2008-01-29 20:04:06 +07:00
|
|
|
/*
|
|
|
|
* 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>
|
|
|
|
|
2015-12-03 21:32:30 +07:00
|
|
|
#include <trace/events/block.h>
|
|
|
|
|
2008-01-29 20:04:06 +07:00
|
|
|
#include "blk.h"
|
|
|
|
|
2015-04-24 12:37:18 +07:00
|
|
|
static struct bio *blk_bio_discard_split(struct request_queue *q,
|
|
|
|
struct bio *bio,
|
2015-10-20 22:13:52 +07:00
|
|
|
struct bio_set *bs,
|
|
|
|
unsigned *nsegs)
|
2015-04-24 12:37:18 +07:00
|
|
|
{
|
|
|
|
unsigned int max_discard_sectors, granularity;
|
|
|
|
int alignment;
|
|
|
|
sector_t tmp;
|
|
|
|
unsigned split_sectors;
|
|
|
|
|
2015-10-20 22:13:52 +07:00
|
|
|
*nsegs = 1;
|
|
|
|
|
2015-04-24 12:37:18 +07:00
|
|
|
/* Zero-sector (unknown) and one-sector granularities are the same. */
|
|
|
|
granularity = max(q->limits.discard_granularity >> 9, 1U);
|
|
|
|
|
|
|
|
max_discard_sectors = min(q->limits.max_discard_sectors, UINT_MAX >> 9);
|
|
|
|
max_discard_sectors -= max_discard_sectors % granularity;
|
|
|
|
|
|
|
|
if (unlikely(!max_discard_sectors)) {
|
|
|
|
/* XXX: warn */
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (bio_sectors(bio) <= max_discard_sectors)
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
split_sectors = max_discard_sectors;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* If the next starting sector would be misaligned, stop the discard at
|
|
|
|
* the previous aligned sector.
|
|
|
|
*/
|
|
|
|
alignment = (q->limits.discard_alignment >> 9) % granularity;
|
|
|
|
|
|
|
|
tmp = bio->bi_iter.bi_sector + split_sectors - alignment;
|
|
|
|
tmp = sector_div(tmp, granularity);
|
|
|
|
|
|
|
|
if (split_sectors > tmp)
|
|
|
|
split_sectors -= tmp;
|
|
|
|
|
|
|
|
return bio_split(bio, split_sectors, GFP_NOIO, bs);
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct bio *blk_bio_write_same_split(struct request_queue *q,
|
|
|
|
struct bio *bio,
|
2015-10-20 22:13:52 +07:00
|
|
|
struct bio_set *bs,
|
|
|
|
unsigned *nsegs)
|
2015-04-24 12:37:18 +07:00
|
|
|
{
|
2015-10-20 22:13:52 +07:00
|
|
|
*nsegs = 1;
|
|
|
|
|
2015-04-24 12:37:18 +07:00
|
|
|
if (!q->limits.max_write_same_sectors)
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
if (bio_sectors(bio) <= q->limits.max_write_same_sectors)
|
|
|
|
return NULL;
|
|
|
|
|
|
|
|
return bio_split(bio, q->limits.max_write_same_sectors, GFP_NOIO, bs);
|
|
|
|
}
|
|
|
|
|
2016-01-23 07:05:33 +07:00
|
|
|
static inline unsigned get_max_io_size(struct request_queue *q,
|
|
|
|
struct bio *bio)
|
|
|
|
{
|
|
|
|
unsigned sectors = blk_max_size_offset(q, bio->bi_iter.bi_sector);
|
|
|
|
unsigned mask = queue_logical_block_size(q) - 1;
|
|
|
|
|
|
|
|
/* aligned to logical block size */
|
|
|
|
sectors &= ~(mask >> 9);
|
|
|
|
|
|
|
|
return sectors;
|
|
|
|
}
|
|
|
|
|
2015-04-24 12:37:18 +07:00
|
|
|
static struct bio *blk_bio_segment_split(struct request_queue *q,
|
|
|
|
struct bio *bio,
|
2015-10-20 22:13:52 +07:00
|
|
|
struct bio_set *bs,
|
|
|
|
unsigned *segs)
|
2015-04-24 12:37:18 +07:00
|
|
|
{
|
2015-09-03 05:46:02 +07:00
|
|
|
struct bio_vec bv, bvprv, *bvprvp = NULL;
|
2015-04-24 12:37:18 +07:00
|
|
|
struct bvec_iter iter;
|
2015-04-28 13:48:34 +07:00
|
|
|
unsigned seg_size = 0, nsegs = 0, sectors = 0;
|
2015-11-24 09:35:30 +07:00
|
|
|
unsigned front_seg_size = bio->bi_seg_front_size;
|
|
|
|
bool do_split = true;
|
|
|
|
struct bio *new = NULL;
|
2016-01-23 07:05:33 +07:00
|
|
|
const unsigned max_sectors = get_max_io_size(q, bio);
|
2015-04-24 12:37:18 +07:00
|
|
|
|
|
|
|
bio_for_each_segment(bv, bio, iter) {
|
|
|
|
/*
|
|
|
|
* If the queue doesn't support SG gaps and adding this
|
|
|
|
* offset would create a gap, disallow it.
|
|
|
|
*/
|
2015-09-03 05:46:02 +07:00
|
|
|
if (bvprvp && bvec_gap_to_prev(q, bvprvp, bv.bv_offset))
|
2015-04-24 12:37:18 +07:00
|
|
|
goto split;
|
|
|
|
|
2016-01-23 07:05:33 +07:00
|
|
|
if (sectors + (bv.bv_len >> 9) > max_sectors) {
|
2016-01-13 05:08:39 +07:00
|
|
|
/*
|
|
|
|
* Consider this a new segment if we're splitting in
|
|
|
|
* the middle of this vector.
|
|
|
|
*/
|
|
|
|
if (nsegs < queue_max_segments(q) &&
|
2016-01-23 07:05:33 +07:00
|
|
|
sectors < max_sectors) {
|
2016-01-13 05:08:39 +07:00
|
|
|
nsegs++;
|
2016-01-23 07:05:33 +07:00
|
|
|
sectors = max_sectors;
|
2016-01-13 05:08:39 +07:00
|
|
|
}
|
2016-01-23 07:05:33 +07:00
|
|
|
if (sectors)
|
|
|
|
goto split;
|
|
|
|
/* Make this single bvec as the 1st segment */
|
2016-01-13 05:08:39 +07:00
|
|
|
}
|
|
|
|
|
2015-09-03 05:46:02 +07:00
|
|
|
if (bvprvp && blk_queue_cluster(q)) {
|
2015-04-24 12:37:18 +07:00
|
|
|
if (seg_size + bv.bv_len > queue_max_segment_size(q))
|
|
|
|
goto new_segment;
|
2015-09-03 05:46:02 +07:00
|
|
|
if (!BIOVEC_PHYS_MERGEABLE(bvprvp, &bv))
|
2015-04-24 12:37:18 +07:00
|
|
|
goto new_segment;
|
2015-09-03 05:46:02 +07:00
|
|
|
if (!BIOVEC_SEG_BOUNDARY(q, bvprvp, &bv))
|
2015-04-24 12:37:18 +07:00
|
|
|
goto new_segment;
|
|
|
|
|
|
|
|
seg_size += bv.bv_len;
|
|
|
|
bvprv = bv;
|
2015-11-24 09:35:29 +07:00
|
|
|
bvprvp = &bvprv;
|
2015-09-17 22:58:38 +07:00
|
|
|
sectors += bv.bv_len >> 9;
|
2015-11-30 15:05:49 +07:00
|
|
|
|
|
|
|
if (nsegs == 1 && seg_size > front_seg_size)
|
|
|
|
front_seg_size = seg_size;
|
2015-04-24 12:37:18 +07:00
|
|
|
continue;
|
|
|
|
}
|
|
|
|
new_segment:
|
|
|
|
if (nsegs == queue_max_segments(q))
|
|
|
|
goto split;
|
|
|
|
|
|
|
|
nsegs++;
|
|
|
|
bvprv = bv;
|
2015-11-24 09:35:29 +07:00
|
|
|
bvprvp = &bvprv;
|
2015-04-24 12:37:18 +07:00
|
|
|
seg_size = bv.bv_len;
|
2015-09-17 22:58:38 +07:00
|
|
|
sectors += bv.bv_len >> 9;
|
2015-11-24 09:35:30 +07:00
|
|
|
|
|
|
|
if (nsegs == 1 && seg_size > front_seg_size)
|
|
|
|
front_seg_size = seg_size;
|
2015-04-24 12:37:18 +07:00
|
|
|
}
|
|
|
|
|
2015-11-24 09:35:30 +07:00
|
|
|
do_split = false;
|
2015-04-24 12:37:18 +07:00
|
|
|
split:
|
2015-10-20 22:13:52 +07:00
|
|
|
*segs = nsegs;
|
2015-11-24 09:35:30 +07:00
|
|
|
|
|
|
|
if (do_split) {
|
|
|
|
new = bio_split(bio, sectors, GFP_NOIO, bs);
|
|
|
|
if (new)
|
|
|
|
bio = new;
|
|
|
|
}
|
|
|
|
|
|
|
|
bio->bi_seg_front_size = front_seg_size;
|
|
|
|
if (seg_size > bio->bi_seg_back_size)
|
|
|
|
bio->bi_seg_back_size = seg_size;
|
|
|
|
|
|
|
|
return do_split ? new : NULL;
|
2015-04-24 12:37:18 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
void blk_queue_split(struct request_queue *q, struct bio **bio,
|
|
|
|
struct bio_set *bs)
|
|
|
|
{
|
2015-10-20 22:13:52 +07:00
|
|
|
struct bio *split, *res;
|
|
|
|
unsigned nsegs;
|
2015-04-24 12:37:18 +07:00
|
|
|
|
|
|
|
if ((*bio)->bi_rw & REQ_DISCARD)
|
2015-10-20 22:13:52 +07:00
|
|
|
split = blk_bio_discard_split(q, *bio, bs, &nsegs);
|
2015-04-24 12:37:18 +07:00
|
|
|
else if ((*bio)->bi_rw & REQ_WRITE_SAME)
|
2015-10-20 22:13:52 +07:00
|
|
|
split = blk_bio_write_same_split(q, *bio, bs, &nsegs);
|
2015-04-24 12:37:18 +07:00
|
|
|
else
|
2015-10-20 22:13:52 +07:00
|
|
|
split = blk_bio_segment_split(q, *bio, q->bio_split, &nsegs);
|
|
|
|
|
|
|
|
/* physical segments can be figured out during splitting */
|
|
|
|
res = split ? split : *bio;
|
|
|
|
res->bi_phys_segments = nsegs;
|
|
|
|
bio_set_flag(res, BIO_SEG_VALID);
|
2015-04-24 12:37:18 +07:00
|
|
|
|
|
|
|
if (split) {
|
2015-10-20 22:13:53 +07:00
|
|
|
/* there isn't chance to merge the splitted bio */
|
|
|
|
split->bi_rw |= REQ_NOMERGE;
|
|
|
|
|
2015-04-24 12:37:18 +07:00
|
|
|
bio_chain(split, *bio);
|
2015-12-03 21:32:30 +07:00
|
|
|
trace_block_split(q, split, (*bio)->bi_iter.bi_sector);
|
2015-04-24 12:37:18 +07:00
|
|
|
generic_make_request(*bio);
|
|
|
|
*bio = split;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(blk_queue_split);
|
|
|
|
|
2009-02-23 15:03:10 +07:00
|
|
|
static unsigned int __blk_recalc_rq_segments(struct request_queue *q,
|
2014-09-02 22:02:59 +07:00
|
|
|
struct bio *bio,
|
|
|
|
bool no_sg_merge)
|
2008-01-29 20:04:06 +07:00
|
|
|
{
|
2013-11-24 08:19:00 +07:00
|
|
|
struct bio_vec bv, bvprv = { NULL };
|
2015-04-24 12:37:18 +07:00
|
|
|
int cluster, prev = 0;
|
2009-02-23 15:03:10 +07:00
|
|
|
unsigned int seg_size, nr_phys_segs;
|
2009-03-06 14:55:24 +07:00
|
|
|
struct bio *fbio, *bbio;
|
2013-11-24 08:19:00 +07:00
|
|
|
struct bvec_iter iter;
|
2008-01-29 20:04:06 +07:00
|
|
|
|
2009-02-23 15:03:10 +07:00
|
|
|
if (!bio)
|
|
|
|
return 0;
|
2008-01-29 20:04:06 +07:00
|
|
|
|
2014-02-08 03:53:46 +07:00
|
|
|
/*
|
|
|
|
* This should probably be returning 0, but blk_add_request_payload()
|
|
|
|
* (Christoph!!!!)
|
|
|
|
*/
|
|
|
|
if (bio->bi_rw & REQ_DISCARD)
|
|
|
|
return 1;
|
|
|
|
|
|
|
|
if (bio->bi_rw & REQ_WRITE_SAME)
|
|
|
|
return 1;
|
|
|
|
|
2009-02-23 15:03:10 +07:00
|
|
|
fbio = bio;
|
2010-12-02 01:41:49 +07:00
|
|
|
cluster = blk_queue_cluster(q);
|
2008-08-15 15:20:02 +07:00
|
|
|
seg_size = 0;
|
2010-06-21 16:02:47 +07:00
|
|
|
nr_phys_segs = 0;
|
2009-02-23 15:03:10 +07:00
|
|
|
for_each_bio(bio) {
|
2013-11-24 08:19:00 +07:00
|
|
|
bio_for_each_segment(bv, bio, iter) {
|
2014-05-29 22:53:32 +07:00
|
|
|
/*
|
|
|
|
* If SG merging is disabled, each bio vector is
|
|
|
|
* a segment
|
|
|
|
*/
|
|
|
|
if (no_sg_merge)
|
|
|
|
goto new_segment;
|
|
|
|
|
2015-04-24 12:37:18 +07:00
|
|
|
if (prev && cluster) {
|
2013-11-24 08:19:00 +07:00
|
|
|
if (seg_size + bv.bv_len
|
2009-05-23 04:17:50 +07:00
|
|
|
> queue_max_segment_size(q))
|
2009-02-23 15:03:10 +07:00
|
|
|
goto new_segment;
|
2013-11-24 08:19:00 +07:00
|
|
|
if (!BIOVEC_PHYS_MERGEABLE(&bvprv, &bv))
|
2009-02-23 15:03:10 +07:00
|
|
|
goto new_segment;
|
2013-11-24 08:19:00 +07:00
|
|
|
if (!BIOVEC_SEG_BOUNDARY(q, &bvprv, &bv))
|
2009-02-23 15:03:10 +07:00
|
|
|
goto new_segment;
|
2008-01-29 20:04:06 +07:00
|
|
|
|
2013-11-24 08:19:00 +07:00
|
|
|
seg_size += bv.bv_len;
|
2009-02-23 15:03:10 +07:00
|
|
|
bvprv = bv;
|
|
|
|
continue;
|
|
|
|
}
|
2008-01-29 20:04:06 +07:00
|
|
|
new_segment:
|
2009-02-23 15:03:10 +07:00
|
|
|
if (nr_phys_segs == 1 && seg_size >
|
|
|
|
fbio->bi_seg_front_size)
|
|
|
|
fbio->bi_seg_front_size = seg_size;
|
2008-10-13 19:19:05 +07:00
|
|
|
|
2009-02-23 15:03:10 +07:00
|
|
|
nr_phys_segs++;
|
|
|
|
bvprv = bv;
|
2015-04-24 12:37:18 +07:00
|
|
|
prev = 1;
|
2013-11-24 08:19:00 +07:00
|
|
|
seg_size = bv.bv_len;
|
2009-02-23 15:03:10 +07:00
|
|
|
}
|
2009-03-06 14:55:24 +07:00
|
|
|
bbio = bio;
|
2008-01-29 20:04:06 +07:00
|
|
|
}
|
|
|
|
|
2009-03-06 14:55:24 +07:00
|
|
|
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;
|
2009-02-23 15:03:10 +07:00
|
|
|
|
|
|
|
return nr_phys_segs;
|
|
|
|
}
|
|
|
|
|
|
|
|
void blk_recalc_rq_segments(struct request *rq)
|
|
|
|
{
|
2014-09-02 22:02:59 +07:00
|
|
|
bool no_sg_merge = !!test_bit(QUEUE_FLAG_NO_SG_MERGE,
|
|
|
|
&rq->q->queue_flags);
|
|
|
|
|
|
|
|
rq->nr_phys_segments = __blk_recalc_rq_segments(rq->q, rq->bio,
|
|
|
|
no_sg_merge);
|
2008-01-29 20:04:06 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
void blk_recount_segments(struct request_queue *q, struct bio *bio)
|
|
|
|
{
|
2014-11-11 23:15:41 +07:00
|
|
|
unsigned short seg_cnt;
|
|
|
|
|
|
|
|
/* estimate segment number by bi_vcnt for non-cloned bio */
|
|
|
|
if (bio_flagged(bio, BIO_CLONED))
|
|
|
|
seg_cnt = bio_segments(bio);
|
|
|
|
else
|
|
|
|
seg_cnt = bio->bi_vcnt;
|
2014-10-09 22:17:35 +07:00
|
|
|
|
2014-11-11 23:15:41 +07:00
|
|
|
if (test_bit(QUEUE_FLAG_NO_SG_MERGE, &q->queue_flags) &&
|
|
|
|
(seg_cnt < queue_max_segments(q)))
|
|
|
|
bio->bi_phys_segments = seg_cnt;
|
2014-05-29 22:53:32 +07:00
|
|
|
else {
|
|
|
|
struct bio *nxt = bio->bi_next;
|
|
|
|
|
|
|
|
bio->bi_next = NULL;
|
2014-11-11 23:15:41 +07:00
|
|
|
bio->bi_phys_segments = __blk_recalc_rq_segments(q, bio, false);
|
2014-05-29 22:53:32 +07:00
|
|
|
bio->bi_next = nxt;
|
|
|
|
}
|
2009-02-23 15:03:10 +07:00
|
|
|
|
2015-07-25 01:37:59 +07:00
|
|
|
bio_set_flag(bio, BIO_SEG_VALID);
|
2008-01-29 20:04:06 +07:00
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(blk_recount_segments);
|
|
|
|
|
|
|
|
static int blk_phys_contig_segment(struct request_queue *q, struct bio *bio,
|
|
|
|
struct bio *nxt)
|
|
|
|
{
|
2013-12-04 04:29:09 +07:00
|
|
|
struct bio_vec end_bv = { NULL }, nxt_bv;
|
2013-08-08 04:30:33 +07:00
|
|
|
struct bvec_iter iter;
|
|
|
|
|
2010-12-02 01:41:49 +07:00
|
|
|
if (!blk_queue_cluster(q))
|
2008-01-29 20:04:06 +07:00
|
|
|
return 0;
|
|
|
|
|
2008-10-13 19:19:05 +07:00
|
|
|
if (bio->bi_seg_back_size + nxt->bi_seg_front_size >
|
2009-05-23 04:17:50 +07:00
|
|
|
queue_max_segment_size(q))
|
2008-01-29 20:04:06 +07:00
|
|
|
return 0;
|
|
|
|
|
2008-08-09 22:42:20 +07:00
|
|
|
if (!bio_has_data(bio))
|
|
|
|
return 1;
|
|
|
|
|
2013-08-08 04:30:33 +07:00
|
|
|
bio_for_each_segment(end_bv, bio, iter)
|
|
|
|
if (end_bv.bv_len == iter.bi_size)
|
|
|
|
break;
|
|
|
|
|
|
|
|
nxt_bv = bio_iovec(nxt);
|
|
|
|
|
|
|
|
if (!BIOVEC_PHYS_MERGEABLE(&end_bv, &nxt_bv))
|
2008-08-09 22:42:20 +07:00
|
|
|
return 0;
|
|
|
|
|
2008-01-29 20:04:06 +07:00
|
|
|
/*
|
2008-08-09 22:42:20 +07:00
|
|
|
* bio and nxt are contiguous in memory; check if the queue allows
|
2008-01-29 20:04:06 +07:00
|
|
|
* these two to be merged into one
|
|
|
|
*/
|
2013-08-08 04:30:33 +07:00
|
|
|
if (BIOVEC_SEG_BOUNDARY(q, &end_bv, &nxt_bv))
|
2008-01-29 20:04:06 +07:00
|
|
|
return 1;
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2013-11-24 08:19:00 +07:00
|
|
|
static inline void
|
2012-08-03 04:42:03 +07:00
|
|
|
__blk_segment_map_sg(struct request_queue *q, struct bio_vec *bvec,
|
2013-11-24 08:19:00 +07:00
|
|
|
struct scatterlist *sglist, struct bio_vec *bvprv,
|
2012-08-03 04:42:03 +07:00
|
|
|
struct scatterlist **sg, int *nsegs, int *cluster)
|
|
|
|
{
|
|
|
|
|
|
|
|
int nbytes = bvec->bv_len;
|
|
|
|
|
2013-11-24 08:19:00 +07:00
|
|
|
if (*sg && *cluster) {
|
2012-08-03 04:42:03 +07:00
|
|
|
if ((*sg)->length + nbytes > queue_max_segment_size(q))
|
|
|
|
goto new_segment;
|
|
|
|
|
2013-11-24 08:19:00 +07:00
|
|
|
if (!BIOVEC_PHYS_MERGEABLE(bvprv, bvec))
|
2012-08-03 04:42:03 +07:00
|
|
|
goto new_segment;
|
2013-11-24 08:19:00 +07:00
|
|
|
if (!BIOVEC_SEG_BOUNDARY(q, bvprv, bvec))
|
2012-08-03 04:42:03 +07:00
|
|
|
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.
|
|
|
|
*/
|
2013-03-20 12:07:08 +07:00
|
|
|
sg_unmark_end(*sg);
|
2012-08-03 04:42:03 +07:00
|
|
|
*sg = sg_next(*sg);
|
|
|
|
}
|
|
|
|
|
|
|
|
sg_set_page(*sg, bvec->bv_page, nbytes, bvec->bv_offset);
|
|
|
|
(*nsegs)++;
|
|
|
|
}
|
2013-11-24 08:19:00 +07:00
|
|
|
*bvprv = *bvec;
|
2012-08-03 04:42:03 +07:00
|
|
|
}
|
|
|
|
|
2014-02-08 03:53:46 +07:00
|
|
|
static int __blk_bios_map_sg(struct request_queue *q, struct bio *bio,
|
|
|
|
struct scatterlist *sglist,
|
|
|
|
struct scatterlist **sg)
|
2008-01-29 20:04:06 +07:00
|
|
|
{
|
2013-12-04 04:29:09 +07:00
|
|
|
struct bio_vec bvec, bvprv = { NULL };
|
2014-02-08 03:53:46 +07:00
|
|
|
struct bvec_iter iter;
|
2008-01-29 20:04:06 +07:00
|
|
|
int nsegs, cluster;
|
|
|
|
|
|
|
|
nsegs = 0;
|
2010-12-02 01:41:49 +07:00
|
|
|
cluster = blk_queue_cluster(q);
|
2008-01-29 20:04:06 +07:00
|
|
|
|
2014-02-08 03:53:46 +07:00
|
|
|
if (bio->bi_rw & REQ_DISCARD) {
|
|
|
|
/*
|
|
|
|
* This is a hack - drivers should be neither modifying the
|
|
|
|
* biovec, nor relying on bi_vcnt - but because of
|
|
|
|
* blk_add_request_payload(), a discard bio may or may not have
|
|
|
|
* a payload we need to set up here (thank you Christoph) and
|
|
|
|
* bi_vcnt is really the only way of telling if we need to.
|
|
|
|
*/
|
|
|
|
|
|
|
|
if (bio->bi_vcnt)
|
|
|
|
goto single_segment;
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (bio->bi_rw & REQ_WRITE_SAME) {
|
|
|
|
single_segment:
|
|
|
|
*sg = sglist;
|
|
|
|
bvec = bio_iovec(bio);
|
|
|
|
sg_set_page(*sg, bvec.bv_page, bvec.bv_len, bvec.bv_offset);
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
for_each_bio(bio)
|
|
|
|
bio_for_each_segment(bvec, bio, iter)
|
|
|
|
__blk_segment_map_sg(q, &bvec, sglist, &bvprv, sg,
|
|
|
|
&nsegs, &cluster);
|
2008-01-29 20:04:06 +07:00
|
|
|
|
2014-02-08 03:53:46 +07:00
|
|
|
return nsegs;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* 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 scatterlist *sg = NULL;
|
|
|
|
int nsegs = 0;
|
|
|
|
|
|
|
|
if (rq->bio)
|
|
|
|
nsegs = __blk_bios_map_sg(q, rq->bio, sglist, &sg);
|
2008-04-11 17:56:52 +07:00
|
|
|
|
|
|
|
if (unlikely(rq->cmd_flags & REQ_COPY_USER) &&
|
2009-05-07 20:24:41 +07:00
|
|
|
(blk_rq_bytes(rq) & q->dma_pad_mask)) {
|
|
|
|
unsigned int pad_len =
|
|
|
|
(q->dma_pad_mask & ~blk_rq_bytes(rq)) + 1;
|
2008-04-11 17:56:52 +07:00
|
|
|
|
|
|
|
sg->length += pad_len;
|
|
|
|
rq->extra_len += pad_len;
|
|
|
|
}
|
|
|
|
|
2008-02-19 17:36:53 +07:00
|
|
|
if (q->dma_drain_size && q->dma_drain_needed(rq)) {
|
2010-08-07 23:20:39 +07:00
|
|
|
if (rq->cmd_flags & REQ_WRITE)
|
2008-02-19 17:36:55 +07:00
|
|
|
memset(q->dma_drain_buffer, 0, q->dma_drain_size);
|
|
|
|
|
2015-08-07 23:15:14 +07:00
|
|
|
sg_unmark_end(sg);
|
2008-01-29 20:04:06 +07:00
|
|
|
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++;
|
2008-03-04 17:17:11 +07:00
|
|
|
rq->extra_len += q->dma_drain_size;
|
2008-01-29 20:04:06 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
if (sg)
|
|
|
|
sg_mark_end(sg);
|
|
|
|
|
2015-11-24 09:35:31 +07:00
|
|
|
/*
|
|
|
|
* Something must have been wrong if the figured number of
|
|
|
|
* segment is bigger than number of req's physical segments
|
|
|
|
*/
|
|
|
|
WARN_ON(nsegs > rq->nr_phys_segments);
|
|
|
|
|
2008-01-29 20:04:06 +07:00
|
|
|
return nsegs;
|
|
|
|
}
|
|
|
|
EXPORT_SYMBOL(blk_rq_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);
|
|
|
|
|
2010-09-11 01:50:10 +07:00
|
|
|
if (req->nr_phys_segments + nr_phys_segs > queue_max_segments(q))
|
|
|
|
goto no_merge;
|
|
|
|
|
2014-09-27 06:20:06 +07:00
|
|
|
if (blk_integrity_merge_bio(q, req, bio) == false)
|
2010-09-11 01:50:10 +07:00
|
|
|
goto no_merge;
|
2008-01-29 20:04:06 +07:00
|
|
|
|
|
|
|
/*
|
|
|
|
* This will form the start of a new hw segment. Bump both
|
|
|
|
* counters.
|
|
|
|
*/
|
|
|
|
req->nr_phys_segments += nr_phys_segs;
|
|
|
|
return 1;
|
2010-09-11 01:50:10 +07:00
|
|
|
|
|
|
|
no_merge:
|
|
|
|
req->cmd_flags |= REQ_NOMERGE;
|
|
|
|
if (req == q->last_merge)
|
|
|
|
q->last_merge = NULL;
|
|
|
|
return 0;
|
2008-01-29 20:04:06 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
int ll_back_merge_fn(struct request_queue *q, struct request *req,
|
|
|
|
struct bio *bio)
|
|
|
|
{
|
2015-09-03 23:28:20 +07:00
|
|
|
if (req_gap_back_merge(req, bio))
|
|
|
|
return 0;
|
2015-09-11 22:03:04 +07:00
|
|
|
if (blk_integrity_rq(req) &&
|
|
|
|
integrity_req_gap_back_merge(req, bio))
|
|
|
|
return 0;
|
2012-09-18 23:19:26 +07:00
|
|
|
if (blk_rq_sectors(req) + bio_sectors(bio) >
|
|
|
|
blk_rq_get_max_sectors(req)) {
|
2008-01-29 20:04:06 +07:00
|
|
|
req->cmd_flags |= REQ_NOMERGE;
|
|
|
|
if (req == q->last_merge)
|
|
|
|
q->last_merge = NULL;
|
|
|
|
return 0;
|
|
|
|
}
|
2008-05-07 14:33:55 +07:00
|
|
|
if (!bio_flagged(req->biotail, BIO_SEG_VALID))
|
2008-01-29 20:04:06 +07:00
|
|
|
blk_recount_segments(q, req->biotail);
|
2008-05-07 14:33:55 +07:00
|
|
|
if (!bio_flagged(bio, BIO_SEG_VALID))
|
2008-01-29 20:04:06 +07:00
|
|
|
blk_recount_segments(q, bio);
|
|
|
|
|
|
|
|
return ll_new_hw_segment(q, req, bio);
|
|
|
|
}
|
|
|
|
|
2008-01-31 19:03:55 +07:00
|
|
|
int ll_front_merge_fn(struct request_queue *q, struct request *req,
|
2008-01-29 20:04:06 +07:00
|
|
|
struct bio *bio)
|
|
|
|
{
|
2015-09-03 23:28:20 +07:00
|
|
|
|
|
|
|
if (req_gap_front_merge(req, bio))
|
|
|
|
return 0;
|
2015-09-11 22:03:04 +07:00
|
|
|
if (blk_integrity_rq(req) &&
|
|
|
|
integrity_req_gap_front_merge(req, bio))
|
|
|
|
return 0;
|
2012-09-18 23:19:26 +07:00
|
|
|
if (blk_rq_sectors(req) + bio_sectors(bio) >
|
|
|
|
blk_rq_get_max_sectors(req)) {
|
2008-01-29 20:04:06 +07:00
|
|
|
req->cmd_flags |= REQ_NOMERGE;
|
|
|
|
if (req == q->last_merge)
|
|
|
|
q->last_merge = NULL;
|
|
|
|
return 0;
|
|
|
|
}
|
2008-05-07 14:33:55 +07:00
|
|
|
if (!bio_flagged(bio, BIO_SEG_VALID))
|
2008-01-29 20:04:06 +07:00
|
|
|
blk_recount_segments(q, bio);
|
2008-05-07 14:33:55 +07:00
|
|
|
if (!bio_flagged(req->bio, BIO_SEG_VALID))
|
2008-01-29 20:04:06 +07:00
|
|
|
blk_recount_segments(q, req->bio);
|
|
|
|
|
|
|
|
return ll_new_hw_segment(q, req, bio);
|
|
|
|
}
|
|
|
|
|
2013-10-30 01:11:47 +07:00
|
|
|
/*
|
|
|
|
* blk-mq uses req->special to carry normal driver per-request payload, it
|
|
|
|
* does not indicate a prepared command that we cannot merge with.
|
|
|
|
*/
|
|
|
|
static bool req_no_special_merge(struct request *req)
|
|
|
|
{
|
|
|
|
struct request_queue *q = req->q;
|
|
|
|
|
|
|
|
return !q->mq_ops && req->special;
|
|
|
|
}
|
|
|
|
|
2008-01-29 20:04:06 +07:00
|
|
|
static int ll_merge_requests_fn(struct request_queue *q, struct request *req,
|
|
|
|
struct request *next)
|
|
|
|
{
|
|
|
|
int total_phys_segments;
|
2008-10-13 19:19:05 +07:00
|
|
|
unsigned int seg_size =
|
|
|
|
req->biotail->bi_seg_back_size + next->bio->bi_seg_front_size;
|
2008-01-29 20:04:06 +07:00
|
|
|
|
|
|
|
/*
|
|
|
|
* First check if the either of the requests are re-queued
|
|
|
|
* requests. Can't merge them if they are.
|
|
|
|
*/
|
2013-10-30 01:11:47 +07:00
|
|
|
if (req_no_special_merge(req) || req_no_special_merge(next))
|
2008-01-29 20:04:06 +07:00
|
|
|
return 0;
|
|
|
|
|
2015-09-03 23:28:20 +07:00
|
|
|
if (req_gap_back_merge(req, next->bio))
|
2015-02-11 22:20:13 +07:00
|
|
|
return 0;
|
|
|
|
|
2008-01-29 20:04:06 +07:00
|
|
|
/*
|
|
|
|
* Will it become too large?
|
|
|
|
*/
|
2012-09-18 23:19:26 +07:00
|
|
|
if ((blk_rq_sectors(req) + blk_rq_sectors(next)) >
|
|
|
|
blk_rq_get_max_sectors(req))
|
2008-01-29 20:04:06 +07:00
|
|
|
return 0;
|
|
|
|
|
|
|
|
total_phys_segments = req->nr_phys_segments + next->nr_phys_segments;
|
2008-10-13 19:19:05 +07:00
|
|
|
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;
|
2008-01-29 20:04:06 +07:00
|
|
|
total_phys_segments--;
|
2008-10-13 19:19:05 +07:00
|
|
|
}
|
2008-01-29 20:04:06 +07:00
|
|
|
|
2010-02-26 12:20:39 +07:00
|
|
|
if (total_phys_segments > queue_max_segments(q))
|
2008-01-29 20:04:06 +07:00
|
|
|
return 0;
|
|
|
|
|
2014-09-27 06:20:06 +07:00
|
|
|
if (blk_integrity_merge_rq(q, req, next) == false)
|
2010-09-11 01:50:10 +07:00
|
|
|
return 0;
|
|
|
|
|
2008-01-29 20:04:06 +07:00
|
|
|
/* Merge is OK... */
|
|
|
|
req->nr_phys_segments = total_phys_segments;
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
2009-07-03 15:48:17 +07:00
|
|
|
/**
|
|
|
|
* 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;
|
|
|
|
}
|
|
|
|
|
2009-03-27 16:31:51 +07:00
|
|
|
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();
|
2011-01-05 22:57:38 +07:00
|
|
|
part = req->part;
|
2009-03-27 16:31:51 +07:00
|
|
|
|
|
|
|
part_round_stats(cpu, part);
|
block: Seperate read and write statistics of in_flight requests v2
Commit a9327cac440be4d8333bba975cbbf76045096275 added seperate read
and write statistics of in_flight requests. And exported the number
of read and write requests in progress seperately through sysfs.
But Corrado Zoccolo <czoccolo@gmail.com> reported getting strange
output from "iostat -kx 2". Global values for service time and
utilization were garbage. For interval values, utilization was always
100%, and service time is higher than normal.
So this was reverted by commit 0f78ab9899e9d6acb09d5465def618704255963b
The problem was in part_round_stats_single(), I missed the following:
if (now == part->stamp)
return;
- if (part->in_flight) {
+ if (part_in_flight(part)) {
__part_stat_add(cpu, part, time_in_queue,
part_in_flight(part) * (now - part->stamp));
__part_stat_add(cpu, part, io_ticks, (now - part->stamp));
With this chunk included, the reported regression gets fixed.
Signed-off-by: Nikanth Karthikesan <knikanth@suse.de>
--
Signed-off-by: Jens Axboe <jens.axboe@oracle.com>
2009-10-07 01:16:55 +07:00
|
|
|
part_dec_in_flight(part, rq_data_dir(req));
|
2009-03-27 16:31:51 +07:00
|
|
|
|
2011-01-07 14:43:37 +07:00
|
|
|
hd_struct_put(part);
|
2009-03-27 16:31:51 +07:00
|
|
|
part_stat_unlock();
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2008-01-29 20:04:06 +07:00
|
|
|
/*
|
|
|
|
* 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;
|
|
|
|
|
2012-09-18 23:19:26 +07:00
|
|
|
if (!blk_check_merge_flags(req->cmd_flags, next->cmd_flags))
|
|
|
|
return 0;
|
|
|
|
|
2008-01-29 20:04:06 +07:00
|
|
|
/*
|
|
|
|
* not contiguous
|
|
|
|
*/
|
2009-05-07 20:24:39 +07:00
|
|
|
if (blk_rq_pos(req) + blk_rq_sectors(req) != blk_rq_pos(next))
|
2008-01-29 20:04:06 +07:00
|
|
|
return 0;
|
|
|
|
|
|
|
|
if (rq_data_dir(req) != rq_data_dir(next)
|
|
|
|
|| req->rq_disk != next->rq_disk
|
2013-10-30 01:11:47 +07:00
|
|
|
|| req_no_special_merge(next))
|
2008-01-29 20:04:06 +07:00
|
|
|
return 0;
|
|
|
|
|
2012-09-18 23:19:27 +07:00
|
|
|
if (req->cmd_flags & REQ_WRITE_SAME &&
|
|
|
|
!blk_write_same_mergeable(req->bio, next->bio))
|
|
|
|
return 0;
|
|
|
|
|
2008-01-29 20:04:06 +07:00
|
|
|
/*
|
|
|
|
* 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;
|
|
|
|
|
2009-07-03 15:48:17 +07:00
|
|
|
/*
|
|
|
|
* 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);
|
|
|
|
}
|
|
|
|
|
2008-01-29 20:04:06 +07:00
|
|
|
/*
|
|
|
|
* 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;
|
|
|
|
|
2009-05-07 20:24:44 +07:00
|
|
|
req->__data_len += blk_rq_bytes(next);
|
2008-01-29 20:04:06 +07:00
|
|
|
|
|
|
|
elv_merge_requests(q, req, next);
|
|
|
|
|
2009-04-22 19:01:49 +07:00
|
|
|
/*
|
|
|
|
* 'next' is going away, so update stats accordingly
|
|
|
|
*/
|
|
|
|
blk_account_io_merge(next);
|
2008-01-29 20:04:06 +07:00
|
|
|
|
|
|
|
req->ioprio = ioprio_best(req->ioprio, next->ioprio);
|
2008-08-26 15:25:02 +07:00
|
|
|
if (blk_rq_cpu_valid(next))
|
|
|
|
req->cpu = next->cpu;
|
2008-01-29 20:04:06 +07:00
|
|
|
|
2009-03-24 18:35:07 +07:00
|
|
|
/* owner-ship of bio passed from next to req */
|
|
|
|
next->bio = NULL;
|
2008-01-29 20:04:06 +07:00
|
|
|
__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;
|
|
|
|
}
|
2011-03-21 16:14:27 +07:00
|
|
|
|
|
|
|
int blk_attempt_req_merge(struct request_queue *q, struct request *rq,
|
|
|
|
struct request *next)
|
|
|
|
{
|
|
|
|
return attempt_merge(q, rq, next);
|
|
|
|
}
|
2012-02-08 15:19:38 +07:00
|
|
|
|
|
|
|
bool blk_rq_merge_ok(struct request *rq, struct bio *bio)
|
|
|
|
{
|
2012-09-18 23:19:25 +07:00
|
|
|
if (!rq_mergeable(rq) || !bio_mergeable(bio))
|
2012-02-08 15:19:38 +07:00
|
|
|
return false;
|
|
|
|
|
2012-09-18 23:19:26 +07:00
|
|
|
if (!blk_check_merge_flags(rq->cmd_flags, bio->bi_rw))
|
|
|
|
return false;
|
|
|
|
|
2012-02-08 15:19:38 +07:00
|
|
|
/* 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 */
|
2013-10-30 01:11:47 +07:00
|
|
|
if (rq->rq_disk != bio->bi_bdev->bd_disk || req_no_special_merge(rq))
|
2012-02-08 15:19:38 +07:00
|
|
|
return false;
|
|
|
|
|
|
|
|
/* only merge integrity protected bio into ditto rq */
|
2014-09-27 06:20:06 +07:00
|
|
|
if (blk_integrity_merge_bio(rq->q, rq, bio) == false)
|
2012-02-08 15:19:38 +07:00
|
|
|
return false;
|
|
|
|
|
2012-09-18 23:19:27 +07:00
|
|
|
/* must be using the same buffer */
|
|
|
|
if (rq->cmd_flags & REQ_WRITE_SAME &&
|
|
|
|
!blk_write_same_mergeable(rq->bio, bio))
|
|
|
|
return false;
|
|
|
|
|
2012-02-08 15:19:38 +07:00
|
|
|
return true;
|
|
|
|
}
|
|
|
|
|
|
|
|
int blk_try_merge(struct request *rq, struct bio *bio)
|
|
|
|
{
|
2013-10-12 05:44:27 +07:00
|
|
|
if (blk_rq_pos(rq) + blk_rq_sectors(rq) == bio->bi_iter.bi_sector)
|
2012-02-08 15:19:38 +07:00
|
|
|
return ELEVATOR_BACK_MERGE;
|
2013-10-12 05:44:27 +07:00
|
|
|
else if (blk_rq_pos(rq) - bio_sectors(bio) == bio->bi_iter.bi_sector)
|
2012-02-08 15:19:38 +07:00
|
|
|
return ELEVATOR_FRONT_MERGE;
|
|
|
|
return ELEVATOR_NO_MERGE;
|
|
|
|
}
|