Return-Path: linux-nfs-owner@vger.kernel.org Received: from mail-vc0-f171.google.com ([209.85.220.171]:52826 "EHLO mail-vc0-f171.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753495AbaIIOiC (ORCPT ); Tue, 9 Sep 2014 10:38:02 -0400 Received: by mail-vc0-f171.google.com with SMTP id id10so17277739vcb.30 for ; Tue, 09 Sep 2014 07:38:00 -0700 (PDT) MIME-Version: 1.0 In-Reply-To: <20140909054929.GA12459@lst.de> References: <1408637375-11343-1-git-send-email-hch@lst.de> <1408637375-11343-4-git-send-email-hch@lst.de> <20140909054929.GA12459@lst.de> Date: Tue, 9 Sep 2014 07:38:00 -0700 Message-ID: Subject: Re: [PATCH 03/19] pnfs: force a layout commit when encountering busy segments during recall From: Trond Myklebust To: Christoph Hellwig Cc: Linux NFS Mailing List Content-Type: text/plain; charset=UTF-8 Sender: linux-nfs-owner@vger.kernel.org List-ID: On Mon, Sep 8, 2014 at 10:49 PM, Christoph Hellwig wrote: > On Mon, Sep 08, 2014 at 05:37:42PM -0700, Trond Myklebust wrote: >> Why wouldn't it make more sense to call pnfs_layoutcommit_inode() >> unconditionally before the call to pnfs_mark_matching_lsegs_invalid()? > > It would minimally reduce the latency, but otherwise be not very different. > The downside is that we'll now need two i_lock roundtrips per recall. > > But if this is your preference I can easily fix it up. Please do. It would make the code a little easier on the eye. -- Trond Myklebust Linux NFS client maintainer, PrimaryData trond.myklebust@primarydata.com