| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|
|
|
| |
Change-Id: I3c964a497e2fcc9106fe469e71a7d61f0d089653
Bug: 322228259
Test: presubmit, and verified performance with stack samples
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Do not use the cached SensorManager if the device association
of the underlying package has changed.
Since the SensorManager is cached per package, if the deviceId
changes, then the native SensorManager will always return the wrong
sensors. This is seen in the failing CTS.
The most straight-forward solution is to refresh the cache when the
device for that UID has changed. It would be better to have an API
that takes a deviceId but that's a longer term fix - b/306575095
This change is effectively flagged by the read-only flag
android.companion.virtual.flags.enable_native_vdm. When it is
disabled, the deviceId will always be 0 (i.e. default device).
Test: atest VirtualSensorNativeTest
Bug: 306575095
Fix: 317495754
Change-Id: If02beb9f3eceec19c104ae03fcb0a8fefb46f85f
|
| |
|
|
|
|
| |
Fix: 315295201
Test: manual with the VDM test app
Change-Id: I096a5af3f3476386e145c4aec23e25719dc78fec
|
| |\ |
|
| | |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| |
| | |
Add device id to SensorManager instances based on the devices
this UID is present on. SensorManager queries the native VDM
instance for the deviceId and the sensor policy of that device.
If the UID is present only on a single virtual device and that
device has custom sensors, then give it that device's sensors by
default. If the UID is not on any virtual devices or is on more
than one, then give it the default device's sensors.
Fix: 303535376
Test: manual with streamed app using sensor NDK to a virtual device
Change-Id: I275a010a7a70b34176deaf4c9dcd61188c83ac32
|
| |/
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
Add a new mode, HAL_BYPASS_REPLAY_DATA_INJECTION, which behaves similar
to Replay Data Injection with the difference that injected sensor data
is not injected into the HAL but simply passed back up to clients in
the platform.
Also, while I was in there, hook up the remaining bits and bobs to get
Replay Data Injection working and accessible from APIs in
SystemSensorManager in the platform.
Bug: 287257057
Test: manual
Change-Id: I9fc33a8bf5b67c02483089f849ba7ff0346d8097
(cherry picked from commit b0df44e2a6a40fb70a6e98a1e4e629945995a5f6)
|
| |
|
|
|
|
| |
Test: atest VtsHalSensorManagerV1_0TargetTest
Bug: 278013275
Change-Id: I091f57de9570b0ace3a8da76f16fe0e83f0aa624
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
When a direct channel is (un-)registered or configured, notify the
runtime sensor callback instead of the HAL and pass the shared memory
region to it. The SensorDirectConnection object is tied to a deviceId
and may only be configured for sensors of that device.
Only ashmem direct channel type is supported for runtime sensors.
Bug: 266042170
Test: atest CtsSensorTestCases
Change-Id: Ie1a628650bd94b6d81e95d3b9f8f25fcf445666c
|
| |
|
|
|
|
|
|
|
|
|
|
| |
In order to ensure that clients don't leak their sensor manager
instance that we currently store in a static map, they need to be able
to remove their instance. Otherwise, this instance is never removed from
the list and will hang around until our SensorManage instance is
destroyed.
Bug: 269014004
Test: Run ./libsensorserviceaidl_fuzzer
Change-Id: I52185f74ae8d28b379440235ca6f03c5089081f5
|
| |
|
|
|
|
|
|
|
|
| |
After catching an error in Sensor::unflatten, there are memory leaks
caught by the fuzzer in the same test case.
Test: libsensorserviceaidl_fuzzer with testcase from bug
Bug: 269014004
Merged-In: I509cceb41f56ca117d9475f6f6674244560fe582
Change-Id: I509cceb41f56ca117d9475f6f6674244560fe582
|
| |
|
|
|
|
|
|
|
|
|
|
| |
The virutal device sensors are runtime sensors, using handle range of [0x5F000000, 0x5FFFFFFFF].
As opposed to the HAL dynamic sensors, they do not rely on the existence of a SENSOR_TYPE_DYNAMIC_SENSOR_META. Instead of sensor events representing registration/unregistration of dynamic sensors, they are handled syncronously in the sensor service.
The virtual dynamic sensors are not exposed via the dynamic sensors API. They are provided to SensorManager through a separate JNI call.
Bug: 237278244
Test: atest cts/tests/sensor
Change-Id: I09e5b089d1ae3bed7a25a5454a31aba9cf594a05
|
| |
|
|
|
|
|
| |
Bug: 217890463
Test: Used uhid-sample to add a dynamic sensor and verified it can be
used with sensor_test.
Change-Id: Ida2ce53c3d94ed530ba4981d299fa4855c35e337
|
| |
|
|
|
|
|
|
| |
Also updates noteOp() to use the attribution tag.
Bug: 171922082
Test: Flash, verify through logs that attribution tags are propagated
Change-Id: I4601f6f8696e8cf1517d2401f8ddfffae4a29d04
|
| |
|
|
|
|
| |
Bug: 144139857
Test: N/A
Change-Id: Id809a8f65b31143759978702d18e6ac944d62ac3
|
| |
|
|
|
|
|
| |
C++ spec requires %p format specifier to correspond to a void pointer.
Test: presubmit
Change-Id: I98c72f94b6ea73607e84f32ddcc2149331e39978
|
| |
|
|
|
|
|
|
|
| |
Avoids potential invalid memory access if system server crashes during
initialization.
Bug: 132927376
Test: confirm sensors initialize
Change-Id: If7421c452b5893ab2567323d265503a1ce304482
|
| |
|
|
|
|
|
|
| |
Fixes -Wzero-as-null-pointer-constant warning.
Test: m
Bug: 68236239
Change-Id: I82e6ef0f074571a9537fa2eeedf56975363ca601
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
The headers that are related to system services are now hidden to
vendors. Note that this does not break anything because vendors have
already been disallowed to talk to system services via sepolicy. Their
use of binder is strictly limited within themselves (via
/dev/vndbinder).
The hidden headers are now guarded with __ANDROID_VNDK__ and including
them from vendors (or vendor variants of other VNDK libraries) will
trigger a build-time error.
Bug: 78113963
Test: m -j
Test: devices boots to the UI
Change-Id: I19be5ca024fc4081b7861d908d75758c1d956a83
|
| |
|
|
| |
Change-Id: Ifa5825b08d5b809865f9066c7a763202cebb987f
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
Set operation parameter to sensor HAL. Possible paramters include
local geomagnetic field, local gravity acceleration, device dock
state, whether device should run in high performance mode, and
if a magnetic field calibration guide is in progress.
Bug: 30958130
Test: Tested with marlin. Modified hal implementation can get local
geomagnetic field.
Change-Id: Idaedd9e616d01383ba244cc64cf3e319ce02191f
|
| |
|
|
|
|
|
|
|
| |
Wait for service before attempting to get a SensorManager object.
Bug: 36389997
Bug: 36404344
Test: system boots, sensors working with sensorlogger test
Change-Id: Ib5d41bc58edb8f6be56f803324a79a99a50da45e
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
Implement creation of sensor direct channel in sensor
service for sensor direct report feature.
Bug: 30985702
Test: cts-tradefed run cts --module CtsSensorTestCases
--test android.hardware.cts.SensorDirectReportTest
Test: cts-tradefed run cts --module CtsSensorTestCases
--test android.hardware.cts.SensorNativeTest
Change-Id: I52350811df538986cea7f5e1fb336fad53234c86
|
|
|
Test: compile & run
Bug: treble cleanup
Change-Id: I5b2c2a14f7d3b364d25695cf0901c5b08cfb0fc9
|