Ignore the last remaining test in ConnectivityServiceTest with spurious
failures. testRequestBenchmark has some intrinsic chances of failure due
to the fact it attempts to assert elapsed time durations against a
reference target.
Bug: 32561414
Test: no functional change
Change-Id: Ib25d76581b47997b2ef84df3e6a9fd9224b85d92
This patch attempts to fix the remaining spurious failures in
ConnectivityServiceTest, which have two causes:
- waitForIdle() does not take into account the NetworkAgents handlers.
- the deadlines in testRequestBenchmark are sometimes exceeded.
To fix the first issue, waitForIdle() is moved to a test level instance
method and also calls waitForIdleHandler on any non null
MockNetworkAgent. This is expected to fix spurious errors for the
following tests:
- testMobildeDataAlwaysOn
- testLingering
- testPacketKeepAlive
- testMMSonWiFi
To fix the second issue, the deadlines for testRequestBenchmark are
extended by 10ms. Also, the failure message is made more actionable by
providing the total time it took for the operation, instead of printing
the number of dispatches that were achieved before the deadline.
Bug: 32561414
Test: tests pass many times in a row (~500).
Change-Id: Id33c6ac1edfb0b89634fa7789dccb2da237e2709
Also, make all tests start with mobile data always on disabled.
This is because some tests only pass when mobile data always on
is disabled. This doesn't cause any problems when running all
the tests in the file, because these tests are always run after
one or more calls to tearDown, which disables mobile data always
on. However, it does cause issues when those tests are run alone.
Test: new test passes 50 times in a row
Test: ConnectivityServiceTest passes
Change-Id: I1eef5d7f5ec5464e0f9a1d7f1130d9ba6dea4557
This patch introduces between ConnectivityManager and
ConnectivityService a mechanism for propagating back to clients
informative errors in the form of error codes and associated custom
runtime exceptions.
Without error code, the service can only throw a limited number of
different exceptions over Binder. Furthermore the throw site stack
traces are always loss. Although for individual instances of a throw,
the error message can be inspected, aggregations of stack traces from
app crashes sanitize error messages and only leaves the stack traces.
This makes debugging dificult for some service calls such as
requestNetwork that can have a variety of failure modes.
In this patch only one failure mode is codified. More can be added later
at a light cost by: 1) defining an error code, 2) defining an
associated exception, 3) mapping the code to the exception. This patch
can serves as a template for doing so.
Test: $ runtest frameworks-net,
#testNetworkRequestMaximum() detects the new exception type.
Bug: 36556809, 36701874
Change-Id: I611fd7915575c9e418f7149fcdc8a879d2a3716d
Wifi Aware networks are per app - i.e. a requestor gets
a dedicated network. Change verifies that the only the
original requestor matches the created network (using UID).
Bug: 36053921
Test: Integration (sl4a) tests
Change-Id: I4ff3994731dd7ccb88e2bea333d1e6905b136f02
This allows an application that knows how to provide seamless
network connectivity (e.g., using QUIC multipath) to find out if
doing so is desired.
(cherry picked from commit 2de4925f5c)
Test: builds, boots, runtest frameworks-net passes.
Bug: 34630278
Change-Id: Ic7fd0b9e1cd879fdfaf84009d7125391895e9087
Tests various combinations of empty / non-empty NetworkSpecifiers
and changing specifiers on the fly.
Test: ConnectivityServiceTest passes
Bug: 27533960
Change-Id: Ibd6b4efc861dc51689e46f3882ba859223411bf0
Test: ConnectivityServiceTest passes on ryu on internal master
Test: ConnectivityServiceTest passes on bullhead
Change-Id: If94102c7df0257ea9e69e72b07a685ae3c2c4022
Test: as follows
- built (bullhead)
- flashed
- booted
- runtest frameworks-net passes
- manual USB tethering toggling between WiFi and mobile
Bug: 32163131
Change-Id: I57edf5114b6361f320577c7870e40f8b3cdf74ce
The state that needs to be transferred includes:
- NetworkCapabilities
- LinkProperties
- whether the network is currently suspended
Additionally:
- Rename notifyNetworkCallback() to notifyNetworkAvailable()
in order to clarify its real function.
- fix previous copy/paste error in unittest
Test: as follows
- built (bullhead)
- flashed
- booted
- runtest frameworks-net passes
- USB tethering with mobile and Wi-Fi upstream toggling
Bug: 32163131
Change-Id: Ib4460bcd5d08863a9feac9e8ab41a238897bb3ea
All the tests are now asynchronous. The only remaining uses of
Thread.sleep() are in the tests that check for NetworkRequest
timeouts.
Test: ConnectivityServiceTest passes 100 times in a row
Bug: 32561414
Change-Id: If420bd66c692a90d5031ee06a888a8cc3b4398a8
The calls in testAvoidBadWifi are no longer necessary now that
waitForIdle is reliable, and the calls in testPacketKeepalives
are replaced with a wait for the NetworkAgent to disconnect.
Test: ConnectivityServiceTest passes 100 times in a row.
Bug: 32561414
Change-Id: Icbb161ca6e343bd14764a1c9ccfdd14b6cd6803f
Rename the class and ensure that all code related to bad wifi
avoidance has names that reflect its purpose. This will allow us
to reuse most of the the code for other purposes in future CLs.
Test: runtest frameworks-net passes
Bug: 34630278
Change-Id: Ida190e5a79e7a12f4eec07e1a3e4d9d138e83bcd
The current implementation of IdleableHandlerThread is based on
the assumption that MessageQueue#isIdle will return true iff the
message loop has finished processing its messages.
Unfortunately, this is incorrect: isIdle returns true iff are no
more messages waiting in the queue; thus, it will also return
true while it is processing the last message before going idle.
Instead of using idle handlers, take the simpler approach of
posting a runnable to the message loop and waiting for it to
be processed.
Test: see other CLs in this patch series
Bug: 31479480
Change-Id: Iae75781f067b762c8653a488a5e4d5ee0c789e01
This patch removes testNotWaitingForIdleCausesRaceConditions() from
ConnectivityServiceTest because it is in nature flaky and prevents using
ConnectivityServiceTest as a patch presubmit check. Estimated failure
rate is 1/15 on Nexus 6p.
This patch also simplifies how ConnectivityServiceTest waits for
handlers to become idle by removing IdleableHandlerThread and turning it
into a signle static method.
Test: $ runtest frameworks-net
Bug: 31479480
Change-Id: I2d78709cbb61d5d01cd59cff326469417f73f1ab
This patch adjusts timeouts in ConnectivityServiceTest to fix tests
known to fail spruriously.
Test: ran the following tests 500 times without observing any failure
- testBackgroundNetworks()
- testMultipleLingering()
- testPacketKeepalives()
- testSatisfiedNetworkRequestDoesNotTriggerOnUnavailable()
- testSatisfiedThenLostNetworkRequestDoesNotTriggerOnUnavailable()
Bug: 32561414
Change-Id: I184fe7189aac768a65b927cc42eefaf8b1b3f930
- CallbackInfo becomes a static class so that it can be instantiated
and looked at more easily than before.
- CallbackInfo gains hashCode() because it has equals().
- effectively final field are qualified final
- fixes bug when polling the next callback: poll() on a LinkedBlockingQueue
returns null when timing out, which was not failing the test cleanly.
Test: ConnectivityServiceTest passes
Bug: 32561414
Change-Id: I64d382a35fe83f21bded98e6906c1ec70a8f1085
This activates all frameworks-test tests in runs of the continuous
platform tests.
Test: $ runtest frameworks-net passes (expect Tether
Bug: 32561414
(cherry picked from commit fa8a6f6220)
Change-Id: I7b0706a7e3368f971d508388e8ad4afc5de9d646
This patch introduces an assertEventuallyThat helper function in
ConnectivityServiceTest which given a boolean function retries until the
function returns true or until a maximum retry time is reached.
This function is used to fix flakyness of testAvoidBadWifiSetting where
the Message posted by reevaluate() could reach the Handler's
MessageQueue after waitForIdle takes effect, resulting in the test to
fail.
Instead of fixing the flakyness by introdcing hard sleep times,
assertEventuallyThat is used to reduce the overall test time.
With this change the test has been observed to pass with 100% success
rate over 50000 invocations.
Test: $ runtest frameworks-net
Bug: 32561414
(cherry picked from commit 99c8294ef0)
Change-Id: If41a24989d5f65aeb439f68741f511b5eb53b18d
This patch fixes flakyness of testRequestBenchmark by adjusting time
limit for callback registration from 100ms to 180ms, and time limits for
onAvailable and onLost triggers from 30ms to 40ms.
With these timeouts the test succeeds 100% over 5000 iterations.
When using 150ms for registration timeout, running the test 5000 times
fails 2 times.
When using 30ms for onLost timeout, running the test 5000 times fails
1 times.
In addition, this patch also cleans testRequestBenchmark and uses the
more stable SystemClock.elapsedRealtime() for duration measurements.
Test: $ runtest frameworks-net
Bug: 32561414
(cherry picked from commit 38be57b438)
Change-Id: I196ab9ef7f5abe456a783eed65db09279d2ecb8c
Guarantees that timeouts are only delivered if a network never
becomes available. Once a network is available the timeout is
canceled.
Bug: 31402633
Test: all timeout related unit tests pass (new one added)
(cherry picked from commit 681fcda589)
Change-Id: I7cd3086544c881915fc6dbf14b87a24ab0cd8748
This will give us a good place to put all the networking tests.
Fix: 31479480
Test: adb shell am instrument -w -e notClass com.android.server.connectivity.tethering.TetherInterfaceStateMachineTest 'com.android.frameworks.tests.net/android.support.test.runner.AndroidJUnitRunner' # PASS
(cherry picked from commit ecc986d430)
Change-Id: I993eeaa5dec001c39389023f355f506129b356e7
