GPU benchmarks should track the frequency of impl-thread vs. main-thread scrolls

cc/CCRenderingStats.h

Index: cc/CCRenderingStats.h
diff --git a/cc/CCRenderingStats.h b/cc/CCRenderingStats.h
index 0203df3872a0fce38692587bdb9cbfbed120c9e0..3292c2c44f406475e72d9d2d0031f655e63ba2ec 100644
--- a/cc/CCRenderingStats.h
+++ b/cc/CCRenderingStats.h
@@ -16,6 +16,7 @@ struct CCRenderingStats {
     double totalRasterizeTimeInSeconds;
     double totalCommitTimeInSeconds;
     size_t totalCommitCount;
+    bool didImplThreadScroll;
 
     CCRenderingStats()
         : numAnimationFrames(0)
@@ -25,6 +26,7 @@ struct CCRenderingStats {
         , totalRasterizeTimeInSeconds(0)
         , totalCommitTimeInSeconds(0)
         , totalCommitCount(0)
+        , didImplThreadScroll(false)
     {
     }
 };