Add ARGBExtractAlpha function

source/row_any.cc

 /*
  *  Copyright 2012 The LibYuv Project Authors. All rights reserved.
  *
  *  Use of this source code is governed by a BSD-style license
  *  that can be found in the LICENSE file in the root of the source
  *  tree. An additional intellectual property rights grant can be found
  *  in the file PATENTS. All contributing project authors may
  *  be found in the AUTHORS file in the root of the source tree.
  */
 
@@ skipping 279 matching lines @@
 ANY21C(NV12ToRGB565Row_Any_AVX2, NV12ToRGB565Row_AVX2, 1, 1, 2, 2, 15)
 #endif
 #ifdef HAS_NV12TORGB565ROW_NEON
 ANY21C(NV12ToRGB565Row_Any_NEON, NV12ToRGB565Row_NEON, 1, 1, 2, 2, 7)
 #endif
 #undef ANY21C
 
 // Any 1 to 1.
 #define ANY11(NAMEANY, ANY_SIMD, UVSHIFT, SBPP, BPP, MASK)                     \
     void NAMEANY(const uint8* src_ptr, uint8* dst_ptr, int width) {            \
       SIMD_ALIGNED(uint8 temp[128 * 2]);                                       \

[magjed_chromium, 2016/05/25 08:19:46]

What's the reason for using 128 instead of (MASK + 1) in these macros?

[fbarchard1, 2016/05/25 18:44:12]

simplicity.  they used to be the width of a single SIMD register, but some code
is unrolled to more than that and some use different BPP, so it was error prone.
If you feel inclined to change it, it could be a macro for max simd alignment,
or use 'mask', which is width - 1 of the SIMD loop alignment, it would be easier
to use the alignment value so you dont have to add + 1.  Then multiply by BPP
and consider subsampling if applicable.

       memset(temp, 0, 128);  /* for YUY2 and msan */                           \
       int r = width & MASK;                                                    \
       int n = width & ~MASK;                                                   \
       if (n > 0) {                                                             \
         ANY_SIMD(src_ptr, dst_ptr, n);                                         \
       }                                                                        \
       memcpy(temp, src_ptr + (n >> UVSHIFT) * SBPP, SS(r, UVSHIFT) * SBPP);    \
       ANY_SIMD(temp, temp + 128, MASK + 1);                                    \
       memcpy(dst_ptr + n * BPP, temp + 128, r * BPP);                          \
     }
@@ skipping 148 matching lines @@
 #endif
 #ifdef HAS_ARGBATTENUATEROW_AVX2
 ANY11(ARGBAttenuateRow_Any_AVX2, ARGBAttenuateRow_AVX2, 0, 4, 4, 7)
 #endif
 #ifdef HAS_ARGBUNATTENUATEROW_AVX2
 ANY11(ARGBUnattenuateRow_Any_AVX2, ARGBUnattenuateRow_AVX2, 0, 4, 4, 7)
 #endif
 #ifdef HAS_ARGBATTENUATEROW_NEON
 ANY11(ARGBAttenuateRow_Any_NEON, ARGBAttenuateRow_NEON, 0, 4, 4, 7)
 #endif
+#ifdef HAS_ARGBEXTRACTALPHAROW_SSE2
+ANY11(ARGBExtractAlphaRow_Any_SSE2, ARGBExtractAlphaRow_SSE2, 0, 4, 1, 7)
+#endif
+#ifdef HAS_ARGBEXTRACTALPHAROW_NEON
+ANY11(ARGBExtractAlphaRow_Any_NEON, ARGBExtractAlphaRow_NEON, 0, 4, 1, 7)
+#endif
 #undef ANY11
 
 // Any 1 to 1 with yuvconstants
 #define ANY11C(NAMEANY, ANY_SIMD, UVSHIFT, SBPP, BPP, MASK)                    \
     void NAMEANY(const uint8* src_ptr, uint8* dst_ptr,                         \
                  const struct YuvConstants* yuvconstants, int width) {         \
       SIMD_ALIGNED(uint8 temp[128 * 2]);                                       \
       memset(temp, 0, 128);  /* for YUY2 and msan */                           \
       int r = width & MASK;                                                    \
       int n = width & ~MASK;                                                   \
@@ skipping 30 matching lines @@
       }                                                                        \
       memcpy(temp, src_ptr + (n >> UVSHIFT) * SBPP, SS(r, UVSHIFT) * SBPP);    \
       memcpy(temp + 128, dst_ptr + n * BPP, r * BPP);                          \
       ANY_SIMD(temp, temp + 128, MASK + 1);                                    \
       memcpy(dst_ptr + n * BPP, temp + 128, r * BPP);                          \
     }
 
 #ifdef HAS_ARGBCOPYALPHAROW_AVX2
 ANY11B(ARGBCopyAlphaRow_Any_AVX2, ARGBCopyAlphaRow_AVX2, 0, 4, 4, 15)
 #endif
-#ifdef HAS_ARGBCOPYYTOALPHAROW_SSE2
+#ifdef HAS_ARGBCOPYALPHAROW_SSE2
 ANY11B(ARGBCopyAlphaRow_Any_SSE2, ARGBCopyAlphaRow_SSE2, 0, 4, 4, 7)
 #endif
 #ifdef HAS_ARGBCOPYYTOALPHAROW_AVX2
 ANY11B(ARGBCopyYToAlphaRow_Any_AVX2, ARGBCopyYToAlphaRow_AVX2, 0, 1, 4, 15)
 #endif
 #ifdef HAS_ARGBCOPYYTOALPHAROW_SSE2
 ANY11B(ARGBCopyYToAlphaRow_Any_SSE2, ARGBCopyYToAlphaRow_SSE2, 0, 1, 4, 7)
 #endif
 #undef ANY11B
 
@@ skipping 279 matching lines @@
 #endif
 #ifdef HAS_UYVYTOUVROW_NEON
 ANY12S(UYVYToUVRow_Any_NEON, UYVYToUVRow_NEON, 1, 4, 15)
 #endif
 #undef ANY12S
 
 #ifdef __cplusplus
 }  // extern "C"
 }  // namespace libyuv
 #endif