SurfaceFlinger: introduce PhaseOffsetsAsDurations

Currently we define phase offset for each refresh rate. This works for
<= 2 refresh rates, but doesn't scale well. This change is introducing
a new way to calculate phase offsets, which is based on duration. Then,
based on the duration and refresh rate, a phase offset is calculated.

The calculation is captured here: https://docs.google.com/spreadsheets/d/1a_5cVNY3LUAkeg-yL56rYQNwved6Hy-dvEcKSxp6f8k/edit#gid=0

Bug: 145561086
Bug: 141329414
Test: jank tests
Test: adb shell /data/nativetest64/libsurfaceflinger_unittest/libsurfaceflinger_unittest
Change-Id: I16aaf7437d30c4b12f955bdaac36582dd100519f
diff --git a/services/surfaceflinger/Scheduler/DispSyncSource.h b/services/surfaceflinger/Scheduler/DispSyncSource.h
index 536464e..328b8c1 100644
--- a/services/surfaceflinger/Scheduler/DispSyncSource.h
+++ b/services/surfaceflinger/Scheduler/DispSyncSource.h
@@ -26,8 +26,7 @@
 
 class DispSyncSource final : public VSyncSource, private DispSync::Callback {
 public:
-    DispSyncSource(DispSync* dispSync, nsecs_t phaseOffset, nsecs_t offsetThresholdForNextVsync,
-                   bool traceVsync, const char* name);
+    DispSyncSource(DispSync* dispSync, nsecs_t phaseOffset, bool traceVsync, const char* name);
 
     ~DispSyncSource() override = default;
 
@@ -55,7 +54,6 @@
 
     std::mutex mVsyncMutex;
     TracedOrdinal<nsecs_t> mPhaseOffset GUARDED_BY(mVsyncMutex);
-    const nsecs_t mOffsetThresholdForNextVsync;
     bool mEnabled GUARDED_BY(mVsyncMutex) = false;
 };