MTK-voice最大音量通话-免提切到耳机模式后音量变小问题总结

问题现象

MTK平台某项目，在voice最大音量通话时，从免提切换到耳机模式后，通话音量变小，此时通过音量键减小一格音量后，通话音量又变大；

根因分析

问题原因：
通话模式从免提切换到耳机模式后，由于MTK上层会设置一个AUDIO_DEVICE_OUT_TELEPHONY_TX 设备的音量，而该设备无对应gaindevice，导致实际在hal层设置对应音量时设置成了speaker的音量，同等级音量下，speaker的音量是比耳机模式小的；

相关代码分析：

在切换设备时，从framework侧会重新设置新通话模式的音量
AudioPolicyManagerCustomImpl::gainTable_applyAnalogGainFromCheckAndSetVolume
调用-->
AudioPolicyManagerCustomImpl::checkAndSetGainTableAnalogGain
调用-->
AudioPolicyManagerCustomImpl::setVolumeParameterToHAL
调用-->
mAudioPolicyManager->mpClientInterface->setParameters
调用-->
AudioALSAStreamManager::setAnalogVolume
调用-->
status_t AudioMTKGainController::setVoiceVolume(int index, int devices, audio_mode_t mode)
调用-->
	GAIN_DEVICE AudioMTKGainController::getGainDevice(audio_devices_t devices)
在这里会根据setparameters传下来的devices 类型选择GAIN_DEVICE，由于AUDIO_DEVICE_OUT_TELEPHONY_TX 没有对应GAIN_DEVICE, 所以使用默认值speaker，从而将speaker对应的等级音量设置到modem dl下行中去，从而将先前一步设置的耳机音量给覆盖掉，导致问题现象出现；
调用-->
AudioMTKGainController::ApplyMdDlGain
调用-->
SpeechDriverNormal::SetDownlinkGain

解决方案

在AudioPolicyManagerCustomImpl.cpp做如下修改，避免通话模式设置AUDIO_DEVICE_OUT_TELEPHONY_TX 设备音量

status_t AudioPolicyManagerCustomImpl::checkAndSetGainTableAnalogGain(audio_stream_type_t stream, int index, const sp<AudioOutputDescriptor>& outputDesc,DeviceTypeSet device,
                                                           int delayMs, bool force)
{
...
        if ((activeStream == AUDIO_STREAM_VOICE_CALL || activeStream == AUDIO_STREAM_BLUETOOTH_SCO) &&
            outputDesc != mAudioPolicyManager->mPrimaryOutput &&
            !(((sp<SwAudioOutputDescriptor>&)outputDesc)->mFlags & AUDIO_OUTPUT_FLAG_VOIP_RX)) {
            // in voice, set to primary only once, skip others
            bool PrimaryVolSetIsSkipped = false;    //  ALPS05140437  We need to check the primary output due to it may not  necessary to do call routing and  update to the new device
            DeviceTypeSet curPrimaryDevices = mAudioPolicyManager->mPrimaryOutput->devices().types();
            DeviceTypeSet curDevices = outputDesc->devices().types();
            PrimaryVolSetIsSkipped = (curDevices.find(
                    Volume::getDeviceForVolume(curPrimaryDevices)) == curDevices.end());
            if (PrimaryVolSetIsSkipped) {
                ALOGD("%s skip. Cur device [%s], primary device [%s]", __FUNCTION__, dumpDeviceTypes(curDevices).c_str(), dumpDeviceTypes(curPrimaryDevices).c_str());
            } else {
                return NO_ERROR;
            }
+
+            if (curDevices.count(AUDIO_DEVICE_OUT_TELEPHONY_TX) != 0) {
+                ALOGD("AUDIO_DEVICE_OUT_TELEPHONY_TX skip set voice volume");
+                return NO_ERROR;
            }
        }
...
}