Simplify ref_yv12[2] in av1_joint_motion_search()

A careful analysis shows that the ref_yv12[2] array in
av1_joint_motion_search() can be replaced by a single variable
orig_yv12, like the orig_yv12 variable in
av1_compound_single_motion_search().

Only xd->plane[plane].pre[0] needs to be backed up and restored, so a
single variable (orig_yv12) suffices.

Change-Id: I46a0ac488678ce617f74580ed12eddc1e56b6f39

av1/encoder/motion_search_facade.c

diff --git a/av1/encoder/motion_search_facade.c b/av1/encoder/motion_search_facade.c
index 63bfe79..5d60d9c 100644
--- a/av1/encoder/motion_search_facade.c
+++ b/av1/encoder/motion_search_facade.c
@@ -588,7 +588,6 @@
   // Allow joint search multiple times iteratively for each reference frame
   // and break out of the search loop if it couldn't find a better mv.
   for (ite = 0; ite < (2 * joint_me_num_refine_iter); ite++) {
-    struct buf_2d ref_yv12[2];
     int bestsme = INT_MAX;
     int id = ite % 2;  // Even iterations search in the first reference frame,
                        // odd iterations search in the second. The predictor
@@ -628,13 +627,10 @@
                    cm->width == scaled_ref_frame[1]->y_crop_width &&
                        cm->height == scaled_ref_frame[1]->y_crop_height));
 
-    // Initialize based on (possibly scaled) prediction buffers.
-    ref_yv12[0] = xd->plane[plane].pre[0];
-    ref_yv12[1] = xd->plane[plane].pre[1];
-
     av1_init_inter_params(&inter_pred_params, pw, ph, mi_row * MI_SIZE,
                           mi_col * MI_SIZE, 0, 0, xd->bd, is_cur_buf_hbd(xd), 0,
-                          &cm->sf_identity, &ref_yv12[!id], interp_filters);
+                          &cm->sf_identity, &xd->plane[plane].pre[!id],
+                          interp_filters);
     inter_pred_params.conv_params = get_conv_params(0, 0, xd->bd);
 
     // Since we have scaled the reference frames to match the size of the
@@ -643,7 +639,11 @@
                                       &inter_pred_params);
 
     // Do full-pixel compound motion search on the current reference frame.
-    if (id) xd->plane[plane].pre[0] = ref_yv12[id];
+    struct buf_2d orig_yv12;
+    if (id) {
+      orig_yv12 = xd->plane[plane].pre[0];
+      xd->plane[plane].pre[0] = xd->plane[plane].pre[id];
+    }
 
     // Make motion search params
     FULLPEL_MOTION_SEARCH_PARAMS full_ms_params;
@@ -680,7 +680,7 @@
                            allow_second_mv;
 
     // Restore the pointer to the first (possibly scaled) prediction buffer.
-    if (id) xd->plane[plane].pre[0] = ref_yv12[0];
+    if (id) xd->plane[plane].pre[0] = orig_yv12;
 
     for (ref = 0; ref < 2; ++ref) {
       if (scaled_ref_frame[ref]) {
@@ -688,13 +688,14 @@
         for (int i = 0; i < num_planes; i++) {
           xd->plane[i].pre[ref] = backup_yv12[ref][i];
         }
-        // Re-initialize based on unscaled prediction buffers.
-        ref_yv12[ref] = xd->plane[plane].pre[ref];
       }
     }
 
     // Do sub-pixel compound motion search on the current reference frame.
-    if (id) xd->plane[plane].pre[0] = ref_yv12[id];
+    if (id) {
+      orig_yv12 = xd->plane[plane].pre[0];
+      xd->plane[plane].pre[0] = xd->plane[plane].pre[id];
+    }
 
     if (cpi->common.features.cur_frame_force_integer_mv) {
       convert_fullmv_to_mv(&best_mv);
@@ -733,7 +734,7 @@
     }
 
     // Restore the pointer to the first prediction buffer.
-    if (id) xd->plane[plane].pre[0] = ref_yv12[0];
+    if (id) xd->plane[plane].pre[0] = orig_yv12;
     if (bestsme < last_besterr[id]) {
       cur_mv[id] = best_mv;
       last_besterr[id] = bestsme;