Received: from eggs.gnu.org ([2001:4830:134:3::10]:51440) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1aMF9K-00059u-Ff for qemu-devel@nongnu.org; Thu, 21 Jan 2016 08:19:43 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1aMF9H-0003Ws-5F for qemu-devel@nongnu.org; Thu, 21 Jan 2016 08:19:42 -0500 From: Markus Armbruster References: <1453219845-30939-1-git-send-email-eblake@redhat.com> <1453219845-30939-20-git-send-email-eblake@redhat.com> <87powvi5x7.fsf@blackfin.pond.sub.org> Date: Thu, 21 Jan 2016 14:19:35 +0100 In-Reply-To: <87powvi5x7.fsf@blackfin.pond.sub.org> (Markus Armbruster's message of "Thu, 21 Jan 2016 11:27:16 +0100") Message-ID: <87bn8ff4t4.fsf@blackfin.pond.sub.org> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable Subject: Re: [Qemu-devel] [PATCH v9 19/37] qmp: Fix reference-counting of qnull on empty output visit List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Eric Blake Cc: marcandre.lureau@redhat.com, qemu-stable@nongnu.org, qemu-devel@nongnu.org, Michael Roth Markus Armbruster writes: > Eric Blake writes: > >> Commit 6c2f9a15 ensured that we would not return NULL when the >> caller used an output visitor but had nothing to visit. But >> in doing so, it added a FIXME about a reference count leak >> that could abort qemu in the (unlikely) case of SIZE_MAX such >> visits (more plausible on 32-bit). (Although that commit >> suggested we might fix it in time for 2.5, we ran out of time; >> fortunately, it is unlikely enough to bite that it was not >> worth worrying about during the 2.5 release.) >> >> This fixes things by documenting the internal contracts, and >> explaining why the internal function can return NULL and only >> the public facing interface needs to worry about qnull(), >> thus avoiding over-referencing the qnull_ global object. >> >> It does not, however, fix the stupidity of the stack mixing >> up two separate pieces of information; add a FIXME to explain >> that issue, which will be fixed shortly in a future patch. >> >> Signed-off-by: Eric Blake >> Cc: qemu-stable@nongnu.org >> Reviewed-by: Marc-Andr=C3=A9 Lureau >> >> --- >> v9: enhance commit message >> v8: rebase to earlier changes >> v7: cc qemu-stable, tweak some asserts, drop stale comment, add more >> comments >> v6: no change >> --- >> qapi/qmp-output-visitor.c | 39 ++++++++++++++++++++++++++++++++--= ----- >> tests/test-qmp-output-visitor.c | 2 ++ >> 2 files changed, 34 insertions(+), 7 deletions(-) >> >> diff --git a/qapi/qmp-output-visitor.c b/qapi/qmp-output-visitor.c >> index d367148..316f4e4 100644 >> --- a/qapi/qmp-output-visitor.c >> +++ b/qapi/qmp-output-visitor.c [...] >> @@ -41,10 +50,12 @@ static QmpOutputVisitor *to_qov(Visitor *v) >> return container_of(v, QmpOutputVisitor, visitor); >> } >> >> +/* Push @value onto the stack of current QObjects being built */ >> static void qmp_output_push_obj(QmpOutputVisitor *qov, QObject *value) >> { >> QStackEntry *e =3D g_malloc0(sizeof(*e)); >> >> + assert(value); >> e->value =3D value; >> if (qobject_type(e->value) =3D=3D QTYPE_QLIST) { >> e->is_list_head =3D true; >> @@ -52,44 +63,51 @@ static void qmp_output_push_obj(QmpOutputVisitor *qo= v, QObject *value) >> QTAILQ_INSERT_HEAD(&qov->stack, e, node); >> } >> >> +/* Grab and remove the most recent QObject from the stack */ Let's stick to the stack terminology you used for qmp_output_push_obj(): /* Pop a value off the stack of QObjects being built, and return it */ >> static QObject *qmp_output_pop(QmpOutputVisitor *qov) >> { >> QStackEntry *e =3D QTAILQ_FIRST(&qov->stack); >> QObject *value; >> + >> + assert(e); >> QTAILQ_REMOVE(&qov->stack, e, node); >> value =3D e->value; >> g_free(e); >> return value; > > @value cannot be null here, because we never push nulls. Worth an > assertion, just to help readers? > >> } [...] >> +/* Grab the most recent QObject from the stack, which must exist */ Stack terminology: /* * Peek at the top of the stack of QObject being built. * The stack must not be empty. */ >> static QObject *qmp_output_last(QmpOutputVisitor *qov) >> { >> QStackEntry *e =3D QTAILQ_FIRST(&qov->stack); >> + >> + assert(e && e->value); >> return e->value; >> } [...]