From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:54636) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1etw1B-0003Yc-TN for qemu-devel@nongnu.org; Thu, 08 Mar 2018 08:55:39 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1etw18-0000Ce-0d for qemu-devel@nongnu.org; Thu, 08 Mar 2018 08:55:37 -0500 Received: from mx3-rdu2.redhat.com ([66.187.233.73]:48766 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 1etw17-0000CD-RK for qemu-devel@nongnu.org; Thu, 08 Mar 2018 08:55:33 -0500 Received: from smtp.corp.redhat.com (int-mx04.intmail.prod.int.rdu2.redhat.com [10.11.54.4]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mx1.redhat.com (Postfix) with ESMTPS id 264054075168 for ; Thu, 8 Mar 2018 13:55:29 +0000 (UTC) Date: Thu, 8 Mar 2018 21:55:22 +0800 From: Peter Xu Message-ID: <20180308135521.GH32252@xz-mi> References: <20180306053320.15401-1-peterx@redhat.com> <20180306053320.15401-10-peterx@redhat.com> <20180307123650.GD20201@redhat.com> <20180308034409.GA32252@xz-mi> <20180308101359.GE4718@redhat.com> <20180308114213.GE32252@xz-mi> <20180308133143.GP4718@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline In-Reply-To: <20180308133143.GP4718@redhat.com> Content-Transfer-Encoding: quoted-printable Subject: Re: [Qemu-devel] [PATCH v2 9/9] chardev: tcp: postpone TLS work until machine done List-Id: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , To: Daniel =?utf-8?B?UC4gQmVycmFuZ8Op?= Cc: qemu-devel@nongnu.org, Paolo Bonzini , Stefan Hajnoczi , =?utf-8?Q?Marc-Andr=C3=A9?= Lureau On Thu, Mar 08, 2018 at 01:31:43PM +0000, Daniel P. Berrang=C3=A9 wrote: > On Thu, Mar 08, 2018 at 07:42:13PM +0800, Peter Xu wrote: > > On Thu, Mar 08, 2018 at 10:13:59AM +0000, Daniel P. Berrang=C3=A9 wro= te: > > > On Thu, Mar 08, 2018 at 11:44:09AM +0800, Peter Xu wrote: > > > > On Wed, Mar 07, 2018 at 12:36:50PM +0000, Daniel P. Berrang=C3=A9= wrote: > > > >=20 > > > > [...] > > > >=20 > > > > > > diff --git a/chardev/char-socket.c b/chardev/char-socket.c > > > > > > index bd40864f87..997c70dd7d 100644 > > > > > > --- a/chardev/char-socket.c > > > > > > +++ b/chardev/char-socket.c > > > > > > @@ -31,6 +31,7 @@ > > > > > > #include "qemu/option.h" > > > > > > #include "qapi/error.h" > > > > > > #include "qapi/clone-visitor.h" > > > > > > +#include "sysemu/sysemu.h" > > > > > > =20 > > > > > > #include "chardev/char-io.h" > > > > > > =20 > > > > > > @@ -722,6 +723,11 @@ static void tcp_chr_tls_init(Chardev *ch= r) > > > > > > Error *err =3D NULL; > > > > > > gchar *name; > > > > > > =20 > > > > > > + if (!machine_init_done) { > > > > > > + /* This will be postponed to machine_done notifier *= / > > > > > > + return; > > > > > > + } > > > > > > + > > > > > > if (s->is_listen) { > > > > > > tioc =3D qio_channel_tls_new_server( > > > > > > s->ioc, s->tls_creds, > > > > > > @@ -1145,6 +1151,10 @@ static int tcp_chr_machine_done_hook(C= hardev *chr) > > > > > > tcp_chr_connect_async(chr); > > > > > > } > > > > > > =20 > > > > > > + if (s->tls_creds) { > > > > > > + tcp_chr_tls_init(chr); > > > > > > + } > > > > >=20 > > > > > This looks questionable - AFAICT, there's no guarantee we have = any > > > > > client connection active when the machine dnoe hook runs. Only = if > > > > > the chardev is set in client mode, and reconnect_time is *not* = set, > > > > > but this seems to be run unconditionally. > > > >=20 > > > > You are right. Thanks for spotting that. > > > >=20 > > > > Then how about this? It's a bit ugly, but I think it should be s= afe: > > >=20 > > > Is it perhaps not possible to just check if 's->ioc' is non-NULL > > > in the tcp_chr_machine_done_hook for your original patch ? > >=20 > > In tcp_chr_new_client() I see that s->ioc will be set no matter what, > > and that function can even be called without TLS enabled I think. > > Does that mean only check against s->ioc would not be enough? >=20 > I mean like this: >=20 > diff --git a/chardev/char-socket.c b/chardev/char-socket.c > index a1966aae51..19e3193817 100644 > --- a/chardev/char-socket.c > +++ b/chardev/char-socket.c > @@ -723,6 +723,11 @@ static void tcp_chr_tls_init(Chardev *chr) > Error *err =3D NULL; > gchar *name; > =20 > + if (!machine_init_done) { > + /* This will be postponed to machine_done notifier */ > + return; > + } > + > if (s->is_listen) { > tioc =3D qio_channel_tls_new_server( > s->ioc, s->tls_creds, > @@ -1146,6 +1151,10 @@ static int tcp_chr_machine_done_hook(Chardev *ch= r) > tcp_chr_connect_async(chr); > } > =20 > + if (s->ioc && s->tls_creds) { > + tcp_chr_tls_init(chr); > + } > + > return 0; > } > =20 >=20 >=20 > s->ioc will only be !=3D NULL, if some codepath in qemu_chr_parse_socke= t I guess here you mean qmp_chardev_open_socket(). > ended up calling tcp_chr_new_client(). If that happened we will have > skipped setup of TLS, so calling tcp_chr_tls_init() based on > s->ioc && s->tls_creds feels right to me. Yes, it looks right to me too. :) I'll post a complete patch soon with your authorship. Thanks! --=20 Peter Xu