During first boot after OTA, additional dexopting has to be done
during PM initialization. Timings for OTA are reported separately,
so we should ignore first boot to avoid skewing the metrics.
The following metrics were updated:
- boot_system_server_init - check added for consistency with other
metrics in SystemServer. The metric is actually unaffected by first boot,
because dexopt will start later
- boot_package_manager_init_ready
- boot_system_server_init
- framework_locked_boot_completed
- framework_boot_completed
Test: manual + UserControllerTest pass
Bug: 32807863
Change-Id: Iff13697b7d4f9ff8439e1e932d7e276f48cd5c37
This API is designed to provide both UID-level stats and overall
summary data for a given storage device, as identified by UUID.
The use of UID-level granularity might appear a bit clunky, but it
matches other usage statistics (such as network and battery), and it
allows us to implement it using an extremely fast quota kernel
feature.
A future CL will wire up the implementation to installd.
Test: builds, boots
Bug: 32206268
Change-Id: I7b51877682d0370c2402c19346f57809f0e7ac53
System services can use it during the boot to submit tasks that can be run in
parallel with the main thread.
Switched PersistentDataBlockService and FingerprintService from FgThread to
the new thread pool.
UiModeManagerService: update initial configurations on init thread. They run
while holding the mLock so no extra synchronization barriers are needed at a
later stage.
Test: manual - device boots without errors
Test: ParallelPackageParserTest passes
Change-Id: I548f34b0a18f61924e09a39afb12e085cde35442
This service will only be enabled if the platform was built with
EMMA_INSTRUMENT=true
Test: Manual
Bug: 31077138
Change-Id: I4ba98b6568d31ded1b66da996b3c2e5a2ee07c75
There are a few major pieces here:
incidentd
---------
This daemon (started by init) runs and accepts incoming requests to take
incident reports. When prompted, it calls into various system services
and fills in an IncidentProto data structure, and then writes the report
into dropbox.
The next steps for incidentd:
- Security review of SELinux policies. These will be a subset of
the dumpstate permissions. Until this is done, incidentd is
not started at boot time.
incident
--------
This shell command calls into incidentd, and can initiate an incident
report and either capture the output or leave for dropbox.
incident_report
---------------
This host side tool can call adb shell with the correct parameters
and also format the incident report as text. This formatting code
was left of the device on purpose. Right now it's pretty small, but
as the number of fields increases, the metadata and code to do the
formatting will start to grow.
The incident_report command also contains a workaround to let it
work before incidentd is turned on by default. Right now, it is
implemented to call adb shell dumpsys <service> --proto directly,
whereas in the future it will go through the full incidentd flow.
incident_section_gen
--------------------
A build-time tool that generates a stripped down set of information
about the fields that are available.
libincident
-----------
This library contains the code to connect to incidentd, and the
meta proto definitions that are used by the framework protos.
The basics are here now, but they are not fully fleshed out yet.
The privacy.proto file contains annotations that can go in the
proto file that we will later use to filter which fields are
uploaded, and which are used by local sources. For example, a
device in a test lab is safe to upload much much more information
than a real user. These will share the same mechanism, but the
user's output will be filtered according to these annotations.
frameworks/core/proto
---------------------
These .proto files contain the definitions of the system's
output. There is one master android.os.IncidentProto file that
is the top level of an incident report, but some other services
(notification, fingerprint, batterystats, etc) will have others
that are used directly by the logging mechanism.
Other files which are shared by several of the services also go
here, such as ComponentName, Locale, Configuration, etc. There
will be many more.
There is also a first iplementation of a dump method handling
--proto in the fingerprint service.
IncidentManager
---------------
The java API to trigger an incident report.
Test: Not written yet
Change-Id: I59568b115ac7fcf73af70c946c95752bf33ae67f
Split systemReady() into 2 callbacks:
1) networkScoreAndNetworkManagementServiceReady - called when required
services are ready. NetworkPolicyManagerService starts its initialization
on the worker thread
2) systemReady - wait for phase (1) to finish so that the service is fully
initialized
NetworkPolicyManagerService is now prepared in parallel with
ConnectivityService and NetworkStatsService
Test: device boots without errors
Test: HostsideRestrictBackgroundNetworkTests and NetworkPolicyManagerServiceTest pass
Bug: 32315581
Change-Id: Ic9755ed096900247fa0bbb43f396d8bcb97ae6db
Other services do not depend on PersistentDatabalockService in the early
stages of system server initialization.
Initialization is now done on the FgThread. The service also ensures that
init is complete before PHASE_ACTIVITY_MANAGER_READY.
Change-Id: I3b9a1994e37c1e995ed04257f271c425d48c2e7c
Test: Manual, device boots, service is initialized and published
Bug: 33199244
Extracted duration measurement functionality into BootTimingTraceLog.
It is now shared between system_server and zygote.
Log the following metrics to tron:
- boot_zygote_init - Time in milliseconds to boot into Zygote init stage.
- boot_android_init - Time in milliseconds to boot into Android init stage.
Test: manual - device boots, metrics are logged to system/event log
Bug: 32780225
Bug: 31115337
Change-Id: I600ac7fc83d35fa226ac92c37cc4b19192b25f59
This CL provides the initial, skeleton implementation of the Auto-Fill
Framework classes:
- Defines the system service and app-based
AIDL (IAutoFillManagerService.aidl and IAutoFillService.aidl respectively).
- Defines the 'adb shell cmd' interface.
- Defines the permission required to access the service.
- Registers the service on SystemServer.
- Adds the code to bind the app-specified service to system_server.
- Defines the service class (AutoFillService) required by providers.
- Implements the initial startSession() method.
This is still a very early, "work-in-progress" change:
- It has many TODOs.
- It does not have unit or CTS tests yet.
- It does not provide a callback method to auto-fill the fields.
- In fact, it has a lot of TODOs.
Despite these adversities, it can be tested by following the steps
below:
1.Create an app with a service extending AutoFillService
2.Implement the onNewSession() method
3.In the manifest:
- Listen to android.service.autofill.AutoFillService intents.
- Require the android.permission.BIND_AUTO_FILL permission.
4.Explicitly set the app as an autofill-service by running:
adb shell settings put secure auto_fill_service MY_APP/.MY_SERVICE
5.Start a session against the top activity:
adb shell cmd autofill start session
BUG: 31001899
Test: manually built and ran it
Change-Id: I00f4822159b31ddddba8f513e57c4474bc74eb89
To protect system stability, any Binder calls leaving the
system_server must carefully be performed using FLAG_ONEWAY (or
the 'oneway' verb in AIDL) which prevents the call from blocking
indefinitely on the remote process.
In this CL, the system_server uses the new Binder.setWarnOnBlocking()
method to enable detection by default for all remote Binder
interfaces. It can also use Binder.allowBlocking() to allow
blocking calls on certain remote interfaces that have been
determined to be safe.
This CL adds the 'oneway' verb to several interfaces and methods
where it should have been added, and marks a handful of system
ContentProviders as being safe to call into. Also, we assume that
any services obtained from ServiceManager are part of the core
OS, and are okay to make blocking calls to.
Test: builds, boots, runs with minimal logs triggered
Bug: 32715088
Change-Id: Ide476e120cb40436a94b7faf7615c943d691f4c0
~Rename only (and any reformatting needed to pass lint) - no
functional changes!
Remove android.net.wifi.nan.STATE_CHANGED from manifest:
redundant/remnant of an older configuration.
(cherry-pick of commit a61b9fb569153917a650f1d48efa20ba8846a9f3)
(cherry-pick of commit b061f21e7e59a99834e163e2baa1c82229e419a6)
Bug: 32263750
Test: All unit tests and integration (sl4a) tests pass.
Change-Id: Ie4ff675fa61041e8fcf6a9bf9900ea835d0a7614
Now display-specific settings, such as dimensions and orientation,
are stored in display override config. For default display it is
mirroring the global config. Each time when global config is updated,
override of the default display should be updated too and vice versa.
Test: Existing and manual tests still pass.
Change-Id: Ic6c2190092d328820f314a05bed43c875db18170
NOTE: Linear blending is currently disabled in this CL as the
feature is still a work in progress
Android currently performs all blending (any kind of linear math
on colors really) on gamma-encoded colors. Since Android assumes
that the default color space is sRGB, all bitmaps and colors
are encoded with the sRGB Opto-Electronic Conversion Function
(OECF, which can be approximated with a power function). Since
the power curve is not linear, our linear math is incorrect.
The result is that we generate colors that tend to be too dark;
this affects blending but also anti-aliasing, gradients, blurs,
etc.
The solution is to convert gamma-encoded colors back to linear
space before doing any math on them, using the sRGB Electo-Optical
Conversion Function (EOCF). This is achieved in different
ways in different parts of the pipeline:
- Using hardware conversions when sampling from OpenGL textures
or writing into OpenGL frame buffers
- Using software conversion functions, to translate app-supplied
colors to and from sRGB
- Using Skia's color spaces
Any type of processing on colors must roughly ollow these steps:
[sRGB input]->EOCF->[linear data]->[processing]->OECF->[sRGB output]
For the sRGB color space, the conversion functions are defined as
follows:
OECF(linear) :=
linear <= 0.0031308 ? linear * 12.92 : (pow(linear, 1/2.4) * 1.055) - 0.055
EOCF(srgb) :=
srgb <= 0.04045 ? srgb / 12.92 : pow((srgb + 0.055) / 1.055, 2.4)
The EOCF is simply the reciprocal of the OECF.
While it is highly recommended to use the exact sRGB conversion
functions everywhere possible, it is sometimes useful or beneficial
to rely on approximations:
- pow(x,2.2) and pow(x,1/2.2)
- x^2 and sqrt(x)
The latter is particularly useful in fragment shaders (for instance
to apply dithering in sRGB space), especially if the sqrt() can be
replaced with an inversesqrt().
Here is a fairly exhaustive list of modifications implemented
in this CL:
- Set TARGET_ENABLE_LINEAR_BLENDING := false in BoardConfig.mk
to disable linear blending. This is only for GLES 2.0 GPUs
with no hardware sRGB support. This flag is currently assumed
to be false (see note above)
- sRGB writes are disabled when entering a functor (WebView).
This will need to be fixed at some point
- Skia bitmaps are created with the sRGB color space
- Bitmaps using a 565 config are expanded to 888
- Linear blending is disabled when entering a functor
- External textures are not properly sampled (see below)
- Gradients are interpolated in linear space
- Texture-based dithering was replaced with analytical dithering
- Dithering is done in the quantization color space, which is
why we must do EOCF(OECF(color)+dither)
- Text is now gamma corrected differently depending on the luminance
of the source pixel. The asumption is that a bright pixel will be
blended on a dark background and the other way around. The source
alpha is gamma corrected to thicken dark on bright and thin
bright on dark to match the intended design of fonts. This also
matches the behavior of popular design/drawing applications
- Removed the asset atlas. It did not contain anything useful and
could not be sampled in sRGB without a yet-to-be-defined GL
extension
- The last column of color matrices is converted to linear space
because its value are added to linear colors
Missing features:
- Resource qualifier?
- Regeneration of goldeng images for automated tests
- Handle alpha8/grey8 properly
- Disable sRGB write for layers with external textures
Test: Manual testing while work in progress
Bug: 29940137
Change-Id: I6a07b15ab49b554377cd33a36b6d9971a15e9a0b
- Update RetailDemoModeService to not do anything outside demo mode.
- Add am command get-started-user-state which is needed for cts tests.
- Update unit tests for RetailDemoModeService.
Bug: 31342350
Test: adb shell am instrument -e class com.android.server.retaildemo.RetailDemoModeServiceTest -w com.android.frameworks.servicestests/android.support.test.runner.AndroidJUnitRunner
Change-Id: Idf50512facd27d47d7111e75cbc2f7b260f49740
in System Server.
Change code to use traceEnd() instead
Trace.traceEnd(Trace.TRACE_TAG_SYSTEM_SERVER); to generate accurate
measurements.
Test: 1-line change. 0-risk change.
bug: 31654120
Change-Id: I24d1468ccf934dbe912822996174eb780a46994c
Added a service that listens whether emergency affordances
are necessary.
If the they are needed, it adds an option to the
global actions dialog that directly launches the
emergency call and also adds a long-press listener
to the keyguard emergency button.
Test: adb shell settings put global force_emergency_affordance 1 && adb shell settings put global emergency_affordance_number 111112
Fixes: 30404490
Change-Id: Ib96a15da2ef4b568a8d77140ebca6aa6f20f5ddb
This patch defines a new metrics service for IpConnectivity events
defined in android.net.metrics, separate from currently existing
MetricsLoggerService.
Similarly to MetricsLoggerService, the new service has an event buffer.
It also implements a dumpsys interface that can be used to flush events
and output a serialized proto.
Bug: 31254800
Change-Id: I0c3faeb4008b283f85d9ba9460371fa68956ea3b
This service proxy (together with the native server) are not needed on
Android Wear, as we don't have any watches with cameras.
Bug: 28560707
Change-Id: Ie4a830a3ba48c90d3e968fc5cdf57ccafcc1f5d8
Build serial is non-user resettable freely available deivice
identifier. It can be used by ad-netowrks to track the user
across apps which violates the user's privacy.
This change deprecates Build.SERIAL and adds a new Build.getSerial()
API which requires holding the read_phone_state permission.
The Build.SERIAL value is set to "undefined" for apps targeting
high enough SDK and for legacy app the value is still available.
bug:31402365
Change-Id: Iddd13430b2bd1d9ab4966e31038ecabdbdcec06d
