libvpx: Pull from upstream

source/libvpx/vp8/common/arm/neon/vp8_subpixelvariance_neon.c

 /*
  *  Copyright (c) 2014 The WebM 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 14 matching lines @@
 
 unsigned int vp8_sub_pixel_variance16x16_neon_func(
         const unsigned char *src_ptr,
         int src_pixels_per_line,
         int xoffset,
         int yoffset,
         const unsigned char *dst_ptr,
         int dst_pixels_per_line,
         unsigned int *sse) {
     int i;
-    DECLARE_ALIGNED_ARRAY(16, unsigned char, tmp, 528);
+    DECLARE_ALIGNED(16, unsigned char, tmp[528]);
     unsigned char *tmpp;
     unsigned char *tmpp2;
     uint8x8_t d0u8, d1u8, d2u8, d3u8, d4u8, d5u8, d6u8, d7u8, d8u8, d9u8;
     uint8x8_t d10u8, d11u8, d12u8, d13u8, d14u8, d15u8, d16u8, d17u8, d18u8;
     uint8x8_t d19u8, d20u8, d21u8;
     int16x4_t d22s16, d23s16, d24s16, d25s16, d26s16, d27s16, d28s16, d29s16;
     uint32x2_t d0u32, d10u32;
     int64x1_t d0s64, d1s64, d2s64, d3s64;
     uint8x16_t q0u8, q1u8, q2u8, q3u8, q4u8, q5u8, q6u8, q7u8, q8u8, q9u8;
     uint8x16_t q10u8, q11u8, q12u8, q13u8, q14u8, q15u8;
@@ skipping 858 matching lines @@
     q5s64 = vmull_s32(vreinterpret_s32_s64(d0s64),
                       vreinterpret_s32_s64(d0s64));
     vst1_lane_u32((uint32_t *)sse, vreinterpret_u32_s64(d1s64), 0);
 
     d10u32 = vshr_n_u32(vreinterpret_u32_s64(vget_low_s64(q5s64)), 8);
     d0u32 = vsub_u32(vreinterpret_u32_s64(d1s64), d10u32);
 
     return vget_lane_u32(d0u32, 0);
 }
 
-enum { kWidth8 = 8 };
-enum { kHeight8 = 8 };
-enum { kHeight8PlusOne = 9 };
-enum { kPixelStepOne = 1 };
-enum { kAlign16 = 16 };
-
 #define FILTER_BITS 7
 
 static INLINE int horizontal_add_s16x8(const int16x8_t v_16x8) {
   const int32x4_t a = vpaddlq_s16(v_16x8);
   const int64x2_t b = vpaddlq_s32(a);
   const int32x2_t c = vadd_s32(vreinterpret_s32_s64(vget_low_s64(b)),
                                vreinterpret_s32_s64(vget_high_s64(b)));
   return vget_lane_s32(c, 0);
 }
 
@@ skipping 31 matching lines @@
   }
 
   *sum = horizontal_add_s16x8(v_sum);
   *sse = (unsigned int)horizontal_add_s32x4(vaddq_s32(v_sse_lo, v_sse_hi));
 }
 
 static unsigned int variance8x8_neon(const uint8_t *a, int a_stride,
                                      const uint8_t *b, int b_stride,
                                      unsigned int *sse) {
   int sum;
-  variance_neon_w8(a, a_stride, b, b_stride, kWidth8, kHeight8, sse, &sum);
-  return *sse - (((int64_t)sum * sum) / (kWidth8 * kHeight8));
+  variance_neon_w8(a, a_stride, b, b_stride, 8, 8, sse, &sum);
+  return *sse - (((int64_t)sum * sum) / (8 * 8));
 }
 
 static void var_filter_block2d_bil_w8(const uint8_t *src_ptr,
                                       uint8_t *output_ptr,
                                       unsigned int src_pixels_per_line,
                                       int pixel_step,
                                       unsigned int output_height,
                                       unsigned int output_width,
                                       const uint16_t *vpx_filter) {
   const uint8x8_t f0 = vmov_n_u8((uint8_t)vpx_filter[0]);
@@ skipping 13 matching lines @@
 }
 
 unsigned int vp8_sub_pixel_variance8x8_neon(
         const unsigned char *src,
         int src_stride,
         int xoffset,
         int yoffset,
         const unsigned char *dst,
         int dst_stride,
         unsigned int *sse) {
-  DECLARE_ALIGNED_ARRAY(kAlign16, uint8_t, temp2, kHeight8PlusOne * kWidth8);
-  DECLARE_ALIGNED_ARRAY(kAlign16, uint8_t, fdata3, kHeight8PlusOne * kWidth8);
+  DECLARE_ALIGNED(16, uint8_t, temp2[9 * 8]);
+  DECLARE_ALIGNED(16, uint8_t, fdata3[9 * 8]);
   if (xoffset == 0) {
-    var_filter_block2d_bil_w8(src, temp2, src_stride, kWidth8, kHeight8,
-                              kWidth8, bilinear_taps_coeff[yoffset]);
+    var_filter_block2d_bil_w8(src, temp2, src_stride, 8, 8,
+                              8, bilinear_taps_coeff[yoffset]);
   } else if (yoffset == 0) {
-    var_filter_block2d_bil_w8(src, temp2, src_stride, kPixelStepOne,
-                              kHeight8PlusOne, kWidth8,
+    var_filter_block2d_bil_w8(src, temp2, src_stride, 1,
+                              9, 8,
                               bilinear_taps_coeff[xoffset]);
   } else {
-    var_filter_block2d_bil_w8(src, fdata3, src_stride, kPixelStepOne,
-                              kHeight8PlusOne, kWidth8,
+    var_filter_block2d_bil_w8(src, fdata3, src_stride, 1,
+                              9, 8,
                               bilinear_taps_coeff[xoffset]);
-    var_filter_block2d_bil_w8(fdata3, temp2, kWidth8, kWidth8, kHeight8,
-                              kWidth8, bilinear_taps_coeff[yoffset]);
+    var_filter_block2d_bil_w8(fdata3, temp2, 8, 8, 8,
+                              8, bilinear_taps_coeff[yoffset]);
   }
-  return variance8x8_neon(temp2, kWidth8, dst, dst_stride, sse);
+  return variance8x8_neon(temp2, 8, dst, dst_stride, sse);
 }