libvpx: Pull from upstream

source/libvpx/vp9/encoder/x86/vp9_subpel_variance.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 252 matching lines @@
 
   add                srcq, src_strideq
   add                dstq, dst_strideq
 %else ; %1 < 16
   movh                 m0, [srcq]
   movh                 m2, [srcq+src_strideq]
 %if %2 == 1 ; avg
 %if mmsize == 16
   movhps               m2, [srcq+src_strideq*2]
 %else ; mmsize == 8
+%if %1 == 4
+  movh                 m1, [srcq+src_strideq*2]
+  punpckldq            m2, m1
+%else
   punpckldq            m2, [srcq+src_strideq*2]
 %endif
+%endif
   movh                 m1, [dstq]
 %if mmsize == 16
   movlhps              m0, m2
 %else ; mmsize == 8
   punpckldq            m0, m2
 %endif
   movh                 m3, [dstq+dst_strideq]
   pavgb                m0, m2
   punpcklbw            m1, m5
   pavgb                m0, [secq]
@@ skipping 250 matching lines @@
   add                srcq, src_strideq
   pavgb                m0, m3
 .x_half_y_half_loop:
   movh                 m2, [srcq]
   movh                 m3, [srcq+1]
 %if %2 == 1 ; avg
 %if mmsize == 16
   movhps               m2, [srcq+src_strideq]
   movhps               m3, [srcq+src_strideq+1]
 %else
+%if %1 == 4
+  movh                 m1, [srcq+src_strideq]
+  punpckldq            m2, m1
+  movh                 m1, [srcq+src_strideq+1]
+  punpckldq            m3, m1
+%else
   punpckldq            m2, [srcq+src_strideq]
   punpckldq            m3, [srcq+src_strideq+1]
 %endif
+%endif
   pavgb                m2, m3
 %if mmsize == 16
   movlhps              m0, m2
   movhlps              m4, m2
 %else ; mmsize == 8
   punpckldq            m0, m2
   pshufw               m4, m2, 0xe
 %endif
   movh                 m1, [dstq]
   pavgb                m0, m2
@@ skipping 721 matching lines @@
 SUBPEL_VARIANCE  4, 1
 INIT_XMM sse2
 SUBPEL_VARIANCE  8, 1
 SUBPEL_VARIANCE 16, 1
 
 INIT_MMX ssse3
 SUBPEL_VARIANCE  4, 1
 INIT_XMM ssse3
 SUBPEL_VARIANCE  8, 1
 SUBPEL_VARIANCE 16, 1