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 687ED3858C5F; Fri,  1 Mar 2024 13:37:38 +0000 (GMT)
DKIM-Filter: OpenDKIM Filter v2.11.0 sourceware.org 687ED3858C5F
DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gcc.gnu.org;
	s=default; t=1709300258;
	bh=YXJlXSInzKO2MRygQuGoAriqDIsNf58ed6cnpYm0S6w=;
	h=From:To:Subject:Date:In-Reply-To:References:From;
	b=pceqMMt1KnSWlGIwmWT4vzk/6d06ioCTs/rlGIDKQY/KnmP+8VMfjKSHPqbRjq1Wf
	 +dQeI0QytcuW1IrCFf4YqzLoVq69v4saTrRCBM3yOiHW66UQ9NRoOSs77tFwSAiR1B
	 uePbUCjLr2wk/J3IfhlHeWPUyEMsOQ8LEh+KWQ0U=
From: "rguenth at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug d/114171] [13/14 Regression] gdc -O2 -mavx generates misaligned
 vmovdqa instruction
Date: Fri, 01 Mar 2024 13:37:37 +0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: d
X-Bugzilla-Version: 14.0
X-Bugzilla-Keywords: 
X-Bugzilla-Severity: normal
X-Bugzilla-Who: rguenth at gcc dot gnu.org
X-Bugzilla-Status: UNCONFIRMED
X-Bugzilla-Resolution: 
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: ibuclaw at gdcproject dot org
X-Bugzilla-Target-Milestone: 13.3
X-Bugzilla-Flags: 
X-Bugzilla-Changed-Fields: 
Message-ID: <bug-114171-4-hEpUMkb9Ey@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-114171-4@http.gcc.gnu.org/bugzilla/>
References: <bug-114171-4@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=3D114171

--- Comment #5 from Richard Biener <rguenth at gcc dot gnu.org> ---
(In reply to Iain Buclaw from comment #4)
> Removed druntime dependency.
> ---
> import gcc.builtins;
> struct Token {
>     string label;
> }
> struct BreakStatement {
>     ulong pad;
>     Token label;
> }
>=20
> pragma(inline, false)
> auto newclass()
> {
>     void *p =3D __builtin_malloc(BreakStatement.sizeof);
>     __builtin_memset(p, 0, BreakStatement.sizeof);
>     return cast(BreakStatement*) p;
> }
>=20
> int main ()
> {
>     auto bn =3D newclass();
>     return bn.label is Token.init;
> }
> ---
>=20
>=20
>=20
> Roughly the equivalent C++
> ---
> struct Token {
>     struct {
>     __SIZE_TYPE__ length;
>     const char *ptr;
>     } label;
> };
> struct BreakStatement {
>     __UINT64_TYPE__ pad;
>     Token label;
> };
>=20
> __attribute__((noinline))
> BreakStatement *newclass()
> {
>     void *p =3D __builtin_malloc(sizeof(BreakStatement));
>     __builtin_memset(p, 0, sizeof(BreakStatement));
>     return (BreakStatement*) p;
> }
>=20
> int main ()
> {
>     auto bn =3D newclass();
>     auto init =3D Token();
>     return *(__uint128_t*)&bn->label =3D=3D *(__uint128_t*)&init;
> }
> ---

Unless gdc somehow guarantees bn->label and init are 128bit aligned
then "casting" this way is broken.  You can of course use
build_aligned_type to build a properly (mis-)aligned type to use
to dereference to.=