[Turbofan] Implement call with spread bytecode in assembly code.

src/interpreter/interpreter.cc

Index: src/interpreter/interpreter.cc
diff --git a/src/interpreter/interpreter.cc b/src/interpreter/interpreter.cc
index 8abc65a3fe3240b77326dfce5f74f22cbf4493c0..26e04c534df6d096627abe4cdba45393a45d326d 100644
--- a/src/interpreter/interpreter.cc
+++ b/src/interpreter/interpreter.cc
@@ -2176,22 +2176,25 @@ void Interpreter::DoCallJSRuntime(InterpreterAssembler* assembler) {
   __ Dispatch();
 }
 
-// CallWithSpread <first_arg> <arg_count>
+// CallWithSpread <callable> <first_arg> <arg_count>
 //
-// Call a JSfunction or Callable in |first_arg| with the receiver in
-// |first_arg + 1| and |arg_count - 2| arguments in subsequent registers. The
+// Call a JSfunction or Callable in |callable| with the receiver in
+// |first_arg| and |arg_count - 1| arguments in subsequent registers. The
 // final argument is always a spread.
 //
 void Interpreter::DoCallWithSpread(InterpreterAssembler* assembler) {
-  Node* first_arg_reg = __ BytecodeOperandReg(0);
-  Node* first_arg = __ RegisterLocation(first_arg_reg);
-  Node* args_count = __ BytecodeOperandCount(1);
+  Node* callable_reg = __ BytecodeOperandReg(0);
+  Node* callable = __ LoadRegister(callable_reg);
+  Node* receiver_reg = __ BytecodeOperandReg(1);
+  Node* receiver_arg = __ RegisterLocation(receiver_reg);
+  Node* receiver_args_count = __ BytecodeOperandCount(2);
+  Node* receiver_count = __ Int32Constant(1);
+  Node* args_count = __ Int32Sub(receiver_args_count, receiver_count);
   Node* context = __ GetContext();
 
   // Call into Runtime function CallWithSpread which does everything.
-  Node* runtime_function = __ Int32Constant(Runtime::kCallWithSpread);
   Node* result =
-      __ CallRuntimeN(runtime_function, context, first_arg, args_count);
+      __ CallJSWithSpread(callable, context, receiver_arg, args_count);
   __ SetAccumulator(result);
   __ Dispatch();
 }