From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-bugs-return-475342-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 26983 invoked by alias); 29 Jan 2015 00:05:00 -0000
Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-bugs.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-bugs/>
List-Post: <mailto:gcc-bugs@gcc.gnu.org>
List-Help: <mailto:gcc-bugs-help@gcc.gnu.org>
Sender: gcc-bugs-owner@gcc.gnu.org
Received: (qmail 26793 invoked by uid 48); 29 Jan 2015 00:04:53 -0000
From: "richard-gccbugzilla at metafoo dot co.uk" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug c/64843] miscompilation of atomic_fetch_add on atomic pointer type
Date: Thu, 29 Jan 2015 00:05:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: c
X-Bugzilla-Version: 4.9.0
X-Bugzilla-Keywords: documentation, wrong-code
X-Bugzilla-Severity: normal
X-Bugzilla-Who: richard-gccbugzilla at metafoo dot co.uk
X-Bugzilla-Status: NEW
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags:
X-Bugzilla-Changed-Fields:
Message-ID: <bug-64843-4-zvbPfjWk50@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-64843-4@http.gcc.gnu.org/bugzilla/>
References: <bug-64843-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 7bit
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
X-SW-Source: 2015-01/txt/msg03336.txt.bz2

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64843

--- Comment #5 from Richard Smith <richard-gccbugzilla at metafoo dot co.uk> ---
(In reply to joseph@codesourcery.com from comment #3)
> The first question is whether this code is actually valid.

To my reading, the C11 standard says that these operations must work on all
atomic integer types, must be disallowed on atomic_bool, and it's unspecified
if they work on any other type. However, since GCC allows them for an
atomic-qualified pointer, the other rules in 7.17.7.5 apply.

I'm not at all confident that is the intended reading, though: in the C++
specification, from which the C specification was derived, there are overloads
of atomic_fetch_add for each atomic-integral type, plus overloads for
atomic<T*>, making it mandatory to handle the pointer case. I think this is
merely a bug in C11's spec, and this wording should also allow pointers, but
AFAICS it's not in C11's DR list.