port NV12ToRGB565Row_AVX2 to gcc

NV12ToRGB565Row for Intel is implemented as a 2 step conversion:
NV12ToARGBRow_SSSE3 and ARGBToRGB565Row_SSE2

NV12ToARGBRow has an AVX2 version, so this CL implements
NV12ToRGB565Row_AVX2 with call to NV12ToARGBRow_AVX2 and
ARGBToRGB565Row_SSE2.

R=harryjin@google.com
BUG=libyuv:554

Review URL: https://codereview.chromium.org/1687953002 .

include/libyuv/row.h

@@ -206,15 +206,19 @@ extern "C" {
 // https://code.google.com/p/libyuv/issues/detail?id=517
 #define HAS_I422ALPHATOARGBROW_AVX2
 #endif
-#define HAS_I444TOARGBROW_AVX2
+#define HAS_I422TOARGB1555ROW_AVX2
+#define HAS_I422TOARGB4444ROW_AVX2
 #define HAS_I422TOARGBROW_AVX2
 #define HAS_I422TORGB24ROW_AVX2
+#define HAS_I422TORGB565ROW_AVX2
 #define HAS_I422TORGBAROW_AVX2
+#define HAS_I444TOARGBROW_AVX2
 #define HAS_INTERPOLATEROW_AVX2
 #define HAS_J422TOARGBROW_AVX2
 #define HAS_MERGEUVROW_AVX2
 #define HAS_MIRRORROW_AVX2
 #define HAS_NV12TOARGBROW_AVX2
+#define HAS_NV12TORGB565ROW_AVX2
 #define HAS_NV21TOARGBROW_AVX2
 #define HAS_SPLITUVROW_AVX2
 #define HAS_UYVYTOARGBROW_AVX2
@@ -225,9 +229,6 @@ extern "C" {
 #define HAS_YUY2TOUV422ROW_AVX2
 #define HAS_YUY2TOUVROW_AVX2
 #define HAS_YUY2TOYROW_AVX2
-#define HAS_I422TOARGB1555ROW_AVX2
-#define HAS_I422TOARGB4444ROW_AVX2
-#define HAS_I422TORGB565ROW_AVX2
 
 // Effects:
 #define HAS_ARGBADDROW_AVX2
@@ -249,7 +250,6 @@ extern "C" {
 #define HAS_ARGBTORGB565ROW_AVX2
 #define HAS_I411TOARGBROW_AVX2
 #define HAS_J400TOARGBROW_AVX2
-#define HAS_NV12TORGB565ROW_AVX2
 #define HAS_RGB565TOARGBROW_AVX2
 #endif
 

source/row_common.cc

@@ -2595,7 +2595,11 @@ void NV12ToRGB565Row_AVX2(const uint8* src_y,
   while (width > 0) {
     int twidth = width > MAXTWIDTH ? MAXTWIDTH : width;
     NV12ToARGBRow_AVX2(src_y, src_uv, row, yuvconstants, twidth);
+#if defined(HAS_ARGBTORGB565ROW_AVX2)
     ARGBToRGB565Row_AVX2(row, dst_rgb565, twidth);
+#else
+    ARGBToRGB565Row_SSE2(row, dst_rgb565, twidth);
+#endif
     src_y += twidth;
     src_uv += twidth;
     dst_rgb565 += twidth * 2;