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 E7D4A38515F4; Thu,  9 Mar 2023 14:56:59 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org E7D4A38515F4
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org;
	s=default; t=1678373819;
	bh=whgBUUqiAhIh1Pvmv7m2eIB0Aoa5nmhrmKwDxTZBmfQ=;
	h=From:To:Subject:Date:In-Reply-To:References:From;
	b=jP9Gz9l7odxBHLPwiP4xvHPohAfn8uubLFQYZVR1wL6V0G+qaQ39lhxW4xggWCW4m
	 KVssYC0inz7/mj5MF5ODjmx3cTo7IwLtco6Lhd0gtD3sW1Se1P7ase/ze8k/8g7mXE
	 lJRZqt6etyQmMj3FcBhd2X9Y33U5S01I58xFAWOY=
From: "jakub at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug c++/109039] [12/13 Regression] Miscompilation with
 no_unique_address and bitfields
Date: Thu, 09 Mar 2023 14:56:59 +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: 12.1.0
X-Bugzilla-Keywords: ABI, wrong-code
X-Bugzilla-Severity: normal
X-Bugzilla-Who: jakub at gcc dot gnu.org
X-Bugzilla-Status: UNCONFIRMED
X-Bugzilla-Resolution: 
X-Bugzilla-Priority: P2
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: 12.3
X-Bugzilla-Flags: 
X-Bugzilla-Changed-Fields: 
Message-ID: <bug-109039-4-fKezCr8roh@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-109039-4@http.gcc.gnu.org/bugzilla/>
References: <bug-109039-4@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
List-Id: <gcc-bugs.sourceware.org>

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=3D109039

--- Comment #2 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
And there doesn't even have to be any NSDMI, even
struct X {
  short x0 : 7;
  short x1 : 8;
  X () : x0 (1), x1 (2) {}
  int get () { return x0 + x1; }
};

struct S {
  [[no_unique_address]] X x;
  char c;
  S () : c (0) {}
};

S s;

int
main ()
{
  if (s.x.x0 !=3D 1 || s.x.x1 !=3D 2 || s.c !=3D 0)
    __builtin_abort ();
}
is miscompiled since that revision.  With sizeof (s) =3D=3D 2 there aren't =
enough
bytes to
fit the 15 bits of bitfields and another byte.  S::x.x0 is on x86_64 at off=
set
0, low 7
bits, then S::x.x1 next 8 bits, and S::c is at offset 1, so overlapping the
high 7 bits of S::x.x1.=