The ServiceTypeClient should be removed after socket destroyed
because it's no longer used by any request. The requests will be
matched to the newly created ServiceTypeClient.

Bug: 278635632
Test: atest FrameworksNetTests CtsNetTestCases
Change-Id: Ia917b14d3666f3bfe8e874606a34800a4ce65c5a

Some events are missed their name in the EVENT_NAMES array and
shows the number on dumpsys logs. Add their name to the event
array to show correct info.

Before:
 rec[0]: time=04-12 15:15:23.553 processed=DefaultState org=EnabledState dest=<null> what=21
 rec[1]: time=04-12 15:15:23.555 processed=EnabledState org=EnabledState dest=<null> what=REGISTER_SERVICE
 rec[2]: time=04-12 15:15:24.375 processed=EnabledState org=EnabledState dest=<null> what=DISCOVER_SERVICES
 rec[3]: time=04-12 15:15:25.338 processed=EnabledState org=EnabledState dest=<null> what=23
 rec[4]: time=04-12 15:15:25.342 processed=EnabledState org=EnabledState dest=<null> what=RESOLVE_SERVICE
 rec[5]: time=04-12 15:15:25.344 processed=EnabledState org=EnabledState dest=<null> what=23
 rec[6]: time=04-12 15:15:25.358 processed=EnabledState org=EnabledState dest=<null> what=UNREGISTER_SERVICE
 rec[7]: time=04-12 15:15:25.551 processed=DefaultState org=EnabledState dest=<null> what=22

After:
 rec[0]: time=04-12 15:19:58.955 processed=DefaultState org=EnabledState dest=<null> what=REGISTER_CLIENT
 rec[1]: time=04-12 15:19:58.958 processed=EnabledState org=EnabledState dest=<null> what=REGISTER_SERVICE
 rec[2]: time=04-12 15:19:59.979 processed=EnabledState org=EnabledState dest=<null> what=DISCOVER_SERVICES
 rec[3]: time=04-12 15:20:00.768 processed=EnabledState org=EnabledState dest=<null> what=MDNS_DISCOVERY_MANAGER_EVENT
 rec[4]: time=04-12 15:20:00.777 processed=EnabledState org=EnabledState dest=<null> what=RESOLVE_SERVICE
 rec[5]: time=04-12 15:20:00.778 processed=EnabledState org=EnabledState dest=<null> what=MDNS_DISCOVERY_MANAGER_EVENT
 rec[6]: time=04-12 15:20:00.785 processed=EnabledState org=EnabledState dest=<null> what=UNREGISTER_SERVICE
 rec[7]: time=04-12 15:20:00.925 processed=DefaultState org=EnabledState dest=<null> what=UNREGISTER_CLIENT

Test: adb shell dumpsys servicediscovery
Bug: 275004614
Change-Id: I9c05efc2d2993af353a2a902fd8810f53ab2cd16

Follow up from aosp/1933093. The test failed because certain
devices use a flattened apex that lacks the NETWORK_STACK
permission for sharing with the GTS apk.
Additionally, the NetworkStatsAccess does not recognize the
GTS apk, which holds the MAINLINE_NETWORK_STACK permission.

Test: TH
Bug: 275532605
Change-Id: Icce80a3d3dbea4a84fbd369d9b5fbe0029ab463e

Remaining comment from aosp/2527124

Test: comment-only change
Change-Id: Ic0418bf0bc15b854e25e211bc7399d83487fe971

And address other comments on aosp/2527139

Bug: 273963543
Test: FrameworksNetTests
Change-Id: I4328aea2881badc2be667b900b7ac2a296af1678

The old behavior was to only return those wifis that have
a null key/subscriber ID (e.g. not carrier wifi). To keep
backward compatibility, restore that behavior.

Also only crash the old methods on devices with initial
SDK U, because it can't be mandated that devices that
upgrade must not call these methods.

Bug: 273963543
Test: FrameworksNetTests
Merged-In: Iefb976ed86a732158964fae38e9e601001c2637a
Change-Id: Iefb976ed86a732158964fae38e9e601001c2637a

Test: atest FrameworksNetTests:android.net.connectivity.android.net.NetworkTemplateTest
Test: atest FrameworksNetTests:android.net.connectivity.android.net.netstats.NetworkTemplateTest
Bug: 273963543
Change-Id: I0a8f94df124147e92d35cf474b3d69d1dee6902c

Change-Id: Ia8b745d674262a312044cd4e2b4320917d550725
Test: TH
Bug: 272150669

Add back methods that were removed in recent refactoring
(aosp/2286634).  Even though they are fully @hide, some usage by
an OEM has been found on T.

Test: atest; the methods should be unused except on the OEM device
Change-Id: Icd67fb8b7e63cc131356a30a3044458ae2f5efff

Fixes: 274725920
Test: Presubmits
Change-Id: I325f686af34720cddbbc63580edec8477a9acc45

Apps targeting sdk < U are considered to use a legacy native
daemon as NsdManager backend, but other apps use a
platform-integration mDNS implementation as backend. So add a
CompatChange flag to enable platform backend for non-legacy
apps.

Bug: 270306772
Test: atest FrameworksNetTests CtsNetTestCases
Change-Id: I7ba58f8a5186fb49ad5f8aeacc8b8234bef1eabe

There's now a shared defaults module for this kind of filegroup. Use
that and remove the previously duplicated prop.

Bug: 271816210
Test: m
Change-Id: Ia4547984a73179fdfc98c427482dd10246f9722c

matchesWifiNetworkKey expects a non-null value of the
wifiNetworkKey, however, it might be null in practice
and the null wifiNetworkKey will be stored into disk.
And then when the matchesWifiNetworkKey is called, the
requireNonNull will crash the system. Thus, remove the
requireNonNull from matchesWifiNetworkKey to avoid
system crash and handle if the wifiNetworkKey is null
then it should not match a template with non-empty
mMatchWifiNetworkKeys.

Check if WifiInfo contains a null network key then skip it
to prevent adding the identity to the network identity set.
Also, add a Log.wtf when setWifiNetworkKey(info.getNetworkKey()),
this might be useful to catch why the wifiNetworkKey is
null.

Bug: 267815242
Bug: 266598304
Test: FrmeworksNetTests
Change-Id: I9c21f7e3dca9482133c7e331741cf808105414e9

HalfSheetUX uses private apis of framework-connectivity-t. It appears
that using just "framework-connectivity-t" will compile against either
the stubs or source of framework-connectivity-t depending on if the
module was included in the same apex as it or not.

aosp/2462194 is adding the ability to compile modules like HalfSheetUX
outside of the context of the apex, at which point it will start
compiling against stubs if we don't explicetly specify to use the
implementation library.

Bug: 254205429
Test: m javac-check with aosp/2462194
Change-Id: I1d41e47aa0e4e3d320b78b28cfc181fc30c4537c
Merged-In: I1d41e47aa0e4e3d320b78b28cfc181fc30c4537c

registerServiceInfoCallback currently only sends updates for
addresses added, but does not handle removes (expiration) and
TXT/SRV record updates. Thus, migrate its backend to
MdnsDiscoveryManager which can support the expiration update.

Bug: 266030646
Test: atest FrameworksNetTests CteNetTestCases
Change-Id: I72add213935dc1beacb6277007868ad30bd89c00

aosp/2241257 and aosp/2241058 are included in the mainline train
release and it causes the T device can’t pass CTS-13_R2/CTS-13_R3.
It’s because some parameters of the constructor have been removed.

Add it back to keep backward compatibility.

Test: 1. atest CtsNetTestCases:android.net.netstats.NetworkTemplateTest
      2. CTS-13_R3
Fix: 269974916
Fix: 269834366
Change-Id: I197fbfb8419e4d4885a97a93a71f13d33c3f02da

Since there's a bug about apps still call hidden
APIs, it's likely some developers might build apps
against hidden API.
To be safe, add these methods back to avoid apps
crash. These methods are still in
hiddenapi-max-target-o-low-priority.txt, so they
should not be removed without considering hidden
API usage. Removing hidden API in general isn't
forbidden, but the ones listed in hiddenapi usage
need extra care.

Bug: 269178029
Test: 1. atest CtsNetTestCases:android.net.netstats.NetworkTemplateTest
      2. CTS-13_R3
Change-Id: I485369274ccacd314230e3d267df665d8083bf97

Test: 1. manual test with test app
      2. atest CtsNetTestCases:android.net.cts.NetworkStatsManagerTest
Fix: 249920382
Fix: 269561252
Change-Id: I4af7fd640d551be3ef62ea248e58bf8bae5e0b87

There's a user still call this method which causes the app
crash and it might be the app was built against hidden API.
The bug was observed after updating the mainline module so
adding this method back to the module to avoid apps crash
due to the method is not found in NetworkTemplate.

Bug: 269178029
Test: build
Change-Id: I06dc694b1ce060488135f5d34138f14abd50ad32

A template with type MATCH_MOBILE could have empty subscriber
IDs but it should not contain one of subscriber IDs is null.

An app might still build a template through the constructor
which is annotated with @UnsupportedAppUsage. NetworkTemplate
will throw an exception when the app trying to construct a
template with type *_WILDCARD. The constructor should be
backward compatible with old version so this change is to add
the backward compatibility to have NetworkTemplate not throw
an exception.

Bug: 267701889
Test: FrameworksNetTests
Change-Id: I23a607dae508e0c53520e2edf187cb611ed36b68

mSubscriberId was removed from NetworkTemplate and equals/hashCode
does not check the mMatchSubscriberIds. This will cause tempates
with different subscriber ID matching that are equal now. Thus
adding the check for mMatchSubscriberIds to fix it.

Bug: 267968247
Test: FrameworksNetTests:NetworkTemplateTest
Change-Id: I2956bf4c8cf2c4d73ebe102d53e755fbcc5d7642

ConnectivityCompatChanges.java becomes the centralized place for all the
CompatChanges used in the Connectivity module. By putting all the
CompatChanges here, we are able to manage them under a single
platform_compat_config.

Bug: 268440216
Test: atest FrameworksNetTests
Change-Id: I3e17af545718073d7d1c96e27298e7790563fd33

Rename the method onResolveStopped(NsdServiceInfo) to
onResolutionStopped(NsdServiceInfo) for consistency as API
review feedback.

Bug: 266811051
Test: atest FrameworksNetTests CtsNetTestCases
Change-Id: I788b5903ecaf78a9aa960dd3b7413a22f171c72f

Currently, the resolution is a one shot query, it only notifies
the first finding service information. There is no way to listen
the service update. Thus, add a new API that can register a
callback to listen to the service updates continuously.

Bug: 245369943
Test: atest FrameworksNetTests CtsNetTestCases
Change-Id: I0e9d92b9028375feb3e344ab6c4acb515c5b2be9

Resolve service may take long time due to network issue or
using wrong service information, but users are not able to stop
it. They can only wait for the callback of resolveService to end,
which sometimes takes a long time. Thus, add the new API that
users can stop the service resolution.

Bug: 245369943
Test: atest FrameworksNetTests CtsNetTestCases
Change-Id: I6b6183c8c73f8db981b9afa51fbc73bf886d9ed3

Service resolved should be notified when receive the
onServiceFound callbacks from MdnsServiceBrowserListener

Bug: 254166302
Test: atest FrameworksNetTests CtsNetTestCases
Change-Id: I681720065084bf3449c5b1ab44cd4ed6a659dcdb

Service found should be notified when receives the
onServiceNameDiscovered callbacks from
MdnsServiceBrowserListener.

Bug: 254166302
Test: atest FrameworksNetTests CtsNetTestCases
Change-Id: I3f41b4fe85cd85ad356fa764663187a88914412c

Register/Unregister the listener to/from MdnsDiscoveryManager
when discovery started/stopped.

Bug: 254166302
Test: atest FrameworksNetTests CtsNetTestsCases
Change-Id: Ibd782029826ac5856c608165928cd942e46dd9a4

This commit exposes APIs to migrate a tunnel mode transform to
new source/destination addresses, as required by MOBIKE.

By calling the exposed API, the caller only caches the new
address in the transform. To complete the migration, caller
MUST apply the tranform to the appropriate tunnel.

This API design is mainly based on the kernel interface and use
cases. The Linux kernel requires Android to provide both the
IpSecTransform and the IpSecTunnelInterface to perform the
migration. And in most cases those two instances are managed
by different entities: IpSecTranform is managed by the key
exchange protocol (e.g. IKE) and IpSecTunnelInterface is managed by
the security tunnel provider (e.g. VPN, VCN, and IWLAN). Thus the
migration process has been designed to have two steps where the
key exchange protocol negotiates and caches the new address,
passes out the updated transform, and the security tunnel
provider applies the transform to a tunnel.

Another benefit of this API is it can also apply to the case
where the network peer does not support MOBIKE and thus cannot
update the existing transforms. In this case, the key exchange
protocol can create a new transform and give it to the security
tunnel provider, and the tunnel provider can still call "apply"
to perform migration without needing to know the details of the
transform update process.

Bug: 169171001
Test: atest FrameworksNetTests, IpSecManagerTunnelTest
Change-Id: I0658cdb09fb31f7e0fb9d0b07f37c2b72b6e705f

Based on documentation, String parameter to CloseGuard.open() should be the name of the method to close the resource (in this case Resource.close()).

As currently written, the debug output from warnIfOpen() will be "Resource failed to call constructor."

Bug: 131838407
Test: TreeHugger
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: I25f005a7944bf3995ae6dc4c021ed1034f2187dd

Based on documentation, String parameter to CloseGuard.open() should be the name of the method to close the resource (in this case Resource.close()).

As currently written, the debug output from warnIfOpen() will be "Resource failed to call constructor."

Bug: 131838407
Test: TreeHugger
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: I9b8a71f9076712014bb883bbde5e5fbdb018b2d5

NetworkTemplate was moved into mainline module and
provide Builder for external users to build template.
MATCH_MOBILE_WILDCARD and MATCH_WIFI_WILDCARD are not
exposed so currently the usage of these two constants
only in module. This change is to remove the constants.

After removing the constants, two methods matchesWifi
and matchesMobile are changed to adapt the original
matches[Wifi|Mobile]Wildcard behavior.

Bug: 238843364
Test: FrameworksNetTests
Change-Id: I9e156b504ce4eb903b39279e443fad27bc921186

Currently NetworkTemplate provides Builder for users to build a
template and it supports multiple subscriber ids. The mSubscriberId
is a local variable in current design so it can be removed from
NetworkTemplate and also can be removed from the parameter of a
constructor as long as the constructor is not annotated with
@UnsupportedAppUsage.

Bug: 238843364
Test: build, FrameworksNetTests
      CtsNetTestCases
Change-Id: Ic6a695e1c1719111bf26ba905a38673952a24736

Currently, mSubscriberIdMatchRule is only used in NetworkTemplate
and it depends on if mMatchSubscriberIds is empty or not. Thus,
remove it since is not really necessary and replace it with
checking matchSubscriberIds if needed.

Bug: 238843364
Test: build, FrameworksNetTests
      CtsNetTestCases
Change-Id: Ic66d2ff2826846778b004bb15a4718a62fa1f470

Bug: 169169973
Test: atest IpSecServiceParameterizedTest (new tests added)
Change-Id: I3dd45b29163cd1e0cdbef08cb8aabdb629cf73bc

This commit adds methods to support migrating tunnel mode
IpSecTransform to new source/destination addresses.

Bug: 169171001
Test: atest FrameworksNetTests (new tests added)
Change-Id: Ic177015fba5b62d3f73009633118109d3631086f

Currently, NET_CAPABILITY_TEMPORARILY_NOT_METERED traffic is
counted as unmetered data usage, and "mobile data usage"
settings screen shows metered data usage which means it doesn't
include TEMPORARILY_NOT_METERED traffic. However, some carriers
show TEMPORARILY_NOT_METERED data usage in customer's bill which
would be confusing if the data usage in the device shows different
results. Thus, revert the change to avoid confusing.

Bug: 183776809
Test: FrameworksNetTests
Change-Id: I4ef25095462adf30b1ecf82134996739f167930d

In current design, mMatchSubscriberIds can be null which has the
same meaning with empty set, and is not easy to maintain since it
need more checks for this variable when mSubscriberId is null.
Thus make mMatchSubscriberIds NonNull for the maintenance.

Bug: 238843364
Test: build, FrameworksNetTests
Change-Id: I6cfc529b5f4a39ded8598283ff968f2f4d1bc89f

NetworkTemplate.Builder was introduced in Android T and the callers
outside of Connectivity module should use this Builder to build a
template instead. Thus remove the buildTemplate* methods from
NetworkTemplate class. Also, deprecate the methods which are
annotated with @UnsupportedAppUsage and also use Builder to build
the template.

Bug: 238843364
Test: build, FrameworksNetTests
Change-Id: I3190325f6663e4771edf5c7a19bf5ecc7780bf12

Bug: 254143771
Test: TreeHugger
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: I9f91fc250e0fa40c9dc9ee3e7a2f7bc3cf2ebc75