Rename A64 port to ARM64 port

src/arm64/lithium-codegen-arm64.cc

Index: src/arm64/lithium-codegen-arm64.cc
diff --git a/src/a64/lithium-codegen-a64.cc b/src/arm64/lithium-codegen-arm64.cc
similarity index 99%
rename from src/a64/lithium-codegen-a64.cc
rename to src/arm64/lithium-codegen-arm64.cc
index 060c1da52c6826f4212148b53b9c6817c103bf9c..c1ac121a3446878487f8c2461a705d2a0be76783 100644
--- a/src/a64/lithium-codegen-a64.cc
+++ b/src/arm64/lithium-codegen-arm64.cc
@@ -27,8 +27,8 @@
 
 #include "v8.h"
 
-#include "a64/lithium-codegen-a64.h"
-#include "a64/lithium-gap-resolver-a64.h"
+#include "arm64/lithium-codegen-arm64.h"
+#include "arm64/lithium-gap-resolver-arm64.h"
 #include "code-stubs.h"
 #include "stub-cache.h"
 #include "hydrogen-osr.h"
@@ -1698,8 +1698,8 @@ void LCodeGen::DoArithmeticD(LArithmeticD* instr) {
       // calling a helper function. With frintz (to produce the intermediate
       // quotient) and fmsub (to calculate the remainder without loss of
       // precision), it should be possible. However, we would need support for
-      // fdiv in round-towards-zero mode, and the A64 simulator doesn't support
-      // that yet.
+      // fdiv in round-towards-zero mode, and the ARM64 simulator doesn't
+      // support that yet.
       ASSERT(left.Is(d0));
       ASSERT(right.Is(d1));
       __ CallCFunction(
@@ -2000,7 +2000,7 @@ void LCodeGen::DoCallWithDescriptor(LCallWithDescriptor* instr) {
     Handle<Code> code = Handle<Code>::cast(ToHandle(target));
     generator.BeforeCall(__ CallSize(code, RelocInfo::CODE_TARGET));
     // TODO(all): on ARM we use a call descriptor to specify a storage mode
-    // but on A64 we only have one storage mode so it isn't necessary. Check
+    // but on ARM64 we only have one storage mode so it isn't necessary. Check
     // this understanding is correct.
     __ Call(code, RelocInfo::CODE_TARGET, TypeFeedbackId::None());
   } else {