We recently changed the behaviour in InputReader to adjust the video
frame data based on the current orientation.
Update the HAL documentation to reflect that.
Also, rearrange the order of width and height, because it is more
natural to have the format of num of rows x num of columns. So
currently, it is reversed, and not as natural.
More information about this is in ag/6414079.
Bug: 123241238
Test: none
Change-Id: I69f184794bd7684e5c0981a6db66542ab914eff2
The VTS tests are needed to test the InputClassifier HAL.
Add some basic tests here.
Test: atest VtsHalInputClassifierV1_0TargetTest
Bug: 117935272
Change-Id: I9a33729ad8dddbacab0500ee2953036c07e566dd
Make the reference implementation for InputClassifier HAL do nothing.
This is to allow cuttlefish to install this implementation.
It is safe for any device to install this implementation, and is
equivalent to not having the HAL running at all.
Test: boot up cuttlefish, and interact with the virtual device to ensure
that input works okay. Then 'ps -A | grep -i input' and ensure that the
HAL is running on the device.
Bug: 122600147
Change-Id: I841811bc0da3d78db8d7c1589cf1c59819d901a1
The reset will be used to prevent an inconsistent stream of
events to be sent to the HAL.
Bug: 111480215
Test: make only
Change-Id: I40c7d671f094065e3fcaff0d83e826c580dcae7a
The types added for input in InputClassifier HAL are not specific to
that HAL. These are common input definitions.
To allow for future reuse in this and other HALs, move these definitions
into a separate, type-only HAL android::hardware::input::common.
This will be similar to such existing HALs as:
hardware/interfaces/graphics/common
hardware/interfaces/camera/common
hardware/interfaces/audio/common
Test: make only
Bug: 111480215
Change-Id: I16d76d6bdb48b24487b232fda45c6146e1003fe9
The new HAL will be responsible for processing touch video frames that
are received from InputClassifier stage.
After processing, the HAL will generate a classification of the current
gesture. This classification will later be provided to the app via
the MotionEvent.getClassification() public API.
The eventual goal is to classify each event stream as a force touch
or not a force touch.
An example implementation of the HAL is presented here. This code was
used for local testing on blueline, and should not be actually installed
or used in any way other than for reference.
Bug: 62940136
Test: observed "detected deep press" logs when pressing with large
finger area on the touchscreen
Change-Id: Id6ac4337435e4ac07877da11ca184b6dd4d64780
