CopyRow_Neon reimplemented with vldm

BUG=none
TEST=none
Review URL: https://webrtc-codereview.appspot.com/412001

git-svn-id: http://libyuv.googlecode.com/svn/trunk@190 16f28f9a-4ce2-e073-06de-1de4eb20be90

README.chromium

 Name: libyuv
 URL: http://code.google.com/p/libyuv/
-Version: 189
+Version: 190
 License: BSD
 License File: LICENSE
 

include/libyuv/version.h

@@ -11,7 +11,7 @@
 #ifndef INCLUDE_LIBYUV_VERSION_H_
 #define INCLUDE_LIBYUV_VERSION_H_
 
-#define LIBYUV_VERSION 189
+#define LIBYUV_VERSION 190
 
 #endif  // INCLUDE_LIBYUV_VERSION_H_
 

source/row.h

@@ -67,16 +67,12 @@ extern "C" {
 #if defined(__ARM_NEON__) && !defined(YUV_DISABLE_ASM)
 #define HAS_MIRRORROW_NEON
 #define HAS_SPLITUV_NEON
+#define HAS_COPYROW_NEON
 #define HAS_I420TOARGBROW_NEON
 #define HAS_I420TOBGRAROW_NEON
 #define HAS_I420TOABGRROW_NEON
 #endif
 
-// The following are available on Neon platforms
-#if defined(__ARM_NEON__) && !defined(__thumb__) && !defined(YUV_DISABLE_ASM)
-#define HAS_COPYROW_NEON
-#endif
-
 #if defined(_MSC_VER)
 #define SIMD_ALIGNED(var) __declspec(align(16)) var
 typedef __declspec(align(16)) signed char vec8[16];

source/row_neon.cc

@@ -187,15 +187,15 @@ void SplitUV_NEON(const uint8* src_uv, uint8* dst_u, uint8* dst_v, int pix) {
 #endif
 
 #if defined(HAS_COPYROW_NEON)
-// TODO(fbarchard): Test with and without pld
-//  "pld        [%0, #0xC0]                    \n"  // preload
+// TODO(fbarchard): Test without pld on NexusS
 // Copy multiple of 64
 void CopyRow_NEON(const uint8* src, uint8* dst, int count) {
   asm volatile (
   "1:                                          \n"
-    "vld1.u8    {q0,q1,q2,q3}, [%0]!           \n"  // load 64
+    "pld        [%0, #0xC0]                    \n"  // preload
+    "vldm       %0!,{q0,q1,q2,q3}              \n"  // load 64
     "subs       %2, %2, #64                    \n"  // 64 processed per loop
-    "vst1.u8    {q0,q1,q2,q3}, [%1]!           \n"  // store 64
+    "vstm       %1!,{q0,q1,q2,q3}              \n"  // store 64
     "bhi        1b                             \n"
     : "+r"(src),   // %0
       "+r"(dst),   // %1