Till Straumann wrote: > > 2) IMO we should review all uses of "m" memory operands in inline > assembly constructs. With current gccs "m" does not what you think > it does! > I remember this discussion. Do you have a reference ? I cannot find it. Thanks Chris