From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1760144Ab2EWN54 (ORCPT ); Wed, 23 May 2012 09:57:56 -0400 Received: from mx2.netapp.com ([216.240.18.37]:6790 "EHLO mx2.netapp.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751051Ab2EWN5u (ORCPT ); Wed, 23 May 2012 09:57:50 -0400 X-IronPort-AV: E=Sophos;i="4.75,645,1330934400"; d="scan'208";a="649770794" From: "Myklebust, Trond" To: Stanislav Kinsbursky CC: "linux-nfs@vger.kernel.org" , "linux-kernel@vger.kernel.org" , "devel@openvz.org" Subject: Re: [PATCH] NFS: init client before declaration Thread-Topic: [PATCH] NFS: init client before declaration Thread-Index: AQHNOBgchN+45TnxmEuX3jdeWgHLc5bWU9EAgAAIiACAAAgtgIAABjUAgAAHbgCAAAb8AIAAQCGAgAD6xICAACkuAA== Date: Wed, 23 May 2012 13:57:44 +0000 Message-ID: <1337781466.3360.12.camel@lade.trondhjem.org> References: <20120522124018.338.20817.stgit@localhost.localdomain> <1337696984.4269.4.camel@lade.trondhjem.org> <1337698816.4269.15.camel@lade.trondhjem.org> <4FBBB0DB.6050708@parallels.com> <1337701904.4269.16.camel@lade.trondhjem.org> <4FBBBC4C.5010103@parallels.com> <1337705000.4269.23.camel@lade.trondhjem.org> <1337718771.4269.29.camel@lade.trondhjem.org> <4FBCCA4F.3020600@parallels.com> In-Reply-To: <4FBCCA4F.3020600@parallels.com> Accept-Language: en-US Content-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: x-originating-ip: [10.104.60.115] Content-Type: text/plain; charset="utf-8" Content-ID: <7D023B575BA06B488F54150793A81E7F@tahoe.netapp.com> MIME-Version: 1.0 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Transfer-Encoding: 8bit X-MIME-Autoconverted: from base64 to 8bit by nfs id q4NDw5u3024874 On Wed, 2012-05-23 at 15:30 +0400, Stanislav Kinsbursky wrote: > On 23.05.2012 00:32, Myklebust, Trond wrote: > > On Tue, 2012-05-22 at 12:43 -0400, Trond Myklebust wrote: > >> On Tue, 2012-05-22 at 20:18 +0400, Stanislav Kinsbursky wrote: > >>> We have another problem here. > >>> nfs4_init_client() will try to create pipe dentries prior to set of NFS_CS_READY > >>> to the client. And dentries will be created since semaphore is dropped and > >>> per-net superblock variable is initialized already. > >>> But __rpc_pipefs_event() relays on the fact, that no dentries present. > >>> Looks like the problem was introduced by me in aad9487c... > >>> So maybe we should not call "continue" instead "__rpc_pipefs_event()", when > >>> client becomes ready? > >>> Looks like this will allow us to handle such races. > >> > >> Let me rework this patch a bit... > > > > The following is ugly, but it should be demonstrably correct, and will > > ensure that __rpc_pipefs_event() will only be called for fully > > initialised nfs_clients... > > > > 8<--------------------------------------------------------------------- > > From 90c3b9fe9faeae32c8f629e8b6cbf5f50bb9b295 Mon Sep 17 00:00:00 2001 > > From: Trond Myklebust > > Date: Tue, 22 May 2012 16:22:50 -0400 > > Subject: [PATCH 1/2] NFSv4: Fix a race in the net namespace mount > > notification > > > > Since the struct nfs_client gets added to the global nfs_client_list > > before it is initialised, it is possible that rpc_pipefs_event can > > end up trying to create idmapper entries on such a thing. > > > > The solution is to have the mount notification wait for the > > initialisation of each nfs_client to complete, and then to > > skip any entries for which the it failed. > > > > Reported-by: Stanislav Kinsbursky > > Signed-off-by: Trond Myklebust > > --- > > fs/nfs/client.c | 10 ++++++++++ > > fs/nfs/idmap.c | 15 +++++++++++++++ > > fs/nfs/internal.h | 1 + > > 3 files changed, 26 insertions(+), 0 deletions(-) > > > > diff --git a/fs/nfs/client.c b/fs/nfs/client.c > > index 60f7e4e..d3c8553 100644 > > --- a/fs/nfs/client.c > > +++ b/fs/nfs/client.c > > @@ -583,6 +583,16 @@ found_client: > > return clp; > > } > > > > +static bool nfs_client_ready(const struct nfs_client *clp) > > +{ > > + return clp->cl_cons_state<= NFS_CS_READY; > > +} > > + > > +int nfs_wait_client_ready(const struct nfs_client *clp) > > +{ > > + return wait_event_killable(nfs_client_active_wq, nfs_client_ready(clp)); > > +} > > + > > /* > > * Mark a server as ready or failed > > */ > > diff --git a/fs/nfs/idmap.c b/fs/nfs/idmap.c > > index 3e8edbe..c0753c5 100644 > > --- a/fs/nfs/idmap.c > > +++ b/fs/nfs/idmap.c > > @@ -530,9 +530,24 @@ static struct nfs_client *nfs_get_client_for_event(struct net *net, int event) > > struct nfs_net *nn = net_generic(net, nfs_net_id); > > struct dentry *cl_dentry; > > struct nfs_client *clp; > > + int err; > > > > +restart: > > spin_lock(&nn->nfs_client_lock); > > list_for_each_entry(clp,&nn->nfs_client_list, cl_share_link) { > > + /* Wait for initialisation to finish */ > > + if (clp->cl_cons_state> NFS_CS_READY) { > > + atomic_inc(&clp->cl_count); > > + spin_unlock(&nn->nfs_client_lock); > > + err = nfs_wait_client_ready(clp); > > > What about NFSv4.1 ? > It's clients NFS_CS_READY status depends on session establishing RPC calls... > Which in turn can hung up pipefs mount call... The alternative then is to wait for cl_cons_state != NFS_CS_INITING. That shouldn't require any upcalls, and so shouldn't be able to deadlock. -- Trond Myklebust Linux NFS client maintainer NetApp Trond.Myklebust@netapp.com www.netapp.com {.n++%ݶw{.n+{G{ayʇڙ,jfhz_(階ݢj"mG?&~iOzv^m ?I