Refactor audio device type in audio flinger.
As audio device type must not be used as a bit mask, there is a
requirement to refactor the code in audio flinger. This CL includes:
1. Use DeviceTypeSet for recording output devices in Threads. The input
and output device types in Threads will only be updated when creating
an audio patch.
2. Use helper functions in audio flinger for checking whether an audio
device type is a certain category.
3. Use DeviceDescriptorBaseVector as vector of DeviceDescriptorBase.
DeviceDescriptorBaseVector will be used in RecordThread to record output
devices, which will be used in audio effect.
4. Do not use set parameters when communicating output devices to
RecordThread. Instead, adding a new event in Thread to update output
devices directly.
Bug: 135621476
Test: atest AudioTrackTest AudioRecordTest AudioPlaybackCaptureTest
Test: atest AudioHostTest AudioServiceHostTest RoutingTest
Test: atest AudioEffectTest BassBoostTest DynamicsProcessingTest
Test: atest AAudioTests
Test: audio smoke test, run OboeTester
Change-Id: I87f3e59fd01b0b100232d61a9abbc471c61110c4
diff --git a/media/libaudiofoundation/AudioContainers.cpp b/media/libaudiofoundation/AudioContainers.cpp
index adc5d40..1bfe3f9 100644
--- a/media/libaudiofoundation/AudioContainers.cpp
+++ b/media/libaudiofoundation/AudioContainers.cpp
@@ -50,6 +50,10 @@
}
bool deviceTypesToString(const DeviceTypeSet &deviceTypes, std::string &str) {
+ if (deviceTypes.empty()) {
+ str = "Empty device types";
+ return true;
+ }
bool ret = true;
for (auto it = deviceTypes.begin(); it != deviceTypes.end();) {
std::string deviceTypeStr;
@@ -83,4 +87,10 @@
return ret;
}
+std::string toString(const DeviceTypeSet& deviceTypes) {
+ std::string ret;
+ deviceTypesToString(deviceTypes, ret);
+ return ret;
+}
+
} // namespace android