Add swizzle for rgb8888.

src/opts/SkNx_sse.h

 /*
  * Copyright 2015 Google Inc.
  *
  * Use of this source code is governed by a BSD-style license that can be
  * found in the LICENSE file.
  */
 
 #ifndef SkNx_sse_DEFINED
 #define SkNx_sse_DEFINED
 
@@ skipping 217 matching lines @@
         union { __m128i v; uint16_t us[8]; } pun = {fVec};
         return pun.us[k&7];
     }
 
     __m128i fVec;
 };
 
 template <>
 class SkNx<4, uint8_t> {
 public:
+    SkNx() {}
     SkNx(const __m128i& vec) : fVec(vec) {}
+    SkNx(uint8_t a, uint8_t b, uint8_t c, uint8_t d)
+        : fVec(_mm_setr_epi8(a,b,c,d, 0,0,0,0, 0,0,0,0, 0,0,0,0)) {}
 
-    SkNx() {}
+
     static SkNx Load(const void* ptr) { return _mm_cvtsi32_si128(*(const int*)ptr); }
     void store(void* ptr) const { *(int*)ptr = _mm_cvtsi128_si32(fVec); }
 
+    uint8_t operator[](int k) const {
+        SkASSERT(0 <= k && k < 4);
+        union { __m128i v; uint8_t us[16]; } pun = {fVec};
+        return pun.us[k&3];
+    }
+
     // TODO as needed
 
     __m128i fVec;
 };
 
 template <>
 class SkNx<16, uint8_t> {
 public:
     SkNx(const __m128i& vec) : fVec(vec) {}
 
@@ skipping 94 matching lines @@
 
 template<> /*static*/ inline Sk4h SkNx_cast<uint16_t, uint8_t>(const Sk4b& src) {
     return _mm_unpacklo_epi8(src.fVec, _mm_setzero_si128());
 }
 
 template<> /*static*/ inline Sk4b SkNx_cast<uint8_t, uint16_t>(const Sk4h& src) {
     return _mm_packus_epi16(src.fVec, src.fVec);
 }
 
 #endif//SkNx_sse_DEFINED