B.42 Function Inline Expansion

Function inline expansion eliminates procedure-call overhead and allows general optimization methods to apply across the expanded code. Function inlining has advantages over macros in that arguments are evaluated only once, parentheses need not be overused to avoid problems with precedence, and the actual expansion can be controlled from the command line.

The following pragmas are provided to control function inline expansion:

#pragma inline (function_name [,function_name....])
#pragma noinline (function_name [,function_name....])

If a function is named in an inline directive, calls to it are expanded as inline code, if the function has the following properties:

• If a function is named in a noinline directive, calls to it are not expanded as inline code.

• If a function is not named in an inline or a noinline directive, the compiler uses a heuristic to perform inline expansion of calls where appropriate.

• The compiler issues an error if a function is named in both an inline and a noinline directive.

  If the noinline compiler option is used, it overrides all inline pragma directives.

Inline functions have the following properties:

• An inline function can be recursive, but only one level of inline expansion is performed if it is.

• Only calls from the source file containing the definition of the inlined function are expanded inline.

• The address of an inline function can be taken and expressions that imply the conversion of the inlined function name to an address are allowed.

• The use of the varargs package (allowing a function to take a variable number of arguments) is not allowed for inline functions.

• An inline function cannot be declared with an ellipsis in its argument list.