From mboxrd@z Thu Jan 1 00:00:00 1970 From: Stanislav Kinsbursky Subject: Re: [PATCH 0/5] NFS: create blocklayout pipe per network namesapce context Date: Tue, 10 Jan 2012 16:58:27 +0400 Message-ID: <4F0C35F3.7060508@parallels.com> References: <20111129090754.20730.64567.stgit@localhost6.localdomain6> <1325285705.10984.4.camel@lade.trondhjem.org> <1325797111.11084.22.camel@lade.trondhjem.org> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: "linux-nfs@vger.kernel.org" , Pavel Emelianov , "neilb@suse.de" , "netdev@vger.kernel.org" , "linux-kernel@vger.kernel.org" , James Bottomley , "bfields@fieldses.org" , "davem@davemloft.net" , "devel@openvz.org" To: Trond Myklebust Return-path: In-Reply-To: <1325797111.11084.22.camel@lade.trondhjem.org> Sender: linux-kernel-owner@vger.kernel.org List-Id: netdev.vger.kernel.org 06.01.2012 00:58, Trond Myklebust =D0=BF=D0=B8=D1=88=D0=B5=D1=82: > On Fri, 2011-12-30 at 17:55 -0500, Trond Myklebust wrote: >> On Tue, 2011-11-29 at 13:10 +0300, Stanislav Kinsbursky wrote: >>> This patch set was created in context of clone of git >>> branch: git://git.linux-nfs.org/projects/trondmy/nfs-2.6.git. >>> tag: v3.1 >>> >>> This patch set depends on previous patch sets titled: >>> 1) "SUNRPC: initial part of making pipefs work in net ns" >>> 2) "SUNPRC: cleanup PipeFS for network-namespace-aware users" >>> 3) "SUNRPC: make RPC clients use network-namespace-aware PipeFS rou= tines" >>> 4) "NFS: create clients and IDMAP pipes per network namespace" >>> >>> This patch set is the final part of making SUNRPC PipeFS and it's u= sers work in >>> network namespace context. >>> >>> The following series consists of: >>> >>> --- >>> >>> Stanislav Kinsbursky (5): >>> NFS: handle blocklayout pipe PipeFS dentry by network namesp= ace aware routines >>> NFS: blocklayout pipe creation per network namespace context= introduced >>> NFS: blocklayout PipeFS notifier introduced >>> NFS: remove RPC PipeFS mount point reference from blocklayou= t routines >>> SUNRPC: kernel PipeFS mount point creation routines removed >>> >>> >>> fs/nfs/blocklayout/blocklayout.c | 154 ++++++++++++++++++++++= ++++++------- >>> fs/nfs/blocklayout/blocklayout.h | 3 - >>> fs/nfs/blocklayout/blocklayoutdev.c | 5 + >>> fs/nfs/blocklayout/blocklayoutdm.c | 7 +- >>> fs/nfs/inode.c | 1 >>> fs/nfs/netns.h | 1 >>> include/linux/sunrpc/rpc_pipe_fs.h | 2 >>> net/sunrpc/rpc_pipe.c | 21 ----- >>> 8 files changed, 137 insertions(+), 57 deletions(-) >> >> >> These patches need rebasing in order to apply on top of 3.2-rc7... > > OK. Further testing seems to indicate that we're going to have to > postpone merging these patches until the 3.4 merge window. > > The problems are twofold: > > As the patches stand now in the linux-next tree, they can (and > occasionally do) Oops on unmount. The reason was that I rejected the > PipeFS notifier patch for the idmapper (due to the reported problem o= f > nfs_idmap_init/nfs_idmap_quit being undefined when CONFIG_NFS_V4 is > undefined), and the fact that it is missing causes the unmount at the > end of our tests to hit the BUG() in fs/dcache.c:905. This suggests t= hat > we will have the same problem with the pNFS block layout driver, sinc= e I > still haven't received a rebased update of the 5 'create blocklayout > pipe per network namesapce context' patches. > Hello, Trond. I've resend the patch set (rebased with fix for nfs_idmap_init/nfs_idma= p_quit). > The second problem that was highlighted was the fact that as they sta= nd > today, these patchsets do not allow for bisection. When we hit the Oo= ps, > I had Bryan try to bisect where the problem arose. He ended up pointi= ng > at the patch "SUNRPC: handle RPC client pipefs dentries by network > namespace aware routine", which is indeed the cause, but which is one= of > the _dependencies_ for all the PipeFS notifier patches that fix the > problem. > I'm confused here. Does this means, that I have to fix patch "SUNRPC: h= andle RPC=20 client pipefs dentries by network namespace aware routine" to make it a= ble to=20 bisect? --=20 Best regards, Stanislav Kinsbursky