Ducking fix #2: Removed the aec state as an input parameter to the FilterFar function.

webrtc/modules/audio_processing/aec/aec_core_sse2.c

 /*
  *  Copyright (c) 2011 The WebRTC 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 11 matching lines @@
 #include "webrtc/modules/audio_processing/aec/aec_rdft.h"
 
 __inline static float MulRe(float aRe, float aIm, float bRe, float bIm) {
   return aRe * bRe - aIm * bIm;
 }
 
 __inline static float MulIm(float aRe, float aIm, float bRe, float bIm) {
   return aRe * bIm + aIm * bRe;
 }
 
-static void FilterFarSSE2(AecCore* aec, float yf[2][PART_LEN1]) {
+static void FilterFarSSE2(int num_partitions,
+                          int xfBufBlockPos,
+                          float xfBuf[2][kExtendedNumPartitions * PART_LEN1],
+                          float wfBuf[2][kExtendedNumPartitions * PART_LEN1],
+                          float yf[2][PART_LEN1]) {
+
   int i;
-  const int num_partitions = aec->num_partitions;
-  for (i = 0; i < num_partitions; i++) {
+  const int num_partitions_local = num_partitions;

[hlundin-webrtc, 2015/11/20 10:59:40]

Same here: unnecessary?

[peah-webrtc, 2015/11/24 06:48:56]

Same comment applies as for Neon. Could you please verify that you think it is
ok also here.

[hlundin-webrtc, 2015/11/25 12:49:40]

Acknowledged.

+  for (i = 0; i < num_partitions_local; i++) {
     int j;
-    int xPos = (i + aec->xfBufBlockPos) * PART_LEN1;
+    int xPos = (i + xfBufBlockPos) * PART_LEN1;
     int pos = i * PART_LEN1;
     // Check for wrap
-    if (i + aec->xfBufBlockPos >= num_partitions) {
-      xPos -= num_partitions * (PART_LEN1);
+    if (i + xfBufBlockPos >= num_partitions_local) {
+      xPos -= num_partitions_local * (PART_LEN1);
     }
 
     // vectorized code (four at once)
     for (j = 0; j + 3 < PART_LEN1; j += 4) {
-      const __m128 xfBuf_re = _mm_loadu_ps(&aec->xfBuf[0][xPos + j]);
-      const __m128 xfBuf_im = _mm_loadu_ps(&aec->xfBuf[1][xPos + j]);
-      const __m128 wfBuf_re = _mm_loadu_ps(&aec->wfBuf[0][pos + j]);
-      const __m128 wfBuf_im = _mm_loadu_ps(&aec->wfBuf[1][pos + j]);
+      const __m128 xfBuf_re = _mm_loadu_ps(&xfBuf[0][xPos + j]);
+      const __m128 xfBuf_im = _mm_loadu_ps(&xfBuf[1][xPos + j]);
+      const __m128 wfBuf_re = _mm_loadu_ps(&wfBuf[0][pos + j]);
+      const __m128 wfBuf_im = _mm_loadu_ps(&wfBuf[1][pos + j]);
       const __m128 yf_re = _mm_loadu_ps(&yf[0][j]);
       const __m128 yf_im = _mm_loadu_ps(&yf[1][j]);
       const __m128 a = _mm_mul_ps(xfBuf_re, wfBuf_re);
       const __m128 b = _mm_mul_ps(xfBuf_im, wfBuf_im);
       const __m128 c = _mm_mul_ps(xfBuf_re, wfBuf_im);
       const __m128 d = _mm_mul_ps(xfBuf_im, wfBuf_re);
       const __m128 e = _mm_sub_ps(a, b);
       const __m128 f = _mm_add_ps(c, d);
       const __m128 g = _mm_add_ps(yf_re, e);
       const __m128 h = _mm_add_ps(yf_im, f);
       _mm_storeu_ps(&yf[0][j], g);
       _mm_storeu_ps(&yf[1][j], h);
     }
     // scalar code for the remaining items.
     for (; j < PART_LEN1; j++) {
-      yf[0][j] += MulRe(aec->xfBuf[0][xPos + j],
-                        aec->xfBuf[1][xPos + j],
-                        aec->wfBuf[0][pos + j],
-                        aec->wfBuf[1][pos + j]);
-      yf[1][j] += MulIm(aec->xfBuf[0][xPos + j],
-                        aec->xfBuf[1][xPos + j],
-                        aec->wfBuf[0][pos + j],
-                        aec->wfBuf[1][pos + j]);
+      yf[0][j] += MulRe(xfBuf[0][xPos + j],
+                        xfBuf[1][xPos + j],
+                        wfBuf[0][pos + j],
+                        wfBuf[1][pos + j]);
+      yf[1][j] += MulIm(xfBuf[0][xPos + j],
+                        xfBuf[1][xPos + j],
+                        wfBuf[0][pos + j],
+                        wfBuf[1][pos + j]);
     }
   }
 }
 
 static void ScaleErrorSignalSSE2(int extended_filter_enabled,
                                  float normal_mu,
                                  float normal_error_threshold,
                                  float *xPow,
                                  float ef[2][PART_LEN1]) {
   const __m128 k1e_10f = _mm_set1_ps(1e-10f);
@@ skipping 643 matching lines @@
   }
 }
 
 void WebRtcAec_InitAec_SSE2(void) {
   WebRtcAec_FilterFar = FilterFarSSE2;
   WebRtcAec_ScaleErrorSignal = ScaleErrorSignalSSE2;
   WebRtcAec_FilterAdaptation = FilterAdaptationSSE2;
   WebRtcAec_OverdriveAndSuppress = OverdriveAndSuppressSSE2;
   WebRtcAec_SubbandCoherence = SubbandCoherenceSSE2;
 }