In practice the NetworkStack was not started after
ActivityManager#systemReady, since ActivityManager was only scheduling
the start (mPendingStarts).
Requesting the start too early causes issues for the PackageManager to
install updates/rollbacks.
Test: flashed AOSP and internal builds, WiFi working
Bug: 128331883
Change-Id: I93315ad31925bd436e21d1eabb21d5967e6aae60
NSS needed it for getting VpnInfo[], NetworkState[] and
activeLinkProperties which it used to query via ConnectivityManager.
For VpnInfo[], this was racy as NSS may ignore intermediate changes to a
VPN's underlying networks. See http://b/123961098 for more context.
It may also lead to deadlocks b/w ConnectivityService and
NetworkStatsService. See http://b/126245192 for more info.
This change will ensure that NSS is never contending on any of
ConnectivityService locks.
Bug: 123961098
Bug: 126245192
Bug: 120145746
Test: atest FrameworksNetTests
Change-Id: Id1da446b54d95ee68ed14079107b1a10318bcf8b
Merged-In: I57e117bb4e9efe491b19d6b5a479f2d58d1c58e6
NetworkStack is only used in services.net or clients of services.net. It
cannot stay in framework.jar because it needs to depend on AIDL
interfaces, which would conflict with app implementations if they were
in framework.jar.
Test: atest FrameworksNetTests NetworkStackTests
Bug: 124033493
Change-Id: Ib1d08a3669983640119d008db7e2990fa798724f
Merged-In: I501b125a388c1100c2182bde4670944c2f0d7a02
Remove the stubbed TimeZoneDetectorService. Priorities
dictated that it not be worked on so it has never been
used.
This reverts commit 7fb88c3928.
Bug: 112827764
Bug: 78217059
Test: build / boot
Change-Id: I39ed2fc7f8de8d4b2c0d883cd264cc126579af27
Merged-In: I39ed2fc7f8de8d4b2c0d883cd264cc126579af27
(cherry picked from commit f4894d391b)
Define the DynamicAndroid with AIDL.
Add a java implementation.
Start a service instance in the system server.
Add a permission test.
Bug: 122015653
Test: Build & Test on a pixel phone with following command \
./frameworks/base/services/tests/runtests.py -e class com.android.server.DynamicAndroidTest
Merged-In: I2e54b6b71fac4a4c5a9c9c25ce6bdac74cddcfb7
Change-Id: I2e54b6b71fac4a4c5a9c9c25ce6bdac74cddcfb7
This commit made the following changes to make the code conform to the
Frameworks style guide:
* Re-named variables
* Re-flowed code
* Organized includes
Topic: zygote-prefork
Test: make & flash & launch apps
Bug: 68253328
Change-Id: I9274b32f1f606f29f6eb3a1e5068ca18f607afe7
Merged-In: I9274b32f1f606f29f6eb3a1e5068ca18f607afe7
One use case for changing settings during initialization is enabling airplane mode in safe mode bootup.
Test: manual
Change-Id: I846f09faca22b35f502a18348b70800d1edabb60
Make it possible to boot in airplane mode when it was not turned on
before shutdown/reboot, either for situations where this is required by
regulation or to work around problems in normal boot. When safe mode is
engaged (Vol-Down), disabled in two stages:
1. Set Settings.Global.AIRPLANE_MODE_ON as early as possible for
subsystems to react to it during their initialization.
2. Trigger the public ConnectiviyService API to enable airplane mode. This
has to be done during later stages in initialization because it requires
sending broadcasts, so some RF leakage may still happen before airplane
mode is enabled (time window of a few seconds).
Bluetooth is not currently disabled and needs to be fixed by another
change. WiFi and NFC are turned off alongside cellular modems.
Test: manual
Bug: 120145930
Change-Id: I1d61990a9d6f7cd046d28b3e49ecf402dd485063
(cherry picked from commit afc9bc80941ea015cb591a0032e7553b5805fcc6)
The API is mostly implemented; except for hooking up the listener
and handling an already running bugreport.
BugreportManager is the handle to the new API exposed to apps.
Generating bugreports requires root privileges. To limit the footprint
of the root access, the actual bugreport generation in Dumpstate binary,
is accessed as a oneshot service and dies after it finishes
running.
System server accesses Dumpstate via a binder interface since it does
not have root privileges.
Starting a oneshot service is done via setting a system property, which
needs to be done from system server. BugreportManagerService is the
new system server service that does this. BugreportManager calls into
BugreportManagerService via a binder interface, since the former is in
the app's process.
Both app to system server as well as system server to native service
calls are via implementations of IDumpstate binder interface.
Bug: 111441001
Test: builds. Flashed & verified it boots.
Test: wrote a test client (not included) and verified invoking
startBugreport works.
Change-Id: I4abeb753388c055c36ae0dd916af1ec8d40b7bf0
Merged-In: I4abeb753388c055c36ae0dd916af1ec8d40b7bf0
The network stack will host updatable network components that are being
moved from the system server.
Test: flashed, booted
Bug: b/112869080
Change-Id: I63f98405b69615e7077aed374b8f34a84fb1e4b2
This reverts commit 2473b08b07.
Reason for revert: Fixed build breakage
sysprop_cpp and sysprop_java directly have been linking against
libprotobuf-cpp-full and do not set proto.type field, which could
cause ASAN error due to linking against both libproto version of full
and lite.
Change-Id: I96eb47444d7b6f6b5fef092efc1514fcdd6f33b6
Merged-In: I8e04f643197b6c8a60cc38c6979e41c5de3469f5
Properties accessed across partitions are now schematized and will
become APIs to make explicit interfaces among partitions.
Bug: 117924132
Test: m -j
Change-Id: I8e04f643197b6c8a60cc38c6979e41c5de3469f5
Merged-In: I8e04f643197b6c8a60cc38c6979e41c5de3469f5
Add a service for reporting runtime debug info via dumpsys.
With expected OEM customization / module versions in play
it makes sense to have runtime and associated "core library"
information in bug reports. The dumpsys information can also
be used from from host-side tests.
Bug: 119026403
Bug: 118242715
Test: adb shell dumpsys runtimeinfo
Change-Id: I62637a1c8bfd65d5de9b3e47b8700ead4bc314f2
Log a wtf in the system server when the runtime does not
use a boot image.
Test: m
Test: system boots
Change-Id: I087c269dd1e95eeeab5db4250133f9c61e9399c7
When VR mode is not supported by removing the feature of
android.hardware.vr.high_performance, do not start VrManagerService.
Otherwise, there will be undesired behavior, such as calling into
PowerHal/VrHal for VR hints.
Test: no undesired behavior observed
Bug: 116427118
Change-Id: I3e3c1f02c2fad958d8659d7957d005d25d823c3e
Merged-In: I3e3c1f02c2fad958d8659d7957d005d25d823c3e
We do this by storing the masking inset in a persistent property.
The boot animation then animates itself to where it would be if that
masking inset were applied, then changes the viewport.
For this to work, we also need to make sure the DisplayManagerService
has the right overlay right at the start.
Bug: 112876936
Test: Hide cutout, then reboot. Verify boot animation is smooth.
Change-Id: I3e988b2340b2e0d2be3939bdc6878704c234ccc8
Test: change pin and immediately crash the kernel with
adb shell 'su root sh -c "echo c >/proc/sysrq-trigger"' and boot
Bug: 112175067
Change-Id: Ia5f43f3118e2297fbea43c805ef2f4577bf8a9bf
(cherry picked from commit 50e00c8dc4)
Merged-in: Ia5f43f3118e2297fbea43c805ef2f4577bf8a9bf
Test: change pin and immediately crash the kernel with
adb shell 'su root sh -c "echo c >/proc/sysrq-trigger"' and boot
Bug: 112175067
Change-Id: Ia5f43f3118e2297fbea43c805ef2f4577bf8a9bf
(cherry picked from commit 50e00c8dc4)
Paves the way for moving more behavior to the time detector
service from the NetworkTimeUpdateService.
The NewNetworkTimeUpdateService is effectively identical to
the OldNetworkTimeUpdateService in this initial commit.
Bug: 78217059
Test: build / boot
Merged-In: Ica9fa9b61c849e034d76c0faae5693b60a4a197f
Change-Id: Ica9fa9b61c849e034d76c0faae5693b60a4a197f
This is a do-nothing TimeZoneDetectorService that can be
populated in following commits. A temporary method has been
added so the service has one method.
Unit tests can be run with:
atest FrameworksServicesTests:TimeZoneDetectorServiceTest
Test: build / boot
Test: See above
Merged-In: I565fb5dd2f18b7aac2e5779d346bfe69e9da02af
Change-Id: I565fb5dd2f18b7aac2e5779d346bfe69e9da02af
This is a do-nothing TimeDetectorService that can be
populated in following commits. A temporary method has been
added so the service has one method.
Unit tests can be run with:
atest FrameworksServicesTests:TimeDetectorServiceTest
Test: build / boot
Test: See above
Merged-In: I9e4eac70b944441f34491315cd1ce7fa2b9ae150
Change-Id: I9e4eac70b944441f34491315cd1ce7fa2b9ae150
(cherry picked from commit feeee682a2)
The service and associated code is unused.
Bug: 80462439
Test: build / boot
Merged-In: Ibdfab1b7d2951a0c45e07bd47850af037990841b
Change-Id: Ibdfab1b7d2951a0c45e07bd47850af037990841b
Wifi services crash if run during crypto bounce and are
unnecessary. Disable.
Test: Cherry-pick ag/3898232. System boots, can set pattern, system
reboots. Wifi works at all points.
Bug: 76452634
Change-Id: Id345082f455617386de7fac23fe84e44a8dee6d9
This reverts commit 6750352248.
Reason for revert: Added SElinux policy to allow the service to be started. Verified by local testing on the latest pi-wear-dev.
Bug: 74018496
Bug: 75974893
Change-Id: I9bd8939f6292be9c160e19ebdf934023792059ba
It tracks binder calls into the system server and reports
statistics about CPU time spent per call on multiple dimensions,
e.g. uid or call description.
Usage: dumpsys binder_calls_stats
Overhead: ~0.5% per binder call (if enabled)
Test: manual
Bug: 75318418
Change-Id: I13b854f67f8fd1c9f985b8e45f74dcba2e73b9cb
It turns out that we are still not ready to publish
DropBoxManagerService at an extremely early boot phase as we attempted
so [1]. Our best right now is to publish DropBoxManagerService
immediatelly after SettingsProvider is installed.
Otherwise, if WTF logs are generated in a too early boot phase, it is
indeed possible that two different instances of SettingsProvider can
be created.
The first SettingsProvider instance, created by the
DropBoxManagerService access to Settings, is still present, and a
reference is held in Settings.ContentProviderHolder, such that
framework API access hits this instance.
The second SettingsProvider instance, created by the SettingsService
startup, also still exists and is referenced by
SettingsService.mProvider -- this instance is hit by "settings" shell
command access via SettingsService.onShellCommand().
This creates a situation where settings modifications made via shell
command cannot be seen by the usual frameworks APIs, and vice-versa.
[1]: Ie71d53fc125ebc47fa08ef59a8b7e4f66f2e805c
d79d203b35
Bug: 72399950
Fix: 72450122
Fix: 73505784
Test: atest CtsTextTestCases:android.text.format.cts.DateFormatTest#test_is24HourFormat
Test: atest CtsInputMethodServiceHostTestCases
Test: Manually verified as follows
1. Boot a certain device that has neither /proc/wakelocks nor /d/wakeup_sources exists
2. adb shell settings get secure enabled_input_methods
-> shows non-empty result
Change-Id: I66b68e7f456c4e0e8f5933e65202c61ed862088d
