xen/blkback: Change fast_flush_area to xen_blkbk_unmap, and tweak xen_blk_map_seg.
authorKonrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Fri, 15 Apr 2011 15:50:34 +0000 (11:50 -0400)
committerKonrad Rzeszutek Wilk <konrad.wilk@oracle.com>
Fri, 15 Apr 2011 15:50:34 +0000 (11:50 -0400)
The previous name ('fast_flush_area') had nothing to do with what it
does right now. Changing the names so that the code dealing with
mapping pages in and out of the guest is called xen_blkbk_[map|unmap].

Signed-off-by: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>
drivers/xen/blkback/blkback.c

index 9598e0f..c645c83 100644 (file)
@@ -249,7 +249,7 @@ struct seg_buf {
  * Unmap the grant references, and also remove the M2P over-rides
  * used in the 'pending_req'.
 */
-static void fast_flush_area(struct pending_req *req)
+static void xen_blkbk_unmap(struct pending_req *req)
 {
        struct gnttab_unmap_grant_ref unmap[BLKIF_MAX_SEGMENTS_PER_REQUEST];
        unsigned int i, invcount = 0;
@@ -282,8 +282,8 @@ static void fast_flush_area(struct pending_req *req)
                }
        }
 }
-static int xen_blk_map_buf(struct blkif_request *req, struct pending_req *pending_req,
-                          struct seg_buf seg[])
+static int xen_blkbk_map(struct blkif_request *req, struct pending_req *pending_req,
+                        struct seg_buf seg[])
 {
        struct gnttab_map_grant_ref map[BLKIF_MAX_SEGMENTS_PER_REQUEST];
        int i;
@@ -361,7 +361,7 @@ static void __end_block_io_op(struct pending_req *pending_req, int error)
         * the proper response on the ring.
         */
        if (atomic_dec_and_test(&pending_req->pendcnt)) {
-               fast_flush_area(pending_req);
+               xen_blkbk_unmap(pending_req);
                make_response(pending_req->blkif, pending_req->id,
                              pending_req->operation, pending_req->status);
                blkif_put(pending_req->blkif);
@@ -540,9 +540,9 @@ static void dispatch_rw_block_io(struct blkif_st *blkif,
        /* If we have failed at this point, we need to undo the M2P override,
         * set gnttab_set_unmap_op on all of the grant references and perform
         * the hypercall to unmap the grants - that is all done in
-        * fast_flush_area.
+        * xen_blkbk_unmap.
         */
-       if (xen_blk_map_buf(req, pending_req, seg))
+       if (xen_blkbk_map(req, pending_req, seg))
                goto fail_flush;
 
        /* This corresponding blkif_put is done in __end_block_io_op */
@@ -606,7 +606,7 @@ static void dispatch_rw_block_io(struct blkif_st *blkif,
        return;
 
  fail_flush:
-       fast_flush_area(pending_req);
+       xen_blkbk_unmap(pending_req);
  fail_response:
        /* Haven't submitted any bio's yet. */
        make_response(blkif, req->id, req->operation, BLKIF_RSP_ERROR);