From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from mx1.redhat.com ([209.132.183.28]:51796 "EHLO mx1.redhat.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1761416AbdDSKlp (ORCPT ); Wed, 19 Apr 2017 06:41:45 -0400 Date: Wed, 19 Apr 2017 18:41:42 +0800 From: Eryu Guan Subject: Re: [PATCH 4/4] fstests: add generic test for file handles Message-ID: <20170419104142.GW8951@eguan.usersys.redhat.com> References: <1492539444-25938-1-git-send-email-amir73il@gmail.com> <1492539444-25938-5-git-send-email-amir73il@gmail.com> <20170419095533.GQ8951@eguan.usersys.redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: Sender: fstests-owner@vger.kernel.org To: Amir Goldstein Cc: Miklos Szeredi , Trond Myklebust , Jeff Layton , "J . Bruce Fields" , fstests , linux-unionfs@vger.kernel.org List-ID: On Wed, Apr 19, 2017 at 01:07:47PM +0300, Amir Goldstein wrote: ... > > So how about remove all the files and call "open_by_handle -c ..." > > before each test? e.g. (I use testdir and don't remove lost+found dir > > because extN needs it for fsck) > > > > Sure. Makes sense. > Going forward, as you might have guessed, I am going to add > _scratch_cycle_mount to the mix. > > > > testdir=$SCRATCH_MNT/testdir > > So I copied using SCRATCH_MNT from xfs/238, but I wonder if > I should go with thsi instead: > testdir=$TEST_DIR/$seq > > Maybe some fs is vulnerable to exporting handles from an ages fs > who knows... Any opinion? Hmm, I have no strong preference on this, seems TEST_DIR should be sufficient for the test and we can get rid of one test requirement. Thanks, Eryu