Fix regression for SoundPool playback

Commit 19ddf0eb introduced a problem with applications (like SoundPool)
relying on an underrun condition to detect end of playback instead of
stopping the track when all data is written.
AudioFlinger would keep waiting for new data in case of partial buffer
filling and never reach the underrun condition.

Added a mechanism to wait no more than once if not enough frames are present
in the track buffer.

Issue 5585490.

Change-Id: I131e605ff6070831a01ddf734e68459e3bf2354b

1 files changed, 7 insertions, 2 deletions

diff --git a/services/audioflinger/AudioFlinger.cpp b/services/audioflinger/AudioFlinger.cpp
index 96e8eb9..ff262f1 100644
--- a/services/audioflinger/AudioFlinger.cpp
+++ b/services/audioflinger/AudioFlinger.cpp
@@ -2066,9 +2066,14 @@ uint32_t AudioFlinger::MixerThread::prepareTracks_l(const SortedVector< wp<Track
         // The first time a track is added we wait
         // for all its buffers to be filled before processing it
         mAudioMixer->setActiveTrack(track->name());
-        // make sure that we have enough frames to mix one full buffer
+        // make sure that we have enough frames to mix one full buffer.
+        // enforce this condition only once to enable draining the buffer in case the client
+        // app does not call stop() and relies on underrun to stop:
+        // hence the test on (track->mRetryCount >= kMaxTrackRetries) meaning the track was mixed
+        // during last round
         uint32_t minFrames = 1;
-        if (!track->isStopped() && !track->isPausing()) {
+        if (!track->isStopped() && !track->isPausing() &&
+                (track->mRetryCount >= kMaxTrackRetries)) {
             if (t->sampleRate() == (int)mSampleRate) {
                 minFrames = mFrameCount;
             } else {