From: Trond Myklebust Subject: RE: [PATCH 1/1] nfs: fix race between renewd, umount, and the state manager in V4.1 Date: Fri, 22 Jan 2010 15:21:20 -0500 Message-ID: <1264191680.30757.33.camel@localhost> References: <1264135871-13083-1-git-send-email-batsakis@netapp.com> <1264186039.30757.7.camel@localhost> <1264189298.30757.16.camel@localhost> <1264191084.30757.28.camel@localhost> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Cc: linux-nfs@vger.kernel.org To: "Batsakis, Alexandros" Return-path: Received: from mail-out1.uio.no ([129.240.10.57]:40501 "EHLO mail-out1.uio.no" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753376Ab0AVUVY (ORCPT ); Fri, 22 Jan 2010 15:21:24 -0500 In-Reply-To: <1264191084.30757.28.camel@localhost> Sender: linux-nfs-owner@vger.kernel.org List-ID: On Fri, 2010-01-22 at 15:11 -0500, Trond Myklebust wrote: > Then the correct thing to do is to fix the renewd kill process. Once > that is done, there should be no reason for any sequence_done calls to > be pending (see reason above). IOW: we should add a call to atomic_inc_not_zero(&clp->cl_count) in nfs4_proc_async_renew()/nfs41_proc_async_sequence(), and have them fail if that attempt to grab a reference fails. Then we should add a call to nfs_put_client() in their .rpc_release method. Trond