Add MSA optimized YUY2ToI422, YUY2ToI420, UYVYToI422, UYVYToI420 functions

source/planar_functions.cc

Index: source/planar_functions.cc
diff --git a/source/planar_functions.cc b/source/planar_functions.cc
index 53248ee34d0d38f1924e93d8afb318f37625559f..55e598fd3e0405c8b281fd60fb7ebb18a9021f5a 100644
--- a/source/planar_functions.cc
+++ b/source/planar_functions.cc
@@ -480,6 +480,18 @@ int YUY2ToI422(const uint8* src_yuy2, int src_stride_yuy2,
     }
   }
 #endif
+#if defined(HAS_YUY2TOYROW_MSA)
+  if (TestCpuFlag(kCpuHasMSA)) {
+    YUY2ToYRow = YUY2ToYRow_Any_MSA;
+    if (width >= 16) {

[fbarchard1, 2016/10/05 22:03:58]

todo - make Any handle remainder better so this test is not necessary.

[manojkumar.bhosale, 2016/10/07 10:42:09]

Done.

+      YUY2ToUV422Row = YUY2ToUV422Row_Any_MSA;
+    }
+    if (IS_ALIGNED(width, 64)) {
+      YUY2ToYRow = YUY2ToYRow_MSA;
+      YUY2ToUV422Row = YUY2ToUV422Row_MSA;
+    }
+  }
+#endif
 
   for (y = 0; y < height; ++y) {
     YUY2ToUV422Row(src_yuy2, dst_u, dst_v, width);
@@ -552,6 +564,18 @@ int UYVYToI422(const uint8* src_uyvy, int src_stride_uyvy,
     }
   }
 #endif
+#if defined(HAS_UYVYTOYROW_MSA)
+  if (TestCpuFlag(kCpuHasMSA)) {
+    UYVYToYRow = UYVYToYRow_Any_MSA;
+    if (width >= 16) {
+      UYVYToUV422Row = UYVYToUV422Row_Any_MSA;
+    }
+    if (IS_ALIGNED(width, 64)) {
+      UYVYToYRow = UYVYToYRow_MSA;
+      UYVYToUV422Row = UYVYToUV422Row_MSA;
+    }
+  }
+#endif
 
   for (y = 0; y < height; ++y) {
     UYVYToUV422Row(src_uyvy, dst_u, dst_v, width);