From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from eggs.gnu.org ([2001:4830:134:3::10]:35875) by lists.gnu.org with esmtp (Exim 4.71) (envelope-from ) id 1erFuY-0006Gl-1L for qemu-devel@nongnu.org; Wed, 28 Feb 2018 23:33:43 -0500 Received: from Debian-exim by eggs.gnu.org with spam-scanned (Exim 4.71) (envelope-from ) id 1erFuT-0002N4-HS for qemu-devel@nongnu.org; Wed, 28 Feb 2018 23:33:42 -0500 Received: from mx3-rdu2.redhat.com ([66.187.233.73]:35920 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 1erFuT-0002Mg-Cq for qemu-devel@nongnu.org; Wed, 28 Feb 2018 23:33:37 -0500 Received: from smtp.corp.redhat.com (int-mx03.intmail.prod.int.rdu2.redhat.com [10.11.54.3]) (using TLSv1.2 with cipher AECDH-AES256-SHA (256/256 bits)) (No client certificate requested) by mx1.redhat.com (Postfix) with ESMTPS id BD3028182D2E for ; Thu, 1 Mar 2018 04:33:33 +0000 (UTC) Date: Thu, 1 Mar 2018 12:33:22 +0800 From: Peter Xu Message-ID: <20180301043322.GO27381@xz-mi> References: <20180228050633.7410-1-peterx@redhat.com> <20180228050633.7410-5-peterx@redhat.com> <20180228091058.GC31550@redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Disposition: inline In-Reply-To: <20180228091058.GC31550@redhat.com> Content-Transfer-Encoding: quoted-printable Subject: Re: [Qemu-devel] [PATCH 04/14] migration: let incoming side use thread context 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 , Juan Quintela , Markus Armbruster , =?utf-8?Q?Marc-Andr=C3=A9?= Lureau , Stefan Hajnoczi , "Dr . David Alan Gilbert" , Laurent Vivier On Wed, Feb 28, 2018 at 09:10:58AM +0000, Daniel P. Berrang=C3=A9 wrote: > On Wed, Feb 28, 2018 at 01:06:23PM +0800, Peter Xu wrote: > > The old incoming migration is running in main thread and default > > gcontext. With the new qio_channel_add_watch_full() we can now let i= t > > run in the thread's own gcontext (if there is one). > >=20 > > Currently this patch does nothing alone. But when any of the incomin= g > > migration is run in another iothread (e.g., the upcoming migrate-reco= ver > > command), this patch will bind the incoming logic to the iothread > > instead of the main thread (which may already get page faulted and > > hanged). > >=20 > > RDMA is not considered for now since it's not even using the QIO APIs= at > > all. >=20 > Errm, yes, it is. >=20 > struct QIOChannelRDMA { > QIOChannel parent; > RDMAContext *rdma; > QEMUFile *file; > size_t len; > bool blocking; /* XXX we don't actually honour this yet */ > }; > .... Ah, you are right. :) I should say that "it's not using QIO watch framework" since it's using qemu_set_fd_handler() so it's always on main thread. Thanks, --=20 Peter Xu