summaryrefslogtreecommitdiffstats
path: root/libaudio2
diff options
context:
space:
mode:
authorEric Laurent <elaurent@google.com>2010-11-03 09:57:24 -0700
committerEric Laurent <elaurent@google.com>2010-11-03 09:57:24 -0700
commitb01135320204346bb7e6ecb87d8f9702cabd90a6 (patch)
tree914021dd9e971bec1ea5010b5b684b1efb5f1dfb /libaudio2
parent9389f2be393210b1c49d6982c2c96d531f9faa81 (diff)
downloaddevice_samsung_crespo-b01135320204346bb7e6ecb87d8f9702cabd90a6.zip
device_samsung_crespo-b01135320204346bb7e6ecb87d8f9702cabd90a6.tar.gz
device_samsung_crespo-b01135320204346bb7e6ecb87d8f9702cabd90a6.tar.bz2
Fix issue 3161227.
When a call is answered, the audio path is configured two times: - a first time by audio HAL when changing audio mode from RINGTONE to IN_CALL, where current device is applied to voice call path as a default configuration - a second time when audio policy manager defines which device should be used. This happens several hundred ms later so that the ringtone can be stopped before audio is routed to the earpiece. The uplink audio disruption apparently occurs because we switch the incall path first to speaker(as this is the current path for ringtone) and then to ear piece. The fix consists in removing the first audio path configuration by audio HAL when entering in call which is in fact redundant with the configuration done by audio policy manager. Change-Id: Ib99674dd3dcbba33e3d766e311ac54e919e64660
Diffstat (limited to 'libaudio2')
-rw-r--r--libaudio2/AudioHardware.cpp3
1 files changed, 0 insertions, 3 deletions
diff --git a/libaudio2/AudioHardware.cpp b/libaudio2/AudioHardware.cpp
index 7cff057..53e9c65 100644
--- a/libaudio2/AudioHardware.cpp
+++ b/libaudio2/AudioHardware.cpp
@@ -332,9 +332,6 @@ status_t AudioHardware::setMode(int mode)
openPcmOut_l();
openMixer_l();
setVoiceRecognition_l(false);
- if (mOutput != 0) {
- setIncallPath_l(mOutput->device());
- }
mInCallAudioMode = true;
}
if (mMode == AudioSystem::MODE_NORMAL && mInCallAudioMode) {