Change SF power hints to use early frame predictions
Currently we issue hints at the end of the frame reporting what happened
that frame. However, this is often too late anyway and lacks critical
information not known until after the frame completes, like gpu timing.
By waiting until the next frame start and using those timings along with
information about gpu duration, the new expected present time, and the
new wakeup time, we can instead use those timings to predict the
duration of the upcoming frame and signal in advance of a frame drop.
This patch also changes how and where frame timing is done and measures
gpu duration so we can predict if a cpu speedup can help gpu composition
happen earlier, preventing a drop that way.
Bug: b/195990840
Test: manual
Change-Id: I5f0ae796198631fced503dce8e6c495885fb256a
diff --git a/services/surfaceflinger/CompositionEngine/src/Output.cpp b/services/surfaceflinger/CompositionEngine/src/Output.cpp
index 4c30f99..3f9b45f 100644
--- a/services/surfaceflinger/CompositionEngine/src/Output.cpp
+++ b/services/surfaceflinger/CompositionEngine/src/Output.cpp
@@ -1087,6 +1087,10 @@
return;
}
+ if (isPowerHintSessionEnabled()) {
+ // get fence end time to know when gpu is complete in display
+ setHintSessionGpuFence(std::make_unique<FenceTime>(new Fence(dup(optReadyFence->get()))));
+ }
// swap buffers (presentation)
mRenderSurface->queueBuffer(std::move(*optReadyFence));
}
@@ -1390,6 +1394,14 @@
// The base class does nothing with this call.
}
+void Output::setHintSessionGpuFence(std::unique_ptr<FenceTime>&&) {
+ // The base class does nothing with this call.
+}
+
+bool Output::isPowerHintSessionEnabled() {
+ return false;
+}
+
void Output::postFramebuffer() {
ATRACE_CALL();
ALOGV(__FUNCTION__);