Bug: 284972285
Test: m & atest NetHttpTests
Change-Id: Id24fa826f617187722b329821dae9d9548bf218b

* changes:
  ethernet: use TrackRecord backtrace for better failure logs
  ethernet: correctly report tethering interface state
  ethernet: add missing assert in eventuallyExpect helper

Allow adding tests or adjusting them to increase coverage on behavior
that is implemented in NsdService.

Test: m
Change-Id: I1b0851bdd220caba68c47d5d50d86a543227817b

Report the nsd metrics event of resolution stop.

Bug: 287546772
Test: atest FrameworksNetTestCases
Change-Id: I2e6de22c946d58a3410f16b2f2689fd3ffc635a4

Those tests are no longer flaky after fix being merged, so add
them back to presubmit.

Fix: 293239828
Test: atest CtsNetTestCases:android.net.cts.Ikev2VpnTest \
      --rerun-until-failure 100
Change-Id: I3e5b0f8053404b8840e8ce67ea82344ab53ce421

* changes:
  StatsValue - define operator+=
  eliminate 'struct Stats' in favour of StatsValue
  eliminate InitStats

- Adds NetHttpCoverageTests to postsubmit. This would allow us
to run tests with @SkipPresubmit in post submit.
- Don't jarjar com.android.testutil so that the annotation can
still be found by the testrunner.

Bug: b/293843426
Test: th
Change-Id: I9e792d338515acba798092bbc4c615aef2cac47a

DeviceConfigUtils has isFeatureEnabled for NetworkStack and Tethering
which were confusing since the difference was only the arguments.
This CL renames isFeatureEnabled for Tethering module to
isTetheringFeatureEnabled to avoid confusion.

Bug: 279108992
Test: m
Change-Id: I05255bb38619afbe621d29524d7910d206a149f1

The created handler thread will stay there for a while until
all the tests are finished. This is considered as a leakage
which will take the resource from the devices. In low end
devices, this may possibly impact the performance. The created
threads should be closed explicitly.

Test: atest FrameworksNetTests
Change-Id: Id5faceec341578f1c9757194c91a433fbb3f657f

The created handler thread will stay there for a while until
all the tests are finished. This is considered as a leakage
which will take the resource from the devices. In low end
devices, this may possibly impact the performance. The created
threads should be closed explicitly.

Test: atest FrameworksNetTests
Change-Id: I332d55d8effdb975fd9a8953b19910a7fb25e426

* changes:
  Use Java code path for bpf map operations in BpfNetMaps
  Rename to isTetheringFeatureNotChickenedOut

Following CLs remove the default value from isFeatureEnabled since the
default value was difficult to use correctly.
For example, there was a code that use isAtLeastT() as default value and
expect feature is always enabled on T+ devices.
However, default value is used only when the device does not have the
flag value or flag value is 0.
So the expectation that the feature is always enabled on T+ devices is
not correct if flag is pushed to the device by mistake or the device is
upgraded from S to T but the flag value is not cleared.
If the feature should be always enabled on T+ devices, `isAtleastT() ||
isFeatureEnabled()` should be used instead of `isFeatureEnabled(
isAtleastT() /* defaultValue*/ )`

After the default value argument is removed from isFeatureEnabled,
isFeatureEnabled is used for default false flag and
isTetheringFeatureNotChickenedOut is used for default true kill switch flag.

automatic_on_off_keepalive_version flag is not configured and devices
don't have this flag value.
So updating to use isTetheringFeatureNotChickenedOut should not have any
behavior change.

Test: m
Bug: 279108992
Change-Id: I43ceab6cc5234bd5a43af3acbd7bd7344844c5a6

Experiment to verify bpf map operations refactoring showed no
change of map size metrics and no system health regressions.
So this CL update BpfNetMaps to always use new code path with kill
switch.

Test: m
Bug: 217624062
Change-Id: Ic17470ac5c075884871ea2df2c13c054f5ac2929

Original name was isTetheringFeatureForceDisabled but it's expected that
the caller of this method will use as `enabled =
!isTetheringFeatureForceDisabled` which is error prune.
So aosp/2711138 renames to isTetheringFeatureNotChickenedOut and updates to
return true if the feature is not disabled.

Bug: 279108992
Test: m
Change-Id: Ia992113e9fb1aee09fcfe48e2b39ca03876063ee

Collect information on resolution success and failure events,
then use this data to build metrics and report them.

Bug: 287546772
Test: atest FrameworksNetTestCases NsdManagerTest
Change-Id: I373ce91fea9d5a278f53fc0c505be90dd39b2d5b

Collect information on discovery start, failure, and stop events,
then use this data to build metrics and report them.

Bug: 287546772
Test: atest FrameworksNetTestCases NsdManagerTest
Change-Id: I2f6a718024870f9658cfeabe8995624a3ad81f05

Mark fixed this test in aosp/2709233

Bug: 295829521
Test: Dashboard will stop showing this failure
Change-Id: I439d4b65b2fc5981af924fb1f1155273bdd52e7a

The testMobileDataPreferredUids was disabled from presubmit
because of flaky test. It's no longer flaky now, so add the test
back to presubmit.

Fix: 292174275
Test: atest android.net.cts.ConnectivityManagerTest\
      #testMobileDataPreferredUids --iterations 1000
Change-Id: I191888069a00942aa4c82f62bd1d62b7fc70b4fa

There may be additional CapabilitiesChanged callbacks before
the test network is torn down. To verify ON_LOST, use
eventuallyExpect instead of expect.

Bug: 295829521
Test: atest ConnectivityManager#testSetOemNetworkPreferenceForTestOnlyPref
Change-Id: I24bb11bf2c2da08e5757c845beeb35b7004f25a4

Ikev2VpnTest uses TestNetworkRunnable to start test networks
to perform VPN related tests. In some low performance devices,
such as cuttlefish, the request to disconnect test network
at the end of the previous test may not be executed completely
yet. It results in the disconnect-pending network being used
by the follow-up test and fail the test. To fix this, the utility
method should wait for onLost callback before performing next
test.

Bug: 293239828
Test: atest CtsNetTestCases:android.net.cts.Ikev2VpnTest \
      --rerun-until-failure 100
Change-Id: I6f3c4f45d3dbadc7e8783c05d6cf67f0896ffa95