libvpx: Pull from upstream

source/libvpx/vp9/common/x86/vp9_copy_sse2.asm

 ;
 ;  Copyright (c) 2010 The WebM project authors. All Rights Reserved.
 ;
 ;  Use of this source code is governed by a BSD-style license
 ;  that can be found in the LICENSE file in the root of the source
 ;  tree. An additional intellectual property rights grant can be found
 ;  in the file PATENTS.  All contributing project authors may
 ;  be found in the AUTHORS file in the root of the source tree.
 ;
 
@@ skipping 115 matching lines @@
   mov                    r4d, dword hm
   lea                    r5q, [src_strideq*3]
   lea                    r6q, [dst_strideq*3]
 .loop4:
   movh                    m0, [srcq]
   movh                    m1, [srcq+src_strideq]
   movh                    m2, [srcq+src_strideq*2]
   movh                    m3, [srcq+r5q]
   lea                   srcq, [srcq+src_strideq*4]
 %ifidn %1, avg
-  pavgb                   m0, [dstq]
-  pavgb                   m1, [dstq+dst_strideq]
-  pavgb                   m2, [dstq+dst_strideq*2]
-  pavgb                   m3, [dstq+r6q]
+  movh                    m4, [dstq]
+  movh                    m5, [dstq+dst_strideq]
+  movh                    m6, [dstq+dst_strideq*2]
+  movh                    m7, [dstq+r6q]
+  pavgb                   m0, m4
+  pavgb                   m1, m5
+  pavgb                   m2, m6
+  pavgb                   m3, m7
 %endif
   movh  [dstq              ], m0
   movh  [dstq+dst_strideq  ], m1
   movh  [dstq+dst_strideq*2], m2
   movh  [dstq+r6q          ], m3
   lea                   dstq, [dstq+dst_strideq*4]
   sub                    r4d, 4
   jnz .loop4
   RET
 %endmacro
 
 convolve_fn copy
 convolve_fn avg