X-Git-Url: http://plrg.eecs.uci.edu/git/?a=blobdiff_plain;f=block%2Fblk-merge.c;h=7225511cf0b48b1644ef3c56bf77d1f58e4911ca;hb=c2b3e8740d8dcadbf90f8e9d947841cd2b7baa4a;hp=de5716d8e525969e7849767a775aabec9e4d8b96;hpb=01d85b9b2b6bec2b0773cf2afc58699dc4b052f8;p=firefly-linux-kernel-4.4.55.git

diff --git a/block/blk-merge.c b/block/blk-merge.c
index de5716d8e525..7225511cf0b4 100644
--- a/block/blk-merge.c
+++ b/block/blk-merge.c
@@ -68,6 +68,18 @@ static struct bio *blk_bio_write_same_split(struct request_queue *q,
 	return bio_split(bio, q->limits.max_write_same_sectors, GFP_NOIO, bs);
 }
 
+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;
+}
+
 static struct bio *blk_bio_segment_split(struct request_queue *q,
 					 struct bio *bio,
 					 struct bio_set *bs,
@@ -76,9 +88,32 @@ static struct bio *blk_bio_segment_split(struct request_queue *q,
 	struct bio_vec bv, bvprv, *bvprvp = NULL;
 	struct bvec_iter iter;
 	unsigned seg_size = 0, nsegs = 0, sectors = 0;
+	unsigned front_seg_size = bio->bi_seg_front_size;
+	bool do_split = true;
+	struct bio *new = NULL;
+	const unsigned max_sectors = get_max_io_size(q, bio);
+	unsigned bvecs = 0;
 
 	bio_for_each_segment(bv, bio, iter) {
-		if (sectors + (bv.bv_len >> 9) > queue_max_sectors(q))
+		/*
+		 * With arbitrary bio size, the incoming bio may be very
+		 * big. We have to split the bio into small bios so that
+		 * each holds at most BIO_MAX_PAGES bvecs because
+		 * bio_clone() can fail to allocate big bvecs.
+		 *
+		 * It should have been better to apply the limit per
+		 * request queue in which bio_clone() is involved,
+		 * instead of globally. The biggest blocker is the
+		 * bio_clone() in bio bounce.
+		 *
+		 * If bio is splitted by this reason, we should have
+		 * allowed to continue bios merging, but don't do
+		 * that now for making the change simple.
+		 *
+		 * TODO: deal with bio bounce's bio_clone() gracefully
+		 * and convert the global limit into per-queue limit.
+		 */
+		if (bvecs++ >= BIO_MAX_PAGES)
 			goto split;
 
 		/*
@@ -88,6 +123,21 @@ static struct bio *blk_bio_segment_split(struct request_queue *q,
 		if (bvprvp && bvec_gap_to_prev(q, bvprvp, bv.bv_offset))
 			goto split;
 
+		if (sectors + (bv.bv_len >> 9) > max_sectors) {
+			/*
+			 * Consider this a new segment if we're splitting in
+			 * the middle of this vector.
+			 */
+			if (nsegs < queue_max_segments(q) &&
+			    sectors < max_sectors) {
+				nsegs++;
+				sectors = max_sectors;
+			}
+			if (sectors)
+				goto split;
+			/* Make this single bvec as the 1st segment */
+		}
+
 		if (bvprvp && blk_queue_cluster(q)) {
 			if (seg_size + bv.bv_len > queue_max_segment_size(q))
 				goto new_segment;
@@ -98,8 +148,11 @@ static struct bio *blk_bio_segment_split(struct request_queue *q,
 
 			seg_size += bv.bv_len;
 			bvprv = bv;
-			bvprvp = &bv;
+			bvprvp = &bvprv;
 			sectors += bv.bv_len >> 9;
+
+			if (nsegs == 1 && seg_size > front_seg_size)
+				front_seg_size = seg_size;
 			continue;
 		}
 new_segment:
@@ -108,16 +161,29 @@ new_segment:
 
 		nsegs++;
 		bvprv = bv;
-		bvprvp = &bv;
+		bvprvp = &bvprv;
 		seg_size = bv.bv_len;
 		sectors += bv.bv_len >> 9;
+
+		if (nsegs == 1 && seg_size > front_seg_size)
+			front_seg_size = seg_size;
 	}
 
-	*segs = nsegs;
-	return NULL;
+	do_split = false;
 split:
 	*segs = nsegs;
-	return bio_split(bio, sectors, GFP_NOIO, bs);
+
+	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;
 }
 
 void blk_queue_split(struct request_queue *q, struct bio **bio,
@@ -412,6 +478,12 @@ int blk_rq_map_sg(struct request_queue *q, struct request *rq,
 	if (sg)
 		sg_mark_end(sg);
 
+	/*
+	 * 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);
+
 	return nsegs;
 }
 EXPORT_SYMBOL(blk_rq_map_sg);