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 6EED23842414; Tue, 16 Feb 2021 11:01:47 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 6EED23842414
From: "redi at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug libstdc++/99117] [9/10/11 Regression] cannot accumulate
 std::valarray
Date: Tue, 16 Feb 2021 11:01:47 +0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: libstdc++
X-Bugzilla-Version: 10.1.0
X-Bugzilla-Keywords: needs-reduction, wrong-code
X-Bugzilla-Severity: normal
X-Bugzilla-Who: redi at gcc dot gnu.org
X-Bugzilla-Status: NEW
X-Bugzilla-Resolution: 
X-Bugzilla-Priority: P2
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: 9.4
X-Bugzilla-Flags: 
X-Bugzilla-Changed-Fields: 
Message-ID: <bug-99117-4-2ueSn8udHw@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-99117-4@http.gcc.gnu.org/bugzilla/>
References: <bug-99117-4@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: <https://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: <https://gcc.gnu.org/mailman/listinfo/gcc-bugs>,
 <mailto:gcc-bugs-request@gcc.gnu.org?subject=subscribe>
X-List-Received-Date: Tue, 16 Feb 2021 11:01:47 -0000

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=3D99117

--- Comment #4 from Jonathan Wakely <redi at gcc dot gnu.org> ---
The problem is that sum + e returns an expression template that holds
references to its operands, so that the sum is done lazily.

When that expression template result is assigned back to sum it evaluates t=
he
sum of each element, which is basically:

struct valarray {
  int* __restrict__ _M_data;
  size_t _M_size;
};
valarray sum{ new int[2]{1,1}, 2 };
valarray rhs{ new int[2]{2,2}, 2 };
int* p =3D sum._M_data;
int* e1 =3D sum._M_data;
int* e2 =3D rhs._M_data;
for (size_t i =3D 0; i < sum._M_size; ++i, ++p)
  *p =3D e1[i] + e2[i];

Where p and e1 alias, but are marked with __restrict__.

Maybe we need to check for such aliasing in __valarray_copy.=