Improve performance of invokevirtual/invokeinterface with embedded imt/vtable

Add an embedded version of imt/vtable into class object. Both tables start at
fixed offset within class object so method/entry point can be loaded directly
from class object for invokeinterface/invokevirtual.

Bug: 8142917
Change-Id: I4240d58cfbe9250107c95c0708c036854c455968
diff --git a/runtime/utils.h b/runtime/utils.h
index 448c591..b47de81 100644
--- a/runtime/utils.h
+++ b/runtime/utils.h
@@ -167,6 +167,10 @@
 
 // For rounding integers.
 template<typename T>
+static constexpr T RoundDown(T x, typename TypeIdentity<T>::type n)
+    __attribute__((warn_unused_result));
+
+template<typename T>
 static constexpr T RoundDown(T x, typename TypeIdentity<T>::type n) {
   return
       DCHECK_CONSTEXPR(IsPowerOfTwo(n), , T(0))
@@ -174,17 +178,27 @@
 }
 
 template<typename T>
+static constexpr T RoundUp(T x, typename TypeIdentity<T>::type n)
+    __attribute__((warn_unused_result));
+
+template<typename T>
 static constexpr T RoundUp(T x, typename TypeIdentity<T>::type n) {
   return RoundDown(x + n - 1, n);
 }
 
 // For aligning pointers.
 template<typename T>
+static inline T* AlignDown(T* x, uintptr_t n) __attribute__((warn_unused_result));
+
+template<typename T>
 static inline T* AlignDown(T* x, uintptr_t n) {
   return reinterpret_cast<T*>(RoundDown(reinterpret_cast<uintptr_t>(x), n));
 }
 
 template<typename T>
+static inline T* AlignUp(T* x, uintptr_t n) __attribute__((warn_unused_result));
+
+template<typename T>
 static inline T* AlignUp(T* x, uintptr_t n) {
   return reinterpret_cast<T*>(RoundUp(reinterpret_cast<uintptr_t>(x), n));
 }