Fix an overflow issue in SSE2 forward ADST

The step that sums three input samples could potentially cause the
intermediate result go beyond 16 bit limit, when operating as the
second 1-D transform. This commit fixes the issue.

Change-Id: Iaf512449ac2d25ddd8a806d760afab362c62a516

vp9/encoder/x86/vp9_dct_sse2.c

@@ -206,12 +206,12 @@ void fadst4_1d_sse2(__m128i *in) {
   const __m128i k__DCT_CONST_ROUNDING = _mm_set1_epi32(DCT_CONST_ROUNDING);
   __m128i u[8], v[8];
   __m128i in7 = _mm_add_epi16(in[0], in[1]);
-  in7 = _mm_sub_epi16(in7, in[3]);
 
   u[0] = _mm_unpacklo_epi16(in[0], in[1]);
   u[1] = _mm_unpacklo_epi16(in[2], in[3]);
   u[2] = _mm_unpacklo_epi16(in7, kZero);
   u[3] = _mm_unpacklo_epi16(in[2], kZero);
+  u[4] = _mm_unpacklo_epi16(in[3], kZero);
 
   v[0] = _mm_madd_epi16(u[0], k__sinpi_p01_p02);  // s0 + s2
   v[1] = _mm_madd_epi16(u[1], k__sinpi_p03_p04);  // s4 + s5
@@ -219,9 +219,10 @@ void fadst4_1d_sse2(__m128i *in) {
   v[3] = _mm_madd_epi16(u[0], k__sinpi_p04_m01);  // s1 - s3
   v[4] = _mm_madd_epi16(u[1], k__sinpi_m03_p02);  // -s4 + s6
   v[5] = _mm_madd_epi16(u[3], k__sinpi_p03_p03);  // s4
+  v[6] = _mm_madd_epi16(u[4], k__sinpi_p03_p03);
 
   u[0] = _mm_add_epi32(v[0], v[1]);
-  u[1] = v[2];
+  u[1] = _mm_sub_epi32(v[2], v[6]);
   u[2] = _mm_add_epi32(v[3], v[4]);
   u[3] = _mm_sub_epi32(u[2], u[0]);
   u[4] = _mm_slli_epi32(v[5], 2);