Processing the large proto tombstones can cause OOM and there is no point adding tombstones larger than dropbox max capacity as they will not get uploaded.


Bug: 343135528
Change-Id: Ib308e48dcfc3282ae7ab0e9fa0827cc8fae97078

TracingServiceProxy depends on external system services to forward its
requests to. There is an edge case where reportTrace was called before
the encrypted storage was unlocked and before external services can be
bound (e.g. occurs when reportTrace is concurrent to system_server
starting up).

This change defers the registration of the service until we are ready to
accept requests. As a side-effect this also means that the device needs
to be unlocked before the service is published.

Bug: 333846370
Test: atest PerfettoReporterTest
Test: boot device without lock and invoke perfetto reporting (verify the report is accepted immediately)
Test: boot device with lock and invoke perfetto reporting (verify the report is accepted after unlocking)
Test: invoke perfetto reporting concurrent to kill -9 `pidof system_server`
Change-Id: Ibb75c3e314d30b18cdc32e3774be1ee8022960b1

Increase the dropbox disk quota to 20Mb for user debug builds as we add more debug to dropbox for user debug.
(cherry picked from https://googleplex-android-review.googlesource.com/q/commit:ea261f6a2a9d1034abfaf2ebdaae72c9869c8c90)
Merged-In: Ib758fa6fa47cfd26e895bf8d7b9d5bde98c99461
Change-Id: Ib758fa6fa47cfd26e895bf8d7b9d5bde98c99461

There's some bad behaviour going on that's exclusive to WatchOS
(discovered on the Pixel Watch 2). The TombstoneWatcher class is
responsible for watching the /data/tombstones/ directory. It is a member
of the NativeTombstoneManager class, which is created when
ActivityManager starts, and lives until the device is powered off.

Unfortunately, some bad behaviour (unclear exactly what at this stage)
means that the TombstoneWatcher class is spuriously garbage collected,
even though the parent NativeTombstoneMenager class is still alive and
reachable. I chatted with hboehm@ from the ART team, and his hypothesis
is that because the TombstoneWatcher member is initialized in the
constructor and never used again, some optimiser is doing the wrong
thing and thinking that it's unused and can be GC'd.

The TombstoneWatcher uses inotify in native code to execute a callback
when a file is added to a directory, so it's clearly in use, but
unfortunately the bug allows this class to be GC'd anyway.

What this means, in practice, is that app native crashes on WatchOS are
not being handled correctly. This includes:
 - Apps cannot get their own crashes through the
   ApplicationExitInfo::getHistoricalProcessExitReasons API.
 - Crash reports are never sent to Google, with the exception of
   tombstones that last until the next reboot, where they're one-time
   scooped up and sent to Google. This has substantially increased
   latency (requires a reboot) and also probably drops quite a number of
   reports because there's a limit of tombstones on the device, and a
   rate limit of how many are sent by GMS.

Bug: 339371242
Test: atest CtsGwpAsanTestCases on eos (Pixel Watch 2, WatchOS)
Change-Id: I9226e4368b03bd4742fccdafde6018f145da63e6

method.

Accessing mDisplayInfos, mInputWindowHandlesOnDisplays and other
collections should add mLock lock according to AOSP requirements.

Test: OEM monkey test
Bug: 344434136
Change-Id: Icb18683680c40939e0ebd352818702f90b79179f

If a display has freeform windowing environments such as Desktop
or XR device, then resumed activity of task fragment will be set
to focused app. However, it can break focused state in embedded
task fragments.

Thus, allow it only if a focused task is different from a next task.

Bug: 343252725
Test: Settings(AE) > Launch SubSettings > Check focused app/window
Change-Id: Ib62ecb7c8e0fe453149f51c610c34d97bbde7cb1

Test: atest
Bug: 343818003
Change-Id: Ia30d818d6ac19c2a3e9ae31b29122fb5bbac338a

Bug: 324085891
Test: atest CtsResumeOnRebootHostTestCases \
FrameworksServicesTests: RebootEscrowDataTest \
FrameworksServicesTests: LockSettingsServiceTests \
FrameworksServicesTests: RecoverySystemServiceTest \
FrameworksServicesTests: RebootEscrowManagerTests \
FrameworksServicesTests: RebootEscrowProviderServerBasedImplTests
Change-Id: Ibba55830bb3168d312d3f92094d7a266931439e6

3bebb903

Change-Id: I605177fabae80ea68399aa42a66419119cf02cd3
Bug: 343583446
Bug: 343690533

rollbackId and originalSessionId info is important for logcat based
debug about session installation with rollback like Mainline.

Bug: 343309166
Test: manual (check RollbackManager log at mainline update)

Change-Id: I20755a4fa8a99b07c4ba12e735af19dba21fcdda

The help/usage text for the notification cmd is not wrapped
correctly when displaying the 'allow_dnd' parameter.

This was introduced in commit 08d381b8.

Test: manual, run cmd notification and check the wrapping
Change-Id: I9a7af1d9e3de91c9124f21eb66817bb73a3512be

This reverts commit ec7f7207.

Reason for revert: Fixed tests

Change-Id: Iadd0ea6f09f7f3159f94489c03f28c326bbbb664

Bug: 342527967
Change-Id: I41811e2dae454c67435c1c99d10bc3a28894fbd1
Test: presubmit

This reduces the latency before we are able to generate a profile from
the server backend.

Test: cmd jobscheduler run -f android 260817
Change-Id: Icdbdf809435580b131f5b6677e039f3043264abe

Test: OEM monkey test
Bug: 341658103
Change-Id: Idfc3db1c0ea3be536cad82bf98fab6d668d26a36

This reverts commit 831334bd.

Reason for revert: Droidmonitor triggered revert due to b/341375758.

Change-Id: I9aaf3e0bca1d173ea7f449339185db5ace780626

For java ANRs and kernel crashes, adding `PageSize: 16384` if device
is using 16KB Page size.

Test: m
Bug 340959006

Change-Id: I4d08ca13ef5cd2e3a807cde2c11452c5e0ccd9a5

Bug: None
Test: presubmit
Change-Id: Iaf8ed5e440b5514c68d09e803a6fd6a6a0944245

In case of crash loops with a short time mainline rollbacks don't finish before the next mitigation which is warm reboot.

Change-Id: I77a487cb27e4254865ba2cc045e379dcb7348994
Test: unit and manual tests
Bug: 310236690

Send the confirm credential intent from the right user.

Bug: 327350831
Test: none
Change-Id: Ife9f2bed7a2095d2aac87a79aafe7868dce3591b

  When the hostsPackageForUser method return true,No nonNull judgment is made on the provider var which may be a null var.

Bug: 339747311

Merged-In: I00f8c12b5dec00eac6e50e2ef755742db3fe4521
Change-Id: I00f8c12b5dec00eac6e50e2ef755742db3fe4521

The root task used for split screen contains two or more sub-tasks. In this case, the activity should not be allowed to be started under the root task, which will cause a null pointer and cause the device to restart. Therefore, when the activity is started by specifying the taskid, check that the task is leaftask.

Change-Id: I095feb549ce2d3a4e61d999a342f64bea8264288

Since the underlying code in connectivity manager is moving to metered
firewalls to denote the uid policy for metered restrictions, NPMS can
start using them now.
Rather than having a single PENALTY_BOX, we have separate firewalls for
user and admin imposed restrictions. Once NPMS uses the appropriate
firewalls for these, it does not need to separately convey the blocked
reasons for metered restrictions on uids to the ConnectivityManager.

Flag: com.android.server.net.use_metered_firewall_chains

Test: atest FrameworksServicesTests:NetworkManagementServiceTest
Test: atest FrameworksServicesTests:NetworkPolicyManagerTest
Test: atest CtsHostsideNetworkPolicyTests

Bug: 336693007
Change-Id: I8da11ff663257d3dafcb5e22ae3313a9255ef0a6
Merged-In: I8da11ff663257d3dafcb5e22ae3313a9255ef0a6

This CL is part of the go/kddm project, to deprecate DDM protocol.
HELO packet currently return the VM name and version, this CL allows
to retrieves the same invo from a tool accessible from shell.

The data is retrieved the same way it is retrieved in DdmHandleHello.
handleHELO (System.getproperty).

Test: NA
Bug: go/kddm

Change-Id: Id5cf9bf768f397bf3595e577f0da7a001b7ec394

We found that if the system application misconfigures the gid in the permissions, -1 will be added to the application's gid list, which will cause the application process or system startup to fail, so consider adding a check to prevent this unexpected situation.

Change-Id: Ic3133067285ecfd9310b4bca56ff402180e7c2aa

WallpaperUpdateReceiver unconditionally updates the wallpaper
to the default_wallpaper when DEVICE_CUSTOMIZATION_READY is
received.

This overwrites live wallpapers and user set wallpapers.
Just clear the wallpaper when there is no live wallpaper to
avoid this from happening.

Test: set a wallpaper and run
adb shell am broadcast \
  -a android.intent.action.DEVICE_CUSTOMIZATION_READY \
  -f 0x01000000

Bug: 159673283
Change-Id: I385c28be84b9146e8620910258495f5d71b9c04e

mLatch.notifyAll() will not help wake the thread which is blocked by
mLatch.await, instead, it will cause system server crash.

So, mark state as failed if any package is failed to be deleted, and let
mLatch.countDown work until it comes to 0 and wake the thread blocked by
await.

Bug: 333845008
Change-Id: Ida1c3c128ddef376aaaacda1358717d0d4e9ca8b

When a device (such as a projection device) is connected, the spatializer available status should be updated regardless of whether this feature is turned on or not. Because other functions (such as the spatializer UI switch) need to change according to the spatializer available status.

Change-Id: I43d6d5d0ea4d69a0d4559390dcf34885b96bc488
Signed-off-by: xialei6 <xialei6@xiaomi.com>

Bug: 332981866
Test: atest com.android.server.audio.MusicFxHelperTest
Flag: NONE, impact of this fix is limited
Change-Id: I996951c7f71678c76afec78efd452b275f17213e
(cherry picked from commit 52c9b652)
Merged-In: I996951c7f71678c76afec78efd452b275f17213e

The apk file which was pinned should be add to the mPinnedFiles,
if not, the unpinApp can not found the PinnedFile from mPinnedFiles,
will not do the unpinFile for the pinned apk file.

Change-Id: I74f30650a590e746de2a98e8f9c5535ee2a9d37e

Test: presubmit
Change-Id: I4472fcf8140a8a9bceaff6b3c5a7abf6bc8df69e

Noticed that Rollback reason during boot loop was getting logged and
unknown because of a missing reason to string map. Updated that.

Test: manual
Bug: 291137901
Change-Id: I2ea6206d07099a413aeeca464e0d1ebd97c04840

1.If an app play media but the usage of attributes is unknown, musicActivecheck should be started.
2.USAGE_UNKNOWN default value is MEDIA.
3.Otherwise, the mMusicActiveMs will not increase for safe volume when above safe volume.
4.Users listen to music until 20 hours later but the volume does not drop.

Change-Id: I5f84b2c6376b283376d65d398e880a8716c98f59
Signed-off-by: liutianxiang <liutianxiang@xiaomi.com>

Package name is more commonly used than appid, so we hope to support obtaining permission and appop status based on package name.

Change-Id: I4f0e9bcb979132d99222cadaf306034bf58e092a

Bug: 335982533
Change-Id: I975641f193917101645cfcf2b78e198801c59a2c

Disable flags in old test, these could be removed during flag cleanup

Bug: 333720513
Test: atest RescuePartyTest
Change-Id: If659c702e3348c1c814f922f55d97776b99f2ff8

AppDataHelper:fix ArrayIndexOutOfBoundsException happen in fixAppsDataOnBoot method due to forget add lock.

   from the code:
    @GuardedBy("mPm.mInstallLock")
    void reconcileAppsDataLI(String volumeUuid, int userId, @StorageManager.StorageFlags int flags,
            boolean migrateAppData) {
        reconcileAppsDataLI(volumeUuid, userId, flags, migrateAppData, false /* onlyCoreApps */);
    }
    we should call the  reconcileAppsDataLI method by mPm.mInstallLock.

Bug: 338600619
Change-Id: I05f8aa84d71d5083c27427e3a9957b456ac57ce4

(cherry picked from https://googleplex-android-review.googlesource.com/q/commit:0ddb53602dee5c5b1d3408e7bda7c09743bda23a)
Merged-In: I44ff4c3f3a68cfbfe85cbb7017a6291f0c5ff9d8
Change-Id: I44ff4c3f3a68cfbfe85cbb7017a6291f0c5ff9d8

Require VCN to be used by a "main user" instead of a "system user"
since "system user" will no longer be a full user on HSUM devices

Bug: 310310661
Test: FrameworksVcnTests, CtsVcnTestCases on non-hsum and hsum devices
Flag: android.net.vcn.enforce_main_user
Change-Id: I5abfa7ba82055cd953083b1ab3f83e61bafd8a55

Audioservice register content observer for MASTER_BALANCE/MONO at the time of initialization/boot up. Registration by default is only applicable for current user which in this case is primary user.
So even when audio balance/mono changes for secondary user, the callback
won't be received in audioservice and hence the settings update won't
have any effect on the audio output.

Bug: 339131898
Test: manual
- Validated that audio output balance and mono varies based on the settings in the secondary user
- atest AudioManagerTest
- atest AudioServiceTest

Change-Id: Ifce17d23e0dd0af61d6e27de7c7da629e9480e77