mCurrentPackage is null when package doesn't exist, which lead test
failed with NPE
Use request.packageName for case BackupTransport.AGENT_UNKNOWN instead
Bug: 77272500
Test: m -j RunFrameworksServicesRoboTests ROBOTEST_FILTER=PerformBackupTaskTest
Change-Id: I27be85f0c3f95e36edd92d0d552189fbcf6952dd
(cherry picked from commit 87f1209ac6)
With the newly created Global setting for backup/restore agent timeouts
(backup_agent_timeout_parameters introduced in ag/3731401), update
references in backup and restore code to get the timeout values from
the setting instead of referencing constants in BMS.
This makes these timeouts configurable outside of the framework.
The default value of the setting is the same as the constants in BMS so
this should have no change on the timeout values used in the backup and
restore flow.
Bug: 70276070
Test: 1) m -j RunFrameworksServicesRoboTests
Change-Id: I0259bba76d0fed48158b8316f430b315ea98086e
Part of push to make backup and restore agent timeouts configurable. Creates
a Global setting for the current static BackupManagerService timeouts so
that they can be overriden with P/H. We keep the current default values,
which will be updated once we investigate what more appropriate values are.
Remame the constants to better reflect what they're used
for. Next, we will update the framework to use these constants.
This depends on the refactor of how we observe changes to key value
backup settings (ag/3709997).
Bug: 70276070
Test: m -j RunFrameworksServicesRoboTests ROBOTEST_FILTER=BackupAgentTimeoutParametersTest
Change-Id: Id506314ce0c8bd5e4d1d8b4001b26cbad0056c99
Extracts an abstract class to observe changes in backup parameter settings that
are stored as a comma-separated key value list. This class is
responsible for registering and unregistering a content observer on the
setting and updating local references to the parameters.
Refactor BackupManagerConstants and LocalTransportParameters to use this
implementation. This will also be used for the new backup timeout
setting.
Bug: 74346317
Test: 1) m -j RunFrameworksServicesRoboTests ROBOTEST_FILTER=BackupManagerConstantsTest
2) gts-tradefed run commandAndExit gts-dev -m GtsBackupHostTestCases -t com.google.android.gts.backup.TransportFlagsHostSideTest
Change-Id: Id4c50fbcf7479c925515887e3fa70e166dd9955c
Added units for http://ag/3709565. Some refactoring.
Test: m -j RunFrameworksServicesRoboTests
Bug: 72485465
Change-Id: Id75a4e0b96936580fd677041e091340b0fff8c1e
(cherry picked from commit cac3a74059)
Retrievable via 'adb shell dumpsys backup transportstats'.
Sample output:
Average connection time: 36.00 ms
Max connection time: 181 ms
Min connection time: 7 ms
Number of connections: 16
Per transport:
com.google.android.gms/.backup.BackupTransportService
Average connection time: 27.71 ms
Max connection time: 139 ms
Min connection time: 13 ms
Number of connections: 14
com.google.android.gms/.backup.component.D2dTransportService
Average connection time: 181.00 ms
Max connection time: 181 ms
Min connection time: 181 ms
Number of connections: 1
android/com.android.internal.backup.LocalTransportService
Average connection time: 7.00 ms
Max connection time: 7 ms
Min connection time: 7 ms
Number of connections: 1
Bug: 72485465
Test: Will follow in another CL if reviewers OK w/ approach.
Change-Id: I133ed423d0b8471d69e3c3631aadee7d42d0ec0e
(cherry picked from commit e5a976404c)
This change sets LOCAL_SDK_VERSION for all packages where
this is possible without breaking the build, and
LOCAL_PRIVATE_PLATFORM_APIS := true otherwise.
Setting one of these two will be made required soon, and this
is a change in preparation for that. Not setting LOCAL_SDK_VERSION
makes the app implicitly depend on the bootclasspath, which is
often not required. This change effectively makes depending on
private apis opt-in rather than opt-out.
Test: make relevant packages
Bug: 73535841
Exempt-From-Owner-Approval: Global cleanup
Change-Id: I26458e41ecb84de91ac9a356a5d4bafb44f463c1
Previously, the transport registration extra was a private
constant. Since GMSCore depends on this value being passed, moving
it to a public API prevents having to define it twice in
framework and GMSCore, and ensures compatibility between the two.
TODO: Update GMSCore with this constant once this drops
into GMSCore.
Bug: 72730566
Test: 1) m -j ROBOTEST_FILTER=TransportManagerTest RunFrameworksServicesRoboTests
2) m -j ROBOTEST_FILTER=TransportClientManagerTest RunFrameworksServicesRoboTests
Change-Id: I8f7a2ca0275047a5d3cc1a530cd86499d0170f2f
Around restoreAll() and restoreSome(). And some small refactorings in
restore code paths.
Test: m -j RunFrameworksServicesRoboTests
Change-Id: I0ff446ef4dcf15eade189c79e90a22c0f2eda0d6
Robolectruc upstreamed this and we picked up in 3.6.1 already :)
Test: m - j RunFrameworksServicesRoboTests
Change-Id: I9df1d1dfc56c5aa1d27e36396ddd21c053a3199d
for <include /> rules in <full-backup-content> specification
Give an app developer the option to include files based on the
transport flags exposed by the transport. This allows conditionally
including files as long as the transport identifies itself as
for instance encrypted or device-to-device.
Extend the parsing mechanism to read optional requiredFlags
attributes, and extend existing structures to encompass
that data for BackupAgent to retrieve and act on it
based on FullBackupDataOutput#getTransportFlags().
-- Changes in robotests/
The old version of this CL (that already got reverted) broke our Robolectric
suite because it added an inner class to FullBackup and a dependency on it from
BackupAgent. FullBackup wasn't being built from Android tree (instead it was in
a prebuilt Robolectric snapshot jar of the framework) but BackupAgent was,
which resulted in not finding the inner class.
So, also in this CL. Changing our tests to include everything under
platform/base/core/java/android/app/backup from Android tree.
`m -j RunFrameworksServicesRoboTests` is green now
Bug: 72484288
Test: m -j RunFrameworksServicesRoboTests
Test: runtest frameworks-core -c android.app.backup.FullBackupTest
Test: make cts -j40 && cts-tradefed run cts -m CtsBackupHostTestCases -t android.cts.backup.FullbackupRulesHostSideTest
Change-Id: Ideaed59f8337257aa6a882ff0ce80c170b17d55e
Added first set around getAvailableRestoreSets() plus some tidy-up
around setup and unused stubs.
Test: m -j RunFrameworksServicesRoboTests
Change-Id: I9b3bf548251f1262907de96407184bd1822b3429
When connecting to a transport, let the transport
know if this connection is for registration. This is to
prevent updating transport attributes during registration
in GMSCore.
Bug: 72730566
Test: 1) m -j ROBOTEST_FILTER=TransportManagerTest RunFrameworksServicesRoboTests
2) m -j ROBOTEST_FILTER=TransportClientManagerTest RunFrameworksServicesRoboTests
3) adb reboot; adb logcat | grep BackupTransportManager; check no errors with "...not registered tried to change description"
4) GMSCore Robo tests
Change-Id: I5adf6ea3e668a8e8ed8c568728d109814b6f8975
Around transport and agent failures. Caught a few possible errors, check
TODOs.
Test: m -j RunFrameworksServicesRoboTests
Change-Id: I9d04f4253511032473885ce6c0856c52639f2957
The interface will be removed in a subsequent CL, it will be
easier to deal with merge conflicts.
Test: 1) m -j RunFrameworksServicesRoboTests
2) runtest -p com.android.server.backup frameworks-services
Bug: 65823538
Change-Id: I37c4c6758c646e1b18889ed05aa1b1d7c6129cf0
Extracted connection in TransportClient to static class with weak
reference to TransportClient so that TransportClient is garbage
collected when left undisposed. Created finalize method that logs,
unbinds TransportClient and warns using CloseGuard.
Also adjusted some logging.
Bug: 17140907
Test: m -j RunFrameworksServicesRoboTests
Test: Manually left a dangling TransportClient and observed logs and
connections
Change-Id: I30f89e7c27579089ba29936483abd1b60c9e8e37
This allows a BackupAgent to check whether the transport has client-side
encryption enabled. It can then use this information to decide whether
to back up more sensitive data.
Bug: 72299360
Test: Manually verified full & kv backup agents receive transport flags
Test: m -j RunFrameworksServicesRoboTests
Change-Id: Ibd9b5f9479815e1721e9d6b7663d892b9ab3fcae
There was a deadlock around the transport lock. We registered transports
with the transport lock held, this kicked-off transport onCreate()
synchronously, which called TransportManager
updateTransportAttributes(), which tried to acquire mentioned lock but
couldn't. This CL removes the lock for any call to the transport or
operation that triggers a call to the transport (it was
TransportClient.connect() or its variants).
Test: Load GMSCore before fix, boot, register transports, check no ANR
Test: m -j RunFrameworksServicesRoboTests
Test: adb shell bmgr transport -c <transport>, being registered & not
Bug: 72147303
Change-Id: I72ca145d7fb73c0ef29c4aa1b620fea4969481db
Let the device owner make backups with a chosen backup
transport mandatory.
BUG: 64012357
Test: make RunFrameworksServicesRoboTests
Test: manually together with the corresponding GmsCore change.
Test: cts-tradefed run cts -m CtsDevicePolicyManagerTestCases --test
com.android.cts.devicepolicy.DeviceOwnerTest#testGetAndSetMandatoryBackupTransport
Test: cts-tradefed run cts -m CtsBackupHostTestCase --test
android.cts.backup.BackupDeviceOwnerHostSideTest#testMandatoryBackupTransport
Change-Id: I9bfae5799beae3459659e697813b75a9b508ae55
This CL is the last of the P1 CLs for binding on-demand. During
transport-usage migration to binding on-demand the permanent-bound
code was still there and we piggybacked on it to register the
transports. Now that everything is migrated we removed the permanent
binding and used the registration code created in the CL #9
(selectTransport) for registering all the transports.
I put a 3s delay on registration after bring-up. Any operation that uses
the transport before that will gracefully fail.
The TransportBoundListener does not return a boolean anymore because it
can't fail anymore (we pass it the data it needs so that it doesn't need
to be exposed to TransportNotRegistered exceptions). It's now called
OnTransportRegisteredListener.
Eligible transports were a similar thing to valid transports from the
permanent binding code, whose only need came from rebinding from what I
could tell. I saw no need for it anymore and removed it. If I missed
something please shout :)
I shuffled methods a bit in TransportManager.
There were a lot of changes to robo test infra to bring together
TransportManager tests and the rest and re-use mocking infra.
Change-Id: If61268228dd0bb724b718abd3dcafdad50e8b3dc
Ref: http://go/br-binding-on-demand
Bug: 17140907
Test: m -j RunFrameworksServicesRoboTest
Test: runtest -p com.android.server.backup frameworks-services
Test: gts-tradefed run commandAndExit gts-dev -m GtsBackupTestCases
Test: gts-tradefed run commandAndExit gts-dev -m GtsBackupHostTestCases
Test: cts-tradefed run commandAndExit cts-dev -m CtsBackupTestCases
Test: adb shell bmgr enable true/false
Test: adb shell bmgr list transports [-c]
Test: adb shell bmgr transport <transport_name>/-c <transport_component>
Test: adb shell bmgr restore <token>/<token> <package>/<package>
Test: adb shell bmgr backup <pacakge>
Test: adb shell bmgr run
Test: adb shell bmgr wipe <transport> <package>
Test: adb shell bmgr fullbackup <package>
Test: adb shell bmgr backupnow --all/<packages>
Test: adb shell cmd jobscheduler run -f android <job_id>
Test: adb shell dumpsys backup
Test: D2D scenario
After Robo 3.5.1 upgrade, these tests should support the new SDK.
Bug: 69153972
Test: m -j RunFrameworkServicesRoboTests
Change-Id: I12be8668b0c8bb7933cc1a30717356d86f1f3737
Migrate selectBackupTransport()/selectBackupTransportAsync() to binding
on-demand. To mimic the bind-if-needed behavior that existed before for
selectBackupTransportAsync we now register-if-needed in the selection.
This means a new registerTransport() method was created in the
TransportManager. I intend to use this method with only few
modifications for the first-binding code.
Change-Id: I39661cff0f7b2f8a27da37905dcd93e0aa9b1178
Ref: http://go/br-binding-on-demand
Bug: 17140907
Test: m -j RunFrameworksServicesRoboTest
Test: adb shell bmgr transport <transport>, observe logs
Test: Robolectric tests for TransportManager will go in registration CL
BackupManagerConstants has a content observer that is meant to be
notified whenever backup settings change. However, that observer is
never registered.
Bug: 71622864
Test: cts-tradefed run cts-dev --module CtsBackupHostTestCase
Change-Id: Icbd90ef3af83dd2f29e26a5d787505bcf5681154
To be able to re-use the TransportClient infra for transport
registration, I need to remove transport dir name property from
TransportClient because it's not available before registration
itself. As a result callsites that used getTransportDirName()
from TransportClient will have to go through the
TransportManager for that. Bryan suggested that the
TransportClient wasn't the best place for the property before.
Ref: http://go/br-binding-on-demand
Bug: 17140907
Test: m -j RunFrameworksServicesTests
Change-Id: I3fa335faf97d63adfad1a929336073a70fc8bc02
Migrate usages of the transport binder to binding on-demand:
* getDestinationString()
* isAppEligibleForBackup()
* dump()
For getDestinationString() we'll be introducing an invisible bug for
people that haven't updated GMSCore to include the usage of
updateTransportAttributes() API introduced in earlier CL. The bug is
that that text won't change, it'll remain constant. It's invisible
because currently only place that uses that method is Settings in some
circumstances that depend on the transport, and those circunstances
don't happen with our transports. Check http://ag/1831025.
For isAppEligibleForBackup(), a new filterAppsEligibleForBackup() is
created and there we bind on-demand.
Change-Id: Idc9e31f0e8eda8531e204c05a84fafdaf0247d08
Ref: http://go/br-binding-on-demand
Bug: 17140907
Test: adb shell dumpsys backup, observe destination of transports
Test: adb shell bmgr backupnow --all, observe only eligible apps got backed-up
Test: Force-loaded settings screen and observed destination string
Test: m -j RunFrameworksServicesRoboTests
