Re: [PATCH] Stack corruption in naked functions.

[Paul Brook <paul@codesourcery.com>]

On Friday 23 May 2008, Richard Guenther wrote:
> On Fri, May 23, 2008 at 3:23 PM, Paul Brook <paul@codesourcery.com> wrote:
> >> I wonder if you start to get ICEs all over the place if you use naked
> >> on a function with addressable local variables or with BLKmode
> >> parameters?  IMHO it would be better to sorry () as soon as a
> >> stack slot is allocated for such a function.
> >
> > I'd make it a proper error.  sorry() is for things that should work, but
> > we haven't got round to implementing.
> > In this case the documentation explicitly says this won't work. Any stack
> > slots are either user error or a compiler bug (like the one we're
> > fixing).
>
> But if it only doesn't work because at -O0 we're not optimizing then this
> case is a sorry () IMHO - of course we probably cannot easily distinguish
> our from the user errors here, so I won't mind using error either.

What are you referring to when you say "it doesn't work"?

The only supported use is that mentioned in the documentation. If that results 
in stack slots then this is an internal compiler error.  

If the stack slots result from code not permitted by the documentation then 
this is user error.  Some not-permitted code may happen to "work" at some 
optimisation levels with some compilers. However when this stops working 
(e.g. at -O0) it is IMHO not defect in the compiler, it is user error.  If 
anything the bug is that we allowed the code in the first place.

Paul