Removes trailing spaces. Removes tabs.

[vlc] / modules / video_chroma / i422_yuy2.h

diff --git a/modules/video_chroma/i422_yuy2.h b/modules/video_chroma/i422_yuy2.h
index 1b0405bb2750f66bd47138a00bcd7ccbda64b67e..52ad7e4c3fd571c65b319714141afcceb4ad58d6 100644 (file)
--- a/modules/video_chroma/i422_yuy2.h
+++ b/modules/video_chroma/i422_yuy2.h
@@ -11,7 +11,7 @@
  * it under the terms of the GNU General Public License as published by
  * the Free Software Foundation; either version 2 of the License, or
  * (at your option) any later version.
- * 
+ *
  * This program is distributed in the hope that it will be useful,
  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
@@ -107,7 +107,7 @@ movq      %%mm1, 8(%0)  # Store high UYVY                                 \n\
     *(uint64_t*)p_line = (uint64_t)mm2;     \
     mm0 = _mm_unpackhi_pi8(mm0, mm1);       \
     *(uint64_t*)(p_line+8) = (uint64_t)mm0;
-    
+ 
 #define MMX_YUV422_YVYU                     \
     mm0 = (__m64)*(uint64_t*)p_y;           \
     mm2 = _mm_cvtsi32_si64(*(int*)p_u);     \
@@ -233,9 +233,82 @@ movdqu    %%xmm1, 16(%0)  # Store high UYVY                               \n\
 
 #include <emmintrin.h>
 
+#define SSE2_CALL(SSE2_INSTRUCTIONS)    \
+    do {                                \
+        __m128i xmm0, xmm1, xmm2;        \
+        SSE2_INSTRUCTIONS               \
+        p_line += 32; p_y += 16;        \
+        p_u += 8; p_v += 8;             \
+    } while(0)
 
 #define SSE2_END  _mm_sfence()
 
+#define SSE2_YUV422_YUYV_ALIGNED                \
+    xmm0 = _mm_load_si128((__m128i *)p_y);      \
+    xmm1 = _mm_loadl_epi64((__m128i *)p_u);     \
+    xmm2 = _mm_loadl_epi64((__m128i *)p_v);     \
+    xmm1 = _mm_unpacklo_epi8(xmm1, xmm2);       \
+    xmm2 = xmm0;                                \
+    xmm2 = _mm_unpacklo_epi8(xmm2, xmm1);       \
+    _mm_stream_si128((__m128i*)(p_line), xmm2); \
+    xmm0 = _mm_unpackhi_epi8(xmm0, xmm1);       \
+    _mm_stream_si128((__m128i*)(p_line+16), xmm0);
+ 
+#define SSE2_YUV422_YUYV_UNALIGNED              \
+    xmm0 = _mm_loadu_si128((__m128i *)p_y);     \
+    xmm1 = _mm_loadl_epi64((__m128i *)p_u);     \
+    xmm2 = _mm_loadl_epi64((__m128i *)p_v);     \
+    xmm1 = _mm_unpacklo_epi8(xmm1, xmm2);       \
+    xmm2 = xmm0;                                \
+    xmm2 = _mm_unpacklo_epi8(xmm2, xmm1);       \
+    _mm_storeu_si128((__m128i*)(p_line), xmm2); \
+    xmm0 = _mm_unpackhi_epi8(xmm0, xmm1);       \
+    _mm_storeu_si128((__m128i*)(p_line+16), xmm0);
+ 
+#define SSE2_YUV422_YVYU_ALIGNED                \
+    xmm0 = _mm_load_si128((__m128i *)p_y);      \
+    xmm2 = _mm_loadl_epi64((__m128i *)p_u);     \
+    xmm1 = _mm_loadl_epi64((__m128i *)p_v);     \
+    xmm1 = _mm_unpacklo_epi8(xmm1, xmm2);       \
+    xmm2 = xmm0;                                \
+    xmm2 = _mm_unpacklo_epi8(xmm2, xmm1);       \
+    _mm_stream_si128((__m128i*)(p_line), xmm2); \
+    xmm0 = _mm_unpackhi_epi8(xmm0, xmm1);       \
+    _mm_stream_si128((__m128i*)(p_line+16), xmm0);
+
+#define SSE2_YUV422_YVYU_UNALIGNED              \
+    xmm0 = _mm_loadu_si128((__m128i *)p_y);     \
+    xmm2 = _mm_loadl_epi64((__m128i *)p_u);     \
+    xmm1 = _mm_loadl_epi64((__m128i *)p_v);     \
+    xmm1 = _mm_unpacklo_epi8(xmm1, xmm2);       \
+    xmm2 = xmm0;                                \
+    xmm2 = _mm_unpacklo_epi8(xmm2, xmm1);       \
+    _mm_storeu_si128((__m128i*)(p_line), xmm2); \
+    xmm0 = _mm_unpackhi_epi8(xmm0, xmm1);       \
+    _mm_storeu_si128((__m128i*)(p_line+16), xmm0);
+
+#define SSE2_YUV422_UYVY_ALIGNED                \
+    xmm0 = _mm_load_si128((__m128i *)p_y);      \
+    xmm1 = _mm_loadl_epi64((__m128i *)p_u);     \
+    xmm2 = _mm_loadl_epi64((__m128i *)p_v);     \
+    xmm1 = _mm_unpacklo_epi8(xmm1, xmm2);       \
+    xmm2 = xmm1;                                \
+    xmm2 = _mm_unpacklo_epi8(xmm2, xmm0);       \
+    _mm_stream_si128((__m128i*)(p_line), xmm2); \
+    xmm1 = _mm_unpackhi_epi8(xmm1, xmm0);       \
+    _mm_stream_si128((__m128i*)(p_line+16), xmm1);
+
+#define SSE2_YUV422_UYVY_UNALIGNED              \
+    xmm0 = _mm_loadu_si128((__m128i *)p_y);     \
+    xmm1 = _mm_loadl_epi64((__m128i *)p_u);     \
+    xmm2 = _mm_loadl_epi64((__m128i *)p_v);     \
+    xmm1 = _mm_unpacklo_epi8(xmm1, xmm2);       \
+    xmm2 = xmm1;                                \
+    xmm2 = _mm_unpacklo_epi8(xmm2, xmm0);       \
+    _mm_storeu_si128((__m128i*)(p_line), xmm2); \
+    xmm1 = _mm_unpackhi_epi8(xmm1, xmm0);       \
+    _mm_storeu_si128((__m128i*)(p_line+16), xmm1);
+
 #endif
 
 #endif