Introduce MoveInteger64 instruction into X64 Macro Assembler

src/x64/code-stubs-x64.cc

Index: src/x64/code-stubs-x64.cc
diff --git a/src/x64/code-stubs-x64.cc b/src/x64/code-stubs-x64.cc
index be8160b7d396124a2eba98eadebe14bbdc65b7da..06afa18bc65f308030993c65c730592800729bff 100644
--- a/src/x64/code-stubs-x64.cc
+++ b/src/x64/code-stubs-x64.cc
@@ -959,7 +959,7 @@ void MathPowStub::Generate(MacroAssembler* masm) {
       Label continue_sqrt, continue_rsqrt, not_plus_half;
       // Test for 0.5.
       // Load double_scratch with 0.5.
-      __ movq(scratch, V8_UINT64_C(0x3FE0000000000000));
+      __ MoveInteger64(scratch, V8_UINT64_C(0x3FE0000000000000));
       __ movq(double_scratch, scratch);
       // Already ruled out NaNs for exponent.
       __ ucomisd(double_scratch, double_exponent);
@@ -969,7 +969,7 @@ void MathPowStub::Generate(MacroAssembler* masm) {
       // Math.pow(-Infinity, 0.5) == Infinity (ECMA spec, 15.8.2.13).
       // According to IEEE-754, double-precision -Infinity has the highest
       // 12 bits set and the lowest 52 bits cleared.
-      __ movq(scratch, V8_UINT64_C(0xFFF0000000000000));
+      __ MoveInteger64(scratch, V8_UINT64_C(0xFFF0000000000000));
       __ movq(double_scratch, scratch);
       __ ucomisd(double_scratch, double_base);
       // Comparing -Infinity with NaN results in "unordered", which sets the
@@ -1001,7 +1001,7 @@ void MathPowStub::Generate(MacroAssembler* masm) {
       // case of Math.pow(-Infinity, -0.5) == 0 (ECMA spec, 15.8.2.13).
       // According to IEEE-754, double-precision -Infinity has the highest
       // 12 bits set and the lowest 52 bits cleared.
-      __ movq(scratch, V8_UINT64_C(0xFFF0000000000000));
+      __ MoveInteger64(scratch, V8_UINT64_C(0xFFF0000000000000));
       __ movq(double_scratch, scratch);
       __ ucomisd(double_scratch, double_base);
       // Comparing -Infinity with NaN results in "unordered", which sets the