[PATCH 6/6] Multipage bvecs
From: Kent Overstreet
Date: Wed Sep 25 2013 - 16:23:23 EST
Convert merging to bio_add_page()/blk_max_segment()
Signed-off-by: Kent Overstreet <kmo@xxxxxxxxxxxxx>
---
arch/x86/include/asm/io.h | 7 --
block/blk-merge.c | 227 ++++++++++++----------------------------------
drivers/xen/biomerge.c | 11 ---
fs/bio.c | 14 ++-
include/linux/bio.h | 25 -----
include/linux/blk_types.h | 7 --
6 files changed, 65 insertions(+), 226 deletions(-)
diff --git a/arch/x86/include/asm/io.h b/arch/x86/include/asm/io.h
index 4799845..9060a25 100644
--- a/arch/x86/include/asm/io.h
+++ b/arch/x86/include/asm/io.h
@@ -333,17 +333,10 @@ extern bool is_early_ioremap_ptep(pte_t *ptep);
#ifdef CONFIG_XEN
#include <xen/xen.h>
-struct bio_vec;
extern bool xen_page_phys_mergeable(const struct page *p1,
const struct page *p2);
-extern bool xen_biovec_phys_mergeable(const struct bio_vec *vec1,
- const struct bio_vec *vec2);
-
-#define BIOVEC_PHYS_MERGEABLE(vec1, vec2) \
- (__BIOVEC_PHYS_MERGEABLE(vec1, vec2) && \
- (!xen_domain() || xen_biovec_phys_mergeable(vec1, vec2)))
#endif /* CONFIG_XEN */
#define IO_SPACE_LIMIT 0xffff
diff --git a/block/blk-merge.c b/block/blk-merge.c
index c9c837b..c356e11 100644
--- a/block/blk-merge.c
+++ b/block/blk-merge.c
@@ -116,10 +116,8 @@ static struct bio *blk_bio_write_same_split(struct request_queue *q,
struct bio *blk_bio_segment_split(struct request_queue *q, struct bio *bio,
struct bio_set *bs)
{
- struct bio_vec bv, bvprv;
- struct bvec_iter iter;
- unsigned seg_size = 0, nsegs = 0;
- int prev = 0;
+ struct bvec_iter iter = bio->bi_iter;
+ unsigned nsegs = 0;
if (bio->bi_rw & REQ_DISCARD)
return blk_bio_discard_split(q, bio, bs);
@@ -127,21 +125,10 @@ struct bio *blk_bio_segment_split(struct request_queue *q, struct bio *bio,
if (bio->bi_rw & REQ_WRITE_SAME)
return blk_bio_write_same_split(q, bio, bs);
- bio_for_each_segment(bv, bio, iter) {
- if (prev && blk_queue_cluster(q)) {
- 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;
- prev = 1;
- continue;
- }
-new_segment:
+ while (iter.bi_size) {
+ struct bio_vec bv = bio_iter_iovec(bio, iter);
+ unsigned nbytes = blk_max_segment(q, &bv);
+
if (nsegs == queue_max_segments(q)) {
struct bio *split;
@@ -159,9 +146,7 @@ new_segment:
}
nsegs++;
- bvprv = bv;
- prev = 1;
- seg_size = bv.bv_len;
+ bio_advance_iter(bio, &iter, nbytes);
}
return NULL;
@@ -170,59 +155,29 @@ new_segment:
static unsigned int __blk_recalc_rq_segments(struct request_queue *q,
struct bio *bio)
{
- struct bio_vec bv, bvprv;
- int cluster, prev = 0;
- unsigned int seg_size, nr_phys_segs;
- struct bio *fbio, *bbio;
- struct bvec_iter iter;
+ unsigned nr_phys_segs = 0;
- 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, iter) {
- if (prev && 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;
- prev = 1;
- continue;
- }
-new_segment:
- if (nr_phys_segs == 1 && seg_size >
- fbio->bi_seg_front_size)
- fbio->bi_seg_front_size = seg_size;
+ struct bvec_iter iter = bio->bi_iter;
+
+ while (iter.bi_size) {
+ struct bio_vec bv = bio_iter_iovec(bio, iter);
+ unsigned nbytes = blk_max_segment(q, &bv);
nr_phys_segs++;
- bvprv = bv;
- prev = 1;
- seg_size = bv.bv_len;
+ bio_advance_iter(bio, &iter, nbytes);
}
- 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);
+ if (rq->bio)
+ rq->nr_phys_segments = __blk_recalc_rq_segments(rq->q, rq->bio);
+ else
+ rq->nr_phys_segments = 0;
}
void blk_recount_segments(struct request_queue *q, struct bio *bio)
@@ -236,82 +191,43 @@ void blk_recount_segments(struct request_queue *q, struct bio *bio)
}
EXPORT_SYMBOL(blk_recount_segments);
-static int blk_phys_contig_segment(struct request_queue *q, struct bio *bio,
- struct bio *nxt)
+static int __blk_segment_map_sg(struct request_queue *q, struct bio *bio,
+ struct scatterlist *sglist,
+ struct scatterlist **sg)
{
- struct bio_vec end_bv, nxt_bv;
- struct bvec_iter iter;
-
- 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;
-
- 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))
- return 0;
+ unsigned nsegs = 0;
- /*
- * bio and nxt are contiguous in memory; check if the queue allows
- * these two to be merged into one
- */
- if (BIOVEC_SEG_BOUNDARY(q, &end_bv, &nxt_bv))
- return 1;
-
- return 0;
-}
-
-static inline 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)
-{
+ for_each_bio(bio) {
+ struct bvec_iter iter = bio->bi_iter;
+
+ while (iter.bi_size) {
+ struct bio_vec bv = bio_iter_iovec(bio, iter);
+ unsigned nbytes = blk_max_segment(q, &bv);
+
+ 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_unmark_end(*sg);
+ *sg = sg_next(*sg);
+ }
- int nbytes = bvec->bv_len;
-
- if (*sg && *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_unmark_end(*sg);
- *sg = sg_next(*sg);
+ sg_set_page(*sg, bv.bv_page, nbytes, bv.bv_offset);
+ nsegs++;
+ bio_advance_iter(bio, &iter, nbytes);
}
-
- sg_set_page(*sg, bvec->bv_page, nbytes, bvec->bv_offset);
- (*nsegs)++;
}
- *bvprv = *bvec;
+
+ return nsegs;
}
/*
@@ -321,22 +237,13 @@ new_segment:
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);
+ struct scatterlist *sg = NULL;
+ unsigned nsegs;
/*
* for each bio in rq
*/
- sg = NULL;
- rq_for_each_segment(bvec, rq, iter) {
- __blk_segment_map_sg(q, &bvec, sglist, &bvprv, &sg,
- &nsegs, &cluster);
- } /* segments in rq */
+ nsegs = __blk_segment_map_sg(q, rq->bio, sglist, &sg);
if (unlikely(rq->cmd_flags & REQ_COPY_USER) &&
@@ -383,24 +290,17 @@ EXPORT_SYMBOL(blk_rq_map_sg);
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;
- struct bvec_iter iter;
-
- nsegs = 0;
- cluster = blk_queue_cluster(q);
+ struct scatterlist *sg = NULL;
+ struct bio *next = bio->bi_next;
+ unsigned nsegs;
- sg = NULL;
- bio_for_each_segment(bvec, bio, iter) {
- __blk_segment_map_sg(q, &bvec, sglist, &bvprv, &sg,
- &nsegs, &cluster);
- } /* segments in bio */
+ bio->bi_next = NULL;
+ nsegs = __blk_segment_map_sg(q, bio, sglist, &sg);
+ bio->bi_next = next;
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);
@@ -471,8 +371,6 @@ 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
@@ -489,13 +387,6 @@ static int ll_merge_requests_fn(struct request_queue *q, struct request *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;
diff --git a/drivers/xen/biomerge.c b/drivers/xen/biomerge.c
index d9fb598..9934266 100644
--- a/drivers/xen/biomerge.c
+++ b/drivers/xen/biomerge.c
@@ -10,14 +10,3 @@ bool xen_page_phys_mergeable(const struct page *p1, const struct page *p2)
return mfn1 + 1 == mfn2;
}
-
-bool xen_biovec_phys_mergeable(const struct bio_vec *vec1,
- const struct bio_vec *vec2)
-{
- unsigned long mfn1 = pfn_to_mfn(page_to_pfn(vec1->bv_page));
- unsigned long mfn2 = pfn_to_mfn(page_to_pfn(vec2->bv_page));
-
- return __BIOVEC_PHYS_MERGEABLE(vec1, vec2) &&
- ((mfn1 == mfn2) || ((mfn1+1) == mfn2));
-}
-EXPORT_SYMBOL(xen_biovec_phys_mergeable);
diff --git a/fs/bio.c b/fs/bio.c
index da8aa81..93aee5c 100644
--- a/fs/bio.c
+++ b/fs/bio.c
@@ -678,8 +678,8 @@ int bio_add_page(struct bio *bio, struct page *page,
if (bio->bi_vcnt > 0) {
bv = &bio->bi_io_vec[bio->bi_vcnt - 1];
- if (page == bv->bv_page &&
- offset == bv->bv_offset + bv->bv_len) {
+ if (bvec_to_phys(bv) + bv->bv_len ==
+ page_to_phys(page) + offset) {
bv->bv_len += len;
goto done;
}
@@ -688,12 +688,10 @@ int bio_add_page(struct bio *bio, struct page *page,
if (bio->bi_vcnt >= bio->bi_max_vecs)
return 0;
- bv = &bio->bi_io_vec[bio->bi_vcnt];
- bv->bv_page = page;
- bv->bv_len = len;
- bv->bv_offset = offset;
-
- bio->bi_vcnt++;
+ bv = &bio->bi_io_vec[bio->bi_vcnt++];
+ bv->bv_page = page;
+ bv->bv_len = len;
+ bv->bv_offset = offset;
done:
bio->bi_iter.bi_size += len;
return len;
diff --git a/include/linux/bio.h b/include/linux/bio.h
index f6f0e99..3d9473d 100644
--- a/include/linux/bio.h
+++ b/include/linux/bio.h
@@ -175,27 +175,6 @@ static inline void *bio_data(struct bio *bio)
#define __bio_kunmap_atomic(addr) kunmap_atomic(addr)
-/*
- * merge helpers etc
- */
-
-/* Default implementation of BIOVEC_PHYS_MERGEABLE */
-#define __BIOVEC_PHYS_MERGEABLE(vec1, vec2) \
- ((bvec_to_phys((vec1)) + (vec1)->bv_len) == bvec_to_phys((vec2)))
-
-/*
- * allow arch override, for eg virtualized architectures (put in asm/io.h)
- */
-#ifndef BIOVEC_PHYS_MERGEABLE
-#define BIOVEC_PHYS_MERGEABLE(vec1, vec2) \
- __BIOVEC_PHYS_MERGEABLE(vec1, vec2)
-#endif
-
-#define __BIO_SEG_BOUNDARY(addr1, addr2, mask) \
- (((addr1) | (mask)) == (((addr2) - 1) | (mask)))
-#define BIOVEC_SEG_BOUNDARY(q, b1, b2) \
- __BIO_SEG_BOUNDARY(bvec_to_phys((b1)), bvec_to_phys((b2)) + (b2)->bv_len, queue_segment_boundary((q)))
-
#define bio_io_error(bio) bio_endio((bio), -EIO)
static inline void bvec_iter_advance(struct bio_vec *bv, struct bvec_iter *iter,
@@ -652,10 +631,6 @@ struct biovec_slab {
#define bip_for_each_page(bvl, bip, iter) \
__bip_for_each(bvl, bip, iter, bvec_iter_page_bytes)
-#define bio_for_each_integrity_vec(_bvl, _bio, _iter) \
- for_each_bio(_bio) \
- bip_for_each_segment(_bvl, _bio->bi_integrity, _iter)
-
#define bio_integrity(bio) (bio->bi_integrity != NULL)
extern struct bio_integrity_payload *bio_integrity_alloc(struct bio *, gfp_t, unsigned int);
diff --git a/include/linux/blk_types.h b/include/linux/blk_types.h
index 8fca6e3..4aebb73 100644
--- a/include/linux/blk_types.h
+++ b/include/linux/blk_types.h
@@ -57,13 +57,6 @@ struct bio {
*/
unsigned int bi_phys_segments;
- /*
- * To keep track of the max segment size, we account for the
- * sizes of the first and last mergeable segments in this bio.
- */
- unsigned int bi_seg_front_size;
- unsigned int bi_seg_back_size;
-
atomic_t bi_remaining;
bio_end_io_t *bi_end_io;
--
1.8.4.rc3
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/