From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:44487) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1bmyzH-0002rD-AM for qemu-devel@nongnu.org; Thu, 22 Sep 2016 04:04:08 -0400 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1bmyzD-000385-Iw for qemu-devel@nongnu.org; Thu, 22 Sep 2016 04:04:07 -0400 Date: Thu, 22 Sep 2016 09:03:58 +0100 From: "Daniel P. Berrange" Message-ID: <20160922080358.GC352@redhat.com> Reply-To: "Daniel P. Berrange" References: <20160921194126.10223-1-marcandre.lureau@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline In-Reply-To: <20160921194126.10223-1-marcandre.lureau@redhat.com> Content-Transfer-Encoding: quoted-printable Subject: Re: [Qemu-devel] [PATCH] qmp: fix object-add assert() without props 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, armbru@redhat.com, qemu-stable@nongnu.org On Wed, Sep 21, 2016 at 11:41:26PM +0400, Marc-Andr=C3=A9 Lureau wrote: > Since commit ad739706bbadee49, user_creatable_add_type() expects to be > given a qdict. However, if object-add is called without props, you reac= h > the assert: "qemu/qom/object_interfaces.c:115: user_creatable_add_type: > Assertion `qdict' failed.", because the qdict isn't created in this > case (it's optional). >=20 > Furthermore, qmp_input_visitor_new() is not meant to be called without = a > dict, and a further commit will assert in this situation. >=20 > If none given, create an empty qdict in qmp to avoid the > user_creatable_add_type() assert(qdict). >=20 > Signed-off-by: Marc-Andr=C3=A9 Lureau > --- > qmp.c | 9 +++++++-- > 1 file changed, 7 insertions(+), 2 deletions(-) >=20 > diff --git a/qmp.c b/qmp.c > index 6733463..8078038 100644 > --- a/qmp.c > +++ b/qmp.c > @@ -665,7 +665,7 @@ void qmp_add_client(const char *protocol, const cha= r *fdname, > void qmp_object_add(const char *type, const char *id, > bool has_props, QObject *props, Error **errp) > { > - const QDict *pdict =3D NULL; > + QDict *pdict; > Visitor *v; > Object *obj; > =20 > @@ -675,14 +675,19 @@ void qmp_object_add(const char *type, const char = *id, > error_setg(errp, QERR_INVALID_PARAMETER_TYPE, "props", "di= ct"); > return; > } > + } else { > + pdict =3D qdict_new(); > } > =20 > - v =3D qmp_input_visitor_new(props, true); > + v =3D qmp_input_visitor_new(QOBJECT(pdict), true); > obj =3D user_creatable_add_type(type, id, pdict, v, errp); > visit_free(v); > if (obj) { > object_unref(obj); > } > + if (!props) { > + qobject_decref(QOBJECT(pdict)); > + } > } > =20 > void qmp_object_del(const char *id, Error **errp) I'd like to see us add a unit test case to cover this scenario. AFAICT, we don't currently have unit tests for specific QMP commands, jus= t the command framework, but we could still do it in tests/test-qmp-command= s.c Regards, Daniel --=20 |: http://berrange.com -o- http://www.flickr.com/photos/dberrange= / :| |: http://libvirt.org -o- http://virt-manager.or= g :| |: http://autobuild.org -o- http://search.cpan.org/~danberr= / :| |: http://entangle-photo.org -o- http://live.gnome.org/gtk-vn= c :|