libvpx: Pull from upstream

source/libvpx/vp9/common/vp9_rtcd_defs.pl

 sub vp9_common_forward_decls() {
 print <<EOF
 /*
  * VP9
  */
 
 #include "vpx/vpx_integer.h"
 #include "vp9/common/vp9_common.h"
 #include "vp9/common/vp9_enums.h"
 
@@ skipping 198 matching lines @@
 add_proto qw/void vp9_dc_left_predictor_32x32/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
 specialize qw/vp9_dc_left_predictor_32x32/, "$sse2_x86inc";
 
 add_proto qw/void vp9_dc_128_predictor_32x32/, "uint8_t *dst, ptrdiff_t y_stride, const uint8_t *above, const uint8_t *left";
 specialize qw/vp9_dc_128_predictor_32x32/, "$sse2_x86inc";
 
 #
 # Loopfilter
 #
 add_proto qw/void vp9_lpf_vertical_16/, "uint8_t *s, int pitch, const uint8_t *blimit, const uint8_t *limit, const uint8_t *thresh";
-specialize qw/vp9_lpf_vertical_16 sse2 neon_asm dspr2/;
+specialize qw/vp9_lpf_vertical_16 sse2 neon_asm dspr2 msa/;
 $vp9_lpf_vertical_16_neon_asm=vp9_lpf_vertical_16_neon;
 
 add_proto qw/void vp9_lpf_vertical_16_dual/, "uint8_t *s, int pitch, const uint8_t *blimit, const uint8_t *limit, const uint8_t *thresh";
-specialize qw/vp9_lpf_vertical_16_dual sse2 neon_asm dspr2/;
+specialize qw/vp9_lpf_vertical_16_dual sse2 neon_asm dspr2 msa/;
 $vp9_lpf_vertical_16_dual_neon_asm=vp9_lpf_vertical_16_dual_neon;
 
 add_proto qw/void vp9_lpf_vertical_8/, "uint8_t *s, int pitch, const uint8_t *blimit, const uint8_t *limit, const uint8_t *thresh, int count";
-specialize qw/vp9_lpf_vertical_8 sse2 neon_asm dspr2/;
+specialize qw/vp9_lpf_vertical_8 sse2 neon_asm dspr2 msa/;
 $vp9_lpf_vertical_8_neon_asm=vp9_lpf_vertical_8_neon;
 
 add_proto qw/void vp9_lpf_vertical_8_dual/, "uint8_t *s, int pitch, const uint8_t *blimit0, const uint8_t *limit0, const uint8_t *thresh0, const uint8_t *blimit1, const uint8_t *limit1, const uint8_t *thresh1";
-specialize qw/vp9_lpf_vertical_8_dual sse2 neon_asm dspr2/;
+specialize qw/vp9_lpf_vertical_8_dual sse2 neon_asm dspr2 msa/;
 $vp9_lpf_vertical_8_dual_neon_asm=vp9_lpf_vertical_8_dual_neon;
 
 add_proto qw/void vp9_lpf_vertical_4/, "uint8_t *s, int pitch, const uint8_t *blimit, const uint8_t *limit, const uint8_t *thresh, int count";
-specialize qw/vp9_lpf_vertical_4 mmx neon dspr2/;
+specialize qw/vp9_lpf_vertical_4 mmx neon dspr2 msa/;
 
 add_proto qw/void vp9_lpf_vertical_4_dual/, "uint8_t *s, int pitch, const uint8_t *blimit0, const uint8_t *limit0, const uint8_t *thresh0, const uint8_t *blimit1, const uint8_t *limit1, const uint8_t *thresh1";
-specialize qw/vp9_lpf_vertical_4_dual sse2 neon dspr2/;
+specialize qw/vp9_lpf_vertical_4_dual sse2 neon dspr2 msa/;
 
 add_proto qw/void vp9_lpf_horizontal_16/, "uint8_t *s, int pitch, const uint8_t *blimit, const uint8_t *limit, const uint8_t *thresh, int count";
-specialize qw/vp9_lpf_horizontal_16 sse2 avx2 neon_asm dspr2/;
+specialize qw/vp9_lpf_horizontal_16 sse2 avx2 neon_asm dspr2 msa/;
 $vp9_lpf_horizontal_16_neon_asm=vp9_lpf_horizontal_16_neon;
 
 add_proto qw/void vp9_lpf_horizontal_8/, "uint8_t *s, int pitch, const uint8_t *blimit, const uint8_t *limit, const uint8_t *thresh, int count";
-specialize qw/vp9_lpf_horizontal_8 sse2 neon_asm dspr2/;
+specialize qw/vp9_lpf_horizontal_8 sse2 neon_asm dspr2 msa/;
 $vp9_lpf_horizontal_8_neon_asm=vp9_lpf_horizontal_8_neon;
 
 add_proto qw/void vp9_lpf_horizontal_8_dual/, "uint8_t *s, int pitch, const uint8_t *blimit0, const uint8_t *limit0, const uint8_t *thresh0, const uint8_t *blimit1, const uint8_t *limit1, const uint8_t *thresh1";
-specialize qw/vp9_lpf_horizontal_8_dual sse2 neon_asm dspr2/;
+specialize qw/vp9_lpf_horizontal_8_dual sse2 neon_asm dspr2 msa/;
 $vp9_lpf_horizontal_8_dual_neon_asm=vp9_lpf_horizontal_8_dual_neon;
 
 add_proto qw/void vp9_lpf_horizontal_4/, "uint8_t *s, int pitch, const uint8_t *blimit, const uint8_t *limit, const uint8_t *thresh, int count";
-specialize qw/vp9_lpf_horizontal_4 mmx neon dspr2/;
+specialize qw/vp9_lpf_horizontal_4 mmx neon dspr2 msa/;
 
 add_proto qw/void vp9_lpf_horizontal_4_dual/, "uint8_t *s, int pitch, const uint8_t *blimit0, const uint8_t *limit0, const uint8_t *thresh0, const uint8_t *blimit1, const uint8_t *limit1, const uint8_t *thresh1";
-specialize qw/vp9_lpf_horizontal_4_dual sse2 neon dspr2/;
+specialize qw/vp9_lpf_horizontal_4_dual sse2 neon dspr2 msa/;
 
 #
 # post proc
 #
 if (vpx_config("CONFIG_VP9_POSTPROC") eq "yes") {
 add_proto qw/void vp9_mbpost_proc_down/, "uint8_t *dst, int pitch, int rows, int cols, int flimit";
 specialize qw/vp9_mbpost_proc_down sse2/;
 $vp9_mbpost_proc_down_sse2=vp9_mbpost_proc_down_xmm;
 
 add_proto qw/void vp9_mbpost_proc_across_ip/, "uint8_t *src, int pitch, int rows, int cols, int flimit";
@@ skipping 12 matching lines @@
 specialize qw/vp9_filter_by_weight16x16 sse2/;
 
 add_proto qw/void vp9_filter_by_weight8x8/, "const uint8_t *src, int src_stride, uint8_t *dst, int dst_stride, int src_weight";
 specialize qw/vp9_filter_by_weight8x8 sse2/;
 }
 
 #
 # Sub Pixel Filters
 #
 add_proto qw/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 qw/vp9_convolve_copy neon dspr2/, "$sse2_x86inc";
+specialize qw/vp9_convolve_copy neon dspr2 msa/, "$sse2_x86inc";
 
 add_proto qw/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 qw/vp9_convolve_avg neon dspr2/, "$sse2_x86inc";
+specialize qw/vp9_convolve_avg neon dspr2 msa/, "$sse2_x86inc";
 
 add_proto qw/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 qw/vp9_convolve8 sse2 ssse3 neon dspr2/, "$avx2_ssse3";
+specialize qw/vp9_convolve8 sse2 ssse3 neon dspr2 msa/, "$avx2_ssse3";
 
 add_proto qw/void vp9_convolve8_horiz/, "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 qw/vp9_convolve8_horiz sse2 ssse3 neon dspr2/, "$avx2_ssse3";
+specialize qw/vp9_convolve8_horiz sse2 ssse3 neon dspr2 msa/, "$avx2_ssse3";
 
 add_proto qw/void vp9_convolve8_vert/, "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 qw/vp9_convolve8_vert sse2 ssse3 neon dspr2/, "$avx2_ssse3";
+specialize qw/vp9_convolve8_vert sse2 ssse3 neon dspr2 msa/, "$avx2_ssse3";
 
 add_proto qw/void vp9_convolve8_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 qw/vp9_convolve8_avg sse2 ssse3 neon dspr2/;
+specialize qw/vp9_convolve8_avg sse2 ssse3 neon dspr2 msa/;
 
 add_proto qw/void vp9_convolve8_avg_horiz/, "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 qw/vp9_convolve8_avg_horiz sse2 ssse3 neon dspr2/;
+specialize qw/vp9_convolve8_avg_horiz sse2 ssse3 neon dspr2 msa/;
 
 add_proto qw/void vp9_convolve8_avg_vert/, "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 qw/vp9_convolve8_avg_vert sse2 ssse3 neon dspr2/;
+specialize qw/vp9_convolve8_avg_vert sse2 ssse3 neon dspr2 msa/;
 
 #
 # dct
 #
 if (vpx_config("CONFIG_VP9_HIGHBITDEPTH") eq "yes") {
   # Note as optimized versions of these functions are added we need to add a check to ensure
   # that when CONFIG_EMULATE_HARDWARE is on, it defaults to the C versions only.
   add_proto qw/void vp9_idct4x4_1_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
   specialize qw/vp9_idct4x4_1_add/;
 
@@ skipping 91 matching lines @@
 
     # dct and add
 
     add_proto qw/void vp9_iwht4x4_1_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
     specialize qw/vp9_iwht4x4_1_add/;
 
     add_proto qw/void vp9_iwht4x4_16_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
     specialize qw/vp9_iwht4x4_16_add/;
   } else {
     add_proto qw/void vp9_idct4x4_1_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
-    specialize qw/vp9_idct4x4_1_add sse2 neon dspr2/;
+    specialize qw/vp9_idct4x4_1_add sse2 neon dspr2 msa/;
 
     add_proto qw/void vp9_idct4x4_16_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
-    specialize qw/vp9_idct4x4_16_add sse2 neon dspr2/;
+    specialize qw/vp9_idct4x4_16_add sse2 neon dspr2 msa/;
 
     add_proto qw/void vp9_idct8x8_1_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
     specialize qw/vp9_idct8x8_1_add sse2 neon dspr2 msa/;
 
     add_proto qw/void vp9_idct8x8_64_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
     specialize qw/vp9_idct8x8_64_add sse2 neon dspr2 msa/, "$ssse3_x86_64";
 
     add_proto qw/void vp9_idct8x8_12_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
     specialize qw/vp9_idct8x8_12_add sse2 neon dspr2 msa/, "$ssse3_x86_64";
 
@@ skipping 11 matching lines @@
 
     add_proto qw/void vp9_idct32x32_34_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
     specialize qw/vp9_idct32x32_34_add sse2 neon_asm dspr2 msa/;
     #is this a typo?
     $vp9_idct32x32_34_add_neon_asm=vp9_idct32x32_1024_add_neon;
 
     add_proto qw/void vp9_idct32x32_1_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
     specialize qw/vp9_idct32x32_1_add sse2 neon dspr2 msa/;
 
     add_proto qw/void vp9_iht4x4_16_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type";
-    specialize qw/vp9_iht4x4_16_add sse2 neon dspr2/;
+    specialize qw/vp9_iht4x4_16_add sse2 neon dspr2 msa/;
 
     add_proto qw/void vp9_iht8x8_64_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride, int tx_type";
     specialize qw/vp9_iht8x8_64_add sse2 neon dspr2 msa/;
 
     add_proto qw/void vp9_iht16x16_256_add/, "const tran_low_t *input, uint8_t *output, int pitch, int tx_type";
     specialize qw/vp9_iht16x16_256_add sse2 dspr2 msa/;
 
     # dct and add
 
     add_proto qw/void vp9_iwht4x4_1_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
-    specialize qw/vp9_iwht4x4_1_add/;
+    specialize qw/vp9_iwht4x4_1_add msa/;
 
     add_proto qw/void vp9_iwht4x4_16_add/, "const tran_low_t *input, uint8_t *dest, int dest_stride";
-    specialize qw/vp9_iwht4x4_16_add/;
+    specialize qw/vp9_iwht4x4_16_add msa/;
   }
 }
 
 # High bitdepth functions
 if (vpx_config("CONFIG_VP9_HIGHBITDEPTH") eq "yes") {
   #
   # Intra prediction
   #
   add_proto qw/void vp9_highbd_d207_predictor_4x4/, "uint16_t *dst, ptrdiff_t y_stride, const uint16_t *above, const uint16_t *left, int bd";
   specialize qw/vp9_highbd_d207_predictor_4x4/;
@@ skipping 901 matching lines @@
 
   add_proto qw/void vp9_highbd_temporal_filter_apply/, "uint8_t *frame1, unsigned int stride, uint8_t *frame2, unsigned int block_width, unsigned int block_height, int strength, int filter_weight, unsigned int *accumulator, uint16_t *count";
   specialize qw/vp9_highbd_temporal_filter_apply/;
 
 }
 # End vp9_high encoder functions
 
 }
 # end encoder functions
 1;