For each camera session, collect capture latency histogram for all streams.
Test: Camera CTS, and observe statsd output
Bug: 154159000
Change-Id: I7741371ca8540cdf9c9a47246c0ef0f0f82df32a
Enables SysUI to draw enroll/auth-specific affordances
If the UI diverges, we can consider enroll or auth specific
UdfpsView subclasses(?) but currently enrollment just shows
an extra circle.
Fixes: 177275693
Test: atest com.android.systemui.biometrics
Change-Id: I5b2aaf00c55279a81348ec7a5dc387b6e62751c0
The only public API (that I know!) that let touches pass-through is
FLAG_NOT_TOUCHABLE, so I've put the rules that relate to app-developers
there.
Please let me know if there are any other public APIs that makes a
window let touches pass-through. The other one I know of is
ViewTreeObserver.addOnComputeInternalInsetsListener() that let's you
specify the touchableRegion, but that's not public.
(The feature works regardless of the API used to make the window
pass-through)
The docs are a shorter version of the "Rules" section of
go/untrusted-touches.
Bug: 158002302
Fix: 173022596
Test: Build
Change-Id: Ic7ba68f03587112db2cc93128ac41cb09d443127
Test: Manually add a listener in another serviec and change state with
`adb shell cmd sensor_privacy enable 0 microphone`.
Bug: 162549680
Change-Id: I4a77bc2bba4405b2ba962e767005ca443a2d732f
For the spaceship mode this is purely an internal refactorization since
all the apis remain the same and the state is kept in sync across users.
For the individual controls there's new calls to set the state of the
toggle per user or per profile group.
Test: Manually change state with `adb shell cmd sensor_privacy enable 0
microphone`
Manually upgrade from old persistence to new persistence form
Manually check that the persisted state remains when rebooted.
Bug: 162549680
Change-Id: I0a5eac7a5c3beb272a456790d999a0282d77bf09
Test: Verified regular fingerprint enrollment correctly logged
enrollment.
Test: Verified that the find fingerprint sensor activity no longer
falsely reports a failed enrollment.
Bug: 175316123
Fingerprint enrollment shows a find sensor screen which makes a call to
FingerprintManager.enroll(), the purpose is to get the user to locate
the sensor. The consequence was that logging built into the framework
was incorrectly logging a failed enrollment after cancellation.
Change-Id: I2ec27120943e1ce3d70f5ca2f3221a4b3b571c39
Bug: 174742989
Test: GCA (basic validity)
Test: Without this CL, capture many images with GCA, profile with
simpleperf, stack traces with java.lang.Float.valueOf in
mm_fault_count kprobes : 21% (most of them coming from
MarshalQueryableArray.unmarshal)
Test: Without this CL, capture many images with GCA, profile with
simpleperf, stack traces with java.lang.Float.valueOf in
mm_fault_count kprobes : 3.92%
Test: camera CTS, no new failures.
Change-Id: If359c828a231df41ba7b586bdda36255656a5a76
Signed-off-by: Jayant Chowdhary <jchowdhary@google.com>
ag/13253804 fixes test we broke in attempt #1 (ag/13228932).
-- Original CL description (update "Test:" line):
After getting enough feedback from dogfooders, the issues left open that
we know of (that will still be addressed ofc) either don't hurt dogfood
experience (eg. b/172627742 touch is blocked only during AGSA's -1
screen <> launcher side animation) or are 3p apps that will have to
change their code at some point. So, turning the feature on in its final
enforcing mode.
Also removing toast since at this point this is more noise than signal.
We're seeing many duplicate bugs (many of which are WAI) that the
dogfooder would likely not notice if it wasn't for the toast. If there
is a real issue the user will still notice by the unresponsiveness and
given a bugreport we will be able to check via logcat.
However, I have not removed the code completely and left that as part
of b/169067926 since we may decide to turn it back on before or on
DP/Beta.
Bug: 158002302
Test: atest WindowUntrustedTouchTest WindowInputTests inputflinger_tests
inputflinger_benchmarks libinput_tests libgui_test
Test: Verify untrusted touches are blocked via logcat but no toast.
Change-Id: I709711e60f5f4f563e018ee079ef89418e161d14
Preallocate sensor event for listener to avoid allocating for each
sensor event notification and garbage collection.
Bug: 161634265
Test: atest InputDeviceSensorManagerTest
Change-Id: I9ef32d57116d0ee4fe7e7b331cb0b2b35d43806f
Add support of sensor manager in InputDevice and allow user to obtain
SensorManager object to query supported sensor types and register sensor
event listeners.
Bug: 161634265
Test: atest InputDeviceSensorManagerTest
Change-Id: Ia7a04a0fcf7e026c49a4fb1a9009b132ff84cf7c
After getting enough feedback from dogfooders, the issues left open that
we know of (that will still be addressed ofc) either don't hurt dogfood
experience (eg. b/172627742 touch is blocked only during AGSA's -1
screen <> launcher side animation) or are 3p apps that will have to
change their code at some point. So, turning the feature on in its final
enforcing mode.
Also removing toast since at this point this is more noise than signal.
We're seeing many duplicate bugs (many of which are WAI) that the
dogfooder would likely not notice if it wasn't for the toast. If there
is a real issue the user will still notice by the unresponsiveness and
given a bugreport we will be able to check via logcat.
However, I have not removed the code completely and left that as part
of b/169067926 since we may decide to turn it back on before or on
DP/Beta.
Bug: 158002302
Test: Verify untrusted touches are blocked via logcat but no toast.
Change-Id: Ic66ce0dc273fae3f1d974dd46d383f046f78e485
This CL relays requests to enable or disable Pointer Capture in
InputManagerService to InputDispatcher.
Windows now request Pointer Capture using their InputChannel token
instead of their IWindow tokens.
Bug: 141749603
Test: manual: flash crosshatch, test pointer capture.
Change-Id: Ic6fc01eae0068dfbba496051d988a71bcabcd0b3
SystemApi complains:
error: Constant field names must be named with only upper case characters:
`android.hardware.hdmi.HdmiControlManager#HDMI_CEC_VERSION_1_4_b`,
should be `HDMI__CEC__VERSION_1_4_B`
Bug: 174731501
Bug: 172637440
Test: atest com.google.android.gts.hdmicec.HdmiControlManagerHostTest
Change-Id: Ia1bfdb44e704c99fa574663f658a2173f6b48d21
