Implement a speedup for Y-only blurs by transposing.

src/opts/opts_check_SSE2.cpp

 /*
  * Copyright 2009 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 "SkBitmapProcState_opts_SSE2.h"
 #include "SkBitmapProcState_opts_SSSE3.h"
 #include "SkBitmapFilter_opts_SSE2.h"
@@ skipping 253 matching lines @@
             return SkErodeX_SSE2;
         case kErodeY_SkMorphologyProcType:
             return SkErodeY_SSE2;
         default:
             return NULL;
     }
 }
 
 bool SkBoxBlurGetPlatformProcs(SkBoxBlurProc* boxBlurX,
                                SkBoxBlurProc* boxBlurY,
-                               SkBoxBlurProc* boxBlurXY) {
+                               SkBoxBlurProc* boxBlurXY,
+                               SkBoxBlurProc* boxBlurYX) {
 #ifdef SK_DISABLE_BLUR_DIVISION_OPTIMIZATION
     return false;
 #else
     if (!cachedHasSSE2()) {
         return false;
     }
-    return SkBoxBlurGetPlatformProcs_SSE2(boxBlurX, boxBlurY, boxBlurXY);
+    return SkBoxBlurGetPlatformProcs_SSE2(boxBlurX, boxBlurY, boxBlurXY, boxBlurYX);
 #endif
 }
 
 SkBlitRow::ColorRectProc PlatformColorRectProcFactory(); // suppress warning
 
 SkBlitRow::ColorRectProc PlatformColorRectProcFactory() {
     if (cachedHasSSE2()) {
         return ColorRect32_SSE2;
     } else {
         return NULL;
     }
 }