From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:38283) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1fByWr-0005sJ-Jb for qemu-devel@nongnu.org; Fri, 27 Apr 2018 04:14:54 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1fByWn-0003XH-Il for qemu-devel@nongnu.org; Fri, 27 Apr 2018 04:14:53 -0400 Received: from mx3-rdu2.redhat.com ([66.187.233.73]:35902 helo=mx1.redhat.com) by eggs.gnu.org with esmtps (TLS1.0:DHE_RSA_AES_256_CBC_SHA1:32) (Exim 4.71) (envelope-from ) id 1fByWn-0003Wy-DR for qemu-devel@nongnu.org; Fri, 27 Apr 2018 04:14:49 -0400 Received: from smtp.corp.redhat.com (int-mx05.intmail.prod.int.rdu2.redhat.com [10.11.54.5]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mx1.redhat.com (Postfix) with ESMTPS id E2325EC01A for ; Fri, 27 Apr 2018 08:14:45 +0000 (UTC) From: Markus Armbruster References: <20180419150145.24795-1-marcandre.lureau@redhat.com> <20180419150145.24795-2-marcandre.lureau@redhat.com> Date: Fri, 27 Apr 2018 10:14:40 +0200 In-Reply-To: <20180419150145.24795-2-marcandre.lureau@redhat.com> (=?utf-8?Q?=22Marc-Andr=C3=A9?= Lureau"'s message of "Thu, 19 Apr 2018 17:01:41 +0200") Message-ID: <87tvrxqd5b.fsf@dusky.pond.sub.org> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable Subject: Re: [Qemu-devel] [PATCH v6 1/5] qobject: ensure base is at offset 0 List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: =?utf-8?Q?Marc-Andr=C3=A9?= Lureau Cc: qemu-devel@nongnu.org, pbonzini@redhat.com Marc-Andr=C3=A9 Lureau writes: > All QObject types have the base QObject as their first field. This > allows the simplification of qobject_to(). > > This explicitly guarantees that existing casts work correctly (even > though we'd prefer to get rid of such casts in any location except the > qobject.h macros) In my opinion, type casts between QObject * and subtypes are plain wrong. I intend to apply my "[PATCH] qobject: Use qobject_to() instead of type cast" first, and drop the paragraph, if you don't mind. > Signed-off-by: Marc-Andr=C3=A9 Lureau > Reviewed-by: Eric Blake > --- > include/qapi/qmp/qobject.h | 5 ++--- > qobject/qobject.c | 9 +++++++++ > 2 files changed, 11 insertions(+), 3 deletions(-) > > diff --git a/include/qapi/qmp/qobject.h b/include/qapi/qmp/qobject.h > index e022707578..5206ff9ee1 100644 > --- a/include/qapi/qmp/qobject.h > +++ b/include/qapi/qmp/qobject.h > @@ -61,9 +61,8 @@ struct QObject { > QEMU_BUILD_BUG_MSG(QTYPE__MAX !=3D 7, > "The QTYPE_CAST_TO_* list needs to be extended"); >=20=20 > -#define qobject_to(type, obj) ({ \ > - QObject *_tmp =3D qobject_check_type(obj, glue(QTYPE_CAST_TO_, type)= ); \ > - _tmp ? container_of(_tmp, type, base) : (type *)NULL; }) > +#define qobject_to(type, obj) \ > + ((type *)qobject_check_type(obj, glue(QTYPE_CAST_TO_, type))) >=20=20 > /* Initialize an object to default values */ > static inline void qobject_init(QObject *obj, QType type) > diff --git a/qobject/qobject.c b/qobject/qobject.c > index 23600aa1c1..87649c5be5 100644 > --- a/qobject/qobject.c > +++ b/qobject/qobject.c > @@ -16,6 +16,15 @@ > #include "qapi/qmp/qlist.h" > #include "qapi/qmp/qstring.h" >=20=20 > +QEMU_BUILD_BUG_MSG( > + offsetof(QNull, base) !=3D 0 || > + offsetof(QNum, base) !=3D 0 || > + offsetof(QString, base) !=3D 0 || > + offsetof(QDict, base) !=3D 0 || > + offsetof(QList, base) !=3D 0 || > + offsetof(QBool, base) !=3D 0, > + "base qobject must be at offset 0"); > + > static void (*qdestroy[QTYPE__MAX])(QObject *) =3D { > [QTYPE_NONE] =3D NULL, /* No such object exists */ > [QTYPE_QNULL] =3D NULL, /* qnull_ is indestructible */ As mentioned elsewhere in this thread, the coding errors this assertion can catch are vanishingly unlikely. I could flip a coin to decide whether to keep it. Instead I'm asking you :) With the commit message rephrased not to give anyone ideas about type casts, and with or without the assertion: Reviewed-by: Markus Armbruster