| Age | Commit message (Collapse) | Author |
|---|
|
Bug: 220074017
Change-Id: Ic33b5a6c518f7aa587e6c7cf0bfd6df4c795782b
|
|
SF would set the desired active mode on the specific DisplayDevice the
mode was requested and use it to swicth the active display mode.
Change-Id: Ib4f01ec63b2087ab2a81490051d1e029afc71854
Test: SF unit tests
Test: refresh rate switching is working on device with
more than one display
Bug: 187539899
Change-Id: Ie0e5ffe364f81775c2101ffd0cf2c596051947fa
Merged-In: Ie0e5ffe364f81775c2101ffd0cf2c596051947fa
|
|
A DisplayMode should include the PhysicalDisplayId it
belongs to we could use that to know which display the
mode belogs to.
Change-Id: I781cf30c5ccf886f061a45073b369b4f04cc9140
Bug: 187539899
Test: SF unit tests
Test: refresh rate switching is working on device with
more than one display
Change-Id: I9c73b43384cbe934f7584695e5fe5c53ab402bfb
Merged-In: I9c73b43384cbe934f7584695e5fe5c53ab402bfb
|
|
Move RefreshRateConfigs to DisplayDevice to be able to maintain
a per-display refresh rate switching policy.
Test: SF unit tests
Test: refresh rate switching is working on device with
more than one display
Bug: 187539899
Change-Id: Ica6a955e8ad0e563a0740f6579b61fc592eb982c
|
|
For devices with multiple internal displays, choose the internal
display the is powered on (if any) to be the default display that
would be used mainly for vsync tracking.
Bug: 187539899
Test: SF unit tests
Change-Id: Icc305cdeec1c8a04729f00ff795dfbb806d8f72c
Merged-In: Icc305cdeec1c8a04729f00ff795dfbb806d8f72c
|
|
Change-Id: I4e775a8f840540a8644ecfe3cf33ba1bffe7b8c5
|
|
Change-Id: I44a5ef0c7858857635265f5dc64a92ebd4b367c3
|
|
BLASTBufferQueue set the max acquired buffers based on SF
vsync configuration (sf.duration and app.duration). This is calculated
based on the max supported refresh rate on the device, and it turn
is propogated to apps via min_undequeued_buffers to the app could
allocate enough buffers for maintaining the pipeline SF expects.
This leads to a higher latency when the device is running in a lower
refresh rate as there are more buffers on the buffer queue then required.
In this change we are holding on the these "extra buffers" and not
releasing them back to the buffer queue so the app would use the number
of buffers it needs based on the current refresh rate, and to avoid having
unnecessary long latency.
Bug: 188553729
Test: Run backpressure based game on 60Hz and 120Hz and collect systraces
Change-Id: I9d4e6278f0ddd28008ac437ab0576aa79d05166a
|
|
In the existing behavior Scheduler checks if the
layerReuquirements have changes since the last time
and only then calls into RefreshRateConfigs to
getBestRefreshRate. There are two problems with that
1. on the first iteration of the algorithm
mFeatures.contentRequirements is empty. If we happen
to have an empty list of current content requirements
(for example if all layers have NoVote), we
won't execute the refresh rate selection algorithm
and we'll end up with a wrong initial refresh rate.
2. the cached value needs to be invalided when one of
these change
- globalSignals
- supported display modes (happens on TV)
- display manager policy
Bug: 188872896
Test: atest SchedulerTest RefreshRateConfigsTest
Change-Id: I101f401522fae8358752e283d8375caa93957b6a
|
|
Currently if Scheduler is created when there is
only one supported display mode, LayerHistory is
not created. This optimization assumes that the
list of supported modes is never going to change
but this is not the case for TV devices.
The typical use case is when a TV dongle or
set-top box is first connected to a display
after it boots. In this cas some devices will
first report only one display mode and later
update to the full list of supported modes.
This CL changes the behaviour to always have an instance
of LayerHistory. If there's only one display mode
recordLayerHistory() and chooseRefreshRateForContent()
are noops.
The alternative to create LayerHistory on demand is
more cumbersome because we also would need to register
all already created layers.
Bug: 188872896
Test: presubmit
Change-Id: Ia236fd4a81cc736c172220ff60762ddbc3cbb83a
|
|
-- Upon receiving app's request for vsync, trigger sf refresh
if display is in idle.
CRs-Fixed: 2786504
Change-Id: Icb49a001f5f6a6f33d52950ecd464a6aac0b5147
|
|
Every class that inherits from RefBase must be held by an sp<>.
Creating a RefBAse object without an sp<> causes a memory leak.
Test: adb shell /data/nativetest64/libsurfaceflinger_unittest/libsurfaceflinger_unittest
Bug: 181807839
Change-Id: I37ec274b01c5212a5a3cecc2e1eb5b4c9d6122f4
|
|
|
|
Expose a function for clients to call and query the number
of extra buffers needed when vsync id is passed with the buffer.
Test: launch an app and observe systrace
Test: SF unit tests
Bug: 178148035
Change-Id: Icbeec66073feeae9768f0dcc45831b26144ab6f6
|
|
This CL continues the raneming of display "configs" to
display "modes". The goal of this is to have the same
names as in the java code and the public display APIs.
Additionally in this CL we move DisplayConfig (from libui)
to the namespace android::ui. This is to prevent conflict
with the SurfaceFlinger's internal android::DisplayMode.
This is consistent with the neighboring classes which
are also in the ui namespace.
Also the type of the parameter defaultMode of
{s,g}etDesiredDisplayModeSpecs is changed to size_t
for consistency with the rest of the code. Appropriate
error handling is added for this.
Bug: 159590486
Bug: 179158858
Test: presubmit
Change-Id: I31e5be1f2223a9ec9340789ce3dc5738eceaf40f
|
|
Bug: 159590486
Bug: 176148651
Test: atest libsurfaceflinger_unittest
Change-Id: Ie177b213bb7a8ab9f67a51d2b5cf27d8fd97b780
|
|
HWC2 is unnecessary stateful stopgap between HWComposer and
ComposerHal. In this CL the following functions are removed from HWC2
* getActiveConfig
* getActiveConfigIndex
* getDisplayVsyncPeriod
and ComposerHal is called directly from HWComposer. This way display
configs are stored only in HWComposer.
Additionally HWC2::Display::Config is renamed to DisplayMode and
it's extracted in its own file.
From the perspective of SurfaceFlinger this CL is not modifying
behaviour.
Bug: 159590486
Bug: 175678215
Test: atest libsurfaceflinger_unittest
Change-Id: I8cb450209adf038d891cff00d1c2690c8e6d94f7
|
|
* changes:
Move LayerHistoryV2 to LayerHistory
SF: Deprecate content detection v1
|
|
Remove the code for the old content detection algorithm,
which is no longer used.
Bug: 174120566
Test: presubmit
Change-Id: I0828bcb886f32ec2ebc896848b72340862613100
|
|
Add a new event to DisplayEventReceiver that conveys a list
of apps whose frame rates have been overriden.
This event is processed by the DisplayManager to advertise the
new frame rate to the overridden app.
Change-Id: I89fce8b5b4d9db65ec5db7dd2393c384c0fcfd82
Bug: 169271059
Bug: 169271062
Bug: 170503758
Test: manual test using SF backdoor
|
|
Add the ability for SurfaceFlinger to be able to throttle down to
a divider of the refresh rate (i.e. for 30/45 for 90Hz)
Change-Id: I6bfd6f43ee1f30e771a136c558d8ae9a6d7fbe0f
Test: Manually via 1039 SF backdoor
Bug: 170502573
Bug: 169270763
Bug: 169271059
|
|
The following happens when we create connection
1. initScheduler()
2. getFactory().createScheduler()
3. mTouchTimer.emplace
4. mScheduler->createConnection
Which means that timer created in step 3 expires exactly
while we are in step 4, so then the EventThread is half baked
in rare cases and we crash.
Also, do not continue to calling functions if the
optional fields do not have a value.
Test: libsurfaceflinger_unittest --gtest_filter=SchedulerTest.testDispatchCachedReportedConfig
Bug: 160926398
Change-Id: Ib2617b914145bc4180cc7ca27203c59dbd625c94
|
|
This change is using the duration given to sf/app without
converting them back to the legacy offset. This allows us to get
the expected vsync associated with a wakeup event.
This change also required some refactoring:
- Move the periodic callback interface out of DispSync and in
to DispSyncSource
- Translate legacy offsets to duration in case that duration
is not specified in the build files
- Add support to VSD to expose the deadline timestamp of when
a frame needs to be ready by
Bug: 162888874
Test: SF unit tests
Test: examine systraces
Change-Id: I87a53cc7dea931d3c195eab6842e003ca4516885
|
|
|
|
EventControlThread was a HWC workaround dating back to K for toggling
VSYNC off the main thread, but as of R it defers back to main only to
incur context switches.
Clean up TestableScheduler construction to skip creating DispSync and
timer threads, which fixes several gMock warnings of unexpected calls.
Remove virtual destructors for non-polymorphic interfaces.
Bug: 160012986
Test: systrace
Test: libsurfaceflinger_unittest
Change-Id: I01360016a7dba79dc905f250753e6fce34af0a90
|
|
Use a type safe wrapper around uint64_t for display ids.
We also modify the format of the EDID generated physical
display IDs by setting bit 62 to 1, which will indicate
that the display id is stable across reboot.
Bug: 160679868
Test: m && flash device
Test: take screnshot on device
Test: atest surfaceflinger_unittest
Change-Id: Ie2c7c2b737e0882fa4208c059caa85b7ded922b2
|
|
Avoid registering, recording, partitioning, and transforming layers
for content detection in the scheduler if there is a single refresh
rate config.
Dump content detection version, and layer history stats.
Bug: 158780872
Test: No layer history on crosshatch, ARC, etc.
Test: dumpsys SurfaceFlinger --vsync
Test: libsurfaceflinger_unittest
Change-Id: Ia23a35cd6aa0cb117b19c98f5d0af8c14836430e
|
|
Do not change refresh rate across configs with multiple resolutions
Bug: 150237118
Test: adb shell /data/nativetest64/libsurfaceflinger_unittest/libsurfaceflinger_unittest
Change-Id: I9b0f355a72343cf120229b98e6ad082ea75d3080
|
|
Bug: 135480141
Test: ./statsd_testdrive 10062
Change-Id: I220f0c96935c96afea194ed8ca69922785cf296e
|
|
ag/9294789 introduce a flag to only allow refresh rate switching on the
devices that have that flag set to true, because it broke devices that
didn't support multiple refresh rates, and Android TV.
LocalDisplayAdapter sets the refresh rate to default when booting the
phone, so no need for that in SF anymore.
Test: Turn off Smooth display (see b/148821456), device doesn't crash.
Test: Install on devices that doesn't have multiple refresh rates.
Test: Run SF unittests.
Test: Turn on all the flags, and run through set of examples:
Video playback stays at 60.
Camera stays at 60.
Maps play at 60.
Low brightness stays at 90.
Swappy successfully switches between 60 & 90.
Test: Turn off all the flags, and rerun the examples from previous.
If applications requested the change, the change is honored,
otherwise we stay at default (set by DM).
Test: Ask Android TV team to test the patch.
Bug: 148427603
Bug: 148821456
Change-Id: I5bb964572d93e5cb78d7b75054b900917be563b0
|
|
This reverts commit aa52b007cb5c96d6124a293ca80c430c7051f7f3.
Reason for revert: The original revert cause a failure: b/148821456. So the revert was created. Which caused a build failure b/148835312. Reverting the original change and fixing the test (that caused build failure) with it.
Bug: 148835312
Bug: 148821456
Bug: 148835894
Test: build, install on coral.
Test: adb shell /data/nativetest64/libsurfaceflinger_unittest/libsurfaceflinger_unittest
Change-Id: I48964c7c27ef53e8fb03d959ef85ee2ec64e1ed1
|
|
This reverts commit 2a215191be4854fcb4b2836929a955157eae52d1.
Reason for revert: Droidcop-triggered revert due to breakage https://android-build.googleplex.com/builds/quarterdeck?branch=git_master&target=hawk-userdebug&lkgb=6180052&lkbb=6180424&fkbb=6180072, bug b/148835312.
Bug: 148835312
Change-Id: Ie8e5d4c683ab7f10051e355a62941bdb37a196b7
|
|
This reverts commit 8c6f3f6bd0c5edce7aa9bfd61b64dd2e7dc43deb.
Reason for revert: Crashes on ToT (b/148822765, b/148821456)
Change-Id: If1bc13b33321727244c3b0f15a0e7b30665d1c43
|
|
ag/9294789 introduce a flag to only allow refresh rate switching on the
devices that have that flag set to true, because it broke devices that
didn't support multiple refresh rates, and Android TV.
LocalDisplayAdapter sets the refresh rate to default when booting the
phone, so no need for that in SF anymore.
Test: Install on devices that doesn't have multiple refresh rates.
Test: Run SF unittests.
Test: Turn on all the flags, and run through set of examples:
Video playback stays at 60.
Camera stays at 60.
Maps play at 60.
Low brightness stays at 90.
Swappy successfully switches between 60 & 90.
Test: Turn off all the flags, and rerun the examples from previous.
If applications requested the change, the change is honored,
otherwise we stay at default (set by DM).
Test: Ask Android TV team to test the patch.
Bug: 148427603
Change-Id: I44085e91a74902d102bcf9f1491164cdb50650cc
|
|
Enhance the refresh rate selection algorithm to allow having multiple
refresh rate. The new process attaches scores to each one of the available
refresh rate and chooses the refresh rate with the highest score.
This behavior is currently controlled by the sysprop flag
'debug.sf.use_content_detection_v2' and currently turned off.
This algorithm stills needs some tunings which will be done in
layer CLs.
Test: adb shell /data/nativetest64/SurfaceFlinger_test/SurfaceFlinger_test
Test: adb shell /data/nativetest64/libsurfaceflinger_unittest/libsurfaceflinger_unittest
Test: go/90hzscenarios manual tests
Bug: 147516364
Fixes: 146068419
Change-Id: I06e07459e469482799ff80fa54fa8dd311325e0e
|
|
Enable global -WConversion in Android.bp and turn it off in
individual files. This is the first step to enable -WConversion
in SurfaceFlinger.
Test: boot
Bug: 129481165
Change-Id: Ia752df39ef380bde71705a32582d618242c466f7
|
|
Composer 2.4 adds a new attribute for configs groups. This change
groups configs according to their group and store them in
RefreshRateConfigs.
Test: rev up composer to 2.4 and test refresh rate switching
Test: adb shell /data/nativetest64/libsurfaceflinger_unittest/libsurfaceflinger_unittest
Test: adb shell /data/nativetest64/SurfaceFlinger_test/SurfaceFlinger_test
Bug: 141329414
Fixes: 139751853
Change-Id: Ic0bcd3da4bf6b73efa11a60c2594948ce030362f
|
|
This CL refactors VSYNC injection to interface with the Scheduler API,
and removes otherwise unused EventThread members.
Bug: 128863962
Test: sffakehwc_test
Change-Id: I2ba143bb78baf3e66a47b90163eacf0d0e431124
|
|
Some devices don't do refresh rate switching, so we should take that
into account when filtering display manager config settings and deciding
scheduling behavior.
This CL adds a sysprop that can be set to indicate if surface flinger
should do refresh rate switching, and modifies surface flinger to have
the correct behavior when we're not doing refresh rate switching.
Bug: 136592946
Bug: 138261472
Test: Ran through various 60/90 switching scenarios on a device with
refresh rate switching.
Test: Set the refresh rate switching sysprop to false, and confirmed we
get a consistent 60Hz.
Test: Inspected dumpsys output and confirmed it looks correct. In
particular, refresh rate stats are output correctly.
Test: Ran automated tests: RefreshRateConfigsTest, RefreshRateStatsTest,
SchedulerTest.
Change-Id: I54cd5be9d2c1b9abc8475c3ce39846cbe9f9fe53
|
|
Remove dynamic allocation for Scheduler::{Connection,ConnectionHandle},
as well as ref-counting for the latter. Also, remove dead code and make
members private.
Bug: 130554049
Test: libsurfaceflinger_unittest
Change-Id: Ibb9dc8d4cb66451a4172c852a36032bbc0a54411
|
|
|
|
Pass negative offsets to DispSync to fix the scheduling
when SurfaceFlinger uses negative offsets.
Change-Id: I1f9544b064305c87f973120cc1bc59a0268b78e5
Bug: 132284303
Test: UI-Bench
|
|
When display refresh rate changes, SF fires DISPLAY_EVENT_CONFIG_CHANGED
thru DisplayEventReceiver. If the other end of the pipe doesn't consume
the events it may lead to pipe full and dropping of events. Furthermore,
The only clients interested in this event in DisplayManager and hwui.
To avoid spamming all clients with this event, this change is adding an
explicit register for DISPLAY_EVENT_CONFIG_CHANGED events.
Bug: 131688378
Test: adb shell /data/nativetest64/libsurfaceflinger_unittest/libsurfaceflinger_unittest
Test: trigger config change and observe logcat
Change-Id: I5973a1ecc1f3e3ff8d8a0cba19db0e49ef0d5341
|
|
Add the notion of Touch Events to Scheduler to enhance the algorithm
to move to Performance refresh rate.
Scheduler selects Performance when:
- There is a touch event
- There is a buffer to HWC
This change also removes the behavior of Scheduler to move to Performance
on Choreographer callbacks.
Test: Switch between apps using gesture navigation
Bug: 131906818
Change-Id: I588cfc32449e87744e829dc7c5261a2e4151a8f8
|
|
1) Each time SF creates a layer, register it with Scheduler and return handle
2) BufferQueueLayer and BufferStateLayer can now send information about buffers
for given layers via layer handle.
Algorithm for detecting content fps:
1) Keep the refresh rate per layer (explicit timestamp, or 0).
2) Keep information about last 10 present or update timestamps. This will be an
indicator for precedence.
3) Choose the MAX refresh rate among last updated layers.
For more info see go/surface-flinger-scheduler and
go/content-fps-detection-in-scheduler
Test: Updating unit tests. Systrace.
Change-Id: I988a7a79e9a9f0f61674c9b637c5142db3336177
Bug: 127727337
|
|
Dispatch timer reset callback from Scheduler's thread. This would
make both Scheduler callbacks (idle timer and reset timer) to be dispatched
from the same context and avoid any potential deadlocks.
Test: surfaceflinger unit tests
Change-Id: Iad7d10b85db090662f9b7c4ffb45ed9110035ea4
|
|
SurfaceFlinger::setRefreshRateTo, called from the Scheduler callbacks,
reads HWC and SF state without locking mStateLock, concurrently with
writes from the main thread. This CL registers ResetIdleTimerCallback
per EventThreadConnection, and locks mStateLock for connections used
off the main thread. Note that ExpiredTimerCallback locks mStateLock
unconditionally, since it is always called from the IdleTimer thread.
This CL also adds a thread annotation to setRefreshRateTo, and refactors
it accordingly.
Bug: 123715322
Test: libsurfaceflinger_unittest
Test: Boot with scheduler enabled
Change-Id: Id62c48ae22da38f292ffc18e8731a1c49a0a083c
|
|
This CL replaces ISurfaceComposer::{eDisplayIdMain,eDisplayIdHdmi} with
the stable 64-bit display IDs generated by SF. Note that the 64-bit IDs
fall back to the old values if the HWC API for display identification is
not supported.
Bug: 74619554
Test: LocalDisplayAdapter and Choreographer receive 64-bit IDs
Test: 64-bit IDs fall back to 0 and 1 on HWC 2.2 and below
Change-Id: I3c08eff6eb8bb179ecce596ab2820a2aa44c8649
|
|
Bug: None
Test: Build
Change-Id: I578fc0d94691edcd905f78bc75e08072d9c49744
|
|
This CL ties the resync callback to an EventThreadConnection instead
of an EventThread. This is a step towards having IDisplayEventConnection
subscribe to a given display rather than the primary display implicitly.
Each display will then have SurfaceFlinger::VsyncState that resyncs
independently at potentially different rates. Callbacks have weak
references to the per-display VsyncState owned by SurfaceFlinger.
Bug: 74619554
Test: Boot and turn display on/off repeatedly
Change-Id: Ic7cc64e2004fa07a5d54431fc330995048a4ed20
|
