Replace the hidden setDefaultNetId and clearDefaultNetId NMS
APIs with accessing INetd directly for the incoming
ConnectivityService mainline.
Bug: 170598012
Test: atest FrameworksNetTests
Test: manually connect and disconnect wifi
Change-Id: I162fae5ca444207a037e5ac4bf8fa0a77a648ca1
ConnectivityService is going to be a mainline module, it can only
use formal APIs or @SystemApi. So use public API -
Context#getSystemService() instead of hidden API -
ServiceManager#checkService().
Bug: 170598012
Test: atest FrameworksNetTests
Change-Id: I9824caa7aec57e70f0ba405fcce39f9bc068732d
ConnectivityService is going to become a mainline module, and
it will not able to use hidden method anymore. Thus, use
alternative new sysprop as API to control the tcp init rwnd
value.
Bug: 170917042
Test: adb shell getprop net.tcp_def_init_rwnd and check if
value is set correctly
Test: atest FrameworksNetTests
Change-Id: If9e99c88de50b6829721b0dfacc430a3b53c7728
Create a new target service-connectivity to split
ConnectivityService from services.core.
Add ConnectivityServiceInitializer for initializing
ConnectivityService and add systemReady() in
ConnectivityManager so that SystemServer can call systemReady()
through ConnectivityManager which won't change current behavior.
Bug: 158268939
Test: make target-java, make host-java
atest FrameworksNetIntegrationTests
atest FrameworksNetTests
make, device can boot,
atest CtsStrictJavaPackagesTestCases
wifi and mobile data work.
Change-Id: I99401772ba9c1c34adca20040da3c7c72d86ddd9
Merged-In: Ie732bfaf381404af0bb599ca2f421a96e7aa4257
Create a new target service-connectivity to split
ConnectivityService from services.core.
Add ConnectivityServiceInitializer for initializing
ConnectivityService and add systemReady() in
ConnectivityManager so that SystemServer can call systemReady()
through ConnectivityManager which won't change current behavior.
Bug: 158268939
Test: make target-java, make host-java
atest FrameworksNetIntegrationTests
atest FrameworksNetTests
make, device can boot,
atest CtsStrictJavaPackagesTestCases
wifi and mobile data work.
Change-Id: Ie732bfaf381404af0bb599ca2f421a96e7aa4257
In current design, crash has been generated when stop function
has been re-entered to catch unexpected behavior. However,
it is possible to re-enter stop function if the network
disconnection occurs after stopping.
Thus, skip stop if keepalive is already in stopping state.
Test: atest ConnectivityServiceTest#testNattSocketKeepalives \
--rerun-until-failure 60000
Bug: 167332570
Change-Id: Ic7068ad3dc990e957c37b8d87d48ebb6469b101f
The existing method is confusing (the argument used to be called
includeDying) and it puts the burden on the caller (which need to
understand what the parameter means).
Furthermore:
- The majority of calls are for getUsers(excludeDying=true).
- The calls for getUsers(excludeDying=false) are equivalent to
calls to getUsers()
Test: m
Test: a VpnTest ConnectivityServiceTest PermissionMonitorTest
Bug: 157921703
Change-Id: Ife767a40b7b7790ba28b5377046de822ddbf275c
Currently, the callbacks of stopping were fired when stop procedure
is started, because the upper layer apps only care about the reason
of stopping instead of stopping result. Thus, there is no need to
wait for the result comes back. However, this behavior generates
races if apps want to re-start keepalive immediately since the
resources are not released yet.
Fix: 134891441
Fix: 140305589
Test: atest com.android.server.ConnectivityServiceTest#testPacketKeepalives \
--rerun-until-failure 1000
Change-Id: I987776a9211a50e964c4675b747bc10e203750f1
Test extra info sent to NetworkMonitor correctly if network
agent is created through new NetworkAgent constructor without
legacy network info taken as parameter.
Bug: 156173829
Test: atest FrameworkNetTests
Change-Id: I4f827664c528bea30cc957a0a617dd37693f4460
This is only necessary when learning the NAT64 prefix from the
RA, because if the NAT64 prefix is learned from DNS, the DNS
resolver already knows the prefix and automatically enables
DNS64 synthesis.
The DNS resolver needs to be informed of the prefix any time
clat is running on a prefix learned from an RA. This is simple to
implement: just set the prefix when starting clat if prefix
discovery is not running, and clear the prefix when stopping clat
if prefix discovery was not running. This ensures that the prefix
is cleared iff it was set.
Bug: 156914456
Test: new unit test coverage
Change-Id: If8ad2d30712a6df3e207c8d3e8a129705242191e
This is not particularly likely to happen unless the pref64 RA is
sent by a different router than the main RA. But more tests are
always good, and this additional coverage will be more useful
in an upcoming change.
Bug: 150648313
Test: test-only change
Change-Id: I3316d49d42100800740afadc4edf0a13a4d8377c
As NetworkAgent is in a transition where all agents need
to include the NOT_SUSPENDED capability as part of their
migration to the system API, ConnectivityService adds it
forcefully to all agents that don't have the CELLULAR
transport. This doesn't include VPNs when VPNs have some
cellular network as their underlying network.
The best way to solve this is to make sure the VPN
capabilities reflect those of the underlying networks as
far as the NOT_SUSPENDED capability is concerned. This
is how they work for other similar capabilities.
This also happens to contain a drive-by fix for an issue
with a spurious capabilities callback is triggered when
a VPN connects and it has any underlying network (which
means almost always, because it will take the default
network if it doesn't declare any). Fixing this was
necessary to have a cogent test of this issue, but it
could be moved to another patch or it could stay unfixed
with some minor ajustment to the tests if judged too
dangerous to include in R at this point.
Test: New tests in this patch. Also manually tested with
tcpdump as described in b/150570873.
Bug: 150570873
Change-Id: I3e4ff990c0d4825b21c7679be29a482a2d1324ec
When a VPN connects and it has any underlying network (which
means almost always, because it will take the default network
if it doesn't declare any), it has default capabilities and
will only take the capabilities of its underlying network
as part of an update happening after making the network
available but before the rematch can take place. This in turn
causes the capabilities callback sent as part of the rematch
to be spuriously sent.
Test: FrameworksNetTests. Also tested together with a
followup that adds tests with drive-by coverage for this.
Bug: 150570873
Change-Id: Id7d8bba486bada1a7ba5b0f152d2aa02e407f249
This change sets the owner and administrator UIDs for test networks when
their initial values match the UID for the app creating the test
network. This ensures that apps registering test networks can only make
themselves owners / administrators of the network.
Bug: 153449964
Test: atest NetworkAgentTest
Change-Id: I3a974700aa1d83cb285295ed1de0aa263e2e5b58
Address issues found during AIDL review:
- Rename clientAddr to singleClientAddr
- Do not use a ParcelableBundle for notifyNetworkTested or
notifyDataStallSuspected; instead use AIDL parcelables for stronger
backwards compatibility guarantees.
Test: atest NetworkMonitorTest ConnectivityServiceTest
ConnectivityServiceIntegrationTest, manual
Bug: 153500847
Change-Id: Id9b71784e5f6294d203230e57737979e063ff0f8
Currently, if a prefix is learned from an RA while prefix
discovery is running, clatd will be correctly started, but
prefix discovery will be stopped.
In order to fix this, make it possible to call
stopPrefixDiscovery without transitioning to IDLE state (which
is obviously necessary in this case), by moving the assignment of
the next state from that method to its callers. For consistency,
do the same for startPrefixDiscovery.
Bug: 150648313
Test: new test coverage
Change-Id: I3803fa3d9806848b331c35ee8bac256934bd1f21
The NAT64 prefix from the RA always takes precedence over the
NAT64 prefix from DNS discovery, because it is detected faster,
and detecting it does not require sending any packets.
Bug: 150648313
Test: new unit test
Change-Id: Ic7452431d2d9aea1ae59b67a9d8383c6cc5b3902
