Return-Path: Received: from discipline.rit.edu ([129.21.6.207]:14855 "HELO discipline.rit.edu" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with SMTP id S932608AbdJaXpG (ORCPT ); Tue, 31 Oct 2017 19:45:06 -0400 From: Andrew W Elble To: Trond Myklebust Cc: "bfields@fieldses.org" , "linux-nfs@vger.kernel.org" Subject: Re: [PATCH 1/5] nfsd: Fix stateid races between OPEN and CLOSE References: <20171031000951.18294-1-trond.myklebust@primarydata.com> <20171031000951.18294-2-trond.myklebust@primarydata.com> <1509492271.3397.8.camel@primarydata.com> Date: Tue, 31 Oct 2017 19:45:06 -0400 In-Reply-To: <1509492271.3397.8.camel@primarydata.com> (Trond Myklebust's message of "Tue, 31 Oct 2017 23:24:33 +0000") Message-ID: MIME-Version: 1.0 Content-Type: text/plain Sender: linux-nfs-owner@vger.kernel.org List-ID: Trond Myklebust writes: > On Tue, 2017-10-31 at 18:33 -0400, Andrew W Elble wrote: >> Not directly related to this patch series, but: >> >> I keep staring at the failure path in nfsd4_process_open2() from the >> call to nfs4_get_vfs_file() and thinking that there's a missing state >> change to the stateid that's still hashed before the mutex is dropped >> and the call to release_open_stateid()? >> > > If the seqid==1, so that we know this OPEN op created that stateid, > then it probably should be unhashed and marked as closed, since then we > know it represents no state. Otherwise, AFAICS it should keep its > current state + seqid. > > Do you want to send a patch, or should I update this patch series? Such > a fix probably does want to be a stable patch, since it will affect > clients that expect compliance with RFC5661/RFC7530. Please feel free to go ahead and update this series. Thanks, Andy -- Andrew W. Elble aweits@discipline.rit.edu Infrastructure Engineer, Communications Technical Lead Rochester Institute of Technology PGP: BFAD 8461 4CCF DC95 DA2C B0EB 965B 082E 863E C912