Implement a speedup for Y-only blurs by transposing.

src/opts/SkBlurImage_opts_SSE2.cpp

 /*
  * Copyright 2013 The Android Open Source Project
  *
  * Use of this source code is governed by a BSD-style license that can be
  * found in the LICENSE file.
  */
 
 
 #include "SkBitmap.h"
 #include "SkColorPriv.h"
@@ skipping 89 matching lines @@
         }
         src += srcStrideY;
         dst += dstStrideY;
     }
 }
 
 } // namespace
 
 bool SkBoxBlurGetPlatformProcs_SSE2(SkBoxBlurProc* boxBlurX,
                                     SkBoxBlurProc* boxBlurY,
-                                    SkBoxBlurProc* boxBlurXY) {
+                                    SkBoxBlurProc* boxBlurXY,
+                                    SkBoxBlurProc* boxBlurYX) {
     *boxBlurX = SkBoxBlur_SSE2<kX, kX>;
     *boxBlurY = SkBoxBlur_SSE2<kY, kY>;
     *boxBlurXY = SkBoxBlur_SSE2<kX, kY>;
+    *boxBlurYX = SkBoxBlur_SSE2<kY, kX>;
     return true;
 }