From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <scottwood@freescale.com>
Received: from az33egw02.freescale.net (az33egw02.freescale.net
	[192.88.158.103])
	(using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits))
	(Client CN "az33egw02.freescale.net",
	Issuer "Thawte Premium Server CA" (verified OK))
	by ozlabs.org (Postfix) with ESMTPS id B40B5DE1CC
	for <linuxppc-dev@ozlabs.org>; Sat, 11 Apr 2009 04:05:18 +1000 (EST)
Message-ID: <49DF89F7.2070403@freescale.com>
Date: Fri, 10 Apr 2009 13:03:35 -0500
From: Scott Wood <scottwood@freescale.com>
MIME-Version: 1.0
To: Segher Boessenkool <segher@kernel.crashing.org>
Subject: Re: [PATCH] Quieten arch/powerpc in a allmodconfig build.
References: <2c4bcf8d1d7083ff53ce5b556765e96676a007fb.1239165378.git.tony@bakeyournoodle.com>	<1239167335.10104.26.camel@localhost>	<20090408055126.GG16602@bilbo.ozlabs.org>	<7ED93BB3-FE0E-4102-89FB-78950A533717@kernel.crashing.org>	<20090409224553.GO16602@bilbo.ozlabs.org>	<20090410091149.5068d8e1.sfr@canb.auug.org.au>
	<1AAD5906-66FA-4850-8C62-8C030816E44E@kernel.crashing.org>
In-Reply-To: <1AAD5906-66FA-4850-8C62-8C030816E44E@kernel.crashing.org>
Content-Type: text/plain; charset=UTF-8; format=flowed
Cc: Stephen Rothwell <sfr@canb.auug.org.au>, linuxppc-dev@ozlabs.org
List-Id: Linux on PowerPC Developers Mail List <linuxppc-dev.ozlabs.org>
List-Unsubscribe: <https://ozlabs.org/mailman/options/linuxppc-dev>,
	<mailto:linuxppc-dev-request@ozlabs.org?subject=unsubscribe>
List-Archive: <http://ozlabs.org/pipermail/linuxppc-dev>
List-Post: <mailto:linuxppc-dev@ozlabs.org>
List-Help: <mailto:linuxppc-dev-request@ozlabs.org?subject=help>
List-Subscribe: <https://ozlabs.org/mailman/listinfo/linuxppc-dev>,
	<mailto:linuxppc-dev-request@ozlabs.org?subject=subscribe>

Segher Boessenkool wrote:
>> Unfortunately -Wno-uninitialized also suppresses the warnings that point
>> to real bugs.
> 
> It's a double-edged sword, yes.  Warnings are always like that:
> if the compiler could know that something _is_ wrong for certain,
> it wouldn't need a warning (it would use an error, instead -- and
> it does do this in certain cases); if it would know something is
> not really wrong, it would just shut up.

The problem is that GCC does not give an error (only a warning) even for 
things like this where it should be trivial to detect that the usage 
*is* uninitialized, not just might be:

int foo(void)
{
    int a;

    return a;
}

And further, there is no separation of warning classes into 
might-be-uninitialized and is-uninitialized-compiler-can-tell-for-sure.

In other words, there should be a way to tell the compiler to err on the 
side of not complaining if it's unsure, but still report the obvious 
ones (or make the latter an error but the former a warning).  That's not 
ESP or DWIM.

-Scott