[PATCHv5 3/9] blk-integrity: properly account for segments

[Keith Busch]

From: Keith Busch <kbusch at kernel.org>

Both types of merging when integrity data is used are miscounting the
segments:

Merging two requests wasn't accounting for the new segment count, so add
the "next" segment count to the first on a successful merge to ensure
this value is accurate.

Merging a bio into an existing request was double counting the bio's
segments, even if the merge failed later on. Move the segment accounting
to the end when the merge is successful.

Reviewed-by: Christoph Hellwig <hch at lst.de>
Reviewed-by: Martin K. Petersen <martin.petersen at oracle.com>
Signed-off-by: Keith Busch <kbusch at kernel.org>
---
 block/blk-integrity.c | 2 --
 block/blk-merge.c     | 4 ++++
 2 files changed, 4 insertions(+), 2 deletions(-)

diff --git a/block/blk-integrity.c b/block/blk-integrity.c
index 010decc892eaa..afd101555d3cb 100644
--- a/block/blk-integrity.c
+++ b/block/blk-integrity.c
@@ -153,8 +153,6 @@ bool blk_integrity_merge_bio(struct request_queue *q, struct request *req,
 	    q->limits.max_integrity_segments)
 		return false;
 
-	req->nr_integrity_segments += nr_integrity_segs;
-
 	return true;
 }
 
diff --git a/block/blk-merge.c b/block/blk-merge.c
index 56769c4bcd799..ad763ec313b6a 100644
--- a/block/blk-merge.c
+++ b/block/blk-merge.c
@@ -639,6 +639,9 @@ static inline int ll_new_hw_segment(struct request *req, struct bio *bio,
 	 * counters.
 	 */
 	req->nr_phys_segments += nr_phys_segs;
+	if (bio_integrity(bio))
+		req->nr_integrity_segments += blk_rq_count_integrity_sg(req->q,
+									bio);
 	return 1;
 
 no_merge:
@@ -731,6 +734,7 @@ static int ll_merge_requests_fn(struct request_queue *q, struct request *req,
 
 	/* Merge is OK... */
 	req->nr_phys_segments = total_phys_segments;
+	req->nr_integrity_segments += next->nr_integrity_segments;
 	return 1;
 }
 
-- 
2.43.5