Revert of try stosd/w

include/core/SkUtils.h

 /*
  * Copyright 2006 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.
  */
 
 #ifndef SkUtils_DEFINED
 #define SkUtils_DEFINED
 
 #include "SkTypes.h"
-#if defined(SK_BUILD_FOR_WIN)
-    #include <intrin.h>
-#endif
 
 namespace SkOpts {
     extern void (*memset16)(uint16_t[], uint16_t, int);
     extern void (*memset32)(uint32_t[], uint32_t, int);
 }
 
 ///////////////////////////////////////////////////////////////////////////////
 
 // The inlining heuristics below were determined using bench/MemsetBench.cpp
 // on a x86 desktop, a Nexus 7 with and without NEON, and a Nexus 9:
 //   - on x86, inlining was never faster,
 //   - on ARMv7, inlining was faster for N<=10.  Putting this check inside the NEON
 //     code was not helpful; it's got to be here outside.
 //   - NEON code generation for ARMv8 with GCC 4.9 is terrible,
 //     making the NEON code ~8x slower that just a serial loop.
 
 /** Similar to memset(), but it assigns a 16bit value into the buffer.
     @param buffer   The memory to have value copied into it
     @param value    The 16bit value to be copied into buffer
     @param count    The number of times value should be copied into the buffer.
 */
 static inline void sk_memset16(uint16_t buffer[], uint16_t value, int count) {
-#if defined(SK_BUILD_FOR_WIN)
-    __stosw(buffer, value, count);
-#elif defined(SK_CPU_X86)
-    __asm__ __volatile__ ( "rep stosw" : "+D"(buffer), "+c"(count) : "a"(value) );
-#elif defined(SK_ARM_HAS_NEON)
-    while (count --> 0) { *buffer++ = value; }
-#else
+#if defined(SK_CPU_ARM64)
+    while (count --> 0) { *buffer++ = value; } return;
+#elif defined(SK_CPU_ARM32)
+    if (count <= 10) { while (count --> 0) { *buffer++ = value; } return; }
+#endif
     SkOpts::memset16(buffer, value, count);
-#endif
 }
 
 /** Similar to memset(), but it assigns a 32bit value into the buffer.
     @param buffer   The memory to have value copied into it
     @param value    The 32bit value to be copied into buffer
     @param count    The number of times value should be copied into the buffer.
 */
 static inline void sk_memset32(uint32_t buffer[], uint32_t value, int count) {
-#if defined(SK_BUILD_FOR_WIN)
-    __stosd((PDWORD)buffer, value, count);
-#elif defined(SK_CPU_X86)
-    __asm__ __volatile__ ( "rep stosl" : "+D"(buffer), "+c"(count) : "a"(value) );
-#elif defined(SK_ARM_HAS_NEON)
-    while (count --> 0) { *buffer++ = value; }
-#else
+#if defined(SK_CPU_ARM64)
+    while (count --> 0) { *buffer++ = value; } return;
+#elif defined(SK_CPU_ARM32)
+    if (count <= 10) { while (count --> 0) { *buffer++ = value; } return; }
+#endif
     SkOpts::memset32(buffer, value, count);
-#endif
 }
 
 
 ///////////////////////////////////////////////////////////////////////////////
 
 #define kMaxBytesInUTF8Sequence     4
 
 #ifdef SK_DEBUG
     int SkUTF8_LeadByteToCount(unsigned c);
 #else
@@ skipping 60 matching lines @@
     }
     ~SkAutoTrace() {
         SkDebugf("--- trace: %s Leave\n", fLabel);
     }
 private:
     const char* fLabel;
 };
 #define SkAutoTrace(...) SK_REQUIRE_LOCAL_VAR(SkAutoTrace)
 
 #endif