fix convertToI420 rowbytes and add assembly calls

BUG=none
TEST=camtool
Review URL: http://webrtc-codereview.appspot.com/343008

git-svn-id: http://libyuv.googlecode.com/svn/trunk@130 16f28f9a-4ce2-e073-06de-1de4eb20be90

README.chromium

 Name: libyuv
 URL: http://code.google.com/p/libyuv/
-Version: 129
+Version: 130
 License: BSD
 License File: LICENSE
 

source/convert.cc

@@ -1293,24 +1293,36 @@ int ConvertToI420(const uint8* sample, size_t sample_size,
                  dst_width, inv_dst_height);
       break;
     case FOURCC_RGBP:
-      src = sample + (src_width * crop_y + crop_x) * 4;
-      RGB565ToI420(src, src_width * 4,
+      src = sample + (src_width * crop_y + crop_x) * 2;
+      RGB565ToI420(src, src_width * 2,
                    y, y_stride,
                    u, u_stride,
                    v, v_stride,
                    dst_width, inv_dst_height);
       break;
+//  V4L2_PIX_FMT_RGB555 'RGBO'
+//               Byte 0                    Byte 1
+//  Bit   7  6  5  4  3  2  1  0    7  6  5  4  3  2  1  0
+//       g2 g1 g0 b4 b3 b2 b1 b0    a r4 r3 r2 r1 r0 g4 g3
+//  Bit 15 14 13 12 11 10  9  8  7  6  5  4  3  2  1  0
+//       a r4 r3 r2 r1 r0 g4 g3 g2 g1 g0 b4 b3 b2 b1 b0
     case FOURCC_RGBO:
-      src = sample + (src_width * crop_y + crop_x) * 4;
-      ARGB1555ToI420(src, src_width * 4,
+      src = sample + (src_width * crop_y + crop_x) * 2;
+      ARGB1555ToI420(src, src_width * 2,
                      y, y_stride,
                      u, u_stride,
                      v, v_stride,
                      dst_width, inv_dst_height);
       break;
+//  V4L2_PIX_FMT_RGB444 'R444'
+//               Byte 0                    Byte 1
+//  Bit   7  6  5  4  3  2  1  0    7  6  5  4  3  2  1  0
+//       g3 g2 g1 g0 b3 b2 b1 b0   a3 a2 a1 a0 r3 r2 r1 r0
+//  Bit 15 14 13 12 11 10  9  8  7  6  5  4  3  2  1  0
+//      a3 a2 a1 a0 r3 r2 r1 r0 g3 g2 g1 g0 b3 b2 b1 b0
     case FOURCC_R444:
-      src = sample + (src_width * crop_y + crop_x) * 4;
-      ARGB4444ToI420(src, src_width * 4,
+      src = sample + (src_width * crop_y + crop_x) * 2;
+      ARGB4444ToI420(src, src_width * 2,
                      y, y_stride,
                      u, u_stride,
                      v, v_stride,

source/row.h

@@ -53,6 +53,11 @@
 #define HAS_REVERSE_ROW_SSE2
 #endif
 
+// The following are available on Windows platforms
+#if defined(_M_IX86) && !defined(YUV_DISABLE_ASM)
+#define HAS_ARGB4444TOARGBROW_SSE2
+#endif
+
 // The following are available on Neon platforms
 #if defined(__ARM_NEON__) && !defined(YUV_DISABLE_ASM)
 #define HAS_REVERSE_ROW_NEON
@@ -166,7 +171,9 @@ void RAWToARGBRow_SSSE3(const uint8* src_bg24, uint8* dst_argb, int pix);
 // TODO(fbarchard): SSE2 565 etc
 //void RGB565ToARGBRow_SSE2(const uint8* src_rgb, uint8* dst_argb, int pix);
 //void ARGB1555ToARGBRow_SSE2(const uint8* src_argb, uint8* dst_argb, int pix);
-//void ARGB4444ToARGBRow_SSE2(const uint8* src_argb, uint8* dst_argb, int pix);
+#endif
+#ifdef HAS_ARGB4444TOARGBROW_SSE2
+void ARGB4444ToARGBRow_SSE2(const uint8* src_argb, uint8* dst_argb, int pix);
 #endif
 void ABGRToARGBRow_C(const uint8* src_abgr, uint8* dst_argb, int pix);
 void BGRAToARGBRow_C(const uint8* src_bgra, uint8* dst_argb, int pix);

source/row_common.cc

@@ -108,10 +108,10 @@ void ARGB1555ToARGBRow_C(const uint8* src_rgb, uint8* dst_argb, int pix) {
 
 void ARGB4444ToARGBRow_C(const uint8* src_rgb, uint8* dst_argb, int pix) {
   for (int x = 0; x < pix; ++x) {
-    uint8 b = src_rgb[0] & 0x1f;
-    uint8 g = src_rgb[0] >> 4;
-    uint8 r = src_rgb[1] & 0x1f;
     uint8 a = src_rgb[1] >> 4;
+    uint8 r = src_rgb[1] & 0x0f;
+    uint8 g = src_rgb[0] >> 4;
+    uint8 b = src_rgb[0] & 0x0f;
     dst_argb[0] = (b << 4) | b;
     dst_argb[1] = (g << 4) | g;
     dst_argb[2] = (r << 4) | r;
@@ -270,7 +270,11 @@ void ARGB1555ToYRow_SSSE3(const uint8* src_argb, uint8* dst_y, int pix) {
 
 void ARGB4444ToYRow_SSSE3(const uint8* src_argb, uint8* dst_y, int pix) {
   SIMD_ALIGNED(uint8 row[kMaxStride]);
+#ifdef HAS_ARGB4444TOARGBROW_SSE2
+  ARGB4444ToARGBRow_SSE2(src_argb, row, pix);
+#else
   ARGB4444ToARGBRow_C(src_argb, row, pix);
+#endif
   ARGBToYRow_SSSE3(row, dst_y, pix);
 }
 #endif
@@ -313,8 +317,13 @@ void ARGB1555ToUVRow_SSSE3(const uint8* src_argb, int src_stride_argb,
 void ARGB4444ToUVRow_SSSE3(const uint8* src_argb, int src_stride_argb,
                            uint8* dst_u, uint8* dst_v, int pix) {
   SIMD_ALIGNED(uint8 row[kMaxStride * 2]);
+#ifdef HAS_ARGB4444TOARGBROW_SSE2
+  ARGB4444ToARGBRow_SSE2(src_argb, row, pix);
+  ARGB4444ToARGBRow_SSE2(src_argb + src_stride_argb, row + kMaxStride, pix);
+#else
   ARGB4444ToARGBRow_C(src_argb, row, pix);
   ARGB4444ToARGBRow_C(src_argb + src_stride_argb, row + kMaxStride, pix);
+#endif
   ARGBToUVRow_SSSE3(row, kMaxStride, dst_u, dst_v, pix);
 }