discard_sectors);
cafe5635	101
458b76ed KO	102	bio_for_each_segment(bv, bio, iter) {
	103	struct bvec_merge_data bvm = {
	104	.bi_bdev = bio->bi_bdev,
	105	.bi_sector = bio->bi_iter.bi_sector,
	106	.bi_size = ret << 9,
	107	.bi_rw = bio->bi_rw,
	108	};
	109
	110	if (seg == min_t(unsigned, BIO_MAX_PAGES,
	111	queue_max_segments(q)))
	112	break;
8e51e414	113
458b76ed KO	114	if (q->merge_bvec_fn &&
	115	q->merge_bvec_fn(q, &bvm, &bv) < (int) bv.bv_len)
	116	break;
cafe5635	117
458b76ed KO	118	seg++;
458b76ed KO	119	ret += bv.bv_len >> 9;
cafe5635 KO	120	}
	121
	122	ret = min(ret, queue_max_sectors(q));
	123
	124	WARN_ON(!ret);
a4ad39b1	125	ret = max_t(int, ret, bio_iovec(bio).bv_len >> 9);
cafe5635 KO	126
	127	return ret;
	128	}
	129
	130	static void bch_bio_submit_split_done(struct closure *cl)
	131	{
	132	struct bio_split_hook *s = container_of(cl, struct bio_split_hook, cl);
	133
	134	s->bio->bi_end_io = s->bi_end_io;
	135	s->bio->bi_private = s->bi_private;
196d38bc	136	bio_endio_nodec(s->bio, 0);
cafe5635 KO	137
	138	closure_debug_destroy(&s->cl);
	139	mempool_free(s, s->p->bio_split_hook);
	140	}
	141
	142	static void bch_bio_submit_split_endio(struct bio *bio, int error)
	143	{
	144	struct closure *cl = bio->bi_private;
	145	struct bio_split_hook *s = container_of(cl, struct bio_split_hook, cl);
	146
	147	if (error)
	148	clear_bit(BIO_UPTODATE, &s->bio->bi_flags);
	149
	150	bio_put(bio);
	151	closure_put(cl);
	152	}
	153
cafe5635 KO	154	void bch_generic_make_request(struct bio bio, struct bio_split_pool p)
	155	{
	156	struct bio_split_hook *s;
8e51e414	157	struct bio *n;
cafe5635 KO	158
	159	if (!bio_has_data(bio) && !(bio->bi_rw & REQ_DISCARD))
	160	goto submit;
	161
	162	if (bio_sectors(bio) <= bch_bio_max_sectors(bio))
	163	goto submit;
	164
	165	s = mempool_alloc(p->bio_split_hook, GFP_NOIO);
8e51e414	166	closure_init(&s->cl, NULL);
cafe5635 KO	167
	168	s->bio = bio;
	169	s->p = p;
	170	s->bi_end_io = bio->bi_end_io;
	171	s->bi_private = bio->bi_private;
	172	bio_get(bio);
	173
8e51e414 KO	174	do {
	175	n = bch_bio_split(bio, bch_bio_max_sectors(bio),
	176	GFP_NOIO, s->p->bio_split);
	177
	178	n->bi_end_io = bch_bio_submit_split_endio;
	179	n->bi_private = &s->cl;
	180
	181	closure_get(&s->cl);
e90abc8e	182	generic_make_request(n);
8e51e414 KO	183	} while (n != bio);
	184
	185	continue_at(&s->cl, bch_bio_submit_split_done, NULL);
cafe5635	186	submit:
e90abc8e	187	generic_make_request(bio);
cafe5635 KO	188	}
	189
	190	/* Bios with headers */
	191
	192	void bch_bbio_free(struct bio bio, struct cache_set c)
	193	{
	194	struct bbio *b = container_of(bio, struct bbio, bio);
	195	mempool_free(b, c->bio_meta);
	196	}
	197
	198	struct bio bch_bbio_alloc(struct cache_set c)
	199	{
	200	struct bbio *b = mempool_alloc(c->bio_meta, GFP_NOIO);
	201	struct bio *bio = &b->bio;
	202
	203	bio_init(bio);
	204	bio->bi_flags \|= BIO_POOL_NONE << BIO_POOL_OFFSET;
	205	bio->bi_max_vecs = bucket_pages(c);
	206	bio->bi_io_vec = bio->bi_inline_vecs;
	207
	208	return bio;
	209	}
	210
	211	void __bch_submit_bbio(struct bio bio, struct cache_set c)
	212	{
	213	struct bbio *b = container_of(bio, struct bbio, bio);
	214
4f024f37 KO	215	bio->bi_iter.bi_sector = PTR_OFFSET(&b->key, 0);
4f024f37 KO	216	bio->bi_bdev = PTR_CACHE(c, &b->key, 0)->bdev;
cafe5635 KO	217
	218	b->submit_time_us = local_clock_us();
	219	closure_bio_submit(bio, bio->bi_private, PTR_CACHE(c, &b->key, 0));
	220	}
	221
	222	void bch_submit_bbio(struct bio bio, struct cache_set c,
	223	struct bkey *k, unsigned ptr)
	224	{
	225	struct bbio *b = container_of(bio, struct bbio, bio);
	226	bch_bkey_copy_single_ptr(&b->key, k, ptr);
	227	__bch_submit_bbio(bio, c);
	228	}
	229
	230	/* IO errors */
	231
	232	void bch_count_io_errors(struct cache ca, int error, const char m)
	233	{
	234	/*
	235	* The halflife of an error is:
	236	* log2(1/2)/log2(127/128) * refresh ~= 88 * refresh
	237	*/
	238
	239	if (ca->set->error_decay) {
	240	unsigned count = atomic_inc_return(&ca->io_count);
	241
	242	while (count > ca->set->error_decay) {
	243	unsigned errors;
	244	unsigned old = count;
	245	unsigned new = count - ca->set->error_decay;
	246
	247	/*
	248	* First we subtract refresh from count; each time we
	249	* succesfully do so, we rescale the errors once:
	250	*/
	251
	252	count = atomic_cmpxchg(&ca->io_count, old, new);
	253
	254	if (count == old) {
	255	count = new;
	256
	257	errors = atomic_read(&ca->io_errors);
	258	do {
	259	old = errors;
	260	new = ((uint64_t) errors * 127) / 128;
	261	errors = atomic_cmpxchg(&ca->io_errors,
	262	old, new);
	263	} while (old != errors);
	264	}
	265	}
	266	}
	267
	268	if (error) {
	269	char buf[BDEVNAME_SIZE];
	270	unsigned errors = atomic_add_return(1 << IO_ERROR_SHIFT,
	271	&ca->io_errors);
	272	errors >>= IO_ERROR_SHIFT;
	273
	274	if (errors < ca->set->error_limit)
	275	pr_err("%s: IO error on %s, recovering",
	276	bdevname(ca->bdev, buf), m);
	277	else
	278	bch_cache_set_error(ca->set,
	279	"%s: too many IO errors %s",
	280	bdevname(ca->bdev, buf), m);
281	}
282	}
283
284	void bch_bbio_count_io_errors(struct cache_set c, struct bio bio,
285	int error, const char *m)
286	{
287	struct bbio *b = container_of(bio, struct bbio, bio);
288	struct cache *ca = PTR_CACHE(c, &b->key, 0);
289
290	unsigned threshold = bio->bi_rw & REQ_WRITE
291	? c->congested_write_threshold_us
292	: c->congested_read_threshold_us;
293
294	if (threshold) {
295	unsigned t = local_clock_us();
296
297	int us = t - b->submit_time_us;
298	int congested = atomic_read(&c->congested);
299
300	if (us > (int) threshold) {
301	int ms = us / 1024;
302	c->congested_last_us = t;
303
304	ms = min(ms, CONGESTED_MAX + congested);
305	atomic_sub(ms, &c->congested);
306	} else if (congested < 0)
307	atomic_inc(&c->congested);
308	}
309
310	bch_count_io_errors(ca, error, m);
311	}
312
313	void bch_bbio_endio(struct cache_set c, struct bio bio,
314	int error, const char *m)
315	{
316	struct closure *cl = bio->bi_private;
317
318	bch_bbio_count_io_errors(c, bio, error, m);
319	bio_put(bio);
320	closure_put(cl);
321	}