ARM Skia NEON patches - 16/17 - Blitmask

src/opts/SkColor_opts_neon.h

Index: src/opts/SkColor_opts_neon.h
diff --git a/src/opts/SkColor_opts_neon.h b/src/opts/SkColor_opts_neon.h
index f81239712ab5bba90fb21b6896f7edbb4e658044..9f25b312dbd5a3393f92417a193c9cd009186656 100644
--- a/src/opts/SkColor_opts_neon.h
+++ b/src/opts/SkColor_opts_neon.h
@@ -2,6 +2,7 @@
 #define SkColor_opts_neon_DEFINED
 
 #include "SkTypes.h"
+#include "SkColorPriv.h"

[djsollen, 2013/11/11 17:21:46]

why do you need this include?

[kevin.petit.not.used.account, 2013/11/11 18:10:34]

This is here to fix a pre-existing mistake. There is code in this file using the
SK_?16_* defines. It didn't show up because all files including this one were
including SkColorPriv.h before.

 
 #include <arm_neon.h>
 
@@ -65,4 +66,17 @@ static inline uint16x8_t SkPixel32ToPixel16_neon8(uint8x8x4_t vsrc) {
     return ret;
 }
 
+static inline uint8x8_t SkBlend32_neon8(uint8x8_t src, uint8x8_t dst, uint16x8_t scale) {

[mtklein, 2013/11/11 17:17:04]

Add a brief comment like for 16->32 and 32->16 above?

[kevin.petit.not.used.account, 2013/11/11 18:10:34]

Done.

+    int16x8_t src_wide, dst_wide;
+
+    src_wide = vreinterpretq_s16_u16(vmovl_u8(src));
+    dst_wide = vreinterpretq_s16_u16(vmovl_u8(dst));
+
+    src_wide = (src_wide - dst_wide) * vreinterpretq_s16_u16(scale);
+
+    dst_wide += vshrq_n_s16(src_wide, 5);
+
+    return vmovn_u16(vreinterpretq_u16_s16(dst_wide));
+}
+
 #endif /* #ifndef SkColor_opts_neon_DEFINED */