libvpx: Pull from upstream

source/libvpx/vp9/encoder/x86/vp9_avg_intrin_sse2.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 43 matching lines @@
   s0 = _mm_adds_epu16(s0, _mm_srli_si128(s0, 4));
   s0 = _mm_adds_epu16(s0, _mm_srli_epi64(s0, 16));
   avg = _mm_extract_epi16(s0, 0);
   return (avg + 8) >> 4;
 }
 
 void vp9_int_pro_row_sse2(int16_t *hbuf, uint8_t const*ref,
                           const int ref_stride, const int height) {
   int idx;
   __m128i zero = _mm_setzero_si128();
-  __m128i src_line = _mm_load_si128((const __m128i *)ref);
+  __m128i src_line = _mm_loadu_si128((const __m128i *)ref);
   __m128i s0 = _mm_unpacklo_epi8(src_line, zero);
   __m128i s1 = _mm_unpackhi_epi8(src_line, zero);
   __m128i t0, t1;
   int height_1 = height - 1;
   ref += ref_stride;
 
   for (idx = 1; idx < height_1; idx += 2) {
-    src_line = _mm_load_si128((const __m128i *)ref);
+    src_line = _mm_loadu_si128((const __m128i *)ref);
     t0 = _mm_unpacklo_epi8(src_line, zero);
     t1 = _mm_unpackhi_epi8(src_line, zero);
     s0 = _mm_adds_epu16(s0, t0);
     s1 = _mm_adds_epu16(s1, t1);
     ref += ref_stride;
 
-    src_line = _mm_load_si128((const __m128i *)ref);
+    src_line = _mm_loadu_si128((const __m128i *)ref);
     t0 = _mm_unpacklo_epi8(src_line, zero);
     t1 = _mm_unpackhi_epi8(src_line, zero);
     s0 = _mm_adds_epu16(s0, t0);
     s1 = _mm_adds_epu16(s1, t1);
     ref += ref_stride;
   }
 
-  src_line = _mm_load_si128((const __m128i *)ref);
+  src_line = _mm_loadu_si128((const __m128i *)ref);
   t0 = _mm_unpacklo_epi8(src_line, zero);
   t1 = _mm_unpackhi_epi8(src_line, zero);
   s0 = _mm_adds_epu16(s0, t0);
   s1 = _mm_adds_epu16(s1, t1);
 
   if (height == 64) {
     s0 = _mm_srai_epi16(s0, 5);
     s1 = _mm_srai_epi16(s1, 5);
   } else if (height == 32) {
     s0 = _mm_srai_epi16(s0, 4);
     s1 = _mm_srai_epi16(s1, 4);
   } else {
     s0 = _mm_srai_epi16(s0, 3);
     s1 = _mm_srai_epi16(s1, 3);
   }
 
-  _mm_store_si128((__m128i *)hbuf, s0);
+  _mm_storeu_si128((__m128i *)hbuf, s0);
   hbuf += 8;
-  _mm_store_si128((__m128i *)hbuf, s1);
+  _mm_storeu_si128((__m128i *)hbuf, s1);
 }
 
 int16_t vp9_int_pro_col_sse2(uint8_t const *ref, const int width) {
   __m128i zero = _mm_setzero_si128();
   __m128i src_line = _mm_load_si128((const __m128i *)ref);
   __m128i s0 = _mm_sad_epu8(src_line, zero);
   __m128i s1;
   int i;
   const int norm_factor = 3 + (width >> 5);
 
@@ skipping 46 matching lines @@
 
   v1  = _mm_srli_si128(sse, 8);
   sse = _mm_add_epi32(sse, v1);
   v1  = _mm_srli_epi64(sse, 32);
   sse = _mm_add_epi32(sse, v1);
 
   mean = _mm_extract_epi16(sum, 0);
 
   return _mm_cvtsi128_si32(sse) - ((mean * mean) >> (bwl + 2));
 }