Return-Path: Received: from verein.lst.de ([213.95.11.211]:42254 "EHLO newverein.lst.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1030316AbbD1SLP (ORCPT ); Tue, 28 Apr 2015 14:11:15 -0400 Date: Tue, 28 Apr 2015 20:11:13 +0200 From: Christoph Hellwig To: Trond Myklebust Cc: Christoph Hellwig , "J. Bruce Fields" , Jeff Layton , Linux NFS Mailing List , Sachin Bhamare , Jeff Layton Subject: Re: [PATCH 2/2] nfsd: fix pNFS return on close semantics Message-ID: <20150428181113.GA6183@lst.de> References: <1430139014-28013-1-git-send-email-hch@lst.de> <1430139014-28013-3-git-send-email-hch@lst.de> <20150427203943.GI4083@fieldses.org> <20150427193008.1195d7dd@tlielax.poochiereds.net> <20150428081456.GA21413@lst.de> <20150428160353.GN4083@fieldses.org> <20150428174934.GA5887@lst.de> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii In-Reply-To: Sender: linux-nfs-owner@vger.kernel.org List-ID: On Tue, Apr 28, 2015 at 01:54:31PM -0400, Trond Myklebust wrote: > The main reason why you may want these patches is that without > return-on-close, you typically can end up with layouts being cached in > knfsd without the filesystem being pinned. When you then unmount the > filesystem... Each layout has a reference on the layout state, and each layout state has a struct file reference.