Porting to .intel_syntax Inline Assembly / "m" Constraint with .intel_syntax / Targeting -fPIC

Ian Lance Taylor iant@google.com
Wed Mar 5 00:37:00 GMT 2008

Daniel Yek <dyek@real.com> writes:

> I would like to find out if it is a good idea to utilize GCC's
> .intel_syntax Inline Assembly support (for non-trivial blocks), or to
> convert them to GAS or NASM assembly functions, .S files.

For a complete function written in assembly code, you might as well
put it in a .s or .S file.  You can put it in a .c file using a simple
asm if that is is more convenient for you.  It doesn't make any big

> Is there a way I can use the memory constraint, "m", effectively in
> .intel_syntax Inline Assembly?

Use the -masm=intel option with gcc.

> (3) Is there any document providing information how I can create a
> single version of assembly code (Inline Assembly or pure assembly) and
> have it compiled with or without -fPIC correctly?

I don't know of any document.  The basic guideline is fairly obvious:
don't use %ebx.  If you must use it, push and pop it in your assembly


More information about the Gcc-help mailing list