libvpx: Pull from upstream

source/libvpx/vpx_dsp/arm/vpx_convolve8_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.
  */
 
 #include <arm_neon.h>
+#include <assert.h>
 
 #include "./vpx_config.h"
 #include "./vpx_dsp_rtcd.h"
 #include "vpx/vpx_integer.h"
 #include "vpx_ports/mem.h"
 
-void vpx_convolve8_horiz_c(const uint8_t *src, ptrdiff_t src_stride,
-                           uint8_t *dst, ptrdiff_t dst_stride,
-                           const int16_t *filter_x, int x_step_q4,
-                           const int16_t *filter_y, int y_step_q4,
-                           int w, int h);
-void vpx_convolve8_vert_c(const uint8_t *src, ptrdiff_t src_stride,
-                           uint8_t *dst, ptrdiff_t dst_stride,
-                           const int16_t *filter_x, int x_step_q4,
-                           const int16_t *filter_y, int y_step_q4,
-                           int w, int h);
-
 static INLINE int32x4_t MULTIPLY_BY_Q0(
     int16x4_t dsrc0,
     int16x4_t dsrc1,
     int16x4_t dsrc2,
     int16x4_t dsrc3,
     int16x4_t dsrc4,
     int16x4_t dsrc5,
     int16x4_t dsrc6,
     int16x4_t dsrc7,
     int16x8_t q0s16) {
@@ skipping 36 matching lines @@
   uint16x4_t d2u16, d3u16, d4u16, d5u16, d16u16, d17u16, d18u16, d19u16;
   int16x8_t q0s16;
   uint16x8_t q1u16, q2u16, q8u16, q9u16, q10u16, q11u16, q12u16, q13u16;
   int32x4_t q1s32, q2s32, q14s32, q15s32;
   uint16x8x2_t q0x2u16;
   uint8x8x2_t d0x2u8, d1x2u8;
   uint32x2x2_t d0x2u32;
   uint16x4x2_t d0x2u16, d1x2u16;
   uint32x4x2_t q0x2u32;
 
-  if (x_step_q4 != 16) {
-    vpx_convolve8_horiz_c(src, src_stride, dst, dst_stride,
-                          filter_x, x_step_q4,
-                          filter_y, y_step_q4, w, h);
-    return;
-  }
+  assert(x_step_q4 == 16);
 
   q0s16 = vld1q_s16(filter_x);
 
   src -= 3;  // adjust for taps
   for (; h > 0; h -= 4,
     src += src_stride * 4,
     dst += dst_stride * 4) {  // loop_horiz_v
     s = src;
     d24u8 = vld1_u8(s);
     s += src_stride;
@@ skipping 147 matching lines @@
   uint8_t *d;
   uint32x2_t d2u32, d3u32;
   uint32x2_t d16u32, d18u32, d20u32, d22u32, d24u32, d26u32;
   int16x4_t d16s16, d17s16, d18s16, d19s16, d20s16, d21s16, d22s16;
   int16x4_t d24s16, d25s16, d26s16, d27s16;
   uint16x4_t d2u16, d3u16, d4u16, d5u16;
   int16x8_t q0s16;
   uint16x8_t q1u16, q2u16, q8u16, q9u16, q10u16, q11u16, q12u16, q13u16;
   int32x4_t q1s32, q2s32, q14s32, q15s32;
 
-  if (y_step_q4 != 16) {
-    vpx_convolve8_vert_c(src, src_stride, dst, dst_stride,
-                         filter_x, x_step_q4,
-                         filter_y, y_step_q4, w, h);
-    return;
-  }
+  assert(y_step_q4 == 16);
 
   src -= src_stride * 3;
   q0s16 = vld1q_s16(filter_y);
   for (; w > 0; w -= 4, src += 4, dst += 4) {  // loop_vert_h
     s = src;
     d16u32 = vld1_lane_u32((const uint32_t *)s, d16u32, 0);
     s += src_stride;
     d16u32 = vld1_lane_u32((const uint32_t *)s, d16u32, 1);
     s += src_stride;
     d18u32 = vld1_lane_u32((const uint32_t *)s, d18u32, 0);
@@ skipping 77 matching lines @@
 
       q8u16 = q10u16;
       d18s16 = d22s16;
       d19s16 = d24s16;
       q10u16 = q13u16;
       d22s16 = d25s16;
     }
   }
   return;
 }