GetDumpFds should return a vector sorted with descending
order by the mtimes of dumps.
Bug: 149242559
Test: atest dumpstate_test
Test: adb bugreport && verify last anr dumped
Change-Id: I1ed9ceb35ec93b10660f2fb31785899d50bd285d
Make sure that we have the __INTRODUCED_IN()s and -- more importantly --
the API levels recorded in the documentation.
Bug: https://github.com/android/ndk/issues/1271
Test: treehugger
Change-Id: I74752f19f165fd5c56b2dfd783298c2da1a926d6
Sensors that derive their output from physical sensors and effectively
apply an irreversible filter onto their data. Requantizing this
information has little benefit and can potentially increase floating
point inaccuracy due to the additional math.
Bug: 157986376
Test: presubmits
Change-Id: I4158739abb6b74d978abe75bd8e03ac32ad4e998
There is a bug where the sensor is not properly deactivated
when ann app becomes idle. This CL fixes the issue, and also
avoids calling batch() if no active clients exist.
Bug: 157934995
Test: Flash, verify sensor is disabled when app goes idle
and re-enabled when app becomes active again
Change-Id: I403b02a20dc89c5ce0630de6ddfd6242bf38ec37
This change includes two separate fixes to make waking up from idle more
robust when using a kernel idle timer:
1) Always reconfirm the refresh rate when we resync to hardware vsync.
This is because if the kernel idle timer is enabled, we may not
immediately be in the refresh rate we expect to be.
2) Ignore the refresh rate reported by the hardware vsync callbacks,
since that refresh rate may not reflect the actual refresh rate
when the kernel idle timer is in effect.
Bug: 154303580
Test: systrace
Test: libsurfaceflinger_unittest
Change-Id: I79c66faeaed262acd9c5925fe2202a9fb3f10b7b
When we still have pending transactions (e.g. we are waiting on the
other side of a deferred transaction), we need to set the transaction
flags on SurfaceFlinger, so SurfaceFlinger will continue to call
doTransaction and give us a chance to process them. However when
we set the transaction flags, we trigger a wake up! This isn't
desirable, as the frame might have not come in. If the frame did come
in we will get to INVALIDATE, and from INVALIDATE observe the
transaction flags. This means we can set the transaction flags without
scheduling a wakeup and rely on the incoming frame to eventually wake
us up. I also considered triggering the transaction processing from
onFrameAvailable but at that point we are too late in the composition
cycle.
Bug: 157685525
Test: Existing tests pass
Change-Id: Ia54bc9f48a80906fc8f21c46debc72fabadd3e07
If a transaction is posted with ISurfaceComposer::eAnimation flag,
record it with layer history. Layers with eAnimation are considered
animating and will vote for peak refresh rate.
Bug: 157695685
Test: Play 24fps video in YouTube PIP mode with setFrameRate and rotate the device - no jank
Test: Chrome playing video - no refresh rate switching
Test: Hide/Show keyboard when inputting text
Test: Running Hay Day and observing refresh rate
Change-Id: If62e11b395c44d9e5fd40b74864fefd068953413
To reduce the delta of the upcoming ANR refactor, add a basic ANR test
here. This will also help highlight the difference in behaviour from the
current code to the new code.
To cause an ANR today, the socket needs to be blocked, which means that
we need to send ~ 50 events to the unresponsive window.
These workarounds will be removed when the ANRs are refactored.
Bug: 143459140
Test: adb shell -t /data/nativetest64/inputflinger_tests/inputflinger_tests --gtest_filter="*InputDispatcherSingleWindowAnr*" --gtest_repeat=1000
--gtest_break_on_failure
Change-Id: I0a1b28c2785d03d8870691641e0f7c6b1ca3b85e
Merged-In: I0a1b28c2785d03d8870691641e0f7c6b1ca3b85e
The current implementation of layer history discards all the history
if a layer is considered inactive. This leads to an issue in the scenario
that a layer is posting infrequent updates so slow, that by the time we
get the next buffer, the layer is considered inactive, and we bump the
refresh rate to Max, instead of detecting that layer as infrequent.
As a fix, we keep around the layer history even if it is inactive.
To avoid cases where detect animations too late, we clear the history
on touch event.
Bug: 156654519
Test: add a new unit test that simulates the buffer pattern
of a Netflix playback
Change-Id: Ie076891806d74dde814fbbb3b1a73c38e0b2ae34
am skip reason: Change-Id Ie7a9546550b70c8834462b06de929472196fe713 with SHA-1 888a6a41ed is in history
Change-Id: Ief69a7d420179ad03b0ab8fff2a0182bae220990
am skip reason: Change-Id Ia7e0c952fcaefd14280322d9b2269f4dba38e8da with SHA-1 01a5613bda is in history
Change-Id: Ib5a9b9cdad619b994181a05037d8b80a22cd5bea
This change allows creating an effect layers without a color fill so
we can avoid an additional sf transactions when creating effect
layers.
Fixes: 157646690
Test: go/wm-smoke
Test: atest SurfaceFlinger_test
Change-Id: I3d8a4f4820d7b8fb05daab697c25cff8def612c1
Adds a -A commandline option to indicate a much smaller set of
dumpstate output safe for ARC++ bug reports.
Implements output of system logcat, event logs, networking, and dropbox
system server crash/system app crashes only.
Bug: 142684959
Bug: 136273873
Test: atest dumpstate_test
Change-Id: I192be7ed841cee0a8847e1057209ef2b164bab07
The snapshots are no longer needed after restore is done.
We do this in keeping with the fact that apk/apex code data is
deleted immediately after the rollback is committed.
See go/rollbackmanager-snapshot-deletion for more details.
Bug: 151805360
Test: N/A. Will be added in the next CL.
Merged-In: I5a7540aa08922c4b0c617c9cb37e1bec2ff23fc1
Change-Id: I5a7540aa08922c4b0c617c9cb37e1bec2ff23fc1
