Fix highbd inter prediction filter sse4 overwriting issue

Properly handle the case where the height is an integer multiple
of 4.

Change-Id: I11ac188c13f78db20902e2e333c60ce76ce837c5

vp10/common/x86/vp10_highbd_convolve_sse4.c

diff --git a/vp10/common/x86/vp10_highbd_convolve_sse4.c b/vp10/common/x86/vp10_highbd_convolve_sse4.c
index e828178..e8e4f77 100644
--- a/vp10/common/x86/vp10_highbd_convolve_sse4.c
+++ b/vp10/common/x86/vp10_highbd_convolve_sse4.c
@@ -276,6 +276,9 @@
     blkHeight--;
   }
 
+  if (blkResidu == 0)
+    return;
+
   for (col = 0; col < w; col += 4) {
     for (i = 0; i < 4; ++i) {
       highbd_filter_horiz(srcPtr, src_stride, verf, tapsNum, temp + (i * 4));