From mboxrd@z Thu Jan 1 00:00:00 1970 From: ebiederm-aS9lmoZGLiVWk0Htik3J/w@public.gmane.org (Eric W. Biederman) Subject: Re: [RFC][PATCH] sunrpc: fix oops in rpc_create() when the mount namespace is unshared Date: Tue, 09 Sep 2008 13:08:29 -0700 Message-ID: References: <48C52B29.4020204@fr.ibm.com> <20080909124311.GA10053@us.ibm.com> <20080909152952.GA21207@us.ibm.com> <869FF00C-4DAF-4D19-80AF-9230617A9223@oracle.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: In-Reply-To: <869FF00C-4DAF-4D19-80AF-9230617A9223-QHcLZuEGTsvQT0dZR+AlfA@public.gmane.org> (Chuck Lever's message of "Tue, 9 Sep 2008 15:00:55 -0400") Sender: linux-nfs-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org To: Chuck Lever Cc: "Serge E. Hallyn" , Cedric Le Goater , Andrew Morton , Trond Myklebust , Linux Kernel Mailing List , Linux Containers , linux-nfs-u79uwXL29TY76Z2rM5mHXA@public.gmane.org List-Id: containers.vger.kernel.org Chuck Lever writes: > On Sep 9, 2008, at Sep 9, 2008, 2:20 PM, ebiederm-aS9lmoZGLiVWk0Htik3J/w@public.gmane.org wrote: >> Chuck Lever writes: >> >>> If the upper layers are responsible for providing the utsname, you will need >>> to >>> fix up lockd and the NFS server's callback client too, at least. >> >> Actually looking at the code. It looks like a proper fix may be even simpler. >> Why do we have both clnt->cl_server and clnt->cl_nodename? Or is cl_server >> the other side of the connection? > > cl_server is the server-side. cl_nodename is the "machine name" of the client. Thanks, Darn. Looks like we need to capture both names at the same or a similar point. I'm wondering if we need to capture a network namespace as well. >>> I don't like the idea of an oops in here. Instead, (for now) it should warn >>> and >>> fail to create the client, IMO. >> >> Which is interesting when the problem happens during NFS unmount. Although >> frankly it could fail anyway. >> >> It seems strange that we are creating a client during unmount anyway. > > It's worth investigating. Just enable RPC tracing (/usr/sbin/rpcdebug -m rpc -s > all) before shutting down the client. > > NFS unmounting is historically difficult because during a system shutdown, > unmounting is the last thing to occur, and usually happens after most system > services (such as the portmapper service) have become unavailable. That's fine > for local file systems, but it makes for a rather dodgy situation for NFS. Interesting. Eric -- To unsubscribe from this list: send the line "unsubscribe linux-nfs" in the body of a message to majordomo-u79uwXL29TY76Z2rM5mHXA@public.gmane.org More majordomo info at http://vger.kernel.org/majordomo-info.html