From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <gcc-bugzilla@gcc.gnu.org>
Received: by sourceware.org (Postfix, from userid 48)
 id 13AA4385DC14; Wed, 15 Apr 2020 06:46:17 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 13AA4385DC14
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org;
 s=default; t=1586933177;
 bh=V0bW8ty+HbtNzdR4niLURIekI2jeCYGPVsJBL7CTsrI=;
 h=From:To:Subject:Date:In-Reply-To:References:From;
 b=PKsmvlDExE8CyY0Fp2kKPGWZU1Czm9U/sPMvhwHn4cvYgNJwDjdWTni5onqQQrw4E
 1L11xK/jp+eeXLhjACwIW67m6nMGOVlI3fuJEJXtODzR0DjD2Ui2VYmtuAAG6k7JIL
 ziw9LOndIg45vrBCvGdGq7EKseIZm3T+ktGSl0Og=
From: "rguenth at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug middle-end/94600] Ignored volatile specifier on loop unrolling
 and bitfield misoptimization
Date: Wed, 15 Apr 2020 06:46:16 +0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: middle-end
X-Bugzilla-Version: 10.0
X-Bugzilla-Keywords: wrong-code
X-Bugzilla-Severity: normal
X-Bugzilla-Who: rguenth at gcc dot gnu.org
X-Bugzilla-Status: NEW
X-Bugzilla-Resolution: 
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-94600-4-QVMc19HQOV@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-94600-4@http.gcc.gnu.org/bugzilla/>
References: <bug-94600-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
X-BeenThere: gcc-bugs@gcc.gnu.org
X-Mailman-Version: 2.1.29
Precedence: list
List-Id: Gcc-bugs mailing list <gcc-bugs.gcc.gnu.org>
List-Unsubscribe: <http://gcc.gnu.org/mailman/options/gcc-bugs>,
 <mailto:gcc-bugs-request@gcc.gnu.org?subject=unsubscribe>
List-Archive: <https://gcc.gnu.org/pipermail/gcc-bugs/>
List-Post: <mailto:gcc-bugs@gcc.gnu.org>
List-Help: <mailto:gcc-bugs-request@gcc.gnu.org?subject=help>
List-Subscribe: <http://gcc.gnu.org/mailman/listinfo/gcc-bugs>,
 <mailto:gcc-bugs-request@gcc.gnu.org?subject=subscribe>
X-List-Received-Date: Wed, 15 Apr 2020 06:46:17 -0000

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=3D94600

--- Comment #3 from Richard Biener <rguenth at gcc dot gnu.org> ---
Confirmed on arm.  The odd thing is that the optimized GIMPLE for foo() is
much nicer:

foo ()
{
  <bb 2> [local count: 153437707]:
  MEM[(volatile struct t0 *)655404B] =3D{v} a0[0];
  MEM[(volatile struct t0 *)655408B] =3D{v} a0[1];
  MEM[(volatile struct t0 *)655412B] =3D{v} a0[2];
  MEM[(volatile struct t0 *)655416B] =3D{v} a0[3];
  MEM[(volatile struct t0 *)655420B] =3D{v} a0[4];
  MEM[(volatile struct t0 *)655424B] =3D{v} a0[5];
  return;

while for bar() we have additional stmts to construct the stack objects
though the actual stores are the same:

  <bb 2> [local count: 1073741824]:
  a01 =3D {};
  a02 =3D {};
  a03 =3D {};
  a04 =3D {};
  a05 =3D {};
  MEM[(struct  *)&a00] =3D 33556023;
  MEM[(volatile struct t0 *)655404B] =3D{v} a00;
  _38 =3D MEM[(struct  *)&a01];
  _39 =3D _38 & 33521664;
  _40 =3D _39 | 33558455;
  MEM[(struct  *)&a01] =3D _40;
  MEM[(volatile struct t0 *)655408B] =3D{v} a01;
  _41 =3D MEM[(struct  *)&a02];
  _42 =3D _41 & 33521664;
  _43 =3D _42 | 167774200;
  MEM[(struct  *)&a02] =3D _43;
  MEM[(volatile struct t0 *)655412B] =3D{v} a02;
  _44 =3D MEM[(struct  *)&a03];
  _45 =3D _44 & 33521664;
  _46 =3D _45 | 33554453;
  MEM[(struct  *)&a03] =3D _46;
  MEM[(volatile struct t0 *)655416B] =3D{v} a03;
  _47 =3D MEM[(struct  *)&a04];
  _48 =3D _47 & 33521664;
  _49 =3D _48 | 33554453;
  MEM[(struct  *)&a04] =3D _49;
  MEM[(volatile struct t0 *)655420B] =3D{v} a04;
  _50 =3D MEM[(struct  *)&a05];
  _51 =3D _50 & 33521664;
  _52 =3D _51 | 33554453;
  MEM[(struct  *)&a05] =3D _52;
  MEM[(volatile struct t0 *)655424B] =3D{v} a05;
  a00 =3D{v} {CLOBBER};
  a01 =3D{v} {CLOBBER};
  a02 =3D{v} {CLOBBER};
  a03 =3D{v} {CLOBBER};
  a04 =3D{v} {CLOBBER};
  a05 =3D{v} {CLOBBER};
  return;

I suspect that the RTL expansion for foo() applies "premature" optimization
via constant folding the aggregate initializer and thus compiling this into
partly initializing the destination which isn't valid for volatile qualified
LHSs.  For bar() there's nothing to optimize for it.=