Add MSA optimized ARGB Attenuate/RGB565/Shuffle/Shader/Gray/Sepia row functions

include/libyuv/row.h

 /*
  *  Copyright 2011 The LibYuv 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 375 matching lines @@
 #define HAS_I422TORGB24ROW_MSA
 #define HAS_ARGBTORGB24ROW_MSA
 #define HAS_ARGBTORAWROW_MSA
 #define HAS_ARGBTORGB565ROW_MSA
 #define HAS_ARGBTOARGB1555ROW_MSA
 #define HAS_ARGBTOARGB4444ROW_MSA
 #define HAS_ARGBTOUV444ROW_MSA
 #define HAS_ARGBMULTIPLYROW_MSA
 #define HAS_ARGBADDROW_MSA
 #define HAS_ARGBSUBTRACTROW_MSA
+#define HAS_ARGBATTENUATEROW_MSA
+#define HAS_ARGBTORGB565DITHERROW_MSA
+#define HAS_ARGBSHUFFLEROW_MSA
+#define HAS_ARGBSHADEROW_MSA
+#define HAS_ARGBGRAYROW_MSA
+#define HAS_ARGBSEPIAROW_MSA
 #endif
 
 #if defined(_MSC_VER) && !defined(__CLR_VER) && !defined(__clang__)
 #if defined(VISUALC_HAS_AVX2)
 #define SIMD_ALIGNED(var) __declspec(align(32)) var
 #else
 #define SIMD_ALIGNED(var) __declspec(align(16)) var
 #endif
 typedef __declspec(align(16)) int16 vec16[8];
 typedef __declspec(align(16)) int32 vec32[4];
@@ skipping 779 matching lines @@
                           const uint8* shuffler,
                           int width);
 void ARGBShuffleRow_AVX2(const uint8* src_argb,
                          uint8* dst_argb,
                          const uint8* shuffler,
                          int width);
 void ARGBShuffleRow_NEON(const uint8* src_argb,
                          uint8* dst_argb,
                          const uint8* shuffler,
                          int width);
+void ARGBShuffleRow_MSA(const uint8* src_argb,
+                        uint8* dst_argb,
+                        const uint8* shuffler,
+                        int width);
 void ARGBShuffleRow_Any_SSE2(const uint8* src_argb,
                              uint8* dst_argb,
                              const uint8* shuffler,
                              int width);
 void ARGBShuffleRow_Any_SSSE3(const uint8* src_argb,
                               uint8* dst_argb,
                               const uint8* shuffler,
                               int width);
 void ARGBShuffleRow_Any_AVX2(const uint8* src_argb,
                              uint8* dst_argb,
                              const uint8* shuffler,
                              int width);
 void ARGBShuffleRow_Any_NEON(const uint8* src_argb,
                              uint8* dst_argb,
                              const uint8* shuffler,
                              int width);
+void ARGBShuffleRow_Any_MSA(const uint8* src_argb,
+                            uint8* dst_argb,
+                            const uint8* shuffler,
+                            int width);
 
 void RGB24ToARGBRow_SSSE3(const uint8* src_rgb24, uint8* dst_argb, int width);
 void RAWToARGBRow_SSSE3(const uint8* src_raw, uint8* dst_argb, int width);
 void RAWToRGB24Row_SSSE3(const uint8* src_raw, uint8* dst_rgb24, int width);
 void RGB565ToARGBRow_SSE2(const uint8* src_rgb565, uint8* dst_argb, int width);
 void ARGB1555ToARGBRow_SSE2(const uint8* src_argb1555,
                             uint8* dst_argb,
                             int width);
 void ARGB4444ToARGBRow_SSE2(const uint8* src_argb4444,
                             uint8* dst_argb,
@@ skipping 98 matching lines @@
 void ARGBToARGB4444Row_NEON(const uint8* src_argb, uint8* dst_rgb, int width);
 void ARGBToRGB565DitherRow_NEON(const uint8* src_argb,
                                 uint8* dst_rgb,
                                 const uint32 dither4,
                                 int width);
 void ARGBToRGB24Row_MSA(const uint8* src_argb, uint8* dst_rgb, int width);
 void ARGBToRAWRow_MSA(const uint8* src_argb, uint8* dst_rgb, int width);
 void ARGBToRGB565Row_MSA(const uint8* src_argb, uint8* dst_rgb, int width);
 void ARGBToARGB1555Row_MSA(const uint8* src_argb, uint8* dst_rgb, int width);
 void ARGBToARGB4444Row_MSA(const uint8* src_argb, uint8* dst_rgb, int width);
+void ARGBToRGB565DitherRow_MSA(const uint8* src_argb,
+                               uint8* dst_rgb,
+                               const uint32 dither4,
+                               int width);
 
 void ARGBToRGBARow_C(const uint8* src_argb, uint8* dst_rgb, int width);
 void ARGBToRGB24Row_C(const uint8* src_argb, uint8* dst_rgb, int width);
 void ARGBToRAWRow_C(const uint8* src_argb, uint8* dst_rgb, int width);
 void ARGBToRGB565Row_C(const uint8* src_argb, uint8* dst_rgb, int width);
 void ARGBToARGB1555Row_C(const uint8* src_argb, uint8* dst_rgb, int width);
 void ARGBToARGB4444Row_C(const uint8* src_argb, uint8* dst_rgb, int width);
 
 void J400ToARGBRow_SSE2(const uint8* src_y, uint8* dst_argb, int width);
 void J400ToARGBRow_AVX2(const uint8* src_y, uint8* dst_argb, int width);
@@ skipping 601 matching lines @@
                                     int width);
 void ARGBToRGB24Row_Any_MSA(const uint8* src_argb, uint8* dst_rgb, int width);
 void ARGBToRAWRow_Any_MSA(const uint8* src_argb, uint8* dst_rgb, int width);
 void ARGBToRGB565Row_Any_MSA(const uint8* src_argb, uint8* dst_rgb, int width);
 void ARGBToARGB1555Row_Any_MSA(const uint8* src_argb,
                                uint8* dst_rgb,
                                int width);
 void ARGBToARGB4444Row_Any_MSA(const uint8* src_argb,
                                uint8* dst_rgb,
                                int width);
+void ARGBToRGB565DitherRow_Any_MSA(const uint8* src_argb,
+                                   uint8* dst_rgb,
+                                   const uint32 dither4,
+                                   int width);
 
 void I444ToARGBRow_Any_NEON(const uint8* src_y,
                             const uint8* src_u,
                             const uint8* src_v,
                             uint8* dst_argb,
                             const struct YuvConstants* yuvconstants,
                             int width);
 void I422ToARGBRow_Any_NEON(const uint8* src_y,
                             const uint8* src_u,
                             const uint8* src_v,
@@ skipping 377 matching lines @@
                            const uint8* src_u,
                            const uint8* src_v,
                            uint8* dst_uyvy,
                            int width);
 
 // Effects related row functions.
 void ARGBAttenuateRow_C(const uint8* src_argb, uint8* dst_argb, int width);
 void ARGBAttenuateRow_SSSE3(const uint8* src_argb, uint8* dst_argb, int width);
 void ARGBAttenuateRow_AVX2(const uint8* src_argb, uint8* dst_argb, int width);
 void ARGBAttenuateRow_NEON(const uint8* src_argb, uint8* dst_argb, int width);
+void ARGBAttenuateRow_MSA(const uint8* src_argb, uint8* dst_argb, int width);
 void ARGBAttenuateRow_Any_SSE2(const uint8* src_argb,
                                uint8* dst_argb,
                                int width);
 void ARGBAttenuateRow_Any_SSSE3(const uint8* src_argb,
                                 uint8* dst_argb,
                                 int width);
 void ARGBAttenuateRow_Any_AVX2(const uint8* src_argb,
                                uint8* dst_argb,
                                int width);
 void ARGBAttenuateRow_Any_NEON(const uint8* src_argb,
                                uint8* dst_argb,
                                int width);
+void ARGBAttenuateRow_Any_MSA(const uint8* src_argb,
+                              uint8* dst_argb,
+                              int width);
 
 // Inverse table for unattenuate, shared by C and SSE2.
 extern const uint32 fixed_invtbl8[256];
 void ARGBUnattenuateRow_C(const uint8* src_argb, uint8* dst_argb, int width);
 void ARGBUnattenuateRow_SSE2(const uint8* src_argb, uint8* dst_argb, int width);
 void ARGBUnattenuateRow_AVX2(const uint8* src_argb, uint8* dst_argb, int width);
 void ARGBUnattenuateRow_Any_SSE2(const uint8* src_argb,
                                  uint8* dst_argb,
                                  int width);
 void ARGBUnattenuateRow_Any_AVX2(const uint8* src_argb,
                                  uint8* dst_argb,
                                  int width);
 
 void ARGBGrayRow_C(const uint8* src_argb, uint8* dst_argb, int width);
 void ARGBGrayRow_SSSE3(const uint8* src_argb, uint8* dst_argb, int width);
 void ARGBGrayRow_NEON(const uint8* src_argb, uint8* dst_argb, int width);
+void ARGBGrayRow_MSA(const uint8* src_argb, uint8* dst_argb, int width);
 
 void ARGBSepiaRow_C(uint8* dst_argb, int width);
 void ARGBSepiaRow_SSSE3(uint8* dst_argb, int width);
 void ARGBSepiaRow_NEON(uint8* dst_argb, int width);
+void ARGBSepiaRow_MSA(uint8* dst_argb, int width);
 
 void ARGBColorMatrixRow_C(const uint8* src_argb,
                           uint8* dst_argb,
                           const int8* matrix_argb,
                           int width);
 void ARGBColorMatrixRow_SSSE3(const uint8* src_argb,
                               uint8* dst_argb,
                               const int8* matrix_argb,
                               int width);
 void ARGBColorMatrixRow_NEON(const uint8* src_argb,
@@ skipping 28 matching lines @@
                     int width,
                     uint32 value);
 void ARGBShadeRow_SSE2(const uint8* src_argb,
                        uint8* dst_argb,
                        int width,
                        uint32 value);
 void ARGBShadeRow_NEON(const uint8* src_argb,
                        uint8* dst_argb,
                        int width,
                        uint32 value);
+void ARGBShadeRow_MSA(const uint8* src_argb,
+                      uint8* dst_argb,
+                      int width,
+                      uint32 value);
 
 // Used for blur.
 void CumulativeSumToAverageRow_SSE2(const int32* topleft,
                                     const int32* botleft,
                                     int width,
                                     int area,
                                     uint8* dst,
                                     int count);
 void ComputeCumulativeSumRow_SSE2(const uint8* row,
                                   int32* cumsum,
@@ skipping 222 matching lines @@
                                  int width,
                                  const uint8* luma,
                                  uint32 lumacoeff);
 
 #ifdef __cplusplus
 }  // extern "C"
 }  // namespace libyuv
 #endif
 
 #endif  // INCLUDE_LIBYUV_ROW_H_