This is a preparation CL that allows followup CLs to add network
metric monitors. This CL does a pure refactoring without
introducing any behavior changes

Specific changes include:
- UnderlyingNetworkEvaluator now encapsulates the calculation of
  mPriorityClass
- mIsSelected is moved from UnderlyingNetworkRecord to
  UnderlyingNetworkEvaluator
- UnderlyingNetworkListener is simplified by moving the network
  candidate out to the UnderlyingNetworkController
- NetworkPriorityClassifier now takes a "isSelected" boolean
  instead of the currently selected network candidate

Bug: 282996138
Bug: 315858972
Test: atest FrameworksVcnTests, CtsVcnTestCases
Change-Id: I8461a68f1c4260b253d84d3efed2b09c41116656

Test: TH
Change-Id: I6760e1da0ceb6d985bdbbf606a334de4c334b2c9

The `--api-lint-ignore-prefix` option adds a package prefix within
which API lint checks will be ignored. However, the options are
not actually necessary. That is because they are only used in the
`android-non-updatable-stubs-defaults` and the non updatable code does
not include any source classes from those packages.

While modules that do use `android-non-updatable-stubs-defaults` do
include classes from those packages on the classpath they are already
excluded from having API lint checks performed on them as they do not
contribute to the API surface being generated by those modules.

Bug: 311690008
Test: m checkapi
Change-Id: Ife750c586edc2b5af92c0b5e9211b23345083f69

Due to a bug in Metalava the `VisiblySynchronized` check would report
an erroneous duplicate baseline entry that referenced some internal
implementation detail, e.g. `PsiThisExpression` instead of an actual
API element. Another entry with the same message and correct API
element was also written.

The Metalava bug has been fixed so this change removes the erroneous
entry.

Bug: 311690008
Test: m checkapi
Merged-In: If5550b8cc0a97f02fb7fb1bd949f8e5517007d4c
Change-Id: I79b3c9802e40e74ba7f0c9b102dc201d01ddeab5

Fixed a typo in an OWNERS file.  The ".java" extension was wrong
because the jni directory only has native code.  The modification
assigns the performance team as owner of all AnrTimer files.

Bug: 282428924
Change-Id: Ib6e053a3b3e0f61f04de18be4ac7dc8158819710

Bug: 316850504

Change-Id: Ie6655bb6634f1a20e9e5c320fd137fd73dab5f16
Signed-off-by: yangjing24 <yangjing24@xiaomi.com>

Test: Presubmit only
Bug: 297373084
Merged-In: I499b6a99e53775abf084f741802463940730091a
Change-Id: I499b6a99e53775abf084f741802463940730091a

Test: automated
Bug: 117306409
Merged-In: If3a3de2325de0c31ebe92ab41631869008425728
Change-Id: If3a3de2325de0c31ebe92ab41631869008425728

Bug: 309649371
Test: N/A
Change-Id: I4adff5cd92042208fa2d695423fdfa0072f916b2

This exemption should have been disabled for Android U+

Bug: 313715896
Test: dumpsys activity
Change-Id: I469e16a8b3b79deaa7ef2f978385efd9df1f145b

Test: N/A
Change-Id: I99744ca971281130dbe57626882ac10e382ea949

We get displayPowerController object from a SparseArray so it would be
better to guard by a lock object or add a null check.

Bug: 315308333
Test: atest, manual
Change-Id: Ic6713d85aa09e660c9dc14f01ac05c4d394922fd

Revert submission 2843682-tombstone-proto

Reason for revert: Android Go build break, b/316637921

Reverted changes: /q/submissionid:2843682-tombstone-proto

Change-Id: Ic9d7b73e97c070fc378becb59799268323d9cead

Merge changes from topic "revert-2870060-revert-2866127-HwNoService-EIABNUAXKH-SMYTAJDJZM" into main

* changes:
  Revert "Revert "Create a HwNoService class that fakes hwservicem..."
  Revert^2 "Add isHidlSupported Java API"

Lint now forces you to explicitly set an explicit
display name for composite quick fixes (it used to
just default to the nested quickfix' display name,
but this is sometimes misleading so now lint checks
need to explicitly do this.)

Test: SimpleManualPermissionEnforcementDetectorTest.kt
Bug: Not filed
Change-Id: Ie46d32833efc7bd9f7051acf6b45327d4fda7927

This reverts commit 4a11b111.

This change was originally reverted due to a bug in the child commit of
a change in the same topic which caused process group kills not to occur
when using cgroup.kill. Now that has been fixed, bring back this change
with the fix.

Bug: 301871933
Change-Id: I81dde53f21ba2e01c01dfd5f7009ea1ef690b582