[AWARE] Handle HAL failure correctly
HAL calls may fail asynchronously or synchronously. The frameworks finds
out about async failures through notification callbacks - already handled.
The synchronous failures happen immediately on HAL API calls. These are
rarer and have not been handled until now. Add code to handle the sync
failures - by calling appropriate failure callbacks. Add unit tests.
Bug: 29061728
Test: unit tests + integration (sl4a) tests pass.
Change-Id: I7ad35c989aa146f74eaad4eaa781d2e628586846
4 files changed