Merge changes I4850b36e,Ic4d7128a into nextgenv2

* changes:
  variance_avx2: sync variance functions with c-code
  Resolve -Wshorten-64-to-32 in variance.

aom_dsp/variance.c

@@ -148,7 +148,7 @@ static void var_filter_block2d_bil_second_pass(const uint16_t *a, uint8_t *b,
                                      uint32_t *sse) {                \
     int sum;                                                         \
     variance(a, a_stride, b, b_stride, W, H, sse, &sum);             \
-    return *sse - (((int64_t)sum * sum) / (W * H));                  \
+    return *sse - (uint32_t)(((int64_t)sum * sum) / (W * H));        \
   }
 
 #define SUBPIX_VAR(W, H)                                                \

aom_dsp/x86/variance_avx2.c

@@ -62,7 +62,7 @@ unsigned int aom_variance32x16_avx2(const uint8_t *src, int src_stride,
   int sum;
   variance_avx2(src, src_stride, ref, ref_stride, 32, 16, sse, &sum,
                 aom_get32x32var_avx2, 32);
-  return *sse - (((int64_t)sum * sum) >> 9);
+  return *sse - (uint32_t)(((int64_t)sum * sum) >> 9);
 }
 
 unsigned int aom_variance32x32_avx2(const uint8_t *src, int src_stride,
@@ -71,7 +71,7 @@ unsigned int aom_variance32x32_avx2(const uint8_t *src, int src_stride,
   int sum;
   variance_avx2(src, src_stride, ref, ref_stride, 32, 32, sse, &sum,
                 aom_get32x32var_avx2, 32);
-  return *sse - (((int64_t)sum * sum) >> 10);
+  return *sse - (uint32_t)(((int64_t)sum * sum) >> 10);
 }
 
 unsigned int aom_variance64x64_avx2(const uint8_t *src, int src_stride,
@@ -80,7 +80,7 @@ unsigned int aom_variance64x64_avx2(const uint8_t *src, int src_stride,
   int sum;
   variance_avx2(src, src_stride, ref, ref_stride, 64, 64, sse, &sum,
                 aom_get32x32var_avx2, 32);
-  return *sse - (((int64_t)sum * sum) >> 12);
+  return *sse - (uint32_t)(((int64_t)sum * sum) >> 12);
 }
 
 unsigned int aom_variance64x32_avx2(const uint8_t *src, int src_stride,
@@ -89,7 +89,7 @@ unsigned int aom_variance64x32_avx2(const uint8_t *src, int src_stride,
   int sum;
   variance_avx2(src, src_stride, ref, ref_stride, 64, 32, sse, &sum,
                 aom_get32x32var_avx2, 32);
-  return *sse - (((int64_t)sum * sum) >> 11);
+  return *sse - (uint32_t)(((int64_t)sum * sum) >> 11);
 }
 
 unsigned int aom_sub_pixel_variance32xh_avx2(const uint8_t *src, int src_stride,
@@ -116,7 +116,7 @@ unsigned int aom_sub_pixel_variance64x64_avx2(const uint8_t *src,
                                       dst + 32, dst_stride, 64, &sse2);
   const int se = se1 + se2;
   *sse = sse1 + sse2;
-  return *sse - (((int64_t)se * se) >> 12);
+  return *sse - (uint32_t)(((int64_t)se * se) >> 12);
 }
 
 unsigned int aom_sub_pixel_variance32x32_avx2(const uint8_t *src,
@@ -126,7 +126,7 @@ unsigned int aom_sub_pixel_variance32x32_avx2(const uint8_t *src,
                                               unsigned int *sse) {
   const int se = aom_sub_pixel_variance32xh_avx2(
       src, src_stride, x_offset, y_offset, dst, dst_stride, 32, sse);
-  return *sse - (((int64_t)se * se) >> 10);
+  return *sse - (uint32_t)(((int64_t)se * se) >> 10);
 }
 
 unsigned int aom_sub_pixel_avg_variance64x64_avx2(
@@ -143,7 +143,7 @@ unsigned int aom_sub_pixel_avg_variance64x64_avx2(
 
   *sse = sse1 + sse2;
 
-  return *sse - (((int64_t)se * se) >> 12);
+  return *sse - (uint32_t)(((int64_t)se * se) >> 12);
 }
 
 unsigned int aom_sub_pixel_avg_variance32x32_avx2(
@@ -152,5 +152,5 @@ unsigned int aom_sub_pixel_avg_variance32x32_avx2(
   // Process 32 elements in parallel.
   const int se = aom_sub_pixel_avg_variance32xh_avx2(
       src, src_stride, x_offset, y_offset, dst, dst_stride, sec, 32, 32, sse);
-  return *sse - (((int64_t)se * se) >> 10);
+  return *sse - (uint32_t)(((int64_t)se * se) >> 10);
 }

aom_dsp/x86/variance_sse2.c

@@ -249,7 +249,7 @@ unsigned int aom_variance32x32_sse2(const uint8_t *src, int src_stride,
                 aom_get16x16var_sse2, 16);
   assert(sum <= 255 * 32 * 32);
   assert(sum >= -255 * 32 * 32);
-  return *sse - (((int64_t)sum * sum) >> 10);
+  return *sse - (unsigned int)(((int64_t)sum * sum) >> 10);
 }
 
 unsigned int aom_variance32x16_sse2(const uint8_t *src, int src_stride,
@@ -260,7 +260,7 @@ unsigned int aom_variance32x16_sse2(const uint8_t *src, int src_stride,
                 aom_get16x16var_sse2, 16);
   assert(sum <= 255 * 32 * 16);
   assert(sum >= -255 * 32 * 16);
-  return *sse - (((int64_t)sum * sum) >> 9);
+  return *sse - (unsigned int)(((int64_t)sum * sum) >> 9);
 }
 
 unsigned int aom_variance16x32_sse2(const uint8_t *src, int src_stride,
@@ -271,7 +271,7 @@ unsigned int aom_variance16x32_sse2(const uint8_t *src, int src_stride,
                 aom_get16x16var_sse2, 16);
   assert(sum <= 255 * 32 * 16);
   assert(sum >= -255 * 32 * 16);
-  return *sse - (((int64_t)sum * sum) >> 9);
+  return *sse - (unsigned int)(((int64_t)sum * sum) >> 9);
 }
 
 unsigned int aom_variance64x64_sse2(const uint8_t *src, int src_stride,
@@ -282,7 +282,7 @@ unsigned int aom_variance64x64_sse2(const uint8_t *src, int src_stride,
                 aom_get16x16var_sse2, 16);
   assert(sum <= 255 * 64 * 64);
   assert(sum >= -255 * 64 * 64);
-  return *sse - (((int64_t)sum * sum) >> 12);
+  return *sse - (unsigned int)(((int64_t)sum * sum) >> 12);
 }
 
 unsigned int aom_variance64x32_sse2(const uint8_t *src, int src_stride,
@@ -293,7 +293,7 @@ unsigned int aom_variance64x32_sse2(const uint8_t *src, int src_stride,
                 aom_get16x16var_sse2, 16);
   assert(sum <= 255 * 64 * 32);
   assert(sum >= -255 * 64 * 32);
-  return *sse - (((int64_t)sum * sum) >> 11);
+  return *sse - (unsigned int)(((int64_t)sum * sum) >> 11);
 }
 
 unsigned int aom_variance32x64_sse2(const uint8_t *src, int src_stride,
@@ -304,7 +304,7 @@ unsigned int aom_variance32x64_sse2(const uint8_t *src, int src_stride,
                 aom_get16x16var_sse2, 16);
   assert(sum <= 255 * 64 * 32);
   assert(sum >= -255 * 64 * 32);
-  return *sse - (((int64_t)sum * sum) >> 11);
+  return *sse - (unsigned int)(((int64_t)sum * sum) >> 11);
 }
 
 unsigned int aom_mse8x8_sse2(const uint8_t *src, int src_stride,
@@ -381,7 +381,7 @@ DECLS(ssse3, ssse3);
       }                                                                        \
     }                                                                          \
     *sse_ptr = sse;                                                            \
-    return sse - (cast_prod(cast se * se) >> (wlog2 + hlog2));                 \
+    return sse - (unsigned int)(cast_prod(cast se * se) >> (wlog2 + hlog2));   \
   }
 
 #define FNS(opt1, opt2)                              \
@@ -452,7 +452,7 @@ DECLS(ssse3, ssse3);
       }                                                                        \
     }                                                                          \
     *sseptr = sse;                                                             \
-    return sse - (cast_prod(cast se * se) >> (wlog2 + hlog2));                 \
+    return sse - (unsigned int)(cast_prod(cast se * se) >> (wlog2 + hlog2));   \
   }
 
 #define FNS(opt1, opt2)                              \