NVMe: Handle physical merging of bvec entries
authorMatthew Wilcox <matthew.r.wilcox@intel.com>
Thu, 10 Feb 2011 18:55:39 +0000 (13:55 -0500)
committerMatthew Wilcox <matthew.r.wilcox@intel.com>
Fri, 4 Nov 2011 19:52:57 +0000 (15:52 -0400)
In order to not overrun the sg array, we have to merge physically
contiguous pages into a single sg entry.

Signed-off-by: Matthew Wilcox <matthew.r.wilcox@intel.com>
drivers/block/nvme.c

index bfdca3a..c0e84b6 100644 (file)
@@ -392,19 +392,25 @@ static struct nvme_prps *nvme_setup_prps(struct nvme_dev *dev,
 static int nvme_map_bio(struct device *dev, struct nvme_bio *nbio,
                struct bio *bio, enum dma_data_direction dma_dir, int psegs)
 {
-       struct bio_vec *bvec;
-       struct scatterlist *sg = nbio->sg;
-       int i, nsegs;
+       struct bio_vec *bvec, *bvprv = NULL;
+       struct scatterlist *sg = NULL;
+       int i, nsegs = 0;
 
-       sg_init_table(sg, psegs);
+       sg_init_table(nbio->sg, psegs);
        bio_for_each_segment(bvec, bio, i) {
-               sg_set_page(sg, bvec->bv_page, bvec->bv_len, bvec->bv_offset);
-               sg++;
-               /* XXX: handle non-mergable here */
-               nsegs++;
+               if (bvprv && BIOVEC_PHYS_MERGEABLE(bvprv, bvec)) {
+                       sg->length += bvec->bv_len;
+               } else {
+                       /* Check bvprv && offset == 0 */
+                       sg = sg ? sg + 1 : nbio->sg;
+                       sg_set_page(sg, bvec->bv_page, bvec->bv_len,
+                                                       bvec->bv_offset);
+                       nsegs++;
+               }
+               bvprv = bvec;
        }
        nbio->nents = nsegs;
-
+       sg_mark_end(sg);
        return dma_map_sg(dev, nbio->sg, nbio->nents, dma_dir);
 }