[builtins] Introduce proper Float64Log2 and Float64Log10 operators.

src/compiler/js-builtin-reducer.h

Index: src/compiler/js-builtin-reducer.h
diff --git a/src/compiler/js-builtin-reducer.h b/src/compiler/js-builtin-reducer.h
index 2b535a5e7e7249873ed734f492c8c78f7f6a2a76..e9ec4e749e734e5b0eb5be3db95c22ac51ca4592 100644
--- a/src/compiler/js-builtin-reducer.h
+++ b/src/compiler/js-builtin-reducer.h
@@ -38,6 +38,8 @@ class JSBuiltinReducer final : public AdvancedReducer {
   Reduction ReduceMathImul(Node* node);
   Reduction ReduceMathLog(Node* node);
   Reduction ReduceMathLog1p(Node* node);
+  Reduction ReduceMathLog2(Node* node);
+  Reduction ReduceMathLog10(Node* node);
   Reduction ReduceMathMax(Node* node);
   Reduction ReduceMathMin(Node* node);
   Reduction ReduceMathRound(Node* node);