[net-next-2.6.git] / block / blk-barrier.c

/*
 * Functions related to barrier IO handling
 */
#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/bio.h>
#include <linux/blkdev.h>
#include <linux/gfp.h>

#include "blk.h"

/*
 * Cache flushing for ordered writes handling
 */
unsigned blk_ordered_cur_seq(struct request_queue *q)
{
	if (!q->ordseq)
		return 0;
	return 1 << ffz(q->ordseq);
}

unsigned blk_ordered_req_seq(struct request *rq)
{
	struct request_queue *q = rq->q;

	BUG_ON(q->ordseq == 0);

	if (rq == &q->pre_flush_rq)
		return QUEUE_ORDSEQ_PREFLUSH;
	if (rq == &q->bar_rq)
		return QUEUE_ORDSEQ_BAR;
	if (rq == &q->post_flush_rq)
		return QUEUE_ORDSEQ_POSTFLUSH;

	/*
	 * !fs requests don't need to follow barrier ordering.  Always
	 * put them at the front.  This fixes the following deadlock.
	 *
	 * http://thread.gmane.org/gmane.linux.kernel/537473
	 */
	if (rq->cmd_type != REQ_TYPE_FS)
		return QUEUE_ORDSEQ_DRAIN;

	if ((rq->cmd_flags & REQ_ORDERED_COLOR) ==
	    (q->orig_bar_rq->cmd_flags & REQ_ORDERED_COLOR))
		return QUEUE_ORDSEQ_DRAIN;
	else
		return QUEUE_ORDSEQ_DONE;
}

bool blk_ordered_complete_seq(struct request_queue *q, unsigned seq, int error)
{
	struct request *rq;

	if (error && !q->orderr)
		q->orderr = error;

	BUG_ON(q->ordseq & seq);
	q->ordseq |= seq;

	if (blk_ordered_cur_seq(q) != QUEUE_ORDSEQ_DONE)
		return false;

	/*
	 * Okay, sequence complete.
	 */
	q->ordseq = 0;
	rq = q->orig_bar_rq;
	__blk_end_request_all(rq, q->orderr);
	return true;
}

static void pre_flush_end_io(struct request *rq, int error)
{
	elv_completed_request(rq->q, rq);
	blk_ordered_complete_seq(rq->q, QUEUE_ORDSEQ_PREFLUSH, error);
}

static void bar_end_io(struct request *rq, int error)
{
	elv_completed_request(rq->q, rq);
	blk_ordered_complete_seq(rq->q, QUEUE_ORDSEQ_BAR, error);
}

static void post_flush_end_io(struct request *rq, int error)
{
	elv_completed_request(rq->q, rq);
	blk_ordered_complete_seq(rq->q, QUEUE_ORDSEQ_POSTFLUSH, error);
}

static void queue_flush(struct request_queue *q, unsigned which)
{
	struct request *rq;
	rq_end_io_fn *end_io;

	if (which == QUEUE_ORDERED_DO_PREFLUSH) {
		rq = &q->pre_flush_rq;
		end_io = pre_flush_end_io;
	} else {
		rq = &q->post_flush_rq;
		end_io = post_flush_end_io;
	}

	blk_rq_init(q, rq);
	rq->cmd_type = REQ_TYPE_FS;
	rq->cmd_flags = REQ_HARDBARRIER | REQ_FLUSH;
	rq->rq_disk = q->orig_bar_rq->rq_disk;
	rq->end_io = end_io;

	elv_insert(q, rq, ELEVATOR_INSERT_FRONT);
}

static inline struct request *start_ordered(struct request_queue *q,
					    struct request *rq)
{
	unsigned skip = 0;

	q->orderr = 0;
	q->ordered = q->next_ordered;
	q->ordseq |= QUEUE_ORDSEQ_STARTED;

	/*
	 * For an empty barrier, there's no actual BAR request, which
	 * in turn makes POSTFLUSH unnecessary.  Mask them off.
	 */
	if (!blk_rq_sectors(rq))
		q->ordered &= ~(QUEUE_ORDERED_DO_BAR |
				QUEUE_ORDERED_DO_POSTFLUSH);

	/* stash away the original request */
	blk_dequeue_request(rq);
	q->orig_bar_rq = rq;
	rq = NULL;

	/*
	 * Queue ordered sequence.  As we stack them at the head, we
	 * need to queue in reverse order.  Note that we rely on that
	 * no fs request uses ELEVATOR_INSERT_FRONT and thus no fs
	 * request gets inbetween ordered sequence.
	 */
	if (q->ordered & QUEUE_ORDERED_DO_POSTFLUSH) {
		queue_flush(q, QUEUE_ORDERED_DO_POSTFLUSH);
		rq = &q->post_flush_rq;
	} else
		skip |= QUEUE_ORDSEQ_POSTFLUSH;

	if (q->ordered & QUEUE_ORDERED_DO_BAR) {
		rq = &q->bar_rq;

		/* initialize proxy request and queue it */
		blk_rq_init(q, rq);
		init_request_from_bio(rq, q->orig_bar_rq->bio);
		if (q->ordered & QUEUE_ORDERED_DO_FUA)
			rq->cmd_flags |= REQ_FUA;
		rq->end_io = bar_end_io;

		elv_insert(q, rq, ELEVATOR_INSERT_FRONT);
	} else
		skip |= QUEUE_ORDSEQ_BAR;

	if (q->ordered & QUEUE_ORDERED_DO_PREFLUSH) {
		queue_flush(q, QUEUE_ORDERED_DO_PREFLUSH);
		rq = &q->pre_flush_rq;
	} else
		skip |= QUEUE_ORDSEQ_PREFLUSH;

	if (queue_in_flight(q))
		rq = NULL;
	else
		skip |= QUEUE_ORDSEQ_DRAIN;

	/*
	 * Complete skipped sequences.  If whole sequence is complete,
	 * return %NULL to tell elevator that this request is gone.
	 */
	if (blk_ordered_complete_seq(q, skip, 0))
		rq = NULL;
	return rq;
}

struct request *blk_do_ordered(struct request_queue *q, struct request *rq)
{
	const int is_barrier = rq->cmd_type == REQ_TYPE_FS &&
				(rq->cmd_flags & REQ_HARDBARRIER);

	if (!q->ordseq) {
		if (!is_barrier)
			return rq;

		if (q->next_ordered != QUEUE_ORDERED_NONE)
			return start_ordered(q, rq);
		else {
			/*
			 * Queue ordering not supported.  Terminate
			 * with prejudice.
			 */
			blk_dequeue_request(rq);
			__blk_end_request_all(rq, -EOPNOTSUPP);
			return NULL;
		}
	}

	/*
	 * Ordered sequence in progress
	 */

	/* Special requests are not subject to ordering rules. */
	if (rq->cmd_type != REQ_TYPE_FS &&
	    rq != &q->pre_flush_rq && rq != &q->post_flush_rq)
		return rq;

	/* Ordered by draining.  Wait for turn. */
	WARN_ON(blk_ordered_req_seq(rq) < blk_ordered_cur_seq(q));
	if (blk_ordered_req_seq(rq) > blk_ordered_cur_seq(q))
		rq = ERR_PTR(-EAGAIN);

	return rq;
}

static void bio_end_empty_barrier(struct bio *bio, int err)
{
	if (err) {
		if (err == -EOPNOTSUPP)
			set_bit(BIO_EOPNOTSUPP, &bio->bi_flags);
		clear_bit(BIO_UPTODATE, &bio->bi_flags);
	}
	if (bio->bi_private)
		complete(bio->bi_private);
	bio_put(bio);
}

/**
 * blkdev_issue_flush - queue a flush
 * @bdev:	blockdev to issue flush for
 * @gfp_mask:	memory allocation flags (for bio_alloc)
 * @error_sector:	error sector
 * @flags:	BLKDEV_IFL_* flags to control behaviour
 *
 * Description:
 *    Issue a flush for the block device in question. Caller can supply
 *    room for storing the error offset in case of a flush error, if they
 *    wish to. If WAIT flag is not passed then caller may check only what
 *    request was pushed in some internal queue for later handling.
 */
int blkdev_issue_flush(struct block_device *bdev, gfp_t gfp_mask,
		sector_t *error_sector, unsigned long flags)
{
	DECLARE_COMPLETION_ONSTACK(wait);
	struct request_queue *q;
	struct bio *bio;
	int ret = 0;

	if (bdev->bd_disk == NULL)
		return -ENXIO;

	q = bdev_get_queue(bdev);
	if (!q)
		return -ENXIO;

	/*
	 * some block devices may not have their queue correctly set up here
	 * (e.g. loop device without a backing file) and so issuing a flush
	 * here will panic. Ensure there is a request function before issuing
	 * the barrier.
	 */
	if (!q->make_request_fn)
		return -ENXIO;

	bio = bio_alloc(gfp_mask, 0);
	bio->bi_end_io = bio_end_empty_barrier;
	bio->bi_bdev = bdev;
	if (test_bit(BLKDEV_WAIT, &flags))
		bio->bi_private = &wait;

	bio_get(bio);
	submit_bio(WRITE_BARRIER, bio);
	if (test_bit(BLKDEV_WAIT, &flags)) {
		wait_for_completion(&wait);
		/*
		 * The driver must store the error location in ->bi_sector, if
		 * it supports it. For non-stacked drivers, this should be
		 * copied from blk_rq_pos(rq).
		 */
		if (error_sector)
			*error_sector = bio->bi_sector;
	}

	if (bio_flagged(bio, BIO_EOPNOTSUPP))
		ret = -EOPNOTSUPP;
	else if (!bio_flagged(bio, BIO_UPTODATE))
		ret = -EIO;

	bio_put(bio);
	return ret;
}
EXPORT_SYMBOL(blkdev_issue_flush);
Commit	Line	Data
86db1e29 JA	1	/*
	2	* Functions related to barrier IO handling
	3	*/
	4	#include <linux/kernel.h>
	5	#include <linux/module.h>
	6	#include <linux/bio.h>
	7	#include <linux/blkdev.h>
5a0e3ad6	8	#include <linux/gfp.h>
86db1e29 JA	9
	10	#include "blk.h"
	11
86db1e29 JA	12	/*
	13	* Cache flushing for ordered writes handling
	14	*/
6f6a036e	15	unsigned blk_ordered_cur_seq(struct request_queue *q)
86db1e29 JA	16	{
	17	if (!q->ordseq)
	18	return 0;
	19	return 1 << ffz(q->ordseq);
	20	}
	21
	22	unsigned blk_ordered_req_seq(struct request *rq)
	23	{
	24	struct request_queue *q = rq->q;
	25
	26	BUG_ON(q->ordseq == 0);
	27
	28	if (rq == &q->pre_flush_rq)
	29	return QUEUE_ORDSEQ_PREFLUSH;
	30	if (rq == &q->bar_rq)
	31	return QUEUE_ORDSEQ_BAR;
	32	if (rq == &q->post_flush_rq)
	33	return QUEUE_ORDSEQ_POSTFLUSH;
	34
	35	/*
	36	* !fs requests don't need to follow barrier ordering. Always
	37	* put them at the front. This fixes the following deadlock.
	38	*
	39	* http://thread.gmane.org/gmane.linux.kernel/537473
	40	*/
33659ebb	41	if (rq->cmd_type != REQ_TYPE_FS)
86db1e29 JA	42	return QUEUE_ORDSEQ_DRAIN;
	43
	44	if ((rq->cmd_flags & REQ_ORDERED_COLOR) ==
	45	(q->orig_bar_rq->cmd_flags & REQ_ORDERED_COLOR))
	46	return QUEUE_ORDSEQ_DRAIN;
	47	else
	48	return QUEUE_ORDSEQ_DONE;
	49	}
	50
8f11b3e9	51	bool blk_ordered_complete_seq(struct request_queue *q, unsigned seq, int error)
86db1e29 JA	52	{
	53	struct request *rq;
	54
	55	if (error && !q->orderr)
	56	q->orderr = error;
	57
	58	BUG_ON(q->ordseq & seq);
	59	q->ordseq \|= seq;
	60
	61	if (blk_ordered_cur_seq(q) != QUEUE_ORDSEQ_DONE)
8f11b3e9	62	return false;
86db1e29 JA	63
	64	/*
	65	* Okay, sequence complete.
	66	*/
	67	q->ordseq = 0;
	68	rq = q->orig_bar_rq;
40cbbb78	69	__blk_end_request_all(rq, q->orderr);
8f11b3e9	70	return true;
86db1e29 JA	71	}
	72
	73	static void pre_flush_end_io(struct request *rq, int error)
	74	{
	75	elv_completed_request(rq->q, rq);
	76	blk_ordered_complete_seq(rq->q, QUEUE_ORDSEQ_PREFLUSH, error);
	77	}
	78
	79	static void bar_end_io(struct request *rq, int error)
	80	{
	81	elv_completed_request(rq->q, rq);
	82	blk_ordered_complete_seq(rq->q, QUEUE_ORDSEQ_BAR, error);
	83	}
	84
	85	static void post_flush_end_io(struct request *rq, int error)
	86	{
	87	elv_completed_request(rq->q, rq);
	88	blk_ordered_complete_seq(rq->q, QUEUE_ORDSEQ_POSTFLUSH, error);
	89	}
	90
	91	static void queue_flush(struct request_queue *q, unsigned which)
	92	{
	93	struct request *rq;
	94	rq_end_io_fn *end_io;
	95
313e4299	96	if (which == QUEUE_ORDERED_DO_PREFLUSH) {
86db1e29 JA	97	rq = &q->pre_flush_rq;
	98	end_io = pre_flush_end_io;
	99	} else {
	100	rq = &q->post_flush_rq;
	101	end_io = post_flush_end_io;
	102	}
	103
2a4aa30c	104	blk_rq_init(q, rq);
28e18d01	105	rq->cmd_type = REQ_TYPE_FS;
8749534f	106	rq->cmd_flags = REQ_HARDBARRIER \| REQ_FLUSH;
16f2319f	107	rq->rq_disk = q->orig_bar_rq->rq_disk;
86db1e29	108	rq->end_io = end_io;
86db1e29 JA	109
	110	elv_insert(q, rq, ELEVATOR_INSERT_FRONT);
	111	}
	112
dd831006 TH	113	static inline struct request start_ordered(struct request_queue q,
dd831006 TH	114	struct request *rq)
86db1e29	115	{
8f11b3e9 TH	116	unsigned skip = 0;
8f11b3e9 TH	117
86db1e29 JA	118	q->orderr = 0;
	119	q->ordered = q->next_ordered;
	120	q->ordseq \|= QUEUE_ORDSEQ_STARTED;
	121
58eea927 TH	122	/*
	123	* For an empty barrier, there's no actual BAR request, which
	124	* in turn makes POSTFLUSH unnecessary. Mask them off.
	125	*/
6958f145	126	if (!blk_rq_sectors(rq))
58eea927 TH	127	q->ordered &= ~(QUEUE_ORDERED_DO_BAR \|
	128	QUEUE_ORDERED_DO_POSTFLUSH);
	129
f671620e	130	/* stash away the original request */
9934c8c0	131	blk_dequeue_request(rq);
86db1e29	132	q->orig_bar_rq = rq;
f671620e	133	rq = NULL;
86db1e29 JA	134
	135	/*
	136	* Queue ordered sequence. As we stack them at the head, we
	137	* need to queue in reverse order. Note that we rely on that
	138	* no fs request uses ELEVATOR_INSERT_FRONT and thus no fs
58eea927	139	* request gets inbetween ordered sequence.
86db1e29	140	*/
58eea927	141	if (q->ordered & QUEUE_ORDERED_DO_POSTFLUSH) {
313e4299	142	queue_flush(q, QUEUE_ORDERED_DO_POSTFLUSH);
f671620e TH	143	rq = &q->post_flush_rq;
f671620e TH	144	} else
8f11b3e9	145	skip \|= QUEUE_ORDSEQ_POSTFLUSH;
86db1e29	146
f671620e TH	147	if (q->ordered & QUEUE_ORDERED_DO_BAR) {
	148	rq = &q->bar_rq;
	149
	150	/* initialize proxy request and queue it */
	151	blk_rq_init(q, rq);
dd831006	152	init_request_from_bio(rq, q->orig_bar_rq->bio);
f671620e TH	153	if (q->ordered & QUEUE_ORDERED_DO_FUA)
f671620e TH	154	rq->cmd_flags \|= REQ_FUA;
f671620e TH	155	rq->end_io = bar_end_io;
	156
	157	elv_insert(q, rq, ELEVATOR_INSERT_FRONT);
	158	} else
8f11b3e9	159	skip \|= QUEUE_ORDSEQ_BAR;
86db1e29	160
313e4299 TH	161	if (q->ordered & QUEUE_ORDERED_DO_PREFLUSH) {
313e4299 TH	162	queue_flush(q, QUEUE_ORDERED_DO_PREFLUSH);
86db1e29 JA	163	rq = &q->pre_flush_rq;
86db1e29 JA	164	} else
8f11b3e9	165	skip \|= QUEUE_ORDSEQ_PREFLUSH;
86db1e29	166
6958f145	167	if (queue_in_flight(q))
86db1e29	168	rq = NULL;
f671620e	169	else
8f11b3e9	170	skip \|= QUEUE_ORDSEQ_DRAIN;
86db1e29	171
8f11b3e9 TH	172	/*
8f11b3e9 TH	173	* Complete skipped sequences. If whole sequence is complete,
dd831006	174	* return %NULL to tell elevator that this request is gone.
8f11b3e9	175	*/
dd831006 TH	176	if (blk_ordered_complete_seq(q, skip, 0))
	177	rq = NULL;
	178	return rq;
86db1e29 JA	179	}
86db1e29 JA	180
dd831006	181	struct request blk_do_ordered(struct request_queue q, struct request *rq)
86db1e29	182	{
33659ebb CH	183	const int is_barrier = rq->cmd_type == REQ_TYPE_FS &&
33659ebb CH	184	(rq->cmd_flags & REQ_HARDBARRIER);
86db1e29 JA	185
	186	if (!q->ordseq) {
	187	if (!is_barrier)
dd831006	188	return rq;
86db1e29	189
8f11b3e9	190	if (q->next_ordered != QUEUE_ORDERED_NONE)
dd831006	191	return start_ordered(q, rq);
8f11b3e9	192	else {
86db1e29	193	/*
a7384677 TH	194	* Queue ordering not supported. Terminate
a7384677 TH	195	* with prejudice.
86db1e29	196	*/
9934c8c0	197	blk_dequeue_request(rq);
40cbbb78	198	__blk_end_request_all(rq, -EOPNOTSUPP);
dd831006	199	return NULL;
86db1e29 JA	200	}
	201	}
	202
	203	/*
	204	* Ordered sequence in progress
	205	*/
	206
	207	/* Special requests are not subject to ordering rules. */
33659ebb	208	if (rq->cmd_type != REQ_TYPE_FS &&
86db1e29	209	rq != &q->pre_flush_rq && rq != &q->post_flush_rq)
dd831006	210	return rq;
86db1e29	211
6958f145 TH	212	/* Ordered by draining. Wait for turn. */
	213	WARN_ON(blk_ordered_req_seq(rq) < blk_ordered_cur_seq(q));
	214	if (blk_ordered_req_seq(rq) > blk_ordered_cur_seq(q))
dd831006	215	rq = ERR_PTR(-EAGAIN);
86db1e29	216
dd831006	217	return rq;
86db1e29 JA	218	}
	219
	220	static void bio_end_empty_barrier(struct bio *bio, int err)
	221	{
cc66b451 JA	222	if (err) {
	223	if (err == -EOPNOTSUPP)
	224	set_bit(BIO_EOPNOTSUPP, &bio->bi_flags);
86db1e29	225	clear_bit(BIO_UPTODATE, &bio->bi_flags);
cc66b451	226	}
f17e232e DM	227	if (bio->bi_private)
	228	complete(bio->bi_private);
	229	bio_put(bio);
86db1e29 JA	230	}
	231
	232	/**
	233	* blkdev_issue_flush - queue a flush
	234	* @bdev: blockdev to issue flush for
fbd9b09a	235	* @gfp_mask: memory allocation flags (for bio_alloc)
86db1e29	236	* @error_sector: error sector
fbd9b09a	237	* @flags: BLKDEV_IFL_* flags to control behaviour
86db1e29 JA	238	*
	239	* Description:
	240	* Issue a flush for the block device in question. Caller can supply
	241	* room for storing the error offset in case of a flush error, if they
f17e232e DM	242	* wish to. If WAIT flag is not passed then caller may check only what
f17e232e DM	243	* request was pushed in some internal queue for later handling.
86db1e29	244	*/
fbd9b09a DM	245	int blkdev_issue_flush(struct block_device *bdev, gfp_t gfp_mask,
fbd9b09a DM	246	sector_t *error_sector, unsigned long flags)
86db1e29 JA	247	{
	248	DECLARE_COMPLETION_ONSTACK(wait);
	249	struct request_queue *q;
	250	struct bio *bio;
fbd9b09a	251	int ret = 0;
86db1e29 JA	252
	253	if (bdev->bd_disk == NULL)
	254	return -ENXIO;
	255
	256	q = bdev_get_queue(bdev);
	257	if (!q)
	258	return -ENXIO;
	259
f10d9f61 DC	260	/*
	261	* some block devices may not have their queue correctly set up here
	262	* (e.g. loop device without a backing file) and so issuing a flush
	263	* here will panic. Ensure there is a request function before issuing
	264	* the barrier.
	265	*/
	266	if (!q->make_request_fn)
	267	return -ENXIO;
	268
fbd9b09a	269	bio = bio_alloc(gfp_mask, 0);
86db1e29	270	bio->bi_end_io = bio_end_empty_barrier;
86db1e29	271	bio->bi_bdev = bdev;
f17e232e DM	272	if (test_bit(BLKDEV_WAIT, &flags))
f17e232e DM	273	bio->bi_private = &wait;
86db1e29	274
f17e232e DM	275	bio_get(bio);
	276	submit_bio(WRITE_BARRIER, bio);
	277	if (test_bit(BLKDEV_WAIT, &flags)) {
	278	wait_for_completion(&wait);
	279	/*
	280	* The driver must store the error location in ->bi_sector, if
	281	* it supports it. For non-stacked drivers, this should be
	282	* copied from blk_rq_pos(rq).
	283	*/
	284	if (error_sector)
	285	*error_sector = bio->bi_sector;
	286	}
86db1e29	287
cc66b451 JA	288	if (bio_flagged(bio, BIO_EOPNOTSUPP))
	289	ret = -EOPNOTSUPP;
	290	else if (!bio_flagged(bio, BIO_UPTODATE))
86db1e29 JA	291	ret = -EIO;
	292
	293	bio_put(bio);
	294	return ret;
	295	}
86db1e29	296	EXPORT_SYMBOL(blkdev_issue_flush);