lists.openwall.net   lists  /  announce  owl-users  owl-dev  john-users  john-dev  passwdqc-users  yescrypt  popa3d-users  /  oss-security  kernel-hardening  musl  sabotage  tlsify  passwords  /  crypt-dev  xvendor  /  Bugtraq  Full-Disclosure  linux-kernel  linux-netdev  linux-ext4  linux-hardening  linux-cve-announce  PHC 
Open Source and information security mailing list archives
 
Hash Suite for Android: free password hash cracker in your pocket
[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Date:	Mon, 5 Aug 2013 18:18:03 +0000
From:	"Myklebust, Trond" <Trond.Myklebust@...app.com>
To:	Jeff Layton <jlayton@...hat.com>
CC:	Nix <nix@...eri.org.uk>,
	Toralf Förster <toralf.foerster@....de>,
	Oleg Nesterov <oleg@...hat.com>,
	NFS list <linux-nfs@...r.kernel.org>,
	Linux Kernel Mailing List <linux-kernel@...r.kernel.org>,
	"dhowells@...hat.com" <dhowells@...hat.com>
Subject: Re: [3.10.4] NFS locking panic, plus persisting NFS shutdown panic
 from 3.9.*

On Mon, 2013-08-05 at 13:37 -0400, Jeff Layton wrote:
> On Mon, 5 Aug 2013 16:15:01 +0000
> "Myklebust, Trond" <Trond.Myklebust@...app.com> wrote:
> 
> > From 3c50ba80105464a28d456d9a1e0f1d81d4af92a8 Mon Sep 17 00:00:00 2001
> > From: Trond Myklebust <Trond.Myklebust@...app.com>
> > Date: Mon, 5 Aug 2013 12:06:12 -0400
> > Subject: [PATCH] LOCKD: Don't call utsname()->nodename from
> >  nlmclnt_setlockargs
> > MIME-Version: 1.0
> > Content-Type: text/plain; charset=UTF-8
> > Content-Transfer-Encoding: 8bit
> > 
> > Firstly, nlmclnt_setlockargs can be called from a reclaimer thread, in
> > which case we're in entirely the wrong namespace.
> > Secondly, commit 8aac62706adaaf0fab02c4327761561c8bda9448 (move
> > exit_task_namespaces() outside of exit_notify()) now means that
> > exit_task_work() is called after exit_task_namespaces(), which
> > triggers an Oops when we're freeing up the locks.
> > 
> > Signed-off-by: Trond Myklebust <Trond.Myklebust@...app.com>
> > Cc: Toralf Förster <toralf.foerster@....de>
> > Cc: Oleg Nesterov <oleg@...hat.com>
> > Cc: Nix <nix@...eri.org.uk>
> > Cc: Jeff Layton <jlayton@...hat.com>
> > ---
> >  fs/lockd/clntproc.c | 5 +++--
> >  1 file changed, 3 insertions(+), 2 deletions(-)
> > 
> > diff --git a/fs/lockd/clntproc.c b/fs/lockd/clntproc.c
> > index 9760ecb..acd3947 100644
> > --- a/fs/lockd/clntproc.c
> > +++ b/fs/lockd/clntproc.c
> > @@ -125,14 +125,15 @@ static void nlmclnt_setlockargs(struct nlm_rqst *req, struct file_lock *fl)
> >  {
> >  	struct nlm_args	*argp = &req->a_args;
> >  	struct nlm_lock	*lock = &argp->lock;
> > +	char *nodename = req->a_host->h_rpcclnt->cl_nodename;
> >  
> >  	nlmclnt_next_cookie(&argp->cookie);
> >  	memcpy(&lock->fh, NFS_FH(file_inode(fl->fl_file)), sizeof(struct nfs_fh));
> > -	lock->caller  = utsname()->nodename;
> > +	lock->caller  = nodename;
> >  	lock->oh.data = req->a_owner;
> >  	lock->oh.len  = snprintf(req->a_owner, sizeof(req->a_owner), "%u@%s",
> >  				(unsigned int)fl->fl_u.nfs_fl.owner->pid,
> > -				utsname()->nodename);
> > +				nodename);
> >  	lock->svid = fl->fl_u.nfs_fl.owner->pid;
> >  	lock->fl.fl_start = fl->fl_start;
> >  	lock->fl.fl_end = fl->fl_end;
> 
> Looks good to me...
> 
> Reviewed-by: Jeff Layton <jlayton@...hat.com>
> 
> Trond, any thoughts on the other oops that Nix posted? The issue there
> seems to be that we're trying to do the pathwalk to the rpcbind unix
> socket from exit_task_work(), but that's happening after we've already
> called exit_fs().
> 
> The trivial answer seems to be to simply call exit_task_work() before
> exit_fs() there, but it seems like we ought to be doing the upcall to
> rpcbind in a mount namespace from which we know we can reach the
> socket...

Isn't it enough to just do the same thing as we did for gss proxy? i.e.
set the RPC_CLNT_CREATE_NO_IDLE_TIMEOUT flag.

See attachment.
-- 
Trond Myklebust
Linux NFS client maintainer

NetApp
Trond.Myklebust@...app.com
www.netapp.com

View attachment "0001-SUNRPC-Don-t-auto-disconnect-from-the-local-rpcbind-.patch" of type "text/x-patch" (1452 bytes)

Powered by blists - more mailing lists

Powered by Openwall GNU/*/Linux Powered by OpenVZ