Add current perf team members shayba@ and jdduke@, and remove
former members jpakaravoor@, balejs@, and cfijalkovich@.

Test: N/A
Change-Id: If786fa38bc5f86d9e64b1672bdcf5e8cde178aec

Skip using the gainmap shader when drawing to a bitmap
with a known SDR colorspace as it'll be a no-op in
rendering results since W=1.0, but it'll be very expensive
in runtime due to CPU evaluation of the SKSL shader.

Before:
createScaledBitmapWithGainmap_median (ns): 179,455,037

After:
createScaledBitmapWithGainmap_median (ns): 15,643,742

Bug: 311085927
Test: benchmark in CL
Change-Id: I3f2506a2bd86bbe29c6e0837d31e4a231a7c182a

Bug: 311064136
Test: m RUN_ERROR_PRONE=true javac-check
Change-Id: Ia6e4162a4244793f772dc880b73be0bc30f26d8b

Fix kotlin nullable errors that were exposed by setting the retention
of android.annotation.NonNull and android.annotation.Nullable to
class retention.

Bug: 294110802
Test: builds
(cherry picked from https://googleplex-android-review.googlesource.com/q/commit:2522501d5b4b4eab8d20986aa1c4b6ed36640211)
Merged-In: Icfec5d0e20b534d7af9aeb1b539a711db76d5e5c
Change-Id: Icfec5d0e20b534d7af9aeb1b539a711db76d5e5c

See https://source.android.com/setup/contribute/respectful-code for reference

For this round, the fixes are only applied to the following to minimize breaking dependencies:
  * comments (excluding javaDoc annotations)
	* private constants
	* private functions
	* parameters within functions

BYPASS_INCLUSIVE_LANGUAGE_REASON=Just updating a few select inclusive language violations.
No-Typo-Check: Changes focused on inclusive language violations.
BUG: 295342157
Change-Id: I1fe31787cc020685f7bc736a217d867844dc6591

Fix kotlin nullable errors that were exposed by setting the retention
of android.annotation.NonNull and android.annotation.Nullable to
class retention.

Bug: 294110802
Test: builds
Change-Id: Icfec5d0e20b534d7af9aeb1b539a711db76d5e5c

Bug: 287081404
Test: atest SurfaceFlingerPerfTests:SurfaceFlingerPerfTest
Change-Id: I8a2e1552956368e721a5b0be69911a569522d90f

Bug: 290757184
Change-Id: Ie897f0d6a8f121d512f2d04aeb213bde6622e73c
Test: N/A

Test: atest ExpensiveObjectsPerfTest
Bug: 284111560
Bug: 286219615
Change-Id: I6e35b65311c970e2c5ed40f5f25a9e771c35e9a2

Bug: 287081404
Test: atest SurfaceFlingerPerfTest

Change-Id: Ib574db0841f2b3dd0bcaa2acdba12c25abd0ca08

Change-Id: I0eb46cddd5b0dfc7e04a04cca9d3da0fb9b8c70b
Test: No behavior to test

Recently, flushing the broadcast loopers behavior, when
"wait-for-broadcast-barrier" command is invoked, is made
optional which resulted in some regressions in multi-user
tests as they same to be relying on flushing the broadcast
loopers behavior. So, explicitly add this option when
invoking the command.

Bug: 282848750
Test: atest android.multiuser.UserLifecycleTests
Change-Id: I6ba4e0858a63653b731c7a24b600ce9af758d39c

Bug: 281807379
Test: atest ExpensiveObjectsPerfTest
Change-Id: Ic4be74dc28c32fcc932d603c003a04e6f7ccb0f1

Calling the API was failing with ActivityManager.START_ABORTED error
code. This CL fixes that issue by starting the test apk via abd shell
command.

Bug: 279403644
Test: atest UserLifecycleTests
Change-Id: I86d0c0f8ccd0f17833ad418c29be19489a2fdc9f

If the user to be started is a profile of a current foreground user,
it should be started via IActivityManager.startProfileWithListener API
instead of IActivityManager.startUserInBackgroundWithListener. This
logic is already implemented in ActivityManagerShellCommand, so rather
than duplicating it, we'll be doing it by running the shell command.

Bug: 279403644
Test: atest UserLifecycleTests
Change-Id: I4b9e6f1e565187e93e13ad94f41d51ee5f7e3e42

Bug: 274491131
Test: atest UserLifecycleTests
Change-Id: I9cfa39f119be60dae98cfab12bfe3f261dec83c0

Bug: 274491131
Test: atest UserLifecycleTests
Change-Id: I2eff4749f562e4e8e733cd991e70b8040c745906

Bug: 274491131
Test: atest UserLifecycleTests
Change-Id: I2401641e5a7afb415cb5d3b0ba7548f8fad71cc4

Instead of freezing the screen during the user switch, with this CL
we are introducing a full screen user switcher, which is showing
target user's user name and profile photo with an animated spinner
around it.

Steps of a user switch:
1. 300ms - dialog show animation
2. 500ms - spinner animation around profile picture
3. Do the actual user switch
4. 300ms - dialog dismiss animation
5. End user switch (call UserSwitchObservers.onUserSwitchComplete, and
   send ACTION_USER_SWITCHED broadcast)

Changes:
* Step1 was already there.
* Step2 is added between 1-3, which shows a nice smooth progress
  animation around profile picture, but increases the user switch
  duration by 500ms.
* Step5 and Step4 were running at the same time before, now Step5 is
  postponed after the completion of Step4. Otherwise, dismiss
  animation becomes jerky. It was also jerky with unfreezing the
  screen, now that jerkiness is gone.
* All animations are disabled for slower devices.
  see: ActivityManager.isLowRamDeviceStatic()

Results:
* We're increasing the user switch duration by 800ms. (Step2 + Step4)
* We've a full screen user switcher with smooth animations.
* There is no more jerkiness in any stage (showing, during, hiding) of
  the dialog.

Notes:
* Our intention was to run Step2 and Step3 simultaneously but it makes
  the spinner animation jerky.
* We're disabling the dialog show/hide and spinner animations when
  running UserLifecycleTests.

Bug: 269237775
Bug: 223090747
Test: atest FrameworksServicesTests:UserControllerTest
Test: atest UserLifecycleTests
Change-Id: I5e0132e19c8da25438c5dbfecdeddf475b18f7d4

* Remove unused dependencies.
* Replace "ub-uiautomator" (v2.2.0) with
  "androidx.test.uiautomator_uiautomator" (v2.3.0).

Bug: 235842600
Test: presubmit
Change-Id: I8ed79b90770a770b9fd772b9b1fbde71a7006fd6

Bug: 269138674
Test: Presubmit
Change-Id: Idf1c0703e619ab8d2aa9c8c9a32ffcc35327738e

Test: Presubmits
Bug: 251818659
Change-Id: Id068a3faeec9790514487f5956af8cc2f18efa26

Bug: 269138674
Test: Presubmit
Change-Id: Ibd1a18f6369396c4ae7b0263baab84f25c88ce06

Test: JSScriptEnginePerfTests
Bug: 272517727
Change-Id: I5ee2919a083fd08bbccd88715f2e513edfe43495

In case killBaselineIme() will result the system busy on doing kill
process stuff and easier happen the locking contention in WM/ATM side
when measuring show IME with cold-lunch activity performance,

Make sure to start atrace after kill IME process being settled down
and dumping the trace for each iteration.

With this CL, following metrics can more accurate in reality without
being affected by unexptected delay:
- testShowImeCold_cv: 70 -> 10-20
- startInputOrWindowGainedFocus_mean: 6-8ms -> 2.8-4ms

Bug: 266708619
Test: atest ImePerfTest#testShowImeCold
Change-Id: Ibd004a36759e557c90bd6f156af3729d39f7721f

Before this CL, the MotionPredictor API was inconvenient to use. The
MotionPredictor.predict returned a list of MotionEvents, with one
MotionEvent per-device.

The goal was to force applications to consider multi-device streams.
However, even if the apps use this API, there's no way they can
currently test this behaviour, since the multi-device feature is not yet
available.

For the multi-device streams feature, the current plan is to provide a
new View callback to get these raw streams. In those streams, the
per-device MotionEvents would continue to look the same. However, the
events may now be interleaved between different devices. For example, a
sequence like this would be possible: DOWN(deviceId=2) ->
DOWN(deviceId=3) -> MOVE(deviceId=2).

That means that the app will likely have to do per-device bookkeeping
anyways. So the app might as well just create a per-device
MotionPredictor object, as well.

Bug: 167946763
Test: (cd frameworks/native/services/inputflinger && atest)
Change-Id: I34b8dc02d4f995146cb3eed33888ae34abde29d6

This is to track performance differences before and after the
cache optimization for null settings.

BUG: 228619157
Test: atest android.provider.SettingsProviderPerfTest
Change-Id: I04590d4d692d12970aa6d53789770fbe426f2c0e

Revert submission 20961419-deviceconfig-restrict

Reason for revert: CTS breakage

Reverted changes: /q/submissionid:20961419-deviceconfig-restrict

Change-Id: Ib6d4bac403e26a60401bebbb8c69f2abce6af6fe
(cherry picked from commit 158097fe)
Merged-In: Ib6d4bac403e26a60401bebbb8c69f2abce6af6fe

Revert submission 20961419-deviceconfig-restrict

Reason for revert: CTS breakage

Reverted changes: /q/submissionid:20961419-deviceconfig-restrict

Change-Id: Ib6d4bac403e26a60401bebbb8c69f2abce6af6fe

Adds these new permissions, and grants them to `adb`:
- `ALLOWLISTED_WRITE_DEVICE_CONFIG`
- `READ_WRITE_SYNC_DISABLED_MODE_CONFIG`

Uses `ALLOWLISTED_WRITE_DEVICE_CONFIG` to restrict writing flags to
DeviceConfig unless they are explicitly allowlisted.

Revokes `WRITE_DEVICE_CONFIG` from `adb`, and grants it
`ALLOWLISTED_WRITE_DEVICE_CONFIG`.

Bug: 251818659
Test: m
Change-Id: I0d6784a9437e0b0344f47d13dab0f838f63eaded

from the timing).

This is needed to ensure the oneway queue isn't overwhelmed, now that the
call is non-blocking.

Bug: 267591473
Test: atest
Change-Id: I99a61e0beae63361f74b64565d3d1efe458131bd

* Use realistic durations and moving events to ensure the model is
  fully exercised.
* The requested prediction time is only a target, and it is valid for
  predictions to fall short of it.

com.android.perftests.core (2 Tests)
[1/2] android.input.MotionPredictorBenchmark#timeRecordAndPredict: PASSED (8.392s)
	perfetto_file_path: /sdcard/test_results/android.input.MotionPredictorBenchmark_timeRecordAndPredict/PerfettoListener/perfetto_android.input.MotionPredictorBenchmark_timeRecordAndPredict-1.perfetto-trace
	timeRecordAndPredict_mean (ns): 84059
	timeRecordAndPredict_median (ns): 82093
	timeRecordAndPredict_min (ns): 77361
	timeRecordAndPredict_standardDeviation: 5402
[2/2] android.input.MotionPredictorBenchmark#timeCreatePredictor: PASSED (7.650s)
	perfetto_file_path: /sdcard/test_results/android.input.MotionPredictorBenchmark_timeCreatePredictor/PerfettoListener/perfetto_android.input.MotionPredictorBenchmark_timeCreatePredictor-1.perfetto-trace
	timeCreatePredictor_mean (ns): 104809
	timeCreatePredictor_median (ns): 105000
	timeCreatePredictor_min (ns): 101148
	timeCreatePredictor_standardDeviation: 2837

Bug: 167946763
Test: atest MotionPredictorBenchmark
Change-Id: I2861a2f377a34eff6571a6ecdc4a34c46fcd753b

This patch adds
* file:/core/java/android/content/res/OWNERS as the owner of
    src/android/content/res/*
* file:/core/java/android/content/om/OWNERS as the owner of
    src/android/content/om/*

Test: change owner only.
Bug: 177360157
Change-Id: Ib8386b0c883f747b31b44eb26a23ed577cda054e

Renaming the start user tests to make clear that it was "started twice/once before in the past, and now we're timing the next start".

Bug: 266555909
Test: atest UserLifecycleTests#startUser_startedOnceBefore
Test: atest UserLifecycleTests#startUser_startedTwiceBefore
Test: atest UserLifecycleTests#startAndUnlockUser_startedTwiceBefore
Change-Id: Ib92e1c69027d05c1dbee7ce6b3620e6e6f8ccc7e

When we create a user, the default is that we ask to initialize it whenever we change from one user to another, without removing that user from Android.
These tests will serve to demonstrate the execution time in the common case of using the start user.

Results:
startUser_uninitializedUser Mean (ms): 2026.75
startUser_startOnceBefore Mean (ms): 924.75

startUser_startTwiceBefore Mean (ms): 438.0

startAndUnlockUser Mean (ms): 2032.75
startAndUnlockUser_startTwiceBefore Mean (ms): 1134.25

switchUser_stopped Mean (ms): 3095.75
switchUser_stopped_realistic Mean (ms): 2403.25

Bug: 266192048
Bug: 266555909
Test: atest UserLifecycleTests#startUser_uninitializedUser
Test: atest UserLifecycleTests#startUser_startOnceBefore
Test: atest UserLifecycleTests#startUser_startTwiceBefore
Test: atest UserLifecycleTests#startAndUnlockUser
Test: atest UserLifecycleTests#startAndUnlockUser_startTwiceBefore
Test: atest UserLifecycleTests#switchUser_stopped
Test: atest UserLifecycleTests#switchUser_stopped_realistic
Change-Id: I1157431eb7f13b21ba2fd230e1148cd5eb09724e

Changing test name to test whether this test's data history will be reset after change.

Bug: 266557361
Test: atest UserLifecycleTest#switchUser_running_initializedUser
Change-Id: I8b7beab7de25dab7ab63d750225d4dc281816aeb

Previously, realistic versions of switchUser, createUser, startUser and
stopUser were added. This commit adds the remaining versions.
The tests were run locally, and some showed no significant changes, and
others improved mean times up to 10% (not enough data to estimate std
deviance difference).

Bug: 261621873
Test: atest UserLifecycleTests
Change-Id: I76f6e87edada4561ac967efde44526e72955e2e3

Here is an example of minimum score (ns) on the Pixel 5
  Short Text,     Regular line break:   131,025
  Short Text,     Phrase line break :   132,650
  Long Text,      Regular line break: 1,145,408
  Long Text,      Phrase line break : 1,083,803
  Long Long Text, Regular line break: 3,957,686
  Long Long Text, Phrase line break : 3,520,376

Bug: 266152097
Test: atest

Change-Id: Id3f789a75eb11c50e144e1fbd4b8ceff24924a27

Current test implementations depend on the test device default
wallpaper. Some devices use LiveWallpapers and some use ImageWallpapers,
so this test aims to make it easier to identify if a regression occured
because of the switchUser operation or due to a slower wallpaper on
certain devices.

Bug: 265011565
Test: atest UserLifecycleTests
Change-Id: Ia587b97cd672f76f9a20481124f47840de9fd331

Bug: 265367022
Test: atest StaticLayoutPerfTest
Change-Id: I54eec5b45b9d7aa69e86b4fb1e1ff51999ad0d46