From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-patches-return-289489-listarch-gcc-patches=gcc.gnu.org@gcc.gnu.org>
Received: (qmail 18721 invoked by alias); 14 Apr 2011 15:48:28 -0000
Received: (qmail 18710 invoked by uid 22791); 14 Apr 2011 15:48:27 -0000
X-SWARE-Spam-Status: No, hits=-2.3 required=5.0	tests=AWL,BAYES_00,DKIM_SIGNED,DKIM_VALID,DKIM_VALID_AU,FREEMAIL_FROM,RCVD_IN_DNSWL_LOW
X-Spam-Check-By: sourceware.org
Received: from mail-qy0-f175.google.com (HELO mail-qy0-f175.google.com) (209.85.216.175)    by sourceware.org (qpsmtpd/0.43rc1) with ESMTP; Thu, 14 Apr 2011 15:48:21 +0000
Received: by qyk35 with SMTP id 35so3742589qyk.20        for <gcc-patches@gcc.gnu.org>; Thu, 14 Apr 2011 08:48:20 -0700 (PDT)
MIME-Version: 1.0
Received: by 10.229.29.5 with SMTP id o5mr665431qcc.141.1302796100024; Thu, 14 Apr 2011 08:48:20 -0700 (PDT)
Received: by 10.229.190.140 with HTTP; Thu, 14 Apr 2011 08:48:19 -0700 (PDT)
In-Reply-To: <Pine.LNX.4.64.1104141721300.1989@wotan.suse.de>
References: <20110414133453.GA18805@intel.com>	<BANLkTim2=MvyCPV8MfSFNyQ6yd4W5NUFrA@mail.gmail.com>	<BANLkTi=tm2Lw0cASOmAKdiSgpvnkZfG=Cw@mail.gmail.com>	<BANLkTim0gCb+_BQNAf2ORWpnYOcmQ4Ew9w@mail.gmail.com>	<BANLkTinqEzz9iyzJXvSqOWhZYoQec_RpKg@mail.gmail.com>	<BANLkTikK-8RfCm6v3gXDnJjcOnFxpwAcag@mail.gmail.com>	<Pine.LNX.4.64.1104141721300.1989@wotan.suse.de>
Date: Thu, 14 Apr 2011 15:48:00 -0000
Message-ID: <BANLkTi=p4eoyxFxj4JTXNWcfZyONLTnWjA@mail.gmail.com>
Subject: Re: PATCH: PR middle-end/48608: Alignment adjust of local variables is lost
From: "H.J. Lu" <hjl.tools@gmail.com>
To: Michael Matz <matz@suse.de>
Cc: Richard Guenther <richard.guenther@gmail.com>, GCC Patches <gcc-patches@gcc.gnu.org>
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: quoted-printable
X-IsSubscribed: yes
Mailing-List: contact gcc-patches-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-patches.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-patches/>
List-Post: <mailto:gcc-patches@gcc.gnu.org>
List-Help: <mailto:gcc-patches-help@gcc.gnu.org>
Sender: gcc-patches-owner@gcc.gnu.org
X-SW-Source: 2011-04/txt/msg01093.txt.bz2

On Thu, Apr 14, 2011 at 8:26 AM, Michael Matz <matz@suse.de> wrote:
> Hi,
>
> On Thu, 14 Apr 2011, Richard Guenther wrote:
>
>> > + =A0if (align > DECL_ALIGN (decl))
>> > + =A0 =A0DECL_ALIGN (decl) =3D align;
>>
>> Shouldn't this unconditionally set DECL_ALIGN in case
>> LOCAL_DECL_ALINGMENT returns something smaller?
>
> Decreasing alignment of DECLs points to a problem elsewhere, so perhaps an
> assert that this doesn't happen is better. =A0Decreasing is a problem
> because it's not conservative: there might have been code generated
> already assuming the once larger alignment that then possibly breaks if it
> turns out the alignment is actually smaller.

ia32 may decrease local variable alignment:

  /* Don't do dynamic stack realignment for long long objects with
     -mpreferred-stack-boundary=3D2.  */
  if (!TARGET_64BIT
      && align =3D=3D 64
      && ix86_preferred_stack_boundary < 64
      && (mode =3D=3D DImode || (type && TYPE_MODE (type) =3D=3D DImode))
      && (!type || !TYPE_USER_ALIGN (type))
      && (!decl || !DECL_USER_ALIGN (decl)))
    align =3D 32;

I am running bootstrap/test on Linux/x86-64 and Linux/ia32.

--=20
H.J.