[deliverable/linux.git] / block / blk-merge.c

/*
 * Functions related to segment and merge handling
 */
#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/bio.h>
#include <linux/blkdev.h>
#include <linux/scatterlist.h>

#include "blk.h"

static unsigned int __blk_recalc_rq_segments(struct request_queue *q,
					     struct bio *bio)
{
	struct bio_vec *bv, *bvprv = NULL;
	int cluster, i, high, highprv = 1;
	unsigned int seg_size, nr_phys_segs;
	struct bio *fbio, *bbio;

	if (!bio)
		return 0;

	fbio = bio;
	cluster = blk_queue_cluster(q);
	seg_size = 0;
	nr_phys_segs = 0;
	for_each_bio(bio) {
		bio_for_each_segment(bv, bio, i) {
			/*
			 * the trick here is making sure that a high page is
			 * never considered part of another segment, since that
			 * might change with the bounce page.
			 */
			high = page_to_pfn(bv->bv_page) > queue_bounce_pfn(q);
			if (high || highprv)
				goto new_segment;
			if (cluster) {
				if (seg_size + bv->bv_len
				    > queue_max_segment_size(q))
					goto new_segment;
				if (!BIOVEC_PHYS_MERGEABLE(bvprv, bv))
					goto new_segment;
				if (!BIOVEC_SEG_BOUNDARY(q, bvprv, bv))
					goto new_segment;

				seg_size += bv->bv_len;
				bvprv = bv;
				continue;
			}
new_segment:
			if (nr_phys_segs == 1 && seg_size >
			    fbio->bi_seg_front_size)
				fbio->bi_seg_front_size = seg_size;

			nr_phys_segs++;
			bvprv = bv;
			seg_size = bv->bv_len;
			highprv = high;
		}
		bbio = bio;
	}

	if (nr_phys_segs == 1 && seg_size > fbio->bi_seg_front_size)
		fbio->bi_seg_front_size = seg_size;
	if (seg_size > bbio->bi_seg_back_size)
		bbio->bi_seg_back_size = seg_size;

	return nr_phys_segs;
}

void blk_recalc_rq_segments(struct request *rq)
{
	rq->nr_phys_segments = __blk_recalc_rq_segments(rq->q, rq->bio);
}

void blk_recount_segments(struct request_queue *q, struct bio *bio)
{
	struct bio *nxt = bio->bi_next;

	bio->bi_next = NULL;
	bio->bi_phys_segments = __blk_recalc_rq_segments(q, bio);
	bio->bi_next = nxt;
	bio->bi_flags |= (1 << BIO_SEG_VALID);
}
EXPORT_SYMBOL(blk_recount_segments);

static int blk_phys_contig_segment(struct request_queue *q, struct bio *bio,
				   struct bio *nxt)
{
	if (!blk_queue_cluster(q))
		return 0;

	if (bio->bi_seg_back_size + nxt->bi_seg_front_size >
	    queue_max_segment_size(q))
		return 0;

	if (!bio_has_data(bio))
		return 1;

	if (!BIOVEC_PHYS_MERGEABLE(__BVEC_END(bio), __BVEC_START(nxt)))
		return 0;

	/*
	 * bio and nxt are contiguous in memory; check if the queue allows
	 * these two to be merged into one
	 */
	if (BIO_SEG_BOUNDARY(q, bio, nxt))
		return 1;

	return 0;
}

static void
__blk_segment_map_sg(struct request_queue *q, struct bio_vec *bvec,
		     struct scatterlist *sglist, struct bio_vec **bvprv,
		     struct scatterlist **sg, int *nsegs, int *cluster)
{

	int nbytes = bvec->bv_len;

	if (*bvprv && *cluster) {
		if ((*sg)->length + nbytes > queue_max_segment_size(q))
			goto new_segment;

		if (!BIOVEC_PHYS_MERGEABLE(*bvprv, bvec))
			goto new_segment;
		if (!BIOVEC_SEG_BOUNDARY(q, *bvprv, bvec))
			goto new_segment;

		(*sg)->length += nbytes;
	} else {
new_segment:
		if (!*sg)
			*sg = sglist;
		else {
			/*
			 * If the driver previously mapped a shorter
			 * list, we could see a termination bit
			 * prematurely unless it fully inits the sg
			 * table on each mapping. We KNOW that there
			 * must be more entries here or the driver
			 * would be buggy, so force clear the
			 * termination bit to avoid doing a full
			 * sg_init_table() in drivers for each command.
			 */
			(*sg)->page_link &= ~0x02;
			*sg = sg_next(*sg);
		}

		sg_set_page(*sg, bvec->bv_page, nbytes, bvec->bv_offset);
		(*nsegs)++;
	}
	*bvprv = bvec;
}

/*
 * map a request to scatterlist, return number of sg entries setup. Caller
 * must make sure sg can hold rq->nr_phys_segments entries
 */
int blk_rq_map_sg(struct request_queue *q, struct request *rq,
		  struct scatterlist *sglist)
{
	struct bio_vec *bvec, *bvprv;
	struct req_iterator iter;
	struct scatterlist *sg;
	int nsegs, cluster;

	nsegs = 0;
	cluster = blk_queue_cluster(q);

	/*
	 * for each bio in rq
	 */
	bvprv = NULL;
	sg = NULL;
	rq_for_each_segment(bvec, rq, iter) {
		__blk_segment_map_sg(q, bvec, sglist, &bvprv, &sg,
				     &nsegs, &cluster);
	} /* segments in rq */


	if (unlikely(rq->cmd_flags & REQ_COPY_USER) &&
	    (blk_rq_bytes(rq) & q->dma_pad_mask)) {
		unsigned int pad_len =
			(q->dma_pad_mask & ~blk_rq_bytes(rq)) + 1;

		sg->length += pad_len;
		rq->extra_len += pad_len;
	}

	if (q->dma_drain_size && q->dma_drain_needed(rq)) {
		if (rq->cmd_flags & REQ_WRITE)
			memset(q->dma_drain_buffer, 0, q->dma_drain_size);

		sg->page_link &= ~0x02;
		sg = sg_next(sg);
		sg_set_page(sg, virt_to_page(q->dma_drain_buffer),
			    q->dma_drain_size,
			    ((unsigned long)q->dma_drain_buffer) &
			    (PAGE_SIZE - 1));
		nsegs++;
		rq->extra_len += q->dma_drain_size;
	}

	if (sg)
		sg_mark_end(sg);

	return nsegs;
}
EXPORT_SYMBOL(blk_rq_map_sg);

/**
 * blk_bio_map_sg - map a bio to a scatterlist
 * @q: request_queue in question
 * @bio: bio being mapped
 * @sglist: scatterlist being mapped
 *
 * Note:
 *    Caller must make sure sg can hold bio->bi_phys_segments entries
 *
 * Will return the number of sg entries setup
 */
int blk_bio_map_sg(struct request_queue *q, struct bio *bio,
		   struct scatterlist *sglist)
{
	struct bio_vec *bvec, *bvprv;
	struct scatterlist *sg;
	int nsegs, cluster;
	unsigned long i;

	nsegs = 0;
	cluster = blk_queue_cluster(q);

	bvprv = NULL;
	sg = NULL;
	bio_for_each_segment(bvec, bio, i) {
		__blk_segment_map_sg(q, bvec, sglist, &bvprv, &sg,
				     &nsegs, &cluster);
	} /* segments in bio */

	if (sg)
		sg_mark_end(sg);

	BUG_ON(bio->bi_phys_segments && nsegs > bio->bi_phys_segments);
	return nsegs;
}
EXPORT_SYMBOL(blk_bio_map_sg);

static inline int ll_new_hw_segment(struct request_queue *q,
				    struct request *req,
				    struct bio *bio)
{
	int nr_phys_segs = bio_phys_segments(q, bio);

	if (req->nr_phys_segments + nr_phys_segs > queue_max_segments(q))
		goto no_merge;

	if (bio_integrity(bio) && blk_integrity_merge_bio(q, req, bio))
		goto no_merge;

	/*
	 * This will form the start of a new hw segment.  Bump both
	 * counters.
	 */
	req->nr_phys_segments += nr_phys_segs;
	return 1;

no_merge:
	req->cmd_flags |= REQ_NOMERGE;
	if (req == q->last_merge)
		q->last_merge = NULL;
	return 0;
}

int ll_back_merge_fn(struct request_queue *q, struct request *req,
		     struct bio *bio)
{
	if (blk_rq_sectors(req) + bio_sectors(bio) >
	    blk_rq_get_max_sectors(req)) {
		req->cmd_flags |= REQ_NOMERGE;
		if (req == q->last_merge)
			q->last_merge = NULL;
		return 0;
	}
	if (!bio_flagged(req->biotail, BIO_SEG_VALID))
		blk_recount_segments(q, req->biotail);
	if (!bio_flagged(bio, BIO_SEG_VALID))
		blk_recount_segments(q, bio);

	return ll_new_hw_segment(q, req, bio);
}

int ll_front_merge_fn(struct request_queue *q, struct request *req,
		      struct bio *bio)
{
	if (blk_rq_sectors(req) + bio_sectors(bio) >
	    blk_rq_get_max_sectors(req)) {
		req->cmd_flags |= REQ_NOMERGE;
		if (req == q->last_merge)
			q->last_merge = NULL;
		return 0;
	}
	if (!bio_flagged(bio, BIO_SEG_VALID))
		blk_recount_segments(q, bio);
	if (!bio_flagged(req->bio, BIO_SEG_VALID))
		blk_recount_segments(q, req->bio);

	return ll_new_hw_segment(q, req, bio);
}

static int ll_merge_requests_fn(struct request_queue *q, struct request *req,
				struct request *next)
{
	int total_phys_segments;
	unsigned int seg_size =
		req->biotail->bi_seg_back_size + next->bio->bi_seg_front_size;

	/*
	 * First check if the either of the requests are re-queued
	 * requests.  Can't merge them if they are.
	 */
	if (req->special || next->special)
		return 0;

	/*
	 * Will it become too large?
	 */
	if ((blk_rq_sectors(req) + blk_rq_sectors(next)) >
	    blk_rq_get_max_sectors(req))
		return 0;

	total_phys_segments = req->nr_phys_segments + next->nr_phys_segments;
	if (blk_phys_contig_segment(q, req->biotail, next->bio)) {
		if (req->nr_phys_segments == 1)
			req->bio->bi_seg_front_size = seg_size;
		if (next->nr_phys_segments == 1)
			next->biotail->bi_seg_back_size = seg_size;
		total_phys_segments--;
	}

	if (total_phys_segments > queue_max_segments(q))
		return 0;

	if (blk_integrity_rq(req) && blk_integrity_merge_rq(q, req, next))
		return 0;

	/* Merge is OK... */
	req->nr_phys_segments = total_phys_segments;
	return 1;
}

/**
 * blk_rq_set_mixed_merge - mark a request as mixed merge
 * @rq: request to mark as mixed merge
 *
 * Description:
 *     @rq is about to be mixed merged.  Make sure the attributes
 *     which can be mixed are set in each bio and mark @rq as mixed
 *     merged.
 */
void blk_rq_set_mixed_merge(struct request *rq)
{
	unsigned int ff = rq->cmd_flags & REQ_FAILFAST_MASK;
	struct bio *bio;

	if (rq->cmd_flags & REQ_MIXED_MERGE)
		return;

	/*
	 * @rq will no longer represent mixable attributes for all the
	 * contained bios.  It will just track those of the first one.
	 * Distributes the attributs to each bio.
	 */
	for (bio = rq->bio; bio; bio = bio->bi_next) {
		WARN_ON_ONCE((bio->bi_rw & REQ_FAILFAST_MASK) &&
			     (bio->bi_rw & REQ_FAILFAST_MASK) != ff);
		bio->bi_rw |= ff;
	}
	rq->cmd_flags |= REQ_MIXED_MERGE;
}

static void blk_account_io_merge(struct request *req)
{
	if (blk_do_io_stat(req)) {
		struct hd_struct *part;
		int cpu;

		cpu = part_stat_lock();
		part = req->part;

		part_round_stats(cpu, part);
		part_dec_in_flight(part, rq_data_dir(req));

		hd_struct_put(part);
		part_stat_unlock();
	}
}

/*
 * Has to be called with the request spinlock acquired
 */
static int attempt_merge(struct request_queue *q, struct request *req,
			  struct request *next)
{
	if (!rq_mergeable(req) || !rq_mergeable(next))
		return 0;

	if (!blk_check_merge_flags(req->cmd_flags, next->cmd_flags))
		return 0;

	/*
	 * not contiguous
	 */
	if (blk_rq_pos(req) + blk_rq_sectors(req) != blk_rq_pos(next))
		return 0;

	if (rq_data_dir(req) != rq_data_dir(next)
	    || req->rq_disk != next->rq_disk
	    || next->special)
		return 0;

	if (req->cmd_flags & REQ_WRITE_SAME &&
	    !blk_write_same_mergeable(req->bio, next->bio))
		return 0;

	/*
	 * If we are allowed to merge, then append bio list
	 * from next to rq and release next. merge_requests_fn
	 * will have updated segment counts, update sector
	 * counts here.
	 */
	if (!ll_merge_requests_fn(q, req, next))
		return 0;

	/*
	 * If failfast settings disagree or any of the two is already
	 * a mixed merge, mark both as mixed before proceeding.  This
	 * makes sure that all involved bios have mixable attributes
	 * set properly.
	 */
	if ((req->cmd_flags | next->cmd_flags) & REQ_MIXED_MERGE ||
	    (req->cmd_flags & REQ_FAILFAST_MASK) !=
	    (next->cmd_flags & REQ_FAILFAST_MASK)) {
		blk_rq_set_mixed_merge(req);
		blk_rq_set_mixed_merge(next);
	}

	/*
	 * At this point we have either done a back merge
	 * or front merge. We need the smaller start_time of
	 * the merged requests to be the current request
	 * for accounting purposes.
	 */
	if (time_after(req->start_time, next->start_time))
		req->start_time = next->start_time;

	req->biotail->bi_next = next->bio;
	req->biotail = next->biotail;

	req->__data_len += blk_rq_bytes(next);

	elv_merge_requests(q, req, next);

	/*
	 * 'next' is going away, so update stats accordingly
	 */
	blk_account_io_merge(next);

	req->ioprio = ioprio_best(req->ioprio, next->ioprio);
	if (blk_rq_cpu_valid(next))
		req->cpu = next->cpu;

	/* owner-ship of bio passed from next to req */
	next->bio = NULL;
	__blk_put_request(q, next);
	return 1;
}

int attempt_back_merge(struct request_queue *q, struct request *rq)
{
	struct request *next = elv_latter_request(q, rq);

	if (next)
		return attempt_merge(q, rq, next);

	return 0;
}

int attempt_front_merge(struct request_queue *q, struct request *rq)
{
	struct request *prev = elv_former_request(q, rq);

	if (prev)
		return attempt_merge(q, prev, rq);

	return 0;
}

int blk_attempt_req_merge(struct request_queue *q, struct request *rq,
			  struct request *next)
{
	return attempt_merge(q, rq, next);
}

bool blk_rq_merge_ok(struct request *rq, struct bio *bio)
{
	if (!rq_mergeable(rq) || !bio_mergeable(bio))
		return false;

	if (!blk_check_merge_flags(rq->cmd_flags, bio->bi_rw))
		return false;

	/* different data direction or already started, don't merge */
	if (bio_data_dir(bio) != rq_data_dir(rq))
		return false;

	/* must be same device and not a special request */
	if (rq->rq_disk != bio->bi_bdev->bd_disk || rq->special)
		return false;

	/* only merge integrity protected bio into ditto rq */
	if (bio_integrity(bio) != blk_integrity_rq(rq))
		return false;

	/* must be using the same buffer */
	if (rq->cmd_flags & REQ_WRITE_SAME &&
	    !blk_write_same_mergeable(rq->bio, bio))
		return false;

	return true;
}

int blk_try_merge(struct request *rq, struct bio *bio)
{
	if (blk_rq_pos(rq) + blk_rq_sectors(rq) == bio->bi_sector)
		return ELEVATOR_BACK_MERGE;
	else if (blk_rq_pos(rq) - bio_sectors(bio) == bio->bi_sector)
		return ELEVATOR_FRONT_MERGE;
	return ELEVATOR_NO_MERGE;
}
Commit	Line	Data
d6d48196 JA	1	/*
	2	* Functions related to segment and merge handling
	3	*/
	4	#include <linux/kernel.h>
	5	#include <linux/module.h>
	6	#include <linux/bio.h>
	7	#include <linux/blkdev.h>
	8	#include <linux/scatterlist.h>
	9
	10	#include "blk.h"
	11
1e428079	12	static unsigned int __blk_recalc_rq_segments(struct request_queue *q,
59247eae	13	struct bio *bio)
d6d48196	14	{
d6d48196	15	struct bio_vec bv, bvprv = NULL;
1e428079 JA	16	int cluster, i, high, highprv = 1;
1e428079 JA	17	unsigned int seg_size, nr_phys_segs;
59247eae	18	struct bio fbio, bbio;
d6d48196	19
1e428079 JA	20	if (!bio)
1e428079 JA	21	return 0;
d6d48196	22
1e428079	23	fbio = bio;
e692cb66	24	cluster = blk_queue_cluster(q);
5df97b91	25	seg_size = 0;
2c8919de	26	nr_phys_segs = 0;
1e428079 JA	27	for_each_bio(bio) {
	28	bio_for_each_segment(bv, bio, i) {
	29	/*
	30	* the trick here is making sure that a high page is
	31	* never considered part of another segment, since that
	32	* might change with the bounce page.
	33	*/
ae03bf63	34	high = page_to_pfn(bv->bv_page) > queue_bounce_pfn(q);
1e428079	35	if (high \|\| highprv)
d6d48196	36	goto new_segment;
1e428079	37	if (cluster) {
ae03bf63 MP	38	if (seg_size + bv->bv_len
ae03bf63 MP	39	> queue_max_segment_size(q))
1e428079 JA	40	goto new_segment;
	41	if (!BIOVEC_PHYS_MERGEABLE(bvprv, bv))
	42	goto new_segment;
	43	if (!BIOVEC_SEG_BOUNDARY(q, bvprv, bv))
	44	goto new_segment;
d6d48196	45
1e428079 JA	46	seg_size += bv->bv_len;
	47	bvprv = bv;
	48	continue;
	49	}
d6d48196	50	new_segment:
1e428079 JA	51	if (nr_phys_segs == 1 && seg_size >
	52	fbio->bi_seg_front_size)
	53	fbio->bi_seg_front_size = seg_size;
86771427	54
1e428079 JA	55	nr_phys_segs++;
	56	bvprv = bv;
	57	seg_size = bv->bv_len;
	58	highprv = high;
	59	}
59247eae	60	bbio = bio;
d6d48196 JA	61	}
d6d48196 JA	62
59247eae JA	63	if (nr_phys_segs == 1 && seg_size > fbio->bi_seg_front_size)
	64	fbio->bi_seg_front_size = seg_size;
	65	if (seg_size > bbio->bi_seg_back_size)
	66	bbio->bi_seg_back_size = seg_size;
1e428079 JA	67
	68	return nr_phys_segs;
	69	}
	70
	71	void blk_recalc_rq_segments(struct request *rq)
	72	{
59247eae	73	rq->nr_phys_segments = __blk_recalc_rq_segments(rq->q, rq->bio);
d6d48196 JA	74	}
	75
	76	void blk_recount_segments(struct request_queue q, struct bio bio)
	77	{
d6d48196	78	struct bio *nxt = bio->bi_next;
1e428079	79
d6d48196	80	bio->bi_next = NULL;
59247eae	81	bio->bi_phys_segments = __blk_recalc_rq_segments(q, bio);
d6d48196	82	bio->bi_next = nxt;
d6d48196 JA	83	bio->bi_flags \|= (1 << BIO_SEG_VALID);
	84	}
	85	EXPORT_SYMBOL(blk_recount_segments);
	86
	87	static int blk_phys_contig_segment(struct request_queue q, struct bio bio,
	88	struct bio *nxt)
	89	{
e692cb66	90	if (!blk_queue_cluster(q))
d6d48196 JA	91	return 0;
d6d48196 JA	92
86771427	93	if (bio->bi_seg_back_size + nxt->bi_seg_front_size >
ae03bf63	94	queue_max_segment_size(q))
d6d48196 JA	95	return 0;
d6d48196 JA	96
e17fc0a1 DW	97	if (!bio_has_data(bio))
	98	return 1;
	99
	100	if (!BIOVEC_PHYS_MERGEABLE(__BVEC_END(bio), __BVEC_START(nxt)))
	101	return 0;
	102
d6d48196	103	/*
e17fc0a1	104	* bio and nxt are contiguous in memory; check if the queue allows
d6d48196 JA	105	* these two to be merged into one
	106	*/
	107	if (BIO_SEG_BOUNDARY(q, bio, nxt))
	108	return 1;
	109
	110	return 0;
	111	}
	112
963ab9e5 AH	113	static void
	114	__blk_segment_map_sg(struct request_queue q, struct bio_vec bvec,
	115	struct scatterlist sglist, struct bio_vec *bvprv,
	116	struct scatterlist *sg, int nsegs, int *cluster)
	117	{
	118
	119	int nbytes = bvec->bv_len;
	120
	121	if (bvprv && cluster) {
	122	if ((*sg)->length + nbytes > queue_max_segment_size(q))
	123	goto new_segment;
	124
	125	if (!BIOVEC_PHYS_MERGEABLE(*bvprv, bvec))
	126	goto new_segment;
	127	if (!BIOVEC_SEG_BOUNDARY(q, *bvprv, bvec))
	128	goto new_segment;
	129
	130	(*sg)->length += nbytes;
	131	} else {
	132	new_segment:
	133	if (!*sg)
	134	*sg = sglist;
	135	else {
	136	/*
	137	* If the driver previously mapped a shorter
	138	* list, we could see a termination bit
	139	* prematurely unless it fully inits the sg
	140	* table on each mapping. We KNOW that there
	141	* must be more entries here or the driver
	142	* would be buggy, so force clear the
	143	* termination bit to avoid doing a full
	144	* sg_init_table() in drivers for each command.
	145	*/
	146	(*sg)->page_link &= ~0x02;
	147	sg = sg_next(sg);
	148	}
	149
	150	sg_set_page(*sg, bvec->bv_page, nbytes, bvec->bv_offset);
	151	(*nsegs)++;
	152	}
	153	*bvprv = bvec;
	154	}
	155
d6d48196 JA	156	/*
	157	* map a request to scatterlist, return number of sg entries setup. Caller
	158	* must make sure sg can hold rq->nr_phys_segments entries
	159	*/
	160	int blk_rq_map_sg(struct request_queue q, struct request rq,
	161	struct scatterlist *sglist)
	162	{
	163	struct bio_vec bvec, bvprv;
	164	struct req_iterator iter;
	165	struct scatterlist *sg;
	166	int nsegs, cluster;
	167
	168	nsegs = 0;
e692cb66	169	cluster = blk_queue_cluster(q);
d6d48196 JA	170
	171	/*
	172	* for each bio in rq
	173	*/
	174	bvprv = NULL;
	175	sg = NULL;
	176	rq_for_each_segment(bvec, rq, iter) {
963ab9e5 AH	177	__blk_segment_map_sg(q, bvec, sglist, &bvprv, &sg,
963ab9e5 AH	178	&nsegs, &cluster);
d6d48196 JA	179	} /* segments in rq */
d6d48196 JA	180
f18573ab FT	181
f18573ab FT	182	if (unlikely(rq->cmd_flags & REQ_COPY_USER) &&
2e46e8b2 TH	183	(blk_rq_bytes(rq) & q->dma_pad_mask)) {
	184	unsigned int pad_len =
	185	(q->dma_pad_mask & ~blk_rq_bytes(rq)) + 1;
f18573ab FT	186
	187	sg->length += pad_len;
	188	rq->extra_len += pad_len;
	189	}
	190
2fb98e84	191	if (q->dma_drain_size && q->dma_drain_needed(rq)) {
7b6d91da	192	if (rq->cmd_flags & REQ_WRITE)
db0a2e00 TH	193	memset(q->dma_drain_buffer, 0, q->dma_drain_size);
db0a2e00 TH	194
d6d48196 JA	195	sg->page_link &= ~0x02;
	196	sg = sg_next(sg);
	197	sg_set_page(sg, virt_to_page(q->dma_drain_buffer),
	198	q->dma_drain_size,
	199	((unsigned long)q->dma_drain_buffer) &
	200	(PAGE_SIZE - 1));
	201	nsegs++;
7a85f889	202	rq->extra_len += q->dma_drain_size;
d6d48196 JA	203	}
	204
	205	if (sg)
	206	sg_mark_end(sg);
	207
	208	return nsegs;
	209	}
d6d48196 JA	210	EXPORT_SYMBOL(blk_rq_map_sg);
d6d48196 JA	211
85b9f66a AH	212	/**
	213	* blk_bio_map_sg - map a bio to a scatterlist
	214	* @q: request_queue in question
	215	* @bio: bio being mapped
	216	* @sglist: scatterlist being mapped
	217	*
	218	* Note:
	219	* Caller must make sure sg can hold bio->bi_phys_segments entries
	220	*
	221	* Will return the number of sg entries setup
	222	*/
	223	int blk_bio_map_sg(struct request_queue q, struct bio bio,
	224	struct scatterlist *sglist)
	225	{
	226	struct bio_vec bvec, bvprv;
	227	struct scatterlist *sg;
	228	int nsegs, cluster;
	229	unsigned long i;
	230
	231	nsegs = 0;
	232	cluster = blk_queue_cluster(q);
	233
	234	bvprv = NULL;
	235	sg = NULL;
	236	bio_for_each_segment(bvec, bio, i) {
	237	__blk_segment_map_sg(q, bvec, sglist, &bvprv, &sg,
	238	&nsegs, &cluster);
	239	} /* segments in bio */
	240
	241	if (sg)
	242	sg_mark_end(sg);
	243
	244	BUG_ON(bio->bi_phys_segments && nsegs > bio->bi_phys_segments);
	245	return nsegs;
	246	}
	247	EXPORT_SYMBOL(blk_bio_map_sg);
	248
d6d48196 JA	249	static inline int ll_new_hw_segment(struct request_queue *q,
	250	struct request *req,
	251	struct bio *bio)
	252	{
d6d48196 JA	253	int nr_phys_segs = bio_phys_segments(q, bio);
d6d48196 JA	254
13f05c8d MP	255	if (req->nr_phys_segments + nr_phys_segs > queue_max_segments(q))
	256	goto no_merge;
	257
	258	if (bio_integrity(bio) && blk_integrity_merge_bio(q, req, bio))
	259	goto no_merge;
d6d48196 JA	260
	261	/*
	262	* This will form the start of a new hw segment. Bump both
	263	* counters.
	264	*/
d6d48196 JA	265	req->nr_phys_segments += nr_phys_segs;
d6d48196 JA	266	return 1;
13f05c8d MP	267
	268	no_merge:
	269	req->cmd_flags \|= REQ_NOMERGE;
	270	if (req == q->last_merge)
	271	q->last_merge = NULL;
	272	return 0;
d6d48196 JA	273	}
	274
	275	int ll_back_merge_fn(struct request_queue q, struct request req,
	276	struct bio *bio)
	277	{
f31dc1cd MP	278	if (blk_rq_sectors(req) + bio_sectors(bio) >
f31dc1cd MP	279	blk_rq_get_max_sectors(req)) {
d6d48196 JA	280	req->cmd_flags \|= REQ_NOMERGE;
	281	if (req == q->last_merge)
	282	q->last_merge = NULL;
	283	return 0;
	284	}
2cdf79ca	285	if (!bio_flagged(req->biotail, BIO_SEG_VALID))
d6d48196	286	blk_recount_segments(q, req->biotail);
2cdf79ca	287	if (!bio_flagged(bio, BIO_SEG_VALID))
d6d48196	288	blk_recount_segments(q, bio);
d6d48196 JA	289
	290	return ll_new_hw_segment(q, req, bio);
	291	}
	292
6728cb0e	293	int ll_front_merge_fn(struct request_queue q, struct request req,
d6d48196 JA	294	struct bio *bio)
d6d48196 JA	295	{
f31dc1cd MP	296	if (blk_rq_sectors(req) + bio_sectors(bio) >
f31dc1cd MP	297	blk_rq_get_max_sectors(req)) {
d6d48196 JA	298	req->cmd_flags \|= REQ_NOMERGE;
	299	if (req == q->last_merge)
	300	q->last_merge = NULL;
	301	return 0;
	302	}
2cdf79ca	303	if (!bio_flagged(bio, BIO_SEG_VALID))
d6d48196	304	blk_recount_segments(q, bio);
2cdf79ca	305	if (!bio_flagged(req->bio, BIO_SEG_VALID))
d6d48196	306	blk_recount_segments(q, req->bio);
d6d48196 JA	307
	308	return ll_new_hw_segment(q, req, bio);
	309	}
	310
	311	static int ll_merge_requests_fn(struct request_queue q, struct request req,
	312	struct request *next)
	313	{
	314	int total_phys_segments;
86771427 FT	315	unsigned int seg_size =
86771427 FT	316	req->biotail->bi_seg_back_size + next->bio->bi_seg_front_size;
d6d48196 JA	317
	318	/*
	319	* First check if the either of the requests are re-queued
	320	* requests. Can't merge them if they are.
	321	*/
	322	if (req->special \|\| next->special)
	323	return 0;
	324
	325	/*
	326	* Will it become too large?
	327	*/
f31dc1cd MP	328	if ((blk_rq_sectors(req) + blk_rq_sectors(next)) >
f31dc1cd MP	329	blk_rq_get_max_sectors(req))
d6d48196 JA	330	return 0;
	331
	332	total_phys_segments = req->nr_phys_segments + next->nr_phys_segments;
86771427 FT	333	if (blk_phys_contig_segment(q, req->biotail, next->bio)) {
	334	if (req->nr_phys_segments == 1)
	335	req->bio->bi_seg_front_size = seg_size;
	336	if (next->nr_phys_segments == 1)
	337	next->biotail->bi_seg_back_size = seg_size;
d6d48196	338	total_phys_segments--;
86771427	339	}
d6d48196	340
8a78362c	341	if (total_phys_segments > queue_max_segments(q))
d6d48196 JA	342	return 0;
d6d48196 JA	343
13f05c8d MP	344	if (blk_integrity_rq(req) && blk_integrity_merge_rq(q, req, next))
	345	return 0;
	346
d6d48196 JA	347	/* Merge is OK... */
d6d48196 JA	348	req->nr_phys_segments = total_phys_segments;
d6d48196 JA	349	return 1;
	350	}
	351
80a761fd TH	352	/**
	353	* blk_rq_set_mixed_merge - mark a request as mixed merge
	354	* @rq: request to mark as mixed merge
	355	*
	356	* Description:
	357	* @rq is about to be mixed merged. Make sure the attributes
	358	* which can be mixed are set in each bio and mark @rq as mixed
	359	* merged.
	360	*/
	361	void blk_rq_set_mixed_merge(struct request *rq)
	362	{
	363	unsigned int ff = rq->cmd_flags & REQ_FAILFAST_MASK;
	364	struct bio *bio;
	365
	366	if (rq->cmd_flags & REQ_MIXED_MERGE)
	367	return;
	368
	369	/*
	370	* @rq will no longer represent mixable attributes for all the
	371	* contained bios. It will just track those of the first one.
	372	* Distributes the attributs to each bio.
	373	*/
	374	for (bio = rq->bio; bio; bio = bio->bi_next) {
	375	WARN_ON_ONCE((bio->bi_rw & REQ_FAILFAST_MASK) &&
	376	(bio->bi_rw & REQ_FAILFAST_MASK) != ff);
	377	bio->bi_rw \|= ff;
	378	}
	379	rq->cmd_flags \|= REQ_MIXED_MERGE;
	380	}
	381
26308eab JM	382	static void blk_account_io_merge(struct request *req)
	383	{
	384	if (blk_do_io_stat(req)) {
	385	struct hd_struct *part;
	386	int cpu;
	387
	388	cpu = part_stat_lock();
09e099d4	389	part = req->part;
26308eab JM	390
26308eab JM	391	part_round_stats(cpu, part);
316d315b	392	part_dec_in_flight(part, rq_data_dir(req));
26308eab	393
6c23a968	394	hd_struct_put(part);
26308eab JM	395	part_stat_unlock();
	396	}
	397	}
	398
d6d48196 JA	399	/*
	400	* Has to be called with the request spinlock acquired
	401	*/
	402	static int attempt_merge(struct request_queue q, struct request req,
	403	struct request *next)
	404	{
	405	if (!rq_mergeable(req) \|\| !rq_mergeable(next))
	406	return 0;
	407
f31dc1cd MP	408	if (!blk_check_merge_flags(req->cmd_flags, next->cmd_flags))
	409	return 0;
	410
d6d48196 JA	411	/*
	412	* not contiguous
	413	*/
83096ebf	414	if (blk_rq_pos(req) + blk_rq_sectors(req) != blk_rq_pos(next))
d6d48196 JA	415	return 0;
	416
	417	if (rq_data_dir(req) != rq_data_dir(next)
	418	\|\| req->rq_disk != next->rq_disk
	419	\|\| next->special)
	420	return 0;
	421
4363ac7c MP	422	if (req->cmd_flags & REQ_WRITE_SAME &&
	423	!blk_write_same_mergeable(req->bio, next->bio))
	424	return 0;
	425
d6d48196 JA	426	/*
	427	* If we are allowed to merge, then append bio list
	428	* from next to rq and release next. merge_requests_fn
	429	* will have updated segment counts, update sector
	430	* counts here.
	431	*/
	432	if (!ll_merge_requests_fn(q, req, next))
	433	return 0;
	434
80a761fd TH	435	/*
	436	* If failfast settings disagree or any of the two is already
	437	* a mixed merge, mark both as mixed before proceeding. This
	438	* makes sure that all involved bios have mixable attributes
	439	* set properly.
	440	*/
	441	if ((req->cmd_flags \| next->cmd_flags) & REQ_MIXED_MERGE \|\|
	442	(req->cmd_flags & REQ_FAILFAST_MASK) !=
	443	(next->cmd_flags & REQ_FAILFAST_MASK)) {
	444	blk_rq_set_mixed_merge(req);
	445	blk_rq_set_mixed_merge(next);
	446	}
	447
d6d48196 JA	448	/*
	449	* At this point we have either done a back merge
	450	* or front merge. We need the smaller start_time of
	451	* the merged requests to be the current request
	452	* for accounting purposes.
	453	*/
	454	if (time_after(req->start_time, next->start_time))
	455	req->start_time = next->start_time;
	456
	457	req->biotail->bi_next = next->bio;
	458	req->biotail = next->biotail;
	459
a2dec7b3	460	req->__data_len += blk_rq_bytes(next);
d6d48196 JA	461
	462	elv_merge_requests(q, req, next);
	463
42dad764 JM	464	/*
	465	* 'next' is going away, so update stats accordingly
	466	*/
	467	blk_account_io_merge(next);
d6d48196 JA	468
d6d48196 JA	469	req->ioprio = ioprio_best(req->ioprio, next->ioprio);
ab780f1e JA	470	if (blk_rq_cpu_valid(next))
ab780f1e JA	471	req->cpu = next->cpu;
d6d48196	472
1cd96c24 BH	473	/* owner-ship of bio passed from next to req */
1cd96c24 BH	474	next->bio = NULL;
d6d48196 JA	475	__blk_put_request(q, next);
	476	return 1;
	477	}
	478
	479	int attempt_back_merge(struct request_queue q, struct request rq)
	480	{
	481	struct request *next = elv_latter_request(q, rq);
	482
	483	if (next)
	484	return attempt_merge(q, rq, next);
	485
	486	return 0;
	487	}
	488
	489	int attempt_front_merge(struct request_queue q, struct request rq)
	490	{
	491	struct request *prev = elv_former_request(q, rq);
	492
	493	if (prev)
	494	return attempt_merge(q, prev, rq);
	495
	496	return 0;
	497	}
5e84ea3a JA	498
	499	int blk_attempt_req_merge(struct request_queue q, struct request rq,
	500	struct request *next)
	501	{
	502	return attempt_merge(q, rq, next);
	503	}
050c8ea8 TH	504
	505	bool blk_rq_merge_ok(struct request rq, struct bio bio)
	506	{
e2a60da7	507	if (!rq_mergeable(rq) \|\| !bio_mergeable(bio))
050c8ea8 TH	508	return false;
050c8ea8 TH	509
f31dc1cd MP	510	if (!blk_check_merge_flags(rq->cmd_flags, bio->bi_rw))
	511	return false;
	512
050c8ea8 TH	513	/* different data direction or already started, don't merge */
	514	if (bio_data_dir(bio) != rq_data_dir(rq))
	515	return false;
	516
	517	/* must be same device and not a special request */
	518	if (rq->rq_disk != bio->bi_bdev->bd_disk \|\| rq->special)
	519	return false;
	520
	521	/* only merge integrity protected bio into ditto rq */
	522	if (bio_integrity(bio) != blk_integrity_rq(rq))
	523	return false;
	524
4363ac7c MP	525	/* must be using the same buffer */
	526	if (rq->cmd_flags & REQ_WRITE_SAME &&
	527	!blk_write_same_mergeable(rq->bio, bio))
	528	return false;
	529
050c8ea8 TH	530	return true;
	531	}
	532
	533	int blk_try_merge(struct request rq, struct bio bio)
	534	{
	535	if (blk_rq_pos(rq) + blk_rq_sectors(rq) == bio->bi_sector)
	536	return ELEVATOR_BACK_MERGE;
	537	else if (blk_rq_pos(rq) - bio_sectors(bio) == bio->bi_sector)
	538	return ELEVATOR_FRONT_MERGE;
	539	return ELEVATOR_NO_MERGE;
	540	}