libvpx: Pull from upstream

source/libvpx/vp9/common/vp9_rtcd_defs.sh

Index: source/libvpx/vp9/common/vp9_rtcd_defs.sh
===================================================================
--- source/libvpx/vp9/common/vp9_rtcd_defs.sh	(revision 219822)
+++ source/libvpx/vp9/common/vp9_rtcd_defs.sh	(working copy)
@@ -7,9 +7,7 @@
 #include "vpx/vpx_integer.h"
 #include "vp9/common/vp9_enums.h"
 
-struct loop_filter_info;
 struct macroblockd;
-struct loop_filter_info;
 
 /* Encoder forward decls */
 struct macroblock;
@@ -22,8 +20,13 @@
 }
 forward_decls vp9_common_forward_decls
 
-[ $arch = "x86_64" ] && mmx_x86_64=mmx && sse2_x86_64=sse2
+# x86inc.asm doesn't work if pic is enabled on 32 bit platforms so no assembly.
+[ "$CONFIG_USE_X86INC" = "yes" ] && mmx_x86inc=mmx && sse_x86inc=sse &&
+  sse2_x86inc=sse2 && ssse3_x86inc=ssse3
 
+# this variable is for functions that are 64 bit only.
+[ $arch = "x86_64" ] && mmx_x86_64=mmx && sse2_x86_64=sse2 && ssse3_x86_64=ssse3
+
 #
 # Dequant
 #
@@ -43,160 +46,160 @@
 #
 # RECON
 #
-prototype void vp9_d27_predictor_4x4 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
-specialize vp9_d27_predictor_4x4
+prototype void vp9_d207_predictor_4x4 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
+specialize vp9_d207_predictor_4x4
 
-prototype void vp9_d45_predictor_4x4 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
-specialize vp9_d45_predictor_4x4
+prototype void vp9_d45_predictor_4x4 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
+specialize vp9_d45_predictor_4x4 $ssse3_x86inc
 
-prototype void vp9_d63_predictor_4x4 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
+prototype void vp9_d63_predictor_4x4 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
 specialize vp9_d63_predictor_4x4
 
-prototype void vp9_h_predictor_4x4 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
-specialize vp9_h_predictor_4x4 ssse3
+prototype void vp9_h_predictor_4x4 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
+specialize vp9_h_predictor_4x4 $ssse3_x86inc
 
-prototype void vp9_d117_predictor_4x4 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
+prototype void vp9_d117_predictor_4x4 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
 specialize vp9_d117_predictor_4x4
 
-prototype void vp9_d135_predictor_4x4 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
+prototype void vp9_d135_predictor_4x4 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
 specialize vp9_d135_predictor_4x4
 
-prototype void vp9_d153_predictor_4x4 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
+prototype void vp9_d153_predictor_4x4 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
 specialize vp9_d153_predictor_4x4
 
-prototype void vp9_v_predictor_4x4 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
-specialize vp9_v_predictor_4x4 sse
+prototype void vp9_v_predictor_4x4 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
+specialize vp9_v_predictor_4x4 $sse_x86inc
 
-prototype void vp9_tm_predictor_4x4 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
-specialize vp9_tm_predictor_4x4 sse
+prototype void vp9_tm_predictor_4x4 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
+specialize vp9_tm_predictor_4x4 $sse_x86inc
 
-prototype void vp9_dc_predictor_4x4 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
-specialize vp9_dc_predictor_4x4 sse
+prototype void vp9_dc_predictor_4x4 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
+specialize vp9_dc_predictor_4x4 $sse_x86inc
 
-prototype void vp9_dc_top_predictor_4x4 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
+prototype void vp9_dc_top_predictor_4x4 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
 specialize vp9_dc_top_predictor_4x4
 
-prototype void vp9_dc_left_predictor_4x4 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
+prototype void vp9_dc_left_predictor_4x4 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
 specialize vp9_dc_left_predictor_4x4
 
-prototype void vp9_dc_128_predictor_4x4 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
+prototype void vp9_dc_128_predictor_4x4 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
 specialize vp9_dc_128_predictor_4x4
 
-prototype void vp9_d27_predictor_8x8 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
-specialize vp9_d27_predictor_8x8
+prototype void vp9_d207_predictor_8x8 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
+specialize vp9_d207_predictor_8x8
 
-prototype void vp9_d45_predictor_8x8 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
-specialize vp9_d45_predictor_8x8
+prototype void vp9_d45_predictor_8x8 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
+specialize vp9_d45_predictor_8x8 $ssse3_x86inc
 
-prototype void vp9_d63_predictor_8x8 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
+prototype void vp9_d63_predictor_8x8 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
 specialize vp9_d63_predictor_8x8
 
-prototype void vp9_h_predictor_8x8 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
-specialize vp9_h_predictor_8x8 ssse3
+prototype void vp9_h_predictor_8x8 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
+specialize vp9_h_predictor_8x8 $ssse3_x86inc
 
-prototype void vp9_d117_predictor_8x8 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
+prototype void vp9_d117_predictor_8x8 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
 specialize vp9_d117_predictor_8x8
 
-prototype void vp9_d135_predictor_8x8 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
+prototype void vp9_d135_predictor_8x8 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
 specialize vp9_d135_predictor_8x8
 
-prototype void vp9_d153_predictor_8x8 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
+prototype void vp9_d153_predictor_8x8 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
 specialize vp9_d153_predictor_8x8
 
-prototype void vp9_v_predictor_8x8 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
-specialize vp9_v_predictor_8x8 sse
+prototype void vp9_v_predictor_8x8 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
+specialize vp9_v_predictor_8x8 $sse_x86inc
 
-prototype void vp9_tm_predictor_8x8 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
-specialize vp9_tm_predictor_8x8 sse2
+prototype void vp9_tm_predictor_8x8 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
+specialize vp9_tm_predictor_8x8 $sse2_x86inc
 
-prototype void vp9_dc_predictor_8x8 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
-specialize vp9_dc_predictor_8x8 sse
+prototype void vp9_dc_predictor_8x8 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
+specialize vp9_dc_predictor_8x8 $sse_x86inc
 
-prototype void vp9_dc_top_predictor_8x8 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
+prototype void vp9_dc_top_predictor_8x8 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
 specialize vp9_dc_top_predictor_8x8
 
-prototype void vp9_dc_left_predictor_8x8 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
+prototype void vp9_dc_left_predictor_8x8 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
 specialize vp9_dc_left_predictor_8x8
 
-prototype void vp9_dc_128_predictor_8x8 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
+prototype void vp9_dc_128_predictor_8x8 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
 specialize vp9_dc_128_predictor_8x8
 
-prototype void vp9_d27_predictor_16x16 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
-specialize vp9_d27_predictor_16x16
+prototype void vp9_d207_predictor_16x16 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
+specialize vp9_d207_predictor_16x16
 
-prototype void vp9_d45_predictor_16x16 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
-specialize vp9_d45_predictor_16x16
+prototype void vp9_d45_predictor_16x16 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
+specialize vp9_d45_predictor_16x16 $ssse3_x86inc
 
-prototype void vp9_d63_predictor_16x16 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
+prototype void vp9_d63_predictor_16x16 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
 specialize vp9_d63_predictor_16x16
 
-prototype void vp9_h_predictor_16x16 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
-specialize vp9_h_predictor_16x16 ssse3
+prototype void vp9_h_predictor_16x16 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
+specialize vp9_h_predictor_16x16 $ssse3_x86inc
 
-prototype void vp9_d117_predictor_16x16 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
+prototype void vp9_d117_predictor_16x16 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
 specialize vp9_d117_predictor_16x16
 
-prototype void vp9_d135_predictor_16x16 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
+prototype void vp9_d135_predictor_16x16 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
 specialize vp9_d135_predictor_16x16
 
-prototype void vp9_d153_predictor_16x16 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
+prototype void vp9_d153_predictor_16x16 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
 specialize vp9_d153_predictor_16x16
 
-prototype void vp9_v_predictor_16x16 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
-specialize vp9_v_predictor_16x16 sse2
+prototype void vp9_v_predictor_16x16 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
+specialize vp9_v_predictor_16x16 $sse2_x86inc
 
-prototype void vp9_tm_predictor_16x16 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
-specialize vp9_tm_predictor_16x16 sse2
+prototype void vp9_tm_predictor_16x16 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
+specialize vp9_tm_predictor_16x16 $sse2_x86inc
 
-prototype void vp9_dc_predictor_16x16 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
-specialize vp9_dc_predictor_16x16 sse2
+prototype void vp9_dc_predictor_16x16 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
+specialize vp9_dc_predictor_16x16 $sse2_x86inc
 
-prototype void vp9_dc_top_predictor_16x16 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
+prototype void vp9_dc_top_predictor_16x16 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
 specialize vp9_dc_top_predictor_16x16
 
-prototype void vp9_dc_left_predictor_16x16 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
+prototype void vp9_dc_left_predictor_16x16 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
 specialize vp9_dc_left_predictor_16x16
 
-prototype void vp9_dc_128_predictor_16x16 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
+prototype void vp9_dc_128_predictor_16x16 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
 specialize vp9_dc_128_predictor_16x16
 
-prototype void vp9_d27_predictor_32x32 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
-specialize vp9_d27_predictor_32x32
+prototype void vp9_d207_predictor_32x32 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
+specialize vp9_d207_predictor_32x32
 
-prototype void vp9_d45_predictor_32x32 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
-specialize vp9_d45_predictor_32x32
+prototype void vp9_d45_predictor_32x32 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
+specialize vp9_d45_predictor_32x32 $ssse3_x86inc
 
-prototype void vp9_d63_predictor_32x32 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
+prototype void vp9_d63_predictor_32x32 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
 specialize vp9_d63_predictor_32x32
 
-prototype void vp9_h_predictor_32x32 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
-specialize vp9_h_predictor_32x32 ssse3
+prototype void vp9_h_predictor_32x32 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
+specialize vp9_h_predictor_32x32 $ssse3 x86inc
 
-prototype void vp9_d117_predictor_32x32 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
+prototype void vp9_d117_predictor_32x32 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
 specialize vp9_d117_predictor_32x32
 
-prototype void vp9_d135_predictor_32x32 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
+prototype void vp9_d135_predictor_32x32 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
 specialize vp9_d135_predictor_32x32
 
-prototype void vp9_d153_predictor_32x32 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
+prototype void vp9_d153_predictor_32x32 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
 specialize vp9_d153_predictor_32x32
 
-prototype void vp9_v_predictor_32x32 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
-specialize vp9_v_predictor_32x32 sse2
+prototype void vp9_v_predictor_32x32 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
+specialize vp9_v_predictor_32x32 $sse2_x86inc
 
-prototype void vp9_tm_predictor_32x32 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
-specialize vp9_tm_predictor_32x32 sse2_x86_64
+prototype void vp9_tm_predictor_32x32 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
+specialize vp9_tm_predictor_32x32 $sse2_x86_64
 
-prototype void vp9_dc_predictor_32x32 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
-specialize vp9_dc_predictor_32x32 sse2
+prototype void vp9_dc_predictor_32x32 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
+specialize vp9_dc_predictor_32x32 $sse2_x86inc
 
-prototype void vp9_dc_top_predictor_32x32 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
+prototype void vp9_dc_top_predictor_32x32 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
 specialize vp9_dc_top_predictor_32x32
 
-prototype void vp9_dc_left_predictor_32x32 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
+prototype void vp9_dc_left_predictor_32x32 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
 specialize vp9_dc_left_predictor_32x32
 
-prototype void vp9_dc_128_predictor_32x32 "uint8_t *ypred_ptr, ptrdiff_t y_stride, uint8_t *yabove_row, uint8_t *yleft_col"
+prototype void vp9_dc_128_predictor_32x32 "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left"
 specialize vp9_dc_128_predictor_32x32
 
 if [ "$CONFIG_VP9_DECODER" = "yes" ]; then
@@ -214,7 +217,7 @@
 # Loopfilter
 #
 prototype void vp9_mb_lpf_vertical_edge_w "uint8_t *s, int pitch, const uint8_t *blimit, const uint8_t *limit, const uint8_t *thresh"
-specialize vp9_mb_lpf_vertical_edge_w sse2
+specialize vp9_mb_lpf_vertical_edge_w sse2 neon
 
 prototype void vp9_mbloop_filter_vertical_edge "uint8_t *s, int pitch, const uint8_t *blimit, const uint8_t *limit, const uint8_t *thresh, int count"
 specialize vp9_mbloop_filter_vertical_edge sse2 neon
@@ -223,7 +226,7 @@
 specialize vp9_loop_filter_vertical_edge mmx neon
 
 prototype void vp9_mb_lpf_horizontal_edge_w "uint8_t *s, int pitch, const uint8_t *blimit, const uint8_t *limit, const uint8_t *thresh, int count"
-specialize vp9_mb_lpf_horizontal_edge_w sse2
+specialize vp9_mb_lpf_horizontal_edge_w sse2 neon
 
 prototype void vp9_mbloop_filter_horizontal_edge "uint8_t *s, int pitch, const uint8_t *blimit, const uint8_t *limit, const uint8_t *thresh, int count"
 specialize vp9_mbloop_filter_horizontal_edge sse2 neon
@@ -265,10 +268,10 @@
 # Sub Pixel Filters
 #
 prototype void vp9_convolve_copy "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h"
-specialize vp9_convolve_copy sse2
+specialize vp9_convolve_copy $sse2_x86inc neon
 
 prototype void vp9_convolve_avg "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h"
-specialize vp9_convolve_avg sse2
+specialize vp9_convolve_avg $sse2_x86inc neon
 
 prototype void vp9_convolve8 "const uint8_t *src, ptrdiff_t src_stride, uint8_t *dst, ptrdiff_t dst_stride, const int16_t *filter_x, int x_step_q4, const int16_t *filter_y, int y_step_q4, int w, int h"
 specialize vp9_convolve8 ssse3 neon
@@ -292,38 +295,35 @@
 # dct
 #
 prototype void vp9_short_idct4x4_1_add "int16_t *input, uint8_t *dest, int dest_stride"
-specialize vp9_short_idct4x4_1_add sse2
+specialize vp9_short_idct4x4_1_add sse2 neon
 
 prototype void vp9_short_idct4x4_add "int16_t *input, uint8_t *dest, int dest_stride"
-specialize vp9_short_idct4x4_add sse2
+specialize vp9_short_idct4x4_add sse2 neon
 
+prototype void vp9_short_idct8x8_1_add "int16_t *input, uint8_t *dest, int dest_stride"
+specialize vp9_short_idct8x8_1_add sse2 neon
+
 prototype void vp9_short_idct8x8_add "int16_t *input, uint8_t *dest, int dest_stride"
 specialize vp9_short_idct8x8_add sse2 neon
 
 prototype void vp9_short_idct10_8x8_add "int16_t *input, uint8_t *dest, int dest_stride"
-specialize vp9_short_idct10_8x8_add sse2
+specialize vp9_short_idct10_8x8_add sse2 neon
 
-prototype void vp9_short_idct1_8x8 "int16_t *input, int16_t *output"
-specialize vp9_short_idct1_8x8
+prototype void vp9_short_idct16x16_1_add "int16_t *input, uint8_t *dest, int dest_stride"
+specialize vp9_short_idct16x16_1_add sse2
 
 prototype void vp9_short_idct16x16_add "int16_t *input, uint8_t *dest, int dest_stride"
-specialize vp9_short_idct16x16_add sse2
+specialize vp9_short_idct16x16_add sse2 neon
 
 prototype void vp9_short_idct10_16x16_add "int16_t *input, uint8_t *dest, int dest_stride"
-specialize vp9_short_idct10_16x16_add sse2
+specialize vp9_short_idct10_16x16_add sse2 neon
 
-prototype void vp9_short_idct1_16x16 "int16_t *input, int16_t *output"
-specialize vp9_short_idct1_16x16
-
 prototype void vp9_short_idct32x32_add "int16_t *input, uint8_t *dest, int dest_stride"
 specialize vp9_short_idct32x32_add sse2
 
 prototype void vp9_short_idct1_32x32 "int16_t *input, int16_t *output"
 specialize vp9_short_idct1_32x32
 
-prototype void vp9_short_idct10_32x32_add "int16_t *input, uint8_t *dest, int dest_stride"
-specialize vp9_short_idct10_32x32_add
-
 prototype void vp9_short_iht4x4_add "int16_t *input, uint8_t *dest, int dest_stride, int tx_type"
 specialize vp9_short_iht4x4_add sse2
 
@@ -343,12 +343,6 @@
 prototype void vp9_short_iwalsh4x4_add "int16_t *input, uint8_t *dest, int dest_stride"
 specialize vp9_short_iwalsh4x4_add
 
-prototype unsigned int vp9_sad32x3 "const uint8_t *src_ptr, int  src_stride, const uint8_t *ref_ptr, int ref_stride, int max_sad"
-specialize vp9_sad32x3
-
-prototype unsigned int vp9_sad3x32 "const uint8_t *src_ptr, int  src_stride, const uint8_t *ref_ptr, int ref_stride, int max_sad"
-specialize vp9_sad3x32
-
 #
 # Encoder functions below this point.
 #
@@ -357,217 +351,214 @@
 
 # variance
 prototype unsigned int vp9_variance32x16 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse"
-specialize vp9_variance32x16 sse2
+specialize vp9_variance32x16 $sse2_x86inc
 
 prototype unsigned int vp9_variance16x32 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse"
-specialize vp9_variance16x32 sse2
+specialize vp9_variance16x32 $sse2_x86inc
 
 prototype unsigned int vp9_variance64x32 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse"
-specialize vp9_variance64x32 sse2
+specialize vp9_variance64x32 $sse2_x86inc
 
 prototype unsigned int vp9_variance32x64 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse"
-specialize vp9_variance32x64 sse2
+specialize vp9_variance32x64 $sse2_x86inc
 
 prototype unsigned int vp9_variance32x32 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse"
-specialize vp9_variance32x32 sse2
+specialize vp9_variance32x32 $sse2_x86inc
 
 prototype unsigned int vp9_variance64x64 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse"
-specialize vp9_variance64x64 sse2
+specialize vp9_variance64x64 $sse2_x86inc
 
 prototype unsigned int vp9_variance16x16 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse"
-specialize vp9_variance16x16 mmx sse2
+specialize vp9_variance16x16 mmx $sse2_x86inc
 
 prototype unsigned int vp9_variance16x8 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse"
-specialize vp9_variance16x8 mmx sse2
+specialize vp9_variance16x8 mmx $sse2_x86inc
 
 prototype unsigned int vp9_variance8x16 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse"
-specialize vp9_variance8x16 mmx sse2
+specialize vp9_variance8x16 mmx $sse2_x86inc
 
 prototype unsigned int vp9_variance8x8 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse"
-specialize vp9_variance8x8 mmx sse2
+specialize vp9_variance8x8 mmx $sse2_x86inc
 
 prototype void vp9_get_sse_sum_8x8 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, int *sum"
 specialize vp9_get_sse_sum_8x8 sse2
 vp9_get_sse_sum_8x8_sse2=vp9_get8x8var_sse2
 
 prototype unsigned int vp9_variance8x4 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse"
-specialize vp9_variance8x4 sse2
+specialize vp9_variance8x4 $sse2_x86inc
 
 prototype unsigned int vp9_variance4x8 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse"
-specialize vp9_variance4x8 sse2
+specialize vp9_variance4x8 $sse2_x86inc
 
 prototype unsigned int vp9_variance4x4 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse"
-specialize vp9_variance4x4 mmx sse2
+specialize vp9_variance4x4 mmx $sse2_x86inc
 
 prototype unsigned int vp9_sub_pixel_variance64x64 "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse"
-specialize vp9_sub_pixel_variance64x64 sse2 ssse3
+specialize vp9_sub_pixel_variance64x64 $sse2_x86inc $ssse3_x86inc
 
 prototype unsigned int vp9_sub_pixel_avg_variance64x64 "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred"
-specialize vp9_sub_pixel_avg_variance64x64 sse2 ssse3
+specialize vp9_sub_pixel_avg_variance64x64 $sse2_x86inc $ssse3_x86inc
 
 prototype unsigned int vp9_sub_pixel_variance32x64 "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse"
-specialize vp9_sub_pixel_variance32x64 sse2 ssse3
+specialize vp9_sub_pixel_variance32x64 $sse2_x86inc $ssse3_x86inc
 
 prototype unsigned int vp9_sub_pixel_avg_variance32x64 "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred"
-specialize vp9_sub_pixel_avg_variance32x64 sse2 ssse3
+specialize vp9_sub_pixel_avg_variance32x64 $sse2_x86inc $ssse3_x86inc
 
 prototype unsigned int vp9_sub_pixel_variance64x32 "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse"
-specialize vp9_sub_pixel_variance64x32 sse2 ssse3
+specialize vp9_sub_pixel_variance64x32 $sse2_x86inc $ssse3_x86inc
 
 prototype unsigned int vp9_sub_pixel_avg_variance64x32 "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred"
-specialize vp9_sub_pixel_avg_variance64x32 sse2 ssse3
+specialize vp9_sub_pixel_avg_variance64x32 $sse2_x86inc $ssse3_x86inc
 
 prototype unsigned int vp9_sub_pixel_variance32x16 "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse"
-specialize vp9_sub_pixel_variance32x16 sse2 ssse3
+specialize vp9_sub_pixel_variance32x16 $sse2_x86inc $ssse3_x86inc
 
 prototype unsigned int vp9_sub_pixel_avg_variance32x16 "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred"
-specialize vp9_sub_pixel_avg_variance32x16 sse2 ssse3
+specialize vp9_sub_pixel_avg_variance32x16 $sse2_x86inc $ssse3_x86inc
 
 prototype unsigned int vp9_sub_pixel_variance16x32 "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse"
-specialize vp9_sub_pixel_variance16x32 sse2 ssse3
+specialize vp9_sub_pixel_variance16x32 $sse2_x86inc $ssse3_x86inc
 
 prototype unsigned int vp9_sub_pixel_avg_variance16x32 "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred"
-specialize vp9_sub_pixel_avg_variance16x32 sse2 ssse3
+specialize vp9_sub_pixel_avg_variance16x32 $sse2_x86inc $ssse3_x86inc
 
 prototype unsigned int vp9_sub_pixel_variance32x32 "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse"
-specialize vp9_sub_pixel_variance32x32 sse2 ssse3
+specialize vp9_sub_pixel_variance32x32 $sse2_x86inc $ssse3_x86inc
 
 prototype unsigned int vp9_sub_pixel_avg_variance32x32 "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred"
-specialize vp9_sub_pixel_avg_variance32x32 sse2 ssse3
+specialize vp9_sub_pixel_avg_variance32x32 $sse2_x86inc $ssse3_x86inc
 
 prototype unsigned int vp9_sub_pixel_variance16x16 "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse"
-specialize vp9_sub_pixel_variance16x16 sse2 ssse3
+specialize vp9_sub_pixel_variance16x16 $sse2_x86inc $ssse3_x86inc
 
 prototype unsigned int vp9_sub_pixel_avg_variance16x16 "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred"
-specialize vp9_sub_pixel_avg_variance16x16 sse2 ssse3
+specialize vp9_sub_pixel_avg_variance16x16 $sse2_x86inc $ssse3_x86inc
 
 prototype unsigned int vp9_sub_pixel_variance8x16 "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse"
-specialize vp9_sub_pixel_variance8x16 sse2 ssse3
+specialize vp9_sub_pixel_variance8x16 $sse2_x86inc $ssse3_x86inc
 
 prototype unsigned int vp9_sub_pixel_avg_variance8x16 "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred"
-specialize vp9_sub_pixel_avg_variance8x16 sse2 ssse3
+specialize vp9_sub_pixel_avg_variance8x16 $sse2_x86inc $ssse3_x86inc
 
 prototype unsigned int vp9_sub_pixel_variance16x8 "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse"
-specialize vp9_sub_pixel_variance16x8 sse2 ssse3
+specialize vp9_sub_pixel_variance16x8 $sse2_x86inc $ssse3_x86inc
 
 prototype unsigned int vp9_sub_pixel_avg_variance16x8 "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred"
-specialize vp9_sub_pixel_avg_variance16x8 sse2 ssse3
+specialize vp9_sub_pixel_avg_variance16x8 $sse2_x86inc $ssse3_x86inc
 
 prototype unsigned int vp9_sub_pixel_variance8x8 "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse"
-specialize vp9_sub_pixel_variance8x8 sse2 ssse3
+specialize vp9_sub_pixel_variance8x8 $sse2_x86inc $ssse3_x86inc
 
 prototype unsigned int vp9_sub_pixel_avg_variance8x8 "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred"
-specialize vp9_sub_pixel_avg_variance8x8 sse2 ssse3
+specialize vp9_sub_pixel_avg_variance8x8 $sse2_x86inc $ssse3_x86inc
 
 # TODO(jingning): need to convert 8x4/4x8 functions into mmx/sse form
 prototype unsigned int vp9_sub_pixel_variance8x4 "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse"
-specialize vp9_sub_pixel_variance8x4 sse2 ssse3
+specialize vp9_sub_pixel_variance8x4 $sse2_x86inc $ssse3_x86inc
 
 prototype unsigned int vp9_sub_pixel_avg_variance8x4 "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred"
-specialize vp9_sub_pixel_avg_variance8x4 sse2 ssse3
+specialize vp9_sub_pixel_avg_variance8x4 $sse2_x86inc $ssse3_x86inc
 
 prototype unsigned int vp9_sub_pixel_variance4x8 "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse"
-specialize vp9_sub_pixel_variance4x8 sse ssse3
+specialize vp9_sub_pixel_variance4x8 $sse_x86inc $ssse3_x86inc
 
 prototype unsigned int vp9_sub_pixel_avg_variance4x8 "const uint8_t *src_ptr, int source_stride, int xoffset, int yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred"
-specialize vp9_sub_pixel_avg_variance4x8 sse ssse3
+specialize vp9_sub_pixel_avg_variance4x8 $sse_x86inc $ssse3_x86inc
 
 prototype unsigned int vp9_sub_pixel_variance4x4 "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse"
-specialize vp9_sub_pixel_variance4x4 sse ssse3
+specialize vp9_sub_pixel_variance4x4 $sse_x86inc $ssse3_x86inc
 #vp9_sub_pixel_variance4x4_sse2=vp9_sub_pixel_variance4x4_wmt
 
 prototype unsigned int vp9_sub_pixel_avg_variance4x4 "const uint8_t *src_ptr, int source_stride, int xoffset, int  yoffset, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse, const uint8_t *second_pred"
-specialize vp9_sub_pixel_avg_variance4x4 sse ssse3
+specialize vp9_sub_pixel_avg_variance4x4 $sse_x86inc $ssse3_x86inc
 
 prototype unsigned int vp9_sad64x64 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int  ref_stride, unsigned int max_sad"
-specialize vp9_sad64x64 sse2
+specialize vp9_sad64x64 $sse2_x86inc
 
 prototype unsigned int vp9_sad32x64 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int max_sad"
-specialize vp9_sad32x64 sse2
+specialize vp9_sad32x64 $sse2_x86inc
 
 prototype unsigned int vp9_sad64x32 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int max_sad"
-specialize vp9_sad64x32 sse2
+specialize vp9_sad64x32 $sse2_x86inc
 
 prototype unsigned int vp9_sad32x16 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int max_sad"
-specialize vp9_sad32x16 sse2
+specialize vp9_sad32x16 $sse2_x86inc
 
 prototype unsigned int vp9_sad16x32 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int max_sad"
-specialize vp9_sad16x32 sse2
+specialize vp9_sad16x32 $sse2_x86inc
 
 prototype unsigned int vp9_sad32x32 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int  ref_stride, unsigned int max_sad"
-specialize vp9_sad32x32 sse2
+specialize vp9_sad32x32 $sse2_x86inc
 
 prototype unsigned int vp9_sad16x16 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int  ref_stride, unsigned int max_sad"
-specialize vp9_sad16x16 mmx sse2
+specialize vp9_sad16x16 mmx $sse2_x86inc
 
 prototype unsigned int vp9_sad16x8 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int  ref_stride, unsigned int max_sad"
-specialize vp9_sad16x8 mmx sse2
+specialize vp9_sad16x8 mmx $sse2_x86inc
 
 prototype unsigned int vp9_sad8x16 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int  ref_stride, unsigned int max_sad"
-specialize vp9_sad8x16 mmx sse2
+specialize vp9_sad8x16 mmx $sse2_x86inc
 
 prototype unsigned int vp9_sad8x8 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int  ref_stride, unsigned int max_sad"
-specialize vp9_sad8x8 mmx sse2
+specialize vp9_sad8x8 mmx $sse2_x86inc
 
 prototype unsigned int vp9_sad8x4 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int max_sad"
-specialize vp9_sad8x4 sse2
+specialize vp9_sad8x4 $sse2_x86inc
 
 prototype unsigned int vp9_sad4x8 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int max_sad"
-specialize vp9_sad4x8 sse
+specialize vp9_sad4x8 $sse_x86inc
 
 prototype unsigned int vp9_sad4x4 "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int  ref_stride, unsigned int max_sad"
-specialize vp9_sad4x4 mmx sse
+specialize vp9_sad4x4 mmx $sse_x86inc
 
 prototype unsigned int vp9_sad64x64_avg "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int  ref_stride, const uint8_t *second_pred, unsigned int max_sad"
-specialize vp9_sad64x64_avg sse2
+specialize vp9_sad64x64_avg $sse2_x86inc
 
 prototype unsigned int vp9_sad32x64_avg "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred, unsigned int max_sad"
-specialize vp9_sad32x64_avg sse2
+specialize vp9_sad32x64_avg $sse2_x86inc
 
 prototype unsigned int vp9_sad64x32_avg "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred, unsigned int max_sad"
-specialize vp9_sad64x32_avg sse2
+specialize vp9_sad64x32_avg $sse2_x86inc
 
 prototype unsigned int vp9_sad32x16_avg "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred, unsigned int max_sad"
-specialize vp9_sad32x16_avg sse2
+specialize vp9_sad32x16_avg $sse2_x86inc
 
 prototype unsigned int vp9_sad16x32_avg "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred, unsigned int max_sad"
-specialize vp9_sad16x32_avg sse2
+specialize vp9_sad16x32_avg $sse2_x86inc
 
 prototype unsigned int vp9_sad32x32_avg "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int  ref_stride, const uint8_t *second_pred, unsigned int max_sad"
-specialize vp9_sad32x32_avg sse2
+specialize vp9_sad32x32_avg $sse2_x86inc
 
 prototype unsigned int vp9_sad16x16_avg "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int  ref_stride, const uint8_t *second_pred, unsigned int max_sad"
-specialize vp9_sad16x16_avg sse2
+specialize vp9_sad16x16_avg $sse2_x86inc
 
 prototype unsigned int vp9_sad16x8_avg "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int  ref_stride, const uint8_t *second_pred, unsigned int max_sad"
-specialize vp9_sad16x8_avg sse2
+specialize vp9_sad16x8_avg $sse2_x86inc
 
 prototype unsigned int vp9_sad8x16_avg "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int  ref_stride, const uint8_t *second_pred, unsigned int max_sad"
-specialize vp9_sad8x16_avg sse2
+specialize vp9_sad8x16_avg $sse2_x86inc
 
 prototype unsigned int vp9_sad8x8_avg "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int  ref_stride, const uint8_t *second_pred, unsigned int max_sad"
-specialize vp9_sad8x8_avg sse2
+specialize vp9_sad8x8_avg $sse2_x86inc
 
 prototype unsigned int vp9_sad8x4_avg "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred, unsigned int max_sad"
-specialize vp9_sad8x4_avg sse2
+specialize vp9_sad8x4_avg $sse2_x86inc
 
 prototype unsigned int vp9_sad4x8_avg "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, const uint8_t *second_pred, unsigned int max_sad"
-specialize vp9_sad4x8_avg sse
+specialize vp9_sad4x8_avg $sse_x86inc
 
 prototype unsigned int vp9_sad4x4_avg "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int  ref_stride, const uint8_t *second_pred, unsigned int max_sad"
-specialize vp9_sad4x4_avg sse
+specialize vp9_sad4x4_avg $sse_x86inc
 
 prototype unsigned int vp9_variance_halfpixvar16x16_h "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse"
-specialize vp9_variance_halfpixvar16x16_h sse2
-vp9_variance_halfpixvar16x16_h_sse2=vp9_variance_halfpixvar16x16_h_wmt
+specialize vp9_variance_halfpixvar16x16_h $sse2_x86inc
 
 prototype unsigned int vp9_variance_halfpixvar16x16_v "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse"
-specialize vp9_variance_halfpixvar16x16_v sse2
-vp9_variance_halfpixvar16x16_v_sse2=vp9_variance_halfpixvar16x16_v_wmt
+specialize vp9_variance_halfpixvar16x16_v $sse2_x86inc
 
 prototype unsigned int vp9_variance_halfpixvar16x16_hv "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse"
-specialize vp9_variance_halfpixvar16x16_hv sse2
-vp9_variance_halfpixvar16x16_hv_sse2=vp9_variance_halfpixvar16x16_hv_wmt
+specialize vp9_variance_halfpixvar16x16_hv $sse2_x86inc
 
 prototype unsigned int vp9_variance_halfpixvar64x64_h "const uint8_t *src_ptr, int source_stride, const uint8_t *ref_ptr, int ref_stride, unsigned int *sse"
 specialize vp9_variance_halfpixvar64x64_h
@@ -679,8 +670,7 @@
 #specialize vp9_sub_pixel_mse16x16 sse2 mmx
 
 prototype unsigned int vp9_mse16x16 "const uint8_t *src_ptr, int  source_stride, const uint8_t *ref_ptr, int  recon_stride, unsigned int *sse"
-specialize vp9_mse16x16 mmx sse2
-vp9_mse16x16_sse2=vp9_mse16x16_wmt
+specialize vp9_mse16x16 mmx $sse2_x86inc
 
 prototype unsigned int vp9_mse8x16 "const uint8_t *src_ptr, int  source_stride, const uint8_t *ref_ptr, int  recon_stride, unsigned int *sse"
 specialize vp9_mse8x16
@@ -702,30 +692,26 @@
 # ENCODEMB INVOKE
 
 prototype int64_t vp9_block_error "int16_t *coeff, int16_t *dqcoeff, intptr_t block_size, int64_t *ssz"
-specialize vp9_block_error sse2
+specialize vp9_block_error $sse2_x86inc
 
 prototype void vp9_subtract_block "int rows, int cols, int16_t *diff_ptr, ptrdiff_t diff_stride, const uint8_t *src_ptr, ptrdiff_t src_stride, const uint8_t *pred_ptr, ptrdiff_t pred_stride"
-specialize vp9_subtract_block sse2
+specialize vp9_subtract_block $sse2_x86inc
 
-[ $arch = "x86_64" ] && ssse3_x86_64=ssse3
-
 prototype void vp9_quantize_b "int16_t *coeff_ptr, intptr_t n_coeffs, int skip_block, int16_t *zbin_ptr, int16_t *round_ptr, int16_t *quant_ptr, int16_t *quant_shift_ptr, int16_t *qcoeff_ptr, int16_t *dqcoeff_ptr, int16_t *dequant_ptr, int zbin_oq_value, uint16_t *eob_ptr, const int16_t *scan, const int16_t *iscan"
 specialize vp9_quantize_b $ssse3_x86_64
 
 prototype void vp9_quantize_b_32x32 "int16_t *coeff_ptr, intptr_t n_coeffs, int skip_block, int16_t *zbin_ptr, int16_t *round_ptr, int16_t *quant_ptr, int16_t *quant_shift_ptr, int16_t *qcoeff_ptr, int16_t *dqcoeff_ptr, int16_t *dequant_ptr, int zbin_oq_value, uint16_t *eob_ptr, const int16_t *scan, const int16_t *iscan"
-specialize vp9_quantize_b_32x32 $ssse3_x86_64
+specialize vp9_quantize_b_32x32
 
 #
 # Structured Similarity (SSIM)
 #
 if [ "$CONFIG_INTERNAL_STATS" = "yes" ]; then
-    [ $arch = "x86_64" ] && sse2_on_x86_64=sse2
-
     prototype void vp9_ssim_parms_8x8 "uint8_t *s, int sp, uint8_t *r, int rp, unsigned long *sum_s, unsigned long *sum_r, unsigned long *sum_sq_s, unsigned long *sum_sq_r, unsigned long *sum_sxr"
-    specialize vp9_ssim_parms_8x8 $sse2_on_x86_64
+    specialize vp9_ssim_parms_8x8 $sse2_x86_64
 
     prototype void vp9_ssim_parms_16x16 "uint8_t *s, int sp, uint8_t *r, int rp, unsigned long *sum_s, unsigned long *sum_r, unsigned long *sum_sq_s, unsigned long *sum_sq_r, unsigned long *sum_sxr"
-    specialize vp9_ssim_parms_16x16 $sse2_on_x86_64
+    specialize vp9_ssim_parms_16x16 $sse2_x86_64
 fi
 
 # fdct functions
@@ -748,7 +734,7 @@
 specialize vp9_short_fdct8x4 sse2
 
 prototype void vp9_short_fdct32x32 "int16_t *InputData, int16_t *OutputData, int pitch"
-specialize vp9_short_fdct32x32
+specialize vp9_short_fdct32x32 sse2
 
 prototype void vp9_short_fdct32x32_rd "int16_t *InputData, int16_t *OutputData, int pitch"
 specialize vp9_short_fdct32x32_rd sse2