cc : Delete LayerImpl::NoteLayerPropertyChangedForSubtree

cc/trees/layer_tree_host.cc

Index: cc/trees/layer_tree_host.cc
diff --git a/cc/trees/layer_tree_host.cc b/cc/trees/layer_tree_host.cc
index 102d39211418a484b5bb2b410fe87942e3abf28a..9967983885fde052799f19268a4b9b2130916965 100644
--- a/cc/trees/layer_tree_host.cc
+++ b/cc/trees/layer_tree_host.cc
@@ -497,6 +497,7 @@ void LayerTreeHost::FinishCommitOnImplThread(LayerTreeHostImpl* host_impl) {
 
   micro_benchmark_controller_.ScheduleImplBenchmarks(host_impl);
   property_trees_.transform_tree.ResetChangeTracking();
+  property_trees_.full_tree_damaged = false;

[jaydasika, 2016/03/02 16:27:08]

I am not calling the property trees reset function here as we don't have to
reset effect tree every time. We can make  ResetAllChageTracking in property
trees take an argument should_reset_effect_tree_also and then call it here also.
Do you think that's better ?

[ajuma, 2016/03/02 16:31:31]

Could you remind me why we don't need to reset the effect tree here as well?

[jaydasika, 2016/03/02 17:19:30]

I meant, we don't have to do it everytime as we don't do any change tracking on
the effect tree on the main thread. We do have to reset the effect tree when we
have back-synced the property trees from active tree. (The comment below in line
501). 

[ajuma, 2016/03/02 17:54:06]

Oh, we don't change-track on the effect tree because opacity and filter changes
on the main thread currently trigger property tree rebuilds even when a Layer
already has an effect node. In that case, how about making ResetChangeTracking
take an enum with allowed values EFFECT_TREE, TRANSFORM_TREE, and ALL_TREES?

[jaydasika, 2016/03/02 19:13:06]

Done.

   // We don't track changes to effect tree on main thread. But, to preserve any
   // change tracking done on active tree's effect tree, we copy it to the main
   // thread's effect tree before we push the main thread property trees to