Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1946602AbWKAGLK (ORCPT ); Wed, 1 Nov 2006 01:11:10 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1946547AbWKAGLK (ORCPT ); Wed, 1 Nov 2006 01:11:10 -0500 Received: from 1wt.eu ([62.212.114.60]:60932 "EHLO 1wt.eu") by vger.kernel.org with ESMTP id S1946602AbWKAGLI (ORCPT ); Wed, 1 Nov 2006 01:11:08 -0500 Date: Wed, 1 Nov 2006 08:11:11 +0100 From: Willy Tarreau To: Chris Wright Cc: linux-kernel@vger.kernel.org, NeilBrown Subject: Re: [PATCH 30/61] knfsd: Fix race that can disable NFS server. Message-ID: <20061101071111.GB543@1wt.eu> References: <20061101053340.305569000@sous-sol.org> <20061101054028.568862000@sous-sol.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20061101054028.568862000@sous-sol.org> User-Agent: Mutt/1.5.11 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 2140 Lines: 57 On Tue, Oct 31, 2006 at 09:34:10PM -0800, Chris Wright wrote: > -stable review patch. If anyone has any objections, please let us know. > ------------------ > > From: NeilBrown > > This is a long standing bug that seems to have only recently become > apparent, presumably due to increasing use of NFS over TCP - many > distros seem to be making it the default. > > The SK_CONN bit gets set when a listening socket may be ready > for an accept, just as SK_DATA is set when data may be available. > > It is entirely possible for svc_tcp_accept to be called with neither > of these set. It doesn't happen often but there is a small race in > svc_sock_enqueue as SK_CONN and SK_DATA are tested outside the > spin_lock. They could be cleared immediately after the test and > before the lock is gained. > > This normally shouldn't be a problem. The sockets are non-blocking so > trying to read() or accept() when ther is nothing to do is not a problem. > > However: svc_tcp_recvfrom makes the decision "Should I accept() or > should I read()" based on whether SK_CONN is set or not. This usually > works but is not safe. The decision should be based on whether it is > a TCP_LISTEN socket or a TCP_CONNECTED socket. > > > Signed-off-by: Neil Brown > Signed-off-by: Greg Kroah-Hartman > Signed-off-by: Chris Wright > > --- > net/sunrpc/svcsock.c | 2 +- > 1 file changed, 1 insertion(+), 1 deletion(-) > > --- linux-2.6.18.1.orig/net/sunrpc/svcsock.c > +++ linux-2.6.18.1/net/sunrpc/svcsock.c > @@ -902,7 +902,7 @@ svc_tcp_recvfrom(struct svc_rqst *rqstp) > return 0; > } > > - if (test_bit(SK_CONN, &svsk->sk_flags)) { > + if (svsk->sk_sk->sk_state == TCP_LISTEN) { > svc_tcp_accept(svsk); > svc_sock_received(svsk); > return 0; This one seems valid for 2.4 too. Neil, do you confirm ? Willy - To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/