From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-return-50916-listarch-gcc=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 5547 invoked by alias); 29 Apr 2002 17:58:55 -0000
Mailing-List: contact gcc-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Archive: <http://gcc.gnu.org/ml/gcc/>
List-Post: <mailto:gcc@gcc.gnu.org>
List-Help: <http://gcc.gnu.org/ml/>
Sender: gcc-owner@gcc.gnu.org
Received: (qmail 5489 invoked from network); 29 Apr 2002 17:58:33 -0000
Received: from unknown (HELO linuxpc1.lauterbach.com) (213.70.137.66)
  by sources.redhat.com with SMTP; 29 Apr 2002 17:58:33 -0000
Received: (qmail 4000 invoked by uid 82); 29 Apr 2002 17:58:30 -0000
Received: from Franz.Sirl-kernel@lauterbach.com by linuxpc1.lauterbach.com with qmail-scanner-1.01 (inocmd32: /. . Clean. Processed in 0.828732 secs); 29 Apr 2002 17:58:30 -0000
Received: from dsl-213-023-062-210.arcor-ip.net (HELO enzo) (213.23.62.210)
  by linuxpc1.lauterbach.com with SMTP; 29 Apr 2002 17:58:29 -0000
Content-Type: text/plain;
  charset="iso-8859-1"
From: Franz Sirl <Franz.Sirl-kernel@lauterbach.com>
To: Jason Merrill <jason@redhat.com>
Subject: Re: [PATCH] Fix PR c/6343 (was: Re: GCC 3.1 Prerelease)
Date: Mon, 29 Apr 2002 11:36:00 -0000
Cc: gcc-patches@gcc.gnu.org,
 Richard Henderson <rth@redhat.com>,
 Alan Modra <amodra@bigpond.net.au>,
 Mark Mitchell <mark@codesourcery.com>,
 gcc@gcc.gnu.org
References: <5.1.1.2.2.20020423130143.04a21008@mail.lauterbach.com> <200204281640.27693@enzo.bigblue.local> <wvl662bmqos.fsf@prospero.cambridge.redhat.com>
In-Reply-To: <wvl662bmqos.fsf@prospero.cambridge.redhat.com>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Message-Id: <200204291959.42699@enzo.bigblue.local>
X-SW-Source: 2002-04/txt/msg01568.txt.bz2

On Sunday 28 April 2002 23:57, Jason Merrill wrote:
> >>>>> "Franz" == Franz Sirl <Franz.Sirl-kernel@lauterbach.com> writes:
> >
> > +  if (SUPPORTS_WEAK
> > +      && DECL_EXTERNAL (newdecl) && DECL_EXTERNAL (olddecl)
> > +      && (TREE_CODE (decl) != VAR_DECL
> > +	  || ! TREE_STATIC (decl))
> > +      && TREE_USED (decl)
> > +      && TREE_SYMBOL_REFERENCED (DECL_ASSEMBLER_NAME (decl)))
> > +    warning_with_decl (decl, "weak declaration of `%s' after first use
> > may result in unspecified behaviour");
>
> I intended that TREE_SYMBOL_REFERENCED should be used rather than
> TREE_USED, not in addition.  I'd probably check DECL_ASSEMBLER_NAME_SET_P
> here, too; if the name hasn't been generated, it hasn't been referenced.
> Perhaps we want a new macro, say
>
> #define DECL_SYMBOL_REFERENCED(DECL) \
>  (DECL_ASSEMBLER_NAME_SET_P (DECL) \
>   && TREE_SYMBOL_REFERENCED (DECL_ASSEMBLER_NAME (DECL)))

But this would warn for these cases, which I think isn't correct:

// case 1
extern void * ffoo1g (void);
extern void * ffoox1g (void);
extern void * ffoo1g (void)  __attribute__((weak, alias ("ffoox1g")));
void * foo1g (void)
{
  if (ffoo1g)
    ffoo1g ();
  return 0;
}

// case 2
extern int vfoo1i __attribute__((weak));
void * foo1i (void)
{
  return (void *)&vfoo1i;
}
extern int vfoo1i __attribute__((weak));
extern int vfoo1i;

// case 3
extern int vfoo1j __attribute__((weak));
void * foo1j (void)
{
  return (void *)&vfoo1j;
}
extern int vfoo1j;
extern int vfoo1j __attribute__((weak));


Actually, looking at handle_alias_attribute(), TREE_USED is exactly the flag 
to test here. The addition of DECL_ASSEMBLER_NAME_SET_P seems reasonable 
though. What do you think?

Franz.