The fuzzer plugin for libaudioflinger is designed based on the understanding of the library and tries to achieve the following:
The configuration parameters are not hardcoded, but instead selected based on incoming data. This ensures more code paths are reached by the fuzzer. The fuzzer covers libaudioflinger APIs as called from libaudioclient through IPC.
libaudioflinger supports the following parameters:
uniqueId
)mode
)sessionId
)encapsulationMode
)portRole
)portType
)streamType
)format
)channelMask
)usage
)contentType
)inputSource
)inputFlags
)outputFlags
)gainMode
)device
)Parameter | Valid Values | Configured Value |
---|---|---|
uniqueId | 0. AUDIO_UNIQUE_ID_USE_UNSPECIFIED 1. AUDIO_UNIQUE_ID_USE_SESSION 2. AUDIO_UNIQUE_ID_USE_MODULE 3. AUDIO_UNIQUE_ID_USE_EFFECT 4. AUDIO_UNIQUE_ID_USE_PATCH 5. AUDIO_UNIQUE_ID_USE_OUTPUT 6. AUDIO_UNIQUE_ID_USE_INPUT 7. AUDIO_UNIQUE_ID_USE_CLIENT 8. AUDIO_UNIQUE_ID_USE_MAX | Value obtained from FuzzedDataProvider |
mode | 0.AUDIO_MODE_INVALID 1. AUDIO_MODE_CURRENT 2. AUDIO_MODE_NORMAL 3. AUDIO_MODE_RINGTONE 4. AUDIO_MODE_IN_CALL 5. AUDIO_MODE_IN_COMMUNICATION 6. AUDIO_MODE_CALL_SCREEN | Value obtained from FuzzedDataProvider |
sessionId | 0. AUDIO_SESSION_NONE 1. AUDIO_SESSION_OUTPUT_STAGE 2. AUDIO_SESSION_DEVICE | Value obtained from FuzzedDataProvider |
encapsulationMode | 0. AUDIO_ENCAPSULATION_MODE_NONE 1. AUDIO_ENCAPSULATION_MODE_ELEMENTARY_STREAM 2. AUDIO_ENCAPSULATION_MODE_HANDLE | Value obtained from FuzzedDataProvider |
portRole | 0. AUDIO_PORT_ROLE_NONE 1. AUDIO_PORT_ROLE_SOURCE 2. AUDIO_PORT_ROLE_SINK | Value obtained from FuzzedDataProvider |
portType | 0. AUDIO_PORT_TYPE_NONE 1. AUDIO_PORT_TYPE_DEVICE 2. AUDIO_PORT_TYPE_MIX 3. AUDIO_PORT_TYPE_SESSION | Value obtained from FuzzedDataProvider |
streamType | 15 values of type audio_stream_type_t | Value chosen from valid values by obtaining index from FuzzedDataProvider |
format | 77 values of type audio_format_t | Value chosen from valid values by obtaining index from FuzzedDataProvider |
channelMask | 83 values of type audio_channel_mask_t | Value chosen from valid values by obtaining index from FuzzedDataProvider |
usage | 22 values of type audio_usage_t | Value chosen from valid values by obtaining index from FuzzedDataProvider |
contentType | 5 values of type audio_content_type_t | Value chosen from valid values by obtaining index from FuzzedDataProvider |
inputSource | 14 values of type audio_source_t | Value chosen from valid values by obtaining index from FuzzedDataProvider |
inputFlags | 9 values of type audio_input_flags_t | Value chosen from valid values by obtaining index from FuzzedDataProvider |
outputFlags | 16 values of type audio_output_flags_t | Value chosen from valid values by obtaining index from FuzzedDataProvider |
gainMode | 3 values of type audio_gain_mode_t | Value chosen from valid values by obtaining index from FuzzedDataProvider |
device | 66 values of type audio_devices_t | Value chosen from valid values by obtaining index from FuzzedDataProvider |
This also ensures that the plugin is always deterministic for any given input.
The plugin tolerates any kind of input (empty, huge, malformed, etc) and doesn't exit()
on any input and thereby increasing the chance of identifying vulnerabilities.
This describes steps to build audioflinger_fuzzer binary.
Build the fuzzer
$ mm -j$(nproc) audioflinger_fuzzer
Create a directory CORPUS_DIR and copy some files to that folder Push this directory to device.
To run on device
$ adb sync data $ adb shell /data/fuzz/arm64/audioflinger_fuzzer/audioflinger_fuzzer CORPUS_DIR