This patch adds a zygote policy parameter to functions in the
ActivityManagerService (and associated classes) to allow for different
launch behavior based on the triggering event. This will allow for
latency sensitive applications to utilize the Unspecialized App Process
Pool, without other processes draining the pool and causing excessive
refill events.
Test: Booted and launched several apps
Bug: 147613193
Bug: 146614089
Change-Id: Ib0a51c8720f95a7eafcab4ad2b5bb08f27a58d8c
Merged-In: Ib0a51c8720f95a7eafcab4ad2b5bb08f27a58d8c
(cherry picked from commit 4d05188588)
In common cases, to resume the next activity we need to wait for the
current one to be paused. Since starting a process for activity is
asynchronous, if we already know the process of next activity has not
started yet, we can start the process earlier so the time waiting for
the pause to complete can be saved.
Also if the launching activity is going to be the top app, we can set
the top schedule group right after its process is started so the start
time before actually launching the activity can be improved.
Although before the current activity is paused, the next top activity
may still change and results some empty processes. That should not be
a common case and the process is still useful when going back the stack,
and the empty background processes are easier to be reclaimed.
Bug: 123043091
Test: AppLaunchTest
Test: Launch calculator from launcher, the event log am_proc_start will
show "pre-top-activity".
Test: Cold launch a top activity, the system log should not show
"not expected top priority".
Test: Use startActivities to start serveral activities in a sequence.
Check "adb shell cat /proc/$pid/task/$pid/cgroup" for each process.
Only the last one has top-app, others are background.
Change-Id: I9601b66e7cc0855fd7c2b573ded31fcf8d0711ae
Merged-In: I9601b66e7cc0855fd7c2b573ded31fcf8d0711ae
IUpdateEngine.cleanupSuccessfulUpdate is now asynchronous with a callback.
Update UpdateEngine.cleanupAppliedPayload accordingly to use the new impl.
Bug: 147696014
Test: apply OTA
Change-Id: I9b308316dcb8e6a507063d3060fed7693da82a19
Use DEFAULT_USERDATA_SIZE as userdata partition size if the calling
Intent desn't specify the userdata size or the specified size is zero.
Bug: 145891687
Test: adb shell am start-activity \
-n com.android.dynsystem/com.android.dynsystem.VerificationActivity \
-a android.os.image.action.START_INSTALL \
-d file:///storage/emulated/0/Download/aosp_arm64-dsu_test.zip
Change-Id: I52767dfefef394403c5e16fcfd40792b14ff9a71
If both BaseBundles are empty, we can infer that without needing to
unparcel any of them.
Test: atest FrameworksCoreTests:android.os.BundleTest
Bug: 146037505
Change-Id: I04c28cdd1293227d9887b0c17e178f61328c1959
Merged-In: I04c28cdd1293227d9887b0c17e178f61328c1959
We now use the sequence number of the atrace tag rather than a Binder
transaction, so zygote can observe changes to the tags.
Test: m
Bug: 137366208
Bug: 147667830
This is a cherry-pick of e38797be84.
Change-Id: I50e32a19c538dd255097c035836504b4746b172b
Merged-In: I50e32a19c538dd255097c035836504b4746b172b
- Get all packages for system user
- Reuse split system user's logic to blacklist packages when headless system
user is enabled
- Add new method in UserManager to get headless sytem user mode (same as
in master)
Bug: 145626101
Test: edit device's sysconfig file, use tag: system-user-blacklisted-app to blacklist app
make services && adb sync system && adb reboot
cts test
Change-Id: I98d1bc33e7dd59ffa3ac6426f95af708671138da
(cherry picked from commit 9c4cf4945d57d45e429bc7d33bf6ffa68eb0838c)
This was previously reserved for keystore only but since credstore
also needs to do attestations, this is needed.
Bug: 111446262
Test: atest android.security.identity.cts
Test: Manually verifying the AttestationApplicationId from credstore
Change-Id: Ie44f9e4c8f2e1bd916ccbe7c7e5537dc498d8154
Provide a mechanism to extend and customize a binder interface without
having to modify the original (base) interface.
See the full motivation here: aosp/1099664
Bug: 147834064
Test: atest FrameworksCoreTests:BinderTest FrameworksCoreTests:BinderProxyTest
Change-Id: Idd3595b1302cf7ac0590d064e795ba0b75811af6
By default, PR_SET_DUMPABLE is 0 for zygote spawned apps, except in the
following circumstances:
1. ro.debuggable=1 (global debuggable enabled, i.e., userdebug or eng builds).
2. android:debuggable="true" in the manifest for an individual application.
3. An app which explicitly calls prctl(PR_SET_DUMPABLE, 1).
4. GraphicsEnv calls prctl(PR_SET_DUMPABLE, 1) in the presence of
<meta-data android:name="com.android.graphics.injectLayers.enable"
android:value="true"/>
in the application manifest.
So checking both ro.debuggable=1 and PR_GET_DUMPABLE is redundant.
Bug: b/144186877, b/148566223
Test: CtsAngleIntegrationHostTestCases
Test: CtsRootlessGpuDebugHostTest
Change-Id: Ica49254df2c7c090808411935cdeb8efd4e3cb51
Merged-In: Ica49254df2c7c090808411935cdeb8efd4e3cb51
(cherry picked from commit 097a3062b9)
Add @WorkerThread to allocateSpace and cleanupAppliedPayload
because they take a non-trivial amount of time.
Bug: 148486809
Test: builds
Change-Id: Ib7daba9c662b9e141430df71401d069577221b8a
