libvpx: Pull from upstream

source/libvpx/third_party/libyuv/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 19 matching lines @@
 #else
 #define align_buffer_64(var, size)                                             \
   uint8* var##_mem = (uint8*)(malloc((size) + 63));               /* NOLINT */ \
   uint8* var = (uint8*)(((intptr_t)(var##_mem) + 63) & ~63)       /* NOLINT */
 #endif
 
 #define free_aligned_buffer_64(var) \
   free(var##_mem);  \
   var = 0
 
-#if defined(__pnacl__) || defined(__CLR_VER) || defined(COVERAGE_ENABLED) || \
-    defined(TARGET_IPHONE_SIMULATOR) || \
-    (defined(__i386__) && !defined(__SSE2__)) || \
-    (defined(_MSC_VER) && defined(__clang__))
+#if defined(__pnacl__) || defined(__CLR_VER) || \
+    (defined(__i386__) && !defined(__SSE2__))
 #define LIBYUV_DISABLE_X86
 #endif
 // True if compiling for SSSE3 as a requirement.
 #if defined(__SSSE3__) || (defined(_M_IX86_FP) && (_M_IX86_FP >= 3))
 #define LIBYUV_SSSE3_ONLY
 #endif
 
+#if defined(__native_client__)
+#define LIBYUV_DISABLE_NEON
+#endif
 // clang >= 3.5.0 required for Arm64.
 #if defined(__clang__) && defined(__aarch64__) && !defined(LIBYUV_DISABLE_NEON)
 #if (__clang_major__ < 3) || (__clang_major__ == 3 && (__clang_minor__ < 5))
 #define LIBYUV_DISABLE_NEON
 #endif  // clang >= 3.5
 #endif  // __clang__
 
 // The following are available on all x86 platforms:
 #if !defined(LIBYUV_DISABLE_X86) && \
     (defined(_M_IX86) || defined(__x86_64__) || defined(__i386__))
 // Conversions:
 #define HAS_ABGRTOUVROW_SSSE3
 #define HAS_ABGRTOYROW_SSSE3
 #define HAS_ARGB1555TOARGBROW_SSE2
 #define HAS_ARGB4444TOARGBROW_SSE2
+#define HAS_ARGBSETROW_X86
 #define HAS_ARGBSHUFFLEROW_SSE2
 #define HAS_ARGBSHUFFLEROW_SSSE3
 #define HAS_ARGBTOARGB1555ROW_SSE2
 #define HAS_ARGBTOARGB4444ROW_SSE2
-#define HAS_ARGBTOBAYERGGROW_SSE2
 #define HAS_ARGBTORAWROW_SSSE3
 #define HAS_ARGBTORGB24ROW_SSSE3
 #define HAS_ARGBTORGB565ROW_SSE2
 #define HAS_ARGBTOUV422ROW_SSSE3
 #define HAS_ARGBTOUV444ROW_SSSE3
 #define HAS_ARGBTOUVJROW_SSSE3
 #define HAS_ARGBTOUVROW_SSSE3
 #define HAS_ARGBTOYJROW_SSSE3
 #define HAS_ARGBTOYROW_SSSE3
 #define HAS_BGRATOUVROW_SSSE3
 #define HAS_BGRATOYROW_SSSE3
 #define HAS_COPYROW_ERMS
 #define HAS_COPYROW_SSE2
 #define HAS_I400TOARGBROW_SSE2
 #define HAS_I411TOARGBROW_SSSE3
 #define HAS_I422TOABGRROW_SSSE3
 #define HAS_I422TOARGB1555ROW_SSSE3
 #define HAS_I422TOARGB4444ROW_SSSE3
 #define HAS_I422TOARGBROW_SSSE3
 #define HAS_I422TOBGRAROW_SSSE3
 #define HAS_I422TORAWROW_SSSE3
 #define HAS_I422TORGB24ROW_SSSE3
 #define HAS_I422TORGB565ROW_SSSE3
 #define HAS_I422TORGBAROW_SSSE3
 #define HAS_I422TOUYVYROW_SSE2
 #define HAS_I422TOYUY2ROW_SSE2
 #define HAS_I444TOARGBROW_SSSE3
-// #define HAS_J422TOARGBROW_SSSE3
+#define HAS_J400TOARGBROW_SSE2
+#define HAS_J422TOARGBROW_SSSE3
 #define HAS_MERGEUVROW_SSE2
 #define HAS_MIRRORROW_SSE2
 #define HAS_MIRRORROW_SSSE3
 #define HAS_MIRRORROW_UV_SSSE3
 #define HAS_MIRRORUVROW_SSSE3
 #define HAS_NV12TOARGBROW_SSSE3
 #define HAS_NV12TORGB565ROW_SSSE3
 #define HAS_NV21TOARGBROW_SSSE3
 #define HAS_NV21TORGB565ROW_SSSE3
 #define HAS_RAWTOARGBROW_SSSE3
 #define HAS_RAWTOYROW_SSSE3
 #define HAS_RGB24TOARGBROW_SSSE3
 #define HAS_RGB24TOYROW_SSSE3
 #define HAS_RGB565TOARGBROW_SSE2
 #define HAS_RGBATOUVROW_SSSE3
 #define HAS_RGBATOYROW_SSSE3
+#define HAS_SETROW_ERMS
 #define HAS_SETROW_X86
-#define HAS_SETROW_ERMS
-#define HAS_ARGBSETROW_X86
 #define HAS_SPLITUVROW_SSE2
 #define HAS_UYVYTOARGBROW_SSSE3
 #define HAS_UYVYTOUV422ROW_SSE2
 #define HAS_UYVYTOUVROW_SSE2
 #define HAS_UYVYTOYROW_SSE2
-#define HAS_YTOARGBROW_SSE2
 #define HAS_YUY2TOARGBROW_SSSE3
 #define HAS_YUY2TOUV422ROW_SSE2
 #define HAS_YUY2TOUVROW_SSE2
 #define HAS_YUY2TOYROW_SSE2
 
 // Effects:
 #define HAS_ARGBADDROW_SSE2
 #define HAS_ARGBAFFINEROW_SSE2
 #define HAS_ARGBATTENUATEROW_SSSE3
 #define HAS_ARGBBLENDROW_SSSE3
@@ skipping 16 matching lines @@
 #define HAS_INTERPOLATEROW_SSE2
 #define HAS_INTERPOLATEROW_SSSE3
 #define HAS_RGBCOLORTABLEROW_X86
 #define HAS_SOBELROW_SSE2
 #define HAS_SOBELTOPLANEROW_SSE2
 #define HAS_SOBELXROW_SSE2
 #define HAS_SOBELXYROW_SSE2
 #define HAS_SOBELYROW_SSE2
 #endif
 
-// The following are available on x64 Visual C:
-#if !defined(LIBYUV_DISABLE_X86) && defined (_M_X64)
+// The following are available on x64 Visual C and clangcl.
+#if !defined(LIBYUV_DISABLE_X86) && defined (_M_X64) && \
+    (!defined(__clang__) || defined(__SSSE3__))
 #define HAS_I422TOARGBROW_SSSE3
 #endif
 
 // GCC >= 4.7.0 required for AVX2.
 #if defined(__GNUC__) && (defined(__x86_64__) || defined(__i386__))
 #if (__GNUC__ > 4) || (__GNUC__ == 4 && (__GNUC_MINOR__ >= 7))
 #define GCC_HAS_AVX2 1
 #endif  // GNUC >= 4.7
 #endif  // __GNUC__
 
 // clang >= 3.4.0 required for AVX2.
 #if defined(__clang__) && (defined(__x86_64__) || defined(__i386__))
 #if (__clang_major__ > 3) || (__clang_major__ == 3 && (__clang_minor__ >= 4))
 #define CLANG_HAS_AVX2 1
 #endif  // clang >= 3.4
 #endif  // __clang__
 
 // Visual C 2012 required for AVX2.
-#if defined(_M_IX86) && defined(_MSC_VER) && _MSC_VER >= 1700
+#if defined(_M_IX86) && !defined(__clang__) && \
+    defined(_MSC_VER) && _MSC_VER >= 1700
 #define VISUALC_HAS_AVX2 1
 #endif  // VisualStudio >= 2012
 
 // The following are available require VS2012.  Port to GCC.
 #if !defined(LIBYUV_DISABLE_X86) && defined(VISUALC_HAS_AVX2)
-// TODO(fbarchard): fix AVX2 versions of YUV conversion.  bug=393
-#define HAS_I422TOABGRROW_AVX2
-#define HAS_I422TOARGBROW_AVX2
-#define HAS_I422TOBGRAROW_AVX2
-#define HAS_I422TORGBAROW_AVX2
-#define HAS_NV12TOARGBROW_AVX2
-#define HAS_NV21TOARGBROW_AVX2
-#define HAS_ARGBTORGB565ROW_AVX2
+#define HAS_ARGB1555TOARGBROW_AVX2
+#define HAS_ARGB4444TOARGBROW_AVX2
 #define HAS_ARGBTOARGB1555ROW_AVX2
 #define HAS_ARGBTOARGB4444ROW_AVX2
-#define HAS_NV12TORGB565ROW_AVX2
-#define HAS_NV21TORGB565ROW_AVX2
-#define HAS_I422TORGB565ROW_AVX2
+#define HAS_ARGBTORGB565DITHERROW_AVX2
+#define HAS_ARGBTORGB565DITHERROW_SSE2
+#define HAS_ARGBTORGB565ROW_AVX2
+#define HAS_I411TOARGBROW_AVX2
 #define HAS_I422TOARGB1555ROW_AVX2
 #define HAS_I422TOARGB4444ROW_AVX2
+#define HAS_I422TORGB565ROW_AVX2
+#define HAS_I444TOARGBROW_AVX2
+#define HAS_J400TOARGBROW_AVX2
+#define HAS_NV12TOARGBROW_AVX2
+#define HAS_NV12TORGB565ROW_AVX2
+#define HAS_NV21TOARGBROW_AVX2
+#define HAS_NV21TORGB565ROW_AVX2
+#define HAS_RGB565TOARGBROW_AVX2
 #endif
 
 // The following are available on all x86 platforms, but
 // require VS2012, clang 3.4 or gcc 4.7.
 // The code supports NaCL but requires a new compiler and validator.
 #if !defined(LIBYUV_DISABLE_X86) && (defined(VISUALC_HAS_AVX2) || \
     defined(CLANG_HAS_AVX2) || defined(GCC_HAS_AVX2))
 #define HAS_ARGBCOPYALPHAROW_AVX2
 #define HAS_ARGBCOPYYTOALPHAROW_AVX2
 #define HAS_ARGBMIRRORROW_AVX2
 #define HAS_ARGBPOLYNOMIALROW_AVX2
 #define HAS_ARGBSHUFFLEROW_AVX2
 #define HAS_ARGBTOUVROW_AVX2
 #define HAS_ARGBTOYJROW_AVX2
 #define HAS_ARGBTOYROW_AVX2
 #define HAS_COPYROW_AVX
+#define HAS_I400TOARGBROW_AVX2
+#define HAS_I422TOABGRROW_AVX2
+#define HAS_I422TOARGBROW_AVX2
+#define HAS_I422TOBGRAROW_AVX2
+#define HAS_I422TORAWROW_AVX2
+#define HAS_I422TORGB24ROW_AVX2
+#define HAS_I422TORGBAROW_AVX2
 #define HAS_INTERPOLATEROW_AVX2
+#define HAS_J422TOARGBROW_AVX2
 #define HAS_MERGEUVROW_AVX2
 #define HAS_MIRRORROW_AVX2
 #define HAS_SPLITUVROW_AVX2
+#define HAS_UYVYTOARGBROW_AVX2
 #define HAS_UYVYTOUV422ROW_AVX2
 #define HAS_UYVYTOUVROW_AVX2
 #define HAS_UYVYTOYROW_AVX2
-#define HAS_YTOARGBROW_AVX2
+#define HAS_YUY2TOARGBROW_AVX2
 #define HAS_YUY2TOUV422ROW_AVX2
 #define HAS_YUY2TOUVROW_AVX2
 #define HAS_YUY2TOYROW_AVX2
 
-// The following require HAS_I422TOARGBROW_AVX2
-#if defined(HAS_I422TOARGBROW_AVX2)
-#define HAS_YUY2TOARGBROW_AVX2
-#define HAS_UYVYTOARGBROW_AVX2
-#endif
-
 // Effects:
 #define HAS_ARGBADDROW_AVX2
 #define HAS_ARGBATTENUATEROW_AVX2
 #define HAS_ARGBMULTIPLYROW_AVX2
 #define HAS_ARGBSUBTRACTROW_AVX2
 #define HAS_ARGBUNATTENUATEROW_AVX2
 #endif
 
-
-// The following are Yasm x86 only:
-// TODO(fbarchard): Port AVX2 to inline.
-#if !defined(LIBYUV_DISABLE_X86) && defined(HAVE_YASM)
-    (defined(_M_IX86) || defined(_M_X64) || \
-    defined(__x86_64__) || defined(__i386__))
-#define HAS_MERGEUVROW_AVX2
-#define HAS_MERGEUVROW_MMX
-#define HAS_SPLITUVROW_AVX2
-#define HAS_SPLITUVROW_MMX
-#define HAS_UYVYTOYROW_AVX2
-#define HAS_UYVYTOYROW_MMX
-#define HAS_YUY2TOYROW_AVX2
-#define HAS_YUY2TOYROW_MMX
-#endif
-
 // The following are disabled when SSSE3 is available:
 #if !defined(LIBYUV_DISABLE_X86) && \
     (defined(_M_IX86) || defined(__x86_64__) || defined(__i386__)) && \
     !defined(LIBYUV_SSSE3_ONLY)
 #define HAS_ARGBATTENUATEROW_SSE2
 #define HAS_ARGBBLENDROW_SSE2
 #define HAS_MIRRORROW_SSE2
 #endif
 
 // The following are available on Neon platforms:
 #if !defined(LIBYUV_DISABLE_NEON) && \
     (defined(__aarch64__) || defined(__ARM_NEON__) || defined(LIBYUV_NEON))
 #define HAS_ABGRTOUVROW_NEON
 #define HAS_ABGRTOYROW_NEON
 #define HAS_ARGB1555TOARGBROW_NEON
 #define HAS_ARGB1555TOUVROW_NEON
 #define HAS_ARGB1555TOYROW_NEON
 #define HAS_ARGB4444TOARGBROW_NEON
 #define HAS_ARGB4444TOUVROW_NEON
 #define HAS_ARGB4444TOYROW_NEON
 #define HAS_ARGBTOARGB1555ROW_NEON
 #define HAS_ARGBTOARGB4444ROW_NEON
-#define HAS_ARGBTOBAYERGGROW_NEON
 #define HAS_ARGBTORAWROW_NEON
 #define HAS_ARGBTORGB24ROW_NEON
 #define HAS_ARGBTORGB565ROW_NEON
 #define HAS_ARGBTOUV411ROW_NEON
 #define HAS_ARGBTOUV422ROW_NEON
 #define HAS_ARGBTOUV444ROW_NEON
 #define HAS_ARGBTOUVJROW_NEON
 #define HAS_ARGBTOUVROW_NEON
 #define HAS_ARGBTOYJROW_NEON
 #define HAS_ARGBTOYROW_NEON
 #define HAS_BGRATOUVROW_NEON
 #define HAS_BGRATOYROW_NEON
 #define HAS_COPYROW_NEON
-#define HAS_I400TOARGBROW_NEON
+#define HAS_J400TOARGBROW_NEON
 #define HAS_I411TOARGBROW_NEON
 #define HAS_I422TOABGRROW_NEON
 #define HAS_I422TOARGB1555ROW_NEON
 #define HAS_I422TOARGB4444ROW_NEON
 #define HAS_I422TOARGBROW_NEON
 #define HAS_I422TOBGRAROW_NEON
 #define HAS_I422TORAWROW_NEON
 #define HAS_I422TORGB24ROW_NEON
 #define HAS_I422TORGB565ROW_NEON
 #define HAS_I422TORGBAROW_NEON
@@ skipping 18 matching lines @@
 #define HAS_RGB565TOYROW_NEON
 #define HAS_RGBATOUVROW_NEON
 #define HAS_RGBATOYROW_NEON
 #define HAS_SETROW_NEON
 #define HAS_ARGBSETROW_NEON
 #define HAS_SPLITUVROW_NEON
 #define HAS_UYVYTOARGBROW_NEON
 #define HAS_UYVYTOUV422ROW_NEON
 #define HAS_UYVYTOUVROW_NEON
 #define HAS_UYVYTOYROW_NEON
-#define HAS_YTOARGBROW_NEON
+#define HAS_I400TOARGBROW_NEON
 #define HAS_YUY2TOARGBROW_NEON
 #define HAS_YUY2TOUV422ROW_NEON
 #define HAS_YUY2TOUVROW_NEON
 #define HAS_YUY2TOYROW_NEON
+#define HAS_ARGBTORGB565DITHERROW_NEON
 
 // Effects:
 #define HAS_ARGBADDROW_NEON
 #define HAS_ARGBATTENUATEROW_NEON
 #define HAS_ARGBBLENDROW_NEON
 #define HAS_ARGBGRAYROW_NEON
 #define HAS_ARGBMIRRORROW_NEON
 #define HAS_ARGBMULTIPLYROW_NEON
 #define HAS_ARGBQUANTIZEROW_NEON
 #define HAS_ARGBSEPIAROW_NEON
@@ skipping 32 matching lines @@
 typedef __declspec(align(16)) int8 vec8[16];
 typedef __declspec(align(16)) uint16 uvec16[8];
 typedef __declspec(align(16)) uint32 uvec32[4];
 typedef __declspec(align(16)) uint8 uvec8[16];
 typedef __declspec(align(32)) int16 lvec16[16];
 typedef __declspec(align(32)) int32 lvec32[8];
 typedef __declspec(align(32)) int8 lvec8[32];
 typedef __declspec(align(32)) uint16 ulvec16[16];
 typedef __declspec(align(32)) uint32 ulvec32[8];
 typedef __declspec(align(32)) uint8 ulvec8[32];
-
 #elif defined(__GNUC__)
 // Caveat GCC 4.2 to 4.7 have a known issue using vectors with const.
 #define SIMD_ALIGNED(var) var __attribute__((aligned(16)))
 #define SIMD_ALIGNED32(var) var __attribute__((aligned(64)))
 typedef int16 __attribute__((vector_size(16))) vec16;
 typedef int32 __attribute__((vector_size(16))) vec32;
 typedef int8 __attribute__((vector_size(16))) vec8;
 typedef uint16 __attribute__((vector_size(16))) uvec16;
 typedef uint32 __attribute__((vector_size(16))) uvec32;
 typedef uint8 __attribute__((vector_size(16))) uvec8;
@@ skipping 460 matching lines @@
 void ARGBShuffleRow_Any_NEON(const uint8* src_argb, uint8* dst_argb,
                              const uint8* shuffler, int pix);
 
 void RGB24ToARGBRow_SSSE3(const uint8* src_rgb24, uint8* dst_argb, int pix);
 void RAWToARGBRow_SSSE3(const uint8* src_raw, uint8* dst_argb, int pix);
 void RGB565ToARGBRow_SSE2(const uint8* src_rgb565, uint8* dst_argb, int pix);
 void ARGB1555ToARGBRow_SSE2(const uint8* src_argb1555, uint8* dst_argb,
                             int pix);
 void ARGB4444ToARGBRow_SSE2(const uint8* src_argb4444, uint8* dst_argb,
                             int pix);
+void RGB565ToARGBRow_AVX2(const uint8* src_rgb565, uint8* dst_argb, int pix);
+void ARGB1555ToARGBRow_AVX2(const uint8* src_argb1555, uint8* dst_argb,
+                            int pix);
+void ARGB4444ToARGBRow_AVX2(const uint8* src_argb4444, uint8* dst_argb,
+                            int pix);
 
 void RGB24ToARGBRow_NEON(const uint8* src_rgb24, uint8* dst_argb, int pix);
 void RAWToARGBRow_NEON(const uint8* src_raw, uint8* dst_argb, int pix);
 void RGB565ToARGBRow_NEON(const uint8* src_rgb565, uint8* dst_argb, int pix);
 void ARGB1555ToARGBRow_NEON(const uint8* src_argb1555, uint8* dst_argb,
                             int pix);
 void ARGB4444ToARGBRow_NEON(const uint8* src_argb4444, uint8* dst_argb,
                             int pix);
 void RGB24ToARGBRow_C(const uint8* src_rgb24, uint8* dst_argb, int pix);
 void RAWToARGBRow_C(const uint8* src_raw, uint8* dst_argb, int pix);
 void RGB565ToARGBRow_C(const uint8* src_rgb, uint8* dst_argb, int pix);
 void ARGB1555ToARGBRow_C(const uint8* src_argb, uint8* dst_argb, int pix);
 void ARGB4444ToARGBRow_C(const uint8* src_argb, uint8* dst_argb, int pix);
 void RGB24ToARGBRow_Any_SSSE3(const uint8* src_rgb24, uint8* dst_argb, int pix);
 void RAWToARGBRow_Any_SSSE3(const uint8* src_raw, uint8* dst_argb, int pix);
+
 void RGB565ToARGBRow_Any_SSE2(const uint8* src_rgb565, uint8* dst_argb,
                               int pix);
 void ARGB1555ToARGBRow_Any_SSE2(const uint8* src_argb1555, uint8* dst_argb,
                                 int pix);
 void ARGB4444ToARGBRow_Any_SSE2(const uint8* src_argb4444, uint8* dst_argb,
                                 int pix);
+void RGB565ToARGBRow_Any_AVX2(const uint8* src_rgb565, uint8* dst_argb,
+                              int pix);
+void ARGB1555ToARGBRow_Any_AVX2(const uint8* src_argb1555, uint8* dst_argb,
+                                int pix);
+void ARGB4444ToARGBRow_Any_AVX2(const uint8* src_argb4444, uint8* dst_argb,
+                                int pix);
+
 void RGB24ToARGBRow_Any_NEON(const uint8* src_rgb24, uint8* dst_argb, int pix);
 void RAWToARGBRow_Any_NEON(const uint8* src_raw, uint8* dst_argb, int pix);
 void RGB565ToARGBRow_Any_NEON(const uint8* src_rgb565, uint8* dst_argb,
                               int pix);
 void ARGB1555ToARGBRow_Any_NEON(const uint8* src_argb1555, uint8* dst_argb,
                                 int pix);
 void ARGB4444ToARGBRow_Any_NEON(const uint8* src_argb4444, uint8* dst_argb,
                                 int pix);
 
 void ARGBToRGB24Row_SSSE3(const uint8* src_argb, uint8* dst_rgb, int pix);
 void ARGBToRAWRow_SSSE3(const uint8* src_argb, uint8* dst_rgb, int pix);
 void ARGBToRGB565Row_SSE2(const uint8* src_argb, uint8* dst_rgb, int pix);
 void ARGBToARGB1555Row_SSE2(const uint8* src_argb, uint8* dst_rgb, int pix);
 void ARGBToARGB4444Row_SSE2(const uint8* src_argb, uint8* dst_rgb, int pix);
 
+void ARGBToRGB565DitherRow_C(const uint8* src_argb, uint8* dst_rgb,
+                             const uint32 dither4, int pix);
+void ARGBToRGB565DitherRow_SSE2(const uint8* src_argb, uint8* dst_rgb,
+                                const uint32 dither4, int pix);
+void ARGBToRGB565DitherRow_AVX2(const uint8* src_argb, uint8* dst_rgb,
+                                const uint32 dither4, int pix);
+
 void ARGBToRGB565Row_AVX2(const uint8* src_argb, uint8* dst_rgb, int pix);
 void ARGBToARGB1555Row_AVX2(const uint8* src_argb, uint8* dst_rgb, int pix);
 void ARGBToARGB4444Row_AVX2(const uint8* src_argb, uint8* dst_rgb, int pix);
 
 void ARGBToRGB24Row_NEON(const uint8* src_argb, uint8* dst_rgb, int pix);
 void ARGBToRAWRow_NEON(const uint8* src_argb, uint8* dst_rgb, int pix);
 void ARGBToRGB565Row_NEON(const uint8* src_argb, uint8* dst_rgb, int pix);
 void ARGBToARGB1555Row_NEON(const uint8* src_argb, uint8* dst_rgb, int pix);
 void ARGBToARGB4444Row_NEON(const uint8* src_argb, uint8* dst_rgb, int pix);
+void ARGBToRGB565DitherRow_NEON(const uint8* src_argb, uint8* dst_rgb,
+                                const uint32 dither4, int width);
 
 void ARGBToRGBARow_C(const uint8* src_argb, uint8* dst_rgb, int pix);
 void ARGBToRGB24Row_C(const uint8* src_argb, uint8* dst_rgb, int pix);
 void ARGBToRAWRow_C(const uint8* src_argb, uint8* dst_rgb, int pix);
 void ARGBToRGB565Row_C(const uint8* src_argb, uint8* dst_rgb, int pix);
 void ARGBToARGB1555Row_C(const uint8* src_argb, uint8* dst_rgb, int pix);
 void ARGBToARGB4444Row_C(const uint8* src_argb, uint8* dst_rgb, int pix);
 
-void ARGBToRGB565DitherRow_C(const uint8* src_argb, uint8* dst_rgb,
-                             const uint8* dither8x8, int pix);
-
-void I400ToARGBRow_SSE2(const uint8* src_y, uint8* dst_argb, int pix);
-void I400ToARGBRow_NEON(const uint8* src_y, uint8* dst_argb, int pix);
-void I400ToARGBRow_C(const uint8* src_y, uint8* dst_argb, int pix);
-void I400ToARGBRow_Any_SSE2(const uint8* src_y, uint8* dst_argb, int pix);
-void I400ToARGBRow_Any_NEON(const uint8* src_y, uint8* dst_argb, int pix);
+void J400ToARGBRow_SSE2(const uint8* src_y, uint8* dst_argb, int pix);
+void J400ToARGBRow_AVX2(const uint8* src_y, uint8* dst_argb, int pix);
+void J400ToARGBRow_NEON(const uint8* src_y, uint8* dst_argb, int pix);
+void J400ToARGBRow_C(const uint8* src_y, uint8* dst_argb, int pix);
+void J400ToARGBRow_Any_SSE2(const uint8* src_y, uint8* dst_argb, int pix);
+void J400ToARGBRow_Any_AVX2(const uint8* src_y, uint8* dst_argb, int pix);
+void J400ToARGBRow_Any_NEON(const uint8* src_y, uint8* dst_argb, int pix);
 
 void I444ToARGBRow_C(const uint8* src_y,
                      const uint8* src_u,
                      const uint8* src_v,
                      uint8* dst_argb,
                      int width);
 void I422ToARGBRow_C(const uint8* src_y,
                      const uint8* src_u,
                      const uint8* src_v,
                      uint8* dst_argb,
@@ skipping 88 matching lines @@
 void I422ToABGRRow_AVX2(const uint8* src_y,
                         const uint8* src_u,
                         const uint8* src_v,
                         uint8* dst_argb,
                         int width);
 void I444ToARGBRow_SSSE3(const uint8* src_y,
                          const uint8* src_u,
                          const uint8* src_v,
                          uint8* dst_argb,
                          int width);
+void I444ToARGBRow_AVX2(const uint8* src_y,
+                        const uint8* src_u,
+                        const uint8* src_v,
+                        uint8* dst_argb,
+                        int width);
 void I422ToARGBRow_SSSE3(const uint8* src_y,
                          const uint8* src_u,
                          const uint8* src_v,
                          uint8* dst_argb,
                          int width);
 void I411ToARGBRow_SSSE3(const uint8* src_y,
                          const uint8* src_u,
                          const uint8* src_v,
                          uint8* dst_argb,
                          int width);
+void I411ToARGBRow_AVX2(const uint8* src_y,
+                        const uint8* src_u,
+                        const uint8* src_v,
+                        uint8* dst_argb,
+                        int width);
 void NV12ToARGBRow_SSSE3(const uint8* src_y,
                          const uint8* src_uv,
                          uint8* dst_argb,
                          int width);
 void NV21ToARGBRow_SSSE3(const uint8* src_y,
                          const uint8* src_vu,
                          uint8* dst_argb,
                          int width);
 void NV12ToARGBRow_AVX2(const uint8* src_y,
                         const uint8* src_uv,
@@ skipping 29 matching lines @@
                         uint8* dst_argb,
                         int width);
 void UYVYToARGBRow_AVX2(const uint8* src_uyvy,
                         uint8* dst_argb,
                         int width);
 void J422ToARGBRow_SSSE3(const uint8* src_y,
                          const uint8* src_u,
                          const uint8* src_v,
                          uint8* dst_argb,
                          int width);
+void J422ToARGBRow_AVX2(const uint8* src_y,
+                        const uint8* src_u,
+                        const uint8* src_v,
+                        uint8* dst_argb,
+                        int width);
 void I422ToBGRARow_SSSE3(const uint8* src_y,
                          const uint8* src_u,
                          const uint8* src_v,
                          uint8* dst_bgra,
                          int width);
 void I422ToABGRRow_SSSE3(const uint8* src_y,
                          const uint8* src_u,
                          const uint8* src_v,
                          uint8* dst_abgr,
                          int width);
@@ skipping 30 matching lines @@
 void I422ToRGB565Row_AVX2(const uint8* src_y,
                           const uint8* src_u,
                           const uint8* src_v,
                           uint8* dst_argb,
                           int width);
 void I422ToRGB24Row_SSSE3(const uint8* src_y,
                           const uint8* src_u,
                           const uint8* src_v,
                           uint8* dst_rgb24,
                           int width);
+void I422ToRGB24Row_AVX2(const uint8* src_y,
+                         const uint8* src_u,
+                         const uint8* src_v,
+                         uint8* dst_rgb24,
+                         int width);
 void I422ToRAWRow_SSSE3(const uint8* src_y,
                         const uint8* src_u,
                         const uint8* src_v,
                         uint8* dst_raw,
                         int width);
+void I422ToRAWRow_AVX2(const uint8* src_y,
+                       const uint8* src_u,
+                       const uint8* src_v,
+                       uint8* dst_raw,
+                       int width);
 void I422ToARGBRow_Any_AVX2(const uint8* src_y,
                             const uint8* src_u,
                             const uint8* src_v,
                             uint8* dst_argb,
                             int width);
 void I422ToBGRARow_Any_AVX2(const uint8* src_y,
                             const uint8* src_u,
                             const uint8* src_v,
                             uint8* dst_argb,
                             int width);
 void I422ToRGBARow_Any_AVX2(const uint8* src_y,
                             const uint8* src_u,
                             const uint8* src_v,
                             uint8* dst_argb,
                             int width);
 void I422ToABGRRow_Any_AVX2(const uint8* src_y,
                             const uint8* src_u,
                             const uint8* src_v,
                             uint8* dst_argb,
                             int width);
 void I444ToARGBRow_Any_SSSE3(const uint8* src_y,
                              const uint8* src_u,
                              const uint8* src_v,
                              uint8* dst_argb,
                              int width);
+void I444ToARGBRow_Any_AVX2(const uint8* src_y,
+                            const uint8* src_u,
+                            const uint8* src_v,
+                            uint8* dst_argb,
+                            int width);
 void I422ToARGBRow_Any_SSSE3(const uint8* src_y,
                              const uint8* src_u,
                              const uint8* src_v,
                              uint8* dst_argb,
                              int width);
 void I411ToARGBRow_Any_SSSE3(const uint8* src_y,
                              const uint8* src_u,
                              const uint8* src_v,
                              uint8* dst_argb,
                              int width);
+void I411ToARGBRow_Any_AVX2(const uint8* src_y,
+                            const uint8* src_u,
+                            const uint8* src_v,
+                            uint8* dst_argb,
+                            int width);
 void NV12ToARGBRow_Any_SSSE3(const uint8* src_y,
                              const uint8* src_uv,
                              uint8* dst_argb,
                              int width);
 void NV21ToARGBRow_Any_SSSE3(const uint8* src_y,
                              const uint8* src_vu,
                              uint8* dst_argb,
                              int width);
 void NV12ToARGBRow_Any_AVX2(const uint8* src_y,
                             const uint8* src_uv,
@@ skipping 24 matching lines @@
                              int width);
 void UYVYToARGBRow_Any_SSSE3(const uint8* src_uyvy,
                              uint8* dst_argb,
                              int width);
 void YUY2ToARGBRow_Any_AVX2(const uint8* src_yuy2,
                             uint8* dst_argb,
                             int width);
 void UYVYToARGBRow_Any_AVX2(const uint8* src_uyvy,
                             uint8* dst_argb,
                             int width);
+void J422ToARGBRow_Any_SSSE3(const uint8* src_y,
+                             const uint8* src_u,
+                             const uint8* src_v,
+                             uint8* dst_argb,
+                             int width);
+void J422ToARGBRow_Any_AVX2(const uint8* src_y,
+                            const uint8* src_u,
+                            const uint8* src_v,
+                            uint8* dst_argb,
+                            int width);
 void I422ToBGRARow_Any_SSSE3(const uint8* src_y,
                              const uint8* src_u,
                              const uint8* src_v,
                              uint8* dst_bgra,
                              int width);
 void I422ToABGRRow_Any_SSSE3(const uint8* src_y,
                              const uint8* src_u,
                              const uint8* src_v,
                              uint8* dst_abgr,
                              int width);
@@ skipping 30 matching lines @@
 void I422ToRGB565Row_Any_AVX2(const uint8* src_y,
                               const uint8* src_u,
                               const uint8* src_v,
                               uint8* dst_rgba,
                               int width);
 void I422ToRGB24Row_Any_SSSE3(const uint8* src_y,
                               const uint8* src_u,
                               const uint8* src_v,
                               uint8* dst_argb,
                               int width);
+void I422ToRGB24Row_Any_AVX2(const uint8* src_y,
+                             const uint8* src_u,
+                             const uint8* src_v,
+                             uint8* dst_argb,
+                             int width);
 void I422ToRAWRow_Any_SSSE3(const uint8* src_y,
                             const uint8* src_u,
                             const uint8* src_v,
                             uint8* dst_argb,
                             int width);
+void I422ToRAWRow_Any_AVX2(const uint8* src_y,
+                           const uint8* src_u,
+                           const uint8* src_v,
+                           uint8* dst_argb,
+                           int width);
 
-void YToARGBRow_C(const uint8* src_y,
-                  uint8* dst_argb,
-                  int width);
-void YToARGBRow_SSE2(const uint8* src_y,
-                     uint8* dst_argb,
-                     int width);
-void YToARGBRow_AVX2(const uint8* src_y,
-                     uint8* dst_argb,
-                     int width);
-void YToARGBRow_NEON(const uint8* src_y,
-                     uint8* dst_argb,
-                     int width);
-void YToARGBRow_Any_SSE2(const uint8* src_y,
-                         uint8* dst_argb,
-                         int width);
-void YToARGBRow_Any_AVX2(const uint8* src_y,
-                         uint8* dst_argb,
-                         int width);
-void YToARGBRow_Any_NEON(const uint8* src_y,
-                         uint8* dst_argb,
-                         int width);
+void I400ToARGBRow_C(const uint8* src_y, uint8* dst_argb, int width);
+void I400ToARGBRow_SSE2(const uint8* src_y, uint8* dst_argb, int width);
+void I400ToARGBRow_AVX2(const uint8* src_y, uint8* dst_argb, int width);
+void I400ToARGBRow_NEON(const uint8* src_y, uint8* dst_argb, int width);
+void I400ToARGBRow_Any_SSE2(const uint8* src_y, uint8* dst_argb, int width);
+void I400ToARGBRow_Any_AVX2(const uint8* src_y, uint8* dst_argb, int width);
+void I400ToARGBRow_Any_NEON(const uint8* src_y, uint8* dst_argb, int width);
 
 // ARGB preattenuated alpha blend.
 void ARGBBlendRow_SSSE3(const uint8* src_argb, const uint8* src_argb1,
                         uint8* dst_argb, int width);
 void ARGBBlendRow_SSE2(const uint8* src_argb, const uint8* src_argb1,
                        uint8* dst_argb, int width);
 void ARGBBlendRow_NEON(const uint8* src_argb, const uint8* src_argb1,
                        uint8* dst_argb, int width);
 void ARGBBlendRow_C(const uint8* src_argb, const uint8* src_argb1,
                     uint8* dst_argb, int width);
@@ skipping 47 matching lines @@
                           uint8* dst_argb, int width);
 void ARGBSubtractRow_Any_NEON(const uint8* src_argb, const uint8* src_argb1,
                               uint8* dst_argb, int width);
 
 void ARGBToRGB24Row_Any_SSSE3(const uint8* src_argb, uint8* dst_rgb, int pix);
 void ARGBToRAWRow_Any_SSSE3(const uint8* src_argb, uint8* dst_rgb, int pix);
 void ARGBToRGB565Row_Any_SSE2(const uint8* src_argb, uint8* dst_rgb, int pix);
 void ARGBToARGB1555Row_Any_SSE2(const uint8* src_argb, uint8* dst_rgb, int pix);
 void ARGBToARGB4444Row_Any_SSE2(const uint8* src_argb, uint8* dst_rgb, int pix);
 
+void ARGBToRGB565DitherRow_Any_SSE2(const uint8* src_argb, uint8* dst_rgb,
+                                    const uint32 dither4, int pix);
+void ARGBToRGB565DitherRow_Any_AVX2(const uint8* src_argb, uint8* dst_rgb,
+                                    const uint32 dither4, int pix);
+
 void ARGBToRGB565Row_Any_AVX2(const uint8* src_argb, uint8* dst_rgb, int pix);
 void ARGBToARGB1555Row_Any_AVX2(const uint8* src_argb, uint8* dst_rgb, int pix);
 void ARGBToARGB4444Row_Any_AVX2(const uint8* src_argb, uint8* dst_rgb, int pix);
 
 void ARGBToRGB24Row_Any_NEON(const uint8* src_argb, uint8* dst_rgb, int pix);
 void ARGBToRAWRow_Any_NEON(const uint8* src_argb, uint8* dst_rgb, int pix);
 void ARGBToRGB565Row_Any_NEON(const uint8* src_argb, uint8* dst_rgb, int pix);
 void ARGBToARGB1555Row_Any_NEON(const uint8* src_argb, uint8* dst_rgb, int pix);
 void ARGBToARGB4444Row_Any_NEON(const uint8* src_argb, uint8* dst_rgb, int pix);
+void ARGBToRGB565DitherRow_Any_NEON(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,
                             int width);
 void I422ToARGBRow_Any_NEON(const uint8* src_y,
                             const uint8* src_u,
                             const uint8* src_v,
                             uint8* dst_argb,
@@ skipping 166 matching lines @@
 void UYVYToUVRow_Any_SSE2(const uint8* src_uyvy, int stride_uyvy,
                           uint8* dst_u, uint8* dst_v, int pix);
 void UYVYToUV422Row_Any_SSE2(const uint8* src_uyvy,
                              uint8* dst_u, uint8* dst_v, int pix);
 void UYVYToYRow_Any_NEON(const uint8* src_uyvy, uint8* dst_y, int pix);
 void UYVYToUVRow_Any_NEON(const uint8* src_uyvy, int stride_uyvy,
                           uint8* dst_u, uint8* dst_v, int pix);
 void UYVYToUV422Row_Any_NEON(const uint8* src_uyvy,
                              uint8* dst_u, uint8* dst_v, int pix);
 
-void ARGBToBayerGGRow_C(const uint8* src_argb, uint8* dst_bayer,
-                        uint32 /* selector */, int pix);
-void ARGBToBayerGGRow_SSE2(const uint8* src_argb, uint8* dst_bayer,
-                           uint32 /* selector */, int pix);
-void ARGBToBayerGGRow_NEON(const uint8* src_argb, uint8* dst_bayer,
-                           uint32 /* selector */, int pix);
-void ARGBToBayerGGRow_Any_SSE2(const uint8* src_argb, uint8* dst_bayer,
-                               uint32 /* selector */, int pix);
-void ARGBToBayerGGRow_Any_NEON(const uint8* src_argb, uint8* dst_bayer,
-                               uint32 /* selector */, int pix);
-
 void I422ToYUY2Row_C(const uint8* src_y,
                      const uint8* src_u,
                      const uint8* src_v,
                      uint8* dst_yuy2, int width);
 void I422ToUYVYRow_C(const uint8* src_y,
                      const uint8* src_u,
                      const uint8* src_v,
                      uint8* dst_uyvy, int width);
 void I422ToYUY2Row_SSE2(const uint8* src_y,
                         const uint8* src_u,
@@ skipping 169 matching lines @@
 void SobelToPlaneRow_SSE2(const uint8* src_sobelx, const uint8* src_sobely,
                           uint8* dst_y, int width);
 void SobelToPlaneRow_NEON(const uint8* src_sobelx, const uint8* src_sobely,
                           uint8* dst_y, int width);
 void SobelXYRow_C(const uint8* src_sobelx, const uint8* src_sobely,
                   uint8* dst_argb, int width);
 void SobelXYRow_SSE2(const uint8* src_sobelx, const uint8* src_sobely,
                      uint8* dst_argb, int width);
 void SobelXYRow_NEON(const uint8* src_sobelx, const uint8* src_sobely,
                      uint8* dst_argb, int width);
+void SobelRow_Any_SSE2(const uint8* src_sobelx, const uint8* src_sobely,
+                       uint8* dst_argb, int width);
+void SobelRow_Any_NEON(const uint8* src_sobelx, const uint8* src_sobely,
+                       uint8* dst_argb, int width);
+void SobelToPlaneRow_Any_SSE2(const uint8* src_sobelx, const uint8* src_sobely,
+                              uint8* dst_y, int width);
+void SobelToPlaneRow_Any_NEON(const uint8* src_sobelx, const uint8* src_sobely,
+                              uint8* dst_y, int width);
+void SobelXYRow_Any_SSE2(const uint8* src_sobelx, const uint8* src_sobely,
+                         uint8* dst_argb, int width);
+void SobelXYRow_Any_NEON(const uint8* src_sobelx, const uint8* src_sobely,
+                         uint8* dst_argb, int width);
 
 void ARGBPolynomialRow_C(const uint8* src_argb,
                          uint8* dst_argb, const float* poly,
                          int width);
 void ARGBPolynomialRow_SSE2(const uint8* src_argb,
                             uint8* dst_argb, const float* poly,
                             int width);
 void ARGBPolynomialRow_AVX2(const uint8* src_argb,
                             uint8* dst_argb, const float* poly,
                             int width);
 
 void ARGBLumaColorTableRow_C(const uint8* src_argb, uint8* dst_argb, int width,
                              const uint8* luma, uint32 lumacoeff);
 void ARGBLumaColorTableRow_SSSE3(const uint8* src_argb, uint8* dst_argb,
                                  int width,
                                  const uint8* luma, uint32 lumacoeff);
 
 #ifdef __cplusplus
 }  // extern "C"
 }  // namespace libyuv
 #endif
 
 #endif  // INCLUDE_LIBYUV_ROW_H_  NOLINT