This new helper method provides a convenient way to validate and
determine whether tethering metrics' data usage can be collected
for a given upstream type.
Bug: 282861165
Test: atest TetheringMetricsTest
Change-Id: I7527ce56815e1f7e9382be1a54b84c91552e1129
Wayne Maauthored and Paul Hucommited 2024-07-15 09:30:43 +00:00
This is a no-op change. Use Dependencies to mock static methods
instead of using spy().
Bug: 282861165
Test: atest TetheringTests
Change-Id: Ia8726375a8da17dab745848b046f4c32b84d5d77
This is a no-op change. Add a new TetheringMetrics constructor
for the subsequent TetheringMetrics TX/RX calculation feature.
Bug: 282861165
Test: atest TetheringTests
Change-Id: I9d368d8ad527d73cc91e02ee72a5731fb4c6c195
These tests are run as part of the MTS suite, setting min_sdk_version
ensures that these tests can be installed on an older device and test
the corresponding mainline module.
Test: atest [<mod_name>...]
Bug: 156476221
Change-Id: Ic94b00ee1818063fc9f13e4ecffaca88f58c30c3
Replace TetheringRequestParcel with TetheringRequest in the service
code. This makes future changes easier since TetheringRequest is a
proper class and not an AIDL object.
Bug: 216524590
Test: atest TetheringTest TetheringManagerTest
Change-Id: I0cdee597d393e53feb9fe9aff4d6c67330c88599
Quang Anh Luongcommited 2024-06-26 16:25:13 +09:00
* changes:
Clean up BpfCoordinatorTest after refactors
Move upstream interface BPF support check to BpfCoordinator
Refactor BPF map update logic of upstream interface change
Centralize IP neighbor monitoring in BpfCoordinator
Get rid of BpfCoordinator#mPollingStarted
KH Shiauthored and Gerrit Code Reviewcommited 2024-06-22 12:34:34 +00:00
* changes:
Add some documentation to EthernetInterfaceStateMachine
Add support for link state tracking
Add unit test class for EthernetInterfaceStateMachine
Add missing call to add states to state machine
Make class testable
Log.wtf if block times out
Wait on IpClient#onQuit in StartedState#exit
Use TAG instead
Introduce CMD_ON_NETWORK_NEEDED and CMD_ON_NETWORK_UNNEEDED messages
Clean up and document state lifecycle
Add convenience method for processMessage with no arguments
Implement NetworkOffer in EthernetInterfaceStateMachine
Add boilerplate to represent interface states in state machine
Fix typo in SyncStateMachineTest
Move SyncStateMachine to staticlibs
Treehugger Robotauthored and Gerrit Code Reviewcommited 2024-06-21 19:01:38 +00:00
Removed the use of a real IpServer instance in BpfCoordinatorTest,
as it is no longer needed after the refactors in aosp/2716773 and
aosp/3113123.
Test: atest TetheringTests
Bug: 294025403
Change-Id: I4760669372275a0165f4980705e328c80e0649a5
Removed upstream interface BPF support check from IpServer and moved
it to BpfCoordinator. The support check can be performed directly in
BpfCoordinator using the interface name.
Test: atest TetheringTests
Bug: 294025403
Change-Id: Ia51bc3ab18dd21afdca9db9199fa060aab68f768
200ms does not seem to be enough for our test infra. There have been test failures where tethering test was started on devices with physical interfaces that were available for tethering (i.e. cuttlefish on Android R and S).
This change increases the timeout to 1000ms, which unfortunately means the tests will take a bit longer.
Test: TH
Bug: 348044867
Change-Id: I5fe541cf1db1097040066baa73ba949533718ef6
This CL centralizes IP neighbor monitoring in the BpfCoordinator,
eliminating the need for redundant monitors in each IpServer.
Test: atest TetheringTests
Bug: 294025403
Change-Id: Ie6fbc3eeaa0ed4a34a6680714f44678e2d5b633a
Refactor IpServer/BpfCoordinator to remove mPollingStarted, and just
maintain the invariant that "data usage polling and conntrack
monitoring are running iff at least one IpServer is serving".
Test: atest TetheringTests
Bug: 294025403
Change-Id: I55064ee78e1cda12ad76b7c31496d760dcffdd76
because the shared r/w lock on the first
conflicts with the exclusive r/w lock on the second
This fixes roughly *half* the BpfMapTest tests.
Test: TreeHugger, atest BpfMapTest less broken
Signed-off-by: Maciej Żenczykowski <maze@google.com>
Change-Id: I9d500597cad16bb7756df3e6feb9a7aa762e59d9
Maciej Żenczykowskicommited 2024-06-15 14:04:03 -07:00
This CL makes the TetheredInterfaceRequester a class member that gets
initialized (and therefore creates a tethered interface request) in
setup, and releases it again in tearDown. This ensures that when a test
interface is created, ethernet never puts said interface in client mode.
In a previous iteration of this test, there was a race where test
interfaces were briefly put in client mode, which started an IpClient on
the interface. One of the first things the IpClient startup code does is
enable IPv6 which causes RS and MLD packets to be transmitted. This
caused a lot of flaky test failures as the tests were not expecting to
see these packets (and they did not drain the interface appropriately).
Making the tethered interface request permanent during the runtime of
this test solves this issue. As a follow up, the test interfaces should
also be drained when first being passed to TetheringTester.
Test: atest EthernetTetheringTest --iterations 50 (-> there was one
flaky failure that is still being investigated).
Bug: 341800184
Change-Id: Ib0a2518b89d8cc05bb800db6ec5e31f527ba2daa
This should speed up BPF map reads by returning the cached value
without needing a system call.
Bug: 343166906
Test: added parameterization to BpfMapTest
Change-Id: Ifc0c5baa80f6d46356434c249749e87fe2f9ec6c
Lorenzo Colitticommited 2024-06-05 06:48:27 +00:00
When running locally, this occasionally fails due to unrelated fds
being opened or closed while the test is running.
Bug: 343166906
Test: test-only change
Change-Id: I437e66077f0471572ec2b215db63a000a4622af5
Lorenzo Colitticommited 2024-06-05 06:48:15 +00:00
If we want to get rid of the flakiness, we may need to change the
production code. You can find a more detailed analysis in
b/318121782/comment#4.
Here's a quick and easy alternative fix: just move setupOnce from
EthernetTetheringTestBase to EthernetTetheringTest. The flakiness
started when I added setupOnce into EthernetTetheringTestBase to
fix flakiness in EthernetTetheringTest.
Bug: 318121782
Test: TH
Change-Id: I0a6fb25773aa60872f71fe4b34e42a0b745a767f
* Struct.java is removed from net-utils-device-common-bpf and
net-utils-device-common-struct's source list
* Created a lib net-utils-device-common-struct-base for Struct.java
* Updated libs that require Struct.java to include
net-utils-device-common-struct-base.
* Replaced net-utils-multicast-forwarding-structs with
net-utils-device-common-struct in service-connectivity
Bug: 323503345
Test: build tethering module.
Change-Id: I3f60715c3d8043b7569db7612dee9d4fa1ccba59
Removing "mWifiTetherRequested" because it might cause Wi-Fi tethering
to not have an upstream if P2P is turned on. The issue could be
recovered when the upstream's capabilities or link properties change,
tethering calls "chooseUpstreamType" to select an upstream. If none of
the above situations happen, Wi-Fi tethering won't have an upstream.
1. When Wifi P2P is enabled and tethering enters TetherModeAliveState.
2. Turning on wifi tethering, TetherModeAliveState would first receive
EVENT_IFACE_SERVING_STATE_INACTIVE when wifi tethering's IpServer
is created and enters IpServer's InitialState. Then it will call
updateUpstreamWanted and change mUpstreamWanted from false to true
before wifi tethering's IpServer actually enters TetheredState.
3. When TetherModeAliveState receives EVENT_IFACE_SERVING_STATE_ACTIVE,
which is triggered by wifi tethering's IpServer entering
TetheredState, it won't call chooseUpstreamType() because
mUpstreamWanted is already true.
case EVENT_IFACE_SERVING_STATE_ACTIVE: {
IpServer who = (IpServer) message.obj;
if (VDBG) Log.d(TAG, "Tether Mode requested by " + who);
handleInterfaceServingStateActive(message.arg1, who);
who.sendMessage(IpServer.CMD_TETHER_CONNECTION_CHANGED,
mCurrentUpstreamIfaceSet);
// If there has been a change and an upstream is now
// desired, kick off the selection process.
final boolean previousUpstreamWanted = updateUpstreamWanted();
if (!previousUpstreamWanted && mUpstreamWanted) {
chooseUpstreamType(true);
}
mWifiTetherRequested is only used to check upstreamWanted() and only be
used in TetherModeAliveState so that it is
safe to remove mWifiTetherRequested and always use mForwardedDownstreams
to check whether upstream is needed for TetherModeAliveState because all
the Tethered IpServer is added to mForwardedDownstreams when it enter
Tethered State and be removed from mForwardedDownstreams when it exit
Tethered State.
This change also move the upstreamWanted() function inside TetherMainSM
to prevent it from being used incorrectly outside of TetherMainSM.
Bug: 329552689
Test: atest TetheringTests
Change-Id: I6842170f455b280dd79aae039a1117d5e6d677f3
If P2P is enabled first then turn on wifi tethering, it doesn't have
upstream.
Bug: 329552689
Test: atest TetheringTests
Change-Id: Ia39031167a3cf7751ae549ea5ce147cb9cb87607
Calling fail() in a callback not running in the test thread will crash
the test process, instead of failing the test.
This affects other tests as the whole process dies.
Avoid that by adding errors to a list, which is verified in tearDown().
Test: atest
Change-Id: I354352b61fb1909bcac519bf58d49b36e81c710f
This CL is created as a best effort to migrate test targets
to the new android ownership model. If you find incorrect or unnecessary
attribution in this CL, please create a separate CL to fix that.
For more details please refer to the link below,
<add g3 doc link>
Bug: 304529413
Test: N/A
Change-Id: I243f17b3f0ad7af9ffa15ca242456e36688733f5
Merged-In: I243f17b3f0ad7af9ffa15ca242456e36688733f5
Add a utility to reconnect mobile data, and refactor toggleWifi to match
it.
The new reconnectWifi utils do not support turning wifi on and back off
anymore, but this was dead code as no test was using it (wifi is
supposed to be connected before starting tests).
Bug: 253698734
Test: atest
Change-Id: I2b8b8d186d05bb57c6a45d872febc54def8c6d59
The logic is currently disabled by default, adding a flag so
that we can compare result with A/B rollout. Tethering only read
this flag once when it is created.
Bug: 319212113
Test: atest TetheringTests
Change-Id: I656b0389e56d29dd691738ee3abfa6418746846f
This reverts commit cf11be9eb2.
Reason for revert: avoid some IpServer(s) running the new sync state machine while others use the async state machine.
Bug: 319212113
Change-Id: Ib086a7ccd102f38e579c79f94aceae9971f1853a
Mark Chienauthored and Markcommited 2024-01-22 11:34:14 +00:00
This change includes:
1. Refactor the logic in Tethering#dump into utils class.
2. Move the utils class to a common place which could be referenced
from other sub-modules.
3. Add @MonitorThreadLeak annotation to enforce
there is no thread leak problem.
Test: atest FrameworksNetTests NetworkStaticLibTests
Test: atest ConnectivityCoverageTests:com.android.networkstack.tethering.TetheringTest#testDumpTetheringLog
Test: adb shell dumpsys tethering (with hardcoded exception)
Fix: 312669345
Change-Id: Ia6fdfeeec2110afa0ec9e056e9db3843748845c3
The first tethering test case might have a tethering restart when
handling IP conflicts. Currently, we start a hotspot once before
tethering, but the test is still flaky in Cuttlefish because it
doesn't support hotspots. Instead, we'll start ethernet tethering.
Bug: 254183718
Test: atest CtsTetheringTest
Change-Id: I695d4e1b3f961893bc30b3144faf7ad7d2d70574
Markauthored and Mark Chiencommited 2023-12-27 08:57:57 +00:00
Wayne Ma
authored and
Paul Hu
commited 
Treehugger Robot
authored and
Gerrit Code Review
commited 
Spandan Das
commited 
Quang Anh Luong
commited 
KH Shi
authored and
Patrick Rohr
commited 
Maciej Żenczykowski
commited 
Lorenzo Colitti
commited 
Mark Chien
authored and
Yang Sun
authored and
Weilun Du
commited 
Remi NGUYEN VAN
commited 
Aditya Choudhary
commited 
Yuyang Huang
authored and
Junyu Lai
authored and