[PATCH] incorrect block layer segment accounting

There's a long standing bug in blk_recount_segments(). Clustering means
physical segment colascing, not hardware segment colascing. This
basically means that we are mapping more segments than here than the bio
+ requests contains, and this causes a bug in the SCSI layer for host
adapters that have CLUSTERING enabled.

This patch makes sure that we are clustering physical segments
correctly, and correctly accounting hardware segments. Please apply.

drivers/block/ll_rw_blk.c

@@ -673,31 +673,22 @@ void blk_recount_segments(request_queue_t *q, struct bio *bio)
 	seg_size = nr_phys_segs = nr_hw_segs = 0;
 	bio_for_each_segment(bv, bio, i) {
 		if (bvprv && cluster) {
-			int phys, seg;
-
-			if (seg_size + bv->bv_len > q->max_segment_size) {
-				nr_phys_segs++;
+			if (seg_size + bv->bv_len > q->max_segment_size)
 				goto new_segment;
-			}
-
-			phys = BIOVEC_PHYS_MERGEABLE(bvprv, bv);
-			seg = BIOVEC_SEG_BOUNDARY(q, bvprv, bv);
-			if (!phys || !seg)
-				nr_phys_segs++;
-			if (!seg)
+			if (!BIOVEC_PHYS_MERGEABLE(bvprv, bv))
 				goto new_segment;
-
-			if (!BIOVEC_VIRT_MERGEABLE(bvprv, bv))
+			if (!BIOVEC_SEG_BOUNDARY(q, bvprv, bv))
 				goto new_segment;
 
 			seg_size += bv->bv_len;
 			bvprv = bv;
 			continue;
-		} else {
-			nr_phys_segs++;
 		}
 new_segment:
-		nr_hw_segs++;
+		if (!bvprv || !BIOVEC_VIRT_MERGEABLE(bvprv, bv))
+			nr_hw_segs++;
+
+		nr_phys_segs++;
 		bvprv = bv;
 		seg_size = bv->bv_len;
 	}