From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from ipmail06.adl2.internode.on.net ([150.101.137.129]) by bombadil.infradead.org with esmtp (Exim 4.80.1 #2 (Red Hat Linux)) id 1ZSX4h-00030k-9d for linux-mtd@lists.infradead.org; Thu, 20 Aug 2015 21:08:40 +0000 Date: Fri, 21 Aug 2015 07:08:12 +1000 From: Dave Chinner To: Andreas =?iso-8859-1?Q?Gr=FCnbacher?= Cc: Dongsheng Yang , fstests@vger.kernel.org, linux-mtd@lists.infradead.org, richard.weinberger@gmail.com Subject: Re: [PATCH] fstests: link .out to correct output when we set USE_ATTR_SECURE=yes Message-ID: <20150820210812.GO3902@dastard> References: <55D52473.7050209@cn.fujitsu.com> <1440032464-17626-1-git-send-email-yangds.fnst@cn.fujitsu.com> MIME-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Disposition: inline Content-Transfer-Encoding: 8bit In-Reply-To: List-Id: Linux MTD discussion mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , On Thu, Aug 20, 2015 at 08:18:28AM +0200, Andreas Grünbacher wrote: > Dongsheng, > > 2015-08-20 3:01 GMT+02:00 Dongsheng Yang : > > > > When we set USE_ATTR_SECURE to yes or no, the expected outputs of generic/062 > > would be different. So we need to link the .out to different file. > > Can the tests for the different namespaces please be separated from > each other here (see below)? I certainly agree with this approach ;) > > diff --git a/tests/generic/062 b/tests/generic/062 > index 194b638..9dd4498 100755 > --- a/tests/generic/062 > +++ b/tests/generic/062 > @@ -142,6 +142,7 @@ for nsp in $ATTR_MODES; do > echo "*** final list (strings, type=$inode, nsp=$nsp)" > getfattr -m '.' -e hex $SCRATCH_MNT/$inode > > + # FIXME: Remove all remaining xattrs > done > done > > Next, can the security namespace tests be put in their own test case > so that we don't need to switch between two different .out files? This > could be done by setting ATTR_MODES and then invoking > tests/generic/062 to avoid duplicating the entire test script. Just not the suggested mechanism :) We can't execute one test from another - that will lead to all sorts of bad juju occurring with results and timing and so on because they are all based on $0. The correct way to do this is to factor the test internals into a common file, then wrap the common functions with a new test file. e.g. see _test_generic_punch(), which started off as corner case testing the XFS_IOC_ZERORANGE ioctl in a single test, Then got extended to testing hole punching in a new test, then got factored into common functions in common/punch, and it is now called by 14 different tests.... Cheers, Dave. -- Dave Chinner david@fromorbit.com