block: add support for limiting gaps in SG lists

Another restriction inherited for NVMe - those devices don't support
SG lists that have "gaps" in them. Gaps refers to cases where the
previous SG entry doesn't end on a page boundary. For NVMe, all SG
entries must start at offset 0 (except the first) and end on a page
boundary (except the last).

Signed-off-by: Jens Axboe <axboe@fb.com>

4 files changed, 28 insertions, 0 deletions

diff --git a/block/bio.c b/block/bio.c
index 8c2e55e39a1b..0ec61c9e536c 100644
--- a/block/bio.c
+++ b/block/bio.c
@@ -746,6 +746,14 @@ static int __bio_add_page(struct request_queue *q, struct bio *bio, struct page
 
 			goto done;
 		}
+
+		/*
+		 * If the queue doesn't support SG gaps and adding this
+		 * offset would create a gap, disallow it.
+		 */
+		if (q->queue_flags & (1 << QUEUE_FLAG_SG_GAPS) &&
+		    bvec_gap_to_prev(prev, offset))
+			return 0;
 	}
 
 	if (bio->bi_vcnt >= bio->bi_max_vecs)
diff --git a/block/blk-merge.c b/block/blk-merge.c
index b3bf0df0f4c2..54535831f1e1 100644
--- a/block/blk-merge.c
+++ b/block/blk-merge.c
@@ -568,6 +568,8 @@ int blk_attempt_req_merge(struct request_queue *q, struct request *rq,
 
 bool blk_rq_merge_ok(struct request *rq, struct bio *bio)
 {
+	struct request_queue *q = rq->q;
+
 	if (!rq_mergeable(rq) || !bio_mergeable(bio))
 		return false;
 
@@ -591,6 +593,14 @@ bool blk_rq_merge_ok(struct request *rq, struct bio *bio)
 	    !blk_write_same_mergeable(rq->bio, bio))
 		return false;
 
+	if (q->queue_flags & (1 << QUEUE_FLAG_SG_GAPS)) {
+		struct bio_vec *bprev;
+
+		bprev = &rq->biotail->bi_io_vec[bio->bi_vcnt - 1];
+		if (bvec_gap_to_prev(bprev, bio->bi_io_vec[0].bv_offset))
+			return false;
+	}
+
 	return true;
 }
 
diff --git a/include/linux/bio.h b/include/linux/bio.h
index f91decbca96b..d2633ee099d9 100644
--- a/include/linux/bio.h
+++ b/include/linux/bio.h
@@ -186,6 +186,15 @@ static inline void *bio_data(struct bio *bio)
 #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)))
 
+/*
+ * Check if adding a bio_vec after bprv with offset would create a gap in
+ * the SG list. Most drivers don't care about this, but some do.
+ */
+static inline bool bvec_gap_to_prev(struct bio_vec *bprv, unsigned int offset)
+{
+	return offset || ((bprv->bv_offset + bprv->bv_len) & (PAGE_SIZE - 1));
+}
+
 #define bio_io_error(bio) bio_endio((bio), -EIO)
 
 /*
diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
index 713f8b62b435..8699bcf5f099 100644
--- a/include/linux/blkdev.h
+++ b/include/linux/blkdev.h
@@ -512,6 +512,7 @@ struct request_queue {
 #define QUEUE_FLAG_DEAD        19	/* queue tear-down finished */
 #define QUEUE_FLAG_INIT_DONE   20	/* queue is initialized */
 #define QUEUE_FLAG_NO_SG_MERGE 21	/* don't attempt to merge SG segments*/
+#define QUEUE_FLAG_SG_GAPS     22	/* queue doesn't support SG gaps */
 
 #define QUEUE_FLAG_DEFAULT	((1 << QUEUE_FLAG_IO_STAT) |		\
 				 (1 << QUEUE_FLAG_STACKABLE)	|	\